Insignificant effects of loss of heterozygosity in HLA in the efficacy of immune checkpoint blockade treatment

BACKGROUND: It is assumed that loss of heterozygosity and allelic copy loss in HLA gene is associated with poor response rates in immune checkpoint inhibitor treatment. H-owever, the accurate extents or consistency in cancer types have not been explored. OBJECTIVE: The goal of this study is to investigate quantitative relationship between HLA allelic copy loss and response rates to immune checkpoint inhibitors. Also, tumor microenvironment was computationally assessed in the tumors with HLA copy loss to provide potential mechanisms for the relationships. METHOD: A total of 282 whole exome sequencing data from three cohorts of patients who received immune checkpoint blockade immunotherapy were analyzed, including Anti-PDL1 treated in metastatic urothelial cancer (N = 216), anti-PD1 treated metastatic melanoma (N = 26), and anti-CTLA4 treated metastatic melanoma (N = 39). The LOHHLA algorithm was used to calculate allelic copy number loss at each HLA-A, -B, and -C locus, and further determine HLA allelic copy loss status. The HLA copy status and ICB response rates were analyzed for association using Fisher’s exact test. The CIBERSORT-absolute algorithm was then used to analyze the patient's immune environment, which represented loss of heterozygosity, using paired matched RNA sequencing data. RESULTS: Unlike the general expectation, HLA allelic copy loss was not significantly associated with the ICB responses. Moreover, the relationship showed a reversed relationship in HLA-A in the urothelial cancer (better ICB response in HLA copy loss). Regardless of the HLA copy status, the proportion of cytotoxic immune cells in the immune environment of patients was correlated with ICB response, which was higher in the loss of heterozygosity group in the urothelial cohort. CONCLUSION: Although the loss of heterozygosity in HLA was generally expected to be an inhibitory factor in the immune treatment response by causing T cell immune evasion, our analysis demonstrates no explicit relationships. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s13258-021-01207-8.