The Potential Regulation of A-to-I RNA Editing on Genes in Parkinson's Disease

Abstract

Parkinson's disease (PD) is characterized by dopaminergic neurodegeneration and an abnormal accumulation of α-synuclein aggregates. A number of genetic factors have been shown to increase the risk of PD. Exploring the underlying molecular mechanisms that mediate PD's transcriptomic diversity can help us understand neurodegenerative pathogenesis. In this study, we identified 9897 A-to-I RNA editing events associated with 6286 genes across 372 PD patients. Of them, 72 RNA editing events altered miRNA binding sites and this may directly affect miRNA regulations of their host genes. However, RNA editing effects on the miRNA regulation of genes are more complex. They can (1) abolish existing miRNA binding sites, which allows miRNAs to regulate other genes; (2) create new miRNA binding sites that may sequester miRNAs from regulating other genes; or (3) occur in the miRNA seed regions and change their targets. The first two processes are also referred to as miRNA competitive binding. In our study, we found 8 RNA editing events that may alter the expression of 1146 other genes via miRNA competition. We also found one RNA editing event that modified a miRNA seed region, which was predicted to disturb the regulation of four genes. Considering the PD-related functions of the affected genes, 25 A-to-I RNA editing biomarkers for PD are proposed, including the 3 editing events in the EIF2AK2, APOL6, and miR-4477b seed regions. These biomarkers may alter the miRNA regulation of 133 PD-related genes. All these analyses reveal the potential mechanisms and regulations of RNA editing in PD pathogenesis.

Keywords: A-to-I RNA editing; Parkinson’s disease; gene expression; miRNA competition; miRNA regulation.

Figure 1

The flowchart of this study.

Figure 2

The overview of RNA editing effects on gene expressions. (A) The distributions of informative RNA editing events in genes, regions, and repeats. (B) Manhattan plots show the effects of 9897 RNA editing events on 6286 genes by QTL (FDR < 0.05) and Pearson correlation analyses (p < 0.05). The top and bottom panels present 137,563 positive and 483,277 negative associations, respectively. (C) The PD-related enrichment pathways of the editing-associated genes by DAVID. (D) The density distributions of editing-associated genes on chromosomes. The color bar shows the number of editing-associated genes in a 10 M bps (base pairs) region. (E) The types of editing–gene associations. Most RNA editing events are associated with distal genes (dist > 1 × 10⁶, trans-associations) and genes in different chromosomes.

Figure 3

A-to-I RNA editing events may alter miRNA regulations of their host genes. (A) Analysis procedures to uncover one mechanism of the associations between RNA editing events and their host genes. (B) The scatter plot shows the RNA editing events, which may alter miRNA regulations of their host genes. The editing example in EIF2AK2 is highlighted. (C) This RNA editing event is negatively associated with EIF2AK2 expression. EIF2AK2 is overexpressed in PD and in more severe PD samples. The RNA editing event shows lower editing frequencies in PD samples and negative associations with tau protein levels. (D) The 16 EIF2AK2-interacted genes are significantly associated with EIF2AK2 and the RNA editing event. Genes known to be implicated in PD are in bold. (E) The 16 genes are enriched in neurodegeneration-related processes by Metascape. (F) Of them, MCL1 is overexpressed in PD and in more severe PD samples. (G) The analyses above may uncover the potential effects of this RNA editing event on five PD-related genes.

Figure 4

A-to-I RNA editing events may alter miRNA competitions between their host genes and other genes. (A) Analysis procedures to uncover one possible mechanism of the trans-associations between RNA editing events and genes. (B) The enrichment results of the genes affected by editing-mediated miRNA competitions. (C) One RNA editing event in the 3′-UTR of APOL6 causes the binding loss of miRNAs and releases the miRNAs to regulate EEF1A1. (D) The frequency of this editing event is positively associated with the expression of APOL6 and negatively correlated to EEF1A1. APOL6 is highly expressed in severe PD samples. The editing event has higher frequencies in the samples with increased H&Y stages or tau protein levels. (E) The editing event may directly result in the overexpressions of APOL6 and indirectly cause the downregulations of 47 genes including EEF1A1 and 8 ribosome proteins. (F) The potential roles of this RNA editing event related to PD pathogenesis.

Figure 5

A-to-I RNA editing events may modify miRNA seed regions to disturb their regulations. (A) Analysis procedures to uncover the possible mechanisms for the associations between RNA editing events in miRNA seed regions and their targets. The editing example in miR-4477b is highlighted. (B) This RNA editing event alters the miRNA seed region, thus leading to the loss of miRNA binding on AGO2. (C) The frequencies of this RNA editing event are positively associated with the expressions of AGO2. The editing event shows higher frequencies in PD samples and positive associations with the levels of amyloid β-protein and α-synuclein. (D) The 46 AGO2-interacted RNAs are significantly associated with AGO2 and the editing event. The red and green circles describe the positive and negative associations, respectively. (E) The associated genes are enriched in the regulations of miRNAs and RNA binding proteins by Enrichr. (F) One PD-related gene, AIMP2, is positively associated with AGO2 and the editing event. (G) The analyses above may uncover the potential effects of this RNA editing event in miRNA seed region on five PD-related genes.

Figure 6

The effects of RNA editing events on genes in other datasets. (A,B) In blood samples of PD patients from PPMI and GEO (GSE165082), the RNA editing event at Chr1:160998027 shows significant associations with F11R. (C–E) In healthy blood samples, the three RNA editing events discussed in detail also show significant associations with corresponding genes. (F) In temporal cortex (TCX) regions of AD patients, the RNA editing event at Chr2:37104057 is significantly associated with EIF2AK2 and MCL1. (G,H) In cerebellum (CER) and dorsolateral prefrontal cortex (DLPFC) regions of AD patients, the RNA editing event (Chr9:63819627) in the seed region of miR-4477b may dysregulate AGO2 and AIMP2.