ALBERT

Description: Abstract 377 Hemophilia B represents a promising model for the development of adeno-associated viral (AAV) vectors-based gene therapeutics. In the first clinical trial for AAV serotype 2 mediated gene transfer of Factor IX (F.IX) to the liver of severe hemophilia B subjects, transgene expression was short-lived with a gradual decline of F.IX levels. The loss of transgene expression was accompanied by a transient transaminitis, which we hypothesized to be the result of the reactivation of a pool of capsid-specific memory CD8+ T cells originated from a previous exposure to wild-type AAV. These results were unanticipated since previous work in small and large animal models showed that AAV administration is uneventful, allowing prolonged expression of F.IX transgene at therapeutic levels. We developed an in vitro cytotoxicity assay using a human hepatocyte cell line expressing HLA-B*0702, a common MHC class I allele for which the AAV capsid immunodominant epitope VPQYGYLTL was identified. Using this model, we demonstrated that HLA-matched AAV-specific effector CD8+ T cells were able to lyse target hepatocytes transduced with AAV-2. We now use this in vitro model of CTL killing of AAV-transduced hepatocytes to demonstrate the efficacy of a novel strategy to circumvent undesirable immune response through the engagement of regulatory T cells. A recently characterized MHC Class II-restricted T cell epitope (Tregitope) in the Fc fragment of IgG has been shown to induce regulatory T cells in vitro and in vivo (Blood, 2008; 112: 3303-3311). AAV-specific HLA-B*0702 effector cells expanded in the presence of a human Tregitope peptide resulted in 79% to 89% inhibition of cytotoxic activity against peptide-pulsed and AAV-transduced target cells, respectively. These results were confirmed using PBMCs from 5 different donors. A similar degree of inhibition of CTL activity was observed for the HLA allele A*0101, which binds to the AAV-derived epitope SADNNNSEY; co-culture of effector cells with the Tregitope inhibited CTL-mediated killing by 60%. Interestingly, the same Tregitope efficiently mediated suppression of CTL activity in subjects carrying different HLA alleles, indicating a high level of promiscuity of Tregitope binding. Staining for the regulatory T cell markers CD4, CD25, and FoxP3 supported the hypothesis that Tregitopes suppress T cell responses by expanding regulatory T cells; 62.2% of the CD4+ population stained positive for CD25 and FoxP3 in PBMCs expanded against AAV epitopes in the presence of Tregitope, compared with PBMCs expanded against an AAV epitope alone (3.63%), or against an AAV epitope and an irrelevant control peptide (1.94%). Polyfunctional analysis for markers for T cell activation showed that CD8+ T cells incubated in the presence of Tregitope had an approximately 5-fold decrease in production of IL-2 and IFN-γand a 2-fold reduction in TNF-α production, indicating levels of activation close to naïve CD8+ T cells. We further characterized the mechanism of action of Tregitopes by showing that Tregitopes are required at the time of CD8+ T cell priming, as CTL activity of AAV-expanded CD8+ T cells against transduced hepatocytes was not inhibited by the CD4+ T cell fraction of PBMC expanded separately in vitro with Tregitopes only. We conclude that the use of Tregitopes represents a promising strategy for antigen-specific, Treg-mediated modulation of capsid-specific T cell responses. Disclosures: Martin: EpiVax: Employment. De Groot:EpiVax, Inc.: Employment, Equity Ownership.

