A Novel Splenocyte Approach for Characterizing T Cell Responses to Adeno-Associated Virus in the Normal Population: Implications on Gene Transfer.

Gene transfer using adeno-associated viral (AAV) vectors offers a promising strategy for the treatment of hemophilia B. However, in the first clinical trial using AAV-2 to deliver Factor IX (F.IX) into the liver via the hepatic artery, transgene expression was short-lived, followed by a gradual decline in F.IX levels and a transient rise in liver enzymes (Manno et al., 2006). We hypothesize this outcome was caused by a memory T cell response to a previous infection with wild-type AAV-2, which led to T cell-mediated destruction of transduced hepatocytes upon recognition of capsid sequences during gene transfer. No immune response to F.IX was detected. In order to further investigate the role that previous exposure to AAV-2 may have on limiting the duration of gene transfer, a normal donor population consisting of 46 adult subjects was tested for AAV-2 neutralizing antibodies and frequencies of circulating capsid-specific T cells by ELISpot assay. More than half (25/46) of the subjects had readily detectable titers of neutralizing antibodies to AAV-2 in serum. However, the prevalence of AAV-2 capsid-specific T cells detected by IFN-γ ELISpot in PBMCs was very low and positive responses could be confirmed in only 2/46 subjects. Attempts to expand the relevant T cell population in vitro, in which PBMCs were stimulated with the relevant AAV-2 peptide epitope, did not increase the frequency of detectable T cell responses. These results suggest that current methods may not be sensitive enough to fully appreciate the prevalence of AAV-2-specific CD8+ memory T cells in humans because they circulate at frequencies that are commonly too low for detection, as indicated by the discrepancy between the presence of AAV neutralizing antibodies and the observed T cell response. Moreover, it is possible that AAV-specific CD8+ T cells fail to circulate but rather reside in lymphatic compartments. To address this problem, we developed an alternative approach using human splenocytes. Splenocytes offer two critical advantages: they permit sampling of a primary lymphoid organ, and they can be obtained in much higher overall cell numbers than PBMCs. Using splenocytes obtained from both normal donor adults and children undergoing splenectomy, T cell responses via ELISpot assay have been detected in 7 out of 27 subjects, reflecting a higher frequency (26%) of observed T cell responses than that seen using PBMCs. In addition, intracellular cytokine staining analysis for IFN-γ confirmed the presence of AAV capsid-specific CD8+ T cells in the samples tested. Interestingly, in the current study T cell responses were seen in donors as young as 5 years of age, while more robust responses were seen in adult patients >37 years of age. These results have important implications for gene transfer with AAV; specifically, that children as young as 5 years old have detectable T cell responses to AAV capsid and could thus mount a memory T cell response on vector infusion. We conclude that memory T-cell-mediated immune responses to viral capsids in the normal population is more common than PBMC-based assays suggest, and given the observed cross-reactivity among AAV serotypes (see abstract by Mingozzi et al.), these results may affect not only AAV-2 clinical trials, but other AAV clinical trials as well.