A comprehensive pan-cancer analysis identifies a novel glycolysis score and its hub genes as prognostic and immunological biomarkers

BACKGROUND: Glycolysis plays significant roles in tumor progression and immune response. However, the exact role of glycolysis in prognosis and immune regulation has not been explored in all cancer types. This study first calculated a novel glycolysis score and screened out 12 glycolytic hub genes, and comprehensively analyzed molecular expression, clinical implications, and immune features of glycolysis among pan-cancer. METHODS: The glycolysis score was derived by the single sample gene set enrichment analysis (ssGSEA) algorithm. The correlations of glycolysis with clinical parameters were analyzed using “limma” package. Downstream pathways of glycolysis were identified by Gene Set Enrichment Analysis (GSEA). The immune cell infiltration was explored and validated by three databases. The association between glycolysis and some immunotherapy biomarkers was explored by Pearson correlation analysis. Single-nucleotide variation (SNV), copy number variation (CNV), DNA methylation, and drug sensitivity analyses of 12 glycolytic hub genes were investigated. IMvigor210 and GSE91061 immunotherapeutic cohorts were retrieved to assess the ability of glycolysis score to predict immunotherapy efficacy. The expression of glycolysis key genes was detected in normal and endometrial cancer cell lines. RESULTS: We found that glycolysis score was generally higher in tumor tissues compared to normal tissues and a high glycolysis score predominated as a risk prognostic factor. A high glycolysis score was associated with decreased immunostimulatory natural killer (NK) cells and CD8(+) T cells infiltration, well increased immunosuppressive M2-tumor-associated macrophages (M2-TAM) cells infiltration. Tumor mutational burden (TMB), microsatellite instability (MSI), and immune checkpoints (ICPs) all closely interacted with glycolysis score and the frequency of gene mutation was confirmed to be higher in colon adenocarcinoma (COAD) patients with higher glycolysis score. The SNV, CNV, and DNA methylation of 12 glycolysis key genes occurred at different frequencies and showed different impacts on survival outcomes. The predictive and prognostic value of glycolysis score for immunotherapy outcomes was validated in two immunotherapy cohorts. The expression levels of key genes differ in normal endometrial and three endometrial cancer cell lines. CONCLUSIONS: This work indicated that glycolysis score and 12 glycolytic hub genes were correlated with an immunosuppressive microenvironment. They could be served as promising biomarkers aiding diagnosis, predicting prognosis and immunotherapy response for some tumor patients.