Abstract
BACKGROUND: Gastric cancer (GC) is a prevalent malignancy with high mortality rate. The process of Epithelial-mesenchymal transition (EMT) significantly contributes to its metastasis and resistance to therapy. This research is designed to develop a survival prediction model for EMT-related genes in GC. METHODS: This study used GC data from public databases and screened core module genes via weighted gene co-expression network analysis (WGCNA). Subsequently, we combined univariate Cox regression and least absolute shrinkage and selection operator (LASSO) regression to identify key feature genes and establish a prognostic model for GC patients. The survival package was used for survival analysis, and the pROC package was employed to generate receiver operating characteristic (ROC) curves for evaluating model performance. The CIBERSORT and ESTIMATE algorithms were applied to assess immune cell infiltration in patients with different risk levels. Gene set enrichment analysis (GSEA) was conducted for pathway enrichment analysis. Finally, in vitro cell experiments were performed to verify the expression levels and potential biological functions of the key genes. RESULTS: We obtained three key genes, NPR3, OLFML2B and GREB1, and established a prognostic RiskScore model consisting of these three key genes, and the area under ROC curve (AUC) > 0.6 confirmed the predictive efficacy of this model. GSEA confirmed that the tumor progression pathways, such as EMT, angiogenesis, and so on, were notably activated in high-risk group. Moreover, patients in high-risk group exhibited a stronger tendency to immune escape and were significantly less sensitive to Afatinib, Gefitinib, and Lapatinib. Finally, in vitro tests displayed that NPR3 knockdown markedly decreased the viability, migration and invasion of GC cells. CONCLUSION: Our study provides a prognostic assessment tool for GC based on EMT-related genes and offers novel insights into understanding the roles of EMT in GC progression and treatment resistance. These findings may aid in the development of precision therapy strategies for GC.