Genomic and transcriptomic analyses of melanoma in Japanese patients reveal candidate biomarkers for immune checkpoint inhibitor responders.

Immune checkpoint inhibitors (ICIs) have greatly improved advanced melanoma prognosis. However, the efficacy of ICIs in Japanese patients has been found to be lower than that in their white counterparts. We aimed to elucidate the genomic and transcriptomic features associated with response to ICIs in Japanese patients with melanoma. A total of 129 tumor samples were collected from 78 patients with melanoma who received therapeutic regimens with or without ICI treatment. We performed exome and RNA sequencing and investigated the association between genomic and transcriptomic factors and the clinical efficacy of ICI. The number of somatic SNVs in Japanese patients with melanoma is lower than that in the TCGA white data owing to the biased distribution of WHO subtypes. The driver subtypes BRAF, NRAS, and NF1 are less prevalent, but the triple wildtype predominantly exists in this cohort. An exome-wide survey reveals no significant association of mutated genes with ICI response; however, transcriptomic analysis reveals inflammation-associated genes, including several chemokines and cytokines, that are highly expressed in clinically benefited patients. Follicular helper T cells, measured by immune cell composition analysis, are significantly enriched in clinically benefited patients (p = 0.0373). Through time-course transcriptome analysis, in addition to several cytotoxic T-cell genes, MARCO on tumor-associated macrophages is found to be induced by ICI treatment in clinically benefited patients (p = 0.0040). Protein expression of these genes is confirmed by immunohistochemical and multiplex immunofluorescence analyses. To our knowledge, this is the first and largest genomic cohort study in Japanese patients with melanoma in which tumor samples were prospectively analyzed. Genomic and transcriptomic analyses reveal candidate biomarkers for ICI in Japan.