Abstract 2168: Multi-regional RNA-Seq reveals pattern of RNA-editing in brain tumor

Abstract

Abstract Introduction: RNA-editing is widespread across human genome. Previous studies have observed that RNA editing level differs significantly across multiple cancer types and may influence cancer development. Increased level of RNA editing has been found to correlate with enhanced tumorigenesis in some cancers but reduced tumorigenesis in others. In brain tumor, the RNA editing landscape has not been well characterized partially due to the difficulty of sample accessibility. We performed multi-regional RNA-Seq on cancerous, peritumoral and distant tissue collected from 40 brain cancer patients. Our aim is to depict the landscape of RNA editing and to comprehend the relevance of RNA editing with other molecular characteristics of brain cancer. Methods: We collected 40 cancerous brain tissue sample (30 LGG and 10 GBM), each coupled with a peritumoral tissue and a distant tissue. These 120 samples were subjected to RNA-Seq sequencing. RNACocktail was applied for identification of RNA editing events. RADAR and ANNOVAR were used for identification and annotation of A-to-I editing sites. RNAEditingIndexer was used to estimate sample-wise editing level denoted as AEI. Pathway enrichment analysis was performed using GSEA. Statistical difference of molecular characteristics between sample groups was calculated using either Wilcoxon signed rank test or Kruskal-Wallis test. Results: A total number of 21,105 and 5,634 informative editing events were identified in LGG and GBM respectively. As expected, most editing events were located within Alu element of intronic and 3'UTR region. In both LGG and GBM, RNA editing level showed decreasing trend from distant to peritumoral to cancerous samples. Compared with normal tissue, we found 5.12‰ up-regulated events and 16.92‰ down-regulated events in LGG samples. Similar pattern was found in GBM as well. We found two editing sites with significant editing differences (adjust p-value < 0.05) within coding region of GRIK2 and NEIL1 gene. In addition, numerous editing sites showed statistical correlation other molecular/pathological characteristics such as IDH1 mutation, existence of seizure, MGMT methylation and pathological classification. At sample level, the same decreasing trend of RNA editing level was consistently observed from distant to cancerous tissue. Enrichment analysis showed that editing level was positively correlated with neural transmission and oncogenic related pathways. In addition, sample level editing was significant different between WHO as well as pathological classification. Citation Format: Cheng Yan, Jun Wu, Dan Wang, Lan Su, Yufei Yang. Multi-regional RNA-Seq reveals pattern of RNA-editing in brain tumor [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 2168.