Discordance between the predicted vs. the actually recognized CD8+ T cell epitopes of HCMV pp65 antigen and aleatory epitope dominance

Abstract

CD8+ T cell immune monitoring aims at measuring the numbers and functions of antigen-specific CD8+ T cell populations engaged during immune responses, providing insights into the magnitude and quality of cell-mediated immunity operational in a test subject. The selection of peptides for ex vivo CD8+ T cell detection is critical, however, because for each restricting HLA class I molecule present in a human individual there is a multitude of potential epitopes within complex antigens, and HLA diversity between the test subjects predisposes CD8+ T cell responses to individualized epitope recognition profiles. We report here on a brute force CD8+ T cell epitope mapping approach for the human cytomegalovirus (HCMV) pp65 antigen on ten HLA-A*02:01-matched HCMV infected human subjects. In this approach, in each test subject, every possible CD8+ T cell epitope was systematically tested; that is 553 individual peptides that walk the sequence of the HCMV pp65 protein in steps of single amino acids. Highly individualized CD8+ T cell response profiles with aleatory epitope recognition patterns were observed. We compared the actually detected epitope utilization in each individual with epitope prediction ranking for the shared HLA-A*02:01 allele, and for additional HLA class I alleles expressed by each individual. No correlation was found between epitopes ranking on the prediction scale and their actual immune dominance. The data suggest that accurate CD8+ T cell immune monitoring might depend on the agnostic reliance on mega peptide pools, or brute force mapping, rather than individualized epitope predictions.