American Association of Immunologists, The Journal of Immunology, 1_Supplement(200), p. 99.14-99.14, 2018
DOI: 10.4049/jimmunol.200.supp.99.14
Full text: Download
Abstract Binding of antigenic peptides to class II MHC molecules (MHCII) and the activity of the “editing” molecule HLA-DM (DM) on the resulting peptide-MHCII complexes are factors critical to the antigen presentation pathway. Different algorithms are available to predict peptide-MHC binding and identify CD4+ T-cell epitopes within pathogenic proteins. Fewer algorithms are focused on MHCII as compared to MHCI., and their reliability is also less solid. Possibly one reason for such difference lies in the greater complexity of peptide binding to MHCII, of which open-ended binding site features a larger flexibility and allows interaction with longer sequences. To validate algorithm predictions, we have compared the experimental IC50 of a library of overlapping 18-mers offset by 5 amino acids spanning the sequence of H1N1 influenza virus HA protein while interacting with HLA-DR1 with the one indicated by different prediction platforms online. The IC50 values for each peptide in the presence and absence of DM were measured using competition binding assay. The ex-vivo CD4+ T cell response to these sequences has already been published, on which basis they have been categorized into immunodominant, subdominant, weak and negative. Comparative analysis shows that measured IC50 is different from predictive IC50, with measured IC50 with DM showing immunodominant peptide as a top hit. Our data also shows that presence of DM reduced the affinity of all the peptides triggering the smaller CD4+ T cell response or negative, whereas the effect on the other peptides was less consistent. Thus, currently available algorithms are still inadequate to predict epitope for MHCII and DM is one important factor that should be considered when designing algorithms for MHCII.