Immuno-epidemiologic mapping of HLA diversity across glioma patient cohorts

Abstract Background Individual-level characteristics underlying population-level variation in glioma risk and outcomes remain incompletely understood. Cancer immunosurveillance, host immunity, and some immunotherapies center on the ability of an individual’s immune cells to recognize antigen epitopes presented on MHC molecules. Inter-individual variation in HLA alleles can elicit distinct repertoires of tumor antigen for presentation to immune cells. Therefore, HLA alleles may impact glioma incidence and prognosis. Methods HLA class I (HLA-I) alleles were identified using sequencing data from four large glioma cohorts and healthy cohorts, matched on ancestry, and race- and age-matched imputed cohorts developed by the Hardy-Weinberg equilibrium were referred to determine odds ratio incidence estimated by logistic regression. HLA prognostication was quantified by Cox regression. Results We analyzed 1,215 cases of glioma patients from non-Hispanic Whites and Asians. The HLA-I allelic frequencies of gliomas generally corresponded to their distribution within each race. However, specific HLA-I alleles were significantly associated with glioma incidence and prognosis, which differ between races but were independent of age and sex. Notably, non-Hispanic White glioma patients exhibited greater HLA homozygosity rates compared with race-matched controls. HLA-C01:02 and HLA-C07:02 displayed opposing effects on glioma prognosis between races. The distinct effects were associated with their capability of presenting specific mutations that appeared at the initial or late phase of glioma progression. Conclusions Expression of specific HLA-I alleles are associated with glioma incidence and prognosis within race. HLA-I-homozygosity is a risk factor for glioma in non-Hispanic Whites. These findings may guide development of precision-guided immunotherapies for glioma.