Datasets:

health360
/

Healix-Shot

Dataset card Viewer Files Files and versions Community

Split (1)

train · 10.8M rows

text stringlengths 4.16k 144k
Bergman, T., and Everitt, E, (1990). Antibodies with specificities against a dispase-produced 15-kilodalton hexon fragment neutralize adenovirus type 2 infectivity. / . V/ro/. 64,4217-4225. 39. Shayakhmetov, D. M., and Lieber, A. (2000). Dependence of adenovirus infectivity on length of the fiber shaft domain./. Virol. 74, 10,274-10,286. 40. Weber, J. M., Gai, F., Murafi, R., and Burnett, R. M. (1994). Sequence and structural analysis of murine adenovirus type 1 hexon./. Gen. Virol. 75, 141-147. 41 . Schoehn, G., Fender, P., Ghroboczek, J., and Hewat, E. A, (1996). Adenovirus 3 pen- ton dodecahedron exhibits structural changes of the base on fibre binding. EMBO J. 15, 6841-6846. 42. Neumann, R., Ghroboczek, J., and Jacrot, B. (1988). Determination of the nucleotide sequence for the penton-base gene of human adenovirus type 5. Gene 69, 153-157. 43. Roberts, R. J., O'Neill, K. E., and Yen, G. T. (1984). DNA sequences from the adenovirus 2 genome./. Biol. Chem. 259, 13,968-13,975. 44. Ghroboczek, J., Bieber, F., and Jacrot, B. (1992). The sequence of the genome of adenovirus type 5 and its comparison with the genome of adenovirus type 2. Virology 186, 280-285. 45. Mathias, P., Wickham, T., Moore, M., and Nemerow, G. (1994). Multiple adenovirus serotypes use alpha v integrins for infection./. Virol. 68, 6811-6814. 46. Davison, A. J., Telford, E. A., Watson, M. S., McBride, K., and Mautner, V. (1993). The DNA sequence of adenovirus type 40 . / . Mol. Biol. 234, 1308-1316. 47. Karayan, L., Hong, S. S., Gay, B., Tournier, J., d'Angeac, A. D., and Boulanger, P. (1997). Structural and functional determinants in adenovirus type 2 penton base recombinant protein. / . Virol. 71, 8678-8689. 48. Hong, S. S., and Boulanger, P. (1995). Protein ligands of the human adenovirus type 2 outer capsid identified by biopanning of a phage-displayed peptide library on separate domains of wild-type and mutant penton capsomers. EMBO J. 14, 4714-4727. 49. Stupack, D. G., Li, E., Silletti, S. A., Kehler, J. A., Geahlen, R. L., Hahn, K., Nemerow, G. R., and Gheresh, D. A. (1999). Matrix valency regulates integrin-mediated lymphoid adhesion via Syk kinase./. Cell. Biol. 144, 777-788. 1. Adenovirus Structure 1 7 50. Acharya, R., Fry, E., Stuart, D., Fox, G., Rowlands, D., and Brown, F. (1989). The three- dimensional structure of foot-and-mouth disease virus at 2.9 A resolution. Nature 337, 709-716. 51. Jackson, T., Sharma, A., Ghazaleh, R. A., Blakemore, W. E., Ellard, F. M., Simmons, D. L., Newman, J. W., Stuart, D. I., and King, A. M. (1997). Arginine-glycine-aspartic acid-specific binding by foot-and-mouth disease viruses to the purified integrin avp3 in vitro. / . Virol. 71, 8357-8361. 52. Roelvink, P. W., Lizonova, A., Lee, J. G., Li, Y., Bergelson, J. M., Finberg, R. W., Brough, D. E., Kovesdi, L, and Wickham, T. J. (1998). The coxsackievirus-adenovirus receptor protein can function as a cellular attachment protein for adenovirus serotypes from subgroups A, C, D, E, and F. J. Virol. 72, 7909-7915. 53. Bergelson, J. M., Cunningham, J. A., Droguett, G., Kurt-Jones, E. A., Krithivas, A., Hong, J. S., Horwitz, M. S., Crowell, R. L., and Finberg, R. W. (1997). Isolation of a common receptor for Coxsackie B viruses and adenoviruses 2 and 5. Science 275, 1320-1323. 54. Tomko, R. P., Xu, R., and Philipson, L. (1997). HCAR and MCAR: The human and mouse cellular receptors for subgroup C adenoviruses and group B coxsackieviruses. Proc. Natl. Acad. Sci. USA 94, 3352-3356. 55. Yeh, H. Y., Pieniazek, N., Pieniazek, D., Gelderblom, H., and Luftig, R. B. (1994). Human adenovirus type 41 contains two fibers. Virus Res. 33, 179-198. 56. Laver, W. G., Banfield Younghusband, H., and Wrigley, N. G. (1971). Purification and properties of chick embryo lethal orphan virus. Virology 45, 598-614. 57. Gelderblom, H., and Maichle-Lauppe, L (1982). The fibers of fowl adenoviruses. Arch. Virol. 72,289-98. 58. Hess, M., Cuzange, A., Ruigrok, R. W., Chroboczek, J., and Jacrot, B. (1995). The avian adenovirus penton: Two fibres and one base. / . Mol. Biol. 252, 379-385. 59. Stouten, P. F., Sander, C., Ruigrok, R. W., and Cusack, S. (1992). New triple-helical model for the shaft of the adenovirus fibre./. Mol. Biol. 226, 1073-1084. 60. Boudin, M. L., D'Halluin, J. C., Cousin, C , and Boulanger, P. (1980). Human adenovirus type 2 protein Ilia. II. Maturation and encapsidation. Virology 101, 144-156. 61. Stewart, P. L., Fuller, S. D., and Burnett, R. M. (1993). Difference imaging of adenovirus: Bridging the resolution gap between X-ray crystallography and electron microscopy. EMBO J. 12,2589-2599. 62. Everitt, E., Lutter, L., and Philipson, L. (1975). Structural proteins of adenoviruses. XII. Loca tion and neighbor relationship among proteins of adenovirion type 2 as revealed by enzymatic iodination, immunoprecipitation and chemical cross-linking. Virology 67, 197-208. 63. Akusjarvi, G., and Persson, H. (1981). Gene and mRNA for precursor polypeptide VI from adenovirus type 2. / . Virol. 38, 469-482. 64. Mangel, W. F., McGrath, W. J., Toledo, D. L., and Anderson, C. W. (1993). Viral DNA and a viral peptide can act as cofactors of adenovirus virion proteinase activity. Nature 361, 274-275. 65. Webster, A., Hay, R. T., and Kemp, G. (1993). The adenovirus protease is activated by a virus-coded disulphide-linked peptide. Cell 72, 97-104. 66. Russell, W. C , and Precious, B. (1982). Nucleic acid-binding properties of adenovirus structural polypeptides./. Gen. Virol. 63, 69-79. 67. GaHbert, F., Herisse, J., and Courtois, G. (1979). Nucleotide sequence of the EcoRI-F fragment of adenovirus 2 genome. Gene 6, 1-22. 68. Herisse, J., Courtois, G., and GaHbert, F. (1980). Nucleotide sequence of the EcoRI D fragment of adenovirus 2 genome. Nucleic Acids Res. 8, 2173-2192. 69. Philipson, L. (1983). Structure and assembly of adenoviruses. Curr. Top. Microbiol. Immunol. 109, 1-52. 70. Colby, W. W., and Shenk, T. (1981). Adenovirus type 5 virions can be assembled in vivo in the absence of detectable polypeptide IX. / . Virol. 39, 977-980. 1 8 Phoebe L. Stev^art 71. Alestrom, P., Akusjarvi, G,, Perricaudet, M., Mathews, M. B., Klessig, D. F., and Pettersson, U. (1980). The gene for polypeptide IX of adenovirus type 2 and its unspHced messenger RNA.C^//19, 671-681. 72. Furcinitti, P. S., van Oostrum, J., and Burnett, R. M. (1989). Adenovirus polypeptide IX revealed as capsid cement by difference images from electron microscopy and crystallography. EMBOJ. 8,3563-3570. 73. Rekosh, D. (1981). Analysis of the DNA-terminal protein from different serotypes of human adenovirus./. Virol. 40, 329-333. 74. Smart, J. E., and Stillman, B. W. (1982). Adenovirus terminal protein precursor. Partial amino acid sequence and the site of covalent linkage to virus DNA. / . Biol. Chem. 257, 13,499-13,506. 75. Alestrom, P., Akusjarvi, G., Pettersson, M., and Pettersson, U. (1982). DNA sequence analysis of the region encoding the terminal protein and the hypothetical N-gene product of adenovirus type 2 . / . Biol. Chem. 257, 13,492-13,498. 76. Gingeras, T. R., Sciaky, D., Gelinas, R. E., Bing-Dong, J., Yen, C. E., Kelly, M. M., Bullock, P. A., Parsons, B. L., O'Neill, K. E., and Roberts, R. J. (1982). Nucleotide sequences from the adenovirus-2 genome./. Biol. Chem. 257, 13,475-13,491. 77. Webster, A., Leith, I. R., Nicholson, J., Hounsell, J., and Hay, R. T. (1997). Role of preterminal protein processing in adenovirus replication./. Virol. 71, 6381-6389. 78. Hosokawa, K., and Sung, M. T. (1976). Isolation and characterization of an extremely basic protein from adenovirus type 5 . / . Virol. 17, 924-934. 79. Alestrom, P., Akusjarvi, G., Lager, M., Yeh-kai, L., and Pettersson, U. (1984). Genes encoding the core proteins of adenovirus type 2. / . Biol. Chem. 259, 13,980-13,985. 80. Weber, J. M., and Anderson, C. W. (1988). Identification of the gene coding for the precursor of adenovirus core protein X . / . Virol. 62, 1741-1745. 81. Vayda, M. E., Rogers, A. E., and Flint, S.J. (1983). The structure of nucleoprotein cores released from adenovirions. Nucleic Acids Res. 11, 441-460. 82. Hannan, C., Raptis, L. H., Dery, C. V., and Weber, J. (1983). Biological and structural studies with an adenovirus type 2 temperature-sensitive mutant defective for uncoating. Intervirology 19,213-223. 83. Gotten, M., and Weber, J. M. (1995). The adenovirus protease is required for virus entry into host cells. Virology 213, 494-502. 84. Greber, U. F., Webster, P., Weber, J., and Helenius, A. (1996). The role of the adenovirus protease on virus entry into cells. EMBO J. 15, 1766-1777. 85. Mangel, W. F., Toledo, D. L., Brown, M. T., Martin, J. H., and McGrath, W.J. (1996). Characterization of three components of human adenovirus proteinase activity in vitro. / . Biol. Chem. 271, 536-543. 86. Webster, A., Leith, I. R., and Hay, R. T. (1994). Activation of adenovirus-coded protease and processing of preterminal protein. / . Virol. 68, 7292-7300. 87. Webster, A., Russell, S., Talbot, P., Russell, W. C., and Kemp, G. D. (1989). Characterization of the adenovirus proteinase: Substrate specificity./. Gen. Virol. 70, 3225-3234. 88. Webster, A., Russell, W. C , and Kemp, G. D. (1989). Characterization of the adenovirus proteinase: development and use of a specific peptide assay. / . Gen. Virol. 70, 3215-3223. 89. Wickham, T. J., Mathias, P., Cheresh, D. A., and Nemerow, G. R. (1993). Integrins alpha v beta 3 and alpha v beta 5 promote adenovirus internalization but not virus attachment. Cell 73,309-319. 90. Berman, H. M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T. N., Weissig, H., Shindyalov, I. N., and Bourne, P. E. (2000). The Protein Data Bank. Nucleic Acids Res. 28, 235-242. 91. KrauHs, P. J. (1991). MOLSCRIPT: A program to produce both detailed and schematic plots of protein structures. / . Appl. Crystallogr. 24, 946-950. C H A P T E R Biology of Adenovirus Cell Entry Glen R. Nemerow^ Department of Immunology The Scripps Research Institute Lo Jolla, California I. PathvNfay of Adenovirus Cell Entry Adenoviruses cause a significant number of acute respiratory, gastroin testinal, and ocular infections in man. While these infections are usually self-limiting they can result in significant morbidity and in immunocompro mised individuals are capable of causing fatal disseminated infections [1]. Among the ^^50 different adenovirus (Ad) serotypes, representing six different subgroups (A-F) [2, 3], the majority of information on the molecular basis of host cell interactions is derived from studies on the closely related types 2 and 5 (subgroup C) [4]. It is, therefore, not surprising that replication-defective forms of Ad5 are currently being used for most in vitro and in vivo gene delivery applications [5, 6]. Despite some reported successes, adenovirus-mediated gene delivery remains hampered due in large part to the host immune response to viral proteins [7, 8]. Increased knowledge of Ad structure [9, 10] and host cell interactions [11] may allov^ redesigning of viral vectors in order to avoid some of the major problems in this area. Ad types 2/5 bind to cells via their fiber protein [12], which recognizes a 46-kDa cell receptor known as CAR (Coxsackie and adenovirus receptor) [13, 14]. However, this high-affinity receptor interaction is unable to promote efficient virus uptake into the host cell. Instead, secondary interactions of the virus penton base protein with avp3 or avP5 integrins facilitates virus internalization [15] (Fig. 1). Adenovirus particles enter cells via ~120 nM clathrin-coated pits and vesicles [16]. Hela cells expressing a mutant form of dynamin; a large GTPase associated with endosome formation, fail to support efficient virus uptake or infection, indicating that clathrin/receptor-mediated endocytosis is the primary pathway of Ad2/5 infection of host cells [17]. Adenovirus internalization also requires the participation of cell signaling ADENOVIRAL VECTORS FOR GENE THERAPY 1 O Copyright 2002, Elsevier Science (USA). All rights reserved. 20 Glen R. Nemerow Actin Reorganization Adenovirus Internalization Figure 1 Schematic diagram of adenovirus cell entry events. Virus attachment is mediated by fiber protein (black) association with CAR. Subsequent interaction of the penton base (light gray) v/ith av integrins (dark gray) promotes Ad internalization. Integrin-mediated Ad internalization also requires the participation of several signaling molecules (c-Src, p i 30CAS, PI 3K, and Rho GTPAses) that mediate actin polymerization. molecules including phosphatidylinositol 3-OH kinase [18], a lipid kinase that regulates a number of important host cell functions. These signaling proteins form a complex that promotes the polymerization of cortical actin filaments needed to efficiently internalize virus particles [16, 19]. Similar processes are used for cell invasion by a number of pathogenic bacteria [20, 21]. While the role of actin in virus or bacteria cell entry has not been clearly defined, polymerized actin filaments may serve as a
scaffold to prolong the half-life of signaling complexes or they may provide the mechanical force necessary for the formation of endocytic vacuoles [22-24]. 2. Biology of Adenovirus Cell Entry 2 1 An important step of Ad entry postinternalization involves disruption of the early endosome allov^ing escape of virions into the cytoplasm prior to degradation by lysosomal proteases [25-27]. As is the case for many nonen- veloped viruses, the precise mechanisms involved in Ad-mediated endosome penetration remain poorly defined. The majority of studies indicate that expo sure of the virus to mildly acidic conditions (~pH 6.0) are sufficient to initiate the loss of key virus coat proteins as w êll as activate the viral encoded cysteine protease. Hov^ever, there is not complete agreement on the requirement for a proton gradient in the early endosome to initiate its disruption [28]. Follow^ing endosome disruption, adenovirus particles are rapidly (30-60 min) translo cated from the cytoplasm to the nucleus. Transmission electron micrographs obtained by several investigators have show^n that viral capsids are docked at the nuclear pore complex [29, 30]. Current information indicates that virus association w îth microtubules [31] may play a key role in nuclear transport. Biochemical analyses have indicated that the majority of the viral capsid remains in the cytoplasm during transport of the viral genome into the nucleus [32]. This latter process appears to require the major host cell factors involved in nuclear import, the heat-shock p70 protein (Hsp70) and perhaps other cellular factors [33]. Viral gene expression and/or viral replication takes place once nuclear transport has occurred and ultimately results in the generation of transgene products in the case of viral vectors or in progeny virions in the case of w^ild-type particles. As is the case w îth many human pathogens, important questions remain to be answ^ered regarding the precise mechanisms involved in each step of adenovirus cell entry. II. Cell Receptors Involved in Attachment A. CAR Studies carried out by Lonberg-Holm et aL [34] first demonstrated that adenovirus and Coxsackie B viruses share the same receptor. A function- blocking monoclonal antibody was subsequently raised against the adenovirus receptor [35]; however, it was only recently that this antibody was used to identify the attachment receptor, CAR [13]. A murine homolog (MCAR) of human CAR (HCAR) is also capable of serving as an Ad attachment receptor [14]. The mechanisms by which CAR expression is regulated in different cell types as well its role in normal host cell functions have not yet been determined. The gene encoding HCAR was recently localized to the short arm of chromosome 21 [36], a finding that may provide further clues to the function and/or regulation of this receptor. The extracellular domain of CAR contains two Ig-like domains but only the most N-terminal domain is necessary for virus interaction [37]. HCAR is anchored in the cell membrane 2 2 Glen R. Nemerow by a single transmembrane domain followed by a relatively long cytoplasmic tail. Previous studies have indicated that only the extracellular domain of CAR is required for adenovirus-mediated gene delivery since recombinant forms of CAR lacking the cytoplasmic tail are fully capable of supporting virus infection [38]. These studies indicate that signaling events are probably not involved in CAR-mediated virus attachment. Several lines of evidence indicate that CAR is a major determinant of Ad infection in vivo. For example, CAR expression is particularly high in cardiac tissue [14] and this correlates with efficient Ad-transgene delivery to the heart in vivo [39, 40]. In contrast, CAR expression is low or absent in primary human fibroblast [41] as well as most hematopoietic cells [42, 43] and these cell types are also difficult to infect with Ad5-based vectors. CAR expression is also limited to the basolateral surface of ciliated airway epithelial cells [44] and this has hampered efficient Ad-mediated gene delivery to the apical surface of these cells [45, 46]. Other investigators have reported that certain cell culturing conditions may also alter CAR expression and, thereby, influence Ad-mediated gene delivery [47]. The recent generation of transgenic mice expressing high levels of HCAR on peripheral blood lymphocytes has allowed efficient transduction of these cells by Ad5 vectors [48]. Ad binding to CAR is mediated by a high-affinity interaction with the fiber knob domain {K^ ~ 1 nM). There are approximately 30-50,000 CAR molecules per epithefial cell depending upon the tissue type [15]. Recent high-resolution structure studies and mutagenesis experiments have shed con siderable light on the molecular basis for fiber-CAR association. Bewley and coworkers have solved the crystal structure of the Ad 12 fiber knob domain in a complex with the first Ig-domain of CAR [49]. The cocrystal structure revealed that CAR interacts with the lateral surface of the fiber knob rather than on top of the fiber as had been predicted in earlier structure studies of the Ad5 fiber [50]. The CAR binding sites, which are composed of multiple regions on extended loop structures, are situated at the interface between individual fiber monomers. As many as three CAR molecules could bind to each fiber knob domain; however, this has not been formally demonstrated. Roelvink and colleagues reported that highly conserved amino acid residues located in the AB loop of the fiber knob domain of adenovirus types are involved in CAR binding as well as in the cocrystal contacts [49]. In contrast, divergent sequences are present in the same region of Ad types (i.e., subgroup B) which do not use CAR [51, 52]. They also showed that site-specific mutations in the AB loop significantly reduced virus binding and infection. The identification of the precise regions in the fiber involved in CAR association has provided an opportunity to generate novel Ad vectors in which the CAR binding sites have been deleted and new receptor binding epitopes inserted (i.e. HI loop). Roelvink et al. have demonstrated the feasibility of this approach by redirecting an Ad5 vector containing a CAR-binding mutation to 2. Biology of Adenovirus Cell Entry 2 3 a novel host cell molecule [52]. Thus, the development of new Ad vectors vî ith increased host cell specificity may be on the horizon. B. Other Adenovirus Receptors Although CAR represents the major host cell receptor for Ad binding and infection, recent studies have suggested that other host cell molecules may also serve as attachment receptors. The a2 domain of MHC class I molecules has been reported to serve as a receptor for Ad5 particles based on competi tion experiments with phage-display peptides [53]. At the present time, these findings have yet to be confirmed by other investigators and thus it remains uncertain as to whether MHC class 1 molecules are specific Ad receptors. Recently Dechecchi et al. provided data suggesting that heparin sulfate pro teoglycans (HSPGs) may also promote cell attachment of subgroup C (Ad2/5) but not subgroup B (Ad3) virus particles [54]. These findings suggest that HSPG may work in concert with CAR to facilitate high-affinity subgroup C virus binding to cells. These investigators also suggested that heparin sulfate proteoglycan interactions may occur via a site(s) located in the fiber shaft rather in the knob domain. Belin and Boulanger have also analyzed host cell proteins capable of interacting with virus particles by cross-linked Ad2 to Hela cells. They showed that cross-hnked virus was bound to three major host cell proteins with molecular weights of 130, 60, and 44 kDa \SS\ They concluded that the 130-kDa protein was a pi integrin subunit; however, they did not identify the lower molecular weight proteins. Based on its apparent mobility on SDS gels, the 44-kDa protein likely is CAR. Ad types belonging to subgroup B that lack the conserved CAR binding residues noted above are very likely to use alternative cell receptors; however, these molecules have yet to be fully characterized. For example, serotypes 3 and 7 have been shown to bind to cells in a CAR-independent manner since the fiber proteins from these types fail to compete Ad5 fiber binding to cells \SG\. While the receptor for subgroup B adenoviruses have not been identified, a partial characterization of a candidate receptor has been reported \S1\, Additional investigations have indicated that Ad serotypes belonging to other subgroups may also use distinct cell receptors for virus attachment. Using virus protein blot assays, Roelvink et a\, demonstrated that Ad serotypes belonging to subgroups A, C, D, E, and F were capable of binding to CAR [51]; however, these investigators did not establish that different virus types were actually capable of associating with CAR on intact cells. This distinction may be important given the fact that there are structural differences in the fiber proteins of different Ad serotypes. For example, adenoviruses from subgroup B and D fibers have relatively short and inflexible fiber shaft domains. These structural features could restrict interaction with CAR on the lateral surface 2 4 Glen R. Nemerow of a short-shafted fiber. In support of this, Shayakhmetov and Lieber found that truncated Ad5 fiber molecules have reduced binding capacity [58]. Huang et al. previously showed that Ad37 (a subgroup D serotype) contains a short- shafted fiber protein displaying the conserved CAR binding residues in its AB loop, but fails to efficiently infect CAR-expressing epithelial cells. Instead, a critical lysine residue at position 240 in the CD loop of the Ad37 fiber knob, mediates association with a cell receptor, expressed on conjunctiva epithelial cells [59]. Arnberg and coworkers also reported that Ad37 does not use CAR but instead recognize sialic acid residues present on one or more unidentified cell membrane proteins [60]. In recent studies, Wu et aL discovered that a 60-kDa protein expressed on diverse cell types is recognized by Ad37 particles and that this association is dependent upon sialic acid [61]. These authors also found that a 50-kDa membrane protein that is preferentially expressed on conjunctiva and supports Ad37 binding in a sialic acid-independent manner. They concluded that the 50-kDa putative receptor represents a portal of entry for pathogenic strains of adenovirus that are associated with severe ocular infections. Further biochemical and molecular biological studies are needed to identify different Ad receptors and determine their precise role in tissue tropism and disease. Despite the fact that alternative Ad receptors have yet to be identified, new viral vectors with altered cell tropism have been generated. For example, several investigators have replaced (pseudotyped) the fiber protein in a first- generation Ad5 vector with an Ad3 or Ad7 fiber [56, 62-64]. The Ad3 pseudotyped vectors were shown to improve gene delivery to several different cell types. Thus, human lymphocytes which are very poorly transduced by Ad5 vectors supported substantially higher levels of infection with vectors equipped with the Ad3 fiber [64], presumably because of higher level of expression of the Ad3 receptor on these cells compared to CAR. Chillon et al also showed that an Ad5 vector pseudotyped with an Adl7 (subgroup D) fiber showed enhanced gene delivery to neuronal cells [65] while vectors retargeted with an Ad35 fiber improve gene delivery to stem cells [66]. It is likely that as new Ad receptors are identified, further knowledge of their tissue expression and structure should lead to improved modifications of standard El A - Ad5 vectors in order to increase host cell specificity. III. Adenovirus Internalization Receptors A. Role of av Integrins as Coreceptors In studies conducted over 40 years ago, Pereira [67] and Everitt et al. [68] described a soluble toxic factor produced during adenovirus infection that caused cell rounding. This toxic factor was later identified as the penton base 2. Biology of Adenovirus Cell Entry 2 5 protein [69]. Wickham et al. subsequently demonstrated that the penton base protein was not actually toxic, although it was capable of inducing epithelial cell detachment from plastic tissue culture surfaces. Cell detachment is due to the presence of an integrin-binding motif (RGD) in the penton base [70] that is able to compete for vitronectin, an extracellular matrix protein. In further studies, Wickham showed that penton base association with the vitronectin binding integrins av^S and avP5 promotes adenovirus internalization rather than virus attachment [15]. While the overall contribution of av integrins in adenovirus infection in vivo has not been firmly established, several lines of evidence suggest that integrins play a significant role. The penton base proteins of most adenoviruses representing different subgroups contain a conserved RGD motif and these viruses also use av integrins for
infection [71, 72]. Interestingly, adenoviruses belonging to subgroup F (types 40, 41) lack an RGD motif and show delayed uptake into cells [73]. Bai et al. also showed that mutations in the penton base RGD motif reduce the kinetics of Ad2 infection in vitro [74]. Huang et al. demonstrated that human B lymphocytes lack av integrins and are not susceptible to infection with Ad5 vectors [43]. In contrast, transformation of B cells with Epstein-Barr virus upregulates av integrin expression and allows infection with Ad5-based vectors. Von Seggern et al. has produced a fiberless adenovirus vector that fails to bind to CAR [7S]. These particles are significantly less infectious that wildtype Ad particles; however, they can infect human monocytic cells in an integrin-dependent manner. Recently, mice genetically deficient (knockouts) in cell integrins have been generated that may allow further investigation of the role of av integrins in adenovirus infection in vivo. Bader and coworkers described the generation of mice genetically deficient in the av integrin subunit which, therefore, lack both avP3 and avp5 [7G\. Unfortunately, the majority of these animals die early during development, thus precluding analyses of adenovirus infection. Huang et al. have reported the generation of p5-integrin-deficient mice and fortunately these animals do not show enhanced developmental lethality [77]. Interestingly, P5-deficient mice did not show decreased susceptibility to adenovirus infection suggesting that expression of this coreceptor is not an absolute requirement for virus infection. However, compensatory cell entry pathways mediated by integrin avP3 or perhaps other as yet unidentified receptors may confound interpretations of these findings. While the precise contribution of integrins to adenovirus infection in vivo remains to be determined, knowledge of integrin interactions has allowed further modification of Ad vectors to take advantage of the integrin/coreceptor pathway to improve gene delivery. For example, Vigne and coworkers showed that a recombinant adenovirus containing RGD motifs inserted into the hexon protein could infect vascular smooth muscle cells in CAR independent manner [78]. Wickham and colleagues have also replaced the penton base RGD motif 2 6 Glen R. Nemerov^ with a pi-integin binding motif (LDV) [79] and suggested that this might be advantageous for expanding the cell tropism of modified Ad vectors since this receptor is broadly distributed on most cell types. B. Structural Features of Penton Base-av Integrin Association A monoclonal antibody (DAV-1) was used to localize the integrin bind ing sites on the penton base protein using cryoelectron microscopy and image reconstruction [80]. This antibody recognizes the Ad2 penton base RGD motif as well as several flanking residues (IRGDTFATR). In more recent studies, Mathias et al. have produced a soluble form of avp5 integrin containing the entire ectodomain of the receptor [81]. This recombinant protein retained ligand-binding activity and was subsequently used to examine the complex of Ad particle and soluble avP5 by cryo-EM [82] (Fig. 2, see color insert). The inte grin ectodomain consists of an N-terminal globular (proximal) region, which is attached to slender stalk-like segments that are intertwined in the cryo-EM images. Approximately four to five soluble integrin molecules were capable of binding to each penton base protein as assessed by surface plasmon resonance analyses (BIAcore), consistent with density measurements obtained in the cryo- EM studies. The integrins form a ring-like structure above the virus surface. Each integrin molecule binds at an approximately 45° angle relative to the fiber shaft, a feature that may allow multimeric receptor association. A small cleft at the base of the integrin proximal domain, which interacts with the 20 A RGD protrusion, could also be visualized (Fig. 3, see color insert). The five RGD protrusions on the penton base are spaced approximately 60 A apart. It is inter esting to note that foot-and-mouth disease virus (FMDV), a nonenveloped RNA virus that also uses integrins for infection, has a similar geometrical arrange ment of its RGD motifs [83]. This observation suggests that the precise display of RGD sites on a nanoscale level plays a key role in promoting integrin cluster ing at the cell surface. In support of this concept, Stupack et al, demonstrated that the multimeric penton base protein but not a monomeric RGD peptide could stimulate B cell signaling and cell adhesion [84]. Maheshawri et al. have also shown that conjugation of RGD peptides on a synthetic substrate with an average spacing of 50 A allows efficient integrin-mediated cell motility [85]. Integrin clustering is intimately associated with signaling processes and actin rearrangement required for efficient virus entry (discussed below). C. Signaling Events Associated with Adenovirus Internalization Association of cell integrins such as avp3 with the extracellular matrix induces the formation of focal adhesion complexes. These integrin complexes contain a number of cytoskeletal associated proteins that recognize specific amino acid sequences located in P integrin cytoplasmic domains as well 2. Biology of Adenovirus Cell Entry 2 7 as tyrosine and mitogen-activated kinases, lipid kinases, and various other adapter molecules [S6, 87]. Integrin-mediated signaling events play a crucial role in several important cell processes including cell motility, tumor cell metastasis, wound healing, and cell grov^th and differentiation [88, 89]. Integrin-mediated signaling events also facilitate host cell invasion by a number of pathogenic bacteria [90] as w êll as other viruses [91]. A general feature observed in integrin signaling is the rearrangement of actin filaments underlying the plasma membrane. Recent studies have indicated that actin assembly may play a significant role in receptor-mediated endocytosis in mammalian cells [22]. Filamentous actin could provide additional mechanical force necessary for endosome formation [24] or it may serve as a platform to stabilize the half-life of signaling complexes needed to induce receptor internalization [92]. Cytochalasin D, an agent that disrupts the actin cytoskeleton also inhibits adenovirus entry and infection [16]. Li and colleagues therefore investigated whether specific signaling events leading to actin reorganization were also involved in adenovirus internalization [19]. They found that adenovirus inter action with cells altered the cell membrane shape, induced polymerized cortical actin filaments as well as activated phosphatidylinositol-3-OH kinase (PI3K). PIP3, a major product of PI3K, acts as a second messenger in many differ ent cell signaling processes, including those regulating cytoskeletal function [93] and bacterial cell invasion [94]. Li et al. found that activation of PI3K was also required for efficient Ad internalization but not virus attachment [18]. PI3K is also capable of activating Rab5, a GTPase associated with early endosome formation. Overexpression of a dominant negative Rab5 in host cells significantly inhibits adenovirus endocytosis and infection [95]. Several lines of evidence suggest that it is the penton base interaction with integrin coreceptors that initiates the key signaling events for virus entry and infection. First, recombinant penton base but not fiber protein is capable of activating PI3K [18]. Second, fiberless adenovirus particles induce similar levels of cell signaling as wild-type fiber-containing virions [96], Finally, Bergelson et al. have shown that mutant forms of CAR that lack a normal transmembrane anchor and cytoplasmic domain support normal levels of adenovirus-mediated gene delivery [38]. In addition to PI3K, the Rho family of small GTPases including Racl, CDC42 and RhoA also are involved in adenovirus cell entry. These small GTPases are tightly regulated molecular switches that control changes in cell shape as well as actin reorganization [97] via interaction with additional downstream effector molecules such as WASP and PAKl [98]. Expression of dominant-negative forms of Rac or CDC42 reduce virus entry and infection [19]. Recently, Li et al. found that pl30CAS is also required for efficient adenovirus entry [96]. This large adapter molecule provides an important functional link between the tyrosine kinase c-SRC [99] and the p85 catalytic 2 8 Glen R. Nemerow subunit of PI3K [96]. The downstream effect molecules downstream of PI3K and CAS have yet to be fully characterized. It is interesting to note that other signaling molecules may become activated upon adenovirus interaction with host cells; however, they may not actually contribute to virus entry. For example, pl25FAK (focal adhesion kinase) becomes tyrosine phosphorylated during adenovirus entry [19], but cells expressing dominant negative forms of FAK exhibit normal levels of Ad uptake [19]. Moreover, mouse embryonic fibroblasts genetically deficient in FAK support very similar levels of Ad infection as expressing cells. Bruder et al. [100] also reported that MAP kinases are activated during Ad infection, whereas inhibitors of ERK1/ERK2 MAP kinases have little if any effect on virus entry [18]. Despite recent progress, the precise mechanisms by which signaling processes regulate virus entry have not been elucidated. Impediments to further advances include the difficulty of studying complex signaling events in live (unfixed) cells. Moreover, signaling processes may vary among different cell types and thus the overall role of a given signaling pathway may differ in different cell types. Finally, the involvement of a specific signaling molecule may be difficult to assess if related molecules (functional homologs) perform similar functions. While further research is needed to fully characterize Ad cell entry mechanisms, the identification of specific signaling molecules involved in adenovirus cell entry may allow improvements in Ad-mediated gene delivery to cells which lack CAR and/or av integrins. For example, ligation of certain growth factor receptors (i.e., epithelial growth factor (EGF)) or cytokines (i.e., tumor necrosis factor alpha (TNFa)) results in activation of remarkably similar signaling pathways as those induced by integrin clustering [101-103]. Li et al. recently investigated whether activation of growth factor receptors could circumvent the need for av integrins/CAR in adenovirus-mediated gene delivery [104]. They generated a bifunctional antibody that recognizes the penton base RGD motif (DAV-1) as well as one of several different cytokine or growth factor receptors. Ad vectors complexed with these bifunctional molecules significantly increased PI3K activation in host cells and improved gene delivery to human melanoma cells that lack avp3 and av^5 integrins. The bifunctional antibody also increased gene delivery by a fiberless adenovirus vector. In addition to having a direct role in adenovirus cell entry, signaling events may also contribute to host immune responses to viral vectors. For example, Bruder and Kovesdi previously reported that adenovirus infection triggers expression of interleukin-8 [100], a response that may enhance the inflammatory reactions associated with in vivo delivery of viral vectors for gene therapy. Zsengeller and coworkers also reported that adenovirus internaliza tion into macrophages involves PI3K-mediated signaling and this is associated with the production of inflammatory cytokines in vivo [105]. Further studies 2. Biology of Adenovirus Cell Entty 2 9 are therefore needed to determine the extent to which the signaUng events ehcited during cell entry influence host immune responses to the virus. These processes are likely to have an impact on vector toxicity as w êll as the duration of transgene expression. IV. Virus-Mediated Endosome Disruption and Uncoating In contrast to enveloped viruses, much less is known about how nonen- veloped viruses traverse cell membranes during the infectious process. Early electron microscopic studies by Chardonnet and Dales [29] and subsequently by Patterson et al. [16] showed that Ad5 particles are rapidly internalized into clathrin-coated vesicles and shortly thereafter are found free in the cytoplasm. The ability of endocytosed Ad particles to escape the early endosome prior to degradation in lysosomes is a key feature of Ad-mediated gene delivery. Although the precise details of adenovirus-mediated endosome penetration remains a mystery; prior studies have provided a few clues that may ultimately lead to further advances in our understanding of Ad entry. A. Role of Penton Base and av Integrins Seth and coworkers first showed that adenovirus interaction with cells alters membrane permeability [25] and that this depends upon association of the penton base with av integrins [15, 26, 106, 107]. Ad-mediated membrane permeabilization occurs at a pH that is very similar to the environment of the early endosome (pH 6.1) [27, 108-110]. The exact nature of the membrane lesion has not yet been revealed; however, it does not appear to be the result of ion channel formation. Further studies indicated that of all the major Ad capsid proteins, the penton base plays a key role in facilitating membrane per meabilization. Interestingly, the penton base of different adenovirus serotypes exhibit different levels of membrane permeabilizing activity. For example, type 3 but not type 2/5 penton base is capable of forming a dodecahedron [111] and Ad3 dodecahedra also directly transduce cDNA into host cells [111], whereas Ad2 penton base monomers do not [107]. Wickham etal. previously demonstrated that although integrins avp3 and avp5 both support adenovirus
internalization, av^5 plays a preferential role in membrane permeabilization and infection [15]. Wang et aL subsequently showed that the cytoplasmic tail of the p5 integrin subunit regulates Ad escape from early endosomes [30]. In these studies, they identified multiple TVD motifs, present in the p5 cytoplasmic tail but not in other integrin subunits, that promote membrane permeabilization. These findings suggest that other as 3 0 Glen R. Nemerow yet unidentified host cell molecules may interact with p5 integrin cytoplasmic tail to promote virus penetration. B. Role of the Adenovirus Cysteine Protease The Ad2 penton base, either in its native form or presented in a multivalent form on latex particles, is unable to directly mediate membrane per- meabilization [15], This suggests that other virus/host cell factors are required for efficient virus penetration. Hannan et al. first described a temperature- sensitive Ad particle, designated ^sl, w^hich failed to cleave five precursor viral proteins at the nonpermissive temperature as w êll as lacked infectivity and uncoating activities associated with wild-type virions [112]. tsl particles can bind and enter host cells but remain trapped inside cell vesicles and eventu ally undergo lysosomal degradation. Cotton and Weber subsequently showed that tsl particles fail to incorporate the adenovirus-encoded 23K cysteine pro tease which is normally present in approximately 10 copies per virion particle and as a consequence, fail to mediate efficient gene delivery of membrane permeabilization [113]. Further biochemical studies by Greber et al. showed that tsl virions also lack the ability to cleave the capsid-stabilizing protein VI [114], a molecule associated with virus uncoating and endosome escape. Interestingly, these investigators reported that interaction of Ad particles with cell integrins was required for activation of the cysteine protease based on competition studies with RGD peptides. While these and other studies have provided some clues as to the events associated with virus penetration and uncoating, further studies are needed to determine the precise mechanisms underlying these events. V. Beyond the Endosome: Trafficking of Viral Capsids and Import of Viral DNA into the Nucleus A. Intracytoplasmic Transport of Viral Capsids An important step in adenovirus ceil entry is the transport of viral capsids to the nucleus following their escape from the early endosome. Early electron microscopic studies by Chardonnet and Dales had suggested that adenovirus particles associate with microtubules during nuclear transport [29]. Unfortunately, it was difficult to discern from these early investigations whether Ad particles were nonspecifically associated with these structure elements during sample preparation. Several investigators have therefore sought to test the validity of these early findings. Using fluorescence-tagged viruses, Greber and colleagues found that adenovirus particles fail to traffic to the 2. Biology of Adenovirus Cell Entry 3 1 nucleus in nocodazole-treated cells, consistent with a role for microtubules [31]. Furthermore, overexpression of p50/dynamitin, a molecule which is known to regulate microtubule motor (dynein)-mediated transport, altered adenovirus movement [31]. Thus, as is the case for some other large DNA viruses, [115] adenovirus appears to use the microtubule apparatus to achieve vectorial movement through the host cell. Matthews and Russell also reported that a cellular protein, p32, may also participate in vectorial transport of Ad capsids [116]. B. Docking at the Nuclear Pore and Translocation of Viral DNA Since adenovirus replicates in the nucleus, it must deliver its genome into this cellular compartment to complete the infectious cycle. Consistent with this concept, electron microscopic studies have revealed partially uncoated adenovirus capsids docked at nuclear pore complexes of infected cells within 1-2 h postinfection. The relatively limited size of the nuclear pore complex, approximately 25 nM in diameter, also indicates that Ad capsids (approx. 90 nM) do not directly enter into the nucleus. Moreover, proteins of greater than 20-40 kDa cannot passively diffuse through the nuclear pore complex and thus the classical nuclear import machinery must then be used to facilitate translocation of the viral genome and associated proteins through pore com plexes. Previous studies have indicated that after exposure of viral particles to low pH, the hexon protein is the major capsid protein that docks at the nuclear pore complex [33]. Using a permeabilized cell system, Saphire et al. showed that purified nuclear transport factors such as importin-a and -P, as well as heat shock 70 (hsp70), are required to facilitate nuclear import of purified hexon proteins but these factors cannot promote import of adenovirus DNA [33]. These findings indicated that other as yet unidentified cellular factors may also be required for DNA translocation. One major question that remains to be addressed is whether nuclear import of the Ad genome requires a protein chaperone(s). In this regard, Greber and coworkers previously showed that protein VII, a protein that is associated with the viral DNA inside the capsid also enters the nucleus along with the viral DNA [114]. In contrast, the vast majority of the hexon outer coat protein remains in the cytoplasm [32]. Further studies are needed to directly demonstrate a role for protein VII or other molecules in facilitating DNA import. VI . Conclusions Adenovirus cell entry requires interactions of multiple host cell receptors with distinct virus capsid proteins. Adenovirus associations with different 3 2 Glen R. Nemerov^ receptors influences cell tropism and undoubtedly plays an important role in determining the efficiency of Ad-mediated gene delivery in vivo. The Ad capsid structure is particularly well designed to mediate multiple receptor events. The elongated and flexible fiber protein of most Ad serotypes mediates high-affinity binding with a receptor (CAR) that is broadly distributed on different host cells. Further studies are needed to determine how the structure of the fiber shaft influences receptor usage and to uncover other host cell receptors that can serve as receptors for different Ad types. The Ad penton base displays five RGD integrin binding motifs and the precise geometrical arrangement of these motifs Ukely facilitates integrin clustering and subsequent signaling events. In particular, integrin coreceptors induce activation of P13K and Rho GTPases that promote virus entry and endosome penetration. Events occurring subsequent to internalization including endosome disruption remain obscure. Other host cell molecules interacting with integrin avp5 may play a key role in this process. Finally, adenovirus may provide important clues as to the mechanisms by which nucleic acids are transported into the nucleus. Increased knowledge of virus structure and host cell interactions has led to reengineering of first-generation Ad vectors to improve tissue targeting, and this may improve transgene expression as well as reduce vector toxicity. Acknov^ledgments I am indebted to members of my laboratory, past and present, who have made several of the major contributions cited in this review. I also express my gratitude to Dr. Phoebe Stewart, whose cryo-EM and image reconstructions of adenovirus complexes are shown in this review. I also thank Joan Gausepohl for preparation of this chapter. This work was supported by NIH Grants EY11431 and HL54352 and is Publication 13698 from The Scripps Research Institute. References 1. Horwitz, M. (1990). The adenoviridae and their replication. In "Virology" (B. Fields and D. Knipe, Eds.)., pp. 1679-1721. Raven Press, New York. 2. Hierholzer, J. C. (1992). Adenoviruses in the immunocompromised host. Clin. Microbiol. Rev. 5, 262-274. 3. Wadell, G., Hammarskjold, M. J., Winberg, G., Varsanyi, T. M., and Sundell, G. (1980). Genetic variability of adenoviruses. Ann. NY Acad. Sci. 354, 16-42. 4. Russell, W. C. (2000). Update on adenovirus and its vectors. / . Gen. Virol. 81, 2573-2604. 5. Wilson, J. M. (1998). Molecular medicine —Adenoviruses as gene-delivery vehicles. Mol. Me^. 334, 1185-1187. 6. Nabel, G. J. (1999). Development of optimized vectors for gene therapy. Proc. Natl. Acad. Sci. USA 96, 324-326. 7. Wilson, J. M. (1995). Gene therapy for cystic fibrosis: Challenges and future directions. / . Clin. Invest. 96, 2547-2554. 2. Biology of Adenovirus Cell Entry 3 3 8. Elkon, K. B., Liu, C.-C, Gall, J. G., Trevejo, J., Marino, M. W., Abrahamsen, K. A., Song, X., Zhou, J.-L., Old, L. J., Crystal, R. G., et al. (1997). Tumor necrosis factor a plays a central role in immune-mediated clearance of adenoviral vectors. Proc. Natl. Acad. Sci. USA 94,9814-9819. 9. Stev^art, P. L., Burnett, R. M., Cyrklaff, M., and Fuller, S. D. (1991). Image reconstruction reveals the complex molecular organization of adenovirus. Cell 67, 145-154. 10. Burnett, R. M. (1990). Structure of the adenovirus virion. In "Use of X-ray Crystallography in the Design of Antiviral Agents" (W. G. Laver and G. M. Air, Eds.), pp. 35-48, Academic Press, San Diego. 11. Nemerow, G. R. (2000). Cell receptors involved in adenovirus cell entry. Virology 274, 1-4. 12. Philipson, L., Lonberg-Holm, K., and Pettersson, U. (1968). Virus-receptor interaction in an adenovirus system./. Virol. 2, 1064-1075. 13. Bergelson, J. M., Cunningham, J. A., Droguett, G., Kurt-Jones, E. A., Krithivas, A., Hong, J. S., Horv^itz, M. S., Crov^ell, R. L., and Finberg, R. W. (1997). Isolation of a common receptor for coxsackie B viruses and adenoviruses 2 and 5. Science 275, 1320-1323. 14. Tomko, R. P., Xu, R., and Philipson, L. (1997). HCAR and MAR: The human and mouse cellular receptors for subgroup C adenoviruses and group B coxsackieviruses. Proc. Natl. Acad. Sci. USA 94, 3352-3356. 15. Wickham, T. J., Mathias, P., Cheresh, D. A., and Nemerov^, G. R. (1993). Integrins ayPs and ttvPi promote adenovirus internalization but not virus attachment. Cell 73, 309-319. 16. Patterson, S., and Russell, W. C. (1983). Ultrastructural and immunofluorescence studies of early events in adenovirus-HeLa cell interactions./. Gen. Virol. 64, 1091-1099. 17. Wang, K., Huang, S., Kapoor-Munshi, A., and Nemerov^, G. R. (1998). Adenovirus inter- nahzation and infection required dynamin. / . Virol. 72, 3455-3458. 18. Li, E., Stupack, D., Cheresh, D., Klemke, R., and Nemerow, G. (1998). Adenovirus endocy- tosis via av integrins requires phosphoinositide-3-OH-kinase./. Virol. 72, 2055-2061. 19. Li, E., Stupack, D., Bokoch, G., and Nemerov^, G. R. (1998). Adenovirus endocytosis requires actin cytoskeleton reorganization mediated by Rho family GTPases. / . Virol. 72, 8806-8812. 20. Cossart, P. (1997). Perspectives series: Host/pathogen interactions. / . Clin. Invest. 100, S33-S37. 21. Alrutz, M. A., and Isberg, R. R. (1998). Involvement of focal adhesion kinase in invasin- mediated uptake. Proc. Natl. Acad. Sci. USA 95, 13,658-13,663. 22. Fujimoto, L. M., Roth, R., Heuser, J. E., and Schmid, S. L. (2000). Actin assembly plays a variable, but not obligatory role in receptor-mediated endocytosis in mammalian cells. Traffic 1,161-171. 23. Garrett, W. S., Chen, L.-M., Kroschewski, R. E. M., Turley, S., Trombetta, S., Galan, J. E., and Mellman, I. (2000). Developmental control of endocytosis in dendritic cells by Cdc42. Cell 102, 325-334. 24. Qualmann, B., Kessels, M. M., and Kelly, R. B. (2000). Molecular Hnks betv^een endocytosis and the actin cytoskeleton. / . Cell. Biol. 150, F111-F116. 25. Seth, P., Pastan, I., and Willingham, M. C. (1987). Adenovirus-dependent changes in cell membrane permeability: Role of Na+, K+- ATPase./. Virol. 61 , 883-888. 26. Seth, P., Fitzgerald, D., Ginsberg, H., Willingham, M., and Pastan, I. (1984). Evidence that the penton base of adenovirus is involved in potentiation of toxicity of Pseudomonas exotoxin conjugated to epidermal growth factor. Mol. Cell. Biol. 4, 1528-1533. 27. Blumenthal, R., Seth, P., Willingham, M. C , and Pastan, I. (1986). pH-Dependent lysis of liposomes by adenovirus. Biochemistry 25, 2231-2237. 28. Rodriguez, E., and Everitt, E. (1996). Adenovirus uncoating and nuclear establishment are not affected by weak base amines. / . Virol. 70, 3470-3477. 3 4 Glen R. Nemerow 29. Chardonnet, Y., and Dales, S. (1970). Early events in the interaction of adenoviruses v^ith HeLa cells. I. Penetration of type 5 and intracellular release of the DNA genome. Virology 40, 462-477. 30. Wang, K., Guan, T., Cheresh, D. A., and Nemerov^, G. R. (2000). Regulation of adenovirus membrane penetration by the cytoplasmic tail of integrin E5. / . Virol. 74, 2731-2739. 31. Suomalainen, M., Nakano, M. Y., Keller, S., Boucke, K., Stidv^ill, R. P., and Greber, U. F. (1999). Microtuble-dependent plus-and minus end-directed motilities are competing processes for nuclear targeting of adenovirus. / . Cell. Biol. 144, 657-672. 32. Greber, U. F., Willetts, M., Webster, P., and Helenius, A. (1993). Stepv^ise dismantHng of adenovirus 2 during entry into cells. Cell 75, 477-486. 33. Saphire, A. C. S., Guan, T., Schirmer, E., Nemerow^, G. R., and Gerace, L. (2000). Nuclear import of adenovirus DNA in vitro involves the nuclear protein import pathv^ay and Hsc70. / . Biol. Chem. 275, 4298-4304. 34. Lonberg-Holm, K., Crowell, R. L., and Philipson, L. (1976). Unrelated animal
viruses share receptors. Nature 259, 679-681. 35. Hsu, K.-H., Lonberg-Holm, K., Alstein, B., and Crow^ell, R. L. (1988). A monoclonal antibody specific for the cellular receptor for the group B coxsackieviruses. / . Virol. 62, 1647-1652. 36. Bowles, K. R., Gibson, J., Wu, J., Shaffer, L. G., Towbin, J. A., and Bowles, N. E. (1999). Genomic organization and chromosomal localization of the human coxsackievirus B- adenovirus receptor gene. Hum. Genet. 105, 354-359. 37. Freimuth, P., Springer, K., Berard, C., Hainfeld, J., Bewley, M., and Flanagan, J. (1999). Coxsackievirus and adenovirus receptor amino-terminal immunoglobulin V-related domain binds adenovirus type 2 and fiber knob from adenovirus type 12. / . Virol. 73, 1392-1398. 38. Wang, X., and Bergelson, J. M. (1999). Coxsackievirus and adenovirus receptor cytoplasmic and transmembrane domains are not essential for coxsackievirus and adenovirus infection. / . Virol. 73, 2559-2562. 39. Rosengart, T. K., Lee, L. Y., Patel, S. R., Kligfield, P. D., Okin, P. M., Hackett, N. R., Isom, O. W., and Crystal, R. G. (1999). Six-month assessment of a phase I trial of angiogenic gene therapy for the treatment of coronary artery disease using direct intramyocardial administration of an adenovirus vector expressing the VEGF121 cDNA. Ann. Surg. 230, 466-472. 40. Kibbe, M. R., Murdock, A., Wickham, T., Lizonova, A., Kovesdi, I., Nie, S., Shears, L., Billiar, T. R., and Tzeng, E. (2000). Optimizing cardiovascular gene therapy. Arch. Surg. 135,191-197. 41. Hidaka, C , Milano, E., Leopold, P. L., Bergelson, J. M., Hackett, N. R., Finberg, R. W., Wickham, T. J., Kovesdi, L, and Roelvink, P. (1999). CAR-dependent and CAR-independent pathways of adenovirus vector-mediated gene transfer and expression human fibroblasts. / . Clin. Invest. 103, 579-587. 42. Leon, R. P., Hedlund, T., Meech, S. J., Li, S., Schaack, J., Hunger, S. P., Duke, R. C , and DeGregori, J. (1998). Adenoviral-mediated gene transfer in lymphocytes. Proc. Natl. Acad. Sci. USA 95, 13159-13164. 43. Huang, S., Stupack, D. G., Mathias, P., Wang, Y., and Nemerow, G. (1997). Growth arrest of Epstein-Barr virus immortalized B lymphocytes by adenovirus-delivered ribozymes. Proc. Natl. Acad. Sci. USA 94, 8156-8161. 44. Pickles, R. J., McCarty, D., Matsui, H., Hart, P. J., Randell, S. H., and Boucher, R. C. (1998). Limited entry of adenovirus vectors into well-differentiated airway epithelium is responsible for inefficient gene transfer. / . Virol. 72, 6014-6023. 45. Zabner, J., Freimuth, P., Puga, A., Fabrega, A., and Welsh, M.J. (1998). Lack of high affinity fiber receptor activity explains the resistance of ciliated airway epithelia to adenovirus infection./. Clin. Invest. 100, 1144-1149. 2. Biology of Adenovirus Cell Entry 3 5 46. Goldman, M., Su, Q., and Wilson, J. M. (1996). Gradient of RGD-dependent entry of adenoviral vector in nasal and intrapulmonary epithelia: Implications for gene therapy of cystic fibrosis. Gene Ther. 3, 811-818. 47. Carson, S. D., Hobbs, J. T., Tracy, S. M., and Chapman, N. M. (1999). Expression of the coxsackievirus and adenovirus receptor in cultured human umbilical vein endothelial cells: Regulation in response to cell density. / . Virol. 73, 7077-7079. 48. Schmidt, M. R., Pickos, B., Cabatingan, M. S., and Woodland, R. T. (2000). Expression of a human coxsackie/adenovirus receptor transgene permits adenovirus infection of primary lymphocytes./. Immunol. 165, 4112-4119. 49. Bewley, M. C , Springer, K., and Zhang, Y.-B. (1999). Structural analysis of the mechanism of adenovirus binding to its human cellular receptor, CAR. Science 286, 1579-1583. 50. Xia, D., Henry, L. J., Gerard, R. D., and Deisenhofer, J. (1994). Crystal structure of the receptor-binding domain of adenovirus type 5 fiber protein at 1.7 A resolution. Structure 2, 1259-1270. 51. Roelvink, P. W., Lizonova, A., Lee, J. G. M., Li, Y., Bergelson, J. M., Finberg, R. W., Brough, D. E., Kovesdi, L, and Wickham, T.J. (1998). The coxsackievirus-adenovirus receptor protein can function as a cellular attachment protein for adenovirus serotypes from subgroups A, C, D, E, and F. / . Virol 71, 7909-7915. 52. Roelvink, P. W., Lee, G. M., Einfeld, D. A., Kovesdi, L, and Wickham, T. J. (1999). Iden tification of a conserved receptor-binding site on the fiber proteins of CAR-recognizing adenoviridae. Science 286, 1568-1571. 53. Hong, S. S., Karayan, L., Tournier, J., Curiel, D. T., and Boulanger, P. A. (1998). Adenovirus type 5 fiber knob binds to MHC class I a2 domain at the surface of human epithelial and B lymphoblastoid cells. EMBO J. 16, 2294-2306. 54. Dechecchi, M. C , Tamanini, A., Bonizzato, A., and Cabrini, G, (2000). Heparan sulfate glycosaminoglycans are involved in adenovirus type 5 and 2-host cell interactions. Virology 268, 382-390. 55. Belin, M.-T., and Boulanger, P. (1993). Involvement of cellular adhesion sequences in the attachment of adenovirus to the HeLa cells surface. / . Gen. Virol. 74, 1485-1497. 56. Stevenson, S. C , Rollence, M., White, B., Weaver, L., and McClelland, A. (1995). Human adenovirus serotypes 3 and 5 bind to tv^o different cellular receptors via the fiber head domain./. Virol. 69, 2850-2857. 57. Di Guilmi, A.M., Barge, A., Kitts, P., Gout, E., and Chroboczek, J. (1995). Human adenovirus serotype 3 (Ad3) and the Ad3 fiber protein bind to a 130-kDa membrane protein on HeLa cells. Virus Res. 38, 71 -81 . 58. Shayakhmetov, D. M., and Lieber, A. (2000). Dependence of adenovirus infectivity on length of the fiber shaft domain. / . Virol. 74, 10,274-10,286. 59. Huang, S., Reddy, V., Dasgupta, N., and Nemerow, G. R. (1999). A single amino acid in the adenovirus type 37 fiber confers binding to human conjunctival cells. / . Virol. 73, 2798-2802. 60. Arnberg, N., Edlund, K., Kidd, A. H., and Wadell, G. (2000). Adenovirus type 37 uses sialic acid as a cellular receptor. / . Virol. 74, 42-48 . 61. Wu, E., Fernandez, J., Fleck, S. K., Von Seggern, D. J., Huang, S., and Nemerow^, G. R. (2000). A 50 kDa membrane protein mediates sialic acid-independent binding and infection of conjunctival cells by adenovirus type 37. Virology 279, 7S-89. 62. Krasnykh, V. N., Mikheeva, G. V., Douglas, J. T., and Curiel, D. T. (1996). Generation of recombinant adenovirus vectors with modified fibers for altering viral tropism. / . Virol. 70, 6839-6846. 63. Gall, J., Kass-Eisler, A., Leinwand, L., and Falck-Pedersen, E. (1996). Adenovirus type 5 and 7 capsid chimera: Fiber replacement alters receptor tropism without affecting primary immune neutralization epitopes./. Virol. 70, 2116-2123. 3 6 Glen R. Nemerov^ 64. Von Seggern, D. J., Huang, S., Fleck, S. K., Stevenson, S. C , and Nemerow, G. R. (2000). Adenovirus vector pseudotyping in fiber-expressing cell lines: Improved transduction of Epstein-Barr virus-transformed B cells. / . Virol. 74, 354-362. 65. Chillon, M., Bosch, A., Zabner, J., Law, L., Armentano, D., Welsh, M. J., and Davidson, B. L. (1999). Group D Adenoviruses infect primary central nervous system cells more efficiently than those from Group C. / . Virol. 73, 2537-2530. 66. Shayakhmetov, D. M., Papayannopoulou, T., Stamatoyannopoulos, G., and Lieber, A. (2000). Efficient gene transfer into human CD34+ cells by a retargeted adenovirus vector./. Virol. 74, 2567-2584. 67. Pereira, H. G. (1958). A protein factor responsible for the early cytopathic effect of aden oviruses. Virology 6, 601-611. 68. Everett, S. F., and Ginsberg, H. S. (1958). A toxin-like material separable from type 5 adenovirus particles. Virology 6, 770-771. 69. Wadell, G., and Norrby, E. (1969). Immunological and other biological characteristics of pentons of human adenoviruses./. Virol. 4, 671-680. 70. Neumann, R., Chroboczek, J., and Jacrot, B. (1998). Determination of the nucleotide sequence for the penton-base gene of human adenovirus type 5. Gene 69, 153-157. 71. Mathias, P., Wickham, T. J., Moore, M., and Nemerow, G. (1994). Multiple adenovirus serotypes use av integrins for infection. / . Virol. 68, 6811-6814. 72. Cuzange, A., Chroboczek, J., and Jacrot, B. (1994). The penton base of human adenovirus type 3 has the RGD motif. Gene 146, 257-259. 73. Albinsson, B., and Kidd, A. H. (1999). Adenovirus type 41 lacks an RGD av-integrin binding motif on the penton base and undergoes delayed uptake in A549 cells. Virus Res. 64, 125-136. 74. Bai, M., Harfe, B., and Freimuth, P. (1993). Mutations that alter an Arg-Gly-Asp (RGD) sequence in the adenovirus Type 2 penton base protein abolish its cell-rounding activity and delay virus reproduction in flat cells. / . Virol. 67, 5198-5205. 75. Von Seggern, D. J., Chiu, C. Y., Fleck, S. K., Stewart, P. L., and Nemerow, G. R. (1999). A helper-independent adenovirus vector with El , E3, and fiber deleted: Structure and infectivity of fiberless particles./. Virol. 73, 1601-1608. 76. Bader, B. L., Rayburn, H., Crowley, D., and Hynes, R. O. (1998). Extensive vasculogenesis, angiogenesis, and organogenesis precede lethality in mice lacking all av integrins. Cell 95, 507-519. 77. Huang, X., Griffiths, M., Wu, J., Farese, R. V., Jr., and Sheppard, D. (2000). Normal Development, wound healing, and adenovirus susceptibility in ^5-deficient mice. Mol. Cell. Biol. 20, 755-759. 78. Vigne, E., Mahfouz, I., Dedieu, J. -F., Brie, A., Perricaudet, M., and Yeh, P. (1999). RGD inclusion in the hexon monomer provides adenovirus type 5-based vectors with a fiber knob-independent pathway for infection. / . Virol. 73, 5156-5161. 79. Wickham, T. J., Carrion, M. E., and Kovesdi, I. (1995). Targeting of adenovirus penton base to new receptors through replacement of its RGD motif with other receptor-specific peptide motifs. Gene Ther. 2, 756. 80. Stewart, P. L., Chiu, C. Y., Huang, S., Muir, T., Zhao, Y., Chait, B., Mathias, P., and Nemerow, G. R. (1997). Cryo-EM visualization of an exposed RGD epitope on adenovirus that escapes antibody neutralization. EMBO J. 16, 1189-1198. 81. Mathias, P., Galleno, M., and Nemerow, G. R. (1998). Interactions of soluble recombinant integrin avP5 with human adenoviruses./. Virol. 72, 8669-8675. 82. Chiu, C. Y., Mathias, P., Nemerow, N. R., and Stewart, P. L. (1999). Structure of adenovirus complexed with its internalization receptor./. Virol. 73, 6759-6768. 83. Acharya, R., Fry, E., Stuart, D., Fox, G., Rowlands, D., and Brown, F. (1989). The three- dimensional structure of foot-and-mouth disease virus at 2.9 A resolution. Nature 337, 709-716. 2. Biology of Adenovirus Cell Entry 3 7 84. Stupack, D., Li, E., Sillett, S. A., Kehler, J. A., Geahlen, R. L., Hahn, K., Nemerow, G. R., and Cheresh, D. A. (1999). Matrix valency regulates integrin-mediated lymphoid adhesion via syk kinase. / . Cell Biol. 144, 777-788. 85. Maheshwari, G., Brown, G., Lauffenburger, D. A., Wells, A., and Griffith, L. G. (2000). Cell adhesion and motility depend on nanoscale RGD clustering. / . Cell Set. 113, 1677-1686. 86. Calderv^ood, D. A., Shattil, S. J., and Ginsberg, M. H. (2000). Integrins and actin filaments: Reciprocal regulation of cell adhesion and signaling. / . Biol. Chem. 275, 22,607-22,610. 87. Clark, E. A., and Brugge, J. S. (1995). Integrins and signal transduction pathv^ays: The road taken. Science 268, 233-239. 88. Hynes, R. O. (1992). Integrins: Versatility, modulation, and signaling in cell adhesion. Cell 69, 11-25. 89. Giancotti, F. G., and Ruoslahti, E. (1999). Integrin signaling. Science 285, 1028-1032. 90. Isberg, R. R., and Tran Van Nhieu, G. (1995). The mechanism of phagocytic uptake promoted by invasion-integrin interaction. Trends Cell Biol. 120, 120-124. 91. Gavrilovskaya, I. N., Shepley, M., Shav^, R., Ginsberg, M. H., and Mackov^, E. R. (1998). p3 integrins mediate the cellular entry of hantaviruses that cause respiratory failure. ?roc. Natl. Acad. Sci. USA 95, 7074-7079. 92. Witke, W., Podtelejnikov, A. V., DiNardo, A., Sutherland, J. D., Gurniak, C. B., Dotti, C , and Mann, M. (1998). In mouse brain profilin I and profilin II associate with regulators of the endocytic pathway and actin assembly. EMBO J. 17, 967-976. 93. Rodriguez-Viciana, P., Warne, P. H., Khwaja, A., Marte, B. M., Pappin, D., Das, P., Waterfield, M. D., Ridley, A., and Downward, J. (1997). Role of phosphoinositide 3-OH kinase in cell transformation and control of the actin cytoskeleton by Ras. Cell 89, 457-467. 94. Ireton, K., Payrastre, B., Chap, H., Ogawa, W., Sakaue, H., Kasuga, M., and Cossart, P. (1996), A role for phosphoinositide 3-kinase in bacterial invasion. Science 274, 780-782. 95. Rauma, T., Tuukkanen, J., Bergelson, J. M., Denning, G., and Hautala, T. (1999). rab5 GTPase regulates adenovirus endocytosis./. Virol. 73, 9664-9668. 96. Li, E., Stupack, D. G., Brown, S. L., Klemke, R., Schlaepfer, D. D., and Nemerow, G. R. (2000). Association of pi30^^^ with phosphatidylinositol-3-OH kinase mediates adenovirus cell entry./. Biol. Chem. 14,729-14,735. 97. Hall, A. (1998). Rho GTPases and the actin cytoskeleton. Science 279, 509-514. 98. Hoffman, G. R., and Cerione R. A. (2000).
Flipping the switch: The structural basis for signaling through the CRIB motif. Cell 102, 403-406. 99. Vuori, K., Hirai, H., Aizawa, S., and Ruoslahti, E. (1996). Induction of pl30'^^' signaling complex formation upon integrin-mediated cell adhesion: A role for Src family kinases. Mole. Cell. Biol. 16, 2606-2613. 100. Bruder, J. T., and Kovesdi, I. (1997). Adenovirus infection stimulates the Raf/MAPK signaling pathway and induces interleukin-8 expression./. Virol. 71, 398-404. 101. Brooks, P. C , Klemke, R. L., Schon, S., Lewis, J. M., Schwartz, M. A., and Cheresh, D. A. (1997). Insulin-like growth factor receptor cooperates with integrin a vp5 to promote tumor cell dissemination in vivo./. Clin. Invest. 99, 1390-1398. 102. Tsakiridis, T., Taha, C , Grinstein, S., and Klip, A. (1996). Insulin activates a p21-activated kinase in muscle cells via phosphatidylinositol 3-kinase./. Biol. Chem. 271,19,664-19,667. 103. Guo, D., and Donner, D. B. (1996). Tumor necrosis factor promotes phosphorylation and binding of insulin receptor substrate 1 to phosphatidylinositol 3-kinase in 3T3-L1 adipocytes. / . Biol. Chem. 271, 615-618. 104. Li, E., Stupack, D. G., Brown, S. L., Klemke, R., Schlaepfer, D. D., and Nemerow, G. R. (2000). Signaling antibodies complexed with adenovirus circumvent CAR and integrin interactions and improve gene delivery. Gene Therapy 7, 1593-1599. 105. Zsengeller, Z., Otake, K., Hossain, S. -A., Berclaz, P. -Y., and Trapnell, B. C. (2000). Internalization of adenovirus by Alveolar macrophages initiates early proinflammatory signaling during acute respiratory tract infection. / . Virol. 74, 9655-9667. 3 8 Glen R. Nemerov^ 106. Seth, P., Wilingham, M. C , and Pastan, I. (1985). Binding of adenovirus and its external proteins to triton X-114. / . Biol. Chem. 260, 14,431-14,434. 107. Wickham, T. J., Filardo, E. J., Cheresh, D. A., and Nemerow, G. R. (1994). Integrin avp5 selectively promotes adenovirus mediated cell membrane permeabilization. / . Cell Biol. 127, 257-264. 108. Seth, P., Willingham, M. C , and Pastan, I. (1984). Adenovirus-dependent release of ^^Cr from KB cells at an acidic pH. /. Biol. Chem. 259, 14,350-14,353. 109. Seth, P., Pastan, I., and Willingham, M. C. (1985). Adenovirus-dependent increase in cell membrane permeability./. Biol. Chem. 260, 9598-9602. 110. Ichimura, T., Hatae, T., and Ishida, T. (1997). Direct measurement of endosomal pH in living cells of the rat yolk sac epithelium by laser confocal microscopy. Eur. J. Cell Biol. 74, 41-48 . 111. Fender, P., Ruigrok, R. W. H., Gout, E., Buffet, S., and Chroboczek, J. (1997). Adenovirus dodecahedron, a new vector for human gene transfer. Nat. Biotech. 15, 52-56. 112. Hannan, C., Raptis, L. H., Dery, C. V., and Weber, J. (1983). Biological and structural studies with an adenovirus type 2 temperature-sens mutant defective for uncoating. Intervirology 19,213-223. 113. Gotten, M., and Weber, J. M. (1995). The adenovirus protease is required for virus entry into host cells. Virology 213, 494-502. 114. Greber, U. F., Webster, P., Helenius, A., and Weber, J. (1996). The role of the adenovirus protease in virus entry into cells. EMBO } . 15, 1766-1777. 115. Sodeik, B., Ebersold, m. W., and Helenius, A. (1997). Microtubule-mediated transport of incoming herpes simplex virus 1 capsids the nucleus. / . Cell Biol. 10, 136-1007. 116. Matthews, D. A., and Russell, W. G. (1998). Adenovirus core protein V interacts with p32 — a protein which is a associated both the mitochondria and the nucleus. / . Gen. Virol. 79, 1677-1685. C H A P T E R Adenovirus Replication Jared D. Evans and Patrick Hearing Department of Molecular Genetics and Microbiology State University of New York School of Medicine Stony Brook, New York I. Introduction Since their discovery, adenoviruses have served the scientific community as a pov\^erful tool for research of important virological as v^ell as cellular events. Adenoviruses w êre first isolated as a result of researchers pursuing the causative agent of the common cold. Rowe and colleagues, in 1953, observed cytopathic effect in primary cell cultures derived from human adenoids [1]. The follov^ing year, the same effect w âs seen in cells exposed to respiratory secretions by Hilleman and Werner, v^ho w êre trying to uncover the cause of acute respiratory disease in Army recruits [2]. It w âs later shov^n that adenoviruses, so named after its source of origin, w êre not the etiologic agent of the common cold, since they cause practically no respiratory morbidity among the general population. Hov^ever, adenovirus (Ad) has been show^n to cause severe respiratory distress in immunocompromised individuals [3]. Ad infection can also result in epidemic conjunctivitis [4] as v^ell as a number of other syndromes, including gastroenteritis [5]. These infections are usually resolved quickly, resulting in lifelong immunity to the virus. The adenovirus family is a large one, containing members that can infect a wide range of animals, including monkeys, livestock, mice, and birds as w êll as humans. All of these viruses consist of a naked icosahedral protein shell (70-100 nm in diameter) that encapsidates a linear, double-stranded DNA molecule. The exact dimension of the virion particle and size of the Ad genome can differ quite greatly betv^een adenoviruses that infect different species. Less than 10 years after their initial discovery, it v^as seen that adenovirus serotype 12 (Adl2) could cause malignant tumors in infected nev\rborn ham sters [6]. This seminal finding by Trentin and colleagues v^as the first evidence ADENOVIRAL VECTORS FOR GENE THERAPY J Q Copyright 2002, Elsevier Science (USA). All rights reserved. 4 0 Evans and Hearing that a human virus could induce cellular transformation. The fact that the transformation occurred in vivo and produced disease made the finding even more profound. Hov\^ever, to date, there has been no significant evidence that w^ould implicate adenovirus in oncogenesis in human beings. Trentin's discov ery thrust adenovirus into the forefront of model systems in the study of cancer as well as basic cellular processes. Adenovirus proved a worthy experimental system due to the ability to grow the virus to high titers in vitro as well as infect a wide variety of cell types. The relative safety and ease with which adenovirus and its genome can be manipulated also make Ad an attractive tool to study basic virology as well as cellular responses to viral infection. Discoveries in adenovirus research has provided a greater understanding of viral and cellular gene expression, DNA replication, cell cycle control, and cellular transformation. A notable example of the impact that the study of adenovirus has had on the scientific field is the discovery of mRNA splicing. It was shown that adenovirus produces a number of mRNAs from a single large transcript [7]. The analysis of the structure of mRNAs by Sharp and colleagues effectively revealed the existence of introns. The existence of splicing sites was then observed. From this finding, it was possible to dissect cellular mRNAs to show the presence of introns and the function of splicing in eukaryotes. II. Classification The family to which adenoviruses belong, Adenoviridiae, is divided into two genera: Mastadenovirus and Aviadenovirus. The Mastadenovirus genus contains viruses that infect a wide range of mammalian species, including human, simian, bovine, ovine, equine, porcine, and opposum. The Aviaden ovirus group infects only bird species (i.e., chicken and turkey). The viruses are classified into six subgroups based on two different criteria: percentage of guanine-cytosine in the DNA molecule and the ability to agglutinate red blood cells [8]. Within these groups are the serotypes of adenovirus. To date, human adenoviruses have been further subdivided into >50 specific types, primarily on the basis of neutralization assays. Type-specific neutralization occurs by antibodies binding the virion capsid hexon protein and, to a much lesser extent, the capsid fiber and penton proteins. IIL Genome Organization The human Ad genome is present as a linear double-stranded DNA molecule approximately 35-36 kbp in length. The genome is contained within the capsid in a highly condensed form, associated with viral proteins V and 3 . Adenov i rus Replication 4 1 VII. These proteins organize the DNA into a nucleosome-hke structure known as the core. The core is tethered to the capsid through the interaction of protein V with protein VI, a protein associated with internal facets of the capsid. The Ad rephcation origins are present in the first 50 base pairs of the '^lOO-bp inverted terminal repeats (ITRs) located at each end of the viral genome. The inverted nature of the ITRs plays a functional role in viral DNA replication (discussed below). A terminal protein is covalently attached to each 5̂ terminus of the viral genome. This protein, along with the Ad DNA polymerase. Ad single-strand DNA binding protein, and cellular factors, are essential for viral DNA replication. A ds-acting packaging sequence is located at one end of Ad genome, conventionally called the left end, which directs the polar encapsidation of the viral DNA into the capsid. The Ad chromosome contains one immediate-early region (ElA), four early transcription units (ElB, E2, E3, and E4), two "delayed" early units (IX and IVa2), and one late unit (major late) that produces five families of mRNAs (LI to L5) (Fig. 1). All of the viral transcription units utilize cellular RNA polymerase II for their transcription. Ad-encoded regulatory proteins participate in the specificity of the transcription program. The viral genome also contains at least one gene that codes for VA RNA (some Ad serotypes have two) which is transcribed by RNA polymerase III. The schematic representation of the genome (Fig. 1) is conventionally drawn with the ElA transcription unit at the left end, adjacent to the packaging sequences (\\|;). The transcription units of the Ad genome are transcribed from both strands of the chromosome: ElA, ElB, pIX, the major late transcription unit, VA RNA, and E3 are transcribed using the rightward reading strand and E4, E2, and IVa2 are transcribed using the leftward reading strand. The Ad genome is an excellent example of the need for viruses to efficiently use limited genetic space and information to produce the maximum number of proteins necessary for virus propagation. In the case of adenovirus, the host cell's RNA producing/processing machinery is manipulated to the advantage of the virus. It would appear that the viral genome is organized in IX ^ L ^^^ ^^ . 19 " -S. L4 ,̂ L5., ^ l > l > MLP ' ^ IM^ I ~ I I I I I I I l ~ 0 10 20 30 40 50 60 70 80 90 . < H E2b E2a E4 IVa2 Figure 1 Schematic representation of the human adenovirus genome. Black arrows depict imme diate-early, early, and delayed-early genes and hatched arrows depict the late genes. The inverted terminal repeats are labeled ITR and the packaging sequence is denoted as \};. MLP corresponds to the major late promoter. 4 2 Evans and Hearing its current fashion as a result of evolution determining the most functionally prudent structure and order. It would also seem obvious that evolution has selected for particular grouping of RNAs, since the majority of Ad transcription units produce proteins with related functions. The grouping of related proteins within the same transcription unit might indicate that replication of the virus requires a logical, stepwise progression of gene expression in order to usurp control of the cellular machinery to direct the efficient production of virus. IV. Virus Infection The primary targets of adenovirus infection are the terminally differenti ated epithelial cells of the upper respiratory tract, gut, and eye. However, it has been shown that adenovirus can infect almost any cell type. Adenovirus entry into a cell is discussed in detail in chapter 2 of this volume. Briefly, adenovirus binds to a cell via a cell surface receptor known as CAR (Coxsackie and adenovirus receptor) through interaction of CAR with the fiber protein [9]. The penton base portion of the fiber structure contains an ROD amino acid sequence that binds integrins on the cell surface. The integrins act as corecep- tors for viral entry. Integrins are not essential for attachment of the virus to the cell, but are necessary for gaining access to the interior of the cell [10]. Virus infection can be blocked by the presence of excess RGD peptides. Recently, it was shown that at least one adenovirus serotype, Ad37, can utilize sialic acid to enter cells [11]. Both CAR and sialic acid are expressed on most, if not all, cell types, which may explain the ability of adenovirus to infect
a wide variety of cells. Once bound to its receptor, adenovirus is internalized via receptor mediated endocytosis in clathrin coated pits. The adenovirus can be visualized in endosomes shortly after infection. The low pH of the endosome facilitates release of the Ad particles that move to the nucleus, apparently via microtubule transport. During transport, the viral capsid is partially degraded, allowing the genome to be inserted into the nucleus through the nuclear pore complexes. Inside the nucleus, the genome positions itself adjacent to particular nuclear organelles through attachment of the terminal protein to the nuclear matrix. This attachment is believed to position the genome in a manner that makes it available for early gene expression and viral DNA replication. V. Early Gene Expression Once the viral genome has entered the nucleus, Ad early gene expression is directed toward achieving three main objectives. First, the host cell must be stimulated to enter S phase of the cell cycle to provide the correct intracellular 3. Adenovirus Replication 4 3 environment for viral replication. Second, the infected cell must be protected from the anti-viral host response to virus infection both from v^ithin the cell and due to the extracellular immune response. Third, viral gene products are produced to be used in concert v^ith cellular proteins to carry out the viral DNA replication program. VI. Early Region l A ( E I A ) Adenovirus encodes over 25 individual early gene products. The early genes are expressed in a temporal and coordinated manner. The first early region expressed after Ad infection is the immediate-early transcription unit ElA since it requires only cellular proteins for its expression. The ElA gene products in turn activate transcription from the other early promoter regions. The ElA gene is composed of tw ô exons and several ElA polypeptides are produced follow^ing alternative splicing of a primary RNA transcript (Fig. 2). The most abundant of the ElA proteins early after infection are referred to as the ElA 12S (243 amino acids) and 13S (289 amino acids) gene products based on the mRNAs that encode them. The ElA 12S and 13S proteins act as major regulators of early viral transcription as well as important modulators of host cell gene expression and proliferation (review^ed in [12-14]. The ElA 12S and 13S proteins share tv^o conserved regions within the 5' exon referred to as CRl and CR2. The two proteins differ only in a 46-residue internal exon segment present in the 13S protein. This region, referred to as conserved region 3 (CR3), is important for the transcriptional transactivation properties of the ElA 13S protein. Both proteins are localized to the nucleus due to a carboxy-terminal Rb/p107/p 130 p400 PCAF TBR ATF, pSOO/CBP Srb CtBP E1A13S [ •jdsIH RsS^HSslSifr^ 1 PRQaa 1 40 80 121 140 185 186 289 E1A12S [ •riSH MSHg-—— 1 ?43aa 1 139 Transformation Apoptosis, Txl repression Txl transactivation Suppression transformation Figure 2 Functional map of El A proteins. The coding sequences of the 12S and 13S El A proteins are shown with conserved regions depicted (CRl, CR2, and CR3). Binding sites for cellular proteins are indicated by bars along with E lA functional activities. 4 4 Evans and Hearing nuclear localization sequence (NLS). The ElA gene products exert their effects by interactions with numerous cellular proteins, most of which are involved in transcriptional regulation [12, 14]. The ElA products interact with a number of important cellular proteins (Fig. 2), including: (1) the retinoblastoma tumor suppressor, pRb, and related family members p l07 and pl30 via CRl and CR2; (2) transcriptional coactivators p300 and CBP via CRl and amino terminal sequences; (3) a number of transcription factors such as TATA- binding protein (TBP), members of the ATF family (e.g., ATF-2, Spl, and c-Jun), and the Srb/mediator complex via CR3 and CtBP via the C-terminus. Rb family members repress the activity of the E2F family of transcription factors, among numerous binding partners (reviewed in [13]). p300/CBP have histone acetyl transferase (HAT) activity and play a role in chromatin remod eling (reviewed in [15]). The ElA 13S product is the major transcriptional activator of viral early gene expression and mediates its function principally through the CR3 domain that acts as a powerful modulator of other proteins involved in transcription [12, 14]. The ElA 13S protein may increase tran scription through stabilization of the transcription factor complex TFIID (via interaction with TBP) at viral and cellular promoter regions. ElA also may increase transcription through stimulation of specific transcription factors. For example, ElA binding to ATF-2 may result in a conformational change result ing in transcriptional activation (reviewed in [16]). Finally, ElA binds to CtBP, a transcriptional corepressor; CtBP binding correlates with ElA suppression of transformation [17]. The most well characterized instance of ElA activation of gene expression involves the E2F family of transcription factors. The E2F family of proteins was initially discovered through studies of Ad E2 promoter regulation [18]. The Ad E2 early promoter contains binding sites for both ATF and E2F transcription factors. E2F transcription factors play a major role in the expression of cellular genes important for the regulation of cell cycle progression [13]. E2Fs exist in the cell as heterodimers containing one of six identified E2F proteins with one of two DP molecules. E2Fs both positively and negatively regulate gene expression. As repressors, E2Fs are bound to DNA at specific sites in complexes with members of the Rb tumor suppressor family (pRb, p l07, pl30). Rb family members interact with histone deacetylase complexes (HDACs), which repress the activity of promoter regions via deacetylation of histones and other promoter-bound transcription factors [19, 20]. Specific members of the Rb family bind to different E2F complexes, determined by the member of the E2F present. E2F binding to Rb members involves the large binding pocket domain of Rb family members which is also the target for ElA protein binding. In uninfected cells, E2F is negatively regulated by binding to Rb family members. Rb family binding to E2Fs is controlled through phosphorylation by cyclin-dependent kinases (cdks) [13]. The hyperphosphorylation of Rb family proteins by Cdks in Gl phase of the cell cycle results in dissociation of Rb from 3. Adenovirus Replication 4 5 the E2F complexes, and derepression of E2F responsive genes. The activation of E2F complexes results in the promotion of Gl and S phase progression. El A acts to subvert the tight control of E2F by binding directly to Rb proteins and sequestering them, freeing E2F heterodimers to activate viral and cellular gene expression [13]. E2Fs activate transcription by the recruitment of HATs to promoter regions [21, 22]. Both the ElA 12S and 13S products direct the release of Rbs from E2Fs, and both ElA proteins are capable of stimulating Ad E2a transcription [12, 14]. As stated, the primary targets of adenoviruses are terminally differentiated epithelial cells. As such, these cells are generally quiescent v^ith low metabolic activity. It is for these reasons that the virus must pressure the infected cell into S phase of the cell cycle in order to augment viral macromolecular synthesis. The interaction of ElA v^ith p300/CBP or Rb family proteins is sufficient to stimulate cellular DNA synthesis (reviev^ed in [23]). It appears that the increase in DNA synthesis may be due, in part, to the activation of E2F transcription through ElA-Rb interactions. The ability of ElA to foster unscheduled DNA synthesis also contributes to its oncogenic potential. Almost all adenoviruses are capable of transforming cells in culture and this ability is primarily attributed to ElA. The regions of ElA responsible for transformation and its oncogenic potential are involved in the binding of p300 and Rb family of tumor suppressor proteins. It appears as though ElA's ability to induce S phase is directly responsible for its ability to cause transformation [23]. ElA also plays a role in the induction of apoptosis in infected cells (reviev^ed in [24, 25]). It has been shown that ElA causes an increase in the level of the tumor suppressor p53. The rise in p53 levels is a result of the stabilization of this usually labile protein by ElA. The presence of p53 is a major obstacle to efficient lytic infection by adenovirus. One function of p53 is to protect the genomic integrity of the cell. Unscheduled DNA synthesis, such as adenovirus DNA production, causes activation of p53. Activated p53 induces gene expression by binding specific promoter sequences, which activates genes that are involved in a number of cellular processes. The presence of p53 can affect cells in primarily two ways [24, 25]. First, p53 can induce Gl arrest, thus inhibiting progression of the cell cycle. This arrest can be facilitated by the transactivation of a gene encoding an inhibitor of Cdks, termed p21^^P~i/^^P-i, which prevents the phosphorylation of Rb family members. p53 also can induce cell death by apoptosis. It does so by inducing the activation of degradative enzymes, caspases, which generate the classic apoptotic pathway. This proteolytic cascade results in a characteristic apoptotic phenotype of shrinkage and rounding of the cell due to breakdown of the cytoskeleton, cleavage of cellular DNA and condensation of the chromatin, cytoplasmic vacuolization and membrane blebbing, and in the final stages, fragmentation of the cell membrane into vesicles or apoptotic bodies that can be taken up by neighboring cells. The activation of p53 and induction 4 6 Evans and Hearing of cellular apoptosis at this stage of infection would be quite deleterious to the virus replication program. Therefore, adenovirus has evolved several mechanisms to decrease or inhibit p53 activity (discussed below). Recently, El A has been shown to suppress p53 transactivation [26]. El A causes the activation of pi9^^^, which leads to the upregulation and stabilization of p53 [27]. El A represses p53 transcriptional activation through the binding and sequestration of p300/CBP, coactivators required for p53-dependent gene expression [28]. VII. Early Region IB(EIB) The second El gene expressed is early region IB that leads to the production of two major species of mRNAs. One mRNA codes for a 19-kDa polypeptide (ElB 19K) and the other codes for a 55-kDa protein (ElB 55K). The two proteins are encoded by alternative reading frames and share no sequence homology. The major roles of these proteins in Ad infection are to inhibit apoptosis and further modify the intracellular environment in order to make the cell more hospitable to viral protein production and viral DNA replication [24, 25]. Viruses with mutations in either or both ElB proteins are significantly reduced in virus yield due to cell death by apoptosis prior to the completion of the replicative cycle. The ElB 55K protein is essential for a variety of important functions in the viral life cycle. One important function is inhibition of the p53 tumor suppressor and inhibition of the induction of p53-dependent apoptosis [24, 25]. The ElB 55K protein binds to the acidic transactivation domain of p53, thus inhibiting p53-induced transcription [29]. However, the binding of ElB 55K to p53 alone cannot inhibit p53 functions. It is theorized that ElB 55K directs repression of promoters when held in a complex with p53 due to strong transcriptional repression by ElB 55K [30]. By doing so, ElB 55K inhibits the activation of p53-responsive promoter regions and blocks cycle arrest and apoptosis programs before they get underway. It is not clear if the ElB 55K protein itself is a transcriptional repressor or it recruits a repressor to the p53-bound complex on DNA. ElB 55K acts in a complex with another early protein, E4 ORE6, which leads to the proteasome-dependent degradation of p53, further decreasing p53 effects on the infected cell [31]. ElB 55K also plays a very important role in producing a cellular environment conducive to viral protein production. A complex containing the ElB 55K protein and the E4 ORF6 product contributes to host cell protein synthesis shutoff by selectively stabiUzing and transporting viral mRNAs from the nucleus to the cytoplasm while inhibiting the transport of host cell mRNAs. This topic will be discussed further in the section on early region 4 (E4). 3. Adenovirus Replication 4 7 The ElB 19K protein is also involved in the inhibition of apoptotis. ElB 19K acts to block apoptotic pathv^ays that do not rely on p53, such as the TNF and Fas ligand cell death pathv^ays [24, 25]. ElB 19K is a functional homolog of a cellular
suppressor of apoptosis, Bcl-2. Homodimers of a pro- apoptotic protein, such as Bax, result in the activation of caspases, leading to cell death. Bcl-2 heterodimerizes v^ith Bax and inhibits its function, preventing the induction of apoptosis. The dimerization occurs through interaction of specific binding regions, Bcl-2 homology or BH domains. ElB 19K acts in the same manner as Bcl-2 by binding Bax and other apoptosis inducers. ElB 19K shares sequence similarity w îth Bcl-2 in tw ô BH domains present in ElB 19K that are necessary to bind Bax. The binding of Bax by ElB 19K leads to inhibition of apoptosis [24, 25]. ElB 19K also plays a role in inhibition of TNF-induced apopto sis by blocking the oligomerization of death-inducing complexes involving FADD [32]. FADD is a protein that is activated by binding Fas via death domains, thus its name (Fas-associated death domain). The exact function of ElB 19K in FADD regulation is not w êll understood. VIII. Early Region 2 (E2) The E2 transcription unit encodes the viral proteins involved in aden ovirus DNA replication: Ad DNA polymerase (Ad Pol), preterminal protein (pTP), and DNA binding protein (DBF). The E2 transcription unit is tran scribed from the E2 early promoter (E2A at genome coordinate l(i)^ v\̂ hich is activated by El A at early times after infection, and the E2 late promoter (E2B at genome coordinate 72), w^hich is activated at intermediate times after infection through an unknow^n mechanism. DBF is expressed by the E2A region (Fig. 1), v^hich shares common RNA leader sequences near genome coordinates IS and 68 v^ith mRNAs for pTP and Ad Pol. Ad Pol and pTP are encoded by the E2B region of the viral genome (Fig. 1) and their mRNAs share a common exon at genome coordinate 39. These short exons are spliced to the main body of the open reading frames (ORFs) for pTP and Ad Pol at genome coordinates 28.9 and 24.1, respectively. The E2 early promoter was shov\̂ n to possess four ds-acting elements that upregulated transcription of the gene: a TBP binding site, two E2F binding sites, and an ATE binding site (reviewed in [33]). The efficient transcription of the E2 early promoter is dependent on the ElAs via Rb binding and E2F derepression and by transactivation via TBP and ATEs. The binding of E2F/DP heterodimers to the E2 early promoter is stabilized by a product of the E4 transcription unit, E4-ORF6/7 (discussed below). The mechanism by which the E2 late promoter is delayed or transactivated is not known. DBF is a nuclear phosphoprotein of apparent molecular weight of 72 kDa that is produced in large quantities in an infected cell. The protein is from 4 8 Evans and Hearing 473 to 529 amino acids in length, according to Ad serotype, and is expressed throughout the infectious cycle. DBP is involved in a number of functions including viral DNA replication, early and late gene expression, host range, transformation, virion assembly, and possibly DNA recombination (reviewed in [34-36]. The N-terminal portion of DBP is highly phosphorylated and contains the NLS. The C-terminal domain is not phosphorylated, but it binds to DNA and is involved in viral DNA replication. DBP binds cooperatively to single-stranded DNA with high affinity and acts to protect the DNA from nuclease digestion. DBP possesses a helix- destabilizing property that is required for unwinding double-stranded DNA in an ATP-independent manner during the elongation phase of viral DNA replication by strand displacement [37]. DBP enhances renaturation of dis placed complementary strands [38]. DBP is also responsible for enhancement of the initiation of DNA replication by facilitating formation of the initiation complex pTP-dCMP as well as increasing NF-I/CTF binding to its recognition site in the auxiliary origin (see below). Finally, DBP increases the processivity of Ad Pol [39]. The pTP protein exists as a stable heterodimer with Ad-Pol and is critical for the initiation of viral DNA replication [34-36]. Ad-Pol catalyzes the covalent linkage of dCMP to serine 580 of pTP. The pTP-dCMP complex functions as the protein primer for Ad DNA synthesis. In adenovirus-infected cells, pTP in a 1:1 stoichiometric ratio with Ad Pol; both of these proteins are expressed at significantly lower levels than DBP. During the initiation of viral DNA replication, pTP binds to the core origin sequences in a phosphorylation- dependent manner. At late times in infection, the 80-kDa pTP is processed to the 55-kDa TP via cleavage by the virus-encoded protease [40]. The 55-kDa TP protein is covalently linked at the 5̂ ends of the genome in the mature virion. The processing of pTP-DNA to TP-DNA is not required for viral DNA replication or virion assembly, but it is necessary for full infectivity of mature virus particles. The presence of pTP or TP at the 5̂ termini may protect the viral genome from exonuclease digestion. Also, covalent attachment of pTP/TP to the genome has been shown to facilitate unwinding of the DNA duplex at the origin of replication. pTP (and TP) is responsible for the attachment of the adenovirus genome to the nuclear matrix [41, 42]. The interaction of pTP and a protein complex that directs pyrimidine biosynthesis known as CAD (carbamyl phosphate synthetase, aspartate transcarbamylase, and dihydroorotase) at active sites of viral DNA replication might suggest that the area in which the genome is deposited is predetermined due to the presence of proteins necessary for DNA synthesis [43]. Ad Pol is a 140-kDa phosphoprotein that is responsible for both the initiation and elongation steps of adenovirus DNA replication (reviewed in [34-36]). Ad Pol is localized to the nucleus via its association with pTP. Ad Pol is a member of a family of proteins known as the alpha DNA polymerases 3. Adenovirus Replication 4 9 and possesses five of the six regions conserved among other members. However, the regions necessary for Ad Pol activity are distributed over the entire length of the molecule and are not limited to the five regions of homology with the other DNA polymerases [44]. Ad Pol possesses an intrinsic 3̂ -^ 5' proofreading exonuclease activity as well as two potential zinc finger motifs important for its DNA binding and viral DNA replication initiation functions. Along with forming a stable heterodimer with pTP that is crucial for viral DNA replication, Ad Pol physically interacts with NF-I/CTF (a cellular factor involved in Ad replication, see below), and this interaction directs the Ad Pol-pTP complex to the origin of replication. Ad Pol is phosphorylated on serine residues, with serine 67 being the major site of phosphorylation. Phosphorylation appears to be important for Ad Pol to initiate replication [45]. The molecular and physical mechanism of Ad replication and the roles of viral and cellular proteins described in this section will be discussed below in the section on viral DNA replication. IX. Early Region 3 (E3) In order to conduct and successfully complete the infectious cycle, ade noviruses have evolved a number of mechanisms to evade the host antiviral defense array. Many of the proteins responsible for counteraction of the host immune response are encoded within the E3 region. The primary host defense against Ad infection is to eliminate the infected cell. The E3 region encodes multiple proteins that function to inhibit multiple pathways of cell death induced by the host innate and cellular immune responses to the infected cell (Fig. 3) (reviewed in [46, 47]). The E3 transcription unit is an early region located at genome position 76-86 whose transcription is induced by the El A 13S protein. The E3 promoter has a TATA box as well as upstream binding sites for the ATF, API, NFl transcription factors as well as NFKB [46]. It is believed that adenovirus-specific cytotoxic T lymphocytes (CTLs) are the major mechanism by which adenovirus-infected cells are eliminated. E3 encodes at least four proteins that are capable of inhibiting CTL killing [46, 47]. For CTL to destroy a virus-infected cell, the T-cell receptor must first recognize viral peptides presented on the cell surface in association with major histocompatibility complex (MHC) class I antigens. Adenovirus encodes a protein, E3 gpl9K, which prevents the transport of MHC I to the cell surface. E3 gpl9K is a membrane glycoprotein that localizes to the endoplasmic reticulum (ER), where it forms a complex with newly synthesized class I antigens, thus preventing their transport to the cell surface. When this protein is expressed, CTL killing of Ad-infected cells is greatly reduced. E3 gpl9K, and subsequently MHC I, are retained in the ER by an ER-retention signal 5 0 Evans and Hearing 3.6K 10.4K 14.7K \;^^m<;^ V/////////A 6.7K 7.5K 12.5K gp19K 11.6K 14.5K II 1 1 1 1 1 1 II V//////////////A kWWN 1 1 1 1 1 1 nt. 1000 2000 3000 gp19K: Integral membrane protein, inhibits kill by CTLs (blocks MHCI presentation) 10.4K (RIDa) and 14.5K (RIDp): Integral membrane protein, together form the RID complex to block FasL and TNF mediated apoptosis by degrading Fas and internalizing TNFR1 14.7K: Inhibits FasL and TNF-mediated apoptosis 11.6K (aka ADP): Integral membrane protein, promotes host cell death and virus release 12.5K, 6.7K and 3.6K: Functions unknown Figure 3 Schematic of the E3 transcription unit. The different proteins encoded by the E3 region are indicated by bars. The functions ascribed to different E3 proteins are listed below the diagram. (KKXX) found at the extreme C-terminus of E3 gpl9K [48]. E3 gpl9K binds to all MHC class I antigens, but with different affinities to which a hierarchy can be assigned [46, 47]. The E3 region also produces proteins that inhibit defenses involving ligand-receptor interactions and activation of cell death pathways [46, 47]. On the surface of most cells, receptors containing death domains (DDs) are expressed. Once CTL are activated, they kill via three main pathways. The primary mechanism of cell killing involves perforin and granzymes that act in concert: perforin forms holes in the target cell and granzymes are then introduced into the cytoplasm of the target cell. One of these enzymes, granzyme B, activates caspases to induce apoptosis. The second pathway involves the receptor Fas expressed on the surface of the infected target cell. The third pathway is mediated through the tumor necrosis factor (TNF) receptor type 1 (TNFRl). The ligands to these receptors are found on the surface of activated CTL: Fas ligand (FasL) and TNF, respectively. The interaction of ligand with receptor triggers a series of protein-protein interactions in the target cell that results in the induction of apoptosis (reviewed in [49]). Upon binding of FasL to Fas, the latter trimerizes and then binds a protein named FADD. This binding is facilitated though the "death domain" (DD), present in both Fas and FADD. FADD also has a "death effector domain" (DED) through which it associates with procaspase 8, thus causing autocleavage and activation of caspase 8. Activated caspase 8 cleaves and activates downstream caspases — a cascade that results in apoptosis. TNF binding to TNFR follows a similar pathway except TNFR binds the DD of TRADD (TNF-receptor-associated death domain) which then binds FADD 3. Adenovirus Replication 5 1 and caspase 8 as well as another DD-containing protein named RIP [50]. RIP is a serine/threonine kinase whose exact function is unclear. E3 RID (for receptor internalization and degradation) is an integral membrane protein composed of two E3 products RIDa (E3 10.4K) and RIDp (E3 14.5K). The RID complex localizes to the plasma membrane, Golgi apparatus, and ER. RID inhibits apoptosis through the TNERl and Fas pathways [51, 52]. Expression of RID leads to the clearance of Fas from the cell surface, which results in degradation in lysosomes. RID-mediated elimination of TNFRl is less efficient, and it is not known if TNFRl is degraded in lysosomes. RID also stimulates internalization of certain other receptors whose activation may result in a less direct inflammatory response. Once RID has deposited the receptor in the lysosome for destruction, RID is recycled back to the cell surface to repeat the internalization process. E3 14.7K also inhibits apoptosis induced by the cytokine activation of receptors. Its effects on TNF pathway is more dramatic than on Fas [53, 54]. Unlike many of the other E3 proteins, E3 14.7K is not associated with a membrane structure, but is present in the cytosol and nucleus. E3 14.7K acts by binding proteins involved in the apoptotic pathway. One protein is FIP-3, which also binds
RIP, a component of TNF-induced activation of NFKB [55]. NFKB activation appears to inhibit apoptosis; FIP3 may activate an apoptotic pathway in conjunction with inhibition of NFKB transactivation. Therefore, the binding of E3 14.7K to FIP-3 allows NFKB to induce transcription of genes that defend against the TNF signal. Also, the presence of NFKB sites in the E3 promoter may lead to the increased expression of E3 proteins [56]. It has also been reported that E3 14.7K may bind caspase 8 directly to inhibit the caspase/protease cascade. Both RID and E3 14.7K prevent TNF-induced release of arachidonic acid (AA). Cytosolic phospholipase Ai (cPLA2) is activated by TNF signaling which causes it to translocate to membranes and cleave phospholipids, producing AA. RID inhibits the translocation of cPLA2 to membranes. This action occurs prior to RID clearing TNFR from the cell surface [54]. The mechanism of E3 14.7K inhibition of AA release is not known. Both RID and E3 14.7K are required to inhibit inflammation and pathology in infected mouse lung. These Ad-encoded E3 proteins inhibit two of the three mechanisms utilized by CTL for cell killing. If Ad also inhibits perforin/granzyme lysis of cells is presently unknown. The E3 11.6K protein, also known as the adenovirus death protein (ADP), is an integral membrane protein localized to the Golgi and ER that is modified with complex oligosaccharides at a single N-linked site. This protein promotes cell death very late in the infectious cycle in order to release mature virions into the surrounding environment. Cells infected with an Ad mutant that does not express the E3 11.6K protein remain viable much longer than cells infected with wild-type adenovirus [57]. This action may seem contradictory to the functions of the other E3 proteins, but E3 11.6K is not produced in significant 52 Evans and Hearing amounts until very late stages of infection when virions accumulate aw^aiting release [58]. X. Early Region 4 (E4) Whether they are early or late, a common theme among the transcription units of adenovirus is that they encode multiple proteins of related functions. How^ever, early region 4 (E4) is the only transcription unit that produces proteins of relatively disparate functions. E4 encodes at least seven proteins according to analysis of ORFs and spliced mRNAs. The gene products exhibit a wide range of activities (Fig. 4). Proteins expressed from the E4 region have been shown to be important for viral DNA replication, viral mRNA transport and splicing, shutoff of host cell protein synthesis, and regulation of apoptosis. Viruses lacking the entire E4 region are extremely compromised for growth, decreased >5 logs in virus growth compared to wild-type Ad [59, 60]. Several of the proteins produced by E4 appear to be cytotoxic to cells. The E4 products cytotoxicity may influence the virus life cycle as well as the decision to include them in gene therapy vectors. The E4 region is transcribed in response to induction by El A. The E4 promoter is regulated to a certain extent by ATE sites, but expression depends more on two sites that bind a transcription factor termed E4E [12]. If El A is not present, the E4 transcription unit is still expressed, but to a much lower extent. The E4 ORFl 14.3-kDa protein is relatively uncharacterized. The E4 ORFl proteins of a number of Ad serotypes are capable of transforming 2 1 vy/////y////////////^^^^^^ ^mmmzzzzzzz 3000 2000 1000 1nt. ORF 1: Transformation, mammary tumors, binds PDZ proteins ORF 2: Function unknown ORF 3: Redundant with 0RF6, virus growth, disrupts PODS, persistence of transgene expression \n vivo ORF 4: Stimulates PP2A, induces apoptosis ORF 6: Redundant with ORF3, host cell shutoff, virus growth ORF 6/7: E2F induction Figure 4 Schematic of the E4 transcription unit. The different proteins encoded by the E4 region are indicated by bars. The functions ascribed to different E4 proteins are listed below the diagram. 3. Adenovirus Replication 5 3 primary rat cells in culture. In addition, the E4 ORFl protein of Ad9 induces mammary carcinomas in rats, independent of El A [61]. The E4 ORFl protein binds to a number of cellular proteins that possess a motif referred to as the PDZ domain, including the cellular dig tumor suppressor protein, and the binding of E4 ORFl to PDZ-containing proteins appears to mediate the oncogenic nature of this viral gene product [62]. The exact role that E4 ORFl plays in the viral replication cycle is currently being analyzed. Nothing is known about the role of the E4 ORF2 14.6-kDa protein in viral replication. This is also true of the E4 ORF3/4 7.1-kDa protein that is the product of a spliced mRNA that fuses the amino terminus of E4 ORF3 to the carboxy terminus of E4 ORF4. The E4 ORF3 11- to 14-kDa protein is expressed early in infection. E4 ORF3 is very tightly associated w îth the nuclear matrix. E4 ORF3 has been shown to have redundant function(s) with another E4 protein, E4 ORF6, with respect to virus growth and splicing of viral mRNAs [59, 60, 63]. A profound defect in Ad growth is observed with mutants that lack all of E4 coding sequences. However, if either E4 ORF3 or E4 ORF6 is expressed with an otherwise E4-deleted virus, growth capacity is restored to within 10-fold of wild type. Further, individual mutation of either the E4 ORF3 or E4 ORF6 proteins has only a modest impact on viral growth, whereas mutation of both protein reading frames results in a significant reduction in virus yield. Thus, either the E4 ORF3 or E4 ORF6 proteins are sufficient to confer the majority of E4 function in an Ad lytic infection in cultured cells. The E4 ORF3 and E4 ORF6 proteins both bind to the ElB 55K product, although to different ends. E4 ORF6 enhances the inhibition of p53 by ElB 55K, whereas E4 ORF3 transiently relieves the repression of p53 by ElB 55K [31, 64, 65]. Yet another function that E4 ORF3 has been proposed to have in common with E4 ORF6 is the ability to bind and inhibit the activity of DNA-protein kinase (DNA- PK), thus resulting in an inhibition of double strand break repair (DSBR) mechanism [66]. Ad DNA replication is likely to induce cellular DSBR. The binding of E4 ORF3 or E4 ORF6 proteins to DNA-PK appears to inhibit DSBR and block the formation of viral DNA concatamers that occurs in the absence of E4 expression. The formation of Ad DNA concatamers would block viral DNA replication and packaging of the genome into the capsid. E4 ORF3 has been shown to localize with discrete nuclear structures known as PODs, PML oncogenic domains, or NDlOs [67]. PODs exist as multiprotein complexes that exhibit a discreet, punctate appearance in the nucleus of an uninfected cell. E4 ORF3 is necessary and sufficient to cause redistribution of these protein complexes into long, track-like structures. PODs have been implicated in a number of cellular processes ranging from transcriptional regulation to the regulation of apoptosis (reviewed in [68]). PODs have also been shown to react to stresses such as heat shock and heavy metals as well as interferon, suggesting a role in cellular defense mechanisms. 5 4 Evans and Hearing A number of DNA viruses express proteins that function to disrupt PODs, i.e., herpesviruses, cytomegalovirus, and papillomavirus [69]. The exact function of PODs is still unknown, as is the purpose for POD reorganization by E4 ORF3, although it has been linked to adenovirus replication [67], E4 ORF3 is also capable of binding a number of other proteins, some of v^hich are involved in transcriptional regulation such as p300 and CBP (Evans and Hearing, unpublished results). Despite considerable research, the exact function(s) of E4 ORF3 in the viral replication cycle is still unclear. E4 ORF4 is a 14-kDa protein that plays a role in several different processes during Ad infection. First, E4 ORF4 binds to the Ba subunit of the serine/threonine phosphatase PP2A [70]. By binding this subunit, the trimeric form of PP2A is activated to dephosphorylate targets such as mitogen-activated protein (MAP) kinases that are important in signal transduction pathways. Increased PP2A activity leads to decreased phosphorylation and inactivation of certain transcription factors, such as E4F, through direct interaction or through the inactivation of MAP kinases. E4 ORF4 expression also results in decreased ElA phosphorylation at MAP kinase consensus sites that are important for E4 transactivation [71]. Through decreasing the activity of ElA and E4F, E4 ORF4 regulates the expression of the E4 region itself through its interaction with PP2A, perhaps to reduce the amount of potentially toxic E4 products [72]. Second, E4 ORF4 plays a role in the regulation of mRNA splicing. Third, much attention has been paid recently to the ability of E4 ORF4 to induce p53-independent apoptosis. The binding to and regulation of PP2A by E4 ORF4 is essential for the induction of cell death. E4 ORF4-dependent apoptosis also requires modulation of Src-family kinases [73]. E4 ORF6/7 is a 17-kDa protein produced from a spliced mRNA that encodes the amino terminus of E4 ORF6 linked to the unique E4-ORF7 sequence. E4 ORF6/7 molecules form stable homodimers that contribute to viral DNA synthesis by enhancing the production of E2 products. E4 ORF6/7 binds free E2F and induces cooperative and stable binding of E2F/DP heterodimers to inverted E2F binding sites in the Ad E2 early promoter [74]. Recently, it was shown that E4 ORF6/7 induces expression from the cellular E2F-1 promoter and is able to functionally compensate for ElA in adenovirus infection [75, 76]. The E4 ORF6 34-kDa protein provides a number of functions that are important in Ad infection. As stated above, E4 ORF6 has been shown to be redundant with E4 ORF3 for a number of roles in the Ad replication cycle. However, E4 ORF6 has a set of unique functions that have led to it being stud ied more intensively than its counterpart. E4 ORF6 binds to and inhibits p53, providing adenovirus yet another defense for p53 effects within the cell [64]. E4 ORF6 enhances ElA-dependent cellular transformation, possibly through the inhibition of p53 [77]. E4 ORF6 also directly binds ElB 55K, and this complex leads to the proteasome-dependent degradation of p53, counteracting 3. Adenovirus Replication 5 5 the induction of p53 stability provided by El A [31]. The E4 ORF6-E1B 55K complex is also important in the replication cycle of adenovirus. These proteins lead to host protein synthesis shutoff by selectively transporting viral mRNAs from the nucleus to the cytoplasm and inhibiting the transport of host mRNAs. E4 ORF6 possesses three targeting signals with its amino acid sequence: an arginine/lysine-rich NLS in its amino terminus, a nuclear export signal in its central region w^here it also binds p53, and a nuclear retention signal (NRS) tov^ard its carboxy terminus. The association of E4 ORF6 causes the relocal- ization of ElB 55K from the perinuclear region to the interior of the nucleus via the NLS and NRS [78]. ElB 55K was also shown to have a shuttling capability independent of its binding to E4 ORF6 [79]. ElB 55K is capable of binding mRNAs in a sequence-independent manner [80]. ElB 55K also has been shown to bind a cellular protein that binds to RNA [81]. The localization of the ElB 55K-E4 ORF6 complex to the viral transcription centers in the nucleus ensures that primarily only adenovirus late mRNA transcripts are bound and selectively transported from the nucleus. The NES of E4 ORF6 mediates the transport the RNA-protein complex out of the nucleus to the cytoplasm for translation. Early gene expression sets the stage for the replication of the viral genome. The accumulation of the replication proteins encoded in the E2 region is necessary to provide the machinery capable of carrying out viral DNA replication while a variety of early proteins attempt to stimulate the cell into S phase or negate the defense systems of the host. If the virus is successful in these pursuits, the virus will replicate very efficiently. XI . Viral DNA Replication Replication of the adenovirus DNA genome has been intensively studied over the past two decades. An in vitro Ad replication system was the first example of a mammalian cell-free DNA replication system which led to a number of discoveries on the mechanics of DNA replication, on the function of nucleoprotein complexes, and on the intricacies of virus-host interactions [34, 36, 82].
Ad DNA replication is the result of an organized interplay between viral proteins, cellular factors, and viral template DNA at distinct sites within the nucleus termed replication factories. DNA synthesis requires three viral proteins (Ad Pol, pTP, and DBP) encoded by the E2 region. Ad replication is significantly stimulated by three cellular proteins (NFI/CTF, NFII, and NFIII/Oct-1) [34, 36, 82]. These cellular factors increase replication initiation up to 200-fold. Ad replication is initiated by a protein priming event, followed by a "jumping back" mechanism, and completion by strand elongation to termination (Fig. 5). The defined origin of Ad DNA replication is located within the first 50 bp of the ITR (Fig. 5A). The terminal 18 bp of the viral genome contains 56 Evans and Hearing AdITR Oct-1 pTP/Ad-Pol TP NF-I Oct-1 NF-I t -^^ I. core on B. Preinitiation complex ^ +dCTP Origin unwinding pTP-dCMP formation NF-I dissociates pTP-CAT formation m Jumping back Oct-1 dissociates Dissociation of pTP/Ad-Pol Elongation ^GTK^^^ Figure 5 Mcdel of adenovirus DNA replication. See text for details. the minimal replication origin (core origin) with an essential triplet repeat at the molecular ends (5^-CATCAT in Ad2/5). Although this region contains the core origin, alone it can support only very limited levels of replication initiation. Immediately adjacent to the core origin is an auxiliary region that contains binding sites for NFI/CTF and NFIII/Oct-1. Binding of these cellular factors to the Ad ITR increases the efficiency vŝ ith w^hich initiation and elongation are undertaken. Nuclear factor I (NFI), a cellular transcription factor also known as CTF, was purified from HeLa cells as a protein that could enhance Ad DNA replication in vitro [34, 36, 82]. NFI/CTF binds as a dimer to the auxiliary origin of replication. This binding is enhanced by DBF, probably via changes in the DNA structure. NFI/CTF interacts with the Ad Pol-pTP complex and recruits this complex to the core origin [83]. The position of the NFI/CTF site 3. Adenovirus Replication 5 7 with relation to the core origin is critical and suggests the spatial distance is necessary for NFI/CTF to position Ad Pol-pTP correctly at the terminus of the genome. The interaction of NFI/CTF with Ad Pol-pTP leads to increased stability of the Ad Pol-pTP complex at the origin, thus increasing stimulation of initiation up to 60-fold. Oct-1, originally identified in adenovirus replication as NFIII, is an extremely well studied transcription factor, which binds to the octamer element present in a variety of promoter and enhancer regions [34, 36, 82]. The only portion of Oct-1 that is required for stimulation of Ad DNA replication is the DNA-binding POU domain [84]. The POU domain is a bipartite sequence of two conserved subdomains separated by a nonconserved or variable linker region. The presence of both subdomains is required for high-affinity DNA binding. In the adenovirus origin, the Oct-1 POU domain binds to a recognition site next to the NFI/CTF site in the auxiliary region and stimulates initiation six- to eightfold, depending on the Ad Pol-pTP concentration. The POU domain contacts the pTP protein in the Ad Pol-pTP complex, whereas NFI/CTF contacts Ad Pol, suggesting a cooperative effect of these proteins to enhance the binding of the initiation complex to the origin. Also, as with NFI/CTF, the spatial relation of the Oct-1 binding site to the core origin is important for stimulation of DNA replication. In order for efficient replication of the entire genome, a third cellular factor is required. NFII is necessary if replication is to proceed beyond 30% of the genome. This protein is a type I DNA topoisomerase and is required for elongation in vitro [85]. The reason for the need for topoisomerase function in vitro is currently unknown. Interestingly, both type I and II topoisomerase activities are required for effective Ad replication in vivo [86]. Inhibition of type I topoisomerase activity leads to an immediate block of all adenovirus replication, while inhibition of type II topoisomerase activity blocks replication after completion of the first round of synthesis. The model of the dynamics of adenovirus DNA replication involves the cooperative efforts of a number of proteins during initiation, jumping back, and elongation (Fig. 5B) [34, 36, 82]. These events need to be carefully orchestrated and organized, in order to be carried out in an efficient manner. Preceding the initiation event, the preinitiation complex composed of Ad Pol-pTP, DBP, NFI/CTF, and Oct-1 is formed at the origin. The assembly of this complex can occur in the absence of nucleotides. Binding of NFI/CTF to its site in the auxiliary origin region is facilitated by DBP [87]. Specific interactions of NFI/CTF and Oct-1 in the auxiliary region recruit and stabilize the interaction of the Ad Pol-pTP complex with the core origin [88]. The binding of the Ad Pol-pTP complex to the core origin is further enhanced by the TP linked to the genome. The phosphorylation state of Ad Pol and pTP is likely to influence the interactions of these proteins with DNA and with other proteins. After recruitment of the Ad Pol-pTP complex to the origin, DNA replication 5 8 Evans and Hearing initiates with the covalent coupUng of the first dCTP to pTP, resuhing in the formation of a pTP-dCMP complex necessary for the protein priming function. The initiation reaction is stimulated by DBP. It is also believed that DBP may be responsible for the unwinding of the origin. After the unwinding, the Ad Pol-pTP complex then forms a pTP- trinucleotide intermediate, pTP-CAT, by Ad Pol using the complementary sequence 3^-GTA located at nucleotides 4 - 6 from the genomic terminus. The presence of nucleotides in the complex causes the dissociation of NFI/CTF from its binding site. This trinucleotide-protein complex then jumps back from positions 4 - 6 by base pairing with template strand nucleotides 1-3 [89]. Following the jumping back event. Ad Pol dissociates from pTP linked to the end of the viral genome and elongation ensues. As Ad Pol replicates the viral DNA, it displaces the Oct-1 from its binding site. Ad Pol carries out DNA replication by displacing the nontemplate strand, with DBP assisting in the unwinding of downstream duplex DNA [37]. DBP also coats the single- stranded DNA resulting from elongation in order to protect it from nuclease digestion. The single-stranded DNA is then available to act as a template for a new round of pTP-primed initiation. However, in order for the Ad Pol-pTP to recognize and bind DNA, the template must be double-stranded. A double- stranded DNA template may be achieved by the annealing of the left and right ITRs on one DNA strand to one another to form a loop or panhandle structure, the end of which looks like the end of an intact genome and contains covalently attached TP. Alternatively, two single-stranded DNA molecules could anneal to form duplex DNA. The pTP attached to the 5̂ end of newly synthesized may protect the genome from nuclease digestion as well as assist in loading of an initiation complex in subsequent rounds of replication. The adenovirus E4 region produces several proteins that are required for efficient DNA replication in cell culture. As described above, the E4 ORF3 and E4 ORF6 proteins display redundant functions in Ad infection. A virus that lacks both of these two proteins is severely delayed in the onset of viral DNA replication, whereas viruses that express either one of these E4 gene products exhibit only a modest delay in the onset of DNA replication [59, 60]. Since neither of these proteins is required for DNA replication in vitro, they probably play a regulatory role in the process rather than have a direct effect in DNA synthesis. The exact nature of their participation in viral DNA replication is still unclear. Another E4 product that affects viral DNA replication is E4 ORF4. E4 ORF4 was shown to have an inhibitory effect on viral DNA replication in the absence of the E4 ORF3 and E4 ORF6 proteins [90]. E4 ORF4 may downregulate DNA synthesis through its interaction with PP2A. Dephosphorylation of ElA may affect the accumulation of E2 products, which would then decrease viral DNA synthesis. Also, PP2a activated by E4 ORF4 may dephosphorylate the viral phosphoproteins involved in DNA replication. 3. Adenovirus Replication 5 9 XII . VA RNA Genes Adenovirus encodes one or two VA (virus-associated) RNAs, depending on serotype, of about 160 nucleotides that are transcribed by host cell RNA polymerase III. VA-RNAj targets the protein kinase named PKR (reviewed in [91]). PKR is activated by the presence of low levels of double-stranded RNA, a likely product from symmetrical transcription of the Ad genome. Upon activation, PKR phosphorylates and inactivates eukaryotic initiation factor 2 alpha (eIF-2a), thus inhibiting protein synthesis in general. VA RNA is produced in large quantities in Ad-infected cells. PKR binds to the significant secondary structure found in VA-RNAj, and the high level of VA-RNAj interferes with PKR activation. This allows for the maintenance of efficient translation in Ad-infected cells [91]. XIII. Late Gene Expression and Virus Assembly Efficient late gene expression commences with the onset of Ad DNA replication. Late transcripts are initiated from the major late promoter (MLP) located at 16 map units (Fig. 1) [92]. Activation of the major late promoter appears to be mediated by both ds-acting changes in the viral genome as well as trans-acting factors. Both cellular and viral trans-acting components have been identified that bind sequences within the major late promoter. Cellular transcription factors TBP/TF-IID, USF/MLTF, and CAT box factor interact with ds-acting elements upstream of the major late promoter initiation site and are important activators of MLP expression [93]. Activation of the MLP also is specified by protein binding sites located downstream from the transcriptional start site [94]. The trans-acting components binding these sites are not fully characterized, but constitute multiprotein complexes containing the virally encoded IVa2 protein [95]. Through a mechanism(s) that is not understood, the Ad replication process significantly stimulates the activity of the MLP. The primary transcript from the major late promoter extends to the right end of the viral genome and is ~30,000 nucleotides (nt) in length. This primary transcript is polyadenylated at one of five sites and undergoes multiple splicing events to generate five families of late mRNAs (LI to L5; Fig. 1) [92]. At least 18 distinct late mRNAs are produced by alternative polyadenylation and splicing of the primary major late transcript. The 5̂ ends of all Ad late mRNAs contain an ~200-nt leader sequence referred to as the tripartite leader (due to the joining of three short exons in the primary late transcript). The tripartite leader sequence directs efficient translation of Ad late mRNAs independent of the host cell initiation factor eIF4F [96, 97]. eIF4F 60 Evans a n d H e a r i n g is a protein complex composed of phosphorylated cap-binding protein eIF4E, eIF4E kinase Mnkl , eIF4A, poly(A)-binding protein, and eIF4 G. Adenovirus infection blocks cellular translation by displacing Mnkl from eIF4F, thereby blocking phosphorylation of eIF4E 198]. The Ad-encoded lOOK late protein binds to eIF4 G in the same region bound by Mnkl and displaces Mnkl from the eIF4F complex. This results in the shut off of translation of host mRNAs and Ad mRNAs that lack the tripartite leader sequence. The translation of Ad late mRNAs that carry the tripartite leader sequence continues, effectively shutting down host mRNA translation w^hile permitting viral late mRNA translation. The mechanism by v^hich translation of Ad late mRNAs continues despite inactivation of eIF4F is not fully understood. Ad late mRNAs encode proteins that are part of the Ad capsid structure (discussed in Chapter 1 of this volume), that are involved in the virus assembly process, or that play other regulatory roles during the late phase of virus infection. Adenovirus DNA packaging into virus particles occurs in a polar manner from left to right and relies on a ds-acting packaging domain located betw^een approximately nt 200 and 380 nt (Fig. 6) (review^ed in [99]). It is thought that as-acting packaging sequences and trans-acting protein components act in conjunction to mediate DNA packaging, similar to a number of bacteriophages like lambda or^29. The formation of Ad particles proceeds through an A. packaging A repeats E1A ITR A1 TTTG GGCGTAAC CG 1 2 3 4 5 6 7 A2 TTTG GCCATTTT 1 1 1 CG 03 194 380 499
E1A enhancer A5 TTTG TCTAGGGC CG A6 TTTG ACCGTTTA CG B. 5'-TTTG NQ C G - 3 ' Al A2 A3 GGATGTTGTAGTAA\|ffni^GCGTAA(^C^GTAAG\|ATTTG\|GCCATT^^ 230 250 270 290 310 A4 A5 A6 A7 AATA/{ffff^GTTACTCATAGCGCGTAAl\|ATTTG[TCTAGGGa(P^ 330 350 370 Figure 6 Schematic of the adenovirus DNA packaging domain. (A) A schematic diagram of the left end of the adenovirus genome including the inverted terminal repeat (ITR), the packaging/enhancer region (nts. 194 to 380), and the El A 5' flanking region. The packaging repeats (A l through A7) are represented by arrows. The El A transcriptional start site is indicated by an arrow at nt 499. (B) The nucleotide sequence of the Ad5 packaging domain is shown. Numbers at the top correspond to nucleotides relative to the left end terminus. A repeats 1 through 7 are encircled. (C) The A repeat consensus sequence is shown. 3. Adenovirus Replication 6 1 ordered series of assembly events. The first virus assembly intermediate is the light particle that appears to be equivalent to a bacteriophage prohead. Light intermediate particles contain all of the major capsid proteins, lack viral DNA, and contain additional proteins that exit the particle during maturation and that may act as scaffolding proteins. Light particles mature to heavy intermediate particles upon the packaging of viral DNA and associated core proteins. As the final maturation step, activation of the virus-encoded and packaged protease results in numerous protein cleavages within the virus particle that result in maturation into the infectious virus [100]. These proteolytic cleavages are absolutely essential for the formation of highly infectious Ad particles. Further, Ad protease plays an important role in the infection process for proper release of the viral core particle from endosomes foUov^ing initial infection (reviewed in Chapter 2 of this volume). The Ad5 packaging domain is depicted in Fig. 6. The Ad5 packaging domain consists of at least seven redundant, although not functionally equiva lent, elements termed A repeats 1 through 7 [101, 102]. The functionally most important A repeats (Al, A2, A5, and A6) share a bipartite consensus motif S^-TTTGNgCG-S', which is conserved among different Ad serotypes [103]. There are spacing constraints between the two conserved parts of the bipartite consensus motif rather than between different A repeats. Multimerized copies of individual packaging repeats can restore viability to a mutant virus lacking a packaging domain. The Ad5 packaging domain displays considerable flexi bility in its position and orientation, but it must be located within ~600 bp of a genomic terminus [104]. Very little is known about the identity of trans-acting packaging com ponents involved in the packaging process. The packaging repeats very likely are binding sites for a trans-acting factor(s) that directs the packaging process. Several cellular DNA binding activities including OCT-1, COUP-TFl, and an unknown activity termed P complex interact with the TTTG half site of the packaging consensus sequence [105]. The functional significance of the binding of these cellular factors is presently under evaluation (Erturk and Hearing, unpublished results). The Ad IVa2 protein interacts with sequences that overlap the CO half of the Ad5 packaging repeat consensus [106]. It is not known if IVa2 is involved in Ad packaging, but the Ad IVa2 protein forms a protein complex with an Ad late gene product, LI 52/55K, that has a clear link to the Ad assembly process [107]. LI 52/55K mutants produce empty Ad particles or particles that are only partly packaged [108, 109]. It is easy to imagine that viral and cellular proteins form a multiprotein complex on Ad packaging signals that direct the encapsidation of the viral genome into virions. The identity and specificity of Ad packaging elements has been used to engineer Ad helper viruses whose packaging may be suppressed and to optimize yields of gutted Ad vectors. These approaches are reviewed in Chapter 15 of this volume. 6 2 Evans and Hearing XIV. Vector Design The ability of adenoviruses to infect a wide range of cell types as well as the ease with which their genomes can be manipulated has made Ad extremely attractive as a gene therapy vector. In order to obtain viral vectors for gene therapy, the viral genome carrying the transgene must be replicated and packaged to relatively high titers. Therefore, the replication events described in this chapter must be carried out to fruition in order for progeny virus to be produced, whether it is through traditional or alternative means. The first generation of Ad vectors lacked functional El and E3 regions, thus making them replication deficient due to the absence of ElA expres sion and greatly reduced transactivation of other early genes including the E2 region. These viruses were propagated in cell lines that provided El pro teins in trans allowing for efficient replication. The lack of E3 expression did not affect virus production in culture. These viruses were very effective in introducing a transgene to cells in culture and in the animal. However, the first-generation viruses elicited a significant host immune response including innate, cell-mediated, and humoral responses, preventing prolonged therapy and efficient reintroduction of the viral vector (reviewed in [110, 111]). Also, the production of these El-deficient viruses in complementing cell lines often resulted in El-positive, replication-competent adenovirus (RCA) due to recom bination with endogenous viral DNA sequences present in the complementing cell line. The contamination of gene therapy vector stocks with essentially wild-type adenovirus is unacceptable due to the outcome of lytic virus infec tion. Production of first generation vectors in complementing cell lines with integrated; nonoverlapping sequences will reduce the instances of recombinant RCAs in virus stocks. It is believed that a low level of expression of Ad proteins using first generation Ad vectors resulted in an immune response to viral infection. This may in part reflect low levels of replication of the viral genome despite the lack of El gene products. Second-generation Ad vectors were developed with additional deletions in genes involved in the replication cycle [110, 111]. The development of El-complementing cells lines that express additional Ad gene products from early region 2 or early region 4 greatly facilitated the development of the second generation Ad vectors. The inclusion of the E2 region proteins required for Ad DNA replication (Ad Pol, DBP, and pTP) in a gene therapy virus was primarily for ease of propagation of the virus due to the lack of complementing cell lines. Recently, several cell lines were described that constitutively or inducibly produce one or more of the E2 proteins, allowing for vectors lacking these sequences [112-114]. Similar lines were developed that express E4 gene products [115]. The removal of the E2 or E4 genes would provide more genetic space for the insertion of larger transgene sequences as well as allow the production of Ad vectors with greatly 3. Adenovirus Replication 6 3 reduced replication potential. The net effect of second-generation Ad vectors was the development of vectors with greater safety and significantly reduced host inflammatory responses and CTL responses to infected cells. However, the deletion of additional Ad gene products may not be without consequence to the utility of the vector in vivo. For example, the presence of the E4 ORF3 gene in Ad vectors that utilize the CMV promoter has been found to significantly contribute to sustained transgene expression both in vitro and in vivo [116, 117]. Thus, Ad vectors that express certain early gene products may be most useful under certain circumstances. An intriguing new approach in the development of adenovirus vectors is the design of oncolytic vectors that replicate in selected cells or types of cells (reviewed in [111]). This approach involves conditionally replicating viruses that undergo lytic infection in tumor cells. An example of conditionally replicating viruses is ONYX-015 [118]. This virus has been shown to replicate more efficiently in cells lacking p53. ONYX-015 is deleted for ElB 55K and it cannot replicate well in p53-positive cells, but is capable of productive infection in cells lacking active p53, such as tumor cells. The ONYX virus is discussed in great detail in Chapter 11 of this volume. Other conditionally replicative cells could be produced with El genes under the control of cell specific promoters, which are discussed in Chapters 9 and 10 of this volume. Other approaches to attack and eliminate p53 mutant cell refractory to other treatments could include vectors possessing E4 ORF4 or the E3 ADP proteins, which induce cell death independent of p53 status. The strategies mentioned above all result in the death of a target cell, such as a tumor cell, which would generally benefit from an inflammatory response and CTL infiltrate. This response would result in clearance of virus- infected cells. However, an immune response would not be beneficial while attempting to treat other diseases, such as metabolic disorders, that may require more than one treatment or prolonged presence of the viral genome. One approach that may delay or evade the immune system would be to include genes from the Ad E3 region that are involved in evasion of host immune responses during viral infection [119]. Additionally or alternatively, other viral or cellular immunomodulatory genes may be incorporated into Ad vectors toward the same goal. These types of approaches are discussed in Chapter 14 of this volume. XV. Conclusion The life cycle of adenovirus represents a complex series of events that must occur in a temporally and stoichiometrically appropriate fashion in order for efficient production of progeny virus. The virus must usurp control of the cellular machinery while controlling the expression and functions of its own 6 4 Evans and Hearing proteins. The manner with which adenovirus exerts these controls is through a myriad of protein-protein interactions. The use of adenovirus as a model system has led to a wealth of information regarding many aspects of viral as well as cellular processes, including transcriptional and posttranscriptional regulation, oncogenic transformation, virus entry into host cells, virus-host immunological interactions, translational regulation, and DNA replication. The future of adenovirus research is still fruitful with respect to both basic as well as clinical studies. A greater understanding of adenovirus infection as a whole will allow for better development and applications of gene therapy vectors through rational design. The result of which would be more effective and safer vectors, including those that are targeted to specific cell types and conditionally replicate. References 1. Rowe, W., Huebner, R. J., Gilmore, L. K., Parrott, R. H., and Ward, T. G. (1953). Isolation of a cytopathogenic agent from human adenoids undergoing spontaneous degeneration in tissue culture. Proc. Soc. Exp. Biol. Med. 84, 570-573. 2. Hilleman, M. a. W. J. (1954). Recovery of new agents from patients with acute respiratory illness. Proc. Soc. Exp. Biol. Med. 85, 183-188. 3. King, J. C , Jr. (1997). Community respiratory viruses in individuals with human immunod eficiency virus infection. Am. J. Med. 102, 19-24. 4. Jawetz, E. (1959). The story of shipyard eye. Br. Med. J. 1, 873-878. 5. Flewett, T. H., Bryden, A. S., and Davies, H. (1973). Letter: Virus particles in gastroenteritis. Lancet 2, 1497. 6. Trentin, J., Yabe, Y., and Taylor, G. (1962). The quest for human cancer viruses. Science 137, 835-849. 7. Berget, S. M., Moore, C., and Sharp, P. A. (1977). Spliced segments at the 5' terminus of adenovirus 2 late mRNA. Proc. Natl. Acad. Sci. USA 74, 3171-3175. 8. Norrby, E., Veen, J. v. d., and Espmark, A. (1970). A new serological technique for identifi cation of adenovirus infections. Proc. Soc. Exp. Biol. Med. 134, 889-895. 9. Bergelson, J. M., Cunningham, J. A., Droguett, G., Kurt-Jones, E. A., Krithivas, A., Hong, J. S., Horwitz, M. S., Crowell, R. L., and Finberg, R. W. (1997). Isolation of a common receptor for Coxsackie B viruses and adenoviruses 2 and 5. Science 275, 1320-1323. 10. Wickham, T. J., Mathias, P., Cheresh, D. A., and Nemerow, G. R. (1993). Integrins alpha v beta 3 and alpha v beta 5 promote adenovirus internalization but not virus attachment. Cell 73,309-319. 11. Arnberg, N., Edlund, K., Kidd, A. H., and Wadell, G. (1999). Adenovirus type 37 uses sialic acid as a cellular receptor. / . Virol. 74, 42-48 . 12. Akusjarvi, G. (1993). Proteins with transcription regulatory properties encoded by human adenoviruses. Trends Microbiol. 1, 163-170. 13. Dyson, N. (1998). The regulation of E2F by pRB-family proteins. Genes Dev. 12, 2245-2262. 14. Flint, J., and Shenk, T. (1997). Viral transactivating proteins. Annu. Rev.
