Exploration of heterogeneity and recurrence signatures in hepatocellular carcinoma.

Abstract

Hepatocellular carcinoma (HCC), the sixth most prevalent cancer globally, is characterized by high recurrence rates and poor prognosis. Investigating the heterogeneity of relapsed HCC and identifying key therapeutic targets may facilitate the design of effective anticancer therapies. In this study, integrative analysis of single-cell RNA sequencing data of primary and early-relapsed HCC revealed increased proportions of infiltrating CD8⁺ T cells along with malignant cells and a decrease in CD4⁺ T cells in relapsed HCC. Cellular interaction and immunohistochemical analysis proposed MIF-(CD74 + CXCR4) signaling pathway as a key mechanism by which malignant cells influence immune cells within the tumor microenvironment. Notably, primary malignant cells showed greater differentiation and proliferation potential, whereas relapsed cells exhibited enhanced epithelial-mesenchymal transition and inflammation, along with upregulated glycogen synthesis and metabolism-related gene expression. Using machine learning techniques on bulk RNA-seq data, we developed a relapsed tumor cell-related risk score (RTRS) that independently predicts overall and recurrence-free survival time with higher accuracy compared with conventional clinical variables. Prognostic biomarkers and potential therapeutic targets were validated via RT-qPCR using mouse implantation models. This comprehensive investigation elucidates the heterogeneity of relapsed HCC and constructs a novel postoperative recurrence prognostic model, paving the way for targeted therapies and improved patient outcomes.