Description: Abstract 3762 AAV vectors have received a great deal of attention for clinical gene therapy (GT), since they transduce many mitotic and quiescent cells and mediate long-term transgene expression. Unfortunately, many of the serotypes of AAV commonly employed in GT procedures ubiquitously infect humans, generating pre-existing immunity against the AAV capsid proteins that precludes efficient transduction or induces CTL responses to the transduced target tissue. At present, highly successful animal studies have not translated into clinical success in humans, due, at least in part, to the paucity of animals which harbor endogenous antibodies which recognize and bind AAV-2 and other AAV serotypes for which humans are the natural host. Sheep have long been used as a model to study a broad range of disease states, and a high degree of clinical predictability has consistently been observed. We therefore examined whether sheep possess antibodies to AAV and could thus serve as a much-needed preclinical model system for evaluating AAV-based GT. ELISAs were performed on sera from a panel of 6 healthy Merino-Rambouillet sheep using AAV-1,-2,-5,-6,-8, -9 particles as the antigen. Our results demonstrate that sheep naturally harbor antibodies to all 6 AAV serotypes tested, yet the titers against the different serotypes varied greatly from sheep-to-sheep. While one sheep exhibited very high level (〉2300ng/ml) IgG against all 6 AAV serotypes tested, others exhibited moderate/low (〉350ng/ml) IgG against all 6 AAV serotypes, and still others exhibited moderate/low level IgG against only 3–4 of the tested serotypes. Despite these differences, all sheep harbored detectable antibodies to AAV 2, 5, & 8. A luciferase-based neutralizing antibody (NAB) assay was then performed on sera from 3 of the sheep exhibiting the highest titer IgG against AAV 2, 8, & 9 to assess the clinical significance of these antibodies in the context of AAV-based GT. All 3 animals harbored relatively high titer (1:100-1:316) NAB to AAV 2, but only 1 animal harbored significant NAB titers against AAV 8 & 9 (1:31, and 1:100, respectively). B cell epitope mapping of these 3 animals with a library of peptides derived from the capsids of AAV 2, 5, 8, and 9 revealed that each individual sheep harbored antibodies recognizing from 17 to 50 of the various capsid-derived peptides, some of which were common to all capsids, and some of which were unique to specific AAV serotypes. Importantly, many of the identified capsid epitopes have also been shown to be recognized by antibodies present in human patients with existing AAV immunity. To our knowledge, this is the first report of an animal disease model harboring naturally occurring functional antibodies to serotypes of human AAV commonly employed as GT vectors. The close parallels between human and sheep physiology, coupled with our recent re-establishment of sheep with severe hemophilia A with a null mutation in the FVIII gene and the presence of these antibodies, suggest that sheep may represent an ideal large animal model system in which to study GT in the context of pre-existing immunity to AAV, and to develop novel strategies for circumventing this immunologic barrier. Disclosures: No relevant conflicts of interest to declare.

Description: Inhibitory antibodies to factor VIII (FVIII) are a major complication in the treatment of hemophilia A, affecting approximately 20% to 30% of patients. Current treatment for inhibitors is based on long-term, daily injections of large amounts of FVIII protein. Liver-directed gene therapy has been used to induce antigen-specific tolerance, but there are no data in hemophilic animals with pre-existing inhibitors. To determine whether sustained endogenous expression of FVIII could eradicate inhibitors, we injected adeno-associated viral vectors encoding canine FVIII (cFVIII) in 2 strains of inhibitor hemophilia A dogs. In 3 dogs, a transient increase in inhibitor titers (up to 7 Bethesda Units [BU]) at 2 weeks was followed by continuous decline to complete disappearance within 4-5 weeks. Subsequently, an increase in cFVIII levels (1.5%-8%), a shortening of clotting times, and a reduction (〉 90%) of bleeding episodes were observed. Immune tolerance was confirmed by lack of antibody formation after repeated challenges with cFVIII protein and normal protein half-life. A fourth dog exhibited a strong early anamnestic response (216 BU), with slow decline to 0.8 BU and cFVIII antigen detection by 18 months after vector delivery. These data suggest that liver gene therapy has the potential to eradicate inhibitors and could improve the outcomes of hemophilia A patients.

Description: Key Points Liver-restricted expression of FIX-Padua induces immune tolerance to the transgene in hemophilia B inhibitor dog models. Long-term toxicity studies show no increased risk of thrombogenicity of FIX-Padua in mice and dogs.

Description: We have recently reported that CD8+ T-cell memory maintenance after immunization with recombinant human adenovirus type 5 (rHuAd5) is dependent upon persistent transgene expression beyond the peak of the response. In this report, we have further investigated the location and nature of the cell populations responsible for this sustained response. The draining lymph nodes were found to be important for primary expansion but not for memory maintenance, suggesting that antigen presentation through a nonlymphoid source was required. Using bone marrow chimeric mice, we determined that antigen presentation by nonhematopoietic antigen-presenting cells (APCs) was sufficient for maintenance of CD8+ T-cell numbers. However, antigen presentation by this mechanism alone yielded a memory population that displayed alterations in phenotype, cytokine production and protective capacity, indicating that antigen presentation through both hematopoietic and nonhematopoietic APCs ultimately defines the memory CD8+ T-cell response produced by rHuAd5. These results shed new light on the immunobiology of rHuAd5 vectors and provide evidence for a mechanism of CD8+ T-cell expansion and memory maintenance that relies upon both hematopoietic and nonhematopoietic APCs.