Genet. 31, 177-212. 15. Brown, C. E., Lechner, T., Howe, L., and Workman, J. L. (2000). The many HATs of transcription coactivators. Trends Biochem. Sci. IS, 15-19. 3. Adenovirus Replication 6 5 16. Hagmeyer,B. M., Angel, P., and van Dam, H. (1995). Modulation of AP-l/ATF transcription factor activity by the adenovirus-ElA oncogene products. Bioessays 17, 621-629. 17. Schaeper, U., Boyd, J. M., Verma, S., Uhlmann, E., Subramanian, T., and Chinnadurai, G. (1995). Molecular cloning and characterization of a cellular phosphoprotein that interacts with a conserved C-terminal domain of adenovirus ElA involved in negative modulation of oncogenic transformation. Proc. Natl. Acad. Set. USA 92, 10,467-10,471. 18. Kovesdi, I., Reichel, R., and Nevins, J. R. (1986). Identification of a cellular transcription factor involved in ElA ^mws-Activation. Cell 45, 219-228. 19. Brehm, A., Miska, E. A., McCance, D. J., Reid, J. L., Bannister, A. J., and Kouzarides, T. (1998). Retinoblastoma protein recruits histone deacetylase to repress transcription. Nature 391,597-601. 20. Ferreira, R., Magnaghi-Jaulin, L., Robin, P., Harel-Bellan, A., and Trouche, D. (1998). The three members of the pocket proteins family share the ability to repress E2F activity through recruitment of a histone deacetylase. Proc. Natl. Acad. Sci. USA 95, 10,493-10,498. 21. Lee, C. W., S. T. S., Shikama, N., and LaThangue, N. B. (1998). Functional interplay between p53 and E2F through coactivator p300. Oncogene 16, 2695-2710. 22. Trouche, D., Cook, A., and Kouzarides, T. (1996). The CBP co-activator stimulates E2F1/DP1 activity. Nucleic Acids Res. 24, 4139-4145. 23. Mymryk, J. S. (1996). Tumour suppressive properties of the adenovirus 5 ElA oncogene. Oncogene n,15Sl-15H9. 24. Chinnadurai, G. (1998). Control of apoptosis by human adenovirus genes. Semin. Virol. 8, 399-408. 25. White, E. (1998). Regulation of apoptosis by ElA and ElB. Semin. Virol. 8, 505-513. 26. Steegenga, W. T., van Laar, T., Riteco, N., Mandarino, A., Shvarts, A., van der Eb, A. J., and Jochemsen, A. G. (1996). Adenovirus ElA proteins inhibit activation of transcription by p53. Mol. Cell Biol. 16, 2101-2109. 27. de Stanchina, E., McCurrach, M. E., Zindy, F., Shieh, S. Y., Ferbeyre, G., Samuelson, A. V., Prives, C , Roussel, M. F., Sherr, C. J., and Lowe, S. W. (1998). ElA signaling to p53 involves the pl9(ARF) tumor suppressor. Genes Dev. 12, 2434-2442. 28. Somasundaram, K., and El-Deiry, W. S. (1997). Inhibition of p53-mediated transactiva- tion and cell cycle arrest by ElA through its p300/CBP-interacting region. Oncogene 14, 1047-1057. 29. Martin, M. E., and Berk, A. J. (1998). Adenovirus ElB 55K represses p53 activation in vitro. / . Virol. 72, 3146-3154. 30. Martin, M. E., and Berk, A. J. (1999). Corepressor required for adenovirus ElB 55,000- molecular-weight protein repression of basal transcription. Mol. Cell Biol. 19, 3403-3414. 31. Steegenga, W. T., Riteco, N., Jochemsen, A. G., Fallaux, F. J., and Bos, J. L. (1998). The large ElB protein together with the E4orf6 protein target p53 for active degradation in adenovirus infected cells. Oncogene 16, 349-357. 32. Perez, D., and White, E. (1998). ElB 19K inhibits Fas-mediated apoptosis through FADD- dependent sequestration of FLICE. / . Cell Biol. 141, 125S-1266. 33. Swaminathan, S. and Thimmapaya, B. (1995). Regulation of adenovirus E2 transcription unit. Curr. Top. Microbiol. Immunol. 199, 177-194. 34. Hay, R. T., Freeman, A., Leith, I., Monaghan, A., and Webster, A. (1995). Molecular interactions during adenovirus DNA replication. Curr. Top. Microbiol. Immunol. 199, 31-48. 35. Ramachandra, M., and Padmanabhan, R. (1995). Expression, nuclear transport, and phos phorylation of adenovirus DNA replication proteins. Curr. Top. Microbiol. Immunol. 199, 50-88. 36. Van der Vliet, P. C. (1995). Adenovirus DNA replication. Curr. Top. Microbiol. Immunol. 199, 1-30. 6 6 Evans and Hearing 37. Dekker, J., Kanellopoulos, P. N., Loonstra, A. K., van Oosterhout, J. A., Leonard, K., Tucker, P. A., and van der Vliet, P. C. (1997). Multimerization of the adenovirus DNA- binding protein is the driving force for ATP-independent DNA unwinding during strand displacement synthesis. EMBO J, 16, 1455-1463. 38. Zijderveld, D. C , Stuiver, M. H., and van der VHet, P. C. (1993). The adenovirus DNA binding protein enhances intermolecular DNA renaturation but inhibits intramolecular DNA renaturation. Nucleic Acids Res. 21, 2591-2598. 39. Lindenbaum, J. O., Field, J., and Hurwitz, J. (1986). The adenovirus DNA binding protein and adenovirus DNA polymerase interact to catalyze elongation of primed DNA templates. / . Biol. Chem. 261, 10,218-10,227. 40. Webster, A., Leith, I. R., Nicholson, J., Hounsell, J., and Hay, R. T. (1997). Role of preter minal protein processing in adenovirus replication./. Virol. 71, 6381-6389. 41. Fredman, J. N., and Engler, J. A. (1993). Adenovirus precursor to terminal protein interacts with the nuclear matrix in vivo and in vitro. / . Virol. 67, 3384-3395. 42. Schaack, J., Ho, W. Y., Freimuth, P., and Shenk, T. (1990). Adenovirus terminal protein mediates both nuclear matrix association and efficient transcription of adenovirus DNA. G^w^sD^z/. 4, 1197-1208. 43. Angeletti, P. C , and Engler, J. A. (1998). Adenovirus preterminal protein binds to the CAD enzyme at active sites of viral DNA replication on the nuclear matrix. / . Virol. 72, 2896-2904. 44. Chen, M., and Horwitz, M. S. (1989). Dissection of functional domains of adenovirus DNA polymerase by linker-insertion mutagenesis. Proc. Natl. Acad. Sci. USA 86, 6116-6120. 45. Ramachandra, M., Nakano, R., Mohan, P. M., Rawitch, A. B., and Padmanabhan, R. (1993). Adenovirus DNA polymerase is a phosphoprotein./. Biol. Chem. 268, 442-448. 46. Wold, W. S., Tollefson, A. E., and Hermiston, T. W. (1995). E3 transcription unit of aden ovirus. Curr. Top. Microbiol. Immunol. 199, 237-274. 47. Wold, W. S., Doronin, K., Toth, K., Kuppuswamy, M., Lichtenstein, D. L., and Tollef son, A. E. (1999). Immune responses to adenoviruses: Viral evasion mechanisms and their implications for the clinic. Curr. Opin. Immunol. 11, 380-386. 48. Jackson, M. R., Nilsson, T., and Peterson, P. A. (1993). Retrieval of transmembrane proteins to the endoplasmic reticulum./. Cell Biol. 121, 317-333. 49. Nagata, S. (1997). Apoptosis by death factor. Cell 88, 355-365. 50. Hsu, H., Huang, J., Shu, H. B., Baichwal, V., and Goeddel, D. V. (1996). TNF-dependent recruitment of the protein kinase RIP to the TNF receptor-1 signaling complex. Immunity 4, 387-396. 51. Gooding, L. R., Ranheim, T. S., Tollefson, A. E., Aquino, L., Duerksen-Hughes, P., Hor- ton, T. M., and Wold, W. S. (1991). The 10,400- and 14,500-dalton proteins encoded by region E3 of adenovirus function together to protect many but not all mouse cell lines against lysis by tumor necrosis factor. / . Virol. 65, 4114-4123. 52. Tollefson, A. E., Hermiston, T. W., Lichtenstein, D. L., CoUe, C. F., Tripp, R. A., Dim- itrov, T., Toth, K., Wells, C. E., Doherty, P. C , and Wold, W. S. (1998). Forced degradation of Fas inhibits apoptosis in adenovirus-infected cells. Nature 392, 726-730. 53. Chen, P., Tian, J., Kovesdi, I., and Bruder, J. T. (1998). Interaction of the adenovirus 14.7-kDa protein with FLICE inhibits Fas ligand-induced apoptosis. / . Biol. Chem. 273, 5815-5820. 54. Krajcsi, P., Dimitrov, T., Hermiston, T. W., Tollefson, A. E., Ranheim, T. S., Vande Pol, 5. B., Stephenson, A. H., and Wold, W. S. (1996). The adenovirus E3-14.7K protein and the E3-10.4K/14.5K complex of proteins, which independently inhibit tumor necrosis factor (TNF)-induced apoptosis, also independently inhibit TNF-induced release of arachidonic acid./. Virol. 70,4904-4913. 55. Li, Y., Kang, J., Friedman, J., Tarassishin, L., Ye, J., Kovalenko, A., Wallach, D., and Hor witz, M. S. (1999). Identification of a cell protein (FIP-3) as a modulator of NF-kappaB 3. Adenovirus Replication 6 7 activity and as a target of an adenovirus inhibitor of tumor necrosis factor alpha-induced apoptosis. Proc. Natl. Acad. Sci. USA 96, 1042-1047. 56. Deryckere, F., and Burgert, H. G, (1996). Tumor necrosis factor alpha induces the adenovirus early 3 promoter by activation of NF-kappaB. / . Biol. Chem. 271, 30,249-30,255. 57. Tollefson, A. E., Scaria, A., Hermiston, T. W., Ryerse, J. S., Wold, L. J., and Wold, W. S. (1996). The adenovirus death protein (E3-11.6K) is required at very late stages of infec tion for efficient cell lysis and release of adenovirus from infected cells. / . Virol. 70, 2296-2306. 58. Tollefson, A. E., Scaria, A., Saha, S. K., and Wold, W. S. (1992). The 11,600-MW protein encoded by region E3 of adenovirus is expressed early but is greatly amplified at late stages of infection./. Virol. 66, 3633-3642. 59. Bridge, E., and Ketner, G. (1989). Redundant control of adenovirus late gene expression by early region 4. / . Virol. 63, 631-638. 60. Huang, M. M., and Hearing, P. (1989). Adenovirus early region 4 encodes two gene products with redundant effects in lytic infection./. Virol. 63, 2605-2615. 61. Thomas, D. L., Shin, S., Jiang, B. H., Vogel, H., Ross, M. A., Kaplitt, M., Shenk, T. E., and Javier, R. T. (1999). Early region 1 transforming functions are dispensable for mammary tumorigenesis by human adenovirus type 9 . / . Virol. 73, 3071-3079. 62. Glaunsinger, B. A., Lee, S. S., Thomas, M., Banks, L., and Javier, R. (2000). Interactions of the PDZ-protein MAGI-1 with adenovirus E4-ORF1 and high-risk papillomavirus E6 oncoproteins. Oncogene 19, 5270-5280. 63. Nordqvist, K., Ohman, K., and Akusjarvi, G. (1994). Human adenovirus encodes two proteins which have opposite effects on accumulation of alternatively spliced mRNAs. Mol. Cell Biol. 14, 437-445. 64. Dobner, T., Horikoshi, N., Rubenwolf, S., and Shenk, T. (1996). Blockage by adenovirus E4orf6 of transcriptional activation by the p53 tumor suppressor. Science 272, 1470-1473. 65. Konig, C., Roth, J., and Dobbelstein, M. (1999). Adenovirus type 5 E4orf3 protein reUeves p53 inhibition by ElB-55-kilodalton protein./. Virol. 73, 2253-2262. 66. Boyer, J., Rohleder, K., and Ketner, G. (1999). Adenovirus E4 34K and E4 U K inhibit double strand break repair and are physically associated with the cellular DNA-dependent protein kinase. Virology 263, 307-312. 67. Doucas, V., Ishov, A. M., Romo, A., Juguilon, H., Weitzman, M. D., Evans, R. M., and Maul, G. G. (1996). Adenovirus replication is coupled with the dynamic properties of the PML nuclear structure. Genes Dev. 10, 196-207. 68. Seeler, J. S., and Dejean, A. (1999). The PML nuclear bodies: actors or extras? Curr. Opin. Genet. Dev. 9, 362-367. 69. Maul, G. G. (1998). Nuclear domain 10, the site of DNA virus transcription and replication. Bioessays 20, 660-667. 70. Kleinberger, T., and Shenk, T. (1993). Adenovirus E4orf4 protein binds to protein phos phatase 2A, and the complex down regulates ElA-enhanced junB transcription./. Virol. 67, 7556-7560. 71. Whalen, S. G., Marcellus, R. C., Whalen, A., Ahn, N. G., Ricciardi, R. P., and Branton, P. E. (1997). Phosphorylation within the transactivation domain of adenovirus El A protein by mitogen-activated protein kinase regulates expression of early region 4. / . Virol. 71, 3545-3553. 72. Bondesson, M., Ohman, K., Manervik, M., Fan, S., and Akusjarvi, G. (1996). Adenovirus E4 open reading frame 4 protein autoregulates E4 transcription by inhibiting El A transactivation of the E4 promoter./. Virol. 70, 3844-3851. 73. Lavoie, J. N., Champagne, C., Gingras, M. C., and Robert, A. (2000). Adenovirus E4 open reading frame 4-induced apoptosis involves dysregulation of Src family kinases. / . Cell Biol. 150, 1037-1056. 6 8 Evans and Hearing 74. Huang, M. M., and Hearing, P. (1989). The adenovirus early region 4 open reading frame 6/7 protein regulates the DNA binding activity of the cellular transcription factor, E2F, through a direct complex. Genes Dev. 3, 1699-710. 75. O'Connor, R. J., and Hearing, P. (2000). The E4-6/7 protein functionally compensates for the loss of ElA expression in adenovirus infection. / . Virol. 74, 5819-5824. 76. Schaley, J., O'Connor, R. J., Taylor, L. J., Bar-Sagi, D., and Hearing, P. (2000). Induction of the cellular E2F-1 promoter by the adenovirus E4-6/7 protein. / . Virol. 74, 2084-2093. 77. Nevels, M., Spruss, T., Wolf, H., and Dobner, T. (1999). The adenovirus E4orf6 protein contributes to malignant transformation by antagonizing ElA-induced accumulation of the tumor suppressor protein p53. Oncogene 18, 9-17. 78. Dobbelstein, M., Roth, J., Kimberly, W. T., Levine, A. J., and Shenk, T. (1997). Nuclear export of the ElB 55-kDa and E4 34-kDa adenoviral oncoproteins mediated by a rev-like signal sequence. EMBO J. 16, 4276-4284. 79. Kratzer, F., Rosorius, O., Heger, P., Hirschmann, N., Dobner, T., Hauber, J., and Stauber, R. H. (2000). The adenovirus type 5 E1B-55K oncoprotein is a highly active shuttle protein and shuttling is independent of E4orf6, p53 and Mdm2. Oncogene 19, 850-857. 80. Horridge, J. J., and Leppard, K. N. (1998). RNA-binding activity of the ElB 55-kilodalton protein from human
adenovirus type 5. J. Virol. 72, 9374-9379. 81. Gabler, S., Schutt, H., Groitl, P., Wolf, H., Shenk, T., and Dobner, T. (1998). ElB 55- kilodalton-associated protein: A cellular protein with RNA-binding activity implicated in nucleocytoplasmic transport of adenovirus and cellular mRNAs. / . Virol. 72, 7960-7971. 82. de Jong, R. N., and van der Vliet, P. C. (1999). Mechanism of DNA replication in eukaryotic cells: cellular host factors stimulating adenovirus DNA replication. Gene 236, 1-12. 83. Armentero, M. T., Horwitz, M., and Mermod, N. (1994). Targeting of DNA polymerase to the adenovirus origin of DNA replication by interaction with nuclear factor I. Proc. Natl. Acad. Sci. USA 91, 11,537-11,541. 84. van Leeuwen, H. C , Strating, M. J., Cox, M., Kaptein, R., and van der Vliet, P. C. (1995). Mutation of the Oct-1 POU-specific recognition helix leads to altered DNA binding and influences enhancement of adenovirus DNA replication. Nucleic Acids Res. 23, 3189-3197. 85. Nagata, K., Guggenheimer, R. A., and Hurwitz, J. (1983). Specific binding of a cellular DNA replication protein to the origin of replication of adenovirus DNA. Proc. Natl. Acad. Sci. 1/5A 80, 6177-6181. 86. Schaack, J., Schedl, P., and Shenk, T. (1990). Topoisomerase I and II cleavage of adenovirus DNA in vivo: Both topoisomerase activities appear to be required for adenovirus DNA replication./. Virol. 64, 78-85 . 87. Stuiver, M. H., and van der Vliet, P. C. (1990). Adenovirus DNA-binding protein forms a multimeric protein complex with double-stranded DNA and enhances binding of nuclear factor I . / . Virol. 64, 379-386. 88. Mul, Y. M., Verrijzer, C. P., and van der Vliet, P. C. (1990). Transcription factors NFI and NFIII/oct-1 function independently, employing different mechanisms to enhance adenovirus DNA replication./. Virol. 64, 5510-5518. 89. King, A. J., and van der Vliet, P. C. (1994). A precursor terminal protein-trinucleotide intermediate during initiation of adenovirus DNA replication: Regeneration of molecular ends in vitro by a jumping back mechanism. EMBO J. 13, 5786-5792. 90. Medghalchi, S., Padmanabhan, R., and Ketner, G. (1997). Early region 4 modulates aden ovirus DNA replication by two genetically separable mechanisms. Virology 236, 8-17. 91. Mathews, M. B., and Shenk, T. (1991). Adenovirus virus-associated RNA and translation control./. Virol. 65, 5657-5662. 92. Shaw, A. R., and Ziff, E. B. (1980). Transcripts from the adenovirus-2 major late promoter yield a single early family of 3' coterminal mRNAs and five late families. Cell 22, 905-916. 93. Reach, M., Xu, L. X., and Young, C. S. (1991). Transcription from the adenovirus major late promoter uses redundant activating elements. EMBO } . 10, 3439-3446. 3. Adenovirus Replication 6 9 94. Mondesert, G., Tribouley, C , and Kedinger, C. (1992). Identification of a novel downstream binding protein implicated in late-phase-specific activation of the adenovirus major late promoter. Nucleic Acids Res. 23, 3881-3889. 95. Lutz, P., and Kedinger, C. (1996). Properties of the adenovirus IVa2 gene product, an effector of late-phase-dependent activation of the major late promoter. / . Virol. 70, 1396-1405. 96. Dolph, P. J., Racaniello, V., Villamarin, A., Palladino, F., and Schneider, R. J. (1988). The adenovirus tripartite leader may eliminate the requirement for cap-binding protein complex during translation initiation./. Virol. 62, 1059-2066. 97. Huang, J. T., and Schneider, R. J. (1991). Adenovirus inhibition of cellular protein synthesis involves inactivation of cap-binding protein. Cell 65, 271-280. 98. Cuesta, R., Xi, Q., and Schneider, R. J. (2000). Adenovirus-specific translation by displace ment of kinase Mnkl from cap-initiation complex eIF4F. EMBO J. 19, 3465-3474. 99. Schmid, S. I., and Hearing, P. (1995). Selective encapsidation of adenovirus DNA. Curr. Top. Microbiol. Immunol. 199, 67-80. 100. Weber, J. M. (1995). Adenovirus endopeptidase and its role in virus infection. Curr. Top. Microbiol. Immunol. 199, 227-235. 101. Grable, M., and Hearing, P. (1990). Adenovirus type 5 packaging domain is composed of a repeated element that is functionally redundant. / . Virol. 64, 2047-2056. 102. Grable, M., and Hearing, P. (1992). cis and trans requirements for the selective packaging of adenovirus type 5 DNA. / . Virol. 66, 723-731. 103. Schmid, S. I., and Hearing, P. (1997). Bipartite structure and functional independence of adenovirus type 5 packaging elements./. Virol. 71, 3375-3384. 104. Hearing, P., Samulski, R. J., Wishart, W. L., and Shenk, T. (1987). Identification of a repeated sequence element required for efficient encapsidation of the adenovirus type 5 chromosome./. Virol. 61, 2555-2558. 105. Schmid, S. I., and Hearing, P. (1998). Cellular components interact v^ith adenovirus type 5 minimal DNA packaging domains. / . Virol. 72, 6339-6347. 106. Zhang, W., and Imperiale, M.J . (2000). Interaction of the adenovirus IVa2 protein v^ith viral packaging sequences./. Virol. 74, 2687-2693. 107. Gustin, K. E., Lutz, P., and Imperiale, M. J. (1996). Interaction of the adenovirus LI 52/55- kilodalton protein v^ith the IVa2 gene product during infection. / . Virol. 70, 6463-6467. 108. Gustin, K. E., and Imperiale, M. J. (1998). Encapsidation of viral DNA requires the aden ovirus LI 52/55-kilodalton protein./. Virol. 72, 7860-7870. 109. Hasson, T. B., Soloway, P. D., Ornelles, D. A., Doerfler, W., and Shenk, T. (1989). Aden ovirus LI 52- and 55-kilodalton proteins are required for assembly of virions. / . Virol. 63, 3612-3621. 110. Benihoud, K., Yeh, P., and Perricaudet, M. (1999). Adenovirus vectors for gene delivery. Curr. Opin. Biotechnol. 10, 440-447. 111. Zhang, W. W. (1999). Development and application of adenoviral vectors for gene therapy of cancer. Cancer Gene Ther. 6, 113-138. 112. GorzigHa, M. I., Kadan, M. J., Yei, S., Lim, J., Lee, G. M., Luthra, R., and Trapnell, B. C. (1996). Elimination of both El and E2 from adenovirus vectors further improves prospects for in vivo human gene therapy. / . Virol. 70, 4173-4178. 113. Hartigan, O. C. D., Amalfitano, A., and Chamberlain, J. S. (1999). Improved production of gutted adenovirus in cells expressing adenovirus preterminal protein and DNA polymerase. / . Virol. 73, 7835-7841. 114. Zhou, H., and Beaudet, A. L., (2000). A new vector system with inducible E2a cell line for production of higher titer and safer adenoviral vectors. Virology 275, 348-357. 115. Brough, D. E., Lizonova, A., Hsu, C , Kulesa, V. A., and Kovesdi, I. (1996). A gene transfer vector-cell line system for complete functional complementation of adenovirus early regions El and E4. / . Virol. 70, 6497-6501. 7 0 Evans and Hearing 116. Brough, D. E., Hsu, C , Kulesa, V. A., Lee, G. M., Cantolupo, L. J., Lizonova, A., and Kovesdi, I. (1997). Activation of transgene expression by early region 4 is responsible for a high level of persistent transgene expression from adenovirus vectors in vivo. / . Virol. 71, 9206-9213. 117. Yew, N. S., Marshall, J., Przybylska, M., Wysokenski, D. M., Ziegler, R. J., Rafter, P. W., Li, C , Armentano, D., and Cheng, S. H. (1999). Increased duration of transgene expression in the lung with plasmid DNA vectors harboring adenovirus E4 open reading frame 3. Hum. GeneTher. 10, 1833-1843. 118. Kirn, D., Hermiston, T., McCormick, F., Brough, D. E., Lizonova, A., Hsu, C , Kulesa, V. A., and Kovesdi, I. (1998). ONYX-015: clinical data are encouraging. Nat. Med. 4, 1341-1342. 119. Bruder, J. T., Jie, T., McVey, D. L., and Kovesdi, I. (1997). Expression of gpl9K increases the persistence of transgene expression from an adenovirus vector in the mouse lung and liver./. Virol. 71, 7623-7628. C H A P T E R Adenoviral Vector Construction I: Mammalian Systems Philip Ng'-^ and Frank L Graham'̂ ^^^ Departments of *Biology, "^Pathology, and "^Molecular Medicine McMaster University Hamilton, Ontario, Canada I. Introduction Adenoviruses (Ads) are excellent gene transfer vectors and are extensively used for high level expression of transgene products in cultured cells, as potential recombinant viral vaccines and for gene therapy. Ads are particularly v^ell suited for these applications because their genome is relatively easy to manipulated, they grow to high titers, they are stable and easy to purify, and they can transduce many cell types from numerous mammalian species including both dividing and nondividing cells in vitro and in vivo [1-4]. A. Adenovirus Biology The adenovirion is a nonenveloped icosohedral capsid containing a linear double-stranded DNA genome of ~30 -40 kb. Of the ~50 serotypes of human Ad, the most extensively characterized are serotypes 2 (Ad2) and 5 (Ad5) of subgroup C (review^ed in [5]). The 36-kb genomes of Ad2 and Ad5 are flanked by inverted terminal repeats (ITRs) which are the only sequences required in cis for viral DNA replication. A c/s-acting packaging signal, required for encapsidation of the genome, is located near the left ITR (relative to the conventional map of Ad). The Ad genome can be roughly divided into tv^o sets of genes (Fig. 1): the early region genes. El A, ElB, E2, E3, and E4, are expressed before DNA replication and the late region genes, LI to L5 are expressed to ^ Present address: Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, Texas. ADENOVIRAL VECTORS FOR GENE THERAPY J \| Copyright 2002, Elsevier Science (USA). All rights reserved. 72 Ng and Graham Late Transcription ^ 1 2 3 X y z [ t EIAEIB 1 E3 ' feVA, ^ _ •'^- ^ N 0 10 20 30 40 50 60 70 80 90 100 1 1 1 I - 1 1 1 1 1 1— 1 » _ i 1 1 1 — 1 1 1 1—1 l^a^ m\'' ^ ^ V _ . V E2 Figure 1 Transcription map of human adenovirus serotype 5. The 100 map unit (~36 kb) genome is divided into four early region transcription units, El - E 4 , and five families of late mRNA, LI -15, which are alternative splice products of a common late transcript expressed from the major late promoter (MPL) located at 16 \|x. Four smaller transcripts, pIX, IVa, and VA RNAs I and II, are also produced. Not shown are the 103-bp inverted terminal repeats located at the termini of the genome involved in viral DNA replication and the packaging signal located from nucleotides 190 to 380 at the left end of the genome involved in encapsidation. high levels after initiation of DNA replication. The ElA transcription unit encodes two major ElA proteins that are involved in transcriptional regulation of the virus and stimulation of the host cell to enter an S phase-like state and is the first early region to be expressed during viral infection. The two major ElB proteins are necessary for blocking host mRNA transport, stimulating viral mRNA transport and blocking ElA-induced apoptosis. The E2 region encodes proteins required for viral DNA replication and can be divided into two subregions; E2a encodes the 72-kDa DNA-binding protein and E2b encodes the viral DNA polymerase and terminal protein precursor (pTP). The E3 region, which is dispensable for virus growth in cell culture, encodes at least seven proteins, most of which are involved in host immune evasion. The E4 region encodes at least six proteins, some functioning to facilitate DNA replication, enhance late gene expression and decrease host protein synthesis. The late region genes are expressed from a common major late promoter (MLP) and are generated by alternative splicing of a single transcript. Most of the late mRNAs encode virion structural proteins. In addition to the early and late region genes, four other small transcripts are also produced. The 4. Adenoviral Vector Construction I: Mammalian Systems 7 3 gene encoding protein IX (pIX) is colinear with ElB but utilizes a different promoter and is expressed at an intermediate time, as is the pIVal gene. Other late transcripts include the RNA polymerase III transcribed VA RNA I and II. Virus infection is initiated through the Ad fiber protein binding to specific primary receptors on the cell surface [6, 7] followed by a secondary interaction between the virion penton base and ayPs and ayps integrins [8]. The efficiency with which Ad binds and enters the cell is directly related to the level of primary and secondary receptors present on the cell surface [9, 10]. Penton-integrin interaction triggers Ad internalization by endocytosis where it escapes from the early endosome into the cytosol prior to lysosome formation [11,12]. The virion is sequentially disassembled during translocation along the microtubule network toward the nucleus where the viral DNA is released into the nucleus [13] where viral DNA replication, beginning 6 to 8 h postinfection, and assembly of progeny virions occur. The entire life cycle takes about 24-36 h, generating
about 10"̂ virions per infected cell. Ads have never been implicated as a cause of malignant disease in their natural host and, in immunocompetent humans, they generally cause only relatively mild, self-limiting illness. The reader is referred to an excellent review by Shenk [5] for a more comprehensive discussion of adenoviruses. B. Adenovirus Vectors Typically, Ads are converted into mammalian gene transfer vectors by replacing the El region with the foreign DNA of interest. This serves two important purposes. First, since the packaging constraint of Ad is 105% of wild type [14], deletion of El increases the cloning capacity to ~5 kb. Second, it renders the vectors replication-deficient, which is important with respect to safety for human gene therapy and other applications. These replication- deficient vectors must be propagated in El-complementing cell lines such as 293 [15]. The E3 region can also be deleted from the vector since it is not required for virus propagation in culture. The combination of El and E3 deletions results in a cloning capacity of ~8 kb, a size that is more than adequate for most expression cassettes. C. Early Methods of Constructing Recombinant Adenoviruses All methods for manipulating Ad genomes for construction of vectors rely on the observation that purified viral DNA is infectious [16]. Early methods for generating recombinant Ad involved direct manipulation of viral DNA extracted from virions. These methods included in vivo homologous recombination between viral DNAs cotransfected into cells [17] and in vitro ligation of viral DNAs cleaved by restriction enzymes [18, 19]. However, a major limitation of these methods was that precisely defined alterations 74 Ng and Graham ^ i^ Viral DNA Xbal Xbal I Cleave DNAs with Xba 11 Ligate in vitro and transfect 293 cells ^f Recombinant virus ^ lA Viral DNA Xbal Cleave with Xba I and cotransfect / 293 cells with shuttle plasmid Homologous recombination X ^^ Recombinant virus Figure 2 Early methods for constructing recombinant Ad vectors. In the method depicted in (A), shuttle plasmids bearing the modified left end of the Ad genome and purified viral DNA are cleaved with a restriction enzyme. The recombinant Ad genome is generated by direct in vitro ligation between the shuttle plasmid and the viral DNA and infectious recombinant viruses are generated by transfecting the ligation product into 293 cells. In the method depicted in (B), shuttle plasmids bearing the modified left end of the Ad genome and purified viral DNA are cotransfected into 293 cells. Recombinant viruses are generated as a result of in vivo homologous recombination between their overlapping region of homology. To minimize production of nonrecombinant parent virus, the viral DNA is cleaved with a restriction enzyme at the left end prior to cotransfection. Thick gray lines represent cloned Ad DNA, thick black lines represent Ad viral DNA, thin black lines represent bacterial plasmid sequences and small arrows represent ITR. \\f, packaging signal. could not be introduced into the genome owing to the difficuhy inherent in manipulating the large linear viral DNA. In 1981, Stow [20] devised a method to overcome this limitation, at least for modifications of the left end of the genome, that employed in vitro ligation between a cloned subgenomic Ad fragment and viral DNA (Fig. 2A). In this study, the left end Hpal E fragment (0 to 4.5 mu) bearing the ElA region of Ad2 was inserted into pBR322, thus 4. Adenoviral Vector Construction I: Mammalian Systems 7 5 permitting the plasmid-borne El A sequences to be easily modified. The plasmid and Ad5 viral DNA (from variant dl309 which has a unique Xbal site [21]) w êre cleaved with Xbal at 3.7 mu and ligated in vitro. Recombinant Ads bearing the modified ElA region were generated by transfecting the ligation product into 293 cells. The significance of this method lay in the ability to reconstruct an infectious viral genome by using a cloned Ad subgenomic fragment as one of the substrates, thus allowing modifications engineered into the cloned sequence to be readily introduced into the viral genome. However, despite this advance, the approach was limited because few unique restriction enzyme sites were available due to the relatively large size of the Ad genome. A method that overcame these limitations was developed by Kapoor and Chinnadurai [22], who demonstrated that recombinant Ads could be generated by in vivo homologous recombination between a cloned Ad subgenomic fragment and viral DNAs (Fig. 2B). As in the method developed by Stow [20], a shuttle plasmid bearing a left end subgenomic Ad fragment was first constructed to permit easy modification of El sequences. The shuttle plasmid along with viral DNA were cotransfected into 293 cells and in vivo homologous recombination between their overlapping sequences resulted in the generation of recombinant Ad bearing the modified El . While this method does not rely on ligation of two restriction enzyme sites, the viral DNA is still cleaved prior to cotransfection to reduce the background of nonrecombinant parental virus. Currently, the unique Xbal site at 3.7 mu in Ad5 variant dl309 and the unique Clal site at 2.5 mu in wt Ad5 are the most useful for manipulation of the left end. While both of these early methods have proven useful, a major limitation is the requirement for viral DNA as a substrate for vector construction. Purification of viral DNA is time consuming and laborious and its use leads to a background of parental nonrecombinant viruses resulting in the need to screen a large number of viruses to isolate the desired recombinant. This can prove especially problematic when the parental virus has a growth advantage over the recombinant vector. Considering the importance and utility of Ad vectors as a tool for mammalian gene transfer, development of improved systems for their efficient and reliable construction was clearly imperative. II. The Two-Plasmid Rescue System One of the first methods that was developed to overcome the limitations of the earlier approaches was the two-plasmid rescue system (Fig. 3A). In this method, recombinant Ad vectors are generated by in vivo homologous recombination between two noninfectious plasmids cotransfected into 293 cells. Since its development, the two-plasmid rescue method has been widely used due to its simplicity. Other methods of constructing Ad vectors have 7 6 Ng and Graham Shuttle plasmid/ Foreign DN homologous recombination ! J3 lA Foreign DNA ™ AEl Adenovirus vector ITR Shuttle plasmid ^• ' • ' 'S^ t^o .^ ign ^^A , lox? Cre-mediated recombination ITR , J ry ITR , _ V̂ LacZ J_^ =t3 AEl /o;cP Adenovirus vector Figure 3 Construction of Ad vectors by (A) in vivo homologous recombination following cotrons- fection of 293 cells with a shuttle plasmid and an Ad genomic plasmid and (B) Cre-mediated site-specific recombination following cotransfection of 293Cre4 cells with a shuttle plasmid bearing a lox P site and pBHGloxAEl ,3. Ad sequences are represented by thick black lines, bacterial plasmid sequences are represented by thin black lines and the position and orientation of the loxP site is represented by a white triangle. Only the relevant portions of the shuttle plasmids are shown. also been developed; however, discussion of these is beyond the scope of this chapter. The reader is encouraged to consult the other chapters in this book for further details regarding these other methods. The remainder of this section w îll focus on the development of the tw^o-plasmid rescue method and the recent improvements that have been made to increase the method's efficiency and expand its versatility. The final section provides detailed protocols for 4. Adenoviral Vector Construction I: Mammalian Systems 7 7 the construction, using the two plasmid rescue method, and propagation of recombinant Ad vectors. A. Development of the Two-Plasmid Rescue System In developing the tw^o-plasmid rescue method, advantage w âs taken of observations made in early studies of Ad. One such observation was made in 1983 by Berkner and Sharp [23], who demonstrated that infectious recombinant Ads could be generated by cotransfecting 293 cells with cloned, noninfectious subgenomic Ad fragments. In this study, subgenomic fragments of Ad were cloned into the EcoRI site of pBR322. Infectious recombinant virus could be generated following cotransfection of 293 cells with these plasmids as a result of in vivo homologous recombination between their overlapping Ad sequences. Generation of infectious virus by this method was dependent on releasing the Ad ITR from at least one of the plasmids by EcoRI cleavage. The significance of this study lies in using only noninfectious plasmids, instead of viral DNA, as the substrates for vector construction, thus avoiding the need to isolate viral DNA and the problem of contaminating nonrecombinant virus. Another key finding, made in 1983 by Ruben et al. [24], was the discovery that up to 10% of Ad viral DNA molecules become circularized following infection of mammalian cells. This permitted cloning of the entire Ad genome as an infectious bacterial plasmid. One such Ad genomic plasmid, pFG140, consisted of a circularized dl309 Ad genome [21] with a 2.2-kb insert in the Xbal site at 3.7 mu containing the ampicillin resistance marker and a bacterial origin of DNA replication [25]. This Ad genomic plasmid could be stably propagated in Escherichia coli and was capable of generating infectious virus following transfection into mammalian cells at efficiencies comparable to purified virion DNA. In 1987, Ghosh-Choudhury et al. [26] observed that an Ad genomic plas mid of wild-type size but, unlike pFG140, bearing a deletion of the Ad protein IX gene was noninfectious. Based on this observation, they hypothesized that pIX was essential for the packaging of full-length Ad genomes. According to this hypothesis, reintroduction of the pIX gene into the noninfectious Ad genomic plasmid should restore its infectivity. To test this, 293 cells were cotransfected with the noninfectious pIX-deleted Ad genomic plasmid and a plasmid bearing the left end of the Ad genome including the pIX gene. Infectious viruses, all bearing the pIX gene, were generated as a result of in vivo homologous recombination between the two cotransfected plasmids thus demonstrating that pIX is essential for packaging full-length genomes. Based on the early studies of Ad described above, McGrory et aL [27] developed the first two-plasmid rescue system designed specifically for con structing recombinant, replication-defective Ad vectors in which the El region was substituted with a foreign transgene as depicted in Fig. 3A. To accomplish 7 8 Ng and Graham this the infectious Ad genomic plasmid pFG140 was modified by replacing the 2.2-kb insert with a 4.4-kb segment to generate the 40-kb plasmid pJM17. The resulting genome exceeded the packaging constraint of Ad and was nonin fectious but could replicate following transfection into 293 cells. To generate infectious recombinant Ads bearing foreign DNA of up to 5.4 kb in place of the El region, 293 cells were cotransfected with pJM17 and a shuttle plasmid bearing the left end of the Ad genome with the desired El region substitution. Since, in principle, neither plasmid was infectious only the desired recombinant El substituted vector should be generated as a result of in vivo homologous recombination between the overlapping Ad sequences in the shuttle plasmid and pJM17. While this system proved to be useful and highly successful, it was observed that pJM17 was not absolutely noninfectious, being able to generate infectious virus, albeit at low frequencies, following single transfection into 293 cells. It was discovered that the restoration of infectivity of pJM17 was due to deletions of the bacterial plasmid sequences following transfection into 293 cells resulting in size reduction to within the packaging constraints of Ad. The two plasmid rescue method was refined in 1994 by Bett et al. [28]. In this iteration of the system, an improved Ad genomic plasmid, pBHGlO, was constructed to address the shortcoming of pJM17 and to introduce additional flexibility into the system. The plasmid pBHGlO contains essentially the entire Ad5 genome joined at the ITRs with two modifications. The first modification is a deletion from 0.5 to 3.7 mu, which removes the El A region as well as the packaging signal required for encapsidation of the adenoviral genome thus rendering the plasmid noninfectious. The second modification was removal of -^2.7 kb from the nonessential E3 region, from 78.3 to 85.8 mu, and introduction of a Fad restriction enzyme site. A series of shuttle plasmids were also developed to be used in conjunction with pBHGlO. These shuttle plasmids contained the
left 15.8 mu of the Ad5 genome including the left end ITR and the packaging signal but with a 3181-bp deletion in El from 0.9 to 9.8 mu into which a polylinker was introduced for transgene insertion. This version of the two-plasmid rescue system offered several improvements over the original system of McGrory et al, [27]. First, the combination of the El region deletion in the shuttle plasmid and the E3 region deletion in the Ad genomic plasmid increased the cloning capacity of the recombinant vector, permitting rescue of up to '^ 8 kb of foreign DNA. Second, recombinant vectors bearing foreign DNA insertions in the E3 region could be easily constructed by utilizing the unique Fad site in the Ad genomic plasmid. To simplify cloning into the large pBHG plasmids, insertion of foreign DNA into the Fad site is facilitated by using the kanamycin resistant pABS series of small shuttle plasmids (www.microbix.com). The pABS plasmids bear two Fad sites which flank a polylinker and the kanamycin resistance gene. The kanamycin resistance gene is flanked by Swal sites. The foreign DNA is first inserted into the polylinker of the pABS plasmid. The Fad fragment bearing 4. Adenoviral Vector Construction I: Mammalian Systems 7 9 the foreign DNA and the kanamycin resistant gene is then cloned into the Pad site of the pBHG plasmid. Following transformation of E. coli, positive clones bearing the E3 insertion are easily identified by their resistance to both ampicillin and kanamycin. Finally, the kanamycin resistance gene is collapsed out of the pBHG plasmid by Swal digestion and religation, leaving behind the foreign DNA in the E3 region. Thus, by using this system, a total of ~8 kb of foreign sequence could be easily rescued into the El and/or E3 regions of the recombinant vector. Third, the deletion of the packaging signal rendered pBHGlO absolutely noninfectious and, thus, all progeny virus generated following cotransfection were the desired recombinant. This version of the two-plasmid rescue system has become very popular for construction of El replacement vectors due to its versatility and simplicity; one need only clone the foreign DNA into the small shuttle plasmid and cotransfect it along with pBHGlO into 293 cells to generated recombinant vectors. However, one limitation of the two plasmid rescue method, especially for those not experienced in adenovirology, was the low efficiency of vector rescue if cells or transfection parameters were suboptimal. B. Refinements to the Two-Plasm id Rescue Method In 1999, Ng et al, [29] hypothesized that the low efficiency of vector res cue was due, in part, to the inefficiency of in vivo homologous recombination. Consistent with this hypothesis was the observation that the plaque-forming efficiency of the infectious Ad genomic plasmid pFG140 was ~100-fold higher than that achieved by a typical cotransfection for vector rescue. To address this limitation the two-plasmid rescue system was modified to make use of high efficiency site-specific recombination catalyzed by bacteriophage PI recombi- nase Cre instead of homologous recombination (Fig. 3B). To accomplish this, a loxP site was inserted into pBHGlO, 5' of the pIX gene and into the shuttle plasmid, y of the foreign transgene. Thus, vector rescue could be achieved by high efficiency Cre-mediated recombination between the two modified plas- mids following their cotransfection into 293 cells expressing Cre recombinase (293Cre4 [30]). Ng et al. [29] demonstrated that the efficiency of vector rescue by Cre-mediated recombination was ~30-fold higher than by homologous recombination. Further improvements were subsequently introduced when Ng et al. [31] demonstrated that replacement of the single ITR in the shuttle plasmid with a head to head ITR junction resulted in a 14-fold increase in the efficiency of homologous recombination mediated vector rescue. Combining Cre-mediated recombination and shuttle plasmids bearing an ITR junction increased the efficiency of vector rescue by ~ 100-fold over the earlier methods of McGrory et al. [27] and Bett et al. [28]. A number of nonmutually exclusive explanations were postulated to account for the effect of ITR junctions on vector rescue 8 0 Ng and Graham efficiency (Fig. 4). Based on the fact that ITR junctions serve as an efficient origin of viral DNA rephcation [25] in contrast to a single ITR linked to plasmid DNA and that the ITRs are the only ds-acting Ad sequences required for viral DNA replication, it w âs postulated that shuttle plasmids bearing an ITR junction, in contrast to shuttle plasmids having only a single ITR, were capable of virus-mediated DNA replication foUov^ing cotransfection of 293 cells with the Ad genomic plasmid which would supply all the trans-acting factors required for viral DNA replication. Thus, the increased vector rescue efficiency may reflect an increase in the substrate pool for recombination. In addition, since Ad DNA replicates as a linear molecule (reviewed in [32]), it is also possible that linearization of the shuttle plasmid by ITR junction-mediated DNA replication may produce a preferred substrate for recombination in contrast to shuttle plasmids bearing a single ITR which remains circular. Also, generation of an infectious genome may be more complex following recombination between a circular and a linear DNA molecule (single ITR shuttle plasmid and replicating Ad genomic plasmid) (Fig. 4A) than between two linear DNA molecules (replicating shuttle and Ad genomic plasmid) (Fig. 4B). In the former case (Fig. 4A), recombination would first result in integration of the circular substrate into the linear substrate. This intermediate is not packagable owing to its size [14] and the distance between the packaging signal and the genome terminus [33]. An infectious packagable genome could be generated from this intermediate by a second step following DNA replication in which the internal ITR is utilized as an origin of replication through the formation of a panhandle structure and repair, a process that is known to occur [34] but which might be less efficient than utilization of the terminal ITRs. In contrast, in the latter case (Fig. 4B), a packagable, infectious genome is generated immediately following recombination between two linear substrates. To further expand the versatility of this method, the system was modified to permit high-efficiency Cre-mediated vector rescue to be achieved using the ubiquitous 293 cells (or any other El-complementing cell line) thus abrogating the need for Cre-expressing cell lines which are not as widely available as the parental 293 cell line. To accomplish this, a Cre expression cassette was inserted into a region of the Ad genomic plasmid which would not contribute to the final recombinant vector genome but permitted transient Cre expression following cotransfection. The vector rescue efficiency following cotransfection of 293 cells using this Ad genomic plasmid was found to be nearly as high as with 293Cre4 cells [31]. One limitation of Cre-mediated vector rescue is that it would be unsuit able for constructing vectors bearing loxP sites elsewhere in the genome designed, for example, to regulate transgene expression [35-37] or to inhibit vector packagability [38] since it would lead to undesired Cre-mediated vector rearrangements. This was addressed by modifying the two plasmid rescue ^ o fl VH _0 3 O a o 3 oo •§7 T>S cCiDi o I <: a>_o S 1 t .E CL - " o E C5D -Q.< to D + -̂^̂ II % o %. % II •t=.2 ^o _1g D5Q. -a I.S O _jc < ? o a n I D L . E o o ^ ^ 1̂ > - D ^ _g) CO D E •T3 t o 0 'c E k ^ U E %, •% S "D "D cD ^ "D CO O Q - 'E«o D ^ Q_ o 41 £ 'E 3 o .0) c 81 8 2 Ng and Graham method to utilize the yeast FLP-mediated site-specific recombination sys tem [39]. With loxP sites in the Ad genomic and shuttle plasmids replaced with frt sites and the Cre-expression cassette replaced with a FLP-expression cassette in the Ad genomic plasmid the efficiency of FLP-mediated vector rescue was comparable to that mediated by Cre. The choice of either Cre- or FLP- mediated recombination further expanded the versatility of the two plasmid rescue method by permitting high-efficiency vector rescue in cases where one of the recombinases is unsuitable or undesirable for vector construction. C. The Ad Genomic Plasmid A variety of Ad genomic plasmids are available for construction of vectors by site-specific recombination (Fig. 5). The plasmids pBFIGloxESCre and pBHGfrtE3FLP bear a wild-type E3 region and are used to generate vectors by Cre-mediated and FLP-mediated recombination, respectively. Owing to the size constraints of Ad [19], the maximum foreign DNA insert that can be rescued into an El-deleted vector with a wild-type E3 region is ^ 5 kb. The plasmids pBHGloxAE3(Xl)Cre and pBHGfrtAE3(Xl)FLP have a 1864-bp deletion in the E3 region and thus permit foreign sequences up to ~7.2 kb to be rescued into vectors, whereas the plasmids pBHGloxAEl,3Cre and pBHGfrtAEl,3FLP have a 2653-bp deletion in the E3 region and permit rescue of up to ~ 8 . While these latter two plasmids offer maximum cloning capacity, vectors bearing this larger E3 deletion may grow slightly slower and result in lower yields ('^2-fold) than vectors bearing the wild-type E3 or smaller E3 deletion (F. L. Graham; unpublished results). As with the earlier pBHGlO based methods, the unique Pad sites in pBHGfrtAEl,3FLP, pBHGfrtAEl,3FLP, pBHGloxAE3(Xl)Cre, and pBHGfrtAE3(Xl)FLP permit insertion of foreign sequences into the E3 deletion for rescue into virus if desired. The choice of these Ad genomic plasmids is dictated by the size of the foreign sequence to be rescued, whether a wild-type or a deleted E3 region is desired and which site-specific recombination system is preferred/necessitated. D. The Shuttle Plasmid A variety of shuttle plasmids are available for insertion and rescue of foreign sequences into Ad vectors by Cre or FLP-mediated recombination (Fig. 6). The shuttle plasmids pDC311 and pDC312 are designed for rescue of expression cassettes into El by Cre-mediated recombination and pDCSll and pDC512 for rescue by FLP-mediated recombination. The shuttle plas mids pDC315, pDC316, pDC515, and pDC516 carry promoters and poly(A) sequences and are designed for insertion of coding sequences. PDC315 and pDC316 use Cre-mediated recombination and pDC515 and pDC516 use FLP- mediated recombination. The polycloning site in these plasmids is flanked by 4. Adenoviral Vector Construction I: Mammalian Systems 83 ITRs Cre ITRs AE3 Cre AE3 (-1864 bp) AiA (-1864 bp) Pad AEl Pad loxF ITRs AE3 Cre AE3 (-2653 bp) (-2653 bp) Pad Pad Figure 5 Ad genomic plasmids used for vector rescue by in vivo site-specific recombination. The plasmids pBHGloxE3Cre, pBHGloxAE3(Xl)Cre, and pBHGloxAEl,3Cre are used to rescue vectors by Cre-mediated recombination bearing a wildtype E3 region, a 1864 bp deletion or a 2653 bp deletion of E3, respectively. Analogous plasmids pBHGfrtE3FLP, pBHGfrtAE3(Xl )FLP and pBHGfrtAEl,3FLP are used to rescue vectors by FLP-mediated recombination. The unique Pad restriction enzyme site in pBHGloxAEl,3Cre, pBHGloxAE3(Xl)Cre, pBHGfrtAE3(Xl)FLP, and pBHGfrtAEl ,3FLP permit insertion of foreign sequences into the E3 deletion. Ad and bacterial plasmid sequences are represented by thick and thin lines, respectively, and loxP or frt sites are represented by ">". 84 Ng and Graham ITRs ITRs polylinker MCMV y promoter loxP polylinker amp SV40 polyA amp loxP ITRs ITRs polylinker y SV40 polyA amp EcoRI Xbal StuI Nhel ^ ,,, Hindlll Ecll36II ACCI Bglll Sad pDCBll 5'TCTAGAGAATTCAGGCCTGCTAGCAGATCTAAGCTTGAGCTCGTCGAC 3 ' EcoRI Nhel Hindlll Ecll36II ^ccl Xbal StuI Bglll Sad BamHI pDC511 5' TCTAGAGAATTCAGGCCTGCTAGCAGATCTAAGCTTGAGCTCGTCGACGGATCC 3>' Sail Acd Ecll36II Hjn^iii E Xbal Sad Bglll Nhel coRI pDC312 StuI BamHI pDC512 5 TCTAGAGTCGACGAGCTCAAGCTTAGATCTGCTAGCAGGCCTGAATTCGGATCC Sail EcoRI Nhel BamHI Acd pDC315 pDC515 5' GAATTCAAGCTGCTAGCAAGGATCCAGCTTGTCGAC 3' Smal Xmal ^ ,̂ ^ Hindlll EC1136II ^^fj EcoRI Bglll Sad pDC316 pDC516 5' GAATTCCCCGGGAGATCTAAGCTTGAGCTCGTCGAC 3' Figure 6 Shuttle plasmids and their polylinker sequences used for vector rescue by /n vivo site-specific recombination. The shuttle plasmids pDC311 (3276 bp), pDC312 (3288 bp), pDC315 (3913 bp), and pDC316 (3913 bp) are used to rescue vectors by Cre-mediated recombination. The shuttle plasmids pDC511 (3277 bp), pDC512 (3277 bp), pDC515 (3957 bp), and pDC516 (3957 bp) are used to rescue vectors by FLP-mediated recombination. The plasmids pDC311, pDC312, pDC511, and pDC512 are used for insertion of expression cassettes (inserts with a pro moter/enhancer and polyadenylation signal as well as coding sequence). The plasmids pDC315, pDC316, pDC515, and pDC516 bear a polylinker flanked by the murine cytomegalovirus immedi ate-early promoter/enhancer and SV40 polyadenylation signal and are used for insertion of coding sequences (e.g., cDNAs). 4.
Adenoviral Vector Construction I: Mammalian Systems 8 5 a murine cytomegalovirus (MCMV) immediate-early promoter and the SV40 polyadenylation signal for high level transgene expression in most cell types. The choice of shuttle plasmid is dictated by v^hether expression from a strong viral promoter is desired, by the orientation of the polylinker and by the site- specific recombination system desired for vector rescue. It has been observed that higher expression levels are obtained when the transcription orientation of the transgene is in the same direction as El and that the MCMV immediate early promoter is stronger in most cell lines than its more commonly used human counterpart [40]. III. Protocols for the Two-Plosmid Rescue System The remainder of this chapter provides detailed protocols for each of the steps involved in the rescue and propagation of recombinant Ad vectors. A flow chart of these steps is presented in Fig. 7. Briefly, 293 cells are cotransfected with the Ad genomic plasmid and the shuttle plasmid. The recombinant vector is generated by in vivo site-specific recombination between the two plasmids and forms a plaque in the cell monolayer. The plaques are isolated, the virus expanded, and the vector DNA is extracted for confirmation by restriction enzyme digestion. The vector is plaque purified by titration and a high titer stock is generated which is then purified by CsCl banding and characterized with respect to concentration, DNA structure, level of RCA contamination and transgene expression. It is recommended that all the steps outlined in Fig. 7 be followed. However, since all infectious viruses generated after cotransfection are the desired recombinant [29, 31, 39], vector production can be expedited if necessary by following one or more shortcuts indicated in Fig. 7 and described in section III.I. A. Preparation of Plasmid DNA The foreign DNA is inserted into an appropriate shuttle plasmid and transformed into £. coli by conventional molecular biology techniques. This section describes the preparation of high-quality plasmid DNA for cotransfec tion. MATERIALS 1. Plasmid DNA: All plasmids described in this chapter and their sequences can be obtained from Microbix Biosystems Inc. (www.microbix.com). 2. Sterile LB broth (Lennox) (Difco) and LB-agar plates supplemented with 50 \|xg/mL ampicillin. Optional: Sterile Super Broth; LB broth 86 Ng and Graham Clone foreign DNA into shuttle plasmid Rescue and isolate Ad vector (section IIIC) Analyze Ad vectors and prepare vector lysate (section HID) Plaque purify Adv ector by titration (section HIE) Analyze Ad vector and prepare vector lysate (section HID) Prepare high titer vector stocks (section IIIF) Purify Ad vector by CsCl banding (section IIIG) Characterize Ad vector (section IIIH) Figure 7 Overview of the steps involved in rescue, propagation, purification, and characterization of Ad vectors. The recommended steps are indicated by thick arrows. Acceptable alternatives to expedite vector production are indicated by thin arrows (see section III.I). 4. Adenoviral Vector Construction I: Mammalian Systems 8 7 supplemented with 22 g/mL peptone, 15 g/mL yeast extract, 1 g/mL D-glucose, 0.005 N NaOH, and 50 \|JLg/mL ampicillin. 3. Solution I: 10 mM EDTA, pH 8.0, 50 mM glucose, 25 mM Tris, pH 8.0, prepared from sterile stock solutions. 4. Solution 11:1% SDS, 0.2 N NaOH, freshly prepared. 5. Solution III: 3 M potassium acetate, 11.5% glacial acetic acid, auto clave sterilized. 6. Isopropanol. 7. TE: 10 mM Tris, pH 8.0, 1 mM EDTA, pH 8.0, autoclave sterilized. 8. Pronase stock solution: 20 mg/mL pronase in 10 mM Tris, pH 7,5-^ preincubate at 56° C for 15 min, followed by 37°C for 1 h. Aliquot and store at —20°C. 9. Pronase-SDS solution: 0.5 mg/mL pronase (above) in 0.5% SDS, 10 mM Tris, pH 7.4, 10 mM EDTA pH 8.0. 10. CsCl (biotechnology grade). 11. 10 mg/mL ethidium bromide. METHOD 1. Inoculate 5 mL of LB supplemented with 50 \|xg/mL ampicillin with bacteria bearing the desired plasmid in the morning. Incubate culture at 37°C with shaking. For the large Ad genomic plasmid, bacterial cultures should be started from well-isolated colonies picked from a bacterial plate less than 1 week old. 2. Inoculate 500 mL of LB supplemented with 50 \|xg/mL ampicillin with the above culture in the late afternoon. Incubate overnight at 37°C with shaking. Optional: For higher yields of plasmid DNA use richer medium such as Super Broth. 3. Transfer culture to a centrifuge bottle and pellet bacteria by spinning at 6000 g for 10 min at 4°C. Resuspend bacterial pellet in 40 mL of cold solution I so that no cell clumps are visible. 4. Add 80 mL of freshly prepared solution II, mix thoroughly but gently by swirling to produce a relatively clear, viscous lysate. 5. Add 40 mL of cold solution III, mix thoroughly but gently by swirling and incubate for 20 min on ice. The viscosity should be greatly reduced and a white precipitate should form. 6. Add 10 mL of dHiO and centrifuge at 4°C for 10 min at 6000 g. 7. Collect the supernatant by filtering it through two to three layers of cheesecloth into a centrifuge bottle. 8. Add 100 mL (0.6 vol) of isopropanol, mix well, and incubate for 30 min at room temperature to precipitate plasmid DNA, centrifuging at 4°C for 10 min at 6000 g to pellet plasmid DNA. 9. Discard the supernatant and air dry the pellets for 15 min. Wipe inside the rim with a clean Kim Wipe to remove all residual isopropanol. 8 8 Ng and Graham 10. Dissolve plasmid DNA pellet in 5 mL TE and transfer to a 50-mL conical tube. 11. Add 2 mL pronase-SDS solution. Mix well and incubate for 30 min at 37°C. 12. Add 8.6 g CsCl, mix to dissolve completely, and incubate on ice for 30 min. 13. Centrifuge at 3000 g for 30 min at 5°C. Slowly collect the supernatant using a 10 cc syringe and 16-gauge needle, avoiding as much of the pellicle as possible. 14. Transfer to a VTi65 ultracentrifuge tube. Add 25 \|xL of 10 mg/mL ethidium bromide and fill the tube with light parafin oil. 15. Seal the tube and mix by inversion. Centrifuge in a Beckman VTi 65.1 rotor at 55,000 rpm for 10-14 h at 14°C. 16. Remove tube and support it with a stand. The supercoiled plasmid DNA band should be the thick red band in the gradient. Puncture the top of the tube to allow entry of air and collect the plasmid DNA through the side of the tube with a 3 cc syringe and 18-gauge needle by puncturing the side of tube just below the band. Except when recovering plasmid DNA bands, keep the tubes in the dark or covered with foil to avoid unnecessary exposure to fluorescent or UV light. 17. Transfer plasmid DNA to a 15-mL polypropylene tube containing 5-mL isopropanol which has been saturated with CsCl in TE. Mix immediately to extract the ethidium bromide into the solvent layer. Allow the phases to separate and discard the ethidium bromide-solvent (pink) layer. Repeat extraction until the solvent layer is colorless. 18. Add TE to bring the volume up to 4 mL, add 8 mL cold 95% ethanol, and mix by inversion to precipitate the DNA. 19. Spin at 3000 g in a table-top centrifuge at room temperature for 15 min to pellet DNA. Wash pellet twice with 5 mL 70% ethanol. 20. Remove as much of the 70% ethanol as possible, allow the pellet to dry, and resuspend with an appropriate volume of TE. Ideally, the concentration should be 1 to 2 \|Jig/\|xL. 21. Determine the plasmid DNA concentration by OD260 and digest a sample with appropriate diagnostic restriction enzymes and confirm the structure by agarose gel electrophoresis. B. Cell Culture Low-passage 293 cells are maintained in 150-mm dishes and are split 1 to 2 or 1 to 3 when they reach confluency (every 2 to 3 days). Generally, a ~90% confluent 150-mm dish of 293 cells is split into 10 60-mm dishes for use the next day for cotransfections. Never allow the cells to become overconfluent or to be seeded too thinly. Change the medium regularly between splits (twice weekly 4. Adenoviral Vector Construction I: Mammalian Systems 8 9 if they are not growing rapidly enough to permit spUtting every 2 - 4 days). A sufficient number of ampoules of the cells should be frozen and stored in liquid N2 to permit initiation of new cultures when the passage number of the lab stocks has reached 40-45 passages or when the cells are no longer behaving well under agar overlays (see sections III.C and III.E). Higher passage or poorly adherent cells which are unsuitable for cotransfections or titrations may still be adequate for virus propagation. 293N3S are suspension-adapted 293 cells and can be used for large-scale vector production instead of 293 cells due to greater ease of handling. MATERIALS 1. Low-passage 293 and 293N3S cells (Microbix Biosystems Inc.). 2. Complete medium: MEM (Gibco BRL 61100) containing 10% fetal bovine serum (FBS) (heat inactivated), 100 units/mL penicillin/strepto mycin, 2 mM L-glutamine, and 2.5 \|xg/mL fungizone. 3. Joklik's modified-MEM (Gibco BRL 22300) supplemented with 10% horse serum (heat inactivated). 4. Citric saline: 135 mM KCl, 15 mM sodium citrate, autoclave steril ized. 5. Spinner flasks (Bellco). Prewarm all cell culture reagents to 37°C prior to use. METHOD 1. Remove medium from 150-mm dish of 293 cells and rinse monolayer twice with 5 mL citric saline. 2. Remove citric saline from step 1, add 0.5 mL citric saline, and leave the dish at room temperature until cells start to round up and detach from the dish (no more than 15 min). 3. Tap the side of the dishes to detach all cells. 4. Resuspend cells with complete medium and distribute into new dishes. 293N3S cells are grown at 37°C in spinner flasks in Joklik's modified MEM supplemented with 10% horse serum (heat inactivated) and should be diluted 1 to 2 or 1 to 3 when the density reaches 5 x 1 0 ^ cells/mL. C. Cotransfection Under optimal conditions, large numbers of plaques are typically gener ated by cotransfecting a single 60-mm dish of 293 cells with 2 \|jig of the shuttle plasmid and 2 \|JLg of the Ad genomic plasmid by site-specific recombination (Table I). However, many factors can influence the efficiency of vector rescue including the quality of the DNA, the efficiency of transfection and especially the state of the 293 cells. Another important consideration is that the plaques 90 Ng and Graham Table 1 Vector Rescue Efficiency by in Vivo Site -Specific Recombination" Shuttle plasmid Ad genomic: plasmid Average plaques/60-mm dish pCA35loxAITR pBHGloxAEl,3Cre 43 pBHGloxAE3(Xl)Cre 63 pBHGloxE3Cre 48 pCA35frtAlTR pBHGfrtAEl,3FLP 41 pBHGfrtAE3(Xl)FLP 27 pBHGfrtE3FLP 25 pFG140^ 103 ^60-mm Dishes of 293 cells were cotransfected with 2 \|jLg of each plasmid and plaques were counted 10 days postcotransfection. ^60-mm Dishes of 293 cells were transfected with 0.5 \xg of pFG140 and plaques were counted 10 days postcotransfection. be well isolated. Thus, it is recommended that a range of DNA amounts be cotransfected to ensure that plaques are obtained and that they are well iso lated. The infectious Ad genomic plasmid pFG140 [18] provides a control for transfection efficiency and under optimal conditions should yield up to ~100 plaques per 0.5 \|jLg. The following is a protocol in which four 60-mm dishes of 293 cells are cotransfected with 0.5, 2, and 5 \|jLg of each plasmid (Fig. 8). MATERIALS 1. Monolayers of low passage 293 cells at ^^80 to 90% confluency in 60-mm dishes. 2. Hepes-buffered saline (HBS): 21 mM Hepes, 137 mM NaCl, 5 mM KCl, 0.7 mM Na2HP04, 5.5 mM glucose, pH 7.1 (adjusted with NaOH), filter sterilized. Store at 4°C in small aliquots in tightly sealed plastic conical tubes. 3. Salmon sperm DNA (2 \|jig/\|JLL in TE). 4. 2.5 M CaCli, filter sterilized. 5. Complete medium (see section III.B) 6. 2x Maintenance medium: 2x MEM (Gibco BRL 61100) sup plement with 10% horse serum (heat inactivated), 200 units/mL penicillin/streptomycin, 4 mM L-glutamine, 5 \|xg/mL fungizone, and 0.2% yeast extract. 7. 1 % Agarose solution, autoclave sterilized. Store at room temperature and melt in a microwave oven prior to use. 8. Ad genomic plasmid DNA (see section II.C). 9. Shuttle plasmid DNA with the desired foreign sequence inserted (see section II.D). 4. Adenoviral Vector Construction I: Mammalian Systems 91 8 ml HBS + 40 \|il salmon sperm
DNA, vortex for 1 min V i ; T 2ml 2ml 2ml 1ml 66 A 99 "B" u u u u Shuttle 20 \| ig plasmid 8 \| ig 2\|^g Ad genomic 20 \| ig l \|^g plasmid 8l ig 2 \| l g (pFG140) 2.5 M CaCl^ 100 III 100 III 100 III 50 III 0.5 ml/60 mm dish i i i I Figure 8 Standard protocol for Ad vector rescue by in vivo site-specific recombination. 10. Phosphate-buffered sahne (PBS): 137 mM NaCl, 8.2 mM Na2HP04, 1.5 mM KH2PO4, 2.7 mM KCl, autoclave sterihzed. 11. PBS++: PBS supplemented with 0.68 mM sterile MgCli and 0.5 mM sterile CaCli. 12. Glycerol, autoclave sterilized. 9 2 Ng and Graham METHOD 1. Label four 60-mm dishes "A", four dishes "B", four dishes "C", and two dishes "D". Seed these dishes with 293 cells to reach ~ 8 0 % confluency in 1 to 2 days for cotransfection. 2. In the late afternoon, 1 hour prior to cotransfection, replace the medium from the 60-mm dishes of 293 cells with 5 mL of freshly prepared complete medium without washing. 3. Meanwhile combine in a 15 mL conical tube 8 mL of HBS and 40 \|JLL of salmon sperm DNA and vortex at maximum setting for 1 min. 4. Add 2 mL of the above solution to each of three polystyrene tube labeled "A", "B", and "C 'and 1 mL to a fourth polystyrene tube labeled "D". 5. Add 2 \|xg of shuttle plasmid DNA and 2 \|xg of Ad genomic plasmid DNA to tube "A" (this will result in 0.5 \|xg of each plasmid per dish). Add 8 \|jLg of shuttle plasmid DNA and 8 \|JLg of Ad genomic plasmid DNA to tube "B" (2 \|xg of each plasmid per dish). Add 20 \|xg of shuttle plasmid DNA and 20 \|xg of Ad genomic plasmid DNA to tube " C " (5 \|jLg of plasmid per dish). Add 1 \|xg of pFG140 DNA to tube "D". Gently mix each tube thoroughly. 6. To tubes ' 'A", "B", and " C " add 100 \|JLL of 2.5 M CaCli dropwise with gentle mixing. To tube " C " add 50 \|xL of 2.5 M CaCli dropwise with gentle mixing. Incubate the tubes at room temperature for 30 min. The solutions should become slightly cloudy. 7. Add 0.5 mL of the contents in tube "A"dropwise to the monolayer in each of the dishes labeled "A" without removing the medium. Repeat for tubes "B", "C", and "D". Distribute the precipitate evenly by rocking the dishes and return to the incubator. 8. The following morning, melt 1% agarose solution in a microwave oven and allow it to equilibrate to 45°C. Equilibrate 2x maintenance medium to 37°C. Prepare overlay solution by combining 7S mL of melted 1 % agarose and 75 mL of 2 x maintenance medium. 9. Remove the medium from each of the cotransfected dishes and add 10 mL of overlay solution prepared in step 8. Perform this step quickly to prevent the overlay solution from prematurely solidifying but gently to prevent disturbing the monolayer. 10. Allow the overlay to solidify at room temperature (10 to 15 min) and then return the dishes to the incubator. Plaques should begin to appear in 5 days and will continue to appear until about 12 to 14 days post-cotransfection. 11. Ten days post-cotransfection, pick well isolated plaques from the monolayer by punching out agar plugs with a sterile cotton plugged Pasteur pipet attached to a rubber bulb. It is recommended that 4. Adenoviral Vector Construction I: Mammalian Systems 9 3 plaques be isolated at about 10 days post-cotransfection to ensure that those chosen are well isolated with no plaques overlapping. 12. Transfer the agar plugs into 0.5 mL PBS"̂ + supplemented with glycerol to 10% in a suitable vial. Vortex briefly and store at —70°C. D. Analysis of Recombinant Vectors and Preparation of Working Vector Stocks Once plaques have been isolated, the viruses are expanded for extraction of vector DNA for analysis and to yield a working vector stock. MATERIALS 1. 90% Confluent 60-mm dishes of 293 cells. 2. TE (see section II.A). 3. Complete medium (see section II.A). 4. Maintenance medium: MEM (Gibco BRL 61100) containing 5% horse serum (HS) (heat inactivated), 100 units/mL penicillin/strepto mycin, 2 mM L-glutamine, and 2.5 \|jLg/mL fungizone. 5. PBS++ (see section II.C). 6. Pronase-SDS solution (see section II.A). METHOD 1. Seed 60-mm dishes of 293 cells (one per plaque) to reach ^^90% confluency on the day of use. 2. Thaw virus plaque picks and vortex briefly. Remove medium from the 60-mm dishes of 293 cells and add 250 \|xL of the plaque pick. Adsorb for 1 h in the incubator rocking the dishes every 10 to 15 min. 3. Following adsorption, add 5 mL of maintenance medium and return dishes to incubator until complete cytopathic effect (CPE) is observed (>90% cells rounded up and detached from dish, usually 4 to 5 days postinfection). It is important that the DNA be extracted following complete CPE so that vector DNA bands are clearly visible above the background of cellular DNA. If complete CPE is not reached by 5 days postinfection (most likely due to low multiplicity) then scrape the monolayer into the medium and transfer the cell suspension into a suitable vial and supplement with glycerol to 10%. Freeze (—70°C)-thaw the cell suspension and infect 60-mm dishes of 90% confluent 293 cells with 0.2 to 0.4 mL as described above. Complete CPE should be observed within 5 days and the vector DNA can be extracted. 4. Once complete CPE is reached, the dishes are processed as follows: Scrape the cells into the medium and transfer 1.5 mL of the cell suspension into an eppendorff tube for vector DNA extraction (see 9 4 Ng and Graham step 5). Transfer the remainder of the cell suspension into a suitable vial, supplement with glycerol to 10% and store at — 70°C. This lysate should contain a significant amount of virus ('^lO^ pfu/mL) and can be used for plaque purification of the vector (section III.E) or can be used in preliminary experiments or for further vector expansion (section III.F). 5. To extract vector DNA, pellet cells by spinning at 3000 rpm in a microcentrifuge for 5 min. 6. Discard supernatant, resuspend the cell pellet in 0.2 mL pronase-SDS solution and incubate tubes at 37°C overnight. 7. Add 0.2 mL dHiO and 1 mL 95% ethanol and mixing by inversion until the DNA precipitate is formed. 8. Pellet DNA by spinning in a microcentrifuge (maximum speed for 2 min) and wash pellet twice with 70% ethanol. Let the pellet dry and resuspend in an appropriate volume of TE (~35 JJLL). Dissolve DNA by heating at 65°C with occasional vortexing. 9. Digest 5 to 10 (JLL of the DNA with an appropriate restriction enzyme. Analyze the DNA structure by agarose gel electrophoresis to verify that the DNA structure of the recombinant virus is correct. If the infection of 293 cells has been complete, viral DNA bands should be readily visible superimposed on a smear of cellular DNA. Once the DNA structure of the vector has been verified the virus can be plaque purified (section IILE). E. Titration of Adenovirus The procedure outlined below is used to plaque purify recombinant vectors as well as to determine the concentration of vector stocks. To accurately determine vector concentration, titrations should be performed in duplicate. MATERIALS 1. 80 to 90% Confluent 60-mm dishes of 293 cells. 2. PBS++ (see section III.C). 3. 1% Agarose solution (see section III.C). 4. 2x Maintenance medium (see section III.C). 5. Glycerol, autoclave sterifized. METHOD 1. Seed 60-mm dishes of 293 cells to reach ^80 to 90% confluency in 1 to 2 days for titration. 2. Prepare serial dilutions of the recombinant virus in PBS"̂ + (10~^ to 10~^ for samples prepared in section III.D and 10~^ to 10~^^ for samples prepared in sections III.F and III.G). 4. Adenoviral Vector Construction I: Mammalian Systems 9 5 3. Remove the medium from the 60-mm dishes of 293 cells and infect with 0.2 mL of the diluted samples. Return dishes to the incubator and adsorb for 1 h, rocking the dishes every 10 to 15 min. 4. During the adsorption period, melt 1% agarose solution in a micro- v^ave oven and equilibrate to 45°C. Equilibrate 2x maintenance medium to 37°C. 5. Follov^ing 1 h adsorption, combine equal volumes of melted agarose solution w îth 2 x maintenance medium, mix well, and gently overlay dishes with 10 mL. Perform this step quickly to prevent the overlay solution from solidifying prematurely but gently to prevent disturbing the monolayer. 6. Allow overlay to solidify for 10 to 15 min at room temperature and then return dishes to the incubator. 7. Plaques should start to appear about 4 days postinfection and should be counted 10 to 12 days postinfection. For isolation of recombi nant virus by plaque purification well isolated plaques should be picked according to steps 11 and 12 of section III.C around 10 days postinfection. The plaque purified vectors are expanded according to section III.D and used as inoculum for the preparation of high-titer viral stocks (section III.F). 8. Determine the vector concentration in plaque forming units per ml (pfu/mL) as follows: titer = (number of plaques)(dilution factor)/(infection volume) Calculate the titer from dishes bearing approximately 20 to 80 plaques. The number of plaques should vary in direct proportion to the dilu tion factor; otherwise, repeat the titration making sure that the samples are thoroughly mixed when setting up the serial dilutions. F. Preparation of High-Titer Viral Stocks (Crude Lysate) Since most of the virus remains associated with the infected cells until very late in infection (i.e., until the cells lyse), high-titer stocks can be easily prepared by concentrating infected 293 cells. The following protocol describes the production of high titer virus preparations using either monolayers of 293 cells or suspension cultures of 293N3S cells. 293N3S cells are preferable for the production of very large amounts of high-titer viral stocks due to the greater ease of handling suspension cultures. The following describes protocols for the preparation of crude lysates of high-titer vector stocks that are suitable for most experiments. Prior to the preparation of high-titer stocks, confirm that enough inoculum is available and if not, prepare an intermediate-scale virus stock by infecting two to three 150-mm dishes of 293 cells. 9 6 Ng and Graham 1. Preparation of High-Titer Viral Stocks (Crude Lysate) from Ceils in Monolayer MATERIALS 1. PBS++ (see section III.C). 2. Glycerol, autoclave sterilized. 3. 150-mm dishes of 80 to 90% confluent 293 cells. 4. Maintenance medium (see section III.D). METHOD 1. Seed 150-mm dishes with 293 cells to be 80-90% confluent at time of infection. The number of dishes is dictated by the amount of vector desired. 2. Dilute vector sample prepared in section III.E, step 7 1:8 v^ith PBS++. 3. Remove medium from the 293 cells and add 1 mL of the diluted vector sample prepared in step 2 to each 150-mm dish of cells (moi of 1-lOpfu/cell). 4. Adsorb for 1 h in the incubator, rocking the dishes every 10 to 15 min. FoUow îng adsorption, add 25 mL maintaince medium and return dishes to the incubator. Examine daily for signs of CPE. 5. When CPE is nearly complete (most cells rounded but not yet detached) harvest by scraping the cells into the medium and centrifuging the cell suspension at 800g for 15 min. 6. Discard the supernatant and resuspend the cell pellet in 2 mL PBS^"^ supplemented w îth glycerol to 10% for each 150-mm dish infected. Freeze (—70°C) and thaw^ the crude virus stock prior to characteriza tion of the vector (section IILH). Store aliquots at —70°C. 2. Preparation of High-Titer Viral Stocks (Crude Lysate) from Cells in Suspension MATERIALS 1. 293N3S cells (Microbix Biosystems Inc.). 2. Joklik's modified MEM supplemented v^ith 10% horse serum (heat inactivated) (see section III.B). 3. Spinner flasks (Bellco) 4. 1% Sodium citrate. 5. Carnoy's fixative: add 25 mL glacial acetic acid to 75 mL methanol. 6. Orcein solution: add 1 g orcein dye to 25 mL glacial acetic acid plus 25 mL dHiO; filter through Whatman No. 1 paper. 4. Adenoviral Vector Construction 1: Mammalian Systems 9 7
METHOD 1. Grow 293N3S cells to a density of 2 - 4 x 10^ cells/mL in 4 L complete Joklik's modified MEM supplemented with 10% HS. Centrifuge cell suspension at 750g for 20 min and save half of the conditioned medium. Resuspend the cell pellet in 0.1 vol fresh medium, and transfer to a sterile 500-mL bottle containing a sterile stir bar. 2. Add virus at an MOI of 1-20 pfu/cell and stir gently at 37°C. After 1 h, return the cells to the 4-L spinner flask and bring to the original volume using 50% conditioned medium and 50% fresh medium. Continue stirring at 37°C. 3. Monitor infection daily by inclusion body staining as follows: (a) Remove 5 mL from the infected spinner culture. Spin for 10 min at 750g and resuspend the cell pellet in 0.5 mL of 1% sodium citrate. (b) Incubate at room temperature for 10 min and then add 0.5 mL Carnoy's fixative and fix for 10 min at room temperature. (c) Add 2 mL Carnoy's fixative and spin 10 min at 750g. Discard supernatant and resuspend the pellet in a few drops of Carnoy's fixative. Add one drop of fixed cells to a slide and air dry for about 10 min Add one drop orcein solution and a coverslip and examine using a microscope. Inclusion bodies appear as densely staining nuclear structures resulting from accumulation of large amounts of virus and viral products at late times postinfection. Include a negative control in initial tests. 4. When inclusion bodies are visible in 80-90% of the cells (~3 days depending on the input MOI), harvest by centrifugation at 75Og for 20 min in sterile 1-L bottles. Combine pellets in a small volume of medium, and spin again. 5. Discard supernatant and resuspend pellet in 20 ml PBS'"+ supplemented with 10% glycerol. Freeze (—70°C)-thaw and then aliquot and store at — 70°C and characterize vector as described in section III.H. G. Purification of Adenovirus by CsCi Banding Many experimental studies can be performed using virus in the form of crude infected cell lysates prepared as described in sections III.E and III.F. However, for some experiments, particularly for animal work, it is desirable to use purified virus. The following protocol describes a method for purifying vectors obtained from 4-L of infected 293N3S cells or 30 x 150-mm dishes of 293 cells by CsCl banding. 9 8 Ng and Graham MATERIALS 1. 10 and 100 mM Tris, pH 8.0, autoclave sterilized. 2. 5% Sodium deoxycholate, filter sterilized. 3. 2 M MgCl2, autoclave sterilized. 4. DNAase I (100 mg bovine pancreatic deoxyribosenuclease I in 10 mL of 20 mM Tris, pH 7.4, 50 mM NaCl, 1 mM dithiothreitol, 0.1 mg/mL bovine serum albumin, 50% glycerol, aliquoted and stored at -20°C). 5. CsCl solutions: Density (g/cc) CsCl (g) 10 mM Tris, pH 8.0 (g) 1.5 90.8 109.2 1.35 70.4 129.6 1.25 54.0 146.0 Dissolve CsCl into 10 mM Tris, pH 8.0, solution in the amounts indicated above to achieve the desired density solution and filter sterilize. Weigh 1.00 mL to confirm density. 6. Glycerol, autoclave sterilized. 7. Beckman SW41 and SW50 rotor and ultraclear tubes. 8. Slide-A-Lyzer dialysis cassettes (Pierce). METHOD 1. Prepare crude cell lysate from infected cells as follows: (a) For 30 x 150-mm dishes as prepared in section IILE.l: when complete CPE is evident, scrape the cells into the medium, transfer the cell suspension to a centrifuge bottle, and spin for 10 min at 750g, Resuspend the cell pellet in 15 ml 0.1 M Tris-Cl, pH 8.0. Sample can be stored at — 70°C. (b) For 4-L spinner cultures prepared in section in.E.2: when inclu sion bodies are visible in 80-90% of the cells, harvest cells by centrifugation at 750g for 20 min in sterile bottles. Resuspend pellet in 15 mL 0.1 M Tris, pH 8.0. Samples can be store at —70°C. 2. Thaw sample and add 1.5 mL 5% Na deoxycholate for each 15 mL of cell lysate. Mix well and incubate at room temperature for 30 min. This results in a highly viscous suspension. 3. Add 150 \|iL 2 M MgCli and 75 \|xL DNAase I solution to each 15 mL of cell lysate, mix well, and incubate at 37°C for 60 min, mixing every 10 min. The viscosity should be greatly reduced. 4. Spin at SOOOg for 15 min at 5°C in the Beckman table-top centrifuge. 5. Meanwhile, prepare CsCl step gradients (one SW41 ultraclear tube for each 5 mL of sample): Add 0.5 mL of 1.5 g/cc solution to each tube. Gently overlay with 3.0 mL of 1.35 g/cc solution. Gently overlay this with 3.0 mL of 1.25 g/cc solution. 4. Adenoviral Vector Construction I: Mammalian Systems 9 9 6. Apply 5 mL of supernatant from step 4 to the top of each gradient. If necessary, top off tubes with 0.1 M Tris, pH 8. 7. Spin at 35,000 rpm in an SW41 rotor at 10°C, for 1 h. 8. Collect virus band (should be at 1.25 d/1.35 d interface) with a needle and syringe by piercing the side of the tube. The volume collected is unimportant at this stage so try to recover as much of the virus band as possible. If more than one tube was used, pool virus bands into a single SW50.1 ultraclear tube. 9. Top off tubes with 1.35 g/cc CsCl solution if necessary and centrifuge in a SW50.1 rotor at 35,000 rpm, 4°C, for 16-20 h. (Alternatively, the pooled virus can be centrifuged in the SW41 rotor at 35,000 rpm, 10°C, 16-24 h.) 10. To collect the virus band, puncture the side of the tube just below the virus band with a needle and syringe. Collect the virus band in the smallest volume possible and transfer to a Slide-A-Lyzer dialysis cassette. Dialyze at 4°C against three changes of 500 mL 10 mM Tris, pH 8.0, for at least 24 h total. 11. After dialysis, transfer the virus to a suitable vial and add sterile glycerol to a final concentration of 10%. Store the purified virus in small aliquots at — 70°C. H. Characterization of Adenoviral Vector Preparations Before the recombinant vector is used for experimentation the concen tration should be determined, the DNA structure should be confirmed and expression of the transgene should be ascertained. MATERIALS 1. All materials Usted in section III.E. 2. TE (see section III.A). 3. lOmMTris , pH8.0 4. 10% SDS 5. Pronase-SDS solution (see section III.A). 6. 3 M Sodium acetate, pH 5.2, autoclave sterilized. 7. 95 and 70% Ethanol. METHOD The concentration in pfu/ml is determined by titration on 293 cells as describe in section III.E. The concentration of virus particles, based on DNA content at OD26O can also be determine spectrophotometrically as follows: 1. Dilute (usually 20-fold) purified virus with TE supplemented with SDS to 0 .1%. Set up blank the same except add virus storage buffer 1 0 0 Ng and Graham (10 mM Tris, pH 8.0, supplemented with glycerol to 10%) instead of virus. 2. Incubate for 10 min at 56°C. 3. Vortex sample briefly. 4. Determine OD260- 5. Calculate the number of particles/mL, based on the extinction coeffi cient of wildtype Ad as determined by Maizel et al, [41] as follows: (OD26o)(dilutionfactor)(l.l x 10^^). The DNA structure of the recombinant vector should be confirmed following large-scale preparation. Virion DNA can be extracted from CsCl banded virus for analysis as follows: 1. An appropriate volume (~25 \|JLL depending on the concentration of the virus) of the purified virus is added to pronase-SDS solution to a final volume of 0.4 mL and incubated overnight at 37°C to lyse the virions and digest virion proteins. 2. Virion DNA is precipitated by adding 1/10 vol 3 M sodium acetate, pH 5.2, and 2.5 vol 95% ethanol and incubating at —20°C for 15 to 30 min. 3. Spin in microcentrifuge for 10 to 15 min at maximum speed. 4. Discard supernatant and wash DNA pellet twice with 70% ethanol. 5. Dry DNA pellet and resuspend in an appropriate volume of TE. For crude preparations, viral DNA can be extracted following infection of 293 cells as described in section III.D. A sample of the vector DNA is digested with the appropriate diagnostic restriction enzyme(s) and the structure of the DNA is analyzed by agarose gel electrophoresis. 293 cells contain nts 1-4344 bp of Ad5 DNA [42] with consequent homology flanking the expression cassette of generation vectors. Therefore, the possibility exists that homologous recombination between the Ad vector and the Ad sequences present in 293 cells may result in the formation of E1+ replication-competent Ad (RCA). The frequency with which Ad vectors recombine with Ad sequences in 293 cells is unknown but in general E1+ RCA replicate faster than El~ vectors. Consequently the proportion of RCA increases with prolonged propagation of the vectors in 293 cells. RCA can act as a "helper" virus resulting in the mobilization of the replication-deficient El-substituted vector in coinfected cells as well as cause tissue damage and pathogenicity. To minimize RCA contamination, vectors should not be serially propagated indefinitely. It is recommended that large-scale vector prepara tions be initiated from a stock prepared immediately after plaque purification (section III.E). If the original plaque purified stock is exhausted, plaque purifi cation can be repeated. The presence and level of RCA contamination in vector stocks should be determined, especially if the vector is to be used for extensive 4. Adenoviral Vector Construction I: Mammalian Systems 1 0 1 experimentation. A number of different approaches has been developed for the detection of RCA, including Southern blot hybridization [43], quantitative PCR [43] and biological assays [44]. I. Alternative Procedures to Expedite Vector Production It is recommended that the steps outlined in Fig. 7 and detailed in the preceding sections be followed as they are well proven. However, since only the correct recombinant vector should be generated following cotransfection [29, 31, 39], several alternative procedures are acceptable to expedite vector production (Fig. 7). 1. Once the vector has been rescued following cotransfection (section III.C), it can be immediately titrated for plaque purification and the DNA structure can be checked afterward. 2. While plaque purification is strongly recommended (section III.E), especially if large quantities of the vector are to be generated for extensive experimentation, this step is not absolutely essential since all infectious virus generated should be the desired recombinant. Optionally, therefore, high-titer stocks can be generated directly from the plaques isolated following cotransfection. 3. As mentioned in section III.F, vector purification by CsCl banding, while recommended, may not be necessary for many experiments. 4. It is strongly recommended that the recombinant vector be isolated from individual plaques following cotransfection using the method described in section III.C. However, vector production can be expe dited by omitting steps 8 through 12 in section III.C. In this case, following overnight cotransfection, remove the medium from the monolayers and add 5 mL of maintenance medium. If complete CPE is observed within 7 days post-cotransfection, then proceed from step 4 in section III.D. If complete CPE is not observed by 7 days post-cotransfection see section III.D, step 3. Acknovsfledgments This work was supported by grants from the National Institutes of Health, the Canadian Institutes of Health Research (CIHR), the National Cancer Institute of Canada (NCIC), and by Merck Research Laboratories. P.N. was supported by a CIHR Postdoctoral Fellowship. References 1. Berkner, K. L. (1988). Development of adenovirus vectors for expression of heterologous genes. Biotechniques 6, 616-629. 1 0 2 Ng and Graham 2. Graham, F. L., and Prevec, L. (1992). Adenovirus-based expression vectors and recombinant vaccines. In "Vaccines: New Approaches to Immunological Problems" (R. W. Ellis, Ed.), pp. 363-389. Butterv^orth-Heinemann, Boston, MA. 3. Hitt, M., Addison, C. L., and Graham, F. L. (1997). Human adenovirus vectors for gene transfer into mammaHan cells. Adv. Pharmacol. 40, 137-206. 4. Hitt, M. M., Parks, R. J., and Graham, F. L. (1999). Structure and genetic organization of adenovirus vectors. In "The Development of Human Gene Therapy" (T. Friedman, Ed.), pp. 61-86. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 5. Shenk, T. (1996). Adenoviridae: The viruses and their replication. In "Fields Viology" (B. N. Fields, D. M. Knipe, and P. M. Hov^ely, Eds.), pp. 2111-2148. Lipponcott-Raven, Philadelphia, PA. 6. Bergelson, J. M., Cunningham, J. A., Droguett, G., et al. (1997). Isolation of
a common receptor for Coxsackie B viruses and adenoviruses 2 and 5. Science 275, 1320-1323. 7. Hong, S. S., Karayan, L., Tournier, J., Curiel, D. T., and Boulanger, P. A. (1997), Adenovirus type 5 fiber knob binds to MHC class I alpha2 domain at the surface of human epitheHal and B lymphoblastoid cells. EmboJ. 16, 2294-2306. 8. Wickham, T. J., Mathias, P., Cheresh, D. A., and Nemerow, G. R. (1993). Integrins alpha v beta 3 and alpha v beta 5 promote adenovirus internalization but not virus attachment. Cell 73,309-319. 9. Wickham, T. J., Segal, D. M., Roelvink, P. W., et al. (1996). Targeted adenovirus gene transfer to endothelial and smooth muscle cells by using bispecific antibodies. / . Virol. 70, 6831-6838. 10. Goldman, M., Su, Q., and Wilson, J. M. (1996). Gradient of RGD-dependent entry of adenoviral vector in nasal and intrapulmonary epithelia: Implications for gene therapy of cystic fibrosis. Gene Ther. 3, 811-818. 11. Mellman, I. (1992). The importance of being acidic: The role of acidification in intracellular membrane traffic./. Exp. Biol. 172, 39-45 . 12. Leopold, P. L., Ferris, B., Grinberg, I., Worgall, S., Hackett, N. R., and Crystal, R. G. (1998). Fluorescent virions: Dynamic tracking of the pathway of adenoviral gene transfer vectors in living cells. Hum. Gene Ther. 9, 367-378. 13. Greber, U. F., Willetts, M., Webster, P., and Helenius, A. (1993). Stepwise dismantling of adenovirus 2 during entry into cells. Cell 75, 477-486. 14. Bett, A. J., Prevec, L., and Graham, F. L. (1993). Packaging capacity and stability of human adenovirus type 5 vectors./. Virol. 67^ 5911-5921. 15. Graham, F. L., Smiley, J., Russell, W. C , and Nairn, R. (1977). Characteristics of a human cell line transformed by DNA from human adenovirus 5. / . Gen. Viol. 36, 59-72. 16. Graham, F. L., and van der Eb, A. J. (1973). A new technique for the assay of infectivity of human adenovirus 5 DNA. Virology 52, 456-467. 17. Chinnadurai, G., Chinnadurai, S., and Brusca, J. (1979). Physical mapping of a large plaque mutation of adenovirus type 2. / . Virol. 32, 623-628. 18. Carlock, L. R., and Jones, N. C. (1981). Transformation-defective mutant of adenovirus type 5 containing a single altered Ela mRNA species. / . Virol. 40, 657-664. 19. Solnick, D. (1981). An adenovirus mutant defective in splicing RNA from early region lA. Nature 291, 50H-510. 20. Stow, N. D. (1981). Cloning of a DNA fragment from the left-hand terminus of the adenovirus type 2 genome and its use in site-directed mutagenesis./. Viol. 37, 171-180. 21. Jones, N., and Shenk, T. (1979). Isolation of adenovirus type 5 host range deletion mutants defective for transformation of rat embryo cells. Cell 17, 683-689. 22. Kapoor, Q. S., and Chinnadurai, G. (1981). Method for introducing site-specific mutations into adenovirus 2 genome: Construction of a small deletion mutant in VA-RNAj gene. Proc. Natl. Acad. Sci. USA 78, 2184-2188. 4. Adenoviral Vector Construction I: Mammalian Systems 1 0 3 23. Berkner, K. L., and Sharp, P. A. (1983). Generation of adenovirus by transfection of plasmids. Nucleic Acids Res. 11, 6003-6020. 24. Ruben, M., Bacchetti, S., and Graham, F. L. (1983). Covalently closed circles of adenovirus 5 DNA. Nature 301, 172-174. 25. Graham, F. L. (1984). Covalently closed circles of human adenovirus DNA are infectious. £MBO 7.3 ,2917-2922. 26. Ghosh-Choudhury, G., Haj-Ahmad, Y., and Graham, F. L. (1987). Protein IX, a minor component of the human adenovirus capsid, is essential for the packaging of full length genomes. EMBO J. 6, 1733-1739. 27. McGrory, W. J., Bautista, D. S., and Graham, F. L. (1988). A simple technique for the rescue of early region I mutations into infectious human adenovirus type 5. Virology 163, 614-617. 28. Bett, A. J., Haddara, W., Prevec, L., and Graham, F. L. (1994). An efficient and flexible system for construction of adenovirus vectors with insertions or deletions in early regions 1 and 3. Proc. Natl. Acad. Sci. USA 91, 8802-8806. 29. Ng, P., Parks, R. J., Cummings, D. T., Evelegh, C. M., Sankar, U., and Graham F. L. (1999) A high efficiency Cre//oxP based system for construction of adenoviral vectors. Hum. Gene Ther. 10, 2667-2672. 30. Chen, L., Anton, M. ., and Graham, F. L. (1996). Production and characterization of human 293 cell lines expressing the site-specific recombinase Cre. Somatic Cell Mol. Genet. 22, 477-488. 31. Ng, P., Parks, R. J., Cummings, D. T., Evelegh, C. M., and Graham, F. L. (2000) An enhanced system for construction of adenoviral vectors by the tv̂ ô plasmid rescue method. Hum. Gene Ther. 11, 693-699. 32. Van Der Vliet, P. C. (1995). Adenovirus DNA replication. Curr. Top. Microbiol. Immunol. 2, 1-27. 33. Hearing, P., Samulski, R. J., Wishart, W. L., and Shenk, T. (1987). Identification of a repeated sequence element required for efficient encapsidation of the adenovirus type 5 chromosome. / . V/ro/. 61,2555-2558. 34. Haj-Ahmad, Y., and Graham, F. L. (1986). Characterization of an adenovirus type 5 mutant carrying embedded inverted terminal repeats. Virology 153, 22-34. 35. Anton, M., and Graham, F. L. (1995). Site-specific recombination mediated by an adenovirus vector expressing the Cre recombinase protein: A molecular switch for control of gene expression. / . Virol 69, 4600-4606. 36. Bilbao, G., Zhang, H., Contreras, J. L., Zhou, T., Feng, M., Saito, I., Mountz, J. D., and Curiel, D. T. (1999). Construction of a recombinant adenovirus vector encoding Fas ligand with a Cre//oxP inducible system. Transplantation Proc. 31, 792-793. 37. Fujino, M., Li, X. -K., Okuyama, T., Funeshima, N., Tamura, A., Enosawa, S., Kita, Y., Amano, T., Yamada, M., Amemiya, H., and Suzuki, S. (1999). On/off switching Fas-ligand gene expression in liver by Cre//oxP adenovirus vector system. Transplantation Proc. 31, 753-754. 38. Parks, R. J., Chen, L., Anton M., Sankar, U., Rudnicki, M. A., and Graham, F. L. (1996). A helper-dependent adenovirus vector system: Removal of helper virus by Cre-mediated excision of the viral packaging signal. Proc. Natl. Acad. Sci. USA 93, 13,565-13,570. 39. Ng, P., Cummings, D. T, Evelegh, C. M., and Graham, F. L. (2000) The yeast recombinase flp functions effectively in human cells for construction of adenovirus vectors. Biotechniques 29, 524-528. 40. Addison, C. L., Hitt, M., Kunsken, D., and Graham, F. L. (1997). Comparison of the human versus murine cytomegalovirus immediate early gene promoters for transgene expression by adenoviral vectors./. Gen. Virol. 78, 1653-1661. 41 . Maizel, J. V., White, D., and Scharff. M. D. (1968). The polypeptides of adenovirus. I. Evidence of multiple protein components in the virion and a comparison of types 2, 7a, and 12. y/ro/og)/36, 115-125. 1 0 4 Ng and Graham 42. Louis, N., Evelegh, C , and Graham, F. L. (1997). Cloning and sequencing of the cellular/viral junction from the human adenovirus type 5 transformed 293 cell line. Virology 233,423-429. 43. Lochmuller, H., Jani, A., Haurd, J. Prescott, S., Simoneau, M., Massie, B., Karpati, G., and Acsadi, G. (1994). Emergence of early region 1-containing replication-competent adenovirus in stocks of replication-defective adenovirus recombinants (AE1+AE3) during multiple passages in 293 cells. Hum. Gene Ther. 5, 1485-1491. 44. Hehir, K. M., Armentano, D., Cardoza, L. M., Choquette, T. L., Berthelette, P. B., White, G. A., Couture, L. A., Everton, M. B., Keegan, J., Martin, J. M., Pratt, D. A., Smith, M. P., Smith, A. E., and Wadsv^orth, S. C. (1996). Molecular characterization of replication- competent variants of adenovirus vectors and genomic modifications to prevent their occurrence./. Virol. 70, 8459-8467. C H A P T E R Adenoviral Vector Construction II: Bacterial Systems M. Lusky,̂ E. Degryse,^ M. Mehtali,^ and C. Chartier^ Department of Genetic Therapy Transgene Strasbourg Cedex, France I. Introduction The use of adenovirus (Ad) as a vector for in vitro and in vivo gene delivery is expanding rapidly. Besides the use of Ad for gene therapy, it is a highly efficient tool to study in vitro and in vivo gene expression in cell types or tissues not easily transduced by other methods. Other purposes include the use of Ad for the production of high levels of recombinant, potentially therapeutic, proteins and for in vivo vaccination [1-6]. In fact, pioneering the applications of Ad as a gene expression vector were studies which demonstrated that high levels of expression of the SV40 large T antigen in Ad vectors could be achieved. This has become an important source for the biochemical analysis of SV40 T antigen [7, 8]. The ability of Ad vectors to efficiently transduce a variety of cell types and many different target organs in vivo, independent of active cell division, is considered an advantage over other vectors. Furthermore, high titers of virus and high levels of transgene expression can easily be obtained [3]. Extensive genetic and molecular analyses of adenovirus have resulted in a detailed knowledge of the viral life cycle and the function of the majority of viral proteins, further stimulating the use and modifications of Ad vectors [9, 10; Chapters 1 and 2, this volume]. The genome of the most commonly used human adenovirus (group C, serotype 5) consists of a linear 36-kb double- stranded DNA molecule. Transcription of the viral genome occurs on both ^ Corresponding author. ^ Present address: Laboratoire Microbiologie, Pernod-Ricard, Creteil Cedex, France. ^ Present address: Deltagen, Illkirch, France. ^ Present address: Children's Hospital, Boston, Massachusetts. ADENOVIRAL VECTORS FOR GENE THERAPY 1 Q C Copyright 2002, Elsevier Science (USA). All rights reserved. 1 0 6 Lusky ef al. strands and viral gene expression is coordinated through a precisely temporally regulated splicing program of almost all the transcripts. Early transcription units (El, E2, E3, E4) are differentiated from late ones, depending on the expression pattern relative to the onset of viral DNA synthesis [9, 10]. The overlapping location of viral genes on the viral genome limits the molecular manipulations for vector constructions to the El , E2, E3, and E4 regions. The earliest, first-generation Ad vectors have the El region deleted (El°), rendering such vectors replication-deficient. In addition, in most AdEl° vectors the viral E3 region is also deleted, as the E3 functions are not required for the viral life cycle in vitro [9, 10]. In most cases a heterologous expression cassette w îth a transgene is inserted in place of the El region. Such AdEl° and AdEl°E3° vectors can be propagated to high yields in permissive El-complementation cell lines, such as 293 cells [11] or PER.C6 cells [12; Chapter 6, this volume]. The latter prevent the occurrence of replication-competent adenovirus by recombination, allow^ing the production of safe, clinical-grade batches of Ad vectors. Hov^ever, the high level of tissue toxicity and inflammation associated with first generation Ad vectors have stimulated further manipulation of the viral genome, resulting in vectors w îth simultaneous deletions of several regulatory regions, AdEl°E3°E2A° or AdEl°E3°E4° [13-25]. Importantly, AdErE3°E4-modified vectors, carrying the E4 ORF3 or E40RF3 + ORF4 functions w êre able to allow persistent transgene expression in vivo, in selected animal models, in the absence of vector-induced toxicity and inflammation [14, 21, 26]. This renders these types of vectors, with a cloning capacity of approximately 11 kb, attractive for certain, such as liver-selective, gene therapy applications [27]. In this context, various studies have shown that high-capacity or gutless vectors, devoid of all viral genes [28, 29], also combine long-term transgene expression with reduced toxicity [30, 31]. The generation of gutless vectors will be described in Chapter 15. This overview summarizes the recent development of novel technologies, which efficiently permit the rapid construction and generation of single or multiply deleted Ad in Escherichia coli. The construction of Ad in £. coli by various recombination techniques, emphasizing homologous recombination, will be summarized and compared to direct in vitro cloning technologies by ligation. II. Generation of Ad: Traditional Approaches Initially Ad El° vectors were generated in eukaryotic cells, such as in 293 cells using two approaches: (i) the in vitro ligation method [32-35] and (ii) the homologous recombination method in 293 cells [36-39]. The in vitro ligation method uses whole viral DNA, cut at a unique site downstream from the viral El region, and ligated directly to a DNA fragment containing the viral left end 5. Adenoviral Vector Construction II: Bacterial Systems 1 0 7 joined to a transgene; the ligation product is used to transfect 293 cells. This method is hampered by the large size of the Ad genome which limits the number of useful restriction sites available for in
vitro ligation and contamination with wild-type virus. Efficient and improved in vitro ligation techniques for the construction of vectors in bacterial systems have been reported [40-42] and will be described below. Alternatively, cotransfection into the complementation cells of the viral genome and plasmid molecules can generate the Ad by homologous recombi nation in vivo. These methods frequently generate a background of parental virus and repeated screening of many plaques is often required to isolate pure recombinant vectors. However, the development of counterselective methods against the parental wild type vector [43-46] has facilitated the screening for the recombinant virus. The homologous recombination method described by Bett et aL [36] uses two plasmids with overlapping sequences of homology that recombine in vivo. The first plasmid carries the entire Ad genome with a deletion of the DNA packaging signal and the El A region. The second plasmid contains the left inverted terminal repeat (ITR), packaging signal, transgene and overlapping sequence with the first plasmid. Both plasmids are cotransfected into 293 cells and pure Ad is then isolated by plaque purification. The major limitation of this approach remains the low frequency of the recombination event and the potential instability of the large plasmid due to the presence of a head-to-head ITR junction [47]. However, due to a variety of novel and improved techniques, highly efficient methods are now available to generate Ad in mammalian systems, reviewed in Chapter 4. Another method is based on the manipulation of the entire viral genome as an infectious yeast artificial chromosome (YAC) [48]. Targeted modifica tions of the viral genome are introduced by homologous recombination in yeast cells and infectious virions are generated after transfection of the ade novirus genome, excised from the YAC vector. Although clearly pioneering the subsequent studies of viral vector construction in bacterial systems, the YAC system requires the use of an additional host (yeast) and DNA yields are relatively low. III. Generation of Ad: Bacterial Systems Recently, several novel methods based on bacterial systems have been developed for the generation of Ad. Three basic methods have evolved to enable the manipulation of the full-length adenoviral genome as a stable plasmid and facilitate the efficient construction of precisely tailored and infectious Ad in £. coli. These methods are based on: (i) homologous recombination, (ii) direct ligation, and (iii) cosmid technology. All three methods offer major advantages over traditional approaches: (i) Manipulation of the viral genome at any point 10 8 Lusky ef a/. is possible, (ii) Recombinant viral DNA is purified from individual bacterial clones and therefore generates homogenous virus preparations, obviating the need for tedious plaque screening and purification, (iii) Importantly, and in contrast to the traditional in vivo approaches, these methods entirely separate viral vector construction from virus production. The first step is performed in bacteria and the second step takes place in the mammalian complementation cell line. Therefore, each step can be carefully controlled and optimized. Trouble shooting is facilitated; for example, failure of producing a virus cannot be associated with the inability to generate the desired genome. The recombination and direct ligation methods are described below. IV. Homologous Recombination in E. coli The use of classical molecular biology techniques for the manipulation of the Ad genome is limited by its large size. Homologous recombination presents an alternative way to engineer DNA. In yeast, homologous recombination is particularly flexible and recombination between linear DNA fragments flanked by short-homology arms and endogenous recipients such as the yeast genome or YACs or a gapped plasmid [49, 50] is routinely used. The concept was also applied in £. coli by the cloning of short DNA fragments [51] and of PCR products [52] into gapped plasmids as targets. We mentioned above the work of Ketner et al. [48], who reported in 1994 the cloning and further manipulation of the Ad genome as an infectious YAC clone, taking advantage of the very efficient Saccharomyces cerevisiae recombination machinery. The availability of specific mutant bacterial strains allowed us and others to transfer this technology into E. coli and to bring the adenovirus genome manipulation back to the level of standard molecular biology. Recombination is an essential process involved in the repair of DNA lesions, such as double-stranded breaks (DSBs) and the restart of replication forks that failed to progress to completion. The E. coli recombination machin ery includes at least 25 different proteins among which the RecBCD enzyme and the RecA protein are major components of the initiation and pairing steps [53]. In wild-type cells, the RecBCD enzyme binds to the end of a dsDNA substrate and initiates unwinding. RecBCD degrades the 3^-terminated strand during unwinding until it reaches a chi site. The chi sequence is a recombination hotspot which modifies the enzymatic activities of RecBCD enzyme [54]. The nuclease activity of RecBCD is attenuated and RecA is loaded on the 3̂ end of ssDNA, allowing the essential steps of pairing and strand exchange. The RecBCD nuclease activity is also responsible for the degradation of foreign DNA that does not contain any chi site. This last property explains why, in contrast to yeast, most bacteria do not recombine transformed DNA readily and are not widely used for plasmid manipulation by homologous recombi nation [55], However, some bacterial strains where the RecBCD enzyme is 5. Adenoviral Vector Construction II: Bacterial Systems 1 0 9 inactivated have been show^n to be recombination-proficient. They harbor an additional suppressor (sbcA or sbcB) mutation that activates an ahernative recombination pathway [56]. The genotype (endA, sbcBC, recBC, galK, met, thi-1, bioT, hsdR, stf) of the BJ5183 bacterial strain [57] used in most of the studies discussed belov^ is RecBC sbcBC and contains in addition an activated RecF pathway. This pathway has been shown to direct nonconservative recom bination which is defined as a homologous recombination event generating one duplex molecule out of two duplex DNA molecules [58]. Subsequently, taking advantage of such recombination-proficient £. coli strains [59], several systems have been developed to manipulate the Ad genome by homologous recombination. The different systems applying homologous recombination for Ad con struction can be put into two groups depending whether the targeted vector is transfected as a linear DNA fragment or as a circle. In the first case, the recombinant plasmid is rescued by recombination between two linear DNA fragments (the donor fragment and the linearized vector). In this approach the only selection necessary is the recircularization of the plasmid. The first application of this technology for Ad construction was described by Chartier et al. [60], demonstrating (i) stable cloning of the entire viral genome into a bacterial plasmid and (ii) the use of such infectious bacterial plasmids to further introduce alterations into the viral genome. Subsequently, the basic protocol was extended and modified by others [61]. A second approach where the plasmid to be modified is transformed as a circle allows more flexibility but requires the development of complex selection systems. Such approaches have been reported and will be discussed [62, 63]. V. Homologous Recombination \N\\\\ Linear Ad Vector Genome Plasmids The work by Chartier et al. [60] showed that stable maintenance of plas mids containing the entire Ad viral genome is achieved through the separation of the viral ITRs by the bacterial plasmid backbone, confirming the observa tions made earlier by Hanahan and Gluzman [64]. This was accomplished by the insertion of the left and right end of the Ad genome in their normal orien tation into a colEI-derived bacterial plasmid ppolyll [65]., using conventional cloning techniques. Such a plasmid linearized between the two Ad ends served as a vector to eventually incorporate the entire Ad viral genome. Cotransfection of the linearized vector DNA and linear Ad5 genomic DNA into the £. coli strain BJ5183 generated a stable circular plasmid containing the full-length Ad5 genome through homologous recombination within the end fragments of Ad5 (Fig. 1). Engineering unique restriction sites such as Pad (noncutting within the Ad5 viral genome) immediately adjacent to the viral ITRs enabled 110 Lusky ef a/. the precise release of a fully infectious viral genome (Fig. 1). In contrast, the closed circular plasmid was unable to generate any infectious virus, confirming that at least one viral ITR extremity has to be in a free configuration to efficiently initiate the replication machinery of Ad [33, 64]. Various single-step replacement strategies exploiting the E. coli homolo gous recombination machinery w êre subsequently designed in our laboratory to selectively modify various genetic regions in Ad5. The principle for these manipulations is simple: the viral region to be modified is first subcloned into a bacterial shuttle plasmid containing sequences of the Ad genome to be targeted. The desired alterations, such as deletions, point mutations, or insertions of transgene-containing expression cassettes, are performed in this shuttle plasmid using standard molecular biology techniques. Subsequently, a restriction fragment containing the modified DNA segment and leaving suffi cient sequences of Ad homology on either side of the modification is prepared. Homologous recombination in E.co/i with linear Ad vectors Cotransform recBC sbcBC BJ5183: -^ homologous recombination transform recA E.coli (DH5a) -^ clonal viral genome construction iiL Production of virus particles in mammalian cells >k Pad digestion, release of the viral genome ^ transfection of permissive mammalian cells ^i homogenous population of Ad5 virus Figure 1 Cloning of infectious full-lengtfi Ad5 genome in E. coli by homologous recombina tion [60]. The vector plasmid pTG3601 contains 935 and 853 bp from the left and right ends of Ad5. Cotransfection of Bglll - linearized pTG3601 with linear Ad5 DNA results in recombinants containing the full-length Ad5 genome. 5. Adenoviral Vector Construction II: Bacterial Systems 111 purified, and cotransfected into BJ5183 along with the plasmid DNA con taining the full length Ad genome to be modified. The Ad genome plasmid is linearized in the targeted region. After recombination occurred between the donor fragment and the linearized Ad plasmid, the expected recombinant is simply rescued by plating bacteria in the presence of the appropriate antibiotic. The circularization of the plasmid is the only selection pressure applied. A detailed example for the targeting of the El region is illustrated in Fig. 2. Unique restriction sites available for the targeting of alterations into various regions of the Ad genome (Fig. 3) are Clal (El), Sg/"!, BamHl (E2), Sr/"!, and Spel (E3). In cases where the double strand break is located outside the targeted region, the percentage of rescue of the expected modification decreases with increasing distance of the linearization site with respect to the targeted region [60, 66]. In order to improve this and to very efficiently modify the E4 Ad5: targeting of the E1 region 1. cloning transgene into transfer plasmid -* homologous recombination (BJ5183) with Clal - linearized pAd5 2. clonal isolation of recombinants (DH5a) 3. virus production : 293, PER.C6 cells Pad Pad recombination Pad MCS Ad5 H+1+- Ad5 Transfer plasmids Pad MCS Ad5 II CMVpH+H+-[pAlf~ Ad5 Figure 2 Targeting of the Ad5 El region by homologous recombination with linear vectors. Representative examples of transfer plasmids are schematically shown. The transfer plasmid is linearized or the expression cassette flanked with sequences of homology is excised. They are then cotransfected with the C/d-l inearized Ad5 genome vector plasmid. Upon homologous recombination in E. coli BJ5183 a recombinant viral genome plasmid is generated with the El region deleted and replaced by the foreign expression cassette. 112 Lusky ef o/. L3 L4 L5 L2- E1A E1B L1. E3 L-ITR pIX MLP R-ITR 3' D= Ad 5 5' n= =n 3' IVa2 E2A (DBP) E4 E2B (pTP, POL) • Release of the viral aenome Pad Clal Swal • linearization • • homoloaous recombination: • E 1 : Clal •E2: Sgfl •E3: l-Scel • L5 / E4: BstBI Figure 3 Schematic representation of the Ad5 genome and map of the basic Ad5 genome plasmid pTGl 5152 used as a standard vector in our laboratory for the construction of Ad vectors. The extent of the deletions in the El and the E3 regions is indicated.. With the exception of these deletions the entire sequence of Ad5 is contained in this plasmid. The viral genome can be released from the plasmid backbone by Swa\ or Pod digestions, located immediately adjacent to the left and right inverted terminal repeats (L-ITR, R-ITR). Unique restriction sites to target the E l , E2, E3 L5/E4 region are indicated. The Clo\ site
is resistant to Dam methylation. To facilitate efficient homologous recombination in the E3 region, the recognition site for the homing endonuclease l-Scel [68] v/as introduced at the deletion point in E3. The 6s/BI site, located exactly between the L5 and E4 transcription units enables the generation of recombinants in both regions. region or introduce mutations into the fiber gene of Ad, we have introduced a Bst^l site just between the poly(A) sites of the fiber gene (L5) and the E4 region [21, 67]. In addition, a recognition site for the homing endonuclease I- Scel [68] was introduced directly at the deletion point in E3 (Fig. 3) to improve the efficiency of targeting the E3 region. The single Clal restriction site in the El region (GATCGATC) is Dam methylation-sensitive. The dam gene of Escherichia coli encodes a DNA methyltransferase that methylates adenine in -GATC-sequences in double-stranded DNA [69]. Thus, the replacement of this Dam-methylation-sensitive Clal site with a methylation-resistant Clal site (T/AATCGATT/A) represented another improvement of the procedure, alleviating the need for a cumbersome step of growing the Ad plasmids in 5. Adenoviral Vector Construction II: Bacterial Systems 113 a dani~ bacterial strain. Multiple transfer plasmids are now available in our laboratory to routinely target the viral El , E2A, E3, L5, and E4 regions using the strategic sites indicated above. Thus, wt have successfully introduced numerous alterations, deletions, as v^ell as insertions of transgenes into the viral El , E2A, E3, E4, and fiber regions [20, 21, 27, 60, 67-^ R. Rooke, unpubfished observations]. The application of homologous recombination to modify the E4 region is illustrated in Fig. 4. Due to the availability of nev^ endonucleases, such as intron-encoded homing endonucleases [70], the generation of further multiply altered Ad viral genomes can be envisioned, by insertion of such endonucleases at further strategic sites in the Ad vector genome. Our experimental evidence suggests that a minimum of ^^50 bp of complete homology on either side is required for significant recombina tion efficiency, consistent v^ith Watt et al. [71]. There is an exponential increase in the frequency of recombination when the length of homol ogous DNA is increased to about 100 bp. Beyond this value, there is an apparent linear increase with longer DNA segments of homology [71; Ad5: targeting of the E4 region E4 modifications AdSErES" E4 wt Ad5ErE3° E4 - modified E1° n= l:0RF4: I lORFS: I l:ORF6;7l \|QRF3;,€Ji Figure 4 Targeting of the E4 region. E4 modifications are introduced in the appropriate transfer plasmid by standard cloning procedures. The vector genome plasmid (such as pTG15152) is linearized at the BsfBI site. Co-transfection into E. coli BJ5183 and homologous recombination introduces the E4 modifications indicated in place of E4. 1 1 4 Lusky ef al. M. Lusky and D. Dreyer, unpublished observations]. If the length of homol ogous DNA segments is too large (1 kb and beyond), double recombination events may result in the parental vector backbone, thus diminishing the yield of true recombinants (M. Lusky and D. Dreyer, unpublished observa tions). The ratios of vector DNA (Ad5 plasmid, linearized) to transfer DNA and the absolute amounts also influence the efficiency of transformation as w êll as the frequency of homologous recombination in BJ1583. Our own experimental evidence suggests that a molar ratio of vector to transfer DNA of about 1:10 appears optimal. In addition, ŵ e routinely use about 3 to 10 ng of vector DNA to achieve optimal transformation and recombination efficiencies. This is consistent w îth the notion that the size of the Ad5 genome constitutes a limiting factor and increasing amounts of Ad vector DNA inhibit the transformation intoE. CO//BJ5183 [60]. One draw^back of the method described above is that tvv̂ o transformation steps in E. colt are required prior to the use of high quality, infectious plasmid DNA for transfection into the appropriate mammalian complementation cells. This is due to the fact that the yield and quality of large plasmids, such as Ad plasmids, in the E. coli strain BJ5183 are low .̂ The nonconservative recombination mechanism exploited for the manipulation of the Ad plasmids is responsible for this feature. During plasmid amplification in a RecBC sbcBC bacterial strain, nonconservative recombination occurs betw^een a replicating circle and another circle. The end of the generated rolling circle undergoes further recombination v^ith circles. This results in the accumulation of linear plasmid multimers and decreases the yield in circular plasmid DNA [58, 72]. Thus, for extensive characterization of a new^ recombinant clone, candidate plasmid DNAs derived upon homologous recombination in BJ5183 are trans formed into the recA~ E. coli strain DH5a [73] in our laboratory. In this strain large plasmids are stable and the yield and quality of plasmid DNA is high. Furthermore, due to the availability of state-of-the-art plasmid preparation kits, this second step becomes routine w^ork, taking one extra day's w^orth of work. Once the new recombinant viral genome is fully characterized, the viral DNA is transfected upon release with Pad into the appropriate human complementation cells. Since production of recombinant virus is clonal there is no need for plaque purification. With simple El-deleted or El replacement vectors complete CPE is usually obtained within 5 to 7 days after transfection into either 293 [11] or PER.C6 cells [12]. Since the screening and molecular analysis have already been carried out in the prokaryotic host, the eukaryotic cells merely serve to amplify the clonal recombinant. This leads to an enormous gain in time. 5. Adenoviral Vector Construction II: Bacterial Systems 1 1 5 Taken together, the basic procedure developed by Chartier et al. [60] allows the rapid cloning and manipulation of full-length infectious Ad genomes in bacterial plasmids. The method combines the powerful genetic engineer ing techniques available in £. colt and the ability of this microorganism to recombine homologous sequences at high frequency. The advantages of this technology are multiple and evident: (i) all cloning and, more importantly, all recombination steps are carried out in E. coli, thus a high degree of control is possible at each step during the procedure, (ii) the frequency of bacterial colonies containing the plasmids with the modified Ad genome is very high (up to 90% efficiency), (iii) any genetic region of the viral genome can be specifically modified or deleted, including the introduction of point mutations, if appropriate restriction sites are available, (iv) plasmids containing full-length and modified Ad genomes can be introduced into appro priate bacterial strains for production of large amounts of infectious viral DNA, and (v) transfection of released recombinant Ad viral DNA into the appropriate complementation cell lines generates homogenous pure virus particles without the time-consuming need for plaque screening and purifi cation. The method also has been successfully applied to the assembly and further modification of a variety of animal adenoviruses into infectious plas mids [74-78]. A direct modification of the technique developed by Chartier et al. [60] uses oligonucleotide site-directed cleavage of DNA and homologous recom bination for the production of recombinant Ad vectors [61]. This procedure, schematically illustrated in Fig. 5, allows the introduction of modifications into the viral genome at virtually any predetermined sequence. Toward this goal, the authors describe the use of torsionally stressed supercoiled Ad plasmid DNA which will allow the stable strand displacement of a targeted sequence by hybridization with a complementary oligonucleotide. The hybridization induces the formation of a stable D-loop structure. The single-stranded DNA displaced in such D-loops is specifically cleaved by a single strand-specific nuclease, such as SI nuclease (Fig. 5). This results in the linearization of the Ad plasmid at the targeted site and therefore provides a perfect template for homologous recombination. The recombination reaction is then performed as described above. The utility of this method was demonstrated by using an oligonucleotide complementary to the E3 transcription unit of Ad5 to incor porate the simian virus 40 (SV40) origin of replication into the E3 region by homologous recombination. The resulting cloning efficiency was about 60%, probably reflecting the relative inefficiency of the vector linearization. The clear advantage of this procedure is the complete independence of unique restriction sites in the vector backbone, thus enabling modification virtually anywhere in the vector genome. 116 Lusky ef a / . D - loop technique \ ^ ^ 0.2MNaC D- loop/S1 linear DNA Supercoiled DNA • primer Homologous recombination homologous Recombinant recombination plasmid BJ5183 Figure 5 Schematic representation of the D-loop technique [61 ]. Linearization of the supercoiled target vector backbone is induced by oligonucleotide-mediated D-loop formation followed by SI nuclease digestion at the target site. D-loop / SI -linearized DNA serves as a vector for subsequent homologous recombination events. VI . Homologous Recombination vsfith Circular Ad Vector Genome Plasmids Another way to circumvent the need for single cutting sites is obviously to use a circular instead of a linear template for the homologous recombination. However, in that case the selection process will necessarily involve a step of counterselection against the parental Ad plasmid, rendering the selection process more complex than the simple recircularization to be selected for in the systems described above. Two different approaches where the viral backbone is used as a supercoiled circular plasmid have been described [62, 63] The method by Crouzet et al. [62] for the clonal production of £. coli- derived Ad genomes (EDRAG) exploits the observation that incP-derived plasmids, but not colEl-derived plasmids, such as pBR322- or pUC-derivatives, replicate and are stably maintained in E. colt polA mutant hosts [79]. As a starting point, an incP plasmid [80, 81] with a tetracycline resistant marker, carrying the extremities of Ad5 was used as a vector to incorporate the entire Ad5 genome to generate a full-length Ad5 genomic infectious plasmid. Sub sequently, any modification can be introduced in this incP-derived full-length viral genome plasmid by homologous recombination events between the incP- Ad5 replicon and a colEl shuttle plasmid. The colEl shuttle plasmid, carrying 5. Adenoviral Vector Construction II: Bacterial Systems 117 a kanamycin resistance marker and a conditional suicide gene (the sacB gene of Bacillus subtilis [82]) is engineered by standard cloning techniques containing the specified modification flanked by appropriate homology sequences to the Ad genome. The colEl-derived shuttle plasmid is transformed into the polA host carrying the incP/Ad5 plasmid. Cointegrates are selected by growing the polA host in the presence of both antibiotics. Resolution, leading to loss of the colEl replicon from the recombinant Ad plasmid, of this cointegrate is subse quently selected by growth of the cells in sucrose activating the B. subtilis sacB gene as suicide gene, leading to the concomitant loss of the sacB conditional suicide marker. Consecutive rounds of this two step recombination procedure allow the introduction of multiple independent modifications within the virus genome, with no requirement for an intermediate virus. The potential of this procedure was demonstrated by the recovery of various E1°E3°E4° vectors. The second system [63] using circular viral genome plasmids as target for the homologous recombination is designed only to replace the viral El region with heterologous transgene expression cassettes (see Fig. 6). The vector backbones (pAdEasy) lack Ad nt 1-3533 and are further deleted of the E4 and/or E3 regions. The various shuttle vectors designed (pShuttle) contain a polylinker or a prepared expression cassette surrounded by adenoviral sequences ("arms"), allowing homologous recombination with the pAdEasy Homologous recombination inE. co//with circular Ad vectors transgene . right arm ^ * t \ "̂ '̂ homologous recombination (BJ5183) Pmel "̂ " " ' " shuttle vector left pAdEasy ^L-LTR / / arm Ad backbc^ne Pacr Pad Pad left arm Clonal isolation in E.coli recA Pad - release • virus production (e.g. 293 cells) L-LTR transgene ĵght arm Ad backbone left arm R-LTR Figure 6 Homologous recombination in E. co// BJ5183 with circular target vectors [63]. The gene of interest is cloned into a shuttle vector. The resulting plasmid is linearized by Pmel and cotransfected into E. coli BJ5183 with the circular Ad backbone plasmid pAdEasy. Selection of recombinant Ad plasmids in the presence of Kanamycin allows the counterselection of the pAdEasy parental backbone. 1 1 8 Lusky ef al. system. The left arm contains Ad nt 34,931 to 35,935 (3^ end of Ad) joined to the plasmid backbone and joined to Ad nt 1-480 (5^ ITR, ori and packaging sequence). The right arm consists of Ad nt 3534-5790 joined to the expression cassette. Pad sites
are engineered immediately adjacent to the ITRs to facilitate the generation of infectious recombinant Ad genomes. Importantly, the shuttle plasmid backbone contains a kanamycin resistance gene for the selection of recombinants in £. colt. Upon linearization of the shuttle plasmid at a single restriction site between the "arms," the DNA is cotransfected with the supercoiled pAdEasy vector into E. colt BJ5183 (Fig. 6). Recombination occurs between the homologous arms, the pAdEasy plasmid backbone that harbors an ampicillir resistance gene is replaced by the pShuttle plasmid backbone that presents the kanonyan resistance gene and the expression cassette is introduced in the El deletion. Recombinant viral genomes are selected in the presence of kanamycin allowing efficient counterselection of the parental, circular pAdEasy vector. Contamination of the reaction with some uncut shuttle DNA can easily be distinguished from the recombinant DNA due to the size differences. Similar to the approach described by Chartier et al, [60], the new recombinant viral plasmid then needs to be transferred into a recA~ strain, such as DHIOB [73] for greater yields of DNA. Using this approach, a cloning efficiency of about 70% was reported. Upon release of the recombinant viral genome from the bacterial plasmid backbone, virus is produced in the appropriate human complementation cell lines as described above. In some of the vectors described by He et al. [63] the green fluorescent protein reporter gene is cointroduced with the expression cassette into the recombinant virus, facilitating the tracing at each step of the viral production. The system as outlined [63] is specifically designed to efficiently and easily introduce heterologous expression cassettes in place of the viral El region. Since this approach relies on the replacement of the plasmid backbone in between the two Ad ITRs for the selection of the recombinants, only viral regions located at the extremities of the genome (El region, E4 region) can easily be targeted. The targeting of any other region requires additional procedures to promote the loss of the prokaryotic selection marker from the viral genome as illustrated in the system described by Crouzet etal. [62]. VII. Ad Vector Construction by Tronsposon-Medioted Recombination Recently a transposon-mediated recombination system has been devel oped which involves the generation of Ad by Tn7-mediated, site-specific transposition in E. colt [83]. The development of this system requires two plasmid components: (i) a low-copy-number, full-length, circular adenoviral plasmid (admid [83]) with a P-galactosidase marker replacing the El region 5. Adenoviral Vector Construction 11: Bacterial Systems 1 1 9 and containing the Tn7 attachment site (lacZatfTn?) and (ii) an admid trans fer vector with a mini-Tn7 containing an expression cassette with the gene of interest. Tn7-mediated transposition of the expression cassette into the lacZattTn? site disrupts the lacZ coding region, resuhing in a P-gal- phe- notype of the newly generated admid. The authors describe a transposition frequency averaging 25%. After clonal isolation of the new recombinants, the Ad genome is released from the admid backbone and transduced into 293 cells for amplification and production. Described to carry out replacements in the El region, the admid system could also be adapted to alter other or additional viral regions. VIII. Ad Vector Construction by in Vitro Ligation The notion of separation of the molecular construction of the rAd vector from the virus production is also guiding the direct cloning procedures described below. Efficient and improved in vitro ligation techniques have recently been described [40-42]. Due to the paucity of unique restriction sites within the adenoviral genome, direct ligation of a transgene expression cassette into the viral vector backbone is facilitated by the use of intron- encoded endonucleases. The technique basically involves two cloning steps in £. coli followed by the transduction of the linear recombinant viral DNA into the permissive mammalian cell-complementation system (see Fig. 7). Intron-encoded endonucleases are enzymes that are encoded within group I introns [70]. An unusually long homing sequence ranging from 15 to 39 bp renders these endonucleases rare cutting and ideal to use for cloning sites in adenoviral backbones. Two groups have described the use of intron-encoded endonucleases, such as l-Ceul [84] and PI-Sc^I [85] for the direct substitution of the viral El region by a reporter gene. Mizuguchi and Kay [40] have generated four basic viral backbones, pAdHMl-4 in a bacterial plasmid; all viral backbones are El° with or without an E3 deletion and the viral DNA is flanked by Pad and Clal sites (pAdHMl, 2) or by Pad sites at both ends (pAdHM3, 4). In each backbone three unique cleavage sites are introduced in to the El-deleted region: l-Ceul, Swal, Vl-Scel. A basic shuttle plasmid (pHM3) contains a multiple cloning site flanked by l-Ceul and Vl-Scel sites, respectively. The expression cassette of interest can be inserted in any site of the multiple cloning site. After cloning, the cassette-containing insert is excised from the shuttle plasmid using the intron-encoded endonucleases. The restricted fragment is then directly cloned into the vector backbone, which is linearized by l-Ceul and Fl-Scel. This scheme permits a directed cloning by ligation of the expression cassette into the viral backbone and lowers the background of parental viral vector (null vector) due to different ends created 120 Lusky ef o/. Ad5 vector construction: in vitro ligation Pad Pad Pl-Scel in vitro ligation ; transform E.coii recA ; Pad release, virus produdion in 293 cells transgene —m — L-ITR E3^ R-ITR Figure 7 Construction of Ad by ligation in vitro [40-42]. A detailed description is found in the text. by the two different intron-encoded endonucleases. Restriction of the ligation product with Swal further diminishes the background of the parental null vector in the cases where the ligation products do not contain a Swal site. The restricted ligation product is then directly transformed into E. coli DH5a. With this method the authors report a 90% cloning efficiency for recombinant viral genomes with the correct restriction pattern. The recombinant viral genome can subsequently be released from the plasmid backbone by Pad or Fad plus Clal digestion followed by the generation of a homogenous Ad population upon transfection into 293 cells. The simplicity of the system is clearly emphasized by the careful choice of unique and strategic restriction sites. The authors have recently added new vector backbones including ElE3E4-deleted vectors as well as modified shuttle vectors to increase the variety of vector construction [41]. The p Advantage system described by Souza and Armentano [42] follows the same logic. Their basic viral El° backbone (pAdyantage) is based on Ad 2 (serotype 2) and contains an l-Ceul site in the region of the El deletion, the ITRs of the viral genome are flanked by Sna^l sites (Sna^l does not 5. Adenoviral Vector Construction II: Bacterial Systems 1 2 1 cleave within the Ad2 viral genome). A shuttle plasmid containing a poly linker flanked by l-Ceul sites is used to introduce the gene expression cassette of interest. The ligation of the l-Ceul restricted expression cassette into the 1-Ceul linearized vector is reported v^ith an ~ 5 0 % cloning efficiency. The authors describing the in vitro ligation approach (Fig. 7) indicate that under optimal conditions the timing from start (cloning of the gene of interest into the shuttle vector) to finish (isolation of plaques, or passage 1 lysate) generally does not exceed 3 weeks. Taken together, the improved, direct in vitro ligation method is simple, straightforward, and appears very efficient, although restricted to the replacement of the El region [42] and to the El and E4 region [40, 41] in its current state. Although it will not allow more precise modification of the Ad genome, the system is clearly amenable to further developments and improvements. For example, additional introduction of noncutting endonucleases, such as intron-encoded or other nonconventional nucleases, into the viral E3 and E4 regions will enable the generation of Ad vectors containing multiple inserts. IX. Conclusion In summary, the concept of the separation of the steps of Ad genome construction in prokaryotic cells and virus amplification in eukaryotic cells has led to the exploitation of powerful methods, including homologous recom bination, direct ligation and transposition technologies to clonally derive Ad constructs in E. coli. The advent of these technologies has enabled the con struction and generation of numerous modified Ad vectors carrying many different transgenes. In addition, our understanding of the in vivo biology of these vectors has been advanced tremendously through the ease of introducing viral genome modifications and studying their effects. The methods in use are simple, highly efficient, and can generate Ad within a very short time period (Fig. 8). The choice of which method to use will depend on the vector region to be targeted and the type of modification to be introduced. Homologous recom bination methods using linear vectors can target virtually any region in the Ad5 backbone and allow precise modification of the Ad genome sequence such as small deletions and point mutations. Homologous recombination methods with circular vectors and the in vitro ligation technique, although very efficient, are currently set up to target the ends of the viral genome (e.g.. El , E4 region) It is clear that these techniques will be further optimized, as research develops, and could be adapted for the generation of other viral vector systems. Many applications of these methods can be envisioned, including the use of custom tailored Ad for the in vivo production of recombinant therapeutic proteins or the generation of custom-made libraries in Ad vectors to enable the high-throughput screening and applied functional analysis of many genes in the context of functional genomics projects. P CL, t o ^ o -a ^4 ^ ' ^ J ^ _y sXtjp e oj 2 sXBp 6 0 U ^ o 9 O C .5 "̂ g-5 w ^ i • .S g - • p - 2 I § § . « O CO ^^ a o CT- _Q U C D. E U 0 >s _yV. ^ V̂ J O (D s/Cep e oj 2 SXBP 5 OJ £ sA^p 6 01Z, _g c E ^ .S o a> > u "Oc o D _Q C si I ̂ _̂ L J L J 0) O SABP £ OJ 2 sAcp 6 oj /. sXipijoTFl 00 O 0) - a iZ E 122 5. Adenoviral Vector Construction II: Bacterial Systems 1 2 3 Acknowledgments We thank various members of our laboratory, particularly Dr. A. Winter for helpful suggestions. The technical assistance of D. Dreyer and M. Gantzer is gratefully acknowledged. References 1. Berkner, K. L. (1988). Development of adenovirus vectors for the expression of heterologous genes. BioTechniques 6, 616-629. 2. Chengalvala, M. V., Lubeck, M. D., Selling, B. J., Natuk, R. J., Hsu, K. H., Mason, B. B., Chanda, P. K., Bhat, R. A., Bhat, B. M., Mizutani, S., Davis, A. R., and Hung, P. P. (1991). Adenovirus vectors for gene expression. Curr. Opin. BiotechnoL 5, 718-722. [Review] 3. Graham, F. L., and Prevec, L. (1992). Adenovirus based expression vectors and recombinant vaccines. In "Vaccines: New Approaches to Immunological Problems" (R. W. Ellis, Ed.), pp. 363-390. Butterworth-Heinemann, Storeham, MA. 4. Morgan, R. A., and Anderson, W. F. (1993). Human gene therapy. Annu. Rev. Biochem. 62, 191-217. [Review] 5. Trapnell, B. C. (1993). Adenoviral vectors for gene transfer. Adv. Drug Deliv. Rev. 12, 185-199. 6. Trapnell, B. C., and Gorziglia, M. (1994). Gene therapy using adenoviral vectors. Curr. Opin. BiotechnoL 5, 617-625. [Review] 7. Thummel, C., Tjian, R., and Grodzicker, T. (1981). Expression of SV40 T antigen under control of adenovirus promoters. Cell 23, 825-836. 8. Thummel, C., Tjian, R., and Grodzicker, T. (1982). Construction of adenovirus expression vectors by site-directed in vivo recombination./. Mol. Appl. Genet. 1, 435-446. 9. Horwitz, M. S. (1990). Adenoviridae and their replication. In "Virology" (B. N. Fields, D. M. Knipe, et al., Eds.), 2nd ed., pp. 1679-1721. Raven Press, New York. 10. Shenk, T. (1996). Adenoviridae: The viruses and their replication. In "Virology" (B. N. Fields, D. M. Knipe, P. M. Howley, et al., Eds.), pp. 2111-2148. Raven Press, Philadelphia, PA. 11. Graham, F. L., Smiley, J., Russell, W. C., and Nairn, R. (1977). Characteristics of a human cell line transformed by DNA from human adenovirus type 5. / . Gen. Virol. 36, 59-74. 12. Fallaux, F. J., Bout, A., van der Velde, I., van den Wollenberg, D. J., Hehir, K. M., Keegan, J., Auger, C
(1990). Differential plasmid rescue from transgenic mouse DNAs into Escherichia coli methylation-restriction mutants. Proc. Natl. Acad. Sci. USA 87, 4645-4659. 74. Zakhartchuk, A. N., Reddy, P. S., Baxi, M., Baca-Estrada, M. E., Mehtali, M., Lome A. Babiuk, L. A., and Tikoo, S. K. (1998). Construction and characterization of E3-deleted 5. Adenoviral Vector Construction II: Bacterial Systems 1 2 7 Bovine adenovirus Type 3 expressing full-length and truncated form of bovine Herpesvirus type 1 glycoprotein gD. Virology 250, 220-229. 75. Reddy, P. S., Idamakanti, N., Hyun, B. H., Tikoo, S. K., and Babiuk, L. A. (1999). Develop ment of porcine adenovirus-3 as an expression vector. / . Gen. Virol. 80, 563-570. 76. van Olphen, A. L., and Mittal, S. K. (1999). Generation of infectious genome of bovine adenovirus type 3 by homologous recombination in bacteria./. Virol. Methods 77, 125-129. 77. Michou, A. I., Lehrmann, H., Saltik, M., and Gotten, M. (1999). Mutational analysis of the avian adenovirus GELO, w^hich provides a basis for gene delivery vectors. / . Virol. 73, 1399-1410. 78. Fujita-Kusano, A., Naito, Y., Saito, I., and Kobayashi, I. (2000). Mutation correction by homologous recombination v^ith an adenovirus vector. Methods Mol. Biol. 133, 101-109, 79. Stachel, S., An, G., Flores, G., and Nester, E. (1985). A Tn3 lacZ transposon for the random generation of beta-galactosidase gene fusions: Application to the analysis of gene expression in Agrobacterium EMBO } . 4, 891-898. 80. Ditta, G., Schmidhauser, T., Yakobson, E., Lu, P., Liang, X. W., Finlay, D. R., Guiney, D., and Helinski, D. R. (1985). Plasmids related to the broad host range vector, pRK290, useful for gene cloning and for monitoring gene expression. Plasmid 13, 149-153. 81. Thomas, G., and Smith, G. (1987). Incompatibility group P plasmids: genetics, evolution, and use in genetic manipulation. Annu. Rev. Microbiol. 41, 77-101. 82. Bloomfield, I. G., Vaughn, V., Rest, R. F., and Eisenstein, B. I. (1991). Allelic exchange in Escherichia coli using the Bacillus subtilis sacB gene and a temperature-sensitive pSGlOl replicon. Mol. Microbiol. 5, 1447-1457. 83. Richards, G. A., Brown, Gh. E., Gogswell, J. P., and Weiner, M. P. (2000). The admid system: Generation of recombinant adenoviruses by Tn7-mediated transposition in E. coli. Biotechniques 29, 146-154. 84. Marshall, P., and Lemieux, G. (1992). The I-Geul endonuclease recognizes a sequence of 19 base pairs and preferentially cleaves the coding strand of the Ghlamydomonas moev^usii chloroplast large subunit rRNA gene. Nucleic Acids Res. 20, 6401-6407. 85. Gimble, F, S., and Thorner J. (1992). Homing of a DNA endonuclease gene by meiotic gene conversion in Saccharomyces cerevisiae. Nature 357, 301-306. C H A P T E R Propagation of Adenoviral Vectors: Use of PER.C6 Cells W. W. Nichols/ R. Lardenoije/ B. J. Ledwith/ K. Brouwer,^ S. Manam/ R. Vogels,^ D. Kaslow,"^ D. Zuidgeest^ A. J. Belt/ L Chen/ M. van der Kaaden,^ S. M. Gal loway/ R. B. Hi l l / S. V. Machotka/ C. A. Anderson,"^ J. Lev^is,'' D. MartineZ/'' J. Lebron,'' C. Russo/ D. Valerio,^ and A. Bout^ *Merck Research Laboratories Merck & Company, Inc. West Point/ Pennsylvania +Crucell NV Leiden, The Netherlands I. Introduction A. Scope of the Chapter The goal of gene therapy is the introduction of genes into human somatic cells for therapeutic purposes. The success of gene therapy is therefore dependent on the efficiency by which a therapeutic gene can be transferred to the patient's target tissues. In many cases, viruses are exploited for gene transfer purposes and in particular gene transfer vectors derived from adenoviruses (adenoviral vectors) are often used to achieve this (for reviev^ see [1]). The reason for this is that adenoviral vectors: • efficiently transfer genes to many different cell types; • can be propagated on well-defined production systems to high yields; and • are very stable, which makes purification and long-term storage possi ble, thereby making pharmaceutical production feasible. ADENOVIRAL VECTORS FOR GENE THERAPY \| 2 9 Copyright 2002, Elsevier Science (USA). All rights reserved. 1 3 0 Nichols ef aL This contribution will focus on the production systems for clinical lots of adenoviral vectors. Particular attention v îll be paid to the generation and use of complementation cell lines that carry the El genes. Particular emphasis will be on the PER.C6 cell line, which was developed to prevent generation of replication-competent adenovirus (RCA) during propagation of El-deleted adenoviral vectors. In addition, safety issues with respect to the use of the cell line for making clinical grade material will be addressed. B. Adenoviruses Human adenovirus was isolated for the first time in 1953 from cultured adenoidal tissue [2, 3]. Since then, 51 different serotypes have been isolated from various tissues and excretions of humans, of which serotypes 42-51 were obtained from immunocompromised individuals [4-6]. A serotype is defined on the basis of its immunological distinctiveness as judged by quantitative neutralization with animal antisera (horse, rabbit). If neutralization shows a certain degree of cross-reaction between two viruses, distinctiveness of serotype is assumed if (i) the hemagglutinins are unrelated, as shown by lack of cross-reaction on hemagglutination-inhibition, or (ii) substantial biophysical/ biochemical differences in DNA exist [7]. Human adenoviruses are subdivided into six different groups (A-F), which are based mainly on differences in hemagglutination, restriction enzyme analysis, and DNA homology [8]. The adenoviruses were found to be associ ated with different disease patterns (see, e.g., [9, 10]). In addition to the human adenoviruses, some 40 different serotypes have been isolated from various animal species [11]. All adenoviruses possess a DNA molecule that is surrounded by a capsid consisting essentially of hexon, penton-base, and fiber proteins. The virion has an icosahedral symmetry and, depending on the serotype, a diameter of 60-90 nm. The well characterized adenovirus serotypes 2 and 5 have a linear double-stranded DNA genome of approximately 36,000 base pairs (Fig. 1). Other adenoviruses have genome sizes ranging from 30 to 38 kb. The genome contains, at both its ends, identical inverted terminal repeats (ITRs) of approximately 90-140 base pairs with the exact length depending on the serotype. The viral origins of replication are within the ITRs exactly at the genome ends. Sequences required for encapsidation (^) of the viral genome are located in a region of approximately 400 bp downstream of the left ITR. The structure of the adenoviral genome is described on the basis of the adenovirus genes expressed following infection of human cells, which are called early (E) and late (L), according to whether transcription of these regions takes place prior to or after onset of DNA replication. 6. Propagation of Adenoviral Vectors 1 3 1 "¥ E1 pIX L1-L3 ITR IVa2 E2B Figure 1 Map of the adenovirus genome. The 36-kb (for adenovirus type 5) double-stranded DNA molecule is usually divided into 100 map units (mu). The early (E) and late (L) regions are indicated on the map. The ITR sequences (inverted terminal repeats) are identical, inverted, terminal repeats of approximately 100 bp, depending on the serotype, which are required for replication. ^ is a stretch of sequences involved in packaging of the viral DNA into particles. El comprises the El A and ElB region, both encoding two proteins, which are described in detail in section I.C. E2A encodes the DNA binding protein, E2B the precursor terminal protein and DNA polymerase. E3 encodes a number of proteins that are predominantly involved in modulating the host's immune response against adenoviral infected cells. E4 proteins (six in total) are involved in modulation of gene expression and viral replication, mainly through interactions with the host cell. IVa2 (transcriptional activator of major late promoter) and pIX (essential for assembly of the virion) are intermediate proteins. LI -15 encode the late proteins, which are mainly capsid proteins, including penton (L2), hexon (L3), hexon-assembly (L4), and fiber (L5) protein. Infection of a target cell starts by interaction of the fiber with a receptor on the surface of the cell. Many, but not all [12], adenoviruses use the Coxsackie- adenovirus receptor (CAR) for this [13,14], which is present on the cell surface. Integrins act as secondary receptors by binding to the viral penton-base protein. Subsequently, the virus is internalized by receptor-mediated endocytosis. The adenoviruses escape from the endocytic vesicles (or receptosomes) by virtue of a change in the configuration of the virion surface due to the low pH in these vesicles. As a consequence, the virus particles are released in the cytoplasm of the cell, where they are further degraded [15], with the DNA ending up in the nucleus, where a complex with histone proteins is formed, which may attach to the nuclear matrix for replication [16]. The adenovirus DNA is usually not integrated into the host cell chromo somal DNA but remains episomal (extrachromosomal) unless transformation or tumorigenesis has occurred. C. Adenovirus Replication As indicated before, a productive adenovirus infection is divided into two distinct phases: the early and the late phases. In the early phase, the so-called early genes (El, E2, E3, and E4) of adenovirus are expressed to prepare the host cell for virus replication. During the late phase, actual viral DNA replication and production of viral structural proteins takes place, leading to the formation of new viral particles. Adenovirus replication requires both host-cell and viral proteins (see [8,16] for reviews). The cellular proteins needed for replication are nuclear factors I, II, and III [16], which are involved in initiation of viral DNA replication and elongation, as well as in increasing the efficiency of replication. 1 3 2 Nichols ef o/. Adenovirus DNA replication starts with expression of the "immedi ate-early" El genes. The El region comprises two different transcription units, ElA and ElB. The main functions of the ElA gene products are (i) to induce quiescent cells to enter the cell cycle and resume cellular DNA syn thesis and (ii) to transcriptionally activate the ElB gene and the other early regions (E2, E3, E4). The ElA region encodes two major RNA products, 12S and 13S, which are generated by one transcription unit and which differ in size due to alternative spHcing. The RNAs encode acidic proteins of 243 and 289 amino acids, respectively (for adenovirus 5). These are phosphorylated proteins, present in the nucleus of the cells. In addition, during lytic infection mRNAs of 9S, lOS, and IIS are produced, but these proteins were found to be not essential for adenoviral replication [17, 18]. The function of these proteins has not yet been resolved. The ElB region codes for one 22S mRNA, which is translated into two proteins, with molecular weights (for adenovirus 5) of 21 and 55 kDa. ElB proteins assist ElA in redirecting the cellular functions to allow viral replication. The ElB 55-kDa protein forms a complex with the E4 open reading frame 6 (ORF6) 34-kDa protein, which is localized in the nucleus [19, 20]. Its main function is to inhibit the synthesis of host proteins and to facilitate the expression of viral genes. In addition, it also blocks the p53 tumor-suppressor protein, thereby inhibiting apoptosis [21]. The ElB 21-kDa protein is important for quenching the cytotoxic effects to the target cells induced by ElA proteins. It has anti-apoptotic functions similar to the human Bcl-2 protein, which is important for preventing premature death of the host cell before the virus life cycle has been completed [22]. Mutant viruses incapable of expressing the ElB 21-kDa gene-product exhibit a shortened infection cycle that is accompanied by excessive degradation of host cell chromosomal DNA (J^g-phenotype) and an enhanced cytopathic effect (cyt- phenotype) [23]. The E2 region encodes three different proteins that function in viral DNA replication: an Ad-specific DNA polymerase, the precursor terminal protein (pTP), and the DNA-binding protein [16]. The DNA-binding protein, which is encoded by the E2A gene, binds to single-stranded DNA and is involved in unwinding duplex DNA. It might also be involved in the regulation of transcription. The precursor of the terminal protein (pTP) and the DNA polymerase, which are present as a heterodimer, are encoded by the E2B region. The pTP is attached to the adenoviral DNA and is cleaved by the viral protease late in infection. It has a function in protection of the DNA from nucleolytic breakdown and in attaching the adenoviral DNA to the nuclear matrix, which may localize the viral genome to areas of the nucleus in which high concentrations of replication and transcription factors are present. The polymerase is involved in the synthesis of
