The major histocompatibility complex (MHC) class I and class II molecules present peptide antigens to T cells to enable immune detection of foreign or mutated peptides. Mutant peptides presented by MHC-I on the cell surface have the potential to activate a cytotoxic CD8+ T-cell response, imposing selective pressure on tumor cells harboring neoantigens. Recently, we demonstrated that patient MHC-I genotype results in individual differences in which cancer-causing mutations can be effectively presented on the cell surface to stimulate CD8+ T cells and is a determinant of the probability of observing a particular mutation (Marty et al., Cell 2017). However, since cytotoxic CD8+ T cell responses require CD4+ T cells that are activated by MHC-II-based antigen presentation, we hypothesized that individual variation in MHC-II genotype must also be an important determinant of immune selection of tumor cells. To investigate the role of MHC-II restriction in shaping the mutational landscape of tumors, we first developed a score that represents the ability of an individual MHC-II genotype to present a mutation on the cell surface. The score maps affinities of mutated peptides for different MHC-II molecules into a single value and was able to distinguish peptides found in complex with MHC-II in vitro by mass spectrometry from random peptides with an AUC of 0.69. Next, we assessed the relationship between MHC-II presentation of 1,018 cancer mutations and their occurence in 7,137 tumors across 30 tissues types from The Cancer Genome Atlas (TCGA). Across all MHC-II genotypes in TCGA, we observed an unexpected bias for recurrent cancer mutations to be less well presented than random mutations (p < 2.2e-16), while foreign pathogenic sequences were more likely to be presented (p < 2.2e-16). Moreover, we found that higher mutation frequency among tumors correlated with overall poorer presentation across cancer patients. Comparing mutation presentation by MHC-II with our previous results for MHC-I, we found only a modest correlation in presentation scores across mutations (rho=0.28). Combining MHC-I and MHC-II genotype-derived presentation scores into a single model increased overall performance for predicting occurrence of mutation. Notably, MHC-II genotype had an even stronger influence over the probability for specific oncogenic mutations to arise than MHC-I genotype (OR 2.12 vs. OR 1.43). We also observed differences in the extent to which MHC-based presentation could predict driver mutation status in different tumor types, with the best performance observed in thyroid cancer (AUC=0.91). These results reveal that immunoediting during tumor development is dependent on MHC-II as well as MHC-I genotype, further implicating the inherited immune system as a key risk factor for cancer.

Citation Format: Rachel Marty, Wesley Thompson, Trey Ideker, Maurizio Zanetti, Hannah Carter. MHC II complements MHC I in shaping the mutational landscape of tumors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 5727.