Prognostic Value of Genomic Instability of m(6)A-Related lncRNAs in Lung Adenocarcinoma

Background: Genomic instability of N6-methyladenosine (m(6)A)–related long noncoding RNAs (lncRNAs) plays a pivotal role in the tumorigenesis of lung adenocarcinoma (LUAD). Our study identified a signature of genomic instability of m(6)A-associated lncRNA signature and revealed its prognostic role in LUAD. Methods: We downloaded RNA-sequencing data and somatic mutation data for LUAD from The Cancer Genome Atlas (TCGA) and the GSE102287 dataset from the Gene Expression Omnibus (GEO) database. The “Limma” R package was used to identify a network of regulatory m(6)A-related lncRNAs. We used the Wilcoxon test method to identify genomic-instability–derived m(6)A-related lncRNAs. A competing endogenous RNA (ceRNA) network was constructed to identify the mechanism of the genomic instability of m(6)A-related lncRNAs. Univariate and multivariate Cox regression analyses were performed to construct a prognostic model for internal testing and validation of the prognostic m(6)A-related lncRNAs using the GEO dataset. Performance analysis was conducted to compare our prognostic model with the previously published lncRNA models. The CIBERSORT algorithm was used to explore the relationship of m(6)A-related lncRNAs and the immune microenvironment. Prognostic m(6)A-related lncRNAs in prognosis, the tumor microenvironment, stemness scores, and anticancer drug sensitivity were analyzed to explore the role of prognostic m(6)A-related lncRNAs in LUAD. Results: A total of 42 genomic instability–derived m(6)A-related lncRNAs were differentially expressed between the GS (genomic stable) and GU (genomic unstable) groups of LUAD patients. Four differentially expressed lncRNAs, 17 differentially expressed microRNAs, and 75 differentially expressed mRNAs were involved in the genomic-instability–derived m(6)A-related lncRNA-mediated ceRNA network. A prediction model based on 17 prognostic m(6)A-associated lncRNAs was constructed based on three TCGA datasets (all, training, and testing) and validated in the GSE102287 dataset. Performance comparison analysis showed that our prediction model (area under the curve [AUC] = 0.746) could better predict the survival of LUAD patients than the previously published lncRNA models (AUC = 0.577, AUC = 0.681). Prognostic m(6)A-related-lncRNAs have pivotal roles in the tumor microenvironment, stemness scores, and anticancer drug sensitivity of LUAD. Conclusion: A signature of genomic instability of m(6)A-associated lncRNAs to predict the survival of LUAD patients was validated. The prognostic, immune microenvironment and anticancer drug sensitivity analysis shed new light on the potential novel therapeutic targets in LUAD.