Neoantigens are somatic mutations, that are unique to the tumor and can be presented on major histocompatibility complex (MHC) molecules and recognized by T cells resulting in a protective anti-tumor response. Next generation sequencing and computational methods have been successfully applied to predict neoantigens that may be presented by MHC molecules, leading to a high interest in using these non-self antigens for personalized cancer vaccination. However prioritization remains challenging since only few candidate neoantigens are immunogenic. To further enlighten the determinants driving T cell responses, we assessed immunogenicity of single amino-acid mutations in mouse tumor models.
We identified two kinds of mutations inducing a CD8 T cell response: the ones at a non-anchor residue for which the absolute binding affinity is predictive of immunogenicity, and the ones at an anchor residue for which the relative affinity (compared to the WT counterpart) is a better predictor. These results, validated on human datasets, demonstrate that incorporating these determinants of immunogenicity will help further prioritize MHC-I neoantigens for immunotherapy.
Although neoantigens were selected using an MHC-I prediction algorithm, we mostly observed mutation-induced CD4 T cell responses, as shown by others. We developed a 2D-LCMS peptide exchange assay allowing to reveal I-Ab epitopes, binding affinity and stability measurements. We then confirmed immunogenicity of identified strong I-Ab binders, and identified tetramer+ CD4 T cells from the spleen of vaccinated mice. These new experimental data and tools will help unraveling the determinants of immunogenicity and favors the selection of immunogenic MHC-II neoantigens.