Genetic control of RNA editing in neurodegenerative disease

Abstract

A-to-I RNA editing diversifies human transcriptome to confer its functional effects on the downstream genes or regulations, potentially involving in neurodegenerative pathogenesis. Its variabilities are attributed to multiple regulators, including the key factor of genetic variants. To comprehensively investigate the potentials of neurodegenerative disease-susceptibility variants from the view of A-to-I RNA editing, we analyzed matched genetic and transcriptomic data of 1596 samples across nine brain tissues and whole blood from two large consortiums, Accelerating Medicines Partnership-Alzheimer's Disease and Parkinson's Progression Markers Initiative. The large-scale and genome-wide identification of 95 198 RNA editing quantitative trait loci revealed the preferred genetic effects on adjacent editing events. Furthermore, to explore the underlying mechanisms of the genetic controls of A-to-I RNA editing, several top RNA-binding proteins were pointed out, such as EIF4A3, U2AF2, NOP58, FBL, NOP56 and DHX9, since their regulations on multiple RNA-editing events were probably interfered by these genetic variants. Moreover, these variants may also contribute to the variability of other molecular phenotypes associated with RNA editing, including the functions of 3 proteins, expressions of 277 genes and splicing of 449 events. All the analyses results shown in NeuroEdQTL (https://relab.xidian.edu.cn/NeuroEdQTL/) constituted a unique resource for the understanding of neurodegenerative pathogenesis from genotypes to phenotypes related to A-to-I RNA editing.

Keywords: A-to-I RNA editing; RNA editing quantitative trait loci; RNA-binding protein; mediation analysis; neurodegenerative disease.

Figure 1

The identification and distribution analysis of RNA-editing QTLs. (A) The pipeline of RNA-editing QTL identification. (B–D) The distributions of RNA-editing QTLs in repeats, regions and chromosomes. (E) The distances of RNA-editing QTLs from the associated editing events. (F) The enriched pathways of the associated edited genes by Enrichr.

Figure 2

An RNA-editing QTL (rs769450) showed its roles in neurodegeneration. (A–C) The Manhattan plot and boxplots showed that this RNA-editing QTL was negatively associated with two RNA-editing events in the cerebellum region of AD patients. (D, E) These two RNA-editing events were abnormally edited in AD samples compared to controls. (F) These analyses and previous literature revealed the alleviation role of this RNA-editing QTL in neurodegeneration. AD: Alzheimer’s disease.

Figure 3

The analysis of RNA-editing QTLs in RBP genes to show their effects on A-to-I RNA editing. (A) The number of RNA-editing QTLs in RBP genes associated with the editing events in their targets. The pubmed IDs shown in this panel are the studies reporting the regulations of RNA-binding proteins on A-to-I RNA editing. (B) The number of pairs between RNA-editing QTLs in RBP genes and the editing events in their targets whose frequencies were significantly associated with RBP expressions. Blue and orange bars represent the negative and positive associations respectively. (C) The fraction of the significant associations in different genotyping groups. (D) The RNA-editing QTL (rs3760252) may interfere in the regulation of DDX42 on the editing event (chr17_61913227_-). (E) It was positively associated with the editing event in the cerebellum region of AD patients. (F) The expressions of DDX42 and the frequencies of the editing event were only significantly associated in the AA genotyping group. (G) The associated editing event was highly edited in AD. RBP: RNA-binding protein.

