Dual-center, dual-platform microRNA profiling identifies potential plasma biomarkers of adult temporal lobe epilepsy

Abstract

BackgroundThere are no blood-based molecular biomarkers of temporal lobe epilepsy (TLE) to support clinical diagnosis. MicroRNAs are short noncoding RNAs with strong biomarker potential due to their cell-specific expression, mechanistic links to brain excitability, and stable detection in biofluids. Altered levels of circulating microRNAs have been reported in human epilepsy, but most studies collected samples from one clinical site, used a single profiling platform or conducted minimal validation.MethodUsing a case-control design, we collected plasma samples from video-electroencephalogram-monitored adult TLE patients at epilepsy specialist centers in two countries, performed genome-wide PCR-based and RNA sequencing during the discovery phase and validated findings in a large (>250) cohort of samples that included patients with psychogenic non-epileptic seizures (PNES).FindingsAfter profiling and validation, we identified miR-27a-3p, miR-328-3p and miR-654-3p with biomarker potential. Plasma levels of these microRNAs were also changed in a mouse model of TLE but were not different to healthy controls in PNES patients. We determined copy number of the three microRNAs in plasma and demonstrate their rapid detection using an electrochemical RNA microfluidic disk as a prototype point-of-care device. Analysis of the microRNAs within the exosome-enriched fraction provided high diagnostic accuracy while Argonaute-bound miR-328-3p selectively increased in patient samples after seizures. In situ hybridization localized miR-27a-3p and miR-328-3p within neurons in human brain and bioinformatics predicted targets linked to growth factor signaling and apoptosis.InterpretationThis study demonstrates the biomarker potential of circulating microRNAs for epilepsy diagnosis and mechanistic links to underlying pathomechanisms.

RNA-sequencing analysis of umbilical cord plasma microRNAs from healthy newborns

“…They were first profiled in human plasma, serum and microvescicles in 2008 [13][14][15] and since then, subsequent studies have found that their levels in peripheral biofluids often fluctuate in patients with various types of cancer, neurological disorders, sepsis, liver and cardiovascular disease (reviewed in [16][17][18]). As such miRNAs have received much interest as potential biomarkers and contain many characteristics which render them ideal biomarker candidates: they are more stable than mRNA as they are resistant to RNase cleavage [19]; expression profiles of miRNAs are often more informative and discriminatory than mRNA profiles; they are abundant and profiling miRNAs is rapid and economical [20]; and their levels often change more rapidly in response to an insult or pathophysiological processes allowing early detection of disease which is critical for progressive illnesses such as cancer, Alzheimer's disease, epilepsy and early life insults [21][22][23][24][25].…”

miRNAs as Biomarkers in Disease: Latest Findings Regarding Their Role in Diagnosis and Prognosis

“…Numerous authors have investigated this opportunity in various medical fields. Evidence suggests that they could play an essential role as biomarkers in cancer through exosome-mediated intercellular communication [1,4,[52][53][54], in neurology for the diagnosis and prognosis of Alzheimer's disease [55], for patients with spinal cord injury [56], epilepsy [57] or neurodegenerative ailments [58]. It could also be used in other fields like cardiology, as a faster and more accurate means of diagnosis for acute cardiovascular disease or heart failure [59][60][61] and in the case of infectious diseases for the diagnosis of sepsis [62].…”

Altered Biogenesis and MicroRNA Content of Hippocampal Exosomes Following Experimental Status Epilepticus

“…For miRNA target identification and Gene Ontology (GO) enrichment analysis, experimentally validated targets were retrieved from miRTarBase Release 7.0 (Chou et al, 2018), TarBase v.8 (Karagkouni et al, 2018), and miRecords (Xiao et al, 2009) while predicted targets were retrieved from TargetScan Release 7.2 (Agarwal et al, 2015) and miRDB Version 6.0 (Liu and Wang, 2019). We calculated a miRNA-target interaction (MTI) score based on combined prediction algorithm scores and the number of publications associated with the validated MTIs, as described previously (Raoof et al, 2018). Enrichment analysis of GO terms was performed on all MTIs with a score >0.1 using ReactomePA R/Bioconductor package (Yu and He, 2016).…”