Glucose metabolism 3-gene signature associates with poor prognosis in hepatocellular carcinoma as a hallmark of tumor microenvironment evolution and progression

Abstract

Background At present, immunotherapy has become an established treatment for cancer. Recently, we have realized that tumor metabolism has a huge impact on the shaping of immune reactivity in the tumor microenvironment (TME). Exploration of the tumor metabolic and immunocellular response reveals metabolic vulnerability as a therapeutic window for intervention to enhance immunotherapy. The clinical significance of glycolysis and its role in tumor-immune evolution have not been fully explored. Understanding the relationships between glycolysis, TME evolution and disease progression is of great significance for optimizing immunotherapy of liver cancer. Methods A glycolysis-related biomarker signature was initially identified in transcriptomic data within The Cancer Genome Atlas (TCGA, n=424). Predicted overall survival (OS) for the gene signature in the GSE54236 dataset (n=161) and our HCC cohort (n=132) was subsequently independently verified. The relative number of immune cells in the microenvironment was calculated using CIBERSORT. CCK-8, IHC, Transwell and Seahorse were used to assess the effects of a 3-gene signature on the malignant phenotype of liver cancer. Results A 3-gene signature of glycolysis significantly associated with poor OS was identified. Based on a multi-omics research strategy, we found the 3-gene signature was closely related to the evolution of liver cancer TME and disease progression, and was associated with liver cancer mutation load and immune evolution. High-risk patients have inhibition and depletion of the immune TME and immune escape through antigen presentation inhibition. Remarkably, the 3-gene signature can predict the clinical outcomes of HCC patients with different clinical subtypes. Conclusions Our investigation of the 3-gene signature provided evidence for tumor metabolism-immune co-evolution along HCC progression. The 3-gene signature for stratification of liver cancer risk has robust predictive power, both at the RNA and protein levels.