A five-gene prognostic signature integrating endothelial senescence and lipid metabolism for risk stratification in hepatocellular carcinoma

Background Hepatocellular carcinoma (HCC) is closely associated with endothelial cell senescence and lipid metabolism (EL), yet shared mechanisms remain unclear. Hence, elucidating the mechanisms linking HCC to EL-related genes (ELRGs) is crucial for timely diagnosis and effective management of HCC. Methods Sequencing datasets related to HCC were obtained from publicly available databases. By intersecting differentially expressed genes (DEGs), endothelial cell senescence-related genes (ESRGs), and lipid metabolism-related genes (LMRGs), candidate genes were obtained. Prognostic genes were obtained through univariate Cox regression analysis and machine learning. Next, HCC samples were divided into two risk groups. Subsequent analyses focused on functional enrichment and others. Single-cell RNA sequencing (scRNA-seq) analysis was further performed to identify the primary cellular sources of the prognostic genes, characterize their expression dynamics along developmental trajectories, and reconstruct intercellular communication networks across distinct cell types. Lastly, the expression of prognostic genes in tumor tissue samples was verified by reverse transcription quantitative polymerase chain reaction (RT-qPCR). Results A total of 52 candidate genes were identified. Subsequently, five prognostic genes were identified and used to establish a prognostic model. Analysis illustrated that two risk groups were co-enriched in pathways related to metabolic pathways. The scRNA-seq analysis demonstrated hepatocyte-predominant expression of all five prognostic genes, with tumor-specific upregulation and stage-specific pseudotime dynamics. Cell-cell communication analysis further identified the NAMPT–(ITGA5 + ITGB1) signaling axis as a key mediator of intercellular crosstalk. RT-qPCR analysis showed significant upregulation of FABP5, FABP6 and UGT1A10, but downregulation of MMP1 and SGPP2 in HCC samples. Conclusions A five-gene prognostic signature (FABP5, FABP6, MMP1, SGPP2, UGT1A10) was established and validated. Among these, FABP5 showed the strongest correlation with activated CD4⁺ T-cell infiltration (cor = 0.52, P < 0.05) and was predominantly expressed in hepatocytes with stage-specific upregulation along pseudotime trajectories. Single-cell analysis further revealed NAMPT–(ITGA5 + ITGB1) as a key intercellular signaling axis. These findings suggest that FABP5 may serve as a link between endothelial senescence, lipid metabolism, and immune microenvironment remodeling in HCC, warranting further functional investigation.