Direct detection of circulating microRNA-122 using dynamic chemical labelling with single molecule detection overcomes stability and isomiR challenges for biomarker qualification

Abstract

Circulating microRNAs are biomarkers reported to be stable and translational across species. miR-122 (miR-122-5p) is a hepatocyte-specific microRNA biomarker for drug-induced liver injury (DILI). Our objective was to develop an extraction-free and amplification-free detection method for measuring miR-122 that has translational utility in context of DILI. We developed a single molecule dynamic chemical labelling (DCL) assay based on miR-122 hybridization to an abasic peptide nucleic acid probe that contained a reactive amine instead of a nucleotide at a specific position in the sequence. The single molecule DCL assay specifically measured miR-122 directly from 10 {micro}L of serum or plasma without any extraction steps, with a fit-for-purpose limit of detection of 1.32 pM. In 192 human serum samples, DCL accurately identified patients at risk of DILI (area under ROC curve 0.98 (95%CI 0.96-1), P<0.0001). The miR-122 assay also quantified liver injury in rats and dogs. When DCL beads were added to serum, the miR-122 signal was stabilised (no loss of signal after 14 days at room temperature). By contrast, there was substantial degradation of miR-122 in the absence of beads ({approx}60% lost in 1 day). RNA sequencing demonstrated the presence of multiple miR-122 isomiRs with DILI that were at low concentration or not present in healthy patient serum. Sample degradation over time produced more isomiRs, particularly rapidly with DILI. PCR was inaccurate when analysing miR-122 isomiRs, whereas the DCL assay demonstrated accurate quantification. In summary, the DCL assay can accurately measure miR-122 directly from serum and plasma to diagnose liver injury in humans and other species, and can overcome important microRNA biomarker analytical and biological challenges.