Identification of Hypoxia- and Circadian Rhythm-Related Prognostic Genes in Hepatocellular Carcinoma Through Integrative Transcriptomic and Single-Cell Analysis

Hypoxia perturbs circadian rhythm gene expression in hepatocellular carcinoma, yet prognostic genes associated with both processes remain insufficiently characterized. Data from the liver hepatocellular carcinoma (LIHC) cohort in The Cancer Genome Atlas (TCGA) and the Gene Expression Omnibus (GEO) were analyzed. Candidate genes were identified by intersecting differentially expressed genes (DEGs) with 27 hypoxia- and circadian rhythm-related genes and were further refined using univariate Cox regression and least absolute shrinkage and selection operator (LASSO) regression; a random survival forest (RSF) risk model was then constructed and validated. Independent prognostic and functional enrichment analyses, immune infiltration profiling, single-cell RNA sequencing (scRNA-seq), and immunohistochemistry (IHC) were subsequently performed. Four prognostic genes (ADM, PFKFB3, SERPINE1, and RORA) were identified through integrative computational analysis. The RSF model stratified LIHC patients into the high-risk group (HRG) and the low-risk group (LRG), which had significantly different survival outcomes, and the risk score and tumor stage were combined in a nomogram. The four genes were co-enriched in four pathways; eight immune cell types, including natural killer T (NKT) cells, differed between the risk groups. scRNA-seq identified stem-like/hypoxic malignant cells whose differentiation trajectory showed distinct expression dynamics of the four genes. IHC confirmed significantly higher ADM and PFKFB3 protein expression in LIHC tissues, whereas protein-level evidence for SERPINE1 and RORA was inconclusive. These findings support further prognostic evaluation of LIHC and investigation of the links among hypoxia, metabolic reprogramming, and stemness.