A Glycolysis-Related Gene Signature Correlates With the Characteristics of the Tumor Immune Microenvironment and Predicts Prognosis in Patients With Hepatocellular Carcinoma

Jun Yang; Yuening Zhang; Jin Duan; Xiaojie Huang; Haibin Yu; Zhongjie Hu

doi:10.3389/fmolb.2022.834976

A Glycolysis-Related Gene Signature Correlates With the Characteristics of the Tumor Immune Microenvironment and Predicts Prognosis in Patients With Hepatocellular Carcinoma

Front Mol Biosci. 2022 Apr 25:9:834976. doi: 10.3389/fmolb.2022.834976. eCollection 2022.

Authors

Jun Yang¹, Yuening Zhang², Jin Duan¹, Xiaojie Huang², Haibin Yu¹, Zhongjie Hu²

Affiliations

¹ Phase Ⅰ Clinical Trial Ward, Beijing Youan Hospital, Capital Medical University, Beijing, China.
² Beijing Youan Hospital, Capital Medical University, Beijing, China.

Abstract

Aim: To develop a glycolysis-related gene signature that correlated with the characteristics of the tumor immune microenvironment and had good predictive power for overall survival (OS) in hepatocellular carcinoma (HCC). Methods: Gene expression profiles, RNA sequencing data, clinical characteristics and survival information for 407 patients with HCC and 58 healthy controls were downloaded from the TCGA database. GSEA 4.1.0 software was used to evaluate the glycolysis-related pathways enriched in HCC compared to normal liver tissue. Univariate Cox, Least Absolute Shrinkage, Selection Operator, and two-step multivariate Cox analyses were used to construct a glycolysis-related gene signature for prognostic prediction. The glycolysis-related gene signature was combined with clinical characteristics to generate a nomogram. Tumor-infiltrating immune cell profiles and PD-L1 protein expression in HCC tissues were investigated. Results: The gene expression profiles of HCC tissues were enriched in glycolysis-related pathways. A glycolysis-related gene signature was used to categorize patients as high-risk or low-risk, where high-risk patients had significantly worse OS. Receiver operating characteristic curves confirmed the predictive capability of the glycolysis-related gene signature for OS (AUC >0.80). There was a significant difference in M0 macrophage (p = 0.017), dendritic cell (p = 0.043), B cell (p = 0.0018), CD4 T cell (p = 0.003), Treg (p = 0.01) and mast cell (p = 0.02) content and PD-L1 protein expression (p = 0.019) between HCC tissues in patients in the high-risk and low-risk groups. Conclusion: We established a glycolysis-related gene signature for OS in HCC that was predictive in training and test TCGA cohorts and correlated with the characteristics of the HCC tumor immune microenvironment. The glycolysis-related gene signature may guide clinical decision-making concerning patient selection for immunotherapy in HCC.

Keywords: PD-L1; hepatocellular carcinoma; immune microenvironment; immune-related glycolysis gene; prognosis.