A demethylation-driven gene signature predicts prognosis and therapeutic vulnerability in hepatocellular carcinoma

去甲基化驱动的基因特征可预测肝细胞癌的预后和治疗敏感性

阅读:2

Abstract

To develop and validate a demethylation-driven gene signature for predicting HCC prognosis, immune microenvironment features, and therapeutic vulnerabilities. We integrated transcriptomic data from TCGA-LIHC (n = 346 tumors) and GEO-GSE112790 (n = 183 tumors) with a curated list of 3,743 demethylation-related genes (DRGs). Tumor-associated DRGs were identified via differential expression analysis and WGCNA. A prognostic six-gene signature was derived using univariate Cox and LASSO regression, and a risk score model was constructed by multivariate Cox analysis. Patients were stratified into high- and low-risk groups. Model performance was evaluated using Kaplan-Meier and time-dependent ROC analyses. Immune infiltration was assessed by ssGSEA, somatic mutations were profiled, and drug sensitivity was predicted via the GDSC database. G6PD expression was validated using immunohistochemistry. A six-gene prognostic signature (CEP41, SUB1, CDC20, G6PD, VPS72, SPINDOC) was established. The risk score significantly stratified patients into high- and low-risk groups with distinct overall survival (p < 0.001). The model showed strong predictive accuracy with AUCs ≥ 0.70 at 1, 3, and 5 years. High-risk patients exhibited enrichment in cell cycle, DNA repair, and metabolic pathways, along with an immunosuppressive microenvironment marked by regulatory T cells and myeloid-derived suppressor cells. Somatic mutation analysis revealed differential TP53 mutation frequencies between risk groups. Drug sensitivity prediction indicated that high-risk patients may respond better to agents such as Tozasertib and Navitoclax. IHC confirmed significant upregulation of G6PD in HCC tissues, supporting its role in metabolic reprogramming. This study establishes a robust, demethylation-driven six-gene signature that effectively stratifies HCC patients into distinct prognostic groups. The model integrates multi-omic insights into tumor biology and therapeutic vulnerability, providing a clinically actionable framework for personalized risk assessment and treatment planning in hepatocellular carcinoma.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。