523. Deep Sequencing of T Cell Receptor in Peripheral Blood and Muscle from Adeno-Associated Virus Vector-Injected Subjects Reveals Differences in T Cell Clonality Between the Two Compartments

Abstract

Long-term correction of the disease phenotype following gene therapy in vivo with adeno-associated virus (AAV) vectors has been demonstrated in several preclinical and clinical studies. With clinical development, however, it has become obvious that the host immune system represents an important obstacle to safe and effective gene transfer with AAV vectors. Indeed T cell responses directed against the viral capsid can have a detrimental effect on the duration of transgene expression and can result in tissue-specific toxicity due to the clearance of vector infected cells. ELISpot and flow cytometry analysis of immune responses to the vector capsid in peripheral blood mononuclear cells (PBMCs) show that both CD4+ and CD8+ T cells reactive to the AAV capsid are detectable in peripheral blood following gene transfer in humans, however little is known about T cell clonality following gene transfer in PBMC vs. locally in the tissue in which the vector is delivered.Here we used TCRbeta deep sequencing of T cells infiltrating muscle injected with AAV vectors and correlated results with a time course of sequencing of TCRbeta performed in fractionated CD8+ and CD8- PBMC from subjects injected with an AAV1 vector expressing γ-sarcoglycan (AAV1-γ-sarco). PBMC and muscle samples from three subjects who received an intramuscular injection of 4.5×1010 vector genomes of AAV1-γ-sarco (Brain 2012; 135:483) were analyzed. In one of these subjects, AAV capsid IFN-γ ELISpot on PBMCs showed activation of T cells responses peaking at day 30 post gene delivery. Consistent with this finding, TCRbeta sequencing of sorted CD8+ T cells showed the de novo expansion of T cell clones in PBMC after gene transfer, which were detectable at day 30 and 180 but not at baseline before vector administration. Additional CD8+ T cell clones present at baselines also expanded between day 0 and day 180. Interestingly, CD8+ T cell clones that expanded in PBMC following gene transfer were not found in biopsies of injected muscle collected at day 30. Conversely, several T cell clones shared between muscle and more in PBMC CD8+ than CD8- population did not expand upon AAV vector injection, possibly representing the underlying inflammation of muscle characteristic of the disease. Finally, we found T cell clones over-represented in muscle following AAV injection compared to PBMC, suggesting specific differences in homing of T cells in different body tissues. For example in one subject, we observed a clear expansion of T cell clones in muscle at day 30 post vector administration compared to baseline; these clones were not detected in the CD8+ T cell compartment in peripheral blood.In conclusion, this work addresses the crucial issue of correlation of T cell responses to AAV in peripheral blood vs. the transduced tissue. Our results suggest that T cell reactivity in peripheral blood does not necessarily represent the state of inflammation in the injected tissue, at the site of antigen presentation.