IMMUNOLOGIC & HOST RESPONSES IN GENE & CELL THERAPY Immunologic & Host Responses in Gene & Cell Therapy 555. AAV2(Y-F) Vector Substantially Reduces Targeting of Transduced Hepatocytes by CapsidSpecific CD8+ T Cells
Ashley T. Martino,1 Etiena Basner-Tschakarjan,2 David M. Markusic,3 Hildegund C. J. Ertl,4 Cox Terhorst,5 Katherine A. High,2,6 Federico Mingozzi,2 Roland W. Herzog.3 1 Pharmaceutical Sciences, Saint John’s University, Queens, NY; 2 Pediatrics, Children’s Hospital of Philadelphia, Philadelphia, PA; 3 Pediatrics, University of Florida, Gainesville, FL; 4Immunology Program, Wistar Institute, Philadelphia, PA; 5Immunology, Beth Israel Deaconess Medical Center, Boston, MA; 6Howard Hughes Medical Center, Philadelphia, PA. CD8+ T cell responses against adeno-associated viral (AAV) vector capsids abrogated long-term liver directed human factor IX (hFIX) correction of hemophilia B in clinical trials, an outcome not predicted in extensive animal studies. Previously, we have shown that an AAV2 triple tyrosine mutant capsid (Y444, 500, 730F), AAV2(Y-F), vector expressing hFIX resulted in sustained therapeutic expression in a murine hemophilia B model. Substitution of surface exposed tyrosine residues in the AAV2 capsid reduces capsid protein poly-ubiquitination, a signal for proteasomal degradation. Since antigen presentation on MHC class I molecules relies on proteasomal processing, we hypothesized that AAV2(Y-F) capsids would be less efficiently presented compared with AAV2. The ability of AAV2(Y-F) capsid vector to avoid destruction by CD8+ T cells was assessed by in vitro CTL assays. A significant reduction in killing was seen in both murine and human hepatocytes pulsed with AAV2(Y-F) compared to AAV2 vector. Use of a reporter CD8+ T cell line confirmed that AAV2(Y-F) derived capsid antigen was presented at significantly lower levels via MHC I at MOIs similar to those used in vivo. Any residual CTL activity was eliminated by proteasome inhibitor (bortezomib) in vitro and in an in vivo model (described below), confirming that MHC I presentation of capsid is dependent on proteasomal degradation. Next, we developed a novel in vivo murine model to assess capsid-specific CD8+ T cell targeting of tranduced hepatocytes. CD8+ T cells specific for a control or capsid epitope (presented by both human B*0207 and murine H2-Ld molecules) were generated and adoptively transferred into immune deficient BALB/c mice transduced with either AAV2 or AAV2(Y-F) hF.IX vectors. At day 7 post transfer, only AAV2-hF.IX injected mice receiving capsid-specific CD8+ T cells experienced a significant elevation in ALT/AST and stained positive for infiltrating CD8+ T cells in the liver, indicating CTL mediated killing of transduced hepatocytes. This resulted in a significant reduction in circulating hF.IX (∼4-fold) and in hF.IX positive hepatocytes (2-fold) by day 28. In contrast, AAV2(Y-F) hF.IX treated mice had no elevation of ALT/AST or infiltration of CD8+ T cells. These showed only a mild reduction in circulating hF.IX levels (1.5 fold), with no change in percent hF.IX expressing hepatocytes. Therefore, AAV capsid can be engineered to markedly reduce in vivo targeting of transduced hepatocytes by capsid-specific CD8+ T cells. Use of capsid sequences that exhibit reduced MHC I presentation and increased efficacy of gene transfer should minimize immune mediated clearance of transduced hepatocytes.
Molecular Therapy Volume 20, Supplement 1, May 2012 Copyright © The American Society of Gene & Cell Therapy
556. Pharmacological Modulation of Humoral Immunity in a Non-Human Primate Model of AAV Gene Transfer for Hemophilia B
Federico Mingozzi,1 Yifeng Chen,1,2 Shangzhen Zhou,1 Samuel Murphy,1 Mark Mezger,3 Robert Donahue,3 Fraser Wright,1 Cynthia Dunbar,3 Katherine High.1,2 1 Children’s Hospital of Philadelphia, Philadelphia, PA; 2Howard Hughes Medical Institute, Philadelphia; PA, 3National Heart, Lung, and Blood Institute, Bethesda, MD. Adeno-associated virus (AAV) vector-mediated factor IX (F.IX) gene transfer to the liver for hemophilia B has shown promising results in terms of safety and efficacy in experimental animals and more recently in human subjects. However, one of the potential and most serious complications of gene- and protein-based replacement therapies for hemophilia is the development of inhibitory antibody against the therapeutic coagulation factor. This risk is low in AAV vector liver-directed gene transfer, as hepatic expression of a transgene is associated with induction of CD4+CD25+FoxP3+ regulatory T cell (Tregs)-mediated antigenic tolerance. Here we describe the development of an inhibitory antibody against the human F.IX transgene product in a rhesus macaque following the intravenous (IV) administration of 2x1013 vector genomes (vg)/kg of an AAV8-F.IX vector. Human F.IX transgene expression was lost ∼4 weeks after vector administration, concomitant with the detection of neutralizing anti-human F.IX antibodies in serum. No immune responses against the endogenous rhesus F.IX were detected, suggesting that the immune response was triggered by the non speciesspecific human F.IX transgene. Administration of a short course of immunosuppression (IS) with cyclosporine A and the anti-B cell antibody rituximab resulted in inhibitor eradication and restoration of human F.IX transgene expression in plasma. IS was associated with transient depletion of B cells in peripheral blood with no effect on the frequency of circulating Tregs. AAV vector readministration with 1x1013 vg/kg of an AAV6-F.IX vector injected IV resulted in an increase of circulating levels of human F.IX transgene form 20 to 40% of normal with no inhibitor formation. These results suggest that IS with cyclosporine A and rituximab may be effective in eradicating inhibitors that may arise in the setting of AAV-F.IX gene transfer to the liver, and does not disrupt the homeostasis of Tregs, thus allowing safe and effective AAV vector readministration.
557. Characterization of AAV T Cell Epitopes Presented by Splenocytes from Normal Human Donors
Daniel J. Hui,1 Etiena Basner-Tschakarjan,1 Yifeng Chen,1 Shyrie C. Edmonson,1 Marcela V. Maus,1 Gregory M. Podsakoff,1 Gary C. Pien,1 Rodney M. Camire,1 Federico Mingozzi,1 Katherine A. High.1,2 1 Pediatrics, Children’s Hospital of Philadelphia, Philadelphia, PA; 2 Howard Hughes Medical Institute, Philadelphia, PA. Adeno-associated viral (AAV) vectors have become increasingly promising as gene delivery vehicles for the correction of monogenic disorders. However, the potential for host immune responses remains a concern given the prevalence of AAV exposure during childhood in the general population. Evidence for pre-existing immune responses to AAV during gene transfer has been observed during clinical trials designed to correct hemophilia B via a liver-directed approach using AAV-2 (Manno et al., 2006) and AAV-8 (Nathwani et al., 2011). We hypothesized that T cell responses observed postadministration were activated upon T cell recognition of AAV capsid particles during transduction, leading to a decline of Factor IX transgene expression and the resultant transient transaminitis. To better characterize the potential for similar immune responses to AAV in future subjects undergoing gene transfer, we previously S215