Datasets:

krvhrv

/

crawl-books

like 0

the methods ultimately used are given in the 1996 review by Newman and are also covered to some extent in the 1991 and 1996 reviews by Pettit. These reviews should be consulted for the details as to sources and methods.4–6 As a result of these endeavors, enough cGMP-grade material was produced from wild collections to provide a source for most of the clinical trials shown in Tables 8.1 and 8.2. The trials in the United Kingdom in the early 1990s were not performed using the batches made under NCI’s auspices but, instead, used batches provided by Pettit; however, subsequent trials have all used the NCI-sourced materials. To date, there have been reports in the literature from seven phase I and 16 phase II trials in which bryostatin 1 was used as monotherapy (Table 8.1). There have also been reports of three phase I trials in which bryostatin 1 was used in conjunction with vinblastine, cytosine arabinoside (AraC), and fludarabine, respectively, and one phase II trial in which paclitaxel was used with bryostatin 1 (Table 8.2). Initially, bryostatin 1 demonstrated some partial responses in phase I trials, with myalgia being the dose-limiting toxicity. Subsequent phase I and phase II trials with bryostatin as monotherapy, using a variety of treatment regimens in patients with carcinomas ranging from solid tumors (mel- anomas, renal and ovarian) to varied leukemias, did not give consistent patterns of responses whether measured as stable disease (SD), or partial (PR) or complete responses (CR). There would be an occasional CR or PR and some SDs, but no consistent responses over the patient population on the trial in the case of the phase IIs. However, when one inspects the patient populations from the aspect of the types of carcinomas, then from the knowledge that is now accruing as to dose-limiting toxicities and methods/timing of administration, there are potential reasons as to why monotherapy by bry- ostatin is not optimal except perhaps in a very selective cohort of leukemia patients (see earlier comments on choice of patients because of potential tumorigenesis in some carcinoma lines). However, when bryostatin is combined with another cytotoxin, such as the vinca alkaloids or nucleosides, and the carcinomas are leukemic in nature, then the response rates, even in phase I trials, begin to demonstrate that such mixed treatments may well be worth further investigation (cf. Table 8.2). Thus, with the combination of high levels of cytosine arabinoside (AraC) and low levels of bryostatin in patients with leukemias, in a population that included patients who had failed high dose AraC or “HiDaC” therapy, 5 of 23 patients presented with complete responses in a phase I trial.32 Similarly, patients with chronic lymphocytic leukemia (CLL) and Non-Hodgkins lymphoma (NHL) treated with fludarabine and bryostatin, were reported to show close to 50% “objective responses” in the trial report.33 Finally, in a phase II trial of bryostatin and paclitaxel, 7 of 11 patients with non-small cell lung cancer demonstrated positive responses (PR/SD), but no CRs.34 At present (January 2005), there are three phase I and two phase II trials underway (see data from the NCI clinical trials Web site, http://clinicaltrials.nci.nih.gov), and in every case, these are combination studies with biologicals such as interleukin 2 or GM-CSF, nucleoside derivatives such as cladribine, or vincristine. These combinations are being tested against leukemias and lymphomas and against ovarian and prostate carcinomas. It is hoped, similar results to those demonstrated in Table 8.2 will be reported in due course. V. FUTURE SOURCES OF BRYOSTATINS A. AQUACULTURE The work reported to date has been from wild collections, made predominately in either the Gulf of California or in the Pacific Ocean off Palos Verdes, California. However, it became obvious, Copyright © 2005 CRC Press, LLC TABLE 8.1 Published Data Summary of Bryostatin-1 Monotherapy Dose Range Tumor No. Year Phase Schedule µg/M2 Type(s) Patients CR PR SD Side Effects Reference 1993 I 1 h infusion; repeated at 14 d; 3 cycles 5–65 Varied 19 0 0 0 Myalgia; phlebitis Prendiville et al.51 1993 I 1 h infusion; 7-, 14-, or 21 d cycle; varied number 25–50 Varied 35 0 2 0 Myalgia Philip et al.24 of cycles 1995 I 24 h infusion; repeated at 7 d; intervals for a 25–50 Varied 19 0 4 0 Myalgia; phlebitis Jayson et al.52 maximum of 8 cycles 1998 I 24–72 h infusions; repeated at 14 d for 4–42 cycles 12–180 NHL, CLL 29 0 0 11 Myalgia Varterasian et al.53 1998 Ib 0.5–24 h infusion; repeated at 8 and 15 d, then 12.5–25 Varied 12 0 0 0 Myalgia Grant et al.54 repeated at 28 d; varied number of cycles 1999 I 1 h infusion; repeated at 8 and 15 d, then repeated 20–57 Varied 22 0 0 4 Myalgia; photophobia Weitman et al.55 at 28 d; varied number of cycles 2002 I 4–14 d infusions; repeated at 14-21 d intervals; 2–3 8–24 Varied 37 0 0 6 Myalgia; fatigue Marshall et al.56 cycles 1998 II 1 h infusion; repeated at 8 and 15 d, then repeated 25 Melanoma 15 0 0 1 Myalgia Propper et al.57 at 28 d; 1–6 cycles 1999 II 1 h infusion; repeated at 8 and 15 d, then repeated 25 Melanoma 17 0 0 1 Myalgia Gonzalez et al.58 at 28 d; 0.67–8 cycles 2000 II 72 h infusion; repeated at 14 d; 4–14 cycles. 120 NHL, CLL 25 1 2 2 Myalgia Varterasian et al.59 2000 II 1 h infusion; repeated at 8 and 15 d, then repeated 25 RCC 30 1 1 0 Anemia; myalgia Pagliaro et al.60 at 28 d; 1–14 cycles 2001 II 24 h infusion; repeated at 8 and 15 d, then repeated 25–35 CRC 28 0 0 0 Myalgia Zonder et al.61 at 28 d; 2–4 cycles Copyright © 2005 CRC Press, LLC 2001 II 24 h infusion; repeated at 8 and 15 d, then repeated 25–40 Melanoma 49 0 1 14 Myalgia Bedikian et al.62 at 28 d; 72 h infusion; repeated at 15 d, then repeated at 28 d; 2–4 cycles 2001 II 24 h infusion; repeated weekly; 1–9 cycles 25 NHL 14 0 0 1 Myalgia Blackhall et al.63 2001 II 72 h infusion; at 14 d intervals; 2–8 cycles 9 MM 9 0 0 0 Myalgia Varterasian et al.64 2001 II 72 h infusion; at 14 d intervals; 3 cycles 120 Sarcoma; HNC 24 0 0 6 Myalgia; hyponatremia Brockstein et al.65 2002 II 24 h infusion; repeated at 8 and 15 d, then repeated 25 HNC 14 0 0 1 Myalgia (grade I) Pfister et al.66 at 28 d; 1–5 cycles 2002 II 24 h infusion (low dose); every 7 d; 72 h infusion 25–120 Melanoma 30 0 0 7 Myalgia Tozer et al.67 (high dose); every 14 d; 1–31 cycles 2003 II 1 h infusion; repeated at 8 and 15 d, then repeated 35–40 RCC 32 0 2 15 Myalgia Haas et al.68 at 28 d; 1–8 cycles. 2003 II 24 h infusion (low dose); 1, 8 and 15 d, repeated 25–120 Ovarian 54 0 1a 19 Myalgia Armstrong et al.69 every 28 d; 72 h infusion (high dose); every 14 d; 1–9 cycles. 2003 II 24 h infusion; repeated every 7 d; 1–9 cycles. 19–25 Ovarian 17 0 0 4a Myalgia Clamp et al.70 2003 II 24 h infusion; 8 and 15 d, then repeated at 28 d; 25 RCC 13 0 0 3a Myalgia Madhusudan et al.71 1–9 cycles. 2003 II 24 h infusion (low dose); repeated at 8 and 15 d, 25–120 Cervical 65 0 2 20 Myalgia Armstrong et a.l72 then repeated at 28 d; 72 h infusion (high dose); every 14 d; varied cycles. a Some question as to response level Copyright © 2005 CRC Press, LLC TABLE 8.2 Combination Therapy with Cytotoxic Drugs No. Year Phase Schedule Dose Range Tumor Type Patients CR PR SD Side Effects Reference 2001 I 24 h infusion and bolus of vincristine; 12.5–62.5 µg.M2 bryostatin; B-cell cancer 25 1 2 4 Myalgia; neuropathy Dowalti et al.73 dose escalation of bryostatin; 1–5 1.4 mg.M2 V cycles 2002 I 24 h infusion; days 1 and 11; AraC on 12.5–50 µg.M2 bryostatin; Leukemia 23 5 1a 0 Myalgia; neutropenia Cragg et al.32 days 2, 3, 9, 10. bryostatin dose 1–3 mg.M2 AraC escalation, fixed AraC; 1–6 cycles 2002 I 24 h infusion; FAra for days 2–6. 16–50 µg.M2 bryostatin; CLL; NHL 53 ? ? ? Neutropenia Roberts et al.33 Repeat at 28 d; or reverse addition 12.5–25 mg/M2 FAra order; 6–9+ cycles 2003 II 1 h infusion of paclitaxel on 1, 8, and 40–50 µg.M2 bryostatin; 90 NSCLC 11 0 2a 5 Myalgia Winegarden et al.34 15 d; 24 h infusion of bryostatin on mg.M2 paclitaxel 2, 9, 16 d; repeated on 28 d cycle; 1–4 cycles a Some question as to response level ? “non-defined objective responses” Copyright © 2005 CRC Press, LLC even as the initial trials were beginning in the early 1990s, that wild collections would not suffice and that production via chemical synthesis would probably not be viable. Thus, in the early 1990s, NCI, using the Small Business Innovation Research program, established a phase I and a subsequent phase II aquaculture program with a small company, CalBioMarine, with the aim of investigating the possibilities for both on-land and in-sea aquaculture for production of B. neritina under conditions that were not at the vagaries of nature. This series of projects was successful, culminating in the proof by CalBioMarine that the organism could be grown under both conditions and also that bryostatins could be isolated from the aquacultured animal in adequate quantities and at costs less than those incurred in wild collections.35 Concomitantly with the phase II aquaculture experiments, NCI also used the Small Business Innovation Research mechanism to investigate, with the Massachusetts company Aphios, the poten- tial for supercritical extraction of bryostatins from sources of B. neritina, both from wild collections and from some of the material from aquaculture. This was successful, and a demonstration of the potential for such methods was shown in the review by Newman,6 in which a comparison was made of the methods used for the earlier cGMP purification versus the Aphios technique. The method used for cGMP production involved a four-stage extraction–concentration process, and then a six-step process repeated 15 times to produce 18 g of bryostatin 1 from 13 metric tonnes of B. neritina. The overall process was multiweek in duration and used massive amounts of solvents. In contrast, the supercritical technique used basically carbon dioxide as the extraction medium on the wet animal mass, followed by supercritical chromatography, and the process was reduced to six simple operations performed within days, yet yielding material of similar purity. B. ACTUAL SOURCE OF THE BRYOSTATINS One interesting observation arising from the search for bryostatin sources was that, despite the ubiquitous occurrence of the nominal producing organism, the number of B. neritina colonies actually producing detectable bryostatin 1–3 levels was very low and geographically spread. One possible solution to this question came from the work of Haygood and her collaborators at the Scripps Institution of Oceanography, which has shown that the bryozoan is actually the host to a symbiotic organism that may well be the actual producer of the compound. In an elegant series of experiments, she and her colleagues have demonstrated the presence of a putative (polyketide synthase) PKS-I gene fragment in colonies that produce bryostatin, and that the fragment is absent in nonproducers.36 In addition, they demonstrated that there are subdivisions within B. neritina samples taken from the same sites, but at different depths. Thus, at depths greater than 9 m, (the D or deep type), bryostatins 1–3 and minor components are found (these are also known as producers of chemotype O for “octa-2,4-dienoatic chain”), whereas at less than 9 m (S or shallow type), only the minor derivatives are seen (chemotype M). The symbiotic organisms (Candidatus Endobugula neritina) isolated from each type differ in the mitochondrial protein CO I sequences by 8%, giving rise to the possibility that the bryozoans are also different taxonomically.37 Later reports demon- strated that the possibility of transferring this particular PKS fragment to other, more amenable microbes was being investigated with the goal of producing bryostatin by fermentative means. Recently Haygood38 suggested that the PKS system resembles that reported

by Piel39 for the Paederus beetle’s pseudomonal symbiont PKS that produces pederine, in that there are no acyl- transferase (AT) domains in the clusters. Further work is ongoing, using “remote” AT domains from another organism. Fuller details of the work have recently been published by Haygood and her collaborators.40 Further work by Haygood and collaborators has demonstrated that closely related symbiotic microbes are also found in Bugula simplex, and that these organisms both form a monophyletic grouping with E. sertula and appear to have PKS sequences that mimic the bryostatin PKS cluster.41 If successful, cultivation of the organism, or a surrogate with the bryostatin PKS system expressed, could probably solve any production problems if bryostatin becomes a viable drug. Copyright © 2005 CRC Press, LLC FIGURE 8.2 Structures of bryostatin analogs. C. CHEMICAL ANALOGS At present, the total synthesis of bryostatin 1 is not a feasible process for the production of this agent. However, synthesis of a simpler analog with comparable activity might well be a viable option. In 1986, Wender et al. analyzed the potential binding site of the phorbol esters on PKC as a guide to the design of simpler analogs of these agents.14 In 1988, this work was expanded by modeling bryostatin 1 onto the same binding site as a result of the initial results indicating that bryostatin 1 interacted with PKC.15 Subsequently, the modeling work was refined to produce three analogs that would maintain the putative binding sites at the oxygen atoms at C1 (ketone), C19 (hydroxyl), and C26 (hydroxyl) in the original molecule. These requirements gave rise to structures 1a, 2a, and 2b (Figure 8.2), which maintained the recognition features but removed a significant amount of the peripheral substituents. These molecules demonstrated nanomolar binding constants when measured in displacement assays of tritiated phorbol esters, with the figures being in the same general range as bryostatin 1, and 1a and 2a had activities in in vitro cell line assays close to those demonstrated by bryostatin 1 itself.42–45 Following on from these examples, modifications were made to the base structure 1a to introduce a second lactone 3 (Figure 8.2) which had 8 nM binding affinity and also inhibited P388 with an ED50 of 113 nM.46 Concomitantly, modifications were made to the base analog 1a, where different fatty acid esters were made (structures not shown). These, too, exhibited binding affinities for PKC isozymes in the 7–232 nM range, depending on the fatty acid used.47 To show the versatility of the base structure, recently Wender et al. published a simple modi- fication in which, by removal of a methyl group in the C26 side chain from compound 1a to produce Copyright © 2005 CRC Press, LLC 1b (Figure 8.2), the binding affinity to PKC was increased to the picomolar level48 and the compound demonstrated greater potency than bryostatin 1 in in vitro cell line assays. Improved syntheses of the molecule that could permit further refinements of the model, and that have the potential for greater overall yields, have been published.49, 50 VI. CONCLUSIONS The unique structural motifs that are part of the bryostatin molecule have now undergone a significant amount of testing in humans. Monotherapy does not appear to be an optimal use, but the potential for usage as part of a combination with other agents is now under intense study, with some promising results — particularly in leukemias — being reported at phase I levels. The problems of supply have been addressed both biologically and chemically. Although aquaculture-based production is feasible, the isolation of a putative producing microbial symbiont from larvae of animals that do produce bryostatins appears to have greater potential, though it has not as yet been grown. The production by total synthesis of simpler “bryologs” exhibiting similar or better levels of biological activity in both cell lines and against PKC isozymes holds promise for the development of more effective analogs. It is hoped, within the next few years the promise of this class of molecule (naturally occurring or synthetic) will be realized and a new series of antitumor drugs will reach the general patient population. REFERENCES 1. Pettit, G. R. et al., Antineoplastic components of marine animals, Nature, 227, 962, 1970. 2. Pettit, G. R. et al., Isolation and structure of bryostatin 1, J. Am. Chem. Soc., 104, 6846, 1982. 3. Suffness, M., Newman, D. J., and Snader, K. M., Discovery and development of antineoplastic agents from natural sources, in Bioorganic Marine Chemistry, Vol. 3, Scheuer, P. J., Ed., Springer-Verlag, Berlin, 1989, pp 131–168. 4. Pettit, G. R., The bryostatins, in Progress in the Chemistry of Organic Natural Products, Vol. 57, Hertz, W., Kirby, G. W., Steglich, W. and Tamm, C., Eds., Springer-Verlag, New York, 1991, pp 135–195. 5. Pettit, G. R., Progress in the discovery of biosynthetic anticancer drugs, J. Nat. Prod., 59, 812, 1996. 6. Newman, D. J., Bryostatin — from bryozan to cancer drug, in Bryozoans in Space and Time, Gordon, D. P., Smith, A. M. and Grant-Mackie, J. A., Eds., NIWA, Wellington, 1996, pp 9–17. 7. Mutter, R. and Wills, M., Chemistry and clinical biology of the bryostatins, Bioorg. Med. Chem., 8, 1841, 2000. 8. Pettit, G. R., Herald, C. L., and Hogan, F., Biosynthetic products for anticancer drug design and treatment: the bryostatins, in Anticancer Drug Development, Baguley, B. C., and Kerr, D. J., Eds., Academic Press, San Diego, CA, 2002, pp 203–235. 9. Hale, K. J. et al., The chemistry and biology of the bryostatin antitumour macrolides, Nat. Prod. Rep., 19, 413, 2002. 10. Pettit, G. R. et al., Relationship of Bugula neritina (Bryozoa) antineoplastic constituents to the yellow sponge Lyssodendoryx isodictyalis, Pure Appl. Chem., 58, 415, 1986. 11. Kamano, Y. et al., An improved source of bryostatin 10, Bugula neritina from the Gulf of Aomori, Japan, J. Nat. Prod., 58, 1868, 1995. 12. Hennings, H. et al., Bryostatin 1, an activator of protein kinase C, inhibits tumor promotion by phorbol esters in SENCAR mouse skin, Carcinogenesis, 8, 1343, 1987. 13. Blumberg, P. M. et al., The protein kinase C pathway in tumor promotion, Prog. Clin. Biol. Res., 298, 210, 1989. 14. Wender, P. A. et al., Analysis of the phorbol ester pharmacophore on protein kinase C as a guide to the rational design of new classes of analogs, Proc. Natl. Acad. Sci. USA, 83, 4214, 1986. Copyright © 2005 CRC Press, LLC 15. Wender, P. A. et al., Modeling of the bryostatins to the phorbol ester pharmacophore on protein kinase C, Proc. Natl. Acad. Sci. USA, 85, 7917, 1988. 16. Lee, J. et al., Protein kinase C ligands based on tetrahydrofuran templates containing a new set of phorbol ester pharmacophores, J. Med. Chem., 42, 4129, 1999. 17. Zhang, G. G. et al., Crystal-structure of the CYS2 activator-binding domain of protein kinase C-delta in complex with phorbol ester, Cell, 81, 917, 1995. 18. Hale, K. J. et al., Synthesis of a simplified bryostatin C-ring analog that binds to the CRD2 of human PKC-alpha and construction of a novel BC-analog by an unusual Julia olefination process, Org. Lett., 5, 499, 2003. 19. Goekjian, P. G. and Jirousek, M. R., Protein kinase C inhibitors as novel anticancer drugs, Expert Opin. Investig. Drugs, 10, 2117, 2001. 20. May, W. S. et al., Antineoplastic bryostatins are multipotential stimulators of human hematopoietic progenitor cells, Proc. Natl. Acad. Sci. USA, 84, 8483, 1987. 21. Hess, A. D. et al., Activation of human T lymphocytes by bryostatin, J. Immunol., 141, 3263, 1988. 22. Berkow, R. L. et al., In vivo administration of the anticancer agent bryostatin 1 activates platelets and neutrophils and modulates protein kinase C activity, Cancer Res., 53, 2810, 1993. 23. Esa, A. H. et al., Bryostatins trigger human polymorphonuclear neutrophil and monocyte oxidative metabolism: association with in vitro antineoplastic activity, Res. Immunol., 146, 351, 1995. 24. Philip, P. A. et al., Phase I study of bryostatin 1: assessment of interleukin 6 and tumor necrosis factor alpha induction in vivo. The Cancer Research Campaign Phase I Committee, J. Natl. Cancer Inst., 85, 1812, 1993. 25. Taylor, L. S. et al., Bryostatin and IFN-gamma synergize for the expression of the inducible nitric oxide synthetase gene and for nitric oxide production in murine macrophages, Cancer Res., 57, 2468, 1997. 26. Battle, T. E. and Frank, D. A., STAT1 mediates differentiation of chronic lymphocytic leukemia cells in response to Bryostatin 1, Blood, 102, 3016, 2003. 27. Hocevar, B., Burns, D. J., and Fields, A. P., Identification of protein kinase C (PKC) phosphorylation sites on human lamin B, J. Biol. Chem., 268, 7545, 1993. 28. Kageyama, M. et al., Synthesis of bryostatin 7, J. Am. Chem. Soc., 112, 7407, 1990. 29. Evans, D. A. et al., Total synthesis of bryostatin 2, J. Am. Chem. Soc., 121, 7540, 1999. 30. Ohmori, K. et al., Total synthesis of bryostatin 3, Angew. Chem. Int. Ed., 39, 2290, 2000. 31. Norcross, R. D. and Paterson, I., Total synthesis of bioactive marine macrolides, Chem. Rev., 95, 2041, 1995. 32. Cragg, L. et al., Phase I trial and correlative laboratory studies of bryostatin 1 (NSC 339555) and high-dose 1-β-D-arabinofuranosylcytosine in patients with refractory acute leukemia, Clinical Cancer Research, 8, 2123, 2002. 33. Roberts, J. D. et al., Phase I study of bryostatin-1 and fludarabine in patients with chronic lymphocytic leukemia and indolent non-Hodgkin’s lymphoma, Clinical Lymphoma, 3, 184, 2002. 34. Winegarden, J. D. et al., A phase II study of bryostatin-1 and paclitaxel in patients with advanced non-small cell lung cancer, Lung Cancer, 39, 191, 2003. 35. Mendola, D., Aquaculture of three phyla of marine invertebrates to yield bioactive metabolites: process development and economics, Biomol. Engin., 20, 441, 2003. 36. Davidson, S. K. et al., Evidence for the biosynthesis of bryostatins by the bacterial symbiont “Can- didatus Endobugula sertula” of the bryozoan Bugula neritina, Appl. Environ. Microbiol., 67, 4531, 2001. 37. Davidson, S. K. and Haygood, M. G., Identification of sibling species of the bryozoan Bugula neritina that produce different anticancer bryostatins and harbor distinct strains of the bacterial symbiont “Candidatus Endobugula sertula,” Biol. Bull., 196, 273, 1999. 38. Haygood, M. G., The role of a bacterial symbiont in the biosynthesis of bryostatins in the marine bryozoan Bugula neritina, Abs. Pap. 6th Int. Mar. Biotech. Conf., Abs. S5, 2003. 39. Piel, J., A polyketide synthase-peptide synthase gene cluster from an uncultured bacterial symbiont of Paederus beetles, Proc. Natl. Acad. Sci. USA, 99, 14002, 2002. 40. Hildebrand, M. et al., Approaches to identify, clone and express symbiont bioactive metabolite genes, Nat. Prod. Rep., 21, 122, 2004. Copyright © 2005 CRC Press, LLC 41. Lim, G. E. and Haygood, M. G., Candidatus Endobugula glebosa, a specific bacterial symbiont of the marine bryozoan Bugula simplex, Appl. Environ. Microbiol., 70, 4921, 2004. 42. Wender, P. A. et al., Synthesis and biological evaluation of fully synthetic bryostatin analogs, Tetra- hedron Lett., 39, 8625, 1998. 43. Wender, P. A. et al., Synthesis of the first members of a new class of biologically active bryostatin analogs, J. Am. Chem. Soc., 120, 4534, 1998. 44. Wender, P. A. et al., The design, computer modeling, solution structure, and biological evaluation of synthetic analogs of bryostatin 1, Proc. Natl. Acad. Sci. USA, 95, 6624, 1998. 45. Wender, P. A. et al., The rational design of potential chemotherapeutic agents: synthesis of bryostatin analogs, Med. Res. Revs., 19, 388, 1999. 46. Wender, P. A. and Lippa, B., Synthesis and biological evaluation of bryostatin analogs: the role of the A-ring, Tetrahedron Lett., 41, 1007, 2000. 47. Wender, P. A. and Hinkle, K. W., Synthesis and biological evaluation of a new class of bryostatin analogues: the role of the C20 substituent in protein kinase C binding, Tetrahedron Lett., 41, 6725, 2000. 48. Wender, P. A. et al., The practical synthesis of a novel and highly potent analog of bryostatin, J. Am. Chem. Soc., 124, 13648, 2002. 49. Wender, P. A., Mayweg, A. V. W., and VanDeusen, C. L., A concise, selective synthesis of the polyketide spacer domain of a potent bryostatin analog, Org. Lett., 5, 277, 2003. 50. Wender, P. A., Koehler, M. T. F., and Sendzik, M., A new synthetic approach to the C ring of known as well as novel bryostatin analogs, Org. Lett., 5, 4549, 2003. 51. Prendiville, J. et al., A phase I study of intravenous bryostatin 1 in patients with advanced cancer, Br. J. Cancer, 68, 418, 1993. 52. Jayson, G.

C. et al., A phase I trial of bryostatin 1 in patients with advanced malignancy using a 24 hour intravenous infusion, Br. J. Cancer, 72, 461, 1995. 53. Varterasian, M. L. et al., Phase I study of bryostatin 1 in patients with relapsed non-Hodgkin’s lymphoma and chronic lymphocytic leukemia, J. Clin. Oncol., 16, 56, 1998. 54. Grant, S. et al., Phase Ib trial of bryostatin 1 in patients with refractory malignancies, Clin. Cancer Res., 4, 611, 1998. 55. Weitman, S. et al., A phase I trial of bryostatin-1 in children with refractory solid tumors: a pediatric oncology group study, Clin. Cancer Res., 5, 2344, 1999. 56. Marshall, J. L. et al., Phase I study of prolonged infusion Bryostatin-1 in patients with advanced malignancies, Cancer Biol. Ther., 1, 409, 2002. 57. Propper, D. J. et al., A phase II study of bryostatin 1 in metastatic malignant melanoma, Br. J. Cancer, 78, 1337, 1998. 58. Gonzalez, R. et al., Treatment of patients with metastatic melanoma with bryostatin-1 — a phase II study, Melanoma Research, 9, 599, 1999. 59. Varterasian, M. L. et al., Phase II trial of bryostatin 1 in patients with relapsed low-grade non-Hodgkin’s lymphoma and chronic lymphocytic leukemia, Clin. Cancer Res., 6, 825, 2000. 60. Pagliaro, L. et al., A phase II trial of bryostatin-1 for patients with metastatic renal cell carcinoma, Cancer, 89, 615, 2000. 61. Zonder, J. A. et al., A phase II trial of bryostatin 1 in the treatment of metastatic colorectal cancer, Clin. Cancer Res., 7, 38, 2001. 62. Bedikian, A. Y. et al., Phase II evaluation of bryostatin-1 in metastatic melanoma, Melanoma Res., 11, 183, 2001. 63. Blackhall, F. H. et al., A phase II trial of bryostatin 1 in patients with non-Hodgkin’s lymphoma, Br. J. Cancer, 84, 465, 2001. 64. Varterasian, M. L. et al., Phase II study of bryostatin 1 in patients with relapsed multiple myeloma, Investig. New Drugs, 19, 245, 2001. 65. Brockstein, B. et al., Phase II studies of bryostatin-1 in patients with advanced sarcoma and advanced head and neck cancer, Investig. New Drugs, 19, 249, 2001. 66. Pfister, D. G. et al., A phase II trial of bryostatin-1 in patients with metastatic or recurrent squamous cell carcinoma of the head and neck, Investig. New Drugs, 20, 123, 2002. 67. Tozer, R. G. et al., A randomized phase II study of two schedules of bryostatin-1 (NSC339555) in patients with advanced malignant melanoma, Investig. New Drugs, 20, 407, 2002. Copyright © 2005 CRC Press, LLC 68. Haas, N. B. et al., Weekly bryostatin-1 in metastatic renal cell carcinoma: A phase II study, Clin. Cancer Res., 9, 109, 2003. 69. Armstrong, D. K. et al., A randomized phase II evaluation of bryostatin-1 (NSC #339555) in recurrent or persistent platinum-sensitive ovarian cancer: a Gynecologic Oncology Group Study, Investig. New Drugs, 21, 373, 2003. 70. Clamp, A. R. et al., A phase II trial of bryostatin-1 administered by weekly 24-hour infusion in recurrent epithelial ovarian carcinoma, Br. J. Cancer, 89, 1152, 2003. 71. Madhusudan, S. et al., A multicentre phase II trial of bryostatin-1 in patients with advanced renal cancer, Br. J. Cancer, 89, 1418, 2003. 72. Armstrong, D. K. et al., A randomized phase II evaluation of bryostatin-1 (NSC #339555) in persistent or recurrent squamous cell carcinoma of the cervix: a Gynecologic Oncology Group Study, Investig. New Drugs, 21, 453, 2003. 73. Dowalti, A. et al., Phase I trial of combination bryostatin-1 and vincristine in B-cell malignancies: final report, Clin. Cancer Res., 7 supplement 3814s, Abs 800, 2001. Copyright © 2005 CRC Press, LLC 9 The Isolation, Characterization, and Development of a Novel Class of Potent Antimitotic Macrocyclic Depsipeptides: The Cryptophycins Rima S. Al-awar and Chuan Shih CONTENTS I. Introduction II. Isolation and Characterization of the Cryptophycins III. Mechanism of Action of the Cryptophycins A. Cellular Mechanism of Action of Cryptophycin 52 (LY355703) B. Effects of Cryptophycin 52 (LY355703) on Microtubule Polymerization In Vitro IV. Synthesis of Cryptophycin 52 (LY355703) A. Retrosynthetic Analysis and Final Assembly of the Fragments B. Preparation of Fragments (A–D) V. SARs of the Cryptophycins A. SAR of the β-Epoxide Region of Fragment A B. SAR of the Phenyl Group of Fragment A C. SAR of the D-Chlorotyrosine of Fragment B D. SAR of the C/D Ester Bond and the C6, C7 Positions VI. Comparison of Cryptophycins 52, 55, and 55-glycinate VII. Cryptophycin 52 (LY355703) A. In Vivo Antitumor Activity of Cryptophycin 52 (LY355703) B. Clinical Evaluation of Cryptophycin 52 (LY355703) VIII. Conclusions Acknowledgments References I. INTRODUCTION The cryptophycins are potent antitumor depsipeptides first isolated from terrestrial blue-green algae. Initial cellular mechanism of action studies showed them to be associated with the inhibition of Copyright © 2005 CRC Press, LLC mitotic spindle function. This characterization puts the cryptophycins into one of the clinically very important classes of anticancer agents, such as the vinca alkaloids (vincristine, vinblastine, vinorel- bine) and the taxanes (taxol and taxotere), that act primarily on the cellular microtubule structure and function. Extensive preclinical studies have demonstrated that the cryptophycins are highly active against a broad spectrum of murine solid tumors and human tumor xenografts in vivo. One very unique pharmacological property of the cryptophycins is that they do not serve as substrates for either Pgp or the MRP multidrug-resistant efflux pumps and are highly active against the resistant tumors that express multiple-drug-resistant phenotypes both in vitro and in vivo. These results indicate that the cryptophycins may be highly active against human solid tumors that are resistant to the taxanes or the vincas and may offer alternative efficacy or survival advantages over the current therapies. Here we summarize the collaborative structure–activity relationship (SAR) studies conducted among three institutions, Lilly Research Laboratories, University of Hawaii, and Wayne State University, in their identification of cryptophycin 52 (LY355703) as the first-generation cryptophycin to undergo clinical evaluation. II. ISOLATION AND CHARACTERIZATION OF THE CRYPTOPHYCINS The cryptophycins are peptolides with a 16-membered macrolide ring structure. The macrocyclic ring is composed of two ester linkages (Fragment A/D and Fragment C/D), two amide linkages (Fragment A/B and Fragment B/C), and seven asymmetric centers (Figure 9.1). Cryptophycin A or 1 was first isolated from the Nostoc sp. ATCC 537981 by researchers at Merck in the early 1990s and was found to be very active against strains of Cryptococcus. The development of cryptophycin A as a potential antifungal agent was unsuccessful mainly because of the toxicity of the agent. During the same period of time, in an effort to identify novel antitumor agents from blue-green algae (cyanobacteria), researchers at the University of Hawaii led by Professor Richard Moore discovered that the lipophilic extract of Nostoc sp. GSV2242 was highly cytotoxic to human nasopharyngeal carcinoma (KB) and human colorectal adenocarcinoma (LoVo) cells. More impor- tantly, this lipophylic extract was found to be highly selective toward solid tumor cells versus leukemia cells and normal fibroblasts in the disk diffusion soft agar colony formation assay developed by Professor Thomas Corbett at Wayne State University.3 To follow up on this activity, a total of 26 naturally occurring secondary metabolites were isolated from the Nostoc sp. GSV224, with cryptophycin 1 accounting for most of the cytotoxic activity in the crude extract. Through rigorous structure elucidation and determination efforts, the University of Hawaii team established O H2N Cl OH OH O HO O Fragment A Fragment B 19 18 16 O 17 O O 1 O HN 10 O Cl 7 3 O 6 N O O O R H OH C O OH C ptophycin A (Cryptophycin 1): R = H HO NH ry 2 Cryptophycin 52 (LY355703): R = CH H 3 Fragment D Fragment C FIGURE 9.1 Chemical structure of crytophycin 1 and its four fragments (A–D). Copyright © 2005 CRC Press, LLC the absolute stereochemistry all the asymmetric centers of cryptophycin 1 and confirmed that cryptophycin 1 was identical to cryptophycin A, reported on earlier by the Merck group.4 The antitumor activity of cryptophycin 1 was then rapidly followed up in vivo by the Wayne State University group (Professor Corbett), using various solid-tumor models developed either in syngeneic or severe combined immunodeficiency (SCID) mice. It was quickly determined that cryptophycin 1 was highly active against a number of murine solid tumors and human tumor xenografts when given by intravenous injection. In contrast, the compound was found to be significantly less active when administered by intraperitoneal or oral routes. The lack of activity via these routes was believed to mainly be a result of poorer metabolic stability. Excellent tumor growth inhibition and antitumor activity were observed with cryptophycin 1 in a number of solid tumors, but more importantly, it was highly active against the murine mammary 16/C tumor that is resistant to paclitaxel because of the overexpression of the multiple-drug-resistant protein Pgp.5 This renders the cryptophycins some of the very few new antitubulin agents that do not serve as substrates for the multiple-drug-resistant protein and, hence, may offer significant advantage for the treatment of resistant human solid tumors. The discovery of the exciting antitumor activity of cryptophycin 1 prompted the Lilly/Hawaii/Wayne State team to develop a unified strategy not only for producing a large number of analogs for the SAR and the selection of a clinical candidate from the series but also for developing technologies that can be used for the potential commercial-scale production of the cryptophycin class of compounds. Critical analyses revealed that a large-scale bioproduction of the cryptophycins from cultured blue-green algae was not only technically very challenging but also economically unattractive. For example, cryptophycin 1 can only be obtained in an estimated 0.4% yield from the dry algae mass after an extensive high-performance liquid chromatography purification process. The bioproduction approach also offered limited chemical diversity for the SAR studies. A total synthetic approach was eventually elected as the major strategy for the chemical investigation of the cryptophycin series of compounds. It offers several advantages over the bioproduction approach, including chemical diversity, synthetic maneuverability, enabling technologies, and last but not least, speed for advancing the project. The majority of the SARs described below were conducted using materials prepared by total synthetic approaches developed jointly between the Lilly Research Laboratories and the University of Hawaii teams. The clinical candidate compound, cryptophycin 52 (LY355703), was prepared via a convergent synthetic approach that was originally developed by Professors Moore and Tius’ groups at the University of Hawaii and was later modified by the Chemical Process Research and Development group at Lilly Research Laboratories, led by Dr. Michael Martinelli.6 The 30-step convergent synthetic sequence of the cryptophycins illustrates both the complexity and the diversity of such an approach for the preparation of this novel chemical series. III. MECHANISM OF ACTION OF THE CRYPTOPHYCINS The cryptophycins are potent antiproliferative and cytotoxic agents that act against a variety of human tumor cell lines derived from both hematopoietic and solid tumors. The mechanism of action behind the potent cytotoxic effects induced by the cryptophycins appears to be associated with inhibition of mitotic spindle function of the cells.7 This inhibition of mitosis is accompanied by accumulation of cells at the metaphase–anaphase transition, followed by spindle disorganization and fragmentation and, finally, apoptosis or cell death.8 Biochemically, it was found that at low concentrations, the cryptophycins interact with microtubules to significantly dampen their dynamic characteristics. At higher concentrations, the process of microtubule polymerization was effectively inhibited by the cryptophycins. Compared with other antitubulin agents such as the vincas and the taxanes, the cryptophycins have a major pharmacological advantage in that they are insensitive to the ABC transporters Pgp and MRP, which are implicated in multidrug resistance.9 The majority of the cellular and biochemical mechanism of action studies were conducted using the synthetic 6,6-gemdimethyl analog (cryptophycin 52) of the natural product cryptophycin 1. Copyright © 2005 CRC Press, LLC A. CELLULAR MECHANISM OF ACTION OF CRYPTOPHYCIN 52 (LY355703) Flow cytometry studies showed that cryptophycin 52 can potently cause the accumulation of cells (CCRF-CEM cells, from 15% to 60%; HT-29 cells, from 24% to 88%) in the G2/M phase of the cell cycle, and this effect was found to be both time and concentration dependent. Microscopic examination of cells treated with cryptophycin 52 showed the cells initially accumulating in mitosis and subsequently undergoing apoptosis. Apoptosis was also evaluated by flow cytometry studies by following the end-labeling of fragmented DNA using bromodeoxyuridine (BrdU). It was found that when CCRF-CEM

cells were treated with 300 pM of cryptophycin 52, 20% of cells were in interphase, 40% were in the mitotic phase, and 40% were in the apoptotic population (at the 24-h time point), which indicated that the cell death mechanism was mediated through the antimitotic effect induced by the cryptophycins. Further studies of cryptophycin 52 in HeLa and HT-29 cells using confocal microscopy showed that cryptophycin 52 caused the mitotic spindle to become fragmented, and multiple microtubule foci were found within a ball-like cluster of chromosomes (HeLa cells at 100 pM, approximately equal to the IC50). At 300 pM of cryptophycin 52, only small residual fragments of spindle microtubules are evident in the chromosome cluster of treated HeLa cells (Figure 9.2). Similar effects that caused spindle fragmentation and multiple microtubule foci were also observed when HT-29 cells were treated with cryptophycin 52 at 300 pM and 1 nM. Abnormal spindle structures appeared within shorter exposure time when a higher concentration (1 nM) of cryptophycin 52 was used. Microtubule bundles were evident in 1-nM compound-treated cells as early as 2 h posttreatment. These bundles are similar to microtubule bundles described in paclitaxel (Taxol®)-treated cells, but not vinblastine- or vincristine-treated cells. The microscopy data clearly supported the notion that microtubules and mitotic spindles are the primary targets of action for cryptophycin 52 and its derivatives. This was further investigated by studying the effect of crytophycin 52 on the dynamics of microtubule polymerization using bovine brain microtubules.10 B. EFFECTS OF CRYPTOPHYCIN 52 (LY355703) ON MICROTUBULE POLYMERIZATION IN VITRO Using a bovine brain microtubule protein preparation that consisted of 70% tubulin and 30% microtubule-associated proteins (MAPs), it was found that cryptophycin 52 potently inhibited the microtubule polymerization (measured by using radiolabeled GTP/GDP bound to tubulin as a sensitive measure of polymerization) with an IC50 of 0.49 µM. This effect was comparable to those exerted by vinblastine (0.35 µM) or vincristine (0.34 µM) and was distinct from paclitaxel, which enhanced (or stabilized) the polymerization of tubulin at concentrations >0.3 µM. Micro- tubules are not simple equilibrium polymers. The hydrolysis of GTP to GDP and inorganic phosphate as tubulin adds to a growing microtubule end creates highly dynamic polymers whose ends grow and shorten stochastically because of the gain and loss of a stabilizing “cap.” The dynamics are extremely rapid both in vitro and in cells, especially during mitosis. Considerable recent evidence strongly indicates that not just the presence of microtubules but also the dynamics of the microtubules are critical for mitotic progression.11 Many of the important antimitotic antitumor drugs such as the vinca alkaloids and paclitaxel exert their powerful antiproliferative actions by suppressing the dynamics of spindle microtubules. Importantly, these antimitotic drugs perturb microtubule dynamics by binding to the ends and along the surface of microtubules at concentrations that are far lower than those required to inhibit microtubule polymerization or induce excess microtubule formation. The effects of cryptophycin 52 on the growing and shortening dynamics of individual microtubules at steady state in real time using video microscopy were thus examined. MAP- free bovine brain tubulin was assembled to polymer mass steady state at the ends of sea urchin Copyright © 2005 CRC Press, LLC The Effects of 355703 on HeLa Cell Mitotic Spindle Structure A B C D FIGURE 9.2 (See color insert ) F luorescence confocal micrographs of HeLa cells. Panel A shows a metaphase mitotic spindle in an untreated control cell. The metaphase spindle is bipolar, with the chromosomes located in a compact metaphase plate at the midpoint between the spindle poles. Panel B shows cells treated with 30 pM of LY355703 (approx. IC50) for 8–10 h. The mitotic spindles look relatively normal. The most obvious abnormality involves the displacement of chromosomes from the compact metaphase plate. In the lower spindle, a chromosome (red) can be seen near the spindle pole (arrow). Panel C (8–10 h; 100 pM of LY355703) and D (8–10 h; 300 pM of LY355703) show increased fragmentation of the spindle microtubules and disorganization of the chromosome mass into ball-like clusters. flagellar axonemal seeds, and the individual dynamics parameters and the changes in length of the microtubules were recorded for approximately 30–45 min. Through such studies, it was found that cryptophycin 52 powerfully suppressed microtubule dynamics at a concentration of 50 nM and at 500 nM, completely shutting down the dynamics of microtubules (Figure 9.3). The studies also indicate that cryptophycin 52 suppresses microtubule shortening rates con- siderably more powerfully than microtubule growing rates. This is a sharp contrast to vinblas- tine, which inhibits both the growing and the shortening of microtubules to similar extents. Overall, the relatively weaker effects of cryptophycin 52 on growing as compared with shortening resembles the action of paclitaxel rather than vinblastine. Thus, cryptophycin 52 has characteristics of both the vinca alkaloids and paclitaxel when it comes to influencing microtubule dynamics.10 Copyright © 2005 CRC Press, LLC Individual Microtubule Life Histories 12 Control 10 8 6 4 2 0 0 1 2 3 4 5 6 12 10 50 nM Cryptophycin 52 8 6 4 2 0 0 1 2 3 4 5 6 12 10 500 nM Cryptophycin 52 8 6 4 2 0 0 1 2 3 4 5 6 Time (minutes) FIGURE 9.3 (See color insert) Microtubule life histories. IV. SYNTHESIS OF CRYPTOPHYCIN 52 (LY355703) A. RETROSYNTHETIC ANALYSIS AND FINAL ASSEMBLY OF THE FRAGMENTS The retrosynthetic analysis of cryptophycin 52 shown in Figure 9.1 represents a straightforward disconnection of the depsipeptide into four corresponding fragments (labeled A–D) through the two ester and two amide linkages. The molecule contains six asymmetric centers, with four contiguous ones present in fragment A, thus making it the most synthetically challenging fragment. For the latter four, the Sharpless asymmetric epoxidation was used as the essential step in controlling two of the four asymmetric centers (16S and 17S), and the final diastereoselective epoxidation of the styrene olefin provided the other two asymmetric centers (18R and 19R) of fragment A. The two distal asymmetric centers (3S and 10R), in contrast, were obtained from optically pure amino acids (D-tyrosine for fragment B, and L-leucine for fragment D, respectively). In general, the cryptophycin nucleus can be assembled efficiently in a convergent fashion by first combining fragments A and B and fragments C and D to give the protected AB and CD fragments. The TBS group of the AB fragments was then removed (aq HF, CH3CN, 95%) to give the secondary alcohol, which was then coupled with the carboxylic acid of the N-BOC C′D fragment (DCC, DMAP, CH2Cl2, 95%) to give the fully protected seco-ABCD compound (Figure 9.4).6 Treatment of the seco compound with TFA (removal of the BOC group) followed by 2- hydroxypyridine effectively catalyzed the macrolactamization to give the cyclized 16-membered depsipeptide in 62% yield (Figure 9.5).12 Diastereoselective epoxidation of the styrylolefin (m- CPBA, CH2Cl2, 95%) gave a 2:1 ratio of epoxides, with the desirable β-epoxide as the major product. The diastereomeric epoxide mixtures can be separated by reverse-phase high-performance Copyright © 2005 CRC Press, LLC Length (µm) O O aq HF OTBS OH OTBS HN Cl CH3CN 95% A O O O HCl H2N Cl Cl3C O O O Ph2P(O)Cl AB i-Pr2NEt B DMF Cl3C 85% O O OH HN Cl O HN Cl O O O O O DCC R2 O O O DMAP O Cl CH R 3C 2Cl2 1 CCl3 95% NHBoc seco-ABCD(C') O O O Pd(PPh3) O OH 4 O morpholine O NHBoc O NHBoc 70% R R 1 R2 1 R2 CD (C'D) ester FIGURE 9.4 Convergent synthetic approach to the seco ABCD compound. liquid chromatography to provide the more active β-epoxide (cryptophycin 52) for biological evaluation. Alternatively, the 2:1 epoxide mixture can be treated with trimethylsilylchloride to give the corresponding chlorohydrins, which then can be readily separated by flash silica gel chroma- tography. Treatment of the pure β-chlorohydrin (cryptophycin 55, 355702) with potassium carbon- ate in acetonitrile converted it back to cryptophycin 52 in high yields. O O HN O O Cl O O HN Cl O TFA O m-CPBA R2 O O O 2-hydroxypyridine O CH2Cl2 O toluene O N O 95% R1 CCl3 62% H NHBoc R1 R2 seco-ABCD(C') Cl O O O TMS-Cl HO O O O HN Cl HN Cl O O O O O O N O K2CO O N O H 3 H CCl3 CH3CN R R1 R 1 R 2 2 α:β = 1:2 Cryptophycin 52: R1 = R2 = CH3 Cryptophycin 55: R1 = R2 = CH3 (LY355703) (355702) FIGURE 9.5 Macrocyclization and the preparation of cryptophycin 52. Copyright © 2005 CRC Press, LLC O CHO 1) HEW O DIBAL OH 3) SAE toluene toluene L-(+)-DET Ti(O-i-Pr)4 t-BuOOH CH2Cl2 –20 °C / 87% OH 4) (CH3)3Al OH OTs O toluene / 85% OH OH 6) TBSOTf OTs 7) NBS/AIBN OTs 9) KCN Et3N / 88% 8) DBU/CH3CN NaHCO OTBS 87% OTBS 3/ 92% O 11 steps CN 10) DIBAL 28% overall 11) HEW / 78% OTBS O ~95% ee OTBS 1 FIGURE 9.6 Sharpless asymmetric epoxidation route to the preparation of Fragment A. B. PREPARATION OF FRAGMENTS (A–D) Fragment A, the synthetically most challenging fragment, with its four contiguous stereogenic centers, was prepared in 11 total linear steps to give methyl(5S,6R)-5-[(tert-butyldimethylsilyl)oxy]-6-methyl- 8-phenylocta-2(E),7(E)-dienoate (1, 28% overall yield, 95% ee). In this sequence, the Sharpless asymmetric epoxidation of the allylic alcohol was used as a key reaction to install two of the four asymmetric centers with ~95% ee (Figure 9.6). Regio- and stereoselective opening of the resulting epoxide with trimethylaluminum installed the desired methyl group and resulted in the diols, where the primary one was selectively tosylated in the presence of dibutyl tin oxide and the secondary one was protected as a TBS ether.13 The styrene moiety was installed via a dehydrobromination, using DBU as the base. Displacement of the tosylate under buffered conditions afforded the crucial nitrile that was reduced to the aldehyde using DIBAL and subsequent Horner-Emmons-Wadsworth olefina- tion gave the final desired unsaturated ester in good yield.14 The other fragments, B, C, and D, were prepared in a more straightforward fashion. Fragment B was obtained in six steps (40% overall yield, 99% ee) from D-tyrosine, as described in Figure 9.7. Fragment C′ (for cryptophycin 52) was obtained in three steps (58% overall) from ethylcyanoacetate, which was then coupled (CDI in THF, 91%) readily with the allyl-L-leucic acid (obtained in two steps from L-leucine) to give the corresponding C′-D fragment (Figure 9.8). Although the total synthesis of cryptophycin 52 required more than 30 discrete synthetic operations, the convergent nature of the process rendered it quite versatile for creating multiple permeations through the combination of different fragments for SAR purposes. This synthetic sequence was also practical enough for the preparation of reasonably large amounts of material for the phase I and II clinical evaluations. V. SARS OF THE CRYPTOPHYCINS Over 500 cryptophycin analogs were prepared via either a total synthetic approach or a semisyn- thetic one, with modifications of the natural cryptophycin 1 or cryptophycin 52, particularly for Copyright © 2005 CRC Press, LLC H2N 1) SO2Cl2 H2N Cl 2) Boc B 2O ocHN Cl HO O OH HCO2H TEA, 85% HO O OH THF HO O OH 90% 3) K2CO BocHN Cl 3 4) 1N NaOH BocHN Cl (CH3)2SO4 THF, Acetone Acetone, MeO O O 92% HO O O CH2Cl2 >95% BocHN Cl HCI.H2N Cl 5) DCC, THF 6) HCI Pyridine O O O EtOAc O O O Cl3CCH2OH 85% 70% Cl3C Cl3C FIGURE 9.7 Preparation of Fragment B. fragment A analogs. Because the cryptophycins exhibited very potent cytotoxicity toward most of the human malignant carcinoma cell lines tested in culture, a 72-h MTT cell-based assay was used to evaluate the newly synthesized analogs. The IC50s for cryptophycins 1 and 52, for example, were in the range of 20–100 pM against several human tumor cell lines (KB, CCRF-CEM, HT-29, SKOV3, LoVo, and GC3). These in vitro cell-based assays provided the required sensitivity and broad range (10 pM ~> 20 µM) of activity essential for differentiating and establishing a firm structure–activity trend for a large number of analogs. In these studies, most of the cytotoxicity data were collected using three cultured human tumor cell lines: KB (Hawaii) and GC3 human colon carcinoma and CCRF-CEM human leukemia (Lilly). The cytotoxicity data cited in the SAR

discussion in this section were mostly derived from the GC3 human colon carcinoma cell line unless stated otherwise. The in vitro cytotoxicity was then used to prioritize the compounds for the initial in vivo antitumor screening, using the Panc 03 (murine pancreatic adenocarcinoma) model, which was then followed by a panel of other murine tumor models and human tumor xenografts (Mam 16C/ADR, Mam 17/ADR, MX-1, GC3, LNCaP, HCT-116, and PC-3).15 The SAR described here will be primarily focused on five regions: the epoxide of fragment A and the prodrugs of its corresponding chlorohydrin, the substitution of the phenyl group of fragment O O O O Cs2CO N 3 Rh/ Al C 2OCN 3 O DMF, Mel O O NHBoc NaOH HO NHBoc H2, THF 71% 81% Boc2O Fragment C' CDI O OH O OH O O THF H2SO4 Allyl Bromide 91% NH NaNO2 OH 2 NaHCO3, H), OH TBAC, CH2Cl2 93% O O O O NHBoc C'D Ester FIGURE 9.8 Preparation of Fragment C′D. Copyright © 2005 CRC Press, LLC Cl O O OH O O O HN 0.009 nM 0.019 nM Cl O Br OH O N O O H H OH Br 0.027 nM 0.004 nM Cl O OH S 4 (>20 nM) 3.06 nM 4.17 nM 2 (9.80 nM) O HN OH 5 (3.13 nM) 71.9 nM 316 nM 3 (3.61 nM) O O Cl F O S OH 6 (32.6 nM) 1160 nM 7 (3.41 nM) OH 8.50 nM FIGURE 9.9 Structure–activity relationship of β-epoxide region of Fragment A. A, the modification of the D-chloromethyltyrosine unit of fragment B, the C/D ester bond, and the C6, C7 positions of fragment C. A. SAR OF THE β-EPOXIDE REGION OF FRAGMENT A Early SAR efforts quickly established the importance of the absolute stereochemistry of the four consecutive stereogenic centers of this fragment. One very important structure–activity feature of the cryptophycins is centered on the configuration of the styrylepoxide portion of the molecule where only the β-epoxide or its masked equivalents (such as the corresponding β-chlorohydrin, β-bromohydrin and the reversed β-bromohydrin) possessed potent cytotoxicity (IC50 = 0.004–0.027 nM) in cell culture assays (Figure 9.9). Epoxides with other configurations (R/S, S/R, and S/S) were much less active. Replacement of the oxygen atom of the epoxide with sulfur (episulfide, 2) or nitrogen (aziridine, 3) led to major loss in activity.16 Removal of the oxygen functionality at C18/C19 also led to inactive phenethyl (4), styryl (5), and chlorobenzyl (6) derivatives of cryptophycin. The corresponding diol and the fluorohydrin (7) derivatives of β-epoxide were also much less active in the in vitro assays. Pharmacokinetic data (vide infra) clearly suggested that the β-chlorohydrin was rapidly con- verting to the corresponding β-epoxide in vivo in rodents and could account for the majority of the antitumor activity seen in murine tumors and human tumor xenografts. The drawback of the chlorohydrins was their lack of long-term stability in a suitable formulation, which prompted a study to stabilize these derivatives and improve their aqueous solubility by preparing several ester prodrugs. Although several of these derivatives showed weaker in vitro activity (CCRF-CEM human leukemia) as compared with cryptophycins 52 and 55, it was necessary to evaluate these derivatives in vivo to determine their potential conversion to the active epoxide species (Figure 9.10).17 B. SAR OF THE PHENYL GROUP OF FRAGMENT A From this SAR it was concluded that the activity of the cryptophycins can be modulated by introducing electron-donating or electron-withdrawing groups on the phenyl ring of the styrene Copyright © 2005 CRC Press, LLC O R O O HN Cl O O N O O H R R Cl Cl O O O O (i) IC50 = 0.25 nM IC50 = 1.3 nM NH2HCl ClH.H NH 2N 2HCl Cl Cl O O O O O NH2HCl IC50 = 0.40 nM IC50 = 3.0 nM HCl.H2N OH OH O Cl Cl O O O IC O O 50 = 0.50 nM IC50 = 2.0 nM O N NH2HCl O O H FIGURE 9.10 IC50s of prodrugs in CCRF-CEM cells. oxide region.18 Introduction of a group larger than a fluorine in the ortho position led to loss of activity, whereas the meta and para positions tolerated many large functional groups. Electron- donating groups such as methyl, hydroxymethyl, and methoxy usually led to highly potent cryptophycin derivatives in vitro (m-methyl, IC50 = 0.15 nM; p-methoxy, IC50 = 0.28 nM; p- hydroxymethyl, IC50 = 0.004 nM; compared with IC50 = 0.022 nM for cryptophycin 52, measured in CCRF-CEM cells). Introduction of p-methyl ester also resulted in an active derivative (IC50 = 0.037 nM; cryptophycin 55, IC50 = 0.05 nM). Along the same trend, introduction of fluorine at either the ortho-, meta-, or para- positions relative to the oxirane ring of the styrene oxide also led to derivatives with potencies ranging between 0.01 and 0.06 nM. Taking advantage of the potential substitution at the para position to improve the aqueous solubility of the molecule, amines 9–11 were prepared, starting with benzyl chloride 8 (Figure 9.11).18 Their potency ranged between 0.001 and 0.092 nM, compared with 0.05 nM for cryptophycin 55, which proved that not only are large groups tolerated but these might provide more potent analogs than cryptophycin 52 or 55. Replacement of the phenyl group of the styrene oxide with other heteroatom-containing aro- matics such as pyridine, thiophene, furan, or thiazole, however, all led to a major loss of cytotoxicity in cell culture. Replacement of the phenyl group with simple alkyl groups (methyl, ethyl, n-propyl, methoxy, formyl, and t-butyl ester) also led to loss of activity in cell culture studies and to completely inactive analogs in in vivo animal studies. C. SAR OF THE D-CHLOROTYROSINE OF FRAGMENT B SAR studies around Fragment B indicated that this is probably the region that is least amenable to structural modification. Various structural modifications of this region of the molecule all led to Copyright © 2005 CRC Press, LLC O O Cl O O HN Cl amine acid O O N H O O 8 9 Cl R = HCI.N O HCI.N 10 R HO O O HN Cl NH.HCI O O N O O H HCI.HN NH2.HCI 11 FIGURE 9.11 Preparation of amines 9–11. significant loss of cytotoxicity in vitro.19 These included replacing the chlorine atom with the fluorine, introduction of additional halogen atoms onto the phenyl ring, modification of the methoxy ether, replacing the chloromethoxyphenyl with other aromatic or nonaromatic groups, and replacing chloromethyltyrosine with other simple natural and unnatural amino acids (such as glycine, valine, proline, p-aminophenylalanine, p-hydroxyphenylglycine, phenylethylglycine, etc.). Methylation of the asymmetric methine carbon at C10 also gave inactive analogs. The only tolerable modification in fragment B was the replacement of the chlorine with a bromine or a methyl group on the phenyl ring of the tyrosine. This led to the corresponding methyl D-tyrosine derivative, which had only a twofold decrease of cytotoxicity in vitro (IC50 = 0.04 nM; cryptophycin 52, IC50 = 0.019 nM, measured in CCRF-CEM cells). D. SAR OF THE C/D ESTER BOND AND THE C6, C7 POSITIONS Some of the early studies of cryptophycin 1 indicated that the ester bond between fragments C and D of the macrocyclic ring may be metabolically labile to nucleophilic attack. This is supported by the observation that arenastatin (a closely related natural product isolated from Japanese sponge Dysidea arenaria) is completely devoid of activity in vivo, although it has potent activity in cell culture.20 Two approaches were then investigated for the stabilization of the C–D ester bond: (a) geminal substitution — to the carbonyl group and (b) isosteric replacement of the ester bond with other functionalities such as an amide. Both approaches produced very active derivatives — in particular, introduction of a gem-dimethyl group at the C6 position gave an analog (cryptophycin 52, LY355703) with a very much improved in vivo antitumor efficacy profile compared with cryptophycin 1. Although geminal disubstitution was effective in protecting the carbonyl group from nucleophilic attack and could be successfully extended to the spirocyclopropyl group (12), however, it was found that potency was rapidly diminished when larger groups (gem-diethyl, gem-di-n-propyl, spirocyclopentyl, and spirocyclohexyl) were introduced into the C6 position (Figure 9.12).21 The amide isosteric replacement of the C–D ester bond also gave very active analogs 13 and 14, regardless of whether there was a monomethyl or a geminal dimethyl group at the C6 position.22 Both agents exhibited potent in vitro activity (IC50 = 0.015 nM); however, it was found that the in vivo potency for the gem-dimethyl amide analog was significantly lower than that of cryptophycin 52 when compared head-to-head in the Panc 03 model. Replacement of the C–D ester with the corresponding ether linkage (15) causes a major loss of activity in cell culture (240-fold).22 As part of the SAR studies of Fragment C, it was also found that moving the substituent (R = CH3, 16) from the C6 to the C7 position retains the activity in vitro. However, similar to what was observed for arenastatin, absence of substituents at the C6 position caused the compound to be completely inactive in vivo presumably because of the metabolic instability issue with the C–D ester bond. Copyright © 2005 CRC Press, LLC O O N NH H R' O O O HN Cl O 13 R' = H, 0.015 nM O N O O 14 R' = CH3, 0.014 nM H H 15 C-D ether analog (IC50= 3.6 nM) O R O O O NH O NH O NH R H R R' 16 R = CH3, 0.4 nM 17 R = Ph, 1.0 nM R = H, Arenastatin Cryptophycin 52 R = R'= CH3, (IC50= 0.017 nM) (IC50= 0.038 nM) inactive in vivo O H R = CH3, Cryptophycin 1 12 R = R'= cyclopropyl O N (IC50= 0.009 nM) (IC50= 0.010 nM) R H 18 R = CH3, 13 nM 19 R = isopropyl, 306 nM FIGURE 9.12 Structure–activity relationship of CD ester bond and C6, C7 positions. Larger substituents at the C7 position (R = phenyl, 17) did not improve the activity of the compound either in vitro or in vivo. Excising the C6 carbon atom of the peptolide ring was also attempted, and the result was disappointing. The smaller 15-membered macrocycles 18 and 19 were signifi- cantly less active in cell culture (1,400-fold) than the natural cryptophycins.23 The key findings from the SAR studies conducted around the various regions of the molecule are summarized in Figure 9.13. Overall, it can be concluded, first, that the absolute stereochemistry of all seven asymmetric centers is essential for eliciting the biological activity of the cryptophycins, and the opposite configuration at these centers led to major loss of activity. Second, the β-epoxides (R/R) at the 18 and 19 positions (Fragment A) are consistently much more active than the other isomers (R/S, S/R, and S/S). Third, substitution of the amide nitrogen of the macrocyclic ring led to inactive analogs; x-ray and nuclear magnetic resonance studies indicate that an intrapeptolide R-CH3 is the best β-epoxide or substituent chlorohydrin CH2CH2 active in vitro Cl, Br and CH3 gave best potency phenyl or OCH3 is required substituted phenyl for activity F, CH3, OCH3, CH2OH R S O R S O O O HN R Cl A-D amide O inactive S R O N O O H isobutyl >> H sec-butyl or substitution of amide NH isopropyl inactive C-D amide tolerated gem- dimethyl substitution or deletion and cyclopropyl of C7 — lost activity FIGURE 9.13 Key structure–activity relationship findings of the cryptophycins. Copyright © 2005 CRC Press, LLC TABLE 9.1 Comparison of the In Vitro Cytotoxicity and the In Vivo Antitumor Activity of Cryptophycins 52, 55 and 55-glycinate in Human Tumor Xenografts Crypto 52 Crypto 55 Crypto 55-glycinate (355703) (355702) (368422) Cytotoxicity (IC50/nM/GC3 0.2–0.7 0.2–0.75 1.0 cells) In Vivo Potency 10–12 mg/kg 40–48 mg/kg 48–64 mg/kg (CDI/nu/nu) (CD1/nu/nu) (CD1/nu/nu) In vivo Antitumor Activitya GC3 Colon + + +++ LNCap Prostate ++++ ++++ ++++ MX-1 Mammary ++ +++ NT H116 Colon +++ ++++ NT a All in vivo antitumor activity are expressed in log cell kill units: Log Kill = T-C/3.2 × Td. +: 0.7–1.2: ++: 1.3–1.9: +++: 2.0–2.8: ++++: >2.8. hydrogen bond may exist between the NH(8) and CO(12) groups for the macrocyclic ring to adopt certain conformations. Fourth, amide replacement of the ester linkage may be tolerated between the

fragment C/D but not between the A/D fragments. Fifth, substituents on the phenyl ring of fragment A are well tolerated; however, substitutions and modifications of the chloromethyl tyrosine unit of Fragment B are limited to a few options. VI. COMPARISON OF CRYPTOPHYCINS 52, 55, AND 55-GLYCINATE A comparison of the in vitro cytotoxicity and the in vivo antitumor activity of cryptophycin 52 (epoxide, LY355703), cryptophycin 55 (chlorohydrin derivative of cryptophycin 52, 355702), and the glycinate derivative of chlorohydrin 55 (analog [i] in Figure 9.10, 368422) are summarized in Table 9.1. All three compounds were found to exert potent cytotoxic activities (200 pM ~ 1 nM) against a number of human tumor cell lines (CCRF-CEM, GC3, HeLa), with the epoxide and chlorohydrin being slightly more potent than the corresponding glycinate derivative. However, it was found that the epoxide compound exhibited the most potent effect in vivo when compared on a milligram-by-milligram basis. To achieve similar levels of antitumor activities in vivo (maximum tolerated doses), it was usually necessary to dose four- to five-fold more of the chlorohydrin or glycinate as compared to the epoxide. This is primarily because both chlorohydrin and glycinate were found to act as prodrugs (vide infra) of the epoxide species in vivo. All three cryptophycins (355703, 355702, and 368422) were found to be active against a broad panel of human tumor xenografts including colon, mammary, prostate, pancreatic, and lung. The level of antitumor activity exhibited by the cryptophycins is superior to most of the commonly used anticancer agents such as the vincas, 5-fluorouracil, cisplatin, and doxorubicin. The only other class of compound that demonstrated comparable levels of in vivo activity as the cryptophycins in these tumors were the taxanes (such as paclitaxel). Various studies demonstrated that the chlorohydrin derivative of cryptophycin 52 (355702) can be converted to the more potent epoxide under in vitro as well as in vivo conditions. It was found that the chlorohydrin derivatives could quite rapidly be transformed into the corresponding epoxide in the formulation (contains 5% ethanol and 5% cremophor) under various storage conditions. This Copyright © 2005 CRC Press, LLC TABLE 9.2 Antitumor Activity (Log Kill*) of LY355703 and Other Oncolytics Against Human Tumor Xenografts HCT-116 H125 LNCaP TSU PC-3 Agent Colon Lung Prostate Prostate Prostate Cryptophycin 52 +++ ++ ++++ ++ + Paclitaxel +++ + ++++ ++++ ++ Doxorubicin NT — — — — Etoposide — + + — — VLB/VCR + +++ ++++ ++++ ++ Cytoxan — — ++++ ++++ — Cisplatin — — — — — 5-FU — — — — — * All in vivo antitumor activity are expressed in log cell kill units: Log Kill = T-C/3.2 × Td. +: 0.7–1.2: ++: 1.3–1.9: +++: 2.0–2.8: ++++: >2.8. conversion was also found to take place in plasma (in vitro) and in animal studies. For example, in mice dosed intravenously with 80 mg/kg of cryptophycin 55, the plasma levels of cryptophycin 52 (converted directly from cryptophycin 55) were almost identical to those found after adminis- tration of 20 mg/kg of cryptophycin 52 itself (a dose relationship that was consistent with the findings of the efficacy studies). This observation indicated that the chlorohydrin derivative cryp- tophycin 55 may act as a prodrug of cryptophycin 52. For the glycinate derivative to convert to the active component epoxide, it requires an enzymatic cleavage step of the ester bond, via the esterases, to generate chlorohydrin 55, which subsequently forms cryptophycin 52. Although the ester hydrolysis of the glycinate (368422) seems quite efficient in mouse plasma, it was discovered that the ester hydrolysis of the glycinate to give the epoxide was significantly reduced in dog and human plasma. VII. CRYPTOPHYCIN 52 (LY355703) A. IN VIVO ANTITUMOR ACTIVITY OF CRYPTOPHYCIN 52 (LY355703) The antitumor activity of cryptophycin 52 in vivo, however, is summarized here to illustrate the broad spectrum of activity of this novel agent against a variety of syngeneic murine solid tumors and human tumor xenografts. Because of the lipophilic nature of the cryptophycins, cryptophycin 52 was formulated in a mixture of 5% Cremophor and 5% ethanol in 90% saline, a formulation that is used clinically for paclitaxel. The treatment schedule varied but was generally consistent, with at least five repeated intermittent intravenous injections, usually every other day (q2d), and the doses used in the xenograft studies are usually in the range of 6–10 mg/kg for cryptophycin 52. The maximum tolerated dose for cryptophycin 52 in nude mice is 10 mg/kg per injection, with a total dose of ca. 50–60 mg/kg. Table 9.2 summarizes the antitumor activity (expressed in Log Kill) of cryptophycin 52 against five human tumor xenografts (HCT-116 colon, H125 lung, LNCaP prostate, TSU prostate, and PC-3 prostate carcinomas). In comparison, the antitumor activity of a group of commonly used oncolytic agents (paclitaxel, doxorubicin, etoposide, vincristine/vinblas- tine, cytoxan, cisplatin, and 5-fluorouracil) is also included.24 It is clear from these studies that cryptophycin 52 is a very active anticancer agent with superior antitumor activity compared to many of the commonly used oncolytic drugs. In addition, cryptophycin 52 also exhibited excellent activity against tumors that are resistant to adriamycin (Mamm16C/Adr, Mamm17/Adr, both are Copyright © 2005 CRC Press, LLC syngeneic murine tumors) and vinblastine (UCLA-P3), whereas paclitaxel, doxorubicin, etoposide, and vinblastine are completely inactive in these multidrug-resistant tumors. B. CLINICAL EVALUATION OF CRYPTOPHYCIN 52 (LY355703) Cryptophycin 52, with its broad antitumor activity and potency, was chosen as the first-generation cryptophycin to undergo clinical evaluation, with 25 patients receiving the drug as a 2-h intravenous infusion on days 1 and 8 every 3 weeks. The doses were escalated from 0.1 to 2.22 mg/m2 where neurological toxicities at or over 1.84 mg/m2 were found to be dose limiting. An alternative twice- weekly schedule was also investigated with 11 patients in an attempt to reduce the neurotoxicity. Unfortunately, doses over 0.75 mg/m2 on a days 1, 4, 8, and 11 were also not tolerated as a result of nausea and constipation.25 As a result of phase I studies showing that toxicity was not schedule dependent but most likely associated with cumulative dose, the dose for phase II was set at 1.5 mg/m2 for a day 1 and 8 schedule. In Phase II clinical trials, LY355703 was used as second-line therapy for metastatic colorectal cancer, for example, or in non–small cell lung cancer patients who had received previous platinum- based chemotherapy. Although partial responses were observed in two patients with platinum- resistant advanced ovarian cancer, no significant activity or stable disease was reported in patients with other tumor types. Toxicities such as arthralgia, constipation, myalgia, and neuropathy were also observed; therefore, these results, along with the lack of efficacy, led to the termination of the clinical trials. VIII. CONCLUSIONS The discovery of the cryptophycins as a new class of antimitotic agents has attracted the attention of many synthetic groups to find better routes for their preparation and to investigate the activity of new related analogs.26 In addition, investigators specializing in antimicrotubule agents have extensively studied their mode of action.27 The research described in this chapter was the collab- orative effort of Eli Lilly, University of Hawaii, and Wayne State University scientists resulting in the identification of cryptophycin 52 as the first generation cryptophycin to undergo clinical evaluation.28 Extensive SAR studies were conducted and facilitated by the support provided by Lilly’s process chemistry group. A convergent and highly efficient synthetic route to cryptophycin 52 was developed and used to both support the SAR and provide the clinical material. The SAR studies led to a large number of cryptophycins with a broad array of structural diversity and activity. They also revealed some critically important areas for structure activity optimization. These included the styrene β-oxide region, the C/D ester bond region, and the phenyl region of fragment A. From the assessment of the collective data of in vitro/in vivo pharmacology, biophar- maceutical properties, ADME/pharmacokinetic evaluation, and toxicological findings, cryptophycin 52 (LY355703) was selected as the first-generation cryptophycin compound for clinical evaluation. The greatest advantage identified during the preclinical investigation of this series was the effec- tiveness of many of the cryptophycins against tumors that overexpressed the multidrug resistance phenotype. This new class of macrocyclic depsipeptides has proven to be quite novel in its mode of action and in its preclinical antitumor activity. ACKNOWLEDGMENTS We wish to thank all our colleagues and collaborators who contributed to this work: Andrew H. Fray, Michael J. Martinelli, Eric D. Moher, Bryan H. Norman, Vinod F. Patel, Richard M. Schultz, John E. Toth, David L. Varie, Daniel C. Williams, John F. Worzalla, Tony Y. Zhang, Thomas H. Corbett (Wayne State University), Marc Tius (University of Hawaii), Richard E. Moore (University of Hawaii), Leslie Wilson, and Mary Ann Jordan (University of California at Santa Barbara). Special Copyright © 2005 CRC Press, LLC thanks to Jenniffer Phillips and Susan Koppleman for their help with the references and formatting of the document. We also wish to thank Drs. Gordon Cragg, David Newman (National Cancer Institute), and David Kingston (Virginia Tech University) for giving us the opportunity to summarize the work of the cryptophycin team and for their critical evaluation of this chapter. REFERENCES 1. Schwartz, R. E. et al., Pharmaceuticals from cultured algae. J. Ind. Microbiol., 5, 118–126, 1990. 2. Patterson, G. M. et al., Antineoplastic activity of cultured blue-green-algae (Cyanophyta). J. Phycol., 27, 530–536, 1991. 3. (a) Corbett, T. H. et al., Cytoxic anticancer drugs: models and concepts for drug discovery and development; Valeriote, F. A., Corbett, T. H., Baker, L. H., Eds., Kluwer Academic Publishers, 35–87, 1992. (b) Valeriote, F. A. et al., Discovery and development of anticancer agents, Valeriote, F. A., Corbett, T. H., Baker, L. H., Eds., Kluwer Academic Publishers, 67–93, 1997. 4. (a) Golakoti, T. et al., Total structures of cryptophycins, potent antitumor depsipeptides from the blue- green alga Nostoc sp. Strain GSV 224. J. Am. Chem. Soc., 116, 4729–4737, 1994. (b) Golakoti, T. et al., Structure determination, conformational analysis, chemical stability studies, and antitumor evaluation of the cryptophycins. Isolation of 18 new analogs from Nostoc sp. strain GSV 224, J. Am. Chem. Soc., 117, 12030–12049, 1995. 5. (a) Corbett, T. H. et al., Anticancer Drug Development Guide. Teicher, B. A., Ed., Humana Press, 75–99, 1997. (b) Corbett, T. H. et al., Discovery of cryptophycin-1 and BCN-183577: examples of strategies and problems in the detection of antitumor activity in mice. Invest. New Drugs, 15, 207–218, 1997. 6. Barrow, R. A. et al., Total synthesis of cryptophycins. Revision of the structures of cryptophycins A and C. J. Am. Chem. Soc., 117, 2479–2490, 1995. 7. (a) Kerksiek, K. et al., Interaction of cryptophycin 1 with tubulin and microtubules. FEBS Lett., 377, 59–61, 1995. (b) Smith, C. D., Zhang, X. Mechanism of action of cryptophycin. Interaction with the Vinca alkaloid domain of tubulin. J. Biol. Chem., 271, 6192–6198, 1996. (c) Bai, R. et al., Charac- terization of the interaction of cryptophycin 1 with tubulin: binding in the Vinca domain, competitive inhibition of dolastatin 10 binding, and an unusual aggregation reaction. Cancer Res., 56, 4398–4406, 1996. (d) Mooberry, S. L., Taoko, C. R., Busquets, L. Cryptophycin 1 binds to tubulin at a site distinct from the colchicine binding site and at a site that may overlap the vinca binding site. Cancer Lett., 107, 53–57, 1996. (e) Panda, D. et al., Mechanism of action of the unusually potent microtubule inhibitor cryptophycin 1. Biochemistry, 36, 12948–12953, 1997. 8. Mooberry, S. L., Busquets, L., Tien, G. Induction of apoptosis by cryptophycin 1, a new antimicro- tubule agent. Int. J. Cancer, 73, 440–448, 1997. 9. (a) Smith, C. D. et al., Cryptophycin: a new antimicrotubule agent active against drug-resistant cells. Cancer Res, 54, 3779–3784, 1994. (b) Wagner, M. M. et al., In vitro pharmacology of cryptophycin 52 (LY355703) in human tumor cell lines. Cancer Chemother. Pharmacol., 43, 115–125, 1999. 10. Panda, D. et al., Antiproliferative mechanism of action of cryptophycin-52: kinetic stabilization of microtubule dynamics by high-affinity binding to microtubule ends. Proc. Nat. Acad. Sci. USA, 95, 9313–9318, 1998. 11. (a) Farrell, K. W., Jordan, M. A. A kinetic analysis of assembly-disassembly at opposite microtubule ends. J. Biol. Chem., 257, 3131–3138, 1982. (b) Farrell, K. W. et al., Phase dynamics at microtubule ends: the coexistence of microtubule length changes and treadmilling. J. Cell Biol., 104, 1035–1046, 1987. (c) Wilson, L., Jordan, M. A. Pharmacological probes of microtubule function. Modern

Cell Biol., 13, 59–83, 1994. (d) Jordan, M. A., Wilson, L. Microtubules and actin filaments: dynamic targets for cancer chemotherapy. Curr. Opin. Cell Biol., 10, 123–130, 1998. (e) Yvon, A-M. C., Wadsworth, P., Jordan, M. A. Taxol suppresses dynamics of individual microtubules in living human tumor cells. Mol. Biol. Cell, 10, 947–959, 1999. (f) Wilson, L., Panda, D., Jordan, M. A. Modulation of microtubule dynamics by drugs: a paradigm for the actions of cellular regulators. Cell Struct. Funct., 24, 329–335, 1999. (g) Jordan, M. A. Mechanism of action of antitumor drugs that interact with microtubules and tubulin. Curr. Med. Chem: Anti-cancer Agents, 2, 1–17, 2002. (h) Jordan, M. A., Wilson, L. Microtubules as a target for anticancer drugs. Nat. Rev. Cancer, 4, 253–265, 2004. Copyright © 2005 CRC Press, LLC 12. Fray. A. H., Intramolecular aminolysis of trichloroethyl esters: a mild macrocyclization protocol for the preparation of cryptophycin derivatives. Tetrahedron Asymmetry, 9, 2777–2781, 1998. 13. (a) Martinelli, M. J. et al., Dibutyltin oxide catalyzed selective sulfonylation of β-chelatable primary alcohols. Org. Lett., 1, 447–450, 1999. (b) Martinelli, M. J., Vaidyanathan, R., Khau, V. V., Selective monosulfonylation of internal 1,2-diols catalyzed by di-n-butyltin oxide. Tetrahedron Lett., 41, 3773–3776, 2000. 14. For other approaches to Fragment A see: (a) Varie, D. L. et al., Bioreduction of (R)-carvone and regioselective Baeyer-Villiger oxidations: application to the asymmetric synthesis of cryptophycin fragment A. Tet. Lett., 39, 8405–8408, 1998. (b) Dhokte, U. P. et al., A novel approach for total synthesis of cryptophycins via asymmetric crotylboration protocol. Tet. Lett., 39, 8771–8774, 1998. (c) Furuyama, M., Shimizu, I. A short enantioselective synthesis of a component of cryptophycin A and arenastatin A. Tetrahedron Asymmetry, 9, 1351–1357, 1998. (d) Liang, J. et al., Synthesis of unit A of cryptophycin via a [2,3]-Wittig rearrangement. J. Org. Chem., 64, 1459–1463, 1999. (e) Liang, J. et al., Synthesis of cryptophycin 52 using the Sharpless asymmetric dihydroxylation: diol to epoxide transformation optimized for a base-sensitive substrate. J. Org. Chem., 65, 3143–3147, 2000. (f) Pousset, C; Haddad, M., Larcheveque, M. Diastereocontrolled synthesis of unit A of cryptophycin. Tetrahedron, 57, 7163–7167, 2001. (g) Raghavan, S., Tony, K. A. Sulfinyl moiety as an internal nucleophile. 1. Efficient stereoselective synthesis of fragment A of cryptophycin 3. J. Org. Chem., 68, 5002–5005, 2003. (h) Phukan, P., Sasmal, S., Maier, M. E. Flexible routes to the 5-hydroxy acid fragment of the cryptophycins. Eur. J. Org. Chem., 9, 1733–1740, 2003. 15. Moore, R. E. et al., The search for new antitumor drugs from blue-green algae. Current Pharmaceutical Design, 2, 317–330, 1996. 16. Moore, R. E. Cyclic peptides and depsipeptides from cyanobacteria: a review. J. Ind. Microbiol., 16, 134–143, 1996. 17. Al-awar, R. S. et al., Preparation of novel cryptophycin pharmaceuticals. PCT Int. Appl., 293 pp. WO 9808505 A1 19980305, 1998. 18. Al-awar, R. S. et al., A convergent approach to cryptophycin 52 analogues: synthesis and biological evaluation of a novel series of fragment a epoxides and chlorohydrins. J. Med. Chem., 46, 2985–3007, 2003. 19. Patel, V. F. et al., Novel cryptophycin antitumor agents: synthesis and cytotoxicity of fragment “B” analogues. J. Med. Chem., 42, 2588–2603, 1999. 20. (a) Kobayashi, M. et al., Arenastatin A, a potent cytotoxic depsipeptide from the okinawan marine sponge Dysidea arenaria. Tetrahedron Lett., 35, 7969–7972, 1994. (b) Kobayashi, M., et al., The absolute stereostructure of arenastatin A, a potent cytotoxic depsipeptide from the Okinawan marine sponge Dysidea arenaria. Chem. Pharm. Bull., 42, 2196–2198, 1994. 21. Varie, E. et al., Synthesis and biological evaluation of cryptophycin analogs with substitution at C-6 (fragment C region). Bioorg. Med. Chem. Lett., 9, 369–374, 1999. 22. Norman, B. H. et al., Total synthesis of cryptophycin analogs. Isosteric replacement of the C-D ester. J. Org. Chem., 63, 5288–5294, 1998. 23. Shih, C. et al., Synthesis and biological evaluation of novel cryptophycin analogs with modification in the β-alanine region. Bioorg. Med. Chem. Lett., 9, 69–74, 1999. 24. (a) Worzalla, J. F. et al., LY355702 and LY355703, new cryptophycin analogues with antitumor activity against human tumor xenografts. Proc. Am. Assoc. Cancer Res., 38, 225, 1997, (Abs. 1516). (b) Polin, L. et al., Preclinical antitumor activity of cryptophycin-52/55 (C-52; C-55) against human tumors in SCID mice. Proc. Am. Assoc. Cancer Res., 38, 225, 1997, (Abs. 1514). (c) Polin, L. et al., Treatment of human prostate tumors PC-3 and TSU-PR1 with standard and investigational agents in SCID mice. Invest. New Drugs, 15, 99–108, 1997. 25. (a) Sessa, C. et al., Phase I and pharmacological studies of the cryptophycin analogue LY355703 administered on a single intermittent or weekly schedule. Eur. J. Cancer, 38, 2388–2396, 2002. (b) Stevenson, J. P. et al., Phase I trial of the cryptophycin analogue LY355703 administered as an intravenous infusion on a day 1 and 8 schedule every 21 days. Clin. Cancer Res., 8, 2524–2529, 2002. 26. (a) Salamonczyk, G. M. et al., Total synthesis of cryptophycins via a chemoenzymatic approach. J. Org. Chem., 61, 6893–6900, 1996. (b) de Muys, J-M. et al., Synthesis and in vitro cytotoxicity of cryptophycins and related analogs. Bioorg. Med. Chem. Lett., 6, 1111–1116, 1996. (c) Rej, R. et al., Total synthesis of cryptophycins and their 16-(3-phenylacryloyl) derivatives. J. Org. Chem., 61, Copyright © 2005 CRC Press, LLC 6289–6295, 1996. (d) Gardinier, K. M., Leahy, J. W. Enantiospecific total synthesis of the potent antitumor macrolides cryptophycins 1 and 8. J. Org. Chem., 62, 7098–7099, 1997. (e) Ali, S. M., Georg, G. I. Formal syntheses of cryptophycin A and arenastatin A. Tetrahedron Lett., 38, 1703–1706, 1997. (f) Georg, G. I., Ali, S. M., Stella, V. J. Halohydrin analogues of cryptophycin 1: synthesis and biological activity. Bioorg. Med. Chem. Lett., 8, 1959–1962, 1998. (g) Eggen, M., Georg, G. I. Enantioselective synthesis of a 3β-dephenylcryptophycin synthon. Bioorg. Med. Chem. Lett., 8, 3177–3180, 1998. (h) White, J. D., Hong, J., Robarge, L. A. Total synthesis of cryptophycins-1, -3, -4, -24 (arenastatin A), and -29, cytotoxic depsipeptides from cyanobacteria of the Nostocaceae. J. Org. Chem., 64, 6206–6216, 1999. (i) Chakraborty, T. K., Das, S. Diastereoselective opening of trisubstituted epoxy alcohols: application in the studies directed toward the synthesis of octenoic acid moiety of cryptophycins. J. Indian Chem. Soc., 76, 611–616, 1999. (j) Christopher, J. A. et al., A synthesis of cryptophycin 4 using a planar chiral molybdenum cationic complex. Synlett., 4, 463–466, 2000. (k) Ghosh, A. K., Bischoff, A. A convergent synthesis of (+)-cryptophycin B, a potent antitumor macrolide from Nostoc sp. Cyanobacteria. Org. Lett., 2, 1573–1575, 2000. (l) Barrow, R. A. et al., Synthesis of 1-aza-cryptophycin 1, an unstable cryptophycin. An unusual skeletal rearrangement. Tetrahedron, 56, 3339–3351, 2000. (m) Eggen, M. et al., Total synthesis of cryptophycin-24 (Aren- astatin A) amenable to structural modifications in the C16 side chain. J. Org. Chem., 65, 7792–7799, 2000. (n) Eggen, MJ., Nair, S. K., Georg, G. I. Rapid entry into the cryptophycin core via an acyl- β-lactam macrolactonization: total synthesis of cryptophycin-24. Org. Lett., 3, 1813–1815, 2001. (o) Smith, A. B. et al., First generation design, synthesis, and evaluation of azepine-based cryptophycin analogues. Org. Lett., 3, 4063–4066, 2001. (p) Martinelli, M. J. et al., Reaction of cryptophycin 52 with thiols. Tetrahedron Lett., 43, 3365–3367, 2002. (q) Li, L-H., Tius, M. A. Stereospecific synthesis of cryptophycin 1. Org. Lett., 4, 1637–1640, 2002. (r) Hoard, D. W. et al., Synthesis of cryptophycin 52 using the Shi epoxidation. Org. Lett., 4, 1813–1815, 2002. (s) Smith, A. B. et al., Design, synthesis, and evaluation of azepine-based cryptophycin mimetics. Tetrahedron, 59, 6991–7009, 2003. (t) Vidya, R. et al., Synthesis of cryptophycins via an N-acyl-β-lactam macrolactonization. J. Org. Chem., 68, 9687–9693, 2003. (u) Ghosh, A. K., Swanson, L. Enantioselective synthesis of (+)-cryptophycin 52 (LY355703), a potent antimitotic antitumor agent. J. Org. Chem., 68, 9823–9826, 2003. 27. (a) Smith, C. D., Zhang, X. Mechanism of action of cryptophycin. Interaction with the Vinca alkaloid domain of tubulin. J. Biol. Chem., 271, 6192–6198, 1996. (b) Kessel, D., Luo, Y. Cells in cryptophycin- induced cell-cycle arrest are susceptible to apoptosis. Cancer Lett., 151, 25–29, 2000. (c) Barbier, P. et al., In vitro effect of cryptophycin 52 on microtubule assembly and tubulin: molecular modeling of the mechanism of action of a new antimitotic drug. Biochemistry, 40, 13510–13519, 2001. (d) Watts, N. R. et al., The cryptophycin-tubulin ring structure indicates two points of curvature in the tubulin dimer. Biochemistry, 41, 12662–12669, 2002. (e) Drew, L et al., The novel antimicrotubule agent cryptophycin 52 (LY355703) induces apoptosis via multiple pathways in human prostate cancer cells. Clinical Cancer Res., 8, 3922–3932, 2002. 28. For other reviews on the cryptophycins see: (a) Shih, C., Teicher, B. A. Cryptophycins: a novel class of potent antimitotic antitumor depsipeptides. Curr. Pharm. Design, 7, 1259–1276, 2001. (b) Shih, C. et al., Synthesis and structure-activity relationship studies of cryptophycins: a novel class of potent antimitotic antitumor depsipeptides. ACS Symposium Series, 796 (Anticancer Agents), 171–189, 2001. (c) Eggen, MJ., Georg, G. I. The cryptophycins: their synthesis and anticancer activity. Med. Res. Rev., 22, 85–101, 2002. (d) Tius, M. A. Synthesis of the cryptophycins. Tetrahedron, 58, 4343–4367, 2002. (e) Li, T., Shih, C. Structure activity relationships of cryptophycins, a novel class of antitumor antimitotic agents. Frontiers Biotech. Pharm., 3, 172–192, 2002. (f) Hong, J., Zhang, Li. Novel approaches to the synthesis of cryptophycin and related depsipeptides. Frontiers Biotech. Pharm., 3, 193–213, 2002. (g) Tius, M. A. Cryptophycin synthesis. Handbook Environ. Chem., 3 (Part P), 265–305, 2003. Copyright © 2005 CRC Press, LLC 10 Chemistry and Biology of the Discodermolides, Potent Mitotic Spindle Poisons Sarath P. Gunasekera and Amy E. Wright CONTENTS I. Introduction II. Natural Sources III. Chemistry IV. Synthesis A. Synthesis of Discodermolide B. Synthesis of Analogs V. Biological Activity A. Immunosuppressive Properties B. Antitumor Properties VI. Structure–Activity Studies VII. Clinical Investigation References I. INTRODUCTION (+)-Discodermolide (1) is a highly functionalized, marine sponge–derived polyketide δ-lactone presumably formed by a combination of propionate and acetate units. 1 was first isolated and characterized by Gunasekera et al.1 at the Harbor Branch Oceanographic Institution (HBOI) from the marine sponge Discodermia dissoluta. Crude extracts of the sponge showed strong potency against the murine P388 lymphocytic leukemia cell line, and bioassay-guided fractionation using this assay resulted in the purification of 1. The structure was defined by spectroscopic methods and was confirmed by X-ray diffraction analysis. The initial reports of (+)-discodermolide’s strong in vitro antiproliferative and in vitro and in vivo immunosuppressive activities2,3 created a wide interest among the synthetic organic chemistry community. The absolute stereochemistry was subsequently defined by total synthesis conducted by the Schreiber group.4 (+)-Discodermolide is a member of the group of antimitotic agents known to act by microtubule stabilization, other members of which include Taxol®5 (2), the epothilones A6 (3), B6 (4), and D7 (5), eleutherobin8 (6), laulimalide9 (7), and recently, dictyostatin 110 (8). Because of its strong microtubule stabilizing properties and its excellent activity against multiple-drug-resistant tumors, HBOI licensed (+)-discodermolide (1) to Novartis Pharmaceutical Corporation in early 1998 for development as an anticancer drug. The discovery that discodermolide’s mechanism of action is similar to that of Taxol (2)5 (Taxol is a registered trademark of Bristol-Myers Squibb Company, NY), coupled with its limited supply Copyright © 2005 CRC Press, LLC from the natural source, prompted the scientific community to prepare synthetic 1 and its analogs via a number of routes. The unique chemical structure with 13 stereogenic centers posed a major synthetic challenge for the preparation of sufficient material to allow for the preclinical and clinical studies. Over 15 publications describe this work. A major milestone in the development of 1 was the progression of a large-scale synthesis of (+)-discodermolide (1) by the Smith group.11 This, coupled with other synthetic schemes, led to a refined process in which over 60 g of synthetic (+)- discodermolide was produced. At present, synthetic (+)-discodermolide is undergoing phase I clinical trials for solid tumor malignancies at the Cancer Therapy and Research Center in San Antonio, Texas. II. NATURAL SOURCES The genus Discodermia du Bocage belongs to the family Theonellidae. This genus is characterized by the presence of discotriaene spicules in addition to the presence of desmas, which are characteristic of all lithistid sponges. The microsclere compliment of the genus Discodermia consists of microxeas and microrhabds.12 Discodermia is represented by at least three valid species, D. dissoluta Schmidt, D. polydiscus du Bocage, and D. verrucosa, in the Caribbean region of the central Atlantic Ocean. Additional as-yet-undescribed species have been collected by

HBOI and are in their collection. (+)-Discodermolide (1) was first isolated1 from a sponge sample collected on March 27, 1987, off Lucaya, Grand Bahamas Island, at a depth of 30 m. Subsequently, sufficient quantity of 1 was isolated from an earlier 1985 collection to allow for the completion of the structure elucidation and preliminary biological evaluations. The 1987 sponge sample was slightly dark in appearance, and the compound responsible for this color coeluted in trace amounts with the first isolate of (+)- discodermolide. This trace impurity gave a pink coloration to the solution of (1), and further purification attenuated the yield of 1. The freshly thawed 434 g of sponge sample from the 1987 collection yielded 7 mg of pure (+)-discodermolide. The samples collected in 1985 and most of the subsequent collections did not contain the pink impurity, and the average yield of 1 was 10 mg/kg of the wet sponge sample. Harbor Branch chemists noted the existence of two chemically distinguishable forms of D. dissoluta. Both sponges contained 1 as a minor secondary metabolite and either the tetramic acid–derived alkaloid discodermide (9)13 or a mixture of depsipeptides trivially named polydisca- mides as the major metabolites. The major metabolites discodermide (9) and polydiscamide A (10)14 occur in gram quantities in 1 kg of the wet sponge sample, but 9 and 10 have not been observed to co-occur in the same sponge sample. The need for larger amounts of discodermolide for preclinical investigation prompted the collection of 20 kg of Discodermia spp. in November 1998. From this material, four naturally occurring analogs have been isolated. One sponge sample (wet weight of 1931 g) was collected by submersible on November 23, 1998, at a depth of 147 m, from the Bell Channel Buoy, Grand Bahama Island, and yielded 12.0 mg of 1 and 0.3 mg of 2- epi-discodermolide15 (11, yield, 0.00001% of wet weight). The morphology of this sponge is described as a ranging from a large cup shape to irregular cups forming branches. A second batch of 3570 g of the Discodermia sp. collected at a depth of 157 m off Lucaya, Grand Bahama Island, furnished 1 (35 mg), 2-des-methyldiscodermolide15 (12, 1.0 mg, yield, 0.00003% wet weight) and 5-hydroxy methyldiscodermolate15 (13, 1.1 mg, yield, 0.00003% wet weight). This sponge sample was tan in color, and the morphology was a cluster of fingers. A third form of Discodermia sp. (2480 g) was described as thin, irregular anastomosing branches that are cream in color. It was collected from the same location as the latter sample and yielded 1 (25 mg) and 19-des- aminocarbonyldiscodermolide15 (14, 1.1 mg, yield, 0.00006% wet weight). A cyclic analog of 1, trivially named 9-13-cyclodiscodermolide15 (15), was isolated from a collection of Discodermia sp. made in May 1993 off Tartar Bank, south of Cat Island, Bahamas, at a depth of 183–198 m. This specimen was cup shaped and white to yellow in color, and the 2500 g sample furnished only 1.2 mg (yield, 0.00005% wet weight) of 9-13-cyclodiscodermolide (15). Copyright © 2005 CRC Press, LLC O R O Ph S HO O NH O N OH H O O H OH O O H O Ph O O HO O O OH NH2 OH O O OH O Ph O OH 3. Epothilone A, R = H 4. Epothilone B, R = Me 1. (+)-Discodermolide 2. Taxol OH H HO O H S O OH OH H N O O O N OH O N O O O O O O H H O O O OH O H H 5. Epothilone D 7. Laulimalide 6. Eleutherobin OH O H H HO N H O O OO OH H H N H O OH OH OH 8. Dictyostatin-1 9. Discodermide H HN O NH NH2 O O H O HN O H H H NH H H N N HO HN N N H N H H H H H O O O O O OH O NH2 HN SO O 3Na O R1 OH NH O H 2 R2 OH H O NH H Br H N N H O H O H O N N O 11. 2-Epi-descidermolide, R1 = Me, R2 = H H O H 12. 2-Des-methyldiscodermolide, R1 = H, R2 = H NH2 O 10. Polydiscamide A STRUCTURES 1–12 Copyright © 2005 CRC Press, LLC III. CHEMISTRY (+)-Discodermolide (1) is a polyhydroxy dodecaketide δ-lactone presumably biosynthesized by a combination of four acetate and eight propionate units. Structurally, it bears 13 stereo centers with 15 rotatable σ-bonds. The δ-lactone ring is tetrasubstituted, and the side chain contains one di- (C8) and one trisubstituted (C13) Z-double bonds, as well as a terminal Z-diene system. Four of the five free secondary hydroxyl groups are located at C3, C7, C11, and C17 positions, and the fifth hydroxyl group is functionalized to a pendant carbamate group. Both in the solid state (as evident from the crystal structure1) and in the solution state (determined by NOE studies in conjunction with Monte Carlo conformational studies16), the structure of 1 exists in a helical conformation, arranging the C1–C19 region into a U-shaped conformation that brings the δ-lactone and the C19 pendant carbamate group in close proximity. The tetrasubstituted δ-lactone preferred a boatlike conformation. This adopted hairpin conformation, in which the two internal (Z)-olefins in the side chain act as conformational locks through minimization of (1,3) strain between their respective substituents in concert with the avoidance of syn-pentane interactions, results in the restricted mobility of the 19-carbon side chain. IV. SYNTHESIS A. SYNTHESIS OF DISCODERMOLIDE The first total synthesis of discodermolide was achieved by the Schreiber group at Harvard University, and their synthesis resulted in the preparation of (–)-discodermolide17 based on the stereochemistry reported in the first publication.1 Subsequently, this group, using the identical route to synthesize the natural antipode (+)-discodermolide (1), established the absolute configuration.4,18 Their synthetic plan was adapted on a disconnection of the 24-member carbon chain into three roughly equivalent segments. The δ-lactone-end fragment C1–C7 was coupled to the middle fragment C8–C15, and the carbon skeleton was completed by attachment of a fragment C15–C24 through the use of enolate alkylation chemistry. Since then, Smith et al.11,19 and Myles et al.20,21 have prepared both antipodes of discodermolide. Marshall and Jones22 and Paterson et al.23 have synthesized (+)-discodermolide. All these groups used similar triply convergent approaches with different coupling and reaction conditions to achieve the total synthesis. A review by Kalesse24 described the various approaches used by these groups to synthesize small quantities of discodermolide. In 1999, the Smith group improved on their previously published method19 and synthesized 1 in 6% overall yield, a procedure that was amenable to gram-scale production.11,25 They improved the triply convergent approach by synthesizing the three subunits from a common precursor, using a modified Negishi coupling using alkyl zinc species, synthesis of the phosphonium salt via ultrahigh pressure, and finally, using a chemoselective Wittig olefination reaction to give the complete carbon chain. Recently, the Novartis group, using a combination of methods developed by Smith, Paterson, and their own research group, synthesized 60 g of clinical grade 1, and this material is now being used in the phase I clinical trials at the Cancer Research Center in San Antonio, Texas. This large- scale synthesis of clinical grade 1 involved 39 steps, with the longest linear sequence having 26 steps and 17 large-scale chromatographic purifications. The Novartis group modified the synthesis of the Smith group Weinreb amide25 that contains the methyl-hydroxy-methyl stereo triad that is repeated in three discodermolide fragments C1–C6, C9–C14, and C15–C21 for pilot plant produc- tion.26 The total synthesis was recently described in a five-part series.26–30 The final paper describes the linkage of the methylketone-containing C1–C6 fragment and the C7–C24 aldehyde fragment using a reagent-controlled stereoselective boron enolate aldol coupling procedure developed by Paterson’s group.31 The total synthesis by the Novartis group of 60 g of crystalline (+)-discoder- molide using good manufacturing processes represents a significant milestone in both natural product drug development and organic synthesis. Copyright © 2005 CRC Press, LLC B. SYNTHESIS OF ANALOGS The Schreiber group was the first to prepare discodermolide analogs for structure–activity relation- ship (SAR) studies.4,18 They prepared C16 (16) and C17 (17) epimers, 16-normethyl (18), epimers of C1 thiophenyl (19, 20), and C1-β thiophenyl 16-normethyldiscodermolide (21). Other analogs prepared by the Schreiber group are the C17 tritium-labeled discodermolide enantiomers (22, 23), the Fragments A (24) and B (25), C17-keto (26), pentyl-NHFmoc ester (27), and C24-hydroxyl- butylester (28). The last two analogs and several similar analogs were prepared to study the binding properties of (+)-discodermolide.18 The Paterson group23 synthesized 5-epi-, 7-epi-, and 5,7-bis- epi-discodermolides (29–31) during their aldol-based total synthesis of (+)-discodermolide. They selectively accessed both C7 epimers in the final C6–C7 aldol coupling, which allowed the synthesis of the above epimers by stereocontrolled reduction of the C5 ketone and subsequent deprotection. The Schreiber group4 observed that changes in the C17–C20 region of the discodermolide backbone resulted in loss of activity, and therefore Smith’s group32 focused their modifications on the C1–C14 segment of 1 and prepared four analogs including 3-deoxy-discodermolide-2-ene (32), 3,7-dideoxy- discodermolide-2-ene (33), 8,9-E-discodermolide (34), and 14-normethyldiscodermolide (35) for SAR studies. Using a small quantity of natural (+)-discodermolide isolated from sponge samples, Gunasekera et al.33 prepared eight acetylated discodermolide analogs (36–43) to study the importance of hydrogen bonding by the hydroxyl groups in the molecule on the cytotoxicity and microtubule stabilizing activity. These analogs were prepared by treating discodermolide with acetic anhydride in pyridine at a controlled temperature and by separating the resulting acetylated mixture by high- performance liquid chromatography methods. Subsequently, the same group has prepared 3-deoxy- discodermolide-2-ene (32) and four acetylated analogs of 3-deoxy-discodermolide-2-ene (44–47), using acetylated analogs of 1 as the starting material.34 Temperature- and time-controlled catalytic hydrogenation of 1 resulted in three hydrogenated discodermolide analogs (48–50) and one 7- deoxy-hexahydro analog (51), which on acetylation, yielded the 7-deoxy-hexahydro-triacetate (52). These compounds were used to study the importance of the double bonds in contributing to the activity. In addition, one 6,7-seco-analog (53) has been generated during the attempted allylic oxidation of the C7 hydroxyl group in 1, using MnO2 as an oxidizing agent. This compound 53 on stirring with titanium(IV) isopropoxide in CH2Cl2 at room temperature for 2 weeks furnished the 19-hydroxy-6,7-seco-analog (54). These compounds were used to ascertain the contribution of the C1–C7 unit on the activity.34 During the final stages of the 60-g synthesis of 1 conducted by the Novartis group,30 the three t-butyldimethylsilanyl protecting groups in (55) were removed by a one-step acid catalyzed pro- cedure. This hydrolysis procedure with concomitant lactonization gave crude 1 in 85% purity and quite a number of various combinations of silyl ether cleavage side products and several cyclized side-products in trace quantities. These minor cyclized side-products (56–61) have been purified, and their structures were determined by nuclear magnetic resonance studies. The formation of five- cyclopentyl (56–60) and one-pyrano (61) analogs indicated that the hairpin confirmation of 1 in solution brings the two isolated double bonds closer together in correct geometry for the proton catalyzed cyclization.35 V. BIOLOGICAL ACTIVITY A. IMMUNOSUPPRESSIVE PROPERTIES (+)-Discodermolide (1) could be considered a unique compound that exhibits both strong immun- osuppressive activity and strong antiproliferative activity.1 Initial testing of 1 indicated potent immunosuppressive activity in the bi-directional murine mixed lymphocyte assay (MLR) and antiproliferative activity toward P388 murine leukemia cells. Further studies by Longley et al. Copyright © 2005 CRC Press, LLC HO 19 HO H O O O OH OH HO OH O O O OH OH NH2 OH OH 14. 19-Des-aminocarbonyldiscodermolide 13. 5-Hydroxymethyldiscodermolate O OH O NH R 2 HO 17 H 16 O O O O OH O O 13 OH NH2 9 OH OH OH 16. R = Me, (C16R, C17R) 17. R = Me, (C16S, C17S) 15. 9-13-Cyclodiscodermolide 18. R = H, (C17R) R HO HO H H PhS O OH O O O T OHO 1 O O OH NH OH NH 2 2 OH OH 19. β-PhS, R = Me 22. 3H-(+)-Discodermolide 20. α-PhS, R = Me 21. β-PhS, R = H HO OH H HO O O H O O T OHO OH O OH NH OH 2 OH 24. Fragment A 23. 3H-(–)-Discodermolide HO H O O O O OH O O O OH NH2 OH NH2 OH 25. Fragment B 26. (+)-17-Oxo-discodermolide

STRUCTURES 13–26 Copyright © 2005 CRC Press, LLC HO HO H O O H OH OR O O OH O H N OH OH 2 O R OH OH 27. R = CO(CH2)5NHFmoc 28. R = –(CH2)4OPv R HO H O O OH O H O O OH O O O OH NH2 OH NH2 OH OH 29. 5-Epi-discodermolide, R = αOH 31. 5,7-Bis-epi-discodermolide, R = βOH 30. 7-Epi-discodermolide HO H H O O OH O O O OH O O O OH NH2 OH NH2 33. 3,7-Dideoxy-discodermolide-2-ene 32. 3-Deoxy-discodermolide-2-ene HO HO H H O O OH O O O OH O O O OH NH2 OH NH2 OH OH 35. 14-Normethyldiscodermolide 34. 8,9-E-Discodermolide R2O H O O OR4 O O OR3 NH2 OR1 36. Discodermolide-3-acetate, R1 = COCH3, R2 = R3 = R4 = H 37. Discodermolide-7-acetate, R2 = COCH3, R1 = R3 = R4 = H 38. Discodermolide-3,7-diacetate, R1 = R2 = COCH3, R3 = R4 = H 39. Discodermolide-3,11-diacetate, R1 = R3 = COCH3, R2 = R4 = H 40. Discodermolide-3,17-diacetate, R1 = R4 = COCH3, R2 = R3 = H 41. Discodermolide-3,7,11-triacetate, R1 = R2 = R3 = COCH3, R4 = H 42. Discodermolide-3,7,17-triacetate, R1 = R2 = R4 = COCH3, R3 = H 43. Discodermolide-3,7,11,17-tetraacetate, R1 = R2 = R3 = R4 = COCH3 STRUCTURES 27–43 Copyright © 2005 CRC Press, LLC R1O HO H O O H OR3 O O O OH O O O OR2 NH OH NH 2 2 44. 3-Deoxy-discodermolide-2-en-11-acetate, R1 = R3 = H, R2 = COCH OH 3 45. 3-Deoxy-discodermolide-2-en-17-acetate, R1 = R2 = H, R3 = COCH3 48. 21,23-Tetrahydrodiscodermolide 46. 3-Deoxy-discodermolide-2-en-11,17-diacetate, R1 = H, R2 = R3 = COCH3 47. 3-Deoxy-discodermolide-2-en-7,11,17-triacetate, R1 = R2 = R3 = COCH3 HO HO H H O O OH O O O OH O O O OH NH OH NH 2 2 OH OH 49. 8,21,23-Hexahydrodiscodermolide 50. 8,13,21,23-Octahydrodiscodermolide H O O OH O H O O O OAc O OH NH O 2 OAc NH2 OH OAc 51. 7-Deoxy-8,21,23-hexahydrodiscodermolide 52. 7-Deoxy-8,21,34-hexhydrodiscodermolide- 3,11,17-triacetate OH OR O O O H N HO RO O O O OR NH2 53. 6,7-Seco-discodermolide, R = CONH2 OR 54. 19-Des-aminocarbonyl-6,7-seco-discodermolide, R = H 55. R = Si(CH3)2C(CH3)3 O O O O OH O O OH O OH OH O NH2 OH OH NH2 56. 7-Deoxy-7, 14-dien-9-13-cyclodiscodermolide 57. 7-Deoxy-7, 14(29)-dien-9-13-cyclodiscodermolide OH O O O O O OH O OH O O O NH OH OH NH 2 2 OH 58. 14(R)-hydroxy-16(S)-9-13-cyclodiscodermolide 59. 14(S)-Hydroxy-16(S)-9-13-cyclodiscodermolide 61. 7-(11)-Oxycyclodiscodermolide 60. 14(S)-Hydroxy-16(R)-9-13-cyclodiscodermolide STRUCTURES 44–61 Copyright © 2005 CRC Press, LLC demonstrated immunosuppressive activity in a number of in vitro assays,2,36 and in an in vivo experimental model of graft-versus-host reaction,3 the Simonsen splenomegaly assay.37 These studies indicated that 1 suppresses the proliferative responses of splenocytes in the murine two- way mixed lymphocyte reaction, and concavalin A (Con A) stimulated cultures, with IC50 values of 0.24 µM and 0.19 µM, respectively, with no detection of cytotoxicity for murine splenocytes at concentrations of 1 as high as 1.26 µM. In further studies, 1 suppressed the proliferative responses of human peripheral blood leukocytes (PBL) in the two-way mixed lymphocyte reaction, Con-A and phytohemagglutinin mitogenesis assays, with IC50 values of 5.65 µM, 28.02 µM, and 30.12 µM, respectively, with no evidence of cytotoxicity toward human PBL at (+)-discodermolide concentrations as high as 80.64 µM.3 (+)-Discodermolide was more potent than the clinically used cyclosporin A, a cyclic decapeptide (CsA) isolated from a fungus Trichoderma inflatum,38 in its ability to suppress the PMA –ino- mycin-induced proliferation of purified murine T-cells. The IC50 values observed in this assay were 9.0 and 14.0 µM for 1 and CsA, respectively. 1 had no effect on interleukin 2 (IL-2) production or IL-2 mRNA expression; however, the expression of the low-affinity IL-2 receptor in PMA–ion- omycin-stimulated, purified, murine T-lymphocytes, and that of PHA-stimulated human PBL blasts, was affected. In vivo studies by Longley et al.3 reported that 1 is effective in the graft-versus-host splenomegaly response of BALB/c → CB6F1 (BALB/c × C57Bl/6J)F1 grafted mice. Mice that were administered daily, intraperitoneal injections with dose levels of 1.25, 0.625, and 0.313 mg/kg remained healthy after 7 days on the regimen and continued to demonstrate suppression of sple- nomegaly with 106%, 72%, and 76% suppression, respectively. Splenocytes obtained from (+)-discodermolide-treated allogeneic grafted mice were suppressed in their ability to respond in vitro to optimal mitogenic concentrations of Con A, and natural killer cell activity directed against YAC-1 tumor cells, compared with vehicle-treated, allogenic grafted control mice. Lower dosages of 1, however, did not affect the subsequent ability of splenocytes obtained from these mice to produce IL-2 following in vitro stimulation with Con A. The unusual in vivo immunosuppressive properties of 1 as compared to CsA indicated that 1 has a specific in vivo mechanism of action. Using the murine DO11.10 T-hybridoma cell line, Longley et al.36 reported that the antiproliferative activity of 1 was a result of its ability to block cell proliferation in the G2/M phase of the cell cycle. Control cultures of DO11.10 cells demonstrated a characteristic pattern of cell cycling. Approximately 68% of the control cells make up the G1 phase of the cell cycle, and approximately 31% of the cells make up the S phase. Only 1% of these cells are demonstrable in the G2/M phase in asynchronous cultures at any one time, mainly because of the high proliferative rate of DO11.10 cells. At 24 h post–culture initiation, after treatment with 1.0 µg/mL 1, the percentage of cells in G1 phase decreased to 25%, and those in S to 16%. However, the percentage of cells in G2/M increased to 58%. The results from this study concluded that 1 acts by blocking cells in the G2/M phase of the cell cycle.36 B. ANTITUMOR PROPERTIES (+)-Discodermolide (1) was originally isolated by Gunasekera et al.1 from the marine sponge Disco- dermia dissoluta, following a cytotoxicity-based bioassay-guided approach. The crude ethanol extracts of the source sponge initially revealed strong antiproliferative activity toward P388 murine leukemia cells in addition to potent immunosuppressive activity in the bidirectional murine mixed lymphocyte assay. (+)-Discodermolide was evaluated for its antiproliferative effects in the National Cancer Insti- tute’s (NCI’s) 60–human tumor cell-line screen.39 The results of the NCI analyses were reported as growth inhibition 50 (GI50) values, in which the growth of 50% of the cells is inhibited; total growth inhibition (TGI) values, in which the growth of all cells is inhibited (i.e., cytostasis); and lethal concentration 50 (LC50), in which 50% of the cells are killed. 1 demonstrated selective cytotoxicity for 32 cell lines for GI50 and 18 cell lines for TGI. According to LC50 measurements for the NCI cell lines, 1 was particularly effective against human non–small cell lung cancer (NCI-H23), human colon Copyright © 2005 CRC Press, LLC cancers (COLO 205 and HCC-2998), human melanoma (M14), two of six central nervous system cancer cell lines (SF 295 and SF 539), and two of eight breast cancer cell lines (MDA-MB-435 and MDA-N). A statistical analysis of this selective toxicity pattern (the COMPARE algorithm)40 indicated that the cytotoxicity pattern of 1 matched that of several previously tested microtubule interactive agents, including the pattern generated by Taxol® (2). Computer-assisted structure–activity analysis by Ernest Hamel at NCI predicted antimitotic properties.41 This finding led to the collaboration of the HBOI group with Billy Day of the University of Pittsburgh and Ernest Hamel of the NCI to begin a study on the antimitotic properties of (+)-discodermolide (1). The group found41 that 1 was equally effective as Taxol (2) in inhibiting the growth of two human breast carcinoma cell lines, estrogen receptor–positive MCF-7 cells (IC50 = 2.4 nM for 1 and 2.1 nM for 2), and estrogen receptor–negative MDA-MB-231 cells (the IC50 values are not reported). Complete inhibition of growth occurred with 10 nM or greater of each drug and was not reversed by removal. (+)- Discodermolide (1) treated cells exhibited highly fragmented condensed nuclei. Flow cytometric studies of cells treated with 1 or 2 at 10 nM showed that both compounds caused cell cycle perturbation and induction of a hypodiploid cell population. 1 caused these effects more extensively and at an earlier time point. Immunofluorescence patterns of cells treated with 1 revealed remarkable rearrangement of the microtubule network, demonstrating promotion of microtubule assembly and indicating a Taxol-like effect (Figure 10.1). FIGURE 10.1 (See color insert) Immunofluorescence images of A549 cells stained with anti-α-tubulin (green) and propidium iodide (red) and observed by confocal microscopy. Cells were exposed to (A) 0.05% ethanol (vehicle control), or (B) (+)-discodermolide at a concentration of 100 nM. Copyright © 2005 CRC Press, LLC 1 was also analyzed for its interactions with purified tubulin. It was found that 1 induced the polymerization of tubulin to form microtubules in both the presence and absence of microtubule- associated proteins and guanosine 5′-triphosphate (GTP), and the resulting microtubules induced polymerization at lower temperatures than 2 and were more stable to the cold (4˚C) than those formed by 2. (+)-Discodermolide-induced polymer differed from Taxol-induced polymer in that it was completely stable at 0˚C in the presence of high concentrations of Ca++ ions. In a quantitative assay designed to select for compounds more potent than Taxol (2) in its ability to polymerize tubulin, 1 showed an EC50 value of 3.2 µM, versus 23 µM for 2.41 Hung et al.4,42 in the Schreiber group reported similar results using synthetic (+)-discodermolide. Their studies demonstrate similar arrest of MG63 human osteosarcoma cells in the M-phase of the cell cycle, induction of polymer- ization of purified tubulin, and “bundling” of the microtubule network in Swiss 3T3 cells. They discovered that the binding of microtubules by 1 and 2 were mutually exclusive in competitive binding assays (i.e., 1 was capable of displacing 2 and vice versa in binding to microtubules). 1 was also found to bind to microtubules with a much higher affinity than 2. Longley et al.39 at HBOI continued further studies on the effect of 1 on cell cycle progression and apoptosis of A549 human lung and GI-101A human breast cells in comparison to Taxol (2). The GI-101A cell line used by the Longley group was provided by Dr. Josephine Hurst, Goodwin Cancer Center, Plantation, Florida, and is an adherent breast cancer cell line derived from the original parent mammary tumor xenograft cell line, GI-101.43 The line was originally derived from a local first recurrence of an infiltrating ductal adenocarcinoma (Stage IIIa, T3N2MX) in a 57-yr-old female who had previously not received any chemotherapy or radiation therapy. Tumors transplanted into murine hosts are positive for normal human breast tumor markers and for the p53 antigen but are negative for c-erbB-2 oncogene. The A549 human lung and GI-101 human breast cells were incubated at 37˚C in 5% CO2 in the presence or absence of 10 nM 1 or 2 for 24 h. Cells were harvested, fixed in ethanol, and stained with 0.5 mg/ml of propidium iodide together with 0.1 mg/ml of RNAse A. Stained preparations were analyzed on a Coulter EPICS ELITE flow cytometer with 488 nM excitation, and the resulting DNA histograms were collected from at least 10,000 propidium iodide (P.I.) stained cells at an emission wavelength of 690 nM. Raw histogram data were further analyzed using a cell cycle analysis program. A549 and GI-101 cells without the drug exhibited a typical pattern of cell cycling, with a large percentage of the cell population comprising the G1 population, with lesser percentages comprising the S and G2/M phases of the cell cycle, as shown in Figure 10.2. Cells treated with 10 nM 1 or 2 for 24 h exhibited decreased percentages of cells comprising the G1 population, with corresponding increased per- centages in both S and G2/M phases of the cell cycle, indicating the ability of 1 and 2 to induce G2/M block, as shown in Figure 10.3. Cells undergoing apoptosis are also evident as a peak immediately to the left of the G1 peak. A5000002.HST PMTS A5control 10.24.95 GI000017.HST PMTS GIcont.24.95 640 640 560 66% A549 Lung GI-101A Breast Control 560 Control 480 480 51% 400 400 320 320 240 240 160 160 12% 80 28% 6% 80 37% 0 0 0 40 80 120 160 200 240 280 0 40 80 120 160 200 240 280 DNA Content DNA Content FIGURE 10.2 Cell cycle analysis of A549 and GI-101

control cells. Copyright © 2005 CRC Press, LLC Cell Number Cell Number A5000014.HST PMT5 A510nm disco 10.24.95 A5000008.HST PMT5 A510nm taxol 10.24.95 200 280 A549 Lung A549 Lung 10nM 240 10 nM 160 Discodermolide Taxol 200 25% 34% 120 160 45% 49% 80 Apoptosis 27% 120 Apoptosis 24% 80 40 43% 48% 40 0 0 0 40 80 120 160 200 240 280 0 40 80 120 160 200 240 280 DNA Content DNA Content GI000122.HST PMT5 GItax10nm 10.18.95 GI000114.HST PMT5 GIdisco10nm 10.18.95 200 240 GI-101A Breast A549 Lung 22% 10 nM 200 19% 10nM 160 Taxol Discodermolide 160 24% 120 27% 40% 120 22% Apoptosis Apoptosis 59% 80 54% 80 40 40 0 0 0 40 80 120 160 200 240 0 40 80 120 160 200 240 DNA Content DNA Content FIGURE 10.3 Cell cycle analysis of A549 and GI-101A treated with (+)-discodermolide (1) and Taxol (2). (+)-Discodermolide (1) indicated strong cytotoxicity in all cancer cell lines tested in vitro, and the activity was comparable to that of Taxol. The results are shown in Table 10.1. However, it was interesting to note that 1 was less cytotoxic to the normal kidney epithelial cell line of the African green monkey. These results indicated that 1 selectively targets dividing cancer cells, and thus suggested microtubule stabilization is probably the only mechanism of action of (+)-discodermolide. TABLE 10.1 In Vitro Cytotoxicity of (+)-Discodermolide (1) in Selected Cancer Cell Lines Cell Line Activity Data (IC50 in µM) Murine P388 leukemia cells 0.033 Human A549 lung cancer cell line 0.015 Human MCF-7 breast cancer cell line 0.0024* Human MDA-MB-231 breast cancer cell line 0.0024* Human NCI/ADR 0.017* PANC-1 0.049 VERO 30 For Taxol* 0.0021 A549 (human lung adenocarcinoma cell line); P388 (cultured murine leukemia cell line); MCF-7 (human breast cancer cell line); NCI/ADR [adriamycin drug resistant cell line]; PANC-1 (human pancreatic cancer cell line); VERO (normal kidney epithelial cell line of African green monkey). Copyright © 2005 CRC Press, LLC Cell Number Cell Number Cell Number Cell Number Longley’s group at HBOI continued studies on the characterization of the molecular basis of apoptosis in (+)-discodermolide-treated tumor cells.39 This study was designed to detect and char- acterize the (+)-discodermolide-induced molecular pathways of apoptosis and to evaluate the differences with Taxol-induced apoptosis, by probing Bcl-2, Bcl-Xl, and Raf-1 proteins in JURKAT and MCF-7 cell lines by Western blot analysis, and following the phosphorylation patterns in response to 1 and 2 treatments. The HBOI group detected that both 1 and 2 phosphorylate the Bcl- 2 family proteins, and this appears to be accompanied by a loss of Bcl-2 function. They also measured the apoptotic responses by detecting and quantitating DNA fragmentation in various cell populations (MCF-7, DO11.10, and JURKAT) under the treatment of drugs, such as Fas-antigen, dexamethasone, Taxol (2), and (+)-discodermolide (1), using a time course (6, 12, 24 h) and concentration-dependent (10, 100, 1000 nM of drug) experiments. The DNA fragmentation analysis, correlated by flow cytometric studies and by microscopy, indicated that 1 induces apoptosis in a dose-dependent manner and appears to do so in the same way as 2. Also, it appears that the dose range for inducing these apoptotic events is more or less the same depending on the cell line used. In comparative studies of 1, with the epothilones (3, 4) and eleutherobin (6) against a Taxol- dependent human lung carcinoma cell line A549-T12, Martello et al.32 found that 1 was unable to act as a substitute for 2, whereas 3, 4, and 6 were able to maintain the viability of the cell line. The presence of low concentrations of 2 amplified the cytotoxicity of 1 20-fold against this cell line and thus suggested that 1 and 2 may constitute a promising chemotherapeutic combination.44 VI. STRUCTURE–ACTIVITY STUDIES The Schreiber group was the first to synthesize the unnatural antipode, (–)-discodermolide and a number of structural analogs, thus laying the ground work for SAR analysis.17,18 Both enantiomers exhibited antiproliferative activity, but at different stages of the cell cycle. The natural (+)-disco- dermolide (1) blocks the cells in the G2/M phase, whereas the (–)-discodermolide inhibits the cell cycle in the S-phase. The two antipodes also showed significantly different potency in cytotoxicity- based assays. In a [3H]-thymidine incorporation assay using NIH3T3 cells, synthetic (1) had an IC50 value of 7 nM, compared to135 nM for the ()-discodermolide, or a 19-fold loss of activity.4,18 Removal of the C16S methyl group (e.g., 16-normethyldiscodermolide [18]) did not affect the potency. However, epimerization at C16 (e.g., C16R [16]) or at C17 (e.g., C16S, C17S [17]) showed a significant reduction in activity (IC50 = 300 and >300 nM, respectively). Interestingly, the thiophenyl derivatives (19, 20), both α and β anomers, as well as the 16-normethyl analog (18) were equally active as 1, indicating that the carbonyl functionality and C16Me are not essential for the cytotoxicity of 1. The other analogs prepared by the Schreiber group,18 the left and right fragments of 1 (24, 25), 17-oxo-discodermolide (26) and pentyl-NHFmoc ester (27) and 24- hydroxybutyl ester (28), retain a low level of activity of IC50 = 70 nM in the [3H]thymidine incorporation assay using MG63 cells. The three epimeric analogs 5,7-bis-epi-, 5-epi-, and 7-epi- discodermolides (29–31) prepared by the Paterson group23 are structurally interesting; however, the biological activities of these compounds were not presented in the synthetic publication. The Smith group32 prepared four interesting discodermolide analogs, 3-deoxy-discodermolide- 2-ene (32), 3,7-dideoxy-discodermolide-2-ene (33), 8,9-E-discodermolide (34), and 14-normethyl- discodermolide (35) — during their total synthesis of 1, which used a highly efficient, triply convergent approach. This was the first report that indicated that microtubule stabilization by 1 is extremely sensitive to simple modifications in the molecular structure. The researchers’ in vitro studies demonstrated that 1 and the structural analogs induced tubulin assembly in the absence of GTP and that the microtubules formed were stable under Ca++-induced depolymerizing conditions. The initial rates of polymerization reflect the nucleation process. It was observed that the rate of tubulin nucleation of the four analogs (32–35) was reduced as compared to that of 1. Analogs 32 and 35 showed similar levels of polymerization to that of 1, and analogs 33 and 34 demonstrated reduced polymerization, attaining approximately 50% of the level of polymerization observed with Copyright © 2005 CRC Press, LLC TABLE 10.2 Biological Data of Taxol (2), (+)-Discodermolide (1) and Discodermolide Acetates (36–43) P388 A549 G2/M Block Microtubule Bundling Compound IC50 nM IC50 nM 100 nM (nM) Taxol (2) 7.7 — + Yes (100) (+)-Discodermolide (1) 35 34 + Yes (10) 3-Acetate (36) 9 9.4 + Yes (10) 7-Acetate (37) 2.5 0.5 + Yes (10) 3,7-Diacetate (38) 0.7 3.7 + Yes (10) 3,11-Diacetate (39) 103 295 — No (100) 3,17-Diacetate (40) 1149 736 — No (100) 3,7,11-Triacetate (41) 166 554 — No (100) 3,7,17-Triacetate (42) >6825 >1307 — No (100) 3,7,11,17-Tetraacetate (43) 837 1307 — No (100) (+)-discodermolide (1). Electron microscopy examination of the microtubules formed in the in vitro assays confirmed that normal microtubules were formed in the presence of the compounds. The microtubules that assembled in the presence of 1 were approximately fivefold shorter than those formed by analogs (32) and (35); 12-fold shorter than those formed by analogs (33) and (34), and fourfold shorter than those formed by Taxol (1). The same group showed that in a drug-binding competition assay, 1 and analogs 32, 33, and 35 are competitive inhibitors of [3H]Taxol binding to preformed microtubules. At 1 µM concentration, Taxol (2), (+)-discodermolide (1), and analogs (32, 33, 35) exhibited very similar inhibition, whereas analog 34 was essentially inactive in displacing [3H]Taxol binding. At 100 µM, 1 displayed 95% inhibition, 2 displayed 80% inhibition, and analog (32) exhibited >90% inhibition of [3H]Taxol binding to preformed microtubules. Analogs (33–35) demonstrated a range of inhibition between 45 and 65% at 100 µM. Using natural (+)-discodermolide (1) isolated from the sponge Discodermia, Gunasekera et al.34 in the Harbor Branch group prepared eight acetylated analogs (36–43) by acetylation of the hydroxyl groups at C3, C7, C11, or C17. These analogs were assayed for biological activity in cultured murine P388 leukemia and human A549 lung adenocarcinoma tumor cell lines to determine the structural requirements for tubulin interaction and cytotoxic effects. The biological data are presented in Table 10.2. The P388 and A549 assay results indicated that acetylation of the hydroxyl groups in the left-hand side of the molecule at C3 and C7 confers a greater cytotoxicity to the discodermolide structure, as seen with analogs (36–38). Interestingly, the discodermolide acetates (39, 41) with acetyl groups at C11 showed reduced cytotoxicity as compared to the parent molecule, whereas compounds (40, 42, 43) that include acetylation at C17 caused a dramatic reduction in cytotoxicity. These results confirmed the previous finding by the Schreiber group18 that acetylation at C17 reduced the activity by tenfold in the [3H]-thymidine incorporation assay, using MG63 cells. Cell cycle analysis by flow cytometry using A549 cells revealed that the highly cytotoxic analogs caused the accumulation of cells in the G2/M phase. The effect on microtubule bundling follows the cytotoxicity as seen by analogs (33–38). The HBOI group recently isolated and reported three natural discodermolide analogs (11–13) that showed modification in the lactone ring system of the molecule. In addition the group prepared through semisynthesis 3-deoxy-discodermolide-2-ene (32) and four 3-deoxy-2-ene acetylated analogs (44–47) for SAR studies. The activity data reported in Table 10.3 indicate that the ring- opened methyl ester (13) resulted in two- and fivefold reduction in activity in the P388 and A549 assays, respectively. Epimerization at C2 (11) results in a fivefold reduction in cytotoxicity and a reduction in the level of tubulin polymerization. Dehydration of the 3-hydroxyl functionality to Copyright © 2005 CRC Press, LLC TABLE 10.3 Biological Data of Natural Analogs (11–13) and 3-Deoxy-Discodermolide-2-ene (32) and Analogs (44–47) G2/M Microtubule P388, A549, Block, Bundling, Compound IC50 nM IC50 nM 100 nM 10 nM 2-Epi-discodermolide (11) 134 67 +Yes + 2-Des-methyldiscodermolide (12) 172 120 +Yes +++ 5-Hydroxymethyldiscodermolate (13) 65 74 +Yes + 3-Deoxy-discodermolide-2-ene (32) 34 21 +Yes +++ 3-Deoxy-discodermolide-2-en-11-acetate (44) 519 519 No — 3-Deoxy-discodermolide-2-en-17-acetate (45) 8103 8103 No — 3-Deoxy-discodermolide-2-en-11,17-diacetate (46) 7587 7587 No — 3-Deoxy-discodermolide-2-en-7-11-17-triacetate (47) 7076 7133 No — yield 3-deoxy-discodermolide-2-ene (32) results in no loss of biological activity compared to that of the parent (+)-discodermolide (1). Interestingly, the 2-des-methyldiscodermolide (12) showed greater than fivefold reduction in cytotoxicity, yet the microtubule bundling properties were comparable to 1 and 32. The analogs (11–13, 32) showed cell cycle arrest at the G2/M checkpoint and induction of microtubule bundling. These SAR studies confirmed that minor changes in the functional groups on the lactone ring do not significantly affect the antimitotic properties of (+)- discodermolide (1). The acetylated analogs of 3-deoxy-discodermolide-2-ene (44–47) showed a significant reduction in cytotoxicity and the complete loss of microtubule-stabilizing properties on acetylation of the C11 or C17 hydroxyl groups, indicating the importance of these two hydroxyl groups for the microtubule stabilizing properties of (+)-discodermolide (1). As shown in Table 10.4, both the 21,23-tetrahydrodiscodermolide (48) in which the terminal conjugated double bond has been hydrogenated and 8,21,23-hexahydrodiscodermolide (49) have similar activity to the parent (+)-discodermolide (1), indicating that the C8 double bond and the terminal diene are not essential for the cytotoxicity and tubulin polymerizing action of (+)-disco- dermolide. The 8,13,21,23-octahydrodiscodermolide (50) shows significantly less cytotoxicity (8.2 µM) and no effect on tubulin polymerization. The retention of the microtubule bundling properties in 21,23-tetrahydrodiscodermolide (48) and 19-des-aminocarbonyl-discodermolide (14) indicated that the terminal conjugated double bond and the C19-carbamate groups are not essential for the microtubule bundling properties of (+)-discodermolide. Further saturation of 48 to 49 showed a TABLE 10.4 Biological Data of Hydrogenated Analogs (48–52) and 19-Des-Aminocarbonyldiscodermolide (14) G2/M Microtubule P388, A549, Block, Bundling, Compound IC50 nM IC50 nM 100 nM 10 nM 21,23-Tetrahydrodiscodermolide (48) 10 8 Yes ++ 8,21,23-Hexahydrodiscodermolide (49) 33 68 Yes ++ 8,13,21,23-Octahydrodiscodermolide (50) 8292 8292 No — 7-Deoxy-8,21,23-hexahydro-discodermolide (51) 309 377 No ++ 7-Deoxy-8,21,23-hexahydro-discodermolide-3,11,17- triacetate(52) 7052 NT No — 19-Des-aminocarbonyldiscodermolide (14) 128 74 Yes ++ Copyright © 2005 CRC Press, LLC TABLE 10.5 Biological Data of 6,7-Seco-Discodermolide Analogs (53, 54) and Cyclodiscodermolide Analogs (56–61) G2/M Microtubule P388, A549, Block, Bundling, Compound IC50 nM IC50 nM 100 nM 10 nM 6,7-seco-Discodermolide (53) 3686 6774 No — 19-des-Aminocarbonyl-6,7-seco-discodermolide (54) 12787 12787 No — (7E,9S,13S,14E)-7-deoxy-7,14-dien-9-(13)- 2130 6640 No NT cyclodiscodermolide (56) (7E,9S,13S)-7-deoxy-7,14(29)-dien-9-(13)- 190 100 NT NT cyclodiscodermolide (57) (7E,9S,13S,14R)-7-deoxy-14-hydroxy-7-en-9-(13)- 2420 840 NT NT

cyclodiscodermolide (58) (7E,9S,13S,14S, 16S)-7-deoxy-14-hydroxy-7-en-9- 2090 2640 NT NT (13)-cyclodiscodermolide (59) (7E,9S,13S,14S,16R)-7-deoxy-14-hydroxy-7-en-9- 7650 >5000 NT NT (13)-cyclodiscodermolide (60) (7R,11S)-7-(11)-Oxycyclodiscodermolide (61) 1000 1460 NT NT fivefold reduction in cytotoxicity in the A549 assay, but comparable activity in the P388 cytotoxicity assay and in the formation of microtubule bundles. These data indicated that the C8 Z-double bond is not essential for the biological activity of 1. Comparison of this finding with the biological activity reported earlier by the Smith group for the C8 E-analog (34) indicated that a C8 Z-double bond or C8 σ-bond type flexibility is required to maintain the high activity of (+)-discodermolide. The 7-deoxy-8, 21,23-hexahydrodiscodermolide (51) showed considerable reduction in cytotoxicity but, interestingly, retained the microtubule stabilizing properties of (+)-discodermolide. A remark- able loss of activity in the P388 and A549 assays on saturation of the C13 Z-double bond, as seen in the 8,13,21,23-octahydrodiscodermolide (50), indicated that the C13 Z-double bond, which partly controls the folding of the molecule, is essential for the cytotoxicity and microtubule stabilizing properties of (+)-discodermolide. As shown in Table 10.5, the reduction of biological activity in 6,7-seco-analogs (53, 54), C9- C13 cyclo-analogs (56–59) except 57, and C7-C11-oxycyclo analog (61) indicate that the combi- nation of the C7 and C11 hydroxy groups are contributing to the optimum biological activity of (+)-discodermolide (1). The C16(R)-analog 60 that has the opposite chirality to that of (+)- discodermolide showed the least cytotoxicity among all assayed cyclo-analogs. This observation is significant because of the earlier finding by the Schreiber group18 that C16(S) is essential for biological activity of (+)-discodermolide (1). The significant loss of biological activity in five of the six cyclo-analogs as shown in Table 10.5 that have the rigid ring structure from C6 through C13 could be attributed to the loss of the hairpin folding observed in (+)-discodermolide (1), and therefore, these observations confirmed the importance of the hairpin conformation for biological activity in (+)-discodermolide (1). These SAR studies confirmed that the C7, C11, and C17 hydroxy groups; the C16(S) stereochemistry; and the C13 Z-double bond that folds the molecule in to the hairpin conformation are essential for the strong microtubule stabilizing properties of (+)-disco- dermolide (1). A number of models have been proposed regarding the binding of (+)-discodermolide (1) to β-tubulin. Inhibition of binding of [3H]-Taxol to microtubules in the presence of 1 indicates that the binding sites are similar or overlap. The core structures (pharmacophores) of Taxol (2), (+)- discodermolide (1), epothilones (3–5), and eleutherobin (6) are all proposed to bind in a pocket Copyright © 2005 CRC Press, LLC formed by Gly368, Thr274, His227, and Asp224 in β-tubulin, and their side chains fit into a pocket formed by His 227 and Asp 224.45 Both the x-ray and solution structures of (+)-discodermolide (1) have demonstrated that the molecule arranges the C1-C19 region into a hairpin or U confor- mation, which brings the δ-lactone and C19 side chain into close proximity. Modeling studies indicate that the backbone of discodermolide mimics the northern region of Taxol (2), and in a preferred model for (+)-discodermolide (1) binding to the Taxol binding site, the δ-lactone and C19 side chain correspond to the C13 and C2 chains of Taxol, respectively.32 None of the models fully accounts for the differences in activity, such as temperature dependence and stability of polymerization or increased nucleation rates, observed between (+)-discodermolide (1) and Taxol (2). Other factors, such as the lack of cross resistance for (+)-discodermolide in the A549-T12 Taxol-resistant cell line and the lack of cross resistance for epothilone-resistant cell lines with class I β-tubulin mutations, indicate that (+)-discodermolide may bind an overlapping site on β-tubulin. Factors such as these, as well as the increased affinity of (+)-discodermolide for β-tubulin, may indicate the presence of additional discrete contact sites with the β-tubulin structure. For example, none of the models proposed to date addresses the near-complete loss of activity when the C17 hydroxyl is functionalized, indicating that this hydroxyl may be an additional contact or binding point with tubulin. VII. CLINICAL INVESTIGATION Based on preclinical studies in which (+)-discodermolide (1, XAA296A) stabilized microtubule polymers more potently than 2 and demonstrated prominent activity against Taxol-refractory xenograft models, 1 was taken into phase I trial in September 2002 at the Cancer Therapy and Research Center in San Antonio, Texas. In the trial, 1 is administered by intravenous infusion once every 22 d to adult patients with advanced solid malignancies. The study is ongoing, but at the time of writing, 21 patients (median age 59.4 yr; 12 males, 9 females) have been treated with the following dose levels (mg/m2): 0.6 (three patients), 1.2 (three patients), 2.4 (four patients), 4.8 (four patients), 9.6 (four patients), and 14.4 (three patients). Karnofsky performance status was as follows 70% (seven patients), 80% (eight patients), 90% (five patients), and 100% (one patient). The drug is well tolerated with no peripheral neuropathy or neutropenia. No objective responses have been observed; however, four of 21 patients have exhibited stable disease as their best response. Phar- macokinetics were evaluated in 18 patients receiving discodermolide over a dose range of 0.6–9.6 mg/m2. Following a short infusion, (+)-discodermolide blood concentrations had decreased to <10% of Cmax (29–460 ng/mL). Nonlinear kinetics, characterized by a second broad peak, were observed at the terminal phase of blood concentration–time profiles. In the majority of patients, the second rising curve started at 24 h postdose, with peak concentrations of 101%–208% of C24h at 48–72 h. Drug concentrations then gradually declined to 56%–173% of C24h at 168 h. The average (mean ± standard deviation) partial mean residence time (last measurable concentration) was 81.77 ± 43.02 h. Additional evaluation of patients at 14.4 mg/m2 dose level is in progress. Accrual will continue to define the recommended phase II dose.46 REFERENCES 1. Gunasekera, S.P. et al., Discodermolide: a new bioactive polyhydroxylated lactone from the marine sponge Discodermia dissoluta, J. Org. Chem., 55, 4912, 1990 (correction: 56.1346, 1991). 2. Longley, R.E. et al., Discodermolide: a new marine-derived immunosuppressive compound. I. In vitro studies, Transplantation, 52, 650, 1991. 3. Longley, R.E. et al., Discodermolide: a new marine-derived immunosuppressive compound. II. In vivo studies, Transplantation, 52, 656, 1991. Copyright © 2005 CRC Press, LLC 4. Hung, D. T., Nerenberg, J.B., and Schreiber, S.L., Distinct binding and cellular properties of synthetic (+)- and (-)-discodermolides, Chem. Biol., 1, 67, 1994. 5. Parness, J. and Horwitz, S.B., Taxol binding to polymerized tubulin in vitro, J. Cell Biol., 91, 479, 1981. 6. Bollag, P.A. et al., Epothilones, a new class of microtubule-stabilizing agents with a Taxol-like mechanism of action, Cancer Res., 55, 2325, 1995. 7. Hardt, I.H. et al., New natural epothilones from Sorangium cellulosum, strains Soce90/B2 and So ce90/D13: isolation, structure elucidation and SAR studies, J. Nat. Prod., 64, 847, 2001. 8. Lindel, T. et al., Eleutherobin, a new cytotoxin that mimics Paclitaxel (Taxol) by stabilizing micro- tubules, J. Am. Chem. Soc., 119, 8744, 1997. 9. Mooberry, S.L. et al., Laulimalide and isolaulimalide, a new paclitaxel-like microtubule-stabilizing agent, Cancer Res., 59, 653, 1999. 10. Isbrucker, R.A. et al., Tubulin polymerizing activity of dictyostatin-1, a polyketide of marine sponge origin, Biochem. Pharmacol., 66, 75, 2003. 11. Smith, III, A.B. et al., Gram-scale synthesis of (+)-discodermolide, Org. Lett., 1, 1823, 1999. 12. Kelly-Borges, M. et al., Species differentiation in the marine sponge genus Discodermia (Demospon- giae: Lithistida): the utility of ethanol extract profiles as species-specific chemotaxonomic markers, Biochem. Syst. Ecol., 22, 353, 1994. 13. Gunasekera, S.P., Gunasekera, M., and McCarthy, P., Discodermide: a new bioactive macrocyclic lactam from the marine sponge Discodermia dissoluta, J. Org. Chem., 56, 4830, 1991. 14. Gulavita, N.K. et al., Polydiscamide A: a new bioactive depsipeptide from the marine sponge Disco- dermia sp., J. Org. Chem., 57, 1767, 1992. 15. Gunasekera, S.P. et al., Five new discodermolide analogs from the marine sponge Discodermia sp., J. Nat. Prod., 65, 1643, 2002. 16. Smith, III, A.B., LaMarche, M.J., and Falcone-Hindley, M., Solution structure of (+)-discodermolide, Org. Lett., 3, 695, 2001. 17. Nerenberg, J.B. et al., Total synthesis of the immunosuppressive agent (-)-discodermolide, J. Am. Chem. Soc., 115, 12621, 1993. 18. Hung, D. T., Nerenberg, J.B., and Schreiber, S.L., Synthesis of discodermolides useful for investigating microtubule binding and stabilization, J. Am. Chem. Soc., 118, 11054, 1996. 19. Smith, III, A.B. et al., Total synthesis of (-)-discodermolide, J. Am. Chem. Soc., 117, 12011, 1995. 20. Harried, S.S. et al., Total synthesis of (-)-discodermolide: an application of a chelation-controlled alkylation reaction, J. Org. Chem., 62, 6098, 1997. 21. Harried, S.S. et al., Total synthesis of (+)-discodermolide, a microtubule-stabilizing agent from the sponge Discodermia dissoluta, J. Org. Chem., 68, 6646, 2003. 22. Marshall, J.A. and Jones, B.A., Total synthesis of (+)-discodermolide, J. Org. Chem., 63, 7885, 1998. 23. Paterson, I. et al., A practical synthesis of (+)-discodermolide and analogs: fragment union by complex aldol reactions, J. Am. Chem. Soc., 123, 9535, 2001. 24. Kalesse, M., The chemistry and biology of discodermolide, Chem. BioChem., 1, 171, 2000. 25. Smith, III, A.B. et al., Evolution of a gram-scale synthesis of (+)-discodermolide, J. Am. Chem. Soc., 122, 8654,. 26. Mickel, S.J. et al., Large-scale synthesis of the anti-cancer marine natural product (+)-discodermolide. Part 1: synthetic strategy and preparation of a common precursor, Org. Process Res. Dev., 8, 92, 2004. 27. Mickel, S. J. et al., Large-scale synthesis of the anti-cancer marine natural product (+)-discodermolide. Part 2: synthesis of fragments C1-6 and C9-14, Org. Process Res. Dev., 8, 101, 2004. 28. Mickel, S.J. et al., Large-scale synthesis of the anti-cancer marine natural product (+)-discodermolide. Part 3: synthesis of fragment C15-21, Org. Process Res. Dev., 8, 107, 2004. 29. Mickel, S.J. et al., Large-scale synthesis of the anti-cancer marine natural product (+)-discodermolide. Part 4: preparation of fragment C7-24, Org. Process Res. Dev., 8, 113, 2004. 30. Mickel, S.J. et al., Large-scale synthesis of the anti-cancer marine natural product (+)-discodermolide. Part 5: linkage of fragments C1-6 and C7-24 and finale, Org. Process Res. Dev., 8, 122, 2004. 31. Paterson, I. et al., Total synthesis of the antimicrotubule agent (+)-discodermolide using boron- mediated aldol reactions of chiral ketones, Angew. Chem. Int. Ed., 39, 377, 2000. 32. Martello, L.A. et al., The relationship between Taxol and (+)-discodermolide: synthetic analogs and modeling studies, Chem. Biol., 8, 843, 2001. Copyright © 2005 CRC Press, LLC 33. Gunasekera, S P., Longley, R.E., and Isbrucker, R.A., Acetylated analogs of the microtubule-stabilizing agent discodermolide: preparation and biological activity, J. Nat. Prod., 64, 171, 2001. 34. Gunasekera, S.P., Longley, R.E., and Isbrucker, R.A., Semisynthetic analogs of the microtubule- stabilizing agent discodermolide: preparation and biological activity, J. Nat. Prod., 65, 1830, 2002. 35. Gunasekera, S.P. et al., Synthetic analogs of the microtubule-stabilizing agent (+)-discodermolide: preparation and biological activity, J. Nat. Prod., 67, 749, 2004. 36. Longley, R.E. et al., Immunosuppression by discodermolide, Ann. N. Y. Acad. Sci., 696, 94, 1993. 37. Simonsen, M., Graft versus host reactions: their natural history and applicability as tools of research, Prog. Allergy, 6, 349, 1962. 38. Borel, J.F. et al., Biological effects of cyclosporin A: a new antilymphocytic agent, Agent Action, 6, 468, 1977. 39. Longley, R.E., Gunasekera, S.P., and Pomponi, S.A., Discodermolide compounds. U. S. Patent 5,840,750 issued November 24, 1998. 40. Boyd, M.R., The NCI in vitro anticancer drug discovery screen, Cancer Lett., 15, 6, 1989. 41. ter Harr, E. et al., Discodermolide, a cytotoxic marine agent that stabilizes microtubules more potently than taxol, Biochemistry, 35, 243, 1996. 42. Hung, D.T., Chen, J., and Schreiber, S.L., (+)-Discodermolide binds to microtubules in stoichiometric ratio to tubulin dimers, blocks taxol binding and results in mitotic arrest, Chem. Biol., 3, 287, 1996. 43. Hurst, J. et al., A novel model of a metastatic breast tumour xenograft line, Br. J. Cancer, 68, 274, 1993. 44. Giannakakou, P. et al., A common pharmacophore for epothilone and taxanes: molecular basis for drug resistance conferred by tubulin mutations in human cancer cells, Proc. Natl. Acad. Sci. USA, 97, 2904, 2000. 45. He, L. et al., A common pharmacophore for Taxol® and the epothilones based on the biological activity of a taxane molecule lacking a C-13 side chain, Biochemistry, 39, 3972, 2000. 46. Mita, C. et al., Phase 1 and pharmacokinetic study of XAA296A (discodermolide) administered once every three weeks in adult patients with advanced solid malignancies, Abstract A252 2003, AACR- NCI-EORTC Meeting, Molecular Target and Cancer therapeutics, Boston, MA, November 17–21, 2003. Copyright © 2005 CRC Press, LLC 11

The Dolastatins: Novel Antitumor Agents from Dolabella auricularia Erik Flahive and Jayaram Srirangam CONTENTS I. Introduction and History II. Origin, Isolation, and Structure of the Dolastatins III. Structure and Synthesis IV. Bioactivity and Mechanisms of Action V. Structural Modifications and SAR VI. Clinical Update References I. INTRODUCTION AND HISTORY Within the field of natural products, marine organisms represent a relatively untapped resource for novel compounds of medicinal interest. The systematic evaluation of biological constituents from marine organisms as potential drug candidates did not begin in earnest until the late 1960s, when the U.S. National Cancer Institute (NCI) initiated a massive screening program directed at the discovery of new anticancer drug leads.1,2 Over the next 30 yr, evaluation of isolates from the marine shell-less mollusk Dolabella auricularia (Aplyidae family) has led to the discovery of an impressive number of novel compounds possessing antineoplastic activity. Among the number of substances characterized to date, dolastatins 10 (1) and 15 (2) and the synthetic compounds derived from them show the most promising antitumor activity and are currently in human clinical trials in the United States, Europe, and Japan. This article will provide highlights of this work and will review recent progress in the discovery and development of the dolastatins as potential cancer chemotherapeutic drugs, with particular emphasis on dolastatins 10 and 15. II. ORIGIN, ISOLATION, AND STRUCTURE OF THE DOLASTATINS D. auricularia (Figure 11.1) and other gastropods of the order Anaspidea (Aplysiomorpha), known as “sea hares” because of the resemblance of the auriculate tentacula of these mollusks to the ears of a hare,3 typically lack the protection of an external shell. A remarkable feature of D. auricularia and other sea hares is their demonstrated ability to sequester and store secondary metabolites from dietary sources in their digestive glands for long periods.4–6 These accumulated substances, which typically occur in vanishingly small quantities, have long been thought to serve as a chemical defense mechanism against predators, though there is some debate as to their true role.6–8 Although Copyright © 2005 CRC Press, LLC FIGURE 11.1 (See color insert) Dollabella auricularia. the dynamics of sequestration and differentiation processes for this apparent chemical defense mechanism are still unclear, the adaptive strategies of D. auricularia and other shell-less mollusks are nonetheless truly remarkable. Some understanding of both the pharmacological utility as well as the potential dangers of extracts of D. auricularia has been known since ancient times.9–13 However, the potential of D. auricularia with respect to modern medical problems was not recognized until the initial report from Pettit and coworkers at Arizona State University concerning the potent antineoplastic activity of crude extracts of the Indian Ocean variety of this inconspicuous marine animal.14 Bioassay- guided (murine P388 leukemia cells [PS system]) isolation of the active principles of the crude extracts of D. auricularia collected off the coast of the Island of Mauritius in the Western Indian Ocean by Pettit’s group eventually led to a considerable number of pure compounds exhibiting an impressive spectrum of antineoplastic and cytostatic activity. Dolastatins 3–15 (Figure 11.2) were among the first group of compounds discovered and were first described in the chemical literature from 1982 to 1989.15 Structurally, the dolastatins are a series of novel linear and cyclic peptides that exhibit remarkable potency against neoplastic cells in vitro (see Table 11.1). These unique peptide constituents possess many interesting structural features including hitherto unknown, modified amino acid residues and were isolated in very low yield (from 106% to 108% yield based on crude wet weight), requiring several recollections of the marine animal. The isolation and structure elucidation16,17 of dolastatins 3–15, which required total synthesis in most cases to verify stereochemistry, has been reviewed.15,18 Dolastatin 10 (1), with an ED50 in the subnanomolar range against a number of cancer cell lines, was the most potent antineoplastic substance known at the time of its discovery.19 The last of the bioactive peptides isolated by Pettit to date were dolastatins 16–18.20–22 These compounds, isolated from D. auricularia specimens collected in Papua New Guinea, possess unique structural elements typical of peptides in this series including the novel alkynyl -β-amino acid of dolastatin 17 (9) and β-branched amino acid dolaphenvaline of dolastatin 16 (8). Among these, dolastatin 16 was the most potent inhibitor of cell growth against a minipanel of the NCI’s human cancer cell lines (mean GI50 2.5 × 104 µg/mL, ~18-fold less potent than dolastatin 10). Isolation of a closely related but less bioactive compound, homodolastatin 16, has recently been Copyright © 2005 CRC Press, LLC FIGURE 11.2 Dolastatins 1–18. reported.23 Yamada and coworkers of Nagoya University initiated a large-scale collection of D. auricularia from Japanese waters (Shima peninsula) in the mid-1980s that led to the discovery of yet another series of unique compounds with antitumor activity (Figure 11.3).24 The majority of these sea hare constituents (Dolastatins C-E,25–28 dolastatin G/nordolastatin G,29,30 dolastatin H/isod- olastatin H,31 dolastatin I,32 aurilide,33,34 dolabellin,35 and doliculide36) are linear or cyclic pseudopeptides and exhibit potent in vitro antitumor activity in some cases. Doliculide (21), containing a novel ortho-iodo-N-methyl-D-tyrosine residue, and dolastatin H (16)/isodolastatin H (17), which bear close structural resemblance to dolastatin 10, are the most potent inhibitors of antineoplastic growth (see Table 11.1). In addition to the pseudopeptide constituents described by Yamada, a number of sea hare–derived macrolides with antineoplastic activity were reported by these same researchers (dolabellides A and B24 and doliculols A and B37). Despite the structural similarity between the dolastatins discovered by Pettit’s and Yamada’s groups, it is interesting to note that neither group has yet to isolate an identical compound from D. auricularia. Copyright © 2005 CRC Press, LLC O N S O H O N N O N O O H N N N N O O NH NH HN O H O O O O NH O N N 2 O O O S Dolastatin C (11) Dolastatin D (12) Dolastatin E (13) O O OH H O N N N N N N N H O O O O O O N O N HO O N O O O OH O O Dolastatin H (16) O OR O OH H Dolastatin G (14, R = CH3) N N N N O Nordolastatin G (15, R = H) O O O O O Isodolastatin H (17) O N O O N H N O OH O O NH HN N N NH N OH N O O O O O S Cl Cl O O N S NH N O O O S Dolastatin I (18) O OH O O O Dolabellin (20) O OH Aurilide (19) I O N O HO O N H Doliculide (21) FIGURE 11.3 Dolastatins isolated from Shima Peninsula collection of Dollabella auricularia.42 The origin of this impressive array of bioactive secondary metabolites found within sea hares has been the subject of much interest. Although direct biosynthesis or biosynthesis by symbiotic microorganisms has been considered, these theories have slowly given way to strong evidence for exogenous procurement of such metabolites from dietary sources.38 It is now generally accepted that cyanobacteria consumed by sea hares are the true biosynthetic source of the rich structural diversity represented by the dolastatins. This idea is supported by a number of observations. Collections of the same species from different locations have led to the isolation of similar, but structurally distinct, dolastatins. Furthermore, the incorporation of structurally unique amino acids that is typical of the dolastatins seems to favor a diet-derived origin based on the prevalence of such novel, nonproteinogenic amino acids among marine algae.39 Moreover, a number of these compounds (dolastatins 3,40 10,41 12–1442,43 and 1644) have now been isolated directly from varieties of the blue-green algae Lyngbya majuscule and Symploca hydnoides, along with new bioactive Copyright © 2005 CRC Press, LLC TABLE 11.1 Summary of Antitumor Agents from the Marine Mollusk Dollabella auricularia79 In vitro activity Compound Collection site (µg/mL) Dolastatin 3 (3) Mauritius 1.6 × 10–1a Dolastatin 10 (1) Mauritius 3.4 × 10–6b Dolastatin 11 (4) Mauritius 1.9 × 10–4b Dolastatin 12 (5) Mauritius 8.8 × 10–2b Dolastatin 13 (6) Mauritius 3.5b Dolastatin 14 (7) Mauritius 2.7 × 10–3b Dolastatin 15 (2) Mauritius 2.8 × 10–3b Dolastatin 16 (8) Mauritius 9.6 × 10–4b Dolastatin 17 (9) Mauritius 4.5 × 10–1b Dolastatin 18 (10) Papua New Guinea 3.9 × 10–1b Dolastatin C (11) Papua New Guinea 17c Dolastatin D (12) Papua New Guinea 2.2c Dolastatin E (13) Shima Peninsula (Japan) 22-40c Dolastatin G (14) Shima Peninsula (Japan) 1c Nordolastatin G (15) Shima Peninsula (Japan) 5.3c Dolastatin H (16) Shima Peninsula (Japan) 2.2 × 10–2c Isodolastatin H (17) Shima Peninsula (Japan) 1.6 × 10–2c Dolastatin I (18) Shima Peninsula (Japan) 12c Doliculide (21) Shima Peninsula (Japan) 1.1 × 10–2c Dolabellin (20) Shima Peninsula (Japan) 6.1c Doliculide (21) Shima Peninsula (Japan) 1 × 10–3c a ED50, P388 (lymphocytic leukemia cells). b GI50, NCI-H460 (non–small cell lung carcinoma). c IC50, HeLa-S3 (cervical carcinoma cells). compounds with closely related structures such as the recently described symplostatins 1–345–47 and somamides A and B.43 This link between the dolastatins and cyanobacteria has important economic and ecological relevance for the future discovery and development of new anticancer drug leads.48 III. STRUCTURE AND SYNTHESIS Though structurally complex in many cases, total synthesis of most of the dolastatins has been achieved to date. The importance of these synthetic endeavors should be underscored, as these efforts have allowed further study of this interesting family of compounds in both in vivo bioassays and human clinical trials. In this section, an overview of progress made in the synthesis of the various dolastatins and related compounds is presented, with particular attention to the methods devised for preparation of dolastatins 10 and 15 and important derivatives thereof. Structure determination for dolastatin 3 (3), the first in the series to be described, required detailed spectral (nuclear magnetic resonance [NMR] and tandem mass spectrometry [MS] sequenc- ing) and hydrolytic studies, as well as total synthesis.15 The structurally interesting thiazole- containing amino acids of dolastatin 3 are related to similar cysteine-derived49 thiazole amino acids found in other biologically active natural products, such as the antineoplastic marine sponge constituents patellamides A–C50, the antineoplastic lower plant metabolite bleomycin,51 and the broad-spectrum antibiotic thiostrepton,52 among others. Several total syntheses of dolastatin 3 have been reported,49,53–56 highlighted by the synthesis of the thiazole amino acids (Gly)Thz and (Gln)Thz. Copyright © 2005 CRC Press, LLC H O HN H N N N N OH + O O O O O N S Dap Dov Val Dil Doe 26 (TFA salt) 27 (TFA salt) 19a O H 19 18 H 10 N N N N 9 N 6 O O O O O N S Dov Val Dil Dap Doe Dolastatin 10 (1) FIGURE 11.4 Dolastatin 10 structural units. The synthesis of the most bioactive of the dolastatins, dolastatin 10 (1), was an extremely challenging undertaking because of the total of nine asymmetric centers of unknown configuration that make up the structure. The most notable structural features of dolastatin 10 are the two novel γ-amino-β-methoxy-acid residues, dolaproine (Dap, 22) and dolaisoleucine (Dil, 23), each con- taining three chiral centers (Figure 11.4). Similar structural units are also found in the marine tunicate antineoplastic components didemnins A–C57 and the lower plant metabolite bleomycin.51 Also present is a novel thiazole-containing C-terminal unit dolaphenine (Doe, 24), derived from phenylalanine, that is also found in dolastatin 18 (10). Several stereoselective syntheses58 of the unique Dap, Dil, and Doe subunits have appeared in the chemical literature. The absolute configuration of dolastatin 10 was confirmed by total synthesis,59 first reported in 1989 by Pettit’s group. Because of the exceptional promise of dolastatin 10 as an anticancer drug, several partial and total syntheses58 of this peptide have been subsequently reported. Assembly of the amino acid constituents in both stepwise (by Shiori,60,61 Koga,62 and Poncet63) and convergent fashion has been successfully demonstrated. The general approach followed by Pettit’s group for the preparation of initial quantities of dolastatin 10 for preclinical studies was convergent, as shown in Figure 11.4. The absence of a chiral center adjacent to the carboxylic acid of tripeptide Dov- Val-Dil (26) and the ease of its coupling with a proline derivative made this approach strategically attractive and free of epimerization concerns. In addition, this strategy was particularly amenable to the synthesis of shortened peptide segments and synthetic analogs that were used for elucidation of

structure–activity relationships (SARs), as well as to provide insights into the mechanism of action for this drug, as discussed in more detail in section V of this chapter. Yamada and coworkers also found this general [3 + 2] coupling approach useful for the preparation of the structurally similar compounds dolastatin H (16) and isodolastatin H (17).31 Although the crude structures of dolastatins 11 (4) and 12 (5) were reported in 1989,64 the absolute stereochemistry of these cyclopeptides remained undetermined until a total synthesis was reported by Bates and coworkers in 1997.65,66 Dolastatins 11 and 12 are closely related cyclodep- sipeptides bearing a novel β-amino acid residue and are structurally very similar to the antifungal natural products majusculamide C67 and 57-normajusculamide C68 — components of the marine blue-green alga Lyngbya majuscula. This synthetic achievement has allowed for the unique bio- chemical effects of dolastatin 11 on tubulin and actin to be studied in more detail (see section IV). Dolastatin 13 (6)69 is a structurally complex cyclodepsipeptide containing a novel piperidine unit Copyright © 2005 CRC Press, LLC Dov-Val-Me-Val-Pro (28) Pro-Hiva-Dpy (29) Petit, 1991 Pro Pro Hiva N O O O N N O 7 O Dov HN N O O O 6 N 1 2 Dpy 4 3 Val MeVal O 1) H , Pd/C H /C, HCHO 2 2, Pd Dolastatin 15 (2) 2) Dov coupling Poncet, 1992 Petit, 1994 Z-Val-Val-Me-Val-Pro-Hiva-Dpy (32) Z-Val-Me-Val-Pro-Hiva-Dpy (34) Z-Val-Val-Me-Val-Pro (30) Hiva-Dpy (31) Z-Val-Me-Val-Pro (33) Pro-Hiva-Dpy (29) FIGURE 11.5 Synthetic approaches to dolastatin 15. and a dehydroamino acid residue. Although no total synthesis has appeared to date, it is presumed that the recent discovery of the structurally analogous cyanobacterial metabolites symplostatin 246 and somamides A and B43 will spark interest in construction of this peptide core. Total synthesis of the moderately active dolastatin 14 (7), a cyclodepsipeptide containing a polyene hydrocarbon fragment, has not been achieved to date. Several groups have published complementary approaches to the four possible isomers of the polyene subunit (dolatrienoic acid).70–72 The absolute configura- tions of both dolastatins 13 and 14 are presently unknown, and their biological effects have not been thoroughly evaluated because of a lack of sufficient material. The second most potent D. auricularia metabolite discovered to date is the linear depsipeptide dolastatin 15 (2). This unique peptide component bears many structural similarities to dolastatin 10 (1), including the presence of an N,N-dimethylvaline (Dov) N-terminus (Figure 11.5). The most interesting structural feature of dolastatin 15 is the novel C-terminal dolapyrrolidone unit (Dpy), which is presumably derived from biosynthetic modification of N-acyl phenylalanine. A total synthesis73 of 2 was achieved by Pettit’s group in 1991, and the absolute configuration was assigned accordingly. A novel synthesis from Poncet’s group appeared in 199274 and featured reductive bis- methylation of the N-terminus as a final step. A second submission from Pettit in 1994 outlined some subtle improvements to the original approach,75 and recently, a solid-phase synthesis of dolastatin 15 was also reported.76 All of the reported solution–phase syntheses of dolastatin 15 are convergent and take advantage of the reduced tendency of peptide segments coupled at proline to epimerize (see Figure 11.5). More recent approaches incorporate the Dov unit in the final step to circumvent purification difficulties associated with early introduction of the N,N-dimethylated N terminus. As mentioned above for dolastatin 10, these complementary, convergent approaches are well suited for rapid preparation of synthetic derivatives useful for uncovering important SARs. The ample supply of synthetic dolastatin 15 has allowed for the biological activity of this unique peptide to be studied in great detail.77 Dolastatin 16 (8) is a unique cyclodepsipeptide containing novel β-amino acid (dolamethylleu- cine) and β-branched valine (dolaphenvaline) units reported by Pettit and coworkers in 1997. The Copyright © 2005 CRC Press, LLC structure and amino acid sequence was determined using high-field NMR and tandem mass spectral interpretations, and the stereochemistry of most of the subunits was determined via chiral high- performance liquid chromatography analysis of the acidic hydrolysate of the natural peptide. This compound exhibits marked antiproliferative effects on cancer cells in vitro, but the lack of available synthetic material has prevented further study of its biochemical effects.20 The structure and partial stereochemistry of dolastatin 17 (9) were elucidated in a manner similar to that of dolastatin 16 above. The absolute stereochemistry of both dolastatin 16 and 17 has not been reported to date. The absolute stereochemistry of the C-terminal dolaphenine unit of dolastatin 18 (10) was recently determined via total synthesis and confirmed by x-ray crystal structure.78 Among the group of dolastatins (and related compounds) described by Yamada (Figure 11.3),24 the absolute configuration for each is known, and a synthesis has been described for nearly every compound. This represents a tremendous achievement given the structural complexity of these substances. The various published syntheses of pseudopeptides 11–21 have been reviewed in detail by Poncet.79 IV. BIOACTIVITY AND MECHANISMS OF ACTION The dolastatins are all strong inhibitors of cancer cell growth, and indeed, the discovery of all these substances was guided by bioassay of extracts of the crude animal matter against actual cancer cells in vitro. Dolastatin 10 is the most potent antineoplastic agent of the dolastatins, followed by dolastatin 15 (on average, dolastatin 15 appears to be about sevenfold less active than dolastatin 10 in most preclinical models80). Dolastatin 10 has been studied most intensively because of its exceptionally potent antiproliferative effect on cancer cells and initial promise as an anticancer drug. Both dolastatins 10 and 15 are currently being pursued as clinical antitumor drug candidates and have reached phase II clinical trials in the United States, Europe, and Japan (see section VI). Dolastatin 10 also exhibits therapeutic potential against other disease indications, as evidenced by its reported antiviral,81 antibacterial,82,83 antifungal,84 and antimalarial83,84 activities. A comparison of the growth inhibitory effects of dolastatins 10–16 against a selection of human cancer cell lines is presented in Table 11.2.15 TABLE 11.2 Comparative ACTIVITIES (GI50 µg/mL) of Dolastatins 10–16 Against a Human Tumor Cell Line Minipanel from the National Cancer Institute Primary Screen Dolastatin OVCAR-3a SF-295b A498c NCI-H460d KM20L2e SK-MEL-5f 10 (1) 9.5 × 10–7 7.61 × 10–6 2.6 × 10–5 3.4 × 10–6 4.7 × 10–6 7.4 × 10–6 11 (4) 0.04 0.031 0.023 1.9 × 10–4 0.037 0.034 12 (5) 0.14 0.37 0.78 0.088 0.29 0.32 13 (6) 2.5 3.9 4.5 3.5 2.4 2.7 14 (7) 4.1 × 10–3 2.0 × 10–3 0.038 2.7 × 10–3 0.020 6.4 × 10–5 15 (2) 1.3 × 10–4 4.3 × 10–4 2.7 × 10–4 2.8 × 10–4 1.1 × 10–4 1.7 × 10–4 16 (8) — 5.2 × 10–3 — 9.6 × 10–4 1.2 × 10–3 3.3 × 10–3 a Ovarian adenocarcinoma. b Brain (glioma). c Renal carcinoma. d Non–small cell lung carcinoma. e Colon carcinoma. f Melanoma. Copyright © 2005 CRC Press, LLC TABLE 11.3 Dolastatin 10 and 15 Activity versus Human Leukemia Cell Lines IC50 (nM) Cell Line Dolastatin 10 (1) Dolastatin 15 (2) BV-173a 4.7 × 10–2 5.9 × 10–1 KM-3a 8.7 × 10–2 8.9 × 10–1 K-562b 1.2 × 10–1 1 U-937c 1.8 × 10–2 1.2 × 10–2 HL-60d 5.5 × 10–2 2.5 × 10–2 KARPAS-299e 9.8 × 10–2 2.8 × 10–2 B-Cells 1.0 >2.0 T-Cells 3.7 × 10–1 >2.0 a Pre-B-cell leukemia. b Erythroleukemia. c Monocytic leukemia. d Myeloid leukemia. e T-cell lymphoma. Both dolastatin 10 and 15 are potent antimitotic agents and, as such, interfere with cell division via disruption of microtubule function.85 Accumulation of cells trapped in metaphase arrest is observed in cancer cells treated with either of these agents, and at higher concentrations, intracellular microtubules disappear completely.86 The cytostatic effects of dolastatins 10 and 15 on cells in vitro are derived from the binding interaction between these peptides and the β-subunit of the dimeric microtubule protein tubulin.87 The role that this biochemical interaction plays in the cytotoxic and antitumor effects of these compounds is discussed below. The impressive performance of dolastatins 10 and 15 against cancer cells in vitro was followed directly by evaluation of these compounds in several murine tumor models. Both compounds were found to be effective against several human tumor xenografts such as the OVCAR-3 human ovary and LOX human melanoma in vivo systems.15 A number of studies concerning the activity of both compounds against human leukemia88–92 and lymphoma93–95 cell lines as well as its efficacy for the treatment of prostate cancer96 have been reported. An interesting result to come out of these studies is that both agents are potently cytostatic (log 3 to log 4 more effective than the clinically established drug vincristine) but not cytotoxic at drug concentrations required to achieve maximum growth inhibition. Importantly, the two dolastatin compounds showed low toxicities to resting or proliferating nonleukemic cells. Furthermore, the observed growth inhibition was reversed once treatment was terminated.90 A summary of results for dolastatins 10 and 15 against human leukemia and lymphoma cells appears in Table 11.3. The antiproliferative effects of dolastatin 10 on purified tubulin and cellular microtubule assem- bly generally parallel those of the well-known vinca alkaloids, but its inhibition of cancer cell growth in vitro is much more potent. Both dolastatin 10 and the structurally unrelated mycotoxin phomopsin A are strong noncompetitive inhibitors of the binding of radiolabeled vinca alkaloids. In addition, dolastatin 10 is a potent inhibitor of nucleotide exchange and tubulin-dependent guanine 5-triphosphate (GTP) hydrolysis (Table 11.4). The hydrolysis of GTP bound to the exchangeable nucleotide site on the β-subunit of tubulin is believed to play an important regulatory role in the polymerization of tubulin. Results of extensive competition experiments with dolastatin 10 and other vinca domain drugs87 have led to a proposed model of the binding sites of this class of antimitotics. Hamel and coworkers have proposed that dolastatin 10 binds to a “peptide site” near both the vinca site and the exchangeable nucleotide site.97 According to the model, when dolastatin 10 is bound to its critical active site, the C-terminal end of the peptide effectively blocks access to Copyright © 2005 CRC Press, LLC TABLE 11.4 Relative Inhibitory Effects of Vinca Domain Drugs on [3H]Vincristine Binding, on the Polymerization of Purified Tubulin,a on Guanine 5-Triphosphate (GTP) Nucleotide Exchange, and on Tubulin-Dependent GTP Hydrolysis Inhibition [3H]Vincristine IC b 50 Tubulin IC50 GTP IC50 GTP Drug Binding (type) Polymerization (µM) Exchange (µM) Hydrolysis (µM) Vinblastine Competitive 1.5 >80 6 Maytansine Competitive 3.4 9 4 Rhizoxin Competitive 6.9 29 13 Phomopsin A Noncompetitive 1.4 5 3 Dolastatin 10 Noncompetitive 1.2 8 3 Dolastatin 15 NIc 23 NIc 33 Halichondrin B Noncompetitived 6.1 17 12 Spongistatin 1 Noncompetitived 5.3 87 — a Standard tubulin assay: preincubation (15 min) of a drug–tubulin mixture (10 µM tubulin) in 1 M monosodium glutamate + 1 mM MgCl2 at 37˚C was followed by the addition of GTP, and the polymerization reaction at 37˚C was monitored, using a spectrophotometer. The tubulin assay used in the current structure–activity study was a modified method as discussed in section 10.3 of reference 87. b IC50 refers to the drug dosage required for 50% inhibition. c Noninhibitory. d Competition experiments with halichondrin B and spongistatin 1 were carried out using [3H]vinblastine rather than [3H]vincristine. the vinca alkaloid binding site and the exchangeable nucleotide site. This idea is supported by several structure–activity observations98 and, in particular, by the fact that a shortened tripeptide segment not containing the latter two C-terminal units (26; Figure 11.4) retains tubulin antiprolif- erative activity but exhibits reduced inhibition of vinca alkaloid binding and nucleotide exchange.86 The effects of dolastatin 15 on the proliferation of cellular microtubules and purified tubulin in vitro have also been studied.97 Similar to dolastatin 10, the structurally similar dolastatin 15 appears to derive its potent antineoplastic effects by interfering with the assembly of microtubules and causing mitotic arrest, though in the case of the latter peptide, the connection is less clear. Dolastatin 15 is roughly 1/7 as active as dolastatin 10 but is nearly seven times more active than vinblastine with respect to cell growth inhibition of L1210 murine leukemia cells.99 Despite its potent antitumor activity, the activity of dolastatin 15 in the tubulin assembly assay is much weaker than for any of the other drugs that bind in the vinca domain (Table 11.4). In addition, dolastatin 15 does not inhibit the binding of the vinca alkaloids to tubulin and does not affect nucleotide exchange to any

degree. Tubulin-dependent GTP hydrolysis is inhibited by dolastatin 15, but to a lesser degree than encountered with dolastatin 10. Hamel et al. have interpreted these results as an indication of a relatively weak drug–tubulin interaction for dolastatin 15. More recently, Jordan and coworkers have shown evidence that the dolastatin 15 derivative cemadotin (38; Figure 11.6) may exert its antitumor activity via suppression of dynamic instability of microtubules assembled to steady state.100 These authors suggest that cemadotin may bind at a novel site in tubulin, distinct from the vinca site. Recent competition studies with radiolabeled dolastatin 15 seem to indicate that the binding sites for antimitotic peptides in the dolastatin family may consist of a series of overlapping domains.101 Dolastatin 10 and 15 are nevertheless presumed to interact with a similar specific binding site in the vinca domain based on structural homology and GTP hydrolysis considerations,97 although other targets cannot be ruled out. With regard to the demonstrated capacity of dolastatin 10 and 15 to disrupt microtubule organization, it is important to note that the exact mechanism by which antimitotic drugs cause Copyright © 2005 CRC Press, LLC O H H N N N N N N H O O O N N O O O O O Dov HN N O O Dov Val Dil Dap PEA BnNH2 Val MeVal Soblidotin (37) (TZ-1027, Auristatin PE) Cemadotin (38) (LU103793) GI50 = 3.0 × 10–6 mg/mL NCI-H460 (Lung-NSC) IC50 = 7.0 × 10–4 mg/mL HeLa-S3 (cervical carcinoma) FIGURE 11.6 (See color insert) Synthetic derivatives of dolastatin 10 (soblidotin) and dolastatin 15 (cematodin) in phase II clinical trials. cell death remains unclear. Some researchers have suggested that apoptosis induced by phospho- rylation of the oncoprotein bcl-2 (an inhibitor of programmed cell death) may play an important role in the mechanism of action of some antitumor agents that are also microtubule poisons,102,103 and some evidence for this explanation has been shown for dolastatins 10 and 15.94,104,105 Some biology of newly discovered, natural analogs of dolastatin 10 was also recently divulged. Cyanobacterial metabolite symplostatin 1 (35; Figure 11.7) reportedly exhibited potent microtubule inhibition and was furthermore found to be effective in vivo against two drug-insensitive, murine tumor models.41 A recent study of the molecular pharmacology of symplostatin 1106 revealed that this closely related substance shares a very similar mechanism of action with dolastatin 10. Symplostatin 3 (36; Figure 11.7) is about 10–100-fold less cytotoxic than dolastatin 10 in vitro.47 Soblidotin (37, TZT-1027; Figure 11.6), a synthetic derivative of dolastatin now in clinical trials, inhibits microtubule polymerization and interferes with microtubule assembly/disassembly equi- libria through interaction with tubulin, as expected from similar work with the parent peptide. Apoptosis by a direct mechanism (cytotoxicity) versus the known antimitotic effect has been observed with soblidotin at higher doses in vitro107 and in vivo with murine B16 melanoma xenografts.108 Of considerable interest is the observation made by Japanese researchers that treat- ment of an advanced murine tumor model (colon 26 adenocarcinoma) with soblidotin induces tumoral vascular collapse and tumor cell death, necrosis, and tumor regression.109 This indicates that soblidotin may exert its antitumor activity by both conventional microtubule disruption and by a unique, antitumoral vascular mechanism of action that could prove to be a powerful combi- nation for cancer chemotherapy. Among other dolastatins whose biology has been studied in detail, dolastatin 11 is particularly interesting because of its unique biochemical effects on tubulin and actin proteins in vitro.110 A O H H H O H N N N N N N N N N N O O O O O O O O O O N S Dov O OH Val Dil Dap Me2IIe Val Dil Dap Doe PhLac Symplostatin 1 (35) Symplostatin 3 (36) IC50 = 3.0 × 10–4 mg/mL IC50 = 3.9 × 10–3 mg/mL KB cells (epidemoid carcinoma) KB cells (epidemoid carcinoma) FIGURE 11.7 (See color insert) Naturally occurring derivatives of dolastatin 10. Copyright © 2005 CRC Press, LLC recent study concerning the mechanism of action of this peptide showed that dolastatin 11 arrests cells at cytokinesis by causing a rapid and massive rearrangement of the cellular actin filament network.110 Dolastatin 16 has shown a mean panel GI50 value of 2.2 × 101 µg/mL against the NCI’s complete panel of 60 cell lines and showed marked inhibitory effects against non–small cell lung carcinoma, colon carcinoma, and melanoma cells in particular.20 Interestingly, dolastatin 16 shows a relatively low GI50-COMPARE111 correlation with dolastatins 10 and 15, which may indicate a different mechanism of action for this cyclic peptide relative to the others. V. STRUCTURAL MODIFICATIONS AND SAR The initial discovery of dolastatin 10 and its potent inhibitory effect on cancer cell growth at unprecedentedly low drug concentrations was soon followed by structure–activity studies on this structurally unique peptide. These efforts were aimed at elucidation of key structural requirements for activity and at reducing structural complexity where possible. The first structural modifications of dolastatin 10 to be prepared and evaluated against tumor cells in vitro were 19 of the possible 128 stereoisomers of the natural peptide, which were synthesized in connection with absolute configuration determinations for the non-proteinogenic Dil, Dap, and Doe residues.98,112 Among the various stereoisomers tested, only two retained antitumor activity comparable to the natural peptide. Change of chirality at most stereocenters resulted in moderate to substantial decrease in activity against the murine P388 cell line, with the exception of the (6R)-isomer [isodolastatin- (6R), 42; Table 11.5] and (19aR)-isomer, where reversal of stereochemistry at the Doe α-carbon (position 6) and Dil side chain (position 19a) was attended by no loss of antitumor activity (Figure 11.8). Moreover, a clear correlation between inhibition of tubulin polymerization and inhibition of L1210 leukemia cell growth was demonstrated for all stereoisomers that retained some antitumor activity. Interestingly, the shortened tripeptide segment Dov-Val-Dil-OMe ester is completely inac- tive against tumor cell growth and yet retains substantial inhibitory effects on tubulin polymeriza- tion. Hamel and coworkers have speculated that the Dil subunit may be essential for the interaction of dolastatin 10 with tubulin, although the exact nature of the link between affinity for tubulin- binding and antineoplastic activity is not completely clear.86 This early work set the stage for more fundamental changes to the peptide backbone of dolastatin 10, carried out mainly by the research groups of Pettit,86,112–114 Poncet,115–117 and Miyazaki.118,119 Most of the attention given to the preparation of dolastatin 10 analogs has centered on modifica- tions of the C terminus (e.g., Doe unit). Modifications at this position allow for structure simplification (e.g., thiazole replacement) without diminution of antineoplastic activity and may furthermore permit fine-tuning of selectivity in some cases. A survey of interesting results from some of these SAR studies is given in Table 11.5. In particular, the phenethylamine core of the C terminus was found to be essential for antitumor activity, and the corresponding derivative, described by both Pettit113,114 and Miyazaki,118,119 has quickly advanced as a drug candidate (soblidotin, 37; Figure 11.6), currently in PII clinical trials. The length of the alkyl chain between the amide and phenyl group does appear to be important for antitumor activity (derivatives 43, 44, 37, and 48), whereas the effect of aryl substitution is less obvious (derivatives 45–47). Several amino acid methyl esters are also adequate replacement groups for the usual C terminus, although the nature of the amino acid employed [e.g., aryl (Phe-OMe, 52), alkyl (Ile-OMe, 55), or sulfur-containing (Met-OMe, 54)] does not appear to have a significant effect on activity. Interestingly, C-terminal replacement groups containing basic functionality (50 and 51) exhibited markedly reduced antineoplastic activity and furthermore showed very little correlation to the activity profile of the parent peptide. Derivatives in which the Dap subunit was either omitted (58) or replaced with proline (57) were ineffective. Modification to certain positions of the other subunits also produced derivatives that showed antitumor activity comparable to that of dolastatin 10. However, none of the modifications described resulted in any meaningful simplification of the core structure. It is interesting to note that one of the active Dov-modified, synthetic analogs (Me2Ile N terminus) described by Pettit et al.114 has since been discovered as a naturally occurring Copyright © 2005 CRC Press, LLC TABLE 11.5 Survey of Dolastatin 10 Modifications and Effect on Inhibition of Tumor Cell Proliferation and Tubulin Polymerization. Mean GI50 Correlation L1210 IC50 Tubulin IC50 No. Compound (×108 M)a Coefficientb (nM)c (µg/mL)d 1 Dolastatin 10 0.012 1.00 0.5 1.3 ± 0.2 42 Isodolastatin-(6R) 0.1e — — — 43 n = 0, R2 = H 1.0 0.828 60 1.8 ± 0.2 44 n = 1, R2 = H 0.174 0.851 80 1.6 ± 0.2 37 n = 2, R2 = H 0.049 0.901 0.6 1.3 ± 0.2 45 n = 2, R2 = ortho-Cl — — 3 1.2 ± 0.2 46 n = 2, R2 = meta-Cl — — 2 1.6 ± 0.2 47 n = 2, R2 = para-Cl 0.076 0.835 0.4 1.5 ± 0.2 48 n = 3, R2 = H 0.182 0.805 20 1.3 ± 0.05 49 0.085 0.847 5 1.2 ± 0.2 50 3.8 0.667 500 2.6 ± 0.5 51 R1= 8.2 <0.6 >1000 2.1 ± 0.3 52 R1 = Phe-OMe 0.020 0.903 5 1.6 ± 0.04 53 R1 = Phe-NH2 0.282 0.852 10 1.0 ± 0.2 54 R1 = Met-OMe 0.040 0.904 6 1.4 ± 0.2 55 R1 = Ile-OMe 0.016 0.899 7 1.7 ± 0.2 56 R1 = OMe — — 30 1.3 ± 0.1 57 Dap-Doe  Pro-PEAf 3.8 0.667 800 2.9 ± 0.3 58 Dov-Val-Dil-Doe (Dap omitted) 19.5 0.711 200 2.1 ± 0.1 a Mean GI50s obtained from quadruplicate screenings of compounds indicated against the National Cancer Institute’s full 60–cancer cell line panel. b Correlation coefficients from COMPARE algorithm analyses of averaged Total Growth Inhibition (TGI) mean graph obtained from quadruplicate screening results against the full 60–cancer cell line panel, using the mean value for dolastatin 10 as the seed for all correlations. c Murine L1210 (lymphocytic leukemia cells). d Inhibition of the polymerization of purified, bovine brain tubulin. e ED50 value for isodolastatin-(6R) against murine P388 leukemia cells. This compounds were not tested on other assays listed above. f Pro-PEA corresponds to the phenethylamide of L-proline. Copyright © 2005 CRC Press, LLC iPr, sBu, or iBu R3 (equipotent) H 19 H 3C 18 N N 10 9 N 6 essential for R ity 2 O O O O activ CH3NH- replacement R O N S 1 (equipotent) NH N O H-replacement (equipotent) FIGURE 11.8 Effects of structural modifications on dolastatin 10 activity. cyanobacterial metabolite (symplostatin 1, 35; Figure 11.7).45 A summary of some SARs for dolastatin 10 is given in Figure 11.8. Preparation of a single crystal of the (6R)-isomer of dolastatin 10 [isodolastatin-(6R), 42] afforded the first opportunity for study of the three-dimensional conformation of this peptide, and a molecular modeling study was undertaken by Pettit and Srirangam. Computer-assisted modeling studies for this isomer of dolastatin 10, supported by data from an x-ray crystal structure, revealed a ~21 kcal/mol energy barrier between cis and trans conformers, visible by NMR at room temperature. Interestingly, the crystalline structure of isodolastatin-(6R) contains all trans-amide bonds except for the Dil–Dap bond, which proved to be cis.112 Following this work on stereoisomer 42, Poncet and coworkers reported a detailed molecular modeling study on the natural stereoisomer of both dolastatin 10 and dolastatin 15, aimed at elucidation of the structural requirements for antitumor activity. The preferred, low-energy conformations of these peptides in DMSO-d6 solution were determined via two-dimensional NMR measurements, followed by molecular dynamics simulations.115 These calculations confirmed that the two conformers observed for dolastatin 10 in solution are a result of cis-trans isomerization of the Dil–Dap amide bond, with an energy barrier of 8 kcal/mol in favor of the cis conformer. The unique, cis-trans isomerization of dolastatin 10 could potentially provide insights into the active conformation of the peptide in association with its presumed intracellular target tubulin. Poncet synthesized a cyclic derivative of the peptide (39), joined through a threonine-like ester linkage between the “Dov” and “Doe” residues, to probe the effect of locking the backbone in a folded conformation (Figure 11.9).117 Although the resulting compound was completely inactive, it is possible that cis-trans isomerization of dolastatin 10 plays a central role in its interaction with tubulin, a mechanism that will likely be more well-defined after further studies. Among the

