Combinatory microRNA serum signatures as classifiers of Parkinson's disease

Abstract

As current clinical diagnostic protocols for Parkinson's disease (PD) may be prone to inaccuracies there is a need to identify and validate molecular biomarkers, such as circulating microRNAs, which will complement current practices and increase diagnostic accuracy. This study identifies, verifies and validates combinatory serum microRNA signatures as diagnostic classifiers of PD across different patient cohorts. Methods 370 PD (drug naïve) and control serum samples from the Norwegian ParkWest study were used for identification and verification of differential microRNA levels in PD which were validated in a blind study using 64 NY Parkinsonism in UMeå (NYPUM) study serum samples and tested for specificity in 48 Dementia Study of Western Norway (DemWest) study Alzheimer's disease (AD) serum samples using miRNA-microarrays, and quantitative (q) RT-PCR. Proteomic approaches identified potential molecular targets for these microRNAs. Results Using Affymetrix GeneChip® miRNA 4.0 arrays and qRT-PCR we comprehensively analyzed serum microRNA levels and found that the microRNA (PARKmiR)-combinations, hsa-miR-335-5p/hsa-miR-3613-3p (95% CI, 0.87–0.94), hsa-miR-335-5p/hsa-miR-6865-3p (95% CI, 0.87–0.93), and miR-335-5p/miR-3613-3p/miR-6865-3p (95% CI, 0.87–0.94) show a high degree of discriminatory accuracy (AUC 0.9–1.0). The PARKmiR signatures were validated in an independent PD cohort (AUC ≤ 0.71) and analysis in AD serum samples showed PARKmiR signature specificity to PD. Proteomic analyses showed that the PARKmiRs regulate key PD-associated proteins, including alpha-synuclein and Leucine Rich Repeat Kinase 2. Conclusions Our study has identified and validated unique miRNA serum signatures that represent PD classifiers, which may complement and increase the accuracy of current diagnostic protocols.