A Systematic Review of Common and Brain-Disease-Specific RNA Editing Alterations Providing Novel Insights into Neurological and Neurodegenerative Disease Manifestations

Abstract

RNA editing contributes to transcriptome diversification through RNA modifications in relation to genome-encoded information (RNA-DNA differences, RDDs). The deamination of Adenosine (A) to Inosine (I) or Cytidine (C) to Uridine (U) is the most common type of mammalian RNA editing. It occurs as a nuclear co- and/or post-transcriptional event catalyzed by ADARs (Adenosine deaminases acting on RNA) and APOBECs (apolipoprotein B mRNA editing enzyme catalytic polypeptide-like genes). RNA editing may modify the structure, stability, and processing of a transcript. This review focuses on RNA editing in psychiatric, neurological, neurodegenerative (NDs), and autoimmune brain disorders in humans and rodent models. We discuss targeted studies that focus on RNA editing in specific neuron-enriched transcripts with well-established functions in neuronal activity, and transcriptome-wide studies, enabled by recent technological advances. We provide comparative editome analyses between human disease and corresponding animal models. Data suggest RNA editing to be an emerging mechanism in disease development, displaying common and disease-specific patterns. Commonly edited RNAs represent potential disease-associated targets for therapeutic and diagnostic values. Currently available data are primarily descriptive, calling for additional research to expand global editing profiles and to provide disease mechanistic insights. The potential use of RNA editing events as disease biomarkers and available tools for RNA editing identification, classification, ranking, and functional characterization that are being developed will enable comprehensive analyses for a better understanding of disease(s) pathogenesis and potential cures.

Keywords: RNA editing; brain disorders; neurodegenerative diseases.

Figure 1

ADAR and APOBEC enzyme members catalyze RNA editing in mammals leading to epi-transcriptomic alterations. (A) The mammalian ADAR family comprises three members: ADAR1, ADAR2 and ADAR3. The first two are catalytically active and widely expressed; ADAR3 is expressed exclusively in the brain, has no proven catalytic activity, and is suggested to act as an ADAR1 and ADAR2 regulator. ADARs catalyze A-I editing in the form of homo- and/or heterodimers, without the requirement of other co-factors. Complementary or partially double-stranded RNAs may be used as ADARs’ substrates. Any dsRNA ≥ 20 bp, including protein-coding exons in pre-mRNAs, repetitive sequence elements, as well as microRNA (miRNA) precursor transcripts, may be ADAR substrates. ADARs deaminate Adenosine (A) to Inosine (I). The cellular transcriptional and translational machinery recognizes Inosine (I) as Guanine (G); thus, processed ADAR edited transcripts display a G at the edited site (A-I-G editing). (B) APOBEC1 is the main C-U editing enzyme in mammals. APOBEC1-mediated editing is highly specific and requires the formation of the editosome, a protein complex that comprises an enzyme homodimer, an essential co-factor (A1CF or RBM47) and auxiliary proteins that regulate enzymatic activity. APOBEC1 targets are ssRNAs and display specific sequence elements, corresponding to the mooring sequence (an 11 nt consensus sequence located downstream the C undergoing deamination, required for A1CF binding) and an AU-rich ‘efficiency region’, located upstream of the edited residue. APOBECs deaminate Cytidine (C) to Uridine (U). The cellular transcriptional and translational machinery recognizes Uridine (U) as Thymine (T); thus, processed APOBEC edited transcripts display a T at the edited site (C-U-T editing). (C) RNA editing events may occur at several sites within a transcript and affect stability, processing, and function of the edited target. Editing in intronic regions or close to splice junction sites may cause the alternative splicing of the edited transcripts. RNA editing events within non-coding regions (5′UTR, 3′UTR) may affect transcript stability and regulation. Editing events within a miRNA seed sequence may re-direct miRNA targeting and cause degradation. Introduction of non-synonymous changes within a coding region of a transcript results in amino acid alterations (protein recoding, stop codon introduction) and may also affect a transcript’s secondary structure and alter its interactions with RNA-binding proteins (RBPs). LncRNA editing can lead to changes in secondary structure affecting its regulatory functions. Figure created with BioRender.com (accessed on 31 January 2022).

Figure 2

Differential editing patterns in different human neuropsychiatric, autoimmune/neurodegenerative and neurological brain disorders compared to healthy individuals. (A) Venn diagrams depicting differentially edited sites identified in schizophrenia (SCZ), genetic ALS (ALS c9orf72) and autism. The intersections represent editing events identified as differentially edited in more than one condition. Twenty-six editing sites have been commonly identified as differentially edited in all three conditions. (Β) Venn diagrams depict differentially edited transcripts (targets) in the same human disorders as in A. The intersections represent editing targets identified as differentially edited in more than one condition. Please note that a higher number of transcripts (187) are differentially edited in all three conditions compared to common, individual editing events (intersection in A), suggesting disease-associated editing that converges in common targets. (C) Gene ontology (GO) analysis of the differentially edited targets in all three human brain disorders (n = 187). Enrichment values are given as –log (p) values (higher –log (p) values indicate greater statistical significance). Differentially edited transcripts are involved in molecular processes associated with hypoxia, synaptic transmission, endosome/lysosome function, cytoskeleton, apoptosis, protein and RNA processing, all of which have been reported as deregulated in these disorders. Differential editing data (p < 0.05) reported in the studies of [35,45,51] were utilized for the analysis. Reported data correspond to different brain regions.

Figure 3

Differentially edited targets in human neuropsychiatric, autoimmune/neurodegenerative, neurological disorders [35,45,51] and in neurological/neurodegenerative disease mouse models [31,43,57] compared to healthy controls. (A) Venn diagram depicting differentially edited transcripts identified in humans and mice compared to healthy controls. The intersection represents editing targets commonly identified as DE in human and mouse brain disorders. (B,C) Gene ontology (GO) analysis of the differentially edited targets in human and mouse brain disorders (n = 34). Enrichment values are given as –log (p) values (higher –log (p) values indicate greater statistical significance). (B) Biological process enrichment. Common DE transcripts are involved in molecular processes associated with synaptic transmission and signaling, regulation of cell communication, pattern recognition receptor signaling, response to thyroid signaling and protein catabolism. (C) Cellular component enrichment. Main axon, glutamate receptor, cation and sodium channel complexes, lytic vacuoles and specifically endo/lysosome lumen, as well as filipodium tip, emerge as cellular component terms enriched in DE transcripts in both human and mouse brain disorders.