new DNA strands. 6. Propagation of Adenoviral Vectors 1 3 3 None of the E3 products are required for virus replication. They do, however, play an important role in virus multiplication in vivo, since they pro tect virus-infected cells from being eradicated by the host's immune response (review^ed in [9]). Several differentially spliced mRNAs are synthesized from the E4 region during infection and six different polypeptides have been iden tified in infected cells [24]. These proteins are involved in modulation of gene expression and viral replication, mainly through interactions w îth the host cell. The E4 ORF3 and E4 ORF6-encoded proteins are involved in post- transcriptional processes that increase viral late protein synthesis. They do so by facilitating the cytoplasmic accumulation of the mRNAs encoding these proteins and by expansion of the pool of late RNAs in the nucleus, most likely by influencing splicing. In addition, the E4 ORF6-encoded protein forms a complex v^ith the ElB 55-kDa protein that selectively increases the rate of export of viral late mRNAs from the nucleus. The complex is located in so-called viral inclusion bodies, the region w^here viral DNA replication, viral late gene transcription, and RNA processing occur [25]. The E4 ORF6 protein, either alone or in a complex w îth the ElB 55-kDa protein, binds the cellular protein p53, thereby blocking its potential to activate the transcription of tumor-suppressing genes [26, 27]. E4 ORFl sequences are related to dUTPase enzymes. It has been hypoth esized that this gene has a role in stimulating quiescent cells [24]. The E4 ORF4 protein binds to protein phosphatase 2A, v^hich results in hypophosphorylation of some proteins, including the adenovirus El proteins. This perhaps limits cytotoxic effects of El A and may lead to a more productive infection. It is also in line w îth the observation that E4 ORF4 mutants are more effective than w^ild-type viruses in killing nonpermissive rodent cells [28]. E4 ORF4 also induces apoptosis in transformed cells like 293 cells [29]. The E4 ORF6/7 modulates the activity of the cellular transcription factor E2F, w^hich may subsequently activate cellular genes v^hich are important for the S phase [30]. The functions of E4 ORFl, ORF2, and ORF3/4 during lytic infection are less clear and are dispensable for grow^th of the virus in laboratory cell lines. After onset of DNA replication, expression of the late genes L2-L5, w^hich are all under the control of one promoter, is sw^itched on. These genes encode the structural components of the virus particles, including L2 the penton, L3 the hexon, L4 the hexon assembly, and L5 the fiber protein. These proteins form the new virus particles into which the adenoviral DNA becomes entrapped. Depending on the serotype, 10,000-100,000 progeny adenovirus particles can be generated in a single cell. The adenoviral replication causes lysis of the cells. 1 3 4 Nichols ef al. II. Cells Expressing El of Adenovirus A. Transformation of Cells by El of Adenovirus In the previous section of this chapter, the function of adenoviral gene products in the repUcation of adenovirus w âs described. There is extensive influence of adenoviral proteins on a large number of cellular functions. In the absence of lytic viral replication, adenoviral genes may have a profound effect on cellular functions, the most striking being transformation by the adenoviral El A and ElB proteins. Clearly, these proteins interfere v^ith the regulatory mechanism of cellular proliferation. Human adenoviruses have a narrow^ host range for productive infections, and can only be propagated in cells of human, chimpanzee [31], pig [32], and cotton rats [33]. In rodent cells, e.g., from rat (w îth the exception of the cotton rat), hamster, or mouse, they bring about an abortive infection, which occasionally leads to transformation [34]. In the transformed cells the adenoviral DNA is integrated into the genome and at least the genes of the viral El region are expressed (review^ed in [35]). The viruses that w êre used for such studies w êre mainly adenovirus serotypes 2, 5, and 12. The various Ad serotypes differ in their ability to induce tumors upon inoculation into newborn hamsters; for example, Ad type 5 (Ad5) is nononcogenic [36], whereas Adl2 is highly oncogenic [34]. However, all Ad serotypes or their DNA can transform rodent cells [37, 38]. Ad5El- transformed cells can form tumors only in immunodeficient mice and rats, whereas Adl2El-transformed cells are oncogenic both in immunodeficient and in immunocompetent animals [39], which correlates with the ability of Adl2El to repress expression of MHC class I genes [40]. In culture, both rodent cells, e.g., from rat, mouse, or hamster, and human cells can be transformed by Ad DNA, although human cells, including fibroblasts and epithelial cells, are relatively refractory to transformation. Ade novirus DNA transformed human cell lines have been made from cultures of human embryonic kidney [41, 42], human embryonic retina [43-46], human embryonic lung [44], and recently, human amniocytes [47]. As described before, the El region consists of two transcriptional units, ElA and ElB. For complete morphological transformation, both regions are needed, but the ElA region by itself can immortalize rodent cells [48] and occasionally human cells [43], albeit with very low efficiency. Expression of ElA usually results in induction of programmed cell death (apoptosis), which can be prevented by coexpression of ElB [49]. The ElA associates with a number of cellular proteins, including the tumor suppressor gene product pRb, as well as pl07, pl30, cyclins A and E, cyclin-dependent kinase 2 (cdk2), and p300 (reviewed in [50-52]). Most of these proteins are involved in cell- cycle control, and, with the exception of p300, regulate the activity of the 6. Propagation of Adenoviral Vectors 1 3 5 transcription factor E2F [51]. The El A proteins do not exert their activity in initiation of transcription by direct, sequence-specific binding to DNA, but rather do so by binding to cellular transcription factors. The ElB 55-kDa [19] and 21-kDa [53] proteins cooperate independently with El A in transformation, and are required to inhibit the apoptotic response initiated by ElA. The 55-kDa ElB protein inhibits apoptosis by blocking the function of the p53 tumor-suppressor protein, which mediates ElA-induced apoptosis [21]. The 21-kDa ElB protein inhibits apoptosis in a way similar to the cellular Bcl-2 protein [22]. B. El -Expressing Cell Lines for Adenoviral Vector Production Most adenoviral vectors currently used in gene therapy have a deletion in the El region, where novel genetic information can be introduced [54]. The El deletion renders the recombinant virus replication-defective, which is a prerequisite for most of the clinical applications. In order to be able to produce El-deleted recombinant adenoviral vectors, complementing cell lines have to be used that express the El proteins of adenovirus. One of the main challenges here is to express sufficient levels of the El protein to achieve this. However, adenovirus El proteins, and in particular ElA proteins, are very toxic to cells. ElA has a profound effect on the transcription of many cellular genes, which leads to alteration of the morphology and growth of the cells and may lead to apoptosis. A few examples have been reported in literature, where cells have been immortalized (but not transformed) with ElA only. This has been described both for rodent [48] and for human cells [43]. It is not known whether cells that express ElA only are able to complement adenoviral vectors that are deleted for both ElA and ElB. Attempts have also been made to express El proteins in established cell lines such as A549. Growth of established cells is not dependent on El expression and the toxicity of El proteins made it difficult to isolate clones that show stable expression of the El proteins, although a few papers report encouraging results [46, 55, 56]. To the best of our knowledge, there is limited use of such cells and therefore this chapter will deal mainly with the group of El-expressing cells that use the transforming capacity of the adenoviral El genes. Typical examples are the cell lines derived from human embryonic kidney (HEK) [41, 42], human embryonic retina (HER) [43-46], and human amniocytes [47]. The advantage of using El for immortalization is that such cells are dependent on El expression for growth, and therefore the levels of El expression are remarkably constant over time. The vast majority of cell lines that were made by immortalization and transformation of primary cells, were made to study immortalization and transformation and were not made for propagation of El-deleted adenoviral 1 3 6 Nichols ef al. vectors. The only documented cell lines based on the El immortalization principle, which were made specifically for use in gene therapy are the PER.C6 cell line [46] and the amniocyte-derived cell line [47]. These cell lines have been tailor-made for the manufacture of clinical lots of adenoviral vectors, with special attention to avoiding generation of RCA (see below). In addition, proper documentation and adequate safety testing are pivotal to ensure manufacture of safe batches of adenoviral vectors. As PER.C6 is the only cell line currently used for making clinical lots of adenoviral vectors, a description of the generation of PER.C6 is given below. Also, the performance of the cell line in production of recombinant adenovirus as well as results of safety and genetic testing are provided. III. PERX6 Prevents RCA during Vector Production A. RCA The majority of preparations of El-deleted adenoviral vectors have been produced on 293 cells. This cell line was generated in Leiden in the group of Prof. Van der Eb, by transfection of El sequences of adenovirus type 5 into primary human embryonic kidney cells [41]. The aim of this experiment was to study the transforming potential of adenoviral El sequences, and the DNA used for it was sheared adenoviral DNA [41]. Precise mapping of the adenoviral sequences present in this cell line indicated that the cell line had integrated bases 1-4137 of the adenoviral DNA [57]. Adenoviral vectors carry a deletion in the El region that runs from approximately nt 400 to nt 3500 of the adenoviral genome. This means that there is a substantial sequence overlap between the El sequences present in the cell line and the adenoviral vector DNA (see Fig. 2). This sequence overlap may result in homologous recombination between the sequences. Due to a double crossover, the El region present in the cellular chromosome may end up into the El-deleted adenoviral vector [58] (Fig. 2). The resulting virus is El-positive and therefore capable of replicating independently in cells that do not contain El sequences in the chromosome. Several reports have described the occurrence of RCA in adenoviral vector batches produced on 293 cells [46, 55^ 58-60]. RCA in clinical preps is unwanted, both from the manufacturing and the safety points of view. Its appearance in batches is a chance process and is therefore unpre dictable and difficult to control. This is a significant problem for GMP manufacturing, in particular if large-scale batches have to be prepared. It is also unwanted from a safety point of view, as upon coinfection of a cell RCA causes the El-deleted adenoviral vector to replicate in an 6. Propagation of Adenoviral Vectors 1 3 7 293 Vector RCA [] [ Figure 2 Mechanism of generation of RCA in 293 cells. Adenoviral vectors contain sequences tfiat overlap with sequences present in the genome of 293 cells, indicated by the crossing lines. Due to the sequence homology, crossover events can occur, which lead to exchange of DNA. El sequences replace the transgene in the adenoviral vectors, resulting in El-containing adenoviruses that are replication-competent. uncontrollable way. It causes shedding of the vector [61]. In addition, RCA has been shown to cause inflammatory responses [59, 62]. Therefore, RCA generation during production of El-deleted adenoviral vectors has to be circumvented. B. PER.C6: Absence of Sequence Overlap Eliminates RCA Generation The strategy to prevent RCA occurrence was to eliminate sequence overlap between the El sequences present in the cellular genome and the adenoviral vector [46]. A potential hurdle to do this is the way the ElB and pIX gene are regulated. Both ElB and pIX use the same poly(A) sequences [63]. Furthermore, the pIX gene is not expressed upon transfection in cultured cells [64], but can be expressed only if present in an adenoviral genome. Therefore, an RCA-free packaging system should consist of two components: (i) an adenoviral vector that is
deleted for ElA and ElB, but contains the pIX expression cassette and (ii) a cell hne that expresses ElA and ElB and is devoid of pIX sequences. 1. El Construct Used for Making PER.C6 To create the novel cell line, the aim was to use only a minimal number of human adenovirus-type-5-derived sequences, i.e., the El protein coding sequences only, to prevent sequence overlap with El-deleted Ad. The El promoter and poly(A) sequences were therefore obtained from nonadenovirus 138 Nichols ef a/. sources. The El promoter was replaced by the human phosphoglycerate kinase (PGK) promoter (see below), which is a known housekeeping promoter [65] and the poly(A) sequences were isolated from hepatitis B virus [66, 67], The construct pIG.ElA.ElB contains, in addition to the ElA and ElB coding regions, sequences upstream of the ElA gene, including ElA enhancer elements and the cap sequence. Untranslated ElA sequences were also retained in the construct. These elements were included since earlier studies indicated that this results in efficient expression of the ElA gene [68]. A map of the construct, designated as pIG.ElA.ElB, is presented in Fig. 3. Despite removal of the splice site at position 3509 of the adenoviral genome [63], which is highly conserved, and truncation of the ElB transcript, high expression levels of both ElB 21 kDa and ElB 55 kDa were obtained [46]. In fact, the expression of the ElB proteins was even higher than in 293 and 911 cells, whereas equal expression levels of ElA were observed [46]. To prevent sequence overlap with El present in PER.C6 cells, adenoviral vectors were constructed that carry a deletion of the complete El region. These vectors were shown to propagate efficiently in PER.C6 cells (see below) and were found to express the pIX gene [46]. 2. Generation of PER.C6 The primary cells selected for making a new El-complementing cell line were human embryonic retinoblasts (HER). The choice for retinoblasts [43] was based on the observation that Adl2 could transform hamster retinal cells in vitro [69] and induce retinoblastomas following intraocular injection into newborn baboons [70]. It has been described that these cells can be immortalized relatively easily by El of human Ad5 [43, 44, 71] and Adl2 [72]. In addition, the 911 cells, which are derived from HER cells, are very efficient in production of recombinant adenoviral vectors and easy to use [71], thus providing a second argument for the use of primary HER cells as the source of primary cells to make a novel cell line. Primary HER cells have a limited life span. Such cells can be cultured for only a few passages, after which the cells senesce. Transfection of HER cells with El constructs results in immortalization and transformation of the cells, reflected by focus formation in the cultures. This is easily recognized by both macroscopic and microscopic examination of the cultures. Such foci '̂î UK ^Pf P^- ''"- "' " , ^ j ' ^ , ,i,^;;fvi^,;^.'.^..-„: • p{'^). Figure 3 The El construct used to generate PER.C6. The pIG.ElA.ElB construct contains aden ovirus type 5 sequences 4 5 9 - 3 5 1 0 . E l A expression is driven by the human PGK promoter. ElB transcription is terminated by hepatitis B virus-derived poly(A) sequences. 6. Propagation of Adenoviral Vectors 1 3 9 can be isolated and cultured further. Therefore, the pIG.ElA.ElB construct was transfected into primary HER cells, and PER.C6 cells were isolated as described in detail before [46]. After propagation of the cells to passage number 29, a research master cell bank was laid down, which was extensively characterized and tested for safety, including sterility testings (see below). Immortalization of primary cells with El sequences of adenovirus guar antees (i) a stable expression of El proteins, as the cells need El expression for growth, and (ii) that no external selection marker is needed to distinguish El expressing from nonexpressing cells. Human adenovirus serotype 5 was taken as the donor for El sequences. C. Frequency of RCA Occurrence In order to test whether PER.C6 cells are able to propagate adenoviral vectors without concomitant generation of RCA, El-deleted adenoviral vectors were propagated on 293 cells and on PER.C6 cells. The adenoviral vectors used did not have any sequence overlap with El sequences in PER.C6. The batches of vector were analyzed for the presence of RCA, using cell culture based assays, as described before [46, 58]. The results (summarized in Table I) clearly indicate that adenoviral vectors when propagated on 293 cells, get contaminated with RCA. On the other hand, the data provided in Table I clearly demonstrate that PER.C6 cells support RCA-free propagation of El deleted adenoviral vectors, even if large-scale batches (produced on 1-3E10 Table I Frequency of RCA Occurrence in 293 Cells and in PER.C6 Cells Helper cell No. of No. of cells per No. of RCA productions production positive batches 2.5E9 lU 2.5E10 lU 293 22 1E8-3E9 13/22 ND PER.C6 8 1E8-3E9 0/8 0/2 PER.C6 3 1E10-3E10 ND 0/3 Note. Batches of El deleted adenoviral vectors, propagated on either 293 and PER.C6 cells, were tested for the presence of RCA at a level of sensitivity of either 1 RCA in 2.5E9 infectious units (lU) or 1 RCA in 2.5E10 lU of El deleted adenoviral vector. The number of batches that were produced on either cell line, as well as the number of cells used for the production, are indicated as well. 1 4 0 Nichols ef al. PER.C6 cells) were tested for RCA in a very sensitive assay (1 RCA/2.5E10 infectious units). In a separate experiment, an El- and E3-deleted Ad5 vector w âs derived and propagated in PER.C6 cells. A master virus seed (MVS), prepared from passage 12, w âs used to generate 8 virus-production lots (passage 13). The unprocessed virus harvest (vector-infected suspension culture) of the MVS and the virus-production lots w êre tested for RCA. In brief, test articles were frozen and thawed and then assayed by inoculation onto the human- lung-carcinoma (A549, ATCC CCL 185) cell line for approximately 1-2 h at 37°C, after which the inoculum was removed and the culture was refed with medium. Cultures were passaged three times to amplify any putative RCA present, with incubation times ranging from 4 to 7 days for the early passages and 2 to 5 days for the final passage. The cultures were exam ined for cytopathic effects at each passage. The virus-production scale was approximately 20 L and a 60 mL volume (diluted to 600 mL to avoid toxicity and interference with detection of RCA) was tested for RCA for each lot. The testing volume was selected on the basis of a worst-case calculation to ensure the testing of at least three dose equivalents of virus. Earlier virus- production studies suggested that the freeze-thaw extract would contain at least 5 X 10^ particles/mL (or 10^^ particles/20 mL). Thus, at least 3 x 10^^ Ad5 particles (three dose equivalents) would be tested. Assuming a random (Poisson) distribution of RCA, if there were an average of one RCA per 1 X 10^^ particles (20 mL), one would predict the probability of not detecting it by testing only 1 x 10^^ particles to be = e~^ or 0.3679 (36.79% chance). By testing 3 x 10^^ particles (60 mL), the (binomial) probability of not detecting 1 RCA/1 X 10^1 particles is reduced to = ^'^ or 0.04979 (4.98% chance). Mathematically, this is equivalent to three independent tests of 20 mL each (60 mL total). No RCA was detected in the MVS or in any of eight virus-production lots assayed. Using the ratio of particle/TCID50 determined for purified virus (15.6 particles/infectious units), the virus-production lots were estimated to have an average of 1.9 x 10^^ particles/mL. It was estimated that the mean probability of not detecting at least one RCA in a dose of 10^^ particles of virus-production lots was 0.000887%. Besides having directly tested the infected cell suspension of the MVS for RCA, the repeated inability to detect RCA in the various clinical batches bodes well for the RCA-free nature of the MVS. For the clinical production runs, 1 mL of MVS is used to inoculate each of 100 roller bottles (RBs). This means a total of 800 mL of MVS have been used for these "clinical lots." Following the same calculation scheme as above, if there were one RCA per 20 mL of the MVS, there would be ^"^^^ or 0.00674 probability (0.674% chance) of not transmitting an RCA when preparing a single clinical batch. Moreover, cumulatively across the eight clinical production runs, there would be only (e'^^'^f or a 4.25 x 10-^^ probability (4.25 x 10"^^%) chance of not 6. Propagation of Adenoviral Vectors 1 4 1 transmitting RCA in the preparation of eight lots. In conclusion, the 60-mL freeze-thaw sample used for RCA testing provided adequate assurance for the detection of RCA in virus-production lots, at a level of one RCA for a 10^^ dose. How^ever, for testing of future Ad5 vector lots, we plan to use a clarified lysate. In this case, the probability estimated for detection of RCA w îll be based on more direct measurement of virus concentration. In summary, eliminating overlap betw^een El sequences in the cell and the El-deleted adenoviral vector eliminates RCA. IV. Production of Adenoviral Vectors A. Vector Stability When constructing El-deleted adenoviral vectors, a number of choices must be made regarding the structure of the vector backbone and the com position of the transgene. One must determine if the size of the El deletion v îll be adequate to accommodate the size of the transgene or if additional deletions, such as in the E3 region, w îll be needed. One must also decide on the placement of the transgene w^ithin the genome (El vs E3) and the orientation of the transgene (El parallel vs El antiparallel). Finally one must decide on the composition of the transgene in terms of the transcrip tional regulation elements that are utilized (promoter and polyadenylation signals). All of these parameters make constructing adenoviral vectors that express the transgene to the desired level, are genetically stable and propagate well enough to allows high-level production, a somew^hat empirical process. The net genome size of the vector, the deletions used, transgene orientation, the composition of the transgene and the transgene product itself can all affect the growAth and productivity of the vector. The degree to which vector and transgene structure can effect genomic stability and productivity is illustrated by our experience with Ad5 vector 1 (Fig. 4). Vector 1 contains an El deletion into which the transgene was introduced in the El antiparallel orientation. The transgene is composed of our gene of interest flanked by the immediate-early gene promoter and intron A from the human cytomegalovirus, and the bovine growth hormone polyadenylation signal sequence. In addition to the deletion of the El region, the vector has an E3 deletion [73]. When the genetic stability of vector 1 was assessed after serial passage in PER.C6 it was found to be unstable. Restriction analysis of purified viral DNA recovered from passages 12 to 19 indicated that the virus population contained genetic variants (Fig. 5). Over this passage series, the proportion and number of variants appeared to increase. An analysis of the novel restriction fragments and close to 1000 individually recovered, circularized viral genomes, indicated that two genetic mechanisms could account for all of the observed 142 Nichols ef o/. Vectorl Hindlll Hindlll HindlllHindlll Hindlll Hindlll Hindlll Hindlll ITR4^BGH Gene Intron A HCMV pA Promoter Figure 4 Genetic structure of Ad5 vector 1. MW P12 P13 P14 P15 P16 P17 P18 P19 pV1 7.0 i 6.0 i' 9tiK^ ^i-i ^ ^ ^ H i^K \|\|g\|gt ̂ H S B J B iw%^S ̂ M ̂ E fli B H Figure 5 Genetic structure of serially passaged vector 1. Viral DNA was purified from passages 12 to 19 of vector 1 digested with H/ndlll and end-labeled with [P^^]-dATP. The end-labeled restriction fragments were then size-fractionated by gel electrophoresis and detected by autoradiography. pVl , the plasmid used to derive Vector 1 is shown for comparison. The position in the vector 1 genome to which the restriction fragments correspond is indicated on the right. The reduction and upward shift in the 6.6-kb transgene-containing restriction fragment (uppermost double arrow) is due to amplification of
the 107-bp sequence in the packaging region. Novel bands seen at approximately 4.8 and 3.2 kb (arrowheads) are due to deletions in the transgene in association with amplification in the packaging region. 6. Propagation of Adenoviral Vectors 1 4 3 RFLPs: (i) deletions of the transgene expression cassette, particularly in the region of the hCMV promotor and intron A, and, in two instances, deletion of only adenovirus sequence; and (ii) amplification (two to four repeats) of a 107-bp sequence in the region containing the viral packaging elements. No rearrangements or insertions in the E3 region were detected. The genetic analysis of vector 1 has led to the development of highly stable vectors that can be easily propagated in PER.C6 cells, suggesting that the genetic instability can be overcome by vector design and is not necessarily related to the use of PER.C6 cells. B. The Production Process To make El-deleted adenoviral vectors for human gene therapy, a scaleable process suitable for commercial manufacturing under GMP condi tions was developed. One of the key factors in the development of cell-culture- based production processes is the culture system. In particular, if scaling of the process is needed, culture of the cells in a bioreactor is highly desired. For robust and scaleable systems, suspension growth of the required cell line is extremely advantageous. PER.C6 cells can be cultured both as adherent cells and in suspension culture. For suspension growth, specific well-defined serum-free media have been developed (e.g., ExCell 525; JRH Biosciences). These media do not contain any protein that is derived from human or animal tissues or specimens. This results not only in many fewer contaminants to be removed during downstream processing but also a favorable safety profile with respect to pathogens which might be introduced by animal/human-derived components. The serum-free culture medium (SF-medium) supports the growth of PER.C6 cells to densities of 1.5-2.5 x 1^ cells/mL in routine T-flask and roller bottle cultures. In perfused bioreactor systems, cell densities up to 1^ cell/mL are easily obtained. An overview of the process of production of El-deleted recombinant adenoviruses is presented in Fig. 6 and is summarized below. After thawing a vial of PER.C6, expansion in a T-flask containing SF- medium is done, followed by transfer of the suspension culture to roller bottles. Then these roller bottles are cultured until sufficient cells are generated to inoculate a bioreactor. In the standard batch-wise production process (e.g., in 2- or 20-L bioreactor) half of the bioreactor working volume is inoculated at 0.5 X 10^ cells/mL. Then PER.C6 is grown in 2 days to 2 x 10^ cells/mL and diluted once to 1 x 10^ cells/mL by adding the same volume of fresh medium. Then the seed virus is added and temperature is lowered from 37° to 35°C, followed by harvest after 3 days by pelleting. The latter is necessary if the purification process consists of ultracentrifugation with CsCl density gradients. After these 3 days, the virus particles become suspended utilizing 144 Nichols ef al. ,iR £^ n U ftraf iltration Capture DOOQO S E C I E X DDODO Final Product Figure 6 Overview of the process of production of El-deleted recombinant adenovirus. The process is described in section 4B of this chapter. cell lysis. The batch process is very robust but not economical since only low cell densities can be obtained due to the rapid consumption of nutrients from the medium. When high cell densities are required a perfusion system can be used. Nutrients are replaced and metabolites removed by perfusion of fresh medium. A suitable perfusion system can be obtained with hollow fiber modules. These modules are operated externally on the bioreactor and can therefore easily be replaced when malfunction occurs. Hollow-fiber technology also has the opportunity for virus retention, easy scale-up, and its potential application as a first step in the virus isolation. To take full advantage of high-density cultures the virus replication should last longer than 3 days to enable the utilization of all cells present because a repeated infection can occur with newly released particles from lysed cells. A typical example of a 20-L bioreactor run is presented in Fig. 7. Because a large part of the total produced virus will be in suspension, the volume of such a culture is too large to enable purification by ultracentrifugation. Hollow-fiber ultrafiltration and chromatography are methods of choice for virus isolation and purification. With these systems directly connected to the bioreactor, thereby ensuring a closed system, all virus can be isolated from the culture medium. After capture of the virus, the bulk product can be further purified utilizing ion exchange chromatography and/or size exclusion chromatography systems. The obtained product is of high purity and infectivity. Final formulation can be done by ultrafiltration, bringing the product to the final concentration in the required buffer. 6. Propagation of Adenoviral Vectors 145 10 + 8 o t 6 1̂ "̂ S 1,5 a > • t 2 144 192 336 time (hours) Figure 7 Example of production of El -deleted adenoviral vectors in PER.C6 in a 20-L bioreactor. PER.C6 cells are seeded at a density of 0.5E6 cells/mL, in ExCell525 culture medium. Perfusion is started 48 h later, at a rate of 1 bioreactor volume/24 h. The glucose concentration remains constant during perfusion. Under these conditions, cell densities of 1 x 10^ cells/ml are obtained. C. Yields of Adenoviral Vectors The yields of virus obtained after propagation in PER.C6 cells in 20-L sus pension cultures ranges from 0.6 x 10^^ to 1.1 x 10^^ vp/mL culture medium with an average yield of 0.8 x 10^^ vp/mL {n = S), The cell density during infection was approximately 3 x 10^ cells/mL. The calculated virus yield per cell is therefore 0.2 x 10^-0.4 x 10^ vp/cell. As the cultures are inoculated at a multiplicity of infection of 40 vp/cell, an amplification factor of 500 was achieved. The loss during isolation and purification can be held to 70-80%. This figure was consistently obtained in multiple runs for three different adenoviral vectors. Similar yields of El-deleted adenoviral vectors obtained on PER.C6 have been obtained by others [74]. D. Scale of Adenoviral Vector Production The estimated scale of the required bioreactor and cell-line stability is calculated as follows. The cell density used for virus production in perfusion mode is 3-6 x 10^ cell/mL. Therefore, assuming at least 20,000 virus particles per cell yield, the overall expected yield in the crude bioreactor harvest is 2 x 10"̂ vp/cell X 5 X 10^ cell/mL = 1 x 10^^ vp/mL. Further, after optimization. 1 4 6 Nichols ef al. maximum expected loss of virus particles after downstream processing (DSP) by column chromatography is 75%, Therefore, from a 20-L perfusion bioreactor 1 X 10^^ vp/mL X 0.25 (recovery) x 10"̂ mL = 5 x 10̂ "̂ vp can be obtained. This gives 5 x 10̂ "̂ vp/1 x 10^^ vp/dose = 50,000 doses (assuming 1 x 10^^ vp/dose). When during product development 40% of the batch is retained for QC and archiving purposes 3000 patients can receive 50,000 x 0.6/3000 = 10 doses each. Therefore, using the currently developed technology, this 20-L bioreactor is sufficient for the generation of material for the first clinical studies. However, to be able to do process development on a larger scale, needed for full commercial production, a larger vessel is required. Full production scale is expected to be about five times larger, and therefore a 100-L bioreactor is expected to be the maximum volume required for application with single doses up to 10^^ vp. To propagate the cells from a working cell bank ampoule, containing 5^ cells, to a 5E6 cell/mL culture in a 100-L bioreactor would take 17 cell doubhngs. So a reliable production process would require a cell line which is at least stable over 20 cell doublings. PER.C6 was shown to be stable with respect to El expression for at least 98 cell doublings. V. Safety Considerations of PER.C6 A. QC Testing of PER.C6 Cells for Use in the Manufacture of Biologicals and Vaccines The safety of vaccines and biologicals manufactured in continuous cell lines of animal or human origin is of paramount importance and must be ensured by the manufacturer through a program of quality control (QC) testing applied to the product before release for human administration. This QC testing is intended to (i) ensure the identity of the product, (ii) ensure the safety and sterility of the product by demonstrating the absence of adventitious microbial agents, and (iii) ensure the safety and sterility of the product by demonstrating the absence of adventitious viral agents. The program for QC testing applied to a biological product, formalized as a release protocol^ is developed as a responsibility of a Department of BioAnalytical Development. The release protocol is developed through an evaluation and integration of (i) relevant compendial literature and precedents, (ii) the origin of the cell line used for production and its development as a master cell bank, (iii) the sourcing and quality control testing of raw materials of animal origin used in manufacture, and (iv) the method of good manufacturing practice (cGMP) manufacture of the bulk and intermediate and final product considering, among other things, the quality of environment in which bioprocessing is conducted, the method of manufacture, in particular the isolation of the culture system from operators, and the consistency of preparation. 6. Propagation of Adenoviral Vectors 1 4 7 The release protocol prescribes the QC testing to be applied not only to final product but, importantly, master cell banks, master virus seeds, and other bioprocess inputs, raw materials of animal origin, and intermediate bulk products developed during dov^nstream processing, purification and formulation. The release protocol specifies testing methods and volumes to be tested relying upon bacterial broth and agar cultures, embryonated eggs, small animals, and in vitro cell culture in a variety of primary and continuous cell lines of mammalian or human origin. These methods are well known to be sensitive to the detection of a variety of bacterial and viral agents and applied in concert provide a comprehensive and sensitive analytical approach upon which to ensure product safety. More recently, with the development of exquisitely sensitive polymerase chain reaction (PCR) methods for the detection of agents which are refractory to animal or cell culture, these classical propagation methods are commonly supplemented with agent-specific testing, using PCR and polymerase-enhanced reverse transcriptase (PERT) assays. The general methods of testing to ensure product safety are presented in illustrated form in Fig. 8. 1. QC Testing for the Release of PER.C6 Master Cell Bank The development of PER.C6 research master cell bank (rMCB) A068-016 to support manufacture of biologies has been previously described. The release protocol to ensure the (i) identity, (ii) sterility, and (iii) viral safety of the rMCB is presented in Table II. The QC testing was conducted by contract at Inveresk Research (Tranent, Scotland) and at MicroSafe (Leiden, The Netherlands). Table 11 Release Protocol for Crucell rMCB A068-0T6 Test Method Identity Isoenzyme analysis Sterility Broth and agar for cultivation of bacteria, fungi, mycoplasma In vitro indicator cells for detection of mycoplasma using Hoechst stain Viral safety In vivo eggs Eggs (allantoic and yolk sac) In vitro cell culture MRC-5, HeLa, Vero, bovine cells Agent-specific testing HBV, HCV, EBV, HHV6, HIV-1, HIV-2, using PCR HTLV-1, HTLV-2, AAV, B19, SV40 Agent-specific testing PERT, S+L-,XC testing for retroviruses 148 Nichols et aL Method Criteria for Evaluation g ^ L::D Turbidity, Sterility Colony Formation Inoculation of Broth and Agar Culture and Cell Cytoplasmic Fluoresence Cultures with Observation of 14-21 days ^ 5> Viability Gross morphology In Vivo Testing in Eggs Hemagluttination Injection of Eggs by Amniotic, Allantoic or Yolk Sac Routes with Obserx'ationfor 7-14 days In Vivo Testing in Animals Viability Fitness Injection of Adult or Suckling mice, Guinea pigs or Evidence of Disease Rabbits by IM, IP, or SC Routes with Observation for 7-60 days In Vitro Testing in Cell Culture Evidence of Cytopathology Hemadsorption Inoculation of Primary or Continuous Cell Lines of Hemagluttination Human, Primate or Animal Origin with Observation for 14-28 days Evidence of Gene Testing for Specific Virus Agents Specific Product Use of Sequence Specific Primers for PCR Evidence of Enzymatic Amplification or PERT, or TEM Activity ofRT Figure 8 Testing methods for the demonstration of product safety. 6. Propagation of Adenoviral Vectors
1 4 9 2. QC Testing for the Release of a PER.C6 Working Cell Bank The release protocol of research working cell bank (rWCB) A068-043W, according to the panel of testing, is presented in Table III. The QC testing was conducted by contract at Inveresk Research and at MicroSafe. This testing included tests for (i) identity, (ii) sterility, and (iii) viral safety in cells of human and simian origin. 3. Development of a Master Cell Bank at the Merck Research Laboratories Cryopreserved vials of the rWCB were obtained from Crucell by the Merck Research Laboratories and expanded under conditions of cGMP manu facture to create a master cell bank (MCB) for future manufacturing use. This MCB has been released for use in the propagation of recombinant adenovirus according to a release protocol presented in Table IV. The preponderance of this QC testing was conducted by Q-One BioTech (Glasgow, Scotland). This release protocol for the rWCB provides persuasive demonstration of the (i) identity, (ii) sterility, and (iii) viral safety of the PER.C6 MCB. This release protocol specifies animal testing in small animals to supplement the egg safety testing applied to the rWCB, expands the variety of primary and continuous cell lines used for viral safety using in vitro cell culture, and greatly broadens the variety of agent-specific testing using PCR-based testing and biochemical testing for retroviruses. The human cell line 293 was included in the panel of tissue culture cell lines in an attempt to detect the presence of any defective adventitious virus that requires the presence of El in the host cell. The direct assay for reverse transcriptase, as well as the detection of RT in cocultivation supernatant fluids, was done with the highly sensitive PCR-based reverse transcriptase (PBRT) assay. The supplemental PCR tests were included Table III Release Protocol for Crucell rWCB A068-043W Test Method Identity Isozyme Sterility Broth and agar for cultivation of bacteria, fungi, mycoplasma In vitro cell culture testing for mycoplasma Viral Safety In vitro cell culture Vero, MRC-5, PER.C6 Agent-specific testing Adeno-associated virus using PCR 1 5 0 Nichols et al. Table IV Protocol for Release for the a PER.C6 Master Cell Bank Test Method Identity Isozyme analysis DNA Fingerprinting PCR-Based Test for El Sterility Broth and agar for cultivation of bacteria, fungi, mycoplasma In vitro cell culture testing for mycoplasma Viral safety In vivo eggs Eggs (allantoic and yolk sac) In vivo animals Guinea pig, adult and suckling mouse In vitro cell cuhure VERO, MRC-5, 293, Rabbit Kidney 13, Vero, bovine turbinate, porcine kidney Agent-specific testing Transmission electron microscopy PERT for RT Raji, RD, H9 cell-cocultivation for retroviruses PCR for HBV, HCV CMV, EBV, HHV6, HHV7, HHV8, HIV-1, HIV-2, HTLV-1 6c HTLV-2, SiFV, SFV, AAV, B19, bovine polyoma, SV40 with due consideration for the human origin of the cell line and the use of bovine serum for the derivation of the cell line. The tumorigenic potential of the cell line w âs tested beyond the anticipated manufacturing cell-passage level. Satisfactory results w êre obtained from all QC testing. The results of the testings are presented in Table V. B. Tumorigenicity 1. Tumorigenicity Studies of PER.C6 Cells Three tumorigenicity studies v^ere carried out on the PER.C6 cell line. The results of these studies are summarized in Table VI. In the first study, nude (nu/nu) mice v^ere injected subcutaneously v^ith 10^ PER.C6 cells. Positive control animals v^ere injected subcutaneously w îth \0^ KB cells. KB is a know^n tumor-producing cell line derived from an epidermoid carcinoma (American Type Culture Collection; CCL-121). The study w âs conducted over 28 days, at w^hich point all animals were necropsied and examined grossly and histologically. All of the positive control animals had growing nodules, and 8 of 10 male mice and 7 of 10 female mice receiving PER.C6 cells had growing nodules, thus producing a positive test (Table VIA). At the time of the first study, 21 or 28 days was the duration that was usually used. Subsequently the Center for Biologies Evaluation and 6. Propagation of Adenoviral Vectors 151 Table V Summary of Testing of PER.C6 Research Master Cell Bank (Passage No. 29) Test Specification Result Sterility (EP) Negative Negative Mycoplasma (broth, agar and DNA Negative Negative staining) In vitro virology for adventitious Negative Negative viruses (28 days, with cytopathic effect and haemadsorption) on Vero, MRC-5, HeLa and PER.C6 cells (PTC) Specific viruses Human immunodeficiency virus Negative Negative types 1 and 2 Human T-lymphotropic virus types Negative Negative 1 and 2 Human hepatitis B + C Negative Negative Human cytomegalovirus Negative Negative Human parvovirus B 19 Negative Negative Human herpes virus 6 Negative Negative Simian virus 40 Negative Negative Adeno-associated virus Negative Negative Epstein-Barr virus Negative Negative Bovine viruses (BVD, IBR and PI3) Negative Negative In vivo virology in suckling mice (i.e. Negative Negative and i.p.), and embryonated eggs. allantoic and yolk sac injections (PTC) Isoenzyme test for human origin Confirmed Confirmed In vivo virology (adult mice, guinea Absence of adventitious Free from infectious pigs and suckling mice) and microbial adventitious transmission electron microscopy contamination microbial (TEM) contamination Reverse transcriptase assay Negative Negative S+L^ focus forming assay and XC Negative Negative plaque assay Tumorigenicity in nude mice Report result Tumorigenic Restriction analysis No evidence of No evidence of mutation or mutation or rearrangements rearrangements Sequencing Report sequence Sequence reported {continued) 152 Nichols et aL Table V {continued) Test Specification Result DNA profiling rMCB (passage 29) Late passage banding Late passage banding and late passage cells (passage 98) pattern resembles pattern resembles rMCB rMCB Karyotyping/chromosomal analysis Report chromosome Modal No. 86. Range numbers 68-106 Fluorescent product enhanced reverse transcriptase (PERT) assay Negative Negative S^L~ focus forming assay and XC plaque assay Negative Negative Multicolor fluorescent in situ hybridization (M-FISH) Report integration site Chromosome 14 Copy no. determination (fiber FISH analysis) Report results 13.6 ± 6 . 1 Prions No evidence for Determination of prions infectious PrP^^ Confirmed Sequence analysis Research (CBER) of the Food and Drug Administration had suggested the observation period be extended to 84 days. This was to give more time for slow growing tumors to appear and for nontumorigenic nodules to regress or disappear. Therefore, the tumorigenicity study on the PER.C6 cells was repeated. The second study was performed in nude (nu/nu) mice over an 84-day period. Thirty nude mice were injected subcutaneously with 10^ PER.C6 cells in 0.2 mL of serum-free medium. As a positive control, 10 mice were injected subcutaneously with 10^ HeLa cells in 0.2 mL of serum-free medium. As a negative control, 30 mice were injected with 0.2 mL of medium. The mice were palpated at the injection site every 3 to 7 days and any nodules found were measured in two dimensions. The PER.C6 cell test arm and the negative control arm had 10 mice necropsied 21, 42, and 84 days postinjections. The positive control arm was necropsied at 42 days postinjection. Gross and histological examinations were performed on all injection sites and nodules if they appeared. During the initial days after injection, palpable nodules were present at the subcutaneous injection sites in all animals inoculated with PER.C6 cells. Between postinjection days 5 and 14, the detectable masses disappeared from the injection sites. However, in several of these mice, the masses subsequently reappeared by around day 21 and continued to enlarge until the animals were necropsied. Of the mice injected with PER.C6 cells, 5 of 10 sacrificed on day 21, 5 of 10 sacrificed on day 42, and 1 of 10 6. Propagation of Adenoviral Vectors 1 5 3 Table VI Tumorigenicity of PER.C6 Cells A. Day 28 tumorigenicity of PER.C6 and KB cells in nude mice Cell type No. of cells Male Female KB 1 X 10^ 10/10 10/10 PER.C6 1 X 10^ 9/10 7/10 B. 84-Day tumorigenicity study of PER.C6 and HeLa cells Cell type No. of cells Day 21 Day 42 Day 84 HeLa 1 X 10^ NA 10/10 NA PER.C6 1 X 10^ 5/10 5/10 1/10 Medium control — 1/10^ 0/10 0/10 C. Titration tumorigenicity study of PER.C6 cells in nude mice Cell type No. of cells Day 21 Day 42 Day 84 PER.C6 1 X 10^ 0/10 0/10 0/10 PER.C6 1 X 10^ 0/10 0/10 0/10 PER.C6 1 X 10^ 5/10 9/10 7/10^ Medium — 0/10 0/10 0/10 Note. Details of the experiment are presented in Section 5B. ^ Benign lung adenoma. Seven animals sacrificed, with tumors on day 56 and leaving 0/3 at day 84. sacrificed on day 84 (actually sacrificed on day 49 due to tumor size) had gross or microscopic evidence of a tumor (Table VI B). Histologically, these recurrent nodules were composed of sheets of large pleomorphic cells with numerous, sometimes abnormal, mitotic figures. These masses compressed the surrounding tissues but were not invasive. No tumors were observed outside the injection sites. The interpretation of the test is that PER.C6 cells are positive for tumorigenicity. In view of the positive tumorigenicity results obtained following injection of 10^ PER.C6 cells, a titration study was performed in which nude mice were injected with PER.C6 cells at doses of 10^, 10^, or 10^ cells per animal. Mice were necropsied 21, 42, or 84 days postinjection. No animals receiving 10^ PER.C6 cells had palpable masses at the injection site from the first palpation day until necropsy. None of these animals had gross or microscopic evidence of nodules or tumor cell collections at any necropsy time point. Two of the 30 mice receiving 10^ PER.C6 cells had palpable nodules on postinjection day 3. 1 5 4 Nichols ef al. These masses disappeared by day 7 and did not recur. Gross and histological examination of the injection sites were negative at all necropsy time points. In the mice that received 10^ PER.C6 cells, 29 of 30 animals had palpable nodules on Day 3 — some of which disappeared or became smaller but most of these recurred and grew progressively until necropsy. At necropsy, 5 of 10 mice on day 21 had tumors, 9 of 10 mice sacrificed on day 42 had tumors, and 7 of 10 in the group scheduled for day 84 had tumors but were sacrificed on day 56 because of tumor size (Table VI C). The histological and gross features of the PER.C6 cell tumors were similar to those described for the previous study (above). No metastatic nodules were found. Thus, the tumorigenicity studies of PER.C6 cells were positive at 10^ cells per animal and negative at 10^ and 10^ cells per animal. This would indicate that not all of the PER.C6 cells are tumorigenic and/or a critical mass of tumorigenic cells are necessary for tumor formation. 2. Tumorigenicity Studies of Residual DNA from PER.C6 Cells In view of the positive tumorigenicity studies with 10^ PER.C6 cells, the oncogenic potential of residual DNA from these cells was tested in both nude mice and newborn hamsters. For these studies, DNA was isolated from passage 61 PER.C6 cells using standard procedures. The DNA preparation was shown to be of high molecular weight (average size ~100 kb) and devoid of significant protein or RNA impurities. In the nude mouse study, 20 female nude (nu/nu) mice were injected subcutaneously with 225 \|JLg of PER.C6 DNA (in a volume of 0.25 mL). For negative controls, two groups of 20 female mice each were injected subcutaneously with 0.25 mL of vehicle. Approximately 5 months after injection, the mice were necropsied and examined histologically for tumor growth. None of the mice in this study exhibited gross or microscopic evidence of tumors at the injection site. One treated mouse had a lymphoma at a distant site. However, nude mice — particularly females — are known to have a high incidence of spontaneous lymphoma [75-78], and the occurrence of a single lymphoma in 20 treated mice is consistent with the spontaneous incidence. Although the lymphoma was almost certainly a spontaneous event, a polymerase chain reaction (PCR) study was performed on the lymphoma DNA to determine if there was any evidence for the presence of the adenovirus El region — the transforming agent of PER.C6 cells. The study was negative, with a sensitivity of approximately one copy of El per 750 tumor cells. Previously, El expression has been shown to be necessary to maintain the transformed state of 293 cells, which, like PER.C6 cells, were transformed by El [79]. The results of the PCR
analysis support the conclusion that the lymphoma was a spontaneous event, not induced by PER.C6 DNA. A second tumorigenicity study using DNA from PER.C6 cells was carried out in newborn hamsters. Between 18 and 36 h after birth, female and male 6. Propagation of Adenoviral Vectors 1 5 5 hamsters (28 total) were injected subcutaneously with approximately 100 \|xg of PER.C6 DNA (in a volume of 110 \|JLL). TWO groups of control hamsters (50, mixed sex, per group) were injected with 100 \|JLL of vehicle. Several pups in each group were lost due to maternal cannibalism, reducing the group sizes to 20 (11 female, 9 male) in the PER.C6 DNA group, 40 (19 female, 21 male) in control group 1, and 45 (27 female, 18 male) in control group 2. After weaning, the hamsters were palpated on a weekly basis. The hamsters were necropsied approximately 5 months after injection and examined grossly and histologically for tumor growth. One female hamster in control group 2 died approximately 21 weeks after injection of a malignant ovarian teratoma. No evidence of tumors was found in the 20 hamsters that were injected with PER.C6 DNA. 3. Concerns about Using a Tumorigenic Cell Substrate The basis for concern about using a tumorigenic cell substrate to produce a vaccine includes three theoretical possibilities. First, DNA from the cells carrying a putative activated oncogene or cancer-causing mutation could be integrated into the recipient's genome and produce a tumor. Second, a transforming protein in the cells could be transmitted and result in a tumor. Third, an adventitious tumor virus may be present and could be transmitted to the recipient and produce a tumor. Concerning residual DNA from a tumorigenic cell substrate, there have now been several reports demonstrating that DNA extracted from tumorigenic cell lines or tumors growing in vivo — and even purified activated onco genes— do not produce tumors when injected into animals at levels up to 1000 \|xg of DNA [80-87]. The negative results obtained with PER.C6 DNA in nude mice and newborn hamsters are consistent with these findings. In the case of the PER.C6 studies, the amount of DNA injected (~100 or 225 \|JLg) rep resents a > 10^-fold excess compared to the amount of residual DNA present in a dose of vaccine produced on this cell substrate. Others have calculated that 100 pg of residual DNA from tumorigenic cells would be equal to less than a billionth of a tumor-producing dose [80-87]. The second concern, transforming proteins or growth factors, has been considered by a WHO study group to be significant only if they are continually produced by cells or have continued administration [80, 81]. The study group did not consider the presence of contaminating known growth factors, in the concentrations that they would be found, to constitute a serious risk in biological products prepared from continuous cell lines. The third category of concern, viruses or other adventitial agents, does present a potential risk. This risk is greatest when primary cells are used because of the frequent need for newly acquired cells that require repeats of the extensive testing for adventitial agents. Human diploid cell lines and continuous tumorigenic cell lines are thoroughly and routinely tested for a 1 5 6 Nichols et al. wide variety of known and unknown adventitial agents in a series of in vitro and in vivo assays, thus providing adequate assurance that adventitial agents will not be transmitted. C. Prion-Related Issues It is now generally accepted that an abnormal form of the cell surface glycoprotein PrP, or prion protein, is the main infectious agent in transmissible spongiform encephalopathies like scrapie, bovine spongiform encephalopathy (BSE), and Creutzfeldt-Jakob disease (CJD) ([88] and reviewed in [89]). The abnormal form of PrP, called PrP^^ or PrP-res, is characterized by a remarkable resistance to denaturing agents and to degradation by Proteinase K (Prot K). Diagnostic tests take advantage of this unusual stability that allows a distinction between PrP^ and PrP^^ using antibodies that recognize both forms ofPrP(e.g., [90]). Human prion diseases occur in sporadic, acquired or inherited forms with different clinical and pathological phenotypes (reviewed in [91]). In 1996 a new variant of CJD (vCJD) was reported in the United Kingdom in relatively young patients with clinical features different from the known CJD forms [92]. It was also found by strain typing that the prion protein of these patients was indistin guishable from the one that causes BSE, thus raising the question whether vCJD could be acquired by consumption of meat from cattle suffering from BSE [93, 94]. The possibility of transmission of PrP^^ from bovine to human raises safety issues for cultured cell lines used for the production of human drugs. Therefore, PER.C6 cells were carefully examined for the PrP phenotype (see below) as well as genotype. It has been found that specific mutations in the PrP gene are associated with hereditary forms of human prion disease (reviewed in [89] and [91]). Furthermore, a common methionine/valine polymorphism at codon 129 of the PrP gene appears to be associated with phenotypic variability and susceptibility to sporadic and iatrogenic CJD. The vast majority of patients suffering from sCJD and also from vCJD were found to be homozygous for 129 M, whereas patients heterozygous at codon 129 were strikingly underrepresented [95-97]. To examine whether the PER.C6 PrP gene contains any of the known mutations associated with susceptibility to prion diseases, the PER.C6 PrP gene was sequenced. For these sequencing studies, genomic DNA from PER.C6 cells was isolated, and used to amplify the PrP gene sequences by PCR. The resulting PCR product was cloned into a vector, and the PrP gene in each of 13 PrP-containing clones was sequenced by BaseClear (Leiden, The Netherlands). Five of these clones contained sequences coding for the 129 Methionine PrP^ protein, while the other eight contained the 129 Valine PrP^ sequence, demonstrating the heterozygosity at this position. To confirm this observation, the resulting PCR product was also sequenced. As expected, a double peak (g/a) was observed in the 129 codon at a position 6. Propagation of Adenoviral Vectors 1 5 7 defining it as a valine (if the nucleotide is a guanine) or as methionine (if the nucleotide is an adenine). The PER.C6 PrP gene sequence was then compared to the wild-type sequence published in GenBank (Accession No. M12899) and was found to be identical to the wild type gene; thus, ruling out the possibility that these cells possessed a hereditary mutation that would be predisposing for prion diseases. The sequence also revealed that PER.C6 cells are heterozygous for methionine/valine at codon 129. PrP^ is constitutively expressed in adult brain [90, 98, 99] and at lower levels in other tissues like liver and spleen [100]. PrP expression has also been found in a variety of rodent and human cell lines. Our studies on PER.C6 and 293 cells have shown that these cells also express the cellular form of PrP. A validated Western blot analysis of Prot K-treated protein extracts of PER.C6 cells and their parental HER cells has failed to detect any Prot K-resistant forms of PrP at passages 33 and 36 of PER.C6 cells and passage 6 of their parental HER cells. In addition to the sequencing of the prion gene and testing for the presence of abnormal prion protein in the PER.C6 cells at an early and late passage level of the culture, serum and trypsin batches that were used were traced to see if any were derived in the United Kingdom. Finally, it has been possible to adopt the PER.C6 cells to serum-free suspension so that bovine sera can be completely avoided in the future if desired. The above-mentioned characteristics of PER.C6 make it a safe manufac turing cell line in this respect. D. Genetic Characterization of PER.C6 Cells 1. Sequence Analysis of El The integrity of the ElA and ElB coding regions present in PER.C6 was tested by sequence analysis. This was done by bidirectional sequencing of PCR fragments generated from these regions, and the sequence of these fragments was compared to the original pIG.ElA.ElB sequence, the construct that was initially used in transfection. No mutations, deletions, or insertions were detected between the sequence of the PCR fragments and pIG.ElA.ElB, indicating that no genetic alterations were introduced in the ElA and ElB regions during transfection and subsequent culture of the cells. 2. Site of Integration of El The chromosomal integration site of the plasmid pIG.ElA.ElB in PER.C6 was determined by using the multicolor fluorescent in situ hybridization (MEISH) technique in combination with the principle of combined binary ratio labeling (COBRA) [101]. This technique combines 24-color COBRA-MFISH 1 5 8 Nichols et al. using chromosome-specific painting probes for all human chromosomes with plasmid probe (pIG.ElA.ElB) visualization (25th color). The pIG.ElA.ElB integration site was determined using PER.C6 cells that are derived from the research master cell bank (passage number 29). Cells were analyzed at passage numbers 31, 41, 55, and 99. Two hundred and fifty metaphases and interphases were studied. pIG.ElA.ElB integration was detected only on chromosome 14 (Fig. 9, see color insert) and in both sister chromatids of the chromosome in all PER.C6 passage numbers screened. Of the 47 metaphases and 203 interphases, 75-80% consisted of integration of pIG.ElA.ElB in one chromosome 14, whereas 2 0 - 2 5 % consisted of integration in two chromosomes 14 [102]. 3. Copy Number of the El Construct The number of copies of pIG.ElA.ElB present in the PER.C6 chro mosome was studied by Southern blot analysis, dot bot analysis and fiber FISH analysis [102]. Southern hybridization revealed the presence of several integrated copies of pIG.ElA.ElB in the genome of PER.C6 [46]. In addition, dot blot analysis showed a pIG.ElA.ElB plasmid copy number of 19 ib 3 (research master cell bank) and 24 ± 16 (extended cell bank, passage number 99) per genome. From the results it was concluded that PER.C6 consists of five to six copies of pIG.ElA.ElB per haploid genome. Fiber FISH enables physical length measurements of in s/Yw-hybridized DNA probes on linearized DNA fibers with a resolution equal to the theoret ical length of a linearized DNA molecule according the model of Watson and Crick (1 kb is 0.34 jxm). Therefore, fiber FISH was conducted to measure the length of the integrated construct in the PER.C6 cell line at pas sage numbers (pns) 31, 41, and 99. Twenty fibers were measured. It was determined that pIG.ElA.ElB was integrated in tandem copies in chromo some 14 of PER.C6. The copy number of these in-tandem integrations was determined to be as follows: pn31, 13.6 ± 6 . 1 ; pn41, 18 ± 4 . 5 ; and pn99, 20.1 ± 7 . 9 . 4. Chromosome Analysis PER.C6 cells from cellular passages 44 and 66 were harvested for chromosome analysis to determine the modal chromosome number and the karyotype in a sample of metaphase plates. Cells were harvested, and slides were prepared and stained using a standard giemsa banding (GTG) technique. At each passage level, the chromosomes in 50 metaphase plates were counted. Also, full karyotypes were prepared from each passage level. At passage level 44, the chromosome number ranged from 43 to 160. The mean number of chromosomes was 72 and the modal number was 6. Propagation of Adenoviral Vectors 1 5 9 61. All metaphase plates examined had structural chromosomal changes and rearrangements. A marker chromosome 19 with additional material in the long arm (19q^) was the most common alteration and was found in 14 of the 20 metaphase plates that were karyotyped. At passage 66, the chromosome number ranged from 42 to 112. The mean number of chromosomes was 63 and the modal number was 64. All metaphase plates karyotyped again were found to have structural changes. The 19q'^ was again the most common change, observed in 15 of 20 karyotypes. There was also a marker chromosome 11 with extra material in the short arm (llp+) in 14 of the 20 karyotypes and a marker chromosome 9 with additional material in the short arm (9p+) in 8 of the 20 karyotypes. Several of the markers differed at the two passage levels, but the most common marker, 19^+, was the same. The continuing changes seen as passage level increases is typical of heteroploid continuous cell lines. 5. DNA Fingerprinting PER.C6 cells were also analyzed on two
occasions by DNA fingerprinting. DNA profile analysis of PER.C6 indicated no changes in the banding pattern obtained between the research master cell bank (pn 29) and an extended cell bank that was laid down at passage number 99. On a second occasion, a consistent DNA fingerprint was obtained between pn 45 and pn 67. There was no evidence of cross contamination with other cell lines. VI . Conclusions At the present time, the PER.C6 cell line is the best substrate for the production of adenoviral vectors for gene therapy or vaccines. This conclusion is based on the ability to obtain good yields and safety considerations. The major safety considerations are the possibility of: i. the production of replication-competent adeno virus (RCA); ii. a tumorigenic risk from the transformed cell line; iii. the presence of abnormal prions; iv. contamination by adventitial agents. As described in this chapter, the lack of any overlap between the genome of the adenoviral vectors that carry the El deletion and the adenoviral El sequences carried in the PER.C6 cells makes homologous recombination impossible, thereby preventing the formation of RCA. It is well known that many transformed cell lines can produce tumors when injected into immunodeficent animals. As described, PER.C6 cells pro duce tumors in nude mice when 10^ cells are injected. They do not produce 1 6 0 Nichols et al. tumors, however, when 10^ or 10^ cells are injected. Since it is not anticipated that there will be any PER.C6 cells in a final product, this leaves the question of possible tumorigenicity of residual PER.C6 cellular DNA. Studies in nude mice and newborn hamsters in which DNA from PER.C6 cells was injected were negative for tumor production. The possibility of the presence of abnormal prions that could produce a neurodegenerative disease was also considered. This could occur if the PER.C6 cells had a mutation in a prion gene or if the cells were contaminated with abnormal prions such as in bovine spongioform encephalopathy. As far as possible, all serum and trypsin batches used from the time of origin of the culture were traced and no contact of serum from British sources was identified. The PER.C6 cell line was also adapted to serum-free suspension cultures. The prion protein gene of PER.C6 cells was sequenced and no mutations were found and the cell line was shown to be heterozygous for the 129 MA^ polymorphism. The cell line was also analyzed for the presence of abnormal prions at an early and late passage and an early passage of the HER parental line and none were found. In total, these studies indicate that the risk of a prion disease from the use of PER.C6 cells is vanishingly small. Finally, extensive studies for known and unknown adventitial agents have been documented and are negative. While there can be no absolute elimination of risk, this body of studies indicates a minimal, if any, risk from the use of this cell substrate for the production of adenoviral vectors. As new studies are developed they will also be apphed to ensure that no hazards are present. It has often been pointed out that a continuous cell line such as PER.C6 permits extensive analysis for adventitial agents and other safety concerns and thus is less hazardous than short-lived primary cell cultures for which testing must be repeated for each newly established culture. References 1. Russell, W. (2000). Update on adenovirus and its vectors./. Gen. Virol. 81, 2573-2604. 2. Rowe, W., Huebner, R., Gilmore, L., Parrott, R., and Ward, T. (1953). Isolation of a cytopathogenic agent from human adenoids undergoing spontaneous degeneration in tissue culture. Proc. Soc. Exp. Biol. Med. 84, 570-573. 3. Hilleman, M., and Werner, J. (1954). Recovery of new agents from patients v^ith acute respiratory illness. Proc. Soc. Exp. Biol. Med. 85, 183-188. 4. Hierholzer, J., Wigand, R., Anderson, L., Adrian, T., and Gold, J. (1988). Adenoviruses from patients with AIDS: A plethora of serotypes and a description of five new serotypes of subgenus D (types 43-47) . / . Infect. Dis. 158, 804-813. 5. Schnurr, D., and Dondero, M. E. (1993). Two new candidate adenovirus serotypes. Intervi- rology 36, 79-83. 6. Jong, J. C. de, Wermenbol, A. G., Verweij-Uijterwaal, M. W., Slaterus, K. W., Wertheim- van Dillen, P., Doornum, G. J., van, Khoo, S. H., and Hierholzer, J. C. (1999). Adenoviruses 6. Propagation of Adenoviral Vectors 1 6 1 from human immunodeficiency virus-infected individuals, including two strains that represent new candidate serotypes Ad50 and Ad51 of species Bl and D, respectively./. Clin. Microbiol. 37, 3940-3945. 7. Francki, R., Fauquet, C , Knudson, D., and Brown, F. (1991). Classification and nomencla ture of viruses. Fifth report of the international committee on taxonomy of viruses. Arch. Virol, suppl. 2, 140-144. 8. Horwitz, M. S. (1990) Adenoviridae and their replication. In "Virology" (B. N. Fields and D. M. Knipe, Eds.), pp. 1679-1740. Raven Press, New York. 9. Wold, W. S. M., Tollefson, A. E., and Hermiston, T. W. (1995) Strategies of immune modulation by adenoviruses. In "Viroceptors, Virokines and Related Immune Modulators Encoded by DNA Viruses" (G. McFadden, Ed.), pp. 147-185. Springer-Verlag, Heidelberg. 10. Wadell, G. (1984). Molecular epidemiology of adenoviruses. Curr. Top. Microbiol. Immunol. 110, 191-220. 11. Ishibashi, M., and Yasue, H. (1984). In "The Adenoviruses" (H. Ginsberg, Ed.), pp. 497-561. Plenum Press, London, New York. 12. Arnberg, N., Kidd, A., Edlund, K., Olfat, F., and Wadell, G. (2000). Initial interactions of subgenus D adenoviruses with A549 cellular receptors: Sialic acid versus av integrins. / . V/ro/. 74, 7691-7693. 13. Roelvink, P., Lizonova, A., Lee, J., Li, Y., Bergelson, J., Finberg, R., Brough, D., Kovesdi, L, and Wickham, T. (1998). The coxsackievirus-adenovirus receptor protein can function as a cellular attachment protein for adenovirus serotypes from subgroups A, C, D, E and F. / . Virol. 72, 7909-7915. 14. Bergelson, J., Cunningham, J., Droguett, G., Kurt-Jones, E., Krithivas, A., Hong, J., Horwitz, M., Crowell, R., and Finberg, R. (1997). Isolation of a common receptor for coxsackie B viruses and adenoviruses 2 and 5. Science 275, 1320-1323. 15. Greber, U. F., Willetts, M., Webster, P., and Helenius, A. (1993). Stepwise dismantling of adenovirus 2 during entry into cells. Cell 75, 477-486. 16. Vliet, P. van der (1995) Adenovirus DNA replication. In "The Molecular Repertoire of Adenoviruses" (W. Doerfler, and P. Bohm, Eds.), Vol. II. Springer Verlag, Berlin. 17. Ulfendahl, P. J., Linder, S., Kreivi, J. P., Nordqvist, K., Sevensson, C , Hultberg, H., and Akusjarvi, G. (1987). A novel adenovirus-2 El A mRNA encoding a protein with transcription activation properties. EMBO J. 6, 2037-2044. 18. Stephens, C , and Harlow, E. (1987). Differential splicing yields novel adenovirus 5 ElA mRNAs that encode 30 kd and 35 kd proteins. EMBO J. 6, 2027-2035. 19. Sarnow, P., Hearing, P., Anderson, C. W., Halbert, D. N., Shenk, T., and Levine, A. J. (1984). Adenovirus early region IB 58,000-dalton tumor antigen is physically associated with an early region 4 25,000-dalton protein in productively infected cells. / . Virol. 49, 692-700. 20. Rubenwolf, S., Schutt, M., Nevels, H., and Dobner, T. (1997). Structural analysis of the adenovirus type 5 ElB 55-kilodalton-E40RF6 protein complex./. Virol. 71, 1115-1123. 21. Debbas, M., and White, E. (1993). Wild-type p53 mediates apoptosis by ElA which is inhibited by ElB. Genes Dev. 7, 546. 22. Tarodi, B., Subramanian, T., and Chinnadurai, G. (1993). Functional similarity between adenovirus ElB 19 K gene and Bcl2 oncogene — mutant complementation and suppression of cell death induced by DNA damaging agents. Int. J. Oncol. 3, 467-472. 23. Telling, G. C , Perera, S., Szatkowski, O. M., and Williams, J. (1994). Absence of an essential regulatory influence of the adenovirus ElB 19-kilodalton protein on viral growth and early gene expression in human diploid WI38, HeLa, and A549 cells. / . Virol. 68, 541-547. 24. Leppard, K. N. (1997). E4 gene function in adenovirus, adenovirus vector and adeno- associated virus infections./. Gen. Virol. 78, 2131-2138. 25. Pombo, A., Ferreira, J., Bridge, E., and Carmo Fonseca, M. (1994). Adenovirus replication and transcription sites are spatially separated in the nucleus of infected cells. EMBO J. 13, 5075-5085. 1 6 2 Nichols ef al. 26. Dobner, T., Horikoshi, N., Rubenwolf, S., and Shenk, T. (1996). Blockage by adenovirus E4orf6 of transcriptional activation by the p53 tumor suppressor. Science 171, 1470-1473. 27. Moore, M., Horikoshi, N., and Shenk, T. (1996). Oncogenic potential of the adenovirus ORF6 protein. Froc. Natl. Acad. Set. USA 93, 11,295-11,303. 28. Muller, U., Kleinberger, T., and Shenk, T. (1992). Adenovirus E4orf4 protein reduces phosphorylation of c-Fos and El A proteins while simultaneously reducing the level of AP-1. / . Virol. 66, 5867-5878. 29. Shtrichman, R., and Kleinberger, T. (1998). Adenovirus type 5 E4 open reading frame 4 protein induces apoptosis in transformed cells./. Virol. 71, 2975-2982. 30. Johnson, D., Ohtani, K., and Nevins, J. (1994). Autoregulatory control of E2F1 expression in response to positive and negative regulators of cell cycle progression. Genes Dev. 8, 1514-1525. 31. Lubeck, M. D., Davis, A. R., Chengalvala, M., Natuk, R. J., Morin, J. E., Molnar, K. K., Mason, B. B., Bhat, B. M., Mizutani, S., Hung, P. P., and Purcell, H. (1989). Immunogenicity and efficacy testing in chimpanzees of an oral hepatitis B vaccine based on live recombinant adenovirus. Proc. Natl. Acad. Sci. USA 86, 6763-6767. 32. Torres, J. M., Alonso, C , Ortega, A., Mittal, S., Graham, F., and Enjuanes, L. (1996). Tropism of human adenovirus type 5-based vectors in swine and their ability to protect against transmissible gastroenteritis coronavirus. / . Virol. 70, 3770-3780. 33. Pacini, D. L., Dubovi, E. J., and Clyde, W. A. (1984). A new animal model for human respiratory tract disease due to adenovirus./. Infect. Dis. 150, 92-97. 34. Trentin, J., Yale, Y., and Taylor, G. (1962). The quest for human cancer viruses. Science 137,835-841. 35. Bernards, R., and Eb, A.J. van der (1984). Adenovirus: Transformation and oncogenicity. Biochim. Biophys. Acta 783, 187-204. 36. Flint, S. J. (1980) Transformation by adenoviruses. In "Molecular Biology of Tumor Viruses: DNA Tumor Viruses" (J. Tooze, Eds.), pp. 547-576. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 37. Freeman, A., Black, P. A. J. van der, Henry, P., Austin, J., and Huebner, R. (1967). Proc. Natl. Acad. Sci. USA 58, 1205-1212. 38. Graham, F. L., Eb, A.J. van der, and Heijneker, H. L. (1974). Size and location of the transforming region in human adenovirus type 5 DNA. Nature 251, 687-691. 39. Bernards, R., and Eb, A. J. van der. (1984). Adenovirus: Transformation and oncogenicity. Biochim. Biophys. Acta 783, 187-204. 40. Schrier, P., Bernards, R., Vaessen, R., Houweling, A., and Eb, A. J. van der. (1983). Expression of class I major histocompatibility antigens switched off by highly oncogenic adenovirus 12 in transformed rat cells. Nature 305, 771-775. 41. Graham, F. L., Smiley, J., Russell, W. C , and Nairn, R. (1977). Characteristics of a human cell line transformed by DNA from adenovirus type 5. / . Gen. Virol. 36, 59-72. 42. Whittaker, J. L., Byrd, P. J., Grand, R. J. A., and Gallimore, P. H. (1984). Isolation and characterization of four adenovirus type 12-transformed human embryo kidney lines. Mol. Cell Biol. 4, 110-116. 43. Gallimore, P. H., Grand, R. J. A., and Byrd, P. J. (1986). Transformation of human embryo retinoblasts with simian virus 40, adenovirus and ras oncogenes. Anticancer Res. 6, 499-467. 44. Vaessen, R. T. M. J., Houweling, A., Israel, A., Kourilsky, P., and Eb, A. J. van der. (1986). Adenovirus ElA-mediated regulation of class I MHC expression. EMBO J. 5, 335-341. 45. Fallaux, F. J., Kranenburg, O., Cramer, S. J., Houweling, A., Ormondt, H. van, and Hoeben, R. C , and A.J. van der. (1996). Characterization of 911: A new helper cell line for the titration and propagation of early region 1-deleted adenoviral vectors. Hum. Gene Ther. 7, 215-222. 6. Propagation of Adenoviral Vectors 1 6 3 46. Fallaux, F. J., Bout, A., Velde, I. v. d., Wollenberg, D. J. M. van den, Hehir, K., Kee- gan, J., Auger, C , Cramer, S. J., Ormondt, H. van, Eb, A. J. van der, Valerio, D., and Hoeben, R. C. (1998). New helper cells and matched early region 1-deleted adenovirus vectors prevent generation of replication-competent adenoviruses. Hum. Gene Ther. 9, 1909-1917. 47. Schiedner, G., Hertel, S., and Kochanek, S. (2000). Efficient transformation of primary human amniocytes by El functions of Ad5: Generation
of new^ cell lines for adenoviral vector production. Hum. Gene Ther. 11, 2105-2116. 48. Houv^eling, A., Elsen, P. v. d., and Eb, A. J. van der (1980). Partial transformation of primary rat cells by the left-most 4.5% of adenovirus 5 DNA. Virology 105, 537-550. 49. Teodoro, J., and Branton, P. (1997). Regulation of apoptosis by viral gene products. / . Virol. 71, 1739-1746. 50. Dyson, N., and Harlow, E. (1992). Adenovirus El A targets key regulators of cell prolifera tion. Cancer Surv. 12, 161-195. 51. Peeper, D. S., and Zantema, A. (1993). Adenovirus-EIA proteins transform cells by seques tering regulatory proteins. Mol. Biol. Rep. 17, 197-207. 52. Peeper, D. S., A.J. van der, and Zantema, A. (1994). The Gl/S cell-cycle checkpoint in eukaryotic cells. Biochim. Biophys. Acta 1198, 215-230. 53. White, E., Cipriani, R., Sabbatini, P., and Denton, A. (1991). Adenovirus ElB 19-kilodalton protein overcomes the cytotoxicity of ElA proteins. / . Virol. 65, 2968. 54. Bout, A. (1999). Towards safe and effective adenoviral vectors for human gene therapy. Eur. Biopharm. Rev. December, 94-99. 55. Imler, J. L., Chartier, C , Dreyer, D., Dieterle, A., Sainte-Marie, M., Faure, T., Pavirani, A., and Mehtah, M. (1996). Novel complementation cell lines derived from human lung carcinoma A549 cells support the growth of El-deleted adenovirus vectors. Gene Ther. 3, 75-84. 56. Brough, D. E., Lizonova, A., and Kovesdi, I. (1996). Stable cell lines for complementation of adenovirus early regions El , E2A and E4. In "Abstract Book CSH Conference on Gene Therapy," p. 42. 57. Louis, N., Evelegh, C , and Graham, F. L. (1997). Cloning and sequencing of the cellular- viral junctions from the human adenovirus type 5 transformed 293 cell line. Virology 233, 423-429. 58. Hehir, K. M., Armentano, D., Cardoza, L. M., Choquette, T. L., Berthelette, P. B., White, G. A., Couture, L. A., Everton, M. B., Keegan, J., Martin, J. M., Pratt, D. A., Smith, M. P., Smith, A. E., and Wadsworth, S. C. (1996). Molecular characterization of replication- competent variants of adenovirus vectors and genome modifications to prevent their occurrence./. Virol. 70, 8459-8467. 59. Lochmiiller, H., Jani, A., Huard, J., Prescott, S., Simoneau, M., Massie, B., Karpati, G., and Acsadi, G. (1994). Emergence of early region 1-containing replication-competent adenovirus in stocks of replication-defective adenovirus recombinants (DEI + DE3) during multiple passages in 293 cells. Hum. Gene Ther. 5, 1485-1492. 60. Zhu, J., Grace, M., Casale, J., Chang, A., Musco, M., Bordens, R., Greenberg, R., Schaefer, E., and Indelicato, S. (1999). Characterization of replication-competent adenovirus isolates from large-scale production of a recombinant adenoviral vector. Hum. Gene Ther. 10, 113-121. 61. Imler, J. L., Bout, A., Dreyer, D., Dieterle, A., Schultz, H., Valerio, D., MehtaU, M., and Pavirani, A. (1995). Trans-complementation of El-deleted adenovirus: A new vector to reduce the possibility of codissemination of wild-type and recombinant adenoviruses. Hum. Gene Ther. 6,711-721. 61. Hermens, W. T. J. M. C , and Verhaagen, J. (1997). Adenoviral-vector-mediated gene expression in the nervous system of immunocompetent Wistar and T cell-deficient nude 1 6 4 Nichols ef al. rats: Preferential survival of transduced astroglial cells in nude rats. Hum. Gene Ther. 8, 1049-1063. 63. Ormondt, H. v., and Galibert, F. (1984) Nucleotide sequences of adenovirus DNAs. In "Current Topics in Microbiology and Immunology," Vol. 110, pp. 73-142. Springer-Verlag, Heidelberg. 64. Vales, L. D., and Darnell, J.J. (1989). Promoter occlusion prevents transcription of aden ovirus polypeptide IX mRNA until after DNA replication. Genes Dev. 3, 49-59. 65. Singer-Sam, J., Keith, D. H., Tani, K., Simmer, R. L., Shively, L., Lindsay, S., Yoshida, A., and Riggs, A. D. (1984). Sequence of the promoter region of the gene for X-hnked 3-phosphoglycerate kinase. Gene 32, 409-417. GG. Valerio, D., Duyvesteyn, M. G. C , Dekker, B. M. M., Weeda, G., Bervens, T. M., Voorn, L. van der, Ormondt, H. van, and Eb, A. J. van der (1985). Adenosine deaminase: charac terization and expression of a gene w îth a remarkable promoter. EMBO J. 4, 437-443. 67. Simonsen, C., and Levinson, A. (1983). Analysis of processing and polyadenylation signals of the hepatitis B Virus surface antigen gene by using simian virus 40-Hepatitis B Virus chimeric plasmids. MoL Cell Biol. 3, 2250-2258. 68. Roberts, B. E., Miller, J. S., Kimelman, D., Cepko, C. L., Lemischka, I. R., and Mulligan, R. C. (1985). Individual adenovirus type 5 early region lA products elicit distinct alterations of cellular morphology and gene expression. /. Virol. 56, 404-413. 69. Albert, D., Rabson, A., and Dalton, A. (1968). In vitro neoplastic transformation of uveal and retinal tissues by oncogenic DNA viruses. Invest. Ophthalmol. 7, 357-365. 70. Mukai, N., Kalter, S., Cummins, L., Mathews, V., Nishida, T., and Nakajima, T. (1980). Retinal tumor induction in the baboon by human adenovirus 12. Science 210, 1023-1025. 71. Fallaux, F. J., Kranenburg, O., Cramer, S. J., Houw^ehng, A., Ormondt, H. v., Hoeben, R. C , and Eb, A. J. van der (1996). Characterization of 911: A new helper cell line for the titration and propagation of early-region-1-deleted adenoviral vectors. Hum. Gene Ther. 7, 215-222. 72. Byrd, P., Brown, K., and Gallimore, P. (1982). Malignant transformation of human embryo retinoblasts by cloned adenovirus 12 DNA. Nature 298, 69 -71 . 73. Bett, A. J., Haddara, W., Prevec, L., and Graham, F. L. (1994). An efficient and flexible system for construction of adenovirus vectors with insertions or deletions in early regions 1 and 3. Proc. Natl. Acad. Sci. USA 91, 8802-8806. 74. Blanche, F., Cameron, B., Barbot, A., Ferrero, L., Guillemin, T., Guyot, S., Somarriba, S., and Bisch, D. (2000). An improved anion-exchange HPLC method for the detection and purification of adenoviral particles. Gene Ther. 7, 1055-1062. 75. Sadoff, D., Jr, W. G., DiGiacomo, R., and Vogel, A. (1988). Neoplasms in NIH Type II athymic (nude) mice. Lab. Anim. Sci. 38, 407-412. 76. Furmanski, P., and Rich, M. (1982) Neoplasms of the hematopoietic system. In "The Mouse in Biomedical Research" {H. Foster, J. Small, and J. Fox, Eds.), pp. 351-371. Academic Press, London. 77. Gershwin, M., Yoshiyuki, P., Castles, J., Ikeda, R., and Ruebner, B. (1983). Anti-\|JL induces lymphoma in germfree congenitally athymic (nude) but not heterozygous (nu/+) mice. / . Immunol. 131, 2069-2073. 78. Frith, C , Ward, J., Frederickson, T., and Harleman, J. (1996) Neoplastic lesions of the hematopoietic system. In "Pathobiology of the Aging Mouse" (U. Mohr, D. Dungsworth, C. Capen, W. Carlton, J. Sundberg, and J. Ward, Eds.), pp. 219-235. ILSI Press, Washington DC. 79. Quinlan, M. P. (1993). Expression of antisense ElA in 293 cells results in altered cell morphologies and cessation of proliferation. Oncogene 8, 257-265. 6. Propagation of Adenoviral Vectors 1 6 5 80. WHO Report (1987). Acceptability of cell substrates for production of biologicals. "Report of WHO Study Group", WHO Technical Report Series 747, tables 1 and 2. World Health Organization, Geneva. 81. WHO Report (1998). WHO requirements for the use of animal cells as in vitro substrates for the production of biologicals. In Requirements for Biological Substances, No. 50 [Reproduced from WHO Technical Report Series 878]. 82. Grachev, V. Advances in biotechnical processes. In "Viral Vaccines" (A. Mizraki, Ed.), Vol. 14. Wiley-Liss, New York. 83. Mufson, R., and Gesner, T. (1985). Lack of tumorigenicity of cellular DNA and oncogene DNA in nev^born Chinese Hamsters. In "Abnormal Cells, New Products and Risk" (H. Hopps, and J. Petricciani, Eds.) pp. 168-169. Tissue Culture Association, Gaithersburg, Maryland. 84. Levinson, A., Svedersky, L., and Jr, M. P. (1985) Tumorigenic potential of DNA derived form mammalian cell lines. In "Abnormal Cells, New Producst and Risks" (H. Hopps, and J. Petricciani, Eds.), pp. 161-165. 85. Palladino, M., Levinson, A., Svedersky, L., and Objeski, J. (1987). Safety issues related to the use of recombinant DNA-derived cell culture products. I Cellular components. Dev. Biol. Standard 66, 13-22. 86. Wierenga, D., Cogan, J., and Petricciani, J. (1995). Administration of tumor cell chromatin to immunosuppressed and non-immunosuppressed non-human primates. Biologicals 23, 221-224. 87. Lower, J. (1990). Risk of tumor induction in vivo by residual cellular DNA: Quantitative considerations./. Med. Virol. 31, 50-53 . 88. Prusiner, S. (1982). Novel proteinaceous infectious particle causes scrapie. Science 216, 136-144. 89. Prusiner, S. (1998). Prions. Proc. Natl. Acad. Sci. USA 95, 13,363-13,383. 90. Oesch, B., Westaway, D., Walchh, M., McKinley, M., Kent, S., Aebersold, R., Barry, R., Tempst, P., Teplow, D., Hood, L., Prusiner, S., and Weissmann, C. (1985). A cellular gene encodes scrapie PrP 27-30 protein. Cell 40, 735-746. 91. Ironside, J. (1998). Prion diseases in man. / . Pathol. 186, 227-234. 92. Will, R., Ironside, J., Zeidler, M., Cousens, S., Estibeiro, K., Alperovitch, A., Poser, S., Pocchiari, M., Hofman, A., and Smith, P. (1996). A new variant of Creutzfeldt-Jakob disease in the U.K. Lancet 347, 921-925. 93. Bruce, M., Will, R., Ironside, J., McConnell, I., Drummond, D., Suttie, A., McCardle, L., Chree, A., Hope, J., Birkett, C , Cousens, S., Eraser, H., and Bostock, C. (1997). Transmissions to mice indicate that 'new variant' CJD is caused by the BSE agent. Nature 389,498-501. 94. CoUinge, J., Sidle, K., Meads, J., Ironside, J., and Hill, A. (1996). Molecular analysis of prion strain variation and the aetiology of "new variant" CJD. Nature 383, 685-690. 95. Palmer, M., Dryden, A., Hughes, J., and CoUinge, J. (1991). Homozygous prion protein genotype predisposes to sporadic Creutzfeld-Jakob disease. Nature 352, 340-342. 96. CoUinge, J., Palmer, M., and Dryden, A. (1991). Genetic predisposition to iatrogenic Creutzfeld-Jakob disease. Lancet 337, 1441-1442. 97. Parchi, P., Castellani, R., CapeUari, S., Ghetti, B., Young, K., Chen, S., Farlow, M., Dickson, D., Sima, A., Trojanowski, J., Petersen, R., and Gambetti, P. (1996). Molecular basis of phenotypic variability in sporadic Creutzfeldt-Jakob disease. Ann. Neurol. 39, 767-778. 98. Chesebro, B., Race, R., Wehrly, K., Nishio, J., Bloom, M., Lechner, D., Bergstrom, S., Robbins, K., Mayer, L., Keith, J., Garon, C , and Haase, A. (1985). Identification of scrapie prion-protein mRNA in scrapie-infected and uninfected brain. Nature 315, 331-335. 99. Meyer, R., McKinley, M., Bowman, K., Braunfeld, M., Barry, R., and Prusiner, S. (1986). Separation and properties of cellular and scrapie prion protein. Proc. Natl. Acad. Sci. USA 83,2310-2314. 1 6 6 Nichols ef al. 100. Caughey, B., Race, R., and Chesebro, B. (1988). Detection of prion protein mRNA in normal and scrapie-infected tissues and cell lines./. Gen. Virol. 69, 711-716. 101. Tanke, H. J., Wiegant, J., van Gijlswijk, R. P., Bezrookove, V., Pattenier, H., Heetebrij, R. J., Talman, E. G., Raap, A. K., and Vrolijk, J. (1999). New strategy for multi-colour fluorescence in situ hybridisation: COBRA: COmbined Binary RAtio labelling. Eur. J. Hum. Genet. 7 , 2 - 1 1 . 102. Wiegant, J., Brouwer, K., Bezrookove, V., Raap, A., Bout, A., and Tanke, H. "Molecular Cytogenetic Characterization of PER.C6 Cells". In preparation. C H A P T E R Purification of Adenovirus Paul Shabram Gary Vellekamp Canji Inc. Schering-Plough Research Institute San Diego, California Kenilworth, New Jersey Carl Scandella Carl Scandella Consulting Bellevue, Washington I. Introduction Since the late 1950s, adenoviruses have been purified using classical methods of density gradient ultracentrifugation. These methods w êre efficient and could supply the quantities of highly purified viral particles necessary for research. The need for larger quantities has arisen v^ith the advent of the use of adenoviral vectors for gene therapy trials. In this chapter, ŵ e discuss the techniques for extracting adenoviral particles from a complex milieu. Selecting the best technique for purification requires an understanding of the physical nature of the particles as well as the nature of contaminants. Knov^ledge of these properties is essential for developing a purification process that is sufficient to supply the commercial market for a therapeutic adenovirus. A. The Physical Characteristics of the Adenovirus Particle in Solution There are numerous adenoviruses that possess specific tropism for many species of animals including human, bovine, ovine, equine, canine, porcine, murine, and simian adenovirus subgenera. Although many of these aden oviruses are capable of delivering a transgene to human tissues, the development of clinical adenoviral agents most often employs human adenoviral vectors derived from human adenovirus serotypes 2 or 5. Consequently, for the pur poses of this discussion, the information given here refers to human adenovirus types 2 and 5. ADENOVIRAL VECTORS FOR GENE THERAPY 1 £.y Copyright 2002, Elsevier Science (USA). All rights reserved. 168 Shabram ef a/. 1. Particle Size Size and shape are key factors involved in purifying any macromolecule. These factors are
equally applicable to adenoviruses, w^hich are much larger than most biomolecules commonly purified. Adenoviruses are icosahedral in shape v^ith fiber-like extensions from each of the 12 vertices. The adenovirus is composed of DNA, protein, and carbohydrate. The viral DNA is packaged in a highly organized protein coat termed the "capsid." Negative staining electron microscopy of the adenovirus capsid was used to estimate a diameter of 73 nm along a fivefold symmetry axis with a vertex-to-vertex diameter of about 83 nm [1]. Freeze-fracture studies demonstrate a slightly larger capsid diameter. Fibers extending from each of 12 vertices increase vertex-to-vertex diameter by about 40 nm for human adenovirus serotype 5. Oliver et al. [2] employed photon correlation spectroscopy to characterize the adenovirus type 5 particles in solution, reporting a molecular w^eight of 167 x 10^ Da and a corresponding particle diameter of 98 nm. 2. Diffusion of Adenovirus Particles The adenovirus particle diffuses very slow l̂y in solution. The diffusion coefficient for Ad-5 is 4.46 x 10~^^ m^/s in serum containing media at 37°C [2, 3]. Figure 1 compares the diffusion constants of some w^ell-knov^n 10^ • Sucrose* ^ 101 Insulin* f s • Horse Hemoblobin* • Urease* Q 1011 Tobacco Mosaic Virus* • Adenovirus Type 5** • 1012 —^ —r- —T" 102 103 104 105 106 107 108 109 Molecular Weight Figure 1 The diffusion coefficients of macromolecules are related to their sizes. Adenoviral particles diffuse very slov/ly in solution. This slow diffusion rate affects mixing, separation, and analytical methods. 7. Purification of Adenovirus 1 6 9 macromolecules to adenovirus. The large size and corresponding slow diffusion of adenovirus particles in solution require consideration in mixing because, given the density of the particle and its SIOMA rate of diffusion, Brovs^nian motion cannot be relied upon to disperse the particle. If left alone the particles in solu tion w^ould take v^eeks to reach equilibrium. Hov^ever, gravity w îll intervene and cause the particles to sediment to the bottom of the container. Conse quently, adenovirus solutions require greater agitation than protein solutions to disperse the particles evenly. The slov^ diffusion rate also complicates analytical methods. For example, the interaction betv^^een the virus and a cell takes much longer than protein-cell interactions because of the SIOVSA diffusion rate. Typical biological methods for quantifying particles depend on Brow^nian motion for bringing about virus-cell interaction. Without accounting for this slow diffusion, the titer of the material tested may be underestimated. This will be discussed in more detail below (see section III.B.2 of this chapter). 3. Capsid Surface The surface of the adenovirus capsid is of particular interest when select ing a separation technique since there are many binding methods available. The adenovirus capsid consists of 252 capsomeres. Two hundred and forty of these capsomeres are hexons and 12 are pentons. Hexons are trimers of protein II and are the main structural component of the capsid. Pentons, constructed from penton base and fiber proteins, are prominent features as these structures protrude and add to the hydrodynamic radius of the virus particle. In addi tion, protein Ilia and protein IX may also contribute to surface characteristics. Protein Ilia is essential and two of these proteins are found at the junction joining adjacent facets together like stitching. Protein IX is not essential for capsid assembly but enhances the stability of the virus at higher temperatures. Four protein IX trimers stabilize a group of nine hexons (ninemers) that have assembled into a facet [4]. The hexon, which comprises about 50% of the total virion protein, dominates the charge characteristics of the particle. Hexon capsomeres possess an isoelectric point (pi) near pH 6 [1]. At physiological pH, the capsomere would be expected to bear a negative charge. The complete adenovirus structure would also be expected to show a net anionic surface charge under physiological conditions. It is generally advisable to avoid exposing the proteins to solutions at or near the pi because many proteins change conformation, degrade, or precipitate when titrated through it. Ninemers, hexons, and complete virions precipitate from solution when titrating the solution from pH 7 to pH 5 [5, 6]. If the particle is titrated through the pi rapidly and further lowered, losses occur but active particles can be recovered from solutions such as acetic acid. The virus is also stable up to around pH 8; the exact threshold being dependent upon the composition of 1 7 0 Shabram ef al. the solution. Exposing particles to a pH greater than 8 generally leads to a loss of activity and particle disruption. 4. Hydration of the Adenovirus Particle The degree of hydration of the particle is an important consideration for both purification and stability of the adenovirus. Sedimentation by ultra- centrifugation using Schlieren optics [2] suggested that the adenovirus particle contains a "hard core" (w^ater excluded) w îth a diameter of 76 nm, similar to that obtained using negatively stained electron microscopy. The difference between the hard core diameter of 76 nm and the light scattering diameter of 98 nm is significant because it suggests that the particle contains w^ater. The amount of v^ater represented by the difference suggests that the virus particle contains 2.3 g of w^ater for every gram of virus [2]. These observations are consistent w îth measurements of other viruses [7, 8]. While proteins are stabilized by the incorporation of a feŵ w^ater molecules ('Vaters of hydration"), the amount of w^ater suggested by these studies is more than 21 million w^ater molecules per virion. This amount far exceeds water typically bound to proteins. This degree of hydration corre sponds to a theoretical density calculation of about 1.4 g/mL, which is very close to the observed density of 1.34 g/mL. The additional water should not be surprising since the virus exists in an aqueous environment. With these data in mind, however, the particle would be expected to show unique properties. The adenovirus particle is generally considered rigid. However, the degree of hydration suggests that a certain amount of flexibility should be considered. Since the particle is not encapsulated in a membrane small ions may have ready access to the core (see section IV). Changes in ionic strength may induce conformational changes in the capsid proteins; some of these may be beneficial for separation and some may be catastrophic. The particle may also be sensitive to rapid changes in salt concentrations. One might also predict that hydrophobic solvents should be avoided. B. Features of the Milieu Effective virus purification capitalizes on the differences between the physical properties of the adenovirus relative to the components of the mixture from which it is being isolated. The exact composition of the milieu varies with the cell culture process, and, to a lesser extent, every batch. In general, the large-scale purification of adenovirus requires the isolation of the virus from infected cell lysate taken from a bioreactor. This mixture consists of a formulated medium usually containing bovine serum, and less frequently antifoaming agents, or anti-clumping agents (pluronics). Significant amounts of additives, however, present difficult challenges for any recovery procedure. Efficient large-scale production requires high cell densities which in turn require 7. Purification of Adenovirus 1 7 1 high gas exchange rates. This can cause severe foaming and necessitate the addition of agents to control it. Other additives, such as anti-clumping agents and hpids, adapt the media for large-scale cell culture. Cell lysis, necessary to release the adenovirus from the host cell, results in the additional release of DNA, protein, lipids, carbohydrates, and other cellular components. Culture conditions, media components, cell derived contaminants, and additives may have a significant impact on dow^nstream processing. 1. Culture Conditions Adenoviruses are produced by infection of cell lines in culture w îth a viral seed stock. The particular cell line used requires a highly developed cell culture method to achieve maximum yield. Flat stock culture, although useful for small-scale work, is generally not sufficient for larger scale applications. Many of the cell lines used in flat stock culture have resisted attempts to adapt them to the suspension and serum free conditions preferred for large-scale processes. A compromise is struck by the culture of attachment-dependent cells using microcarriers in a bioreactor. This microcarrier-based process introduces yet another component that must be separated from the adenovirus. Similarly, if serum is utilized, it will be necessary to consider the effective removal of its components. 2. Construct Induced Contaminants and Considerations The viral construct may also contribute to the milieu as the viral DNA backbone may lead to the contribution of many more contaminants. For down stream purification, higher titers favor better recovery and cleaner preparations because recovery and purification is an enrichment process. Even with max imum productivity, however, adenoviral particles represent a small fraction of molecular entities produced by the end of the culture process. Therefore, factors affecting the end titer can also affect the process. The majority of adenoviral vectors for gene therapy are serotype 5 and have been rendered deficient for replication in most cells. With the exception of replication-competent adenoviruses, most vectors have been crippled to eliminate their replication in normal human cells. In general, when compared to wild-type virus, deletions or mutations in the early genes tend to attenuate viral replication in all cell lines. Attenuation for replication is typically achieved by large deletions in the immediate-early region El . These vectors require specialized packaging cell lines for efficient production. Cell lines such as HEK 293 [9] or PER.C6 [10] have been transformed with adenoviral DNA and provide sufficient El function in trans to enable replication. In addition to El deletions, many vectors possess deletions for much of the E3 region. The deleted E3 genes are considered nonessential for viral replication and these deletions allow for larger transgene packaging capacity. Other deletions have been made to reduce the frequency of recombination 1 7 2 Shabram ef o/. during culture. So-called "second generation" viruses may have additional early gene deletions (e.g., in E4) as w êll as a deletion of protein IX encoding sequences [11]. Elimination of certain essential genes from the virus requires that the cell line be able to complement these protein functions in trans to package the virus [12]. A significant unw^anted by-product of adenoviral replication is DNA. Wild-type human adenoviruses are able to replicate in a variety of both quiescent and proliferating human cells due to the function of adenoviral immediate-early genes. El a proteins can be observed within 1 h after infection, as cellular transcription factors are sufficient to transcribe the El a genes. El expression initiates the adenoviral life cycle by altering the cell cycle machinery to induce cellular DNA replication even in quiescent cells. Viral and cellular proteins activate subsequent viral transcription. New^ copies of viral DNA are synthesized and viral production proceeds in a replication cascade. By the end of viral DNA replication, a large amount of DNA is present in the infected cell. The purpose of a gene therapy vector is to convey a therapeutic effect by the delivery and expression of therapeutic genes. Many of these trans- genes have a significant effect on the cells and adenoviral life cycle. Some genes, such as the retinoblastoma protein and the cyclin-dependent kinase inhibitor p21 directly affect the levels of activated E2F. E2F is a cellular transcription factor that is necessary for the transactivation of the adenovi ral E2 promoter and thus the expression of viral DNA replication proteins. Other transgene products, such as proapoptotic proteins, can overcome the adenoviral block to apoptosis leading to early cell death and can interfere v^ith adenoviral particle assembly. Secreted pleiotropic transgene products, such as growth factors, can trigger undesired effects in the packaging cell and severely attenuate production. Before adenoviral DNA can be coated with viral core proteins, the DNA is available for transcription. In this way, some adenoviral genes, especially those encoding capsid components, are not expressed until DNA replication occurs [13]. During this phase, the expression of transgene product is enhanced by the replication cycle itself by expanding the copy number of the transgene with each copy becoming available for transcrip tion. Strong exogenous promoters may also sequester transcription factors and the cellular protein synthesis machinery can become clogged with transgene expression leading to attenuated adenoviral protein production. If the sequence of events leading to particle maturation is disturbed, even by an imbalance in protein production, large
quantities of viral proteins, incomplete particle assemblies, transgene products, abnormal cellular structures, and in some cases extreme amounts of extracellular proteins can be added to the milieu. These complications can significantly impede purification and add to the analytical requirements. 7. Purification of Adenovirus 1 7 3 C. Summary of Characteristics The attributes of the adenoviral particle, the culture process, components of the media and properties of the vector itself have significant impacts on the design of a purification process. Table I outlines the salient aspects of the particle and the lysate. II. Recovery and Purification of Adenoviral Particles Purification must take place in the context of a complicated lysate. Table II summarizes the key features of the adenovirus particle and suggests recovery techniques that may be employed. Together, the features of the particle and the milieu point to a sequence of process steps that yield purified adenovirus (Table I). Since the particles are produced in cells, the first recovery step is harvesting of the infected cells. The next step requires lysis of the cell to release the virus. The cell lysate contains cell debris so a clarification step is necessary to protect dov^nstream steps. A substantial amount of DNA is present and must be eliminated early in the process. The clarified lysate is too crude for high-resolution purification so an initial purification is needed. Once the preparation has been simplified by the initial purification, a fine separation step removes the remaining contamination. The purification may utilize salts or buffers that are undesirable for use in the clinic. These components may need to be exchanged for a final formulation. The follow^ing sections provide techniques to accomplish this sequence. A. Harvest Methods Cells grow^n and infected in large-scale flat stock culture eventually detach from the surface. Alternatively, trypsin may be used to detach cells before the onset of cytopathic effect (CPE). Cells free in the medium may be collected by centrifugation or filtration. Cells grow^n in suspension are also harvested in the same manner. Cells grow^n on microcarriers may be harvested by allow^ing the cells to settle so the spent medium can be decanted. Infected cells may be removed from the microcarriers by trypsinization or processed v^hile still on the carrier. Infected cells remain suspended and therefore can be decanted with the spent medium. In order to maximize the harvest yield, one should consider the point at vŝ hich harvest occurs. The life cycle of the adenovirus vŝ as thought to terminate at the time cytopathic effect (CPE) is observed. Analysis using AEHPLC (section III.F of this chapter), how^ever, demonstrated that the peak 174 Shabram ef aL Table I Properties of Adenoviral Particles Property Data Consideration Issues Diffusion 4.46 X 10-12 j^2/5 Filtration, • Slow diffusion rate may centrifugation. provide a means for chromatography, separation on filters and freezing, and chromatography resins. thawing • Centrifugation not counter acted by diffusion. Mixing will be problematic especially during freeze-thaw operations. • Slow diffusion may decrease concentration dependent aggregation. • Assays are hindered. Water High compared to Fragility, salt • Particle may be swelled with content proteins concentration, water. small ions, solvents. • High salt and solvents may freezing, and result in degradation. thawing • Sensitivity to shearing forces. Viral genome Viral, cellular, and Early gene alterations • Cell lysate may contain many alterations transgene and deletions incorrectly assembled expression overexpression of particles, adenoviral protein transgene products structures, and unusual cellular structures that are similar in size to adenoviral particles. Transgene product may be a contaminant which could confound potency assays. High titers result in less difficult purification Lysate Complex Contaminants Particle assay for crude materials. DNA, lipids, BSA, antifoaming agents, anti-clumping agents, and other contaminants may bind to the particle and co- purify or foul filters and resins. of particle production occurs before cells begin to detach. Figure 2 shows the particle count taken at various time points during a bioreactor process. In most bioreactor runs the particle concentration drops slightly from the peak, but in some cases the particle concentration falls to as low as 10% of the peak value 7. Purification of Adenovirus 175 Table II Physical Properties of Adenovirus Particles That Can Be Exploited for Purification Property Data Method Issues Density 1.34 g/mL Density gradient Classical method — small scale. uhra-centrifugation Size ~100 nm Fihration Some cell debris similar size, lipids, and cell culture additives may interfere. Size exclusion Particle larger than resin pore chromatography sizes. Could be used for buffer exchange. Surface Protein: ionic, Ion exchange, Scaleable methods employed for hydrophobic, hydrophobic protein purification. specific surface interaction, affinity Well established literature to chemistry chromatography predict behavior of cell culture hgand binding components. Particle size larger fihration than pore sizes. Aggregation may occur. Reversed-phase Solvents may be problematic. chromatography, solvent extraction in just a few hours. This phenomenon underscores the need to monitor the process carefully to obtain maximum yield. B. Lysis Methods Following recovery of the infected cells, the adenovirus must be released from the cell. This is accomplished by lysing the cell. There are many methods available for both large- and small-scale cell lysis. The most useful of these methods are discussed below. 1. Freeze-Thaw Cells burdened with a full viral load are fragile and easily disrupted. Freezing and thawing the infected cells achieves the release of virus. This method is attractive at small scale because it does not require specialized equipment. It is less attractive at large scale because the freezing and thawing of a large sample is difficult to control. Consequently other methods are preferred for large scale. The freeze-thaw lysis method requires that one observe how the solution freezes and how it thaws. During the freezing process, solutes, such as salts, proteins, and free viral particles, depress the freezing point of the solution. 176 S h a b r a m ef al. 40000 35000 30000 25000 C/3 20000 u 15000 10000 H 5000 A Hours Post Infection Figure 2 Adenoviral particle production can be followed using anion-exchange high-performance liquid chromatography (HPLC) as discussed in section III.F. The per cell productivity was monitored during a 5 L bioreactor run with a replication deficient type 5 adenovirus grown in HEK 293 cells. The completion of particle production occurs before obvious signs of cytopathic effects are manifested. Small ice crystals of pure water begin to form. The solutes excluded from the ice tend to concentrate in spaces between the ice crystals. These areas of concentrated solutes experience freezing point depression. If the freezing process is too slow, the virus may be found in highly concentrated bands. Once the thawing process begins, low-molecular-weight solutes are free to diffuse away rapidly, but adenovirus particles remain roughly in the same place, owing to their large size and slow diffusion constant. Given that nearly all proteins precipitate from solution at a critical concentration, one would expect that the virus particle would also be similarly limited. At temperatures above 0°C, frequent collisions among particles lead to an aggregation cascade. While freeze-thaw releases more than 90% of the virus in three cycles under favorable conditions, improper control may lead to greater than 50% loss. Consideration for damage and loss of the particles represents the greatest concern. Damage may not be obvious at first when a structure as large as a virus is involved; rather, damage suffered during early recovery steps may manifest itself as reduced stability of the purified virus. Particles are packed into the cell in a tight array; once released the viral particle can aggregate. Collisions leading to aggregation in the cell may be limited by mediating proteins, which impede movement and hold the particle in the soluble array. 7 . Purif ication of Adenov i rus 177 The control of pH and salt concentration is also critical for freeze-thaw. Some buffers such as phosphate do not maintain buffering capacity when the solution freezes. This may be because sodium phosphates precipitate at low temperature. 2. Homogenization A useful method for cell disruption used for recovery of recombinant proteins involves passage of bacterial or yeast cells through a small orifice under pressures up to 20,000 psi. Upon passage through the orifice, the cells expand and rupture as they experience a sudden drop in pressure. The French press has been used for this purpose at laboratory scale for many years [14]. Large-scale processing using this principle may be accomplished using a Gaulin homogenizer (APV Gaulin, Wilmington, MA). Lysis of infected mammalian cells can also be achieved using a similar device. However, at the pressures under which this method is normally used to rupture bacterial cells viral loss is observed. To protect the virus the pressures must be minimized. Figure 3 illustrates this affect by comparing the recovery of adenovirus particles after lysis at different fluidization pressures in a microfluidizer. \fV - • • 80 - u > • O QJ 60 - Pi u 40 - • PH 20 - • 0 - 1 1 1 1 ! 1000 2000 3000 4000 5000 6000 Microfluidizer Pressure (psi) Figure 3 Methods to lyse infected cells vary from repeated cycles of freezing and thawing to microfluidization. The cell can be lysed by a sudden pressure drop generated by several different means. This experiment used anion-exchange HPLC to measure the recovery of particles over a range of pressures in a microfluidizer. Samples were also submitted for titer and showed the same trend. Particle recovery declines at differential pressures above 600 psi. Similar results were obtained using a nitrogen bomb apparatus (data not shown). 178 Shabram ef al. 3. Sonication Sonication is widely used for breaking small quantities of cells for research [14] because it is rapid and convenient for small samples. This method results in excellent release of virus from infected cells. Disadvantages of this method include the generation of heat, production of free radicals, and attendant chemical damage, as well as the lack of equipment suitable for large-scale applications. 4. Simultaneous Harvest and Lysis Using a Continuous Flow Centrifuge Bacterial and yeast cells can be readily harvested at large scale by the use of continuous flow centrifuges. The use of these devices to harvest intact mammalian cells is less certain. These centrifuges are configured to concentrate and harvest cells by different means. To aid in the separation these centrifuges often are fitted with a series of conical discs inside the rotor. All systems expose the cells and liquid to a centrifugal field allowing the cells to be concentrated in the interior of a hollow rotor. Supernatant, which has a lower buoyant density, flows out of the rotor. The "pellet," which remains liquid, either collects in the rotor, exits through restrictions on the outer edge of the rotor, or is ejected in a discharge cycle as the rotor opens. This third method is to briefly open the rotor on the outer edge while it is spinning. If the rotor is opened for just a fraction of a second, the pellet is discharged and little supernatant is lost. Generally, these centrifuges develop between 10,000 and 20,000 g. Fragile cells, such as mammalian cells, are particularly at risk of lysis in both the discharge and nozzle-type disc stack machines because of the shear forces and rapid pressure changes. For the same reason these systems are ideal for large-scale concentration and lysis of infected cells [15]. Shear from the rotating discs provide some cellular disruption and the rapid pressure drop from discharge finishes the job. The resulting lysate may be further clarified. 5. Lysis during Filtration Another method used for lysis is cross-flow filtration. Cross-flow filtra tion, also known as tangential-flow filtration, separates particles in solution by passing the solution along the surface of a membrane. Liquid passes through the membrane because of the pressure differential across the membrane. Par ticles and solutes are retained if they are larger than the "cutoff size" of the membrane. Cross-flow systems are configured to allow for recirculation of material along the membrane surface. In this way, larger components are retained (retentate) and smaller components are collected in the passed liq uid (permeate). Membranes may be configured in flat plate, spiral wound, or hollow fiber systems. Cutoff sizes vary from particles visible to the eye down to molecules as small as 300 molecular weight. A typical clarification operation can be achieved by using either a 0.45- or a 0.2-^im nominal cutoff 7.
Purification of Adenovirus 1 7 9 size. Infected cells in this system are exposed to both shear and rapid pressure drops. The advantage of these systems is that the viral particles, w^hich are slightly less than 0.1 \|xm in diameter, are not allowed to traverse the filter until they have been released from the cell. Cellular debris is retained and therefore separated from the virus. C. Clarification The lysis procedure releases the virus from the packed array inside the cell into the medium. All procedures produce cell debris, which must be removed before purification. A common procedure for lysis includes concentration of the intact cells followed by lysis. This method works well at the laboratory scale and may be necessary for production with viral vectors that exhibit poor per cell productivity. With optimized culture methods overconcentration becomes a concern. Highly concentrated lysates exhibit significant losses during lysis and clarification. The loss seems to be associated with aggregation of the virus particle. 1. Centrifugation Centrifugation at low relative centrifugal force (RCF, ~1000g-min.) is sufficient for cell debris removal. Centrifugation in a swinging bucket centrifuge is a common method and can be efficient for volumes below 5 L. The disadvantages include performing this operation using an aseptic technique and the possible generation of an aerosol of virus particles. Centrifugation at higher RCF can lead to the loss of virus. Figure 4 shows relative yield loss over time of particles after centrifugation for 5 min at increasing RCF. 2. Filtration Filtration is another clarification method. Either cross-flow filtration or dead end single-pass filtration can be used to remove debris [16]. A study employing a variety of membranes with varying compositions is necessary for optimizing yield. Some membrane materials bind proteins and may also bind adenovirus. Membranes composed of polyethersulfone possess low protein binding characteristics. Adenovirus will pass through these membranes with excellent yields if the concentration of the adenoviral particles is kept below 5 X 10^^ particles/mL. Most filters are rated by their performance in passing dyes or particles of standard sizes. Flowever, manufacturers generate membrane pores in different ways. Some membranes possess pore sizes similar to the cutoff size; while most possess pores much larger than the cutoff. These types of membranes rely on the torturous path the solute must follow to get through the membrane. Proteins and other contaminants can interfere with filtration by forming a barrier that effectively reduces the pore size. Adenovirus can be filtered through some 180 Shabram ef aL 00 - • • • o 90 - • 80 - 4> • 70 - 60 - • 50 - 1 1 r- 1 1 - I 1 1 2000 4000 6000 8000 10000 12000 14000 16000 18000 RCF Figure 4 The density of adenoviral particles is significantly higher than most lysate components. Samples of infected cell lystate were spun in an Eppendorf Microfuge model 5415c at 4°C for 5 min at different speeds. Recovery of particles was measured using anion-exchange HPLC. Yields were plotted against relative centrifugal force (RCF). Particle loss occurred when the sample was subject to RCF greater than 3000 g min. 0.22-\|xm membranes with better than 90% yield; but the membrane must be selected by experimentation. Fortunately, adenoviral preparations may be sterilized by 0.22-\|xm filtration with filters available from many manufacturers such as Millipore, Gelman, or Sartorius. 3. Expanded Bed Chromatography Expanded bed chromatography removes cell debris using an upward- flowing chromatography column partially filled with large or dense beads [17-19]. "Expanded bed" differs from "fluidized bed" in that the suspended bed is stabilized by a gradient of bead densities so that mixing is minimized. Resin of this type is available from Pharmacia (Piscataway, NJ). In this mode of operation, the column is fed from the bottom at a flow rate sufficient to suspend the beads throughout the column but not cause the resin to pack at the top. Supernatant exits through the top frit. After absorption of the entire volume to be clarified, the direction of flow reverses and the column is packed, washed, and eluted. The main advantage of this technique is that it accomplishes debris removal and column chromatography in one unit operation. Drawbacks include the need for a special column and a special resin, limiting flexibility. The flow rate for loading the column is restricted to 7. Purification of Adenovirus 181 a rather narrow range by the requirements to keep the bed suspended but not packed at the top of the column. 4. Digestion of DNA Lysis releases a large amount of DNA in both large and small fragments. Some of the DNA associates with DNA binding proteins and are found in large structures. DNA (and RNA) digestion is necessary because this con taminant promotes aggregation and complicates downstream processing. The adenovirus particle possesses sufficient surface area that significant amounts of DNA bind to the capsid in spite of the anionic surface charge of the particle. Much of this DNA is viral and if not separated can be taken up by the target cell causing abnormal replication. In some cases, El can be cotrans- fected with an El-deficient particle and can give rise to the generation of replication-competent adenovirus. Consequently, it is important to eliminate as much of the exogenous DNA as is practical. Fortunately, nucleases such as Benzonase (Merck KGaA, Darmstadt, Germany) are available in highly pure forms able to digest the majority of the complicating nucleic acids. The enzymes function best at 37°C near pH 7 in the presence of magnesium ion. The salt concentration is also critical as high salt inhibits the enzymes. Aggregation of the adenovirus, which depends upon collisions, occurs more rapidly at warmer temperatures, while the enzymes function poorly below 15°C. The lysate must be cleared of debris since the quantity of host cell and free viral DNA competes for enzyme and results in incomplete digestion. A compromise is to perform a modest clarification (see above), adjust the salt concentration by dilution, buffer the solution to maintain physiological pH, add magnesium ion, and then perform the digestion at room temperature. The final concentration of nuclease can be increased to accelerate the process. D. Purification Once the clarified lysate is free of exogenous DNA, substantial purifi cation can proceed by a variety of techniques. At small scale both chromato graphic or density gradient centrifugation methods are effective. Large scale, however, favors chromatography. 1. Ultracentrifugation Meselson et al. [20] presented a new method of determining the molecular weight and partial specific volume of macromolecules by density gradient centrifugation. This technique has been particularly useful for macromolecules such as DNA and viruses. Salts, such as CS2SO4 and CsCl, form density gradients when subjected to a strong centrifugal field. Macromolecules separate from contaminates on the basis of their respective buoyant densities and collect in bands at their own density if the sample is centrifuged to equilibrium. 1 8 2 Shabram ef a/. Adenoviruses can be purified using this technique since the buoyant density of the particles is approximately 1.34 g/mL. A typical purification scheme is a three-step process where the infected cell is lysed and the DNA is digested. The sample is then applied to a step gradient of CsCl in a tube where the density of the bottom layer of CsCl is around 1.4 g/mL and that of the top layer is around 1.25 g/mL (both layers are buffered with Tris to approximately pH 8). After spinning at approximately 150,000 g for 1-2 h, the virus separates from cellular debris and collects in a band between the CsCl layers. The band is collected by puncturing the tube and drawing the material out with a syringe. This collected band is then mixed with CsCl at 1.35 g/mL, placed in a centrifuge tube and subjected to 200,000 g overnight. The intact virus separates from DNA, proteins, and defective particles and is collected as before. CsCl is then removed by dialysis. This method is easy to perform and yields high-purity virus preparations. Unfortunately, the process time required is long, CsCl must be removed from the final product, and specialized equipment is required. The main disadvantage is that this process cannot be performed at the large scales demanded for pivotal clinical trial or market use. 2. Purification by Chromatography Column chromatography is by far the most versatile and powerful method for purification of viruses. The methods described above serve to prepare the viral preparation for chromatography by freeing it from cells, cell debris, and interfering substances. The clarified lysate must be in a buffer suitable for application to a chromatography column. Modes of chromatography applica ble to viruses include ion exchange, reversed phase, hydrophobic interaction, size exclusion (gel filtration), immobilized metal chelate, affinity, etc. For each mode, one could choose from many commercially available resins and mobile phases. The selection of an optimal sequence of chromatography steps has been made easier by commercial instruments that are able to perform a sys tematic search of columns and gradient conditions. Books and review articles describe these modes of chromatography and offer strategies for selecting the best combination [21-27]. Specific purification methods for adenoviruses are also found in the literature [28]. Fundamental differences distinguish analytical and preparative chro matography [29]. Analytical runs are performed by injecting a small amount of sample onto a column with high resolving power. Such columns typically have very small particle size beads (3-10 \|xm), high theoretical plate numbers, and rapid run times under high pressure. The result of the run is judged by the appearance of the chromatogram; that is, the peaks should be symmetrical, narrow, and well resolved. Preparative runs, in contrast, are carried out by applying a large sample load (usually near the maximum for the column) using columns of 1 to 500-L bed volume packed with larger beads (20-90 \|xm 7. Purification of Adenovirus 1 8 3 or larger). The cost of the packing becomes more significant for such large columns. The result of the preparative run is measured by the ability to recover pure fractions from the column with high yield. This means that an analytical technique is needed to judge purity and yield of the fractions. Anion exchange HPLC and reversed-phase HPLC [30-32] serve as the tv^o most pow^erful analytical techniques (see section III of this chapter). Without such information, the outcome of a preparative separation is not known because the chromatograms for many preparative separations are complex and difficult to interpret. Method scouting is conveniently done on small columns in high-pressure systems (HPLC, FPLC) in order to speed development and conserve material. Media with smaller bead sizes may be used provided larger bead sizes are also available. Preliminary screening should identify two modes of chromatography able to resolve the virus of interest from the contaminants. The first column step usually employs a resin with high binding capacity and/or high selectivity for the product; otherwise, a very large column may be needed. Anion exchange chromatography is often selected as the first step. Adenoviruses do not adsorb to cation exchange resins at physiological pH; but this type of resin may be used as a first step. While the adenovirus passes through cation exchange columns, some protein contaminants will be removed from the viral solution. For aden ovirus, this method is not needed. Resins for each step are then selected based on resolution, recovery, speed and cost, and possibly other factors such as freedom from extractable materials and availability of documentation required for the production of clinical material for human trials under current Good Manufacturing Practices (cGMP, Part 21CFR). The mobile phase is selected and the gradient optimized. Sample volume and concentration influence resolu tion in column chromatography so both of these parameters must be optimized as well. All of this work can be done on relatively small columns. The product produced at a small-scale should meet all of the purity requirements desired for the final product. Scale-up of chromatography steps is performed by maintaining the media bed height and linear flow rate of the mobile phase while increasing the cross- sectional area (hence column volume) of the column. Fine-tuning the process is usually done at the production scale; only minor adjustments should be needed. Column packing instructions depend on the particle size and nature of the resin [25]. The resin manufacturer's instructions should be followed and then checked by measuring theoretical plates, w,
for the column with acetone (UV detection) or sodium acetate or sodium chloride (conductivity detection). The theoretical plate number can be measured from the chromatogram by the formula [29] 1 8 4 Shabram et al. where t is the retention time and Wi/2 is the width of the peak at half of its maximum height. Sometimes the width is measured at the basehne. In this case, the constant in the equation changes from 5.54 to 16. The plate number increases with column length. Often it is useful to correct the plate number of column length, yielding a parameter known as "height equivalent theoretical plate" (HETP), given by H E T P ^ ^ . N These parameters offer a simple way to monitor column performance over time. Care should be taken to avoid introducing air into the column because air pockets degrade performance and may necessitate repacking of the column. Columns should be cleaned after use and stored in a suitable bacteriostatic environment [25] following the manufacturer's directions. Most process resins can be cleaned and sanitized with sodium hydroxide solutions in the range of 0.1-1 N (exception: silica-based materials dissolve at alkaline pH). a. Ion-Exchange Chromatography of Adenovirus Ion-exchange chro matography offers a powerful method for adenoviral fractionation because of its high capacity and resolution. Ion-exchange chromatography exploits the charge that proteins carry on their surface. The net charge of these groups varies with pH and the amino acids exposed in the protein surface [27, 33]. Adenoviral capsids are highly anionic in nature making anion exchange ideal for purifying them. Anion exchange resins carry positively charged groups such as diethylaminoethyl (DEAE) or quaternary amino ethyl (QAE), which bind anionic proteins in a manner that depends on pH. Elution may also be accomplished by changing pH to eliminate the ionic interaction with the protein. For proteins, it is helpful to know both the isoelectric point of interest and how the protein charge varies with pH. These properties can be measured by isoelectric focusing and electrophoretic titration [25]. Good binding and elution characteristics are often obtained about 1-1.5 pH units above the iso electric point for anion exchange or an equal increment below the isoelectric point for cation exchange. The predominate capsid protein is the hexon which possesses an isoelec tric point near pH 6. As mentioned above, the particle is only stable in a narrow pH range near pH 7. Ion exchange resins also bind protons (cation exchange) or hydroxyl ions (anion exchange). Increasing salt concentration may lead to large changes in pH because salt competes with protons or hydroxyl ions for binding sites on the resin. One should be alert to the possibility of pH changes, possibly as much as one pH unit during chromatography. Maintaining the pH with the correct buffer at sufficient concentration is important for stabilizing the pH during elution. Several ion exchange resins should be tested for binding capacity and resolution at a constant flow rate because resins with the same functional 7. Purification of Adenovirus 1 8 5 group may differ considerably in these properties owing to differences in their backbone, density of substitution, or other factors. After selection of the resin and mobile phase the other critical parameters can be optimized: sample load and volume, flow rates for absorption and elution, and elution gradient. Hexon is a noncovalent trimer that is anionic at pH 7. The capsid is composed of 240 of these capsomeres and gives the particle a large number of negative charges on the surface at neutral pH. Proteins bind to ion exchange resins in low salt (5-50 mM NaCl) and elute with high salt (0.1-1 M). Concerted binding of capsomeres in the capsid to the resin allows the particle to adsorb at higher salt concentrations than those used to elute endotoxins and most proteins. This allows easy separation of viral particles from proteins. Most chromatographic resins are optimized for different classes of ligands by making the resin particle with various pore sizes. Proteins and other ligands have access to a substantial amount of resin surface area inside the pores. Adenoviral particles do not have access and are limited to the outer surface of the resin. Large fragments of DNA, however, are also highly anionic but with a higher charge density. Consequently, DNA elutes at higher salt concentrations than adenovirus. These properties result in an order of binding and elution for the constituents of a clarified lysate. At buffered salt concentrations as high as 350 mM NaCl, the particle binds to the column while nearly no free proteins bind under these conditions. Large DNA-protein complexes such as incorrectly assembled particles and some cellular structures bind under these conditions. DNA binds tightly. Using a linear salt gradient the order of elution will generally be proteins first, followed by complex contaminants, viral particles, other cellular-derived structures, and, last, DNA that has escaped digestion. The elution of these components results in excellent separation between peaks (Fig. 5). Purification yields are as great as 99% but can be lower if the peak must be trimmed to improve purity. b. Immobilized Metal Afftnity Chromatography of Adenovirus In 1975, Porath showed that metal ions could be linked to a column in a 1:1 com plex with a chelating ligand, iminodiacetic acid, bound to the column 134, 35]. These columns had unique properties for fractionation and provide orthogonal methods for purification. This technique is referred to as "immobilized metal affinity chromatography" or IMAC. Beaded agarose is the most common sup port and iminodiacetic acid remains the most popular chelating ligand. Such columns can be charged with a variety of divalent metal ions, Zn^^ and Cu^^ being preferred for protein chromatography 136]. Adenovirus particles bind readily to Zn^+ charged resin, whereas Cu^+ is not as efficient. Excess metal ion is removed by washing before applying protein. Bound metal ion forms a coordination complex, leaving some coordination sites free to interact with proteins. Protein binding typically occurs through histidine residues [23, 36], which occupy the free coordination sites. However, coordination with epsilon 186 Shabram ef a/. 15 Column Volume Figure 5 Anion exchange chromatography is the most robust method to recover and purify adenoviral particles from crude stocks. Monitoring the optical density at 280 and 260 nm allows the chromatographer to easily recognize the fractions containing adenovirus by taking a ratio of A260 ^o A280- The ratio for pure virus is around 1.25. The chromatogram above is plotted as the optical density vs the number of column volumes of materials that have been pumped through the column. In this example, DEAE Fractogel 650 M (EM Sciences) was used. The column was buffered in 50 mM Hepes, pH 7.5, at room temperature throughout the process. Infected cell lysate was loaded onto the column with an adjusted salt concentration of approximately 350 mM NaCl. The load produced significant absorbance as the majority of contaminants passed through the column. More contaminants were eluted during a post loading wash with equilibration buffer. Elution was with a linear salt gradient from 350 to 600 mM NaCl. The adenovirus peak is well resolved at the end of the chromatogram. The approximate salt concentration of the collected peak was about 450 mM NaCl. Column cleaning was achieved with 1 M NaCl and 0.5 N NaOH (data not shown). The column height for this chromatography was about 5 cm. The chromatography looks the same, however at 10 cm bed height. Scale up produces the same chromatogram if the column diameter is increased using the same bed height and the flow rate is adjusted accordingly. amino groups is also probable. Elution can be achieved either by changing pH or by adding competitors such as imidazole or glycine, for the binding sites. Ethylenediaminetetraacetic acid (EDTA) can be used to elute the column, as EDTA strips the metal from the column and the protein. IMAC works well as a polishing step for purification since it removes residual host cell contaminants. Since IMAC can be operated in high salt conditions, fractions from an initial recovery column, such as anion exchange eluate, can be loaded directly on the equilibrated and charged column. The 7 . Purif ication of Adenov i rus 187 buffer selected must not strongly chelate metals, of course. The yields typically fall in the range of 60-80% and the purity is greater than with CsCl ultra- centrifugation methods. Figure 6 shows a chromatogram of adenovirus type 5 purified with zinc charged IMAC and eluted with a step gradient of glycine. Modifications to the fiber can also affect the chromatography. Inter estingly, "fiberless" adenovirus, vectors that have been altered to express truncated fiber protein, bind normally to anion exchange resin, but do not bind at all to IMAC charged with zinc. Presumably the residues serving as zinc binding sites have been removed from the surface of the fiber by muta tion. Alternatively, the adenovirus fiber can be engineered to provide histidine repeats that will bind to Ni^+ very effectively [37]. c. Reversed-Phase Chromatography Reversed-phase HPLC can be used for analysis of adenovirus particles or as a polishing step after initial purification 0.35 Adenovirus 0.30 0.25 0.20 •^ 0.15 O 0.10 0.05 Contaminants 0.00 I 4 6 10 Column Volume Figure 6 High-resolution techniques, such as zinc metal affinity chromatography are needed to complete the purification of adenoviral particles. TosoHaas AF chelate 650 M immobilized metal affinity resin was charged with divalent zinc. The column was equilibrated at room temperature with 450 mM NaCi in 50 mM Hepes at pH 7.5. DEAE adenovirus fractions (Fig. 5) were loaded onto the column followed by a wash with equilibration buffer. Remaining contaminants eluted from the column during the load and wash. The adenovirus was eluted from the column with a 500 mM glycine step gradient. After elution, the column was stripped with EDTA followed by 1 M NaCi then 0.5 N NaOH (data not shown). 1 8 8 Shabram ef o/. by anion exchange chromatography. The recoveries for the preparative method run in the range of 20%, which is poor compared to other methods. The lower recovery may be related to the presence of organic solvents in the mobile phase as high-molecular-weight proteins tend to denature or precipitate. Recoveries may be improved through careful selection of column, solvent, ion-pairing agent, and pH. d. Hydrophobic Interaction Chromatography of Adenovirus The dis covery of hydrophobic interaction chromatography (HIC) resulted from an attempt to make affinity columns [38]. This fortunate accident uncovered a unique mode of protein chromatography. On the surface, HIC resembles reversed-phase chromatography [27, 39, 40] in that the protein binds to the column through hydrophobic interactions in an aqueous solvent. Both resin types consist of a stationary phase with a hydrophobic surface. There after, the two techniques diverge [23]. HIC resins are typically constructed from polysaccharide or polymeric material. Reversed-phase resins are typically bonded silicas. HIC resins have a lower density of substitution and they tend to be less hydrophobic than reversed-phase media. Typically, the conformational changes are driven by high salts (such as ammonium sulfate). The salt presents an ionic environment that is favorable to hydrophilic surfaces. Hydrophobic surfaces are driven together so that exposure to the environment is reduced. In this way, the proteins are partially "salted out" and adsorb to the resin. The use of ammonium sulfate is relatively gentle because most proteins are stabilized in the presence of high concentra tions of ammonium sulfate. High salt, however, may destabilize the particle (see section IV, below), possibly because of its high water content or because the capsid proteins are twisted into destabilized conformations. The loading material may be adjusted to high salt prior to application to the column. Alter natively, small amounts may be applied to the column repeatedly, washing with equilibration buffer, or the material could also be diluted in equilibration buffer inline with the load. The advantage of the two later methods is that protein precipitation occurs slowly from the time of salt addition. Limiting the time of exposure to high salt may improve the chromatography and mitigate yield loss. Elution is achieved by reducing the salt concentration with a reverse gradient. With the exception of the direction of the gradient, HIC columns are optimized and operated along the same lines as ion exchange columns. Resi dence time on the column should be minimized because of the possibility
of denaturation [23]. Yields of virus from this type of chromatography typically range between 20 and 60%. e. Size-Exclusion Chromatography (SEC) of Adenovirus Size exclu sion, also called gel filtration or gel permeation, is the only mode of 7. Purification of Adenovirus 1 8 9 chromatography that is not intended to involve binding of proteins to the resin [41, 42]. The pore structure of the resin provides a molecular sieve, w^here smaller molecules can access the entire volume of the pores and large molecules are excluded from the pores. If a mixture of proteins differing in size is applied to a size exclusion column, the largest proteins v îll emerge first and smallest last. Molecules above a certain size do not penetrate the pores at all. Normally, these are the first to elute from the column; the volume at which they elute is termed the "excluded volume." Adenovirus particles, perhaps owning to the slow rate of diffusion may not elute in the excluded volume of the column. Instead, the particles may elute as much as a column volume beyond the excluded volume. Late elution from SEC often results from charge interactions between the sample and the column. This effect may be reversed by raising salt concentration. Similarly, molecules below a certain size all elute at the "included volume." All other species elute between the excluded volume and the included volume. This property limits the resolving power of size-exclusion chromatog raphy because the number of peaks that can fit into the volume allowed is small. A further limitation of size exclusion chromatography is that resolution deteriorates if sample volume exceeds 4 - 5 % of the bed volume of the col umn. Another limitation is that resolution deteriorates with increasing sample viscosity. The maximum protein concentration allowing good resolution is usually in the range of 5-10 mg/mL protein. Taken together, these two factors mean that a size exclusion column has 1-5% as much protein capacity as an ion-exchange column of the same size! Hence, size exclusion has limited utility for purifying adenoviral particles or proteins and is usually reserved as a last step. Sample load volume, maximum protein concentration and flow rate should be determined by experiment. Careful column packing technique is critical for good results. Plate number should be measured for a new column and at regular intervals during use. Despite these limitations, size-exclusion chromatography has an important place in the arsenal. It is gentle and rapid so yields are nearly quantitative. Additionally, it provides an opportunity to exchange the buffer to the desired formulation because SEC is compatible with a wide range of aqueous buffers. E. Buffer Exchange Every process confronts the problems of removing low-MW species and/or concentrating the desired fractions. Dialysis [43-45] or gel filtration (section II.D.l.e, above) may be used to remove small molecules or exchange buffers when the sample volume is in the range of 1 L or less. Several types of ultrafiltration devices are available for concentration of proteins on a laboratory scale, including pressurized stirred cells and filters driven by centrifugal force or other means. 1 9 0 Shabram et al. Concentration of protein solutions at a process scale is usually done by ultrafiltration using a tangential flow filter [16, 27, 46], Buffer exchange and removal of low-MW species are usually done by diafiltration; both concentration and diafiltration may be done on the same device and combined as a unit operation. Diafiltration is more efficient than dialysis in that less buffer and less time is needed to achieve a given level of solute removal. The concentration of solute remaining in the retentate after 5 vol of continuous diafiltration is given by [47] w^here a represents the rejection of the solute by the membrane. For solutes freely permeable to the membrane, a = 0. Under these conditions, diafiltration with 3 vol of buffer reduces the concentration of solute by 95%. In contrast, 20 vol of buffer would be required for the same result by standard dialysis. In either case, buffer exchange of viral particles can be achieved. Care must be taken to avoid foaming or excessive shear. Special attention should be devoted to pH control. The process should be monitored to avoid overconcentration and possible loss of product through precipitation. III. Analytical Methods for Process Development and Process Tracking Analytical methods are as important for purification as the process steps themselves. Analytical methods are essential for following the process and assessing the purity of adenovirus particles throughout the process. The methods must be rapid, reliable, and informative about the quantity and quality of adenovirus particles. They should be sensitive enough to detect subtle changes. As with the process techniques, characteristics of the virus are useful for selecting analytical techniques. In this section we have highlighted several assays not only because their usefulness for a process, but because an understanding of what these assays mean is critical in producing vector with the quality required for use in humans. A. Plaque-Forming Titer Assays Plaque forming assays have been in use as biological assays since early in the 20th century. A common method for many viruses, this type of assay has been employed with adenovirus since they were discovered. The plaque assay is performed using many variations but generally consists of diluting the virus preparation to a point that a thin layer placed over sensitive cells will result in a countable number of infection events. This is usually accomplished in a petri dish or in six-well plates. Once cells have been exposed, the viral 7. Purification of Adenovirus 1 9 1 solution is removed and a layer of warm agar applied on top of the cells. After 1 to 2 weeks of incubation, the cells are stained with a dye, such as neutral red, and plaques of lysed cells become visible. It is assumed that because of the extreme dilution, each plaque is the result of a single viral particle infecting a cell. The plaque arises following the replication of that viral particle and subsequent infection of adjacent cells by virus progeny. The number of plaques in a well divided by the inoculum volume and corrected for dilution yields a titer. This method is simple but relatively insensitive. If the cells are robust, the inoculum layer may be very thin and the exposure reasonably long. The slow diffusion rate of the particle and the formation of a meniscus in the well limit the method to sampling about 10% of the virions in a sample. B. Adenovirus 96-Well Titer Plate Assay A convenient method for estimating adenoviral titers uses cells (of an appropriate cell line) in the wells of a 96-well plates. The cells are plated such that they reach 50% confluence after 1 day of growth. The sample is then diluted so that the final particle concentration falls between 5 and 1000 particles/mL. Several different dilutions are prepared. It is best to perform an initial dilution in the original sample tube using the whole sample because the freeze-thaw process concentrates the virus into bands that will not disperse without substantial mixing. The largest error in dilution usually occurs with this first dilution. The initial dilution should be limited so that the concentration after the initial dilution can be verified using a particle assay. Thereafter, dilution of the sample is not problematic. To infect the cells the entire medium is first removed from the seeded wells. Each different dilution of virus is pipetted into at least 10 wells per dilution. When using a diffusion-adjusted calculation (see section III.2, below) the wells should be filled to the top. It is common in practice, however, to inoculate 100-200 \|xL of medium with 10-50 \|xL microliters of virus solution. For diffusion-corrected calculation the infection time is limited to 1 h or less (15 min is convenient) and then the virus solution is replaced with medium. For methods using Spearman-Karber titer calculations, such as the Lynn Titerpint analysis [48], analysts typically leave the virus solution on the plate for the duration of the assay. Incubation of cells takes place at 37°C, 7-10% CO2, 90-100% humidity for varying times depending on the method of detection. One method is to fix the cells after 3 to 5 days with methanol and acetone followed by staining with a FITC-conjugated anti-adenovirus antibody. This method of detection requires microscopic examination of each well under ultraviolet light. A well is counted if one or more fluorescent cells as positive. If cytopathic effect is used for detection, 1 to 2 weeks of incubation will be required. The dilutions that produce fewer than 100% positive wells are used in the titer calculations. 1 9 2 Shabram ef al. 1. Spearman - Karber Analysis Spearman-Karber analysis, based on Finney [49], essentially converts data such that graphing the data as log dilution vurses positive v^ells approaches a straight line. Spearman-Karber performs an interpolation to a midpoint. Thus, Spearman-Karber gives a log dilution w^here 50% of the v^ells v^ould have been positive. Titer is expressed as a negative log of the dilution. The Lynn program transforms the number by taking the reciprocal of the dilution (10 raised to the powder of the Spearman-Karber number) and divides by the inoculum amount ostensibly to get the inoculum concentration. How^ever, this value is expressed as ED50 (or TCID50) per milliliter and is often substituted for a virus concentration. This assumes that everything that was put into the well is measured by the assay and the fact that positive wells follow a Poisson distribution is ignored. A more appropriate analysis accepts the Poisson distribution in that even if the average number of virions per well is one, not every well would get a virion; but the Poisson distribution is applicable for this kind of assay. The Poisson distribution is given by where p is probability or fraction of positive wells, S is the event density or average virion per well, and r is the number of virion in a particular well. For the number of wells that get zero particles, r = 0, Ŝ = 1, and r! = 1 so the fraction of positive wells is given by 1 — e~^. The Spearman-Karber analysis gives us a convenient way to take all the data into account, calculate a standard error, and then apply the Poisson distribution to get a concentration. Since the value obtained gives the dilution where the fraction of positive wells is 0.5, p = 0.5. Solving for S yields S = 0.69 virion per well. This is always the case when 50% of the wells are positive. If, for example, the inoculum were 50 \|xL, then the concentration of the inoculum would be 13.8 particles per milliliters. Using the dilution obtained by the Spearman-Karber number the original concentration can be calculated. As in the plaque assay, the key assumption is that no virion in a well escapes detection. In the time frame of these assays, this is clearly not possible. A more precise analysis must take into account Brownian motion (diffusion). 2. Diffusion-Normalized Calculation In the older animal virus literature the methodology used for measuring infectious titer was simple: a thin-layer of a viral preparation was placed over the target cells for as long as practical. It was intuitively understood that the infection process was diffusion limited. This methodology would help minimize underrepresenting the titer. Done properly, these assays may underestimate by 10 to 100-fold; but the values obtained in a given experiment were useful in a relative sense. However, many of the cells used for replication-deficient 7. Purification of Adenovirus 1 9 3 adenoviruses cannot be maintained with very lov^ media levels. Unfortunately, the adenoviral particle, or any particle of similar size, diffuses very slov^ly in solution. Adding more media mitigated the sensitivity of the cells, but resulted in substantially greater underestimations of titer. This is because the probability of infecting a given cell is dependent on the concentration of the virus and the time of exposure. The discrepancy between adenoviral titers and the more precise particle assays (see below) has furthered the concept of particle to infectious unit ratio, or PIU. In the case of adenoviral vectors, this concept is based on the supposition that, in a population of intact and otherwise complete particles, most are not infective. However, a search of the literature does not support this supposition.
