Landscape and functional impact of variants of unknown significance of immune response genes in human cancer.

Immune checkpoint blockade (ICB) is a standard treatment for several types of human cancer, yet we still lack a deep understanding of the mechanisms underlying primary resistance. Tumor-intrinsic defects in immune recognition and interferon-gamma (IFNγ) signaling pathways facilitate immune evasion and may limit the efficacy of ICB. Here, we delineate the mutational landscape and functional consequences of amino acid substitutions in key immune-related genes, B2M, CALR, IFNGR1, IFNGR2, JAK1, and JAK2, across more than 12,000 primary tumors and cancer cell lines. Genomic alterations affecting the coding regions of at least one of these genes were identified in approximately 11% of cancers, with missense variants accounting for 55% of these events. B2M, encoding the invariant light chain of the heavy chain-I (HLA-I) complex, exhibited the highest mutation frequency per base pair, the mutations predominantly involving truncating variants. A curated set of 2156 missense mutations in B2M and in components of the IFNγ-signaling pathway (IFNGR1, IFNGR2, and JAK2) was analyzed using SIFT, PolyPhen-2, and AlphaMissense, yielding predicted pathogenicity rates of 52%, 35%, and 27%, respectively. The functional assays, performed in lung cancer cells, revealed JAK2 and IFNGR1 variants that impaired IFNγ-mediated transcriptional activation and growth suppression, and B2M variants that disrupted HLA class I complex formation. Notably, AlphaMissense predictions showed the highest concordance with experimental data. These findings provide a detailed mutational map of antigen presentation and IFNγ-response components in cancer. Overall, our results provide a resource of specific mutations in genes involved in immune pathways that compromise tumor immunogenicity and will serve for support in patient selection for response to ICB.