Polyamine-related 4-gene prognostic signature in hepatocellular carcinoma

This study aimed to explore the biological and clinical roles of polyamine-related genes (PRGs) in hepatocellular carcinoma (HCC) and to construct a PRG-based prognostic signature for prognostic assessment and precision therapy guidance in HCC. Twenty-two PRGs were retrieved from the Molecular Signatures Database. Transcriptome, clinical, and somatic mutation data of HCC were integrated from The Cancer Genome Atlas, International Cancer Genome Collaboratory, National Omics Data Encyclopedia, and Gene Expression Omnibus. Consensus clustering identified PRG subtypes, and gene set variation analysis screened differentially-expressed genes. Univariate Cox, Least Absolute Shrinkage and Selection Operator, and multivariate Cox regression analyses were used to construct a prognostic signature, validated in internal and 4 external cohorts. Immune infiltration was analyzed via single-sample gene set enrichment analysis, CIBERSORT, and ESTIMATE. The signature's predictive value for chemotherapy, transcatheter arterial chemoembolization, and immunotherapy was assessed with GDSC, tumor immune dysfunction and exclusion and relevant therapy cohorts, and its performance was compared with 4 published HCC signatures. The expression of signature genes was verified at the mRNA, protein, and single-cell levels. Two PRG subtypes were identified in HCC. Cluster A correlated with poor prognosis, high immune infiltration, and low metabolic activity. A 4-gene signature (KIF20A, SPP1, PON1, and SERPINE1) was established, effectively stratifying HCC patients into high- and low-risk groups with significantly different overall survival. It was an independent prognostic factor with better predictive performance than published signatures, and the high-risk group had a higher TP53 mutation rate, distinct immune infiltration patterns, and was predicted to have poor therapy responses. The 4 signature genes showed abnormal expression in HCC tissues at multiple levels. This study identified 2 PRG subtypes and a robust 4-gene prognostic signature for HCC, which accurately predicts prognosis, reflects tumor immune microenvironment features, and guides precision therapy selection, highlighting PRGs' potential as HCC biomarkers and therapeutic targets.