To the contrary, work by Nyberg-Hoffman et al. [3] used a model derived from Pick's Laws of Diffusion to demonstrate that most, if not all, adenovirus particles are indeed infective. That work, and others [50], showed the importance of diffusion mechanics for virus binding and demonstrated the dependence of titer determination on often ignored experimental conditions. The defective particle misconception neglects the important factor that most of the added particles in a well never contact the target cell during the critical period of the assay. Changes in assay conditions such as particle con centration or exposure time can have a dramatic effect the results. A significant outcome of the misconception has been the requirement by regulatory agen cies to demand that there be at least one infective particle per hundred total particles (FDA Guidelines). Substantial resources have been spent in attempts to purify away the "defective particles." Unfortunately, in many cases these particles are merely experimental artifacts and thus cannot be removed by purification. Other ramifications of this misconception have been discussed in the literature [50, 51]. Diffusion can be accounted for by the use of diffusion-normalized analy sis [3]. This analysis takes into account the diffusion of the particle under the conditions of the assay by deriving normalization equations from Pick's Laws of Diffusion. Por a titer plate assay the equation is given by ' " ( • - T ) V — ^̂ —n X dilution factor, where pw is the number of positive wells per dilution, n is the total number of wells per dilution, A^ is the area of the bottom of the well in cm^, Cw is the confluence of the well at the time of infection, I is a constant incorporating the diffusion coefficient and is equal to 2.38 x 10~^ cm/particles s^/^, and t is the exposure time in seconds. Prom the equation, one can see that p^ must be less than n and greater than zero. Optimally, the number of positive wells should be between 20 and 80% of the total wells in a dilution. This method yields titers that are up to 50% of the particle concentration. 1 9 4 Shabram et al. C. Flow Cytometry Fluorescence activated cell sorting (FACS) offers another sensitive method to assess infective titer [52]. Sufficient quantities of permissive cells for analysis can be grown and infected in six-well plates. Infected cells are harvested, fixed, and stained with a FITC-conjugated anti-adenovirus antibody. Infected cells are brightly stained while uninfected cells are not. FACS analysis determines the fraction of cells infected at the time of harvest. The fraction of infected cells at the time of infection is diluted at the time of harvest because uninfected cells continue to divide. By performing cell counts at infection and at harvest times, one can then calculate the proportion of cells infected at infection time by using the fact that the number of infected cells does not increase during the incubation period. Multiplying the proportion of infected cells by the number of cells at infection time yields the number of cells originally infected. Since the total number of cells at infection time is also known, the proportion of infected cells at infection time is also known. This value can be used to calculate a titer. Cells are infected with a virus concentration high enough to get 5-10% of cells to stain positive. Exposure time ranges from 30 s to 60 min. Incubation is up to 50 h to avoid secondary infections. The titer is given by the diffusion- adjusted equation: y — — X dilution factor. Iip^/t The average cell area (two-dimensional footprint), symbolized by cp, is a variable that must be determined for each cell line. Subclones of HEK 293 cells, for example can display morphology differences from the parent stock. Cell area can be determined using image analysis to analyze micrographs of cells. HEK 293 cells obtained from ATCC and at low passage number possess a cell area of approximately 6.3 x 10~^ cm^. Other subclones of 293 cells can have larger or smaller areas. F is the final adjusted fraction of positive cells detected. / is 2.38 x lO""^ cm/particles s /̂̂ as above. This method yields titers that are 50 to 80% of the particle concentration. This is probably the best value that can be obtained considering that no step in the infection process is 100% efficient. D. Particle Concentration Determination by Ultraviolet Absorbance The most common method for measuring particle concentration is to dis rupt the particles using SDS followed by absorbance measurement at 260 nm. Maizel [53] determined that the absorptivity of adenovirus was 1.1 x 10^^ particles/mL/absorbance unit at 260 nm. This method is convenient and rapid but not without limitations. The sample must be pure and free of particulates and aggregates in order to obtain an accurate reading. The buffer formulation of the sample can affect the reading as salt concentration partially determines 7. Purification of Adenovirus 1 9 5 the concentration at which SDS may form micelles. This method cannot dis tinguish disrupted particles from intact particles. Contaminating DNA may increase the absorbance and lead to an overestimation of particle concentration. Absorbance readings at several wavelengths provide a check on the validity of the assay. Absorbance in the longer ultraviolet and visible regions indicates light scattering. The ratio of 260 nm (DNA) to 280 nm (protein) should fall between 1.2 and 1.3. A ratio outside this range indicates contamination. E. Analytical Reverse-Phase HPLC Re versed-phase high-performance liquid chromatography (RPHPLC) first achieved prominence as an analytical technique because of its wide applicability and ability to resolve a large number of components in a sin gle chromatographic run [54]. This excellent technique also works well as a preparative method for some proteins, mainly those of lower molecular weights (<30,000) [25, 55], RPPHPLC is the dominant method for the purification of peptides and protein of MW < 10,000 [56]. For example, human insulin is produced at a level of tons per year using RPHPLC. Application to larger proteins is limited by the denaturing tendency of organic solvents. For this reason RPHPLC is considered a denaturing technique. The binding of proteins to reversed-phase columns results from a hydrophobic interaction between exposed regions of the protein and the hydrophobic surface of the stationary phase (the resin). Denaturation by the mobile phase exposes hydrophobic regions buried within the protein. Elution is achieved by applying a gradient of increasing concentration of organic solvent, usually acetonitrile. Proteins tend to elute as broad, asymmetrical peaks unless an ion pairing agent, such as trifluoroacetic acid (TFA) is included. TFA is thought to bind to the positive charges on the protein, masking negative charges on the resin matrix and providing additional hydrophobic surface to interact with the column [54]. RPHPLC is a powerful analytical tool to use for adenoviral samples because it resolves the proteins contained in the virus particle. This method, described in Lehmberg et al. [31], has provided detailed insights into the nature of adenovirus preparations. The method consists of injecting the sample onto a C4 reversed-phase column that has been equilibrated in 20% acetonitrile at a constant TFA concentration of 0 .1%. The column temperature, a critical parameter for RPHPLC, is kept at 40°C. These conditions dissociate the particle into proteins that in turn bind to the column. The column is then eluted with an acetonitrile gradient beginning at 20% and ending at 60%. The TFA concentration is maintained at 0 .1%. The absorbance is monitored at 214 nm and can be monitored at 260 and 280 nm. The resultant chromatogram gives a characteristic fingerprint of the adenovirus proteins. Mass spectrometry and N-terminal sequencing have identified 14 major peaks. These proteins can be recovered quantitatively enabling this assay to be used as a quantitative 1 9 6 Shabram ef al. method for determining particle concentration. The relative peak areas of the proteins can be compared to a known standard to assess particle quality. F. Analytical Anion-Exchange HPLC The anionic nature of the adenoviral particle lends itself to analysis by anion-exchange HPLC [30, 32]. This method is nondestructive and yields a w^ealth of information. Before this method, monitoring the production and purification of adenovirus particles w âs limited to infectious titer assays. For the purposes of process development and monitoring, infectious titer assays were too slow, resulting in low throughput. They were not sensitive enough to distinguish small differences among samples. Protein analysis methods, such as SDS-PAGE, could not quantitate virus in crude samples because most of the viral protein is not incorporated into whole virus [1]. Absorbance techniques to assess viral particles based upon the UV absorption of DNA were not applicable to crude samples. This is because free viral and host cell DNA is present in the sample. These limitations are overcome by AEHPLC. A 1-mL Resource Q anion exchange column (Pharmacia Biotech, Piscat- away, NJ) is convenient for HPLC analysis of samples. Lysates are prepared by treatment with nuclease. Semipure or pure virus can be injected directly because they do not require nuclease treatment. As long as the column is not reequilibrated during injection, the assay is independent of the injection volume. After sample loading, the column is washed with equilibration buffer (Hepes or Tris, at approximately pH 7.5) followed by linear salt gradient elu- tion. The chromatography can be monitored on a standard ultraviolet detector. Significantly, more information can be gleaned by the use of a photo diode array (PDA) detector scanning from 210 to 300 nm. The retention time of the adenovirus peak varies with the serotype of the virus. The AEHPLC chromatogram reveals information about purity, particle integrity, and particle quantity. Figure 7 shows the chromatograms of purified adenovirus type 5 and infected cell lysate. Other serotypes give slightly different peak elution times. Provided that the HPLC system remains below the pressure limits of the resin, the chromatography can be performed in less than 6 min. Monitoring the process by anion exchange HPLC enables the production staff to rapidly obtain a picture of the progress of the purification. IV. Formulation and Stability Process development must determine product stability over a wide range of conditions. The time course of the purification must be scrutinized for excessive run times and delays because rapid processing favors high yields. A well-designed process includes holding points selected such that the viral 7 . Purif ication of Adenov i rus 197 020 Purified AdenovirusType 5 0.15 0.10 Adenovirus 0.05 0.00 200 400 &00 aoo laoo 1200 i4oo laoo laoo 2aoo 1.80 i.eo Adenovirus Infected Cell Lysate 1.40 1.20 1.00 aso aeo 0.40 UNA a20 aoo 1 200 400 aoo aoo moo 1200 i400 laoo laoo 2000 Minutes Figure 7 The most useful method for analyzing both crude and pure samples of adenovirus is anion-exchange HPLC. Pure adenovirus elutes in a nearly symmetrical peak. The adenoviral peak from lysate elutes with baseline separation from contaminants such as hexon and undigested DNA. The method is detailed in Shabram etal. [30]. preparation can be safely stored in the event of a planned or unplanned delay. Stability studies are critical to identify those steps at which the virus is at risk. Additionally, the lessons gleaned from these studies point to a stabilizing formulation needed for clinical trials and beyond. Aggregation of adenovirus particles is w êll know^n to those in the field. The adenovirus demonstrates low^ stability at 4°C. This phenomenon has yet to be explained, although there are examples of cold-sensitive enzymes in the literature [57, 58]. The instability seems to be related to the tendency of adenovirus particles to aggregate. Aggregation requires that particles bind to each other after a collision. As the collisions proceed the aggregating particle groves. A systematic analysis of aggregation v^as published in 1917 by Smoluchowski [59]. Using Smoluchow^ski's coagulation model the aggregation frequency of adenoviral particle can be roughly estimated at one per every 50 198 Shabram ef al. collisions! This may seem surprisingly frequent until one realizes that collisions between particles are relatively rare due to the slow diffusivity of the particle. It also explains the observation that aggregation is dependent on particle concentration. Aggregation may also be the end result of damage that occurs early in the production process but does not manifest itself until the sample is concentrated. Aggregation can be mitigated by
changing the conditions of the formulation such that aggregation events are not favored. Collision frequency can be reduced by increasing the viscosity of the solution. Figure 8 shows the effect of concentration on the stability of virus in phosphate-buffered saline with 2% sucrose at 4°C. Aggregation was measured as disappearance on AEHPLC. Aggregation accelerated at concentrations greater than 5 x 10^^ particles per milliliter. The addition of glycerol increased the virus stability (Fig. 9). This is likely due to the ability of glycerol to cause preferential hydration of protein surfaces [60, 61], leading to a tighter association of the capsid subunit structure. Additionally, an increase in viscosity reduces the number of collisions between particles. Salt concentration plays a significant role in the stability of the particle in that stability is reduced in salt solutions above 300 mM. Stability at these salt levels was increased in the presence of glycerol. The putative damage may be due to an anion-specific effect and probably follows the Hoffmeister series. Studies with potassium chloride, sodium chloride, and cesium chloride showed that while chloride concentrations above 1.5 M were not destabiliz ing, concentrations between approximately 0.4 and 1.0 M could be harmful 100 H 80 2 c 60 ] o U 40 ^ 20 ^ Days at 4° C Figure 8 Adenovirus stability is affected by concentration. Recombinant adenovirus was diluted to 4 X 10^1 (•), 5 x IQ i i ( A ) , 6 x IQi^ ( • ) , o r S x 10^^ ( • ) particles/mL and was stored at 4°C. On the indicated days aliquots were assayed for virus concentration with the anion exchange HPLC assay. The buffer was 20 mM sodium phosphate, pH 8, TOO mM NaCi, 2 mM MgCl2, and 2% 7 . Purif ication of Adenov i rus 199 10 4' o 1 OH Days at 4° C Figure 9 Glycerol stabilizes the adenovirus in solution. Recombinant adenovirus at 10^^ parti- cles/mL was stored at 4°C in aliquots: undiluted ( • ) , diluted 20% with water ( A ) , or diluted with 50% glycerol to a final concentration of 10% (v/v) glycerol (O). On the indicated days aliquots were assayed for virus concentration with the anion exchange HPLC assay. The buffer was 20 mM sodium phosphate, pH 8, 100 mM NaCi, 2 mM MgCl2, and 2% sucrose. regardless of the cation. In contrast high concentrations of sodium sulfate, sodium phosphate or potassium phosphate were not destabilizing. A major consideration is the exposure to pH outside the physiologi cal range. As mentioned earlier, phosphate buffer at pH 7.2 lacks buffering capacity during freeze-thaw. Hence, the rate of freezing and thawing are critical. In general, fast freezing and fast thawing improves the stability of the material. Buffers such as Hepes or Tris maintain buffering capacity during freeze-thaw and therefore are preferable for stabilizing pH. One might suspect that the lower stability at 4°C may be related to pH stability and therefore the buffering capacity of the formulation. While phosphate may be problematic for freeze-thaw it is interesting that substituting Tris for phosphate does not affect the stability at 4°C. Freezing and thawing was mentioned earlier as a potential risk for the viral preparation. Cryoprotection agents are often used to mitigate the risk by disturbing ice crystal formation and providing for an amorphous frozen solid. Typically carbohydrates are use to accomplish this. Sucrose and mannitol are often found in formulations where the freezing process is 2 0 0 Shabram ef al. critical. Typically, mannitol provides slightly better protection with proteins than sucrose and is preferred for its superior cake formation in a lyophilized product. With adenovirus, however, the opposite is true. Sucrose provides moderate protection but mannitol has a clear negative affect. V. Conclusions The use of adenovirus vectors for gene therapy has placed increased demands upon the technology for production, purification, and characteriza tion of virus particles. Some of the classic technology has been reexamined and improved. A new class of methods based upon column chromatography has added a powerful set of tools to this array. In large part, the chromatographic methods are based upon modes of chromatography and resins originally devel oped for protein purification. With proper consideration for the size and other characteristics of adenovirus particles column chromatography may be applied with considerable success. The past decade has witnessed rapid advances in this area. Column chromatography is now a preferred method for adenovirus purification because of its versatility and ability to purify large amounts of virus to a high state of purity while retaining biological activity. Acknov^ledgments The authors acknowledge the indispensable help of Rich Murphy, Beth Hutchins, Daniel Giroux, and Muralidhara Ramachandra for input, advice, and editing of the manuscript. The work presented here is a compilation of effort by many individuals over the past 7 years. In particular, we thank and acknowledge Bernie Huyghe, Daniel Giroux, Mark Horn, Xioa-Dong Lui, Suganto Sutjipto, Erno Pungor, Estuardo Aguilar-Cordova, Douglas Cornell, Mary Nunnaly, Kai Tam, Ann Goudreau, Nancy Connelly, Cassandra Nyberg-Hoffman, Jing Zhou, Fred Porter, Andreas Frei, Barry Sugarman, Margarita Nodelman, Josefina Beltran, Anastasia Sofianos, Karli Watson, Thomas Schluep, Susan Miller, Shu-Fen Wen, Tattanahalli Nagabhushan, and Michael Shepard. References 1. Philipson, L. (1984). Structure and assembly of adenoviruses. Curr. Top. Microbiol. Immunol. 109, 1-52. 2. Oliver, C. J., Shortridge, K. F., and Belyavin, G. (1976). Diffusion coefficient and molecular weight of type 5 adenovirus by photon-correlation spectroscopy. Biochim. Biophys. Acta 437,589-598. 3. Nyberg-Hoffman, C , Shabram, P., Li, W., Giroux, D., and Aguilar-Cordova, E. (1997). Sensitivity and reproducibility in adenoviral infectious titer determination. Nat. Med. 3, 808-811. 7. Purification of Adenovirus 2 0 1 4. Stewart, P. L., Fuller, S. D., and Burnett, R. M. (1993). Difference imaging of adenovirus: Bridging the resolution gap between X-ray crystallography and electron microscopy. EMBO / . 12,2589-2599. 5. Pettersson, U., Philipson, L., and Hoglund, S. (1967). Structural proteins of adenoviruses. I. Purification and characterization of the adenovirus type 2 hexon antigen. Virology 33, 575-590. 6. Boulanger, P. A., Flamencourt, P., and Biserte, G. (1969). Isolation and comparative chemical study of structural proteins of the adenoviruses 2 and 5: Hexon and fiber antigens. Eur. J. Biochem. 10, 116-131. 7. Pusey, P. N., Koppel, D. E., Schaefer, D. W., Camerini-Otero, R. D., and Koenig, S. H. (1974). Intensity fluctuation spectroscopy of laser light scattered by solutions of spherical viruses: R17, Q beta, BSV, PM2, and T7. I. Light-scattering technique. Biochemistry 13, 952-960. 8. Sakaki, Y., Maeda, T., and Oshima, T. (1979). Bacteriophage phiNSll: A lipid-containing phage of acidophilic thermophilic bacteria. IV. Sedimentation coefficient, diffusion coefficient, partial specific volume, and particle weight of the phage./. Biochem. (Tokyo) 85,1205-1211. 9. Graham, F. L., Smiley, J., Russell, W. C., and Nairn, R. (1977). Characteristics of a human cell line transformed by DNA from human adenovirus type 5. / . Gen. Virol. 36, 59-74. 10. Fallaux, F. J., Bout, A., van der Velde, I., van den WoUenberg, D. J., Hehir, K. M., Keegan, J., Auger, C., Cramer, S. J., van Ormondt, H., van der Eb, A. J., Valerio, D., and Hoeben, R. C. (1998). New helper cells and matched early region 1-deleted adenovirus vectors prevent generation of replication-competent adenoviruses. Hum. Gene Ther. 9, 1909-1917. 11. Hehir, K. M., Armentano, D., Cardoza, L. M., Choquette, T. L., Berthelette, P. B., White, G. A., Couture, L. A., Everton, M. B., Keegan, J., Martin, J. M., Pratt, D. A., Smith, M. P., Smith, A. E., and Wadsworth, S. C. (1996). Molecular characterization of replication- competent variants of adenovirus vectors and genome modifications to prevent their occurrence./. Virol. 70, 8459-8467. 12. Branton, P. E. (Ed.) (1999). "Early Gene Expression." R.G. Landes, Austin, TX. 13. Vales, L. D., and Darnell, J. E. (1989). Promoter occlusion prevents transcription of aden ovirus polypeptide IX mRNA until after DNA replication. Genes Dev. 3, 49-59. 14. Penefsky, H. S., and Tzagoloff, A. (1971). Extraction of water soluble enzymes and proteins from membranes. Methods Enzymol. 22, 204-219. 15. Monica, T. (2000). In "Viral Vectors and Vaccines." WiUiamsburg Bioprocessing Foundation, Lake Tahoe, NV. 16. McGregor, W. C. (1986). Membrane separation in biotechnology. In "Bioprocessing Tech nology" (W. C. McGregor, Ed.). Dekker, New York. 17. Chase, H. A. (1998). The use of affinity adsorbents in expanded bed adsorption. / . Mol. Recognit. 11 ,217-221 . 18. Hjorth, R. (1997). Expanded-bed adsorption in industrial bioprocessing: Recent develop ments. Trends Biotechnol. 15, 230-235. 19. Thommes, J. (1987). Fluidized bed adsorption as a primary recovery step in protein purifica tion. Adv. Biochem. Eng. Biotechnol. 58, 185-230. 20. Meselson, M., Stahl, F. W., and Vinograd, J. (1957). Equilibrium sedimentation of macro- molecues in density gradients. Proc. Natl. Adam. Sci. USA 43, 581-588. 21. Burgess, R. R. (1987). "Protein Purification: Micro to Macro. In UCLA Symposia on Molecular and Cellular Biology." R. Liss, New York. 22. Deutscher, M. P. (1990). Guide to Protein Purification. In "Methods in Enzymology" (J. N. A. a. M. I. Simon, Ed.). Academic Press, New York. 23. Gagnon, P. (1996). "Purification Tools for Monoclonal Antibodies." Validated Biosystems, Tuscon, AZ. 24. Jacoby, W. B. (1984). Enzyme purification. In "Methods in Enzymology" (P. P. C. a. N. O. Chapman, Ed.), p. 528. Academic Press, New York. 2 0 2 Shabram et al. 25. Janson, J. C , and Ryden, L. (1989). Protein purification. In "Protein Purification: Principles, High ResolutionMethods, and Applications" (J. J. C. and R. L. Eds.). VCH, New York. 26. Kenny, A., and Powell, S. (1992). Practical protein chromatography. Methods Mol. Biol. 11, 1-327. 27. Wheelwright, S. M. (1991). "Protein Purification: Design and Scale up of Downstream Processing". Hanser Publishers, Munich. 28. Huyghe, B. G., Liu, X., Sutjipto, S., Sugarman, B. J., Horn, M. T., Shepard, H. M., Scandella, C. J., and Shabram, P. (1995). Purification of a type 5 recombinant adenovirus encoding human p53 by column chromatography. Hum. Gene Ther. 6, 1403-1416. 29. Snyder, L. R., and Kirkland, J. J. (1979). "Introduction to Modern Liquid Chromatography." Wiley, New York. 30. Shabram, P. W., Giroux, D. D., Goudreau, A. M., Gregory, R. J., Horn, M. T., Huyghe, B. G., Liu, X., Nunnally, M. H., Sugarman, B. J., and Sutjipto, S. (1997). Analytical anion- exchange HPLC of recombinant type-5 adenoviral particles. Hum. Gene. Ther. 8, 453-465. 31. Lehmberg, E., Traina, J. A., Chakel, J. A., Chang, R. J., Parkman, M., McCaman, M. T., Murakami, P. K., Lahidji, V., Nelson, J. W., Hancock, W. S., Nestaas, E., and Pungor, E. (1999). Reversed-phase high-performance liquid chromatographic assay for the adenovirus type 5 proteome./. Chromatogr. B 732, 411-423. 32. Blanche, F., Cameron, B., Barbot, A., Ferrero, L., Guillemin, T., Guyot, S., Somarriba, S., and Bisch, D. (2000). An improved anion-exchange HPLC method for the detection and purification of adenoviral particles. Gene Ther. 7, 1055-1062. 33. Karlsson, R. M. (1989). Ion exchange chromatography. In "Protein Purification. Protein Purification: Principles, High ResolutionMethods, and Applications" (J. C. Janson and L. Ryden, Eds.). VCH, New York. 34. Porath, J., and Olin, B. (1983). Immobilized metal ion affinity adsorption and immo bilized metal ion affinity chromatography of biomaterials. Serum protein affinities for gel-immobilized iron and nickel ions. Biochemistry 22, 1621-1630. 35. Porath, J., Carlsson, J., Olsson, I., and Belfrage, G. (1975). Metal chelate affinity chromatog raphy, a new approach to protein fractionation. Nature 258, 598-599. 36. Kagedal, L. (1989). "Immobilized Metal Ion Affinity Chromatography." VCH, New York. 37. Douglas, J. T., Miller, C. R., Kim, M., Dmitriev, I., Mikheeva, G., Krasnykh, V., and Curiel, D. T. (1999). A system for the propagation of adenoviral vectors with genetically modified receptor specificities. Nat. Biotechnol. 17, 470-475. 38. Shaltiel, S. (1984). Hydrophobic chromatography. In "Enzyme Purification and Related Techniques" (W. B. Jacoby, Ed.). Academic Press, New York. 39. Eriksson, K. O. (1989). Hydrophobic interaction chromatography. In "Protein Purification: Principles, High Resolution Methods, and Applications" (J. C. Janson and L. Ryden, Eds.). VCH, New York. 40. Kennedy, R. M. (1990). Hydrophobic chromatography. In "Guide to Protein Purification" (M. P. Deutscher, Ed.). Academic Press, New York. 41. Fischer, L. (1980). "Gel Filtration Chromatography." Elsevier, Amsterdam. 42. Hagel, L. (1989). Gel filtration. In "Protein Purification: Principles, High Resolution Methods, and Applications" (J. C. Janson and L. Ryden, Eds.). VCH, New York. 43. Coligan, J. E., Dunn, B. M., Ploegh, H. L., Speicher, D. W., and Wingfield, P. T. (1996). In "Current Protocols" (V. B. Chanda, Ed.). Wiley, New York. 44. Craig,
L. C. (1967).Techniquesforthestudy of peptides and proteins by dialysis and diffusion. In "Methods in Enzymology" (S. P. C. a. N. O. Kaplan, Ed.), pp. 870-905. Academic Press, New York. 45. McPhie, P. (1971). Dialysis. In "Methods in Enzymology" (S. P. C. a. N. O. Kaplan, Ed.), pp. 23-32. Academic Press, New York. 46. Wiseman, A. (1985). "Handbook of Enzyme Biotechnology." Halsted PressAViley, New York. 7. Purification of Adenovirus 2 0 3 47. Cheryan, M. (1986). "Ultrafiltration Handbook." Technomic, Lancaster. 48. Lynn, D. E. (1992). A BASIC computer program for analyzing endpoint assays. Biotechniques 12,880-881. 49. Finney, D. J. (1962). "Probit Analysis." Cambridge University Press, Cambridge, UK. 50. Andreadis, S., Lavery, T., Davis, H. E., Le Doux, J. M., Yarmush, M. L., and Morgan, J. R. (2000). Tov^ard a more accurate quantitation of the activity of recombinant retroviruses: Alternatives to titer and multiplicity of infection [corrected and republished article originally printed in / . ViroL 2000 Feb;74(3):1258-1266]./. Virol. 74, 3431-3439. 51. Shabram, P., and Aguilar-Cordova, E. (2000). Multiplicity of infection/multiplicity of confu sion. MoL Ther. 2, 420-421. 52. Musco, M. L., Cui, S., Small, D., Nodelman, M., Sugarman, B., and Grace, M. (1998). Comparison of flow cytometry and laser scanning cytometry for the intracellular evaluation of adenoviral infectivity and p53 protein expression in gene therapy. Cytometry 33,290-296. 53. Maizel, J. V., White, D. O., and Scharff, M. D. (1968). The polypeptides of Adenovirus. L Evidence for multiple protein components in the virion and a comparison of types2, 7A, and 12. V/Vo/og); 36, 115-125. 54. Snyder, L. R., Kirkland, J. J., and Glajch, J. L. (1997). "Practical HPLC Method Develop ment." Wiley, New York. 55. Hearn, M. T. W. (1984). Reversed phase high performance liquid chromatography. In "Enzyme Purification and Related Techniques" (W. B. Jakoby, Ed.), Part C, pp. 190-212, Academic Press, New York. 56. Rivier, J. (1984). Reversed-phase high-performance liquid chromatography: Preparative purification of synthetic peptides. / . Chromatogr. 288, 303-328. 57. Lee, M. L., and Muench, K. H. (1969). Prolyl transfer ribonucleic acid synthetase of E. coH. / . Biol. Chem. 244, 223-230. 58. Frieden, C. (1971). Protein-protein interaction and enzymatic activity. Annu. Rev. Biochem 40, 653-696. 59. Smoluchowski, M. V. (1917). Versuch einer mathematischentheorie der koagulationskinetic koUoider losungen. Z. Phy. Chem 92, 129-168. 60. Gekko, K., and Timasheff, S. N. (1981). Mechanism of protein stabilization by glycerol: Preferential hydration in glycerol-water mixtures. Biochemistry 20, 4667-4676. 61. Gekko, K., and Timasheff, S. N. (1981). Thermodynamic and kinetic examination of protein stabilization by glycerol. Biochemistry 20, 4677-4686. C H A P T E R Targeted Adenoviral Vectors I: Transductional Targeting Victor Krasnykh'̂ '̂ and Joanne T. Douglas"̂ *Division of Human Gene Therapy Departments of Medicine, Pathology, and Surgery and the Gene Therapy Center University of Alabama at Birmingham Birmingham, Alabama WectorLogics, Inc. Birmingham, Alabama I. Introduction The extensive use of human adenoviruses (Ads) for gene therapy is largely due to the fact that the biology of these viruses has been extensively studied for decades. Therefore, at the time when gene therapy emerged, these viral agents represented a rational choice as a candidate system for delivery of therapeutic genes to diseased tissues. Compared to other gene transfer vectors, both viral and nonviral, adenoviruses possess a number of properties which make them a preferred means of cell transduction. In addition to the well understood biology of Ad and extensively developed methods for the generation, propagation, and purification of these vectors, the in vivo stability of Ad, the capacity to accommodate significant amounts of heterologous DNA, and the ability to efficiently infect a wide variety of different cell types at various points of the cell cycle, have significantly facilitated progress in Ad-based gene therapy in the early stages of its development. A critical overview of the results achieved by gene therapy during the past decade clearly shows that although it has not yet lived up to most of the expectations raised at the time it was conceived and even has resulted in the death of a human being, it has, in fact, shown quite impressive results in some applications. Importantly, being a new field of experimental biology and a promising new direction in modern medical science, gene therapy has attracted a substantial additional workforce, talent ADENOVIRAL VECTORS FOR GENE THERAPY 2 0 5 Copyright 2002, Elsevier Science (USA). All rights reserved. 2 0 6 Krasnykh and Douglas and funds to not only use the knowledge previously generated by classical virologists, but also to foster further efforts in studying the biology of Ad, thereby creating a positive feedback to basic science. Despite recent successes in employing Ad vectors for developing nev^ treatment modalities, progress in this direction has clearly been hampered by a number of deficiencies of therapeutic Ad vectors. The list of these limitations includes at least three major points. First of all, Ad vectors are not as safe as desired. Even though Ads, which are considered mild pathogens in the first place, are used for gene delivery in the form of highly attenuated derivatives of the wild-type viruses, when used at high doses they can potentially lead to quite deleterious side effects, including lethal outcomes. This is due in part to the fact that even replication-deficient versions of these vectors are capable of low-level expression of viral genes in the infected cells, thereby causing vector-associated cell toxicity and death. In addition, since Ad virions contain hundreds of copies of viral proteins, they are highly immunogenic and as such can cause acute immune responses in a patient, which result in rapid clearance of both the injected agent and the virus-infected cells. Finally, the promiscuous tropism of Ad vectors may result in the widespread dissemination of the vector upon delivery to the patient and may potentially lead to random infection of normal tissues, thereby further complicating the safety and toxicity issues. The purpose of this chapter is to demonstrate that by addressing the deficiency of Ad vectors related to their natural tropism, one can potentially improve the specificity, immunogenicity, and safety of these agents and thus further increase their overall utility for gene therapy. li. The Pathv^ay of Adenoviral Infection Strategies to retarget Ad vectors should be based on an understanding of the biology of Ad infection and the roles played by the viral capsid proteins involved in determining tropism. These topics are reviewed elsewhere in this volume and will therefore be described only briefly here. Adenoviruses are nonenveloped viruses with double-stranded DNA genomes packaged into icosahedral capsids. The major protein component of the Ad particle is the hexon protein, which forms the planes of the capsid and appears to play a predominantly structural role in the virion. At each of the 12 vertices of the capsid is a penton complex consisting of a pentameric penton base associated with a trimeric fiber protein which projects from the viral surface. The entry of Ad into susceptible cells requires two distinct, sequential steps — binding and internalization — each mediated by the interaction of a specific component of the Ad penton complex with a cellular receptor. The initial high-affinity binding of Ad to the primary cellular receptor occurs via 8. Targeted Adenoviral Vectors I: Transductional Targeting 2 0 7 the carboxy (C)-terminal knob domain of the fiber capsid protein [1, 2]. The knob domain also initiates and maintains the trimeric configuration of the fiber molecule [3,4], which is critical for the ability of the fiber to associate with the penton base, since fiber monomers cannot be incorporated into mature viral particles [5]. The central shaft domain of the fiber serves to extend the knob away from the surface of the virion, thereby facilitating its interaction with the primary receptor. The amino (N)-terminal tail domain is associated with the penton base and also contains a nuclear localization signal, which directs the newly synthesized fiber polypeptides to the nucleus of the infected cell, where the assembly of the viral particle occurs [6]. Thus, the three domains of the fiber protein fulfill well-defined and distinct functions. The cellular fiber receptor for the two human Ad serotypes which are most commonly used as gene delivery vectors, Ad2 and Ad5 from subgroup C, has been identified as the coxsackievirus and adenovirus receptor, CAR [7, 8]. CAR also serves as the primary receptor for Ad serotypes from subgroups A, D, E, and F, but not subgroup B [9]. CAR appears to function purely as a high-affinity docking site for Ad on the cell surface: the cytoplasmic and transmembrane domains of the molecule are not essential for Ad infection [10, 11]. CAR has two extracellular immunoglobulin-fike domains, of which the N-terminal domain, D l , which is distal to the cell surface, is responsible for binding the Ad knob [12]. Following attachment, the next step in Ad infection is internalization by receptor-mediated endocytosis potentiated by the interaction of Arg-Gly-Asp (RGD) peptide sequences at the apex of protruding loops of the Ad penton base with secondary host cell receptors, integrins avP3 and av^S [13, 14]. After internalization, the virus is localized within the cellular vesicle system, initially in clathrin-coated vesicles and then in endosomes. Acidification of the endosomes allows the virions to escape and enter the cytosol. The virions are then translocated along microtubules to nuclear pore complexes where the capsid is disassembled and the DNA genome is imported into the nucleus [15]. III. Strategies and Considerations The capacity of an Ad vector to infect a given cell is therefore dictated by the CAR- and integrin-expression levels of the cell. It has been shown that cells expressing both receptors below a certain threshold level are refractory to Ad infection [16]. Recent studies have also demonstrated that a number of cell types such as endothelial, smooth muscle cells, differentiated airway epithelium cells, lymphocytes, fibroblasts, hematopoietic cells, and some others demonstrate either complete or partial resistance to Ad infection [10, 17-23]. Importantly, the employment of Ad vectors for cancer gene therapy has revealed that many types of tumor cells express CAR at marginal or even 2 0 8 Krasnykh and Douglas undetectable levels and are thus Ad-refractory [24-26]. An interesting finding in this regard was recently published by Okegawa et al.^ who demonstrated a striking inverse correlation between the level of CAR expression by prostate cancer cell lines and their tumorigenicity, thereby suggesting that in general the most aggressive tumors may be CAR-deficient and therefore refractory to therapeutic intervention utilizing unmodified Ad2- or Ad5-derived vectors [27]. The authors have also observed the same phenomenon on human breast cancer cells. If the results of this work are further corroborated by data from other laboratories, CAR deficiency in tumors may become a major obstacle in employing Ad vectors for cancer gene therapy, therefore necessitating the derivation of Ad vectors capable of infecting these tumors in a CAR- independent fashion. Another reason to develop tropism-modified Ad vectors is the fact that many normal human tissues express high levels of CAR [7] and may thus become random targets for therapeutic Ad agents. As the products of some therapeutic genes may be toxic or otherwise deleterious to normal cells, such uncontrolled transduction may result in destructive side-effects which can compromise the efficiency of the therapy. An overview of the native cell-entry pathway utilized by Ad suggests that it may be modified by altering the mechanism of the virus-cell interaction. Theoretically, this goal may be achieved by modifying the structure of the receptor-binding components of the Ad virion, the fiber and the penton base, in a way which promotes interactions of the modified capsids with cell surface- localized molecules distinct from the native Ad receptors. The accomplishment of this goal would result in Ad vectors possessing expanded tropism, which would be able to achieve cell entry by either of two routes, the natural or newly created pathway. Obviously, although such infectivity-enhanced vectors would be of utility in those clinical applications where tropism to CAR is not a confounding issue, they would still be a suboptimal means of cell-specific gene delivery in most therapeutic strategies requiring stringent control over vector dissemination in patients. Therefore, in order to achieve the maximum targeting gain, the development of truly targeted Ad vectors will necessitate the ablation of the native CAR tropism of the vector. These two goals may be realized by a variety of strategies, which differ from each other in the means utilized for
engineering the novel viral tropism and the ablation of CAR tropism. In essence, there are two conceptually different approaches, which may be referred to as conjugate-based targeting and genetic targeting. These strategies are similar in that they are both based on establishing a physical link between the Ad virion and a targeting molecule, or ligand, such that binding of the ligand to a target receptor attaches the virion to the cell expressing that receptor. The basic difference between these strategies is that whereas the conjugate-based approach employs methods of complexing the Ad vector with the targeting moieties which do not usually require any modifications of the Ad virion, and results in a multicomponent vector, in 8. Targeted Adenoviral Vectors I: Transductional Targeting 2 0 9 genetic targeting no extraneous complexes or conjugates are involved as the targeting is achieved by genetic modification of the Ad virion itself, thereby resulting in a single-component vector. A variety of different types of molecules may be employed as targeting ligands in these tw ô approaches. Perhaps w îth the exception of small inorganic molecules which possess specificity to selected cell surface receptors, any natural receptor-binding ligand can be linked to an Ad capsid. This covers a w îde spectrum of ligands ranging from relatively simple organic substances such as folate to complex chemical conjugates or genetic fusions of antibodies (see "Conjugate-Based Targeting" belov^). How^ever, this spectrum of naturally available targeting moieties, although quite broad, cannot meet the needs of Ad targeting in the most general sense, as it is not applicable to cell surface molecules w^hich do not perform any receptor functions and thus do not have any natural ligands. To direct Ad vectors to this type of molecule, relevant targeting ligands should be engineered de novo. This task may be achieved by developing mono- or polyclonal antibodies against the target molecule and using these antibodies for Ad targeting. How^ever, this approach can be used only in a conjugate-based strategy, since the incorporation of an entire antibody molecule into the Ad capsid is not yet possible (see discussion below). Alternatively, a more versatile approach which is compatible with both targeting strategies may be employed for the identification of ligands. Specifically, phage libraries which are designed to display an enormous diversity of random peptides or single-chain antibodies may be utilized for the identification of the ligands of interest in a so-called "biopanning" procedure. Such biopanning usually involves several rounds of interactions between the phages constituting the library with the target, which may be represented either by purified target molecules or cells expressing these molecules or, in the extreme, the entire organism [28-34]. Each round of selection leads to the isolation of an enriched subpopulation of phage particles demonstrating some degree of binding to the target, which is then used in a subsequent round of selection. After being repeated several times, this sequential procedure normally results in the identification of ligands possessing specificity to selected targets, which may be used for the Ad rerouting strategies. IV. Conjugate-Based Targeting A. Bispecific Chemical Conjugates In order to restrict gene delivery exclusively to the target cells, it is necessary to prevent the interaction between the knob domain of the Ad fiber and its cellular receptor, CAR, which plays the major role in the determination of Ad tropism. Douglas et al. were the first to show that it is possible 210 Krasnykh and Douglas to redirect Ad infection by employing the Fab fragment of a neutralizing anti-knob monoclonal antibody (mAb) chemically conjugated to a receptor- specific ligand, in this case folate [35]. When complexed with Ad vector particles, the bispecific conjugates simultaneously ablate endogenous viral tropism and introduce novel tropism, thereby resulting in a truly targeted Ad vector (Fig. 1). In this approach, the Fab fragment is employed in preference to the intact anti-knob mAb, in order to prevent the two antigen-binding arms of the parent antibody cross-linking different viruses to form large complexes which might prove refractory to cellular uptake. Since native Ad entry is a two-step process in which the primary cellular receptor serves merely as a docking site for Ad on the cell surface, by analogy it is possible to retarget an Ad vector simply by redirecting binding to an alternative cellular receptor, with subsequent internalization mediated by the interaction between the penton base and cellular integrins. Flence, it has been possible to retarget Ad vectors by conjugating the anti-knob Fab fragment to a wide variety of ligands. In many cases, such bispecific conjugates have been demonstrated to enhance Ad-mediated gene transfer to A. B. / Figure 1 Strategy for targeting of adenoviral vectors using Fab fragment of a neutralizing anti-knob monoclonal antibody (mAb) chemically conjugated to a receptor-specific ligand. (A) Ad attachment to cells is accomplished by the high-affinity binding of the knob domain of the fiber to the primary cellular receptor, CAR. (B) When complexed with the Fab fragment of a neutralizing anti-knob antibody conjugated to a receptor-specific ligand, the Ad vector is unable to bind CAR and is directed to a novel target receptor on the cell surface. 8. Targeted Adenoviral Vectors I: Transductional Targeting 2 1 1 target cells which are refractory to native Ad infection due a low level of CAR. For example, a bispecific conjugate in which the anti-knob Fab fragment is chemically conjugated to basic fibroblast growth factor, FGF2, has been shown to mediate enhanced Ad infection of Kaposi's sarcoma cell Unes [36] and vascular endothelial and smooth muscle cells [37], which are only poorly infected by the unmodified vector. In these instances, retargeting of the Ad vector with FGF2 permits a given level of gene transfer to be achieved with a lower dose of virus. The Fab-FGF2 conjugate has also been employed to demonstrate that enhancement of Ad infection can be exploited for therapeutic advantage. To this end, it is well recognized that Ad vectors produce a dose-dependent inflammatory response in rodents and primates. Vector-associated toxicity has also been observed in human clinical trials and threatens to prevent Ad from realizing its full potential as a vector for human gene therapy. This suggests that it would be advantageous to reduce the number of Ad particles required for a given level of gene transfer in vivo. Preliminary in vitro and in vivo studies demonstrated that Ad-mediated gene transfer to the human ovarian cancer cell line SKOV3.ipl could be significantly enhanced by using basic fibroblast growth factor (FGF2) as the targeting ligand, permitting the transduction of a given number of target cells to be achieved by a lower dose of virus [38-40]. Rancourt et al. subsequently demonstrated that intraperitoneal administration of an FGF2-redirected Ad vector carrying the gene for herpes simplex virus thymidine kinase, AdTK, resulted in a significant prolongation of survival in a murine model of human ovarian cancer compared to the same number of particles of the unmodified vector [39]. In addition, the enhanced Ad infection permitted an equivalent therapeutic effect using a 10-fold lower dose of the vector. Similar results have been reported by other investigators [41, 42]. Moreover, intravenous administration of the FGF2-targeted AdTK vector led to markedly decreased hepatic toxicity and liver transgene expression compared with the untargeted vector [41, 42]. Thus, these findings suggest that strategies to enhance the efficiency of infection of recombinant Ad vectors may be of general clinical utility, by permitting therapeutically significant levels of gene transfer while minimizing the toxicity associated with high numbers of virus particles. These benefits are not limited to the field of cancer gene therapy: FGF2-mediated augmentation of gene transfer by an Ad vector encoding platelet-derived growth factor-B has been shown to enhance infection of target cells involved in tissue repair, resulting in an improved therapeutic outcome and potentially overcoming the safety and efficacy limitations of unmodified Ad vectors [43]. Bispecific moieties consisting of the anti-knob Fab fragment chemically conjugated to a mAb directed against the epidermal growth factor receptor (EGFR) have been employed to retarget Ad vectors to primary and established glioma cells and squamous cell carcinoma of the head and neck (SCCHN) 2 1 2 Krasnykh and Douglas cells [25, 44]. In addition to EGFR-specific infection, the retargeted vectors increased gene transfer to CAR-deficient cancer cells by up to 66-fold relative to the unmodified vector. Furthermore, retargeting enhanced the selectivity of Ad infection for tumor tissue relative to normal tissue from the same patient [44]. The value of retargeting Ad vectors to achieve efficient and specific gene transfer to cancer cells has further been demonstrated by means of a bispecific conjugate targeted to the pan-carcinoma antigen EpCAM [45]. The conjugation of the anti-knob Fab fragment to mAb CC49 permits enhanced Ad infection of primary ovarian carcinoma cells, which express the cognate TAG-72 receptor, while decreasing gene transfer to normal peritoneal mesothelial cells, relative to untargeted Ad [46]. In this case, the selectivity of the targeted vector for cancer versus normal cells was enhanced up to more than 200-fold relative to the unmodified vector. In the field of cancer immunotherapy, the relative resistance of CAR- deficient dendritic cells (DCs) to Ad infection has limited the application of gene-based vaccination. Using a bispecific antibody consisting of the anti-knob Fab fragment conjugated to an anti-CD40 mAb, Tillman et al, observed highly augmented Ad-mediated gene transfer to monocyte-derived dendritic cells [47]. Importantly, this efficient gene transfer was accompanied by the maturation of the DCs, resulting in an enhancement in the efficacy of DC-based vaccination against human papilloma virus 16-induced tumor cells in a murine model [48]. The anti-knob Fab fragment has also been chemically conjugated to the He fragment of tetanus toxin, permitting Ad vectors to be retargeted to neurons following intramuscular injection into mouse tongues [49]. This result further demonstrates the versatility of this targeting strategy, the aim of which is to redirect Ad binding to the surface of the target cells. The universality of this approach is further exemplified by the chemical conjugation of an anti-knob Fab fragment to a peptide with specificity for the lung endothelium, which was identified by in vivo phage display [50]. In an in vitro study, the redirected Ad vector exhibited specificity for cells expressing the target receptor for this peptide. Reynolds et al. have recently reported that a bispecific antibody consisting of the anti-knob Fab fragment chemically conjugated to a mAb directed against angiotensin-converting enzyme (ACE), which is preferentially expressed on pulmonary capillary endothelium, was able to mediate targeted Ad infection of pulmonary endothelial cells following tail vein injection into rats [51]. This shows that an Ad vector complexed with a bispecific conjugate maintains its targeting fidelity upon systemic vascular administration, a result which clearly has important and encouraging implications. Ad vectors have also been retargeted by bispecific antibodies in which the anti-Ad mAb is directed against the penton base. In this case, the full-length mAbs are employed, the combined length of which allows the retargeting moiety to extend beyond the fiber knob. Wickham et al, incorporated the 8. Targeted Adenoviral Vectors I: Transductional Targeting 2 1 3 FLAG peptide epitope into the penton base and then generated bispecific antibodies comprising an anti-FLAG mAb conjugated to a mAb specific for the target receptor. The retargeting of Ad binding to av integrins augmented gene transfer to CAR-deficient endotheUal and smooth muscle cells by 7- to 9-fold [52], while retargeting of Ad to E-selectin increased gene transfer to endothelial cells by 20-fold [53]. An anti-FLAG x anti-CD3 bispecific antibody resulted in a 100- to 500-fold increase in gene transfer to T cells, another cell type poorly infected by unmodified Ad vectors [54]. Yoon et al. have described a strategy to target Ad to human hepatocellular carcinoma cells by means of a bispecific antibody comprising the Fab fragment of an anti-hexon antibody chemically conjugated to a mAb that binds to an antigen that is highly expressed on the target cells [55]. Hov^ever, the Ad vector in this case v^as not truly targeted since it retained the ability to bind the native receptor, CAR. The relatively large number of published studies in v^hich the tropism of Ad vectors has been modified by means of bispecific moieties consisting of an anti-knob Fab fragment chemically conjugated to a ligand largely reflects the historical primacy of this strategy. The chief advantage of this approach is that a variety of ligands, including vitamins, grov^th factors, antibodies, and
peptides, can be chemically conjugated to the anti-knob Fab fragment and used to redirect Ad binding. Hov^ever, the chemical conjugation results in a heterogeneous population of molecules, w^hich presents a problem in obtaining regulatory approval for a clinical trial employing these conjugates. Moreover, the yield of appropriately conjugated bispecific molecules can be low .̂ B. Bispecific Recombinant Fusion Proteins In recognition of the disadvantages associated v^ith chemical conjugation strategies, a number of groups have generated bispecific targeting moieties in the form of recombinant fusion proteins. This permits the expression and purification of a homogenous population of retargeting molecules. The principle on v^hich the design of these bispecific proteins is based is the same as that underlying the construction of chemically cross-linked targeting agents: one site of the protein is directed against an Ad capsid protein, v^hile a second site is specific for a cell surface molecule. Again, the derivation of a truly targeted vector requires that the bispecific molecule block the binding of Ad to its native primary receptor, CAR. One class of bispecific fusion proteins used to retarget Ad vectors consists of a neutralizing anti-Ad knob single-chain antibody (scFv) genetically fused to a cell receptor-specific ligand or scFv. Watkins et al, v^ere the first to describe this type of fusion protein, for w^hich they coined the term "adenobody"[56]. They isolated a neutralizing anti-Ad5 knob scFv, designated s l l , from a phage library and then fused epidermal grow^th factor (EGF) to its C-terminal. The resultant fusion protein w âs expressed in bacteria and purified from the 2 1 4 Krasnykh and Douglas periplasmic fraction. As expected, this adenobody bound both the Ad5 knob and the EGFR on target cells, and was therefore able to redirect Ad infection via this target receptor [56]. Nicklin et al. fused this anti-Ad5 knob scFv with a heptapeptide identified by biopanning a phage display library on human umbilical vein endothelial cells (HUVECs) [57]. The fusion protein retargeted Ad infection of HUVECs, resulting in a 15-fold increase in the efficiency of transduction of these CAR-deficient cells, relative to the unmodified vector. Haisma et al. constructed a bispecific scFv by fusing the anti-Ad5 knob scFv s l l to an scFv directed against EGFR [58]. Two versions of the bispecific scFv were constructed, with the anti-EGFR scFv at either the N- or C-terminal of s l l . In this case, the bispecific scFvs were expressed in mammalian COS-7 cells and purified from the conditioned medium. Both forms of the scFv were able to retarget Ad infection via EGFR. One disadvantage of the adenobody approach is that the incorporation of different targeting moieties can have a big impact on the solubility properties on the resultant fusion protein. Hence, it can be difficult to predict whether the fusion of a given targeting scFv or ligand to the anti-Ad5 knob scFv will yield a soluble molecule which can be purified and will be functional. A second class of bispecific fusion proteins permitting the derivation of a truly targeted Ad vector consists of the extracellular domain of CAR genetically fused to a receptor-targeting moiety. By definition, once complexed with a CAR-ligand fusion protein, an Ad vector will not be able to bind to its native primary receptor on the cell surface. Dmitriev et al. genetically fused EGF at the C-terminal of the extracellular domain of CAR [59]. The soluble CAR-EGF fusion protein was expressed in insect cells using a baculovirus expression system. The bispecific fusion protein mediated EGFR-specific, CAR- independent Ad infection of target cells [59]. In a similar approach, Ebbinghaus et al. fused the extracellular domain of CAR to the Fc region of human immunoglobulin 1 [60]. When complexed with an Ad vector, this fusion protein mediated up to a 250-fold increase of transgene expression in CAR- negative. Fey receptor I-positive human monocyte cell lines. The third class of bispecific fusion protein employed to retarget Ad infection is rather different from those described above. Li et al. chose to exploit the common signaling pathways triggered by ligation of av integrins and growth factor receptors [61]. Consequently, they fused a mAb specific for the integrin-binding site on the Ad2/5 penton base to recombinant growth factors and cytokines (TNF-a, IGF-1, and EGF) which trigger the activation of phosphatidylinositol-3-OH kinase (PI3K), a signaling molecule involved in Ad internalization. The bifunctional antibodies were expressed in insect cells as secreted proteins. Ad vectors complexed with these bispecific molecules increased gene delivery 10- to 50-fold to human melanoma cells lacking av integrins [61]. Thus, whereas other strategies to retarget Ad vectors are based on redirecting Ad binding, with internalization mediated by av integrins, the 8. Targeted Adenoviral Vectors I: Transductional Targeting 2 1 5 use of bispecific molecules to trigger alternative cell signaling pathways enabled the native secondary Ad receptors to be bypassed. How^ever, it would appear that the range of receptors which could be targeted in this manner is somewhat limited and lacking in cell specificity. C. Bispecific Peptides Hong et al. have retargeted Ad infection by means of a bispecific 35- mer oligopeptide comprising two distinct peptide domains [62]. One domain was a 20-mer peptide isolated from a phage library on the basis of its ability to recognize the receptor-binding region of the Ad5 knob, while the second domain corresponded to the gastrin-releasing peptide (GRP). The authors demonstrated that the relative orientation of the two domains of the bispecific peptide was crucial: only the peptide with the GRP domain at the N- terminal was capable of mediating Ad infection in a GRP receptor-dependent manner [62]. In contrast to other targeting strategies in which the bispecific molecules were complexed with the Ad vector prior to infection, in this case the bifunctional peptide was bound to the target cellular receptors prior to the addition of Ad. It is not clear whether this approach was mandated by the small size of the oligopeptide. D. Polymer-Mediated Coupling of Ligands to Ad Capsid Proteins An alternative strategy to link receptor-specific targeting peptides to the Ad capsid involves the use of polyethylene glycol (PEG). Romanczuk et aL used bifunctional PEG to couple Ad to a peptide identified by biopanning a phage display library against differentiated, ciliated airway epithelial cells [63]. Similarly, Drapkin et a\, used PEG to couple Ad to a seven-residue peptide derived from urokinase plasminogen activator [64]. In both cases, infection of the CAR-deficient target cells by the modified vector was significantly enhanced over the unmodified vector. However, the modified vectors were still able to infect HeLa cells in a CAR-dependent manner: hence, the native tropism of the modified vectors had not been abolished. Nevertheless, now that the residues of the Ad5 knob responsible for binding CAR have been identified, a truly targeted vector could readily be generated by coupling peptides to the capsid of a virion whose knob domain had been mutagenized to prevent binding to CAR. A significant advantage of the use of PEG to couple ligands to the Ad capsid is that PEGylation partially protects the virus from neutralizing antibodies both in vitro and in vivo \(iS\. In a similar approach, Fisher et al. have shown that incorporation of targeting ligands such as basic FGF and vascular endothelial growth factor on to Ad vectors coated with a multivalent hydrophilic polymer leads to ligand-mediated, CAR-independent gene transfer to target cells \(i(i\. Importantly, the polymer-coated, retargeted vector was resistant to neutralizing antibodies. 2 1 6 Krasnykh and Douglas E. Biotinylated Ad/Avidin Bridge/Biotinylated Ligand Whereas the targeting complexes described above have comprised two components, three-component complexes have also been described. In this case, an avidin or streptavidin molecule serves as a bridge to link a biotinylated Ad vector to a biotinylated ligand. Smith et al. used this strategy to enhance Ad infection of CAR-negative primitive hematopoietic cells which express the c-Kit receptor on the cell surface [67]. Biotinylated Ad vectors were linked via an avidin bridge to the biotinylated cognate ligand for the c-Kit receptor, stem cell factor. Ad-mediated gene transfer was targeted specifically to c- Kit-positive hematopoietic cell lines, resulting in up to a 2440-fold increase in gene expression. In a second example of the versatility of this strategy. Smith et al. infected CAR-negative primary T cells with Ad vectors targeted with biotinylated antibodies to CD44 (resting and activated T cells) or with biotinylated IL-2 (activated T cells only) [67]. Kreda et al. have employed a streptavidin bridge to link a biotinylated Ad vector to a small molecule agonist of the G-protein-coupled P2Y2 receptor [68]. The tropism-modified vector was able to infect Ad-resistant, well-differentiated airway epithelia cells in a VlYi receptor-specific manner. A key attribute of this targeting strategy is that it is fairly straightforward to biotinylate a range of molecules. One disadvantage is the potential problem of scaling up the procedure, but the major limitation is the fact that three components are involved, each of which would have to meet the standards laid down by the regulatory authorities before the targeted vectors could be employed in a clinical context. Overall, conjugate-based targeting of Ad vectors possesses the advantage that major structural alterations of the Ad capsid are not involved. Conse quently, a preexisting Ad vector can be complexed with a variety of bispecific targeting molecules in order to redirect infection to a number of cell or receptor types. However, the major problem is that this targeting approach employs at least two components, vector and targeting moiety, which must be produced independently and then complexed together. This can lead to significant variation between batches of tropism-modified vector. In contrast, a single-component targeted vector can be derived by genetic modification of the Ad capsid. V. Genetic Targeting As mentioned above, this Ad targeting strategy involves genetic ablation of the virus' tropism to CAR and simultaneous engineering of an alternative receptor-binding specificity to the vector by genetic engrafting of target ing ligands into protein components of the Ad capsid. Whereas the means for accomplishing the first of these tasks was a real challenge until just a 8. Targeted Adenoviral Vectors I: Transductional Targeting 2 1 7 couple of years ago, nowadays it may be achieved rather trivially. This has become possible largely due to systematic efforts undertaken at the Brookhaven National Laboratory (BNL) and GenVec, Inc., to identify those amino acids in the Ad fiber knob domain w^hich mediate the interaction v^ith CAR. Specif ically, the BNL team used recombinant forms of the Adl2 fiber knob and the Dl domain of CAR protein to obtain crystal structures of knob-Dl com plexes [69]. The resultant three-dimensional model of the complex has allow^ed the identification of CAR-binding residues v^ithin the knob. The importance of those amino acids in CAR recognition by Ad v^as then confirmed by the genera tion of Ad virions incorporating mutated fiber proteins. The same goal has been reached by Roelvink et aL in a large-scale project on Ad5 knob mutagenesis v\rhich w âs rationalized by the identification of amino acid residues conserved in the fiber knobs of CAR-binding human Ad serotypes [70]. Although these groups utilized totally different approaches, the results of the tw ô studies corroborate each other quite nicely. It has been shov^n that the CAR-binding site w^ithin the Ad5 knob domain involves the AB- and DE- loop, as v\̂ ell as P-strands B, E, and F (Fig. 2, see color insert). Mutations of the key amino acids identified in these studies quite dramatically decrease the capacity of recombinant knob proteins or complete Ad virions bearing such mutations to bind CAR, thereby providing a simple means to ablate native Ad tropism to CAR. While this knob mutagenesis represented a universal approach to CAR ablation, engineering of novel Ad tropism may be achieved in a variety of distinct strategies. These targeting maneuvers are distinguished by genetic modifications of different components of the Ad capsid. It has been demon strated that genetic modifications of the three major proteins constituting an Ad virion, the hexon, the fiber and the penton base, may be employed to redirect the virus. A. Ad-Targeting Strategies Involving Genetic Manipulations of the Fiber Protein Logically, as the fiber normally plays the role of primary attachment of the virion to the cell, the majority of Ad targeting efforts have been based on fiber modification. To date, at least three distinct strategies of Ad targeting involving fiber modification have been used:
fiber or knob shuffling, fiber modifications via the incorporation of targeting ligands, and fiber- or knob-replacement strategies. 1. Fiber- and Knob-Shuffling Approaches This is the most obvious and straightforv^ard strategy, w^hich is based on the overall structural similarity of the fibers of different Ad serotypes and is further rationalized by the fact that representatives of these serotypes 2 1 8 Krasnykh and Douglas use different receptors to infect permissive cells. In this regard, the high degree of homology between fiber tail domains suggests that the penton base protein of a given Ad serotype may associate quite efficiently with fibers derived from different serotypes. Furthermore, the junctions between the tail and shaft, and shaft and knob, domains within the fiber protein represent convenient fusion points in those instances when a mosaic fiber protein incorporating individual domains derived from different Ad serotype fibers is designed. These considerations have resulted in a number of studies employing Ad vectors whose fibers have been either completely replaced with heterologous fiber proteins or modified to contain knobs originating from other Ad serotype fibers. The feasibility of Ad retargeting via fiber shuffling was demonstrated in pioneering work by Gall et al. who showed that by replacing the fiber gene in Ad5 genome with the gene from Ad7, a chimeric Ad vector incorporating Ad7 fibers may be generated [71]. By using two types of competition assays, the authors demonstrated that this fiber substitution resulted in the alteration of the tropism of the Ad vector. It was thus shown that exchange of the fiber is a strategy that may be used to manipulate native Ad tropism. This approach was later utilized in collaborative work conducted by researchers at the University of Iowa and Genzyme Corp. to develop better Ad vectors for human gene therapy. To this end, by screening a number of Ad serotypes for an enhanced ability to infect either the well-differentiated ciliated human airway epithelia (CHAE) or fetal rat central nervous system (CNS) cells, the authors identified Ad 17 as one of the best agents for transduction of both types of cell targets [72, 73]. Based on these data, an Ad2 vector was then genetically modified to contain the Ad 17 fiber in place of the native Ad2 fiber. The resultant virus possessed the receptor specificity and infection capacity of Ad 17. Specifically, the levels of reporter gene expression directed by this modified vector in CHAE cells were 15- to 95-fold higher than those achieved by the parental Ad2 vector [73]. Additionally, this tropism-modified vector was sevenfold more efficient than its unmodified Ad2 counterpart in transducing CNS cells in vitro [72]. These findings provide a rationale for the utilization of similar Ad vectors incorporating Adl7 fiber proteins for gene therapy approaches to the treatment of cystic fibrosis and CNS disorders. The poor transducibility of human hematopoietic stem cells (HSCs) by commonly used Ad vectors was addressed by the utilization of the fiber swapping approach in work by Shayakhmetov et al. [22]. An analysis of the binding and internalization of six different Ad serotypes (3, 4, 5, 9, 35, and 41) performed on CD34^ cells identified Ad35 as the best overall vehicle for transduction of HSCs. The subsequent generation of an Ad5 vector incorporating the Ad35 fiber protein and evaluation of its transduction capacity on CD34+ cells showed that this vector was several-fold more efficient than a control vector incorporating Ad5 fibers. 8. Targeted Adenoviral Vectors I: Transductional Targeting 2 1 9 Another example of the successful use of the fiber shuffling strategy is the result of extensive studies originated at Crucell Holland (Leiden, The Netherlands). A panel of Ad5-based vectors incorporating fibers from other serotypes v^as employed in order to identify those Ad fibers which v^ould be the most efficient in mediating infection of human DCs and cardiovascular or synovial tissues [22, 74-76], An initial screening resulted in the Ad35 fiber being the most efficient in directing Ad binding to DCs, while synovial and cardiovascular tissues were most sensitive to a chimeric vector containing Adl6 fibers. A chimeric Ad5/35 vector proved to be 100-fold more potent than an Ad5 control for gene transfer and expression in human DCs [76]. However, no maturation of the DCs was observed, in marked contrast to the studies in which Ad was targeted to CD40 [47]. In a parallel study, an Ad5 vector carrying the fiber of Adl6, showed on average, an 8- and 64-fold increase in the reporter gene expression in endothelial and smooth muscle cells, compared to the parental Ad5 vector [75]. Therefore, these findings suggest that Ad5/16 may serve as a prototype for the generation of derivative vectors carrying relevant therapeutic transgenes, which may be quite efficient as a means to treat cardiovascular disorders. According to a study by Goosens et al.^ the same vector platform shows promise as an efficient system for gene delivery to synoviocytes which do not express CAR and thus are refractory to infection by Ad5-based vectors [74]. Specifically, the authors demonstrated that an Ad5/16 vector was more potent in transducing cultured synoviocytes compared to Ad5. An observed 150-fold increase in transgene expression was caused by both the transduction of a higher percentage of synoviocytes and higher level of gene expression per transduced cell. An approach similar to fiber shuffling was realized in studies which employed genetic replacement of the Ad5 fiber knob domain with the Ad3 knob as a way to achieve the goal of Ad5 rerouting via an alternative Ad receptor [77^ 78]. In these reports, the retargeting of the vector to the, as yet unidentified, Ad3 fiber receptor suggested the utility of this knob-swapping approach in those instances where preferential expression of the Ad3 fiber receptor favors infection of Ad5-refractory cells. For instance, the vector- containing chimeric fibers outperformed the control vector incorporating Ad5 fibers in gene delivery to human fibroblasts and head and neck cancer cells [78]. Similarly, in their efforts to develop a gene delivery system for treatment of lymphoproliferative disorders. Von Seggern et al. also derived an Ad vector containing Ad5/Ad3 fiber chimeras and demonstrated its superior gene transfer properties on Epstein-Barr virus-transformed B lymphocytes [79]. Despite their attractive simplicity, the fiber- and knob-shuffling approaches are limited in their utility for gene therapy and may hardly be viewed as universal strategies with the potential to direct Ad vectors to any given type of cell target. This is due to the fact that the repertoire of target receptors is dictated by the diversity of targeting ligands available. However 2 2 0 Krasnykh and Douglas significant the transduction efficiency gains achieved in the pubfished studies, they cannot be further improved after all available Ad serotypes have been tested in the cell system of interest and the best overall candidate fiber has been identified. Whereas in the context of the fiber- or knob-shuffling approaches an investigator is strictly limited in the capacity to improve upon the vector specificity, this problem may be overcome by expanding the repertoire of targeting ligands beyond the limits of the natural diversity of Ad fibers. This concept is realized in strategies of Ad targeting based on the incorporation of targeting ligands into the Ad fiber protein or fiber-replacing molecules. 2. Ad Fiber Modifications The most significant advantage of this approach compared to the fiber- or knob-shuffling strategies is that it potentially permits the utilization of a wide range of ligands for Ad targeting. As mentioned above, these may be the natural ligands for target receptors, or alternatively they may be identified via the use of phage display technology. From a theoretical standpoint, if these tw ô methodologies are utilized in a rational manner, a virtually unlimited number of targeted Ad vectors may be derived. There are, how^ever, practical considerations v^hich may limit the usefulness of this approach. Evidently, the genetic fusion of a targeting ligand w îth the fiber protein should be accomplished in a v^ay w^hich would retain a functional configuration of both fusion partners. Therefore, the specific site in the fiber protein chosen for ligand insertion should not play any vital function which could potentially be abrogated by the ligand insertions. Furthermore, the configuration of such a site should favor the presentation of a ligand on the surface of the fiber molecule, thereby facilitating its interaction with the cognate receptor. In addition, the architecture of an Ad virion in general and the receptor-binding function of the knob in particular imply that of the three structural domains of the fiber protein, the knob is the most logical locale for presentation of targeting ligands. Thus, the configuration of the knob may apply additional limitations on the design of functional fiber-ligand fusions. Although these consideration are quite obvious, they were of little use at the time when the first attempts to modify the fiber were undertaken, as the work was initiated well before the three-dimensional model of the fiber knob domain became available. Modification of the Ad5 fiber protein was first endeavored by Michael et al.^ who demonstrated that a short peptide ligand, the gastrin-releasing peptide (GRP), genetically fused to the C-terminal of a recombinant fiber protein does not interfere with fiber trimerization and is available for binding with a GRP-specific antibody [80]. In the absence of any structural information about the knob domain, the choice of the C- terminal was rationalized by previous findings that the N-terminal of the protein is embedded in the capsid, while the C-terminal is exposed outside 8. Targeted Adenoviral Vectors I: Transductional Targeting 2 2 1 the virion. Wickham et al. subsequently derived an Ad vector incorporating a fiber extended at the C-terminal v^ith a stretch of lysine residues in order to target this vector to heparin-containing cellular receptors [17]. This tropism expansion maneuver resulted in a vector which was able to efficiently transduce a number of Ad5-refractory cells, thereby establishing the feasibility of this genetic approach to Ad targeting. Specifically, the transduction capacity of this new vector on endothelial cells, glioblastomas, smooth muscle cells, and fibroblasts was 9- to 311-fold higher than that of unmodified control vector bearing wild-type Ad5 fibers. The same group of investigators subsequently redesigned the oligo-lysine modified vector by making AdZ.F(pK7), a vector containing seven C-terminal lysine residues, and also constructed a similar vector, designated AdZ.F(RGD), incorporating an RGD-containing peptide at the C-terminal of the fiber [18]. The latter vector was designed to target av-integrins expressed by a number of cell targets whose infection with unmodified Ad vectors is inefficient. Evaluation of these two vectors on a panel of cells which are poorly infected with Ad5 shbwed that they both were equally efficient in transducing endothefial cells, while AdZ.F(pK7) clearly outperformed AdZ.F(RGD) on smooth muscle cells and macrophages [18]. The highly augmented transduction efficiency of AdZ.F(pK7) on smooth muscle cells in tissue culture encouraged the authors to employ this vector to transduce pig iliac arteries injured by a balloon catheter. As expected from the data obtained in the in vitro studies, AdZ.F(pK7) proved to be more efficacious than its unmodified counterpart in delivering a transgene to the target tissue. This proof of concept work led to a number of spin-off studies utilizing AdZ.F(pK7) and AdZ.F(RGD) for gene delivery to a variety of CAR-deficient tissues. For instance, AdZ.F(RGD) showed superior in vivo gene transfer to the cortical vasculature in rats [81], whereas AdZ.F(pK7) proved to be useful for transduction of muscle cells at all stages of differentiation as well as mature skeletal muscle [82], myeloma cells [83], myeloid leukemic cells [84], and malignant glioma cells [85]. However, the magnitude of enhancement achieved by AdZ.F(pK7) in vivo was somewhat less than might have been anticipated from the in vitro augmentation. A series of Ad vectors similar to AdZ.F(pK7) designed by Hamada's group was used in extensive work aimed to develop gene therapy for gliomas. All the vectors used in these studies were designed on an Ad5 platform and incorporated a fiber protein with a stretch of 20 lysine residues fused via a peptide linker to the C-terminal of the fiber. The employment of this prototype vector for delivery of genes encoding cytokines IL-2 and IL-12 [86], the p53 tumor suppressor [86], prodrug-converting herpes simplex virus thymidine kinase [87], or a conditionally replicative Ad genome [88] has led to significant improvements of transgene expression and Ad-mediated killing of glioma cells. Despite the fact that the utility of C-terminal
modifications of the fiber for Ad targeting has been clearly demonstrated, it has also been reported that some 2 2 2 Krasnykh and Douglas of the attempted modifications employing rather long targeting peptides did not result in viable Ad virions [18]. Moreover, fiber modification studies performed by Hong and Engler shov^ed that the addition of a 27-amino-acid residue peptide sequence to the C-terminal of the fiber resulted in a protein incapable of assembly into trimers [4]. This led to the hypothesis that ligands exceeding a threshold of about 30 amino acids in length cannot be successfully incorporated into this locale in the fiber protein w^ithout detrimental consequences for the structure of the fiber. Evidently, these findings w^ould limit the applicability of the fiber modi fication approach should the C-terminal of the fiber be the only locale within the molecule suitable for the presentation of ligands. Fortunately, the three- dimensional model of the Ad5 fiber knob proposed by Xia et al, [89] provided the rationale for the evaluation of other sites w îthin this domain for the purpose of ligand incorporation. According to this model, the loops which connect the P-sheets R and V within each fiber knob monomer may function well as ligand- presenting structures. These loops are localized on the surface of the knob and are thus readily accessible for interactions with potential receptors (Fig. 2). The flexibility of the loops suggests that ligands incorporated within the loops would be able to assume the proper configuration required for the interac tion with their cognate receptors. The loops are not involved in intramolecular interactions; therefore, modifications of their structure should not affect the sta bility of the fiber. Additionally, alignment of the primary sequences of the knob domains of various Ad serotypes reveals that the length and amino acid compo sition of the loops varies quite significantly, implying that the incorporation of targeting ligands into the loops would be well tolerated by the knob structure. These considerations encouraged Krasnykh et ah to conduct a proof of concept study aimed to show the feasibility of Ad targeting via genetic modification of the loops within the fiber knob [90]. This was achieved by genetic incorpora tion of a ligand-mimicking octapeptide FLAG tag into the HI loop of the knob domain. First, it was shown that a fiber protein incorporating the FLAG tag expressed in baculovirus-infected insect cells retains its native trimeric config uration and binds to an affinity matrix containing an anti-FLAG antibody. It was then shown that incorporation of the peptide did not affect the ability of the modified fiber to bind to CAR-positive cells. Hence, the two key features of the fiber, trimerization and receptor binding, were both preserved in the fiber- FLAG protein, thereby rationalizing the generation of an Ad vector containing such fibers. A virus containing chimeric fiber-FLAG proteins was shown to be viable and was able to infect cells via CAR-mediated pathway. Importantly, the FLAG peptide engrafted in the virion retained its ability to bind an anti-FLAG mAb, suggesting that a peptide of a similar size possessing targeting properties should function well in the context of the HI loop of the fiber. This concept was proved in a subsequent study by the same team of investigators [91], who derived an Ad vector, Ad5lucRGD, incorporating 8. Targeted Adenoviral Vectors I: Transductional Targeting 2 2 3 within the HI loop of the fiber knob the targeting peptide CDCRGDCFC, known as RGD-4C. This Hgand was chosen for Ad targeting based on its small size and well-documented ability to bind quite efficiently with a number of cellular integrins, which were used as target receptors in these studies. Therefore, the rationale behind this work was to target the virus directly to integrins, which normally function as the secondary Ad receptors. It was expected that the resultant virus would thus possess the capacity to enter the cell via a CAR-independent, integrin-mediated pathway. The employment of the resultant virus for gene transfer to a panel of cell targets expressing various levels of CAR and integrins proved that the HI loop-localized RGD-4C peptide was indeed able to direct the vector to integrins, thereby expanding its tropism. Importantly, it was demonstrated that as a result of this tropism expansion, the virus gained a significant advantage over its unmodified counterpart in transducing cell targets which are low in CAR expression. This infectivity enhancement was most dramatic when both viruses were applied to primary tumor cells isolated from patients with ovarian cancer. In this experiment, the reporter gene expression detected in Ad5lucRGD-transduced cells was two orders of magnitude higher than that in the cells infected with the control vector. The superior transduction efficiency of Ad5lucRGD was subsequently exploited in more extensive studies in established cell lines and primary samples of ovarian tumors 192]. Furthermore, it was later demonstrated that this vector offers another advantage in the treatment of ovarian cancer by circumventing the inhibition of Ad infection of tumor cells by neutralizing anti-fiber antibodies present in ascitic fluids in the patient's peritoneum [93]. The successful utilization of AdilucRGD to augment gene transfer to squamous cell carcinoma of head and neck cell lines [94], myelomonocytic leukemia cells [95], rhabdomyosarcoma cells [96], and glioma cells [97] by up to three orders of magnitude suggests that this vector may serve as a prototype for the derivation of agents suitable for gene therapy of various types of cancer, where CAR deficiency of target tumors undermines the efficacy of unmodified Ad5 vectors. A recent report by Asada-Mikami et al. showed another use for Ad vectors based on the AdSlucRGD platform by demonstrating the improved capacity of this vector to transduce dendritic cells [98]. A further demonstration of the utility of HI loop modifications for Ad tar geting, as well as the compatibility of this approach with the ligand definition strategy based on phage library biopanning, was provided by Xia et al. [99]. In order to design an Ad vector suitable for transduction of brain microcapillary endothelium (BME) via the transferrin receptor (TfR)-mediated pathway, these investigators screened a phage library displaying linear, nonconstrained non- apeptides on a recombinant form of the extracellular domain of human TfR and isolated a total of 42 phage clones demonstrating significant binding to 2 2 4 Krasnykh and Douglas this target. Ten of these peptides were then incorporated into the HI loop of the fiber as potential targeting ligands. Notably, the authors succeeded in rescuing only 7 of the 10 viral vectors; moreover, 2 of the rescued vectors w êre only amplified quite poorly on 293 cells. The rescued vectors v^ere then employed for gene transfer to CAR-negative cells expressing TfR. These experiments show^ed that by using TfR as a primary binding receptor, the peptide-modified Ad vectors w êre 3- to 34-fold more efficient than unmodified vectors in trans ducing the target TfR-positive cells. The successful use of these vectors for gene delivery to human BME cells proved the suitability of these agents for the purposes of gene therapy of inherited metabolic disorders causing malfunction of the central nervous system. Since the w^ork v^ith Ad vectors incorporating fibers w îth C-terminal modifications revealed the limitations relating to the size of the figands w^hich could be incorporated at this locale, it w^ould be logical to address this ligand size issue in the context of HI loop modifications of the fiber. In this regard, a pilot study using incremental increases in the size of this loop via the incorporation of heterologous protein sequences v^ith the subsequent characterization of the yields and infectivities of the resultant vectors v^ould be of high relevance. This task has been achieved in a recent study by Belousova and Krasnykh [100], w ĥo generated a panel of Ad vectors incorporating w îthin the HI loop a series of fragments of the RGD-containing loop of the Ad5 penton base protein. The results of this study shovŝ ed that heterologous protein sequences up to at least 83 amino acids long may be incorporated into the HI loop of the fiber protein v^ îthout any significant negative consequences on the viability, yield, and infectivity of the resultant vector. How^ever, these studies also revealed that the stepw^ise increments in the insert size have some adverse effects on these properties of the vector. These findings led to the conclusion that although the capacity of the HI loop to accommodate targeting ligands of moderate size is superior to that of the C-terminal of the fiber, ligands vŝ hose size exceeds a certain limit v îll not fit into this loop. Additionally, not only the size but also the configuration of a targeting ligand may become an issue w^hen the HI loop is used for ligand presentation. For instance, the functional configuration of some ligands may conflict with the framework of the HI loop. This conflict may arise when the C-terminal of a ligand needs to be directly involved in receptor binding: if this is the case, then the covalent bond between the ligand and the loop would prevent the efficient interaction of the modified Ad vector with the target cell. These considerations, together with the problems with the rescue and propagation of the Ad vectors observed in the work by Xia et al. [99], suggest that in order to increase the likelihood of the successful generation of fiber- modified Ad vectors, it will be necessary to develop an approach which would allow for the easy and fast high-throughput evaluation of newly identified targeting ligands in the context of the Ad fiber knob. The development of such 8. Targeted Adenoviral Vectors I: Transductional Targeting 2 2 5 a method would streamline the generation of targeted Ad vectors by narrowing the range of candidate ligands to those which have maximum probability of functioning upon engrafting into the Ad capsid. Perhaps the most rational way to solve this problem would involve the generation of ligand display libraries using Ad as the vector. In such libraries, a variety of ligands would be randomly incorporated into specific locales of the Ad fiber protein and, upon rescue, the resultant diverse repertoire of modified Ad vectors would be screened for specificity for the receptors or cells of interest. However, at the present stage of development of the technology for the generation and characterization of recombinant Ad vectors, the derivation and screening of such a library presents a challenge of enormous complexity and is hardly achievable. Nevertheless, recent advances in the field have shown that there may exist some alternative approaches potentially useful in the rationalization and facilitation of the selection of ligands for Ad targeting via fiber modification. A first step in this direction was taken by Jakubczak et al. who described a method which allows the fast and easy generation of fiber-modified Ad vectors without the need to construct recombinant viral genomes and rescue recombinant virions [101]. This approach capitalizes on two key advances previously made in the Scripps Research Institute by Von Seggern et al. [79, 102]. First, the strategy takes advantage of a plasmid vector designed to express high levels of Ad fiber protein in eukaryotic cells [79]. This plasmid is used to direct the expression of the candidate modified fiber protein. Second, a helper Ad vector containing a fiber gene-deleted genome packaged into the wild type Ad5 capsid by propagation in the fiber-complementing cell line (for details see below) is used to produce fiberless Ad capsids into which the plasmid-encoded modified fiber proteins will be incorporated. The method works in the following manner: (i) a candidate fiber gene is cloned into the expression plasmid; (ii) the plasmid is then used for transfection of 293 cells, resulting in the expression of the fiber protein of interest; (iii) the transfected cells are infected with wild-type Ad virions encapsidating fiber gene-deleted genomes. If the fiber protein expressed by the plasmid retains the configuration necessary for its efficient incorporation into an Ad capsid, it transcomplements the deletion in the genome of the helper virus and is incorporated into Ad virions. The resultant virions may then be subjected to screening on a target of interest. Although the throughput of the system cannot match that of the phage display library, this strategy provides an excellent means for the generation of genetically modified fibers and their preliminary characterization in the context of complete Ad particles. Another promising approach addressing the same
issue of the ligand-fiber compatibility was developed in a recent study by Pereboev et al. [103]. The rationale for this work is to expand the utility of a traditional phage display system for the identification of targeting ligands by tailoring its format such that 2 2 6 Krasnykh and Douglas it closely mimics the ligand presentation by the Ad fiber protein. In contrast to the classical approach, where randomized ligands are genetically incorporated into one of the phage coat proteins, in this novel strategy a diversity of targeting ligands is created within the Ad5 fiber knob domain attached to the surface of the phage particle by Jun and Fos leucine zippers. Therefore, the subsequent screening of the library leads to the identification of ligands which demonstrate the ability to bind to a target receptor while engrafted directly in the fiber knob. This approach thereby counterselects against those ligands which could be identified in a biopanning experiment employing a traditional phage library, but would then fail to recognize the target upon incorporation into the Ad fiber. Further development of this proof of concept study should result in the derivation of phage libraries which would meet the most stringent criteria of selection for Ad targeting ligands. 3. Fiber- and Knob-Replacement Strategies While the new approaches developed to facilitate the identification of fiber-compatible targeting ligands may be very efficient and useful, they cannot solve the problem of the structural incompatibility of the ligand and the fiber. Furthermore, extensive use of these methods may soon show that a significant proportion of ligands is not suitable for fiber modification because of serious structural conflicts between the fiber and the ligand. The high rate of failures in documented attempts to incorporate a targeting ligand into the rather complex framework of the fiber knob domain makes these expectations quite real. Therefore, it is rational to hypothesize that the tropism of the Ad vector might be manipulated more easily and much more efficiently, if the receptor binding function in the resultant fiber molecule were structurally disengaged from the trimerization function. This goal may be achieved by generating a fiber-derived molecule in which the functions of receptor recognition and trimerization would be delegated to distinct domains within the protein. Therefore, the trimerization of such a fiber would be secured by a protein moiety introduced into the design of the "platform" fiber and would not be affected by subsequent modifications of the other domain of the chimera, which would define the vector tropism. The practical realization of this strategy would dramatically expand the range of targeting ligands compatible with the fiber modification strategy, thereby diversifying the repertoire of target receptors and target cells. Obviously, the knob domain, whose complex structure seems to restrict the range of targeting ligands and targeting approaches, should be deleted from the resultant protein. In a more general sense, only the fiber tail domain which anchors the fiber in the capsid and is thus indispensable, should be retained in the modified protein, implying that the shaft of the fiber may be replaced too. In order to maintain the trimeric structure of a knob-deleted protein, the loss of the trimerization function normally provided by the knob should be compensated by the incorporation of a heterologous protein moiety 8. Targeted Adenoviral Vectors I: Transductional Targeting 2 2 7 into this protein. This may be achieved by designing a knob-deleted fiber containing a protein or peptide motif known to form stable homotrimers upon self-association. Two recent studies illustrate this general concept. The work by Krasnykh et al. [104] involved the generation of an Ad vector whose capsid lacked wild-type fibers but instead incorporated chimeric molecules designed to fulfill the functions normally performed by the Ad fiber. These chimeras comprised the N-terminal portion of the Ad5 fiber protein, including the tail domain and two pseudorepeats of the shaft domain, genetically fused to the bacteriophage T4 fibritin protein deleted at the N- terminal. The entire knob domain and most of the shaft domain of the fiber were deleted in this protein. Truncated fibritin was incorporated into this chimeric protein in order to provide trimerization of the resultant molecule. Previous studies on the structure of fibritin showed that, owing to the presence of the C-terminal "foldon" domain, fibritin is capable of forming homotrimers which are extremely stable under a variety of different conditions. Most importantly, it had also been demonstrated that the trimeric structure of fibritin is not compromised by either extensive N-terminal deletions or extensions of its C-terminal, thereby making fibritin an ideal "stuffer" between the fiber tail and the targeting ligand positioned at the C-terminal of the resultant chimera. A sequence of six histidine residues connected to the fibritin protein via a short peptide linker was used to demonstrate the feasibility of targeting of fibritin- containing Ad vectors to alternative cell-surface receptors by directing the modified vector to an artificial receptor, whose extracellular domain consists of an anti-His scFv [105]. First, it was shown that the fiber-fibritin 6H chimera expressed in bacteria is trimeric and binds to affinity matrix via the 6His ligand present in the protein. The subsequent incorporation of this chimera into the Ad5 capsid resulted in a stable virion capable of infecting target cells expressing the complementary receptor in a CAR-independent manner. Although fiber replacement did not affect the stability of the virus or its yield, the efficiency of infection by the new virus was lower than that by the control Ad vector containing wild-type Ad5 fibers. This decrease in infectivity may be due to the previously reported low affinity of interaction between the targeting ligand, 6His, and the ligand-binding scFv component of the artificial receptor used in the study. Fortunately, the ligand-accommodating capacity of the described fiber-fibritin chimera extends well beyond the 6His sequence, thereby allowing utilization of a wide variety of targeting ligands in the context of this prototype molecule. To this end, it has been previously demonstrated that C-terminal insertions up to at least 163 amino acids long are well tolerated by the fibritin structure [106]. A similar study was reported by Van Beusechem et al. [107], who attempted to replace the fiber knob domain with trimerizing a-helical coiled- coil domain of the Moloney murine leukemia virus p i5 envelope protein. Since p l5 is known to have low thermostability, the resultant fiber chimeras were 2 2 8 Krasnykh and Douglas rather unstable with only 5-10% of the protein being assembled into trimers. Expression of these proteins by the El-deleted Ad vectors, which also expressed the wild-type fiber, demonstrated their nuclear localization and some degree of incorporation into complete Ad virions. Importantly, the knob- or fiber-replacement strategies allow for the simultaneous fulfillment of the two tasks required for the generation of truly targeted Ad vectors: ablation of native Ad tropism to CAR and introduction of novel tropism. Equally importantly, these approaches may be applied for the generation of truly targeted vectors derived from any Ad serotype without the prior identification and ablation of the receptor-binding site within a given fiber. Although the studies described herein have demonstrated only the fea sibility of the fiber- and knob-replacement approaches for Ad targeting, the further development of these novel technologies may eventually result in significant improvements to the utility of the present generation of Ad vectors. 4. Strategies to Rescue and Propagate Truly Targeted Ad Vectors Regardless of the approach chosen for the derivation of truly targeted Ad vectors, the resultant virus should somehow be rescued and propagated. This constitutes a serious technical problem, as such a vector is, by definition, not able to infect cells via the native cell entry pathway and thus cannot be amplified in any of the cell lines normally used for this purpose. It would theoretically be possible to solve this problem on a case-by-case basis by first deriving an El-complementing cell line expressing the target receptor and then using this cell line for the rescue of the targeted Ad vector of interest. However, the practical execution of this strategy would be extremely laborious, cumbersome and thus highly inefficient. In addition, this approach would not work in those instances when an Ad vector targeted to unknown cell surface molecule was being derived. Therefore, a universal solution to the problem would be highly desirable. At least two distinct approaches have been developed to address this problem. The first strategy involves the generation of a packaging cell line which expresses the wild-type Ad fiber and may be used in the rescue and initial propagation of truly targeted Ad vectors (Fig. 3). Ideally, this cell fine should be designed to express an Ad fiber whose receptor is naturally produced by this cell line. The utilization of such a cell line will result in the production of mosaic Ad virions randomly incorporating both wild-type and modified fibers. The presence of the wild-type fiber in the resultant virions will allow the efficient infection of any cell line expressing native Ad fiber receptor, including the packaging line. After a sufficient amount of mosaic virus has been generated, this vector may then be converted into the truly targeted configuration by a final amplification step on a cell line which does not express any Ad fiber, for 8. Targeted Adenoviral Vectors I: Transductional Targeting 229 Infection Figure 3 Strategy employing a packaging cell line which expresses the wild-type Ad fiber to enable the rescue and initial propagation of truly targeted Ad vectors. In this approach, 211B cells, a derivative of 293 cells which constitutively express the Ad5 fiber protein, are transfected with an Ad5 genome containing a modified fiber gene. This results in the production of mosaic Ad virions randomly incorporating both wild-type and modified fibers. The presence of the wild-type fiber in the resultant virions allows the efficient infection of any cell line expressing native Ad fiber receptor, including the packaging line. After a sufficient amount of mosaic virus has been generated, this vector may then be converted into the truly targeted configuration by a final amplification step on a cell line which does not express any Ad fiber, for example 293 cells. example 293 [108], 911 [109], or PerC6 [110]. The feasibility and utility of this approach was first demonstrated by two groups of investigators [102, 111] who designed derivatives of the El-complementing 293 cell line, designated 211 and 293-Fib, respectively, which constitutively express the wild-type Ad5 fiber protein to enable the propagation of fiber-deleted Ad vectors. Both cell lines allowed the efficient rescue and amplification of Ad virions containing fiber gene-deleted genomes. Therefore, this strategy is quite efficient and may be used for the generation of Ad targeting vectors derived from virtually any Ad serotype. The only drawback of this approach, although purely hypothetical at this point, is that homology between the modified fiber gene contained in the genome of the targeted Ad vector and the wild-type fiber gene incorporated into the genome of the packaging cell may result in recombination and restoration of the wild-type fiber gene in the viral genome, thereby negating the whole targeting effort. In an alternative strategy, the targeted Ad vector is designed to have two different tropisms. One of the receptor specificities engineered into the virion provides virus binding to an artificial receptor, which is expressed by the 230 Krasnykh and Douglas A. B. Wild type Targeted Targeted Ad vector Ad vector Ad vector Cell expressing artificial receptor Target cell Figure 4 Strategy for utilization of an artificial receptor to provide a CAR-independent pathway of cellular entry to enable propagation of truly targeted Ad vector lacking native tropism. (A) In contrast to the wild-type vector, targeted Ad cannot utilize CAR for cell attachment and, therefore, achieves cell entry during rescue and amplification in cell culture via binding to artificial receptor. This attachment is mediated by a propagation ligand (five-pointed star) incorporated into the Ad capsid. (B) For gene delivery to target cells, the vector employs a targeting ligand (four-pointed star), which recognizes cognate receptor molecule distinct from the artificial receptor. Reproduced in modified form with permission from [141]. correspondingly modified packaging cell line (Fig. 4). This feature of the vector is only used during vector amplification in the laboratory. The other receptor specificity defines the vector tropism in the context of targeted gene delivery for gene therapy purposes. Therefore, the resultant vector contains tw ô targeting ligands of v^hich
one (the "propagation ligand") is needed during vector rescue and amplification, w^hereas the second (the "targeting ligand") directs the vector to a receptor naturally expressed by the target tissue. Importantly, the receptor chosen to facilitate Ad propagation should be entirely artificial and should not have any natural analogs. Othervs^ise, the propagation ligand may compromise the truly targeted status of the vector by randomly binding in vivo to receptors analogous to the artificial receptor expressed by the packaging cells. This approach has been realized by Douglas et al. [105], w ĥo generated a number of cell lines expressing an artificial receptor (AR), w^hich binds proteins containing C-terminal 6His tags. This receptor v^as designed by genetically fusing an anti-6His single chain antibody, scFv, v^ith the transmembrane domain of the platelet-derived grov^th factor receptor. The functional utility of this receptor w âs demonstrated by constructing an Ad vector containing 6His tag at the C-terminal of the fiber and using this vector for gene delivery to AR-expressing cells as w êll as to the parental cells lacking the AR. These studies shov^ed that the 6His-modified Ad vector can efficiently infect AR-expressing cells in a CAR-independent fashion. 8. Targeted Adenoviral Vectors I: Transductional Targeting 2 3 1 A similar system employing a hemagglutinin (HA) epitope incorporated into the Ad capsid and a complementary receptor embodying an anti-HA scFv was generated by Einfeld et al. and used for the rescue and amplification of Ad vectors lacking tropism to CAR [112]. The disadvantage of the approach using artificial receptor-expressing cell lines is that it requires additional modification of the Ad vector to carry a propagation ligand. How^ever, this is just a minor hmitation of an otherw^ise very efficient strategy, as the propagation tag may first be incorporated into the capsid of the prototype Ad vector, v^hich can subsequently be used for the generation of various targeted derivatives. B. Ad-Targeting Strategies Involving Genetic Manipulations of the Hexon and Penton Base Proteins The genetic targeting of Ad virions via modifications of the hexon or penton base proteins has not been studied as extensively as fiber-based targeting. This is primarily due to the fact that these proteins do not play a significant role in the attachment of Ad to the primary cellular receptors. Therefore their utility for Ad targeting is not immediately apparent. However, a number of considerations imply that attempts to modify Ad vector tropism by alterations to these proteins may be successful. Although the accessibility of both proteins for binding to a putative target receptor may be an issue because of structural interference with the fiber protein, it may be significantly improved by shortening the shaft of the fiber by genetic means, thereby reducing the length of the entire fiber protein. The direct binding with cellular integrins of the penton base protein of human Ad9 [113], whose fibers are significantly shorter than those in Ad5 or Ad2 virions, suggests that shortening of the fiber may be a general strategy to facilitate interactions between the penton base and a cellular receptor. The fact that Ad2 binds to P2 integrins via its penton base protein [114] indicates that under certain circumstances the shortening of the fiber is not even required for such direct binding to occur. This is further supported by the finding by Einfeld et al. that a peptide ligand incorporated into the RGD-containing loop within the penton base of Ad5 vector binds quite efficiently to an artificial receptor expressed by the target cell [112]. It has also been previously reported that genetic modifications of recombinant penton base proteins result in alterations of their binding specificity [115]. Although the hexon protein does not play a documented role in the cell entry pathway used by Ad, its abundance in the Ad capsid makes the hexon a very attractive candidate as a ligand-presenting molecule. In addition, a comparison of the amino acid sequences of several known hexons reveals the presence of a number of hypervariable regions (HVRs) in these otherwise highly conserved proteins [116]. There are significant differences in the length and amino acid composition of these regions, strongly suggesting that they 2 3 2 Krasnykh and Douglas may be used as sites for genetic alterations of the protein. A recent study by Rux and Burnett further rationahzed the use of these HVRs as potential sites for the incorporation of targeting ligands by showing their localization on the surface of the Ad virion [117]. A practical demonstration of the feasibility of Ad targeting via hexon modification was performed by Vigne et al.^ who replaced HVR5 in the Ad5 hexon protein with an RGD-containing peptide flanked with flexible linkers and demonstrated the ability of an Ad vector incorporating the modified hexon to achieve fiber-independent transduction of vascular smooth muscle cells [118]. In the aggregate, it appears that genetic modifications of the penton base and hexon proteins may eventually develop into an alternative strategy of Ad targeting, which may be used instead of, or in addition to, the fiber-based targeting approaches. VI. Transductionally Targeted Ad Vectors for Clinical Gene Therapy Applications As discussed above, the poor efficiency of Ad-mediated gene transfer in several human gene therapy trials has been correlated with a low level of expression of CAR by the target cells. Strategies to accomplish efficient cell-specific gene transfer by Ad vectors in vivo merely by exploiting physical methods to confine vector administration to isolated body compartments have proven inadequate. For example, locally administered Ad vectors carrying the herpes simplex virus thymidine kinase (HSV-TK) gene have been shown to disseminate, probably as a result of leakage into the bloodstream, resulting in a high level of liver-associated toxicity [119]. Substantial hepatic toxicity related to the absence of tumor cell-specific targeting has also been demonstrated in Ad- mediated transfer of the HSV-TK gene in an ascites model of human breast can cer [120]. Thus, targeted Ad vectors capable of efficient and cell-specific CAR- independent gene transfer are required for clinical gene therapy applications. The benefits accrued in preclinical studies using tropism-modified Ad vectors provide a strong rationale for the immediate employment of these vectors in clinical trials. As discussed above. Ad vectors modified to contain the integrin-targeting RGD motif within the HI loop of the fiber knob have permitted levels of gene transfer to CAR-deficient primary cells to be enhanced more than two orders of magnitude over unmodified vectors. Based on these observations, the University of Alabama at Birmingham is currently employing this vector backbone in Phase I clinical trials for ovarian cancer and recurrent cancer of the oral cavity and oropharynx. These trials are the first to employ tropism-modified viral vectors in human patients. It is hypothesized that the tropism-modified vectors will allow augmented transfer of the herpes simplex virus thymidine kinase and cytosine deaminase genes, respectively. 8. Targeted Adenoviral Vectors I: Transductional Targeting 2 3 3 at lower vector doses, thereby leading to increased efficacy and reduced toxicity. These two diseases represent ideal opportunities to perform the initial studies of tropism-modified Ad vectors in the clinical context. In this regard, ovarian cancer is generally confined to the peritoneal cavity, permitting vector administration by injection into that body compartment. Cancer of the oral cavity and oropharynx is a locoregional disease accessible to direct intratumoral injection of Ad vectors. Thus, the anatomical isolation of the disease targets facilitates vector administration. However, it is apparent that additional requirements will be imposed upon targeted Ad vectors designed for clinical use in disease settings for which systemic vector administration is mandated. It has been reported that intravenous administration of untargeted Ad5 vectors delivers more than 90% of the input virus to the liver, thereby reducing the titer of virus particles available for transduction of the target disease cells [121-123]. Importantly, several studies have shown that the intravenous administration of Ad vectors leads to liver toxicity [124, 125]. Thus, one of the barriers to intravenous delivery of Ad vectors in vivo is the high degree of sequestration by the liver. Zinn et al, have demonstrated that the liver uptake of intravenously administered technetium (Tc)-99m-labeled recombinant Ad5 knob in mice is significantly reduced upon coinjection of unlabeled Ad5 knob, but is not affected by Ad3 knob, which recognizes a different primary receptor [126]. This indicates that the liver possesses specific receptors for the Ad5 knob, an observation supported by the subsequent reports of high levels of CAR mRNA in the liver [7, 8]. These findings seemed to suggest that successful strate gies to reduce liver sequestration and achieve cell-specific targeting following intravenous injection of Ad vectors will necessitate modifications to the knob domain to prevent recognition of CAR. In support of this, Printz et al. [42] and Reynolds et al, [51] have observed significantly reduced transgene expression in the livers of mice injected with Ad5 vectors which are retargeted by bispecific conjugates which prevent binding to CAR. However, the effect of the conju gates in reducing hepatocytes transgene expression may not be due to CAR blockade alone. It is possible that the size of the antibody-complexed vector contributes to the reduction in hepatocyte transduction by effectively enlarging the vector particle such that it less readily transverses the small fenestrations of the mouse liver sinusoidal epithelium. The reasons for hypothesizing a mechanism other than (or in addition to) CAR blockade stem from the emerging results of studies using Ad vectors whose fibers have been genetically modified so that they no longer recognize CAR. Somewhat surprisingly, Leissner et al. observed that hepatic transgene expression mediated by CAR-ablated vectors following intravenous adminis tration into the tail vein of mice was not significantly reduced compared to unmodified vectors [127]. While the CAR-ablated vectors used in this study were suboptimal in that they did not contain a targeting ligand with specificity 2 3 4 Krasnykh and Douglas for an alternate receptor (and thus may eventually have accumulated in the liver "by default"), these results have called into question the notion that liver transduction by Ad vectors is purely due to the high level of CAR on hepato- cytes. We have in fact found that complexing Ad w îth the Fab fragment of a neutralizing anti-knob mAb was not sufficient to reduce liver transgene expres sion, w^hereas w^hen conjugated to a ligand, the same Fab fragment achieved the desired reduction. Whether this is due to particle size or the need for an effective alternate ligand is as yet unclear. Hence, modification of Ad vectors to avoid hepatocyte transduction does not appear to be as straightforw^ard as first thought. Additional mutations such as ablating the RGD motif in the penton base to avoid interaction w îth cellular integrins have been proposed and are currently under evaluation, as are studies using vectors genetically modified to increase particle size by extending shaft length with a view to diminishing viral penetration of the hepatic fenestrations. The combination of "liver untargeting" approaches with a genetically incorporated, truly specific ligand are eagerly awaited. It is clear that the development of rational strategies to facilitate the clinical application of systemically administered Ad vectors will be dependent on a better understanding of the biological basis of hepatic vector localization. The problem of liver sequestration of Ad vectors is not an issue which relates only to hepatocytes. In this regard, Reynolds et al. showed that while transductional targeting of Ad led to a reduction in hepatic transgene expres sion, the biodistribution of viral DNA 1.5 h after intravenous administration was not significantly altered [51]. The authors hypothesized that this could reflect nonspecific phagocytic uptake by Kupffer cells. It has previously been shown that 90% of Ad DNA is eliminated by the liver within 24 h in an early innate immune response and does not lead to transgene expression [128]. Inhibition of Kupffer cells reduces the elimination of Ad DNA from the liver and leads to a three- to fourfold increase in hepatic transgene expression [129, 130]. This suggests that further improvements in the use of targeted Ad vec tors for systemic gene delivery might necessitate strategies to mitigate against nonspecific sequestration of the vector by the reticuloendothelial system (RES). In accordance with this, Tao et al. have generated data in mice suggesting that low doses of Ad (1-3 x 10^^ viral particles) are efficiently taken up by the RES/Kupffer cells, whereas high doses (1 x 10^^ viral particles) saturate these cells [131]. While
the hepatic sinusoids and their fenestrations constitute a highly favorable anatomic environment for Ad entry and are thus a problem in the context of liver sequestration of the vector, anatomical factors in other tissues may actually impede Ad transduction. In support of this idea, Fechner et al. have reported that expression of CAR and av integrins does not correlate with Ad vector-mediated gene delivery in vivo [132], suggesting that anatom ical barriers, in particular the endothelium and the subendothelial matrix. 8. Targeted Adenoviral Vectors I: Transductional Targeting 2 3 5 need to be overcome in order to achieve organ-specific gene delivery. Thus, efficient gene transfer by targeted Ad vectors might require the implementation of additional methods to permeabilize anatomical barriers. In this regard, Maillard et al. have shov^n that pretreatment of the rabbit iliac artery w îth elastase could enhance Ad-mediated gene transfer to arterial smooth mus cle cells after balloon abrasion [133]. In a similar approach, Kuriyama et al. have reported that the administration of proteases to degrade the fibrous proteins of the extracellular matrix prior to intratumoral injection of Ad vec tors leads to increased Ad infection [134]. An in vitro study by Nevo et al, demonstrated that the endothelial cell monolayer presents a physical barrier to Ad infection of myocytes, which could be partially overcome by increasing endothelial permeability v^ith a-thrombin [135]. Protease digestion might also prove a rational strategy to increase the permeability of the basal lamina, which has been shown by Huard's group to present a physical barrier to the transduction of mature skeletal muscle by both untargeted and tropism- expanded Ad vectors [136, 137]. In a quite different approach, Cho et al. demonstrated that the efficiency of transduction of mature skeletal muscle could be enhanced by administering Ad vectors in a large solvent volume, thereby increasing the hydrostatic pressure and favoring vector egress out of the intravascular compartment [138]. In contrast to the anatomical situations described above, the "leaky" vasculature associated with solid tumors [139] is hypothetically favorable for the vascular egress of Ad tumor-targeted Ad vectors. Thus, it is apparent that the success of systemic administration of targeted Ad vectors will depend on a greater understanding of the receptor-independent biological factors such as vascular pharmacodynamics and anatomical barriers limiting their utility, which should, in turn, facilitate the development of rational strategies whereby these obstacles might be overcome. VII. Conclusion Key to the realization of the full potential of gene therapy is the devel opment of gene delivery vectors possessing the requisite level of efficiency and specificity. Despite a number of important biological features that make human adenoviruses a promising vector system for gene therapy, the CAR-dependence of Ad infection has been recognized as one of the major disadvantages of this vector system. In order to overcome this limitation, the concept of targeted Ad vectors capable of delivering therapeutic genes to specific subsets of cells affected by a disease was proposed in the early 1990s. A number of suc cessful studies in recent years have shown the feasibility of the concept and provided the rationale for further improvements to the currently available vectors. Moreover, the degree of specificity and efficiency of Ad-mediated 2 3 6 Krasn/kh and Douglas gene transfer from currently available vectors could be refined by combining existing strategies. For example, Ad-mediated gene transfer to a heteroge neous population of cells, such as found in a tumor, could be increased by simultaneous targeting with tw ô or more vectors targeted against distinct cell surface receptors, as described by Grill et al. [97]. Furthermore, a more exquisite level of specificity could be imposed on a transductionally targeted Ad vector by placing the expression of the transgene under the control of a tumor- or tissue-selective promoter, as described elsew^here in this volume. Additionally, v^hile this chapter has focused on the transductional targeting of replication-defective Ad vectors, it is also recognized that strategies to redirect replicating adenoviruses to achieve CAR-independent infection w îll be nec essary to realize the full potential of replicating adenoviruses in the clinical setting [140]. The studies described in this chapter have taken the field of Ad targeting from its conceptual stage to the state of a rapidly evolving branch of exper imental virology. It is novŝ becoming apparent that anatomical barriers w îll need to be overcome to enhance the utility of tropism-modified Ad vectors in the context of systemic delivery. Future w^ork should lead to the derivation of truly targeted Ad vectors suitable for a variety of clinical applications requiring highly efficient, cell-specific delivery of therapeutic genes in vivo. References 1. Henry, L. J., Xia, D., Wilke, M. E., Deisenhofer, J., and Gerard, R. D. (1994). Characteriza tion of the knob domain of the adenovirus type 5 fiber protein expressed in Escherichia coH. / . Virol 68, 5239-5246. 2. Louis, N., Fender, P., Barge, A., Kitts, P., and Chroboczek, J. (1994). Cell-binding domain of adenovirus serotype 2 fiber. / . Virol. 68, 4104-4106. 3. Novelli, A., and Boulanger, P. A. (1991). Deletion analysis of functional domains in baculovirus-expressed adenovirus type 2 fiber. Virology 185, 365-376. 4. Hong, J. S., and Engler, J. A. (1996). Domains required for assembly of adenovirus type 2 fiber trimers./. Virol 70, 7071-7078. 5. Novelli, A., and Boulanger, P. A. (1991). Assembly of adenovirus type 2 fiber synthesized in cell-free translation system./. Biol Chem. 266, 9299-9303. 6. Hong, J. S., and Engler, J. A. (1991). The amino terminus of the adenovirus fiber protein encodes the nuclear localization signal. Virology 185, 758-767. 7. Bergelson, J. M., Cunningham, J. A., Droguett, G., Kurt-Jones, E. A., Krithivas, A., Hong, J. S., Horv^itz, M. S., Crov^ell, R. L., and Finberg, R. W. (1997). Isolation of a common receptor for Coxsackie B viruses and adenoviruses 2 and 5. Science 275, 1320-1323. 8. Tomko, R. P., Xu, R., and Philipson, L. (1997). HCAR and MCAR: The human and mouse cellular receptors for subgroup C adenoviruses and group B coxsackieviruses. Proc. Natl Acad. Scu USA 94, 3352-3356. 9. Roelvink, P. W., Lizonova, A., Lee, J. G., Li, Y., Bergelson, J. M., Finberg, R. W., Brough, D. E., Kovesdi, I., and Wickham, T. J. (1998). The coxsackievirus-adenovirus receptor protein can function as a cellular attachment protein for adenovirus serotypes from subgroups A, C, D, E, and F. / . Virol 72, 7909-7915. 8. Targeted Adenoviral Vectors I: Transductional Targeting 2 3 7 10. Leon, R. P., Hedlund, T., Meech, S. J., Li, S., Schaack, J., Hunger, S. P., Duke, R. C , and DeGregori, J. (1998). Adenoviral-mediated gene transfer in lymphocytes. Proc. Natl Acad. ScL USA 95, 13,159-13,164. 11. Wang, X., and Bergelson, J. M. (1999). Coxsackievirus and adenovirus receptor cytoplasmic and transmembrane domains are not essential for coxsackievirus and adenovirus infection. / . Virol. 73, 2559-2562. 12. Freimuth, P., Springer, K., Berard, C , Hainfeld, J., Bew l̂ey, M., and Flanagan, J. (1999). Coxsackievirus and adenovirus receptor amino-terminal immunoglobulin V- related domain binds adenovirus type 2 and fiber knob from adenovirus type 12. / . Virol. 73, 1392-1398. 13. Wickham, T. J., Mathias, P., Cheresh, D. A., and Nemerov^, G. R. (1993). Integrins alpha v beta 3 and alpha v beta 5 promote adenovirus internalization but not virus attachment. Cell 73,309-319. 14. Bai, M., Campisi, L., and Freimuth, P. (1994). Vitronectin receptor antibodies inhibit infection of HeLa and A549 cells by adenovirus type 12 but not by adenovirus type 2. / . Virol. 68, 5925-5932. 15. Dales, S., and Chardonnet, Y. (1973). Early events in the interaction of adenoviruses with HeLa cells. IV. Association with microtubules and the nuclear pore complex during vectorial movement of the inoculum. Virology 56, 465-483. 16. Freimuth, P. (1996). A human cell line selected for resistance to adenovirus infection has reduced levels of the virus receptor, / . Virol. 70, 4081-4085. 17. Wickham, T. J., Roelvink, P. W., Brough, D. E., and Kovesdi, L (1996). Adenovirus targeted to heparan-containing receptors increases its gene delivery efficiency to multiple cell types. Nat. Biotechnol. 14, 1570-1573. 18. Wickham, T. J., Tzeng, E., Shears, L. L., Roelvink, P. W., Li, Y., Lee, G. M., Brough, D. E., Lizonova, A., and Kovesdi, I. (1997). Increased in vitro and in vivo gene transfer by adenovirus vectors containing chimeric fiber proteins. / . Virol. 71, 8221-8229. 19. Zabner, J., Freimuth, P., Puga, A., Fabrega, A., and Welsh, M.J . (1997). Lack of high affinity fiber receptor activity explains the resistance of ciliated airway epithelia to adenovirus infection./. Clin. Invest. 100, 1144-1149. 20. Pickles, R. J., McCarty, D., Matsui, H., Hart, P. J., Randell, S. H., and Boucher, R. C. (1998). Limited entry of adenovirus vectors into well-differentiated airway epithelium is responsible for inefficient gene transfer./. Virol. 72, 6014-6023. 21. Nalbantoglu, J., Pari, G., Karpati, G., and Holland, P. C. (1999). Expression of the primary coxsackie and adenovirus receptor is downregulated during skeletal muscle maturation and limits the efficacy of adenovirus-mediated gene delivery to muscle cells. Hum. Gene Ther. 10, 1009-1019. 22. Shayakhmetov, D. M., Papayannopoulou, T., Stamatoyannopoulos, G., and Lieber, A. (2000). Efficient gene transfer into human CD34(+) cells by a retargeted adenovirus vector. / . Virol. 74, 2567-2583. 23. Wan, Y. Y., Leon, R. P., Marks, R., Cham, C. M., Schaack, J., Gajewski, T. F., and DeGregori, J. (2000). Transgenic expression of the coxsackie/adenovirus receptor enables adenoviral-mediated gene delivery in naive T cells. Proc. Natl Acad. Sci. USA 97, 13,784-13,789. 24. Hemmi, S., Geertsen, R., Mezzacasa, A., Peter, I., and Dummer, R. (1998). The presence of human coxsackievirus and adenovirus receptor is associated with efficient adenovirus- mediated transgene expression in human melanoma cell cultures. Hum. Gene Ther. 9, 2363-2373. 25. Miller, C. R., Buchsbaum, D. J., Reynolds, P. N., Douglas, J. T., Gillespie, G. Y., Mayo, M. S., Raben, D., and Curiel, D. T. (1998). Differential susceptibility of primary and estab lished human glioma cells to adenovirus infection: targeting via the epidermal growth factor receptor achieves fiber receptor-independent gene transfer. Cancer Res. 58, 5738-5748. 2 3 8 Krasnykh and Douglas 26. Li, Y., Pong, R. C , Bergelson, J. M., Hall, M. C , Sagalowsky, A. I., Tseng, C. P., Wang, Z., and Hsieh, J. T. (1999). Loss of adenoviral receptor expression in human bladder cancer cells: a potential impact on the efficacy of gene therapy. Cancer Res. 59, 325-330. 27. Okegawa, T., Li, Y., Pong, R. C , Bergelson, J. M., Zhou, J., and Hsieh, J. T. (2000). The dual impact of coxsackie and adenovirus receptor expression on human prostate cancer gene therapy. Cancer Res. 60, 5031-5036. 28. O'Neil, K. T., Hoess, R. H., Jackson, S. A., Ramachandran, N. S., Mousa, S. A., and DeGrado, W. F. (1992). Identification of novel peptide antagonists for GPIIb/IIIa from a conformationally constrained phage peptide library. Proteins 14, 509-515. 29. Doorbar, J., and Winter, G. (1994). Isolation of a peptide antagonist to the thrombin receptor using phage display. / . Mol. Biol. 244, 361-369. 30. Goodson, R. J., Doyle, M. V., Kaufman, S. E., and Rosenberg, S. (1994). High-affinity urokinase receptor antagonists identified with bacteriophage peptide display. Proc. Natl. Acad. Sci. USA 91, 7129-7133. 31. Barry, M. A., Dower, W. J., and Johnston, S. A. (1996). Toward cell-targeting gene therapy vectors: Selection of cell-binding peptides from random peptide-presenting phage libraries. Nat. Med. 2, 299-305. 32. Pasqualini, R., Koivunen, E., and Ruoslahti, E. (1997). Alpha v integrins as receptors for tumor targeting by circulating ligands. Nat. BiotechnoL 15, 542-546. 33. Arap, W., Pasqualini, R., and Ruoslahti, E. (1998). Cancer treatment by targeted drug delivery to tumor vasculature in a mouse model. Science 279, 377-380. 34. Pasquahni, R., Koivunen, E., Kain, R., Lahdenranta, J., Sakamoto, M., Stryhn, A., Ashmun, R. A., Shapiro, L. H., Arap, W., and Ruoslahti, E. (2000). Aminopeptidase N is a receptor for tumor-homing peptides and a target for inhibiting angiogenesis. Cancer Res. 60, 111-111. 35. Douglas, J. T., Rogers, B. E., Rosenfeld, M. E., Michael, S. I., Feng, M., and Curiel, D. T. (1996). Targeted gene delivery by tropism-modified adenoviral vectors. Nat. BiotechnoL 14, 1574-1578. 36. Goldman, C. K., Rogers, B. E., Douglas, J. T., Sosnowski, B. A., Ying, W., Siegal, G. P., Baird, A., Campain, J. A., and Curiel, D. T. (1997). Targeted gene delivery to Kaposi's sarcoma cells via the fibroblast growth factor receptor. Cancer Res. 57, 1447-1451. 37. Reynolds, P. N., Miller, C. R., Goldman, C. K., Doukas, J.,
Sosnowski, B. A., Rogers, B. E., Gomez-Navarro, J., Pierce, G. F., Curiel, D, T., and Douglas, J. T. (1998). Targeting adenoviral infection with basic fibroblast growth factor enhances gene delivery to vascular endothelial and smooth muscle cells. Tumor Target. 3, 156-168. 38. Rogers, B. E., Douglas, J. T., Ahlem, C , Buchsbaum, D. J., Frincke, J., and Curiel, D. T. (1997). Use of a novel cross-linking method to modify adenoviral tropism. Gene Ther. 4, 1387-1392. 39. Rancourt, C , Rogers, B. E., Sosnowski, B. A., Wang, M., Piche, A., Pierce, G. F., Alvarez, R. D., Siegal, G. P., Douglas, J. T., and Curiel, D. T. (1998). Basic fibroblast growth factor enhancement of adenovirus-mediated delivery of the herpes simplex virus thymidine kinase gene results in augmented therapeutic benefit in a murine model of ovarian cancer. Clin. Cancer Res. 4 ,1455-1461. 40. Rogers, B. E., Douglas, J. T., Sosnowski, B. A., Ying, W., Pierce, G., Buchsbaum, D. J., DellaManna, D., Baird, A., and Curiel, D, T. (1998). Enhanced in vivo gene delivery to human ovarian cancer xenografts utilizing a tropism-modified adenovirus vector. Tumor Target. 3,15-31. 41. Gu, D. L., Gonzalez, A. M., Printz, M. A., Doukas, J., Ying, W., D'Andrea, M., Hoganson, D. K., Curiel, D. T., Douglas, J. T., Sosnowski, B. A., Baird, A., Aukerman, S. L., and Pierce, G. F. (1999). Fibroblast growth factor 2 retargeted adenovirus has redirected cellular tropism: Evidence for reduced toxicity and enhanced antitumor activity in mice. Cancer Res. 59,2608-2614. 8. Targeted Adenoviral Vectors I: Transductional Targeting 2 3 9 42. Printz, M. A., Gonzalez, A. M., Cunningham, M., Gu, D.-L., Ong, M., Pierce, G. F., and Aukerman, S. L. (2000). Fibroblast growth factor 2-retargeted adenoviral vectors exhibit a modified biolocalization pattern and display reduced toxicity relative to native adenoviral vectors. Hum. Gene Ther. 11, 191-204. 43. Chandler, L. A., Doukas, J., Gonzalez, A. M., Floganson, D. K., Gu, D. L., Ma, C , Nesbit, M., Crombleholme, T. M., Herlyn, M., Sosnowski, B. A., and Pierce, G. F. (2000). FGF2- targeted adenovirus encoding platelet-derived growth factor-B enhances de novo tissue formation. Mol. Ther. 2, 153-160. 44. Blackwell, J. L., Miller, C. R., Douglas, J. T., Li, H., Peters, G. E., Carroll, W. R., Strong, T. v. , and Curiel, D. T. (1999). Retargeting to EGFR enhances adenovirus infection efficiency of squamous cell carcinoma. Arch. Otolaryngol. Head Neck Surg. 125, 856-863. 45. Haisma, H., Pinedo, H., Rijswijk, A., der Meulen-Muileman, I., Sosnowski, B., Ying, W., Beusechem, V., Tillman, B., Gerritsen, W., and Curiel, D. (1999). Tumor-specific gene transfer via an adenoviral vector targeted to the pan-carcinoma antigen EpCAM. Gene Ther. 6, 1469-1474. 46. Kelly, F. J., Miller, C. R., Buchsbaum, D. J., Gomez-Navarro, J., Barnes, M. N., Alvarez, R. D., and Curiel, D. T. (2000). Selectivity of TAG-72-targeted adenovirus gene transfer to primary ovarian carcinoma cells versus autologous mesothelial cells in vitro. Clin. Cancer Res. 6, 4323-4333. 47. Tillman, B. W., Gruijl, T. D., Bakker, S. A., Scheper, R. J., Pinedo, H. M., Curiel, T. J., Gerritsen, W. R., and Curiel, D. T. (1999). Maturation of dendritic cells accompanies high-efficiency gene transfer by a CD40-targeted adenoviral vector. / . Immunol. 162, 6378-6383. 48. Tillman, B. W., Hayes, T. L., DeGruijl, T. D., Douglas, J. T., and Curiel, D. T. (2000). Adenoviral vectors targeted to CD40 enhance the efficacy of dendritic cell-based vaccination against human papillomavirus 16-induced tumor cells in a murine model. Cancer Res. 60, 5456-5463. 49. Schneider, H., Groves, M., Muhle, C , Reynolds, P. N., Knight, A., Themis, M., Carvajal, J., Scaravilli, F., Curiel, D. T., Fairweather, N. F., and Coutelle, C. (2000). Retargeting of adenoviral vectors to neurons using the He fragment of tetanus toxin. Gene Ther. 7, 1584-1592. 50. Trepel, M., Grifman, M., Weitzman, M. D., and Pasqualini, R. (2000). Molecular adaptors for vascular-targeted adenoviral gene delivery. Hum. Gene Ther. 11, 1971-1981. 51. Reynolds, P. N., Zinn, K. R., Gavrilyuk, V. D., Balyasnikova, I. V., Rogers, B. E., Buchs baum, D. J., Wang, M., Miletich, D. J., Grizzle, W. E., Douglas, J. T., Danilov, S. M., and Curiel, D. T. (2000). A targetable, injectable adenoviral vector for selective gene delivery to pulmonary endothefium in vivo. Mol. Ther. 2, 562-578. 52. Wickham, T. J., Segal, D. M., Roelvink, P. W., Carrion, M. E., Lizonova, A., Lee, G. M., and Kovesdi, L (1996). Targeted adenovirus gene transfer to endothelial and smooth muscle cells by using bispecific antibodies./. Virol. 70, 6831-6838. 53. Harari, O. A., Wickham, T. J., Stocker, C. J., Kovesdi, L, Segal, D. M., Huehns, T. Y., Sarraf, C , and Haskard, D. O. (1999). Targeting an adenoviral gene vector to cytokine-activated vascular endothelium via E-selectin. Gene Ther. 6, 801-807. 54. Wickham, T. J., Lee, G. M., Titus, J. A., Sconocchia, G., Bakacs, T., Kovesdi, L, and Segal, D. M. (1997). Targeted adenovirus-mediated gene delivery to T cells via CD3. / . Virol. 71, 7663-7669. 55. Yoon, S. K., Mohr, L., O'Riordan, C. R., Lachapelle, A., Armentano, D., and Wands, J. R. (2000). Targeting a recombinant adenovirus vector to HCC cells using a bifunctional Fab-antibody conjugate. Biochem. Biophys. Res. Commun. 272, 497-504. 56. Watkins, S. J., Mesyanzhinov, V. V., Kurochkina, L. P., and Hawkins, R. E. (1997). The 'adenobody' approach to viral targeting: Specific and enhanced adenoviral gene delivery. Gene Ther. 4, 1004-1012. 2 4 0 Krasnykh and Douglas 57. Nicklin, S. A., White, S. J., Watkins, S. J., Hawkins, R. E., and Baker, A. H. (2000). Selective targeting of gene transfer to vascular endothelial cells by use of peptides isolated by phage display. Circulation 102, 231-237. 58. Haisma, H. J., Grill, J., Curiel, D. T., Hoogeland, S., van Beusechem, V. W., Pinedo, H. M., and Gerritsen, W. R. (2000). Targeting of adenoviral vectors through a bispecific single-chain antibody. Cancer Gene Ther. 7, 901-904. 59. Dmitriev, I., Kashentseva, E., Rogers, B. E., Krasnykh, V., and Curiel, D. T. (2000). Ectodomain of coxsackievirus and adenovirus receptor genetically fused to epidermal grov^th factor mediates adenovirus targeting to epidermal grow^th factor receptor-positive cells. / . Virol. 74, 6875-6884. 60. Ebbinghaus, C., Al-Jaibaji, A., Operschall, E., Schoffel, A., Peter, L, Greber, U. F., and Hemmi, S. (2001). Functional and selective targeting of adenovirus to high-affinity Fcgamma receptor I-positive cells by using a bispecific hybrid adapter. / . Virol. 75, 480-489. 61. Li, E., Brown, S. L., Von Seggern, D. J., Brown, G. B., and Nemerow, G. R. (2000). Signaling antibodies complexed with adenovirus circumvent CAR and integrin interactions and improve gene delivery. Gene Ther. 7, 1593-1599. 62. Hong, S. S., Galaup, A., Peytavi, R., Chazal, N., and Boulanger, P. (1999). Enhancement of adenovirus-mediated gene delivery by use of an oligopeptide with dual binding specificity. Hum. Gene Ther. 10, 2577-2586. 63. Romanczuk, H., Galer, C. E., Zabner, J., Barsomian, G., Wadsworth, S., and O'Riordan, C. R. (1999). Modification of an adenoviral vector with biologically selected peptides: A novel strategy for gene delivery to cells of choice. Hum. Gene Ther. 10, 2615-2626. 64. Drapkin, P. T., O'Riordan, C. R., Yi, S. M., Chiorini, J. A., Cardella, J., Zabner, J., and Welsh, M. J. (2000). Targeting the urokinase plasminogen activator receptor enhances gene transfer to human airway epitheHa. / . Clin. Invest. 105, 589-596. 65. O'Riordan, C. R., Lachapelle, A., Delgado, C , Parkes, V., Wadsworth, S. C , Smith, A. E., and Francis, G. E. (1999). PEGylation of adenovirus with retention of infectivity and protection from neutralizing antibody in vitro and in vivo. Hum, Gene Ther. 10, 1349-1358. 66. Fisher, K. D., Stallwood, Y., Green, N. K., Ulbrich, K., Mautner, V., and Seymour, L. W. (2001). Polymer-coated adenovirus permits efficient retargeting and evades neutralising antibodies. Gene Ther. 8, 341-348. 67. Smith, J. S., Keller, J. R., Lohrey, N. C , McCauslin, C. S., Ortiz, M., Cowan, K., and Spence, S. E. (1999). Redirected infection of directly biotinylated recombinant aden ovirus vectors through cell surface receptors and antigens. Proc. Natl Acad. Sci. USA 96, 8855-8860. 68. Kreda, S. M., Pickles, R. J., Lazarowski, E. R., and Boucher, R. C. (2000). G-protein- coupled receptors as targets for gene transfer vectors using natural small-molecule ligands. Nat. Biotechnol. 18, 635-640. 69. Bewley, M. C , Springer, K., Zhang, Y. B., Freimuth, P., and Flanagan, J. M. (1999). Structural analysis of the mechanism of adenovirus binding to its human cellular receptor, CAR. Science 286, 1579-1583. 70. Roelvink, P. W., Mi Lee, G., Einfeld, D. A., Kovesdi, I., and Wickham, T.J. (1999). Identification of a conserved receptor-binding site on the fiber proteins of CAR-recognizing adenoviridae. Science 286, 1568-1571. 71. Gall, J., Kass-Eisler, A., Leinwand, L., and Falck-Pedersen, E. (1996). Adenovirus type 5 and 7 capsid chimera: Fiber replacement alters receptor tropism without affecting primary immune neutralization epitopes./. Virol. 70, 2116-2123. 72. Chillon, M., Bosch, A., Zabner, J., Law, L., Armentano, D., Welsh, M. J., and Davidson, B. L. (1999). Group D adenoviruses infect primary central nervous system cells more efficiently than those from group C. / . Virol. 73, 2537-2540. 8. Targeted Adenoviral Vectors I: Transductional Targeting 2 4 1 73. Zabner, J., Chillon, M., Grunst, T., Moninger, T. O., Davidson, B. L., Gregory, R., and Armentano, D. (1999). A chimeric type 2 adenovirus vector v^ith a type 17 fiber enhances gene transfer to human airway epithelia. / . Virol. 73, 8689-8695. 74. Goossens, P. H., Havenga, M. J., Pieterman, E., Lemckert, A. A., Breedveld, F. C., Bout, A., and Huizinga, T. W. (2001). Infection efficiency of type 5 adenoviral vectors in synovial tissue can be enhanced with a type 16 fiber. Arthritis Rheum. 44, 570-577. 75. Havenga, M. J., Lemckert, A. A., Grimbergen, J. M., Vogels, R., Huisman, L. G., Vale- rio, D., Bout, A., and Quax, P. H. (2001). Improved adenovirus vectors for infection of cardiovascular tissues./. Virol. 75, 3335-3342. 76. Rea, D., Havenga, M. J., van Den Assem, M., SutmuUer, R. P., Lemckert, A., Hoeben, R. C., Bout, A., Melief, C. J., and Offringa, R. (2001). Highly efficient transduction of human monocyte-derived dendritic cells with subgroup B fiber-modified adenovirus vectors enhances transgene-encoded antigen presentation to cytotoxic T cells. / . Immunol. 166, 5236-5244. 77. Krasnykh, V. N., Mikheeva, G. V., Douglas, J. T., and Curiel, D. T. (1996). Generation of recombinant adenovirus vectors with modified fibers for altering viral tropism. / . Virol. 70, 6839-6846. 78. Stevenson, S. C., Rollence, M., Marshall-Neff, J., and McClelland, A. (1997). Selective targeting of human cells by a chimeric adenovirus vector containing a modified fiber protein. / . Virol. 71, 4782-4790. 79. Von Seggern, D. J., Huang, S., Fleck, S. K., Stevenson, S. C., and Nemerow, G. R. (2000). Adenovirus vector pseudotyping in fiber-expressing cell lines: Improved transduction of epstein-barr virus-transformed B cells. / . Virol. 74, 354-362. 80. Michael, S. L, Hong, J. S., Curiel, D. T., and Engler, J. A. (1995). Addition of a short peptide ligand to the adenovirus fiber protein. Gene Ther. 2, 660-668. 81. McDonald, G. A., Zhu, G., Li, Y., Kovesdi, I., Wickham, T. J., and Sukhatme, V. P. (1999). Efficient adenoviral gene transfer to kidney cortical vasculature using a fiber modified vector. / . Gene Med. 1,103-110. 82. Bouri, K., Feero, W. G., Myerburg, M. M., Wickham, T. J., Kovesdi, L, Hoffman, E. P., and Clemens, P. R. (1999). Polylysine modification of adenoviral fiber protein enhances muscle cell transduction. Hum. Gene Ther. 10, 1633-1640. 83. Gonzalez, R., Vereecque, R., Wickham, T. J., Facon, T., Hetuin, D., Kovesdi, I., Banters, F., Fenaux, P., and Quesnel, B. (1999). Transduction of bone marrow cells by the AdZ.F(pK7) modified adenovirus demonstrates preferential gene transfer in myeloma cells. Hum. Gene Ther. 10,2709-2717. 84. Gonzalez, R., Vereecque, R., Wickham, T. J., Vanrumbeke, M., Kovesdi, I., Banters, F., Fenaux, P., and Quesnel, B. (1999). Increased gene transfer in acute myeloid leukemic cells by an adenovirus vector containing a modified fiber protein. Gene Ther. 6, 314-320. 85. Staba, M. J., Wickham, T. J., Kovesdi, I., and Hallahan, D. E. (2000). Modifications of the fiber in adenovirus vectors increase tropism for malignant glioma models. Cancer Gene Ther. 7 ,13-19 . 86. Yoshida, Y., Sadata, A., Zhang, W., Saito, K., Shinoura, N., and Hamada, H. (1998). Generation of fiber-mutant recombinant adenoviruses for gene therapy of malignant glioma. Hum. Gene Ther. 9, 2503-2515. 87. Shinoura, N., Sakurai, S., Asai, A., Kirino, T., and Hamada, H. (2000). Transduction of a fiber-mutant adenovirus for the HSVtk gene highly augments the cytopathic
effect towards gliomas.//?«./. Cancer Res. 91 , 1028-1034. 88. Shinoura, N., Yoshida, Y., Tsunoda, R., Ohashi, M., Zhang, W., Asai, A., Kirino, T., and Hamada, H. (1999). Highly augmented cytopathic effect of a fiber-mutant ElB-defective adenovirus for gene therapy of gUomas. Cancer Res. 59, 3411-3416. 89. Xia, D., Henry, L. J., Gerard, R. D., and Deisenhofer, J. (1994). Crystal structure of the receptor-binding domain of adenovirus type 5 fiber protein at 1.7 A resolution. Structure 2, 1259-1270. 2 4 2 Krasnykh and Douglas 90. Krasnykh, V., Dmitriev, I., Mikheeva, G., Miller, C. R., Belousova, N., and Curiel, D. T. (1998). Characterization of an adenovirus vector containing a heterologous peptide epitope in the HI loop of the fiber knob. / . Virol. 72, 1844-1852. 91. Dmitriev, I., Krasnykh, V., Miller, C. R., Wang, M., Kashentseva, E., Mikheeva, G., Belousova, N., and Curiel, D. T. (1998). An adenovirus vector with genetically modified fibers demonstrates expanded tropism via utilization of a coxsackievirus and adenovirus receptor-independent cell entry mechanism./. Virol. 72, 9706-9713. 92. Vanderkwaak, T. J., Wang, M., J, G. m.-N., Rancourt, C , Dmitriev, I., Krasnykh, V., Barnes, M., Siegal, G. P., Alvarez, R., and Curiel, D. T. (1999). An advanced generation of adenoviral vectors selectively enhances gene transfer for ovarian cancer gene therapy approaches. Gynecol. Oncol. 74, 227-234. 93. Blackwell, J. L., Li, H., Navarro, J., Dmitriev, I., Krasnykh, V., Richter, C. A., Shaw, D. R., Alwarez, R. D., Curiel, D. T., and Strong, T. V. (2000). Using a tropism-modified adenoviral vector to circumvent inhibitory factors in ascites fluid. Hum. Gene Ther. 11, 1657-1669. 94. Kasono, K., Blackwell, J. L., Douglas, J. T., Dmitriev, I., Strong, T. V., Reynolds, P., Kropf, D. A., Carroll, W. R., Peters, G. E., Bucy, R. T., Curiel, D. T., and Krasnykh, V. (1999). Selective gene delivery to head and neck cancer cells via an integrin targeted adenovirus vector. Clin. Cancer Res. 5, 2571-2579. 95. Garcia-Castro, J., Segovia, J. C , Garcia-Sanchez, F., Lillo, R., Gomez-Navarro, J., Curiel, D. T., and Bueren, J. A. (2001). Selective transduction of murine myelomonocytic leukemia cells (WEHI-3B) with regular and RGD-adenoviral vectors. Mol. Ther. 3, 70-77. 96. Gripe, T. P., Dunphy, E. J., Holub, A. D., Saini, A., Vasi, N. H., Mahller, Y. Y., Collins, M. H., Snyder, J. D., Krasnykh, V., Curiel, D. T., Wickham, T. J., DeGregori, J., Bergelson, J. M., and Currier, M. A. (2001). Fiber knob modifications overcome low, heterogeneous expression of the coxsackievirus-adenovirus receptor that limits adenovirus gene transfer and oncolysis for human rhabdomyosarcoma cells. Cancer Res. 61, 2953-2960. 97. Grill, J., Van Beusechem, V. W., Van Der Valk, P., Dirven, C. M., Leonhart, A., Pherai, D. S., Haisma, H. J., Pinedo, H. M., Curiel, D. T., and Gerritsen, W. R. (2001). Combined targeting of adenoviruses to integrins and epidermal growth factor receptors increases gene transfer into primary glioma cells and spheroids. Clin. Cancer Res. 7, 641-650. 98. Asada-Mikami, R., Heike, Y., Kanai, S., Azuma, M., Shirakawa, K., Takaue, Y., Krasnykh, v., Curiel, D. T., Terada, M., Abe, T., and Wakasugi, H. (2001). Efficient gene transduction by RGD-fiber modified recombinant adenovirus into dendritic cells. Jpn. J. Cancer Res. 92, 321-327. 99. Xia, H., Anderson, B., Mao, Q., and Davidson, B. L. (2000). Recombinant human ade novirus: Targeting to the human transferrin receptor improves gene transfer to brain microcapillary endothelium. J. Virol. 74, 11359-11366. 100. Belousova, N., and Krasnykh, V., unpublished data. 101. Jakubczak, J. L., Rollence, M. L., Stewart, D. A., Jafari, J. D., Von Seggern, D. J., Nemerow, G. R., Stevenson, S. C , and Hallenbeck, P. L. (2001). Adenovirus type 5 viral particles pseudotyped with mutagenized fiber proteins show diminished infectivity of coxsackie B-adenovirus receptor-bearing cells./. Virol. 75, 2972-2981. 102. Von Seggern, D. J., Kehler, J., Endo, R. I., and Nemerow, G. R. (1998). Complementation of a fibre mutant adenovirus by packaging cell lines stably expressing the adenovirus type 5 fibre protein./. Gen. Virol. 79, 1461-1468. 103. Pereboev, A., Pereboeva, L., and Curiel, D. T. (2001). Phage display of adenovirus-5 fiber knob as a tool for specific ligand selection and vafidation./. Virol. 75, 7107-7113. 104. Krasnykh, V., Belousova, N., Korokhov, N., Mikheeva, G., and Curiel, D. T. (2001). Genetic targeting of an adenovirus vector via replacement of the fiber protein with the phage T4 fibritin./. Virol. 75,4176-4183. 8. Targeted Adenoviral Vectors I: Transductional Targeting 2 4 3 105. Douglas, J. T., Miller, C. R., Kim, M., Dmitriev, I., Mikheeva, G., Krasnykh, V., and Curiel, D. T. (1999). A system for the propagation of adenoviral vectors v^ith genetically modified receptor specificities. Nat. Biotechnol. 17, 470-475. 106. Mesyanzhinov, V., pers. commun. 107. van Beusechem, V. W., van Rijsv^ijk, A. L., van Es, H. H., Haisma, H. J., Pinedo, H. M., and Gerritsen, W. R. (2000). Recombinant adenovirus vectors w îth knobless fibers for targeted gene transfer. Gene Ther. 7, 1940-1946. 108. Graham, F. L., Smiley, J., Russell, W. C., and Nairn, R. (1977). Characteristics of a human cell line transformed by DNA from human adenovirus type 5. / . Gen. Virol. 36, 59-74. 109. Fallaux, F. J., Kranenburg, O., Cramer, S. J., Houv^eling, A., Van Ormondt, H., Hoeben, R. C , and Van Der Eb, A. J. (1996). Characterization of 911: a new helper cell line for the titration and propagation of early region 1-deleted adenoviral vectors. Hum. Gene Ther. 7, 215-222. 110. Fallaux, F. J., Bout, A., van der Velde, I., van den WoUenberg, D. J., Hehir, K. M., Keegan, J., Auger, C , Cramer, S. J., van Ormondt, H., van der Eb, A. J., Valerio, D., and Hoeben, R. C. (1998). Nev^ helper cells and matched early region 1-deleted adenovirus vectors prevent generation of replication-competent adenoviruses. Hum. Gene Ther. 9, 1909-1917. 111. Legrand, V., Spehner, D., Schlesinger, Y., Settelen, N., Pavirani, A., and Mehtafi, M. (1999). Fiberless recombinant adenoviruses: Virus maturation and infectivity in the absence of fiber. / . Virol. 73, 907-919. 112. Einfeld, D. A., Brough, D. E., Roelvink, P. W., Kovesdi, L, and Wickham, T.J. (1999). Construction of a pseudoreceptor that mediates transduction by adenoviruses expressing a ligand in fiber or penton base. / . Virol. 73, 9130-9136. 113. Roelvink, P. W., Kovesdi, I., and Wickham, T. J. (1996). Comparative analysis of adenovirus fiber-cell interaction: Adenovirus type 2 (Ad2) and Ad9 utilize the same cellular fiber receptor but use different binding strategies for attachment. / . Virol. 70, 7614-7621. 114. Huang, S., Kamata, T., Takada, Y., Ruggeri, Z. M., and Nemerow, G. R. (1996). Adenovirus interaction v^ith distinct integrins mediates separate events in cell entry and gene delivery to hematopoietic cells. / . Virol. 70, 4502-4508. 115. Wickham, T. J., Carrion, M. E., and Kovesdi, I. (1995). Targeting of adenovirus penton base to new receptors through replacement of its RGD motif with other receptor-specific peptide motifs. Gene Ther. 2, 750-756. 116. Crawford-Miksza, L., and Schnurr, D. P. (1996). Analysis of 15 adenovirus hexon proteins reveals the location and structure of seven hypervariable regions containing serotype-specific residues./. Virol. 70, 1836-1844. 117. Rux, J. J,, and Burnett, R. (2000). Type-specific epitope locations revealed by X-ray crystal- lographic study of adenovirus type 5 hexon. Mol. Ther. 1, 18-30. 118. Vigne, E., Mahfouz, I., Dedieu, J. F., Brie, A., Perricaudet, M., and Yeh, P. (1999). RGD inclusion in the hexon monomer provides adenovirus type 5-based vectors with a fiber knob-independent pathway for infection./. Virol. 73, 5156-5161. 119. Brand, K., Arnold, W., Bartels, T., Lieber, A., Kay, M. A., Strauss, M., and Dorken, B. (1997). Liver-associated toxicity of the HSV-tk/GCV approach and adenoviral vectors. Cancer Gene Ther. 4, 9-16. 120. Yee, D., McGuire, S. E., Brunner, N., Kozelsky, T. W., Allred, D. C , Chen, S. H., and Woo, S. L. (1996). Adenovirus-mediated gene transfer of herpes simplex virus thymidine kinase in an ascites model of human breast cancer. Hum. Gene Ther. 7, 1251-1257. 121. Herz, J,, and Gerard, R. D. (1993). Adenovirus-mediated transfer of low density lipoprotein receptor gene acutely accelerates cholesterol clearance in normal mice. Froc. Natl Acad. Sci. U5A 90, 2812-2816. 122. Kass-Eisler, A., Falck-Pedersen, E., Elfenbein, D. H., Alvira, M., Buttrick, P. M., and Leinwand, L. A. (1994). The impact of developmental stage, route of administration and the immune system on adenovirus-mediated gene transfer. Gene Ther. 1, 395-402. 2 4 4 Krasnykh and Douglas 123. Huard, J., Lochmuller, H., Acsadi, G., Jani, A., Massie, B., and Karpati, G. (1995). The route of administration is a major determinant of the transduction efficiency of rat tissues by adenoviral recombinants. Gene Ther. 2, 107-115. 124. O'Neal, W. K., Zhou, H., Morral, N., Aguilar-Cordova, E., Pestaner, J., Langston, C., Mull, B., Wang, Y., Beaudet, A. L., and Lee, B. (1998). Toxicological comparison of E2a-deleted and first-generation adenoviral vectors expressing alpha 1-antitrypsin after systemic delivery. Hum. Gene Ther. 9, 1587-1598. 125. Lozier, J. N., Metzger, M. E., Donahue, R. E., and Morgan, R. A. (1999). Adenovirus- mediated expression of human coagulation factor IX in the rhesus macaque is associated with dose-limiting toxicity. Blood 94, 3968-3975. 126. Zinn, K. R., Douglas, J. T., Smyth, C. A., Liu, H. G., Wu, Q., Krasnykh, V. N., Mountz, J. D., Curiel, D. T., and Mountz, J. M. (1998). Imaging and tissue biodistribution of 99mTc-labeled adenovirus knob (serotype 5). Gene Ther. 5, 798-808. 127. Leissner, P., Legrand, V., Schlesinger, Y., Hadji, D. A., van Raaij, M., Cusack, S., Pavirani, A., and Mehtali, M. (2001). Influence of adenoviral fiber mutations of viral encapsidation, infectivity and in vivo tropism. Gene Ther. 8, 49-57. 128. Worgall, S., Wolff, G., Falck-Pedersen, E., and Crystal, R. G. (1997). Innate immune mechanisms dominate elimination of adenoviral vectors following in vivo administration. Hum. Gene Ther. 8, 37-44. 129. Lieber, A., He, C. Y., Meuse, L., Schowalter, D., Kirillova, I., Winther, B., and Kay, M. A. (1997). The role of Kupffer cell activation and viral gene expression in early liver toxicity after infusion of recombinant adenovirus vectors. / . Virol. 71, 8798-8807. 130. Wolff, G., Worgall, S., van Rooijen, N., Song, W. R., Harvey, B. G., and Crystal, R. G. (1997). Enhancement of in vivo adenovirus-mediated gene transfer and expression by prior depletion of tissue macrophages in the target organ. / . Virol. 71, 624-629. 131. Tao, N., Gao, G.-P., Parr, M., Johnston, J., Baradet, T., Wilson, J. M., Barsoum, J., and Fawell, S. E. (2001). Sequestration of adenoviral vector by Kupffer cells leads to a nonlinear dose response of transduction in liver. Mol. Ther. 3, 28-35 . 132. Fechner, H., Haack, A., Wang, H., Wang, X., Eizema, K., Pauschinger, M., Schoemaker, R. G., van Veghel, R., Houtsmuller, A. B., Schultheiss, H.-P., Lamers, J. M. J., and Poller, W. (1999). Expression of Coxsackie adenovirus receptor and alphav-integrin does not correlate with adenovector targeting in vivo indicating anatomical vector barriers. Gene Ther. 6, 1520-1535. 133. Maillard, L., Ziol, M., Tahlil, O., Le Feuvre, C , Feldman, L. J., Branellec, D., Bruneval, P., and Steg, P. (1998). Pre-treatment with elastase improves the efficiency of percutaneous adenovirus-mediated gene transfer to the arterial media. Gene Ther. 5, 1023-1030. 134. Kuriyama, N., Kuriyama, H., Julin, C. M., Lamborn, K., and Israel, M. A. (2000). Pre- treatment with protease is a useful experimental strategy for enhancing adenovirus-mediated cancer gene therapy. Hum. Gene Ther. 11, 2219-2230. 135. Nevo, N., Chossat, N., Gosgnach, W., Mercadier, J.-J., and Michel, J.-B. (2001). Increasing endothelial cell permeability improves the efficiency of myocyte adenoviral vector infection. / . Gene Med. 3, 42-50. 136. van Deutekom, J. C. T., Cao, B., Pruchnic, R., Wickham, T. J., Kovesdi, I., and Huard, J. (1999). Extended tropism of an adenoviral vector does not circumvent the maturation- dependent transducibility of mouse skeletal muscle. / . Gene Med. 1, 393-399. 137. Feero, W. G., Rosenblatt, J. D., Huard, J., Watkins, S. C , Epperly, M., Clemens, P. R., Kochanek, S., Glorioso, J. C , Partridge, T. A., and Hoffman, E. P. (1997). Viral gene delivery to skeletal muscle: insights on maturation-dependent loss of fiber infectivity for adenovirus and herpes simplex type 1 viral vectors. Hum. Gene Ther. 8, 371-380. 138. Cho, W. K., Ebihara, S., Nalbantoglu, J., Gilbert, R., Massie, B., Holland, P., Karpati, G., and Petrof, B. J. (2000). Modulation of Starling forces and muscle fiber maturity permits 8. Targeted Adenoviral Vectors I: Transductional Targeting 2 4 5 adenovirus-mediated gene transfer to adult dystrophic (mdx) mice by the intravascular route. Hum. Gene Ther. 11, 701-714.
139. Brown,]. M., and Giaccia, A. J. (1998).Theuniquephysiology of solid tumors: opportunities (and problems) for cancer therapy. Cancer Res 58, 1408-16. 140. Douglas, J. T., Kim, M., Sumerel, L. A., Carey, D. E., and Curiel, D. T. (2001). Efficient oncolysis by a replicating adenovirus (Ad) in vivo is critically dependent on tumor expression of primary Ad receptors. Cancer Res. 61, 813-817. 141. Krasnykh, V. N., Douglas, J. T., and van Beuschem, V. W. (2000). Genetic targeting of adenoviral vectors. Mol. Ther. 1, 391-405. C H A P T E R Targeted Adenoviral Vectors II: Transcriptional Targeting Sudhanshu P. Raikvsfar, Chinghai H. Kao, and Thomas A. Gardner Department of Urology, Microbiology, and Immunology Indiana University Medical Center Indianapolis, Indiana 46202 I. Introduction: Rationale of Transcriptional Targeting Gene therapy is an innovative approach aimed at introducing genetic material into an organism for therapeutic intent. Still in its infancy, this novel concept has witnessed fundamental preclinical success with numerous ongoing clinical trials to confirm these findings. Critical to the success of gene therapy trials are issues relating to specific delivery of physiologically active biomolecules at therapeutically significant concentrations. Initially this was achieved by using direct intralesional injections of vectors to localize the delivery to the target tissue and universal promoters to maximize expression at that site. Over the past several years, we and several other investigators have investigated the potential of tumor-specific promoters to transcriptionally regulate gene expression in the laboratory and in clinical trials. The safety demonstrated by these trials using tumor/tissue-specific promoters has led to the recent approval of a trial administering a conditionally replicative adenovirus systemically for the treatment of metastatic prostate cancer. In order for gene therapy to be widely applicable, there is an urgent need to develop a new generation of viral vectors capable of achieving these goals of targeted delivery and controlled gene expression at the target site. The aim of this chapter is to discuss various potential strategies that have been utilized to achieve tissue/tumor-specific expression using adenoviral vectors. A better understanding of tissue specific gene expression necessitates a basic review of the eukaryotic transcription process at the molecular level. Consequently, we ADENOVIRAL VECTORS FOR GENE THERAPY 2 4 7 Copyright 2002, Elsevier Science (USA). All rights reserved. 2 4 8 Gardner ef al. begin by examining the molecular architecture of DNA and its relationship with the transcriptional mechanism. li. Regulation of Transcription in Eukaryotes To fully understand the complexity underlining transcriptional targeting a brief review of the mammalian transcriptional process follows: A. Molecular Organization of DNA During interphase the genetic material in association with proteins is dispersed throughout the nucleus in the form of chromatin. At the onset of mitosis, chromatin condensation takes place and during prophase it undergoes further compression into recognizable chromosomes. The associated proteins are basic, positively charged (lysine- and arginine-containing) histones and less positively charged nonhistones including high-mobility group (HMG) proteins. Histones play a key role in chromatin structural organization and are subject to various posttranslational modifications like acetylation, phosphorylation, and ubiquitination. Histones constitute nearly half of all the chromatin protein by weight and can be divided into six types: HI , H2A, H2B, H3, H4, and H5. DNA is incorporated into a 100 A nucleosomal fiber comprising of two molecules each of H2A, H2B, H3, and H4 which form the core histone octamer along with one linker histone HI or H5. The nucleosome core particle consists of 146 base pairs of DNA while the core histone octamer interacts with about 200 base pairs of DNA. While the histones function by interacting with DNA to form nucleosome, the nonhistone proteins are responsible for performing diverse functions including tissue-specific transcription. The 100-A nucleosomal fiber is arranged into a higher order structure termed a 300-A supercoiled filament or solenoid. Evidence indicates that certain nonhistone proteins including topoisomerase II bind to chromatin every 60-100 kilobases and tether the supercoiled, 300-A filament into structural loops. Further interaction with other nonhistone proteins leads to gathering of loops into rosettes, which in association with additional nonhistones undergo condensation forming a scaffold. This is known as the radical loop-scaffold model of compaction. Special, irregularly spaced repetitive base sequences associate with nonhistone proteins to define chromatin loops. These stretches of DNA are known as scaffold-associated regions (SARs). In order to be competent for transcription, the 300-A chromatin filament must undergo decondensation. B. The Central Dogma According to the central dogma, the genetic information flows from (1) DNA to DNA during genomic replication and (2) DNA to protein during 9. Targeted Adenoviral Vectors II: Transcriptional Targeting 2 4 9 gene expression. Gene expression can be simply defined as a phenomenon by which the genetic information stored in DNA is transferred to a protein. It involves tw ô distinct processes. The process by w^hich cells convert genetic information from DNA to RNA is called transcription and the decoding of RNA information to generate a specific sequence of amino acids is called translation. In addition, the flow^ of genetic information from RNA to DNA has been demonstrated in the case of retroviruses. Thus, the flov\̂ of genetic information from DNA to RNA is sometimes reversible. How^ever, this flow^ is unidirectional from RNA to protein and irreversible since, normally, the genetic information v^ithin the messenger RNA (mRNA) intermediate is not altered. How^ever, a itw exceptions in the form of RNA editing seem to challenge the present concept. RNA editing has been shov^n to alter the information content of the gene transcripts by changing the structures of individual bases and by inserting or deleting uridine monophosphate residues. Gene expression in eukaryotes is a spatially and temporally regulated process. Gene expression is regulated at multiple levels including transcription, posttranscriptional processing, nucleocytoplasmic transport, mRNA stability, translation, posttranslational modification, and intracellular trafficking of the protein. C. Transcription (Fig. 1) In eukaryotes, transcription occurs in the nucleus with the help of RNA polymerase to generate a single-stranded RNA molecule that is complementary in base sequence to the DNA template strand. There are three different types of RNA polymerases for the transcription of different types of genes. RNA polymerase I functions to transcribe ribosomal RNA (rRNA) genes to generate a large rRNA primary transcript which undergoes processing within the nucleolus to generate a 28S rRNA, a 5.8S rRNA, and an 18S rRNA. RNA polymerase II transcribes all of the protein coding genes into primary transcripts called pre-mRNAs that upon posttranscriptional processing generate mRNAs. While RNA polymerase III is known to transcribe transfer RNAs (tRNAs), 5S rRNA, and small nuclear RNAs (snRNAs). There are two main types of c/s-acting elements in all polll-transcribed genes: promoters and enhancers. The promoter, which is in close proxim ity to the protein coding region, consists of nucleotide sequences spanning approximately —40 and +50 nucleotides relative to a transcription initia tion site. A typical core promoter consists of four distinct elements: (1) A unique sequence called Goldberg-Hogness or the TATA box which has a consensus sequence TATAAAA and is located about —25 to —30 nucleotides upstream of the transcription initiation site. The TATA box alone is sufficient for independently directing a low-level polll-mediated transcription. (2) An initiator element that is functionally analogous to the TATA box and directly 250 Gardner ef o/. Figure 1 The transcriptional process of the RNA polymerase I, II, and transcribing unwound DNA to rRNA, mRNA, and tRNA, respectively. overlaps the transcription start site and has the loose consensus sequence PyPyA+lNT/APyPy. (3) The downstream core promoter element which is located approximately at position +30 downstream of the initiation site and acts in conjunction with the initiator element to direct transcription initia tion. (4) The TFIIB recognition element, which has the consensus sequence ^ / C ^ / C ^ / A C G C C and is located from - 3 2 to - 3 8 upstream of the TATA box. Another cis-acting element called the CAAT box has a consensus sequence GGCCAATCT and is located near position —70 to —80 relative to transcrip tion initiation site. Mutagenesis studies suggest a critical role of the CAAT box in modulating the promoter's ability to facilitate transcription. In addition, polll promoters often contain two conserved sequences, the SPl or GC box (GGGCGG) at about position —110 and the octamer box (ATTTGCAT); however, their positions are variable and they may occur either singly or in multiple copies. These consensus sequences are known to influence 9. Targeted Adenoviral Vectors II: Transcriptional Targeting 2 5 1 the efficiency of the promoter in initiating transcription. In addition, the regu latory regions termed enhancers are located farther upstream, downstream, or within the gene. The activity of the enhancers is independent of the location, ori entation, and gene type. Although they may not be involved directly in template binding, they are capable of modulating highly efficient transcription initiation. The promoter regions are normally sequestered within the nucleosome and thus are rarely able to bind to basal transcription factors and RNA poly merase, thereby leading to transcriptional repression or silencing. In order for transcription initiation to occur, the sequestered promoter must be exposed so that it can readily bind basal factors and this is achieved by chromatin remodeling. The DNA in highly compacted chromatin is relatively resistant to nuclease DNasel digestion. Thus, sensitivity of the DNA to DNasel reveals the degree of chromatin condensation and is directly proportional to the transcrip tional activity of a particular gene. Chromatin remodeling by acetylation and deacetylation of the histone proteins represents a major regulatory mechanism during gene activation and repression, respectively. The acetylation of histones by histone acetylase causes neutralization of the lysine basic charge, which in turn causes relaxation of contacts between the histones and the DNA. Thus, acetylated histones are preferentially found in active or potentially active genes where the chromatin is less tightly packed. Further, treatment of cultured cells with compounds like sodium butyrate, which enhances histone acetylation, leads to activation of previously silenced cellular genes. The normal chromatin in the nucleosomal conformation can be converted into highly condensed heterochromatin which is transcriptionally inactive by the addition of methyl groups to a series of cytosine residues in the CpG din- ucleotides found in tissue-specific genes. Thus, methylation and demethylation may play a crucial role in tissue-specific gene regulation. Locus control regions (LCRs) are specialized regulatory sequences located several kilobases upstream of the gene and capable of modulating transcription of gene clusters by influenc ing the chromatin structure. An assembled LCR-transcription factor complex is called an enhanceosome and if any of the components of this complex are missing, transcriptional activation of the gene cluster cannot occur. Insula tors or boundary elements are regulatory sequences located in the vicinity of junctions between condensed and decondensed chromatin, which repre sent transcriptionally active and inactive loci, respectively. Insulators do not enhance transcription and are responsible for position-independent effects, but can prevent transcription when placed between an enhancer and a promoter. D. Mechanism of Transcription Eukaryotic transcription by RNA polll involves five stages: (a) formation of the preinitiation complex, (b) initiation, (c) promoter clearance, (d) elonga tion, and (e) termination. RNA polll cannot interact directly with the promoter to initiate transcription but requires recruitment to the promoter by interacting 2 5 2 Gardner et aL with transcription factors. Transcription initiation is precisely controlled by the binding of a variety of trans-acting proteins termed transcription factors to the promoter and the enhancer. Transcription factors that assist the binding of RNA polymerase II to the promoter and initiate low levels of transcription are called basal factors, while other transcription factors are termed activators and repressors by binding to the enhancers. The transcription factors that bind to the TATA box are known as the TATA-binding protein (TBP) and are essential to the initiation of transcription from all pol II genes. A number of other basal factors that associate with TBP are called TBP-associated factors (TAF//s) and help in the assembly and binding of the complex to the promoter, which in turn leads to transcription initiation. The first event in the formation of preinitiation complex involves recogni tion of the TATA box by a multisubunit TFIID complex. A complex consisting of TBP and TAF/js called TFIID specifically binds to the TATA box to induce conformational changes that favor the binding of other transcription factors like TFIIA and TFIIB, both of which can interact directly with TFIID. TFIIB
serves two critical roles in transcription initiation: (a) It acts as a bridge and recruits TFIIF/RNA polll to the promoter; and (b) it aids in the selection of the transcriptional start site. TFIIB interacts asymmetrically with TFIID-DNA and contacts the phosphodiester backbone of DNA both upstream and down stream of TATA box. The position of the amino terminus of TFIIB in the DNA-TFIID-TFIIB complex is located near the transcription start site, which might explain the role of TFIIB in stabilizing the melting of the promoter prior to RNA synthesis. Following the assembly of the DNA-TFIID-TFIIA-TFIIB complex, RNApolII is recruited to the promoter by TFIIF. TFIIF has two subunits: (1) the larger subunit, RAP74, which has an ATP-dependent DNA helicase activity which may catalyze the local unwinding of the DNA to initiate transcription, and (2) the smaller subunit, RAPS8, by which it binds tightly to the RNA polll. This is followed by binding of TFIIF to the DNA downstream from the transcriptional start point. Two other factors, TFIIH and TFIIJ are recruited to the initiation complex but their locations in the complex are unknown. The interaction of the preinitiation complex with the core promoter alone is not sufficient to initiate transcription. A sequence of events beginning with the phosphorylation of the carboxy-terminal domain of RNApolII by TFIIF followed by ATP hydrolysis set the stage for DNA melting, initiation of synthesis and promoter clearance. Most of the TFII factors dissociate before RNApolII leaves the promoter. The carboxy-terminal domain coordinates processing of RNA with transcription. The general process of transcription initiation is similar to that catalyzed by bacterial RNA polymerase. Binding of the RNApolII generates a closed complex, which is converted at a later stage to an open complex in which the DNA strands have been separated. TFIIF and TFIIH are involved in an 9. Targeted Adenoviral Vectors II: Transcriptional Targeting 2 5 3 extension of the unwound region of the DNA to allow the polymerase to begin transcription elongation. Several elongation factors including TFIIF, SII, SIII, ELL, and P-TEFb function to suppress or prevent premature pausing of RNApolII as it traverses the DNA template. Early in the elongation process when the growing RNA chains are about 30 nucleotides long, the 5̂ ends of the pre-mRNAs are modified by the addition of 7-methyl guanosine caps. The 7- methyl guanosine cap contains an unusual 5^-5' triphosphate linkage and two methyl groups are added posttranscriptionally. The 7-methyl guanosine caps are recognized by protein factors involved in the initiation of translation and also help by protecting the growing RNA chains from degradation by nucleases. The 3̂ ends of the RNA transcripts are produced by endonucleolytic cleavage of the primary transcripts rather than by the termination of tran scription. The transcription termination occurs at multiple sites located 1000 to 2000 nucleotides downstream from the site that will eventually become the 3̂ end of the mature transcript. The endonucleolytic cleavage occurs 11 to 30 nucleotides downstream from the conserved consensus sequence AAUAAA, which is located near the end of the transcription unit. Following endonucleolytic cleavage, the enzyme poly(A) polymerase adds about a 200- nucleotide-long poly(A) tail to the 3̂ ends of the transcript in a process termed polyadenylation. E. Structural Motifs (Fig. 2) The transcription factors are modular in nature and contain characteristic structural motifs. The DNA binding domain as the name implies, binds to the DNA sequences present in the promoters and enhancers while the trans-actwsition domain is responsible for the activation of transcription via protein-protein interactions. The DNA binding domains have characteristic three-dimensional motifs, which result from associations between amino acids present within the polypeptide chains. Thus far, at least five types of DNA binding motifs have been extensively characterized. These include (1) helix-turn-helix (HTH) motif, (2) leucine zipper motif, (3) helix-loop-helix (HLH) motif, (4) zinc-finger motif, and (5) steroid hormone-binding motif. 1. Helix-turn-helix motif (HTH) was first discovered as the DNA- binding domain of phage repressor proteins. It is characterized by a geometric conformation that consists of two a-helical regions separated by a turn of several amino acids, which enable it to bind to DNA. Unlike other DNA binding motifs, HTH cannot function alone, but as part of a larger DNA- binding domain it fits well into the major groove of the DNA. The HTH motif has been identified in a 180-bp sequence called the homeobox, which specifies a 60-amino-acid homeodomain sequence in a large number of eukaryotic transcription factors involved in developmentally regulated genes. 254 Gardner ef a/. a. HTH b> Zinc fingers Sheet Zinc Ion a Helix Recognition helix c. bLZip Leucine side chain Leucine zipper domain Loop (dimerization domain) DNA Basic region (contacts DNA) Basic region (contacts DNA) Figure 2 The structural nnotifs exhibited by transcriptional factors. 2. Leucine zipper motif consists of a stretch of amino acids with a leucine residue in every seventh position. The leucine-rich regions form an ot-hehx v^ith a leucine residue protruding at every other turn and when two such molecules dimerize, the leucine residues zip together. The dimer contains two alpha- helical regions adjacent to the zipper, which bind to phosphate residues and specific bases in DNA, giving it a scissors-like appearance. The transcription factor API has two major components: Jun and Fos polypeptides encoded by c-jun and c-fos genes, respectively. Both Jun and Fos contain leucine zippers in their dimerization domains. A Jun leucine zipper can interact with another Jun leucine zipper to form a homodimer or with a Fos leucine zipper to form a heterodimer; however, a Fos leucine zipper is unable to interact with another Fos leucine zipper to form a homodimer. Neither Jun nor Fos alone can bind to DNA and thus in their monomeric forms, they are unable to act as transcription factors. However, Jun-Jun homodimers or Jun-Fos heterodimers are both transcription factors and bind to DNA with the same target specificity but with different affinities. The ability to form homo- or heterodimers greatly increases the repertoire of potential transcription factors a cell can assemble from a limited number of gene products. 9. Targeted Adenoviral Vectors II: Transcriptional Targeting 2 5 5 3. Helix-loop-helix motif consists of a stretch 40-50 amino acids con taining two amphipathic a-helices separated by a 12- to 28-amino-acid long nonhelical loop. The proteins bearing HLH form both homodimers and het- erodimers by means of interactions between the hydrophobic residues on the corresponding faces of the two helices. The HLH proteins that contain a stretch of highly basic amino acids adjacent to the HLH motif are termed bHLH proteins. These bHLH proteins are of two types: Class A consists of proteins that are ubiquitously expressed (e.g., mammalian E12/E47), while, Class B consists of proteins that are expressed in a tissue-specific manner (e.g., mammalian MyoD, myogenin, and Myf-5). 4. Zinc-finger motif was first recognized in the Xenopus RNA polIII transcription factor TFIIIA. There are several types of zinc-finger proteins, however, the classic zinc-finger consists of about 23 amino acids with a loop of 12 to 14 amino acids between the Cys and His residues and a 7-8-amino-acid linker between the loops. The consensus sequence of a typical zinc finger is Cys-X2-4-Cys-X3-Phe-X3-Leu-X2-His-X3-His. The interspersed cysteine and histidine residues covalently bind a single zinc ion to form a tetrahedral structure thereby folding the amino acids into loops. The crystal structure analysis of DNA bound by zinc fingers suggests that the C-terminal part of each finger forms a-helices that bind DNA while the N-terminal part forms a P-sheet. Three a-helices fit into one turn of the major groove and each a-helix makes two sequence-specific contacts with DNA. A zinc finger transcription factor may contain anywhere from 2 to 13 zinc fingers. Thus zinc fingers bind to DNA and also control the specificity of dimerization. Therefore, a zinc finger motif offers a novel strategy to design an artificial sequence-specific DNA-binding protein aimed at regulating specific gene expression. Recent studies indicate that it is possible to engineer zinc finger protein- based gene switches for precise and specific regulation of gene expression. Beerli et al[l] have utilized zinc-finger domains to design a polydactyl protein specif ically recognizing 9- or 18-bp sequences in the 5̂ untranslated region of the erbB-2/HER-2 promoter. They have evaluated the efficacy of gene regulation by converting the polydactyl finger into a transcriptional repressor by fusion with Kruppel-associated box (KRAB), ERF repressor domain (ERD), or mSIN3 interaction domain (SID) repressor domains. Transcriptional activators were generated by fusion with the HSV VP16 activation domain or with a tetrameric repeat of VP16's minimal activation domain, termed VP64. Their results indi cate that both gene repression and activation can be achieved by targeting designed proteins to a single site within the transcribed region of a gene. Kang and Kim [2] examined the ability of designer zinc-finger transcription factors to regulate transcription in vitro using an ecdysone-inducible system. They con structed a 268/NRE chimeric peptide by linking the three-finger peptide from Zif268, which recognizes the site 5^-GCGTGGGCG-3\ and the three-finger NRE peptide (a variant of the Zif268 peptide), which binds specifically to part 2 5 6 Gardner et aL of a nuclear hormone response element 5^-AAGGGTTCA-3^ By incorporating a 19-bp binding site for the 268/NRE near the transcriptional start site in the luciferase reporter vectors >99% repression of activated transcription was observed in vivo. Earlier studies have shown that 268/NRE peptide binds to the 19-bp recognition sequence about 6000-fold more tightly than the Zif268 peptide [3]. Imanishi et al. [4] utilized zinc fingers to create six-zinc-finger pro teins SplZF6(Gly)n by connecting two DNA-binding domains of transcription factor Spl with flexible polyglycine peptide linkers. These peptides were capa ble of inducing specific DNA bending by binding to two GC boxes and may provide an optimized approach to control gene expression by changing the DNA bending direction. Corbi et al [5] engineered a novel gene, "Jazz," that encodes for a three-zinc-finger peptide capable of binding the 9-bp DNA sequence 5 -̂ GCTGCTGCG-3^ present in the promoter region of the human and murine utrophin genes. Chimeric transcription factors Gal4-Jazz and Spl-Jazz were able to drive the expression of luciferase from the human utrophin promoter. Moore et al. [6] addressed the issue of zinc-finger DNA-binding specificity by altering the way in which zinc-finger arrays are constructed. Their results suggest that by linking three two-finger domains rather than two three-finger units, far greater target specificity and binding with picomolar affinity can be achieved through increased discrimination against mutated or closely related sequences. Taken together, the overall results suggest the potential utility of zinc-finger-based designer transcription factors in achieving regulation of gene- specific expression in diverse applications including gene therapy, functional genomics, and transgenic organisms. F. Regulation of Adenoviral DNA Transcription Process The adenovirus is a double-stranded DNA virus that has evolved to infect a host cell, transport its DNA into the nucleus of the host, replicate its DNA, use the host transcriptional apparatus to produce necessary structural proteins for replication, assemble itself, and destroy the host to release the newly formed infectious particles to perpetuate the process further. This process has been described in detail in Chapter 3. III. Approaches of Transcriptional Regulation A. Prior Rationale: Universal Promoters Several universal promoters have been utilized to attempt to maximize gene expression. The LTR, CMV, and RSV promoters were isolated from Maloney retrovirus, cytomegalovirus, and Rous sarcoma virus, respectively. 9. Targeted Adenoviral Vectors II: Transcriptional Targeting 2 5 7 These promoter elements were used because of the universal transcriptional activation over a broad host range. This universal transcription allowed for excellent but nondiscriminatory gene transcription and subsequent transgene expression. Because of the high levels of gene expression within several DNA constructs (i.e., viruses, cosmids, plasmids, etc.), these promoters are still used daily throughout the scientific community to test hypotheses which require uniform and high-level gene transcription. These were the promoters utilized in the first wave of gene therapy clinical trials, which focused on maximal gene expression and used local injection techniques to control the region of gene expression achieved. The LTR promoter was used to control herpes simplex virus thymidine kinase (HSV-TK) expression in a retroviral vector by placing retroviral producer cells into residual brain tumors to confer TK expression to the brain tumor, which could lead to conversion of a prodrug and subsequent tumor cell death. The CMV promoter was used in a replication- deficient adenovirus to deliver p53
