840 Liver Directed AAV Mediated Homologous Recombination Is Independent of Serotype induced a CTL response against AAV peptidcs using intcrferon y (IFN y) ELISPOT assay Based on these in vitro studies[.]
induced a CTL response against AAV peptidcs using intcrferon-y (IFN-y) ELISPOT assay Based on these in vitro studies , transient immune suppression has been proposed for human clinical trial However, there was no study about the direct relationship ofCTLs with AAV2 transduced target cells To directly address the role of AAV capsid specific CTLs in elimination of AAV transduce cells in vitro and in vivo , we established a cell line (CT26/Cap-luc) with endogenous AAV2 capsid expression It demonstrated that AAV2 capsid specific CTLs could be induced by dendritic cells with endogenous AAV2 capsid expression or plused with AAV2 vectors These CTLs had the capacity to kill CT26/cap-luc cells in vitro and in vivo (tumor xenograft), or CT26 cells transduced with AAV2 vectors in vitro To determine the effect of CTLs on the elimination oftarget cells transduced by AAV2 vectors in vivo , we carried out adoptive transfer experiments The CTLs eliminated liver cells with endogenous AAV2 capsid expression, but not liver cells transduced by AAV2 vectors regardless of the reporter genes (luciferase, human F9 and human AAT) Also, similar result was observed in AAV2 mediated muscle transduction, no transgene expression was inhibited by capsid specific CTLs after muscular injection of AAV2/AAT vectors Based on these data, it strongly suggests that AAV vector transduced cells can rarely be eliminated by AAV2 capsid specific CTLs in vivo even though AAV vector shell capsid can be cross-presented to MHC class I molecular to induce a CTL response In conclusion, AAV capsid specific CTLs not appear to playa role in elimination ofrAAV transduced cells in vivo The reason may be poor immunogenic of AAV2 capsid in mice It is imperative to establish a mouse model, in which theAAV2 capsid can induce a strong cellular immune response , to investigate the mechanism ofAAV2 capsid antigen presentation and exploit a novel approach to evade AAV2 capsid mediated cellular immune response in gene therapy AAV APPLICATION II 839 Distinct Immunoglobulin G Subclass Responses Following Natural Exposure or Vector Infusion of Adeno-Associated Virus in Humans Samuel L Murphy,' Hojun Li,2 Shyrie Edmonson; Katherine A High.'-' Children s Hopsital 0/ Philadelphia Philadelphia PA; 2Medicine, University ofPennsylvania Philadelphia PA; "Hematology, Howard Hughes Medical Institute Philadelphia, PA IHematology Adeno-associated virus subtype (AAV2) is a vector commonly used in the development of protocols for gene transfer for human disease Recent gene transfer studies have shown that the human immune response is amajor regulator ofthe success or failure ofany attempt to treat disease via gene transfer In this study we analyze the humoral component of the human immune response to AAV2 While previous publications on humoral responses to AAV have been limited to analysis of distribution between IgM and IgG anti-AAV antibodies, we further categorized the distribution of anti-AAV2 IgG antibodies by characterizing antibody responses of individual IgG subclasses 1-4, as well as IgM in human subjects We initially studied a population of 18 normal donors and found that of 18 possessed pre-existing IgM anti-AAV2 antibodies, 10 of 18 possessed IgG I antibodies, of 18 IgG2 antibodies, 13 of 18 IgG3 antibodies, and II of 18 IgG4 antibodies In addition, we found that one normal donor had pre-existing T-cells responding to AAV2 capsid peptides, and this donor was also positive for pre-existing anti-AAV2 antibodies of all subclasses, IgM and IgG 1-4 We also found that of the donors positive for pre-existing neutralizing antibody (NAB) titers greater than 1:30, all were positive for both anti-AAV2 IgG I and IgG3, were positive for IgG2, were positive for IgM, and Molecular Therapy VolumeJ5 Supplement l, ~ br 2007 Copyright © The American Soci ety ot G ene "1l1f:r:lpy were positive for IgG4 We then studied