Description: Abstract 3765 Adeno-associated viral (AAV) vectors are one of the most extensively studied vector platforms for gene therapy applications. Our group is currently developing AAV vectors for the therapeutic treatment of hemophilia B (HB) in humans. The first clinical trial using an AAV2 vector to express human Factor IX (hFIX) (AAV2-hFIX16) from the liver of HB patients revealed a cytotoxic T lymphocyte (CTL) response directed against AAV capsid that occurred 4–6 weeks following treatment that was associated with a decline in transgene expression. Thus, immunosuppressive (IS) therapies may be required during AAV2 vector administration at high doses to prevent or to halt the immune mediated destruction of transduced hepatocytes. Previous work in murine and non-human primate (NHP) models has shown that sustained AAV-mediated expression of transgenes can induce tolerance, and that this is in part, dependent on CD4+ CD25+ FoxP3+ regulatory T cells (Tregs). Here we investigate the safety of a Treg sparing anti-T cell IS regimen in the context of liver mediated AAV2 gene transfer. Rabbit anti-thymocyte globulin (rATG) is an immune suppressive drug that is used in solid organ transplant and autoimmune disease. rATG has been shown to dramatically deplete the majority of T-cells, however some studies have shown that rATG spares Tregs and can induce tolerance in human T cells. rATG was administered to rhesus macaques (along with an 8-week course of Mycophenolate Mofetil (MMF) and sirolimus) either at the time of AAV vector administration (AAV2-hFIX16), or 5 weeks post-vector administration (rescue therapy). The administration of ATG at week 5 had no detrimental effect on hFIX expression and was not associated with inhibitor formation (n=3) indicating that rATG might be safe to use as an IS ‘rescue' agent, after the detection of an ongoing immune response against transduced cells. Interestingly we observed that early administration of rATG prevented tolerance induction and resulted in inhibitor formation in 2 of 3 animals upon withdrawal of IS. The inhibitor formation was associated with transient elevations in circulating levels of IL-2, IL-4, IL-10 and IFN-g. These results are comparable to previous findings in NHP using an anti-CD25 IS regimen (Daclizumab) at the time of vector administration (Blood 2007, 110(7):2334-41). We conclude that the timing of IS regimens is critical, and that IS regimens that alter the numbers, frequency, and/or function of T-cells at the time of vector administration can result in neutralizing antibodies (inhibitors) to the transgene product (hFIX). These data suggest that there might be multiple mechanisms responsible for maintaining tolerance in this model, and that Tregs alone might not be sufficient. This study highlights the critical need for safety studies in large animal models of potential immune suppressive regimens in the context of gene transfer before translating to the clinic. Disclosures: High: Genzyme, Inc: Consultancy, Patents & Royalties; Third Rock Ventures: Consultancy; Novo-Nordisk: Consultancy; Shire, Inc.: Consultancy.

Description: Recent clinical trial successes of AAV-based gene therapy for hemophilia B utilizing the hyperactive factor (F) IX variant, Padua (Arg338Leu), demonstrate the potential of this strategy to achieve disease ameliorating FIX activity levels at lower vectors doses. However, the underlying molecular mechanisms responsible for the 8-fold increase in specific activity remain unanswered. To address this question, we have undertaken a series of biochemical comparisons between recombinant FIX-Padua and FIX wild-type (WT). We observe FIX-Padua demonstrates increased aPTT-based clotting activity and increased thrombin generation compared to FIX-WT, as both a zymogen and activated (FIXa) enzyme. Furthermore, FIX-Padua and FIX-WT are activated by FXIa at similar rates. Combined, these results suggest that the hyperactivity of FIX-Padua is due to alterations of the activated enzyme and not due to differences in the zymogen or its activation. We therefore investigated the protease ability of FIXa-Padua and FIXa-WT. In the absence of FVIIIa, FIXa-Padua and FIXa-WT display similar rates of cleavage of peptide substrates. Similarly, FIXa-Padua and FIXa-WT are inactivated by the suicide substrate antithrombin at similar rates. However, when combined with FVIIIa within the intrinsic Xase complex, FIXa-Padua, demonstrates a 3-fold increased catalytic rate (kcat) of FX activation compared to FIXa-WT. Thus, the hyperactivity of FIX-Padua appears to be due to enhanced enzymatic activity with its incorporation into the intrinsic Xase during clotting, and not due to differences in activation or inactivation. The additional fold increase in plasma-based clotting activity may be due to a combination of amplification and feedback reactions