EHMT2 and INMT as methyltransferase related biomarkers predicting prognosis in HBV associated hepatocellular carcinoma.

HBV-related HCC remains a major cause of cancer mortality, and conventional staging systems do not fully reflect its molecular heterogeneity. This study evaluated the prognostic value of EHMT2 and INMT and explored the potential link between antiviral therapy and host epigenetic remodeling. Because EHMT2 and INMT showed opposite prognostic directions and represented complementary methylation-related biological processes, they were further evaluated as a biologically informed methyltransferase-related axis. TCGA transcriptomic analysis identified 26 methyltransferase-related differentially expressed genes, 13 of which were prognostically relevant by univariate Cox analysis. LASSO Cox regression further established a two-gene EHMT2-INMT risk model, which stratified patients into high- and low-risk groups with significantly different survival outcomes based on the 3-year ROC cutoff. In multivariate analysis, the risk score remained significantly associated with overall survival after adjustment for pathological stage and other clinicopathological variables, supporting its preliminary independent prognostic value. High-risk patients also exhibited lower stromal and immune scores, indicating a relatively immune-depleted tumor phenotype. Based on these findings, and because TDF is a first-line antiviral agent widely used in HBV management, we performed preliminary in vitro experiments to explore whether TDF treatment was associated with host epigenetic remodeling rather than to establish direct therapeutic causality. In HepAD38 cells, TDF suppressed HBV replication, downregulated EHMT2, upregulated INMT, and reduced global DNA methylation, including INMT gene-body hypomethylation. These findings suggest that the EHMT2-INMT axis may provide a methyltransferase-related prognostic framework in HBV-related HCC and that TDF may influence host epigenetic regulation beyond viral suppression, although causality requires functional validation.