gene expression after intralesional delivery to patients with both lung and head and neck tumors and is still under clinical investigation. The RSV promoter was employed to express HSV-TK after intralesional delivery in patients with several different tumor types. B. Current Rationale of Tissue-Specific Promoters A major challenge facing gene therapy is to generate vectors capable of achieving tissue- or tumor-specific expression. Initial gene therapy strate gies utilized universal promoters that demonstrated gene transfer, but were associated with toxicity associated with nonspecific gene transduction (section III.A, above). Tissue-specific promoters offer a novel approach to developing transcriptionally targeted viral vectors with enhanced potential for human gene therapy applications as described below. Several important characteristics are required to develop a tissue/tumor-specific strategy for a particular disease. Fortunately, the recent explosion in our understanding of molecular events that are present in a variety of disease processes has simplified the identification of suitable promoters. Additionally the completion of the genome project and the utilization of microarray technology have enhanced the development of tissue- or tumor-specific promoters by allowing for the identification of novel but specific molecules associated with a particular disease (e.g., cancer). The advancements in molecular cloning techniques (e.g., PCR) has allowed the investigator to extract regulatory sequences from genomic DNA and evaluate each component through site directed mutagenesis analysis in plasmid expres sion vectors. Additionally, the development of luciferase and green fluorescent protein as well as other quantifiable transgenes has enabled the investigator to test the tissue- or tumor-specific nature of a particular promoter. To illustrate the concept and utility of a tissue/tumor-specific promoter five such promoters have been selected from Table I. The basic rationale for 258 Gardner ef o/. Table I Gene Therapy Applications Of Tissue-Specific Promofers for Transcriptional Targeting Promoter Tissue-Specificity Transgene Vector References AFP HCC HSV-TK Adenoviral [27,126, 134] CD, IL-2 Adenoviral [28, 30] ElA Adenoviral [133] Albumin Liver factor VIII Adenoviral [135, 136] a-Actin Muscle GHRH Nonviral [137] a-Lactalbumin Breast cancer CD Adenoviral [138] p-Lactoglobulin Breast cancer HSV-TK Adenoviral [139] P-Globin Erythroid cells p-globin Retroviral [140] c-erbB2 Breast and HSV-TK Adenoviral [141, 142] pancreatic cancer CEA Breast, pancreatic, HSV-TK, Cre Adenoviral [15,19,21] lung, and H-ras mutant colorectal carcinoma Egr-1 Radiation induced TNF-a, LacZ Adenoviral [143, 144] E-Selectin Tumor endothelium TNF-a Retroviral [145] Flt-1 Vascular endothelial Luciferase Adenoviral [146] growth factor receptor type-1 GFAP Glial cells FasL Adenoviral [147, 148] TH Retroviral Grp78 (BIP) Anoxic/acidic tumor HSV-TK Adenoviral [149, 150] tissue HSV-TK Retroviral hAAT Hepatocytes FactorIX Nonviral [151] HGH and HGPH-a Pituitary HSV-TK, Adenoviral [152, 153] HIF-la/HRE Hypoxia inducible Erythropoietin Nonviral [154] hK2 Prostate EGFP, ElA, ElB Adenoviral [132, 155, 156] HSP Heat induced p53, TNF-a Nonviral [157] Hybrid ERE-HRE Breast Cancer Harakiri Adenoviral [158] L-Plastin Epithelial tumors LacZ Adenoviral [159] MBP Oligodendrocytes Caspase 8 Adenoviral 1160] GFP AAV [161] MCK Undifferentiated LacZ Adenoviral [162] muscle MMTV-LTR Prostated cancer antisense Retroviral [163] c-myc MN/CA9 Renal cell carcinoma ElA Adenoviral [164, 165] MUCl (DF3) Breast cancer ElA Adenoviral [166] HSV-TK Retroviral [167] Nestin Glioma, Cre, LacZ Adenoviral [168] glioblastoma 9 . Targeted Adenoviral Vectors II: Transcriptional Targeting 259 Table 1 {continued) Promoter Tissue-Specificity Transgene Vector References NSE Neurons FasL, Adenoviral [169] BDNF AAV [170] Osteocalcin Osteosarcoma HSV-TK Adenoviral [77, 79, 171-173] Prostate HSV-TK Adenoviral [174-176] ElA Adenoviral [177] PEPCK Hepatocytes Insulin Adenoviral [178] PSA Prostate Nitroreductase Adenoviral [46, 48, 49] HSV-TK, PNP Adenoviral Preproenkephalin CNS LacZ HSV [179] Probasin Prostate ElA Adenoviral [156, 180] Caspase 9 Prolactin Pituitary HSV-TK Adenoviral [181] lactotrophic cells SLPI Ovarian, cervical HSV-TK Nonviral [182] carcinoma SM22a Smooth muscle cells LacZ Adenoviral [183] Surfactant protein C Respiratory HSV-TK Adenoviral [184] epithelium Tyrosinase Melanocytes Luc, PNP Nonviral [185, 186] GALV-FMG Retroviral Tyrosine hydroxylase Sympathetic nervous LacZ HSV [122] system selection, in vitro and in vivo laboratory investigation and the clinical testing associated with each, will be briefly reviewed below. 1. Carcinoembryonic Antigen (CEA) Promoter a. Rationale Carcinoembryonic antigen is a 180-kDa cell surface gly coprotein overexpressed in 90% of gastrointestinal malignancies, including colon, gastric, rectal, and pancreatic tumors, 70% of lung cancers, and about 50% of breast cancers [7]. Thompson et al. [8] initially reported on the molec ular cloning of the CEA gene from a human genomic library. Subsequently, Schrewe et al. [9] also isolated and characterized a cosmid clone containing the entire coding region of the CEA gene including its promoter. The CEA promoter region encompasses 400 bp upstream of the translational start site and is known to confer tissue-specific CEA expression. Hauck and Stan- ners [10] demonstrated that the CEA promoter region located between —403 and —124 bp upstream of the translational initiation site is capable of directing high levels of gene expression in CEA-expressing human colon cancer CRC cells. Chen et al. [11] showed the CEA promoter region to lie between —123 and —28 bp upstream from the transcriptional start site and have demonstrated 2 6 0 Gardner et al. the presence of SPl and upstream stimulatory factor binding sites. According to Richards et al. [12] the CEA promoter is located between —90 and +69 bp upstream from the transcriptional start site and the essential sequences of the CEA promoter reside between —90 and —17 bp upstream of the transcriptional start site of the CEA gene. Cao et al. [13] compared the CEA core promoter regions between —135 and +69 bp isolated from human colorectal carcinoma and normal adjacent mucosa and found that both the sequences were identical and without any mutations. Taking advantage of this fact, various studies have suggested the potential utility of the CEA promoter for restricted expression of heterologous genes (14, 10, 12). b. In Vitro and in Vivo Experiments with CEA Promoter Takeuichi et al. [15] demonstrated that an adenoviral vector encoding a CEA promoter- driven N116Y dominant-negative H-Ras mutant was capable of suppressing liver metastasis by the human pancreatic cancer cell line PCI-43 in a nude mice model. Lan et al. [16, 17] demonstrated successful adenoviral-mediated transduction of E. coli cytosine deaminase (CD) in vitro as well as in an immunodeficient in vivo model of MKN45 gastric carcinoma. As compared to an adenoviral vector in which CD expression is driven by the constitutive CAG promoter, the expression of CD under the control of CEA promoter was confined to tumor xenografts. However, the reduction in tumor burden by AdCEA-CD/5-fluoro-cytosine (5FC), although significant, was not as great as that induced by AdCAG-CD/5FC. In fact, the CEA promoter was shown to be 200 times less active than the CAG promoter. Similar results have been described in mice bearing xenografts that were transfected with CEA-CD constructs and subsequently treated with 5-FC (18, and 12). Tanaka and colleagues [19] have used the CEA promoter sequence located between —424 and —2 bp upstream of translational start site to generate an adenoviral vector expressing HSV-TK and examined its efficacy in killing CEA-producing cancer cells in vitro and in vivo. By employing intratumoral Ad-CEA-TK injection and gancyclovir (GCV) administration, the growth of the tumors was inhibited by 20% as compared to untreated tumors. Brand et al. [20] have used the CEA promoter (—296 to +102 bp with respect to transcriptional start site) to drive the expression of HSV-TK in an adenoviral vector. Their results indicate that the CEA promoter was active in several human and rat tumor-derived cell lines but not in rat primary hepatocytes and in mouse liver, while the CMV promoter was highly active in all cell types. Although the CEA promoter-driven TK expression was less, it was sufficient to kill 100% of cancer cells, indicating a significant bystander effect. Treatment of subcutaneous tumors in SCID mice with Ad-CEA-TK was able to significantly reduce tumor growth and the tail vein injection of a high dose of this virus caused no side-effects in the liver. Kijima et al. [21] have utilized a novel Cre-lox-based strategy to achieve enhanced antitumor effect against CEA-producing human lung and colon 9. Targeted Adenoviral Vectors II: Transcriptional Targeting 2 6 1 cancer cell lines. Their strategy involved generation of tw ô recombinant aden oviral vectors: one expressing the Cre recombinase gene under the control of the CEA promoter w^hile the second adenoviral vector is designed to express HSV-TK gene from the CAG promoter only after Cre excises the neomycin resistance gene (inserted betw^een the CAG promoter and HSV-TK) in a loxP site-specific manner. (Cre recombinase derived from bacteriophage PI mediates site-specific excisional deletion of a DNA sequence that is flanked by a pair of loxP sites composed of 34 nucleotides.) This novel approach requires simulta neous infection of a cell by the tv^o adenoviral vectors. Using this approach, a CEA-producing human cancer cell line w âs rendered 8.4-fold more sensitive to GCV than infection by Ad-CEA-TK alone. Intratumoral injection of Ad-CEA- Cre along v^ith Ad-lox-TK follow^ed by GCV treatment almost completely eradicated CEA-producing tumors in an athymic subcutaneous tumor model, v^hereas intratumoral injection of Ad-CEA-TK with GCV treatment shov^ed reduced tumor grow^th. 2. a-Fetoprotein (AFP) Promoter a. Rationale The human AFP gene is developmentally regulated and is expressed at high levels in the fetal liver but its transcription declines rapidly after birth and is barely detectable in adult life [22, 23]. However, overexpression of the AFP gene is a characteristic feature of human hepatocell ular carcinoma. The human AFP gene is about 20 kb long and contains 15 exons and 14 introns [24]. The cap site is located 44 nucleotides upstream of the translation initiation site and the TATA box is located 27 nucleotides upstream from the cap site and is flanked by sequences with dyad symmetry. Other sequences in the S' untranslated region include a CCAAC pentamer, a 14-bp enhancer-like sequence, a 9-bp sequence homologous to the gluco corticoid responsive element, a 90-bp direct repeat, and several alternating purine/pyrimidine sequences. The AFP promoter is 200 bp upstream of the transcriptional start site. It is regulated by hepatocyte nuclear factor 1 (HNFl), nuclear factor 1 (NFl), and CCAAT/enhancer binding protein (C/EBP). The human AFP enhancer is located between —4.9 and — 3.0 kb upstream of the transcriptional start site and consists of at least two functional domains designated A and B which have binding sites for at least four transcription factors, including HNFl , HNF3, HNF4, and C/EBP. The domain B is located at —3.7 to —3.3 kb upstream of the transcriptional start site and is solely responsible for typical enhancer effects, but maximum enhancer activity is observed together with domain A located at —5.1 to —3.7 kb. A hepatoma-specific nuclear factor termed AFPl is known to bind to an AT-rich sequence, TGATTAATAATTACA, in the B domain of the human AFP enhancer. The AFP enhancer plays a critical role in enhancing AFP gene expression in the fetal liver as well as in hepatocellular carcinoma. The AFP silencer, which is a negative c/s-acting element with a consensus sequence, 5^-CTTCATAACCTAATACTT- 3^ has been identified [25]. Two 2 6 2 Gardner ef aL transcriptional silencer elements have been identified: the proximal silencer which contains a single copy of the consensus sequence at —0.31 kb and the distal silencer at —1.75 kb which carries four copies of the consensus sequence. Of the two silencers, the distal silencer, exhibits a higher suppressive activity than the proximal silencer. The silencer activity is manifested only when the silencer is located downstream of the enhancer and upstream of the promoter. An inverse correlation exists between the silencer activity and the AFP expression levels in hepatocellular carcinoma cell lines, thereby suggesting the role of the silencer in downregulating the level of AFP expression. b. In Vitro and in Vivo Experiments with the AFP Promoter Because of its tissue-specific nature, the AFP promoter has been used in adenoviral vectors for transcriptional targeting of suicide genes in AFP-producing hepatocellular carcinoma (HCC) cells in vitro as well as in vivo. Kaneko et al, [26] developed adenoviral vectors using either the 4.9-kb AFP promoter (AvlAFPTKl) or RSV promoter (AvlTKl) to express HSV-TK gene. In vitro and in vivo cell-specific killing was observed in AFP-producing HuH7 hepatocellular carcinoma cells transduced with AvlAFPTKl and treated with GCV. In contrast to HuH7 tumors, AFP-nonproducing hepatocellular carcinoma SK-Hep-1 cells did not show complete regression when treated with AvlAFPTKl. AvlTKl was able to cause complete regression in SK-Hep-1 tumors. Using a similar approach, Kanai et al. [27] developed adenoviral vectors by incorporating AFP enhancer domains A and B (—4.0 to —3.3 kb) and a
0.17-kb AFP promoter to drive the expression of HSV-TK. These vectors conferred cell-specific killing in AFP- producing HuH-7 and HepG2 cell lines but not in non-AFP-producing HLE and HLF cell lines. Kanai et al. [28] have also reported on the development of adenoviral vectors in which the expression of £. coli CD is driven by the AFP promoter. These vectors were capable of causing regression of HCC xenografts following treatment with 5FC. Arbuthnot et al. [29] analyzed in vitro and in vivo cell-specific expression of the nuclear p-galactosidase using adenoviral vectors containing transcriptional elements derived from either rat AFP or the human insulin-like growth factor II genes. Their results indicate hepatoma cell- specific expression using AFP promoter; however, primary hepatoma cells were poorly infected by these adenoviral vectors. Bui et al. [30] compared adenoviral vector-mediated expression of IL-2 under the transcriptional control of murine AFP promoter and CMV promoters for the treatment of established human hepatocellular xenografts in CB-17/SCID mice. Intratumoral injection of these adenoviral vectors resulted in growth retardation and regression in a majority of animals but with a wider therapeutic index and less systemic toxicity using the AFP vector. Using the AFP promoter and cre-lox based approach Sato et al. [31] were able to achieve strictly tissue-specific expression of LacZ in AFP- producing cells in vitro as well as in vivo in nude mice bearing AFP-producing tumor xenografts. 9. Targeted Adenoviral Vectors II: Transcriptional Targeting 2 6 3 3. Prostate-Specific Antigen (PSA) Promoter a. Rationale The gene for prostate-specific antigen, a member of the glandular kallikrein family, was independently characterized by Riegman et al. [32, 33] and Lundwall [34] from a human genomic library. The gene contains five exons and is located on the long arm of chromosome 19, in the region ql33-qtQr [33]. The gene is 7130 bp long and includes 633 bp of 5̂ and 639 bp of 3̂ flanking sequence. The promoter region contains a variant TATA box (TTTATA) at position - 2 8 to - 2 3 , a GC box at - 5 3 to - 4 8 , a CACCC box at - 1 2 9 to - 1 2 5 . An imperfect palindromic sequence (AGAACAGCAAGTGCT) closely related to the reverse complement of the consensus sequence for steroid hormone receptor binding (TGTACANNNTGTC/TCT), is found at position — 170 to —156. In addition, GGGAGGG and CAGCCTC repeats are located in the region —123 to —72. Expression of PSA is primarily detected in human prostate [35-37]. Further, PSA expression has been shov^n to be androgen- responsive [38]. This is achieved by several transcription factors that are involved in regulating prostate-specific antigen gene. Tw ô functionally active androgen receptor-binding sites or androgen response elements (AREs) have been identified at positions —170 (ARE- I) and - 3 9 4 (ARE-II) [38-41]. Cleutjens et al. have identified a complex, androgen-regulated 440-bp enhancer (—4366 to —3874) w^hich contains a high-affinity AR-binding site, ARE-III (5^-GGAGGAACATATTGTATCGAT- 3^), at position —4200. In subsequent studies, a 6-kb PSA promoter fragment has been show^n to confer prostate-specific and androgen-regulated expression of P-galactosidase in transgenic mice [42]. Pang et al. [43] identified an 822-bp PSA gene regulatory sequence, PSAR w^hich w^hen combined v^ith the PSA promoter (PCPSA-P) exhibited an enhanced luciferase activity in LNCaP cells. Upon stimulation v^ith 10 to 100 nM dihydrotestosterone, a more than 1000- fold increase in expression w âs observed as compared to androgen-negative controls. Their studies further suggest that this 822-bp sequence alone could serve as a promoter, thereby indicating that the complete PSA promoter contains tw ô functional domains: a proximal promoter and a distal promoter, w^hich can also function as an enhancer. Yeung et al. [44] have identified tv^o ds-acting elements w^ithin the 5.8- kb PSA promoter that are essential for the androgen-independent activity of the PSA promoter in prostate cancer cells. Their studies provide evidence that androgen-independent activation of the PSA promoter in the androgen- independent prostate cancer cell line C4-2 involves tw ô distinct regions, a 440-bp AREc and a 150-bp pN/H, which are responsible for upregulation of the PSA promoter activity by employing two different pathways. AREc confers high basal PSA promoter activity in C4-2 cells, while pN/H is a strong AR-independent positive-regulatory element of the PSA promoter in both LNCaP and C4-2 cells. Further, a 17-bp RI fragment within the pN/H region was identified as the key cis element, which interacts with a 45-kDa 2 6 4 Gardner ef al. prostate cancer cell-specific transcription factor to mediate androgen- and AR- independent transcriptional activation of the PSA promoter. By juxtaposing AREc and pN/H, a chimeric PSA promoter has been created that exhibits 2- to 3-fold higher activity than wild-type PSA promoter in both LNCaP and C4-2 cells. Oettgen et al. [45] have identified a novel prostate epithelial-specific Ets transcription factor, PDEF, that is involved in PSA gene regulation and acts as a coregulator of AR. PDEF acts as an androgen-independent transcriptional activator of the PSA promoter. It also directly interacts with the DNA- binding domain of AR and enhances androgen-mediated activation of the PSA promoter. Thus, strong tissue-specificity of the PSA promoter makes it an ideal candidate for prostate cancer gene therapy. Latham [46] compared tissue- specific expression of luciferase reporter vectors by employing PSA, human glandular kallikrein (hKLK2), and CMV promoters in PSA-positive LNCaP and PSA-negative CoLo320, DG75, A2780, and Jurkat cells. Their studies revealed that minimal 628-bp PSA and hKLK2 promoters showed only low-level androgen-independent expression in both PSA-positive and PSA-negative cell lines. Tandem duplication of the PSA promoter slightly increased expression in LNCaP cells. Addition of the CMV enhancer upstream of the PSA or hKLK2 promoter led to substantially enhanced and nonspecific luciferase expression in all the cell lines. By placing a 1455-bp PSA enhancer sequence upstream of either the PSA or the hKLK2 promoter, a 20-fold increase in tissue-specific luciferase expression was observed. Tandem duplication of the PSA enhancer increased the expression 50-fold higher than either promoter while retaining tissue specificity. The expression from all the enhancer constructs was 100-fold above the basal levels upon induction with androgen dihydrotestosterone. b. In Vitro and in Vivo Experiments with the FSA Promoter These enhancer sequences were incorporated in adenoviral vectors to express enhanced green fluorescent protein (EGFP) and nitroreductase. The results indicate low-level expression of EGFP by PSA enhancer promoter in LNCaP cells and no expression in non-PSA-producing EJ cells when compared with CMV promoter-driven EGFP. However, the PSA enhancer promoter was able to direct expression of comparable levels of nitroreductase in a tissue-specific manner in LNCaP cells alone. These transduced LNCaP cells upon treatment with CB1954 exhibited cytotoxicity. A replication-competent adenoviral vector CN706 in which the El A gene is under the transcriptional control of the PSA enhancer/promoter has been shown to exhibit selective toxicity toward PSA- expressing prostate cancer cells [47]. Martinello-Wilks et al. [48] examined the efficacy of adenoviral vectors with a 630-bp PSA promoter-driven HSV-TK and E. coli purine nucleoside phosphorylase (PNP) genes for their ability to kill androgen-insensitive prostate cancer cell line PC-3 tumor xenografts in a nude mouse model. Both HSV-TK and E. coli PNP-expressing adenoviral vectors were able to achieve significant tumor regression in vivo following 9. Targeted Adenoviral Vectors II: Transcriptional Targeting 2 6 5 GCV or 6MPDR treatment. Gotoh et al. [49] developed transcriptionally targeted recombinant adenoviral vectors by incorporating either 5837-bp long or 642-bp short PSA promoter elements to drive the expression of HSV- TK. The long PSA promoter w âs shown to have superior activity over the short promoter and was more active in C4-2 cells than in LNCaP cells. In vitro expression of TK conferred marked killing of C4-2 cells upon acyclovir treatment. Administration of this virus in an in vivo subcutaneous C4-2 tumor model, followed by acyclovir treatment, revealed significant inhibition of tumor burden. Lee et al. [50] demonstrated tissue-specific growth suppression of PSA- positive and -negative cell lines by transfecting PSA promoter enhancer-driven p53 tumor-suppressor genes. Recently, human prostate cancer- and tissue- specific genes P503, P540S, and P510S have been identified using a combination of cDNA library subtraction and high-throughput microarray screening by Xu et al. [51]. It would be interesting to characterize the promoter region of these genes and use it in developing transcriptionally targeted adenoviral vectors. 4. Osteocalcin (OC) Promoter (Fig. 3) a. Rationale Osteocalcin (bone y-carboxyglutamic acid (Gla)- containing protein (BGP)) is a 50-amino-acid, 5.8-kDa, major noncollagenous protein found in adult bone and has been shown to be transcriptionally regulated by 1,25-dihydroxyvitamin D3 [52, 53]. The human, rat, and murine osteocalcin genes have been cloned and each consists of four exons and three introns [54-57]. Montecino [58] reported that the key promoter elements are located in two DNase I-hypersensitive sites. The proximal hypersensitive site ( — 170 to —70) includes sequence motifs that specifically interact with basal transcription factors such as Msx [59-61], HLH protein Id-1 [62], AP-1 [63], a bone-specific nuclear-matrix-associated protein, NMP-2 [64], Nonprostatic Cell Prostate Cancer Cell ^ T - - . Trgi^ Tumor-specific ^^""^""^^^^;;;'' ' " ^ ^ S . transcription # ' ' ' :" - - ^ . factors f OCpmmolar ' - TK - m \ f OC promoter Figure 3 The specific ability of a tissue specific promoter such as the osteocalcin promoter to produce HSV-TK in a prostate cancer allowing cell death on prodrug (ACV) administration while sparing nonprostate cell by not allowing osteocalcin promoter activation. 2 6 6 Gardner ef al. and a member of the AML family of transcription factors [65^ 66, 71]. The distal hypersensitive domain (—600 to —400) contains the vitamin D-responsive element (VDRE, —465 to —437), which interacts with the VDR- RXRa complex in a ligand-dependent manner [67-69]. Montecino et al, [70] have demonstrated that the promoter segment —343 to —108 is critical for inducing both proximal nuclease hypersensitivity and basal transcriptional activity and the DNase I hypersensitivity at —600 is not essential for vitamin D-dependent transcriptional upregulation. Two additional NMP-2 sites (site A, - 6 0 4 to -599 ; site B, - 4 4 0 to -435) have been identified in the sequences flanking the distal DNase I-hypersensitive domain that might support specific interactions between the nuclear matrix and the OC gene promoter [64, 71]. Analysis of the 5' flanking sequence of rat osteocalcin gene reveals a modular organization of the promoter consisting of the TATAAAA sequence between —31 and —25 and the CCAAT sequence between —92 and — 88 [72]. Lian et al. [55] identified a 24-nucleotide regulatory sequence, 5'- ATGACCCCCAATTAGTCCTGGCAG-3^ in the proximal promoter region with a CAAT motif as a central element, and have designated this sequence as an osteocalcin (OC) box since only two nucleotide substitutions are found in the rat and human osteocalcin genes in this region. Hoffman et al. [59] reported that the OC box is located at nucleotide positions between —99 and —76 and TATA box containing a consensus glucocorticoid-responsive element (GRE) between —44 and —31. The stimulation of osteocalcin gene expression by 1,25- dihydroxyvitamin D3 is associated with sequence-specific binding of nuclear factors to a 26-bp sequence, 5^-CTGGGTGAATGAGGACATTACTGACC- 3', located between —462 and —437. This sequence contains a region of hyphenated dyad symmetry and shares homology with consensus steroid- responsive elements.The promoter region has been shown to contain two sites of an E-box motif (a consensus binding site for HLH proteins) termed OCEl (CACATG at -102) and OCE2 (CAGCTG at -149) [62]. Mutagenesis studies have indicated that osteoblastic-specific gene transcription is regulated via the interaction between certain E-box binding transcription factors in osteoblasts and the OCEl sequence in the promoter region of the osteocalcin gene. Banerjee et al. [63] demonstrated that an AML-1 binding sequence within the proximal promoter (nt —138 to —130) contributes to 75% of the level of osteocalcin gene expression. The promoter region is not GC-rich and does not contain a consensus sequence for the SPl binding site [73]. Theofan et al. [74] performed a detailed analysis of the BGP promoter region. Three regulatory elements that share partial homology with the consensus sequence for the GRE have been identified at nucleotide positions —356, —178, and —68, respectively. In addition, two sequences related to the consensus sequence for the metal ion-responsive element (MRE) have been identified at positions — 190 and —143. An octanucleotide sequence, TGCAGTCA, is located directly adjacent 3̂ to the second MRE. Two
other sequences that share homology 9. Targeted Adenoviral Vectors II: Transcriptional Targeting 2 6 7 with the cAMP-responsive element are found at —437 (TGAGGACA) and —392 (TCACGGCA). The BGP promoter region also contains several pairs of inverted repeat sequences that form regions of dyad symmetry. Three particularly long regions of imperfect dyad symmetry are located between - 5 2 3 and -504 , - 2 3 4 and -214 , and - 5 1 and - 2 8 . An octanucleotide palindromic sequence from —134 to —127 partially overlaps both a putative MRE and a cAMP-responsive element. A short sequence, GCAG, or its complement, CTGC, is repeated 17 times. A region of alternating purines and pyrimidines at location —90 to —81 from the CAT box has the potential to form a Z-DNA structure which may be important in gene regulation. A 7-bp osteocalcin silencer element, 5^-TGGCCCT-3^, has been located between +29 and +35 position in the first exon of the human osteocalcin gene, while two silencer elements, 5'-CCTCCT-3^ (nt +106 to +111 and +135 to +140) and 5'-TTTCTTT-3' (nt +118 to +124), have been located in the first intron of the rat osteocalcin gene [75^ 76]. b. In Vitro and in Vivo Experiments with the OC Promoter Ko et al, [77] developed an osteocalcin-promoter-driven TK-expressing recom binant adenoviral vector to achieve tissue-specific killing of osteosarcoma cells in experimental animal model. Administration of this vector followed by acyclovir treatment led to a significant growth inhibition of osteosarcoma in an experimental animal model. Cheon et al. [78] used a chemogene therapy approach by combining OC-promoter-driven TK expression and acyclovir with a methotrexate treatment regimen in nude mice bearing either subcutaneous human osteosarcoma (MG-63) or rat osteosarcoma (ROS) xenografts. Their results indicate that osteosarcoma tumor growth was more efficiently inhibited due to synergistic effects of combined methotrexate and acyclovir treatment. Shirakawa et al. [79] further demonstrated the potential utility of an aden oviral osteocalcin promoter-mediated suicide gene therapy for osteosarcoma pulmonary metastasis in nude mice. Hou et al. [80] demonstrated osteoblast- specific gene expression in adherent bone marrow cells using a 1.7-kb rat OC-CAT gene. Recipient mice were shown to be positive for osteoblast-specific expression following bone marrow transplantation. Using a replication-defective adenovirus, Ad-OC-TK, we have completed a phase I clinical trial that demonstrated the expected safety profile and gene transfer that we expected. Eleven men with recurrent or metastatic prostate cancer were enrolled in a phase I intralesional dose-escalating trial, combining two Ad-OC-TK injections with 3 weeks of valacyclovir administration. In sum mary, this was well tolerated at all doses reaching a maximum of 5 x 10^^ pfu (or 1 X 10^^ vp) in patients in the high-dose group. Viral distribution stud ies revealed that after intralesional administration the patients demonstrated a measurable viremia for 2 - 3 days. Despite the presence of viral particles at these time points, no patient demonstrated hepatotoxity with valacyclovir admin istration. This is in direct contrast to intralesional delivery of Ad-RSV-TK 2 6 8 Gardner ef al. to the prostatic recurrence, in which patients experience hepatotoxicity upon prodrug administration. Finally, comparison of biopsy specimens prior to the first (day 0) and second (day 7) injection and at the end of the study (day 30) revealed successful gene transfer at day 7 by immunohistochemical stain ing for HSV-TK and some evidence of tumor destruction by day 30. These expected and encouraging results have led us to propose a phase I trial to test the transcriptional ability of the osteocalcin promoter to regulate adenoviral replication in a similar format. 5. MN/CA9 Promoter a. Rationale The human MN/CA9 gene has been isolated, sequenced, and characterized by Opavsky [81]. This gene is a member of the carbonic anhydrase (CA) family, which codes for a diverse group of catalysts of the reversible conversion of carbon dioxide to carbonic acid. MN/CA9 expression has been detected in several types of carcinomas including renal, ovarian, and cervical, as well as in normal gastric mucosa [82-85]. The complete genomic sequence of the MN/CA9 gene including the 5^-flanking region encompasses 10.9 kb with a coding sequence comprising of 11 exons. The MN/CA9 protein contains 459 amino acids with a molecular weight ranging from 54 to 58 kDa. MN displays CA activity and binds zinc [86]. The nucleotide sequence close to the 5' end shows 91.4% sequence homology to the U3 region of the long terminal repeats (LTRs) of the human HERV-K endogenous retroviruses [87]. This LTR-like sequence is 222 bp long with an A-rich tail at its 3̂ end. Analysis of the MN/CA9 promoter region between —507 and +1 upstream of the transcription initiation site indicates that despite the presence of 60% GC residues, the additional features of TATA-less promoters are absent, but the presence of consensus sequences for API, AP2, and p53 transcription factor binding sites has been demonstrated [88-90]. Functional characterization of the 3.5-kb MN 5' upstream region by deletion analysis led to the identification of —173 to +31 fragment as the MN promoter. The promoter region lacks the CpG-rich islands that are typical for TATA-less promoters but contains two nonoverlapping consensus initiator sequences required for promoter activity. b. In Vitro and in Vivo Experiences with MN Initial in vitro studies with this promoter driving luciferase expression demonstrated tumor specificity for both renal cell carcinoma and cervical carcinoma. Based on the expression assays, we have constructed an oncolytic adenovirus with the MN promoter which has demonstrated 40- to 100-fold increased killing in human renal cell carcinomas compared to control cell lines not expressing this promoter activity. We are currently evaluating this oncolytic vector in animal models of human renal cell carcinoma. 9. Targeted Adenoviral Vectors II: Transcriptional Targeting 2 6 9 C. Inducible Transcription The ability to precisely regulate spatial and temporal expression of a particular gene is likely to have a significant impact in the field of human gene therapy. In order to be effective, such an approach must necessarily fulfill several criteria, including: (1) biological safety, (2) ease of administration, (3) low^ basal expression, (4) high and gene-specific inducibility, (5) reversibility, and (6) (preferably) of human origin to minimize immunogenicity. A w îde variety of inducible systems for regulating gene expression have been developed. These include the use of metal response promoter [91], heat-shock promoter [92], the glucocorticoid-inducible promoter [93], IPTG-inducible lac repressor/operator system [94, 95], tetracycline-inducible system [96], RU486-inducible system [97], ecdysone- inducible system [100], FK506/rapamycin-inducible system [101], hypoxia- inducible factor 1 system [102], radiation-inducible system [103], and the tamoxifen-inducible system [104]. It is beyond the scope of this chapter to provide in-depth information on all of the above-mentioned inducible systems. Consequently, v̂ e v îll focus on those inducible systems that might have the greatest potential for human gene therapy applications. 1. Tetracycline-Inducible System The tet-inducible system originally developed by Bujard and cow^ork- ers [94, 105] is v^idely used to regulate gene expression. The tet-inducible system is based on the tetracycline resistance operon of E. colt. The system utilizes the specificity of the tet repressor (tetR) for the tet operator sequence (tetO), the sensitivity of tetR to tetracycline, and the potent transactivator function of herpes simplex virus protein VP16. The system is based upon the concept of negatively regulating the transcription of the bacterial resistance gene by tetR protein binding to tetO DNA sequences. Addition of tetracycline or doxycycline causes derepression by binding to the tetR protein, thereby allow^ing transcription to proceed. This has been achieved by employing a tet transactivator (tTA) v\̂ hich is a chimeric tetracycline-repressed transactivator generated by fusing the carboxy terminal of tetR protein to the carboxy ter minal 127 amino acids of VP16. The tTA, w^hen bound to tetracycline, is prevented from binding to seven copies of tetO sequences, v^hich are jux taposed upstream of a minimal human cytomegalovirus promoter, thereby selectively turning off the transcription of the gene in question. Removal of tetracycline results in binding of tTA to the tetO sequences in the tet-inducible promoter, follov^ing which the VP16 moiety of tTA transactivates the target gene by promoting assembly of a transcriptional initiation complex, thereby selectively turning on the gene expression. A recent modification of this system allow^s for selective induction of gene expression in the presence of tetracycline. In this strategy, a mutated tetR, called reverse tTA (rtTA), has been generated by incorporating 4 amino acid changes into tTA, thereby facilitating rtTA 2 7 0 Gardner ef a/. binding to the tetO sequence in the presence of tetracycHne. Another variation involves fusion of tTA v^ith the KRAB repressor domain of the human zinc- finger protein Koxl. Upon binding to tetO sequences, this protein is capable of blocking transcription as far as 3 kb dov^nstream [106]. A further variation has revealed that by placing two minimal promoters in opposite orientations on either side of the tetO sequences, it is possible to simultaneously regulate the expression of tw ô genes from a single plasmid [107]. Massie et al. [108] used the tet-inducible system to generate a recombinant adenoviral vector encoding a deletion in the Rl subunit of the herpes simplex virus type 2 ribonucleotide reductase. Topical and tetracycline-inducible gene expression in transgenic mice carrying a gene under the tet-inducible promoter has been achieved by adenovirus mediated gene transfer and expression of tTA [109]. Rubinchik et al. [110] developed a tet-inducible, double recombinant adenoviral vector expressing a fusion of murine FasL and green fluorescent protein. In this virus, the tet-responsive element and the transactivator element are built into opposite ends of the same vector to avoid enhancer interference. The in vitro expression of FasL-GFP in various cell lines could be conveniently regulated by tetracycline or doxycycline in a dose-dependent manner. 2. FK506/Rapamycin-Inducible System The latest in the armamentarium of inducible gene expression systems are the chemical dimerizers that rely upon drug-dependent recruitment of a transactivation domain to a basal promoter to drive the expression of the ther apeutic gene. The strategy is based upon generating a genetic fusion composed of a heterologous DNA-binding domain and an activation domain v^ith the drug binding domain, thereby enabling a bivalent drug to crosslink the tw ô proteins and reconstitute an active transcription factor. This is achieved by using small cell-permeable immunosuppressive molecules, FK506, rapamycin, and cyclosporine, to bind members of the immunophilin family. The FK506 molecule binds tightly to the cellular protein, FKBP12, w^hile FK1012, a synthetic dimer of FK506, causes dimerization of several chimeric proteins containing FKBP12 [111]. Another synthetic compound, FKCsA, created by fusion betw^een FK506 and cyclosporine A, binds vŝ ith high affinity to FKBP12 and cyclophilin and has been used for inducible transcription of exogenous genes [112]. Hov^ever, the most promising results have been obtained using the heterodimerizer rapamycin, w^hich binds simultaneously to the human proteins FKBP and FRAP [113, 114]. In this system, transcriptional activa tion is achieved through rapamycin induced reconstitution of a transcription factor complex formed by coupling of (a) a unique DNA-binding domain, ZFHD, genetically fused to FKBP and (b) the activation domain of the p65 subunit of nuclear factor kappa B (NFKB), fused with the rapamycin-binding domain of FRAP. This novel approach has been successfully utilized for sta ble in vivo delivery of secreted alkaline phosphate, murine erythropoietin 9. Targeted Adenoviral Vectors II: Transcriptional Targeting 2 7 1 and human growth hormone using eukaryotic expression vectors, adenoviral, retroviral, and adeno-associated viral vectors [115-117]. One of the limita tions of this approach is the growth inhibitory and immunosuppressive activity of rapamycin which is due to the inhibition of endogenous FRAP activity [114]. This limitation can be overcome by nonimmunosuppressive analogs (rapalogs) of rapamycin by incorporating mutations in the FRAP domain that accom modate modified drugs [118, 119]. Considerable progress has also been made in designing novel synthetic dimerizers of the ligand for human FKBP12 and mutated FKBP [119-121]. These studies are suggestive of the potential utility of this novel approach for human gene therapy applications. 3. RU 486 Wang et al. [97] developed a novel regulated transcriptional activator consisting of a truncated ligand-binding domain of the human progesterone receptor, the DNA-binding domain of yeast transcriptional activator GAL4, and a C-terminal fragment of the herpes simplex virus VP16 transcriptional regulator protein. This novel transcriptional activator binds with high affinity to the synthetic progesterone antagonist RU 486 but binds very poorly to progesterone. In conjunction with the target gene containing
four copies of the consensus GAL4 binding site, the gene expression was activated only in the presence of RU 486 [97, 98]. Wang et al. [99] also developed an inducible repressor system by substituting the KRAB transcriptional repressor domain for the VP16 transactivation domain. In addition to RU 486, this system can be activated by other synthetic progesterone antagonists at low concentration. The efficacy of this system has been demonstrated using an ex vivo transplantation approach in which cells containing stably integrated chimeric regulator GLVP and a target gene (tyrosine hydroxylase) were grafted in rats. One of the caveats of this system is the low but distinctive basal activity of the GAL4-responsive promoter in the absence of RU 486. Consequently, this system has been refined by designing a synthetic transcription factor which contains a 35-amino- acid truncation of the progesterone receptor rather than the 42-amino-acid truncation [123]. This system exhibits two- to threefold lower basal activities as compared to the earlier version. IV. Enhanced Control of Tronsgene Expression A. Safety Improvements Prior to initiating our clinical trial with Ad-OC-TK, we performed a distribution study that measured TK activity in a variety of organs har vested 3 days after intravenous (iv) injection of Ad-CMV-TK ( 2 x 1 0 ^ pfu) or 2 7 2 Gardner ef o/. Ad-OC-TK (2 X 10^ pfu) with three mice per group. TK enzymatic activity was detected only in the AdCMV-TK group (hver and spleen only), but not the Ad-OC-TK group. Next we performed a comparative study in which 10 C57/BL mice received one iv injection of 2 x 10^ pfu of Ad-OC-TK or Ad-CMV-TK and intraperitoneal (ip) GCV. Significant mortality with severe hepatic histopathology was observed in the Ad-CMV-TK/GCV group (90% mortality), while the Ad-OC-TK/GCV administration did not affect survival of the treated animals (100% survival). These data and the above tissue distribu tion studies support the hypothesis that, in syngeneic hosts, the OC promoter is tissue-specific for tumors, since Ad-OC-TK inhibits tumor growth as effectively as do RSV-TK and CMV-TK, but without the generalized toxicity observed with these universal promoters. These findings paralleled the formal GLP toxi cology study in mice and our toxicology profile in our cfinical trial. Others have demonstrated the lethal effects of both universal promoter HSV-TK viruses in mice and rats and hepatotoxicity in humans after intraprostatic injections. B. Potency Concerns The initial concern with a tumor-specific promoter is that the magnitude of the transgene expression would decrease because of the specificity of the promoter. To address this issue, we compared the in vivo growth inhibition associated with intralesional administration of Ad-OC-TK with that of Ad CMV-TK using a rat osteosarcoma (ROS 17/2.8) subcutaneous model. Ten athymic nude mice were injected with 1 x 10^ ROS cells per site in four subcutaneous locations. After establishment of tumor growth at greater than 5 mm diameter, Ad-CMV-TK or Ad-OC-TK were injected intralesionally into five animals (or 20 tumors) each. After viral injection, the animals received ip GCV (three mice, 12 tumors) or phosphate-buffered saline (PBS; two mice, 8 tumors) for a 2-week period. The animals received one additional adenoviral injection 7 days after the first. The tumors were measured weekly and the animals were sacrificed after the second week of GCV or PBS administration. Both Ad-OC-TK and Ad-CMV-TK forms of therapy demonstrated a greater growth-inhibitory effect than was observed with PBS administration. The growth inhibition was superior with the Ad-OC-TK adenovirus. Therefore, the OC promoter has high intrinsic activity rivaling that of the strong universal CMV promoter, at least in ROS cells. V. Future Directions A. Enhancement of Weak But Specific Promoters A wide variety of highly tissue-specific promoters have been evaluated for achieving transcriptional targeting, however, their applicability has been 9. Targeted Adenoviral Vectors II: Transcriptional Targeting 2 7 3 hampered due to weak transcriptional activity. Enhancement of weak tissue specific promoters can be achieved by employing several different strate gies. One of the simplest approach involves (a) deletion of those sequences from the promoter that do not contribute to tissue specificity or transcrip tional activity and (b) incorporation of multiple copies of the enhancer and positive regulatory elements. This approach has been successfully used in the case of PSA promoter [43], tyrosinase promoter [124, 125], and CEA promoter [12]. Another approach involves generation of activating point mutations within the promoter region as has been in the case of AFP promoter [126] and the MDR 1 promoter [127]. Yet another strategy involves selective com bination of multiple positive regulatory and tissue-specific elements to achieve enhancement of weak promoters. This strategy has shown promising results in augmenting melanoma-specific gene expression when the tyrosinase promoter, either alone or in combination with single or dual, tandem melanocyte-specific enhancer, was used to drive the expression of luciferase and the £. colt purine nucleoside phosphorylase gene. Transient expression studies indicated 5- to 500-fold increase in luciferase activity following incorporation of either single or tandem enhancer elements. In another example, when 5-20 muscle-specific transcriptional elements were randomly assembled and linked to the minimal chicken a-actin promoter, sixfold higher activity was observed as compared to the CMV promoter [128]. In case of adenoviral vectors it might be possible to selectively increase specific expression from exogenous promoters by coexpression of modified VAI genes. Using this approach, Eloit et al. [129] were able to achieve 12.5- to 502-fold increased reporter gene expression. The fact that activity of certain E2F- responsive promoters in tumor cells exceeds that achieved in mitotically active normal cells has been exploited for tumor-selective transgene expression using an adenoviral vector in a malignant glioma model [130]. A novel approach involves development of dual-specificity promoters that are both cell-type- specific and cell-cycle-regulated. In this approach the transgene is under the transcriptional control of an artificial heterodimeric transcription factor whose DNA binding domain is expressed from a tissue-specific promoter, whereas the transactivating subunit is transcribed from a cell-cycle-regulated promoter. The feasibility of this approach has been successfully tested in a transient transfection system [131]. Transcriptional targeting of viral replication for selective killing of tumor cells can be achieved by deletion of adenoviral ElB/55-kDa protein which is essential for viral replication but is dispensable in p53-deficient tumor cells. An alternate approach involves generation of a replication-competent adenoviral vector in which El A or El A and ElB genes are under the transcriptional control of tumor-specific promoters like PSA, kallikrein-2, or AFP [47, 132, 133]. 2 7 4 Gardner ef aL B. Improving Specificity with Multiple Promoter Segments Several investigators have placed combinations of promoter sequences in tandem to derive more specific transgene expression. The authors of the following chapter in this book have both laboratory and clinical experience with this approach and this topic is well covered in their chapter. C. Tumor-Specific Oncolysis Several different approaches have been designed to achieved cancer- cell-specific adenoviral replication and subsequent tumor lysis. Based on our previous work in the laboratory and the clinic, we have designed an adenoviral vector that would only replicate in cells, which could activate the osteocalcin promoter. We have recently received approval for OBA #426 using the osteo calcin promoter to transcriptionally regulate adenoviral replication for the treatment of men with metastatic and recurrent prostate cancer. This approach is thoroughly reviewed elsewhere in this volume (see Chapter 10). D. Combined Targeting Approaches The preceding chapter describes elegant methods to achieve trans- ductional targeting. These approaches will allow for the concentration of adenovirus at metastatic tumor deposits after a systemic administration. In collaboration with these investigators we have begun to combine both transductional and transcriptional targeting to allow for both tumor-specific concentration and tumor-specific oncolysis. This approach combines many of the individual strides achieved in adenoviral gene therapy in the past decade and holds great promise for the future of adenoviral cancer gene therapy. VI. Summary In summary, we believe that the success of gene therapy and its general applicability to medicine will be partially linked to the development of effective transcriptional targeting strategies. The main purpose of this chapter was to illustrate to the reader the benefits of transcriptional targeting and how this approach can be used to generate tumor- or tissue-specific gene expression. The main example of the osteocalcin promoter was used because of our laboratory's significant investigation of this promoter. Acknovs^ledgments The authors are supported by the following grant and awards K08 CA 79544-01A2 (TAG), DOD DAMD17-00-1-0027 (TAG), New York Academy of Medicine Edwin Beer Award (TAG), 9. Targeted Adenoviral Vectors II: Transcriptional Targeting 2 7 5 CaPCURE Foundation (TAG, CK), DOD DAMD17-98-1-8643 (CK, TAG), Walther Oncology Center (TAG, CK), and Department of Urology Research Fund (TAG, CK, SPR). References 1. Beerli, R. R., Segal, D. J., Dreier, B., and Barbas, C. F. (1998). Toward controlling gene expression at will: Specific regulation of the erbB-2/HER-2 promoter by using polydactyl zinc finger proteins constructed from modular building blocks. Proc. Natl. Acad. Set. USA 95, 14,628-14,633. 2. Kang, J. S., and Kim, J. S. (2000). Zinc finger proteins as designer transcription factors. / . Biol. Chem. 275, 8742-8748. 3. Kim, J. S., and Pabo, C. O. (1998). Getting a handhold on DNA: Design of poly-zinc finger proteins with femtomolar dissociation constants. Proc. Natl. Acad. Sci. USA 95,2812-2817. 4. Imanishi, M., Hori, Y., Nagaoka, M., and Sugiura, Y. (2000). DNA-bending finger: Artificial design of 6-zinc finger peptides with polyglycine linker and induction of DNA bending. Biochemistry 39, 4383-4390. 5. Corbi, N., Libri, V., FanciuUi, M., Tinsley, J. M., Davies K. E., and Passananti, C. (2000). The artificial zinc finger coding gene 'Jazz' binds the utrophin promoter and activates transcription. Gene Ther. 7, 1076-1083. 6. Moore, M., Klug, A., and Choo, Y. (2001). From the Cover: Improved DNA binding specificity from polyzinc finger peptides by using strings of two-finger units. Proc. Natl. Acad. Sci. USA 98, 1437-1441. 7. Gold, P., Shuster, J., and Freedman, S. O. (1978). Carcinoembryonic antigen (CEA) in clinical medicine: Historical perspectives, pitfalls and projections. Cancer 42, 1399-1405. 8. Thompson, J. A., Pande, H., Paxton, R. J., Shively, L., Padma, A., Simmer, R. L., Todd, C. W., Riggs, A. D., and Shively, J. E. (1987). Molecular cloning of a gene belonging to the carcinoembryonic antigen gene family and discussion of a domain model. Proc. Natl. Acad. Sci. USA 84, 2965-2969. 9. Schrewe, Fl., Thompson, J., Bona, M., Hefta, L. J., Maruya, A., Hassauer, M., Shively, J. E., von Kleist, S., and Zimmermann, W. (1990). Cloning of the complete gene for carcinoem bryonic antigen: Analysis of its promoter indicates a region conveying cell type-specific expression. Mol. Cell Biol. 10, 2738-2748. 10. Hauck, W., and Stanners, C. P. (1995). Transcriptional regulation of the carcinoembryonic antigen gene. Identification of regulatory elements and multiple nuclear factors. / . Biol. Chem. 270, 3602-3610. 11. Chen, C. J., Li, L. J., Maruya, A., and Shively, J. E. (1995). In vitro and in vivo footprint analysis of the promoter of carcinoembryonic antigen in colon carcinoma cells: Effects of interferon gamma treatment. Cancer Res. 55, 3873-3882. 12. Richards, C. A., Austin, E. A., and Huber, B. E. (1995). Transcriptional regulatory sequences of carcinoembryonic antigen: Identification and use with cytosine deaminase for tumor- specific gene therapy. Hum. Gene Ther. 6, 881-893. 13. Cao, G., Kuriyama, S., Gao, J., Mitoro, A., Cui, L., Nakatani, T., Zhang, X., Kikukawa, M., Pan, X., Fukui, H., and Qi, Z. (1998). Comparison of carcinoembryonic antigen promoter regions isolated from human colorectal carcinoma and normal adjacent mucosa to induce strong tumor-selective gene expression. Int. J. Cancer 78, 242-247. 14. Osaki, T., Tanio, Y., Tachibana, I., Hosoe, S., Kumagai, T., Kawase, I., Oikawa, S., and Kishimoto, T. (1994). Gene therapy for carcinoembryonic antigen-producing human lung cancer cells by cell type-specific expression of herpes simplex virus thymidine kinase gene. Cancer Res. 54, 5258-5261. 2 7 6 Gardner ef o/. 15. Takeuchi, M., Shichinohe, T., Senmaru, N., Miyamoto, M., Fujita, H., Takimoto, M., Kondo, S., Katoh, H., and Kuzumaki, N. (2000). The dominant negative H-ras mutant, N116Y, suppresses growth of metastatic human pancreatic cancer cells in the liver of nude mice. Gene Ther. 7, 518-526. 16. Lan, K. H., Kanai, F., Shiratori, Y., Okabe, S., Yoshida, Y., Wakimoto, H., Hamada, H., Tanaka, T., Ohashi, M., and Omata, M. (1996). Tumor-specific gene expression in carci- noembryonic antigen — Producing gastric cancer cells using adenovirus vectors. Gastroen terology 111, 1241-1251. 17. Lan, K. H., Kanai,
F., Shiratori, Y., Ohashi, M., Tanaka, T., Okudaira, T., Yoshida, Y., Hamada, H., and Omata, M. (1997). In vivo selective gene expression and therapy mediated by adenoviral vectors for human carcinoembryonic antigen-producing gastric carcinoma. Cancer Res. 57, 4279-4284. 18. DiMaio, J. M., Clary, B. M., Via, D. F., Coveney, E., Pappas, T. N., and Lyerly, H. K. (1994). Directed enzyme pro-drug gene therapy for pancreatic cancer in vivo. Surgery 116,205-213. 19. Tanaka, T., Kanai, F., Lan, K. H., Ohashi, M., Shiratori, Y., Yoshida, Y., Hamada, H., and Omata, M. (1997). Adenovirus-mediated gene therapy of gastric carcinoma using cancer- specific gene expression in vivo. Biochem. Biophys. Res. Commun. 231, 775-779. 20. Brand, K., Loser, P., Arnold, W., Bartels, T., and Strauss, M. (1998). Tumor cell-specific transgene expression prevents liver toxicity of the adeno-HSVtk/GCV approach. Gene Ther. 5, 1363-1371. 21. Kijima, T., Osaki, T., Nishino, K., Kumagai, T., Funakoshi, T., Goto, H., Tachibana, I., Tanio, Y., and Kishimoto, T. (1999). Application of the Cre recombinase/loxP system further enhances antitumor effects in cell type-specific gene therapy against carcinoembryonic antigen-producing cancer. Cancer Res. 59, 4906-4911. 22. Belanger, L., Baril, P., Guertin, M., Gingras, M. C., Gourdeau, H., Anderson, A., Hamel, D., and Boucher, J. M. (1983). Oncodevelopmental and hormonal regulation of alpha 1- fetoprotein gene expression. Adv. Enzyme Regul. 21, 73-99. 23. Nahon, J. L., Danan, J. L., Poiret, M., Tratner, L, Jose-Estanyol, M. and Sala-Trepat, J. M. (1987). The rat alpha-fetoprotein and albumin genes. Transcriptional control and compari son of the sequence organization and promoter region. / . Biol. Chem. 262, 12,479-12,487. 24. Sakai, M., Morinaga, T., Urano, Y., Watanabe, K., Wegmann, T. G., and Tamaoki, T. (1985). The human alpha-fetoprotein gene. Sequence organization and the 5' flanking region./. Biol. Chem. 260, 5055-5060. 25. Nakabayashi, H., Hashimoto, T., Miyao, Y., Tjong, K. K., Chan, J., and Tamaoki, T. (1991). A position-dependent silencer plays a major role in repressing alpha- fetoprotein expression in human hepatoma. Mol. Cell Biol. 11, 5885-5893. 26. Kaneko, S., Hallenbeck, P., Kotani, T., Nakabayashi, H., McGarrity, G., Tamaoki, T., Anderson, W. F., and Chiang, Y. L. (1995). Adenovirus-mediated gene therapy of hepatocellular carcinoma using cancer-specific gene expression. Cancer Res. 55, 5283-5287. 27. Kanai, F., Shiratori, Y., Yoshida, Y., Wakimoto, H., Hamada, H., Kanegae, Y., Saito, L, Nakabayashi, H., Tamaoki, T., Tanaka, T., Lan, K. H., Kato, N., Shiina, S., and Omata, M. (1996). Gene therapy for alpha-fetoprotein-producing human hepatoma cells by adenovirus- mediated transfer of the herpes simplex virus thymidine kinase gene. Hepatology 23, 1359-1368. 28. Kanai, F., Lan, K. H., Shiratori, Y., Tanaka, T., Ohashi, M., Okudaira, T., Yoshida, Y., Wakimoto, H., Hamada, H., Nakabayashi, H., Tamaoki, T., and Omata, M. (1997). In vivo gene therapy for alpha-fetoprotein-producing hepatocellular carcinoma by adenovirus- mediated transfer of cytosine deaminase gene. Cancer Res. 57, 461-465. 29. Arbuthnot, P. B., Bralet, M. P., Le Jossic, C , Dedieu, J. F., Perricaudet, M., Brechot, C , and Ferry, N. (1996). In vitro and in vivo hepatoma cell-specific expression of a gene transferred with an adenoviral vector. Hum. Gene Ther. 7, 1503-1514. 9. Targeted Adenoviral Vectors II: Transcriptional Targeting 2 7 7 30. Bui, L. A., Butterfield, L. H., Kim, J. Y., Ribas, A., Seu, P., Lau, R., Glaspy, J. A., McBride, W. H., and Economou, J. S. (1997). In vivo therapy of hepatocellular carcinoma with a tumor-specific adenoviral vector expressing interleukin-2. Hum. Gene Ther. 8, 2173-2182. 31. Sato, Y., Tanaka, K., Lee, G., Kanegae, Y., Sakai, Y., Kaneko, S., Nakabayashi, H., Tamaoki, T., and Saito, I. (1998). Enhanced and specific gene expression via tissue-specific production of Cre recombinase using adenovirus vector. Biochem. Biophys. Res. Commun. 244, 455-462. 32. Riegman, P. H., Vlietstra, R. J., van der Korput,J. A., Romijn, J. C , and Trapman, J. (1989). Characterization of the prostate-specific antigen gene: A novel human kallikrein-like gene. Biochem. Biophys. Res. Commun. 159, 95-102. 33. Riegman, P. H., VHetstra, R. J., Klaassen, P., van der Korput, J. A., Geurts van Kessel, A., Romijn, J. C., and Trapman, J. (1989). The prostate-specific antigen gene and the human glandular kalHkrein-1 gene are tandemly located on chromosome 19. FEBS Lett. 247, 123-126. 34. Lundwall, A, (1989). Characterization of the gene for prostate-specific antigen, a human glandular kallikrein. Biochem. Biophys. Res. Commun. 161, 1151-1159. 35. Wang, M. C , Valenzuela, L. A., Murphy, G. P., and Chu, T. M. (1979). Purification of a human prostate specific antigen. Invest. Urol. 17, 159-163. 36. Wang, M. C , Papsidero, L. D., Kuriyama, M., Valenzuela, L. A., Murphy, G. P., and Chu, T. M. (1981). Prostate antigen: A nev^ potential marker for prostatic cancer. Prostate 2, 89-96. 37. Gallee, M. P., van Vroonhoven, C. C , van der Korput, H. A., van der Kwast, T. H., ten Kate, F. J,, Romijn, J. C , and Trapman, J. (1986). Characterization of monoclonal antibod ies raised against the prostatic cancer cell line PC-82. Prostate 9, 33-45. 38. Riegman, P. H., Vlietstra, R. J., van der Korput, H. A., Romijn, J. C , and Trapman, J. (1991). Identification and androgen-regulated expression of two major human glandular kallikrein-1 (hGK-1) mRNA species. Mol. Cell Endocrinol. 76, 181-190. 39. Cleutjens, C. B., Steketee, K., van Eekelen, C. C , van der Korput, J. A., Brinkmann, A. O., and Trapman, J. (1997). Both androgen receptor and glucocorticoid receptor are able to induce prostate-specific antigen expression, but differ in their growth- stimulating properties of LNCaP cells. Endocrinology 138, 5293-5300. 40. Cleutjens, K. B., van der Korput, H. A., van Eekelen, C. C , van Rooij, H. C , Faber, P. W., and Trapman, J. (1997). An androgen response element in a far upstream enhancer region is essential for high, androgen-regulated activity of the prostate-specific antigen promoter. Mol. Endocrinol. 11, 148-161. 41. Cleutjens, K. B., van Eekelen, C. C , van der Korput, H. A., Brinkmann, A. O., and Trap- man, J. (1996). Two androgen response regions cooperate in steroid hormone regulated activity of the prostate-specific antigen promoter./. Biol. Chem. 271, 6379-6388. 42. Cleutjens, K. B., van der Korput, H. A., Ehren-van Eekelen, C. C , Sikes, R. A., Fasciana, C , Chung, L. W., and Trapman, J. (1997). A 6-kb promoter fragment mimics in transgenic mice the prostate-specific and androgen-regulated expression of the endogenous prostate-specific antigen gene in humans. Mol. Endocrinol. 11, 1256-1265. 43. Pang, S., Dannull, J., Kaboo, R., Xie, Y., Tso, C. L., Michel, K., deKernion, J. B., and Bellde- grun, A. S. (1997). Identification of a positive regulatory element responsible for tissue- specific expression of prostate-specific antigen. Cancer Res. 57, 495-499. 44. Yeung,F., Li, X., Ellett, J., Trapman,]., Kao, C , and Chung, L. W. (2000). Regions of prostate-specific antigen (PSA) promoter confer androgen-independent expression of PSA in prostate cancer cells./. Biol. Chem. 275, 40,846-40,855. 45. Oettgen, P., Finger, E., Sun, Z., Akbarali, Y., Thamrongsak, U., Boltax, J., Grail, F., Dube, A., Weiss, A., Brown, L., Quinn, G., Kas, K., Endress, G., Kunsch, C , and 2 7 8 Gardner ef a/. Libermann, T. A. (2000). PDEF, a novel prostate epithelium-specific ets transcription factor, interacts with the androgen receptor and activates prostate-specific antigen gene expression. / . Biol. Chem. 275, 1216-1225. 46. Latham, J. P., Searle, P. F., Mautner, V., and James, N. D. (2000). Prostate-specific antigen promoter/enhancer driven gene therapy for prostate cancer: Construction and testing of a tissue-specific adenovirus vector. Cancer Res. 60, 334-341. 47. Rodriguez, R., Schuur, E. R., Lim, H. Y., Henderson, G. A., Simons, J. W., and Hender son, D. R. (1997). Prostate attenuated replication competent adenovirus (ARCA) CN706: A selective cytotoxic for prostate-specific antigen-positive prostate cancer cells. Cancer Res. 57,2559-2563. 48. Martinello-Wilks, R., Garcia-Aragon, J., Daja, M. M., Russell, P., Both, G. W., Mol- loy, P. L., Lockett, L. J., and Russell, P. J. (1998). In vivo gene therapy for prostate cancer: Preclinical evaluation of two different enzyme-directed prodrug therapy systems delivered by identical adenovirus vectors. Hum. Gene Ther. 9, 1617-1626. 49. Gotoh, A., Ko, S. C., Shirakawa, T., Cheon, J., Kao, C., Miyamoto, T., Gardner, T. A., Ho, L. J., Cleutjens, C. B., Trapman, J., Graham, F. L., and Chung, L. W. (1998). Devel opment of prostate-specific antigen promoter-based gene therapy for androgen-independent human prostate cancer. / . Urol. 160, 220-229. 50. Lee, S. E., Jin, R. J., Lee, S. G., Yoon, S. J., Park, M. S., Heo, D. S., and Choi, H. (2000). Development of a new plasmid vector with PSA-promoter and enhancer expressing tissue- specificity in prostate carcinoma cell lines. Anticancer Res. 20, 417-422. 51. Xu, L. L., Srikantan, v. , Sesterhenn, L A., Augustus, M., Dean, R., Moul, J. W., Carter, K. C , and Srivastava, S. (2000). Expression profile of an androgen regulated prostate specific homeobox gene NKX3.1 in primary prostate cancer. / . Urol. 163, 972-979. 52. Pan, L. C , and Price, P. A. (1984). The effect of transcriptional inhibitors on the bone gamma-carboxyglutamic acid protein response to 1,25-dihydroxyvitamin D3 in osteosar coma cells./. Biol. Chem. 259, 5844-5847. 53. Price, P. A., and WiUiamson, M. K, (1985). Primary structure of bovine matrix Gla protein, a new vitamin K- dependent bone protein./. Biol. Chem. 260, 14,971-14,975. 54. Kerner, S. A., Scott, R. A., and Pike, J. W. (1989). Sequence elements in the human osteo calcin gene confer basal activation and inducible response to hormonal vitamin D3. Proc. Natl. Acad. Sci. USA 86, 4455-4459. 55. Lian, J., Stewart, C , Puchacz, E., Mackowiak, S., Shalhoub, V., Collart, D., Zambetti, G., and Stein, G. (1989). Structure of the rat osteocalcin gene and regulation of vitamin D- dependent expression. Proc. Natl. Acad. Sci. USA 86, 1143-1147. 56. Desbois, C , Hogue, D. A., and Karsenty, G. (1994). The mouse osteocalcin gene cluster contains three genes with two separate spatial and temporal patterns of expression. / . Biol. Chem. 269, 1183-1190. 57. Desbois, C , Seldin, M. F., and Karsenty, G. (1994). Localization of the osteocalcin gene cluster on mouse chromosome 3. Mamm. Genome 5, 321-322. 58. Montecino, M., Pockwinse, S., Lian, J., Stein, G., and Stein, J. (1994). DNase I hypersensitive sites in promoter elements associated with basal and vitamin D dependent transcription of the bone-specific osteocalcin gene. Biochemistry 33, 348-353. 59. Hoffmann, H. M., Catron, K. M., van Wijnen, A. J., McCabe, L. R., Lian, J. B., Stein, G. S., and Stein, J. L. (1994). Transcriptional control of the tissue-specific, developmentally regu lated osteocalcin gene requires a binding motif for the Msx family of homeodomain proteins. Proc. Natl. Acad. Sci. USA 91, 12,887-12,891. 60. Towler, D. A., Rutledge, S. J., and Rodan, G. A. (1994). Msx-2/Hox 8.1: A transcriptional regulator of the rat osteocalcin promoter. Mol. Endocrinol. 8, 1484-1493. 61. Towler, D. A., Bennett, C. D., and Rodan, G. A. (1994). Activity of the rat osteocalcin basal promoter in osteoblastic cells is dependent upon homeodomain and CPl binding motifs. Mol. Endocrinol. 8, 614-624. 9. Targeted Adenoviral Vectors II: Transcriptional Targeting 2 7 9 62. Tamura, M., and Noda, M. (1994). Identification of a DNA sequence involved in osteoblast- specific gene expression via interaction w îth helix-loop-helix (HLH)-type transcription factors./. Cell Biol. 126, 773-782. 63. Banerjee, C , Hiebert, S. W., Stein, J. L., Lian, J. B., and Stein, G. S. (1996). An AML-1 consensus sequence binds an osteoblast-specific complex and transcriptionally activates the osteocalcin gene. Proc. Natl. Acad. Set. USA 93, 4968-4973. 64. Merriman, H. L., van Wijnen, A. J., Hiebert, S., Bidv^ell, J. P., Fey, E., Lian, J., Stein, J., and Stein, G. S. (1995). The tissue-specific nuclear matrix protein, NMP-2, is a member of the AML/CBF/PEBP2/runt domain transcription factor family: Interactions with the osteocalcin gene promoter. Biochemistry 34, 13,125-13,132. 65. Geoffroy, v. , Ducy, P., and Karsenty, G. (1995). A PEBP2 alpha/AML-1-related factor increases osteocalcin promoter activity through its binding to an osteoblast-specific cis-acting element./. Biol. Chem. 270, 30,973-30,979. 66. Banerjee, C., Hiebert, S. W., Stein, J. L., Lian, J. B., and Stein, G. S. (1996). An AML-1 consensus sequence binds an osteoblast-specific complex and transcriptionally activates the osteocalcin gene. Proc. Natl. Acad. Set. USA 93, 4968-4973. 67. Markose, E. R., Stein, J. L., Stein, G. S., and Lian, J. B. (1990). Vitamin D-mediated mod ifications in protein-DNA interactions at tw ô promoter elements of the osteocalcin gene. Proc. Natl. Acad. Sei. USA 87, 1701-1705. 68. Demay, M. B., Gerardi, J. M., DeLuca, H. F., and Kronenberg, H. M. (1990). DNA sequences in
the rat osteocalcin gene that bind the 1,25- dihydroxyvitamin D3 receptor and confer responsiveness to 1,25- dihydroxyvitamin D3. Proc. Natl. Acad. Sei. USA 87, 369-373. 69. Breen, E. C., van Wijnen, A. J., Lian, J. B., Stein, G. S., and Stein, J. L. (1994). In vivo occupancy of the vitamin D responsive element in the osteocalcin gene supports vitamin D-dependent transcriptional upregulation in intact cells. Proc. Natl. Acad. Set. USA 91, 12,902-12,906. 70. Montecino, M., Frenkel, B., Lian, J., Stein, J., and Stein, G. (1996). Requirement of distal and proximal promoter sequences for chromatin organization of the osteocalcin gene in bone-derived cells. / . Cell Biochem. 63, 221-228. 71. Bidwell, J. P., Van Wijnen, A. J., Fey, E. G., Dworetzky, S., Penman, S., Stein, J. L., Lian, J. B., and Stein, G. S. (1993). Osteocalcin gene promoter-binding factors are tissue-specific nuclear matrix components. Proc. Natl. Acad. Sei. USA 90, 3162-3166. 72. Yoon, K. G., Rutledge, S. J., Buenaga, R. F., and Rodan, G. A. (1988). Characterization of the rat osteocalcin gene: Stimulation of promoter activity by 1,25-dihydroxyvitamin D3. Biochemistry 27, 8521-8526. 73. Briggs, M. R., Kadonaga, J. T., Bell, S. P., and Tjian, R. (1986). Purification and biochemical characterization of the promoter-specific transcription factor Spl. Science 234, 47-52. 74. Theofan, G., Haberstroh, L. M., and Price, P. A. (1989). Molecular structure of the rat bone Gla protein gene and identification of putative regulatory elements. DNA 8, 213-221. 75. Goto, K., Heymont, J. L., Klein-Nulend, J., Kronenberg, H. M., and Demay, M. B. (1996). Identification of an osteoblastic silencer element in the first intron of the rat osteocalcin gene. Biochemistry 35, 11,005-11,011. 76. Kearns, A. E., Goto, K., Gianakakos, G., Lippmann, W., and Demay, M. B. (1999). Tran scriptional repression of the rat osteocalcin gene: Role of two intronic CCTCCT motifs. Endocrinology 140, 4120-4126. 77. Ko, S. C., Cheon, J., Kao, C., Gotoh, A., Shirakawa, T., Sikes, R. A., Karsenty, G., and Chung, L. W. (1996). Osteocalcin promoter-based toxic gene therapy for the treatment of osteosarcoma in experimental models. Cancer Res. 56, 4614-4619. 78. Cheon, J., Ko, S. C , Gardner, T. A., Shirakawa, T., Gotoh, A., Kao, C , and Chung, L. W. (1997). Chemogene therapy: Osteocalcin promoter-based suicide gene therapy in combina tion with methotrexate in a murine osteosarcoma model. Cancer Gene Ther. 4, 359-365. 2 8 0 Gardner ef al. 79. Shirakawa, T., Ko, S. C , Gardner, T. A., Cheon, J., Miyamoto, T., Gotoh, A., Chung, L. W., and Kao, C. (1998). In vivo suppression of osteosarcoma pulmonary metastasis with intravenous osteocalcin promoter-based toxic gene therapy. Cancer Gene Ther. 5, 274-280. 80. Hou, Z., Nguyen, Q., Frenkel, B., Nilsson, S. K., Milne, M., van Wijnen, A. J., Stein, J. L., Quesenberry, P., Lian, J. B., and Stein, G. S. (1999). Osteoblast-specific gene expression after transplantation of marrov^ cells: Implications for skeletal gene therapy. Proc. Natl. Acad. Set. USA 96, 7294-7299. 81. Opavsky, R., Pastorekova, S., Zelnik, V., Gibadulinova, A., Stanbridge, E. J., Zavada, J., Kettmann, R., and Pastorek, J. (1996). Human MN/CA9 gene, a novel member of the carbonic anhydrase family: Structure and exon to protein domain relationships. Genomics 33,480-487. 82. Der, C. J., and Stanbridge, E. J. (1981). A tumor-specific membrane phosphoprotein marker in human cell hybrids. Cell 26, 429-438. 83. Zavada, J., Zavadova, Z., Pastorekova, S., Giampor, P., Pastorek, J., and Zelnik, V. (1993). Expression of MaTu-MN protein in human tumor cultures and in clinical specimens. Int. J. Cancer 54,26^-274. 84. Liao, S. Y., Brewer, C , Zavada, J., Pastorek, J., Pastorekova, S., Manetta, A., Berman, M. L., DiSaia, P. J., and Stanbridge, E. J. (1994). Identification of the MN antigen as a diagnostic biomarker of cervical intraepithelial squamous and glandular neoplasia and cervical carcinomas. Am. J. Pathol. 145, 598-609. 85. Pastorekova, S., Parkkila, S., Parkkila, A. K., Opavsky, R., Zelnik, V., Saarnio, J., and Pas torek, J. (1997). Garbonic anhydrase IX, MN/CA IX: Analysis of stomach complementary DNA sequence and expression in human and rat alimentary tracts. Gastroenterology 112, 398-408. 86. Pastorek, J., Pastorekova, S., Gallebaut, I., Mornon, J. P., Zelnik, V., Opavsky, R., Zat'ovicova, M., Liao, S., Portetelle, D., Stanbridge, E. J., et al. (1994). Cloning and characterization of MN, a human tumor-associated protein with a domain homologous to carbonic anhydrase and a putative helix-loop-helix DNA binding segment. Oncogene 9, 2877-2888. 87. Ono, M. (1986). Molecular cloning and long terminal repeat sequences of human endogenous retrovirus genes related to types A and B retrovirus genes. / . Virol. 58, 937-944. 88. Locker, J., and Buzard, G. (1990). A dictionary of transcription control sequences. DNA Seq. 1 ,3 -11 . 89. Imagawa, M., Chiu, R., and Karin, M. (1987). Transcription factor AP-2 mediates induction by two different signal- transduction pathways: Protein kinase C and cAMP. Cell 51, 251-260. 90. el-Deiry, W. S., Kern, S. E., Pietenpol, J. A., Kinzler, K. W., and Vogelstein, B. (1992). Defi nition of a consensus binding site for p53. Nat. Genet. 1, 45-49. 91. Searle, P. F., Stuart, G. W., and Palmiter, R. D. (1985). Building a metal-responsive promoter with synthetic regulatory elements. Mol. Cell Biol. 5, 1480-1489. 92. Fuqua, S. A., Blum-Salingaros, M., and McGuire, W. L. (1989). Induction of the estrogen- regulated "24 K" protein by heat shock. Cancer Res. 49, 4126-4129. 93. Hirt, R. P., Easel, N., and Kraehenbuhl, J. P. (1994). Inducible protein expression using a glucocorticoid-sensitive vector. Methods Cell Biol. 43, 247-262. 94. Figge, J., Wright, C , Collins, C. J., Roberts, T. M., and Livingston, D. M. (1988). Stringent regulation of stably integrated chloramphenicol acetyl transferase genes by E. coli lac repressor in monkey cells. Cell 52, 713-722. 95. Baim, S, B., Labow, M. A., Levine, A. J., and Shenk, T. (1991). A chimeric mammalian transactivator based on the lac repressor that is regulated by temperature and isopropyl beta-D-thiogalactopyranoside. Proc. Natl. Acad. Sci. USA 88, 5072-5076. 9. Targeted Adenoviral Vectors II: Transcriptional Targeting 2 8 1 96. Gossen, M., and Bujard, H. (1992). Tight control of gene expression in mammalian cells by tetracycline- responsive promoters. Proc. Natl. Acad. Set. USA 89, 5547-5551. 97. Wang, Y., O'Malley, B. W., and Tsai, S. Y. (1994). A regulatory system for use in gene transfer. Proc. Natl. Acad. Sci. USA 91, 8180-8184. 98. Wang,Y., DeMayo, F. J., Tsai, S. Y., and O'Malley, B. W. (1997). Ligand-inducible and liver-specific target gene expression in transgenic mice. Nat. Biotechnol. 15, 239-243. 99. Wang, Y., Xu, J., Pierson, T., O'Malley, B. W., and Tsai, S. Y. (1997). Positive and negative regulation of gene expression in eukaryotic cells w îth an inducible transcriptional regulator. Gene Ther. 4,432-441. 100. No, D., Yao, T. P., and Evans, R. M. (1996). Ecdysone-inducible gene expression in mam malian cells and transgenic mice. Proc. Natl. Acad. Sci. USA 93, 3346-3351. 101. Rivera, V. M., Clackson, T., Natesan, S., Pollock, R., Amara, J. F., Keenan, T., Magari, S. R., Phillips, T., Courage, N. L., CerasoU, F., Holt, D. A., and Oilman, M. (1996). A humanized system for pharmacologic control of gene expression. Nat. Med. 2,1028-1032. 102. Dachs, G. U., Patterson, A. V., Firth, J. D., Ratcliffe, P. J., Townsend, K. M., Stratford, I. J., and Harris, A. L. (1997). Targeting gene expression to hypoxic tumor cells. Nat. Med. 3, 515-520. 103. Scott, S. D., Marples, B., Hendry, J. H., Lashford, L. S., Embleton, M. J., Hunter, R. D., Howell, A., and Margison, G. P. (2000). A radiation-controlled molecular sw îtch for use in gene therapy of cancer. Gene Ther. 7, 1121-1125. 104. Putzer, B. M., Stiev^e, T., Crespo, F., and Esche, H. (2000). Improved safety through tamoxifen-regulated induction of cytotoxic genes delivered by Ad vectors for cancer gene therapy. Gene Ther. 7, 1317-1325. 105. Gossen, M., Freundlieb, S., Bender, G., Muller, G., Hillen, W., and Bujard, H. (1995). Tran scriptional activation by tetracyclines in mammalian cells. Science 268, 1766-1769. 106. Deuschle, U., Meyer, W. K., and Thiesen, H. J. (1995). Tetracycline-reversible silencing of eukaryotic promoters. Mol. Cell Biol. 15, 1907-1914. 107. Baron, U., Freundlieb, S., Gossen, M., and Bujard, H. (1995). Co-regulation of tv^o gene activities by tetracycline via a bidirectional promoter. Nucleic Acids Res. 23, 3605-3606. 108. Massie, B., Couture, F., Lamoureux, L., Mosser, D. D., Guilbault, C , Jolicoeur, P., Belanger, F., and Langelier, Y. (1998). Inducible overexpression of a toxic protein by an adenovirus vector with a tetracycline-regulatable expression cassette. / . Virol. 72, 2289-2296. 109. Ghersa, P., Gobert, R. P., Sattonnet-Roche, P., Richards, C. A., Merlo Pich, E., and Hooft van Huijsduijnen, R. (1998). Highly controlled gene expression using combinations of a tissue- specific promoter, recombinant adenovirus and a tetracycline-regulatable transcription factor. Gene Ther. 5, 1213-1220. 110. Rubinchik, S., Ding, R., Qiu, A. J., Zhang, F., and Dong, J. (2000). Adenoviral vector which delivers FasL-GFP fusion protein regulated by the tet-inducible expression system. Gene Ther. 7, 875-885. 111. Liang, X., Hartikka, J., Sukhu, L., Manthorpe, M., and Hobart, P. (1996). Novel, high expressing and antibiotic-controlled plasmid vectors designed for use in gene therapy. Gene Ther. 3, 350-356. 112. Belshaw, P. J., Spencer, D. M., Crabtree, G. R., and Schreiber, S. L. (1996). Controlling pro grammed cell death with a cyclophifin-cyclosporin-based chemical inducer of dimerization. Chem.Biol. 3 ,731-738. 113. Standaert, R. F., Galat, A., Verdine, G. L., and Schreiber, S. L. (1990). Molecular cloning and overexpression of the human FK506-binding protein FKBP. Nature 346, 671-674. 114. Brown, E. J., Albers, M. W., Shin, T. B., Ichikawa, K., Keith, C. T., Lane, W. S., and Schreiber, S. L. (1994). A mammalian protein targeted by Gl-arresting rapamycin-receptor complex. Nature 369, 756-758. 2 8 2 Gardner ef o/. 115. Magari, S. R., Rivera, V. M., luliucci, J. D., Gilman, M., and Cerasoli, F. (1997). Pharmaco logic control of a humanized gene therapy system implanted into nude mice. / . Clin. Invest. 100,2865-2872. 116. Ye, X., Rivera, V.M., Zoltick, P., Cerasoli, F., Schnell, M. A., Gao, G., Hughes, J. V., Gilman, M., and Wilson, J. M. (1999). Regulated delivery of therapeutic proteins after in vivo somatic cell gene transfer. Science 283, 88 -91 . 117. Rivera, V. M., Ye, X., Courage, N. L., Sachar, J., Cerasoh, F., Wilson, J. M., and Gilman, M. (1999). Long-term regulated expression of growth hormone in mice after intramuscular gene transfer. Proc. Natl. Acad. Sci. USA 96, 8657-8662. 118. Liberies, S. D., Diver, S. T., Austin, D. J., and Schreiber, S. L. (1997). Inducible gene expres sion and protein translocation using nontoxic ligands identified by a mammalian three-hybrid screen. Proc. Natl. Acad. Sci. USA 94, 7825-7830. 119. Clackson, T., Yang, W., Rozamus, L. W., Hatada, M., Amara, J. F., Rollins, C. T., Steven son, L. F., Magari, S. R., Wood, S. A., Courage, N. L., Lu, X., Cerasoh, F., Gilman, M., and Holt, D. A. (1998). Redesigning an FKBP-ligand interface to generate chemical dimerizers with novel specificity. Proc. Natl. Acad. Sci. USA 95, 10,437-10,442. 120. Amara, J. F., Clackson, T., Rivera, V. M., Guo, T., Keenan, T., Natesan, S., Pollock, R., Yang, W., Courage, N. L., Holt, D. A., and Gilman, M. (1997). A versatile synthetic dimer- izer for the regulation of protein-protein interactions. Proc. Natl. Acad. Sci. USA 94, 10,618-10,623. 121. Rollins, C. T., Rivera, V. M., Woolfson, D. N., Keenan, T., Hatada, M., Adams, S. E., Andrade, L. J., Yaeger, D., van Schravendijk, M. R., Holt, D. A., Gilman, M., and Clackson, T. (2000). A ligand-reversible dimerization system for controlling protein-protein interactions. Proc. Natl. Acad. Sci. USA 97, 7096-7101. 122. Wang, Y., Yu, L., and Geller, A. L (1999). Diverse stabilities of expression in the rat brain from different cellular promoters in a helper virus-free herpes simplex virus type 1 vector system. Hum. Gene Ther. 10, 1763-1771. 123. Delort, J. P., and Capecchi, M. R. (1996). TAXI/UAS: A molecular switch to control expres sion of genes in vivo. Hum. Gene Ther. 7, 809-820. 124. Siders, W. M., Halloran, P. J., and Fenton, R. G. (1996). Transcriptional targeting of recom binant adenoviruses to human and murine melanoma cells. Cancer Res. 56, 5638-5646. 125. Siders, W. M., Halloran, P. J., and Fenton, R. G. (1998). Melanoma-specific cytotoxicity induced by a tyrosinase promoter-enhancer/herpes simplex virus thymidine kinase aden ovirus. Cancer Gene Ther. 5, 281-291. 126. Ishikawa, H., Nakata, K., Mawatari, F., Ueki, T., Tsuruta, S., Ido, A., Nakao, K., Kato, Y., Ishii, N., and Eguchi, K. (1999). Utilization of variant-type of human alpha-fetoprotein
promoter in gene therapy targeting for hepatocellular carcinoma. Gene Ther. 6, 465-470. 127. Stein, U., Walther, W., and Shoemaker, R. H. (1996). Vincristine induction of mutant and wild-type human multidrug- resistance promoters is cell-type-specific and dose-dependent. / . Cancer Res. Clin. Oncol. Ill, TIS-1%1. 128. Li, X., Eastman, E. M., Schwartz, R. J., and Draghia-Akli, R. (1999). Synthetic muscle pro moters: Activities exceeding naturally occurring regulatory sequences. Nat. Biotechnol. 17, 241-245. 129. Eloit, M., Adam, M., Gallais, L, and Fournier, A. (1997). High level of transgene expression in cell cultures and in the mouse by replication-incompetent adenoviruses harboring modified VAI genes./. Virol. 71, 5375-5381. 130. Parr, M. J., Manome, Y., Tanaka, T., Wen, P., Kufe, D. W., Kaelin, W. G., and Fine, H. A. (1997). Tumor-selective transgene expression in vivo mediated by an E2F- responsive adenoviral vector. Nat. Med. 3, 1145-1149. 9. Targeted Adenoviral Vectors II: Transcriptional Targeting 2 8 3 131. Nettelbeck, D. M., Jerome, V., and MuUer, R. (1999). A dual specificity promoter system combining cell cycle-regulated and tissue-specific transcriptional control. Gene Ther. 6, 1276-1281. 132. Yu, D. C , Chen, Y., Seng, M., Dilley, J., and Henderson, D. R. (1999). The addition of adenovirus type 5 region E3 enables calydon virus 787 to eliminate distant prostate tumor xenografts. Cancer Res. 59, 4200-4203. 133. Hallenbeck, P. L., Chang, Y. N., Hay, C , Golightly, D., Stewart, D., Lin, J., Phipps, S., and Chiang, Y. L, (1999). A novel tumor-specific replication-restricted adenoviral vector for gene therapy of hepatocellular carcinoma. Hum Gene Ther. 10, 1721-1733. 134. Bilbao, R., Gerolami, R., Bralet, M. P., Qian, C , Tran, P. L., Tennant, B., Prieto, J., and Brechot, C. (2000). Transduction efficacy, antitumoral effect, and toxicity of adenovirus- mediated herpes simplex virus thymidine kinase/ ganciclovir therapy of hepatocellular carcinoma: the woodchuck animal model. Cancer Gene Ther. 7, 657-662. 135. Balague, C , Zhou, J., Dai, Y., Alemany, R., Josephs, S. F., Andreason, G., Hariharan, M., Sethi, E., Prokopenko, E., Jan, H. Y., Lou, Y. C , Hubert-Leslie, D., Ruiz, L., and Zhang, W. W. (2000). Sustained high-level expression of full-length human factor VIII and restoration of clotting activity in hemophilic mice using a minimal adenovirus vector. Blood 95, 820-828. 136. Brann, T., Kayda, D., Lyons, R. M., Shirley, P., Roy, S., Kaleko, M., and Smith, T. (1999). Adenoviral vector-mediated expression of physiologic levels of human factor VIII in nonhuman primates. Hum. Gene Ther. 10, 2999-3011. 137. Draghia-AkH, R., Li, X., and Schwartz, R. J. (1997). Enhanced growth by ectopic expression of growth hormone releasing hormone using an injectable myogenic vector. Nat. Biotechnol. 15,1285-1289. 138. Anderson, L. M., Krotz, S., Weitzman, S. A., and Thimmapaya, B. (2000). Breast cancer- specific expression of the Candida albicans cytosine deaminase gene using a transcriptional targeting approach. Cancer Gene Ther. 7, 845-852. 139. Anderson, L. M., Swaminathan, S., Zackon, I., Tajuddin, A. K., Thimmapaya, B., and Weitzman, S. A. (1999). Adenovirus-mediated tissue-targeted expression of the HSVtk gene for the treatment of breast cancer. Gene Ther. 6, 854-864. 140. Li, Q., Emery, D. W., Fernandez, M., Han, H., and Stamatoyannopoulos, G. (1999). Development of viral vectors for gene therapy of beta-chain hemoglobinopathies: Optimization of a gamma-globin gene expression cassette. Blood 93, 2208-2216. 141. Vassaux, G., Hurst, H. C , and Lemoine, N. R. (1999). Insulation of a conditionally expressed transgene in an adenoviral vector. Gene Ther. 6, 1192-1197. 142. Ring, C. J., Harris, J. D., Hurst, H. C , and Lemoine, N. R. (1996). Suicide gene expression induced in tumour cells transduced with recombinant adenoviral, retroviral and plasmid vectors containing the ERBB2 promoter. Gene Ther. 3, 1094-1103. 143. Staba, M. J., Mauceri, H. J., Kufe, D. W., Hallahan, D. E., and Weichselbaum, R. R. (1998). Adenoviral TNF-alpha gene therapy and radiation damage tumor vasculature in a human malignant gUoma xenograft. Gene Ther. 5, 293-300. 144. Manome, Y., Kunieda, T., Wen, P. Y., Koga, T., Kufe, D. W., and Ohno, T. (1998). Transgene expression in malignant glioma using a replication-defective adenoviral vector containing the Egr-1 promoter: activation by ionizing radiation or uptake of radioactive iododeoxyuridine. Hum. Gene Ther. 9, 1409-1417. 145. Jaggar, R. T., Chan, H. Y., Harris, A. L., and Bicknell, R. (1997). EndotheUal cell-specific expression of tumor necrosis factor-alpha from the KDR or E-selectin promoters following retroviral delivery. Hum. Gene Ther. 8, 2239-2247. 146. Reynolds, P. N., Nicklin, S. A., Kafiberova, L., Boatman, B. G., Grizzle, W. E., Balyasnikova, I. v . . Baker, A. H., Danilov, S. M., and Curiel, D. T. (2001). Combined transductional and transcriptional targeting improves the specificity of transgene expression in vivo. Nat. Biotechnol. 19, 838-842. 2 8 4 Gardner ef aL 147. Smith-Arica, J. R., Morelli, A. E., Larregina, A. T., Smith, J., Lowenstein, P. R., and Castro, M. G. (2000). Cell-type-specific and regulatable transgenesis in the adult brain: adenovirus- encoded combined transcriptional targeting and inducible transgene expression. MoL Ther. 2, 579-587. 148. Cortez, N., Trejo, F., Vergara, P., and Segovia, J. (2000). Primary astrocytes retrovirally transduced with a tyrosine hydroxylase transgene driven by a glial-specific promoter elicit behavioral recovery in experimental parkinsonism./. Neurosci. Res. 59, 39-46. 149. Gazit, G., Hung, G., Chen, X., Anderson, W. F., and Lee, A. S. (1999). Use of the glucose starvation-inducible glucose-regulated protein 78 promoter in suicide gene therapy of murine fibrosarcoma. Cancer Res. 59, 3100-3106. 150. Chen, X., Zhang, D., Dennert, G., Hung, G., and Lee, A. S. (2000). Eradication of murine mammary adenocarcinoma through HSVtk expression directed by the glucose-starvation inducible grp78 promoter. Breast Cancer Res. Treat. 59, 81-90. 151. Miao, C. H., Ohashi, K., Patijn, G. A., Meuse, L., Ye, X., Thompson, A. R., and Kay, M. A. (2000) Inclusion of the hepatic locus control region, an intron, and untranslated region increases and stabilizes hepatic factor IX gene expression in vivo but not in vitro. Mol. Ther. 1, 522-532. 152. Lee, E. J., Anderson, L. M., Thimmapaya, B., and Jameson, J. L. (1999). Targeted expression of toxic genes directed by pituitary hormone promoters: a potential strategy for adenovirus- mediated gene therapy of pituitary tumors./. Clin. Endocrinol. Metab. 84, 786-794. 153. Lee, E. J., Thimmapaya, B., and Jameson, J. L. (2000). Stereotactic injection of adenoviral vectors that target gene expression to specific pituitary cell types: implications for gene therapy. Neurosurgery 46, 1461-1468; discussion 1468-1469. 154. Rinsch, C , Regulier, E., Deglon, N., Dalle, B., Beuzard, Y., and Aebischer, P. (1997). A gene therapy approach to regulated delivery of erythropoietin as a function of oxygen tension. Hum. Gene Ther. 8, 1881-1889. 155. Xie, X., Zhao, X., Liu, Y., Young, C. Y., Tindall, D. J., Slawin, K. M., and Spencer, D. M. (2001). Robust prostate-specific expression for targeted gene therapy based on the human kallikrein 2 promoter. Hum. Gene Ther. 12, 549-561. 156. Yu, D. C , Chen, Y., Seng, M., Dilley, J., and Henderson, D. R. (1999). The addition of adenovirus type 5 region E3 enables calydon virus 787 to eliminate distant prostate tumor xenografts. Cancer Res. 59, 4200-4203. 157. Luna, M. C , Ferrario, A., Wong, S., Fisher, A. M., and Gomer, C. J. (2000). Photodynamic therapy-mediated oxidative stress as a molecular sw îtch for the temporal expression of genes ligated to the human heat shock promoter. Cancer Res. 60, 1637-1644. 158. Hernandez-Alcoceba, R., Pihalja, M., Nunez, G., and Clarke, M. F. (2001). Evaluation of a new dual-specificity promoter for selective induction of apoptosis in breast cancer cells. Cancer Gene Ther. 8, 298-307. 159. Chung, I., Schwartz, P. E., Crystal, R. G., Pizzorno, G., Leavitt, J., and Deisseroth, A. B. (1999). Use of L-plastin promoter to develop an adenoviral system that confers transgene expression in ovarian cancer cells but not in normal mesothelial cells. Cancer Gene Ther. 6, 99-106. 160. Shinoura, N., Koike, H., Furitu, T., Hashimoto, M., Asai, A., Kirino, T., and Hamada, H. (2000). Adenovirus-mediated transfer of caspase-8 augments cell death in gliomas: implication for gene therapy. Hum. Gene Ther. 11, 1123-1137. 161. Chen, H., McCarty, D. M., Bruce, A. T., and Suzuki, K. (1998). Gene transfer and expression in oligodendrocytes under the control of myelin basic protein transcriptional control region mediated by adeno-associated virus. Gene Ther. 5, 50-58. 162. Hauser, M. A., Robinson, A., Hartigan-O'Connor, D., Williams-Gregory, D. A., Buskin, J. N., Apone, S., Kirk, C. J., Hardy, S., Hauschka, S. D., and Chamberlain, J. S. (2000). 9. Targeted Adenoviral Vectors II: Transcriptional Targeting 2 8 5 Analysis of muscle creatine kinase regulatory elements in recombinant adenoviral vectors. Mol. Ther. 2, 16-25. 163. Steiner, M. S., Anthony, C. T., Lu, Y., and Holt, J. T. (1998). Antisense c-myc retroviral vector suppresses established human prostate cancer. Hum. Gene Ther. 9, 747-755. 164. Ou, Y. C , Yeung, F., Yang, L., Kao, C , Chung, L. W. K., and Gardner, T. A. (1998). MN promoter activity in renal cell carcinoma cell lines: A potential for tissue restrictive gene therapy. / . Urol. 161, 554. 165. Gardner, T. A., Shalhav, M., Kim, H.-S., Ou, Y.-C, Kim, S.-J., and Kao, C. (2001). Preclinical evaluation of renal cell carcinoma specific-oncolysis using MN promoter mediated transcriptional targeting. Mol. Therapy 3, S982. 166. Kurihara, T., Brough, D. E., Kovesdi, I., and Kufe, D. W. (2000). Selectivity of a replication- competent adenovirus for human breast carcinoma cells expressing the MUCl antigen. / . Clin. Invest. 106, 763-771. 167. Ring, C. J., Blouin, P., Martin, L. A., Hurst, H. C , and Lemoine, N. R. (1997). Use of transcriptional regulatory elements of the MUCl and ERBB2 genes to drive tumour-selective expression of a prodrug activating enzyme. Gene Ther. 4, 1045-1052. 168. Kurihara, H., Zama, A., Tamura, M., Takeda, J., Sasaki, T., and Takeuchi, T. (2000). Glioma/glioblastoma-specific adenoviral gene expression using the nestin gene regulator. Gene Ther. 7, 686-693. 169. MoreUi, A. E., Larregina, A. T., Smith-Arica, J., Dev^ey, R. A., Southgate, T. D., Ambar, B., Fontana, A., Castro, M. G., and Lov^enstein, P. R. (1999). Neuronal and glial cell type-specific promoters v^ithin adenovirus recombinants restrict the expression of the apoptosis-inducing molecule Fas ligand to predetermined brain cell types, and abolish peripheral liver toxicity./. Gen. Virol. 80, 571-583. 170. Klein, R. L., Meyer, E. M., Peel, A. L., Zolotukhin, S., Meyers, C , Muzyczka, N., and King, M. A. (1998). Neuron-specific transduction in the rat septohippocampal or nigrostriatal pathway by recombinant adeno-associated virus vectors. Exp. Neurol. 150, 183-194. 171. Cheon, J., Ko, S. C , Gardner, T. A., Shirakawa, T., Kao, C , and Chung, L.W.K. (1997). Osteosarcoma: Osteocalcin promoter based suicide gene therapy in combination v^ith methotrexate in a murine osteosarcoma model. Cancer Gene Ther. 4, 359-365. 172. Gardner, T. A., Ko, S. C , Kao, C , Cui, Q., Balian, G., and Chung, L.W.K. (1998). Exploiting prostate cancer-bone stromal interaction v^ith gene therapy: Application of osteoblastic tissue specific promoter and bone homing mechanism to deliver therapeutic toxic genes for the treatment of prostate cancer. / . Urol. 159, Abstract 16. 173. Wada, Y., Gardner, T. A., Ko, S. C , Kao, C , Kim, S. J., Shirakawa, T., and Chung, L.W.K. (1999). Widening the therapeutic window of thymidine kinase/GCV gene therapy with a tumor-restricted promoter: Osteocalcin based TK gene therapy for osteosarcoma. Proc. Am. Assoc. Cancer Res. 40, Abstract 87. 174. Gardner, T. A., Ko, S.-C, Kao, C , Shirakawa, T., Cheon, J., Gotoh, A., Wu, T. T., Sikes, R. A., Zhau, H. E., Balian, G., and Chung, W. K. (1998). Exploiting Stromal-epitehlial interaction for model development and new strategies of gene therapy for prostate cancer and osteosarcoma metastases (review). Gene Ther. Mol. Biol. 2, 41-58 . 175. Gardner, T. A., Chung, L.W.K,, Koenemen, K., Gillenwater, J. Y., Myers, C , Fisch, M., Kallmes, D., Wada, Y., Yang, L., and Kao, C. (2000). Preliminary safety report of an osteocalcin-based adenoviral gene therapy (Ad-OC-TK) for metastatic and recurrent prostate cancer: Tumor-restrictive promoter confers safety of toxic gene expression. Mol. Therapy 1, S236. 176. Koeneman, K. S., Kao, C , Ko, S. C , Yang, L., Wada, Y., Kallmes, D. F., Gillenwater, J. Y., Zhau, H. E., Chung, L. W., and Gardner, T. A. (2000). Osteocalcin-directed gene therapy for prostate-cancer bone metastasis. World J. Urol. 18, 102-110. 177. Matsubara, S., Wada, Y., Gardner, T. A., Egawa, M., Park, M. S., Hsieh, C. L., Zhau, H. E., Kao, C , Kamidono, S., Gillenwater, J. Y., and Chung, L. W. (2001). A conditional 2 8 6 Gardner ef a/. replication-competent adenoviral
vector, Ad-OC-Ela, to cotarget prostate cancer and bone stroma in an experimental model of androgen-independent prostate cancer bone metastasis. Cancer Res. 61, 6012-6019. 178. Lu, D., Tamemoto, H., Shibata, H., Saito, I., and Takeuchi, T. (1998). Regulatable production of insulin from primary-cultured hepatocytes: insulin production is up-regulated by glucagon and cAMP and down-regulated by insulin. Gene Ther. 5, 888-895. 179. Kaplitt, M. G., Kwong, A. D., Kleopoulos, S. P., Mobbs, C. V., Rabkin, S. D., and Pfaff, D. W. (1994). Preproenkephalin promoter yields region-specific and long-term expression in adult brain after direct in vivo gene transfer via a defective herpes simplex viral vector. Proc. Natl. Acad. Sci. USA 91, 8979-8983. 180. Xie, X., Zhao, X., Liu, Y., Zhang, J., Matusik, R. J., Slav^in, K. M., and Spencer, D. M. (2001). Adenovirus-mediated tissue-targeted expression of a caspase-9-based artificial death switch for the treatment of prostate cancer. Cancer Res. 61, 6795-6804. 181. Southgate, T. D., Windeatt, S., Smith-Arica, J., Gerdes, C. A., Perone, M. J., Morris, I., Davis, J. R., Klatzmann, D., Lowenstein, P. R., and Castro, M. G. (2000). Transcriptional targeting to anterior pituitary lactotrophic cells using recombinant adenovirus vectors in vitro and in vivo in normal and estrogen/sulpiride-induced hyperplastic anterior pituitaries. Endocrinology 141, 3493-3505. 182. Robertson, M. W., 3rd, Wang, M., Siegal, G. P., Rosenfeld, M., Ashford, R. S., 2nd, Alvarez, R. D., Garver, R. I., and Curiel, D. T. (1998). Use of a tissue-specific promoter for targeted expression of the herpes simplex virus thymidine kinase gene in cervical carcinoma cells. Cancer Gene Ther. 5, 331-336. 183. Kim, S., Lin, H., Barr, E., Chu, L., Leiden, J. M., and Parmacek, M. S. (1997). Transcriptional targeting of replication-defective adenovirus transgene expression to smooth muscle cells in vivo./. Clin. Invest. 100, 1006-1014. 184. Harrod, K. S., Hermiston, T. W., Trapnell, B. C., Wold, W. S., and Whitsett, J. A. (1998). Lung-specific expression of adenovirus E3-14.7K in transgenic mice attenuates adenoviral vector-mediated lung inflammation and enhances transgene expression. Hum. Gene Ther. 9, 1885-1898. 185. Park, B. J., Brown, C. K., Hu, Y., Alexander, H. R., Horti, J., Raje, S., Figg, W. D., and Bartlett, D. L. (1999). Augmentation of melanoma-specific gene expression using a tandem melanocyte-specific enhancer results in increased cytotoxicity of the purine nucleoside phosphorylase gene in melanoma. Hum. Gene Ther. 10, 889-898. 186. Emiliusen, L., Gough, M., Bateman, A., Ahmed, A., Voellmy, R., Chester, J., Diaz, R. M., Harrington, K., and Vile, R. (2001). A transcriptional feedback loop for tissue-specific expression of highly cytotoxic genes which incorporates an immunostimulatory component. Gene Ther. 8, 987-998. C H A P T E R Development of Attenuated Replication Competent Adenoviruses (ARCAs) for the Treatment of Prostate Cancer Daniel R. Henderson and De-Chao Yu Calydon, Incorporated Sunnyvale, California I. Introduction The specificity, or therapeutic index, of anticancer chemotherapy agents has long been problematic. The majority of cancer chemotherapy agents, such as alkylating agents, antimetabolites, antibiotics, plant alkyloids, and other cytotoxic agents, nonspecifically injure or kill dividing cells [1]. These agents are noted for their poor specificity and low therapeutic ratio of toxicity toward target cancer cells compared to normal cells (e.g., therapeutic ratios of 2:1 to 6:1). In some instances, hormonal anticancer agents offer improved specificity [2]. The few biologic response modifiers [3], particularly humanized monoclonal antibodies, also offer greater anti-cancer specificity. However, cytotoxic agents remain the mainstay of cancer chemotherapy. The unwanted toxicity problems, most notably the myeloid stem cell suppression characteristic of cytotoxic drugs, are so great that drugs designed to recover patients from the side-effects of cytotoxic anticancer agents such as G-CSF, GM-CSF, and erythropoetin [4] represent as significant a commercial market as the cytotoxic chemotherapy agents themselves. Intense efforts to increase specific cancer cell cytotoxicity of anticancer agents have frustrated researchers for decades. One such effort is gene ther apy [5, 6]. In experimental models of gene therapy using replication defective adenoviruses (Ads), the use of prodrug converting enzymes such as herpes simplex virus thymidine kinase [7-11] and cytosine deaminase [10] under the control of transcriptional response elements (TREs), has shown anticancer activity in animal models with significantly increased specificity. However, to ADENOVIRAL VECTORS FOR GENE THERAPY 2 8 7 Copyright 2002, Elsevier Science (USA). All rights reserved. 2 8 8 Henderson and Yu destroy a solid tumor in a human, replication-defective adenoviruses must deliver a therapeutic gene and initiate a significant bystander effect all before the host immune response to the adenovirus coat proteins limits further treatment. Unfortunately, in humans even w^hen gene transfer v^as success ful, gene expression from replication-defective vectors has been inadequate or too short-lived. Thus, clinically the limiting issue has been a lack of efficacy. To address some of these shortcomings we have tried to design therapeu tics v^ith sufficient specificity and efficacy so that the short-term expression of adenovirus v îll be successful in killing enough target cancer cells to be med ically useful. Physically, replicating adenoviruses can infect a broad range of human cells and produce infectious progeny that could attack adjacent tumor cells, leading to destruction of a solid tumor w^ithin a short period of time from a single virus treatment. In human patients, tumor-specific replicating adenoviruses w^ould be expected to induce a strong cytotoxic T-cell response confined to the target tumor cells precipitating a cancer vaccine-like response that could help eliminate tumor cells. Replicating viruses have been proposed to treat cancer for nearly a century 112]. The first sustained attempts to treat tumors in animal models occurred in the late 1940s and early 1950s w^hen infections w êre induced v^ith viruses such as avian pest, Russian Far East encephalitis, St. Louis encephalitis, Coxsackie, foot and mouth, herpes simplex, influenzae. West Nile, dengue, Nev^castle disease, vaccinia, and rabies [13]. A significant early attempt was made to explore the cell-killing properties of replicat ing cytolytic adenoviruses for the treatment of cancer in humans. The first isolates of adenovirus were shown to grow "luxuriantly" on HeLa cells, cells originally derived from cervical cancer. It was proposed that perhaps adenoviruses would preferentially replicate in and destroy cervical cancers. Smith et al. [14] tested 10 different wild-type adenovirus serotypes, includ ing adenovirus type 5 (Ad5), as a treatment for locally advanced cervical carcinoma. Virus was administered via intratumoral injection or intraarte rial injection. The virus stocks used were unpurified lysates of tissue culture cells; the number of infectious viral particles (plaque-forming units) and the total number of virus particles in the injected dose were not determined. Although long-term clinical benefit was not achieved, tumor necrosis and cavity formation was observed in 6S% of treated patients via intratumoral injection, and these effects were limited to the carcinoma tissue. Side effects, detected primarily in patients receiving immunosuppression with cortisone, included febrile illness and malaise; in all cases, the symptoms resolved in 7-9 days. Infectious virus was not recovered from any biopsy specimens or vaginal smears, but the titers of neutralizing antibodies were uniformly ele vated by 5 -7 days postinjection [14]. This study is significant, for it illustrates the promise and limitations of oncolytic adenoviruses while describing the 10. Development of ARCAs for the Treatment of Prostate Cancer 2 8 9 limited toxicity to be expected of replicating adenoviruses that contain the E3 region and do not contain transgenes encoding foreign proteins at these intermediate dose levels. However, adenoviruses were subsequently shown to replicate in many cell types and lacked the hoped-for specificity for cervical cancer cells. A resurgence of interest in replicating adenoviruses has occurred in the past decade due to the ability to genetically manipulate viruses. In 1996, Bishoff et al, introduced the use of Ad5 deleted in the ElB-55kD protein so that the virus (ONYX-015 = dll520) [15] preferentially replicates in p53" cells as compared to p53+ cells by a factor of 100-fold. However, the mechanism of antitumor specificity of the ONYX-015 virus has come under criticism [16-20]. In 1997, Rodriquez et al. introduced transcriptional target ing of adenovirus using the enhancer/promoter of the human prostate-specific antigen gene to drive the Ad5 El A gene [21]. We have focused on the use of transcriptional response elements to control the expression of virus genes required for virus replication [21-23]. To test this idea we initially chose prostate cancer and the regulatory enhancer and promoter elements (PSE) of prostate-specific antigen (PSA). PSA is the most widely used serum marker for the diagnosis and management of any form of cancer. It is produced in prostate cancer cells and normal prostate ductal epitheha (which represents less than 5% of the cells of the prostate); it is also produced in much smaller amounts in the periurethral glands (less than 0.01% compared to prostate epitheha) and very rarely in tumors of the skin and salivary gland, but frequently in tumors of the female breast [24]. Since the prostate is an accessory organ, removal or ablation of the entire gland has no serious health repercussions [2, 25-28] , although the side-effects of incontinence and impotence are legendary. A virus destroying all PSA- producing cells would not be expected to attack cells leading to incontinence and impotence. Thus, the regulatory regions of the prostate-specific antigen gene are a reasonable choice for such an approach. We reasoned that placing adenovirus genes under the control of the PSE would create host range mutants or a virus in which replication would be restricted primarily to PSA-producing (PSA+) ductal epithelial cells within the prostate, and PSA+ prostate cancer (PCA) cells. We refer to our geneti cally engineered viruses using transcriptional response elements as attenuated replication-competent adenoviruses (ARCAs). We describe CV706 (PSE driv ing the Ad5 ElA genes and deleted in the Ad5 E3 region), which is currently in clinical trials for localized prostate cancer. Since taking CV706 to clinic, we have focused on improving the specificity and efficacy of the ARCA platform. Below we describe additional prostate-specific viruses on the developmental pathway from CV706, leading to CV787 (probasin driving the ElA gene, PSE driving the ElB gene, and reintroduction of the Ad5 E3 region) and their toxicity and explore the possibility of achieving a better antitumor efficacy 2 9 0 Henderson and Yu by combining ARCA with conventional therapies including radiotherapy and chemotherapy. CV787 is also currently in the clinic. II. ARCAs for Prostate Cancer: CV706 and CV787 A. Adenovirus: Gene Expression and Regulation Members of the human Adenoviridae family were first cultured from the tonsils and adenoids of children in 1953 [29]. They represent 51 different serotypes which are distinguishable by antibody reactivity to epitopes on the virion surface. Each serotype is assigned to one of five subgroups (A-E). Adenovirus type 5 (Ad5), a member of Subgroup C, is associated with a self-limiting, febrile respiratory illness and ocular disease in humans; infectious virus can be recovered from the throat, sputum, urine, and rectum. Ad5 is also associated with renal impairment, hepatic necrosis, and gastric erosions in immunosuppressed individuals [30, 31]. Ad5 and the other Subgroup C viruses have little or no oncogenic potential in mammals [32]. The adenovirus type 5 genome is a double-stranded DNA molecule of 35,935 base pairs [33] containing short inverted terminal repeats [34]. Expression of the genome is a regulated cascade which is arbitrarily divided into early (E) and late (L) phases, with viral DNA replication required for maximal L gene expression. Related RNA transcripts are grouped according to the region of the genome from which they are transcribed as well as by the timing (E or L) of their expression. Viral gene expression is regulated at the levels of transcription, posttranscriptional modification (splicing), translation, and posttranslational modification. Products of the El region are essential for efficient expression of the other regions of the adenovirus genome. The ElA transcription unit is the first Ad sequence to be expressed during viral infection and its products play a crucial role in a number of important biological functions in adenovirus-infected cells. The E2 region encodes several proteins that are required for viral DNA replication. These include a DNA- binding protein [35], the viral DNA-dependent DNA polymerase, and the DNA terminal protein that are required for DNA replication [36-38]. The E3 region is not essential for replication in tissue culture and this region is deleted from most first-generation therapeutic adenoviruses [39, 40]. Proteins encoded by the E3 region modulate host immune responses to infection by inhibiting transport of the MHC class I protein to the
cell surface, thereby impairing the cytotoxic T lymphocyte (CTL) response [41-43], and by blocking TNFA- induced cytolysis of infected cells [44-46]. Significantly, all natural isolates of adenovirus contain the E3 region. Seven transcripts of the E4 region have been identified. Some of the encoded proteins interact with and/or modulate the activity of El region proteins. 10. Development of ARCAs for the Treatment of Prostate Cancer 2 9 1 The onset of viral DNA replication signals the switch from E to L gene expression. Although the precise mechanisms are not fully understood, this transition requires both cis- and trans-acting factors [47-49]. Late genes primarily encode the structural components of the virion and the nonstructural scaffolding proteins that are essential for the assembly of infectious virus. It is estimated that up to 10,000 adenovirus virions accumulate per cell and most remain cell-associated [50]. The entire adenovirus replication cycle is complete in approximately 32-36 h [51]. Host range mutants of adenovirus have played a significant role in elucidation of virus functions. ARCAs using transcriptional response elements create host range mutants where replication is restricted to a particular cell type. The cytotoxicity associated with virus replication (lysis) and the vaccine nature of expressing highly visible foreign capsid antigens should be limited to a certain type of cell. B. Tissue Specificity of ARCA We hypothesized that tropism of a virus could be redirected if expression of an essential viral gene could be controlled. Viruses generated from this approach would have the same capsid as its parental virus and they should be able to penetrate all cell types that express the CAR receptor. Presumably, in all cells containing the CAR receptor, these viruses would follow the normal cell entry process: they would penetrate the endosome, fuse with the endosome membrane, reach the cytoplasm, find transport to the nucleus, and uncoat the viral DNA. In a normal adenovirus replication cycle the ElA gene is the only gene expressed during the first 2.5 h of infection [52-55]. In turn, the ElA proteins as transcription factors upregulate expression of the impending cascade of viral genes. However, we have genetically engineered prostate tissue- specific promoters and enhancers so as to drive the ElA genes. Viral replication should preferentially take place in cells that express the necessary transcription factors, thus enabling activation of the tissue or tumor-specific transcription regulatory elements. Thus, the ElA proteins should be preferentially expressed and the virus preferentially replicate in prostate cells. There are several criteria important in regard to the transcriptional response element (TRE) necessary for the successful engineering of a thera peutic adenovirus: (1) the tissue-specific regulatory specificity must be tightly regulated, and transcription should be limited to tumor cells, or accessory cells with as few other sites of expression as medically tolerable, (2) the TRE must regulate the initiation of transcription of the adjacent gene, (3) the promoter must be strong enough to drive sufficient expression of essential viral genes, and (4) the TRE must be small enough to fit within the packaging limits of adenovirus. We chose prostate cancer and the TREs of PSA as our initial target. Expression of the PSA gene is modulated by the prostate-specific enhancer (PSE) element that is located several thousand nucleotides upstream of the PSA 292 Henderson and Yu promoter [56]. When fused to a fragment (position -230 to +7, relative to the start of transcription) containing the PSA promoter, the PSE (position —5322 to —3875, relative to the start of transcription) confers tissue-specific expression on the reporter gene chloramphenicol acetyl-transferase (CAT) [56]. Sequence analysis of the PSE reveals the presence of regions v^ith homology to steroid-response elements (SREs) and to binding sites for several cellular transcription factors including c-Fos and AP-1 [23, 56., 57]. A functional androgen-response element (ARE) w îthin the PSE increases expression up to 100-fold in the presence of testosterone or the nonmetabolized testosterone analog Rl881. 1. ARC As Containing One Prostate-Specific Transcriptional Response Element To test the feasibility of the ARCA technology, v ê engineered the PSE fragment into the adenovirus genome and generated a first generation virus, CV706. CV706 contains the PSE fragment (PSA promoter and enhancer) inserted immediately upstream of the El A region and transcription of the El A region is regulated by the PSE (Table I). Virus characterization show^ed that CV706 was able to efficiently replicate in PSA+ prostate carcinoma cell lines but not in the other PSA" human cell lines HBL-100, MCF-7, PANC-1, OVCAR-3. CV706 also does not replicate efficiently in DU-145, a prostate cancer cell line which does not express PSA and does not contain the androgen receptor [21]. Further study indicated that the transcription of the ElA mRNA was regulated by the PSE. ElA mRNA was detectable in PSA+ LNCaP cells, Table I ARCAs for Prostate Cancer Virus E l A E l B E3 E4 Targeting cell driven by driven by region driven by CV702 wt wt Deleted wt N/A CV706 PSE wt Deleted wt Prostate cancer CV711 wt PSA Deleted wt Prostate cancer CV716 PSE PSE Deleted wt Prostate cancer CV730 No ElA wt Deleted wt N/A CV737 PB wt Deleted wt Prostate cancer CV738 wt PB Deleted wt Prostate cancer CV739 PB PSE Deleted wt Prostate cancer CV740 PB PB Deleted wt Prostate cancer CV757 wt wt Deleted PSE Prostate cancer CV763 H K 2 wt Deleted wt Prostate cancer CV764 PSE H K 2 Deleted wt Prostate cancer CV787 PB PSE Full-length wt Prostate cancer CV802 wt wt Full-length wt N/A 10. Development of ARC As for the Treatment of Prostate Cancer 2 9 3 but was not detectable in PSA~ cells. El A protein was also reduced by 99% in PSA~ cells, compared to that in the PSA+LNCaP cells [21]. This indicates that the inserted PSE has successfully controlled expression of the El A gene and the host range of this adenovirus mutant has been confined to a particular cell type. We also showed that the tropism of adenovirus could also be changed when the ElB gene or the E4 gene was placed under the control of the PSE TRE. CV711, whose ElB gene is placed under the control of PSE, and CV757, whose E4 genes are driven by PSE, both replicate similarly to wild- type adenovirus in PSA+ cells but are highly attenuated in PSA~ cells. Cell specificity of CV711 viruses is similar to CV706 and replicates similarly to wild-type virus in PSA+cells. In contrast, CV757, shows significantly greater specificity for PSA"^cells. While CV757 grows similarly to wild-type in PSA+ cells, it suffers a very large reduction in the ability to replicate in PSA~ cells (data not shown). Thus, adenovirus mutants can be generated to target PSA+ cells when any one of the ElA, ElB, of E4 genes are driven by the PSE. These observations have been confirmed with other prostate-specific TREs including the TREs for probasin and hK2. The rat probasin (PB) gene is developmentally regulated in the prostate by androgens. Induction of the rat probasin gene by androgens was shown to involve the partic ipation of two different cis-acting DNA elements that bind the androgen receptor. An expression cassette carrying 426 bp of the PB gene promoter and 28 bp of the 5^-untranslated region was found to be sufficient to tar get expression of a bacterial CAT reporter gene specifically to the prostate epithelium [58]. It was also shown that the same 5^-flanking region of PB gene promoter fragment fused to the SV40 TAg gene could lead to the devel opment of progressive forms of prostate disease that histologically resemble human prostate cancer in transgenic animals [58]. The promoter of the rat probasin gene was engineered into adenovirus to drive the expression of either the ElA gene or the ElB gene to generate CV737 and CV738, respec tively. Both CV737 and CV738 showed significant specificity to PSA+prostate carcinoma cells. We also recently cloned the TRE of the human glandular kallikrein (hK2) gene. The hK2 gene is located 12 kb downstream from the PSA gene in a head- to-tail fashion, whereas the hKl gene is located 30 kb upstream of the PSA gene in head-to-head fashion [59]. The PSA and hK2 gene share DNA (80%) and amino acid (78%) sequence homologies that suggests they evolved by gene dupfication from the same ancestral gene [60, 61]. Interestingly, the hK2 protein was recently shown to be expressed in every prostate cancer, and the expression of hK2 protein incrementally increased from benign epithelium, to high-grade prostatic intraepithelial neoplasia, to adenocarcinoma. We recently described CV763 containing the hK2 promoter and enhancer driving the Ad5 El gene. CV763 behaved identically to CV706 [23]. 2 9 4 Henderson and Yu Thus, the repHcation of adenovirus can be restricted to prostate cancer cells when one of the essential adenovirus genes El A, ElB, or E4 is placed under the control of any one of three different prostate-specific TREs. 2. ARC As Containing Tw ô Prostate-Specific Transcriptional Response Elements Since both the El A and ElB genes are essential for adenovirus replication, we reasoned that it was possible to create a virus with significantly higher specificity if both the El A and ElB genes were under independent control of two TREs. To test this hypothesis, we generated an adenovirus mutant CV716, in which both the El A gene and the ElB gene were under the control of PSE. In vitro study showed that CV716 replicated well in the PSA-producing prostate cancer cells. However, replication of CV716 was highly attenuated in nonprostate human cell lines. Compared to CV706, the efficiency of CV716 replication in nonprostate cancer cells has been further reduced by another 100-fold, giving specificity for PSA^ cells compared to PSA" cells of nearly 10,000:1 (data not shown). The high degree of specificity for PSA"̂ cells of CV716 as compared to PSA~cells was found to be universally true [22, 23]. CV740, containing duplicate copies of the rat probasin promoter, also showed this high level of specificity. Unfortunately, CV716 and CV740 are genetically unstable, resulting in self-inactivation of the virus. The ElA gene and one copy of the tissue-specific TRE inserts are deleted during replication. Southern blot analysis of stocks of CV716 indicated a new band when annealed with an EIB-labeled probe. DNA sequence analysis of the cloned deletion mutant indicated that self-inactivation is due to homologous recombination between two identical inserted TREs. In order to make a stable tissue-specific adenovirus we employed two different TREs to drive expression of early essential viral genes. In CV739 the ElA gene and the ElB genes are under the control of the TRE of the rat probasin gene and PSA gene, respectively. CV739 replicates well in PSA"̂ prostate cancer cells, but poorly in nonprostate human cancer cell lines. The cell specificity of CV739 was similar to that of CV716, again showing the roughly 10,000:1 selectivity for PSA+ cells as compared to PSA" cells. However, CV739 is stable. No replication-defective mutants with deleted genomes were found after extensive passages. The same is true for other CV739-like viruses including CV764. CV764 is a stable ARCA variant containing the PSE driving the ElA genes and the hK2 promoter and enhancer driving the ElB genes. The sequences of the PSE and hK2 promoter and enhancer are 80% identical, yet the virus is genetically stable. Again, CV764 has the high therapeutic index of the other viruses containing two prostate-specific TREs with a cell specificity of 10,000:1 for PSA+ cells compared to PSA-cells [22, 23]. Taken together, these adenovirus variants show that tropism of aden ovirus can be redirected by placing essential viral genes under the control of 10. Development of ARCAs for the Treatment of Prostate Cancer 2 9 5 tissue-specific regulators. The cell selectivity of a stable oncolytic virus can be over 10,000:1 w^hen the expression of more than one viral gene is driven by two different tissue-specific TREs. The phenomenal success of creating adenovirus host range mutants v^ith specificity for target cells compared to nontarget cells of over 10,000:1 is one of the major achievements of the ARCA technology. C. Antitumoral Efficacy of ARCA In vivo studies evaluating intratumoral and intravenous administration of prostate-specific adenoviruses w êre conducted in the nu/nu mouse containing human tumor xenografts. Tumors were produced by subcutaneous injection of PSA+-producing prostate cancer LNCaP cells into each flank of each mouse, and after establishment of palpable tumors (mean
tumor volume 300 mm^), the tumors were directly injected with purified virus or vehicle (PBS and 10% glycerol). Tumor growth was then followed for 6 weeks, at which time the mean tumor volume in each group was determined. A significant antitumoral activity was observed in the in vivo study for CV706. Tumor volume dropped by more than 80% in the animal group that was treated with CV706 by a single intratumoral injection. These residual tumor masses were shown by histology to be scar tissue devoid of PSA+ cells. After 6 weeks, 5 of 10 mice were visually free of tumor [21]. In contrast, DU145 is a prostate cancer cell line that is PSA~ and does not produce the androgen receptor. Tumors of DU145 cells were induced in nude mice and challenged with buffer, wild-type Ad5 but E3 virus CV702 and CV706. The results showed that CV702 inhibited growth of DU145 tumors, whereas CV706 has no effect on tumor growth. Thus, the prostate-specific CV706 virus not only shows efficacy but also selectivity for PSA"^cells in vivo [21]. The E3 region has long been considered unnecessary for replication of adenovirus in vitro. It has been universally deleted from Ad5 gene therapy constructs until recent efforts to prolong transgene expression from replication- defective Ad5 gene therapy constructs [39,40, 62-64]. To test the possibility of increasing virus cytotoxicity, we created CV787 from its parent virus CV739 by engineering the full-length E3 region back into the viral genome. Thus CV787 contains the rat probasin promoter driving the ElA gene the PSE driving the ElB gene. Otherwise, CV787 is identical to the recombinant wild-type adenovirus CV802. CV787 retained the high specificity of characteristics of two TRE-containing viruses driving the ElA and ElB genes. Cell viability assay and virus yield assay demonstrated that addition of E3 aids virus replication and increases virus cytotoxicity. Thus CV787 has a stronger cytotoxicity than CV739 [22]. The increased cytotoxicity due to the Ad5 E3 region was also confirmed in vivo in the LNCaP xenograft animal model. A single intratumoral injection 2 9 6 Henderson and Yu of CV739 and CV787 yielded identical reduction of LNCaP xenografts. However, CV739 required 100-fold more virus to achieve the same effect as CV787 [22]. A single intratumoral CV787 at a dose of 1 x 10^ particles/mm^ was curative for animals 6 weeks after treatment (^ = 8). A single intravenous injection of CV739 at a dose of 5 x 10^^ particles could stop tumor growth, whereas CV787 at this dose level caused a fourfold reduction in tumor volume [22]. Six weeks following a single intravenous injection of 1 x 10^^ particles, the sizes of tumors were reduced to less than 5% of their original size, and 8 of 14 mice were visually free of tumors. The residual tumors measurably present were immunohistologically devoid of PSA [22]. The serum PSA levels in mice injected intravenously with CV787 decreased to 5% of their starting values within 4 weeks. Intravenous administration of CV787 designed to treat LNCaP xenografts showed that 1 x 10^^ particles could eliminate 300 mm^ preexistent LNCaP xenografts, whereas 1 x 10^^ particles of CV706 administered intravenously only stabilizes tumors. A dose-response curve of 1 X 10^ and 1 x 10^^ CV787 particles administered as a single intravenous dose can stabilize and regress tumors, respectively, but not eliminate tumors. These data indicate that CV787 has a significantly improved antitumor activity and a single dose of intratumoral or intravenous administration can eliminate pre-existent tumors in animal models. D. Mechanism for Cell-Killing of ARCA Infection with adenovirus causes profound changes in host-cell macro- molecular synthesis that ultimately lead to cell death. Virion fiber protein inhibits macromolecular synthesis when applied directly to cells bearing the adenovirus receptor [65]; soluble penton protein causes cytopathic effects (CPEs) in susceptible cells that are similar to those caused by infectious virus [66]. Cell-specific DNA synthesis, export of cellular mRNAs from the nucleus to the cytoplasm, and cell-specific translation are all inhibited after infection, but the precise mechanisms are not completely understood. The 243E1A protein induces the full range of classical apoptotic events by increasing the level of the host cellular tumor suppressor protein p53. The 289E1A protein induces apoptosis by a p53-independent mechanism that requires a product of the E4 region [67, 6S]. The ElA-induced activation of the apoptosis path way (s) must be modulated by ElB proteins to ensure efficient virus replication prior to cell death [69]. Activation of the interferon-inducible RNase L pathway by the adenovirus-associated type I (VAI) RNA [70] may also contribute to the stimulation of apoptotic pathways in adenovirus-infected cells [71]. The E3 11.6-kDa adenovirus death protein also has a role in cell- killing and promotes the release of progeny virions from the cell [72, 73]. We have investigated how our oncolytic viruses kill tumors in the nuinu mouse model. Immunohistochemical analyses were performed to assay for the 10. Development of ARCAs for the Treatment of Prostate Cancer 2 9 7 de novo synthesis of CV787-encoded proteins in tumor xenografts and to examine the effects of treatment with CV787 on tumor morphology in vivo. The mice bearing human LNCaP tumors were injected intravenously on day 0 with 1 X 10^^ particles of CV787 per animal. Tumors were excised from two animals on days 1, 3, 7, 14, 21, and 28. The tumors were cut into six pieces and each piece fixed, embedded in paraffin, and sectioned. Sections were stained for the presence of adenovirus protein by a double-antibody protocol with rabbit anti-Ad antibodies and Fast Red stain followed by a hematoxylin counterstain. On day 1, intracellular staining for adenovirus protein was detected in less than 1% of the tumor cells examined in 12 sections from two tumors. Occasional small clusters of stained cells, as well as dispersed single stained cells, were visible. By day 3, large clusters of cells expressing adenovirus proteins were detected in one of the two excised tumors. In some instances, areas of tumor necrosis were adjacent to clusters of adenovirus protein positive cells. On day 7, intracellular staining for adenovirus proteins was detected in greater than 10% of the tumor cells examined in 12 sections from both excised tumors. Virus-infected cells within the tumor sections were prominent on Day 21 and increased to more than 90% of the microscopic field of the section by Day 28. These results demonstrated that CV787 replicated in and expressed virus-encoded gene products in the LNCaP xenografts. The increased distribution of virus protein-positive cells indicated that infectious progeny CV787 spread to adjacent cells within the tumor which was associated with progressive necrosis in vivo [22]. Adenovirus-induced apoptosis causes cell death in vitro., specifically at the late stage of infection [67]., and this process may contribute to the therapeutic effect of oncolytic virus in vivo. LNCaP xenografts in athymic mice were treated on day 0 with vehicle alone or a total dose of 3.2 x 10^ particles of CV706 per mm^ of tumor. Tumor biopsy specimens were taken on day 14, and 5-\|xm sections were prepared and examined for apoptosis. Extensive areas containing apoptotic nuclei were detected in sections of tumors treated with CV706. More than 25% of nuclei were apoptotic in some sections from CV706-treated tumors. In contrast, less than 2% of nuclei were apoptotic in sections from tumors treated with vehicle alone. Additionally, a visual change of tumor color has been documented when the animals receive tumor specific adenovirus variants. For example, the color of the established human LNCaP xenografts is black. However, the tumors will become white within 1 week after receiving a dose of CV706 or CV787 through either intratumoral or intravenous administration. Histological H&E staining analysis found that the virus-treated tumors had significantly fewer numbers of blood vessels when compared to the tumors treated with vehicle. It is unclear at this time as to the precise mechanism by which this reduction in blood vessel number is achieved. This can achieved either though direct 298 Henderson and Yu A. Vehicle g cwOB .:'.> V-\-.. / <. -, l?^..v rr. ::r': • - ^'^"'!® V.!" '^?^ '*°'" '"9 f"'' "^^'y developed blood vessel. Tumors were harvested 14 days after receiving (A) vehicle or (B) 1 x 10^ particles/mm3 tumor of CV706 and stained with anti-CD31 monoclonal antibody by immunohistochemical analysis. damage of endothelial cells or indirectly through the destruction of tumor vasculature by extensive necrosis. CD31 is expressed constituitively on the surface of adult and embryonic endothelial cells and has been used as a marker to detect angiogenesis [74]. Immunohistochemical staining was performed to examme the effect of virus treatment on tumor angiogenesis by using monoclonal antibody against CD31. Tumors treated with CV706 showed a sigmficantly lower level of CD31-positive cells in the vessel when compared to vehicle treated tumors (Fig. 1). This observation suggests that CV706 may be mhibitmg tumor angiogenesis to a significant extent. The change of tumor color from black to white has been a reliable early indication for antitumor ethcacy of a virus. III. Synergy of ARCA and Conventional Therapy Although conventional cancer therapies (surgery, chemotherapy, and radiation) are effective at curing early-stage disease, few human cancers are curable with a single modality. The utilization of a replication-competent cytolytic adenovirus as a therapeutic modality shows much promise. In our laboratory and clinical trials, CV706 and CV787 have been shown to be effective agamst human prostate tumors [21-23]. However, single-dose total eradication of human LNCaP xenografts in our animal model required as much as one-third of the lethal dose of virus alone [22, 75]. To set the bar for efficacy high, namely single-dose efficacy, is deliberate. While multiple doses 10. Development of ARCAs for the Treatment of Prostate Cancer 2 9 9 always allowed total eradication of tumors with lower doses, in the clinic the patient immune response will tend to limit the number of doses that can be effectively delivered despite the SIAPA technology described below. Certainly, strategies to improve the efficacy of these viruses are desirable. However, in preclinical experimental animal models, it can be readily appreciated that genetic engineering allows creation of many different viruses with an associated rationale for why the new mutant may be an improvement with greater efficacy over its predecessor. Experimentally it is impossible to compare different viruses for their efficacy in animal models unless a standardized treatment regimen is used. Thus, we use only single-dose treatment regimens in our animal studies. Small changes in efficacy do not warrant replacing current clinical candidates; large changes in efficacy are most interesting. The largest changes in efficacy we have seen came with the addition of the E3 region and the synergy afforded by radiation and chemotherapy. We studied neoadjuvant therapy consisting of combining ARC A with conventional therapy for additional reasons as well: (1) combinations of agents with different toxicological profiles can result in increased efficacy without increasing overall toxicity to the host; (2) overlapping resistance between ARCA and conventional therapeutics has not been described previously and a combination of agents may thwart the development of drug resistance; and (3) tumor cell populations have different drug sensitivity profiles, which allow the physician to take advantage of possible synergies between drugs, resulting in increased anticancer efficacy in patients. A. Synergy of CV706 and Irradiation Radiation therapy, either administered through external beam or seed implantation (brachytherapy), is one of the most widely used therapeutic modalities for treatment of clinically localized prostate cancer [76, 77]. How ever, nearly 30% of patients treated with potentially curative radiation doses suffer relapse as defined as three serial rises in serum PSA levels (biochemical failure). Specifically, the eradication of locally advanced or high-risk prostate cancer with radiation has proven more difficult than believed previously. The gains with high radiation dose have been modest and fraught with signifi cant side-effects [78]. It is apparent that there is a need for novel methods of radiosensitization. We have examined the effects of a combination treat ment involving CV706 and irradiation for localized prostate cancer. We have demonstrated a very high synergy when irradiation treatment is complemented with CV706. Initial in vitro experiments demonstrated a synergistic cytotoxicity to LNCaP prostate cancer cells when CV706 was combined with radiation. LNCaP cells treated with CV706 and radiation had significantly decreased viability compared to cells treated with either agent alone. LNCaP cells 300 Henderson and Yu treated with radiation exhibited
a significantly greater burst size of CV706 (m.o.i. = 0.1) with more than 100 times more virus produced per cell. In addition, the combination treatment of CV706 with radiation did not alter the specificity of replication-mediated cytotoxicity. A similar synergistic anti-tumor efficacy of combination therapy with CV706 and radiation was also observed in vivo in prostate LNCaP xenografts. Previous in vivo studies demonstrated that prostate tumors were completely eliminated within 6 weeks by a single intratumoral administration of CV706 at a dose 5 x 10^ particles per mm^. Combination treatment of CV706 at a dose of 1 X 10'̂ particles per mm^ with radiation 10 Gy eliminated tumors within 6 weeks. Under similar conditions CV706 or radiation alone could only slow down or inhibit further tumor growth (Fig. 2). Thus, a 50-fold lower dose of virus (1 x 10^ particles) can be used in the combination treatment modality to achieve the same curative effect. Statistical analysis of the in vivo studies indicated that CV706 and the radiation combination group showed a significant anti-tumor synergy with a 1.5- to 6.7-fold higher inhibition of tumor Weeks ""•-Vehicle -"CV708 m) and EBRT (d1) - ^ C V 7 0 6 (1x10'̂ p/mm3) (dO) XBRT(10Gy)(d1) -Q~ E8RT (dO) and CV706 (d1) Figure 2 /n v/vo efficacy of intratumorally administered CV706 and radiation against nude mouse LNCaP xenografts. Tumor volume of LNCaP xenografts treated witfi either vehicle (diamond), CV706 alone (1 x 10'̂ particles per mm^ of tumor, circle), radiation alone (10 Gy, triangle), CV706 (1 X 10'̂ particles per mm^ of tumor) at day 0 plus radiation (10 Gy) at day 1 (close square), and radiation at day 0 plus CV706 at day 1 (open square). 10. Development of ARCAs for the Treatment of Prostate Cancer 3 0 1 growth over the additive effect during the entire treatment period. Subsequent studies have show^n that a synergistic antitumor efficacy could be achieved w^hen 1 X lO"" particle CV706 w âs combined with only 5 Gy radiation (unpublished data, Chen etal.). Further in vivo studies are in progress to determine the effective minimum dose of CV706 in combination with radiation required for complete regression of tumors. A series of experiments was then designed to examine the effects of the sequencing of the agents, the timing of radiation following virus admin istration, and radiation fractionation. Efficacy was highly dependent on the sequencing of the agents; treatment with CV706 24 h prior to radiation was significantly superior to radiation followed by CV706. The antitumor activity was decreased when the tumors were treated with radiation 7 days after CV706 administration. However, there was no significant difference in antitumor effi cacy when the tumors were treated with CV706 followed by a single dose of radiation or four sequential daily fractional doses of radiation as long as the total does of radiation was the same. A preliminary assessment of synergistic activity in the CV706 and radi ation combination treatment reveals several mechanistic possibilities. First, radiation at the synergistic dose significantly increases virus replication. One- step growth curve study shows that although synergistic doses of irradiation did not alter virus replication kinetics, the irradiation significantly increases the burst size of CV706 in LNCaP cells. For example, burst size of CV706 in LNCaP cells treated with CV706 for 24 h followed by irradiation is 500-fold higher than that in cells treated with CV706 alone (0.01 m.o.i.). Irradiation kills mammalian cells in the reproductive (also known as clonogenic) death pathway, and breaks down DNA strands. Most radiation induced DNA double-stranded breaks are rapidly repaired by constituitively expressed DNA repair mechanisms [79]. DNA repair machinery becomes more active in irradi ated cells, thus allowing for greater replication/multiplication of the episomal adenoviral DNA. Because of its small target size, the adenoviral genome (36 kb) is far less likely to sustain radiation-induced damage as it is 10^-fold smaller than that of human cells (3 x 10^ kb). Therefore, the more active cellular DNA synthesis machinery in the irradiated cells may facilitate viral DNA synthesis and virus replication. Second, CV706 may be augmenting the anti-tumor activity of radiation. The adenovirus El A gene is the only viral gene expressed during the first 2.5 h of infection and encodes a multifunctional tran scriptional factor also known to induce apoptosis [32, 80]. It is believed that the adenovirus ElA gene is a potent inducer of radiosensitivity through p53- dependent and -independent mechanisms. Malignant tumors, when expressing adenovirus ElA, are very sensitive to treatment with DNA-damaging agents in vivo., including irradiation [81, 82]. In the tumors treated with CV706 and radiation, the histological features included intravascular thrombosis and massive necrosis positioned more centrally within tumors. This supports a 3 0 2 Henderson and Yu growing concept for the involvement of an effect on the destruction of the vasculature leading to tumor reduction and elimination. This seems to be in agreement w îth a recent finding that inhibition of angiogenesis led to increased tumor radiosensitivity [83]. Indeed, CV706 in combination v^ith radiation exerted a significant reduction of CD31 positive blood vessels, indicating an anti-angiogenesis effect of the combination treatment [84, 85]. As human tumors are composed of a mixture of cells having different genetic makeup, this inherent intratumoral heterogeneity may need treatment w îth multiple therapeutic modalities. As demonstrated here, radiation can be successfully combined with cytolytic adenoviral therapy. The advantages are as follows: the combination of CV706 with irradiation still limits the damaging effects of CV706 to the irradiated cells; the combination of CV706 with irradiation leads to significantly increased necrosis, marked decrease of blood vessel number, and inhibition of angiogenesis; The combination of CV706 with irradiation allows at least a 50-fold reduction in the amount of virus to achieve the same curative effect; animals receiving combination treatment appear healthier, characterized by a significant weight gain compared to the groups treated with either agent alone. Thus, combining cytolytic adenoviral therapy with radiation may augment the efficacy of standard cancer modalities. B. Synergy of CV787 and Chemotherapy Once prostate cancer enters a metastatic stage, the current treatment is androgen ablation therapy, which in 70% of men provides relief from otherwise uncontrollable bone pain and increases life expectancy by 6-18 months [86, 87]. For many years, chemotherapy was felt not to play any role in the treatment of advanced prostate cancer. However, this negative impression is apparently changing, as significant activity is being seen with new drugs and drug combinations [87]. Interim analysis of an ongoing Phase III trial of docetaxel in end-stage hormone refractory prostate cancer patients shows an improvement in life span to 30 months (Dan Petrylak, pers. commun.). A synergistic antiproliferative effect in LNCaP prostate cancer cells was observed when CV787 was combined with various chemotherapeutic agents including cisplatin (Platinol), docetaxel (Taxotere), doxorubicin (Adri- amycin), estramustine (Emcyt capsules), etoposide (VePesid), gemcitabine HCl (Gemzar), mitoxantrone (Novantrone), and paclitaxel (Taxol). In this chapter, we will discuss the synergistic antitumor efficacy of CV787 only when com bined with the taxanes. In vitro experiments demonstrated that LNCaP cells cultured with paclitaxel or docetaxel for 24 h before or after infection with CV787 had significantly lower rates of proliferation than cells treated with either agent alone [88]. Also, LNCaP cells exhibited a greater burst size of CV787, whereas no significant effect on viral growth kinetics was seen. No significant difference 10. Development of ARCAs for the Treatment of Prostate Cancer 3 0 3 in the effectiveness of the combined therapy of taxane and CV787 infection was observed by varying the time of taxane administration in vitro^ v^hereas varying the administration schedule of pacHtaxel w îth Ad-p53 gene therapy can modulate the synergistic activity between these two agents in ovarian can cer [89]. Furthermore, in vitro combination treatment of CV787 with taxane did not alter the specificity of replication-mediated cytotoxicity. CV787 has been shown previously to replicate preferentially in PSA-producing human prostate cancer cells 10,000 times more efficiently than in non-PSA-producing cells [22]. In the presence of paclitaxel or docetaxel in the culture medium, CV787 replicates to the same degree of specificity in the non-PSA-producing human cell lines versus PSA+ cells. Cell viability assays with MTT further indicate that CV787 in combination with taxane remains fully selective. The antitumor effects of combination therapy with CV787 and taxane were also evaluated in vivo. Previous studies have demonstrated that tumors were eliminated within 6 weeks by a single intravenous administration of CV787 at a dose of 1 x 10^^ particles [22]. Clinically, docetaxel is adminis tered intravenously at 12.5 mg/kg once per week for 6 weeks. A single cycle of docetaxel is not toxic, whereas the full course of six cycles is highly toxic. Combination treatment of CV787 and taxane (three courses of 5 mg/kg) elim inated tumors within 4 weeks, whereas CV787 alone at the same dose of 1 X 10^^ particles per animal could only slow down tumor growth (Fig. 3A). The experiments described below are with a single cycle of docetaxel. In combination with the higher clinical dose of 12.5 mg/kg docetaxel, 1 x 10^ particles and of 1 x 10^ particles of CV787 led to a complete elimination of tumors within 4 weeks. Thus, the dose of CV787 required for complete remission has been reduced by 1000 times from 1 x 10^^ particles to 1 x 10^. The LDo and LDioo single bolus dose of CV787 for Balb/c, nuinu is 1.0 x 10^^ particles and 2.5 x 10^^ particles, respectively. Thus, the single-dose combina tion of docetaxel with CV787 has increased the potential curative therapeutic window, from 1 to 1000. Furthermore, healthier animals, characterized by body weight, were observed in the combination treatment group as compared to groups treated with either agent alone (data not shown). The mechanism(s) of synergistic activity in the combination of taxane with CV787 is unknown at this time; however, our experiments suggest a few hypotheses. First, taxane at the synergistic dose may be augmenting viral replication. It has been previously shown that a low concentration of paclitaxel (1-14 nM) increased the number of cells transduced by recombinant adenovirus 3 -35% in a dose-dependent manner [89]. Recently, it was found that taxane increases intracellular receptor trafficking. For example, more Coxsackie and adenovirus receptor (CAR) moved to outer membrane in the paclitaxel-treated cells (Nielsen Loretta, pers. commun.). It is suggested that adenovirus transduction efficiency may increase after the cells are treated with taxane. Indeed, our data show that although the synergistic dose of paclitaxel 304 Henderson a n d Y u A 1400 1200 0) 1000 ....^^ _. 1 E "oo BOO > O E 600 ZJ L-'T'̂ î ,̂ . ^ \ 400 i*f<^' i { < ; 200 ^ - ^ » _ , 1 2 3 4 6 6 7 Weeks After Treatment " Control A CV787 (1 x1 O^Op) 'CV787(1x10^^p) - Docetaxel (5 mg/kg) - CV787/Docetaxe[ B Docetaxel CV787 CV787/Docetaxel Mm Figure 3 In vivo efficacy of intravenously administered CV787 and taxane against nude mouse LNCoP xenografts. (A) LNCaP tumor volume in mice treated with either vehicle, CV787 (1 x 10 ' ^ 1 X 10^^ particles per animal, iv), docetaxel (5 mg/kg at day 2, 5 and 8, iv), or CV787 (1 x 10^^ particles per animal) and docetaxel (5 mg/kg at days 2, 5, and 8, iv), (n = 6). Tumor volumes measured weekly. Error bars represent the standard error of the mean. (B) Histological features (H&E staining) of LNCaP tumors treated with either docetaxel (12.5 mg/kg) (left), CV787(1 x 10^0 particles per animal) (middle), and CV787 (1 x 10^^ particles per animal) plus docetaxel (12.5 mg/kg) (right) at X 400 original magnification. LNCaP xenografts were harvested 14 days after treatment. The open arrow indicates cells with abnormal mitosis, the filled arrow indicates necrotic cells. 10. Development of ARC As for the Treatment of Prostate Cancer 3 0 5 or docetaxel did not alter virus replication kinetics, the chemotherapy drugs slightly increased the burst size of CV787 in LNCaP cells. Second, CV787 may be augmenting the anti-tumor activity of taxane. El A gene expression has been show^n to increase cellular sensitivity to chemotherapeutic agents [90] and this enhanced sensitivity is partially caused by the induction of p53-dependent apoptosis by the ElA-induced sensitization of the cells [91]. Recently, Ueno et al. found that human ovarian cancer cells that w êre originally resistant to
paclitaxel, became paclitaxel sensitive in El A dow^nregulated HER-2/neu cells \GG\. In CV787, the ElA gene is intact and may be overexpressed in PSA- producing LNCaP cells, w^hich may enable significant interaction betv\̂ een ElA and taxane. In addition, ElA is a potent inducer of p53 protein expression in infected cells [92]. p53 levels may increase follov^ing infection, thereby increasing cell sensitivity to chemotherapy-induced apoptosis. This is consistent w îth the observation that more apoptotic cells v\̂ ere seen in the LNCaP tumors that received combination treatment than in tumors that received either agent alone (Fig. 3B). Interestingly, in the combination-treated tumors, most of the cells v^ere empty and virtually devoid of cellular content. Finally, the actions of the tw ô agents may be occurring at tvŝ o distinct points in the same apoptotic pathv^ay, analogous to the activity of cyclosporine A and rapamycin on tw ô distinct points in the T-cell activation signal transcription pathw^ay [93]. Further investigation of the possible mechanism(s) of synergistic activity in the combination of taxane w îth CV787 is under w ây. IV. Toxicity of Intravenously Administered ARCAs in the Absence or Presence of Docetaxel As a prerequisite for a Phase I trial, vŝ e evaluated the biodistribution, persistence, and toxicity of CV787 follow^ing intravenous administration. Cotton rats v^ere chosen over the mouse for Cotton rats are semipermissive for adenoviral replication [32, 94]. The objective of the first study was to determine the toxicity of CV787, after a single tail vein intravenous administration into the male Cotton rat, followed by a 3-day observation period. CV787 was administered at 3 x 10^^ particles, 1 x 10^^ particles, and 3 x 10^^ particles per animal. There were no deaths or treatment-related clinical observations over the study period. In addition, there were no effects of treatment noted on body weight, food consumption, or clinical pathology or at the macroscopic pathological examination at all dose levels. As 3 x 10^^ particles per animal was the highest achievable dose, this dose was considered the no-effect level when administered as a single bolus intravenous injection to the Cotton rat. To investigate the biodistribution and persistence of CV787 following intravenous administration, we conducted a 21-day study with Cotton rats. 3 0 6 Henderson and Yu Twenty-seven male Cotton rats received either vehicle, replication-defective Ad5 CV730, an ElA deleted adenovirus mutant, and CV787 at 3 x 10^^ particles per animal in a single intravenous tail vein injection on day 1. Three animals from each group were sacrificed on days 3, 8, and 22. DNA was extracted from plasma, testes, brain, spleen, liver, heart, lungs, kidneys, and bone marrow. The DNA obtained from tissues was quantified by optical density measurement and 100 ng was subjected to quantitative PCR using a TaqMan procedure specific for the Ad5 packaging site present in both CV730 and CV787. There were no ill effects seen upon intravenous delivery of 3 x 10^^ particles of CV787 per Cotton rat. As similar distribution of adenovirus seen in other rodents [94], CV787 accumulated primarily in the liver and spleen, and to a lesser extent in the heart, kidney, lung, and bone marrow. Viral DNA levels were very low in the testes and undetectable in the brain. The tissue distribution was similar for Ad5 CV730, a replication-defective control virus that was detected primarily in the spleen, liver, and kidney. The potential for nonspecific replication of CV787 was found to be very low, resulting in no more than a two- to fivefold increase in viral copy number during the course of the study. A third study was designed to compare the acute toxicity and inflam matory potential of CV787 and CV706 by subcutaneous and intradermal administration in immunocompetent mice, C57BL/6. Four groups of five mice each were treated with a single dose of 1 x 10^^ particles of CV787 or CV706 by subcutaneous (sc) or intradermal (id) injection on day 1. Four control groups were treated either intradermally or subcutaneously with the vehicle for CV787 or the vehicle for CV706. Two additional groups were treated intravenously (iv) with either virus at the same dose to compare the acute liver toxicity between CV787 and CV706. All animals were scarified on day 4. Blood was collected for serum chemistry analysis. Macroscopic examination for signs of toxicity was performed at necropsy. Portions of the liver, lung, spleen draining lymph nodes, and sc or id injection sites were collected in buffered formalin and examined for histopathological effects. All animals survived until the scheduled day of sacrifice. No signs of toxicity were observed after treatment or during macroscopic examination in any of the id or sc groups. Serum chemistry analysis showed no significant changes following subcutaneous or intradermal administration of either virus preparation. With both viruses, histopathological analysis revealed a mild chronic inflammation in the areas surrounding the subcutaneous injection site in some of the animals, but no effects on any of the body organs were observed. Some inflammation was observed in the dermas following intradermal administration with both viruses, but the inflammation remained localized to the injection sites and the overall effects were no more severe than those observed after subcutaneous treatment. As expected, intravenous 10. Development of ARCAs for the Treatment of Prostate Cancer 3 0 7 delivery of the virus at a dose of 1 x 10^^ particles per animal resulted in significant liver toxicity, as evidenced by 5- to 100-fold increase of ALT and AST enzyme levels. The ALT and AST response w âs more pronounced for CV706 than for CV787. Intravenous treatment w îth CV706 also resulted in statistically significant effects on BUN, serum cholesterol, glucose, and calcium levels, w^hich v^ere not as apparent follov\^ing the intravenous administration of CV787. The subcutaneous and intradermal effects observed in this study suggest that CV787 poses no additional risk of inflammation over CV706. Results from intravenous treatment of C57BL/6 mice indicate that CV787 may shov^ less systemic toxicity than CV706. Also, it does not appear that Cotton rats add any nev^ information on toxicity compared to the use of more common and easier to use laboratory mouse strains such as C57BL/6. A synergistic antitumor efficacy w âs observed w^hen CV787 w âs combined w îth taxane in the prostate tumor xenograft [88]. To examine the toxicity of CV787 in combination w îth the chemotherapeutic agent docetaxel, we conducted a fourth toxicology study in C57BL/6. The 4-day and 28-day effects w êre evaluated in mice given a single daily dose of dexamethasone (Decadron) each day for 3 consecutive days, a single injection of CV787 at lovv̂ (1 x 10^ particles per animal), medium (3 x 10^ particle per animal), or high (1 x 10^^ particles per animal) dose, and a single dose of the chemotherapeutic agent, docetaxel (100 mg/m^). Six different treatment groups v^ere established using identical treatment regimens for animals in subgroups sacrificed at 4 days and at 28 days. Each treatment subgroup consisted of eight randomly assigned mice. The ARCA virus vehicle and all three doses of virus were given intravenously via tail vein injection. All virus injections v^ere given on day 1 at 1-2 h prior to Decadron administration. Dexamethasone and docetaxel vŝ ere both given separately by intraperitoneal injection. At the end of the 4-day and 2 8-day treatment periods, respectively, blood samples w êre taken for complete blood count (CBC), including platelet count, and serum chemistry analysis just prior to sacrifice. Tissues w êre taken for organ v^eights and histopathologic evaluation of liver, lung, spleen, kidney, brain, heart, mesenteric lymph nodes, bone marrov^, any enlarged lymph nodes, and any gross lesions. Acute virus-associated changes at day 4 v\̂ ere characterized by hepato cellular necrosis and elevated serum leakage (liver) enzymes (ALT, AST) seen in the groups w ĥo received a high dose of CV787 alone or high dose of CV787 in combination with docetaxel, as well as an increasing incidence of lymphoid hyperplasia from mid-dose (3 x 10^ particles per animal) to high- dose (1 X 10^^ particles per animal) virus subgroups. These findings indicate dose-related viral effects without evidence of augmentation by docetaxel in combination with dexamethasone. In contrast, the severity of the lymphoid hyperplasia in the high-dose virus subgroup may have been ameliorated some by the docetaxel combination. Evidence of acute (high-dose) virus-associated 3 0 8 Henderson and Yu hepatocellular toxicity was resolved by day 28 and replaced by a viral dose- dependent increase in mild multifocal hepatic mononuclear cell inflammatory infiltrates in subgroups w ĥo received CV787 at a dose of 3 x 10^ particles per animal or 1 x 10^^ particles per animal with or without docetaxel. These inflammatory changes are most likely to be associated with chronic effects of viral replication and antigen stimulation and do not appear to be ameliorated by docetaxel combination. There was acute mild/moderate bone marrow necrosis at day 4 from high-dose virus plus docetaxel combination in two of eight mice, but not from high-dose virus alone or from lesser-dose virus plus docetaxel combination. There were no bone marrow lesions seen in any animals at day 28, so if the effect is real, it appears to be a reversible effect of the high-dose virus plus docetaxel combination. Acute mild to moderate bone marrow hypoplasia was found at day 4 in all eight mice only in the high-dose virus plus docetaxel combination. Hypoplasia was not seen at day 28. The occurrence of bone marrow necrosis (2/8) and hypoplasia (8/8) in the day 4 high-dose virus plus docetaxel combination subgroup constitutes evidence for an additive but reversible effect of the combination. The 25% bone marrow necrosis and 100% hypoplasia were the only additive toxic effects ascribed to the combination of docetaxel with virus. These toxicities appeared to be dose-dependent with respect to the virus and are reversible. V. Effects of Preexisting Adenovirus Antibody on Antitumor Activity and Immunoopheresis for Human Therapy Preexistent antibodies to adenoviruses are highly prevalent in the human population. By 7 months of age 23% of infants have circulating antibody titers to adenovirus [95], a percentage that increases with age to essentially 100%[96,97]. Adenovirus antibodies are readily divided into total antibodies (TAb) and neutralizing antibodies (NAb). Almost all adults have antibodies (TAb) to adenovirus type 5, but only Sl^/o of adults have neutrahzing anti bodies \1S^ 98]. These antibodies are of significant interest in the design of therapeutic strategies based on adenoviruses. Transient immunosuppression during initial adenovirus administration with cyclosporine [7, 99], FK506 [100-103], cyclophosphamide [104, 105], deoxyspergualin [105], IL-12 [106], CTLA4Ig [107, 108], anti-CD4 antibody [109-111], anti-CD40 ligand antibody [112, 113], and dexamethasone [113, 114] has been used to blunt the formation of anti-Ad Abs in animal mod els. This strategy has successfully allowed repeat adenovirus treatments in animals and should extrapolate extremely well to the clinic. Clinically, this strategy could prevent a rise in antibody titer, thereby maintaining preexistent 10. Development of ARCAs for the Treatment of Prostate Cancer 3 0 9 anti-Ad5 Ab titers at pretreatment levels. However, treatment in these ani mal models was initiated in immunologically naive animals, animals without preexistent circulating anti-Ad antibody [115-117]. This does not represent the situation in the clinic where many patients already have preexistent anti-Ad antibodies [118-120]. Transient immunosuppression would not be expected to suppress the levels of these preexistent circulating antibodies. Three approaches have been explored to overcome preexistent adenovirus antibody: inducing immunologic tolerance [120, 121], the use of another of the 51+ different adenovirus serotypes [122-124], or the use of polyethylene glycol (PEG) to block adenovirus virions from antibody binding [125-127]. All three approaches show promise, but all may also be encumbered by practical considerations. The induction of tolerance may be time-consuming, sporadic, and unpredictable. Changing serotypes changes the drug, implying separate costly drug approval pathways. The conjugation of PEG to purified virions to successfully block antibody interactions incurs questions of also blocking the virus from the interaction with the CAR [128,129] as well as virus pen etration, internalization, and uncoating [127, 129]. In addition, large-scale purification of conjugates from nonreacted components during manufacturing will be required [130]. Surprisingly, preexistent adenovirus antibody has not impaired the ther apeutic activity of adenoviruses directly injected into tissues in animal studies or in human clinical trials [118, 131-134]. Presumably, the therapeutic dose of viruses directly injected into tissues overwhelms the level of free aden ovirus NAb available. We have been
interested in systemic administration of adenovirus for the treatment of metastatic prostate cancer. Such treatment requires intravascular injection of adenovirus to achieve virion distribution to the multiple distant tumor sites. In the vascular compartment, adenovirus will be subject to the opsonizing activity of TAb and the neutralizing activity of NAb; both could adversely affect efficacy [115, 131]. There is considerable evidence that repeat administrations of adenovirus-based therapeutics in the face of mounting specific anti-adenovirus immune responses eliminates efficacy [98,112,115,135]. A. Preexisting Adenovirus Antibodies Inhibit Systemic Toxicity and Antitumor Activity In order to better understand the effect of preexisting antibody on ARCA- mediated toxicity and anti-tumor efficacy, we first developed an animal model controlling both the dose of the intravascularly administered adenovirus and the level of preexistent antibody while measuring the virus-induced toxicity and the anti-tumor activity. To this end we prepared hyperimmune sera to purified adenovirus in rabbits, purified the rabbit total IgG containing high titer adenovirus antibody, and passively injected the purified rabbit anti-Ad5 3 1 0 Henderson and Yu antibody into Balb/C nu/nu mice bearing LNCaP prostate cancer xenografts. This established a measurable humoral anti-Ad5 antibody titer. The prostate specific adenovirus, CV706, was injected into the tail vein (iv) of animals, and toxicity and anti-tumor activity was measured. First, preexisting neutralizing antibodies were found to be protective of nude mice from liver toxicity. Adenoviruses injected intravenously into mice have been shown to cause liver toxicity from a rapid induction of chemokines and neutrophils that lead to hepatic necrosis and apoptosis [44, 86,136-138]. In the absence of preexistent antibody, the lethal dose (LDioo) was 2.5 x 10^^ CV706 particles, whereas 1 X 10^^ CV706 particles was not lethal. Yet, in the presence of a 1:80 preexistent titer of Ad5 NAb, iv injection of 5 x 10^^ CV706 particles was no longer lethal. Serum ALT levels started to elevate 48 h after viral injection in the animals that had no or low titers of preexistent antibodies (1:10 or 1:20), and increased progressively, reaching peak levels within 72-96 h. However, the animals injected with the same lethal dose of CV706 but containing a 1:80 level of preexistent anti-Ad5 NAb exhibited normal or only slightly elevated ALT levels during the entire treatment period (Table II). In this study, a maximum tolerated dose of CV706 with a 1:80 pre-existent NAb titer was not found. Secondly, no effect of preexistent antibodies on the uptake of CV706 in different organs was found in the animals after receiving virus [76]. Balb/C nu/nu mice, pretreated with purified IgG (NAb titer 1:80) or buffer (no neutralizing antibodies), were administered 2.5 x 10^^ particles of CV706 by a single tail vein injection. As expected from prior observations [139], the majority of Ad DNA was accumulated in liver and a small percentage of DNA was found in spleen, whereas Ad DNA was undetectable in blood, lung, heart, or kidney. The level of Ad genome in liver was equivalent with or without preexistent anti-Ad5 Ab at 30 min and at 2 h after CV706 injection. However, at 24 h postinjection. Ad DNA was undetectable in the liver of mice with preexistent antibodies, whereas a significant amount of Ad DNA was still present in the liver of animals devoid of preexistent Ad5 Ab. No Ad DNA was detectable in liver for both treatment groups 72 h after virus administration. This observation suggests that the reduced amount of Ad DNA in liver of mice with preexistent antibodies at 24 h is not due to virus redistribution, but may be related to accelerated virus DNA degradation in the liver. Most likely, virus complexed with antibody remains in the endosomes and is subjected to lysosomal degradation processes, an outcome that may not trigger the adenovirus toxicity response. Third, preexistent anti-Ad5 Ab was found to be critical to the outcome of CV706 anti-tumor efficacy. In our previous work, we have shown the curative antitumoral activity of a single intratumoral injection of CV706 [21]. In the same animal model following a single intravenous tail vein injection we showed CV706 could inhibit additional tumor growth. With our second generation virus, CV787, a single intravenous administration could eliminated distant S S ^ ^ t \ <N o ro ^ ro O O < g O o < ^ ^ o E Ji ^ 'c < = s C < ^ c ^ ^ O) ^ B .SP -^ .5P -^ .5? c IZ) > C/5 • > C^ > C/5 S c/5 > (/3 > <U . tU ^ <U ^ > ^ ^ ^ J:̂ ^ ^ ^ 5 c^ O c^ O 03 O - ^ 3 ^ ^ ^ K 0) Z TJ '15 "cS ^o^ < B 'B _> 0) '> "oT c < C/5 i ^ «A *>< bJD 0) _C £ (/) JU 0 . • > < < a -I o o X X l O l o 311 3 1 2 Henderson and Yu preexistent tumors [22]. These reports showed that the nonspecific immunity, most notably the uptake of adenovirus by hver Kupffer cells [44, 139], does not eliminate the anti-tumor activity follow^ing intravascular administration of these viruses. Our in vivo studies demonstrated that 1 x 10^^ particles of CV706 inhibited additional grow^th of LNCaP mouse tumor xenografts in the absence of preexistent anti-Ad5 NAb. This anti-tumor activity w âs also show^n using the same dose but in the presence of 1:20 preexistent anti-Ad5 Ab. However, in the presence of preexistent antibodies (1:80), tumors in animals that were iv injected with 1 x 10^^particles CV706 grew progressively, a growth similar to the growth found in the group that received 1:80 antibody but no CV706. These results indicate that at a low preexistent anti-Ad5 antibody titer of 1:20, 1 x 10^^ CV706 particles can overwhelm the available antibody and retain anti-tumor activity. However, higher preexistent anti-Ad5 Ab (1:80) can eliminate the anti-tumor activity of intravenously administered CV706. The anti-tumor activity of CV706 can be restored in the face of a preexistent 1:80 antibody titer by a threefold higher dose of CV706 (3 x 10^^ particles/mouse). The amount of infectious CV706 in the circulation of mice showed an antibody titer-dependent pattern. In animals that had no preexistent antibodies or had low levels (1:10 or 1:20) of preexistent anti-Ad5 NAb and had received 1 x 10^^ infectious CV706 particles, infectious virus was detectable in the circulation 4 h after administration of CV706. In contrast, no infectious CV706 was found in mice that received a higher level of neutralizing antibodies (1:80) but the same dose of CV706. Interestingly, infectious CV706 was once again found in mice with a preexistent antibody (1:80) following a threefold higher dose of CV706 (3 x 10^^ particles). These results suggest that both the level of circulating neutralizing antibodies and the dose of administered viruses play significant roles in determining therapeutic toxicity and efficacy of intravenously administered CV706. Clearly, preexistent humoral antibody to Ad5 could present a serious barrier to the clinical application of systemically administered adenovirus. However, our data also indicate that therapeutic efficacy could be achieved by administering a higher dose of viruses, or by decreasing the levels of circulating neutralizing antibodies. The December 1999 RAC meeting, convened after the death in Philadel phia of Jesse Gelsinger, explored the then available data with adenovirus administered intravenously. It was apparent from that meeting that while investigators at the University of Pennsylvania, Onyx, Aventis, and Schering- Plough were monitoring for the existence of preexistent antibody to adenovirus, none were using preexistent antibody levels as a patient inclusion/exclusion criteria. The data presented above caused us to adopt a strategy we call SIAPA, for screening and immunoapheresis of preexistent antibody (described below). Unfortunately, the data also showed that not prescreening patients for preex istent antibody levels and proceeding with a dose escalation Phase I clinical 10. Development of ARCAs for the Treatment of Prostate Cancer 3 1 3 trial could lead to an uncontrolled clinical trial. Namely, if a patient with a high anti-Ad5 antibody received a given dose of adenovirus administered intravenously shov^ed no efficacy and no toxicity, the clinical answ^er is to proceed to the next higher dose. Hov^ever, if the next patient had a lov^ or undetectable antibody titer, the next dose of virus could be lethal. B. SIAPA: Screening and Immunoapheresis of Preexistent Antibody for Monitoring and Removing Preexistent Ad5 Antibodies from Blood Based on the observation that toxicity and anti-tumor efficacy of intravas cular CV706 treatment is inversely proportional to the titer of circulating anti-Ad5 antibodies, v̂ e became interested in the clinical use of an affinity column consisting of recombinant capsid proteins to specifically remove anti- adenovirus antibodies from human sera. Such an affinity column could be used in conjunction v^ith apheresis to become immunoapheresis. During the apheresis procedure, the serum could pass through a column of Ad5 antigen, specifically removing anti-adenovirus antibodies. Clinically, patients could be screened for the titer of preexistent anti-Ad5 Ab w îth the rapid 5̂ disposable immunoassay. Those patients v^ith a preexistent antibody titer above a given cutoff value would be candidates for adenovirus antibody immunoapheresis. After several hours of immunoapheresis, the antibody screening test could be repeated, confirming the removal of anti-Ad5 antibodies. Since anti-Ad5 antibodies would be expected to repopulate the vascular compartment at the rate of 5% per hour, a temporal window of several hours for intravascular adenovirus therapy could be created [140, 141]. The relevant preexistent anti-Ad5 antibodies are directed against the three principal capsid protein components: hexon, penton, and fiber [142, 143]. The Ad5 fiber, hexon and penton genes were cloned into the bacterial expression vector pQE-30. The recombinant proteins were expressed in E. coli M15 cells with IPTG induction and visualized as 120 IcDa (hexon), 82 kDa (penton), and 62 kDa (fiber) polypeptide proteins by denaturing SDS-PAGE and Western blotting. These results agreed well with sizes of hexon, fiber, and penton proteins derived from the purified virions. To test the feasibility of removing anti-Ad antibodies from serum with the recombinant Ad5 antigens, we built three types of column by individually coupling the purified hexon, penton, or fiber proteins to the Ni-NTAmatrix. A human clinical serum sample collected during the CV706 clinical trial at Johns Hopkins Oncology Center was used to test the ability of the columns to remove anti-Ad5 antibody. The patient serum collected 15 days after administration of CV706 had an anti- Ad5 NAb titer of 1:3200. The patient serum was passed through each column separately or sequentially. The processed serum was tested for the efficiency and specificity of removing anti-Ad5 antibodies and found that total anti-Ad5 3 1 4 Henderson and Yu Table III Efficiency of Ad Antigen Column to Remove Ad Antibodies from Plasma Patient NAb tier TAb (ixgy' ml) sera Before After Before After Serum-1 1:400 <1:10 15.6 0 Serum-2 1:1600 <1:10 75.6 3.7 Serum-3 1:400 <1:10 158.5 9.6 Serum-4 1:1600 <1:10 498.4 8.7 Serum-5 1:400 <1:10 604.4 19.8 Serum-6 1:400 <1:10 65.1 0.56 IgG was significantly reduced after passing through the columns. Table III showed that after passage through all three columns, neutralizing antibodies in sera were completely eliminated in all six samples and total anti-Ad IgG was reduced to 0.5 to 5% of initial levels (Table III). Further study demonstrated that the depletion of anti-Ad antibodies by an Ad antigen column was specific to Ad capsid proteins. This immunoapheresis column for specific removal of Ad antibody is currently under development at Calydon. VI. Clinical Development of CV706 and CV787 CV706 and CV787 are novel therapeutic agents with a novel mechanism of action. As of this writing, a phase I/II clinical trial of CV706 for recurrent local prostate cancer has been completed and CV787 has entered a multicenter Phase I/II clinical trial for metastatic, hormonal refractory prostate cancer. A brief summary for CV706 trial result and factors impacting clinical efficacy and safety are discussed. A. CV706 Phase I/II Trial for Locally Recurrent Prostate Cancer A Phase I trial of CV706 was initiated in 1998 at the Brady Urological Institute of the Johns Hopkins Oncology Center under the direction of Jonathan Simons, MD, and Ted DeWeese, MD. The patient population consists of men with locally recurrent prostate cancer with rising PSA levels following definitive external beam irradiation. Men in this category are usually left untreated or receive androgen ablation therapy
as serum PSA levels rise significantly above 10 ng/mL. On average, these men have a life expectancy of 3 years. The virus was administered under spinal anesthesia using the brachytherapy template 10. Development of ARCAs for the Treatment of Prostate Cancer 3 1 5 and ultrasound 3D imaging using the MMS Terapac Plus 6.6 B3DTUI (Char lottesville, VA) treatment-planning software for implantation of radioactive seeds. Virus w âs initially administered w îth 0.1 mL aliquots from up to 40 brachytherapy needles. PSA levels were determined and biopsies obtained. Systemic toxicity was minimal and limited to brief Grade 1 fever with or without an associated chill. These episodes were self-limited, responded to routine anti-pyretics, and no patient required antibiotics. This phenomenon is consistent with previously reported series using intratumoral injections of replication-competent adenovirus (Onyx-015) [144, 145] and likely represents cytokine release (e.g., IL-1, IL-6, TNF-a) in response to the adenovirus [146, 147]. A transient non-clinically significant lymphopenia confined to the normal range was noted in a majority of patients (95%) within 24 h of viral instillation. Full recovery of cell counts occurred within 4 to 7 days posttreatment. As with the transient fever, the timing of this decrement combined with the quick recovery are consistent with an acute-phase reaction mediated by transient cytokine release, as occurs with a variety of agents including bisphosphonates, and are not consistent with viral induced bone marrow suppression [148, 149]. Importantly, treatment with CV706 was not associated with significant hepatic or coagulation abnormalities. No patient experienced >Grade 2 elevation of liver transaminase levels and no patient had evidence of alteration in PT or PTT or a decrement in fibrinogen. This safety was evident even at the highest dose level of 1 x 10^^ viral particles and with viral shedding into the blood as documented in this study. Taken together, these data reveal a high degree of safety and tolerability of CV706 when administered by intraprostatic injection. The analysis of secondary study end points provided compelling evidence of CV706 activity. Serum PSA is well known to be a marker of both disease activity as well as disease burden and the use of serum PSA as a marker of therapeutic efficacy has become increasingly well defined. Several investigators have correlated declines in serum PSA of greater than 50% with prolongation of survival in men with hormone refractory prostate cancer [150, 151]. In addition, other investigators have found that a slowly rising PSA following definitive management with radiation or surgery is associated with an increased time to clinically evident metastatic disease when compared to patients with a more rapid PSA doubling time [152, 153]. Moreover, there was a statistically significant reduction in the PSA velocity following treatment with CV706, most pronounced for patients in dose levels 4 and 5, again suggestive of a dose-response relationship. In the final two dose levels, 50% of treated patients achieved a PSA partial remission (PR). It also appears that treatment with CV706 resulted in a prolongation of the time required for the serum PSA to double, suggesting a slowing of cancer growth within the prostate even among individuals not achieving a PR as defined by the protocol. It is well known that biopsy of the prostate results in significant elevations in serum PSA for 2 weeks [154, 155]. It is 3 1 6 Henderson and Yu likely that while the design of this study, with frequent posttreatment biopsies, aided in the documentation of viral replication, these same invasive procedures prevented a full analysis of the PSA response to therapy with CV706. Thus, it is possible that significant reductions in serum PSA could have been obscured by these frequent prostatic manipulations. Despite this possibility, the evidence gathered on PSA levels following treatment with CV706 are encouraging and suggest that at the higher dose levels, a clinically meaningful treatment effect may be achievable. This treatment effect at higher doses is associated with histologic evidence of viral replication. The viral inclusions seen on electron microscopy are consistent with viral replication in prostate epithelial cells. The positive staining for hexon protein seen on immuohistochemistry from day 4 biopsy materials is also confined to prostatic epithelial cells, is greatest in the highest dose levels, and, like the electron microscopy, is highly suggestive of intraprostatic replication of CV706 in these patients. Therefore, there is appropriate rationale for optimism given these findings in men treated with CV706, particularly at the high dose level. We believe these data to be significant and warrant CV706 evaluation in a Phase II study. Importantly, we were able to rigorously document viral circulation in the blood following intraprostatic delivery of CV706 without significant associated clinical sequale. The quantitative PCR assay is very specific for CV706 and is capable of detecting 1300 copies per milliliter of plasma. These results confirm that a small but significant amount of the intraprostatically administered virus reached the circulation. The amount of virus released in the first "peak" varied between patients and did not appear to be related to the dose level or neutralizing antibody titer. The highest total amount of virus detected was in two patients (patients 12 and 14) with an estimate of less than 2% of the dose being detected. Circulating virus then became undetectable analogous to a virus "eclipse." A significant secondary "peak" of circulating CV706 was observed in most patients within 3 days of treatment, suggestive of viral replication in these patients. The appearance and size of the secondary "peak" seemed to correlate best inversely with the anti-Ad5 antibody titer at the time of treatment. These data are consistent with those derived from electron microscopy and immunohistochemistry, and are highly suggestive of CV706 replication in the human prostate. Response of PSA to CV706 delivered directly into the prostate was not correlated with the presence of preexisting Ad5 neutralizing antibod ies. As expected, following CV706 administration, most patients developed Ad5-neutralizing antibodies. However, development of neutralizing antibod ies failed to correlate with response to treatment. Moreover, our data also reveal that the presence of preexisting anti-Ad5 antibodies was not correlated with treatment-related toxicity. These data extend the previously reported 10. Development of ARCAs for the Treatment of Prostate Cancer 3 1 7 work on intratumoral delivery of adenovirus by revealing a lack of asso ciation betvv êen preexisting neutralizing antibodies and treatment efficacy and toxicity [118, 133]. While circulating anti-Ad5 antibody may significantly impact on the efficacy and toxicity of systemically administered adenovirus [75, 86], it is not clear that these antibodies have the same access to the tumor-bearing prostate and thus may have a limited impact on direct intratumoral injections [156]. In summary, these data reveal that CV706 is safe and not associated v^ith irreversible serious short- or long-term side effects w^hen delivered by intratumoral injection using a planned, stereotactic approach. These data also suggest that CV706 replicates selectively in prostatic epithelial cells, i.e., those prostate cells that make PSA, and does so in a time frame consistent w îth an adenovirus replication cycle. These data suggest that CV706 has significant biologic activity as evidenced by significant durable dose-related decreases in patient serum PSA. Indeed, three of five men treated w îth 1 x 10^^ particles of CV706 experienced a PSA partial response. Thus, CV706 delivered by brachytherapy is an excellent candidate for the treatment of organ-confined prostate cancer. B. Factors Impacting Clinical Efficacy and Safety The pathogenesis of adenoviral infections is influenced by a large number of factors, some pertaining to the virus and others pertaining to the host defenses of the virus. Important issues for the virus include the route of infec tion, the size of the virus inoculum, the tropism of the virus for different-cell types, and w^hether the virus spreads directly from cell to cell or through extra cellular fluid. Clearly, the vascularization of tumors, the leakiness of capillaries to virus, and the physical size of virus particle w îll affect intratumoral virus distribution. In the replication efficiency of the virus in prostate tumor cells, both the time of the replication cycle and the burst size are also important. Host defenses include mechanical defenses (epithelia, mucosal, liver Kupffer cells, or the blood-brain barrier), nonspecific immune defenses (interferons, recog nition of infected cells by natural killer cells, release of cytokines, macrophage recruitment and activation, and triggering of complement and kinin cascades), and specific immune defenses (humoral immunity, mostly IgM and IgG but also IgA, IgD, and IgE, and finally cell-mediated immunity) [152]. In adenovirus-mediated prostate cancer therapy, the virus can be either injected directly into the tumor or administered by intravenous injection. In either case, the dose of virus is massive (10^^-10^^ particles) compared to natural, vaccine-induced adenovirus infections (10^-10^ particles) [153, 157, 158], or the clinical trials w îth w^ild-type adenovirus (10^-10^ particles) [14]. Very little is knov^n about the human host response to large doses of ade noviruses [14, 159] and nothing is knov^n about the human host response 3 1 8 Henderson and Yu to using the intravenous route of administration of large doses of replicating adenoviruses. Liver toxicity of virion proteins may be limiting at these high doses. Therapeutic antibody studies have indicated that antibodies do not effec tively penetrate the core of a solid tumor; extravasation is limited to the tumor periphery. This suggests that the accessibility of replicating virus to antibody binding should be minimal foUow îng direct intratumoral injection [98, 131]. Cell-mediated immunity directed tow^ard infected tumor cells may actually enhance the efficacy of replicating viruses in cancer patients if enough replica tion and spread occur initially. How^ever, a systemically delivered replicating adenovirus is going to face several potential hurdles: (1) the nonspecific removal of adenovirus by liver Kupffer cells, (2) the inactivation of virus by preexisting circulating antibodies to adenovirus, (3) a limitation of viral replication medi ated by a vigorous CTL response to virally infected cells, and (4) a limitation of the efficacy of repeat dosage by primary or secondary induction of humoral immunity. Incorporation of the Ad5 genome into germ cells has been expressed as a concern but has not been found for any of the Ad5 gene therapy constructs. Indeed, adenovirus gene expression is characterized as transient in nature due to a lack of viral DNA integration. Virus shedding has been expressed as a concern but it has not been detected in any Ad5 clinical trial to date. In our clinical trial, virus replication w âs detected after 2 - 8 days but w âs undetectable after 2 weeks. It is difficult to estimate the increased cytolytic activity in humans of CV787 compared to CV706. However, replicating adenoviruses containing hepatitis B surface antigen (HbsAg), with and without the E3 region, have been tested in chimpanzees, a system permissive for infection by human adenoviruses [160]. In this study, the addition of the E3 region resulted in 10- to 100-fold increase in virus shedding and a 10- to 100-fold increase in titer to HBsAg. However, one should not lose sight of the fact that adenoviruses are ubiquitous. Twenty-three percent of normal healthy infants are seropositive for adenoviruses by 7 months of age [95] and CV787 is attenuated 10,000:1 compared to the wild-type virus. We believe the therapeutic use of CV787 will be safe; the major question is whether or not there is sufficient efficacy to be medically useful. VII. Summary Safety of administrating wild-type Ad5 either intratumorally and intra venously was demonstrated at intermediate doses (10̂ ^ to 10^ particles) over 40 years ago [14]. None of the treated patients had significant side-effects. Safety and efficacy will be the major issues as adenovirus doses escalate from 10^^ to lO^^particles. CV787 is a replication-competent adenovirus atten uated 10,000:1 compared to the wild-type virus in PSA~ cells. This is an 10. Development of ARCAs for the Treatment of Prostate Cancer 3 1 9 unprecedented therapeutic index for a cytotoxic agent as measured in vitro. CV787 is capable of eliminating distant mouse xenograft tumors with a single intravenous injection. Synergy of prostate cancer-specific adenovirus variants has been demonstrated w^hen combined with conventional therapies including radiotherapy and chemotherapy. Synergy of CV787 and docetaxel has opened a single-dose curative therapeutic window in excess of 1,000:1 in vivo. This progress, coupled with screening and immunoapheresis to control and remove preexistent antibody to adenovirus [75^ 161], may overcome the hurdles and achieve clinical benefits with intravenous administration of adenovirus-based therapeutics.
