Development and validation of a combined glycolysis and immune prognostic signature for lung squamous cell carcinoma

Background: The involvement of glycolysis in the regulation of the tumor immune microenvironment has become a novel research field. In this study, the specific functions and clinical significance of glycolysis-related genes (GRGs) and immune-related genes (IRGs) were systematically characterized in lung squamous cell carcinoma (LUSC). Methods: We evaluated the prognostic value, interactions, somatic mutations, and copy-number variations of GRGs and IRGs in LUSC from a dataset of The Cancer Genome Atlas (TCGA). An integrated glycolysis–immune score (GIS) model was generated by random forest algorithm and stepwise Cox regression analysis. The predictive power of the GIS was examined by survival analysis, receiver operating characteristics, univariate and multivariate analyses, and subgroup analysis. The correlations between GIS and biological functions, glycolysis, immune activity, immune cell infiltration, and genomic changes were analyzed, and the potential of GIS to guide clinical treatment decisions was evaluated. Results: A total of 54 prognostic GRGs and IRGs were identified, and a strong correlation was noted among them. However, most of them had somatic mutations and a high incidence of CNV. The GIS model that contained two GRGs (PYGB and MDH1) and three IRGs (TSLP, SERPIND1, and GDF2) was generated and a high GIS indicated poor survival. Moreover, we found that low GIS was associated with immune pathway activation, M1 macrophage infiltration, and higher immune scores. Finally, patients with low GIS were more sensitive to chemotherapy and immunotherapy. Conclusion: An integrated model based on glycolysis and immune genes can distinguish the biological functions and immune infiltration patterns of individual tumors, quantitatively estimate the prognosis of patients with LUSC, and guide chemotherapy and immunotherapy decisions.