Identification of TNF-related gene subtypes, development of a prognostic model and exploration of the tumor immune landscape in gastric cancer

Abstract Background Gastric cancer (GC) is a common cancer in the world, with a high prevalence in East Asia. Tumor necrosis factor (TNF) is considered a high correlation to the development of tumor. This study aim to establish a prognostic model based on TNF-related genes (TNFRG), and to analyze the role of TNFRG in immune function. Method RNA sequencing data and information on clinical features of GC samples were extracted from TCGA-STAD and GEO (GSE84437) databases. Molecular and gene subtypes were identified and derived out using unsupervised clustering analysis. Prognostic models were determined and constructed later by univariate and multivariat cox regression, and LASSO regression. Reliability of prognostic models was verified using ROC curves and Kaplan-Meier analysis. The nomogram was used to quantify the probability of survival. Immune-related functions were analyzed using CIBERSORT and ssGSEA. Finally, RT-qPCR determined the expression of risk genes in GC. Result We confirmed two molecular subtypes and three gene subtypes by two clustering analyses. Both molecular cluster A and gene cluster C had higher levels of immune cell infiltration and better prognosis than others. A prognostic model comprised of four risk genes was constructed, which had different responses to TME, immune checkpoints, immune scores, immune cell infiltration and chemotherapy drug sensitivity, respectively. Conclussion This study enhances our understanding of TNFRG in GC, and provides a theoretical basis for predicting tumor prognosis and clinical treatments.