Deciphering Angiogenic Drivers in Hepatocellular Carcinoma: From Prognostic Signature Construction to...

Hepatocellular carcinoma (HCC) is an aggressive malignancy with poor prognosis and a lack of reliable biomarkers. In this study, we integrated weighted gene co-expression network analysis, machine-learning-based prognosticmodelling, single-cell RNA sequencing, spatial transcriptomics, and molecular docking to identify angiogenesis-related drivers in HCC. SCAF1 was identified as a hub gene and was significantly upregulated in HCC tissues compared with normal liver tissues. High SCAF1 expression was closely associated with advanced clinicopathological features and unfavourable survival outcomes. The prognostic model constructed by machine learning showed robust predictive performance in both training and validation cohorts, and decision curve analysis supported its clinical utility. Single-cell and spatial transcriptomic analyses further suggested that SCAF1 was mainly associated with endothelial and immune-related features of the tumour microenvironment. Functional assays showed that SCAF1 knockdown suppressed HCC cell migration, invasion, VEGFA expression, and angiogenic capacity in vitro. Mechanistically, SCAF1 was negatively correlated with anti-angiogenic chemokines, including CCL19, CCL14, and CCL11, indicating a role in shaping a pro-angiogenic microenvironment. In addition, molecular docking identified genistein as a promising compound with stable binding to SCAF1. Experimental validation showed that genistein reduced SCAF1 and VEGFA expression and inhibited HCC cell proliferation, migration, invasion, and tube formation. Collectively, these findings suggest that SCAF1 is a novel prognostic and angiogenesis-related biomarker in HCC and a potential therapeutic target for genistein-based intervention.