plasma samples from subjects who underwent liver-directed gene transfer with AAV2 In the one subject with documented pre-existing levels ofAAV2 NABs (I: 17), we observed no significant increase in IgG 1-4and IgM levels following gene transfer In the three remaining subjects we found that the rise in anti-AAV2 antibody levels following gene transfer is independent of dose, with the subject that received the smallest dose having the quickest and largest rise in antibody level of all subclasses In this subject, we also found that a rapid rise and fall in all IgG levels was followed by a steady increase of IgG2-4, but not IgG I The other subjects whose plasma samples we studied showed intermediate responses in IgG subclass levels following gene transfer For both normal donors and clinical trial subjects, we determined that IgG I was the dominant subclass in determining totallgG response to AAV2, and that IgG4 responses impacted total IgG response the least We also analyzed subjects who underwent muscle-directed gene transfer with AAV2 and similar results were obtained These findings from the AAV2-hFIX infused subjects are in close agreement with those from an AAVI-LPL injected cohort of subjects Overall, this study demonstrated a large diversity in IgG subclass distributions of subjects naturally exposed to AAV2 Additionally, this study was also able to elucidate trends in IgG subclass responses following in-vivo gene transfer that would have been masked had only total IgG response been measured 840 Liver Directed AAV-Mediated Homologous Recombination Is Independent of Serotype Karsten Wursthorn ,' Nieole Paulk; Terry Storm,' Milton Finegold/ Mark Kay,2 Markus Grompe.' 'Oregen Stem Cell Center; Oregon Health and Sciences University, Portland, OR; "Department a/Pediatrics and Genetics Stanford University-School ofMedicine, Stanford CA; JDepartment ofPathology Texas Children s Hospital Houston , TX AAV serotypes (AAV2) and (AAV8) are widely used for hepatic gene transfer applications in animal models Episomal gene expression is possible in quiescent and dividing cells Given the singlestranded nature ofAAV, gene targeting to correct mutations through homologous recombination have proven successful in vitro and in vivo AAV8-mediated homologous recombination has previously been shown to be capable of correcting the metabolic liver disease hereditary tyrosinemia I (HTI) in vivo We have since expanded our analysis to include AAV2, examined integration stability and analyzed the frequency in relation to the vector dose administered We designed a vector carrying 4500 nucleotides homologous to the mouse fumarylacetoacetatc hydrolase (Fah) genomic sequence with a central point mutation leading to a premature stop codon (pAAV4k) Viral particles ofAAV2 and AAV8 were produced AAV-4k was inj ected into Fah 5% IS8 mice, a model for HT I Targeted recombination events can be detected by Fah immunohistology and clinical course monitoring When injected as newborns at Ix 1011 particles per mouse, AAV2-4k, similar to AAV8-4k, led to BTl correction after selecting for Fah-positive hepatocytes for >4 weeks as measured by weight gain To test for event stability, a 2/3 partial hepatectomy was performed after weeks of selection Hepatectomized mice exhibited clinical recovery simi lar to untreated controls indicating stable homologous recombination Immunohistological analysis of recovered tissue shows that nearly 50% ofthe liver ofthese AAV2-4k injected mice were repopulated with healthy Fah-positive hepatocytes They form nodules expanding from at least one recombination event each The minimum recombination frequency is 1-2 per 10,000 hepatocytes Recipient Fah 5% ' S8 mice undergoing serial transplantation with 300,000 hepatocytes from clinically restored mice show Fah-positive hepatocyte engraftrnent after 5-6 weeks of selection Serial transplantation of corrected liver cells is another means to induce cell division and allows for loss of unintegrated vector In a S321 doseresponse study, varyingdosesfrom3xlOs to 2x10" ofeachserotypewere injectedinto newbornsand