N6-methyladenosine RNA methylation regulator-related alternative splicing gene signature as prognostic predictor and in immune microenvironment characterization of patients with low-grade glioma

Abstract

Background: N6-methyladenosine (m6A) RNA methylation is an important epigenetic modification affecting alternative splicing (AS) patterns of genes to regulate gene expression. AS drives protein diversity and its imbalance may be an important factor in tumorigenesis. However, the clinical significance of m6A RNA methylation regulator-related AS in the tumor microenvironment has not been investigated in low-grade glioma (LGG). Methods: We used 12 m6A methylation modulatory genes (WTAP, FTO, HNRNPC, YTHDF2, YTHDF1, YTHDC2, ALKBH5, YTHDC1, ZC3H13, RBM15, METTL14, and METTL3) from The Cancer Genome Atlas (TCGA) database as well as the TCGA-LGG (n = 502) dataset of AS events and transcriptome data. These data were downloaded and subjected to machine learning, bioinformatics, and statistical analyses, including gene ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis. Univariate Cox, the Least Absolute Shrinkage and Selection Operator (LASSO), and multivariable Cox regression were used to develop prognostic characteristics. Prognostic values were validated using Kaplan-Maier survival analysis, proportional risk models, ROC curves, and nomograms. The ESTIMATE package, TIMER database, CIBERSORT method, and ssGSEA algorithm in the R package were utilized to explore the role of the immune microenvironment in LGG. Lastly, an AS-splicing factor (SF) regulatory network was examined in the case of considering the role of SFs in regulating AS events. Results: An aggregate of 3,272 m6A regulator-related AS events in patients with LGG were screened using six machine learning algorithms. We developed eight AS prognostic characteristics based on splice subtypes, which showed an excellent prognostic prediction performance. Furthermore, quantitative prognostic nomograms were developed and showed strong validity in prognostic prediction. In addition, prognostic signatures were substantially associated with tumor immune microenvironment diversity, ICB-related genes, and infiltration status of immune cell subtypes. Specifically, UGP2 has better promise as a prognostic factor for LGG. Finally, splicing regulatory networks revealed the potential functions of SFs. Conclusion: The present research offers a novel perspective on the role of AS in m6A methylation. We reveal that m6A methylation regulator-related AS events can mediate tumor progression through the immune-microenvironment, which could serve as a viable biological marker for clinical stratification of patients with LGG so as to optimize treatment regimens.