Neuronally enriched microvesicle RNAs are differentially expressed in the serums of Parkinson's patients

Abstract

Background: Circulating small RNAs (smRNAs) originate from diverse tissues and organs. Previous studies investigating smRNAs as potential biomarkers for Parkinson's disease (PD) have yielded inconsistent results. We investigated whether smRNA profiles from neuronally-enriched serum exosomes and microvesicles are altered in PD patients and discriminate PD subjects from controls.

Methods: Demographic, clinical, and serum samples were obtained from 60 PD subjects and 40 age- and sex-matched controls. Exosomes and microvesicles were extracted and isolated using a validated neuronal membrane marker (CD171). Sequencing and bioinformatics analyses were used to identify differentially expressed smRNAs in PD and control samples. SmRNAs also were tested for association with clinical metrics. Logistic regression and random forest classification models evaluated the discriminative value of the smRNAs.

Results: In serum CD171 enriched exosomes and microvesicles, a panel of 29 smRNAs was expressed differentially between PD and controls (false discovery rate (FDR) < 0.05). Among the smRNAs, 23 were upregulated and 6 were downregulated in PD patients. Pathway analysis revealed links to cellular proliferation regulation and signaling. Least absolute shrinkage and selection operator adjusted for the multicollinearity of these smRNAs and association tests to clinical parameters via linear regression did not yield significant results. Univariate logistic regression models showed that four smRNAs achieved an AUC ≥ 0.74 to discriminate PD subjects from controls. The random forest model had an AUC of 0.942 for the 29 smRNA panel.

Conclusion: CD171-enriched exosomes and microvesicles contain the differential expression of smRNAs between PD and controls. Future studies are warranted to follow up on the findings and understand the scientific and clinical relevance.

Keywords: Parkinson’s disease; biomarkers; exosome; microvesicle; random forest; regression; small RNA.

Figure 1

Validation of exosomes isolated from serum sample via bioanalyzer, transmission electron microscopy (TEM), and nanoparticle Tracking Analysis (NTA). (A) Schematic of the study design. (B) Morphological characterization of exosomes isolated from serum samples by transmission electron microscopy. Bar, 200 nm. (C) Bioanalyzer report of smRNA fragment length distribution. (D) Graph generated by NanoSight showing concentration and average size of the vesicles after ExoQuick precipitation showing exosomes collected in the 100–400 nm range.

Figure 2

Altered smRNAs profiles in neuronally-enriched serum-derived exosomes of PD by RNA-Seq Analysis. (A) MDS analyzes the smRNA read counts from PD and control subjects. (B) Unsupervised hierarchical clustering of the 29 differentially expressed smRNAs (rows) for the 45 PD and 28 control samples (columns). The sample ID and the top column in the annotation bars indicating the disease state of each subject in a color code (grey for controls and black for PD cases) are shown at the top of the plot. The other annotation bars illustrate age, sex, education, smoking status, disease stages, disease duration, MDS-UPDRS, MoCA, twin pairs, and the expression level of smRNAs across all samples. Not available values are shown in white. (C) qRT-PCR was used to confirm the level of miR-26b-5p. Relative levels of one miRNA (miR-26b-5p) in the serum of PD cases and controls were determined by qRT-PCR. Each point represents the mean of triplicate samples. ***p < 0.001.

Figure 3

Random forest model diagnostic values and AUC curves. Variable importance plots obtained from Random Forest (RF) in R show the top smRNAs ranked based on the mean decrease in Gini coefficients. The ROC curve plot and the AUC value display the RF model’s performance in discriminating between PD and controls.

Figure 4

Gene pathways. The network displays the biological effects of the smRNAs on genes and other smRNAs. The edges with a solid arrowhead represent the direction of effects, and a blunted arrowhead represents inhibition. The input for the gene path analyses were the 29 differentially expressed smRNAs that were statistically significant.