Multi-omics analysis reveals microbiome-associated subtypes of esophageal cancer with distinct immune profiles and therapeutic responses.

[BACKGROUND] Esophageal cancer is an aggressive malignancy with poor prognosis. Comprehensive molecular characterization is crucial for identifying novel therapeutic targets and improving outcomes.

[METHODS] We performed an integrative multi-omics analysis of esophageal tumors, profiling the microbiome, transcriptome, epigenome, and clinical data. Unsupervised consensus clustering using ten different algorithms identified robust molecular subtypes, which were further characterized by functional enrichment analyses and experimental validation to uncover subtype-specific features and therapeutic vulnerabilities.

[RESULTS] Microbiome profiling revealed distinct taxonomic differences between normal and tumor tissues, with Proteobacteria and Pseudomonas being more abundant in tumors. Two molecular subtypes, C1 (n = 92) and C2 (n = 57), were identified, exhibiting unique clinical, prognostic, and molecular features. Subtype C2 showed higher mutation burden, increased abundance of Pseudomonas species, and an immunologically active microenvironment, with greater infiltration of effector memory CD4 + T cells, regulatory T cells, and M2 macrophages. Subtype C2 also exhibited higher predicted immunotherapy response rates and lower T cell exclusion scores. In contrast, subtype C1 was characterized by elevated expression of specific RNAs and DNA methylation, as well as higher sensitivity to chemotherapy agents and EGFR inhibitors, which was experimentally validated using representative cell lines. Notably, STAT3 inhibitors potentiated the efficacy of PD-L1 blockade in subtype C2 cells by enhancing apoptosis, downregulating PD-L1 expression, suppressing EMT, and attenuating invasion and migration.

[CONCLUSION] This multi-omics analysis reveals esophageal cancer's molecular heterogeneity and identifies microbiome and immune signatures associated with prognosis and therapeutic response, providing a framework for developing personalized treatments targeting specific subtypes.