Identification of a basement membrane-related genes signature to predict prognosis, immune landscape and guide therapy in gastric cancer

Abstract

The basement membrane is an essential defense against cancer progression and is intimately linked to the tumor immune microenvironment. However, there is limited research comprehensively discussing the potential application of basement membrane-related genes (BMRGs) in the prognosis evaluation and immunotherapy of gastric cancer (GC). The RNA-seq data and clinical information of GC patients were collected from the TCGA and GEO database. Prognosis-associated BMRGs were filtered via univariate Cox regression analysis. The 4-BMRGs signatures were constructed by lasso regression. Prognostic predictive accuracy of the 4-BMRGs signature was appraised with survival analysis, receiver operating characteristic curves, and nomogram. Gene set enrichment analysis (GSEA), gene ontology, and gene set variation analysis were performed to dig out potential mechanisms and functions. The Estimate algorithm and ssGSEA were used for assessing the tumor microenvironment and immunological characteristics. Identification of molecular subtypes by consensus clustering. Drug sensitivity analysis using the “pRRophetic” R package. Immunotherapy validation with immunotherapy cohort. A 4-BMRGs signature was constructed, which could excellently predict the GC patient prognosis (5-year AUC value of 0.873). Kaplan–Meier and Cox regression analyses showed that the 4-BMRGs signature was an OS-independent prognostic factor, and that higher risk scores were associated with shorter OS. The high-risk subgroup exhibits a higher abundance of immune cell infiltration, such as macrophages. Additionally, we observed a strong correlation between 2 BMRGs (LUM, SPARC) and immune cells such as CD8 + T cells and macrophages. The high-risk subgroup appears to be more sensitive to Axitinib, DMOG, Gemcitabine and Docetaxel by pRRophetic analysis. Furthermore, the validation of the cohort that received immune therapy revealed that patients in the high-risk group who underwent immune checkpoint inhibitor treatment exhibited better response rates. Pan-cancer analysis also shows that risk scores are strongly associated with immune and carcinogenic pathways. The 4-BMRGs signature has demonstrated accuracy and reliability in predicting the GC patient’s prognosis and could assist in the formulation of clinical strategies.