Identification of M2-like macrophage-related signature for predicting the prognosis, ecosystem and immunotherapy response in hepatocellular carcinoma

Abstract

Background Hepatocellular carcinoma is one of the most common malignancies worldwide, representing a big health-care challenge globally. M2-like macrophages are significantly correlated with tumor progression, metastasis and treatment resistance. Methods Integrative 10 machine learning algorithms were performed to developed a M2-like macrophage related prognostic signature (MRPS). Single-cell RNA-sequencing analysis was performed to dissect the ecosystem of HCC. Several approaches, including TIDE score, immunophenoscore, TMB score and tumor escape score were used to evaluate the predictive role of MRPS in immunology response. Results The optimal MRPS constructed by the combination of stepCox + superPC algorithm served as an independent risk factor and showed stable and powerful performances in predicting the overall survival rate of HCC patients with 2-, 3-, and 4-year AUCs of 0. 763, 0.751, and 0.699 in TCGA cohort. HCC patients with low risk score possessed a more interaction of immunoactivated cells, including NK, CD8+ cytotoxic T, and activated B, and a less interaction of immunosuppressive cells, including Treg, CD4+ exhauster T, and M2-like macrophage. Low risk score indicated a higher PD1&CTLA4 immunophenoscore, higher TMB score, lower TIDE score and lower tumor escape score in HCC, suggesting a better immunotherapy response. The IC50 value of docetaxel, gemcitabine, crizotinib and Osimertinib in HCC with high risk score were lower versus that with low risk score. HCC patients with high risk score had a higher score of cancer-related hallmarks, including angiogenesis, DNA repair, EMT, glycolysis, and NOTCH signaling. Conclusion Our study proposed a novel MRPS for predicting the prognosis, ecosystem and immunotherapy response in HCC.