MicroRNA editing alters target selection and regulated pathways in human temporal lobe epilepsy

Temporal lobe epilepsy (TLE) is a brain disease that is characterized by spontaneous recurrent seizures originating from the temporal lobe region. TLE is one of the most common form of drug-resistant epilepsy in adults and is frequently associated with specific neuropathological changes within the hippocampus including neuron loss and gliosis (hippocampal sclerosis). There is now extensive evidence that small noncoding RNAs called microRNAs (miRNAs) contribute to the molecular and cellular changes in experimental and human drug-resistant TLE. However, the regulatory mechanisms governing the change of these miRNAs in TLE are poorly understood. Adenosine deaminase acting on RNA (ADAR) is a well-characterised RNA editing mechanism in mammals that primarily edits adenosine (A) residues to inosine (I) in double-stranded RNA. Recently, RNA editing by ADARs has emerged as one of the mechanisms that modulate miRNA functions at a post-transcriptional level. Dysregulation of ADAR editing has been implicated in various neurological diseases, including epilepsy. The central hypothesis in the present thesis was that dysregulation of ADARs and their editing actions on miRNAs is a feature of the molecular pathophysiology of human TLE. 

In results chapter one (Chapter 3), the expression of ADAR family proteins was investigated in surgically-obtained neocortical and hippocampal samples from TLE patients and in mouse samples from the intra-amygdala kainic-acid model of TLE. This analysis identified brain region and ADAR isoform-specific differences in expression of ADARs in brain tissue from TLE patients, but not in the experimental model. This included higher level of ADAR1 in the neocortex and hippocampus of the TLE patients, while ADAR2 and ADAR3 were lower in the neocortex of TLE patients. We also used immunohistochemistry to examine the cell type(s) that express ADAR isoforms. This revealed that ADAR isoforms were expressed predominantly in neurons with limited or no expression in astrocytes and microglia. 

In results chapter two (Chapter 4), A-to-I miRNA editing was profiled by performing bioinformatic analysis on small RNA sequencing data from the hippocampus of TLE patients. This identified evidence of editing of 45 miRNAs detected in the hippocampus, ranging from 0.8 to 85 % editing levels detected in these miRNAs. Analysis of the location of these events revealed that miRNA editing was enriched within the seed region most critical for target engagement. However, there were very few differences between control and epilepsy samples in miRNA editing. We identified one miRNA that was differentially edited within the seed region, which was miR-376a-3p at site +6, where the editing level was significantly reduced in patients compared to controls. Target analysis predicted shift in miRNA target repertoire when the edited site +6 became unedited. Taken together, these studies represent the first identification of differentially edited miRNAs in TLE and suggest single base changes to miRNAs can bring about extensive reorganisation of the target pool. Thus, miRNA editing represents a highly efficient cellular mechanism to produce small changes to a wide range of gene transcripts. 

In results chapter three (Chapter 5), we validated the reduction of editing activity at site +6 of miR-376a-3p in a separate cohort of TLE patients using Sanger sequencing and TaqMan assay. Next, an in vitro assay using knockdown ADARs revealed that ADAR1 is likely to be the primary deaminase enzyme responsible for the editing at site +6 of miR-376a-3p. In the final experiments, antisense inhibitors (AntagomiRs) were delivered into human induced-pluripotent stem cells (hiPSCs) derived neurons to selectively deplete either the edited (anta-E) or unedited (native, anta-N) forms of the miRNA. RNA sequencing was then used to map the impact of the gene expression landscape. This revealed enrichments of mitochondrial function-related genes among up-regulated transcripts in samples from neurons treated with anta-E. Thus, the reduced editing of miR-376a-3p may result in metabolic reprogramming which may influence the enduring state of hyperexcitability or neuropathological changes in human TLE. 

In summary, these data provide strong evidence that ADAR proteins are altered in the brain tissue of TLE patients and miRNAs undergo specific differential ADAR editing in the hippocampus of TLE patients. Furthermore, reductions of edited miR-376a-3p in TLE dysregulates sets of genes that are important for mitochondrial functions in maintaining normal brain physiology. The findings expand our understanding of the molecular landscape of the human brain in epilepsy and point to novel mechanisms of disease, potential tissue-based biomarkers and therapeutic targets.