livertissue wasanalyzedby immunohistochemistry for Fah-positive hepatocytes after2 (AAV8) or weeks (AAV2) without selection Frequenciesof homologous recombination appear unrelated to doses given, ranging from to >5 per 1000cells Disrupted mismatchrepair systems can increase the frequencyof homologousrecombination Wetherefore injected AAV8-4k into adult FahS%'SB mice that were either heterozygousor mutant for MlhI Detailed analyses are underway Liver directed AAV-mediated homologous recombination is feasible and stable withAAV2 andAAV8 Recombination frequenciesarc independent for a wide rangeof vector doses Studiesto enhancethe rateof gene targeting arc currently under investigation 841 Gender as a Factor for Gene Transfer to Rodent Salivary Glands Antonis Voutctakis,' Changyu Zheng,I Jianghua Wang, I Corinne M Goldsmith,' Martin L wenk,' Molly Vallant,' Richard D Irwin,' Bruce J Baum.' IGTTB, NIDCR, NIH DHHS, Bethesda, MD; "Toxicology Division, Biokeliance Invitrogen Bioservices, Rockville, MD; 3NTp, NIEHS, NIH, DHHS, Research Triangle Park, NC Sexualdimorphism maybe manifested as difl'erences instructure, function, responses to external signals, and other characteristics Salivary glands (SGs) show a high degree of structural variability between the sexes, especially in rodents One of the components of the SG ductal system, the granular convoluted tubules (GCTs), is particularly prominent in male rodents conferring a pronounced sexual dimorphism on the entire organ For example, in mouse submandibularglands the proportionofGCT cells in males is 45% compared to 12% in the females The androgen-dependent GCT cells are of interest for SG gene transfer since they are a major target ofrAAV2 vectors rAAV2s have proven useful for SG gene therapeuticssince they can mediate sustainedtransgeneexpression and leadto highserum levelsof certaintransgenicsecretoryproteins after retro-ductal vector administrationto male mice Interestingly, heparan sulfate proteoglycan, a primary attachment receptor for rAAV2s, is reported to be significantly lower in female submandibular SGs Moreover, murine female glands are reported to have more mucous and seromucous cells, and substantial differences in cell adhesion, extracellular matrix, and basement membrane, compared to their male counterparts Gender-related differences, e.g., higher (-I DO-fold) liver transduction in male mice compared to females, have been observed after intravenous administration of rAAV2 vectors (Davidoff et al., 2003; Berraondo et al., 2005) We therefore wanted to determine the effect of gender on rAAV2 mediatedgene transferto murinesubmandibularSGs.ln this study, whichconformedto the FDAGood Laboratory Practiceregulations, 84 Balb/c mice were evaluated 21 male and 21 female mice were administered with lOelO partieles of a rAAV2 encoding human erythropoietin to bothsubmandibularSGs.An equal numberof both male and femalecontrols (saline administered)were also ineluded Mice were sacrificed on day (n=6 for all groups), and day 30, 55 and 92 (n=5 for all groups) At each time point, submandibular SGs and other major organs were harvested, DNA extracted and Q-PCR performed in order to determine vector copy numbers for each tissue On day after administration, the mean vector copy number (per ug of genomic DNA) in male submandibular SGs was 55xIOe5, while in female mice it was 21x IOe3 (-260x less) Conversely, copy numbers in the liver were comparable: 7xlOe3 and 21xlOe3 for male and female mice, respectively On days 30, 55 and 92 vector copy numbers in male mice were relativelystable: 53x IOe3, SxIOe3, and 12xIOe3, respectively In contrast, no vector could be detected in female SGs Vectorcopy numbers in the liver on day 92 were comparable in both genders (3xlOe2 and Ix lOe2, S322 respectively) No vectorwas foundinany controlmice.Weconelude that the sexual dimorphism encountered in murine submandibular SGs significantly affects rAAV2 gene transfer and suggest that gender differences be considered for both pre-clinical and clinical salivary gland gene transfer studies 842 Gene Correction