Document resource
Background Cancer immunoediting critically influences tumorigenesis, progression, and therapeutic resistance, highlighting involvement of the immune system in shaping the tumor mutational landscape. 1 Human Leukocyte Antigens (HLA) are proposed to be central to this process, as they present cancer-derived neoantigens that elicit anti-tumor immunity. While earlier studies reported evidence of positive selection for neoantigens with low HLA presentation scores,2 3 a recent reanalysis4 of the same pan-cancer dataset from the Cancer Genome Atlas (TCGA) challenged these findings, attributing them to statistical artifacts. To resolve this controversy, we reassessed the relationship between germline HLA genotypes and somatic cancer driver genes using the independent, well-powered Genomics England (GEL) cancer cohort.Methods The availability of germline whole-genome sequencing (WGS) data in GEL enabled high-quality and high-resolution HLA typing, with likely greater accuracy than that used by the TCGA (based on genotyping arrays and whole-exome sequencing data). Our analyses focused on European-ancestry patients, which comprised most of the study sample. Rather than relying on generic canonical cancer driver genes, we analyzed patient-specific driver mutations identified within GEL. 5 To overcome statistical limitations of prior methods that used Patient Harmonic-mean Best Rank (PHBR) scores to estimate HLA-neoantigen binding affinity, we directly tested associations between germline HLA Class I/II and presence of driver genes.Results Using multivariable logistic regression assuming additive genetic effects, we analyzed associations across the four most common cancers in GEL (N=4,932): colorectal (N=1,965), breast (N= 1,610), kidney (N=764), and lung (N= 593). Models were adjusted for age, sex, and the top five ancestry principal components. No associations reached significance after stringent Bonferroni correction (thresholds ranging from 7.2E-06–2.6E-05). The lowest p-values observed included particular HLA alleles with KRAS gene (p=8.0E-04; colorectal cancer), SPEN (p=3.E-04; breast cancer), ARID1A (p=2.2E-04; kidney cancer), and TRIM51 (p=2.3E-04; lung cancer). A pooled analysis adjusting for cancer type similarly yielded no significant associations.Conclusions Our findings corroborate the recent report 4 that challenged evidence of a strong role of germline HLA genotypes in shaping somatic driver mutation landscapes. By circumventing PHBR-based limitations and leveraging high-quality WGS-derived HLA typing with cancer subtype-stratified analyses in a large, homogeneous cohort, we provide robust evidence against significant HLA-driven contribution to cancer driver gene selection. However, small-effect associations may exist, warranting further investigation in larger cohorts. These results refine current models of cancer immunoediting and have important implications for next-generation immunotherapy target discovery, as well as personalized cancer screening and treatment strategies.References Dunn GP, Bruce AT, Ikeda H, Old LJ, Schreiber RD. Cancer immunoediting: from immunosurveillance to tumor escape. Nat Immunol. 2002;3:991–8.Marty R, Kaabinejadian S, Rossell D, Slifker MJ, Haar J van de, Engin HB, et al. MHC-I genotype restricts the oncogenic mutational landscape. Cell. 2017;171:1272–1283.e15.Pyke RM, Thompson WK, Salem RM, Font-Burgada J, Zanetti M, Carter H. Evolutionary pressure against MHC class II binding cancer mutations. Cell. 2018;175:416–428.e13.Kherreh N, Cleary S, Seoighe C. No evidence that HLA genotype influences the driver mutations that occur in cancer patients. Cancer Immunol, Immunother. 2022;71:819–27.Kinnersley B, Sud A, Everall A, Cornish AJ, Chubb D, Culliford R, et al. Analysis of 10,478 cancer genomes identifies candidate driver genes and opportunities for precision oncology. Nat Genet. 2024;1–10.Ethics Approval Genomics England has obtained written informed consent from all participants.