occurring in plasma-based assays since the relative difference in clotting activity between FIX-Padua and FIX-WT increases with increasing FX concentration. These biochemical studies support the safety of FIX-Padua as a therapeutic since they suggest that FIX-Padua is regulated similarly to FIX-WT. As such, thrombotic complications would only be anticipated at supratherapeutic FIX activity levels and would be similar for FIX-Padua and FIX-WT at a given activity level. The implication of these results is that the active site of FIXa in the intrinsic Xase is improved by the Padua substitution. We also observe that most substitutions in a 20 amino acid screen at position 338 result in FIX variants with at least equal activity compared to WT. FIX-Padua is the most active variant, while surprisingly, FIX-WT is one the least functional that does not cause hemophilia. Since Arg-338 is highly conserved within FIX mammalian orthologs, these results suggest that FIX-WT may have evolved to have limited procoagulant activity; indeed, our biochemical studies suggest that the FIX-WT active site is sub-optimal for FX activation. We, therefore, examined if there were additional positions in FIX where amino acid substitutions would lead to variants with increased clotting activity. To direct our search, we focused on positions within the protease domain that were 1) likely structurally important for interactions with FVIIIa and 2) where hemophilia B causing mutations have not been reported. Using this strategy, we identified 3 additional positions where single amino acid substitutions resulted in FIX variants with 2 - 6 fold increased clotting activity compared to FIX-WT. Moreover, the most active novel substitution can be additively combined with the Padua substitution: recombinant FIX of this two amino-acid substituted variant displayed 15-fold increased specific activity compared to FIX-WT and almost 2-fold increased specific activity compared to FIX-Padua. Given the proven benefit of the enhanced activity of FIX-Padua now being employed in clinical trials, novel hyperactive substitutions of FIX could further enhance the safety and efficacy of these therapeutics. Disclosures Camire: Pfizer: Consultancy, Patents & Royalties, Research Funding; sparK: Membership on an entity's Board of Directors or advisory committees, Patents & Royalties; Novo Nordisk: Research Funding; Bayer: Consultancy. Arruda:Pfizer: Patents & Royalties, Research Funding.

Description: Emerging data from early phase clinical studies of AAV gene therapy for hemophilia B (HB) (factor IX [FIX] deficiency) show sustained expression of therapeutic levels of FIX and phenotypic improvement. However, the safety and efficacy of in vivo gene therapy is limited by the vector dose. Recently, we reported a naturally occurring, hyperfunctional FIX (FIX Padua) caused by a single amino acid change of arginine 338 to leucine that exhibits an 8-fold increase in specific activity in humans (N Engl J Med 2009), making it a potential candidate for HB gene therapy with reduced vector doses. However, to take advantage of FIX Padua for HB gene therapy, it is critical to first define the risk of immunogenicity of this variant in preclinical models of severe HB. We have previously shown that delivery of AAV-cFIX-Padua to skeletal muscle in HB dogs with a missense mutation in the canine (c) F9 gene resulted in no anti-FIX neutralizing antibodies (inhibitors), non-neutralizing antibodies (IgG) or FIX-specific T-cell response (Blood 2012). While promising, these dogs express FIX RNA and have a pre-existing tolerance to cFIX due to the nature of their mutation, and so do not represent the most rigorous model for immunogenicity studies. Here, we tested the efficacy and immunogenicity of cFIX Padua in a severe HB dog colony with an early stop codon mutation. This mutation results in no FIX RNA transcript, and the dogs are prone to develop cFIX inhibitors upon exposure to protein concentrates. Three dogs were infused peripherally with a liver-specific AAV8-cFIX-Padua at two different doses, and monitored for cFIX antigen and activity levels and inhibitors. The first dog, which received 3 x 1012 vg/kg, showed average plateaued expression levels of 3.98 ± 1.44% antigen and 24.5 ± 4.1% activity, with no development of anti-cFIX inhibitors or IgG antibodies. Whole blood clotting time (WBCT) and aPTTs returned to normal by day 3 post-vector administration and have remained stable for 〉20 months (ongoing observations). A second dog was treated with a lower dose of 1 x 1012 vg/kg and showed average plateaued expression levels of 2.41 ± 0.05% antigen and 22.0 ± 0.4% activity, with no development of anti-cFIX inhibitors or IgG antibodies. WBCT and aPTTs returned to normal by day 3 post-vector administration and have remained stable for 〉3 months. An additional dog, upon previous