D10/D15 hybrid compounds 40 and 41 prepared by Poncet (Figure 11.9), both peptides retained partial Dil Val O O N HN N O N Dap O Dov O O point of cyclization N NH (26% yield) O O S Doe Pro 39 N H Pro O O O N N N Hiva Dov HN H O N O O O N S N N N N O O Dap O O O O Val MeVal Dov O N Doe Dpy Val Dil O 40 41 FIGURE 11.9 (See color insert) Dolastatin 10 cyclic derivative and D-10/D-15 hybrid compounds. Copyright © 2005 CRC Press, LLC TABLE 11.6 Survey of Dolastatin 15 N-Terminus Modifications and Effect on Inhibition of Tumor Cell Proliferation and Tubulin Polymerization L1210 P388 ED50 NCI-H460 ED50 Tubulin IC50 No. N-Terminus IC50 (nM)a (ng/mL)b (ng/mL)c (µg/mL)d 1 (7S)-Hiva-(4S)-Dpy 0.15 2.4 0.28 5.2 59 (7R)-Hiva-(4S)-Dpy 0.30 — — — 60 (7S)-Hiva-(4R)-Dpy 0.31 — — — 61 (7S)-Hiva-(4S)-Dpy 1.78 — — — 62 (7S)-Hiva-anilide 3 0.52 0.032 26 63 (7S)-Hiva-benzamide 1 0.42 0.25 20 64 (7S)-Hiva-phenethylamide 2 0.26 0.044 15 65 (7S)-Hiva-phenpropylamide — 0.38 1.1 — 66 (7S)-Lac-phenethylamide — 0.46 0.37 — 67 L-Val-phenethylamide — 0.59 2.5 — 68 L-MeVal-phenethylamide — >1000 3600 — a Murine L1210 (lymphocytic leukemia cells). b Murine P388 (lymphocytic leukemia cells). c Human non–small cell lung carcinoma. d Inhibition of the polymerization of purified, bovine brain tubulin. antitumor activity, whereas only peptide 41 (Dov-Val-Dil-Pro-Hiva-Dpy) retained capacity to inhibit tubulin polymerization in vitro.117 This result appears to support Hamel’s assertion that the Dil residue is critical for molecular interactions between the peptide and tubulin.86 Because dolastatins 11–14, dolastatins 16–18, and Yamada’s series of D. auricularia metabolites have not undergone extensive biochemical evaluation to date, less is presently known about SARs for these compounds. Recently, Russo and coworkers in Italy began preliminary studies aimed at elucidation of SAR for dolastatins 11 and 12 using molecular mechanics and dynamics calculations in aqueous solution.120 Now that Bates and coworkers have devised a total synthesis of dolastatins 11 and 12,66 the door remains open for further elucidation of its unique biochemical effects on actin through preparation of synthetic derivatives. These and other advances should help provide more insight into the specific mechanisms of action for these as-yet largely unstudied peptides. Similar to dolastatin 10, efforts toward understanding SARs for dolastatin 15 are further advanced than for other members of this series, and a number of research groups have prepared derivatives and published studies. In contrast to dolastatin 10, dolastatin 15 was found to be surprisingly insensitive to structural and stereochemical modifications at the C terminus (Table 11.6; Figure 11.10). Change in the configuration of the stereogenic centers that comprise the Hiva and Dpy unit was found to minimally affect in vitro antineoplastic activity (compounds 59–61).121 As with dolastatin 10, a slight trend favoring the phenethylamine C terminus over other arylalkyl chain lengths was observed, but the effect was much more subtle in the case of dolastatin 15 (compounds 62–65).122 Results of SAR studies for dolastatin 15 clearly indicate that the structurally complex Dpy C terminus plays a limited role in the cell growth inhibitory effects of this peptide. The interaction of dolastatin 15 with purified tubulin in vitro is considerably weaker than that of dolastatin 10 (about 20-fold) and other antimitotic drugs that bind in the vinca domain of tubulin,99 despite its potent antiproliferative effect on cellular microtubules (one-seventh as active as dolastatin 10, but 10 times more potent than the clinically useful anticancer drug vinblastine). Thus far, results from SAR studies on dolastatin 15 have not established a clear connection between cancer cell growth inhibition and a tubulin-based mechanism of action. Although all structural modifications of the C terminus of dolastatin 15 exhibited potent antineoplastic activity, most showed modest reduction of Copyright © 2005 CRC Press, LLC D-Me-Ala (less potent) ester or amide (equipotent) N-Me amide (loss of activity) Bn2N (less potent) N O O N N O O O HN O O N R1 O N 1 Hiva-Dpy structure 2 non-essential iPr, sBu (equipotent) O iBu (less potent) N - (CH2)n - Ar (n = 2 optimal) FIGURE 11.10 (See color insert) Effects of structural modifications on dolastatin 15 activity. inhibitory activity against tubulin polymerization. The opposite has been observed for dolastatin 10 derivatives, where potent inhibitory effect on tubulin polymerization (1–4-µM range) has been retained in spite of significant loss of antitumor activity (>1000-fold in some cases).122 It has been suggested that the potent cytotoxicity and inhibitory effects on microtubule assembly exhibited by dolastatin 15 are not a result of the parent molecule but of a more active metabolite derived from it.99 In this context, the reduced activity of dolastatin 15 in the purified tubulin assay could be explained by the absence of proteolytic enzymes normally present in cells. Furthermore, the fact that dolastatin 15 is active against prokaryotic cells that lack tubulin indicates a second cellular target. Derivatives 66–68 were prepared by Pettit’s group to probe whether intracellular hydrolysis of the ester bond of this depsipep- tide might be a necessary condition for in vitro activity.123 Interestingly, replacement of the ester with an N-methyl amide function (compound 68) led to complete loss of antitumor activity. Although the N-methyl amide modification would be expected to impart greater stability of the peptide bond toward enzymatic cleavage, conformational changes in the peptide backbone induced by N-methylation cannot be ruled out as the cause for the lack of potency in this compound. Among the most important discoveries made in the dolastatin series is that made by researchers at BASF, who found that the Hiva-Dpy units of dolastatin 15 may be completely replaced with a simple benzamide group without loss of activity (cemadotin, LU103793, 38; Figure 11.6).124 The progress of this water-soluble D15 derivative through clinical trials in humans is discussed in the next section. Hu and coworkers have described a series of derivatives based on the structure of cemadotin, in which modifications of the Dov (Bn2Val) and Pro (D-MeAla) residues as well as aromatic ring substitutions were evaluated.125 Although aromatic ring substituents altered the cyto- static properties only slightly, other modifications described led to more pronounced loss of activity in this series. Kessler and coworkers have prepared a peptidomimetic compound based on the structure of dolastatin 15, using N-substituted glycine monomers, but the resulting peptoid was devoid of cytostatic properties.126 A summary of dolastatin 15 structural requirements for antitumor activity is shown in Figure 11.10. The minimum-energy, solution-phase conformation of dolastatin 15 has been predicted by Poncet et al., using two-dimensional NMR techniques and molecular dynamics simulations.115 Despite the known tendency for proline residues to form peptide turn structures because of a decreased energy barrier between cis and trans amide bond conformations, dolastatin 15, which contains two contiguous proline units, appears to adopt a less-folded, linear shape (all trans amide bonds) in solution.121 VI. CLINICAL UPDATE Of the naturally occurring dolastatins, only dolastatin 10 has progressed to human clinical trials. Although dolastatin 15 has not been studied directly, synthetic derivatives of both dolastatins 10 Copyright © 2005 CRC Press, LLC (soblidotin, 37) and 15 (cemadotin, 38) are currently under evaluation in phase II clinical trials. The progress of these studies and the future potential of the dolastatins and its congeners are discussed in this section. Among the naturally occurring peptides of D. auricularia, dolastatin 10 was selected over dolastatin 15 for further evaluation in clinical trials because of its inhibitory effect on cancer cell growth at extremely low concentrations. Preclinical studies on both peptides indicated that myelo- toxicity would be a potentially limiting factor, as both compounds were found to inhibit bone marrow cell colony formation at drug concentrations ~100-fold lower than that required for anti- tumor activity in human tumor cell lines in vitro.127,128 Preclinical animal toxicity evaluations in mice, rats, and dogs showed that myelotoxicity was dose-limiting in all three species, with mice being the least sensitive, and a phase I clinical trial starting dose of 65 µg/M2 was recommended.127 Among phase I patients with advanced solid tumors, a maximum tolerated dose was reached at 300 µg/M2, and granulocytopenia (myelosuppression) was the dose-limiting toxicity, observed in 33% of patients at this dose.129 Other mild-to-moderate toxicities were also noted such as fatigue, nausea, and peripheral neuropathy (40% of patients).128 Clinical pharmacokinetic studies with five patients detected a single dolastatin 10 metabolite, corresponding to a more water-soluble, N- desmethyl derivative.129,130 Dolastatin 10, administered as an intravenous bolus injection, showed a plasma elimination half-life of 5–7 hours.80 Phase II clinical trials for dolastatin 10 in a number of cancer indications (metastatic prostate adenocarcinoma,131 metastatic melanoma,132 metastatic colorectal cancer,133 advanced non–small cell lung carcinoma,134 and platinum-sensitive ovarian carcinoma135) have been completed. The MTD has varied from patient to patient from 300 to 400 µg/M2. No significant antitumor activity has been observed to date for dolastatin 10, administered as a single agent to patients with advanced solid tumors. Dolastatin 10 does act synergistically with other anticancer substances (e.g., vincristine, bryostatin 1),95 providing some hope that this highly potent peptide may find use in drug combination therapies. Structure/activity work by Miyazaki118,119 and Pettit136 led to the discovery of the dolastatin 10 synthetic derivative soblidotin (37) that is currently in phase II clinical trials. Although less infor- mation concerning the results of preclinical work and phase I clinical trials is presently available for soblidotin, a pattern similar to that of dolastatin 10 would be expected because of their close structural analogy. Some preliminary results from a few phase I studies have been recently sum- marized.137,138 Soblidotin (single dose, intravenous feed at 1.35–2.4 mg/M2) was generally well tolerated among patients with advanced solid tumors, and induced antitumor effects were observed.139 A more recent study conducted with non–small cell lung carcinoma patients has generated some promising results. The drug (single dose, intravenous infusion) was reportedly well tolerated up to 4.8 mg/M2, and complete or partial responses were seen in 4 of 44 patients.140,141 These results are a positive sign for how this promising drug may fare in phase II clinical trials, although no further data are presently available. The third and final drug in the dolastatin family to reach phase II clinical trials in humans as of this writing is the dolastatin 15 derivative cemadotin (38). As with dolastatin 10, myelotoxicity (neutropenia) was found to be the principal dose-limiting toxicity, and a maximum tolerated dose of 12.5 mg/M2 was established for phase II patients.142 In an earlier phase I clinical study, cardio- vascular toxicity and hypertension were the dose-limiting toxicities, but these were later easily controlled via modification of the treatment schedule.143 Other nonhematologic toxicities were nausea and fatigue and were moderate. Interestingly, the primary metabolite detected for cemadotin is not due to modification of the N terminus (as with dolastatin 10) but, rather, from loss of the C- terminal benzamide group. This may constitute further evidence that the bioactive species in the case of dolastatin 15 is actually the peptide metabolite resulting from enzymolysis of the Pro-Hiva bond (for dolastatin 15) or Pro-benzamide bond (for cemadotin), as discussed in section V. Phase II results have been reported for trials involving patients with non–small cell lung carcinoma and metastatic breast cancer, but no positive responses have been noted with the current dose and schedule.144,145 Researchers at BASF have also pursued the use of cemadotin in drug combination Copyright © 2005 CRC Press, LLC therapies (e.g., with taxanes).146 Although the dolastatins have yet to demonstrate clinical efficacy against cancer, it should be noted that these compounds have been studies in biological domains for a relatively short period of time, and in this regard, much of what the dolastatins may ultimately offer for the treatment of human diseases such as cancer remains to be determined. REFERENCES 1. Chemistry of marine natural products, Scheuer, P. J. Ed., Academic Press, New York, NY, Vols. I and II, 1973; Vol. III 1980; Vol. IV, 1981; Vol. V. 1983. 2. Faulkner, D. J., Marine natural products, Nat. Prod. Rep., 12, 223, 1995. 3. Roberts, M., in A popular history of the mollusca, Reeve & Benham, London, 1851. 4. Paul, V. J. and Pennings, S. C., Diet-derived chemical defenses in the sea hare Stylocheilus longicauda (Quoy and Gaimard 1824), J. Expl. Mar. Biol. Ecol., 151, 227, 1991. 5. Pennings, S. C., Nadeau, M.

T., and Paul, V. J., Selectivity and growth of the generalist herbivore Dolabella auricularia feeding upon complementary resources, Ecology, 74, 879, 1993. 6. Pennings, S. C. and Paul, V. J., Sequestration of dietary secondary metabolites by three species of sea hares: location, specificity and dynamics, Mar. Biol., 117, 535, 1993. 7. De Nys, R. et al., Quantitative variation of secondary metabolites in the sea hare Aplysia parvula and its host plant Delisea pulchra, Mar. Ecol. Prog. Ser., 130, 135, 1996. 8. Faulkner, D. J., Chemical defenses of marine molluscs, in Ecological roles of marine natural products, Paul V. J., Ed., Comstock Publishing Associates, Ithaca, NY, 1992, 119. 9. Pliny, in Historica naturalis, 60 AD, Lib. 9, Lib. 32. 10. Donati, G. and Porfirio, B., Marine pharmacology and toxicology. The dolastatins, Conchiglia, 16,1984. 11. Eales, N. B., Typical british marine plants and animals, aplysia, in L.M.B.C. Memoirs Hardmann, W. A. and Johnstone, J., Eds, Liverpool, England, 1921. Vol. 24. 12. Halstead, B. W., Poisonous sea hares, in Poisonous and venemous marine animals of the world, Government Printing Office, Washington, DC, 1965. Vol. 1, 709 13. Sorokin, M., Human poisoning by ingestion of a sea hare (Dolabella auricularia), Toxicon, 26, 1095, 1998. 14. Pettit, G. R. et al., Antineoplastic components of marine animals, Nature, 227, 962, 1970. 15. Pettit, G. R., The dolastatins, Fortschr. Chem. Org. Naturst., 70, 1, 1997. 16. Pettit, G. R. et al., Anti-neoplastic agents. 69. The isolation of loliolide from an Indian Ocean opisthobranch mollusk, J. Nat. Prod., 43, 752, 1980. 17. Pettit, G. R. et al., Anti-neoplastic agents. 72. Marine animal biosynthetic constituents for cancer chemotherapy, J. Nat. Prod., 44, 482, 1981. 18. Pettit, G. R. et al., Isolation of dolastatins 10-15 from the marine mollusc Dolabella auricularia, Tetrahedron, 49, 9151, 1993. 19. Pettit, G. R., The isolation and structure of a remarkable marine animal antineoplastic constituent: dolastatin 10, J. Am. Chem. Soc., 109, 6883, 1987. 20. Pettit, G. R. et al., Isolation and structure of the human cancer cell growth inhibitory cyclodepsipeptide dolastatin 16, J. Nat. Prod., 60, 752, 1997. 21. Pettit, G. R. et al., Antineoplastic agents. 369. Isolation and structure of dolastatin 17, Heterocycles, 47, 491, 1998. 22. Pettit, G. R. et al., Antineoplastic agents. 370. Isolation and structure of dolastatin 18, Bioorg. Med. Chem. Lett, 7, 827, 1997. 23. Davies-Coleman, M. T. et al., Isolation of homodolastatin 16, a new cyclic depsipeptide from a Kenyan collection of Lyngbya majuscula, J. Nat. Prod., 66, 712, 2003. 24. Yamada, K. and Kigoshi, H., Bioactive compounds from the sea hares of two genera: Aplysia and Dolabella, Bull. Chem. Soc. Japan, 70, 1479, 1997. 25. Sone, H. et al., Isolation, structure, and synthesis of dolastatin C, a new depsipeptide from the sea hare Dolabella auricularia, Tetrahedron Lett., 34, 8445, 1993. 26. Sone, H. et al., Isolation, structure, and synthesis of dolastatin D, a cytotoxic cyclic depsipeptide from the sea hare Dolabella auricularia, Tetrahedron Lett., 34, 8449, 1993. Copyright © 2005 CRC Press, LLC 27. Ojika, M. et al., Dolastatin E, a new cyclic hexapeptide isolated from the sea hare Dolabella auric- ularia, Tetrahedron Lett., 36, 5057, 1995. 28. Nakamura, M. et al., Stereochemistry and total synthesis of dolastatin E, Tetrahedron Lett., 36, 5059, 1995. 29. Mutou, T. et al., Isolation and stereostructures of dolastatin G and nordolastatin G, cytotoxic 35- membered cyclodepsipeptides from the Japanese sea hare Dolabella auricularia, J. Org. Chem., 61, 6340, 1996. 30. Mutou, T. et al., Synthesis of dolastatin G and nordolastatin G, cytotoxic 35-membered cyclodep- sipeptides of marine origin, Tetrahedron Lett., 37, 7299, 1996. 31. Sone, H. et al., Dolastatin H and isodolastatin H, potent cytotoxic peptides from the sea hare Dolabella auricularia: isolation, stereo structures, and synthesis, J. Am. Chem. Soc., 118, 1874, 1996. 32. Sone, H., Kigoshi, H., and Yamada, K., Isolation and stereostructure of dolastatin I, a cytotoxic cyclic hexapeptide from the Japanese sea hare Dolabella auricularia, Tetrahedron, 53, 8149, 1997. 33. Suenaga, K. et al., Isolation and stereostructure of aurilide, a novel cyclodepsipeptide from the Japanese sea hare I, Tetrahedron Lett., 37, 6771, 1996. 34. Mutou, T. et al., Enantioselective synthesis of aurilide, a cytotoxic 26-membered cyclodepsipeptide of marine origin, Synlett, 199, 1997. 35. Sone, H. et al., Dolabellin, a cytotoxic bisthiazole metabolite from the sea hare Dolabella auricularia: structural determination and synthesis, J. Org. Chem., 60, 4774, 1995. 36. Ishiwata, H. et al., Isolation and stereostructure of doliculide, a cytotoxic cyclodepsipeptide from the Japanese sea hare Dolabella auricularia, J. Org. Chem., 59, 4710, 1994. 37. Ojika, M., Nemoto, T., and Yamada, K., Doliculol-A and doliculol-B, the nonhalogenated C-15 acetogenins with cyclic ether from the sea hare Dolabella-auricularia, Tetrahedron Lett., 34, 3461, 1993. 38. Faulkner, D. J., Marine natural-products — metabolites of marine-algae and herbivorous marine mollusks, Nat. Prod. Rep., 1, 251, 1984. 39. Fattorusso, E. and Piattelli, M., Amino acids from marine algae, in Marine Natural Products, Chemical and Biological Perspectives, Scheuer, P. J., Ed., Academic Press, New York, 1980, Vol. III, 95. 40. Mitchell, S. S. et al., Dolastatin 3 and two novel cyclic peptides from a Palauan collection of Lyngbya majuscula, J. Nat. Prod., 63, 279, 2000. 41. Luesch, H. et al., Isolation of dolastatin 10 from the marine cyanobacterium Symploca species VP642 and total stereochemistry and biological evaluation of its analogue symplostatin 1, J. Nat. Prod., 64, 907, 2001. 42. Harrigan, G. G. et al., Isolation, structure determination, and biological activity of dolastatin 12 and lyngbyastatin I from Lyngbya majuscula Schizothrix calcicola cyanobacterial assemblages, J. Nat. Prod., 61, 1221, 1998. 43. Nogle, L. M., Williamson, R. T., and Gerwick, W. H., Somamides A and B, two new depsipeptide analogues of dolastatin 13 from a Fijian cyanobacterial assemblage of Lyngbya majuscula and Schizo- thrix species, J. Nat. Prod., 64, 716, 2001. 44. Nogle, L. M. and Gerwick, W. H., Isolation of four new cyclic depsipeptides, antanapeptins A-D, and dolastatin 16 from a Madagascan collection of Lyngbya majuscula, J. Nat. Prod., 65, 21, 2002. 45. Harrigan, G. G. et al., Symplostatin 1: a dolastatin 10 analogue from the marine cyanobacterium Symploca hydnoides, J. Nat. Prod., 61, 1075, 1998. 46. Harrigan, G. G. et al., Symplostatin 2: a dolastatin 13 analogue from the marine cyanobacterium Symploca hydnoides, J. Nat. Prod., 62, 655, 1999. 47. Luesch, H. et al., Symplostatin 3, a new dolastatin 10 analogue from the marine cyanobacterium Symploca sp VP452, J. Nat. Prod., 65, 16, 2002. 48. Luesch, H. et al., The cyanobacterial origin of potent anticancer agents originally isolated from sea hares, Curr. Med. Chem., 9, 1791, 2002. 49. Schmidt, U. et al., Amino acids and peptides; 60. Synthesis of biologically active cyclopeptides; 10. Synthesis of 16 structural isomers of dolastatin 3; ii: synthesis of the linear educts and the cyclopep- tides, Synthesis, 236, 1987. 50. Hamada, Y., Kato, S., and Shioiri, T., New methods and reagents in organic-synthesis 51. A synthesis of ascidiacyclamide, a cyto-toxic cyclic peptide from ascidian — determination of its absolute- configuration, Tetrahedron Lett., 26, 3223, 1985. Copyright © 2005 CRC Press, LLC 51. Arai, H. et al., Synthesis of the pyrimidine moieties of bleomycin and epibleomycin, J. Am. Chem. Soc., 102, 6631, 1980. 52. Bodanszky, M., Scozzie, J. A., and Muramatsu, I., Dehydroalanine residues in thiostrepton, J. Antibiot., 23, 9, 1970. 53. Pettit, G. R. et al., Antineoplastic agents. 150. The structure and synthesis of dolastatin-3, J. Am. Chem. Soc., 109, 7581, 1987. 54. Schmidt, U. and Utz, R., Synthetic studies on the elucidation of the structure and configuration of dolastatin 3, Angew. Chem. Int. Ed. Eng., 23, 725, 1984. 55. Hamada, Y., Kohda, K., and Shioiri, T., Proposed structure of the cyclic peptide dolastatin-3, a powerful cell-growth inhibitor, should be revised, Tetrahedron Lett., 25, 5303, 1984. 56. Holzapfel, C. W. and Van Zyl, W. J., The synthesis and conformation in solution of cyclo[l-pro-l-leu- (gln)thz-(gly)thz-l-val] (dolastatin 3), Tetrahedron, 46, 649, 1990. 57. Rinehart, Jr., K. L. et al., Biologically-active peptides and their mass-spectra, Pure Appl. Chem., 54, 2409, 1982. 58. Pettit, G. R., Dolastatin 10, Fortschr. Chem. Org. Naturst., 70, 14, 1997. 59. Pettit, G. R. et al., Antineoplastic agents. 189. The absolute-configuration and synthesis of natural (-)- dolastatin-10, J. Am. Chem. Soc., 111, 5463, 1989. 60. Hamada, Y., Hayashi, K., and Shioiri, T., Efficient stereoselective synthesis of dolastatin-10, an antineoplastic peptide from a sea hare, Tetrahedron Lett., 32, 931, 1991. 61. Shioiri, T., Hayashi, K., and Hamada, Y., Stereoselective synthesis of dolastatin 10 and its congeners, Tetrahedron, 49, 1913, 1993. 62. Tomioka, K., Kamai, M., and Koga, K., An expeditious synthesis of dolastatin-10, Tetrahedron Lett., 32, 2395, 1991. 63. Roux, F. et al., Synthesis of dolastatin-10 and [r-doe]-dolastatin-10, Tetrahedron, 50, 5345, 1994. 64. Pettit, G. R. et al., Antineoplastic agents. 173. Isolation and structure of the cell growth inhibitory depsipeptides dolastatins 11 and 12, Heterocycles, 28, 553, 1989. 65. Bates, R. B. and Gangwar, S., Asymmetric syntheses of 3-amino-2-methylpentanoic acids. Configu- rations of the beta-amino acid in majusculamide C, 57-normajusculamide C and dolastatins 11 and 12, Tetrahedron Asymm., 4, 69, 1993. 66. Bates, R. B. et al., Dolastatins. 26. Synthesis and stereochemistry of dolastatin 11, J. Am. Chem Soc., 119, 2111, 1997. 67. Carter, D. C. et al., Structure of majusculamide-C, a cyclic depsipeptide from lyngbya-majuscula, J. Org. Chem., 49, 236, 1984. 68. Mynderse, J. S., Hunt, A. H., and Moore, R. E., 57-normajusculamide-C, a minor cyclic depsipeptide isolated from Lyngbya-majuscula, J. Nat. Prod., 51, 1299, 1988. 69. Pettit, G. R. et al., Antineoplastic agents. 174. Isolation and structure of the cytostatic depsipeptide dolastatin 13 from the sea hare Dolabella auricularia, J. Am. Chem. Soc., 111, 5015, 1989. 70. Sih, C. J. et al., Biocatalytic methods for enantioselective synthesis, Stereocontrolled Organic Syn- thesis, 399, 1994. 71. Moune, S. et al., Total synthesis of dolatrienoic acid: A subunit of dolastatin 14, J. Org. Chem., 62, 3332, 1997. 72. Duffield, J. J. and Pettit, G. R., Synthesis of (7s,15s)- and (7r,15s)-dolatrienoic acid, J. Nat. Prod., 64, 472 2001. 73. Pettit, G. R. et al., Antineoplastic agents. 220. Synthesis of natural (-)-dolastatin 15, J. Am. Chem. Soc., 113, 6692, 1991. 74. Patino, N. et al., Total synthesis of the proposed structure of dolastatin-15, Tetrahedron, 48, 4115, 1992. 75. Pettit, G. R. et al., The dolastatins-20 — a convenient synthetic route to dolastatin-15, Tetrahedron, 50, 12097, 1994. 76. Akaji, K. et al., Convergent synthesis of dolastatin 15 by solid phase coupling of n-methylamino acid, Peptide Science, 35, 9, 1999. 77. Pettit, G. R., Structural modifications of dolastatin 10, Fortschr. Chem. Org. Naturst., 70, 42, 1997. 78. Pettit, G. R., Hogan, F., and Herald, D. L., Synthesis and x-ray crystal structure of the Dolabella auricularia peptide dolastatin 18, J. Org Chem., 69, 4019, 2004. 79. Poncet, J., The dolastatins, a family of promising antineoplastic agents, Current Pharm Design, 5, 139, 1999. Copyright © 2005 CRC Press, LLC 80. Aherne, G. W. et al., Antitumour evaluation of dolastatins 10 and 15 and their measurement in plasma by radioimmunoassay, Cancer Chemother. Pharmacol., 38, 225, 1996. 81. Pietra, F., A secret world, in The natural products of marine life, Basel, Switzerland, 1990, 39. 82. Yamazaki, M., Iijima, R., and Kosuna, K., Antibacterial and antifungal peptide from Dolabella auricularia, Japanese Patent JP10251297, 22 SEP 1998. 83. Donia, M. and Hamann, M. T., Marine natural products and their potential applications as anti-infective agents, Lancet Infect. Dis., 3, 338, 2003. 84. Fennell, B. J. et al., Effects of the antimitotic natural product dolastatin 10, and related peptides, on the human malarial parasite Plasmodium falciparum, J. Antimicrobial Chemother., 51, 833, 2003. 85. Hamel, E., Interaction of tubulin with small ligands, in Microtubule proteins, Avila, J., Ed., CRC Press, Boca Raton, FL, 1990, 89. 86. Bai, R., Pettit, G. R., and Hamel, E., Structure activity studies with chiral isomers and with segments of the antimitotic marine peptide dolastatin-10, Biochem. Pharmacol., 40, 1859, 1990. 87. Bai, R., Pettit, G. R., and Hamel, E., Binding of dolastatin-10 to tubulin at a distinct site for peptide antimitotic agents near the exchangeable nucleotide and vinca alkaloid sites, J. Biol. Chem., 265, 17141, 1990. 88. Quentmeier, H. et al., Cytostatic effects of dolastatin 10 and dolastatin 15 on human leukemia cancer cell lines, Leukemia Lymphoma, 6, 245, 1992. 89. Steube, K. G. et al., Inhibition of cellular proliferation by the natural peptides dolastatin 10 and dolastatin 15, Mol. Biol. Haematapoiesis, 2, 567, 1992. 90.

Steube, K. G. et al., Dolastatin 10 and dolastatin 15: Effect of two natural peptides on growth and differentiation of leukemia cells, Leukemia, 6, 1048, 1992. 91. Hu, Z.-B. et al., Effects of dolastatins on human B-lymphocytic leukemia cell lines, Leukemia Res., 17, 333, 1993. 92. Mohammad, R. M. et al., Successful treatment of human chronic lymphocytic leukemia xenografts with combination biological agents auristatin PE and bryostatin 1, Clin. Cancer Res., 4, 1337, 1998. 93. Beckwith, M., Urba, W. J., and Longo, D. L., Growth inhibition of human lymphoma cell lines by the marine natural products dolastatins 10 and 15, J. Natl. Cancer Inst., 85, 483, 1993. 94. Maki, A. et al., Effect of dolastatin 10 on human non-Hodgkin’s lymphoma cell lines, Anti-Cancer Drugs, 7, 344, 1996. 95. Mohammad, R. M. et al., Synergistic interaction of selected marine animal anticancer drugs against human diffuse large cell lymphoma, Anti-Cancer Drugs, 9, 149, 1998. 96. Turner, T. et al., Treatment of human prostate cancer cells with dolastatin 10, a peptide isolated from a marine shell-less mollusc, Prostate, 34, 175, 1998. 97. Hamel, E., Natural products which interact with tubulin in the vinca domain: Maytansine, rhizoxin, phomopsin a, dolastatins 10 and 15 and halichondrin B, Pharmacol. Ther., 55, 31, 1992. 98. Pettit, G. R. et al., Antineoplastic agents. 205. Chiral modifications of dolastatin-10 — the potent cytostatic peptide (19ar)-isodolastatin-10, J. Med. Chem., 33, 3132, 1990. 99. Bai, R. et al., Dolastatin-15, a potent antimitotic depsipeptide derived from Dolabella-auricularia — interaction with tubulin and effects on cellular microtubules, Biochem Pharmacol., 43, 2637, 1992. 100. Jordan, M. A. et al., Suppression of microtubule dynamics by binding of cemadotin to tubulin: possible mechanism for its antitumor action, Biochemistry, 37, 17571, 1998. 101. Cruz-Monserrate, Z. et al., Dolastatin 15 binds in the vinca domain of tubulin as demonstrated by Hummel-Dreyer chromatography, Eur. J. Biochem., 270, 3822, 2003. 102. Haldar, S., Basu, A., and Croce, C. M., Bcl2 is the guardian of microtubule integrity, Cancer Res., 57, 229, 1997. 103. Haldar, S., Basu, A., and Croce, C. M., Serine-70 is one of the critical sites for drug-induced bcl2 phosphorylation in cancer cells, Cancer Res., 58, 1609, 1998. 104. Maki, A. et al., The bcl-2 and p53 oncoproteins can be modulated by bryostatin-1 and dolastatins in human diffuse large-cell lymphoma, Anti-Cancer Drugs, 6, 392, 1995. 105. Ali, M. A. et al., Dolastatin 15 induces apoptosis and bcl-2 phosphorylation in small cell lung cancer cell lines, Anticancer Res., 18, 1021, 1998. 106. Mooberry, S. L. et al., The molecular pharmacology of symplostatin 1: a new antimitotic dolastatin 10 analog, International J. Cancer, 104, 512, 2003. Copyright © 2005 CRC Press, LLC 107. Watanabe, J. et al., Induction of apoptosis in human cancer cells by TZT-1027, an antimicrotubule agent, Apoptosis, 5, 345, 2000. 108. Ikeda, R. et al., Induction of apoptosis in mice with B16 melanoma by treatment with the antimicro- tubule agent TZT-1027, Acta Histochem. Cytochem., 33, 341, 2000. 109. Otani, M. et al., TZT-1027, an antimicrotubule agent, attacks tumor vasculature and induces tumor cell death, Jpn. J. Cancer Res, 91, 837, 2000. 110. Bai, R. et al., Dolastatin 11, a marine depsipeptide, arrests cells at cytokinesis and induces hyperpo- lymerization of purified actin, Mol. Pharmacol., 59, 462, 2001. 111. Boyd, M. R. and Paull, K., Some practical considerations and applications of the National-Cancer- Institute in-vitro anticancer drug discovery screen, Drug Dev. Res., 34, 91, 1995. 112. Pettit, G. R. et al., The dolastatins. 21. Synthesis, x-ray crystal-structure, and molecular modeling of (6r)-isodolastatin-10, J. Org. Chem., 59, 6127, 1994. 113. Pettit, G. R. et al., Antineoplastic agents. 337. Synthesis of dolastatin 10 structural modifications, Anti-Cancer Drug Design, 10, 529, 1995. 114. Pettit, G. R. et al., Antineoplastic agents 365. Dolastatin 10 SAR probes, Anti-Cancer Drug Design, 13, 243, 1998. 115. Alattia, T. et al., Conformational study of dolastatin-10, Tetrahedron, 51, 2593, 1995. 116. Poncet, J. et al., Synthesis and antiproliferative activity of a cyclic analog of dolastatin 10, Bioorg. Med. Chem., 8, 2855, 1998. 117. Poncet, J. et al., Synthesis and biological activity of chimeric structures derived from the cytotoxic natural compounds dolastatin 10 and dolastatin 15, J. Med. Chem., 41, 1524, 1998. 118. Miyazaki, K. et al., Synthesis and antitumor-activity of novel dolastatin-10 analogs, Chem. Pharm. Bull., 43, 1706, 1995. 119. Kobayashi, M. et al., Antitumor activity of TZT-1027, a novel dolastatin 10 derivative, Jpn. J. Cancer Res., 88, 316, 1997. 120. Alcaro, S. et al., Theoretical comparison between structural and dynamical features of dolastatins 11 and 12 antineoplastic depsipeptides, SAR QSAR Environ. Res., 14, 475, 2003. 121. Roux, F. et al., Synthesis and in-vitro cytotoxicity of diastereoisomerically modified dolastatin-15 analogs, Bioorg. Med. Chem. Lett., 4, 1947, 1994. 122. Pettit, G. R. et al., Antineoplastic agents 360. Synthesis and cancer cell growth inhibitory studies of dolastatin 15 structural modifications, Anti-Cancer Drug Design, 13, 47, 1998. 123. Flahive, E. J., The dolastatins: synthesis and structural modification of dolastatin 15, Ph.D. Disserta- tion, Arizona State University, Tempe, AZ, 1996. 124. De Arruda, M. et al., Lu103793 (NSC D-669356): a synthetic peptide that interacts with microtubules and inhibits mitosis, Cancer Res., 55, 3085, 1995. 125. Hu, M. K. and Huang, W.-S., Synthesis and cytostatic properties of structure simplified analogues of dolastatin 15, J. Peptide Res., 54, 460, 1999. 126. Schmitt, J. et al., Synthesis of dolastatin 15 mimetic peptoids, Bioorg. Med. Chem. Lett., 8, 385, 1998. 127. Mirsalis, J. C. et al., Toxicity of dolastatin 10 in mice, rats and dogs and its clinical relevance, Cancer Chemother. Pharmacol., 44, 395, 1999. 128. Schwartsmann, G. et al., Marine-derived anticancer agents in clinical trials, Exp. Opin. Investig. Drugs, 12, 1367, 2003. 129. Madden, T. et al., Novel marine-derived anticancer agents: a phase I clinical, pharmacological and pharmacodynamic study of dolastatin 10, Clinical Cancer Res., 6, 1293, 2000. 130. Garteiz, D. A. et al., Quantitation of dolastatin 10 using HPLC/electrospray ionization mass spec- trometry: application in a phase I clinical trial, Cancer Chemother. Pharmacol., 41, 299, 1998. 131. Vaishampayan, U. et al., Phase II study of dolastatin 10 in patients with hormone-refractory metastatic prostrate adenocarcinoma, Clinical Cancer Res., 6, 4205, 2000. 132. Margolin, K. et al., Dolastatin 10 in metastatic melanoma: a phase II and pharmacokinetic trial of the California Cancer Consortium, Investig. New Drugs, 19, 335, 2001. 133. Aguayo, A. et al., Phase II study of dolastatin 10 administered intravenously every 21 days to patients with metastatic colorectal cancer, Proc. Am. Soc. Clin. Oncol., 1125 (Abs), 2000. 134. Krug, L. M., Phase II study of dolastatin 10 in patients with non-small-cell lung carcinoma, Ann. Oncol., 11, 227, 2000. Copyright © 2005 CRC Press, LLC 135. Hoffman, M. A., Blessing, J. A., and Lentz, S. S., A phase II trial of dolastatin 10 in recurrent platinum- sensitive ovarian carcinoma: A gynecologic oncology group study, Gynecol. Oncol., 89, 95, 2003. 136. Pettit, G. R., Williams, M. D., and Srirangam, J. K., Preparation of dolastatin analog pentapeptide methyl esters as anticancer agents, Eur. Pat. Appl. EP 95-305130, 1995. 137. Bayés, M., Rabasseda, X., and Prous, J. R., Gateways to clinical trials, Methods Find. Exp. Clin. Pharmacol., 25, 387, 2003. 138. Bayés, M., Rabasseda, X., and Prous, J. R., Gateways to clinical trials, Methods Find. Exp. Clin. Pharmacol., 26, 53, 2004. 139. Verweij, J. et al., Phase I and pharmacologic study of TZT-1027, a novel dolastatin 10 derivative, administered in a day 1 and 8 iv schedule every 3 weeks in patients (pts) with advanced solid tumors, Clin. Cancer Res., 16 (Suppl.), A255 (Abs.), 2003. 140. Horti, J., Juhasz, E., and Bodrogi, I., Preliminary results of a phase I trial of TZT-1027, an inhibitor of tubulin polymerization, in patients with advanced non-small cell lung carcinoma, Proc. Am. Assoc. Cancer Res., 43, 2744 (Abs.), 2002. 141. Horti, J. et al., A phase I trial of TZT-1027, an inhibitor of tubulin polymerization, in patients with advanced non-small cell lung cancer (NSCLC), Clin. Cancer Res., 16 (Suppl.), A256 (Abs.), 2003. 142. Supko, J. G. et al., A phase I clinical and pharmacokinetic study of the dolastatin analogue cemadotin administered as a 5-day continuous intravenous infusion, Cancer Chemother. Pharmacol., 46, 319, 2000. 143. Mross, K. et al., Phase I clinical and pharmacokinetic study of lu103793 (cemadotin hydrochloride) as an intravenous bolus injection in patients with metastatic solid tumors, Onkologie, 19, 405, 1996. 144. Marks, R. S. et al., A phase II study of the dolastatin 15 analogue lu103793 in the treatment of advanced non-small cell lung cancer, Am. J. Clin. Oncol., 26, 336, 2003. 145. Kerbrat, P. et al., Phase II study of LU103793 (dolastatin analogue) in patients with metastatic breast cancer, Eur. J. Cancer, 39, 317, 2003. 146. Barlozzari, T. and Haupt, A., Dolastatin-15 derivatives in combination with taxanes, U.S. Patent 6,103, 618, 15 AUG 2000. Copyright © 2005 CRC Press, LLC 12 Ecteinascidin 743 (ET-743; Yondelis™), Aplidin, and Kahalalide F Rubén Henríquez, Glynn Faircloth, and Carmen Cuevas CONTENTS I. Introduction II. Ecteinascidin 743 (ET-743; Yondelis) A. Background B. Mechanism of Action C. Preclinical Drug Development 1. In Vitro and In Vivo Testing 2. Preclinical Toxicology D. Chemical Synthesis 1. Synthetic Routes to Ecteinascidin 743 (Yondelis) E. Clinical Studies 1. Yondelis (ET-743) in STS (Soft Tissue Sarcoma) 2. Yondelis in Ovarian Cancer 3. Safety Profile III. Aplidin A. Background B. Mechanism of Action C. Preclinical Drug Development 1. In Vitro and In Vivo Testing 2. Preclinical Toxicology D. Chemical Synthesis E. Clinical Studies IV. Kahalalide F A. Background B. Mechanism of Action C. Drug Development 1. In Vitro and In Vivo Testing 2. Preclinical Toxicology D. Synthesis of Kahalalide F E. Clinical Trials References I. INTRODUCTION The natural world, both terrestrial and marine, is a varied and bountiful source for new drugs. Yet until recently only natural products from terrestrial sources have been developed into, or have Copyright © 2005 CRC Press, LLC FIGURE 12.1 (See color insert) Structure of Ecteinascidin 743 (Yondelis™) and source organism, Ecteinascidia turbinata. formed the basis for, modern medicines. The advent of more sophisticated and efficient means to collect marine organisms over the last few decades has allowed marine biologists, natural products chemists, and medicinal scientists from academia and the pharmaceutical industry to collaborate in unique and enterprising ways to sample the great variety and bounty of intertidal, shallow, and deep-water sea life in the world. Recently these efforts have resulted in the development of a first generation of drugs from the sea into clinical trials that includes the PharmaMar compounds, Yondelis™, Aplidin®, Kahalalide, and ES 285. II. ECTEINASCIDIN 743 (ET-743; YONDELIS) A. BACKGROUND Ecteinascidia turbinata is a tunicate species from the Caribbean and the Mediterranean that belongs to the class Ascidiacea within the subphylum Tunicata (also called Urochordata). Ascidians, or sea squirts, are small, bottom-dwelling soft-bodied marine animals that form colonies comprising many individuals, called zooids. The name tunicate is derived from their characteristic protective covering, or tunic, which functions to a certain extent as an external skeleton and consists of some cells, blood vessels, and a secretion of a variety of proteins and carbohydrates, including cellulose — an unusual finding in animals. Within the tunic is the muscular body wall, which controls the opening of the siphons used for feeding. Colonies are formed by asexual reproduction through budding; that is, outgrowths on the parent break off as new individuals. Sexual reproduction, in contrast, leads to a fertilized egg that develops into a free-swimming tadpole larva. The larval stage is brief and is used to find an appropriate place for the adult to live. In 1969, early assays with crude aqueous ethanol extracts of Ecteinacidia turbinata showed impressive in vitro cytotoxicity against a variety of cancer cell lines. Many unsuccessful attempts at isolation of the active compounds were followed by concerted efforts that led to the characterization of six alkaloids called ecteinas- cidins 729, 743, 745, 729A, 759B, and 770 by Kenneth Rinehart and his coworkers from the University of Illinois, Urbana-Champaign, of which ET 743 was the most abundant (0.0001% yield) (Figure 12.1).1 Ecteinascidin-743 (ET-743; molecular formula: C39H43N3O11S; relative molecular mass: 761.8) has the international nonproprietary name of trabectedin, is being developed under the trademark name of Yondelis, and the full chemical name of [6R-(6α, 6aβ, 7β, 13β, 14β, 16α, 20R*)]-5- (acetyloxy)-3,4,6,6a,7,13,14,16,-octahydro-6,8,14-trihydroxy-7,9-dimethoxy-4,10,23-trimethyl- Copyright © 2005 CRC Press, LLC spiro[6,16(epithiopropanoxymethano)-7,13-imino-12H-1,3-dioxolo[7,8]-isoquino[3,2-b][3]benza- zocine-20, 1 (2H)-isoquinolin]-19-one. It is a white to pale yellow amorphous powder, is

hydro- phobic, and is very soluble in methanol, ethanol, chloroform, ethyl acetate, and acetone. The novel and unique chemical structure of ecteinascidins is formed by a monobridged penta- cyclic skeleton composed of two fused tetrahydroisoquinoline rings linked to a 10-member lactone bridge through a benzylic sulfide linkage. Most ecteinascidins have an additional tetrahydroiso- quinoline ring attached to the rest of the structure through a spiro ring. This is one of the features distinguishing these molecules from saframycins, safracins, and renieramycins,2 compounds iso- lated from bacteria and sponges. B. MECHANISM OF ACTION Ecteinascidin 743 (ET743; trabectedin) is the first of a new class of DNA binding agents with a complex, transcription-targeted mechanism of action. Ecteinascidin 743 inhibits tumor cell prolif- eration via a promoter-specific mechanism and transcription-dependent nucleotide excision repair. Cell cycle studies on tumor cells in vitro reveal that ET-743 decreases the rate of progression of the cells through S phase toward G2 and causes a prolonged blockade in G2/M at biologically relevant concentrations (20–80 nM).3,4 The cell cycle block is p53 independent and eventually leads to a strong apoptotic response. The major findings that shed light on the primary mechanisms of action are detailed below. Ecteinascidin 743 binds to DNA through the minor groove and, through its C21 carbon, alkylates the N2 position of the guanine residue, preferentially flanked by guanine or cytosine on the 3 side and a pyrimidine on the 5 side.5–7 Remarkably, the binding of ET 743 to DNA is reversible under nondenaturing conditions, and the rate of reversion is greatest at nonpreferred sequences, allowing it to “walk” from its less favored target sequences toward its most favored target sequences.8 Ecteinascidin 743 strongly inhibits the activation of the transcription of certain genes without affecting basal levels of transcription.9,10 It appears to selectively target a set of cellular genes in a promoter-specific manner, of which notable examples are p21(WAF1/CIP1), a fundamental regulator of cell division that is very frequently found mutated in tumors; c-Fos and c-Jun, which are oncogenes functioning as transcription factors central to cell regulation; and MDR1, a gene critically involved in resistance to many chemotherapeutic agents.11 The expression of genes targeted by ET-743 is normally under the control of transcription factor Sp1, and current studies implicate this drug as a transcription factor–specific, CCAAT box–dependent inhibitor of tran- scription activation.12 ET-743 has also been shown to counteract the activation of the steroid and xenobiotic receptor (SXR).13 The orphan nuclear SXR controls MDR1 and the liver enzyme P450 3A4, thereby providing a possible mechanism for modulating degradation and efflux of xenobiotic drugs in normal liver cells and cells lining the intestine, respectively. Finally, ET-743 also blocks transcription activation of reporter genes. It has been shown that ET-743 blocked Trichostatin A- induced activation of the Sp1-regulated p21 gene at concentrations that had a minimal effect on uninduced expression.18 Through a mechanism not yet fully understood, the transcription-dependent nucleotide excision repair system is also involved in mediating some of the actions of this drug.14–17 Unexpectedly, mutations rendering defects in the cellular transcription–coupled DNA repair machinery were reported to result in resistance to ET-743, and reduced cytotoxic activity of Yondelis against cells that are deficient in the nucleotide repair machinery was observed, indicating that some component of this machinery is involved in mediating tumor cell killing by ET 743. A model has been proposed explaining the role of the TC-NER process in the production of abortive intermediates of an incomplete repair process, leading to cell death. Other anticancer drugs such as cisplatin, in clinical use, and Irofulven, in trials, exhibit the opposite behavior with regard to TC-NER condition. These compounds are more active against cells with a faulty TC-NER, as is expected for traditional DNA alkylating agents. Copyright © 2005 CRC Press, LLC Taking all these data into account, it is very tempting to propose a mechanistic link between ET-743 and the transcription and repair protein complexes associated with RNA pol II while plowing through coding regions of activated genes. TC-NER or other transcriptional targets may be important in the drug’s specific cellular toxicity profile. C. PRECLINICAL DRUG DEVELOPMENT 1. In Vitro and In Vivo Testing Preclinical data generated during the development of ET-743 have provided important insight for the selection and design of the clinical trials. Early in vitro studies carried out by PharmaMar and the National Cancer Institute (NCI) identified potent (1 pM to 10 nM) activity of ET-743 against cell line subpanels containing murine leukemias and human solid tumors, such as melanoma, breast, non–small cell lung (NSCL), and ovarian cancer. ET-743 was subsequently evaluated in a human tumor colony forming unit (TCFU) assay derived from surgically ablated primary tumor speci- mens.18 Over a range of concentrations (1 nM to 10 µM), ET-743 activity was seen against cancers of the head and neck, ovarian, mesothelioma, breast, kidney, peritoneum, NSCL, melanoma, sarcomas, and primary tumor specimens. The NCI COMPARE analysis of ET-743 (NSC648766) and the Standard Anticancer Agent Database showed significant correlation coefficients (0.962–0.876) to that of the DNA intercalators such as the anthracyclines and the morpholino compounds (NCI data). ET-743 has been tested in a great variety and number of models against tumors of murine origin (P388 leukemia and B16 melanoma) and human xenografts (melanoma, MEXF 989; NSLC, LXFL 529; breast, MX-1 early and advanced; and ovarian, HOC 22 and HOC 18) (Table 12.1).19–22 Drug-resistant lines have also been tested that include melanoma (MEXF 514), NSLC (LXFL 629), and ovarian (HOC 18) tumor lines. These first efforts showed that ET-743 has a broad spectrum of antineoplastic activity, with several tumor types showing selectivity; namely, melanoma, NSCL, and ovarian carcinomas. As a further example of strong activity and long-lasting antitumor effects, the action of ET-743 on human endometrial carcinoma xenografts (HEC-1-B) results in complete regression lasting for more than 125 days. These in vivo findings correlate strongly with the results from in vitro cross-resistance studies in which ET-743 was very active against many resistant tumors and only moderately cross-resistant to several of the standard agents.18 The determination that soft tissue sarcomas (STSs) are more sensitive to ET-743 than other solid tumors was not predicted during the preclinical development of the drug. The finding was serendipitous and came from the prevalence of responding or stable STS patients in the clinical trials. As a result, there has been a considerable effort to confirm this finding to supplement the extensive nonclinical profile already characterized for the antineoplastic effect of ET-743 against other solid tumors. In fact, a variety of sarcomas are differentially sensitive to ET-743, showing IC50 potencies in the picomolar and subpicomolar range compared to the nanomolar concentrations established against non-STS solid tumors (Table 12.2).23,24 During its clinical development, several in vivo studies of ET-743 were performed in represen- tative sarcoma tissues.19–31 ET-743 shows a strong activity against several sarcoma types, especially chondrosarcomas, where tumor growth can be inhibited as much as 90% at relatively low doses (about 20% Maximum Tolerated Dose [MTD] in mice) and correlates with the in vitro findings.24 In contrast, the in vivo effects of ET-743 against some soft-tissue sarcoma models, such as SW634 and UW2237 fibrosarcomas, appear to be less marked than in chondrosarcomas or osteosarcomas.25 2. Preclinical Toxicology Early preclinical toxicology studies were independently carried out both by the NCI (Bethesda, MD)32,33 and by PharmaMar in collaboration with the New Drug Development Office (Amsterdam). More specific nonclinical studies are currently being performed in collaboration with Johnson & Copyright © 2005 CRC Press, LLC TABLE 12.1 In vivo Antitumor Activity of ET-743 in Murine and Human Tumors Optimal Dose Tumor-free Cancer Type (Cell Line) Regimen (µg/kg per d) Antitumor Effect Survivors Leukemia i.p. P388 i.p. qd×5 27 48% ILS None Melanoma i.p.B16 i.p. qd×9 13.4 78% ILS 1/10 (d60) s.c. MEXF 989a i.v. q4d×3 100 0.2% T/C 6/6 (d35) i.v. (qd×3) × 2 50 1.8% T/C none s.c. MEXF 514a i.v. q4d × 3 100 37% T/C none DTIC resistant i.v. (qd×3) × 2 50 76% T/C (inactive) none NSCL s.c. LXFL 529a i.v. q4d×3 150 0.1% T/C 3/4 (d63) i.v. (qd×3)×2 50 0.1% T/C 2/4 (d40) s.c. LXFA 629b i.v. q4d×3 200 70% T/C (inactive) none Ifos resistant i.v. qd×3 200 90% T/(inactive) none Ovarian s.c. HOC 22 i.v. q4d×3 200 <1.0% T/C 5/6 (d102) i.v. (qd×3) × 3 100 2.0% T/C 3/5 (d102) s.c. HOC 18 i.v. q4d×3 200 5.8% T/C 4/8 (d100) CDDP resistant i.v. (qd×3) × 2 11.8% T/C 2/6 (d100) Breast s.c. MX-1 early i.p. qd×9 9 no activity none i.p. qd×5 18 no activity none i.v. qd×5 40 >160% growth delay 7/10 i.v. qd×5 60 0.1% T/C 10/10 (d47) i.v. q4d×3 75 0.1% T/C 9/10 (d23) s.c. MX-1 advancedc i.v. qd×1 (d9) 250 203% growth delay 4/10 (d58) i.v. qd×5 50 224% growth delay 3/10 (d58) i.v. q3d×4 41.7 161% growth delay 3/10 (d58) s.c. G101metastaticd i.v. qd×1 250 50% (d70) delay 0/10 (d30) Endometrial s.c. HEC-1-Be i.v. q4d×3 200 < 1% T/C 3/6 (d125) Prostate s.c. PC-3 i.v. q4d×3 200 30% T/C 0/6 (d30) Soft Tissue Sarcomas Chondrosarcoma (CHSA) i.v. qd×2 100 24% T/C ND Fibrosarcoma (UV2237) i.v. qd×1 200 37% T/C ND Osteosarcoma (OSA-FH) i.v. qd×5 40 33% T/C ND Rhabdomyosarc (TE-671) i.v. qd×1 300 [toxic] ND Reticulocell of ovary (M5076) i.v. q7d×2 150 36% T/C ND Note: d = study day, i.p. = intraperitoneal, ILS = % increase in life span, i.v. = intravenous, s.c. = subcutaneous, T/C = % of control versus treated tumor volumes. a Fiebig, Freiburg, Germany. b Langdon, Edinburgh, Scotland. c Jacqueline Plowman, National Cancer Institute, Frederick, Maryland. d Josephine Hurst, Goodwin Institute for Cancer Research, Plantation, Florida (unpublished data). e Raffaella Giavazzi, Mario Negri Institute, Milan, Italy. Copyright © 2005 CRC Press, LLC TABLE 12.2 Comparative In Vitro Activity of ET-743 in Human Cancer Cell Lines23 Tumor Type Cell Lines IC50 Potency Malignant fibrous histiocytoma HS-90M/M-8805/M-9005 < 0.1 pM Fibrosarcoma HT-1080 < 0.1 pM Mesenchymal chondrosarcoma HS-16 4 pM–100 pM Liposarcoma HS-18 4 pM–100 pM Hemangiopericytoma HS-30 4 pM–100 pM Mesenchymoma HS-42 4 pM–100 pM Colon HCT-8 10 nM Colon HT-29 3 nM Colon HCT-16 3 nM Breast MCF-7 20 nM Johnson Pharmaceutical Research & Development in anticipation of drug registration in the United States and Europe. Acute toxicity studies in mice (CD1, MF1), rats (WISTAR, Fisher 344), dogs (beagles), and monkeys (cynomolgus) by intravenous (i.v.) administration, either as a single-bolus dose or using an intermittent five times daily (d×5) schedule, provided the safe starting dose for phase I clinical trials as 50 µg/m2 for a single dose and 10 µg/m2 per day for fractionated dose (d×5) schedules. Furthermore, chronic toxicity studies (3 cycles) using several infusion schedules in rats (Sprague–Dawley) and monkeys (Cynomolgus) have been performed as a part of the integrated development plan. ET-743 toxicity, regardless of the animal species tested, is characterized by a common pattern that includes: hepatotoxicity; bone marrow toxicity, both in vitro32,33 and in vivo; and damage at the injection site. Qualitative and quantitative differences of this common pattern can be noted depending on the animal species tested. In rats, gender differences were detectable (although not in other species), with female animals displaying more severe toxicity. Gender differences in rats have also been detected in metabolic behavior.42 Liver toxicity in rats was characterized by proliferative/inflammatory and epithelial changes in the bile ducts. Cholangitis is mainly found in rats, but not in humans. Bone marrow toxicity has also been detected in rats, and this toxicity shows a tendency to recover after 3 weeks. In contrast to rats, monkeys express metabolic profile similarities to humans42 and are thus considered to be a more relevant and predictive animal model of hepatoxicity than are rats, and particularly than female rats. The pattern of hepatotoxicity observed in monkey was also similar to that observed in patients (noncumulative, reversible, predominance of transaminase increase). Bone marrow toxicity was also detected, and it was characterized by myelosuppression (lym- phoid depletion in the spleen and thymus, hypocellularity in the bone marrow), with recovery after 3 weeks. D. CHEMICAL SYNTHESIS Ecteinascidia turbinata has been successfully grown and harvested in aquaculture facilities located along the Mediterranean coast. The purification of the active ingredient is then accomplished on an industrial scale, using chromatographic procedures that represent a more practical and environ- mentally sound practice than harvesting the creature from the wild. Nevertheless, in recent years

several synthetic schemes have been developed for industrial production of ET-743 in the quantities and quality required for a drug product (Yondelis) that will be used in clinical studies worldwide and manufacturing for commercialization. Copyright © 2005 CRC Press, LLC 1. Synthetic Routes to Ecteinascidin 743 (Yondelis) To date, two distinct total synthetic routes to Yondelis (ET-743) have been reported. The seminal work of E. J. Corey and coworkers34 provided for the first time a total synthesis of this complex molecule. This breakthrough scheme resolved one of the main roadblocks to the synthesis — the cyclization to obtain the 10-member lactone bridge through a benzylic sulfide linkage. More recently, Fukuyama et al.35 were able to successfully direct previous work on synthetic routes to the saframycins to converge with the general approach established by Corey for the later stages of the synthesis. The procedures outlined above represent some of the most outstanding work in recent organic chemistry. However, the long and involved procedures for total synthesis of the molecule represent a tremendous barrier to industrial manufacture of the drug, which is particularly challenging in the face of regulatory requirements for pharmaceuticals. This problem was finally solved with the development of a semisynthetic procedure36,37 representing the first industrially feasible route to the manufacture of the drug on a large scale. This semisynthetic approach provides access not only to Yondelis (ET-743) but also to other members of the ecteinascidins and analogous compounds. The procedure uses safracin B, an antibiotic available through fermentation of the bacteria Pseudomonas fluorescens, as the starting point. Optimization of the fermentation process, followed by its transformation according to Scheme 12.1, provided a robust, easily scaled-up procedure for manufacturing the drug. Cyanosafracin B was converted into 2 via a five-step sequence. The amino and phenol groups were protected as the BOC and MOM derivatives, respectively. The hydrolysis of the methoxy-p- quinone and subsequent reduction of the p-quinone afforded the unstable hydroquinone, which was treated with bromochloromethane to give the methylendioxy ring. Alkylation of the remaining phenol gave compound 2. Removal of the BOC and MOM groups from 2 was followed by amide cleavage via an Edman degradation protocol providing 3 in 68% yield. Protection of the phenol allowed for the diazotization of the primary amine for conversion to alcohol 4. Next, esterification with the protected cysteine and protection of the phenol with the MEMCl produces compound 5. The synthesis of ET-743 was completed using the chemistry from Corey et al. on similar substrates. A five-step sequence was used to form 9. Deprotection of the allyl group and oxidation of the phenol, followed by dehydration and cyclization under Swern conditions, gave the intermediate 7. Removal of the MEM and BOC protecting groups was followed by ketone formation. Finally, Pictet-Spengler reaction and carbinolamine formation produces ET-743. E. CLINICAL STUDIES Yondelis (ET-743) is currently in the advanced stages of clinical development. Worldwide, more than 2000 patients have already been treated with this innovative drug. ET-743 is currently being studied in phase II clinical trials for ovarian, STS, breast, endometrial, prostate, NSCLC, and pediatric indications. Yondelis was granted Orphan Drug designation by the European Commission for the indications of STS and ovarian cancer. 1. Yondelis (ET-743) in STS (Soft Tissue Sarcoma) STS is an aggressive type of cancer with unfavorable prognosis and limited treatment options. The current survival rate after 5 years of chemotherapy treatment in metastatic disease is 8%, highlight- ing the urgent need for improved therapeutic options. Results of an exploratory, pooled analysis that assessed the efficacy of Yondelis and the tumor growth rate in patients with advanced STS participating in three pivotal phase II trials have been released.38 These trials used 1500 µg/m2 of Yondelis in a 24-h infusion regimen every 3 weeks. Of the 183 patients with progressive disease evaluated, 7.7% (n = 14) achieved objective response (>50% tumor shrinkage), a further 7.7% (n = 14) achieved minor response (25% to 50% tumor Copyright © 2005 CRC Press, LLC SCHEME 12.1 Semisynthesis of ET-743 from Saframycin. Copyright © 2005 CRC Press, LLC shrinkage), and 36.1% (n = 66) had stabilization of the disease with a median duration of 9 months. Therefore, in total, 51.4% (n = 94) of patients showed clinical benefit in terms of tumor growth control from treatment with Yondelis. The 6-month progression-free survival (PFS) rate, an endpoint that assesses the percentage of patients whose disease has not progressed by a defined time point, was 19.8%, which is favorable according to results of an European Organization for Research and Treatment of Cancer study that considers drugs with a 6-month PFS rate above 14% as efficacious, and those below 8% as inactive.39 Following Yondelis treatment, 47.5% and 29.3% of patients with advanced, metastatic disease were alive after 1 and 2 years, respectively. Median overall survival time was 10.3 months, with 29% of patients surviving at 2 years or beyond. Yondelis proved particularly effective in cases of advanced sarcoma that had relapsed or that were resistant to conventional therapy. Phase I and II trials showed that a shorter treatment schedule (3-h infusions at 1300 µg/m2 every 3 weeks) was feasible. 2. Yondelis in Ovarian Cancer Ovarian cancer is one of the deadliest gynecological cancers. Unfortunately, detection of ovarian cancer is difficult, and the disease is often diagnosed too late for successful treatment. The overall 5-year survival rate of patients with advanced disease (stage III and IV) is approximately 20%, thus, there is an urgent need for improved therapy. In a study of Yondelis administered using a 3- h i.v. infusion regimen in patients with ovarian cancer who had previously been treated with platinum-taxane combination therapy, 10 of 29 patients (34.5%) achieved an objective response (tumor shrinkage), with a recurrence of 6 months or longer after the platinum regimen.40 Those with recurrence before 6 months had a confirmed objective response rate of 6.7% (2 of 30 patients).40 First results of a phase I study of a 3-h i.v. infusion of Yondelis administered in combination with cisplatin every 3 weeks in patients with a variety of cancers revealed no evidence of severe hematological and nonhematological toxicities, and the pharmacokinetic data demonstrated that Yondelis does not influence the plasma disposition of cisplatin, indicating that the combination regime should be feasible for clinical use.41 This study of Yondelis in combination with cisplatin, as well as eight other studies of Yondelis in combination with doxorubicin, liposomal doxorubicin, docetaxel, carboplatin, paclitaxel, capecitabine, gemcitabine and oxaliplatin, are currently ongoing. Yondelis has also been tested in advanced breast cancer in patients failing to respond to standard chemotherapy (anthracyclines and taxanes). This heavily pretreated population has a very poor prognosis, with few responses to available agents. Yondelis has been tested in this setting with a response rate of 13.6% (3 out of 22 patients, 9 patients had stable disease and 8 had a greater than 50% decrease of tumor marker CA 15.3 levels in blood).103 3. Safety Profile In all clinical studies, Yondelis has shown a good safety and tolerability profile. In STS patients treated with Yondelis, no mucositis, alopecia, neurotoxicity, cardiotoxicity, or cumulative toxicities have been observed. The most frequent adverse event appears to be neutropenia, which is reversible. Transaminase elevations were also reported but were transient. III. APLIDIN A. BACKGROUND Aplidin (dehydrodidemnin B) is a potent depsipeptide that was isolated from a Mediterranean tunicate, Aplidium albicans.43 This compound is structurally related to the Didemnins, agents that were previ- ously isolated from the Caribbean marine tunicate Trididemnum solidum.44,45 The structures of the didemnins were reported to contain a 23-membered macrocycle with an attached side chain. The macrocycle is made up of six subunits, (S)-Leu, (S)-Pro, (1S, 2R)-Thr, (S)-N(Me)-O(Me)-Tyr, Copyright © 2005 CRC Press, LLC FIGURE 12.2 (See color insert) Structure and source organism of Aplidin. (3S,4R,5S)-isostatin (Ist), and (2S,4S)-3-oxo-4-hydroxy-2,5-dimethylhexanoic acid, also known as α-(α-hydroxyisovaleryl)propionic acid. The side chain, whose first amino acid is always (R)-N(Me)- Leu, is joined to the Thr of the macrocycle, and its structure differentiates the various didemnins. Aplidin contains pyruvil-L-Pro as the side chain (Figure 12.2). Aplidin (molecular formula: C57H87N7O15; relative molecular mass: 1110.3) has the registered name Aplidin®, the common name dehydrodidemnin B, and the chemical name (2S)-N-{(1R)-1- [({(3S,6R,7S,10R,11S,15S,17S,20S,25aS)-11-hydroxy-20-isobutyl-15-isopropyl-3-(4-methoxyben- zyl)-2,6,17-trimethyl-10-[(1S)-1-methylpropyl]-1,4,8,13,16,18,21-heptaoxodocosahydro-15H-pyr- rolo[2,1-f][1,15,4,7,10,20]dioxatetraazacyclotricosin-7-yl}amino)carbonyl]-3-methylbutyl}-N- methyl-1-pyruvoyl-2-pyrrolidinecarboxamide. Several investigators have reported their results of various biological activities for a wide variety of didemnins.46–48 Among them Aplidin exhibited enhanced cytotoxic activity in vitro against several solid tumor and leukemic cell lines with IC50 values that were about 10-fold more potent than those for didemnin B. This enhanced level of in vitro potency focused attention on further use of the Aplidin congener. The compound was then chosen for development as an antineoplastic agent because of its antitumor activity, and in early 1999, Aplidin entered phase I clinical trials in Europe and Canada. B. MECHANISM OF ACTION The primary mechanism of action of Aplidin is still unknown and is the subject of active investiga- tion.49–62 Aplidin induces a G1 block or delay in the S phase and a G2 blockade in nonsynchronized cells.49 New cells are blocked in G1 or become apoptotic. S phase–synchronized cells treated with Aplidin proceed to G1 and then either die (apoptosis) or stop growing. This effect is clear from 1–100 nM at between 1 and 24 h of exposure. Proportionately stronger apoptosis occurs with higher concen- trations or length of exposure to Aplidin. An NCI COMPARE analysis between Aplidin (NSC:638719) and the Standard Anticancer Agent Database showed moderate (0.7–0.8) but statistically significant correlation coefficients to pancratistatin, bactobolin, Didemnin B, bisantrene HCl, and anguidine. Early studies on the effects of Didemnins identified at least three effects on cellular biochemistry. First, inhibition of protein synthesis by modulating the GTP-dependent elongation factor 1-α.50 Copyright © 2005 CRC Press, LLC Second, inhibition of palmitoyl protein thioesterase, which modulates G-protein-dependent signal transduction pathways for cellular proliferation.51 Aplidin inhibits palmitoyl protein thioesterase in an uncompetitive manner through a selective complex formation with substrate-bound palmitoyl protein thioesterase similar to Didemnins A and B and with similar inhibition IC50 values. Third, Aplidin reduces cellular ornithine decarboxylase activity. However, the effect is not directly on enzyme activity but some degree of inhibition of the level of transcriptional expression of the gene, as observed in the DU145 human prostate carcinoma cell line.52,53 In these experiments, one course of Aplidin significantly (twofold or more) down-regulated the expression of 81 (<1%) genes encoding proteins involved in lipid metabolism, receptors and other membrane proteins, and DCMTase. The latter finding may be relevant because DCMTase is strongly implicated in carcinogenesis, and a link between DCMTase and ODC has been postulated.54 Twelve genes were up-regulated by a single course of Aplidin. Two courses of Aplidin significantly down-regulated 34 genes, including the angiotensin receptor, while transglutaminase, a marker of cellular differentiation, was significantly up-regulated. In a separate microchip array assay using MOLT-4 (p53+) leukemic cells exposed for 1 h to the drug, Aplidin decreased cyclin E levels by 24 h coincident to a G1 block of the cell cycle. There is no effect on the expression of CDK2, CDC2, CDK4, CYCB, CYCD1, p21, or p53.55 Several new mechanistic proposals for Aplidin are emerging to explain its strong cytotoxicity and apoptogenic activity to human leukemic and solid tumor cell lines.57–62 In human T-cell leukemia (MOLT-4) cells, Aplidin inhibits the secretion of vascular endothelial growth factor (VEGF) by an unexplained mechanism in which liberation to the extracellular media rather than VEGF synthesis is blocked58 (Figure 12.3). At the same time, Aplidin downregulates the VEGF receptor (flt-1), thereby blocking an essential autocrine loop for cell proliferation. Ectopic expression or addition to the culture media of VEGF partially palliates cytotoxicity of Aplidin in this system, further supporting the relevance of the VEGF autocrine loop. Aplidin also seems to exert a direct antiangiogenic activity (unpublished data). In vitro, Aplidin inhibited endothelial cell proliferation. In vivo, Aplidin dose-dependently inhibited angiogenesis in the chick embryo chorioallantoic membrane assay (up to 80% inhibition at 10 nM). Moreover, Aplidin was highly active in the process of angiogenesis induced in the chick embryo chorioallantoic membrane by VEGF. These data indicate that, in addition to its direct antitumor activity, Aplidin also affects tumor angiogenesis by preventing the production of angio- genic factors by tumor cells and by directly inhibiting the angiogenic process. In fact, in studies on solid tumor cell types Aplidin was found to significantly reduce VEGF production in the

human ovarian tumor cell line (IGROV-1) and inhibited the intraperitoneal growth of human ovarian carcinoma cells xenografted into nude mice (unpublished data). 60 50 40 30 20 10 0 0 6 24 Time (h) FIGURE 12.3 Vascular endothelial growth factor concentration in the medium of ALL-PO cells treated for 1 h with 20 nM Aplidin and evaluated at 0.6 and 24 h after drug washout. The values express the percentage of VEGF concentration in the medium of treated cells with regard to control cells.57 Copyright © 2005 CRC Press, LLC % control In promyelocytic leukemia (HL-60 and primary) and breast cancer (MCF-7) cells, Aplidin cytotoxicity occurs by a selective apoptotic action along the signal pathway mediated through Fas/CD95 (reversed by use of an antagonistic anti-Fas antibody) and early mitochondrial release of cytochrome c. Overexpressed bcl-2 blocked the onset of these actions by Aplidin,61 and in breast (MDA-MB-231) and renal (A-498 and ACHN) cancer cells, Aplidin induces apoptosis via glu- tathione depletion and sustained activation of the EGFR, Src, JNK, and p38 MAP kinases along the stress response pathway.62,63 Apoptosis was not induced by Aplidin in breast cancer cells (MCF- 7) that do not express caspase-3, known for its proapoptotic action. However, Aplidin did induce apoptosis in cell lines normally expressing caspase-3, and its action was inhibited when caspase- 3 activity was blocked. Identification of new antileukemic agents is essential for improving the survival of acute lymphoblastic leukemia (ALL) patients with high-risk or refractory leukemia. Aplidin shows potent in vitro and in vivo activity against both human hematological and solid tumor cells.46,47,58,59 The antileukemia activity of Aplidin was investigated on a number of continuously growing ALL cell lines, ALL cell lines resistant to standard therapy, and leukemia cells freshly isolated from bone marrow of children with ALL at diagnosis (nine cases) or at the time of relapse (five cases).58 A dose-dependent growth inhibitory effect was observed at nanomolar concentrations in all ALL cell lines (IC50 ranging from 5 to 20 nM). Aplidin acts by triggering apoptosis in the ALL cell lines as assessed by TdT-dUTP flow cytometric analysis. Taken together, these data indicate that Aplidin is a potent antileukemic drug to be evaluated in vitro in refractory and relapsing ALL patients. Analyzed within ALL only, Aplidin did not show a significant cross-resistance to any of eight cytotoxic drugs commonly used in leukemia. In a separate line of study, investigators have been measuring the in vitro sensitivity of childhood leukemias to Aplidin.59 The dose-limiting toxicity of Aplidin in phase I/II trials has been muscular toxicity, with a remarkable lack of severe myelosuppression. Studies of the cytotoxicity of Aplidin on bone marrow and peripheral blood lymphocytes samples of healthy children and on samples from children with different types of leukemia have been performed to assess the potential use of Aplidin as a cytotoxic agent in pediatric leukemia. When the sensitivity to Aplidin of a number of leukemic cells from patients suffering from acute lymphoblastic or myelogenous leukemia was compared with that of normal bone marrow and peripheral blood lymphocytes from healthy children, a significant difference in the level of sensitivity was observed — an indication of the potential for this drug in the clinical treatment of these severe hematological diseases (Figure 12.4). In the context of hematological tumors, Aplidin has also been reported to have potent in vitro activity in multiple myeloma tumor cell lines.60 In these studies, Aplidin is active at pharmacologically Aplidin LC75 data 0.20 iALL rALL 0.15 iAML + + BM + Pb + 0.10 + rAML + + + 0.05 + + 0.00 FIGURE 12.4 Aplidin LC75 values and median in different subgroups of leukemia and normal bone marrow and peripheral blood (PB) samples. 59 Copyright © 2005 CRC Press, LLC Aplidin uM iALL rALL iAML BM Pb rAML relevant concentrations (<10 nM) against most of the cell lines in a multiple myeloma panel that includes cells resistant to conventional or novel agents used in multiple myeloma therapies. Taken together, these separate lines of investigation indicate the great potential relevance to the ongoing clinical studies, particularly to determine the effects of Aplidin on VEGF levels in phase II clinical trials. C. PRECLINICAL DRUG DEVELOPMENT 1. In Vitro and In Vivo Testing The preclinical data including in vitro, in vivo, and toxicology results were important for the selection and design of the phase I and II programs. The selection of tumor types and design of the studies have been directly related to the observation that Aplidin has consistent cytotoxic activity against human leukemia, multiple myeloma, and lymphoma cell lines, as well as selected human solid tumors. In early drug development, Aplidin was shown to be broadly cytotoxic in the range of IC50 concentrations between 0.16 and 0.51 nM to murine L1210 lymphocytic leukemia cells (0.29 nM), murine P388 lymphoid neoplasm (0.16 nM), murine B16 melanoma (0.23 nM), and human HeLa cervix epithelioid carcinoma (0.51 nM). Moreover, there was noted a selectivity toward human A549 NSCLC (0.08 nM) and less selectivity, but adequate potency, to human KB oral epidermoid carcinoma (5.3 nM).46–48,64,65 A further evaluation in the human in vitro screening panel of tumor cell lines at the NCI confirmed that Aplidin was strongly cytotoxic against a broad spectrum of tumor types with selectivity for melanomas, NSCL, prostate, ovarian, and colorectal cancer cell lines, with IC50s in the range of 0.18 to 0.45 nM (NCI data). Corollary studies have extended the profile by evaluating in vitro selectivity against prostate tumors showing species-independent specificity for this tumor type: rat MATLyLu (3.40 nM, IC50), human LNCaP (0.32 nM, IC50), PC- 3 (27.0 nM, IC50), and DU 145 (10.23 nM, IC50).66 Aplidin demonstrated significant schedule-dependent antitumor activity in a modified in vitro human tumor colony-forming assay, with potent activity noted against gastric carcinoma and lymphomas.67,68 These studies indicate that Aplidin is a phase-specific agent, a conclusion subse- quently supported by data obtained with a human LoVo colon carcinoma cell line showing a cell cycle block at G1, G2/M.69 Moreover, cell cycle perturbation studies on LoVo (0.1 pM, IC50) and the adriamycin-resistant MDR overexpressing subline LoVo/DX (1 µM, IC50) indicate that Aplidin is a Pgp substrate affected by the classical multidrug-resistant phenotype. Aplidin has been tested in several tumors of murine origin, transplantable human tumor xenografts, and human cancer cell lines in hollow fibers.45–47,64,65,70,71 Activities were seen against leukemias and solid tumors, with dose–response relationships evident in most models. The optimal modes of admin- istration are i.v. infusion and the intraperitoneal route. From the earliest antitumor studies, significant increases in the life span were observed in mice implanted intraperitoneally with murine P388 leukemia and B16 melanoma, followed by intraperitoneal administration of the drug on days 1–9. Significant antitumor activity was also seen using the same dosing regimen against subcutaneously implanted Lewis lung tumors, with complete remissions recorded at the highest dose. The full spectrum of in vivo antitumor activity from the early studies of Aplidin is summarized below (Table 12.3). During this time it was found that i.v. bolus activity was not a route to observe optimal antitumor activity (not published). However, follow-up studies demonstrated that i.v. infusions could restore the activity of the drug and, hence, form the basis for proceeding to clinical trials. 2. Preclinical Toxicology Preliminary toxicology studies were carried out in mice (CD1) and rats (WISTAR). The most significant observations from acute i.v. toxicity studies of Aplidin in mice and rats are bone marrow and reticulo- endothelial system toxicities when given either as a single injection on an intermittent (d×5) schedule. In both species, local injection site effects were also seen (tail lesions and discoloration). To varying Copyright © 2005 CRC Press, LLC TABLE 12.3 In vivo Antitumor Activity against Murine and Human Tumors Opt. Antitumor Tumor Type Regimen Dose Effect Score Leukemia Intraperitoneal P388 Intraperitoneally, qd×5 MTDa 88% ILS ++++ Melanoma Intraperitoneal B16 Intraperitoneally, qd×9 MTDa 98% ILS ++++ Ovary Intraperitoneal M5076 Intraperitoneally, q4d×5 MTDa 48% ILS +++ NSC Lung Subcutaneous Lewis Intraperitoneally, qd×9 MTDa 0% T/C ++++ Breast Subcutaneous MX-1 Intraperitoneally, q4d×3 MTDb 37% T/C + Prostate Subcutaneous PC-3 Intraperitoneally, q4d×3 MTDa 34% T/C + 1/2 MTDa 38% T/C + Intraperitoneally, qd×9 MTDc 25% T/C ++ 1/2 MTDc 30% T/C ++ Burkitts Subcutaneous P3HR1 Intraperitoneally, q4d×3 MTDc 32% T/C + 1/2 MTDc 48% T/C +/- Subcutaneously, 7days MTDd 36% T/C + Gastric Subcutaneous MRIH254 Intraperitoneally, Q d×9 MTDb 19% T/C +++ 1/2 MTDb 17% T/C +++ Intraperitoneally, Q4 d×3 MTDc 18% T/C +++ 1/2 MTDc 38% T/C + a MTD = 2.10 mg/kg. b MTD = 0.80 mg/kg. c MTD = 1.25 mg/kg (drug product). d 1.80 mg/kg, and ILS = increases in the life span. degrees by species and schedule, there was evidence of leukopenia and gastrointestinal tract inflam- mation and transient signs of hepatotoxicity. The depletion of circulating lymphocytes is more persistent in rats. Effects on the stomach were seen in both species — crypt cell necrosis in the rat and hyperplasia of the squamous epithelium in mice. In all cases, recovery was apparent toward the end of the studies. A summary table of the early preclinical MTD levels is given below (Table 12.4). The MTD for formulated product in mice is 3300–3750 µg/m2 and 3420–4200 µg/m2 in rats. The in vitro toxicity of Aplidin on cultured human bone marrow hematopoietic progenitors was evaluated comparatively to doxorubicin on each of three pre–stem cell lineages: myeloid, erythroid, and megakaryocytic cells.72 Regardless of the origin of the hematopoietic progenitors, the toxicity of Aplidin was lower than that observed in tumor cell lines (IC50: 0.0002–0.027). Significantly, plasma concentrations in humans receiving maximum tolerated doses in phase I clinical trials were below the IC50 for hematopoietic progenitors (maximum 0.0034 µM), and no myelotoxicity has been observed to date, indicating that Aplidin is less toxic than doxorubicin in all hematopoietic progenitors studies. TABLE 12.4 MTD Determinations in Mice and Rats for Aplidin MTD Weight MTD Area Safe Clinical Starting Regimen Species (µg/kg) (µg/m2) Dose (µg/m2)a Intravenous, d×1 Mouse 1250 3750 342–375 Rat 570 3420 350 a 1/10 of mean MTDs (µg/m2). Copyright © 2005 CRC Press, LLC D. CHEMICAL SYNTHESIS Aplidin belongs to the didemnin family,73 and as a result of the biological activity of these compounds, several research laboratories have made them synthetic targets. Rinehart et al. reported the first total synthesis of the didemnins in 1987. Subsequently, different research groups published several syntheses, differing mainly in the selection of the two amino acids to achieve macrocy- clization (Figure 12.5). To date, the highest-yielding cyclization was reported by Schmidt et al.74–76 and by Jou et al.77 The first group synthesized didemnin A, B, and C, and Jou et al. described a process to obtain Didemnin A and Aplidin (dehydrodidemnin B); the uniqueness of this synthesis is the use of new coupling reagents based on uranium and phosphonium salts. With this process, it is now possible to prepare a large number of analogs. The synthesis published by Jou et al. involved the synthesis of the Boc-protected didemnin macrocycle, followed by its coupling with Boc-(R)-N(Me)-Leu-OH to give Boc-didemnin A, which was coupled with Pyr-Pro-OH to afford Aplidin (Scheme 12.2). E. CLINICAL STUDIES Phase I trials with Aplidin focused on the evaluation of its safety and pharmacokinetic profile and on the optimization of the infusion schedule.78–82 Results of phase I clinical trials presented at the 2001 European Conference on Clinical Oncology showed promising antitumor activity of Aplidin against pretreated colorectal and renal carcinoma, melanoma, and different neuroendocrine tumors (specifically in medullary thyroid carcinoma and carcinoid), using doses below the MTD.83 Updated results on a pooled analysis on medullary thyroid carcinoma have been presented.84 In further phase I studies, Aplidin has also demonstrated activity against bronchial carcinoid and NSCLC and has been shown to cause tumor shrinkage and improve symptoms. From these phase I studies, an i.v. infusion of 5 mg/m2 Aplidin every 2 weeks has emerged as the preferred dosage schedule, and this schedule is being used in most of the ongoing phase II studies.85 Aplidin is currently in phase II trials for melanoma, pancreatic cancer, head and neck tumors, non-Hodgkin lymphoma, NSCLC, and SCLC. Phase II trials in gastric, bladder, and prostate tumors, Shiori et al. BOP-CI: 49% Jou et al. HATU, HOAt: 76% Schmidt et al. NMe-L-Tyr6 PFP ester: 75% L-Pro5 O O L-Leu4 N O Joullié et al. N

L-Thr1 DPPa: 42% O O FDPP: 60% O NH O O O Hip3 O O NH N H R NH O NMe-D-Leu7 Rhinehart et al. Ist2 (isostatine) EDC, HOBt: 18% Jouin et al. BOP: 54% R= didemnin side chains FIGURE 12.5 Amino acids to achieve the cyclization of Aplidin. Copyright © 2005 CRC Press, LLC O OSi OBn H O N HO N O N O O O O O 1. HBTU/HOBt, DIEA, HN Boc-lst(TBDMS)-OH NH + 3 , CI– THF, 65% O Boc-Leu-Pro-OBzl 2. TBAF, THF, 95% 3. DCC, CH2CI2, CMAP O TBDMSO 4. HCI/ Dioxane HO OH O O Hip-Leu-Pro-OBzI TBDMS-Hip-OH N OBn O N O O O HN O O N O 5. HBTU, HOBt, DIEA O H O O CH2CI2/DMF, 69% O OH O O OH N Boc-Thr(Z-N(Me)-O(Me)-Tyr)-OH O NH2.HCI NH O O O O Boc-Ist(TBDMS)-Hip-Leu-Pro-BzI HN O O O Boc-Thr(Z-N(Me)-O(Me)-Tyr)-Ist-Hip-Leu-Pro-BzI 6. H2, Pd(OH)2-C, THF, 100% HATU, HOAt, NMM, CH3CN, 68% 7. HCI dioxane HATU, HOAt, NMM, CH2CI2, DMF Z-N(Me)-D-Leu-OH O O O O N N N O HO N O O N H O O O N O O O NH O O O O O O O OH NH N Z O O 8. H2, Pd(OH), THF O OH NH N N O NH 9. DIPCDI, CH2CI2, 93% O O NH O O Z-didemnin A Aplidin SCHEME 12.2 Synthesis of Aplidin. multiple myeloma, and acute lymphoblastic leukemia are ongoing. In 2003, based on preclinical evidence and the urgent need for new and effective treatments, the European Commission granted Aplidin Orphan Drug designation for acute lymphoblastic leukemia. This type of leukemia is the leading cause of death from cancer under the age of 35 years. A phase I trial in pediatric solid tumors and acute leukemias is currently ongoing. To date, more than 325 patients have been treated with Aplidin in nine clinical studies across seven trial centers in Europe and Canada, and it has demonstrated a good safety profile and broad- spectrum antitumor activity (PharmaMar data on file). IV. KAHALALIDE F A. BACKGROUND Kahalalide F is a partially cyclic depsipeptide with the unusual feature of having a short chain fatty acid amide at its amino terminal residue (Figure 12.6).86,87 A number of other natural Kahalalides have also been described.88 This compound, together with a number of other natural Kahalalides, was Copyright © 2005 CRC Press, LLC FIGURE 12.6 (See color insert) Structure and source organism of Kahalalide F. discovered in the mollusc Elysia rufescens (a sea slug; a marine gastropod of the ophistobranchia subclass) collected at Kahala (Muanalua Bay) in O’ahu (Hawai’i). Several species of animals of this class are known to acquire, process, and accumulate chemicals produced by the algae on which they feed. Surprisingly, some species are able to retain chloroplasts from these algae that remain active in photosynthesis within the animal for prolonged periods of time.89 In the case of Kahalalide F, the compound is found in the algae (Bryopsis pennata) on which the Elysia molluscs feed, albeit in much reduced concentration.90 In this natural setting a biological role has been proposed for Kahalalides as a deterrent to the feeding behavior of predators of the Elysia nudibranchs. Kahalalide F (molecular formula: C75H124N14O16; relative molecular mass: 1477.8) has the chemical name, (2R)-N-{(1S)4-amino-1-[({(1R,2R)-1-[({(3S,6Z,9S,12R,15R,18R,19S)-9-benzyl-6 eth- ylidene-3,12-diisopropyl-19-methyl-15-[(1R)-1-methylpropyl]-2,5,8,11,14,17-hexaoxo-1-oxa- 4,7,10,13,16-pentaazacyclononadecan-18-yl}amino)carbonyl]-2-methylbutyl}amino) carbonyl] butyl}- 1-{(2R,5S,8S,11R)-8-[(1R)-1-hydroxyethyl]-2,5,11-triisopropyl-17-methyl-4,7,10,13-tetraoxo- 3,6,9,12-tetraazaoctadec-1-anoyl}-2-pyrrolidinecarboxamide. B. MECHANISM OF ACTION The primary mechanism of action of Kahalalide F has not been identified; however, early experi- mental results have provided some insight into the physiological events that correlate with tumor cell killing by this compound. No significant correlations to any other standard chemotherapeutic agents were discovered when the actions of Kahalalide F on the NCI tumor cell line panel were analyzed through the COMPARE algorithm, indicating that the compound may possess a unique way to achieve its biological effectiveness. In these in vitro studies, selectivity was noted to colon, central nervous system, melanoma, prostate, and breast tumor cell lines, where PC-3 and DU-145 prostate tumor cells are the most sensitive at LC50 concentrations below 1 µM (NCI data). Neither protein nor nucleic acid synthesis are found to be inhibited in cultured cells by sublethal concentrations of Kahalalide F. Topoisomerase enzymatic activities (I or II) are also not affected, and no damage to DNA has been specifically correlated with exposure to Kahalalide F. A cell cycle block in G0/G1 has been identified in a variety of tumor cell lines that include prostate (DU145), cervical (HeLa), colon (HT29), head and neck (HN30), and NSCL (HOP62), all with IC50s in the 1-µM range.91 Kahalalide F is strongly cytotoxic to both wild-type p53 and mutated p53 tumor cells in the NCI panel. A number of cell lines overexpressing MDR (PC-3 prostate, CACO-2 colon, UO-31 renal, and MCF7 breast), as well as cell lines resistant to topoisomerase II inhibitors, are sensitive to Kahalalide F, indicating that it acts independent of the respective resistance mechanisms. Cultured cells exposed to biologically relevant concentrations of Kahalalide F detach from their substrate and become markedly swollen, which is associated with the formation of large intracellular vacuoles.92 Copyright © 2005 CRC Press, LLC FIGURE 12.7 (See color insert) Effect on the microtubule network and membrane- bound organelles.92 Within minutes, these engorged vesicles move from the periphery of the cell to a perinuclear location, as observed by confocal laser scanning microscopy. Confocal fluorescence microscopy studies have identified major and immediate effects on lysosomes (Figure 12.7). Moreover, there is an increase in lysosomal pH. The morphologies of the endoplasmic reticulum and the Golgi apparatus appear to be unaffected by the action of Kahalalide F. The effect on cell adhesion has not been explained by changes in organization of cytoskeletal structures, and in particular the microtubule network, which appear intact. It has been suggested that the subcellular effect might be explained if Kahalalide F is inserted as an ionophore in membranes favoring an increase in cation permeability, thus causing a passive water influx and resulting in cisternal dilation. The hydrophobic nature of the compound would not be incompatible with a model in which membrane-associated events trigger cell death. These effects would be similar to those of compounds like the carboxylic ionophore, monensin.93 The compound has also been suggested to block the EGF receptor and inhibit TGF-β gene expression.94 In vitro studies have shown that Kahalalide F is selectively cytotoxic to neu+ cells overexpressing Her2, indicating that it may interfere with erbB2 transmembrane tyrosine kinase activity. However, it does not inhibit autophosphorylation of the receptors or MEK kinase activity (unpublished data). More recently, Kahalalide F has been shown to cause rapid and potent cytotoxic effects in ErbB2 (HER2/neu) overexpressing breast cancer cell lines (SKBR3, BT474), which was associated with the induction of a hypodiploid cell population, dramatic cell swelling, and perme- abilization of the plasma and lysosomal membranes; however, overexpression of HER2 does not affect the sensitivity of MCF7 breast cancer cells to Kahalalide F.95 Several markers of caspase- dependent apoptosis were negative after Kahalalide F exposure, including the externalization of phosphatidyl serine, release of cytochrome c out of mitochondria, and cleavage of caspase-3 and PARP. Moreover, molecular or chemical inhibition of caspases by ectopic overexpression of Bcl-2 or a pan-caspase inhibitor (zVAD-fmk), respectively, failed to protect against Kahalalide F cyto- toxicity. Specific inhibitors of cathepsin B (CA-074 Me, zFA-fmk) or D (pepstatin A) also failed to protect against Kahalalide F–induced cell death. Together, these results indicate that Kahalalide F–induced cytotoxicity is predominantly a result of a process of a necrotic cell death involving oncosis rather than apoptosis. This effect has also been reported in hepatoma cell lines.96 The sensitivity to Kahalalide F in a panel of human tumor cell lines derived from breast (SKBR3, BT474, MCF7), vulval (A431), NSCLC (H460, A549, SW1573, H292), and hepatic carcinoma Copyright © 2005 CRC Press, LLC (SKHep1, HepG2, Hep3B), significantly correlated with protein expression levels of ErbB3 (HER3), but not other ErbB receptors. Downregulation of ErbB3 expression was observed in Kahalalide F–sensitive cell lines within 4 h exposure to Kahalalide F, as well as inhibition of the PI3K-Akt/PKB signaling pathway, which is directly linked to ErbB3.97 Moreover, ectopic expression of a constitutively active Akt mutant had a protective effect against Kahalalide F cytotoxicity. This suggests that ErbB3 and the Akt pathway are major determinants of Kahalalide F action on these cell lines. C. DRUG DEVELOPMENT 1. In Vitro and In Vivo Testing Early preclinical data identified Kahalalide F as a potent new chemical entity showing significant cytotoxic activity below 10 µM (IC50) against solid tumor cell lines. Further evaluation demonstrated that this activity is selective for, but not restricted to, prostate tumor cells. Subsequent studies have identified tumor cells that overexpress the Her2/neu and Her3/neu oncogenes as potentially sensitive targets for Kahalalide F. Moreover, Kahalalide F cytotoxicity is not schedule dependent (not published), and it is not a strong MDR substrate, as it is effective against many multidrug-resistant tumor cell lines (NCI data). Preliminary in vitro screening studies identified micromolar activity of Kahalalide F against murine leukemia (P388) and two human solid tumors: NSCL (A549) and colon (HT29). Gastric (HS746T, 0.01 µM) and prostatic (PC-3, 0.08 µM) tumors were shown to be very sensitive as well. In vitro studies in cell lines of human origin evaluated by the NCI confirmed these results and identified selective activity against colon, NSCL, CNS, melanoma, prostate, and breast cancer cells, with potencies ranging from 200 nM (prostate) to 10 µM (leukemia). Extended in vitro selectivity studies reveals that Kahalalide F is active against neu+ (Her2 overexpressing) human breast tumor cells95 and some primary sarcoma lines, but not against hormone-sensitive LNCAP tumor cells (Table 12.5). In vivo antitumor activity was observed against breast, colon, prostate, and NSCL human tumor cells xenografted into athymic mice. Interestingly, DU-145 (hormone refractory prostate) tumors, resistant to most agents, responded to Kahalalide F at the MTD and 1/2 MTD levels. Similar studies were completed with PC-3 xenografted tumors, producing the same effect of tumor growth inhi- bition (between 31% and 48% T/C; unpublished data). 2. Preclinical Toxicology The pharmacokinetic behavior of Kahalalide F was characterized in mice and confirmed in rats, in conjunction with in vitro and in vivo antitumor activity studies, to discern the pattern of systemic exposure to the drug associated with efficacious dosing regimens.98–100 TABLE 12.5 In Vitro Cytotoxic Activity of Kahalalide F against Solid Tumors Tumor Line IC50 (Molar) Chondrosarcoma CHSA 1.58 µM Osteosarcoma OSA-FH 1.65 µM Prostate PC-3 1.02 µM Prostate DU-145 1.78 µM Prostate LNCAP Not active Breast SK-BR-3 2.50 µM Breast BT-474 2.00 µM Copyright © 2005 CRC Press, LLC A dose of 278 µg/kg given as a rapid i.v. injection afforded an initial plasma concentration of 1.55 µM. The plasma concentration–time profile was distinctly biexponential, with half-lives of 15.8 min and 4.4 h for the initial and terminal disposition phases, respectively. The apparent volumes of distribution of the drug were very large, more than 100 times the body weight, indicating that the compound distributes extensively into peripheral tissues. The total body clearance, 14.5 mL/min per kilogram, was only 23% of hepatic blood flow. The Cmax of the MTD i.v. dose is comparable to the in vitro IC50 values of the most sensitive human tumor cell lines. No drug was found in mouse plasma 24 h after i.v. injection. There was no accumulation on repeated i.v. injection at an interval of 24 h. Surprisingly, although a slightly greater than MTD dose is too toxic when given as an i.v. single bolus, daily serial i.v. injections of the MTD dose are not toxic. These preliminary data indicate that Kahalalide F is rapidly eliminated from plasma with limited binding to extravascular tissues. This suggests tolerance to repeat dosing without acute cumulative adverse effects, which would be favorable to clinical development of the drug when confirmed in the human setting. Conditions in which a number of tumor cell lines were exposed in vitro to Kahalalide F for 24 h exhibited IC50 values ranging from 0.01 to 2.51 µM. Bone marrow hemopoietic progenitors show no significant inhibition by Kahalalide F after 24 h of exposure at this range of concentrations; in fact, myeloid and erythroid bone marrow hemopoietic progenitors show no significant inhibition by Kahalalide F after 24 h of exposure at higher (10 µM) concentrations.99 The results indicate sensitivity of tumor cells to doses of Kahalalide F that are not toxic

to normal hematopoietic cells, indicating that neutropenias and thrombocytopenias as a consequence of Kahalalide F treatment would not be expected. A dose of 250 µg/kg of a single bolus injection was established to be the maximal tolerated dose in rodents, although there are gender differences. The dose-limiting toxicity is predominantly renal toxicity, reversible by the day 29 necropsy, with minimal histologic evidence of nephrotoxicity at the 1/2 MTD. Signs of liver function alteration were seen on day 4 at the MTD dose, with recovery by day 29. If the maximum tolerated dose is exceeded there are signs of neurotoxicity, responsible for the lethalities. In intermittent dose studies in beagle dogs, slight nonregenerative anemia was seen in males and females at all dose levels. The effect was typically most severe on day 8, with general recovery by days 15 and 22. No other hematological changes were observed. No correlative histologic changes, for example, in the bone marrow, spleen, or liver, were seen. No other indices of toxicity were seen, and drug-related clinical signs or histologic lesions were not apparent. D. SYNTHESIS OF KAHALALIDE F Only a single synthetic approach has been published to date to produce Kahalalide F.87 This procedure makes use of solid-phase chemistry to prepare a linear chain using the Fmoc/tBu strategy on 2-chlorotrityl chloride-resin, allowing the cleavage of the peptide under mild acid conditions. The amino acids, D-allo-Thr, and the Thr precursor of the Z-Dhb, were both introduced without protection of the hydroxyl function. HATU/DIEA was used to form the amide functional groups. With the linear peptide in hand, cyclization in solution phase and final deprotection allows the preparation of the natural compound in a straightforward way. Before detaching the peptide from the resin, the alloc group is removed under standard conditions. The cyclization reaction is per- formed with PyBOP/DIEA, using DMF as solvent, and finally the deprotection of the Boc group yields the natural compound. This solid-phase methodology is easily scaled up and applicable to the generation of a wide variety of new analogs (Scheme 12.3). E. CLINICAL TRIALS Results from a dose-finding phase I study in patients with advanced androgen-resistant prostate cancer were presented at the 2002 annual meeting of the American Society for Clinical Copyright © 2005 CRC Press, LLC SCHEME 12.3 Synthesis of Kahalalide F. Oncology.101 In this study, Kahalalide F was found to have a rapid plasma clearance, which could be beneficial for serial dosing and could further enhance its efficacy in patients. In addition, the dose could be escalated up to 930 µg/m2 per day. Kahalalide F also demonstrated a very favorable safety profile, and treatment-related side effects were noncumulative and rapidly reversible. Data from a phase I study in patients with advanced solid tumors that had failed to respond to previous chemotherapies were presented at the 2002 joint meeting of the European Organization for Research and Treatment of Cancer, the National Cancer Institute, and the American Association for Cancer Research.102 In this study, Kahalalide F was administered as weekly 1-h i.v. infusions, and the dose could be escalated up to 1200 µg/m2 per week. Signs of activity in a variety of cancers were observed at 400–1200 µg/m2 per week. The data support the favorable safety profile of Kahalalide F, with no bone marrow or renal toxicities, mucositis, alopecia, or cumulative toxicity reported. Copyright © 2005 CRC Press, LLC REFERENCES 1. Rinehart, K.L., et al., Ecteinascidins 729, 743, 745, 759A, 759B, and 770: potent antitumor agents from the Caribbean tunicate Ecteinascidia turbinata [Erratum to document cited in CA113(9):75189d], J. Org. Chem., 55, 4512–4515, 1990. 2. He, H., Faulkner, D.J., Renieramycins E and F from the sponge Reniera sp.: reassignment of the stereochemistry of the renieramycins, J. Org. Chem., 54, 5822–5824, 1989. 3. Erba, E., et al., Mode of action of Ecteinascidin 743, a natural marine compound with antitumor activity, Ann. Oncol., 9, 535, 1998. 4. Erba, E., et al., Ecteinascidin 743, a natural marine compound with a unique mechanism of action, Eur. J. Cancer, 37, 97–105, 2001. 5. Moore, B.M., et al., Mechanism for the catalytic activation of ecteinascidin 743 and its subsequent alkylation of guanine N2, J. Am. Chem. Soc., 120, 2490–2491, 1998. 6. Seaman, F.C., Hurley, L.H., Molecular basis for the DNA sequence selectivity of Ecteinascidin 736 and 743: evidence for the dominant role of direct readout via hydrogen bonding, J. Am. Chem. Soc., 120, 13028–13041, 1998. 7. Hurley, L.H., DNA and its associated processes as targets for cancer therapy, Nat. Rev. Cancer, 2, 188–200, 2002. 8. Zewail-Foote, M., Hurley, L.H., Differential rates of reversibility of Ecteinascidin 743-DNA covalent adducts from different sequences lead to migration to favoured bonding sites, J. Am. Chem. Soc., 123, 6485–6495, 2001. 9. Sheng, K., et al., Ecteinascidin 743, a transcription-targeted chemotherapeutic that inhibits MDR1 activation, Proc. N.Y. Acad. Sci., 97, 6775–6779, 2000. 10. Minuzzo, M., et al., Interference of transcriptional activation by the antineoplastic drug Ecteinascidin 743, Proc. N.Y. Acad. Sci., 97, 6780–6784, 2000. 11. Jin, S., Hu, Z., Scotto, K., The antitumor agent Ecteinascidin 743 inhibits transcriptional activation of the MDR1 gene by multiple inducers, Clin. Cancer Res., 5, 302, 1999. 12. Mantovani, R., et al., Effect on ET-743 on the interaction between transcription factors and DNA, Ann. Oncol., 9, 534, 1998. 13. Synold, T.W., Dussault, I., Forman, B.M., The orphan nuclear receptor SXR coordinately regulates drug metabolism and efflux, Nat. Med., 7, 584–590, 2001 14. Takebayashi, Y., et al., Antiproliferative activity of ecteinascidin 743 is dependent upon transcription- coupled nucleotide-excision repair, Nat. Med., 7, 961–966, 2001. 15. Zewail-Foote, M., et al., The inefficiency of incisions of Ecteinascidin 743-DNA adducts by the UvrABC nuclease and the unique structural feature of the DNA adducts can be used to explain the repair-dependent toxicity of this antitumor agent. Chem. Biol., 8, 1033–1049, 2001. 16. Bonfanti, M., et al., Effect of Ecteinascidin 743 on the interaction between DNA binding proteins and DNA, Anticancer Drug Des., 14, 179–186, 1999. 17. Friedman, D., et al., Ecteinascidin 743 inhibits activated by not constitutive transcription. Cancer Res., 62, 3377–3381, 2002. 18. Izbicka, E., et al., Incomplete cross-resistance between ET-743 and standard chemotherapeutic agents against primary tumors in human tumor cloning assay, Ann. Oncol., 9, 130, 1998. 19. Valoti, G., et al., Ecteinascidin-743, a new marine natural product with potent antitumor activity on human ovarian carcinoma xenografts, Clin. Cancer Res., 4, 1977–1983, 1998. 20. Hendriks, H.R., et al., High antitumor activity of ET-743 against human tumor xenografts from melanoma, non-small cell lung and ovarian cancer, Ann. Oncol., 10, 1233–1240, 1999. 21. Riccardi, R., et al., Effective combinations of ET-743 and doxorubicin for tumor growth inhibitions against murine and human sarcomas in athymic mice, Proc. Am. Assoc. Cancer Res., 42, 1132, 2001. 22. D’Incalci, M., et al., The combination of Yondelis and cisplatin is synergistic against human tumor xenografts, Eur. J. Cancer, 39, 1920–1926, 2003. 23. Li, W.W., et al., Sensitivity of soft tissue sarcoma cell lines to chemotherapeutic agents: identification of ecteinascidin 743 as a potent cytotoxic agent, Clin. Cancer Res., 7, 2908–2911, 2001. 24. Takahashi, N., et al., Sequence-dependent enhancement of cytotoxicity produced by ecteinascidin 743 with doxorubicin or paclitaxel in soft tissue sarcoma cells, Clin. Cancer Res., 7, 3251–3257, 2001. Copyright © 2005 CRC Press, LLC 25. Faircloth, G.T., In vivo study using tumor containing hollow fibers in the nude mouse, PharmaMar USA report no. 02E01HF, January 15, 2001. 26. Hornicek, F., Weissbach, L., Invasiveness of chondrosarcoma cells in vitro is altered by ET743, Proc. Am. Assoc. Cancer Res., 43, 539, 2002. 27. Faircloth, G.T., Grant, W., Hornicek, F., In vivo combinations of chemotherapeutic agents with ET743 against solid tumors, Clin. Cancer Res., 7, 387, 2001. 28. Faircloth, G.T., et al., Dexamethasone potentiates the activity of Ecteinascidin 743 in preclinical melanoma and osteosarcoma models, Proc. Am. Assoc. Cancer Res., 43, 73, #379, 2002. 29. Morioka, H., et al., Antiangiogenesis treatment combined with chemotherapy produces chondrosar- coma necrosis, Clin. Cancer Res., 9, 1211–1217, 2003. 30. Morioka, H., et al., Antiangiogenesis treatment combined with chemotherapy produces chondrosar- coma necrosis, Clin. Cancer Res., 8, 1211–1217, 2003. 31. Faircloth, G.T., Grant, W., Jimeno, J., Enhancing the preclinical in vivo antitumor activity of Ectein- ascidin 743, a marine natural product currently in phase II clinical trials, Clin. Cancer Res., 5, 306, 1999. 32. Ghielmini, M., et al., In vitro schedule dependancy of myelotoxicity and cytotoxicity of Ecteinascidin 743, Ann. Oncol., 9, 989–993, 1998. 33. Albella, B., et al., In vitro toxicity of ET743 and Apldine, two marine-derived antineoplastics, on human bone marrow haematopoietic progenitors: comparison with the clinical results, Eur. J. Cancer, 38, 1395–1404, 2002. 34. Corey, E.J., Gin, D.Y., Kania, R.S., Enantioselective total synthesis of Ecteinascidin 743, J. Am. Chem. Soc., 118, 9202, 1996. 35. Endo, A., et al., Total synthesis of Ecteinascidin 743, J. Am. Chem. Soc., 124, 6552, 2002. 36. Cuevas, C., et al., Synthesis of Ecteinascidin ET-743 and Phthalascidin Pt-650 from Cyanosafracin B, Org. Lett., 2, 2545, 2000. 37. Menchaca, R., et al., Synthesis of natural Ecteinascidins (ET-729, ET-745, ET-759B, ET-736, ET-637, ET-594) from Cyanosafracin B. J. Org. Chem., 68, 8859, 2003. 38. Lopez-Martin, A.J., et al., An exploratory analysis of tumor growth rate (TGR) variations induced by trabectedin (ecteinascidin-743, ET-743) in patients (pts) with pretreated advanced soft tissue sarcoma (PASTS), Proc. Am. Soc. Clin. Oncol., 22, 819, abst. 3293, 2003. 39. Van Glabbeke, M., et al., On behalf of the EORTC Soft Tissue and Bone Sarcoma Group. Progression- free rate as the principal end-point for phase II trials in soft-tissue sarcomas, Eur. J. Cancer, 38, 543–549, 2002. 40. Colombo, N., et al., Phase II and pharmacokinetics study of 3-hr infusion ET-743 in ovarian cancer patients failing platinum-taxanes, Proc. Am. Soc. Clin. Oncol., 21, 221a, abst. 880, 2002. 41. Grasselli, G., et al., Phase I and pharmacokinetic (PK) study of ecteinascidin-743 (ET, Trabectedin) and cisplatin (P) combination in pre-treated patients (pts) with selected advanced solid tumors, Proc. Am. Soc. Clin. Oncol., 22, 135, abst. 542, 2003. 42. Vermier, M., et al., An Interspecies Comparison of the Metabolism of the Anticancer Agent YondelisTM (Trabectedin, ET-743). Cytochrome P450, Biochemistry, Biophysics and Drug Metabolism. 13th Inter- national Conference on Cytochromes P450, Prague, Czech Republic, June 29–July 3, 2003. 43. Rinehart, K.L., Lithgow-Berelloni, A.M., Novel Antiviral and Cytotoxic Agent, PCT Int. Pat. Appl. WO 91.04985, Apr. 18, 1991; GB Appl. 89/22,026, Sept. 29, 1989; Chem. Abstr., 1991, 115, 248086q. 44. Rinehart, K.L., et al., Structures of the didemnins, antiviral and cytotoxic depsipeptides from a caribbean tunicate, J. Am. Chem. Soc., 103, 1857–1859, 1981. 45. Rinehart, K.L., et al., Didemnins and tunichlorin: novel natural products from the marine tunicate Trididemnum solidum, J. Nat. Prod., 51, 1–21, 1988. 46. Urdiales, J.L., et al., Antiproliferative effect of dehydrodidemnin B (DDB), a depsipeptide isolated from Mediterranean tunicates, Cancer Lett., 102, 31–37, 1996. 47. Faircloth, G.T., et al., Dehydrodidemnin B, a new marine derived antitumor agent with activity against experimental tumor models, Ann. Oncol., 7, 34, 1996. 48. Sakai, R., et al., Structure activity relationships of didemnins, J. Med. Chem., 39, 2819–2834, 1996. 49. Erba, E., et al., Mechanism of antileukemic activity of Aplidin, Proc. Amer. Assoc. Cancer Res., 40, 3, 1999. Copyright © 2005 CRC Press, LLC 50. Crews, C.M., et al., GTP-dependent binding of the antiproliferative agent didemnin to elongation factor 1-alpha, J. Biol. Chem., 269, 15411–15414, 1994. 51. Crews, C.M., Lane, W.S., Schreiber, S.L., Didemnin binds to the palmitoyl protein thioesterase responsible for infantile neuronal ceroid lipofuscinosis, Proc. N.Y. Acad. Sci., 93, 4316–4319, 1996. 52. Erba, E., et al., Is Aplidin acting as an ornithine decarboxylase (ODC) inhibitor? AACR-EORTC-NCI Conference, 312, 1999. 53. Izbicka, E., et al., Evaluation of molecular targets for Aplidin, a novel anticancer agent, 11th NCI- EORTC-AACR Symposium on New Drugs in Cancer Therapy, 213, 2000. 54. Duranton, B., et al., Concomitant changes in polyamine pools and DNA methylation during growth inhibition of human colonic cancer cells, Exp. Cell Res., 243, 319–325, 1998. 55. Erba, E., et al., Mechanism of antileukemic activity of Aplidin, Proc. Amer. Assoc. Cancer Res., 40, 3, 1999. 56. Broggini, M., et al., Changes in gene expression in tumor cells exposed to the two marine compounds ET-743 and Aplidin by using cDNA microarrays, 10th NCI-EORTC-AACR Symposium on New Drugs in Cancer Therapy, 310, 1998. 57. Broggini, M., et al., Aplidin-induced apoptosis in Molt-4 cells is mediated by its ability to block VEGF secretion, 12th NCI-EORTC-AACR Symposium on New

Drugs in Cancer Therapy, 384, 2002. 58. Erba, E., et al., Effect of Aplidin in acute lymphoblastic leukaemia cells, Br. J. Cancer, 89, 763–773, 2003. 59. Bresters, D., et al., In vitro cytotoxicity of Aplidin and crossresistance with other cytotoxic drugs in childhood leukemias and normal bone marrow and blood samples: a rational basis for clinical development, Leukemia, 17, 1338–1343, 2003. 60. Mitsiades, C.S., et al., Preclinical studies of the clinical development of Aplidin® for the treatment of multiple myeloma, American Society of Hematology, 45th Ann. Meeting and Exposition, abst. 250, 2003. 61. Gajate, C., An, F., Mollinedo, F., Rapid and selective apoptosis in human leukemia cells induced by Aplidin through a Fas/CD95- and mitochondrial-mediated mechanism, Clin. Cancer Res., 9, 1535–1545, 2003. 62. Cuadrado, A., et al., Aplidin™ induces apoptosis in human cancer cells via glutathione depletion and sustained activation of the epidermal growth factor receptor, Src, JNK, and p38 MAPK, J. Biol. Chem., 278, 241–250, 2003. 63. Garcia-Fernandez, L.F., et al., Aplidin™ induce the mitochondrial apoptotic pathway via oxidative stress-mediated JNK and p38 and protein kinase C, Oncogene, 21, 7533–7544, 2002. 64. Lobo, C., et al., Effect of Dehydrodidemnin B on human colon carcinoma cell lines, Anticancer Research, 17, 333–336, 1997. 65. Faircloth, G.T., et al., Schedule-dependency of Aplidin, a marine depsipeptide with antitumor activity, 10th NCI-EORTC-AACR Symposium on New Drugs in Cancer Therapy, 394, 1998. 66. Mastbergen, S.C., et al., Cytotoxicity and neurocytotoxicity of Aplidin, a new marine anticancer agent evaluated using in vitro assays, 10th NCI-EORTC-AACR Symposium on New Drugs in Cancer Therapy, 131, 1998. 67. Faircloth, G.T., et al., Preclinical characterization of Aplidin, a new marine anticancer depsipeptide, Proc. Am. Assoc. Cancer Res., 38, 692, 1997. 68. Depenbrock, H., et al., In vitro characterization of Aplidin, a new marine derived anticancer compound, on freshly explanted clonogenic human tumor cells and haematopoietic precursor cells, Br. J. Cancer, 78, 739–744, 1998. 69. D’Incalci, M., et al., New drugs of marine origin, Ann. Oncol., 7, 19, 1996. 70. Faircloth, G.T., et al., Preclinical development of Aplidin, a novel marine-derived agent with potent antitumor activity, 10th NCI-EORTC-AACR Symposium on New Drugs in Cancer Therapy, 129, 1998. 71. Faircloth, G.T., et al., Aplidin is a novel, marine derived depsipeptide with in vivo antitumor activity, Proc. Am. Assoc. Cancer Res., 39, 227, 1986. 72. Gomez, S.G., et al., In vitro hematotoxicity of Aplidin on human bone marrow and cord blood progenitor cells, Toxicology In Vitro, 15, 347–350, 2001. 73. Vera, M.D., Joullié, M.M., Natural products as probes of cell biology: 20 years of didemnin research, Med. Res. Rev., 22, 102–145, 2002. Copyright © 2005 CRC Press, LLC 74. Schmidt, U., Kroner, M., Griesser, H., Total synthesis of the didemnins — I. Synthesis of the peptolide ring, Tet. Lett., 29, 3057, 1988. 75. Schmidt, U., Kroner, M., Griesser, H., Total synthesis of the didemnins — II. Synthesis of didemnins A, B, C and prolyldidemnin A, Tet. Lett., 29, 4407, 1988. 76. Schmidt, U., Kroner, M., Griesser, H., Total synthesis of the didemnins — III. Synthesis of proptected (2R, 3S)-allo-isoleucine and (3S, 4R, 5S)-isostatine derivatives-amino acids from hydroxyl acids, Synthesis, 832–835, 1989. 77. Jou, G., et al., Total synthesis of dehydrodidemnin B. Use of uranium and phosphonium salt coupling reagents in peptide synthesis in solution, J. Org. Chem., 62, 354, 1997. 78. Anthoney, A., Paz-Ares, L., Twelves, C., Phase I And Pharmacokinetic (PK) Study of Aplidin (APL) using A 24-hour, weekly schedule, Proc. Am. Soc. Clin. Oncol., abst. 734, 2000. 79. Ciruelos, E.M., et al., Phase I clinical and pharmacokinetic study of the marine compound Aplidin (APL) administered as a 3 hour infusion every 2 weeks, Proc. Am. Soc. Clin. Oncol., abst. 422, 2002. 80. Bowman, A., et al., Phase I clinical and pharmacokinetic (PK) study of the marine compound Aplidin (APL), administered as A 1 hour weekly infusion, Proc. Am. Soc. Clin. Oncol., abst. 476, 2001. 81. Armand, J.P., et al., Phase I and pharmacokinetic study of Aplidin (APL) given as a 24-hour continuous infusion every other week (Q2w) in patients (Pts) with solid tumor (ST) and lymphoma (NHL), Proc. Am. Soc. Clin. Oncol., abst. 477, 2001. 82. Maroun, J.A., et al., Phase I study of Aplidin in a 5 day bolus Q 3 weeks in patients with solid tumors and lymphomas, Proc. Am. Soc. Clin. Oncol., abst. 2082, 2001. 83. Raymond, E., et al., Phase I (PI) trials with aplidin (APL), a new marine derived anticancer compound, Eur. J. Cancer, 37, S32, 2001. 84. Raymond, E., et al., Activity of APL, a new marine compound, against medullary thyroid carcinoma (MTC): phase I trials as screening tools for rare tumors, Ann. Oncol., 13, 22, 2002. 85. Jimeno, J., et al., Progress in the clinical development of new marine-derived anticancer compounds, Anticancer Drugs, 15, 321–329. 2004. 86. Hamann, M.T., Scheuer, P.J., Kahalalide F., A bioactive depsipeptide from the sacoglossan mollusk Elysia rufescens and the green alga Bryopsis sp., J. Am. Chem. Soc., 115, 5825–5826, 1993. 87. López-Macia, A., et al., Synthesis and structural determination of Kahalalide F, J. Am. Chem. Soc., 123, 11398–11401, 2001. 88. Hamann, M.T., et al., Kahalalides, bioactive depsipeptides from the sacoblassan mollusk Elysia rufescens and the green alga Bryopsis sp., J. Org. Chem., 61, 6594–6600, 1996. 89. Green, B.J., et al., Mollusc-algal chloroplast endosymbiosis. Photosynthesis, thylakoid protein main- tenance, and chloroplast gene expression continue for many months in the absence of the algal nucleus, Plant Physiol., 124, 331–342, 2000. 90. Bercero, M.A., et al., Chemical defences of the sacoglossan mollusk Elysia rufescens and its host alga Briopsis sp., J. Chem. Ecology, 27, 2287–2299, 2001. 91. Córdoba, S., et al., In vitro chemosensitivity, cell cycle redistribution and apoptosis induction by Kahalalide F, a new marine compound in a panel of human tumoral cells, Eur. J. of Cancer Suppl., 1, S173, 2003. 92. García-Rocha, M., Bonay, P., Avila, J., The antitumoral compound Kahalalide F acts on cell lysosomes. Cancer Lett., 99, 43–50, 1996. 93. Tartakoff A.M., Perturbations of vesicular traffic with the carboxylic ionophore monensin, Cell, 32, 1026–1028, 1983. 94. Wosikowski, K., Schuurhuis, D., Johnson, K., Identification of epidermal growth factor receptor and c-erbB2 pathway inhibitors by correlation with gene expression patterns, J. Nat. Cancer Inst., 89, 1505–1515, 1997. 95. Suarez, J., et al., A new marine-derived compound induces oncosis in human prostate and breast cancer cells, Mol. Cancer Ther., 2, 863–872, 2003. 96. Sewell, J.M., et al., Kahalalide F appears to promote necrotic cell death in hepatoma cell lines, Proc. Am. Assoc. Cancer Res., abst. 1509, 2004. 97. Janmaat, M.L., et al., Kahalalide F induces caspase-independent cytotoxicity that correlates with HER2 and/or HER3 expression levels and is accompanied by down-regulation of AKT signaling, Proc. Am. Assoc. Cancer Res., abst. 5328, 2004. Copyright © 2005 CRC Press, LLC 98. Nuijen, B., et al., Development of a lyophilized, parenteral pharmaceutical formulation of the inves- tigational polypeptide marine anticancer agent Kahalalide F, Drug Dev. Ind. Pharm., 27, 767–780, 2001. 99. Gómez, S.G., et al., In vitro toxicity of three new antitumoral drugs (trabectedin, aplidin and Kahalalide F) on hematopoietic progenitors and stem cells, Exp. Hemat., 31, 1104–1111, 2003. 100. Brown, A.P., et al., Preclinical toxicity of Kahalalide F, a new anticancer agent: single and multiple dosing regimens in the rat, Cancer Chemother. Pharmacol., 50, 333–340, 2002. 101. Schellens, J.H., et al., Phase I and pharmacokinetic study of Kahalalide F in patients with advanced androgen refractory prostate cancer, Proc. Am. Soc. Clin. Oncol., 451, 2002. 102. Circuelos, C., et al., A phase I clinical and pharmacokinetic study with Kahalalide F in patients with advanced solid tumors with a continuous weekly 1-hour iv infusion schedule, Eur. J. Cancer, 38, 33, 2002. 103. Zelek, L., et al., Preliminary results of Phase II study of ecteinascidin 743 (ET-743) with the 24-hour continuous infusion q3 week schedule in pretreated advanced/metastatic breast cancer patients [Abstract], Proc. 11th NCI-EORTC-AACR Symposium, 6, 45085, 2000. Copyright © 2005 CRC Press, LLC 13 Discovery of E7389, a Fully Synthetic Macrocyclic Ketone Analog of Halichondrin B Melvin J. Yu, Yoshito Kishi, and Bruce A. Littlefield CONTENTS I. Introduction II. Background III. Halichondrin Structure–Activity Relationship IV. Total Synthesis and Compound Supply V. Structure–Activity Relationship of Structurally Simplified Analogs VI. Synthesis of HB Analogs VII. In Vitro Characterization of E7389 VIII. In Vivo Characterization of E7389 IX. Conclusion Acknowledgments References I. INTRODUCTION E7389, a fully synthetic analog of the structurally complex marine natural product halichondrin B (HB), exhibits unique effects on tubulin dynamics, along with striking anticancer activity in preclinical models of human cancer. The nature of E7389’s antitumor activity in such models, combined with its novel tubulin-based effects and superiority over paclitaxel in several animal models, indicates that this agent may exhibit a unique spectrum of antitumor activities in the clinic that will differentiate it from existing anticancer drugs. Limited supply of natural halichondrins from the wild, a serious obstacle common to other structurally complex marine natural products isolated in low yield, represented a critical issue that all but precluded development of these compounds as drug candidates. Three issues in particular affected the halichondrins. These included securing the required absolute amount of material with a consistent impurity profile to support drug development activities, separating the desired hali- chondrin from other members of the family closely related in structure, and removing highly toxic coisolated metabolites such as okadaic acid (a known contaminant in extracts from Halichondria okadai). The limited supply of halichondrins and strong interest on the part of the U.S. National Cancer Institute (NCI) to develop HB as a new anticancer drug created the need to identify alternative sources of this important class of marine natural products. Despite the chemical complexity of the halichondrin family, total synthesis represented the most attractive option to not only identify an optimized analog for drug development but also provide the means for a reliable and renewable source of material. With the ability to address Copyright © 2005 CRC Press, LLC problems at the molecular level, to fine-tune pharmacological properties, and to adjust physico- chemical parameters, E7389 was identified as the most promising analog, with biological charac- teristics matching or surpassing those of the natural product that inspired it. Although the chemical structure of E7389 is significantly less complex than HB, it nevertheless remains a challenging synthetic target. However, contemporary synthetic organic chemistry has both the capacity and the potential to meet this challenge. With preclinical development work completed, clinical evaluation of this novel agent is now underway to establish the therapeutic value of halichondrin-based drugs as anticancer therapeutics. II. BACKGROUND HB is a polyether macrolide that exhibits remarkable anticancer activity in preclinical animal models of human cancer. First isolated in the mid-1980s from the western Pacific marine sponge Halichondria okadai Kadota,1,2 members of the halichondrin family were subsequently found in Axinella sp.,3 Phakellia carteri,4 and most recently, Lissodendoryx sp.5 Trace quantities of hali- chondrins were also detected in an extract of the black New Zealand shallow-water sponge Raspalia agminata.6 Members of the halichondrin family include halichondrins B and C, norhalichondrins A, B, and C, homohalichondrins A, B, and C, halistatins 1 and 2, isohomohalichondrin B, neo- norhalichondrin B, neohomohalichondrin B, and 55-methoxyisohomohalichondrin B (Figure 13.1). Differences between members of this family lie primarily in the level of oxygenation at C.10, C.12, and C.13 and in the length of the carbon backbone. Of these compounds the most potent include HB (1a) and homohalichondrin B (4b). The name halipyran was proposed by Pettit and coworkers for the conserved unsubstituted C.1–C.38 carbon skeleton common to the halichondrin family of marine natural products.4 Mechanistically, the halichondrins make up a subtype of tubulin-interactive antimitotic agent with subtle differences in mechanism of action from those of other antimitotics such as the vinca alkaloids.7 Although classified as a tubulin depolymerizer, the pattern of tubulin interaction by the halichondrins is highly specific.8 Testing against the 60 human tumor cell lines of the NCI’s drug screening program demonstrated that HB exhibited subnanomolar growth-inhibitory activity against a wide variety of cancer cell types in vitro, consistent with a tubulin-based antimitotic mechanism.9 These conclusions were further supported by findings that low micromolar concentrations of the drug inhibited various aspects of tubulin function in vitro, including tubulin polymerization, micro- tubule assembly, βs-tubulin crosslinking, GTP and vinblastine binding to tubulin, and tubulin- dependent GTP hydrolysis.10,11 Although such in vitro findings are consistent with a vinca alka- loid–like antitubulin mechanism, they must

still be viewed in light of the ~104-fold discrepancy between the micromolar concentrations required for antitubulin effects in vitro and the subnano- molar concentrations that inhibit cell growth. This discrepancy is shared by other tubulin-active drugs and may reflect direct or indirect effects on microtubule dynamics at low drug concentrations rather than net changes in microtubule content seen at higher drug concentrations.12,13 Other members of the halichondrin family reported to disrupt microtubule networks by preventing micro- tubule assembly include isohomohalichondrin B.14 Recent studies by Jordan and coworkers with E7389, a synthetic analog of HB, helped shed further light on the unique mechanism of microtubule disruption by the halichondrin-class mole- cules.15 Although many tubulin-based antimitotic agents exert their biological effects through direct interactions with microtubules, the halichondrins represented by E7389 suppressed microtubule dynamics by inducing nonproductive tubulin aggregates, leading to suppression of spindle microtu- bule growth events and reduced microtubule polymer mass. This mechanism differs significantly from that of other tubulin-interactive drugs such as vinblastine or paclitaxel and is consistent with the highly specific pattern of tubulin interaction by HB reported by Hamel and coworkers. The novel mechanism of E7389 relative to existing tubulin-based drugs further supports the halichondrins as Copyright © 2005 CRC Press, LLC 10 38 H H H O O H O O 31 O 1 O "A" Series 12 H O H 13 HO O O O O O HO OH O 45 H H H H O 10 X HO H O 26 O O 54 O 12 10 HO O 13 O O 19 O "C" Series 12 1a,b: Halichondrin B-C "B" Series 13 OH 2a: Halistatin 1 (X=OH) H H H H H H HO O HO H H O O O O O O O O 5 H HO 4 O 45 H RO O 45 H H OH H H H HO2C 3a-c: Norhalichondrin A-C 4a-c: Homohalichondrin A-C 5: Isohomohalichondrin B (R=H) 2b: Halistatin 2 (X=OH) 6: 55-Methoxyisohomohalichondrin B (R=Me) H H OH H O OH O HO O HO O OH O 45 H 45 H H OH O 7: Neohomohalichondrin B 8: Neonorhalichondrin B FIGURE 13.1 Halichondrin family (X = H unless otherwise noted). a unique subtype of tubulin-based antimitotics and raises the possibility that compounds from this class may exhibit unique characteristics in the clinic as cancer chemotherapeutic agents. The high in vitro potency of HB is mirrored by its extraordinary anticancer activity in a variety of preclinical animal models. HB was reported by Hirata and Uemura in 1986 to increase mean survival times of mice bearing murine B16 melanoma and P388 and L1210 leukemias.2 Efficacy in these animal models was also reported in the patent literature by Fujisawa scientists that same year.16 In 1996, Fodstad and coworkers described the remarkable activity of halichondrins isolated from natural sources in a number of human tumor xenograft models including melanoma, osteogenic sarcoma, lung cancers, and colon cancers.7 Synthetic HB prepared by Kishi and coworkers at Harvard University similarly exhibited impressive activity in human tumor xenograft models.17 For example, dose-dependent inhibition of subcutaneous LOX human melanoma xenografts was observed in nude mice following intraperitoneal administration (5–25 µg/kg). At 50 µg/kg, tumor growth was completely suppressed, with no visible signs of toxicity.17 The extraordinary in vitro potency, in vivo efficacy, and distinct antitumor profile of the halichondrins in human tumor models led the NCI to select HB for preclinical development as an anticancer chemotherapeutic agent in 1992. III. HALICHONDRIN STRUCTURE–ACTIVITY RELATIONSHIP Based on limited cytotoxicity data of naturally occurring halichondrins A–C, norhalichondrin A, and homohalichondrin B in B16 melanoma cells in vitro, Hirata and Uemura2 proposed that the 2,6,9-trioxatricyclo[3.3.2.03,7]decane system should be relatively lipophilic (HB-type) and that the Copyright © 2005 CRC Press, LLC terminal moiety should contain two or three hydroxyl groups, but not a carboxylate, for maximum expression of antitumor activity. Because the length of the halichondrin molecules was estimated to be 30–35 Å, corresponding to half the distance of a lipid biomembrane, Hirata and Uemura speculated that the 2,6,9-trioxatricyclo[3.3.2.03,7]decane portion of the molecule may be important for insertion into the lipid bilayer. Activity in the NCI 60–human tumor cell line panel of several naturally occurring and semi- synthetic halichondrin derivatives was summarized in a review by Munro.6 Hydrolysis of the macrolactone ring to the corresponding seco acid and reduction of the exo-olefin groups at C.19 and C.26 were both detrimental to in vitro biological activity. Acid-induced rearrangement of homohalichondrin B similarly provided products with diminished biological activity. Treatment with excess trifluoroacetic acid in CH2Cl2 at 20˚C, for example, led to spiroacetal epimerization at C.38 or rearrangement of the 2,6,9-trioxatricyclo[3.3.2.03,7]decane ring system to a furan deriva- tive.18 Of the two acid-induced rearrangement pathways, epimerization at C.38 was reportedly more detrimental to biological activity, affording compounds with a two-log shift in GI50 values in the NCI 60–human tumor cell line panel, along with correspondingly major changes in molecular conformation. Extrapolating these results to other naturally occurring members of the halichondrin family, Munro speculated that epimerization at C.38 would, in general, lead to halichondrin deriv- atives with diminished biological activity. For example, the C.38 epimer of isohomohalichondrin was less active against murine leukemia P388 cells in vitro compared with natural material.5 However, minor modifications to the terminal region of the halichondrins were well tolerated, with minimal impact on in vitro potency. Using HB as the seed compound, correlation coefficients from a COMPARE analysis using a GI50-centered mean graph profile as a qualitative index of antimitotic ability were plotted against GI50 values for CCRF-CEM cells as an index of cytotoxicity. The resulting scatter did not allow Munro and coworkers to draw meaningful conclusions regarding structure–activity relationships for these compounds based on subtle differences in the antimitotic mean graph profile relative to HB. However, significant differences in the average solution conformation in CDCl3 were noted for the less cytotoxic compounds despite similar COMPARE correlation coefficients. Thus, Munro concluded that conformational changes resulting from structural modifications diminished biolog- ical activity without altering the antimitotic profile characteristic of HB. IV. TOTAL SYNTHESIS AND COMPOUND SUPPLY Because of the extraordinary in vivo potency of the halichondrins, the amount of material required to support preclinical development and eventual clinical investigation was estimated to be approx- imately 10 g.6 Assuming success in clinical trials, supply of the halichondrins to meet commercial needs was similarly estimated to be relatively small, at 1–5 kg/yr. However, securing even these modest amounts through collection from the wild was determined to be unlikely because of the extremely low yield of halichondrins isolated from natural sources, the low abundance of known HB-producing sponges in the world, and the environmental effect of subjecting deep sea sponges to large-scale harvesting, with risk of extinction. As a result, compound supply represented a significant obstacle for both development and possible future commercialization of HB as an anticancer drug. Although initial aquaculture feasibility trials on Lissodendoryx sp. 1 demonstrated the potential for halichondrins to be isolated from a cultured sponge,6 the ability of scale-up operations to meet commercial demands of bulk drug substance that satisfy purity specifications in a reliable, economic, and consistent manner remains unknown. The difficulties associated with securing the needed absolute amount of material for devel- opment and possible commercialization are not unique to the halichondrins19 and represent only one aspect of the supply issue surrounding natural products isolated in extremely low yield from the producing organism. For example, separation of the desired halichondrin (e.g., HB) from other closely related family members, and the need to remove structurally unrelated but highly Copyright © 2005 CRC Press, LLC potent coisolated bioactive metabolites, added purification concerns to the list of compound supply challenges associated with collection from the wild or through aquaculture techniques. These three issues made development of the halichondrins difficult; unfortunately, NCI’s collaborative efforts to identify more widespread sources of the natural product met with only limited success, placing the program at risk of stalling at an early stage of preclinical development and leaving compound supply a serious concern for the NCI Drug Development Group (http://dtp. nci.nih.gov/docs/ddg/ddg_descript.html). Hirata and Uemura proposed that symbiotic microor- ganisms rather than the sponge itself may be responsible for producing the halichondrins as secondary metabolites.2 Although the isolation of halichondrins from several marine sponges supported this proposal, no reports regarding the isolation, tissue culture, or fermentation of the putative producing organism or organisms have appeared in the literature, making this approach an unlikely near-term solution. With compound supply an obstacle for preclinical development of this promising class of anticancer agents, alternative sources of the natural product were critically needed. The extraordi- nary biological activity of the halichondrins coupled with the extremely challenging and unique structural architecture of these compounds attracted the attention of chemistry research groups throughout the world. Although a number of laboratories reported synthetic studies on the hali- chondrins,20–23 the only successful total synthesis to date was reported by Kishi and coworkers in 1992 for the total synthesis of HB and norhalichondrin B.20 Kishi’s approach was highly convergent, bringing together four major fragments corresponding to the C.1–C.13, C.14–C.26, C.27–C.38, and C.39–C.54 regions of the molecule for final assembly (Figure 13.2). The convergent nature of this approach allowed the chemistry of individual fragments to be further optimized independent of one another, with minimal effect on late-stage coupling or protecting group strategies.20 In addition, Kishi’s synthesis was extremely flexible, allowing structure modifications to be made in a modular fashion, thereby facilitating use of key stockpiled intermediates to prepare different members of the halichondrin family (e.g., synthesis of HB or norhalichondrin B from common “right half” intermediate 11). The availability of pure natural product via total synthesis was an important factor in the NCI Decision Network Committee’s decision to select HB for preclinical development in March 1992.24 Total synthesis provided not only a reliable source of material free from potentially toxic coisolated contaminants such as okadaic acid but also the tools for directed structure modification and the opportunity to identify novel analogs with optimized biopharmaceutical and physicochemical properties not accessible through semisynthesis or derivatization of the natural product itself. The ability to address problems at the molecular level, to fine-tune pharmacological properties as well as metabolic and chemical stability, aqueous solubility, and compound purity, were all essential factors for identifying and bringing the best possible analog forward for clinical investigation. Although a number of synthetic intermediates and model systems were generated in connection with the synthetic efforts launched by research groups worldwide, biological data for only a handful of compounds were reported in the literature. In 1992, synthetic HB and several synthetic intermediates from the Kishi group at Harvard University were kindly provided to Eisai Research Institute for in vitro and in vivo biological evaluation as anticancer agents. As expected, potent cell growth-inhibitory activity was observed for synthetic HB. Surprisingly, however, synthetic intermediate 11 corresponding to the C.1–C.38 “right half” fragment of the natural product exhibited in vitro activity within an order of magnitude of the natural product in a 3–4-d growth- inhibition assay, using DLD-1 human colon cancer cells.20,25,26 This compound represented the first example of a structurally simplified synthetic analog of HB exhibiting potent antiproliferative activity, thereby providing experimental evidence that biological activity was associated with the macrocyclic moiety. Both HB and 11 blocked cells at the G2/M phase of the cell cycle, both disrupted normal mitotic spindle architecture through microtubule destabilization with few if any effects on other phases of the cell cycle, and both exhibited a similar profile in the NCI 60–cell line screen. The exciting discovery of a structurally simplified HB fragment that retains the potent Copyright © 2005 CRC Press, LLC H TBSO O OMPM MeO2C H O TBSO O H O H OTBS O TBSO TBSO 9 O OPv 10 I TBS OMPM O O H TBSO HO O O O I H TBSO 12 HO O O O H O H TBSO O 11 O O TBS OMPM O TBSO O O O O I MeO2C 13 Right Half Left Halves Halichondrin B or Norhalichondrin B FIGURE 13.2 Kishi total synthesis of halichondrin B and norhalichondrin B. cell growth-inhibitory activity of the natural product formed the basis for a research program at Eisai Research Institute to identify novel halichondrin-based anticancer agents using chemistry technology developed in the Kishi laboratory at Harvard University. Results obtained

with biotinylated macrolactone analog 14 (Figure 13.3) provided the first direct proof of any halichondrin class molecule binding directly to tubulin itself.27 Under the affinity chromatography conditions, it was not possible to identify the tubulin subtype targeted by the halichondrin macrolactone, as binding to either the α- or the β-tubulin subtype present in α/β- tubulin heterodimers would result in retention of both. The pattern of cell growth inhibition, cell cycle effects, mitotic spindle disruption, and affinity chromatography results taken together indi- cated that the structurally simplified macrolactone analogs share a similar or identical mechanism to parental HB. Thus, the structural elements necessary for anticancer biological activity reside in the macrocyclic sector of the molecule, providing a clear and compelling starting point for a structure–activity relationship investigation to identify the minimum pharmacophoric substructure for this family of marine natural products. The discovery that a structurally simplified and synthetically accessible fragment of HB retained potent anticancer activity created the possibility that supply problems associated with the natural product itself could be circumvented. Specifically, if readily accessible halichondrin analogs that retain the full spectrum of remarkable biological activity associated with this family of compounds could be identified through total synthesis, then clinical evaluation of the halichondrin class of Copyright © 2005 CRC Press, LLC FIGURE 13.3 Biotinylated derivative of compound 11. marine natural products could proceed with an optimized derivative to determine their value as chemotherapeutic agents in humans. Horita and Yonemitsu in 1997 described a similar approach from their laboratory that bifurcated HB into partially overlapping polyether and macrolactone units (Figure 13.4).28 Although the synthesis for each (compounds 15 and 16, respectively) was reported in the literature, biological activity was presented only for early precursors rather than for the target compounds themselves FIGURE 13.4 Yonemitsu synthetic study. Copyright © 2005 CRC Press, LLC FIGURE 13.5 Representative synthetic intermediates reported by Yonemitsu et al.28 (Figure 13.5 and Table 13.1). Interestingly, activity was noted for intermediates associated with both sections, although those corresponding to the polyether unit exhibited a higher level of cytotoxicity against KB and L1210 cells relative to those corresponding to the macrolactone region. Polyether analogs corresponding to the C.37–C.54 region of HB were described in the patent literature as antitumor agents with micromolar activity against a panel of 39 human tumor cell lines29 and were shown in a university publication to have modest activity against MCF-7 breast and KE-4 esophagus cancer cells.30 Because no details regarding mechanism of action for these compounds were provided, it is unclear whether any of the fragments inhibited cell growth through a tubulin-based antimitotic mechanism associated with the natural product. V. STRUCTURE–ACTIVITY RELATIONSHIP OF STRUCTURALLY SIMPLIFIED ANALOGS Despite the in vitro similarities, and in contrast to HB, compound 11 was not active in the in vivo LOX human melanoma xenograft model. Cell growth-inhibitory activity for both HB and 11 were similar under continuous exposure conditions in the 3–4-d cell growth assay. Flow cytometric analysis of U937 human histiocytic lymphoma cells, however, revealed a clear difference in ability of the two compounds to maintain a complete mitotic block (CMB) 10 h after drug washout.31 TABLE 13.1 In Vitro Activity of Halichondrin B Synthetic Intermediates Reported by Yonemitsu et al.28 IC50 (nM) Compound KB L1210 17 8.3 0.64 18 3.9 3.1 19 >100 28 Copyright © 2005 CRC Press, LLC FIGURE 13.6 Representative octahydropyrano[3,2-b]pyran macrolactone analogs. Although a 10-nM concentration of HB was sufficient to induce and maintain a CMB, concentra- tions of 11 as high as 1 µM (highest tested concentration) were ineffective against this challenge. Thus, in contrast to the natural product, the antimitotic effects of 11 were reversible on washout. Cell viability dose–response curves 5 d postwashout tracked 10-h postwashout rather than 0-h prewashout IC99 values, indicating that long-term cell viability was dependent on continued main- tenance of a CMB, rather than its initial induction. The reversible nature of 11 indicated that cell growth inhibition would only occur under continuous exposure conditions and that tumor inhibitory activity in vivo would cease when drug concentrations in blood fell below a certain critical threshold. On the basis of this working hypothesis, a proprietary cell-based assay was established at Eisai Research Institute that allowed measurement of a compound’s ability to maintain a CMB following drug washout as an index of potential antitumor efficacy in in vivo settings, using standard intermittent dosing schedules.32 Macrolactone analogs modified at the C.30–C.38 region of the molecule were subsequently screened for both intrinsic cell growth-inhibitory activity under continuous exposure conditions as well as the ability to induce an irreversible CMB (Figure 13.6 and Table 13.2).33 Although intrinsic potency was relatively insensitive to variations in this region of the molecule, the ability to induce an irreversible CMB was structure specific. For example, compound 20, prepared by the Kishi group at Harvard University in connection with an improved synthesis of the macrolactone fragment of HB, was the first structurally simplified macrolactone analog found to induce an irreversible CMB (postwashout IC99 = 220 nM). With a shortened side chain at C.36 and an epimeric hydroxyl group at C.35, this compound represented two structural changes from lead compound 11, which exhibited a postwashout IC99 value greater than 1000 nM. Inverting the C.35 hydroxyl group of 11 alone (cf. 11 and 21) was not sufficient to improve activity in the mitotic block reversibility assay. Consistent with the idea that induction of an irreversible CMB would permit manifestation of in vivo activity, compound 20 exhibited good antitumor activity in the LOX human melanoma Copyright © 2005 CRC Press, LLC TABLE 13.2 In Vitro Biological Activity of Octahydropyrano[3,2-b]Pyran Macrolactone Analogs Mitotic Block Reversibility Intrinsic Potency, U937 IC99 0 h U937 IC99 10 h Reversibility Compound DLD-1 IC50 (nM)a (nM)b (nM)c Ratio 11 4.6 30 >880 >29 20 3.4 9 220 24 21 3.6 10 >1000 >100 22 2.6 20 >1000 >50 23 1.2 7 >650 >93 24 6.0 10 1000 100 25 610.0 — — — 26 9.1 30 >1000 >33 27 59 300 >1000 >3 28 650 — — — a Growth-inhibitory potency against DLD-1 cells after 3–4 d continuous drug exposure b Induction of CMB in U937 cells after 12 h drug exposure (0 h of washout) c Maintenance of CMB 10 h after drug washout xenograft model.34 In contrast, analog 22, a direct substructure of the natural product, exhibited a nearly identical in vitro profile to diol 11, as did methoxyl derivative 23. Increasing the size of the C.31 methyl substituent to an ethyl group (e.g., 24) led to a modest loss of potency, but removing this group altogether to give compound 25 reduced activity by over two orders of magnitude. In addition to improving the in vitro profile of macrolactone 11, structural simplicity and compound accessibility were also major factors driving the structure–activity relationship investi- gation. For example, if the C.19, C.25, or C.26 substituents of compound 11 could be removed without significantly affecting biological activity (Figure 13.7), then the synthesis of the compounds could be simplified with enhanced access to these compounds. Analogs 26 and 27 indicated that complete removal of the C.19 and C.26 exo-olefins was tolerated, but removal of the C.25 methyl FIGURE 13.7 Representative deletion macrolactone analogs. Copyright © 2005 CRC Press, LLC FIGURE 13.8 Representative tetrahydropyran macrolactone analogs. group to give macrolactone 28 reduced activity 100-fold (Table 13.2). Because modifying centers contained directly within the macrocyclic core uniformly led to analogs with equivalent or reduced activity (these results apply only to the particular macrolactone analogs synthesized and tested. The possibility exists that other modifications [e.g., substitution patterns or conversion of the macrolactone ring to the corresponding macrolactam, macrocyclic ketone, carbocyclic ring, etc.] may lead to analogs where the same changes may not negatively affect activity), it was hypothesized that the macrocyclic ring may represent the minimum pharmacophoric substructure of the hali- chondrins for cell growth-inhibitory activity. As a result, the macrolactone ring was fixed and held constant as modifications to the fused C.29–C.32 pyran sector were explored. In support of this hypothesis, compound 29 (Figure 13.8) with a single tetrahydropyran ring in place of the C.29–C.36 octahydropyrano[3,2-b]pyran ring system exhibited in vitro biological activity comparable to lead compound 20 in terms of both intrinsic potency and ability to induce an irreversible CMB (Table 13.3). Modifications to the tetrahydropyran ring substituents, especially with regard to the C.32 hydroxyl group, however, had little effect on intrinsic potency, but a significant effect on the compound’s reversibility characteristics. The earlier structure–activity relationship investigation around the original octahydropyr- ano[3,2-b]pyran series of analogs demonstrated the importance of the C.31 substituent. Because small changes in substituent size (e.g., methyl to ethyl) were tolerated, extrapolation to the tetrahy- dropyran series suggested that replacing the C.31 methyl group with a methoxyl substituent would allow more ready access from commercially available carbohydrate precursors without compro- mising the biology activity of the final compound. The C.31 methoxyl derivative 30 was found to be equipotent with the corresponding C.31 methyl analog 29, supporting the interchangeability of these two groups at this center of the molecule. As observed with previous macrolactone analogs, modifications to the tetrahydropyran ring had no effect on intrinsic cell growth-inhibitory potency. Individually converting each of the hydroxyl groups of diol 30 to methyl ethers 31 and 32, in contrast, provided highly reversible compounds, Copyright © 2005 CRC Press, LLC TABLE 13.3 In Vitro Biological Activity of Tetrahydropyran and Tetrahydrofuran Macrolactone Analogs Mitotic Block Reversibility Intrinsic Potency, U937 IC99 0 h U937 IC99 10 h Reversibility Compound DLD-1 IC50 (nM)a (nM)b (nM)c Ratio Tetrahydropyran Series 29 2.5 10 170 17 30 1.8 10 300 30 31 2.1 10 >1000 >100 32 1.8 10 >1000 >100 33 2.6 10 300 30 34 2.0 3 65 22 35 1.8 2 100 50 36 >1000 — — — Tetrahydrofuran Series 37 1.0 3 100 33 38 >1000 — — — 39 0.97 7 100 14 40 6.4 10 100 10 41 1.4 3 30 10 42 >1000 — — — 43 220 — — — 44 0.67 1.5 15 10 a Growth-inhibitory potency against DLD-1 cells after 3–4 d continuous drug exposure. b Induction of CMB in U937 cells after 12 h drug exposure (0 h of washout). c Maintenance of CMB 10 h after drug washout. as measured by the U937 mitotic block reversibility assay. Shortening the C.33 side chain to a one- carbon-lower homolog or inverting the hydroxyl stereochemistry at C.32 had little effect (e.g., 33–35), but removing the C.33 side chain altogether (i.e., 36) led to complete loss of activity. Because incremental changes to the C.29–C.33 tetrahydropyran ring system did not provide a clear direction for improvement, an alternative strategy was pursued based on a working hypothesis that the C.31 substituent played a critical role in stabilizing the bioactive conformation of the macrolactone. As noted earlier, presence of the C.31 substituent was critical for good in vitro biological activity, albeit with a degree of steric tolerance. Thus, if the bioactive conformation could be reinforced by an alternative ring system, then the apparent potency limitations associated with the tetrahydropyran analogs might be circumvented. The x-ray crystal structure of norhalichondrin A p-bromophenacyl ester reported in 1985 provided an empirically derived starting point for conformational analysis of the C.29–C.33 tet- rahydropyran ring.1 Although the solid-state conformation may not necessarily represent the bio- active conformation of the halichondrins, it does represent a reasonable low-energy structure for examining the octahydropyrano[3,2-b]pyran ring conformation. 1H-nmr studies reported by Munro indicate that the average solution conformation of HB in CDCl3 may be similar to the solid-state conformation of norhalichondrin A p-bromophenacyl ester.6 In the x-ray structure, the C.29–C.33 tetrahydropyran ring adopts a twist boat conformation that places the C.31 methyl group in a pseudoequatorial position.1,35 Although a tetrahydropyran ring can adopt a similar twist boat con- formation, the tetrahydrofuran ring would be expected to do so with less conformational freedom (Figure 13.9). By stabilizing the pseudoequatorial orientation of the C.31 methyl group relative to Copyright © 2005 CRC Press, LLC O O O O O O O O FIGURE 13.9 Conformational analysis of tetrahydropyran and tetrahydrofuran analogs. the macrocyclic ring, the tetrahydrofuran analogs might be expected to exhibit superior biological activity relative to the tetrahydropyran derivatives based on favorable entropic factors. Consistent with this working hypothesis, the in vitro biological activity of tetrahydrofuran 37 (Figure 13.10) was

virtually indistinguishable from that of tetrahydropyran 34. Inverting the ste- reocenter at C.29 to give compound 38, however, provided a compound that was inactive at the highest concentration tested, consistent with the need to maintain a particular low-energy macro- lactone ring conformation (these results apply only to the particular macrolactone analogs synthe- sized and tested). Extending the C.32 hydroxymethyl side chain by either one carbon (e.g., 39) or to a vicinal propanediol (e.g., 40 and 41) had little effect, albeit showing a modest difference between the C.34 R- and S-isomers in the DLD-1 cell growth assay. As expected, based on conformational arguments, activity was critically dependent on the configuration at C.30 and C.31 (e.g., 42 and 43, respectively; these results apply only to the particular macrolactone analogs synthesized and tested). Further structural and stereochemical optimization ultimately led to tet- rahydrofuran 44, whose intrinsic potency and ability to maintain a CMB exceeded those of the tetrahydropyran derivatives. FIGURE 13.10 Representative tetrahydrofuran macrolactone analogs. Copyright © 2005 CRC Press, LLC Compounds 30 and 44, representatives from the tetrahydropyran and tetrahydrofuran series, respectively, were evaluated in the LOX human melanoma xenograft model. Unexpectedly, both failed to demonstrate in vivo efficacy despite an equal or superior in vitro profile relative to lead compound 20. Although several explanations for this lack of in vivo efficacy in the mouse xenograft model were possible, instability of the ester bond in mouse serum was regarded as the most likely explanation. In support of this hypothesis, activity of the octahydropyrano[3,2-b]pyran series of macrolactone analogs (e.g., 11) was for the most part unaffected by inclusion of up to 1% (v/v) mouse serum in the 3–4-d growth-inhibition assay using DLD-1 human colon carcinoma cells. However, activities of the simplified tetrahydropyran series of macrolactones were completely abrogated by mouse serum concentrations as low as 0.01%–0.05% in the cell culture medium.36 Elimination of biological activity did not occur with either bovine or human sera, regardless of heat inactivation status, consistent with high levels of nonspecific esterases present in mouse serum (18 U/mg protein) compared to bovine (0.3 U/mg) or human sera (0.6 U/mg). Because the seco acid of macrolactone 11 was not active in vitro, lack of in vivo efficacy for compounds 30 and 44 in the mouse xenograft model was attributed to sensitivity of the macrolactone ring to nonspecific esterases present in mouse serum. The 3–4-d DLD-1 human colon carcinoma cell growth assay was adapted to provide a functional measure of compound stability in the presence of mouse serum. At a fixed drug concentration of 100 nM, analogs were examined for cell growth-inhibitory activity in the presence of varying concentrations of mouse serum. The “mouse serum IC50 value” from this assay corresponded to the percentage of mouse serum needed to suppress 100-nM drug-mediated cell growth inhibition by 50%. Both the tetrahydropyran and tetrahydrofuran series of compounds were found to be highly sensitive to mouse serum inactivation, with compounds 35 and 37 exhibiting mouse serum IC50 values of just 0.002%. However, the reasons underlying the stability of 11 to nonspecific esterases present in mouse serum relative to the sensitivity seen in structurally simplified tetrahydropyran or tetrahydrofuran analogs were not well understood. Although conformational factors may play a role, molecular modeling failed to provide a clear explanation for the striking loss of cell growth- inhibitory activity in the presence of mouse serum esterases, although speculative, protein-binding calculations indicate that differential low-affinity binding of these compounds to serum proteins with nonspecific esterase activity may account at least in part for their observed behavior. Attempts to further improve the in vitro biological profile of the macrolactones met with limited success. As summarized in Figure 13.11, most changes were detrimental to biological activity, with greater latitude observed for those centers external to the C.1–C.30 macrolactone ring (e.g., C.32–C.38). Because minimal opportunities existed to modify the 2,6,9-trioxatricy- clo[3.3.2.03,7]decane ring system without significantly altering the global conformation of the molecule, and as members of the halichondrin family with oxygenation at C.10, C.12, or C.13 (e.g., halichondrin A, halistatins 1–2) were less potent relative to their nonhydroxylated counter- parts, the C.1–C.13 fragment of HB was held constant as structure–activity relationships were explored around the remainder of the molecule. Despite extensive investigation, instability of the macrolactone ring remained problematic. Lacking a predictive model with sufficient resolution to guide structure modification efforts, nonhydrolyzable ester bioisosteres were evaluated to identify derivatives that retain activity in the presence of mouse serum. As expected, the C.1 amide, hydroxyl, ether, and ketone analogs exhibited improved stability against nonspecific esterases. However, of the surrogate groups examined, only the ketone bioisostere provided analogs with sufficient in vitro potency to warrant further investigation.32 From these studies, compound 45 (ER-076349, NSC 707390) emerged as the prototypical macrocyclic ketone for in vivo evaluation (Figure 13.12). This analog exhibited antitumor effects in a variety of human tumor xenograft models including MDA-MB-435 breast carcinoma, COLO- 205 colon carcinoma, LOX melanoma, and NIH:OVCAR-3 ovarian carcinoma, with no evidence of toxicity based on body weight or water consumption at the doses tested.27 Copyright © 2005 CRC Press, LLC FIGURE 13.11 Closed circles indicate carbon centers that were modified to explore structure–activity rela- tionships. The arrows indicate whether changes ultimately led to an increase, decrease or no effect on in vitro potency. Consistent with the bifurcation model of HB, in which the macrocyclic region represented the minimum pharmacophoric substructure for biological activity, structure modifications to regions of the molecule distant from the macrocyclic ring (e.g., the C.32 side chain) had little effect on in vitro biological activity. For example, treatment of 45 with p-toluenesulfonyl chloride and several nucleophiles afforded a variety of derivatives differing at C.35 but exhibiting a nearly identical in vitro biological profile (Table 13.4).37 Inverting the C.34 stereocenter (e.g., 46) similarly had little effect on potency, as did cleavage of the C.34–C.35 vicinal diol to furnish alcohol 50. Treatment of 45 with p-toluenesulfonyl chloride followed by ammonia, however, provided an amine (compound 51; E7389, previously ER-086526, also NSC-707389) that for the first time exhibited a reversibility ratio of one, indicating that even a 10-h drug washout had no effect on the compound’s ability to maintain a CMB (Figure 13.13). Several other amine derivatives were subsequently prepared and also found to exhibit a low reversibility ratio (e.g., 52). In addition to conferring superior reversibility characteristics to this series, the amino group also provided a convenient handle for late-stage derivatization (e.g., acylation). Although simple amides of 51 were FIGURE 13.12 Representative macrocyclic ketone analogs. Copyright © 2005 CRC Press, LLC TABLE 13.4 In Vitro Biological Activity of Macrocyclic Ketone Analogs Mitotic Block Reversibility Intrinsic Potency, U937 IC99 0 h U937 IC99 10 h Reversibility Compound DLD-1 IC50 (nM)a (nM)b (nM)c Ratio 45 1.0 3 38 13 46 1.9 3 100 33 47 0.49 2.0 30 15 48 0.53 3 30 10 49 0.71 3 30 10 50 0.66 3 100 33 51 (E7389) 20 11 11 1 52 3.3 10 10 1 53 1.2 1 10 10 54 0.47 1 10 10 55 0.43 3 10 3.3 56 1.3 3 30 10 57 0.69 3 20 6.7 58 0.72 2 10 5 59 0.87 3 10 3.3 60 1.8 3 30 10 a Growth-inhibitory potency against DLD-1 cells after 3–4 d continuous drug exposure. b Induction of CMB in U937 cells after 12 h drug exposure (0 h of washout). c Maintenance of CMB 10 h after drug washout. FIGURE 13.13 Representative C.35 amine analogs. Copyright © 2005 CRC Press, LLC FIGURE 13.14 Representative C.35 amide analogs. highly potent in vitro (Figure 13.14), they exhibited diminished in vivo activity in human tumor xenograft models (data not shown). From the analogs prepared, compound 51 (E7389) emerged as the most promising candidate and was therefore selected for preclinical development by Eisai Co., Ltd., as a new anticancer chemotherapeutic agent. VI. SYNTHESIS OF HB ANALOGS The macrocyclic ketone analogs were prepared in a highly convergent manner, taking advantage of key building blocks, protecting group strategies, and carbon–carbon bond disconnections developed by Kishi and coworkers for the total synthesis of HB and norhalichondrin B (Figure 13.15). Having established the minimum pharmacophore for cytotoxic activity, the C.1–C.13 and C.14–C.26 syn- thetic intermediates were fixed and stockpiled, allowing final coupling of the three key fragments to proceed under a standard set of conditions. In this manner, the synthesis of new analogs was reduced to preparation of various C.27–C.35 tetrahydrofuran derivatives for conversion to final product. For example, aldehyde 61 was coupled with vinyl iodide 62 under Nozaki–Hiyama–Kishi conditions38,39 and cyclized with base. Removal of the MPM protecting group and introduction of the sulfone moiety afforded intermediate 63. Coupling with C.1–C.13 aldehyde 64 and functional group modification around the C.1 ketone provided an intermediate that underwent a smooth intramolecular Nozaki–Hiyama–Kishi reaction to close the macrocyclic ring between C.13 and C.14. Finally, treatment with tetrabutylammonium fluoride and pyridinium p-toluenesulfonate gen- erated the 2,6,9-trioxatricyclo[3.3.2.03,7]decane system to furnish 45 in good yield. Modification of the C.32 side chain subsequently afforded the desired macrocyclic ketone analogs. VII. IN VITRO CHARACTERIZATION OF E7389 The halichondrins exhibited a unique pattern of interaction with tubulin compared with other antimitotic drugs,10 indicating that E7389 may similarly demonstrate unique interactions with tubulin and microtubules. Both E7389 and HB shared low- to subnanomolar cell growth-inhibitory activity against multiple human tumor cell lines in vitro, both performed similarly in the NCI 60- cell line screen, both induced G2/M cell cycle arrest with mitotic spindle disruption consistent with a tubulin-based antimitotic mechanism, and both directly inhibited tubulin polymerization with nearly identical IC50 values. Copyright © 2005 CRC Press, LLC FIGURE 13.15 Synthesis of macrocyclic ketone analogs. Recent work by Jordan and coworkers revealed that E7389 suppressed microtubule dynamics through a unique mechanism that differed significantly from those of other tubulin-active drugs such as paclitaxel and vinblastine.15 For example, unlike other antimitotic drugs, this compound induced the formation of nonproductive tubulin aggregates from soluble α/β-tubulin heterodimers that effectively reduced the concentration of soluble tubulin available for productive microtubule growth. Interestingly, there was no effect on treadmilling of microtubules in vitro, and in living cells E7389 did not alter catastrophe and rescue frequencies or the shortening events of microtubule dynamic instability at the concentrations tested. The unique characteristics of E7389 in these studies further emphasized the mechanistic differences that set the halichondrin class of compounds apart from other tubulin-interactive agents. Because the role of microtubule-associated proteins and microtubule-regulatory proteins in tubulin aggregate formation induced by E7389 remains unclear at this time, further studies will be needed to elucidate what effects, if any, this agent may have on the interactions between such proteins and microtubules in the context of drug-induced alter- ations in microtubule function and dynamics. In addition, studies with radio-labeled E7389 may help clarify whether it binds directly to microtubules or only to soluble tubulin heterodimers. In vitro metabolism studies using a series of human enzyme preparations including liver microsomal preparations, S9 fractions, recombinant drug-metabolizing cytochrome P450 and phase 2 conjugative enzymes, and primary human hepatocytes indicate that E7389 is mainly metabolized by CYP3A4 at pharmacologically relevant concentrations. Suppression of CYP3A4 activity was modest, with an apparent Ki of 3–10 µM, as determined by form-specific nifedipine dehydration, testosterone 6ß-hydroxylation, and R-warfarin 10-hydroxylation. No induction of CYP1A and CYP3A enzymes was observed immunochemically and enzymatically using primary human hepa- tocytes. Thus, based on a low anticipated therapeutic dose in humans, clinical drug–drug interactions with E7389 are anticipated to be unlikely.40 VIII. IN VIVO CHARACTERIZATION OF E7389 The subnanomolar growth-inhibitory activity of E7389 in vitro was reflected in the marked anti- tumor activity of this agent in vivo at 0.05–1 mg/kg against several human cancer xenografts, Copyright © 2005 CRC Press, LLC including MDA-MB-435 breast cancer, COLO 205 colon cancer, LOX melanoma, and NIH:OVCAR-3 ovarian cancer (Figure 13.16).27 In the MDA-MB-435 breast cancer model (Figure 13.16A), treatment with E7389 led to regres- sion of measurable tumors that were present at the start of dosing. No evidence of toxicity was observed at any dose based on body weight losses or decreased water consumption. Notably, all doses of E7389 over a fourfold range were either equally efficacious or superior to paclitaxel, which was dosed at its empirically determined MTD. In contrast, paclitaxel exhibited complete tumor suppression only over a narrow 1.7-fold dose range in this model (data

not shown). The ability of E7389 to completely suppress tumor growth at several doses below the maximum tolerated dose (MTD) is unusual for a cytotoxic agent and may be a consequence of its unique mechanism of action. In the COLO 205 colon cancer xenograft model (Figure 13.16B), E7389 also showed frank regression and long-term suppression of tumor regrowth superior to that seen with paclitaxel at its empirically determined MTD. As in the MDA-MB-435 model, no evidence of toxicity based on body weight losses or decreased water consumption was observed. In the LOX melanoma model (Figure 13.16C), treatment with E7389 led to complete tumor suppression over a fivefold dose range, again representing an unusually wide therapeutic window for a cytotoxic agent. In this model, tumor regrowth rates in the 0.5 mg/kg E7389 group were delayed significantly beyond that seen with 12.5 mg/kg paclitaxel. Finally, in the NIH:OVCAR-3 ovarian cancer xenograft model (Figure 13.16D), E7389 led to complete tumor suppression equivalent to paclitaxel, both during dosing and throughout the regrowth phase. The ability of E7389 to induce complete tumor suppression over a relatively wide dose range below the MTD is in sharp contrast to the behavior of standard cytotoxic drugs, irrespective of mechanistic class. Expression of in vivo anticancer efficacy by paclitaxel, doxorubicin, or the vinca alkaloids, for example, required dosing at or near the MTD, reflecting a delicate balance between efficacy and toxicity. Although speculative, the exceptionally wide therapeutic window for E7389 in preclinical xenograft models may be related to its novel mode of interaction with tubulin and raises the possibility that this agent may exhibit unique characteristics in the clinic that will distinguish it from existing tubulin-based chemotherapeutic agents. IX. CONCLUSION In the 18 years since Hirata and Uemura first described the halichondrins, each of the challenges associated with bringing this class of structurally complex marine natural products through the discovery and preclinical development pipeline was successfully met with dramatic results. The first — and to date only — total synthesis of HB reported by Kishi and coworkers provided the basis for directed structure modifications by Eisai scientists to transform the original polyether macrolactone to a structurally simplified macrocyclic ketone through a rational step-wise series of modifications that did not compromise the natural product’s remarkable biological profile. Starting from synthetic intermediate 11 prepared in connection with Kishi’s total synthesis, the minimum pharmacophoric substructure of the halichondrins was determined to be the C.1–C.34 macrolactone fragment. Optimization of the C.29–C.38 ring system ultimately led to structurally simplified macrolactones that retained the natural product’s extraordinary biological activity. Stabilization of the macrolactone ring against serum inactivation and final structure optimization afforded a series of highly potent analogs from which E7389 emerged as the most promising, with outstanding pharmacological, toxicological, and physicochemical characteristics. With the identification of an optimized fully synthetic drug candidate, the obstacles commonly encountered during the development of structurally complex marine natural products were circum- vented.19 For example, limited compound supply was a critical issue that all but precluded devel- opment of HB as an anticancer agent in the early 1990s. Access to sufficient quantity of the producing organism to secure the needed absolute amount of material in a reliable, consistent, and Copyright © 2005 CRC Press, LLC (A) (B) FIGURE 13.16 Inhibition of human tumor xenograft growth in vivo by E7389 and paclitaxel. Nude mice bearing MDA-MB-435 breast cancer (A), COLO 205 colon cancer (B), LOX melanoma (C), and NIH:OVCAR- 3 ovarian cancer (D). Periods of dosing are indicated by gray bars underlying the x axes. COLO 205 and MDA-MB-435 models: E7389 was administered intravenously (MDA-MB-435) or intraperitoneally (COLO 205), using a q2d×3 schedule for four weekly cycles, beginning on day 13. LOX model: E7389 was admin- istered by daily intraperitoneal injection on days 3–7 and 10–14 inclusive. NIH:OVCAR-3 model: E7389 was administered intravenously beginning on day 40 and continuing on a q2d×3 schedule for three weekly cycles. Control groups in all of the studies received appropriate vehicle injections. Paclitaxel was administered at its empirically determined MTD for each treatment regimen in all of the experiments. Plotted mean tumor volumes occurring after onset of any complete remissions represent averages of only those animals continuing to bear measurable tumors. (Modified from Towle, et al., Cancer Res., 61, 1013, 2001. With permission.) (Continued) economic manner; separation of the desired drug substance from other structurally related members of the halichondrin family; and removing trace amounts of highly toxic coisolated contaminants to ensure safety were contributing factors that made collection from the wild or through aquaculture techniques impractical, difficult, or unknown. Total synthesis, however, allowed ready access to halichondrin analogs with a defined and consistent impurity profile in sufficient quantity to support Copyright © 2005 CRC Press, LLC (C) (D) FIGURE 13.16 (Continued) all preclinical research activities, satisfy quality criteria, and prevent compound supply from becom- ing a critical obstacle in the path for clinical evaluation. Despite significant simplification relative to the natural product that inspired it, the chemical structure of E7389 remains a synthetically challenging target that exceeds the structural com- plexity of any other totally synthetic drug on the market today. Whereas success in clinical trials will unequivocally establish the potential of E7389 to become an exciting and important drug in the war on cancer, it will be the development of an economically feasible route for possible commercial production that will determine the long-term success of this program. Although hurdles remain, contemporary synthetic organic chemistry has both the capacity and the potential to meet this challenge.20 Tubulin-based antimitotics such as paclitaxel represent a class of agents that is clinically proven. In several human tumor xenograft animal models, E7389 exhibited superior efficacy to paclitaxel, with antitumor activity observed at dose levels well below its MTD. The exceptionally wide therapeutic window of E7389 in preclinical animal models, combined with its novel mechanism of tubulin heterodimer aggregation and superior efficacy to paclitaxel, raises the possibility that E7389 may exhibit a unique profile in the clinic that will differentiate it from other anticancer drugs. Copyright © 2005 CRC Press, LLC With preclinical development completed and work well underway to address future compound supply needs, the stage is finally set to establish the therapeutic value of halichondrin-based agents in humans. E7389 is currently undergoing phase I clinical trials as an anticancer chemotherapeutic agent with first-in-man studies initiated collaboratively between Eisai Co., Ltd., and NCI.41 If results in the clinic parallel the outstanding preclinical activity observed in animal models, then E7389 will offer new hope for cancer patients and their families and will help bring the dream of getting drugs from the sea another step closer to reality. ACKNOWLEDGMENTS The discovery and development of E7389 was possible only through the dedication, vision, and spirit of innovation shared by the many Eisai scientists who made this human healthcare program a reality. In addition, we express our appreciation for valuable assistance of the NCI through a collaborative research and development agreement to help bring the halichondrin class of marine natural products from laboratory concept to clinical investigative drug. REFERENCES 1. Uemura, D., et al., Norhalichondrin A: an antitumor polyether macrolide from a marine sponge, J. Am. Chem. Soc., 107, 4796, 1985. 2. Hirata, Y. and Uemura, D., Halichondrins — antitumor polyether macrolides from a marine sponge, Pure Appl. Chem., 58, 701, 1986. 3. Pettit, G. R., et al., Isolation and structure of the cell growth inhibitory constituents from the western pacific marine sponge Axinella sp., J. Med. Chem., 34, 3339, 1991. 4. Pettit, G. R., et al. Isolation and structure of halistatin 1 from the eastern Indian Ocean marine sponge Phakellia carteri, J. Org. Chem., 58, 2538, 1993. 5. (a) Litaudon, M., et al., Antitumor polyether macrolides: new and hemisynthetic halichondrins from the New Zealand deep-water sponge Lissodendoryx sp., J. Org. Chem., 62, 1868, 1997; (b) Gravelos, D.G., et al., Halichondrins: cytotoxic polyether macrolides, European Patent Application, EP 0 572 109, 1993. 6. Lill, J.B., et al., The halichondrins: chemistry, biology, supply and delivery, in Drugs from the Sea, Fusetani, N., Ed., Karger, New York, 2000, 134. 7. Fodstad, Ø., et al., Comparative antitumor activities of halichondrins and vinblastine against human tumor xenografts, J. Exp. Ther. Oncol., 1, 119, 1996. 8. Hamel, E., Natural products which interact with tubulin in the vinca domain: maytansine, rhizoxin, phomopsin A, dolastatins 10 and 15 and halichondrin B, Pharmacol. Ther., 55, 31, 1992. 9. Bai, R.L., et al., Halichondrin B and homohalichondrin B, marine natural products binding in the vinca domain of tubulin. Discovery of tubulin-based mechanism of action by analysis of differential cytotoxicity data, J. Biol. Chem., 266, 15882, 1991. 10. Ludueña, R.F., et al., Interaction of halichondrin B and homohalichondrin B with bovine brain tubulin, Biochem. Pharmacol., 45, 421, 1993. 11. Bai, R., et al., Spongistatin 1, a highly cytotoxic, sponge-derived, marine natural product that inhibits mitosis, microtubule assembly, and the binding of vinblastine to tubulin, Mol. Pharmacol., 44, 757, 1993. 12. Jordan, M.A., Thrower, D., and Wilson, L. Mechanism of inhibition of cell proliferation by vinca alkaloids, Cancer Res., 51, 2212, 1991. 13. Jordan, M.A., et al., Mechanism of mitotic block and inhibition of cell proliferation by taxol at low concentrations, Proc. Natl. Acad. Sci. USA, 90, 9552, 1993. 14. García-Rocha, M., García-Gravalos, M.D., and Avila, J., Characterization of antimitotic products from marine organisms that disorganize the microtubule network: ecteinascidin 743, isohomohalichondrin- B and LL-15, Br. J. Cancer, 73, 875, 1996. 15. Kamath, K., et al., E7389, a synthetic analog of halichondrin B, suppresses microtubule dynamics in living MCF7 cells by a novel mechanism, presented at the annual meeting of the American Association for Cancer Research, Washington, DC, July 11–14, 2003, LB-58. Copyright © 2005 CRC Press, LLC 16. Tsukitani, Y., et al., Japanese Patent Application JP 61191687, 1986; Japanese Patent 92001751, 1992. 17. Towle, M.J., et al., In vitro and in vivo anticancer properties and cell cycle effects of synthetic halichondrin B, presented at the annual meeting of the American Association for Cancer Research, Toronto, Ontario, March 18–22, 1995, 2342. 18. Hart, J.B., Blunt, J.W., and Munro, M.H.G., Acid-catalyzed reactions of homohalichondrin B, a marine sponge-derived antitumor polyether macrolide, J. Org. Chem., 61, 2888, 1996. 19. Fusetani, N., Introduction, in Drugs from the Sea, Fusetani, N., Ed., Karger, New York, 2000, 1. 20. (a) Aicher, T.D. and Kishi, Y., Synthetic studies towards halichondrins, Tetrahedron Lett., 28, 3463, 1987; (b) Aicher, T.D., et al., Total synthesis of halichondrin B and norhalichondrin B. J. Am. Chem. Soc., 114, 3162, 1992; (c) Aicher, T.D., et al., Synthetic studies towards halichondrins: synthesis of the C.27-C.38 segment, Tetrahedron Lett., 33, 1549, 1992; (d) Buszek, K.R., Synthetic studies towards halichondrins: synthesis of the left half of halichondrins, Tetrahedron Lett., 33, 1553, 1992; (e) Fang, F.G., et al., Synthetic studies towards halichondrins: synthesis of the left halves of norhalichondrins and homohalichondrins, Tetrahedron Lett., 33, 1557, 1992; (f) Duan, J.J. and Kishi, Y., Synthetic studies on halichondrins: a new practical synthesis of the C-1–C-12 segment, Tetrahedron Lett., 34, 7541, 1993; (g) Chen, C., Tagami, K., and Kishi, Y., Ni(II)/Cr(II)-mediated coupling reaction: an asymmetric process, J. Org. Chem., 60, 5386, 1995; (h) Stamos, D.P. and Kishi, Y., Synthetic studies on halichondrins: a practical synthesis of the C.1-C.13 segment, Tetrahedron Lett., 37, 8643, 1996; (i) Stamos, D.P. and Kishi, Y., Synthetic studies on halichondrins: a practical synthesis of the C.1- C.13 segment, Tetrahedron Lett., 37, 8643, 1996; (j) Stamos, D.P., Taylor, A.G., and Kishi, Y., A mild preparation of vinyliodides from vinylsilanes, Tetrahedron Lett., 37, 8647, 1996; (k) Stamos, D.P., Chen, S.S., and Kishi, Y., New synthetic route to the C.14 – C.38 segment of halichondrins, J. Org. Chem., 62, 7552, 1997; (l) Stamos, D.P., et al., Ni(II)/Cr(II)-mediated coupling reaction: beneficial effects of 4-tert-butylpyridine as an additive and development of new and improved workup proce- dures, Tetrahedron Lett., 38, 6355, 1997; (m) Stamos, D.P., Chen, S.S., and Kishi, Y., New synthetic route to the C.14-C.38 segment of halichondrins, J. Org. Chem., 62, 7552, 1997; (n) Xie, C., Nowak, P., and Kishi, Y., Synthesis of the C20-C26 building block of halichondrins via a regiospecific and stereoselective SN2’ reaction, Org. Lett., 4, 4427, 2002; (o) Wan, Z.-K., et al., Asymmetric Ni(II)/Cr(II)-mediated coupling reaction: stoichiometric process, Org. Lett., 4, 4431, 2002; (k) Choi, H., et al., Asymmetric Ni(II)/Cr(II)-mediated coupling reaction: catalytic process, Org. Lett., 4, 4435, 2002; (p) Choi, H., et al.,

