Abstract
Neurodegenerative diseases (NDs) are age-dependent; among them, Alzheimer's disease (AD) and Parkinson's disease (PD) are the most frequent. Similarly, cerebrovascular damage can induce the development of vascular-related disorders that share common features with AD and PD, respectively, named vascular dementia (VD) and vascular parkinsonism (VP). To date, ND diagnosis is mainly clinical; therefore, since these disorders show similar symptoms, their correct discrimination may be difficult. We detected 23 ND-associated microRNAs (miRNAs) by literature mining and investigated their serum expression in a cohort of 139 patients including AD, PD, VD, and VP patients and healthy controls. TaqMan RT-PCR data showed that miR-23a upregulation was associated with an ongoing neurodegenerative process, similar to miR-22* and miR-29a, while let-7d, miR-15b, miR-24, miR-142-3p, miR-181c, and miR-222 showed an altered expression in Parkinson-like phenotypes, as well as miR-34b, miR-125b, and miR-130b in Alzheimer-like disorders. By computing logistic regression models and ROC curves, we identified signatures of neuro-miRNAs specific for each disease, showing good diagnostic performance. Interestingly, we found that miR-23a, miR-29a, miR-34b, and miR-125b exhibited a different distribution between exosomes and vesicle-free serum, suggesting a heterogeneity of secretion for these miRNAs. Our results suggest that miRNA signatures could discriminate in a non-invasive manner neurodegenerative disorders, thus improving clinical diagnoses.