exposure to recombinant human (h) FIX protein, had developed inhibitors that cross-reacted with cFIX. This immune response was ongoing at the time of vector administration (3 x 1012 vg/kg). Anti-cFIX antibodies peaked at day 14 post-AAV, with 4.7 BUs and 3643 ng/mL IgG2, but dropped to undetectable levels by day 70. There was a concurrent rise in cFIX Padua expression levels, suggesting successful tolerization to the cFIX Padua. Antigen levels plateaued at 14.6 ± 4.3% and activity at 51.7 ± 23.5%, with ongoing normalization of WBCT and aPTTs for 〉18 months. In all three dogs, cholesterol, albumin and total protein were within normal limits with no clinical or laboratory evidence of nephrotic syndrome (a potential complication in FIX inhibitor patients that have undergone immune tolerance induction with frequent FIX protein injections). The safety of FIX Padua was further confirmed using a mouse model of HB. Mice (n=8-12/group) were treated with 5 x 1010 vg/kg liver-directed AAV8-hFIX-WT or AAV8-hFIX-Padua, resulting in expression levels of 1076 ± 343 ng/mL (21.5 ± 6.9% antigen, 67.5 ± 10.1% activity) and 797 ± 255 ng/mL (15.9 ± 5.1% antigen, 274.8 ± 73.8% activity), respectively. In cross-over experiments, the mice were then were immunologically challenged 10-15 weeks after gene delivery with 100 ug/kg of the reciprocal recombinant protein (ie mice expressing hFIX Padua were challenged with hFIX WT, and visa versa). Challenges were administered subcutaneously alone or with adjuvant (CFA) weekly for 4 weeks. In no instance did mice develop antibodies to either FIX, suggesting that tolerance was successfully induced in all cases. Together, these date show that FIX Padua shows no increase in immunogenicity compared to FIX WT and is capable not only of preventing inhibitor formation, but also of eradicating pre-existing inhibitory antibodies to FIX in an inhibitor-prone HB dog model. Thus, FIX Padua is an attractive transgene that will allow for decreased vector doses in human HB gene therapy, improving the safety profile of AAV liver gene therapy without increased immunogenicity. Disclosures: High: Alnylam Pharmaceuticals: Consultancy; BioMarin: Consultancy; bluebirdbio, Inc.: Consultancy, Equity Ownership, Membership on an entity’s Board of Directors or advisory committees; BristolMyersSquibb: Consultancy, membership on a Data Safety and Monitoring Board, membership on a Data Safety and Monitoring Board Other; Elsevier, Inc.: royalties from textbook, royalties from textbook Patents & Royalties; Genzyme, Inc.: Membership on an entity’s Board of Directors or advisory committees; Intrexon: Consultancy; Novo Nordisk: Consultancy, Member of a grant review committee, Member of a grant review committee Other; Shire : Consultancy; Benitec: Consultancy.

Description: Abstract 695 Adeno-associated viral (AAV) vectors are one of the most extensively studied and highly used vector platforms for gene therapy applications. We have recently provided evidence for AAV capsid-derived antigen presentation through MHC class I on the surface of AAV-transduced cells, supporting the hypothesis that in the first clinical trial using AAV to treat Hemophilia B, AAV capsid proteins were presented on the surface of transduced hepatocytes, resulting in clearance by antigen-specific CD8+ T cells and consequent loss of therapeutic transgene expression. Proteasome inhibitors are small molecule compounds that are able to specifically inhibit the activity of the proteasome, resulting in a buildup of ubiquitinated proteins, increased intracellular reactive oxygen species, and a general decrease in presentation of MHCI-peptide complexes. It has previously been shown that proteasome inhibitors can have a dramatic effect on AAV transduction in vitro and in vivo. Here we describe using the FDA approved proteasome inhibitor, bortezomib, to decrease capsid antigen presentation on hepatocytes in vitro, while at the same time, enhancing gene expression in vivo. Using an AAV capsid specific T cell reporter line to analyze effects of proteasome inhibitor on antigen presentation, we demonstrated capsid antigen presentation at low MOI's, as well as inhibition of antigen presentation at clinically relevant levels of bortezomib. We also demonstrate that bortezomib can enhance FIX expression from an AAV2 vector in C57Bl/6 mice, however does not appear to enhance expression of AAV8. Based on the data presented here, it appears as if future studies using proteasome inhibitors in large animal models may be warranted. A pharmacological agent that can enhance AAV transduction, decrease T-cell activation/proliferation, and decrease antigen presentation would be a promising solution to many of the obstacles to successful translation of AAV-mediated, liver-directed gene transfer to the clinic. Disclosures: No relevant conflicts of interest to declare.

