The impact of de novo lipogenesis on predicting survival and clinical therapy: an exploration based on a multigene prognostic model in hepatocellular carcinoma.

BACKGROUND: Hepatocellular carcinoma (HCC) ranks among the most aggressive malignancies worldwide, with poor outcomes attributed to delayed diagnosis and therapeutic limitations. Emerging evidence suggests that de novo lipogenesis (DNL) plays a crucial role in HCC progression and its interaction with the immune microenvironment. METHODS: We systematically analyzed DNL-related gene expression profiles from TCGA, GEO, ICGC-LIRI datasets, and our Xiangya HCC cohort (n = 106) to construct a prognostic risk model. Through LASSO-Cox regression analysis, we identified six signature genes (G6PD, LCAT, SERPINE1, SOAT2, CYP2C9, and UGT1A10) that effectively stratified patients into distinct risk groups. We evaluated clinical characteristics, immune cell infiltration patterns, and differential therapeutic responses between high-risk and low-risk groups. Comprehensive validation included immunohistochemical analysis and Western blotting to assess expression levels of key model genes, along with multiplex immunofluorescence staining and single-cell RNA sequencing(scRNA-seq) to characterize immune microenvironmental differences between risk groups. RESULTS: We successfully established a robust six-gene prognostic signature (G6PD, LCAT, SERPINE1, SOAT2, CYP2C9, and UGT1A10) based on de novo lipogenesis pathways, which demonstrated excellent predictive performance (AUC: 0.78-0.82). The model revealed significant differences in immune infiltration patterns between risk groups, with the high-risk group exhibiting immunosuppressive characteristics characterized by increased Treg cell infiltration, while the low-risk group showed greater NK cell retention. Integrated scRNA-seq and our cohort validation further demonstrated that high-risk scores were associated with poorer response to immunotherapy but greater sensitivity to targeted therapies. These findings suggest that de novo lipogenesis-mediated immune evasion contributes to therapy resistance and worse prognosis in high-risk HCC patients, whereas low-risk HCC patients maintain an immunologically active microenvironment more amenable to immunotherapy. CONCLUSIONS: This study provided a novel prognostic model for HCC, incorporating 6 representative DNLs. The model demonstrated the potential for predicting HCC prognosis and highlighted the involvement of immune cell infiltration and the association between risk scores and clinical therapy. Validation of model genes further supported the association between de novo lipogenesis and HCC development.