Figure 4

The analysis of RNA-editing QTLs in RBP targets to show their effects on A-to-I RNA editing. (A) The number of RNA-editing QTLs in RBP targets. The pubmed IDs shown in this panel are the studies reporting the regulations of RNA-binding proteins on A-to-I RNA editing. (B) The enrichment (expected versus observed) of the RNA-editing QTLs in RBP targets by GREGOR. The enriched ADAR-binding sites from two different studies showed similar results, revealing the reliability of the CLIPseq analyses. (C) The number of RNA-editing QTLs in RBP targets potentially regulating the editing events whose frequencies were significantly associated with the corresponding RBP expressions in the genotyping groups (AA or aa). (D–F) An RNA-editing QTL potentially disrupted NOP58 regulation on an RNA-editing event of FGD5-AS1 in the STG region of AD patients. (G) The statistically positive or negative effects of several RNA-binding proteins on editing events in different groups. For more detailed statistics results, please refer to Table S5 available online at http://bib.oxfordjournals.org/. (H) The number of RBPs showing potentially positive or negative effects on the editing events in the four groups. For blood samples of PD and AC samples of AD, most RBPs showed negative effects on the RNA-editing events. While in another two groups, most RBPs presented positive effects. (I, J) For the 22 RBPs showing opposite effects on 81 RNA-editing events between blood samples of PD patients and CER samples of AD patients, we compared the levels of these editing events in the two groups. All the editing events were significantly down-regulated in PD, partially, since these RBPs probably conferred negative effects in this group. GREGOR: genomic regulatory elements and GWAS overlap algorithm, CLIPseq: cross-linking immunoprecipitation associated to high-throughput sequencing, DLPFC: dorsolateral prefrontal cortex, AC: head of caudate nucleus, PCC: posterior cingulate cortex, CER: cerebellum, TCX: temporal cortex, FP: frontal pole, IFG: inferior frontal gyrus, PG: parahippocampal gyrus, STG: superior temporal gyrus, PD: Parkinson’s disease.

Figure 5

The shared genetic architecture of RNA editing across different regions and diseases. (A) π1 statistics for the significant pairs in the first dataset (row) that were also shared by the second dataset (column). The statistics values were calculated by two procedures. First, we selected the pairs of RNA-editing QTLs and editing events in the second dataset that were significantly correlated in the first dataset (FDR < 0.05). Second, for the P values of the selected associations, we calculated π1 values using the R function of qvalue_trunc with the bootstrap estimation method. (B) The π1 statistics results were partially attributed to the number of significant pairs in the second dataset. (C) The number of significant pairs was associated with the number of samples in each group. (D) Pearson correlation coefficients of beta values between two groups for the overlapped significant pairs of RNA-editing QTLs and editing events (P < 0.05). (E, F) AD and PD shared the genetic regulation of rs36118024 on the editing event in FGD5-AS1.

Figure 6

The effects of RNA-editing QTLs on proteins. (A, B) RNA-editing QTLs which themselves or whose associated RNA-editing events may affect amino acid sequences and protein functions. (C) The enriched biological functions and pathways of the neurodegeneration related proteins affected by RNA-editing QTLs (P < 0.05, Q < 0.2). (D, E) The genetic variant (rs35024632) may play its roles in neurodegeneration through the down-regulation of an RNA-editing event in the PCC regions of AD patients.

Figure 7

Shared genetic architecture of RNA-editing events and gene expressions. (A) The models showed the associations between genetic variations, RNA-editing events, and gene expressions. (B) The distributions of these associations in each group. (C, D) One example in the CER region of AD patients showed that the edQTL (rs2240783) associated with the expression QTL in the adjacent locations of FUS targets may affect the expressions of SLC6A7, accompanied by the altered RNA-editing event (chr5_149575166_+) in it. (E, F) One example in the AC region of AD patients showed that the eQTL (rs7184655) may affect the expressions of neurodegeneration-related RPL13, through its regulation on an RNA-editing event (chr16_89630026_+) in the 3′-UTR of this gene.

Figure 8

Shared genetic architecture of RNA editing and alternative splicing. (A) The models showed the associations between genetic variants, RNA-editing events and alternative splicing patterns. (B) The distributions of these associations in each group. (C, D) One example in the DLPFC region of AD patients showed that the edQTL (rs1049115) may disturb LIN28 regulations on the exon skipping and editing event in adjacent two genes. (E, F) One example in the CER region of AD patients showed that the eQTL (rs202639) was associated with an intron retention event though partial mediation of an RNA-editing event in ZC3H7B. For the potential mechanisms of the associations between editing and splicing, we explored the possible regulations of three RNA-binding proteins on these two events, supported by their significant correlations.