Description: The formation of antibodies (or inhibitors) to FVIII is a major complication in the treatment of humans with hemophilia A (HA), affecting up to 30% of individuals with severe or moderate disease. Inhibitors develop mainly in young boys, and render the treatment with infused protein suboptimal or completely ineffective. Inhibitor patients exhibit significant delays in their maturation and physical development. Thus, inhibitor formation negatively affects both hematological and systemic outcomes in these patients and, therefore the prevention or eradication of inhibitors is of fundamental importance. To eliminate the presence of inhibitors in hemophilia, an immune tolerance induction (ITI) is followed consisting of large amounts of FVIII injected on a daily basis for long periods of time (months to years) with variable success rates. Although liver-directed gene therapy has been used to deliver therapeutic transgenes and can induce tolerance to the expressed protein, to date there have been no large animal studies using liver gene therapy to eradicate inhibitors to FVIII. The HA dog colonies are an excellent model for studying tolerance induction as both colonies have the same mutation found in the majority of the human population (inversion of intron 22) resulting in severe hemophilia. We hypothesize that sustained expression of cFVIII could mimic the effects of ITI in eradicating inhibitors to canine FVIII in dogs that have a history of inhibitors, thus demonstrating a potential alternative to the current ITI protocol. We injected AAV-8 vectors encoding the light and heavy chain of canine FVIII (cFVIII) (two separate vectors) driven by a liver specific promoter (2.5 e13 vg/kg per vector) into the peripheral vein of four adult (7–36 months old) HA dogs with inhibitors. The first dog treated (K03, 12 months old) was from a new strain from the Chapel Hill (UNC) HA dog colony prone to inhibitor formation. K03 had a historical maximum inhibitor titer of 12 B.U. and a titer of 3 B.U. at the time of vector infusion. Post AAV injection, we documented a slow shortening of whole blood clotting time (WBCT) to normal values (13 min) and inhibitors decreased to undetectable levels by week 3. K03 has demonstrated cFVIII antigen levels (~30 ng/mL) that are stable out to 250 days post treatment. When challenged with 4 weekly injections of 125 μg of highly purified recombinant cFVIII he failed to generate an inhibitor, indicating induction of immune tolerance. K01 (UNC, 20 months old) had a historical maximum inhibitor titer of 12–13 B.U. and a titer of 3 B.U. at the time of vector infusion. We observed an initial transient rise in cFVIII antigen, followed by a decrease to baseline that corresponded to an increase in inhibitor titer to 5–6 B.U. before decreasing to undetectable levels within 4 weeks of gene transfer. cFVIII antigen levels have slowly increased over time, and have reached 14 ng/mL by day 150. WBCT has decreased corresponding with decreasing inhibitor titers and has normalized to 13.5 min. L44 (UNC, 8 months old) had a historical maximum inhibitor titer of 3 B.U. and a titer of 1.5 at the time of infusion. After vector infusion we observed a rapid onset of cFVIII expression (peaking at 50 ng/mL) that decreased to near baseline by day 8 before increasing and stabilizing at 15–20 ng/ml (out to 40 days) and inhibitor titers have decreased while WBCT has decreased to 18 min. The fourth dog (Wembley, 3 years old) is from the HA dog colony at Queens University and had previously been exposed to both canine and human FVIII and presented inhibitors to both human (10 B.U.) and canine (3.6 B.U.) FVIII. After vector infusion Wembley showed a transient increase in cFVIII levels, followed by a decrease in cFVIII expression and no indication of tolerance induction after 10 weeks, ongoing observation. Notably, this dog presented high titer IgG1 and IgG2 anti-cFVIII antibodies, while the other three dogs had mostly IgG2 (equivalent to human IgG4). No abnormalities in liver or renal functions were observed in these dogs. Factors such as exposure to xeno-antigens, the nature of the antibody response, duration of inhibitor or age may influence the outcome of the tolerance induction protocol. Collectively, these data demonstrate the potential of liver directed, AAV mediated gene delivery to no only treat genetic deficiencies such as hemophilia, but to induce immune tolerance to the transgene in the setting of pre-existing inhibitory antibodies.