Synthetic studies on the marine natural product halichondrins, Pure Appl. Chem., 75, 1, 2003. 21. (a) Kim, S. and Salomon, R.G., Total synthesis of halichondrins: highly stereoselective construction of a homochiral pentasubstituted H-ring pyran intermediate from -D-glucose, Tetrahedron Lett., 30, 6279, 1989; (b) Cooper, A.J. and Salomon, R.G., Total synthesis of halichondrins: enantioselective construction of a homochiral pentacyclic C1-C15 intermediate from D-ribose, Tetrahedron Lett., 31, 3813, 1990; (c) DiFranco, E., Ravikumar, V.T., and Salomon, R.G., Total synthesis of halichondrins: enantioselective construction of a homochiral tetracyclic KLMN-ring intermediate from D-mannitol, Tetrahedron Lett., 34, 3247, 1993; (d) Cooper, A.J., Pan, W., and Salomon, R.G., Total synthesis of halichondrin B from common sugars: an F-ring intermediate from D-glucose and efficient construction of the C1 to C21 segment, Tetrahedron Lett., 34, 8193, 1993. 22. (a) Burke, S.D., Buchanan, J.L., and Rovin, J.D., Synthesis of a C(22) C(34) halichondrin precursor via a double dioxanone-to-dihydropyran rearrangement, Tetrahedron Lett., 32, 3961, 1991; (b) Burke, S.D., et al., An expeditious synthesis of the C(1)-C(14) subunit of halichondrin B, Tetrahedron Lett., 35, 703, 1994; (c) Burke, S.D., Zhang, G., and Buchanan, J.L, Enantioselective synthesis of a halichondrin B C(20)–C(36) precursor, Tetrahedron Lett., 36, 7023, 1995; (d) Burke, S.D., et al., Synthetic studies toward complex polyether macrolides of marine origin, Spec. Publ. R. Soc. Chem., 198 (Anti-Infectives), 73, 1997; (e) Burke, S.D., Austad, B.C., and Hart, A.C., An expeditious synthesis of the C(38)-C(54) halichondrin B subunit, J. Org. Chem., 63, 6770, 1998; (f) Burke, S.D., Quinn, K.J., and Chen, V.J., Synthesis of a C(22)-C(34) halichondrin B precursor via ring opening-double ring closing metathesis, J. Org. Chem., 63, 8626, 1998; (g) Burke, S.D., et al., Halichondrin B: synthesis of the C(1)-C(15) subunit, J. Org. Chem., 65, 4070, 2000; (h) Austad, B.C., Hart, A.C., and Burke, S.D., Halichondrin B: synthesis of the C(37)-C(54) subunit, Tetrahedron, 58, 2011, 2002; (i) Jiang, L. and Burke, S.D., A novel route to the F-ring of halichondrin B. Diastereoselection in Pd(0)- mediated meso and C2 diol desymmetrization, Org. Lett., 4, 3411, 2002; (j) Jiang, L., Martinelli, J.R., Copyright © 2005 CRC Press, LLC and Burke, S.D., A practical synthesis of the F-ring of halichondrin B via ozonolytic desymmetrization of a C2-symmetric dihydroxycyclohexene. J. Org. Chem., 68, 1150, 2003; (k) Lambert, W.T. and Burke, S.D., Halichondrin B: synthesis of a C1-C14 model via desymmetrization of (+)-conduritol E, Org. Lett., 5, 515, 2003. 23. (a) Horita, K., et al., Chiral synthesis of polyketide-derived natural products. 46. Synthetic studies of halichondrin B, an antitumor polyether macrolide isolated from a marine sponge. 1. Stereoselective synthesis of the C1-C13 fragment via construction of the B and A rings by kinetically and thermo- dynamically controlled intramolecular Michael reactions, Synlett, 38, 1994; (b) Horita, K., et al., Chiral synthesis of polyketide-derived natural products. 47. Synthetic studies of halichondrin B, an antitumor polyether macrolide isolated from a marine sponge. 2. Efficient synthesis of C16-C26 fragments via construction of the D ring by a highly stereocontrolled iodoetherification, Synlett, 40, 1994; (c) Horita, K., Chiral synthesis of polyketide-derived natural products. 49. Synthetic studies of halichondrin B, an antitumor polyether macrolide isolated from a marine sponge. 4. Synthesis of the C37-C54 subunit via stereoselective construction of three consecutive J, K, and L rings, Synlett, 46, 1994; (d) Horita, K., et al., Chiral synthesis of polyketide-derived natural products. 48. Synthetic studies of halichondrin B, an antitumor polyether macrolide isolated from a marine sponge. 3. Synthesis of C27-C36 subunit via completely stereoselective C-glycosylation to the F ring, Synlett, 43, 1994; (e) Horita, K., et al., Synthetic studies of halichondrin B, an antitumor polyether macrolide isolated from a marine sponge. 5. A highly concise and efficient synthesis of the C37-C54 tricyclic JKL-ring part, Heterocycles, 42, 99, 1996; (f) Horita, K., et al., Synthetic studies of halichondrin B, an antitumor polyether macrolide isolated from a marine sponge. 6. Synthesis of the C1-C15 unit via stereoselective construction of the B and A rings by kinetically and thermodynamically controlled Michael reactions with the aid of computational search for dominant conformers, Chem. Pharm. Bull., 45, 1265, 1997; (g) Horita, K., et al., Synthetic studies of halichondrin B, an antitumor polyether macrolide isolated from a marine sponge. 7. Synthesis of two C27-C36 units via construction of the F ring and completely stereoselective C-glycosylation using mixed Lewis acids, Chem. Pharm. Bull., 45, 1558, 1997; (h) Horita, K., et al., Synthetic studies of halichondrin B, an antitumor polyether macrolide isolated from a marine sponge. 8. Synthesis of the lactone part (C1-C36) via Horner- Emmons coupling between C1-C15 and C16-C36 fragments and Yamaguchi lactonization, Tetrahedron Lett., 38, 8965, 1997; (i) Horita, K., et al., Synthetic studies on halichondrin B, an antitumor polyether macrolide isolated from a marine sponge. 9. Synthesis of the C16-C36 unit via stereoselective construction of the D and E rings, Chem. Pharm. Bull., 46, 1199, 1998; (j) Horita, K., Nishibe, S., and Yonemitsu, O., Research on antitumor active site of marine source natural product, halichondrin B. International Congress Series, 1157 (Towards Natural Medicine Research in the 21st Century), 327, 1998; (k) Yonemitsu, O., Yamazaki, T., and Uenishi, J.-I., On the stereoselective construction of the B and A rings of halichondrin B. A PM3 study, Heterocycles, 49, 89, 1998; (l) Horita, K., Nishibe, S., and Yonemitsu, O., Synthetic study of a highly antitumorigenic marine phytochemical, halichondrin B, in Phytochemicals and Phytopharmaceuticals, Shahidi, F. and Ho, C.-T., Eds., AOCS Press, Champaign, IL, 2000, chap. 35. 24. Michael R. Boyd, Personal communication, 1992. 25. Towle, M. J., Kishi, Y., and Littlefield, B. A., Unpublished observations, 1992. 26. Yoon, S.K., et al., U.S. Patent Application 849769, 1992; U.S. Patent 5338865, 1994; U.S. Patent 5436238, 1995. 27. Towle, M.J., et al., In vitro and in vivo anticancer activities of synthetic macrocyclic ketone analogs of halichondrin B, Cancer Res., 61, 1013, 2001. 28. Horita, K., Nishibe, S., and Yonemitsu, O., Research on anti-tumor active site of marine source natural product, halichondrin B, International Congress Series, 1157 (Towards Natural Medicine Research in the 21st Century), 327, 1998. 29. Horita, K., Takahashi, N., and Nishizawa, M., Japanese Patent JP 2003261447, 2003. 30. Horita, K., et al., Research on anti-tumor active site of marine source compound, halichondrin B. Inhibitory activity of C37-C54 part derivatives toward the breast and the esophagus cancer cell lines, Bull. Marine Biomed. Inst. Sapporo Med. Univ., 5, 35, 2002. 31. Towle, M.J., et al., In vivo anticancer activity of synthetic halichondrin B macrocyclic ketone analogs ER-076349 and ER-086526 correlates with ability to induce irreversible mitotic blocks, presented at Copyright © 2005 CRC Press, LLC the annual meeting of the American Association for Cancer Research, New Orleans, LA, March 24–28, 2001. Abstract 1976. 32. (a) Littlefield, B.A., et al., U.S. Patent 6214865, 2001; (b) Littlefield, B.A., et al., U.S. Patent 6365759, 2002; Littlefield, B.A., et al., U.S. Patent 6469182, 2002; (c) Littlefield, B.A. and Towle, M.J., U.S. Patent 6653341, 2003. 33. Wang, Y., et al., Structure-activity relationships of halichondrin B analogs: modifications at C.30-C38, Bioorg. Med. Chem. Lett., 10, 1029 (2000). 34. Eisai Research Institute, 1995. 35. Supplementary crystallographic data for norhalichondrin A p-bromophenacyl ester (CCDC 231897). http://www.ccdc.cam.ac.uk/data_request/cif. 36. Towle, M.J., et al., Halichondrin B macrocyclic ketone analog E7389: medicinal chemistry repair of lactone ester instability generated during structural simplification to clinical candidate, presented at the annual meeting of the American Association for Cancer Research, San Francisco, CA, April 6–10, 2002, 5721. 37. Zheng, W., et al., Synthetic macrocyclic ketone analogs of halichondrin B: structure-activity relation- ships, presented at the annual meeting of the American Association for Cancer Research, San Fran- cisco, CA, April 1–5, 2000, 1915. 38. (a) Jin, H, et al., Catalytic effect of nickel(II) chloride and palladium(II) acetate on chromium(II)- mediated coupling reaction of iodo olefins with aldehydes, J. Am. Chem. Soc., 108, 5644, 1986; (b) Takai, K., et al., Reactions of alkenylchromium reagents prepared from alkenyl trifluoromethane- sulfonates (triflates) with chromium(II) chloride under nickel catalysis, J. Am. Chem. Soc., 108, 6048, 1986. 39. (a) Nozaki, H. and Takai, K., Nucleophilic addition of organochromium reagents to carbonyl com- pounds, Proc. Jpn. Acad., 76, 123, 2000; (b) Fürstner, A., Carbon-carbon bond formations involving organochromium(III) reagents, Chem. Rev., 99, 991, 1999; (c) Wessjohann, L.A. and Scheid, G., Recent advances in chromium(II)- and chromium(III)-mediated organic synthesis, Synthesis 1, 1999; (d) Avalos, M., et al., Synthetic variations based on low-valent chromium: new developments, Chem. Soc. Rev., 28, 169, 1999; (e) Saccomano, N.A., Organochromium reagents, in Comprehensive Organic Synthesis, Trost, B. M., Fleming, I., Eds., Pergamon Press, Oxford, UK, 1991; vol. 1, chap. 1.6. 40. Zhang, Z.-Y., et al., Characterization of in vitro metabolism of an anticancer agent E7389: prediction of the potential risk of clinical drug-drug interactions, presented at the 12th North American meeting of the International Society for the Study of Xenobiotics, Providence, RI, October 12–16, 2003, 367. 41. Synold, T.W., et al., Human pharmacokinetics of E7389 (halichondrin B analog), a novel anti- microtubule agent undergoing phase I investigation in the California Cancer Consortium (CCC), presented at the American Society of Clinical Oncology, Chicago, IL, May 31–June 3, 2003, 575. Copyright © 2005 CRC Press, LLC 14 HTI-286, A Synthetic Analog of the Antimitotic Natural Product Hemiasterlin Raymond J. Andersen and Michel Roberge CONTENTS I. Introduction II. Natural Product Lead Structures III. Biological Activity of the Natural Products IV. Synthesis of Hemiasterlin V. Synthesis of Analogs and SAR VI. Biological Activity of HTI-286 VII. Preclinical and Clinical Development of HTI-286 VIII. Summary Acknowledgment References I. INTRODUCTION HTI-286 (1) is a synthetic experimental anticancer drug currently in phase II clinical trials for the treatment of non–small cell lung cancer.1 It shows activity in a wide variety of tumor xenograft models, including several multidrug-resistant tumors. The lead structure for the development of HTI-286 was the sponge tripeptide hemiasterlin (2), a microtubule depolymerizing agent that kills cells by causing mitotic arrest, leading to apoptosis. The sequence of discoveries that led to the development of HTI-286 and a profile of its biological activities are described here. II. NATURAL PRODUCT LEAD STRUCTURES The hemiasterlins, criamides, and milnamides are a small family of tri- and tetrapeptide cytotoxins isolated from a variety of marine sponges. Hemiasterlin (2) was first isolated by Kashman and coworkers in 1994 from the sponge Hemiasterella minor (Kirkpatrick) collected in Sodwana Bay, South Africa.2 At virtually the same time, Crews and coworkers reported the isolation of milnamide A (3) from the sponge Auletta cf. constricta collected in Milne Bay, Papua New Guinea.3 Both source sponges also contained cyclic depsipeptide cytotoxins. H. minor yielded jaspamide (jasplak- inolide) (4) and geodiamolide TA (5) in addition to hemiasterlin (2), and A. constricta contained jaspamide (4) and milnamide A (3). Hemiasterlin (2) is a tripeptide composed of the rare amino acid tert-leucine and the two novel amino acids 4-amino-2,5-dimethylhex-2-enoic acid and N,N,β,β-tetramethyltryptophan. Tert-Leu had been previously reported only from the discodermins,4 isolated from the marine sponge Discodermia kiiensis, and the bottromycin peptides, produced by cultured actinomycetes.5 Copyright © 2005 CRC Press, LLC 24 22 R1 23 O O 14 15 O 30 O 5 18 N N 3 11 20 26 28 OH 10 N N 12 19 25 29 OH H H HN O 7 16 9 HN O N 31 34 1 17 32 33 13 R 1 2 R=R1=Me 6 R=H, R1=Me 7 R=R1=H 17 R=Me, R1=H H N NH2 NH O O O O OH N N N N OH N H H H N O HN O N O N R 3 8 R=H 9 R=Me O O N N OH H N N O 18 STRUCTURES 1–3, 6–9, 17, 18. Milnamide A (3) has a β-carboline, formally derived from N,N,β,β-tetramethyltryptophan via a Pictet Spengler condensation with formaldehyde, in place of the N,N,β,β-tetramethyltryptophan residue in hemiasterlin (2). The 4-amino-2,5-dimethylhex-2-enoic acid residue in 2 and 3 appears to have a biogenetic origin involving N-methylvaline and propionic acid building blocks. There were no relative or absolute configurations assigned to the amino acids in hemiasterlin (2) or milnamide A (3) in the original reports of their structures. Kashman noted there was an NOE observed between the olefinic methyl (H-34) and tryptophan N-methyl (H-17) proton resonances in hemiasterlin (2), indicating a cyclic conformation of the C11 to C29 peptide backbone, probably stabilized by an electrostatic attraction between the spatially close C-terminal carboxylate anion and N-terminal N-methyl ammonium cation of the zwitterionic linear peptide.2 In 1995, Andersen

and coworkers reported the isolation of hemiasterlin (2), hemiasterlins A (6) and B (7), criamides A (8) and B (9), and geodiamolides A (10) to G (16) from the sponge Cymbastela sp. collected on reefs off of Madang, Papua New Guinea.6 All three amino acids in hemiasterlin were shown to have the L configuration by a combination of Marfey’s analysis on the natural product and single-crystal x-ray diffraction analysis of hemiasterlin methyl ester.7 It was also shown that the arginine residue in the criamides had the L configuration. The solid-state conformation of hemiasterlin methyl ester has an extended linear peptide backbone, which supports Kashman’s suggestion that ionic interactions play an important role in the formation of a cyclic peptide backbone conformation for hemiasterlin in solution. Boyd and coworkers reported finding hemiasterlins in two sponges also collected in Papua New Guinea.8 An Auletta sp. yielded the known compounds hemiasterlin (2), hemiasterlin A (6), and several geodiamolides, along with the new tripeptide hemiasterlin C (17). Two Siphonochalina spp. were also found to contain the same hemiasterlin tripeptides 2, 6, and 17, along with geodiamolides. No evidence for the presence of milnamide A (3) or hemiasterlin B (7) was found in any of the extracts studied. Ireland’s group reported isolating jaspamide, hemiasterlin (2), milnamide A (3), and the new compound milnamide D (18) from specimens of Cymbastela sp. collected in Milne Bay, Papua New Guinea.9 Copyright © 2005 CRC Press, LLC HO O R' O O HN Br HN X HN O O HO N O O N O O R N N H H 5 R' = isopropyl, R = methyl, X=I 4 10 R'=R=Me, X=I 11 R'=R=Me, X=Br 12 R'=R=Me, X=CI 13 R'=Me, R=H, X=I 14 R'=Me, R=H, X=Br 15 R'=Me, R=H, X=CI HO O HN I O CI HN O O O HO N O HN O O N O N O H 19 N 16 H STRUCTURES 4, 5, 10–16, 19. Table 14.1 lists the taxonomic classification of the sponges that have been the source of hemiasterlins and milnamides.10 The source sponges in the genera Auletta and Cymbastela belong to the same family Axinellidae and are, therefore, closely related taxonomically. In contrast, the sponges Hemiasterella minor and Siphonochalina sp. belong to different orders, and each is in a different order than Auletta and Cymbastela. Therefore, the genera Hemiasterella, Siphonochalina, and Cymbastela/Auletta are taxonomically very distant, which indicates a microbial rather than sponge cell origin for the hemiasterlins/milnamides. To date, all occurrences of hemiasterlins/mil- namides have been accompanied by cyclic depsipeptides in the jaspamide/geodiamolide family. There is a report of cultured myxobacteria making chondramides (e.g., chondramide A (19)), which are close analogs of jaspamide.11 This strongly supports a microbial origin for the jaspamide/geo- diamolide cyclic depsipeptides. The absolute fidelity in the cooccurrence of the hemiasterlin/mil- namide and jaspamide/geodiamolide classes of peptides perhaps indicates that there is a single microorganism in each sponge that makes both families. The variation in metabolite content from one source sponge to another might then result either from microbial strain or growth condition differences at each site. III. BIOLOGICAL ACTIVITY OF THE NATURAL PRODUCTS Kashman reported that hemiasterlin (2) showed in vitro cytotoxicity against murine leukemia P388 cells at a concentration of about 19 nM.2 However, Kashman also cautioned that his sample might have contained impurities — namely the cytotoxic depsipeptide geodiamolide TA (5) — so the assays needed to be repeated when additional hemiasterlin became available. Andersen and cowork- ers isolated sufficient quantities of hemiasterlin from Cymbastela to allow the first determination of accurate IC50s for its in vitro cytotoxicity.6 The researchers reported an IC50 of 87 pM against P388, which is over two orders of magnitude lower than Kashman’s value. Potent activity was also observed against a small panel of human cancer cell lines (breast cancer MCF-7 IC50, 169 nM; glioblastoma/astrocytoma U373 IC50, 23 nM; ovarian carcinoma HEY IC50, 3 nM). The tetrapeptide criamides (8 and 9) were found to be roughly 100-fold less potent than hemiasterlin. Crews reported Copyright © 2005 CRC Press, LLC TABLE 14.1 Source Sponges Taxonomy Research Group Collection Site Genus/species Class Order Family Compounds Kashman (1994) Sodwana Bay, Hemiasterella minor Demospongiae Hadromerida Hemiasterellidae Hemiasterlin (2) South Africa Geodiamolide TA (5) Crews (1994) Milne Bay, Auletta cf. constricta Demospongiae Halichondrida Axinellidae Milnamide A (3) Papua New Guinea Jaspamide (4) Andersen (1995) Madang, Papua New Cymbastela sp. Demospongiae Halichondrida Axinellidae Hemiasterlins (2,6,7) Guinea Criamides (8, 9) Geodiamolides (10-16) Boyd (1999) Papua New Guinea Auletta sp. Demospongiae Halichondrida Axinellidae Hemiasterlins (2,6,17) Geodiamolides Boyd (1999) Papua New Guinea Siphonochalina sp. Demospongiae Haplosclerida Callyspongiidae Hemiasterlins (2,6,17) Geodiamolides Ireland (2003) Milne Bay, Cymbastela sp. Demospongiae Halichondrida Axinellidae Hemiasterlin (2) Papua New Guinea Milnamide D (18) Jaspamide (4) Copyright © 2005 CRC Press, LLC that milnamide A (3) also had in vitro cytotoxicity against P388 (IC50, 1.4 µM) and a small panel of human tumor cell lines (IC50s, 5.4–7.9 µM), but its potency was modest compared with hemi- asterlin.3 Ireland found that milnamide D (18) (IC50, 67 nM) was more potent than milnamide A (3) (IC50, 1600 nM) against human colon carcinoma HCT-116 cells in vitro, but that both were less potent than hemiasterlin (2) (IC50, 6.8 nM), confirming that the presence of the β-carboline N terminus in 3 and 18 decreased activity relative to 2.9 Andersen and Kashman isolated the hemiasterlins as cytotoxic agents with potent subnano- molar to nanomolar activity against P388 mouse leukemia and other cell lines.2,6 However, the structure of these compounds and their potent cytotoxic activity gave no clue to their mechanism of action. Roberge and coworkers examined the effects of hemiasterlins on cell cycle progression in human mammary carcinoma MCF-7 cells and found that hemiasterlin, hemiasterlin A, and hemiasterlin B all block cells in mitosis at the same concentrations that showed cytotoxicity.12 Examination of the hemiasterlin mitotic arrest phenotype using immunofluorescence microscopy showed that the cells arrested at a metaphase-like stage, with effects on the morphology of the mitotic spindle that were similar to those caused by tubulin-targeting agents such as vinblastine and nocodazole. High concentrations of hemiasterlin caused complete microtubule depolymeriza- tion. At low hemiasterlin concentrations, cells displayed bipolar spindles with long astral micro- tubules or multiple asters. Chromosomes failed to align at metaphase and were unable to proceed through later stages of mitosis because of activation of the spindle checkpoint. This study con- cluded that hemiasterlins exert their cytotoxic effects by inhibiting spindle microtubule dynamics at mitosis. The study by Anderson12 implied, but did not demonstrate, that hemiasterlin targets microtubules directly. Lassota, at Wyeth, subsequently showed that hemiasterlin inhibits the in vitro assembly of tubulin into microtubules and noncompetitively inhibits vincristine binding to tubulin.13 Researchers at the National Cancer Institute (NCI) then examined in detail the interaction of hemiasterlin with tubulin in vitro. Gamble et al. showed that hemiasterlin, hemiasterlin A, and hemiasterlin C potently inhibited tubulin polymerization in vitro (IC50, 1µM).8 Bai et al.14 confirmed that hemiasterlin inhibits noncompetitively the binding of vinblastine to tubulin and inhibits com- petitively the binding of dolastatatin 10, indicating that hemiasterlin binding causes steric interfer- ence with the binding of vinca alkaloids to tubulin.15 The researchers further demonstrated that hemiasterlin inhibits nucleotide exchange on β-tubulin and that it induces the formation of tubulin aggregates with a ring-like structure and a diameter of about 40 nm. Overall, these effects resembled, but were not identical to, those of dolastatin and cryptophycin, which bind to the same site in tubulin. The tubulin aggregates induced by hemiasterlin in vitro were recently shown to be closed, single-walled, ring-like nanostructures containing 14 α,β−tubulin dimers.16 These rings form at high concentrations and are unstable — they dissociate even at the relatively high hemiasterlin concentration of 100 nM. This is unlike the rings induced by cryptophycin, which are very stable, and those induced by dolastatin, which have an intermediate stability. Because the tubulin rings are induced only at hemiasterlin concentrations about 100-fold above the concentration necessary to induce mitotic arrest and to inhibit cell proliferation, they likely do not contribute to hemiasterlin’s cytotoxicity.16 The first in vivo studies of hemiasterlin were carried out by Theresa Allen at the University of Alberta. As reported by Ireland et al.,17 Allen found high antileukemic activity in an in vivo P388 murine leukemia model (%T/C 308 with five doses at 0.45 µg/mouse). Hemiasterlin also showed activity in an in vivo Gzhi (metastatic murine breast cancer) model: 60% of the mice were long- term survivors with five doses at 1.0 µg/mouse. IV. SYNTHESIS OF HEMIASTERLIN The potent in vitro and in vivo cytotoxic activity of hemiasterlin, its ability to depolymerize microtubules, a validated anticancer target, and its relatively simple but novel tripeptide structure Copyright © 2005 CRC Press, LLC SCHEME 14.1 Synthesis of Boc-tetramethyltryptophan. made it an attractive target for total synthesis and structure–activity relationship (SAR)-driven modification. A reasonably practical synthesis of hemiasterlin and structural analogs was also required to provide sufficient quantities of compounds for preclinical evaluation and, ultimately, clinical trials. Andersen and Piers reported the first total synthesis of hemiasterlin in 1997.18,19 The synthesis employed a convergent approach that involved independent preparation of appro- priately protected forms of the novel N-terminal tetramethyltryptophan and C-terminal 4-amino- 2,5-dimethylhex-2-enoic acid residues, followed by peptide coupling with the commercially available L-tert-leucine. The synthesis of Boc-protected tetramethyltryptophan, starting from indole-3-acetic acid, is outlined in Scheme 14.1. Methylation, first with diazomethane and then with potassium hexame- thylsilazide and methyl iodide, introduced the side chain gem-dimethyl and indole N-methyl substituents present in the desired tetramethyltryptophan residue. Homologation of the permethy- lated indole acetic acid was accomplished in a straightforward fashion to give an intermediate indole-3-propionic acid derivative that contained the carbon skeleton of the final amino acid. Introduction of the α-amino functionality with the correct L configuration was carried out using Evan’s methodology. The Evan’s chiral auxiliary was introduced by first making a mixed anhydride with 2,2-dimethylpropionyl chloride, which was then treated with the lithiated isopropyl oxazoli- done to give an amide. Sequential treatment of the amide with base and 2,4,6-triisopropylbenze- nesulfonyl azide produced an α-azide, which was reduced in situ to an amine, and then reacted with di-tert-butyl dicarbonate to give the Boc-protected α-amine functionality with the correct configuration. The chiral auxiliary was removed with lithium hydroperoxide, and the resulting acid was bismethylated. Hydrolysis of the ester formed in the methylation reaction gave Boc-protected (L)- N,N,β,β-tetramethyltryptophan, ready for peptide coupling. Scheme 14.2 shows the synthesis of Boc-protected 4-amino-2,5-dimethylhex-2-enoic acid ethyl ester. Commercially available (L)-N-Boc-N-methylvaline was first converted into the Weinreb amide, followed by reduction with LiAlH4, to give the corresponding aldehyde. A Wittig reaction on the aldehyde gave the fully protected C-terminal amino acid residue with high stereoselectivity. SCHEME 14.2 Synthesis of Boc-4-amino-2,5-dimethylhex-2-enoic acid ethyl ester. Copyright © 2005 CRC Press, LLC SCHEME 14.3 Andersen synthesis of hemiasterlin.18,19 The reactions involved in coupling the N-Boc-amino acids and deprotection to form the desired tripeptide hemiasterlin (2) are shown in Scheme 14.3. In the discussions that follow, the N-terminal N,N,β, β-tetramethyltryptophan will be designated as the “A” residue, tert-leucine as “B,” and 4-amino-2,5-dimethylhex-2-enoic acid as “C.” Formation of the N-Boc-B-C-OEt dipep- tide was accomplished by removal of the N-Boc group of the fully protected C residue with TFA, followed by coupling with N-Boc-amino acid B as either the mixed anhydride or in the presence of the uronium coupling agent HATU (not shown19). Fully protected hemiasterlin resulted from treatment of the TFA salt of the dipeptide B-C-OEt with the N-Boc protected amino acid A, using PyBOP® as the coupling agent. Deprotected hemiasterlin (2) was obtained by saponification of the ethyl ester, followed by removal of the Boc group with TFA. A second, shorter, and more efficient synthesis of hemisterlin (2) that takes advantage of N- benzothiazol-2-sulfonyl (Bts)–protected amino acid derivatives was subsequently reported by Vedejs in 2001.20 The Vedejs approach to hemiasterlin (2), outlined in Scheme 14.4, parallels the Andersen synthesis, passing through many of the same intermediates. Commercially available (S)-valinol is the starting point for the Vedejs synthesis of the ethyl ester of 4-amino-2,5- dimethylhex-2-enoic acid. Bts-protected (S)-valinol was readily N-methylated under very mild conditions by treatment with iodomethane and potassium carbonate in DMF. Oxidation using either Dess-Martin periodinane or Swern conditions gave the corresponding aldehyde that was subjected to the same ester-stabilized Wittig reaction employed by Andersen. Removal of the Bts protecting group by treatment with PhSH/K2CO3

afforded the C residue ethyl ester ready for peptide coupling. L-tert-Leucine was protected as its Bts amide, and the free carboxylic acid was converted to an acid chloride in preparation for peptide coupling. Biphasic conditions (NaHCO3–Na2CO3, CH2Cl2–H2O) were used to couple the activated L-tert-leucine with the C residue amine. This coupling reaction, involving a N-methyl amine, proceeded in higher yield (86%) than the mixed anhydride coupling (69%) employed by Andersen to make the same peptide bond. Removal of the Bts protecting group by treatment with PhSH/K2CO3 gave the BC dipeptide ethyl ester, ready for coupling to the A residue. Synthesis of a protected N,N,β,β-tetramethyltryptophan A residue started with the trimethylated indole-3-acetaldehyde intermediate from the Andersen synthesis. An asymmetric Strecker sequence Copyright © 2005 CRC Press, LLC SCHEME 14.4 Vedejs synthesis of hemiasterlin.20 using (R)-2-phenylglycinol as the chiral auxiliary and Bu3SnCN as a source of cyanide was employed to introduce the α-amino group with the required stereochemistry. The resulting nitrile was hydrolyzed to a primary amide with H2O2/K2CO . 3 1.5H2O in 10:1 MeOH:DMSO, and the chiral auxiliary was removed by hydrogenolysis to give a primary α-amine. Treatment of the amine with BtsCl followed by methylation with MeI/K2CO3 gave the N-Bts-N-methyl amide, which was converted to the bis-Boc derivative by reaction with Boc2O/DMAP in CH3CN. The bis-Boc amide was reactive enough that with DMAP catalysis it directly formed a peptide bond with the amino group of the BC dipeptide ethyl ester (97% yield). Straightforward removal of the Bts (PhSH/K2CO3) and ethyl ester (LiOH) protecting groups gave hemiasterlin (2). Durst’s group has reported two approaches to a convergent asymmetric synthesis of the N,N,β,β-tetramethyltryptophan A residue in hemiasterlin (2) (Scheme 14.5).21 The key sequence of reactions in the first approach are a SnCl4-mediated ring opening of a cyanoepoxide in the SCHEME 14.5 Syntheses of the Boc-N,N-tetramethyltryptophan residue of hemiasterlin. Copyright © 2005 CRC Press, LLC presence of N-methylindole to give a cyanohydrin, followed by hydrolysis with NaOH to give the trimethylated indole-3-acetaldehyde intermediate from the Andersen and Vedejs syntheses. This intermediate was then converted into the N-Boc methyl ester derivative of N,N,β, β-tetram- ethyltryptophan by a chiral Strecker synthesis similar to that employed by Vedejs. Durst’s second approach involved the use of a glycidic ester in place of the cyanoepoxide to directly generate an α-hydroxy methyl ester (Scheme 14.5). Mitsunobu azidization of the α-hydroxy ester followed by reduction with triphenyphosphine in wet THF gave the methyl ester of N,N,β, β-tetramethyl- tryptophan. Reaction of optically active glycidic ester with N-methyl indole gave the α-hydroxy methyl ester with the same optical purity as the glycidic ester. Durst proposed that further transformation of this compound via Mitsunobu azidization and reduction would lead to enan- tiomerically enriched N,N,β, β-tetramethyltryptophan methyl ester, although this was not verified experimentally. V. SYNTHESIS OF ANALOGS AND SAR The convergent nature of the Andersen synthesis18 of hemiasterlin made it ideally suited for preparing analogs to probe the structural requirements for potent antimitotic activity. Using this synthesis, each of the A, B, and C amino acid residues was altered in turn to prepare a small library of roughly 30 tripeptides that were assessed for biological activity.19 The synthetic modifications focused on the unique structural features of the natural product hemiasterlin. These included the extensive methylation of the N-terminal tryptophan residue — and in particular the β,β-dimethy- lation — the presence of the 4-amino-2,5-dimethylhex-2-enoic acid residue at the C terminus, and occurrence of the rare tert-leucine residue in the center of the tripeptide. At the outset, it was felt that these structural features might play important roles in protecting hemiasterlin from proteolysis and in stabilizing a biologically active conformation. A major objective of the analog synthesis program at the University of British Columbia (UBC) was to identify a compound that was more potent and easier to synthesize than the natural product hemiasterlin (2). The development of a convenient cell-based assay for the quantitative determination of cells arrested in mitosis22 enabled determination of the antimitotic activity of the synthetic hemiasterlin analogs prepared at UBC. A comparison of the IC50 of these compounds for antimitotic activity with their IC50 for cytotoxicity showed a remarkably linear relation (r = 0.958) over a concentration range of almost six orders of magnitude (Figure 14.1).19 This quantitative relation between antim- itotic activity and cytotoxic activity provided strong evidence that the sole mechanism of cytotox- icity of this class of compounds is antimitotic activity (i.e., that the hemiasterlins probably have no secondary cellular targets of relevance to cell proliferation or survival). The synthetic analog study identified structural elements that are critical for potent activity and regions in which single structural changes could be made without loss of potent activity (Figure 14.2). It was found that the tryptophan β-methyl substituents, the N-methyl substituent on the tryptophan α- amino group, and the C26 isopropyl substituent were all extremely important for potent biological activity. In addition, L configurations for the N,N,β,β-tetramethyltryptophan and 4-amino-2,5-dimeth- ylhex-2-enoic acid residues, as well as the presence of the ∆27,28 alkene, were required for maximal activity. Conversely, the C-terminal carboxylic acid could be converted to a methyl ester, the N-25 methyl substituent could be removed, or the tert-leucine residue could be replaced with valine or butyrine as single structural changes without any significant loss in potency. Replacing the N-methylindole fragment of the tetramethyltryptophan residue with either methyl or phenyl groups was also well tolerated. Only one of the synthetic tripeptide analogs tested was more potent than hemiasterlin against MCF-7 mp53 cells. This compound, designated SPA110 (synthetic peptide analog 110), was the phenyl alanine analog of hemiasterlin.19,23 The synthesis of SPA110 was considerably shorter and more efficient than the synthesis of hemiasterlin, primarily because the carbon skeleton of the trimethylphenylalanine N-terminal residue could be prepared in a single, high-yield step from Copyright © 2005 CRC Press, LLC FIGURE 14.1 Plot of in vitro cytotoxicity versus antimitotic activity against human breast cancer MCF-7 cells expressing mutant p53, for natural hemiasterlins, HTI-286, and other synthetic analogues (unlabeled circles) generated by Nieman et al.19 benzene and 3-methyl-2-butenoic acid, as shown in Scheme 14.6. SPA110 was roughly threefold more potent than hemiasterlin in Roberge’s cell-based antimitotic assay (SPA110 IC50, 0.08 nM; hemiasterlin IC50, 0.3 nM). On the basis of its relative ease of synthesis and enhanced potency, Wyeth Pharmaceuticals licensed SPA110 from UBC, and they gave it the code number HTI-286 (Hemiasterlin Tubulin Inhibitor 286). other alkyl OK dimethyl N-methyl required not required olefin required O O N best R N OH H HN O OH or OMe N good isopropyl and N-methyl amino L configuration and L-configuration required required FIGURE 14.2 Structural requirements for optimal cytotoxicity (IC50 < 1 nM) and antimitotic activity (IC50 < 1 nM) of hemiasterlin analogs. O O OH AICI3, benzene OH 94% SCHEME 14.6 Synthesis of trimethylphenylalanine N-terminal residue of HTI-286. Copyright © 2005 CRC Press, LLC VI. BIOLOGICAL ACTIVITY OF HTI-286 Loganzo et al. recently published an in-depth study of the tubulin-binding and antiproliferative properties of HTI-286.24 HTI-286 potently inhibited tubulin polymerization in vitro (about 50% at 0.1 µM and totally at 1 µM), about tenfold more potently than reported by others for hemiasterlin (IC50 = 0.98 µM).8 The effects of HTI-286 on the morphology of mitotic cells were similar to those described for hemiasterlin by Anderson.12 Flow cytometry showed that HTI-286 caused a pro- nounced accumulation of cells in the G2 or M phases of the cell cycle after 24 h, consistent with mitotic arrest as the mechanism of action. After 48 h there was an increase in apoptotic cells. The antiproliferative activity of HTI-286 was evaluated against a panel of 18 tumor cell lines.24 The compound showed broad antiproliferative activity, independent of tumor origin, with an average IC50 of 2.5 ± 2.1 nM. HTI-286 was more potent than paclitaxel (IC50 = 128 ± 369 nM) for all cell lines. Importantly, cells expressing P-glycoprotein, and consequently resistant to many drugs includ- ing paclitaxel, retained nearly complete sensitivity to HTI-286. In addition, the KB-8-5 cell line that has been selected for expression of drug efflux pumps by chronic drug exposure was resistant to paclitaxel, docetaxel, vinblastine, vinorelbine, dolastatin-10, colchicine, and doxorubicin but retained sensitivity to HTI-286. The KB-V1 cell line that expresses very high levels of P-glyco- protein showed resistance to HTI-286 (81-fold), which was much less than the resistance shown to paclitaxel (1400–1800-fold). HTI-286 was selected for further preclinical development over other hemiasterlin analogs in large part because of this ability to inhibit the growth of cell lines expressing moderate to high levels of P-glycoprotein. Researchers at Wyeth25 found that tritium-labeled benzophenone photoaffinity analogs of HTI- 286 specifically and solely photolabel α-tubulin. The binding site maps to sheet 8- helix 10 region of α-tubulin, a region of lateral and longitudinal contacts with other tubulin subunits. In contrast, the binding sites for colchicine are at the α–β-tubulin interface, and the binding site for paclitaxel is within β-tubulin. This is the first evidence that a tubulin-binding peptide interacts with α-tubulin and is consistent with the observation that mutations in α-tubulin have been reported in HTI-286- resistant cells. VII. PRECLINICAL AND CLINICAL DEVELOPMENT OF HTI-286 The initial encouraging in vivo efficacy of hemiasterlin in the mouse leukemia P388 model reported by Coleman et al.,6 combined with the availability of HTI-286, a potent, water-soluble, and more readily synthesized analog, prompted Wyeth to examine the ability of HTI-286 to inhibit tumor growth in a variety of mouse models.24 HTI-286 administered intravenously has in vivo efficacy against a variety of xenograft tumor models, including LOX-IMV melanoma, KB-3-1 epidermoid carcinoma, PC3-MM2 prostate carcinoma, SW620 colon carcinoma, U87 glioma, MCF-7 breast carcinoma, and LOVO colon carcinoma. Moreover, HTI-286 was able to inhibit the growth of even large established tumors derived from LOX-IMV, KB-3-1, LOVO, and PC3MM2. HTI-286 admin- istered for 4 days inhibited the growth of 2.5 g LOX-IMV tumors, and subsequent dosing caused 93% tumor regression. In keeping with results observed with cell lines in culture, HTI-286 also displayed efficacy against tumors derived from cell lines with inherent or acquired multidrug resistance; KB-8-5 tumors, MX-1W human breast carcinoma, DLD-1 human colon carcinoma, and HCT-15 human colon carcinoma all overexpressed P-glycoprotein and showed resistance to paclitaxel or vincristine but were sensitive to HTI-286.24 Interestingly, ovarian carcinoma cells resistant to paclitaxel or epothilone, and an A549 lung carcinoma line selected for epothilone resistance because of mutations in the α-tubulin gene, also retained sensitivity to HTI-286.24 Wyeth also examined the metabolism of HTI-286 in vitro in liver microsomes from a variety of animal species, and in vivo in rats and dogs. HTI-286 was metabolized to a major biologically inactive N-demethylated species.26 Wyeth’s preclinical studies also showed that HTI-286 has efficacy Copyright © 2005 CRC Press, LLC when administered orally.24 These favorable preclinical parameters, combined with the relative ease of synthesis and water solubility of HTI-286, indicated that HTI-286 may have clinical utility in patients, particularly those with intrinsic or acquired resistance to paclitaxel. A phase I clinical trial was carried out in 2002 with patients with metastatic or advanced-stage malignant solid tumors to assess the safety, tolerability, and pharmacokinetics of HTI-286. HTI- 286 dissolved in saline was administered intravenously over 30 min once every 21 d at doses ranging from 0.06 to 2.0 mg/m2. This investigation has been published in abstract form by Ratain et al.27 Pain, hypertension, and neutropenia were observed as dose-limiting toxicities. Preliminary pharmacokinetic data indicated considerable interindividual variability in clearance, steady-state volume of distribution, and half-life. Serum concentrations above 0.5 ng/mL were associated with neutropenia. The study recommended a phase II dose of 1.5 mg/m2 and additional studies to determine the basis for the marked interindividual variability in clearance. An additional phase I dose escalation study of HTI-286 administered in combination with carboplatin is underway. A phase II open-label study of HTI-286 as a single agent for the treatment of non–small cell lung cancer for disease reoccurrence following platinum-based therapy is also currently underway.1 VIII. SUMMARY Natural products continue to be an important source of lead structures for the development of new experimental drugs for the treatment of cancer.28 The development path leading to HTI-286 illus- trates some of the challenges and opportunities encountered in bringing a marine natural prod- uct–derived drug to clinical trials.29 Marine sponges, which are the most primitive metazoans in the oceans, are the richest source of

new marine natural product structures.29,30 Secondary metabolites isolated from sponge tissues can be either true metabolites of the sponge’s own enzymes or metabolites of microorganisms living in association with the sponge. Large-scale wild harvest of the source sponge is environmentally unacceptable, and the problems associated with sponge aquaculture and with laboratory culture of sponge-associated microorganisms make it difficult to routinely use nature’s biosynthetic machinery to make sponge-derived compounds in bulk. Therefore, the major practical source of sponge secondary metabolites for drug development at the moment is via laboratory synthesis. As a consequence, the current goal of examining sponge-derived natural products as part of drug dis- covery programs is primarily to find “chemical inspiration” about new pharmacophores. Research aimed at overcoming the limitations of culturing sponges and their associated microorganisms, partly involving genetic approaches to cloning entire biosynthetic pathways, is ongoing, so this situation might well change in the near future. As suggested above, hemiasterlin, which was the chemical inspiration for the HTI-286 antimitotic pharmacophore, is presumably a microbial sec- ondary metabolite. Several attempts by the Andersen group at UBC to isolate and culture a hemiasterlin-producing microorganism from Cymbastela tissues have all failed. A number of properties of hemiasterlin made it stand out as an extremely exciting natural product lead structure for an anticancer drug development program. The first things that attracted attention were its exceptionally potent in vitro and in vivo cytotoxicities against P388 and its relatively simple tripeptide structure. Identification of tubulin as its molecular target elevated the interest in the compound and really provided the impetus for initiating the synthetic program. It is interesting to speculate whether a tripeptide antimitotic agent such as HTI-286 would have ever emerged from a combinatorial chemistry screening program. Some of the first combinatorial libraries comprised small peptides made from protein amino acids, and a Trp-Val-Val tripeptide would have been a logical compound to make as part of a small peptide library. However, the SAR studies on hemiasterlin showed that such an unmodified tripeptide would not have shown significant activity in a cytotoxin or cell-based antimitotic screen. Nature revealed through the natural product hemiasterlin the additional requirements of extensive β- and N-methylation, along with modification Copyright © 2005 CRC Press, LLC of the C-terminal valine residue, that are necessary to create potent antimitotic activity and likely proteolytic stability starting from this simple tripeptide template. These combined modifications are unlikely to have been part of a combinatorial synthesis program before the chemical inspiration provided by the discovery of the natural product hemiasterlin. Lack of a reliable industrial-scale source of the natural compound or bioactive synthetic analogs is often a major impediment to developing a natural product lead into a drug. The relatively easy synthetic accessibility to hemiasterlin and analogs played an important role in moving the devel- opment of this class of antimitotic peptides forward. Synthesis provided hemiasterlin and selected analogs in sufficient quantities for more detailed preclinical in vitro and in vivo evaluation and guaranteed an unlimited supply of compound for human trials and eventual clinical use. It is important to recognize that the natural product hemiasterlin already had very potent in vitro and in vivo activity against human tumor cell lines, so nature had done most of the work of optimizing the SAR for this pharmacophore. Analog synthesis did identify HTI-286, which shows an incre- mental improvement in activity profile, but all other analogs that have been made and tested at UBC showed a loss of activity relative to the natural product.19 Two important milestones in the preclinical decision process to move ahead with HTI-286 were its superior in vitro cytotoxicity profile versus multidrug-resistant human cancer cells lines relative to the natural product hemiasterlin, and the fact that it retained activity against paclitaxel and vinca alkaloid-resistant cell lines in mouse tumor xenograft models. Other desirable attributes were its chemical stability and solubility in saline, making formulation straightforward, and its oral activity in mouse models. The discovery that it interacts with α-tubulin distinguished it mechanistically from other tubulin depolymerization agents and raised its development profile. HTI-286 appears to be a very promising experimental anticancer drug. It represents a new “natural product–inspired” chemical structural class of antimitotic agents that target tubulin and inhibit its polymerization into microtubules. HTI-286 shows excellent intravenous and oral activity against drug-sensitive and multidrug-resistant human tumors in mouse xenograft models, both preventing the growth of newly implanted tumors and also regressing well-established tumors. The compound has successfully completed a phase I clinical trial, which established a maximum tolerated dose in humans and identified the dose limiting toxicities. Phase II clinical trials with HTI-286 are ongoing, and as with all experimental drugs, there is the hope that the high level of efficacy observed in preclinical animal models will translate into real benefit for human cancer patients. ACKNOWLEDGMENT Financial support for the research at UBC was provided by the National Cancer Institute of Canada (MR, RJA), the U.S. National Institutes of Health (RJA), and the Natural Sciences and Engineering Research Council of Canada (RJA, EP). We thank Theresa Allen, Ed Piers, Hilary Anderson, John Coleman, Mike LeBlanc, Dilip de Silva, Debra Wallace, Lynette Lim, and Jim Nieman, who all played important roles in the research leading to HTI-286 (SPA-110) at UBC. Wyeth Pharmaceuticals has been a wonderful partner in taking HTI-286 from a basic laboratory discovery to clinical trials. Many individuals at Wyeth, too numerous to mention, made major contributions to the basic science of HTI-286 and its preclinical development. Several that deserve special mention are Phil Frost, Lee Greenberger, Frank Loganzo, Carolyn Discafani, Arie Zask, and Semiramis Ayral-Kaloustian. REFERENCES 1. Margaret and Charles Juravinski Cancer Centre. Cancer Services. Available at: http://www. hrcc.on.ca/ActiveTrials.asp. 2. Talpir, R., et al., Hemiasterlin and geodiamolide TA: two new cytotoxic peptides from the marine sponge Hemiasterella minor (Kirkpatrick), Tetrahedron Lett., 35, 4453, 1994. Copyright © 2005 CRC Press, LLC 3. Crews, P., et al., A highly methylated cytotoxic tripeptide from the marine sponge Auletta cf constricta, J. Org. Chem., 59, 2932, 1994. 4. Matsunaga, S., Fusetani, N., and Konosu, S., Bioactive metabolites VI. Structure elucidation of discodermin A, an antimicrobial peptide from the marine sponge Discodermia kiiensis, Tetrahedron Lett., 25, 5165, 1984. 5. Kaneda, M., Studies on the Bottromycins. I. proton and carbon-13 NMR assignments of bottromycin A2, the main component of the complex, J. Antibiot., 45, 792, 1992. 6. Coleman, J. E., et al., Cytotoxic peptides from the marine sponge Cymbastela sp, Tetrahedron, 51, 10653, 1995. 7. Coleman, J. E., et al., Hemiasterlin methyl ester, Acta Crystal., 52, 1525, 1996. 8. Gamble, W. R., et al., Cytotoxic and tubulin-interactive hemiasterlins from Auletta sp. and Siphonoch- alina spp. sponges, Bioorg. Med. Chem., 7, 1611, 1999. 9. Chevallier, C., et al., A new cytotoxic and tubulin-interactive milnamide derivative from a marine sponge Cymbastela sp., Org. Lett., 5, 3737, 2003. 10. Hooper, J. N. A. and van Soest, R. W. M., Systema Porifera, a Guide to the Classification of Sponges, Kluwer Academic/Plenum, New York, 2002. 11. Kunze, B., et al., Chondramines A-D, new antifungal and cytostatic depsipeptides from Chondromyces crocatus (Myxobacteria). Production, physico-chemical and biological properties, J. Antibiot., 48, 1262, 1995. 12. Anderson, H. J., et al., Cytotoxic peptides hemiasterlin, hemiasterlin A and hemiasterlin B induce mitotic arrest and abnormal spindle formation, Cancer Chemother. Pharmacol., 39, 223, 1997. 13. Personal communication from Peter Lassota. 14. Bai, R., et al., Interactions of the sponge-derived antimitotic tripeptide heniasterlin with tubulin: comparison with dolastatin 10 and cryptophycin 1, Biochemistry, 38, 14302, 1999. 15. Hamel, E., Interactions of antimitotic peptides and depsipeptides with tubulin, Biopolymers, 66, 142, 2002. 16. Boukari, H., Nossal, R., and Sackett, D. L., Stability of drug-induced tubulin rings by fluorescence correlation spectroscopy, Biochemistry, 42, 1292, 2003. 17. Ireland, C. M., et al., Anticancer agents from unique natural products sources, Pharmaceutical Biol., in press, 2004. 18. Andersen, R. J., et al., Total synthesis of (-)-hemiasterlin, a structurally novel tripeptide that exhibits potent cytotoxic activity, Tetrahedron Lett., 38, 317, 1997. 19. Nieman, J., et al., Synthesis and antimitotic/cytotoxic activity of hemiasterlin analogs, J. Nat. Prod., 66, 183, 2003. 20. Vedejs, E. and Kongkittingam, C., A total synthesis of (-)-hemiasterlin using N-Bts methodology, J. Org. Chem., 66, 7355, 2001. 21. Reddy, R., et al., Asymmetric synthesis of the highly methylated tryptophan portion of the hemiasterlin tripeptides, Org. Lett., 4, 695, 2002. 22. Roberge, M., et al., Cell-based screen for antimitotic agents and identification of analogues of rhizoxin, eleutherobin and paclitaxel in natural extracts, Cancer Res., 60, 5052, 2000. 23. Andersen, R. J., et al. WO Patent 99/32509, 1999. 24. Loganzo, F., et al., HTI-286, a synthetic analog of the tripeptide hemiasterlin, is a potent antimicro- tubule agent that circumvents P-glycoprotein-mediated resistance in vitro and in vivo, Cancer Res., 63, 1838, 2003. 25. Nunes, M., et al., Two photoaffinity analogs of HTI-286, a synthetic analog of hemiasterlin, interact with alpha-tubulin, Eur. J. Cancer, 38, S119, 2002. 26. Wang, C. P., et al., In vitro and in vivo metabolism of HTI-286, an antimicrotubule anticancer agent, Drug Metabol. Rev., 35 Suppl. 2, 42, 2003. 27. Ratain, M. J., et al., Phase 1 and pharmacological study of HTI-286, a novel antimicrotubule agent: correlation of neutropenia with time above a threshold serum concentration, Proc. Am. Soc. Clin. Oncol., 22, 129, 2003. 28. Newman, D. J., Cragg, G. M., and Snader, K. M., Natural products as sources of new drugs over the period 1981–2002, J. Nat. Prod., 66, 1022, 2003. 29. Andersen, R. J. and Williams, D. E., in Chemistry in the Marine Environment, Vol. 13, R. E. Hester and R. M. Harrison, Eds., The Royal Society of Chemistry, Cambridge, 2000, 55. 30. Blunt, J., et al., Marine natural products, Nat. Prod. Rep., 20, 1, 2003. Copyright © 2005 CRC Press, LLC 15 The Actinomycins Anthony B. Mauger and Helmut Lackner CONTENTS I. Historical Introduction II. Separation and Nomenclature III. Structures IV. Conformation V. Mechanism of Biological Action VI. In Vitro Antitumor Activity VII. In Vivo Antitumor Activity and Toxicity VIII. Synthesis IX. Analogs A. Directed Biosynthesis B. Partial Synthesis C. Total Synthesis X. Structure–Activity Relationships A. Introduction B. Natural Actinomycin Variants C. Chromophoric Analogs D. Peptidic Analogs XI. Clinical Applications of Actinomycins References I. HISTORICAL INTRODUCTION The actinomycins, a family of structurally related chromopeptide antibiotics with a common phenoxazinone chromophore attached to two pentapeptide lactone moieties (Figure 15.1), vary in their amino acid content (Table 15.1). They emerged from the pioneering work of Selman Waksman on soil microorganisms, and in particular from his work on the Streptomyces. The first actinomycin, isolated in 1940 from Streptomyces antibioticus,1,2 was the first of several antibiotics discovered by Waksman, the first crystalline antibiotic, and the first to display antitumor activity. Many other actinomycins have since been isolated from other Streptomyces species, and it was also found in an unrelated genus, Micromonospora.3 The first actinomycin was described as a red crystalline substance active against Gram-positive microorganisms,2 but its toxicity4 precluded its clinical use as an antibiotic. Interest in the actino- mycins revived in 1952 when the antitumor activity of another actinomycin (C) in mouse and rat tumors was found by Hackmann5 and when Schulte reported the first clinical studies.6 The discovery that actinomycin inhibits DNA-primed RNA synthesis7,8 resulted in its widespread use in studies of macromolecular biosynthesis and virus replication.9 Copyright © 2005 CRC Press, LLC FIGURE 15.1 Structures of 18 naturally occurring actinomycins (see Table 15.1). II. SEPARATION AND NOMENCLATURE With the first actinomycin preparation designated A, subsequent isolates from different species or strains of Streptomyces were termed B,10 C,11 D,12 I,13 X,14 Z,15 and so forth. That they were mixtures was first shown when actinomycin C separated into components C1, C2, and C3 in countercurrent distribution.16,17 Paper partition chromatographic studies confirmed that most other actinomycin preparations were also multicomponent complexes.16,18–21 Techniques for their sepa- ration have been reviewed.22 TABLE 15.1 Structures of Natural Actinomycins (See Figure 15.1) Actinomycin A B C D E F I=X0β Thr D-Val D-Val Pro Hyp MeVal X0δ Thr D-Val D-Val Pro aHyp MeVal II=F8 Thr D-Val D-Val Sar Sar MeVal III*=F9 Thr D-Val D-Val Pro Sar MeVal IV=X1=C1=D Thr D-Val D-Val Pro Pro MeVal V=X2 Thr D-Val D-Val Pro 4-oxo-Pro MeVal X1a Thr D-Val D-Val Sar 4-oxo-Pro MeVal X0γ Thr D-Val D-Val Sar Pro MeVal VI=C2 Thr D-Val D-aIle Pro Pro MeVal i-C2 Thr D-aIle D-Val Pro Pro MeVal VII=C3 Thr D-aIle D-aIle

Pro Pro MeVal Z1 4-OH-Thr D-Val D-Val 3-OH-5-MePro 4-oxo-5-MePro MeAla Z2 Thr D-Val D-Val 3-OH-5-MePro 4-oxo-5-MePro MeAla Z3 4-Cl-Thr D-Val D-Val 3-OH-5-MePro 4-oxo-5-MePro MeAla Z4 Thr D-Val D-Val 5-MePro 4-oxo-5-MePro MeAla Z5 4-Cl-Thr D-Val D-Val 5-MePro 4-oxo-5-MePro MeAla ZP Thr D-Val D-Val 5-MePro 5-MePro MeVal G† 4-OH-Thr D-Val D-Val Pro 3-OH-5-MePro MeAla Note: Nonstandard amino acid abbreviations: MeVal = N-methylvaline; MeAla = N-methylala- nine; 5-MePro = 5-methylproline; all amino acids are L except where prefixed by “D”. * Actinomycin III is a mixture of two isomers with Pro and Sar interchangeable at sites D and E. † Actinomycin G refers to the product from Streptomyces HKI-0155. Copyright © 2005 CRC Press, LLC This situation was simplified when it was found that different complexes had components in common. Thus, actinomycins A, B, and X contained the same five compounds, albeit in different proportions, and a system of nomenclature was proposed23 that named them actinomycins I through V in the order of increasing lipophilicity. The actinomycin C components were named IV, VI, and VII, of which only IV was also present in the A, B, and X complexes. Actinomycin D was found to consist solely of IV. The five components of the actinomycin Z complex lie outside of this system and are termed Z1 through Z5. III. STRUCTURES The structures of the best-characterized actinomycins are shown in Figure 15.1 and Table 15.1. The elucidation of the structure of actinomycin C3 (VII) in 1956 by Brockmann et al. was a notable achievement in the era before nuclear magnetic resonance and mass spectrometry were available.24 The structure comprises a chromophoric 2-aminophenoxazin-3–one ring system bearing two methyl groups and two identical pentapeptide lactone moieties. The chromophore, including the two carboxyls, is termed actinocin, so an actinomycin can be described as an actinocinyl-bis(pentapep- tide lactone). The structure of actinomycin D (IV), which has been abbreviated “AMD,” was found to be identical to that of C3, except that both D-allo-Ile residues are replaced by D-Val.25 As distinct from iso-actinomycins such as D (IV) and C3 (VII), the aniso-actinomycin C2 (VI)26 has a minor regioisomer i-C2 (originally termed C2a).27 The two isomers were distinguished via oxidative degradation of the chromophore, permitting separation of the two peptides.28 The structures of actinomycins I and V emerged from studies on the reduction of actinomycin X2 in various ways to generate X0β (I), X0δ, and X1 (IV).29 The oxygenated proline residues were located in the β-peptide by oxidative degradation of the chromophore. These and other experiments indicated that X2 contains 4-oxoproline, which is usually destroyed during hydrolysis, and this was confirmed30 by its isolation. Minor components of the X complex termed X0γ and X1a were also described31 (Table 15.1). Actinomycins II and III are trace components of the A, B, and X complexes and were characterized after addition of sarcosine to the culture medium for S. antibioticus markedly increased their relative concentrations.32 The added sarcosine acted as a biosynthetic precursor, replacing one or both proline residues.33 That this involved no change in amino acid sequence was also confirmed.34 Chromatog- raphy revealed that the aniso-actinomycin III was a mixture of major and minor regioisomers.35 The actinomycin Z complex contains at least five components comprising a far greater diversity of amino acid content than actinomycins I–VII. Early studies15 revealed that they differ from the other actinomycins in that they all contain N-methylalanine, but no proline. Instead, there were several unusual proline congeners found in these actinomycins, including cis-5-MePro,36,37 4-oxo- 5-MePro,38,39 and trans-3-hydroxy-cis-5-MePro.40,41 Also, 4-hydroxy-Thr42 was found in Z1. More recently, two-dimensional nuclear magnetic resonance and mass spectrometry techniques permitted the elucidation of the structures of actinomycins Z1 through Z5 and unexpectedly revealed that Z3 and Z5 contained 4-chloro-Thr.43 In addition, actinomycin G from S. HKI-0155 has been found to have an α-peptide, which is similar to that of AMD, and a Z-type β-peptide.44 Another actinomycin named ZP has 5-MePro in both peptides but otherwise is identical to AMD.45 Numerous other actinomycins have been described, but they are less well characterized struc- turally than those included in Table 15.1. Information on them is available in several reviews.46–49 IV. CONFORMATION The first information on the conformation of an actinomycin (D) emerged from an x-ray crystallo- graphic study of its complex with deoxyguanosine (Figure 15.2).50 Later, x-ray studies of uncomplexed Copyright © 2005 CRC Press, LLC FIGURE 15.2 High-resolution x-ray crystallographic structure of actinomycin Z3, shown from the chro- mophoric side of the molecule with the β-peptide ring on the right.52 Abbreviations: MOPro = cis-5-methyl- 4-oxo-L-proline; HMPro = trans-3-hydroxy-cis-5-methyl-L-proline (Table 15.1). actinomycins D,51,52 X2,53 and Z3 (Figure 15.3)52 indicated that their conformations were similar except for an orientation change in the ester linkages. The molecule possesses pseudo-C2 symmetry, and the two peptide moieties are held in unique juxtaposition by two antiparallel hydrogen bonds between the D-Val NH in one peptide and the D-Val C=O in the other, and vice versa. The D-Val-Pro and Pro-Sar peptide bonds are cis. Nuclear magnetic resonance studies of actinomycins54 and of actino- mycin-DNA complexes55 indicate an essentially similar conformation in solution that is remarkably stable in solvents of different polarity. V. MECHANISM OF BIOLOGICAL ACTION The biological activity of the actinomycins is exerted by inhibition of DNA-dependent RNA synthesis,7,8 which thereby inhibits protein synthesis.8 This effect is explained by the strong binding of actinomycin to double-helical DNA.56,57 This complexation has usually been detected by absorption spectroscopy (bathochromic shift of about 20 nm in the visible maximum of an actinomycin normally at 440 nm)7,58,59 or by elevation in the thermal denaturation temperature (Tm) of the DNA.60,61 It is required that the DNA be helical62,63 and that it contains guanine residues.62,64,65 An intercalative model for the actinomycin-DNA complex was proposed66 in which the actinomycin chromophore is inserted adjacent to a G–C pair, and this hypothesis was supported by several solution studies involving equilibrium, kinetic and hydrodynamic techniques, and sedimentation coefficients.67 Interaction of deoxyguanosine with AMD in solution produces spectral shifts similar to those with DNA.68 It was therefore of much interest when the 2:1 complex was obtained in crystalline form for x-ray studies.50 The structure that emerged (Figure 15.2) has approximate two-fold symmetry, with the AMD chromophore “sandwiched” between the two guanine ring systems. The complex is held together by π-complex interactions between the three ring systems, by Copyright © 2005 CRC Press, LLC Sar MeVal Pro Thr D-Val FIGURE 15.3 X-ray crystallographic structure of the 1:2 complex of actinomycin D with deoxyguanosine.50 The actinomycin chromophore and the deoxyguanosine molecules are shown in bold. Hydrogen bonds are shown as dotted lines. Reprinted from Journal of Molecular Biology, Vol 61, Jain and Sobell, pp.1, 1972, with permission from Elsevier. hydrogen bonds between the guanine 2-amino groups and the threonine carbonyl oxygens, and by hydrophobic interactions. This structure provided a model for the geometry of the AMD–DNA interaction, and this model is supported by x-ray structures of DNA complexes of AMD with d(GC)69 and d(ATGCAT).70 The hydrophobic interactions that enhance the binding of actinomycins in the narrow groove of DNA depend on the hydrophobic inner surface of the peptide units, and the nature of the outer surfaces, when the binding occurs near the pause or rho-dependent termination sites of the DNA, terminates transcription by RNA polymerase.71 Also, actinomycins belong to a class of antineo- plastic agents that inhibit both topoisomerase I and topoisomerase II. It forms a ternary complex of topoisomerase II, inducing DNA strand breaks,72 and it also stimulates topoisomerase I–mediated DNA cleavage.73 Studies have also been reported on the binding of actinomycin D to single-stranded DNA.74 Recently, evidence has emerged of the capability of actinomycin to down-regulate the effect of the transcription factors TBP and Sp1 by blocking access to their specific binding sites.75 VI. IN VITRO ANTITUMOR ACTIVITY Early studies (1958) showed that AMD displays equal growth inhibitory activity against both normal and malignant cells.76,77 The cytotoxic effects of several actinomycins were compared in HeLa cells.78 More recently, AMD and several other actinomycins have been evaluated in the 60 human tumor cell line screening system79 of the National Cancer Institute (NCI). For AMD, the mean molar log GI50 (50% growth inhibition), averaged over many experiments, was –8.73. The corre- sponding numbers for the various panels of cell lines (mean values from several lines) were leukemia, –9.2; non–small cell lung, –8.6; small cell lung, –8.8; colon, –8.7; CNS, –9.0; melanoma, –8.9; ovarian, –8.4; renal, –8.5; prostate, –8.6; and breast, –8.7. Thus, the most sensitive cell lines were those of leukemia, CNS cancers, and melanoma. In the COMPARE program,80 AMD most closely resembles other topoisomerase II inhibitors. Copyright © 2005 CRC Press, LLC MVH HVM MV VM Z OBzl Z OH HO Z Z Z Y Y Y Y Y X (i) X X (ii) X X T (O) T T (O) (O) T T (O) CO CO CO CO CO NO N NH 2 2 N NH2 OBzl O O O O T = Threonine X = D-allo-Isoleucine Y = Proline Z = Sarcosine MV = Methylvaline FIGURE 15.4 First total synthesis of an actinomycin (C3 = VII).89 Reagents: (i) Cbz-MeVal-OH/carbonyldi- imidazole; H2/Pd, then K3Fe(CN)6 ; (ii) ClCOOEt/Bu3N. VII. IN VIVO ANTITUMOR ACTIVITY AND TOXICITY The first observations of antitumor activity of an actinomycin (C) involved the suppression of the Ehrlich carcinoma in mice and the Wilms tumor in rats.5,81 These were quickly followed by numerous studies in various laboratories; the early work has been reviewed.82 Among the best responders were the Ridgway osteogenic carcinoma in mice83 and the Wilms tumor in rats,84 especially when combined with radiation.85 Other experimental tumors susceptible to actinomycins include the P388 and L1210 leukemias, B16 melanoma, and adenocarcinoma 755. 86 Actinomycins C1 (AMD), C2, and C3 were equally efficacious. Actinomycins are highly toxic4; the intravenous LD50s of AMD in mice and rats are 1.2 and 0.6 mg/kg per day, respectively, and the corresponding subcutaneous figures are 1.4 and 0.80.87 In common with many other antitumor agents, the toxic effects manifest primarily in rapidly prolif- erating tissues such as the bone marrow, intestinal mucosa, and lymphoid organs. VIII. SYNTHESIS The first synthesis of an actinomycin (C3) involved the oxidative coupling of two molecules of a 3- hydroxy-4-methylanthraniloyl O-pentapeptide, followed by cyclization of both peptide moieties at the Sar-MeVal peptide bonds88,89 (Figure 15.4). Interestingly, this route failed to distinguish between the accepted structure for actinomycins and an alternate version that had been proposed,90,91 in which the C terminus of one peptide is lactonized by the Thr OH of the other, and vice versa, forming a cyclode- capeptide dilactone. Subsequently, several unambiguous syntheses of actinomycins have been described involving a pentapeptide lactone derivative as the key intermediate. In the first example92 (Figure 15.5), the 3-benzyloxy-4-methylanthraniloyl-pentapeptides were lactonized in syntheses of actinomycins C1, C2, i-C2, and C3. Analogous peptide lactones were subsequently synthesized in higher yields via cycliza- tion at the Pro-Sar93 or D-Val-Pro94 peptide bonds. This intermediate can also be constructed from the MeVal-OH MeVal MeVal MeVal Sar Sar Sar Sar Pro Pro Pro Pro A (i) A (ii) A A Thr Thr (O) (O) Thr Thr (O) CO CO CO CO NO NO N NH 2 2 2 OBzl OBzl O O FIGURE 15.5 Synthesis of actinomycins C1, C2, i-C2 and C3.92 A = D-Val or D-aIle. Reagents: (i) AcCl/imi- dazole; (ii) H2/Pd, then K3Fe(CN)6. Copyright © 2005 CRC Press, LLC cyclic pentapeptide by acid-catalyzed N,O-acyl shift followed by acylation of the Thr amino group.95 Regioselective syntheses of actinomycins C2 and i-C2 have been reported,96,97 as well as that of X0β.53 IX. ANALOGS A. DIRECTED BIOSYNTHESIS The amino acid content of actinomycins can be manipulated by the addition of certain amino acids to the medium in which the producing organism is cultured, a method also known as “controlled biosynthesis.”98 Addition of isoleucine to cultures of Streptomyces chrysomallus generated new actinomycins, termed E1 and E2,99 containing one and two N-Me-allo-Ile residues (respectively) in place of N-MeVal.100 Similar studies with Streptomyces antibioticus and Streptomyces parvulus uncovered a more complex situation, in which D-Ile and D-allo-Ile were both able to replace D- Val.101 Addition of sarcosine to S. chrysomallus cultures produced a mixture in which one or both prolines of C2 and C3 were replaced by sarcosine, and the products were termed (respectively) F3 and F1 (analogs of C2) and F4 and F2 (analogs of C3).99,102 One or both of the

proline residues in actinomycins can be replaced by several Pro analogs using directed biosynthesis. They include the cis and trans isomers of 4-methyl-,103 4-chloro-, and 4-bromo-prolines35; thiazolidine-4-carboxylic acid104; azetidine-2-carboxylic acid; and pipecolic acid.105 When one Pro is replaced, the product is a mixture of two difficultly separable regioisomers. In the case of the 4-substituted prolines, only the four analogs formed using cis- and trans-4-MePro were characterized.103 Following the early observation105 that new actinomycins were formed by S. antibioticus in the presence of azetidine-2-carboxylic acid, two analogs, named azetomycins I and II, were isolated,106 with one or both prolines (respectively) replaced by azetidine-2-carboxylic acid. When pipecolic acid is added to S. antibioticus cultures, not only does it replace Pro but actino- mycins containing 4-hydroxy- and 4-oxo-pipecolic acid are generated as well107 (Table 15.2). This is analogous to the presence of Hyp and 4-oxo-Pro (respectively) in the natural actinomycins I and V. B. PARTIAL SYNTHESIS Mild acid hydrolysis of actinomycins replaces the chromophoric NH2 by OH, and subsequent treatment with thionyl chloride produces 2-chloro-2-deaminoactinomycin.108 Treatment of this com- pound with a variety of primary and secondary amines provided 34 2-N-substituted actinomy- cins.109–113 In addition to its hydrogenation to 2-deaminoactinomycin,108 the 2-chloro derivative is also an intermediate in the preparation of analogs substituted with Cl and Br in the 7-position.114 Other groups substituted in the 7-position include nitro,109,115 amino,110,115 hydroxy,110 and methoxy.116. TABLE 15.2 Biosynthetic Actinomycin Analogsa from Streptomyces antibioticus with Pipecolic Acid105 Actinomycin Proline Pipecolic Acid 4-OH-Pipecolic Acid 4-Oxo-Pipecolic Acid Pip 1α 0 1 0 1 Pip 1β 1 1 0 0 Pip 1γ 0 1 1 0 Pip 1δ 1 0 0 1 Pip 1ε 1 0 1 0 Pip 2 0 2 0 0 a The number of residues of each amino acid at the two 3-sites are shown; the remainder of each structure is similar to that of actinomycin D. Copyright © 2005 CRC Press, LLC Peptide Lactones Peptide Lactones Peptide Lactones CO CO CO CO H H H CO CO N N R N N N N O O O O2N O O O O O O Peptide Lactones Peptide Lactones CO CO H CO CO N N R' N N R' O O O O O Peptide Lactones Peptide Lactones CO CO H CO CO O2N N N R" H2N N N R" HO O O O O O FIGURE 15.6 Tetracyclic analogs of actinomycin D. R = H, CH3, C2H5, CH2F, n-C6H13, CH2COOH, CH2CH2COOH, C6H5, C6F5, o-ClC6H4, m-ClC6H4, p-ClC6H4, 2,4-Cl2C6H3, 2-naphthyl; R′ = CH3, C6H5, C6F5; R″ = CH3, C6H5, CH2COOH, CH2CONH(CH2)4NH2. N-Acetyl, N-pivaloyl, and N-stearoyl derivatives of 7-amino-AMD have been described,67 as well as nine N-alkyl and N-aryl derivatives.117 7-Hydroxy-8-amino-AMD118 and nine O-alkyl and four O-acyl derivatives of 7-hydroxy-AMD have also been reported.113 Hydrogenation of actinomycin produces dihydroactinomycin, which, unlike actinomycin, can be acetylated. Subsequent reoxidation furnishes N-acetyl-actinomycin.119 Dihydroactinomycin reacts with α-ketoacids to generate tetracyclic analogs with a fused oxazine ring.109,120,121 Reaction of AMD with aldehydes leads to tetracyclic analogs with a fused oxazole ring, and further manip- ulations produced related compounds substituted in the 7 and 8 positions.118,121,122 Some of these tetracyclic analogs are shown in Figure 15.6. The lactone rings of actinomycins were opened with alkali to produce actinomycinic acids123 from which dimethyl esters and di-O-acetyl derivatives have also been prepared.119 One or both lactone rings can also be opened by microbial degradation by Actinoplanes species.124 The peptide lactones of actinomycin C3 have been cleaved at the Sar-MeVal peptide bond with cold, concentrated HCl to produce the bis-seco-actinomycin.125 The OH groups of Hyp and allo-Hyp in actinomycins X0β and X0δ have been converted to their O-acetyl and O-hexadecanoyl derivatives.29,126 Likewise, an O-acetate and di-O-acetate have been prepared from actinomycin Z1.39 The 4-oxo-5-MePro residue in this actinomycin has been reduced to two diastereoisomers of 4-OH-5-MePro, and a triacetate of the resulting actinomycins was prepared.127 The 4-oxo-Pro in actinomycin X1α has been reduced to Hyp, of which the O-acetyl derivative has also been produced.31 Actinocinyl-gramicidin S has been described128; the two couplings were effected at the ornithine δ-NH2 groups. C. TOTAL SYNTHESIS The two chromophoric methyl groups of AMD have been replaced by H, Br, OMe, Et, and t-Bu.129 Also, the 6-Me and both 4- and 6-Me groups have been replaced by CF3.130 An analog termed Copyright © 2005 CRC Press, LLC pseudo-actinomycin C1 has been synthesized in which the chromophoric methyls and the peptide moieties are exchanged.131 In another analog, the actinocinyl chromophore is replaced by 4-meth- ylphenazine-1,9-dicarboxylic acid.132 A number of peptidic analogs of actinomycin have been synthesized by methods similar to those used for the natural actinomycins (Table 15.3); they include enantio-AMD.133 X. STRUCTURE–ACTIVITY RELATIONSHIPS A. INTRODUCTION For biological activity mimicking the natural actinomycins, the chromophore and both intact cyclopeptide moieties are required. Even the scission of one peptide lactone abolishes biological activity.96,124 Some actinocinyl derivatives bind to DNA but are inactive. In the case of actinomine (actinocinyl-bis[diethylaminoethylamide]),67 this phenomenon was shown to derive from the rapid rate of dissociation from its complex with DNA, which was 1000 times faster than that of AMD. Thermodynamic and other studies62,70 indicated an important hydrophobic interaction between the peptide lactones of actinomycins and DNA, and their biological activity depends on the slow dissociation of the complex. The number of actinomycin analogs that have been described is very large (>100), and the review that follows is not exhaustive. A more detailed review of the work reported through 1977 has appeared.48 Natural actinomycins and their analogs have been evaluated in a variety of ways, including binding to DNA, inhibition of RNA synthesis, and antimicrobial and antitumor activities. Because antitumor activity is of primary interest, for those compounds for which such data are available, other activities are usually omitted. For other compounds, antimicrobial data, albeit a poor predictor of antitumor activity, is provided. Some analogs described above, which are biolog- ically inactive, are not discussed below. In many cases, analogs were produced in insufficient quantities for in vivo evaluation, and none reached clinical trial, although some partially synthetic analogs surpassed AMD in in vivo antitumor activity. TABLE 15.3 Amino Acid Content of Synthetic Peptidic Actinomycin Analogs Site 1 Site 2 Site 3 Site 4 Site 5 References Thr D-Val α-Hyp, β-Pro Sar MeVal 53 Thr D-Val Hyp Sar MeVal 53 Ser D-Val Pro Sar MeVal 134 Dpr D-Val Pro Sar MeVal 135 Dbu D-Val Pro Sar MeVal 136 Thr D-Ala Pro Sar MeVal 134 Thr D-Leu Pro Sar MeVal 134 Thr D-Thr Pro Sar MeVal 137 Thr D-Val Meg Sar MeVal 137 Thr D-Val Pro Gly Val 138 Thr D-Val Pro Sar MeAla 139 Thr D-Val Pro Sar MeLeu 137 D-Thr Val D-Pro Sar D-MeVal 133 Note: Nonstandard amino acid abbreviations: Dpr = L-2,3-diaminopropi- onic acid; Dbu = L-2,3-diamino-n-butyric acid; Meg = N-[2-(methoxycar- bonyl)ethyl]-glycine. Copyright © 2005 CRC Press, LLC TABLE 15.4 Biological Activities of Actinomycins I–VIIa Antitumor Streptomyces Streptomyces Actinomycin Toxicity Efficacyb aureusc subtilisc I = X0β 10 + 25 5–15 II 10 ++ 45 35 III 50 ++ 35 25–33 IV = AMD = X1 = C1 100 ++ 100 100 V = X2 800 ± 200 200 VI = C2 100 ++ 70 90 VII = C3 70 ++ 95 100 a Numerals indicate activity relative to AMD = 100. b Evaluated by increased life span in mice implanted with leukemia P388, leukemia L1210, or B16 melanoma. c Antimicrobial activities. B. NATURAL ACTINOMYCIN VARIANTS Comparative biological data for actinomycins I–VII are shown in Table 15.4.140,141 In addition, comparison of II, III, and IV in several mouse ascitic tumors revealed that in the most sensitive tumor, the Gardner lymphosarcoma, these actinomycins had comparable efficacy.142 These data indicate that replacement of the β-Pro by Hyp or by sarcosine reduces toxicity, antitumor potency, and antimicrobial activity, whereas antitumor efficacy remains about the same. In contrast, replace- ment by 4-oxo-Pro increases toxicity and antimicrobial activity but reduces antitumor efficacy. More recent work43 comparing actinomycins Z1, Z3, and Z5 with AMD is shown in Table 15.5. TABLE 15.5 Cytotoxicities (Micromolar) in Three Human Tumor Cell Lines and Minimal Inhibitory Concentrations (MICs) (µg/mL) against Bacillus subtilis for Actinomycins D, Z1, Z3, and Z5 Actinomycin HMO2 HEP G2 MCF 7 Antimicrobial and Level (Stomach) (Liver) (Breast) MIC D GI50 0.2 1.0 0.5 0.78 TGI 0.8 4.0 2.2 LC50 >50 >50 >50 Z1 GI50 0.75 0.95 <0.5 12.5 TGI 5.8 5.5 >50 LC50 >50 >50 >50 Z3 GI50 <0.1 <0.1 <0.1 0.20 TGI <0.1 1.4 <0.1 LC50 0.28 >50 0.5 Z5 GI50 <0.1 1.5 <0.1 0.75 TGI <0.1 10 0.12 LC50 0.50 >50 0.5 Copyright © 2005 CRC Press, LLC Actinomycin Z3 is several times as potent as AMD against all three human tumor cell lines studied, as well as against Bacillus subtilis. Z1 is considerably less active, and Z5 is roughly comparable with AMD. In earlier studies,143 Z1 and Z5 were less effective than AMD in inhibiting RNA synthesis in B. subtilis and in HeLa cells. The lower activity of Z1 probably results from the presence of 4- hydroxythreonine in a location where its OH group is close to the chromophoric amino group.52 C. CHROMOPHORIC ANALOGS The importance of the chromophoric 2-NH2 group is illustrated by the lack of biological activity when it is replaced by H, OH, or Cl108 or is acylated.119 Most 2-N-alkylated actinomycins were only tested against B. subtilis and were at least 10-fold less active that the natural variants.109,111 However, 2-N-γ-hydroxypropyl-AMD, although possessing only 10% of the antimicrobial activity of AMD, was equal in antitumor activity.112 Also, in P388 leukemia (mice), 2-N-3′-amino-n-propyl- AMD surpassed AMD in terms of increased life span (ILS).113 Replacement of the chromophoric Me groups by H or OMe weakened binding of the analog to DNA and reduced antimicrobial activity to 1%–5% of that of AMD, whereas their replacement by Et reduced antimicrobial activity to 30%–35%, and the di-t-Bu analog lacked activity, presumably because these bulky groups obstruct DNA intercalation.129 Despite the nearly equal size of Me and CF3 groups, antibacterial and cytotoxic activities of the 6-CF3 and 4, 6-di-CF3 analogs were reduced to 25% and almost 0% (respectively) of the parent actinomycin, probably because of electronic effects.130 Substitution in the chromophoric 7-position produces a variety of effects. 7-Chloroactinomycin C 114 3 and 7-nitroactinomycin C 110 2 have about half the antimicrobial activity of the parent com- pounds, whereas that of 7-bromoactinomycin C 114 3 is 50% higher. 7-Nitro- and 7-amino-AMD were roughly comparable with AMD in in vivo antitumor activity.144 In vivo antitumor studies (mice, P388) of nine N-substituted 7-amino-AMDs gave ILS values superior to those of AMD for 7- amino-AMD and 7-N-(3′,4′-dichlorophenyl)amino-AMD.117 In similar studies of 7-hydroxy-AMD and its O-alkyl and O-acyl derivatives, superior ILS values were seen in 7-hydroxy-AMD and 7- adamantoyloxy-AMD.113 No rationale was proposed for the enhanced efficacy resulting from these bulky substitutions in the chromophoric 7-position. Chromophoric analogs of AMD bearing a fourth oxazinone ring hydrolyze to AMD above pH7. In P388 murine leukemia, the compounds with R = CH2F and C6H5 (Figure 15.6) had ILS values superior to that of AMD.120 These analogs presumably act as AMD prodrugs, but their enhanced efficacy has not been explained. Dihydro-AMD analogs bearing a fourth oxazole ring were inactive unless the A ring was oxidized to the iminoquinone, which is metabolized to 7-hydroxy-AMD.121 The 8-amino-iminoquinones had antitumor activity superior to that of AMD when R″ = C6H5 or CH2CONH(CH2)4NH2 (Figure 15.6), presumably resulting from their metabolism to 7-hydroxy-8- amino-AMD.122 D. PEPTIDIC ANALOGS Synthetic variations at the Thr sites of actinomycin involving replacement by Ser, Dpr, or Dbu (Table 15.3) gave antimicrobial activities of 20%, 10%, and 10% (respectively) of that of AMD.134–136 Antitumor potency of the Dbu analog was 10–20 times lower than that of AMD. At the D-Val sites, replacement by D-Ala or D-Leu reduces antimicrobial activities to 1% and 10% (respectively) of that of AMD,134 and D-Thr produces a reduction to 10%–20% of that of AMD.137 Replacement of one Pro by pipecolic acid (Pip 1β) has little effect on antimicrobial activity, whereas replacement of both (Pip 2) reduces this activity 5- to 15-fold, and replacement by one 4- ketopipecolic acid (Pip 1δ) reduces it further.145 The in vivo antitumor efficacy of Pip 1 is comparable with that of AMD at an eightfold higher dosage, whereas Pip 2 is less active.86 The in vivo antitumor activity of the compounds

in which one or both prolines were replaced by azetidine-2-carboxylic acid, termed azetomycins I and II, respectively, have been extensively evaluated; azetomycin I, the Copyright © 2005 CRC Press, LLC more active of the two, was found equivalent to AMD in some experiments146 and somewhat superior in others.147 Actinomycins in which one or both prolines were replaced by cis- or trans-4-MePro (termed K1c, K2c, K1t, and K2t respectively) had lower antimicrobial activity than AMD in the sequence AMD > K1t > K1c > K2t > K2c.103 Replacement of both MeVal residues by MeLeu137 produced much stronger binding to DNA than AMD, and this analog was at least 10-fold more potent than AMD in the NCI in vitro screen comprising 60 human tumor cell lines.79 (This analog was initially reported137 to be approximately 100 times as potent as AMD in the NCI screen, but evaluation of data from several experiments has since required a revision of that ratio.) Presumably the MeLeu side-chains enhance hydrophobic interaction with DNA. In contrast, the AMD analog with both MeVal residues replaced by MeAla binds to DNA less strongly than AMD and has approximately 10-fold lower antimicrobial activity,139 indicating a poorer hydrophobic interaction with DNA. This effect does not prevent actinomycins Z3 and Z5, which both have one MeVal replaced by MeAla, from being at least as active as AMD, but there are other structural differences, including additional methyl groups in the 5-position of both prolines.43 XI. CLINICAL APPLICATIONS OF ACTINOMYCINS Many early clinical studies with actinomycins C and D in a variety of cancers demonstrated that their effectiveness was most prominent in trophoblastic tumors in females148 (including methotrex- ate-resistant choriocarcinoma149), in metastatic carcinoma of the testis,148,150,151 and in Wilms’ tumor (nephroblastoma) in children.152 In rhabdomyosarcoma, efficacy was enhanced by combination with vincristine.153 Cures were obtained in Wilms’ tumor by a combination of surgery, radiotherapy, and actinomycin therapy.154,155 The most effective route of administration is intravenous. Actinomycin D, also known as Dactinomycin and Cosmegen, soon became the only form of actinomycin in clinical use, as it is produced by S. parvulus as a single component, whereas actinomycin C is a mixture, and they are equally effective. Actinomycin D is still in use in therapy of the above-mentioned tumors, as well as in treatment of Ewing’s sarcoma. Side-effects include bone marrow depression, oral and intestinal lesions, skin reactions, alopecia, and immunosuppres- sion. As a consequence, in common with other classical anticancer drugs, its effectiveness is constrained by dose-limiting toxicity. Nonetheless, many complete remissions have occurred. REFERENCES 1. Waksman, S.A. and Woodruff, H.B., Bacteriostatic and bacteriocidal substances produced by a soil actinomyces, Proc. Soc. Exptl. Biol. Med., 45, 609, 1940. 2. Waksman, S.A. and Woodruff, H.B., Actinomyces antibioticus, a new soil organism antagonistic to pathogenic and nonpathogenic bacteria, J. Bacteriol., 42, 231, 1941. 3. Fisher, W.P., Charney, J. and Bolhofer, W.A., An actinomycin from a species of the genus Micromono- spora, Antibiot. Chemother., 1, 571, 1951. 4. Waksman, S.A. et al., Toxicity of actinomycin, Proc. Soc. Exptl. Biol. Med., 47, 261, 1941. 5. Hackmann, C., Experimentelle untersuchungen über die wirkung von actinomycin C, Z. Krebsforsch., 58, 607, 1952. 6. Schulte, G., Erfahrungen mit neuen zytostatischen mitteln bei hämoblastosen und carcinomen und die abgrenzung ihrer wirkungen gegen roentgentherapie, Z. Krebsforsch, 58, 500, 1952. 7. Kirk, J.M., The mode of action of actinomycin D, Biochim. Biophys. Acta, 42, 167, 1960. 8. Slotnick, I.J., Mechanism of action of actinomycin D in microbiological systems, Ann. N.Y. Acad. Sci., 89, 342, 1960. 9. Shatkin, A.J., The effects of actinomycin on virus replication, in Actinomycin, Waksman, S.A., ed., Interscience Publishers, New York, 1968, 69. Copyright © 2005 CRC Press, LLC 10. Lehr, K. and Burger, J., The isolation of a crystalline actinomycin-like antibiotic, J. Arch. Biochem., 23, 503, 1949. 11. Brockmann, H. and Grubhofer, N., Actinomycin C, Naturwiss., 36, 376, 1949. 12. Manaker, R.A. et al., Actinomycin. III. The production and properties of a new actinomycin, Antibiot. Annu., 853, 1954/55. 13. Brockmann, H. and Gröne, H., Reine actinomycine, Naturwiss., 41, 65, 1954. 14. Brockmann, H., Linge, H. and Gröne, H., Zur kenntnis des actinomycins X, Naturwiss., 40, 224, 1953. 15. Bossi, R. et al., Stoffwechselprodukte von actinomyceten, 14. Mitt. actinomycin Z, Helv. Chim. Acta, 41, 1645, 1958. 16. Brockmann, H. et al., Über das actinomycin C (antibiotica aus actinomyceten, V. Mitt.), Chem. Ber., 84, 260, 1951. 17. Brockmann, H. and Pfennig, N., Auftrennung von actinomycin C durch gegenstromverteilung, Natur- wiss., 39, 429, 1952. 18. Vining, L.C. and Waksman, S.A., Paper chromatographic identification of the actinomycins, Science, 120, 389, 1954. 19. Gregory, F.J., Vining, L.C. and Waksman, S.A., Actinomycin. IV. Classification of the actinomycins by paper chromatography, Antibiot. Chemother., 5, 409, 1955. 20. Vining, L.C., Gregory, F.J. and Waksman, S.A., Actinomycin. V. Chromatographic separation of the actinomycin complexes, Antibiot. Chemother., 5, 417, 1955. 21. Roussos, G.G. and Vining, L.C., Isolation and properties of pure actinomycins, J. Chem. Soc., 2469, 1956. 22. Mauger, A.B. and Katz, E., Actinomycins, in Antibiotics: isolation, separation and purification, J. Chromatogr. Library, Wagman, G.H. and Weinstein, M.J., Eds., Elsevier, New York, 1978, 15, 1. 23. Waksman, S.A., Katz, E. and Vining, L.C., Nomenclature of actinomycins, Proc. Natl. Acad. Sci. U.S.A., 44, 602, 1958. 24. Brockmann, H. et al., Zur konstitution der actinomycine, Angew. Chem., 68, 70, 1956. 25. Bullock, E. and Johnson, A.W., The structure of actinomycin D, J. Chem. Soc, 3280, 1957. 26. Brockmann, H., Boldt, P. and Petras, H.-S., Die aminosäuresequenz von actinomycin C1 und actino- mycin C2, Naturwiss., 47, 62, 1960. 27. Brockmann, H. and Franck, B., Actinomycin C2a, ein isomeres des actinomycins C2, Naturwiss., 47, 15, 1960. 28. Brockmann, H. and Boldt, P., Oxidative abspaltung des β-pentapeptidlactonringes aus actinomycin C2 und C3; ein beweis für die bis-pentapeptidlacton-struktur der actinomycine, Chem. Ber., 101, 1940, 1968. 29. Brockmann, H. and Manegold, J.H., Ueberfuehrung von actinomycin X2 in die actinomycine C1, X0β und X0β; actinomycine XXIII, Chem. Ber., 93, 2971, 1960. 30. Diegelman, R., Mauger, A. and Katz, E., Isolation and identification of oxoproline in actinomycin X2 (V) hydrolysates, J. Antibiot., 22, 85, 1969. 31. Brockmann, H and Manegold, J.H., Actinomycin X1a and X0β, Chem. Ber., 95, 1081, 1962. 32. Waksman, S.A. et al., Directed formation of the actinomycins — chemical structure and biological activities, Science, 129, 1290, 1959. 33. Johnson, A.W. and Mauger, A.B., The isolation and structure of actinomycins II and III, Biochem. J., 73, 535, 1959. 34. Mauger, A.B., Degradation of peptides to diketopiperazines: application of pyrolysis-gas chromatog- raphy to sequence determination in actinomycins, J. Chem. Soc. Chem. Commun., 39, 1971. 35. Katz, E. and Mauger, A.B., unpublished data, 1977. 36. Brockmann, H. and Stähler, E.A., Zur konstitution des actinomycins Z5, Tetrahedron Lett., 2567, 1973. 37. Katz, E., Mason, K.T. and Mauger, A.B., Identification of cis-5-methylproline in hydrolysates of actinomycin Z5, Biochem. Biophys. Res. Commun., 52, 819, 1973. 38. Brockmann, H. and Stähler, E.A., 4-Oxo-5-methylprolin, ein baustein des actinomycins Z1, Naturwiss., 52, 391, 1965. 39. Brockmann, H. and Manegold, J.H., Actinomycine, XXV; antibiotica aus actinomyceten, LIV. Acti- nomycin Z1; zur quantitativen aminosäureanalyse der actinomycine, Hoppe-Seyler’s Z. Physiol. Chem., 343, 86, 1965. Copyright © 2005 CRC Press, LLC 40. Katz, E., Mason, K.T. and Mauger, A.B., 3-Hydroxy-5-methylproline, a new amino acid identified as a component of actinomycin Z1, Biochem. Biophys. Res. Commun., 63, 502, 1975. 41. Mauger, A.B. et al., Synthesis and stereochemistry of 3-hydroxy-5-methylproline, a new naturally occurring imino acid, J. Org. Chem., 42, 1000, 1977. 42. Katz, E., Mason, K.T. and Mauger, A.B., The presence of α-amino-β,γ-dihydroxybutyric acid in hydrolysates of actinomycin Z1, J. Antibiot. 27, 952, 1974. 43. Lackner, H. et al., Structures of five components of the actinomycin Z complex from Streptomyces fradiae, two of which contain 4-chlorothreonine, J. Nat. Prod., 63, 352, 2000. 44. Lackner, H. et al., A new actinomycin-type chromopeptide from Streptomyces Sp. HKI-0155, J. Antibiot., 53, 84, 2000. 45. Hanada, M. et al., Protactin, a new antibiotic metabolite and a possible precursor of the actinomycins, J. Antibiot., 45, 20, 1992. 46. Brockmann, H., Die actinomycine, Fortschr. Chem. Org. Naturst., 18, 1, 1960. 47. Umezawa, H., Index of antibiotics from Actinomycetes, Japan Scientific Societies Press, State College Pennsylvania, 1967 (Vol.I) and Univ. Park Press, Baltimore, MD, 1978 (Vol. II). 48. Meienhofer, J. and Atherton, E., Structure-activity relationships in the actinomycins, in Structure- activity relationships among the semisynthetic antibiotics, Perlman, D., ed., John Wiley & Sons, New York, 1977, 427. 49. Mauger, A.B., The actinomycins, in Topics in antibiotic chemistry, Vol. 5, Sammes, P.G., ed., Ellis Horwood, Chichester, 1980, 223. 50. Jain, S.C. and Sobell, H.M., Stereochemistry of actinomycin binding to DNA. I. Refinement and further structural details of the actinomycin-deoxyguanosine crystalline complex, J. Mol. Biol., 68, 1, 1972. 51. Ginell, S., Lessinger, L. and Berman, H.M., The crystal and molecular structure of the anticancer drug actinomycin D — some explanations for its unusual properties, Biopolymers, 27, 843, 1988. 52. Schäfer, M., Sheldrick, G.M. et al., Crystal structures of actinomycin D and actinomycin Z3, Angew. Chem. Internat. Ed., 37, 2381, 1998. 53. Lifferth, A. et al., Synthese und struktur von prolinring-modifizierten actinomycinen des X-typs, Z. Naturforsch., 54b, 681, 1999. 54. Lackner, H., Three-dimensional structure of the actinomycins, Angew. Chem. Int. Ed., 14, 375, 1975. 55. Liu, X., Chen, H. and Patel, D.J., Solution structure of actinomycin-DNA complexes: drug intercalation at isolated G-C sites, J. Biomol. NMR, 1, 323, 1991. 56. Kawamata, J. and Imanishi, M., Interactions of actinomycin with deoxyribonucleic acid, Nature, 187, 1112, 1960. 57. Kersten, H., Kersten, W. and Rauen, H.M., Action of nucleic acids on the inhibition of growth by actinomycin of Neurospora crassa, Nature, 187, 60, 1960. 58. Rauen, H.M., Kersten, H. and Kersten, W., Zur wirkungsweise von actinomycinen, Hoppe-Seyler’s Z. Physiol. Chem., 321, 139, 1960. 59. Kawamata, J. and Imanishi, M., Mechanism of action of actinomycin, with special reference to its interaction with deoxyribonucleic acid, Biken J., 4, 13, 1961. 60. Reich, E. et al., Actinomycin: Effect of actinomycin D on cellular nucleic acid synthesis and virus production, Science, 134, 556, 1961. 61. Haselkorn, R., Actinomycin D as a probe for nucleic acid secondary structure, Science, 143, 682, 1964. 62. Reich, E. and Goldberg, I.H., Actinomycin and nucleic acid function, Prog. Nucleic Acid Res. Mol. Biol., 3, 183, 1964. 63. Gellert, M. et al., Actinomycin binding to DNA: mechanism and specificity, J. Mol. Biol., 11, 445, 1965. 64. Goldberg, I.H., Rabinowitz, M. and Reich, E., Basis of actinomycin action. I. DNA binding and inhibition of RNA-polymerase synthetic reactions by actinomycin, Proc. Natl. Acad. Sci. U.S.A., 48, 2094, 1962. 65. Cerami, A. et al., The interaction of actinomycin with DNA: requirement for the 2-amino group of purines, Proc. Natl. Acad. Sci. U.S.A., 57, 1036, 1967. 66. Kahan, E., Kahan, F.M. and Hurwitz, J., The role of deoxyribonucleic acid in ribonucleic acid synthesis. VI. Specificity of action of actinomycin D, J. Biol. Chem., 238, 2491, 1963. 67. Müller, W. and Crothers, D.M., Studies on the binding of actinomycin and related compounds to DNA, J. Mol. Biol., 35, 251, 1968. Copyright © 2005 CRC Press, LLC 68. Kersten, W., Interaction of actinomycin C with constituents of nucleic acids, Biochim. Biophys. Acta, 47, 610, 1961. 69. Takusagawa, F. et al., The structure of a pseudo intercalated complex between actinomycin and the DNA binding sequence d(GpC), Nature, 296, 466, 1982. 70. Takusagawa, F. et al., Crystallization and preliminary X-ray study of a complex between d(ATGCAT) and actinomycin D, J. Biol. Chem., 259, 4714, 1984. 71. Takusagawa, F., The role of the cyclic depsipeptide rings in antibiotics, J. Antibiot., 38, 1596, 1985. 72. Ross, W.E., Glaubiger, D. and Kohn, K.W., Qualitative and quantitative aspects of intercalator-induced DNA strand breaks, Biochim. Biophys. Acta, 562, 41, 1979. 73. Wasserman, K. et al., Effects of morpholinyl doxorubicins, doxorubicin and actinomycin D on mammalian DNA topoisomerases I and II, Mol. Pharmacol., 38, 38, 1990. 74. Wadkins, R.M. et al., Actinomycin D binding to single-stranded DNA: sequence specificity and hemi- intercalation model from fluorescence and 1H NMR spectroscopy, J. Mol. Biol., 262, 53, 1996. 75. Gniazdowski, M. et al., Transcription factors as targets for DNA-interacting drugs, Curr. Med. Chem. 10, 909, 2003. 76. Foley, G.E. and Eagle, H., The cytotoxicity of anti-tumor agents for normal human and animal cells in first tissue culture passage, Cancer Res., 18, 1012, 1958. 77. Eagle, H. and Foley, G.E., Cytotoxicity in human cell cultures as a primary screen for the

detection of anti-tumor agents, Cancer Res., 18, 1017, 1958. 78. Reich, E., Goldberg, I.H. and Rabinowitz, M., Structure-activity correlations of actinomycins and their derivatives, Nature, 196, 743, 1962. 79. Monks, A. et al., Feasibility of a high-flux anticancer drug screen utilizing a diverse panel of human tumor cell lines in culture, J. Natl. Cancer Inst., 83, 757, 1991. 80. Paul, K.D. et al., Display and analysis of patterns of differential activity of drugs against human tumor cell lines: development of mean graph and COMPARE algorithm, J. Natl. Cancer Inst., 81, 1088, 1989. 81. Hackmann, C., HBF 386 (Actinomycin C), ein cytostatisch wirksamer naturstoff, Strahlentherapie, 90, 296, 1953. 82. Stock, J.A., Antitumor antibiotics, Exptl. Chemother., 4, 239, 1966. 83. Sugiura, K. and Schmidt, M.S., Effects of antibiotics on the growth of a variety of mouse and rat tumors, Proc. Am. Assoc. Cancer Res., 2, 151, 1956. 84. Hirono, I. et al., Antitumor effect of actinomycin D on transplantable Wilm’s tumor in rats, Gann., 59, 473, 1968. 85. Priestly, J.B., The laboratory basis for the clinical treatment of Wilms tumor, J. Urol, 107, 696, 1972. 86. Goldin, A. and Johnson, R., Evaluation of actinomycins in experimental systems, Cancer Chemother. Reps., Part 1, 58, 63, 1974. 87. Philips, F.S. et al., The toxicity of actinomycin D, Ann. N.Y. Acad. Sci., 89, 348, 1960. 88. Brockmann, H. and Lackner, H., Totalsynthese von actinomycin C3, Naturwiss., 47, 230, 1960. 89. Brockmann, H. and Lackner, H., Totalsynthese von actinomycin C3 über bis-seco-actinomycin C3, Chem. Ber., 100, 353, 1967. 90. Bachmann, H.G. and Müller, W., X-ray diffraction of actinomycin C3, Nature, 201, 261, 1964. 91. Perutz, M.F., X-ray diffraction of actinomycin C3, Nature, 201, 814, 1964. 92. Brockmann, H. and Lackner, H., Totalsynthese der actinomycine C1, C2, i-C2 und C3 über N-(2-amino- 3-hydroxy-4-methylbenzoyl)-pentapeptidlactone; ein weiterer beweis für die bispentapeptidstruktur der actinomycine, Chem. Ber., 101, 2231, 1968. 93. Meienhofer, J., Synthesis of actinomycin and analogs. III. A total synthesis of actinomycin D (C1) via peptide cyclization between proline and sarcosine, J. Am. Chem. Soc., 92, 3771, 1970. 94. Vlasov, G.P., Lashkov, V.N. and Glibin, E.N., Synthesis of actinomycin D and its peptide analogs. IV. Synthesis of de(di-3’-proline)actinomycin D, Zh. Org. Khim. S.S.S.R., 15, 983, 1979. 95. Mauger, A.B. and Stuart, O.A., Synthesis of an actinomycin-related peptide lactone from the corre- sponding cyclic peptide by N,O-acyl shift, Int. J. Peptide Protein Res., 34, 196, 1989. 96. Lackner, H., Totalsynthese von aniso-actinocinyl-peptiden und aniso-actinomycinen über deuterium- markierte vorstufen, Chem. Ber., 103, 2476, 1970. 97. Lackner, H., Synthesen von selectiv deuterierten pentapeptidlactonen und stellungsspezifisch im α- oder β-peptidring markierten actinomycinen, Chem. Ber., 104, 3653, 1971. Copyright © 2005 CRC Press, LLC 98. Katz, E., Controlled biosynthesis of actinomycins, Cancer Chemother. Rep., Part 1, 58, 83, 1974. 99. Schmidt-Kastner, G., Actinomycin E und actinomycin F, zwei neue biosynthetische actinomycinge- mische, Naturwiss., 43, 131, 1956. 100. Katz, E., Kawai, Y. and Shoji, J., Configuration of the N-methylisoleucine in the actinomycins, J. Biochem. Biophys. Res. Commun., 43, 1035, 1971. 101. Yajima, T., Grigg, M.A. and Katz, E., Biosynthesis of antibiotic peptides with isoleucine isomers, Arch. Biochem. Biophys., 151, 565, 1972. 102. Schmidt-Kastner, G., The production of actinomycins by controlled biosynthesis: the F actinomycins, Ann. N.Y. Acad. Sci., 89, 299, 1960. 103. Katz, E. et al., Novel actinomycins formed by biosynthetic incorporation of cis- and trans-4-methyl- proline, Antimicrob. Agents Chemother., 11, 1056, 1977. 104. Nishimura, J.S. and Bowers, W.F., Evidence for the incorporation of L-thiazolidine-4-carboxylate into actinomycins by Streptomyces antibioticus, Biochem. Biophys. Res. Commun., 28, 665, 1967. 105. Katz, E., Biogenesis of the actinomycins, Ann. N.Y. Acad. Sci., 89, 304, 1960. 106. Formica, J.V. and Apple, M.A., Production, isolation and properties of azetomycins, Antimicrob. Agents Chemother., 9, 214, 1976. 107. Formica, J.V. and Katz, E., Isolation, purification, and characterization of pipecolic acid-containing actinomycins, Pip 2, Pip 1α and Pip 1β, J. Biol. Chem., 248, 2066, 1973. 108. Brockmann, H., Gröne, H. and Pampus, G., Chloractinomycine, Chem. Ber., 91, 1916, 1958. 109. Brockmann, H., Pampus, G. and Mecke, R., Actinomycine, XXII; antibiotica aus actinomyceten, XLIV: N-Alkyl-actinomycine, Chem. Ber., 92, 3082, 1959. 110. Brockmann, H., Müller, W. and Peterssen-Borstel, H., Actinomycin-derivate aus dihydroactinomy- cinen, Tetrahedron Lett., 3531, 1966. 111. Brockmann, H., Hocks, P. and Müller, W., N-2-Substituierte actinomycine, Chem. Ber., 100, 1051, 1967. 112. Moore, S. et al., Synthesis and antitumor activity of 2-deamino- and N2-(γ-hydroxypropyl)actinomycin D, J. Med. Chem., 18, 1098, 1975. 113. Sengupta, S.K. et al., N2- and C-7-substituted actinomycin D analogues: Synthesis, DNA-binding affinity, and biochemical and biological properties. Structure-activity relationship, J. Med. Chem., 24, 1052, 1981. 114. Brockmann, H., Ammann, J. and Müller, W., 7-Chlor- und 7-brom-actinomycine, Tetrahedron Lett., 3595, 1966. 115. Sengupta, S.K. et al., Synthesis and DNA-binding properties of 7-amino-actinomycin D — a selective fluorescent binding agent, Fed. Proc., Fed. Am. Soc. Exptl. Biol., 30, 342, 1971. 116. Müller, W., Bindung von actinomycin und actinomycin-derivaten an desoxyribonucleinsäuren, Natur- wiss., 49, 156, 1962. 117. Madhavarao, M.S., Chaykovsky, M. and Sengupta, S.K., N7-Substituted 7-aminoactinomycin D ana- logs. Synthesis and biological properties, J. Med. Chem., 21, 958, 1978. 118. Sengupta, S.K., Kelly, C. and Sehgal, R.K., “Reverse” and “Symmetrical” analogs of actinomycin D: metabolic activation and in vitro and in vivo tumor growth inhibitory activities, J. Med. Chem., 28, 620, 1985. 119. Brockmann, H. and Franck, B., Hydrierende acetylierung von actinomycin, Angew. Chem., 68, 68, 1956. 120. Sengupta, S.K. et al., Actinomycin D oxazinones as improved antitumor agents, J. Med. Chem., 22, 797, 1979. 121. Sengupta, S.K. et al., Tetracyclic chromophoric analogs of actinomycin D: Synthesis, structure elu- cidation and interconvertability from one form to another, antitumor activity, and structure-activity relationships, J. Med. Chem., 26, 1631, 1983. 122. Sengupta, S.K., et al. New actinomycin D analogs as superior chemotherapeutic agents against primary and advanced colon tumors and colon xenografts in nude mice, J. Med. Chem., 31, 768, 1988. 123. Brockmann, H. and Franck, B., Aufspaltung der actinomycine zu actinomycinsäuren, Angew. Chem., 68, 68, 1956. 124. Perlman, D., Mauger, A.B. and Weissbach, H., Microbial transformations of peptide antibiotics. I. Degradation of actinomycins by Actinoplanes species, Antimicrob. Agents Chemother., 581, 1966. 125. Brockmann, H. and Sunderkötter, W., Abbau von actinomycin C3 mit konzentrierter salzsäure, Natur- wiss., 47, 229, 1960. Copyright © 2005 CRC Press, LLC 126. Brockmann, H., Pampus, G. and Manegold, J.H., Actinomycin X0β; zur systematik und nomenklatur der actinomycine, Chem. Ber., 92, 1294, 1959. 127. Brockmann, H. and Stähler, E.A., 3-Hydroxy-4-oxo-5-methylprolin als baustein von actinomycin Z1, Tetrahedron Lett., 3585, 1973. 128. Mauger, A.B. and Wade, R., The synthesis of actinomycin analogs. Part II. Actinocyl-gramicidin S, J. Chem. Soc., 1406, 1966. 129. Brockmann, H. and Seela, F., 4,6-Didesmethyl-actinomycin C1 und dessen 4,6-dimethoxy-, 4,6- diäthyl- und 4,6-di-tert-butyl-derivat, Chem. Ber., 104, 2751, 1971. 130. Giencke, A. and Lackner, H., Desmethyl(trifluormethyl)-actinomycine, Liebigs Ann. Chem., 569, 1990. 131. Brockmann, H. and Schulze, E., Synthese von pseudo-actinomycin C1, Tetrahedron Lett., 1489, 1971. 132. Mosher, C.W. et al., A phenazine analog of actinomycin D, J. Med. Chem., 22, 918, 1979. 133. Brockmann, H. and Schramm, W., Synthese von actinomycin-(thr-Val-pro-sar-meval), dem antipoden von actinomycin C1, Tetrahedron Lett., 2331, 1966. 134. Brockmann, H. and Lackner, H., Aufbau von actinomycinen über actinomycinsäuren; ein zweiter syntheseweg, Chem. Ber., 101, 1312, 1968. 135. Moore, S. et al., Synthesis and some properties and antitumor effects of the actinomycin lactam analog [di(1′-L-α,β-diaminopropionic acid)]actinomycin D, J. Med. Chem., 19, 766, 1976. 136. Atherton, E. et al., Actinomycin D lactam, [1′1′-bis(threo-α,β-diaminobutyric acid)]actinomycin D, J. Med. Chem., 16, 355, 1973. 137. Mauger, A.B. and Stuart, O.A., Synthesis and properties of some peptide analogs of actinomycin D, J. Med. Chem., 34, 1297, 1991. 138. Mosher, C.W. and Goodman, M., Synthesis of [4′,4′-bis(glycine), 5′,5′-bis(valine)]actinomycin D, a tetra-N-demethylactinomycin, J. Org. Chem., 37, 2928, 1972. 139. Mauger, A.B. et al., Conformation and dimerization of actinomycin-related peptide lactones in solution and in the solid state, J. Am. Chem. Soc., 107, 7154, 1985. 140. Pugh, L.H., Katz, E. and Waksman, S.A., Antibiotic and cytostatic properties of the actinomycins, J. Bacteriol, 72, 660, 1956. 141. Katz, E. and Pugh, L.H., Variations on a theme. The actinomycins, Appl. Microbiol., 9, 263, 1961. 142. Pugh, L.H. and Solotorovsky, M., Comparative effects of actinomycins II, III and IV on several ascitic tumors in mice, Ann. N.Y. Acad. Sci., 89, 373, 1960. 143. Mason, K., Katz, E. and Mauger, A.B., Studies on the biological activities of actinomycins Z1 and Z5, Arch. Biochem. Biophys., 160, 402, 1974. 144. Modest, E.J. and Sengupta, S.K., 7-Substituted actinomycin D (NSC-3053) analogs as fluorescent DNA-binding and experimental antitumor agents, Cancer Chemother. Reps. Part I, 58, 35, 1974. 145. Formica, J.V., Shatkin, A.J. and Katz, E., Actinomycin analogs containing pipecolic acid: relationship of structure to biological activity, J. Bacteriol., 95, 2139, 1968. 146. Johnson, R.K., private communication. 147. Rose, W.C. et al., Comparative antitumor activity of actinomycin analogs in mice bearing Ridgeway osteogenic sarcoma or P388 leukemia, Cancer Treatment Reps., 62, 779, 1978. 148. Li, M.C., Management of choriocarcinoma and related tumors of uterus and testis, Med. Clin. N. Am., 45, 661, 1961. 149. Ross, G.T., Stohlbach, L.L. and Hertz, R., Actinomycin in the treatment of methotrexate-resistant trophoblastic disease in women, Cancer Res., 22, 1015, 1962. 150. Li, M.C. et al., Effects of combined drug therapy of metastatic cancer of the testis, J. Am. Med. Assoc., 174, 1291, 1960. 151. MacKenzie, A.R., Chemotherapy of metastatic testis cancer — results in 154 patients, Cancer, 19, 1369, 1966. 152. Tan, C.T.C., Dargeon, H.W. and Burchenal, J.H., The effect of actinomycin D in cancer in childhood, Pediatrics, 24, 544, 1959. 153. James, D.H. et al., Childhood malignant tumors. Concurrent chemotherapy with dactinomycin and vincristine sulfate, J. Am. Med. Assoc., 19, 1043, 1966. 154. Farber, S., Chemotherapy in the treatment of leukemia and Wilms’ tumor, J. Am. Med. Assoc., 198, 826, 1966. 155. Fernbach, D.J. and Martyn, D.T., Role of dactinomycin in the improved survival of children with Wilms’ tumor, J. Am. Med. Assoc., 195, 1005, 1966. Copyright © 2005 CRC Press, LLC 16 Anthracyclines Federico Maria Arcamone CONTENTS I. Introduction II. Development of the Anthracycline Drugs A. Daunorubicin B. Doxorubicin C. Second-Generation Antitumor Anthracyclines III. Mechanism of Action A. DNA Complex and Topoisomerase II Poisoning B. Natural and Acquired Resistance to Anthracyclines C. Cardiotoxicity IV. Synthesis of the Drugs A. Doxorubicin B. Epirubicin C. Idarubicin V. Medicinal Chemistry of Antitumor Anthracyclines A. Analogs Modified in the Aglycone Moiety 1. Deoxyderivatives 2. 8-Fluoroderivatives 3. 14-α-Fluorohydrins B. Analogs Modified in the Sugar Moiety 1. Monosaccharide Analogs 2. Disaccharide Analogs C. Anthracycline–Triplex Helix-Forming Oligonucleotide Conjugates VI. Clinical Activity of Anthracyclines A. Daunorubicin B. Doxorubicin C. Epirubicin D. Idarubicin E. Aclacinomycin F. Sabarubicin G. Liposomal Anthracyclines References I. INTRODUCTION Anthracycline aminoglycosides represent a wide class of antibiotics obtained by submerged aerobic fermentation of different microorganisms belonging to the genus Streptomyces and are called anthracyclines because the aglycone moiety is a tetracyclic system bearing an anthraquinone chro- mophore.1 The anthracyclines are microbial metabolites belonging, from the biogenetic standpoint, Copyright © 2005 CRC Press, LLC to the large family of the polyketide natural products, comprising a wide range of biologically active structural groups: the antibacterial antibiotics, tetracyclines, and the macrolides of the erythromycin type; immunosuppressants such as FK506; antiparasitic compounds such as monensin; and the avermectins. These molecules are synthesized by multifunctional polyketide synthase enzymes, which catalyze repeated condensation cycles between simple acylthioesters with the formation of a growing carbon chain containing β keto groups. These keto groups may undergo a number of intramolecular reactions, leading to aromatic derivatives, such as the aglycone of the anthracyclines, the anthracyclinones, and a series of reductive steps. The final products are characterized by a considerable molecular diversity.2 The antitumor activity of a biosynthetic anthracycline was first reported in 1959.3 In fact, different samples of a mixture of red pigments isolated from the cultures of a Streptomyces strain at Farmitalia Research Laboratories were found to be active against murine Ehrlich carcinoma and sarcoma 180 test systems.3 The presence of compounds possessing the hydroxyanthraquinone chromophore was deduced on the basis of the typical electronic spectrum — red in acid and blue in alkaline solutions. The pigments were isolated by countercurrent distribution, and their rela- tionship with the anthracyclines — previously studied and described by different authors, and more precisely with the rhodomycin complex1 — was established. Subsequently, useful pharma- cological properties were found that were associated with the compounds produced by a novel species, Streptomyces peucetius, and related strains. Doxorubicin

(Ia), the best known component of this group, is the 14-hydroxylated derivative of the main fermentation product daunorubicin (Ib), and since its registration in the early 1970s, doxorubicin has been one of the most widely used drugs in cancer chemotherapy.4 (The trademark of doxorubicin formulations in the United States is Adriamycin®.) The successful therapeutic application of Ia stimulated a considerable effort in the chemical and biological study of the antitumor anthracyclines aimed at the development of better analogs by chemical modification of the parent drugs or by total synthesis of new structurally related compounds. Although Ia remains one of the most effective agents in the medical treatment of a range of solid tumors, different new members of this chemical group, among which epirubicin (II) and idarubicin (III) are the most important, are presently used in the medical practice. Such new agents were already described in previous books5,6 and are currently known as second- generation anthracyclines. In 1992 it was observed that an analog truly superior to doxorubicin had not been identified. In fact, after the synthesis and biological testing of hundreds of new anthracycline glycosides in different laboratories, neither an analog with a really improved therapeutic index nor one with a substantially different spectrum of activity had been found.7 It should be noted that analogs with considerable structural variations were submitted to testing procedures that were mostly limited to murine models of transplantable leukemias, which is possibly not the best way to take an appropriate selection of new interesting compounds. The present tumor type–oriented pharmacological evalu- ation, based on the use of a large number of human tumor cell lines and on human tumor heterotransplants in nude mice, coupled with rational drug design and a closer attention to aspects such as drug bioavailability, tissue distribution, and pharmacokinetics, might increase the probability of obtaining results closer to the desired goal.8 Antitumor activity also has been associated with other biosynthetic anthracyclines generally differing from the daunorubicin–doxorubicin group in the chemical structure of both the tetracyclic aglycone and the sugar moiety, with the latter often present as an oligosaccharide fragment. One such compound, aclacinomycin, has been introduced in clinical use.9 The biosynthetic anthracy- clines have been reviewed, together with the products of microbial transformations of individual anthracycline antibiotics, by Oki.10 More recently, an overview of the different biosynthetic anthra- cyclines, describing the producing microorganisms, the chemical structure, and when available, the pharmacological data of over 250 compounds has been published.11 A series of outstanding review articles on the fascinating molecular biology of anthracycline biosynthesis is also available.12-14 Copyright © 2005 CRC Press, LLC II. DEVELOPMENT OF THE ANTHRACYCLINE DRUGS A. DAUNORUBICIN In December 1958 and in February 1959, two similar crude preparations, obtained from a different Streptomyces species isolated from a soil sample collected in Apulia at Caste del Monte close to the ancient castle of King Frederick II, were shown to be endowed with significantly superior antitumor properties in terms of tumor size reduction and time of survival of the treated animals versus untreated ones in the above-mentioned mouse tests. The new microorganism and the main compound obtained from the active samples were named S. peucetius and daunomycin, respectively, because of the names of the ancient populations living in the region in early historical times, the Peucetii and the Daunii. The significant antitumor activity of daunomycin was confirmed in a wider spectrum of experimental systems.15 These results stimulated a detailed investigation of the chemical structure of daunomycin to determine the differences between it and the classical anthracyclines — differences that were responsible for the markedly improved pharmacological properties. The products of acid hydrolysis, namely, aglycone, daunomycinone, and the sugar moiety, daunosamine, were studied by means of chemical degradation reactions and of physicochemical analyses. In brief, both fragments were new compounds16,17 indicating that the microorganism produced a novel family of anthracycline glycosides. The complete structure and stereochemistry of daunomycin were finally established as represented in Ib,18–20 and they were eventually confirmed by the x-ray diffraction analysis of the corresponding N-bromoacetylderivative.21 In the meantime, daunomycin was iden- tified with rubidomycin, isolated by French researchers from Streptomyces coeruleorubidus.22 The World Health Organization suggested “daunorubicin” for the compound’s name in recognition of the independent contributions from Farmitalia and Rhone-Poulenc laboratories. Clinical studies, showing its high activity in acute leukemias, were presented in a meeting held at St. Louis Hospital in Paris in 1967.23 B. DOXORUBICIN Doxorubicin (Ia, originally named Adriamycin), a compound with marked activity in experimental systems and a wide spectrum of antitumor efficacy, especially in solid tumors, was isolated as a pigment coproduced with daunorubicin in the culture medium of S. peucetius–derived strains.24 The structure of doxorubicin was determined, inter alia, by nuclear magnetic resonance analysis of the aglycone, adriamycinone, and by direct semisynthesis from daunorubicin.25 Because of the very small quantities of doxorubicin present in the fermentation broths of the wild strains, the semisynthetic process was developed at a pilot plant scale and eventually was scaled up to the industrial scale, starting in the early 1970s. The publications from the Istituto Nazionale Tumori, Milan, concerning laboratory26 and clinical27 results aroused high interest, and the efficacy of doxorubicin in a spectrum of solid tumors was highlighted in an International Symposium held in Milan in September 1971. Early experience with doxorubicin in the treatment of human cancer diseases was reviewed by S.K. Carter.28 C. SECOND-GENERATION ANTITUMOR ANTHRACYCLINES Daunosamine, 3-amino-2,6 dideoxy-L-lyxo-hexose,17 is the natural aminosugar present in all anthra- cycline glycosides produced by the strains belonging to the species S. peucetius — the original producer of the daunorubicin group of antibiotics — and related strains.11 Daunosamine has never been identified as a constituent of other anthracyclines or other natural products. An early approach to new analogs consisted of the synthesis of related aminosugars, differing from daunosamine in stereochemistry or in structure, and of the coupling of the same with the aglycone moiety, with the aim of modifying the pharmacokinetics or metabolism of the natural glycosides. This approach afforded a number of compounds with a wide variation in bioactivity, out of which the analog of Copyright © 2005 CRC Press, LLC O OH O O OH O O OH O R 9 CH2OH OH14 OH 7 4 OH O O OH O O O OH O O OH O O O O HO 4' OHNH2 NH2 OHNH2 Ia: R=OH II III Ib: R=H STRUCTURES I–III structure II, containing the aminosugar acosamine, namely, 3-amino-2,6-dideoxy-L-arabino-hexose, was selected for further development because of a significantly higher chemotherapeutic index when compared with doxorubicin.4 The compound, known by the generic name of epirubicin, exhibited superior performance when compared with doxorubicin in experimental models and in clinical studies and was registered in most countries in the 1980s as Farmorubicin® or Pharmoru- bicin®. However, registration in the United States was obtained only more recently, and the com- pound is currently marketed in this country as Ellence®. Daunomycinone, the aglycone of daunorubicin, is characterized by the tetracyclic trihydroxy- anthraquinone chromophore typical of natural anthracyclinones, albeit containing specific structural features that identify the daunomycin group of metabolites; namely, the ketone function in the C9 side chain, the absence of a substituent at C10, and the presence of a methoxyl group instead of a hydroxyl at C4. In the closely related carminomycin, the C4 hydroxyl is not methylated. As methods for the total synthesis of daunomycinone and related compounds were made available, it is reason- able that analogs showing no substitution at C.4 were prepared and tested. In fact, their synthesis was easier in comparison with that of the natural compounds because of the absence of regioisomers in the construction of the anthraquinone system. The first compound synthesized in this series, 4- demethoxydaunorubicin (III) (idarubicin), has been developed to the clinical stage because of its powerful activity in experimental leukemia models and reduced cardiotoxicity. Idarubicin exhibits features rendering this compound unique among clinically available anthracyclines. The results of preclinical studies indicate that the higher lipophilicity of III leads to faster accumulation in the nuclei, superior DNA-binding capacity, and consequently greater cytotoxicity compared to Ib. A major advantage over the 4-methoxylated parent drug is its ability to partially overcome multidrug resistance. Its major metabolite, idarubicinol, is as active as the parent compound.29 Idarubicin is marketed with the trademark Zavedos®. As seen below, a 4-demethoxy aglycone has become a favorable structural feature in more recently designed analogs. III. MECHANISM OF ACTION Anthracyclines, as with several other classes of antitumor agents, work by DNA intercalation, a specific mode of binding in which the drug chromophore is inserted between adjacent DNA base pairs.30 However, the mechanism of action cannot be described in such a simple way, and the picture appears to be rather more complex. A. DNA COMPLEX AND TOPOISOMERASE II POISONING The binding of anthracyclines to DNA has been extensively studied by a variety of biochemical and biophysical techniques.31,32 The drug intercalates at a pyrimidine–purine step with a preference for the cytidine-guanosine (CG) sequence and acts as a groove binder, as the sugar moiety interacts with the minor groove.33–35 DNA binding is a necessary but not sufficient condition for drug activity. Copyright © 2005 CRC Press, LLC In other words, there is no simple relationship between DNA binding affinity and cytotoxicity, and a number of different mechanisms have been proposed. Briefly, poisoning of topoisomerases is likely to be the mechanism most closely associated with their cytoxicity, although other molecular inter- actions may play an important role as well. Topoisomerases are enzymes that are involved in the topological interconversions of DNA during DNA transcription and replication; their functions are numerous and important because they are critical for DNA functions and cell survival. It was found that the DNA breaks induced by doxorubicin were associated with the enzyme topoisomerase II.36 In a study including different anthracyclines, it was found that for all bioactive anthracyclines in the entire SV40 DNA a single region of prominent cleavage sites lies between nucleotides 4237 and 4294. A good correlation was recorded between cytotoxicity and intensity of topoisomerase II–medi- ated DNA breakage.37 All of the doxorubicin-stabilized sites had an adenosine at the 3 terminus of at least one member of each pair of strand breaks that would constitute a topoisomerase II double- strand scission. Conversely, none of the enzyme-only sites had an adenosine simultaneously at the corresponding positions on opposite strands. The requirement of a 3-adenosine for doxorubicin- stabilized cleavage is therefore incompatible with enzyme-only cleavage and explains the mutual exclusivity of the two classes of sites.38 The strength of DNA binding does not correlate with the stimulatory effect of anthracyclines on topoisomerase II–mediated DNA cleavage and with the cytotoxic potency. However, drug intercalation is still required for optimal drug activity, indicating that the specific mode of DNA interaction, rather than the strength of binding, is important in determining the cytotoxic potency.39 Again, on the basis of a study including different semisynthetic doxorubicin analogs, it was concluded that cytotoxic potency of anthracyclines might be the result of an interplay of different factors, namely, the level, the persistence, and the genomic localization of topoisomerase II–mediated DNA cleavage.40 In conclusion, the main steps of the mode of action of antitumor anthracyclines would be the formation of a drug-DNA-enzyme ternary complex, in which the enzyme is covalently linked to the broken DNA strand, and the consequent DNA damage consisting of protein-associated double-strand breaks that trigger apoptotic cell death.41,42 B. NATURAL AND ACQUIRED RESISTANCE TO ANTHRACYCLINES Drug resistance is a major problem in cancer chemotherapy, and it is also clearly apparent in the case of antitumor anthracyclines. A number of important tumors of great clinical importance do not respond to currently available anthracycline drugs. These tumors include colon cancers, lung cancers, pancreatic and renal cancers, and malignant melanoma, to cite only some examples of “naturally resistant” tumors. Other diseases, such as gastric and small cell lung cancers and advanced ovary and breast tumors, are only partially responsive, and often the benefit of drug treatment is marginal. So far, only the classic multidrug resistance phenotype, which is a result of the presence of P-glycoprotein in the plasma membrane (a “pump” that can extrude a wide range of anticancer drugs), has been shown to contribute to resistance in clinical conditions. The other possible mechanisms of resistance are a non-P-glycoprotein-mediated multidrug resistance, a phenotype characterized by expression of other proteins in the

plasma membrane that are also able to extrude anticancer drugs; changes in the intracellular distribution of the drug; glutathione transferases and detoxification mechanisms; low levels of expression or alterations in topoisomerase II; and finally, an increased DNA repair. However, evidence for these mechanisms of resistance in cancer patients is still missing.43 A relationship was found between doxorubicin resistance and glutathione transferase activity, but not between multidrug resistance–associated protein and P-glycoprotein, in human colon ade- nocarcinoma cell lines,44 and enzyme-catalyzed formation of the adduct, 7-deoxy-7-S-glutathionyl- adriamycinone, may be relevant to tumor cell resistance to Adriamycin.45 Introduction of HER-2 oncoprotein gene in vitro induces resistance to several anticancer drugs, including taxanes, cisplatin, and 5-fluorouracil, in breast cancer cells. Interestingly, in clinical studies, patients with HER-2 oncoprotein overexpression, an important prognostic factor associated with a poor prognosis in breast Copyright © 2005 CRC Press, LLC cancer, responded better to an anthracycline-based regimen than did patients with low HER-2 expression, and their overall survival was also superior. In addition, treatment with trastuzumab, an anti-HER-2 antibody, increased drug sensitivity to anthracyclines.46 C. CARDIOTOXICITY Together with the onset of resistance, a major dose-limiting factor in the repeated treatment with doxorubicin is the development of cardiotoxicity at cumulative dosages higher than 400 mg/m2. In the mouse, the most prominent and persistent myocardial change induced by repeated doxorubicin administrations was an increased reactivity of the sarcoplasmic reticulum with the zinc iodide–osmium tetroxide reagent that appeared to be related with the observed cardiotoxic effects.47 The zinc iodide–osmium tetroxide reagent allows detection of disulfide bonds in polypeptides, cystine, and oxidized glutathione, substances that react more promptly than other potential sub- strates such as unsaturated lipids and catecholamines. This observation indicates the presence of an oxidative stress that, as already demonstrated by different authors,48,49 may be caused by a higher then normal oxygen tension in the heart tissue. Excess oxygen could be the result of the markedly long-lasting inhibition of nucleic acid synthesis found in the heart tissue of laboratory animals treated with the drug,50 as well as a consequent reduction of metabolic consumption of oxygen. Also, inhibition by doxorubicin of mitochondrial DNA transcription leads to a diminished ATP production and consequently to cell damage in high-energy-demanding tissues.51 In any case, current evidence appears to be consistent with a fundamental role for free radicals, whatever their origin, in the cardiotoxic actions of the anthracyclines.41 IV. SYNTHESIS OF THE DRUGS A. DOXORUBICIN Doxorubicin Ia is prepared by semisynthesis starting from biosynthetic daunorubicin Ib, currently obtained in bulk quantities by fermentation of different Streptomyces strains, mostly derived from S. peucetius. The reaction sequence comprises an electrophilic bromination of Ib followed by an acid treatment to give the 14-bromoketone that is converted to Ia via the intermediate 14-O-formate.4 As with all commercial anthracyclines, the compound is crystallized to higher than 98% purity as the hydrochloride and is formulated in freeze-dried form with addition of lactose. Doxorubicin ready-to-use solutions, both with and without stabilizers, have been patented.52,53 B. EPIRUBICIN Epirubicin (II) is prepared by a similar reaction sequence starting from 4-epidaunorubicin, originally prepared by the glycosidation of daunorubicin aglycone, daunomycinone, with N,O-di-trifluoroacetyl- L-acosaminyl chloride and silver triflate, followed by removal of the protecting trifluoroacetyl groups.4 A direct conversion of biosynthetic daunorubicin to the 4-epi derivative VII has also been carried out.54 It involves oxidation at C4 of N-trifluoroacetyl derivative IV to ketone V and selective reduction of the latter with sodium borohydride to VI, followed by deprotection with base (Scheme 16.1). This procedure is the basis of current industrial production of II. C. IDARUBICIN Idarubicin (III) was originally prepared by glycosidation of 4-demethoxydaunomycinone (idaru- bicinone), which was obtained by total synthesis using a modification of the Wong procedure.4 A stereoselective synthesis of Wong’s tetralin has been published very recently.55 In contrast, a semisynthetic process useful for the industrial preparation of idarubicinone starting from dauno- mycinone VIII, easily prepared on acid hydrolysis of biosynthetic daunorubicin, has been carried Copyright © 2005 CRC Press, LLC SCHEME 1 O OH O O OH O OH OH a,b c lb O O OH O O O OH O O O NHCOCF3 OH O NHCOCF3 IV V O OH O O OH O d e,f OH OH g,h,i,j II O O OH O O O OH O O O HO HO NHCOCF NH 3 2.HCI VI VII a: Tf2O b: MeOH, CHO - 3 c: DMSO, Tf2O, DBN d: NaBH2 e: OH- f: HCI/MeOH g: Br2/MeOH, dioxane h: HBr/H2O i: HCO2Na, H2O j: HCI/MeOH SCHEME 16.1 Synthesis of epirubicin. out.56 As shown in Scheme 16.2, this process involves demethylation of VIII to carminomycinone IX and the conversion of the latter to the 13-ethylidene ketal X and then to the 4-O-triflate XI. The subsequent step, leading to 4-deoxy derivative XII, is based on the deoxygenation reaction of phenols discovered by Cacchi.57 Removal of the protecting group provides idarubicinone XIII, which is glycosylated to III. The method is particularly interesting because daunorubicin is also a source of the aminosugar daunosamine — the other starting material for the synthesis of III. V. MEDICINAL CHEMISTRY OF ANTITUMOR ANTHRACYCLINES A. ANALOGS MODIFIED IN THE AGLYCONE MOIETY 1. Deoxyderivatives The substitution of a hydroxyl group with a hydrogen atom has been extended to positions 6 and 11, as specific interest in the C6 and C11 deoxyderivatives stems from the hypothesis that antitumor activity or cardiotoxicity of anthracyclines might be related with the redox behavior or with the ability to form metal complexes. Synthesis of the corresponding daunorubicin and doxorubicin analogs led to the conclusion that such modifications, although bearing noticeable consequences on the polarographic behavior or on the chemical outcome of anaerobic quinone reduction, did not result in a marked change of bioactivity.59 A significant structural modification in terms of structure–activity relationships was instead the substitution of the C9 hydroxyl group on ring A with a hydrogen atom. In fact 9-deoxy- daunorubicin, although showing the same affinity for DNA and similar electrochemical behavior as the parent daunorubicin, was two orders of magnitude less cytotoxic on cultured HeLa cells. Therefore, it Copyright © 2005 CRC Press, LLC SCHEME 2 O OH O O OH O O OH O O 4 OH a OH b OH O O OH O OH O OH OH OH O OH OH VIII IX X c O OH O O OH O O O OH O O OH e OH d OH O OH OH O OH OH OTf O OH OH XIII XII XI a: AICI3, CH2CI2, reflux b: (CH2OH)2, cat. PTSA, toluene/i-PrOH, reflux c: iPr2EtN, DMAP, (CF3SO2)2O, Py d: Pd(OAc)4, 1,1'-bis(diphenylphosphine)ferrocene, HCO2H, Et3N, DMF, 40° e: CF3CO2H, 0° SCHEME 16.2 Synthesis of idarubicinone. was deduced that the 9-hydroxyl group, which according to the well-known molecular model of the daunorubicin–DNA complex protrudes outside the double helix, might be involved in a specific inter- action relevant to the bioactivity of the drug in terms of topoisomerase II blockade in the ternary complex.60 2. 8-Fluoroderivatives According to the established importance of the 9-hydroxyl group mentioned above, a study aimed at the improvement of doxorubicin anticancer activity via introduction of a fluorine atom at the α position with respect to the 9 position has been carried out. The derivatives substituted at C8 and at C14 (the latter in the doxorubicin series) resulted in the most interesting compounds. The substitution of a hydrogen atom with a fluorine at position 8 gave rise to compounds with markedly different chemical and biological properties depending on the stereochemistry at the new chiral center. 8(S)- Fluoro aglycones XIVa and XIVb were obtained from 8,9-oxirane intermediates. These where either prepared by total synthesis when R = H or, in the case of XIVb, in enantiomeric (8S) form by semisynthesis from “pink compound” XV via the corresponding 7,8-epoxide, XVIa, and finally XVIb. Glycosidation of XIVa and XIVb, followed by diastereomeric separation when R = H, afforded glycosides XVIIa and XVIIb, respectively. Compound XVIIb was eventually converted to 8(S)- fluorodoxorubicin, albeit in low yields because of an inertness toward electrophilic bromination at C14 (possibly a sign of the strong electrostatic effect of the fluorine substitution). A significantly lower cytotoxic potency exhibited by XVIIa compared to unsubstituted idarubicin was related to the β conformation of ring A, predominating in the solution of XVIIa over the α conformation typical of idarubicin and the other antitumor anthracyclines. To obtain the corresponding epimeric cis 8R,9S- fluorohydrins, Diels-Alder adduct XVIII was treated with N-bromosuccinimide and tetrabutylam- monium dihydrogentrifluoride to give, together with a 8-bromo-9-fluoro derivative, the desired 8- fluoro-9-bromo isomer XIX that was converted to the 9,13-epoxide with base. Opening of the oxirane ring with acid and oxidation of the side chain alcohol gave a ketone that was functionalized at C7 via radical bromination and hydrolysis to give the desired 8,9-cis-fluorohydrin XIVc. Copyright © 2005 CRC Press, LLC STRUCTURES XIV–XXI Glycosidation of the latter, followed by separation of diastereomeric glycosides and deprotec- tion, gave 8(R)-fluoroidarubicin XVIIc, which was found to be as potent as idarubicin in the inhibition of the growth of tumor cell lines in vitro. Interestingly, XVIIc displayed marked cytotoxic activity against doxorubicin-resistant cell lines (Table 16.1). When delivered intravenously at a dosage of 1.5 mg/kg to immunodepressed mice bearing A2780 human ovarian carcinoma xenografts, the new compound inhibited tumor growth by 93%, whereas in the same system, doxorubicin at 7 mg/kg inhibited by 83% and idarubicin at 1.2 mg/kg inhibited by only 36%. Values of “log cell kill” (difference in mean time required for tumors of treated and control mice to reach 2 g divided by the product of mean doubling time of the tumor and by 3.32) were equal to 3 for the 8R-fluoro derivative, and 1.8 for doxorubicin.61,62 3. 14-α-Fluorohydrins Synthesis of 14-fluorodoxorubicin XXI did not appear to be an easy task. However, this α- fluorohydrin, a compound with an unusual atomic arrangement, has been obtained starting from daunomycinone VIII, brominated to known XXa, whose treatment with AgBF4 and dimethylsul- foxide gave, interestingly, fluoroderivative XXb. Bromination and reaction of compound XXc with AgBF4 afforded XXd. Desired XXI was eventually obtained through glycosylation of XXc with 1-p-nitrobenzoyl-3,4-N,O-diallyloxycarbonyl daunosamine and treatment of the resulting protected glycoside with the AgBF4-DMSO reagent, followed by removal of the protecting groups. Bioactivity of XXI is currently under investigation.63 Copyright © 2005 CRC Press, LLC TABLE 16.1 Inhibition of Growth of Human Tumor Cell Lines by Doxorubicin (Ia), Idarubicin (III), and Epimeric 8(S) and 8(R)-Fluoroidarubicins (XVIIa and XVIIc, Respectively) Cell line I III XVIIa XVIIc A2780a 13 2.8 14 5 A2780/DX 250 10 60 9 LOVOb 41 3 12 3 LOVO/DX 989 20 147 11 MCF-7c 13 2 6 3 MCF-7/DX 2460 27 291 16 POVDd 37 4 6 9 POVD/DX 350 48 70 9 POGBd 16 2 20 2 POGB/DX 390 24 140 4 Note: Results are expressed as IC50 values (ng/ml) (adjusted from Ref. 62). a Ovarian tumor. b Colon tumor. c Breast tumor. d Small cell lung cancer. B. ANALOGS MODIFIED IN THE SUGAR MOIETY 1. Monosaccharide Analogs 4-Deoxydaunosamine was obtained from daunosamine and used for the preparation of 4-deoxy- doxorubicin (esorubicin) — a compound showing activity comparable to that of the parent, doxorubicin — and lower cardiotoxicity in experimental studies5 (albeit not confirmed in clinical phase II studies64). Similarly, 4-iodo-4-deoxydoxorubicin, showing promising results in preclinical studies,65 was also developed up to clinical phase II but did not exhibit therapeutic properties suggesting improved anticancer activity.66 Instead, the compound was found to be the lead for a novel class of drugs for amyloid-related diseases.67,68 A doxorubicin derivative named pirarubicin, namely, 4-O-dihydropyranyl-doxorubicin, was studied in Japan and has entered into clinical use as a doxorubicin analog endowed with lower side effects.9 Annamycin XXII, a lipophilic iodinated analog, devoid of the sugar amino group and of the methoxy group at C4, was shown, especially when entrapped in liposomes, to be partially effective in cell cultures and in mouse tests against O OH O OH OH O OH O O 2' HO OH I XXII STRUCTURE XXII Copyright © 2005 CRC Press, LLC O OH O OH OH O O OH O O OHN O O XXIII STRUCTURE XXIII doxorubicin-resistant tumors.69 However, no efficacy was found in breast cancer patients not responding to doxorubicin.70 An interesting group of analogs is represented by the morpholinyl anthracyclines. These new derivatives were discovered by Acton et al. and found to be

about two orders of magnitude more cytotoxic than the known antitumor anthracyclines — very likely owing their high potency to an alkylating mechanism of action.71 Methoxymorpholinyl derivative XXIII was developed up to the clinical stage, but a low response rate was observed in patients with tumors endowed with intrinsic resistance to chemotherapy.72 A different, not DNA-related, bio- chemical mechanism of action has been suggested for the very lipophilic doxorubicin derivative, N-benzyldoxorubicin 14-O-valerate.73 2. Disaccharide Analogs More recently, attention has been given to analogs containing two sugar residues. Anthracycline disaccharides have been obtained both as biosynthetic products or by semisynthesis. The disaccha- ride 4-O-daunosaminyldaunorubicin was present in the medium of the daunorubicin fermentation, and the same, as well as the corresponding analogs in which the second sugar was acosamine or 2-deoxy-L-rhamnose, was also obtained by semisynthesis and found to be cytotoxic, albeit less potent against murine transplantable leukemias than daunorubicin.4 Other related glycosides of daunorubicin have been synthesized74,75; however, the first bearing a nonaminated sugar directly linked to the aglycone, namely, 7-O-(3-O-α,L-daunosaminyl-2-deoxy- α,L-rhamnosyl)-daunomyci- none, was biologically inactive, apparently because of the different position of the second glycosidic linkage.76 In fact, as suggested by the DNA binding geometry and by the conclusions concerning the topoisomerase-related mechanism of action of antitumor anthracyclines, a rational design of disaccharide analogs might be based on the hypothesis that it would be beneficial to increase the molecular recognition properties through a functional enrichment of the portion of the anthracycline molecule involved in the stabilization of the “cleavable ternary complex.” Structure activity rela- tionships pointed to the importance of ring A as a “scaffold” determining the orientation of the C7 and C9 substituents, in particular that of the sugar moiety, whose structure and stereochemistry are critical for the stabilization of the ternary complex.77,78 Synthesis of disaccharide derivatives of the daunorubicin/doxorubicin and of the 4-demethoxy (idarubicin) series in which the first sugar moiety linked to the aglycon was a nonaminated sugar, namely, 2-deoxy-L-fucose or 2-deoxy-L-rhamnose, and the second moiety was either daunosamine or acosamine or a 3-hydroxy or 4-deoxy analog of these aminosugars, was performed starting from the corresponding aglycon that was glycosylated with the appropriately protected and activated disaccha- ride.79 The desired final products were obtained by sequential deprotections and purification. Disac- charide glycosides containing 2-deoxy-L-fucose XXIVa,b exhibited superior pharmacological prop- erties in respect to the stereoisomers XXVa,b containing 2-deoxy-L-rhamnose.80 7-O-(α-L- Daunosaminyl-α(1-4)-2-deoxy-L-fucosyl)-4-demethoxyadriamycinone (XXIVc) was synthesized as shown in Scheme 16.3 and was developed to the clinical stage because of its favorable pharmacological Copyright © 2005 CRC Press, LLC O OH O O OH O CH2R' OH OH R O OH O R O OH O O O O OH O OH O O OHNH2 OHNH2 XXIVa: R=H; R'=H XXVa: R=H XXIVb: R=OCH3; R'=H XXVb: R=OCH3 XXIVc: R=H; R'=OH STRUCTURES XXIV–XXV properties. The high activity of XXIVc was related to the activation of p53-independent apoptosis.81 The new compound exhibited, especially at selected DNA sites, a more marked topoisomerase II–mediated cleavage that was accompanied by a superior antitumor efficacy in the experimental systems when compared with doxorubicin. Very significant activity was found against a spectrum of human tumors such as breast, ovary, and lung cancer xenografts in athymic nude mice, where the SCHEME 3 OR O OCO2AII OH pNBzO PhS O O O NHCO2AII a b pNBzO NHCO2AII NHCO2AII pNBzO pNBzO XXIX (R=pMBn) c XXX (R=H) XXVII d XXVIII XXXI (R=pNBz) e O OH O O OH O OH O OH f,g OH O O OH O O OH O O O O OH OCO2AII O O O OHNH2 NHCO2AII OpNBz XXIVc XXXII a: PhSTMS, TMSOTf, CH2CI2 b: p-methoxybenzyl 3-O-allyloxycarbonyl–2-deoxy-L-fucoside, IDCP, CH2CI2, Et2O c: Ce(NH4)(NO3)6, MeCN, H2O d: pNBzCI, py e: 14-acetoxyidarubicinone, TMSOTf, MS, CH2CI2, Et2O, –10° (R=pNBz) f: 0.5M K2CO3 g: (Ph3P)Pd, Me2NTMS, AcOTMS, CH2CI2 SCHEME 16.3 Synthesis of 7-O-(-L-daunosaminyl-(1-4)-2-deoxy-L-fucosyl)-4-demethoxyadriamycinone. Copyright © 2005 CRC Press, LLC compound appeared markedly superior to doxorubicin in inhibiting tumor growth and in terms of an increased number of disease-free survivors among treated animals.82,.83 The generic name selected for XXIVc by World Health Organization is “sabarubicin.” Spectrophotometric and fluorescence titrations with calf thymus DNA,85 and with the d(CGATCG) hexamer of XXIVc, have revealed spectral patterns very similar to those obtained with doxorubicin, implying that the binding mechanism and the stability of the resulting com- plexes are nearly the same. X-ray diffraction data of the orange–red crystals of the sabarubi- cin–hexamer complex86 allowed the definition of the asymmetric unit containing two DNA strands forming one duplex, two bound drugs, and 35 solvent molecules. Similar to the well-known complexes of other antitumor anthracyclines with the same DNA hexamer, intercalation occurs at each end of the DNA duplex in the CpG steps, with the drug chromophoric system perpendicular to the long axis of stacked base pairs. However, the two drug molecules appear to have different conformations, and the two binding sites show large differences. In one site, both sugar rings lie in the minor groove, whereas in the other site, the arrangement of the drug is such that the first sugar resides in the minor groove and, although rotated by about 45˚ around the O7- C1 bond with respect to the corresponding doxorubicin complex, is in the chair conformation with the 3- hydroxyl group still in van der Waals contact with the guanine residue. The second sugar protrudes out in the solvent region where the amino and hydroxyl groups interact with two water molecules (Figure 16.1). Interestingly, a tight interaction takes place between the amino group of the disaccharide moiety and a guanine residue of a second DNA molecule, different from the one where the drug is intercalated. As a consequence, in the crystals, the end-over-end stacked duplexes do not run parallel, as in other crystallized anthracycline-DNA complexes, but are crossing each other at right angles along the tetragonal c axis, leading to the formation of two layers of stacked duplexes mutually perpendicular to one another. The packing contacts include the amino group on the second sugar FIGURE 16.1 (See color insert) Representation of the sabarubicin intercalation complex with d(CGATCG)2 according to a x-ray crystal diffraction study. The drawing was kindly provided by Dr. Giovanni Ughetto, Laboratorio del CNR di Montelibretti (Roma). Copyright © 2005 CRC Press, LLC O OH O OH O O OH O O NH2 OH S pTpGpGpGpTpGpGpGpTpGpGp(CH2)3OH (5') (3') XXVI STRUCTURE XXVI that is directly hydrogen bonded to N7 and, through a bridging water, to O6 of a guanine, providing the first example of an anthracycline/DNA complex in which a cross-link with a second DNA helix has been observed. It seems, therefore, that the second sugar may exert interactions with other cellular targets in close proximity with DNA. C. ANTHRACYCLINE–TRIPLEX HELIX-FORMING OLIGONUCLEOTIDE CONJUGATES Triplex helix-forming oligonucleotides (TFOs) that bind to homopurine:homopyrimidine sequences in DNA in a sequence-specific manner can be used to selectively inhibit gene expression and provide the opportunity for rational design of novel DNA sequence–specific cancer therapeutics. Chemical modifications of the TFO that would enhance triplex stability without affecting sequence specificity, such as conjugation with an anthracycline molecule, further increase the potential efficacy of this approach.87 The resulting conjugates may correctly position the TFO moiety in the major groove of the duplex and allow it to form Hoogsteen hydrogen bonds with the complementary bases of the purine-rich strand of the target DNA while allowing ring A of the intercalated anthracycline, together with the sugar moiety, to protrude into the minor groove, as in the well- known model of the daunomycin–DNA complex. When the aminosugar is lacking (i.e., the TFO is conjugated with a nonglycosylated anthracyclinone), the stabilization of the triple helix is still obtained, albeit one order of magnitude lower.88 In a recent study, DNA binding and antigene activity of anthracycline-conjugated TFOs targeted to a sequence near the major P2 promoter in the c-myc gene were synthesized. GT-rich 11-mer TFOs were covalently attached through a hexamethylene bridge at the hydroxyl on position 4 of carminomycin. As shown by footprinting and electrophoretic experiments, as well as by spectroscopic measurements, conjugate XXVI had high binding affinity and formed a much more stable triplex than the nonconjugated TFO. Binding of XXVI was also highly sequence specific, as indicated by the fact that it protected a single region corresponding to the 11-basepair target sequence in the c-myc promoter. Conjugates with different TFOs, which were partially complementary to the c-myc target sequence, bound with much lower affinity than XXVI. The latter inhibited transcription from the P2 promoter in in vitro transcription assays and also inhibited transcription in luciferase reporter assays when cotransfected into prostate cancer cells along with a pGL-myc promoter reporter plasmid, but it did not affect the activity of the control pRL-SV40 plasmid. These data indicate that attachment of the anthracycline to the TFO increased its binding affinity and allowed formation of a very stable complex with duplex DNA. Moreover, XXVI was able to selectively inhibit transcription of the c-myc gene both in vitro and in cells. These results encourage further testing of this approach for the development of novel antigene cancer therapeutics.89 Copyright © 2005 CRC Press, LLC VI. CLINICAL ACTIVITY OF ANTHRACYCLINES Since the first clinical trials, daunorubicin and doxorubicin have gained rapid acceptance as major therapeutic agents in the medical treatment of cancer. Both drugs, but especially doxorubicin, have been widely used as part of combination regimens for the effective treatment of acute nonlympho- cytic leukemia, Hodgkin and non-Hodgkin lymphomas, breast cancer, and sarcomas.90 Epirubicin, idarubicin, and liposomal doxorubicin offer less toxic and, in some instances, more effective alternatives to older anthracyclines for leukemia, breast cancer, ovarian cancer, and other diseases.91 A. DAUNORUBICIN Daunorubicin (Ib), the first anthracycline introduced into therapeutic use, is endowed with a potent antitumor activity whose spectrum is narrower than that of doxorubicin and is accompanied by a significant hematological toxicity. The compound exhibits delayed cardiotoxicity. Main indications are acute leukemias, but significant activity was recorded against neuroblastomas and rhabdomy- osarcomas. Induction of remission in over 70% of patients with acute lymphoblastic leukemia is obtained on treatment with daunorubicin, vincristine, and prednisone, with disease-free survival at 5 yr in up to 45% of patients receiving a consolidation/intensification therapy after remission.92 Intensive induction chemotherapy using 7-d courses of cytarabine and daunorubicin or amsacrine produces remission in 60%–85% of patients with acute myelogenous leukemia, a condition that was invariably fatal 30 yr ago. Median remission duration is 9–16 mo. In some series, 20%–40% of patients are in continuous remission for 2 yr or more; many of these patients remain in remission for 5 yr or longer, and some may be cured.93 B. DOXORUBICIN A notable feature of the antitumor properties of doxorubicin (Ia) is the wide spectrum of activity including (as particularly responsive) acute leukemias, Hodgkin and non-Hodgkin lymphomas, lung microcytomas, soft-tissues sarcomas, Ewing’s tumor, thyroid carcinoma, ovarian cancer, breast cancer, rhinopharyngeal tumors, Wilms’ tumor, and other pediatric malignancies. It was recognized early on that in these diseases, doxorubicin shows efficacy when administered alone as a slow intravenous injection every 3 wk at a dosage of 65 mg/m3, with its most usual administration being, however, in combination with other active agents.94 Out of 113 drug combinations containing doxorubicin (Adriamycin) already used in clinical studies worldwide at the end of the first decade after the original demonstration of therapeutic efficacy of the drug, cyclophosphamide, an alkylating agent, was the most frequent component (54 combinations), followed by the spindle poison vinc- ristine or other vinca alkaloids such as vinblastine (42 combinations), the metabolic inhibitor 5- fluorouracyl (24 combinations), bleomycin and cis-platinum (22 combinations each), and methotr- exate (12 combinations), with other less frequent components being mitomycin, etoposide, or nitrosoureas.95 Doxorubicin was frequently accompanied by cyclophosphamide and vincristine or by 5-fluorouracyl in the treatment of breast cancer (the acronyms used for the main combinations are FAC and VAC). Cyclophosphamide and vincristine or etoposide were the most frequent com- ponents in the treatment of small cell lung tumors, whereas doxorubicin was accompanied by cyclophosphamide and methotrexate or either cis-platinum or mitomycin or a nitrosourea in the case of the less responsive non–small cell lung cancer. In the low respondent gastrointestinal tract malignancies, doxorubicin was combined with 5-fluorouracyl and mitomycin or a nitrosourea and with cyclophosphamide, a vinca alkaloid, bleomycin, and prednisone in the generally more respon- sive lymphomas. Cis-platinum and cyclophosphamide were mainly associated with doxorubicin in the treatment of genital

tract tumors. Doxorubicin is characterized by a rapid distribution phase and a slow elimination phase. The successive half-lives of doxorubicin in plasma are approximately 5 min, 1 h, and 30 h. Total plasma clearance is about 30 L/h/m2, and its total volume of distribution is approximately 15 L/kg. As the Copyright © 2005 CRC Press, LLC other anthracyclines, the drug is excreted mostly through the bile, and special care is taken in patients with hepatic dysfunction.96 Doxorubicinol, the product of the reduction of the 13-keto group to a secondary alcohol function, is the major human metabolite of doxorubicin, albeit present in plasma at concentrations well below those of the parent drug.97 C. EPIRUBICIN Phase II studies of epirubicin (II) indicated that the drug produces a pattern of acute toxicity qualitatively similar to but quantitatively lower than that of doxorubicin at identical doses, the most frequently employed dose schedules being 75 or 90 mg/m2 intravenously every 3 wk. Epirubicin showed activity in a variety of tumors such as breast carcinoma, soft tissue sarcomas, lymphomas, leukemias, ovarian cancer, and gastric cancer, and preliminary evidence of activity was found in melanoma, rectal cancer, and pancreatic cancer, indicating a broad spectrum of activity. As to chronic cardiac toxicity, preliminary findings from a randomized comparison of the two drugs in breast cancer indicated that epirubicin might have a lower cumulative cardiotoxicity than doxorubicin. According to the first clinical studies, pharmacokinetics of epirubicin in human patients showed a multiexponential decrease of plasma levels, with the same pattern being observed for the metabolite epidoxorubicinol, together with a high plasma clearance (0.9–1.4L/min), a terminal half-life of about 30–40 hr, and a large volume of distribution.98 According to a subsequent review,99 to achieve hematological equitoxicity of the two drugs, the dose of epirubicin should be approximately 20% higher than that of doxorubicin, giving rise to a higher dosage level for epirubicin. In contrast, epirubicin is significantly less cardiotoxic than doxorubicin. Thus, the maximum cumulative dose, which for doxorubicin is 500 mg/m2, for epirubicin is close to 1000 mg/m2. More recent conclusions indicate that epirubicin, used alone or in combination with other cytotoxic agents in the treatment of a variety of malignancies, is better tolerated than, and achieves equivalent objective response rates and overall median survival to, similar doxorubicin-containing regimens in the treatment of advanced and early breast cancer, non–small cell lung cancer, small cell lung cancer, non-Hodgkin lymphoma, ovarian cancer, gastric cancer, and nonresectable primary hepatocellular carcinoma. Dose-intensive epirubicin-containing regimens, which have been designed because of the drug’s lower myelosuppression and cardiotoxicity and the definite dose–response relationship of anthra- cyclines, have produced high response rates in early breast cancer, a potentially curable malignancy, as well as advanced breast and lung cancers. Furthermore, such improved response rates have improved quality of life in some clinical settings, but whether this leads to prolonged survival is yet to be determined.100 The metabolism of epirubicin is characterized by a biotransformation to relatively or totally inactive metabolites, including a 13-dihydro derivative, epirubicinol, two glu- curonides, and four aglycones. Quantitatively, the glucuronides of epirubicin and epirubicinol are very important, and this pathway, which is unique to epirubicin metabolism in humans, might explain the better tolerability of this drug compared with doxorubicin. Epirubicin pharmacokinetics may be described by a three-compartment model, with median half-life values of 3.2 min, 1.2 h, and 32 h for each phase. Total plasma clearance is 46 L/h/m2, and the volume of distribution at steady state is 1000 L/m2. The pharmacokinetics of epirubicin appear to be linear for doses up to 150–180 mg/m2, which is the maximal tolerated dose.101 D. IDARUBICIN Idarubicin (III) is currently used in the medical treatment of acute nonlymphocytic leukemia in adults. Idarubicin is at least as effective as daunorubicin in this disease, and two studies indicate greater activity and longer survival with an idarubicin–cytarabine regimen than with a daunorubi- cin–cytarabine regimen. The drug is active in acute lymphocytic leukemia, lymphomas, and breast cancer. Reportedly, idarubicin is more potent and less cardiotoxic than daunorubicin or doxorubicin. However, a maximum cumulative dose, to prevent chronic cardiomyopathy, has not yet been defined. Unlike the other clinically useful anthracyclines, idarubicin has significant oral bioavailability Copyright © 2005 CRC Press, LLC (average 28%), and an oral dosage form has been developed.102 The oral formulation has demon- strated efficacy in advanced breast cancer, non-Hodgkin lymphoma, and myelodysplastic syndromes as well as in first-line induction therapy of acute myelogenous leukemia. It also has potential for ameliorating blast crisis of chronic myelogenous leukemia and in multiple myeloma. The most frequent acute adverse effects associated with oral idarubicin are those commonly found with anthracyclines, but there appears to be minimal significant cardiotoxicity. Available data concerning oral idarubicin in hematological malignancies and advanced breast cancer are sufficiently encour- aging to warrant further research.103 E. ACLACINOMYCIN Aclacinomycin (ACM) has been extensively evaluated in patients with relapsed leukemia and advanced malignant lymphoma. Analysis of results compiled from Europe, Japan, and the United States shows that ACM is probably equivalent to doxorubicin for remission induction of patients with relapsed acute nonlymphoblastic leukemia. The antitumor activity of ACM in patients with acute lymphoblastic leukemia or malignant lymphoma who have previously received doxorubicin or daunorubicin is low, and the issue of whether ACM lacks clinical cross-resistance to other anthracyclines is unresolved. Acute cardiac arrhythmias have been observed following administra- tion of ACM, but congestive cardiomyopathy has been uncommon. It has been reported that ACM has fulfilled its early expectations of antileukemic activity and reduced toxicity.104 F. SABARUBICIN Clinical tolerability of weekly dosed sabarubicin (XXIVc) has been determined by researchers at EORTC, who reported the results of a phase I study and recommended a intravenous dosages of 30 and 40 mg/m2 for three consecutive weeks followed by 1 wk rest in, respectively, pretreated and naïve patients.105 Pharmacokinetics of sabarubicin were studied in patients from two different phase I dose schedules, one in which the drug was administered once every 3 wk (the current schedule of anthracyclines), and the other in which it was administered once every week for 3 wk, followed by 1 wk rest. The plasma peak concentration levels and AUC after each dosage showed linear relation- ship with the administered dose. Plasma concentrations were particularly high, in the ranges of 0.474–0.612 µg/mL for the lowest (4 mg/m2) and 9.538–21.587 µg/mL for the highest (110 mg/m2) dose used. There was no accumulation of the drug before the next infusion in the weekly regimen. The mean elimination half-life was 20.7 h, which is significantly shorter than that for doxorubicin or epirubicin. The volume of distribution, much smaller than that of the said anthracyclines, was 95.6 L/m2 and showed a standard deviation of 43.4 L/m2. In fact, a particularly high interpatient variation of pharmacokinetic parameters was found.106 A dosage of 80–90 mg/m2 every 3 wk was used in poor-prognosis patients with advanced or metastatic ovarian cancer who had received prior chemotherapy. Out of 14 evaluable patients, one partial response and eight stabilizations of the disease were recorded.107 The activity of sabarubicin has been demonstrated acting as a single agent in a poor-prognosis population of patients with lung cancers. The drug was well tolerated when used at an intravenous dose of 80–90 mg/m2 every 3 wk in patients with non–small cell lung cancer, as second line therapy, and with small cell lung cancer as first line therapy, with dose-limiting toxicity being hematological. One partial response and five stable diseases were observed in 16 evaluable patients with non–small cell lung cancer, and one complete response, nine partial responses, and three stable diseases were observed in 21 evaluable, pretreated small cell lung cancer patients.108,109 G. LIPOSOMAL ANTHRACYCLINES Liposomes are closed vesicular lipidic structures that envelop water-soluble molecules. Liposomal anthracyclines have been extensively studied in humans with a variety of cancer types, including Copyright © 2005 CRC Press, LLC one containing daunorubicin and two preparations containing doxorubicin. Results have indicated that liposomal encapsulation may lead to toxicity attenuation while retaining or even enhancing the efficacy of the parent anthracyclines. Approved indications have been achieved for liposomal daunorubicin against AIDS-related Kaposi sarcoma and for liposomal doxorubicin against AIDS- related Kaposi sarcoma and ovarian cancers.110 In particular, pegylated liposomal doxorubicin shows a pharmacokinetic profile characterized by an extended circulation time and a reduced volume of distribution. Clinical studies in humans have included patients with AIDS-related Kaposi sarcoma and with a variety of solid tumors, including ovarian, breast, and prostate carcinomas. The phar- macokinetic profile in humans at doses between 10 and 80 mg/m2 is biphasic, exhibiting an initial phase with a half-life of 1–3 h and a second phase with a half-life of 30–90 h. The AUC after a dose of 50 mg/m2 is reported to be approximately 300-fold greater than that with free drug. Clearance and volume of distribution are drastically reduced (at least 250-fold and 60-fold, respec- tively). The toxicity profile is characterized by dose-limiting mucosal and cutaneous toxicities, mild myelosuppression, and decreased cardiotoxicity compared with free doxorubicin and minimal alopecia. Interestingly, pegylated liposomal doxorubicin represents a new class of chemotherapy delivery system that may significantly improve the therapeutic index of doxorubicin.111 A retro- spective study on 40 patients indicated that pegylated liposomal doxorubicin at a dosage of 40–45 mg/m2 every 4 wk is clinically active and well tolerated in women with metastatic breast cancer.112 REFERENCES 1. Brockmann, H., Anthracyclinone und anthracycline (Rhodomycinone, Pyrromycinone und Ihre Gly- coside), in Forschritte der Chemie Organische Naturstoffe, Zechmeister, L, Ed., Springer, Vienna, 21, 121-182, 1963. 2. O’Hagan, D., The polyketide metabolites, Ellis Horwood, Chichester, 1991. 3. Arcamone, F. et al., Isolamento ed attività antitumorale di un antibiotico da Streptomyces sp., Giorn. Microbiol., 9, 83, 1961. 4. Arcamone, F., Doxorubicin anticancer antibiotics, De Stevens, G. Ed., Academic Press, New York, 1981. 5. Arcamone, F., The development of new antitumor anthracyclines, in Anticancer agents based on natural products models. Cassady, J. and Douros, J. Eds., Academic Press, New York, 1980, chapter 1. 6. Arcamone, F. and Penco, S., Synthesis of new doxorubicin analogs, in Anthracycline and anthracene- dione-based anticancer agents, Lown, J.W. Ed., Elsevier, Amsterdam, 1988, chapter 1. 7. Weiss, R.B., The anthracyclines: will we ever find a better doxorubicin? Semin. Oncol. 19, 670, 1992. 8. Arcamone, F. et al., New developments in antitumor anthracyclines, Pharmacol. Ther. 76, 117, 1997. 9. Takeuchi, T. Antitumor antibiotics discovered and studied at the Institute of Microbial Chemistry. J. Cancer Res. Clin. Oncol., 121, 505, 1995. 10. Oki, T., Antitumor anthracycline antibiotics from microbial origin, in ref. (5), p.103. 11. Arcamone, F. and Cassinelli, G., Biosynthetic anthracyclines, Curr. Med. Chem. 5, 391, 1998. 12. Fujii, I. and Ebizuka, Y., Anthracycline biosynthesis in Streptomyces galilaeus, Chem. Rev., 97, 2511, 1997. 13. Hutchinson, C.R., Biosynthetic studies of daunorubicin and tetracenomycin C, Chem. Rev., 97, 2525, 1997. 14. Hopwood, D.A., Genetic contributions to understanding polyketide synthase, Chem. Rev., 97, 2465, 1997. 15. Di Marco, A. et al., Studi sperimentali sull’attività antineoplastica del nuovo antibiotico daunomicina, Tumori, 49, 203, 1963. 16. Arcamone, F. et al., Daunomycin. I. The structure of daunomycinone, J. Am. Chem. Soc., 86, 5334, 1964. 17. Arcamone, F. et al., Daunomycin. II. The structure and stereochemistry of daunosamine, J. Am. Chem. Soc., 86, 5335, 1964. 18. Arcamone, F. et al., The structure of daunomycin, Tetrahedron Lett., 30, 3349, 1968. Copyright © 2005 CRC Press, LLC 19. Arcamone, F. et al., The total absolute configuration of daunomycin, Tetrahedron Lett., 30, 3353, 1968- 20. Arcamone, F. et al., Struttura e stereochimica della daunomicina, Gazz. Chim. Ital., 100, 848, 1970. 21. Angiuli, R. et al., Structure of daunomycin; x-ray analysis of N-Br-acetyl-daunomycin solvate, Nat. New Biol., 234, 78, 1971. 22. Dubost, M. et al., Rubidomycin, a new antibiotic with cytostatic properties, Cancer Chemother. Rep., 41, 35, 1964. 23. Bernard, J. et al., Recent results in cancer research: rubidomycin, Springer, Berlin, 1969. 24. Arcamone, F. et al., Adriamycin, 14-hydroxydaunomycin, a new antitumor antibiotic from S. peucetius var. caesius. Biotechnol. Bioeng., 11, 1101, 1969. 25. Arcamone, F. et al., Adriamycin(14-hydroxydaunomycin), a novel antitumor antibiotic, Tetrahedron Lett., 13, 1007, 1969. 26. Di Marco, A., Gaetani, M. and Scarpinato, B.M., A new antibiotic with antitumor activity, Cancer Chemother. Rep., 53, 33, 1969. 27. Bonadonna, G. et al., Clinical evaluation of Adriamycin a new antitumor antibiotic, Brit. Med. J., 3, 503, 1969. 28. Carter, S.K., Adriamycin — a review, J. Nat. Cancer Inst., 55,