Ongoing clinical trials of agents such as CV787 will resolve the issues of safety and efficacy and hopefully point to a new mode of cancer chemotherapy, one that includes the use of targeted cytolytic adenoviruses. References 1. Donehoweer, R. C , Abeloff, M. D., and Perry, M. C. (1995). Chemotherapy. In "Chni- cal Oncology" (M. D. Abeloff, J. O. Armitage, A. S. Lichter, and J. E. Niederhuber, Eds.), pp. 201-218. Churchill Livingston, New York. 2. Scher, H. I., Isaacs, J. T., Fuks, Z., and Walsh, P. C. (1995). Prostate. In "Clinical Oncology" (M. D. Abeloff, J. O. Armitage, A. S. Lichter, and J. E. Niederhuber, Eds.), pp. 1439-1472. Churchill Livingston, New^ York. 3. Golomb, H. M., Fefer, A., Golde, D. W., Ozer, H., Portlock, C , Silber, R., Rappeport, J., Ratain, M. J., Thompson, J., Bonnem, E., et al. (1988). Report of a multi-institutional study of 193 patients with hairy cell leukemia treated with interferon-alfa2b. Semin. Oncol. 15, 7 -9 . 4. Demetri, G. D., and Anderson, K. C. (1995). Bone Marrow Failure. In "Clinical Oncology" (M. D. Abeloff, J. O. Armitage, A. S. Lichter, and J. E. Niederhuber, Eds.), pp. 433-456. Churchill Livingston, New York. 5. Anderson, W. F. (1994). Gene therapy for cancer. Hum. Gene Ther. 5, 1-2. 6. Wilson, J. M. (1996) Adenoviruses as gene-delivery vehicles. N. Engl. J. Med. 334, 1185-1187. 7. Elshami, A, A., Kucharczuk, J. C , Sterman, D. H., Smythe, W. R., Hwang, H. C , Amin, K. M.,Litzky, L. A., Albelda, S. M., and Kaiser, L. R. (1995). The role of immunosuppression in the efficacy of cancer gene therapy using adenovirus transfer of the herpes simplex thymidine kinase gene. Ann. Surg. 222, 298-307; 307-310. 8. Kaneko, S., Hallenbeck, P., Kotani, T., Nakabayashi, H., McGarrity, G., Tamaoki, T., Ander son, W. F., and Chiang, Y. L. (1995). Adenovirus-mediated gene therapy of hepatocellular carcinoma using cancer-specific gene expression. Cancer Res. 55, 5283-5287. 9. Kanai, F., Shiratori, Y., Yoshida, Y., Wakimoto, H., Hamada, H., Kanegae, Y., Saito, L, Nakabayashi, H., Tamaoki, T., Tanaka, T., Lan, K. H., Kato, N., Shiina, S., and Omata, M. (1996). Gene therapy for alpha-fetoprotein-producing human hepatoma cells by adenovirus-mediated transfer of the herpes simplex virus thymidine kinase gene. Hepatology 23,1359-1368. 10. Kanai, F., Lan, K. H., Shiratori, Y., Tanaka, T., Ohashi, M., Okudaira, T., Yoshida, Y., Wakimoto, H., Hamada, H., Nakabayashi, H., Tamaoki, T., and Omata, M. (1997). In vivo gene therapy for alpha-fetoprotein-producing hepatocellular carcinoma by adenovirus- mediated transfer of cytosine deaminase gene. Cancer Res. 57y 461-5 . 3 2 0 Henderson and Yu 11. Su, H., Lu, R., Chang, J. C , and Kan, Y. W. (1997). Tissue-specific expression of herpes simplex virus thymidine kinase gene dehvered by adeno-associated virus inhibits the grov^th of human hepatocellular carcinoma in athymic mice. Proc. Natl. Acad. Set. USA 94, 13,891-13,896. 12. Dock, G. (1904). Influence of complicating diseases upon leukemia. Am. J. Med. Sci. 127, 536-592. 13. Moore, A. E. (1954). Effects of viruses on tumors. Annu. Rev. Microbiol. 8, 393-410. 14. Smith, R. R., Huebner, R. J., Rowe, W. P., Schatten, W. F., and Thomas, L. B. (1956). Studies on the use of viruses in the treatment of carcinoma of the cervix. Cancer 9, 1211-1218. 15. Bischoff, J. R., Kirn, D. H., Williams, A., Heise, C , Horn, S., Muna, M., Ng, L., Nye, J. A., Sampson-Johannes, A., Fattaey, A., and McCormick, F. (1996). An adenovirus mutant that replicates selectively in p53-deficient human tumor cells. Science 274, 373-376. 16. Hall, A. R., Dix, B. R., O'Carroll, S. J., and Braithwaite, A. W. (1998). p53-dependent cell death/apoptosis is required for a productive adenovirus infection. Nat. Med. 4, 1068-1072. 17. Goodrum, F. D., and Ornelles, D. A. (1998). p53 status does not determine outcome of ElB 55-kilodalton mutant adenovirus lytic infection. / . Virol. 71, 9479-9490. 18. Horridge, J. J., and Leppard, K. N. (1998). RNA-binding activity of the ElB 55-kilodalton protein from human adenovirus type 5 . / . Virol. 72, 9374-9379. 19. Gabler, S., Schutt, H., Groitl, P., Wolf, H., Shenk, T., and Dobner, T. (1998). ElB 55- kilodalton-associated protein: A cellular protein with RNA-binding activity implicated in nucleocytoplasmic transport of adenovirus and cellular mRNAs./. Virol. 71, 7960-7971. 20. Zhang, J. F., Hu, C., Geng, Y., Selm, J., Klein, S. B., Orazi, A., and Taylor, M. W. (1996). Treatment of a human breast cancer xenograft with an adenovirus vector containing an interferon gene results in rapid regression due to viral oncolysis and gene therapy. Proc. Natl. Acad. Sci. VSA 93, 4513-4518. 21. Rodriguez, R., Schuur, E. R., Lim, H. Y., Henderson, G. A., Simons, J. W., and Henderson, D. R. (1997). Prostate attenuated replication competent adenovirus (ARCA) CN706: A selective cytotoxic for prostate-specific antigen-positive prostate cancer cells. Cancer Res. 57, 2559-2563. 22. Yu, D. C., Chen, Y., Seng, M., Dilley, J., and Henderson, D. R. (1999). The addition of adenovirus type 5 region E3 enables calydon virus 787 to eliminate distant prostate tumor xenografts. Cancer Res. 59, 4200-4203. 23. Yu, D. C , Sakamoto, G. T., and Henderson, D. R. (1999). Identification of the transcrip tional regulatory sequences of human kallikrein 2 and their use in the construction of calydon virus 764, an attenuated replication competent adenovirus for prostate cancer therapy. Cancer Res. 59, 1498-504. 24. Aumuller, G., Seitz, J., Lilja, H., Abrahamsson, P. A., von der Kammer, H., and Scheit, K. H. (1990). Species- and organ-specificity of secretory proteins derived from human prostate and seminal vesicles. Prostate 17, 31-40. 25. Scher, H. I., and Fossa, S. (1995). Prostate cancer in the era of prostate-specific antigen. Curr. Opin. Oncol. 7, 281-291. 26. Wilding, G. (1995). Endocrine control of prostate cancer. Cancer Surv. 23, 43-62. 27. Smith, P. H. (1995). Hormone therapy: an overview. Cancer Surv. 23, 171-181. 28. Walsh, P. C. (1994). Radical prostatectomy: A procedure in evolution. Sentin. Oncol. 21, 662-671. 29. Rowe, W. P., Huebner, R. J., Gilmore, L. K., Parrott, R. H., and Ward, T. G. (1953). Isola tion of a cytopathogenic agent from human adenoids undergoing spontaneous degeneration in tissue culture. Proc. Soc. Exp. Med. 84, 570-573. 30. Zahradnik, J. M., Spencer, M. J., and Porter, D. D. (1980). Adenovirus infection in the immunocompromised patient. Am. J. Med. 68, 725-732. 10. Development of ARCAs for the Treatment of Prostate Cancer 3 2 1 31. Shields, A. F., Hackman, R. C , Fife, K. H., Corey, L., and Meyers, J. D. (1985). Adenovirus infections in patients undergoing bone-marrow transplantation. N. Engl. } . Med. 312, 529-533. 32. Horwitz, M. S. (1990). "Adenovirus and Their Replication." Raven Press, Ltd., Nev^ York, pp. 1679-1721. 33. Chroboczek, J., Bieber, F., and Jacrot, B. (1992). The sequence of the genome of adenovirus type 5 and its comparison with the genome of adenovirus type 2. Virology 186, 280-285. 34. Garon, C. F., Berry, K. W., and Rose, J. A. (1972). A unique form of terminal redundancy in adenovirus DNA molecules. Proc. Natl. Acad. Set. USA 69, 2391-2395. 35. Linne, T., Jornvall, H., and Philipson, L. (1977). Purification and characterization of the phosphorylated DNA-binding protein from adenovirus-type-2-infected cells. Eur.}. Biochem. 76,481-490. 36. Lichy, J. H., Field, J., Horwitz, M. S., and Flurwitz, J. (1982). Separation of the adenovirus terminal protein precursor from its associated DNA polymerase: Role of both proteins in the initiation of adenovirus DNA replication. Proc. Natl. Acad. Set. USA 79, 5225-5229. 37. Lichy, J. H., Nagata, K., Friefeld, B. R., Enomoto, T., Field, J., Guggenheimer, R. A., Ikeda, J. E., Horwitz, M. S., and Hurwitz, J. (1983). Isolation of proteins involved in the replication of adenoviral DNA in vitro. Cold Spring Harbor Symp. Quant. Biol. 47, 731-740. 38. Friefeld, B. R., Lichy, J. H., Hurwitz, J., and Horwitz, M. S. (1983). Evidence for an altered adenovirus DNA polymerase in cells infected with the mutant H5tsl49. Proc. Natl. Acad. Sci. USA 80, 1589-1593. 39. Ghosh-Choudhury, G., Haj-Ahmad, Y., Brinkley, P., Rudy, J., and Graham, F. L. (1986). Human adenovirus cloning vectors based on infectious bacterial plasmids. Gene 50, 161-171. 40. Graham, F. L. (1990). Adenoviruses as expression vectors and recombinant vaccines. Trends Biotechnol. 8, 85-87. 41. Feuerbach, D., and Burgert, H. G. (1993). Novel proteins associated with MHC class I antigens in cells expressing the adenovirus protein E3/19K. EMBO J. 12, 3153-3161. 42. Beier, D. C., Cox, J. H., Vining, D. R., Cresswell, P., and Engelhard, V. H. (1994). Associ ation of human class I MHC alleles with the adenovirus E3/19K protein./. Immunol. 152, 3862-3872. 43. Lee, M. G., Abina, M. A., Haddada, H., and Perricaudet, M. (1995). The constitutive expression of the immunomodulatory gpl9k protein in E1-, E3- adenoviral vectors strongly reduces the host cytotoxic T cell response against the vector. Gene Ther. 2, 256-262. 44. Lieber, A., He, C. Y., Meuse, L., Schowalter, D., Kirillova, L, Winther, B., and Kay, M. A. (1997). The role of Kupffer cell activation and viral gene expression in early liver toxicity after infusion of recombinant adenovirus vectors. / . Virol. 71, 8798-8807. 45. Gooding, L. R., Aquino, L., Duerksen-Hughes, P. J., Day, D., Horton, T. M., Yei, S. P., and Wold, W. S. (1991). The ElB 19,000-molecular-weight protein of group C adenoviruses prevents tumor necrosis factor cytolysis of human cells but not of mouse cells. / . Virol. 65, 3083-3094. 46. Gooding, L. R., Ranheim, T. S., Tollefson, A. E., Aquino, L., Duerksen-Hughes, P., Hor ton, T. M., and Wold, W. S. (1991). The 10,400- and 14,500-dalton proteins encoded by region E3 of adenovirus function together to protect many but not all mouse cell lines against lysis by tumor necrosis factor. / . Virol. 65, 4114-4123. 47. Thomas, G. P., and Mathews, M. B. (1980). DNA replication and the early to late transition in adenovirus infection. Cell 22, 523-533. 48. Shaw, A. R., and Ziff, E. B. (1980). Transcripts from the adenovirus-2 major late promoter yield a single early family of 3' coterminal mRNAs and five late families. Cell 22, 905-916. 49. Shaw, A. R., and Ziff, E. B. (1982). Selective inhibition of adenovirus type 2 early region II and III transcription by an anisomycin block of protein synthesis. Mol. Cell Biol. 2, 789. 3 2 2 Henderson and Yu 50. Green, M., and Daesch, G. E. (1961). Biochemical studies on adenovirus multiplication. I. Kinetics: If nucleic acid and protein synthesis in suspension cultures. Virology 13,169-176. 51. Wold, S. M., Green, M., and Buttner, W. (1978). "The Molecular Biology of Animal Viruses." Dekker, New York. 52. Berk, A.J. (1986). Adenovirus promoters and ElA transactivation. Annu. Rev. Genet. 20, 45-79. 53. Nevins,J. R. (1991). Transcriptional activation by viral regulatory proteins. Trends Biochem. Set. 16, 435-439. 54. Berk, A. J., and Sharp, P. A. (1978). Structure of the adenovirus 2 early mRNAs. Cell 14, 695-711. 55. Nevins, J. R., Ginsberg, H. S., Blanchard, J. M., Wilson, M. C., and Darnell, J. E. (1979). Regulation of the primary expression of the early adenovirus transcription units. / . Virol. 32, 727-733. 56. Schuur, E. R., Henderson, G. A., Kmetec, L. A., Miller, J. D., Lamparski, H. G., and Hender son, D. R. (1996). Prostate-specific antigen expression is regulated by an upstream enhancer. / . Biol. Chem. 271, 7043-7051. 57. Brookes, D. E., Zandvliet, D., Watt, F., Russell, P. J., and MoUoy, P. L. (1998). Relative activity and specificity of promoters from prostate-expressed genes. Prostate 35, 18-26. 58. Greenberg, N. M., DeMayo, F. J., Sheppard, P. C., Barrios, R., Lebovitz, R., Finegold, M., Angelopoulou, R., Dodd, J. G., Duckworth, M. L., Rosen, J. M., et al. (1994). The rat probasin gene promoter directs hormonally and developmentally regulated expression of a heterologous gene specifically to the prostate in transgenic mice. Mol. Endocrinol. 8, 230-239. 59. Riegman, P. H., Vlietstra, R. J., Suurmeijer, L., Cleutjens, C. B., and Trapman, J. (1992). Characterization of the human kallikrein locus. Genomics 14, 6 -11 . 60. Schedlich, L. J., Bennetts, B. H., and Morris, B.J. (1987). Primary structure of a human glandular kallikrein gene. DNA 6, 429-437. 61. Morris, B.J. (1989). hGK-1: A kallikrein gene expressed in human prostate. Clin. Exp. Pharmacol. Physiol. 16, 345-351. 62. Haj-Ahmad, Y., and Graham, F. L. (1986). Development of a helper-independent human adenovirus vector and its use in the transfer of the herpes simplex virus thymidine kinase gene. / . Virol. 57, 267-274. 63. Robbins, P. D., Tahara, H., and Ghivizzani, S. C. (1998). Viral
vectors for gene therapy. Trends Biotechnol. 16, 35-40. 64. Wivel, N. A., and Wilson, J. M. (1998). Methods of gene delivery. Hematol. Oncol. Clin. North. Am. 12 ,483-501. 65. Levine, A.J., and Ginsberg, H. S. (1967). Mechanism by which fiber antigen inhibits multi plication of type 5 adenovirus./. Virol. 1, 747-757. 66. Ueno, N. T., Bartholomeusz, C., Herrmann, J. L., Estrov, Z., Shao, R., Andreeff, M., Price, J., Paul, R. W., Anklesaria, P., Yu, D., and Hung, M. C. (2000). ElA-mediated paclitaxel sensitization in HER-2/neu-overexpressing ovarian cancer SKOV3.ipl through apoptosis involving the caspase-3 pathway. Clin. Cancer Res. 6, 250-259. 67. Teodoro, J. G., Shore, G. C., and Branton, P. E. (1995). Adenovirus ElA proteins induce apoptosis by both p53-dependent and p53-independent mechanisms. Oncogene 11,467-474. 68. Marcellus, R. C., Teodoro, J. G., Wu, T., Brough, D. E., Ketner, G., Shore, G. C., and Branton, P. E. (1996). Adenovirus type 5 early region 4 is responsible for ElA-induced p53-independent apoptosis./. Virol. 70, 6207-6215. 69. Han, J., Sabbatini, P., Perez, D., Rao, L., Modha, D., and White, E. (1996). The ElB 19K protein blocks apoptosis by interacting with and inhibiting the p53-inducible and death-promoting Bax protein. Genes Dev. 10, 461-477. 10. Development of ARCAs for the Treatment of Prostate Cancer 3 2 3 70. Desai, S. Y., Patel, R. C , Sen, G. C , Malhotra, P., Ghadge, G. D., and Thimmapaya, B. (1995). Activation of interferon-inducible 2^-5' oligoadenylate synthetase by adenoviral VAI RNA./. BioL Chem. 270, 3454-3461. 71. Diaz-Guerra, M., Rivas, C., and Esteban, M. (1997). Activation of the IFN-inducible enzyme RNase L causes apoptosis of animal cells. Virology 236, 354-363. 72. Tollefson, A. E., Scaria, A., Hermiston, T. W., Ryerse, J. S., Wold, L. J,, and Wold, W. S. (1996). The adenovirus death protein (E3-11.6K) is required at very late stages of infection for efficient cell lysis and release of adenovirus from infected cells. / . Virol, 70, 2296-2306. 73. Tollefson, A. E., Ryerse, J. S., Scaria, A., Hermiston, T. W., and Wold, W. S. (1996). The E3- 11.6-kDa adenovirus death protein (ADP) is required for efficient cell death: Characterization of cells infected vŝ ith adp mutants. Virology 220, 152-162. 74. Chu, J. S., Huang, C. S., and Chang, K.J. (1998). The prognostic significance of tumor angiogenesis in Taiv^anese patients v^ith invasive ductal breast carcinomas. Cancer Lett. 134, 7-14. 75. Chen, Y., Yu, D. C , Charlton, D., and Henderson, D. R. (2000). Pre-existent adenovirus antibody inhibits systemic toxicity and antitumor activity of CN706 in the nude mouse LNCaP xenograft model: Implications and proposals for human therapy. Hum. Gene Ther. 11,1553-1567. 76. Pollack, A., and Zagars, G. K. (1997). External beam radiotherapy dose response of prostate cancer. Int. J. Radiat. Oncol. Biol. Phys. 39, 1011-1018. 77. Hanks, G. E., Hanlon, A. L., Schultheiss, T. E., Pinover, W. H., Movsas, B., Epstein, B. E., and Hunt, M. A. (1998). Dose escalation with 3D conformal treatment: Five year outcomes, treatment optimization, and future directions. Int. J. Radiat. Oncol. Biol. Phys. 41,501-510. 78. Zelefsky, M. J., Leibel, S. A., Gaudin, P. B., Kutcher, G. J., Fleshner, N. E., Venkatra- men, E. S., Renter, V. E., Fair, W. R., Ling, C. C , and Fuks, Z. (1998). Dose escalation v^ith three-dimensional conformal radiation therapy affects the outcome in prostate cancer. Int. J. Radiat. Oncol. BioL Phys. 41, 491-500. 79. Garzotto, M., Haimovitz-Friedman, A., Liao, W. C , White-Jones, M., Huryk, R., Hes- ton, W. D., Cardon-Cardo, C , Kolesnick, R., and Fuks, Z. (1999). Reversal of radiation resistance in LNCaP cells by targeting apoptosis through ceramide synthase. Cancer Res. 59, 5194-5201. 80. Jones, N., and Shenk, T. (1979). An adenovirus type 5 early gene function regulates expression of other early viral genes. Proc. Natl. Acad. Sci. USA 76, 3665-3669. 81. Martin-Duque, P., Sanchez-Prieto, R., Romero, J., Martinez-Lamparero, A., Cebrian- Sagarriga, S., Guinea-Viniegra, J., Dominguez, C , Lleonart, M., Cano, A., Quintanilla, M., and Ramon, Y. C. S. (1999). In vivo radiosensitizing effect of the adenovirus El A gene in murine and human malignant tumors. Int. J. Oncol. 15, 1163-1168. 82. Rogulski, K. R., Freytag, S. O., Zhang, K., Gilbert, J. D., PaieUi, D. L., Kim, J. H., Heise, C. C , and Kirn, D. H. (2000). In vivo antitumor activity of ONYX-015 is influenced by p53 status and is augmented by radiotherapy. Cancer Res. 60, 1193-1196. 83. Gorski, D. H., Beckett, M. A., Jaskowiak, N. T., Calvin, D. P., Mauceri, H. J., Salloum, R. M., Seetharam, S., Koons, A., Hari, D. M., Kufe, D. W., and Weichselbaum, R. R. (1999). Blockage of the vascular endothelial growth factor stress response increases the antitumor effects of ionizing radiation. Cancer Res. 59, 3374-3378. 84. Giatromanolaki, A., Sivridis, E., Koukourakis, M. I., Georgoulias, V., Gatter, K. C , and Harris, A. L. (1999). Intratumoral angiogenesis: A new prognostic indicator for stage I endometrial adenocarcinomas? Oncol. Res. 11, 205-212. 85. Giatromanolaki, A., Koukourakis, M. I., Theodossiou, D., Barbatis, K., O'Byrne, K., Harris, A. L., and Gatter, K. C. (1997). Comparative evaluation of angiogenesis assessment with anti-factor-VIII and anti-CD31 immunostaining in non-small cell lung cancer. Clin. Cancer. Res. 3, 2485-2492. 3 2 4 Henderson and Yu 86. Nemunaitis, J., Swisher, S. G., Timmons, T., Connors, D., Mack, M., Doerksen, L., Weill, D., Wait, J., Lawrence, D. D., Kemp, B. L., Fossella, F., Glisson, B. S., Hong, W. K., Khuri, F. R., Kurie, J. M., Lee, J. J., Lee, J. S., Nguyen, D. M., Nesbitt, J. C , Perez-Soler, R., Pisters, K. M., Putnam, J. B., Richli, W. R., Shin, D. M., Walsh, G. L., et al. (2000). Adenovirus-mediated p53 gene transfer in sequence with cisplatin to tumors of patients with non-small-cell lung cancer./. Clin. Oncol. 18, 609-622. 87. Oh, W. K., and Kantoff, P. W. (1999). Treatment of locally advanced prostate cancer: Is chemotherapy the next step? / . Clin. Oncol. 17, 3664-3675. 88. Yu, D. C., Chen, Y., Dilley, J., Embry, M., Zhang, H., Nguyen, N., Amin, P., Oh, J., and Henderson, D. R. (2001). Antitumor synergy of CV787, a prostate cancer-specific adenovirus, and pachtaxel and docetaxel. Cancer Res. 61, 517-525. 89. Nielsen, L. L., Lipari, P., Dell, J., Gurnani, M., and Hajian, G. (1998). Adenovirus-mediated p53 gene therapy and paclitaxel have synergistic efficacy in models of human head and neck, ovarian, prostate, and breast cancer. Clin. Cancer. Res. 4, 835-46. 90. Sanchez-Prieto, R., Quintanilla, M., Cano, A., Leonart, M. L., Martin, P., Anaya, A., and Ramon y Cajal, S. (1996). Carcinoma cell lines become sensitive to DNA-damaging agents by the expression of the adenovirus El A gene. Oncogene 13, 1083-1092. 91. Lowe, S. W., Ruley, H. E., Jacks, T., and Housman, D. E. (1993). p53-dependent apoptosis modulates the cytotoxicity of anticancer agents. Cell 74, 957-967. 92. Debbas, M., and White, E. (1993). Wild-type p53 mediates apoptosis by ElA, which is inhibited by ElB. Genes Dev. 7, 546-554. 93. Molnar-Kimber, K. L. (1996). Mechanism of action of rapamycin (Sirolimus, Rapamune). Transplant Proc. 28, 964-969. 94. Paielli, D. L., Wing, M. S., Rogulski, K. R., Gilbert, J. D., Kolozsvary, A., Kim, J. H., Hughes, J., Schnell, M., Thompson, T., and Freytag, S. O. (2000). Evaluation of the biodistribution, persistence, toxicity, and potential of germ-line transmission of a replication-competent human adenovirus following intraprostatic administration in the mouse. Mol. Ther. 1, 263-274. 95. Nelson, J. K., Shields, M. D., Stewart, M. C , and Coyle, P. V. (1999). Investigation of seroprevalence of respiratory virus infections in an infant population with a multiantigen fluorescence immunoassay using heel-prick blood samples collected on filter paper. Pediatr. Res. 45,799-802. 96. D'Ambrosio, E., Del Grosso, N., Chicca, A., and MiduUa, M. (1982). Neutralizing antibodies against 33 human adenoviruses in normal children in Rome./. Hyg. (London) 89, 155-161. 97. Piedra, P. A., Poveda, G. A., Ramsey, B., McCoy, K., and Hiatt, P. W. (1998). Incidence and prevalence of neutralizing antibodies to the common adenoviruses in children with cystic fibrosis: Implication for gene therapy with adenovirus vectors. Pediatrics 101, 1013-1019. 98. Schulick, A. H., Vassalli, G., Dunn, P. F., Dong, G., Rade, J. J., Zamarron, C , and Dichek, D. A. (1997). Established immunity precludes adenovirus-mediated gene transfer in rat carotid arteries. Potential for immunosuppression and vector engineering to overcome barriers of immunity. / . Clin. Invest. 99, 209-219. 99. Fang, B., Eisensmith, R. C , Wang, H., Kay, M. A., Cross, R. E., Landen, C. N., Gordon, G., Bellinger, D. A., Read, M. S., Hu, P. C , et al. (1995). Gene therapy for hemophilia B: host immunosuppression prolongs the therapeutic effect of adenovirus-mediated factor IX expression. Hum. Gene Ther. 6, 1039-1044. 100. Ilan, Y., Jona, V. K., Sengupta, K., Davidson, A., Horwitz, M. S., Roy-Chowdhury, N., and Roy-Chowdhury, J. (1997). Transient immunosuppression with FK506 permits long-term expression of therapeutic genes introduced into the liver using recombinant adenoviruses in the rat. Hepatology 26, 949-956. 101. Lochmuller, H., Petrof, B. J., Allen, C , Prescott, S., Massie, B., and Karpati, G. (1995). Immunosuppression by FK506 markedly prolongs expression of adenovirus-delivered trans- 10. Development of ARCAs for the Treatment of Prostate Cancer 3 2 5 gene in skeletal muscles of adult dystrophic [mdx] mice. Biochem. Biophys. Res. Commun. 213, 569-574. 102. Lochmuller, H., Petrof, B. J., Pari, G., Larochelle, N., Dodelet, V., Wang, Q., Allen, C , Prescott, S., Massie, B., Nalbantoglu, J., and Karpati, G. (1996). Transient immunosup pression by FK506 permits a sustained high-level dystrophin expression after adenovirus- mediated dystrophin minigene transfer to skeletal muscles of adult dystrophic (mdx) mice. Gene. Ther. 3, 706-716. 103. Vilquin, J. T., Guerette, B., Kinoshita, I., Roy, B., Goulet, M., Gravel, C., Roy, R., and Tremblay, J. P. (1995). FK506 immunosuppression to control the immune reactions triggered by first-generation adenovirus-mediated gene transfer. Hum. Gene Ther. 6, 1391-1401. 104. Jooss, K., Yang, Y., and Wilson, J. M. (1996). Cyclophosphamide diminishes inflammation and prolongs transgene expression follov^ing delivery of adenoviral vectors to mouse liver and lung. Hum. Gene Ther. 7, 1S5S-\SG6. 105. Smith, T. A., White, B. D., Gardner, J. M., Kaleko, M., and McClelland, A. (1996). Tran sient immunosuppression permits successful repetitive intravenous administration of an adenovirus vector. Gene Ther. 3, 496-502. 106. Yang, Y., Greenough, K., and Wilson, J. M. (1996). Transient immune blockade prevents formation of neutralizing antibody to recombinant adenovirus and allov^s repeated gene transfer to mouse liver. Gene Ther. 3, 412-420. 107. Jooss, K., Turka, L. A., and Wilson, J. M. (1998). Blunting of immune responses to adenoviral vectors in mouse liver and lung v^ith CTLA4Ig. Gene Ther. 5, 309-319. 108. Kay, M. A., Holterman, A. X., Meuse, L., Gov^n, A., Ochs, H. D., Linsley, P. S., and Wilson, C. B. (1995). Long-term hepatic adenovirus-mediated gene expression in mice following CTLA4Ig administration. Nat. Genet. 11, 191-197. 109. Kolls, J. K., Lei, D., Odom, G., Nelson, S., Summer, W. R., Gerber, M. A., and Shellito, J. E. (1996). Use of transient CD4 lymphocyte depletion to prolong transgene expression of El-deleted adenoviral vectors. Hum. Gene Ther. 7, 489-497. 110. Lei, D., Lehmann, M., Shellito, J. E., Nelson, S., Siegling, A., Volk, H. D., and Kolls, J. K. (1996). Nondepleting anti-CD4 antibody treatment prolongs lung-directed El-deleted adenovirus-mediated gene expression in rats. Hum. Gene Ther. 7,1T73-1T79. 111. Scaria, A., St George, J.A., Gregory, R. J., Noelle, R. J., Wadsworth, S. C , Smith, A. E., and Kaplan, J. M. (1997). Antibody to CD40 ligand inhibits both humoral and cellular immune responses to adenoviral vectors and facilitates repeated administration to mouse airv^ay. Gene Ther. ^,611-617. 112. McClane, S. J., Chirmule, N., Burke, C. V., and Raper, S. E. (1997). Characterization of the immune response after local delivery of recombinant adenovirus in murine pancreas and successful strategies for readministration. Hum. Gene Ther. 8, 2207-2216. 113. Kay, M. A., Meuse, L., Gov^n, A. M., Linsley, P., Hollenbaugh, D., Aruffo, A., Ochs, H. D., and Wilson, C. B. (1997). Transient immunomodulation with anti-CD40 ligand antibody and CTLA4Ig enhances persistence and secondary adenovirus-mediated gene transfer into mouse liver. Froc. Natl. Acad. Sci. USA 94, 4686-4691. 114. Cichon, G., and Strauss, M. (1998). Transient immunosuppression with 15-deoxyspergualin prolongs reporter gene expression and reduces humoral immune response after adenoviral gene transfer. Gene Ther. 5, 85-90. 115. Juillard, v. , Villefroy, P., Godfrin, D., Pavirani, A., Venet, A., and Guillet, J. G. (1995). Long-term humoral and cellular immunity induced by a single immunization with replication- defective adenovirus recombinant vector. Eur.
J. Immunol. 25, 3467-3473. 116. Otake, K., Ennist, D. L., Harrod, K., and Trapnell, B. C. (1998). Nonspecific inflammation inhibits adenovirus-mediated pulmonary gene transfer and expression independent of specific acquired immune responses. Hum. Gene Ther. 9, 2207-2222. 117. Li, H., Griscelli, P., Lindenmeyer, F., Opolon, P., Sun, L. Q., Connault, E., Soria, J., Soria, C , Perricaudet, M., Yeh, P., and Lu, H. (1999). Systemic delivery of antiangiogenic adenovirus 3 2 6 Henderson and Yu AdmATF induces liver resistance to metastasis and prolongs survival of mice. Hum. Gene Ther. 10, 3045-3053. 118. Molnar-Kimber, K. L., Sterman, D. H., Chang, M., Kang, E. H., ElBash, M., Lanuti, M., Elshami, A., Gelfand, K., Wilson, J. M., Kaiser, L. R., and Albelda, S. M. (1998). Impact of preexisting and induced humoral and cellular immune responses in an adenovirus-based gene therapy phase I clinical trial for locaHzed mesotheUoma. Hum. Gene Ther. 9, 2121-2133. 119. Harvey, B. G., Hackett, N. R., El-Saw^y, T,, Rosengart, T. K., Hirschov^itz, E. A., Lieberman, M. D., Lesser, M. L., and Crystal, R. G. (1999). Variability of human systemic humoral immune responses to adenovirus gene transfer vectors administered to different organs. / . Virol 73, 6729-6742. 120. Harvey, B. G., Worgall, S., Ely, S., Leopold, P. L., and Crystal, R. G. (1999). Cellular immune responses of healthy individuals to intradermal administration of an El-E3-adenovirus gene transfer vector. Hum. Gene Ther. 10, 2823-2837. 121. Kagami, H., Atkinson, J. C , Michalek, S. M., Handelman, B., Yu, S., Baum, B. J., and O'Connell, B. (1998). Repetitive adenovirus administration to the parotid gland: Role of immunological barriers and induction of oral tolerance. Hum. Gene Ther. 9, 305-313. 122. Gall, J. G., Crystal, R. G., and Falck-Pedersen, E. (1998). Construction and characteri zation of hexon-chimeric adenoviruses: Specification of adenovirus serotype. / . Virol. 71, 10,260-10,264. 123. Mack, C. A., Song, W. R., Carpenter, H., Wickham, T. J., Kovesdi, L, Harvey, B. G., Magovern, C. J., Isom, O. W., Rosengart, T., Falck-Pedersen, E., Hackett, N. R., Crystal, R. G., and Mastrangeli, A. (1997). Circumvention of anti-adenovirus neutralizing immunity by administration of an adenoviral vector of an alternate serotype. Hum. Gene Ther. 8, 99-109. 124. Mastrangeli, A., Harvey, B. G., Yao, J., Wolff, G., Kovesdi, L, Crystal, R. G., and Falck- Pedersen, E. (1996). "Sero-switch "adenovirus-mediated in vivo gene transfer: Circumvention of anti-adenovirus humoral immune defenses against repeat adenovirus vector administration by changing the adenovirus serotype. Hum. Gene Ther. 7, 79-87. 125. Beer, S. J., Matthev^s, C. B., Stein, C. S., Ross, B. D., Hilfinger, J. M., and Davidson, B. L. (1998). Poly (lactic-glycolic) acid copolymer encapsulation of recombinant adenovirus reduces immunogenicity in vivo. Gene Ther. 5, 740-746. 126. Chillon, M., Lee, J. H., Fasbender, A., and Welsh, M. J. (1998). Adenovirus complexed with polyethylene glycol and cationic lipid is shielded from neutralizing antibodies in vitro. Gene Ther. 5,995-1002. 127. O'Riordan, C. R., Lachapelle, A., Delgado, C , Parkes, V., Wadsworth, S. C , Smith, A. E., and Francis, G. E. (1999). PEGylation of adenovirus with retention of infectivity and protection from neutralizing antibody in vitro and in vivo. Hum. Gene Ther. 10,1349-1358. 128. Bewley, M. C , Springer, K., Zhang, Y. B., Freimuth, P., and Flanagan, J. M. (1999). Structural analysis of the mechanism of adenovirus binding to its human cellular receptor, CAR. Science 286, 1579-83. 129. Nemerow, G. R., and Stewart, P. L. (1999). Role of alpha(v) integrins in adenovirus cell entry and gene delivery. Microbiol. Mol. Biol. Rev. 63, 725-734. 130. Henderson, D. R., Friedman, S. B., Harris, J. D., Manning, W. B., and Zoccoh, M. A. (1986). CEDIA, a new homogeneous immunoassay system. Clin. Chem. 32, 1637-1641. 131. Bramson, J. L., Hitt, M., Gauldie, J., and Graham, F. L. (1997). Pre-existing immunity to adenovirus does not prevent tumor regression following intratumoral administration of a vector expressing IL-12 but inhibits virus dissemination. Gene Ther. 4, 1069-1076. 132. Li, Z., Rakkar, A., Katayose, Y., Kim, M., Shanmugam, N., Srivastava, S., Moul, J. W., McLeod, D. G., Cowan, K. H., and Seth, P. (1998). Efficacy of multiple administrations of 10. Development of ARCAs for the Treatment of Prostate Cancer 3 2 7 a recombinant adenovirus expressing wild-type p53 in an immune-competent mouse tumor modd.Gene Ther. 5, 605-613. 133. Smith, J. G., Raper, S. E., Wheeldon, E. B., Hackney, D., Judy, K., Wilson, J. M., and Eck, S. L. (1997). Intracranial administration of adenovirus expressing HSV-TK in combination v^ith ganciclovir produces a dose-dependent, self-limiting inflammatory response. Hum. Gene Ther. 8 ,943-954. 134. Sterman, D. H., Treat, J., Litzky, L. A., Amin, K. M., Coonrod, L., Molnar-Kimber, K., Recio, A., Knox, L., Wilson, J. M., Albelda, S. M., and Kaiser, L. R. (1998). Adenovirus- mediated herpes simplex virus thymidine kinase/ganciclovir gene therapy in patients with localized malignancy: Results of a phase I clinical trial in malignant mesothelioma. Hum. Gene Ther. 9, 10H3-1092. 135. Yang, Y., Li, Q., Ertl, H. C , and Wilson, J. M. (1995). Cellular and humoral immune responses to viral antigens create barriers to lung-directed gene therapy with recombinant adenoviruses./. Virol. 69,2004-2015. 136. Muruve, D. A., Barnes, M. J., Stillman, I. E., and Libermann, T. A. (1999). Adenoviral gene therapy leads to rapid induction of multiple chemokines and acute neutrophil-dependent hepatic injury in vivo. Hum. Gene Ther. 10, 965-976. 137. Muruve, D. A., Nicolson, A. G., Manfro, R. C., Strom, T. B., Sukhatme, V. P., and Liber mann, T. A. (1997). Adenovirus-mediated expression of Fas ligand induces hepatic apoptosis after Systemic administration and apoptosis of ex vivo-infected pancreatic islet allografts and isografts. Hum. Gene Ther. 8, 955-963. 138. Postlethwaite, R. (1979). Liver damage induced in mice by human adenovirus type 5. Scot. Med. J. 18,131. 139. Worgall, S., Wolff, G., Falck-Pedersen, E., and Crystal, R. G. (1997). Innate immune mechanisms dominate elimination of adenoviral vectors following in vivo administration. Hum. Gene Ther. 8, 37-44. 140. Dau, P. C. (1995). Immunologic rebound./. Clin. Apheresise 10, 210-217. 141. Dau, P. C , Callahan, J., Parker, R., and Golbus, J. (1991). Immunologic effects of plasma pheresis synchronized with pulse cyclophosphamide in systemic lupus erythematosus. / . Rheumatol. 18, 270-276. 142. Gahery-Segard, H., Farace, F., Godfrin, D., Gaston, J., Lengagne, R., Tursz, T., Boulanger, P., and Guillet, J. G. (1998). Immune response to recombinant capsid proteins of adenovirus in humans: Antifiber and anti-penton base antibodies have a synergistic effect on neutralizing activity./. Virol. 72,2388-2397. 143. Wohlfart, C. (1988). Neutralization of adenoviruses: Kinetics, stoichiometry, and mecha nisms. / . Virol. 62, 2321-2328. 144. Ganly, I. (1997). Phase I trial of intratumoral injection with an ElB-attenuated adenovirus, ONYX-015, in patients with recurrent p53(-) head and neck cancer. Proc. Am. Soc. Clin. Oncol. 16, 382a. 145. Kirn, D. H., and McCormick, F. (1996). Replicating viruses as selective cancer therapeutics. Mol. Med. Today 2, 519-527. 146. Khuri, F. R., Nemunaitis, J., Ganly, I., Arseneau, J., Tannock, I. F., Romel, L., Gore, M., Ironside, J., MacDougall, R. H., Heise, C , Randlev, B., Gillenwater, A. M., Bruso, P., Kaye, S. B., Hong, W. K., and Kirn, D. H. (2000). a controlled trial of intratumoral ONYX-015, a selectively-replicating adenovirus, in combination with cisplatin and 5-fluorouracil in patients with recurrent head and neck cancer. Nat. Med. 6, 879-885. 147. Ganly, I., Eckhardt, S. G., Rodriguez, G. I., Soutar, D. S., Otto, R., Robertson, A. G., Park, O., Gulley, M. L., Heise, C , Von Hoff, D.D., and Kaye, S.B. (2000). A phase I study of Onyx-015, an ElB attenuated adenovirus, administered intratumorally to patients with recurrent head and neck cancer. Clin. Cancer Res. 6, 798-806. 3 2 8 Henderson and Yu 148. van Elsacker-Niele, A. M., and Kroes, A. C. (1999). Human parvovirus B19: relevance in internal medicine. Neth. J. Med. 54, 221-230. 149. Segovia, J. C , Gallego, J. M., Bueren, J. A., and Almendral, J. M. (1999). Severe leukopenia and dysregulated erythropoiesis in SCID mice persistently infected v^ith the parvovirus minute virus of mice./. Virol. 73, 1774-1784. 150. Kelly, W. K., Scher, H. I., Mazumdar, M., Vlamis, V., Schwartz, M., and Fossa, S. D. (1993). Prostate-specific antigen as a measure of disease outcome in metastatic hormone-refractory prostate cancer./. Clin. Oncol. 11, 607-615. 151. Smith, D. C , Dunn, R. L., Strav^derman, M. S., and Pienta, K.J. (1998). Change in serum prostate-specific antigen as a marker of response to cytotoxic therapy for hormone-refractory prostate cancer./. Clin. Oncol. 16, 1835-1843. 152. Pollack, A., Lankford, S., Zagars, G. K., and Babaian, R. J. (1996). Prostate specific antigen density as a prognostic factor for patients v^ith prostate carcinoma treated v^ith radiotherapy. Cancer 77, 1515-1523. 153. Pound, C. R., Partin, A. W., Eisenberger, M. A., Chan, D. W., Pearson, J. D., and Walsh, P. C. (1999). Natural history of progression after PSA elevation follov^ing radical prostatec tomy./A MA 281, 1591-1597. 154. Herschman, J. D., Smith, D. S., and Catalona, W. J. (1997). Effect of ejaculation on serum total and free prostate-specific antigen concentrations. Urology 50, 239-243. 155. Ornstein, D. K., Rao, G. S., Smith, D. S., Ratliff, T. L., Basler, J. W., and Catalona, W.J. (1997). Effect of digital rectal examination and needle biopsy on serum total and percentage of free prostate specific antigen levels./. Urol. 157, 195-198. 156. Jain, R. K. (1990). Physiological barriers to delivery of monoclonal antibodies and other macromolecules in tumors. Cancer Res. 50, 814s-819s. 157. Cunningham, T. (1997). Pathogenesis of viral infection. In "Antiviral Agents and Human Viral Diseases" (J. Glasso, R. J. Whitley, and T. C. Merigan, Eds.), pp. 45-78 . Lippincott- Raven, Philadelphia. 158. Gutekunst, R. R., White, R. J., Edmondson, W. P., and Chanock, R. M. (1967). Immuniza tion with live type 4 adenovirus: Determination of infectious virus dose and protective effect of enteric infection. Am. J. Epidemiol. 86, 341-349. 159. Russi, T. J., Hirschowitz, E. A., and Crystal, R. G. (1997). Delayed-type hypersensitivity response to high doses of adenoviral vectors. Hum. Gene Ther. 8, 323-330. 160. Chengalvala, M. V., Bhat, B. M., Bhat, R. A., Dheer, S. K., Lubeck, M. D., Purcell, R. H., and Murthy, K. K. (1997). Replication and immunogenicity of Ad7-, Ad4-, and Ad5-hepatitis B virus surface antigen recombinants, with or without a portion of E3 region, in chimpanzees. Vaccine 15, 335-339. 161. Kass-Eisler, A., Leinwand, L., Gall, J., Bloom, B., and Falck-Pedersen, E. (1996). Cir cumventing the immune response to adenovirus-mediated gene therapy. Gene Ther. 3, 154-162. C H A P T E R Replicatiofi'Selective Oncolytic Adenovirus El-Region Mutants: Virotherapy for Cancer David Kirn Imperial Cancer Research Fund Program for Viral and Genetic Therapy of Cancer Hammersmith Hospital Imperial College School of Medicine London, United Kingdom I. Introduction Chemotherapy for metastatic sohd tumors generally fails due to an insuf- fiicient therapeutic index and/or insufficient antitumoral potency. Standard agents target a variety of different structures within cancer cells, but almost all of them are thought to kill cancer cells through the induction of apoptosis. As a result, apoptosis-resistant clones develop follov^ing standard therapies, even if numerous high-dose chemotherapeutic agents are used in combination. New treatment approaches should therefore have not only greater potency and greater selectivity than currently available treatments, they should also have novel mechanisms of action that will not be subject to cross-resistance with standard treatments (i.e., efficacy should not be dependent on apoptosis induction in cancer cells exclusively). Replication-selective oncolytic adenoviruses appear to have these char acteristics. These viruses have evolved to infect cells, replicate, induce cell death, release viral particles, and finally spread in human tissues. Replication in tumor tissue leads to amplification of the input dose, while a block in replication in normal tissues can lead to efficient clearance and reduced toxic ity (Fig. 1). Selective replication within tumor tissue can theoretically increase the therapeutic index of these agents dramatically. In addition, viruses kill cells by a number of unique mechanisms. In addition to direct lysis at the conclusion of the repUcative cycle, viruses can kill cells through expression of ADENOVIRAL VECTORS FOR GENE THERAPY 3 2 9 Copyright 2002, Elsevier Science (USA). All rights reserved. 330 David Kirn toxic proteins, induction of both inflammatory cytokines and T-cell-mediated immunity, and enhancement of cellular-sensitivity to their effects. Therefore, cross-resistance with standard chemotherapeutics or radiotherapy should be much less likely. Remarkable advances in molecular biology and genetics have led to a new understanding of both (1) the replication and pathogenicity of viruses and (2) carcinogenesis. The resulting techniques and knowledge have
allowed novel agents to be engineered to enhance their safety and/or their antitumoral potency. Genetically engineered viruses in development over the past decade have included adenoviruses, herpesviruses, and vaccinia. Viruses with inher ent tumor-selectivity have been evaluated and include reovirus, autonomous parvoviruses, Newcastle disease virus, measles virus strains, and vesicular stomatitis virus [1]. Each of these agents has shown tumor selectivity in vitro and/or in vivo^ and efficacy has been demonstrated in murine tumor mod els with many of these agents following intratumoral, intraperitoneal, and/or intravenous routes of administration. Preclinical data reported with these agents has been encouraging, but a number of important questions have awaited results from clinical trials. Viral agents like adenovirus have complex biologies, potentially including NORMAL cell viral replication blocked CANCER cell cancer-specific viral replication genetic alteration Figure 1 Schematic representation of tumor-selective viral replication and cell killing (A) and tumor-selective tissue necrosis (B). Reprinted with permission from Journal of Clinical Investigation. 1 1 . Replication-Selective Oncolytic Adenovirus E1 -Region Mutants 331 Normal Figure 1 (continued) species-specific interactions with host cell machinery and/or immune response effectors [2, 3]. Antitumoral efficacy and safety studies with these viruses have generally been performed in rodent or primate models. All published animal tumor model data with replication-selective adenoviruses has come from immunodeficient mouse-human tumor xenograft models [4-6]. Data from cancer patients has therefore been eagerly awaited. Now, after more than 5 years of clinical research with dll520 (Onyx-015 or CI-1042, Pfizer Corp., Groton, CT), approximately 15 clinical trials have been completed and recently analyzed involving approximately 250 patients. This chapter reviews the discovery and development of replication- selective oncolytic adenoviruses, with an emphasis on recently acquired data from phase I and II clinical trials with J/1520. The goal will be to summarize (1) the genetic targets and mechanisms of selectivity for these agents; (2) clinical trial data and what it has taught us to date about the promise but also the potential hurdles to be overcome with this approach; (3) future approaches to overcome these hurdles. 3 3 2 David Kirn II. Attributes of Replication-Selective Adenoviruses for Cancer Treatment A number of efficacy, safety, and manufacturing issues need to be assessed when considering a virus species for development as an oncolytic therapy [1]. First, by definition the virus must replicate in and destroy human tumor cells. An understanding of the genes modulating infection, replication or pathogenesis is necessary for rational engineering of the virus. Since most solid human turnors have relatively low growth fractions, the virus should infect both cycling and noncycling cells. In addition, receptors for viral entry must be expressed on the target tumor(s) in patients [7]. From a safety standpoint, the parental wild- type virus should ideally cause only mild, well-characterized human disease(s). Nonintegrating viruses have potential safety advantages as well. A genetically stable virus is desirable from both safety and manufacturing standpoints. Finally, the virus must be amenable to high-titer production and purification under Good Manufacturing Practices (GMP) guidelines for clinical studies. Human adenoviruses have these characteristics and are therefore excellent oncolytic virus candidates [8]. IIL Biology of Human Adenovirus Adenovirus biology is reviewed in detail elsewhere [9]. Roughly 50 dif ferent serotypes of human adenovirus have been discovered; the two most commonly studied are types 2 and 5 (group C). Adenoviruses have lin ear, double-stranded DNA genomes of approximately 38 kb. The capsid is nonenveloped and is composed of the structural proteins hexon, fiber (binds Coxsackie and adenovirus receptor (CAR)), and penton (binds ay^s^s integrins for virus internalization). The adenovirus life-cycle includes the following steps: (1) virus entry into the cell following CAR and integrin bind ing, (2) release from the endosome and subsequent entry into the nucleus, (3) expression of early region gene products, (4) cell entry into S-phase, (5) prevention of p53-dependent and -independent apoptosis, (6) shut-off of host cell protein synthesis, (7) viral DNA replication, (8) viral structural protein synthesis, (9) virion assembly in the nucleus, (10) cell death, and (11) virus release. The E3 region encodes a number of gene products respon sible for immune response evasion [10, 11]. The gp 19-kDa protein inhibits MHC-class I expression on the cell surface (i.e. avoidance of cytotoxic T- lymphocyte-mediated killing) [12], and the E3 10.4/14.5-kDa (RID complex) and 14.7-kDa proteins inhibit apoptosis mediated by FasL or tumor necrosis factor (TNF) [11, 13]. 1 1 . Replication-Selective Oncolytic Adenovirus El-Region Mutants 3 3 3 IV. Mechanisms of Adenovirus-Mediated Cell Killing Adenovirus replication within a target tumor cell can lead to cell destruc tion by several mechanisms (Table I). Viral proteins expressed late in the course of infection are directly cytotoxic, including the E3 11.6-kDa adenovirus death protein [14] and E40RF4 [Branton, 1999 #1920]. Deletion of these gene products results in a significant delay in cell death. In addition, El A expression early during the adenovirus life cycle induces cell sensitivity to TNF-mediated killing [15]. This effect is inhibited by the E3 proteins 10.4/14.5 and 14.7; deletion of these E3 proteins leads to an increase in TNF expression in vivo and enhanced cell sensitivity to TNF [3]. Finally, viral replication in and lysis of tumor cells has been show^n to promote the induction of cell-mediated immunity to uninfected tumor cells in model systems w îth other viruses [16, 17]; v^hether this v îll occur in patients and v^ith adenovirus remains to be determined. V. Approaches to Optimizing Tumor-Selective Adenovirus Replication Two broad approaches are currently being used to engineer tumor- selective adenovirus replication. One is to limit the expression of the ElA Table I Pofential Mechanisms of Antitumoral Efficacy v^ith Replication-Selective Adenoviruses Mechanism Examples of adenoviral genes modulating effect I. Direct cytotoxicity due to viral proteins • E3 11.6 kDa • E40RF4 II. Augmentation of antitumoral immunity CTL infiltration, killing • E3 gp 19 kDa^ Tumor cell death, antigen release • E3 11.6 kDa Immunostimulatory cytokine induction • E3 10.4/14.5, 14.7 kDa^ Antitumoral cytokine induction (e.g., TNF) • E3 10.4/14.5, 14.7 kDa^ Enhanced sensitivity to cytokines (e.g., TNF) • ElA Unknov^n (.̂ ElA, others) III. Sensitization to chemotherapy IV. Expression of exogenous therapeutic genes Note. CTL, cytotoxic T-lymphocyte; TNF, tumor necrosis factor; NA, not applicable. ^Viral protein inhibits antitumoral mechanism. 3 3 4 David Kirn gene product to tumor tissues through the use of tumor- and/or tissue-specific promoters. ElA functions to stimulate S-phase entry and to transactivate both viral and cellular genes that are critical for a productive viral infection [18]. A second broad approach to optimizing tumor selectivity is to delete gene functions that are critical for efficient viral replication in normal cells but are expendable in tumor cells (described below^). Tissue- or tumor-specific promoters can replace endogenous viral sequen ces in order to restrict viral replication to a particular target tissue. For example, the prostate-specific antigen (PSA) promoter/enhancer element has been inserted upstream of the ElA gene; the result is that viral replication correlates with the level of PSA expression in a given cell [4]. This virus, CN706 (Calydon Pharmaceuticals, CA), is currently in a Phase I clinical trial of intratumoral injection for patients w îth locally recurrent prostate carcinoma. A second prostate-specific enhancer sequence has subsequently been inserted upstream of the ElB region [19]; the use of these two prostate-specific enhancer elements to drive separate early gene regions has led to improved selectivity over the first-generation virus [19]. A similar approach has been pursued by other groups using tissue-specific promoters to drive ElA expression selectively in specific carcinomas (e.g., alpha-fetoprotein, carcinoembryonic antigen, MUC- 1)[20,21]. A second general approach is to complement loss-of-function mutations in cancers with loss-of-function mutations within the adenovirus genome. Many of the same critical regulatory proteins that are inactivated by viral gene products during adenovirus replication are also inactivated during car cinogenesis [22-25]. Because of this convergence, the deletion of viral genes that inactivate these cellular regulatory proteins can be complemented by genetic inactivation of these proteins within cancer cells [26, 78]. The deletion approach was first described by Martuza et al. with herpesviruses; the thymi dine kinase gene (dlsptk) [27] and subsequently the ribonucleotide reductase gene (G207) were deleted [28]. Two adenovirus deletion mutation approaches have subsequently been described (see below). VI. El A-CR2 Region Deletion Mutants Mutants in the ElA conserved region 2 (CR2) are defective in pRB bind ing [29, 30] (Fig. 2). These viruses are being evaluated for use against tumors with pRB pathway abnormalities (e.g., loss of the Gl-S checkpoint) [26, 31, 32]. The delta-24 E1A-CR2 mutant virus was able to efficiently replicate in tumor cell lines lacking functional pRB, while replication was significantly inhibited by reintroduction of wild-type RB protein into a tumor cell line lack ing functional pRB; both in vitro and in vivo efficacy were demonstrated [32]. With J/922/947, a very similar E1A-CR2 mutant, S-phase induction and viral 1 1 . Replication-Selective Oncolytic Adenovirus El-Region Mutants 3 3 5 Adenovirus El A 289R •Exonl 11 Unique •Exon2- 243R 139 186 289 \| C R 1 CE2 oo 40 eo 120 188 pRb family-binding regions pRB 30 60 121 127 pl30 30 60 121 139 pl07 30 60 124127 ^922-947 92:?/947 i 122 129 Figure 2 Diagram of the structure of the adenovirus El A RNA (12S, 13S) and pRB protein family-binding regions showing deletion in c//922-947 mutant adenovirus. replication are significantly inhibited in quiescent normal cells, whereas repli cation and cytopathic effects proceed efficiently in tumor cells; interestingly, dl922/947 demonstrates significantly greater potency than dllSlO both in vitro and in vivo [26, 31], and in a nude mouse-human tumor xenograft model, intravenously administered d/922/947 had significantly superior efficacy to even wild-type adenovirus [31]. The El A mutant adenoviruses described by these two groups may in fact behave very similarly, although to date they have been tested in different fashions. Unlike the complete deletion of ElB 55 kDa in dll520, these mutations in El A are targeted to a single conserved region and may therefore leave intact other important functions of the gene product; therefore, viral potency is not attenuated. VII. E1B 55-kDa Gene Deletion Mutant: d/1520 J/1520 (Onyx-015) was the first adenovirus described to mirror the gene deletion approach pioneered by Martuza with herpesvirus. McCormick et al. 336 David Kirn hypothesized that an adenovirus with deletion of a gene encoding a p53- binding protein, ElB 55-kDa, would be selective for tumors that already had inhibited or lost p53 function. p53 function is lost in the majority of human cancers through mechanisms including gene mutation, overexpression of p53- binding inhibitors (e.g., mdm2, human papillomavirus E6), and loss of the p53-inhibitory pathway modulated by pl4^^^ [33-35]. However, the precise role of p53 in the inhibition of adenoviral replication has not been defined to date. In addition, other adenoviral proteins also have direct or indirect effects on p53 function (e.g., E4orf6, ElB 19-kDa, ElA) [36]. Finally, ElB 55-kDa itself has important viral functions that are unrelated to p53 inhibition (e.g., viral mRNA transport, host cell protein synthesis shut-off) [37] (Fig. 3). Not surprisingly, therefore, the role of p53 in the replication-selectivity of dll520 has been difficult to confirm despite extensive in vitro experimentation by many groups. ElB 55-kDa gene deletion was associated with decreased replication and cytopathogenicity in p53(-h) tumor cells versus matched p53(-) tumor cells, relative to wild-type adenovirus, in RKO and H1299 cells [38-40]. However, conflicting data on the role of p53 in modulating J/1520 replication and/or cytopathic effect (CPE) has come from different cell systems; no p53 effect was demonstrated in matched U20S cells, for example [40]. Although p53 can clearly inhibit J/1520 in many cell systems, it has become clear that many other cellular factors independent of p53 play critical roles in determining E1B49K E^ltS K >S>1 ^ ^ \m^- Figure 3 Diagram of both p53 pathway interactions with adenoviral gene products and functions of El B 55 kDa: complexity of cancer cell and adenoviral biology. (A) Note that adenoviral proteins (light gray) target multiple components of this pathway at sites upstream of p53, downstream of p53 and at the level of p53 itself. Examples of p53-regulated cell functions are shown (black boxes). In addition, the known functions of ElB 55 kDa are shown (dark gray). (B) Graphic representation of the consequences of ElB 55-kDa gene deletion. In addition to the loss of p53 binding when ElB 55 kDa was deleted in c//1520 (Onyx-015), other important viral functions
are also lost. 1 1 . Replication-Selective Oncolytic Adenovirus E1 -Region Mutants 337 Figure 3 (continued) the sensitivity of cells to dll520 [39, 41-44] . Clinical trials were ultimately necessary to determine the selectivity and clinical utility of J/1520 (see below). Clinical trial data confirmed the tumor-selectivity of dll520. VIII. Clinical Trial Results with Replication-Competent Adenoviruses in Cancer Patients A. Clinical Trial Results with Wild-Type Adenovirus Over the past century a diverse array of viruses were injected into cancer patients by various routes, including adenovirus, Bunyamwara, Coxsackie, dengue, feline panleukemia, Ilheus, mumps, Newcastle disease virus, vaccinia, and West Nile [1, 45-47]. These studies illustrated both the promise and the hurdles to overcome with oncolytic viral therapy. Unfortunately, these previous clinical studies were not performed to current clinical research standards, and therefore none give interpretable and definitive results. At best, these studies are useful in generating hypotheses that can be tested in future trials. Although suffering from many of the trial design flaws listed below, a trial with wild-type adenovirus is one of the most useful for hypothesis generation but also for illustrating how clinical trial design flaws severely curtail the utility of the study results. The knowledge that adenoviruses could eradicate a variety of tumor cells in vitro led to a clinical trial in the 1950s with wild-type adenovirus. Ten different serotypes were used to 3 3 8 David Kirn treat 30 cervical cancer patients [47]. Forty total treatments were administered by either direct intratumoral injection {n = 23), injection into the artery perfusing the tumor {n = 10), treatment by both routes (n = 6), or intravenous administration (n = 1). Characterization of the material injected into patients was minimal. The volume of viral supernatant injected is reported, but actual viral titers/doses are not; injection volumes (and by extension doses) varied greatly. When possible, the patients were treated with a serotype to which they had no neutralizing antibodies present. Corticosteroids were administered as nonspecific immunosuppressive agents in roughly half of the cases. Therefore, no two patients were treated in identical fashion. Nevertheless, the results are intriguing. No significant local or systemic toxicity was reported. This relative safety is notable given the lack of preexisting immunity to the serotype used and concomitant corticosteroid use in many patients. Some patients reported a relatively mild viral syndrome lasting 2 - 7 days (severity not defined); this viral syndrome resolved spontaneously. Infectious adenovirus was recovered from the tumor in two-thirds of the patients for up to 17 days postinoculation. Two-thirds of the patients had a "marked to moderate local tumor response" with necrosis and ulceration of the tumor (definition of "response" not reported). None of the seven control patients treated with either virus-free tissue culture fluid or heat-inactivated virus had a local tumor response (sta tistical significance not reported). Therefore, clinically evident tumor necrosis was only reported with viable virus. Neutralizing antibodies increased within 7 days after administration. Although the clinical benefit to these patients is unclear, and all patients eventually had tumor progression and died, this study did demonstrate that wild-type adenoviruses can be safely administered to patients and that these viruses can replicate and cause necrosis in sofid tumors despite a humoral immune response. The maximally tolerated dose, dose-limiting toxicity, objective response rate, and time to tumor progres sion, however, remain unknown for any of these serotypes by any route of administration. B. A Novel Staged Approach to Clinical Research with Replication-Selective Viruses: The Example of c//1520(Onyx-015) For the first time since viruses were first conceived as agents to treat cancer over a century ago, we now have definitive data from numerous phase I and II clinical trials with a well-characterized and well-quantitated virus. J/1520 (Onyx-015, a.k.a., CI-1042, Pfizer, Inc.) is a novel agent with a novel mechanism of action. This virus was to become the first virus to be used in humans that had been genetically engineered for replication selectivity. We predicted that both toxicity and efficacy would be dependent on multiple 1 1 . Replication-Selective Oncolytic Adenov i rus E1 -Region Mutan ts 339 Staged Development approach: Replication-selective agents for cancer Intra tumoral Intra peritoneal /w^ra-arterial/hepatic artery Intra venous Figure 4 Staged clinical research and development approach used in research and development with c//1520 (Onyx-015). Once safety and biological activity was demonstrated by the intratumoral route, clinical trials were initiated sequentially to study intracavitary instillation (initially intraperi toneal), intraarterial infusion (hepatic artery) and eventually intravenous administration. Only patients with advanced and incurable cancers were enrolled on trials initially. Once safety was demonstrated in these patients, trials were initiated in patients with premalignant lesions. Finally, clinical trials of combinations with chemotherapy were initiated only after the safety of c//1520 as a single agent had been documented by the relevant route of administration. Reprinted with permission from Gene Therapy. factors including (1) the inherent abihty of a given tumor to rephcate and shed the virus, (2) the location of the tumor to be treated (e.g., intracranial vs peripheral), and (3) the route of administration of the virus. In addition, we felt it w^ould be critical to obtain biological data on viral replication, antiviral immune responses and their relationship to antitumoral efficacy in the earliest phases of clinical research. We therefore designed and implemented a novel staged clinical research and development approach with this virus (Fig. 4). The goal of this approach was to sequentially increase systemic exposure to the virus only after safety with more localized delivery had been demonstrated. Following demonstration of safety and biological activity by the intratumoral route, trials were sequen tially initiated to study intracavitary instillation (initially intraperitoneal), intraarterial infusion (initially hepatic artery), and eventually intravenous administration. In addition, only patients with advanced and incurable cancers were initially enrolled on trials. Only after safety had been demonstrated in terminal cancer patients were trials initiated for patients with premalignant conditions. Finally, clinical trials of combinations with chemotherapy were initiated only after the safety of dllSlO as a single agent had been documented by the relevant route of administration. 3 4 0 David Kim IX. Results from Clinical Trials with d/1520 (Onyx-015orCI -1042) A. Toxicity No maximally tolerated dose or dose-limiting toxicities were identified at doses up to 2x10^^ particles administered by intratumoral injection. Flu like symptoms and injection-site pain were the most common associated toxicities [48]. This safety is remarkable given the daily or even twice-daily dosing that was repeated every 1-3 weeks in the head and neck region or pancreas [49]. Intraperitoneal, intraarterial, and intravenous administration were also remarkably well tolerated in general. Intraperitoneal administration was feasi ble at doses up to 10^^ particles divided over 5 days [50]. The most common toxicites included fever, abdominal pain, nausea/vomiting, and bowel motil ity changes (diarrhea, constipation). The severity of the symptoms appeared to correlate with tumor burden. Patients with heavy tumor burdens reached a maximally tolerated dose at 10^^ particles (dose-limiting toxicities were abdominal pain and diarrhea), whereas patients with a low tumor burden tolerated 10^^ without significant toxicity. No dose-limiting toxicities were reported following repeated intravascu lar injection at doses up to 2 x 10^^ particles (hepatic artery) [51] or 2 x 10^^ particles (intravenous) [52]. Fever, chills, and asthenia following intravascular injection were more common and more severe than after intratumoral injec tions (grade 2 - 3 fever and chills vs grade 1). Dose-related transaminitis was reported infrequently. The transaminitis was typically transient (<10 days) and low-grade (grade 1-2) and was not clinically relevant. Further dose escalation was limited by supply of the virus. B. Viral Replication Viral replication was documented at early time points after intratumoral injection in head and neck cancer patients [49, 53]. Roughly 70% of patients had evidence of replication on days 1-3 after their last treatment (Table II). In contrast, day 14-17 samples were uniformly negative. Patients with injected pancreatic tumors, in contrast, showed no evidence of viral replication by plasma PCR (indirect evidence) or fine-needle aspiration. Similarly, intraperi toneal dllSlO could not be shown to reproducibly infect ovarian carcinoma cells within the peritoneum. Therefore, different tumor types can vary dramat ically in their permissiveness for viral infection and replication (Table II). Proof-of-concept for tumor infection following intraarterial [51] or intra venous [52] administration with human adenovirus has also been achieved. Approximately half of the roughly 25 patients receiving hepatic artery infusions c vH O C! o o bJD o o o ^ ^ O o o O _fi - - - vx) ^-^ ON oh^- or o^ ON rsl (N ^ \o 'o ON rH oo -̂' o ^ o —~ -̂̂ bJD ^ ON O o ^ O O " ^ Z o o 1 —̂1 O ^ P O -T3 O CD c Q 13 O x' ^2 nJ (>U (L) 3 > o ^:i> u oo ^ O ^ OJ OH ^ oo c " d _> o '35 oo c:i r o Q Q ^r n oCî •A ?^ Q 0) Q. O ^ o o t \ uo — O ^^ -^ o s I c^ "Ĵ -^ r-. ^ ^ lO (N O o ro O K £ I I o 3 n3 ^ o O H £ 0) UO <!-> !/5 = o*J o O IJ O <u O 00 ro O iW r̂ " ^ ^ (L) C -T3 3 0 3 T 3 - c/5 J ^ 'bi) ^ 'bJD ^ o cs o ( ^ -̂ "bJD ^ " •S >̂ ^ Lo O o ^ X X rsl 'o O I I 1 00 X O "o O ^ 2 c Q ^ o X 0) o 1 u (U CT3 t-l .sC/3 ^ o 0 B c r ^ n c _c ^ "o H OS 2 'o (/3 (/3 B C3 OS ,0 o 0> O o 1C3U >̂,̂̂ u "5- C O CT3 0 .'u2 • M V-l ^ -I > OS 341 3 4 2 David Kim of 2 X 10^^ particles were positive by PCR 3-5 following treatment. Since input virus genomes are cleared to undetectable levels within 6-12 h, these late recurrent peaks of viral genomes are highly suggestive of viral replication and shedding into the blood. Three of four patients with metastatic carcinoma to the lung treated intravenously with >2 x 10^^ particles were positive for genomes in the blood on day 3 (±1). Therefore, it appears to be feasible to infect distant tumor nodules following intravenous or intra-arterial administration. C. Immune Response Neutralizing antibody titers to the coat (Ad5) of J/1520 were positive but relatively low in roughly 50-60% of all clinical trial patients at baseline [49]. Antibody titers increased uniformly following administration of J/1520 by any of the routes tested, in some cases to levels > 1:80,000. Antibody increases occurred regardless of evidence for replication or shedding into the blood stream [49]. Acute inflammatory cytokine levels were determined prior to treatment (by hepatic artery infusion), 3 h posttreatment and 18 h posttreat- ment: IL-1, IL-6, IL-10, interferon-gamma, tumor necrosis factor. Significant increases were demonstrated within 3 h for IL-1, IL-6, tumor necrosis fac tor, and to a lesser extent interferon-gamma; all cytokines were back down to pretreatment levels by 18 h [54]. In contrast, IL-10 did not increase until 18 h. However, cytokine levels varied greatly from patient to patient and from treatment cycle to treatment cycle. D. Efficacy with cf/1520 (Onyx-Ol 5) as a Single Agent Two Phase II trials enrolled a total of 40 patients with recurrent head and neck cancer [49, 53]. Tumors were treated very aggressively with 6-8 daily needle passes for 5 consecutive days (30-40 needle passes per 5-day cycle; n = 30) and 10-15 per day on a second trial (50-75 needle passes per cycle; n = 10). The median tumor volume on these studies was approximately 25 cm^; an average^ cm^of tumor therefore received an estimated 4 - 5 nee dle passes per cycle. With this dose-intensive local treatment regimen, the unconfirmed response rate at the injected site was 14% and the confirmed local response rate was 7%. Interestingly, there was no correlation between evidence of antitumoral activity and neutralizing antibody levels at baseline or posttreatment [49]. No objective responses were demonstrated in patients with tumor types that could not be so aggressively injected (due to their deep locations), although some
evidence of shrinkage or necrosis was obtained. In summary, single agent responses across all studies were uncommon, and therefore combinations with chemotherapy were explored. E. Efficacy in Combination with Chemotherapy: Potential Synergy Discovered Evidence for a potentially synergistic interaction between adenoviral therapy and chemotherapy has been obtained on multiple trials. Encouraging 1 1 . Replication-Selective Oncolytic Adenovirus E1 -Region Mutants 3 4 3 clinical data has been obtained in patients with recurrent head and neck cancer treated with intratumoral J/1520 in combination with intravenous cispaltin and 5-fluorouracil [55]. Thirty-seven patients were treated and 19 responded (54%, intent-to-treat; 63%, evaluable); this compares favorably with response rates to chemotherapy alone in previous trials (30-40%, generally). The time-to-tumor progression was also superior to previously reported studies. However, comparisons to historical controls are unreliable. We therefore used patients as their own controls whenever possible (n = 11 patients). Patients with more than one tumor mass had a single tumor injected with dll520 while the other mass(es) was left uninjected. Since both masses were exposed to chemotherapy, the effect of the addition of viral therapy to chemotherapy could be assessed. The J/1520-injected tumors were significantly more likely to respond (? = 0.017) and less likely to progress (P = 0.06) than were nonin- jected tumors. Noninjected control tumors that progressed on chemotherapy alone were subsequently treated with Onyx-015 in some cases; two of the four injected tumors underwent complete regressions. This data illustrates the potential of viral and chemotherapy combinations. The clinical utility of J/1520 in this indication will be definitively determined in a phase III randomized trial. A phase I/II trial of J/1520 administered by hepatic artery infusion in combination with intravenous 5-fluorouracil and leukovorin was carried out {n = 33 total) [54]. Following phase I dose escalation, 15 patients with colorec tal carcinoma who had previously failed the same chemotherapy were treated with combination therapy after failing approximately to respond to J/1520 alone; 1 patient underwent a partial response (following initial progression on virus alone) and 10 had stable disease (2-7+ months). Combination virus plus chemotherapy-induced responses in colorectal liver metastases was therefore possible via hepatic artery infusions, although the magnitude and frequency of this effect remains to be determined. In addition, the optimal combination regimen has not yet been defined. In contrast, data from a phase I/II trial studying the combination of J/1520 and gemcitabine chemotherapy were dis appointing (n = 21); the combination resulted in only two responses, and these patients had not received prior gemcitabine [56]. Therefore, potential synergy was demonstrated with J/1520 and chemotherapy in two tumor types that supported presumed viral replication (head and neck, colorectal), but not in a tumor type that was apparently resistant to viral replication (pancreatic). X. Clinical Trial Results with c/f1520 (Onyx-015): Summary J/1520 was well-tolerated at the highest practical doses that could be administered (2 x 10^^-2 x 10^^ particles) by intratumoral, intraperitoneal, intraarterial, and intravenous routes. The lack of clinically significant toxicity in the liver or other organs was notable. Flu-like symptoms (fever, rigors. 3 4 4 David Kirn asthenia) were the most common toxicities and were increased in patients receiving intravascular treatment. Acute inflammatory cytokines (especially IL-1 and IL-6) increased within 3 h following intraarterial infusion, although these changes were extremely variable. Neutralizing antibodies increased in all patients, regardless of dose, route, or tumor type. Viral replication was documented directly (by biopsy) or indirectly (by delayed viral genome peaks in blood) in head and neck and colorectal tumors following intratumoral or intraarterial administration. Neutralizing antibodies did not block antitumoral activity in head and neck cancer trials of intratumoral injection. Single agent antitumoral activity was minimal (=15%) in head and neck cancers that could be repeatedly injected. No objective responses were documented with single agent therapy in phase I or I/II trials in patients with pancreatic, colorectal, or ovarian carcinomas. A favorable and potentially synergistic interaction with chemotherapy was discovered in some tumor types and by different routes of administration. XI . Future Directions: Why Has c//1520 (Onyx-015) Failed to Date as a Single Agent? Future improvements with this approach will be possible if the reasons for J/1520 failure as a single agent, and success in combination with chemotherapy, are elucidated. Factors that are specific to this particular adenoviral mutant, as well as factors that are generalizable to other adenoviruses, should be considered. Regarding this particular ElB 55-kDa gene mutant, it is important to remember that this virus is significantly attenuated relative to wild-type adenovirus in most tumor cell lines in vitro and in vivo, including even p53 mutant tumors [31, 39, 40, 43, 57]. This is an expected finding since this virus does not have critical ElB 55-kDa functions that are unrelated to p53, including viral mRNA transport and shut-off of host protein synthesis. This attenuated potency is not apparent with the adenovirus mutant J/922/947 [31]. The deletion in the E3 gene region of genes for the 10.4/14.5 complex is likely to make this virus more sensitive to the antiviral effects of tumor necrosis factor; an immunocompetent animal model will need to be identified in order to resolve this issue. Factors likely to be important for any adenovirus include barriers to intratumoral spread (e.g., fibrosis), antiviral immune responses and inadequate viral receptor expression (e.g., CAR, integrins). Viral coat modifications may be beneficial if inadequate CAR expression plays a role in the resistance of particular tumor types [58, 59]. 1 1 . Replication-Selective Oncolytic Adenovirus El-Region Mutants 3 4 5 XII . Improving the Efficacy of Replication-Selective Oncolytic Adenoviral Agents Alterations within the adenoviral genome can be used to enhance selectiv ity and/or potency. The promising adenoviral ElA CR-2 mutant (dl922/947) has been described that demonstrates not only tumor selectivity (based on the Gl-S checkpoint status of the cell) but also significantly greater antitumoral efficacy in vivo compared to dllSlO (all models tested) and even w^ild-type adenovirus (in a breast cancer metastasis model) [26]. A very similar ElA CR-2 mutant adenovirus has demonstrated replication and cytopathic effects based on the pRB status of the target cell, in addition to encouraging in vivo antitumoral efficacy [32]. Deletion of the ElB 19-kDa gene (antiapoptotic bcl-2 homolog) is knov^n to result in a "large plaque" phenotype due to enhanced speed of cell killing [60]. This observation has now^ been extended to multiple tumor cell lines and primary tumor cell cultures [61, 62]. A sim ilar phenotype resulted from overexpression of the E3 11.6-kDa adenovirus death protein [63]. It remains to be seen vs^hether these in vitro observa tions are foUow êd by evidence for improved efficacy in vivo over w^ild-type adenovirus. "Arming" viruses w îth therapeutic genes can also result in improved potency (e.g., prodrug-activating enzymes and cytokines) [64-67]. Prodrug- activating enzyme conversion of nontoxic prodrugs to active cytotoxic agents w^ithin the tumor is an attractive strategy that has been pursued; how^ever, this approach may result in virus inactivation and decreased viral oncolysis, in addition to beneficial "bystander effects." Viral coat modifications may be ben eficial if inadequate CAR expression plays a role in the resistance of particular tumor types [58, 59]. Improved systemic delivery may require novel formu lations or coat modifications, as v^ell as suppression of the humoral immune response. Determination of the viral genes (e.g., E3 region) and immune response parameters mediating efficacy and toxicity w îll lead to immunomod ulatory strategies. Finally, identification of the mechanisms leading to the potential synergy betw^een replicating adenoviral therapy and chemotherapy may allov\̂ augmentation of this interaction [68]. XIIL Summary Replication-selective oncolytic adenoviruses represent a novel cancer treatment platform. Clinical studies w îth J/1520 and now^ other viruses have demonstrated the safety and feasibility of this approach, including the delivery of adenovirus to tumors through the bloodstream [5, 51, 69]. The discovery of 3 4 6 David Kirn the inherent capacity of repHcation-competent adenoviruses to sensitize tumor cells to chemotherapy led to chemosensitization strategies. Clinical research is anticipated with novel adenoviral agents, including constructs expressing exogenous therapeutic genes to enhance both local and systemic antitumoral activity [8, 64, 70]. In addition to adenovirus, other viral species are being developed, including herpesvirus, vaccinia, reovirus and measles virus [1, 17, 27, 71-75]. Since intratumoral spread also appears to be a substantial hurdle for viral agents, inherently motile agents such as bacteria may hold great promise for this field [76^ 77]. Although data from in vitro cell-based assays and murine tumor model systems will be important for testing and generating hypotheses, it is vital that encouraging adenoviral agents are tested in well- designed clinical trials as soon as possible. Data from clinical trials must be used to guide future laboratory approaches, as well as the converse. This "iterative loop" between laboratory and clinic may result in major cancer treatment advances. Acknov^ledgments The following individuals have been instrumental in making this chapter possible: John Nemunaitis, Stan Kaye, Tony Reid, Fadlo Khuri, James Abruzzesse, Eva Galanis, Joseph Rubin, Antonio Grillo-Lopez, Carla Heise, Larry Romel, Chris Maack, Sherry Toney, Nick LeMoine, Britta Randlev, Patrick Trown, Fran Kahane, Frank McCormick, and Margaret Uprichard. References 1. Kirn, D. (2000a)./. Clin. Invest. 105, 836-838. 2. Wold, W. S., Hermiston, T. W., and Tollefson, A. E. (1994). Trends Microbiol. 2, 437-443. 3. Sparer, T. E., Tripp, R. A., Dillehay, D. L., Hermiston, T. W., Wold, W. S., and Gooding, L. R. (1996)./. Virol. 70, 2431-2439. 4. Rodriguez, R., Schuur, E. R., Lim, H. Y., Henderson, G. A., Simons, J. W., and Henderson, D. R. (1997). Cancer Res. 57, 2559-2563. 5. Heise, C., Williams, A., Xue, S., Propst, M., and Kirn, D. (1999b). Cancer Res. 59,2623-2628. 6. Heise, C., Williams, A., Olesch, J., and Kirn, D. (1999a). Cancer Gene Ther. 6. 7. Wickham, T. J., Segal, D. M., Roelvink, P. W., Carrion, M. E., Lizonova, A., Lee, G. M., and Kovesdi, I. (1996)./. Virol. 70, 6831-6838. 8. Heise, C , and Kirn, D. (2000)./. Clin. Invest. 105, 847-851. 9. Shenk, T. (1996). In "Fields Virology" (K. Howley, Ed.), pp. 2135-2137. Lippincott-Raven: Philadelphia. 10. Wold, W. S., Tollefson, A. E., and Hermiston, T. W. (1995). Curr. Top. Microbiol. Immunol. 199, 237-274. 11. Dimitrov, T., Krajcsi, P., Hermiston, T. W., Tollefson, A. E., Hannink, M., and Wold, W. S. (1997)./. Virol. 71,2830-2837. 12. Hermiston, T. W., Tripp, R. A., Sparer, T., Gooding, L. R., and Wold, W. S. (1993). / . Virol. 67, 5289-5298. 1 1 . Replication-Selective Oncolytic Adenovirus E1 -Region Mutants 3 4 7 13. Shisler, J., Duerksen, H. P., Hermiston, T. M., Wold, W. S., and Gooding, L. R. (1996). / . Virol. 70, 68-77. 14. Tollefson, A. E., Ryerse, J. S., Scaria, A., Hermiston, T. W., and Wold, W. S. (1996). Virology 220, 152-62. 15. Gooding, L. R. (1994). Infect. Agents Dis. 3, 106-115. 16. Toda, M., Rabkin, S., Kojima, H., and Martuza, R. (1999). Hum. Gene Ther. 10, 385-393. 17. Martuza, R. (2000)./. Clin. Invest. 105, 841-846. 18. Whyte, P., Ruley, H., and Harlow, E. (1988). / . Virol. 62, 257-265. 19. Yu, D., Sakamoto, G., and Henderson, D. R. (1999). Cancer Res. 59, 1498-1504. 20. Hallenbeck, P. Chang, Y., Hay, C., Phipps, S., and Chiang, Y. (1999). Hum. Gene Ther. 10(10), 1721-1733. 21. Kurihara, T., Brough, D. E., Kovesdi, I., and Kufe, D. W. (2000). / . Clin. Invest. 106, 763-771. 22. Barker, D. D., and Berk, A. J. (1987). Virology 156, 107-121. 23. Nielsch, U., Fognani, C , and Babiss, L. E. (1991). Oncogene 6, 1031-1036. 24. Sherr, C. J. (1996). Science 274, 1672-1677. 25. Olson, D. C , and Levine, A. J. (1994). Cell Growth Differ, 5, 61 -71 . 26. Kirn, D., Heise, C , Williams, M., Propst, M., and Hermiston, T. (1998a). "Cancer Gene Therapy" Sobol (Ed.). San Diego. 27. Martuza, R. L., Malick, A., Markert, J. M., Ruffner, K. L., and Coen, D. M. (1991). Science ISl, 854-856. 28. Mineta, T., Rabkin, S. D., Yazaki, T., Hunter, W. D., and Martuza, R. L. (1995). Nat. Med. 1, 938-943. 29. Whyte, P., Williamson, N., and Harlow, E. (1989). Cell 56, 67-75 . 30. Hu, Q. J., Dyson, N. and Harlow, E. (1990). EMBO J. 9, 1147-1155. 31. Heise, C , Hermiston, T., Johnson, L., Brooks, G., Sampson-Johannes, A., Williams, A., Hawkins, L., and Kirn, D. (2000a). Nat. Med. 6, 1134-1139. 32. Fueyo, J., Gomez-Manzano, C , Alemany, R., Lee, P., McDonnell, T., Mitlianga, P., Shi,
certain microbial surface structures (e.g., via the alternative pathway, a phylogenetically older system); (2) recognition of microbe-bound antibody by a complement component called Clq (e.g., via the classic pathway); and (3) direct cleavage of C3 by a protease associated with complexes of collectins bound to microbes (see below). Subsequent to 12. Innate Immune Responses to in Vivo Adenovirus Infection 3 5 3 recognition, complement activation can neutralize viruses and other microbes by several mechanisms, including direct killing by microbial membrane per foration, agglutination of the microbes v^ith a net loss in infectivity, and opsonization resulting in clearance by C3-receptor-bearing phagocytes. Acti vation of each of these three pathw^ays (classical, alternative, and collectin) has been demonstrated in response to viral infection. The soluble protein effector component of innate immunity also includes a heterogeneous collection of other members that serve opsonic and lytic roles in microbial clearance. Collectins are one such group present in blood and in tissues v^hose family members include conglutinin, mannose-binding protein (MBP), and the surfactant proteins A (SP-A) and D (SP-D) [5]. Members of this family share a common basic structure consisting of collagen-like and lectin domains separated by a short neck region. The lectin moiety binds to microbial surface carbohydrates in a calcium-dependent manner and thus, collectins share both structural and functional homology v^ith Clq. Although surfactant proteins found in the lung (SP-A, SP-B, SP-C, SP-D) v^ere initially regarded as functioning to maintain alveolar structural integrity and patency, emerging data has defined an important role for SP-A and SPD in innate immune mechanisms of lung host defense against bacterial and viral pathogens. Other soluble protein effectors include defensins (a family of proteins with bacteriocidal properties secreted from neutrophils and epithelial cells) lysozyme, elastase, cathepsin G, phospholipase A2, lactoferrin, and transferrin. At present, the potential role of these molecules in viral infection is not particularly clear. Finally, natural antibodies, w^hich are constitutively expressed in the absence of microbial infection, recognize and bind components of the microbial surface and serve as opsonins and can activate complement. Natural antibodies are of fixed specificity because they are expressed from germline genes v^hich do not undergo genetic rearrangements subsequent to microbial infection. Examples include antibodies directed against blood group antigens and those that recognize species-specific cell surface carbohydrate structures. Natural antibodies directed at species-specific murine cell surface carbohydrates present on retroviral vectors derived from murine producer cells are largely responsible for the rapid, complement-mediated lysis of these vectors by human serum. Alveolar macrophages (AMs) play a central role in lung host defense by providing a critical primary barrier to microbial infection mediated by both intrinsic and extrinsic pathv^ays of resistance [6, 7]. Intrinsic resistance mech anisms, which lie completely within the realm of innate immunity, include the ability of macrophages to internalize, degrade, and thus restrict the replica tion of microbial pathogens. Extrinsic resistance mechanisms include release of cytokine mediators that recruit and activate other inflammatory cells (e.g., neutrophils, NK cells) or stimulate antiviral resistance in neighboring cells (e.g., IFNa/IFNp). Both of these mechanisms are also clearly components of innate 3 5 4 Trapnell and Shanley immunity. Extrinsic resistance pathways also include mechanisms by which innate immunity reprograms adaptive immune responses to alter the nature or strength of microbe-specific responses (e.g., cytokines that modulate THI/TH2 development and the balance of cellular and humoral immune responses. How ever, while macrophages help to provide a barrier against infection by some viruses (e.g., vesicular stomatitis, encephalomyocarditis virus and influenza virus), with other viruses (e.g., human immunodeficiency virus) they do not provide a barrier but instead serve as a reservoir of latent infection and facili tate recrudescent disease. Recent data from preclinical and clinical studies with replication-deficient recombinant adenoviral vectors in humans and various animal models strongly support the notion that alveolar macrophages provide a critical innate immune barrier to infection by adenovirus. These data will constitute the focus of the remainder of this chapter. II. Distribution and Clearance of Adenovirus from the Respiratory Tract A. Clinical Aspects of Natural Adenoviral Infection in Humans Adenovirus is an important respiratory pathogen affecting individuals of all ages with an annual incidence of between 5 to 10 million in the United States. Infections can occur sporadically, epidemically and nosocomially but most individuals are infected at a young age; adenovirus accounts for 7 to 10% of all respiratory illnesses in infants and children [8, 9]. Although adenovirus frequently causes a mild, acute upper respiratory illness, e.g., the "common cold," respiratory infections occur as a broad spectrum of distinct clinical syndromes ranging from self-limited acute pharyngitis to fatal pneumonia [10-12]. Adenovirus has also been identified an etiological factor of exacerbations in individuals with chronic obstructive lung diseases and infections can be especially problematic in immunocompromised individuals. Examples of the latter include persistent bladder infections in individuals with chemotherapy-induced neutropenia, fatal pneumonia in neonates, and exacerbation of graft rejection and bronchiolitis obliterans in lung transplant recipients [13, 14]. Natural adenovirus infections are typically initiated by deposition of aerosol droplets containing adenovirus on the mucosal surface of the respira tory epithelium [15]. Some virions then diffuse to the cell surface and enter by receptor-mediated endocytosis. Once inside the cell nucleus, wild-type (replica tion competent) adenovirus DNA overtakes, reprograms, and eventually kills the infected cell ultimately releasing up to 10,000 virions per infected cell. Newly replicated and released virions then infect neighboring cells, repeating the process and thus spreading the infection through the epithelial sheet. The 12. Innate Immune Responses to in Vivo Adenovirus Infection 3 5 5 ensuing clinical course is determined by the race between virus replication and spread and the successful mounting of host innate and adaptive immune responses. B. Distribution of Recombinant, Replication-Deficient Adenoviral Vectors In considering infection of the respiratory tract by recombinant ade noviruses, it is important to recognize tw ô important and fundamental differences from natural infections by v^ild-type adenovirus. First, recom binant human adenoviral vectors used to date for in vivo gene transfer have generally been deleted of El region sequences and are thus relatively replication- deficient in human cells (review^ed in [16, 17]). Further, in mice and primates, tw ô species most commonly used in preclinical gene therapy studies, human adenoviral vectors have a species-related host range restriction that prevents viral replication [15]. These restrictions on viral replication, however, do not affect the infection of a given cell by the vector, i.e., virion internalization and transgene expression proceed normally. Second, and very importantly, there is an enormous difference in the size of the infecting inoculum of adenovirus in the two scenarios. A typical natural infection is thought be caused by less than 1000 adenovirus virions. In contrast, human trials have been conducted in which the virus dose administered was up to one billion times higher (~10^^ virions/individual). Studies in rodent models utilizing transtracheal, liquid bolus administra tion of adenovirus demonstrated that: (1) gene transfer to and expression in airway epithelium was dependent on the dose of virus administered; (2) all cell types could be infected; and (3) gene transfer occurred throughout the bronchial tree but was patchy [18, 19]. Studies in nonhuman primates uti lizing bronchoscopic or aerosol delivery also demonstrated gene transfer but confirmed the overall inefficiency of gene transfer [20-23]. Finally, both direct liquid-based or aerosol-based vector administration in humans also demon strated only low level infection and transduction of the respiratory epithelium. In vitro studies demonstrated, unexpectedly, that while intact airway epithe lium was poorly transduced, damaged epithelium, immature epithelial cells, or differentiating airway epithelium were all easily transduced. These findings were reconciled by studies utilizing a bronchial epithelial cell xenograft model which demonstrated that integrin aj;P5, a coreceptor required for adenoviral virion internalization, was not expressed on the apical membrane of mature airway epithelium and was expressed only on the basolateral surface of the cell [24]. In the context that physical access to the basolateral membrane of epithelial cells of intact epithelium is restricted by tight junctions which connect adjacent airway epithelial cells just below the apical (luminal) surface [25], this data provided early insight into one mechanism by which airway epithilium 3 5 6 Trapnell and Shanley present an innate immune barrier to adenovirus infection. The apical mem brane surface glycocalyx represents another barrier [26]. Evaluation of various organs of cotton rats or monkeys following intrapulmonary adenovirus vector administration using a sensitive polymerase chain reaction technique demon strated that vector does not escape from the lung [21, 27]. Recently, studies in rodents have shown that a large portion of adenovirus administered to the res piratory tract is distributed to alveolar macrophages rapidly after pulmonary administration (see below). C. Kinetics and Mechanisms of Clearance of Adenovirus The initial report of adenovirus-mediated, in vivo transfer of CFTR to the lungs of cotton rats demonstrated the presence of adenoviral vector DNA in the lung as late as 6 weeks after vector administration [28]. Subsequent studies car ried out in various rodent models, nonhuman primates, and humans, however, have demonstrated that most of the adenoviral vector DNA initially adminis tered to the respiratory tract is eliminated from the lung within several weeks in the context of an intact immune system (reviewed in [29]). Data showing that adenovirus-mediated pulmonary transgene expression in athymic mice lasted for more than 3 months implicated the cell-mediated adaptive immunity in pulmonary clearance of adenovirus [30]. This conclusion was supported by the demonstration of prolonged transgene expression in mice depleted of CD4+ cells. The mechanism of this T lymphocyte-mediated clearance was shown to be direct lysis of adenovirus vector-transduced cells by cytotoxic lymphocytes (CTLs) directed at both adenoviral- and transgene-derived proteins. Adaptive immune CTL-mediated clearance is a delayed mechanism because CTL are detected only after several days and are most abundant for 7 to 14 days after infection. The importance of innate immunity in clearance of adenovirus from the lung was first demonstrated by the finding that ^ 7 0 % of the adenoviral DNA present immediately after pulmonary administration in mice was eliminated by degradation within 24 h [31] (Fig. 1). Since this "early phase" clearance was well before a significant adaptive immune response could have been mounted, an innate immune mechanism was sought. Similarity in the pattern of clearance in athymic and normal mice demonstrated independence from lymphocyte-based adaptive mechanisms and alveolar macrophages were pos tulated to be the mechanism of this early clearance. Several findings support this hypothesis. First, in vitro studies demonstrated that infection of human, rat, and murine alveolar macrophages led to loss of approximately two-thirds of the viral DNA within 24 h, whereas similar infection of epithelial cells resulted in no significant loss of viral DNA. Second, pretreatment of the lungs with clodronate-laden liposomes to deplete phagocytic cells significantly impaired the rapid clearance of adenovirus. However, because rapid and significant neu trophil influx occurs during the first 24 h of infection [32], thus overlapping 12. Innate Immune Responses to in Vivo Adenovirus Infection 3 5 7 Acute Intermediate 8 12 16 20 24 28 Time after infection (days) Figure 1 Clearance of viral DNA after adenovirus infection of the respiratory tract. Adenoviral vector DNA is cleared from the lung in a biphasic pattern. Most adenoviral DNA is cleared from the lung very early over the first 24 h after lung infection. The remainder is cleared more slowly over the following several weeks. Clearance during the acute phase is due to internalization and degradation within phagocytes, mostly resident alveolar macrophages (primary clearance). Clearance during the intermediate phase is probably mostly due to recruitment and activation of innate immune NK cells. Clearance during the late phase is principally mediated by the cytotoxic T-cell response. Thus, viral clearance is due to both innate (acute, intermediate) and adaptive (late) immune mechanisms. the early phase of adenovirus eUmination [31], neutrophil-mediated clearance cannot be excluded as an important mediator of viral clearance in these exper iments. This concern is supported by the direct demonstration of the uptake of fluorescently labeled adenovirus by neutrophils recruited to the lung using con- focal microscopy (Zsengeller and Trapnell, unpublished observations). Third, infectious, fluorescently labeled adenovirus is rapidly internalized by alveolar macrophages in vivo as early as 1 min foUov^ing pulmonary administration in mice [33]. The mechanism by v^hich alveolar macrophages internalize adenovirus in vivo is not knov^n but may involve endocytosis and/or phagocytosis and may involve other factors within the local milieu.
In order to better understand this mechanism, it is useful to first consider the mechanism of adenovirus infection in epithelial cells which has been well studied (reviewed in [34]). The virion is internalized by receptor-mediated endocytosis and can be summarized as fol lows: (i) high-affinity binding of the virion to the cell mediated by attachment of the adenovirus fiber knob to its 46-kDa cell surface receptor, CAR [35]; (ii) receptor clustering and rapid virion internalization via a clathrin-coated vesicle mediated by interaction of the adenovirus penton base with integrins avPi or avPs [36-40]; (iii) release of clathrin to generate an endocytotic vesicle; (iv) endosome acidification mediated by an endogenous vesicular membrane proton pump [41]; (v) penetration of the endosome membrane (endosome lysis) and release of the virion into the cytoplasm mediated by the TVD motif- containing cytoplasmic tail portion of integrin P5 [42]; (vi) virion translocation 3 5 8 Trapnell and Shanley to the nuclear membrane mediated by microtubules [39, 43]; (vii) virion bing- ing to the nuclear pore [44]; (viii) capsid disassembly (continued) at the nuclear pore [34]; and (ix) translocation of viral chromatin into the nucleus through the nuclear pore [44]. Adenovirus uptake by mononuclear phagocytes has been studied to some extent in vitro. In contrast to highly susceptible Coxsackie and adenovirus receptor (CAR)+ epithelial cells, hematopoietic lineage cells includ ing alveolar macrophages, monocytes, and related cell lines do not express CAR and internalize adenovirus about 100 to 1000-fold less w êll [45-48]. Internal ization of adenovirus by these cells in vitro requires cell surface integrin ay, similar to CAR"^ epithelial cells and upregulation of integrin a^^s and aypa on human monocytes facilitates infection [45]. Studies using RAW264.7 murine macrophages have show^n that the internalization of adenovirus by these cells is temperature-sensitive and calcium-dependent and requires phosphatidylinosi- tol 3-OH kinase [33]. Data from administration of adenovirus to mice in vivo suggest the potential involvement of other factors or an alternative mechanism of uptake. For example, in vivo uptake of adenovirus by alveolar macrophages is reduced in mice deficient in surfactant protein A [49]. Mice deficient in GM- CSF due to targeted gene ablation are unable to clear adenovirus from the lung and alveolar macrophages in these mice are unable to internalize adenovirus efficiently [50] due to a generalized defect in phagocytosis/endocytosis [51]. In contrast, mice deficient in M-CSF (osteopetrotic mice) have no apparent defect in uptake of adenovirus (Zsengeller and Trapnell, unpublished observa tions). Thus, further studies w'\\\ be required to determine the mechanism by v^hich alveolar macrophages internalize and degrade adenovirus in vitro and in vivo. A second important innate immune mechanism of clearance of adenovirus DNA has recently been demonstrated to be the clearance of virus-transduced cells by recruited NK cells [52]. Intravenous administration of adenovirus results in detectable levels of NK cells in infected tissues by 7-10 days and depletion of NK cells prolonged the duration of transgene expression. Interest ingly, variation in transgene expression betw^een different strains of mice w âs associated with significant differences in levels of IL-12 and IFNy production and NK cell activation. In summary, multiple innate immune barriers block adenovirus infection of the lung and several, redundant innate immune mechanisms of clearance contribute to elimination. Barriers to uptake include production of a mucous layer that traps virions, the ciliary escalator that ejects trapped virions, the epithelial cell glycocalyx that traps virions, and epithelial tight junctions that sequester required adenovirus cell surface receptors aw ây from the luminal surface of the airw^ay epithelial cells. Innate immune mechanisms of clearance include rapid phagocyte-mediated internalization and destruction of the virion (i.e., by primary alveolar macrophages and secondarily recruited neutrophils) and NK cells that destroy adenovirus-infected cells. 12. Innate Immune Responses to in Vivo Adenovirus Infection 3 5 9 III. Molecular Mediators of Inflammation Until recently, little was known about the molecular inflammatory responses triggered by adenovirus infection of the lung. In contrast, bacte rial infection and sepsis has been extensively studied in humans and a variety of animal models and has provided a context for the evaluation of these responses to adenovirus infection. Recent studies of adenovirus pneumonia in children and a great number of studies in animal models and humans receiving replication-deficient adenoviral vectors have provided important details of the molecular signaling responses to adenovirus infection of the respiratory tract. The latter are the result of intense interest and efforts to develop adenovirus- based strategies for in vivo gene therapy for lung diseases such as cystic fibrosis (reviewed in [29]). Consequently, it is now known that pulmonary infection by either wild-type (replication competent) or replication-deficient adenovirus initiates expression of a complex cascade of cytokine mediators that accom panies cellular infiltration of the lung. The precise temporal relationship of expression of these mediators and their relationship to parenchymal infiltra tion by the various leukocyte populations are beginning to define their role in regulation of both innate and adaptive immune events (see below). Before discussing in detail the molecular responses to adenovirus, it is first useful to briefly review some general features of the principal cytokines involved. We will then discuss the cytokine responses during adenoviral pneumonia and the cytokine responses to pulmonary administration of replication-deficient adenoviral vectors in humans, nonhuman primates, and mice. Cytokines are proteins secreted by cells of the innate and adaptive immune systems in response to microbes and antigens that mediate many of the functions of their cellular components [4]. For historical reasons, various terms including monokine, lymphokine, and IL have been used to refer to these molecules. For simplicity, we will use only the term "cytokine," which does not make reference to the cellular origin or target. Cytokine responses to microbes are generally brief, redundant, and often include pleiotropic local and systemic effects. The effects of cytokines are mediated by binding to specific cell surface receptors resulting in alteration of gene expression and acquisition of new target cell functions or cell proliferation. Cytokine responsiveness can be regulated by expression of its receptor on the target cell and cytokines often alter the expression or effects of other cytokines, i.e., some cytokines antagonize while others synergize the effects of another cytokine. Cytokines serve as important regulators of innate immunity (e.g., TNFa, IL-1, IL-12), adaptive immunity (e.g., IL-2, IL-4, IL-5), and hematopoiesis (e.g., granulocyte macrophage- colony stimulating factor (GM-CSF), M-CSF, IL-3, stem cell factor (SCF)). Other cytokines play an important role in both innate and adaptive immunity (e.g., IFNy). While such functional grouping is useful, it is not absolute. For example, while GM-CSF is relevant to hematopoiesis, recent findings have 3 6 0 Trapnell and Shanley shown that it plays a critical role in innate immunity in the lung [53], but has only a noncritical or redundant role in hematopoiesis [54]. TNFa is a principal mediator of the acute inflammatory response to microbial infection and is responsible for many local and systemic effects of infection. The main function of TNFa is enhancement of neutrophil and monocyte recruitment to sites of infection and stimulation of their microbial clearance functions. TNFot is mainly produced by activated mononuclear phagocytes. The principal action and source of IL-1 is similar to that of TNFa. Chemokines (chemotactic cytokines) are a large family of small, structurally similar cytokines that stimulate migration of leukocytes from the blood to sites of local production in tissues (reviewed in [55]). Chemokines have been divided into four groups on the basis of cysteine motifs in their primary structure: C-chemokines (e.g., lymphotactin); CC-chemokines (e.g., MlP-la); CXC-chemokines (e.g., IL-8, MIP-2); and CXXXC-chemokines (e.g., fractalkine). In terms of inflammatory cell recruitment, CXC chemokines mainly promote neutrophil chemotaxis while CC chemokines act on monocytes, lymphocytes, and eosinophils. MIP- l a also functions in an autocrine manner to enhance TNFa expression [56]. Other cytokines are primary mediators of innate immune responses and serve to stimulate adaptive immunity. For example, both IL-12 and IL-18 are released by macrophages and stimulate NK and THI cell release of IFNy, a potent stimulator of cellular immunity. In contrast to the many cytokines that stimulate innate immune cell functions, IL-10 functions as an important inhibitor of macrophage inflammatory cytokine release (e.g., TNFa release), and thus functions as a homeostatic regulator of both innate and adaptive immune responses to microbial infection. IL-6 is a cytokine with pleiotropic functions affecting both innate and adaptive immunity. Among its diverse functions are the stimulation of acute-phase proteins by hepatocytes and stimulation of B lymphocyte proliferation and function. This background will serve as a brief context for considering the cytokine responses to adenovirus infection of the respiratory tract. A. Clinical Adenovirus Infections in Humans Cytokine responses have not been adequately studied in mild cases of natural adenovirus infection, however, a recent study of moderate to severe infections has provided important insights relating the severity of lung infection to the production of certain cytokines [57]. In this study of apparently healthy children, ages 3 weeks to 19 months, who were hospitalized for acute adenovirus pneumonia, the infection was mild in 10, moderate in 12, and fatal in 16. While neither IL-6 nor TNFa were detected in the serum of mild-moderate cases, both were detected in the serum of most fatal cases (13/16, 9/12, respectively). IL-8 was detected in all three groups with serum levels correlating with disease severity. The more severe cases were also noted to have reduced levels of complement, increased levels of circulating immune 12. Innate Immune Responses to in Vivo Adenovirus Infection 3 6 1 complexes, and a decrease in IgG consistent with a consumptive process. IL-1 and IL-4 were rarely detected in any patients and all patients demonstrated by IgM and IgG directed against adenoviral epitopes. Finally, development of septic shock was associated with markedly increased serum levels of IL-6, IL-8, and TNFa. This study demonstrated a clear association between the level of severity of adenoviral infection of the lung, morbidity/mortality and production of certain proinflammatory cytokines. Cytokine responses have also been studied in individuals with cystic fibro sis in which recombinant, replication-deficient, El-,E3-region-deleted adenovi ral vectors were administered to either nasal or bronchial epithelium or both. In one double-blind, vehicle-controlled, dose-escalation study of nasal administra tion, nostrils receiving the adenoviral vector had a greater increase in IL-1, IL-6, IL-8, and IL-10 than vehicle control nostrils in the same patient [58]. In another study of bronchoscopic delivery of a similar vector to the lungs of individuals with cystic fibrosis, IL-6 was increased in the serum 4 h after vector adminis tration with values tapering over 8-24 h [59]. Further, the maximum increase in serum IL-6 levels was proportional to the dose of adenovirus administered. IL-6 was also detected in the bronchoalveolar lavage fluid of these individuals. Taken together, these data show that infection of the respiratory tract by either replication-competent or replication-deficient adenovirus results in release of similar proinflammatory cytokines. Further, cytokines can be detected locally and systemically and appear to be released in proportion to the magnitude of the infection. B. Adenovirus Infections in Animal Models Adenovirus infection of the respiratory tract has been recently studied in a variety of animal models including mice [30, 32, 60], Cotton rats (Sigmodon Hispidus) [15, 19, 27, 61], and primates [20-23] as part of efforts to develop adenovirus-based vectors for human gene therapy for cystic fibrosis (reviewed in [29, 62]). In nonhuman primates receiving an El-,E3-deleted vector by bronchoscopic delivery, IL-lp and IL-8 were both elevated in lung lavage fluid compared to animals receiving only vehicle control. IL-ip levels were mildly increased in the lungs of animals receiving 10^^ plaque-forming units (pfus) at 3 days, but progressively higher levels were seen at 21 and 28 days. In contrast, IL-8 levels were consistently elevated at 3,10, 21, and 28 days following vector administration. No increases were seen in monkeys receiving only low doses (10^ pfu/animal). Thus, studies in nonhuman primates are consistent with human studies and demonstrate a dose-dependent increase in lung cytokine expression following adenovirus infection. Cytokine response studies have been most thoroughly studied in mice. Inbred C57BL/6N mice infected intranasally with 10^^ pfu of wild-type human serotype 5 adenovirus (Ad5) showed elevated lung levels of TNFa, IL-1, and 362 Trapnell and Shanley IL-6 24 h after infection. While TNFa and IL-1 were not elevated in serum at any time, IL-6 was elevated as early as 24 h after infection. TNFa levels were baseline in the
lung by day 3 in contrast to IL-1 levels which rose progressively through 7 days. IL-6 levels peaked on day 1 and fell progressively but were still elevated at 7 days. Subsequent studies have evaluated cytokine responses in BALB/c mice receiving intratracheal administration of an E1-,E3- deleted adenoviral vector [32, 33]. These studies more completely defined the chronology and levels of cytokine responses and demonstrated that the different cytokines could be grouped in terms of patterns of expression at different times after infection: acute (elevated by 6 h), intermediate (not elevated at 6 h, but elevated at 24 h), and late (not elevated at 6 h, but progressively increasing thereafter through 72 h) (Fig. 2). The observation that these responses are similar in nude and control mice demonstrates that these early cytokine responses to adenovirus are independent of adaptive immunity. C. Acute Cytokine Responses Intratracheal administration of an El-,E3-deleted adenoviral vector in BALB/c mice results in elevation of TNFa, IL-6, MIP-2, and MlP-la levels in Acute 24 48 72 Time after infection (hr) Figure 2 Patterns of innate immune cytokine responses in the lung after adenovirus infection of the respiratory tract. Patterns of cytokine responses to adenovirus lung infection can be grouped into "phases" of expression including an acute phase (e.g., TNFa, IL-6) or hyperacute (MIP-2, striped area), an intermediate phase (e.g., IL-lp and IFNy), and a late phase (e.g., MCP-1). Specific cytokines elicited during these phases coordinate a number of innate immune defenses and initiate regulatory pathways that modulate the level of specific adaptive immune responses. See text for details. 12. Innate Immune Responses to In Vivo Adenovirus Infection 3 6 3 bronchoalveolar lavage (BAL) 6 h after infection [32]. In another study TNFa mRNA was detected as early as 30 min after pulmonary infection, whereas TNFa protein was detected in lung by 3 h [33]. In the later study, in situ hybridization analysis demonstrated that adenovirus-induced TNFa mRNA expression was localized to alveolar macrophages; however, expression was not seen in either respiratory epithelium or vascular endothelium for the duration of the study (6 h). The importance of TNFa in adenoviral clearance was demonstrated using mice deficient in TNFa due to targeted gene ablation (TNFa-/-) [63]. Intravenous administration of an adenoviral vector expressing a chloramphenicol transgene resulted in prolonged transgene expression in TNFa-/- mice compared to controls. Abrogation of the adenovirus-induced TNFa response by coadministration of corticosteroids or the suppressive cytokine, IL-10, also increased the duration of adenoviral vector-mediated transgene expression. Blocking TNFa receptor function prolongs adenoviral vector-mediated transgene expression as demonstrated by prolonged transgene expression after coadministration of an adenoviral vector expressing a TNFa receptor decoy with one expressing a conventional marker [52]. Together, these studies support the concept that TNFa, which is derived primarily from alveolar macrophages, plays an important role in elimination of adenovirus following pulmonary infection. MIP-2, an important neutrophil chemoattractant, is elevated within 3 h after pulmonary administration of an El-,E3-deleted adenoviral vector in mice. Levels peak at 6-8 h and then rapidly decline to baseline by 24 h. Consistent with this, MIP-2 mRNA levels are elevated by 30 min in both lung and alveolar macrophage total RNA of adenovirus-infected but not vehicle-exposed mice. These studies demonstrate that the alveolar macrophages can be a potent effector of neutrophil recruitment following adenovirus lung infection. MlP-la expression is initiated with kinetics similar to MIP-2, but in contrast, expression remained elevated at 24 h. Mice deficient in MlP-la due to targeted gene ablation also demonstrate substantially reduced inflammation and delayed viral clearance after Coxsackie virus or influenza virus infection [64]. IL-6 expression after adenoviral infection of the respiratory tract was similar to that of TNFa at both the mRNA and protein levels [33]. Similarly, in situ hybridization localized IL-6 mRNA to alveolar macrophages but not epithelium nor vascular endothelium after adenoviral vector infection of the lung [33]. Despite this chronological description of its response, the role of IL-6 in adenovirus infection of the lung is not currently known. D. Intermediate Cytokine Responses The "second wave" of cytokine responses includes IL-ip and IFNy, both of which are at baseline 6 h after adenovirus lung infection, but are increased by 24 h. While much is known about the general functions of IL-1 and IFNy in other systems, studies in IL-1 or IFNy gene-ablated mice have not been done 3 6 4 Trapnell and Shanley and specific data regarding their role in adenoviral lung infection is lacking. Interestingly, coadministration of corticosteroids abrogated the IFNy response to adenoviral lung infection [32] and vŝ as associated w îth increased adenoviral vector-mediated gene expression. Hov\^ever, since other cytokines were also altered by this treatment, the relevance of changes in IFNy w îll require further evaluation. IFNy production by NK and THI cells can be stimulated by both IL-12 and IL-18, w^hich are both stimulated by adenovirus infection of the lung [65]. However, only IL-18 is necessary for optimal production of IFNy, independent of IL-12. Further, THI cell development was unaffected by blocking both IL-12 and IL-18, demonstrating that other factors may be required for THI differentiation following adenovirus infection. Differences in IFNy levels have been demonstrated between BALB/c and C57BL mice after adenovirus infection and have been related to differences in the duration of adenoviral vector transgene expression in these two mouse strains. E. Late Cytokine Responses MCP-1 was not elevated at 6 h after adenovirus lung infection but progressively increased thereafter and was still at the highest value 3 days after infection. This pattern of expression is different from the immediate cytokines because it is of substantially longer duration. In fact, a decline in expression after adenovirus lung infection has not been defined. MCP-1 is capable of recruiting multiple mononuclear cell subsets, however, its precise role in adenoviral infection is currently unclear. In other lung inflammation models (e.g., silicosis) a network of cytokine/ chemokine signaling has been observed linking TNFa expression to MCP-1 production and T-cell recruitment (Tamez and Shanley, unpublished observations). MCP-1 is also important in recruitment of monocytes. Thus, MCP-1 may have multiple roles in adenoviral lung infection. Interestingly, overexpression of MCP-1 in the lung results in skewing toward TH2 development with increased IL-10 and IL-13 levels and decreased IL-12 and IFNy levels [66]. IV. Inflammatory Cell Recruitment While the clinical manifestations of natural adenovirus infection in the respiratory tract range from inapparent infection to overwhelming pneumonia, sepsis, and death, the histological responses have been well-described in humans only in cases of fatal pneumonia which result in necrotizing bronchiolitis [67]. In human clinical trials with replication-deficient adenoviral vectors, virus administration to the nasal epithelium has also produced mixed results. In one study in which adenoviral vector was applied to one nostril and saline to the other, viral infection was associated with an increase in the percentage of 12. Innate Immune Responses to in Vivo Adenovirus Infection 3 6 5 neutrophils recovered by swabbing compared to the sahne-exposed nostril [68]. Neutrophilia resolved by 1 week in this study. Another study involving direct nasal administration of various doses of a similar adenoviral vector showed no differences in the number of inflammatory cells in epithilium or submucosa between biopsy specimens from adenovirus- or saline-treated nostrils [58]. Histological data are not available from clinical trials in which adenoviral vectors were administered to the lungs of patients. Histological responses to wild-type adenovirus have also been studied in cotton rats and mice in an effort to develop an animal model of human adenovirus diseases and more recently as part of preclinical studies in attempts to develop adenoviral vectors for human gene therapy. It was previously reported [61] that despite evaluation of multiple laboratory animals, clinical disease from wild-type human adenovirus infection of the respiratory tract was identified only in these cotton rats and mice. However, in recent studies, administration of large doses of replication- deficient El-,E3-deleted adenoviral vectors has produced histological evidence of cellular inflammation in several animal models including Cotton rats, mice, nonhuman primates, sheep, and others. Data suggest that the histological responses in these different models are similar. Since they have been best characterized in rodent models we will focus on that data here. Intranasal administration of Ad5 in C57BL/6N mice resulted in devel opment of pneumonia characterized by perivascular, peribronchiolar, and intraalveolar leukocyte [60]. The infiltrate was present on day 1 and was still present on day 7 after infection and consisted primarily of mononuclear cells. The infiltrate was described as consisting of an early phase noted on days 2 through 5 and an overlapping delayed phase peaking on days 5 to 7. The early phase consisted of infiltration by monocytes/macrophages and lymphocytes with scattered neutrophils while the delayed phase consisted of a very promi nent lymphocytic perivascular and peribronchial infiltration. In nude mice which lack thymus-derived T lymphocytes, the delayed lymphocytic phase was markedly reduced in all locations especially in the perivascular region [60]. In contrast, the early phase was not affected in athymic mice. Together, these data suggested that cytotoxic T cells were an important component of the delayed but not the early phase of cellular infiltration. Nasal administration of Ad5 in cotton rats, which are partially permissive for replication of Ad5, produced a similar histological response which was proportional to the dose of virus in the infecting inoculum [61]. Two phases were also noted: an early phase con sisting of monocytes-macrophages, neutrophils, and occasional lymphocytes, and a delayed phase consisting almost exclusively of lymphocytes which was especially prominent in the peribronchiolar and perivascular space. Histological responses to infection of the respiratory tract by replication- deficient adenoviral vectors has been fairly well studied in cotton rats and mice. In Cotton rats, direct comparison of histopathology following intratracheal administration of Ad5 and Ad5-derived El-,E3-deleted replication deficient 366 Trapnell and Shanley vectors showed a qualitatively similar histopathologic pattern of virus dose- dependent leukocyte infiltration [27]. Quantitatively, however, the response to Ad5 was markedly more pronounced than the response to the replication- deficient vector. This study also showed a prominent neutrophilic component early after infection, including infiltration of airway epithelium as early as 24 h after infection. This neutrophilic component was evaluated more carefully in a subsequent study in mice infected by the replication-deficient adenoviral vector, which demonstrated a very rapid neutrophil infiltration which peaked at around 6 h and resolved over the ensuing 3 to 4 days. Interestingly, this coincided with expression of the major neutrophil chemoattractant MIP-2 in these mice [32]. Recently, NK cells have been demonstrated to be an important com ponent of the delayed phase inflammatory cell infiltrate present in the liver after intravenous administration of a replication-deficient adenoviral vector in mice [52]. Increased NK cell numbers were apparent on days 7 and 10 after infection and were associated with increased levels of IL-12 and IFNy. Differences in levels of NK cell accumulation in hepatic tissues and hepatocel lular damage were strongly inversely correlated with the duration of transgene expression in BALB/c and C57BL/6 mice. These finding suggest that differences in NK cell responses may explain strain differences in the duration of transgene expression independent of CTL responses. Cytotoxic lymphocytes 4 6 8 10 Time after infection (days) Figure 3 Pattern of inflammatory cell accumulation in the lung following adenovirus infection of the respiratory tract. Infiltration of the lungs following adenovirus infection has been characterized in several animal models and the general features are similar in each. Neutrophils appear very early and recede fairly quickly. An increase in the macrophage-monocyte population probably begins quite early but is apparent by 2 - 3 days and slowly recedes over the course of a week or so. NK cells have been shown to be present in increased numbers on about days 7 - 1 0 in infected liver tissues and is presumably similar in lung. Cytotoxic lymphocytes are seen late and are prominent on days 7 - 14. These patterns of leukocyte recruitment can be correlated with specific patterns of chemokines known to be chemotactic for the specific leukocyte subsets. See text for details. 12. Innate Immune Responses to in Vivo Adenovirus Infection 3 6 7 Taken together, these studies suggest that massive adenovirus infection of the respiratory tract as would be the case in current strategies for in vivo adenovirus-mediated gene transfer, eHcits a histopathological response consisting of "acute," "intermediate," and "delayed" phases (Fig. 3). The acute phase begins within
hours of infection, peaks within 6-24 h, and resolves thereafter. Both the peak and the duration of the inflammatory infiltrate, which consists primarily of neutrophils, are strongly dependent on the dose of infecting virus. Release of potent neutrophil chemoattractants (e.g., MIP-2 and MlP-la) from alveolar macrophages are likely part of the molecular mechanism of this recruitment. The intermediate phase begins within 1-2 days, continues for upto 7-10 days, and consists of accumulation of monocytes-macrophages and NK cells. Increased levels of MlP-lot, MCP-1, IFNy, and other factors are involved in recruitment and activation of cells during the intermediate phase, although more studies are required to delineate the specific role of these and other factors. The late phase is evident by 7 days, is still present on day 14 but probably subsides thereafter and consists primarily of adenovirus-specific lymphocytes. The cytokines responsible for this phase are not yet well-studied but multiple cytokines are likely to be involved. V. Innate Immunity and Programming of Adaptive Responses Phagocyte-based mechanisms of innate immunity are primarily respon sible for clearing the bulk of adenovirus administered to the lungs during massive infections. However, clearance of adenovirus-infected cells is due to both NK and CTL-based mechanisms belonging to innate and adaptive cel lular immunity, respectively. Further, successful reinfection by adenovirus is inversely proportional to the presence of neutralizing antibodies produced by the humoral adaptive immune system [19, 69]. While these topics are covered in detail in other chapters of this book, it is worth noting how innate immune responses to adenovirus infection in vivo regulate these adaptive immune responses. Cytokines, produced by cells of both innate and adaptive immu nity have a profound effect on the nature and magnitude of specific adaptive immune responses. TNFot, the principal mediator of the acute inflammatory response ade noviral infection of the lung, also has been shown to profoundly affect both cellular and humoral adaptive immune responses to adenovirus infection in vivo. Mice deficient in TNFa due to targeted gene ablation (TNFa -/-) had markedly reduced CTL infiltration of the liver compared to normal controls after intravenous administration of an El-,E3-deleted adenoviral vector [63]. Importantly, the absence of TNFa expression was associated with prolonged expression of the viral transgene. These mice also had substantially reduced 3 6 8 Trapnell and Shanley levels of anti-adenoviral antibodies following infection compared to normal controls. Mice deficient in both TNFa and its related protein, lymphotoxin (TNFp) had reduced inflammation and impaired production of anti-adenoviral IgG and IgA antibodies after intravenous adenovirus administration 170]. The impairment of humoral response in these TNFa/TNF^-deficient mice v^as suf ficient to permit successful reinfection by the vector. In a separate strategy targeted at the TNFa receptor, prior administration of soluble TNF receptor (type I) resulted in prolonged adenoviral vector-mediated gene expression 171]. In a similar approach, mice genetically deficient in both type I and type II TNF receptors demonstrated a significantly reduced humoral antibody response, but no significant prolongation of transgene expression [72]. In the latter study, adenoviral vector-mediated gene transfer of human IL-10, a potent deactivator of macrophages that suppresses TNFa synthesis, also abrogated the neutralizing antibody response to adenoviral vector infection, but had no effect on prolonging transgene expression. Together, these observations shov^ that TNFa, a critical regulator of innate immunity to adenovirus, is also an important regulator of the adaptive immune responses to adenovirus. Other cytokine mediators of innate immunity are also important in reg ulating the adaptive responses to adenovirus. For example, IFNy is important in activation of alveolar macrophages and also directs T helper lymphocyte development toward Tni-type responses. IL-18 has recently been defined as a critical component of adenovirus-induced IFNy production after in vivo infec tion which is independent of IL-12-induced IFNy expression [GS]. Although compelling, these studies demonstrating that cytokines produced by innate immune cells regulate both the nature and strength of adaptive immune responses represent only the beginning of our understanding of the com plex interplay between innate and adaptive immune responses to adenoviral infection. VI. Innate Immunity and In Vivo Gene Therapy Recent progress in our understanding of the significance of innate immune responses to adenovirus infection in vivo have important implications for strategies of in vivo gene therapy using adenovirus. First, the phagocyte- based immune system is present and active prior to vector administration in contrast to adaptive immune responses. Therefore, any strategy designed to deliver large amounts of vector to tissues must take into account the loss of vector due to phagocyte-mediated clearance. Second, clearance of adenovirus by macrophages results in potent inflammatory signals which profoundly influence both molecular and cellular inflammation and likely determine the extent of tissue damage and systemic effects due to adenoviral infection of the 12. Innate Immune Responses to in Vivo Adenovirus Infection 3 6 9 lung. Consequently, if clearance of adenovirus cannot be prevented, uncoupling of phagocytic internalization from release of large amounts of proinflammatory mediators w^ould be useful. Third, because release of inflammatory cytokines determines the level of clearance of adenovirally infected cells by NK cells (e.g., IL-12, IFNy) and CTL (TNFa, TNFp), innate immunity also has a major influence on the duration of transgene expression. Thus, immunosuppressive strategies should be aimed at both innate and adaptive clearance mechanisms or their regulation cytokines. Finally, innate immune barriers to infection of epithelials cells (mucous barrier, glycocalyx, sequestration of required receptors) must be overcome in order to achieve adequate initial levels of adenoviral vector-mediated gene transfer. VII. Future Directions Recent efforts to better characterize the in vivo host responses and barriers to adenovirus infection have defined a number of hurdles to the use of adenoviral vectors for in vivo human gene therapy. With this groundw^ork laid, systematic exploration of these barriers v îll hopefully lead to development of the means to circumvent them. This is already a very active area of research and is likely to continue to be productive. Adenoviral vectors also provide excellent tools v\Aith which to study various biological processes in vivo and to dissect the specific molecular and cellular innate immune responses to viral infection of the lung. Thus, separate from the utility v^ith respect to development of adenovirus-based gene therapy, such studies are clearly increasing our know^ledge regarding the basic biology of an important respiratory pathogen vv̂ hich infects most individuals by adulthood. Acknovs^iedgments This work was supported by The Children's Hospital Research Foundation, Cincinnati, Ohio (BCT), the Cystic Fibrosis Foundation (BCT) and by NIH K08 HL/AI 04291-01 (TPS). References 1. van Furth, R., and Cohn, Z. A. (1968). The origin and kinetics of mononuclear phagocytes. / . Exp. Med. 128(3), 415-435. 2. Kennedy, D. W., and Abkowitz, J. L. (1998). Mature monocytic cells enter tissues and engraft. Froc. Natl. Acad. Set. USA 95(25), 14,944-14,949. 3. Biron, C. A. (1997). Activation and function of natural killer cell responses during viral infections. Curr. Opin. Immunol. 9(1), 24-34. 4. Abbas, A. K., Lichtman, A. H., and Pober, J. S. (2000). "Cytokines. Cellular and Molecular Immunology," 4th ed., pp. 235-269. Saunders Philadelphia. 3 7 0 Trapnell and Shanley 5. Crouch, E. C. (1998). CoUectins and pulmonary host defense. Am. J. Respir. Cell Mol. Biol. 19(2), 177-201. 6. Pels, A. O., and Cohn, Z. A. (1986). The alveolar macrophage. / . Appl. Physiol. 60(2), 353-369. 7. Welsh, R. M., and Sen, G. C. (1997). Nonspecific host responses to viral infections. In "Viral Pathogenesis" (N. Nathason, Ed.), pp. 109-141. Lippincott-Raven, Philadelphia. 8. Eraser, R. S., Muller, N. L., Colman, N., and Pare, P. D. (1999). "Viral Respiratory Diseases. Eraser and Pare's Diagnosis of Diseases of the Chest," 4th ed., pp. 994-996. Saunders, Philadelphia. 9. Horowitz, M. S. (1996). Adenoviruses. In "Eields Virology" (B. N. Eields, D. M. Knipe, P. M. Howley, Eds.), 3rd ed., pp. 2149-2171. Lippincott-Raven, New York. 10. Macek, v. , SorH, J., Kopriva, S., and Marin, J. (1994). Persistent adenoviral infection and chronic airway obstruction in children. Am. J. Respir. Crit. Care Med. 150(1), 7-10. [see comments] 11. Teichtahl, H., Buckmaster, N., and Pertnikovs, E. (1997). The incidence of respiratory tract infection in adults requiring hospitalization for asthma. Chest 112(3), 591-596. 12. Videla, C , Carballal, G., Misirhan, A., and Aguilar, M. (1998). Acute lower respiratory infections due to respiratory syncytial virus and adenovirus among hospitalized children from Argentina. Clin. Diagn. Virol. 10(1), 17-23. 13. Zahradnik, J. M., Spencer, M. J., and Porter, D. D. (1980). Adenovirus infection in the immunocompromized patient. Am. ]. Med. 68, 725. 14. Bridges, N. D., Spray, T. L., Collins, M. H., Bowles, N. E., and Towbin, J. A. (1998). Ade novirus infection in the lung results in graft failure after lung transplantation. / . Thorac. Cardiovasc. Surg. 116(4), 617-623. 15. Ginsberg, H. S., and Prince, G. A. (1994). The molecular basis of adenovirus pathogenesis. Infect. Agents Dis. 3(1), 1-8. 16. Trapnell, B. C , and Gorziglia, M. (1994). Gene therapy using adenoviral vectors. Curr. Opin. Biotechnol. 5(6), 617-625. 17. Trapnell, B. C. (1993). Adenoviral vectors for gene transfer. Adv. Drug Delivery Rev. 12, 185-199. 18. Mastrangeli, A., Danel, C , Rosenfeld, M. A., Stratford-Perricaudet, L., Perricaudet, M., Pavi- rani. A., Lecocq, J. P., and Crystal, R. G. (1993). Diversity of airway epithelial cell targets for in vivo recombinant adenovirus-mediated gene transfer. / . Clin. Invest. 91(1), 225-234. 19. Yei, S., Mittereder, N., Tang, K., O'Sullivan, C , and Trapnell, B. C. (1994). Adenovirus- mediated gene transfer for cystic fibrosis: Quantitative evaluation of repeated in vivo vector administration to the lung. Gene Ther. 1(3), 192-200. 20. Zabner, J., Petersen, D. M., Puga, A. P., Graham, S. M., Couture, L. A., Keyes, L. D., Luka- son, M. J., St. George, J. A., Gregory, R. J., Smith, A. E., et al. (1994). Safety and efficacy of repetitive adenovirus-mediated transfer of CFTR cDNA to airway epithelia of primates and cotton rats. Nat. Genet. 6(1), 75-83. 21. Wilmott, R. W., Amin, R. S., Perez, C. R., Wert, S. E., Keller, G.,Boivin, G. P.,Hirsch, R.,De Inocencio, J., Lu, P., Reising, S. F., Yei, S., Whitsett, J. A., and Trapnell, B. C. (1996). Safety of adenovirus-mediated transfer of the human cystic fibrosis transmembrane conductance regulator cDNA to the lungs of nonhuman primates. Hum. Gene Ther. 7(3), 301-318. 22. Simon, R. H., Engelhardt, J. F., Yang, Y., Zepeda, M., Weber-Pendleton, S., Grossman, M., and Wilson, J. M. (1993). Adenovirus-mediated transfer of the CFTR gene to lung of nonhu man primates: Toxicity study. Hum. Gene Ther. 4(6), 771-780. 23. Brody, S. L., Metzger, M., Danel, C , Rosenfeld, M. A., and Crystal, R. G. (1994). Acute responses of non-human primates to airway delivery of an adenovirus vector containing the human cystic fibrosis transmembrane conductance regulator cDNA. Hum. Gene Ther. 5(7), 821-836. 12. Innate Immune Responses to in Vivo Adenovirus Infection 3 7 1 24. Grubb, B. R., Pickles, R. J., Ye, H., Yankaskas, J. R., Vick, R. N., Engelhardt, J. F., Wil son, J. M., Johnson, L. G., and Boucher, R. C. (1994). Inefficient gene transfer by adenovirus vector to cystic fibrosis airw^ay epithelia of mice and humans. Nature 371(6500), 802-806. 25. Walters, R. W., Grunst, T., Bergelson, J. M., Finberg, R. W., Welsh, M. J., and Zabner, J. (1999). Basolateral localization of fiber receptors limits adenovirus infection from the apical surface of airway epithelia./. Biol. Chem. 274(15), 10,219-10,226. 26. Pickles, R. J., Fahrner, J. A., Petrella, J. M., Boucher, R. C., and Bergelson, J. M. (2000). Retargeting the coxsackievirus and adenovirus receptor to the apical surface of polarized epithelial cells reveals the glycocalyx as a barrier to adenovirus-mediated gene transfer. / . Virol 74(13), 6050-6057. 27. Yei, S., Mittereder, N., Wert, S., Whitsett, J. A., Wilmott, R. W., and Trapnell, B. C. (1994). In vivo evaluation of the safety of adenovirus-mediated transfer of the human cystic fibrosis transmembrane conductance regulator cDNA to the lung. Hum. Gene Ther. 5(6), 731-744. 28. Rosenfeld, M. A., Yoshimura, K., Trapnell, B. C., Yoneyama, K., Rosenthal, E. R., Dale- mans, W., Fukayama, M., Bargon, J., Stier, L. E., Stratford-Perricaudet, L., et al. (1992). In vivo transfer of the human cystic fibrosis transmembrane conductance regulator gene to the airv^ay epithelium. C^//68(l), 143-155. 29. Trapnell, B. C. (Ed.) (1997). "Gene Therapy for Cystic Fibrosis Lung Disease, The Pediatric Lung." Birkhauser Verlag, Basel. 30. Yang, Y., Li, Q., Ertl, H. C., and Wilson,]. M. (1995). Cellular and humoral immune responses to viral antigens create barriers to lung-directed gene therapy with
recombinant adenoviruses./. Virol. 69(4), 2004-2015. 31. Worgall, S., Leopold, P. L., Wolff, G., Ferris, B., Van Roijen, N., and Crystal, R. G. (1997). Role of alveolar macrophages in rapid elimination of adenovirus vectors administered to the epithehal surface of the respiratory tract. Hum. Gene Ther. 8(14), 1675-1684. 32. Otake, K., Ennist, D. L., Harrod, K., and Trapnell, B. C. (1998). Nonspecific inflammation inhibits adenovirus-mediated pulmonary gene transfer and expression independent of specific acquired immune responses. Hum. Gene Ther. 9(15), 2207-2222. 33. Zsengeller, Z. K., Otake, K., Hossain, S. -A., Berclaz, P. -Y., and Trapnell, B. C. (2000). Inter nalization of adenovirus by alveolar macrophages initiates early proinflammatory signaling during acute respiratory tract infection./. Virol. 74(20), 9655-9667. 34. Greber, U. F., Willetts, M., Webster, P., and Helenius, A. (1993). Stepwise dismantling of adenovirus 2 during entry into cells. Cell 75(3), 477-486. 35. Bergelson, J. M., Krithivas, A., CeH, L., Droguett, G., Horwitz, M. S., Wickham, T., Crow- ell, R. L., and Finberg, R. W. (1998). The murine CAR homolog is a receptor for coxsackie B viruses and adenoviruses./. Virol. 72(1), 415-419. 36. Wickham, T. J., Mathias, P., Cheresh, D. A., and Nemerow, G. R. (1993). Integrins alpha v beta 3 and alpha v beta 5 promote adenovirus internalization but not virus attachment. Cell 73(2), 309-319. 37. Li, E., Stupack, D., Bokoch, G. M., and Nemerow, G. R. (1998). Adenovirus endocytosis requires actin cytoskeleton reorganization mediated by Rho family GTPases./. Virol. 72(11), 8806-8812. 38. Li, E., Stupack, D., Klemke, R., Cheresh, D. A., and Nemerow, G. R. (1998). Adenovirus endocytosis via alpha(v) integrins requires phosphoinositide-3-OH kinase. / . Virol. 72(3), 2055-2061. 39. Leopold, P. L., Ferris, B., Grinberg, I., Worgafl, S., Hackett, N. R., and Crystal, R. G. (1998). Fluorescent virions: Dynamic tracking of the pathway of adenoviral gene transfer vectors in living cefls. Hum. Gene Ther. 9(3), 367-378. 40. Wang, K., Huang, S., Kapoor-Munshi, A., and Nemerow, G. (1998). Adenovirus internaliza tion and infection require dynamin./. Virol. 72(4), 3455-3458. 41. Perez, L., and Carrasco, L. (1994). Involvement of the vacuolar H(+)-ATPase in animal virus entry./. Gen. Virol. 75(Pt. 10), 2595-2606. 3 7 2 Trapnell and Shanley 42. Wang, K., Guan, T., Cheresh, D. A., and Nemerow, G. R. (2000). Regulation of adenovirus membrane penetration by the cytoplasmic tail of integrin beta5. / . Virol. 74(6), 2731-2739. [In process citation] 43. Suomalainen, M., Nakano, M. Y., Keller, S., Boucke, K., Stidwill, R. P., and Greber, U. F. (1999). Microtubule-dependent plus- and minus end-directed motilities are competing pro cesses for nuclear targeting of adenovirus. / . Cell Biol. 144(4), 657-672. 44. Greber, U. F., Suomalainen, M., Stidwill, R. P., Boucke, K., Ebersold, M. W., and Hele- nius, A. (1997). The role of the nuclear pore complex in adenovirus DNA entry. EMBO J. 16(19), 5998-6007. 45. Huang, S., Endo, R. I., and Nemerow, G. R. (1995). Upregulation of integrins alpha v beta 3 and alpha v beta 5 on human monocytes and T lymphocytes facilitates adenovirus-mediated gene delivery./. Virol. 69(4), 2257-2263. 46. Huang, S., Kamata, T., Takada, Y., Ruggeri, Z. M., and Nemerow, G. R. (1996). Adenovirus interaction with distinct integrins mediates separate events in cell entry and gene delivery to hematopoietic cells./. Virol. 70(7), 4502-4508. 47. Von Seggern, D. J., Chiu, C. Y., Fleck, S. K., Stewart, P. L., and Nemerow, G. R. (1999). A helper-independent adenovirus vector with El , E3, and fiber deleted: Structure and infectivity of fiberless particles./. Virol. 73(2), 1601-1608. 48. Kaner, R. J., Worgall, S., Leopold, P. L., Stolze, E., Milano, E., Hidaka, C., Ramahngam, R., Hackett, N. R., Singh, R., Bergelson, J., Finberg, R., Falck-Pedersen, E., and Crystal, R. G. (1999). Modification of the genetic program of human alveolar macrophages by adenovirus vectors in vitro is feasible but inefficient, limited in part by the low level of expression of the coxsackie/adenovirus receptor. Am. J. Respir. Cell Mol. Biol. 20(3), 361-370. 49. Harrod, K. S., Trapnell, B. C., Otake, K., Korfhagen, T. R., and Whitsett, J. A. (1999). SP-A enhances viral clearance and inhibits inflammation after pulmonary adenoviral infection. Am. J. Physiol. 277(3 Pt. 1), L580-L588. 50. Trapnell, B. C., Zsengeller, Z., Chroneos, Z. C., Whitsett, J. A., and Berclaz, P. -Y. (2000). Alveolar macrophages from granulocyte-monocyte colony stimulating factor (GM-CSF) deficient mice are arrested at an early differentiation stage. Respir. Crit. Care Med. 162, AOOO. 51. Shibata, Y., Berclaz, P-Y., Chroneos, Z., Yoshida, H., Whitsett, J. A., and Trapnell, B. C. (submitted for publication). "GM-CSF Regulates Alveolar Macrophage Differentiation and Innate Immunity in the Lung Through PU.l." 52. Peng, Y., Falck-Pedersen, E., and Elkon, K. B. (2001). Variation in adenovirus transgene expression between BALB/c and C57BL/6 mice is associated with differences in interleukin-12 and gamma interferon production and NK cell activation. / . Virol. 75(10), 4540-4550. 53. LeVine, A. M., Reed, J. A., Kurak, K. E., Cianciolo, E., and Whitsett, J. A. (1999). GM-CSF- deficient mice are susceptible to pulmonary group B streptococcal infection. / . Clin. Invest. 103(4), 563-569. 54. Dranoff, G., Crawford, A. D., Sadelain, M., Ream, B., Rashid, A., Bronson, R. T., Dickersin, G. R., Bachurski, C. J., Mark, E. L., Whitsett, J. A., et al. (1994). Involvement of granulocyte-macrophage colony-stimulating factor in pulmonary homeostasis. Science 264(5159), 713-716. 55. Zlotnik, A., and Yoshie, O. (2000). Chemokines: A new classification system and their role in immunity. Immunity 12(2), 121-127. 56. Shanley, T. P., Schmal, H., Friedl, H. P., Jones, M. L., and Ward, P. A. (1995). Role of macrophage inflammatory protein-1 alpha (MIP-1 alpha) in acute lung injury in rats. / . Immunol. 154(9), 4793-4802. 57. Mistchenko, A. S., Diez, R. A., Mariani, A. L., Robaldo, J., Maffey, A. F., Bayley- Bustamante, G., and Grinstein, S. (1994). Cytokines in adenoviral disease in children: Association of interleukin-6, interleukin-8, and tumor necrosis factor alpha levels with clinical outcome./. Pediatr. 124(5 Pt. 1), 714-20. 12. Innate Immune Responses to in Vivo Adenovirus Infection 3 7 3 58. Knowles, M. R., Hohneker, K. W., Zhou, Z., OlsenJ. C , Noah, T. L., Hu, P. C , Leigh, M. W., Engelhardt, J. F., Edwards, L. J., Jones, K. R., et aL (1995). A controlled study of adenoviral-vector-mediated gene transfer in the nasal epithelium of patients with cystic fibrosis. N. Engl. J. Med. 333(13), 823-831. [See comments] 59. McElvaney, N. G., and Crystal, R. G. (1995). IL-6 release and airway administration of human CFR cDNA adenovirus vector. Nat. Med. 1(3), 182-184 [Letter]. 60. Ginsberg, H. S., Moldawer, L. L., Sehgal, P. B., Redington, M., Kihan, P. L., Chanock, R. M., and Prince, G. A. (1991). A mouse model for investigating the molecular pathogenesis of adenovirus pneumonia. Proc. Natl. Acad. Sci. USA 88(5), 1651-1655. 61. Prince, G. A., Porter, D. D., Jenson, A. B., Horswood, R. L., Chanock, R. M., and Gins berg, H. S, (1993). Pathogenesis of adenovirus type 5 pneumonia in cotton rats (Sigmodon hispidus)./. Virol. 67(1), 101-111. 62. Boucher, R. C. (1999). Status of gene therapy for cystic fibrosis lung disease. / . Clin. Invest. 103(4), 441-445. [Comment] 63. Elkon, K. B., Liu, C. C , Gall, J. G., Trevejo, J., Marino, M. W., Abrahamsen, K. A., Song, X., Zhou, J. L., Old, L. J., Crystal, R. G., and Falck-Pedersen, E. (1997). Tumor necrosis factor alpha plays a central role in immune-mediated clearance of adenoviral vectors. Proc. Natl. Acad. Sci. USA 94(18), 9814-9819. 64. Cook, D. N., Beck, M. A., Coffman, T. M., Kirby, S. L., Sheridan, J. F., Pragnell, L B., and Smithies, O. (1995). Requirement of MIP-1 alpha for an inflammatory response to viral infection. Science 269(5230), 1583-1585. 65. Xing, Z., Zganiacz, A., Wang, J., Divangahi, M., and Nawaz, F. (2000). IL-12-independent Thl-type immune responses to respiratory viral infection: Requirement of IL-18 for IFN- gamma release in the lung but not for the differentiation of viral-reactive Thl-type lympho cytes./. Immunol. 164(5), 2575-2584. 66. Matsukawa, A., Flogaboam, C. M., Lukacs, N. W., Lincoln, P. M., Strieter, R. M., and Kunkel, S. L. (2000). Endogenous MCP-1 influences systemic cytokine balance in a murine model of acute septic peritonitis. Exp. Mol. Pathol. 68(2), 77-84. 67. Sly, P. D., Soto-Quiros, M. E., Landau, L. I., Hudson, L, and Newton-John, H. (1984). Fac tors predisposing to abnormal pulmonary function after adenovirus type 7 pneumonia. Arch. Dis. Child 59(10), 935-939. 68. Zabner, J., Couture, L. A., Gregory, R. J., Graham, S. M., Smith, A. E., and Welsh, M. J. (1993). Adenovirus-mediated gene transfer transiently corrects the chloride transport defect in nasal epithelia of patients with cystic fibrosis. Cell 75(2), 207-216. 69. Smith, T. A., Mehaffey, M. G., Kayda, D. B., Saunders, J. M., Yei, S., Trapnell, B. C , McClelland, A., and Kaleko, M. (1993). Adenovirus mediated expression of therapeutic plasma levels of human factor IX in mice. Nat. Genet. 5(4), 397-402. 70. Benihoud, K., Saggio, I., Opolon, P., Salone, B., Amiot, F., Connault, E., Chianale, C , Dautry, F., Yeh, P., and Perricaudet, M. (1998). Efficient, repeated adenovirus-mediated gene transfer in mice lacking both tumor necrosis factor alpha and lymphotoxin alpha. / . Virol. 72(12), 9514-9525. 71. Zhang, H.G. , Zhou, T., Yang, P., Edwards, C. K., 3rd, Curiel, D. T., and Mountz,J. D. (1998). Inhibition of tumor necrosis factor alpha decreases inflammation and prolongs adenovirus gene expression in lung and liver. Hum. Gene Ther. 9(13), 1875-1884. 72. Minter, R. M., Rectenwald, J. E., Fukuzuka, K., Tannahill, C. L., La Face, D., Tsai, V., Ahmed, I., Hutchins, E., Moyer, R., Copeland, E. M., 3rd, and Moldawer, L. L. (2000). TNF- alpha receptor signaling and IL-10 gene therapy regulate the innate and humoral immune responses to recombinant adenovirus in the lung. / . Immunol. 164(1), 443-451. C H A P T E R Humoral Immune Response Catherine O^Riordan Genzyme Corporation Framingham, Massachusetts I. introduction Replication-deficient adenovirus vectors are useful tools for the transfer and expression of therapeutic genes into various types of cells and tissues. They have rapidly emerged as the most efficient system of in vivo gene delivery and are w îdely used for gene therapy in basic research and clinical trials. It is now^ well recognized, how^ever, that there are many limitations to the use of adenovirus vectors in gene therapy applications. In particular there is a grov^- ing body of evidence that suggests that immune responses against components of viral vectors limit gene expression and may induce immunopathological consequences. For the most part immune responses consist of conventional antigen-specific lymphocyte responses; in addition, however, there is a nonspe cific and innate response inherent to cells such as macrophages, natural killer cells and organ parenchymal cells. Antivector immune responses can be broadly divided into cellular and humoral. While the cellular immune response limits duration of transgene expression, the humoral response may reduce the therapeutic efficacy of repeated vector administration. It is also emerging that the humoral immune response to viral vectors is not limited to responses against the viral vector components but also includes anti-transgene responses that can affect the persistence of transgene expression. If gene therapy is to be successful the host immune response to the viral vector needs to be overcome. In the case of adenoviral vectors there is a robust immune response following systemic or local administration of vector. Most notably there is a significant rise in neutralizing antibodies directed against the virus which can preclude repeat administration of the vector. This may be a disadvantage in definitively correcting genetic defects where repeated administration of vector would be required. An additional problem potentially ADENOVIRAL VECTORS FOR GENE THERAPY 3 7 5 Copyright 2002, Elsevier Science (USA). All rights reserved. 3 7 6 Catherine O'Riordan exists for human gene therapy apphcations in that most individuals have preexisting immunity to adenovirus; thus, even the first administration of an adenovirus vector to some patients is Ukely to be ineffective. In this chapter the humoral immune response to adenovirus-based vectors is discussed, in particular the many strategies that have been developed to overcome this problem are review^ed. Such strategies include modifications to the viral vector in addition to approaches that modulate the host immune response of the recipient organism. II. Adenovirus Structure and Serotype A. Classification of Adenoviruses The human adenoviruses (Ads) belong to the genus Mastadenovirus and to date 51 serotypes of human Ads have been recognized [1] and grouped into six groups (A to F) [2] on the basis of their hemagglutinating properties and biophysical and biochemical criteria.
