The (tremendous) gap between predicted epitopes and those actually recognized by CD8 T cells

Abstract

Abstract In order to monitor CD8 cell immunity it is essential to know the peptide determinant(s) of the antigen that these T cells recognize. Relevant peptide determinants of antigens were initially identified through characterization of T cell lines or hybridomas, but more recently they are increasingly predicted based on MHC-peptide binding algorithms. Systematic, brute force testing of the actually engaged CD8 T cell repertoire ex vivo is an alternative approach, but has been challenging due to (a) the large number of candidate peptides for even simple antigens, (b) the volume of blood required for testing these peptides, and (c) the scope of effort being at the border of feasibility. To highlight the utility of ELISPOT to perform such epitope mapping, we performed a feasibility study in which PBMC from HCMV infected, HLA-A*02:01 subjects were assessed for reactivity against the 553 individual nonamer peptides that encompass the entire HCMV pp65 protein. Reactivity against a single peptide of the HCMV pp65 protein, pp65495–503, was of particular interest because numerous reports have indicated that this peptide is immunodominant in HLA-A*02:01 subjects. Strikingly, our epitope mapping showed that pp65495–503 was immunodominant in only a fraction of our HLA-A*02:01, HCMV infected test subjects. Instead, responses to pp65495–503 were either subdominant or absent in other HLA-A*02:01, HCMV infected test subjects. In the latter subjects, other peptides of pp65 elicited robust responses, providing further confirmation of HCMV status. Collectively, our findings suggest that systematic testing against all potential epitopes of the antigen of interest is both feasible and is needed for reliably monitoring of T cell immunity.