1265, 1975. 29. Borchmann, P. et al., Idarubicin: a brief overview on pharmacology and clinical use, Int. J. Clin. Pharmacol. Ther., 3, 80, 1997. 30. Wang, A.H.J., Intercalative drug binding to DNA, Curr. Opin. Struct. Biol., 2, 361, 1992. 31. Stutter, E., Schuetz, H. and Berg, H., Quantitative determination of cooperative anthracycline-DNA binding, ref. 5, p. 245. 32. Ughetto, G., X.-Ray diffraction analysis of anthracycline-oligonucleotide complexes, ref. 5, p. 296. 33. Pullman, B., Binding affinities and sequence selectivity in the interaction of antitumor anthracyclines and anthracenediones with double stranded polynucleotides and DNA, ref. 5, p. 371. 34. Chaires, J.B., Herrera, J.E., and Waring, J.M., Preferential binding of daunomycin to 5′ at CG and 5′ at GC sequences revealed by footprinting titration experiments, Biochemistry, 29, 6145, 1990. 35. Frederick, C.A. et al., The structural comparison of anticancer drug-DNA complexes: Adriamycin and daunomycin, Biochemistry, 29, 2538, 1990. 36. Tewey, K.M. et al, Adriamycin induced DNA damage mediated by mammalian DNA topoisomerase II, Science, 226, 466, 1984. 37. Capranico, G. et al., Sequenze selective topoisomerase Ii inhibition by anthracycline derivatives in SV 40 DNA: relationship with DNA binding affinity and cytotoxicity, Biochemistry, 29, 562, 1990. 38. Capranico, G., Kohn, K.W. and Pommier, Y., Local sequence requirements for DNA cleavage by mammalian topoisomerase II in the presence of doxorubicin, Nucleic Acids Res., 22, 6611, 1990. 39. Zunino, F. and Capranico G., DNA topoisomerase II as the primary target of anti-tumor anthracyclines, Anticancer Drug Design, 4, 307, 1990. 40. Binaschi, M. et al., Relationship between lethal effects and topoisomerase II-mediated double-stranded DNA breaks produced by anthracyclines with different sequence specificity, Mol. Pharmacol., 6, 1053, 1997. 41. Gerwitz, D.A., A critical evaluation of the mechanisms of action proposed for the antitumor effects of the anthracyclines antibiotics Adriamycin and daunorubicin, Biochem. Pharmacol., 57, 727, 1992. 42. Pommier, Y., DNA topoisomerase II inhibitors, in Cancer therapeutics, experimental and clinical agents, Techter B.A. Ed., Humana Press, Totowa, NJ, 153, 1997. 43. Nielsen, D., Maare, C. and Skovsgaard, T., Cellular resistance to anthracyclines, Gen. Pharmacol., 27, 255, 1996. 44. Beaumont, P.O. et al., Role of glutathione S-transferases in the resistance of human colon cancer cell lines to doxorubicin, Cancer Res., 58, 947, 1998. 45. Gaudiano, G., Resing, K. and Koch, T.H., Reaction of reduced anthracyclines with glutathione — formation of aglycone conjugates, J. Am. Chem. Soc., 116, 6537, 1994. 46. Kim, R., The role of HER-2 oncoprotein in drug-sensitivity in breast cancer, Oncol. Rep., 9, 3, 2002. 47. Bellini, O. and Solcia, E., Early and late sarcoplasmic reticulum changes in doxorubicin cardiomy- opathy, Virchows Arch. (Cell Pathol.), 49, 137, 1985. 48. Ishikawa, T. and Sies, H., Cardiac transport of glutathione disulfide and S-conjugates. Studies with isolated perfused rat heart during hydroperoxide metabolism, J. Biol. Chem., 239, 3838, 1984. Copyright © 2005 CRC Press, LLC 49. Olson, R.D. et al., Regulatory role of glutathione and soluble sulfidryl groups in the toxicity of Adriamycin, J. Pharmacol. Exp. Ther., 215, 450, 1980. 50. Formelli, F. et al., Effect of Adriamycin on DNA synthesis in mouse and rat heart, Cancer Res., 38, 3286, 1978. 51. Ellis, C.N., Ellis, M.B. and Blakemore, W.S., Effect of Adriamycin on heart mitochondrial DNA, Biochem. J., 245, 309, 1987. 52. Gatti, G. et al., Injectable stabilized solutions containing an antitumor anthracycline glucoside, BP 2178311 (Feb. 11, 1987), Chem. Abs. 107, 223275, 1997. 53. Gatti, G. et al., Ready to use injections containing anthracycline glycosides in acidic solution, Germ. Offen. 3741037 (Jun. 9, 1988), Chem. Abs. 110, 141560, 2000. 54. Suarato, A. et al., Anthracycline chemistry: direct conversion of daunorubicin into the L-arabino, L- ribo, and L-lyxo analogues, and selecrive deoxygenation at C-4, Carbohydr. Res., 98, C1-C3, 1981. 55. Badalassi, F. et al., Efficient enantioselective synthesis of (R)-2-hydroxy-5,8-dimethoxy-1,2,3,4-tet- rahydronaphthalene, the key intermediate in the synthesis of anthracycline antibiotics, Tetrahedron Asymmetry, 12, 3155, 2001. 56. Cabri, W. et al., Process for preparing anthracyclinones, U.S. Patent 5,180,758 (Jan. 19, 1993). 57. Cacchi, S. et al., Palladium-catalyzed triethylammonium formate reduction of aryl triflates, a selective method for the deoxygenation of phenols, Tetrahedron Lett., 27, 5541, 1986. 58. Arcamone, F. and Penco, S., Chemical derivatives of anticancer antibiotics with different DNA binding properties, in Molecular mechanisms of carcinogenic and antitumor activity, Chagas, C. and Pullman B. Eds. Pontificiae Academiae Scientiarum Scripta Varia, Vatican City, 70, 225, 1987. 59. Arcamone, F. and Penco, S., Relationship of structure to anticancer activity and toxicity in anthracy- clines, Gann Monogr. Cancer Res., 36, 81, 1989. 60. Capranico, G. et al., Role of DNA breakage to cytotoxicity of doxorubicin, 9-deoxydoxorubicin, and 4-demethyl-6-deoxydoxorubicin in murine leukemia P388 cells, Cancer Res., 49, 2022, 1989. 61. Pasqui, F. et al., Synthesis of ring A fluorinated anthracyclines, Tetrahedron, 52, 185, 1996. 62. Animati, F. et al., Biochemical and pharmacological activity of novel 8-fluoroanthracyclines: influence of stereochemistry and conformation, Mol. Pharmacol., 50, 603, 1996. 63. Berettoni, M. et al., Synthesis of 14-fluorodoxorubicin, Tetrahedron Lett., 43, 2867, 2002. 64. Ringenberg, Q.S. et al., Clinical cardiotoxicity of esorubicin (4-deoxydoxorubicin, DxDx): prospective studies with serial gated heart scans and reports of selected cases. A Cancer and Leukemia Group B report, Invest. New Drugs, 8, 221, 1990. 65. Barbieri, B. et al., Chemical and biological characterization of 4-iodo-4-deoxydoxorubicin, Cancer Res., 47, 4001, 1987. 66. Sessa, C. et al., Phase II studies of 4-iodo-4-deoxydoxorubicin in advanced non-small cell lung, colon and breast cancers, Ann. Oncol., 2, 727, 1991. 67. Merlini, G. et al., Interaction of the anthracycline 4-iodo-4-deoxydoxorubicin with amyloid fibrils: inhibition of amyloidogenesis, Proc. Natl. Acad. Sci. U.S.A., 92, 2959, 1995. 68. Gertz, M.A. et al., A multicenter phase II trial of 4-iodo-4deoxydoxorubicin (IDOX) in primary amyloidosis (AL), Amyloid, 9, 24, 2002. 69. Priebe, W. et al., Non-cross-resistant anthracyclines with reduced basicity and increased stability of the glycosidic bond, in ACS symposium series 574: Anthracycline antibiotics, new analogues, methods of delivery, and mechanism of action, Priebe, W. Ed., American Chemical Society, Washington, DC, 1995. 70. Booser, D.J. et al., Phase II study of liposomal annamycin in the treatment of doxorubicin-resistant breast cancer, Cancer Chemother. Pharmacol., 50, 6, 2002. 71. Acton, M., Wasserman, K. and Newman, R.A., Morpholinyl anthracyclines, ref. 6, chapter 2. 72. Bakker, M. et al., Broad phase II and pharmacokinetic study of methoxy-morpholino doxorubicin (FCE 23762-MMRDX) in non small cell lung cancer, renal cancer and other solid tumor patients, Br. J. Cancer, 77, 139, 1998. 73. Roaten, J.B., Interaction of the novel anthracycline antitumor agent N-benzyladriamycin-14-valerate with the C1-regulatory domain of protein kinase C: structural requirements, isoform specificity, and correlation with drug cytotoxicity. Mol. Cancer Therapeutics, 1, 483, 2002. 74. Boivin J., Monneret, C. and Pais, M., Approche de la synthèse d’anthracyclines oligosaccharidiques. Hemisynthèse de la 4-O-(2.deoxy-L-fucosyl)-daunorubicine, Tetrahedron, 37, 4219, 1981. Copyright © 2005 CRC Press, LLC 75. El Khadem, H.S. and Matsura, D., Synthesis of 7-(3-amino-2,3,6-trideoxy-4-O-(2,6 dideoxy-α-L-lyxo- hexopyranosyl)-α-L-lyxo-hexopyranosyl)-daunomycinone, Carbohydr. Res., 101, C1, 1982. 76. Horton, D. et al., Synthesis and antitumor activity of anthracycline disaccharide glycosides containing daunosamine, J. Antibiotics, 46, 1720, 1993. 77. Capranico, G. et al., Influence of structural modifications at the 3 and 4 positions of doxorubicin on the drug ability to trap topoisomerase II and to overcome drug resistance, Mol. Pharmacol., 45, 908, 1994. 78. Capranico, G., Butelli, E. and Zunino, F., Change of sequence specificity of daunorubicin-stimulated topoisomerase II DNA cleavage by epimerisation of the amino group of the sugar moiety, Cancer Res., 55, 312, 1995. 79. Animati, F. et al., New anthracycline disaccharides. Synthesis of L-daunosaminyl-α(1-4)-2-deoxy-L- rhamnosyl and of L-daunosaminyl-α(1-4)-2-deoxy-L-fucosyl daunorubicin analogues, J. Chem. Soc. Perkin Trans. I, 1327, 1996. 80. Arcamone, F. et al., Configurational requirements of the sugar moiety for the pharmacological activity of anthracycline disaccharides, Biochem. Pharmacol., 57, 1133, 1999. 81. Arcamone, F. et al., A doxorubicin disaccharide analog: apoptosis related improvement of efficacy in vivo, J. Natl. Cancer Inst., 89, 1217, 1997. 82. Pratesi, G. et al., Improved efficacy and enlarged spectrum of activity of a novel anthracycline disaccharide analogue of doxorubicin against human tumor xenografts, Clin. Cancer Res., 4, 2833, 1998. 83. Pratesi, G. et al., A comparative study of cellular and molecular pharmacology of doxorubicin and MEN 10755, a disaccharide analogue, Biochem. Pharmacol., 62, 63, 2001. 84. Cirillo, R. et al., Comparison of doxorubicin and MEN 10755-induced long term progressive car- diotoxicity in the rat, J. Cardiovasc. Pharmacol., 35, 100, 2000. 85. Messori, L. et al., Solution chemistry and DNA binding properties of MEN 10755, a novel disaccharide analogue of doxorubicin, Bioorg. Med. Chem., 9, 938, 2001. 86. Temperini, C. et al., The crystal structure of the complex between a disaccharide anthracycline and the DNA hexamer d(CGATCG) reveals two different binding sites involving two DNA duplexes, Nucleic Acids Res., 31, 1464, 2003. 87. Garbesi, A. et al., Synthesis and binding properties of conjugates between oligodeoxynucleotides and daunorubicin derivatives, Nucleic Acids Res., 25, 2121, 1997. 88. Capobianco, M.L. et al., New TFO conjugates containing a carminomycinone derived chromophore, J. Bioconjugate Chem., 12, 523, 2001. 89. Catapano C. et al., Synthesis of daunomycin conjugated triple-forming oligonucleotides that act as selective repressors of c-myc gene expression, in Proc. 92nd Annual Meeting Am. Ass. Cancer Res., Abstract 4563, New Orleans LA, 2001. 90. Young, R.C., Ozols, R.F., and Myers, C.E., Medical progress. The anthracycline antineoplastic drugs, New Engl. J. Med., 305, 149, 1981. 91. Geffen, D.B. and Man, S., New drugs for the treatment of cancer, 1990-2001, Isr. Med. Assoc. J., 4, 1124, 2002. 92. Hoelzer, D. and Gale, R.P., Acute lymphoblastic leukemia in adults: recent progress, future directions, Semin. Hematol., 24, 27, 1987. 93. Gale, R.P. and Foon, K.A., Therapy of acute myelogenous leukaemia, Semin. Hematol., 24, 40, 1987. 94. Bonadonna, G., Principi di chemioterapia antitumorale, in Manuale di Oncologia Medica, Bonadonna, G. and Robustelli della Cuna, G. Eds., Masson Italia, Milan, 1981, chap. 18. 95. Beretta, G., Cancer chemotherapy regimens, Farmitalia-Carlo Erba SpA, Milan, 1983. 96. Robert, J. and Gianni, L., Pharmacokinetics and metabolism of anthracyclines, Cancer Surv., 17, 219, 1993. 97. Gill, P. et al., Time dependency of Adriamycin and adriamycinol kinetics, Cancer Chemother. Pharmacol., 10, 120, 1983. 98. Ganzina, F., 4′-Epidoxorubicin, a new analogue of doxorubicin: a preliminary overview of preclinical and clinical data, Cancer Treat Rev., 10, 1, 1983. 99. Mouridsen, H.T. et al., Current status of epirubicin (Farmorubicin) in the treatment of solid tumours, Acta Oncol., 29, 257, 1990. Copyright © 2005 CRC Press, LLC 100. Plosker, G.L. and Fauld, D., Epirubicin. A review of its pharmacodynamic and pharmacokinetic properties, and therapeutic use in cancer chemotherapy, Drugs, 45, 788, 1993. 101. Robert, J., Clinical pharmacokinetics of epirubicin, Clin Pharmacokinet., 26, 428, 1994. 102. Cersosimo, R.J., Idarubicin: an anthracycline antineoplastic agent, Clin. Pharm., 11, 152, 1992. 103. Buckley, M.M. and Lamb, H.M., Oral idarubicin. A review of its pharmacological properties and clinical efficacy in the treatment of haematological malignancies and advanced breast cancer, Drugs Aging, 11, 61, 1997. 104. Warrell Jr., R.P., Aclacinomycin A: clinical development of a novel anthracycline antibiotic in the haematological cancers, Drugs Exp. Clin. Res., 12, 275, 1986. 105. Schrivers, D. et al., Phase I study of MEN-10755, a new anthracycline in patients with solid tumors: a report from the European Organization for Research and Treatment of Cancer, Early Clinical Studies Group, Ann. Oncol., 13, 385, 2002. 106. Bos, A.M.E. et al., Pharmacokinetics of MEN-10755, a novel anthracycline disaccharide analogue, in two phase I studies in adults with advanced solid tumours, Cancer Chemother. Pharmacol., 48, 361, 2001. 107. Caponigro, F. et al., A phase II study of MEN 10755 in patients with advanced or metastatic ovarian cancer. EORTC-NCI-AACR Meeting, (Abstr. 42), Eur. J. Cancer, 38 (Suppl. 7) S18, 2002. 108. Tjan-Heijnen, V.C.G. et al., A phase II study with MEN 10755 as second line therapy in advanced non small cell lung cancer (NSCLC), 27th ESMO Congress, Nice, France, Oct. 2002. Abstr. 542P, Ann. Oncol., 13, 55, 2002. 109. Comandini, A. et al., MEN 10755 in lung cancers: a survey of phase II studies in patients with advanced or metastatic NSCLC and SCLC-ED. 24th Winter Meeting of the EORTC-PAMM Group, Florence, Italy, Feb. 5-8, 2003. 110. Muggia, F.M., Liposomal encapsulated anthracyclines: new therapeutic horizons, Curr Oncol Rep., 3, 156, 2001. 111. Gabizon, A., Shmeeda, H. and Barenholz, Y., Pharmacokinetics of pegylated liposomal Doxorubicin: review of animal and human studies, Clin Pharmacokinet., 42, 419, 2003. 112. Perez, A.T et al., Pegylated liposomal doxorubicin (Doxil) for metastatic breast cancer: the Cancer Research Network, Inc. experience, Cancer Invest., 20 (Suppl.

2), 22, 2002. Copyright © 2005 CRC Press, LLC 17 Ansamitocins (Maytansinoids) Tin-Wein Yu and Heinz G. Floss CONTENTS I. Introduction II. Naturally Occurring Maytansinoids III. Biological Activity and Mechanism of Action IV. Total Synthesis V. Medicinal Chemistry: Structure Modification, Structure–Activity Relationships, and Metabolism VI. Preclinical and Clinical Developments VII. Biosynthesis References I. INTRODUCTION The maytansinoids represent a family of antitumor agents of extraordinary potency. Following the discovery of the parent compound, maytansine, in 1972,1 the chemistry and biology of the may- tansinoids were the subject of intense studies in the 1970s and early 1980s, culminating in clinical trials of maytansine against a variety of human cancers. After the disappointing outcome and eventual termination of the phase II clinical trials, interest in these compounds subsided consider- ably. The last decade has brought a certain resurgence of interest in the maytanisinoids, which were shown to have considerable potential as “warheads” in immunoconjugates with tumor-specific antibodies — an application that is currently under intense investigation. Other studies are address- ing the biosynthesis of the maytansinoids, its genetic control, and the role of plants versus micro- organisms in their formation. The earlier work, from the initial discovery to about 1984, has been summarized in several authoritative reviews.2–5 This chapter only briefly covers some of this earlier work and focuses primarily on the more recent developments in this field, which have occurred since these first reviews were published. II. NATURALLY OCCURRING MAYTANSINOIDS In 1972, Kupchan and colleagues, working under the auspices of the National Cancer Institute (NCI), reported the isolation and structure elucidation of maytansine (1) (Figure 17.1).1 They followed up on the observation of activity in vitro against cells of the human epidermoid carcinoma of the nasopharynx (KB) and in vivo against mouse L1210 and P388 leukemia, mouse sarcoma 180, and Lewis lung carcinoma and rat Walker 256 carcinosarcoma of extracts of Maytenus serrata (Celastraceae) collected in Ethiopia in 1961/1962. Bioassay-guided fractionation led to the isolation of 1 as the principal active constituent in a yield of 0.2 mg/kg of dried plant material. The targeted isolation of such a trace compound was only possible because of its potent cytotoxic activity (ED50 104–105 µg/mL against KB cells). The structure of 1 was determined primarily by x-ray crystallography. Subsequently, other Copyright © 2005 CRC Press, LLC H O O H O O N CI O O H N O CI O O H O 19 N O 1 3 O 20 18 2 4 5 6 N 21 H 17 16 8 7 H O O 13 11 15 109 14 N O 12 OHH HO N O O OHH O Maytansine (1) Colubrinol (2) HO N C H O O=C O O CI CH O H CI O O H 2 O O O O O N N H H CH3 O O R N O R N O OHH HO H O O R=OCH3, N-Methyltrenudone (3) R=H, Ansamitocin P-3 (4) R=H, Maytanbicycolinol (8) R=OCH3, 15-Methoxyansamitocin P-3 (5) H O R O 1 N R2 CI O O H O O N H O N O OHH O R1=CH2CH3, R2=CH3, Mallotusine (6) R1=CH3, R2=CH2CH3, Isomallotusine (7) FIGURE 17.1 Structures of naturally occurring maytanisinoids. Maytenus species (e.g., Maytenus buchananii), as well as another Celastraceae, Putterlickia verrucosa, were established as better sources of 1, containing seven to eight times higher levels of the compound.3 Following the discovery of maytansine, substantial numbers of analogs were also isolated from the same or related plant sources. Most of these contain different C3 ester side-chains, lack the ester side-chain or the N-methyl group, or carry a hydroxy/methoxy substituent at C15. Members of two other plant families, Colubrina texensis (Rhamnaceae) and Trewia nudiflora (Euphorbiaceae), also yielded maytanisinoids — the former some 15-oxygenated compounds, such as colubrinol (2),6 and the latter structures with cyclized ester moieties, such as N-methyltrenudone (3).7 The structures of the maytansinoids are unusual for higher plant products but closely resemble microbial metabolites of the ansamycin family.8 This raised questions as to the true biosynthetic origin of these compounds and led to a search for microbial producers of maytansinoids. This search culminated in a report by scientists at Takeda Chemical Industries9 on the isolation of a microorganism originally designated a Nocardia species but later reclassified as Actinosynnema pretiosum (Actinomycetes),10 which produces a series of compounds — the ansamitocins (e.g., Copyright © 2005 CRC Press, LLC ansamitocin P-3, 4) — with structures closely resembling and in some cases identical to those of plant-derived maytansinoids. Again, these compounds, to the extent that they carry ester side chains at C3, such as 4, are very potent cytotoxic agents. To date, 49 fully characterized, naturally occurring maytansinoids have been reported in the literature.11 Most of these were isolated during the 1970s and early 1980s, and their structures are listed in the earlier reviews.3–5 Since then, only five new compounds have been discovered, and these represent minor variations of known structures. Two groups reported the isolation of 4 from several species of mosses,12,13 together with three related maytansinoids including one new com- pound: 15-methoxyansamitocin P-3 (5).12 Feng et al.14 reported two new compounds, mallotusine (6) and isomallotusine (7), from Mallotus anomalus, and Sneden and coworkers15 isolated 2-N- demethylmaytanbutine (normaytanbutine) and, notably, maytanbicyclinol (8), from M. buchananii. There are also allusions to additional microbial maytansinoids in the literature (e.g., ansamitocin S-3 carrying a carbamoylated pentose moiety in place of the N-methyl group of 4),16 but details of their characterization have apparently not been published. III. BIOLOGICAL ACTIVITY AND MECHANISM OF ACTION The antitumor activity of maytansine and the ansamitocins has been discussed extensively in previous reviews.2–5 1 showed significant inhibitory activity against solid tumors, Lewis lung carcinoma, and B16 melanocarcinoma and antileukemic activity against P388 lymphocytic leuke- mia over a large (50–100-fold) dose range.17,18 In the 60–human cancer cell line screen of the NCI,19 1 showed an activity pattern indicative of tubulin-interactive agents.20 Related compounds carrying the C3 ester function (e.g., maytanprine, maytanbutine, and normaytansine), exhibited a similar activity spectrum and potency as 1.21 The 15-oxygenated maytansinoids colubrinol (2) and colu- brinol acetate isolated from C. texensis also showed potencies comparable to those of 1 against KB cells in vitro and P388 leukemia in vivo.6 Similarly, the bicyclic maytansinoids isolated from T. nudiflora and M. buchananii retain activity against P388 leukemia and KB cells.7,15 The ansam- itocins P-3 and P-4 exhibit a very similar activity spectrum and dose range as maytansine (e.g., antitumor activity against P388 leukemia in vivo at daily doses as low as 0.8 µg/kg), and they are also active against B16 melanoma, sarcoma 180, Ehrlich carcinoma, and P815 mastocytoma, though not very active against L1210 leukemia.9 It was established early on that maytansine22–24 and ansamitocin P-325 inhibit cell division by binding to tubulin, blocking the assembly of microtubules. The cells are arrested in metaphase; the cytological changes resemble those induced by vincristine. Maytansine competitively inhibits vincristine binding to tubulin, and vincristine — but not colchicine — competitively inhibits tubulin binding of 1.26 Thus, maytansinoids apparently bind to tubulin at a site overlapping the vinca alkaloids binding site and different from the colchicine binding site. The binding site of the vinca alkaloids has been mapped to the β-subunit of tubulin,27 and maytansine also binds to this subunit.28 Several studies have explored features of the β-tubulin structure that are important for maytansinoid binding/activity (e.g., a conserved asparagine residue29,30 and key sulfhydryl groups31). Fluorescent and photoaffinity probes based on ansamitocin P-3 have been described as reagents for probing the tubulin structure.32 Because tubulin binding of maytansinoids correlates with their activity against P388 leukemia33,34 and no other primary effects were apparent, the interaction with tubulin is considered the principal mechanism of antitumor action of the maytansinoids.3,22 In addition to their antitumor activity, the maytansinoids also show inhibitory activity against other eukaryotic systems, such as fungi and yeasts, protozoa, insects, and plants, but they have little if any antibacterial activity.35 Among a large number of fungi and yeasts tested, the yeast Filobasidium uniguttulatum (IFO 0699) was selected as a particularly sensitive test organism for the bioassay of maytansinoids. An inhibition zone of 20 mm was obtained with 45 ng of ansamitocin P-4.36 The antitubulinic activity of maytansinoids can be tested using protozoa, such as Tetrahymena pyriformis, which have cilia containing microtubule bundles. Deciliated Tetrahymena naturally Copyright © 2005 CRC Press, LLC recover their cilia, and hence motility, within 60 min, and this process is inhibited by maytansinoids (e.g., 100% by 1 µM 4).37,38 A similar assay observes axopod retraction in Actinophris sol or Actinosphaerium eichhorni.39 Insecticidal activity against members of the Lepidoptera and Coleoptera and against the chicken body louse was reported for extracts and isolated maytansinoids from T. nudiflora seeds.40 IV. TOTAL SYNTHESIS The decade following the original report on the isolation of 1 saw intense efforts aimed at the structure modification of the naturally occurring maytansinoids and at the total synthesis of this class of compounds. The work published up to 1983 has been summarized,3,5 and in particular detail in the review by Reider and Roland.4 Two groups, those of Meyers and of Corey, reported the total synthesis of (±) and (–)-N-methylmaysenine (not a naturally occurring maytansinoid).41–43 The first synthesis of a natural maytansinoid, (±)-maysine, was reported by Meyers et al. in 1979,44 followed in short order by syntheses of (±)-maytansinol (Meyers)45 and (–)-maytansine (Corey)46 in 1980. A third group, that of Isobe and Goto, also completed a total synthesis of (±)-maytansinol in 1982.47 These approaches all converged on an aminoaldehyde or aminoester intermediate con- taining the complete functionalized structural framework except carbons 1 and 2, which were introduced and ring-closed to the macrocyclic lactam in different ways.4 In a stereocontrolled synthesis of (–)-maytansinol reported by Isobe et al. in 1984, such an intermediate was prepared stereospecifically from D-mannose.48 Other approaches to the assembly of the macrolactam struc- ture of the maytansinoids were explored (e.g., by the groups of Ho, 4 Barton,4 Fried,49 Ganem,4 Confalone,50 Goodwin,51 and Hodgson et al.52) but did not lead to completed syntheses. Subsequent to the early synthetic work, two more total syntheses were reported. Gao’s group in Shanghai completed a synthesis of maytansine in 1988 (Figure 17.2),53 and Bénéchie and Khuong-Huu, building on the earlier efforts of the group at Gif-sur Yvette, reported a new synthesis of (–)- maytansinol in 1996 (Figure 17.3).54 However, all these syntheses, although challenging and creative, are lengthy and low yielding and neither solved the supply issue nor contributed signifi- cantly to the generation of structural analogs for structure–activity studies. V. MEDICINAL CHEMISTRY: STRUCTURE MODIFICATION, STRUCTURE–ACTIVITY RELATIONSHIPS, AND METABOLISM Despite the small amounts of material available from the plant sources, the Kupchan group explored the chemistry of maytansine to generate some analogs for preliminary structure–activity relationship studies. The 9-OH group was converted into various ethers, all of which showed no significant cytotoxic activity. Attempts to cleave the C3 ester group by hydrolysis led instead to elimination to give the olefin, maysine. The ester was cleaved successfully by LiAlH4 reduction at low temperature to give the alcohol, maytansinol. LiAlH4 reduction under more vigorous conditions led to the corresponding deschloro compound. None of the compounds lacking the ester side chain had significant activity, whereas numerous naturally occurring analogs with different ester groups all showed potent activity. All this chemistry has been described in detail in the previous reviews.3–5 With access to more starting material — compound 4 from the microbial fermentation — scientists at Takeda Chemical Industries explored the structure–antitumor activity of the maytansi- noids in more detail. Through chemical and microbial transformations, the researchers prepared a large number of analogs of 4 and tested their antitumor activity, as reported in a series of publications5,55–60 and a large number of patents. 11 Microbial transformations of 4 led to N- demethylation,55 20-O-demethylation,56 15-hydroxylation with either R or S stereochemistry,57 and 3-O-deacylation.55 Except for the deacylated compounds, all of these transformation products Copyright © 2005 CRC Press, LLC CI CI O NH O NH (C2H5O) CO C 2 P O H3 O O O O OSiPh2C(CH3)3 CI O CI OSiPh2C(CH3)3 O N CF3 O O N CF3 CH3 O S S O OMEM H S O S OH O CI O O O O O CI O N H P (OC ) O 3H5 2 O NHO OMEM CH3 OMEM S S S O S O CI

O OTBDMS O N 1 OMEM S S O FIGURE 17.2 Total synthesis of maytansine by Gu et al.53 retained significant antitumor activity. The 3-, 15-, and 20-hydroxy compounds served as starting materials for the synthesis of large numbers of alkyl and acyl derivatives at these positions, most of them showing varying degrees of biological activity.60 Interestingly, modifications of the 3-O- ester moiety modulated the activity of the molecule, with some compounds of low in vitro activity having high in vivo activity. One compound, the 3-O-phenylglycinate ester, showed a two to four times better than that of 1.60 Removal of the epoxide function to give an E-4,5-double bond did not abolish the activity,58 but inversion of the stereochemistry at C3 by oxidation of maytansinol to the 3-ketone, reduction with NaBH4, and reesterification gave 3-epi-ansamitocin P-3, which was devoid of cytotoxic activity.59 Deschloro compounds were also prepared by reduction of 4 with excess LiAlH4 followed by reesterification, with little effect on the biological activity.61 More recently, Sneden and coworkers explored further the reductive chemistry of maytansinoids (Figure 17.4).62,63 Reduction of maytanbutine, the analog of 1 with an N-isobutyryl instead of the N- acetyl group, with lithium borohydride/lithium triethylborohydride (10:1) cleaved the carbinolamide function to give the 7-carbamoyloxy-9-hydroxy derivative with a 1000-fold decrease in cytotoxic activity.62 In contrast, hydrogenation of 1 over Pd/C generated 11,12,13,14-tetrahydromaytansine, with little effect on the biological activity.63 The structure–antitumor activity relationships of the Copyright © 2005 CRC Press, LLC OTBDMS OMEM OTBDMS (EtO)2HC (EtO) S 2HC 1. n-BuLi, TMeDA S OMEM 2. DMF H (7 steps from 2(S),3(S)-epoxy-1-(tosyloxy)-butane) S S O (EtO) OTBDMS 1 eq, LDA C 2HC I DME/pentane, –78°C O N O separated by CH3I CI O OMEM chromatography Ag2O O N O S O S OR S Ar R = H, (1:1 to 4:1 mix of epimers) O (from vanillin in 12 steps) O (EtO) OTBD C 2HC MS HO OR I CI O N O O N five steps Mukayama’s O OMEM OMEM procedure S S S S O O CI O OTBDMS CI OH H O O N HO N five steps H OMEM O S (based on earlier routes) N O S OHH O O maytansinol FIGURE 17.3 Total synthesis of (–)-maytansinol by Bénéchie and Khuong-Huu.54 maytansinoids revealed by all these studies can be summarized, as shown in Figure 17.5. An ester group at the 3-position and an unblocked carbinolamide function at C9 are absolute requirements for potent antitumor activity; most other functional group modifications have either no or only a modu- lating effect on the biological activity. Suchocki and Sneden also examined the long-term stability of 1, showing that in outdated clinical samples 40% of the maytansine had decomposed, mostly by elimination of the C3 side chain to give maysine.64 One report65 and a patent66 deal with improved methods for the preparation of maytansinol (Figure 17.3) — an important starting material for the synthesis of immunoconju- gates of maytansinoids (vide infra).67 Although maytansine progressed all the way to phase II clinical trials, there are no data in the literature on the mammalian metabolism of the compound. The reasons probably are that the amounts of material available at the time were extremely small and that the analytical techniques then available were not sensitive enough to follow the metabolism of a compound that is active at such low concentrations. Chan and colleagues have recently begun to address this problem.11,67 Ansamitocin Copyright © 2005 CRC Press, LLC O HO O N LiAIH(OCH3)3 (68%) Inactive O O R N O CI O O H O O NH LiBH4/LIBHEt 2 O 3 N H OH O O Inactive N O HO H O CI H2, PD/C O O Active FIGURE 17.4 Reductive modifications of maytanisinoids.62,63,65 P-3 was incubated with rat and human liver microsomes, with a rat liver homogenate, and admin- istered to a whole rat. The metabolites were analyzed by electrospray tandem mass spectrometry, identifying relevant compounds by the presence of the chlorine isotopes and determining their likely structures by comparing their fragmentation patterns to those of the parent compound. Early results indicated that the most prominent route of metabolism in microsomes was demethylation of the C10 methoxy group, with several other oxidative metabolites detectable. Human liver microsomes metab- olized 70% of the administered dose, compared to 20% for rat liver microsomes. In contrast, the most prominent metabolite in rat liver homogenate was maytansinol, indicating the presence of an esterase in the soluble enzyme fraction. In a similar study, the metabolism of maytansine in human liver microsomes was found to involve most prominently the oxidation and removal of the N-methyl group in the ester side chain. These preliminary results are summarized in Figure 17.6. Ester side chain essential, Configuration but nature of ester group at C-3 critical not important CH3 not essential for activity N Chlorine not essential CH C for activity Epoxide not essential, 27 C O CI O O but modulates activity CH3 not essential O H O for activity 20 19 O N 1 3 5 23 28 18 2 4 6 21 17 7 H 16 8 O Hydroxy- and acyloxy 15 13 11 24 groups can be tolerated 14 12 10 9 N O OHH 26 O Free OH group 25 essential for activity Double bonds not essential Configuration at C-10 not important FIGURE 17.5 Summary of the structure–antitumor activity relationships of maytanisinoids. Copyright © 2005 CRC Press, LLC 10 O-demethylation (major) 26-hydroxylation (major) Lesser metabolites: 20-O- and N-demethylation, di-demethylations, 15-hydroxylation, 26- or 15- human liver hydroxylation+demethylation microsomes 70% of administered dose metabolized O C CI O O O CI OH O 20 N O O rat liver O N O homogenate 15 O 10 15 N O 10 OHH N O OH 26 O H O 4 rat liver 10 O-demethylation (major) microsomes Lesser metabolites: 20-O- and N-demethylation, di-demethylations, 26- or 15-hydroxylation 20% of administered dose metabolized R N O CH N O CH C CI O O C O O CI O O O N O O O N human liver O microsomes O N O OHH N O O OHH 1 O R = H and CHO FIGURE 17.6 (See color insert) Preliminary results on the mammalian metabolism of ansamitocin P-3 and maytansine.67 VI. PRECLINICAL AND CLINICAL DEVELOPMENTS Given the high potency of maytansine against various tumor types, including some solid tumors, the development of the compound for clinical trials was a high priority of the NCI. Following the isolation of larger, although still limited, quantities of 1 and satisfactory preclinical phar- macological and toxicological evaluation, the compound was entered into phase I clinical trials in patients with advanced disease refractory to conventional therapy in 1975.2,3 The tolerated dose was established as 1–2 mg/m2; observed side effects were primarily gastrointestinal toxicity, neurotoxicity, nausea, vomiting, and diarrhea; and a few patients showed promising partial responses. This led to the scheduling of a large number of phase II trials of 1 alone and in combination with other agents. These gave rather disappointing results, showing no substantial clinical benefits.68–84 As summarized in a 1984 report by the NCI to the Food and Drug Administration on the Investigational New Drug Application (INDA) for maytansine,85 the evaluation of the compound as a single agent against 36 different types of tumors in 819 patients showed only one complete response and 20 partial responses while at the same time revealing substantial gastrointestinal and neurological toxicity. At this point the NCI terminated the clinical trials of 1. The 1990s have seen a certain resurgence of clinical interest in the maytansinoids, specifically as cytotoxic agents (“warheads”) for antibody-directed targeted delivery to tumor cells. The Copyright © 2005 CRC Press, LLC H O O N SS H O O CI O O H H O N SS O N Ansamitocin P-3 reductive HO H or Maytansine Maytansinol O cleavage DCC, ZnCI2 N O OHH O DM1 H O O N SH CI O O H H O O N DTT Ab S-S-R H Ab S-S-Maytansinoid Conjugates O N O OHH O FIGURE 17.7 Structural modification of maytansinoids for conjugation to antibodies and receptor-specific carriers.86–89 antibody-targeted delivery approach, although conceptually appealing, has had limited practical success because the amount of drug that can be delivered to the target by an antibody is too small for most clinically used anticancer drugs to be effective. To overcome this problem, the warhead should be a highly potent drug, such as a maytansinoid. At the same time, the target-specific delivery holds the promise of minimizing the general toxic effects of the drug. Two approaches have been explored. Iwasa and coworkers at Takeda16 developed a bispecific monoclonal antibody that recognizes both ansamitocin P-3 and the human transferrin receptor, which is highly expressed on various human tumor cells. The 4 targeted to human A431 epidermoid carcinoma xenografts in nude mice was more effective than unconjugated 4 and eventually eradicated this tumor while at the same time showing reduced side effects. An alternative approach was pursued by Chari et al.86 at ImmunoGen, Inc., who modified maytansine for conjugation to antibodies or receptor agonists through disulfide linkages to the ester side chain. As shown in Figure 17.7, maytansinol generated either from 1 or 4 was converted by procedures detailed in the patent literature87,88 into a compound, designated DM1, with a methyl disulfide function in the ester side chain. DM1, on reduction with dithiothreitol gives the corre- sponding thiol, which can be conjugated to an antibody through either thioether or, preferably, disulfide linkages. The effectiveness of this approach was demonstrated with conjugates to the TA.1, 3E9, and A7 antibodies directed against the c-erb-2 oncogene protein, the human transferrin receptor, and human colon cancer cell lines, respectively. Multiple DM1 molecules could be conjugated to the TA.1 antibody (one to six molecules per antibody) and gave cytotoxicities against SB-BR-3 breast cancer cells up to IC50 1.2 × 1011 M, which were antagonized by the unconjugated antibody and dependent on expression of the target epitope.86 Subsequently, encouraging preclinical results were presented with a DM1 conjugate to the monoclonal antibody C242, which recognizes a glycoprotein expressed in human colorectal cancers.89 The C242-DM1 conjugate cured mice carrying xenografts of the COLO 205 human colon tumor — even very large tumors — at doses that showed very little toxicity. Effectiveness against two other colon tumor types was also dem- onstrated. The ImmunoGen technology was considered clinically promising and has been licensed to a number of major pharmaceutical companies.90 Several different DM1 conjugates are now undergoing clinical trials.91–94 The ImmunoGen group also reported a DM1 conjugate to a folate derivative designed to target cancer cells that overexpress the folate receptor.95 Copyright © 2005 CRC Press, LLC VII. BIOSYNTHESIS The biosynthesis of ansamitocins has been studied in A. pretiosum by feeding experiments with labeled precursors, which established the building blocks from which the molecule is generated. The maytansinoids are ansamacrolides that are assembled on a polyketide synthase (PKS) from an aromatic starter unit, 3-amino-5-hydroxybenzoic acid (AHBA),96 by chain extension with three acetate units, three propionate units, and one “glycolate” extender unit.97 Functional groups added to the polyketide backbone include three methyl groups from methionine,97 a cyclic carbinolamide from the carbamoyl group of citrulline97 and the ester side chain at C3, which is derived from the corresponding fatty acid or its precursor amino acid.98 The spectrum of the maytansinol esters produced in the fermentation could be strongly influenced by feeding the respective acid (e.g., isobutyric acid for ansamitocin P-3) or one of its precursors (e.g., isobu- tyraldehyde, isobutanol, or valine) to direct the fermentation towards predominant (up to 90%) production of a single ansamitocin.98 On the basis of knowledge about the biosynthesis of AHBA in the rifamycin producer, Amy- colatopsis mediterranei,99–101 the gene cluster encoding ansamitocin biosynthesis was cloned from A. pretiosum, using the AHBA synthase gene as a probe.102 Analysis of the cloned DNA revealed the presence of the relevant genes in two clusters, I and II, each carrying an AHBA synthase homologue (Figure 17.8A). The clusters are separated by a 30-kilobase, nonessential DNA segment. Cluster I contains the majority of the biosynthetic genes, including four large type I polyketide synthase genes (asmA–D), which together with the cognate downloading enzyme, encoded by asm9, are predicted to catalyze the assembly of the required linear octaketide and its release and cyclization to proansamitocin (Figure

17.8B). Also present are some, but not all, of the AHBA synthesis genes (asm22–24) — the genes required to synthesize the substrate for the third chain extension step incorporating the “glycolate” extender unit (asm13–17), genes for the post-PKS modifications of proansamitocin into 4 (asm7,10–12,19,21), and potential regulatory (asm2,8,18,29,31,34) and trans- port/resistance genes (asm4,5,35). Cluster II only contains the remaining genes essential for AHBA biosynthesis (asm43–45,47); potential regulatory (asm39,40,48) and transport/resistance genes (asm41,42) present can only affect the AHBA supply, as deletion of cluster II gave a nonproducing mutant in which full 4 production could be restored by supplementation with AHBA.102 On the basis of analogy to the other chain extension substrates, malonyl-CoA for acetate and methylmalonyl-CoA for propionate units, one would predict the substrate for incorporation of “glycolate” units to be 2-hydroxy- or 2-methoxymalonyl-CoA. This, however, is not the case, as the corresponding N-acetylcysteamine thioesters were not incorporated into 4.103 By compar- ison with the gene cluster for FK520,104 which also contains “glycolate” extender units, the products encoded by asm13–17 were recognized as probably involved in forming the substrate for this unusual chain extension step. Asm14 encodes a unique acyl carrier protein, which could substitute for the coenzyme A moiety as a carrier of the hydroxy- or methoxymalonate unit. Inactivation of each of these genes and phenotypic analysis of the mutants confirmed this deduction, showing loss of 4 production in each mutant, but formation of 10-demethoxy-4 in a low yield in the asm13, asm15, asm16, and asm17 mutants. 10-Demethoxy-4 was not formed in the asm14 mutant, indicating that the unique asm14-encoded ACP is necessary for the aberrant incorporation of malonyl-CoA in the absence of the “glycolate” substrate.11,103 Asm13–17 are not only indispensable but also sufficient to generate this “glycolate” substrate, as shown in Strep- tomyces lividans by coexpression of an asm13–17 cassette105 with a cassette of the erythromycin PKS, modified by replacement of the methylmalonate-specific AT6 with the presumably hydroxy- malonate-specific AT8 domain of the FK520 PKS.106 The cotransformant produced a new com- pound, 2-methoxy-2-desmethyl-6-deoxyerythronolide B, resulting from the incorporation of a “glycolate” in place of the terminal propionate unit. Deletion of the methyltransferase gene asm17 from the asm13–17 cassette gave no new product, indicating that the hydroxymalonate moiety must be methylated before incorporation.105 Hence the substrate for the third chain extension Copyright © 2005 CRC Press, LLC A I 1 2 3 4 5 6 7 8 25 2627 28 29 30 31 323334 35 36 37 C D A B 9 10 11 12 1314 15 16 17 18 19 20 21 22 23 24 II 38 39 40 4142 43 44 45 46 47 48 B asmA asmB asmC asmD L domain Module 1 Module 2 Module 3 Module 4 Module 5 Module 6 Module 7 ADE ACP KS AT DH ER KR ACP KS AT DH KR ACP KS AT DH KR ACP KS AT ACP KS AT KR ACP KS AT DH KR ACP KS AT KR ACP S S O S S O O O S O S S O S N2CO O O HO O N NH 2CO OH OH 2 HO O OH HO N2CO O 1 NH2 N NH 2CO OH 2 HO 2 O 3 NH N 2HO 2CO asm22 OCH3 HO asm23 4 NH2 HOOC-CH-CO~S-ACP asm24 5 NH HO 2 NH asm43 6 2HO asm19 7 NH asm44 2 asm10 asm45 asm12 COC(CH3)asm47 2 asm11 asm9 CI CH O O 3 H O CH O OH 3 H3CO N HO HN asm13 H CH asm14 3 O OH asm21 asm7 asm15 N O O asm16 CH OHH OCH3 3 H3CO asm17 Proansamitocin Ansamitocin P-3 FIGURE 17.8 (See color insert) Biosynthesis of ansamitocin P-3: (A) The ansamitocin biosynthetic gene cluster from Actinosynnema pretiosum, (B) assembly of proansamitocin on a type I polyketide synthase and post-PKS modification to ansamitocin.102 step must be 2-methoxymalonyl-ACP, which is formed from an as-yet-unidentified intermediate of the glycolytic pathway, as shown in Figure 17.9. The enzymes catalyzing the post-PKS modification steps from proansamitocin to 4 and the order in which they function were examined by inactivation of each corresponding gene and, in some cases, by heterologous expression in Escherichia coli and characterization of the enzymatic activ- ity.107,108 Halogenation and carbamoylation, catalyzed by Asm12 and Asm21, respectively, appear to be the first and second step in the sequence. As predicted by the results of Hatano et al.,98 the acyltransferase Asm19 catalyzing the addition of the ester side chain is rather promiscuous regarding its acyl substrate. Surprisingly, this reaction is not the terminal biosynthetic step, as expected from the natural occurrence of maytansinol, but occurs before the epoxidation by Asm11 and N-methy- lation by Asm10.107 The predominant pathway from proansamitocin to 4 is depicted in Figure 17.10. However, many of these post-PKS modifying enzymes are not highly substrate specific; hence, alternate minor pathways to 4 also operate, resulting in a metabolic grid.108 The results delineated above provide the tools for the genetic manipulation of the biosynthetic pathway to maytansinoids to generate structural analogs for biologic evaluation, which would be Copyright © 2005 CRC Press, LLC FIGURE 17.9 (See color insert) Formation of 2-methoxymalonyl-ACP, the substrate incorporation of the “glycolate” chain extension unit into ansamitocin .105 difficult to synthesize chemically. Such genetic manipulations could increase the structural diver- sity by altering the polyketide backbone and its functionality, or the downstream modifications of the macrolide. The question of the biosynthetic source of the maytansinoids isolated from higher plants has been investigated by Leistner and colleagues in Putterlickia species.39 Several lines of indirect evidence indicate that at least the polyketide framework of 1 is not biosynthesized by the plant, although the plants may have a role in determining the final structures elaborated: greenhouse- grown plants do not contain maytansine39; cell cultures of Maytenus109,110 and Putterlickia39 species do not produce maytansinoids; the 1-content of individual Putterlickia plants (three species) varied dramatically — a phenomenon apparently also encountered by the U.S. Department of Agriculture workers who recollected Maytenus species for large-scale isolation of 139; and a H OH CI O H OH HO O N HO AHBA N 3 Malonyl-CoA PKS Asm12 Asm21 3 Methylmalonyl-CoA Methoxymalonyl-ACP (AmsABCD + 9) OH OH O O O O CI H OH CI CI O H OH H O O HO O O N O N HO N H Asm19 H O Asm7 H O O N O N O N O OHH O OHH OHH O O CI H O O CI O O O O H O H HO N HO N Asm11 H Asm10 H O O N O N O OHH OHH O O Ansamitocin P-3 FIGURE 17.10 (See color insert) The sequence of post-PKS modification reactions converting proansamitocin into ansamitocin P-3.108 Copyright © 2005 CRC Press, LLC polymerase chain reaction–based search for AHBA synthase-homologous genes in the plant genome gave negative results.39 The focus is now on the isolation and screening of microorganisms found in, on, or near 1- producing plants. None of a substantial number of fungal endosymbionts isolated from three Putterlickia species carried an AHBA synthase homologue in its genome. Among the bacterial community of P. verrucosa and P. retrospinosa, one promising candidate was isolated from the rhizosphere and identified as Kitasatospora putterlickiae sp. nov.111 Several ansamitocin-specific gene homologues were found in its genome, but the ability to produce maytansinoids has not yet been demonstrated for this organism.11 REFERENCES 1. Kupchan, S.M. et al., Maytansine, a novel antileukemic ansa macrolide from Maytenus ovatus, J. Am. Chem. Soc., 94, 1354-1356, 1972. 2. Issell, B.F. and Crooke, S.T., Maytansine, Cancer Treat. Rev., 5, 199-207, 1978. 3. Komoda, Y. and Kishi, T., Maytansinoids, in Anticancer Agents Based on Natural Product Models, Douros, J. and Cassady, J.M., Eds., Academic Press, New York, 1980, pp. 353-389. 4. Reider, P.J. and Roland, D.M., Maytansinoids, in The Alkaloids, Brossi, A., Ed., Academic Press, New York, 1984, Vol. 23, pp. 71-156. 5. Smith Jr., C.R. and Powell, R.G., Chemistry and pharmacology of maytansinoid alkaloids, in Alkaloids, Pelletier, S.W., Ed., John Wiley and Sons, New York, 1984, Vol. 2, pp. 149-204. 6. Wani, M.C., Taylor, H.L., and Wall, M.E., Plant antitumor agents: Colubrinol acetate and colubrinol, antileukaemic ansa macrolides from Colubrina texensis, J. Chem. Soc. Chem. Commun., 390, 1973. 7. Powell, R.G. et al., Treflorine, trenudine, and N-methyltrenudone: Novel maytansinoid tumor inhibitors containing two fused macrocyclic rings, J. Am. Chem. Soc., 104, 4929-4934, 1982. 8. Rinehart Jr., K.L. and Shield, L.S., Chemistry of the ansamycin antibiotics, Fortschr. Chem. Org. Nat., 33, 231-307, 1976. 9. Higashide, E. et al., Ansamitocins, a group of novel maytansinoid antibiotics with antitumor properties from Nocardia, Nature, 270, 721-722, 1977. 10. Hasegawa, T. et al., Motile actinomycetes: Actinosynnema pretiosum subsp. pretiosum sp. nov., subsp. nov., and Actinosynnema pretiosum subsp. auranticum subsp. nov., Int. J. Syst. Bacteriol., 33, 314- 320, 1983. 11. Cassady, J.M. et al., Recent developments in the maytansinoid antitumor agents, Chem. Pharm. Bull., 52, 1-26, 2004. 12. Sakai, K. et al., Antitumor principles in mosses: The first isolation and identification of maytansinoids, including a novel 15-methoxyansamitocin P-3, J. Nat. Prod., 51, 845-850, 1988. 13. Suwanborirux, K. et al., Ansamitocin P-3, a maytansinoid, from Claopodium crispifolium and Ano- modon attenuatus or associated actinomycetes, Experientia, 46, 117-120, 1990. 14. Feng, S.C. et al., Studies on the chemical constituents of Mallotus anomalus. II. Structures of the antitumor components from Mallotus anomalus, Chinese Chem. Lett., 5, 743-746, 1994. 15. Larson G.M., Schaneberg B.T., and Sneden A.T., Two new maytansinoids from Maytenus buchananii, J. Nat. Prod., 62, 361-363, 1999. 16. Okamoto, K. et al., Therapeutic effect of ansamitocin targeted to tumor by a bispecific monoclonal antibody, Jpn. J. Cancer Res., 83, 761-768, 1992. 17. Kupchan, S.M. et al., Novel maytansinoids. Structural interrelations and requirements for antileukemic activity, J. Am. Chem. Soc., 96, 3706-3708, 1974. 18. Kupchan, S.M. et al., Maytanprine and maytanbutine, new antileukemic ansa macrolides from May- tenus buchananii, J. C. S. Chem. Comm., 1065, 1972. 19. Boyd, M.R. and Paull, K.D., Some practical considerations and applications of the National Cancer Institute in vitro anticancer drug discovery screen, Drug Dev. Res., 34, 91-109, 1995. 20. Cragg, G.M., personal communication, 2002. 21. Kupchan, S.M. et al., The maytansinoids. Isolation, structural elucidation, and chemical interrelation of novel ansa macrolides, J. Org. Chem., 42, 2349-2357, 1977. Copyright © 2005 CRC Press, LLC 22. Wolpert-DeFilippes, M.K. et al., Initial studies on the cytotoxic action of maytansine, a novel ansa macrolide, Biochem. Pharmacol., 24, 751-754, 1975. 23. Remillard, S. et al., Antimitotic activity of the potent tumor inhibitor maytansine, Science, 189, 1002- 1005, 1975. 24. Wolpert-DeFilippes, M.K. et al., Initial studies on maytansine-induced metaphase arrest in L1210 murine leukemia cells, Biochem. Pharmacol., 24, 1735-1738, 1975. 25. Ootsu, K. et al., Effects of new antimitotic antibiotics, ansamitocins, on the growth of murine tumors in vivo and on the assembly of microtubules in vitro, Cancer Res., 40, 1707-1717, 1980. 26. Mandelbaum-Shavit, F., Wolpert-DeFilippes, M.K., and Johns, D.G., Binding of maytansine to rat brain tubulin, Biochem. Biophys. Res. Commun., 72, 47-54, 1976. 27. Rai, S.S. and Wolff, J., Localization of the vinblastine-binding site on beta-tubulin, J. Biol. Chem., 271, 14707-14711, 1996. 28. Hamel E., Natural products which interact with tubulin in the vinca domain: Maytansine, rhizoxin, phomopsin A, dolastatin 10 and 15 and halichondrin B, Pharmacol. Therap., 55, 31-51, 1992. 29. Takahashi, M. et al., Molecular basis for determining the sensitivity of eucaryotes to the antimitotic drug rhizoxin, Mol. Gen. Genet., 222, 169-175, 1990. 30. Li, Y. et al., Binding selectivity of rhizoxin, phomopsin A, vinblastine, and ansamitocin P-3 to fungal tubulins: Differential interactions of these antimitotic agents with brain and fungal tubulins, Biochem. Biophys. Res. Comm., 187, 722-729, 1992. 31. Luduena, R.F. and Roach, M.C., Tubulin sulfhydryl groups as probes and targets for antimitotic and antimicrotubule agents, Pharmacol. Ther., 49, 133-152, 1991. 32. Sawada, T. et al., A fluorescent probe and a photoaffinity labeling reagent to study the binding site of maytansine and rhioxin on tubulin, Bioconjugate Chem., 4, 284-289, 1993. 33. Ikeyama, S. and Takeuchi, M., Antitubulin activities of ansamitocins and maytansinoids, Biochem. Pharmacol., 30, 2421-2425, 1981. 34. York, J. et al., Binding of maytansinoids to tubulin, Biochem. Pharmacol., 30, 3239-3243, 1981. 35. Tanida, S. et al., Ansamitocins, maytansinoid antitumor antibiotics. Producing organism, fermentation, and antimicrobial activities, J. Antibiot., 33, 192-198, 1980. 36. Hatano, K., Higashide, E., and Yoneda, M., Bioassay

of ansamitocin P-3, an antitumor antibiotic, Agric. Biol. Chem., 48, 1889-1890, 1984. 37. Tanida, S., Higashide, E., and Yoneda, M., Inhibition of cilia regeneration of Tetrahymena by ansam- itocins, new antitumor antibiotics, Antimicrob. Agents Chemother., 16, 101-103, 1979. 38. Tanida, S., Hasegawa, T., and Higashide, E., A simple method for determining antitubulinic activities of ansamitocins and related compounds using a cilia regeneration system with deciliated Tetrahymena, Agric. Biol. Chem., 44, 1847-1853, 1980. 39. Pullen, C.B. et al., Occurrence and non-detectability of maytansinoids in individual plants of the genera Maytenus and Putterlickia, Phytochemistry, 62, 377-387, 2003. 40. Freedman, B. et al., Biological activities of Trewia nudiflora extracts against certain economically important insect pests, J. Chem. Ecol., 8, 409-418, 1982. 41. Corey, E.J. et al., The total synthesis of (±)-N-methylmaysenine, J. Am. Chem. Soc., 100, 2916-2918, 1978. 42. Meyers, A.I. et al., Progress toward the total synthesis of maytansinoids. Synthesis of (±)-4,5- deoxymaysine (N-methylmaysenine), J. Am. Chem. Soc., 101, 4732-4734, 1979. 43. Corey, E.J. et al., Total synthesis of ()-N-methylmaysenine, J. Am. Chem. Soc., 102, 1439-1441, 1980. 44. Meyers, A.I. et al., Total synthesis of (±)-maysine, J. Am. Chem. Soc., 101, 7104-7105, 1979. 45. Meyers, A.I., Reider, P.J., and Campbell, A.L., Total synthesis of (±)-maytansinol. The common precursor to the maytansinoids, J. Am. Chem. Soc., 102, 6597-6598, 1980. 46. Corey, E.J. et al., Total synthesis of maytansine, J. Am. Chem. Soc., 102, 6613-6615, 1980. 47. Isobe, M., Kitamura, M., and Goto, T., Stereocontrolled total synthesis of (±)-maytansinol, J. Am. Chem. Soc., 104, 4997-4999, 1982. 48. Kitamura, M. et al., Stereocontrolled total synthesis of (-)-maytansinol, J. Am. Chem. Soc., 106, 3252- 3257, 1984. 49. Petrakis, K.S. and Fried, J., A novel synthesis of the cyclic carbinolamide moiety of the maytansinoids, Tetrahedron Lett., 24, 3065-3066, 1983, and references therein. Copyright © 2005 CRC Press, LLC 50. Confalone, P.N., The use of heterocyclic chemistry in the synthesis of natural and unnatural products, J. Heterocyclic Chem., 27, 31-46, 1990, and references therein. 51. Goodwin T.E. et al., Synthesis of two new maytansinoid model compounds from carbohydrate precursors, J. Carbohydr. Chem., 17, 323-339, 1998, and references therein. 52. Hodgson, D.M., Parsons, P.J., and Stones, P.A., A short and efficient synthesis of the C-3 to C-10 portion of the maytansinoids, Tetrahedron, 47, 4133-4142, 1991. 53. Gu, X. et al., Studies on the total synthesis of maytansinoids (IV) — synthesis of maytansine, Sci. Sinica, Ser. B., 31, 1342-1351, 1988, and references therein. 54. Bénéchie, M. and Khuong-Huu, F., Total synthesis of ()-maytansinol, J. Org. Chem., 61, 7133-7138, 1996. 55. Nakahama, K. et al., Microbial conversion of ansamitocin, J. Antibiot., 34, 1581-1586, 1981. 56. Izawa, M. et al., Demethylation of ansamitocins and related compounds, J. Antibiot., 34, 1587-1590, 1981. 57. Izawa, M. et al., Hydroxylation of ansamitocin P-3, J. Antibiot., 34, 1591-1595, 1981. 58. Kawai, A. et al., Chemical modification of ansamitocins. I. Synthesis and properties of 4,4-deoxy- maytansinoids, Chem. Pharm. Bull., 32, 2194-2199, 1984. 59. Akimoto, H. et al., Chemical modification of ansamitocins. II. Synthesis of 3-epimaytansinoids via 3-maytansinones, Chem. Pharm. Bull., 32, 2565-2570, 1984. 60. Kawai, A. et al., Chemical modification of ansamitocins. III. Synthesis and biological effects of 3- acyl esters of maytansinol, Chem. Pharm. Bull., 32, 3341-3351, 1984. 61. Miyashita, O. and Akimoto, H., Dechloromaytansinoids, Takeda Chemical Industries, Ltd., Japan, European Patent EP 11277 19800528, 1980. 62. Suchocki, J.A. and Sneden, A.T., New maytansinoids: Reduction products of the C(9)-carbinolamide, J. Org. Chem., 53, 4116-4118, 1988. 63. Schaneberg, B.T. and Sneden, A.T., Abstracts, 218th ACS Natl. Meeting, New Orleans, Aug. 22-26, 1999. 64. Suchocki, J.A. and Sneden, A.T., Characterization of decomposition products of maytansine, J. Pharm. Sci., 76, 738-743, 1987. 65. Sneden, A.T., Synthetic modifications of biologically interesting natural products, Synlett, 5, 313-322, 1993. 66. Terfloth, G., SmithKline Beecham Corp., USA, Process for preparation and purification of maytansi- nol, Int. Appl. WO 02/074775 A1, 2002. 67. Liu, Z., Floss, H.G., Cassady, J.M., and Chan, K.K., unpublished work, 2003. 68. O’Connell, M.J. et al., Phase II trial of maytansine in patients with advanced colorectal carcinoma, Cancer Treat. Rep., 62, 1237-1238, 1978. 69. Cabanillas, F. et al., Results of a phase II study of maytansine in patients with breast carcinoma and melanoma, Cancer Treat. Rep., 63, 507-509, 1979. 70. Creagen, E.T. et al., Phase II evaluation of maytansine in patients with advanced head and neck cancer, Cancer Treat. Rep., 63, 2061-2062, 1979. 71. Neidhart, J.A. et al., Minimal single-agent activity of maytansine in refractory breast cancer: A Southwest Oncology Group study, Cancer Treat. Rep., 64, 675-677, 1980. 72. Ahmann, D.L. et al., Phase II study of maytansine and chlorozotocin in patients with disseminated malignant melanoma, Cancer Treat. Rep., 64, 721-723, 1980. 73. Franklin, R. et al., A phase I-II study of maytansine utilizing a weekly schedule, Cancer, 46, 1104- 1108, 1980. 74. Rosenthal, S. et al., Phase II study of maytansine in patients with advanced lymphomas: An Eastern Cooperative Oncology Group pilot study, Cancer Treat. Rep., 64, 1115-1117, 1980. 75. Edmonson, J.H. et al., Phase II study of maytansine in advanced breast cancer, Cancer Treat. Rep., 65, 536-537, 1981. 76. Borden, E.C. et al., Phase II evaluation of dibromodulcitol, ICRF-159, and maytansine for sarcomas, Am. J. Clin. Oncol., 5, 417-420, 1982. 77. Creech, R.H. et al., Phase II study of cisplatin, maytansine, and chlorozotocin in small cell lung carcinoma (EST 2578), Cancer Treat. Rep., 66, 1417-1419, 1982. 78. Ratanatharathorn, V. et al., Phase II evaluation of maytansine in refractory non-Hodgkin’s lymphoma: A Southwest Oncology Group study, Cancer Treat. Rep., 66, 1587-1588, 1982. Copyright © 2005 CRC Press, LLC 79. Thigpen, J.T. et al., Phase II study of maytansine in the treatment of advanced or recurrent adenocar- cinoma of the ovary: A Gynecology Oncology Group study, Am. J. Clin. Oncol., 6, 273-275, 1983. 80. Edmonson, J.H. et al., Phase II study of maytansine in advanced sarcomas, Cancer Treat. Rep., 67, 401-402, 1983. 81. Thigpen, J.T. et al., Phase II study of maytansine in the treatment of advanced or recurrent squamous cell carcinoma of the cervix, Am. J. Clin. Oncol., 6, 427-430, 1983. 82. Sabio, H. et al., Maytansine in refractory childhood acute lymphocytic leukemia: A Pediatric Oncology Group study, Cancer Treat. Rep., 67, 1045, 1983. 83. Ravry, M.J., Omura, G.A., and Birch, R., Phase II evaluation of maytansine (NSC 153858) in advanced cancer: A Southwestern Cancer Study Group trial, Am. J. Clin. Oncol., 8, 148-150, 1985. 84. Kalser, M.H. et al., Phase II trials of maytansine, low-dose chlorozotocin, and high-dose chlorozotocin as single agents against advanced measurable adenocarcinoma of the pancreas, Cancer Treatment Rep. 69, 417-420, 1985. 85. Anon., Annual Report to the FDA by DCT, NCI, on Maytansine, NSC 153858, IND 11857, February 1984. 86. Chari, R.J. et al., Immunoconjugates containing novel maytansinoids: Promising anticancer drugs, Cancer Res., 52, 127-131, 1992. 87. Chari, R.J. et al., Cytotoxic agents comprising maytansinoids and their therapeutic use, ImmunoGen, Inc., USA, U.S. Patent 5,208,020 (1993). 88. Chari, R.J. and Widdison, W.C., Process for the preparation and purification of thiol-containing maytansinoids, ImmunoGen, Inc., USA, U.S. Patent 6,333,410 (2001). 89. Liu, C. et al., Eradication of large colon tumor xenografts by targeted delivery of maytansinoids, Proc. Natl. Acad. Sci. USA, 93, 8618-8623, 1996. 90. ImmunoGen Web site. Available at: http://www.immunogen.com. 91. Liu, C. and Chari, R.J., The development of antibody delivery systems to target cancer with highly potent maytansinoids, Expert Opin. Investig. Drugs, 6, 169-171, 1997. 92. Johnson, R., Progress in targeted anticancer therapy using immunotoxins, Pharmaceut. J., 265, 490, 2000. 93. Erickson, S. and Schwall, R., Humanized anti-ErbB2 antibody-maytansinoid conjugates and uses thereof in cancer therapy, Genentech, Inc., WO 2000, US 17229, 2001. 94. Ross, S. et al., Prostate stem cell antigen as therapy target: Tissue expression and in vivo efficacy of an immunoconjugate, Cancer Res., 62, 2546-2553, 2002. 95. Ladino, C.A. et al., Folate-maytansinoids: Target-selective drugs of low molecular weight, Int. J. Cancer, 73, 859-864, 1997. 96. Hatano, K. et al., Biosynthetic origin of aminobenzenoid nucleus (C7N-unit) of ansamitocin, a group of novel maytansinoid antibiotics, J. Antibiot., 35, 1415-1417, 1982. 97. Hatano, K. et al., Biosynthetic origin of the ansa-structure of ansamitocin P-3, Agric. Biol. Chem., 49, 327-333, 1985. 98. Hatano, K. et al., Selective accumulation of ansamitocins P-2, P-3 and P-4, and biosynthetic origins of their acyl moieties, Agric. Biol. Chem., 48, 1721-1729, 1984. 99. Yu, T.-W. et al., Mutational analysis and reconstituted expression of the biosynthetic genes involved in the formation of 3-amino-5-hydroxybenzoic acid, the starter unit of rifamycin biosynthesis in Amycolatopsis mediterranei S699, J. Biol. Chem., 276, 12546-12555, 2001, and references therein. 100. Guo, J. and Frost, J.W., Kanosamine biosynthesis: A likely source of the aminoshikimate pathway’s nitrogen atom, J. Am. Chem. Soc., 124, 10642-10643, 2002. 101. Arakawa, K. et al., Characterization of the early stage aminoshikimate pathway in the formation of 3-amino-5-hydroxybenzoic acid: The RifN protein specifically converts kanosamine into kanosamine 6-phosphate, J. Am. Chem. Soc., 124, 10644-10645, 2002. 102. Yu, T.-W. et al., The biosynthetic gene cluster of the maytansinoid antitumor agent ansamitocin from Actinosynnema pretiosum, Proc. Natl. Acad. Sci. USA, 99, 7968-7973, 2002. 103. Carroll, B.J. et al., Identification of a set of genes involved in the formation of the substrate for the incorporation of the unusual “glycolate” chain extension unit in ansamitocin biosynthesis, J. Am. Chem. Soc., 124, 4179-4177, 2002. 104. Wu, K. et al., The FK520 gene cluster of Streptomyces hygroscopicus var. ascomyceticus (ATTC 14891) contains genes for biosynthesis of unusual polyketide extender units, Gene, 251, 81-90, 2000. Copyright © 2005 CRC Press, LLC 105. Kato, Y. et al., Functional expression of genes involved in the biosynthesis of the novel polyketide chain extension unit, methoxymalonyl-acyl carrier protein, and engineered biosynthesis of 2-desme- thyl-2-methoxy-6-deoxyerythronolide B, J. Am. Chem. Soc., 124, 5268-5269, 2002. 106. Reeves, C.D. et al., A new substrate specificity for acyl transferase domains of the ascomycin polyketide synthase in Streptomyces hygroscopicus, J. Biol. Chem., 277, 9155-9159, 2002. 107. Moss, S.J. et al., Identification of Asm19 as an acyltransferase attaching the biologically essential ester side chain of ansamitocins using N-desmethyl-4,5-desepoxymaytansinol, not maytansinol, as its substrate, J. Am. Chem. Soc., 124, 6544-6545, 2002. 108. Spiteller, P. et al., The post-polyketide synthase modification steps in the biosynthesis of the antitumor agent ansamitocin by Actinosynnema pretiosum, J. Am. Chem. Soc., 125, 14236-14237, 2003. 109. Kutney, J.P. et al., Isolation and characterization of natural products from plant tissue cultures of Maytenus buchananii, Phytochemistry, 20, 653-657, 1981. 110. Dymowski, W. and Furmanowa, M., Searching for cytostatic substances in plant tissue of Maytenus molina by in vitro culture. III. Release of substances from active biological fractions from the callus extract of Maytenus wallichiana R. and B., Acta Pol. Pharm., 47, 51-54, 1990. 111. Groth, I. et al., Kitasatospora putterlickiae sp. nov., isolated from rhizosphere soil, transfer of Strep- tomyces kifunensis to the genus Kitasatospora as Kitasatospora kifunensis comb. nov., and amended description of Streptomyces aureofaciens Duggar 1948, Int. J. Syst. Evol. Microbiol., 53, 2033-2040, 2003. Copyright © 2005 CRC Press, LLC 18 Benzoquinone Ansamycins Kenneth M. Snader CONTENTS I. Introduction II. Mechanism of Action III. Synthesis of the Drug IV. Medicinal Chemistry of the Drug V. Clinical Applications References I. INTRODUCTION Ansamycins have been defined as cyclic molecules with long aliphatic chains that connect two opposite points of an aromatic ring, like a handle or ansa (hence the name).1 Most commonly the ansa ring is a macrocyclic lactam whose biosynthesis resembles that of other macrolides by being acetate derived. The aromatic rings are usually divided into two groups: those containing naphtha- lene rings, whose members include rifamycins, streptovaricins, tolypomycin, and naphthomycin, and those containing benzene rings. Furthermore, the most common, and usually the most biolog- ically active, aromatic rings are quinones or hydroquinones. This discussion will be limited to the biologically active benzoquinone ansamycins, whose members include macbecins, herbimycins, and geldanamycin, with particular emphasis on geldanamycin and its analogs. Geldanamycin derivatives especially are enjoying a resurgence of interest as a result of the discovery of their novel mechanism of cytotoxicity and potential utility as anticancer agents. The first of the benzoquinoid ansamycins to be discovered was geldanamycin (GA). It was isolated from a strain of Streptomyces hygroscopicus by the group at Upjohn,2 and the structure was published the same year by the Rinehart group.3 GA showed modest activity against Gram- positive organisms and some fungi and parasites, but most interesting was the cytotoxicity that was also demonstrated against various

cancer cell lines in vitro. Indeed, GA has shown a broad range of strong activity against the National Cancer Institute (NCI) panel of 60 human tumor cell lines.4 The structure was determined,3 and the absolute configuration has been confirmed by x-ray crystallographic determination5 to be 6(S)-, 7(S)-, 10(S)-, 11(R)-, 12(S)-, 14(R)-, as shown in Figure 18.1. Its strong in vivo activity in several models prompted its initial clinical trial, but that was discontinued because of hepatotoxicity, which was further elaborated on in a pharmacological evaluation at the NCI.6 Although the positions of the substituents listed in Figure 18.1 are reported in a consistent manner in the papers published on the benzoquinone ansamycins, the designation of the relative stereochemistry of some of the substituents, particularly those at C14 and C15, varies between papers. These structural variations are shown in Scheme 18.1 (note that the references cited in Scheme 18.1 are not in numerical sequence) and are also reflected in the discussions of the synthetic approaches presented in this review. Not long after the initial discovery, two other benzoquinoid ansamycins were discovered. In 1979, herbimycin A was discovered in another strain of Streptomyces hygroscopicus var. geldanus,7 Copyright © 2005 CRC Press, LLC O R5 17 O R4 15 N H O R R 2 1 R3 11 6 OCONH2 R1 (C6) R2 (C11) R3 (C12) R4 (C15) R5 (C17) Geldanamycin -OMe -OH -OMe -H -OMe Herbimycin A (TAN 420F) -OMe -OMe -OMe -OMe -H Herbimycin B -OMe -OH -OMe -H -H Herbimycin C (TAN 420D) -OMe -OH -OMe -OMe -H Macbecin I -Me -OMe -OMe -OMe -H Macbecin II (hydroquinone) -Me -OMe -OMe -OMe -H TAN 420A (hydroquinone) -OMe -OH -OH -OMe -H TAN 420B -OMe -OH -OH -OMe -H TAN 420C (hydroquinone) -OMe -OH -OMe -OMe -H TAN 420E (hydroquinone) -OMe -OMe -OMe -OMe -H FIGURE 18.1 Geldanamycin and related compounds. and x-ray crystallographic studies8 proved it to be quite similar to GA, with the same macrolide ring system but differing in the substituents at C11, C15, and C17. Not unsurprising was the retention of chirality between the two compounds in all shared chiral centers. Two other isomers, herbimycin B and herbimycin C, were also isolated and differed from the parent by variations at C11 and C15, as shown in Figure 18.1. Herbimycin B is also known as 17-demethoxygeldanamycin. Although somewhat less potent in in vitro studies, the herbimycins still share many of the same biological activities with GA.9 A third benzoquinoid ansamycin to be discovered was macbecin I, together with its hydro- quinone form, macbecin II. This compound was produced by a Nocardia sp. and showed a biological activity profile similar to GA.10,11 Once again the macrolide ring system and stereochemistry were the same as GA, with differences only at C6, C11, C15, and C17, as shown in Figure 18.1. It is not clear whether the hydroquinone is deliberately produced by the organism or is an artifact of the fermentation and isolation procedure. Because the quinones are more stable chemical entities when exposed to air, they have been most commonly studied, but the hydroquinones can be isolated under the appropriate conditions. In 1986 another series of ansamycins were isolated from a Streptomyces hygroscopicus strain and reported as the TAN 420 series.12 Isomers A, C, and E were isolated as the hydroquinones, and their corresponding quinone forms were reported as B, D, and F. TAN 420D was shown to be identical to herbimycin C, and TAN 420F was shown to be identical to herbimycin A. Only TAN 420A and its quinone, TAN 420B, were new isomers, differing from the reported herbimycin structures by having an hydroxy group at C12 instead of the more usual methoxy group. II. MECHANISM OF ACTION The original discovery of the antitumor activity of GA prompted a study of its mechanism of action. Initial reports were that the benzoquinone ansamycins were tyrosine kinase inhibitors4,13–15 although they also demonstrated several other activities, including virucidal,16 ischemia protection,17 herbi- cidal,7 antiprotozoal,2 and antimalarial.18 The presence of a quinone ring as part of the GA structure Copyright © 2005 CRC Press, LLC O O O O 17 O O O O N O N 15 H N 15 H O 15 H O O NH2 O OH NH 12 O NH 2 2 O 12 O 11 O O O O O O 11 O O Geldanamycin after Andrus et al.39 Macbecin I Macbecin I, Antibiotic C14919E1 DNP56 has 6 & 7 as Herbimycin B after Panek et al.41 after Baker et al.33 & DNP Geldanamycin numeration OH OH O 17 O O 17 O O O N N 15 H 15 H O N OH OH O NH O NH 15 H 2 2 O 12 12 O NH O 2 O 11 O O O 11 O O O 11 O O Macbecin II Macbecin II, Antibiotic C14919E2 Herbimycin A, Antibiotic TAN 420F after Panek et al.41 after Baker et al.33 & DNP after Martin et al.57 Geldanamycin numeration DNP has 6 & 7 as Herbimycin B O O O O 17 O O O 15 N H O O 15 N 15 N OH NH H H O 2 12 O O O 11 OH NH OH NH 2 2 O O 12 O 12 O O 11 O O O 11 O O Herbimycin A Herbimycin B Herbimycin B after Panek et al.41 after Panek et al.41 DNP56 O OH OH 17 R5 17 O O O O O R4 15 N 15 N H H 15 N OH OH H OH NH O 1 O 2 OH NH 12 O 2 2 HO 11 O O O 11 O O R R 2 R1 3 11 6 Antibiotic TAN 420A Antibiotic TAN 420C OCONH2 DNP56 DNP56 Antibiotic TAN 420B DNP56 OH O 17 17 O O O O 5 N 15 N 1 H H OH O O NH2 OH NH O O 2 O 11 O O O 11 O O Herbimycin C, Antibiotic TAN 420D Antibiotic TAN 420E DNP56 DNP56 SCHEME 18.1 Benzoquinone ansamycins. Copyright © 2005 CRC Press, LLC prompted a detailed study on the generation of intracellular free radicals through redox cycling.19 However, the major change in emphasis occurred with the discovery of the unique ability of GA to specifically bind to and antagonize the function of the chaperone protein, heat shock protein 90 (Hsp90).4 This stimulated a large volume of research into the role of these mediators in replicating cells. The fact that Hsp90 is constitutively expressed at from 2 to 10 times higher levels in tumor cells, compared to their normal counterparts, makes it even more important as an anticancer target.20,21 The various studies on the inhibition of Hsp90 by GA and the other members of the benzo- quinoid ansamycins have been thoroughly reviewed,20,22 but a few important conclusions and the current hypothetical model can be restated here. Chaperone proteins are regulatory molecules. They control the correct assembly of other polypeptides into active molecules without becoming incor- porated into the final structure. They are generally thought of as being involved with the folding of newly synthesized proteins or in their unfolding once their function has been completed. Hsp90 is one of the most abundant of these chaperone proteins, and perhaps because of this, the greatest amount of knowledge has been acquired about this protein. It has become clear that the Hsp90 family of chaperone proteins are modulators of many of the proteins involved in cell proliferation, including many tyrosine and serine/threonine kinases. The action of GA and other benzoquinoid ansamycins on Hsp90 proteins would then also show an effect on these kinases, which was their originally observed action. In their reviews, both Neckers and Ochel note that Hsp90 also regulates several other “client proteins” that are intimately involved with the proliferation and survival of cancer cells (Figure 18.2). These include the estrogen receptor, the androgen receptor, mutant p53, the Src family kinases, p185erbB2, Raf-1/v-Raf, and CDK4/CDK6. It thus becomes more understandable why the earlier publications on the biological activity reported for the benzoquinoid ansamycins could implicate such a wide variety of oncologic markers. Indeed, the very earliest reports of activity for GA were that they were inhibitors of p60v-src tyrosine kinase activity.23,24 It now fits the model not as a primary activity but by viewing the kinase inhibition as an indirect effect, in which disruption of the p60v- src-Hsp90 complex mimics Hsp90 depletion and causes destabilization and loss of the kinase protein.4 It also helps to explain the very high potency of these compounds, which are frequently active at nanograms per milliliter in in vitro cell systems such as rapidly proliferating leukemia cell lines like P388, L1210, and KB cells. Affecting numerous targets provides an amplification factor in GA’s action. The extent to which the heat shock proteins regulate all the proteins in the cell replication cycle is still evolving and will undoubtedly provide research opportunities for some years to come. There has been significant effort in attempting to understand the exact nature of the interaction between the benzoquinoid ansamycins and Hsp90. The group at Memorial Sloan Kettering25 was able to cocrystallize GA with an amino terminal fragment of human Hsp90 and to publish the structure of this complex. At approximately the same time, another group26 described a structural portion of bacterial DNA gyrase that contained the ATP binding sites and has an amino acid sequence that is nearly identical to that contained in the Hsp90 sequence, which binds ansamycins. In addition, the ATP binding region is also fully contained within the benzoquinoid ansamycin binding region. This information, together with the binding study data from Grenert’s laboratory,27 has led to the current working model in which ATP binds to Hsp90 in the hydrophobic pocket studied by Stebbins, and the conformation adopted by the model indicates that the ATP/GA binding site acts as a conformational switch region that regulates the assembly of Hsp90-containing mul- tichaperone complexes. This model also implies that any Hsp90 binding phenomena that are not ATP dependent would not be affected by the benzoquinoid ansamycins. BA's ATP Bind & Stabilize "Client" Proteins ADP "Inactive" Hsp90 "Active" Hsp90 FIGURE 18.2 Mechanistic role of Hsp90. Copyright © 2005 CRC Press, LLC OH O O HO O CI FIGURE 18.3 Radicicol. One other observation that has the potential to provide substantial medicinal chemistry devel- opment is the discovery28–30 of a non-ansamycin compound, radicicol (Figure 18.3), a fungal product that also binds to Hsp90 and competes with GA for binding to the amino-terminal binding pocket of Hsp90. This entry into the molecular modeling pool indicates that substantial latitude is available for more synthetic modifications and possibly some less complex structures — an ideal model for combi-chem approaches. Further encouragement can be derived from the recent report that a small purine, PU-3 (Figure 18.4), competes with GA in binding assays to Hsp90, although the relative affinity has only reached the 15–20 µM level so far.31 A full discussion of the structure of heat shock proteins is not within the scope of this review but can be found in one of the two excellent reviews already published.20,22 Some specifics relating to individual proteins identified with cancer cells are appropriate. Hsp90 exists in two forms, α and β. These forms are dimeric in the cytosol of unstressed cells but accumulate in the nucleus when the cell is stressed. Hsp90 interacts with 14 proteins that have been identified as important in cancer because of their involvement in the proliferation or survival of cancer cells. The original reference to the interaction of benzoquinoid ansamycins with Hsp90 was from the National Institutes of Health.4 They began with identifying a high correlation between biological effects and the ability to bind to Hsp90. Three groups independently identified the Hsp90 binding site for benzoquinoid ansamycins. Serendipitously, the pooled results of these studies showed that the hydrophobic binding pocket for GA-Hsp90 also fully contains the ATP-binding sites and thus offered an explanation for the effect of nucleotide binding (ATP) to Hsp90. The key to benzoquinoid ansamycin sensitivity is target protein destabilization, which is generally mediated by proteosomes. One of the major proteins in defining cancer cells is p53 protein. The current working hypothesis suggests that normal (wild-type) p53 has a tumor suppressant effect. A mutated p53 protein reacts with Hsp90/p23

complexes to achieve their correct conformation. GA interferes with this process and thus destabilizes mutant p53 to help restore the normal transcriptional activity of wild-type p53. The transforming protein of Rous sarcoma virus (p60v-src) is a tyrosine kinase. It is complexed with Hsp90 and another protein and is inactive as a kinase in this form to allow transport to the plasma membrane, where it then becomes active. Presence of GA disrupts this complex, causing destabilization and loss of the kinase protein. This has been offered as a partial explanation of the original observations that GA was a tyrosine kinase inhibitor, although it may be an insufficient reason. It has been shown32 that a hybrid of the phosphoinositol-3 kinase (PI3K) inhibitor LY294002 with GA is a direct inhibitor of DNA-dependent protein kinase (DNA-PK) and is over two orders of magnitude more active against DNA-PK than PI3K (see Section IV). NH2 N N N N O O O PU3 FIGURE 18.4 PU-3. Copyright © 2005 CRC Press, LLC In prostate and breast cancers, p185erbB2 (a receptor tyrosine kinase) is overexpressed, and GAs were shown to stimulate its rapid degradation. They cause rapid polyubiquitination of this kinase, followed by proteosome-dependent degradation. The p185erbB2 protein associates with another protein, Grp 94 (a Hsp90 homolog), and GAs bind to the Grp94. The full effect of Hsp90 in cancer cells is not yet fully elaborated; however, several important points can be made: Hsp90 is up-regulated in many tumor cells both in vivo and in vitro; Hsp90 has been correlated with poor survival in breast cancer; Hsp90 is essential for normal eukaryotic cell survival, and cancer cells seem to be especially sensitive to interference; and Hsp90 interference may also be of value in those instances in which steroid hormone antagonism is indicated, such as hormone-dependent prostate cancer and breast cancer. Ochel22 notes that there are at least three functions for Hsp90 proteins: they prevent aggregation and promote refolding of proteins after various stresses, they suppress phenotypic traits that only become expressed after certain stresses, and they stabilize certain key regulatory proteins such as progesterone and gluco-corticosteroid receptors. He also notes that the role of p53 in cancer cells is to act as a tumor suppressor but is mutated or absent in about one-half of all human cancers. It has been clearly demonstrated that treatment with GA results in destabilization of mutated p53, although it has no effect on wild-type p53. Thus, the benzoquinone ansamycins are able to selectively destabilize mutated p53. III. SYNTHESIS OF THE DRUG The synthesis of all three of the benzoquinoid ansamycins has been accomplished. The first total synthesis of (+)-macbecin I was published by Baker in 199033 and was rapidly followed by Evans34 and Panek.35 In addition, a formal total synthesis was also published by Martin.36 Similarly, herbimycin was also prepared by Tatsuta et al.37,38 However, it took until 2002 for the synthesis of GA to be accomplished.39 With seven stereogenic sites on the macrocyclic lactam, an isolated trisubstituted double bond, and a (Z,E) conjugated diene lactam, these molecules are a synthetic challenge that bring out the most creative solutions. There are at least three different methods for introducing the chiral centers, and a brief outline of how these centers were introduced is presented here. None of the syntheses were intended to replace the fermentation process for production of the parent antibiotic, but all of them offer the opportunity to examine the individual elements of the molecule for their relevance in the molecular binding to Hsp90 and subsequent enhancement of activity. Presented here is a drastically shortened description of the synthetic strategies of three of these syntheses intended to illustrate the different chiral reagents used for introduction of the stereochemical centers. For the construction of macbecin I, the Evans group34,40 used chiral boron enolate methodology to introduce the chiral centers. The molecule was broken down into three subunits, each containing appropriate stereochemistry, as shown in Figure 18.5. The first synthon was prepared from the benzaldehyde 4 plus the chiral oxazolidinone 5 as the boron enolate made with di-n-butyl boronate triflate (n-Bu2BOTf). The chirality derived from the selection of 4R,5S-norephedrine as starting material for the oxazolidinone established the centers at C15 and C14 of the ansa ring. Further elaboration of this intermediate led to the first synthon, 1 (Figure 18.6). Using the same oxazolidinone and boron enolate condensation with trans-cinnamaldehyde, the chirality around C11 and C10 was established as shown. The chain was extended with appropriate phosphonate chemistry to lead to an intermediate aldehyde, 6, which was again condensed with a third use of the oxazolidinone to fix the stereochemistry at C7 and C6. Cleavage of the double bond in 7 originally present in the cinnamaldehyde, followed by conversion to an aldehyde produced the second synthon 2 (Figure 18.7). Coupling the two synthons, 1 and 2, with TiCl4, followed by Dess Martin oxidation, gave the precursor 8 for macbecin with all the stereochemistry from C15 through C5. All that remained was to extend the chain by building the cis,trans diene (C1–C4)–conjugated ester using appropriate Copyright © 2005 CRC Press, LLC O O O OTBS O S N N 2O 15 13 H O 11 5 N O S O O 1 2 O O P O O F O F F F F F 3 FIGURE 18.5 Evans synthesis synthons. vinlogous phosphorus-based reagents to produce the open-chain intermediate 9. Closure was accom- plished by reduction of the nitro to amine, followed by ring closure using BOPCl and Hunig’s base, as previously used by Baker,33 to produce the des-carbamoyl macbecin I (10) (Figure 18.8). Although the first method used a partial convergent approach, Panek’s group35,41 assembled the entire chain sequentially using three crotyl silation steps. Starting with the aromatic aldehyde dimethylacetal (11) and the (E)-crotylsilane reagent 12, the stereochemistry at C15 and C14 was established. Hydroboration of the olefin led to the correct hydroxylation at the C12 position in a 10:1 ratio, and the chain was shortened by converting the terminal ester to an alcohol followed by cleavage of the diol to the aldehyde 13 (synthon 1) (Figure 18.9). O O O O O OH O O O2N + N O O2N 15 13 N O O Ph Ph O 4 5 O O O O 2N S 15 13 O S O 1 FIGURE 18.6 Evans synthesis synthon 1. Copyright © 2005 CRC Press, LLC O O O O O O O N O Ph H + Ph IM Ph H Ph 6 O O O OH O O OTBSO 6 + N O Ph H O IM N O Ph 7 2 FIGURE 18.7 Evans synthesis synthon 2. S S O O O OTBSO O N O N 1+2 O O O OTBSO 2 15 11 5 H O O 2N 15 11 5 N O OH O 8 O O O O O O OTBS O 5 15 N H N 4 H 2 O O 12 O O O O OH OH 9 10 FIGURE 18.8 Evans synthesis: descarbamoyl-macbecin I. Condensation of synthon 1 (13) with a second crotylsilane (14) gave the intermediate with stereochemistry established at C12 and C11 in a 12:1 diastereomeric ratio. Cleavage of the olefin by ozonolysis to the aldehyde followed by appropriate phosphorane olefin extension with pyruvate gave synthon 2 (15) (Figure 18.10). Using a third crotylsilane (16) on synthon 2 (15), plus an exchange of protective groups, led to the intermediate 17 with stereochemistry established at C7 OBn O O O O n O O B O O 2N O O2N 15 11 Si O O Ph O O hydroboration 11 12 10:1 anti:syn O O O O2N H O O OBn O O2N OH O O 13 O FIGURE 18.9 Panek synthesis synthon 1. Copyright © 2005 CRC Press, LLC FIGURE 18.10 Panek synthesis synthon 2. and C6. Cleavage of the olefin once again led to the aldehyde synthon 3 (18) (Figure 18.11). Once again, introduction of the olefins was done with appropriate phosphorane chemistry to give the open-chain nitro ester. Following the procedures that have precedence for final macrocyclic ring closure, developed for the maytansine ansamitocins, then completed the synthesis of descarbam- oylmacbecin (19) (Figure 18.12). As a final example, the recent preparation of GA 39,42,43 represents an even more difficult challenge. The chiral synthesis of geldanmycin is more complicated than that of macbecin and herbimycin by a further two additional features. First, the lack of an oxygen-bearing functionality at C15 discourages the use of an aldol reaction to form the C15–C14 bond. Second, the functionality at C11 is an hydroxyl rather than a methoxyl group, which implies that the synthesis must take care to protect and distinguish this hydroxyl from the other oxygenated function at C12. Finally, the unexpected behavior of the trimethoxy benzene precursor made the final oxidation to quinone much more complicated than the syntheses of the other benzoquinoid ansamysins. Nevertheless, O O OPAB O O2N 1 11 6 O 5 O O Si Ph O 17 16 O O O PMBO O O2N 15 5 H O O 18 FIGURE 18.11 Panek synthesis synthon 3. Copyright © 2005 CRC Press, LLC FIGURE 18.12 Panek synthesis; descarbamoyl-macbecin I. the synthesis was successfully completed in 41 steps and produced several new compounds whose pharmacological properties should contribute significantly to the developing structure–activity studies between GA and Hsp90. Once again, the synthesis was a sequential construct beginning with the trimethoxy aromatic 20. The first chiral center was then introduced through the use of asymmetric Evans alkylation. The resulting oxazolidinone was subsequently converted to the chiral aldehyde, synthon 1 (21) (Figure 18.13). The next chiral center was introduced by aldol condensation with the dioxanone 22. After rearranging protecting groups, methylation of the hydroxyl at C12 and reduction, the aldehyde synthon 2 (23) was obtained (Figure 18.14). The aldehyde was then homologated once again with an antiglycolate aldol condensation and then converted to the terminal methylene 24. Careful hydroboration, followed by deprotection, gave the S configuration at the C10 position in 25. Oxidation led to the aldehyde which, after condensation with pyruvate, provided the aldehyde synthon 3 (26) (Figure 18.15). The aldehyde was then condensed once again to complete the chain, using a new (–)-norephe- drine-derived glycolate, which was subjected to boron–enolate aldol condensation to give the syn adduct 27 in a ratio better than 20:1. Introduction of the Z and E olefin constructs followed routes similar to those of earlier investigators in macbecin and herbimycin syntheses and led to a carbox- ylate ester. The carboxylate residue was then elaborated into the protected allyl ester (synthon 4) 28 (Figure 18.16), which became the precursor for asymmetric hydroboration to the diol with the S configuration at C10. O O O O O O Br O H N O O O NO Bn 2 NO2 20 21 FIGURE 18.13 Andrus synthesis synthon 1. Copyright © 2005 CRC Press, LLC FIGURE 18.14 Andrus synthesis synthon 2. FIGURE 18.15 Andrus synthesis synthon 3. FIGURE 18.16 Andrus synthesis synthon 4. Copyright © 2005 CRC Press, LLC O O OTBS OTES O Ar N O H O O OTBS O O O O O O NH2 29 FIGURE 18.17 Andrus synthesis synthon 5. Ring closure using the now-familiar macrocyclic lactamization procedures smoothly gave the trimethoxy macrocycle. The carbamoyl was attached using trichloroacetylisocyanate to avoid pre- mature deprotection to give synthon 5 (29) (Figure 18.17). All that remained was the required oxidation to quinone. Conversion of the trimethoxy benzene to the p-quinone proved to be much more complex than the other benzoansamycins. Most oxidation techniques led to either azaquinone 30 or the orthoquinone 31 (Figure 18.18). Ultimately, GA was obtained through a very brief exposure (~1 min) to nitric acid, followed by rapid quenching of the reaction in NaHCO3. A careful examination of the electronic conformation of the hydroquinone system offered some explanation for this bias but did not lead to a good oxidation system. The side reactions led to at least two new ansamycin analogs for biological studies, although all of the other analogs were less cytotoxic in a cell-based assay known to be dependent on Hsp90 kinase signaling. The azageldanamycin showed significant activity and may be important in future design studies. IV. MEDICINAL CHEMISTRY OF THE DRUG Early studies on GA attempted to improve solubility of this

hydrophobic molecule. A variety of prodrug esters and quinone derivatives were prepared44 in an attempt to improve solubility and drug delivery. Unfortunately, none of these early modifications approached the potency of the parent and so were abandoned. One of the first published, systematic evaluations of the structure–activity relationships of GA came from the group at Pfizer.45 Using both GA and 4,5-dihydrogeldanamycin, which are readily prepared by fermentation, the group exploited the opportunity offered by the facile Michael addition that is possible at the C17 location on the quinone ring. They noted that it was possible to treat both GA and dihydrogeldanamycin with various amines to produce the 17-amino-17-desmethoxy analogs (32) (Figure 18.19). It was only some time later that it was shown that the binding with Hsp90 is not affected by substitutions in this location of the molecule.25 The group also prepared a series of derivatives at O O O O O O O O O N N N O H H O O OH O OH O OH O O O O O O O O NH2 O NH2 O NH2 Geldanamycin Azaquinone ortho-Geldanamycin 30 31 FIGURE 18.18 Andrus synthesis: geldanamycin and analogs. Copyright © 2005 CRC Press, LLC O O O RHN O O N N H R-NH2 H O O OH O OH O O O O O O NH2 O NH2 32 FIGURE 18.19 Pfizer aminogeldanamycin analogs. both the C17 and C19 position by condensation with diamines. Activity was evaluated in vitro by measuring depletion of erB2 in SKBR cells. With this model they found that C17 substituents were best tolerated and so prepared a series of more than 80 derivatives. In vitro activity was followed by in vivo evaluation, and although the researchers noted that the two activities did not correlate well, 17-allylamino-17-desmethoxylgeldanamycin was one of the best candidates for further study. Indeed this analog is currently in phase I clinical trial in spite of difficulty in intravenous admin- istration because of the poor solubility characteristics. The group also studied a series of fused ring compounds similar to geldanoxazine (33) and geldanazine (34) that were originally reported by Rinehart (Figure 18.20).44 In summary, the researchers found that these fused-ring compounds were not significantly active in their in vivo systems. In a second discussion,46 the authors expanded their description of the 17-amino derivatives including 17-AAG. Careful evaluation of the N-22 acylated analogs was also explored but was abandoned when it was found that they were readily cleaved in vitro. The researchers also explored the decarbamoylated products at C7, the 11-O-acyl derivatives, the C7 ketone, and the open ansa ring compound. From their studies, several conclusions were made: simple alkylation of the N22 substituent generally reduced activity, except for the phenacyl derivative, where the x-ray supported a hydrogen-bonded structure that preserved the ansa ring configuration; decarbamoylation at C7 generally reduced the in vitro activity; acylation of the C11 hydroxyl generally reduced the in vitro activity; opening the ansa ring always destroyed any biological activity; the ketone formed at C7 was inactive, but a ketone at C11 was quite potent in vitro — almost as good as the corresponding hydroxy analog; and herbimycins A and C (containing a methoxy group at C15 and a proton at C17) are active in vitro but are about fourfold less potent than GA. Unfortunately, the Pfizer group did not have the benefit of x-ray crystal studies, which later demonstrated the binding of their analogs to a highly conserved pocket in the molecular chaperon Hsp90.25,47,48 Among other observations, these studies showed that substitution at C17 of the GA structure was well tolerated by the Hsp90 binding. The group at Sloan-Kettering49 exploited this tolerance by creating several dimers with molecules that were selected to improve their selective N N O N O O N N H H O OH OH O OH O O O O O O NH2 O NH2 Geldanoxazine Geldanazine 33 34 FIGURE 18.20 Geldanazine and derivatives. Copyright © 2005 CRC Press, LLC O O Br Br Br TBSO TBSO 1. NaN O 3 2. LAH/THF NH OH 2 TBSO H O N HO O N H O OH O O O O NH2 35 FIGURE 18.21 Memorial Sloan-Kettering estradiol heterodimers. adsorption by specific cell types. The first attempt involved the construction of heterodimers linking GA to estradiol through an aminoalkyl group, which was prepared as shown in Figure 18.21. A series of linkage chains were prepared and tested for their ability to affect the steady-state levels of HER2, ER (estrogen receptor), and Raf in MCF-7 breast cancer cells. HER2 is a trans- membrane kinase that is amplified and overexpressed in a significant number of breast cancers. The researchers’ studies showed that all of the heterodimers were less potent than the parent GA. The nature of the linker was important for selective activity. The best compound incorporated the 1,4-(2-butenyl) linker (35), which showed quite reasonable activity against the three marker proteins but was essentially inactive against IGF1R marker protein and had only marginal activity against Raf-1, both of which are not associated with the estrogen receptor. Also supporting the researchers’ hypothesis was the demonstration that the analog was essentially inactive against a prostate cancer cell line, whereas GA induced the loss of an androgen receptor. With this success in hand, the group50 prepared a series of heterodimers with a testosterone attachment, as shown in Figure 18.22. In this case, activity was evaluated in vitro against LNCaP prostate cells. Once again, the nature of the linker was important for activity, and it was also important that the attachment of the testosterone molecule was linked to the C17α position of the steroid. The best analog of this series was (36; n = 4), which contained a six-carbon chain. It was substantially less active than GA against AR-independent cells (up to 7.5-fold) but was more potent that GA against the wild-type prostate cells (LAPC4), which strongly expresses the androgen receptor. A third study31 prepared a series of hybrid dimers between the phosphoinositol kinase (PI3K) inhibitors LY 294002 or LY292223 (Figure 18.23) and GA, with the intention that it would then affect the complex formed between the LY-binding protein and Hsp90 and would produce a better inhibitor of PI3Kinase. PI3Ks are critical in signal transduction, regulating p53 and certain DNA damage-repair mechanisms. The test system for evaluating biological effect involved direct inhi- bition of PI3K, which was purified from MCF-7 breast cancer cells. A series of dimers were prepared according to the scheme shown in Figure 18.24. On testing the various analogs, the dimers (37) showed inhibition of MCF-7-derived PI3K, which was weaker than their parents, while GA did not inhibit the kinases at all. Some of the dimers were actually more potent than the parent LY 294002. Addition of Hsp90 to the system followed by measurement of the inhibition showed dramatic selectivity of the dimer against the DNA-PK with virtually no activity against the ubiquitous PI3K, whereas the parent was equally Copyright © 2005 CRC Press, LLC OH O OTBS n HC O O O HO OTBS n O O HO O HN O O N H O O OH O O O O NH2 36 FIGURE 18.22 Memorial Sloan-Kettering testosterone heterodimers. active against both systems demonstrating the desired selectivity. As further evidence for selectivity, the authors showed direct binding of the dimers to Hsp90, indicating that hybridization did not interfere with this association. A similar study with GA dimers51 was made in an attempt to circumvent the broad toxicity of the parent GA. The objective was to improve the selectivity of GA against HER-kinases that may require association with Hsp90 chaperones and thus reduce overall toxicity. Analogs were prepared by the route shown in Figure 18.25. The best analog in this case was the four-carbon link (38; n = 4), which was almost as potent against HER2 as GA but was much less effective against Raf-1. The researchers also prepared several unsymmetrical derivatives containing only a single GA residue, but these were generally no more selective than the parent GA. Once again, the opening of the ansa ring destroyed any activity. Further mechanism studies indicated that the dimer, similar to GA, affects HER2 expression by inducing protein degradation. Further tests on cell lines for which the HER-kinase family are not expressed showed that GA was still a potent inhibitor at 3 nM but the dimer was not. Further evaluation of all of these dimers in in vivo systems will have to done, but clearly the possibility of tailoring the molecular structure to build enzyme specificity is a viable possibility and should produce new therapeutics. V. CLINICAL APPLICATIONS Many of the various 17-aminogeldanamycin analogs were evaluated at the NCI using their 60–human tumor cell line panel and in vivo models. A combination of potency, reduced toxicity, O O O N O O N Ph O LY 294002 LY 292223 FIGURE 18.23 LY 294002 and LY 29223. Copyright © 2005 CRC Press, LLC FIGURE 18.24 Memorial Sloan-Kettering LY heterodimers. and the discovery of a novel mechanism of activity (Hsp90 binding) prompted the selection of 17- allylamino-17-desmethoxygeldanamycin (17-AAG) for clinical studies,52 and it is currently in six phase I clinical trials.53 These trials are directed toward solid tumors, epithelial cancer, malignant lymphoma, sarcoma, prostate, and chronic myelogenous leukemia. One of the greatest obstacles to the development of 17-AAG is its very poor solubility, which prompted the development of a microdispersion preparation for drug delivery.54 This pharmaco- kinetic problem with the GA analogs has prompted other groups to pursue improvements in their physical properties.55 Subsequent examination of other analogs suggested the selection of another, more soluble analog — 17-dimethylaminoethylamino-17-desmethoxygeldanamycin (17-DMAG) — for clinical development. This analog is currently undergoing preclinical evaluation at the NCI. O O O O O H H N O N N n O H H2N n NH O 2 N N H H OH O O O O O O HO OH O O NH2 O O O O H2N O O NH2 38 FIGURE 18.25 Memorial Sloan-Kettering GA dimers. Copyright © 2005 CRC Press, LLC REFERENCES 1. Lancini, G. and Parenti, F., Antibiotics: An Integrated View, Springer, New York, 1982, 10. 2. DeBoer, C. et al., Geldanamycin, a new antibiotic, J. Antibiot., 23, 442, 1970. 3. Rinehart, K.L. et al., Geldanamycin I. Structure assignment, J. Am. Chem. Soc., 92, 7591, 1970. 4. Whitesell L. et al., Inhibition of heat shock protein HSP90-pp60v-src heteroprotein complex formation by benzoquinone ansamycins: essential role for stress proteins in oncogenic transformation, Proc. Natl. Acad. Sci. USA, 91, 8324, 1994. 5. Schnur, R.C. and Corman, M.L., Tandem [3,3]-sigmatropic rearrangements in an ansamycin: Ste- reospecific conversion of an (S)-allylic alcohol to an (S)-allylic amine derivative, J. Org. Chem., 59, 2581, 1994. 6. Supko, J.G. et al., Preclinical pharmacologic evaluation of geldanamycin as an antitumor agent, Cancer Chemother. Pharmacol., 36, 305, 1995. 7. Omura, S. et al., Herbimycin, a new antibiotic produced by a strain of Streptomyces, J. Antibiot., 32, 255, 1979. 8. Furusaki, A. et al., Herbimycin A: an ansamycin antibiotic; x-ray crystal structure, J. Antibiot., 33, 781, 1980. 9. Fukazawa, H. et al., Specific inhibition of cytoplasmic protein tyrokinases by herbimycin A in vitro, Biochem. Pharmacol., 42, 1661, 1991. 10. Tanida, S., Hasegawa, T., and Higashida, E. Macbecins I and II, new antitumor antibiotics. I., J. Antibiot., 33, 199, 1980. 11. Muroi, M. et al., Macbecins I and II, new antitumor antibiotics. II., J. Antibiot., 33, 205, 1980. 12. Shibata, K. et al., TAN420 Antibiotics, J. Antib., 39, 1630, 1986. 13. Miller, P. et al., Depletion of the erbB-2 gene product p185 by benzoquinoid ansamycins, Cancer Res. 54, 2724, 1994. 14. Miller, P. et al., Binding of benzoquinoid ansamycins to p100 correlates with their ability to deplete the erbB2 gene product p185, Biochem. Biophys. Res. Comm., 210, 1313, 1994. 15. Sepp-Lorenzino, L. et al., Herbimycin A induces the 20S proteasome and ubiquitin dependent deg- radation of receptor tyrosine kinases, J. Biol. Chem., 270, 16580, 1995. 16. Li, L.H. et al., Effects of geldanamycin and its derivatives on RNA-directed DNA polymerase and infectivity of Rausher leukemia virus, Cancer Treat. Rep., 61, 815, 1977. 17. Conde, A.G., et al., Induction of heat shock proteins by tryosine kinase inhibitors in rat cardiomylo- cytes and myogenic cells confers protection against simulated ischemia, J. Mol. Cell Cardiol., 29, 1927, 1997. 18.

Kumar, R., Musiyenko, A., and Barik, S., The heat shock protein 90 of Plasmodium falciparum and antimalarial activity of its inhibitor, geldanamycin, Malaria J., 2, 1, 2003. 19. Benchekroun, N.M., Myers, C.E., and Sinha, B.K., Free radical formation by ansamycin benzoquinone in human breast tumor cells; implications for cytotoxicity and resistance, Free Radic. Biol. Med., 17, 191, 1994. 20. Neckers, L., Schulte, T.W., and Mimnaugh, E., Geldanamycin as a potential anti-cancer agent: Its molecular target and biochemical activity, Investig. New Drugs, 17, 361, 1999. 21. Ferarini, M. et al., Unusual expression and localization of heat-shock proteins in human tumor cells, Int. J. Cancer, 51, 613, 1992. 22. Ochel, H-J., Eichhorn, K., and Gademann, G., Geldanamycin: the prototype of a class of antitumor drugs targeting the heat shock protein 90 family of molecular chaperones, Cell Stress Chaperones, 6, 105, 2001. 23. Uehara, Y., et al., Phenotypic change from transformed to normal induced by benzoquinoid ansamycins accompanies inactivation of p60src in rat kidney cells infected with Rous sarcoma virus, Mol. Cell Biol., 6, 2198, 1986. 24. Uehara, Y. et al., Effects of herbimycin derivatives on src oncogene function in relation to antitumor activity, J. Antibiot. (Tokyo), 41, 831, 1988. 25. Stebbins, C.E. et al., Crystal structure of an Hsp90-geldanamycin complex: targeting of a protein chaperone by an antitumor agent, Cell, 89, 239, 1997. 26. Bergerat, A. et al., An atypical topoisomerase II from Archaea with implications for meiotic recom- bination, Nature, 386, 414, 1997. Copyright © 2005 CRC Press, LLC 27. Grenert et al., The amino-terminal domain of heat shock protein 90 (hsp90) that binds geldanamycin is an ATP/ADP switch domain that regulates hsp90 conformation, J. Biol. Chem. 272, 23843, 1997. 28. Schulte, T.W. et al., Antibiotic radicicol binds to the N-terminal domain of Hsp90 and shares important biologic activities with geldanamycin, Cell Stress Chaperones, 3, 100, 1998. 29. Sharma, S.V., Agatsuma, T., and Nakano, H., Targeting of the protein Hsp90 by the transformation suppressing agent radicicol, Oncogene, 16, 2639, 1998. 30. Roe, S.M. et al., Structural basis for inhibition of the Hsp90 molecular chaperone by the antitumor antibiotics radicicol and geldanamycin, J. Med. Chem., 42, 260, 1999. 31. Chiosis, G. et al., A small molecule designed to bind to the adenine nucleotide pocket of Hsp90 causes Her2 degradation and the growth arrest and differentiation of breast cancer cells, Chem. Biol., 8, 289, 2001. 32. Chiosis, G., Rosen, N., and Sepp-Kirenzino, L., LY294002-geldanamycin heterodimers as selective inhibitors of the PI3K and PI3K-related family, Bioorg. Med. Chem. Lett., 11, 909, 2001. 33. Baker, R and Castro, L., Total synthesis of (+)-macbecin I, J. Chem. Soc., Perkin Trans. I. 47, 1990. 34. Evans, D.A. et al., Asymmetric synthesis of macbecin I, J. Org. Chem., 57, 1067, 1992. 35. Panek, J.S. and Xu, F., Total synthesis of (+)-macbecin I, J. Am. Chem. Soc., 117, 10587, 1995. 36. Martin, S.F. et al., A formal total synthesis of (+)-macbecin I, J. Org. Chem., 57,1070, 1992. 37. Nakata, M. et al., Total synthesis of herbimycin A, Tetrahedron Lett., 32, 6015, 1991. 38. Nakata, M. et al., The total synthesis of herbimycin-a, Bull. Chem. Soc. Jpn., 65, 2974, 1992. 39. Andrus, M.B. et al., Total synthesis of (+)-geldanamycin and (-)-o-quinogeldanamycin with use of asymmetric anti- and syn-glycolate aldol reactions, Org. Lett., 4, 3549, 2002. 40. Evans, D.A., Miller, S.J. and Ennis, M.D., Asymmetric synthesis of the benzoquinoid ansamycin antitumor antibiotics: Total synthesis of (+)-macbecin, J. Org. Chem., 58, 471, 1993. 41. Panek, J.S., Xu, F., and Rondon, A.C., Chiral crotylsilane-based approach to benzoquinoid ansamycins: Total synthesis of (+)-Mabecin I, J. Am. Chem. Soc., 120, 4113, 1998. 42. Andrus, M.B. et al., Total synthesis of (+)-geldanamycin and (-)-quinogeldanamycin: asymmetric glycolate aldol reactions and biological evaluation, J. Org. Chem., 68, 8162, 2003. 43. Andrus, M.B., Meredith, E.L., and Soma Sekhar, B.B.V., Synthesis of the left-hand portion of geldanamycin using an anti-glycolate aldol reaction, Org. Lett., 3, 259, 2001. 44. Rinehart, K.L. et al., Synthesis of phenazine and phenoxazinone derivatives of geldanamycin as potential polymerase inhibitors, Bioorg. Chem., 6, 353, 1977. 45. Schnur, R.C. et al., Inhibition of the oncogene product p185erbB-2 in vitro and in vivo by geldanamycin and dihydrogeldanamycin derivatives, J. Med. Chem., 38, 3806, 1995. 46. Schnur, R.C. et al., erbB-2 Oncogene inhibition by geldanamycin derivatives: synthesis, mechanism of action and structure-activity relationships, J. Med. Chem., 38, 3813, 1995. 47. Prodromou, C. et al., Identification and structural characterization of the ATP/ADP binding site in the Hsp90 molecular chaperone, Cell, 90, 65, 1997. 48. Prodromou, C. et al., A molecular clamp in the crystal structure of the N-terminal domain of the yeast Hsp90 chaperone, Nat. Struct. Biol, 4, 477, 1997. 49. Kudkuk, S.D. et al., Synthesis and evaluation of geldanamycin-estradiol hybrids, Bioorg. Med. Chem. Lett., 9, 1233, 1999. 50. Kudkuk, S.D. et al., Synthesis and evaluation of geldanamycin-testosterone hybrids, Bioorg. Med. Chem. Lett., 10, 1303, 2000. 51. Zheng, F.F., et al., Identification of a geldanamycin dimer that induces the selective degradation of HER-family tyrosine kinases, Cancer Res., 60, 2090, 2000. 52. Sausville, E.A., Tomaszewski, J.E., and Ivy, P., Clinical development of 17-allylamino-17-demethox- ygeldanamycin, Curr. Cancer Drug Targets, 3, 377, 2003. 53. A Phase I and Pharmacologic Study of 17-(Allylamino)-17-demethoxygeldanamycin (AAG, NSC 330507) in Adult Patients with Solid Tumors, NIH Clinical Research Studies, Protocol Number: 99- C-0054. Available at: http://clinicaltrials.gov. 54. Tabibi, S.E. et al., Water-insoluble drug delivery system, U.S. Patent, 6,682,758 B1, 2004. 55. Kasuya, Y. et al., Improved synthesis and evaluation of 17-substituted aminoalkylgeldanamycin deriv- atives applicable to drug delivery systems, Bioorg. Med. Chem. Lett., 11, 2089, 2001. 56. Dictionary of Natural Products on CD-ROM, Version 12:2. Chapman & Hall, Boca Raton, FL, 2004. 57. Martin, S.F. et al., Novel approach to the ansamycin antibiotics Macbecin I and Herbimycin A. A formal total synthesis of (+)-Macbecin I, Tetrahedron, 52, 3229, 1996. Copyright © 2005 CRC Press, LLC 19 Bleomycin Group Antitumor Agents Sidney M. Hecht CONTENTS I. Introduction II. Mechanism of Action A. DNA Degradation B. RNA Degradation 1. Transfer RNAs and tRNA Precursor Transcripts 2. Other RNA Substrates for BLM 3. Chemistry of RNA Cleavage 4. Nonoxidative Cleavage of RNA C. Protein Synthesis Inhibition III. Development of Novel Antitumor Agents IV. Synthesis of Bleomycin V. Medicinal Chemistry A. Study of Individual “Amino Acid” Constituents 1. Bithiazole Moiety and C Substituent 2. Linker Domain 3. Metal-Binding Domain B. A BLM Combinatorial Library VI. Clinical Applications Acknowledgment References I. INTRODUCTION The bleomycins are structurally related antitumor antibiotics first isolated from Streptomyces verticillus as Cu(II) chelates by Umezawa and his colleagues.1,2 As illustrated in Figure 19.1, the bleomycins may be considered to be glycosylated oligopeptide-derived species, albeit composed of some rather unusual amino acid constituents. Some of these may actually be derived biosyn- thetically from proteinogenic amino acids, such as the bithiazole moiety, which is apparently formed by dehydrative cyclization and oxidation of β-alanylcysteinylcysteine.3 However, recent studies have revealed that the biosynthesis also depends in part on transformations used to assemble polyketide antibiotics.4 As discussed in some detail below, the bleomycins exhibit characteristic sequence-selective DNA binding and degradation properties. DNA affinity resides to a significant extent in the “tripeptide S”5,6 portion of bleomycin, which is composed of the contiguous threonine, bithiazole, and C-terminal moieties. The N terminus of bleomycin consists of the pyrimidoblamic acid and glycosylated β-hydroxyhistidine moieties, which are responsible for metal chelation and oxygen Copyright © 2005 CRC Press, LLC H2N O NH H H 2 linker region N NH2 metal binding O domain N N O O H H H O HO N 2N NH S N R H O HN O N HO N S H H O N HO OH H DNA binding domain O N O H OH O OH carbohydrate OH O moiety OH O NH2 H bleomycin A2 R = N S NH H bleomycin B2 R = N N NH2 H bleomycin A5 R = NH HN NH2 bleomycin A6 R = NH HN N NH H 2 FIGURE 19.1 Structures of some naturally occurring bleomycins. activation. These domains also bind to DNA and constitute the primary determinant of the sequence selectivity of DNA cleavage by bleomycin.7–10 The N- and C-terminal domains of bleomycin are connected by a linker, (2S, 3S, 4R)-4-amino- 3-hydroxy-2-methylvaleric acid. Although possible functions of this domain were unclear for some time, it has now been shown convincingly that this linker assumes a compact, folded structure that is essential for efficient sequence-selective DNA cleavage by bleomycin.11,12 Although the bleomycins have been used clinically for a number of years, the complex structures of agents in this class have thus far precluded the identification of synthetic congeners worthy of clinical development on the basis of improved properties. This chapter focuses in part on the unrealized potential of the bleomycins and the newly implemented synthetic strategies that may permit bleomycin analogs having improved properties to be prepared and characterized with facility. II. MECHANISM OF ACTION Early mechanistic studies of bleomycin explored the effects of this class of compounds on DNA and RNA polymerases, DNA and RNA nucleases, and DNA ligase.13-16 However, the finding that bleomycin could mediate damage to DNA in vitro17–20 and in vivo21–23 focused efforts on a detailed characterization of this process. A. DNA DEGRADATION Bleomycin-mediated DNA degradation requires a metal ion, which forms a complex with the N- terminal domain of bleomycin (cf. Figure 19.1). The first metallobleomycin studied was Fe(II)•BLM,24–26 but Cu(I)•BLM,27,28 Mn(II)•BLM,29,30 and VO(IV)•BLM31 have now been shown Copyright © 2005 CRC Press, LLC C HO O O – O P O O G O 5' O e CGCTAGCG Fe BLM Crieg e - type – GCGATCGC O2 O P O rearrangement 3' C O O HOO O – O P O OTAGCG C HO O C O HO – pTAGCG O P O O O G O O – NH anti - O P O 2 O G O N – elimination O P O O C O N O O O H – O P O O OOH O OH H O O – O P O 2 1 OTAGCG products formed from oxidative cleavage of cytidine3 FIGURE 19.2 Pathway leading to frank strand scission of DNA by Fe•BLM. to mediate DNA degradation as well. To the extent that information is available, these species appear to effect DNA degradation in the same fashion as Fe(II)•BLM. DNA degradation by Fe(II)•BLM also depends on the availability of O2, which is used to form an activated complex with the metallobleomycin. The nature of the activated Fe•BLM has been the subject of considerable debate as regards metal ion ligation sites, stoichiometry of the complex, and oxidation state of the metal ion.32 One useful insight regarding the nature of activated Fe•BLM derives from the obser- vation that activated Fe•BLM mediates chemical transformations of small molecules in a fashion most similar to that of cytochrome P450 and its analogs.28,33,34 This includes the oxidation and oxygenation of compounds such as stilbene and styrene and the NIH shift of deuterium in p- deuterioanisole.34 In further analogy with cytochrome P450 model compounds, Fe(III)•BLM can be activated with H2O 35,36 2 and with oxygen surrogates such as periodate and iodosylbenzene.34 The actual chemistry of DNA degradation by activated Fe•BLM has been studied using both DNA and DNA oligonucleotides. Because the cleavage of DNA by Fe•BLM is sequence selective, the use of DNA oligonucleotides containing a limited number of cleavage sites has permitted a detailed analysis of the nature of DNA degradation.17–19 Two pathways for Fe•BLM-mediated DNA degradation have been defined and are outlined in Figures 19.2 and 19.3. Shown in Figure 19.2 is a pathway that leads to frank DNA strand scission.37 The self-complementary octanucleotide d(CGCTAGCG) was employed as a substrate for Fe•BLM based on the earlier finding that the cleavage of DNA is sequence selective and involves primarily a subset of 5-GT-3 and 5-GC-3 Copyright © 2005 CRC Press, LLC HO C O O – O P O O G O CpGp O 5' Fe BLM OH– CGCTAGCG O pTAGCG GCGATCGC O OH 2 – 3' O P O O H O NH2 O OH OH N O O N – O P O H OTAGCG FIGURE 19.3 BLM-mediated formation and base-induced decomposition of the alkali–labile lesion. sequences. In fact, the substrate shown in Figure 19.2 was cleaved quite efficiently by activated Fe•BLM and predominantly at the two 5-GC-3 sequences. The observed

products, verified by independent chemical synthesis, were trans-3-(cytosin-1-yl)propenal (1), 5-GMP and a CpG deriv- ative terminating at its 3-end with a 3-(phosphoro-2-O-glycolate) moiety (2).7,31,38 The yield of base propenal was approximately equal to the sum of the yields of 5-GMP and 2.7,36 Although the relative amounts of 5-GMP and 2 formed could be altered by changing the BLM congener employed, in all cases the amount of 1 formed closely approximated the combined yields of 5-GMP and 2. The mechanistic scheme shown in Figure 19.2 rationalizes the nature and the amounts of products obtained. Initial abstraction of a H atom would afford a C4 deoxyribose radical that could combine with O2 to form a C4 hydroperoxy radical. Reduction and protonation of this intermediate, or recombination with a H atom, would afford a C4 hydroperoxide, the latter of which would be expected to undergo a Criegee-type rearrangement, as shown, with scission of the C3-C4 deoxyri- bose bond. Conversion of the resulting hemiacetal to the aldehyde, and elimination of elements of glycolic acid, would then afford the observed products. It may be noted that in the example shown in Figure 19.2, CpGpCH2 COOH (2) would result only from cleavage at cytidine3, while 5-GMP would result only from cleavage at cytidine7. Two other oligomeric products would be expected to form (one oligomeric 5-phosphate and one oligomeric 3-phosphoroglycolate) but were not conveniently analyzed. It may be noted that a reasonable alternative for the scheme shown in Figure 19.2 has been proposed.39 In addition to the products shown in Figure 19.2, another set of products was also formed.40–43 These included free cytosine and a DNA lesion denoted the alkali labile lesion, as subsequent treatment with alkali resulted in strand scission. Strand scission was found to occur at the same sites as those outlined in Figure 19.2. In fact, the proportion of frank strand scission products and alkali labile lesions was a function of oxygen tension; higher oxygen tension favored frank strand scission products.43–45 The structure determined for the initially formed alkali labile lesion was that of a C4-OH apurinic acid, which presumably arose from depurination of 4-OH deoxycytidine (Figure 19.3). Because the ratio of products formed via the two pathways was influenced by the amount of oxygen available, the two pathways likely have a common intermediate.44,45 It seems likely that in the absence of sufficient O2, the initially formed C4 deoxyribose radical (Figure 19.2) undergoes a further one-electron oxidation to afford the carbocation.46–48 The latter could then react with water, affording the putative 4-OH deoxycytidine intermediate. Independent support for the intermediacy of a C4-deoxyribose radical has been obtained by the use of a C4-deuterated DNA oligonucleotide as a substrate for Fe(II)•BLM.49 A primary kinetic isotope effect was observed at individual thymidine residues, indicating that abstraction of the C4-H is actually rate limiting for Fe•BLM-mediated DNA degradation. Copyright © 2005 CRC Press, LLC It is interesting to note that although activated Fe•BLM is capable of effecting both the oxidation and oxygenation of small molecules,50,51 the degradation of DNA according to the pathways shown in Figures 19.2 and 19.3 involves only DNA substrate oxidation.48 Dioxygen is required both for Fe(II)•BLM activation and for reaction with the putative C4 deoxyribose radical but is apparently not transferred to DNA from the oxygenated metalloBLM. The catalytic cycle for Fe(II)•BLM was studied further using both electrochemical activation at a glassy carbon electrode52 and 17O nuclear magnetic resonance spectroscopy.53 Electrochemical activation permitted correlation of the number of electrons consumed and DNA oligonucleotide products produced under conditions in which each Fe•BLM molecule was employed up to about 13 times in producing DNA lesions. In the electrochemical system, it was also clear that Fe(II)•BLM could not be activated reductively once it had bound to DNA. Thus, the obligatory order of events must have been Fe(II)•BLM binding by O2, reduction to afford activated Fe•BLM, and then DNA binding and cleavage.52 Fe(II)•BLM activation was also studied by 17O nuclear magnetic resonance spectroscopy in the absence and presence of a DNA substrate.53 These studies verified that in the absence of any DNA substrate, Fe(II)•BLM catalyzes a 4 e- reduction of O2 to H2O. The production of H 17 2 O in the presence of DNA was entirely consistent with the mechanisms outlined in Figures 19.2 and 19.3. An additional noteworthy feature of DNA cleavage by Fe(II)•BLM is the formation of double- strand DNA breaks. The proportion of double-strand cleavage is a function both the specific BLM congener54 and the DNA substrate employed55–58; under favorable circumstances, such breaks can constitute up to 20% of all DNA lesions. It has been demonstrated convincingly that double-strand cleavage results from a specific set of mechanistic events at a DNA sequence conducive to such cleavage55–58 and not from the random accumulation of single-strand breaks. For some DNA sequences, it has been argued that a significant rearrangement of structure must occur after the first break to orient the activated BLM in a fashion conducive to the cleavage of a second strand.58,59 However, at least for one substrate, this would seem to be unnecessary. The oligonucleotide shown in Figure 19.4 underwent 43% double-strand cleavage from a bound conformation in which bleomycin was found to be oriented close to the cleavage sites on both strands.60 Although the individual metallobleomycins noted above all cleave DNA in a superficially similar fashion, for most metallobleomycins other than Fe•BLM there is a paucity of information on which to base mechanistic comparisons. Those species that have been studied to some extent include Cu(I)•BLM, which has been shown to form a metal complex with BLM quite different than that formed by Fe(II),27,61 to undergo aerobic activation much less quickly than Fe(II)•BLM,28 and to produce sequence specific DNA lesions in different ratios.28 Co(III)•BLM has a signifi- cantly higher DNA affinity than other metalloBLMs,62 requires light but not oxygen for DNA cleavage,63–67 binds and cleaves DNA at some sites distinct from those targeted by Fe(II)•BLM,68 and cleaves DNA with minimal production of the compounds formed in Figure 19.2.67 This metalloBLM has also been postulated to employ a reactive species different than activated Fe•BLM.68 Accordingly, mechanistic generalizations between metalloBLMs should be made only with great caution. FIGURE 19.4 Oligonucleotide substrate that underwent efficient double strand cleavage by Fe•BLM. Copyright © 2005 CRC Press, LLC B. RNA DEGRADATION 1. Transfer RNAs and tRNA Precursor Transcripts Following an initial report by Magliozzo et al.69 in which a high (300 µM) concentration of activated Fe•BLM was shown to effect limited degradation of yeast tRNAPhe and a few other tRNA isoac- ceptors, the Hecht laboratory carried out a systematic study of RNA degradation by Fe(II)•BLM.70–74 For transfer RNAs and tRNA precursor transcripts, most of the species studied were refractory to activated Fe•BLM at all tested concentrations. A few tRNAs were cleaved with modest efficiency at intermediate concentrations, notably a Schizosaccharomyces pombe amber suppressor tRNASer construct and mature Escherichia coli tRNA His.73 1 By far the most efficient cleavage was achieved with a Bacillus subtilis tRNAHis precursor (Figure 19.5), which was cleaved with surprising effi- ciency at U35 and weakly at several other positions.70,73 Although examples of cleavage occurred in many positions of the cloverleaf structure of tRNA, and at sequences not well represented in DNA cleavage studies, in the aggregate the majority of cleavage sites involved 5-G•pyr-3 sequences and were located at what are believed to be the junctions between single- and double-stranded regions of the tRNAs.10,70–73 Two tRNA substrates that were cleaved with reasonable efficiency by Fe•BLM were particularly helpful in defining the RNA microenvironments conducive to cleavage. These were yeast tRNAPhe and cytoplasmic yeast tRNAAsp,75 the structures of which have been defined crystallographically.76–80 By systematic variation of these two RNA structures, Giegé, Florentz, and coworkers have defined the identity elements for these two RNAs required for recognition by their cognate aminoacyl–tRNA synthetases.81–85 The use of these hybrid tRNA structures as substrates for Fe•BLM revealed a surprisingly complex set of cleavage patterns.75 Perhaps more than any other study, this report underscored the ability of Fe•BLM to recognize subtle changes in RNA structure. 2. Other RNA Substrates for BLM Although most of the studies of RNA cleavage by bleomycin have involved tRNAs, several other RNAs have been tested as substrates. Perhaps the most successful of these has been yeast 5S ribosomal RNA, which was found to be cleaved quite efficiently at three sites by Fe(II)•BLM.73 Interestingly, each of the three cleavage sites involved the uridine moiety in a 5-GUA-3 sequence, and each had a one-base bulge on the cleaved strand one or two nucleosides to the 3-side of the cleavage site. However, it has not been shown that any rRNA within an intact ribosome is cleaved by Fe(II)•BLM, and studies of the effects of Fe(II)•BLM on a eukaryotic protein synthesizing system indicated that ribosomes are not a primary target for BLM (vide infra).86 FIGURE 19.5 A Bacillus subtilis tRNAHis precursor transcript that undergoes efficient cleavage by Fe•BLM. Copyright © 2005 CRC Press, LLC Several messenger RNAs have been studied as substrates for Fe(II)•BLM. RNAs found to undergo cleavage have included HIV-1 reverse transcriptase mRNA,70 the iron regulatory element of bullfrog ferritin mRNA,87 firefly luciferase mRNA,86 and E. coli dihydrofolate reductase mRNA.88 Also studied has been the cleavage of an RNA–DNA heteroduplex prepared by reverse transcription of E. coli 5S rRNA.74 Both the RNA and DNA strands were cleaved under comparable conditions and at a limited number of sites. Significantly, cleavage of the RNA strand of the heteroduplex involved cleavage sites different than those observed for the original rRNA, indicating that cleavage must actually have involved a heteroduplex structure. 3. Chemistry of RNA Cleavage The chemistry of RNA cleavage by Fe(II)•BLM has been studied less intensively than that of DNA, but a few useful observations have been made. By the use of yeast 5S rRNA 32P-labeled alternatively at the 5 and 3 ends, it was shown that the mobility of individual bands was consistent with that expected if the chemistry of RNA cleavage was analogous to that of DNA.73 This conclusion was also supported by experiments using the self-complementary oligonucleotide CGCTAGCC in which deoxycytidine3 was replaced by ribocytidine.70 However, the latter experiment indicated the presence of greater amounts of free cytosine than would have been anticipated on the basis of products resulting from initial abstraction of C4 H (cf. Figures 19.2 and 19.3). In the belief that an altered conformation at position 3 in the minor groove might lead to H atom abstraction from C1, the initially formed products were treated with diaminobenzene in the hope of capturing the putative intermediate (i) resulting from Criegee-type rearrangement of a C1 hydroperoxynucleotide (Figure 19.6). In fact, a product having chromatographic properties identical with synthetic quinoxaline 3 was observed. Assuming that this product arose by the pathway shown in Figure 19.6, about 10% of the products resulting from admixture of C3-ribo CGCTAGCG and Fe(II)•BLM were formed by initial abstraction of C1 H from ribocytidine3.89,90 Repetition of this experiment using C3-ara CGCTAGCG gave an amount of 3 corresponding to 58% C1 H abstraction.89 This is consistent with the known conforma- tional alteration of the latter species.91 Not yet resolved experimentally is the issue of the formation of an alkali labile lesion from RNA analogous to the intermediate formed from DNA (Figure 19.3). Although the intrinsic lability of the RNA internucleotide linkage to alkali precludes the use of this reagent to unmask the alkali labile lesion selectively, it has been shown by admixture of hydrazine to Fe•BLM-treated B. subtilis tRNAHis precursor transcript that some currently uncharacterized lesion is present.90 4. Nonoxidative Cleavage of RNA One particularly surprising finding has been reported by Keck and Hecht.92 In the absence of any added metal ion, BLM A2 was found to mediate RNA strand scission at a number of sites. Cleavage was found to be suppressed by admixture of metal ions that bind to BLM. As illustrated in Figure 19.7A, cleavage was observed at all unmodified pyrimidine–purine sites in the tRNA. Analysis of the nature of the formed cleavage products indicated that each contained a 2,3-cyclic phosphate at the 3 end. Logically, these might be thought to have resulted from a phosphoryl transfer reaction in which the internucleotide phosphate linkage at certain positions underwent nucleophilic attack by the 2-OH group of ribose. Oligonucleotide 2,3-cyclic phosphate moieties could then form concomitant with RNA strand scission (Figure 19.7B).

