Identification and validation of a novel redox- related differentially expressed lncRNA prognostic signature for predicting clinical immunotherapy response in gastric cancer

Abstract

Abstract Redox responses modulated by intracellular long noncoding RNA (lncRNA) can be involved in tumorigenesis and progression. However, the role of redox-related lncRNAs (RRlncRNAs) in gastric cancer (GC) development remains mostly unknown. Our research aims to establish and validate novel prognostic and immune infiltration markers for GC by constructing a prognostic model of RRlncRNAs. We downloaded the transcriptomic and mutational data for 407 GC pa-tients from The Cancer Genome Atlas (TCGA) database and randomized them 1:1 into a training and validation set to show that redox-related lncRNAs affect GC patients' prognosis. Subse-quently, the prognostic model was constructed for the screened RRlncRNAs using the Least Absolute Shrinkage and Selection Operator (LASSO) and the multivariate COX regression algo-rithm. Then, Survival analyses were performed on the train and test sets. The overall survival rate of GC patients was significantly correlated with the signatures of eight RRlncRNAs, including AC103702.2, AL138756.1, AL356417.2, CFAP61-AS1, RHPN1-AS1, CDK6-AS1, LINC02864, and AL355574.1. Meanwhile, we validated the model's accuracy through nomograms, Decision Curve Analysis (DCA), and comparisons using models from other studies. The results demonstrated that our model is more effective and outperforms the signature of Jiang et al. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis of gene enrichment in high-risk patients shows significant enrichment in immune-related pathways. Waterfall plots of gene mutations, tumor mutation burden (TMB), and tumor immune dysfunction and exclusion (TIDE) showed significant differences in immune function between high- and low-risk groups. Then, we divided the 407 GC patients into two clusters using a consensus clustering algorithm and found significant differences in their immune microenvironment through immune cell difference anal-ysis, ESTIMATEScore, and gene set enrichment analysis (GSEA). Taken together, we conclude that the prognostic model constructed by RRlncRNAs can significantly affect the prognosis of GC patients and may alter their tumor progression by modulating the immune microenvironment in vivo. Our study found eight RRlncRNA-associated signatures, representing promising new markers for immunotherapy and diagnosis in GC patients.