Genera of the adenovirus (Ad) family have been further subdivided into numerous serotypes (Table I). Serotype is defined on the basis of its immunological distinctiveness as determined by quantitative neutralization with animal antisera [3]. The most common serotypes used for gene therapy applications are Ad2 and Ad5 both belonging to group C. Virus is only neutralized by antibodies raised against antigens or virus of the same type. Thus for example, types 2 and 5 are in the same species group, but antibody to type 2 w îll only neutralize type 2 virus. Early studies on neutralization of adenovirus by antibodies demonstrated that although most of the antibody made is group-specific and can bind intact virus, only type-specific antibody can neutralize the virus [4, 5]. B. Adenoviral Structural Proteins and Type-Specific Epitopes Adenoviruses share a common architecture consisting of an nonenveloped icosahedral capsid surrounding a linear dsDNA genome of approximately Table I Adenovirus Serotypes Group Serotypes A 12,18,31 B 3,7 , 11, 14, 16 ,21 ,34 ,35 ,50 C 1 ,2 ,5 ,6 D 8, 9, 10, 13, 15, 17, 19, 20, 22, 23, 24, 25, 26, 27, 28, 29, 30, 32, 33, 36, 37, 38, 39, 42, 43, 44, 45, 46, 47, 48, 49, 51 E 4 F 40,41 13. Humoral Immune Response 3 7 7 36 kbp. The viral capsid is composed of three structural proteins: hexon, fibre, and penton base. Hexon is the major structural component, forming the 20 facets of the icosahedron, whereas pentons, being complexes of penton base with fibre, form the 12 vertices [6]. The Ad structural proteins, hexon, fibre and penton base, are the antigenic determinants for both group- and type- specific antibodies. The a determinant demonstrable on hexon is responsible for cross-reacting group antigen found on most of the known human and animal Ads, while the 8 determinant of fibre give rise to type-specific neu tralization antibodies. In addition, hexon possesses a complex arrangement of antigenic determinants, including those with genus, type, intersubgenus, and intrasubgenus specificities [7]. 1. Hexon Ad2 hexon and the closely related Ad5 have been extensively studied. Hexon was the first animal viral particle to be crystallized [8-11] and the Ad2 structure has been resolved by X-ray crystallography to 2.9 A resolution [12, 13]. These studies revealed the Ad2 hexon trimer as a hexagonal "pedestal" base from which a "tower" region projects outward into the solvent. Three surface loops, LI, L2 and L4 from each monomer interdigitate to form the tower domain (Fig. la, see color insert). Sequencing studies completed on adenoviruses from groups A, B, and C have confirmed that there are common sequences within the hexon protein coding domain that are extensively shared. More than 99% of the variability between the hexons of different Ad serotypes is accounted for in seven hypervariable regions [14] that map to the exterior of the protein and include the serotype-specific epitopes [11]. In contrast, the sequences that code for the pedestal are highly conserved. In the first reported X-ray crystallography of the structure of Ad2 hexon [13], there were some unexplained findings which have now been resolved by X-ray cyrstallographic analysis of Ad5 hexon [11]. In particular, the Ad5 hexon model provides new insights into the location of the type-specific epitopes; Ad2 and Ad5 hexon sequences are homologous (86% identity), so their molecular structures are closely related. The structure of the Ad5 hexon monomer (Fig. lb , see color insert) is in general similar to the previously described Ad2 hexon structure [13]. However, the most significant difference is the localization of seven hypervariable regions (HVRs), five of which mapped to the surface of the Ad2 hexon structure and two of which were buried. The new model for Ad5 hexon clears up this anomaly, with all the HVRs locating to the top of the molecule [13]. In efforts to define the hexon residues responsible for type specificity, several analyses have been made of adenovirus hexon primary sequences and of anti-peptide sera that can neutralize adenovirus in a type-specific manner [12-19). Results from these studies suggest that adenovirus-type determinants are localized to sequences within DEI and FGl in the Ad5 3 7 8 Catherine O'Riordan hexon (Fig. lb). For Ad2, type-specific domains have been mapped to unique sequences in loop 1 (amino acids 281-292) of the hexon and loop 2 (amino acids 441-455) (Fig. la), by generating neutralizing antibodies to peptides from each of these regions [19]. 2. Fibre Pentons are the vertex structures of adenovirus capsids. They consist of a penton base protein v^hich anchors a single fibre. The fibre is a trimeric protein [20-22] w^hich allows virus attachment to cells [23], prior to penton base-mediated entry [24, 25]. The fibre has a tail, a slender shaft of variable length, and a globular head [26, 27]. The head domain facilitates attachment and there is appreciable variation in head domain amino acid sequences, w îth several amino acids fully conserved [28]. Variation in head domain sequences may explain differences in receptor specificity betw^een serotypes. For instance human adenovirus serotypes 3 and 5 have been shown to bind to two different cellular receptors via their fibre knob domain [29]. Until recently little was known about the receptor responsible for Ad attachment. However, a 46-kDa protein. Coxsackievirus and adenovirus receptor (CAR), was identified as the receptor responsible for Coxsackie B virus infection of human and mouse cells. In addition, this protein also functions in adenovirus attachment and adenovirus-mediated gene delivery [30]. Fligh-resolution X-ray crystallography of adenovirus type 5 (Ad5) head has revealed much information on the organization of this domain [31]. From these studies it was determined that strongly recognized antigenic determinants containing linear epitopes map to the outer loops or uppermost ^-sheets in the fibre head or knob domain. Studies with peptide-based epitope mapping using polyclonal or monoclonal antibodies against both native Ad2 fibre, Ad2 or Ad5 head domain further support this observation [32-34] and demonstrate that the Ad2 knob domain contains a neutralizing type-specific epitope, a neutral izing group-specific epitope, and a nonneutralizing type-specific epitope [32]. 3. Penton Base Although hexon and fibre contain most of the epitopes recognized by neutralizing antibodies, there are epitopes on penton base that are also recog nized by some neutralizing antibodies [35]. Recent studies on the identification of immunoreactive domains on penton base show that there are three major immunoreactive regions [36]. One is located at the N-terminal domain, whereas the two others are symmetrically displayed on both sides of a conserved RGD motif, one overlaps the fibre-binding site. Interestingly, MAbs against these immunoreactive domains are not neutralizing. However, the RGD motif of penton base, which is involved in binding to cell integrins and internalization of virus, is thought to escape neutralization due to steric hindrance by the fibre protein which prevents IgG binding to all RGD sites on penton base within the intact virus [37]. 13. Humoral Immune Response 3 7 9 C. chimeric Adenovirus Vectors The value of the detailed structural studies on adenovirus hexon is of significance in the design of chimeric adenoviruses. Alteration of type-specific epitopes in hexon can be achieved by exchange of epitopes betv^een adenovirus serotypes. The practical significance of changing the type determinants of a vector v^ould be to allov^ readministration of a gene to a host with circulating neutralizing antibodies. Chimeric adenoviruses were created by replacing the entire hexon in Ad5 or just the hexon FGl loop (Fig. lb) with that from Ad2 [38], the rationale behind this being that type-determining epitopes are primarily associated with this loop. However, in spite of the serotype distinc tiveness of the chimeric hexon viruses, epitope similarity between the vectors resulted in a low level of cross-reactive neutralizing antibody. Thus exchanging the Ad5 hexon with the Ad2 hexon protein did not allow readministration to animals that had previously been exposed to Ad5 virus and thus had anti-Ad5 neutralizing antibodies [38]. In contrast attempts to isolate Ad5/Ad7 chimeras failed, suggesting that it may be difficult to exchange hexons between adenoviruses of greater evolutionary distance [38]. However, a recent study demonstrates that it is indeed possible to package the Ad5 genome into virus particles that contain hexon proteins from other Ad serotypes [39]. Using viruses that represent subtypes B (Ad3), D (Ad9), and E (Ad4), it was shown that Ad5 DNA could be packaged into capsids that contain the hexons of more divergent subtypes. An alternative strategy was employed in the generation of an Ad5/Adl2 chimera where rather than replacing the entire Ad5 hexon, the external loops of hexon were just replaced [40]. With this approach an Ad5/Adl2 hexon chimera was created by replacing all four loop domains of Ad5 hexon with those of Adl2 hexon [40]. Ad 12 is in subgroup A, which is evolutionary quite distant from the Ad5/Ad2 containing subgroup C [41]. This chimeric virus efficiently evaded host immunity in mice immunized previously with an Ad5 vector. The alteration of the type-specific epitopes is not limited to the exchange of epitopes between adenovirus serotypes. For example, an eight-amino-acid sequence from the major antigenic site in the VPl capsid protein of poliovirus type 3 was inserted into two regions of Ad2 hexon [42]. Antisera specific for the poliovirus sequence efficiently neutralized the modified adenovirus and antisera raised against the modified adenovirus recognized the VPl capsid protein of poliovirus type 3. The ability to switch hexons is most Ukely due to the considerable conservation of the amino acid sequence of hexon across subtypes [39]. As already discussed, more than 99% of the variability between hexons of different Ad serotypes is accounted for in seven hypervariable regions that map to the exterior of the protein and include the serotype specific epitopes (section II.B.l). Since hexon protein is a major antigenic determinant of the Ad capsid, this approach of switching capsids or replacing epitopes may prove 3 8 0 Catherine O'Riordan useful in reducing the antigenicity of therapeutic Ad vectors to allow repeated vector administrations. Thus modifications to the hexon structure can be used as a strategy to modulate the immune response to adenoviral vectors. D. Influence of Serotypic Variations on Adenoviral Cell Interactions Entry into and trafficking through target cells has been studied mainly with subgroup C viruses. Binding to target cells occurs via a high-affinity inter action between the fibre protein and the coxsackie-Ad receptor (CAR) on the cell surface [30]. Subgroup C adenovirus then rapidly enter cells by endocytosis through interaction of the penton base protein with vitronectin binding inte- grins on the cell surface, including avp3, avP5, am^l , and a5pi integrins [24, 43-46]. Endosomal membranes are lysed by adenovirus, allowing the escape of capsids to the cytosol [24, 47, 48]. Finally, adenovirus translocates to the nucleus by using microtubules in the cytoplasm, binds to the nuclear envelope, and inserts its genome through nuclear pore complexes [49, 50]. There are three viral interactions with the host cell entry pathway which can be influenced by serotypic variations in the capsid proteins: (i) different serotypes of adenovirus bind through different cell surface receptors [29, 51], (ii) the secondary virus-cell interaction mediated through penton base varies by serotype [24, 25, 43], and (iii) escape from the endosome may involve distinct mechanisms of vesicle disruption, again in a serotype-dependent manner [46, 51]. At least two points along the pathway of gene transfer by aden ovirus are susceptible to blockage by circulating antibodies, blocking the fibre protein-cellular receptor interaction (CAR) and escape from the endo some. Escape of the viral genome from the endosome can be inhibited by capsid-bound antibodies against both hexon, the major capsid component, and penton base [35]. Antihexon antibody is considered to be the dominant neutralizing antibody in response to adenovirus infection [35]. Neutralization by anti-hexon antibodies has single hit kinetics with an average of 1.4 antibody molecules bound per virion, and it is thought that bound antibody inhibits a low pH-induced conformational change that takes place in the acidic endo- somes [35]. During this conformational change the N-terminal region of the protein is exposed and antibodies directed against an N-terminal 15 K pro teolytic fragment can neutralize virus infectivity [52]. In contrast anti-penton base antibodies play a far less significant role in the neutralization of adenovirus compared to anti-hexon antibodies, inhibiting infection by only 50% [35]. As already mentioned (section II.B.3), this may be due to the architecture of the penton-base-fibre
complex where steric hindrance by fiber may prevent IgG binding to penton base. Although antifibre antibodies can neutralize adenovirus infection [35], there are reports claiming that fibre is not an important immunogen [4, 53]. Additionally, the immunogenicity of fibre varies; it is a relatively weak 13. Humoral Immune Response 3 8 1 immunogen if it is delivered as a purified protein and a relatively strong immunogen v^hen delivered in the context of a v^hole virus particle [54]. It is thought that anti-fibre antibodies do not possess true neutralizing activity until present at very high titer [55] and only then do they prevent transduction follow^ing a systemic administration of a gene therapy vector [35]. Use of an Ad chimera, adenovirus type 5 v^ith a fibre gene from Ad7A, [55] has helped define the contribution of the antifiber humoral immune response against adenovirus infection. The chimeric virus clearly demonstrated that under standard repeat administration conditions, in the rat, the antibody generated against the fibre gene is inconsequential with regard to neutraliz ing function, suggesting that the only capsid protein that is functioning as a dominant neutralizing epitope is the hexon trimer [55]. In addition, results from a human clinical trial follow^ing administration of an Ad vector suggest that anti-fibre antibodies are only neutralizing when acting in synergy with anti-penton base antibodies [56], III. Host Response to Gene Therapy Vectors Three phases have been described in the elimination of adenovirus after intravenous delivery [57]. Phase one involves innate immune mechanisms, which occurs within 24 h postinfection and accounts for the elimination of 90% of the adenoviral DNA. Phase two is mediated by the adaptive immune system and consists of a cytotoxic T lymphocyte (CTL) and/or antibody response to the transgene and/or to the viral proteins. Phase three is characterized by a slow and constant decrease of the transgene expression even in the absence of an immune response to the transgene and is thought to be the result of leaky expression of viral gene products which leads eventually to clearance of the transduced cells. A. Innate Immune Response Although immune-related mechanisms play a significant role in eliminat ing the recombinant Ad genome following in vivo administration of Ad vectors, it is not clear how the immune response to Ad vectors is initiated, i.e., for the immune system to be sensitized to Ad antigens. Studies by Worgall et al. [58] suggest that there is some initial destruction of the Ad vectors by innate immune mechanisms. In contrast to the antigen-specific, adaptive immune response, innate immune mechanisms comprise the immediate, nonantigen- specific events which include tissue macrophages, which act as scavengers to clear incoming pathogens. In addition tissue macrophages influence the ini tiation of the adaptive immune response. Innate pathways of virus clearance are also mediated by Kupffer cells within the liver [58]. Following the innate immune response a second phase of the immune response occurs, the adaptive immune response. This results in the generation of antibodies against viral 3 8 2 Catherine O'Riordan capsid proteins and transgene, i.e., the humoral immune response, in addition CTLs against viral proteins and transgene products ensue to generate the cellular immune response. B. Adaptive Immune Response: B-T Cell Interactions Activated T cells play a critical role in the generation of humoral and cellular immune responses. In general there are tv^o major classes of T cells, HelperTCell IL4 IL-2 IL.5 Î IFN Activate B^ells AdvateCTL's Figure 2 (a) Antigen-presenting cells (APCs) such as dendritic cells and macrophages ingest foreign proteins (viral particles) nonspecifically. Ingested antigens are processed intracellularly and presented as peptide fragments in the context of the MHC class II complex. Activation of the helper T cell occurs following recognition of the foreign peptide/MHC class II molecule by the T-cell receptor, TCR. In addition a second signal is required for helper T-cell activation which occurs when the plasma membrane bound signaling molecules B7-1/B7-2 are recognized by a coreceptor protein CD28/CTLA4-lg which is present on the surface of the helper T cell. Cellular immune response: At the same time viral vectors can infect (APCs) and deposit their genomes into the nucleus. The genome encodes viral and transgene proteins that are expressed and presented by MHC class 1 molecules to CD8+ cells or cytotoxic T lymphocytes (CTLs). (b) Foreign antigens are taken up from the extracellular fluid by receptor mediated endocytosis following binding to the B cell receptor. They are then degraded and recycled to the cell surface in the form of peptides bound to MHC class H molecules. Thus the helper T cell activates the B cell that displays the same antigen in the context of the MHC class II molecule as originally activated it. in addition, interaction of CD40 ligand with CD40 activates B cells to proliferate and mature into memory and antibody-secreting cells. In secondary antibody responses, memory B cells themselves may act as APCs and activate helper T cells as well as being the subsequent targets of the helper T cells. 13. Humoral Immune Response 383 Figure 2 (continued) cytotoxic CD8+ T cells (CTLs) and helper T cells (CD4+ T cells). In the context of a viral infection, CTLs eliminate virally infected cells expressing neoantigens such as viral proteins and the transgene product. In contrast, CD4+ T cells help activate the responses to extracellular antigens by stimulating B cells to proliferate and secrete antibodies. Activation of antigen-presenting cells (APCs) and B cells by input viral capsid proteins underlies the mechanism responsible for the production of the humoral immune response to Ad vectors. Administration of UV-inactivated virus leads to a full humoral response v^ithout any CTL involvement, v^hich is consistent v^ith the role of exogenous viral capsid proteins in the activation of B cells and of endogenously produced antigens in the activation of primary CD8+ cells [59, 59a]. Hov^ever, activation of CD4+ T cells by viral capsid proteins has been shov^n to contribute to CTL- mediated clearance of Ad-transduced cells in addition to stimulating B cells to produce neutralizing antibodies [59-60]. Activation of T cells by antigens requires a complex program of molecular interactions betv^een the T cell and an APC, each of which could be a target for immune blockade in gene therapy. Initiation of a T-cell-mediated response requires APCs v^hich present short peptides derived from ingested foreign antigens (e.g., a virus particle) in association v^ith major histocompatability (MHC class II) molecules to interact with the T-cell receptor (Fig. 2a). In 3 8 4 Catherine O'Riordan addition, there are other signals that are needed for successful stimulation of the T-cell response. These include B7-1 (CD80) and B7-2 (CD86) ligands present on antigen presenting cells which bind to the CD28 / CTLA4 receptors on T cells and elicit a costimulatory response needed for this activation [61]. Such costimulatory responses result in cytokine secretion and full T-cell activation (Fig. 2a). There are two functionally distinct subclasses of helper T cells that can be distinguished by the interleukins that they secrete upon activation. Thl cells secrete IL-2 and gammainterferon and are concerned mainly with helping cytotoxic T cells and macrophages, while Th2 cells secrete interleukin (IL)-4 and IL-5 and are concerned mainly with helping B cells (Fig. 2a). Once activated, the helper T cell can stimulate a B cell that specifically displays the same complex of foreign antigen and MHC class II protein on its surface (Fig. 2b). T-cell dependent activation of B cells in turn leads to upregulation of CD40 ligand (CD40L) on the T cell, promoting the engagement of CD40 on the cognate B cell [62] (Fig. 2b). IV. Strategies to Overcome the Humoral Immune Response The humoral and cellular immune response to recombinant adenoviral vectors, as described in several animal models, result in extinction of transgene expression, severe local inflammation, and production of neutralizing antibod ies that prevent readministration [59, 63, 64]. A direct correlation between neutralizing antibody and the block to readministration of vector has been established by passive transfer of immunity by sera from treated to naive animals [59]. One approach to enhance adenoviral-mediated gene transfer is to modulate the host immune response by immunosuppression of the recipient organism. A. General Immunosuppression Chronic immune suppression with drugs such as cyclosporine and cyclophosphamide has improved the stability of adenovirus-encoded trans- gene expression in animal models of liver-, lung-, and muscle-directed gene therapy [64-66]. Cyclophosphamide is a commonly used immune suppressive agent for the treatment of autoimmune diseases and prevention of rejec tion following allograft organ transplantation [67]. It is activated by hepatic cytochrome p450 to metabolites that exhibit toxicity primarily to dividing cells, including activated T and B cells [68]. Administration of cyclophosphamide with intravenous infusion of aden oviral vector blocked activation of both CTL and T helper cells, resulting in prolonged transgene expression in the liver with reduced anti-Ad neutralizing antibody production [66]. A similar effect was seen in the lung; however, a 13. Humoral Immune Response 3 8 5 much lower dose of cyclophosphamide was needed to prevent neutralizing antibody formation. In contrast, stabilization of transgene expression was achieved only at a high dose. This difference may be a consequence of dif ferences in the route of administration of the vector, which could result in differences in presentation of antigens. For example the intravenous route more likely deposits larger quantities of antigens to tissue enriched with antigen- presenting cells such as the spleen. In addition neutralization of virus in the lung is restricted to the Th2-dependent isotope, which is easier to ablate than ablation of both Thl and Th2 subsets which contribute to formation of antiviral responses when vector is delivered systemically [66]. In contrast cyclosporin (CSA) alone failed to reduce the production of neutralizing antibodies to cFIX in hemophilia B dogs but was effective at pro longing gene expression of FIX [64, 65]. CSA reportedly inhibits early events in T-cell activation such as activation of interleukin-2 gene expression [69], which may explain why CSA most likely affected the cellular rather than the humoral immune response following adenovirus-mediated gene therapy in the hemophilia B dogs. One of the main concerns with the protocols used in animal models for general immunosuppression is the high dose necessary to successfully obtain readministration of gene therapy vectors. This is substantially higher than approved doses for use in humans. Thus it remains to be established whether clinically acceptable doses (presumably lower doses) may indeed have the same effect on immunosuppression and allow readministration of gene therapy vectors in a clinical setting. Bouvet et al. [70] report that etoposide at clinically acceptable doses suppresses the formation of neutralizing antibodies and CTLs to adenovirus and results in successful intratumor transgene expression in immunized mice. Etoposide is a semisynthetic derivative of podophyllotoxin that causes an arrest at G2 of the cell cycle. It inhibits DNA synthesis by interfering with the enzyme topoisomerase II and leads to cell death by apoptosis [71]. Thus repeated adenoviral-mediated gene therapy may be achievable in cancer patients who are concurrently undergoing treatment with chemotherapy. Most of the immunosupressants discussed so far have the distinct dis advantage of causing general immunosuppression that may not be desirable in some clinical settings. An alternative immunosuppressant, deoxyspergualin, (DSG), with more selective properties has been shown to be useful in read ministration of systemically delivered viral vectors expressing Factor IX [72] or lung-directed viral vectors expressing the human cystic fibrosis conductance regulator (hCFTR) [73]. Deoxyspergualin interferes with the differentiation of B and T cells and also with antigen processing. An important property of DSG is that it does not induce a general suppression of the immune system, but rather results in a selective lack of response to specific antigens presented at the time of drug treatment. 3 8 6 Catherine O'Riordan B. Transient Selective Immunosuppression The central role of the CD4"^ T cell provides a strategy to prevent humoral and cellular responses to adenovirus vectors through a transient blockade of CD4+ T-cell activation at the time of vector administration. The rationale for this approach is that chronic immune suppression should not be necessary if the primary stimulus for activation is the input capsid proteins. In support of this hypothesis, it has been shown that depletion of CD4^ T cells w îth a monoclonal antibody (GK1.5) at the time of vector administration can effectively prevent CTL and B-cell responses in murine models of liver- and lung-directed gene therapy [74-76]. 1. Cytokine
Treatment Selective inhibition of the TH2 subset of T helper cells by administration of the cytokine interleukin 12 or gamma interferon (IFNy) v^ith adenovirus vector has prevented the humoral immune response in mouse lung tissue [77]. The success of this approach, how^ever, depends on the relative contribution that Th2-dependent immunoglobulin (Ig) isotypes play in virus neutralization, the profile of which may be affected by strain and species of animal as well as routes of vector administrations. Th2-specific ablation with IL-12 is an effective approach for lung-directed gene therapies in the mouse where IgA is the primary source of neutralizing antibodies. However, in the case of the mouse liver both Thl and Th2 cells contribute to the production of virus- specific antibodies and although IL-12 reduced the total amount of neutralizing antibody in this organ it was not enough to allow effective readministration of the virus [75]. 2. CTLA4 Ig Interfering with the distal pathway of CD4+ activation by administering CTLA4 immunoglobulin (Ig) with adenovirus vector improved the stabil ity of recombinant gene expression in mouse liver but did not significantly impact neutralizing antibody production or allow systemic vector readmin istration [78]. muCTLA4Ig is a chimeric protein of murine immunoglobulin IgG2a fused to murine CTLA4 and is an inhibitor of the CD28-B7 pathway (Fig. 2a). In contrast in the lung huCTLA4-Ig treatment significantly blocked the formation of neutralizing antibodies allowing efficient readministration of virus, whereas transgene expression was only moderately prolonged [79]. Differences between the effects of this drug on humoral and cellular responses in lung versus liver are surprising. It is known that secretory IgA, which depends on Th2 cells, contributes to neutralization in the lung, whereas antibodies of other isotypes, such as IgGl and IgG2a, neutralize virus adminis tered systemically. As previously discussed, different routes of virus instillation may result in different mechanisms of antigen presentation, which could affect 13. Humoral Immune Response 3 8 7 inhibition of B- and T-cell activation by CTLA4Ig. For example, local injection of an adenovirus vector expressing CTLA4 Ig into the brain suppressed not only local cell infiltration in this tissue but also reduced the humoral immune response to adenovirus [80]. 3. Anti-CD40 Ligand Antibody The expression of CD40L by activated helper T cells (Th) triggers B cell cycling through binding to CD40 (Fig. 2b). CD40L is expressed transiently at high levels on activated CD4+ T cells [81, 82]. The costimulation provided by CD40L w îth CD40 is essential for thymus-dependent humoral immunity [82, 83] and is also thought to play an important role in the generation of cellular immune responses through the production of helper cytokines [83, 84]. A transient block of costimulation between T cells and B cells and other antigen- presenting cells using a monoclonal antibody against CD40 ligand (MR-1) suppressed the development of antibodies against Ad delivered to mouse or nonhuman primate lung, in addition to decreasing the cellular immune response to the vector [85-87]. This in turn resulted in an increase in persistence of transgene expression. Furthermore, when MRl was administered with a second dose of Ad vector to mice preimmunized against vector, it was able to interfere with the development of a secondary antibody response and allowed for high levels of transgene expression upon a third administration of vector to the mouse airway [86]. Similarly in the nonhuman primate lung administration of a humanized anti-CD40 ligand MAb (hu5C8) at the time of vector instillation, markedly suppressed adenovirus-induced lymphoproliferation and cytokine responses, in addition there was a marked suppression of IgA and neutralizing antibodies which permitted vector readministration [87]. In the case of systemic delivery of vector it is thought that only CD40 ligand blockade inhibits anti-Ad antibody generation sufficiently to allow redosing to the liver [88]. Thus a combination of anti-CD40 ligand and murine CTLA4Ig was necessary to allow transduction after secondary vector administration in mouse liver, whereas neither agent alone was sufficient [89]. C. Oral Tolerance Orally administered antigens have been shown to induce systemic unre sponsiveness to a subsequent exposure to the antigen. Oral tolerance primarily results in active suppression by regulatory T cells, clonal anergy, or clonal deletion. What activates one mechanism over the other is not altogether clear, although it is generally agreed that lower doses of antigen more likely lead to suppression and higher doses to clonal anergy or deletion [90]. In addition, the form of antigen and frequency of feeding influences the type of tolerance induced [91]. Long-term adenovirus-based gene expression was observed in the liver of rats with preexisting anti-adenovirus antibodies that were tolerized by feeding 3 8 8 Catherine O'Riordan viral proteins (11 x 1-mg dose) [92]. In the tolerized rats the anti-adenovirus humoral immune response was downregulated allowing systemic readminis- tration of vector. Moreover, in the tolerized rats vector readministration did not lead to a secondary humoral immune response, suggesting that repeated adenovirus-directed gene transfer may be possible despite the presence of a residual antibody titer from a previous exposure to vector [92]. Similarly, this approach has been used to enhance gene transfer to the rat parotid gland which is within the mucosal immune system [93]. As it is more difficult to induce complete mucosal tolerance compared to systemic tolerance only partial tolerance to mucosally applied viral vectors was achieved. A different feeding regimen of adenovirus (5 x 50 \|JLg dose) along with a different form of antigen, UV inactivated vector, in addition to a difference in route of viral challenge may explain the differences in the results from the two studies. Nonetheless, it is possible to induce some degree of tolerance to mucosally applied adenovirus by feeding animals the virus. In clinical situations, some concern may exist about tolerizing a host toward adenoviruses, which can be pathogenic in humans. D. Serotype Switching One strategy to circumvent Ad vector-specific neutralizing immunity is to switch the serotype of the Ad vector [94, 95]. As already discussed in section II. A, there are 51 serotypes that are classified on the basis of biological, chemical, immunological, and structural properties into six subgroups and then into serotypes based on neutralization by antisera to other Ad serotypes. Following an initial administration of adenovirus, serotype-specific antibodies are generated against the major viral capsid proteins (section II.B). Group C adenoviruses include Ad2 and Ad5, the more commonly used serotypes for gene therapy vectors. While the capsid proteins of the group C adenoviruses are highly conserved, viruses from a different subgroup have capsid proteins which are only weakly homologous to the group C viral capsid proteins. Thus the immune response against the non-group-C viruses in many instances does not block infection by a group C virus [94]. This was demonstrated in a study where Sprague-Dawley rats were injected intraperitoneally (ip) with wild-type Group B virus, WT Ad7, either alone or sequentially with WTAd4 (group E) prior to intracardial administration of an Ad5-based gene transfer vector [94]. Transgene expression in all animals that received non-group-C viruses prior to Ad5 was equivalent to naive animals. In contrast, animals that received WT Ad5 prior to the Ad5-based gene transfer vector had greatly reduced levels of transgene expression compared to naive animals. Similarly, in the development of gene therapy vectors for the lung and the treatment of cystic fibrosis, it was shown that intratracheal administration of an immunizing does of wild-type Ad 4 (subgroup E) or Ad30 (subgroup D) 13. Humoral Immune Response 3 8 9 did not affect the subsequent expression of human CFTR from an Ad5 based gene transfer vector [95]. More importantly the alternate use of Ad vectors from different serotypes (Ad2, Ad5) within the same subgroup (C) can also circumvent anti-Ad humoral immunity and permit effective gene transfer to the lung [96] and to the liver [97] upon repeat administration. In the context of future clinical applications for this approach it is relevant that Ad2- and Ad5-based vectors can be administered alternately as these are the Ad serotypes that are in current use in human clinical trials. E. Masking Neutralizing Epitopes Alternative approaches have been developed for circumventing antibody neutralization of Ad vectors that are centered on modification of the Ad virion rather than on immunosuppressive treatment of recipient animals [98-101]. One such approach involves the covalent attachment of the polymer PEG (polyethylene glycol) to the surface of adenovirus [99^ 100]. Covalent mod ification of Ad virions using chemically reactive PEG renders the virus less susceptible to neutralization, due to shielding of neutralizing epitopes on the surface of the virus by PEG molecules. The components of the capsid that elicit a neutralizing immune response, i.e., hexon, fibre, and penton base (see section II.B) are the main targets for PEGylation. Importantly the covalent attachment of a PEG polymer to the surface of the adenovirus can be achieved with retention of infectivity, while PEG-modified adenovirus have been shown to be protected from antibody neutralization in the lungs of mice with high antibody titers to adenovirus [99, 100]. Similarly, Beer et al, [98] demonstrated that adenovirus vectors could be formulated in a polymer preparation of PLGA (poly(lactic/glycolic acid) with retention of bioactivity. Mice immunized subcutaneously with encapsulated recombinant adenoviral vectors show a greater than 45-fold reduction in anti- adenovirus titers relative to nonencapsulated vectors. Although the authors do not show any in vivo data they postulate that the process of encapsulation of a vector in a polymer preparation may potentially mask the adenovirus from circulating antibodies. This is also based on the observation that encapsulated vectors are less susceptible to neutralization than nonencapsulated vectors in vitro. The possible disadvantages of this approach include the efficiency at which the Ad vector is released from the polymer and the true demonstration that this approach has any advantage in vivo. More recently another nongenetic strategy to modify the surface of the virion has been described. This involves a covalent coating using a multiva lent hydrophilic polymer based on poly-[N-(2-hydroxypropyl)methacrylamide] (pHPMA). Multivalent polymeric modification of adenovirus rendered the virus less susceptible to neutralization by anti-adenovirus antibodies. As with the studies with PLGA this approach was shown only to be effective in vitro and may not be applicable in an in vivo setting. 3 9 0 Catherine O'Riordan F. Immunoapheresis Another novel approach to overcoming the problem of serum neutralizing antibodies w âs described by Chen et al. [102], and involves the principle of immunoapheresis. An affinity column consisting of cloned recombinant capsid proteins v^as generated to specifically remove anti-adenovirus antibodies from human clinical serum samples. The authors postulate that such an affinity column could be used in conjunction w îth apheresis in a technique called immunoapheresis. During the apheresis procedure, patients' serum could pass through this immunoaffinity column removing anti-adenovirus antibodies. Anti-adenovirus antibodies would be expected to repopulate the vascular compartment eventually but a temporal v^indow of several hours for intravascular adenovirus therapy could be created [103]. V. Factors Modulating Host Responses to Gene Transfer Vectors The immune response generated to adenoviral vectors can be broadly divided into cellular and humoral responses. The cellular immune response results in cytotoxic T cells against viral proteins and transgenes that can lead to the destruction of vector-transduced cells and reduced persistence of transgene expression. In addition, there is the humoral immune response against both capsid proteins and transgene. The design of the vector backbone has been shown to influence the persistence of transgene expression by modulating the cellular immune response. Similarly, the use of tissue-specific promoters to drive transgene expression can modulate the humoral immune responses to the transgene product. A. Viral Vector Backbone 1. First-Generation Adenoviral Vectors First-generation adenoviral vectors are typically deleted for El and some times for E3, which result in decreased expression of early and late genes and deficient replication of the virus (Fig. 3). Multiple studies performed with these vectors in mice. Cotton rats, and nonhuman primates have shown successful gene transfer in a variety of tissues with high levels of expression of trans- genes [104-107]. However, a major shortcoming of these El-deleted vectors is the general loss of transgene expression at 3-4 weeks after administration [59^ 63, 108], although administration to neonatal or immune-deficient animals frequently results in more persistent expression [64, 63, 109]. The replication defect of El-deleted viruses can be overcome in part because cellular factors complement the function of the El gene, leading to the expression of viral 1 3 . H u m o
r a l I m m u n e Response 391 A d 2 D N A G e n o m e 3 6 k b p E 3 E 1 A E 1 B ITR ITR E 2 E 4 R e c o m b i n a n t A d E 3 ITR E 2 E 4 Promoter T ransgene Figure 3 Schematic of adenovirus genome. The bar indicates the DNA genome of 36 kb. The black arrows indicate the transcription units: El A immediate-early transcription unit and the delayed-early ElB, E2, E3, and E4 transcription units. The expression cassette replacing the El region is depicted. The cassette contains a promoter with transcriptional elements used to drive expression of the transgene. The transgene is the open reading frame in the recombinant transcript that translates the desired protein. First-generation adenoviruses are made by substituting an expression cassette for the El and/E3 regions. Second-generation adenovirus vectors are generated by the additional deletion of genes necessary for viral replication, e.g., E2a DNA-binding protein [114, 115]; E4 region [113]; the E2b-encoded terminal protein and viral DNA polymerase [120]. Third-generation vectors are vectors deleted for all viral genes, but retain the c/s-acting sequences necessary for viral replication and packaging (see Chapter 15). proteins that are presented in the context of MHC class 1 molecules to elicit a cytotoxic T-cell response [63]. For these reasons additional changes have been introduced into the backbone of adenovirus vectors to render them more replication-defective and thus further reduce their potential for viral gene expression. 2. Second-Generation Adenoviral Vectors Tvŝ o regions of adenovirus, E2 and E4, v\̂ hich play critical roles in viral DNA synthesis and late gene expression have been targeted for dele tion [110-113]. A temperature-sensitive mutation, ts l25, and a deletion have been introduced into the E2a region [111, 114, 115], vectors containing this mutation show^ed prolonged transgene expression in CBA mice. Cotton rats, and nonhuman primates. Hov^^ever, contrary to these reports, a recent study using BALB/C mice and hemophilia B dogs demonstrates that this E2a mutation 3 9 2 Catherine O'Riordan is insufficient for achieving persistent expression [116]. In the case of E4-deleted vectors long-term gene expression is dependent on both the promoter used to control expression and the context of the E4 region [110]. Interpretation of the earlier v^ork on deletions in the E2 and E4 regions were complicated by the immunogenic transgenes w^hich w êre used in these studies [117, 118]. Another version of a second-generation vector w âs generated using a nonimmunogenic protein the hCFTR [119]. The vector contained w^ild-type E2 and E4 w îth a partial deletion in the E3 region and w^hen instilled into the lungs of various strains of immunocompetent mice persistent transgene expression w âs mea sured in lung tissue up to 70 days. In this vector the persistence of transgene expression w âs attributed in part to the CMV enhancer-promoter used for transgene expression in conjunction with a wild-type E4 region [110, 119]. Other groups have also reported on the optimization of vectors deleted for El and DNA polymerase. This significantly modified vector expressing the highly immunogenic P-gal transgene was shown to persist in the livers of immunocompetent mice for up to 2 months [120]. Such a vector could have broad benefits for use in human gene therapy in which the encoded transgene may be seen as a neoantigen by the human immune system. 3. Helper-Dependent Vectors More recently, vectors deleted for all viral coding sequences (helper- dependent or "gutless" Ad vectors) have been developed, so that leaky expression of viral protein is eliminated [121-124] (see Chapter 15). Such a helper-dependent vector has been generated which contains the entire human alpha 1-anti-trypsin gene under the control of a tissue (liver)-specific pro moter. This vector results in more than 1 year of stable expression, provides supraphysiological levels of hAAT in the mouse, and demonstrates less hepato- toxicity compared with first-generation vectors [125, 126]. Although many of the advantages of this gutless vector can be attributed to elimination of leaky viral gene expression it was later shown that the inclusion of a tissue-specific promoter also helped reduce the development of a host immune response to the transgene [127]. 4. Tissue-Specific Promoters The use of tissue-specific promoters may be helpful in avoiding host immune responses to transgenic proteins in human gene therapy, the rationale behind this being that expression in antigen-presenting cells would be greatly reduced. Thus the transcriptional unit responsible for expression of the trans gene in an adenoviral vector could have a substantial effect on the nature of the ensuing immune response. Most experiments have used constitutively active promoters that may express efficiently in dendritic cells. A reduced immune response is seen with vectors that contain more specific promoters, which may not express efficiently in antigen presenting cells [127]. For example, when a helper-dependent Ad vector expressing the hAAT cDNA from a liver-specific 13. Humoral Immune Response 3 9 3 promoter was used to express hAAT in C3H/HeJ mice, anti-hAAT antibodies did not develop and long-term expression of hAAT resulted [128]. In contrast, use of a non-liver-specific promoter to drive expression of the same transgene resulted in antibody production to hAAT in the same mouse strain. Thus, vector-specific differences in transgene expression w^ithin APCs due to choice of promoter could explain some of the variation in immune responses that have characterized in vivo applications of gene therapy vectors. 5. E3 Region In addition to making progressive deletions of the adenoviral backbone other groups have coexpressed the Ad early region (E3) v^ith the transgene of interest [128]. Injection of Ad-overexpressing E3-encoded gene products leads to inhibition of cytotoxic activity tov^ard Ad-infected cells in addition to marked down regulation of antibody formation to structural viral proteins. Genes encoded by the E3 region downregulate surface MHC class I expression, which in turn interferes with presentation of viral peptides and reduced CD8-h cytotoxic T lymphocytes. Thus the absence of Ad-specific CTLs in animals injected with this vector was expected. What was unexpected was the inhibition of a humoral antibody response to this E3 overexpressing vector. The authors suggest that transduction of an E3 containing Ad vector into liver cells in the absence of CTL or TNFa-induced cytolysis may result in poor antigen release, consequently there is little antigen presentation by APCs to initiate an antibody response. TNFa is one of the cytokines that controls dendritic cell maturation and migration; thus, early antigen presentation may be downregulated by inhibition of this cytokine by E3 proteins [128]. However, other studies using an Ad vector overexpressing the herpes simplex ICP47 gene suggest that use of any vector that downregulates MHC class I presentation should be assessed carefully [129]. Coexpression of ICP47 has a similar result to expression of E3 in that there is downregulation of MHC class I presentation. Administration of this vector to the lungs of rhesus mon keys inhibited the generation of Ad-specific CTLs. However, natural killer cell activity was enhanced, suggesting that strategies to protect the Ad-transduced cell without interfering with MHC class 1 expression should also be explored. B. Species and Strain To date the majority of animal studies evaluating the efficacy of aden ovirus vectors have been performed with vectors expressing bacterial or human proteins. Clearly these proteins also constitute potential antigens recognized as foreign by the host immune response. Administration of vectors encoding the murine erythropoietin resulted in long-lasting elevated hematocrit lev els in mice [130]. In contrast, injection of adenoviruses carrying the human erythropoietin induced a strong immune response directed against the human protein, which resulted in transient expression of the transgene. 3 9 4 Catherine O'Riordan Most in vivo studies are performed in inbred mouse strains of various MHC haplotypes whose immune systems might react differently to a given antigen. Inbred immunocompetent C57BL/6 mice have been a favored strain to study transgene expression of human blood coagulation Factor IX from viral vectors. This is in part because systemic expression of the secreted protein is not limited by antibody responses following intravenous (iv) injection of vector. Importantly iv injection of an Ad vector results in sustained expression of human FIX in normal or hemophilic C57B1/6 mice, while antibodies against FIX develop in other strains [131, 132]. A similar observation was seen with an Ad vector encoding human Factor VIII under the control of a liver-specific promoter following treatment of hemophilic C57B1/6 mice. High-level human FVIII expression was detected in the serum of the mice for over 5 months with no antibody production against the transgene [133]. In contrast treatment of FVIII-deficient hemophilic dogs with an Ad vector encoding human FVIII resulted in a strong antibody response directed to the human protein [133]. A reason for the difference in antibody response between different mouse strains was thought to be due to the MHC haplotype. C57BL/6 mice (haplotype H-2b) lack MHC class II allele IE, and may therefore have some deficiency in humoral immune responses. However, mice of another strain with the same haplotype, and therefore the same lack of the IE allele, did mount an immune response to human FIX following systemic administration of a similar adenoviral vector. This suggests that the data produced in studies based on C57BL/6 mice often cannot be extrapolated to other species. The mechanism of tolerance to FIX or FVIII by iv injection of adenoviral vector in C57BL/6 mice remains elusive, but illustrates the difficulty of extrapolation of results obtained in inbred strains of mice and highlights the importance of studies in other animal models [132]. Similarly, others have reported [109, 134] that intravenous administra tion of an El-deleted adenovirus vector carrying the human alpha-antitrypsin (hAAT) cDNA leads to a strain-related variation in persistence of expression of transgene. Transient expression of hAAT was seen in C3H/HeJ and Balb C mice with longer persistence of expression seen in C57B1/6 mice [109, 134]. Persistence was shown to correlate with poor anti-hAAT antibody formation in these mouse strains while Balb C and C3 H mice developed significant levels of anti-hAAT antibodies which resulted in a corresponding disappearance of hAAT in the serum [134]. In contrast, when a helper-dependent Ad vector expressing the hAAT cDNA from a liver-specific promoter was used C3H/HeJ mice failed to develop antibodies and demonstrated long-term expression of hAAT [127]. Careful identification and characterization of the host factors involved in the formation of anti-transgene antibody responses should provide insight into the development of useful gene therapy systems for the treatment of patients with genetic diseases involving null mutations. A good understanding 13. Humoral Immune Response 3 9 5 of these immune responses is critical to the appropriate interpretation of many previous gene therapy studies and to the design of future studies. Ahhough new generations of adenoviral vectors may offer many advantages compared to first-generation vectors in terms of persistence of transgene expression, it is important to establish an improved experimental paradigm to evaluate the effect of vector modifications on transgene expression, especially in the context of a highly immunogenic transgene. C. Route of Delivery 1. Intravenous versus Intraperitoneal Studies by Gahery-Segard et al. [135] investigating the humoral immune response to Ad capsid components demonstrated that routes of immunization modulate virus-induced Ig subclass shifts. Two routes of immunization, intra venous (iv) and intraperitoneal (ip) w êre compared for the response induced against the adenovirus particle in particular the three major components of the viral capsid, hexon, penton base, and fiber. The molecular components of the viral capsid are differentially recognized depending on the route of administration. The sera from mice immunized ip recognized only the hexon protein and a preferential switch to the IgG2a subclass was obtained. The sera from mice immunized iv had a more complex response. At the beginning of the response an isotype bias toward the IgG2a subclass was observed, but the isotype distribution changed during the period of the response. Neutralizing activity was maximum 45 days after immunization by both routes. However, iv serum displayed a higher neutralizing activity than ip serum, while the two routes of immunization did not induce the same IgG isotypes to neutralize viral infectivity. 2. Lung Instillation The delivery of adenoviral vectors to the lung has received much attention due to the concerted efforts of many groups to develop gene therapy vectors for the treatment of lung disorders such
as cystic fibrosis. To correct the CFTR defect, CFTR cDNA needs to be delivered to the respiratory epithelium in situ and must direct gene expression independent of cell division. Many groups have shown that Ad vectors can deliver CFTR cDNA to airway epithelial cells, leading to protein expression [105, 108, 136] and correction of the CF phenotype in vivo and in vitro [137-139], although the efficiency of repeated Ad administrations is diminished by the development of serum and mucosal neutralizing antibodies to the Ad vector [140]. In a study with repeat administrations to the nonhuman primate lung both IgG and sIgA antibodies against Ad2 were detected and it was shown that sIgA alone can contribute to neutralization of the infectivity of Ad particles entering the lung [105, 140, 141]. 3 9 6 Catherine O^Riordan 3. Delivery to the Brain Interestingly, injection of an El-deleted adenovirus into the brain trig gered a humoral immune response to the adenovirus and its gene products but no neutralizing antibody was detected thus repeat administration of the adenovirus w âs possible [142]. The authors claim that one reason for the lack of neutralizing activity may be due to the relative immunological privilege in the brain. Similarly, when adenovirus vectors are injected into the subretinal space, which is also considered to be immunologically privileged, they do not elicit humoral immune responses and repeated administration of adenovirus vectors is possible [143]. Thus there are regions within the body that may be more resistant to immune clearance on the basis of their anatomic structure. The retina for example, which is a derivative of the central nervous system, has the equivalent of a blood-brain barrier. In another report intranasal immunization of mice with wild-type adenovirus 1 month before intratumoral administration with an Ad vector of the same serotype did not efficiently inhibit repeat administration to the tumor [144]. It is likely that the structural integrity of the tumor or the extracellular matrix around the tumor presented a barrier to the neutralizing antibodies. 4. Intramuscular Delivery More recently it was shown that effective repeat administration of adenovirus vectors to muscle was not hindered by the presence of neutralizing antibodies in the serum [145]. The authors reasoned that the concentration of adenovirus-specific neutralizing antibodies in the muscle may be considerably lower than in the serum, thus permitting effective multiple dosing to the muscle. The ability to repeat dose to the muscle has significant implications for cardiovascular gene therapy where it has been shown that intramuscular administration of adenovirus vectors expressing vascular endothelial growth factor (VEGF)-stimulated angiogenesis in hind-limb ischemia in rats [146]. Thus, the ability to repeat dose to the muscle for both peripheral and coronary vascular disease could significantly improve the efficacy of gene therapies for cardiovascular disorders. VI . Immune Response to Adenoviral-Based Vectors in Humans Host immune response can play a significant role in the outcome of in vivo gene therapy. Experiments with adenoviral vectors clearly demonstrate the development of neutralizing antibodies that block readministration and cellular responses that extinguish gene expression. However, most of the work described in this field relates to animal models that are naive to the virus. This will not be the case in humans, many of whom have been exposed 13. Humoral Immune Response 3 9 7 to Ad due to a naturally acquired infection. A study performed by Chirmule et al. [147] surveyed normal subjects and cystic fibrosis patients to demonstrate the relevance of pre-existing immunity to Ad to the outcome of in vivo gene therapy. They found that antibodies reactive to Ad capsid proteins were present in 97% of individuals; however, serum from only SS% of subjects actually neutralized Ad infection in vitro. Due to this discrepancy between seropositivity and neutralization of virus in vitro, the authors suggest that human trials should include all patients irrespective of in vitro measurements of preexisting immunity. This conclusion is also based on results from a Phase 1 gene therapy clinical trial for localized mesothelioma where it was shown that preexisting humoral immune responses did not preclude gene transfer [148], Similarly, in human clinical trials where an adenovirus vector encoding hCFTR was repeatedly administered to the nasal epithelium of patients with cystic fibrosis [149], a complex immune response ensued which varied from patient to patient. Importantly, the pattern of the immune response did not differentiate patients with either large or absent correction of the CF defect following exposure to an Ad/CFTR vector [149]. Thus, what are the strengths and the limits of using experimental animals to predict human responses to gene transfer vectors? The intensity and the nature of the anti-Ad humoral immune response in experimental animals is dependent on the dose and on the route of administration of the vector. But is this the case in humans? To address this question a study was designed to determine the variability of human systemic humoral immune responses to adenovirus administered to different organs [150]. The study aimed to determine (a) if the administration of Ad vectors to humans always produced systemic anti-Ad neutralizing antibodies, (b) if the extent of the neutralizing antibody response depended on the route of administration, (c) if the systemic anti-Ad humoral response was dose dependent, and (d) how much preexisting anti-Ad antibodies influence the subsequent humoral response to Ad vector administrations. Vectors were administered to the airway epithelium of individuals with cystic fibrosis, or individuals with liver metastatic tumors from colon cancer, or the skin of healthy normal individuals or the myocardium of individuals with coronary artery disease. Interestingly, the administration of the Ad vector to the bronchial epithelium of CF patients yielded the lowest antibody response, while direct injection of the Ad vector to colon carcinoma metastases in the liver resulted in the most vigorous antibody response. For those individuals that received Ad vector intradermally or intramyocardially there was a varying response, with some individuals having no increase in anti-Ad neutralizing anti body titer with others having a robust response. The most significant observa tion was the strong correlation between preexisting neutralizing antibodies and the likelihood that an individual would mount a higher titer following vector administration. Irrespective of route of administration and underlying disease 3 9 8 Catherine O'Riordan state individuals with higher baseUne anti-Ad neutraUzing antibodies mounted a higher neutraUzing antibody response after exposure to an Ad vector. How^ much does disease state modify the systemic humoral host response to Ad vectors? In the case of individuals v^ith CF they yielded the low^est antibody response following intrabronchial administration of an Ad vector. The authors suggest that the minimal responses to bronchial administration of Ad in CF patients could be due to the fact that the airway epithelium of these individuals is covered by mucus, which can preclude efficient Ad vector infection of the airway epithelial cells. To address this concern the same authors have just recently reported the results of a clinical trial using an El-E3-deleted Ad vector, which was repeat edly administered to the lung airway of six normal individuals [151], There were minimal to no systemic or local (epithelial lining fluid) anti-Ad neutraliz ing antibodies in the normal individuals following vector administration. This is a significant observation as in this situation there are no adverse conditions in the normal lung unlike the CF lung, that might affect host immune responses. Thus, in contrast to experimental animals where it has been shown that there is a robust anti-Ad response following repeated vector administration to the lung, it appears that for Ad-based vectors for use in human lung gene therapy this maybe less of a problem. VII. Conclusion There are several important conclusions on host responses to adenovirus- based vectors that can be made on the basis of extensive studies in experimental animals and some limited studies in humans. In particular, genetic composition of the Ad vector, dosing regimen, and routes of administration all modulate the humoral immune response generated to an adenoviral vector. In addition, fac tors other than the Ad-specific adaptive immune responses play a significant role in modulating gene expression following in vivo Ad transduction. These factors are related to host species and the immediate innate host responses following vector administration. Importantly, responses in experimental animals are not always predictive of responses in a human and care should be taken in extrap olating data from animal studies to predict outcomes in human clinical trials. Several strategies that may be used to overcome host humoral immune responses to Ad-based vectors have been developed. Administration of different-serotype Ad vectors may circumvent the local host immunity elicited by a first vector administration. Vectors with capsid components modified at specific capsid sites known to be targets for the anti-Ad-neutralizing antibodies might be useful in overcoming preexisting, serotype-specific anti-Ad immunity. In addition strategies to suppress humoral immune response in the host such as cytokine treatment, use of monoclonal antibodies, T-cell depletion strategies, immunosuppressive drugs have all shown efficacy as being effective in allowing 13. Humoral Immune Response 3 9 9 repeat administration of Ad vectors. To date, all of these strategies have been evaluated only in experimental animals, while their efficacy and safety have yet to be tested in humans. References 1. De Jong, J. C , Wermenbol, A. G., Verweij-Uijterwaal, M. W., Slaterus, K. W., Wertheim- Van Dillen, P., Van Doornum, G. J., Khoo, S.H., and Hierholzer, J. C. (1999). Adenoviruses from human Immunodeficiency virus-infected individuals, including two strains that rep resent new^ candidate serotypes Ad50 and Ad51 of species Bl and D respectively. / . Clin. Microbiol. 37, 3940-3945. 2. Benko, M., Harrach, B., and Russell, W. C. (1999). Family adenoviridae. In "Virus Tax onomy" (M. H. Van Regenmortel, C. M. Fauquet, D. H. L. Bishop, E. B. Carstens, M. K. Estes, S. M. Lemon, J. Maniloff, M. A. Mayo, D. J. McGeoch, C. R. Pringle, and R. B. Wick- ner, Eds.), Seventh Report of the International Committee on Taxonomy of Viruses, pp. 227-238. Academic Press, New York. 3. Francki, R. I. B., Fauquet, C. M., Knudson, D. L., and Brow^n, F. (1991). Classification and nomenclature of viruses. Fifth report of the International Committee on Taxonomy of viruses. Arch. Virol. 1991 (Suppl. 2), 140-144. 4. Kjellen, L., and Pereira, H. G. (1968). Role of adenovirus antigens in the induction of virus neutralizing antibody./. Gen. Virol. 2, 177-185. 5. Wilcox, N., and Mautner, V. (1976). Antigenic determinants of adenovirus capsids. II. Homogeneity of hexons, and accessibility of their determinants, in the virion. / . Immunol. 116,25-29. 6. Ginsberg, H. S. (Ed.) (1984). "The Adenoviruses." Plenum, Nev̂ ^ York. 7. Norrby, E., and Wadell, G. (1969). Immunological relationship betw^een hexons of certain human adenoviruses./. Virol. 4, 663-670. 8. Cornick, G., Sigler, P. B., and Ginsberg, H. S. (1971). Characterization of crystals of type 5 adenovirus hexon./. Mol. Biol. 57, 397-401. 9. Franklin, R., Petterson, U., Akervall, K., Strandberg, B., and Philipson, L. (1971). Structural proteins of adenovirus V. Size and structure of the adenovirus type 2 hexon. / . Mol. Biol. 57,383-395. 10. Pereira, H. G., Valentine, R. C , and Russell, W. C. (1968). Crystallization of an adenovirus protein (the hexon). Nature 219, 946-947. 11. Rux, J. J., and Burnett, R. M. (2000). Type-Specific epitope locations revealed by X-ray crystallographic study of adenovirus type 5 hexon. Mol. Ther. 1, 18-29. 12. Athappilly, F. K., Murali, R., Rux, J.J. Cai, Z., and Burnett, R. M. (1994). The refined crystal structure of hexon, the major coat protein of adenovirus type 2, at 2.9 A resolution. / . Mol. Biol. 242, 430-452. 13. Roberts, M. M., White, J. L., Grutter, M. G., and Burnett, R. M. (1986). Three dimensional structure of the adenovirus major coat protein hexon. Science 232, 1148-1151. 14. Crav\^ford-Micksa, L., and Schnurr, D. P. (1996). Analysis of 15 adenoviruses hexon proteins reveals the location and structure of seven hypervariable regions containing serotype-specific residues./. Virol. 70, 1836-1844. 15. Eiz, B., Adrian, T., and Pring-Akerblom, P. (1995). Immunological adenovirus variant strains of subgenus D: Comparison of the hexon and fiber sequences. Virology 213, 313-320. 16. Kinloch, R., Mackay, N., and Mautner, V. (1984). Adenovirus hexon: Sequence comparison of subgroup C serotypes 2 and 5. / . Biol. Chem. 259, 6431-6436. 17. Pring-Akerblom, P., and Adrian, T. (1993). The hexon genes of adenoviruses of subgenus C: Comparison of variable regions. Res. Virol. 144, 117-127. 4 0 0 Catherine O'Riordan 18. Pring-Akerblom,
P., Trijssenaar, F. E., and Adrian, T. (1995). Sequence characterization and comparison of human adenovirus subgenus B and E hexons. Virology 212, 232-236. 19. Toogood, C. I. A., Crompton, J., and Hay, R. T. (1992). Antipeptide antisera define neu- trahzing epitopes on the adenovirus hexon. / . Gen. Virol. 73, 1429-1435. 20. Sundquist, B., Petterson, U., Thelander, R., and Philipson, L. (1973). Structural proteins of adenoviruses. IX. Molecular weight and subunit composition of adenovirus type 2 fibre, y/ro/ogy 51, 252-256. 21. Chatellard, C. and Chroboczek, J. (1989) Synthesis of human adenovirus type 2 fibre protein in E. Coh. cells. Gene 81, 267-274. 22. Stouten, P. F., Sander, C , Ruigrok, R. W. H., and Cusack, S. (1992). A new triple-helical model for the adenovirus fibre shaft. / . Mol. Biol. 226, 1073-1084. 23. Philipson, L., Lonberg-Holm, K., and Petterson, U. (1968). Virus-receptor interaction in an adenovirus system./. Virol. 35, 204-215. 24. Wickham, T. J., Mathias, P., Cheresh, D. A., and Nemerow, G. R. (1993). Integrins avp3 and OLv^S promote adenovirus internalization but not virus attachment. Cell 73, 309-319. 25. Bai, M., Campisi, L., and Freimuth, P. (1994). Vitronectin receptor antibodies inhibit infection of HeLa and A549 cells by adenovirus type 12 but not by adenovirus type 2. / . Virol. 68, 5929-5932. 26. Valentine, R. C , and Pereira, H. G. (1965). Antigens and the structure of the adenoviruses. / . Mol. Biol. 13, 13-20. 27. Green, N. M., Wrigley, H. G., Russell, W. C., Martin, S. R., and McLachlan, A. (1983). Evidence for a repeating cross-beta sheet structure in the adenovirus fibre. EMBO J. 2, 1357-1365. 28. Kidd, A. H., Chroboczek, J., Cusack, S., and Ruigrok, R. W. H. (1993). Adenovirus type 40 virions contain two distinct fibres. Virology 192, 73-84. 29. Stevenson, S. C, RoUence, M., White, B., Weaver, L., and McClelland, A. (1995). Human adenovirus serotypes 3 and 5 bind to two different cellular receptors via the fiber head domain./. Virol. 69, 2850-2857. 30. Bergelson, J. M., Cunningham, J. A., Droguett, G., Kurt-Jones, E. A., Krithivas, JA., Hong, J. S., Horwitz, M. S., Crowell, R. L., and Finberg, R. W. (1997). Isolation of a common receptor for coxsackie B viruses and adenoviruses 2 and 5. Science 275, 1320-1323. 31. Xia, D., Henry, L. J., Gerard, R. D., and Deisenhofer, J. (1994). Crystal structure of the receptor-binding domain of adenovirus type 5 fibre protein at 1.7 A resolution. Structure 2, 1259-1270. 32. Watson, G., Burdon, M. G., and Russell, W. C. (1988). An antigenic analysis of the adenovirus type 2 fibre polypeptide./. Gen. Virol. 69, 525-535. 33. Fender, P., Kidd, A. H., Brebant, R., Oberg, M., Drouet, E., and Chroboczek, J. (1995). Antigenic sites on the receptor binding domain of human adenovirus type 2 fiber. Virology 214,110-117. 34. Liebermann, H., Lotz, K., Mentel, R., Bauer, U., and Seidel, W. (2001). Mapping of linear epitopes on fibre knob of human adenovirus serotype 5. Virus Res. 73, 145-151. 35. Wohlfart, C , (1988). Neutralization of adenoviruses: kinetics, stoichiometry, and mecha nisms./. Virol. 62, 2321-2328. 36. Hong, S. S., Bardy, M., Monteil, M., Gay, B., Denesvre, C , Tournier, J., Martin, G., Eliot, M., and Boulanger, P. (2000), Immunoreactive domains and integrin-binding motifs in adenovirus penton base capsomer. Virol. Immunol. 3, 353-371. 37. Stewart, P. L., Chiu, C. Y., Huang, S., Muir, T., Zhao, Y., Chait, B., Mathias, P., and Nemerow, G. R. (1997). Cryo-EM visualization of an exposed RGD epitope on adenovirus that escapes antibody neutralization. EMBO ]. 16, 1189-1198. 38. Gall, J. G. D., Crystal, R. G., and Falck-Pedersen, E. (1998). Construction and character ization of hexo-chimeric adenoviruses: Specification of adenovirus serotype. / . Virol. 72, 10,260-10,264. 13. Humoral Immune Response 4 0 1 39. Ostapchuk, P., and Hearing, P. (2001). Pseudopackaging of adenovirus type 5 genomes into capsids containing the hexon proteins of adenovirus serotypes B, D, or E. / . Virol. 75, 4 5 - 5 1 . 40. Roy, S., Shirley, P. M., McClelland, A., and Kaleko, M. (1998). Circumvention to Immunity to the adenovirus major coat protein hexon. / . Virol. 72, 6875-6879. 41. Bailey, A., and Mautner, V. (1994). Phylogenetic relationships among adenovirus serotypes, y/ro/og)/205, 438-452. 42. Crompton, J., Toogood, C. I., Wallis, N., and Hay, R. T. (1994). Expression of a foreign epitope on the surface of the adenovirus hexon./. Gen. Virol. 75, 133-139. 43. Bai, M., Harfe, B., and Freimuth, P. (1993). Mutations that alter an Arg-Gly-Asp (ROD) sequence in the adenovirus type 2 penton base protein abolish its cell-rounding activity and delay virus reproduction in flat cells. / . Virol. 67, 5198-5205. 44. Davison, E., Diaz, R. M., Hart, I. R., Santis, G., and Marshall, J. F. (1997). Integrin a5bl- mediated adenovirus infection is enhanced by the integrin activating antibody TS2/16. / . Virol. 71, 6204-6207. 45. Huang, S., Kamata, T., Takada, Y., Ruggeri, Z. M., and Nemerow, G. R. (1996). Adenovirus interaction with distinct integrins mediates separate events in cell entry and gene delivery to hematopoietic cells. / . Virol. 70, 4502-4508. 46. Wickham, T. J., Filardo, E. J., Cheresh, D. A., and Nemerow, G. R. (1994). Integrin alpha V beta 5 selectively promotes adenovirus mediated cell membrane permeabilization. / . Cell. Biol. 127, 257-264. 47. Blumenthal, R., Seth, P., Willingham, M. C , and Pastan, I. (1986). pH-dependent lysis of liposomes by adenovirus. Biochemistry 25, 2231-2237. 48. Leopold, P. L., Ferris, B., Grinberg, I., Worgall, S., Hackett, N. R., and Crystal, R. G. (1998). Fluorescent virionic dynamic tracking of the pathway of adenoviral gene transfer vectors in living cells. Hum. Gene Ther. 9, 367-378. 49. Wisnivesky, J. P., Leopold, P. L., and Crystal, R. G. (1999). Specific binding of the adenovirus capsid to the nuclear envelope. Hum. Gene Ther. 10, 2187-2195. 50. Saphire, A. C , Guan, T., Schirmer, E. C , Nemerow, G. R., and Gerace, L., (2000). Nuclear import of adenovirus DNA in vitro involves the nuclear protein import pathway and hsc 70. / . Biol. Chem. 275, 4298-4304. 51. Defer, C , Belin, M. T., Caillet-Boudin, M. L., and Boulanger, P. (1990) Human adenovirus host cell interactions comparative study with members of subgroups B and C. / . Virol. 64, 3561-3673. 52. Varga, M.J,, Bergman, T., and Everitt, E., (1990). Antibodies with specificities against a dispase-produced 15-kilodalton hexon fragment neutralize adenovirus type 2 infectivity. / . V/ro/. 64,4217-4225. 53. Pettersson, U., Philipson, L., and Hoglund, S. (1968). Structural proteins of adenovirus 11. Purification and characterization of the adenovirus type 2 fibre antigen. Virology 35, 204-215. 54. Mautner, V., and Wilcox, H. N. (1974). Adenovirus antigens: a model system in mice for subunit vaccination./. Gen. Virol. 25, 325-336. 55. Gall, J., Kass-Eisler, Leinwand, L., and Falck-Pedersen, E. (1996). Adenovirus type 5 and 7 capsid chimera: Fiber replacement alters receptor tropism without affecting primary immune neutralization epitopes./. Virol. 70,2116-2123. 56. Gahery-Segard, H., Farace, F., Godfrin, D., Gaston, J., Lengagne, R., Tursz, T., Boulanger, P., and Guillet, J. (1998). Immune response to recombinant capsid proteins of adenovirus in humans: Antifiber and anti-penton base antibodies have synergistic effect on neutralizing activity./. Virol. 72,2388-2397. 57. Pastore, L., Morral, N., Zhou, H., Garcia, R., Parks, R. J. Kochanek, S., Graham, F., Lee, B., and Beaudet, A. L. (1999). Use of a liver-specific, promoter reduces immune response to the transgene in adenoviral vectors. Hum. Gene Ther. 10, 1773-1781. 4 0 2 Catherine O'Riordan 58. Worgall, S., Wolff, G., Falck-Pedersen, E., and Crystal, R. (1997). Innate immune mecha nisms dominate elimination of adenoviral vectors follov^ing in vivo administration. Hum. Gene Ther. 8, 37-44. 59. Yang, Y., Li, Q., Ertl, H. C. J., and Wilson, J. M. (1995). Cellular and humoral immune responses to viral antigens create barriers to lung-directed gene therapy with recombinant adenoviruses./. Virol. 69,2004-2015. 59a. Kaplan, J. M., Armentano, D., Sparer, T. E., Wynn, S. G., Peterson, P. A., Wadsw^orth, S. C , Couture, K. K., Pennington, S. E., St. George, J. A., Gooding, L. R., and Smith, A. E. (1997) Characterization of factors involved in modulating persistence of transgene expression from recombinant adenovirus in the mouse lung. Hum. Gene Ther. 8, 45-56. 60. Yang, Y., Xiang, Z., Ertl, H., and Wilson, J. M. (1995). Upregulation of class I major histocompatibility complex antigens by interferon gamma is necessary for T-cell mediated elimination of recombinant adenovirus-infected hepatocytes in vivo. Proc. Natl. Acad. Sci. USA 92, 7257-7261. 61. Linsley, P. S., and Ledbetter, J. A. (1993). The role of the CD28 receptor during T cell responses to antigen. Annu. Rev. Immunol. 11, 191-212. 62. Durie, F. H., Foy, T. M., Masters, S. R., Laman, J. D., and Noelle, R. J. (1994). The role of CD40 in the regulation of humoral and cell-mediated immunity. Immunol. Today 15, 406-411. 63. Yang, Y., Nunes, F. A., Berencsi, K., Furth, E. E, Gonczol, E., and Wilson, J. M. (1994). Cellular immunity to viral antigens limits El-deleted adenoviruses for gene therapy. Proc. Natl. Acad. Sci. USA 91, 4407-4411. 64. Dai, Y., Schwarz, E. M., Gu, D., Zhang, W., Sarvetnick, N., and Verma, I. M. (1995). Cellular and humoral immune responses to adenoviral vectors containing factor IX gene: Tolerization of factor IX and vector antigens allows for long-term expression. Proc. Natl. Acad. Sci. USA 92, 1401-1405. 65. Fang, B., Eisensmith, H., Wang, H., Kay, M. A., Cross, R. E., Landen, C. N., Gordon, G., Bellinger, D. A., Read, M. S., Hu, P. C , Brinkhous, K. M., and Woo, S. L. C. (1995). Gene therapy for hemophilia B: host immunosuppression prolongs the therapeutic effect of adenovirus-mediated factor IX expression. Hum. Gene Ther. 6, 1039-1044. 66. Joos, K., Yang, Y., and Wilson, J. M. (1996). Cyclophosphamide diminishes inflammation and prolongs transgene expression following delivery of adenoviral vectors to mouse liver and lung. Hum. Gene Ther. 7, 1555-1566. 67. Berge, R. J. (1994). Immunosuppressive drugs in clinical medicine. Netherlands./. Med. 45, 329-338. 68. Clarke, L., and Waxman, D. J. (1989). Oxidative metabolism of cyclophosphamide: Iden tification of the hepatic monooxygenase catalysts of drug activation. Cancer Res. 49, 2344-2350. 69. Borel, J. K., and Gunn, H. C. (1986). Cyclosporine as a new approach to therapy of autoimmune diseases. Ann. NY Acad. Sci. 475, 307-318. 70. Bouvet, M., Fang, B., Ekmekcioglu, S., Ji, L., Bucana, C. D., Hamada, K., Grimm, E. A., and Roth, J. A. (1998). Suppression of the immune response to an adenovirus vector and enhancement of intratumoral transgene expression by low-dose etoposide. Gene Ther. 5, 189-195. 71. Lock, R. B., and Stribinskine, L. (1996). Dual modes of death induced by etoposide in human epithelial tumor cells allow Bcl-2 to inhibit apoptosis without affecting clonogenic survival. Cancer Res. 56, 4006-4012. 72. Smith, T. A. G., White, B. D., Gardner, J. M., Kaleko, M., and McClelland, A. (1996). Transient immunosuppression permits successful repetitive intravenous administration of an adenovirus vector. Gene Ther. 3, 496-502. 13. Humoral Immune Response 4 0 3 73. Kaplan, J. M., and Smith, A. E. (1997) Transient immunosuppression with deoxysperguaUn improves longevity of transgene expression and ability to readminister adenoviral vector to the mouse lung. Hum. Gene Ther. 8, 1095-1104. 74. Kolls, J. K., Lei, D., Odom, G., Nelson, S., Summer, W., Gerber, M. A., and Shellito, J. E. (1996). Use of transient CD4 lymphocyte depletion to prolong transgene expression of El-deleted adenoviral vectors. Hum. Gene Ther. 7, 489-497. 75. Yang, Y., Greenough, K., and Wilson, J. M. (1996). Transient immune blockade prevents formation of neutralizing antibody to recombinant adenovirus and allov^̂ s repeated gene transfer to mouse liver. Gene Ther. 3, 412-420. 76. Ye, X., Robinson, M. B., Pabin, C., Batshav^, M. L., and Wilson, J. M. (2000). Transient depletion of CD4 lymphocyte improves efficacy of repeated administration of recombinant adenovirus in the ornithine transcarbamylase deficient sparse fur mouse. Gene Ther. 7, 1761-1767. 77. Yang, Y., Trinchieri, G., and Wilson, J. M. (1995). Recombinant IL-12 prevents formation of blocking IgA antibodies to recombinant adenovirus and allows repeated gene therapy to mouse lung. Nat. Med. 1, 890-893. 78. Kay, M. A., Holterman, A-X., Meuse, L., Gown, A., Ochs, H. D., Linsley, P. S., and Wilson, C. B. (1995). Long-term hepatic adenovirus-mediated gene expression in mice following CTLA4Ig administration. Nat. Genet. 11, 191-197. 79. Jooss, K., Turka, L. A., and Wilson,}. M. (1998). Blunting of immune responses to adenoviral vectors in mouse liver and lung with CTLA4Ig. Gene Ther. 5, 309-319. 80. Ideguichi, M., Kajiwara, K., Yoshikawa, K., Uchida,
Michou, A. L, Santoro, L., Christ, M., Julliard, V., Pavirani, A., and Mehtafi, M. (1997). Adenovirus-mediated gene transfer: Influence of transgene, mouse strain, and type of immune response on persistence of transgene expression. Gene Ther. 4, 473-482. 132. Fields, P. A., Armstrong, E., Hagstrom, J. N., Arruda, V. R., Murphy, M. L., Farrell, J. P., High, K. A., and Herzog, R. W. (2001). Intravenous administration of an El/E3-deleted adenoviral vector induces tolerance to Factor IX in C57B1/6 mice. Gene Ther. 8, 354-361. 133. Connelly, S., Gardner, J. M., McClelland, A., and Kaleko, M. (1996). High-level tissue specific expression of functional human Factor VIII in mice. Hum. Gene Ther. 7, 183-195. 134. Schowalter, D. B., Himeda, C, L., Winther, B. L., Wilson, C. B., and Kay, M. (1999). Implication of interfering antibody formation and apoptosis as two different mechanisms leading to variable duration of adenovirus mediated transgene expression in immune- competent mice. / . Virol. 73, 4755-4766. 135. Gahery-Segard, H., Juillard, V., Gaston, J., Lengagne, R., Pavirani, A., Boulanger, P., and Guillet, J-G. (1997). Humoral immune response to the capsid components of recombinant adenoviruses: Routes of immunization modulate virus-induced Ig subclass shifts. Eur. J. Immunol. 27, 653-659. 136. Crystal, R. G., McElvaney, N. G., Rosenfeld, M. A., Chu, C. S., Mastrangeli, A., Hay, J. G., Brody, S. L., Jaffe, H. A., Eissa, N. T., and Danel, C. (1994) Administration of an adenovirus containing the human CFTR cDNA to the respiratory tract of individuals with cystic fibrosis. Nat. Genet. 8, 8-9. 137. Zabner, J., Couture, L. A., Gregory, R. J., Graham, S. M., Smith, A. E., and Welsh, M. J. (1993). Adenovirus-mediated gene transfer transiently corrects the chloride transport defect in nasal epithelia of patients with cystic fibrosis. Cell 75, 207-216. 13. Humoral Immune Response 4 0 7 138. Zabner, J., Couture, L. A., Smith, A. E., and Welsh, M.J . (1994). Correction of cAMP- stimulated fluid secretion in cystic fibrosis airway epithelia: Efficiency of adenovirus-mediated gene transfer in vitro. Hum. Gene Ther. 5, 585-593. 139. Rosenfeld, M.A., Yoshimura, K., Trapnell, B.C., Yoneyama, K., Rosenthal, E.R., Dalemans, W., Fukayama, M., Bargon, J., Stier, L.E., Stratford-Perricaudet, L., Perricaudet, M., Guggino, W. B., Pavirani, A., Lecocq, J-P., and Crystal, R. (1992). In vivo transfer of the human cystic fibrosis transmembrane conductance regulator gene to the airw^ay epithelium. Ce//68, 143-155. 140. Kaplan, J. M., St. George, J. A., Pennington, S. E., Keyes, L. D., Johnson, R. P., Wadsv^orth, S. C , and Smith, A. E. (1996). Humoral and cellular immune responses of nonhuman primates to long-term repeated lung exposure to Ad2-CFTR-2. Gene Ther. 3, 117-127. 141. Van Ginkel, F. W., Liu, C , Simecka, J. W., Dong, J-Y., Greenway, T., Frizzell, R. A., Kiyono, H., McGhee, J. R., and Pascual, D. W. (1995). Intratracheal gene delivery with adenoviral vector induces elevated systemic IgG and mucosal IgA antibodies to adenovirus and p-galactosidase. Hum. Gene Ther. 6, 895-903. 142. Kajiwara, K., Byrnes, A. P., Ohmoto, Y., Charlton, H. M., Wood, M. J. A., and Wood, K. J. (2000). Humoral immune responses to adenovirus vectors in the brain. / . Neurol. Immunol. 103 ,8-15 . 143. Bennett, J., Pakola, S., Zeng, Y., and Maguire, A. (1996). Humoral response after admin istration of El deleted adenoviruses: Immune privilege of the subretinal space. Hum. Gene Ther. 7, 1763-1769. 144. Bramson, J. L., Hitt, M., Gauldie, J., and Graham, F. L. (1997). Pre-existing immunity to adenovirus does not prevent tumor regression following intratumoral administration of a vector expressing IL-12 but inhibits virus dissemination. Gene Ther. 4, 1069-1076. 145. Chen, P., Kovesdi, I., and Bruder, J. T. (2000). Effective repeat administration with aden ovirus vectors to the muscle. Gene Ther. 7, 587-595. 146. Mack, C. A, Magovern, C. J., Budenbender, K. T., Patel, S. R., Schwarz, E. A., Zanzonico, P., Ferris, B., Sanborn, T,, Isom, P., Ferris, B., Sanborn, T., Isom, O. W., Crystal, R. G., and Rosengart, T. K. (1998). Salvage angiogenesis induced by adenovirus mediated gene transfer of vascular endothelial growth factor protects against vascular occlusion. / . Vase. Surg. 4, 699-709. 147. Chirmule, N., Propert, K. J., Magosin, S. A., Qian, Y., Qian, R., and Wilson, J. M. (1999). Immune responses to adenovirus and adeno-associated virus in humans. Gene Ther. 6, 1574-1583. 148. Molnar-Kimber, K. L., Sterman, D. H., Chang, M., Kang, E. H., ElBash, M., Lanuti, M., Elshami, A., Gelfand, K., Wilson, J. M., Kaiser, L. R., and Albeda, S. M. (1998). Impact of preexisting and induced humoral and cellular immune responses in an adenovirus-based gene therapy phase I clinical trial for localized mesotheHoma. Hum. Gene Ther. 9, 2121-2133. 149. Zabner, J., Ramsey, B. W., Meeker, D. P., Aitken, M. L., Balfour, R. P., Gibson, R. L., Launspach, J., Moscicki, R. A., Richards, S. M., Standaert, T. A., WiUiams-Warren, J., Wadsworth, S. C , Smith, A. E., and Welsh, M.J . (1996). Repeat administration of an adenovirus vector encoding cystic fibrosis transmembrane conductance regulator to the nasal epithelium of patients with cystic fibrosis./. Clin. Invest. 97, 1504-1511. 150. Harvey, B-G., Hackett, N. R., El-Sawy, T., Rosengart, T. K., Hirschowitz, E. A., Lieberman, M. D., Lesser, M. L., and Crystal, R. G. (1999). Variability of human systemic humoral immune responses to adenovirus gene transfer vectors administered to different organs. / . Virol. 73, 6729-6742. 151. Harvey, B-G., Hackett, N. R., Ely, S., and Crystal, R. C. (2001). Host responses and persistence of vector following intrabronchial administration of an El~/E3~ adenovirus gene transfer vector to normal individuals. Mol. Ther. 3, 206-215. C H A P T E R Novel Methods to Eliminate the Immune Response to Adenovirus Gene Therapy Huang-Ge Zhang/'^ Hui-Chen Hsu/ and John D. Mountz"̂ '̂ *Department of Medicine Division of Clinical Immunology and Rheumatology University of Alabama at Birmingham and "'"Birmingham Veterans Administration Medical Center Birmingham, Alabama I. Introduction The immune response to adenovirus (Ad) vectors or their transgene is a hmiting factor in the successful appHcation of gene therapies [1-3] (Fig. 1). Components of the adenovirus are recognized for their abihty to eUcit either a strong antigenic response or to subvert the immune response. Early proteins (E), w^hich are produced continuously by adenovirus, can both enhance and inhibit the immune response. El (including ElA and ElB) and E2 are necessary for Ad replication but also evoke a strong immune response. E4 is a p70 antigen that promotes early and late transportation of viral mRNA and is highly immunogenic. E3 can be immunogenic but also produces a 19 K product that inhibits major histocompatibility complex-I (MHC-I) transportation to the surface of antigen-presenting cells (APCs) and therefore subverts the immune response [4, 5]. Adenovirus late proteins include the penton base which is in a cryptic position but can evoke an immune response. The neutralizing antibody response is formed against the fiber knob which is also a strong antigen on the surface of adenovirus. For applications to gene therapy, the transgene expressed by the Ad vector can also evoke an immune response, especially when combined in the environment of viable transduced adenovirus. ADENOVIRAL VECTORS FOR GENE THERAPY 4 0 9 Copyright 2002, Elsevier Science (USA). All rights reserved. 410 Zhang et aL Adanovlrus CytotoKic T Cells f ^ Helper T Cells B Cells Neutralizing Ab t j ^ Prcrtitii \|MiNi# Onai}! Figure 1 Components of adenovirus that can evoke an immune response. Adenovirus expresses early proteins including E l , E2, E3, and E4, most of which can provoke a strong immune response. E3 is known to inhibit MHC class I transportation of antigens to the surface of antigen-presenting cells (APCs), and minimizes the immune response to adenovirus. Late proteins include the penton base and the fiber-knob of the adenovirus. The fiber knob can elicit a strong antigenic response and promote production of neutralizing antibody against the fiber knob. The gene therapy transgene can also evoke an immune response. II. Immune Suppression Both cellular and humoral immune responses have been implicated in the shortening of the time span of transgene expression. Transgene expression is limited by eradication of the transfected cells. Induction of anti-Ad neutral izing antibodies precludes the opportunity to readminister the gene therapy. Immunosuppressive drugs including cyclophosphamide [6], cyclosporine [7], and FK506 [8-10], have reduced the T-cell immune response. Lochmuller et al. [8] used FK506, resulting in prolonged expression of adenovirus-mediated dystrophin gene transfer in mdx adult mice for at least 2 months, even though the FK506 treatment was discontinued after 1 month. There w âs a marked reduction in inflammation and reduced levels of nitric oxide synthesase in macrophages in the muscles of such treated animals. Howell et al. [9] used the dystrophin-deficient golden retriever dog model and showed that cyclosporine significantly prolonged transgene expression after Ad-mediated expression of a truncated human dystrophin gene. Ilan et al. [10] showed that com bined immunotherapy in humans resulted in prolonged transgene expression. 14. Novel Methods to Eliminate the immune Response 4 1 1 Combined therapy with cyclosporine, azathioprin, and prednisone has been shown to reduce the immune response to adenovirus-mediated gene therapy and prolonged transgene expression. Other strategies reported to control the immune response include reduc tion of T-cell response by oral tolerance [11], thymic tolerance, anti-T-cell therapy, and anti-CD4 monoclonal antibody therapy [12-15]. Modulation of T-cell subset development after adenovirus therapy has successfully been attempted using recombinant interleukin 12 (IL-12). IL-12 activates Th-1 cells to secrete gamma interferon (IFNy) which diminishes Th-2 T-cell formation and reduces formation of neutralizing antibodies [16]. III. Immune Modulation Other specific strategies include reduction of the costimulatory signal ing activity using the CTLA4 immunoglobulin Fc (CTLA4-Ig) and blocking CD40-CD40L interaction (Fig. 2). Interaction of APC with processed ade novirus antigens or adenovirus transgene antigens with T cells through the MHC-T-cell receptor (TCR) ligation constitutes signal 1, which is an incom plete signal and can induce T-cell nonresponsiveness or anergy. A second signal can be provided by interaction with B7-1 or B7-2 (CD80/CD86) on the APC with CTLA4 (CD152) or CD28 on the T cell (Fig. 2). A soluble form of CTLA4-Ig (sCTLA4-Ig) can bind to B7-1 and B7-2 and block interactions of this molecule with membrane bound CTLA4 on T cells. The anti-adenovirus response can be decreased and there is prolonged expression of the aden ovirus transgene in the presence of either sCTLA4-Ig or in the presence of an adenovirus that expresses CTLA4-Ig (AdsCTLA4-Ig). Kay et aL [17, 18] demonstrated that systemic coadministration of recombinant adenovirus with sCTLA4-Ig leads to persistent adenovirus gene expression in mice without long-term immunosuppression. Ideguchi et al. [19] utilized local administration of adenovirus that expressed p-galactosidase as well as CTLA4-Ig (AdsCTLA4-Ig) in the central nervous system and showed that this combination decreased T-cell infiltration and also decreased the anti- adenovirus antibody titer. Expression of p-galactosidase at the injection site in the striatum and corpus callosum peaked at day 6 and remained until day 60 in both control and treated groups at about the same level despite suppression of the inflammatory response. Guibinga et al. [20] showed that the combination of CTLA4-Ig plus FK506 resulted in prolonged adenovirus vector-mediated production of dystrophin compared to treatment with either immunosup pressive alone. Schowalter et al. [21] demonstrated that murine CLTA4-Ig markedly prolonged adenovirus transgene expression in the liver and dimin ished formation of neutralizing antibodies as well as decreasing the proliferative response without causing irreversible immune suppression. Ali et al. [11] used AdsCTLA4-Ig administration in association with intraocular administration 412 Zhang ef al. sCD40«Ig <CDI52) (CD86) sCTLA4-Ig Figure 2 Role of costimulatory molecules to induce cytotoxic or helper T-cell response and promote anti-Ad antibody production. Signal 1 consists of the processed antigen presented by the MHC expressed by APCs stimulation of T cells. These antigens are recognized by specific T-cell receptors (TCRs) on T cells. This interaction is also assisted by CDS expressed on cytotoxic T cells or CD4 expressed on helper T cells that interact with MHC. In addition to this central, specific T-cell-APC interaction, costimulatory molecules are necessary for optimal T-cell response. These consist of CD28 and CTLA4 (CD! 52) expressed on T cells that interact with B7.1 and B7.2 (CD80 and CD86), expressed on APCs. This interaction can be blocked by soluble CTLA4-lg (sCTLA4-lg) that tightly binds to B7.1 and B7.2 and prevents the interaction of this molecule expressed on APCs with CD28/CTLA-4 expressed on T cells. A second costimulatory molecule pathway consists of CD40
ligand (CDl 54) expressed on T cells that interact with CD40 expressed on APCs. This interaction can be blocked by administration of a soluble CD40-lg (sCD40-lg) which binds with the CD40 ligand and prevents the interaction between CD40 and CD40 ligand. of adenovirus encoding p-galactosidase and demonstrated reduced immune response to adenovirus, as w êll as prolonged expression of the transgene in retinal cells. Kay et al. [17] show^ed that combined treatment v^ith sCTLA4-Ig and anti-CD40 ligand resulted in prolonged adenovirus-mediated gene expres sion for up to 1 year in the liver and the ability to readminister adenovirus in 50% of mice. Follow^ing readministration, there w âs persistent secondary gene expression lasting 200-300 days, and diminished spleen proliferative response, tumor necrosis factor (TNF)-a and IFNy production and decreased production of neutralizing antibodies. Chirmule et al. [23] showed that despite the absence of CD40-CD40 ligand interaction in CD40 ligand knockout mice, after administration of LacZ into the mouse lung, these mice devel oped a functional humoral response to the vector evidenced by germinal center formation and anti-adenovirus IgGl and IgA that resulted in effective neutralization of virus and prevented effective readministration of the virus. 14. Novel Methods to Eliminate the immune Response 4 1 3 Wilson et al. [24, 25] used combined treatment with an adenovirus vector expressing sCTLA4-Ig to block CD28 stimulation and a monoclonal antibody against CD40 ligand to demonstrate prolonged adenovirus transgene expres sion after intratracheal administration. In addition, secondary administration and transgene expression after secondary administration w âs prolonged in the lung, but there w âs increased reaction from the liver. These results indicate that the mechanisms limiting transgene expression in the airv^ays and the alveoli are different to those of the liver. Stein et al, [26] have shov^n that combined treatment of Ad-human factor IX (FIX) w îth an anti-CD40 ligand antibody MR-1 as w êll as depletion of macrophage liposomes resulted in prolonged expression of AdFIX as vŝ ell as higher levels of plasma FIX. This persistence Mras accompanied by inhibition of anti-adenovirus IgG, and decreased IL-10 and IFN-y production from spleen lymphocytes follov^ing reexposure to virus particles in vitro. This treatment regimen also enabled secondary and tertiary infusions of AdFIX w^hich w âs superior to treatment w îth CD40 ligand block ade alone. Kuzmin et al. [27] utilized macrophage depletion in combination w îth blockade of CD40 ligation to demonstrate the prolonged expression of transgene after administration of El-deleted adenovirus. This resulted in a decreased cellular and humoral response as w êll as the induction of transgene tolerance in the animals. Animals that v^ere rendered immunologically unre sponsive to vector and transgene antigens regained their ability to mount a productive immune response against the vector after recovery of immune func tion but remained unresponsive to the transgene product. Stein et al, [28] used an anti-CD40 ligand (anti-CD 154) in combination v^ith adenovirus-mediated low^-density-lipoprotein receptor (LDLR) gene transfer in LDLR-deficient mice to demonstrate that these mice express LDLR on hepatocytes and maintain cholesterol levels below^ or w^ithin the normal range for at least 92 days. It has been more difficult to eliminate B-cell responses. B-cell produc tion of neutralizing antibodies is decreased after treatment w îth anti-CD40 or soluble CD40 [29] and deoxyspergualin [30-32]. Readministration of ade novirus vector has been achieved in the lungs of nonhuman primate by blocking of CD40-CD40 ligand interactions. A humanized anti-CD40 ligand antibody hu5C8 w âs used to treat primates in the presence of administra tion of adenovirus vectors [23]. These animals produced IgM but did not develop secretory IgA or neutralizing antibodies. This is significant since this is the first demonstration that anti-Ad neutralizing antibodies could be inhibited in a primate system by inhibiting the CD40 interaction with CD40 ligand. A third approach includes modification of the adenovirus vector to reduce the immune response [33-36]. A more universal strategy for decreas ing response to adenovirus vectors includes production of the "gutless" adenovirus which greatly reduces the immune response to Ad and its trans gene [37-40]. It was initially demonstrated that constitutive expression of the 4 1 4 Zhang efo/. immune modulatory gpl9 K protein in adenovirus vectors reduced the cyto toxic response. Further refinement of vectors including the removal of E4 also resulted in prolonged transgene expression [34]. A gutless Ad that v^as depleted of all adenovirus genes, showed prolonged expression of P-galactosidase in muscle. This prolonged expression correlated w îth a decrease in the infiltration of CD4+ and CD8"^ lymphocytes. However, in LacZ transgenic mice, which was predicted to result in immunologic tolerance to p-galactosidase expression, there was prolonged expression of the vector DNA, indicating that the immune response to this "gutless" Ad was primarily against p-galactosidase and that the response to the adenovirus vector lacking all genes was minimal [37]. Gene therapy expression using adenovirus vectors with deletions of the El , E2A, E3, and E4 regions could be prolonged when combined with immunosuppressive drugs including cyclophosphamide and FK506. One limitation of immunomodulating therapy has been that it is not specific for the adenovirus or transgene. The present review will focus on our attempts to reduce the immune response mediated by TNFot and other cytokines. A second strategy to prolong gene therapy expression is to ablate the immune response to cells that are the target of gene therapy. Such apoptosis-inducing factors include TNFa but also Fas ligand and TNF recep tor apoptosis-inducing ligand (TRAIL). We have also developed methods to specifically reduce the T-cell response to Ad and also new methods to prevent B cell responses including blocking of the TNF receptor (TNFR) homolog trans membrane activator and calcium modulator and cyclophilin ligand (CAML) interactor (TACI) signal in B cells. These strategies strongly suggest that it will be possible to develop strategies to ablate the immune response to adenovirus including the cytotoxic response that leads to the loss of cells carrying the transgene. IV. Treatment with Soluble TNFR1 to Eliminate Ad Inflammation in Lung and Liver The rationale for use of soluble TNF receptor as a modulator of ade novirus inflammation stems from the observation that TNFa is one of the principal mediators of inflammation after adenovirus gene therapy [41-43]. Neutralization of TNFa with TNFa inhibitors, such as soluble TNFRl (sTNFRl) greatly reduces tissue injury and cell death after endotoxin and other inflammatory agents. This is a rational approach, since adenovirus takes advantage of TNFa as an immune mediator to promote expression of several immunosubversive proteins supporting its escape from immunosurveil- lance [5]. The interaction of TNFa with its receptor is a strong virulence factor for inflammation and elimination of virus infection. The E3-gpl9 K protein not only prevents CTL recognition of Ad-infected fibroblasts by sequestering 14. Novel Methods to Eliminate the Immune Response 4 1 5 MHC class I proteins in the endoplasmic reticulum, but also E3 proteins 10.4 K, 14.5 K, and 14.7 K function to protect infected cells from TNFa cytolysis. Transgenic mice that express the E3 gene encoding these proteins have been shown to exhibit decreased pulmonary infiltration after intranasal inoculation [42]. Peng et al. [43] showed that adenovirus gene transfer of an sTNFR results in effective blockade of tumor necrosis factor activity and also prolongs the gene therapy. Therefore, neutralization of TNFa is a rational approach to decrease chronic inflammation as well as prolonged transgene expression. Certain cytokines such as TNFa can result in rapid clearance of ade novirus or viral therapy. Administration of anti-inflammatory cytokines, such as IL-10, can reduce inflammation and prolong gene therapy [44]. We have evaluated the effect of the treatment with a novel TNF-binding protein (TNF-bp), a polyethylene-glycol (PEG)-linked dimer of sTNFRl, on inflammation of the lung and viral clearance after intranasal administration of AdCMVLacZ (1 x W^ pfu) [45, 46] (Fig. 3, see color insert). Three days after intranasal administration, there was a moderate inflammatory infiltrate in the lungs of control (CT)-treated C57BL/6-+/-h mice, which peaked at day 7 and was nearly resolved by day 30 (Fig. 3A). In contrast, 3 days after admin istration of AdCMVLacZ, there was no evidence of an inflammatory infiltrate in the lungs of TNF-bp-treated C57BL/6—h/-h mice and only minimal evidence of infiltration was observed from day 3 through day 30. We next determined the expression of LacZ adenovirus gene-therapy product. The results indicate that the expression of ^-galactosidase (P-Gal) in control-treated C57BL/6—h/+ mice was high at day 3, but was considerably reduced by day 30 (Fig. 3B). The expression of the p-Gal in TNF-bp-treated C57BL/6-+/+ mice was equivalent at day 3 but, in contrast to the control-treated mice, the expression of p-Gal remained high in the lung through day 30. These results indicate that there is greatly decreased inflammatory disease and prolonged gene expression in AdCMVLacZ-infected mice treated with TNF-bp compared to vehicle-treated mice. The results also indicate that TNFa is a key factor in the pathogene sis of inflammation in AdCMVLacZ-virus-infected mice. Thus, the TNF-bp PEG-linked dimer may be therapeutically useful in reducing the inflammatory response to adenovirus gene therapy. V. Inhibition of Cell Cytolysis Which Combines Treatment vŝ ith Soluble DR5, Soluble Fas, and Soluble TNFR1 A major factor limiting prolonged transgene expression is due to the elimination of the cells infected by the adenovirus gene therapy that expresses the gene therapy product. Elimination of cells by either nonspecific "bystander" mediators or specific induction of cell death by T cells and other inflammatory 416 Zhang ef aL cells is carried out by the process of either necrosis or apoptosis. Necrosis results in lysis of the cell and is mediated by TNFa as well as other cytokines. Apoptosis results in elimination of the cell by triggering programmed cell death followed by phagocytosis of the cell into nearby reticulo-endothelial cells. The primary molecules that contain an intracellular death domain include Fas, TNF receptor (TNFR), and death domain receptor (DR3, -4, and -5) mediated by TRAIL (Fig. 4). The relative contribution of TNFa-mediated necrosis compared to Fas ligand (FasL) or TRAIL-mediated apoptosis was investigated using systemic treatment or adenovirus that expressed soluble forms of receptors capable of neutralizing these factors. We have previously shown that Fas-Fc is capable of neutralizing Fas ligand and preventing Fas ligand-mediated apoptosis [47]. Similarly, a soluble form of the death domain receptor 5 (sDR5-Fc) can neutralize TRAIL and inhibit TRAIL-mediated apoptosis [48]. Adenoviruses were constructed that expressed soluble forms of DR5-Fc (sDR5), AdsDR5, and soluble forms of Fas-Fc (sFas), AdsFas. The pretreatment with AdsFas can prevent liver cell apoptosis after iv administration of anti-Fas Jo-2 [49]. AdsDR5 was shown to protect TRAIL-mediated apoptosis of Jurkat T cells. TNFRI Figure 4 Death domain receptor family. Death domain family members are exemplified by Fas and TNF receptor 1 (TNFRI). These have three and four extracellular cystine-rich repeat domains, respectively. There is an 9- to 31-amino-acid linker between the extracellular domain and the transmembrane domain. Both molecules have a homologous intracellular death domain represented by a rectangle. Other members include cytotoxic apoptosis receptor 1 (CAR-1) and also death domain receptors that bind TNF-related apoptosis-inducing ligand (TRAIL). These receptors include death domain receptor 3 (DR3), DR4, and DR5. Also, there is a decoy receptor (DcRI) that can bind TRAIL but lacks the intracellular death domain and therefore binds to TRAIL but does not introduce apoptosis. 14. Novel Methods to Eliminate the Immune Response 417 S AdsDRS S AdsFas D STNFR1 S AdsDRS+AdsFas • AdsDR5+sTNFR1 • AdsFas+sTNR1 DPBS 7 14 Time (days) Figure 5 Inhibition of liver apoptosis by soluble TNFRl, soluble Fas, and soluble DR5. Mice were treated with adenovirus-expressing soluble Fas (AdsFas) and soluble DR5 (AdsDR5) as well as soluble TNFRl (100 ixg/mouse, iv). Mice were also given an Ad-luciferase. The luciferase activity was measured at 3, 7, 14, 30, and 50 days in mice treated with AdsDR5, AdsFas, sTNFRl protein, AdsDR5 plus AdsFas, AdsDRS plus sTNFRl, or AdsFas plus sTNFRl. As a control, mice were given Ad-luciferase plus PBS. To determine if adenovirus gene expression can be prolonged by protecting liver cells from apoptosis, mice vŝ ere pretreated v^ith either AdsDRS, AdsFas, or sTNFRl, or different combinations of these cytoprotective therapies. Pre- treatment v^ith AdsDR5 alone resulted in a greatly prolonged expression by the liver after subsequent administration of an Ad vector expressing luciferase (AdLuc) (Fig. 5). Consistent v^ith our previous results, the second most effec tive molecule to prolong luciferase expression v^as treatment v^ith
sTNFRl. Pretreatment v^ith AdsFas provided only modest prolongation of the luciferase gene expression. Combined treatment v^ith AdsDR5 and sTNFRl provided the greatest protective effect after administration of AdLuc and the greatest prolongation of gene expression. These results indicate that liver gene therapy is limited primarily by expression of TRAIL by either infected liver cells or the immune response to this adenovirus gene therapy and TNF and FasL play a lesser role. VI . Immune Privilege One physiologic mechanism for induction or maintenance of tolerance to self-antigens in the body is for the antigens to be present in an immune privileged site [50-52]. Although anatomical barriers and soluble mediators 4 1 8 Zhang efo/. have been implicated in immune privilege, it appears that apoptotic cell death of Fas-positive cells by tissue-associated FasL is an important component. Constitutive expression of FasL occurs in immune privilege sites including the retina, ciliary body, iris and cornea of the eye, and the testes, w^hich have been knov^n to be an immune privilege site. T cells that interact w îth antigens in immune privilege site upregulate Fas and enable Fas apoptosis signaling and are killed by FasL present in these sites. This prevents inflammation of these sites since the mouse cornea expresses abundant FasL and immune privilege has been implicated in the success of these corneal transplants. The ability of the eye to kill invading inflammatory cells helps maintain immune privilege and minimize bystander tissue damage, v^hile tolerance regulates dangerous inflammatory reactions to prevent autoimmunity. Adenovirus delivered to the eye fails to elicit an immune response [53-55]. The immune privilege site has been proposed to prevent reactivity in other tissues including the thyroid gland and pancreatic P-islet cells. Belgreu et al. [50] demonstrated that FasL expression by testicular Sertoli cells pro tected P-islet from rejection when transplanted heterotopically into the kidney capsule. Similarly, Griffith et al, [51, 52], w êre able to prevent rejection of allo geneic pancreatic islet cells by cotransplantation w îth FasL positive myoblast which were also transplanted to the kidney capsule. These observations were presumably the result of induction of apoptotic cell death of Fas-positive cells invading the graft from the Fas-positive graft tissue. The implication of these studies is that manipulation of the Fas-FasL system might provide a mechanism to prevent local inflammation. We have utilized the immune privilege concept to prolong gene therapy expression in muscle [56, 57]. For this purpose, we have produced a binary adenovirus system consisting of an AdLoxpFasL plus an AxCANCre [56]. AdLoxpFasL can be grown to high titer in 293 cells and does not produce Fas ligand in the absence of AxCANCre. Therefore, these viruses can be grown to high titers separately. However, when transfected into the same cell line, this leads to high-titer production of FasL. To create a local immune privilege site, we used gene therapy to reproduce the immune privilege site created by muscle cells as described above [56, 57], BALB/c mice were injected intraglossally with either AdLoxpFasL plus AxCANCre plus AdCMVLuciferase or with AdLoxpFasL plus AdCMVTK (thymidine kinase) plus AdCMVLuciferase. As a control, mice were injected with luciferase. The mice were analyzed at days 7, 21, 35, and 50 postinjection and luciferase was determined in the harvested tissue. There was no increase in the prolongation of the expression of the vector-encoded transgene by attempting to produce an immune privilege site in muscle and diminish the immune eradication of these vector-transduced cells. This brings up several issues related to the use of FasL for prolongation of gene therapy. First, it is possible that, in the attempt to produce an immune privilege site, ectopic expression of FasL can actually elicit an inflammatory response. 14. Novel Methods to Eliminate the Immune Response 4 1 9 Second, the coexpression of Fas with an ectopic FasL can induce an autocrine loop which might induce target cell apoptosis. In addition, the magnitude and temporal pattern of FasL expression may be critical to determine its efficacy in creation of an immune privilege site. For these reasons, creation of a local immune privilege site in muscle by FasL does not result in prolonged transgene expression. VII. APC-AdFasL Prolongs Transgene Expression and Specifically Minimizes T-Cell Response Adenovirus gene therapy is limited by induction of an immune response to the virus or gene therapy protein product. APCs lead to antigen processing and presentation of T cells which can be highly immunogenic or tolerogenic, depending on costimulatory molecules and production of other cytokines, such as FasL. T-cell tolerance after an immune response to adenovirus is also maintained by activation-induced cell death (AICD) of T cells mediated by Fas/FasL interactions. We surmise that APCs such as macrophages, which express FasL, might directly induce apoptosis of T cells that express Fas as cell therapy resulting in an adenovirus-specific T-cell tolerance without toxic effects of FasL (Fig. 6). The AdLoxpFasL can be used to infect APCs from Ipr/lpr mice and does not kill these APCs since these APCs lack Fas expression [58, 59]. These AdLoxpFasL plus AxCANCre-infected APCs (APC-AdFasL) resuh in high production of FasL capable of lysing A20 target cells. The macrophages infected by the adenovirus exhibited at least a 50- to 100-fold higher FasL titer compared to PMA-activated T cells. High levels of FasL expression by the macrophages were sustained for at least 7 days of in vitro culture. These results indicate that adenovirus can deliver FasL into the primary-culture macrophages from Fas mutant Ipr/lpr mice, and this leads to a high level of FasL expression by the macrophages without toxicity to the macrophages. To determine if APC-AdFasL could inhibit an immune response to Ad, mice were pretreated with APC-AdFasL every 3 days for five doses. After 7 days, the mice were inoculated intravenously with 10^^ pfu of AdCMVLacZ. P-Gal staining was determined up to 50 days later (Fig. 7) [58]. The levels of LacZ gene expression in the liver of control-treated mice decreased rapidly after pretreatment with APC-Ad control. In contrast, in mice treated with APC- AdFasL, the levels of LacZ expression did not decrease but were sustained for at least 50 days after infection. These results indicate that pretreatment with APC-AdFasL significantly prolonged AdCMVLacZ transgene expression. To determine if the APC-AdFasL therapy resulted in a preferential specific deletion of Ad-reactive T cells, but not T cells reactive with other viruses, wild-type C57BL/6-+/-h mice were treated with APCs, APC-Ad control, or APC-AdFasL 420 Zhang ef aL Ad-FasL FasL Apoptosis Ad-specific T cells Ad-Ag Figure 6 Specific apoptosis of T cells by APC-Ad-FasL therapy. Macrophages are infected with an AdFas ligand (Ad-FasL) gene expression system. This results in high expression of FasL on the surface of macrophages. Macrophages do not undergo autocrine apoptosis since they are derived either from Ipr mice (Fas mutant) or express an anti-apoptosis gene. In addition, macrophages will process and present adenovirus antigens (Ad-Ag) specifically to T cells. Activation by processing Ad-Ag by macrophage upregulates Fas expression and also Fas apoptosis signaling, facilitating apoptosis of Ad-specific T cells in the presence of FasL produced by the macrophage. c 'S n APC-AdFasL O 100000 Q. k. H APC-AdControl 0) • PBS > o •5; 10000 o O) E c I 1000 4 'E 3 0) > 2 100 I bt tt 7 14 21 30 50 Days post-injection (IV) with AdCI\/IVLacZ Figure 7 AdLuc transgene expression after APC-AdFasL followed by AdCMVLacZ. Wild-type C57BL/6 -+ /+ mice were treated with 1 x 1C^ of APCs cotransfected with AdLoxpFasL plus AxCAN- Cre (APC-AdFasL), or APCs cotransfected with AdLoxpFasL plus AdCMVLUC (APC-AdControl), or phosphate-buffered saline (PBS) every 3 days until five closes were given. After 7 days, the mice were inoculated iv with 1 x 10^^ pfu of AdCMVLacZ and p-Gal was determined up to 50 days later. The error bars represent the mean ± standard error of the mean (SEM) for three mice analyzed separately in triplicates. 14. Novel Methods to Eliminate the Immune Response 421 ou - ? 60- ^ 3 40- 1 T • Control CM • Ad ^ 20- n _ 1 J nMCMV APC-AdControl APC-AdFasL Treatment Figure 8 IL-2 production by T cells stimulated with APC plus MCMV after APC-AdFasL. Wild-type mice were treated with either APC-AdFasL or APC-Ad Control. Seven days later, mice were challenged in vitro with either AdCMVLacZ or MCMV. Seven days after splenic T cells were stimulated in vitro with either APCs, AdCMVLacZ transfected APCs, or MCMV-infected APCs. IL-2 production in the supernatants was determined by ELISA. every 3 days until five doses as above. To determine w^hether the T-cell tolerance induced by APC-AdFasL vŝ as specific, the T-cell response of APC-AdFasL- and APC-Ad control-tolerized mice to murine cytomegalovirus (MCMV) infection was evaluated. C57BL/6-+/+ mice v^ere treated as described above and then challenged 7 days later v^ith either adenovirus or MCMV. Although there was a reduction in the T-cell response through adenovirus vector, the T-cell response to MCMV was not impaired as demonstrated by comparable levels of IL-2 produced by T cells from both APC-Ad control- and APC-AdFasL-treated mice (Fig. 8). These results indicate that inhibition of the T-cell response in APC-AdFasL-tolerized mice is specific for the adenovirus vector. VIII. Production of AdsTACI Prolongs Gene Expression and Minimizes B-Cell Response The TNF receptor family includes apoptosis-signaling molecules as described above including TNFRl, TNFR2, Fas, CD40, DR4, and DR5 [60-64] (Fig. 4]). In addition, the TNF receptor family contains factors related to B-cell growth including TNF- and apoptosis ligand-related lymphocyte-expressed ligand 1 (TALL-1)/B lymphocyte stimulator (BlyS) factor that belongs to the TNF family. TALL-1 is a potent B cell costimulatory factor and acts by direct binding and activating its cell surface receptor on B cells, referred to as transmembrane activator and calcium modulator and cyclophilin ligand (CAML) interactor (TACI). The interaction between TALL-1 and TACI 422 Zhang et aL is important for regulation of B-cell growth and humoral immunity. Stimulation through this pathway promotes production of antibody and autoantibodies. We surmise that blocking this pathway with a soluble TACI-Fc (sTACI-Fc) that binds to TALL-1 would inhibit the B-cell response to adenovirus gene therapy. We therefore constructed an adenovirus expressing sTACI-Fc in the El A site of adenovirus. C57BL/6-+/+ mice were injected on day 0 with AdsTACI (5 x 10^ pfu, iv) and AdLacZ (5 x 10^ pfu iv). On day 3, the mice were either untreated or treated with APC-AdFasL to modulate the T-cell response to adenovirus. On days 7,14, 30, and 50, the antibody production to Ad was determined using an Ad neutralization assay. The Ad neutralization assay was carried out using two fold dilution of sera, which was then incubated with an Ad vector expressing green fluorescent protein (AdGFP) for 1 h (Fig. 9). The AdGFP was then inoculated with 293 cells for an additional hour and the unbound AdGFP was washed with PBS. The infected 293 cells were incubated at 37°C for another 3 days, and the percent of GFP-positive cells was assayed as an indicator of the level of adenovirusneutralizing antibody. In control mice, there were high levels of anti-Ad IgG immunoglobulin that reached a maximum titer at day 30 in 800 ^ 700 Days post-treatment o 1 • 0 S 600-\| • 7 O 500^ N 400 S 300-j I m 14 0 30 ^ 5 0 3 O 200-\| 100-1 1 I Control APC/FasL APC/FasL/sTACI sTACI 0 Treatment Figure 9 Reduction of onti-adenovirus antibody response by treatment with AdsTACI. C57BL/6 mice were injected on day 0 with AdsTACI (5 x 10^ pfu, iv) and AdLacZ (5 x 10^ pfu, iv). On day 3, the mice were either untreated or treated with AdAPC-FasL to modulate the T-cell response to adenovirus. The Ad neutralization assay was carried out using a twofold solution of sera which was incubated with AdGFP for 1 h. Sera were collected on days 0, 7, 14, 30, and 50 after administration of AdLacZ or AdsTACI. 14. Novel Methods to Eliminate the immune Response 4 2 3 the absence of any immunosuppressive treatment. In the presence of AdsTACI administered on day 0, the IgG anti-Ad neutraUzation titers remained very lov^ throughout the time course of the study. APC-AdFasL therapy reduced the peak titer especially on days 30 and 50, but the neutralization titer at these time points w âs higher than AdsTACI treatment alone. The combined treatment vŝ ith AdsTACI and APC-AdFas ligand vv̂ as similar
to treatment v^ith AdsTACI alone. These results indicate that treatment v^ith APC-AdFasL or blocking B cells signaling w îth AdsTACI can greatly inhibit the peak anti-Ad immunoglobulin production and also prevent long-term antibody production against adenovirus. IX. Summary Suppression of the T-cell main response to adenovirus, to date, has gen erally been achieved using the same paradigms to reduce the cellular immune response as applied in other systems. This includes reducing the T-cell response either by inhibiting T-cell activation at the surface by blocking costimula- tory molecules such as the CD28/CTLA4 or B7.1/B7.2 ligand pathw^ay or by blocking intracellular signaling using cyclosporine or FK506. More general immunosuppressants such as prednisone and cyclophosphamide also suppress the immune response and are synergistic v̂ îth more specific T-cell surface or intracellular signaling pathv^ays. Similarly, more general immunosuppres sants such as TNF receptor Fc or sTNF receptor can block proinflammatory cytokines after stimulation of an immune response by adenovirus and are synergistic v^ith direct immunosuppressants of the T cells. Methods to elimi nate these cells rather than to suppress them include treatment w îth anti-CD4 or anti-CD3, or more specific cell gene therapy methods using APC-AdFasL v^hich kill T cells that interact v̂ îth APCs that express Ad virus or transgene. All of these three approaches can be used together to exhibit synergistic effects to reduce the number of Ad-reactive T cells, the activation of Ad-specific T cells, and the inflammatory mediators produced by these Ad-specific T cells. A second approach is to reduce the antigenicity of the adenovirus or the transgene. Such methods are analogous to, for example, decreasing the cel lular immune response to organ transplant by tissue typing. In the case of adenovirus, certain components of the virus are know^n to be more antigenic than others and can be eliminated to produce replication-deficient Ad. The ultimate result is the "gutless" adenovirus that only contains the transgene v^ith the Ad inverted terminal repeats (ITRs). The immune response to the transgene is similar to the immune response to any biological administered reagent such as factor IX or soluble TNF receptor-Fc. These responses lead to neutralizing antibodies to the transgene or administered biologic protein and mechanisms to eliminate antibody response can also be successfully reduced or eliminated. 4 2 4 Zhang et at. References 1. Yang, Y., Nunes, F. A., Berencsi, K., Furth, E. E., Gonczol, E., and Wilson, J. M. (1994). Cellular immunity to viral antigens limits El-deleted adenoviruses for gene therapy. Proc. Natl. Acad. Sci. USA 91, 4407-4411. 2. Yang, Y., Ertl, H. C , and Wilson, J. M. (1994). MHC class I-restricted cytotoxic T lym phocytes to viral antigens destroy hepatocytes in mice infected with El-deleted recombinant adenoviruses. Immunity 1, 433-442. 3. Chirmule, N., Hughes, J. V., Gao, G. P., Raper, S. E., and Wilson, J. M. (1998). Role of E4 in eliciting CD4 T-cell and B-cell responses to adenovirus vectors delivered to murine and nonhuman primate lungs./. Virol. 72, 6138-6145. 4. Bennett, E. M., Bennink, J. R., Yewdell, J. W., and Brodsky, F. M. (1999). Cutting edge: Adenovirus El9 has two mechanisms for affecting class I MHC expression./. Immunol. 162, 5049-5052. 5. Wold, W. S., Doronin, K., Toth, K., Kuppuswamy, M., Lichtenstein, D. L., and Tollefson, A. E. (1999). Immune responses to adenoviruses: Viral evasion mechanisms and their impli cations for the clinic. Curr. Opin. Immunol. 11, 380-386. 6. Jooss, K., Yang, Y., and Wilson, J. M. (1996). Cyclophosphamide diminishes inflammation and prolongs transgene expression following delivery of adenoviral vectors to mouse liver and lung. Hum. Gene Ther. 7, \555-1566. 7. Fang, B., Eisensmith, R. C , Wang, H., Kay, M. A., Cross, R. E., Landen, C. N., Gordon, G., Bellinger, D. A., Read, M. S., Hu, P. C , et al. (1995). Gene therapy for hemophilia B: Host immunosuppression prolongs the therapeutic effect of adenovirus-mediated factor IX expression. Hum. Gene Ther. 6, 1039-1044. 8. Lochmuller, H., Petrof, B. J., Pari, G., Larochelle, N., Dodelet, V., Wang, Q., Allen, C , Prescott, S., Massie, B., Nalbantoglu, J., and Karpati, G. (1996). Transient immunosuppres sion by FK506 permits a sustained high-level dystrophin expression after adenovirus-mediated dystrophin minigene transfer to skeletal muscles of adult dystrophic (mdx) mice. Gene Ther. 3, 706-716. 9. Howell, J. M., Lochmuller, H., O'Hara, A., Fletcher, S., Kakulas, B. A., Massie, B., Nalban toglu, J., and Karpati, G. (1998). High-level dystrophin expression after adenovirus-mediated dystrophin minigene transfer to skeletal muscle of dystrophic dogs: Prolongation of expression with immunosuppression. Hum. Gene Ther. 9, 629-634. 10. Ilan, Y., Jona, V. K., Sengupta, K., Davidson, A., Horwitz, M. S., Roy-Chowdhury, N., and Roy-Chowdhury, J. (1997). Transient immunosuppression with FK506 permits long-term expression of therapeutic genes introduced into the liver using recombinant adenoviruses in the rat. Hepatology 26, 949-956. 11. Ilan, Y., Prakash, R., Davidson, A., Jona, Droguett, G., Horwitz, M. S., Chowdhury, N. R., and Chowdhury, J. R. (1997). Oral tolerization to adenoviral antigens permits long-term gene expression using recombinant adenoviral vectors. / . Clin. Invest. 99, 1098-1106. 12. DeMatteo, R. P., Raper, S. E., Ahn, M., Fisher, K. J., Burke, C , Radu, A., Widera, G., Clay- tor, B. R., Barker, C. F., and Markmann, J. F. (1995). Gene transfer to the thymus. A means of abrogating the immune response to recombinant adenovirus. Ann. Surg. Ill, 229-239. 13. Sawchuk, S. J., Boivin, G. P., Duwel, L. E., Gall, W., Bove, K., Tarpnell, B., and Hirsch, R. (1996). Anti-T cell receptor monoclonal antibody prolongs transgene expression follow ing adenovirus-mediated in vivo gene transfer to mouse synovium. Hum. Gene Ther. 7, 499-506. 14. DeMatteo, R. P., Markmann, J. F., Kozarsky, K. F., Barker, C. F., and Raper, S. E. (1996). Prolongation of adenoviral transgene expression in mouse liver by T lymphocyte subset depletion. Gene Ther. 3, 4 -12 . 14. Novel Methods to Eliminate the Immune Response 4 2 5 15. Chirmule, N., Truneh, A., Haecker, S. E., Tazelaar, J., Gao, G. P., Raper, S. E., Hughes, J. V., and Wilson, J. M. (1999). Repeated administration of adenoviral vectors in lungs of human CD4 transgenic mice treated v^ith a nondepleting CD4 antibody./. Immunol. 163, 448-455. 16. Yang, Y., Trinchieri, G., and Wilson, J. M. (1995). Recombinant IL-12 prevents formation of blocking IgA antibodies to recombinant adenovirus and allows repeated gene therapy to mouse lung. Nat. Med. 1, 890-893. 17. Kay, M. A., Holterman, A. X., Meuse, L., Gov^n, A., Ochs, H. D., Linsley, P. S., and Wil son, C. B. (1995). Long-term hepatic adenovirus-mediated gene expression in mice following CTLA4Ig administration. Nat. Genet. 11, 191-197. 18. Kay, M. A., Meuse, L., Gown, A. M., Linsley, P., HoUenbaugh, D., Aruffo, A., Ochs, H. D., and Wilson, C. B. (1997). Transient immunomodulation with anti-CD40 ligand antibody and CTLA4Ig enhances persistence and secondary adenovirus-mediated gene transfer into mouse liver. Froc. Natl. Acad. ScL USA 94, 4686-4691. 19. Ideguchi, M., Kajiwara, K., Yoshikawa, K., Uchida, T., and Ito, H. (2000). Local adenovirus- mediated CTLA4-immunoglobulin expression suppresses the immune responses to adenovirus vectors in the brain. Neuroscience 95, 217-226. 20. Guibinga, G. H., LochmuUer, H., Massie, B., Nalbantoglu, J., Karpati, G., and Petrof, B. J. (1998). Combinatorial blockade of calcineurin and CD28 signaling facilitates primary and secondary therapeutic gene transfer by adenovirus vectors in dystrophic (mdx) mouse muscles. / . V/Vo/. 72,4601-4609. 21. Schowalter, D. B., Meuse, L., Wilson, C. B., Linsley, P. S., and Kay, M. A. (1997). Constitutive expression of murine CTLA4Ig from a recombinant adenovirus vector results in prolonged transgene expression. Gene Ther. 4, 853-860. 22. Ah, R. R., Reichel, M. B., Byrnes, A. P., Stephens, C. J., Thrasher, A. J., Baker, D., Hunt, D. M., and Bhattacharya, S. S. (1998). Co-injection of adenovirus expressing CTLA4-Ig prolongs adenovirally mediated lacZ reporter gene expression in the mouse retina. Gene Ther. 5, 1561-1565. 23. Chirmule, N., Raper, S. E., Burkly. L., Thomas, D., Tazelaar, J., Hughes, J. V., and Wilson, J. M. (2000). Readministration of adenovirus vector in nonhuman primate lungs by blockade of CD40-CD40 ligand interactions./. Virol. 74, 3345-3352. 24. Wilson, C. B., Embree, L. J., Schowalter, D., Albert, R., Aruffo, A., HoUenbaugh, D., Linsley, P., and Kay, M. A. (1998). Transient inhibition of CD28 and CD40 ligand interactions prolongs adenovirus-mediated transgene expression in the lung and facilitates expression after secondary vector administration./. Virol. 72, 7542-7550. 25. Jooss, K., Turka, L. A., and Wilson, J. M. (1998). Blunting of immune responses to adenoviral vectors in mouse liver and lung with CTLA4Ig. Gene Ther. 5, 309-319. 26. Stein, C. S., Pemberton, J. L., van Rooijen, N., and Davidson, B. L. (1998). Effects of macrophage depletion and anti-CD40 ligand on transgene expression and redosing with recombinant adenovirus. Gene Ther. 5, 431-439. 27. Kuzmin, A. L, Galenko, O., and Eisensmith, R. C. (2000). An immunomodulatory procedure that stabilizes transgene expression and permits readministration of el-deleted adenovirus vectors. Mol. Ther. 3, 293-301. 28. Stein, C. S., Martins, L, and Davidson, B. L. (2000), Long-term reversal of hypercholes terolemia in low density lipoprotein receptor (LDLR)-deficient mice by adenovirus-mediated LDLR gene transfer combined with CD154 blockade. / . Gene Med. 2, 4 1 - 5 1 . 29. Yang, Y., Su, Q., Grewal, L S., Schilz, R., Flavell, R. A., and Wilson, J. M. (1996). Transient subversion of CD40 ligand function diminishes immune responses to adenovirus vectors in mouse liver and lung tissues. / . Virol. 70, 6370-6377. 30. Cichon, G., and Strauss, M. (1998). Transient immunosuppression with 15-deoxyspergualin prolongs reporter gene expression and reduces humoral immune response after adenoviral gene transfer. Gene Ther. 5, 85-90. 4 2 6 Zhang ef al. 31. Kaplan, J. M., and Smith, A. E. (1997). Transient immunosuppression with deoxysperguahn improves longevity of transgene expression and ability to readminister adenoviral vector to the mouse lung. Hum. Gene Ther. 8, 1095-1104. 32. Smith, T. A., White, B. D., Gardner, J. M., Kaleko, M., and McClelland, A. (1996). Transient immunosuppression permits successful repetitive intravenous administration of an adenovirus vector. Gene Ther. 3, 496-502. 33. Lee, M. G., Abina, M. A., Haddada, H., and Perricaudet, M. (1995). The constitutive expression of the immunomodulatory gpl9 k protein in E1-, E3- adenoviral vectors strongly reduces the host cytotoxic T cell response against the vector. Gene Ther. 2, 256-262. 34. Wang, Q., Greenburg, G., Bunch, D., Farson, D., and Finer, M. H. (1997). Persistent transgene expression in mouse liver follov^ing in vivo gene transfer with a delta El/delta E4 adenovirus vector. Gene Ther. 4, 393-400. 35. Yang, Y., Su, Q., and Wilson, J. M. (1996). Role of viral antigens in destructive cellu lar immune responses to adenovirus vector-transduced cells in mouse lungs. / . Virol. 70, 7209-7212. 36. Harvey, B. G., Worgall, S., Ely, S., Leopold, P. L., and Crystal, R. G. (1999). Cellular immune responses of healthy individuals to intradermal administration of an El-E3-adenovirus gene transfer vector. Hum. Gene Ther. 10, 2823-2837. 37. Chen, H. H., Mack, L. M., Kelly, R., Ontell, M., Kochanek, S., and Clemens, P. R. (1997). Persistence in muscle of an adenoviral vector that lacks all viral genes. Proc. Natl. Acad. Set. USA 94, 1645-1650. 38. Christ, M., Lusky, M., Stoeckel, F., Dreyer, D., Dieterle, A., Michou, A. I., Pavirani, A., and Mehtali, M. (1997). Gene therapy with recombinant adenovirus vectors: Evaluation of the host immune response. Immunol. Lett. 57, 19-25. 39. Molinier-Frenkel, V., Gahery-Segard, H., Mehtali, M., Le Boulaire, C , Ribault, S., Boulanger, P., Tursz, T., Guillet, J. G., and Farace, F. (2000). Immune response to recombinant adenovirus in humans: Capsid components from viral input are targets for vector-specific cytotoxic T lymphocytes./. Virol. 74, 7678-7682. 40. Harvey, B. G., Worgall, S., Ely, S., Leopold, P. L., and Crystal, R. G. (1999). Cellular immune responses of healthy individuals to intradermal administration of an El-E3-adenovirus gene transfer vector. Hum. Gene Ther. 10, 2823-2837. 41. Kolls, J. K., Lei, D., Nelson, S., Summer, W. R., Greenberg, S., and Beutler, B. (1995). Adenovirus-mediated blockade of tumor necrosis factor in mice protects against endotoxic shock yet impairs pulmonary host defense. / . Infect. Dis. 171, 570-575. 42. Harrod, K. S., Hermiston, T. W., Trapnell, B. C , Wold, W. S., and Whitsett, J. A. (1998). Lung-specific expression of adenovirus E3-14.7 K in transgenic mice attenuates adenoviral vector-mediated lung inflammation and enhances transgene expression. Hum. Gene Ther. 9, 1885-1898. 43. Peng, Y., Trevejo, J., Zhou, J., Marino, M. W., Crystal, R. G., Falck-Pedersen, E., and Elkon, K. B. (1999).
Inhibition of tumor necrosis factor alpha by an adenovirus-encoded soluble fusion protein extends transgene expression in the liver and lung./. Virol. 73, 5098-5109. 44. Minter, R. M., Rectenwald, J. E., Fukuzuka, K., Tannahill, C. L., La Face, D., Tsai, V., Ahmed, I., Hutchins, E., Moyer, R., Copeland, E. M., 3rd, and Moldawer, L. L. (2000). TNF- alpha receptor signaling and IL-10 gene therapy regulate the innate and humoral immune responses to recombinant adenovirus in the lung. / . Immunol. 164, 443-451. 45. Zhang, H. G., Zhou, T., Yang, P., Edwards, C. K., 3rd, Curiel, D. T., and Mountz, J. D. (1998). Inhibition of tumor necrosis factor alpha decreases inflammation and prolongs adenovirus gene expression in lung and liver. Hum. Gene Ther. 9, 1875-1884. 46. Zhang, H. G., Xie, J., Yang, P., Wang, Y., Xu, L., Liu, D., Hsu, H. C , Zhou, T., Edwards, C. K., 3rd, and Mountz, J. D. (2000). Adeno-associated virus production of soluble tumor necrosis factor receptor neutralizes tumor necrosis factor alpha and reduces arthritis. Hum. Gene Ther. 11,2431-2442. 14. Novel Methods to Eliminate the Immune Response 4 2 7 47. Cheng, J., Zhou, T., Liu, C , Shapiro, J. P., Brauer, M. J., Kiefer, M. C , Barr, P. J., and Mountz, J. D. (1994). Protection from Fas-mediated apoptosis by a soluble form of the Fas molecule. Science 263, 1759-1762. 48. Strehlow, D., Jodo, S., and Ju, S. T. (2000). Retroviral membrane display of apoptotic effector molecules. Proc. Natl. Acad. Sci. USA 97, 4209-4214. 49. Matsuki, Y., Li, L., Hsu, H. - C , Yang, P., Zheng, R., Edwards, C. K. Ill, Chaudry, I. H., Mountz, J. M., Zhang, H. -G., and Mountz, J. D. (2002). Soluble Fas gene therapy protects against Fas-mediated apoptosis of hepatocytes but not the lethal effects of Fas-induced TNF-a production by Kupffer Cells. Cell Death Diff., in press. 50. Bellgrau, D., Gold, D., Selawry, H., Moore, J., Franzusoff, A., and Duke, R. C. (1995). A role for CD95 ligand in preventing graft rejection. Nature 377, 630-632. 51. Griffith, T. S., Brunner, T., Fletcher, S. M., Green, D. R., and Ferguson, T. A. (1995). Fas ligand-induced apoptosis as a mechanism of immune privilege. Science 270, 1189-1192. 52. Griffith, T. S., Yu, X., Herndon, J. M., Green, D. R., and Ferguson, T. A. (1996). CD95- induced apoptosis of lymphocytes in an immune privileged site induces immunological tolerance. Immunity 5, 7-16. 53. Bennett, J., Pakola, S., Zeng, Y., and Maguire, A. (1996). Humoral response after administra tion of El-deleted adenoviruses: Immune privilege of the subretinal space. Hum. Gene Ther. 7, 1763-1769. 54. Reichel, M. B., Ah, R. R., Thrasher, A. J., Hunt, D. M., Bhattacharya, S. S., and Baker, D. (1998). Immune responses limit adenovirally mediated gene expression in the adult mouse eye. Gene Ther. 5, 1038-1046. 55. Rubinchik, S., Ding, R., Qiu, A. J., Zhang, F., and Dong, J. (2000). Adenoviral vector v̂ ĥich delivers FasL-GFP fusion protein regulated by the tet-inducible expression system. Gene Ther. 7, 875-885. 56. Bilbao, G., Zhang, H., Contreras, J. L., Zhou, T., Feng, M., Saito, I., Mountz, J. D., and Curiel, D. T. (1999). Construction of a recombinant adenovirus vector encoding Fas ligand with a CRE/Loxp inducible system. Transplant Proc. 31, 792-793. 57. Zhang, H. G., Bilbao, G., Zhou, T., Contreras, J. L., Gomez-Navarro, J., Feng, M., Saito, I., Mountz, J. D., and Curiel, D. T. (1998). Application of a Fas ligand encoding a recombinant adenovirus vector for prolongation of transgene expression./. Virol. 72, 2483-2490. 58. Zhang, H. G., Liu, D., Heike, Y., Yang, P., Wang, Z., Wang, X., Curiel, D. T., Zhou, T., and Mountz, J. D. Induction of specific T-cell tolerance by adenovirus-transfected. Fas ligand-producing antigen presenting cells. Nat. Biotechnol. 16, 1045-1049. 59. Zhang, H. G., Zhou, T., Yang, P., Edwards, C. K., 3rd, Curiel, D. T., and Mountz, J. D. (1998). Inhibition of tumor necrosis factor alpha decreases inflammation and prolongs adenovirus gene expression in lung and liver. Hum. Gene Ther. 9, 1875-1884. 60. Rennert, P., Schneider, P., Cachero, T. G., Thompson, J., Trabach, L., Hertig, S., Holler, N., Qian, F., Mullen, C , Strauch, K., Browning, J. L., Ambrose, C , and Tschopp, J. A. (2000). Soluble form of B cell maturation antigen, a receptor for the tumor necrosis factor family member APRIL, inhibits tumor cell growth. / . Exp. Med. 192, 1677-1684. 61. von Bulow, G. U., and Bram, R. J. (1997). NF-AT activation induced by a CAML-interacting member of the tumor necrosis factor receptor superfamily. Science 278, 138-141. 62. Gross, J. A., Johnston, J., Mudri, S., Enselman, R., Dillon, S. R., Madden, K., Xu, W., Parrish-Novak, J., Foster, D., Lofton-Day, C , Moore, M., Littau, A., Grossman, A., Haugen, H., Foley, K., Blumberg, H., Harrison, K., Kindsvogel, W., and Clegg, C. (2000). TACI and BCMA are receptors for a TNF homologue implicated in B-cell autoimmune disease. Nature 404, 995-943. 63. Xia, X. Z., Treanor, J., Senaldi, G., Khare, S. D., Boone, T., Kelley, M., Theill, L. E., Colombero, A., Solovyev, I., Lee F., McCabe, S., Elliott, R., Miner, K., Hawkins, N., Guo, J., Stolina, M., Yu, G., Wang, J., Delaney, J., Meng, S. Y., Boyle, W. J., and Hsu, H. (2000). 4 2 8 Zhang ef al. TACI is a TRAF-interacting receptor for TALL-1, a tumor necrosis factor family member involved in B cell regulation./. Exp. Med. 192, 137-143. 64. Wu, Y., Bressette, D., Carrell, J. A., Kaufman, T., Feng, P., Taylor, K., Gan, Y., Cho, Y. H., Garcia, A. D., Gollatz, E., Dimke, D., LaFleur, D., Migone, T. S., Nardelli, B., Wei, P., Ruben, S. M., Ullrich, S. J., Olsen, H. S., Kanakaraj, P., Moore, P. A., and Baker, K. P. (2000). Tumor necrosis factor (TNF) receptor superfamily member TACI is a high affinity receptor for TNF family members APRIL and BLyS. / . Biol. Chem. 275, 35,478-35,485. C H A P T E R High-Capacity ''Gutless'' Adenoviral Vectors: Technical Aspects and Applications Gudrun Schiedner/ Paula R. Clemens,^ Christoph Volpers,'' and Stefan Kochanek'' *Center for Molecular Medicine University of Cologne Cologne, Germany "'"Department of Neurology University of Pittsburgh Pittsburgh, Pennsylvania •• Introduction Successful somatic gene therapy fundamentally depends on the availabil ity of vectors that allow the efficient and nontoxic delivery of nucleic acids into the appropriate target cells. El-deleted first-generation adenoviral vectors have been used in a number of clinical trials for the treatment of neoplastic and inherited disorders. So far these vectors have been based on adenovirus serotypes 2 and 5 and have usually been produced in the El-complementing 293 cell line [1]. These vectors may still find an application in the treatment of cancer diseases or for vaccination in which "immunostimulation" by viral functions may act as a beneficial adjuvant to the function of the transgene. However, it is unlikely that in the future these vectors will still be used for the treatment of inherited recessive or dominant disorders that would require durable expression of the therapeutic gene. Immune responses to viral proteins expressed from the vector have been observed in various systems. Immediate toxic effects following gene transfer with high vector doses have been attributed both to the viral capsid and to viral gene expression. Chronic toxicity, apparently unrelated to a specific antiviral immune response has been noted [2]. A DNA capacity of 7-8 kb allows the expression of many cDNAs ADENOVIRAL VECTORS FOR GENE THERAPY 4 2 9 Copyright 2002, Elsevier Science (USA). All rights reserved. 4 3 0 Schiedner ef al. but not of all. However, the main reason why first-generation vectors should not be used for the treatment of inherited disorders in which long-term gene expression is required lies in the realistic appreciation that our knowledge of potential interactions between functions from different viruses from the same or different species or between viral functions and exogenous factors is very slim. As has been discussed before [3], it is likely that there could and would be interactions between viral functions of the gene transfer vector and of other virus species. Because the consequences of such interactions are currently largely unpredictable, gene transfer studies in humans that are based on vectors that still carry viral genes have to be subjected to extremely careful risk-benefit considerations. In an attempt to address several of the disadvantages of first and second- generation adenoviral vectors, a new vector has been developed [4-11] that has been variably named "high-capacity (HC)," "gutless," "gutted," "mini," "deleted," "third-generation," "delta (A)," or "helper-dependent (HD)" ade noviral vector. For simplicity the term HC-Ad vector is used throughout this text. This chapter is divided into two parts. In the first part, technical aspects are discussed as they relate to the production and the design of HC-Ad vectors. In the second part the results of gene transfer experiments are summarized. II. Technical Aspects A. Vector Production The production of first-generation adenovirus vectors is relatively simple: only the El functions that are absent from the vector have to be complemented in a producer cell line. The 293 cell line [1] has been extremely valuable in serv ing the production needs for gene transfer vectors for many years. Vectors with additional mutations in the E2 and/or E4 genes (second-generation vectors) can still be relatively easily complemented by cell lines that provide the missing functions in trans. The production is considerably more complicated with vec tor genomes that have increasingly large deletions. The successful completion of a productive infection cycle and the generation of a large number of infec tious particles during production require the precise coordination of a complex viral transcription and replication program. The current production of HC-Ad vectors is based on earlier studies in which the accidental generation of hybrid vector genomes was observed. These consisted of both adenovirus and human or SV40 DNA, respectively, and were dependent on the presence of a wild-type helper virus [12, 13]. Based on these studies, several research groups success fully rescued recombinant adenoviral vector particles that did not contain any viral coding sequences and expressed different transgenes [4, 5, 8, 14]. The 15. High-Capacity '^Gutless'' Adenoviral Vectors 4 3 1 first vector that expressed a reporter gene and that was deleted in LI, L2, VAI+II, and pTP was produced by using wild-type adenovirus type 2 as helper virus [4]. It was possible to rescue, serially propagate, and partially purify a recombinant, although not fully deleted adenoviral vector. In one system [4, 15] a replication-deficient helper virus was engineered to carry a partially defective packaging signal in order to impair the packaging of the helper virus and thereby to enhance vector production. The packaging signal of Ad5 has been characterized in a series of elegant studies that involved the generation and analysis of a large number of packaging-deficient adenoviral mutants [16-19]. The adenoviral packaging element is located between nucleotides 230 and 370 and consists of seven elements, the so-called consensus A repeats AI-AVII. The detailed molecular mechanism of adenoviral DNA packaging is still not clear [20]. To attenuate the packaging capability of the helper virus, 91 bp of the packaging element involving AII-AV were deleted [5] using the packaging impaired adenoviral mutant dl309-267/358 as a template [17]. Compared to wild-type Ad5, this mutant can be grown in cell culture with about 90-fold reduced titer. Using this mutated packaging signal in an El-deleted helper virus it was possible to serially propagate a plaque isolate containing the vector and helper virus genome on 293 cells, and to separate helper virus and vector particles by CsCl equilibrium centrifugation. The HC-Ad vector was obtained in fairly high titers with a 1% contamination by helper virus. In these early studies an HC-Ad vector was generated that contained the 13.8-kb full-length murine dystrophin cDNA under the control of a 6.5-kb muscle-specific pro moter and a lacZ reporter gene. The only viral elements retained on the vector genome are the inverted terminal repeats (ITRs), which is the viral origin of replication, and the packaging signal. Using a comparable strategy, an HC-Ad vector that expressed the human Factor VIII gene was generated [15]. In these systems the recombinant vector DNA with the wild-type packaging signal is preferentially packaged into capsids. Because of differences in the densities of the particles it is possible to separate vector particles from helper virus by CsCl equilibrium centrifugation so that contamination of the vector with the helper virus is around 1%.