Although the possible relevance of this transformation to the mechanism of action of BLM is unclear, it may be mentioned that the facility of this transformation is entirely comparable to that involving oxidative RNA cleavage by Fe(II)•BLM (vide infra). C. PROTEIN SYNTHESIS INHIBITION The ability of Fe(II)•BLM to cleave some tRNAs, mRNAs, and rRNAs suggested that this metallob- leomycin might be able to inhibit the synthesis of proteins. This was studied in a DNA-independent Copyright © 2005 CRC Press, LLC O O – O P O O G O 5' egee - type CGCTAGCG Fe BLM O Cri – GCGATCGC O O P O rearrangement 3' 2 O C O OOH O OH – O P O OTAGCG OH C O O OH C – TAGCG O P O O p O G H2N O O H O – 2 NH 2N C O P O 2 O O G – O P O O N O C O N O H O – O O P O O – CHO O O P O N H OTAGCG N i 3 FIGURE 19.6 Products potentially accessible from RNA by abstraction of the C1 H of deoxyribose by Fe•BLM. C represents aracytidine or ribocytidine. cell-free protein biosynthesizing system containing rabbit reticulocyte lysate and programmed with luciferase mRNA.86 As shown in Table 19.1, treatment with 100 µM BLM A5 alone reduced luciferase synthesis by ~26%, either because of adventitious metal ions or because of nonoxidative RNA cleavage, as described above. Treatment with 50 µM Fe(II)•BLM A5 reduced protein synthesis by about 68%, and 100 µM Fe(II)•BLM A5 caused ~86% reduction. To determine which RNAs had been affected by Fe(II)•BLM A5, the reaction mixtures were treated with additional ribosomes, luciferase mRNA, or a mixture of tRNA isoacceptors following Fe(II)•BLM treatment. As indicated in the table, only admixture of tRNAs had any effect in retaining luciferase synthesis. Production of this protein was completely restored for the treatment involving 50 µM Fe(II)•BLM and was partially restored for the 100-µM treatment. Thus, one or more transfer RNAs may represent a critical cellular target for BLM. In this context, it is interesting to note that the antitumor ribonuclease onconase93–97 has been employed in an experiment analogous to the one summarized in Table 19.1, and virtually identical results were obtained.98 Because onconase and BLM have been shown to have remarkably similar cytotoxicity profiles toward a panel of cultured tumor cells,86 the specific tRNAs cleaved by both agents were compared. The two antitumor agents were found to effect cleavage of several tRNAs in common, although a comparison of cleavage patterns with one isoacceptor revealed cleavage at quite different sites in that tRNA. Copyright © 2005 CRC Press, LLC (A) (B) FIGURE 19.7 (A) Sites of cleavage of tRNAPhe by metal-free BLM. Major cleavage sites are in boxes; minor cleavage sites in circles. (B) Chemistry of tRNA degradation by metal-free BLM. Cleavage between C72 and A73 is exemplified. Although the specific tRNA(s) exhibiting maximal sensitivity to these agents have yet to be reported, it is interesting to note that following treatment of uninfected or HIV-1 infected H9 lymphocytes with onconase, tRNA Lys 3 was found to be cleaved by onconase.99 Although the role of RNA cleavage in the cancer chemotherapeutic effects of BLM remains to be established, the foregoing experiments support the possible involvement of one or more tRNA isoacceptors. III. DEVELOPMENT OF NOVEL ANTITUMOR AGENTS The earliest members of the bleomycin family to be identified were the phleomycins, which contain a thiazolinylthiazole moiety rather than the bithiazole present in bleomycin (cf. Figures 19.1 and 19.8). Copyright © 2005 CRC Press, LLC TABLE 19.1 Effect of tRNAs on Luciferase Synthesis Following Treatment with Fe(II)•BLM A5 Addition Luciferase Synthesis (%) –– 100 100 µM Fe2+ 113 100 µM BLM A5 74 50 µM Fe(II)•BLM A5 32 50 µM Fe(II)•BLM A5, then 40 µg/mL tRNAs 120 100 µM Fe(II)•BLM A2 14 100 µM Fe(II)•BLM A2, then 40 µg/mL tRNAs 53 The phleomycins and bleomycins each contain several different C substituents, as illustrated in Figures 19.1 and 19.8. Although the members of both series each cleave DNA with similar characteristics with regard to sequence selectivity, individual congeners do display different biochemical properties, even at the level of cell culture.100 Although discovered before the bleomycins, the phleomycins exhibited unacceptable toxicity when tested as a mixture of congeners. Further, the thiazoline moiety is susceptible to hydrolytic ring opening and possibly also to racemization. For these reasons, attention was focused on the bleomycins as clinical candidates. The complexity of the bleomycin structure precluded total chemical synthesis at the time of their initial isolation. Accordingly, most of the congeners studied were naturally occurring variants that differed exclusively at the C terminus (Figure 19.1). It was also found that additional congeners H O 2N NH H H 2 N NH2 N N O O H O H O H H N H N 2N O H O NH S R HN O N HO N S H H O N H HO OH O N O H OH O OH OH O OH O NH2 H NH N NH2 phleomycin D1 R = NH NH H H NH N N HN NH2 phleomycin E R = NH NH H H H NH N N HN N NH NH 2 phleomycin G R = NH NH FIGURE 19.8 Structures of some naturally occurring phleomycins. Copyright © 2005 CRC Press, LLC H2N O NH H H 2 N NH H O 2 2N H N N O NH2 O H2N N N H O 2N NH H H 2 Cu2+ O H2N N NH2 epi BLM O depyruvamide BLM N N O O H H O H HO N 2N O NH S N NHR H HN O N HO N S H H H O 2N NH O N O H H 2 BLM H N OH hydrolase HO OH OH– O O OH N O H OH N N O H OH O OH 2O OH O O H N OH H NH 2 O 2O 2 O OH O NH iso BLM O deamido BLM 2 O OH H O 2N NH OH H H 2 OH + N NH H3 O OH 2 decarbamoyl BLM N N O O H O H O H N HO N 2 O NH S N NHR H HN O N HO N S H H HO N H N H deglyco BLM FIGURE 19.9 Analogs of bleomycin accessible by semisynthesis from BLM itself. differing at the C terminus could be prepared via fermentation by including in the incubation medium a high concentration of the amine intended for incorporation into the C terminus of bleomycin.101 Semisynthetic bleomycins differing at the C terminus were also accessible following initial chemical removal of the C substituent of bleomycin A2; the latter was realized by means of a series of chemical transformations to afford bleomycinic acid.102–104 Treatment of bleomycin B2 with an acylagmatine hydrolase also provided access to bleomycinic acid.105 Condensation of bleomycinic acid with different polyamines, for example, then afforded new analogs of bleomycin. Although most of the early structural modifications of the bleomycins were limited to alterations at the C terminus, several other types of transformations of the natural product were reported. As summarized in Figure 19.9, this included the preparation of epi-BLM, iso and decarbamoyl BLMs, and deamido and depyruvamide BLMs.106 These were found to differ somewhat in their animal organ distribution profiles, as well as in their intrinsic biological properties.106 One of the most remarkable analogs was deglycoBLM, the simplified analog lacking the carbohydrate moiety of BLM.107,108 This species was found to mediate DNA cleavage with almost the same sequence selectivity as BLM itself and to produce the same chemical products.108,109 Other naturally occurring variants of the BLMs were also identified. Perhaps the best known are the tallysomycins, which have also been used in clinical trials. As shown in Figure 19.10, the tallysomycins have a third carbohydrate moiety, in the form of a talose residue, which is attached to the peptide backbone of tallysomycin via a structurally novel glycosylcarbinolamide linkage. The tallysomycins cleave DNA predominantly at 5-GC-3 and 5-GT-3 sequences, as do the bleomycins, but the patterns of preferred sites of cleavage are somewhat altered.110 Copyright © 2005 CRC Press, LLC O NH2 NH H 2 HO NH2 N NH2 O OH O H X– O N N O O S N NH S H H O OH H N 2N O N N S H H HN O OH N HO H H O N HO HO H O O N H HO O OH OH O OH O NH2 FIGURE 19.10 Structure of tallysomycin S2B. In the United States, Europe, and Japan, bleomycin is used clinically as a mixture of congeners that differ at the C terminus. The major constituents are bleomycin A2 (~60%–65%) and bleomycin B2 (~30%). In animal tumor models, this mixture was reported to exhibit efficacy better than single congeners.1 It has been reported that BLM congeners having free amino groups within the C substituent exhibited enhanced pulmonary toxicity — an observation that led to the preparation and evaluation of analogs such as liblomycin111 and BAPP.104,106 Interestingly, BLM A5 (which has free amino groups within the C substituent in the form of a spermidine) is used clinically in Russia and China. IV. SYNTHESIS OF BLEOMYCIN Three laboratories have reported the synthesis of naturally occurring bleomycins. The first two syntheses were reported in 1982 by the Umezawa112 and Hecht113 laboratories. The development of methodology relevant to the synthesis of bleomycin has continued to this time. As a result, two additional syntheses of bleomycin have been described recently by the Boger114 and Hecht115 laboratories, as have the syntheses of closely related species.54,115–117 Of special interest in this context has been the recent report of the total synthesis of deamidobleomycin, the sole reported catabolite formed following therapeutic administration of bleomycin.117 An important advance in the synthesis of bleomycin congeners was realized in 2000 with the first report of the preparation of a deglycobleomycin by solid-phase synthesis.118 The synthesis involved the use of Tentagel resin functionalized with a trityl linker. The “amino acid” constituents of bleomycin were condensed stepwise onto the resin-bound linker from the C terminus. Following condensation of Boc pyrimidoblamic acid,119 the formed deglycoBLM A5 was deblocked and cleaved from the resin by treatment with trifluoroacetic acid/trisopropylsilane/water/dimethylsulfide.118 Although the use of a linker incorporating a trityl group did provide synthetic access to deglycoBLM, there were some limitations to this strategy. One critical limitation was the obligatory cleavage of deglycoBLM from the resin concomitant with removal of the Boc protecting group. While reducing the number of steps in the overall procedure, this simultaneous deblocking/cleavage precluded the elaboration of fully deblocked deglycoBLM still attached to the synthesis resin. Because it had previously been shown that natural bleomycins conjugated to solid supports via their C substituents retained the ability to effect sequence-selective DNA cleavage,120 it seemed desirable to be able to test the properties of newly synthesized deglycoBLMs before their removal from the synthesis resin. The use of a dimedone-based resin linker, cleavable by treatment with hydrazine, proved successful in permitting the elaboration of fully protected deglycoBLMs attached to a Tentagel Copyright © 2005 CRC Press, LLC 1. Boc H N 2N N NH2 Boc (i-Pr)2NEt, DMF 1. 20% piperidine/DMF 1. 20% piperidine/DMF O O O O N N O H H S N OH H FmocHN FmocHN OH HOO N S FmocHN HO OH H HBTU, (i-Pr)2NEt, DMF HBTU, HOBt, (i-Pr)2NEt, DMF HBTU, HOBt, (i-Pr)2NEt, DMF 1. 20% piperidine/DMF 1. 20% piperidine/DMF 90:5:5 TFA:i-Pr3SiH/Me2S 2. HO 2. O NH2 FmocHN NHBoc OH H H N N NH2 N N N O Tr H OH HATU, HOBt, (i-Pr)2NEt, DMF 2N O H O 2N NH H H 2 BOP, (i-Pr)2NEt, DMF N NH2 N N O O H H O O HO H N N 2N NH S NH HN H O H HN O N HO N S NH HN N H H N O O O N H 2% H2NNH2 deglycobleomycin A6 DMF FIGURE 19.11 Scheme employed for the solid-phase synthesis of deglycoBLM. support.121 The synthetic scheme is outlined in Figure 19.11. As shown, after attachment of each Fmoc-protected fragment to the resin-bound linker, subsequent treatment

with piperidine effected Fmoc cleavage and formation of a dibenzofulvene-piperidine adduct, which has been characterized spectroscopically.122 This permitted stepwise coupling yields to be determined. The synthesis of deglycoBLM A2-a was realized in an overall yield of 68% using this procedure.121 In a parallel study, optimization of the C substituent permitted determination of the C substituent optimal to permit on-resin assay of the synthesized deglycoBLM.123 A further study addressed the issue of preparing by solid-phase synthesis BLMs containing the disaccharide moiety of BLM itself, or other carbohydrate analogs. As outlined in Figure 19.12, this was realized by preparing a suitably protected β-hydroxyhistidine derivative containing the disac- charide moiety present in BLM. The synthesis was based on earlier syntheses described by Ohno124 and Boger125 but employed protecting groups compatible with the solid-phase synthesis outlined in Figure 19.11. In this fashion, bleomycin A5 was synthesized in reasonable yield,126 and with much greater facility than the respective solution phase synthesis.115 As described in a section below, this strategy has also been employed successfully to prepare carbohydrate analogs of bleomycin. It may be mentioned that the stepwise linear construction realized by solid-phase synthesis was different than the coupling scheme employed in any of the solution phase syntheses. As regards the preparation of other bleomycin group antibiotics, the absolute stereochemistry of the methine H in the thiazoline moiety of PLM has been described,127 potentially enabling its total synthesis. However, the synthesis of phleomyin itself has not been reported as yet. In contrast, Sznaidman and Hecht recently reported a strategy for elaborating the glycosylcarbinolamide moiety of tallysomycin128 in spite of the fact that the absolute stereochemistry has not yet been defined at some of the stereocenters. Copyright © 2005 CRC Press, LLC OH HN N 1. RdCI 2. MnO2 49% H O O O O 1. nBuLi B Tr N N O NH 2. bromoacetyl bromide O N Br Bu2BOTf O O 53% O N Br 41% 1. LiOH OH O O 2. 10%Pd/C; H O H OH O O 2 3. FmocOSu FmocHN OBn N NaN N N O 3 N O 4. BnN2, EtOAc HO N N Br 60% N N3 46% H N Tr Tr Tr O H O FmocHN OH O AcO OAc O H O O P Ph OPh 1. TMSOTf, CH2CI O 2 O N FmocHN OBn 2. 20% Pd(OH)2/C, H2 AcO OAc H AcO O N OAc O HO N O OAc 49% Tr H AcO O AcO N O OAc O OAc Tr NH AcO O 2 O NH2 FIGURE 19.12 Synthesis of the carbohydrate moiety of BLM protected in a fashion suitable for use in the solid phase synthesis of the natural product. V. MEDICINAL CHEMISTRY A. STUDY OF INDIVIDUAL “AMINO ACID” CONSTITUENTS 1. Bithiazole Moiety and C Substituent In contrast to efforts on the total synthesis of bleomycin, which have been pursued for 30 years, systematic evaluation of those structural elements in bleomycin essential for its nucleic acid degradation properties has been addressed only more recently. Alteration of the bithiazole moiety in a systematic fashion has yet to be reported, although several studies have contributed to our understanding of the way this moiety functions in DNA binding. These include photochemical transformations of the bithiazole moiety129,130 and replacement with mono and bishalogenated bithiazoles.131 Although neither of these types of alterations effected obvious changes in the DNA- binding properties of the derived (deglyco)BLMs under dark conditions, irradiation of the degly- coBLMs containing halogenated thiazoles did permit confirmation that these structural elements can reside in the minor groove of DNA.131 The ability of the tripeptide S moiety of BLM (consisting of the threonine, bithiazole, and C- terminal substituent) to bind DNA with essentially the same affinity as BLM5,6 argues that much of the DNA binding affinity of BLM resides within this domain. Fully consistent with this inter- pretation were the findings that deglycoBLM analogs containing monothiazole constituents cleaved DNA inefficiently and without demonstrable sequence selectivity.132 As might have been expected, Copyright © 2005 CRC Press, LLC the replacement of the bithiazole moiety with a pyrrole-based heterocyclic system resulted in a change in sequence selectivity to afford cleavage at sequences reminiscent of the sequence-selective binding of distamycin and netropsin.133,134 On the basis of the foregoing observations, it might have been reasonable to conclude that the sequence selectivity of DNA cleavage by the (deglyco)BLMs is determined by the bithiazole moiety, as actually suggested in an early study.135 However, there are now a few lines of experimentation that make it clear that the actual situation is more complex. These include the findings that photolysis of halogenated bithiazoles in the presence of DNA resulted in sequence-selective DNA damage, albeit not with a sequence selectivity characteristic of BLM,136,137 whereas an analog of the pyrim- idoblamic acid moiety effected sequence selective cleavage of DNA in the same fashion as BLM.9 Also, BLM congeners having the same bithiazole + C terminus, but differing in their metal binding domains, exhibited altered strand selectivity of DNA cleavage.7 Perhaps the most compelling evidence was obtained for a series of analogs of deglycoBLM in which the threonine moiety was replaced by glycine or oligoglycine, thus systematically altering the distance between the metal binding and bithiazole domains.8 On the assumption that increasing the number of glycines between the two binding domains constrained them on average to be at greater distances from each other when bound to a DNA substrate, the pattern of DNA cleavage should then reflect which binding partner (i.e., the metal binding domain or bithiazole) is the predominant structural element in defining the sequence selectivity of DNA binding. It was shown experimentally that all of the analogs were cleaved at the same sites, arguing that the metal binding domain is the primary determinant of the sequence selectivity of DNA binding by BLM.8 The selectivity for 5-GC-3 and 5-GT-3 sequences may well reflect the reality that the metal binding domain of BLM is able to bind to DNA the most effectively at the widest, shallowest part of the minor groove.10 This suggestion is supported by modeling studies138,139 and by the observation that BLM cleaves DNA and RNA duplexes most efficiently at sites having a widened minor groove. Examples might be thought to include DNA140 and RNA bulges,73 sites of DNA platination,141 the 5-junction of a DNA triplex structure with its DNA duplex target,142 and the putative junctions between single- and double-stranded structures in RNAs.70,73 Although not yet established experimentally, it seems likely that the most efficient cleavage of DNA by BLM occurs at those sites that permit efficient binding and H atom abstraction by the metal binding domain of BLM while also allowing the bithiazole to bind to its own preferred sites. Naturally occurring BLMs have several different C substituents, most of which are polyamines, or polyamines containing guanidine moieties. Virtually all of the naturally occurring BLMs have C substituents that are charged at physiological pH, and the charged C substituent is clearly important for productive binding to the (negatively charged) nucleic acid substrates for BLM. BLM congeners lacking positively charged C substituents, notably (deglyco)BLM demethyl A2, have been demonstrated to bind to and cleave DNA less efficiently. Although not studied in a thorough fashion, it appears that two positively charged substituents within the C substituent123 that are appropriately positioned143 are sufficient to support DNA cleavage by that (deglyco)BLM congener in an optimal fashion. Although the nature of the C substituent has often been ignored in studies of (deglyco)BLM, cell culture studies make it clear that this substituent is of considerable importance from the perspective of medicinal chemistry.100 2. Linker Domain Several lines of evidence have suggested that (deglyco)BLM assumes a folded conformation when it is bound to a DNA duplex substrate.11,12,139,144,145 The structural elements that control this folded conformation have now been defined convincingly by studies from the Boger and Stubbe labora- tories.11,12 Alterations of the threonine moiety did not affect the sequence selectivity of DNA cleavage but did affect efficiency of cleavage, as shown earlier by the Hecht laboratory.8 The ratio of single- to double-strand cleavage was also affected by changes in the threonine moiety.11 Careful Copyright © 2005 CRC Press, LLC analysis of the effect of substituents within the methylvalerate moiety indicated that substituents in the 2- and 4-positions are critical in facilitating folding into a stable conformation optimal for DNA cleavage.11,12 Recently, an effort has been made to exploit the putative folded conformation of (deglyco)BLM by the preparation of conformationally rigid analogs containing structurally altered methylvalerate moieties.146,147 Although the prepared analogs were capable of binding Fe2+ and activating oxygen for transfer to low–molecular weight substrates, none mediated sequence- selective DNA cleavage in the same fashion as BLM. The most likely interpretation of these findings is that none of the analogs was a sufficient mimic of the preferred folded conformation of BLM to permit efficient DNA cleavage to proceed. It may be noted, however, that it is possible that productive DNA binding and cleavage by BLM may actually require some flexibility within the ligand for mechanistic reasons. It seems possible that this structural domain of BLM may be of use in differentiating the binding of different classes of substrates (e.g., DNA and RNA) by BLM. 3. Metal-Binding Domain The metal-binding domain of BLM is composed of pyrimidoblamic acid and (glycosylated) β- hydroxyhistidine moieties. Although many of the early analogs of BLM, prepared by semisynthesis from naturally occurring BLMs, were altered within this domain, relatively little has been done to systematically vary this domain. Changes made within the β-hydroxyhistidine domain of (deglyco)BLM have mainly involved those reported by Boger and coworkers. These have been important in defining the role of the π- nitrogen of the imidazole moiety in sequence selective DNA cleavage,148 consistent with the role of this N atom as a metal ion ligand, and the lack of importance of the OH group of β-hydroxy- histidine to the function of deglycoBLM as a DNA cleaving agent.149 Recently, some reports have appeared that are pertinent to the roles of the disaccharide moiety of BLM. These have included two studies in which monosaccharide derivatives have been prepared that explored the role of the L-gulose moiety. Both studies found that monosaccharide analogs of BLM containing L-gulose attached to β-hydroxyhistidine were the most effective in maintaining potency of DNA cleavage.116,126 This monosaccharide analog was also found to be more efficient in cleaving a good RNA substrate for BLM than congeners containing D-mannose or L-rhamnose.126 It seems clear that the carbohydrate moiety of BLM is exceptionally important in determining the efficiency of RNA cleavage by individual BLM analogs. B. A BLM COMBINATORIAL LIBRARY Although the studies described above have provided a good starting point for a program of systematic modification of BLM structure to identify analogs with improved properties, they provide little insight into the potential effects of preparing analogs that differ from BLM in multiple positions. Given the complexity of the structure of BLM, multiple alterations represent a particularly attractive opportunity to derive congeners having properties substantially different than those of the parent antitumor agent. To address this opportunity, the Hecht laboratory used its methodology for solid-phase BLM synthesis (cf. Figures 19.11 and 19.12) to prepare a library of 108 deglycoBLM analogs through parallel solid-phase synthesis.150 The library, prepared with the spermine C substituent of BLM A6, involved the use of three analogs of the bithiazole moiety, three analogs of threonine, three analogs of the methylvalerate moiety, and four analogs of β-hydroxyhistidine (Figure 19.13). Each of the analogs was purified by preparative reversed-phase high performance liquid chromatography, char- acterized by 1H nuclear magnetic resonance spectroscopy and mass spectrometry, and then used to assess DNA cleavage. A few important lessons were learned from this initial library, including the observations that a number of the analogs retained the ability to effect sequence-selective DNA cleavage, and two Copyright © 2005 CRC Press, LLC O CI O H O H O FmocNH S N OH FmocNH S N OH FmocNH OH FmocNH OH N S N S O HO CH H S O H 3 FmocNH OH S FmocNH S N N OH N S tBuO bithiazole analogues threonine analogues H H H O H O HO OH HO OH FmocNH OH FmocNH OH FmocNH O FmocNH O N H H O H N O S H HO OH Tr H FmocNH OH FmocNH

OH FmocNH O H S N Boc methylvalerate analogues histidine analogues FIGURE 19.13 Building blocks used for the preparation of a deglycoBLM library. of the analogs were more potent than the parent molecule from which they were derived in mediating relaxation of a supercoiled plasmid DNA. Critically, for any amino acid constituent whose inclusion rendered the parent molecule poorly functional in DNA cleavage, no combination of changes elsewhere in the molecule were capable of restoring function. The ability to prepare parallel libraries of BLMs on beads should be readily extensible to mix- and-split libraries [i.e., libraries containing mixtures of beads, each of which contains multiple copies of a single (deglyco)BLM congener]. In concert with the use of DNA molecular beacons already described as substrates for BLM,151 in principle it should be possible to identify BLM congeners with improved properties by a process of selection rather than design. Clearly, the realization of this strategy could dramatically accelerate the development of our understanding of the medicinal chemistry of BLM. VI. CLINICAL APPLICATIONS Bleomycin has been employed in the clinic as an antitumor agent for more than 25 years; the early studies and major findings have been summarized in two books.152,153 Bleomycin was found to have clinically apparent activity as a single agent when used to treat squamous cell carcinomas, including those of the skin and head and neck.154 Although initial studies involved limited numbers of patients, good response rates were also observed in the treatment of primary brain tumors and superficial bladder tumors. A more usual application for bleomycin at this time is in combination with other antitumor agents. Bleomycin is effective as part of a regimen for the treatment of Hodgkin’s lymphomas, testicular cancers, germ cell ovarian cancers, and some non-Hodgkin lymphomas.155,156 One feature that has promoted the study of bleomycin as an agent in combination therapy is its lack of myelosuppressive activity. Pulmonary fibrosis, observable in about 10% of patients, is generally the dose-limiting toxicity for bleomycin in the clinic.157,158 Bleomycin is well established in the clinic as an antitumor agent in certain regimens, where it has been shown to contribute to favorable clinical outcomes. A good example is its activity in the treatment of germ cell carcinomas.159 In spite of its long history of clinical trials and applications, new uses continue to be sought for this agent. Numerous clinical trials have been reported within the last several years for a variety of indications. Recent trials with potentially promising outcomes include those involving BLM treatment of symptomatic hemangiomas160 in Copyright © 2005 CRC Press, LLC children and craniopharyngiomas.161–163 Bleomycin has also been used recently with cisplatin and vinblastine in a comparative study for the treatment of epithelial ovarian cancer164 and with etoposide and cisplatin for the treatment of advanced disseminated germ cell tumors.165,166 Other recent studies include the use of bleomycin in combination with radiotherapy for the treatment of patients with inoperable head and neck cancer.167 In a somewhat different vein, BLM has also been used to treat resistant warts by intralesional injection.168,169 Following reports that the toxicity of BLM in cultured cells was increased dramatically by electroporation,170 this technique has also been incorporated into clinical trials.171,172 Not unexpect- edly, a significant increase in toxicity has been observed in some cases. The uptake of bleomycins into cells, and the factors that control cellular and nuclear uptake, is an important issue whose resolution could provide direct clinical benefit. One striking feature of the clinical studies with BLM is that all have involved naturally derived BLMs, and most of these have differed only in the nature of the C-terminal substituent. Some clinical studies were conducted with the natural product tallysomycin,173,174 which differs structur- ally from the other BLM group antibiotics (cf. Figures 19.1 and 19.8) and produces a somewhat different spectrum of DNA damage.110,175 The absence of clinical trials of any synthetic or semisynthetic bleomycin no doubt reflects the difficulty of obtaining such materials in quantities sufficient for preclinical and clinical evaluation. Hopefully, that situation may now have changed with the development of more efficient synthetic methodology and the use of solid-phase synthesis. Another key element that needs to be addressed is the development of assays that permit the facile characterization and understanding of the therapeutic importance of key parameters of BLM analog function, including tissue distribution and cell uptake, catabolism by BLM hydrolase and possibly other mechanisms, and the actual cellular locus of action of the drug (i.e., DNA, RNA, or both). ACKNOWLEDGMENT Work in our laboratory on bleomycin is supported by National Institutes of Health research grants CA76297 and CA77284, awarded by the National Cancer Institute. REFERENCES 1. Umezawa, H., Advances in bleomycin studies, in Bleomycin: Chemical, Biochemical and Biological Aspects, Hecht, S. M., Ed., Springer, New York, 1979, pp 24-36. 2. Crooke, S. T., Bleomycin: a brief review, in Carter, S. K., Crooke, S. T., and Umezawa, H., Eds. Bleomycin: Current Status and New Developments, Academic, New York, 1978, pp 1-8. 3. Fujii, A., Biogenetic aspects of bleomycin-phleomycin group antibiotics, in Bleomycin: Chemical, Biochemical and Biological Aspects, Hecht, S. M., Ed., Springer, New York, 1979, pp 75-91. 4. Shen, B. et al., Cloning and characterization of the bleomycin biosynthetic gene cluster from Strep- tomyces verticillus ATCC15003, J. Nat. Prod., 65, 422, 2002. 5. Chien, M., Grollman, A. P., and Horwitz, S. B., Bleomycin-DNA interactions: fluorescence and proton magnetic resonance studies, Biochemistry, 16, 3641, 1977. 6. Boger, D. L. et al., Total synthesis of bleomycin A2 and related agents. 1. Synthesis and DNA binding properties of the extended C-terminus: tripeptide S, tetrapeptide S, pentapeptides S, and related agents, J. Am. Chem. Soc., 116, 5607, 1994. 7. Sugiyama, H. et al., DNA strand scission by bleomycin: catalytic cleavage and strand selectivity, J. Am. Chem. Soc., 108, 3852, 1986. 8. Carter, B. J. et al., A role for the metal binding domain of bleomycin in determining the DNA sequence selectivity of Fe-bleomycin, J. Biol. Chem., 265, 4193, 1990. 9. Guajardo, R. J. et al., [FePMA]n+n = 1,2: good models of Fe-bleomycins and examples of mononuclear non-heme iron complexes with significant O2-activation capabilities, J. Am. Chem. Soc., 115, 7971, 1993. Copyright © 2005 CRC Press, LLC 10. Hecht, S. M., RNA degradation by bleomycin, a naturally occurring bioconjugate, Bioconjugate Chem., 5, 513, 1994. 11. Boger, D. L. et al., A systematic evaluation of the bleomycin A2 L-threonine side chain: its role in preorganization of a compact conformation implicated in sequence-selective DNA cleavage, J. Am. Chem. Soc., 120, 9139, 1998. 12. Boger, D. L et al., Definition of the effect and role of the bleomycin A2 valerate substituents: preorganization of a rigid, compact conformation implicated in sequence-selective DNA cleavage, J. Am. Chem. Soc., 120, 9149, 1998. 13. Tanaka, N., Yamaguchi, H., and Umezawa, H., Mechanism of action of phleomycin. I. Selective inhibition of the DNA synthesis in E. coli and HeLa cells, J. Antibiot., 16A, 86, 1963. 14. Falaschi, A. and Kornberg, A., Phleomycin, an inhibitor of DNA polymerase, Fed. Proc., 23, 940, 1964. 15. Mueller, W. E. and Zahn, R. K., Effect of bleomycin on DNA, RNA, protein, chromatin and on cell transformation by oncogenic RNA viruses, Prog. Biochem. Pharmacol., 11, 28, 1976. 16. Ohno, T. et al., Actions of bleomycin on DNA ligase and polymerases, Prog. Biochem. Pharmacol., 11, 48, 1976. 17. Hecht, S. M., The chemistry of activated bleomycin, Acc. Chem. Res., 19, 383, 1986. 18. Stubbe, J. and Kozarich, J. W., Mechanisms of bleomycin-induced DNA degradation, Chem. Rev., 87, 1107, 1987. 19. Natrajan, A. and Hecht, S. M., Bleomycin: mechanism of polynucleotide recognition and oxidative degradation, in Molecular Aspects of Anticancer Drug–DNA Interactions, Neidle, S. and Waring, M., Eds., MacMillan, London, 1994, pp 197-242. 20. Hecht, S. M., Bleomycin: new perspectives on the mechanism of action, J. Nat. Prod., 63, 158, 2000. 21. Barranco, S. C. and Humphrey, R. M., Effects of bleomycin on survival and cell progression in Chinese hamster cells, Cancer Res., 31, 1218, 1971. 22. Hittelman, W. N. and Rao, P. N., Bleomycin-induced damage in prematurely condensed chromosomes and its relation to cell cycle progression in CHO cells, Cancer Res., 34, 3433, 1974. 23. Barlogie, B. et al., Pulse cytophotomeric analysis of cell cycle perturbation with bleomycin in vitro, Cancer Res., 36, 1182, 1976. 24. Ishida, R. and Takahashi, T., Increased DNA chain breakage by combined action of bleomycin and superoxide radical, Biochem. Biophys. Res. Commun., 66, 1432, 1975. 25. Sausville, E. A., Peisach, J., and Horwitz, S. B., Effect of chelating agents and metal ions on the degradation of DNA by bleomycin, Biochemistry, 17, 2740, 1978. 26. Sausville, E. A. et al., Properties and products of the degradation of DNA by bleomycin and iron II, Biochemistry, 17, 2746, 1978. 27. Ehrenfeld, G. M. et al., Copper I-bleomycin: structurally unique complex that mediates oxidative DNA strand scission, Biochemistry, 24, 81, 1985. 28. Ehrenfeld, G. M et al., Copper dependent cleavage of DNA by bleomycin, Biochemistry, 26, 931, 1987. 29. Ehrenfeld, G. M., Murugesan, N., and Hecht, S. M., Activation of oxygen and mediation of DNA degradation by manganese-bleomycin, Inorg. Chem., 23, 1496, 1984. 30. Burger, R. M. et al., DNA degradation by manganese II-bleomycin plus peroxide, Inorg. Chem., 23, 2215, 1984. 31. Kuwahara, J., Suzuki, T., and Sugiura, Y., Effective DNA cleavage by bleomycin-vanadium IV. complex plus hydrogen peroxide, Biochem. Biophys. Res. Commun, 129, 368, 1985. 32. Claussen, C. A. and Long, E. C., Nucleic acid recognition by metal complexes of bleomycin, Chem. Rev., 99, 2797, 1999. 33. Murugesan, N., Ehrenfeld, G. M., and Hecht, S. M., Oxygen transfer from bleomycin-metal com- plexes, J. Biol. Chem., 257, 8600, 1982. 34. Murugesan, N. and Hecht, S. M., Bleomycin as an oxene transferase. Catalytic oxygen transfer to olefins, J. Am. Chem. Soc., 107, 493, 1985. 35. Burger, R. M., Peisach, J., and Horwitz, S. B., Activated bleomycin. A transient complex of drug, iron, and oxygen that degrades DNA, J. Biol. Chem., 256, 11636, 1981. 36. Natrajan, A. et al., A study of O2 vs H2O2-supported activation of Fe•bleomycin, J. Am. Chem. Soc., 112, 3997, 1990. Copyright © 2005 CRC Press, LLC 37. Sugiyama, H. et al., An efficient, site-specific DNA target for bleomycin, J. Am. Chem. Soc., 107, 7765, 1985. 38. Sugiyama, H. et al., DNA strand scission by bleomycin group antibiotics, J. Nat. Prod., 48, 869, 1985. 39. McGall, G. H. et al., New insight into the mechanism of base propenal formation during bleomycin- mediated DNA degradation, J. Am. Chem. Soc., 114, 4958, 1992. 40. Sugiyama, H. et al., Structure of the alkali-labile product formed during FeII•bleomycin-mediated DNA strand scission, J. Am. Chem. Soc., 107, 4104, 1985. 41. Rabow, L. E. et al., Identification of the alkali-labile product accompanying cytosine release during bleomycin-mediated degradation of dCGCGCG., J. Am. Chem. Soc., 108, 7130, 1986. 42. Sugiyama, H. et al., Chemistry of the alkali-labile lesion formed from ironII-bleomycin and dCGCTT- TAAAGCG., Biochemistry, 27, 58, 1988. 43. Burger, R. M., Peisach, J., and Horwitz, S. B., Stoichiometry of DNA strand scission and aldehyde formation by bleomycin, J. Biol. Chem., 257, 8612, 1982. 44. Wu, J. C., Kozarich, J. W., and Stubbe, J., The mechanism of free base formation from DNA by bleomycin. A proposal based on site specific tritium release from polydA.dU, J. Biol. Chem., 258, 4694, 1983. 45. Wu, J. C., Stubbe, J., and Kozarich, J. W., Mechanism of bleomycin: evidence for 4-ketone formation in polydA-dU associated exclusively with free base release, Biochemistry, 24, 7569, 1985. 46. Rabow, L. E., Stubbe, J., and Kozarich, J. W., Identification and quantitation of lesion accompanying base release in bleomycin-mediated DNA degradation, J. Am. Chem. Soc., 112, 3196, 1990. 47. Rabow, L. E. et al., Identification of the source of oxygen in the alkaline-labile product accompanying cytosine release during bleomycin-mediated oxidative degradation of dCGCGCG., J. Am. Chem. Soc., 112, 3203, 1990. 48. Kane, S. A. and Hecht, S. M., Polynucleotide recognition and degradation by bleomycin, Prog. Nucleic Acid Res. Mol. Biol., 49, 313, 1994. 49. Kozarich, J. W. et al., Sequence-specific isotope effects on the cleavage of DNA by bleomycin, Science, 245, 1396, 1989. 50. Heimbrook, D. C., Mulholland, R. L., and Hecht, S. M, Multiple pathways in the oxidation of cis- stilbene by Fe•bleomycin, J. Am. Chem. Soc., 108, 7839, 1986. 51. Heimbrook, D. C. et al., On the mechanism of oxygenation of cis-stilbene by Fe•bleomycin, Inorg. Chem., 26, 3835, 1987.

52. Van Atta, R. B. et al., Electrochemical activation of oxygenated Fe•bleomycin, J. Am. Chem. Soc., 111, 2722, 1989. 53. Barr, J. R. et al., FeII•BLM-mediated reduction of O2 to H2O: an 17O-NMR study, J. Am. Chem. Soc., 112, 4058, 1990. 54. Boger, D. L. et al., Total synthesis of bleomycin A2 and related agents. 3. Synthesis and comparative evaluation of deglycobleomycin A2, epideglycobleomycin A2, deglycobleomycin A1, and desaceta- mido-, descarboxamido-, desmethyl-, and desimidazolyldeglycobleomycin A2, J. Am. Chem. Soc., 116, 5631, 1994. 55. Povirk, L. F., Han, Y.-H., and Steighner, R. J., Structure of bleomycin-induced DNA double-strand breaks: predominance of blunt ends and single-base 5 extensions, Biochemistry, 28, 5808, 1989. 56. Steighner, R. J. and Povirk, L. F., Bleomycin-induced DNA lesions at mutational hot spots: implica- tions for the mechanism of double-strand cleavage, Proc. Natl. Acad. Sci. U.S.A., 87, 8350, 1990. 57. Absalon, M. J., Kozarich, J. W., and Stubbe, J., Sequence specific double-strand cleavage of DNA by Fe-bleomycin. 1. The detection of sequence-specific double-strand breaks using hairpin oligonucle- otides, Biochemistry, 34, 2065, 1995. 58. Absalon, M. J. et al., Sequence-specific double-strand cleavage of DNA by Fe-bleomycin. 2. Mech- anism and dynamics, Biochemistry, 34, 2076, 1995. 59. Hoehn, S. T. et al., Solution structure of CoIII-bleomycin-OOH bound to a phosphoglycolate lesion containing oligonucleotide: implications for bleomycin-induced double-strand DNA cleavage, Bio- chemistry, 40, 5894, 2001. 60. Keck, M. V., Manderville, R. A., and Hecht, S. M., Chemical and structural characterization of the interaction of bleomycin A2 with dCGCGAATTCGCG.2. Efficient, double-strand DNA cleavage accessible without structural reorganization, J. Am. Chem. Soc., 123, 8690, 2001. Copyright © 2005 CRC Press, LLC 61. Oppenheimer, N. J. et al., Copper I•bleomycin. A structurally unique redox-active complex, J. Biol. Chem., 256, 1514, 1981. 62. Wu, W. et al., Studies of Co•bleomyicn A2 green: its detailed structural characterization by NMR and molecular modeling and its sequence-specific interaction with DNA oligonucleotides, J. Am. Chem. Soc., 118, 1268, 1996. 63. Chang, C.-H. and Meares, C. F., Light-induced nicking of deoxyribonucleic acid by cobaltIII•bleo- mycins, Biochemistry, 21, 6332, 1982. 64. Chang, C.-H. and Meares, C. F., Cobalt-bleomycins and deoxyribonucleic acid: sequence-dependent interactions, action spectrum for nicking, and indifference to oxygen, Biochemistry, 23, 2268, 1984. 65. Wensel, T. G., Chang, C.-H., and Meares, C. F., Diffusion-enhanced lanthanide energy-transfer study of DNA-bound cobaltIII. bleomycins: comparisons of accessibility and electrostatic potential with DNA complexes of ethidium and acridine orange, Biochemistry, 24, 3060, 1985. 66. Subramanian, R. and Meares, C. F., Photosensitization of cobalt bleomycin, J. Am. Chem. Soc., 108, 6427, 1986. 67. Saito, I. et al., Photoinduced DNA strand scission by cobalt bleomycin green complex, J. Am. Chem. Soc., 111, 2307, 1989. 68. McLean, M. J., Dar, A., and Waring, M. J., Differences between sites of binding to DNA and strand cleavage for complexes of bleomycin with iron or cobalt, J. Mol. Recog., 1, 184, 1989. 69. Magliozzo, R. S., Peisach, J., and Ciriolo, M. R., Transfer RNA is cleaved by activated bleomycin, Mol. Pharmacol., 35, 428, 1989. 70. Carter, B. J. et al., Site-specific cleavage of RNA by FeII•bleomycin, Proc. Natl. Acad. Sci. U.S.A., 87, 9373, 1990. 71. Carter, B. J., Reddy, K. S., and Hecht, S. M., Polynucleotide recognition and strand scission by Fe•bleomycin, Tetrahedron, 47, 2463, 1991. 72. Carter, B. J. et al., Metal-catalyzed polynucleotide strand scission, Nucleosides Nucleotides, 10, 215, 1991. 73. Holmes, C. E., Carter, B. J., and Hecht, S. M., Characterization of ironII.-bleomycin-mediated RNA strand scission, Biochemistry, 32, 4293, 1993. 74. Morgan, M. and Hecht, S. M., IronII•bleomycin-mediated degradation of a DNA-RNA heteroduplex, Biochemistry, 33, 10286, 1994. 75. Holmes, C. E. et al., Fe•bleomycin as a probe of RNA conformation, Nucleic Acids Res., 24, 3399, 1996. 76. Kim, S. H. et al., Three-dimensional tertiary structure of yeast phenylalanine transfer RNA, Science, 185, 435, 1974. 77. Robertus, J. D. et al., Structure of yeast phenylalanine tRNA at 3Å resolution, Nature, 250, 546, 1974. 78. Stout, C. D. et al., Crystal and molecular structure of yeast phenylalanyl transfer RNA. Structure determination, difference Fourier refinement, molecular conformation, metal and solvent binding, Acta Crystallogr., 354, 1529, 1978. 79. Moras, D. et al., Crystal structure of yeast tRNAAsp, Nature, 288, 699, 1980. 80. Westhof, E., Dumas, P., and Moras, D., Crystallographic refinement of yeast aspartic acid transfer RNA, J. Mol. Biol., 184, 119, 1985. 81. Giegé, R. et al., Exploring the aminoacylation function of transfer RNA by macromolecular engineer- ing approaches-involvement of conformational features in the charging process of yeast transfer RNAAsp, Biochimie, 72, 453, 1990. 82. Perret, V. et al., Conformation in solution of yeast transfer RNAAsp transcripts deprived of modified nucleotides, Biochimie, 72, 735, 1990. 83. Pütz, J. et al., Identity elements for specific aminoacylation of yeast transfer RNAAsp by cognate aspartyl-transfer RNA synthetase, Science, 252, 1696, 1991. 84. Perret, V. et al., Effect of conformational features on the aminoacylation of transfer RNAs and consequences on the permutation of transfer RNA specificities, J. Mol. Biol., 226, 323, 1992. 85. Rudinger, J. et al., Determinant nucleotides of yeast tRNAAsp interact directly with aspartyl-tRNA synthetase, Proc. Natl. Acad. Sci. U.S.A., 89, 5882, 1992. 86. Abraham, A. T. et al., RNA cleavage and inhibition of protein synthesis by bleomycin, Chem. Biol., 10, 45, 2003. Copyright © 2005 CRC Press, LLC 87. Dix, D. J. et al., The influence of the base-paired flanking region on structure and function of the ferritin mRNA iron regulatory element, J. Mol. Biol., 231, 230, 1993. 88. Snow, A., Ph.D. Thesis, University of Virginia, 1995. 89. Duff, R. J. et al., Evidence for C-1’ hydrogen abstraction from modified oligonucleotides by Fe•ble- omycin, J. Am. Chem. Soc., 115, 3350, 1993. 90. Holmes, C. E. et al., On the chemistry of RNA degradation by FeII•BLM, Bioorg. Med. Chem., 5, 1235, 1997. 91. Pieters, J. M. L. et al., Hairpin structures in DNA containing arabinofuranosylcytosine. A combination of nuclear magnetic resonance and molecular dynamics, Biochemistry, 29, 788, 1990. 92. Keck, M. V. and Hecht, S. M., Sequence-specific hydrolysis of yeast tRNAPhe mediated by metal free bleomycin, Biochemistry, 34, 12029, 1995. 93. Ardelt, W., Mikulski, S. M., and Shogen, K., Amino-acid-sequence of an antitumor protein from rana- pipiens oocytes and early embryos-homology to pancreatic ribonucleases, J. Biol. Chem., 266, 245, 1991. 94. Darzynkiewicz, Z. et al., Cystostatic and cytotoxic effects of pannon P-30. A novel anticancer agent, Cell Tissue Kinet., 21, 169, 1988. 95. Mikulski, S. M. et al., Tamoxifen and trifluoroperazine (stelazine) potentiate cystostatic cytotoxic effects of P-30 protein. A novel protein possessing antitumor activity, Cell Tissue Kinet., 23, 237, 1990. 96. Mikulski, S. M. et al., Striking increase of survival of mice bearing M109 madison carcinoma treated with a novel protein from amphibian embryos, J. Natl. Cancer Inst., 82, 151, 1990. 97. Mikulski, S. M. et al., Phase-1 human clinical trial of onconase R (P-30 protein) administered intravenously on a weekly schedule in cancer patients with solid tumors, Int. J. Oncol., 3, 57, 1993. 98. Lin, J.-J. et al., Characterization of the mechanism of cellular and cell free protein synthesis inhibition by an anti-tumor ribonuclease, Biochem. Biophys. Res. Commun., 204, 156, 1994. 99. Saxena, S. K. et al., Entry into cells and selective degradation of tRNAs by a cytotoxic member of the RNase A family, J. Biol. Chem., 277, 15142, 2002. 100. Berry, D. E., Chang, L.-H., and Hecht, S. M., DNA damage and growth inhibition in cultured human cells by bleomycin congeners, Biochemistry, 24, 3207, 1985. 101. Hecht, S. M., Bleomycin group antitumor agents, in Cancer Chemotherapeutic Agents, Foye, W. O., Ed., American Chemical Society, Washington, DC, 1995, pp 369-388. 102. Takita, T. et al., Chemical cleavage of bleomycin to bleomycinic acid and synthesis of new bleomycins, J. Antibiot., 26, 252, 1973. 103. Tanaka, W. and Takita, T., Pepleomycin, the second generation bleomycin chemically derived from bleomycin A2, Heterocycles, 13, 469, 1979. 104. Umezawa, H., Recent chemical studies of bioactive microbial products: genetics, active structures, development of effective agents with potential usefulness, Heterocycles, 13, 23, 1979. 105. Umezawa, H. et al., Preparation of bleomycinic acid: hydrolysis of bleomycin B2 by a Fusarium acylagmatine amidohydrolase, J. Antibiot., 26, 117, 1973. 106. Tanaka, W., Development of new bleomycins with potential clinical utility, J. Antibiot., 30, S-41, 1977. 107. Saito, S. et al., A new synthesis of deglyco-bleomycin A2 aiming at the total synthesis of bleomycin, Tetrahedron Lett., 23, 529, 1982. 108. Aoyagi, Y. et al., Deglycobleomycin: total synthesis and oxygen transfer properties of an active bleomycin analog, J. Am. Chem. Soc., 104, 5237, 1982. 109. Oppenheimer, N. J. et al., Deglyco-bleomycin. Degradation of DNA and formation of a structurally unique FeII•CO complex, J. Biol. Chem., 257, 1606, 1982. 110. Kross, J. et al., Specificity of deoxyribonucleic acid cleavage by bleomycin, phleomycin and tallyso- mycin, Biochemistry, 21, 4310, 1982. 111. Takahashi, K. et al., Liblomycin, a new analog of bleomycin, Cancer Treat. Rev., 14, 169, 1987. 112. Takita, T. et al., Total synthesis of bleomycin A2, Tetrahedron Lett., 23, 521, 1982. 113. Aoyagi, Y. et al., Total synthesis of bleomycin, J. Am. Chem. Soc., 104, 5537, 1982. 114. Boger, D. L. and Honda, T., Total synthesis of bleomycin A2 and related agents. 4. Synthesis of the disaccharide subunit 2-O-3-O-carbamoyl-α-D-mannopyranosyl-L-gulopyranose and completion of the total synthesis of bleomycin A2, J. Am. Chem. Soc., 116, 5647, 1994. 115. Katano, K. et al., Total synthesis of bleomycin group antibiotics. Total synthesis of bleomycin demethyl A2, bleomycin A2, and decarbamoyl bleomycin demethyl A2, J. Am. Chem. Soc., 120, 11285, 1998. Copyright © 2005 CRC Press, LLC 116. Boger, D. L., Teramoto, S., and Zhou, J., Key synthetic analogs of bleomycin A2 that directly address the effect and role of the disaccharide: demannosylbleomycin A2 and α-D-mannopyranosyldeglycob- leomycin A2, J. Am. Chem. Soc., 117, 7344, 1995. 117. Zou, Y. et al., Total synthesis of deamido bleomycin A2, the major catabolite of the antitumor agent bleomycin, J. Am. Chem. Soc., 124, 9476, 2002. 118. Leitheiser, C. J. et al., Solid phase synthesis of bleomycin group antibiotics. Elaboration of degly- cobleomycin A5, Org. Lett., 2, 3397, 2000. 119. Aoyagi, Y. et al., Synthesis of pyrimidoblamic acid and epipyrimidoblamic acid, J. Org. Chem., 55, 6291, 1990. 120. Abraham, A. T., Zhou, X., and Hecht, S. M., DNA cleavage by FeII•bleomycin conjugated to a solid support, J. Am. Chem. Soc., 121, 1982, 1999. 121. Smith, K. L. et al., Deglycobleomycin: solid-phase synthesis and DNA cleavage by the resin-bound ligand, Org. Lett., 4, 1079, 2002. 122. Gordeev, M. F. et al., Combinatorial chemistry of natural products: solid phase synthesis of D-and L-cycloserine derivatives, Tetrahedron, 54, 15879, 1998. 123. Tao, Z.-F. et al., Solid phase synthesis of deglycobleomycins: a C-terminal tetramine linker that permits direct evaluation of resin-bound bleomycins, Bioconjugate Chem., 13, 426, 2002. 124. Owa, T., Otsuka, M., and Ohno, M., An efficient synthesis of erythro-β-hydroxy-L-histidine, the pivotal amino acid of bleomycin-FeII•O2 complex, Chem. Lett., 1873, 1988. 125. Boger, D. L. and Menezes, R. F., Synthesis of tri- and tetrapeptide S: the extended C-terminus of bleomycin A2, J. Org. Chem., 57, 4331, 1992. 126. Thomas, C. J. et al., Solid phase synthesis of bleomycin A5 and three monosaccharide analogs. Exploring the role of the carbohydrate moiety in RNA cleavage, J. Am. Chem. Soc., 124, 12926, 2002. 127. Hamamichi, N. and Hecht, S. M., Determination of the absolute configuration of the thiazolinylthiazole moiety of phleomycin, J. Am. Chem. Soc., 115, 12605, 1993. 128. Sznaidman, M. L. and Hecht, S. M., Studies on the total synthesis of tallysomycin. Synthesis of the threonylbithiazole moiety containing a structurally unique glycosylcarbinolamide, Org. Lett., 3, 2811, 2001. 129. Morii, T. et al., Phototransformed bleomycin antibiotics. Structure and DNA cleavage activity, J. Am. Chem. Soc., 108, 7089, 1986. 130. Morii, T. et al., New lumibleomycin-containing thiazolylisothiazole ring, J. Am. Chem. Soc., 109, 938, 1987. 131. Zuber, G., Quada Jr., J. C., and Hecht, S. M., Sequence selective cleavage of a DNA octanucleotide by chlorinated bithiazoles and bleomycins, J. Am. Chem. Soc., 120, 9368, 1998. 132. Hamamichi, N., Natrajan, A., and Hecht, S. M., On the role of individual bleomycin thiazoles in oxygen activation and DNA cleavage, J. Am. Chem. Soc., 114, 6278, 1992. 133. Otsuka, M. et al., Man-designed bleomycin with altered sequence specificity in DNA cleavage, J. Am. Chem. Soc., 112, 838, 1990. 134. Owa, T. et al., Man-designed bleomycins: significance of the binding sites as enzyme models and of the stereochemistry of the linker moiety, Tetrahedron, 48,

1193, 1992. 135. Kuwahara, J. and Sugiura, Y., Sequence-specific recognition and cleavage of DNA by metallobleo- mycin: minor groove binding and possible interaction mode, Proc. Natl. Acad. Sci. U.S.A., 85, 2459, 1988. 136. Quada Jr., J. C., Zuber, G. F., and Hecht, S. M., Interaction of bleomycin with DNA, Pure Appl. Chem., 70, 307, 1998. 137. Quada Jr., J. C., Boturyn, D., and Hecht, S. M., Photoactivated DNA cleavage by compounds structurally related to the bithiazole moiety of bleomycin, Bioorg. Med. Chem., 9, 2303, 2001. 138. Manderville, R. A., Ellena, J. F., and Hecht, S. M., Interaction of ZnII•bleomycin with dCGCTAGCG2. A binding model based on NMR experiments and restrained molecular dynamics calculations, J. Am. Chem. Soc., 117, 7891, 1995. 139. Sucheck, S. J., Ellena, J. F., and Hecht, S. M., Characterization of ZnII•deglycobleomycin A2 and interaction with dCGCTAGCG2. Direct evidence for minor groove binding of the bithiazole moiety, J. Am. Chem. Soc., 120, 7450, 1998. 140. Williams, L. D. and Goldberg, I. H., Selective strand scission by intercalating drugs at DNA bulges, Biochemistry, 27, 3004, 1988. Copyright © 2005 CRC Press, LLC 141. Gold, B. et al., Alteration of bleomycin cleavage specificity in a platinated DNA oligomer of defined structure, J. Am. Chem. Soc., 110, 2347, 1988. 142. Kane, S. A. et al., Specific cleavage of a DNA triple helix by FeII•bleomycin, Biochemistry, 34, 16715, 1995. 143. Kross, J. et al., Structural basis for the DNA affinity of bleomycins, Biochemistry, 21, 3711, 1982. 144. Lui, S. M. et al., Structural characterization of Co•bleomycin A2 brown: free and bound to dCCAG- GCCTGG., J. Am. Chem. Soc., 199, 9603, 1997. 145. Wu, W. et al., NMR studies of Co•deglycobleomycin A2 green and its complex with dCCAGGC- CTGG., J. Am. Chem. Soc., 120, 2239, 1998. 146. Rishel, M. J. et al., Conformationally constrained analogs of bleomycin A5, J. Am. Chem. Soc., 125, 10194, 2003. 147. Cagir, A. et al., Solid-phase synthesis and biochemical evaluation of conformationally constrained analogs of deglycobleomycin A5, Bioorg. Med. Chem., 11, 5179, 2003. 148. Boger, D. L., Ramsey, T. M., and Cai, H., Synthesis and evaluation of potential Nπ and Nσ metal chelation sites with the beta-hydroxy-L-histidine subunit of bleomyin A2: functional characterization of imidazole Nπ metal complexation, Bioorg. Med. Chem., 4, 195, 1996. 149. Boger, D. L., Teramoto, S., and Cai, H., Synthesis and evaluation of deglycobleomycin A2 analogs containing a tertiary N-methyl amide and simple ester replacement for the L-histidine secondary amide: direct functional characterization of the requirement for secondary amide metal complexation, Bioorg. Med. Chem., 4, 179, 1996. 150. Leitheiser, C. J. et al., Solid phase synthesis of bleomycin group antibiotics. Construction of a 108- member deglycobleomycin library, J. Am. Chem. Soc., 125, 8218, 2003. 151. Hashimoto, S., Wang, B., and Hecht, S. M., Kinetics of DNA cleavage by FeII•bleomycins, J. Am. Chem. Soc., 123, 7437, 2001. 152. Bleomycin: Current Status and New Developments, Carter, S. K., Crooke, S. T., and Umezawa, H., Eds., Academic Press, Orlando, FL, 1978. 153. Bleomycin Chemotherapy, Sikic, B. I., Rozencweig, M., and Carter, S. K., Eds., Academic Press, Orlando, FL, 1985. 154. Bennett, J. M. and Reich, S. D., Bleomycin, Ann. Intern. Med., 90, 945, 1979. 155. Einhorn, L. H. and Donohue, J., Cis-diamminedichloroplatinum, vinblastine, and bleomycin combi- nation chemotherapy in disseminated testicular cancer, Ann. Intern. Med., 87, 293, 1977. 156. Carlson, R. W. et al., Combination cisplatin, vinblastine, and bleomycin chemotherapy (PVB) for malignant germ-cell tumors of the ovary, J. Clin. Oncol., 1, 645, 1983. 157. Comis, R. L. et al., Role of single-breath carbon monoxide-diffusing capacity in monitoring the pulmonary effects of bleomycin in germ cell tumor patients, Cancer Res., 39, 5076, 1979. 158. Muggia, F. M., Louie, A. C., and Sikic, B. I., Pulmonary toxicity of antitumor agents, Cancer Treat. Rev., 10, 221, 1982. 159. Levi, J. A. et al., The importance of bleomycin in combination chemotherapy for good-prognosis germ cell carcinoma, J. Clin. Oncol., 11, 1300, 1993. 160. Kullendorff, C. M., Efficacy of bleomycin treatment for symptomatic hemangiomas in children, Ped. Surg. Int., 12, 526, 1997. 161. Mottolese, C. et al., Intracystic chemotherapy with bleomycin in the treatment of craniophrryngiomas, Childs Nervous System, 17, 724, 2001. 162. Alen, J. F. et al., Intratumoural bleomycin as a treatment for recurrent cystic craniopharyngioma. Case report and review of the literature, Neurocirugia, 13, 479, 2002. 163. Jiang, R., Liu, Z., and Zhu, C., Preliminary exploration of the clinical effect of bleomycin on craniopharyngiomas, Stereotactic Funct. Neurosurg., 78, 84, 2002. 164. Tokuhashi, Y. et al., A randomized trial of cisplatin, vinblastine, and bleomycin versus cyclophospha- mide, aclacinomycin, and cisplatin in eipthelial ovarian cancer, Oncology, 54, 281, 1997. 165. Nichols, C. R et al., Randomized comparison of cisplatin and etoposide and either bleomycin or ifosfamide in treatment of advanced disseminated germ cell tumors: an Eastern Cooperative Oncology Group, Southwest Oncology Group, and Cancer and Leukemia Group B study, J. Clin. Oncol., 16, 1287, 1998. 166. Christian, J. A. et al., Intensive induction chemotherapy with CBOP/BEP in patients with poor prognosis germ cell tumors, J. Clin. Oncol., 21, 871, 2003. Copyright © 2005 CRC Press, LLC 167. Minatel, E. et al., Combined radiotherapy and bleomycin in patients with inoperable head and neck cancer with unfavourable prognostic factors and severe symptoms, Oral Oncol., 34, 119, 1998. 168. Munn, S. E. et al., A new method of intralesional bleomycin therapy in the treatment of recalcitrant warts, Br. J. Dermatol., 135, 969, 1996. 169. Pollock, B. and Sheehan-Dare, R., Pulsed dye laser and intralesional bleomycin for treatment of resistant viol hand warts, Lasers Surg. Med., 30, 135, 2002. 170. Poddevin, B. et al., Very high cytotoxicity of bleomycin introduced into the cytosol of cells in culture, Biochem. Pharmacol., 42, 567, 1991. 171. Horiuchi, A. et al., Enhancement of antitumor effect of bleomycin by low-voltage in vivo electropo- ration: a study of human uterine leiomyosarcomas in nude mice, Int. J. Cancer, 88, 640, 2000. 172. Nanda, G. S. et al., Electroporation enhances therapeutic efficacy of anticancer drugs: treatment of human pancreatic tumor in animal models, Anticancer Res., 18, 1361, 1998. 173. Nicaise, C. et al., Phase II study of tallysomycin S10b in patients with advanced colorectal cancer, Cancer Chemother. Pharmacol., 26, 221, 1990. 174. Nicaise, C. et al., Phase II study of tallysomyin S10b in patients with advanced head and neck cancer, Invest. New Drugs, 8, 325, 1990. 175. Mirabelli, C. K., Huang, C. H., and Crooke, S. T., Comparison of DNA damage and single- and double-strand breakage activities on PM-2 DNA by tallysomycin and bleomycin analogs, Cancer Res., 40, 4173, 1980. Copyright © 2005 CRC Press, LLC 20 Biochemical and Biological Evaluation of (+)-CC-1065 Analogs and Conjugates with Polyamides Rohtash Kumar and J. William Lown CONTENTS I. Introduction II. Drug–DNA Interactions III. Modification in the Pharmacophore Unit A. CPI Analogs B. Enantiomers of CPI Analogs C. CC-1065 Analogs IV. Interaction of CC-1065 with Chromatin V. Polyamide Conjugates and Related Sequence-Directed Structures VI. Bisalkylators (Dimers) VII. Structural Characterization of Drug DNA Complexes VIII. Cellular and Pharmacological Studies A. In Vitro Studies B. In Vivo Studies IX. Clinical Studies X. Conclusions and Prospects References I. INTRODUCTION CC-1065 (1) is an extremely potent antitumor antibiotic isolated from Streptomyces zelensis by the Upjohn company.1,2 It is one of the most potent anticancer compounds having a wide spectrum of activities against tumor cells in vitro and in vivo as well as against microbial organisms.3,4 Struc- turally, the CC-1065 molecule consists of three repeating pyrroloindole subunits, one of which contains the DNA reactive cyclopropylpyrroloindole (CPI) moiety (2) while the other two subunits mediate noncovalent binding interactions with DNA. CC-1065 alkylates the N3 position of adenine with its reactive cyclopropyl group in the minor groove of double helical DNA in a sequence specific manner with a consensus sequence of 5-(T/A)(T/A) A* (where the asterisk indicates the site of alkylation). The preferred site of alkylation by CC-1065 was determined to be the 5-TTA* sequence.4-10 Despite the highly potent activity of CC-1065 in vitro3,11 and in vivo,2 clinical devel- opment was precluded by an unusual hepatotoxicity, which led to delayed death in mice at therapeutic doses.12 Subsequently, an aggressive and highly successful synthetic program at the Copyright © 2005 CRC Press, LLC STRUCTURES 1–8 Upjohn company provided a series of three new compounds that appeared to be more suitable for clinical development. These agents, adozelesin (3), U-71184 (4), and U-80224 (5), have been shown to have improved antitumor efficacy over CC-1065 and do not show delayed lethality.13,14 Duocarmycin (6) and pyrindamycin (7) are structurally related natural products, with (6) possessing the CC-1065-DNA reactive cyclopropane ring.15,16 The structural element of CC-1065 that produces delayed toxicity has been elucidated, but its biochemical mechanism is still unknown. A number of CC-1065 analogs, which maintain potent antitumor activity without exhibiting delayed death, were subsequently synthesized, and the unique feature of all these synthetic analogs is the cyclopropyl indole moiety (CPI; 2). Bizelesin (8) is a bifunctional analog, shows good antitumor efficacy both in vitro and in vivo, and is generally 20–30-fold more potent than adozelesin (a monofunctional analog) when tested against human carcinoma cells.17 Adozelesin 3 is currently in phase II clinical trials, and bizelesin (8) is currently in phase I clinical trials. Certain synthetic analogs of CC-1065 and duocarmycin, including dimeric structures, show promise and are also currently undergoing clinical trials,17 indicating that synthetic efforts in this area can be useful. A new class of agents bearing the 1,2,9,9a-tetrahydrocyclopropa[c] benz [e] indole-4-one (9) pharmacophore (CBI) has been synthesized.18 In this chapter, in addition to delineating the cytotoxic and antitumor activities of CC-1065, we also examine the most recent analogs of CPI and CBI and their cytotoxicities. Certain CPI–polyamide conjugates have shown good cytotoxic potency19,20; therefore, we also describe various new CPI and Copyright © 2005 CRC Press, LLC STRUCTURES 9–10 CBI polyamide conjugates. Since CPI–CPI dimers were very active against tumor cells, a large number of CBI–CBI dimers with varying lengths of the aliphatic chain linkers and with specific polyamides were synthesized in our group, and their cytotoxic properties against tumor cells are summarized below. II. DRUG–DNA INTERACTIONS The mechanisms of cytotoxic action of CC-1065 and its analogs have been shown to involve binding to the minor groove of DNA and alkylation at the N-3 position of adenine with concomitant opening of the reactive CPI unit (Figure 20.1). Thermal treatment of this covalently modified DNA results in depurination of the alkylated bases followed by strand cleavage, producing a 5-phosphate and 3-modified deoxyribose.4,5,21 FIGURE 20.1 The mechanisms of cytotoxic action of CC-1065 and its analogs. Copyright © 2005 CRC Press, LLC Both enantiomers primarily alkylate N-3 of adenine, but the orientation of the drug is different. For the natural (+) enantiomer, the B and C subunits lie on the 5 site of alkylation and on the 3 site for the (–) enantiomer, as shown by DNA footprinting. They also show different sequence selectivity, although both prefer A•T rich sites. However, (+) and (–) analogs that contain only the A subunit do not show different sequence selectivity, demonstrating the importance of the B and C subunits in determining the site of alkylation. The unnatural enantiomer appears to be less reactive and provides less stabilization of the DNA helix as judged by the change in melting temperature of the alkylated DNA. A slight increase in the amount of alkylation on guanine was observed for (–)-CC-1065. Only the natural enantiomer appears to cause delayed death.7,12,22,23 Recent studies have shown that many analogs of CC-1065 reversibly alkylate DNA. However, CC-1065 itself appears to alkylate DNA irreversibly under the conditions tested. This difference is attributed to the strong noncovalent binding of CC-1065, which either prevents the reverse reaction after alkylation has occurred or keeps the drug bound so that it realkylates the same site. Boger and Johnson have suggested that irreversibility of the alkylation may be related to the delayed toxicity of some of these compounds.24 Several studies have shown that the unnatural enantiomers or epimers of this class of compounds also alkylate DNA and possess cytotoxic activity. Terashima’s group synthesized (+)-duocarmycin A and its three possible stereoisomers and compared their in vitro cytotoxicities against P388 murine leukemia, demonstrating that both compounds bearing natural configurations at the cyclo- propane moiety were more potent than those having unnatural configurations. This result indicated the importance of chirality in the cyclopropyl ring in

duocarmycin A.25,26 Boger and Johnson24 synthesized both enantiomers of CC-1065 and duocarmycin A, duocar- mycin SA, and epimers of duocarmycin A and compared cytotoxic potencies against L1210 and relative DNA alkylation efficiencies. In the case of duocarmycin A, both (–) forms, which had opposite chirality in the cyclopropyl ring to that of natural (+)-duocarmycin A, had almost no effect on the DNA alkylation and cytotoxic potency. In the case of duocarmycin SA, the natural enantiomer was found to alkylate DNA at concentrations approximately 10 times that required for the nonnatural products.27 III. MODIFICATION IN THE PHARMACOPHORE UNIT A. CPI ANALOGS Compounds derived from the pharmacophore of CC-1065 but bearing different electron-withdraw- ing constituents in the benzene ring and different leaving groups on the C3 methylene have been synthesized and evaluated.28 These synthetic agents were less reactive at the common preferred sites than CC-1065, and higher concentrations are required to produce detectable cleavage. On the basis of alkylation intensities and densitometric data, the overall binding intensity is in the order of CC-1065 > 12 > 13 > 11. From those experiments, it appears that there are two factors that are critical for the formation of the cyclopropane ring by cyclization: The first is the electron density of the benzoyl moiety where an electron-withdrawing group retards the cyclization, whereas an electron-donating group promotes the cyclization, and the second factor is the leaving ability of the group on the C3 methylene carbon. A better leaving group will favor the cyclization. The evidence from gel electrophoresis showed that when supercoiled PM2 DNA is incubated with 11, 12, 13, and CC-1065 at 80 µM for 20h, the CCC DNA form was completely converted to OC DNA. This indicated that, like CC-1065, 11, 12, and 13 also exert their cytotoxicity through covalent binding to DNA. However, additional binding at the G residues by these synthetic analogs may represent an alternative or additional mechanism for this recognition of DNA.29,30 In this connection, the DNA sequence specificity of the pyrrolo [1,4] benzodiazepine antitumor antibiotics, which bind through N2 of guanine, has been supported by theoretical calculations.31 Copyright © 2005 CRC Press, LLC O O O O O O S S S O O O Br O2N HO N HO N HO N 11 O 12 O 13 O S S S O O O O O O STRUCTURES 11–13 A series of N-substituted CPI-analogs (14–40) has been synthesized, and certain structural features have been identified that affect reactivity toward nucleophiles and those influencing non- covalent interactions with DNA. Those features have then been related to biological activity and potency.13 Table 20.1 lists the biological activities of CC-1065 and the racemic CPI analogs. Solvolytic reactivity is indicated by the pseudo-first-order rate constant for acid-catalyzed opening of the cyclopropyl ring in methanolic buffer. When CC-1065 is incubated with calf thymus DNA for 24 h at 25˚C (Table 20.1), the observed induced circular dichroism (ICD) is entirely caused by the binding. Although the racemic analogs display no circular dichroism on their own, in the presence of DNA some of them show significant ICD. From the ICD it appears that two structural features of the CPI moiety strongly influence DNA binding, the size of the ring attached to the pyrrolidine nitrogen of CPI, and the length of the aromatic amide chain. An ICD is almost exclusively associated with structures having a five-membered heteroaryl ring attached to CPI. Methyl substituents on the five-membered ring significantly diminish the ICD, as is also the case when the nitrogen atom is replaced with oxygen or sulfur. Analogs with six-membered rings attached to CPI fail to show an appreciable ICD. Table 20.1 shows that cytotoxic potency is qualitatively correlated to DNA binding. The DNA alkylation selectivity of the analogs of (+) CC-1065 has been studied.32,33 The sites of alkylation of double-stranded DNA were examined for simple derivatives of 7-methyl-1, 2,8,8a- tetrahydrocycloprop[1,2-C] pyrrolo [2,2-e] indole-4 (5H)-one (CPI; 2). (+) CC-1065 (1) and 1,2,7,7a- tetrahydrocycloprop [1,2-C] indole-4-one (CI; 41) left-hand subunits. It was found that the CI subunit of the agents is a much more reactive alkylating agent than the natural CPI alkylation subunit of CC-1065. The simple derivatives 42, 43, 44, and 45 of CI were found to alkylate double- stranded DNA under milder conditions than did simple derivatives of CPI, and the marked simi- larities in CI and CPI DNA alkylation profiles indicate that CI represents the minimum pharma- cophore of CPI. By use of the Fe (III)-CDPI3 EDTA (FeIII-42, -43), they demonstrated that the noncovalent binding is a prerequisite for DNA alkylations. The relative cleavage efficiency observed with Fe(III)-46 was found to parallel the relative alkylation efficiency of (+)- or net (–)-CC-1065. B. ENANTIOMERS OF CPI ANALOGS Boger’s group34 made a detailed evaluation of the DNA-alkylation selectivity of (+)-CC-1065 and its enantiomer (–) CC-1065 and a series of analogs (47–51) possessing the CPI alkylation subunit. The natural enantiomers bind in the minor groove in the 3 → 5 direction starting from the adenine N3 alkylation site across a 2-base (N-BocCPI 47; i.e., 5-AA), 3.5-base (CPI-CDPI1 48/CPI-PDEI1 51; i.e., 5-AAA), 5-base (CC-1065/CPI-CDPI2 49; i.e., 5-AAAAA), or 6.5-base (CPI-CDPI3 50; i.e., 5- AAAAAA) AT-rich site. In contrast, the unnatural enantiomers bind in the reverse 5 → 3 direction in the minor groove, and the binding site starts at the first 5 base preceding the adenine N3 alkylation site and extends across the alkylation site to the adjacent 3 base covering an AT-rich site of 2 bases (NBoc-CPI 47; i.e., 5AA), 5 bases (CC-1065/CPI-CDPI2; i.e., 5-AAAAA), or 6.5 bases (CPI-CDPI3; i.e., 5-AAAAAA). Three-dimensional models of natural and unnatural enantiomer alkylations clearly Copyright © 2005 CRC Press, LLC TABLE 20.1 Physicochemical and Biological Properties of Racemic CC-1065 and CPI Analogs P388 In Vivoe No. R Formula R a K × 105a ICD × 103c ID d ODf 2 50 %ILSg N R HN O 14 (PDE-I)2 C37H33N7O8 0.39 280 0.07 0.1 62 15 SO2Me C13H14N2O3S 0.82 0.3 0 100 25 60 16 SO2Ph C18H16N2O3S 0.58 900 17 H C12H12N2O 0.87 Stable 0 900 10(HDT) NA 18 CO2C(CH3)3 C17H20N2O3 0.35 0.5 0 300 19 COCH3 C14H14N2O2 0.90 0 20 3 45 20 CO(CH2)8CH3 C22H30N2O2 0.07 0 100 40(HDT) NA 21 COPh C19H16N2O2 0.63 0.9 <1 3 12 71 22 CO-4-(NHCOPh)C6H4 C26H21N3O3 0.42 11 0.5 0.5 83 23 CO-2-pyridyl C18H15N3O2 0.76 1.2 0 10 15 69 24 CO-2-naphthyl C23H18N2O2 0.46 0 7 6 73 25 CO-2-quinolyl C22H17N3O2 0.50 1.2 1 10 6 142 26 CO-6-(NHCO(CH2)4CH3)-2-quinolyl C28H28N4O3 0.15 h 100 21 127 27 CO-2-pyrrolyl C17H15N3O2 0.70 10 2 1 71 28 CO-2-indolyl C21H17N3O2 0.45 1.0 36 0.2 0.4 90 29 CO-1-Me-2-indolyl C22H19N3O2 0.43 1.0 3 10 8 44 30 CO-2-benzofuranyl C21H16N2O3 0.47 1.1 6 0.04 0.25 110 31 CO-2-benzothiophenyl C21H16N2O2S 0.41 1.1 7 0.3 0.80 100 32 CO-3-Me-2-indenyl C23H20N2O2 0.39 4 30 17 NA (HDT) N HN N R H O O 33 5-OMe C22H19N3O3 0.47 1.0 22 0.01 0.05 95 34 6-OH, 7-OMe C22H19N3O4 0.63 1.0 37 0.1 0.05 55 35 5-NHCONH2 C22H19N5O3 0.89 1.0 47 0.6 0.06 65 36 5-NHCOPh C28H22N4O3 0.43 36 0.07 0.05 82 37 5-NHCO-2-indolyl C30H23N5O3 0.30 i 70 0.01 0.05 159(4) 38 5-NHCO-2-benzofuranyl C30H22N4O4 0.30 1.0 50 0.02 0.05 170(1) 39 5-NHCO-5-(NHCONH2)-2-indolyl C31H25N7O4 75 0.3 0.1 121(2) 40 5-NHCO-5-(NHCOPh)-2-indolyl C37H28N6O4 0.30 33 0.01 0.1 77 Note: HDT, highest dose tested, toxicity not reached; NA, no activity. a Chromatographic measure of lipophilicity determined by liquid–liquid chromatography on C18-bonded silica gel. b Pseudo-first-order rate constant, in sec1, for solvolytic ring opening at pH3, followed spectrophotometrically. c Molar ellipticity, in the presence of calf thymus DNA. d ID50, the nanomolar concentration of drug required to inhibit, by 50%, the growth of murine L1210 leukemia cells in a 3-day assay. e Drug given by interperitoneal injection to mice implanted interperitoneally with 106 P388 leukemia cells. f Optimum dose in milligrams per kilogram per day, given on days 1, 5, and 9. g Percentage increase in life span of treated animals over that of control mice bearing tumor, when treated at the optimal dose. Numbers in parentheses are the numbers of mice that survived for 30 days. h Extreme insolubility of the drug gave anomalous and irreproducible results. i The drug was not soluble in this medium in the concentration range used. Copyright © 2005 CRC Press, LLC OH O N HO N O O N 42 R = OtBu R 43 R = CH O O 41 CI 3 O N O NH NH NH2 2 HN N N n 3 O N O O O N O N H H 46 CDPI3-EDTA 44 n = 1; [(+)CI-CDPI1] 45 n = 2; [(+)CI-CDPI2] STRUCTURES 41–46 illustrate the offset binding sites. A simple model for the CC-1065 DNA alkylation reaction that accommodates the behavior of both enantiomers is provided, in which the sequence selectivity is derived from the noncovalent binding selectivity of the agents in the narrower sterically more acces- sible AT-rich minor groove. The inherent steric accessibility to the adenine N3-alkylation site accom- panies deep penetration of the agent into the minor groove within the AT-rich site (Figure 20.2). In the DNA binding and cytotoxic activity studies of two enantiomers of bis-indole analogs of CC-1065, the isomer with the same stereochemical configuration as natural (+)-CC-1065 was a potent cytotoxic agent, but its enantiomer was essentially inactive.35 Both enantiomers showed significant binding to DNA, but the biologically less active isomer showed less overall binding. In all cases the agents preferred AT-rich DNA, and all bound to similar regions in DNA, as evidenced by the positions of drug-initiated thermal breaks in single end-labeled fragments of Phi X174 RF DNA. The overall similarity in site specificity for binding of the structurally diverse agents indicates that much of the specificity observed in the binding of the agent to DNA lies in the DNA itself. C. CC-1065 ANALOGS A simple method to detect CPI binding sites on double-stranded DNA has been devised.8 The technique uses a modified form of bacteriophage T7 polymerase to synthesize a radiolabeled nascent strand from double-stranded DNA that has been allowed to react in vitro with the CC-1065 analog U-73,975 (adozelesin) 3. The reaction products were electrophoresed on sequencing gels containing 8 M urea and visualized by autoradiography. The transit of this DNA polymerase is inhibited at the sites where CPIs are bound to the template strand. The positions of polymerase inhibition can be determined by comparison of CPI-treated and untreated DNA reactions. This modified dideox- ynucleotide technique has been used to establish the sequence selectivity of U-73,975. Bizelesin (U-77,779) 8 is a symmetrical dimer of the spirocyclopropyl alkylating subunit of (+)- CC-1065 in which the linker consists of two indole subunits separated by a ureido group. A comparison HN NH HN 2 HN O N N n NH2 O O N N O O N N O H O N O H O 48 n = 1; (+)-CPI-CDPI1 O 46 (-)-N-Boc CPI 49 n = 2; (+)-CPI-CDPI2 50 n = 3; (+)-CPI-CDPI 51 (+)-CPI-PDEI 3 1 STRUCTURES 47–51 Copyright © 2005 CRC Press, LLC FIGURE 20.2 (See color insert) Comparison stick and space-filling models of the (+)- duocarmycin SA (left) and ent-()-duocarmycin SA (right) alkylation at the same site within w794DNA: duplex 5-d(GACTAATTTTT). The natural enantiomer binding extends in the 3 → 5 direction from the adenine N3 alkylation site 5-CTAA. The unnatural enantiomer binding extends in the 5 → 3 direction across the site 5-AATT. of bizelesin with a monoalkylating analog of (+)-CC-106536 shows that it appears to have increased sequence selectivity such that monoalkylating compounds like (+)-CC-1065 react at more than one site. In contrast, however, bizelesin reacts only at sites where there are two suitably positioned alkylation sites. Bizelesin generally forms interstrand cross-links with adenine primarily spaced 6 base pairs apart (including the covalently modified adenines) while occupying the intervening minor groove. Bizelesin alkylates DNA through guanine in restriction fragments in which there is a suitably positioned adenine contained in a highly reactive monoalkylation sequence 5-TTTTT N*, and in Copyright © 2005 CRC Press, LLC which N was G, C, or T. Such fragments were prepared to evaluate the cross-linking potential of bizelesin at nonadenine bases. Kinetic analysis of monoalkylation and cross-linking events dem- onstrates that it is the reaction at N (guanine or cytosine) that results in the cross-link that is the slow step. On

the basis of this analysis and the normal unreactivity of guanine and cytosine to alkylation by the CPI alkylating moiety of (+)-CC-1065, the molecular mechanism for this type of cross-linking reaction most probably involves a covalent immobilization of the second alkylating arm in a proximity-driven reaction. The sequence selectivity for DNA interstrand cross-links (ISCs) by bizelesin was determined by sequencing gel analysis37 and showed that bizelesin induced two distinct forms of DNA-ISC, one of which spanned six nucleotides and linked two adenine N3- positions within an AT-rich sequence, and the second spanned seven nucleotides, also linking two adenine N3- positions, with a preference for contiguous runs of adenines. Three major six-nucleotide DNA ISCs were identified and found to occur within 5-TAATTA-3, 5-TAAATA-3, and 5-TAAAAA-3 sequences. The major seven-nucleotide DNA-ISC was found to occur within 5-TAAAAAA-3 sequences, and within this sequence, the formation of the seven-nucleotide DNA-ISC was preferred over the six-nucleotide DNA-ISC by a ratio of approximately 2:1. DNA ISC formation within adenine tracts eliminated the inherent DNA bending associated with such sequences. Further, chemical probing of each isolated DNA ISC with diethyl pyrocarbonate (A-specific) and potassium permanganate (T-specific) showed that the major DNA conformational changes, such as helical distortion, were localized within the cross-linked sequence. These results indicate that a significant degree of DNA distortion may also be demonstrated38 and that covalent DNA adducts induced by bizelesin at the adenine N3- position undergo two subsequent competing reactions. One reaction causes DNA strand cleavage via depurination, and one reaction proceeds through loss of the DNA adduct (adduct reversal with restoration of DNA integrity). The results were obtained by studying the chemical stability of synthetic oligonucleotides, which contained either a distinct monofunctional adduct or DNA interstrand cross-links. Quantifi- cation of adduct reversal was performed on the basis that drug-modified DNA, on exposure to heat followed by hot piperidine treatment, was resistant to strand cleavage at the site of alkylation. The rate of adduct reversal was found to increase with increasing temperature and was found to be maximum at 70˚–80˚C. The rate of adduct reversal was also found to increase with increasing pH and ionic strength. Adduct reversal was favored in DNA containing cross-links, whereas rapid depurination occurred preferentially within monofunctionally alkylated DNA. The ligand-mediated polymerase chain reaction technique has been used to map drug-induced DNA alkylation sites in single-copy genes at the nucleotide level in human cells.39–41 This technique has been used to map the alkylation sites induced by adozelesin and bizelesin within DNA in human cells.42 The pattern of the drug alkylation sites is mapped within the human PGK1 and P53 genes, with the monofunctional alkylating agent, adozelisin, alkylating genomic DNA predominantly within 5-(A/T)(A/T) A* (A* = alkylated adenine) sequences, and additional sites of alkylation were observed within 5-(A/T)(G/C)(A/T) A* sequences; however, these were considered to represent sites of medium to low preference. In contrast, bizelesin, a bifunctional analog capable of both DNA monofunctional alkylation and DNA interstrand cross-link formation, was also found to alkylate 5-(A/T) (A/T) A* sequences. Putative bizelesin DNA interstrand cross-link sites indicated that AT-rich sequences are preferred in the intervening sequence between the two cross-linked adenines. Both six- and seven-nucleotide regions were identified as putative sites of DNA interstrand cross-link formation, with 5-TTTTTTA*, 5- TTTATCA*, and 5-GTACTAA* sequences being preferred. Nonadenine bases were not observed as potential intracellular sites of either DNA interstrand cross-linking formation or monofunctional alkylation. The DNA base pair preferences of the antitumor antibiotic CC-1065 and two analogs of CC- 1065 were studied by following the rate of covalent bond formation (N3 adenine adduct) with DNA oligomers containing the 5-NNTTA* and 5-NNAAA* sequences (N = nucleotide).43 The rate of adduct formation of CC-1065 is greatly affected by DNA base changes at the fourth and fifth Copyright © 2005 CRC Press, LLC positions of the binding sites for the 5-NNAAA sequences, but not the 5-NNTTA sequences. However, an analog of the CC-1065 ring system, with additional methylene and oxygen substituents, shows similar rates of adduct formation for all sequences. A second analog of CC-1065, containing a fused three-ring system, but not the methylene and oxygen substituents of CC-1065, shows similar rates of adduct formation, with the same sequence dependence as CC-1065, but does not distinguish between the sequence to the degree shown by CC-1065. Adduct formation of CC-1065, but not the analogs, competes with a reversibly bound species. Thymine bases to the 3 side of a potentially reactive adenine or cytosine base at the fifth position from the bonding adenine create reversible binding sites that decrease the rate of adduct formation of CC-1065. The sequence 5-GCGAATT binds CC-1065 only reversibly. This sequence can compete for CC-1065 with covalent binding sequences if the sites are located in different oligomers, or if the sites are located (overlapped or not overlapped) in the same oligomer. The results of these competitive binding experiments indicate that the transfer of CC-1065 from the reversible binding site to the covalent bonding site with both sites located on a single DNA duplex, not overlapped, occurs through an equilibrium of CC-1065 in solution, not by migration of CC-1065 in the minor groove. A number of indirect observations that indicate a conformational acceleration in the action of CC-1065 and its analogs have been summarized by Boger et al.44,45 The DNA alkylation reaction is derived in part from a DNA binding–induced conformational change in the agents, which substantially increases their inherent reactivity. The ground-state destabilization results from a binding-induced twist in the linkage N2 amide and requires a rigid extended N2-amide substituent, which disrupts the vinylogous amide stabilization and activates the agents for DNA alkylation. The calf thymus DNA (CT-DNA) and poly (dI-dC). poly (dI-dC) binding properties of the natural antitumor antibiotic and selected analogs of CC-1065, U-66,664, U-68,819, U-66,694, and U-65,415 were studied by circular dichroism (CD) and absorbance methods.46 The results indicate that the intense long-wavelength DNA–induced CD band of these molecules originates from a chiral electronic transition, which is delocalized over the whole molecule. Both the covalently bound species exhibit the characteristic spectral behavior of an inherently disymmetric chromophore when these agents bind within the minor groove of B-form DNA. This mechanism of optical activity accounts for why CC- 1065 shows a weak CD in buffer but a very intense induced CD at long wavelength when bound to DNA, why the intensity of the induced CD of CC-1065 analogs depends on how many fused ring systems the analog contains, and why covalently bound analogs, having the mirror image configuration of the natural configuration also exhibit an intense, positive, induced CD at long wavelength. IV. INTERACTION OF CC-1065 WITH CHROMATIN Most of the CC-1065 studies have used natural DNA or synthetic oligonucleotides for binding studies. However, nuclear DNA does not exist as naked DNA but is packaged as chromatin. Chromatin consists of repeating nucleosome subunits composed of an interior core of histones, around which are wrapped approximately 140 base pairs of the B-form DNA. Moy et al.47 reported on the interactions of CC-1065 and adozelesin with DNA, chromatin, and histones. CD binding studies show that CC-1065 and adozelesin bind to P388 chromatin and that induced CD spectra are similar to those obtained when CC-1065 binds to P388 DNA and many synthetic polynucle- otides. The intensity of the induced CD spectra and the rate of conversion of reversibly bound drug to irreversibly bound drug are less for chromatin than for naked DNA. Because the histones themselves do not bind CC-1065, this indicates that histones act indirectly to restrict the binding of CC-1065 to chromatin. This could occur because fewer binding sites are available on chromatin as compared to free DNA or because of a less dissymmetric environment of chromatin. It was assumed that a good fit in the minor groove leads to a more chiral environment for the drug chromophore and, hence, a more intense induced CD. A new type of CC-1065 analog, 52–58, where the cyclopropylpyrroloindole (CPI) moiety is replaced by 1,2-dihydro-1-(chloromethyl)-5-hydroxy-8-methyl-3H-furano [3,2-c] indole (CFI), has Copyright © 2005 CRC Press, LLC O H N B N O O H HO N N N O A O H HO O N 52 A = B = NH 55 A = S, B = NH H 58 53 A = NH, B = O 56 A = S, B = O 54 A = NH, B = S 57 A = B = S STRUCTURES 52–58 been synthesized.48 The major site of alkylation with W794 duplex DNA for 55 and (+)-CC-1065 proved to be identical. 5-d(AATTA) 3- alkylation occurs at the 3 adenine and agent binding is in the 3′→5′ direction. The profile of DNA alkylation selectivity for 52 also proved to be identical to that of adozelesin (3). However, compound 58 exhibited a different and less selective profile of DNA alkylation. The high-affinity site for 58 with W794 duplex DNA was 5′-(AATTT)-3′, with alkylation at the 5 adenine and agent binding in 5′→3′ direction. The profile of DNA alkylation exhibited by 58 proved essentially identical to that of the unnatural enantiomers (–)-CC-1065 and (–)-CPI-CDPI3.49 A more significant distinction between 52 and 58 is the relative efficiency or intensity of DNA alkylation. Consistent with the relative in vitro cytotoxic potency of 52, it proved to be comparable or slightly less effective (one to five times) at alkylating the duplex DNA than (+)-CC-1065 and much more effective than compound 58. A new CBI (1,2,9,9a-tetrahydrocyclopropa[c] benz [e] indole-4-one; 9) analog of CC-1065, CBI-PDEI2 (59) has been synthesized and evaluated.18 Compound 59 was shown to have similar DNA sequence selectivity and structural effects on DNA as (+)-CC-1065. The effect of CBI-PDE- I dimer was also compared with (+)-CC-1065 in the inhibition of duplex unwinding by helicase II and nick sealing by T4 ligase and found to be quantitatively similar. The in vitro and in vivo potencies of CBI compounds correspond very closely to the corresponding CPI derivatives. A number of 1,2,9,9a-tetrahydro cyclopropa[c]benz[e]indole-4-one (CBI; 9)–based and 9a- chloromethyl 1,2,9,9a-tetrahydro cyclopropa[c]benz[e]indole-4-one (C2BI; 60)–based analogs of CC-1065 (50) and 61 have been synthesized and evaluated by Boger et al.50,51 It was found that CBI analogs are four times more stable and potent than the corresponding CPI analogs. Similarly, the CBI-based agents alkylate DNA with unaltered sequence selectivity at an enhanced rate with greater efficiency than the corresponding CPI analogs. In comparison with the CBI-based agents, the C2BI-based agents proved to be approximately 100–10,000 times less effective at DNA alky- lation and 100–10,000 times less potent in cytotoxicity assays. These effects are suggested to be the consequence of a significant steric deceleration of the adenine N3-alkylation reaction attributed to the additional 9a-chloromethyl substituent. Consistent with this interpretation, the noncovalent binding constant of C2BI-CDPI2 (61) for poly [dA]-poly [DT] proved nearly identical to CDPI3 under kinetic binding conditions, and prolonged incubation of C2BI-CDPI2 (61) with poly [dA]- poly [dT] (72 h, 25˚C) provided covalent complexes with a helix stabilization comparable to that observed with (+) or (–)-CPI-CDPI2, indicating that the size of the C2BI subunit inhibits, but does not preclude, productive DNA alkylation. NH2 NH N N 2 O O O O N H 9 CBI 59 CBI-PDEI O 2 STRUCTURES 9, 59 Copyright © 2005 CRC Press, LLC CI CI NH NH 2 N N 2 O O O O N H 60 C2BI 61 C2BI-CDPI2 STRUCTURES 60–61 A library of 132 CBI (1,2,9,9a-tetrahydrocyclopropa[c]benzo[e]indole-4-one) analogs of CC- 1065 and the duocarmycins was prepared using the solution-phase technology of acid–base liq- uid–liquid extraction for the isolation and purification steps (Figure 20.3).52 The 132 analogs constituted a systematic study of the DNA binding domain with the incorporation of dimers composed of monocyclic, bicyclic, and tricyclic (hetero) aromatic subunits. From their examination, clear trends in cytotoxic potency and DNA alkylation efficiency emerged, highlighting the impor- tance of the first-attached DNA-binding subunit (A subunit): tricyclic is more active than bicyclic, which in turn is far more active than monocyclic (hetero) aromatic subunits. Notably, the trends observed in the cytotoxic potencies paralleled those observed in the relative efficiencies of DNA alkylation. Their interpretation of these results is that the trends represent the partitioning of the role of the DNA-binding subunits into two distinct contributions, the first of which is derived from an increase in DNA-binding selectivity and affinity, which leads to property enhancements of 10–100-fold and is embodied in the monocyclic series. The second,

which is additionally embodied in the bicyclic and tricyclic heteroaromatic subunits, is a contribution to the catalysis of the DNA alkylation reaction that provides additional enhancements of 100–1000-fold. The total enhance- ments thus exceed 25,000-fold. Aside from the significance of these observations in the design of future CC-1065/duocarmycin analogs, their importance to the design of hybrid structures containing the CC-1065/duocarmycin alkylation subunit should not be underestimated. Those that lack an attached bicyclic or tricyclic A subunit, that is, duocarmycin/distamycin hybrids, can be expected to be intrinsically poor or slow DNA alkylating agents. V. POLYAMIDE CONJUGATES AND RELATED SEQUENCE-DIRECTED STRUCTURES In a novel approach to explore the properties of cyclopropylindole (CPI), Lown et al. envisaged certain CC-1065 analogs, 62-71, in which different numbers of N-methylpyrrole amide (a DNA sequence recognizing moiety) groups are attached with different linkers to the CPI moiety.19 They postulated that the positively charged functionality of the protonated dimethylamino group in the Cl N Boc N N N HO O H O N H B seco-CBI A 11 substitutions 12 substitutions 132 compounds (12x11) FIGURE 20.3 Solution-phase synthesis of 132 CC-1065 analogs based on CBI. All the compounds were purified by liquid–liquid acid–base extraction. Copyright © 2005 CRC Press, LLC N HN HN n HN HN R N O N O O N O NH O 2 N O 62 n = 1 65 R = CH3 63 n = 2 66 R = CH2CH3 64 n = 3 67 R = CH2CH2CH3 HN 68 L = no linker N 69 L = CH2 L 70 L = CH2CH O 2 N 71 L = trans (CH=CH)2 O STRUCTURES 62–71 oligopeptide might contribute to reduced cytotoxic potency by reducing intracellular accessibility and, therefore, should be replaced by an uncharged moiety. It was also considered likely that the linker between the CPI unit and the oligopeptide would be important to the binding potency of the conjugate molecules to DNA to maintain proper contact over the whole of the molecular recognition region.53 The extent and the relative rates of DNA cleavage following alkylation and thermal treatment of these conjugates were determined by agarose gel mobility shift assay and high- resolution polyacrylamide gel electrophoresis and contrasted with that of CC-1065. The CPI-N methyl pyrrole agents avoid the major alkylation sites of CC-1065, but all alkylate the minor CC- 1065 site of 5ATAA and exhibit a consensus sequence of 5-NA/TA/TA. The cytotoxicities of these compounds were determined against KB human tumor cells in vitro. Compound 67 bearing a 4- butyramide group in the N-methyl pyrrole is 100 times more potent than compound 68, which lacks an amide group, whereas compound 71, which bears a rigid trans double-bond linker is 1000 times more potent than its flexible ethyl-linked counterpart 70. Sugiyama’s group synthesized the novel hybrid molecules 72–73, and they alkylate the 3 end of A in AT-rich sequences, as does the parent duocarmycin. More significantly, these hybrids alkylate G residues of predetermined DNA sequences effectively and with high specificity by formation of a heterodimer with distamycin.54 They also synthesized compounds 74 and 75,55 which were designed to alkylate the target sequences (A/T) G (A/T)2N(A/G) and (A/T) G (A/T) CN (A/G), respectively, according to Dervan’s ring-pairing rule. High-resolution denaturing gel electrophoresis indicated that compound 74 exclusively alkylated the 5-TGTAAAA-3 within a 400–base pair DNA fragment. Similarly, alkylation of 75 occurred exclusively at the G of the 5AGTCAGA-3 sequence with efficiency at the nanomolar concentration (Figure 20.4) H N H H O N N O H X N N O O H N O O H N N N O O N N O O N O NH HN N O O N O O O O N N O 74 X = N 72 X = N H N 75 X = CH 73 X = CH N O H X NH STRUCTURES 72–75 Copyright © 2005 CRC Press, LLC FIGURE 20.4 DNA alkylation sites of 74 and 75. Recently, Bando et al.56 developed a novel type of DNA interstrand cross-linking agent by synthesizing dimers of a pyrrole (Py)/imidazole (Im)-diamide-CPI conjugate, ImPyLDu86 (76), using seven different linkers for connection (Figure 20.5). The tetramethylene linker compound, 77, efficiently produces DNA interstrand cross-links at the 9–base pair sequence, 5- PyGGC(T/A)GCCPu-3, but only in the presence of a partner triamide, ImImPy. For efficient cross- linking by 77 with ImImPy, one AT base pair between two recognition sites was required to accommodate the linker region. Elimination of the AT base pair and insertion of an additional AT base pair, and substitution with the GC base pair, significantly reduced the degree of cross-linking. The sequence specificity of the interstrand cross-linking by 77 was also examined in the presence of various triamides. The presence of ImImIm slightly reduced the formation of a cross-linked product compared with ImImPy. The mismatch partners ImPyPy and PyImPy did not produce an interstrand cross-link product with 77, whereas ImPyPy and PyImPy induced efficient alkylation at their matching site with 77. The interstrand cross-linking abilities of 77 were further examined using denaturing polyacry- lamide gel electrophoresis with 5-Texas Red–labeled 400-and 67–base pair DNA fragments. The sequencing gel analysis of the 400–base pair DNA fragment with ImImPy demonstrated that 77 alkylates several sites on the top and bottom strands, including one interstrand cross-linkage on longer DNA fragments. A simple method using biotin-labeled complementary strands was devel- oped (Figure 20.6) and produced a band corresponding to the interstrand cross-linked site on both top and bottom strands. Densitometric analysis indicated that the interstrand cross-links in the observed alkylation bands was approximately 40% of the total. This compound efficiently cross- linked both strands at the target sequence. This system consisted of a 1:2 complex of the alkylating agent and its partner ImImPy and caused an interstrand cross-linking in a sequence-specific fashion according to the base pair recognition rule of Py-Im polyamides (Figure 20.7). H N N H O O X HN N O H O N N O N N N N O N O N O NH O NH 2 ImPyLDu86; 76 O 77 X = -CO(CH2)4CO- O STRUCTURES 76–77 (See color insert) Copyright © 2005 CRC Press, LLC (a) 5'-PyGACPu-3' : Im 3'-PuCTGPy-5' : Py (b) 5'-PyGGCAGCCPu-3' 3'-PuCCGTCGGPy-5' FIGURE 20.5 (See color insert) (A) The chemical structure of ImPyLDu86 76 and a schematic representation of the recognition of the 5-PyGACPu-3 sequence by the homodimer of 76. The arrows indicate the site where alkylation takes place. (B) A putative binding mode of the 1:2 complex of covalent dimer of 76, 77, with ImImPy to 5-PyGGCAGCCPu-3 sequence. 5'-Biotin label 400bp DNA 5'-Texas Red label Magnet + Avidin cross-linker Wash with alkaline Heat and hot piperidine Removed Sequencing Interstrand Cross-linking Site FIGURE 20.6 (See color insert) Determination of interstrand cross-linking sites. Copyright © 2005 CRC Press, LLC (a) (b) 3' 5' C G MeO C N H 2 A O O T N H C N O N N G O N H C N N N G H N O O N H G N N C N H N R A O O T R N N H C N N G H N O O N H G N N C N H N O G N N C N H N T O O A CO Me H N 2 C G 5' 3' R= = CONHCH2CH2CH2NMe2 FIGURE 20.7 (See color insert) (A) Energy-minimized structure of 77-ImImPy 2-d(CTG- GCTGCCAC)/d(GTGGCAGCCAG) interstrand cross-linked complex. DNA is drawn in white. Py and Im residues of 77 and ImImPy are drawn in blue and red, respectively. (B) A DNA interstrand cross-linking mode of the 1:2 complex of 77 with ImImPy. A series of hybrid agents, 78–84, of CC-1065 has been synthesized and evaluated and incor- porate the CBI analog of the DNA-alkylation subunit of the CC-1065 linked to the C-terminus di- and tripeptide S-DNA binding domain of bleomycin A2.57 The attachment of the DNA binding domain of bleomycin A2 did not alter the DNA-alkylation selectivity of the CBI alkylation’s subunit (5AA> 5TA), nor did it enhance the selectivity that approaches the 5 base pairs’ AT-rich alkylation’s selectivity of CC-1065. A new class of hybridization-triggered cross-linkable oligodeoxyribonucleotides (ODNs) that are conjugated to the reactive cyclopropylindole (CPI) has been synthesized and studied for their properties.58 Here, the authors conjugated racemic CPI to ODNs via a terminal phosphorothioate at either the 3 or 5 end of ODNs (85). This replaces the minor groove binding B and C subunits of CC-1065 with many more sequence-specific binding agents (ODN). These conjugates were stable in aqueous solution at neutral pH even in the presence of strong nucleophiles. When a 3- CPI-ODN conjugate was hybridized to a complementary DNA strand at 37˚C, the CPI moiety alkylated nearby adenine bases of the complement efficiently and rapidly, with a half-life of a few minutes. CPI-ODN conjugates are highly effective sequence-specific inhibitors of single-stranded viral DNA replication or gene-selective inhibitors of transcription initiation. Lown and coworkers59–61 synthesized a bis functionalized precursor of CBI, 86, and with this precursor synthesized bis-polyamide CBI conjugates of symmetrical 87 and unsymmetrical 88 types by incorporating pyrrole rings on both sides for AT specificity and pyrrole on one side and imidazole on the other side of CBI for mixed sequence recognition. The authors also explored62,63 solid-phase Copyright © 2005 CRC Press, LLC STRUCTURES 78–85 techniques for the synthesis of seco-CBI 90 and conjugates, which bears two polyamide moieties on either side of the pharmacophore 89 through an intermediate 91. Lown et al.19,20 have demon- strated that incorporation of a vinyl linker between CPI or CBI and the carrier dramatically enhances the efficiency of DNA alkylation as well as cytotoxicity. In confirmation of this significant finding, a diamide CPI conjugate possessing a vinyl linker 92 was also synthesized by Tao et al.64 Molecular modeling indicated that the insertion of a vinyl linker (L) between polyamide and CPI adjusts the location of the cyclopropane ring of the conjugates and in this way improves the alkylation efficacy STRUCTURES 86–92 Copyright © 2005 CRC Press, LLC of the conjugates. Sequence-selective alkylation of double-stranded DNA by 92 was investigated by high-resolution denaturing gel electrophoresis, using 400-bp DNA fragments. Highly efficient alkylation predominantly occurs simultaneously at the purines of the 5-PyG(A/T) Cpu-3 site on both strands at nanomolar concentrations of 92. These results indicate that the homodimer of conjugate 92 dialkylates both strands according to Dervan’s pairing rule,65 together with a new mode of recognition in which the Im-vinyl linker (L) pair target G/C base pairs. In addition to the major dialkylation sites, a minor alkylation site was also observed at 5-GT(A/T) GC-3. This alkylation can be explained by an analogous slipped homodimer recognition mode in which the L–L pair recognizes the A/T base pair. High performance liquid chromatography analysis revealed that the conjugate 92 simultaneously alkylates GN3/AN3 of the target sequences on both strands of ODNs. Eight-ring hairpin polyamides conjugated with seco-CBI were prepared by Chang et al.66 Alkylation yields and specificity were determined on a restriction fragment containing a 6–base pair match and mismatch series. Alkylation was observed at a single adenine flanking the polyamide binding site, and strand-selective cleavage could be achieved on the basis of the enantiomer of seco-CBI conjugate; a near quantitative cleavage was observed after 12 h. VI. BISALKYLATORS (DIMERS) The CPI bisalkylator, U-77,779 (bizelesin), in which the two alkylating moieties are connected with a rigid linker 1,3-bis (2-carbonyl-1H-indole-5-yl) urea group, showed promising antitumor activity,17,67 and this compound is in clinical trials. Boger et al.68 synthesized a series of four dimers, 93–96, derived from head-to-tail coupling of the two enantiomers of duocarmycin SA alkylation subunits. All the agents proved to be potent in cytotoxicity assays and displayed a two- to threefold higher activity than duocarmycin SA. CPI dimers have shown good cytotoxic and antitumor activities, and now it is well documented that the activity of the dimeric drugs is strongly related to the length and the position of the linker. To investigate the structure–activity relationship systematically, Lown and coworkers61,62,69–71 have designed three types of seco-CBI dimers, viz. C7-C7, N3-N3, and N3-C7 97-102, which contain two racemic CBI moieties linked from two different positions by a flexible methylene chain of variable length and with n numbers of pyrrole, imidazole, or thiazole moieties with the polyamides. Dimers, which are connected by a flexible methylene chain,

are generally more active against human cancer cell lines as compared to the corresponding monomer, connected to a polyamide. Recently, Lown and coworkers developed the glycosylated water-soluble version of seco-CBI polyamide conjugates (unpublished work), 103–104, which are more active against the human cancer cell lines when compared with their non-water-soluble version. A set of 10 compounds, each combining the seco-1, 2,9,9a-tetrahydrocyclo- propa[c]benz[e]indol-4-one (seco-CBI) and pyrrolo[2,1-c]benzodiazepine pharmacophores (105), was designed and prepared.72 These compounds were anticipated to cross-link between N3 of adenine and N2 of guanine in the minor groove of DNA. The compounds, which differ in the chain length separating the two alkylation subunits, and the configuration of the CBI portion, showed great variation in cellular toxicity (over four orders of magnitude in a cell line panel), with the most potent example exhibiting IC50s in the picomolar range. Cytotoxicity correlated with the ability of the compounds to cross-link naked DNA. Cross-linking was also observed in living cells, at much lower concentrations than for a related symmetrical pyrrolo[2,1-c]benzodiazepine dimer. A thermal cleavage assay was used to assess sequence selectivity, demonstrating that the CBI portion controlled the alkylation sites, whereas the pyrrolo[2,1-c]benzodiazepine substituent increased the overall efficiency of alkylation. Several compounds were tested for in vivo activity using a tumor growth delay assay against WiDr human colon carcinoma xenografts, with one compound (the most cytotoxic and most efficient cross-linker) showing a statistically significant increase in survival time following a single intravenous dose. Copyright © 2005 CRC Press, LLC STRUCTURES 93–102 Copyright © 2005 CRC Press, LLC CI O N H O N O H N N O H N OH O OH OH n H O 103 n = 1, 2, 3 OH HO H O N CI H N N O N O OH O OH HN n H O OH OH HO 104 n = 1, 2, 3 STRUCTURES 103–104 CI O O N H N n O N O OH 105 n = 1-5 STRUCTURE 105 VII. STRUCTURAL CHARACTERIZATION OF DRUG DNA COMPLEXES Bizelesin forms an adduct with d-(CGTAATTACG)2 and 1H nuclear magnetic resonance (NMR) analysis of this adduct indicated that adenines 6 base pairs apart on opposite DNA strands are cross-linked, yielding two major adducts differing in the central duplex region (5AATT3).73 In the major product, both adenines are syn oriented and Hoogsteen base paired to thymines (5HG model), whereas the other adduct contains an anti-oriented AT- step adenine that shows no evidence of hydrogen bonding with pairing thymines (50P model). Bizelesin’s size, rigidity, and cross-linking properties restrict the DNA adduct’s range of motion and freeze out DNA conformations adopted during the cross-linking process. In several sequences, bizelesin cross-links and shows overwhelm- ing preference for a 7–base pair sequence over a possible 6–base pair sequence74 (Figure 20.8). The role of the central GC base pairs in the formation of the 7–base pair cross link was investigated using two-dimensional 1H NMR studies that concentrated on the 7–base pair cross- link formed with the sequence 5TTAGTTA-3. 1H NMR analysis coupled with restrained molecular FIGURE 20.8 Crosslinking of adenines 6 base pairs apart on opposite DNA strands. Copyright © 2005 CRC Press, LLC FIGURE 20.9 (See color insert) Stereoviews of (A) the initial model involving docking of the CPI-CDPI2 (dotted van der Waals surface is shown) to a regular B-DNA decamer and (B) NOESY distance restrained, MORASS hybrid matrix/molecular dynamics final refined structure of the decamer-CPI-CDPI2 complex starting from the initial model. The A strand C1 is depicted in the upper left of the structures, with the minor groove at A17 facing away or to the right of the viewer. Note the sharp bend in the duplex near the site of alkylation in B. dynamics provided evidence for distortion around the covalently modified adenines. The modified bases are twisted toward the center of the duplex adduct because of distortion, and this effectively reduces the cross-linked distance. The restrained molecular dynamics study also showed that a hydrogen bond is formed between the exocyclic amine of the central guanine and the carbonyl of the ureylene linker. It is possible to speculate on the role of the central GC bases in this sequence preference and to propose a mechanism by which bizelesin forms a 7–base pair rather than a 6–base pair cross-link with 5-TTAGTTA-3 on the basis of the observation of the distortion in the duplex and the hydrogen bonding between the drug and DNA. Multidimensional NMR studies of the CPI-CDPI2.d-(CGCTTAAGCG)2 complex indicate that CPI-CDPI2 binds to the decamer in the same manner as CC-106575 (Figure 20.9). The effect of CPI- CDPI2 on the 1H and 31P spectra of the decamer was consistent with a minor groove-binding motif, with the drug alkylating at A17 with CDPI rings oriented toward the 5-end of the alkylated strand. NMR data indicate formation of one major and one minor adduct. The minor adduct formation could represent drug alkylation of the DNA at a secondary site or alternative orientation of the rings. VIII. CELLULAR AND PHARMACOLOGICAL STUDIES A. IN VITRO STUDIES Adozelesin is an extremely potent cytotoxic agent that causes 90% lethality after 2 h exposure in vitro of Chinese hamster ovary (CHO) and lung (V79), mouse melanoma (B16), and human Copyright © 2005 CRC Press, LLC carcinoma (A2780) cells at 0.33, 0.19, 0.2, and 0.25 ng/mL, respectively. The relative drug sensi- tivity of the cell lines (A2780 > V79, B16, CHO) was correlated to the relative amounts of [3H] adozelesin that alkylated DNA. The greater sensitivity of A2780 resulted in greater cell kill at comparable DNA alkylation levels. Phase-specific toxicity studies show that adozelesin was least lethal to CHO cell in mitosis and very early G1. Lethality increased as the cell progressed through G1 and was maximum in late G1 and early S. Adozelesin had three different effects on progression of CHO, V79, B16, and A2780 through the cell cycle: it slowed progression through S, which resulted in significantly increasing the percentage of S-phase cells (though this effect was transient); cell progression was blocked in G2 for a long time period; and the response of the cell lines to G2 block differed. CHO and V79 cells escaped G2 block by dividing and became tetraploid. In contrast, B16 and A2780 cells remained blocked in G2 and did not become tetraploid. Subsequent G2 blockade was followed by a succession of events (polyploidy and unbalanced growth) that resulted in cell death. The cell cycle kinetics have been studied by flow cytometry76 and were performed on five human gynecological cancer cell lines, AN3, AE7, BG1, HEC1A, and SKUT1B, with exposure to adozelesin at near confluency at 0, 0.1, 0.2, 0.5, 1.0, and 5X, with X = 10 pg/mL as a reference concentration for 90 min. Cell samples were taken by trypsinization at 0, 24, 48, 72, 96, and 168 h for flow cytometry. The results revealed that there was a spectrum of cell-cycle perturbations that included biphasic S and G2 blocks, reverse dose-dependent G2 blocks, and a sequential relationship of S and G2 blocks. This study demonstrated that the cell kinetics response to adozelesin depended on several variables such as cell lines, drug sensitivity, concentrations, and sampling time. The effects of CC-1065 and its analogs U-66,664, U-66,819, U-66,694, and U-71,184 have been studied on inhibition of CHO cell survival, cell progression, and their phase-specific toxic- ity.77,78 Lethality of these compounds after 2 h drug exposure was in the following order (50% lethal dose in nanomolars in parentheses), CC-1065 (0.06) > U-71,184 (1.3) > U-66,694 (3.2) > U-68,819 (171) > U- 66,664 (greater than 1200). The phase-specific cytotoxicity of U-71,184 and U-66,694 was different from that of CC-1065. CC-1065 was most cytotoxic to cells in M and early G1, and toxicity decreased as cells entered late G1 and S. In contrast, U-66,694 and U-71,184 were most toxic to cells in late G1. The biochemical and cellular effects of U-71,184 were then studied in detail; after 2 h exposure to 3 ng/mL, 90% cell kill or growth inhibition was observed, whereas 100 ng/mL was needed for similar inhibition of DNA and RNA synthesis. This discrepency between the doses indicated that inhibition of nucleic acid synthesis may not be causally related to lethality. Further studies showed that when the drug was removed after 2 h exposure, DNA synthesis continued to be inhibited, whereas RNA and protein synthesis reached levels higher than the control. Therefore, it is likely that the stimulation of RNA combined with protein synthesis leads to unbalanced growth and cell death. The adozelesin cytotoxic potency was compared with doxorubicin, cis-platin, 5-fluorouracil, and cytoxan by the adenosine triphosphate chemosensitivity assay in 10 gynecologic-cancer cell lines.79 By comparing the mean drug concentrations required to reduce the adenosine triphosphate concentrations by 50%, adozelesin was found to be approximately 100 times more potent than doxorubicin, cisplatin, 5-fluorouracil, or cytoxan. Similarly, experiments designed to investigate the cytotoxicity of adozelesin on fresh cervical and ovarian carcinoma samples demonstrated that at least a 4000-fold lower level of this agent was required compared to cisplatin in the same experiment,80 with similar results being found when the adenosine triphosphate assay referred to earlier was used with 10 gynecologic cancer cell lines.81 The cytotoxicity against CHO cells of several agents combined with adozelesin has been reported.82 The additive, synergistic, or antagonistic nature of the combined drug effect was deter- mined for most combinations using the median effect principle. In experiments with alkylating agents, combinations of adozelesin with melphalan or cisplatin were usually additive or slightly synergistic. Adozelesin–tetraplatin combinations were synergistic at several different ratios of exposure to drug. By using methylxanthines, adozelesin combined synergistically with noncytotoxic Copyright © 2005 CRC Press, LLC doses of caffeine and pentoxifylline and resulted in several logs increase in adozelesin cytotoxicity. In experiments with hypomethylating agents, adozelesin combined synergistically with 5-azacyti- dine (5-aza-CR) and 5-aza-2-deoxy cytidine (5-aza-2-CdR), and combinations of adozelesin with tetraplatin or 5-aza-2-CdR were also tested against B16 melanoma cell in vitro and were found to be additive and synergistic, respectively. Hematopoietic clonal assays were used to evaluate the effects of bizelesin on granulocyte- macrophage (CFU-gm) colony formation.83 Marrow cells were exposed in vitro to bizelesin (0.001–1000 nM) for 1 or 8 h and were then assayed for colony formation. There was a 3-log difference in drug concentration at which 100% colony inhibition occurred (1or 8 h) for murine CFU-gm versus human or canine CFU-gm. The IC70 value after an 8 h bizelesin exposure for human CFU-gm (0.006 ± 0.002 nM) was 2220 times lower than for murine CFU-gm (13.32 ± 8.31 nM). At any given concentration, an 8-h drug exposure resulted in greater colony inhibition than a 1-h exposure for all species (P < 0.05). Increasing exposure time from 1 to 8 h increased toxicity to human and canine CFU-gm much more than to murine CFU-gm. The results of these in vitro clonal assays were quantitatively consistent with those seen in whole animal studies, indicating that bizelesin could be a potent myelosuppressive agent in the clinic. Because the dose-limiting toxicity in preclinical models is myelosuppression, and because the in vitro sensitivity of human and canine CFU-gm is similar, the canine maximum tolerated dose is better than the murine maximum tolerated dose to determine a safe starting dose for phase I clinical trials. The DNA damage and differential cytotoxicity produced in human carcinoma cells by adoze- lesin and bizelesin have been studied.84 The concentration of adozelesin required to produce a 1- log cell kill in six human tumor cell lines varied from 20 to 60 pM, whereas with bizelesin, the comparable concentrations ranged from 1 to 20 pM. The cytotoxicity of adozelesin and bizelesin was found to be independent of the guanine O6-alkyl transferase phenotype. The sensitivities of the BE and HT-29 human colon carcinoma cells were increased when the time of drug exposure was varied from 2 to 6 h, with DNA interstrand cross-links, as measured by the technique of alkaline elution, only detected when HT-29 or BE cells were exposed to extremely high concentrations of bizelesin for 6 h. The cytotoxic actions of adozelesin against the human colon (HT-29, DLD-1) and lung (SK) carcinoma cell lines have been studied.85 The concentrations of adozelesin that produced 50% cell kill for 4- and 24-h exposures were in the range of 0.001–0.02 ng/mL for both colon and lung carcinoma cells, indicating that this analog is

a very potent cytotoxic agent. Because most clinical regimens for tumor therapy consist of several drugs, the antineoplastic action of adozelesin in combination with 5-aza-2-CdR, a potent inhibitor of DNA methylation, or cytosine arabinoside (Ara-C), a potent inhibitor of DNA synthesis, was examined. The adozelesin plus 5-Aza-CdR combination showed a synergistic effect on cytotoxicity against DLD-1 colon carcinoma cells for both a 6- and 24-h exposure. However, the combination of adozelesin and Ara-C for a 6-h exposure showed an antagonistic effect, whereas 24-h exposure showed a synergistic effect. These preclinical results provided some preliminary data on possible drugs that might be useful in combination with adozelesin in clinical trials. Bis-indolyl-(seco)-1,2,9a-tetrahydrocyclopropa[c] benz [e] indole-4-one compounds are syn- thetic analogs of CC-1065 that are highly cytotoxic toward a broad spectrum of tumor cell lines. One of these compounds, DC1, was conjugated to antibodies via novel cleavable disulfide linkers. Conjugates of DC1 with the murine mAbs anti-Bu and N901 directed against tumor-associated antigens CD19 and CD56, respectively, proved to be extremely potent and antigen selective in killing target cells in culture.86 DC1 conjugates with humanized versions of anti-B4 and N901 antibodies were also examined and shown to be as cytotoxic and selective as the respective murine antibody conjugates. The anti-B4-DC1 conjugate showed antitumor efficacy in an aggressive metastatic human B-cell lymphoma survival model in SCID mice and completely cured animals bearing large tumors. Anti-B4-DC1 was considerably more effective in this tumor model than doxorubicin, cyclophosphamide, etoposide, or vincristine at their maximum tolerated doses. Copyright © 2005 CRC Press, LLC The mutagenicity for V-79 cell (6-thioguanine resistance and salmonella histidine auxotropy or aza guanine resistance) of selected CPI analogs was compared with DNA-binding activity, and the structure–activity relationship was determined. The compounds CC-1065, U-62,736, U-66,866, U-66,694, U-67,786, and U-68,415 all have an A segment with an intact cyclopropyl group and different B segments,87 whereas the cyclopropyl group is absent from U-66,226 and U-63,360. This absence of the cyclopropyl ring diminished the cytotoxic and mutagenic potency of the compounds such that U-63,360 was nearly three orders of magnitude less potent than CC-1065 in V-79 cells. For the compounds with an intact cyclopropyl group, the order of cytotoxic and mutagenic potency in V79 cells generally correlated with binding to calf thymus DNA and increased with the length of the B segment; the order of cytotoxicity being CC-1065> U-68,415> U-66,694 > U-66,866> U-62,736. These results show that an electrophilic carbon afforded by an intact cyclopropyl group of this type is necessary but not sufficient to account for the high cytotoxic and mutagenic potency of CC- 1065 and U-68,415. The size and characteristics of the B segment also affect the potency. The authors speculate that the more cytotoxic analogs are less mutagenic because they may have greater structure directed binding to less mutable DNA sites in the minor groove. B. IN VIVO STUDIES The analog of CPI or its kinetically equivalent ring opened seco precursor, carzelesin (5), was evaluated against a series of human xenografts growing at the subcutaneous sites.88 The model consisted of seven colon adenocarcinomas and six pediatric rhabdomyosarcomas. Carzelesin was given a single intravenous injection, and the tumor volumes were determined at 7-day intervals. At the highest dose (0.5 mg/kg), the dose producing 10% lethality (LD10), carzelesin significantly inhibited growth in four of the colon tumor lines; in contrast, on lowering the dose to 0.25 mg/kg, only two of the seven colon adenocarcinomas exhibited significant growth inhibition. However, with the rhabdomyosarcoma lines, all were significantly arrested in growth. There was no apparent cross-resistance to carzelesin in two rhabdomyosarcomas selected for vincristin resistance (RH 12/VCR, Rh 18/VCR) or in Rh 28/LPAM xenografts selected for primary resistance to the bifunc- tional alkylating agent, melphalan. Carzelesin maintained full activity against Rh 18/Topo tumors selected in situ for resistance to topotecan. Adozelesin (3), which was selected from a series of analogs of CC-1065 for its superior in vivo and antitumor activity, was highly active when administered intravenously against intraperi- tonel or subcutaneous implanted murine tumors, including L1210 leukemia, B16 melanoma, MS076 sarcoma, and colon 38 carcinoma, and produced long-term survivors in mice bearing intravenous inoculated L1210 and Lewis lung carcinomas.14 In addition, modest activity was shown against the highly drug-resistant pancreas 02 carcinoma. Adozelesin was also highly effective against human tumor xenografts, implanted subcutaneously in athymic (nude) mice, including colon CX-1 adenocarcinoma, lung LX-1 tumor, clear cell caki-1 carcinoma, and ovarian 2780 carcinoma. Its broad spectrum of in vivo activity compared favorably with the three widely used antitumor drugs: cisplatin, cyclophosphamide, and doxorubicin. Adozelesin appeared to be more effective than these drugs in the treatment of various resistant tumors such as subcutaneously implanted mouse B16 melanoma, pancreatic 02 carcinoma, human colon CX-1, and human lung LX-1 tumor xenografts, and based on its high potency and high efficacy, it was chosen for clinical investigation and development. The mutagenic potential of the antitumor compounds CC-1065 and adozelesin were assessed89 to track the in vivo fate of their unique modifications at the nucleotide level. Mice were inoculated with a single therapeutic dose of these agents and sacrificed at 18 h, 3 d, or 15 d for extraction and analysis of liver DNA. Although undetectable at 18 h posttreatment, by 72 h a threefold increase in mutant frequency was observed in drug-treated animals such that sequence analysis of drug- induced mutations could be performed and a direct comparison made between in vitro and in vivo Copyright © 2005 CRC Press, LLC DNA alkylation. Base substitution involving guanine or cytosine accounted for 64% of the 41 mutations that occurred at a cyclopropylindole alkylation site, whereas 23 mutations occurred 1–4 nucleotides from a potentially alkylated adenine. Following intravenous administration of bizelesin (15 µg/kg) to male CD2F1 mice, the plasma elimination of cytotoxic activity resulting from bizelesin in the L1210 cell bioassay was described by a two-compartment open model90; the A-phase (t1/2 A) and B-phase (t1/2 B) half lives, steady- state volume distribution, and total body clearance were 3.5 min, 7.3 h, 7.641 mL/kg, and 16.3 mL/min per kilogram, respectively. The drug level following intraperitoneal administration was at least 10 times lower than that resulting from intravenous infusion. Following intravenous or intraperitoneal administration, the recovery of material in urine was <0.1% of the delivered dose. The low urine recovery indicated an extensive interaction of the parent drug and other reactive species with macromolecules or further metabolism or degradation to inactive species. IX. CLINICAL STUDIES Adozelesin is a potent synthetic analog that was chosen for clinical development because it had a preclinical antitumor spectrum that was similar to CC-1065 at therapeutic doses but did not produce deaths. Phase I evaluations using a variety of adozelesin treatment schedules have been reported.91,92 Adozelesin was given as a 10-min intravenous infusion for five consecutive days every 21 d on recovery from toxicity. The dose range evaluated was 6–30 µg/m2 per day. All patients had refractory solid tumors and had received prior cytotoxic drug treatment. Thirty-three patients (22 men, 11 women) were entered in the study, and 87 courses of treatment were initiated. Dose-limiting toxicity was cumulative myelosuppression (leukopenia, thrombocytopenia). The maximum tolerated dose was 30 µg/m2 per day. The only other significant toxicity was an anaphylactoid syndrome that occurred in two patients. A partial response was observed in a patient with refractory soft tissue sarcoma. The recommended phase II starting dose of adozelesin using 10-min intravenous infusion for five consec- utive days was 25 µg/m2 per day, to be repeated every 4–6 weeks to allow recovery from myelotoxicity. As of late September 2004, the only agent that is still listed as being in clinical trials is carzelesin, which is quoted in the Prous Integrity™ database as being in phase II and under “active develop- ment.” The corresponding National Cancer Institute (NCI) database does not list the compound as being in any current trial. The differences could be a result of the NCI database not yet containing all clinical trial data from pharmaceutical companies. X. CONCLUSIONS AND PROSPECTS The sequence selectivity inherent in even the simplest minor groove ligands has enabled them to be used as probes of DNA structure and dynamics. From these data has arisen the concept of devising sequence-specific agents such as polyamides capable of single-gene recognition, and thus of being targeted to single deleterious genes. An extensive series of studies has led to the devel- opment of synthetic approaches to the natural products CC-1065 and the duocarmycins. The present understanding of the origin of the properties of the natural products emerged from the design and evaluation of synthetic substrates and agents containing systematic structural modifications, in addition to extensive studies conducted on the natural product themselves. As a consequence of these efforts, and despite expectations of the unique behavior of the natural products, potent and efficacious antitumor activity has not only been observed with such analogs but both their potency and efficacy may even exceed those of the natural products. This illustrates that their useful properties may be enhanced by well-founded and well-designed structural modifications. The active units, when coupled with certain appropriate linkers as dimers, have proven to be even more potent and are in clinical trials. Conjugating the pharmacophore unit of CC-1065 with the DNA-binding polyamides indicates a promising additional approach for developing new types of sequence-specific Copyright © 2005 CRC Press, LLC DNA alkylating agents. The evident advantages of single-gene agents for molecular medicine will determine much of the future emphasis of this area, provided the formidable problems of converting them into pharmacologically acceptable drugs can be overcome. REFERENCES 1. Hanka, L. J. et al., CC-1065 (NSC-298223), a new antitumor antibiotic: Production, in vitro biological activity, microbiological assays and taxonomy of the producing microorganisms, J. Antibiot., 31, 1211, 1978. 2. Martin, D. G. et al., CC-1065 (NSC 298223), a potent new antitumor agent: Improved production and isolation, characterization and antitumor activity, J. Antibiot., 34, 1119, 1981. 3. Bhuyan, B. K. et al., CC-1065 (NSC-298223), a most potent antitumor agent, kinetics of inhibition of growth DNA synthesis, and cell survival, Cancer Res., 42, 3532, 1982. 4. Reynolds, V. L. et al., Reaction of the antitumor antibiotic CC-1065 with DNA location of the site of thermally induced strand breakage analysis of DNA sequence specificity, Biochemistry, 24, 6228, 1985. 5. Hurley, L. H. et al., Reaction of the antitumor antibiotic CC-1065 with DNA: Structure of a DNA adduct with DNA sequence specificity, Science, 226, 843, 1984. 6. Hurley, L. H. et al., Molecular basis for sequence-specific DNA alkylation by CC-1065, Biochemistry, 27, 3886, 1988. 7. Hurley, L. H. et al., Sequence specificity of DNA alkylation by the unnatural enantiomer of CC-1065 and its synthetic analogs, J. Am. Chem. Soc., 112, 4633, 1990. 8. Weiland, K. L. and Dooley, J. P., In vitro and in vivo DNA binding by the CC-1065 analogue U- 73975, Biochemistry, 30, 7559, 1991. 9. Boger, D. L., Munk, S. A., and Zarrinmayeh, H., (+)-CC-1065 DNA alkylation: Key studies demon- strating a no covalent binding selectivity contribution to the alkylation selectivity, J. Am. Chem. Soc., 113, 3980, 1991. 10. Boger, D. L. et al., An alternative and convenient strategy for generation of substantial quantities of singly5-32P-end labeled double stranded DNA for binding studies. Development of a protocol for examination of functional features of (+)-CC-1065 and ducaromycins that contribute to their sequence selective DNA alkylation properties, Tetrahedron, 47, 2661, 1991. 11. Li, H. et al., CC-1065 (NSC 298223), a novel antitumor agent that interacts strongly with double- stranded DNA, Cancer Res., 42, 999, 1982. 12. McGovren, J. P. et al., Preliminary toxicity studies with the DNA-binding antibiotic, CC-1065, J. Antibiot., 37, 63, 1984. 13. Warpehoski, M. A. et al., Stereoelectronic factors influencing the biological activity and DNA inter- action of synthetic antitumor agents modeled on CC-1065, J. Med. Chem., 31, 590, 1988. 14. Li, L. H. et al., Adozelesin, a selected lead among cyclopropyl pyrroleindole analogs of the DNA binding antibiotic, CC-1065, Invest. New Drugs, 9, 137, 1991. 15. Ichimura, M. et al., Duocarmycin SA, a new antitumor antibiotic from Streptomyces sp., J. Antibiot., 43, 1037, 1990. 16. Sugiyama, H. et al., Covalent alkylation of DNA with duocarmycin A. Identification of a basic site structure, Tetrahedron Lett., 31, 7197, 1990. 17. Mitchell, M. A. et al., Synthesis and DNA cross-linking by a rigid CPI dimer, J. Am. Chem. Soc., 113, 8994, 1991. 18.

Aristoff, P. A. et al., Synthesis and biochemical evaluation of the CBI-PDE-I dimer, a benzannelated analog of (+)-CC-1065 that also produces delayed toxicity in mice, J. Med. Chem., 36, 1956, 1993. 19. Wang, Y. et al., Design, synthesis, cytotoxic properties and preliminary DNA sequencing evaluation of cpi-n-methyl pyrrole hybrids. Enhancing effect of a trans double bond linker and role of the terminal amide functionality on cytotoxic potency, Anticancer Drug Des., 11, 15, 1996. 20. Iida, H. and Lown, J. W., Design and synthesis of cyclopropylpyrroloindole (CPI)-lexitropsin conju- gates and related structure, Synth. Org. Chem., 1, 1998. 21. Reynolds, V. L., McGovren, J. P., and Hurley, L. H., The chemistry, mechanism of action and biological properties of CC-1065, a potent antitumor antibiotic, J. Antibiot., 39, 1986. Copyright © 2005 CRC Press, LLC 22. Boger, D. L. et al., Synthesis and evaluation of aborted and extended CC-1065 functional analogs: (+)- and ()-cpi-pde-i1, (+)- and ()-cpi-cdpi1, and (+-.)-, (+)-, and ()-CPI-CDPI3. Preparation of key partial structures and definition of an additional functional role of the CC-1065 central and right-hand subunits, J. Am. Chem. Soc., 112, 4623, 1990. 23. Boger, D. L., Johnson, D. S., and Yun, W., (+)- and ent-(–)-duocarmycin SA and (+)- and ent-(–)-n- boc-dsa DNA alkylation properties. Alkylation site models that accommodate the offset at-rich adenine N3 alkylation selectivity of the enantiomeric agents, J. Am. Chem. Soc., 116, 1635, 1994. 24. Boger, D. L. and Johnson, D. S., CC-1065 and the duocarmycins: Understanding their biological function through the mechanistic studies, Angew. Chem. Int. Ed. Engl., 35, 1439, 1996. 25. Fukuda, Y. et al., Synthetic studies on duocarmycin. 1. Total synthesis of dl-duocarmycin A and its 2-epimer, Tetrahedron, 50, 2793, 1994. 26. Fukuda, Y., Nakatani, K., and Tershima, S., Synthetic studies on duocarmycin. 2. Synthesis and cyto- toxicity of natural (+)-duocarmycin A and its three possible stereoisomers, Tetrahedron, 50, 2809, 1994. 27. Boger, D. L. and Mesini, P., DNA alkylation properties of CC-1065 and duocarmycin analogs incorporating the 2,3,10,10a-tetrahydrocyclopropa[d]benzo[f]quinol-5-one alkylation subunit: Identi- fication of subtle structural features that contribute to the regioselectivity of the adenine N3 alkylation reaction, J. Am. Chem. Soc., 117, 11647, 1995. 28. Wang, Y. et al., CC-1065 functional analogues possessing different electron withdrawing substituents and leaving groups: Synthesis, kinetics and sequence specificity of reaction with DNA and biological evaluation, J. Med. Chem., 36, 4172, 1993. 29. Lown, J. W. et al., Molecular recognition between oligopeptides and nucleic acids: Novel imidazole containing oligopeptides related to netropsin that exhibited altered DNA sequence specificity, Bio- chemistry, 27, 3886, 1983. 30. Sugiyama, H. et al., A novel guanine N3 alkylation by antitumor antibiotic duocarmycin a, Tetrahedron Lett., 34, 2179, 1993. 31. Dickerson, R. E., Kopka, M. L., and Pjura, P. E., The major and minor grooves of the DNA helix as conduits for information transfer, in DNA ligand interaction, from drugs to proteins, Guschlbauer, W. and Saenger, W., Eds., Plenum Press, New York, 1987, 45. 32. Boger, D. L. et al., Demonstration of a pronounced effect of nonvovalent binding selectivity on the (+) CC-1065 DNA alkylation and identification of the pharmacophore of the alkylation subunit, Proc. Natl. Acad. Sci. USA, 88, 1431, 1991. 33. Boger, D. L., Zhou, J., and Cai, H., Demonstration and defination of the noncovalent binding selectivity of agents related to CC-1065 by an affinity cleavage agent: Non covalent binding coincidental with alkylation, Bioorg. Med. Chem., 4, 859, 1996. 34. Boger, D. L. et al., Molecular basis for sequence selective DNA alkylation by (+)- and enant-(–)-CC- 1065 and related agents: Alkylation site models that accommodate the offest at-rich adenine n3 alkylation selectivity, Bioorg. Med. Chem., 2, 115, 1994. 35. Swenson, D. H. et al., Evaluation of DNA binding characteristics of some CC-1065 analogues, Chem. Biol. Interactions, 67, 199, 1985. 36. Ding, Z. M. and Hurley, L. H., DNA interstrand cross-linking, DNA sequence specificity, and induced conformational changes produced by dimeric analogue of (+)-CC-1065, Anticancer Drug Design, 6, 427, 1991. 37. Lee, C. S. and Gibson, N. W., Nucleotide preferences for DNA interstrand crosslinking induced by the cyclopropyl pyrrole indole analogue U-77, 779, Biochemistry, 32, 2592, 1993. 38. Lee, C. S. and Gibson, N. W., DNA interstrand cross-links induced by the cyclopropyl pyrroloindole antitumor agent bizelisin are reversible upon exposure to alkali, Biochemistry, 32, 9108, 1993. 39. Muller, P. R. and Wold, B., In vivo footprinting of a muscle specific enhancer by ligation mediated PCR, Science, 246, 780, 1989. 40. Pfeifer, G. P. et al., Genomic sequencing and methylation analysis by ligation mediated PCR, Science, 246, 810, 1989. 41. Garrity, P. A. and Wold, B. J., Effects of different DNA polymerases in ligation mediated PCR: Enhanced genomic sequencing and in vivo footprinting, Proc. Natl. Acad. Sci. USA, 89, 1021, 1992. 42. Lee, C. S., Pfeifer, G. P., and Gibson, N. W., Mapping of DNA alkylation sites induced by adozelesin and bizelesin in human cells by ligation-mediated polymerase chain reaction, Biochemistry, 33, 6024, 1994. Copyright © 2005 CRC Press, LLC 43. Krueger, W. C. et al., DNA sequence recognition by the antitumor antibiotic CC-1065 and analogues of CC-1065, Chem. Biol. Interactions, 79, 265, 1991. 44. Boger, D. L. and Johnson, D. S., CC-1065 and duocarmycins: Unraveling the keys to a new class of naturally derived DNA alkylating agents, Proc. Natl. Acad. Sci. USA, 92, 3642, 1995. 45. Boger, D. L. and Garbaccio, R. M., Catalysis of the CC-1065 and duocarmycin DNA alkylation reaction: DNA binding induced conformational change in the agent results inactivation, Bioorg. Med. Chem., 5, 263, 1997. 46. Krueger, W. C. and Prairie, M. D., The origin of the DNA induced circular dichroism of CC-1065 and analogues, Chem. Biol. Interact., 79, 137, 1991. 47. Moy, B. C. et al., Interaction of CC-1065 and its analogues with mouse DNA and chromatin, Cancer Res., 49, 1983, 1989. 48. Mohamdi, F. et al., Total synthesis and biological properties of novel antineoplastic (chlorome- thyl)furanoindolines, an asymmetric hydroboration mediated synthesis of the alkylation subunit, J. Med. Chem., 37, 232, 1994. 49. Coleman, R. S. et al., Synthesis and evaluation of aborted and extended CC-1065 functional analogues: (+) and (–) CPI-PDEI3 preparation of key partial structures and definition of an additional functional role of CC-1065 central and right hand subunits, J. Am. Chem. Soc., 112, 4623, 1990. 50. Boger, D. L. et al., Evaluation of functional analogues of CC-1065 and the duocarmycins incorporating the cross-linking 9a-chloromethyl-1, 2,9,9a-tetrahydrocyclopropa[c] benz [e] indole-4-one (C2BI) alkylation subunit, Bioorg. Med. Chem, 1, 27, 1993. 51. Boger, D. L., Yun, W., and Han, N., 1,2,9,9a-tetrahydro cyclopropa[c] benz [e] indole-4-one (CBI) analogues of CC-1065 and the duocarmycins: Synthesis and evaluation, Bioorg. Med. Chem., 3, 1429, 1995. 52. Boger, D. L. et al., Parallel synthesis and evaluation of 132 (+)-1,2,9,9a-tetrahydrocyclo- propa[c]benz[e]indol-4-one (CBI) analogues of CC-1065 and the duocarmycins defining the contri- bution of the DNA-binding domain, J. Org. Chem., 66, 6654, 2001. 53. Reddy, B. S. P., Sharma, S. K., and Lown, J. W., Recent developments in sequence selective minor groove DNA effectors, Curr. Med. Chem., 8, 475, 2001. 54. Tao, Z. F. et al., Sequence-specific DNA alkylation by hybrid molecules between segment A of duocarmycin A and pyrrole/imidazole diamide, Angew. Chem. Int. Ed. Engl., 38, 650, 1999. 55. Tao, Z. F. et al., Rational design of sequence-specific DNA alkylating agents based on duocarmycin A and pyrrole-imidazole hairpin polyamides, J. Am. Chem. Soc., 121, 4961, 1999. 56. Bando, T. et al., Highly efficient sequence-specific DNA interstrand cross-linking by pyrrole/imidazole CPI conjugates, J. Am. Chem. Soc., 125, 3471, 2003. 57. Boger, D. L. and Han, N., CC-1065/duocarmycin and bleomycin A2 hybrid agents. Lack of enhance- ment of DNA alkylation by attachment to non-complementary DNA binding subunits, Bioorg. Med. Chem., 5, 233, 1997. 58. Lukhatanov, E. A. et al., Rapid and efficient hybridization-triggered cross linking within a DNA duplex by an oligodeoxyribonucleotide bearing a conjugated cyclopropyl pyrrole indole, Nucleic Acids Res., 24, 683, 1996. 59. Jia, G., Iida, H., and Lown, J. W., Synthesis of bis-lexitropsin-1,2,9,9a-tetrahydrocyclo- propa[c]benz[e]indole-4-one (CBI) conjugates, Heterocyclic Comm., 4, 1998. 60. Jia, G., Iida, H., and Lown, J. W., Synthesis of an unsymmetrical bis-lexitropsin-1,2,9,9a-tetrahydro- cyclo-propa[c]benz[e]indole-4-one (CBI) conjugates, Chem. Comm., 119, 1999. 61. Jia, G., Iida, H., and Lown, J. W., Design and synthesis of 1,2,9,9a-tetrahydrocyclo- propa[c]benz[e]indole-4-one (seco-CBI) dimers, Heterocyclic Comm., 5, 497, 1999. 62. Jia, G., Iida, H., and Lown, J. W., Solid-phase synthesis of 1-chloromethyl-1,2-dihydro-3h- benz[e]indole (seco-CBI) and polyamide conjugate, Synlett., 603, 2000. 63. Jia, G. and Lown, J. W., Design, synthesis and cytotoxicity evaluation of 1-chloromethyl-5-hydroxy- 1,2-dihydro-3h-benz[e]indole (seco-CBI) dimers, Bioorg. Med. Chem., 8, 1607, 2000. 64. Tao, Z. F., Saito, I., and Sugiyama, H., Highly cooperative DNA dialkylation by the homodimer of imidazole-pyrrole diamide-CPI conjugate with vinyl linker, J. Am. Chem. Soc., 122, 1602, 2000. 65. Wemmer, D. E. and Dervan, P. B., Targeting the minor groove of DNA, Curr. Opin. Struct. Biol., 7, 355, 1997. Copyright © 2005 CRC Press, LLC 66. Chang, A. Y. and Dervan, P. B., Strand selective cleavage of DNA by diastereomers of hairpin polyamide-seco-CBI conjugates, J. Am. Chem. Soc., 122, 4856, 2000. 67. Mitchell, M. A. et al., Interstrand DNA cross-linking with dimers of the spirocyclopropyl alkylating moiety of CC-1065, J. Am. Chem. Soc., 111, 6428, 1989. 68. Boger, D. L. et al., Bifunctional alkylating agents derived from duocarmycin SA: Potent antitumor activity with altered sequence selectivity, Bioorg. Med. Chem. Lett., 10, 495, 2000. 69. Kumar, R. and Lown, J. W., Design and synthesis of bis 1-chloromethyl-5-hydroxy-1, 2-dihydro-3h- benz [e] indole (seco-CBI)-pyrrole polyamide conjugates, Org. Lett., 4, 1851, 2002. 70. Kumar, R. et al., Synthesis of novel thiazole-contaning bis 1-chloromethyl-5-hydroxy-1, 2-dihydro- 3h-benz [e] indole (seco-CBI)-polyamide conjugates, as anticancer agents, Heterocyclic Commun., 8, 521, 2003. 71. Kumar, R. and Lown, J. W., Synthesis and cytotoxicity evaluation of novel c7-c7, c7-n3 & n3-n3 dimers of 1-chloromethyl-5- hydroxy-1,2-dihydro-3h-benz [e]indole (seco-CBI) with pyrrole and imidazole polyamide conjugates, Org. Biomol. Chem., 1, 2630, 2003. 72. Tercel, M. et al., Unsymmetrical DNA cross-linking agents: Combination of the CBI and PBD pharmacophores, J. Med. Chem., 46, 2132, 2003. 73. Seaman, F. C. and Hurley, L. H., Interstrand cross-linking by bizelesin produces a Watson-Crick to Hoogsteen base pairing transition region in d(CGTAATTACG)2, Biochemistry, 32, 12577, 1993. 74. Thompson, A. S. et al., Determination of structural role of the internal guanine-cytosine base pair in recognition of a seven base pair sequence cross-linked by bizelesin, Biochemistry, 34, 11005, 1995. 75. Powers, R. and Gorenstein, D. G., Two-dimensional 1H and 31P NMR spectra restrained molecular dynamics structure of covalent CPI-CDPI2-oligodeoxy ribonucleotide decamer complex, Biochemis- try, 29, 9994, 1990. 76. Nguyen, H. N. et al., Spectrum of cell-cycle kinetics of alkylating agent adozelesin in gynecological cancer cell lines: Correlation with drug induced cytotoxicity, J. Cancer Res. Clin. Oncol., 118, 515, 1992. 77. Adams, E. G., Bardiner, G. J., and Bhuyan, B. K., Effects of U-71, 184 and several other CC-1065 analogues on cell survival and cell cycle of Chinese hamster ovary cells, Cancer Res., 48, 109, 1988. 78. Zsido, T. J. et al., Resistance of CHO cells expressing p-glycoprotein to cyclopropylpyrroloindole (CPI) alkylating agents, Biochem. Pharmacol., 43, 1817, 1992. 79. Nguyen, H. N. et al., In vitro evaluation of novel chemotherapeutic agent, adozelesin, in gynecologic- cancer cell lines, Cancer Chemother. Pharmacol., 30, 37, 1992. 80. Hightower, R. D. et al., Comparison of U-73,975 and cisplatin cytotoxicity in fresh cervical and ovarian carcinoma specimens with ATP-chemosensitivity assay, Gynecol. Oncol., 47, 186, 1992. 81. Hightower, R. D. et al., In vitro evaluation of the novel chemotherapeutic agents U-73, 975, U-77, 779, and U-80, 244 in gynecologic cancer cell lines, Cancer Invest., 11, 276, 1993. 82. Smith, K. S. et al., Synergistic and additive combinations of several antitumor drugs and other agents with the potent alkylating agent adozelesin, Cancer Chemother. Pharmacol., 35, 471, 1995. 83. Volpe, D. A. et al., Myelotoxic effects of the bifunctional alkylating agent bizelesin on human, canine and murine myeloid progenitor cells, Cancer Chemother. Pharmacol., 39, 143, 1996. 84. Lee, C. S. and Gibson, N. W., DNA damage and differential cytotoxicity produced in human carcinoma cells by CC-1065 analogues, U-73, 975 and U-77, 779, Cancer. Res., 51, 6586, 1991. 85. Cote, S. and Momparler, R. L., Evaluation of the antineoplastic activity of adozelesin alone and in combination with 5-aza-2-deoxycytidine and cytosine arabinoside on DLD-1 human colon carcinoma cells, Anticancer Drugs, 4, 327, 1993. 86. Chari, R. V. et al., Enhancement of the selectivity and antitumor efficacy of a CC-1065 analogue through immuno conjugate

formation, Cancer Res., 55, 4079, 1995. 87. Harbach, P. R. et al., Mutagenicity of the antitumor antibiotic CC-1065 and its analogues in mammalian (V79) cells and bacteria, Cancer Res., 48, 32, 1988. 88. Houghton, P. J. et al., Therapeutic efficacy of the cyclopropylpyrroloindole, carzelesin, against xenografts derived from adult and childhood solid tumors, Cancer Chemother. Pharmacol., 36, 45, 1995. 89. Monroe, T. J. and Mitchell, M. A., In vivo mutagenesis induced by CC-1065 and adozelesin DNA alkylation in a transgenic mouse model, Cancer Res., 53, 5690, 1993. Copyright © 2005 CRC Press, LLC 90. Walker, D. L., Reid, J. M., and Ames, M. M., Preclinical pharmacology of bizelesin, a potent bifunctional analogue of the DNA-binding antibiotic CC-1065, Cancer Chemother. Pharmacol., 34, 317, 1994. 91. Shamdas, G. J. et al., Phase I study of adozelesin (U-73, 975) in patients with solid tumors, Anticancer Drug Des, 5, 10, 1994. 92. Foster, B. J. et al., Phase I trial of adozelesin using the treatment schedule of daily x 5 every three weeks, Invest. New Drugs, 13, 321, 1996. Copyright © 2005 CRC Press, LLC 21 Epothilone, a Myxobacterial Metabolite with Promising Antitumor Activity Gerhard Höfle and Hans Reichenbach CONTENTS I. Introduction II. Natural Epothilones III. Mechanism of Action IV. Total Synthesis and Semisynthesis A. Total Synthesis B. Semisynthesis V. Structure–Activity Relationships VI. Preclinical Studies VII. Clinical Applications Note Added in Proof Acknowledgments References I. INTRODUCTION Epothilone is a secondary metabolite of the myxobacterium Sorangium cellulosum. It is a mac- rolactone of a novel structural type characterized by an epoxy and a ketogroup in the lactone ring and a side chain with a thiazole ring (Figure 21.1) — functional groups from which the name epo-thi-lone was coined.1 Two main variants, epothilone A and epothilone B, are normally isolated from the bacterial culture.2,3 The B variant is a homolog with an extra methyl group at the epoxide, which makes it by a factor of 10 more active. Similarly, minor and trace components are always observed as pairs of homologs, with the B-series being invariably more active.4 Epothilone, originally spelled epothilon, was discovered by us in 1986 because of its antifungal activity. Its structure was first disclosed in 1991 in the Annual Scientific Report of our institute, Gesellschaft für Biotechnologische Forschung (GBF).1 The structure was elucidated mainly by spectroscopic methods and corroborated by x-ray crystallography, which also provided the absolute configura- tion.3 Besides antifungal activity, strong cytotoxicity against mammalian cells was also reported. A German patent was filed in November 1991 (granted in May 1994), and an international one November 1992,5 though it was withdrawn in April 1994 because the impressive antitumor activity of the epothilones was not, as yet, known, and there was limited interest in purely cytotoxic compounds. The patent strain, Sorangium cellulosum So ce90, was deposited at the German Culture Collection (Deutsche Sammlung von Mikroorganismen und Zellkulturen, DSMZ) in Braunschweig in October 1991 under number DSM 6773 and has been available to the public since then. Copyright © 2005 CRC Press, LLC 26 O 27 12 10 O S 19 9 S 21 13 11 15 8 N 14 OH N OH 17 16 7 2 O 1 3 5 O 4 6 O OH O O OH O Epothilone A 1a Epothilone B 1b FIGURE 21.1 Structures of the major epothilones. The discovery of epothilone was the result of a decade of basic research at the GBF with the aim of identifying new producers of diverse secondary metabolites among bacteria. This has indeed been achieved, with the discovery that myxobacteria synthesize a wealth of compounds, most of which possess diverse and novel structures and unusual and interesting mechanisms of action. Thus far, we have isolated and characterized more than 600 different compounds from those organisms, representing around 120 structural types,6 and it is quite obvious that our many publications have stimulated more extensive investigation of myxobacteria for secondary metabolites. In the 1980s and 1990s Taxol (paclitaxel) was developed into a very successful antitumor drug with a new mechanism of action, viz, stimulation of tubulin polymerization and stabilization of microtubuli. As Taxol initially had to be produced from the bark of slow-growing Pacific yew trees, and there was little chance of development of an economically viable total synthesis, several phar- maceutical companies started screening for Taxol mimics from diverse natural sources, at first without any success in spite of the enormous numbers of compounds and extracts tested.7 Colleagues at Merck, Sharp, and Dohme screened extracts from myxobacteria, having obtained access to the collection of Sorangium strains of Dr. John E. Peterson, a retired botanist and professor at Emporia State University in Kansas, who had worked on those organisms 30 years ago. They soon found the desired activity, and with the small quantities of substance they could recover from plate cultures they elucidated the structure, only to find out that they had rediscovered epothilone. They retained our name, just adding the “e,” and cited our patent. The publication of their results, including the mechanism of action,8 which we had failed to determine, started an avalanche of patents and publications on total synthesis, biosynthesis, genetics, molecular biology, biochemistry, and preclinical and clinical work, which still is increasing in volume at a rapid pace,9 including numerous reviews.10 It soon turned out that epothilone has several properties that could make it superior to Taxol. It acts at very low doses (in the nano- to picomolar range), is active against multidrug-resistant (MDR), including Taxol-resistant, cancer cells, and is more water soluble, so that certain derivatives can even be administered orally. Further, it can be produced by fermentation without difficulty, and as a smaller and simpler molecule, is also accessible by total synthesis, enabling the production of many variants (so far about 1000). The epothilone story following publication of the 1995 Merck article8 is a good example of the often-overlooked fact that basic research and its reduction to practice is far from a straightfor- ward process. When the Merck group discovered our patents, they terminated the project without contacting us regarding possible collaboration. Later, we ourselves tried to interest four major German, one Swiss, and three U.S. pharmaceutical companies in the possibility of collaboration in epothilone development, supplying test material to each company. After over 18 months of consideration, we finally came to an agreement with Bristol-Myers Squibb. The Swiss (Novartis) and one of the German companies (Schering) soon started independent epothilone projects, Schering using total synthesis, and Novartis with material produced by fermentation, using our patent strain, So ce90. By 1997 we had already developed fermentation and separation processes with mutant strains, allowing us to produce 10 g amounts of pure epothilone A and B.10c The yields could later be increased using traditional methods to several hundred milligrams per liter, so that, contrary to Copyright © 2005 CRC Press, LLC 1985 Dr. Reichenbach at GBF isolates Sorangium cellulosum strain So ce90 from a soil sample 1987 The strain is screened positive for antifungal activity, epothilone A and B were isolated, and their structures elucidated by Dr. Gerth and Dr. Bedorf 1991 Immunosuppressive activity studied by Ciba Geigy First publication of structure in Sci. Ann. Rep. of GBF 1992 Application tests and field trials for plant protection by Ciba-Geigy 1994 Work on epothilones terminated at GBF, international patent application abandoned Good activity in NCl 60 cell line antitumor screening 1995 Rediscovery of epothilone A and B as taxol mimics at MSD Disclosure of the absolute configuration as personal communications by Dr. Höfle 1996 First total syntheses by the Danishefsky, Nicolaou, and Schinzer groups 1997 Large scale production at GBF by fermentation Joint semisynthesis program started at BMS and GBF 1999 Clinical trials with natural, synthetic, and semisynthetic epothilones commence 2000 Biosynthesis genes cloned by Novartis and Kosan 2003 Phase II/III clinical trials with epothilone B-lactam (Ixabepilone) FIGURE 21.2 Timeline for epothilone discovery and development. statements in the literature, it was possible to produce all the material needed for development and future applications using Sorangium cellulosum, which can be cultivated without problem on cheap media based on soy meal in industrial bioreactors on a 60–100 m3 scale. The time course of epothilone discovery and development is shown in Figure 21.2. II. NATURAL EPOTHILONES The epothilones were discovered in cultures of the myxobacterium Sorangium cellulosum. Myxo- bacteria are Gram-negative bacteria that move by gliding along surfaces and are notable for their highly developed intercellular communication systems and their ability to reproduce, under unfa- vorable living conditions, in a cooperative action involving hundreds of thousands of cell fruiting bodies of a remarkably sophisticated shape and structure.11,12 The producing organism (Figure 21.3), strain So ce90, is a cellulose degrader and was isolated at the GBF in 1985 from a soil sample collected on the banks of the Zambesi river in southern Africa. In addition to epothilones A and B, the strain synthesized in large scale a family of novel spiroketal polyene polyketides, named spirangiens,1,13 which also showed antifungal traits and cytotoxicity. The epothilones aroused immediate interest because of their strong immunosuppressive effect and good activity against oomycetes — important parasites of agricultural plants.5 Both effects lacked practical applications, however, because of the high toxicity of the compounds. The need for gram amounts of epothilones for relevant tests had already led at that time to the establishment of a production and isolation process at the GBF; however, because of the presence of an excess of spirangiens, the isolation of pure compounds was cumbersome and only achieved by multistep chromatography. The discovery in 1995 that S. cellulosum strain SMP44 also synthesizes epothilones made a second producer strain available.8 This strain is used for practically all studies in the Unites States, with the exception of the BMS/GBF project. Epothilone production with SMP44 was performed in large (150 mm) Petri dishes on agar for 10 d at 28˚C. From 80 plates, 2.7 mg of pure epothilone A and 0.9 mg of epothilone B could be recovered, which was a sufficient amount to elucidate the structures by nuclear magnetic resonance and mass spectroscopy.8 Similarly, the initial yields of epothilones from the GBF strain were rather low. As a first step in yield improvement, strain So ce90 was mutated to eliminate spirangien synthesis, which competed for precursors (the first spirangien negative mutant was obtained in October 1996).14 This was possible only after the strain Copyright © 2005 CRC Press, LLC 10 µm A 1000 µm B 100 µm C FIGURE 21.3 The myxobacterium, Sorangium cellulosum, producer of epothilone. (A) Vegetative cells; (B) vegetative, spreading swarm colony, (C) fruiting bodies, consisting of tiny sporangioles. had been adapted to grow in liquid media in single cell suspension and to develop colonies from individual cells after plating, which are both time-consuming procedures in myxobacteria. A first optimization of the fermentation medium was then undertaken. Cellulose-degrading S. cellulosum is much more versatile metabolically than most myxobac- teria, notably Myxococcus xanthus, and grows on complex as well as on very simple media. Our first production medium consisted of defatted soybean meal, potato starch, glucose, yeast extract, and mineral salts.2 As epothilones are excreted into the broth, a hydrophobic adsorber resin, Amberlite XAD-16, was added during fermentation. The resin increased yields by a factor of five by avoiding feedback inhibition by the end product, and by prevention of its degradation.9,15 The medium was later improved, among others, by adding skim milk powder.16 Production was during log and early stationary phase, with yields of 22 mg/L epothilone A and 11 mg/L epothilone B.2 The generation time was around 16 h, and cell densities of up to 2 × 109 were achieved. After 7 d of fermentation, the resin was harvested by sieving, and the epothilones were extracted with methanol and processed further. Later, the yields could be increased substantially by further optimizing the medium, and especially by a mutation program during which more than 24,000 mutant clones were generated by treatment either with ultraviolet light or nitroso guanidine (NTG) Copyright © 2005 CRC Press, LLC R R 12 O S 13 S 21 N OH HO N OH O O O OH O O OH O Epothilone C 2a R = H Epothilone E 3a R = H Epothilone D 2b R = CH3 Epothilone F 3b R = CH3 FIGURE 21.4 Structures of the minor epothilones. and were characterized by high performance liquid

chromatography analysis.17 Thereby, not only high producers but also clones with defective biosynthetic pathways were obtained,15–17 among them strains that produced epothilone A only, but none that produced pure epothilone B. The same was observed with about 40 further epothilone producers discovered among the 1700 S. cellulosum strains of the GBF collection. The positive strains came from soil samples collected on four continents and represented between 1% and 2.5% of all strains, depending on the area of origin.9 The various strains differed not only in the epothilone pattern but also in the physiology of production. Thus, for example, epothilone synthesis is stimulated by glucose in So ce90 but inhibited in strain So ce1198.9 So far, epothilones have not been found in any other myxobacterial species, nor in any other organism. From high-producing S. cellulosum strains, small amounts of epothilones C (2a), D (2b), E (3a), and F (3b) (Figure 21.4) were isolated, in addition to epothilones A and B.4,18 Whereas in C and D the epoxide is replaced by a cis double bond, an additional hydroxyl group is present at the thiazole methyl group in the variants E and F. Later, both desoxyepothilones 2 and hydroxye- pothilones 3 became available by fermentation of a P450 defective mutant and by biotransformation with S. cellulosum, respectively. This greatly facilitates their isolation, but separation of the homo- logous pairs still requires careful reversed-phase chromatography. From a large-scale fermentation (700 L) of So ce90 mutant B2, 35 variants could be isolated in very small amounts in addition to epothilones A–F.4 Of those, only epothilones C7 and C9 may be regarded as regular biosynthetic products carrying extra hydroxyl groups at C14β and C27. The majority of the other variants result from aberrant biosynthesis. They lack methyl groups, carry extra methyl groups, have an oxazole instead of a thiazole ring and extra double bonds, and have reduced (by two C-atoms) or expanded ring sizes (epothilones K and I). Several small open-chain molecules are obviously derived from biosynthetic intermediates that had escaped from the enzyme complex.4 The epothilones A–F are colorless solids, readily soluble in polar organic solvents and stable at ambient temperature in the pH range of 5–9. Epothilones A, B, and D are crystalline, with melting points in the range of 76˚–128˚C, depending on the solvents of crystallization (the correct melting point of epothilone A is 76˚–78˚C [ethyl acetate/toluene], and 85˚–87˚C [metha- nol/water].3,19–21 Several crystal structures are available for crystals from lipophilic and polar solvents exhibiting different conformations (Figure 21.7; Section V).3,20 Aqueous solubility depends crucially on the presence of epoxide and C12 methyl groups and ranges from 700 mg/L for epothilone A and 200 mg/L for B to 16 mg/L for epothilone D.3,21 Thus, epothilone D is about as poorly water soluble as paclitaxel (6–30 mg/L),22 posing similar problems with its administration at therapeutically useful doses. Nevertheless, because of its promising therapeutic window, epothilone D was identified very early as a candidate for clinical studies by the Danishefsky group.23 Total synthesis was upscaled and production by fermentation optimized. Under favorable conditions, our So ce90 P450 defective mutant produces 70 mg/L of epothilone D. At Kosan Biosciences, the genes for epothilone synthesis were cloned from strain SMP44, first into Streptomyces coelicolor24 and then into another myxobacterium, M. xanthus.25 Both strains, however, produced less than 0.2 mg/L of epothilones. By eliminating the P450 epoxidase gene, Copyright © 2005 CRC Press, LLC 26 O 12 O S S 11 N OH N 2 OH 22 5 17 23 O O 24 O OH O O OH O Epothilone A Epothilone B FIGURE 21.5 Incorporation of acetate, propionate, methionine, and cysteine in epothilones A and B. clones that produced only epothilones C and D (0.45 µg/mL) were obtained.26 Later, the yields could be increased to 23 mg/L and to 90 mg/L26,27 by optimizing the fermentation process. This process requires the addition of methyl oleinate, which complicates work-up; thus, only 63 mg/L of epothilone D could be recovered. Biogenesis of epothilones has been studied in S. cellulosum on three levels: actual biosynthesis using feeding experiments and blocked mutants; analysis of the genes coding for the relevant enzymes; the genes were cloned and sequenced from both producer strains, So ce90 and SMP44; and in vitro synthesis using part of the enzymatic apparatus cloned and expressed in Escherichia coli. By feeding 13C and radioactively labeled precursors to strain So ce90, it was quickly established that the carbon chain is synthesized from acetate, propionate, and cysteine (Figure 21.5; the correct stereochemical assignment of the 22-Me and 23-Me groups is given in references 184 and 185).16 Methyl C21 comes from acetate, methyl C22 from S-adenosyl methionine, and all other methyls arise from propionate. This is also the case with methyl C26 at the epoxide site. The incorporation of propionate at this site in place of acetate is a result of a low specificity of the enzyme in this region. So far it has not been possible to isolate strains or mutants that synthesize exclusively the more active epothilone B, although strains restricted to epothilone A are available.16 Synthesis starts by joining cysteine and acetate to yield methyl thiazole. This requires participation of a nonribosomal peptide synthetase (NRPS). The rest of the molecule is essentially provided by polyketide synthase (PKS) modules.16 As was corroborated by the genetic approach, the whole biosynthetic machinery consists of a huge hybrid multienzyme complex composed of one NRPS and eight PKS modules. The oxygen of the epoxy group is introduced by a post-PKS monooxygenase using molecular oxygen. The PKS module responsible for this section of the molecule has, however, no dehydratase sequence that could produce the double bond required as the substrate of the monooxygenase. As we could isolate two epothilone variants, one with a hydroxyl on C13 and the other with a C12/C13 double bond, which is perhaps produced by a dehydratase (DH) in another module, the substrate of the monooxygenase may arise in this way.17 The dehydration must happen at an early stage because the hydroxy precursors of epothilones C and D could not be detected.17 Increasing amounts of epothilones A or B in the medium inhibit de novo synthesis of those molecules. At the same time, epothilones C and D, present only in traces in the control, rise to very high levels.17 Obviously, the monooxygenase responsible for the intro- duction of the epoxy group is feedback inhibited by epothilones A and B, and epothilones C and D are the end products of the PKS. When epothilones C or D are fed to the nonproducing So ce90 C2 mutant, the compounds are bound immediately to the cells and transformed into epothilones A/B. The monooxygenase is found only in epothilone-producing strains and was never seen in other Sorangium strains.17 It appears that epothilone C has a higher affinity for the monooxygenase than epothilone D, as it is preferentially epoxidized, and even slight and remote changes in the molecule may prevent epoxidation. The block mutant D48 is defective in its monooxydase and is thus a producer of epothilones C and D.17 In cultures of So ce90 without XAD resin, yields of epothilones A/B go down dramatically, and a small proportion of the material becomes hydroxylated on C21 to give epothilones E/F.15 Degradation starts with the opening of the lactone ring by an esterase, which can be demonstrated Copyright © 2005 CRC Press, LLC EPOS A EPOS P EPOS B EOPS C EPOS D EPOS E EPOS F I Module 0 NRPS Module 1 Module 2 Module 3 Module 4 Module 5 Module 6 Module 7 Module 8 P450 EPO A EPO B EPO C EPO D EPO E EPO F EPO K II Loading Module 1 Module 2 Module 3 Module 4 Module 5 Module 6 Module 7 Module 7 Module 9 P450 NRPS SCHEME 21.1 Analysis and assignment of epothilone biosynthesis genes by the Novartis (I) and Kosan groups (II). in the culture supernatant. The hydroxylation is performed by another epothilone A/B–induced monooxygenase, which appears to be a P450 enzyme, for it is inhibited when iconazole is added to the culture.15 A screening of 95 wild strains of S. cellulosum showed that this monooxygenase is widely distributed among nonproducing Sorangium strains. Comparable biosynthetic studies are not available for strain SMP44 and the heterologous host, M. xanthus. For those strains, it was shown that the composition of the medium may have an influence not only on epothilone yields but also on the pattern of variants produced.28,29 In M. xanthus, addition of serine resulted in a substantial increase of epothilones that have an oxazole instead of the thiazole ring. Oxazole epothilones are normally found only in trace amounts, also in So ce90 fermentations.4 Their synthesis is obviously a result of a lack of absolute specificity of the NRPS module taking occasionally a serine instead of a cysteine precursor. The epothilone genes have been cloned and sequenced from both producer strains, So ce9030 and SMP44.31 Both sequences are patented.32,33 As mentioned, the SMP44 genes have been modified and used to produce epothilones heterologously. Good summaries of the organization of the epothilone gene cluster are available30,31,34 (Scheme 21.1), and because of space limitations, details can not be discussed here. Briefly, in SMP44, all seven genes required for epothilone synthesis, epoA to epoF and epoK, are clustered together in a 56–kilobasepair stretch of DNA.31 EpoA is the loading domain, EpoB the NRPS module, and EpoC to EpoF are PKS modules. Overall, there are nine modules plus the loading domain and EpoK, a P450 monooxygenase responsible for the introduction of the epoxy group. All genes are transcribed in one direction and probably in one operon. A methyl transferase domain in module 8 (EpoE) may be responsible for the introduction of one of the methyl groups on C4. When EpoK was expressed in E. coli, the isolated protein could be used to epoxidize epothilone D into epothilone B.31 There is an anomaly in the sequence of PKS module 4: the DH required for the introduction of the C-12/13 double bond of the PKS end product is lacking. The hypothesis that the DH of the neighboring module 5 may be responsible for creating that double bond has recently been proven by destroying that domain.35 Although the exact mechanism of the enzymatic reaction is not yet elucidated, a forward–backward shuffling of the growing chain beween modules 4 and 5 is suggested. Another anomaly is found in module 9 (Epo F), where the DH and ER are inactive31 ([3-13C, 3-18O] labeled epothilones were isolated after feeding of [1-13C, 1- 18O] propionate21). The group sequencing the So ce90 epothilone genes used a slightly different nomenclature for the same general organization of the gene cluster30 (Scheme 21.1). In this case also, transcription is unidirectional. It is suggested that all those genes form one operon of more than 56 kilobasepair. The finished polyketide, epothilone C/D, appears to be released from the PKS by a thioesterase; a sequence showing conserved motifs of that enzyme has been identified next to module 8 in epoE.30 A methyltransferase domain in epoD is required to produce one of the gem methyl groups at C4. The sequence of epoF indicates a cytochrome P450 oxygenase. A short distance in front and behind Copyright © 2005 CRC Press, LLC the epothilone genes, two ORFs, orf3 and orf14, seem to code for transport proteins that may be involved in export of epothilone from the producing cell.30 In summary, the genetic apparatus for epothilone synthesis is remarkably similar, in fact, almost identical, in two producing strains of S. cellulosum, although these are of completely different origin. Among 56,000 basepairs, only 407 (= 0.7%) differ in the two gene clusters, which is an astonishing and very unusual correspondence not yet seen with any other Sorangium compounds (Rolf Müller, personal communication). The epothilone gene cluster from SMP44 was cloned and expressed in heterologous hosts that actually produced epothilones.24,25 The M. xanthus system was also used to make structural variants by manipulating the gene cluster. Thus, by inactivating the monooxygenase gene, strains were obtained that synthesized preferentially epothilone D.25–27 Using site-specific mutation, epoD strains were further modified by inactivating the enoyl reductase domain in module 5 to make 10,11-didehydro-epoD (=Epo490),36 and by inactivating the ketoreductase domain in module 6 to synthesize 9-keto-epothilone D.37a In addition, 4-desmethyl epothilones, molecules with an addi- tional

ether ring, and two fragmentary epothilones were obtained.37a Recently, it was demonstrated that an E. coli strain engineered to express the last half of the epothilone biosynthetic pathway is able to process an advanced precursor to epothilone C.37b In the future, this approach possibly may allow for the exclusive production of epothilone D and its analogs from simple synthetic precursors, thus avoiding tedious separation of C/D homologs. By biotransformation using the oxygen-intro- ducing bacterium, Amycolata autotrophica (formerly Nocardia autotrophica), hydroxylated vari- ants were produced from epothilone D and 10,11-didehydro-epothilone D.38 Several of the analogs created by those different methods are also known from So ce90 fermentations.4 The genes of the first three modules of the cluster, EpoA, EpoB, and EpoC, were expressed in E. coli, the resulting proteins purified, and their enzymatic reactions studied in vitro.34,39–43 Products were usually obtained in minute quantities but could be identified by comparison with synthetic material. The system is of particular interest because it allows us to study PKS/NRPS/PKS interfaces in this hybrid multienzyme complex. It turned out that the individual enzymes of the epothilone cluster show considerable flexibility with respect to their substrate specificity, which could be useful for combinatorial biosynthesis. However, before application, the problem of producing reasonable yields has to be solved. In a recent study, it was demonstrated that the base sequence of the KS domain of epoB in So ce90 (corresponds to epoC in SMP44) differs in several respects from the KS domains of all other epothilone modules.44 Although those, in a phylogenetic analysis, cluster closely together, as is the rule in PKS genes, the epoB KS goes with PKS modules of other organisms. Also, the epoB KS deviates significantly in its GC composition. This was taken as a hint that the epoB KS sequence may have been introduced into a rudimentary epothilone gene sequence by horizontal gene transfer. Also recently, the exact position of epothilone bound to a protein could be analyzed for the first time.45 The crystal structure of the P450epoK monooxygenase was analyzed in the free form and again with the substrate and product bound. The enzyme has a very large binding cavity, but otherwise it resembles other P450 enzymes, particularly P450eryF from the erythromycin cluster (30% identity, 48% similarity). III. MECHANISM OF ACTION The biological and biochemical effects of the epothilones have been studied in many laboratories, applying diverse experimental methods, in the United States and in Europe, and results are not always strictly comparable. Clearly, individual cell lines and the various epothilone analogs behave quite differently. Nevertheless, the overall picture is fairly consistent, and as a rule, safe generali- zations can be made. In the ensuing discussion, frequent reference is made to synthetic and structure- activity studies discussed in detail in Sections IV and V, respectively. Epothilone does not act on bacteria, and among the filamentous fungi and yeasts tested, only Mucor hiemalis (Zygomycetes) responded (minimal inhibitory concentration 20 µg/mL).2 Copyright © 2005 CRC Press, LLC In greenhouse experiments, various oomycetes were also inhibited. The cytotoxicity of the epothilones in animal cell cultures was noted immediately after their discovery,1,2,5 but the effects they produce on the tubulin cytoskeleton were only realized much later.8 In the beginning, only the two main fermentation products, epothilones A and B, were available for study. Although epothilone B differs only in an additional methyl group on C13, its efficacy is consistently higher by a factor of 5–10.4,8,10e,m,46–51 This difference also remains with analogs of the two (e.g., epothilones C and D4,17,46,49,52 or the oxazole derivatives4,53,54) and is even seen in binding and polymerization experiments with tubulin.47 The epothilones act on a wide variety of cell types, including many human tumor cell lines. The concentration range for cytotoxicity is IC50 = 0.3–4 ng/mL10n or 1.3–39 nM8,10m,46–48 for epothilone A, and 0.1–1.5 ng/mL4 or 0.06–40 nM for epothilone B.8,10m,46–48 The figures for Taxol are 0.2–95 nM.8,46,47 (The molecular weights are Taxol, 854; epothilone A, 494; epothilone B, 508; i.e., Taxol is 1.73 and 1.68 times larger than epothilone A and B, respectively). Efficacy of Taxol equals that of epothilone A and is clearly less than that of epothilone B.8,46,47 For instance, with 0.05–0.06 vs. 0.22–0.6 nM, epothilone B is 6–10-fold more active than Taxol against human prostrate cancer cells.52 It came as a big surprise and is of considerable practical relevance that the epothilones still act on Taxol- resistant and MDR cells in general.8,47 Although the resistant cell lines tolerated 100- to several thousand–fold higher concentrations of the selective agent, such as Taxol, vinblastine, or adriamycin, resistance to epothilones rose only by a factor of 1.5–10, and rarely higher. For epothilones C and D, it was even found to decrease in some instances.46 It was reported that epothilone B–lactam and epothilone C–lactam lose activity on drug-resistant cell lines,48,55,56 but this was not corroborated by another study, using several drug-resistant human cancer cell lines.57 In vivo studies also did not support that statement. Epothilone B–lactam indeed shows about the same activity as epothilone A. Resistance to epothilone itself will be discussed below. The epothilones promote tubulin polymerization and stabilize microtubules (MTs), as does Taxol. The morphological changes produced in treated cells resemble those effected by Taxol. In interphase cells, massive bundles of MTs arise in the cytoplasm, apparently randomly and inde- pendent from centrosomes, originating from multiple nucleation centers.8,47,58,59 This transformation can already be seen 2 h after application, and after 24 h, virtually all MTs are in bundles (epothilone B).47 When cells are kept at 4˚C, their MTs decay completely within 1 h, but in the presence of 25 µM epothilone A or B or Taxol, they remain stable.8 Already, 2 h after the addition of epothilones, mitotic defects can be observed, reaching a maximum after 6 h.47 At concentrations of 10 µM, normal spindles are no longer seen.47 The organization of MTs in mitotic cells differs from that in interphase cells, so that a considerable turnover of tubulin still must occur in spite of the MT-stabilizing effect of the drug.8,47 An explanation could be that before mitosis, MT-destabilizing proteins, such as katanin or stathmin/op18, are activated, or that MTs are stabilized less rigorously, perhaps because MT-associated proteins (MAPs) have dissociated after having become phosphorylated.60 Multipolar spindles arise in a concentration range between 5 nM and 1 µM.8 The spindles produced in the presence of epothilone A, epothilone B, and Taxol seem to differ somewhat in their morphology.47 The result of those mitotic aberrations is an arrest of the cell cycle at the G2–M stage (e.g., with 100 nM epothilone A or B after 72 h).8,47,52,57 The DNA content is then 4n.52 The mitotic blockade is complete after 8 h (at the ID50, 7.5 nM of epothilone B–lactam,57 or in another system, after 20 h at 3.5 nM epothilone B60). Mitotic arrest and cytotoxicity are seen at much lower doses than aberrations in interphase MTs (e.g., at 3–30 nM vs. 6–40 nM vs. more than 1 µM, respectively).8 A certain percentage of cells seem to pass through aberrant mitosis and become arrested in the following G1 phase as multimininucleate cells.8 With epothilone B at its IC50, 38% of cells become multimininucleate, and with epothilone A and Taxol, 10%–15% do so. Delayed mitosis takes place in those cells 24–30 h after the addition of the drug, and although they die, too, they do so after a longer time. The phenomenon is preceded by an aberrant chromosome alignment in metaphase and is probably related to the apoptotic fragmentation of nuclei discussed below. In MCT7 cell cultures, 2.5% of control cells are in Copyright © 2005 CRC Press, LLC mitosis.60 When epothilone B is added, many cells enter mitosis but do not progress beyond metaphase. The ratio of anaphase to metaphase cells in the control was 0.18; with 1 nM epothilone B, 0.03; and at 3 nM, 0. Many of the cells that were not blocked in mitosis became multinucleate, probably because of an abortive mitosis; the behavior of cells exposed to epothilone B at a dose too low to lead to mitotic arrest is of considerable interest. In this case, multipolar spindles are induced, resulting in aneuploid cells which also are not viable.61 The epothilones lead to apoptosis, as indicated by DNA ladders. The DNA is nicked specifically in G2–M blocked cells.8 Further, fragmentation of the cell nuclei is seen. This can be monitored automatically by an image analyzer.62 With L926 mouse fibroblasts, fragmenta- tion of nuclei can be seen already after 1 d and reaches its maximum after 3 d (100 ng/mL epothilone B). Fragmentation increases linearly with the logarithm of concentration between 1 and 1000 ng/mL epothilone A.62 Treated cells appear to remain vital for some time, at least the reduction of tetrazolium salts continues in mouse macrophages unrestricted for 48 h (epothilone B, 0.02–20 ng/mL).59 A comparison of epothilones B and D showed that total uptake of epothilone B is higher by a factor of 4.63 In addition, the intracellular distribution of the two compounds differs dramatically. Although nearly all of epothilone D is found in the protein fraction of the cytosol, almost 50% of the epothilone B resides in the nucleus.63 This different behavior was corroborated by a more detailed study using tritiated 6-propyl-epothilones B and D (pEpts B and D), which show almost the same inhibitory effects as the natural 6-methyl analogs.64 The study compared two cell lines, A431 (human epidermoid) and NCI/Adr (human mammary carcinoma, Taxol resistant up to more than 1µM). Epothilone D and pEpt D acted much more slowly than epothilone B and pEpt B, which produced almost maximal and lasting growth inhibition after only 4 h exposure in A431 cells. Drug uptake and intracellular pools turned out to be dose dependent. Efflux of pEpt D was faster (almost complete after 60 min) than that of pEpt B, 60%–70% of which was still inside the cells after 4 h when the loading dose was high (70 nM). When it was low (3.5 nM), NCI/Adr, in contrast to A431 cells, lost pEpt B quickly. This demonstrates that the two cell lines have different transport mechanisms for the two analogs. Apparently, the export pump of NCI/Adr cells becomes blocked when pEpt B is present in high concentrations. In both cell lines there seems to exist a MDR1-independent, yet verapamil-sensitive, export. Verapamil, a Ca antagonist and inhibitor of the MDR1 P-glycoprotein (Pgp) shuttle, stimulates uptake of both drugs at low drug concentrations, but at high concentrations stimulates uptake only of pEpt D. Intracellular pools may reach up to 100 pmol/106 cells. Drug distribution was the same in both cell lines: pEpt D (and Taxol) was mainly in the soluble part of the cytosol (80%), and only little was in the nucleus (5%–15%), whereas half of the pEpt B was in the nucleus (40%–50% vs. 40%–60% in the cytosol); there it was located on nuclear proteins, presumably tubulin, and nothing was on the DNA. This distribution may contribute to the differences in drug transport and efficacy of the two analogs. The influence of epothilone B on MT dynamics was directly measured on time-lapse photographs of MTC7 cells that were stably transfected with green-fluorescent- protein α-tubulin.60 Cells were analyzed 6 h after the addition of the drug, when the intracellular equilibrium was reached. The growth rate of MTs in control cells was 13.3 µm/min, the shortening rate 21 µm/min, and dynamicity (length of growth + length of shortening during total observation period per minute) 10.4 µm/min. With 2 nM epothilone B, those figures were reduced to 8.3, 19.3, and 6.3 µm/min, and with 3.6 nM, to 7.5, 14.9, and 3.9 µm/min. Thus, while under all conditions MTs shortened twice as fast as they grew, epothilone B clearly reduced MT dynamics. Also, at 3.5 nM most MTs were completely stabile, although their ends showed minute (below 5 µm/min) changes in length (they “chatted”). The few MTs that still were dynamic grew and shortened more slowly and for shorter lengths of time; also, pause times, at which no changes occurred, became longer. The concentration and timescales of the MT effects correlate closely with those of mitotic arrest, and it was concluded that reduction of MT dynamics is the direct cause of mitotic arrest. Copyright © 2005 CRC Press, LLC Spindle MTs may be particularly sensitive to interference with their dynamics

because they have to grow for a longer time and to greater length than interphase MTs. Essentially the same reactions as just described for epothilone B were seen with Taxol, only at somewhat higher concentrations. However similar the reactions of cells to Taxol and epothilone are, they still are not identical. An example is the behavior of mouse macrophages (in vitro).59 At very high doses (0.1–10 µg/mL) Taxol stimulates the macrophages to produce endotoxin-like effects; for example, to synthesize nitric oxide (Salmonella typhimurium lipopolysaccharide does so at 1–10 ng/mL). Epothilone B (up to 10 µg/ml) neither induces macrophages to release nitric oxide nor prevents it when they are treated with lipopolysaccharide, yet epothilone B acts normally on MTs inside the cell (at 12.5 ng/mL). Also, the degrees of resistance of normal cells to epothilone and Taxol do not always match.65 Studies on the effects of epothilones on tubulin in vitro showed that they make tubulin poly- merize under conditions not applicable in the absence of the drug; for instance, in the absence of GTP8,47 or MAPs or both, and in the presence of Ca++ (3 mM)8 and at low temperatures (0°C).47 They stabilize existing MTs at cold temperatures (4°C for 30 min,8 0° for at least 1 h47), with epothilone B doing so even after the addition of 5 mM CaCl2 (not, however, epothilone A or Taxol).47 Although initial studies suggested that epothilone A is more effective than epothilone B,8 this was later corrected when purified tubulin was used.47 The observed effects resemble those produced by Taxol, and indeed it was shown that epothilones displace [3H]-Taxol from polymerized tubulin8,47 (apparent Ki values, depending on the method of data analysis: epothilone A, 1.4/0.6 µM; epothilone B, 0.7/0.4 µM; docetaxel, 1.2 µM).47 This indicates both that the epothilones dock to the binding site of Taxol (i.e., β-tubulin), although initially allosteric effects or overlapping binding sites could not be excluded, and that epothilone B binds more tightly than epothilone A and Taxol. More thorough studies showed that epothilone A is close to Taxol in efficacy while there are still subtle differences.47 Thus, for example, in experiments without GTP but with MAPs, epothilone A causes better polymerization than Taxol (both epothilones allow assembly at 10°C, maximum at 25°C, whereas Taxol produces no MTs below 25°C, maximum at 37°C), or with GTP but without MAPs at 25°C, Taxol is more active than epothilone A (epothilone B is already near its maximum) and reaches its maximum at 37°C, when it is twice as active as epothilone A. In the absence of both MAPs and GTP, significant polymerization takes place with all three drugs, only at higher concen- trations (i.e., more than 10 µM per 10 µM tubulin, which was the ratio in the former experiments). Polymerization in the absence of GTP is somewhat enigmatic. There are two GTPs bound to the tubulin dimer: one on α-tubulin at the α , β-interface, which is neither exchangeable nor hydrolyzable, and one near the carboxy end of β-tubulin that is exchangeable and becomes hydrolyzed to GDP soon after the addition of the heterodimer to the plus end of the growing MT (which is capped with GTP–tubulin). It appears that the GTP hydrolysis changes the shape of the dimer from straight to (in the free form) bean shaped.66 This in turn may introduce some stress in the MT, produce the helical arrangement of the protofilaments, and may be a prerequisite for decay of the MT. Epothilone also reduces the number of protofilaments in MTs formed in vitro from 55% with 14 (control) to 67% with 13 (20 µM epothilone B, 10 µM tubulin).66 The idea that the epothilones stimulate nucleation of tubulin polymerization is supported by two further experiments.47 The critical concentration of tubulin, [T]cr (i.e., the lowest concen- tration at which polymerization begins under standard conditions), is regarded to be proportional to nucleation. The variable [T]cr was, without drug, 3 µM and in the presence of any one of the three drugs, less than 1 µM. In the presence of only MAPs or only GTP, [T]cr, with Taxol or epothilone A, was 1.4–1.7/4.3–4.5 µM, and with epothilone B, 0.8/2 µM. Without both MAPs and GTP, [T]cr, with Taxol, was 22 µM, with epothilone A was 19 µM, and with epothilone B was 7.4 µM. In addition, the average length of the MTs formed in the presence of the three drugs was measured in electronmicroscopic pictures. With MAPs and GTP but without drugs, the MTs were 4.3 µm long, with epothilones A or B, they were 2.8 and 1.1 µm, and with Taxol, they were 1.9 µm (i.e., shorter than with epothilone A). When GTP was lacking, the MTs became longer, and without MAPs even more so, yet with epothilone B, they were always half as long as with the Copyright © 2005 CRC Press, LLC other drugs. The morphology of the MTs was the same with or without drug, viz, tubules with occasional ribbon-like open sections. In summary, the experiments just described demonstrate that the epothilones bind to, or close to, the Taxol site on β-tubulin, that they (also) bind to polymerized tubulin, that they seem to stimulate nucleation, and that MAPs are not required for drug binding. There is an obvious discrepancy between the concentrations effective in producing cytotoxicity and mitotic arrest (lower nanomole range) and those interfering with MT turnover in vitro (micromole range). This may be explained by a several hundred–fold accumulation of epothilones inside the cells.63,67 When HeLa cells, containing about 25 µM tubulin, were exposed to 10 nM epothilone, the drug reached intracellular saturation levels after 2 h. The cells then contained 4.22 µM epothilone A, or 2.55 µM epothilone B, respectively. Exposure to 100 nM drug reversed the relative internal levels to 17 µM epothilone A or 26 µM epothilone B.67 Thus, even at relatively low doses, both drugs reached intracellular concentrations sufficient to disturb MT dynamicity, which may lead to a lasting mitotic arrest after a short exposure of the cells to the drug. In the last few years a much better understanding of the molecular interaction between Taxol/epothilone and β-tubulin has been gained. The yeast Saccharomyces cerevisiae is not inhibited by epothilone B (up to 150 µM), which is likely a result of a lack of uptake of the drug,68,69 as purified yeast tubulin is induced by epothilone B to polymerize in vitro (EC50 = 1.3 µM in the presence of 5 µM tubulin and 50 µM GTP; with bovine brain tubulin, the EC50 is 1.2 µM).69 Interestingly, Taxol has no such effect. Epothilone B reduces the dynamicity of yeast tubulin by a factor of nine.69 In the presence of epothilone B, yeast tubulin also assembles without GTP. Yeast contains only one β-tubulin gene, and its β-tubulin differs in 124 amino acids from mammalian brain tubulin. With 75% similarity, it is close enough to coassemble into MTs. By analyzing available data on Taxol binding to bovine brain β-tubulin, five amino acids that differed in yeast tubulin were assumed to be responsible for preventing Taxol binding. Those amino acids were exchanged by site-specific mutation, viz, Ala19 to Lys19, Thr23 to Val23, Gly26 to Asp26, Asn227 to His227, and Tyr270 to Phe270.68 The mutated yeast tubulin now polymerized with Taxol (EC50 1.55 µM, at 5 µM tubulin under conditions that do not allow spontaneous assembly) even more efficiently than bovine brain tubulin, which is perhaps a consequence of the 5%–10% lower critical concentration of yeast tubulin. The EC50 of epothilone B remained the same (1.45 µM) with yeast wild type and mutant tubulin. The molar binding ratio of Taxol:tubulin for wild-type yeast tubulin was 0.13, and for mutant tubulin it was 1.01, which was identical to that for bovine brain tubulin.68 Radioactive Taxol bound to mutant MTs could be replaced almost completely by a 10-fold excess of epothilone B.68 Clearly, the five inserted amino acids are not required for epothilone binding but are essential for interaction of Taxol with β-tubulin. This shows that the binding site of epothilone need not be exactly identical with that of Taxol. The unaltered EC50 of epothilone B for the mutant tubulin seems to indicate that epothilone indeed binds at a different location to Taxol. Efforts have been made to identify the binding site of epothilone by mapping point mutations conferring resistance. The studies concentrated on the highly variable carboxy end of the otherwise conserved 450 amino acids of α- and β-tubulin, because this is the location of the nucleotide and MAP binding sites. Also, this part of the molecule is important for MT stability, because it is responsible for lateral contact of the protofilaments. In addition, most residues involved in Taxol binding have been located in this area. The mutants showed an increase in epothilone tolerance by a factor of 20 to 470.70–73 Several of them had at the same time become hypersensitive to MT- destabilizing drugs (colchicine, vinblastine) or become outright epothilone dependent.71,73 This was explained by a destabilizing effect of the mutation that is counteracted by the stabilizing activity of epothilone (or Taxol).74 The drug-resistant cell lines grow more slowly than the parent lines.71 Expression of the MDR1 (Pgp) and the MPR (multidrug-resistance associated protein, another membrane protein) export systems are not increased71,73 or even absent.72 Also, there is no reduction of intracellular [3H]- Taxol accumulation.73 All this means that resistance is most likely caused by reduced epothilone Copyright © 2005 CRC Press, LLC TABLE 21.1 Amino Acids in β-Tubulin Suspected to Be Involved in Epothilone Binding Method Responsible Amino Acids Isotype Reference Resistant mutants Thr274 to Ile, Arg282 to Gln 70 Resistant mutants Pro173 to Ala, Gln292 to Lys, Tyr422 to Tyr/Lys βI 71 (heterozyg.) Resistant mutants Pro173 to Pro/Ala, Tyr422 to Tyr/Lys βI 75 Resistant mutants Ala231 to Thr, Gln292 to Glu βI 73 Predicted from epothilone binding Leu217, His229, Thr276, Arg284 45 to P450epoK Not required in yeast tubulin Lys19, Val23, Asp26, His227, Phe270 68 (in contrast to Taxol) binding. This has been shown to be the case by determining the percentage of tubulin polymerized; the tubulin decreased with increasing resistance from 28% (parental) to 6.9%.73 In one cell line selected against 300 nM epothilone D, the mutated tubulin did not bind epothilone to any extent and also could not be stimulated to polymerize by the drug.73 As a rule, epothilone resistance goes parallel to Taxol resistance,73 but sometimes considerable discrepancies were observed — and even more so with respect to docetaxel.70,71,73 Usually, only one point mutation was discovered in resistant cell lines, but occasionally two could arise.73 Most mutations were expressed on the protein level.73 The identified amino acid exchanges were usually close to the M loop and on helices H7 and H9. Thus, the mutated residues were indeed near the Taxol binding site. Thr274 appears to be particularly critical for epothilone binding; it lies exactly at the site of interaction between β-tubulin and the C7–OH of Taxol.72 However, the mapped mutations concerned a considerable variety of residues (see Table 21.1). This may be because epothilone resistance is a more complex phenomenon than it may seem at first look. For one thing, resistance may vary with different cell types, among others, because they vary in their β-tubulin isotype composition, including posttranslational modifications. Only recently it became possible to determine the exact isotype profile of an experimental cell line by liquid chromatography combined with electrospray ionization mass spectrometry.75 This permits assurance that the isotypes present in the genome are really expressed under the conditions of the experiment, which is not necessarily the case.76 The isotypes are encoded by different genes, and the various genes are preferentially expressed in specific cell types. They are more than 80% homologous and vary mainly at their carboxy ends. The isotypes differ in conformation, assembly, dynamics, and ligand binding.63 Among the seven known isotypes of β-tubulin, I appears to be the preferred target of epothilone,71,73,75 and II that of Taxol.63 In HeLa cells, the prominent isotypes of β-tubulin were I and IVb. Two epothilone-resistant sublines contained the same isotypes with a separate heterozygous mutation, each in I. It could be shown that the mutant alleles were expressed, and apparently at a higher level than the wild type.75 Altered expression of isotypes74 and changes in posttranslational modification of β-tubulin,75 which could influence binding of MAPs, have

to be considered as possibilities in producing resistance, although this has not yet been shown for epothilone. Nor can one be sure that tubulin mutations that lead to drug resistance are located in the β-isomer. In a Taxol-resistant, Taxol- dependent A549 cell line, the β1-tubulin was unaltered, and Taxol still bound to it, but there was a Ser379 to Arg exchange in the Kα1 isotype of β-tubulin.74 Interestingly, there was an increased expression of βIII-tubulin, which also seemed to contribute to Taxol resistance, apparently by changing MT dynamicity.74,75 A major contribution to resistance may come from altered binding of MAPs to tubulin. MAPs may stabilize or destabilize MTs, and mutations in tubulin could influence the interaction between the two proteins, but of course, there could also be mutations in the MAPs themselves; in the Copyright © 2005 CRC Press, LLC enzymes that activate or deactivate them, usually kinases and phosphatases; or in the systems that control the expression of MAPs. There is experimental evidence that those assumptions are not mere theory. In the Taxol-resistant (and Taxol-dependent) cell lines, A549-T12 and A549-T24, a rise in MT dynamicity is seen in the absence of Taxol.74,75 In this case, the responsible mutation is in the α-tubulin. The point mutation is near the carboxy end and presumably at a site of MAP4 and stathmin interaction. In response to rising Taxol concentrations (2 nM is required for survival), the status of the two regulatory proteins is also changed. Whereas the active, hyperphosphorylated form of stathmin (which destabilizes MTs) was barely detectable, the intracellular level of the nonphos- phorylated, inactive form increased twofold, and the phosphorylated, and therefore inactive, form of MAP4 (which is a MT stabilizer) had clearly risen.74 In another study it was shown that tubulin- interactive drugs, like epothilone A, lead to phosphorylation of MAP4, probably by a mitotic kinase.76 The phosphorylated MAP4 dissociates from the MTs, which then become disorganized. Drug-resistant cells do not phosphorylate MAP4, and their MTs remain intact with MAP4 associated. At very high doses, however, the same events as in sensitive cells take place, which shows that the interaction between tubulin and MAP4 is undisturbed. Also, expression of MAP4 is not influenced by the drugs in sensitive and resistant cells. Thus, it appears that MAP4 takes part in the drug-induced killing process, and resistance is somehow connected with the failure to phosphorylate the molecule. Although the exact mechanism of this sequence of events is not yet understood, it is obvious that the binding of epothilone to tubulin activates downstream events that contribute to cytotoxicity. In a more general approach, genes were identified by analyzing cDNA microarrays, which were overexpressed in epothilone A–resistant cell lines.65,77,78 The STAT1 gene, interferon-inducible genes, and a MT-associated GTPase were found to be upregulated. The latter contributes perhaps to resistance by destabilizing MTs in the resistant cells. A recent study, also using cDNA microarrays, demonstrated that 41 genes had an altered expression after epothilone B (10 nM) treatment of cells, and that many of them were connected with the TNF (tumor necrosis factor) stress response pathway.77 Another important determinant in epothilone resistance could be the regulation of transport systems (i.e., of the membrane proteins involved in uptake and thus intracellular accumulation of the drug) and of the MDR1 and MPR machinery responsible for export. The latter are of particular interest in connection with epothilone, because they appear to be much less active on epothilone than on other cytotoxic drugs, including Taxol. Studies on epothilone resistance repeatedly demonstrated that intracellular accumulation of tritium-labeled Taxol was unimpaired.73 Thus, in those cases, resistance could not be attributed to impeded uptake of the drug. Similar experiments have been done with tritium-labeled epothilone analogs.63,64 [3H]-6-Propyl epothilones B and D were used to study uptake by nonresistant A431 (human epidermoid) and doxorubicin (Adriamycin)-resistant NCI/Adr tumor cells (which also are resistant to epothilones).63 The resulting picture turned out to be rather complex, as pointed out already. In several epothilone B–resistant A549 and HeLa cell lines, the MDR1 gene was not expressed at all, and although the MRP system was expressed, its level of expression did not increase after the cells became epothilone resistant.71 Also, in epothilone D–resistant leukemia cell lines, no rise in the levels of the MDR1 and MRP systems was seen, which could have explained resistance.73 In fact, it already had been discovered that epothilones are not substrates for Pgp.8,47 A human neuroblastoma cell line that was constitutive for MDR1 still responded to epothilone A, but not to Taxol.58 The level of Pgp even rose during drug treatment (i.e., MDR1 was induced by epothilone A). In contrast to epothilone, calcein was exported from the cells ever more efficiently the higher the Pgp levels became, which proved that the transport system was fully functional. Thus, it seems that the cellular export systems do not play a major role in epothilone resistance, nor is there evidence that impaired uptake may be a decisive factor. How does epothilone kill cells? All studies dealing with the killing mechanism of epothilone agree that the drug binds to MTs within cells, which leads to mitotic arrest by disorganization of the normal turnover of tubulin, to an accumulation of the cell population in the G2–M phase of the cell cycle, to a decay of the cell nuclei, and to apoptosis. However, there still remains to be Copyright © 2005 CRC Press, LLC explained how exactly apoptosis is achieved, and whether apoptosis is the only way by which epothilones become cytotoxic. The answer is of more than academic interest, as it may have implications for the treatment of cancer. Thus far, no fully consistent picture can be presented. Among other reasons, this may be because investigators dealing with that problem used different cell types and various epothilone analogs. It appears that the tumor suppressor protein, p53, does not play a role in the killing of cells by epothilone, as human neuroblastoma cell lines, which constitutively sequester p53 in the cytoplasm and thus prevent its effect as an intranuclear transcription factor, still were fully sensitive to epothilone A (and Taxol).58 In contrast, it was shown that, although cells with mutated p53 normally become less sensitive to cytostatic compounds, their sensitivity to epothilone B (and Taxol) rises, perhaps because a higher proportion of cells with mutant p53 is in mitosis — the target of epothilone.79 Apoptosis is a rather complex phenomenon induced by a receptor-mediated extrinsic or a mitochondrial pathway, and it is modulated by a large number of proteins. Very briefly (for more details see, e.g., refs. 80 and 81), the extrinsic pathway starts with death receptors (e.g., DR4 and DR5) in the cell membrane. They are activated by binding a death ligand (e.g., Apo2- 2L/TRAIL=TNF-related apoptosis-inducing ligand) and then, after several more steps, form a death- initiating signaling complex, which results in autocatalytic activation of initiator caspase 8, and the following effector caspase cascade finally triggers apoptosis. The mitochondrial pathway begins with release of cytochrome c from mitochondria into the cytosol. This, which again involves several steps, leads to the formation of apoptosomes (consisting of oligomers of APAF-1 [apoptotic protease-activating factor 1]), which activate initiator caspase 9 and, in turn, executioner caspase 3 (and caspase 7). Those caspases cleave a number of cellular proteins and finally lead to apoptosis. The sequence is modulated by the Bcl-2 and IAP protein families. One study applying (water-soluble) epothilone B–amine (BMS 310705) on ovarian carcinoma cells taken from a patient who no longer responded to treatment with platinum and Taxol demonstrated that the drug killed the cells via the mitochondrial pathway of apoptosis.80 The cells were exposed for 1 h to 10–500 nM of the drug and then incubated in drug-free medium. After 24 h, more than 25% of cells showed apoptosis, and after 96 h, 90% were dead. An increase in activity of initiator caspase 9 and effector caspase 3, but not of initiator caspase 8 (typical for the extrinsic pathway) was seen. Also, 12 h after treatment, cytochrome c was detected in the cytosol. Another study using epothilone B–lactam on cell cycle–synchronized human ovarian carcinoma cells found indicators of the mitochondrial pathway (leakage of cytochrome c into the cytosol, activation of caspase 3, and downregulation of the protective proteins survivin, cIAP1, and XIAP, which inhibit caspase 3, caspase 7, and caspase 9), but also a rise of the death receptors DR4 and DR5 in the cytoplasmic membrane, which sensitized the cells to the death ligand, Apo2-2L/TRAIL.81 Thus, in this case the extrinsic pathway seemed also to be in operation. The study also showed that epothilone B–lactam is active on Taxol-resistant cells. Still another study showed a role of the Bcl-2 family of apoptosis-modulating proteins in the killing process.82 Again, epothilone B–lactam was used. The drug induced a conformational change in the proapoptotic Bax protein, which resulted in its transfer from the cytosol into mitochondria and a concomitant release of cytochrome c from mitochondria into the cytosol, with the conse- quences pointed out above. Apoptosis could be mitigated by overexpression of Bcl-2 and enhanced by the Bcl-2 antagonist, BAK-BH3 peptide. The conclusion was that epothilone B–lactam induces apoptosis by a Bcl-2-suppressible pathway, and that treatment of breast cancer can perhaps be improved by a combined therapy with epothilone B–lactam and a Bcl-2 antagonist. Involvement of the Bcl-2 family was also seen when human ovarian carcinoma cells were treated with epothilone A.76 The epothilone A–tubulin interaction, which must be unimpaired, results in inactivation of the prosurvival proteins Bcl-2 and Bcl-xL by phosphorylation, and in lowering of the Mcl-1 level by increased turnover. As Mcl-1 inactivates the proapoptotic Bax protein by forming a complex with it, reduced Mcl-1 and a concomitant overexpression of Bax clearly would support apoptosis. Typically, in resistant cells Mcl-1 remains high in the presence of epothilone A, Bax is lowered, and Bcl-xL is not phosphorylated. Copyright © 2005 CRC Press, LLC In contrast to the studies just discussed, the authors of an investigation of the effects of epothilone B on non–small cell lung cancer cells concluded that epothilone kills the cells by an as-yet-unknown, caspase-independent mechanism.83 At low doses, cytotoxic effects (disruption of the MT cytoskeleton, arrest of cells in the G2–M phase, aberrant mitosis, multinucleated cells) are seen rather soon, before 24 h after the addition of the drug, yet symptoms of apoptosis [nuclear condensation, apoptotic bodies, cytochrome c release, cleavage of poly (ADP-ribose) polymerase, activation of caspase 3 and caspase 9] appear only much later, and therefore may be just side effects. Furthermore, the cells would still die when the mitochondrial or the extrinsic pathway of apoptosis were neutralized by overexpression of the antiapoptotic proteins, Bcl-2 or Bcl-xL, or of the dominant negative FAS-associated death domain. Even the caspases could be blocked by natural (response modifier A) or artificial (Z-Val-Ala-Asp-fluoromethylketone) antagonists without inter- fering with cell death. If this hypothesis is correct, drugs may be found that induce that mechanism more efficiently and, thus, may be even more beneficial in cancer therapy. Several studies report synergistic effects between epothilone and other drugs, which may become of interest in cancer therapy. Many solid tumors show interstitial hypertension, which impedes transfer of drugs from capillaries to tumor tissue. This high interstitial fluid pressure appears to be mediated by the activity of platelet-derived growth factor, the receptors of which are expressed in many solid tumors. The tyrosine kinase inhibitor, STI571 (Novartis Pharma), blocks the kinase activity of those receptors, which results in lowered interstitial fluid pressure and improves uptake of drugs by the tumor. When severe combined immunodeficiency (SCID) mice with human anaplastic thyroid carcinoma were treated with a combination of epothilone B and STI571, the tumors became more than 40% smaller than after treatment with epothilone B alone.84 The level of epothilone B in the tumor increased threefold but remained unchanged in liver, kidney, and the intestinal tract. However, STI571 (100–150 mg/kg) had to be administered on three consecutive days at a minimum, and its effects faded away 2 d after the last application. The synthetic flavonoid, flavopiridol, has aroused considerable interest in recent years on account of its antitumor activity. It arrests cells in the G1–S (or G2–M) phase of the cell cycle and induces apoptosis. The mechanism appears to be inhibition of mRNA synthesis, perhaps by blocking transcriptional elongation or by binding to duplex

