Emerging evidence has revealed the pivotal role of epigenetic modifications in shaping the tumor microenvironment (TME). However, crosstalk between different modification types and their clinical relevance in cancers remain largely unexplored. In this study, using ChIP/MeRIP-seq data of seven human gastric cell lines, we systematically characterized the crosstalk of four epigenetic modification types including H3K4me1, H3K4me3, H3K27ac, and N6-methyladenosine (m6A) and identified a recurrent subtype with high FTO expression and low HDAC1 expression across three independent gastric cancer (GC) cohorts, which we named the epigenetic-modification-dysregulated (EMD) subtype. Patients of the EMD subtype were featured with poor survival, stromal activation, and immune suppression. Extensive relevance to clinical characteristics was observed in the EMD subtype, including the Lauren classification, MSI status, histological grade, TNM stage, the Asian Cancer Research Group classification, and the immune/fibrotic classification. An EMD score was then constructed using WGCNA and ssGSEA algorithms, to precisely recognize the EMD subtype and indicate prognosis and response to immunotherapy in multiple independent GC cohorts. Correlations of the EMD score with tumor mutation burden, tumor purity, aneuploidy score, tumorigenic pathways, TME characteristics, and FTO/HDAC1 ratio were measured. In vitro experiments were performed to demonstrate the correlation between FTO and the epithelial-mesenchymal transition pathway, which suggested FTO as a targetable vulnerability for GC patients with a high EMD score. Altogether, by comprehensively analyzing the epigenetic modification patterns of 1518 GC patients, we identified a novel stromal-activated subtype with poor survival and resistance to immunotherapy, which might benefit from the combined immune checkpoint inhibition therapy with FTO inhibition.
Keywords: gastric cancer; histone modification; immunotherapy; m6A; prognosis; tumor microenvironment.
Copyright © 2022 Yuan, Zhang, Deng, Xia, Li, Meng, Jin, Cheng, Li, Zhang and He.