Chronic alcohol consumption is an established risk factor for esophageal squamous cell carcinoma (ESCC). Elucidating the underlying mechanisms linking metabolic stress to esophageal carcinogenesis could help identify prevention and treatment strategies. Through a trans-ancestry meta-analysis encompassing more than 200,000 individuals from East Asia, we identified a germline variant (rs1051921 C>T) at the MLXIPL locus (encoding ChREBP) that markedly elevated ESCC risk among alcohol drinkers. Functionally, the risk allele facilitated N6-methyladenosine (m6A) modification of ChREBP transcripts, enhancing their stability through recognition by the m6A reader YTHDF1. Alcohol further potentiated this regulatory axis by promoting ACSS2-driven acetyl-CoA synthesis and increasing H3K27ac enrichment at promoters of key m6A regulators. Increased ChREBP expression activated the transcription factor ATF3, which triggered endoplasmic reticulum stress and epithelial-mesenchymal transition, thereby conferring anoikis resistance. In parallel, ATF3 promoted an immunosuppressive tumor microenvironment through upregulation of PD-L1 and VEGFA, leading to exclusion of CD8+ T cells and expansion of granulocytic myeloid-derived suppressor cells. In vivo intervention via ATF3 knockdown, anti-PD-L1 treatment, or pharmacological perturbation of ChREBP-related metabolism with metformin or an ACSS2 inhibitor significantly suppressed ESCC progression and restored intratumoral CD8+ T-cell infiltration. These findings establish ChREBP as an alcohol-sensitive metabolic-epitranscriptomic switch that integrates genetic susceptibility with immune evasion to drive esophageal carcinogenesis.
Figure S1. Fine-mapping of the ESCC risk locus rs1051921 and single-cell analysis of ChREBP expression dynamics in tumor progression trajectories and invasive epithelial subclusters.
Supplementary Figure S4 shows identification and functional analysis of regulatory variants in alcohol-responsive genes. (A) The flowchart of identifying the regulatory variants of alcohol-responsive genes. (B) A significant correlation was found between different genotypes of rs10223516 and the expression levels of FOXP4 in 497 esophageal mucosal tissue samples from the GTEx database. (C) The construction of the pGL3-promoter and pGL3-basic plasmids for the dual luciferase reporter gene assay experiments.
Supplementary Figure S5 shows FOXP4 transcriptionally upregulates CYP26B1 and MYC. (A) The RNA level of FOXP4 in FOXP4-overexpressing KYSE30 and FOXP4-overexpressing KYSE150 cells. (B-C) The protein level of FOXP4 in ESCC cells with FOXP4 overexpression or knockdown. (C) The pathway enrichment analysis of genes regulated by FOXP4 in ESCC cells. (D) The expression of MYC was correlated with the expression of FOXP4 in esophageal mucosal tissues from the GTEx database. Shade represented 95% confidence intervals. (E) The H3K27ac CUT&Tag qPCR experiment results of FOXP4, CYP26B1, and MYC in alcohol-treated KYSE30 and KYSE150 cells. (F) The protein expression levels of FOXP4, CYP26B1 and MYC were higher in alcohol-treated KYSE30 and KYSE150 cells than the control.