DNA. In consequence of this, cyclins and antiapoptotic members of the IAP and Bcl-2 families of proteins are downregulated.85 Further, flavopiridol blocks cyclin-dependent kinases and brings about conformational changes of Bax, with the sequels outlined above. In human breast cancer cells, flavopiridol (100–500 nM) acted syner- gistically with epothilone B (20 nM) in inducing apoptosis, but only when applied after, and not in advance of, that drug.85 In a Bcl-2 overexpressing cell line, epothilone B and flavopiridol, alone or in sequence, induced only minor apoptosis, which demonstrates again that Bcl-2 acts as an epothilone antagonist. Specific farnesyl transferase inhibitors are potent antitumor drugs. The explanation is that they prevent posttranslational farnesylation of certain proteins (e.g., of the p21ras family), which is a prerequisite of their insertion into the cell membrane. With human breast cancer cells, it was shown that those inhibitors (e.g., L-744,822; Merck) increase sensitivity to antimitotic drugs, such as Taxol and epothilone.86 The synergistic effect was additive. It is speculated that this happens because a farnesylated protein controls the checkpoint of mitosis. Also, human prostate cancer cells, including a line that was rather resistant to L-744,822, became considerably more sensitive to the drug when they were pretreated with epothilones.52 This was regarded as a promising strategy for therapy of advanced prostate cancer. A broad synergistic effect was seen in colon cancer cells treated with a combination of epothilone D and 5-fluorouracil.87 Histone-deacetylase inhibitors, like LAQ824 (Novartis), lead to altered gene expression, with changes in the levels of several important regulatory proteins. As a result they enhance epothilone B–induced apoptosis in human breast cancer cells.88 Epothilone B–lactam sensitizes human lung cancer cells to radiation, in vitro as well as in vivo (athymic nude mice).89 In vitro, the induction of apoptosis by drug and radiation was additive, and a Copyright © 2005 CRC Press, LLC higher dose of the drug for a shorter exposure time was more efficient than a lower dose for a longer time. The in vivo effect is considered to be a result of a combination of mitotic arrest and reoxygenation of the tumor, so that the timing of the irradiation after drug exposure becomes a critical factor. IV. TOTAL SYNTHESIS AND SEMISYNTHESIS A. TOTAL SYNTHESIS The discovery of the epothilones as the first Taxol mimics, and reports of their promising pharma- cological properties, triggered enormous worldwide synthetic activities in 1995. With seven stereo- genic centers in a 16-membered macrocycle, their total synthesis appeared, though challenging, to be far less difficult than that of Taxol. Thus, even smaller groups entered the field and made epothilone probably the most often synthesized natural product in recent years. Apart from preliminary synthetic studies, more focused work started in late 1995, when the absolute configuration became available from our lab (before publication in July 1996, the absolute configuration of epothilone A and B had been communicated to interested scientists from October 1995 on). First total syntheses of epothilone A and, later, of epothilone B were reported by the groups of Danishefsky, Nicolaou, and Schinzer in 1996 and 1997.10a Key bond-forming steps in these early syntheses were Yamaguchi macrolac- tonization, a highly diastereoselective C6–C7 aldol reaction, and olefin metathesis at C12–C13. Soon after, macroaldolization forming the C2–C3 bond and a novel B-alkyl Suzuki coupling of C11–C12 were introduced by the Danishefsky group. Over the years, these basic approaches have been developed further by the original authors and adopted by many other groups for their own synthetic strategies. Many more alternatives were developed, omitting only a few carbon–carbon bonds not used in the assembly of the macrocycle and the attachment of the thiazole side-chain (Figure 21.6). Thanks to the enormous efforts of many groups, the early synthetic strategies mentioned above proved their value and were adapted to plant-scale syntheses. It is certainly beyond the scope of this review to discuss in detail the more than 20 direct and formal total syntheses published and reviewed already10a,b,d,f,j–p; however, a few points on the bond-forming strategies should be made. In most syntheses, the first strategic goals are, as in biosynthesis, the 12,13-olefins epothilone C and D (occasionally named also dEpoA and dEpoB, 2a, 2b), which are epoxidized in a more or less stereoselective manner to epothilone A and B.125,126 Alternatively, the sensitive epoxide is olefin,94,95,96,97 alkyne98 metathesis, Wittig reaction,99 12,13-diol cyclization90,91,92,93 epoxidation95,100,101 Nozaki-Hiyama-Kishi coupling102 Suzuki coupling,94,100 Pd-mediated coupling104 Nozaki-Kishi macrocyclisation105 alkyne-epoxide opening103 olefin metathesis,108,109 Castro-Stephens, Stille coupling110 sulfone alk ylation112 nitrile oxide cycloaddition106,107 olefin metathesis,114,115 Wittig reaction110 cyanosilylation111 Horner olefination90 Horner olefination99,118 R O S crotylboronate addition116 N OH aldol addition95,96,117,118 O O OH O Stille coupling113,119,120 nitrile oxide cycloaddition106 Nozaki-Kishi coupling120,121 dimethylallylborane addition,122 Cr-Reformatsky reaction123 aldol addition,94,124 allylboronation,117 Cr-Reformatsky reaction123 Yamaguchi macrolactonisation94,95,99 esterification95,96 FIGURE 21.6 C,C- and C,O-bond-forming strategies in the total syntheses of epothilones. Copyright © 2005 CRC Press, LLC introduced at an early stage with high stereoselectivity, and with the necessary precautions, carried through the synthesis by the Mulzer group.90 Soon after completion of the synthesis of the four major epothilones, A–D, the groups of Danishefsky and Nicolaou and others started to explore structure–activity relationships for cytotoxicity and tubulin binding. Following their own philoso- phies, either a limited number of selected analogs or large libraries were synthesized via combi- natorial chemistry. In this work, all stereocenters were probed for their relevance for biological activity, the alkyl groups on C8127 and C16128,129 were removed, or new ones were added at C10130 and C14.130,131 Further, the C664 and C12 methyl groups were replaced by larger alkyl and modified alkyl groups.49,115,119,128,132–136 C17–C18 Stille coupling allowed the introduction of a great variety of five- and six-membered heterocyclic side-chains, optionally carrying additional substitu- ents.113,119,129 Another obvious modification, the ring closure of C27 with the thiazole ring, was first realized by the Novartis group only in 2000 with the synthesis of benzthiazole, benzoxazole, and quinoline analogs.137 The 12,13-cis epoxide was changed to trans and, moreover, replaced by a cyclopropane,120,121,138–141 cyclobutane,121,140 or aziridin ring.142 Whereas 12,13-trans olefins and the epoxides derived from them were regularly unwanted by- products in the ring-closing olefin metathesis, Altmann et al. synthesized them in a stereoselective manner to prove their unexpected biological activity,143 which had been observed earlier by Nicolaou and coworkers.144 To rigidify the macrocycle and enforce certain conformations, 12,13-benzo,145 10,11,12-benzo,64,146 and 10,11-cyclopropano108 analogs, and a C23, C24 ethanol bridged analog,91 were synthesized. Similarly, the C4 geminal dimethyl group was integrated in a cyclopropane ring,53,54 11-hydroxy and 11-fluoro substituents were introduced,105 and oxygen was introduced as isosteric replacement of C10.147 Epothilones with extra double bonds at C9, C10101,115b,148,149 and C10, C11,108 or with a C9, C10 alkyne bond,150 became available en route to epothilone B and D. Occasionally, natural product building blocks were introduced, such as nerol123,151 and carene152 in the northern ring segment. The use of pantolactone64,153 for the C1–C5 ring segment and malic acid64 for the C13–C16 fragment provided the C3 and C15 stereocenters. In one and the same synthesis, both northwestern and southeastern ring segments were constructed from glucose by extensive func- tional group manipulations performed by Potier et al.154 Chemo-enzymatic methods were applied to establish the desired configuration at the C3, C7, and C15 hydroxyl groups. Thus, resolution was achieved by commercial esterases and lipases,155,156 by a genetically engineered lipase,157 or in case of C15, by a catalytic aldolase antibody.158–160 An aldolase-catalyzed asymmetric synthesis was developed by Wong et al.,161 providing building blocks with the C7 and C15 stereocenters in enanti- omerically pure form. In a polymer-supported synthesis of an epothilone C library, Nicolaou et al. used an olefinic linker to the polymer support, which is cleaved in the final step by participation in an olefin metathesis cyclization.144,162 In this way, only successfully cyclized molecules are released from the polymer to give immediately pure products. In an alternative approach, Ley et al.163 synthe- sized epothilone B, using polymer-supported reagents and scavengers while the growing product was kept in solution. Thus, tedious extraction and purification steps were avoided, making this approach amenable for large-scale synthesis. However, the advantages of the overall process may be doubtful in terms of cost effectiveness when costs for regeneration of the polymer reagents are included. In summary, during the past eight years of epothilone synthesis, around 160 papers have been published and more than 100 patents filed and, in part, granted. Apart from academic syntheses of natural epothilones and the development of new synthetic methods, around 1000 epothilone analogs may have been synthesized for structure–activity relationships studies. In two cases, epothilone D and epothilone C6–modified epothilone B analogs, large-scale processes were developed, taken to plant-level production by the Danishefsky and Schering groups, and introduced in clinical trials. B. SEMISYNTHESIS In the course of strain and production optimization at GBF, epothilone A and B became available in larger amounts in 1996 and 1997 and were used in a semisynthesis program in collaboration with Copyright © 2005 CRC Press, LLC R X R R S X X a, b 1,2 O c N H 7 OH OH OH HO OH O OH O O OH O 5 5 6 4 a. Pig liver esterase, DMSO/buffer, 37°C R = H or CH3 b. LiOH, CH3OH/H2O, RT X = O c. 2,4,6-trichlorobenzoyl chloride, N(C2H5)3, DMAP, Toluene SCHEME 21.2 Bristol-Myers Squibb. Later, a block-mutant strain was added, supplying epothilones C and D.17 Early one-step transformations were performed with epothilone A, including mono- and diacylation of the 3,7-dihydroxy groups, Swern oxidation to 3,5- and 5,7-diketones, and borohydride reduction of the C5 ketone to a 1:1 mixture of epimeric alcohols.164 Similarly, solvolysis of the epoxide produced mixtures of stereoisomeric diols (cat. H2SO4), chlorohydrins (HCl), or rearranged 14- membered lactones (TFA, BF3OEt2). Dihydroxylation of epothilone C with OsO4/NMO gave a 2:1 mixture of epimeric cis diols, dimethyl dioxirane, and MCPBA mixtures of natural epothilone A — its β-epoxy isomer — along with other oxidation products.165,166 12,13-Dihydro-epothilone C was formed selectively with diimine, whereas catalytic reduction produced complex mixtures because of hydrogenolysis of the allylic ester moiety.165 Disappointingly, all these derivatives were devoid of significant biological activity. Hydrolysis of the lactone in epothilones A–D was achieved by LiOH in MeOH/H2O to give seco-acids 4a–d,21,167 and depending on reaction time, by retroaldol reactions, the C7 aldehyde and C5 isopropyl-ketone fragments (Scheme 21.2).21 A very clean cleavage to seco-acids was achieved with pig liver esterase.21,51,167 Interestingly, epothilone A was restored by Yamaguchi macrolacton- ization of 3,7-nonprotected seco-acid 4a in 25% yield.21 Although C3 and C7 hydroxyl groups occupy β-positions to carbonyl groups, they are not eliminated under these and other more rigorous conditions. Only the 3,7-diformyl epothilones 7 on treatment with DBU gave in good yield trans- 2,3-olefins 8, which act as Michael acceptors (e.g., for cyanide to give a 1:1 mixture of 3α- and 3β-cyano epothilones 9a and 9b; Scheme 21.3).168 Another ideal target for structural modification was the 12,13-epoxide. Thus, epothilone A was selectively opened to bromohydrin 11 with MgBr2 in ether, oxidation to the bromoketone, α addition of cyanide followed by ring closure gave the trans-like cyanoepoxide 12.168 Its structure was proven by x-ray crystallography. Alternatively, the bromohydrin was converted to the azido alcohol, and Mitsunobu inversion and O-mesylation set the stage for phosphine reduction and cyclization to aziridin 10a with the natural α-configuration.142 Subsequently, the nitrogen was alkylated and acylated with a broad range of residues to 10b (Scheme 21.4). R R O O R S O S N O H OH 2 N OH O O a,b O c - e 3 O O O O O O H 8 OH CN O 7 O 9a (αCN) a. DBU, CH2CI2 c. (C2H5)3SiCI, N(C2H5)3, DMAP 9b (βCN) b. NH3, CH3OH d. KCN, 18-crown–6, DMF e. CH3Co2H, THF, H2O SCHEME 21.3 Copyright © 2005 CRC Press, LLC R N OH O S Br OH CN OH OP O a - h N j - l O O O O OH O O OP O O OH 10a R =H 11 12 i 10b R = alkyl, acyl a. (C2H5)3SiCl, N(C2H5)3, DMAP, DMF j. PCC, Py, CH2Cl2 b. MgBr2, (C2H5)2O/CH2Cl2 -20 to -5 oC k. KCN, 18-crown-6, THF c. NaN3, DMF, 42 oC l. TFA/CH2Cl2, -10 oC d. p-NBA, DEAD, Ph3P e. NH3, CH3OH f. MsCl, N(C2H5)3, CH2Cl2 g. P(CH3)3, THF/H2O h. TFA/CH2Cl2, -10 oC i. e.g. (CH3)2SO4, proton sponge, THF or RCOCl/Hunig's base SCHEME 21.4 Whereas epoxidation of the C12, C13 double

bond is the final step in epothilone A and B biosynthesis and in many total syntheses, it was desirable to use the rich supply of epothilone A and B from fermentation for the reverse reaction. This was achieved by the BMS group in good yield and retention of (Z)-configuration by reduction of unprotected 1a and 1b with TiCp2Cl2/Mg and WCl6/n-butyl lithium to epothilone C and D, respectively.139 Cyclopropanation of the C12, C13 double bond of epothilones C and D turned out to be a tough problem but was eventually achieved in low yield with dibromo- and dichlorocarbene generated under phase transfer conditions to give exclusively α-cyclopropane 13a.139 Subsequently, the halogens were removed by tinhydride reduc- tion to 13b (Scheme 21.5). Cleavage of the C12, C13 double bond of epothilone C and D by either ozonation or with OsO4/Pb(OAc)4 or, alternatively, diol formation and periodate cleavage of epothilone A and B gave valuable C12 aldehyde and ketone building blocks ready for, for example, Wittig condensation with synthetic ring segments.169,170 Inspired by the ring-closing olefin metathesis in the total synthesis of epothilones, the reverse reaction was investigated in presence of an excess of ethylene and second-generation Grubbs’s catalysts. The expected seco-diene 14 was obtained in good yield and processed further by replacement of the C13–C15 ring segment with a synthetic analog to give 15. Ring-closing olefin metathesis afforded, for example, the alkyne analog 16 of epothilone C along with the expected stereoisomer at C12/C13 (Scheme 21.6). Subsequently, 16 was epoxidized to the alkyne analog of epothilone A.171 More direct modifications of the linker group C16,C17 were catalytic hydrogenation, epoxidation followed by reduction to a 16-hydroxyl group,164 and its elimination to a C16 exomethylene derivative.21 Several epothilones, including the 16-desmethyl analog of epothilone A, could be isomerized to 6:4 mixtures with their Z-isomers 17 by benzophenone-sensitized irradiation (Scheme 21.7).171 C16-ketones 18 were obtained from epothilone A and B by ozonolysis and used as synthetic building blocks.164 Apart from standard carbonyl derivatives of the acetyl group, new side-chains were introduced by aldol condensation to, for example, 19, whereas Wittig-like olefinations only worked well with the boron analog (C2H4O2B)2CHLi yielding vinyl boronic acid 20 (Scheme 21.8). Suzuki coupling of 20 required reactive iodo compounds, such as iodo benzene; otherwise, it was trans- formed by N-iodosuccinimide (NIS) to the corresponding iodovinyl epothilone and introduced in Stille couplings,166 as described by Nicolaou et al.119,120 Initially, modification of the thiazole ring other than by replacement seemed hardly possible in the presence of the sensitive functional groups in the macrocycle. However, with sec-butyl and Copyright © 2005 CRC Press, LLC R R O OH OH a O O b O OH O O OH O 1a, 1b 2a, 2b b. R = H; TBSOTf, 2,6-lutidine, CH2Cl2. 0°C Bnzl(C2H5)3NCl, 50% NaOH(aq), CHBr3, 45°C a. R = H; TiCp2Cl2, Mg(s), THF R = CH3; WCl6, nBuLi, THF X X OH O O OH O 13a X = Br c 13b X = H c. Bu3SnH, AlBN, C6H12, 70°C SCHEME 21.5 t-butyl lithium at low temperatures, deprotonation of C19 and, to a lesser extent, C21 methyl was possible, and after quenching with deuterium chloride or carbon and heteroatom electrophiles, the corresponding derivatives were obtained in moderate to low yields.172 An unexpected and selective access to C21 derivatives was opened by thiazole N-oxidation using m-chloroperbenzoic acid (MCPBA). Expectedly, on treatment of the N-oxide 21 with acetic anhydride, a Polonovsky-like R S OH b N OH a 2a O c O OH O S 14 N S N R OTBS OH d, e O O O O OTBS O O OH 15 16 a. (H2I Mes2)(PCy3)Cl2Ru = CHPh, C2H4, CH2Cl2 b. TBSOTf, 2,6-lutidine c. DCC, DMAP, CH2Cl2 d. (PCy3)2Cl2Ru = CHPh, CH2Cl2 e. TFA, CH2Cl2, 0 oC SCHEME 21.6 Copyright © 2005 CRC Press, LLC 1 R1 R X X R2 S R2 OH N OH hν O O N S O OH O O OH O benzophenone/benzene high pressure mercury lamp 17 0°C, 10 min. Epothilones A - D, A8 x = O or ∆ SCHEME 21.7 R O OTMS O O O OTMSO d a - c 18 R R O O O HO B OTMS OH HO O N N O O OTMSO O OH O 19 20 (E:Z = 7:3) a. LiTMP, THF, –70°C d. (C2H4O2B)2CHLi, CH2CI2/THF 1:1 b. 1-methyl-3-imidazocarbaldehyde c. citric acid SCHEME 21.8 rearrangement transposed the oxygen to the C21 methyl, and on mild hydrolysis, epothilones E and F were obtained.50 It should be mentioned that all these steps were carried out without 3,7- OH protection (Scheme 21.9). As benzylic alcohols, epothilones E and F were easily oxidized to the corresponding aldehydes, and both alcohols and aldehydes were further modified by the introduction of halogen and a great variety of N, O, S, and carbon substituents.50,173 One of these, 21-amino epothilone B 22, was introduced into clinical trials by the BMS. An alternative to the N- oxide route, C21 hydroxylation, was achieved by biotransformation using Amycolata or Actinomy- ces strains.174 In addition to C21, Am. autotrophica also hydroxylates C9, C14, and the C12 methyl.38 O O S a, b S c, d S 21 N+ OH HO N H2N N OH O- O O O OH O O OH O Epothilone F 3b 21 22 (BMS-310705) a. TFA anhydride c. DPPA, DBU, DMF b. 1%NaOH in CH3OH/H2O d. H2/Pt, C2H5OH SCHEME 21.9 Copyright © 2005 CRC Press, LLC O S O S N OH a N OH O L3Pd HO O OH O O OH O 1a, 1b N - 3 O O S S b, c OH N OH N N3 HN HO O OH O O OH O 23 (BMS-247550) a. Pd(PPh3)4, NaN3, THF/H20, 45 oC 24 b. P(CH3)3, THF; then NH3/H2O, 45 oC c. EDC, HOBT, CH3CN/DMF SCHEME 21.10 Although obvious, isosteric replacements of the carboxyl group were not covered in the earlier total syntheses but were first realized by semisynthesis. In a very short and elegant reaction sequence, Vite and coworkers at BMS transformed epothilone B lactone to the lactam. Using palladium chemistry, the allylic lactone was opened and the C15 oxygen replaced by azide with overall retention of configuration. Reduction of the azide to an amine, and cyclization by standard peptide chemistry, produced the lactam in good yield (Scheme 21.10).51 Remarkably, the whole sequence did not require 3,7-OH protection and could be performed as a one-pot reaction. Because of its favorable cytotoxic and pharmacological profile, epothilone B–lactam 23 (BMS-247550, Ixabepilone) was introduced into clinical trials in 2000.10q V. STRUCTURE–ACTIVITY RELATIONSHIPS Investigation of structure–activity relationships started very early at GBF, with more than 39 natural epothilones isolated from large-scale fermentations and simple one-step semisynthetic derivatives. Shortly after, the Danishefsky and Nicolaou groups published biological data for a great variety of analogs, and industrial groups at Novartis and Schering followed. Unfortunately, the data generated for cytotoxicity/growth inhibition by different groups are hardly comparable on a quantitative basis because of the broad range of sensitivity of the various cell lines used. In vitro tubulin polymerization depends even more on the experimental conditions used.138,175 In this review, we classify activity as high (>50% of the parent compound), moderate (5%–50%), low (1%–5%), and inactive. It is generally accepted that a 16-membered macrocycle and the stereochemical assignment of substit- uents within the box (Figure 21.7) are a prerequisite for biological activity.113 However, good to moderate activity was observed with certain 17- and 18-membered epothilones.113,176 The substit- uents in the box may be modified to a certain extent. Epothilones with an (E)-2,3-double bond from elimination of the 3-hydroxyl are still moder- ately active, whereas 3β-cyano analogs are highly active.168 This means that the 3-hydroxyl is not a hydrogen bridge donor but, rather, an acceptor in the tubulin bound state of epothilone. Both 4- desmethylepothilones A1 and A 4 2 and 16-desmethylepothilone129 are highly active, whereas replace- ment of 8-Me by hydrogen drastically reduces activity.127 Extension of the C6 methyl to ethyl increases activity fourfold compared with epothilone B, and allyl and propyl derivatives are only slightly less active than the parent compound.64 An extra methyl in the 14α position reduces activity Copyright © 2005 CRC Press, LLC CH2, S, NH, N-Acyl, N-Alkyl oxazole benzene pyridine ring small alkyl, alkenyl, haloalkyl R2 O 10 (E) - ∆9,10 or (E) - ∆10,11 H, small alkyl, alkenyl, alkynyl S H R4 11 9 O-alkyl, S-alkyl R5 CH2OH, CH2Hal, CH2N3, CH2NH2 N OH R3 HC=O, HC=NNH2 essential part of O the molecule R1 N-oxide HN O OH O (E) - ∆2,3 small alkyl, alkenyl, alkynyl CN FIGURE 21.7 Structure–activity relationships of epothilone. slightly, whereas a 14β methyl causes refolding of the corresponding ring segment and loss of activity.131 The vinyl spacer C16/C15 in the side-chain and the presence and position of the sp2 nitrogen are of crucial importance for biological activity (Figure 21.8).49 Shorter spacers or those extended by one atom,166 (Z)-configuration of the double bond,171 an alkyne spacer,171 epoxides, or conformationally flexible derivatives with a C16/C17 single bond164 are inactive. The thiazole may be replaced by oxazole or α-pyridyl (25, X = O, 26), without loss of activity whereas β- and γ-pyridyl analogs are only moderately active or inactive.4,53,54,64,119 Likewise, thiazoles linked by their C2 or C5 position are inactive.113 Integration of the C16, C17 spacer in a benzene ring to give benzothiazole, benzoxazole (28), and quinoline (27) analogs confers high biological activity.64,137 In contrast, introduction of any substituent at the thiazole C19 (29) or alkylation of the nitrogen (30) abolishes activity.172 The C20 methyl on the thiazole may be replaced by hydrogen, small alkyl, alkenyl, or alkynyl groups173 or heteroatom substituents with retention of high activity.141,178 In general, three-atom (except hydrogen) residues are still highly to moderately active, and bigger ones are weakly active or inactive.173 The methyl group C21 may be modified by a variety of halogen, oxygen, nitrogen, and sulfur substituents or oxidized to the aldehyde and further derivatized without loss of activity, provided the space demand of the entire residue does not exceed three to four atoms.173 The thiazole-2-carboxylic acid derivative is inactive.173 Another part of great structural flexibility is the C12, C13 epoxide. Cyclopropane,120,139,140 cyclobutane,140 episulfide,21,177 and aziridin analogs including N-alkyl or N-acyl modifications142 are highly active, whereas the corresponding olefins, epothilone C and D, are moderately active.97,113,179 X X N N N N O O O O O O O O 25 26 27 28 S S O N S N N+ 19 BF - O R 4 O O O O O 29 30 31 FIGURE 21.8 Side-chain modified epothilones. Copyright © 2005 CRC Press, LLC FIGURE 21.9 (See color insert) Crystal structures of epothilone B (1b) from dichlo- romethane/petroleum ether (I), methanol/water (II), and epothilone A-N-oxide (III). Remarkably, the latter are almost as active as their epoxides in the in vitro tubulin assay,162,179 a fact that rules out an early hypothesis of covalent binding to tubulin protein via epoxide opening. Whereas β-epoxides and their analogs are entirely inactive, inversion of the configuration only at C12, corresponding to a C12, C13-trans configuration, retains activity.162 The same applies for the intro- duction of trans double bonds in the C9, C10,115,148,149 and C10, C11 position.4,108,109 The C12 methyl is extremely important for high activity of the epothilone B series. However, it may be replaced with retention of activity by small alkyl, alkenyl, and alkynyl groups or modified by introduction of halogen and oxygen substituents.49,132,136 Interestingly, the fluorescent epothilone D analog with a large m-dimethylaminobenzoyl residue is still about as active as Taxol.180 Last but not least, replace- ment of the lactone group of epothilone B, D, and F by a lactam reduces cytotoxic activity and tubulin binding slightly, while other properties such as the therapeutic window or p.o. efficacy are improved.51,57 Remarkably, lactam analogs obtained by total synthesis rather than semisynthesis seem to be less active.48,181 Studies on the solution conformation of epothilone A started even before the absolute config- uration was published. Thus, Georg’s group predicted182 the relative configuration and conformation of epothilone A on the basis of nuclear magnetic resonance measurements and molecular dynamics calculations. One of the proposed structures showed the correct relative configuration for the independent western and eastern ring segments. We found NOEs and coupling constants of epothilone B in DMSO/D2O solution in good agreement with a conformation similar to that

in the crystal.3 Only the thiazole side-chain is apparently rotating around both its single bonds. In a detailed conformational analysis, Taylor and Zajicek183 added a second minor conformer, designated B, with the 3-OH rotated from an axial to an equatorial position, accompanied by rotations in the C6/C7 segment. Interestingly, an equatorial 3-OH is also found in the crystal structure of epothilone B-N-oxide, where it is stabilized by a hydrogen bridge to the N-oxide, while the conformation around C6-C8 is not influenced.50 Apparently, a weak intramolecular hydrogen bond of the 3-OH to the thiazole nitrogen is also able to stabilize this 3-OHeq conformation in solution, as we found in (21R,S)-diastereomeric C21-substituted epothilone A and B derivatives.173 Thus, in the 20- epoxyethyl derivative 31 (Figure 21.8), the 3-OH nuclear magnetic resonance signal appears duplicated, and a strong NOE is observed between 17H and 3H. The latter indicates a C16–C19 syn-periplanar conformation as it is observed in the crystal structure of epothilone B from an aqueous solvent (Figure 21.9). Most remarkably, the same conformational changes were recently proposed for the tubulin- bound state of epothilone A from sophisticated nuclear magnetic resonance measurements (trans- ferred NOEs and transferred cross-correlation rates) by the Griesinger group.184,185 However, the unusual experimental conditions employed and the lack of appropriate controls cast some doubts on the relevance of these findings.186 In fact, changes may have been induced by the aqueous solvent of high ionic strength rather than specific binding to tubulin. Earlier results on the bioactive conformation of epothilones based on computational methods were published in a series of papers in 1999 and 2000. Inspired by some intriguing structural similarities of epothilone and Taxol and by the fact that Taxol is replaced from its binding site by epothilone, common pharmacophore models were postulated and corroborated by mapping of cross-resistance to certain amino acid replacements in β-tubulin.187–189 Recently, a tubulin point mutation outside the Taxol binding site was also observed that confers high resistance to both Taxol and epothilone.73 Even though the Copyright © 2005 CRC Press, LLC binding models proposed are contradictory, there are some similarities. Thus, after discovering that the C13 side-chain of Taxol is not absolutely necessary for tubulin binding, Kingston and Horwitz189 proposed that the C2 benzoyl residue of Taxol and the thiazole side-chain of epothilone occupy the same region of the protein. The same applies for the (less-favored) binding mode I proposed by Giannakakou et al.70 Surprisingly, little use is made of conformationally restrained epothilone analogs and the highly active C12,C13 trans isomers in these modeling studies. From replacement of the thiazole by a pyridine and modifications of the C16–C17 linker we know that the nitrogen has to be exactly positioned in the binding site serving presumably as hydrogen bond acceptor. The bioactive conformation of the thiazole ring is clearly C16–C19 syn-periplaner from the fact that no substituents are tolerated at C19166,172 and from the high activity of the conformationally syn-locked benzothiazole and quinoline analogs.64,137 Interestingly, the opposite, a C16–C19 anti- periplaner conformation of epothilone D and B, is adopted in the crystal structure of the complex with cytochrome P450epok,45 the biosynthetic enzyme that introduces the epoxide. Clearly, major advances of our knowledge of the epothilone–tubulin binding have to come from crystallographic and photoaffinity labeling studies, as was the case with the taxoids. VI. PRECLINICAL STUDIES Epothilone effects, as far as they are published, have been determined mostly in athymic nude mice but occasionally also in severe combined immunodeficiency (SCID) mice, nude rats, and beagle dogs, in most cases with xenografts of human cancer cells. The data are even more heterogeneous than those with cell cultures because of differences of the method and schedule of application and because of the kinds of cancer cells, epothilone analogs, and experimental animals used. The studies established, nevertheless, that epothilones act in vivo; that they inhibit human cancers, often resulting in cures of the animals; and that they are superior to Taxol and other cytostatic drugs in many cases, especially with MDR tumors. Severe combined immunodeficiency mice tolerate epothilone B well at a dose of 0.3 mg/kg; at 1 mg/kg, weight loss is observed; and at 3 mg/kg the animals show severe toxic effects (once per week, subcutaneously, vehicle 30% PEG + 70% 0.9% NaCl solution).84 In a preliminary study on nude mice, a dose of 0.6 mg/kg epothilone B, daily for 4 days, intraperitoneally, killed all animals, whereas all survived treatment with 25 mg/kg, daily for 5 days, of epothilone D.46 It was suggested that the epoxy group at C12/C13 may be responsible for high in vivo toxicity.55 Apart from leukopenia (43% decrease of lymphocytes), no hematological or other pathological changes in various organs were seen with either drug.46 Toxicity of epothilone B could be mitigated by an altered application schedule: at 0.6 mg/kg, 5 times every second day, intraperi- toneally, in DMSO, only three of seven mice died, and with the same dosage, but intravenously, all animals survived. Although epothilone D treatment resulted in appreciable regression and even cures of human mammary and ovarian carcinoma xenografts, epothilone B had too narrow a therapeutic window for good antitumor effects. The mode of application was later optimized intravenously by infusion over 6 h rather than by injection, using the vehicle cremophore-ethanol.23a,56 The benefits of epothilones are most impressive on MDR tumors. Thus, in nude mice epothilone D was >35,000-fold more active than Taxol on MDR human adenocarcinoma. It also was fully curative for human T-cell leukemia.23a A comparison of several epothilones, Taxol, and other cytostatic drugs showed superior performance of epothilone D and its C21–OH analog (d Epothione F) in many cases.23a,b Epothilone D (30 mg/kg, intravenously) was better than Taxol in controlling xenografts in nude mice of human solid lung, mammary, and one of two ovarian adenocarcinomas, and in particular of MDR sublines. It was inferior to Taxol, however, with colon, colon adeno, prostate adeno, and one of the ovarian adenocarcinoma lines. Complete remission was only achieved with MX-1 mammary carcinoma (as well by Taxol). Epothilone D was superior to Taxol, with complete remission, in four human leukemias, including MDR lines, and was inferior to Taxol only in the HL-60 line. Epothilone B and epothilone B–lactam were much less successful because of their narrow therapeutic windows (but see below). The same study showed that resistance to Copyright © 2005 CRC Press, LLC epothilone D developed only very slowly and to low levels (twofold in 21 months).56 In mouse plasma, epothilone D was quickly inactivated (t1/2 = 20 min), probably because of esterase activity, but was much more stable in human plasma (t1/2 = >3 h).56 Beagle dogs fully tolerated a single dose of 2–6 mg/kg, intravenously (40–120 mg/m2); they developed diarrhea and leukopenia and lost weight at 12 mg/kg (240 mg/m2); and 20 mg/kg (400 mg/m2) produced severe toxic effects (e.g., necrosis of the intestinal mucosa, cell decay in the bone marrow) and was lethal.56 At 6 mg/kg, t1/2 was >5 h, and plasma concentrations after 24 h were 0.045 µg/mL (0.092 µM), which is 10 times the IC50 (0.0095 µM) for leukemia cells.56 Recently, more effective epothilone analogs have been tested.136 Introduction of a C9, C10 double bond in epothilone D makes the compound 10 times more active on MX-1 mammary carcinoma xenografts in nude mice, but also more toxic (with a maximum tolerated dose [MTD] of 3 mg/kg, versus that of epothilone D [30 mg/kg]). Also, although the compound may lead to a complete remission of tumors, they often later come back. Performance could be much improved by further replacing the CH3 on C12 by CF3, which bestows higher stability (t1/2 = 212 min in mouse plasma, 10.5 h in human plasma; epothilone D, 46 min) and bioavailability (solubility in water: 20 µg/mL; epothilone D: 9.4 µg/mL). Thus, for example, with xenografts of human slow- growing lung carcinoma, A549, 25 mg/kg, 6 times every second day twice 8 days apart, resulted in 99.5% remission and, after two more doses, in complete elimination (epothilone D could not eradicate the tumor). Growth of a Taxol-resistant subline was completely stopped, and tumor size reduced by 24% at 20 mg/kg, 7 times every second day.136 Another novel compound, semisynthetic C-21 thiomethyl epothilone B (ABJ879), has just been proposed by Novartis for a phase I clinical study.190 It shows about the same tubulin-polymerizing activity as Taxol but is clearly more potent in blocking cell proliferation in cultures (average IC50 for seven human carcinomas: 0.09 nM vs. 4.7 nM of Taxol). The drug is fully active on MDR1 cells and lines with resistance resulting from β-tubulin mutations. In nude mice, single-dose treatment (2–3 mg/kg, intravenously, 6–7 d after implantation of the tumor) produces transient but long-lasting remission (25%–95%). Positive effects were seen with slow-growing lung adenocar- cinoma H-596, colon tumor HT-29, and large-cell lung tumor H-460. A preliminary pharmacoki- netics–pharmacodynamics drug-response model has also been presented.191 Doses of 1.5 and 1.8 mg/kg, every second week, were acceptable, but 1.5 and 2 mg/kg, once a week, caused intolerable weight losses. The drug penetrates tissues, where its level becomes 10 times higher than in plasma, and is retained there for a long time, so that plasma concentrations are not representative of tissue exposure — a fact that may be true also for other epothilone analogs. The controversy about the efficacy of epothilone B–lactam (BMS-247550) has been fueled by decidedly positive results reported by BMS scientists.57 In addition to various established tumor cell lines, three primary xenografts from biopsies were tested. A human ovarian carcinoma, Pat-7, from a patient who no longer responded to Taxol and other cytostatic drugs was highly responsive to epothilone B–lactam in nude mice at the optimal dose of 4.8–6.3 mg/kg, intravenously, 5 times every second day (the optimal dose of Taxol is 24–36 mg/kg). The log cell kill (LCK) was 2.9 (Taxol, 0.8). The same tumor in nude rats at the optimal dose of 3 mg/kg, intravenously, two times eight days apart, could be eradicated in four of six animals (LCK > 5; Taxol, LCK = 2.2, no cures [0/6]). Good LCK values and growth delay were also seen in nude mice with Taxol-resistant human ovarian and colon carcinomas. Also, in several other instances of Taxol-sensitive tumors in nude mice, the animals could be cured, and high LCK values (>6.3) were seen, even when the compound was not better than Taxol. Furthermore, epothilone B–lactam gave excellent results when administered orally. At the MTD (60–80 mg/kg, 5 times every second day), epothilone B–lactam was highly active on the Pat-7 xenograft (LCK = 3.1, growth delay of 32.8 d; in comparison, Taxol, intravenously, LCK = 1.3, growth delay of 9.8 d, orally inactive). HCT116 colon carcinoma was eradicated in all eight mice treated with 90 mg/kg, 5 times every second day (equivalent to the best intravenous regimen). In contrast to epothilone A and B, epothilone B–lactam is completely stable in rodent and human plasma and in liver microsomal preparations and shows clearly reduced toxicity.57 Copyright © 2005 CRC Press, LLC Epothilone D, too, was reported to be active in rats and beagle dogs after oral administra- tion.192,193 The drug was well tolerated, and >50% (dogs) and 10%–20% (rats) was found in the bloodstream, with half-lives of 9–11 h (dogs) and 6 h (rats), respectively. A study by the Sloan- Kettering Institute for Cancer Research yielded far fewer positive results with respect to epothilone β lactam in comparison to their own epothilone D.55 At 6 mg/kg, 4 times every day, intravenous infusion, MX-1 mammary tumor xenografts in nude mice were somewhat inhibited in growth, but no remission was seen, and the effect was not much improved by treatment with doses near the MTD (9 mg/kg). Similar results were obtained with K562 leukemia, which could be cured, however, when the animals subsequently were treated with C-21-OH epothilone D (30 mg/kg). A study, to date only in vitro, indicates that the dosing regime could have a massive influence on tumor angiogenesis and metastasis.194 Human vascular endothelial cell lines were more efficiently killed by a “metronomic” administration of cytostatic drugs, including epothilone B; that is, by extended (144 h) application of low (10–100 pM) concentrations rather than high doses for a short time (24 h). IC50 values were thus reduced to 25–140 pM, although those for tumor cells and fibroblasts remained high (550 pM

to >1 nM), so that this strategy appears to selectively affect endothelial cells. Not much has been published about the metabolic fate of epothilones in animals. Degradation of epothilone B (10 mg/kg, intravenously) in the liver of nude mice was studied by capillary high performance liquid chromatography combined with mass spectrometry.195 One hour after admin- istration, three metabolites could be recovered and identified guided by a key fragment at m/z166, an acylium ion derived from the very stable thiazole side-chain. The following degradation pathway was reconstructed: opening of the epoxide, opening of the macrolactone ring to the hydroxycar- boxylic acid, reduction of the C5-keto group, and reaction of the C5-hydroxy group with the carboxylic acid group to give a six-membered lactone ring or, after conjugation with taurine, to yield a six-membered lactam ring at the end of the chain. VII. CLINICAL APPLICATIONS Five epothilone analogs have been introduced into clinical studies (Figure 21.10). Epothilone B (1b, EPO906, Patupilone, Novartis) started phase I, beginning early in 1999, followed by phase II.10n,196 Epothilone B lactam (23, BMS-247550, Bristol-Myers Squibb) started phase I beginning October 1999 and completed September 2000; phase I in children started in April 2002; phase II trials have been operative since February 2001, and in 2002, there were 15 National Cancer Institute–sponsored phase II studies; it is presently in phase III.10q,197 Epothilone B C21-amine (22, BMS-310705) is in phase I (2002),10s,198 and epothilone D (2b, KOS-862, Kosan) is in phase I.136,192,193 Also in phase I is a synthetic epothilone of undisclosed structure, ZK-EPO (32, Schering). Epothilone B C2-thiomethyl (33, ABJ879, Novartis) is a recent candidate for a phase I study.190,191 Although most data are only published as abstracts, several excellent reviews summarize the results,10n,q,s,196,199 so that details need not be repeated here. The epothilone B lactam studies are by far the best documented ones. The epothilones have been tested in patients with many different carcinomas, mostly solid tumors, including MDR cases and others that are difficult to treat (e.g., cervical adenocarcinomas of the squamous type).200 The responses have varied between complete and partial remission, tumor stabilization, reduction of CA 125 levels (cancer antigen found at elevated levels in the blood of many cancer patients, especially those with ovarian cancer), and continued growth. The individual epothilone analogs appear to differ in the MTD, the dose-limiting toxicity (DLT), the side effects, half-life in the patient, and efficacy. It may still be too early to assess the quality of the drugs conclusively. For epothilone B lactam, the MTD was estimated to be 6 mg/m2, intravenously, over 1 h, daily for 5 days every 3 wk.201 At 8 mg/m2 dose-limiting neutropenia was seen, but many patients tolerated higher doses when the drug was applied in combination with filgrastim (granulocyte colony-stimulating Copyright © 2005 CRC Press, LLC O O S S S OH OH N OH N N O O HN O OH O O OH O O OH O Epothilone D 2b Epothilone B 1b Ixabepilon e 23 (KOS-862) (EPO-906) (BMS-247550) O ? O O S S S S N OH N OH N OH H2N O O HN O OH O O OH O O OH O 22 (BMS-310705) 32 (ZK-EPO) 33 (ABJ879) FIGURE 21.10 Epothilones in clinical trials. factor, which mitigates myelosuppression and neutropenia). Other side effects were peripheral neur- opathy, fatigue, nausea, stomatitis, myalgia, and arthralgia, usually in mild forms.10q,201,202 DLT was neutropenia. Two patients died of sepsis, probably as a consequence of neutropenia. The acceptable weekly dose may be 20–30 mg/m2. The drug was quickly absorbed by the tissue, where its half-life was 24–48 h, whereas t1/2 in the plasma was only 16.8 h. This was also supported by a study that analyzed effects in peripheral blood mononuclear cells in patients.197 It showed that MTs formed bundles 1 h after application (at the already toxic dose of 50 mg/m2, intravenously, over 1 h, once a week every 3 wk, in 70% of cells; at 40 mg/m2 in 63%), but after 24 h, the effect was much reduced (to 25% and 16%–23%, respectively). Cytotoxic effects followed with a time delay on MT bundling but were expressed within 24 h: Cleavage of poly(ADP-ribose)polymerase, indicating cell death (apoptosis and necrosis), was at base level after 1 h (3%) but rose to 28% of cells after 24 h. For treatment of progressive metastatic prostate cancer, a dose of 35 mg/m2, intravenously, over 3 h, once every 3 weeks (five cycles) in combination with estramustine phosphate (an estradiol analog without affinity to the estradiol receptor) is recommended.202 Positive responses and decline in prostate-specific antigen levels were seen in 40%–60% of patients. It appears that in aqueous solution the epoxide ring of epothilone B lactam opens to give the inactive diol.198 For epothilone B,199 the MTD is 6 mg/m2, intravenously, once every 3 weeks, or 2.5 mg/m2, intravenously, once every week. The DLT was diarrhea. Other toxic effects (neuropathy, nausea) were mild to moderate, and significant myelosuppression, mucositis, and alopecia were not regis- tered. Blood distribution was multiphasic. An initial half-life of 7 min was followed by a terminal half-life of 4–5 d, which indicates fast absorption by the tissue. There was no excretion of unaltered compound via the kidneys (<0.1%). For epothilone D, reports on clinical performance still are very preliminary.193 Five patients were treated with doses of 9 and 18 mg/m2 (intravenously, once every 3 weeks), at which no DLT was seen (mild to moderate emesis and anemia). The elimination half-life of the drug was 5–10 h, and plasma concentrations declined rapidly after infusion in a biphasic pattern. NOTE ADDED IN PROOF Since the completion of this article in July 2004 a considerable number of papers on biology and chemistry of epothilones were published. Among the most important ones is that by Nettles et al. on the structure of the epothilone A/tubulin complex.203 As with taxol, the structure was solved Copyright © 2005 CRC Press, LLC from two-dimensional crystals by electron crystallography and molecular modelling. Expectedly, epothilone occupies part of the taxol binding site, however, its conformation and contacts to the protein are profoundly different from previous proposals.204 ACKNOWLEDGMENTS We thank all former group members at the GBF for their contributions and the colleagues at Bristol- Meyers Squibb for a fruitful collaboration in the epothilone project. We further thank Mrs. Monica Kirchner for taking care for literature references and typing the manuscript. Financial support from Bristol-Myers Squibb and the Fonds der Chemischen Industrie is gratefully acknowledged. REFERENCES 1. Höfle, G., Biologically active secondary metabolites from myxobacteria — isolation and structure elucidation, in Scientific Annual Report of the GBF, Walsdorff, H.-J., Ed., Gesellschaft für Biotech- nologische Forschung, Braunschweig, 1991, 65-68. 2. Gerth, K. et al., Epothilones A and B: Antifungal and cytotoxic compounds from Sorangium cellulosum (Myxobacteria), J. Antibiot., 49, 560, 1996. 3. Höfle, G. et al., Epothilone A and B — Novel 16-membered macrolides with cytotoxic activity: isolation, crystal structure, and conformation in solution, Angew. Chem. Int. Ed., 35, 1567, 1996. 4. Hardt, I.H. et al., New natural epothilones from Sorangium cellulosum, strains So ce90/B2 and So ce90/D13: isolation, structure elucidation, and SAR studies, J. Nat. Prod., 64, 847, 2001. 5. Höfle, G. et al., German Patent 4138042, filed Nov. 11, 1991, granted Oct. 14, 1993; World Patent 9310121 filed Nov. 19, 1992. 6. See, e.g., Reichenbach, H. and Höfle, G., Biologically active secondary metabolites from myxobac- teria, Biotech. Adv., 11, 219, 1993; Höfle, G. and Reichenbach, H., The biosynthetic potential of the myxobacteria, in Sekundärmetabolismus bei Mikroorganismen, Kuhn, W. and Fiedler, H.-P., Eds., Attempto, Tübingen, 1995, 61-78; Reichenbach, H. and Höfle, G., Myxobacteria as producers of secondary metabolites, in Drug Discovery from Nature, Grabley, S. and Thiericke, R., Eds., Springer, Berlin, 1999, 149-179. 7. Lavelle, F., What’s new about new tubulin/microtubules binding agents? Exp. Opin. Invest. Drugs, 4, 771, 1995. 8. Bollag, D.M. et al., Epothilones, a new class of microtubule-stabilizing agents with a Taxol-like mechanism of action, Cancer Res., 55, 2325, 1995. 9. Gerth, K. et al., Myxobacteria, proficient producers of novel natural products with various biological activities past and future biotechnological aspects with the focus on the genus Sorangium, J. Biotech- nol., 106, 233, 2003. 10. a) Nicolaou, K.C., Roschangar, F., and Vourloumis, D., Chemical biology of epothilones, Angew. Chem. Int. Ed. Engl. 37, 2014, 1998; b) Mulzer, J., Martin, H.J., and Berger, M., Progress in the synthesis of chiral heterocyclic natural products: epothilone B and tartrolon B, J. Heterocyclic Chem., 36, 1421, 1999; c) G. Höfle, Epothilone — a natural product on the road to becoming a medicine, in Scientific Annual Report of the GBF, Jonas, R., Ed., 1999/2000, 21; d) Altmann, K.-H. et al., Epothilones and their analogs — potential new weapons in the fight against cancer, Chimia, 54, 612, 2000; e) Altmann, K.-H., Wartmann, M., and O’Reilly, T., Epothilones and related structures — a new class of microtubule inhibitors with potent in vivo antitumor activity, Biochim. Biophys. Acta, 1470, M79, 2000; f) Mulzer, J., Epothilone B and its derivatives as novel antitumor drugs: total and partial synthesis and biological evaluation, Monatshefte für Chemie, 131, 205, 2000; g) Winssinger, N. and Nicolaou, K.C., Epothilones and sarcodictyns: from combinatorial libraries to designed analogs, ACS Symposium Series 796, Anticancer Agents, Frontiers in Cancer Therapy, Ojima, I., et al., Eds., American Chemical Society, 2001, 148; h) Altmann, K.-H., Microtubule-stabilizing agents: a growing class of important anticancer drugs, Curr. Opin. Chem. Biol., 5, 424, 2001; i) Stachel, S.J., Biswas, K., and Danishefsky, S.J., The epothilones, eleutherobins, and related types of molecules, Curr. Pharm. Design 7, 1277, 2001; j) Nicolaou, K.C., Ritzén, A., and Namoto, K., Recent development in the Copyright © 2005 CRC Press, LLC chemistry, biology and medicine of the epothilones, Chem. Commun., 1523, 2001; k) Flörsheimer, A. and Altmann, K.-H., Epothilones and their analogs — a new class of promising microtubule inhibitors, Expert Opin. Ther. Patents, 11, 951, 2001; m) Wartmann, M. and Altmann, K.-H., The biology and medicinal chemistry of epothilones, Curr. Med. Chem. AntiCancer Agents, 2, 123, 2002; n) Borzilleri, R.M. and Vite, G.D., Epothilones: new tubulin polymerization agents in preclinical and clinical development, Drugs Future, 27, 1149, 2002; o) Altmann, K.-H., Epothilone B and its analogs — a new family of anticancer agents, Med. Chem., 3, 149, 2003; p) Mulzer, J. and Martin, H.J., Chem. Rec., 3, 258, 2004; q) Lin, N., Brakora, K., and Seiden, M., BMS-247550 Bristol-Myers Squibb/GBF, Curr. Opin. Investig. Drugs, 4, 746, 2003; r) Rivkin, A. et al., Application of ring-closing metathesis reactions in the synthesis of epothilones, J. Nat. Prod., 67, 139, 2004; s) Goodin, S., Kane, M.P., and Rubin, E.H., Epothilones: mechanism of action and biological activity, J. Clin, Oncol., 10, 2015, 2004. 11. Reichenbach, H., Order Myxococcales, in Bergey’s Manual of Systematic Bacteriology, 2nd edition, Garrity, G., Ed., Springer, New York, 2004. 12. Myxobacteria II, Dworkin, M. and Kaiser, D., Eds., American Society for Microbiology, Washington, DC, 1993, 1-404 13. Höfle, G. et al., (GBF) German Patent 4211056, 1993. 14. K. Gerth, personal communication. 15. Gerth, K. et al., Studies on the biosynthesis of epothilones: hydroxylation of epo A and B to epothilones E and F, J. Antibiot., 55, 41, 2002. 16. Gerth, K. et al., Studies on the biosynthesis of epothilones: The biosynthetic origin of the carbon skeleton, J. Antibiot., 53, 1373, 2000. 17. Gerth, K. et al., Studies on the biosynthesis of epothilones: the PKS and epothilone C/D monooxy- genase, J. Antibiot., 54, 144, 2001. 18. Reichenbach, H. et al., (GBF) German Patent 9647580 (1996). 19. Hofmann, H. et al., (Novartis) World Patent 9942602 (1999). 20. Hecht, H.-J. and Höfle, G., unpublished results; Crystallographic data of the structure have been deposited with the Cambridge Crystallographic Data Centre as supplementary publication no. CCDC- 241333 and CCDC-241334 — copies of the data can be obtained free of charge on application to CCDC, 12 Union Road, Cambridge CB2 1EZ, UK [fax: (+44)1223-336-033; e-mail: deposit@ccdc.cam.ac.uk]. 21. Höfle, G. et al., unpublished results. 22. Straubing, R.M., Biopharmaceuticals of paclitaxel, in Taxol, Science Application, Suffness, M., Ed., CRC Press, Boca Raton, FL, 1995, 238. 23. a) Chou, T.-C. et al., Desoxyepothilone B is curative against human tumor xenografts that are refractory to paclitaxel, Proc. Natl. Acad. Sci. USA, 95, 15798, 1998; b) Lee, C.B. et al., Total synthesis and antitumor activity of 12,13-desoxyepothilone F: An unexpected solvolysis problem at C15, mediated by

remote substitution at C21, J. Org. Chem., 65, 6525, 2000. 24. Tang, L. et al., Cloning and heterologous expression of the epothilone gene cluster, Science, 287, 640, 2000. 25. Julien, B. and Shah, S., Heterologous expression of epothilone biosynthetic genes in Myxococcus xanthus, Antimicrob. Agents Chemother., 46, 2772, 2002. 26. Lau, J. et al., Optimizing the heterologous production of epothilone D in Myxococcus xanthus, Biotechnol. Bioeng., 78, 280, 2002. 27. Arslanian, R.L. et al., Large-scale isolation and crystallization of epothilone D from Myxococcus xanthus cultures, J. Nat. Prod., 65, 570, 2002. 28. Regentin, R. et al., Nutrient regulation of epothilone biosynthesis in heterologous and native production strains, Appl. Microbiol. Biotechnol., 61, 451, 2003. 29. Frykman, S. et al., Modulation of epothilone analog production through media design, J. Ind. Micro- biol. Biotechnol., 28, 17, 2002. 30. Molnar, I. et al., The biosynthetic gene cluster for the microtubule-stabilizing agents epothilones A and B from Sorangium cellulosum So ce90, Chem. Biol., 7, 97, 2000. 31. Julien, B. et al., Isolation and characterization of the epothilone biosynthetic gene cluster from Sorangium cellulosum, Gene, 249, 153, 2000. 32. Schupp, T. et al., World Patent 9966028-A (Dec. 23, 1999), (Novartis AG). Copyright © 2005 CRC Press, LLC 33. Julien, B. et al., World Patent 200031247-A (June 2, 2000), (Kosan Bioscience). 34. Walsh, C.T. and O’Connor, S.E., Polyketide-nonribosomal peptide epothilone antitumor agents: the EpoA, B, C subunits, J. Ind. Microbiol. Biotechnol., 30, 448, 2003. 35. Tang, L. et al., Elucidating the mechanism of cis double bond formation in epothilone biosynthesis, J. Am. Chem. Soc., 126, 46, 2004. 36. Arslanian, R.L. et al., A new cytotoxic epothilone from modified polyketide synthases heterologously expressed in Myxococcus xanthus, J. Nat. Prod., 65, 1061, 2002. 37. a) Starks, C.M. et al., Isolation and characterization of new epothilone analogs from recombinant Myxococcus xanthus fermentations, J. Nat. Prod., 66, 1313, 2003; b) Boddy, C.N. et al., Precursor- directed biosynthesis of epothilone in Escherichia coli, J. Am. Chem. Soc., 126, 7436, 2004. 38. Tang, L. et al., Generation of novel epothilone analogs with cytotoxic activity by biotransformation, J. Antibiot., 56, 16, 2003. 39. O’Connor, S.E., Walsh, C.T., and Liu, F., Biosynthesis of epothilone intermediates with alternate starter units: engineering polyketide-nonribosomal interfaces, Angew. Chem. Int. Ed., 42, 3917, 2003. 40. Schneider, T.L., Walsh, C.T., and O‘Connor, S.E., Utilization of alternate substrates by the first three modules of the epothilone synthetase assembly line, J. Am. Chem. Soc., 124, 11272, 2002. 41. Chen, H. et al., Epothilone biosynthesis: assembly of the methylthiazolylcarboxy starter unit on the EpoB subunit, Chem. Biol., 8, 899, 2001. 42. O’Connor, S. E., Chen, H., and Walsh, C. T., Enzymatic assembly of epothilones: the epoC subunit and reconstitution of the epoA-ACP/B/C polyketide and nonribosomal peptide interfaces, Biochem- istry, 41, 5685, 2002. 43. Schneider, T.L., Shen, B., and Walsh, C.T., Oxidase domains in epothilone and bleomycin biosynthesis: thiazoline to thiazole oxidation during chain elongation, Biochemistry, 42, 9722, 2003. 44. Lopez, J.V., Naturally mosaic operons for secondary metabolite biosynthesis: variability and putative horizontal transfer of discrete catalytic domains of the epothilone polyketide synthase locus, Mol. Gen. Genomics, 270, 420, 2003. 45. Nagano, S. et al., Crystal structures of epothilone D-bound, epothilone B-bound, and substrate-free forms of cytochrome P450epoK, J. Biol. Chem., 278, 44886, 2003. 46. Chou, T.-C. et al., Desoxyepothilone B: An efficacious microtubule-targeted antitumor agent with a promising in vivo profile relative to epothilone B, Proc. Natl. Acad. Sci. USA, 95, 9642, 1998. 47. Kowalski, R.J., Giannakakou, P., and Hamel, E., Desoxyepothilone B is curative against human tumor xenografts that are refractory to paclitaxel, J. Biol. Chem., 272, 2534, 1997. 48. Schinzer, D. et al., Synthesis and biological evaluation of Aza-epothilones, Chem. Bio. Chem., 1, 67, 2000. 49. Su, D.-S. et al., Structure–activity relationships of the epothilones and the first in vivo comparison with paclitaxel, Angew. Chem. Int. Ed., 36, 2093, 1997. 50. Höfle, G. et al., N-Oxidation of epothilone A–C and O-Acyl rearrangement to C-19- and C21- substituted epothilones, Angew. Chem. Int. Ed., 38, 1971, 1999. 51. Borzilleri, R.M. et al., A novel application of a Pd(0)-catalyzed nucleophilic substitution reaction to the regio- and stereoselective synthesis of lactam analogs of the epothilone natural products, J. Am. Chem. Soc., 122, 8890, 2000. 52. Sepp-Lorenzino, L. et al., The microtubule-stabilizing agents epothilones A and B and their desoxy- derivatives induce mitotic arrest and apoptosis in human prostate cancer cells, Prostate Cancer Prost. Dis., 2, 41, 1999. 53. Nicolaou, K.C. et al., Total synthesis of oxazole- and cyclopropane-containing epothilone A analogs by the olefin metathesis approach, Chem. Eur. J. 3, 1957, 1997. 54. Nicolaou, K.C. et al., Total synthesis of oxazole- and cyclopropane-containing epothilone B analogs by the macrolactonization approach, Chem. Eur. J., 3, 1971, 1997. 55. Stachel, S.J. et al., On the interactivity of complex synthesis and tumor pharmacology in the drug discovery process: total synthesis and comparative in vivo evaluations of the 15-Aza epothilones, J. Org. Chem., 66, 4369, 2001. 56. Chou, T.-C. et al., The synthesis, discovery, and development of a highly promising class of micro- tubule stabilisation agents: curative effects of desoxyepothilones B and F against human tumor xenografts in nude mice, Proc. Natl. Acad. Sci. USA, 98, 8113, 2001. 57. Lee, F.Y.F. et al., BMS-247550: a novel epothilone analog with a mode of action similar to paclitaxel but possessing superior antitumor efficacy, Clin. Cancer Res., 7, 1429, 2001. Copyright © 2005 CRC Press, LLC 58. Wolff, A., Technau, A., and Brandner, G., Epothilone A induces apoptosis in neuroblastoma cells with multiple mechanisms of drug resistance, Int. J. Oncol., 11, 123, 1997. 59. Mühlradt, P.F. and Sasse, F., Epothilone B stabilizes microtubuli of macrophages like Taxol without showing Taxol-like endotoxin activity, Cancer Res., 57, 3344, 1997. 60. Kamath, K. and Jordan, M.A., Suppression of microtubule dynamics by epothilone B is associated with mitotic arrest, Cancer Res., 63, 6026, 2003. 61. Chen, J.G. and Horwitz, S.B., Differential mitotic responses to microtubule-stabilizing and -destabilizing drugs, Cancer Res., 62, 1935, 2002. 62. Iffert, R. and Sasse, F., Analyse des Apoptischen Zellkernzerfalls durch Bildverarbeitung, BIOforum, 22, 631, 1999. 63. Lichtner, R.B. et al., Subcellular distribution of epothilones in human tumor cells, Proc. Natl. Acad. Sci. USA, 98, 11743, 2001. 64. Klar, U. et al., Synthesis and biological activity of epothilones, ACS Symposium Series 796, Anticancer Agents, Frontiers in Cancer Therapy, Ojima, Iwao et al., Eds., American Chemical Society, 2001, 131–146. 65. Atadja, P., Isolation and characterisation of epothilone A resistant cells, Proc. Am. Assoc. Cancer Res., 41, 803, 2000. 66. Meurer-Grob, P., Kasparian, J., and Wade, R.H., Microtubule structure at improved resolution, Bio- chemistry, 40, 8000, 2001. 67. Wartmann, M. et al., Epothilones A and B accumulate several hundred fold inside cells, Proc. Am. Assoc. Cancer Res. 41, 213, 2000, Abstr. No. 1362. 68. Gupta, M.L. et al., Understanding tubulin-Taxol interactions: mutations that impart Taxol binding to yeast tubulin, Proc. Natl. Acad. Sci. USA, 100, 6394, 2003. 69. Bode, C.J. et al., Epothilone and paclitaxel: Unexpected differences in promoting the assembly and stabilisation of yeast microtubules, Biochemistry, 41, 3870, 2002. 70. Giannakakou, P. et al., A common pharmacophore for epothilone and taxanes: molecular basis for drug resistance conferred by tubulin mutations in human cancer cells, Proc. Natl. Acad. Sci. USA, 97, 2904, 2000. 71. He, L., Huang Yang, C.-P., and Horwitz, S.B., Mutations in β-tubulin map to domains involved in regulation of microtubule stability in epothilone-resistant cell lines, Mol. Cancer Ther., 1, 3, 2001. 72. Giannakakou, P. et al., A β-tubulin mutation confers epothilone resistance in human cancer cells, Proc. Am. Assoc. Cancer Res., 40, 284, 1999, Abstr. No. 1885. 73. Verrills, N.M. et al., Microtubule alterations and mutations induced by desoxyepothilone B: implica- tions for drug-target interactions, Chem. Biol., 10, 597, 2003. 74. Martello, L.A. et al., Elevated levels of microtubule destabilizing factors in a Taxol-resistant/dependent A549 cell line with an β-tubulin mutation, Cancer Res., 63, 1207, 2003. 75. Verdier-Pinard, P. et al., Direct analysis of tubulin expression in cancer cell lines by electrospray ionization mass spectrometry, Biochemistry, 42, 12019, 2003. 76. Poruchynsky, M.S. et al., Accompanying protein alterations in malignant cells with a microtubule- polymerizing drug-resistance phenotype and a primary resistance mechanism, Biochem. Pharmacol., 62, 1469, 2001. 77. Khabele, D. et al., Tumor necrosis factor-α related gene response to epothilone B in ovarian cancer, Gynecol. Oncol., 93, 19, 2004. 78. Atadja, P. et al., Gene expression profiling of epothilone A–resistant cells, in Mechanisms of Drug Resistance in Epilepsy: Lessons from Oncology, Novartis Foundation Symposium, 2002, pp. 119-136. 79. Ioffe, M. et al., Epothilone induced cytotoxicity is dependent on p53 in prostate cells, Proc. Am. Assoc. Cancer Res., 44, Abstr. No. 829, 2003. 80. Uyar, D. et al., Apoptotic pathways of epothilone BMS 310705, Gynecol. Oncol., 91, 173, 2003. 81. Griffin, D. et al., Molecular determinants of epothilone B derivative (BMS 247550) and Apo- 2L/TRAIL-induced apoptosis of human ovarian cancer cells, Gynecol. Oncol., 89, 37, 2003. 82. Yamaguchi, H. et al., Epothilone B analog (BMS-247550)-mediated cytotoxicity through induction of bax conformational change in human breast cancer cells, Cancer Res., 62, 466, 2002. 83. Bröker, L.E. et al., Late activation of apoptotic pathways plays a negligible role in mediating the cytotoxic effects of discodermolide and epothilone B in non-small cell lung cancer cells, Cancer Res., 62, 4081, 2002. Copyright © 2005 CRC Press, LLC 84. Pietras, K. et al., STI571 enhances the therapeutic index of epothilone B by a tumor-selective increase of drug uptake, Clin. Cancer Res., 9, 3779, 2003. 85. Wittmann, S. et al., Flavopiridol down-regulates antiapoptotic proteins and sensitizes human breast cancer cells to epothilone B-induced apoptosis, Cancer Res., 63, 93, 2003. 86. Moasser, M.M. et al., Farnesyl transferase inhibitors cause enhanced mitotic sensitivity to Taxol and epothilones, Proc. Natl. Acad. Sci. USA, 95, 1369, 1998. 87. Zhou, Y. et al., KDS-862 (epothilone D) with 5-fluorouracil in human colon cancer cell lines: synergistic antitumor effects, Proc. Am. Assoc. Cancer Res., 44, Abstr. No. 2746, 2003. 88. Fuino, I. et al., Histone deacetylase inhibitor IAQ824 down-regulates Her-2 and sensitizes human breast cancer cells to trastuzumab, taxotere, gemcitabine, and epothilone B, Mol. Cancer Ther., 2, 971, 2003. 89. Kim, J.-C. et al., Potential radiation-sensitizing effect of semisynthetic epothilone B in human lung cancer cells, Radiother. Oncol., 68, 305, 2003. 90. Martin, H.J., Drescher, M., and Mulzer, J., How stable are epoxides? A novel synthesis of epothilone B, Angew. Chem. Int. Ed., 39, 581, 2000. 91. Martin, H.J. et al., The 12,13-diol cyclization approach for a truly stereocontrolled total synthesis of epothilone B and the synthesis of a conformationally restrained analog, Chem. Eur. J., 7, 2261, 2001. 92. Mulzer, J., Karig, G., and Pojarliev, P., A novel highly stereoselective total synthesis of epothilone B and of its (12R, 13R) acetonide, Tetrahedron Lett., 41, 7635, 2000. 93. Meng, D. et al., Studies toward a synthesis of epothilone A: use of hydropyran templates for the management of acyclic stereochemical relationships, J. Org. Chem., 61, 7998, 1996. 94. Meng, D. et al., Total syntheses of epothilones A and B, J. Am. Chem. Soc., 119, 10073, 1997. 95. Schinzer, D. et al., Total synthesis of (–)-epothilone A, Angew. Chem. Int. Ed., 36, 523, 1997. 96. Nicolaou, K.C. et al., The olefin metathesis approach to epothilone A and its analogues, J. Am. Chem. Soc., 119, 7960, 1997. 97. Meng, D. et al., Remote effects in macrolide formation through ring-forming olefin metathesis: an application to the synthesis of fully active epothilone congeners, J. Am. Chem. Soc., 119, 2733, 1997. 98. Fürstner, A., Mathes, C., and Grela, K., Concise total syntheses of epothilone A and C based on alkyne metathesis, Chem. Commun., 1057, 2001. 99. Nicolaou, K.C. et al., Total syntheses of epothilones A and B via a macrolactonization-based strategy, J. Am. Chem. Soc., 119, 7974, 1997. 100. Balog, A. et al., Total synthesis of (–)-epothilone A, Angew. Chem. Int. Ed., 35, 2801, 1996. 101. Yang, Z. et al., Total synthesis of epothilone A: the olefin metathesis approach, Angew. Chem. Int. Ed., 36, 166, 1997. 102. Taylor, R.E. and Chen, Y. Total synthesis of epothilones B and D, Org. Lett., 3, 2221, 2001. 103. Liu, Z.-Y., Yu, C.-Z., Yang, J.-D., Chiral synthesis of the C3-13 segment of epothilone A, Synlett, 12, 1383, 1997. 104. Schinzer, D., Bauer, A., and Schieber, J., Synthesis of epothilones: stereoselective routes to epothilone B, Synlett, 8, 861, 1998. 105. Njardson, J.T., Biswas, K., and Danishefsky, S.J., Application

of hitherto unexplored macrocyclization strategies in the epothilone series: novel epothilone analogs by total synthesis, Chem. Commun., 2759, 2002. 106. Bode, J.W. and Carreira, E.M., Stereoselective syntheses of epothilones A and B in nitrile oxide cycloadditions and related studies, J. Org. Chem., 66, 6410, 2001. 107. Bode, J. W. and Carreira, E.M., Epothilone B, J. Am. Chem. Soc., 123, 3611, 2001. 108. Biswas, K. et al., Highly concise routes to epothilones: the total synthesis and evaluation of epothilone 490, J. Am. Chem. Soc., 124, 9825, 2002. 109. Yamamoto, K. et al., Effects of temperature and concentration in some ring closing metathesis reactions, Tetrahedron Lett., 44, 3297, 2003. 110. White, J.D. et al., Total synthesis of epothilone B, epothilone D, and cis- and trans-9,10-dehydroe- pothilone D, J. Am. Chem. Soc., 123, 5407, 2001. 111. Sawada, D., Kanai, M., and Shibasaki, M., Enantioselective total synthesis of epothilones A and B using multifunctional asymmetric catalysis, J. Am. Chem. Soc., 122, 10521, 2000. 112. Mulzer, J., Mantoulidis, A., and Öhler, E., Easy access to the epothilone family — synthesis of epothilone B, Tetrahedron Lett., 39, 8633, 1998. Copyright © 2005 CRC Press, LLC 113. Nicolaou, K.C. et al., Total synthesis of epothilone E and analogs with modified side chains through the Stille Coupling Reaction, Angew. Chem. Int. Ed., 37, 84, 1998. 114. Bertinato, P. et al., Studies toward a synthesis of epothilone A: stereocontrolled assembly of the acyl region and models for macrocyclization, J. Org. Chem., 61, 8000, 1996. 115. a) Rivkin, A. et al., Complex target-oriented total synthesis in the drug discovery process: the discovery of a highly promising family of second generation epothilones, J. Am. Chem. Soc., 125, 2899, 2003; b) Sun, J. and Sinha S.C., Stereoselective total synthesis of epothilones by the metathesis approach involving C9-C10 bond formation, Angew. Chem. Int. Ed., 41, 1381, 2002. 116. May, S.A. and Grieco, P.A., Total synthesis of ()-epothilone B, Chem. Commun., 1597, 1998. 117. Mulzer, J. and Mantoulidis, A., Synthesis of the C(1)-C(9) segment of the cytotoxic macrolides epothilon A and B, Tetrahedron Lett., 37, 9179, 1996. 118. Schinzer, D., Limberg, A., and Böhm, O.M., Studies towards the total synthesis of epothilones: asymmetric synthesis of the key fragments, Chem. Eur. J., 2, 1477, 1996. 119. a) Nicolaou, K.C. et al., Total synthesis of epothilone E and related side-chain modified analogs via a Stille Coupling based strategy, Bioorg. Med. Chem., 7, 665, 1999; b) Nicolaou, K.C. et al., Chemical synthesis and biological properties of pyridine epothilones. Chem. Biol., 7, 593, 2000. 120. Nicolaou, K.C. et al., Chemical synthesis and biological evaluation of novel epothilone B and trans- 12,13-cyclopropyl epothilone B analogs, Tetrahedron, 58, 6413, 2002. 121. Nicolaou, K.C. et al., Synthesis and biological evaluation of 12,13-cyclopropyl and 12,13-cyclobutyl epothilones, Chem. Bio. Chem., 1, 69, 2001. 122. Ramachandran, P.V. et al., Preparative-scale synthesis of both antipodes of B-γ,γ-dimethylallyldiiso- pinocampheylborane: application for the synthesis of C1–C6 subunit of epothilone, Tetrahedron Lett., 45, 1011, 2004. 123. Gabriel, T. and Wessjohann, L., The chromium-reformatsky reaction: asymmetric synthesis of the aldol fragment of the cytotoxic epothilons from 3-(2-bromoacyl)-2-oxazolidinones, Tetrahedron Lett., 38, 1363, 1997. 124. Panicker, B., Karle, J.M., and Avery, M.A., An unusual reversal of stereoselectivity in a boron mediated aldol reaction: enenantioselective synthesis of the C1–C6 segment of the epothilones, Tetrahedron, 56, 7859, 2000. 125. Stachel, S.J. and Danishefsky, S.J., Chemo- and stereoselective epoxidation of 12,13-desoxyepothilone B using 2,2-dimethyldioxirane, Tetrahedron Lett., 42, 6785, 2001. 126. Liu, Z.-Y. et al., Total synthesis of epothilone A through stereospecific epoxidation of the p-methoxy- benzyl ether of epothilone C, Chem. Eur. J., 8, 3747, 2002. 127. Balog, A. et al., Stereoselective syntheses and evaluation of compounds in the 8-desmethylepothilone A. Series: some surprising observations regarding their chemical and biological properties, Tetrahedron Lett., 38, 4529, 1997. 128. Nicolaou, K.C. et al., Synthesis of 16-desmethylepothilone B: improved methodology for the rapid, highly selective and convergent construction of epothilone B and analogs, Chem. Commun., 519, 1999. 129. Nicolaou, K.C. et al., Total synthesis of 16-desmethylepothilone B, epothilone B10, epothilone F, and related side chain modified epothilone B analogs, Chem. Eur. J., 6, 2783, 2000. 130. Taylor, R.E. et al, Conformation–activity relationships in polyketide natural products. Towards the biologically active conformation of epothilone, Org. Biomol. Chem., 2, 127, 2004. 131. Taylor, R.E., Chen, Y., and Beatty, A., Conformation–activity relationships in polyketie natural products: a new perspective on the rational design of epothilone analogs, J. Am. Chem. Soc., 125, 26, 2003. 132. Nicolaou, K.C. et al, Total synthesis of 26-hydroxyepothilone B and related analogs, Chem. Commun., 2343, 1997. 133. Nicolaou, K.C. et al., Total synthesis of 26-hydroxy-epothilone B and related analogs via a macro- lactonization based strategy, Tetrahedron, 54, 7127, 1998. 134. Chappell, M.D. et al., Probing the SAR of dEpoB via chemical synthesis: a total synthesis evaluation of C26-(1,3-dioxolanyl)-12,13-desoxyepothilone B, J. Org. Chem., 67, 7730, 2002. 135. Koch, G., Loiseleur, O., and Altmann, K.-H., Total synthesis of 26-fluoro-epothilone B, Synlett, 4, 693, 2004. 136. Chou, T.-C. et al., Design and total synthesis of a superior family of epothilone analogs, which eliminate xenograft tumors to a nonrelapsable state, Angew. Chem. Int. Ed., 42, 4762, 2003. Copyright © 2005 CRC Press, LLC 137. Altmann, K.-H. et al., Synthesis and biological evaluation of highly potent analogs of epothilones B and D, Bioorg. Med. Chem. Lett., 10, 2765, 2000. 138. Buey, R.M. et al., Interaction of epothilone analogs with the paclitaxel binding site: relationship between binding affinity, microtubule stabilisation, and cytotoxicity, Chem. Biol., 11, 225, 2004. 139. Johnson, J. et al., Synthesis, structure proof, and biological activity of epothilone cyclopropanes, Org. Lett., 2, 1537, 2000. 140. Nicolaou, K.C. et al., Chemical synthesis and biological evaluation of cis- and trans-12,13-cyclopropyl and 12,13-cyclobutyl epothilones and related pyridine side chain analogs, J. Am. Chem. Soc., 123, 9313, 2001. 141. Nicolaou, K.C. et al., Design, synthesis, and biological properties of highly potent epothilone B analogs, Angew. Chem. Int. Ed., 42, 3515, 2003. 142. Regueiro-Ren, A. et al., Synthesis and biological activity of novel epothilone aziridines, Org. Lett., 3, 2693, 2001. 143. Altmann, K.-H. et al., The total synthesis and biological assessment of trans-epothilone A, Helvetica Chim. Acta, 85, 4086, 2002. 144. Nicolaou, K.C. et al., Synthesis of epothilone A and B in solid and solution phase, Nature, 387, 268, 1997. 145. Glunz, P.W. et al., The synthesis and evaluation of 12,13-benzodesoxyepothilone B: a highly conver- gent route, Tetrahedron Lett., 40, 6895, 1999. 146. End, N. et al., Synthetic epothilone analogs with modifications in the northern hemispere and the heterocyclic side-chain-synthesis and biological evaluation, Fourth International Electronic Confer- ence on Synthetic Organic Chemistry (ECSOC-4), September 1–30, 2000. 147. Quintard, D. et al., Enantioselective synthesis of 2,3-dehydro-3-desoxy-10-oxa epothilone D, Synlett, 13, 2033, 203. 148. Rivkin, A. et al., On the introduction of a trifluoromethyl substituent in the epothilone setting: chemical issues related to ring forming olefin metathesis and earliest biological findings, Org. Lett., 4, 4081, 2002. 149. Yoshimura, F. et al., Synthesis and conformational analysis of (E)-9,10-dehydroepothilone B: a suggestive link between the chemistry and biology of epothilones, Angew. Chem. Int. Ed., 42, 2518, 2003. 150. White, J.D. and Sundermann, K.F., Synthesis, conformational analysis, and bioassay of 9,10-didehy- droepothilone D, Org. Lett., 4, 995, 2002. 151. Zhang, H.S., Zhong, C.F., and Bao, X.P., Studies of the total synthesis of epothilone B and D: a facile synthesis of C7-C14 and C15-C21 fragments, Chinese Chem. Lett., 14, 115, 2003. 152. Akbutina, F.A. et al., Approaches to epothilone carboanalogs starting from ∆3-carene, Russ. J. Org. Chem., 39, 75, 2003. 153. Akbutina, F.A. et al., Chiral synthetic block based on (R)-pantolactone, Mendeleev Commun., 151, 2003. 154. Ermolenko, M.S. and Potier, P., Synthesis of epothilones B and D from D-glucose, Tetrahedron Lett., 43, 2895, 2002. 155. Machajewski, T.D. and Wong, C.-H., Chemoenzymatic synthesis of Key epothilone fragments, Syn- thesis, 1469, 1999. 156. Shioji, K. et al., Synthesis of C1-C6 fragment for epothilone A via lipase-catalyzed optical resolution, Synth. Comm., 31, 3569, 2001. 157. Bornscheuer, U.T., Altenbuchner, J., and Meyer, H.H., Directed evolution of an esterase for the stereoselective resolution of a key intermediate in the synthesis of epothilones, Biotechnol. Bioeng., 58, 554, 1998. 158. Sinha, S.C. et al., Synthesis of epothilone analogs by antibody-catalyzed resolution of thiazole aldol synthons on a multigram scale. Biological consequences of C-13 alkylation of epothilones, Chem. Bio. Chem, 2, 656, 2001. 159. Sinha, S.C. at al., Sets of aldolase antibodies with antipodal reactivities. Formal synthesis of epothilone E by large-scale antibody-catalyzed resolution of thiazole aldol, Org. Lett., 1, 1623, 1999. 160. Sinha, S.C., Barbas III, C.F., and Lerner, R.A., The antibody catalysis route to the total synthesis of epothilones, Proc. Natl. Acad. Sci. USA, 95, 14603, 1998. 161. Liu, J. and Wong, C.-H., Aldolase-catalyzed asymmetric synthesis of novel pyranose synthons as a new entry to heterocycles and epothilones, Angew. Chem. Int. Ed., 41, 1404, 2002. Copyright © 2005 CRC Press, LLC 162. Nicolaou, K.C. et al., Designed epothilones: combinatorial synthesis, tubulin assembly properties, and cytotoxic action against Taxol-resistant tumor cells, Angew. Chem. Int. Ed., 36, 2097, 1997. 163. Storer, R.J. et al., A total synthesis of epothilones using solid-supported reagents and scavengers, Angew. Chem. Int. Ed., 42, 2521, 2003; Storer, R.I. et al., Multi-step application of immobilized reagents and scavengers: a total synthesis of epothilone C, Chem. Eur. J., 10, 2529, 2004. 164. Sefkow, M. et al., Oxidative and reductive transformations of epothilone A, Bioorg. Med. Chem. Lett., 8, 3025, 1998. 165. Sefkow, M., Kiffe, M., and Höfle, G., Derivatization of the C12-C13 functional groups of epothilones A, B and C, Bioorg. Med. Chem. Lett., 8, 3031, 1998. 166. Höfle, G. et al., Epothilone A–D and their thiazole-modified analogs as novel anticancer agents, Pure Appl. Chem., 71, 2019, 1999. 167. Boddy, C.N. et al., Epothilone C macrolactonization and hydrolysis are catalyzed by the isolated thioesterase domain of epothilone polyketide synthase, J. Am. Chem. Soc., 125, 3428, 2003. 168. Regueiro-Ren, A. et al., SAR and pH stability of cyano-substituted epothilones, Org. Lett., 4, 3815, 2002. 169. Vite, G.D. et al., (BMS), US Patent, 6399638 B1 (June, 4, 2002). 170. Dong, S.D. et al., Rapid access to epothilone analogs via semisynthetic degradation and reconstruction of epothilone D, Tetrahedron Lett., 45, 1945, 2004. 171. Karama, U. and Höfle, G., Synthesis of epothilone 16,17-alkyne by replacement of the C13-C15(O)- ring segment of natural epothilone C, Eur. J. Org. Chem., 1042, 2003. 172. Sefkow, M. and Höfle, G., Substitutions at the thiazole moiety of epothilone, Heterocyles, 48, 2485, 1998. 173. Glaser, N., Doctoral Thesis, Technical University of Braunschweig, 2001. 174. Matson, J.A. et al., (BMS), International Patent Application, WO 00/39276 (July 6, 2000). 175. Hamel, E., Evaluation of antimitotic agents by quantitative comparisons of their effects on the polymerization of purified tubulin, Cell Biochem. Biophys., 38, 1, 2003. 176. a) Nicolaou, K.C. et al., Variation der Ringgröße von Epothilonen — Totalsynthese von [14]-, [15], [17]- und [18]Epothilon A, Angew. Chem., 110, 85, 1998; b) Rivkin, A. et al., Total syntheses of [17]- and [18]dehydrodesoxyepothilones B via a concise ring-closing metathesis-based strategy: correlation of ring size with biological activity in the epothilone series, J. Org. Chem., 67, 7737, 2002. 177. U.S. Patent Application Serial No. 09/280,210, U.S. Patent No. 6,498,257 B1, 2002. 178. Nicolaou, K.C. et al., Synthesis and biological properties of C12,13-cyclopropyl-epothilone A and related epothilones, Chem. Biol., 5, 365, 1998. 179. Su, D.-S. et al., Total synthesis of (-)-epothilone B: an extension of the Suzuki coupling method and insights into structure-activity relationships of the epothilones, Angew. Chem. Int. Ed., 36, 757, 1997. 180. Ganesh, T. et al., Synthesis and biological evaluation of fluorescently labeled epothilone analogs for tubulin binding studies, Tetrahedron, 59, 9979, 2003. 181. Stachel, S. J. et al., On the total synthesis and preliminary biological evaluations of 15(R) and 15(S) Aza-dEpoB: a Mitsunobu inversion at C15 in pre-epothilone fragments, Org. Lett., 2, 1637, 2000. 182. Victory, S.F. et al., Relative stereochemistry and solution conformation of the novel paclitaxel-like antimitotic agent epothilone A, Bioorg. Med. Chem. Lett., 6, 893, 1996. 183. Taylor, R.E. and Zajicek, J., Conformational properties of epothilone, J. Org. Chem., 64, 7224, 1999; ibid. 65, 5449, 2000. 184. Carlomagno, T. et al., The high-resolution solution structure of epothilone A bound to tubulin: an understanding of the structure-activity relationships for a powerful class of antitumor agents, Angew. Chem. Int. Ed., 42, 2511, 2003. 185. Carlomagno, T. et al., Derivation of dihedral angles from CH-CH dipolar-dipolar cross-correlated

relaxation rates: a C-C torsion involving a quaternary carbon atom in epothilone A bound to tubulin, Angew. Chem. Int. Ed., 42, 2515, 2003. 186. The spectral data for the supposed epothilone/tubulin complex in aqueous buffer are compared with that of epothilone in dichloromethane, and the tubulin is stated not to polymerise even in the presence of a 100 fold molar excess of epothilone. 187. Ojima, I. et al., A common pharmacophore for cytotoxic natural products that stabilize microtubules, Proc. Natl. Acad. Sci. USA, 96, 4256, 1999. 188. Wang, M. et al., A unified and quantitative receptor model for the microtubule binding of paclitaxel and epothilone, Org. Lett., 1, 43, 1999. Copyright © 2005 CRC Press, LLC 189. He, L. et al., A common pharmacophore for Taxol and the epothilones based on the biological activity of a taxane molecule lacking a C13 side chain, Biochemistry, 39, 3972, 2000. 190. Wartmann, M., Preclinical pharmacological profile of ABJ879, a novel epothilone B analog with potent and protracted anti-tumor activity, AACR Ann. Meet. Orlando, March 27–31, 2004, Abstr. No. 5440. 191. DiLea, C. et al., A PK-PD dose optimization strategy for the microtubule stabilizing agent ABJ879, AACR Ann. Meet. Orlando, March 27–31, 2004, Abstr. No. 5132. 192. Johnson et al., Oral bioavailability of KOS-862 (epothilone D), a potent stabilizer of microtubulin polymerization in rats and Beagle dogs. Proc. Am. Assoc. Cancer Res., 44, Abstr. No. 5339, 2003. 193. Rosen, P.J. et al., KOS-862 (epothilone D): results of a phase I dose-escalating trial in patients with advanced malignancies, Proc. Am. Soc. Clin. Oncol. 21: Abstr. No. 413, 2002. 194. Bocci, G., Nicolaou, K.C., and Kerbel, R.S., Protracted low-dose effects on human endothelial cell proliferation and survival in vitro reveal a selective antiangiogenic window for various chemothera- peutic drugs, Cancer Res., 62, 6938, 2002. 195. Blum, W. et al., In vivo metabolism of epothilone B in tumor-bearing nude mice: identification of three new epothilone B metabolites by capillary high-pressure liquid chromatography/mass spectrom- etry/tandem mass spectrometry, Rapid Commun. Mass Spectrom., 15, 41, 2001. 196. Altaha, R. et al., Epothilones: a novel class of non-taxane microtubule-stabilizing agents, Curr. Pharm. Design, 8, 1707, 2002. 197. McDaid, H.M. et al., Validation of the pharmacodynamics of BMS-247550, an analog of epothilone B, during a phase I clinical study, Clin. Cancer Res., 8, 2035, 2002. 198. Borman, S., First disclosures of clinical candidates, Chem. Eng. News, 80, 35, 2002. 199. Rothermel, J. et al., EPO906 (epothilone B): A promising novel microtubule stabilizer, Sem. Oncol., 30, Suppl. 6, 51, 2003. 200. Agrawal, M. et al., Treatment of recurrent cervical adenocarcinoma with BMS-247550, an epothilone B analog, Gynecol. Oncol., 90, 96, 2003. 201. Abraham, J. et al., Phase I trial and pharmacokinetic study of BMS-247550, an epothilone B analog, administered intravenously on a daily schedule for five days, J. Clin. Oncol., 21, 1866, 2003. 202. Smaletz, O. et al., Pilot study of epothilone B analog (BMS-247550) and estramustine phosphate in patients with progressive metastatic prostate cancer following castration, Ann. Oncol., 14, 1518, 2003. 203. Nettles, J.H. et al., The Binding Mode of Epothilone A on α,β-Tubulin by Electron Crystallography, Science, 305, 866, 2004. 204. Heinz, D.H. et al., Much Anticipated — The Bioactive Conformation of Epothilone and Its Binding to Tubulin, Angew. Chem. Int. Ed., 44, 1298, 2005. Copyright © 2005 CRC Press, LLC 22 Enediynes Philip R. Hamann, Janis Upeslacis, and Donald B. Borders CONTENTS I. Introduction II. History III. Structural Classes and Producing Organisms IV. Biosynthesis V. Molecular Mechanisms of Action VI. Cellular Mechanisms of Action VII. Preclinical Studies VIII. Clinical Experience IX. Summary References I. INTRODUCTION The enediyne antitumor antibiotics represent some of the most potent biologically active natural products ever discovered. These compounds possess both antibacterial and antitumor activity and are related by containing enediyne functionality constrained within 9- or 10-member rings, with varying degrees of chemical complexity in the remainder of the molecules. Most of these antitumor antibiotics are produced by microorganisms classified as the actinomycetes. The challenging chem- ical structures of these enediynes, their mechanisms of action, and their biological activity have attracted considerable scientific attention. Two of these compounds, neocarzinostatin and esperam- icin A1, were advanced into anticancer clinical trials as single agents. Esperamicin A1 was even- tually abandoned from further development because of toxicity issues. However, neocarzinostatin has been approved for human use in Japan for a number of cancer indications, both as the natural product itself as well as a part of a polymer-based conjugate. A monoclonal antibody conjugate of calicheamicin was successful in clinical trials and has been commercialized for the treatment of acute myeloid leukemia in adults. Some of the previous reviews covering the enediynes have included a comprehensive summary in book form1 as well as a number of focused reviews detailing structure determinations, mechanism of action, synthetic approaches, biosynthesis, and biological activity.2–8 A book devoted to neo- carzinostatin has also been published.9 This review is limited to a selective coverage of highlights, with an emphasis on recent developments that have produced, or have the potential to produce, marketed products based on the enediynes. II. HISTORY The structures of the major components of the calicheamicins and esperamicins were published as back-to-back articles in a 1987 issue of the Journal of the American Chemical Society based on Copyright © 2005 CRC Press, LLC FIGURE 22.1 Structures of representative 10-membered ring enediyne antibiotics. research at two different pharmaceutical companies.10–13 It is fortunate that these publications appeared together because, taken individually, the structures with 10-membered ring enediynes were difficult to believe. These remarkable compounds were structurally quite different from one another, but surprisingly, both had the same novel 10-member ring enediyne system, an allylic methyl trisulfide group, a hydroxyamino-linked disaccharide unit, and unique amino- and thio- containing carbohydrates (Figure 22.1). In addition, the calicheamicins possessed a highly substi- tuted halobenzene unit. Both classes of compounds showed extremely potent biological activity with a mechanism of action involving free-radical damage to DNA at selective sites. In the case of the calicheamicins, a high percentage of the damage involved double-strand cleavage of DNA.14 In time, the structures of additional members of both families of compounds were elucidated.15,16 A key factor in the discovery of the calicheamicin complex in fermentation extracts was a newly developed in vitro test termed the biochemical induction assay that detected DNA damaging agents.17 The assay was extremely sensitive for detecting the very low concentrations of the calicheamicin complex present in the early fermentation. Even very crude preparations of the complex were significantly more potent in this assay than the reference compounds, and this provided the first suggestion that the compounds were novel. Pure calicheamicin γ I 1 was extremely active in the biochemical induction assay and against Gram-positive bacteria in the picogram per milliliter range.18,19 It was also very active against Gram-negative bacteria. Copyright © 2005 CRC Press, LLC Once the structures of calicheamicin and esperamicin were firmly established, the very elegant and earlier structural work with the more remarkable nine-membered ring enediyne chromophoric component and associated apoprotein of the antitumor antibiotic neocarzinostatin could be fully appreciated.20 This compound could only be isolated along with the apoprotein, which provided stability to the highly reactive 9-member ring enediyne chromophoric component. It became apparent that calicheamicin and esperamicin represented 10-member ring enediyne antitumor anti- biotics that were stable without an associated protein, whereas neocarzinostatin represented another type, the 9-member chromoprotein enediynes, requiring a protein for stabilization. A number of the enediyne antibiotics listed in Table 22.1 have various synonyms. The numeric designations were given by the pharmaceutical companies at the initial stages of their investiga- tions and then were changed to more conventional names if the compounds showed promise for further development. Thus, the esperamicin complex was originally designated BMY 28175 by Bristol-Myers. Two other companies working on components of the same complex at approxi- mately the same time as Bristol-Myers designated their compounds as FR900405 and FR900406 (Fujisawa) or veractamycins A and B (Parke-Davis). The BMY 28175 designation was subse- quently changed to esperamicin and eventually shown to be identical to the products isolated by the other two companies.21 Lederle Laboratories originally named the LL-E33288 complex “calichemicin,” and that name appeared in the initial structure papers. However, a conflict caused by the resemblance of the name of “calichemicin” to that of a chemical company required changing the name to “calicheamicin.” TABLE 22.1 Summary of Enediyne Antibiotic Complexes Produced by Various Organisms Name Synonyms Producing Organism Structure Typea Calicheamicin LL-E33288 Micromonospora echinospora, 10-mr enediyne with allylic calichemicin ssp. calichensis MeSSS Esperamicin BMY 28175 Actinomadura verrucosospora 10-mr enediyne with allylic FR900405 MeSSS FR900406 Veractamycins A and B Namenamicin Polysyncraton lithostrotum 10-mr enediyne with allylic (ascidian) MeSSS Shishijimicin Didemnum proliferum 10-mr enediyne with allylic (ascidian) MeSSS Dynemicin BU 3420T Micromonospora chersina 10-mr enediyne with Micromonospora globosa anthraquinone Endynamicin Micromonospora globosa 10-mr enediyne with anthraquinone Actinoxanthin Streptomyces globisporus Presumed 9-mr enediyne with apoprotein Auromomycin Macromomycin I Streptomyces macromyceticus Presumed 9-mr enediyne with apoprotein Kedarcidin Streptoalloteichus sp. 9-mr enediyne with apoprotein Lidamycin C-1027 Streptomyces globisporus 9-mr enediyne with apoprotein Maduropeptin BBM 1644 Actinomadura madurae 9-mr enediyne with apoprotein Neocarzinostatin Zinostatin Streptomyces carzinostaticus 9-mr enediyne with apoprotein NSC 69856 Antibiotic N1999A2 Streptoverticillium sp. 9-mr enediyne a 9- and 10-membered rings are abbreviated as 9-mr and 10-mr, respectively. Copyright © 2005 CRC Press, LLC FIGURE 22.2 Structures of representative nine-membered ring enediyne antibiotics. III. STRUCTURAL CLASSES AND PRODUCING ORGANISMS The enediyne antitumor antibiotics fall into two general classes. Representative examples wherein the conjugated enediyne system is a part of a 10-member macrocycle are shown in Figure 22.1, and examples of the nine-member enediynes identified so far are shown in Figure 22.2. Most of the 10- member ring enediynes are related to calicheamicin and esperamicin and have an allylic methyl trisulfide group that can act as a trigger to initiate formation of a diradical through cycloaromatization of the core ring. These diradicals are responsible for the resultant DNA damage. Although dynemicin and endynamicin, differing from each other by virtue of stereochemistry, also possess a 10-member ring enediyne, they contain no methyl trisulfide group to initiate radical formation. Instead, these compounds contain an epoxide and bridging anthraquinone unit that is involved in redox activation, putting them into a different subclass than the other 10-member ring enediynes. With the exception of antibiotic N1999A2,22 the 9-member ring enediynes are molecules that are associated noncovalently with coproduced apoproteins (Figure 22.2) that stabilize the chromophores by binding them in a hydrophobic cavity.23 Although the apoproteins examined so far exhibit significant homology, they differ for each individual antibiotic. For example, the Copyright © 2005 CRC Press, LLC neocarzinostatin, kedarcidin, and lidamycin apoproteins contain 113, 114, and 110 amino acids, respectively.24–26 The x-ray crystal structure of the neocarzinostatin apoprotein was determined to a resolution of 0.15 nm. The structure of this apoprotein was then compared to crystal structure data for the apoproteins of actinoxanthin and auromomycin to show a three-dimensional simi- larity to these homologous proteins.27 The apoproteins stabilize the enediynes in a cellular environment by excluding glutathione from the reaction center through electronic repulsions.28 Supra C-1027, a significantly stabilized analog of the lidamycin chromoprotein, was prepared by incorporation of deuterated glycine into the apoprotein.29 This was accomplished by express- ing the dideutero-Gly apoprotein in Escherichia coli containing a vector with the lidamycin apoprotein coding sequence. The dideutero-Gly apoprotein was then complexed with the chro- mophore to obtain the chromoprotein. Peptidase activity has also been attributed to the apopro- teins, which may help expose the DNA to the enediyne chromophores.30,31 The structures of the enediynes have been determined by spectral methods, mainly two- dimensional nuclear magnetic resonance and mass spectrometry techniques, chemical degrada- tions, and in some cases, x-ray crystallography. Because of the considerable interest in this class of compounds and the challenge of synthesizing these unusual structures, a number of the enediyne antibiotics have been prepared by total syntheses. This has either confirmed the structures, par- ticularly individual points of asymmetry, or resulted in modifications of the proposed structures. Calicheamicin γ I 1 has one of the most complex and elaborate structures, yet interestingly it was the first enediyne antibiotic prepared by total synthesis.32–34 The structure of the kedarcidin chromophore was originally deduced mainly from nuclear magnetic resonance and mass spec- trometry data but was recently revised to the structure shown in Figure 22.2 as a result of

total synthesis studies.35–37 Total syntheses have been achieved for several additional enediynes after their structures were elucidated, including the neocarzinostatin chromophore,20,38 antibiotic N1999A2,22,39,40 and dyne- micin.41–43 Enediyne antibiotics whose structures have been solved but have only led to partial synthesis studies are lidamycin chromophore44,45 and maduropeptin chromophore.46 Thus, the struc- ture of maduropeptin shown in Figure 22.2 is one of several possibilities. Structures for namenam- icin and the shishijimicin complex (Figure 22.1) have been determined47,48 but no synthetic efforts have been reported. Only limited structural information has been published for actinoxanthin and auromomycin chromophores.27,49–51 The apoprotein for auromomycin is macromomycin. Most of the enediyne antitumor antibiotics are produced by microorganisms classified as actinomycetes. Within this general class, however, there is considerable diversity of genus and species for the organisms that produce the natural products (Table 22.1). Recently, macroscopic organisms have been implicated in producing 10-member ring enediynes related to calicheamicin and esperamicin. Marine ascidians were credited for producing namenamicin and shishijimicin.47,48 However, the ascidian that produced namenamicin was later found to coexist with several Micromonospora species in a symbiotic relationship, which led to speculation that one of the Micromonospora sp. may have actually produced the enediynes rather than the ascidian. IV. BIOSYNTHESIS Most of the enediyne antibiotics are generated as complexes of related components by the producing organisms. The various components of the calicheamicin complex — at last count about 20 — differ not only by the halogen on the aromatic ring but also by the presence or absence of the terminal sugars. In addition, the amine of the terminal amino sugar is elaborated with a methyl, ethyl, or isopropyl group. The natural components of the calicheamicin complex contained a bromine atom, which is an unusual halogen for a terrestrial organism. Attempts to improve fermentation yields by adding potassium iodide to the media resulted in a significant improvement and gave new components containing iodine instead of bromine.52 Surprisingly, addition of iodide to the fermen- tation media of esperamicin and dynemicin also resulted in increased fermentation yields even Copyright © 2005 CRC Press, LLC though these products do not contain a halogen.21 This indicates that iodide or one of the other halides may be involved in intermediate steps in the biosynthesis of the enediyne antibiotics. Biosynthetic incorporation studies using labeled acetate and other potential polyketide precur- sors revealed the pathway for assembly of the esperamicin, neocarzinostatin chromophore, and dynemicin biosynthetic precursors.53–56 Cloning, sequencing, and characterization of the complete calicheamicin and lidamycin chromophore biosynthesis gene clusters have been reported.8,57,58 The polyketide synthases for these enediynes are quite unique from previously reported polyketide synthases. A comparison of the lidamycin chromophore locus with that of calicheamicin revealed that the enediyne polyketide synthase is highly conserved in spite of producing 9- and 10-member ring enediynes, respectively. This has led to a proposed general polyketide pathway for the bio- synthesis of both classes of enediyne antibiotics. In addition, a phylogenetic tree has been suggested that relates the synthesis of both 9- and 10-member enediynes and yields a genomic method to screen for other enediynes.59 Various unprecedented enzymes have been uncovered in investigating the biosynthesis of the enediynes. For example, exploration of the biosynthesis of the unique structural elements in calicheamicin has led to various unique enzymes involved in those pathways.60 The gene cluster responsible for the biosynthesis of lidamycin also was found to contain a gene that encodes for a new type of aminomutase that converts L-tyrosine into (S)-β-tyrosine.61 Further biochemical pro- cessing of the (S)-β-tyrosine converts it to the subunit of lidamycin. In addition, the genes that encode the production of the lidamycin chromophore in Streptomyces globisporus are clustered with the cagA gene that encodes the lidamycin apoprotein.62 This suggests that some type of genetic coordination exists between production of the chromophore and the apoprotein in this organism. It has been speculated that the lidamycin chromophore is sequestered by binding to a preapo- protein to form a complex that is transported out of the cell by an efflux pump and processed by removing a leader peptide sequence to yield the chromoprotein.62 This may serve to protect the producing organism from the toxic effects of the lidamycin. Another mechanism has been discovered for calicheamicin that may prove to be a general phenomenon. A protein, CalC, has been discovered that is coproduced with calicheamicin. It binds to the triggered natural product before the Bergman cyclization better than it binds to the intact form, thereby scavenging the triggered form through “self-sacrifice” before it can damage the DNA.63 Previous studies with kedarcidin showed that exposure of the producing organism to the chromophore resulted in significantly upregulated apoprotein production by as much as 10- to 20- fold.51 However, decoupling of the joint biosynthesis of apoprotein and chromophore has also been achieved. When the neocarzinostatin-producing organism was fermented in a synthetic medium containing MgSO4, the free chromophore was observed in the culture filtrate.64 V. MOLECULAR MECHANISMS OF ACTION As a class, the enediynes cause single- and double-strand cleavage of DNA by rearrangement of their core to an aromatic diradical through a Bergman (10-mr) or Bergman-like (9-mr) cycliza- tion.65–67 How the molecules interact with DNA differs somewhat. Calicheamicin is the first known DNA binder that uses a carbohydrate sequence to recognize DNA and binds to the minor groove in a nonintercalative manner. In addition, the carbohydrate portion of calicheamicin has a novel pyrimidine selectivity as a result of a shape-dependent induced-fit interaction.68–72 Esperamicin binds through a combination of carbohydrate-guided minor groove fit as well as base-pair interca- lation, and dynemicin and antibiotic N1999A2 bind through purely intercalative mechanisms.22,39,73 Triggering of the methyl trisulfide-containing enediynes within cells requires a reductive cleav- age of the trisulfide by glutathione and subsequent rearrangements to form the diradical. This chemistry is illustrated for calicheamicin γ I 1 in Figure 22.3. The reduction does not necessarily have to occur with the enediyne prebound to DNA. For example, the pre-Berman cyclization intermediate for calicheamicin γ I 1 has a half-life of about 4 sec.74 However, how the enediyne and product diradical Copyright © 2005 CRC Press, LLC FIGURE 22.3 Diradical formation from calicheamicin γ I 1 /neocarzinostatin and their interaction with DNA. are positioned in the DNA determines the type of damage that occurs. A diradical that spans both strands produces double-strand breaks, whereas a diradical with access to only one strand leads to single-strand breaks.14,74–76 The structure of a calicheamicin γ I 1 -DNA duplex complex was studied in aqueous solution using two-dimensional nuclear magnetic resonance techniques and molecular dynamics calculations. A predominate species of the hairpin duplex complex with DNA containing (TCCT) (AGGA) was observed and gave further insight into the binding and sequence specificity.77 The Bergman cyclization of the dynemicin-like enediynes is the same as illustrated for cali- cheamicin, but the trigger in this case is reduction of the anthraquinone, which allows the electrons on nitrogen to open the epoxide vinylogously through the aromatic ring.78,79 Triggering of the 9-member enediynes involves glutathione, just like most of the 10-member ring enediynes, but in a conjugate addition to the poly-unsaturated system. They all owe their existence to an epoxide functionality (e.g., neocarzinostatin and kedarcidin chromophores) or a bridging chain across the enediyne (e.g., maduropeptin chromophore), or both, that keep the diyne portion of the core separated. The triggering of neocarzinostatin is illustrated in Figure 22.3. As is also true for the 10-mr enediynes, the triggering is achieved by structural changes, such as opening of the epoxide, disruption of the bridging structure, or binding to DNA, that permit the reacting ends of an enediyne to approach within bonding distance of 2.9 to 3.4 Å.67 A review of the interaction of enediynes with DNA has recently been published.80 Lidamycin causes sequence-specific double-strand DNA cleavage.81 However, when oxygen is depleted from the reaction of lidamycin with DNA, sequence-specific covalent DNA drug adducts and DNA Copyright © 2005 CRC Press, LLC interstrand cross links are formed with the drug.82 These types of reactions may be important in the central regions of large tumors where relatively anaerobic conditions prevail. Lidamycin cleaved DNA more efficiently in cells than in a cell-free environment. Mixed single- and double- strand breaks were observed in intracellular episomal, mitochondrial, and genomic DNA at low nanomolar concentrations.83 Calicheamicin γ I 1 cleavage of plasmid DNA was not inhibited by excess tRNA or protein, showing that this enediyne specifically targets DNA.84 The ratio of DNA double-strand breaks to single-strand breaks caused by calicheamicin γ I 1 in cellular DNA of human fibroblasts was 1:3, close to the 1:2 ratio observed with purified plasmid DNA.84 In addition, calicheamicin γ I 1 induced a strong double-strand break response in human fibroblasts at concentrations that would deliver fewer than 1000 molecules per cell. In these and other studies, however, the cells were fully capable of repairing these breaks.85 VI. CELLULAR MECHANISMS OF ACTION Enediynes exert their biological effects through free-radical damage to DNA, and as a result they have been called radiomimetics. It is clear, however, that the cellular response to each of the various enediynes is unique and differs significantly from the effects of radiation. Although the traditional responses to DNA damage such as apoptosis (programmed cell death) have been observed, other types of responses occur as well. The specific cellular response is in part guided by the type of DNA damage induced, but it is also clearly dependent on the inherent make-up of the cells and the resultant tendencies to follow certain biological pathways. Lidamycin appears to cause normal apoptosis in HL-60 leukemia cells86 but abnormal apoptosis at higher concentrations in BEL-7402 hepatoma cells.87 Although laddering of DNA and caspase involvement were seen, the nuclear membrane remained intact, and no apoptotic bodies were observed. This was ascribed to direct DNA shredding by lidamycin preceding the onset of the actual apoptotic signals. Consistent with this, a separate study indicated that the DNA laddering appears to occur before caspase involvement.88 Many of the responses that were ataxia–telangiectasia mutation–dependent for ionizing radiation and bleomycin, such as p53 phosphorylation, were shown to occur by ataxia–telangiectasia muta- tion–independent mechanisms with lidamycin.89 However, Mylotarg® (Wyeth), an antibody conju- gate of calicheamicin, has been shown to cause G2 arrest associated with the ataxia–telangiectasia mutation/Chk1 and Chk2 phosphorylation pathway and caspase-3 mediated apoptosis or G2 arrest without apoptosis.90 Lidamycin, as well as neocarzinostatin, has been reported to inhibit DNA replication by an indirect mechanism involving the hyperphosphorylation of replication protein A.91 Dynemicin A induces apoptosis.92 Curiously though, a series of simplified synthetic analogs of this enediyne appear to induce apoptosis by a non-DNA-dependent mechanism,92 although it is possible that initial DNA damage triggers a secondary mechanism. Other related analogs not capable of generating free radicals were also shown to block apoptosis induced by a variety of unrelated agents, such as the anthracyclines, camptothecin, and Ara-C. Calicheamicin γ I 1 , as well as other members of this family, show different effects on different cell lines. SH-SY5Y N-type neuroblastoma cells undergo apoptosis at subnanomolar concentrations of the compound, whereas SH-EP1 S-type neuroblastoma cells proceed to differentiation.93 The same differential effects were seen for esperamicin A1, dynemicin, and neocarzinostatin. Work in the same labs with this latter agent has demonstrated the involvement of the Bcl-2/caspase-3 pathway in apoptosis of PC12 pheochromocytoma cells.94 The synthetic enediyne calicheamicin θ differs from γ I 1 only by having a thioester replacement for the methyl trisulfide. Thus, hydrolysis activates this analog as opposed to the reductive mech- anism required for the natural calicheamicins. This compound causes apoptosis in Molt-4 leukemia cells, as indicated by chromatin condensation and DNA degradation.95 However, when examined over a broader concentration range in LS neuroblastoma cells, it showed a variety of effects.96 At Copyright © 2005 CRC Press, LLC femtomolar concentrations, cytostasis with upregulation of p53 occurred. At picomolar concentra- tions, it showed signs of typical apoptosis, and at nanomolar concentrations the derivative produced rapid cell lysis. Other studies have indicated that calicheamicin θ–induced apoptosis is p53 inde- pendent and Bax dependent, involving the usual cytochrome C release and caspase-3 and caspase- 9 involvement.97 Upregulation of p53 that correlated with DNA damage has also been reported for calicheam- icin γ I.98 1 Studies with Mylotarg, an antibody conjugate of calicheamicin, on bone marrow samples from acute myeloid leukemia (AML) patients indicated a general lack of intrinsic apoptotic potential in

these cells and a low rate (25%) of caspase involvement after drug expo- sure.99 This may indicate that nonapoptotic mechanisms predominate in this case, or that the inherently low apoptotic potential in many of the samples may be responsible for a lack of response in some patients. One similarity has been noted between radiation damage and the cellular response to esperam- icin and neocarzinostatin. Cytotoxicity for all these treatments shows a dependence on oxygen, making hypoxic cells that occur in the center of solid tumors more resistant.100 Radiosensitive cells are also more responsive to calicheamicin and neocarzinostatin as well as to bleomycin, although no mechanistic details have been investigated.101 Although poly(ADP-ribose) polymerase has been reported to be involved in double-strand break repair caused by radiation, it does not appear to be involved in repairing double-strand breaks caused by neocarzinostatin.102 Similar to radiation sensitivity, the cytotoxicity of one enediyne, calicheamicin in the form of its antibody conjugate Mylotarg, has been shown to be cell-cycle dependent.103 Causing cells to leave G0 phase and cycle by treatment with INF or GM-CSF resulted in more rapid cell death and increased sensitivity. In marked contrast to radiation sensitivity, neocarzinostatin is more cytotoxic to neuroblastoma cells when Bcl-2 is upregulated, even though Bcl-2 normally renders cells resistant to apoptosis.104 This has been correlated to an increase in glutathione in these cells, which is responsible for activating neocarzinostatin.105 Such an increase in sensitivity does not occur for a synthetic enediyne that is activated by a nonreductive pathway. This may not be general for all enediynes or all cell lines. In related work, 6-mercaptodopamine, a thiol taken up by the dopamine pathway, increases the efficacy of neocarzinostatin in neuroblastoma cell lines both in vitro and in vivo.106 VII. PRECLINICAL STUDIES The enediynes are among the most cytotoxic compounds known, but they tend to have rather narrow therapeutic windows. As a class they also exhibit poor pharmacokinetic properties because of their lipophilicity. There has only been limited preclinical pharmacology published for most of the unmodified natural products. Availability of the enediynes themselves has been a deterrent in this regard because fermentation production levels, even after considerable refinement, have remained low compared with most natural products. Studies with semisynthetic derivatives have also been hampered by the same constraints, and even though totally synthetic routes have been developed for many of the natural enediynes, these have been undertaken for the sheer intellectual challenge of assembling these molecules from available starting materials rather than with any sourcing goals in mind. Two notable exceptions exist to these generalities. The first is calicheamicin θ, a totally synthetic derivative further discussed below. The second is dynemicin A, structurally one of the simplest members of the natural enediynes. Even though the total syntheses reported were well over 20 steps each and were still not a practical source for these compounds, they led to a variety of studies designed to identify the minimal structural units necessary for high biological activity. A relatively recent review summarizes both the chemistry and biology of these efforts,107 so only derivatives designed for tumor targeting are summarized here. It has been such attempts at targeting the enediynes, particularly by attaching them to antibodies, that have given this class of natural products significant medicinal promise. Copyright © 2005 CRC Press, LLC Lidamycin has been attached to an RGD-containing nonapeptide. This targeted construct inhibited pulmonary metastases in a PG giant-cell lung carcinoma model.108 Antibody conjugates have also been reported and were made with a monosulfide bond to native antibody (3G11) or a Fab or scFv antibody fragment against type-IV collagenase.109–111 These constructs inhibited KB epidermoid cells at 2 fM and decreased the growth of hepatoma 22 xenograft tumors by 90%. An antigastric cancer antibody, 3H11, has also been used for targeting this enediyne.112 The conjugate had an IC50 of 0.02 fM with 400-fold selectivity versus a nontargeting control conjugate and was active against BGC823 gastric cell xenografts. Dynemicin A has been examined by a different strategy. ADEPT and GDEPT are approaches that use a less cytotoxic prodrug that is activated at the tumor site by an enzyme that is either pretargeted by an antibody or expressed by an introduced or induced gene, respectively. A simplified version of dynemicin missing the anthraquinone and containing an alternate carbamate activating group has been reported.113 The nitrobenzyl carbamate is activated by nitroreductases.114 This derivative showed a 90-fold increase in potency in the presence of the enzyme, and it was also 21- to 135-fold more cytotoxic to cell lines that express a nitroreductase.115 Related simplified versions of dynemicin have been attached to lexitropsins and were cytotoxic to a panel of cancer cell lines in the 20–50 µM range.116 Nuclear uptake was indicated by confocal microscopy. Calicheamicin and esperamicin have been used to make numerous targeted constructs. Espe- ramicin has been conjugated to alpha-fetoprotein to make a conjugate that showed activity both in vitro and in vivo, where subcutaneous P388 tumors were completely eliminated in over 90% of the animals.117 Calicheamicin has also been conjugated to the same protein.118 This conjugate was cytotoxic to the QOS T-lymphoma cell line at 0.15 nM and was more potent than calicheamicin itself. The structure of the actual alpha-fetoprotein-calicheamicin conjugate is not evident from these reports, however. The conjugation methodology and structures reported are not entirely compatible with what is known about the chemistry of calicheamicin, though a disulfide-linked construct attached through lysines can be assumed. An antisense oligodeoxynucleotide conjugate of calicheamicin has also been prepared,119 but no further details are available. Nearly all of the antibody conjugates prepared from calicheamicin analogs have been based on semisynthetic modification of the natural products themselves. An N-acetyl derivative of calicheam- icin γ I 1 has been used most extensively for preparing such constructs, even though this derivative is about one order of magnitude less potent than the corresponding parent molecule.120 An unusual displacement of the methyl trisulfide functionality by thiols has been used to introduce linkers for eventual antibody conjugation.121 This transformation has had three beneficial consequences. First, conversion to a disulfide stabilizes the molecule. Second, a wide variety of thiols is useful for this displacement, allowing further tailoring of how readily the reductive activation step can occur, modulating the stability of the construct in serum, and allowing introduction of a functional group for antibody conjugation.121 And third, the cleavage of the disulfide releases calicheamicin γ I 1 in “triggered” form, significantly limiting the potential half-life of the resultant triggered intermediates.74 For conjugates of N-acetyl calicheamicin γ I 1 possessing stabilized (“α,α-dimethyl”) disulfides, three different methods of making conjugates have been reported. The first, direct conjugation to aldehyde groups generated by oxidation of carbohydrate residues on the antibodies, gives hydrazone linkages that are capable of releasing drug by mild acid hydrolysis (“carbohydrate” conjugates). The second method is the conjugation of activated esters to lysines on the antibodies to produce constructs with only the disulfide available as a point of drug release (“amide” conjugates). The third method also attaches drug to the lysines, but these conjugates contain an added hydrolytic site of release by virtue of a hydrazone linkage in the center of the linker (“hybrid” conjugates). The carbohydrate-based conjugates were the first to be reported and were capable of curing a number of xenograft tumors.122 Two specific examples of the calicheamicin amide conjugates have been published. One was made from an anticanine lymphoma antibody designated 231.123 This conjugate was curative of canine lymphoma xenografts in mice with no apparent lethality, and all control groups (antibody, Copyright © 2005 CRC Press, LLC free calicheamicin derivative, and admixtures of the two) were essentially inactive. The second was a conjugate of the anti-PEM (MUC1) antibody CT-M-01. This conjugate, CMB-401, was curative of various solid tumor xenografts.124 In addition, CMB-401 was superior to the corresponding carbohydrate-based conjugate against multidrug resistant (MDR)-resistant cell lines.125 Three examples of “hybrid” calicheamicin conjugates have been reported. One is Mylotarg (formerly CMA-676), the first Food and Drug Administration–approved antibody-targeted chemo- therapeutic agent that recognizes the CD33 antigen expressed on differentiated myeloid cells and is used to treat AML.126,127 N-acetyl-calicheamicin γ I 1 was converted to the disulfide derivative of α,α-dimethylpropionyl hydrazide, and this in turn was reacted with 4-(4-acetylphenoxy)-butyric acid (AcBut). Activation of the carboxylic acid as a hydroxysuccinimide ester followed by conju- gation to the antibody hP67.6 produced Mylotarg. The product, containing the AcBut “hybrid” linker, is relatively stable in circulation but releases the calicheamicin derivative in the acidic environment of the lysosomes.128 In vitro, Mylotarg was superior to the corresponding amide conjugate versus HL-60 promye- locytic leukemia cells125 and was also an improvement over the corresponding carbohydrate con- jugate.128 It showed an IC50 in cell culture against HL-60 cells of 0.46 pg of calicheamicin equivalents per milliliter, was 78,000-fold selective versus control cells that were equally sensitive to N-acetyl calicheamicin γ I 1 , and was curative of HL-60 xenografts in mice over a wide dose range. Clinical support studies have shown that sensitivity to Mylotarg is cell cycle dependent and that some internalization can occur through nonspecific endocytosis.129 Such studies have also largely confirmed the intended mechanism of CD33 targeting and internalization.130 The other two “hybrid” calicheamicin conjugates differ primarily in the antibody used to produce the constructs. CMC-544 targets the CD22 lymphoid antigen, makes use of the G544 antibody, and is to B-cell lymphomas what Mylotarg is to leukemia.131 It had an IC50 of 10–500 pM against a panel of lymphoma cell lines and was 20-fold more potent than a control conjugate made from a nonbinding antibody. It produced long-term, tumor-free survivors in various xenograft models and was active in a disseminated model of B-cell lymphoma. The other conjugate targets the LewisY antigen and uses the antibody designated Hu3S193.132 This conjugate was more effica- cious in several different solid-tumor xenograft models (gastric, colon, and prostate) than a control conjugate, producing cures in the gastric model. Four conjugates of calicheamicin θ have been reported that target three different antigens. This enediyne was also linked to antibodies through a disulfide-containing spacer, but attachment was through the amino sugar residue, which was first modified with N-hydroxysuccinimido-3-(2- pyridyldithio)propionate. This activated intermediate was then reacted with antibodies that had been modified with iminothiolane to produce the conjugates.133 Two of these conjugates targeted gan- glioside GD2 expressed on neuroblastomas.134 One was produced using the chimeric CH 14.18 antibody, where IC50 values against three different cell lines ranged between 0.1 and 6 nM. The second, more thoroughly studied, conjugate was prepared from antibody 14G2a.133 This conjugate exhibited an IC50 against NXS2 cells of around 0.1 pM, or about 100-fold less cytotoxic than calicheamicin θ itself, and had some activity in a model of neuroblastoma metastatic to the liver. No data were reported for nontargeting conjugates or control cell lines. An anti-CD19 conjugate of calicheamicin θ has also been described.135 The IC50 was between 1 and 100 pM and tripled the life span of mice implanted with the BCR/Abl-positive pre-B cell leukemia Nalm-6. Last, an antirenal gamma-glutamyl transferase conjugate made using antibody 138H11 had an IC50 of 50 pM against Caki-1 cells, with 40-fold selectivity over a control conjugate and a control cell line.136 Suppression of tumor growth for at least 3 week was seen in a xenograft model of Caki-1 cells at the maximum tolerated dose, where some weight loss was also seen. At a half-maximal dose, a growth delay of 2 wk was seen, whereas no significant activity was observed with a nonbinding control conjugate. Mild reduction of the same antibody cleaved some of the surface-accessible disulfide bonds to generate thiols. A calicheamicin θ conjugate has also been constructed through these thiols, showing an IC50 of 16 nM against Caki-2 cells.137 Copyright © 2005 CRC Press, LLC Numerous publications describe neocarzinostatin conjugates, all prepared with the chromophore inside the apoprotein, using the antiadenocarcinoma antibody A7. These have included not only the whole antibody as the targeting unit but also smaller antigen-recognizing fragments such as F(ab)2s and Fabs, as well. All such constructs have been prepared using traditional N-hydroxysuc- cinimido-3-(2-pyridyldithio)propionate (SPDP)–based cross-linking strategies. Whole-antibody conjugates of A7 and neocarzinostatin were active in a xenograft model of pancreatic cancer when injected either intravenously or intratumorally.138 The same conjugate also showed activity against gastric cell lines.139 In vitro, the conjugate was 10-fold more active against antigen positive MKN45

cells than antigen negative MKN1 control cells with an IC50 of 70 ng/mL in neocarzinostatin equivalents. Tumor nodule formation was inhibited by 90% in a disseminated intraperitoneal model. However, the effect was not significantly different than that seen with neocarzinostatin alone, although the diminished toxicity data indicated a wider therapeutic window for the conjugate. This was confirmed in a subsequent report where the MTD was improved by about fivefold, with the conjugate being less toxic, even though the significantly better pharmacokinetic properties would be expected to increase the total exposure to neocarzinostatin.140 More than twice as much antibody A7 and its conjugate localized to xenografts compared to a control antibody, yet both localized to an antigen negative tumor in equal amounts. A neocarzinostatin conjugate prepared from Fab fragments of A7 inhibited colony formation of antigen positive colon141 and pancreatic142 cells in an antigen-specific manner, but the specificity was only about twofold. Modest antigen-specific activity was also seen in xenografts of colorectal cancer,143 but complete, antigen-dependent tumor suppression was seen in a pancreatic xenograft model144 and was greater than the response seen with a whole-antibody conjugate.145 The tumor-targeting ability of the Fab construct has been studied relative to whole antibody and a F(ab)2 construct using 125I radiolabeling.146–148 The Fab conjugate targeted the tumor faster, and more radiolabel was associated with the implant at the earlier time points. Of the three, whole antibody showed the least amount of tumor targeting at 2 h,148 so even though Fab fragments are usually considered to be cleared too rapidly for drug targeting, this may be appropriate for neocarzi- nostatin, which itself has a relatively short half-life in circulation. When this Fab conjugate (chFabA7- neocarzinostatin) was administered to cancer patients, kidney toxicity was observed,141,143 probably because of the rapid clearance in general of Fab fragments through the kidneys. In an approach involving targeting through an avidin–biotin system, both neocarzinostatin and an anti-CEA antibody were first biotinylated. Cells were treated with the modified antibody, followed by avidin, and finally the modified neocarzinostatin.149,150 The multivalent avidin protein binds tightly to both biotinylated molecules stepwise in sandwich fashion. When colon adenocar- cinoma cells were treated for 72 h, no difference was seen with or without the biotinylated antibody. However, when the cells were washed 7 min after the final exposure, a fivefold increase was seen in the cytotoxicity with the biotinylated antibody compared to controls.151 TES-23 is an antibody raised against rat tumor–derived endothelial cells.152 As such, it targets tumor vasculature in human xenografts about 20 times more effectively than a control antibody. A neocarzinostatin conjugate was active in suppressing KMT-17 fibrosarcoma tumors in rats (the same tumor used for making the antibody) at 17 µg/kg, injected three times.152 Tumor hemorrhagic necrosis was observed. Less toxicity was seen with the conjugate, even though the activity was not significantly different from neocarzinostatin itself. Antigen-specific distribution of the conjugate to KMT-17 rat and Sarcoma-180 murine tumors was demonstrated,153 as well as to Meth-A and Colon 26 murine tumors.154 The conjugate was active against Meth-A tumors implanted subcutaneously in mice with three out of four survivors in a group treated with 50 µg/kg of neocarzinostatin equivalents.154 Four other antibodies have been conjugated to neocarzinostatin.155 These were prepared using thiol–maleimide coupling to give a monosulfide linkage instead of the disulfides produced by SPDP–based couplings. The antibodies used were 791T/36 (anti-gp72, targeting osteosarcoma and colorectal carcinoma), BW 431/26 (CEA, targeting colon and stomach cancer), SWA11 (anti-SCLC- Ag cluster w4/gp45, targeting small cell lung cancer), and OV-TL3 (anti-OA3/p85, targeting ovarian Copyright © 2005 CRC Press, LLC carcinoma). The in vitro selectivity of the first three conjugates relative to antigen-negative cells was low, whereas the last one was about eight-fold selective over background. The potency of these conjugates ranged from 16 to 5100 ng/mL. Neocarzinostatin has also been conjugated to transferrin, in this case, through a monosulfide between dithiolane and a maleimide.156 An increase in cytotoxicity was seen compared to neocarzi- nostatin alone; however, additional experiments with modified incubation conditions and with excess transferrin indicated that the effects were not caused by internalization of the conjugate but by neocarzinostatin that had disassociated from the conjugate before cell entry. VIII. CLINICAL EXPERIENCE Esperamicin A1 has been examined in phase I and II clinical trials for the treatment of advanced adenocarcinoma of the pancreas.157–159 No clearly demonstrable cases of response were seen in the 14 patients treated. Nausea/vomiting, malaise/fatigue, thrombocytopenia, and hepatic toxicity were the most common side effects. Neither calicheamicin γ I 1 itself nor other members of this family have been examined in the clinic. Instead, all the clinical experience regarding this enediyne has been based on antibody- targeted forms. Clinical data are available for two such conjugates. CMB-401, which targets the PEM (MUC-1) antigen with the hCT-M-01 antibody, has been evaluated through phase II clinical trials. An imaging study was conducted first with 111In labeled antibody to confirm tumor targeting in ovarian cancer.160 A 7.6-fold increase of radioactivity in tumor versus normal tissues was seen at 0.1 mg/kg protein dose, and a 14-fold increase was seen at 1 mg/kg. In a phase I trial of CMB- 401, 34 patients were pretreated with 35 mg/m2 of unmodified hCT-M-01 antibody to clear circu- lating antigen, followed by doses of the calicheamicin conjugate ranging from 2 to 16 mg/m2.161 The maximum tolerated dose was defined by malaise, hematological toxicities, and gastrointestinal hemorrhage. Four patients had >50% decrease in CA125 levels, and three patients showed reduction in tumor bulk. In this trial, the half-life of the conjugate was estimated to be 36 h, and up to four cycles of treatment led to no allergic reactions with only three increases in antibody titer.162 Ovarian cancer patients who had relapsed after standard chemotherapy were treated with 35 mg/m2 of native antibody followed by 16 mg/m2 of CMB-401 in a phase II trial.163 Of the 19 evaluable patients, only four had >50% decreases in CA125 levels, and there were no objective responses. Two reasons cited for possible failure were complications arising from the attempts to clear circulating antigen and from use of an amide-based linking system for calicheamicin, relying on only the reductive mechanism of drug release. There are no data given to support either of these hypotheses, and numerous other possibilities certainly exist. Mylotarg (gemtuzumab ozogamicin) is the first antibody-targeted chemotherapeutic agent reg- istered for humans. It was approved by the Food and Drug Administration in May 2000 for use in relapsed AML in patients over the age of 60 yr who no longer respond to conventional chemother- apy.126,127 Approval was based on phase II data, in which about 30% of patients achieved a complete remission. Myelosuppression was the dose-limiting toxicity, along with transient, reversible liver effects. A common infusional syndrome seen with most antibody therapies was also observed. Mylotarg is significantly less toxic than traditional chemotherapy, which has resulted in a savings regarding hospitalization time and the associated costs.164,165 However, a low incidence of veno- occlusive disease, an unusual and sometimes fatal toxicity, has been seen with Mylotarg. This may be CD33 antigen related.166 The occurrence of veno-occlusive disease appears to depend on the history of the patient as well as other coadministered therapies.167-169 Reviews of Mylotarg efficacy and toxicities have recently been published.170,171 The lack of a universal response to Mylotarg in AML may have various causes. Some possi- bilities include P-glycoprotein expression,172–174 other efflux pumps,175 variations in CD33 lev- els,176–178 and in apoptosis.179 Copyright © 2005 CRC Press, LLC Although AML most commonly affects the elderly, Mylotarg is also being examined for pediatric AML as well as acute lymphoblastic leukemia.180,181 The drug has been used before hematopoietic stem cell transplant182 and appears to be useful in relapse after hematopoietic stem cell transplant.183,184 However, the occurrence of veno-occlusive disease may be increased if Mylo- targ is used pre–hematopoietic stem cell transplant 185 or if sufficient time has not elapsed since the stem cell transplant.186 Mylotarg has also been evaluated as a treatment for myelodysplastic syndrome, but with less than encouraging results so far.187 However, there is an indication of benefit when it is used in CD33-positive acute lymphoblastic leukemia patients.188 Mylotarg has been evaluated in combinations with various high-dose regimens involving tra- ditional anticancer agents. These include cytarabine,189,190 all-trans retinoic acid,191 IL-11,192,193 topotecan/Ara-C,194,195 fludarabine/Ara-C/G-CSF/idarubicin,196 troxatyl,197 daunorubicin/Ara-C with or without thioguanidine,198 Ara-C/mitoxantrone/amifostine,199 daunorubicin/Ara-C/thioguani- dine/mitoxantrone/etoposide,169 mitoxantrone/Ara-C/etoposide,200 imatinib,201 fludarabine/Ara- C/cyclosporin,202 Ara-C/idarubicin,203,204 and daunorubicin/Ara-C/cyclosporin.205 Although most of the data on these combinations are preliminary, these studies have the potential to increase the usefulness of Mylotarg. Neocarzinostatin has been examined in the clinic both by itself and as several forms of polymer- based conjugates. The natural product is approved for a number of cancers in Japan, but not elsewhere. These cancers include stomach, pancreas, liver, bladder, and brain, as well as leukemia. The main side effect with neocarzinostatin is myelosuppression. More recently it failed to outperform intravesical Bacille Calmette-Guerin therapy for recurrent superficial bladder cancer in combination with cytarabine.206 However, it has shown some promise in refractory or recurrent AML in combi- nation with enocitabine and aclarubicin, with 57% of patients achieving complete remission.207 Conjugates of neocarzinostatin and the anticolon antibody A7 have been in clinical trials. Patients given between 15 and 90 mg of whole-antibody conjugate showed no adverse effects; three out of eight patients with liver metastases secondary to colon cancer showed a decrease in tumor size, and three experienced pain relief, but this treatment had no effect on other metastases.208 However, most of these patients mounted a human anti-mouse response against the murine antibody. Experiments with sera from these patients showed that the human anti-mouse response blocked conjugate cytotoxicity.209 Styrene maleic acid copolymer neocarzinostatin conjugate is approved in Japan for patients with liver cancer or liver metastases and is being explored for cancers of the kidney, lung, and pancreas. It is coadministered as a mixture with Lipiodol, a heavily iodinated poppy seed oil. Recent reviews covering the discovery, development, and clinical use of this therapy are available.210–213 A somewhat related approach that has not progressed to clinical trials yet is the use of chitin derivatives to deliver neocarzinostatin.214 IX. SUMMARY The isolation and preliminary biological activity of neocarzinostatin as an anticancer agent was first reported in 1965.215 During subsequent years, the characterization and x-ray crystal structure of this natural product were reported, and the mechanism of DNA cleavage was elucidated. A variety of conjugates were prepared, including styrene maleic acid copolymer neocarzinostatin conjugate as well as various antibody-based constructs, in an effort to overcome the short biological half-life of this compound. Clinical trials were initiated with neocarzinostatin, in Japan as well as in the United States, yet these studies were all done with the conviction that neocarzinostatin was, in fact, an anticancer protein. It was not until 1979 that the protein was shown to be associated with a much smaller chromophoric subunit that was responsible for the observed biological activ- ity,216 and it took another 6 yr before an enediyne macrocyclic structure was assigned to the chromophore, though without much of the stereochemistry.20 The last chapter relating to the Copyright © 2005 CRC Press, LLC structure of the molecule was closed in 1998, with the report of the total synthesis of the neocarzi- nostatin chromophore.38 In the meantime, similarities between the neocarzinostatin apoprotein and several other “anti- cancer proteins” soon led to work that showed that some of these sequestered enediynes as well. However, these early enediynes were soon upstaged by the more stable 10-member ring analogs: isolation and biological activity for esperamicin was published in 1985217 and its structure in 1987,11 along with that of calicheamicin γ I 1 the same year.13 The complex and extremely novel structures of these compounds captured the attention and resources of many major academic total synthesis groups in a race to be first to construct these molecules. Biochemists and biologists helped elucidate the activity and the mechanisms of action for these compounds. As more members of this family were isolated and their structures were assigned, the mechanisms and pathways by which micro- organisms assembled these molecules became an important and ongoing challenge. What remained, however, was to make the enediynes into useful drugs. In spite of extreme potency, their physical characteristics, their short half-lives, and their small therapeutic windows precluded using most of these natural products as drugs themselves. Neocarzinostatin and cali-

cheamicin γ I 1 were eventually converted to anticancer agents useful in humans via conjugation to a polymer and an antibody, respectively. We have tried to demonstrate in this review that the research into enediynes continues in an effort to understand the biology of these molecules better and thereby expand their clinical applications. REFERENCES 1. Enediyne Antibiotics as Antitumor Agents, Borders, D.B. and Doyle, T.W., Eds., Marcel Dekker, New York, 1995. 2. Lee, M.D., Ellestad, G.A., and Borders, D.B., Calicheamicins: Discovery, structure, chemistry, and interaction with DNA, Acc. Chem. Res., 24, 235, 1991. 3. Nicolaou, K.C. and Dai, W.M., Chemistry and biology of the enediyne anticancer antibiotics, Angew. Chem., Int. Ed. Engl., 30, 1387, 1991. 4. Smith, A.L. and Nicolaou, K.C., The enediyne antibiotics, J. Med. Chem., 39, 2103, 1996. 5. Nicolaou, K.C., Sorensen, E.J., and Winssinger, N., The art and science of organic and natural products synthesis, J. Chem. Educ., 75, 1225, 1998. 6. Xi, Z. and Goldberg, I.H., DNA-damaging enediyne compounds, in Comprehensive Natural Products Chemistry, Barton, D.S., Nakanishi, K., and Meth-Cohn, O., Eds., Pergamon Press, Oxford, UK, 1999, 7, 553. 7. Jones, G.B. and Fouad, F.S., Designed enediyne antitumor agents, Curr. Pharmaceut. Design, 8, 2415, 2002. 8. Shen, B., Liu, W., and Nonaka, K., Enediyne natural products: Biosynthesis and prospect towards engineering novel antitumor agents, Curr. Med. Chem., 10, 2317, 2003. 9. Neocarzinostatin: The Past, Present and Future of an Anticancer Drug, Maeda, H., Edo, K., and Ishida, N., Eds., Springer, New York, 1997. 10. Golik, J. et al., Esperamicins, a novel class of potent antitumor antibiotics. 2. Structure of esperamicin X., J. Am. Chem. Soc., 109, 3461, 1987. 11. Golik, J. et al., Esperamicins, a novel class of potent antitumor antibiotics. 3. Structures of esperamicins A1, A2, and A1b, J. Am. Chem. Soc., 109, 3462, 1987. 12. Lee, M.D. et al., Calichemicins, a novel family of antitumor antibiotics. 1. Chemistry and partial structure of calichemicin γ I 1 , J. Am. Chem. Soc., 109, 3464, 1987. 13. Lee, M.D. et al., Calichemicin, a novel family of antitumor antibiotics. 2. Chemistry and structure of calichemicin γ I 1 , J. Am. Chem. Soc., 109, 3466, 1987. 14. Zein, N. et al., Calichemicin γ I 1 : An antitumor antibiotic that cleaves double-stranded DNA site specifically, Science, 240, 1198, 1988. 15. Lee, M.D. et al., Calicheamicins, a novel family of antitumor antibiotics. 4. Structure elucidation of calicheamicins β Br 1 , γ Br 1 , α I 2 , α I I 3 , β1 , γ I 1 , and δ I 1 , J. Am. Chem. Soc., 114, 985, 1992. Copyright © 2005 CRC Press, LLC 16. Golik, J.A. et al., Antitumor Antibiotics, U.S. patent 5,086,045, 1992. 17. White, R.J., Maiese, W.M., and Greenstein, M., in Manual of Industrial Microbiology and Biotech- nology, Demain, A.L. and Solomon, N.A., Eds., American Society of Microbiology, Washington, DC, 1986, 24. 18. Lee, M.D. et al., Calichemicins, a novel family of antitumor antibiotics. 3. Isolation, purification and characterization of calicheamicins β Br, γ Br I 1 , α I 2 , α3 , β I I 1 , γ1 , and δ I 1 1 , J. Antibiot., 42, 1070, 1989. 19. Maiese, W.M. et al., Calicheamicins, a novel family of antitumor antibiotics: Taxonomy, fermentation and biological properties, J. Antibiot., 42, 558, 1989. 20. Edo, K. et al., The structure of neocarzinostatin chromophore possessing a novel bicyclo-7.3.0- dodecadiyne system, Tetrahedron Lett., 26, 331, 1985. 21. Lam, K.S. and Forenza, S., Fermentation and isolation of esperamicins, in Enediyne Antibiotics as Antitumor Agents, Borders, D.B. and Doyle, T.W., Eds., Marcel Dekker, New York, 1995, chap. 10. 22. Ando, T. et al., A new non-protein enediyne antibiotic N1999A2: Unique enediyne chromophore similar to neocarzinostatin and DNA cleavage feature, Tetrahedron Lett., 39, 6495, 1998. 23. Takashima, H., Amiya, S., and Kobayashi, Y., Neocarzinostatin: Interaction between the antitumor- active chromophore and the carrier protein, J. Biochem., 109, 807, 1991. 24. Hofstead, S.J. et al., Kedarcidin, a new chromoprotein antitumor antibiotic. II. Isolation, purification and physico-chemical properties, J. Antibiot., 45, 1250, 1992. 25. Kim, K.H. et al., Crystal structure of neocarzinostatin, an antitumor protein-chromophore complex, Science, 262, 1042, 1993. 26. Okuno, Y., Otsuka, M., and Sugiura, Y., Computer modeling analysis for enediyne chromophore- apoprotein complex of macromolecular antitumor antibiotic C-1027, J. Med. Chem., 37, 2266, 1994. 27. Teplyakov, A. et al., Crystal structure of apo-neocarzinostatin at 0.15-nm resolution, Eur. J. Biochem., 213, 737, 1993. 28. Chin, D.H., Rejection by neocarzinostatin protein through charges rather than sizes, Chem.-Eur. J., 5, 1084, 1999. 29. Usuki, T. et al., Rational design of supra C-1027: Kinetically stabilized analog of the antitumor enediyne chromoprotein, J. Am. Chem. Soc., 126, 3022, 2004. 30. Zein, N. et al., Selective proteolytic activity of the antitumor agent kedarcidin, Proc. Natl. Acad. Sci. U.S.A., 90, 8009, 1993. 31. Sakata, N. et al., Aminopeptidase activity of an antitumor antibiotic, C-1027, J. Antibiot., 45, 113, 1992. 32. Nicolaou, K.C. et al., Total synthesis of calicheamicin γ I 1 : 3. The final stages, J. Am. Chem. Soc., 115, 7625, 1993. 33. Hitchcock, S.A. et al., A convergent total synthesis of calicheamicin γ I 1 , Angew. Chem., Int. Ed. Engl., 33, 858, 1994. 34. Nicolaou, K.C. et al., Total synthesis of calicheamicin γ I 1 , J. Am. Chem. Soc., 114, 10082, 1992. 35. Leet, J.E. et al., Kedarcidin, a new chromoprotein antitumor antibiotic: Structure elucidation of kedarcidin chromophore, J. Am. Chem. Soc., 114, 7946, 1992. 36. Leet, J.E. et al., Kedarcidin chromophore: Structure elucidation of the amino sugar kedarosamine, Tetrahedron Lett., 33, 6107, 1992. 37. Kawata, S. and Hirama, M., Synthetic study of the kedarcidin chromophore: Efficient construction of the aryl alkyl ether linkage, Tetrahedron Lett., 39, 8707, 1998. 38. Myers, A.G. et al., Total synthesis of (+)-neocarzinostatin chromophore, J. Am. Chem. Soc., 120, 5319, 1998. 39. Kobayashi, S. et al., The first total synthesis of N1999-A2: Absolute stereochemistry and stereochem- ical implications into DNA cleavage, J. Am. Chem. Soc., 123, 11294, 2001. 40. Kobayashi, S. et al., Investigation of the total synthesis of N1999-A2: Implication of stereochemistry, J. Am. Chem. Soc., 123, 2887, 2001. 41. Konishi, M. et al., Crystal and molecular structure of dynemicin A: A novel 1,5-diyn-3-ene antitumor antibiotic, J. Am. Chem. Soc., 112, 3715, 1990. 42. Myers, A.G. et al., A convergent synthetic route to (+)-dynemicin A and analogs of wide structural variability, J. Am. Chem. Soc., 119, 6072, 1997. 43. Shair, M.D. et al., The total synthesis of dynemicin A leading to development of a fully contained bioreductively activated enediyne prodrug, J. Am. Chem. Soc., 118, 9509, 1996. Copyright © 2005 CRC Press, LLC 44. Yoshida, K.-I. et al., Structure and cycloaromatization of a novel enediyne, C-1027 chromophore, Tetrahedron Lett., 34, 2637, 1993. 45. Iida, K.-I. et al., Synthesis and absolute stereochemistry of the aminosugar moiety of antibiotic C- 1027 chromophore, Tetrahedron Lett., 34, 4079, 1993. 46. Schroeder, D.R. et al., Isolation, structure determination and proposed mechanism of action for artifacts of maduropeptin chromophore, J. Am. Chem. Soc., 116, 9351, 1994. 47. McDonald, L.A. et al., Namenamicin, a new enediyne antitumor antibiotic from the marine ascidian Polysyncraton lithostrotum, J. Am. Chem. Soc., 118, 10898, 1996. 48. Oku, N., Matsunaga, S., and Fusetani, N., Shishijimicins A-C, novel enediyne antitumor antibiotics from the ascidian Didemnum proliferum, J. Am. Chem. Soc., 125, 2044, 2003. 49. Samy, T.S. et al., Primary structure of macromomycin, an antitumor antibiotic protein, J. Biol. Chem., 258, 183, 1983. 50. Suzuki, H. et al., Biological activities of non-protein chromophores of antitumor protein antibiotics: Auromomycin and neocarzinostatin, Biochem. Biophys. Res. Commun., 94, 255, 1980. 51. Doyle, T.W. and Borders, D.B., Enediyne antitumor antibiotics, in Enediyne Antibiotics as Antitumor Agents, Borders, D.B. and Doyle, T.W., Eds., Marcel Dekker, New York, 1995, chap. 1. 52. Fantini, A.A. and Testa, R.T., Taxonomy, fermentation and yield improvement, in Enediyne Antibiotics as Antitumor Agents, Borders, D.B. and Doyle, T.W., Eds., Marcel Dekker, New York, 1995, chap. 3. 53. Hensens, O.D., Giner, J.-L., and Goldberg, I.H., Biosynthesis of NCS Chrom A, the chromophore of the antitumor antibiotic neocarzinostatin, J. Am. Chem. Soc., 111, 3295, 1989. 54. Lam, K.S. et al., Biosynthesis of esperamicin A1, an enediyne antitumor antibiotic, J. Am. Chem. Soc., 115, 12340, 1993. 55. Lam, K.S. et al., Inhibition of esperamicin A1 synthesis by cerulenin, Abstr. Annu. Meet. Am. Soc. Microbiol., 90, 266, 1990. 56. Tokiwa, Y. et al., Biosynthesis of dynemicin A, a 3-ene-1,5-diyne antitumor antibiotic, J. Am. Chem. Soc., 114, 4107, 1992. 57. Ahlert, J. et al., The calicheamicin gene cluster and its iterative type I enediyne PKS, Science, 297, 1173, 2002. 58. Liu, W. et al., Biosynthesis of the enediyne antitumor antibiotic C-1027, Science, 297, 1170, 2002. 59. Wen, L. et al., Rapid PCR amplification of minimal enediyne polyketide synthase cassettes leads to a predictive familial classification model, Proc. Natl. Acad. Sci. U. S. A., 100, 11959, 2003. 60. Thorson, J.S. et al., Understanding and exploiting nature’s chemical arsenal: The past, present and future of calicheamicin research, Curr. Pharmaceut. Design, 6, 1841, 2000. 61. Christenson, S.D. et al., A novel 4-methylideneimidazole-5-one-containing tyrosine aminomutase in enediyne antitumor antibiotic C-1027 biosynthesis, J. Am. Chem. Soc., 125, 6062, 2003. 62. Liu, W. and Shen, B., Genes for production of the enediyne antitumor antibiotic C-1027 in Strepto- myces lobisporus are clustered with the cagA gene that encodes the C-1027 apoprotein, Antimicrob. Agents Chemother., 44, 382, 2000. 63. Biggins, J.B., Onwueme, K., and Thorson, J.S., Resistance to enediyne antitumor antibiotics by CalC self-sacrifice, Science, 301, 1537, 2003. 64. Kudo, K. et al., Production of a free chromophore component of neocarzinostatin (NCS) in the culture filtrate of Streptomyces carzinostaticus var. F-41, J. Antibiot., 35, 1111, 1982. 65. Bergman, R.G., Reactive 1,4-dehydroaromatics, Acc. Chem. Res., 6, 25, 1973. 66. Chen, W.C., Chang, N.Y., and Yu, C.H., Density functional study of Bergman cyclization of enediynes, J. Phys. Chem. A, 102, 2584, 1998. 67. Schreiner, P.R., Cyclic enediynes — relationship between ring size, alkyne carbon distance, and cyclization barrier, Chem. Comm., 483, 1998. 68. Biswas, K. et al., The molecular basis for pyrimidine-selective DNA binding: Analysis of calicheamicin oligosaccharide derivatives by capillary electrophoresis, J. Am. Chem. Soc., 122, 8413, 2000. 69. Walker, S.L., Andreotti, A.H., and Kahne, D.E., NMR characterization of calicheamicin bound to DNA, Tetrahedron, 50, 1351, 1994. 70. Walker, S. et al., Analysis of hydroxylamine glycosidic linkages: Structural consequences of the NO bond in calicheamicin, J. Am. Chem. Soc., 116, 3197, 1994. 71. Paloma, L.G. et al., Interaction of calicheamicin with duplex DNA: Role of the oligosaccharide domain and identification of multiple binding modes, J. Am. Chem. Soc., 116, 3697, 1994. Copyright © 2005 CRC Press, LLC 72. Ikemoto, N. et al., Calicheamicin-DNA complexes: Warhead alignment and saccharide recognition of the minor groove, Proc. Natl. Acad. Sci. U.S.A., 92, 10506, 1995. 73. Miyagawa, N. et al., DNA cleavage characteristics of non-protein enediyne antibiotic N1999-2A, Biochem. Biophys. Res. Commun., 306, 87, 2003. 74. De Voss, J.J., Hangeland, J.J., and Townsend, C.A., Characterization of the in-vitro cyclization chemistry of calicheamicin and its relation to DNA cleavage, J. Am. Chem. Soc., 112, 4554, 1990. 75. Zein, N. et al., Exclusive abstraction of nonexchangeable hydrogens from DNA by calicheamicin γ I 1 , J. Am. Chem. Soc., 111, 6888, 1989. 76. De Voss, J.J. et al., Site-specific atom transfer from DNA to a bound ligand defines the geometry of a DNA-calicheamicin γ I 1 complex, J. Am. Chem. Soc., 112, 9669, 1990. 77. Kumar, R.A., Ikemoto, N., and Patel, D.J., Solution structure of the calicheamicin γ I 1 -DNA complex, J. Mol. Biol., 265, 187, 1997. 78. Sugiura, Y. et al., Reductive and nucleophilic activation products of dynemicin A with methyl thiogly- colate. A rational mechanism for DNA cleavage of the thiol-activated dynemicin A, Biochemistry, 30, 2989, 1991. 79. Nicolaou, K.C. et al., Molecular design, chemical synthesis, kinetic studies, calculations, and biolog- ical studies of novel enediynes equipped with triggering, detection, and deactivating devices: Model dynemicin A epoxide and cis-diol systems, J. Am. Chem. Soc., 115, 7944, 1993. 80. Cosgrove, J.P., and Dedon, P.C., Binding and reaction of calicheamicin and other enediyne antibiotics with DNA, in Small Molecule DNA and RNA Binders, Demeunynck, M., Bailly, C., and Wilson, W.D., Eds., Wiley-VCH, Weinheim, Germany, 2003, 2, 609. 81. Xu, Y.-J., Zhen, Y.-S., and Goldberg, I.H., C1027 chromophore, a potent new enediyne antitumor antibiotics,

induces sequence-specific double-strand DNA cleavage, Biochemistry, 33, 5947, 1994. 82. Xu, Y.J., Zhen, Y.S., and Goldberg, I.H., Enediyne C1027 induces the formation of novel covalent DNA interstrand cross-links and monoadducts, J. Am. Chem. Soc., 119, 1133, 1997. 83. Cobuzzi, R.J., Jr. et al., Effects of the enediyne C-1027 on intracellular DNA targets, Biochemistry, 34, 583, 1995. 84. Elmroth, K. et al., Cleavage of cellular DNA by calicheamicin γ I 1 , DNA Repair, 2, 363, 2003. 85. Susanne, M. et al., Activation of the DNA-dependent protein kinase by drug-induced and radiation- induced DNA strand breaks, Radiation Research, 160, 291, 2003. 86. Jiang, B., Li, D.D., and Zhen, Y.S., Induction of apoptosis by enediyne antitumor antibiotic C1027 in HL-60 human promyelocytic leukemia cells, Biochem. Biophys. Res. Commun., 208, 238, 1995. 87. He, Q. and Li, D., Interaction of direct breaking DNA strands and initiating of apoptotic pathway contributed to high potent cytotoxicities of enediyne antibiotic lidamycin (C1027) to human tumor cells, Proc. Am. Assoc. Cancer Res., 42, 643, 2001. 88. Wang, Z. et al., Non-caspase-mediated apoptosis contributes to the potent cytotoxicity of the enediyne antibiotic lidamycin toward human tumor cells, Biochem. Pharmacol., 65, 1767, 2003. 89. Dziegielewski, J. and Beerman, T.A., C-1027-induced activation of DNA-damage responses is ATM- independent, Proc. Am. Assoc. Cancer Res., 42, 199, 2001. 90. Amico, D. et al., Differential response of human acute myeloid leukemia cells to gemtuzumab ozogamicin in vitro: Role of Chk1 and Chk2 phosphorylation and caspase 3, Blood, 101, 4589, 2003. 91. McHugh, M.M. et al., The cellular response to DNA damage induced by the enediynes C-1027 and neocarzinostatin includes hyperphosphorylation and increased nuclear retention of replication protein A (RPA) and trans inhibition of DNA replication, Biochemistry, 40, 4792, 2001. 92. Hiatt, A. et al., Regulation of apoptosis in leukemic cells by analogs of dynemicin A, Bioorg. Med. Chem., 2, 315, 1994. 93. Hartsell, T.L. et al., Determinants of the response of neuroblastoma cells to DNA damage: The roles of pre-treatment cell morphology and chemical nature of the damage, J. Pharmacol. Exp. Ther., 277, 1158, 1996. 94. Liang, Y. et al., Early events in Bcl-2-enhanced apoptosis, Apoptosis, 8, 609, 2003. 95. Nicolaou, K.C. et al., Calicheamicin θI 1: A rationally designed molecule with extremely potent and selective DNA cleaving properties and apoptosis inducing activity, Angew. Chem., Int. Ed. Engl., 33, 183, 1994. 96. Wrasidlo, W. et al., Modulation of p53 expression in neuroblastoma cell lines by the DNA cleaving agent calicheamicin theta, Proc. Am. Assoc. Cancer Res., 38, 610, 1997. Copyright © 2005 CRC Press, LLC 97. Prokop, A. et al., Induction of apoptosis by enediyne antibiotic calicheamicin θ proceeds through a caspase-mediated mitochondrial amplification loop in an entirely Bax-dependent manner, Oncogene, 22, 9107, 2003. 98. Siegel, J. et al., Enhanced p53 activity and accumulation in response to DNA damage upon DNA transfection, Oncogene, 11, 1363, 1995. 99. Andreeff, M. et al., Caspase independent cell death in AML: An in vivo study in patients undergoing chemotherapy, Blood, 98, 208b, 2001. 100. Batchelder, R.M. et al., Oxygen dependence of the cytotoxicity of the enediyne anti-tumour antibiotic esperamicin A-1, Br. J. Cancer, 74, S52, 1996. 101. Van Duijn-Goedhart, A. et al., Differential responses of Chinese hamster mutagen sensitive cell lines to low and high concentrations of calicheamicin and neocarzinostatin, Mutat. Res., 471, 95, 2000. 102. Noel, G. et al., Poly(ADP-ribose) polymerase (PARP-1) is not involved in DNA double-strand break recovery, BMC Cell Biology, http://www.biomedcentral.com/content/pdf/1471-2121-4-7.pdf, 2003. 103. Jedema, I. et al., Sensitivity of acute myeloid leukemia cells to anti-CD33-calicheamicin (gemtuzumab ozogamicin) is cell cycle-dependent and can be increased by interferon gamma, Blood, 98, 120a (abst. 503), 2001. 104. Cortazzo, M. and Schor, N.F., Potentiation of enediyne-induced apoptosis and differentiation by Bcl- 2, Cancer Res., 56, 1199, 1996. 105. Kappen, L.S. and Goldberg, I.H., Activation and inactivation of neocarzinostatin-induced cleavage of DNA, Nucleic Acids Res., 5, 2959, 1978. 106. Schor, N.F. et al., The use of 6-mercaptodopamine to target neuroblastoma cells for enediyne-induced apoptosis and antineoplastic effect, Proc. Am. Assoc. Cancer Res., 35, 316, 1994. 107. Maier, M.E., Bosse, F., and Niestroj, A.J., Design and synthesis of dynemicin analogs, Eur. J. Org. Chem., 1, 1999. 108. Zhen, Y.S. et al., Lidamycin and its RGD-containing peptide conjugate inhibit angiogenesis and metastasis, Proc. Am. Assoc. Cancer Res., 41, 645 (abst. 4098), 2000. 109. Liu, X.Y. et al., Antitumor effects of novel immunoconjugates with downsized-molecule prepared by linking lidamycin to Fab′ and scFv antibody, Proc. Am. Assoc. Cancer Res., 41, 290 (abst. 1848), 2000. 110. Wang, F., Shang, B., and Zhen, Y., Antitumor effects of the molecule-downsized immunoconjugate composed of lidamycin and Fab′ fragment of monoclonal antibody directed against type IV collage- nase, Sci. China, Ser. C Life Sci., 47, 66, 2004. 111. Feng-Qiang, W., Bo-Yang, S., and Yong-Su, Z., Antitumor effects of the immunoconjugate composed of lidamycin and monoclonal antibody 3G11, Acta Pharmaceutica Sinica, 38, 515, 2003. 112. Zhen, Y.S. et al., Conjugate of antibiotic C1027 and monoclonal antibody 3H11 shows potent antitumor effect against gastric cancer, Proc. Am. Assoc. Cancer Res., 35, 511 (abst. 3047), 1994. 113. Wrasidlo, W. et al., In vivo efficacy of novel synthetic enediynes, Acta Oncol., 34, 157, 1995. 114. Hay, M.P., Wilson, W.R., and Denny, W.A., A novel enediyne prodrug for antibody-directed enzyme prodrug therapy (ADEPT) using E. coli B nitroreductase, Bioorg. Med. Chem. Lett., 5, 2829, 1995. 115. Hay, M.P., Wilson, W.R., and Denny, W.A., Nitrobenzyl carbamate prodrugs of enediynes for nitrore- ductase gene-directed enzyme prodrug therapy (GDEPT), Bioorg. Med. Chem. Lett., 9, 3417, 1999. 116. Xie, Y. et al., Enediyne-lexitropsin DNA-targeted anticancer agents. Physicochemical and cytotoxic properties in human neoplastic cells in vitro, and intracellular distribution, Anti-Cancer Drug Des., 12, 169, 1997. 117. Severin, S.E. et al., Targeted delivery of esperamicin A1 by using oncofetal protein alpha-fetoprotein, Eur. J. Cancer, 33, S89, 1997. 118. Severin, S.E. et al., Alpha-fetoprotein-mediated targeting of anti-cancer drugs to tumor cells in vitro, Biochem. Mol. Biol. Int., 37, 385, 1995. 119. Capson, T. and Ruffner, D., Preparation of calicheamicin-oligodeoxynucleotide conjugates for use as antisense therapeutics, Pharm. Res. (New York), 13, S153, 1996. 120. Durr, F.E. et al., Biological activities of calicheamicin, in Enediyne Antibiotics as Antitumor Agents, Borders, D.B. and Doyle, T.W., Eds., Marcel Dekker, 1995, chap. 8. 121. Ellestad, G.A. et al., Reactions of the trisulfide moiety in calicheamicin, Tetrahedron Lett., 30, 3033, 1989. 122. Hinman, L.M. et al., Preparation and characterization of monoclonal antibody conjugates of the calicheamicins: A novel and potent family of antitumor antibiotics, Cancer Res., 53, 3336, 1993. Copyright © 2005 CRC Press, LLC 123. Jeglum, K.A. et al., Antitumor effects of a calicheamicin immunoconjugate with anti-canine lymphoma monoclonal antibody 231, Proc. Am. Assoc. Cancer Res., 35, 506 (abstr. 3016), 1994. 124. Hinman, L.M. et al., Lysine-linked conjugates of calicheamicin γ I 1 with a fully humanized anti- polymorphic epithelial mucin antibody show potent antitumor effects in ovarian tumor xenografts, Proc. Am. Assoc. Cancer Res., 34, 479 (Abs 2860), 1993. 125. Hamann, P.R. et al., An anti-CD33 antibody-calicheamicin conjugate for treatment of acute myeloid leukemia. Choice of linker, Bioconj. Chem., 13, 40, 2002. 126. Bross, P.F. et al., Approval summary: Gemtuzumab ozogamicin in relapsed acute myeloid leukemia, Clin. Cancer Res., 7, 1490, 2001. 127. Bross, P.F. et al., Approval summary: Gemtuzumab ozogamicin in relapsed acute myeloid leukemia (correction to previous reference), Clin. Cancer Res., 8, 300, 2002. 128. Hamann, P.R. et al., Gemtuzumab ozogamicin, a potent and selective anti-CD33 antibody-calicheam- icin conjugate for treatment of acute myeloid leukemia, Bioconj. Chem., 13, 47, 2002. 129. Jedema, I. et al., Internalization and cell cycle-dependent killing of leukemic cells by gemtuzumab ozogamicin: Rationale for efficacy in CD33-negative malignancies with endocytic capacity, Leukemia, 18, 316, 2004. 130. van der Velden, V.H.J., Berger, M.S., and Van Dongen, J.J.M., Mylotarg therapy in acute myeloid leukemia: Mechanism of action and implications for future treatment protocols, Haematol. Blood Transfusion, 41, 169, 2003. 131. DiJoseph, J.F. et al., Antibody-targeted chemotherapy with CMC-544: A CD22-targeted immunocon- jugate of calicheamicin for the treatment of B-lymphoid malignancies, Blood, 103, 1807, 2004. 132. Boghaert, E.R. et al., Antibody-targeted chemotherapy with the calicheamicin conjugate hu3S193-N- acetyl gamma calicheamicin dimethyl hydrazide targets Lewisy and eliminates Lewisy-positive human carcinoma cells and xenografts, Clinical Cancer Research, 10, 4538, 2004. 133. Lode, H.N. et al., Targeted therapy with a novel enediyne antibiotic calicheamicin θI 1 effectively suppresses growth and dissemination of liver metastases in a syngeneic model of murine neuroblas- toma, Cancer Res., 58, 2925, 1998. 134. Burchardt, C.A. et al., Cytotoxic effects of new immunoconjugates in different neuroblastoma cell- lines, J. Cancer Res. Clin. Oncol., 124, 137, 1998. 135. Bernt, K.M. et al., CD19 targeted calicheamicin theta is effective against high risk ALL, Blood, 96, 467a, 2000. 136. Knoll, K. et al., Targeted therapy of experimental renal cell carcinoma with a novel conjugate of monoclonal antibody 138H11 and calicheamicin θI 1, Cancer Res., 60, 6089, 2000. 137. Schmidt, C.S. et al., Chemoimmunoconjugates with the monoclonal antibody 138H11 for targeting the cytotoxic prodrug calicheamicin θ to renal cell carcinomas, Tumor Target., 4, 271, 1999. 138. Otsuji, E. et al., The effect of intravenous and intra-tumoural chemotherapy using a monoclonal antibody-drug conjugate in a xenograft model of pancreatic cancer, Eur. J. Surg. Oncol., 21, 61, 1995. 139. Okamoto, K. et al., Targeted chemotherapy in mice with peritoneally disseminated gastric cancer using monoclonal antibody-drug conjugate, Cancer Lett., 122, 231, 1998. 140. Kitamura, K. et al., The role of monoclonal antibody A7 as a drug modifier in cancer therapy, Cancer Immunol. Immunother., 36, 177, 1993. 141. Yamaguchi, T. et al., Production, binding and cytotoxicity of human/mouse chimeric monoclonal antibody-neocarzinostatin conjugate, Jpn. J. Cancer Res., 84, 1190, 1993. 142. Otsuji, E. et al., Antitumor effect of neocarzinostatin conjugated to human/mouse chimeric Fab fragments of the monoclonal antibody A7 on human pancreatic carcinoma, J. Surg. Oncol., 57, 230, 1994. 143. Yamaguchi, T. et al., In vivo efficacy of neocarzinostatin coupled with Fab human/mouse chimeric monoclonal antibody A7 against human colorectal cancer, Jpn. J. Cancer Res., 85, 167, 1994. 144. Otsuji, E. et al., Effects of neocarzinostatin-chimeric Fab conjugates on the growth of human pancreatic carcinoma xenografts, Br. J. Cancer, 73, 1178, 1996. 145. Matsumura, H. et al., Increased effects of neocarzinostatin-chimeric Fab conjugates on the growth of human pancreatic carcinoma xenografts, Cancer Biother. Radiopharm., 13, 60, 1998. 146. Otsuji, E. et al., Biodistribution of neocarzinostatin conjugated to chimeric Fab fragments of the monoclonal antibody A7 in nude mice bearing human pancreatic cancer xenografts, Jpn. J. Cancer Res., 85, 530, 1994. Copyright © 2005 CRC Press, LLC 147. Yamaguchi, T. et al., Production of human/mouse chimeric monoclonal antibody-neocarzinostatin conjugate and distribution in tumor bearing athymic mice, Proc. Controlled Release Soc., 22, 544, 1995. 148. Otsuji, E. et al., Applicability of monoclonal antibody Fab fragments as a carrier of neocarzinostatin in targeting chemotherapy, J. Surg. Oncol., 61, 149, 1996. 149. Otsuji, E. et al., Distribution of neocarzinostatin conjugated to biotinylated chimeric monoclonal antibody Fab fragments after administration of avidin, Br. J. Cancer, 74, 597, 1996. 150. Otsuji, E. et al., Decreased renal accumulation of biotinylated chimeric monoclonal antibody-neo- carzinostatin conjugate after administration of avidin, Jpn. J. Cancer Res., 88, 205, 1997. 151. Nakaki, M., Takikawa, H., and Yamanaka, M., Targeting immunotherapy using the avidin-biotin system for a human colon adenocarcinoma in vitro, J. Int. Med. Res., 25, 14, 1997. 152. Tsunoda, S. et al., Specific binding of TES-23 antibody to tumour vascular endothelium in mice, rats and human cancer tissue: A novel drug carrier for cancer targeting therapy, Br. J. Cancer, 81, 1155, 1999. 153. Makimoto, H. et al., Tumor vascular targeting using a tumor-tissue endothelium-specific monoclonal antibody as an effective strategy for cancer chemotherapy, Biochem. Biophys. Res. Commun., 260, 346, 1999. 154. Wakai, Y. et al., Effective cancer targeting using an anti-tumor tissue vascular endothelium-specific monoclonal antibody (TES-23), Jpn. J. Cancer Res., 91, 1319, 2000. 155. Maibucher, A. et al., Neocarzinostatin immunoconjugates: In vitro evaluation of therapeutic potential, Pharm. Sci. Commun., 4, 253, 1994. 156. Schonlau, F. et al., Mechanism of free and conjugated neocarzinostatin activity: Studies on chro- mophore and protein uptake using a transferrin-neocarzinostatin conjugate, Z. Naturforsch., C: Biosci., 52, 245, 1997. 157. Brown, T. et al., A phase II trial of BMY-28175 in advanced pancreatic carcinoma, Proc. Am. Soc. Clin. Oncol., 13, 219, 1994. 158. Melink, T. et al., A phase I trial of BMY-28175 on a single dose schedule, Proc.

