Identification of the AAV2 Capsid CD8+ T Cell Epitope in C57BL/6 Mice.
Abstract Recently, a clinical trial for adeno-associated virus serotype 2 (AAV2) mediated liver directed gene transfer of human Factor IX to subjects with severe hemophilia B revealed that two patients developed transient asymptomatic transaminitis following vector administration. Immunology studies in the second patient demonstrated a transient T cell response to AAV2 capsid peptides suggesting that the immune response to the AAV capsid may be related to the transient transaminitis. We hypothesized that the observations made in the human subjects were due to a CD8 T cell response to AAV2 capsid protein. Preclinical studies in mice and dogs, which are not naturally infected by wild type AAV2 viruses, did not predict these findings in the clinical study. Thus, we developed a mouse model in which we were able to mimic this phenomenon (Blood 102:493a). In an effort to further characterize the immune responses to AAV2 capsid proteins in this mouse model, we identified the T cell epitope in the AAV capsid protein recognized by murine C57Bl/6 CD8 T cells. A peptide library of AAV2 VP1 capsid peptides (n=145) that were synthesized as 15mers overlapping by 10 amino acids were divided into 6 pools each containing 24–25 peptides. C57Bl/6 mice were immunized intramuscularly with an adenovirus expressing AAV2 capsid protein. Nine days later the spleen was harvested and intracellular cytokine staining (ICS) was used to assess release of IFN-γ from CD8 T cells in response to 6 AAV2 capsid peptide pools. ICS demonstrated CD8 cells from mice immunized with Ad-AAV2 produced IFN-γ (3.5% of the CD8 cells) in response to Pool F (amino acid 119–145) while no IFN-γ release in CD8 cells was detected with Pool A to E (mean 0.28%±0.25%) compared to the media control (0.16%). This detection of IFN-γ release from CD8 T cells indicates a specific proliferation to a peptide(s) within this peptide pool (Pool F). A matrix approach was used to further define which peptide(s) contained the immunodominant epitope. Eleven small peptide pools of Pool F were created in which each peptide was represented in 2 pools. ICS of splenocytes from immunized (Ad-AAV2 capsid) C57Bl/6 mice demonstrated IFN-γ response from CD8 cells to 3 of the matrix pools corresponding to peptide 140 (PEIQYTSNYNKSVNV) and 141 (TSNYNKSVNVDFTVD) compared with media controls. To determine the exact peptide sequence that binds to the MHC Class I molecule, 9 amino acid peptides (n=7) were created that overlap peptide 140 and 141. Peptide SNYNKSVNV showed positive staining for both CD8 and IFN- γ(3.2%) compared with the six other peptides (0.14%±0.08%), media control (0.08%) and mice that were not immunized (0.11%). This epitope lies in the C terminus of the AAV2 VP1 capsid protein. Current studies using strains of mice with different MHC H2 haplotypes will allow us to determine which of the C57Bl/6 MHC alleles the epitope binds. These findings will provide us with a powerful tool for assessing immune responses to AAV capsid in the context of gene therapy. Specifically, they will allow us to determine how long immunologically detectable capsid sequences persist in an animal injected with AAV vectors. This in turn will provide a basis for a clinical study in which subjects are transiently immunosuppressed, from the time of vector injection until capsid epitopes are no longer detectable by the immune system.