Transcriptome-wide Landscape of RNA Modification N6-Methyladenosine in Chromophobe Renal Cell Carcinoma

Abstract Background As the most abundant internal mRNA modification, N6-methyladenosine (m6A) is associated with various cancers. However, RNA modification m6A has not been studied in chromophobe renal cell carcinoma (chRCC). The present study aimed to comprehensively analyze the global m6A modification pattern in chRCC. Methods Three subjects with chRCC were enrolled in our study. Transcriptome-wide m6A methylone and transcriptome analysis in chRCC and tumor-adjacent normal tissues were detected via m6A-SEAL-seq and RNA-sEq. m6A-modified mRNAs were further validated by m6A-immunoprecipition followed by quantitative real-time PCR (m6A-IP-qPCR). The least absolute shrinkage and selection operator (LASSO) Cox regression and multivariate Cox proportional hazards regression analysis were used to determine the candidate gene. Results We performed qPCR in six subjects of chRCC and found that the expression levels of m6A writer subunit WTAP, m6A erasers FTO and ALKBH5, and m6A reader YTHDF2 were significantly downregulated in chRCC tissues compared with corresponding tumor-adjacent normal tissues. In all three subjects, 12,841 confident m6A peaks representing 10,102 transcripts and 15,024 confident m6A peaks representing 11,396 transcripts were respectively identified in chRCC and tumor-adjacent normal tissues. Analysis of differential m6A levels identified 644 hypermethylated m6A peaks and 1,304 hypomethylated m6A peaks in chRCC compared with tumor-adjacent normal tissues. Gene Ontology (GO) analysis revealed that genes with hypomethylated m6A peaks (1,137) were associated with pathways in cancer. Compared with tumor-adjacent normal tissues, 3,911 genes were significantly dysregulated in chRCC, including 2,344 downregulated and 1,567 upregulated mRNAs. Functional enrichment analysis revealed that the dysregulated genes in chRCC were significantly enriched in multiple metabolic processes. We identified two hypomethylated genes NOTCH1 and FGFR1, which might respectively act as a tumor suppressor and an oncogene in chRCC. Three m6A-dependent signatures were identified using Cox regression screen and LASSO regression. Based on the significant prognostic signatures, we build a m6A-dependent prognostic model, with the Concordance index (C-index) = 0.96. The Kaplan-Meier survival curve and log-rank between the high-risk and low-risk group showed significant difference. Conclusions This study presented the first m6A transcriptome-wide profile of human chRCC, which may provide clues for the m6A methylation-based research on chRCC epitranscriptomic etiology and pathogenesis.