Abstract
Hepatocellular carcinoma (HCC) is a cancer caused by abnormal cell growth due to faulty signal transduction. Cells secrete tumor suppressor factors in response to potential carcinogenic signals, inducing cellular senescence (CS) as a countermeasure. However, accurately measuring CS levels in different types of tumors is challenging due to tumor heterogeneity and the lack of universal and specific CS markers. Machine learning has revealed unique molecular traits in HCC patients, leading to clinical advantages. More research is needed to understand senescence-related molecular features in these patients. In this study, the gene expression profile features of patients with HCC were analyzed by integrating single-cell RNA sequencing and bulk RNA-seq datasets from HCC samples. The analysis identified the senescence-related pathways exhibiting HCC specificity. Subsequently, genes from these pathways were used to identify senescence-related molecular subtypes in HCC, showing significant variations in biological and clinical attributes. An HCC-specific CS risk model developed in this study revealed substantial associations between the patients' CS scores and prognosis grouping, clinical staging, immune infiltration levels, immunotherapy response, and drug sensitivity levels. Within the constructed model, G6PD was identified as a key gene, potentially serving as a senescence-related target in liver cancer. Molecular biology experiments demonstrated that overexpression of G6PD effectively promotes the proliferative, invasive, and migration capacities of HepG2 and SK-HEP-1 cells. In conclusion, this analysis offers a valuable framework for understanding senescence in HCC and introduces a new biomarker. These findings improve our understanding of senescence in HCC and have potential for future research.