by Different Repair Substrates Matthew L Hirsch,' Richard J Samulski ' [Gene Therapy Center, University ofNorth Carolina, Chapel Hill Manipulation ofthe humangenomeat a defectivelocusloomsas a formal possibility for the treatmentof a varietyofgeneticdiseasesex vivoand perhapsin vivo.Currentgene correction approaches vary in efficiency and the single-strand adcno-associated virus(AAV) seems to be enhanced for this process Correction following transduction can be as high as I% without selection, a phenomenon dependent uponthe specificmutationand other factors Gene correctionis dramaticallyenhanced by a specificdouble-strandbreak in the vicinity of the mutation(s) for bothviralandexogenous repairDNA.Thesitespecific I-SceI endonuclease of yeast works well for this purpose Here we investigated the ability of different DNA repair substrates to relieve a 35 bp insertion(includinga unique I-Sce! site; Porteus and Baltimore03) inactivatinga g/j:J reporter presentat single copy in humanembryonickidney cells First, we demonstratenear equal gene targetingusinga 1.1 kb homologousg/j:J fragment packaged in AAV asa single-strand or self-complemental)' genome Co-infection experiments withsingle-strand AAV/Sce increased targeting10-fold while self-complementary delivered See increasedcorrection 100fold (4%) at a similar MOL Gfp+ cells were sorted and sequence analysisdemonstratedthe high fidelity of repair,although a mutant sequence(which maintainedphenotype)was recovered The use of a slightly mutant DNA repair substrate (modifiedyJP)gave rise to clones that were genetically g!p or yJP-gfp chimeras We extended our analysis to a panel of DNA oligonucleotidesand demonstrate a ncar 4-fold correction preference for the sense sequence (0.7%) and gene correctionby a 20-mer Experimentsanalyzingadditional DNAsubstrates includingsingle-strandcircles and no-end or openend linear duplexes for gene correction are underway as well as an effort to define their relative stabilities 843 Interaction of Host and Vector-Dependent Factors in Determining the Risk of Germline Transmission of AAV Vectors Patricia Favaro,' Harre D Downey,' Federico Mingozzi,' Frederico Xavier,I Fraser Wright, I Katherine High.P ValderR Arruda.1.2 [Hematology, The Children s Hospital ofPhiladelphia, Philadelphia PA; 2Pediatrics, University ofPennsylvania, Philadelphia PA The growingenrolment in gene-basedclinical trials of patientsat reproductiveage with long-lifeexpectanciesand of children raises several safety concerns, notably, the risk of germline transmission of vector sequences Recombinant adeno-associated viral (AAV) vectoris a promising gene-based strategyforthe treatmentof several inherited disease already tested in adults and pediatric early-phase trial are underway Using a rabbit model, we have determined that the risk of germline transmission by AAV serotype depends on the routeof administration, vector-dose and time fromthe injection Here we sought to exploit whether the host genitourinatytract and viral tissue tropism would modulate reproductive toxicology of AAV vectors We carried out experiments using AAVof different scrotypcs and rabbits in which the germ cell access to the semen was surgically interrupted Adult rabbits underwent to bilateral vasectomy (n=12) and serial semen analysis demonstrated that no spermatogonia were present in the ejaculate of 10/12 rabbits Six Molecular Therapy Volume15.Supplement t, \by 2007 Copyright © '111C American Society of Gene Therapy ... analyses are underway Liver directed AAV- mediated homologous recombination is feasible and stable withAAV2 andAAV8 Recombination frequenciesarc independent for a wide rangeof vector doses Studiesto... 2x10" ofeachserotypewere injectedinto newbornsand livertissue wasanalyzedby immunohistochemistry for Fah-positive hepatocytes after2 (AAV8 ) or weeks (AAV2 ) without selection Frequenciesof homologous. .. long-lifeexpectanciesand of children raises several safety concerns, notably, the risk of germline transmission of vector sequences Recombinant adeno-associated viral (AAV) vectoris a promising gene-based