Am. Soc. Clin. Oncol., 10, 120, 1991. 159. Sessa, C. et al., Phase I study of BMY-28175 (esperamicin A1) on a single intermittent schedule, Proc. Am. Soc. Clin. Oncol., 9, 66, 1990. 160. Van Hof, A.C. et al., Biodistribution of 111indium-labeled engineered human antibody CTM01 in ovarian cancer patients: Influence of protein dose, Cancer Res., 56, 5179, 1996. 161. Gillespie, A.M. et al., Phase I open study of the effects of ascending doses of the cytotoxic immuno- conjugate CMB-401 (hCTM01-calicheamicin) in patients with epithelial ovarian cancer, Ann. Oncol., 11, 735, 2000. 162. Broadhead, T.J. et al., CMB-401 — a new agent for immunotherapy of epithelial ovarian cancer: Immune response and pharmacokinetics, Br. J. Cancer, 78, 44, 1998. 163. Chan, S.Y. et al., A phase 2 study of the cytotoxic immunoconjugate CMB-401 (hCTM01-calicheam- icin) in patients with platinum-sensitive recurrent epithelial ovarian carcinoma, Cancer Immunol. Immunother., 52, 243, 2003. 164. Mallick, R., Sievers, E., and Berger, M., A decision analysis of expected cost savings with outpatient administration of gemtuzumab ozogamicin (MylotargTM, CMA-676) in patients in first relapse of acute myeloid leukemia (AML), Blood, 96, 439a, 2000. 165. Lang, K. et al., Outcomes in patients treated with gemtuzumab ozogamicin for relapsed acute myel- ogenous leukemia, Am. J. Health-Syst. Pharm., 59, 941, 2002. 166. Rajvanshi, P. et al., Hepatic sinusoidal obstruction after gemtuzumab ozogamicin (Mylotarg) therapy, Blood, 99, 2310, 2002. 167. Erba, H.P. et al., Final results of a multivariate logistic regression analysis to determine factors contributing to the risk of developing hepatic veno-occlusive disease (VOD) following treatment with gemtuzumab ozogamicin, Blood, 100, 339a (abstr. 1313), 2002. 168. Erba, H. et al., Results of a multivariate logistic regression analysis to determine factors contributing to the risk of developing hepatic veno-occlusive disease (VOD) following treatment with gemtuzumab ozogamicin, Proc. Am. Soc. Clin. Oncol., 21, 270a, 2002. 169. Kell, W.J. et al., A feasibility study of Mylotarg (gemtuzumab ozogamicin, GO) in combination with standard induced therapy in acute myeloid leukaemia (AML), Br. J. Haematol., 117, 19, 2002. Copyright © 2005 CRC Press, LLC 170. Stadtmauer, E.A., Trials with gemtuzumab ozogamicin (Mylotarg®) combined with chemotherapy regimens in acute myeloid leukemia, Clinical Lymphoma, 2, S24, 2002. 171. Sievers, E.L., Antibody-targeted chemotherapy of acute myeloid leukemia using gemtuzumab ozo- gamicin (Mylotarg), Blood Cells, Molecules, & Diseases, 31, 7, 2003. 172. Naito, K. et al., Calicheamicin-conjugated humanized anti-CD33 monoclonal antibody (gemtuzumab ozogamicin, CMA-676) shows cytocidal effect on CD33-positive leukemia cell lines, but is inactive on p-glycoprotein-expressing sublines, Leukemia, 14, 1436, 2000. 173. Linenberger, M.L. et al., Multidrug-resistance phenotype and clinical responses to gemtuzumab ozogamicin, Blood, 98, 988, 2001. 174. Matsui, H. et al., Reduced effect of gemtuzumab ozogamicin (CMA-676) on p-glycoprotein and/or CD34-positive leukemia cells and its restoration by multidrug resistance modifiers, Leukemia, 16, 813, 2002. 175. Walter, R.B. et al., Multidrug resistance protein (MRP) attenuates gemtuzumab ozogamicin (GO)- induced cytotoxicity in acute myeloid leukemia (AML), Blood, 100, 224b (abstr. 4415), 2002. 176. Jilani, I. et al., Differences in CD33 intensity between various myeloid neoplasms, Am. J. Clin. Pathol., 118, 560, 2002. 177. Jilani, I. et al., Quantitative differences in CD33 intensity between various myeloid neoplasms, Blood, 98, 586a (abstr. 2458), 2001. 178. Becker, M.W., Qian, D., and Clarke, M.F., CD33 is variably expressed on leukemic stem cells in acute myelogenous leukemia, Blood, 98, 587a (abstr. 2459), 2001. 179. Introna, M. et al., Differential response of acute myeloid leukaemia cells to gemtuzumab ozogamicin (Mylotarg®) in vitro. Role of Chk1 and 2 phosphorylation and caspase 3, Blood, 100, 319a (abstr. 1237), 2002. 180. Sievers, E. et al., Preliminary report of an ascending dose study of gemtuzumab ozogamicin (Mylo- targTM, CMA-676) in pediatric patients with acute myeloid leukemia, Blood, 96, 217b, 2000. 181. Zwaan, Ch. M. et al., Gemtuzumab ozogamicin in pediatric CD33-positive acute lymphoblastic leukemia: First clinical experiences and relation with cellular sensitivity to single agent calicheamicin, Leukemia, 17, 468, 2003. 182. Sievers, E. et al., Final report of prolonged disease-free survival in patients with acute myeloid leukemia in first relapse treated with gemtuzumab ozogamicin followed by hematopoietic stem cell transplantation, Blood, 100, 89a (abstr. 327), 2002. 183. Cohen, A.D. et al., Gemtuzumab ozogamicin (Mylotarg) monotherapy for relapsed AML after hemato- poietic stem cell transplant: Efficacy and incidence of hepatic veno-occlusive disease, Bone Marrow Transplant., 30, 23, 2002. 184. Sievers, E.L. et al., Gemtuzumab ozogamicin (Mylotarg) as a single agent to evaluate safety and determine maximum tolerated dose (MTD) in post hematopoietic stem cell transplant patients with relapsed, CD33+ acute myeloid leukemia (AML), Blood, 100, 336a (abstr. 1304), 2002. 185. Goldberg, S.L. et al., Gemtuzumab ozogamicin (Mylotarg) prior to allogeneic hematopoietic stem cell transplantation increases the risk of hepatic veno-occlusive disease, Blood, 100, 415a (abstr. 1611), 2002. 186. Stadtmauer, E. et al., Analysis of predisposing factors for hepatic veno-occlusive disease after treatment with gemtuzumab ozogamicin (Mylotarg®, CMA-676), Blood, 98, 124a (abstr. 520), 2001. 187. Raza, A. et al., Preliminary analysis of a randomized phase 2 study of the efficacy of 1 vs. 2 doses of gemtuzumab ozogamicin (Mylotarg®) in patients with high risk myelodysplastic syndrome, Blood, 100, 795a (abstr. 3140), 2002. 188. Zwaan, Ch. M. et al., Gemtuzumab ozogamicin in pediatric CD33-positive acute lymphoblastic leukemia: First clinical experiences and relation with cellular sensitivity to single agent calicheamicin, Leukemia, 17, 468, 2003. 189. Leopold, L.H. et al., Comparative efficacy and safety of gemtuzumab ozogamicin monotherapy and high-dose cytarabine combination therapy in the treatment of patients with acute myeloid leukemia in first relapse, Blood, 96, 504a, 2000. 190. Baccarani, M. et al., Preliminary report of a phase 2 study of the safety and efficacy of gemtuzumab ozogamicin (Mylotarg®) given in combination with cytarabine in patients with acute myeloid leukemia, Blood, 100, 341a (abstr. 1322), 2002. Copyright © 2005 CRC Press, LLC 191. Estey, E.H. et al., Experience with gemtuzumab ozogamicin (“Mylotarg”) and all-trans retinoic acid in untreated acute promyelocytic leukemia, Blood, 99, 4222, 2002. 192. Estey, E. et al., Mylotarg +/- IL-11 in patients age 65 and above with newly-diagnosed AML/MDS: Comparison with idarubicin + Ara-C, Blood, 98, 720a (abstr. 3007), 2001. 193. Estey, E.H. et al., Gemtuzumab ozogamicin with or without interleukin 11 in patients 65 years of age or older with untreated acute myeloid leukemia and high-risk myelodysplastic syndrome: Comparison with idarubicin plus continuous-infusion, high-dose cytosine arabinoside, Blood, 99, 4343, 2002. 194. Alvarez, R.H. et al., Pilot trial of Mylotarg, topotecan and Ara-C in relapsed/refractory acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS), Blood, 100, 340a (abstr. 1321), 2002. 195. Cortes, J. et al., Mylotarg combined with topotecan and cytarabine in patients with refractory acute myelogenous leukemia, Cancer Chemother. Pharmacol., 50, 497, 2002. 196. Raj, K. et al., Gemtuzumab ozogamicin is not associated with an increased risk of veno-occlusive disease in relapsed or refractory AML when used alone or in combination with flag chemotherapy, Br. J. Haematol., 117 (Suppl. 1), 20, 2002. 197. Giles, F. et al., Fatal hepatic veno-occlusive disease in a phase I study of Mylotarg and troxatyl in patients with refractory acute myeloid leukemia or myelodysplastic syndrome, Acta Haematol. (Basel), 108, 164, 2002. 198. Kell, J.W. et al., Mylotarg (gemtuzumab ozogamicin: GO) given simultaneously with intensive and/or consolidation therapy for AML is feasible and may improve the response rate, Blood, 100, 199a (abstr. 746), 2002. 199. Venugopal, P. et al., Phase II study of gemtuzumab ozogamicin (Mylotarg) combined with intensive induction chemotherapy using high dose Ara-C and mitoxantrone followed by amifostine in poor prognosis acute myeloid leukemia: Preliminary results, Blood, 100, 341a (abstr. 1323), 2002. 200. Amadori, S. et al., Sequential administration of gemtuzumab ozogamicin (GO) and intensive chemo- therapy for remission induction in previously untreated patients with AML over the age of 60: Interim results of the EORTC leukemia group AML-15A phase II trial, Blood, 98, 587a (abstr. 2465), 2001. 201. Sallah, A.S. et al., Combination imatinib and gemtuzumab ozogamicin in patients with chronic myeloid leukemia in myeloid blast crisis, Blood, 100, 319b (abstr. 4824), 2002. 202. Tsimberidou, A.M. et al., Mylotarg, fludarabine, cytarabine (Ara-C) and cyclosporine (MFAC) com- bination regimen in patients with CD33-positive primary resistant or relapsed acute myeloid leukemia (AML), Blood, 100, 339a (abstr. 1317), 2002. 203. Haran, V., Hiemenz, J., and Ballester, O.F., A phase I-II study of intensified induction and consolidation therapy with gemtuzumab ozogamicin for patients with high-risk acute myeloid leukemia: The AIM protocol, Blood, 100, 269b (abstr. 4608), 2002. 204. Alvarado, Y. et al., Pilot study of Mylotarg, idarubicin and cytarabine combination regimen in patients with primary resistant or relapsed acute myeloid leukemia, Cancer Chemother. Pharmacol., 51, 87, 2003. 205. Apostolidou, E. et al., Mylotarg, liposomal daunorubicin, Ara-C and cyclosporine (MDAC) regimen in patients with relapsed/refractory acute myelogenous leukemia (AML), Blood, 100, 268b (abstr. 4600), 2002. 206. Tanaka, M. et al., Prophylactic intravesical instillation of BCG in recurrent superficial bladder cancer: Is BCG treatment superior to other chemotherapeutic agents? J. Urol., 151, Suppl., 1994. 207. Takeshita, T., Salvage combination chemotherapy with enocitabine, aclarubicin and neocarzinostatin for recurrent or refractory acute myelogenous leukemia, Med. J. Kagoshima Univ., 52, 33, 2000. 208. Takahashi, T. et al., Clinical application of monoclonal antibody-drug conjugates for immunotargeting chemotherapy of colorectal carcinoma, Cancer, 61, 881, 1988. 209. Otsuji, E. et al., Effects of idiotypic human anti-mouse antibody against in vitro binding and antitumor activity of a monoclonal antibody-drug conjugate, Hepato-Gastroenterology, 50, 380, 2003. 210. Fang, J., Sawa, T., and Maeda, H., Factors and mechanism of “EPR” effect and the enhanced antitumor effects of macromolecular drugs including SMANCS, Adv. Exp. Med. Biol., 519, 29, 2003. 211. Maeda, H., Sawa, T., and Konno, T., Mechanism of tumor-targeted delivery of macromolecular drugs, including the EPR effect in solid tumor and clinical overview of the prototype polymeric drug SMANCS, J. Controlled Release, 74, 47, 2001. 212. Maeda, H., SMANCS and polymer-conjugated macromolecular drugs: Advantages in cancer chemo- therapy, Adv. Drug Delivery Rev., 46, 169, 2001. Copyright © 2005 CRC Press, LLC 213. Greish, K. et al., Macromolecular therapeutics: Advantages and prospects with special emphasis on solid tumor targeting, Clin. Pharmacokinetics, 42, 1089, 2003. 214. Tokura, S. et al., Induction of drug-specific antibody and the controlled release of drug by 6-O- carboxymethyl-chitin, J. Controlled Release, 28, 235, 1994. 215. Nishikawa, T. et al., Cytological study on the effect of neocarzinostatin on sarcoma 180 in vivo, J. Antibiot., 18, 223, 1965. 216. Napier, M.A. et al., Neocarzinostatin: Spectral characterization and separation of a non-protein chromophore, Biochem. Biophys. Res. Commun., 89, 635, 1979. 217. Konishi, M. et al., Esperamicins, a novel class of potent antitumor antibiotics. I. Physico-chemical data and partial structure, J. Antibiot., 38, 1605, 1985. Copyright © 2005 CRC Press, LLC 23 The Mitomycins William A. Remers CONTENTS I. Introduction II. Mechanism of Action III. Analogs and Derivatives A. Structural Analogs and Derivatives B. Conjugated Derivatives IV. Structure–Activity Relationships V. Synthesis VI. Clinical Applications References I. INTRODUCTION In 1956 Hata and colleagues at the Kitasato Institute isolated mitomycins A and B by chromatog- raphy of a mixture compounds from cultures of Streptomyces caespitosis.1 A subsequent investiga- tion on the same organism by Wakaki’s group at Kyowa Fermentation Industry yielded mitomycin C, which proved to be the most active anticancer agent among these, and subsequent, natural mitomycins.2 Mitomycin C (Chart 23.1) was introduced as an anticancer drug by Kyowa, and it caught on rapidly.3 It was given to an estimated 60% of patients receiving chemotherapy in Japan during the 1960s. In contrast, the first clinical trials in the United States were very unsatisfactory, with 11 patients dying from the effects of myelosuppression.4 The reason for this serious toxicity was that the mitomycin C was given on the schedule used for 5-fluorouracil — daily injection for five consecutive days — but it has cumulative and delayed toxicity to bone marrow. Bristol-Myers finally received approval in 1974 for clinical use of mitomycin C based on a revised dosage schedule.5 The earliest mitomycins all had the azirino[2,3:3,4]pyrrolo[1,2-a]indole ring system as their basic skeleton.6 They were divided into three structural types based on the nature and stereochem- istry of their side chains, as shown in Chart 23.1. Thus, mitomycin A and B types differ in the stereochemistry of the side chain,7 whereas mitomycin G types have only a methylene group.8 Two mitomycin-like structures (mitomycinoids) with somewhat different skeletons, isomytomycin A and albomitomycin A, were discovered by Kono and colleagues in 1987,9 and the somewhat related compound FR-900482 was isolated in 1987 (Chart 23.1).9 The

complicated systematic chemical names of the mitomycins led investigators to develop a distinctive nomenclature for them. Compounds with the complete skeleton including the aziridine ring, saturation at carbons 9 and 9a, and the quinone oxidation state are called mitosanes,6 and the corresponding hydroquinones are called leucomitosanes.10 If the structures are the same, except that there is a 9,9a-double bond, the compounds are called aziridinomitosenes and leucoaziridi- nomitosenes. If these compounds lack the aziridine ring, they are mitosenes (Figure 23.1).6 The structures of mitomycins A, B, and porfiromycin (Chart 23.1) were elucidated by Webb and colleagues at American Cyanamid in 1962 and confirmed by Tulinsky’s x-ray diffraction analysis.11 Copyright © 2005 CRC Press, LLC CHART 23.1 Natural Mitomycins One noteworthy finding in the course of degradative studies on mitomycins was the formation of 1- hydroxy-2-aminomitosenes when the mitomycins were treated with acid (Scheme 23.1).6 The ste- reochemistry of these “apomitomycins” and other acid hydrolysis products were determined by Taylor and Remers and by Shaber and Rebek.12,13 Other chemical transformations of mitomycins that became important in later mode of action studies included selective hydrolysis of the 7-amino group of mitomycin C to a hydroxyl group14 and selective hydrolysis of the carbamate group to a hydroxyl group (Scheme 23.1).15 Acylation of the aziridinyl nitrogen of certain mitomycins was used in preparing many conjugates, as described below. Catalytic reduction of the mitomycin quinone ring produces a leucoaziridinomitosene, which can be reoxidized to an aziridinomitosene.10 Other two-electron-reduction processes typically afford 1,2-disubstituted mitosenes with the 1-substituent and the stereochemistry dependent on the nucleo- phile.16 Reduction of mitomycins in the presence of reactive thiols resulted in replacement of the carbamate group in the side chain, as well as opening of the aziridine ring.17 Hornemann found that sodium dithionite reduction of mitomycin C afforded a mitosane in which the 9a-methoxy group was replaced by sodium bisulfite (Scheme 23.2).18 When mitomycin C was reduced at lower pH (5.0), it formed a 2,7-diaminomitosene by trapping a proton from the solvent. Kohn and coworkers reported that proton capture at C(1) occurs with high stereoselectivity: the proton is Copyright © 2005 CRC Press, LLC O HO 10 8 7 9 CH2OCONH2 CH2OCONH2 6 9a N 1 5 4 N NH NH O 3 2 1a HO Mitosane Leucomitosene O HO CH2OCONH2 CH2OCONH2 N NH N O NH HO Aziridinomitosene Leucoaziridinomitosane O CH2OCONH2 N O Mitosene FIGURE 23.1 Nomenclature for mitomycins and derivatives. added from the opposite side of the ring to the protonated 2-amino group. They suggested further that a quinone methide intermediate participates in this process (Scheme 23.3).19 Another interesting discovery by Kohn’s group was that mitomycin C analogs containing a second amino group in the C(7) substituent undergo cyclization to give a C(8) imine, provided that there is an appropriate spacer between the two nitrogens (Scheme 23.4).20 When the two nitrogens were separated by a three-methylene chain in which the central atom was dimethylated, the product was an iminoquinone with an albomitomycin structure (Scheme 23.4).21 O O O O H O NH2 H 2N 2N O NH2 OCH3 H+ N N OH NH O O NH2 NaOCH Apomitomycin 3 OH– O O O H2N O NH2 H2N O OCH OCH 3 3 N N NH NH O O 10-decarbamoyl mitomycin C 7-hydroxy-9a-methoxymitosene SCHEME 23.1 Chemical transformations of mitomycin C. Copyright © 2005 CRC Press, LLC O O O O O NH NH H2N O 2 H 2 2N OCH H2/Pt 3 N N NH NH O O Na2S2O4 [O] O O O O O NH H O NH2 H 2N 2N 2 SO3Na N N NH NH O O Na2S2O4 KSCO2Et O H2N SCO2Et SCO N 2Et O NH2 SCHEME 23.2 Reductive transformations of mitomycin C. Mitomycin C reduction by biological agents such as xanthine oxidase or NADPH -cytochrome C P450 reductase involves an initial one-electron transfer to give a semiquinone radical, which can be detected by electron paramagnetic resonance.22 Electochemical or catalytic reduction of mito- mycin C in inert solvents also produced semiquinone radicals.23 In biological systems, these species undergo disproportionation to the hydroquinones and quinones.24 O O O O H NH 2N O 2 H O N 2N H2 OCH Na2S2O4 3 N N NH NH O O O O O O H O 2N NH2 H O 2N NH2 H+ N N H O H NH O 3 NH3 2,7-diaminomitosene SCHEME 23.3 Formation of 2,7-diaminomitosene. Copyright © 2005 CRC Press, LLC O O O N H O NH O NH H 2 N 2 2N N H OCH OCH 3 3 N N NH NH O O O O O H2N H O NH N 2 N O NH N 2 OCH H 3 N OCH NH N 3 O O N SCHEME 23.4 Cyclization reactions of mitosanes with amino groups on the c(7)-substituent. Early studies on the biosynthesis of mitomycins showed that the O-methyl and N-methyl groups, but not the C-methyl group, are derived from L-methionine.25 Hornemann and coworkers then dem- onstrated that D-glucosamine is incorporated intact into mitomycins, and that L-citrulline donates the intact CONH2 group.26 More recently, Sherman and colleagues determined that the aminomethylben- zoquinone unit of mitomycin C comes from 3-amino-5-hydroxybenzoic acid (Figure 23.2).27 Floss and coworkers had previously shown that 3-amino-5-hydroxybenzoic acid is formed by a modified shikimate pathway involving 3,4-dideoxy-4-amino-D-arabinoheptulosonic acid-7-phosphate.28 In 1990 Islam and Skibo prepared pyrrolobenzimidazole analogs of mitosenes.29 The compound with 6-aziridino and 1-carbamoyloxy substituents (Scheme 23.5) was highly potent against human solid-tumor cell lines. It resembled doxorubicin in its DNA-strand cleaving ability and cardiotox- icity.30 The S-enantiomer was selectively toxic over the R-enantiomer for ovarian cancer cells.31 II. MECHANISM OF ACTION The antimicrobial and antitumor activities of mitomycins result in part from alkylation of DNA. Approximately 10% of their covalent interactions result in cross-linking of double-helical DNA, and the remaining linkages are monofunctional.32,33 Although alkylation of DNA occurred readily in living cells, this process could not be demonstrated in vitro until Iyer and Szybalski discovered that NADPH-dependent cell extracts or chemical reducing agents were required.34 Under these conditions, only marginal amounts of mitomycin C were incorporated, and there was no cross-linking. An ingenious solution to this problem in mitomycin activation was provided by Tomasz, who postulated O citrulline O H H O N 2N 2 AHBA OCH3 N NH D-glucosamine O FIGURE 23.2 Biosynthetic precursors of mitomycin C. Copyright © 2005 CRC Press, LLC SCHEME 23.5 Preparation of 6-aziridino-1-carbamoyloxy-2,3-dihydro-7-methyl-[1,2-a]benzimidazole-5,8- dione. that the initial binding of mitomycins involved semiquinone radicals rather than hydroquinones and devised a method of maximizing the concentration of semiquinone radicals by adding the reducing agent, sodium dithionite, in portions and maintaining an excess of the mitomycin.35 These conditions produced a high binding ratio of mitomycins, and cross-links were obtained. The effects of different oxidoreductase enzymes on the activation of mitomycin C were studied. Gustafson and Pritsos found that xanthine dehydrogenase activated it under both aerobic and anaerobic conditions, and they proposed that mitomycin reduction occurs at the NAD+ site on this enzyme.36 Enhancement of mitomycin C toxicity by compounds that induce D-T diaphorase was found to occur in tumor cells by Begleiter and coworkers.37 In contrast, D-T diaphorase down regulation in MKN28 gastric cancer cells produced resistance to mitomycin C.38 Sartorelli and colleagues found that in Chinese hamster ovary cells that overexpressed human NADPH:cyto- chrome C reductase, mitomycin C was highly active under both oxygenated and hypoxic conditions, whereas porfiromycin was markedly more cytotoxic under hypoxic than aerobic conditions.39 In 1964 Iyer and Szybalski proposed a mechanism of action for mitomycins based on bifunc- tional alkylation and cross-linking of DNA. It involved reduction of the quinone ring to a hydro- quinone followed by loss of the elements of methanol to give a mitosene hydroquinone. In this hydroquinone, the aziridine ring opens to provide a carbonium ion that can alkylate DNA. Cross-linking occurs by alkylation of DNA by C(10) of the mitomycin following elimination of the carbamoyloxy group.34 Moore modified this mechanism to include the intermediacy of quinone methides (Scheme 23.6).40 Kohn’s group found that mitomycin D and mitomycin-9a-sulfonate gave binding profiles comparable to that of mitomycin C under reductive conditions. These results indicated that neither the leaving group at C(9a) nor the stereochemistry at C(9) and C(9a) influence the sites of DNA binding and supported the concept of quinone methide formation before bonding at C(1).41 Boryah and Skibo determined the pKa values of mitomycin C in its quinone and hydroquinone forms. The quinone form has pKa values as follows: 1.2 for the protonated quinone ring, 2.7 for the protonated indole nitrogen, and 7.6 for the protonated aziridine nitrogen. In comparison, the hydroquinone form has pKa values of 1.2 for the protonated 7-amino group, 5.1–6.1 for the protonated indole nitrogen, and 9.1 for the protonated aziridine nitrogen. Protonation of the indoline nitrogen in the quinone form prevents the elimination of methanol; however, there is a 105-fold increase in the rate constant for methanol elimination in the hydroquinone form. The carbocation resulting from methanol elimination (pKa = 7.1) dissociates to a quinone methide. At pH values below 7.1 the carbocation traps water, and at pH values above 7.1 it traps both water and a proton.42 Copyright © 2005 CRC Press, LLC O O O O H O 2N NH2 H2N O NH2 OCH3 OCH3 -CH3OH N N NH NH O HO O O HO O H O NH2 2N H NH 2N O 2 DNA N N NH HO HO NH2 O HO HO H O 2N NH2 H2N DNA DNA N N HO HO NH NH 2 2 HO HO H2N DNA H2N DNA [O] N DNA N DNA HO NH2 HO NH2 SCHEME 23.6 Proposed mechanism of bifunctional alkylation of DNA by mitomycin C. The specific atoms involved in the alkylation of DNA by mitomycin C were determined by Tomasz, Nakanishi, and coworkers. They hydrolyzed the mitomycin-bound DNA with enzymes including snake venom diesterase and alkaline phosphatase to obtain the mitomycin-bound nucleo- sides and then elucidated their structures by a variety of analytical methods.43–48 Monoalkylation occurs preferentially at C(1) of the mitomycin, with the covalent bond formed with the 2-amino group of a guanine residue. Among the three different products isolated in the initial experiments, the main product had the amino groups attached to C(1) and C(2) of the mitosene moiety in a cis orientation. The minor products were the C(1) epimer and the 10-decarbamoyl derivative.49 Dialkylation with cross-linking involves alkylation of the 2-amino group of guanine in one strand of the DNA double helix by C(1) of mitomycin C and alkylation of the 2-amino group of guanine in the opposite strand by C(10) of the same mitomycin C molecule (Figure 23.3).50,51 Tomasz and coworkers isolated a O N N O N N N H H 2N ribose H N N N N ribose O NH2 N N H O FIGURE 23.3 Main product from cross-linking of DNA by mitomycin C. Copyright © 2005 CRC Press, LLC product that had an intrastrand cross-link between the DNA and mitosene. The covalent bonds were between 2-amino groups of guanines and C(1) and C(10) of the mitosene, just as in the intermolecular cross-linked products; however, the DNA double helix was bent at 14.6 degrees near the cross-link site.50,52A further monoadduct was isolated from mouse mammary tumor cells. It had the covalent bond formed between C(10) of the mitosene and the 2-amino group of a guanine residue.53 The preferred 5-CpG sequence for mitomycin binding and cross-linking DNA was first sug- gested by molecular modeling studies,54 and it was confirmed by chemical and spectroscopic studies in a number of laboratories.51,55,56 The laboratories of Tomasz and Kohn independently showed that the initial covalent bonding event occurs at C(1) of the mitomycin in the preferred 5-CpG sequence and is independent of the second bonding step.57–59 Kohn pointed out that a hydrogen bond between the C(10) carbonyl oxygen of an activated mitomycin and the 2-amino group on a guanine residue in the precovalent complex contributes significantly to the sequence specificity.59 Substitutions such as methoxycarbonyl and methanesulfonyl on the aziridine nitrogen of mitomycin C changed the sequence specificity for DNA binding.60 The mechanism of enzymatic reductive activation of mitomycin C was revised by Tomasz and colleagues in 1997. They found that when mitomycin C was reduced by a variety of chemical and enzymic agents in the presence of Micrococcus lysodeicticus, the

overall yield of adducts was higher at pH 7.4 than at 5.8, except for DT-diaphorase reduction. The cross-links were always greater at lower pH. From these results they proposed that three competing pathways for activation afford three different DNA-reactive electrophiles that generate three unique sets of DNA adducts.61 DNA alkylation by certain mitomycin C analogs was investigated by Tomasz and coworkers, who reported that in the absence of reduction 2,7-diaminomitosene bound to DNA, whereas mitomycin C, mitomycin A, porfiromycin, and certain analogs did not. Protonation of the 2-amino group of 2,7-diaminomitosene strongly affected binding, although the presence of functional groups including 1-OH and N-methyl did not.62 Under reductive activation, C(10) of 2,7-diami- nomitosene alkylated DNA on N(7) of a guanine residue.63 Decarbamoyl mitomycin C was surprisingly more cytotoxic than mitomycin C to hypoxic EMT6 mouse mammary tumor cells and CHO cells. The frequency of total adducts was 20–30-fold higher with decarbamoyl mitomycin C, although the frequency of cross-links was equal. These results are in contrast to the alkylation of calf thymus DNA in vitro, wherein mitomycin C is more reactive.64 The unnatural enantiomer of mitomycin C formed both a mono and a bis adduct on reduction in the presence of DNA. These adducts were stereoisomers of the main adducts formed by natural mitomycin C.65 Reduction of racemic 1,10-bisacetoxy-7-methoxymitosene in the presence of DNA gave five products. In the two major adducts, the stereochemistry at C(1) was the same as that found for natural mitomycin C adducts, but the C(2) chirality was reversed.66 Alkylation sites on reduced mitosenes were assessed by Ouyang and Skibo, using 13C nuclear magnetic resonance. They found that reduction by dithionite gave sulfite esters as well as sulfonates, both the 6-amino group of adenine and the 2-amino group of guanine are alkylated, both nitrogen and oxygen centers in purines residues are alkylated, and the main fate of the iminium ions formed after reduction is head-to-tail polymerization (Scheme 23.7).67 Zhu and Skibo found that when indoloquinones with 5-aziridino substituents were reduced to hydroquinones, the fate of this aziridino group and the cytotoxicity were strongly influenced by the adjacent 6-substituent. In the 6-unsubstituted compound the aziridine ring may assume a conformation in which a 1,5-sigmatropic shift occurs (Scheme 23.8). This shift results in loss of the aziridine ring and drastically reduces cytotoxicity. In contrast, a 6-methyl group provides steric hindrance that prevents the sigmatropic shift and allows the aziridine ring to alkylate DNA, thus retaining cytotoxicity. Similar results were observed for the corresponding benzoquinones and benzimidazolequinones.68 The binding of mitomycin C to modified oligonucleotides was investigated. Millard and Beachy found that cross-linking was increased when the cytosines in CpG-CpG were methylated. They suggested that this enhancement resulted from an electronic effect.69 Oligonucleotides in which 2,6-diaminopurine was substituted for guanine were readily alkylated by reduced mitomycin C. Copyright © 2005 CRC Press, LLC O O O 13 CH O CH2O C CH 13 13 3 2 CH2SSO CH2SO Na2S2O4 O O N N N O O CH3O N O 13 CH2 13 pentamer O CH3 CH3O N H O SCHEME 23.7 Products from dithionite reduction of a 13C-labeled mitosene. The sequence specificity was TpD (analogous to CpG), and cross-links occurred in TpD-TpD and TpG.CpD regions.70 Glutathione does not reduce mitomycin C or form conjugates with it; however, in the presence of chemical or enzymic reducing systems, three monoconjugates and two bisconjugates are formed. Similar results were found with mercaptoethanol or N-acetylcysteine. Glutathione accelerated the reduction of mitomycin C by “slow” agents such as cytochrome C reductases. Glutathione was considered by Tomasz to be involved in further reduction of the initially formed semiquinone radical.71 A ternary glutathione–mitomycin–DNA adduct was isolated from reduction of mitomycin C in the presence of DNA or synthetic duplex oligonucleotides. The mitosene residue was linked to the nucleotide by its C(1) atom, and the glutathione residue was linked to C(10). Overall, glutahione has an inhibitory effect on mitomycin C alkylation of DNA, which may be a factor in mitomycin resistance by tumor cells with high glutathione levels.72 In support of this idea, Xu and coworkers found that the cytotoxicity of mitomycin C increased significantly in cells pretreated with ethacrynic acid, which inhibits glutathione transferase and depletes glutathione levels.73 The effects of DNA alkylation by mitomycin C on other biological processes were examined by Tomasz and collaborators. The action of DNA polymerases was terminated nearly quantitatively at the nucleotide on the 3 side of each mitomycin monoadduct site on the DNA. These monoadducts do not distort the DNA, but the resulting increase in thermodynamic stability may prevent replication.74 In selected tissue culture cell lines, mitomycin C and decarbamoylmitomycin C, but not 2,7-diami- nomitosene, induced p53 protein levels and increased levels of p21am and Gadd45 mRNA. They also H O O HO N N H2N N HO H HO HO SCHEME 23.8 1,5-Sigmatropic shift in a 5-aziridinoindolohydroquinone. Copyright © 2005 CRC Press, LLC Nu O O Nu N N O R O Nu P O O O O O O Nu O P O O O O H O N N R O N O SCHEME 23.9 Alkylation of a phosphate group in DNA by a 6-aziridinopyrrolo[1,2-a]imidazole. induced apoptosis in ML-1 cells. This marked the first observation that adducts of a multiadduct type of DNA-damaging agent were differentially recognized by DNA damage sensor pathways.75 Benzimidazole analogs with 6-aziridino substituents (Scheme 23.9) cleave DNA under reducing conditions, specifically at G + A bases.76 Skibo and Schultz showed that after reduction of the quinone ring, the aziridino group of these compounds alkylate phosphate groups in the DNA backbone to form phosphotriesters that are labile to hydrolysis. They also alkylate N(7) of guanine and adenine residues. The observed base specificity was attributed to Hoogsteen-type hydrogen bonding of the hydroquinone in the major groove of the DNA double helix.77 A second important mechanism of action for mitomycins is the generation under aerobic conditions of superoxide and hydroxyl radicals that can cause DNA strand cleavage.78-82 This process is promoted by metal ions such as Cu(II) and inhibited by enzymes such as catalase and superoxide dismutase, as well as by free-radical scavengers and metal-ion sequestering agents.83 Pritsos and Sartorelli showed that the free radicals generated by quinone antibiotics, including mitomycins, cause cytotoxicity that depends on the extent of DNA damage and the ability of cells to repair the damage.84 This damage is easier to repair than that caused by mitomycin cross-linking of DNA. III. ANALOGS AND DERIVATIVES A. STRUCTURAL ANALOGS AND DERIVATIVES The rather broad antitumor spectrum of mitomycin C, together with its serious toxicity, spurred the quest for analogs that might have better therapeutic indices. Hundreds of mitosanes, aziridi- nomitosenes, and mitosenes, as well as indoloquinones, were synthesized and screened.85 Despite the development of numerous compounds with very impressive activity in animal screens, there have been no new clinical agents approved to date. Among the mitosanes, the position most readily substituted is C(7) in the quinone ring. In particular, the replacement of the 7-methoxy group, as found in mitomycin A, with amines and alkoxides has been much investigated (Scheme 23.10). The earliest studies were by Kinoshita and colleagues.86–88 Mitomycin A is not produced in quantity by fermentation, but it is readily obtained by hydrolysis of mitomycin C to the 7-hydroxy derivative (Scheme 23.1), followed by O-methy- lation using diazomethane (Scheme 23.10).14 The corresponding compounds with a methyl group on the aziridine nitrogen can be prepared from porfiromycin. An important feature of the 7-amino analogs is that both the quinone-reduction potential and the lipophilicity can be controlled by varying the substituents on the nitrogen. A large variety of substituted amines, including ethy- lamines, secondary amines, arylamines, and amidines were prepared and tested by Remers and coworkers and by chemists at Bristol-Myers Squibb.89–98 As discussed in the following sections, the arylamines and amidines were highly potent, and ethylamines bearing disulfide groups were Copyright © 2005 CRC Press, LLC O O O O HO O NH2 CH3O O NH2 OCH CH2N2 3 OCH3 N N NH NH O O – RO O O O O RO O NH2 R1R2N O NH2 OCH OCH3 3 N N NH NH O O SCHEME 23.10 Preparation of 7-substituted mitosanes. efficacious. New 7-alkoxymitosanes were prepared by alkoxide exchange on mitomycin A or treatment of 7-hydroxymitosane with 3-substituted 1-phenyltriazenes (Scheme 23.10). Many of them showed high potency.99 Mitomycin C analogs with structural changes at other positions included acyl derivatives on the aziridine nitrogen,86 the potent 10-decarbamoyl derivative,15 which was converted into 10-chloro and 10-bromoderivatives,100 and the 6-carboxaldehyde.101 Aziridinomitosenes corresponding to mitomycin A and its N-methyl derivative were prepared from the parent compounds by catalytic reduction and air oxidation. The methoxy group was then replaced by a variety of amines. As described below, potency depended on the ease of quinone reduction.102 Mitosenes without the aziridine ring, but with substituents at C(1) or at C(1) and C(2), were prepared by total synthesis by Remers and coworkers.103,104 1-Acetoxy-mitosenes with a variety of substituents in the quinone ring, and 7-methoxymitosenes with new substituents at C(1) and C(10), were also investigated.105,106 None of these mitosenes were comparable in activity to the mitosanes. The earliest work on synthesis and evaluation of mitomycin analogs was conducted by Allen and coworkers on indoloquinones and directed toward the development of new antibacterial agents. It provided important information on the relative importance of various functional groups in mitomycin-related structures.107–110 B. CONJUGATED DERIVATIVES A variety of conjugates has been prepared from mitomycin C in attempts to improve its therapeutic properties. The earliest one, which was prepared by Iwata and Goldberg, involved coupling con- conavilin A with N(1a)-succinylated mitomycin C in the presence of a carbodiimide. An improved antitumor effect was claimed.111 Conjugation with bovine serum albumin required prior acetylation to prevent protein polymerization in the presence of carbodiimides. The conjugate of mitomycin C with glutarated albumin had high mitomycin content, and it was stable.112 Many conjugates were based on the delivery of mitomycin C to selected targets using mono- clonal antibodies. In general, these conjugates were prepared by substituting the aziridino nitrogen with acyl groups having a free carboxyl group at the other end of the chain and then coupling with an antibody (Scheme 23.11). Sometimes the antibody is linked to an intermediary, which provides many binding sites for the mitomycin. Acylated aziridines are less stable than ordinary amides, and they undergo gradual hydrolysis to liberate mitomycin C at the target. The following conjugates were formed in this manner: monoclonal anti-MM46 immunoglobulin M antibody, with or without Copyright © 2005 CRC Press, LLC O O O O O O H O H O NH 2N NH2 2N 2 OCH OCH 3 O N N CO2H NH NC O O O H2NAb carbodiimide O O H O H 2N N 2 OCH O N C NC NHAb O O SCHEME 23.11 Synthesis of monoclonal antibody conjugates of mitomycin C. serum albumin as intermediary113; anti-alpha-fetoprotein antibody114; monoclonal antibody MGb2 with dextran T-40 as intermediary115; and anti–neural cell adhesion molecule monoclonal anti- body.116 None of these conjugates has become an established therapeutic agent to date, although the MGb2 monoclonal antibody conjugate was reported to have enhanced activity against human gastric cancer cells in nude mice.115 Two types of conjugates were directed toward binding with DNA. Kohn’s group treated decar- bamoylporfiromycin with carbonyldiimidazole followed by H2N(CH2)6P=S(OH)R, wherein R is the base sequence for human A-raf-1-gene or human FGFR1 gene.117 Tomasz and colleagues prepared conjugates in which 1, 2, or 3 N-methylpyrrolecarboxamide units were tethered by five methylene groups to the 7-amino group of mitomycin C. These N-methylpyrrole units are found in distamycin, and they direct molecules to AT-rich regions of DNA.118 Some derivatives of mitomycin C were prepared to modify its physical properties for improved absorption and distribution. Thus, alkyl and alkenylazacycloalkananone derivatives exhibit improved penetration through rat skin.119 Another lipophilic substituent was the docosahexenoyl group.120 Derivatives with hydrophilic properties were prepared by attaching a linear (1-3)-β-D- glucan to the aziridine nitrogen,121 or a glucopyranosyl group linked through a phenolic group to N(7).122 A conjugate with dextran, prepared by using CO(CH2)4CH2NHC(N=H) as the linker, showed strong affinity to lymph cells.123 Other hydrophilic mitomycin C derivatives were formed by condensing poly[N5-(2-hydroxyethyl)-L-glutamine] through a tetrapeptide spacer124 and by coupling a N(1a)succinyl mitomycin C with a chitosan. The latter derivative had high drug content, pH-dependent drug release, and

long duration in plasma.125 A novel prodrug containing a 1,2-diacyl lipid linked to mitomycin C by a thiolytically cleavable dithiobenzyl carbamate (Figure 23.4) was formulated in STEALTH liposomes. It was reported to be well retained in the liposomes and to show increased tumor inhibition over mitomycin C.126 O O O O NH H OCC17H35 2N 2 OCH S S 3 CH2OCC17H35 N NCO O O O FIGURE 23.4 Thiolytically cleavable lipophilic prodrug of mitomycin C. Copyright © 2005 CRC Press, LLC O NH2 O O O H2N O NH2 N H2N O NH2 O O N N O P O O N O NH O HO 3 O NH3 HO NH2 N O – O N O P O O –O HO HO SCHEME 23.12 Prodrug for mitomycin and 5-ara-CMP. Conjugates of 2,7-diaminomitosenes with cytidine 5-monophosphate and 5-ara-CMP were prepared as prodrugs that would liberate both the nucleotides and an alkylating intermediate on reduction. This process actually worked, as shown by liberation of the nucleotides, and the 5-ara- CMP conjugate prepared from mitomycin C was more potent than either parent compound (Scheme 23.12).127,128 IV. STRUCTURE–ACTIVITY RELATIONSHIPS One key to understanding structure–activity relationships among mitomycins and their analogs is that the functional groups are so distributed on the pyrrolo[1,2-a]indole nucleus that they interact with each other in a chain of events that produces potent alkylating functionalities. Another key to understanding is that at least three factors determine antitumor potency. They include the ability to enter cells, which is determined in part by their partition coefficient; the ease of their reduction to reactive intermediates, which is determined by the quinone reduction potential; and the presence of optimal alkylating groups such as the aziridine ring and the carbamoyloxymethyl substituent.86,87,95 Mitomycins are relatively stable in their quinone state; however, as described in Section II, reduction of the quinone ring strongly activates the molecule. The nitrogen atom common to two rings becomes more electron rich and is able to participate in the loss of the 9a-substituent to give a leukoaziridinomitosene, which can undergo opening of the aziridine ring and loss of the carbamoyl group (Scheme 23.6). The ease of reduction of the quinone ring is obviously important in initiating the action of mitomycins, and this can influence antitumor potency. Thus mitomycin A, which has a 7-methoxy substituent and half-wave reduction potential (E1/2) of –0.21 V, is more readily reduced and a more potent cytotoxic agent than mitomycin C, which has a 7-amino group and E1/2 of –0.45 V.91 Kunz and coworkers subjected a set of 30 mitomycin C and mitomycin A analogs to a QSAR study using multiple linear regression analysis based on cytotoxic potency in a panel of human tumor cells in culture (log 1/C), E1/2, and the logarithm of the partition coefficient (log P). A statistically significant correlation with an equation log (1/C) = 10.1 + 6.59 E1/2 + 0.35 log P was obtained. The parameter E1/2 accounted for 48% of the variance, and log P accounted for 21%.95 When the analysis was limited to mitomycin C analogs, E1/2 was the only significant variable, whereas when the analysis was limited to mitomycin A analogs, log P was the only significant variable. These results were explained by the consideration that all of the mitomycin A analogs were 7-alkoxymitosanes with identical low E1/2 values, meaning that cell uptake is the controlling Copyright © 2005 CRC Press, LLC TABLE 23.1 Mitomycins Receiving Clinical Trials O O X O NH2 OCH3 N N-Y O Identification X Y Mitomycin C H2N H Porfiromycin H2N CH3 M83 HO NH H BMY-25282 (CH3)2N–C=N– H BMY-25067 O2N SSCH2CH2NH H H2N H KT 6149 H HO2C (CH2)2CONH(CH2)2SS(CH2)2NH factor in their potency, whereas the mitomycin C analogs, being harder to reduce in biological systems, are limited by reduction potential rather than by cell uptake. The nature of the assay system may be crucial in determining the relative potencies of mitomycins, and presumably other types of drugs. Thus, in the murine leukemia assay, the most potent mitomycin C and mitomycin A analogs were the most hydrophilic ones, whereas in the cell culture study described above, the most lipophilic analogs were the most potent ones.129,130 This result indicates that the more lipophilic compounds may be concentrating in fatty tissues in the mice. It also indicates that optimizing a set of analogs in an in vitro assay might not be valuable for subsequent in vivo assays. An ethyleneimino substituent at C(7) in the quinone ring gave especially potent antitumor activity against P-388 leukemia in mice.91 Skibo and coworkers subsequently showed that this substituent alkylates phosphate groups in the DNA backbone, giving enhanced covalent binding energy.76 Quinones bearing an ethyleneimine group are reduced more readily than those with alkylamines.92 Another C(7) substituent that confers increased antitumor potency to mitomycins is the amidine group, as found in BMY-25282 (Table 23.1).98,131 Elimination of the elements of water or methanol from C(9) and C(9a) of mitomycins produces aziridinomitosenes. They have all of the appropriate functionality for efficient DNA cross-linking, but they are less potent than the corresponding mitosanes because their indoloquinone chromophores are more difficult to reduce than the corresponding benzoquinone chromophores of mitosanes.132 Mitosenes without the aziridine ring are also less potent than mitosanes, although the presence of two alkylating functionalities at C(1) and C(10) permits bisalkylation with cross-linking of DNA. For example, the 1,10-diacetoxymitosene prepared by Orlemans et al. possessed antitumor activity, although the spectrum was more limited than that of mitomycin C.133 Even indoloquinones with good leaving groups at C(2), which is equivalent to C(1) in mitosenes, have potent antibacterial activity in culture.134 Mitosanes lacking the carbamoyl substituent are surprisingly active antitumor agents.64 Their potency may be explained by facile loss of hydroxide from an activated intermediate to give the Copyright © 2005 CRC Press, LLC O OH O O OCH O NH 3 CH3O 2 CH O 3O CH3 HBF OCH3 4 NR N NR N O H O SCHEME 23.13 Formation of the mitosane nucleus in the Kishi synthesis. same quinonemethide that mitomycins provide. The antitumor activity of mitomycin G, which has only a methylene group for C(10),8 can be rationalized in the same way. V. SYNTHESIS The total synthesis of mitomycins has presented a formidable challenge to chemists because of the highly interactive functionality arranged on a compact skeleton. Allen and colleagues at American Cyanamid made early syntheses of indoloquinines and mitosenes related to the mitomycins, with the goal of producing new antibacterial agents.135–137 This research was extended by Remers and coworkers to 1-substituted and 1,2-disubstituted mitosenes.103,104,138 Groups led by Hirata and by Franck showed how to form the aziriding ring on the pyrrolo[1,2-a]indole nucleus.139,140 Kametani and coworkers introduced the 9a-methoxy group into the pyrrolo[1,2-a]indole system.141 The real breakthrough in mitomycin synthesis was the landmark total synthesis of mitomycin C published by Kishi’s group in 1977.142,143 Although this synthesis was quite long, it showed how to solve most of the vexing problems in handling the reactive functionalities of mitomycins. A key step was the transannular closure of the two five-membered rings of the pyrrolo[1,2-a] indole system from an eight-membered ring (Scheme 23.13). In 1989, Fukuyama and Lang published an elegant and practical synthesis of mitomycin C by a route that went through isomitomycin A (Scheme 23.14).144 Danishefsky and colleagues have reported a number of ingenious, but unsuccessful, attempts to synthesize mitomycin C.145 They did, however, succeed in the total syntheses of mitomycin K and the mitomycin-related compound FR-900482.146,147 VI. CLINICAL APPLICATIONS Mitomycin C was very widely used in cancer chemotherapy in Japan during the 1960s, and it subsequently received some important applications in the United States and other countries. It showed a broad spectrum of antitumor activity, although it was seldom the most effective agent against specific tumors. More recently, its use has declined because of its high toxicity and the introduction of newer agents of other chemical types. Nevertheless, it is still effective in cancer chemotherapy, especially in combinations with other anticancer drugs. In 1968 Mortel and colleagues found that mitomycin C produced objective responses in patients with tumors of the colon, pancreas, stomach, liver, and gallbladder.148 It also showed activity against non–small cell lung cancer, some head and neck cancers, breast cancer, and chronic myelogenous leukemia.149–151 Early and coworkers showed that mitomycin C is active in topical administration to superficial bladder tumors.152 Furthermore, systemic activity was found for mitomycin C in advanced biliary, ovarian, and cervical squamous cell carcinoma.153 Combination regimens including mitomycin C have been used against a variety of cancers. One of the most common combinations is the FAM regimen, which contains 5-fluorouracil, Adri- amycin, and mitomycin C and is used in treating gastric and pancreatic carcinomas.154 Mitomycin C and bleomycin in combination produced response rates up to 80% in cancers of the uterine cervix.149 The MOB regimen includes mitomycin C, Oncovin, and bleomycin. Copyright © 2005 CRC Press, LLC SCHEME 23.14 Total synthesis of mitomycin C. Mitomycin C (Mutamycin®) is supplied as lyophilized powder also containing mannitol in 5- and 20-mg vials. It is diluted with sterile water and administered by slow intravenous push or by continuous intravenous infusion. Extravasation produces severe local necrosis. It is cleared rapidly from the vascular compartment, and the highest level appears in the kidneys, with lower detectable levels found in muscle, intestine, stomach, and eye.155 The most serious toxicity of mitomycin C is bone marrow suppression, which reduces blood levels of platelets, leukocytes, and erythrocytes.151 It is delayed and cumulative, which requires careful minimization of lifetime doses to under 50 to 60 mg/m2. Gastrointestinal distur- bances, renal toxicity, and alopecia also occur, and fatigue and drowsiness have been observed.151,153,155 The troublesome toxicity of mitomycin C has prompted the synthesis of hundreds of analogs. Among them, four compounds with novel substituents at C(7) have been introduced into clinical trials (Table 23.1). The p-hydroxyanilino derivative known as M83 was introduced by Kyowa Hakko Kogyo, and it showed excellent activities in animal models;90 however, it was not superior to mitomycin C in the clinic. Bristol-Myers Squibb developed the highly potent dimethylamidino derivative, BMY–25282,156–157 but it was cardiotoxic. Following interesting preclinical studies on the supposed mercaptoethylamino derivative, which was shown to actually be the corresponding disulfide,98,158,159 a variety of disulfides cleavable to mercaptoethylamino derivatives was prepared.160 Bristol-Myers Squibb conducted clinical trials on the [2-(4-nitrophenyl)dithio]ethylamino deriva- tive, BMY-25067,161 but eventually decided not to pursue it further. The 2-[2-(L-glutamy- lamino)ethyl]dithioethylamino derivative, KT 6149,96 was introduced into the clinic by Kyowa Fermentation Industries and is still under investigation. Porfiromycin was discovered at Upjohn,162 and they gave it a brief clinical trial. It has much the same pattern of activity and toxicity as mitomycin C, but it is less potent. Copyright © 2005 CRC Press, LLC REFERENCES 1. Hata, T. et al., Mitomycin, a new antibiotic from streptomyces, J. Antibiot. (Jpn.), Ser. A, 9, 141, 1956. 2. Wakaki, S. et al., Isolation of new fractions of antitumor mitomycins, Antibiot. Chemother. 8, 288, 1958. 3. Frank, W. and Osterberg, A.E., Mitomycin C (NSC-26988) — an evaluation of the Japanese reports, Cancer Chemother. Rep., 9, 114, 1960. 4. Jones, Jr., R., Mitomycin C. A preliminary report of studies of human pharmacology and initial therapeutic trial, Cancer Chemother. Rep. 2, 3, 1959. 5. Sutow, W.W. et al., Evaluation of dosage schedules of mitomycin C (NSC-26988) in children, Cancer Chemother. Rep., Part 1, 55, 285, 1971. 6. Webb, J.S. et al., The structures of mitomycins A, B, C, and porfiromycins — parts I and II, J. Am. Chem. Soc., 84, 3185, 3186, 1962. 7. Yahashi, R. and Matsubara, I., The molecular structure of 7-demethoxy-7-p-bromoanilino mitomycin B, J. Antibiot (Tokyo), 29, 10, 1976. 8. Urakawa, C. et al., Mitomycin derivatives, Jpn. Kokai Tokkyo Koho, 1980. 9. Kono, M. et al., Albomitomycin A and isomitomycin A: products of novel intramolecular rearrange- ment of mitomycin A, J. Am. Chem. Soc., 109, 7224, 1987. 10. Feigelson, G.B. et al, On the reaction of leucomitosenes with osmium tetroxide. A route to novel mitomycins, J. Org. Chem., 1988, 53, 3390. 11. Tulinsky, A. and van den Hende, J.H., The crystal and molecular structure of n-brosylmitomycin A, J. Am. Chem. Soc., 89, 2905, 1967. 12. Taylor, W.G. and Remers, W.A., Structures and stereochemistry of mitomycin hydrolysis products, Tetrahedron Lett., 2483, 1974. 13. Rebek, J. et al., Total synthesis of a mitosene, J. Org. Chem., 49, 5164, 1984. 14. Garrett, E.R.,

The physical chemical characterization of the products, equilibria, and kinetics of the complex transformations of the antibiotic porfiromycin, J. Med. Chem., 6, 488, 1963. 15. Kinoshita, S. et al., Chemical transformations of mitomycins: structure-activity relation of mitomycin derivatives, Prog. Antimicrob. Anticancer Chemother. II, 112, 1970. 16. Underberg, W.J.M. and Lingeman, H., Aspects of chemical stability of mitomycin and porfiromycin in solution, J. Pharm. Sci., 72, 549, 1974. 17. Schiltz, P. and Kohn, H., Studies on the reactivity of reductively activated mitomycin C, J. Am. Chem. Soc., 115, 10510, 1993. 18. Hornemann, Ho, Y.-K., Mackey, Jr., J.K., and Srivastava, S.C., Studies on the mode of action of mitomycin antibiotics. Reversible conversion of mitomycin C into sodium 7-aminomitosane-9a- sulfonate, J. Am. Chem. Soc., 98, 7069, 1976. 19. Han, I., Russell, D.J., and Kohn, H., Studies on the mechanism of mitomycin C(1) electrophilic transformations: structure-activity relationships, J. Org. Chem., 57, 1799, 1992. 20. Wang, S. and Kohn, H., Studies on the mechanism of activation of C(7) ethylenediamine substituted mitomycins. Relevance of the proposed mode of action of BMY-25067 and KW-2149, Tetrahedron Lett., 37, 2337, 1996. 21. Wang, S. and Kohn, H., Mitomycin betaines: synthesis, structure and solvolytic reactivity, J. Org. Chem., 61, 9202, 1996. 22. Pan, S.-S. et al., Reductive activation of mitomycin C and mitomycin C metabolites catalyzed by NADPH-cytochrome P-450 reductase and xanthine oxidase, J. Biol. Chem., 259, 959, 1984. 23. Danishefsky, S.J. and Egbertson, M., The characterization of intermediates in the mitomycin activation cascade: a practical synthesis of an aziridinomitosene, J. Am. Chem. Soc., 108, 4648, 1986. 24. Hoey, B.M., Butler, J., and Swallow, A.J., Reductive activation of mitomycin C, Biochemistry, 27, 2608, 1988. 25. Kirsch, E.J. and Korshalla, J.D., Influence of biological methylation on the biosynthesis of mitomycin A, J. Bacteriol., 87, 247, 1964. 26. Hornemann, U. et al., D-Glucosamine and L-citrulline, precursors in mitomycin biosynthesis by Streptomyces verticillatus, J. Am. Chem. Soc., 96, 320, 1974. 27. Mao, Y., Varoglu, M., and Sherman, D.H., Mitomycin resistance in Streptomyces lavendulae includes a novel drug-binding-protein-dependent export system, J. Bact., 181, 2199, 1999. Copyright © 2005 CRC Press, LLC 28. Kim, C.G. et al., Formation of 3-amino-5-hydroxybenzoic acid, the precursor of mC7 units in ansamycin antibiotics, by a new variant of the shikimate pathway, J. Am. Chem. Soc., 111, 4941, 1992. 29. Islam, I. and Skibo, E.B., Synthesis and physical studies of azamitosene and iminoazamitosene reductive alkylating agents. Iminoquinone hydrolytic stability, syn/anti isomerization, and electro- chemistry, J. Org. Chem., 55, 3195, 1990. 30. Islam, I. et al, Structure-activity studies of antitumor agents based on pyrrolo[1,2-a]benzimidazoles: new reductive alkylating DNA cleaving agents, J. Med. Chem., 34, 2994, 1991. 31. Huang, X., Suleman, A., and Skibo, E.B., Rational design of pyrrolo[1,2-a]benzimidazole based antitumor agents targeting the DNA major groove, Bioorg. Chem., 28, 324, 2000. 32. Mercado, C.M. and Tomasz, M., Inhibitory effects of mitomycin-related compounds lacking the C1- C2 aziridine ring, Antimicrob. Agents Chemother., 1, 73, 1972. 33. Szybalski, W. and Iyer, V.N., The mitomycins and porfiromycin, in Antibiotics I, Mechanism of Action, Gottlieb, D. and Shaw, P.D., Eds., Springer, New York, 1967, pp. 221-245. 34. Iyer, V.N. and Szybalski, W., Mode of interaction of mitomycin C with deoxyribonucleic acid and other polynucleotides in vitro, Science, 145, 55, 1964. 35. Tomasz, M. et al., Mode of interaction of mitomycin C with deoxyribonucleic acid and other poly- nucleotides in vitro, Biochemistry, 13, 4878, 1974. 36. Gustafson, D.L. and Pritsos, C.A., Kinetics and mechanism of mitomycin C bioactivation by xanthine dehydrogenase under aerobic and hypoxic conditions, Cancer Res., 53, 5470, 1993. 37. Wang, X. et al., Enhanced cytotoxicity of mitomycin C in human tumor cells with inducers of DT- diaphorase, Br. J. Cancer, 80, 1223, 1999. 38. Sagara, N. and Katoh, M., Mitomycin C resistance induced by TCF-3 overexpression in gastric cancer cell line MKN28 is associated with DT-diaphorase down regulation, Cancer Res., 60, 5959, 2000. 39. Belcourt, M.F. et al., Differential toxicity of mitomycin C and porfiromycin to aerobic and hypoxic Chinese hamster ovary cells overexpressing human NADPH:cytotcrome (P-540) reductase, Proc. Natl. Acad. Sci. USA, 93, 456, 1996. 40. Moore, H.W., Bioactivation as a model for drug design bioreductive alkylation, Science, 197, 527, 1977. 41. Li, V.-S. et al., Structural requirements for mitomycin C DNA binding, Biochemistry, 34, 7120, 1995. 42. Boruah, R.C. and Skibo, E.B., Mitomycin C redox couple pKa values studied by spectrophotometric titration, pH rate profiles, and Nernst-Clark plots, J. Org. Chem., 60, 2232, 1995. 43. Tomasz, M. et al., Isolation and characterization of a major adduct between mitomycin C and DNA, J. Am. Chem. Soc., 110, 5892, 1988. 44. Tomasz, M., Chawla, A.K., and Lipman, R., Mechanism of monofunctional and bifunctional alkylation of DNA by mitomycin C, Biochemistry, 27, 3182, 1988. 45. Tomasz, M. et al., Full structure of a mitomycin C dinucleotide adduct. Use of differential FT-IR spectroscopy in microscale structural studies, J. Am. Chem. Soc., 105, 2059, 1983. 46. Verdine, G.L. and Nakanishi, K., Use of differential second-derivative UV and FTIR spectroscopy in structural studies with multichromophoric compounds, J. Am. Chem. Soc., 107, 6118, 1985. 47. Tomasz, M. et al., Reassignment of the guanine-binding mode of reduced mitomycin C. Biochemistry, 25, 4337, 1986. 48. Tomasz, M. et al., Reaction of DNA with chemically or enzymatically activated mitomycin C: isolation and structure of the major covalent adduct, Proc. Natl. Acad. Sci. USA, 83, 670, 1986. 49. Borowy-Borowski, H. et al., Duplex oligodeoxyribonucleotides crosslinked by mitomycin C at a single site: synthesis, properties, and crosslink, Biochemistry, 19, 2992, 1990. 50. Norman, B. et al., NMR and computational characterization of mitomycin cross-linked to adjacent deoxyguanosines in the minor groove of the d(TACGA).d(TACGA) duplex, Biochemistry, 29, 2861, 1990. 51. Weidner, M.F., Millard, J.T., and Hopkins, P.B., Determination at single-nucleotide resolution of the sequence specificity of DNA interstrand cross-linking agents in DNA fragments, J. Am. Chem. Soc., 111, 9720, 1989. 52. Rao, S.N., Singh, U.C., and Kollman, P.A., Conformations of the noncovalent and covalent complexes between mitomycins A and C and d(GCGCGCGCGC)2, J. Am. Chem. Soc., 108, 2058, 1986. 53. McGuiness, B.F. et al., Reductive alkylation of DNA by mitomycin A, a mitomycin with high redox potential, Biochemistry, 30, 6444, 1991. Copyright © 2005 CRC Press, LLC 54. Remers, W.A. et al., Conformations of complexes between mitomycins and decanucleotides. 3. Sequence specificity, binding at C-10, and mitomycin analogs, J. Med. Chem., 31, 1612, 1988. 55. Teng, S.P., Woodson, S.A., and Crothers, D.M., DNA sequence specificity for mitomycin crosslinking, Biochemistry, 28, 3901, 1989. 56. Borowy-Borowski, H., Lipman, R., and Tomasz, M., Recognition between mitomycin C and specific double-strand DNA sequences for cross-link formation, Biochemistry, 29, 2999, 1990. 57. Kumar, S., Johnson, W.S., and Tomasz, M., Orientation isomers of the mitomycin C interstrand cross- links in non-self-complementary DNA. Differential effect of two isomers on restriction endonuclease cleavage at a nearby site, Biochemistry, 32, 1364, 1993. 58. Kumar, S., Lipman, R., and Tomasz, M., Recognition of specific DNA sequences by mitomycin C for alkylation, Biochemistry, 31, 1399, 1992. 59. Li, V.-S. et al., Role of the C-10 subsituent in mitomycin C-1-DNA binding, J. Am. Chem. Soc., 118, 2326, 1996. 60. Kohn. H. et al., On the origins of the DNA sequence selectivity of mitomycin monoalkylation transformations, J. Am. Chem. Soc., 114, 9218, 1992. 61. Kumar, S.G. et al., Mitomycin C-DNA adducts generated by DT-diaphorase. Revised mechanism of the enzymatic reductive activation of mitomycin C, Biochemistry, 36, 14128, 1997. 62. Kumar, G.S. et al., Binding of 2,7-diaminomitosene to DNA: Model for recognition of DNA by activated DNA, Biochemistry, 34, 2662, 1995. 63. Kumar, G.S. et al., 2,7-Diaminomitosene, a monofunctional mitomycin C derivative, alkylates DNA in the major groove. Structure and base-sequence specificity of the DNA adduct and mechanism of alkylation, J. Am. Chem. Soc., 118, 9209, 1996. 64. Palom, Y. et al., Relative toxicities of DNA cross-links and monoadducts: new insights from studies of decarbamoyl mitomycin C and mitomycin C, Chem. Res. Tox., 15, 1398, 2002. 65. Gargulio, D. et al., Alkylation and crosslinking of DNA by the unnatural enantiomer of mitomycin C: mechanism of the DNA-sequence specificity of mitomycins, J. Am. Chem. Soc., 117, 9388, 1995. 66. Maliepaard, M. et al., Mitosene-DNA adducts. Characterization of two major DNA adducts formed by 1,10-bis(acetoxy)-7-methoxymitosene upon reductive activation., Biochemistry, 36, 9211, 1997. 67. Ouyang, A. and Skibo, E., The iminium ion chemistry of mitosene DNA alkylating agents. Enriched 13C-NMR studies, Biochemistry, 39, 5718, 2000. 68. Xing, C. and Skibo, E.B., Sigmatropic reactions of the aziridinyl semiquinone species. Why aziridinyl benzoquinones are metabolically more stable than aziridinyl indoloquinones, Biochemistry, 39, 10770, 2000. 69. Millard, J.T. and Beachy, T.M., Selective recognition of m5Cpg dinucleotide sequence in DNA by mitomycin C for alkylation and crosslinking, Biochemistry 32, 12850, 1993. 70. Tomasz, M. et al., The purine 2-amino group as the critical recognition element for sequence-specific alkylation and crosslinking of DNA by mitomycin C, J. Am. Chem. Soc., 120, 11581, 1998. 71. Sharma, M. and Tomasz, M., Conjugation of glutathione and other thiols with bioreductively activated mitomycin C. Effect of thiols on reductive activation rate, Chem. Res. Toxicol., 7, 390, 1994. 72. Sharma, M., He, Q.-Y., and Tomasz, M., Effects of glutathione on alkylation and crosslinking of DNA by mitomycin C. Isolation of a ternary glutathione-mitomycin-DNA adduct, Chem. Res. Toxicol., 7, 401, 1994. 73. Xu, B.H., Gupta, V., and Singh, S.V., Mitomycin sensitivity in human bladder cancer cells: possible role of glutathione and glutathione transferase in resistance, Arch. Biochem. Biophys., 308, 164, 1994. 74. Basu, A.K. et al., Effect of site specifically located mitomycin C-DNA monoadducts on in vitro DNA synthesis by DNA polymerases, Biochemistry, 32, 4708, 1993. 75. Abbas, T. et al., Differential activation of p53 by the various adducts of mitomycin C, J. Biol. Chem., 277, 40513, 2002. 76. Skibo, E.B. and Schultz, W.G., Pyrrolo[1,2-a]benzimidazole based aziridinyl quinones, J. Med. Chem., 36, 3050, 1993. 77. Schultz, W.G., Nieman, R.A., and Skibo, E.B., Evidence for DNA phosphate backbone alkylation and cleavage by pyrrolo[1,2-a]benzimidazoles, small molecules capable of causing sequence specific phosphodiester bond hydrolysis, Proc. Natl. Aced. Sci. USA, 92, 11854, 1995. Copyright © 2005 CRC Press, LLC 78. Bachur, N.R., Gordon, S.L., and Gee, M.V., A general mechanism for microsomal activation of quinone anticancer agents to free radicals, Cancer Res., 38, 1745, 1978. 79. Tomasz, M., Hydrogen peroxide generation during the redox cycle of mitomycin C and DNA-bound mitomycin C, Chem. Biol. Interact., 13, 89, 1976. 80. Lown, J.W., The molecular mechanism of antitumor action of the mitomycins, in Mitomycin C, Current Status and New Developments, Carter, S.T. and Crooke, S.T., Eds., Academic Press, Orlando, FL, 1979, pp. 5-26. 81. Ueda, K. et al., Inactivation of bacteriophage NX174 by mitomycin C in the presence of sodium hydrosulfite and cupric ions, Chem. Biol. Interact., 29, 145, 1980. 82. Pan, S.-S., Andrews, P.A., and Bachur, N.R., Reduction of mitomycin C by NADPH cytochrome p450 reductase and xanthine oxidase, Fed. Proc., 41, 1541, 1982. 83. Hashimoto, Y., Shudo, K., and Okamoto, T., Structures of modified nucleotides isolated from calf thymus DNA alkylated with reductively activated mitomycin C, Tetrahedron Lett., 23, 677, 1982. 84. Pritsos, C.A. and Sartorelli, A.C, Generation of reactive oxygen radicals through bioactivation of mitomycin antibiotics, Cancer Res., 46, 3528, 1986. 85. Remers, W.A. and Dorr, R.T. in The Alkaloids: Chemical and Biological Perspectives, Pelletier, S.W., Ed., Wiley, New York, 1988, Vol. 6. 86. Kinoshita, S. et al., Mitomycin derivatives. 1. Preparation of mitosane and mitosene compounds and their biological activities, J. Med. Chem., 14, 103, 1971. 87. Kinoshita, S. et al., Mitomycin derivatives. 2. Derivatives of decarbamoylmitosane and decarbam- oylmitosene, J. Med. Chem., 14, 109, 1971. 88. Kojima, R. et al., Some structure-activity relationships for mitomycin C (NSC-26980) derivatives in the treatment of l1210 leukemia, Cancer Chemother. Rep., 3, 121, 1972. 89. Iyengar, B.S. et al., Mitomycin C and porfiromycin analogs with substituted ethylamines at position 7 J. Med. Chem., 26, 16, 1983. 90. Imai, R. et al., Antitumor activity of 7-N-phenyl derivatives of mitomycin C in the leukemia p388 system, Gann, 71, 560, 1980. 91. Iyengar, B.S. et al., Development of new mitomycin C and porfiromycin analogs. J. Med. Chem., 24, 975, 1981. 92. Iyengar, B.S. et al., Mitomycin C analogs with secondary amines at position 7, J. Med. Chem., 26, 1453, 1983. 93. Sami, S.M. et al., Mitomycin C analogs with aryl substituents on the 7-amino group, J. Med. Chem., 27, 701, 1984. 94. Iyengar, B.S. et al., Metal

complexes of mitomycins, J. Med. Chem., 29, 144, 1986. 95. Kunz, K.R. et al., Structure-activity relationships for mitomycin C and mitomycin A analogs, J. Med. Chem., 34, 2281, 1991. 96. Ishioka, C. et al., Comparative studies on the action of 7-N-[2-[[2-(-L-glutamy- lamino)ethyl]dithio]ethyl]mitomycin C and of mitomycin C on cultured NL60-cells and isolated phage and plasmid DNA, Cancer Chemother. Pharmacol., 26, 117, 1990. 97. Bradner, W.T. et al, Antitumor activity and toxicity in animals of N-7[2-(4-nitrophenyldithio)ethyl]mit- omycin C (BMY-25067), Invest. New Drugs, 8, 51, 1990. 98. Bradner, W.T. et al., Antitumor activity and toxicity in animals of RR-150 (7-cysteaminomitosane), a new mitomycin derivative, Cancer Res., 44, 5619, 1984. 99. Sami, S.M. et al., Preparation and antitumor activity of new mitomycin A analogs, J. Med. Chem., 30, 168, 1987. 100. Li, V.-S. et al., The role of the C-10 substituent in mitomycin C-1-DNA binding, J. Am. Chem. Soc., 118, 2326, 1966. 101. Takanagi, H. et al., One step preparation of 6-demethyl-6-formylmitomycin C and cytotoxicity of its derivatives, J. Antibiot., 45, 1815, 1992. 102. Iyengar, B.S., Remers, W.A., and Bradner, W.T., Preparation and antitumor activities of 7-substituted- 1,2-aziridinomitosenes, J. Med. Chem., 29, 1864, 1986. 103. Fost, D.L., Ekwuribe, N.N., and Remers, W.A., Synthesis of 1-substituted-7-methoxymitosenes, Tet- rahhedron Lett., 131, 1973. 104. Taylor, W.G. et al., Mitomycin antibiotics. Synthesis and antineoplastic activity of mitosene analogs of the mitomycins, J. Med. Chem, 24, 1184, 1981. Copyright © 2005 CRC Press, LLC 105. Casner, M.L., Remers, W.A., and Bradner, W.T., Synthesis and biological activity of 6-substituted analogs of the mitomycins, J. Med. Chem, 28, 921, 1985. 106. Hodges, J.C., Remers, W.A., and Bradner, W.T., Synthesis and antineoplastic activity of mitosene analogs of the mitomycins, J. Med. Chem, 24, 1184, 1981. 107. Allen, Jr., G.R., Poletto, J.F., and Weiss, M.J., The mitomycin antibiotics. Synthetic studies. II. Synthesis of 7-methoxymitosene, an antibacterial agent, J. Am. Chem. Soc., 86, 3878, 1964. 108. Allen, Jr., G.R., Poletto, J.F., and Weiss, M.J., The mitomycin antibiotics. Synthetic studies. XV. The preparation of a related indoloquinone, J. Med. Chem., 10, 1, 1967. 109. Allen, Jr., G.R. and Weiss, M.J., The mitomycin antibiotics. Synthetic studies XVII. Preparation of 1-(β-substituted ethyl)indoloquinone analogs, J. Med. Chem., 10, 23, 1967. 110. Remers, W.A., Roth, R.H., and Weiss, M.J., The mitomycin antibiotics. Synthetic studies. VIII. Indoloquinone analogs with variations at C-5, J. Org. Chem., 31, 1012, 1966. 111. Iwata, H. and Goldberg, E.P., Tissue-binding macromolecular antitumor drugs for localized therapy: mitomycin C-concanavalin A conjugates, Polymer Reprints, 24, 73, 1983. 112. Kaneko, Y., Tanaka, T., and Iguchi, S., Preparation and properties of a mitomycin C-albumin conjugate, Chem. Pharm. Bull., 38, 2614, 1990. 113. Umemoto, N. et al., Conjugates of mitomycin C with the immunoglobulin M monomer fragment of a monoclonal anti-MM46 immunoglobulin M antibody with or without serum albumin as intermediary, J. Appl. Biochem., 6, 297, 1984. 114. Tsukada, Y., Characterization of human AFP producing stomach cancer xenotransplanted in nude mice and the effect of mitomycin C and antibody to human AFP of this tumor, Nippon Shokakibyo Gakki Zasshi, 82, 18, 1985. 115. Li, S. et al., Preparation of antigastric cancer monoclonal antibody MGb2-mitomycin C conjugate with improved antitumor activity, Bioconjugate Chem., 1, 245, 1990. 116. Tanaka, J.-I. et al., Preparation of a conjugate of mitomycin C and anti-neural cell adhesion molecule monoclonal antibody for specific chemotherapy against biliary tract carcinoma, Surg. Today, 28, 1217, 1998. 117. Huh, N. et al., Design, synthesis, and evaluation of mitomycin-tethered phosphorothioate oligodeox- ynucleotides, Bioconjugate Chem., 7, 659, 1996. 118. Paz, M.M, Das, A., and Tomasz, M., Mitomycin C linked to minor groove binding agents: synthesis, reductive activation, DNA binding and cross-linking properties and in vitro antitumor activity, Bioorg. Med. Chem., 7, 2713, 1999. 119. Okamoto, H. et al., Effect of 1-alkyl or 1-alkenyl-azocycloalkanone derivatives on penetration of mitomycin C through rat skin, Chem. Pharm. Bull., 35, 4605, 1987. 120. Shikano, M. et al., 1a-Docosahexaenoyl mitomycin C: a novel inhibitor of protein tyrosine kinase, Biochem. Biophys. Res. Commun., 248, 858, 1998. 121. Nagai, K. et al., Synthesis and antitumor activities of mitomycin C (1-3)-beta-D-glucan conjugate, Chem. Pharm. Bull., 40, 986, 1992. 122. Ghiorghis, A., Talebaum, A., and Clarke, R., In vitro antineoplastic activity of C-7-substituted mito- mycin C analogs MC-77 and MC-62 against human breast-cancer cell lines, Cancer Chemother. Pharmacol., 29, 290, 1992. 123. Yan, Z. et al., Synthesis of the anticancer drug mitomycin C-dextran conjugate with affinity to lymph and studies on its biological properties, Zhongguo Yaowu Huaxue Zashi, 9, 240, 1999. 124. Hoste, K., Schacht, and E., Rihova, B., Synthesis and biological evaluation of PEG-substituted macromolecular prodrugs from mitomycin C, J. Bioactive Compatible Polymers, 17, 123, 2002. 125. Kato, Y., Onishi, H., and Machida, Y., A novel water-soluble N-succinyl-chitosan-mitomycin C conjugate prepared by direct carbodiimide coupling: physicochemical properties, antitumor charac- teristics, and systemic retention, S.T.P. Pharma Sci., 10, 133, 2000. 126. Zalipsly, S. et al., New liposomal prodrug of mitomycin C, Proceedings of the 28th International Symposium on Controlled Release of Bioactive Materials and 4th Consumer Diversified Products Conference, San Diego, CA, June 23–27, 2001. 127. Iyengar, B.S. et al., Nucleotide derivatives of 2,7-diaminomitosene, J. Med. Chem., 31, 1612, 1988. 128. Iyengar, B.S., Dorr, R.T., and Remers, W.A., Additional nucleotide derivatives of mitosenes. Synthesis and activity against parenteral and multidrug resistant L1210 leukemia, J. Med. Chem., 34, 1947, 1991. Copyright © 2005 CRC Press, LLC 129. Sami, S.M. et al., Preparation and antitumor activity of new mitomycin A analogs, J. Med. Chem., 30, 168, 1987. 130. Bradner, W.T., Remers, W.A., and Vyas, D.M., Structure-activity comparison of mitomycin C and mitomycin A analogs, Anticancer Res., 9, 1095, 1989. 131. Bradner, W.T. et al., Antitumor activity and toxicity in animals of BMY-25282, a new mitomycin derivative, Cancer Res., 45, 6475, 1985. 132. Remers, W.A., Mitomycins, in Anticancer Agents Based on Natural Product Models, Cassady, J.M and Douros, J.D., Eds., Academic Press, New York, 1980, Chapter 4. 133. Orlamans, E.O. et al., Synthesis, mechanism of action, and biological evaluation of mitosenes, J. Med. Chem., 32, 1612, 1989. 134. Weiss, M.J. et al., The mitomycin antibiotics. Synthetic studies. XXII. Antibacterial structure-activity relationships in the indoloquinone series, J. Med. Chem., 11,742, 1968. 135. Allen, Jr., G.R., Poletto, J.F., and Weiss, M.J., Mitomycin antibiotics. Synthetic studies. III. Related indoloquinones, active antibacterial agents, J. Am. Chem. Soc., 86, 3878, 1964. 136. Allen, Jr., G.R., Poletto, J.F., and Weiss, M.J., Mitomycin antibiotics, Synthetic studies. XVII. Indolo- quinones with variants at the 2 position, J. Org. Chem., 30, 2897, 1965. 137. Poletto, J.F., Allen, Jr., G.R., and Weiss, M.J., Mitomycin antibiotics. Synthetic studies. XIX. Synthesis of indoloquinone analogs with certain C-3 variants, J. Med. Chem., 10, 95, 1967. 138. Leadbetter, G. et al., Mitomycin antibiotics. Synthesis of 1-substituted 7-methoxymitosenes, J. Org. Chem., 39, 3508, 1974. 139. Hirata, T., Yamada, Y., and Matsui, M., Synthetic studies on mitomycins. Synthesis of aziridinopyr- rolo[1,2-a]indoles, Tetrahedron Lett., 19, 1969. 140. Siuta, G.J., Franck, R.W., and Kempton, R.J., Studies directed toward a mitomycin synthesis, J. Org. Chem., 39, 3739, 1974. 141. Kametani, T. et al., Photooxygenation of 9-keto-7-methoxy-6-methyl-9H-pyrrolo[1,2-a]indole as a synthetic approach to mitomycins, Heterocycles, 4, 1637, 1976. 142. Nakatsubo, F. et al., Synthetic studies toward mitomycins. 2. Total synthesis of dl-porfiromycin, J. Am. Chem. Soc., 99, 8115, 1977. 143. Fukuyama, T. et al., Synthetic studies toward mitomycins. III. Total synthesis of mitomycins A and C, Tetrahedron Lett., 4295, 1977. 144. Fukuyama, T. and Yang, L., Practical synthesis of (±)–mitomycin C, J. Am. Chem. Soc., 111, 8303, 1989. 145. Danishefsky, S., Chemical explorations driven by an enchantment with mitomycinoids — a twenty year account, Synlett, 475, 1995. 146. Benbow, J.W., McClure, K.F., and Danishefsky, S.J., Intramolecular cycloaddition reactions of dienyl nitroso compounds: application to the synthesis of mitomycin K, J. Am. Chem. Soc., 115, 12305, 1993. 147. McClure, K.F. and Danishefsky, S.J., A novel Heck arylation reaction: rapid access to congeners of FR 900482, J. Am. Chem. Soc., 115, 6094, 1993. 148. Moertel, C.G., Reitemeir, R.J., and Hahn, R.G., Mitomycin C therapy in advanced gastrointestinal cancer. J. Am. Med. Assoc., 204, 1045, 1968. 149. Doll, D.C., Weiss, R.B., and Issel, B.F., Mitomycin: ten years after approval for marketing, J. Clin. Oncol., 3, 276, 1985. 150. Miller, T.P., McMahon, L.J., and Livingston, R.B., Extensive adenocarcinoma and large cell undif- ferentiated carcinoma of the lung treated with 5-FU, vincristine, and mitomycin C. Cancer Treat. Rep., 64, 1241, 1980. 151. Crooke, S.T. and Bradner, W.T., Mitomycin C: a review. Cancer Treat. Rev., 3, 121, 1976. 152. Early, K. et al., Mitomycin C in the treatment of metastatic transitional cell carcinoma of urinary bladder. Cancer, 31, 1150, 1973. 153. Baker, L.H., Opipari, M.I., and Izbicki, R.M., Phase II study of mitomycin C, vincristine, and bleomycin in advanced squamous cell carcinoma of the uterine cervix, Cancer, 38, 2222, 1976. 154. MacDonald, J. et al., 5-Fluorouracil (5-FU), mitomycin C (MMC), and Adriamycin (ADR) — FAM: a new combination chemotherapy program for advanced gastric carcinoma (abstract C-11), Proc. Am. Soc. Clin. Oncol., 17, 244, 1976. 155. Dorr, R.T. and Von Hoff, D.D., Cancer Chemotherapy Handbook, Appleton and Lange, Norwalk, CT, 1994, p. 719. Copyright © 2005 CRC Press, LLC 156. Kaplan, M.A. and Vyas, D.M., Crystalline 7-(dimethylaminomethylene)amino-9a-methoxymitosane, Ger. Offen. DE 3,605,523, 1986. 157. Bradner, W.T. et al., Antitumor activity and toxicity in animals of BMY-25282, a new mitomycin derivative, Cancer Res., 45, 6475, 1985. 158. Bradner, W.T. et al., Antitumor activity and toxicity in animals of RR-150 (7-cysteaminomitosane), a new mitomycin derivative, Cancer Res., 44, 5619, 1984. 159. Senter, P.D. et al., Reassignment of the structure of the antitumor agent RR-150, J. Antibiot., 41, 199, 1988. 160. Vyas, D.M. et al., Novel disulfide mitosanes as antitumor agents, Recent Adv. Chemother, Proc. Int. Congr. Chemother 14th, Anticancer Sect. 1, 517, 1985. 161. Bradner, W.T. et al., Antitumor activity and toxicity in animals of N-7-[2-(4-nitrophe- nyldithio)ethyl]mitomycin C (BMY-25067), Invest. New Drugs, 8, S1-S7, 1990. 162. De Boer, C. et al., Porfiromycin, a new antibiotic, Antimicrobial. Agents Annual, 1960, Plenum Press, New York, 1961, pp. 17-22. Copyright © 2005 CRC Press, LLC 24 Staurosporines and Structurally Related Indolocarbazoles as Antitumor Agents Michelle Prudhomme CONTENTS I. Introduction II. Mechanisms of Action A. Staurosporine and UCN-01 B. K-252a C. Rebeccamycin D. AT2433-A1 and B1 III. Total Syntheses of the Natural Products IV. Development of Drugs from the Natural Products V. Clinical Applications VI. Conclusions References I. INTRODUCTION Bacteria and primitive organisms are important sources of various compounds of biological interest. Among the diversity of bacterial metabolites are maleimido- or maleamido-indolocarbazoles. Some of them do not bear any sugar moiety (arcyriaflavins, tjipanazole J, K-252c, BE-13793C), and others have a carbohydrate moiety linked to only one indole nitrogen (rebeccamycin, AT2433-A1 and B1) or to both indole nitrogens (staurosporines, K-252a and b) (Figure 24.1).1–6 Staurosporine was isolated from Streptomyces staurosporeus7,8 in 1977, and 3 yr later arcyriaflavins were obtained from the fruiting bodies of slime molds Arcyria denudata.9 Later, more than 30 microbial stauro- sporine analogs were isolated from bacterial cultures. Indolocarbazoles have attracted considerable interest since the discovery in 1986 of the inhibitory properties of staurosporine toward protein kinase C (PKC), a kinase involved in many signal transduction pathways leading to a variety of cell responses, such as cell proliferation, differentiation, and gene expression.10 K-252a and analogs bearing a five-membered ring sugar moiety have been isolated from Nocardiopsis and Streptomyces species, and they are also PKC inhibitors.11–13 Staurosporine is a nonselective ATP-competitive kinase inhibitor that inhibits the different PKC isoenzymes, cyclin-dependent kinases (CDKs), and tyrosine kinases involved in cell proliferation.14–17 More recently, it has been shown that UCN-01, a 7-hydroxy staurosporine isolated from cultures of a Streptomyces strain,18 abrogates cell cycle arrest caused by DNA damage. This results in premature activation of mitosis in DNA-damaged cells, which leads to cell death.19,20 Copyright © 2005 CRC Press, LLC H H O N O O N X R2 R1 R R N N H H N N O O HO N N R H H H H R3 OH tjipanazole J R = CI, X = OH arcyriaflavin B : R = H OH K-252c : R = X = H arcyriaflavin C : R = OH tjipanazole E R1 = CI, R3 = CH2OH tjipanazole F2 R1 = CI, R2 = R3 = H tjipanazole F1 R2 = CI, R1 = R3 = H H R H N O N O H O N O N N N

N O O N N H H H H O HO OH HO NH BE-13793C O OR K-252a R = CH3 staurosporine R = H K-252b R = H UCN-01 R = OH H O N O O N O N N N N H H CI I O CI R O C OH OH HO OH O O OH O O Rebeccamycin AT2433-A1 R = CI OH AT2433-B1 R = H NH FIGURE 24.1 Indolocarbazoles from natural sources. The cytotoxicity of rebeccamycin and BE-13793C is linked to their capacity to inhibit topoi- somerases, nuclear enzymes that control the topological state of DNA and are involved in various vital processes such as replication and transcription. Rebeccamycin is only a topoisomerase I inhibitor, whereas BE-13793C inhibits both topoisomerase I and II.21,22 The antibiotics, AT2433- A1 and B1, structurally related to rebeccamycin, do not inhibit topoisomerases, and their cytotox- icity may be a consequence of their tight intercalative binding to DNA.23 Because of the promising biological activities of indolocarbazoles, large structure–activity relationship studies were carried out, which led to the discovery of more potent and more selective molecules. In this chapter, only indolocarbazole compounds bearing a sugar moiety are described. II. MECHANISMS OF ACTION A. STAUROSPORINE AND UCN-01 Staurosporine, initially described as a PKC inhibitor, proved to be a potent and nonselective adenosine triphosphate–competitive kinase inhibitor. It strongly inhibits serine/threonine kinases Copyright © 2005 CRC Press, LLC Lys33 Ala31 Phe80 Asp145 Leu134 2.7 Glu81 O H N N CH Phe82 Val18 Glu12 3 O CH O 3 NH 2.6 O CH3 N O N Gly13 Leu83 H H ⊕ 2.8 Ile10 2.9 O O Hls84 Gln85 O Gln131 Asp86 CDK-2-staurosporine Complex Chk-1-UCN-01 Complex FIGURE 24.2 (See color insert) Crystal structures of staurosporine and UCN-01 complexes with kinases. such as PKC, PKA, PKG, CDKs, CaMK, MLCK, S6-kinase, ERK-1, and aurora kinases, as well as tyrosine kinases such as PDGFR, c- and v-Src, EGFR, and many others.14,24,25 The majority of them are extremely sensitive to staurosporine, with IC50 values in the low nanomolar range. The crystal structure of staurosporine in complex with the cyclin-dependent kinase CDK2 provides insight into the interactions with a variety of kinases. The carbonyl and the NH of the lactam heterocycle seem to be always involved in hydrogen bonds with amino acid residues of all kinases (Figure 24.2).26 Staurosporine has also been cocrystallized with cAMP-dependent protein kinase PKA and Src tyrosine kinase CSK.27,28 More interesting is UCN-01, which exhibits greater selectivity toward protein kinases. UCN- 01 inhibits “classic” PKC isoenzymes but has no effect against the “atypical” PKC isoforms, which are only activated on their regulatory domain by phosphatidylserine and phosphatidylinositol-3,4,5- triphosphate.29 UCN-01 inhibits the growth of various human tumor cell lines. In contrast to staurosporine, UCN-01 is a relatively nonpotent direct inhibitor of CDK1 and CDK2, and it induces preferential G1-phase accumulation and reduction of cellular CDK2 activity by preventing the phosphorylation of Thr160 residue.30–32 Moreover, UCN-01 abrogates G2 and S checkpoints.33,34 In the cell cycle of normal cells, three principal checkpoints can be activated in response to DNA damage in the G1, S, and G2 phases. These checkpoints control the ability of cells to arrest the cell cycle, allowing time to repair the DNA. The G1 checkpoint allows DNA repair before DNA replication, the S checkpoint is involved in replication fork regulation, and the G2 checkpoint allows DNA repair before mitosis. The G1 checkpoint is dependent on the activity of p53. The p53 gene is mutated in nearly half of all human malignancies and in large proportions of premalignant Copyright © 2005 CRC Press, LLC Detection of DNA damage 2 Checkpoint Rad17 3 Loading ATR 4 Complex 5 (CLC) ATRIP P P Rad1 Checkpoint R P a Sliding P d s1 9 Hu Clamp BRCTs (CSC) Activation of Chk1 P – Cdc25 Cdc2 kept P P inactive Chk1 Cyclin P Y15 Cdc2 T161 + P P Wee1 FIGURE 24.3 Activation of Chk1. precursor lesions.35–37 In p53 mutated cancer cells, the G1 checkpoint is lacking, and the G2 checkpoint provides cancer cells with the opportunity to repair their DNA after damage. The combination of a DNA-damaging agent with molecules that can abrogate S and G2 arrest would enhance cytotoxicity selectively in the tumor cells. The arrest in the G2 phase after DNA damage seems to be mediated by Chk1 kinase inhibition. In the presence of DNA damage, Chk1 is activated by phosphorylation by ATM or ATR (Figure 24.3). Chk1 then phosphorylates and inactivates the Cdc25C phosphatase, thereby preventing activation of the cyclinB1/CDK1 mitotic complex through phosphorylation of Thr-14 and Tyr-15 residues con- trolled by Wee-1 kinase and Cdc25C phosphatase. Because Wee-1 kinase is relatively resistant to UCN- 01,14,38 this staurosporine analog should inhibit one or several kinases that phosphorylates Cdc25C on the Ser-16 residue (Chk1, Chk2, and cTAK1).39 It has been shown that UCN-01 inhibits Chk1 and cTAK1 with IC50 values of 11 and 27 nM, respectively.33 Initially, it was suggested that Chk2 was weakly inhibited by UCN-01 (IC50 1040 nM). A more recent report has shown that UCN-01 efficiently inhibits both Chk1 and Chk2 kinase activity immunoprecipitated from human colon carcinoma HCT116 cells (IC50 about 10 nM); however, inhibition of recombinant Chk2 requires much higher concentrations of UCN-01 (IC50 about 300 nM).40 In p53 mutant cells, low concentrations of UCN-01 caused S phase cells to progress to G2 before undergoing mitosis and death, whereas high concentrations caused a direct S-to-M premature mitosis.34 X-ray structures of Chk1 in complex with staurosporine and UCN- 01 have been determined. Both compounds bind in the ATP-binding pocket of the kinase (Figure 24.2).41 The selectivity of UCN-01 toward Chk1 over cyclin-dependent kinases can be a result of the interaction of the hydroxyl group on the lactam heterocycle with the ATP-binding pocket. The inhibition activities of UCN-01 and staurosporine (Ki, app) against Chk1, CDK1, CDK2, and CDK4 have been determined. The Ki, app values for UCN-01 in nanomoles are 5.6, 95, 30, and 3600, respectively, and for staurosporine, the values are 7.8, 1.0, 2.9, and 41, respectively.41 An additional target for the apoptosis-modulating activity of UCN-01 has recently been identified. UCN-01 inhibits 3-phosphoi- nositide-dependent protein kinase 1 (PDK1), which plays a central role in the regulation of the Akt- PKB survival pathway by phosphorylating Akt on Thr208 residue.42 The PDK1-Akt survival pathway represents a new and attractive target for cancer chemotherapy. B. K-252A K-252a was originally identified as a PKC inhibitor, but its effects are wide-ranging and include tyrosine kinase receptor TrkA and MLK3 inhibition and activation of PI3K and MEK signaling pathways. It also induces cell cycle arrest and apoptosis by inhibiting CDK1 and Cdc25C.11,43–45 Copyright © 2005 CRC Press, LLC C. REBECCAMYCIN Rebeccamycin was isolated in 1987 from cultures of Saccharothrix aerocolonigenes.46 Its antipro- liferative activity against various tumor cell lines is linked to its ability to inhibit topoisomerase I.47 Although rebeccamycin is structurally related to staurosporine, the biological targets of these microbial metabolites are quite different. Rebeccamycin is inactive against PKC, PKA, and topoi- somerase II but inhibits topoisomerase I,21 whereas staurosporine is a nonselective kinase inhibitor without activity against topoisomerases. Like camptothecin, rebeccamycin is a topoisomerase I poison, stabilizing the enzyme–DNA interaction in the form of a covalent intermediate called “cleavable complex” and preventing the religation of the cleaved DNA strand.48 More potent and soluble rebeccamycin analogs were prepared either by semisynthesis from rebeccamycin or by total synthesis.49-51 D. AT2433-A1 AND B1 AT2433-A1 and B1 (Figure 24.1) have been isolated from cultures of Actinomadura melliaura, with two other molecules bearing a primary amine instead of the methylamino group on the deoxysugar moiety.52 Their antitumor activity was evaluated against transplantable mouse leukemia P388. AT2433-A1 was at least eight times more potent than AT2433-B1.Their cytotoxicity may be a consequence of their tight intercalative binding to DNA, preferentially to GC-rich sequences, but contrary to rebeccamycin, these disaccharide indolocarbazoles have no inhibitory effect on topoi- somerase I.53,54 In contrast to AT2433-B1, its isomer, JDC-277, bearing an axial hydroxyl group instead of an equatorial N-methylamino group on the deoxysugar moiety, stimulates topoisomerase I-mediated DNA cleavage mainly at TG sites, showing that sequence-selective DNA interaction and topoisomerase I inhibition depends on the stereochemistry, and the substituents on the disac- charide unit. III. TOTAL SYNTHESES OF THE NATURAL PRODUCTS Because of the small quantities generally available from the natural sources and because of the possibility to have access to various analogs, total syntheses of natural indolocarbazoles have triggered considerable efforts.55 Early work centered on aglycon construction.56–63 The first total synthesis of staurosporine was achieved in 1995 by Danishefsky et al. in 27 steps (Figure 24.4), using sequential indole-Grignard additions to N-(benzyloxymethyl)-dibromomaleimide, followed by glycosylation with a 1,2-anhydrosugar, and then photocyclization. Coupling of the sugar moiety with the second indole nitrogen was performed from the exo-glycal intermediate, and after removal of the protective groups, the last step was the reduction of a carbonyl group of the imide hetero- cycle.64,65 In 1997, Wood et al.66,67 reported the synthesis of staurosporine and several natural compounds such as TAN-1030A, K-252a, RK286, and MLR-52 from the same intermediate obtained by cyclofuranosylation of the lactam aglycone in the presence of camphorsulfonic acid. A ring expansion approach was developed that allowed the transformation of a furanosylated indolocarbazole to a pyranosylated derivative (Figure 24.4). The 23-step stereocontrolled total synthesis of K-252a was described in 1999 starting from indole-3-acetic acid (Figure 24.5).68 The glycosylation was carried out before the introduction of the second indole moiety provided by tryptamine. The carboxylate substituent of the sugar was introduced via the addition of hydrogen cyanide on a carbonyl group. Four total syntheses of rebeccamycin have been reported (Figure 24.6).69-71 Kaneko et al. proposed two synthetic approaches to the chlorinated indolocarbazole aglycone, to which a α -bromo sugar was attached via the Koenigs–Knorr method. The chlorinated aglycone was obtained from N-benzyloxymethyl- dibromomaleimide and either chloroindole or 7-7-dichloroindigo. Danishefsky’s method used a bis-indolylmaleimide aglycone that was coupled to an anhydro sugar before photocyclization. Faul et al. performed the glycosylation of 7-chloro-indole-3-acetamide using the anhydro sugar described Copyright © 2005 CRC Press, LLC BOM H BOM O N N O O N O O Br Br PMBO O N N N N O O H O H O N NBOM O MgBr O NH O N staurosporine BOM O Danishefsky H N O N N O H H3CO N O OH O K-252a H N O N N O H DMB N N O N O O NH H O NH staurosporine N N O RK-286c H H 3C H3CO H H N O O N O N N O N N O H H O OH O OH N HO TAN-1030a MLR-52 Wood FIGURE 24.4 Total syntheses of staurosporine and related microbial metabolites. Copyright © 2005 CRC Press, LLC O H H N O N O N O OH O O N N N N N Br N N O O Ac H H H O O H TolO CI TolO HO O O O TolO TolO K-252a FIGURE 24.5 Total synthesis of K-252a by Kobayashi et al. by Danishefsky before condensation with methyl 7-chloroindole-3-glyoxylate. The oxidative cyclization of the bisindolylmaleimide was carried out using palladium(II) triflate. The synthesis of tjipanazole E was achieved by Bonjouklian et al. (Figure 24.7).72 Tjipanazole D, obtained by condensation of p-chlorophenylhydrazine and 1,2-cyclohexanedione followed by a Fischer indolization, was further glycosylated using α-bromo-tetraacetylglucose. Tjipanazoles F1 and F2 were synthesized by Gilbert et al. (Figure 24.7).73,74 Tjipanazole F2 was prepared from 1,2- bis(3-indolyl)ethane via a Mannich cyclization, followed by sequential bromination, glycosylation, oxidation, and halogen exchange. Tjipanazole F1 was synthesized via the same sequence of reactions from 1-(5-chloro-indolyl)-2-indolyl-ethane; the halogen atom was introduced at the begin- ning of the synthesis using 5-chloro-indole. Chisholm and Van Vranken described in 2000 the first total synthesis of AT2433-A1 and B1 via the glycosylation of the indoline intermediate prepared from N-methyl-dibromomaleimide and 7-chloro-indolylmagnesium bromide and indolylmagne- sium bromide, respectively (Figure 24.7).75 OBn OBn O N O H O N O O N O Kaneko et al. N Br Br CI H N N OBn O CI N N H CI OAc CI H O O N O O O CI OH AcO O HO AcO N N Rebeccamycin Br CI H H CI OH O OBn OBn Kaneko et al. O O N O BnO O NH2 N

N O CI H SEM CI N O CI O O Danishefsky OH O BnO N OBn H CI O Faul et al. FIGURE 24.6 Total syntheses of rebeccamycin. Copyright © 2005 CRC Press, LLC CI CI CI CI CI N N N N NHNH2, HCI O H H O OH Tjipanazole D O HO OAc OH AcO O OH AcO Tjipanazole E OAc Br CI H Br H N N N N N N H H OH H H H O OH D-xylose OH Tjipanazole F2 CI CI H CI H N N N N N N H H OH H H H O OH D-xylose OH Tjipanazole F1 O OBn O N O O N O O O teoc N O OBn BnO OBn N N N N H H H teoc = 2-trimethylsilylethyl carbamate O CI H H CI OH O O OH O O N O N MgBr AT2433-A1 Br OH Br CI NH FIGURE 24.7 Syntheses of tjipanazoles E, F1, F2, and AT2433-A1. IV. DEVELOPMENT OF DRUGS FROM THE NATURAL PRODUCTS Structural analogs of staurosporine were synthesized, aiming at higher selectivity. Among them, CGP 41 251 was developed by Novartis (Figure 24.8). CGP 41 251 was less potent but much more selective toward kinases than staurosporine. The antiproliferative activity of staurosporine was three to nine times stronger than that of CGP 41 251. The increased specificity of CGP 41 251 has not resulted in a cell-type-specific antiproliferative effect. In vivo, staurosporine and CGP 41 251 were active in the growth inhibition model, with a higher therapeutic index for CGP 41 251 in comparison to staurosporine.76,77 Like staurosporine, K-252a and UCN-01, CGP 41 251 has no effect on topoisomerase I.78 Copyright © 2005 CRC Press, LLC R1 R1 N O N O O N O R R N N O N N N N O O H H3C H H O HO HO N R N 2 O R O staurosporine R1 = R2 = H KT6528 R = OH, R' = CH2OH CGP 41251 R1 = H, R2 = COC6H5 K-252a R = H KT6124 R = NH2, R' = CH2OH KT6006 R = OH ICP-1 R = H, R' = COOCH3 H N O H N O N N O N N H O O H O O O O NH H N 2 O H CEP-2583 CEP-701 H2N H H H N O N O O N O S S N N N N N N O O O H H O HO O HO SB-218078 O O 3'-epi K-252a CEP1347 FIGURE 24.8 Indolocarbazoles structurally related to staurosporine and K-252. On the contrary, the K-252a derivatives KT6124, KT6528, and KT6006 (Figure 24.8) are topoisomerase I inhibitors and exhibit quite different effects on the cell cycle of HR-3Y1 cells.78,79 Staurosporine, CGP 41 251, and K-252a induce G2/M phase accumulation, whereas UCN-01 exerts a preferential G1 phase block. KT6124 and KT6528, which are weaker PKC inhibitors, induce an S phase delay, as observed with the topoisomerase I inhibitor camptothecin. KT6006 induces not only the S phase delay but also G2/M phase block. SB-218078, in which a tetrahydrofuranyl unit replaces the sugar unit (Figure 24.8), was found to enhance the cytotoxicity of DNA-damaging agents by preventing G2 arrest. As with UCN-01, SB-218078 is a potent inhibitor of human Chk1 kinase (Ki,app = 15 nM),41 and it abrogates G2 arrest caused by either γ-irradiation or a topoi- somerase I inhibitor.80 Avid binding of UCN-01 to human plasma proteins, not predicted from the animal studies, limits access to the tumor.81,82 A screen of related indolocarbazoles showed that ICP-1 (Figure 24.8), a K- 252a analog, was also able to abrogate S and G2 phase arrest and to enhance cytotoxicity induced by cisplatin only in p53 defective cells.83 In contrast to UCN-01, checkpoint abrogation by ICP-1 was Copyright © 2005 CRC Press, LLC only slightly inhibited by human plasma. CEP-2583 and CEP-701 (Figure 24.8), tyrosine kinase receptor inhibitors developed by Cephalon, are currently undergoing phase II clinical trials.84,85 This group has recently obtained, by semisynthesis from K-252a, the 3-epi diastereoisomer (Figure 24.8). Inverting the 3-alcohol resulted in a 20-nM inhibitor of VEGFR2 and a 1-nM inhibitor of TrkA tyrosine kinase compared with 43 nM and 13 nM for K-252a, respectively.86 CEP1347 is a semisynthetic K-252a derivative (Figure 24.8) bearing two methylthioethyl substituents in the 3 and 9 positions on the planar indolocarbazole. Both K-252a and CP1347 are neuroprotective compounds, and they mediate neuroprotective effects through the activation of neurotrophic signaling pathways. Both compounds inhibit JNK (c-Jun N-terminal kinase) pathway activation through similar mechanisms. Unlike K-252a, however, CP1347 does not inhibit the majority of common kinases such as PKC, PKA, MLCK, and PI-3 kinase, but it directly inhibits PAKs, a family of serine/threonine kinases, in a quite selective manner.87–89 BE-13793C (Figure 24.1), a topoisomerase I inhibitor, was isolated from cultures of Strep- toverticillium species.90 Modifications were carried out to enhance the solubility and to obtain compounds exhibiting more potent antitumor activities in vivo. Glycosylated derivatives ED-110 and NB-506 were prepared, as well as analogs such as J-107088 bearing the hydroxyl groups in various positions on the indolocarbazole framework (Figure 24.9).91–94 ED-110 and NB-506 stim- ulate DNA cleavage by topoisomerase I without any effect on topoisomerase II. The replacement of the NHCHO substituent of NB-506 with a longer side chain [NHCH(CH2OH)2] reinforces the interaction with DNA without affecting the intercalative binding process.95,96 A NB-506 regioisomer, with the two hydroxyl groups at positions 2 and 10 instead of 1 and 11, has lost its capacity to intercalate into DNA but remains an extremely potent topoisomerase I poison, showing that inter- calation into DNA is not required for inhibition of topoisomerase I by indolocarbazoles. NB-506 and its regioisomer stimulate cleavage at different sites on DNA. The 2,10 isomer is up to 100- fold more toxic to tumor cells than the parent NB-506.97 NB-506 induces apoptosis in leukemia cells via alterations of mitochondrial and caspases functions.98 The metabolism of NB-506 has been studied, both in vitro and in vivo, in animals and humans (Figure 24.10).99,100 In rodent plasma, NB-506 was converted to its deformylated derivative ED-501, whereas in dog and human plasma, it was converted to its anhydride ED-551, an inactive metabolite. It has been suggested that the conversion of NB-506 into anhydride ED-551 is catalyzed by metal ions with the formation of a complex in which the metal ion interacts with the three carbonyl groups. The replacement of the NHCHO group by a NHCH(CH2OH)2 substituent (compound D, Figure 24.10) significantly reduces the rate of metabolization into the inactive anhydride form.101 During the course of clinical trials of NB-506, Banyu Pharmaceuticals continued to synthesize NB-506 analogs, and from among several hundred compounds, J-107088 (Figure 24.9), which showed greater inhibitory activity than NB-506 toward topoisomerase I, has recently replaced NB-506 in the clinical trials.93 R R H O N O O N O O N O N N N N HO N N OH H H H HO OH HO OH H O O OH OH BE-13793C HO HO ED110 R = H NB506 R = CHCHO OH OH OH OH J-107088 R = NHCH(CH2OH)2 R = NHCHO Regioisomer of NB-506 FIGURE 24.9 Chemical structures of BE-13793C, semi-synthetic derivatives, and synthetic analogs. Copyright © 2005 CRC Press, LLC NHCHO NH2 O O O O N O O N O Human and dog plasma rodent plasma N N N N N N H H H HO OH HO OH HO OH O O O OH OH OH HO HO HO ED-551 NB506 ED-501 OH OH OH OH OH OH OH OH HN O N O N N H HO OH O OH HO D OH OH FIGURE 24.10 Metabolism of NB-506. Large structure–activity relationship studies have also been carried out on rebeccamycin analogs prepared either by semisynthesis or by total synthesis.50,51,102 Biosynthetic studies revealed that rebeccamycin is from two tryptophan, one glucose, and one methionine molecules.102 Addition of 6-, 5-, or 4-fluorotryptophan to the fermentation broths resulted in rebeccamycin analogs bearing, instead of chlorine atoms, fluorine atoms in the 2 and 10, 3 and 9, and 4 and 8, positions, respectively, on the indolocarbazole unit. Maximal topoisomerase I–induced cytotoxicity is imparted by 3,9- dihalogenated compounds. The 2,10-dihalogenated compounds exhibit less topoisomerase I selec- tive cytotoxicity, and the 1,11-disubstituted analogs have the least topoisomerase I–dependent cytotoxicity. These observations may be explained by interaction with other cytotoxic targets such as DNA or kinases. The 4,8-dihalogenated compounds are the least cytotoxic and the least active toward topoisomerase I or other targets. The methylation of the 4-hydroxyl group of the sugar moiety has little effect on topoisomerase I inhibition but results in increased cytotoxicity, very likely because of better penetration into the cells. The semisynthetic analog synthesis program developed by Bristol-Myers culminated in the discovery of the 6-N-diethylaminoethyl derivative, a water-soluble analog of rebeccamycin.49 NSC 655649, a tartrate salt, entered phase I studies (Figure 24.11). This compound has a broad spectrum of activity against solid tumors and, in contrast to rebeccamycin, inhibits topoisomerase II. NCS 655649 is currently undergoing phase II clinical trials for the treatment of hepatobiliary and gall bladder cancers.102 Interestingly, the introduction of an amino group at the 6-position on the sugar unit (Figure 24.11) strongly enhances the capacity of the drugs to interact with DNA but almost abolishes their poisoning effect on topoisomerase I, showing that DNA and topoisomerase I represent two independent targets that can both be used for the development of antitumor rebeccamycin derivatives.103 The main differences between the structures of rebeccamycin and staurosporine are the sugar moiety linked to one indole nitrogen in rebeccamycin and to both indole nitrogens in staurosporine and the imide function in the upper heterocycle of rebeccamycin instead of an amide function in Copyright © 2005 CRC Press, LLC N H O N O H O N O O N O R R N N H OH N N O HO H OH N N OH R H O OH O HO OH HO OH O OH OH OH H O 6'-amino rebeccamycins NSC 655649 (tartrate salt) R = NH2, NHCH3, N(CH3) N 2 R = CHO, CH2OH, NO2, NH2, OH, COOCH3 H R R R O N O O N O O N O R' N N N N N N N N N N H H H O O OH OH O OH HO HO HO OH OH OH OH OH OH R = CH3 or H R = CH3 or H, R' = H R = CH3 or H R = H, R' = Br or NO2 FIGURE 24.11 Semi-synthetic rebeccamycin analogs and aza-rebeccamycins obtained by total synthesis. staurosporine. These differences should be responsible for their enzyme selectivities. Therefore, structural modifications were performed on rebeccamycin to obtain staurosporine analogs, with the sugar moiety linked to both indole nitrogens (Figure 24.12).104,105 The anti-topoisomerase I activities of the unsubstituted imide and rebeccamycin are in the same range, whereas the activity of the regioisomeric mixture of amides is about 10-fold weaker. There is no doubt that the carbonyl group that distinguishes these compounds must be important for both DNA binding and topoisomerase I inhibition. None of these bridged compounds inhibits PKC. The two cationic molecules, bearing a N,N-diethylamino chain at the imide nitrogen or an amino substituent on the sugar moiety, exhibit enhanced affinity for DNA, and this translates at the biological level into a 10-fold higher cytotoxic index; however, they do not inhibit topoisomerase I, in contrast to the parent uncharged derivative (R′ = R″ = OH), which stimulates DNA single-strand breaks by topoisomerase I.106 In contrast to rebeccamycin and dechlorinated rebeccamycin, which exhibit nonselective cytotoxicity toward various tumor cell lines, some compounds in this bridged series, such as 3,9-dinitro, 3,9-dihydroxy, and 3,9-diformyl compounds, exhibit a marked selectivity, indicating targets that are differently expressed in the various tumor cell lines. Substitutions in 3,9-positions on the indolocarbazole framework of rebeccamycin have been carried out (Figure 24.11).107 Depending on the structural modifications, the rebeccamycin deriv- atives may or may not exhibit selectivity toward the various tumor cell lines tested. 3,9-Substituents Copyright © 2005 CRC Press, LLC R' H H O N O O N N O R R N N N N N N O O OH OH O OH R'' HO O HO O O regioisomeric mixture R'

= H, R'' = OH, R = H, CHO, CH2OH, OH, Br, NO2, NH2, COOCH3 R = H, R'' = OH, R' = OH, (CH2)2NEt2.HCI R = R' = H, R'' = CI, N3, NH2.HCI FIGURE 24.12 Staurosporine analogs from rebeccamycin. can enhance or abolish the cytotoxicity, but they can also induce selectivity. The 3,9-dihydroxy derivative is a DNA-binding topoisomerase I poison, but it also inhibits CDK1/cyclinB and CDK5/p25. It induces accumulation of L1210 cells in the G2+M phases, which seems to be correlated with CDK1/cyclinB inhibition. This multitarget drug represents a new lead for subsequent drug design in the rebeccamycin series. Another interesting series is that of the 7-aza-rebeccamycins obtained by total synthesis.108,109 Surprisingly, DNA binding experiments in the monoazaindole series showed major differences between the compounds bearing the carbohydrate moiety on either the azaindole moiety or the indole unit. Compounds with the sugar attached to the indole exhibit higher affinity for DNA and stronger topoisomerase I inhibition than the parent, dechlorinated rebeccamycin. When the sugar is linked to the azaindole, however, DNA binding properties are abolished and topoisomerase I inhibitory properties are strongly weakened. Even the introduction of a bromo or a nitro group is not sufficient to restore the affinity for DNA; however, these substitutions induced marked topoi- somerase I inhibition. Both monoaza or diaza rebeccamycins showed a selective action toward certain cell lines. Their cytotoxic profiles are quite different from those of the parent non-aza compounds, which are nonselective and display similar cytotoxicity against all the tumor cell lines, but like most of rebeccamycin derivatives, they induce a G2+M arrest. It can be concluded from these studies that if topoisomerase I remains the first target for rebeccamycin derivatives, they certainly have other biological targets. The first results show that kinases that control the cell cycle should also be targets for rebeccamycin analogs. V. CLINICAL APPLICATIONS Though there is no indolocarbazole on the market at present, some of them are undergoing clinical trials. Phase I and II trials of UCN-01 have been reported.110 In a phase I study, where the doses were administered by continuous infusion for 72 h, a prolonged half-life of 600 h was observed. The side effects were various, and among them, hyperglycemia needed insulin treatment. However, avid plasma protein binding of UCN-01 dictated a change in dose escalation and administration schedules. Phase I trials with shorter infusions are being completed. In a phase II study, one patient with melanoma achieved a partial response and encouraging results were obtained in a patient with refractory anaplastic large cell lymphoma.111–113 The maximal tolerated and recommended phase II dose was 42.5 mg/m2 per day for 3 d. A phase II clinical trial of the staurosporine analog CGP41251 (PKC412) has been reported. The patients with lymphoproliferative disorders were treated at three oral dose levels of 25, 150, and 225 mg/day for 14 d. The results show that the treatment is well tolerated and reduces the tumor load in chronic B-cell malignancies.114 Rebecca- mycin derivative NCS 655649 has entered into phase I and II studies. This drug may be safely given at a maximal tolerated dose of 572 mg/m2 in both single-dose and multiple-dose formats. Copyright © 2005 CRC Press, LLC Antitumor activity was observed in two heavily pretreated ovarian cancer patients and in one patient with a soft tissue sarcoma refractory to etoposide and doxorubicin. Preliminary data of a phase I study demonstrated antitumor activity in colorectal carcinoma. A phase II trial evaluated the efficacy of NSC 655649 in patients with advanced previously minimally treated metastatic colorectal cancer. At 500 mg/m2 once every 3 wk, the treatment was inactive.115–118 A phase II study of NCS-655649 in metastatic renal cell cancer showed a modest antitumor activity of this drug in patients with advanced this cancer.119 VI. CONCLUSIONS Indolocarbazoles represent a promising class of antitumor agents. Minimal structural modifications are able to modify the biological targets and induce selectivity toward tumor cell lines. Staurosporine is a nonselective kinase inhibitor, whereas the main target for rebeccamycin is topoisomerase I. Rebeccamycin exhibits no selectivity toward the various tumor cell lines, whereas 7-aza-rebeccamy- cins in which an indole moiety is replaced by a 7-azaindole unit are highly selective. Other structural modifications on the rebeccamycin framework such as substitutions in 3,9-positions may induce selectivity toward the tumor cell lines and strong inhibitory properties toward kinases involved in the progression of the cell cycle. No indolocarbazole has yet been launched on the market, but a staurosporine analog UCN-01 and a rebeccamycin derivative NCS 655649 have entered into phase I and phase II trials. UCN-01 and a K-252a analog, ICP-1, abrogate the DNA-damage induced checkpoints to cell-cycle progression in the G2 and in the S phases. This is probably the most important point because, in this manner, indolocarbazoles may exhibit highly selective activity against tumor cells without effect on normal cells, thereby achieving the ultimate goal for antitumor drugs. REFERENCES 1. Prudhomme, M., Indolocarbazoles as anti-cancer agents, Curr. Pharm. Des., 3, 265, 1997, and references provided therein. 2. Williams, D.E. et al., Holyrines A and B, possible intermediates in staurosporine biosynthesis produced in culture by a marine actinomycete obtained from the North Atlantic Ocean, Tetrahedron Lett., 40, 7171, 1999. 3. Schupp, P. et al., Staurosporine derivatives from the ascidian Eudistoma toealensis and predatory flatworm Pseudoceros sp., J. Nat. Prod., 62, 959, 1999. 4. Cantrell, C.H. et al., A new staurosporine analog from the prosobranch mollusk Coriocella nigra, Nat. Prod. Lett., 14, 39, 1999. 5. Hernandez, L.M.C. et al., 4-N-methyl-5-hydroxystaurosporine and 5-hydroxystaurosporine, new indolocarbazole alkaloïds from a marine Micromonospora sp. strain, J. Antibiot., 53, 895, 2000. 6. Schupp, P., Proksch, P. and Wray, V., Further new staurosporine derivatives from the ascidian Eudis- toma toealensis and its predatory flatworm Pseudoceros sp., J. Nat. Prod., 65, 295, 2002. 7. Omura, S. et al., A new alkaloid AM-2282 of Streptomyces origin. Taxonomy, fermentation, isolation and preliminary characterization, J. Antibiot., 30, 275, 1977. 8. Furusaki, A. et al., X-ray crystal structure of staurosporine: a new alkaloid from a Streptomyces strain, J. Chem. Soc. Chem. Commun., 800, 1978. 9. Steglich, W. et al., Indole pigments from the fruiting bodies of the slime mold Arcyria denudata, Angew. Chem. Int. Ed. Engl., 19, 459, 1980. 10. Tamaoki, T. et al., Staurosporine, a potent inhibitor of phospholipid/Ca++ dependent protein kinase, Biochem. Biophys. Res. Commun., 135, 397, 1986. 11. Kase, H., Iwahashi, K. and Matsuda, Y., K-252a, a potent inhibitor of protein kinase C from microbial origin, J. Antibiot., 39, 1059, 1986. 12. Cai, Y. et al., The staurosporine producing strain Streptomycetes longisporoflavus produces metabolites related to K-252a. Proposal for biosynthetic intermediates of K-252a, J. Antibiot., 49, 1060, 1996. Copyright © 2005 CRC Press, LLC 13. Yasuzawa, T. et al., The structures of the novel protein kinase C inhibitors K-252a, b, c, and d, J. Antibiot., 39, 1072, 1986. 14. Gray, N. et al., ATP-directed inhibitors of cyclin-dependent kinases, Curr. Med. Chem., 6, 859, 1999. 15. Hoehn, P., et al., 3-Demethoxy-3-hydroxystaurosporine, a novel staurosporine analog produced by a blocked mutant, J. Antibiot., 48, 300, 1995. 16. Rialet, V. and Meijer, L., A new screening test for antimitotic compounds using the universal M phase- specific protein kinase, p34cdc2/cyclin Bcdc13, affinity-immobilized on p13suc1-coated microtitration plates, Anticancer Res., 11, 1581, 1991. 17. Al-Obeidi, F.A. and Lam, K.S., Development of inhibitors for protein tyrosine kinases, Oncogene, 19, 5690, 2000. 18. Takahashi, I. et al., UCN-01 and UCN-02, new selective inhibitors of protein kinase C II. Purification, physico-chemical properties, structural determination and biological activities, J. Antibiot., 42, 571, 1989. 19. Jackson, J.R. et al., An indolocarbazole inhibitor of human checkpoint kinase (Chk1) abrogates cell cycle arrest caused by DNA damage, Cancer Res., 60, 566, 2000. 20. Yu, L. et al., UCN-01 abrogates G2 arrest through a Cdc2-dependent pathway that is associated with inactivation of the Wee1Hu kinase and activation of the Cdc25C phosphatase, J. Biol. Chem., 273, 33455, 1998. 21. Rodrigues Pereira, E. et al., Structure-activity relationships in a series of substituted indolocarbazoles: topoisomerase I and protein kinase C inhibition and antitumoral and antimicrobial properties, J. Med. Chem., 39, 4471, 1996. 22. Kojiri, K. et al., A new antitumor substance, BE-13793C, produced by a Streptomycete, taxonomy, fermentation, isolation, structure determination and biological activity, J. Antibiot., 44, 723, 1991. 23. Facompré, M. et al., DNA binding and topoisomerase I poisoning activities of novel disaccharide indolocarbazoles, Mol. Pharmacol., 62, 1215, 2002. 24. Meggio, F. et al., Different susceptibility of protein kinases to staurosporine inhibition. Kinetic studies and molecular bases for the resistance of protein kinase CK2, Eur. J. Biochem., 234, 317, 1995. 25. Mahedevan, D.; Bearss, D.J. and Vankayalapati, H., Structure-based design of novel anti-cancer agents targeting aurora kinases, Curr. Med. Chem.-Anti-Cancer Agents, 3, 25, 2003. 26. Lawrie, A.M. et al., Protein kinase inhibition by staurosporine revealed in details of the molecular interaction with CDK2, Nat. Struct. Biol., 4, 796, 1997. 27. Prade, L. et al., Staurosporine-induced conformational changes of cAMP-dependent protein kinase catalytic subunit explain inhibitory potential, Structure, 5, 1627, 1997. 28. Lamers, M.B.A.C. et al., Structure of the protein tyrosine kinase domain of C-terminal Src kinase (CSK) in complex with staurosporine, J. Mol. Biol., 285, 713, 1999. 29. Roy, K.K. and Sausville, E.A., Early development of cyclin dependent kinase modulators, Curr. Pharm. Des., 7, 1669, 2001. 30. Akiyama, T. et al., G1-checkpoint function including a cyclin-dependent kinase 2 regulatory pathway as potential determinant of 7-hydroxystaurosporine (UCN-01)-induced apoptosis and G1-phase accu- mulation, Jpn. J. Cancer Res., 90, 1364, 1999. 31. Sausville, E.A., Cyclin-dependent kinases: novel targets for cancer treatment, in Perry, M. Ed., 1999 Educational Book. American Society for Clinical Oncology, Alexandria, VA, 1999, 9. 32. Abe, S. et al., UCN-01 (7-hydroxystaurosporine) inhibits in vivo growth of human cancer cells through selective perturbation of G1 phase checkpoint machinery, Jpn. J. Cancer Res., 92, 537, 2001. 33. Busby, E.C. et al., The radiosensitizing agent 7-hydroxystaurosporine (UCN-01) inhibits the DNA damage checkpoint kinase hChk1, Cancer Res., 60, 2108, 2000. 34. Kohn, E.A. et al., Abrogation of the S phase DNA damage checkpoint results in S phase progression or premature mitosis depending on the concentration of 7-hydroxystaurosporine and the kinetics of Cdc25C activation, J. Biol. Chem., 277, 26553, 2002. 35. Zhou, B.-B.S. and Elledge, S.J., The DNA damage response: putting checkpoints in perspective, Nature, 408, 433, 2000. 36. Cuddihy, A.R. and O’Connell, M.J., Cell-cycle responses to DNA damage in G2, Int. Rev. Cytol., 222, 99, 2003. 37. Nyberg, K.A. et al., Toward maintaining the genome: DNA damage and replication checkpoints, Annu. Rev. Genet., 36, 617, 2002. Copyright © 2005 CRC Press, LLC 38. Graves, P.R. et al., The Chk1 protein kinase and the Cdc25C regulatory pathways are targets of the anticancer agent UCN-01, J. Biol. Chem., 275, 5600, 2000. 39. Tenzer, A. and Pruschy, M., Potentiation of DNA-damage-induced cytotoxicity by G2 checkpoint abrogators, Curr. Med. Chem.-Anti-Cancer Agents, 3, 35, 2003. 40. Yu, Q. et al., Inhibition of Chk2 activity and radiation-induced p53 elevation by the cell cycle checkpoint abrogator 7-hydroxystaurosporine (UCN-01), Proc. Am. Assoc. Cancer Res., 42, 800, 2001. 41. Zhao, B. et al., Structural basis for Chk1 inhibition by UCN-01, J. Biol. Chem., 277, 46609, 2002. 42. Sato, T.; Fujita, N. and Tsuruo, T., Interference with PDK1-Akt survival signaling pathway by UCN- 01 (7-hydroxystaurosporine), Oncogene, 21, 1727, 2002. 43. Roux, P.P. et al., K-252a and CEP1347 are neuroprotective compounds that inhibit mixed-lineage kinase-3 and induce activation of Akt and ERK, J. Biol. Chem., 277, 49473, 2002. 44. Angeles, T.S., Kinetics of trkA tyrosine kinase activity and inhibition by K-252a., Arch. Biochem. Biophys., 349, 267, 1998. 45. Chin, L.S. et al., K-252a induces cell cycle arrest and apoptosis by inhibiting Cdc2 and Cdc25c, Cancer Invest., 17, 391, 1999. 46. Bush, J.A. et al., Production and biological activity of rebeccamycin, a novel antitumor agent, J. Antibiot., 40, 668, 1987. 47. Marminon, C. et al., Syntheses and antiproliferative activities of 7-azarebeccamycin analogs bearing one 7-azaindole moiety, J. Med. Chem., 46, 609, 2003. 48. Bailly, C. et al., DNA cleavage by topoisomerase I in the presence of indolocarbazole derivatives of rebeccamycin, Biochemistry, 36, 3917, 1997. 49. Kaneko, T. et al., Water soluble derivatives of rebeccamycin, J. Antibiot., 43, 125, 1990. 50. Prudhomme, M., Recent developments of rebeccamycin analogs as topoisomerase I inhibitors and antitumor agents, Curr. Med. Chem., 7, 1189, 2000. 51. Prudhomme, M., Rebeccamycin analogs as anti-cancer agents, Eur. J. Med. Chem., 38, 123, 2003.

52. Matson, J.A. et al., AT2433-A1, AT2433-A2, AT2433-B1, and AT2433-B2 novel antitumor antibiotic compounds produced by Actinomadura melliaura. Taxonomy, fermentation, isolation and biological properties, J. Antibiot., 42, 1547, 1989. 53. Carrasco, C. et al., DNA sequence recognition by the indolocarbazole antibiotic AT2433-A1 and its diastereoisomer, Nucleic Acids Res., 30, 1774, 2002. 54. Facompré, M. et al., DNA binding and topoisomerase I poisoning activities of novel disaccharide indolocarbazoles, Mol. Pharmacol., 62, 1215, 2002. 55. Knölker, H-J. and Reddy, K.R., Isolation and synthesis of biologically active carbazole alkaloïds, Chem. Rev., 102, 4303, 2002. 56. Sarstedt, B. and Winterfeldt, E., Reactions with indole derivatives XLVIII, a simple synthesis of staurosporin aglycon, Heterocycles, 20, 469, 1983. 57. Magnus, P.D., Exon, C. and Sear, N.L., Indole-2,3-quinodimethanes. Synthesis of indolocarbazoles for the synthesis of the fused dimeric indole alkaloïd staurosporinone, Tetrahedron, 39, 3725, 1983. 58. Joyce, R.P., Gainor, J.A. and Weinreb, S.M., Synthesis of the aromatic and monosaccharide moieties of staurosporine, J. Org. Chem., 52, 1177, 1987. 59. Brenner, M. et al., Synthesis of arcyriarubin B and related bisindolylmaleimides, Tetrahedron, 44, 2887, 1988. 60. Hughes, I., Nolan, W.P. and Raphael, R.A., Synthesis of the indolo[2,3-a]carbazole natural products staurosporinone and arcyriaflavin B, J. Chem. Soc. Perkin Trans. I, 2475, 1990. 61. Moody, C.J. et al., Synthesis of the staurosporin aglycon, J. Org. Chem., 57, 2105, 1992. 62. Mc Combie, S.W. et al., Indolocarbazoles. 1. Total synthesis and protein kinase inhibiting character- istics of compounds related to K-252c, Bioorg. Med. Chem. Lett., 3, 1537, 1993. 63. Link, J.T., Gallant, M. and Danishefsky, S.J., The first synthesis of a fully functionalized core structure of staurosporine: sequential indolyl glycosidation by endo and exo glycals, J. Am. Chem. Soc., 115, 3782, 1993. 64. Link, J.T., Raghavan, S. and Danishefsky, S.J., First total synthesis of staurosporine and ent-stauro- sporine, J. Am. Chem. Soc., 117, 552, 1995. 65. Link, J.T. et al., Staurosporine and ent-staurosporine: the first total syntheses, prospects for a regiose- lective approach, and activity profiles, J. Am. Chem. Soc., 118, 2825, 1996. Copyright © 2005 CRC Press, LLC 66. Wood, J.L. et al., Design and Implementation of an efficient synthetic approach to furanosylated indolocarbazoles: total synthesis of (+)- and (-)-K252a, J. Am. Chem. Soc., 119, 9641, 1997. 67. Wood, J.L. et al., Design and Implementation of an efficient synthetic approach to pyranosylated indolocarbazoles: total synthesis of (+)-RK286c, (+)-MLR-52, (+)-staurosporine and (-)-TAN-1030a, J. Am. Chem. Soc., 119, 9652, 1997. 68. Kobayashi, Y.; Fujimoto, T. and Fukuyama, T., Stereocontrolled total synthesis of (+)-K252a, J. Am. Chem. Soc., 121, 6501, 1999. 69. Kaneko, T. et al., Two synthetic approaches to rebeccamycin, Tetrahedron Lett., 26, 4015, 1985. 70. Gallant, M., Link, J.T. and Danishefsky, S.J., A stereoselective synthesis of indole-b -N-glycosides: an application to the synthesis of rebeccamycin, J. Org. Chem., 58, 343, 1993. 71. Faul, M.M., Winneroski, L.L. and Krumrich, C.A., Synthesis of rebeccamycin and 11-dechlororebec- camycin, J. Org. Chem., 64, 2465, 1999. 72. Bonjouklian, R. et al., Tjipanazoles, new antifungal agents from the blue-green alga Tolypothrix tjipanasensis, Tetrahedron, 37, 7739, 1991. 73. Gilbert, E.J. and Van Vranken, D.L., Control of dissymmetry in the synthesis of (+)-tjipanazole F2, J. Am. Chem. Soc., 118, 5500, 1996. 74. Gilbert, E.J., Ziller, J.W. and Van Vranken, D.L., Cyclizations of unsymmetrical bis-1,2-(3- indolyl)ethanes: synthesis of (-)-tjipanazole F1, Tetrahedron, 53, 16553, 1997. 75. Chisholm, J.D. and Van Vranken, D.L., Regiocontrolled synthesis of the antitumor antibiotic AT2433- A1, J. Org. Chem., 65, 7541, 2000. 76. Meyer, T. et al., A derivative of staurosporine (CGP 41 251) shows selectivity for protein kinase C inhibition and in vitro anti-proliferative as well as in vivo anti-tumor activity, Int. J. Cancer, 43, 851,1989. 77. Hoehn, P. et al, 3-Demethoxy-3-hydroxystaurosporine, a novel staurosporine analog produced by a blocked mutant, J. Antibiot., 48, 300, 1995. 78. Akinaga, S., Diverse effects of indolocarbazole compounds on the cell cycle progression of ras- transformed rat fibroblast cells, J. Antibiot., 46, 1767, 1993. 79. Yamashita, Y. et al., Induction of mammalian DNA topoisomerase I mediated DNA cleavage by antitumor indolocarbazole derivatives, Biochemistry, 31, 12069, 1992. 80. Jackson J.R. et al., An indolocarbazole inhibitor of human checkpoint kinase (Chk1) abrogates cell cycle arrest caused by DNA damage, Cancer Res., 60, 566, 2000. 81. Fuse, E. et al., Unpredicted clinical pharmacology of UCN-01 caused by specific binding to human α1-acid glycoprotein, Cancer Res., 58, 3248, 1998. 82. Sausville, E. et al., Phase I trial of 72-hour continuous infusion of UCN-01 in patients with refractory neoplasms, J. Clin. Oncol., 19, 2319, 2001. 83. Eastman, A. et al., A novel indolocarbazole, ICP-1, abrogates DNA damage-induced cell cycle arrest and enhances cytotoxicity: similarities and differences to the cell cycle checkpoint abrogator UCN- 01, Molecular Cancer Therapeutics, 1, 1067, 2002. 84. Traxler, P. et al., Tyrosine kinase inhibitors: from rational design to clinical trials, Med. Res. Rev., 21, 499, 2001. 85. Laird, A.D. and Cherrington, J.M., Small molecule tyrosine kinase inhibitors: clinical development of anticancer agents, Expert Opin. Investig. Drugs, 12, 51, 2003. 86. Gingrich, D.E. and Hudkins, R.L., Synthesis and kinase inhibitory activity of 3-(S)-epi-K-252a, Bioorg. Med. Chem. Lett., 12, 2829, 2002. 87. Maroney, A.C. et al., CEP1347 (KT7515), a semi-synthetic inhibitor of the mixed lineage kinase family, J. Biol. Chem., 276, 25302, 2001. 88. Roux, P.P. et al., K252a and CP1347 are neuroprotective compounds that inhibit mixed-lineage kinase- 3 and induce activation of Akt and ERK, J. Biol. Chem., 277, 49473, 2002. 89. Nheu, T.V. et al., The K252a derivatives, inhibitors for the PAK/MLK kinase family selectively block the growth of RAS transformants, Cancer J., 8, 238, 2002. 90. Kojiri, K. et al., A new antitumor substance, BE-13793C, produced by a Streptomycete. Taxonomy, fermentation, isolation, structure determination and biological activity, J. Antibiot., 44, 723, 1991. 91. Tanaka, S. et al., A new indolopyrrolocarbazole antitumor substance, ED-110, a derivative of BE- 13793C, J. Antibiot., 45, 1797, 1992. Copyright © 2005 CRC Press, LLC 92. Kanzawa, F. et al., Antitumor activities of a new indolocarbazole substance, NB-506, and establishment of NB-506-resistant cell lines, SBC-3/NB, Cancer Res., 55, 2806, 1995. 93. Yoshinari, T. et al., Mode of action of a new indolocarbazole anticancer agent, J-107088, targeting topoisomerase I, Cancer Res., 59, 4271, 1999. 94. Ohkubo, M. et al., Synthesis and biological activities of NB-506 analogs modified at the glucose group, Bioorg. Med. Chem. Lett., 10, 419, 2000. 95. Yoshinari, T. et al., Induction of topoisomerase I-mediated DNA cleavage by a new indolocarbazole, ED-110, Cancer Res., 53, 490, 1993. 96. Bailly, C. et al., Substitution at the F-ring N-imide of the indolocarbazole antitumor drug NB-506 increases the cytotoxicity, DNA binding, and topoisomerase I inhibition activities, J. Med. Chem., 42, 2927, 1999. 97. Bailly, C. et al., Intercalation into DNA is not required for inhibition of topoisomerase I by indolo- carbazole antitumor agents, Cancer Res., 59, 2853, 1999. 98. Facompré, M. et al., Apoptotic response of HL-60 human leukemia cells to the antitumor drug NB- 506, a glycosylated indolocarbazole inhibitor of topoisomerase I, Biochem. Pharmacol., 61, 299, 2001. 99. Takenaga, N. et al., In vitro metabolism of a new anticancer agent, 6-N-formylamino-12,13-dihydro- 1,11-dihydroxy-13-(β-D-glucopyranosyl)-5H-indolo[2,3-a]pyrrolo[3,4-c]-5,7(6H)-dione (NB-506), in mice, rats, dogs and humans, Drug Metabol. Disposition, 27, 213, 1999. 100. Takenaga, N. et al., In vivo metabolism of a new anticancer agent, 6-N-formylamino- 12,13-dihydro- 1,11-dihydroxy-13-(β-D-glucopyranosyl)-5H-indolo[2,3-a]pyrrolo[3,4-c]-5,7(6H)-dione (NB-506), in rats and dogs: pharmacokinetics, isolation, identification, and quantification of metabolites, Drug Metabol. Disposition, 27, 205, 1999. 101. Goossens, J.-F. et al., Plasma stability of two glycosyl indolocarbazole antitumor agents, Biochem. Pharmacol., 65, 25, 2003. 102. Long, B.H. et al, Discovery of antitumor indolocarbazoles: rebeccamycin, NSC 655649, and fluor- oindolocarbazoles, Curr. Med. Chem.-Anti-Cancer Agents, 2, 255, 2002. 103. Anizon, F. et al., Rebeccamycin analogs bearing amine substituents or other groups on the sugar moiety, Bioorg. Med. Chem., 11, 3709, 2003. 104. Anizon, F. et al., Synthesis, biochemical and biological evaluation of staurosporine analogs from the microbial metabolite rebeccamycin, Bioorg. Med. Chem., 6, 1597, 1998. 105. Marminon, C. et al., Syntheses and antiproliferative activities of new rebeccamycin derivatives with the sugar unit linked to both indole nitrogen, J. Med. Chem., 45, 1330, 2002. 106. Facompré, M. et al., DNA targeting of two new antitumor rebeccamycin derivatives, Eur. J. Med. Chem., 37, 925, 2002. 107. Moreau, P. et al., Semi-synthesis, topoisomerase I and kinases inhibitory properties, and antiprolifer- ative activities of new rebeccamycin derivatives, Bioorg. Med. Chem., 11, 4871, 2003. 108. Marminon, C. et al., Syntheses and antiproliferative activities of 7-azarebeccamycin analogs bearing one 7-azaindole moiety, J. Med. Chem., 46, 609, 2003. 109. Marminon, C. et al., Syntheses and antiproliferative activities of rebeccamycin analogs bearing two 7-azaindole moieties, Bioorg. Med. Chem., 11, 679, 2003. 110. Dancey, J. and Sausville, E.A., Issues and progress with protein kinase inhibitors for cancer treatment, Nat. Rev. Drug Discovery, 2, 297, 2003. 111. Walder, S., Perspectives for cancer therapies with CDK2 inhibitors, Drug Resistance Update, 4, 347, 2001. 112. Senderowicz, A.M., Preclinical and clinical development of CDK inhibitors, Proceedings 1st Inter- national Symposium on Signal Transduction Modulators in Cancer Therapy, Amsterdam, 2002, p 41. 113. Sausville, E.A. et al., Signal transduction-directed cancer treatments, Ann. Rev. Pharmacol. Toxicol., 43, 199, 2003. 114. Ganeshaguru, V.A. et al., A novel treatment approach for low grade lymphoproliferative disorders using PKC412 (CGP41251), an inhibitor of protein kinase C, Hematol. J., 3, 131, 2002. 115. Merchant, J. et al., Phase I clinical and pharmacokinetic study of NSC 655649, a rebeccamycin analog, given in both single-dose and multiple dose-formats, Clin. Cancer Res., 8, 2193, 2002. 116. Tolcher, A.W. et al., Phase I and pharmacokinetic study of NSC 655649, a rebeccamycin analog with topoisomerase inhibitory properties, J. Clin. Oncol., 19, 2937, 2001. Copyright © 2005 CRC Press, LLC 117. Dowlati, A. et al., Phase I and pharmacokinetic study of rebeccamycin analog NSC 655649 given daily for five consecutive days, J. Clin. Oncol., 19, 2309, 2001. 118. Goel, S. et al., A phase II study of rebeccamycin analog NSC 655649 in patients with metastatic colorectal cancer, Invest. New Drugs, 21, 103, 2003. 119. Hussain, M. et al., A phase II study of rebeccamycin analog (NSC-655649) in metastatic renal cell cancer, Invest. New Drugs, 21, 465, 2003. Copyright © 2005 CRC Press, LLC 25 Combinatorial Biosynthesis of Anticancer Natural Products Michael G. Thomas, Kathryn A. Bixby, and Ben Shen CONTENTS I. Introduction II. Requirements for Combinatorial Biosynthesis III. Examples of Combinatorial Biosynthesis of Anticancer Natural Products A. Daunorubicin and Doxorubicin (Type II Polyketide Synthase) 1. Biosynthetic Gene Cluster and Biosynthetic Mechanisms 2. Combinatorial Biosynthesis B. Mithramycin (Type II PKS) 1. Biosynthetic Gene Cluster and Biosynthetic Mechanisms 2. Combinatorial Biosynthesis C. Ansamycins (Type I PKS) 1. Biosynthetic Gene Clusters and Biosynthetic Mechanisms 2. Combinatorial Biosynthesis D. Epothilones (Hybrid PKS-NRPS) 1. Biosynthetic Gene Cluster and Biosynthetic Mechanisms 2. Combinatorial Biosynthesis E. Leinamycin (Hybrid NRPS-PKS Featuring AT-Less PKS) 1. Biosynthetic Gene Clusters and Biosynthetic Mechanisms 2. Combinatorial Biosynthesis F. Enediynes (Iterative Type I PKS) 1. Biosynthetic Gene Clusters and Biosynthetic Mechanisms 2. Combinatorial Biosynthesis G. Aminocoumarins (Mixed Biosynthesis Featuring Atypical NRPS) 1. Biosynthetic Gene Clusters and Biosynthetic Mechanisms 2. Combinatorial Biosynthesis IV. Summary, Challenges, and Prospects Acknowledgments References I. INTRODUCTION Natural products have an excellent track record as sources of anticancer drugs, with 60% of today’s available anticancer drugs originating from natural products, their derivatives, or mimics.1 Tradi- tional natural product drug discovery approaches depend primarily on bioassay-guided screening for lead identification and chemical synthesis and medicinal chemistry for lead development. Two historical shortcomings of such endeavors are the small quantities of materials available from nature and the difficulty of their chemical total synthesis and structural derivatization. Total synthesis has Copyright © 2005 CRC Press, LLC limited practical value for the preparation of complex natural products. Analog generation by chemical modification remains problematic because of the rich variety of functional groups that are characteristic of natural products and often requires multiple protection and deprotection steps, making the process quite difficult and laborious. If these hurdles of traditional natural products drug discovery can be overcome, we will have many more successful drugs derived from natural products. Combinatorial biosynthesis offers promising solutions to both of those problems. Specific structural alteration in the presence of abundant functional groups can often be achieved by rational engineering of the biosynthetic machinery, and the target molecules can therefore be produced by a recombinant organism that is amenable for large-scale fermentation, providing sufficient quantities of natural products and their structural analogs by mass production, thereby lowering the production cost and

reducing the environmental concerns associated with conventional chemical syntheses.1–18 The field of combinatorial biosynthesis started to become a reality with the seminal work of Sir David Hopwood and colleagues in the 1980s. The heterologous expression of the actinorhodin biosynthetic gene cluster19 provided a basis for novel metabolite production as well as introducing the concept of “hybrid” natural products.20 Since this work, the definition of combinatorial biosyn- thesis as a field has come to embrace more “traditional” methodologies of metabolic engineering and all aspects of molecular biological manipulations used to alter the production of the natural product of interest. Therefore, in this chapter we include targeted gene disruption/replacement, heterologous expression, hybrid pathway construction, enzyme and pathway evolution, and com- binations of all these techniques under the umbrella of combinatorial biosynthesis. Page limitations prevent us from summarizing all of the work on combinatorial biosynthesis of anticancer natural products, and so we have chosen a selected set of natural products to showcase the effectiveness and potential utilities of combinatorial biosynthesis in the production and development of novel natural product-based anticancer drugs. Although significant progress has also been made in plant natural products by cell culture technology21–23 and fungal natural products by combinatorial biosynthesis methods,24–26 we have limited the scope of this chapter to bacterial natural products. II. REQUIREMENTS FOR COMBINATORIAL BIOSYNTHESIS Four prerequisites must be satisfied before any attempt to implement combinatorial biosynthesis strategies for natural product structural diversity. They are the availability of the gene clusters encoding the production of a particular natural product or family of natural products, genetic and biochemical characterizations of the biosynthetic machinery for the targeted natural products to a degree that the combinatorial biosynthesis principles can be rationally applied to engineer the novel analogs, expedient genetic systems for in vivo manipulation of genes governing the production of the target molecules in their native producers or heterologous hosts, and production of the natural products or their engineered analogs to levels that are appropriate for detection, isolation, and structural and biological characterization.27 Although each of these requirements is essential to realize the full potential of combinatorial biosynthesis, establishing an expedient genetic system for in vivo manipulation of the biosynthetic machinery of the targeted metabolites is of paramount importance. This requires the development or engineering of the necessary vectors (e.g., plasmids, cosmids, and Escherichia coli–Streptomyces artificial chromosomes) and hosts (e.g., Streptomyces coelicolor CH999 and Streptomyces lividans K4-114), mechanisms for the introduction of DNA into the host (e.g., polyethylene glycol- or CaCl2- assisted transformation, electroporation, and E. coli–Streptomyces conjugation), precursor availabil- ity (e.g., methylmalonyl–CoA, propionyl–CoA, and nonproteinogenic amino acids), posttransla- tional modifications (e.g., phosphopantetheinylation), cofactor availability (e.g., coenzyme F420), and regulatory issues (e.g., butyrolactone production). Empirical circumstances may dictate that a genetic system be developed for each natural product; however, a more desirable goal would be the establishment of a limited number of heterologous hosts whereby designed genetic engineering experiments can be carried out for any natural product or family of natural products. Copyright © 2005 CRC Press, LLC III. EXAMPLES OF COMBINATORIAL BIOSYNTHESIS OF ANTICANCER NATURAL PRODUCTS Selected in this section are only a few examples of combinatorial biosynthesis of anticancer natural products, illustrating the principle and practice of combinatorial biosynthesis of anticancer natural products. We have based our choice of the natural products on the following three criteria: the biosynthetic machinery and fundamental chemistry and biochemistry that governs their biosynthe- sis, the combinatorial biosynthesis strategies that have been applied to their engineering, and the clinical significance or potential of the natural products as anticancer drugs. Readers are directed to Table 25.1 for a more comprehensive summary of combinatorial biosynthesis of anticancer natural products of microbial origin covering the literature up to early 2004. A. DAUNORUBICIN AND DOXORUBICIN (TYPE II POLYKETIDE SYNTHASE) Daunorubicin (DNR) and doxorubicin (DXR), members of the anthracycline family of antitumor antibiotics, have been widely used clinically in various cancer chemotherapies, and DXR has been considered one of the most successful anticancer drugs ever developed.28 The drugs intercalate into DNA noncovalently, thus preventing DNA template function as well as stabilizing the DNA–topoi- somerase II cleavable complex.29–31 The stabilization of this complex leads to DNA strand scission. Although anthracyclines have been shown to have many different effects on cellular physiology, this DNA strand-breakage is generally accepted as their primary activity.29–31 DNR was first isolated from a Streptomyces peucetius strain and was subsequently found in a number of other Streptomyces species. In contrast, DXR was isolated only from mutant strains that were derived from the DNR-producing S. peucetius wild-type strain upon chemical mutagenesis.32 This was somewhat surprising because DXR has an additional HO- group at C-14 compared to DNR. DNR and DXR are structurally characterized by a napthacenequinone aglycon that is glycosylated with a deoxy aminosugar, L-daunosamine, at C-7-O (Figure 25.1A). Isotopic labeling experiments established that the DNR and DXR aglycones are derived from one propionate and nine acetates, whereas the L-daunosamine moiety is derived from L-glucose.28,33 1. Biosynthetic Gene Cluster and Biosynthetic Mechanisms The Hutchinson group began isolating and sequencing the DNR/DXR biosynthetic gene clusters in 1989 from S. peucetius and S. peucetius subsp. caesius, with the latter strain producing DXR.34 This effort continued for the next 10 yr and was complemented by parallel studies by the Strohl group on the DNR biosynthetic gene cluster from Streptomyces sp. strain C5,28 leading to the eventual cloning, sequencing, and characterization of the entire DNR/DXR cluster. Central to the cluster are the dps genes that encode a type II polyketide synthase (PKS) for aglycone biosynthesis, the dnm genes that encode enzymes for TDP-L-daunosamine biosynthesis, and the dnr and dox genes for additional modifications of the DNR/DXR aglycone (Figure 25.1A and 25.1B). Many of these genes have been exploited successfully for the production of novel DNR/DXR analogs and other metabolites by combinatorial biosynthesis methods.28,35,36 DpsA-G and DpsY together catalyzes the assembly of the DNR/DXR aglycone backbone from one unit of propionyl-CoA and nine units of malonyl-CoA (Figure 25.1A). The process is initiated by DpsCDG to form the first intermediate of β-ketovaleryl-S-DpsG, with DpsC selecting for propionyl- CoA as a starter unit.35 Further elongation of β-ketovaleryl-S-DpsG by DpsAB affords the linear decaketide intermediate, and DpsAB together dedicate the number of the malonyl-CoA extender units added to the growing acyl-S-DpsG intermediate.37 The resultant reactive linear decaketide intermediate undergoes a series of facile transformations including regiospecific reduction (by DpsE) and multiple intramolecular cyclizations (by DpsFY), yielding the first isolatable biosynthetic intermediate, 12- deoxyaklanonic acid,33,38–41 a key intermediate found in the biosynthesis of many other anthracyclines such as rhodomycin and aclacinomycin A.42 Additional modifications of 12-deoxyaklanonic acid by Copyright © 2005 CRC Press, LLC TABLE 25.1 Anticancer Natural Productsa Biosynthetic Gene Cluster (Whole Heterologous Anticancer Antibiotic Producing Organism or Partial) Expression Combinatorial Biosynthesis Aclacinomycin Streptomyces galilaeus ATCC 31615 182 66 66,183 Actinomycin D Streptomyces chrysomallus 184 185 186 Ansamitocin Actinosynnema pretiosum 70 187 81 Aureothin Streptomyces thioluteus HKI-227 188 188 — Bleomycin Streptomyces verticillus ATCC15003 189-191 107,192-195 — Borrelidin Streptomyces parvulus Tü4055 196 — — C-1027 Streptomyces globisporus 131 — 131 Calicheamicin Micromonaspora echinospora ssp. calichensis 130 120 — Clorobiocin Streptomyces roseochromogenes DS 12976 143 197-199 152-154 Coumermycin A1 Streptomyces rishiriensis DSM 40489 144 197,200 150,151,201 Daunorubicin/Doxorubicin Streptomyces peucetius (spp casius, dxr producer) 28,35,36,48 37,202,203 43,183,203 Dynemicin Micromonospora chersina 128 — — Elloramycin Streptomyces olivaceus Tü2353 204-206 204,205,207,208 209,210 Epothilones Sorangium cellulosum So ce90 96-98 97,101,107,108,211 211-213 Esperamicin Actinomadura verrucosospora 128 — — Fredericamycin Streptomyces griseus 214 214 214 Geldanamycin Streptomyces hygroscopicus var. geldanus NRRL 3602 79 — — Gilvocarcin V Streptomyces griseoflavus Gö 3592 215 215 — Hedamycin Streptomyces griseoruber (ATCC15422) 216 — — Jadomycin Streptomyces venezuelae ISP5230 217-222 203,223 203,220 Landomycin A Streptomyces cyanogenus S136 (A) 224 225 226,227 Copyright © 2005 CRC Press, LLC Leinamycin Streptomyces atroolivaceus S-140 114,115 — 115 Maduropeptin Actinomadura madurae 128 — — Mithramycin Streptomyces argillaceus 57-59,63,228 66,229-231 51,65,66,227,232,233 Mitomycin C Streptomyces lavendulae NRRL 2564 234,235 236,237 — Macrotetrolide Streptomyces griseus DSM40695 238,239 238 238,240,241 Neocarzinostatin Streptomyces carzinostaticus 128,136 — — Nogalamycin Streptomyces nogalater 242,243 66,244 66,183,245 Novobiocin Streptomyces spheroides NCIB 11891 142,246 246-249 150,250 Oleandomycin Streptomyces antibioticus ATCC 11891 251-259 252,253,255,259-261 258-260,262-264 Puromycin Streptomyces alboniger 265 266 — Rebeccamycin Saccharothrix (Lechevalieri) aerocolonigenes ATCC 39243 267,268 269 269,305 Rifamycin B Amylcolatopsis mediterranei S699 69,72,78 187 81,270,271 Saframycin Myxococcus xanthus 272,273 — — Simocyclinone D8 Streptomyces antibioticus Tu6040 145,274 — 145 Sparsomycin Streptomyces sparsogenes 275,276 — — Staurosporine Streptomyces sp. TP-A0274 277 277 — Tetracenomycin C Streptomyces glaucescens 278,279 208,278,280-288 48,67,201,203,282,286,289-292 Urdamycin Streptomyces fradiae Tü2717 293-296 293 226,227,297 294,295,298,299 Valanimycin Streptomyces viridifaciens MG456-hF10 300 300 — Vicenistatin Streptomyces halstedii 301 301 — a Numbers in final three columns are reference numbers. Copyright © 2005 CRC Press, LLC (A) O DpsCDG H3C O O O COOH DpsAB DpsEFY O CH + S-CoA H 3 3C CO-S-DpsG S-DpsG O O O O O O O CH3 HO S-CoA HO S-CoA O O O O O OH O OH O (8 x) 12-deoxyaklanonic acid O COOR O COOCH3 O R COOCH3 DnrGC CH DnrD O 3 CH DnrEF DnmS OH 3 OH OH O OH O OH O OH O OH O OH OH TDP-daunosamine aklanonic acid (R=H) aklavinone aklaviketone (R=H) aklanonic acid methyl ester (R=CH3) ε-8-rhodomycinone (R=OH) O OH COOCH3 O OH R1 R2 O OH O 14 CH DnrP/DoxA/DnrK 3 CH3 DoxA CH2OH OH 7 OH 7 OH OH O OH O OR3O OH O OCH3O OH O H3C O H O H O 3C 3C H OH 2N NH NH OH 2 OH 2 Rhodomycin D 13-deoxycarminomycin (R2=H/H; R2=CH3) doxorubicin carminomycin (R1=O; R2=H) daunorubicin (R1=O; R2=CH3) (B) CH3 O DnmLMUTJ O DnmV O OH O CH 4' CH DnmS 3 3 Rhodomycin D daunorubicin HO OPO3 OTDP HO OTDP doxorubicin OH NH NH 2 2 ε-ehodomycinone TDP-daunosamine O OH O HO R O DnmS OH 4' CH3 4'-epi-Rhodomycin D AvrE or EryBIV OTDP OCH O OH O NH 3 2 ε-rhodomycinone H3C O TDP-4'-epi-daunosamine HO 4' NH2 4'-epi-daunorubicin (R=CH3) 4'-epi-doxorubicin (R=CH2OH) (C) OH OH OH H3C OH H3C OH OH O OH CH OH O 3 O O HO CH HO OH OH 3 OH CH O OH 3 O O HO CH3 HO O O CH3 HO HO O O O O O O OH O OH O O OH O OH O OH O OH O OH UWM5 SEK15 SEK15b SEK4 SEK4b SEK43 SEK34 FIGURE 25.1 (A) Structures and biosynthesis of daunorubicin and doxorubicin. (B) Structures and biosyn- thesis of TDP-daunosamine and engineered biosynthesis of 4′-epi-daunorubicin and 4′-epi-doxorubicin. (C) Structures of novel metabolites produced by hybrid type II PKSs containing subunits of the daunorubicin/dox- orubicin type II PKS. the so-called tailoring enzymes that act after the Dps PKS enzymes finally furnish the myriad of functional groups characteristic of aromatic polyketide metabolites such as DNR/DXR. These fun- damental genetic, biochemical, and chemical characterizations of the DNR/DXR biosynthetic machin- ery set the stage to formulate strategies, design experiments, and develop tools for combinatorial biosynthesis. The choice of the starter unit (as dedicated by DpsC), the number of the elongation units (as controlled by DpsAB), the regiochemistry of reduction and cyclization of the nascent linear polyketide (as determined by DpsEFY), and the multitude of post-PKS modifications (as furnished by DnrCDEFGKP and DoxA) represent outstanding opportunities for combinatorial biosynthesis and have all been successfully exploited to engineer structural diversity of aromatic polyketides such as DNR/DXR (Figure 25.1A and 25.1B). Copyright © 2005 CRC Press, LLC The biosynthesis of TDP-L-daunosamine proceeds via six catalytic steps, starting from D- glucose-1-phosphate, and DnmLMJVUT are sufficient to catalyze this process.43 Subsequent attach- ment of TDP-L-daunosamine to ε-rhodomycinone by the DnmS glycosyltransferase yields the first glycosylated product rhodomycin D (Figure 25.1B).35 Many of the deoxy sugars found in natural products are biosynthesized by analogous pathways and often share common intermediates, and the glycosyltransferases show significant flexibility in TDP-deoxy sugar recognition and transfer. Thus, complementing aglycone engineering, the deoxy sugar biosynthetic pathways and the cor- responding glycosyltransferases provide another dimension for natural production structural diver- sity by combinatorial biosynthesis. 2. Combinatorial Biosynthesis One of the most informative experiments for determining the role a particular enzyme plays in the biosynthesis of a metabolite is the inactivation of the gene encoding an enzyme of interest, followed by isolation and structural characterization of the biosynthetic intermediate accumulated by the mutant strains. Thus, using either targeted gene disruption or random

mutagenesis to inactivate genes, the roles that many, if not all, of the encoded enzymes play in biosynthesis can be deciphered. An additional benefit of this approach is that the intermediates accumulated by the mutants greatly facilitate the structural–activity relationship analysis of the natural products, and some of these intermediates may have biological activity distinct from the parent natural product. Although the intermediates of the DNR/DXR pathway have not proven fruitful in generating compounds with improved biological activities, the thorough analysis of this pathway serves an excel- lent example illustrating the potential and effectiveness of identifying biosynthetic intermediates via targeted gene disruption. The details of these studies have been reviewed elsewhere,27,28,35 and numerous compounds have been identified, including aklaviketone,44 aklavinone,45 ε-rhodomycinone,45,46 and rhodomycin D47,48 (Figure 25.1A). Although many of these intermediates were isolated from strains that were generated by random mutagenesis, the availability and characterization of the DNR/DXR biosynthetic gene cluster provided an alternative to do so rationally by targeted gene disruption. Gene inactivation can be combined with gene complementation with homologs from other path- ways to provide alternative functions. An excellent example of this is the conversion of the DNR/DXR- producing S. peucetius ATCC 29050 strain into an 4-epi-DNR/4-epi-DXR-producing strain.49 Hutch- inson and colleagues recognized that the stereochemistry of the 4-OH of TDP-L-daunosamine is controlled by DnmV, and DnmV homologs from the avermectin (AveE for TDP-oleandrose biosyn- thesis) or erythromycin (EryBIV for TDP-mycarose biosynthesis) pathways catalyze the analogous chemistry but with opposite stereochemistry. They reasoned that substitution of dnmV with avrE or eryBIV might afford the biosynthesis of TDP-4-epi-daunosamine, which could eventually lead to the production of 4-epi-DNR or 4-epi-DXR if the DnmS glycosyltransferase and other downstream enzymes can recognize and process the 4-epi-diastereomers of the biosynthetic intermediates. Remark- ably, complementation of the ∆dnmV mutant with either avrE or eryBIV indeed resulted in the pro- duction of 4-epi-DNR and 4-epi-DXR, providing a fermentation alternative to these important anti- cancer drugs that are currently manufactured by low-yielding semisynthesis methods (Figure 25.1B).49 The DNR/DXR type II PKS has also been extensively exploited by combinatorial biosynthesis methods with other type II PKS systems such as actinorhodin,36,50 tetracenomycin (TCM),37 and jadomycin,37 leading to the production of numerous novel polyketide metabolites (Figure 25.1C). Although these studies do not focus on generating new DNR/DXR analogs, they do highlight the feasibility and versatility of producing novel aromatic polyketide molecular scaffolds based on the type II PKS machinery. B. MITHRAMYCIN (TYPE II PKS) Mithramycin (MTM), isolated from Streptomyces argillaceus and several other Streptomyces spe- cies,51 is a member of the aureolic acid group of anticancer drugs that includes the chromomycins, Copyright © 2005 CRC Press, LLC olivomycins, chromocyclomycins, durhamycins, and UCH9.52 Structurally, members of the aureolic acid group of anticancer drugs (with the exception of chromocyclomycin) share a similar aglycone with two alkyl substitutions at C3 and C7 but differ in their glycosylation pattern at C6-O and C2- O. MTM has proven clinically important for the treatment of various cancers.53 The anticancer activity of MTM is presumed to be a result of a Mg2+-MTM complex binding to GC-rich DNA and disrupting DNA-dependent RNA synthesis,52 in addition to inducing the surrounding AT-rich sequences to be sensitive to DNase II attack.54,55 1. Biosynthetic Gene Cluster and Biosynthetic Mechanisms Although the polyketide origin of the aureolic acid group of anticancer natural products was established by Rosazzas and coworkers by incorporating isotope-labeled precursors into chromo- mycins A3,56 it was the reinterpretation of these data and subsequent cloning and sequencing of the MTM and chromomycins biosynthetic gene clusters by Salas, Rohr, and coworkers that unam- biguously established the biosynthetic pathway for this family of metabolites, involving a single linear dekaketide as a key intermediate.52,57–61 Key features of the MTM clusters include genes encoding the MTM type II PKS and associated enzymes that catalyze the formation of the tetracyclic decaketide intermediate premithramycinone from 10 units of malonyl-CoA, genes encoding deox- ysugar biosynthesis enzymes that catalyze the biosynthesis of the five deoxy sugars and their attachment at C2-O and C6-O of the MTM aglycone, and genes encoding various tailoring enzymes that catalyze modifications of the MTM aglycone, including the unique ring fission oxidation, to finally afford the characteristic tricyclic aglycone for this group of anticancer natural products (Figure 25.2A).52 In contrast to the DNR/DXR type II PKS that requires the DpsCD to select propionyl-CoA as a special starter unit, the MTM type II PKS does not contain the DpsCD homologs; it instead derives the acetate starter unit from malonyl-CoA, a common feature believed to be true for all type II PKSs that catalyze the biosynthesis of polyketide natural products with acetate as a starter unit.6 One of the most fascinating aspects for MTM biosynthesis is the interplay between the glyc- osyltransferases and the tailoring enzymes. MTM contains a trisaccharide (D-olivose-D-oliose-D- mycarose) at C2-O (C-12a-O in premithramycinone) and a disaccharide (D-olivose-D-olivose) at C6-O (C8-O of premithramycinone), but only four glycosyltransferases (MtmGI/II/III/IV) have been identified within the MTM cluster. The timing of the glycosylation steps was deciphered by a series of gene disruption experiments followed by structural determination of the accumulated intermediate(s).52,62,63 Thus, D-olivose is attached to premithramycinone at C12a-O by MtmGIV, which is followed by D-oliose attachment to the D-olivose by MtmGIII. The addition of D-mycarose to complete the trisaccharide moiety is still an open question, with either MtmGIII catalyzing the addition of both D-oliose and D-mycarose to D-olivose or MtmGIV catalyzing D-mycarose addi- tion. The disaccharide (D-olivose-D-olivose) attached to premithramycinone at C8-O is synthesized first by the action of MtmGI or MtmGII and subsequently attached by either of these enzymes. Once the five sugar moieties are attached, forming premithramycin B, ring D is oxidatively cleaved by MtmOIV to finally afford MTM (Figure 25.2A). 2. Combinatorial Biosynthesis Salas, Rohr, and coworkers have extensively characterized the MTM biosynthetic pathway by targeted gene disruption.58,60,62–64 The resultant mutants accumulated a number of novel MTM analogs including 4-demethylpremithramycinone, premithramycinone, premithramycin A1, premithramycin A2, 9-demethylpremithramycin A3, premithramycin A3, 7-demethyl-mithramycin, premithramycin B, 4A-keto-D-olivosyl-premithramycin A2, 4A-keto-D-olivosyl-9-demethyl- premithramycin A2, and 4A-keto-D-olivosyl-demycarosyl-mithramicin (Figure 25.2A and 25.2B). By mixing and matching genes from the MTM pathway with genes from various other biosyn- thetic pathways in either the MTM-producing Streptomyces agrillaceus or other polyketide-producing Copyright © 2005 CRC Press, LLC OR OCH H H 3 O O O S-MtmS (A) MtmPKS HO 8 OH HO OH MtmGIV HO S-CoA O O O O 12a CH CH 3 MtmMII CH3 (10 x) 3 R OH O O O O O OH OH O O O OH OH O O H C O HO 3 O O premithramycinone (R=CH3) HO 4-demethylpremithramycinone (R=H) premithramycin A1 (R=H) OCH3 OCH premithramycin A1' (R=CH H H 3 3) MtmGIII HO OH MtmGIII, GIV? HO OH CH CH H R 3 H3C HO O 3C 3 OH OH OH O O O CH3 H O 3C HO HO O O H C OH OH OH O HO O 3 O O CH H 3 3C O O OTDP HO O O HO O H3C O MtmGI or GII OH premithramycin A2 premithramycin A3 (R=CH3) H3C O H3C 9-demethylpremithramycin A3 (R=H) HO HO O HO O MtmGI or GII OTDP H3C H3C O H3C OCH3OH HO O H3C OCH3 HO O H HO HO O H HO O O 6 CH3 HO O O OH 3 D 2 CH3 R O OH MtmOIV 7 O H3C H O 3C H OH OH O others O H H C OH OH OH O O O CH O 3C 3 HO 3 O O H O CH3O H3C HO HO O H O O 3C O H C O OH 3 O OH mithramycin (R=CH3) premithramycin B 7-demethylmithramycin (R=H) (B) OCH H H 3 3C H HO O 3C OCH3 OH HO OH HO O H HO O O CH3 CH R 3 O OH H3C O OH OH OH O OH OH OH O O O CH H3C O O O 3 O CH3 O H3C O O HO O HO O 4A-keto-D-olivosyl-9-demycarosyl-mithramycin 4A-keto-D-olivosyl-premithramycin A2 (R=CH3) 4A-keto-D-olivosyl-9-demethyl-premithramycin A2 (R=H) O CH O O O COOCH3 (C) 3 OH O O OH HO OCH3 CH3 HO O H3CO OH O RO OCH CO2H 3 OH O OH OH O O OH O OH OH OH OH OH HO CH3 TCM C (R=CH3) H3C HO O 8-demethyl-8-β-D-divosyl-TCM C ( R= C H3C ) H3 OH premithramycinone H Auramycinone TCM M H3C 8-demethyl-8-β-D-mycarosyl-TCM C ( R= H O O ) HO H3C H3C 8-demethyl-8-β-D-olivo-3'-1''-β-D-divosyl-TCM C ( R= H O O HO O ) HO O FIGURE 25.2 (A) Structures and biosynthesis of mithramycin. (B) Structures of novel mithramycin analogs produced by engineering the mitramycin pathway. (C) Structures of novel metabolites produced by mixing and matching genes from the mithramycin, tetracenomycin, and elloramycin pathways. Streptomyces strains, Salas, Rohr, and coworkers have also systematically explored the potential of using the MTM biosynthetic machinery for the production of novel metabolites by combinatorial biosynthesis methods (Figure 25.2C). Introduction of the tcmGH genes from the TCM pathway, both of which encode oxygenases, into an S. agrillaceus ∆mtmD mutant resulted in the production of premithramycinone H.65 Expression of the minimal MTM type II PKS (mtmPKS) along with a cyclase (mtmX) in an Streptomyces galilaeus mutant strain produced auramycinone, an intermediate in aclacinomycin biosynthesis,66 while introduction of the same set of genes into Streptomyces glauc- escens yielded TCM M.67 Production of the anthracycline antibiotic elloramycin in S. agrillaceus resulted in the glycosylation of the elloramycin aglycon with sugars from the MTM biosynthetic Copyright © 2005 CRC Press, LLC pathway, leading to the production of hybrid metabolites such as 8-β-D-olivosyl-tetracenomycin C, 8-demethyl-8-β-D-mycarosyl-tetracenomycin C, and 8-demethyl-8-β-D-olivo-3-1-β-D-olivosyl-tet- racenomycin C.68 These studies highlight how enzymatic components from pathways producing different anticancer natural products can be mixed and matched to further extend the biosynthetic potential of the characterized biosynthetic machineries. C. ANSAMYCINS (TYPE I PKS) The ansamycins all include an aliphatic ansa bridge, giving the antibiotic family their name, and are divided into two structural groups based on whether their macrocycle contains a benzenic or naphthalenic chromophore. The benzoquinone ansamycins include the anticancer drugs geldana- mycin, herbimycin A, maytansine, and ansamitocin P-3, whereas the napthoquinone ansamycins include rifamycin B, the most well-known and best characterized member of this antibiotic family69 (Figure 25.3A). All of the ansamycins contain a 3-amino-5-hydroxybenzoic acid (AHBA) moiety that is oxidized to the benzoquinone or converted to the naphthoquinone via cyclization and oxidation with other moieties within the molecule. On the basis of results from feeding experiments with isotope-labeled precursors, the biosynthesis of both the benzoquinone and naphthoquinone ansamycins can be predicted to involve a type I PKS specifying AHBA as the starter unit.69–72 Although none of the benzoquinone ansamycins are currently in clinical use, the promising anticancer activities they possess have spurred interest in the development of analogs of these compounds that lack the dose-limiting toxicity in humans.73 The biological target of these molecules is Hsp-90, an important protein chaperone in mammalian cells.74 Geldanamycin has been shown to bind to the ATP-binding site on Hsp-90, thereby inhibiting the ATP-dependent activity of this chaperone.75,76 In the absence of Hsp-90 function, associated protein kinases that are essential (A) CH3 CH3 CH3 Pr-i OCONH2 CH H3C 3CO2 Cl CH O 3 O O H H3CO OH OH H3CO N O CH H3CO CH3OCH 3 3 CH OR 3 HO NH H 1 O 2 CH CH H3C O OH OH 3 3 O O CH3 H3C NH NH N O R2 CO2H CH O 3 H O HH O C 3 R3 O OCH2COOH AHBA O H O 3C ansamitocin P-3 geldanamycin (R1=R2=H, R3=OCH3) rifamycin B herbimycin A (R1=CH3, R2=OCH3, R3=H) (B) Pr-i O R O 1 CH3 OH R1 R O 3 OR4 R1 R O 3 O O OH H3C CH3 R N CH 2O 3 R N CH 2O 3 R2O N O CH H2N CH3 3 H3C CH OH 3 CH3 CH3 CH OH O 3 O OH H3C OH COOH O OH O N N H O H O HO H3C O OH CH3 CH3O CH3 CH OOH CH3 CH H 3OO CH3 3 R proansamitocin N-demethyl-desepooxymytansinol 19-deschloro-N-demethyl-AP-3 P8/1-OG 6-deoxyerythronilide B (R1=R2=H) (R (6DEB, R=CH3)19-chloroproansamitocin 1=Cl, R2=CH3, R3=R4=H) (R1=R3=H, R2=CH3) (R N-demethyl-desepoxy-AP-3 20-O,N-didemethyl-AP-3 1=Cl, R 2-desmethyl-2-methoxy-6-DEB 2=H) (R (R=OCH 20-O-methyl-19-chloroproansamitocin 1=Cl, R2=CH3, R3=H, R4=isobutyryl) (R1=Cl, R2=R3=H) 3) (R1=Cl, R2=CH3) 3-O-isobutyryl-carbamoylproansamitocin N-demethyl-AP-3

