Integrative Analysis Identifies a TNFα-Derived Gene Signature for Predicting Prognosis, Tumor Immunity, and Treatment Sensitivity in Gastric Cancer

Objective: TNF-α is an essential pro-inflammatory cytokine in the tumor microenvironment of gastric cancer (GC), possessing a key biological and clinical impact. Here, we conducted an integrative analysis of the role of TNFα-derived genes in GC prognosis and precision medicine. Methods: We pooled transcriptome and clinical features of GC patients from TCGA and GSE15459 projects. TNFα signaling was quantified through the ssGSEA algorithm, and TNFα-derived genes were screened with WGCNA. Thereafter, a LASSO model was established. The somatic mutation was analyzed across GC specimens. Immune cell infiltrations were inferred through ESTIMATE and ssGSEA algorithms, followed by measuring the immune checkpoint expression. AKR1B1, CPVL, and CTSL expressions were measured in gastric mucosal cells GES-1 and GC cells (HGC-27, MKN-28, and AGS) through RT-qPCR and Western blotting. Results: A TNFα-derived gene signature (containing AKR1B1, CPVL, and CTSL) was developed for GC. A high-risk score indicated more undesirable OS, DFS, DSS, and PFS outcomes. Time-independent ROC curves and multivariate cox regression models confirmed that the signature reliably and independently predicted GC prognosis. Additionally, risk scores displayed significant correlations to more severe histological grades and pathological stages. A low-risk score was characterized by increased somatic mutation, while a high-risk score was characterized by immune and stromal activation, enhanced immune cell infiltrations, and increased expression of immune checkpoint molecules. Experimental results confirmed the significant upregulation of AKR1B1, CPVL, and CTSL in GC cells. Conclusion: Collectively, stratification based on the TNFα-derived gene signature might enable GC patients to predict prognosis, benefit from immunotherapy, and assist in formulating novel therapeutic regimens.