A detailed comparison of peptides presented by different HLA class I loci: an in silico approach

Abstract

Human Leukocyte Antigen (HLA) class I is a group of genes located on human chromosome 6 which play a crucial role in initiating potentially protective immune responses, by presenting pathogen-derived peptides to CD8+ T cells and thus targeting infected cells for elimination. Compare to other HLA class I loci, HLA-B molecules show some specific characteristics: 1) HLA-B alleles have been associated with opposing disease outcomes, namely susceptibility to, or protection from, parasitic and viral infection. 2) HLA-B elicits more dominant T cell responses with higher frequencies or superior magnitude. In order to discover potential mechanisms which may explain why HLA-B restricted T cell responses tend to be immunodominant, the focus of my PHD project has been on an extensive comparison of HLA-A and HLA-B antigen presentation by analysing their ligands. Besides the experimental verified epitopes downloaded, filtered and analyzed from several public databases, I also selected and applied in silico predictions based on different research questions for main steps of antigen presentation pathway : MHC binding (SMM, NetMHC, NetMHCpan), TAP binding probability (SMM) and proteasome cleavage (SMM or NetChop), as a compensatory of potential bias made by experimental data. Since the differences in CD8+ T cell responses mediated via HLA-A and HLA-B molecules might already begin at the antigen presentation level, in chapter 2 we have analyzed the epitope diversity and epitope binding affinity of this two major HLA class I loci. Opposite to our expectation, the amount of epitopes presented by HLA-A molecules is more than that of their HLA-B counterparts with higher binding affinities. Alternatively, targeting more constrained regions of a pathogen genome might be favourable for the outcome of an infection. Thus in chapter 3 we compared CTL epitopes restricted by HLA-A and HLA-B molecules and found that HLA-B molecules prefer to present more epitopes from certain conserved proteins. Moreover the residues targeted by HLA-B alleles in HIV-proteome were significantly more conserved than the ones targeted by HLA-A alleles. In order to explore if similar mechanism occurs in other infectious disease, in chapter 4 a similar analysis was applied on HCV proteome. As a result, we found that HLA molecules associated with HCV clearance preferentially present epitopes from the conserved proteins, like NS5B and Core. In addition, we found that less promiscuous peptide presentation might contribute to viral control during HIV-1 infection (chapter 5). To test whether the HLA promiscuity across loci has an effect on the frequencies of HLA class I haplotypes, we estimated the ligand repertoires for all HLA class I haplotypes (HLA-A-B) by in silico prediction and found that the most common HLA-A-B haplotypes are enriched in HLA-A and HLA–B pairs with distinct peptide binding motifs (chapter 6). In summary, this research implies that factors other than antigen presentation might be more essential for the protective and immunodominant nature of HLA-B-restricted T cell responses. It will also give mre support information for identifying interesting epitopes which could be good candidates for vaccine design.