However, our own experiences indicate that by using this production system it would be very difficult if not impossible to produce clinical grade material in large amounts. A different and improved produc tion scheme, which takes advantage of the Cre-loxP recombination system of bacteriophage PI [21] considerably increased the ease of vector production and resulted in an increased vector yield and purity. This system utilizes a helper virus with the packaging signal being flanked by two loxP-recognition sites [9, 10]. As in the earlier production system the helper virus is El deficient and corresponds, therefore, to a replication-defective first-generation aden ovirus vector. The recombinant vector carrying the transgene is produced in 293 cells constitutively expressing Cre recombinase [22]. Following infection of producer cells the packaging signal of the helper virus is excised with high 432 Schiedner ef al. efficiency without affecting viral replication. Having lost the packaging signal by Cre-mediated excision, the helper virus is excluded from the capsids while the recombinant vector is efficiently packaged. Although the Cre-mediated removal of the packaging signal was shown to efficiently suppress helper virus contamination, sometimes overgrowth of the helper virus during amplifica tion has been observed [23]. The helper virus could escape suppression if one of the loxP sites flanking the packaging signal was lost by an intermolecular recombination event that occurred between the two identical packaging signals that are present on the helper virus and on the vector genome. Exchanging the DNA sequences between the consensus A repeats reduced the chances of homologous recombination between vector and helper virus genomes [23]. In a modification of the original Cre-loxP system, additional gene functions (DNA polymerase and preterminal protein) were deleted from the helper virus genome and supplied in trans in a packaging cell line [24, 25]. For clinical-grade production of HC-Ad vectors it is likely that Cre- recombinase-expressing cell lines will be used that are not based on 293 cells. As with first-generation adenovirus vectors there is sequence overlap between the El region in 293 cells and the helper virus genome. Therefore, a generation of replication-competent adenovirus (RCA) by homologous recombination is expected to be likely, especially if large amounts of vector are produced. Cell lines that exclude the generation of RCA have been developed [26, 27]. When Cre recombinase is expressed in these cells, efficient production of HC-Ad vectors with low helper virus contamination is possible [G.S., unpublished observation]. For construction of HC-Ad vector genomes, plasmid cloning strategies have been implemented. Most plasmids for HC-Ad vector construction harbor the left and right adenoviral ITRs, including the packaging signal, and different sizes of stuffer DNAs to accommodate different insert sizes. In these plasmids the ITRs are flanked by unique restriction sites that are used to release the plasmid backbone prior to the rescue of vector in the producer cell line. Following transfection of the vector plasmid into El and Cre expressing producer cells that are coinfected with a loxP helper virus, the vector titer is increased through four to six serial amplifications. In every amplification the cells are coinfected with loxP helper virus. Although excision of the packaging signal of the helper virus is not 100% efficient, the final helper virus contamination in this system can be less than 0 .1%. The vector yield per cell can be as high as 1000-2000 infectious units [G.S., unpubfished observation]. B. Stuffer DNA Earlier experiences with Ad5-SV40 hybrid vectors had suggested that the lower size limit for efficient production of adenoviral genomes is about 25 kb [66]. However, most expression cassettes that are currently in use for gene transfer are of much smaller sizes. Rearrangements and/or amplifications 15. High-Capacity '^Gutless'' Adenoviral Vectors 4 3 3 of the vector genome resulting in concatamerization of a starting monomer has been the rule in several studies in w^hich small expression cassettes v^ere rescued as deleted adenoviral vectors. The vector preparations consisted of viral particles that contained both monomeric and dimeric genomes and frequently w êre mixtures of particles v^ith head-to-head, head-to-tail, or tail-to-tail DNA concatemers [6, 14, 28, 29]. This size-dependence of stable vector production vŝ as confirmed with the loxP production system that w âs used to rescue and propagate HC-Ad vectors w îth differently sized vector genomes as starting material. Only vectors w îth genome sizes of at least 27 kb allov^ed efficient and stable vector amplification [30]. Thus, "stuffer" DNA has to be added to the therapeutic gene cassette to bring the total vector genome size to at least 27 kb. Some practical considerations are outlined here. Since approximately 30-kb plasmids are typically used for vector construction, the stuffer DNA should not contribute to instability during plasmid propagation in Escherichia coli. Larger stretches of repetitive elements might increase the likelihood of vector instability during cloning and production procedures and therefore should be avoided. Likewise, stuffer DNA should support stability and growth during viral vector amplification. Stuffer DNA should not interfere with transgene expression in vivo and should be transcriptionally silent. Other elements like matrix or scaffold attachment regions (MARs or SARs) may have positive influences on vector stability in the transduced cells. Stuffer DNA has the potential to promote recombination between HC-Ad vectors and the recipient cell genome since these vectors may share large stretches of homology with the genomic DNA of the target cell. This could increase the possibility of vector integration by homologous recombination. However, experimental evidence from in vitro studies suggests that, compared to first-generation adenoviral vectors, integration frequencies might be somewhat increased but are unlikely to be high [31]. There is some evidence that the source of the stuffer sequences may have an impact on the levels of transgene expression from the vector. An HC-Ad vector containing CpG-rich stuffer DNA derived from phage lambda resulted in significantly reduced and only short-term hepatic expression of a lacZ transgene when compared to an HC-Ad vector containing noncoding stuffer DNA from the human hypoxanthine-guanine phosphoribosyltransferase (HPRT) locus [32]. One explanation that could account for this observation was a possible inadvertent expression of phage lambda genes in eukaryotic cells resulting in toxicity or immunogenicity. In addition, lambda DNA harbors a high number of immunostimulatory CpG motifs which could contribute to the immunogenicity of the transgene protein. Human DNA sequence is probably the best source of stuffer DNA. Various HC-Ad vectors carrying human stuffer DNA have demonstrated high and long-lasting transgene expression in vivo [2, 29, 33-36]. These vectors either contained noncoding stuffer DNA from the human HPRT locus and/or from the human cosmid C346. Even though these 4 3 4 Schiedner ef al. stuffer DNAs were stable through cloning and amplification in most HC-Ad vectors, some reports have indicated instability of the stuffer DNA that was derived from the HPRT gene [23]. In these analyses, HC-Ad vectors containing stuffer DNAs from other human genomic loci seemed to have some growth advantages during amplification and also showed improved expression levels when compared to vectors containing HPRT stuffer DNAs. Future experiments will likely add to the understanding of the impact of human stuffer DNA on expression levels and stability. C. Vector Capsid Modification Experimental strategies directed toward the improvement of efficacy and safety of adenoviral, including HC-Ad vector-mediated, gene transfer by modification of the vector capsid involve three different aspects: first, attempts to circumvent the neutralizing immune response raised within the recipient following the initial vector delivery and preventing repeated administration; second, efforts to abolish the native adenoviral tropism in order to minimize transduction of nontarget tissues; and third, introduction of new ligands or binding domains to target the vector to specific cell types. The capsid of HC-Ad vectors, whose protein components are encoded by the helper virus genome, is not different from that of first-generation vectors. Therefore, neutralizing antibodies produced as a consequence of the first vector delivery still represent a significant problem in readministration schemes. Based on the observation that neutralizing antibodies are Ad type-specific [37], Parks et al. recently demonstrated that this problem can be overcome by the sequential use of HC-Ad vectors of alternative serotypes [38]. In addition to the Ad5-based helper virus originally used in the Cre/loxP system [9], a new helper virus was constructed that was based on serotype 2. An HC-Ad vector with an Ad2 capsid was injected into mice, followed 3 months later by administration of a HC-Ad vector that had either an Ad2 capsid or an Ad5 capsid. The repeat administration of the HC-Ad vector of the same serotype resulted in a 30- to 100-fold reduction in reporter gene expression in the liver, compared with unimmunized animals, whereas no decrease in transgene expression was observed when the second HC-Ad vector was of the different serotype. No Ad5-cross-reactive antibodies were produced in mice immunized with the Ad2- based vector [38]. Similarly, successful repeat vector delivery was achieved in baboons by sequential administration of Ad5- and Ad2-based first-generation Ad vectors [39]. These data indicate that such an approach, taking advantage of the availability of different adenoviral serotypes, might allow repeated gene transfer in immunocompetent individuals. With respect to strategies that aim at a modification of the tropism of adenoviral vectors, HC-Ad vector technology will build on the experience and results collected with first-generation vectors. Considerable progress has 15. High-Capacity ''Gutless'' Adenoviral Vectors 4 3 5 recently been made to develop infectious vector particles v^ith reduced or no affinity for the native coxsackie-and-adenovirus receptor, CAR, by site-directed mutagenesis of the CAR-binding region in the fiber knob domain [40, 41], by the design of knobless vector particles [42], or by production of completely fiberless particles in specialized producer cells [43]. These modifications, when applied to HC-Ad vectors in the future, could further add to their targeting efficiency and safety by reducing undesired infection of nontarget cells and increasing vector concentration at target sites in vivo. Retargeting of first- generation vectors to cell surface molecules of interest has been achieved by the use of bispecific "adapter" molecules like bispecific recombinant antibod ies [44] or CAR fusion proteins [45], by chemical cross-linking of binding moieties to the vector capsid [46] or by genetic insertion of ligands either into the fiber knob protein [47] or the hexon protein [48]. (For comprehensive description of these strategies, see Chapter 8 in this volume). For retargeting of HC-Ad vectors, we have recently constructed a new Ad5-based helper virus containing two unique restriction sites in the fiber gene which facilitate inser tion of binding ligands into the fiber knob HI loop. In one line of experiments, an RGD peptide motif was inserted into this reengineered HI loop site for redirecting vectors to av integrins. HC-Ad reporter vectors produced using this helper virus transduced ovarian carcinoma cells as well as primary endothelial and smooth muscle cells with a 2- to 20-fold higher efficiency, depending on the cell type, than unmodified vectors [49], providing proof-of-concept experiments for the powerful combination of HC-Ad vector technology and retargeting strategies. III. Applications A. Liver Gene Transfer The liver possesses a variety of characteristics that make this organ very attractive for gene therapy. Because of the fenestrated structure of its endothelium, the liver parenchymal cells are readily accessible to large particles such as viruses present in the blood. With respect to blood circulation, the liver can serve as a secretory organ for the systemic delivery of many therapeutic proteins. In addition, in many inborn errors of metabolism the liver is the mainly affected organ. Adenoviral vectors gained considerable interest for liver gene therapy owing to their capacity to very efficiently transduce quiescent hepatocytes in vivo. In fact, upon intravenous injection into the tail vein of mice, a large proportion of adenovirus particles preferentially localizes to the liver. However, in immunocompetent animals and with first-generation adenoviral vectors, transgene expression in general is transient both due to the loss of transduced hepatocytes and to promoter inactivation. Immunological 4 3 6 Schiedner ef al. and toxic effects in transduced cells due to viral gene expression significantly limit the use of first-generation vectors for hepatic gene transfer in vivo. An HC-Ad vector expressing the human a 1-antitrypsin gene w âs used in several instructive experiments. Using the loxP helper virus production system, an HC-Ad vector w âs generated containing the 19-kb genomic human a 1-antitrypsin locus that included both the macrophage and liver- specific promoters, all exons and introns, and the natural polyadenylation signal [2]. al-Antitrypsin antagonizes
neutrophilic elastase and is abundantly expressed in hepatocytes and at a low^er level in macrophages. Expression in the tw ô cell types is regulated by different tissue-specific promoters. Currently, al-antitrypsin-deficient patients have a shortened life expectancy due to emphy sema. Patients are treated with v^eekly injections of human a 1-antitrypsin purified from human plasma. Gene transfer of 2 x 10^^ particles of this vector in immunocompetent C57BL/6J mice resulted in tissue-specific and stable gene expression for longer than 1 year. Transcription of the human a 1-antitrypsin RNA in the liver of transduced animals v^as initiated from the liver-specific promoter, but not from the macrophage-specific promoter. Gene transfer w îth increasing vector doses resulted in high and stable al-antitrypsin levels in serum. Significantly, w îth increasing vector doses, serum levels of a 1-antitrypsin w êre obtained that w^ould be considered supraphysiological in humans. Even these very high vector doses w êre not accompanied by liver toxicity. Mice that received the same dose of a first-generation vector carrying the human a 1-antitrypsin cDNA under the control of the murine phosphogylcerate kinase (PGK) promoter experienced liver damage as documented by histological abnormalities and elevated liver enzymes detected in the serum of transduced mice [50]. Gene transfer of this vector in baboons resulted in relatively stable transgene expression for longer than 16 months in tw ô of three baboons [39]. In these animals only a slow decline was observed to 19% and 8% of peak levels at 16 and 24 months, respectively. This was not surprising for two reasons. First, hepatocytes are not postmitotic and there is a regular, albeit slow, turnover in this cell type. Second, the animals were young and still growing when they were injected. Therefore, a decline of a 1-antitrypsin levels correlated with animal growth. In a third baboon, the generation of anti-al-antitrypsin antibodies was associated with a short duration of expression of only 2 months. Transgene expression in all three animals injected with a first-generation vector was limited to 3 to 6 months. The lack of anti-al-antitrypsin antibodies in these animals and further immunological studies suggested that cellular immune responses against viral proteins might have resulted in the elimination of vector-transduced hepatocytes. In summary, these studies demonstrated the main advantages of HC-Ad vectors: increased capacity allowing the incorporation of large DNA fragments and even some genes in the genomic context, improved levels and persistence of transgene expression, and significantly reduced toxicity. 15. High-Capacity "Gutless'' Adenoviral Vectors 4 3 7 Improved expression and decreased liver toxicity has also been observed follov^ing gene transfer v^ith an HC-Ad vector expressing the murine leptin cDNA from the human cytomegalovirus (HCMV) promoter [29]. Leptin is a potent modulator of weight and food intake. In leptin-deficient ob/ob mice, daily delivery of recombinant leptin protein suppresses appetite, induces weight reduction, and decreases blood insulin and glucose levels. Results from gene transfer experiments with a first-generation vector suggested that delivery of the leptin cDNA might provide therapeutic benefit equivalent to daily leptin protein treatment. However, the effects were only transient in both lean and ob/ob mice due to the loss of DNA and due to significant inflammatory changes in liver. Using an HC-Ad vector carrying the same expression cassette, leptin expression and physiological consequences were analyzed following gene transfer. In lean mice, tail vein injection of 1-2 x 10^^ particles of the HC-Ad vector resulted in long-term leptin expression. Gene expression in ob/ob mice (which are leptin-deficient and therefore not tolerant to leptin) following gene transfer with the same dose of an HC-Ad vector was improved, prolonged, and associated with increased weight loss. However, even in HC-Ad vector transduced ob/ob mice leptin serum levels declined and finally disappeared due to the generation of anti-leptin antibodies. Relatively realistic disease targets for HC-Ad vectors are the clotting disorders hemophilias A and B. The hemophilias are characterized by sponta neous and prolonged bleeding into joints, muscle, and internal organs. Current treatment of the hemophilias, which are often life-threatening and frequently associated with disabling arthropathy due to recurring joint bleeding, consists of protein-replacement therapy with infusion of plasma-derived or recom binant factor VIII (FVIII) or factor IX (FIX). The hemophilias are attractive candidates for gene therapy since they are due to single gene defects. A signif icant advantage is the fact that the therapeutic window is relatively broad. In addition, tissue-specific expression and precise control of the transgene expres sion is probably not required. Importantly, even moderate increases of FVIII or FIX levels would be sufficient to convert a severe hemophilia to a milder form. Intravenous injection of first-generation adenoviral vectors expressing the human or canine B-domain-deleted FVIIII cDNA in normal or hemophilic mice and dogs resulted in therapeutic and physiological levels of biologically active FVIII that was accompanied by a correction of bleeding tendency. However, both in hemophilic mice and dogs FVIII levels gradually declined, resulting in only short-term phenotypic correction. In mice transduced with a first-generation adenoviral vector expressing the human FVIII gene, anti-FVIII antibodies were not detectable. However, in hemophilic dogs, neutralizing FVIII antibodies were generated upon gene transfer of first generation vectors expressing either the human or canine FVIII cDNA [for review see 51]. Recently, an HC-Ad vector that carries the full-length human FVIII cDNA under the control of the 12.5-kb albumin promoter was injected into 4 3 8 Schiedner ef aL hemophilic mice, resulting in efficient hepatic gene transfer and therapeutic FVIII expression which led to the correction of the phenotype. However, FVIII levels declined, possibly due to the generation of inhibitory antibodies to the human FVIII protein. Histopathological findings of vector-induced toxicity were not observed [52]. Therapeutic expression levels could only be observed with relatively high vector doses (2 x 10^^ viral particles per mouse). With a 10-fold lower vector dose FVIII could not be detected in the serum. These results suggested a nonlinear "threshold" effect which also has been observed with first-generation vectors [53]. Two further examples of liver gene transfer by HC-Ad vectors are men tioned since they point to additional advantages of this new vector type. In one instance an HC-Ad vector was generated to express murine erythro- poetin (mEPO), a glycoprotein regulating erythropoiesis [35]. EPO is mainly secreted by kidney peritubular cells in response to hypoxia and promotes late erythroid precursor proliferation and terminal differentiation of erythrocytes. Patients suffering from chronic renal failure show anemia as a major compli cation resulting from the destruction of EPO-secreting cells. These patients are treated with administration of recombinant human EPO protein. As an alterna tive treatment, delivery of the human EPO gene via an HC-Ad vector was tested and compared to a first-generation adenovirus vector with the same expression cassette. Relatively low amounts of an HC-Ad vector (3 x 10^ infectious units or 3 X 10^ particles per mouse) were sufficient to elevate hematocrit levels significantly, although with varying efficiencies, in different immunocompetent mouse strains. In this system the HC-Ad vector was at least 100-fold more efficient than a first generation vector. Because the low vector doses did not initiate any detectable neutralizing antibody response, intravenous readminis- tration of the vector was possible without a need for immunosuppression. In contrast, a second injection of a first-generation virus into mice that had been previously transduced with the same vector induced a much smaller and only transient hematocrit increase. A second example concerns the use of the mifepristone inducible gene expression system within the HC-Ad vector context [34]. In this system a chimeric ^mws-activator was used consisting of a mutated progesterone receptor ligand-binding domain, part of the activation domain of the human p65 subunit of the NF-KB complex, and a GAL4 DNA-binding domain. Expression of the ^r^ws-activator was under the transcriptional control of the liver-specific transthyretin (TTR) promoter. A second expression cassette was located on the same vector and consisted of a 17-mer GAL4-binding site just upstream of a minimal TATA box containing the promoter and cDNA of human growth hormone (hGH). In the presence of the progesterone antagonist mifepristone the transactivator dimerizes, binds to the Gal4 DNA binding site and induces hGH expression. In vitro studies in HepG2 cells and in vivo experiments in mice demonstrated extremely tight control of gene expression and very strong 15. High-Capacity ''Gutless'' Adenoviral Vectors 4 3 9 induction of hGH expression upon administration of mifepristone. Following liver gene transfer, repetitive induction was possible for longer than 1 year [34, and unpublished data]. B. Gene Transfer into Skeletal Muscle The first in vivo application of HC-Ad vectors was for gene transfer studies toward a treatment for Duchenne muscular dystrophy (DMD), an inherited muscular dystrophy caused by mutations in the dystrophin gene. The dystrophin cDNA is 14-kb in length; thus, only shortened versions of this cDNA could be accommodated by first-generation or second-generation adenoviral vectors. Therefore, HC-Ad vectors provided the potential to deliver the full-length dystrophin cDNA with an adenoviral vector. DMD is the most common form of muscular dystrophy with an incidence of 1:3500 male births. Mutations in the dystrophin gene result in the absence of the cytoskeletal dystrophin protein that is normally located at the cytoplasmic face of the cell membrane in skeletal and cardiac muscle. In normal muscle, dystrophin serves as a link in a network of proteins that span from actin within the muscle cell to laminin in the extracellular matrix. The absence of dystrophin results in a secondary loss of dystrophin-associated proteins, increased fragility of the muscle membrane, and cycles of degeneration followed by regeneration. Ultimately, the regenerative process fails and muscle fibers are replaced with fibrosis. HC-Ad vectors encoding the dystrophin cDNA were developed by sev eral groups [5, 8, 14]. Two groups incorporated a muscle-specific muscle creatine kinase (MCK) promoter [5, 8], allowing demonstration of striated muscle-specific expression of dystrophin from the vector. Direct intramuscular injection of these dystrophin-encoding HC-Ad vectors in the dystrophin- deficient mdx mouse model resulted in expression of recombinant dystrophin that properly localized to the muscle sarcolemma [7, 14]. Furthermore, dystrophin-associated proteins, which are lost in DMD and mdx muscle secondary to the primary absence of dystrophin, were restored in muscle fibers expressing HC-Ad vector-delivered dystrophin [54]. The prevention of dystrophic morphologic changes in muscle of mdx mice receiving an intramus cular injection of dystrophin-encoding HC-Ad vector was a second indicator of normal function provided by the recombinant dystrophin that was expressed from the HC-Ad vector [7]. One HC-Ad vector encoding a MCK-driven murine dystrophin cDNA and an HCMV-controUed lacZ gene, called AdDYSPgal, resulted in a profound cellular infiltrate composed primarily of CD4+ and CD8+ T cells when injected intramuscularly in nondystrophic, normal mice, even when gene delivery was performed during the neonatal period [54]. Expression of P-galactosidase was identified as the principal cause of the observed cellular immune response 4 4 0 Schiedner ef al. by performing parallel intramuscular injections of AdDYSPgal in neonatal mice with a germline lacZ transgene on the same genetic background. LacZ- transgenic mice did not develop a cellular infiltrate in skeletal muscle at any time point after intramuscular AdDYSPgal delivery [54]. Further studies demonstrated that dystrophin expression from AdDYSPgal in skeletal muscle of mdx mice also could induce at least an antibody-mediated immune response to dystrophin antigens (P.R.C., unpublished observations). When immunity to the vector v^as largely eliminated in direct muscle gene transfer studies, the AdDYSPgal vector DNA w âs stably maintained in skeletal muscle for at least 5 months [33]. Furthermore, the integrity of vector DNA remained intact [33]. This provided assurance that HC-Ad vector DNA could remain as a stable episome in transduced muscle cells. These studies clearly show^ the utility of HC-Ad vectors for muscle gene transfer. An important issue to address in future studies is the nature of immunity induced by transgene proteins and adenoviral capsid antigens in the context of specific disease applications. It is likely that the underlying pathology of a muscle disorder v îll influence immunity induced or augmented by HC-Ad vector-mediated gene delivery. The low efficiency and extent of gene delivery to muscle is a second issue that currently prevents clinical applications of HC-Ad vectors. Targeting of HC-Ad vectors for muscle gene delivery may permit systemic administration that could result in transduction of muscle tissue widespread throughout the body. C. Gene Transfer into the Eye and into the CNS Adenoviral vectors have successfully been used for transgene delivery to different anatomic compartments and cell
types of the eye, in vitro and in vivo. Several groups have demonstrated efficient transduction of retinal cells with first-generation adenoviral vectors expressing reporter or therapeutic genes [see for example 55-61]. The eye is considered a site of immune privilege, which is immunologically tolerant to foreign antigens similar to the testis, ovary, and uterus [62]. However, following adenoviral-mediated gene transfer into different ocular cell types, gene expression has always been transient. The short duration of gene expression obtained, together with the limited insertion capacity of first-generation Ad vectors, recently prompted studies that aimed at developing HC-Ad vectors for somatic gene therapy of human retinal degenerative diseases. R. Kumar-Singh et al. constructed an "encapsi- dated adenovirus mini-chromosome" containing a full-length murine cDNA encoding the P-subunit of the guanosine 3^,5^-monophosphate (cyclic GMP) phosphodiesterase (P-PDE) under control of a human p-PDE promoter which is transcriptionally active in photoreceptor cells of the neuronal retina [28, 63]. This vector was prepared by cotransfection of 293 cells with helper virus DNA and a circular plasmid with head-to-head-oriented adenoviral ITRs generat ing linear adenoviral "mini-chromosomes" following rescue in 293 cells. The 15. High-Capacity ''Gutless'' Adenoviral Vectors 4 4 1 vector particles contained either monomers of the 13-kb starting material, or dimers in a head-to-head, head-to-tail, or tail-to-tail configuration [28, 63]. The P-PDE HC-Ad vector was delivered to the subretinal space of homozy gous rd mice. These mice, w^hich show^ a similar retinal phenotype as retinitis pigmentosa patients, suffer from an early-age onset of degeneration of reti nal photoreceptors due to a loss-of-function mutation in the P-PDE gene. Expression of p-PDE in transgenic rd mice is know^n to rescue photoreceptor degeneration in this model [64]. In the P-PDE HC-Ad vector-treated animals, expression of the transgene in the neuronal retina v^as demonstrated by RT- PCR, Western blot analysis and functional enzymatic assays [28, 63]. When the thickness of the outer nuclear layer, as a marker of photoreceptor cell rescue, v^as evaluated at l-week intervals, significant differences v̂ êre observed betw^een mice injected w îth the p-PDE HC-Ad vector and control vector up to 12 v̂ êeks postinfection [28, 63]. Despite these encouraging results the expres sion of the P-PDE Ad vector was transient and loss of expression w âs complete at 120 days follov^ing subretinal injection. Whether the loss of expression was due to an immune response directed against contaminating first-generation helper virus or against the transgenic protein, to promoter shutdow^n, or sim ply to instability of the vector DNA is not clear at the time of this w^riting. Since quiescent cells of the CNS allows efficient gene transfer by adenoviral vectors, glial and neuronal cells are very interesting target cells for HC-Ad vectors. In an in vitro study primary neuronal cells isolated from the cerebellum of 8- to 9-day-old mice v^ere transduced w îth either a first-generation or an HC-Ad vector expressing £. coli P-galactosidase [65]. Compared to gene transfer with a first-generation vector, transduction of these primary cells vŝ ith the HC-Ad vector resulted in a marked decline in vector-mediated toxicity as assessed by morphological and metabolic studies. In particular, this was evident at moder ate vector doses, corresponding to up to 50 multiplicities of infection (m.o.i.), a vector dose that resulted in an 85% transduction rate. However, at very high doses, the HC-Ad vector exhibited cytotoxicity, though not as severe as could be observed with a first-generation vector control. A problem of clinical significance that has been rarely addressed concerns the fate of a viral vector following the superinfection by a virus of the same or a closely related serotype. Stereotactic injection in rats into the striatum of the brain of both a first-generation and an HC-Ad vector expressing lacZ resulted in stable gene expression over at least 60 days with both vectors [36]. However, challenge by peripheral subcutanous injection of a first-generation vector expressing an immunologically unrelated transgene resulted in a strong inflammatory response in the brain of rats that had received the first-generation vector but not the HC-Ad vector. Gene expression was completely abolished in rats that were injected with the first-generation vector while expression from the HC-Ad vector was stable. This experimental setup is mirrored by a 4 4 2 Schiedner ef al. clinical situation in which therapeutic gene transfer is followed at a later time by infection with a virus of the same or a closely related serotype. IV. Conclusion Studies to date convincingly demonstrate the utility of HC-Ad vectors for gene transfer into different tissues. Safety and expression features of HC-Ad vectors are improved over earlier-generation adenoviral vectors. The increased capacity may allow coexpression of different therapeutic genes and improved control of gene expression. A critical issue that stands between the current status of HC-Ad vector development and clinically useful applications for human patients is at the level of vector production. It is felt, however, that with recently improved producer cell lines, vectors, and helper viruses the production under GMP will be possible in the near future. Retargeting strategies will increase the spectrum of therapeutic applications and will further add to the safety of HC-Ad vectors. References 1. Graham, F. L., Smiley, J., Russel, W. C , and Nairn, R. (1977). Characteristics of a human cell line transformed by DNA from human adenovirus type 5. J. Gen. Virol, 36, 59-74. 2. Schiedner, G., Morral, N., Parks, R. J., Wu, Y., Koopmans, S. C., Langston, C., Graham, F. L., Beaudet, A. L., and Kochanek, S. (1998). Genomic DNA transfer with a high-capacity adenovirus vector results in improved in vivo gene expression and decreased toxicity. Nat. Genet. 18, 180-183. 3. Kochanek S. (1999). High-capacity adenoviral vectors for gene transfer and somatic gene therapy. Hum. Gene Ther. 10, 2451-2459. 4. Mitani, K., Graham, F. L., Caskey, C. T., and Kochanek, S. (1995). Rescue, propagation and partial purification of a helper virus-dependent adenovirus vector. Proc. Natl. Acad. Set. USA 92, 3854-3858. 5. Kochanek, S., Clemens, P. R., Mitani, K., Chen, H. -H., Chan, S., and Caskey, C. T. (1996). A new^ adenoviral vector: replacement of all viral coding sequences v^ith 28 kb of DNA independently expressing both full-length dystrophin and ?-galactosidase. Proc. Natl. Acad. Sci. U5A 93, 5731-5736. 6. Fisher, K. J., Choi, H., Burda, J., Chen, S.-J., and Wilson, J. M. (1996). Recombinant adenovirus deleted for all viral genes for gene therapy of cystic fibrosis. Virology 217, 11-22. 7. Clemens, P. R., Kochanek, S., Sunada, Y., Chan, S., Chen, H. -H., Campbell, K. P., and Caskey, C. T. (1996). In vivo muscle gene transfer of full-length dystrophin w îth an adenoviral vector that lacks all viral genes. Gene Ther. 3, 965-972. 8. Kumar-Singh, R., and Chamberlain, J. S. (1996). Encapsidated adenovirus minichromosomes allow delivery and expression of a 14 kb dystrophin cDNA to muscle cells. Hum. Mol. Genet. 5 ,913-921 . 9. Parks, R. J., Chen, L., Anton, M., Sankar, U., Rudnicki, M. A., and Graham F. L. (1996). A helper-dependent adenovirus vector system: Removal of helper virus by Cre-mediated excision of the viral packaging signal. Proc. Natl. Acad. Sci. USA 93, 13,565-13,570. 15. High-Capacity ^'Gutless'' Adenoviral Vectors 4 4 3 10. Hardy, S., Kitamura, M., Harris-Stansil, T., Dai, Y., and Phipps, M. L. (1997). Construction of adenovirus vectors through Cre-lox recombination./. Virol. 71, 1842-1849. 11. Alemany, R., Dai, Y., Lou, Y. C , Sethi, E., Prokopenko, E., Josephs, S. F., and Zhang, W. W. (1997). Complementation of helper-dependent adenoviral vectors: Size effects and titer fluctuations./. Virol. Methods 68, 147-159. 12. Deuring, R., Klotz, G., and Doerfler, W. (1981). An unusual symmetric recombinant betw^een adenovirus type 12 DNA and human cell DNA. Proc. Natl. Acad. Set. USA 78, 3142-3146. 13. Gluzman, Y., and van Doren, K. (1983). Palindromic adenovirus type 5-simian virus 40 hybrid./, y/ro/. 45, 91-103. 14. Haecker, S. E., Stedman, H. H., Balice-Gordon, R. -J., Smith, D. B. J, Greelish, J. P., Mitchell, M. A., Wells, A., Sweeney, H. L., and Wilson, J. M. (1996). In vivo expression of full-length human dystrophin from adenoviral vectors deleted of all viral genes. Hum. Gene Ther. 7, 1907-1914. 15. Zhang, W. -W., Josephs, S. F., Zhou, J., Fang, X., Alemany, R., Balague, C , Dai, Y., Ayares, D., Prokopenko, E., Lou, Y. - C , Sethi, E., Hubert-Leslie, D., Kennedy, M., Ruiz, L., and Rockov^-Magnone, S. (1999). Development and application of a minimal adenoviral system for gene therapy of hemophilia A. Thromb. Haemostasis 82, 562-571. 16. Hearing, P., Samulski, R. J., Wishart, W. L., and Shenk, T. (1987). Identification of a repeated sequence element required for efficient encapsidation of the adenovirus type 5 chromosome. / . Virol. 61, 2555-2558. 17. Graeble, M., and Hearing, P. (1990). Adenovirus type 5 packaging domain is composed of a repeated element that is functionally redundant. / . Virol. 64, 2047-2056. 18. Graeble, M., and Hearing, P. (1992). Cis and trans requirements for the selective packaging of adenovirus type 5 DNA. / . Virol. 66, 723-731. 19. Schmid, S. I., and Hearing, P. (1997). Bipartite structure and functional independence of adenovirus type 5 packaging elements./. Virol. 71, 3375-3384. 20. Schmid, S. I., and Hearing, P. (1998). Cellular components interact with adenovirus type 5 minimal DNA packaging domains. / . Virol. 72, 6339-6347. 21. Sauer, B. (1987). Functional expression of the Cre-lox site specific recombination system in the yeast Saccharomyces cerevisiae. Mol. Cell. Biol. 7, 2087-2096. 22. Chen, L., Anton, M., and Graham, F. L. (1996). Production and characterization of human 293 cell lines expressing the site-specific recombinase Cre. Somat. Cell. MoL Genet. 22, 477-488. 23. Sandig, V., Youil, R., Bett, A. J., Franlin, L. L., Oshima, M., Maione, D., Wang, F., Metzker, M. L., Savino, R., and Caskey, C. T. (2000). Optimization of the helper-dependent adenovirus system for production and potency in vivo. Proc. Natl. Acad. Sci. USA 97, 1002-1007. 24. Amalfitano, A., and Chamberlain, J. S. (1997). Isolation and characterization of packaging cell lines that coexpress the adenovirus El , DNA polymerase, and preterminal proteins: Implications for gene therapy. Gene Ther. 4, 258-263. 25. Hartigan-O'Connor, D., Amalfitano, A., and Chamberlain, J. S. (1999). Improved production of gutted adenovirus in cells expressing adenovirus preterminal protein and DNA polymerase. / . Virol. 73, 7835-7841. 26. Fallaux, F. J., Bout, A., van der Velde, I., van den Wollenberg, D. J. M., Hehir, K. M, Keegan, J., Auger, C , Cramer, S. J., van Ormondt, H., van der Eb., A., Valerio, D., and Hoeben, R. C. (1998). New Helper cells and matched early region 1-deleted adenovirus vectors prevent generation of replication-competent adenoviruses. Hum. Gene Ther. 9, 1909-1917. 27. Schiedner, G., Hertel, S., and Kochanek, S. (2000). Efficient transformation of primary human amniocytes by El functions of Ad5: Generation of new cell lines for adenoviral vector production. Hum. Gene Ther. 11, 2105-2116. 28. Kumar-Singh, R., and Farber, D. B. (1998). Encapsidated adenovirus minichromosome- mediated delivery of genes to the retina: application to the rescue of photoreceptor degeneration. Hum. Mol. Genet. 7, 1893-1900. 4 4 4 Schiedner ef al. 29. Morsy, M. A., Gu, M., Motzel, S., Zhao, J., Lin, J., Su, Q., Allen, H., Franklin, L., Parks, R. J., Graham, F. L., Kochanek, S., Bett, A. J., and Caskey, C. T, (1998). An adenoviral vector deleted for all viral coding sequences results in enhanced safety and extended expression of a leptin transgene. Proc. Natl. Acad. Sci. USA 95y 7866-7871. 30. Parks, R. J., and Graham, F. L. (1997). A helper-dependent system for adenovirus production helps define a low^er limit for efficient DNA packaging. / . Virol. 71, 3293-3298. 31. Harui, A., Suzuki, S., Kochanek, S., and Mitani, K. (1999). Frequency and stability of chromosomal integration of adenovirus vectors. / . Virol. 73, 6141-6146. 32. Parks, R. J., Bramson, J. L., Wan, Y., Addison, C. L., and Graham, F. L. (1999). Effects of stuffer DNA on transgene expression from helper-dependent adenovirus vectors. / . Virol. 73, 8027-8034. 33. Chen, H. -H., Mack, L. M., Choi, S. Y., Ontell, M, Kochanek, S., and Clemens, P. R. (1997). DNA form both high-capacity and first-generation adenoviral vectors remains intact in skeletal muscle. Hum. Gene Ther. 10, 365-373. 34. Burcin, M. M., Schiedner, G., Kochanek, S., Tsai, S. Y., and O'Malley, B. W. (1999). Adenovirus-mediated regulable target gene expression in vivo. Proc. Natl. Acad. Sci. USA 96, 355-360. 35. Maione, D., Wiznerowicz, M., Delmastro, P., Cortese, R., Cifiberto, G., La Monica, N., and Savino, R. (2000). Prolonged
expression and effective readministration of erythropoietin delivered with a fully deleted adenoviral vector. Hum. Gene Ther. 11, 859-868. 36. Thomas, C. E., Schiedner, G., Kochanek, S., Castro, M. G., and Loewenstein, P. R. (2000). Peripheral infection with adenovirus causes unexpected long-term brain inflammation in animals injected intracranially with first-generation, but not with high-capacity, adenovirus vectors: towards realistic long-term neurological gene therapy for chronic diseases. Proc. Natl. Acad. Sci. USA 97, 7482-7487. 37. Gall, J. G. D., Crystal, R. G., and Falck-Pedersen, E. (1998). Construction and character ization of hexon-chimeric adenoviruses: Specification of adenovirus serotype. / . Virol. 72, 10,260-10,264. 38. Parks, R. J., Evelegh, C. M., and Graham, F. L. (1999). Use of helper-dependent adenovi ral vectors of alternative serotypes permits repeat vector administration. Gene Ther. 6, 1565-1573. 39. Morral, N., O'Neal, W., Rice, K., Leland, M., Kaplan, J., Piedra, P. A., Zhou, H., Parks, R. J., Velji, R., Aguilar-Cordova, E., Wadsworth, S., Graham, F, L., Kochanek, S., Carey, K. D., and Beaudet, A. L. (1999). Administration of helper-dependent adenoviral vectors and sequential delivery of different vector serotype for long-term liver-directed gene transfer in baboons. Proc. Natl. Acad. Sci. USA 96, 12816-12821. 40. Kirby, L, Davison, E., Beavil, A. J., Soh, C. P. C , Wickham, T. J., Roelvink, P. W., Kovesdi, L, Sutton, B. J., and Santis, G. (1999). Mutations in the DG loop of adenovirus type 5 fiber knob protein abolish high-affinity binding to its cellular receptor CAR. / . Virol. 73, 9508-9514. 41. Kirby, I., Davison, E., Beavil, A. J., Soh, C. P. C , Wickham, T. J., Roelvink, P. W., Kovesdi, L, Sutton, B. J., and Santis, G. (2000). Identification of contact residues and definition of the CAR-binding site of adenovirus type 5 fiber protein. / . Virol. 74, 2804-2813. 42. van Beusechem, V. W., van Rijswijk, A. L. C. T., van Es, H. H. G., Haisma, H. J., Pinedo, H. M., and Gerritsen, W. R. (2000). Recombinant adenovirus vectors with knobless fibers for targeted gene transfer. Gene Ther. 7, 1940-1946. 43. von Seggern, D. J., Chiu, C. Y., Fleck, S. K., Stewart, P. L., and Nemerow, G. R. (1999). A helper-independent adenovirus vector with El , E3, and fiber deleted: Structure and infectivity of fiberless particles./. Virol. 73, 1601-1608. 44. Wickham, T. J., Segal, D. M., Roelvink, P. W., Carrion, M. E., Lizonova, A., Lee, G. M., and Kovesdi, I. (1996). Targeted adenovirus gene transfer to endothelial and smooth muscle cells by using bispecific antibodies./. Virol. 70, 6831-6838. 15. High-Capacity ^^Gutless'' Adenoviral Vectors 4 4 5 45. Dmitriev, I., Kashentseva, E., Rogers, B. E., Krasnykh, V., and Curiel, D. T. (2000). Ectodomain of coxsackievirus and adenovirus receptor genetically fused to epidermal grow^th factor mediates adenovirus targeting to epidermal grow^th factor receptor-positive cells. / . Virol. 74, 6875-6884. 46. Romanczuk, H., Galer, C, E., Zabner, J., Barsomian, G., Wadsv^^orth, S. C , and O'Riordan, C. R. (1999). Modification of an adenoviral vector v^ith biologically selected peptides: A novel strategy for gene delivery to cells of choice. Hwm. Gene Ther. 10, 2615-2626. 47. Dmitriev, L, Krasnykh, V., Miller, C. R., Wang, M., Kashentseva, E., Mikheeva, G., Belousova, N., and Curiel, D. T. (1998). An adenovirus vector vŝ ith genetically modified fibers demonstrates expanded tropism via utilization of a coxsackievirus and adenovirus receptor-independent cell entry mechanism. / . Virol. 72, 9706-9713. 48. Vigne, E., Mahfouz, I., Dedieu, J. -F., Brie, A., Perricaudet, M., and Yeh, P. (1999). RGD inclusion in the hexon monomer provides adenovirus type 5-based vectors with a fiber knob-independent pathv^ay for infection. / . Virol. 73, 5156-5161. 49. Biermann, V., Volpers, C., HuEmann, S., Stock, A., Kewes, H., Schiedner, G., Herrmann, A., and Kochanek, S. (2001). Targeting of high-capacity adenoviral vectors. Hum. Gene Ther. 12, 1757-1769. 50. Morral, N., Parks, R. J,, Zhou, H., Langston, C., Schiedner, G., Quinones, J., Graham, F. L., Kochanek, S., and Beaudet, A. L. (1998). High doses of a helper-dependent adenoviral vector yield supraphysiological levels of ?1-antitrypsin v\̂ ith negligible toxicity. Hum Gene Ther. 9, 2709-2716. 51. Chuah, M. K. L., Collen, D., and VandenDriessche, T. (2001). Gene therapy for hemophilia. / . G^w^M^^. 3, 3-20. 52. Balague, C., Zhou, J., Dai, Y., Alemany, R., Josephs, S. F., Andreason, G., Hariharan, M., Sethi, E., Prokopenko, E., Jan, H. -Y., Lou, Y. -C., Hubert-Leslie, D., Ruiz, L., and Zhang, W. -W. (2000). Sustained high-level expression of full-length human factor VIII and restoration of clotting activity in hemophilic mice using a minimal adenovirus vector. Blood 95, 820-828. 53. Gallo-Penn, A. M., Shirley, P. S., Andrev^s, J. L., Kayda, D. B., Pinkstaff, A. M., Kaloss, M., Tinlin, S., Cameron, C , Notley, C , Hough, C , Lillicrap, D., Kaleko, M., and Conelly, S. (1990). In vivo evaluation of an adenoviral vector encoding canine factor VIII: High-level, sustained expression in hemophilic mice. Hum. Gene Ther. 10, 1791-1802. 54. Chen, H. -H., Mack, L. M., Kelly, R., Ontell, M., Kochanek, S., and Clemens, P. R. (1997). Persistence in muscle of an adenoviral vector that lacks all viral genes, Proc.Natl.Acad.Sci. USA 94, 1645-1650. 55. Bennett, J., Wilson, J., Sun, D., Forbes, B., and Maguire, A. (1994). Adenovirus vector- mediated in vivo gene transfer into adult murine retina. Invest. Ophthalmol. Vis. Sci. 35, 2535-2542. 56. Li, T., Adamian, M., Roof, D. J., Berson, E. L., Dryja, T. P., Roessler, B. J., Davidson, B. L. (1994) In vivo transfer of a reporter gene to the retina mediated by an adenoviral vector. Invest. Ophthalmol. Vis. Sci. 35, 2543-2549. 57. Bennett, J., Tanabe, T., Sun, D., Zeng, Y., Kjeldbye, H., Gouras, P., and Maguire, A. M. (1996). Photoreceptor cell rescue in retinal degeneration (rd) mice by in vivo gene therapy. Nat. Med. 2, 649-654. 58. Anglade, E., and Csaky, K. G. (1998) Recombinant adenovirus-mediated gene transfer into the adult rat retina. Curr. Eye Res. 17, 316-321. 59. da Cruz, L., Robertson, T., Hall, M. O., Constable, I. J., and Rakoczy, P. E. (1998). Cell polarity, phagocytosis and viral gene transfer in cultured human retinal pigment epithelial cells. Curr. Eye Res. 17, 668-672. 60. Akimoto, M., Miyatake, S., Kogishi, J., Hangai, M., Okazaki, K., Takahashi, J. C , Saiki, M., Iw^aki, M., and Honda, Y. (1999). Adenovirally expressed basic fibroblast grov^th factor rescues photoreceptor cells in RCS rats. Invest. Ophthalmol. Vis. Sci. 40, 273-279. 4 4 6 Schiedner ef al. 61. Cayouette, M., and Gravel, C. (1997). Adenovirus-mediated gene transfer of ciliary neu rotrophic factor can prevent photoreceptor degeneration in the retinal degeneration (rd) mouse. Hum. Gene Ther. 8, 423-430. 62. Streilein, J. W. (1995). Unraveling immune privilege. Science 270^ 1158-1159. 63. Kumar-Singh, R., Yamashita, C. K., Tran, K., and Farber, D. B. (2000). Construction of encap- sidated (gutted) adenovirus minichromosomes and their application to rescue of photoreceptor degeneration. Methods Enzymol. 316, 724-743. 64. Lem, J., Flannery, J. G., Li, T., Applebury, M. L., Farber, D. B., and Simon, M. I. (1992). Reti nal degeneration is rescued in transgenic rd mice by expression of the cGMP phosphodiesterase beta subunit. Proc. Natl. Acad. Set. USA 89, 4422-4426. 65. Cregan, S. P., MacLaurin, J., Gendron, T. F., Callaghan, S. M., Park, D. S., Parks, R. J., Graham, F. L., Morley, P., and Slack, R. S. (2000). Helper-dependent adenovirus vectors: Their use as a gene delivery system to neurons. Gene Ther. 7, 1200-1209. 66. Hassell, J. A., Cukanidin, E., Fey, G., and Sambrook, J. (1978). The structure and expression of two defective adenovirus 2/simian virus 40 hybrids. / . Mol. Biol. 120, 209-247. C H A P T E R Xenogenic Adenoviral Vectors Gerald W. Both Molecular Science CSIRO North Ryde, New South Wales, Australia !• Impetus and Rationale Although human adenoviruses (HAdVs) have been extensively studied over the past four decades, it is only in the past 10 years or so that studies on animal adenoviruses have begun to approach the same level of molecular analysis. This w âs partly driven by the desire to characterize viruses that clearly had very different properties and host ranges compared with HAdV, but it was also recognized that natural infection of human populations would very likely induce a level of immunity that might curtail the effective use of HAdV vectors. Molecular studies of xenogenic AdVs have substantially expanded our knowledge. Understanding their biology will ultimately lead to an increased choice of gene delivery vectors, providing more options in therapeutic strategy and design. IL Classification of Adenoviruses Adenoviruses were classified originally on the basis of serological tests and hemagglutination ability (reviewed in [1, 2] but the availability of genetic data has enhanced the ability to assess viral relatedness. The great majority of AdVs are classified as members of the mastadenovirus genus. This group includes all known human and many AdVs of animal origin. Bovine, porcine, canine, murine, equine, simian, and ovine viruses are all represented, some as multiple serotypes [3]. The genus aviadenovirus has also been known for many years. This group consists exclusively of viruses of avian origin, as the name suggests. Again, multiple serotypes of fowl AdV occur, with the prototype virus being the FAdVl isolate known as CELO. A third group, ADENOVIRAL VECTORS FOR GENE THERAPY AA 7 Copyright 2002, Elsevier Science (USA). • • • * # All rights reserved. 4 4 8 Gerald W. Both proposed as a new genus called the atadenoviruses [4-7], comprises viruses from bovine, ovine, and avian species and, tentatively, viruses from goats, deer, and possum (B. Harrach and D. Thomson and H. Lehmkuhl, pers. commun.). The OAdV7 isolate 287 has been proposed as the prototype of this group [5, 8]. Turkey hemorrhagic enteritis virus (HEV) and frog virus (FrAdVl) may constitute a fourth genus [9]. To assist in defining the potential uses of each vector it is important to understand the host range and biology of each virus. Although some information has been gleaned from genetic data, for the nonmastadenoviruses especially, feŵ studies of the nonstructural viral gene products have been done. III. Factors Affecting Vector Design and Utility A. Host Range and Pathogenicity A driving force behind the development of HAdV vectors w âs the knowledge that they are not associated with significant disease in healthy individuals [1]. The production of defective vectors in complementing cell lines has provided an additional margin of safety [10]. Several of the xenogenic AdVs reviewed here are being adapted for use as vaccine vectors in the homologous host. Thus, it is important that wild-type BAdV3, PAdV3, CAdV2, FAdVl, and OAdV7 cause only mild or subclinical symptoms upon experimental infection of the species from which they were isolated [11-15]. When considering viral host range it is important to distinguish between host range defined by viral replication and host range defined by the ability to transduce cells. Transduction is influenced largely by the interaction between the fiber protein and a primary cellular receptor. Some avi- and mastaden- oviruses have a second fiber protein [16, 17]. The major primary receptor for HAdVs has been identified as Coxsaclcie and adenovirus receptor (CAR) [18, 19]. This is probably also used by SAdVs (Table I) because they grow well in human cells and were propagated in human embryonic kidney 293 cells [20]. For most other xenogenic AdVs no primary receptor has been characterized, nor is it clear whether secondary receptors such as integrins [21] are involved in virus uptake. Indeed, xenogenic AdVs lack identifiable or functionally con firmed integrin-binding sequences in their penton proteins [22-27]. For fiber, the coiled coil, trimeric structure of the stalk [28] is conserved, but the distinct sequences of the cell binding domains for the avi and atadenoviruses suggest that they utilize primary receptors that are distinct from CAR. Consistent with this, although HAdV5 and OAdV7 can both infect CSL503 ovine lung cells, they do not compete with each other for entry [29]. CAdV2 is the only xenogenic mastadenovirus that has been examined with respect to cell binding and uptake. Despite the differences between the HAdV5 and CAdV2 capsids 16. Xenogenic Adenoviral Vectors 449 Table 1 Complete Nucleotide Sequences of Xenogenic Adenovirus Vectors Virus type Isolate GenBank Accession No. Canine adenovirus RI261 NC_001734 CAdVl Canine adenovirus Toronto A26/61 U77082 CAdVl Bovine adenovirus WBR-1 AF030154 BAdV3 Porcine adenovirus 6618 AF083132 PAdV3 Murine adenovirus NC_000942 MAdVl Simian adenovirus US patent SAdV21 CI 6,083,176 SAdV25 C68 Fowl adenovirus CELO; Phelps U46933 FAdVl (ATCC VR-432) Fov^l adenovirus ATCC strain A-2A AF083975 FAdV8 Ovine
adenovirus 287 U40839 OAdV7 Duck adenovirus EDS strain 127 Y09598 DAdVl Frog adenovirus VR-896 AF224336 FrAdVl Turkey adenovirus Hemorrhagic AF074946 TAdV3 enteritis virus the kinetics of uptake and trafficking of the two viruses in dog kidney cells was surprisingly similar [30]. CAdV2 shares some features of AdV2/5 tropism but also exhibits distinct characteristics. For example, CAdV2-infected Chinese hamster ovary (CHO) cells that expressed human or mouse CAR but it did not bind to human dendritic cells that were efficiently infected by HAdV5. Uptake of CAdV2 in susceptible cells must be augmented principally by CAR because the Arg-Gly-Asp (RGD) motif that binds to a^^s integrin is absent from the CAdV2 penton. However, CAdV2 also appears capable of binding to other cell surface proteins [31]. Identifying the receptors for xenogenic adenoviruses and defining the mechanisms of virus uptake is important as it will allow target and nontarget cells to be identified, thus suggesting potential uses for each vector. 4 5 0 Gerald W. Both However, it is possible that amino acid variation between a natural viral recep tor and its counterpart on heterologous cells may alter virus binding affinity. B. Neutralization HAdVs are ubiquitous in the human population. As a result of natural infection most individuals develop immunity to adenoviruses by the time they reach maturity. Antibodies against multiple serotypes are common [32] and a substantial portion have neutralizing activity [33]. Nonneutralizing antibodies can also bind to virus particles, leading to their indirect inactivation via the com plement system [34]. In addition, individuals commonly develop a long-lived CD4+ T-cell response against multiple serotypes of human adenovirus [35] which may mitigate the strategy of using human adenoviruses from alternative serotypes as vectors [36, 37]. Apart from preexisting immunity, administra tion of a HAdV at high dose can elicit an inflammatory response [38]. The vector may also induce an immune response that can reduce the efficacy of subsequent doses, although the extent of this effect may vary with the route of administration [39, 40]. A variety of methods have been used to overcome these problems, including transient immunosuppression, blocking of antibod ies with agents such as polyethylene glycol and removal of antibodies from serum by immunoapheresis [41, and references therein]. The use of xenogenic adenovirus vectors is expected to avoid neutraliza tion due to preexisting immunity to HAdVs. To investigate this, random human sera were examined for the presence of antibodies that neutralized OAdV7 or CAdV2. Of a panel of 57 sera, most of which neutralized HAdV5 to high titer, only three also neutralized CAdV2 [42, 43]. Similarly, 13 individual and two pools of human sera that neutralized HAdV5 did not neutralize OAdV7 [44]. SAdVs were also not neutralized by antisera that neutralized HAdVs [45]. These data suggest that xenogenic adenoviruses will provide an advantage upon initial administration although it is not expected that the vector will be immuno logically silent. However, whether vector is given locally or systemically may determine whether it is possible to administer more than a single dose [39, 40]. C. Genome Structure and Function Of the xenogenic mastadenoviruses, complete nucleotide sequences have been determined for bovine (BAdV2 and 3) [24], porcine (PAdV3) [46], murine (MAdVl) [27], canine viruses (CAdVl and 2) [26], and simian viruses [20] (Table I). For the aviadenoviruses, FAdVl [23] was the first genome sequenced but FAdV8 [47] is now also completed. Among the atadenoviruses, ovine (OAdV7) [22], bovine (BAdV4) (B. Harrach, pers. commun.), and duck (DAdVl) [48] genomes are sequenced. The turkey (TAdVl) [49] and frog (FrAdVl) genomes [9] have also been characterized. All of these viruses 1 6 . Xenogenic A d e n o v i r a l Vectors 451 are potential vectors for gene delivery because they can nov^ be rationally engineered, but not all are being developed as vectors at this stage. The viruses described above represent the extreme ranges of genome size, the largest being ^43.8 and 45 kb for FAdVl and FAdV8, respectively, and the smallest being -26 .3 kb for TAdVl and -29 .5 kb for OAdV7. The Mas- tadenovirus genomes range in size from —30.9 (MAdVl) to 34.4 kb (BAdV3). 1. Central Core In comparing the nucleotide sequence for prototype viruses in each genus it is apparent that there is a central core in each genome bounded by the pVIII and IVai genes (Fig. 1). This codes for the DNA replication, structural proteins, and accessory polypeptides required for their assembly. Most capsid proteins have homologs in each genus but proteins V and IX are unique to mastadenoviruses. Instead, OAdV7 has a gene for the structural Rep Gam ,4- dUTPase CU D PVIII fib 1 fib2 P I =11 I a FAdVI D O d Z I pol ca c=3 cz IVa2 22 a I I AVIADENO 2kb I- — - - I E1A E1B IX PVIII E3 fib •=3 PI 31 HAdV5 MASTADENO IVa2 pol E4 Site I I Sites ill II E1B pVlll k• fib DPI I P ii 0AdV7 a D„ 0 D„ ATADENO p32K IVa2 pol ' 1 0 D =3 0 3-1 E4 6-1RH Figure 1 Comparison of the genome structures of prototype viruses from the ovi-, most-, and atadenoviruses. The central core of each genome (filled rectangle) flanked by the IVa2 and pVIII genes is essentially conserved in arrangement and is truncated for simplicity. Other major open reading frames are indicated by open rectangles. Arrows indicate sites for insertion of foreign gene cassettes. The solid and broken lines indicate regions that can be provided in trans and regions that can be deleted, respectively. Note that E4 and E2 sequences have also been deleted in HAdV5 and SAdV vectors but this has not been demonstrated for other xenogenic mastadenoviruses. 4 5 2 Gerald W. Both protein, p32 K, that lies at the extreme left end of the genome (Fig. 1). This capsid protein complement correlates with the observation that FAdVl and OAdV7 are more heat-stable than the mastadenoviruses [50, 51]. It will be of interest to determine whether a functionally equivalent protein exists for the aviadenoviruses. Also in the central core of HAdVs are one or two copies of VA RNA genes [52]. Except for PAdV3 [46] and SAdVs [53], these are not present in xenogenic mastadenoviruses or OAdV7 [54], but a single copy is present near the right-hand end of FAdVl [55] and DAdVl (Fig. 1) [23, 48]. 2. Right-End Sequences To the right of the central core the genomes vary greatly in structure and gene complement. In the mastadenoviruses the E3 region varies in size and complexity but is located between the pVIII and fiber genes (Fig. 1). HAdV2 and —5 have an E3 region of ^2.5 kb that codes for numerous polypeptides, many of which interact with components of the immune system [56]. For the xenogenic mastadenoviruses the least complex E3 region from MAdVl appears to encode a single reading frame that may be variably spliced [57., 58]. BAdVl, —2, and —3, CAdV2, and PAdV3 and —5 have E3 regions of intermediate complexity, ranging in size from ^^1.2 to 2.3 kb. These code for a variable number of putative proteins that show some homology within a species and occasionally across species [59-63]. The BAdV3 E3 codes for a 284-residue glycoprotein and a 14.7-kDa polypeptide that appears to be the homolog of the HAdV5 14.7-kDa protein. The BAdV3 gene can functionally substitute for the human gene to protect cells against tumor necrosis factor (TNF)-induced lysis [62, 64]. E3 sequences are nonessential for replication in vitro [65] and were some of the first sequences deliberately deleted in the construction of recombinant HAdVs [66]. However, it was shown that retention of E3 sequences in a HAdV5 vector dampened the immune response in a rat model, thus extending the time of transgene expression [67]. Consistent with this, a HAdV5 virus in which E3 sequences were deleted showed an enhanced inflammatory response in a Cotton rat model [68]. It remains to be determined whether these results will translate to xenogenic vectors with less complex E3 regions. However, the timing and duration of gene expression that is required is a factor to be considered in vector design. In contrast to the mastadenoviruses, the avi- and atadenoviruses lack E3 regions between pVIII and fiber and instead have small intergenic regions of ^200 and ~400 bp, respectively, that contain signals for transcription termination and splicing of fiber RNA [69]. To the right of the fiber gene in mastadenoviruses lies the E4 region. Like HAdV2/5, a single promoter in BAdV3 and MAdVl produces seven transcripts that encode multiple polypeptides, some of which are homologous to HAdV proteins [70, 71]. In particular, homologs of HAdV5 E4 ORF6 carry a short 16. Xenogenic Adenoviral Vectors 4 5 3 amino acid motif that is highly conserved in many adenoviruses. Based on the conservation of this motif in OAdV7, v^here it w âs first recognized [22], the proposed E4 region in the atadenoviruses is penuhimate to the right end of the genome (Fig. 1). Tw ô promoters apparently control the expression of three open reading frames (ORFs), tv^o of w^hich contain the motif [48, 72]. No E4 region has been identified in the right-hand portion of aviadenoviruses. Indeed, the function of most reading frames in the right hand ~ 2 5 % of the genome remains to be determined. In FAdVl, the products of GAM-1 and ORF22 (Fig. 1) have been identified as proteins that interact w îth pRb [73]. However, in comparing the related FAdVl and FAdV8 genomes, 5 of 13 unassigned ORFs are unique to FAdV8 [74]. At the extreme right hand end of FAdVl are 3 ORFs that can be deleted and replaced v^ith a luciferase reporter gene cassette w^ithout affecting virus viability [51]. The extreme right ends of the avi- and atadenovirus genomes carry genes that are species specific. For example, DAdVl has numerous ORFs of unknov^n function that have no counterpart in OAdV7 [22, 23, 48, 75]. Within the right-hand-end region of OAdV7 lies a series of six short reading frames (RHl to RH6) (Fig. 1), four of which (RHl, - 2 , - 4 , and - 6 ) are closely related to each other. This is surprising in a compact genome of only ^29.6 kb. In DAdVl there are two ORFs that are related to each other and to those in OAdV7 [72, 76]. For OAdV7 only two transcripts from the region were detected by RT-PCR and these were spliced such that RHl and RH6 were the only ORFs that could be translated. The apparent redundancy of these ORFs was confirmed by the fact that the reading frames RH2 to RH5 could be deleted without seriously affecting virus viability [75]. The function of these ORFs remains to be determined. 3. Left End Sequences Left of the central core the genome structures also differ significantly (Fig. 1). For the xenogenic mastadenoviruses there are three ORFs at the left end that show homology with HAdVs [77-80]. The genome packaging signal is also present within the first ^500 nucleotides of the HAdV5 genome [81], but until recently this had not been defined for any xenogenic virus. For CAdV2, however, it was shown that the packaging region consists of a ^200- bp region that contains redundant, but not functionally equivalent sequences. The consensus sequence for HAdVs [81] is present only once and is of minor importance [82]. For the avi and atadenoviruses, some ORFs at the left end are unique to individual viruses or have homologs only within the genus. Two ORFs from the atadenoviruses show some homology with the HAdV5 ElB 19- and 55-kDa genes, suggesting that these functions are conserved. However, no homolog of the El A gene was identified [22,48]. An additional ORE that could encode a ^9.6-kDa protein is present in OAdV7 and BAdV4 (B. Harrach, pers. 4 5 4 Gerald W. Both commun.) but is missing from DAdVl [48]. The gene for the p32 K structural protein is also present near the left end. The promoter for ORFs LHl and LH2 is also on the opposite strand within this gene [72]. The packaging signal for atadenoviruses has not been defined but it may incorporate the ~160-bp region between the C-terminus of p32 K and the ITR. For the aviadenoviruses, ORFs with homologies to dUTPase and the REP protein of adeno-associated virus have
been found [23] but there are distinct differences between FAdVl and FAdV8 with three of eight ORFs in the left end being unique to FAdVS [74]. It was also reported recently [83] that the cysteines and several other residues in the conserved sequence motif of E4 ORF6 are conserved in FAdVl ORF14, which lies near the left end of the genome [23]. 4. Transcription Maps The determination of transcription maps for some xenogenic viruses has assisted vector design by complementing the data on genome structure. The major transcription units have been described for BAdV3 and PAdV3 [24, 46, 71], FAdVl [84] and OAdV7 [72]. No transcription map has been reported for CAdV2. More detailed data is available for the El , E3, and E4 regions of MAdVl [58, 70, 85] and for the BAdV3 El \13, 86] and E3 regions [60]. For BAdV3 and PAdV3, there are minor differences in the splicing pattern within some transcription units but on a broader scale the basic units described for AdV2/5 are completely conserved. Studies of FAdVl identified many transcripts for ORFs in the genome and a major transcription unit that is controlled by the MLP. However, at the left and right ends of the genome there are 5 and 15 kb, respectively, for which the promoters and transcriptional organization is undefined [84]. In the OAdV7 genome, the left (LHl to LH3)- and right-hand ends (RHl to RH6), E2 and the proposed E4 region (E4.1 to E4.3), as well as the structural protein genes constitute individual transcription units. The IVa2 and p32 K ORFs also appear to be transcribed from their own promoters. The LH and E4 regions each appear to be regulated by two promoters [72]. The identification of promoter regions and transcription termination sites has identified possible sites for gene insertion that are less likely to interfere with viral functions. D. Transforming Ability Many AdVs are known to carry oncogenes. Members of the mastaden- oviruses readily transform cells in culture [11, 87-89], although these viruses differ in their ability to induce tumor formation in animals. Among HAdVs, the group A viruses such as AdV12 are highly oncogenic, while group C (including HAdV5) and E viruses are not known to be tumourigenic (reviewed in [2, GS^ 90]). BAdV3 can induce tumor formation in hamsters [91] but there are no reports of tumor induction by other animal mastadenoviruses. FAdVl 16. Xenogenic Adenoviral Vectors 4 5 5 also transforms cell in vitro [92, 93] and rapidly induces tumors in newborn rodents [94, 95], For the atadenoviruses there are conflicting reports of tumor induction in hamsters. In one study, tumor formation was reported in ham sters inoculated with BAdV8 [96]. In a second study, none of BAdV4 to -10 produced tumors [97]. More recent studies showed that OAdV7 was unable to transform cells that were transformed by HAdV5 [98]. Primary rat embryo cells were infected with HAdV5 or OAdV7 but only the former produced colonies with a transformed phenotype. Similarly, baby rat kidney cells were transformed by HAdV5 El A/B sequences but not by the nonstructural genes of OAdV7. The apparent absence of oncogenes in the OAdV7 genome suggests that the virus interacts with the cell cycle machinery in a way that differs from the mast and aviadenoviruses, although this is yet to be defined. The presence of oncogenes in vector genomes has important implications for vector design in that it is customary to delete these sequences for safety reasons. Continuous cell lines that express the deleted genes in trans are established to permit virus propagation. The transforming properties of the mastadenoviruses reside primarily in the ElA and ElB genes at the left end of the genome (reviewed in [2, GS^ 90]). The ElA products bind to proteins of the cellular retinoblastoma (pRb) protein family [99], thereby releasing E2F transcription factors that regulate cell cycle progression into S phase [100]. The ElB 55-kDa protein binds to the tumor suppressor protein, p53, and blocks p53-mediated apoptosis [101]. The ElB 19-kDa protein is also anti-apoptotic [102]. Thus, animal adenoviruses typified by BAdV3, PAdV3, CAdV2, SAdV, and MAdVl have ElA and ElB homologs that have similar transforming and oncogenic potential. The E4 ORF3 and ORF6 products of HAdV5 can also augment the transforming activity of the ElA and ElB genes [103-106]. However, the E4 regions of human and animal mastadenoviruses vary in sequence and complexity. Homology with HAdV5 ORF6 is always evident, especially in a cysteine-rich motif [22] that is thought to mediate ORF6/p53 interaction [83]. Furthermore, a complex between the E4 ORF6 and ElB 55-kDa proteins promotes the selective nuclear export of late viral transcripts [107] and references therein). This ORE may be therefore be conserved as it provides a core function for replication in all adenoviruses. However, other E4 ORFs in the xenogenic viruses are unique [22,26,108-110] and their function/transforming potential is not clear. In FAdVl there are no identifiable ElA/B or E4 regions in the genome [23], but recently two proteins, GAM-1 and ORF22, that interact with pRb were identified [73]. In addition, GAM-1 has been identified as an anti- apoptotic protein [111] and one that can activate the cellular heat-shock response, the latter being required for viral replication. The Hsp40 gene is a primary target [112]. GAM-1 may also functionally substitute for the ElB 19 kDa [111]. FAdVl therefore appears to share with the mastadenoviruses an ability to disrupt complexes between pRb and the E2F transcription 4 5 6 Gerald W. Both factors to modulate the cell cycle, albeit via different effector proteins [99, 113]. In contrast, OAdV7, the prototype atadenovirus, lacks an identifiable ElA homolog, although it appears to carry ElB 19- and 55-kDa genes. Penultimate to the right end is a transcription unit that contains a unique ORF (E4.1) of unknown function and two ORFs (E4.2 and E4.3) which contain the conserved ORF6 cysteine-rich motif mentioned above [22, 72, 98]. These ORFs otherwise appear unrelated. Similar features are found in the DAdVl and BAdV4 genomes [48 and B. Harrach, pers. commun.]. However, OAdV7 so far lacks oncogenic activity as the complete OAdV7 genome did not transform primary rodent cells under conditions where transformation was achieved with control HAdV5 sequences [98]. These findings invite the hypothesis that OAdV7 lacks the ability to activate the cell cycle in quiescent cells, instead taking advantage of the cycle as it progresses. The presence or absence of transforming sequences strongly influences the design of xenogenic adenovirus vectors for gene delivery. Based on HAdV2 and —5, vectors derived from BAdV3, PAdV3, SAdV, and CAdV2 were designed such that the potentially oncogenic ElA/B homologs were deleted [20, 42, 43, 114, 115]. A similar approach could be applied to MAdVl [116]. Such vectors are replication-defective in cells lines that do not express the deleted genes [42, 43], but in some cases, homologs from HAdV5 can substitute [114, 115]. Some vectors derived from OAdV7, avian, and PAdV3 viruses retain potential transforming genes and carry foreign DNA inserts in nonessential regions of the genome [51, 75^ 117-120]. This strategy may be acceptable for vectors that are intended for gene delivery in the homologous animal or avian host but is unlikely to be acceptable for gene therapy purposes, except perhaps in the case of OAdV7, where the vector apparently lacks transforming genes. E. Cell Lines Successful rescue of a virus requires a cell line that can be transfected with high efficiency to initiate infection. The cells should also have abundant copies of the primary and secondary receptors to facilitate spread and the production of high titers of virus. Depending on the recombinant, the cells may or may not carry viral sequences to complement a deletion in the viral genome. 1. Primary Cell Lines The general strategy has been to identify a cell line that is permissive for the wild-type virus and then adapt it for more specialized purposes. For prop agation of BAdV3, MDBK, buffalo lung, primary kidney, and bovine cornea endothelial cells have all been tried, with MDBK cells being preferred [114, 121]. CAdV2 was grown in MDCK, dog kidney (ATCC CRL6247) or grey hound kidney [43, 121], MAdVl in mouse 3T6 [116] and PAdV3 in swine testis cells [115]. FAdVl recombinants were rescued in leghorn male hep atoma (LMH) cells [51]. FAdVl can be grown in embryonic chicken kidney 16. Xenogenic Adenoviral Vectors 4 5 7 cells but, for reasons of cost, is often grown in embryonated chicken eggs [23]. OAdV7 has a narrow host range and failed to grow in several ovine cell types [15]. However, it grew to high titre in CSL503 cells, a primary ovine fetal lung cell line [122] and a fetal ovine skin fibroblast line HVO-156 (C. Hofmann and P. Loser, pers. commun.). 2. Transformed Cell Lines Primary cells are adequate for growing replication-competent recombi nants. However, there was a need to produce cell lines that would complement genomic deletions and an expectation that transformed cell lines would ensure a continuous supply of cells. This encouraged attempts to develop lines equiv alent to 293 cells [123]. Note that SAdVs grow in 293 cells [20]. Based on this and similar precedents [124], the ElA/B sequences of BAdV3 were used to stably transfect MDBK cells [114, 125, 126]. These grew poorly and expressed undetectable amounts of the BAdV3 El proteins [114] but nevertheless com plemented the growth of an ElA-deleted HAdV5/lacZ recombinant [114, 125]. Attempts were also made, unsuccessfully, to transfect foetal bovine retinal cells (FBRCs) with BAdV3 ElA/B sequences. Because BAdV3 comple mented HAdV5/ElA-defective replication [125], it was expected that HAdV5 ElA/B sequences would complement BAdV3/El deleted vectors. Transfection of FBRCs with HAdV5 ElA/B sequences in which El A and ElB were controlled by the mouse PGK and ElB promoters, respectively, produced morphologically distinct clones, one of which was single-cell cloned and characterized as the VIDO R2 line. These cells expressed detectable levels of El A and ElB 19-kDa, but not ElB 55-kDa protein, supported plaque formation by BAdV3 and HAdV5, and were transfected more efficiently than MDBK cells. Transfection of El-deleted recombinant genomes into VIDO R2 cells resulted in the rescue of several viruses that carried expression cassettes [114]. For propagation of PAdV3 vectors a transformed fetal porcine retinal cell line (VIDO Rl) was also produced by transfection of swine testis cells with HAdV5 El sequences [115]. These cells were also morphologically distinct from the parental cells. ElA and ElB 19-kDa proteins were produced, as shown by Western blots, but ElB 55-kDa protein was not detected. While PAdV3 grew well in these cells, for reasons that are not understood, an E1/E3-deleted vector and a similar virus that carried a GFP cassette in El grew two logs less efficiently [115]. Similarly, the ElA/B region of CAdV2 was used to transform MDBK and DK cells [42]. Again, low levels of ElA transcripts were produced and ElB transcripts were not reliably detected. Nevertheless, the cells were morphologi cally and phenotypically distinct from parental MDCK cells. A second series of clones was produced by transfecting DK cells with CAdV2 sequences in which ElA and ElB were controlled by the HCMV and ElB promoters, respectively. Cells produced in this way expressed detectable ElA and ElB transcripts and ElB 19-kDa protein [42] and allowed the rescue and propagation of ElA/B-deleted CAdV2 vectors [43]. 4 5 8 Gerald W. Both Attempts were also made to produce a transformed derivative of CSL503 cells, which are permissive for OAdV7, using the left end (~4 kb) of the OAdV7 genome [9S], The sequences used incorporated the proposed ElB homologs of OAdV7 and a 9.6-kDa ORF of unknown function. No ElA homolog was identified [72, 98]. Only two clones that grew well enough to prepare frozen stocks were obtained and these were morphologically similar to the parental cells. In contrast, transfection of CSL503 cells with HAdV5 ElA/B sequences produced morphologically distinct clones. Growth of OAdV7 in these cell lines appeared to be retarded compared with its growth in wild type CSL503 cells (Xu and Both, unpublished results). F. Strategies for Vector Construction and Rescue A huge amount of work carried out over some 30 years on HAdV2 and —5 has defined viral promoters, transcripts and their splice sites and genes that could be deleted or that would function in trans (reviewed in [2, 65]). The packaging capacity of the viral capsid was also shown to
be ^^105% of the viral genome [127]. The strategic design of bovine, canine, porcine, simian, and murine adenovirus vectors, although based on new genetic information, has drawn extensively on historic precedents. As precedents did not exist for the avi- and atadenoviruses it was necessary to identify intergenic regions within genomic sequences and to use mutagenesis to identify nonessential reading frames for the insertion of gene cassettes. Vector design and virus rescue was also confounded initially by the absence of transcription maps and the lack of knowledge concerning the packaging capacity of these viruses. Construction of xenogenic adenovirus vectors first required the identifi cation of an insertion site(s) that could stably accommodate a gene cassette without affecting virus growth. For the mastadenoviruses, vector construction strategies followed those for human Ad vectors. Genes were inserted into the nonessential E3 region of BAdV3 [126, 128] or PAdV3 [118] or between the E4 promoter and the right ITR of PAdV3 [118, 120] to generate viruses that were replication competent in noncomplementing cells. More recently, ElA/B region replacements that generated replication-deficient viruses were produced for BAdV3 [114, 121], PAdV3 [115], SAdV [20], and CAdV2 [42, 43]. It is likely that a similar strategy would be successful for MAdVl where an infec tious clone is now available [116]. For the aviadenoviruses, a mutation strategy was used to identify nonessential regions of the genome or regions that could be complemented in trans [51]. Deletions between nucleotides 938 and 2900 were complemented by cotransfection of a plasmid that carried the left hand ^5.5 kb of the genome. Deletion of three ORFs adjacent to the right end of the FAdVl genome did not require transcomplementation, identifying these genes as nonessential for replication in vitro. Similarly, replication-competent FAdV8 16. Xenogenic Adenoviral Vectors 4 5 9 vectors were constructed by inserting a gene cassette into sites near the right end of the genome (Fig. 1) [119]. For the atadenovirus, OAdV7, genes were initially inserted at site I (Fig. 1) in the pVIII and fiber intergenic region [44, 50, IS^ 117, 129], but additional sites were subsequently identified by a mutation strategy. It was found that foreign DNA could be inserted into a unique ?ial\ site (Site II) within ORF RH2, ^\ kb from the right end, and that ORFs in the vicinity could be deleted \1S\. In addition, unique cloning sites were tolerated between the right-hand end and E4 transcription units (Site III) [72, 117]. The identification of permissible insertion sites in the genome required the construction of plasmids that enabled the rescue of infectious viruses. The first BAdV3 recombinant was constructed by recombination between a plasmid that carried BAdV3 sequences flanking the luciferase gene inserted into E3 and BAdV3 genomic DNA that had been cut with fvu\ to reduce background. DNAs were transfected into MDBK cells that also expressed BAdV3 El sequences. However, this method was inefficient and produced relatively few plaques [126]. Similarly, a CAdV2 recombinant that expressed the lacZ gene was produced by recombination between the CAdV2 (Manhattan strain) genome and a plasmid that carried the expression cassette. However, this recombinant was contaminated by wild-type CAdV2 that could not be eliminated [42]. A more favorable approach was to construct a plasmid in which sequences required for the propagation of plasmid DNA in Escherichia coli were cloned into a unique restriction enzyme site that linked the ITRs of the viral genome [117]. There was one precedent for this approach [130], although others had reported that perfect palindromes longer than 30 bp were often unstable in E. coli [131] and plasmids with large palindromes based on HAdV5 were subject to rearrangement [132]. Unique restriction sites were also introduced into appropriate locations in the OAdV7 genome to allow cloning of gene cassettes. This plasmid design allowed the genome to be released intact by restriction enzyme digestion prior to its transfection into susceptible cells for virus rescue [75, 117]. Subsequently, it was discovered that such plasmids could be constructed using recombination in £. coli [133]. Infectious recombinant clones have now been constructed for BAdV3 [71, 114, 134], PAdV3 [118], CAdV2 (Toronto strain) [43], MAdVl [116], FAdVl [51], and OAdV7 [44]. The specific infectivity of these naked DNAs in the permissive cell line is usually low (often only a few plaques per microgram) and depends on the transfection efficiency of the cells. However, a significant advantage of this approach is that transfection of purified plasmid DNA almost invariably yields the corresponding virus without the need for extensive plaque purification that may accompany other approaches where background viruses can be generated. Many xenogenic recombinant viruses have now been rescued. New viruses that first appeared with the formation of plaques or a cytopathic effect in the appropriate transfected cell line were amplified on fresh permissive cells to produce an infectious stock. Viruses were then characterized by restriction 4 6 0 Gerald W. Both enzyme, Southern blot [75,115,118,126], or PCR analysis [43, 51] to confirm the integrity of the genome and expression cassette. For the vectors where an insertion strategy was pursued, it was particularly important to check the genome integrity because the packaging capacity of the new vectors was undefined. Mastadenoviruses can package 105 to ^107% of the wild-type genome [43, 120, 127], OAdV7 has a capacity of 114%, presumably because of its smaller genome and similar capsid volume [75], while the capacity of aviadenoviruses is undefined. Despite the increased packaging capacity of OAdV7, some viral genomes in which expression cassettes (ranging from 1.8 to 3.1 kb) were inserted into site I of the genome proved to be unstable upon passaging. By passage three, the genomic BamHl profile of viruses that combined the HCMV promoter with a reporter gene sometimes displayed smaller fragments [50]. In contrast, a virus that carried 4.3 kb of "stuffer" DNA was successfully rescued [75] and with the RSV promoter, two viruses with site I cassettes in opposite orientations were stable to at least passage four [44; unpublished results]. Site I instability appears to vary with sequence and possibly orientation and may reflect the need to produce adequate amounts of fiber transcript and protein. Events that lead to transgene deletion with improved fiber production may generate viruses that have a growth advantage. The stability after passage of genomes for other xenogenic recombinant vectors has not been adequately reported. The propagation of a mixed population of CAdV2 wild-type and deleted vector [42] illustrated the potential for producing gutless vectors based on xenogenic AdVs. The principles established with HAdV5 [135, 136] will further assist this process. It will be necessary to define the packaging signal [82] and a minimum permissible genome size for a particular virus, provide a suitable a helper virus for propagation, and devise a means to purify defective particles. The benefits may be greater safety and more efficient gene delivery in a naive host and prolonged transgene gene expression. IV. Utility of Xenogenic Vectors Xenogenic AdV vectors can potentially be used as gene delivery vectors for a range of purposes. However, it is necessary to understand the advan tages and disadvantages of vectors in particular situations so as to identify their most appropriate uses. The next section discusses the first attempts to determine the safety and utility of xenogenic vectors for vaccination or gene delivery. The following section reviews the properties and behavior of vectors in heterologous situations. A. Veterinary Studies Within the limits of the testing done so far, the viruses discussed in this review are of low pathogenicity in the host from which they were iso lated [11-15]. Vectors designed for use in those hosts are often replication 16. Xenogenic Adenoviral Vectors 4 6 1 competent to facilitate vaccination by a live viral vector. In the first studies, carried out v^ith BAdV3, the luciferase reporter v^as inserted directly into the E3 region where a small deletion had been introduced. Expression did not require an exogenous promoter and the vector remained replication competent in bovine cells, although its titer was reduced 10-fold [126]. In contrast to HAdV5 vectors that lacked part of the E3 region [6S]^ this BAdV3 recombi nant did not show increased pathogenicity in a Cotton rat model compared with the wild-type virus [137]. Similar replication-competent viruses that car ried various forms of the bovine herpesvirus gD gene were shown to express the antigen [71] in an immunogenic form [128]. Intranasal vaccination of calves with these viruses induced gD-specific neutralizing antibodies, primed a cellular immune response and protected against viral challenge, despite the presence of preexisting serum antibodies to BAdV3 [138,139]. El/E3-deleted replication-defective BAdV3 vectors that carried gD in the El region were also constructed [114]. These viruses allowed the parameters for vaccination of cattle by replicating and nonreplicating vectors to be compared. Adminis tration of each vector at the same dose twice via the intra tracheal route and once subcutaneously showed that the replication-competent vector induced superior levels of serum IgG antibodies against gD. Partial protection against challenge was obtained with the replication-competent vector. However, with the replication-defective vector challenge with BHVl dramatically boosted the levels of serum IgG and IgA antibodies, suggesting that animals had been primed for gD-specific antibody responses [140]. Similar BAdV3 recombinants were constructed in which the bovine diarrhea virus E2 glycoprotein linked to the BHVl gD signal peptide was expressed from the BAdV3 E3/MLP [141]. The 53-kDa protein that was expressed formed dimers and was recognized by E2 specific monoclonal antibodies. Intranasal immunization of Cotton rats with the recombinant induced E2-specific IgA and IgG responses at mucosal surfaces and in the serum. In contrast, attempts to construct vectors that expressed the bovine coronavirus hemagglutinin esterase gene from the E3 region using the strategy for the BHVl gD gene were unsuccessful. The addi tion of exogenous control elements comprising an intron and the HCMV or SV40 promoter increased the level of expression but altered the kinetics. The recombinant virus also replicated less efficiently than wild-type BAdV3 [142]. Replication-competent PAdV3 vectors that express the pseudorabies gD protein or the classical swine fever virus (CSFV) gp55 protein were also constructed. The gD gene was inserted into a partially deleted E3 region without flanking sequences. In contrast to similar BAdV3 vectors, expression of gD was observed at early but not late times pi [118]. The gp55 gene linked to the PAdV3 MLP and tripartite leader sequence (TLS) was inserted at the right end between the ITR and E4 promoter. Vaccination of outbred pigs with a single dose of recombinant virus induced complete protection from lethal challenge with CSFV [120]. 4 6 2 Gerald W. Both A FAdV8 recombinant that expressed chicken gamma interferon from the viral MLP/TLS sequences was also constructed by inserting the cassette at sites near the right end [119]. Depending on the insertion site, the recombinants displayed differing growth characteristics in chicken kidney monolayers. Inser tion of the cassette adjacent to FAdV8 ORF7, about 7.2 kb from the right end, produced a recombinant with wild-type growth characteristics. In contrast to the FAdVl viruses discussed below, deletion of the FAdVl 36-kDa homolog in FAdV8 caused a significant reduction in growth. Interferon was produced in supernatants as early as 24 h pi in proportion to the growth characteristics of each virus in vitro. Interferon levels peaked at 48 h and were maintained for at least 10 days. Chickens treated with the recombinant showed increased weight gains compared to controls and suffered reduced weight loss when challenged with a coccidial parasite [119]. An OAdV7 vector was constructed in which the 45 W antigen of Taenia ovis was expressed from the viral MLP/TLS elements [143], the cassette being inserted at site I (Fig. 1) [75]. This vector was used alone, or in tandem with DNA or purified 45W protein to vaccinate sheep. Prime/boost strategies where vaccination was initiated with protein or DNA and boosted with the OAdV7 vector were effective in stimulating an immune response that protected animals against challenge with the parasite [144]. The above examples illustrate that with further refinement, xenogenic vec tors may have utility for vaccination and gene delivery in their respective hosts. B. Vector Biology Ideally, vectors for gene transfer into human cells should be capable of transgene expression without replication or detrimental expression of viral genes. Infection of human cell lines with intact xenogenic adenoviruses estab lished the principle that these viruses are replication defective at the inputs
tested [42, 51, 76^ 116, 118, 121, 126], although the molecular basis for defective replication is not understood. Studies in animal models have also allowed biodistribution profiles to be determined for some viruses. 1. Transduction of Cells Selected cell lines have been used to examine viral transduction. However, it is sometimes difficult to compare data from different laboratories because, especially in early studies, the input virus was not characterized with respect to both particle number and infectivity. BAdV3 recombinants in which a HCMV/lacZ or HCMV/GFP gene cassette was expressed from the El or E3 region, respectively [114, 121], were used to infect human and other cell types. The GFP recombinant replicated in cells of bovine origin and in Cotton rat lung fibroblasts, but not in cells from other species. When cells were infected with more than 5 pfu/cell of BAdV3/GFP, some GFP expression was observed at 16. Xenogenic Adenoviral Vectors 4 6 3 3 days pi in 293 and HeLa but not in A549 or HepG2 cells [114]. In contrast, others found that at an m.o.i of 10 pfu/cell, at 65 h pi A549 and MRC5 cells were efficiently transduced by a BAdV3/lacZ recombinant while HeLa and 293 and primary human muscle cells were transduced with lower efficiency [121]. Since both studies used the HCMV promoter and a similar multiplicity of infection, the reason for the difference with A549 cells is unclear. The host range of CAdV2 vectors was also investigated. Human 293, HeLa, primary myocyte, and HIB cells were infected with 10^ transduction units of CAdV2/RSVlacZ in the presence of wild-type CAdV2. All cell types showed P-gal expression when examined at 1 to 2 days pi [42]. In addition, replication-deficient CAdV2 vectors expressing GFP or lacZ from the HCMV and RSV promoters, respectively, were tested for their ability to transduce a range of human cell types in comparison with HAdV5/HCMV/GFP [43]. At 2 days pi HeLa, A172, and HT 1080 cells were transduced with similar efficiency by both viruses. In vivo^ the CAdV2 vectors also transduced mouse airway epithelia cells with similar efficiency to a comparable HAdV5 vector. Similarly, a replication-deficient PAdV3 recombinant carrying a HCMV/GFP cassette in El was used to determine the ability of this vector to infect human and animal cells in vitro. At a m.o.i. of 1 pfu/cell PAdV3 apparently entered, but did not replicate in canine kidney, ovine skin fibroblasts, bovine (MDBK), and human (293, A549) cells [115]. Although an infectious clone of MAdVl now exists, recombinant viruses have not yet been made. However, it was demonstrated by RT PCR that human 293 and primary umbilical endothelial cells were infected, the latter at low efficiency [116]. Replication-competent aviadenovirus vectors that express luciferase from the HCMV promoter [51] were constructed by inserting cassettes at the right end of FAdVl to replace nonessential ORFs. Vectors replicated in LMH cells with kinetics similar to wild-type FAdVl. When compared to a HAdV5/luciferase recombinant for its ability to transduce human cell types, the FAdVl recombinant showed a similar ability to express luciferase in HepG2, A549 and primary human fibroblasts [51]. Several recombinants that carried reporter genes at site I of the genome (Fig. 1) were used to investigate the host range of OAdV7 [29, 44, 50, 76,129]. These studies showed that OAdV7 can infect, but not replicate in a variety of human cell types, including breast (MCF7, T47D2) and prostate cancer (PC3), liver carcinoma (HepG2), and retinal (911), foreskin (HFF), and lung (MRC5) fibroblasts [76], Reporter gene expression increased proportionally with the m.o.i. Monkey (COS) and mouse prostate (RMl) cells were also infected efficiently in vitro [50 and unpublished results]. Considering the quite broad host range of OAdV7, it will be of consider able interest to identify the receptor(s) that mediates infection. In principle it is also possible to redirect the vector via an alternative receptor as was done for 4 6 4 Gerald W. Both HAdVs [36, 37]. It was shown [29] ttiat ttie cell-binding domain of OAdV7 fiber protein could be replaced with the equivalent binding domain from HAdV5. This was the only change in the viral capsid but it profoundly altered the cell tropism of OAdV7, apparently independent of any integrin/penton RGD interaction, since this motif is absent from OAdV7 [22]. Although the hybrid virus grew less well, this result confirmed that the two viruses use distinct receptors and demonstrated that targeting of xenogenic viruses may be possible. 2. Abortive Replication in Vitro Abortive replication of xenogenic adenoviruses probably reflects viral promoter function in human cell types. The function of early and late BAdV3 promoters in human cells was examined by RT PCR and Southern blot ting [121]. In A549 and 293 cells ElA transcripts were detectable for at least 5 days. At very high m.o.i. hexon mRNA was detectable at day 3 in primary human muscle, MRC5 human lung fibroblasts, and nasal septum epithelial cells. It was also shown that CAdV2 replicated to a limited extent in some human cells, as judged by higher virus output compared with input and some expression of capsid proteins. However, this was observed only at the first passage [121]. For human cells infected with OAdV7 at m.o.i. 20 pfu/cell the situation was polarized, depending on the cell type. On the one hand, in MRC5 cells, all early promoters in the genome that were examined were active, as monitored by RT PCR amplification of selected transcripts. On the other hand, in HepG2 liver carcinoma cells, none of the early promoters had detectable activity. In most other cell types, e.g., MCF7 and T47D2 breast cancer and PC3 prostate cells, some promoters, typically E2, were active. Interestingly, in all human cell types tested, and even when the early promoters were active, transcripts from the OAdV7 major late promoter (MLP) could not be detected [72, 76]. This may be related to key events that occur in the transition from early to late protein synthesis. For HAdV2, accumulation of early gene products is not sufficient for MLP activity. DNA replication is also required for late gene expression. High-level transcription from the MLP is further dependent on a c/s-acting change in the viral chromatin [145]. In addition, HAdV2 MLP activity is stimulated by ^mws-activating factors DBP and IVa2 [146-148]. At a gross level there is little or no DNA replication in OAdV7 infected human, compared with permissive ovine cells. However, the OAdV7 E2 promoter was active in several human cell types and large amounts of DBP transcript (and presumably, transcripts for DNA polymerase and Terminal protein) were produced [7G\, Cellular factors also cooperate with viral proteins during genome replication (reviewed in [2]). The apparent absence of DNA replication may be due to the incompatibility of one or more human cell factors with binding sites on the OAdV7 ITR sequences or with other viral proteins involved in the process. There are significant differences 16. Xenogenic Adenoviral Vectors 4 6 5 in putative binding sites for transcription factors between the ITRs of human and xenogenic viruses [149]. The inactivity of the OAdV7 MLP could further be due to a missing trans-actWsiting factor, such as IVa2, whose expression in human cells has not been investigated. Such abortive replication makes it unlikely that conditionally replication-competent vectors [150, 151] based on xenogenic vectors will be developed in the near future. 3. Biodistribution and Persistence in Vivo Few studies on the biodistribution and persistence of xenogenic AdVs in vivo have been reported, but some have been carried out with MAdVl and OAdV7. In the homologous situation, mice were injected intraperitoneally (ip) or intranasally with 10^ pfu of MAdVl and the localization of virus was monitored histologically during acute infection [152]. Endothelial cells of the brain and spinal cord showed extensive evidence of infection. Endothelial cells in lungs, kidneys, and other organs gave a positive signal, indicating a widespread involvement of the systemic circulation. Some lymphoid tissues were also positive. In experiments that examined persistence of OAdV7 it was found that 5 x 1 0 ^ pfu of a recombinant OAdV7 vector injected intravenously (iv) into SCID mice produced hAAT expression that persisted for at least 60 days. However, the same vector dose in BALB/c mice was cleared by 20 days. Thus, the vector did not persist in the normal host and a substantial dose of virus (2 X 10^^ particles) did not cause significant toxicity in normal or immunocompromised animals. The distinct nature of the OAdV7 receptor was reflected in the biodis tribution of OAdV7 following iv or ip administration of the vector to mice. OAdV7 was evenly distributed between liver, heart, spleen, and kidney [44], whereas HAdV5 vectors given iv concentrated predominantly in the liver [153]. OAdV7 given via the intraportal vein led to a greater accumulation of vector in the liver, but the vector was still found in all tissues examined [50]. In addition, when virus was injected directly into mouse skeletal muscle, cells were trans duced and high levels of hAAT reporter protein were secreted in vivo [129]. By judicious adjustment of the first dose of vector it was shown that a second dose that resulted in substantial reporter gene expression could be given, raising the prospect that the vector may be suitable for prime/boost vaccination strategies. The vector was not detected in liver and spleen, indicating that it did not spread via the circulation. Expression, however, was transient and the vector DNA had essentially disappeared by day 14. Clearance occurred in the absence of detectable OAdV7 gene expression as assayed by RT PCR. As proposed for HAdV5 vectors [154] clearance may occur via presentation of antigen using an MHC class I independent mechanism. Experiments utilizing HAdV5 and OAdV7 recombinants demonstrated a perceived advantage of xenogenic AdV, showing that OAdV7 could deliver a reporter gene in vivo in the face of preexisting antibodies against human 4 6 6 Gerald W. Both HAdV5 [44]. This result was encouraging from a clinical viewpoint and should be mimicked by other xenogenic AdV. It may be possible eventually to use different vectors in tandem to deliver multiple doses of the same gene [155]. C. Gene Therapy Studies To date no gene therapy applications have been reported for xenogenic Adv. However, work is in progress in this laboratory to assess OAdV7 as a gene delivery vector for prostate cancer. The strategy is based on gene- directed enzyme prodrug therapy (GDEPT). This is a two-component cell killing system: a gene that encodes an enzyme not present in mammalian cells and a nontoxic prodrug that is converted to a toxic product by cells that produce the enzyme. Although there are several GDEPT systems [156], in this case purine nucleoside phosphorylase (PNP), an E. colt enzyme, and the prodrug fludarabine are being used [157]. OAdV7 vectors that express the PNP gene under the control of the constitutive RSV, or a tissue-specific prostate promoter, were constructed and tested for cell killing in vitro and in mouse models of prostate cancer. Viruses were injected directly into human PC3 or LN3 tumors grown subcutaneously in nude mice or into mouse RMl tumors grown subcutaneously (sc) in immunocompetent animals. Prodrug was given systemically [158 and Voeks et al (in preparation)]. Evidence of tumor shrinkage and prolongation of mouse survival indicate that this vector and GDEPT system has potential for prostate cancer therapy. This work has also highlighted other important issues that must be addressed for OAdV7, and for xenogenic vectors in general, if they are to be developed for clinical application. These especially include biosafety and vector growth, purification, and scaleup. V. Biosafety Most work with xenogenic vectors is still firmly based in the laboratory and while this is appropriate to demonstrate the utility of a vector the amount of work required for eventual exploitation of a vector in the clinic should not be underestimated. A. Complementation and Recombination Although the xenogenic AdV undergo abortive replication in human cells, one hypothetical situation concerning the clinical application of these vectors is their potential for interaction with opportunistic, replication-competent human adenoviruses in a patient. This may involve complementation of a replication- deficient virus or recombination between genomes to create a hybrid with 16. Xenogenic Adenoviral Vectors 4 6 7 undesirable properties. A priori^ such events seem more likely to occur between viruses that are closely related, particularly if they share a common receptor to facilitate coinfection. Evidence v^as sought for
interaction between HAdV5 and CAdV2. However, coinfection of HeLa or A549 cells with CAdV2 (m.o.i. 10) and HAdV5 (m.o.i. 2) had no effect on the production of CAdV2 over five passages, compared with CAdV2 infection alone. DNA extracted from the cells was also digested and analyzed by Southern hybridization using a whole genome CAdV2 probe to track the DNA and look for the appearance of hybrid genomes. In coinfected HeLa and A549 cells CAdV2 DNA disappeared after one to three passages. HAdV5 DNA became visible by passage four and its restriction enzyme profile was identical to HAdV5 alone. No CAdV2 sequences were detected in these samples by hybridization [121]. Similar experiments have been done to determine whether any pro ductive interaction occurred between OAdV7 and HAdV5, a typical human adenovirus. No complementation of OAdV7 replication was detected in the presence of wild-type HAdV5 in MCF 7 cells, although both viruses infect these cells [76] and HAdV5 replicated with high efficiency. Similarly, when DNA from several passages of cells that were coinfected by OAdV7 and HAdV5 was analyzed by Southern blot using whole genome OAdV7 or HAdV5 probes, no hybrid genomes were detected [158a]. Considering the differences in genome structure between the two viruses (Fig. 2), the apparent lack of viable hybrid LHE LHIEIB III pVIII Fiber ixR xp̂ ^ > ^ ^ - • - • 46bp OAdV7 a::::::::::::>-----:-:-:-:-:-:':':-:-;-:v 33.6% ir 4-4- G/C P32 E4? RHE EIAEIBIX III V pVIIIE3 Fiber j^ j^ - • • • > > - • > — • > 103bp HAdV5 Ê 55.2% " E4 G/C Figure 2 Difference map between HAdV2/5 and OAdV7. The stippled rectangles indicate the genomes with distinct G / C content, striped boxes at each end show the ITRs of different length, and sequence and ORFs with bold type are unique. The packaging signal is shown in (^) . 4 6 8 Gerald W. Both virus formation and the absence of complementation was not surprising. First, the G/C content of the two genomes is vastly different, indicating low nucleotide sequence homology. Next, the ITRs of each genome differ in length and sequence, suggesting that neither would be compatible with the DNA replication machinery of the other. Third, each virus has a distinct complement of capsid proteins, including unique proteins and distinct fibers as well as non structural genes (Fig. 1). In addition, the packaging signals for each genome are likely to be incompatible. Thus, vectors such as OAdV7 may offer a greater margin of safety over those that are more closely related to HAdVs, such as SAdVs, with respect to potential for unwanted interactions. It is significant, therefore, that no human atadenoviruses have yet been described. B. Oncogenes in Viral and Cellular DNA As discussed above, replication-deficient El-deleted vectors are rescued and propagated in continuous cell lines that were derived from primary cell lines by transformation with adenovirus ElA/B genes [42, 43,114,115]. While this is an advantage for cell growth and virus production it is a disadvantage for downstream processing and purification. Regulatory agencies impose strict limits on the permissible levels of contaminating DNA (10 ng/dose) in purified vector preparations [159]. A rigorous purification process is therefore required to remove potentially oncogenic DNA. Thus, an advantage of OAdV7 vectors is that they grow in a primary fetal ovine lung cell line. The trade-off is that the cells grow more slowly and have a life span of 50 to 70 doublings [122]. Oncogenes must also be removed from the vector genome. This may be more straightforward for the mastadenoviruses, where precedents exist from HAdV2/5 studies, but within this genus some viruses are more oncogenic than others [2, 65] and some ORFs exhibit unexpected transforming proper ties [160-162]. Progress toward oncogene identification in FAdVl has also been made [73, 111], but others may exist. Ultimately, the regulatory authori ties will require tests to be conducted on the residual oncogenicity of xenogenic vectors prior to clinical application. The apparent lack of transforming ability of OAdV7 in systems that have been used as a benchmark for such assays was therefore encouraging [98]. C. Virus/Cell Interactions Adenoviruses undergo a lytic infection cycle in permissive cells. The mechanism behind cell lysis is not well defined in all cases but for FlAdV5 it is due to the production of a "death protein" late in the infectious cycle [163]. Other mechanisms that may be involved in selective killing of tumor cells are being investigated [164]. These observations highlight the potential for interactions between a virus and a cell that may be undesirable in the context of extended gene expression or from a biosafety perspective. 16. Xenogenic Adenoviral Vectors 4 6 9 Despite the inactivity of its MLP, in some cell types, typified by MRC5 lung fibroblasts, OAdV7 produced an apparent cytopathic effect (CPE) that was limited by the m.o.i. CPE was not due to viral replication because virus passaged twice on MRC5 cells failed to produce CPE in permissive CSL503 cells [76]. Thus, the effect is likely to involve an early gene product. This is currently under investigation. In this regard it is intriguing that the induction of rapid cell death following infection by certain HAdVs appears to be due to an interaction between p53 and the ElB 55-kDa product [164]. The response was abrogated by the absence of either protein due to mutation or lack of expression. Given the many genes of unknown function that exist in the expanding range of xenogenic AdVs the potential to discover other unwanted interactions exists. It may prove necessary to engineer vectors to remove deleterious genes and to grow them in complementing cell lines, but that raises complementation risks. D. Replication Competent Viruses A significant problem with the production of HAdV5 vectors has been the emergence of replication-competent viruses from cells that were designed to prevent their formation. Sequence overlap between the viral vector and integrated genes and subsequent recombination between them has generally been the cause [165]. Thus, PERC6 cells and matching vectors in which sequence overlaps were eliminated were specifically designed to overcome the problem [166]. An advantage offered by the xenogenic vectors is that all of them are replication-deficient in all human cell lines that have been tested. Additional work with particular vectors and cell types to understand the molecular basis for abortive replication would be very helpful in assessing the safety of new vectors. VI . Vector Production and Purification For vector production at the laboratory level the availability of a cell line or egg system [23, 48] for virus rescue and propagation is sufficient. Virus can be purified using methods based on cell lysis and CsCl centrifugation similar to those described for HAdV5 [10, 15, 43, 49, 137]. However, increasing success with a vector brings increasingly stringent requirements as work proceeds toward production for veterinary applications or a clinical trial. Strategic decisions taken early to facilitate subsequent steps in vector development and exploitation could save substantial time and effort later on. A key requirement for vector production is the availability of a cell line that, having been expanded and laid down as master and working cell banks (MCBA)C^CB), is tested and shown to be free of adventitious agents. Attention to detail in the creation and 4 7 0 Gerald W. Both documentation of such a cell line would pay dividends in the long term. A master virus seed stock also needs to be established. This dictates that the viral genome, including the transgene, must be stable upon serial passage such that biological activity and potency are maintained. This stock must also be free of other agents. The issue of vector yield from the WCB should also be considered. For veterinary applications w^here the vector may be replication competent in the host, low^ yields may be less important. Hov^ever, if gene therapy is being considered as an application a purified virus yield of >10'^ particles per cell is probably required for cost effective production of a vector. For clinical applications in particular, a robust scheme for vector purifi cation is required. While this might involve CsCl gradient centrifugation to produce quantities of vector for preclinical and perhaps phase I studies, such methodology is unlikely to be appropriate for producing larger amounts of vector. Methods involving chromatography may be more advantageous [167]. It is recognized that the above provides a very brief summary of issues that might be substantial for particular vector systems. How^ever, the intention is to alert the reader contemplating the use of a new^ vector system to the many chal lenges that lie ahead in the process of chaperoning it through production and regulatory processes. The correct strategic decisions taken early can facilitate subsequent steps in vector development and exploitation. Acknov^ledgments The author extends his appreciation to the colleagues who provided preprints and infor mation that helped to ensure the timeliness and relevance of this chapter and to Pamela Russell and Keith Smith for comments on the manuscript. References 1. Horwitz, M. S. (1996). Adenoviruses. Third ed. In "Fields Virology" (B. N. Fields, D. M. Knipe, and P. M. Howley, Eds.), Vol. 2, pp. 2149-2171. Lippincott-Raven, Philadelphia. 2. Shenk, T. (1996). Adenoviridae: The viruses and their replication. In "Fields Virology" (B. N. Fields, D. M. Knipe, and P. M. Howley, Eds.), 3rd ed.. Vol. 2, pp. 2111-2148. Lippincott-Raven, Philadelphia. 3. Benko, M., Harrach, B., and Russell, W. C. (1999). Family adenoviridae. In "Virus Taxonomy. Seventh Report of the International Committee on Taxonomy of Viruses" (M. H. V. Van Regenmortel, C. M. Fauquet, D. H. L. Bishop, E. B. Carstens, M. K. Estes, S. M. Lemon, J. Maniloff, M. A. Mayo, D. J. McGeoch, C. R. Pringle, and R. B. Wickner, Eds.), pp. 227-238. Academic Press, New York, San Diego. 4. Harrach, B., Meehan, B. M., Benko, M., Adair, B. M., and Todd, D. (1997). Close phylogenetic relationship between egg drop syndrome virus, bovine adenovirus serotype 7, and ovine adenovirus strain 287. Virology 229, 302-308. 16. Xenogenic Adenoviral Vectors 4 7 1 5. Benko, M., and Harrach, B. (1998). A proposal for a new (third) genus within the family Adenoviridae. Arch. Virol. 143, 829-837. 6. Dan, A., Ruzsics, Z., Russell, W. C , Benko, M., and Harrach, B. (1998). Analysis of the hexon gene sequence of bovine adenovirus type 4 provides further support for a new adenovirus genus (Atadenovirus)./. Gen. Virol. 79, 1453-1460. 7. Harrach, B., and Benko, M. (1999). Phylogenetic analysis of adenovirus sequences — Proof of the necessity of establishing a third genus in the Adenoviridae family. In "Adenovirus Methods And Protocols" (W. S. M. Wold, Ed.), Vol. 21 , pp. 309-339. Humana Press, Totowa, NJ. 8. Both, G. (2001). Atadenoviruses. In "The Springer Index of Viruses" (C. Tidona, and G. Darai, Eds.), pp. 11-17. Springer-Verlag, Berlin. 9. Davison, A. J., Wright, K. M., and Harrach, B. (2000). DNA sequence of frog adenovirus. / . Gen. Virol. 81,2431-2439. 10. Graham, F. L., and Prevec, L. (1992). Adenovirus based expression vectors and recombinant vaccines. In "Vaccines: New Approaches to Immunological Problems" (R. W. ElHs, Ed.), pp. 363-390. Butterworth-Heinemann, Stoneham, MA. 11. Darbyshire, J. H., Jennings, A. R., Dawson, P. S., Lamont, P. H., and Omar, A. R. (1966). The pathogenesis and pathology of infection in calves with a strain of bovine adenovirus type 3. Res. Vet. Sci. 7, 81-93. 12. Derbyshire, J. B., Clarke, M. C., and Collins, A. P. (1975). Serological and pathogenicity studies with some unclassified porcine adenoviruses./. Comp. Pathol. 85, 437-443. 13. Koptopoulos, G., and Cornwell, H. (1981). Canine adenoviruses: A review. Vet. Bull. 51, 135-142. 14. Cowen, B., Calnek, B. W., Menendez, N. A., and Ball, R. F. (1978). Avian adenoviruses: Effect on egg production, shell quality, and feed consumption. Avian Dis. 22, 459-470. 15. Boyle, D. B., Pye, A. D., Kocherhans, R., Adair, B. M., Vrati, S., and Both, G. W. (1994). Characterisation of AustraUan ovine adenovirus isolates. Vet. Microbiol. 41, 281-291. 16. Kidd, A. H., Chroboczek, J., Cusack, S., and Ruigrok, R. W. H. (1993). Adenovirus Type-40 Virions Contain 2 Distinct Fibers. Virology 192, 73-84. 17. Hess, M., Cuzange, A., Ruigrok, R. W., Chroboczek, J., and Jacrot, B. (1995). The avian adenovirus penton: Two fibres and one base. / . Mol. Biol. 252, 379-385. 18. Tomko, R. P., Xu, R. L., and Philipson, L. (1997). HCAR and MCAR: The human and mouse cellular receptors for subgroup
C adenoviruses and group B coxsackieviruses. Proc. Natl. Acad. Sci. USA 94, 3352-3356. 19. Bergelson, J. M., Cunningham, J. A., Droguett, G., Kurt-Jones, E. A., Krithivas, A., Hong, J. S., Horwitz, M. S., Crowell, R. L., and Finberg, R. W. (1997). Isolation of a common receptor for coxsackie B viruses and adenoviruses 2 and 5. Science 275, 1320-1323. 20. Farina, S. F., Fisher, K. J., and Wilson, J. M. (2000). Chimpanzee adenovirus vector useful in gene therapy and genetic engineering in general. US patent 6,083,716. 21. Wickham, T. J., Mathias, P., Cheresh, D. A., and Nemerow, G. R. (1993). Integrin-alpha- v-beta-3 and integrin-alpha-v-beta-5 promote adenovirus internalization but not virus attachment. Cell 73, 309-319. 22. Vrati, S., Brookes, D. E., Strike, P., Khatri, A., Boyle, D. B., and Both, G. W. (1996). Unique genome arrangement of an ovine adenovirus: Identification of new proteins and proteinase cleavage sites. Virology 220, 186-199. 23. Chiocca, S., Kurzbauer, R., Schaffner, G., Baker, A., Mautner, V., and Gotten, M. (1996). The complete DNA sequence and genomic organization of the avian adenovirus CELO. / . Virol. 70, 2939-2949. 24. Reddy, P. S., Idamakanti, N., Zakhartchouk, A. N., Baxi, M. K., Lee, J. B., Pyne, C , Babiuk, L. A., and Tikoo, S. K. (1998). Nucleotide sequence, genome organization, and transcription map of bovine adenovirus type 3. / . Virol. 72, 1394-1402. 4 7 2 Gerald W. Both 25. Reddy, P. S., Idamakanti, N., Song, J. Y., Lee, J. B., Hyun, B. H., Park, J. H., Cha, S. H., Tikoo, S. K., and Babiuk, L. A. (1998). Sequence and transcription map analysis of early region-1 of porcine adenovirus type-3. Virus Res. 58, 97-106. 26. Morrison, M. D., Onions, D. E., and Nicolson, L. (1997). Complete DNA sequence of canine adenovirus type 1 . / . Gen. Virol. 78, 873-878. 27. Meissner, J. D., Hirsch, G. N., La Rue, E. A., Fulcher, R. A., and Spindler, K. R. (1997). Completion of the DNA sequence of mouse adenovirus type 1: Sequence of E2B, LI, and L2 (18-51 map units). Virus Res. 51, 53-64. 28. Stouten, P. F. W., Sander, C , Ruigrok, R. W. H., and Cusack, S. (1992). New Triple-Helical Model for the Shaft of the Adenovirus Fibre. / . Mol. Biol. 226, 1073-1084. 29. Xu, Z. Z., and Both, G. W. (1998). Altered tropism of an ovine adenovirus carrying the fiber protein cell binding domain of human adenovirus type 5. Virology 248, 156-163. 30. Chillon, M., and Kremer, E. (2001). Trafficking and propagation of canine adenovirus vectors lacking a known integrin-interacting motif. Hum. Gene Ther. 12, 1815-1823. 31. Soudais, C , Boutin, S., Hong, S., Chillon, M., Danos, O., Bergelson, J., Boulanger, P., and Kremer, E. (2000). Canine adenovirus type 2 attachment and internalization: CAR, alternative receptors and an RGD-independent pathway. / . Virol. 74, 10,639-10,649. 32. D'Ambrosio, E., Del Grosso, N., Chicca, A., and MiduUa, M. (1982). Neutralizing antibodies against 33 human adenoviruses in normal children in Rome. / . Hyg. (London) 89, 155-161. 33. Schulick, A. H., Vassalli, G., Dunn, P. F., Dong, G., Rade, J. J., Zamarron, C , and Dichek, D. A. (1997). Established immunity precludes adenovirus-mediated gene transfer in rat carotid arteries. Potential for immunosuppression and vector engineering to overcome barriers of immunity./. Clin. Invest. 99, 209-219. 34. Cooper, N. (1998). Complement dependent virus neutrafisation. In "The Complement System" (K. Rother, G. Till, and G. Hansch, Eds.), pp. 302-309. Springer Verlag, Berlin-Heidelberg-New York. 35. Flomenberg, P., Piaskowski, V., Truitt, R. L., and Casper, J. T. (1995). Characterization of human proliferative T cell responses to adenovirus. / . Infect. Dis. 171, 1090-1096. [see comments! 36. Mastrangeli, A., Harvey, B. G., Yao, J., Wolff, G., Kovesdi, I., Crystal, R. G., and Falck-Pedersen, E. (1996). "Sero-switch" adenovirus-mediated in vivo gene transfer: Cir cumvention of anti-adenovirus humoral immune defenses against repeat adenovirus vector administration by changing the adenovirus serotype. Hum. Gene Ther. 7, 79-87. 37. Mack, C. A., Song, W. R., Carpenter, H., Wickham, T. J., Kovesdi, L, Harvey, B. G., Magovern, C. J., Isom, O. W., Rosengart, T., Falck-Pedersen, E., Hackett, N. R., Crystal, R. G., and Mastrangefi, A. (1997). Circumvention of anti-adenovirus neutralizing immunity by administration of an adenoviral vector of an alternate serotype. Hum. Gene Ther. 8, 99-109. 38. Russi, T. J., Hirschowitz, E. A., and Crystal, R. G. (1997). Delayed-type hypersensitivity response to high doses of adenoviral vectors. Hum. Gene Ther. 8, 323-330. 39. Bramson, J. L., Hitt, M., Gauldie, J., and Graham, F. L. (1997). Pre-existing immunity to adenovirus does not prevent tumor regression following intratumoral administration of a vector expressing IL-12 but inhibits virus dissemination. Gene Ther. 4, 1069-1076. 40. Harvey, B. G., Hackett, N. R., El Sawy, T., Rosengart, T. K., Hirschowitz, E. A., Lieber- man, M. D., Lesser, M. L., and Crystal, R. G. (1999). Variability of human systemic humoral immune responses to adenovirus gene transfer vectors administered to different organs./. Virol. 73, 6729-6742. 41. Chen, Y., Yu, D. C , Charlton, D., and Henderson, D. R. (2000). Pre-existent adenovirus antibody inhibits systemic toxicity and antitumor activity of CN706 in the nude mouse LNCaP xenograft model: ImpUcations and proposals for human therapy. Hum. Gene Ther. 11,1553-1567. 16. Xenogenic Adenoviral Vectors 4 7 3 42. Klonjkowski, B., Gilardi-Hebenstreit, P., Hadchouel, J., Randrianarison, V., Boutin, S., Yeh, P., Perricaudet, M., and Kremer, E.J. (1997). A recombinant El-deleted canine adenoviral vector capable of transduction and expression of a transgene in human-derived cells and in vivo. Hum. Gene Ther. 8, 2103-2115. [see comments] 43. Kremer, E. J., Boutin, S., Chillon, M., and Danos, O. (2000). Canine adenovirus vectors: An alternative for adenovirus-mediated gene transfer./. Virol. 74, 505-512. 44. Hofmann, C , Loser, P., Cichon, G., Arnold, W., Both, G. W., and Strauss, M. (1999). Ovine adenovirus vectors overcome preexisting humoral immunity against human adenoviruses in vivo. / . Virol. 73, 6930-6936. 45. Wigand, R., Mauss, M., and Adrian, T. (1989). Chimpanzee adenoviruses are related to four subgenera of human adenoviruses. Intervirology 30, 1-9. 46. Reddy, P. S., Idamakanti, N., Song, J. Y., Lee, J. B., Hyun, B. H., Park, J. H., Cha, S. H., Bae, Y. T., Tikoo, S. K., and Babiuk, L. A. (1998). Nucleotide sequence and transcription map of porcine adenovirus type 3. Virology 251, 414-426. 47. Ojkic, D., and Nagy, E. (2000). The complete nucleotide sequence of fowl adenovirus type 8. / . Gen. Virol 81, 1833-1837. 48. Hess, M., Blocker, H., and Brandt, P. (1997). The complete nucleotide sequence of the egg drop syndrome virus: An intermediate between mastadenoviruses and aviadenoviruses. y/ro/og}/238, 145-156. 49. Pitcovski, J., Mualem, M., Rei-Koren, Z., Krispel, S., Shmueli, E., Peretz, Y., Gutter, B., Gallih, G. E., Michael, A., and Goldberg, D. (1998). The complete DNA sequence and genome organization of the avian adenovirus, hemorrhagic enteritis virus. Virology 249, 307-315. 50. Loser, P., Cichon, G., Jennings, G., Both, G., and Hofmann, C. (1999). Ovine adenovirus vectors promote efficient gene delivery in vivo. Gene Ther. Mol. Biol. 4, 33-43 . 51. Michou, A. I., Lehrmann, H., Saltik, M., and Gotten, M. (1999). Mutational analysis of the avian adenovirus CELO, which provides a basis for gene delivery vectors. / . Virol. 73, 1399-1410. 52. Ma, Y., and Mathews, M. B. (1993). Comparative analysis of the structure and function of adenovirus virus-associated RNAs./. Virol. 67, 6605-6617. 53. Kidd, A. H., Garwicz, D., and Oberg, M. (1995). Human and simian adenoviruses: Phylogenetic inferences from analysis of VA RNA genes. Virology 207, 32-45 . 54. Venktesh, A., Watt, F., Xu, Z. Z., and Both, G. W. (1998). Ovine adenovirus (OAV287) lacks a virus-associated RNA gene./. Gen. Virol. 79, 509-516. 55. Larsson, S., Bellett, A., and Akusjarvi, G. (1986). VA RNAs from avian and human adenoviruses: Dramatic differences in length, sequence, and gene location. / . Virol. 58, 600-609. 56. Wold, W. S. M., and Gooding, L. R. (1991). Region E3 of adenovirus: A cassette of genes involved in host immunosurveillance and virus-cell interactions. Virology 184, 1-8. 57. Raviprakash, K. S., Grunhaus, A., el Kholy, M. A., and Horwitz, M. S. (1989). The mouse adenovirus type 1 contains an unusual E3 region./. Virol. 63, 5455-5458. 58. Beard, C. W., Ball, A. O., Wooley,E. H., and Spindler, K. R. (1990). Transcription mapping of mouse adenovirus type 1 early region 3. Virology 175, 81-90. 59. Esford, L. E., and Haj-Ahmad, Y. (1994). Sequence analysis of the putative E3 region of bovine adenovirus type 2. Intervirology 37, 277-286. 60. Evans, P. S., Benko, M., Harrach, B., and Letchworth, G.J. (1998). Sequence, transcriptional analysis, and deletion of the bovine adenovirus type 1 E3 region. Virology 244, 173-185. 61. Reddy, P. S., Idamakanti, N., Derbyshire, J. B., and Nagy, E. (1996). Porcine adenoviruses types 1, 2 and 3 have short and simple early E-3 regions. Virus Res. 43, 99-109. 62. Idamakanti, N., Reddy, P. S., Babiuk, L. A., and Tikoo, S. K. (1999). Transcription mapping and characterization of 284R and 12IR proteins produced from early region 3 of bovine adenovirus type 3. Virology 256, 351-359. 4 7 4 Gerald W. Both 63. Tuboly, T., and Nagy, E. (2000). Sequence analysis and deletion of porcine adenovirus serotype 5 E3 region. Virus Res. 68, 109-117. 64. Zakhartchouk, A., Godson, D., Babiuk, L., and Tikoo, S. (2001). 121R protein from the E3 region of bovine adenovirus-3 inhibits cytolysis of mouse cells by human TNF. Intervirol. 44 ,29 -35 . 65. Horwitz, M. S. (1990). Adenoviridae and their replication. In "Virology" (B. N. Fields, and D. M. Knipe, Eds.), 2nd ed., pp. 1679-1721. Raven Press, New York. 66. Morin, J. E., Lubeck, M. D., Barton, J. E., Conley, A. J., Davis, A. R., and Hung, P. P. (1987). Recombinant adenovirus induces antibody responses to hepatitis B virus surface antigen in hamsters. Proc. Natl. Acad. Sci. USA 84, 4626-4630. 67. Ilan, Y., Droguett, G., Chowdhury, N. R., Li, Y. A., Sengupta, K., Thummala, N. R., Davidson, A., Chov^dhury, J. R., and Horwitz, M. S. (1997). Insertion of the adenovi ral E3 region into a recombinant viral vector prevents antiviral humoral and cellular immune responses and permits long-term gene expression. Proc. Natl. Acad. Sci. USA 94, 2587-2592. 68. Ginsberg, H. S., Lundholm-Beauchamp, U., Horswood, R. L., Pernis, B., Wold, W. S. M., Chanock, R. M., and Prince, G. A. (1989). Role of early region 3 (E3) in pathogenesis of adenovirus disease. Proc. Natl. Acad. Sci. USA 86, 3823-3827. 69. Vrati, S., Boyle, D., Kocherhans, R., and Both, G. W. (1995). Sequence of ovine adenovirus homologs for 100 K hexon assembly, 33 K, pVIII, and fiber genes: Early region E3 is not in the expected location. Virology 209, 400-408. 70. Kring, S. C., Ball, A. O., and Spindler, K. R. (1992). Transcription mapping of mouse adenovirus Type-1 early region-4. Virology 190, 248-255. 71. Baxi, M. K., Babiuk, L. A., MehtaU, M., and Tikoo, S. K. (1999). Transcription map and expression of bovine herpesvirus-1 glycoprotein D in early region 4 of bovine adenovirus-3. V/ro/og); 261, 143-152. 72. Khatri, A., and Both, G. W. (1998). Identification of transcripts and promoter regions of ovine adenovirus OAV287. Virology 245, 128-141. 73. Lehrmann, H., and Gotten, M. (1999). Characterization of CELO virus proteins that modulate the pRb/E2F pathway./. Virol. 73, 6517-6525. 74. Cao, J. X., Krell, P. J., and Nagy, E. (1998). Sequence and transcriptional analysis of terminal regions of the fowl adenovirus type 8 genome./. Gen. Virol. 79, 2507-2516. 75. Xu, Z. Z., Hyatt, A., Boyle, D. B., and Both, G. W. (1997). Construction of ovine ade novirus recombinants by gene insertion or deletion of related terminal region sequences. V/ro/og}; 230, 62 -71 . 76. Khatri, A., Xu, Z. Z., and Both, G. W. (1997). Gene expression by atypical recombinant ovine adenovirus vectors during abortive infection of human and animal cells in vitro. Virology 239, 226-237. 77. Kleiboeker, S. B. (1995). Identification and sequence analysis of the El genomic region of a porcine adenovirus. Virus Res. 36, 259-268. 78. Elgadi, M., Rghei, N., and Haj-Ahmad, Y. (1993). Sequence and sequence analysis of El and pIX regions of the BAV3 genome. Intervirology 36, 113-120. 79. Reddy, P. S., Chen, Y., Idamakanti, N., Pyne, C , Babiuk, L. A., and Tikoo, S. K. (1999). Characterization of early region 1 and pIX of bovine adenovirus-3. Virology 253,299-308. 80. Aggarwal, N., and Mittal, S. K. (2000). Sequence analysis of porcine adenovirus type 3 El region, pIX and pIVa2 genes,
and two novel open reading frames (Vol 43, pg 6, 2000). Intervirology 43, 128. 81. Schmid, S. I., and Hearing, P. (1998). Cellular components interact with adenovirus type 5 minimal DNA packaging domains./. Virol. 72, 6339-6347. 82. Soudais, C , Boutin, S., and Kremer, E. (2001). Characterisation of cis-acting sequences involved in the canine adenovirus packaging domain. Mol. Ther. 3, 631-640. 16. Xenogenic Adenoviral Vectors 4 7 5 83. Nevels, M., Rubenwolf, S., Spruss, T., Wolf, H., and Dobner, T. (2000). Two distinct activities contribute to the oncogenic potential of the adenovirus type 5 E4orf6 protein. / . V/Vo/. 74,5168-5181. 84. Payet, V., Arnauld, C , Picault, J. P., Jestin, A., and Langlois, P. (1998). Transcriptional organization of the avian adenovirus CELO./. Virol. 72, 9278-9285. 85. Ball, A. O., Beard, C. W., Redick, S. D., and Spindler, K. R. (1989). Genome organization of mouse adenovirus type 1 early region 1: A novel transcription map. Virology 170, 523-536. 86. Zheng, B. J., Graham, F. L., and Prevec, L. (1999). Transcription units of Ela, Elb and pIX regions of bovine adenovirus type 3. / . Gen. Virol. 80, 1735-1742. 87. McAllister, R., Nicolson, M., Levels AM, J., Macpherson, I., and Huebner, R. (1969). Transformation of rat embryo cells by adenovirus type 1. / . Gen. Virol. 4, 29-36. 88. Trentin, J. J., Yabe, Y., and Taylor, G. (1962). The quest for human cancer viruses. Science 137, 835. 89. Kinjo, T., Nishi, T., and Yanagawa, R. (1969). In vitro transformation of hamster cells by infectious canine hepatitis virus. Jpn. J. Vet. Res. 17, 128-135. 90. Flint, S. (1980). Cell transformation induced by adenoviruses. In "DNA Tumor Viruses" (J. Tooze, Ed.), 2nd ed.. Part 2, pp. 574-576. Cold Spring Harbor Press, Cold Spring Harbor, NY. 91. Darbyshire, J. (1966). Oncogenicity of bovine adenovirus type 3 in hamsters. Nature 211, 102. 92. Anderson, J., Yates, V. J., Jasty, V., and Mancini, L. O. (1969). In vitro transformation by an avian adenovirus (CELO). I. Hamster-embryo fibroblastic cultures./. Natl. Cancer Inst. 42, 1-7. 93. Anderson, J., Yates, V. J., Jasty, V., and Mancini, L. O. (1969). In vitro transformation by an avian adenovirus (CELO). 11. Hamster kidney cell cultures. / . Natl. Cancer Inst. 43, 65-70. 94. Sarma, P. S., Huebner, R. J., and Lane, W. T. (1965). Induction of tumors in hamsters with an avian adenovirus (CELO), Science 149, 1108. 95. Mancini, L. O., and Yates, V. J. (1973). Attempts to produce tumors in mice with CELO virus. Arch. Gesamte Virusforsch. 40, 174-175. 96. Rondhuis, P. R. (1973). Induction of tumors in hamsters with a bovine adenovirus strain (serotype 8). Arch. Gesamte Virusforsch. 41, 147-149. 97. Mohanty, S. B. (1971). Comparative study of bovine adenoviruses. Am. J. Vet. Res. 32, 1899-1905. 98. Xu, Z. Z., Nevels, M., MacAvoy, E. S., Lockett, L. J., Curiel, D., Dobner, T., and Both, G. W. (2000). An ovine adenovirus vector lacks transforming ability in cells that are transformed by AD5 ElA/B sequences. Virology 270, 162-172. 99. Whyte, P., Buchkovich, K. J., Horowitz, J. M., Friend, S. H., Raybuck, M., Weinberg, R. A., and Harlow, E. (1988). Association between an oncogene and an anti-oncogene: The adenovirus El A proteins bind to the retinoblastoma gene product. Nature 334, 124-129. 100. Nevins, J. R. (1992). E2F: A link between the Rb tumor suppressor protein and viral oncoproteins. Science 258, 424-429. 101. White, E. (1998). Regulation of apoptosis by adenovirus ElA and ElB oncogenes, Sem. Virol. 8, 505-513, 102. Rao, L., Debbas, M,, Sabbatini, P., Hockenbery, D., Korsmeyer, S., and White, E. (1992). The adenovirus ElA proteins induce apoptosis, which is inhibited by the ElB 19-kDa and Bcl-2 proteins. Proc. Natl. Acad. Sci. USA 89, 7742-7746. 103. Moore, M., Horikoshi, N., and Shenk, T. (1996). Oncogenic potential of the adenovirus E4orf6 protein, Proc. Natl. Acad. Sci. USA 93, 11,295-11,301. 4 7 6 Gerald W. Both 104. Nevels, M., Rubenwolf, S., Spruss, T., Wolf, H., and Dobner, T. (1997). The adenovirus E4orf6 protein can promote ElA/ElB-induced focus formation by interfering with p53 tumor suppressor function. Proc. Natl. Acad. Set. USA 94, 1206-1211. 105. Nevels, M., Spruss, T., Wolf, H., and Dobner, T. (1999). The adenovirus E4orf6 protein contributes to malignant transformation by antagonizing ElA-induced accumulation of the tumor suppressor protein p53. Oncogene 18, 9-17. 106. Nevels, M., Tauber, B., Kremmer, E., Spruss, T., Wolf, H., and Dobner, T. (1999). Transforming potential of the adenovirus type 5 E4orf3 protein. / . Virol. 73, 1591-1600. 107. Dobbelstein, M., Roth, J., Kimberly, W. T., Levine, A. J., and Shenk, T. (1997). Nuclear export of the ElB 55-kDa and E4 34-kDa adenoviral oncoproteins mediated by a rev-like signal sequence. EMBO J. 16, 4276-4284. 108. Ball, A. O., Beard, C. W., Villegas, P., and Spindler, K. R. (1991). Early region 4 sequence and biological comparison of two isolates of mouse adenovirus type 1. Virology 180, 257-265. 109. Reddy, P. S., Idamakanti, N., Derbyshire, J. B., and Nagy, E. (1997). Characterization of the early region 4 of porcine adenovirus type 3. Virus Genes 15, 87-90. 110. Lee, J. B., Baxi, M. K., Idamakanti, N., Reddy, P. S., Zakhartchouk, A. N., Pyne, C , Babiuk, L. A., and Tikoo, S. K. (1998). Genetic organization and DNA sequence of early region 4 of bovine adenovirus type 3. Virus Genes 17, 99-100. 111. Chiocca, S., Baker, A., and Gotten, M. (1997). Identification of a novel antiapoptotic protein, GAM-1, encoded by the GELO adenovirus./. Virol. 71, 3168-3177. 112. Glotzer, J. B., Saltik, M., Ghiocca, S., Michou, A. I., Moseley, P., and Gotten, M. (2000). Activation of heat-shock response by an adenovirus is essential for virus replication. Nature 407,207-211 . 113. Kovesdi, I., Reichel, R., and Nevins, J. R. (1987). Role of an adenovirus E2 promoter binding factor in ElA-mediated coordinate gene control. Proc. Natl. Acad. Sci. USA 84, 2180-2184. 114. Reddy, P. S., Idamakanti, N., Ghen, Y., Whale, T., Babiuk, L. A., Mehtah, M., and Tikoo, S. K. (1999). Replication-defective bovine adenovirus type 3 as an expression vector. / . Virol. 73, 9137-9144. 115. Reddy, P. S., Idamakanti, N., Babiuk, L. A., Mehtali, M., and Tikoo, S. K. (1999). Porcine adenovirus-3 as a helper-dependent expression vector. / . Gen. Virol. 80, 2909-2916. 116. Nguyen, T. T., Nery, J. P., Joseph, S., Rocha, G. E., Garney, G. E., Spindler, K. R., and Vil- larreal, L. P. (1999). Mouse adenovirus (MAV-1) expression in primary human endotheHal cells and generation of a full-length infectious plasmid. Gene Ther. 6, 1191-1197. 117. Vrati, S., Macavoy, E. S., Xu, Z. Z., Smole, G., Boyle, D. B., and Both, G. W. (1996). Gonstruction and transfection of ovine adenovirus genomic clones to rescue modified viruses. Virology 110, 200 -3 . 118. Reddy, P. S., Idamakanti, N., Hyun, B. H., Tikoo, S. K., and Babiuk, L. A. (1999). Devel opment of porcine adenovirus-3 as an expression vector. / . Gen. Virol. 80, 563-570. 119. Johnson, M. A., Pooley, G., and Lowenthal, J. W. (2000). Delivery of avian cytokines by adenovirus vectors. Dev. Comp. Immunol. 24, 343-354. 120. Hammond, J. M., McGoy, R. J., Jansen, E. S., Morrissy, G. J., Hodgson, A. L., and John son, M. A. (2000). Vaccination with a single dose of a recombinant porcine adenovirus expressing the classical swine fever virus gp55 (E2) gene protects pigs against classical swine fever. Vaccine 18, 1040-1050. 121. Rasmussen, U. B., Benchaibi, M., Meyer, V., Schlesinger, Y., and Schughart, K. (1999). Novel human gene transfer vectors: Evaluation of wild-type and recombinant animal adenoviruses in human-derived cells. Hum. Gene Ther. 10, 2587-2599. 122. Pye, D. (1989). Gell lines for growth of sheep viruses. Aust. Vet. J. 66, 231-232. 123. Graham, F. L., Smiley, J., Russell, W. G., and Nairn, R. (1977). Gharacteristics of a human cell line transformed by DNA from human adenovirus type 5. / . Gen. Virol. 36, 59-72. 16. Xenogenic Adenoviral Vectors 4 7 7 124. Fallaux, F. J., Kranenburg, O., Cramer, S. J., Houweling, A., van Ormondt, H., Hoeben, R. C , and van der Eb, A. J. (1996). Characterization of 911: A nev^ helper cell line for the titration and propagation of early region 1-deleted adenoviral vectors. Hum. Gene Ther. 7, 215-222. 125. Zheng, B., Mittal, S. K., Graham, F. L., and Prevec, L. (1994). The El sequence of bovine adenovirus type 3 and complementation of human adenovirus type 5 ElA function in bovine cells. Virus Res. 31, 163-186. 126. Mittal, S. K., Prevec, L., Graham, F. L., and Babiuk, L. A. (1995). Development of a bovine adenovirus type 3-based expression vector./. Gen. Virol. 76, 93-102. 127. Bett, A. J., Prevec, L., and Graham, F. L. (1993). Packaging capacity and stability of human adenovirus type- 5 vectors./. Virol. 67, 5911-5921. 128. Zakhartchouk, A. N., Reddy, P. S., Baxi, M., Baca-Estrada, M. E., Mehtali, M., Babiuk, L. A., and Tikoo, S. K. (1998). Construction and characterization of E3-deleted bovine adenovirus type 3 expressing full-length and truncated form of bovine herpesvirus type 1 glycoprotein gD. Virology 250, 220-229. 129. Loser, P., Hillgenberg, M., Arnold, W., Both, G. W., and Hofmann, C. (2000). Ovine aden ovirus vectors mediate efficient gene transfer to skeletal muscle. Gene Ther. 7, 1491-1498. 130. Hanahan, D., and Gluzman, Y. (1984). Rescue of functional replication origins from embedded configurations in a plasmid carrying the adenovirus genome. Mol. Cell. Biol. 4, 302-309. 131. Leach, D. R., and Stahl, F. W. (1983). Viability of lambda phages carrying a perfect palindrome in the absence of recombination nucleases. Nature 305, 448-451. 132. Graham, F. L. (1984). Covalently closed circles of human adenovirus DNA are infectious. £ M B O / . 3,2917-2922. 133. Chartier, C , Degryse, E., Gantzer, M., Dieterle, A., Pavirani, A., and Mehtali, M. (1996). Efficient generation of recombinant adenovirus vectors by homologous recombination in Escherichia coH. / . Virol. 70, 4805-4810. 134. van Olphen, A. L., and Mittal, S. K. (1999). Generation of infectious genome of bovine adenovirus type 3 by homologous recombination in bacteria. / . Virol. Methods 77, 125-129. 135. Morsy, M. A., Gu, M. C , Motzel, S., Zhao, J., Su, Q., Allen, H., Franlin, L., Parks, R. J., Graham, F. L,, Kochanek, S., Bett, A. J., and Caskey, C. T. (1998). An adenoviral vector deleted for all viral coding sequences results in enhanced safety and extended expression of a leptin transgene. Proc. Natl. Acad. Sci. USA 95, 7866-7871. 136. Kochanek, S., Clemens, P. R., Mitani, K., Chen, H. H., Chan, S., and Caskey, C. T. (1996). A new adenoviral vector: Replacement of all viral coding sequences with 28 kb of DNA independently expressing both full-length dystrophin and beta-galactosidase. Proc. Natl. Acad. Sci. USA 93, 5731-5736. 137. Mittal, S. K., Middleton, D. M., Tikoo, S. K., Prevec, L., Graham, F. L., and Babiuk, L. A. (1996). Pathology and immunogenicity in the cotton rat (Sigmodon hispidus) model after infection with a bovine adenovirus type 3 recombinant virus expressing the firefly luciferase gene./. Gen. Virol. 77, 1-9. 138. Zakhartchouk, A. N., Pyne, C , Mutwiri, G. K., Papp, Z., Baca-Estrada, M. E., Griebel, P., Babiuk, L. A., and Tikoo, S. K. (1999). Mucosal immunization of calves with recombinant bovine adenovirus-3: Induction of protective immunity to bovine herpesvirus-1. / . Gen. Virol. 80, 1263-1269. 139. Babiuk, L. A., and Tikoo, S. K. (2000). Adenoviruses as vectors for delivering vaccines to mucosal surfaces./. Biotechnol. 83, 105-113. 140. Reddy, P., Idamakanti, N., Pyne, C , Zakhartchouk, A., Godson, D., Papp, Z., Baca- Estrada, M., Babiuk, L., Mutwiri, G., and Tikoo, S. (2000). The immunogenicity and efficacy of replication-defective and replication-competent bovine adenovirus-3 expressing bovine herpes-1 glycoprotein gD in cattle. Vet. Immunol. Immunopathol. 76, 267-268. 4 7 8 Gerald W. Both 141. Baxi, M., Deregt, D., Robertson, J., Babiuk, L., Schlapp, T., and Tikoo, S. (2000). Recombinant bovine adenovirus type 3 expressing bovine diarrhea virus glycoprotein E2 induces an immune response in cotton rats. Virology 278, 234-243. 142. Reddy, P., Idamakanti, N., Zakhartchouk, A., Babiuk, L., Mehtali, M., and Tikoo, S. (2001). Optimization of bovine coronavirus hemagglutinin-esterase glycoprotein expression in E3 deleted bovine adenovirus 3. Virus Res. 70, 65-73 . 143. Vrati, S., Brookes, D. E., Boyle, D. B., and Both, G. W. (1996). Nucleotide sequence of ovine adenovirus tripartite leader sequence and homologues of the IVa2, DNA polymerase and terminal proteins. Gene 177, 35 -41 . 144. Rothel, J. S., Boyle, D. B., Both, G. W., Pye, A. D., Waterkeyn,
J. G., Wood, P. R., and Lightow^lers, M. W. (1997). Sequential nucleic acid and recombinant adenovirus vaccination induces host-protective immune responses against Taenia ovis infection in sheep. Parasite Immunol \9,111-117. 145. Thomas, G., and Mathews, M. (1980). DNA replication and the early to late transition in adenovirus infection. Cell 22, 523-533. 146. Tribouley, C., Lutz, P., Staub, A., and Kedinger, C. (1994). Product of the adenovirus intermediate gene IVa2 is a transcriptional activator of the major late promoter. / . Virol. 68, 4450-4457. 147. Zijderveld, D. C., Difagagna, F. D., Giacca, M., Timmers, H. T. M., and van der Vhet, P. C. (1994). Stimulation of the adenovirus major late promoter in vitro by transcription factor USE is enhanced by the adenovirus DNA binding protein. / . Virol. 68, 8288-8295. 148. Chang, L.-S., and Shenk, T. (1990). The adenovirus DNA-binding protein stimulates the rate of transcription directed by adenovirus and adeno-associated virus promoters. / . Virol. 64,2103-2109. 149. Dan, A., Elo, P., Harrach, B., Zadori, Z., and Benko, M. (2001). Four new inverted terminal repeat sequences from bovine adenoviruses reveal striking differences in the length and content of ITRs. Virus Genes 22, 175-179. 150. Bischoff, J. R., Kim, D. H., Williams, A., Heise, C., Horn, S., Muna, M., Ng, L., Nye, J. A., Sampson-Johannes, A., Fattaey, A., and McCormick, F. (1996). An adenovirus mutant that replicates selectively in p53-deficient human tumor cells. Science 274, 373-376. 151. Rodriguez, R., Schuur, E. R., Lim, H. Y., Flenderson, G. A., Simons, J. W., and Henderson, D. R. (1997). Prostate attenuated replication competent adenovirus (ARCA) CN706: A selective cytotoxic for prostate-specific antigen-positive prostate cancer cells. Cancer Res. 57,2559-2563. 152. Kajon, A. E., Brown, C. C., and Spindler, K. R. (1998). Distribution of mouse adenovirus type 1 in intraperitoneally and intranasally infected adult outbred mice. / . Virol. 71, 1219-1223. 153. Smith, T. A. G., Mehaffey, M. G., Kayda, D. B., Saunders, J. M., Yei, S., Trapnell, B. C., McClelland, A., and Kaleko, M. (1993). Adenovirus mediated expression of therapeutic plasma levels of human factor IX in mice. Nat. Genet. 5, 397-402. 154. Kafri, T., Morgan, D., Krahl, T., Sarvetnick, N., Sherman, L., and Verma, I. (1998). Cellular immune response to adenoviral vector infected cells does not require de novo viral gene expression: Implications for gene therapy. Proc. Natl. Acad. Sci. USA 95, 11,377-11,382. 155. Moffatt, S., Hays, J., HogenEsch, H., and Mittal, S. K. (2000). Circumvention of vector- specific neutralizing antibody response by alternating use of human and non-human adenoviruses: Implications in gene therapy. Virology 171, 159-167. 156. Aghi, M., Hochberg, F., and Breakefield, X. O. (2000). Prodrug activation enzymes in cancer gene therapy. / . Gene Med. 2, 148-164. 157. Sorscher, E. J., Peng, S., Bebok, Z., Allan, P. W., Bennett, L. L., Jr., and Parker, W. B. (1994). Tumor cell bystander killing in colonic carcinoma utilizing the Escherichia coli DeoD gene to generate toxic purines. Gene Ther. 1, 233-238. 16. Xenogenic Adenoviral Vectors 4 7 9 158. Martiniello-Wilks, R., Garcia-Aragon, J., Daja, M. M., Russell, P., Both, G. W., MoUoy, P. L., Lockett, L. J., and Russell, P.J. (1998). In vivo gene therapy for prostate cancer: Preclinical evaluation of two different enzyme-directed prodrug therapy systems delivered by identical adenovirus vectors. Hum. Gene Ther. 9, 1617-1626. 158a. Lockett, L. J., and Both, G. W. (2002). Complementation of a defective human adenovirus by an otherwise incompatible ovine adenovirus recombinant carrying a functional El A gene. Virology, in press. 159. Brown, F., et al. (Eds.) (1998). "WHO Requirements for the Use of Animal Cells as in Vitro Substrates for the Production of Biologicals. Safety of Biological Products Prepared from Mammalian Cell Culture," Vol. 93. Basel, Karger. 160. Yamashita, T., Ren, C. S., Yoshida, K., Shinagawa, M., Masuda, K., and Fujinaga, K. (1985). Two distinctive transforming DNA regions on the canine adenovirus type 1 genome. Jpn. J. Cancer Res. 76, 436-440. 161. Weiss, R. S., Lee, S. S., Prasad, B. V., and Javier, R. T. (1997). Human adenovirus early region 4 open reading frame 1 genes encode growth-transforming proteins that may be distantly related to dUTP pyrophosphatase enzymes./. Virol. 71, 1857-1870. 162. Thomas, D. L., Shin, S., Jiang, B. H., Vogel, H., Ross, M. A., Kaplitt, M., Shenk, T. E., and Javier, R. T. (1999). Early region 1 transforming functions are dispensable for mammary tumorigenesis by human adenovirus type 9 . / . Virol. 73, 3071-3079. 163. Tollefson, A. E., Ryerse, J. S., Scaria, A., Hermiston, T. W., and Wold, W. S. M. (1996). The E3-11.6-kDa adenovirus death protein (ADP) is required for efficient cell death: Characterization of cells infected with adp mutants. Virology 220, 152-162. 164. Dix, B. R., O'CarroU, S. J., Myers, C. J., Edwards, S. J., and Braithwaite, A. W. (2000). Efficient induction of cell death by adenoviruses requires binding of Elb55 k and p53. Cancer Res. 60, 2666-2672. 165. Zhu, J. D., Grace, M., Casale, J., Chang, A. T. L, Musco, M. L., Bordens, R., Greenberg, R., Schaefer, E., and Indelicato, S. R. (1999). Characterization of replication-competent adenovirus isolates from large-scale production of a recombinant adenoviral vector. Hum. Gene Ther. 10 ,113-121 . 166. Fallaux, F. J., Bout, A., van der velde, L, van den Wollenberg, D. J. M., Hehir, K. M., Keegan, J., Auger, C , Cramer, S. J., van Ormondt, H., van der Eb, A. J., Valerio, D., and Hoeben, R, C. (1998). New helper cells and matched early region 1-deleted adenovirus vectors prevent generation of replication-competent adenoviruses. Hum. Gene Ther. 9, 1909-1917. 167. Blanche, F., Cameron, B., Barbot, A., Ferrero, L., Guillemin, T., Guyot, S., Somarriba, S., and Bisch, D. (2000). An improved anion-exchange HPLC method for the detection and purification of adenoviral particles. Gene. Ther. 7, 1055-1062. C H A P T E R Hybrid Adenoviral Vectors Stephen J. Murphy and Richard G. Vile Molecular Medicine Program Mayo Clinic and Foundation Rochester, Minnesota I. Introduction The characterization of disease at the genetic level facilitates potential genotypic and/or phenotypic correction by gene therapy. Although the concept of gene therapy has been extensively established over the past tv^o decades, the development of effective clinical protocols to facilitate efficacious reversal of disease has proven highly problematic. The development of an effective gene delivery system to the site of therapeutic significance has proven to be the major hurdle to the advancement of gene therapies. Many questions currently remain unanswered and these raise major debates over the best vector systems to treat a specific clinical disorder, and, at a more fundamental level, the choice of gene to be applied. The ultimate goal of a gene therapy protocol is efficient targeted delivery of a therapeutic transgene, whose expression can be sufficiently regulated, in a defective tissue. Vector delivery would ideally involve a single, lifetime treatment by a simple, noninvasive, and safe protocol, which can be incorporated into clinical practice. The vast array of diseases, for which gene therapy presents clinical promise, demands a multitude of different requirements for a vector system to meet. Ideologies for gene therapy vectors will differ considerably among dif ferent disorders. The treatment of severely disabling genetic disorders such as Duchenne muscular dystrophy would require lifelong genetic complemen tation of the defective gene in an immense amount of both skeletal and smooth muscular tissues, as well as brain tissues to correct cognitive functions. Whereas somatic gene therapy for hemophilia B holds out greater potential for treatment; only a few percent of normal, reversed-phenotype cells would be sufficient to provide a constant level of factor IX in plasma, offering patients significant clinical improvements. In contrast to the aim of preservation of host ADENOVIRAL VECTORS FOR GENE THERAPY ^ g \| Copyright 2002, Elsevier Science (USA). All rights reserved. 4 8 2 Murphy and Vile Table I Comparison of the Ideologies of a Gene Therapy Vector for Genetic Disorders and Cancer Genetic disorder Aim: Cell preservation Targeting diseased tissues Efficient transduction of affected cells Therapeutic levels of transgene expression Adequate maintenance of gene expression levels Long term stable transgene expression Minimal vector toxicity Cancer Aim: Cell eradication Targeting diseased tissues Efficient transduction of tumor cells Therapeutic levels of transgene expression Transient vector expression for tumor clearance Vector toxicity — danger signals attack tumor cells physiology for inherited disorders, gene therapy for cancer focuses on efficient cell killing (Table I). Hence, genetic cancer therapies require different vector functions, requiring initial high local transduction of primary tumor masses to effect clinical removal, follow^ed by subsequent systemic vector surveillance to eliminate metastatic disease. In essence, ideological concepts are rarely fully achieved and the current minimal aim of gene therapy is reversal of clinical phenotypes to the extent of easy maintenance, facilitating improvements in standards of life for patients. Despite the development of increasingly complex nonviral gene delivery systems, it is virally derived vector systems which still offer most promise to the clinic. Viruses have throughout evolution developed highly skilled methods of entering cells, evading the host immune defense, and delivering their viral payloads. Hence, phenomenal amounts of research have been directed at harnessing the finely tuned transduction functions and obligate parasite lifestyles of viruses. A plethora of genetically modified viral vector systems has now been reported, all ingeniously subverting the parasitic viral life cycles for the presentation of therapeutic transgenes aimed at reversal of disease phenotype. The development of viruses as cHnical vectors will revolutionize the medical world, providing an invaluable new tool for the treatment of disease. Our present understanding of the molecular genetics of many viruses renders possible their manipulation as cloning vectors for gene transfer both in cell culture and directly in patients. As the major objective is usually long-lasting 17. Hybrid Adenoviral Vectors 4 8 3 gene transfer, deletion of the key regulatory viral genes was deemed essential to manipulate the genetic program of the virus and to ensure that infection of the target cell does not lead to cell death. Conversely, for the treatment of cancer, more recent strategies have reversed this thinking and selectively retain the replicative functions of the virus to enhance tumor cell killing. Viruses have thus been designed with predictable biological properties, retaining the beneficial targeting/infectivity properties, while dissociating them from the major virulent determinants of pathology in normal tissues. Currently, four classes of viral vector have presented most promise as gene delivery vehicles: retroviruses (RVs), adenoviruses (Ads), adeno-associated viruses (AAVs) and herpes-simplex-based viruses (HSVs). Although retro viruses embodied the pioneering vector when the concept of gene therapy began to emerge as a reality in the early 1980s, adenoviruses have since become the major vector choice in the chnic. More recent advances in the pro duction technologies of HSV- and AAV-based vectors have greatly increased their clinical potentials. Additionally, the lentiviral (LV) subclass of retroviral vectors, with distinct biological properties, has emerged with great potential and has gained individual acclaim from the rest of the group. The major properties of each viral vector are presented in Table II, as well as being briefly discussed below. A. Retroviral Vectors Retroviruses are enveloped RNA viruses, whose genomes consist of three core genetic units termed gag^ pol^ and env (Fig. lA) [1]. Retroviruses stably transduce cells by integrating their genomes into the host-cell chromosomes and subsequently release progeny virus by continuously budding viral particles from the cell membrane. The gag gene encodes proteins which form the viral core, while the pol gene encodes reverse transcriptase (RT), the viral integrase (INT), and a viral protease which acts on the gag gene products. The env gene encodes the glycosylated envelope proteins that determine the tropism of the virus. These genetic elements are flanked by the long-terminal repeat (LTR) sequences and a packaging signal (\\|;) which directs the assembly of the genome into the viral particles (Fig. lA) [1]. The LTR sequences contain the ds-acting elements required to regulate viral genome replication and transcription and mediate stable integration into the host genome [1]. Retroviral vectors have been principally based on the well-studied Moloney murine leukaemia virus (MoMuLV). Recombinant MoMuLV vectors are engineered by replacing the gag^ pol^ and ^/^t'-coding units with a transgene of interest, while retaining the LTRs and packaging ds-acting sequences. Producer cell lines stably trans formed with independent gag/pol and env expression cassettes are used to fully complement the viral polypeptides for packaging of the vector proviruses [2, 3]. Hence by transfecting these packaging cell lines with plasmid-based
LTR- flanked retroviral cassettes, retroviral particles efficiently bud from the host >̂ > bJD -a ^ - O H-1 s < ^-T^ -Q (U Z G ^> P^ w o Q a> C/5 < o a. o 2 OJ aus C/3 > c -a w ^ ' >H ^ c > 1 OS >s Q. P <L> I—H JT^ -D J^ V3 I PL] \|\ ^ "o '̂ >̂ I > O ^ S ^ } 0^ 2 I I ON III! ^ 1 c ro ^ _o 1 3 -O 3 oo ĉ s bJD i> 1 1 ^ £ K-1 bf 6̂ bJD a o c p 2 c3 O 2 484 ^ > -^3 O C/3 c3 tJD o S 2 g S ^ i r3 :3 OJ c^ OJ cj ^ (U 'U-i O^ -j-H q i ! o i r !̂ s tr 'S j : ^ -^ •—I "H ^ .ti ^ c ^ • • • 2 § ^ H ^ y .§ O (u IEJII (U .§ 3 ^ ci Ŝ '7^ s g s i ^ I 11o (u <î ^ . 5 5̂ rH O ^ § 8 bJD >- r v ^2 «:§" B O C3 <7i l l l ^ l . ^1 ^1 1CI ^ 2 S I j \| g StjD .ĝS S. f̂5t̂ 1s S^ J^ ^ .̂ •D ^ -g i s fS :i: 8 u If > CI . 2 ci ?5 c &J0 CD o .. (U <U S.^ ^P si o '~̂ bt, C Y <u 2 B s ••̂ ^ n O H ^ K g ^ ^ o hS OS O ^i O i"̂t li p̂ i -̂ 2 H-l cj > ::3 U i H § ^ .S W 2 o O ^ bJD g 7u ^ S S . ^ IT) en u* C ^r '̂ 3 -^ 5 - 1 2H OS O '3 X W) O . > - o ix -rU C • 2 K cl (U _^ "+3 -O CH JJ • ' - ' ^3 ^ •̂ e ^ij r;" 5-1 s-i j - j ^ ^ 3 JC as t j o bJO O S ^ c o ^ O H pti o o U 485 486 Murphy and Vile 5'LTR »F iMPS^BlittKftBill SWIBI 3'LTR /:,-l^^ltftiildfie €5^^'=} v; ii'v:;:;;; •' 1111111 t Late Genes ^ E 3 5'rTR * 3'ITR t *• E2 E4 ' f iraiag^-.iD^^:, B Ori,b acterial Transgene Ampicilin' D Cassette (s) Figure 1 Vector genome structures. The wild-type viral genomes and the strategy of transgene substitution are presented for (A) retrovirus, (B) adenovirus, and (C) adeno-associated virus. (D) The minimal structure of the HSV-1 -based amplicon vector. 17. Hybrid Adenoviral Vectors 4 8 7 cells containing the recombinant retroviral genome. These retroviral particles are capable of infecting cells and directing the expression of the transgene of interest, but cannot replicate or generate progeny virus. B. Adenoviral Vectors Adenoviral particles consist of lipid-free "spiked" regular icosahedra of 60-90 nm in diameter, consisting of three main structural proteins, termed hexon, penton base, and fiber [4]. The genome consists of a double-stranded linear DNA molecule of approximately 36 kb in length, functionally divided into two major noncontiguous overlapping regions, early and late, defined by the onset of transcription after infection (Fig. IB) [5]. There are five distinct early regions (ElA, ElB, E2, E3, and E4) and one major late region (MLR) with five principal coding units (LI to L5), plus several minor intermediate and/or late regions. At the extremities of the viral chromosome are the inverted terminal repeats (ITRs) and the encapsidation signal (\\|;), encompassing the cis elements necessary for viral DNA replication and packaging [5]. Recombinant Ad vectors are constructed by deleting the essential early genes ElA and ElB, whose expression enables transformation of the host cell and trans-actiyatQS expression of the other early viral genes, as well as some host factors [6]. Transgenes are inserted into this deleted region (Fig. IB) and can be assembled into infectious adenoviral particles in cell lines which trans- complement the ElA/B functions [7]. Additional deletions in the nonessential E3 region are also often performed to increase cloning capacities [8], Thus infection of cells with the Ad vector enables expression of the transgene in the absence of expression of viral proteins. Further incapacitation of the Ad vector genomes, limiting leaky expression of viral proteins by further deletions in the E2 or E4, has also proved advantageous, further enhancing the cloning capacities, but requiring further complementation functions in packaging cell lines [9-12]. The development of so-called "gutless" or helper-dependent (HD) adenoviral vectors has also greatly expanded the potential of Ad vectors. These vectors retain just the terminal ITRs and ijf required for replication and packaging of adenoviral genomes, greatly increasing the cloning capacity [13]. C. Adeno-associated Viruses Adeno-associated viruses have recently become attractive candidates for gene transfer. AAVs belong to the family parvoviridae and consist of nonenveloped icosahedral virions of 18-26 nm diameter, with linear single- stranded DNA genomes of 4680 nucleotides for the most characterized AAV2 strain [14, 15]. The genome consists of two coding regions, cap and rep, which are flanked by ITRs and encapsidation signals (\\|/) at either end of the genome (Fig. IC). The cap gene encodes the capsid (coat) proteins and rep encodes proteins involved in replication and integration functions [15]. 4 8 8 Murphy and Vile After infection, AAV genomes can persist extrachromosomally in an episomal form [16, 17] or integrate into the cellular genome [18, 19]. AAV has been demonstrated to preferentially integrate into human chromosome 19 at site ^13.4 (AAVSl), directed by the rep genes, facilitating latent infection for the life of the cell [20]. AAV is, however, naturally replication-incompetent and requires additional genes from a helper virus infection, v^hich in nature is generally complemented by Ad or HSV coinfection [21]. AAV-based vectors generally involve replacement of the rep and cap genes v\̂ ith a transgene of interest (Fig. IC), retaining the terminal repeats and packaging sequences essential to direct replication and packaging of the genome [15]. These AAV vectors can be packaged into infectious AAV particles upon complementation of the rep/cap genes and Ad/HSV helper functions in trans. Deletion of the rep genes, however, eliminates targeted integration of the AAV cassettes at AAVSl. The nonpathogenic nature of AAV, having not been associated with any disease or tumor in humans, makes it a potentially powerful clinical vector. D. Herpes Simplex Viruses Herpes simplex virus belongs to the herpesvirus family, a diverse family of large DNA viruses, all of which have the potential to establish lifelong latent infection [22, 23]. HSV consists of 110-nm-diameter particles comprising an icosahedral nucleocapsid, surrounded by a protein matrix, the tegument, which in turn is surrounded by a glycolipid-containing envelope [24]. The HSV-1 genome consists of a giant linear double-stranded DNA molecule of 152 kb encoding 81 known genes, 38 of which are essential for virus pro duction in vitro [24]. First-generation HSV-based vectors involve replacement of one or more of the seven immediate-early (IE) genes whose functions are ^raws-complemented by packaging cell lines [24]. Second-generation HSV- amplicon vectors consist of plasmids containing just the HSV-1 origin of replication (Oris) for replication in packaging cell lines by the roUing circle mechanism, and the cleavage/packaging signal (pac) (Fig. ID). These amplicon vectors can accommodate inserts of up to 15 kb, enabling the assembly of concatemer structures of up to 10 genomes, reconstituting the packaging size of 150 kb [25]. The future construction of "full-size" gutless HSV vectors could accommodate up to 150 kb of insert DNA [26]. E. Lentiviral Vectors Lentiviruses are a subclass of retroviral vectors which have become infamous in world affairs by the HIV family members. LV vectors are charac terized by the presence of additional accessory genes to the gag/pol/env-hased genomes [27]. These accessory genes extend the functions of the viruses, with 17. Hybrid Adenoviral Vectors 4 8 9 the major gene therapy focus being on their abiUty to infect nondividing as well dividing cells, in distinct contrast to other retroviral family members such as MoMuLV. These karyotropic properties of lentiviruses provide a promising tool to direct retro virus-mediated gene therapies to nondividing cells [28]. LV vectors are constructed by an analogous mechanism to conventional MoMuLV vectors. F. The Choice of Gene Therapy Vector No single vector system can presently provide the necessary flexibility for all the possible clinical applications of gene therapy. Vast variabilities exist in vector host range and uptake potentials for the many tissues of the human body, which together with the many biological barriers to reaching the target tissues make a universal vector unlikely. Thus disease-specific gene targeting strategies are likely to be required, involving the development of multiple gene delivery systems. Hence the technology of gene therapy stands to benefit from the vast range of clinical vectors being designed, each system having distinct properties which can complement each other in the clinic. Extensive research has focused on the potential of adenoviruses as transducing viruses for use in gene therapy. The translation of laboratory- derived viral vectors as practical pharmaceutical tools is a major determinant of gene therapy interest in the clinic. In essence, the ease of generating Ad vectors, the efficiency of purification, and the superior titers which can be obtained (>10^^ pfu/mL) have made Ad the vector of choice for many applications of in vivo gene therapy [4]. The rapid technical advances in the construction and purification of alternative viral vector systems has, however, expanded clinical interests. The vastly improved techniques of helper-free AAV production have significantly increased the potential of these vectors. Titers of AAV vectors equivalent to those of Ad vectors are now routinely achievable, which are free of the once problematic helper virus contamination [29]. The production procedures, however, are still relatively laborious and problematic. The comparatively low titers of the MoMuLV-, HSV-, and LV-based vectors, generally greater than 2 logs lower stable titers, limit the effectiveness of these vector systems especially upon translation to the clinic. However, current immune system barriers preclude the beneficial attributes of administration of Ad vectors at their maximal titers, with significant safety concerns apparent with the maximal doses of Ad vectors in the clinic [30]. The generation of large-scale, high-titer vector preparations with stable shelf lives is essential for clinical applications. The stable pharmaceutical prop erties of Ad virions, as well as the similarly encapsidated AAV and HSV virions, present significant advantages over the much less stable enveloped retrovirus- based vectors. The integrative functions of retroviral vectors, however, confer on them the potential of long-term stable expression, fulfilling an additional 4 9 0 Murphy and Vile highly desirable vector property. These integrative functions, together with rapidly advancing methods of enhancing viral titers using concentration proce dures [31], maintain major clinical interest in retroviral vectors. The integrative functions of AAV vectors are also highly desirable, specifically the chromoso mal targeting mechanism in the presence of the Rep protein [32]. The absence of any cellular retention mechanisms for Ad and HSV vectors presents a distinct disadvantage to many gene therapy applications. In the context of tumor erad ication, however, high-titer vector transduction is unlikely to require long-term maintenance of vectors. In deciding the most appropriate vector for treatment of a clinical disorder, the main selection criterion for vector choice comes in the ability of a specific vector to efficiently transduce the target tissue. Ad vectors have a wide distribution of their target receptors dispersed throughout the body tissues. AAV and HSV have similar diverse tropism to most cells in the human body, with HSV-1 vectors having a major selective tropism for neuronal tissues. MoMuLV viruses are, however, severely limited by their dependence on host cell mitosis to enable stable transduction of a cell, limiting their efficacy in quiescent cell populations [33]. These cell cycle restrictions are not apparent with the LV subclass of retroviral vectors, which possess the additional nuclear targeting functions [34], The additional nuclear targeting property of LV vectors, together with the integration functions, has significantly raised the clinical interest in respect of gene therapy. Additionally, the ability of MoMuLV vectors to infect only dividing cells can be deemed an advantage in targeting actively dividing tumor cells which are surrounded by nondividing
normal tissues. The ideal vector system is thus very much dependent on the diseased tissue to be treated. The extent of genetic material that is required to be delivered to a specific tissue is also a major influence on the vector system. Ad vectors offer a wide range of insert potentials from 7 to 8-kb insert capacities for first-generation vectors and up to 36-kb inserts in the "gutless" HD vector system [6, 35]. Whereas the relatively small packageable genome sizes of retroviral vectors (-^8 kb), but more significantly of AAV vectors (~4.5 kb), severely limit their applications to some gene therapy protocols [15], specifically where the delivery of multiple genes or the insertion of large regulatory elements is deemed essential. It is, however, the HSV-1-based vectors that offer the superior transgene delivery potentials with inserts of up to 150 kb feasible in a gutless vector [36]. Additionally, in the alternative HSV-1 amplicon vector system, as well as providing an insert capacity of up to 15 kb, the assembly of concatemers vastly increases the copy number of transgene cassettes being delivered to target cells [25]. The immune system is a perpetual barrier to viral transduction. The compromised state of many diseases would be severely stressed by fur ther immunological effects/inflammation induced by a "therapeutic" vector 17. Hybrid Adenoviral Vectors 4 9 1 challenge. The exception again is cancer gene therapy where activation of local immune responses can be advantageous in tumor recognition and possibly aid in breaking immune tolerance [37]. Viral vectors are designed to exploit specific biological properties of viruses, such as recognition of cell receptors for entry and mechanisms of host genome integration, that have evolved over time in relationship w îth the host. The natural response of the host has, hov^ever, also developed to eliminate disease-inducing viral pathogens. Current strate gies of viral vector design are working to engineer viruses with predictable biological properties, maintaining the biological advantages of the virus that have been selected by nature while reducing the immunogenicity of the viral components. The majority of Ad vector-derived immunogenicity was deemed to be due to the leaky expression of retained viral transcripts in the vector genome [30, 38]. For AAV and HSV amplicon vectors, contaminating helper virus was also deemed highly immunogenic. The more recent improvements in Ad vector design [9-12] and generation of "helper-free" packaging systems for AAV and HSV amplicon vectors [29, 39] has stunted this immunogenicity to some extent. However, the immune system still stands as a major barrier to gene-therapy efficacy. The mere physical presence of the virus can induce significant cytopathology. The current requirement of repeated administration to boost expression levels further augments the immune memory responses to the presence of the virus until eventual complete immunity is developed to the applied vector [40]. The power of the immune system is emphasized by practically 95% of Ad virions being eliminated by the natural nonspecific innate immune response on each administration [41]. G. How to Maintain Stable Transgene Expression The transient natures of Ad and HSV-1 vectors, as well as the rapid loss of transgene expression upon stable integration of AAV and RV vectors due to nuclear effects on the transgene cassettes, have dramatically limited the efficacy of each vector system. Hence the question remains: how do we maintain stable transgene expression following recombinant viral vector transduction.^ One solution may come from looking closer at the wild-type mechanisms of preservation evolved by the parental viruses. Viruses have developed diverse mechanisms of self-preservation and maintenance to enable them to infect cells and direct self-replication and propagation. Mechanisms of maintenance vary according to the life cycle of the virus. Viruses such as retroviruses have developed life cycles that live in harmony with the host cell. They utilize the host cellular machinery to enable continuous shedding of the virus and thus require stable preservation of the viral genetic material. Retroviruses facilitate this function by stable integration into the host genome, permitting continuous replication/maintenance of the viral genome in the context of host cell replication [1]. Conversely, lytic viruses 4 9 2 Murphy and Vile such as adenoviruses subvert the host's cellular functions solely for their ow n̂ preservation. Infected cells become short-term factories of virus production, amassing viral particles until host cell saturation is achieved and cell lysis occurs in less than 36 h [5]. The short-term association of virus and host does not therefore necessitate mechanisms for long-term persistence of the viral genome. The Ad genome is thus maintained extrachromosomally w îth a very efficient mechanism of replication to enable large-scale genome packaging into the vast numbers of viral particles generated. Herpes viruses, such as HSV, Epstein-Barr virus (EBV), and cytomegalovirus (CMV), have developed more complex mechanisms of self-preservation [42]. Upon infection, a lysogenic life cycle enables the virus to live in harmony with the cell, maintaining the genome in an extrachromosomal state, w^here methylation and histone binding to the viral genome keep viral gene expression essentially quiescent [22]. The sv^itch of the life cycle from the quiescent latent state to the major virulent lytic phase, upon signals of cell stress, rapidly reveals the viral presence. This terminal lytic stage of rapid viral genome reproduction and mass assembly of virions enables the virus to rapidly multiply and abandon the host. The AAV life cycle is a further intriguing evolutionary mechanism, being naturally dependent on helper Ad or HSV coinfection to effect lytic AAV virion assembly and viral progeny release. In the absence of such helper functions, AAV remains lysogenic by either stable integration into the host genome or independent episomal replication in the infected cell [14]. II. Hybrid Viral Vectors The inadequacies of each viral vector system are illustrated in Table II. The negative attributes of one vector, however, generally emphasize the positive attributes of another. Thus most of the criteria defined for a hypothetical perfect gene therapy might actually be met by considering defined properties of the currently available vectors defined in Table II. Hence, although at present no individual virus system alone can meet all the criteria, current research is focusing on combining individual viral properties into single vector constructs, termed "hybrid" or "chimeric" vectors. Adenoviral vectors are currently the major vector choice for a variety of clinical disorders, despite the limited efficacy due to the transient nature of the vector. Mechanisms of enhancing the pharmaceutical properties of Ad vectors are thus highly desirable. The incorporation of other viral vector functions that could enhance the duration of Ad-directed transgene expression and/or target the vectors to a specific disease tissue would be extremely beneficial. In essence, whether the aim is to kill or cure the target cell, a vector encompassing the advantageous properties of high titer, broad host range, and infectivity of an Ad vector, together with the low immunogenicity and potential for long-term 17. Hybrid Adenoviral Vectors 4 9 3 stable expression of a retrovirus, AAV, or EBV vector v^ould be extremely useful for gene therapy for a wide range of genetic and acquired disorders. Hence the main focus of this chapter is to review the properties of other viral vectors which have been utilized to generate hybrid adenoviral vectors in the aim of enhancing vector efficacy in the clinic. A. Are Hybrid Vectors Truly New Technology? The formation of hybrid adenoviruses is not a new technology and has been extensively reported to occur naturally in nature. Adenoviral/simian virus 40 (SV40) hybrids have been documented to occur in nature [43, 44]. Although human adenoviruses do not normally replicate in primate cells, upon coinfection with SV40, Ad genomes acquired sequences from the SV40 genomes (large T antigen) which permitted replication and assembly of hybrid genomes into wild-type Ad capsid particles [43]. Additionally it may be that the helper-dependent AAV genome represents a segment of an extinct or undiscovered virus that was selected upon coinfection with an Ad or an HSV. Perhaps the parental virus was too virulent to coexist in a human host, thereby explaining the nonpathogenic nature of the dependovirus. The development of hybrid viral vectors is fundamentally not a new technology in gene therapy. Since the dawn of gene therapy, scientists have utilized alternative as-acting sequences from other viruses, specifically pro moters and enhancers, to drive transgene expression. Most significantly, the cytomegalovirus (CMV) immediate-early promoter and enhancer has been utilized in almost every viral vector reported to date and is well characterized as an extremely strong constitutive promoter in most tissues [45, 46]. Other well utilized viral promoters have included the Rous sarcoma virus (RSV) LTR promoter, the SV40 early promoter, hepatitis B virus (HBV), and the EBV promoter [45, 46]. Additionally, application of the picornaviral func tions of "cap-independent" initiation of translation has also been extensively exploited in viral vectors. These translational regulatory elements, termed internal ribosomal entry site (IRES) sequences, enable bicistronic expression from a single mRNA transcript [47]. The application of these elements greatly complemented the limited insert capacities of viral vectors, thereby negating the need for separate promoters to drive two transgene cassettes. Retroviral vectors have been studied in hybrid vector systems since the early 1980s, "pseudotyping" them with functions from other retroviral vectors. Specifically heterotropic viral glycoproteins from other retroviral env genes have been stably incorporated into MoMuLV vector particles. The incorporation of vesicular somatic virus G (VSV-G) glycoprotein [48], gibbon ape leukemia virus (GALV) and HIV-1 glycoproteins [49] into murine leukemia virus particles has been reported. These hybrid MoMuLV virions attain the tropism of the pseudotyped env proteins, retargeting or broadening the host 4 9 4 Murphy and Vile range of the MoMuLV vector. Additionally, incorporation of VSV-G env has been demonstrated to increase the stability of the virions, enabling higher titer-yielding purification techniques to be applied [50, 51]. Hybrid retroviral vectors have also been constructed, incorporating different c/s-acting elements contained in the U3 region of the LTR, w^hich direct the transcriptional activity of the virus. Replacement of these U3 regulatory elements can impart tissue- specific transcriptional activity on the RV vector [52, 53]. Hence the concept of hybrid vectors is not a new^ technology, but the nev^ strategies proposed could vastly expand the repertoire of viral vectors available to the clinic. III. Hybrid Adenoviral Vector Systems A number of hybrid adenoviral vector systems have been reported in the literature, combining the properties of RV, AAV, and EBV vectors, as w êll as elements of other Ad serotypes, to enhance the therapeutic efficacy of Ad vectors in vivo. The principal aim of these new^ hybrid vectors is to overcome the limitations of transient Ad vector retention in infected cells. In addition to the w^ell-documented limitations of Ad vectors (Table II), some initially perceived advantageous properties of Ad vectors do actually limit their effectiveness toward therapy for some diseases. The broad host range of Ad vectors induces significant disadvantages v^hen tissue targeting is required and compromises systemic administration. Additionally, the \ow pathogenicity of adenoviruses in humans has resulted in many serotypes, including the conventional vector strains of Ad2 and Ad5, being endemic. Hence a potent natural anti-adenoviral immunity is fashioned generally at a very early age. The highly immunogenic nature of the proteinous Ad virion further confounds the system, v^ith a rapid and highly effective host humoral response being developed to the Ad vector. Research is thus being channelled into both retargeting Ad vectors to specific tissues and silencing the structural immune stimuli to facilitate enhanced Ad vector transduction. A. Pseudotyping and Retargeting Adenoviral Vectors As targeting and humoral immunity are connected in essence to the same surface moieties of the Ad particles, both disciplines are fundamentally interlinked. Methods applied to limit the humoral responses have focused on two main strategies: application of alternative "immune silent" Ad serotypes or display of alternative ligands on the surface of the virions, which is also the major strategy for retargeting the vector. The use of alternative serotypes enables the consecutive application of immunologically distinct Ad particles, enabling avoidance of specific humoral responses to previously applied vectors [54, 55]. This system has presented some success in vivo [56]., although the presence of cross-reacting antibodies 17. Hybrid Adenoviral Vectors 4 9 5 is problematic due to the evolutionary similarities of Ad serotypes. The application of alternative Ad serotypes w îth different surface markers also provides a mechanism of alternative targeting, as different serotypes possess tropism for different tissues in
the human body. For instance, the conventional gene therapy subtypes Ad2 and Ad5 have natural tropism for the gut epithelial layer. Hence, in terms of gene therapy for cystic fibrosis, initial vectors proved disappointing due to their low infectivity of the airv^ay epithelia. To overcome this restriction, Zabner and colleagues investigated other Ad serotypes for airway epithelia tropism [57]. A number of other Ad serotypes, specifically Ad 17, were found to infect the airway epithelia with increased efficiency to wtAd2 [57], They therefore proceeded to generate Ad2 hybrid vectors pseudotyped with the Adl7 fiber, where the endogenous Ad2 fiber gene was replaced with the Ad 17 fiber gene. The resultant chimeric vector displayed increased efficiency of binding and gene transfer to well differentiated human epithelial cells. A similar study by Croyle and colleagues demonstrated that wild-type Ad41 had enhanced transduction properties in intestines compared to Ad5 [58]. These studies emphasize the potential of alternative Ad serotypes with tropism for different tissues in the human body. Pseudotyping also provides an invaluable mechanism of integrating alternative serotype fiber (and/or penton base) genes from other Ad serotypes into the currently well- researched Ad vectors, without having to reconstruct the vector backbones. The use of nonhuman adenoviruses as vectors for gene therapy is also under investigation, with bovine, ovine, canine, feline, and avian adenoviruses being researched [59-62]. As well as being potentially unexposed to the immune system, they may also have specific tropism for selective tissues in humans. The potential of pseudotyping nonhuman Ad vector components with conventional human Ad vectors is therefore of interest. The use of targeted viral vectors to localize gene therapy to specific cell types introduces significant advances over vectors with conventional natural tropism. As well as the safety aspects of reduced immunogenicity and toxicity, the reduced uptake by nontargeted cell types may enable application of systemic delivery with feasible viral titers and loads. In order to retarget Ad vectors, first the natural tropism of the virus must be removed and, second, novel, tissue-specific ligands introduced [63]. Two main mechanisms have been used to retarget Ad vectors. First, the use of external molecules with affinities for both the Ad surface structural moieties as well as a cell-type-specific surface ligand. These bispecific molecules act as bridges between the virions and the cell. A neutralizing antibody or high-affinity peptide for the fiber or penton base can act as the Ad-binding moiety, which can be covalently linked to a high-affinity ligand for a tissue-specific receptor [63]. A drawback of the bridging molecule approach is that native receptor binding is never 100% blocked. To truly block native Ad binding to its cognitive receptor, removal of the intrinsic receptor binding domains is required. 4 9 6 Murphy and Vile A second approach involves creation of hybrid Ad vectors, pseudotyped w îth novel receptor recognition functions. Genetic modification of the Ad genome by incorporating targeting ligands inside the genome, while deleting or ablating sequences of the penton and fiber involved in receptor recogni tion, has been reported. High-affinity peptide motifs have been subsequently demonstrated to be functionally incorporated into Ad particles. These "proof- of-concept" studies focused on the incorporation of ligands w^ithout ablating natural receptor interactions and resulted in expanding the vector tropism, which has proved beneficial in vivo in transducing both vascular smooth mus cle and some tumor types [64-66]. Future studies will focus on honing the targeting functions to specific cell types. High-affinity ligands have been stably inserted into the HI loop or on the C-terminus of the fiber or into the integrin- binding RGD domain of the penton base [63]. However, the size, location, and type of ligand to be inserted are currently under debate and remain to be deter mined. Wickham and colleagues demonstrated 10- to 1000-fold reductions in transduction of cells expressing the coxsackievirus and adenovirus receptor (CAR) with CAR-ablated vectors [63], the residual transduction being penton- base-mediated, emphasizing the requirement for additional ablation of penton base binding [63]. The further requirement of novel packaging cell lines to facil itate infection and propagation of the CAR/integrin-binding ablated particles also remains an issue. B. Adenoviral/Retroviral Hybrid Vector Technologies A hybrid vector system incorporating the advantageous long-term stable integrative functions of retroviral vectors into adenoviral vectors could provide a major clinical advancement to gene therapy. Hybrid vector systems are thus being investigated, incorporating retroviral components into the backbones of adenoviral vectors. Initial studies have focused on utilizing adenoviral vectors as directors of retroviral vector production, delivering the gag^ pol, and env genes as well as retroviral LTR cassettes to cell populations both in vitro and in vivo. Conventional retroviral packaging cell lines are stably transformed with gag^ pol, and env functions and release retroviral particles upon plasmid transfection of a retroviral LTR transgene cassette [2]. High-titer retroviral stocks of greater than 10^ infectious units (iu)/mL can now be obtained from conventional stable producer cell lines [3]. To achieve the highest vector titer, it is necessary to select clones of vector-transduced cells individually due to the varying titers of producer cell clones [67]. Direct injection of retroviral vectors in vivo has, however, yielded limited efficiencies due to the limited transducing titers and poor infectivity. Application of retroviral vectors in the clinic has thus focused on ex vivo protocols. This involves the removal of patient tissues, which can be cultured for a brief period in the laboratory, transduction with the RV vector, and reimplantation back into the patient. The ex vivo approach has 17. Hybrid Adenoviral Vectors 4 9 7 yielded some success, although the procedure is cumbersome and costly, and in most cases, it can only transduce a small fraction of the target cells [68, 69]. The establishment of retroviral producer cells in situ provides a further mechanism of enhancing the efficacy of retroviral gene therapy. Transient transfection of target cells in vivo v^ith the retroviral vector and packaging plasmids, previously used to generate producer cell lines in vitro^ by direct DNA injection has been reported [70]. Although stable integration of subsequently generated retroviral particle genomes could be detected, the efficiency w âs very lov^. The implan tation of retroviral producer cell lines into patients has presented a far greater potential for the in situ production of retroviral vectors. Gene therapy using MoMuLV-based producer cells to treat brain tumors [71] has been carried out in a clinical trial, but no clear clinical benefit has been reported to date. The infectivity of Ad vectors both in vitro and in vivo provides great potential in increasing the efficiencies of retroviral production technology. The group of David Curiel pioneered the development of hybrid retrovi ral/adenoviral vectors by using the infectivity of adenoviral vectors to efficiently deliver the requisite retroviral packaging and vector functions to target cells in vivo^ thereby rendering them retroviral producer cells in situ (Fig. 2). The sub sequent release of high local concentrations of retroviral particles in situ v^ould enable stable transduction of neighboring tissues, for the transient period of adenovirus transduction. The Ad/RV hybrid system reported by Feng and col leagues utilized a tvs^o-adenovirus delivery strategy [72]. The first adenovirus contained an LTR-flanked retroviral vector cassette encompassing the GFP marker and neomycin resistance genes: Ad/RV-vector. The second adenovirus contained the replication-defective retroviral helper machinery, carrying the gag^ pol, and env genes of MoMuLV: Ad-gag/pol/env. High-titer adenoviral vectors could be generated containing the RV cassettes, w^hich could efficiently direct the in vitro packaging of RV particles at titers similar to conventional packaging cell lines [72, 67]. These studies clearly demonstrate the compati bility of both the adenoviral and the retroviral life cycles in the context of a hybrid vector configuration. Upon infection of cells in vitro w îth the Ad/RV vector alone, high initial levels of GFP expression v^ere observed but gradual loss of expression w âs documented over a period of 60 days as the nonintegrated adenovirus v^as lost from dividing target cells. Conversely, upon application of both adenoviruses to cells in vitro, GFP expression was persistent for extended periods of time. How^ever, the persistent level of gene expression w âs reduced beyond the time at v^hich expression could be solely attributed to the Ad/RV vector. The stable integration of the retroviral cassette in surrounding cells was believed to be responsible for this extended expression. The longer term GFP-expressing cells in cultures transduced with both Ad vectors, were present in clustered outgrowths, suggesting local retroviral spreading and/or clonal origin. Subsequent demonstration of proviral integration was confirmed by the 498 Murphy and Vile \| l T R \| ¥ i ^ ^ ; J }.'• Adbetmm '': • ' \|ITR\| 1ITR1 'P \;\ 'MI0^X • M Genome ITR\| 1 ITR\| W lIlTraasi^aeli.i, .3 •:7:liM!(^^t<m-L^.. MITR] Ad Gag/Pol Ad LTR-Transgene AdEnv Ad Infection and Nuclear Delivery Producer Cell RV Particles Figure 2 Hybrid-Ad/RV vector-mediated production of RV particles. Hybrid adenoviral vectors expressing the gag/pol and env RV genes either together or on split adenoviral constructs (as shown) are coinfected with the Ad-LTR transgene vector into cells in vitro or in vivo. Subsequent expression of the gog/pol and env genes in the cells establishes in situ retroviral producer cells which direct the packaging of the expressed retroviral genonnes. RV particles expressing the transgene cassettes subsequently bud from the cells and are released into the surrounding environment. presence of retroviral transgene sequences in high-molecular-weight cellular DNA [72]. No replication-competenet retroviruses (RCRs) were detected v^ith the Ad/RV chimera, despite the large genome copy numbers associated w îth adenovirus production in vitro [72]. 17. Hybrid Adenoviral Vectors 4 9 9 The ex vivo efficacy of the Ad/RV hybrid vector system was investigated by transducing the ovarian carcinoma cell line SKOV3 in vitro with the Ad/RV vector alone or in combination with the kd-gaglpollenv vector at an m.o.i. of 50 pfu/cell. The infected cells were then mixed at a ratio of 3:1 with uninfected SKOV3 cells and subcutaneously implanted in athymic nude mice to allow tumor formation. Tumors were assessed 20 days posttransplantation for GFP expression. Large expansive clusters of GFP-expressing cells were observed only in tumors treated with both vectors. Further in vivo studies involved direct intraperitoneal injection of 5-day-old established SKOV3 tumors in nude mice (1 x 10^ cells) with the single- or two-adenoviral strategy (1 x 10^ pfu/mouse). Sixteen days post-Ad infection, the two-virus-treated tumors were observed to have islands of GFP-positive cells (10-15% transduction), consistent with secondary retroviral transduction. In contrast, single virus- treated tumors revealed very limited (<1%) GFP-positive cells. This pioneering study thus established the great potential of hybrid Ad/RV vectors, whose pros and cons are presented in Table III and discussed further in the concluding remarks of this chapter. Following the initial proof of concept, a number of other laborato ries have further investigated the concept of Ad-mediated establishment of retroviral producer cell lines in situ, Duisit and colleagues in collaboration Table III Pros and Cons of Hybrid Adeno-/Retroviral Vectors for Gene Therapy Pros • Exploit high-efficiency adenoviral infection to deliver retroviral assembly machinery • Utilize stable high titer adenoviral vectors • Increase the duration of biological activity of delivered transgene • Avoid initial limitations of retroviral infectivity to nondividing tissues • Utilize adenoviral vector tropism • Therapeutic gene expressed by retroviral cassette will still be expressed in context of sole delivery of the Ad-RV cassette vector • Initial burst of transgene expression can be converted to a stable low^er level expression • Delivery deep into cell layers Cons • Progeny retroviral vector can still only infect dividing cells • In situ released retroviral vector limited according to RV infectivity and tropism • Requires codelivery of two or more adenoviral vectors • Risk of RCR • Safety? • Rescue of endogenous retroviral elements • Interactions with host cell functions? • Diffusion may still be very limited around the initial needle tract • In situ titers of retroviral particles may be limited • Different cells have different intrinsic potentials for retrovirus production 5 0 0 Murphy and Vile with Fran^ois-Loic Cosset, reported on an extension of the hybrid vector system [67]. These studies further restricted the potential for RCR by sep arating the gag/pol core particle-expressing elements from the env surface glycoprotein gene, which they supply on a separate Ad vector (Ad-gag/pol and Ad~env; Fig. 2), to minimize retroviral sequence overlaps. Additionally, in the context of pseudotyping retroviral vectors, they replaced the