Uncovering epitranscriptome-wide mRNA m6A methylation and expression patterns in human glioma

Abstract

Glioblastoma (GB) continue to be a devastating disease despite advances in treatment therapies. High invasiveness, intra-tumoral heterogeneity and poor prognosis are key barriers to successful GB treatment. The most prevalent m6A modification in RNA serves as a biomarker for cancer diseases. Recent findings implicate the valuable role of m6A modification in tumor progression and tumorigenesis of gliomas. We profiled m6A modifications within mRNA across glioma stem cells NCH421k (GSCs) using methylated RNA immunoprecipitation (MeRIP) technique and glioblastoma tumor tissues using direct Nanopore RNA sequencing (dRNA-seq). A total of 1,570 distinct differentially modified genes were found after MeRIP-seq data analysis of which 740 genes were hypermethylated, and 830 genes were hypomethylated in GSCs. Four-quadrant plot was produced to evaluate the distribution of differentially expressed mRNAs with statistically significant m6A modifications. After, dRNA-seq data showed 4,340 RRACH motifs linked to hyper-methylated upregulated genes and 9,106 linked to hyper-methylated downregulated genes across all patients’ samples. To find statistically significant 8 RRACH motifs through seven genes and narrow down the list we used logistic regression model and Chi-square test: AAACA|2129|OS9, AGACA|1210|PAGR1, GGACA|2173|OS9, GGACT|2187|TOB1, AAACC|3283|PIK3R2, GAACC|3068|GP1BB, GGACA|3110|RETREG1 and GGACT|3122|LUC7L3. Analysis showed that adenosine in selected motifs in LGG samples undergo modifications about 3.4 times more frequently compared to GB, and subsequent hierarchical clustering analysis classified patient samples according to pathology. Next, based on 8 RRACH motifs methylation we constructed total m6A methylation score which significantly differed between LGG and GB clusters (p=0.0002) being higher in LGG cluster. Low m6A methylation score was associated with a worse survival prognosis in glioma patients (p=0.016). The expression patterns of selected genes did not show any significant differences in m6A-based clusters (p=0.08), as well as the Kaplan-Meier survival analysis (p=0.79). To summarize, m6a modifications in mRNA can serve as potential biomarkers in gliomas.