(R1=R2=R3=H, R4=isonutyryl) (R1=Cl, R2=CH3, R3=H) FIGURE 25.3 (A) Structures of four representatives of the ansamycin family of natural products. The part that is derived from the methoxymalonyl-ACP extender unit is boxed in ansamitocin P-3, geldanamycin, and herbimycin A. (B) Structures of novel ansamitocin P-3 biosynthetic intermediates produced by engineering the ansamitocin P-3, erythromycin, and FK520 pathways. Copyright © 2005 CRC Press, LLC components of signal transduction pathways cannot function properly. The failure of these protein kinases to function can result in a cytostatic effect on cancer cells.73 Importantly, Hsp-90 and one or more of its protein kinase associates are overproduced in certain types of human cancers, indicating that Hsp-90 is a promising target for the development of anticancer drugs. 1. Biosynthetic Gene Clusters and Biosynthetic Mechanisms The cloning and sequencing of the genes encoding the biosynthesis of the AHBA moiety of the ansamycins by Floss and coworkers77 gave a unique handle to specifically clone the ansamycin biosynthetic gene clusters from their producing organisms. The first ansamycin biosynthetic gene cluster sequenced was the archetype of this family of molecules, rifamycin B, which was cloned from two different strains of Amycolatopsis mediterranei.69,78 Bioinformatic analysis of the cloned clusters revealed that the rifamycin backbone is assembled by the Rif PKS (RifA-E), a type I PKS that specifies for AHBA as a starter unit (by RifA) and that uses two malonyl-CoA and eight methylmalonyl-CoA as extender units (by RfiA-E). Intramolecular cyclization of the full-length polyketide intermediate (by RifF), followed by subsequent modifications of the resultant macro- lactam intermediate by the tailoring enzymes, finally affords the rifamycins. The insights gained from the rifamycin B biosynthetic gene cluster, particularly the genes encoding the AHBA biosynthetic enzymes (RifG-N), greatly facilitated the cloning and sequencing of the ansamitocin70 and geldanamycin79 biosynthetic gene clusters from Actinosynnema pretiosum and Streptomyces hygroscopicus NRRL 3602, respectively. The Asm and Gdm type I PKS consist of not only a loading module specific for AHBA (by AsnA and GdmA1), similar to the Rif PKS, but also extending modules specific for the unusual extender unit of methoxymalonyl-acyl carrier protein (ACP), a finding that is consistent with their structures (Figure 25.3A). A dedicated set of genes encoding the biosynthesis and activation of the rare methoxymalonyl-ACP extender unit was also identified within both clusters. 2. Combinatorial Biosynthesis Floss and coworkers have meticulously inactivated each of the candidate genes, encoding the tailoring steps for ansamitocin P-3 biosynthesis in A. pretiosum followed by isolation and structural elucidation of the accumulated metabolites.80 These studies not only allowed them to identify the functions for many of the gene products and to establish the order of the tailoring steps catalyzed by these enzymes but also demonstrated the feasibility of producing novel ansamitocin analogs by combinatorial biosynthesis in the native producing organisms (Figure 25.3B). Alternatively, pro- duction of an ansamycin polyketide intermediate has also been achieved by expressing a hybrid biosynthetic pathway in E. coli.81 Khosla and coworkers introduced the following components from various biosynthetic machin- eries into a single E. coli strain: the sfp phosphopantetheinyl transferase gene from Bacillus subtilis to ensure posttranslation modification of PKS or nonribosomal peptide synthetase (NRPS) mod- ules,82,83 the rifH/K/L/M/N genes from the rifamycin producing A. mediterranei and the asm23/47 genes from the ansamcyin P-3 producing A. pertiosum to produce the AHBA starter unit,81 the pccB and accA1 gene from Streptomyces coelicolor to provide the methymalonyl-CoA extender units,84 and the rifA gene from A. mediterranei to catalyze the initiation and the first three cycles of polyketide elongation for a tetraketide intermediate in ansamycin biosynthesis.81,85 To overcome the difficulty of producing the four-module RifA as an intact protein, specific polypeptide linkers from 6-deoxy- erythronolide B synthase (DEBS), the archetype type I PKS catalyzes the biosynthesis of erythro- mycins from Saccharopolyspora erythraea,86–88 were fused to a divided RifA synthase to facilitate interpolypeptide transfer of the growing biosynthetic intermediate.81 The resultant E. coli recombinant strain indeed produced the expected tetraketide intermediate, 2,6-dimethyl-3,5,7-trihydroxy-7-(3- amino-5-hydroxyphenyl) 2,4-heptadienoic acid (P8/1-OG) (Figure 25.3B). This work is one of the best examples of rational construction of a hybrid pathway from various biosynthetic machineries Copyright © 2005 CRC Press, LLC and represents an important first step toward the reconstitution of the ansamycin biosynthetic machin- ery in a fast-growing, genetically tractable host for engineered production of novel ansamycins. Complementary to ansamycin production in a heterologous host, parts of the ansamycin bio- synthetic machinery have also been combined with other polyketide biosynthetic machinery for the production of novel structures. For example, Floss and coworkers introduced the asm13-17 genes from A. pertiosum, which encode the biosynthesis of the unusual methoxymalonyl-ACP extender unit, into a recombinant S. lividans strain. The latter carried an engineered DEBS whose methylmalonyl-specific PKS module-6 has been replaced by the methoxymalonyl-specific PKS module of the FK520 pathway from S. hygroscopicus.79 Production of methoxymalonyl-ACP by the Asm13–17 enzymes, and incorporation of this unusual extender unit by the engineered DEBS, resulted in the production of the anticipated 2-desmethyl-2-methoxy-6-deoxyerythronolide B (Fig- ure 25.3B), serving as another example of producing novel structures by combining features from multiple natural product biosynthetic pathways.89 With the growing list of gene clusters sequenced for the ansamycin family of metabolites, the extensive genetic and biochemical groundwork con- ducted by Floss and coworkers on rifamycin biosynthesis,90 and the enormous success at repro- gramming type I PKS such as DEBS,15 it is not difficult to envisage engineering additional novel ansamycin analogs by combinatorial biosynthesis methods in the near future. D. EPOTHILONES (HYBRID PKS-NRPS) The epothilones (EPOs) were first isolated from culture extracts of the myxobacterium Sorangium cellulosum strain So ce90, exhibiting antifungal and cytotoxic activities, and were subsequently rediscovered in a target-driven screening for molecules that stabilize tubulin by a mechanism similar to the anticancer drug paclitaxel (Taxol).91,92 By stabilizing tubulin, the EPOs induce cell cycle arrest, leading to apoptosis, and most significantly the EPOs are active against paclitaxel-resistant cell lines, as well as cell lines resistant to other anticancer agents.92 EPO B showed the most promising in vitro activity; however, this activity was markedly reduced when analyzed in vivo. In addition to the two-major EPOs (A and B), numerous minor analogs have also been isolated and characterized, including most notably EPO C and D (Figure 25.4A). EPO D has better activity than EPO B and paclitaxel and did not result in the same acute toxic effects as EPO B.93 Several variants of the EPOs are currently in various stages of clinical trials.94 The EPOs contain a methylthiazolyl moiety attached to a 16-membered macrolactone. The main structural variations among the EPOs are the modification at C12/C13 (a double bond, as in EPO C and D vs. an epoxide as in EPO A and B) and the substitution at C12 (H for EPO A and C and CH3 for EPO B and D). Feeding experiments with isotopic-labeled precursors have estab- lished a hybrid peptide–polyketide origin for EPO biosynthesis with the EPO backbone specifically derived from acetate, propionate, methionine, and cysteine.95 These results are consistent with the hypothesis that EPO biosynthetic machinery is characterized by a hybrid nonribosomal peptide synthetase (NRPS)-PKS system (Figure 25.4A).96–98 1. Biosynthetic Gene Cluster and Biosynthetic Mechanisms Molnar and coworkers and Julien and coworkers independently cloned and sequenced the EPO biosynthetic gene cluster from S. cellulosum strain So ce9096 and S. cellulosum strain SMP44,97 respectively. The gene nomenclature proposed by Julien and coworkers is used here.97,98 Central to the EPO cluster are five PKS genes (epoA-E) encoding nine PKS modules, one NRPS gene (epoB) encoding one NRPS module, and epoK encoding a cytochrome P-450 monooxygenase. EpoA-F constitutes a hybrid PKS/NRPS/PKS megasynthase that activates, condenses, and modifies five malonyl, four methylmalonyl, and one cysteinyl unit to form the macrolactone core of EPO C. Module 4 of the Epo hybrid PKS/NRPS/PKS megasynthase apparently can accept both malonyl- CoA or methylmalonyl-CoA as extender units, resulting in the production of EPO D, in addition to EPO C. Further oxidation of EPO C and D by EpoK finally affords EPO A or B (Figure 25.4A). Copyright © 2005 CRC Press, LLC O (A) (1 x) H3C O SCoA R R O O2C (4 x) S CH3 S CH3 O SCoA 12 EpoA-F OH EpoK 13 H3C N H3C H C N H3C OH O2C SCoA (4 x) H C CH 3 H C CH O 3 3 O 3 3 CH O CH3 CH 3 3 (1 x) O OH O O OH O HS O NH O 3 epothilone C (R=H) epothilone A (R=H) H3CS epothilone D (R=CH epothilone B (R=CH 3) 3) O (1 x) NH3 (B) (C) O R X 2 EpoB/EpoC CH3 R1 = CH3, Et, i-Pr, Ph CH3 R1 S-ACP S-ACP R E 1 N poC EpoA R2=H, CH3 R R R2 O S CH 1 2 X = O, S 3 O H C OH HX O H3C N 3 H C CH 3 3 NH3 CH3 CH3 O OH O O CH3 EpoC R1 S-ACP S 10,11-didehydroepothilone D (R1=CH=CH, R2=H/H) R1 S-ACP EpoB EpoB R1 = or 9-keto-epothilone D (R1=CH2CH2, R2=O) O O H3C N N O2C SCoA CH 3 FIGURE 25.4 (A) Structures and biosynthesis of the epothilones. (B) Structures of novel epothilones pro- duced by engineering the epothilone pathway. (C) Structures of novel analogs of epothilone biosynthesis intermediates produced by purified and/or engineered epothilone type I PKS enzymes. 2. Combinatorial Biosynthesis Combinatorial biosynthesis of EPOs serves as an excellent example to illustrate overexpression and manipulation of an entire biosynthetic pathway in a more amenable heterologous host than its native producer. The impetus for this work was the finding that the natural EPO-producing myx- obacterial strains are slow growing and that methods and tools for the genetic manipulation of these strains have yet to be fully developed. Furthermore, EPO D, the derivative that shows the greatest therapeutic index among the EPOs,99,100 is only a minor component produced by the native strains. Functional expression of the EPO biosynthetic pathway into a faster-growing host that is genetically more tractable than the native producer therefore could be advantageous for both titer improvement of the natural products and rational engineering of the EPO pathway to selectively produce the desired EPO analog. Julien and coworkers first cloned the entire EPO cluster (epoA-K) into two compatible expres- sion vectors and introduced them into S. coelicolor CH999, an engineered Streptomyces host extensively used for combinatorial biosynthesis.97 The resulting recombinant strain produced mod- erate amounts (50–100 µg/L) of EPOs, far less than required for industrial scale-up; however, this system could be amenable to yield improvement via traditional and contemporary means. The researchers then moved the entire EPO cluster into the well-studied myxobacterium Myxococcus xanthus.101 Optimization of fermentation conditions for the resultant recombinant strain has sub- sequently led to EPO titer exceeding 20 mg/L, an acceptable yield for industrial scale production by fermentation.93,101–103 Although these studies successfully afforded the production of EPO A and B in S. coelicolor CH999 and M. xanthus, Julien and coworkers also constructed recombinant strains that lack a functional epoK.98,101 Because EpoK catalyzes the epoxidation of EPO C and D, the absence of a functional EpoK enabled the latter strains to produce EPO C and D exclusively. As a proof of principle that rational engineering of the hybrid Epo NRPS-PKS megasynthase could potentially introduce additional structural diversities into the EPO hybrid peptide–polyketide backbone, Carney, Katz, and coworkers inactivated several domains within the EpoD PKS subunits. They took advantage of the genetic tools available in M. xanthus and introduced the engineered Copyright © 2005 CRC Press, LLC hybrid Epo NRPS-PKS into M. xanthus to screen for the production of novel EPO analogs.93 Although inactivation of the enoylreductase domain of module 5 of EpoD yielded 10,11-didehy- droepothilone C and D, double inactivation of the ketoreductase domain of module 6 of EpoD and EpoK resulted in the production of 9-keto-epothilone D (Figure 25.4B).104 Complementary to these in vivo studies, Walsh and coworkers have performed a number of elegant in vitro studies to probe the enzymatic flexibility of EpoB and EpoC to accept alternative substrates for EPO biosynthesis105–108 (Figure 25.4C). They first showed that EpoB and EpoC are quite flexible, accepting various acyl-ACPEpoA as substrates, including propionyl-, isobutyryl-, and benzoyl-ACPEpoA, resulting in the production of a series of 2-alkylthiazolylmethylacrylyl-ACPEpoC derivatives. They then revealed that EpoB can activate and modify not only the natural L-cysteine substrate but also L-serine and L-threonine,

with each subsequently being incorporated and modified by the reconstituted ACPEpoA/EpoB/EpoC complex.107 They finally demonstrated that EpoA or EpoB can be replaced by proteins from the rapamycin, enterobactin, and yersiniabactin pathways, with efficient catalysis made possible by manipulating protein–protein recognition components at the PKS/NRPS and NRPS/PKS interfaces.106 These studies demonstrated the potential to engineer additional EPO structural diversity by applying combinatorial biosynthesis methods to the hybrid Epo NRPS-PKS megasynthase. E. LEINAMYCIN (HYBRID NRPS-PKS FEATURING AT-LESS PKS) Leinamycin (LNM), produced by several Streptomyces species,109,110 contains an unusual 1,3- dioxo-1,2-dithiolane moiety that is spiro-fused to a thiazole-containing 18-membered lactam ring (Figure 25.5), a molecular architecture that has not been seen in any other natural product. LNM shows potent antitumor activity in vitro and in various tumor models in vivo.111,112 Most signifi- cantly LNM is active against tumors that are resistant to clinically important anticancer drugs, such as cisplatin, DOX, MTM C, or cyclophosphamide.110 LNM causes alkylative DNA cleavage in the presence of thiol agents as cofactors, and this process is mediated by an episulfonium ion intermediate.113 However, development of LNM into a clinical anticancer drug has so far been hampered by its in vivo instability.111,112 1. Biosynthetic Gene Clusters and Biosynthetic Mechanisms Shen and coworkers developed a polymerase chain reaction method to clone genes encoding thiazole- forming NRPS and applied it to the cloning and localization of the LNM biosynthetic gene cluster from Streptomyces atroolivaceus.114 Sequencing and subsequent characterization of the LNM cluster revealed that the LNM macrolactam backbone is derived from one molecule each of D-alanine and L-cysteine and six molecules of malonyl-CoA, with the two CH3 groups at C6 and C4 derived from methionine. Strikingly, the biosynthesis of the LNM backbone from the carboxylic acid and amino acid precursors is catalyzed by a hybrid NRPS (LnmQP) and AT-less type I PKS (LnmGIJ) megasyn- thase, representing a novel architecture for both NRPS and PKS.115,116 The biosynthetic origin of the 1,3-dioxo-1,2-dithiolane moiety, however, remains to be established. O O 6 H CH 3C OH 3 OH H O 3C 4' O CH3 CH CH 3 S S 3 SH OH N N N O O N H H O S S leinamycin TG-25 FIGURE 25.5 Structures of leinamycin and TG-25. TG-25 was produced by combinatorial biosynthesis of the leinamycin pathway. Copyright © 2005 CRC Press, LLC 2. Combinatorial Biosynthesis To produce LNM analogs by combinatorial biosynthesis methods, Shen and coworkers first devel- oped an expedient genetic system for in vivo manipulation of LNM biosynthesis in S. atrooliva- ceus.114 Systematic inactivation of genes from the sequenced region not only defined the precise boundaries of the LNM cluster but also generated a series of mutants that accumulated numerous novel metabolites.116 For example, inactivation of lnmE abolished LNM production, but the ∆lnmE mutant strain produced a new compound TG-25 (Figure 25.5) (Yun, B.-S., Tang, G.-L., Cheng, Y.- Q., and Shen, B., unpublished data). TG-25 represents a novel structural variant of the LNM scaffold, featuring a free –SH group. Remarkably, although episulfonium ion formation from LNM requires a reductive environment, needing reduced thiols, preliminary studies indicated that TG- 25 can generate a similar episulfonium ion under an oxidative environment, requiring O2. TG-25 could therefore potentially be used as anticancer drug via oxidative activation in situ in tumors under high reactive oxygen species while remaining relatively nontoxic to normal tissues. F. ENEDIYNES (ITERATIVE TYPE I PKS) Since the unveiling of the neocarzinostatin (NCS) chromophore structure in 1985,117 the enediyne family has grown steadily.118 Over 20 enediyne natural products are currently known, and they are classified into two subcategories according to the enediyne core structures. Members of the 9-membered enediyne core subcategory are chromoproteins consisting of an apo-protein and the enediyne chromophore, including C-1027 from Streptomyces globisporus, NCS from Streptomyces carzinostaticus, and maduropeptin (MAD) from Actinomadura madurea. The 10-membered ene- diyne core subcategory includes calicheamicin (CAL) from Micromonospora echinospora spp. calichenisis, dynemicin (DYN) from Micromonospora chersina sp. nov. No. M965-1, and espe- ramicin (ESP) from Actinomadura verrucosospora (Figure 25.6). Although the 10-membered enediynes are all isolated as discrete small molecules, the gene product of calC from the CAL biosynthetic gene cluster confers CAL resistance in vivo.119 Although no binding between CalC and CAL has been experimentally demonstrated, CalC apparently confers resistance by a self- sacrificing mechanism.120 This finding raised the interesting question of whether self-sacrificing is a general mechanism for all 10-membered enediynes and whether there is an evolutionary relationship (mechanistic or structural) between CalC and the apo-proteins of the nine-membered O H C O R N 2 3 O O O N O OCH O O Cl H 3 O O O O OH H C HO 3 O O H3C O OH (CH3) O 2N O R OH 1 H H 3CO OH O C 3 O OH O OH CH3HN O NHCH3 maduropetin HO R O 2 NH HO CH CH 3 HO 3 2 C-1027 (R1=OH, R2=Cl, R3=OCH3) neocarzinostatin deshydroxy-C-1027 (R1=H, R2=Cl, R3=OCH H3C 3) SSSCH3 O deschloro-C-1027 (R1=OH, R2=H, R3=OCH3) NHCO2CH3 SSSCH CH OH 3 desmethyl-C-1027 (R1=OH, R2=Cl, R3=OH) HO O O 3 O CH O O 3 NHCO desmethoxy-C-1027 (R1=OH, R2=Cl, R3=H) O N O OH H SCH 2CH3 3 OH HO O O CH3 O CH O O N O S3 O H3C OCH O 3 OH NHPr-i H CH O OCH H3C OCH H 3 HO 3 O NH3Et CO OH O HN 2H H3CO O esperamicin A1 I OCH O 3 OCH3 H H 3CO NH 3Ccalicheamicin i I O 1 HO O OCH3 CH3O OH OH O OH dynemicin FIGURE 25.6 Structures of representatives of the 9-membered (neocarzinostatin, C-1027, maduropetin) and 10-membered (calicheamicin, dynemicin, esperamicin) enediynes and novel C-1027 analogs produced by engineering the C-1027 pathway. Copyright © 2005 CRC Press, LLC enediyne chromoproteins, even through no sequence homology exists between CalC and the nine- membered enediyne apo-proteins. The enediynes are extremely potent anticancer agents with a unique molecular architecture. As a family, they are the most potent, highly active anticancer agents ever discovered. All members of this family contain a unit consisting of two acetylenic groups conjugated to a double bond or incipient double bond within a 9- or 10-membered ring (i.e., the enediyne core). As a consequence of this structural feature, these compounds share a common mechanism of action: The enediyne core undergoes an electronic rearrangement (Bergman or Myers rearrangement) to form a transient benzenoid diradical, which is positioned in the minor groove of DNA so as to damage DNA by abstracting hydrogen atoms from deoxyriboses on both strands. Reaction of the resulting deoxyri- bose carbon-centered radicals with molecular oxygen initiates a process that results in both single- stranded and double-stranded DNA cleavage.118,121,122 Although the natural enediynes have seen limited use as clinical drugs mainly because of substantial toxicity, various polymer-based delivery systems or monoclonal antibody (mAB)-enediyne conjugates have shown great clinical success or promise in anticancer chemotherapy.119,123–125 For example, the poly(styrene-co-maleic acid)-conjugated NCS (SMANCS®) was approved in Japan in 1993 and has been marketed since 1994 for use against hepatoma.122 A CD33 mAB-CAL conjugate (Mylotarg®) was approved by the FDA in 2000 for the treatment of acute myeloid leukemia.124,126 Several anti- hepatoma mAB-C-1027 conjugates have also been prepared and displayed high tumor specificity with a strong inhibitory effect on the growth of established tumor xenografts.125 These examples clearly demonstrate that the enediynes can be developed into powerful drugs when their extremely potent cytotoxicity is harnessed and specifically delivered into the target tumor cells. 1. Biosynthetic Gene Clusters and Biosynthetic Mechanisms Feeding experiments with isotope-labeled precursors for NCS, DYN, and ESP biosynthesis clearly established that the enediyne cores of both the 9- and 10-membered subcategories are derived from the head-to-tail condensation of (at minimum) eight acetate units.9,127–129 This indicated a PKS mechanism of synthesis, but the exact mechanism of polyketide biosynthesis remained controversial until the Shen and Thorson groups simultaneously published the cloning, sequencing, and charac- terization of the CAL130 and C-1027131 biosynthetic gene clusters, respectively. Surprisingly, both gene clusters encode an iterative type I PKS, enediyne PKS, which is predicted to synthesize a linear polyketide intermediate. This linear intermediate is presumably modified by a series of enediyne PKS-associated enzymes to generate the enediyne core scaffolds. Both gene clusters also encode the necessary enzymes for synthesis and attachment of the appropriate deoxy aminosugar and other peripheral moieties. These studies laid the foundation for combinatorial biosynthesis of this important family of anticancer natural products. 2. Combinatorial Biosynthesis To set the stage to apply combinatorial biosynthesis methods to the enediyne biosynthetic machinery for generating novel anticancer drugs, the Shen and Thorson groups first developed genetic systems for S. globisporus (C-1027), S. carzinostaticus (NCS), and M. calichensis (CAL).128,130–132 With these tools in hand, Shen and coworkers have systematically inactivated many of the genes identified within the C-1027 and NCS genes gene clusters, unveiling many mechanistic details into enediyne biosynthesis.128,131–136 On the basis of these studies, they have demonstrated both yield improvement for C-1027 and production of novel C-1027 analogs by combinatorial biosynthesis methods. For example, overexpression of sgcA1 and its flanking genes in the wild-type S. globisporus has resulted in two- to fourfold increase in C-1027 production,135 and inactivation of sgcC1, sgcC3, sgcD3, and sgcD4 has afforded S. globisporus recombinant strains that produced four novel C-1027 analogs131 (Liu, W., Ju, J., Shen, B., unpublished data). Copyright © 2005 CRC Press, LLC Genetic and biochemical characterization of the C-1027 biosynthetic gene cluster has previously predicted that SgcC1 and SgcC3 catalyze the hydroxylation and chlorination steps in β-amino acid moiety biosynthesis, while SgcD3 and SgcD4 are responsible for the hydroxylation and methylation steps for the benzoxazolinate moiety.131 Mutants of corresponding genes indeed accumulated the predicted new metabolites: deshydroxy-C-1027 (∆sgcC3), deschloro-C-1027 (∆sgcC3), desmethyl- C-1027 (∆sgcD4), and desmethoxy-C-1027 (∆sgcD3) (Figure 25.6). These results clearly demon- strated the feasibility and effectiveness to produce novel enediyne analogs by combinatorial bio- synthesis methods. Strikingly, deshydroxy-C-1027 is at least fivefold more stable than C-1027 in respect to undergoing the Bergman cyclization, a property that could be potentially explored in C- 1027 engineering.131 The latter result indicates that chemical and physical properties of C-1027 could be fine-tuned by varying the peripheral moieties of the enediyne core, supporting the wisdom of diversifying the C-1027 molecular scaffold for drug discovery. Expedient methods to access and clone enediyne gene clusters have also been developed, and the growing list of enediyne clusters available will surely accelerate the pace of applying combinatorial biosynthesis methods to the enediyne biosynthetic machinery for novel anticancer drug discovery. Farnet and coworkers have developed a high-throughput genome scanning method to detect and analyze gene clusters involved in natural product biosynthesis. This method was applied successfully to uncover numerous biosynthetic pathways encoding enediyne biosynthesis in variety of actino- mycetes, including many strains that were not known previously as enediyne producers.132 Inspired by the biosynthetic potential for enediyne natural products, as evidenced by the DNA sequence, Farnet and coworkers demonstrated that enediyne production in these organisms can be induced on optimi- zation of their fermentation conditions. This genomics-guided approach for discovering and expressing cryptic metabolic pathways provides an outstanding opportunity to fully exploit the biosynthetic potential of microorganisms that we now know is much greater than previously appreciated.137,138 Complementary to the genome scanning method, the Shen and Thorson groups developed a polymerase chain reaction method to quickly access additional enediyne gene clusters on the basis of the emerging paradigm for enediyne biosynthesis featuring the enediyne PKS. This method was applied to clone the MAD, DYN, and ESP enediyne PKS loci from their respective producing organisms.128 A phylogenetic analysis of these bona fide enediyne PKS genes has revealed a clear genotypic distinction between 9- and 10-membered enediyne core subcategories (i.e., C-1027, NCS, MAD vs. CAL, DYN, ESP). This finding is consistent with the postulation that the minimal enediyne PKS helps define the structural divergence of the enediyne cores and could potentially be used to predict unknown enediyne structures on the basis of their enediyne PKS sequences or to guide rational construction of hybrid enediyne biosynthetic pathways for novel enediyne production.128 G. AMINOCOUMARINS (MIXED BIOSYNTHESIS FEATURING ATYPICAL NRPS) The aminocoumarin antibiotics, novobiocin from Streptomyces spheroides, coumermycin A1 from Streptomyces risluriensis, clorobiocin from Streptomyces roseochromogenes, and simocyclinone D8 from Streptomyces antibioticus, were initially identified for their antibiotic activity against Gram- positive bacteria. The interest in their role in the treatment of

cancer is a result of their ability to potentiate the effects of several anticancer drugs such as etoposide and teniposide.139,140 In addition, novobiocin was recently found to reverse drug resistance that was caused by a specific multidrug efflux mechanism in breast cancer cells, indicating a model for anticancer drug augmentation.141 The aminocoumarin antibiotics share a 3-amino-4,7-dihydroxycoumarin moiety that is highly decorated. Novobiocin and clorobiocin share a prenylated 4-hydroxybenzoate moiety but differ in two respects. The CH3 at C-8 and the carbamoyl moieties at C-3-O in novobiocin are replaced by Cl and a 5-methyl-2-pyrrolecarbonyl moieties in clorobiocin, respectively. Coumermycin A1 contains two coumarin-noviose moieties joined by a 3-methyl-2,4-dicarboxylpyrrole, with noviose deoxysugars containing 5-methyl-2-pyrrolecarbonyl substitutions at C-3-O. Finally, simocycli- none D8 is a hybrid natural product with an angucyclinone polyketide ring system connected to Copyright © 2005 CRC Press, LLC (A) OH OH CH CH HO 3 H HO O 3 HO N HO CH OH H3CO 3 HO CH OTDP 3 O 8 O CH OPO3 HO OTDP O O O 3 OHOH 3" novobiocin HO OH CH3OH R1 CH3 CO2H NH R 2 2O HO2C novobiocin (R1=CH3, R2=carbamoyl) HO HO NH2 HO O O CH clorobiocin (R1=Cl, R2=5-methylpyrrole-2-carbonyl) H 3 N R R=Cl or CH O H3CO 3 CH CH CH 3 R 3 3 O O O O N CO H R N CO R=H or CH3 2 2H R R = CH3OH CH N 3 H H H H3C N O HO2C CH O H 3 O O CH coumermycin A1 O 3 ? O HO N CO2H O CH O 3 H3CO CH HO O 3 OH H HO O O OH O O O OH O OH HN O SCoA O 4 4 O OH OH OH O OH OH OH HO O O simocyclinone D8 Cl (B) O HO OH H HO CH N H 3 R R2O N R = CH3 R O N O O O O CH3 H CH HO O O O 3OH R1 OR3 CH3 novobiocic acid CH3 coumermycin D (R1=R2=CH3, R3=5-methylpyrrole-2-carbonyl) O bis-carbamoylated coumermycin D (R1=R2=CH3, R3=carbamoyl) coumermycin LW1 (R1=H, R2=CH3, R3=5-methylpyrrole-2-carbonyl) CO2H coumermycin LW2 (R1=CH3, R2=H, R3=5-methylpyrrole-2-carbonyl) HO OH OH HO 2,4-dihydroxy-α-oxy-phenylacetic acid H N R4O O CH3 CH O O 3 O O CH3 HO O O O novobiocin H3C CH3OR3 R1 CH O CH O 3 3 O O OH R2O O R novclobiocin 101 (R1=R3=H, R2=5-methylpyrrole-2-carbonyl, R4=CH3) 4 O OH OH OH novclobiocin 102 (R1=CH3, R2=5-methylpyrrole-2-carbonyl, R3=H, R4=CH3) novclobiocin 104 (R1=Cl, R2=R3=H, R4=CH3) HO H novclobiocin 114 (R1=Cl, R2=carbamoyl, R3=H, R4=CH3) simocycline C5 (R=H) N novclobiocin 105 (R1=Cl, R2=R3=R4=H) novclobiocin 115 (R1=Cl, R2=carbamoyl, R3=R4=H) simocycline D-met R = novclobiocin 109 (R1=Cl, R2=pyrrole-2-carbonyl, R3=H, R4=CH3) HO O O novclobiocin 110 (R1=Cl, R2=H, R3=pyrrole-2-carbonyl, R4=CH3) CH3 FIGURE 25.7 (A) Structures and biosynthesis of the aminocoumarins novobiocin, clorobiocin, coumermycin A1, and simocyclinone D8. (B) Structures of novel aminocoumarins and/or biosynthetic intermediates pro- duced by combinatorial biosynthesis of the four aminocoumarin pathways. the 3-amino-4,7-dihydroxycoumarin moiety by an octatetraene dicarboxylic acid and a deoxy sugar olivose (Figure 25.7A). 1. Biosynthetic Gene Clusters and Biosynthetic Mechanisms Heide and coworkers have cloned and sequenced the biosynthetic gene clusters for novobiocin,142 clorobiocin,143 coumermycin A1,144 and simocylcinone D8,145 setting an excellent stage to study aminocoumarin biosynthesis by a comparative genomics approach and to develop new aminocou- marins by combinatorial biosynthesis methods. The enzymes encoded by these clusters can be divided into three groups that are involved in the biosynthesis of the core moieties of the antibiotics, including TDP-noviose and TDP-olivose from L-glucose-1-phosphate, aminocoumarin from L- tyrosine, 5-methyl-2-pyrrolecarboxylic acid from L-proline, prenylated 4-hydroxybenzoate from Copyright © 2005 CRC Press, LLC L-tyrosine, angucyclinone and the octatetraene dicarboxylic acid from acetates; the assembly of the various moieties into the characteristic molecular scaffolds of the aminocoumarin family of antibiotics; and tailoring steps such as carbamoylation, glycosylation, chlorination, and methylation (Figure 25.7A). Although the biosynthetic machineries for the deoxy sugar and polyketide moieties of the aminocoumarin family of antibiotics are similar to those known for other natural products in Strep- tomyces, those for the aminocoumarin, 5-methyl-2-pyrrolecarboxylate, hydroxybenzoate moieties from the corresponding amino acid precursors are certainly unique, and are characterized by a family of atypical NRPS enzymes. In contrast to NRPSs known for nonribosomal peptide biosynthesis, these atypical NRPS enzymes function to tether the proteinogenic amino acid precursors to peptidyl carrier protein as a series of modifications occur to generate the aminocoumarin, 5-methyl-2-pyrrole- carboxylate, or hydroxybenzoate moiety.142–145 Tethering a proteinogenic amino acid as an aminoacyl- S–peptidyl carrier protein for modification by tailoring enzymes has been speculated as a general strategy to sequester, and thereby divert, a fraction of the proteinogenic amino acid to provide building blocks for secondary metabolite biosynthesis.146–149 Thus, it is not apparent how these atypical NRPS could be mixed and matched with other NRPS enzymes in combinatorial biosynthesis. Rather, the tailoring steps have turned out to be the most rewarding targets of the aminocoumarin biosynthetic machinery for structural diversities by combinatorial biosynthetic methods. 2. Combinatorial Biosynthesis Given their mixed biosynthetic origin, targeted gene inactivation has been instrumental in deci- phering the molecular logic of how the aminocoumarin family of antibiotics is assembled. These studies established the roles many of the enzymes play in their respective biosynthetic pathways, resulted in the isolation of a series of mutants and the accumulated biosynthetic intermediates, and set the stage to engineer novel metabolites by mixing and matching biosynthetic components among the four aminocoumarin gene clusters. Mutants generated and biosynthetic intermediates isolated could also be exploited by combinatorial biosynthetic methods such as directed biosynthesis and chemoenzymatic synthesis to further expand the structural diversity of the aminocoumarins. Leading the effort of combinatorial biosynthesis of the aminocoumarins, Heide and coworkers have generated more than 50 new analogs (Figure 25.7B). Thus, inactivation of the couN3 and couN4 genes of the coumermycin A1 pathway yielded S. rishiriensis mutants that accumulated coumermycin D,150 and that of couO and couP resulted in mutants that produced coumermycin LW1 and coumermycin LW2, respectively.151 Introduction of novN, encoding the carbamoyltrans- ferase from the novobiocin biosynthetic pathway, into the ∆couN3 mutant strain generated the novel aminocoumarin, bis-carbamoylated coumermycin D.150 Investigations into the function of genes involved in chlorobiocin biosynthesis by S. roseochromogenes var oscitans DS12.986 resulted in mutants that produced novclobiocins 104 and 105,152 novclobiocin 109,153 and novclobiocin 101.154 Complementation of a ∆clo-hal mutant with the novO methyl transferase gene from the novobiocin biosynthetic gene cluster resulted in the formation of novclorobicin 102.154 Expression of novH-N in S. lividans TK24 resulted in the production of novobiocic acid — the aglycone of novobiocin, and expression of novD-K afforded 2,4-dihydroxy-α-oxy-phenylacetic acid, demonstrating the feasibility to manipulate aminocoumarin biosynthesis in heterologous hosts.155 As a final example of combinatorial biosynthesis of aminocoumarin antibiotics, Heide and coworkers combined various combinatorial biosynthetic methods, including directed biosynthesis and mutational and chemoenzymatic synthesis for novel aminocoumarin production (Figure 25.7B). Thus, inactivation of simJ1 of the simocyclinone pathway yielded an S. antibioticus mutant strain that accumulated simocyclinone C5. Fermentation of the ∆simJ1 mutant strain in the presence of the aminocoumarin intermediate for novobiocin biosynthesis resulted in the production of the novel simocyclinone derivative simocyclinone D-met.155 Alternatively, various 3-unsubstituted aminocou- marin derivatives were isolated from mutants of the clorobiocin producer S. roseochromogenes, Copyright © 2005 CRC Press, LLC and they were readily converted to the corresponding 3-carbamoylated products in vitro, using the purified recombinant NovN carbamoyltransferase as exemplified by novclobiocin 104 and 105 to novclobiocin 114 and 115, respectively.156 Similar efforts in combining biochemical, genetic, and synthetic approaches to take advantage of the substrate promiscuity of CloL, NovL, and CouL resulted in the production of an additional 32 new aminocoumarins.157 IV. SUMMARY, CHALLENGES, AND PROSPECTS Recent advances in cloning, and thus rapid access to biosynthetic gene clusters; fundamental understanding of secondary metabolic pathways; and emerging recombinant DNA technologies have fueled the exponential growth of the field of combinatorial biosynthesis in the last decade. Combinatorial biosynthesis has advanced from a limited set of proof-of-principle experiments into a more mature scientific discipline. With the ever-growing list of gene clusters for natural product biosynthesis becoming available, the continuous advancement of recombinant DNA technologies, and the accelerated pace in our understanding of the fundamental chemistry, biochemistry, and genetics governing natural product biosynthesis, combinatorial biosynthesis will surely play an increasing role in accessing natural products and generating natural product diversity for anticancer drug discovery.10,11 We have presented several examples of how combinatorial biosynthesis has been used to introduce structural diversity into anticancer natural products. We have by no means been complete in our summary. Rather, we have tried to highlight examples that show the enormous potential that combinatorial biosynthesis gives a researcher to develop or extend the metabolic capabilities of so- called “secondary metabolism.” The techniques used in combinatorial biosynthesis are not neces- sarily new. Rather, the most significant advancement in the field comes from the sequencing of natural product biosynthetic gene clusters at a breathtaking pace. With this information in hand, a researcher has a blueprint for how a natural product is biosynthesized and also has an understanding of how the blueprint can be altered to build new structural diversity into a natural product. Although there has been much success, there are still a number of hurdles that need to be overcome before the full potential of combinatorial biosynthesis can be reached. As stated previ- ously, the importance of developing genetic systems for the manipulation of the biosynthetic pathways cannot be stressed enough. Although it is possible that genetic techniques can be devel- oped for each organism that produces a natural product of interest, it would be more desirable to work toward the development of a series of genetically tractable universal heterologous hosts that can produce functional natural product biosynthetic pathways. This has been accomplished with some success for natural products biosynthesized by the actinomycetes. As shown by examples in this chapter, S. coelicolor and S. lividans have proven useful for the heterologous production of functional biosynthetic components of noncognate pathways from a variety of actinomycetes and also for the epothilone pathway from a myxobacterium. Development of metabolically engineered E. coli and M. xanthus strains for the production of complicated natural products84,101,158 indicates that these bacteria can also be developed into general hosts for natural product production. Although not discussed, development and demonstration of Bacillus subtilis as a heterologous host to express large biosynthetic gene clusters from related bacteria have also been reported.159,160 These strains, however, may not be the heterologous hosts of choice for natural product bio- synthetic pathways from more distantly related organisms. It is essential that heterologous hosts be developed to express gene clusters from cyanobacteria, pseudomonads, photorhabdus, and other bacterial genera known to be rich in metabolic potential. In addition, the development of strains for the heterologous production of natural products produced by filamentous fungi, plants, and animals will require the appropriate heterologous hosts to be developed. In addition to cultured organisms, harnessing the metabolic potential of uncultured microorganisms from sources such as soil (metage- nomic DNA161–165) and marine waters166–168 will also push the need for appropriate heterologous hosts. Many of these ecology niches are known to be rich sources of anticancer natural products. Copyright © 2005 CRC Press, LLC The development of these heterologous hosts requires the parallel development of the necessary vectors for the cloning, mobilization, and expression of the heterologous genes. The generation of bacterial artificial chromosomes makes it possible to clone entire biosynthetic gene clusters onto a single plasmid.169 New bacterial artificial chromosome derivates that replicate in E. coli and that can be moved by conjugation or transformation into Streptomyces170–172 or Bacillus173 pave the way for producing a natural product in more than one heterologous host using a single clone. The development of a universal bacterial artificial chromosome vector for mobilization into a series of engineered organisms would provide the ultimate vehicle for screening potential heterologous hosts. Although the enabling recombinant DNA technologies are necessary, a better understanding of the fundamental chemistry, biochemistry, and genetics governing the biosynthetic machinery is critical to direct the engineering of the enzymes to have increased or altered substrate specificity or better catalytic efficiency. Knowledge of the essential domains needed for the appropriate protein–protein interactions between biosynthetic enzymes will enable more efficient catalysis between hybrid constructs. Biochemical and mechanistic characterization of natural product bio- synthetic machinery requires the enzymes of

interest to be overproduced, purified in their functional forms, and studied in vitro, posing great challenges for which we so far have few solutions. Although the approaches outlined above focus on the rational engineering of natural products biosynthesis, mechanisms must also be investigated for protein and pathway evolution. Nature has evolved a stunning diversity of natural products through gene-shuffling processes over millions of years. With the recent advances in gene174,175 and genome176,177 shuffling, a future goal will be to develop genetic screens and selection that enable us to “evolve” new natural product biosynthetic pathways from the large library of natural product biosynthetic gene clusters that DNA sequencing has uncovered. Finally, combinatorial biosynthesis alone will not reach the full potential of natural product structural diversity. It will be the combination of this approach with other established and emerging technologies178–181 that will ultimately yield the maximal structural diversity of natural products for anticancer drug discovery.302–304 ACKNOWLEDGMENTS This work was supported in part by grant IRG-58-011-46-07 from the Amercian Cancer Society to MGT. Studies on natural product biosynthesis in Shen laboratory were supported in part by an IRG from American Cancer Society and the School of Medicine, University of California, Davis; the Searle Scholars Program/the Chicago Community Trust; the University of California BioSTAR Program and Kosan Biosciences, Hayward, CA (Bio99-10045); and National Institutes of Health grants CA78747, CA94426, CA106150. BS is the recipient of National Science Foun- dation CAREER Award MCB9733938 and National Institutes of Health Independent Scientist Award AI51689. REFERENCES 1. Newman, D. J., Cragg, G. M., and Snader, K. M., Natural products as sources of new drugs over the period 1981-2002, J. Nat. Prod. 66, 1022, 2003. 2. Cane, D. E., Walsh, C. T., and Khosla, C., Harnessing the biosynthetic code: combinations, permu- tations, and mutations, Science 282, 63, 1998. 3. Shen, B., Aromatic Polyketide Biosynthesis, in Topics in current chemistry: biosynthesis, Leeper, F. J. and Vederas, J. C. Springer, Berlin, 2000, pp. 1. 4. Staunton, J. and Weissman, K. J., Polyketide biosynthesis: a millennium review, Nat. Prod. Rep. 18, 380, 2001. 5. Staunton, J. and Wilkinson, B., Combinatorial biosynthesis of polyketides and nonribosomal peptides, Curr. Opin. Chem. Biol. 5, 159, 2001. Copyright © 2005 CRC Press, LLC 6. Du, L. and Shen, B., Biosynthesis of hybrid peptide-polyketide natural products, Curr. Opin. Drug Discov. Devel. 4, 215, 2001. 7. Walsh, C. T., Combinatorial biosynthesis of antibiotics: challenges and opportunities, ChemBioChem 3, 124, 2002. 8. Mootz, H., Schwarzer, D., and Marahiel, M., Ways of assembling complex natural products on modular nonribosomal peptide synthetases, ChemBioChem 3, 490, 2002. 9. Shen, B., Polyketide biosynthesis beyond the type I, II and III polyketide synthase paradigms, Curr. Opin. Chem. Biol. 7, 285, 2003. 10. Walsh, C. T., Polyketide and nonribosomal peptide antibiotics: modularity and versatility, Science 303, 1805, 2004. 11. Shen, B., Accessing natural products by combinatorial biosynthesis, Sci. STKE 2004, pe14, 2004. 12. Strohl, W. R., Biochemical engineering of natural product biosynthesis pathways, Metab. Eng. 3, 4, 2001. 13. Kennedy, J. and Hutchinson, C. R., Nurturing nature: engineering new antibiotics, Nat. Biotechnol. 17, 538, 1999. 14. Khosla, C., Combinatorial chemistry and biology: an opportunity of engineers, Curr. Opin. Biotechnol. 7, 219, 1996. 15. Khosla, C., Harnessing the biosynthetic potential of modular polyketide synthases, Chem. Rev. 97, 2577, 1997. 16. Kerwin, S., Toward bioengineering anticancer drugs, Chem. Biol. 9, 956, 2002. 17. Hutchinson, C. R. et al., Drug discovery and development through the genetic engineering of antibiotic- producing microorganisms, J. Med. Chem. 32, 929, 1989. 18. Hutchinson, C. R. and Fujii, I., Polyketide synthase gene manipulation: a structure-function approach in engineering novel antibiotics, Annu. Rev. Microbiol. 49, 201, 1995. 19. Malipartida, F. and Hopwood, D., Molecular cloning of the whole biosynthetic pathway of a Strep- tomyces antibiotic and its expression in a heterologous host, Nature 309, 462, 1984. 20. Hopwood, D. A. et al., Production of “hybrid” antibiotics by genetic engineering, Nature 314, 642, 1985. 21. Kutney, J. P., Plant cell culture combined with chemistry: a powerful route to complex natural products, Acc. Chem. Res. 26, 559, 1993. 22. Roberts, S. C. and Shuler, M. L., Large-scale plant cell culture, Curr. Opin. Biotechnol. 8, 154, 1997. 23. Bourgaud, F. et al., Production of plant secondary metabolites: a historical perspective, Plant Sci. 161, 839, 2001. 24. Kennedy, J. et al., Modulation of polyketide synthase activity by accessory proteins during lovastatin biosynthesis, Science 284, 1368, 1999. 25. Miao, V. et al., Genetic approaches to harvesting lichen products, Trends Biotechnol. 19, 349, 2001. 26. Nicholson, T. P. et al., Design and utility of oligonucleotide gene probes for fungal polyketide synthases, Chem Biol 8, 157, 2001. 27. Hutchinson, C. R., Antibiotics from genetically engineered microorganisms, in Biotechnology of antibiotics, 2 ed., Strohl, W. R. Marcel Dekker, New York, 1997, pp. 683. 28. Strohl, W. R. et al., Anthracyclines, in Biotechnology of antibiotics, 2 ed., Strohl, W. R. Marcel Dekker, New York, 1997, pp. 577. 29. Cullinane, C., van Rosmalen, A., and Phillips, D. R., Does adriamycin induce interstrand cross-links in DNA? Biochemistry 33, 4632, 1994. 30. Chen, A. Y. and Liu, L. F., DNA topoisomerases: essential enzymes and lethal targets, Annu. Rev. Pharmacol. Toxicol. 34, 191, 1994. 31. Drlica, K. and Franco, R. J., Inhibitors of DNA topoisomerases, Biochemistry 27, 2253, 1988. 32. Arcamone, F. et al., Adriamycin, 14-hydroxydaunorubicin, a new antitumor antibiotic from S. peuce- tius var. caesius, Biotechnol. Bioeng. 11, 1101, 1969. 33. Hutchinson, C. R., Anthracyclines, in Biochemistry and genetics of commercially important antibiotics Butterworths, Boston, 1995, pp. 331. 34. Otten, S. L., Stutzman-Engwall, K. J., and Hutchinson, C. R., Cloning and expression of daunorubicin biosynthesis genes from Streptomyces peucetius and S. peucetius subsp. caesius, J. Bacteriol. 172, 3427, 1990. 35. Hutchinson, C. R. and Colombo, A. L., Genetic engineering of doxorubicin production in Streptomyces peucetius: a review, J. Ind. Microbiol. Biotechnol. 23, 647, 1999. Copyright © 2005 CRC Press, LLC 36. Rawlings, B. J., Biosynthesis of polyketides (other than actinomycete macrolides), Nat. Prod. Rev. 16, 425, 1999. 37. Meurer, G. and Hutchinson, C. R., Daunorubicin type II polyketide synthase enzymes DpsA and DpsB determine neither the choice of starter unit nor the cyclization pattern of aromatic polyketides, J. Am. Chem. Soc. 117, 5899, 1995. 38. Grimm, A. et al., Characterization of the Streptomyces peucetius ATCC 29050 genes encoding doxorubicin polyketide synthase, Gene 151, 1, 1994. 39. Ye, J. et al., Isolation and sequence analysis of polyketide synthase genes from the daunomycin- producing Streptomyces sp. strain C5, J. Bacteriol. 176, 6270, 1994. 40. Rajgarhia, V. B. and Strohl, W. R., Minimal Streptomyces sp. strain C5 daunorubicin polyketide biosynthesis genes required for aklanonic acid biosynthesis, J. Bacteriol. 179, 2690, 1997. 41. Gerlitz, M. et al., Effect of daunorubicin dpsH gene on the choice of starter unit and cyclization pattern reveals that type II polyketide synthetases can be unfaithful yet intriguing, J. Am. Chem. Soc. 119, 7392, 1997. 42. Kantola, J. et al., Expanding the scope of aromatic polyketides by combinatorial biosynthesis, Comb. Chem. High Throughput Screen. 6, 501, 2003. 43. Olano, C. et al., A two-plasmid system for the glycosylation of polyketide antibiotics: bioconversion of epsilon-rhodomycinone to rhodomycin D, Chem. Biol. 6, 845, 1999. 44. Eckardt, K. et al., Biosynthesis of anthracyclinones: isolation of a new early cyclization product aklaviketone, J Antibiot (Tokyo) 41, 788, 1988. 45. Grein, A., Antitumor anthracyclines produced by Streptomyces peucetius, Adv. Appl. Microbiol. 32, 203, 1987. 46. McGuire, J. C. et al., Biosynthesis of daunorubicin glycosides: role of epsilon-rhodomycinone, Antimicrob. Agents Chemother. 18, 454, 1980. 47. Cassinelli, G. et al., 13-Deoxycarminomycin, a new biosynthetic anthracycline, J. Nat. Prod. 48, 435, 1985. 48. Hutchinson, C. R., Biosynthetic studies of daunorubicin and tetracenomycin C, Chem. Rev. 97, 2525, 1997. 49. Madduri, K. et al., Production of the antitumor drug epirubicin (4-epidoxorubicin) and its precursor by a genetically engineered strain of Streptomyces peucetius, Nat. Biotechnol. 16, 69, 1998. 50. Hopwood, D. A., Genetic contributions to understanding polyketide synthases, Chem. Rev. 97, 2465, 1997. 51. Remsing, L. L. et al., Mithramycin SK, a novel antitumor drug with improved therapeutic index, mith- ramycin SA, and demycarosyl-mithramycin SK: three new products generated in the mithramycin pro- ducer Streptomyces argillaceus through combinatorial biosynthesis, J. Am. Chem. Soc. 125, 5745, 2003. 52. Rohr, J., Mendez, C., and Salas, J. A., The biosynthesis of aureolic acid group antibiotics, Bioorg. Chem. 27, 41, 1999. 53. Hughes, S., Peel-White, A. L., and Peterson, C. K., Paget’s disease of bone — current thinking and management, J. Manipulative Physiol. Ther. 15, 242, 1992. 54. Cons, B. M. and Fox, K. R., Effects of the antitumor antibiotic mithramycin on the structure of repetitive DNA regions adjacent to its GC-rich binding site, Biochemistry 30, 6314, 1991. 55. Cons, B. M. and Fox, K. R., The GC-selective ligand mithramycin alters the structure of (AT)n sequences flanking its binding sites, FEBS Lett. 264, 100, 1990. 56. Montanari, A. and Rosazza, J. P., Biogenesis of chromomycin A3 by Streptomyces griseus, J. Antibiot. 43, 883, 1990. 57. Lombo, F. et al., Characterization of Streptomyces argillaceus genes encoding a polyketide synthase involved in the biosynthesis of the antitumor mithramycin, Gene 172, 87, 1996. 58. Lombo, F. et al., Cloning and insertional inactivation of Streptomyces argillaceus genes involved in the earliest steps of biosynthesis of the sugar moieties of the antitumor polyketide mithramycin, J. Bacteriol. 179, 3354, 1997. 59. Prado, L. et al., Analysis of two chromosomal regions adjacent to genes for a type II polyketide synthase involved in the biosynthesis of the antitumor polyketide mithramycin in Streptomyces argilla- ceus, Mol. Gen. Genet. 261, 216, 1999. 60. Gonzalez, A. et al., The mtmVUC genes of the mithramycin gene cluster in Streptomyces argillaceus are involved in the biosynthesis of the sugar moieties, Mol. Gen. Genet. 264, 827, 2001. Copyright © 2005 CRC Press, LLC 61. Menendez, N. et al., Biosynthesis of the antitumor chromomycin A3 in Streptomyces griseus: analysis of the gene cluster and rational design of novel chromomycin analogs, Chem. Biol. 11, 21, 2004. 62. Prado, L. et al., Oxidative cleavage of premithramycin is one of the last steps in the biosynthesis of the antitumor drug mithramycin, Chem. Biol. 6, 19, 1999. 63. Fernandez, E. et al., Identification of two genes from Streptomyces argillaceus encoding glycosyl- transferases involved in transfer of a disaccharide during biosynthesis of the antitumor drug mithra- mycin, J. Bacteriol. 180, 4929, 1998. 64. Fernadez Lozano, M. et al., Characterization of two polyketide methyltransferases involved in the biosynthesis of the antitumor drug mithramycin by Streptomyces argillaceus, J. Biol. Chem. 275, 3065, 2000. 65. Lombo, F. et al., The novel hybrid antitumor compound premithramycinone H provides indirect evidence for a tricyclic intermediate of the biosynthesis of the aurolic acid antibiotic mithramycin, Angew. Chem. Int. Ed. 39, 796, 2000. 66. Kantola, J. et al., Folding of the polyketide chain is not dictated by minimal polyketide synthase in the biosynthesis of mithramycin and anthracycline, Chem. Biol. 4, 751, 1997. 67. Kunzel, E. et al., Tetracenomycin M, a novel genetically engineered tetracenomycin resulting from a combination of mithramycin and tetracenomycin biosynthetic genes, Chem.-A Eur. J. 3, 1675, 1997. 68. Wohlert, S.-E. et al., Novel hybrid tetracenomycins through combinatorial biosynthesis using a glycosyltransferase encoded by the elm genes in cosmid 16F4 and which shows a broad sugar substrate specificity, J. Am. Chem. Soc. 120, 10596, 1998. 69. August, P. R. et al., Biosynthesis of the ansamycin antibiotic rifamycin: deductions from the molecular analysis of the rif biosynthetic gene cluster of Amycolatopsis mediterranei S699, Chem. Biol. 5, 69, 1998. 70. Yu, T. W. et al., The biosynthetic gene cluster of the maytansinoid antitumor agent ansamitocin from Actinosynnema pretiosum, Proc. Natl. Acad. Sci. USA 99, 7968, 2002. 71. Hatano, K. et al., Biosynthetic origin of aminobenzenoid nucleus (C7N-unit) of ansamitocin, a group of novel maytansinoid antibiotics, J. Antibiot. 35, 1415, 1982. 72. Tang, L. et al., Characterization of the enzymatic domains in the modular polyketide synthase involved in rifamycin B biosynthesis by Amycolatopsis mediterranei, Gene 216, 255, 1998. 73. Reider, P. J. and Roland, D. M., in The Alkaloids, Pelletier, S. W. Academic Press, New York, 1984, pp. 71. 74. Richter, K. and Buchner, J., Hsp90: chaperoning signal transduction, J. Cell Physiol. 188, 281, 2001. 75. Whitesell, L. et al., Inhibition of heat shock protein HSP90-pp60v-src heteroprotein complex formation by benzoquinone ansamycins: essential role for stress proteins in oncogenic transformation, Proc. Natl. Acad. Sci. USA 91,

8324, 1994. 76. Blagosklonny, M. V., Hsp-90-associated oncoproteins: multiple targets of geldanamycin and its ana- logs, Leukemia 16, 455, 2002. 77. Kim, C. G. et al., 3-Amino-5-hydroxybenzoic acid synthase, the terminal enzyme in the formation of the precursor of mC7N units in rifamycin and related antibiotics, J. Biol. Chem. 273, 6030, 1998. 78. Schupp, T. et al., Cloning and sequence analysis of the putative rifamycin polyketide synthase gene cluster from Amycolatopsis mediterranei, FEMS Microbiol. Lett. 159, 201, 1998. 79. Rascher, A. et al., Cloning and characterization of a gene cluster for geldanamycin production in Streptomyces hygroscopicus NRRL 3602, FEMS Microbiol. Lett. 218, 223, 2003. 80. Spiteller, P. et al., The post-polyketide synthase modification steps in the biosynthesis of the antitumor agent ansamitocin by Actinosynnema pretiosum, J. Am. Chem. Soc. 125, 14236, 2003. 81. Watanabe, K. et al., Engineered biosynthesis of an ansamycin polyketide precursor in Escherichia coli, Proc. Natl. Acad. Sci. USA 100, 9774, 2003. 82. Lambalot, R. H. et al., A new enzyme superfamily - the phosphopantetheinyl transferases, Chem. Biol. 3, 923, 1996. 83. Walsh, C. T. et al., Post-translational modification of polyketide and nonribosomal peptide synthases, Curr. Opin. Chem. Biol. 1, 309, 1997. 84. Pfeifer, B. et al., Biosynthesis of complex polyketides in a metabolically engineered strain of E. coli, Science 291, 1790, 2001. Copyright © 2005 CRC Press, LLC 85. Ghisalba, O. and Nuesch, J., A genetic approach to the biosynthesis of the rifamycin-chromophore in Nocardia mediterranei. IV. Identification of 3-amino-5-hydroxybenzoic acid as a direct precursor of the seven-carbon amino starter-unit, J. Antibiot. 34, 64, 1981. 86. Tsuji, S., Cane, D., and Khosla, C., Selective protein-protein interactions direct channeling of inter- mediates between polyketide synthase modules, Biochemistry 40, 2326, 2001. 87. Wu, N., Cane, D. E., and Khosla, C., Quantitative analysis of the relative contributions of donor acyl carrier proteins, acceptor ketosynthases, and linker regions to intermodular transfer of intermediates in hybrid polyketide synthases, Biochemistry 41, 5056, 2002. 88. Wu, N. et al., Assessing the balance between protein-protein interactions and enzyme-substrate interactions in the channeling of intermediates between polyketide synthase modules, J. Am. Chem. Soc. 123, 6465, 2001. 89. Kato, Y. et al., Functional expression of genes involved in the biosynthesis of the novel polyketide chain extension unit, methoxymalonyl-acyl carrier protein, and engineered biosynthesis of 2-desme- thyl-2-methoxy-6-deoxyerythronolide B, J. Am. Chem. Soc. 124, 5268, 2002. 90. Floss, H. G. and Yu, T. W., Lessons from the rifamycin biosynthetic gene cluster, Curr. Opin. Chem. Biol. 3, 592, 1999. 91. He, L., Orr, G. A., and Horwitz, S. B., Novel molecules that interact with microtubules and have functional activity similar to Taxol, Drug Discov. Today 6, 1153, 2001. 92. Hoefle, G. et al., Antibiotics from gliding bacteria. 77. Epothilone A and B — novel 16-membered macrolides with cytotoxic activity: isolation, crystal structure, and conformation in solution., Angew. Chem. Int. Ed. 35, 1567, 1996. 93. Arslanian, R. L. et al., Large-scale isolation and crystallization of epothilone D from Myxococcus xanthus cultures, J. Nat. Prod. 65, 570, 2002. 94. McCarthy, A. A., Kosan biosciences. Better chemistry through genetics, Chem. Biol. 9, 849, 2002. 95. McAllister, K. A. et al., Biochemical and molecular analyses of the Streptococcus pneumoniae acyl carrier protein synthase, an enzyme essential for fatty acid biosynthesis, J. Biol. Chem. 275, 30864, 2000. 96. Molnar, I. et al., The biosynthetic gene cluster for the microtubule-stabilizing agents epothilones A and B from Sorangium cellulosum So ce90, Chem. Biol. 7, 97, 2000. 97. Tang, L. et al., Cloning and heterologous expression of the epothilone gene cluster, Science 287, 640, 2000. 98. Julien, B. et al., Isolation and characterization of the epothilone biosynthetic gene cluster from Sorangium cellulosum, Gene 249, 153, 2000. 99. Chou, T. C. et al., Desoxyepothilone B: an efficacious microtubule-targeted antitumor agent with a promising in vivo profile relative to epothilone B, Proc. Natl. Acad. Sci. USA 95, 9642, 1998. 100. Chou, T. C. et al., Desoxyepothilone B is curative against human tumor xenografts that are refractory to paclitaxel, Proc. Natl. Acad. Sci. USA 95, 15798, 1998. 101. Julien, B. and Shah, S., Heterologous expression of epothilone biosynthetic genes in Myxococcus xanthus, Antimicrob. Agents Chemother. 46, 2772, 2002. 102. Frykman, S. et al., Modulation of epothilone analog production through media design, J. Ind. Micro- biol. Biotechnol. 28, 17, 2002. 103. Lau, J. et al., Optimizing the heterologous production of epothilone D in Myxococcus xanthus, Biotechnol. Bioeng. 78, 280, 2002. 104. Starks, C. M. et al., Isolation and characterization of new epothilone analogues from recombinant Myxococcus xanthus fermentations, J. Nat. Prod. 66, 1313, 2003. 105. Chen, H. et al., Epothilone biosynthesis: assembly of the methylthiazolylcarboxy starter unit on the EpoB subunit, Chem. Biol. 8, 899, 2001. 106. O’Connor, S. E., Chen, H., and Walsh, C. T., Enzymatic assembly of epothilones: the EpoC subunit and reconstitution of the EpoA-ACP/B/C polyketide and nonribosomal peptide interfaces, Biochem- istry 41, 5685, 2002. 107. Schneider, T. L., Shen, B., and Walsh, C. T., Oxidase domains in epothilone and bleomycin biosyn- thesis: thiazoline to thiazole oxidation during chain elongation, Biochemistry 42, 9722, 2003. 108. Walsh, C. T., O’Connor, S. E., and Schneider, T. L., Polyketide-nonribosomal peptide epothilone antitumor agents: the EpoA, B, C subunits, J. Ind. Microbiol. Biotechnol. 30, 448, 2003. Copyright © 2005 CRC Press, LLC 109. Hara, M. et al., DC 107, a novel antitumor antibiotic produced by a Streptomyces sp., J. Antibiot. 42, 333, 1989. 110. Hara, M. et al., Leinamycin, a new antitumor antibiotic from Streptomyces: producing organism, fermentation and isolation, J. Antibiot. 42, 1768, 1989. 111. Kanda, Y. et al., Synthesis and antitumor activity of leinamycin derivatives: modifications of C-8 hydroxy and C-9 keto groups, Bioorg. Med. Chem. Lett. 8, 909, 1998. 112. Kanda, Y. et al., Synthesis and antitumor activity of novel thioester derivatives of leinamycin, J. Med. Chem. 42, 1330, 1999. 113. Gates, K. S., Mechanisms of DNA damage by leinamycin, Chem. Res. Toxicol. 13, 953, 2000. 114. Cheng, Y. Q., Tang, G. L., and Shen, B., Identification and localization of the gene cluster encoding biosynthesis of the antitumor macrolactam leinamycin in Streptomyces atroolivaceus S-140, J. Bac- teriol. 184, 7013, 2002. 115. Cheng, Y. Q., Tang, G. L., and Shen, B., Type I polyketide synthase requiring a discrete acyltransferase for polyketide biosynthesis, Proc. Natl. Acad. Sci. USA 100, 3149, 2003. 116. Tang, G. L., Cheng, Y. Q., and Shen, B., Leinamycin biosynthesis revealing unprecedented architectural complexity for a hybrid polyketide synthase and nonribosomal peptide synthetase, Chem. Biol. 11, 33, 2004. 117. Edo, K. et al., The structure of neocarzinostatin chromophore possessing a novel bicyclo[7,3,0]dodec- adiyne system, Tetrahedron Lett. 26, 331, 1985. 118. Shen, B., Liu, W., and Nonaka, K., Enediyne natural products: biosynthesis and prospect towards engineering novel antitumor agents, Curr. Med. Chem. 10, 2317, 2003. 119. Thorson, J. S. et al., Understanding and exploiting nature’s chemical arsenal: the past, present and future of calicheamicin research, Curr. Pharm. Des. 6, 1841, 2000. 120. Biggins, J. B., Onwueme, K. C., and Thorson, J. S., Resistance to enediyne antitumor antibiotics by CalC self-sacrifice, Science 301, 1537, 2003. 121. Doyle, T. W., Enediyne antibiotics as antitumor agents. Marcel Dekker, New York, 1995. 122. Maeda, H., Edo, K., and Ishida, N., Neocarzinostatin: The past, present, and future of an anticancer drug. Springer, New York, 1997. 123. Jones, G. B. and Fouad, F. S., Designed enediyne antitumor agents, Curr. Pharm. Des. 8, 2415, 2002. 124. Sievers, E. L. et al., Selective ablation of acute myeloid leukemia using antibody-targeted chemother- apy: a phase I study of an anti-CD33 calicheamicin immunoconjugate. Blood 93, 3678, 1999. 125. Brukner, I., C-1027 Taiho Pharmaceutical Co. Ltd., Curr. Opin. Oncologic, Endocrine Metabol. Invest. Drugs 2, 344, 2002. 126. Bross, P. F. et al., Approval summary: gemtuzumab ozogamicin in relapsed acute myeloid leukemia, Clin. Cancer Res. 7, 1490, 2001. 127. Thorson, J. S. et al., Enediyne biosynthesis and self-resistance: a progress report, Bioorg. Chem. 27, 172, 1999. 128. Liu, W. et al., Rapid PCR amplification of minimal enediyne polyketide synthase cassettes leads to a predictive familial classification model, Proc. Natl. Acad. Sci. USA 100, 11959, 2003. 129. Hensens, O. D., Giner, J. L., and Golberg, I. H., Biosynthesis of NCS Chrom A, the chromophore of the antitumor antibiotic neocarzinostatin, J. Am. Chem. Soc. 111, 3295, 1989. 130. Ahlert, J. et al., The calicheamicin gene cluster and its iterative type I enediyne PKS, Science 297, 1173, 2002. 131. Liu, W. et al., Biosynthesis of the enediyne antitumor antibiotic C-1027, Science 297, 1170, 2002. 132. Zazopoulos, E. et al., A genomics-guided approach for discovering and expressing cryptic metabolic pathways, Nat. Biotechnol. 21, 187, 2003. 133. Liu, W. and Shen, B., Genes for production of the enediyne antitumor antibiotic C-1027 in Strepto- myces globisporus are clustered with the cagA gene that encodes the C-1027 apoprotein, Antimicrob. Agents Chemother. 44, 382, 2000. 134. Christenson, S. D. et al., A novel 4-methylideneimidazole-5-one-containing tyrosine aminomutase in enediyne antitumor antibiotic C-1027 biosynthesis, J. Am. Chem. Soc. 125, 6062, 2003. 135. Murrell, J. M., Liu, W., and Shen, B., Biochemical characterization of the SgcA1 alpha-D-glucopy- ranosyl-1-phosphate thymidylyltransferase from the enediyne antitumor antibiotic C-1027 biosyn- thetic pathway and overexpression of sgcA1 in Streptomyces globisporus to improve C-1027 production, J. Nat. Prod. 67, 206, 2004. Copyright © 2005 CRC Press, LLC 136. Liu, W. et al., The neocarzinostatin biosynthetic gene cluster from Streptomyces carzinostaticus ATCC15944 inolving two iterative type I polyketide synthase, Chem. Biol., 12, 293–302, 2005. 137. Omura, S. et al., Genome sequence of an industrial microorganism Streptomyces avermitilis: deducing the ability of producing secondary metabolites, Proc. Natl. Acad. Sci. USA 98, 12215, 2001. 138. Bentley, S. D. et al., Complete genome sequence of the model actinomycete Streptomyces coelicolor A3(2), Nature 417, 141, 2002. 139. Rappa, G., Lorico, A., and Sartorelli, A. C., Potentiation by novobiocin of the cytotoxic activity of etoposide (VP-16) and teniposide (VM-26), Int. J. Cancer 51, 780, 1992. 140. Lorico, A., Rappa, G., and Sartorelli, A. C., Novobiocin-induced accumulation of etoposide (VP-16) in WEHI-3B D+ leukemia cells, Int. J. Cancer 52, 903, 1992. 141. Shiozawa, K. et al., Reversal of breast cancer resistance protein (BCRP/ABCG2)-mediated drug resistance by novobiocin, a coumermycin antibiotic, Int. J. Cancer 108, 146, 2004. 142. Steffensky, M. et al., Identification of the novobiocin biosynthetic gene cluster of Streptomyces spheroides NCIB 11891, Antimicrob. Agents Chemother. 44, 1214, 2000. 143. Pojer, F., Li, S. M., and Heide, L., Molecular cloning and sequence analysis of the clorobiocin biosynthetic gene cluster: new insights into the biosynthesis of aminocoumarin antibiotics, Microbi- ology 148, 3901, 2002. 144. Wang, Z. X., Li, S. M., and Heide, L., Identification of the coumermycin A(1) biosynthetic gene cluster of Streptomyces rishiriensis DSM 40489, Antimicrob. Agents Chemother. 44, 3040, 2000. 145. Galm, U. et al., Cloning and analysis of the simocyclinone biosynthetic gene cluster of Streptomyces antibioticus Tü 6040, Arch. Microbiol. 178, 102, 2002. 146. Chen, H. et al., Formation of beta-hydroxy histidine in the biosynthesis of nikkomycin antibiotics, Chem. Biol. 9, 103, 2002. 147. Chen, H. et al., Aminoacyl-S-enzyme intermediates in β-hydroxylations and α,β-desaturations of amino acids in peptide antibiotics, Biochemistry 40, 11651, 2001. 148. Chen, H. and Walsh, C. T., Coumarin formation in novobiocin biosynthesis: beta-hydroxylation of the aminoacyl enzyme tyrosyl-S-NovH by a cytochrome P450 NovI, Chem. Biol. 74, 1, 2001. 149. Thomas, M. G., Burkart, M. D., and Walsh, C. T., Conversion of L-proline to pyrrolyl-2-carboxyl-S- PCP during undecylprodigiosin and pyoluteorin biosynthesis, Chem. Biol. 9, 171, 2002. 150. Xu, H. et al., Genetic analysis of the biosynthesis of the pyrrole and carbamoyl moieties of coumer- mycin A1 and novobiocin, Mol. Genet. Genom. 268, 387, 2002. 151. Li, S. M. et al., Methyltransferase genes in Streptomyces rishiriensis: new coumermycin derivatives from gene-inactivation experiments, Microbiology 148, 3317, 2002. 152. Xu, H. et al., CloN2, a novel acyltransferase involved in the attachment of the pyrrole-2-carboxyl moiety to the deoxysugar of clorobiocin, Microbiology 149, 2183, 2003. 153. Westrich, L., Heide, L., and Li, S. M., CloN6, a novel methyltransferase catalysing the methylation of the pyrrole-2-carboxyl moiety of clorobiocin, ChemBioChem 4, 768, 2003. 154. Eustaquio, A. S. et al., Clorobiocin biosynthesis in streptomyces. Identification of the halogenase and generation of structural analogs, Chem. Biol. 10, 279, 2003. 155. Merchant, J. et al., Phase I clinical and pharmacokinetic study of NSC 655649, a rebeccamycin analogue, giving in both single-dose and multiple-dose formats, Clin. Cancer Res. 8, 2193, 2002. 156. Xu, H., Heide, L., and Li,

S. M., New aminocoumarin antibiotics formed by a combined mutational and chemoenzymatic approach utilizing the carbamoyltransferase NovN, Chem. Biol. 11, 655, 2004. 157. Galm, U. et al., In vitro and in vivo production of new aminocoumarins by a combined biochemical, genetic, and synthetic approach, Chem. Biol. 11, 173, 2004. 158. Pfeifer, B. and Khosla, C., Biosynthesis of polyketides in heterologous hosts, Micro. Mol. Biol. Rev. 65, 106, 2001. 159. Doekel, S., Eppelmann, K., and Marahiel, M. A., Heterologous expression of nonribosomal peptide synthetases in B. subtilis: construction of a bi-functional B. subtilis/E. coli shuttle vector system, FEMS Microbiol. Lett. 216, 185, 2002. 160. Eppelmann, K., Doekel, S., and Marahiel, M. A., Engineered biosynthesis of the peptide antibiotic bacitracin in the surrogate host Bacillus subtilis, J. Biol. Chem. 276, 34824, 2001. 161. Schloss, P. D. and Handelsman, J., Biotechnological prospects from metagenomics, Curr. Opin. Biotechnol. 14, 303, 2003. 162. Gillespie, D. E. et al., Isolation of antibiotics turbomycin A and B from a metagenomic library of soil microbial DNA, Appl. Environ. Microbiol. 68, 4301, 2002. Copyright © 2005 CRC Press, LLC 163. Rondon, M. R. et al., Cloning the soil metagenome: a strategy for accessing the genetic and functional diversity of uncultured microorganisms, Appl. Environ. Microbiol. 66, 2541, 2000. 164. Rondon, M. R., Goodman, R. M., and Handelsman, J., The Earth’s bounty: assessing and accessing soil microbial diversity, Trends Biotechnol. 17, 403, 1999. 165. Handelsman, J. et al., Molecular biological access to the chemistry of unknown soil microbes: a new frontier for natural products, Chem. Biol. 5, R245, 1998. 166. Stein, J. L. et al., Characterization of uncultivated prokaryotes: isolation and analysis of a 40-kilobase- pair genome fragment from a planktonic marine archaeon, J. Bacteriol. 178, 591, 1996. 167. Schleper, C. et al., Genomic analysis reveals chromosomal variation in natural populations of the uncultured psychrophilic archaeon Cenarchaeum symbiosum, J. Bacteriol. 180, 5003, 1998. 168. Beja, O. et al., Construction and analysis of bacterial artificial chromosome libraries from a marine microbial assemblage, Environ. Microbiol. 2, 516, 2000. 169. Shizuya, H. et al., Cloning and stable maintenance of 300-kilobase-pair fragments of human DNA in Escherichia coli using an F-factor-based vector, Proc. Natl. Acad. Sci. USA 89, 8794, 1992. 170. Sosio, M. et al., Artificial chromosomes for antibiotic-producing actinomycetes, Nat. Biotechnol. 18, 343, 2000. 171. Sosio, M., Bossi, E., and Donadio, S., Assembly of large genomic segments in artificial chromosomes by homologous recombination in Escherichia coli, Nucl. Acids Res. 29, E37, 2001. 172. Alduina, R. et al., Artificial chromosome libraries of Streptomyces coelicolor A3(2) and Planobispora rosea, FEMS Microbiol. Lett. 218, 181, 2003. 173. Handelsman, J. et al., Cloning the metagenome: culture-independent access to the diversity and functions of the uncultured microbial world, in Methods in Microbiology, Wren, G. and Dorrell, N. Academic Press, New York, 2002, pp. 241. 174. Stemmer, W. P., DNA shuffling by random fragmentation and reassembly: in vitro recombination for molecular evolution, Proc. Natl. Acad. Sci. USA 91, 10747, 1994. 175. Stemmer, W. P., Rapid evolution of a protein in vitro by DNA shuffling, Nature 370, 389, 1994. 176. Zhang, Y. X. et al., Genome shuffling leads to rapid phenotypic improvement in bacteria, Nature 415, 644, 2002. 177. Patnaik, R. et al., Genome shuffling of Lactobacillus for improved acid tolerance, Nat. Biotechnol. 20, 707, 2002. 178. Barton, W. A. et al., Structure, mechanism and engineering of a nucleotidylyltransferase as a first step toward glycorandomization, Nat. Struct. Biol. 8, 545, 2001. 179. Barton, W. A. et al., Expanding pyrimidine diphosphosugar libraries via structure-based nucleotidy- lyltransferase engineering, Proc. Natl. Acad. Sci. USA 99, 13397, 2002. 180. Fu, X. et al., Antibiotic optimization via in vitro glycorandomization, Nat. Biotechnol. 21, 1467, 2003. 181. Thorson, J. S. et al., Structure-based enzyme engineering and its impact on in vitro glycorandomiza- tion, ChemBioChem 5, 16, 2004. 182. Raty, K. et al., Cloning and characterization of Streptomyces galilaeus aclacinomycins polyketide synthase (PKS) cluster, Gene 293, 115, 2002. 183. Kantola, J. et al., Elucidation of anthracyclinone biosynthesis by stepwise cloning of genes for anthracyclines from three different Streptomyces spp., Microbiology 146 (Pt 1), 155, 2000. 184. Schauwecker, F. et al., Molecular cloning of the actinomycin synthetase gene cluster from Strepto- myces chrysomallus and functional heterologous expression of the gene encoding actinomycin syn- thetase II, J. Bacteriol. 180, 2468, 1998. 185. Pfennig, F., Schauwecker, F., and Keller, U., Molecular characterization of the genes of actinomycin synthetase I and of a 4-methyl-3-hydroxyanthranilic acid carrier protein involved in the assembly of the acylpeptide chain of actinomycin in Streptomyces, J. Biol. Chem. 274, 12508, 1999. 186. Schauwecker, F. et al., Construction and in vitro analysis of a new bi-modular polypeptide synthetase for synthesis of N-methylated acyl peptides, Chem. Biol. 7, 287, 2000. 187. Yu, T. W. et al., Mutational analysis and reconstituted expression of the biosynthetic genes involved in the formation of 3-amino-5-hydroxybenzoic acid, the starter unit of rifamycin biosynthesis in Amycolatopsis mediterranei S699, J. Biol. Chem. 276, 12546, 2001. 188. He, J. and Hertweck, C., Iteration as programmed event during polyketide assembly; molecular analysis of the aureothin biosynthesis gene cluster, Chem. Biol. 10, 1225, 2003. Copyright © 2005 CRC Press, LLC 189. Du, L. et al., The biosynthetic gene cluster for the antitumor drug bleomycin from Streptomyces verticillus ATCC15003 supporting functional interactions between nonribosomal peptide synthetases and a polyketide synthase, Chem. Biol. 7, 623, 2000. 190. Shen, B. et al., The biosynthetic gene cluster for the anticancer drug bleomycin from Streptomyces verticillus ATCC15003 as a model for hybrid peptide-polyketide natural product biosynthesis, J. Ind. Microbiol. Biotechnol. 27, 378, 2001. 191. Shen, B. et al., Cloning and characterization of the bleomycin biosynthetic gene cluster from Strep- tomyces verticillus ATCC15003, J. Nat. Prod. 65, 422, 2002. 192. Sanchez, C. et al., Cloning and characterization of a phosphopantetheinyl transferase from Strepto- myces verticillus ATCC15003, the producer of the hybrid peptide-polyketide antitumor drug bleomy- cin, Chem. Biol. 8, 725, 2001. 193. Du, L. and Shen, B., Identification and characterization of a type II peptidyl carrier protein from the bleomycin producer Streptomyces verticillus ATCC 15003, Chem. Biol. 6, 507, 1999. 194. Du, L. et al., An oxidation domain in the BlmIII non-ribosomal peptide synthetase probably catalyzing thiazole formation in the biosynthesis of the anti-tumor drug bleomycin in Streptomyces verticillus ATCC15003, FEMS Microbiol. Lett. 189, 171, 2000. 195. Du, L. et al., BlmIII and BlmIV nonribosomal peptide synthetase-catalyzed biosynthesis of the bleo- mycin bithiazole moiety involving both in cis and in trans aminoacylation, Biochemistry 42, 9731, 2003. 196. Olano, C. et al., Biosynthesis of the angiogenesis inhibitor borrelidin by Streptomyces parvulus Tü4055: cluster analysis and assignment of functions, Chem. Biol. 11, 87, 2004. 197. Schmutz, E. et al., Resistance genes of aminocoumarin producers: two type II topoisomerase genes confer resistance against coumermycin A1 and clorobiocin, Antimicrob. Agents Chemother. 47, 869, 2003. 198. Pojer, F. et al., CloR, a bifunctional non-heme iron oxygenase involved in clorobiocin biosynthesis, J. Biol. Chem. 278, 30661, 2003. 199. Pojer, F. et al., CloQ, a prenyltransferase involved in clorobiocin biosynthesis, Proc. Natl. Acad. Sci. USA 100, 2316, 2003. 200. Schmutz, E. et al., An unusual amide synthetase (CouL) from the coumermycin A1 biosynthetic gene cluster from Streptomyces rishiriensis DSM 40489, Eur. J. Biochem. 270, 4413, 2003. 201. Shen, B. and Hutchinson, C. R., TetracenomycinF2 cyclase: intramolecular aldol condensation in the biosynthesis of tetracenomycin C in Streptomyces glaucens, Biochemistry 32, 11149, 1993. 202. Rajgarhia, V. B., Priestley, N. D., and Strohl, W. R., The product of dpsC confers starter unit fidelity upon the daunorubicin polyketide synthase of Streptomyces sp. strain C5, Metab. Eng. 3, 49, 2001. 203. Meurer, G. et al., Iterative type II polyketide synthases, cyclases and ketoreductases exhibit context- dependent behavior in the biosynthesis of linear and angular decapolyketides, Chem. Biol. 4, 433, 1997. 204. Decker, H. et al., Identification of Streptomyces olivaceus Tü 2353 genes involved in the production of the polyketide elloramycin, Gene 166, 121, 1995. 205. Blanco, G. et al., Identification of a sugar flexible glycosyltransferase from Streptomyces olivaceus, the producer of the antitumor polyketide elloramycin, Chem. Biol. 8, 253, 2001. 206. Rafanan, E. R., Jr. et al., Cloning, sequencing, and heterologous expression of the elmGHIJ genes involved in the biosynthesis of the polyketide antibiotic elloramycin from Streptomyces olivaceus Tü 2353, J. Nat. Prod. 64, 444, 2001. 207. Patallo, E. P. et al., Deoxysugar methylation during biosynthesis of the antitumor polyketide ellora- mycin by Streptomyces olivaceus. Characterization of three methyltransferase genes, J. Biol. Chem. 276, 18765, 2001. 208. Summers, R. G. et al., Sequencing and mutagenesis of genes from erythromycin biosynthetic gene cluster of Saccharopolyspora erythraea that are involved in L-mycarose and D-desomine production, Microbiology 143, 3251, 1997. 209. Rodriguez, L. et al., Generation of hybrid elloramycin analogs by combinatorial biosynthesis using genes from anthracycline-type and macrolide biosynthetic pathways, J. Mol. Microbiol. Biotechnol. 2, 271, 2000. 210. Fischer, C. et al., Digitoxosyltetracenomycin C and glucosyltetracenomycin C, two novel elloramycin analogues obtained by exploring the sugar donor substrate specificity of glycosyltransferase ElmGT, J. Nat. Prod. 65, 1685, 2002. Copyright © 2005 CRC Press, LLC 211. Arslanian, R. L. et al., A new cytotoxic epothilone from modified polyketide synthases heterologously expressed in Myxococcus xanthus, J. Nat. Prod. 65, 1061, 2002. 212. Schneider, T. L., Walsh, C. T., and O’Connor, S. E., Utilization of alternate substrates by the first three modules of the epothilone synthetase assembly line, J. Am. Chem. Soc. 124, 11272, 2002. 213. O’Connor, S. E., Walsh, C. T., and Liu, F., Biosynthesis of epothilone intermediates with alternate starter units: engineering polyketide-nonribosomal interfaces, Angew. Chem. Int. Ed. 42, 3917, 2003. 214. Wendt-Pienkowski, E. et al., Cloning, sequencing, and analysis of the fredericamycin biosynthetic gene cluster from Streptomyces griseus, Submitted, 2005. 215. Fischer, C., Lipata, F., and Rohr, J., The complete gene cluster of the antitumor agent gilvocarcin V and its implication for the biosynthesis of the gilvocarcins, J. Am. Chem. Soc. 125, 7818, 2003. 216. Bililign, T. et al., The hedamycin locus implicates a novel aromatic PKS priming mechanism, Chem. Biol. 11, 959–969, 2004. 217. Han, L. et al., Cloning and characterization of polyketide synthase genes for jadomycin B biosynthesis in Streptomyces venezuelae ISP5230, Microbiology 140, 3379, 1994. 218. Han, L. et al., An acyl-coenzyme A carboxylase encoding gene associated with jadomycin biosynthesis in Streptomyces venezuelae ISP5230, Microbiology 146, 903, 2000. 219. Yang, K., Han, L., and Vining, L. C., Regulation of jadomycin B production in Streptomyces venezuelae ISP5230: involvement of a repressor gene, jadR2, J. Bacteriol. 177, 6111, 1995. 220. Yang, K. et al., Accumulation of the angucycline antibiotic rabelomycin after disruption of an oxy- genase gene in the jadomycin B biosynthetic gene cluster of Streptomyces venezuelae, Microbiology 142 (Pt 1), 123, 1996. 221. Wang, L., McVey, J., and Vining, L. C., Cloning and functional analysis of a phosphopantetheinyl transferase superfamily gene associated with jadomycin biosynthesis in Streptomyces venezuelae ISP5230, Microbiology 147, 1535, 2001. 222. Wang, L., White, R. L., and Vining, L. C., Biosynthesis of the dideoxysugar component of jadomycin B: genes in the jad cluster of Streptomyces venezuelae ISP5230 for L-digitoxose assembly and transfer to the angucycline aglycone, Microbiology 148, 1091, 2002. 223. Wohlert, S. et al., Insights about the biosynthesis of the avermectin deoxysugar L-oleandrose through heterologous expression of Streptomyces avermitilis deoxysugar genes in Streptomyces lividans, Chem. Biol. 8, 681, 2001. 224. Westrich, L. et al., Cloning and characterization of a gene cluster from Streptomyces cyanogenus S136 probably involved in landomycin biosynthesis, FEMS Microbiol. Lett. 170, 381, 1999. 225. von Mulert, U. et al., Expression of the landomycin biosynthetic gene cluster in a PKS mutant of Streptomyces fradiae is dependent on the coexpression of a putative transcriptional activator gene, FEMS Microbiol. Lett. 230, 91, 2004. 226. Trefzer, A. et al., Elucidation of the function of two glycosyltransferase genes (lanGT1 and lanGT4) involved in landomycin biosynthesis and generation of new oligosaccharide antibiotics, Chem. Biol. 8, 1239, 2001. 227. Trefzer, A. et al., Rationally designed glycosylated premithramycins: hybrid aromatic polyketides using genes from three different biosynthetic pathways, J. Am. Chem. Soc. 124, 6056, 2002. 228. Lombo, F. et al., The mithramycin gene cluster of Streptomyces argillaceus contains a positive regulatory gene and two repeated DNA sequences that are located at both ends of the cluster, J. Bacteriol. 181, 642, 1999. 229. Blanco, G. et al., Deciphering the biosynthetic origin of the aglycone of the aureolic acid group of anti-tumor agents, Chem. Biol. 3, 193, 1996. 230.

Fernandez, E. et al., An ABC transporter is essential for resistance to the antitumor agent mithramycin in the producer Streptomyces argillaceus, Mol. Gen. Genet. 251, 692, 1996. 231. Lozano, M. J. et al., Characterization of two polyketide methyltransferases involved in the biosynthesis of the antitumor drug mithramycin by Streptomyces argillaceus, J. Biol. Chem. 275, 3065, 2000. 232. Rix, U. et al., Modification of post-PKS tailoring steps through combinatorial biosynthesis, Nat. Prod. Rep. 19, 542, 2002. 233. Remsing, L. L. et al., Ketopremithramycins and ketomithramycins, four new aureolic acid-type compounds obtained upon inactivation of two genes involved in the biosynthesis of the deoxysugar moieties of the antitumor drug mithramycin by Streptomyces argillaceus, reveal novel insights into post-PKS tailoring steps of the mithramycin biosynthetic pathway, J. Am. Chem. Soc. 124, 1606, 2002. Copyright © 2005 CRC Press, LLC 234. Mao, Y., Varoglu, M., and Sherman, D. H., Molecular characterization and analysis of the biosynthetic gene cluster for the antitumor antibiotic mitomycin C from Streptomyces lavendulae NRRL 2564, Chem. Biol. 6, 251, 1999. 235. Mao, Y., Varoglu, M., and Sherman, D. H., Genetic localization and molecular characterization of two key genes (mitAB) required for biosynthesis of the antitumor antibiotic mitomycin C, J. Bacteriol. 181, 2199, 1999. 236. Sheldon, P. J. et al., Mitomycin resistance in Streptomyces lavendulae includes a novel drug-binding- protein-dependent export system, J. Bacteriol. 181, 2507, 1999. 237. He, M., Sheldon, P. J., and Sherman, D. H., Characterization of a quinone reductase activity for the mitomycin C binding protein (MRD): Functional switching from a drug-activating enzyme to a drug- binding protein, Proc. Natl. Acad. Sci. USA 98, 926, 2001. 238. Kwon, H. J. et al., Cloning and heterologous expression of the macrotetrolide biosynthetic gene cluster revealed a novel polyketide synthase that lacks an acyl carrier protein, J. Am. Chem. Soc. 123, 3385, 2001. 239. Shen, B. and Kwon, H. J., Macrotetrolide biosynthesis: A novel type II polyketide synthase, Chem. Rec. 2, 389, 2002. 240. Smith, W. C., Xiang, L., and Shen, B., Genetic localization and molecular characterization of the nonS gene required for macrotetrolide biosynthesis in Streptomyces griseus DSM40695, Antimicrob. Agents Chemother. 44, 1809, 2000. 241. Kwon, H. J. et al., C-O bond formation by polyketide synthases, Science 297, 1327, 2002. 242. Ylihonko, K. et al., A gene cluster involved in nogalamycin biosynthesis from Streptomyces nogalater: sequence analysis and complementation of early-block mutations in the anthracycline pathway, Mol. Gen. Genet. 251, 113, 1996. 243. Torkkell, S. et al., The entire nogalamycin biosynthetic gene cluster of Streptomyces nogalater: characterization of a 20-kb DNA region and generation of hybrid structures, Mol. Genet. Genom. 266, 276, 2001. 244. Ylihonko, K. et al., Production of hybrid anthracycline antibiotics by heterologous expression of Streptomyces nogalater nogalamycin biosynthesis genes, Microbiology 142 (Pt 8), 1965, 1996. 245. Kunnari, T. et al., Isolation and characterization of 8-demethoxy steffimycins and generation of 2,8- demethoxy steffimycins in Streptomyces steffisburgensis by the nogalamycin biosynthesis genes, J. Antibiot. 50, 496, 1997. 246. Steffensky, M., Li, S. M., and Heide, L., Cloning, overexpression, and purification of novobiocic acid synthetase from Streptomyces spheroides NCIMB 11891, J. Biol. Chem. 275, 21754, 2000. 247. Eustaquio, A. S. et al., Novobiocin biosynthesis: inactivation of the putative regulatory gene novE and heterologous expression of genes involved in aminocoumarin ring formation, Arch. Microbiol. 180, 25, 2003. 248. Freel Meyers, C. L. et al., Initial characterization of novobiocic acid noviosyl transferase activity of NovM in biosynthesis of the antibiotic novobiocin, Biochemistry 42, 4179, 2003. 249. Albermann, C. et al., Substrate specificity of NovM: implications for Novobiocin biosynthesis and glycorandomization, Org. Lett. 5, 933, 2003. 250. Freel Meyers, C. L. et al., Characterization of NovP and NovN: Completion of novobiocin biosynthesis by sequential tailoring of the noviosyl ring, Angew. Chem. Int. Ed. 43, 67, 2004. 251. Swan, D. G. et al., Characterisation of a Streptomyces antibioticus gene encoding a type I polyketide synthase which has an unusual coding sequence, Mol. Gen. Genet. 242, 358, 1994. 252. Rodriguez, A. M. et al., A cytochrome P450-like gene possibly involved in oleandomycin biosynthesis by Streptomyces antibioticus, FEMS Microbiol. Lett. 127, 117, 1995. 253. Olano, C. et al., A second ABC transporter is involved in oleandomycin resistance and its secretion by Streptomyces antibioticus, Mol. Microbiol. 16, 333, 1995. 254. Quiros, L. M. and Salas, J. A., Biosynthesis of the macrolide oleandomycin by Streptomyces antibi- oticus. Purification and kinetic characterization of an oleandomycin glucosyltransferase, J. Biol. Chem. 270, 18234, 1995. 255. Quiros, L. M. et al., Two glycosyltransferases and a glycosidase are involved in oleandomycin modification during its biosynthesis by Streptomyces antibioticus, Mol. Microbiol. 28, 1177, 1998. 256. Olano, C. et al., Analysis of a Streptomyces antibioticus chromosomal region involved in oleandomycin biosynthesis, which encodes two glycosyltransferases responsible for glycosylation of the macrolac- tone ring, Mol. Gen. Genet. 259, 299, 1998. Copyright © 2005 CRC Press, LLC 257. Shah, S. et al., Cloning, characterization and heterologous expression of a polyketide synthase and P-450 oxidase involved in the biosynthesis of the antibiotic oleandomycin, J. Antibiot. 53, 502, 2000. 258. Tang, L., Fu, H., and McDaniel, R., Formation of functional heterologous complexes using subunits from the picromycin, erythromycin and oleandomycin polyketide synthases, Chem. Biol. 7, 77, 2000. 259. Aguirrezabalaga, I. et al., Identification and expression of genes involved in biosynthesis of L- oleandrose and its intermediate L-olivose in the oleandomycin producer Streptomyces antibioticus, Antimicrob. Agents Chemother. 44, 1266, 2000. 260. Doumith, M. et al., Interspecies complementation in Saccharopolyspora erythraea: elucidation of the function of oleP1, oleG1 and oleG2 from the oleandomycin biosynthetic gene cluster of Streptomyces antibioticus and generation of new erythromycin derivatives, Mol. Microbiol. 34, 1039, 1999. 261. Rodriguez, L. et al., Functional analysis of OleY L-oleandrosyl 3-O-methyltransferase of the olean- domycin biosynthetic pathway in Streptomyces antibioticus, J. Bacteriol. 183, 5358, 2001. 262. Long, P. F. et al., Engineering specificity of starter unit selection by the erythromycin-producing polyketide synthase, Mol. Microbiol. 43, 1215, 2002. 263. Gaisser, S. et al., Parallel pathways for oxidation of 14-membered polyketide macrolactones in Saccharopolyspora erythraea, Mol. Microbiol. 44, 771, 2002. 264. Rodriguez, L. et al., Engineering deoxysugar biosynthetic pathways from antibiotic-producing micro- organisms. A tool to produce novel glycosylated bioactive compounds, Chem. Biol. 9, 721, 2002. 265. Tercero, J. A. et al., The biosynthetic pathway of the aminonucleoside antibiotic puromycin, as deduced from the molecular analysis of the pur cluster of Streptomyces alboniger, J. Biol. Chem. 271, 1579, 1996. 266. Tercero, J. A., Espinosa, J. C., and Jimenez, A., Expression of the Streptomyces alboniger pur cluster in Streptomyces lividans is dependent on the bldA-encoded tRNALeu, FEBS Lett. 421, 221, 1998. 267. Sanchez, C. et al., The biosynthetic gene cluster for the antitumor rebeccamycin: characterization and generation of indolocarbazole derivatives, Chem. Biol. 9, 519, 2002. 268. Onaka, H. et al., Characterization of the biosynthetic gene cluster of rebeccamycin from Lechevalieria aerocolonigenes ATCC 39243, Biosci. Biotechnol. Biochem. 67, 127, 2003. 269. Hyun, C. G. et al., The biosynthesis of indolocarbazoles in a heterologous E. coli host, ChemBioChem 4, 114, 2003. 270. Watanabe, K. et al., Understanding the substrate specificity of polyketide synthatse modules by generating hybrid multimodular synthases, J. Biol. Chem. 278, 42020, 2003. 271. Hu, Z. et al., A host-vector system for analysis and manipulation of rifamycin polyketide biosynthesis in Amycolatopsis mediterranei, Microbiology 145, 2335, 1999. 272. Pospiech, A. et al., A new Myxococcus xanthus gene cluster for the biosynthesis of the antibiotic saframycin Mx1 encoding a peptide synthetase, Microbiology 141, 1793, 1995. 273. Pospiech, A., Bietenhader, J., and Schupp, T., Two multifunctional peptide synthetases and an O- methyltransferase are involved in the biosynthesis of the DNA-binding antibiotic and antitumour agent saframycin Mx1 from Myxococcus xanthus, Microbiology 142, 741, 1996. 274. Trefzer, A. et al., Biosynthetic gene cluster of simocyclinone, a natural multihybrid antibiotic, Anti- microb. Agents Chemother. 46, 1174, 2002. 275. Parry, R. J. and Hoyt, J. C., Purification and preliminary characterization of (E)-3-(2,4-dioxo-6-methyl- 5-pyrimidinyl)acrylic acid synthase, an enzyme involved in biosynthesis of the antitumor agent sparsomycin, J. Bacteriol. 179, 1385, 1997. 276. Lazaro, E. et al., Characterization of sparsomycin resistance in Streptomyces sparsogenes, Antimicrob. Agents Chemother. 46, 2914, 2002. 277. Onaka, H. et al., Cloning of the staurosporine biosynthetic gene cluster from Streptomyces sp. TP- A0274 and its heterologous expression in Streptomyces lividans, J. Antibiot. 55, 1063, 2002. 278. Motamedi, H. and Hutchinson, C. R., Cloning and heterologous expression of a gene cluster for the biosynthesis of tetracenomycin C, the anthracycline antitumor antibiotic of Streptomyces glaucescens, Proc. Natl. Acad. Sci. USA 84, 4445, 1987. 279. Bibb, M. J. et al., Analysis of the nucleotide sequence of the Streptomyces glaucescens tcmI genes provides key information about the enzymology of polyketide antibiotic biosynthesis, EMBO J. 8, 2727, 1989. 280. Decker, H., Motamedi, H., and Hutchinson, C. R., Nucleotide sequences and heterologous expression of tcmG and tcmP, biosynthetic genes for tetracenomycin C in Streptomyces glaucescens, J. Bacteriol. 175, 3876, 1993. Copyright © 2005 CRC Press, LLC 281. McDaniel, R. et al., Engineered biosynthesis of novel polyketides, Science 262, 1546, 1993. 282. Zawada, R. J. and Khosla, C., Heterologous expression, purification, reconstitution and kinetic analysis of an extended type II polyketide synthase, Chem. Biol. 6, 607, 1999. 283. Shen, B. et al., Purification and characterization of the acyl carrier protein of the Streptomyces glaucescens tetracenomycin C polyketide synthase, J. Bacteriol. 174, 3818, 1992. 284. Summers, R. G. et al., The tcmVI region of the tetracenomycin C biosynthetic gene cluster of Streptomyces glaucescens encodes the tetracenomycin F1 monooxygenase, tetracenomycin F2 cyclase, and, most likely, a second cyclase, J. Bacteriol. 175, 7571, 1993. 285. Shen, B. and Hutchinson, C. R., Triple hydroxylation of tetracenomycin A2 to tetracenomycin C in Streptomyces glaucescens. Overexpression of the tcmG gene in Streptomyces lividans and character- ization of the tetracenomycin A2 oxygenase, J. Biol. Chem. 269, 30726, 1994. 286. McDaniel, R. et al., Engineered biosynthesis of novel polyketides: influence of a downstream enzyme on the catalytic specificity of a minimal aromatic polyketide synthase, Proc. Natl. Acad. Sci. USA 91, 11542, 1994. 287. Bao, W., Wendt-Pienkowski, E., and Hutchinson, C. R., Reconstitution of the iterative type II polyketide synthase for tetracenomycin F2 biosynthesis, Biochemistry 37, 8132, 1998. 288. Kulowski, K. et al., Functional characterization of the jadI gene as a cyclase forming angucyclinones., J. Am. Chem. Soc. 121, 1786, 1999. 289. Khosla, C. et al., Genetic construction and functional analysis of hybrid polyketide synthases con- taining heterologous acyl carrier proteins, J. Bacteriol. 175, 2197, 1993. 290. Kramer, P. J. et al., Rational design and engineered biosynthesis of a novel 18-carbon aromatic polyketide, J. Am. Chem. Soc. 119, 635, 1997. 291. Fu, H. et al., Engineered biosynthesis of novel polyketides: stereochemical course of two reactions catalyzed by a polyketide synthase, Biochemistry 33, 9321, 1994. 292. Fu, H. et al., Engineered biosynthesis of novel polyketides: regiospecific methylation of an unnatural substrate by the tcmO O-methyltransferase, Biochemistry 35, 6527, 1996. 293. Decker, H. and Haag, S., Cloning and characterization of a polyketide synthase gene from Streptomyces fradiae Tü2717, which carries the genes for biosynthesis of the angucycline antibiotic urdamycin A and a gene probably involved in its oxygenation, J. Bacteriol. 177, 6126, 1995. 294. Hoffmeister, D. et al., The NDP-sugar co-substrate concentration and the enzyme expression level influence the substrate specificity of glycosyltransferases: cloning and characterization of deoxysugar biosynthetic genes of the urdamycin biosynthetic gene cluster, Chem. Biol. 7, 821, 2000. 295. Trefzer, A. et al., Function of glycosyltransferase genes involved in urdamycin A biosynthesis, Chem. Biol. 7, 133, 2000. 296. Faust, B. et al., Two new tailoring enzymes, a glycosyltransferase and an oxygenase, involved in biosynthesis of the angucycline antibiotic urdamycin A in Streptomyces fradiae Tü2717, Microbiology 146 (Pt 1), 147, 2000. 297. Hoffmeister, D. et al., Engineered urdamycin glycosyltransferases are broadened and altered in substrate specificity, Chem. Biol. 9, 287, 2002. 298. Kunzel, E. et al., Inactivation of urdGT2 gene, which encodes a glycosyltransferase responsible for the C-glycosyltransfer of activated D-olivose, leads to the formation of the novel urdamycins I, J, and K., J. Am. Chem. Soc. 121, 11058, 1999. 299. Rix, U. et al., Urdamycin L: a novel metabolic shunt product that provides evidence for the role of the urdM gene in the urdamycin A biosynthetic pathway of Streptomyces fradiae Tü 2717, ChemBio- Chem 4, 109, 2003. 300. Garg, R. P. et al., Molecular characterization and analysis of the biosynthetic gene cluster for the azoxy antibiotic valanimycin, Mol. Microbiol. 46, 505, 2002.

301. Ogasawara, Y. et al., Cloning, sequencing, and functional analysis of the biosynthetic gene cluster of macrolactam antibiotic vicenistatin in Streptomyces halstedii, Chem. Biol. 11, 79, 2004. 302. Clardy, J. and Walsh, C. T., Lessons from natural molecules, Nature, 432, 7–15, 2004. 303. Galm, U. et al., Antitumor antibiotics: Bleomycin, enediynes, and mitomycin, Chem. Rev., 105, 739–758, 2005. 304. Chen, J. and Stubbe, J., Bleomycins: Towards better therapeutics, Nat. Rev. Cancer, 5, 102–112, 2005. 305. Sanchez, C. et al., Combinatorial biosynthesis of antitumor indolocarbazole compounds, Proc. Natl. Acad. Sci. USA, 102, 461–466,2005. Copyright © 2005 CRC Press, LLC 26 Developments and Future Trends in Anticancer Natural Products Drug Discovery* David J. Newman and Gordon M. Cragg CONTENTS I. Introduction II. Plant Sources A. Combretastatins B. Flavopiridol C. Olomucine, Roscovitine, and the Purvalanols D. Benzopyrans and Privileged Structures E. Microbial Associates and Endophytic Fungi F. Old Molecules, New Uses 1. Bruceantin 2. Triterpenoid Acids 3. β-Lapachone III. Marine Sources A. Microbial Associates 1. Dolastatins 2. Ecteinascidins 3. Jaspamide/Jasplakinolide 4. Geodiamolides and Chondramides 5. Pederins, Mycalamides, and Related Structures 6. Manzamines B. Independent Microbes 1. Salinosporamide C. Marine Invertebrates 1. NVP-LAQ824 and Psammaplin IV. Terrestrial Microbes A. Leptomycins B. Palmarumycin CP1 V. Plant–Microbe–Marine Interface A. Indigo and the Indirubins VI. Conclusion References * Not including those products discussed in other chapters in this volume. Copyright © 2005 CRC Press, LLC I. INTRODUCTION Important, naturally derived chemotypes currently in clinical use or development as anticancer agents have been discussed in earlier chapters, and the purpose of this chapter is to provide an overview — by no means exhaustive — of developments and future trends in the generation of novel agents from nature. For reference, a list of approved antitumor agents (as of the end of 2002) is shown in Table 26.1, using the data from Newman et al.1 The organization of the chapter is according to the putative sources of agents where these are known or assumed, but the chapter ends with a discussion, together with examples, of what appears to be an evolving trend about the actual source organisms. As more is learned about the genetics of a “producing organism,” frequently it is realized that the actual organism from which the original material was isolated and identified is either a host for microbial biosynthetic machinery or may perform a tailoring function on materials produced by a microbe. II. PLANT SOURCES On the basis of recent experience using the National Cancer Institute (NCI) 60–cell line screen, the potential for the discovery of a new cytotoxic chemotype from plants, comparable to the taxanes or camptothecins, appears to be relatively low. In spite of work covering vascular plants from extensive collections in the tropics over the last 18+ yrs, no novel chemotypes have, as yet, been reported. Nevertheless, a number of variations on known chemotypes or their sources have been reported, including camptothecin derivatives, materials that resemble bleomycin in their activities at the molecular level, and compounds, such as the jatrophane esters, that have now been shown to have paclitaxel-like activities. A. COMBRETASTATINS What is evident however, is that at least one of the classes of natural products currently in clinical trials, the combretastatins, either as the original compound or as its phosphate ester (combretastatin A4 phosphate), has generated a significant amount of interest as a basic chemotype (Chapter 3). At present, there are multiple variations on the CA4 basic structure (trimethoxy-stilbene) either in clinical trials or entering various stages of preclinical development as a drug candidate. The compounds range from very close chemical analogs to topologically equivalent molecules from a receptor perspective, but significantly different when viewed in two dimensions, though resembling the basic molecular structure of CA4 when carefully examined (1–7). The structures of a significant number of these natural product mimics are given in the recent review of Cragg and Newman, which should be consulted by the interested reader.2 B. FLAVOPIRIDOL Approximately contemporaneous with the initial work on CA4, Flavopiridol (8) was prepared by the Indian subsidiary of Hoechst (now Sanofi-Aventis) following the isolation and synthesis of the plant-derived natural product, rohitukine (9). It was the most active of approximately 100 analogs made when assayed against cyclin-dependent kinases (CDKs) and showed about 100-fold more selectivity compared with its activity versus tyrosine kinases. Although it showed roughly compa- rable activity (IC50 100–400 nM) against CDKs, it turned out to be the first compound at NCI identified as a potential antitumor agent that subsequently was proven to be a relatively specific CDK inhibitor.3 The initial report4 on its CDK2 inhibitory activity was made in 1994, followed by data demonstrating antitumor activity, provided by Czech et al.,5 in 1995. Flavopiridol is currently in phase III clinical trials as an inhibitor of cyclin-dependent kinase 2 (CDK2), both as a single agent and as a modulator in combination with other agents, particularly paclitaxel and cisplatinum. It has been reported to lead to partial or complete remissions in a number of phase I patients, leading Copyright © 2005 CRC Press, LLC TABLE 26.1 All Anticancer Drugs (1940s–2002) Generic Name Year Sourcea OCT-43 1999 B Alemtuzumab 2001 B Celmoleukin 1992 B Denileukin diftitox 1999 B Interferon alfa2a 1986 B Interferon, gamma-1a 1992 B Interleukin-2 1989 B Pegaspargase 1994 B Rituximab 1997 B Tasonermin 1999 B Teceleukin 1992 B Trastuzumab 1998 B BEC 1989 N Aclarubicin 1981 N Actinomycin D Pre-1981 N Angiotensin II 1994 N Arglabin 1999 N Asparaginase Pre-1981 N Bleomycin Pre-1981 N Daunomycin Pre-1981 N Doxorubicin Pre-1981 N Masoprocol 1992 N Mithramycin Pre-1981 N Mitomycin C Pre-1981 N Paclitaxel 1993 N Pentostatin 1992 N Peplomycin 1981 N Solamargine 1987 N Streptozocin Pre-1981 N Testosterone Pre-1981 N Vinblastine Pre-1981 N Vincristine Pre-1981 N Alitretinoin 1999 ND Amrubicin HCl 2002 ND Cladribine 1993 ND Cytarabine ocfosfate 1993 ND Docetaxel 1995 ND Dromostanolone Pre-1981 ND Elliptinium acetate 1983 ND Epirubicin HCI 1984 ND Estramustine Pre-1981 ND Ethinyl estradiol Pre-1981 ND Etoposide Pre-1981 ND Etoposide phosphateb 1996 ND Exemestane 1999 ND Fluoxymesterone Pre-1981 ND Formestane 1993 ND Fulvestrant 2002 ND Gemtuzumab ozogamicin 2000 ND Hydroxyprogesterone Pre-1981 ND Copyright © 2005 CRC Press, LLC TABLE 26.1 (Continued) All Anticancer Drugs (1940s–2002) Generic Name Year Sourcea Idarubicin hydrochloride 1990 ND Irinotecan hydrochloride 1994 ND Medroxyprogesterone acetate Pre-1981 ND Megesterol acetate Pre-1981 ND Methylprednisolone Pre-1981 ND Methyltestosterone Pre-1981 ND Miltefosine 1993 ND Mitobronitol Pre-1981 ND Pirarubicin 1988 ND Prednisolone Pre-1981 ND Prednisone Pre-1981 ND Teniposide Pre-1981 ND Testolactone Pre-1981 ND Topotecan HCl 1996 ND Triamcinolone Pre-1981 ND Triptorelin 1986 ND Valrubicin 1999 ND Vindesine Pre-1981 ND Vinorelbine 1989 ND Zinostatin stimalamer 1994 ND Aminoglutethimide 1981 S Amsacrine 1987 S Arsenic trioxide 2000 S Bisantrene hydrochloride 1990 S Busulfan Pre-1981 S Camostat mesylate 1985 S Carboplatin 1986 S Carmustine Pre-1981 S Chlorambucil Pre-1981 S Chlortrianisene Pre-1981 S Cis-diamminedichloroplatinum Pre-1981 S Cyclophosphamide Pre-1981 S Dacarbazine Pre-1981 S Diethylstilbestrol Pre-1981 S Flutamide 1983 S Fotemustine 1989 S Heptaplatin /SK-2053R 1999 S Hexamethylmelamine Pre-1981 S Hydroxyurea Pre-1981 S Ifosfamide Pre-1981 S Levamisole Pre-1981 S Lobaplatin 1998 S Lomustine Pre-1981 S Lonidamine 1987 S Mechlorethanamine Pre-1981 S Melphalan Pre-1981 S Mitotane Pre-1981 S Copyright © 2005 CRC Press, LLC TABLE 26.1 (Continued) All Anticancer Drugs (1940s–2002) Generic Name Year Sourcea Mustine hydrochloride Pre-1981 S Nedaplatin 1995 S Nilutamide 1987 S Nimustine hydrochloride Pre-1981 S Oxaliplatin 1996 S Pipobroman Pre-1981 S Porfimer sodium 1993 S Procarbazine Pre-1981 S Ranimustine 1987 S Sobuzoxane 1994 S Thiotepa Pre-1981 S Triethylenemelamine Pre-1981 S Uracil mustard Pre-1981 S Zoledronic acid 2000 S Aminogluethimide Pre-1981 S* Capecitabine 1998 S* Carmofur 1981 S* Cytosine arabinoside Pre-1981 S* Doxifluridine 1987 S* Enocitabine 1983 S* Floxuridine Pre-1981 S* Fludarabine phosphate 1991 S* Fluorouracil Pre-1981 S* Gemcitabine HCl 1995 S* Goserelin acetate Pre-1981 S* Leuprolide Pre-1981 S* Mercaptopurine Pre-1981 S* Methotrexate Pre-1981 S* Mitoxantrone HCI 1984 S* Tamoxifen Pre-1981 S* Thioguanine Pre-1981 S* Bexarotene 2000 S*/NM Raltitrexed 1996 S*/NM Temozolomide 1999 S*/NM Anastrozole 1995 S/NM Bicalutamide 1995 S/NM Camostat mesylate 1985 S/NM Fadrozole HCl 1995 S/NM Gefitinib 2002 S/NM Imatinib mesilate 2001 S/NM Letrazole 1996 S/NM Toremifene 1989 S/NM Bcg live 1990 V Melanoma theraccine 2001 V a Definitions as Newman et al.1 b Prodrug (not counted). Copyright © 2005 CRC Press, LLC to phase II studies in patients with paclitaxel-resistant tumors. There have been a number of relatively recent reports of this agent and the various combinations with other drugs and drug candidates — a significant number of which are either natural products or are derived from natural products.6–8 C. OLOMUCINE, ROSCOVITINE, AND THE PURVALANOLS Continuing on the theme of plant-derived products being leads to cell cycle modulators, Meijer, one of the early researchers in pharmacologic intervention in the cell cycle,9 demonstrated, using the simple starfish oocyte model of cell division, that substituted purines, particularly 6-dimethy- laminopurine and isopentenyladenine (from Castanea sp.) were found to inhibit the mitotic histone H1 kinase, which is now better known as CDK1/cyclin B, at the 50–100 µM level in vitro. On searching for other purine-derived compounds, the group tested a plant secondary metabolite, subsequently named olomucine (10), that was originally isolated from the cotyledons of the radish but had been synthesized in Australia in 1986 by Parker et al.10 This material disproved the current dogma about adenosine triphosphate binding sites because it demonstrated an improved efficacy (IC50 of 7 µM) and selectivity for CDKs and, to some extent, MAP kinases by direct competition with adenosine triphosphate. Further development of this series led to roscovitine (11), and then following synthesis of a very focused library via combinatorial chemistry techniques, the purval- anols, which were even more potent, with IC50 values in the 4–40-nM range (compared with 450 nM for roscovitine).11 The (R)-isomer of roscovitine was selected for clinical development under the code CYC202 under license to Cyclacel in Dundee, Scotland, and currently is in phase II clinical trials in Europe. As with other signal transduction inhibitors in clinical trials, sequential treatment with cytotoxins is also being used or considered. Though some beneficial effects are seen with the signal transduc- tion inhibitors, complete or partial responses tend to only be demonstrated when sequential treat- ments of signal transduction inhibitor/cytotoxin are used. D. BENZOPYRANS AND PRIVILEGED STRUCTURES This type of chemical modification of an existing basic “privileged structure” is documented in the combinatorial chemistry literature, particularly in the case of the formally plant-derived base structure, the benzopyran. The term, “privileged structures,” was originally introduced by Evans et al.12 for benzazepines and benzdiazepines because of their ability to bind to a multitude of unrelated classes of peptide receptors with high affinity. The term was later extended by Mason et al.13 to other structural classes such as benzamidines, biphenyltetrazoles, spiropiperidines, indoles, and benzylpiperidines and was reviewed by Patchett and Nargund in 2000.14 In a series of elegant papers, Nicolaou et al.15–17 dramatically extended the concept to include benzopyrans, as a result of the researchers’ hypothesis: “We were particularly intrigued by the possibility that using scaffolds of natural origin, which presumably have undergone evolutionary selection over time, might confer favorable bioactivities and bioavailabilities to library members.”15 A search of the literature showed almost 4000 2,2-dimethyl-2H-benzopyran moieties (12); a slight structural modification whereby the pyran ring is reduced (13) identified another 8000 structures. Examples of the multiplicity of bioactive derivatives of this general structure are given in detail in Figure 2 of the first paper in the series,15 which should be consulted by the interested reader. Using a technique of solid-phase combinatorial chemistry that used a selenenyl bromide sub- stituted polystyrene resin,18 the group was able to generate what were effectively “libraries from libraries” by using a massively parallel synthetic technique on the pyrans, which were produced from the seleno-linked libraries by epoxidation of the pyran double bond, followed by the opening of the epoxide by nucelophilic attack. These libraries, and subsequent derivatives based on the benzopyrans, were then tested against a variety of pharmacologic screens, and the leads were then reoptimized, using the screening data. Copyright © 2005 CRC Press, LLC To date (mid-2004), four publications have come from these initial libraries, with activities in the inhibition of NADH:ubiquinone oxidoreductase giving rise to benzopyrans with IC50 values in the 18–55-nM range and having predominately cytostatic effects on tumor lines, thus indicating a possible potential in chemoprevention. The interested reader should consult Figure 4 and Figure 5 in the relevant publication for more information on the actual

structures and activities in both types of assay.19 This report, relevant to antitumor studies, was then followed by others in different disease states covering antibiotics, energy metabolism, and genetic dis- section of a cell pathway.20–22 E. MICROBIAL ASSOCIATES AND ENDOPHYTIC FUNGI As mentioned in the introduction, there is now a realization that materials that were often thought to be from — in this case — plants are actually produced in large part by another organism, and the plant host may only perform some final tailoring of the molecule. Such is the case with maytansine since, as reported by Floss (Chapter 17), the actual source of the base structure is a bacterium, with perhaps the modification leading to maytansine being performed by the plant itself. There are other examples given later in the chapter that continue on this theme. In addition to the above example of what may be a simple transfer of material to the host from a microbe, there are some extremely interesting examples in the recent review by Strobel et al.23 of materials that are produced by plant epiphytic microbes, predominately from the fungi. Among the organisms reported are a series of fungal epiphytes from the well-known genus Pestalotiopsis and a report that Pestalotiopsis micromonospora, a very common epiphytic fungus, produces paclitaxel at low levels. This followed on from the initial report, which was greeted with a fair amount of skepticism, that the novel fungus Taxomyces andreanae, isolated from Taxus brevifolia, produced paclitaxel at very low levels. Since the original report in 1993, many further examples of paclitaxel production at low levels have been reported in the scientific literature, not just from Taxus spp., but from trees of many other genera. These findings have led Strobel to suggest that paclitaxel production by epiphytic fungi is a protective mechanism for the plant host against plant pathogens, such as Pythium spp. and Phytophora spp., as paclitaxel is known to be very active as a fungicide against these particular pathogens. F. OLD MOLECULES, NEW USES 1. Bruceantin An example of an “old” drug of the same vintage as Taxol® and camptothecin, and having a possibility of revival, is bruceantin (14), which was first isolated from a tree, Brucea antidysenterica, used in Ethiopia for the treatment of “cancer.” Activity was observed in animal models bearing a range of tumors, but no objective responses were observed in clinical trials, and further development was terminated. Interest has been revived by the observation of significant activity against panels of leukemia, lymphoma, and myeloma cell lines, as well as in animal models bearing early and advanced stages of the same cancers. This activity has been associated with the downregulation of a key oncoprotein (c-myc), and these data are being presented as strong evidence supporting the development of bruceantin as an agent for the treatment of hematological malignancies.24 2. Triterpenoid Acids Triterpenoid acids such as oleanolic (15) and ursolic acid, which are common plant constituents, are known to possess weak antiinflammatory and antitumor activities. Attempts to synthesize new analogs having increased potencies have led to the synthesis of 2-cyano-3,12-dioxoolean-1,9-dien- 28-oic acid (16) and its methyl ester, which have potent in vitro and in vivo antitumor activity against a wide range of tumors, including breast carcinomas, leukemias, and pancreatic carcinomas. Copyright © 2005 CRC Press, LLC Of particular interest is the significant activity of 2-cyano-3,12-dioxoolean-1,9-dien-28-oic acid against epithelial ovarian carcinoma cell lines, including lines that were resistant to clinically used agents such as cisplatin. Because epithelial ovarian carcinoma is the leading cause of death from gynecologic cancers, further evaluation of 2-cyano-3,12-dioxoolean-1,9-dien-28-oic acid in the treatment of these cancers is being pursued.25 3. β-Lapachone The relatively simple naphthoquinone β-lapachol (17) is a well-known compound obtained from the bark of the lapacho tree, Tabebuia avellanedae, and other species of the same genus native to South America, and it, together with other components of the plant, have extensive usage as an ethnobotanical treatment in the Amazonian region. Although advanced to clinical status by the NCI in the 1970s, this compound showed unacceptable levels of toxicity and was dropped. Recently, however, a close relative, β-lapachone (18), has shown very interesting molecular target activity, with induction of apoptosis in transformed cells being one of the mechanisms of action. What is of prime import, however, is that one of the possible routes of apoptotic induction is via the transcription factor E2F126 and appears to only function in transformed cells. In contrast, no comparable effect was shown in nontransformed cells at comparable concentrations of the agent. Evidence of further involvement in the transcription processes was the recent demonstration by Choi et al.27 that this agent induced activation of caspase-3, inhibition of NF-κB, and subsequent downregulation of bcl-2. At present, β-lapachone is in phase I clinical trials in the United States under the aegis of ArQule, who licensed it from the originators, Cyclis Pharmaceuticals. Further information on the background of these agents is given in the review by Ravelo et al.28 III. MARINE SOURCES A. MICROBIAL ASSOCIATES Perhaps it is in this area that the interplay of microbes and a host invertebrate/alga is becoming a significant part of the overall equation. For many years, it has been speculated that a number of potent marine-derived molecules that were active as antitumor agents might be the product of interactions between the organism from which the agent was isolated and associated microbes. 1. Dolastatins The driving force behind such suppositions was the number of reports in the marine natural product literature over the last 10–15 yr, where the structures of compounds isolated from an invertebrate or, in an increasing number of cases, from multiple different invertebrate phyla in distinct geo- graphical areas of the oceans were similar to each other or to metabolites from terrestrial bacteria. Some relevant examples are the dolastatins from Dolabella auricularia and simplostatins, and now dolastatins, from the cyanophyte, Simploca sp., on which the nudibranch feeds (Chapter 11). 2. Ecteinascidins When the structure of Et743 was established by the groups of Rinehart and Wright in 1990 (Chapter 12), it became obvious that this compound had a base structure29 that was well known from microbial sources on land, from invertebrates in the ocean, and from a marine bacterium. Thus, Et743 resembled the naphyridinomycins and saframycins from terrestrial Streptomycetes; the renieramycins from oceanic sponges, with the latest variation, renieramycin J being recently reported by Oku et al.30; jorumycin from the nudibranch, Jorunna funebris31; and cyanosafracin B from a marine-derived Pseudomonas fluorescens. The ecteinascidins all contain a fourth Copyright © 2005 CRC Press, LLC STRUCTURES 1–19 Copyright © 2005 CRC Press, LLC exocyclic isoquinoline group when compared to their putative siblings, together with a bridging cysteinyl-derived grouping. At present, there is work ongoing on identifying the endosymbionts of Ecteinascidia turbinata, with the ultimate aim of identifying, cloning, and expressing the putative Et743 biosynthetic genes if the potential producing organism, known currently as Candidatus Endoecteinascidia frumenten- sis, is not amenable to cultivation.32 Further examples of the potential involvement of microbes in the production of sponge-isolated secondary metabolites with antitumor potential (either directly or as a warhead on a delivery system) include, but are not limited to, the following. 3. Jaspamide/Jasplakinolide Jaspamide (19) is the major metabolite33 of the Pacific sponge, Dorypleres splendens (also identified as a Jaspis sp.), and was also reported as jasplakinolide when it was isolated from a Fijian-derived sample of Jaspis johnstoni.34 In addition, modified jaspamides were also reported from the sponge Jaspis splendens collected in Vanuatu.35 Jaspamide was initially reported to be a potent insecticide and antifungal agent,33,34 but subsequent work by NCI scientists demonstrated that it induces actin polymerization in vitro, leading to disruption of the actin skeleton.36,37 However, in spite of many attempts at NCI, no reproducible in vivo antitumor activity could be demonstrated because the therapeutic index was very close to unity. In this respect, it is similar to the effects seen with other actin inhibitors such as the cucurbitacins, latrunculins, and cytochalasins. However, with a suitable delivery system, these molecules may well be of utility. 4. Geodiamolides and Chondramides The possible involvement of microbes as a source of these compounds was raised by the reports of the structures of the closely related geodiamolides,38 from a Geodia sp. in 1987 and from other taxonomically distant sponges in following years.39 The potential for microbial involvement was reinforced by the reports from the Reichenbach group in 1995, of the isolation and purification of similar molecules, the chondramides (20), with an 18-membered macrolide ring from a terrestrial myxobacterium.40 Analogous to the jaspamides, their first reported activities were as antifungal compounds, but subsequent investigation demonstrated their actin-stabilizing activity.41 5. Pederins, Mycalamides, and Related Structures An early entrant to the potential for microbial involvement was the recognition in the late 1980s that the highly cytotoxic sponge metabolites known as the mycalamides42 and subsequently other derivatives of the basic structure, the onnamides, theopederins, and the ring-opened form known as irciniastatin A (also reported later as psymberin),43,44 all contained the basic ring structure of the toxin from the Paederus beetle, pederin (21). These molecules (which currently number over 3044) are very interesting from the perspective of potential antitumor drug leads, but development has been bedeviled by sourcing problems, even though a formal synthesis of mycalamide A (22) was published in 2004 by Trost et al.45 Recently, however, work by Kellner46 on the source of pederin from the Paederus sabaeus beetle from Turkey demonstrated that the actual source was a putative pseudomonad. This work was extended from a molecular genetic aspect by Piel et al.,47 demonstrating a symbiosis island in the microbe that indicates the potential for horizontal gene transfer to other organisms and, most important, the presence in the microbe of a PKS-PS cluster that formally codes for the biosynthetic pathway for pederin.48 If this organism can either be cultured or the cluster expressed in another host, then the sourcing problems of this class of molecules may well be on the way to being solved, allowing further investigation of their potential. Copyright © 2005 CRC Press, LLC 6. Manzamines Although one could always make a circumstantial case that a microbe or microbes were involved in the biosynthesis of either the agents listed above or of precursors that might then be further host- modified (analogous to maytansine; see Section IIE above), it was not until 2003 that there was actual hard evidence for production of the same compound by an associated microbe and by the host invertebrate. An Indonesian sponge, collected from approximately 100 m depth, using rebreather collection techniques, produced a series of antitumor compounds known as the man- zamines; on culturing the associated microbes, a Micromonospora sp. was purified, which on fermentation in selected media, produced the base molecule, manzamine A, and a derivative, 8- hydroxymanzamine A, both of which were isolable from the host sponge. Thus, for the first time, the thesis mentioned above was proven and the potential for production of this agent by fermentation was demonstrated.49,50 On the basis of these observations and data, it is probably correct to assume that each of these series of compounds has a common genetic ancestor (a microbe). Thus, it behooves natural products chemists and biologists to look beyond the “nominal sources” and to actually track down, as far as is possible, the actual progenitor. If a series of compounds is going to be useful as a drug entity, or lead thereto, fermentation is still the easiest production technique on any significant scale above a small laboratory preparation level. B. INDEPENDENT MICROBES Although pioneering work was done in the middle of the last century in isolating and culturing marine microbes from a microbiological aspect, it was not until the 1980s that significant time and effort was devoted to the isolation, identification, and fermentation of marine microbes (or microbes from marine sources) in order to evaluate their potential to produce secondary metabolites with antitumor activity that would be comparable to those known since the 1940s from their terrestrial equivalents. 1. Salinosporamide In retrospect, what probably had the most effect on the isolation techniques was the realization that the media used for work with terrestrial microbes was much too rich in carbon for their marine- based cousins. Once a series of suitable media and storage/recovery techniques was established, in large part because of the work of Fenical and Jensen

and their colleagues at The Scripps Institute of Oceanography, the field opened up, with reports of active and chemically unique molecules being published at frequent intervals. What is of prime import is that as a result of their systematic searching for marine bacteria, the Fenical group recently described a totally new genus51 within the order Actinomycetales, for which they have proposed the genus name of Salinospora.52 This bacterium, on fermentation in suitable media in seawater, produced the omuralide (23) derivative salinosporamide A (24). Omuralide (23) is the spontaneous reaction product from the terrestrial Streptomycete metabolite, lactacystin (25), first reported in 1991 by Omura et al.53 and synthesized, together with analogs, by Corey’s group.54 Subsequently, lactacystin (actually omu- ralide) was found to be a specific inhibitor of the 20S proteosome, modifying a subunit-specific threonine.55 That such an inhibitor has drug potential is demonstrated by the now approved antitumor agent, Velcade®, a synthetic peptide boronate, so the search for other molecules with similar activities is worthwhile. Earlier work on natural products and synthetic agents has been covered by Kisselev and Goldberg.56 Salinosporamide itself, which differs from lactacystin/omuralide in terms of its substitution pattern around the pyrrolid-2-one ring, is also a potent proteosome inhibitor and is currently under preclinical development by Nereus Pharmaceuticals in San Diego, California, with a recent publi- cation from Corey’s group giving details of a simple synthetic route to the compound.57 Copyright © 2005 CRC Press, LLC TABLE 26.2 Status of Marine-Derived Natural Products in Clinical and Preclinical Trials (Not discussed in other chapters) Name Source Status (Disease) Comment ES-285 (Spisulosine) Spisula polynyma Phase I (Cancer) Rho-GTP inhibitor KRN-7000 Agelas mauritianus Phase I (Cancer) An Agelasphin derivative Squalamine Squalus acanthias Phase II (Cancer) Antiangiogenic activity as well NVP-LAQ824 Synthetic Phase I (Cancer) Derived from Psammaplin, Trichostatin, and Trapoxin structures Laulimalide Cacospongia mycofijiensis Preclinical (Cancer) Synthesized by a variety of investigators Vitilevuamide Didemnum cucliferum and Preclinical (Cancer) Polysyncraton lithostrotum Diazonamide Diazona angulata Preclinical (Cancer) Synthesized and new structure elucidated. Eleutherobin Eleutherobia sp Preclinical (Cancer) Synthesized and derivatives made by combi-chem; can be produced by aquaculture Sarcodictyin Sarcodictyon roseum Preclinical (Cancer) Combi-chem synthesis performed around (derivatives) structure Peloruside A Mycale hentscheli Preclinical (cancer) Salicylihalimides A Haliclona sp Preclinical (cancer) First marine Vo-ATPase inhibitor. Similar materials from microbes, synthesized Thiocoraline Micromonospora marina Preclinical (Cancer) DNA polymerase inhibitor Ascididemnin Preclinical (Cancer) Reductive DNA-cleaving agents Variolins Kirkpatrickia variolosa Preclinical (Cancer) CDK inhibitors C. MARINE INVERTEBRATES At present, there is a considerable number of formally marine invertebrate–derived molecules that are in preclinical studies that have not been covered in other chapters in this book. Table 26.2 lists a selection of those that are in preclinical development, and we will comment on one of them at a greater length below. The citations to the compounds in Table 26.2, together with descriptions of their sources and comments on their activities as antitumor agents, and so forth, are given in detail in the 2004 review by Newman and Cragg, which should be consulted for further information.58 1. NVP-LAQ824 and Psammaplin One compound listed in Table 26.2 is in fact a totally synthetic molecule but was derived from information from natural products, from both a marine invertebrate and from terrestrial microbes. This is the Novartis agent known currently as NVP-LAQ824. The basic structure (26) was derived from work with both natural products and synthetic derivatives. The marine natural product, psammaplin A (27), which had originally been identified in 1987 by the groups of Schmitz59 at the University of Oklahoma and Crews60 at the University of California, Santa Cruz, was screened, together with congeners and with the microbial products trapoxin B (28) and trichostatin A (29), for their activity as histone deacetylase inhibitors by Novartis (then Ciba-Giegy) as part of a National Cooperative Drug Discovery Group program. The psammaplins were found to be extremely potent histone deacetylase inhibitors. The synthetic path from the psammaplin, trapoxin, and trichostatin structures to the compound now known as NVP-LAQ824 was described in three papers.61–63 These papers, and in particular the review by Remiszewski,63 should be consulted for the chemical rationales that led from these natural products to the current clinical candidate. This agent demonstrated significant in vitro activity Copyright © 2005 CRC Press, LLC against cells derived from human multiple myeloma patients, inducing apoptotic signaling and also exhibiting histone deacetylase inhibition. At present, the material is in a phase I clinical trial against hematologic malignancies64 at the Dana-Farber Cancer Institute. IV. TERRESTRIAL MICROBES At this time, most agents in advanced preclinical development as direct antitumor agents are not from terrestrial microbes, but what is becoming more important is the use of older molecules as the basis for treatments with newer agents, as well as the use of the old molecule as a lead structure for combinatorial chemistry (see Section IIF above). We will give two examples, one from each category. A. LEPTOMYCINS Reports have been coming out in the biological literature of the utility of the well-known microbial products, the leptomycins, and the effect that they have on tumor cells when combined with specific molecular target inhibitors. The effect of the fungal metabolite, leptomycin B (30), on the cell cycle was first described by Yoshida et al.,65 when they demonstrated that this compound induced elon- gation of the cells of the fission yeast Schizosaccharomyces pombe and also stopped proliferation of rat 3Y1 cells at both G1 and G2. Removal of the leptomycin then produced tetraploid cells that could proliferate without passage through the M phase. Subsequent probing of the genetic machinery of a leptomycin-resistant strain of S. pombe demonstrated that the probable target of leptomycin in this organism and, by inference, in human cells was located in the regulatory cascade of crm1, a gene that was required for maintenance of higher-order chromosome structures, and hence cell growth in the yeast.66 Confirmation of crm1 as the target of leptomycin B in both the yeast and in human cells was provided by the subsequent identification of a cysteine residue (Cys-529) in this protein, as the site of covalent binding resulting from Michael addition of the cysteinyl–SH group across the α, β-unsaturated δ-lactone that is common to all of the leptomycin-like molecules. This led to a total inhibition of its function as part of a trimeric complex with RanGTP and a protein containing a nuclear export sequence whose function was to move proteins out of the nucleus into the cytoplasm.67–69 In 2001, Vigneri and Wang70 reported that leptomycin B, in conjunction with the synthetic PTK inhibitor STI571 (Gleevec®), stopped egress of the oncoprotein bcr-abl from the nucleus of CML cells and induced apoptosis. Using target mutants, the researchers confirmed that the effect was caused by inhibition of crm1, which stopped the export of the bcr-abl/RanGTP/protein complex and led to the sequestration of the bcr-abl within the nucleus and induction of apoptosis. This finding had a potential clinical significance, as a fair proportion of patients treated with Gleevec® become refractory in a relatively short period of time. Work on this approach is ongoing, with over 80 citations to the initial report in the literature. An updated review by Yoshida’s group on the potential for nuclear transport inhibitors was recently published, covering leptomycins and other similar agents.71 B. PALMARUMYCIN CP1 Palmarumycin CP1 (31) from the terrestrial fungus Coniothyrium, first reported by Krohn et al.72 in 1994, was used by Wipf at the University of Pittsburgh as a base structure to synthesize a variety of compounds containing the base naphthoquinone–spiroketal structure. This study was part of a search for compounds that inhibited the thioredoxin–thioredoxin reductase system at similar levels to the activity demonstrated by the fungal metabolite, pleurotin (32), whose in vitro and in vivo activities in this and in related pathways were recently reported by Welsh et al.73 In a series of Copyright © 2005 CRC Press, LLC recent papers, Wipf and his colleagues have presented evidence demonstrating that a simple modification of the base CP1 structure gives compounds S11 and S12 (33, 34) with excellent enzyme inhibitory activity and in vitro cytotoxicities, comparable to those of the natural products.74–76 V. PLANT–MICROBE–MARINE INTERFACE Finally, there is one class of compounds that has recently come into play as molecular target inhibitors that have a very long provenance as chemical entities — in some cases as mixtures and in others as single compounds — and whose listed activities range from dyestuffs, through a traditional Chinese medical treatment for leukemia, to very specific inhibition of an essential cellular enzyme cascade. A. INDIGO AND THE INDIRUBINS The compounds in question are all based on the simple heterocycle, indole. The essential aminoacid tryptophan is formed from the chorismate pathway, via cyclization of anthranilate, and indole is formed from tryptophan catabolism. If indole is hydroxylated in the 3 position, presumably by a suitable cytochrome P450, then it is tautomeric with the 3 keto analog, indoxyl. Various levels of oxidation then lead to the mixture of indigo, indirubin, and their isomers that is commonly used as the source of indigo dyestuffs for either clothing or as war paint (the ancient Celtic “woad” being a mixture from the plant Isatis tinctora77). Although usually thought of as being plant products, indigo and the indirubins have been reported from four nominally independent sources: a variety of plants,77 a number of marine mollusks (usually belonging to the Muricidae family of gastropods),78 natural or recombinant bacteria,79 and human urine.80 It is highly probable, however, that in all cases, the indirubins and indigo are the terminal oxidation products of tryptophan/indole catabolism — perhaps even a method of removing excess indole from the organism. However, irrespective of the reason for the production, the indirubins have been identified as the major active ingredient of the traditional Chinese medical recipe known as Danggui Longhui Wan, used to treat chronic myelogenous leukemia81,82 in China for many years. What is of import from both a natural product and a pharmacological perspective is the recognition by Meijer’s group that the indirubins as a class were both inhibitors of several CDKs and also potent inhibitors of glycogen synthase kinase-3 (GSK-3). Thus, in a recent paper, Meijer et al.83 described the early work on this enzyme and then demonstrated that simple indirubins, including a brominated version (35) that had never been isolated from natural sources before their work with the mollusk Hexaplex trunculus and its chemically modified oxime derivative (36), were potent and selective inhibitors of both variants of GSK-3, with IC50 values of 22 and 5 nM, respectively. There was at least a fivefold specificity versus CDK1/Cyclin B or CDK/p25 and significantly more specificity against a wide range of other kinases. A slightly later paper from the same group gave full details of the chemistry involved and established structure–activity relationships using x-ray crystallography and molecular model- ing techniques.84 By using immobilized indirubins, Meijer’s group was able to conclusively identify GSK-3 as the target of these molecules in cell lyzates, to demonstrate that they had a significant effect in vivo on Xenopus development — consonant with the suggested mechanism, and to derive x-ray crystallographic evidence for the binding of the agents at the active site (adenosine triphosphate binding site) of GSK-3. They also demonstrated that the activation of the enzyme via phosphory- lation by an as-yet-unidentified kinase on two specific tyrosine residues (GSK-3 is a Ser/Thr kinase) is blocked when indirubins are bound, perhaps because of a conformational shift in the bound versus free protein. Copyright © 2005 CRC Press, LLC STRUCTURES 20–36 Copyright © 2005 CRC Press, LLC These papers were then followed by another from the same group in which they demonstrated that the indirubins actually had yet another, independent action at the cell receptor level. By using the same basic suite of compounds, they demonstrated that they serve as ligands for the “orphan receptor” known as the aryl hydrocarbon receptor (AhR). As yet, no natural ligands have been identified for AhR, even though, contrary to earlier beliefs, it has existed for over 450 million years (though there have been proposals that indole-containing compounds are among the natural ligands85). From a series of elegant biochemical and biological studies, including the use of cells that were null–null mutants for the AhR, the researchers were able to differentiate between the cytostatic effects of some indirubins following their activation of the AhR from kinase inhibition, which appears to be

the main mechanism underlying the cytotoxicity of these molecules.86 VI. CONCLUSION Although one generally thinks of the organism from which a compound is isolated and purified as being the source of the natural product, it is becoming more and more difficult to actually state that this is so. From the information presented in the case of plants and marine invertebrates, it can be seen that in many instances, microbes (perhaps from all three kingdoms) are intimately involved at some stage in the production of those secondary metabolites that are being evaluated for their potential as drug leads or entities. We deliberately did not give any details of the work now ongoing in studies of the metagenomes87 of either soils or marine sediments, because, as yet, no compounds have been reported in the open literature that are acting as leads to new antitumor agents. However, because the microbial diversity is many orders of magnitude greater than any other source of biological diversity, such agents will be forthcoming. REFERENCES 1. Newman, D. J., Cragg, G. M., and Snader, K. M., Natural products as sources of new drugs over the period 1981-2002, J. Nat. Prod., 66, 1022, 2003. 2. Cragg, G. M. and Newman, D. J., A tale of two tumor targets: Topoisomerase I and tubulin. The Wall and Wani contribution to cancer chemotherapy, J. Nat. Prod., 67, 232, 2004. 3. Sielecki, T. et al., Cyclin-dependent kinase inhibitors: Useful targets in cell cycle regulation, J. Med. Chem., 43, 1, 2000. 4. Losiewicz, M. D. et al., Potent inhibition of cdc2 kinase activity by the flavanoid, l86-8275, Biochem. Biophys. Res. Comm., 201, 589, 1994. 5. Czech, J. et al., Anti-tumoral activity of flavone l 86-8275, Int. J. Oncol., 6, 31, 1995. 6. Dai, Y. and Grant, S., Cyclin-dependent kinase inhibitors, Curr. Opin. Pharmacol., 3, 362, 2003. 7. Dancey, J. and Sausville, E. A., Issues and progress with protein kinase inhibitors for cancer treatment, Nature Rev. Drug Disc., 2, 296, 2003. 8. Senderowicz, A. M., Small-molecule cyclin-dependent kinase modulators, Oncogene, 22, 6609, 2003. 9. Meijer, L. and Raymond, E., Roscovitine and other purines as kinase inhibitors. From starfish oocytes to clinical trials, Acc. Chem. Res., 36, 417, 2003. 10. Parker, C. W., Entsch, B., and Letham, D., Inhibitors of two enzymes which metabolize cytokinins, Phytochemistry, 25, 303, 1986. 11. Chang, Y. T. et al., Synthesis and application of functionally diverse 2,6,9-trisubstituted purine libraries as CDK inhibitors, Chem. Biol., 6, 361, 1999. 12. Evans, B. E. et al., Methods for drug discovery: Development of potent, selective, orally effective cholecystokinin antagonists, J. Med. Chem., 31, 2235, 1988. 13. Mason, J. S. et al., New 4-point pharmacophore method for molecular similarity and diversity applications: Overview of the method and applications, including a novel approach to the design of combinatorial libraries containing privileged sub-structures, J. Med. Chem., 42, 3251, 1999. Copyright © 2005 CRC Press, LLC 14. Patchett, A. A. and Nargund, R. P., Privileged structures — an update, in Ann. Repts. Med. Chem., Doherty, A. M., Ed., Academic Press, San Diego, CA, 2000, Vol. 35, pp 289–298. 15. Nicolaou, K. C. et al., Natural product-like combinatorial libraries based on privileged structures. 1. General principles and solid-phase synthesis of benzopyrans, J. Am. Chem. Soc., 122, 9939, 2000. 16. Nicolaou, K. C. et al., Natural product-like combinatorial libraries based on privileged structures. 2. Construction of a 10,000-membered benzopyran library by directed split-and-pool chemistry using nanokans and optical encoding, J. Am. Chem. Soc., 122, 9954, 2000. 17. Nicolaou, K. C. et al., Natural product-like combinatorial libraries based on privileged structures. 3. The “libraries from libraries” principle for diversity enhancement of bezopyran libraries, J. Am. Chem. Soc., 122, 9968, 2000. 18. Nicolaou, K. C. et al., Polymer-supported selenium reagents for organic synthesis, Chem. Comm., 1947, 1998. 19. Nicolaou, K. C. et al., Combinatorial synthesis of novel and potent inhibitors of NADH: Ubiquinone oxidoreductase, Chem. Biol., 7, 979, 2000. 20. Nicolaou, K. C. et al., Discovery of novel antibacterial agents active against methicillin-resistant Staphylococcus aureus from combinatorial benzopyran libraries, ChemBioChem, 460, 2001. 21. Nicolaou, K. C. et al., Discovery and optimization of non-steroidal FXR agonists from natural product- like libraries, Org. Biomol. Chem., 1, 908, 2003. 22. Downes, M. et al., A chemical, genetic, and structural analysis of the nuclear bile acid receptor FXR, Mol. Cell, 11, 1079, 2003. 23. Strobel, G. et al., Natural products from endophytic microorganisms, J. Nat. Prod., 67, 257, 2004. 24. Cuendet, M. and Pezzuto, J. M., Antitumor activity of bruceantin. An old drug with new promise, J. Nat. Prod., 67, 269, 2004. 25. Melichar, B. et al., Growth-inhibitory effect of a novel synthetic triterpenoid, 2-cyano-3,12-dioxoalean- 1,9-dien-28-oic acid, on ovarian carcinoma cell lines not dependent on peroxisome proliferators- activated receptor-γ expression, Gynecol. Oncol., 93, 149, 2004. 26. Li, Y. et al., Selective killing of cancer cells by β-lapachone: Direct checkpoint activation as a strategy against cancer, Proc. Natl. Acad. Sci., USA, 100, 2674, 2003. 27. Choi, B. T., Cheong, J., and Choi, Y. H., β-lapachone-induced apoptosis is associated with activation of caspase-3 and inactivation of NF-κB in human colon cancer HCT-116 cells, Anti-Cancer Drugs, 14, 845, 2003. 28. Ravelo, A. G. et al., Recent studies on natural products as anticancer agents, Curr. Topics Med. Chem., 4, 241, 2004. 29. Scott, J. D. and Williams, R. M., Chemistry and biology of the tetrahydroisoquinoline antitumor antibiotics, Chem. Rev., 102, 1669, 2002. 30. Oku, N. et al., Renieramycin J, a highly cytotoxic tetrahydroisoquinoline alkaloid, from a marine sponge Neopetrosia sp., J. Nat. Prod., 66, 1136, 2003. 31. Fontana, A. et al., A new antitumor isoquinoline alkaloid from the marine nudibranch Jorunna funebris, Tetrahedron, 56, 7305, 2000. 32. Moss, C. et al., Intracellular bacteria associated with the ascidian Ecteinascidia turbinata: Phylogenetic and in situ hybridization analysis, Mar. Biol., 143, 99, 2003. 33. Zabriskie, T. M. et al., Jaspamide, a modified peptide from a Jaspis sponge, with insecticidal and antifungal activity, J. Am. Chem. Soc., 108, 3123, 1986. 34. Crews, P., Manes, L. V., and Boehler, M., Jasplakinolide, a cyclodepsipeptide from the marine sponge, Jaspis sp., Tetrahedron Lett., 27, 2797, 1986. 35. Zampella, A. et al., New jaspamide derivatives from the marine sponge Jaspis splendans collected in Vanuatu, J. Nat. Prod., 62, 332, 1999. 36. Bubb, M. R. et al., Jasplakinolide, a cytotoxic natural product, induces actin polymerization and competitively inhibits the binding of phalloidin to f-actin, J. Biol. Chem., 269, 14869, 1994. 37. Senderowicz, A. M. et al., Jasplakinolide’s inhibition of the growth of prostate carcinoma cells in vitro with disruption of the actin cytoskeleton, J. Nat. Can. Inst., 87, 46, 1995. 38. Chan, W. R. et al., Stereostructures of geodiamolides A and B, novel cyclodepsipeptides from the marine sponge Geodia sp., J. Org. Chem., 52, 3091, 1987. Copyright © 2005 CRC Press, LLC 39. Coleman, J. E., Van Soest, R. W. M., and Andersen, R. J., New geodiamolides from the sponge Cymbastela sp. collected in Papua New Guinea, J. Nat. Prod., 62, 1137, 1999. 40. Kunze, B. et al., Chondramides A-D, new antifungal and cytostatic depsipeptides from Chondromyces crocatus (myxobacteria). Production, physico-chemical and biological properties, J. Antibiotics, 48, 1262, 1995. 41. Sasse, F. et al., The chondramides: Cytostatic agents from myxobacteria acting on the actin cytosk- eleton, J. Nat. Can. Inst., 90, 1559, 1998. 42. Perry, N. B. et al., Mycalamide A, an antiviral compound from a New Zealand sponge of the genus Mycale, J. Am. Chem. Soc., 110, 4850, 1988. 43. Pettit, G. R. et al., Antineoplastic agents. 520. Isolation and structure of irciniastatins A and B from the Indo-Pacific marine sponge Ircinia ramosa, J. Med. Chem., 47, 1149, 2004. 44. Cichewicz, R. H., Valeriote, F. A., and Crews, P., Psymberin, a potent sponge-derived cytotoxin from Psammocinia distantly related to the pederin family, Org. Lett., 6, 1951, 2004. 45. Trost, B. M., Yang, H., and Probst, G. D., A formal synthesis of (-)-mycalamide A, J. Am. Chem. Soc., 126, 48, 2004. 46. Kellner, R. L. L., Molecular identification of an endosymbiotic bacterium associated with pederin biosynthesis in Paederus sabaeus (coleoptera: Staphylinidae), Insect Biochem. Mol. Biol., 32, 389, 2002. 47. Piel, J., Hofer, I., and Hui, D., Evidence for a symbiosis island involved in horizontal acquisition of pederin biosynthetic capabilities by the bacterial symbiont of Paederus fuscipes beetles, J. Bact., 186, 1280, 2004. 48. Piel, J., A polyketide synthase-peptide synthase gene cluster from an uncultured bacterial symbiont of Paederus beetles, Proc. Natl. Acad. Sci. USA, 99, 14002, 2002. 49. Kasanah, N. et al., Biotransformation and biosynthetic studies of the manzamine alkaloids, Abs. Pap. 6th Int. Mar. Biotech. Conf., Abs S14, 2003. 50. Yousaf, M. et al., Solving limited supplies of marine pharmaceuticals through the rational and high- throughput modification of high yielding marine natural producer scaffolds, Abs. Pap. 6th Int. Mar. Biotech. Conf., Abs. S14, 2003. 51. Mincer, T. J. et al., Widespread and persistent populations of a major new marine actinomycete taxon in ocean sediments, Appl. Environ. Microbiol., 68, 5005, 2002. 52. Feling, R. H. et al., Salinosporamide A; a highly cytotoxic proteosome inhibitor from a novel microbial source, a marine bacterium of the new genus Salinospora, Angew. Chem. Int. Ed., 42, 355, 2003. 53. Omura, S. et al., Lactacystin, a novel microbial metabolite, induces neuritogenesis of neuroblastoma cells, J. Antibiot., 44, 113, 1991. 54. Corey, E. J. and Li, W. D., Total synthesis and biological activity of lactacystin, omuralide and analogs, Chem. Pharm. Bull., 47, 1, 1999. 55. Fenteany, G. et al., Inhibition of proteosome activities and subunit-specific amino-terminal threonine modification by lactacystin, Science, 268, 726, 1995. 56. Kisselev, A. F. and Goldberg, A. L., Proteosome inhibitors: From research tools to drug candidates, Chem. Biol., 8, 739, 2001. 57. Reddy, L. J., Saravanan, P., and Corey, E. J., A simple stereocontrolled synthesis of salinosporamide A, J. Am. Chem. Soc., 126, 6230, 2004. 58. Newman, D. J. and Cragg, G. M., Marine natural products and related compounds in clinical and advanced preclinical trials, J. Nat. Prod., 67, 1216, 2004. 59. Arabshahi, L. and Schmitz, F. J., Brominated tyrosine metabolites from an unidentified sponge, J. Org. Chem., 52, 3584, 1987. 60. Quinoa, E. and Crews, P., Phenolic constituents of Psammaplysilla, Tet. Lett., 28, 3229, 1987. 61. Pina, I. C. et al., Psammaplins from the sponge Pseudoceritina purpurea: Inhibition of both histone deacetylase and DNA methyltransferase, J. Org. Chem., 68, 3866, 2003. 62. Remiszewski, S. W. et al., N-hydroxy-3-phenyl-2-propenamides as novel inhibitors of human histone deacetylase with in vivo antitumor activity: Discovery of (2e)-n-hydroxy-3-[4-[[2-hydroxyethyl)[2- (1h-indol-3-yl)ethyl]amino]methyl]-phenyl]-2-propenamide (NVP-LAQ824), J. Med. Chem., 46, 4609, 2003. 63. Remiszewski, S. W., The discovery of NVP-LAQ824: From concept to clinic, Curr. Med. Chem., 10, 2393, 2003. Copyright © 2005 CRC Press, LLC 64. Catley, L. et al., NVP-LAQ824 is a potent novel histone deacetylase inhibitor with significant activity against multiple myeloma, Blood, 102, 2615, 2003. 65. Yoshida, M. et al., Effects of leptomycin B on the cell cycle of fibroblasts and fission yeast cells, Exp. Cell Res., 187, 150, 1990. 66. Nishi, K. et al., Leptomycin B targets a regulatory cascade of crm1, a fission yeast nuclear protein, involved in control of higher order chromosome structure and gene expression, J. Biol. Chem., 269, 6320, 1994. 67. Kudo, N. et al., Leptomycin B inhibition of signal-mediated nuclear export by direct binding to crm1, Exp. Cell Res., 242, 540, 1998. 68. Kudo, N. et al., Leptomycin B inactivates crm1/exportin 1 by covalent modification at a cysteine residue in the central conserved region, Proc. Natl. Acad. Sci., USA, 96, 9112, 1999. 69. Henderson, B. R. and Eleftheriou, A., A comparison of the activity, sequence specificity, and crm1- dependence of different nuclear export signals, Exp. Cell Res., 256, 213, 2000. 70. Vigneri, P. and Wang, J. Y. J., Induction of apoptosis in chronic myelogenous leukemia cells through nuclear entrapment of bcr-abl tyrosine kinase, Nat. Med., 7, 228, 2001. 71. Yashiroda, Y. and Yoshida, M., Nucleo-cytoplasmic transport of proteins as a target for therapeutic drugs, Curr. Med. Chem., 10, 741, 2003. 72. Krohn, K. et al., Palmarumycins C1-C16 from Coniothyrium: Isolation, structure elucidation, and biological activity, Liebigs Ann. Chem., 1099, 1994. 73. Welsh, S. J. et al., The thioredoxin redox inhibitors 1-methylpropyl 2-imidazoyl disulfide and pleurotin inhibit hypoxia-induced factor 1-α and vascular endothelial growth factor formation, Mol. Can. Therap., 2, 235, 2003. 74. Wipf, P. et al., New inhibitors of the thioredoxin-thioredoxin reductase system based on a naph- thaquinone spiroketal natural