Common signatures of differential microRNA expression in Parkinson’s and Alzheimer’s disease brains

Abstract Dysregulation of microRNA gene expression has been implicated in many neurodegenerative diseases, including Parkinson’s disease. However, the individual dysregulated microRNAs remain largely unknown. Previous meta-analyses have highlighted several microRNAs being differentially expressed in post-mortem Parkinson’s disease and Alzheimer's disease brains versus controls, but they were based on small sample sizes. In this study, we quantified the expression of the most compelling Parkinson’s and Alzheimer’s disease microRNAs from these meta-analyses (‘candidate miRNAs’) in one of the largest Parkinson’s/Alzheimer’s disease case–control post-mortem brain collections available (n = 451), thereby quadruplicating previously investigated sample sizes. Parkinson’s disease candidate microRNA hsa-miR-132-3p was differentially expressed in our Parkinson’s (P = 4.89E−06) and Alzheimer’s disease samples (P = 3.20E−24) compared with controls. Alzheimer’s disease candidate microRNAs hsa-miR-132-5p (P = 4.52E−06) and hsa-miR-129-5p (P = 0.0379) were differentially expressed in our Parkinson’s disease samples. Combining these novel data with previously published data substantially improved the statistical support (α = 3.85E−03) of the corresponding meta-analyses, clearly implicating these microRNAs in both Parkinson’s and Alzheimer’s disease. Furthermore, hsa-miR-132-3p/-5p (but not hsa-miR-129-5p) showed association with α-synuclein neuropathological Braak staging (P = 3.51E−03/P = 0.0117), suggesting that hsa-miR-132-3p/-5p play a role in α-synuclein aggregation beyond the early disease phase. Our study represents the largest independent assessment of recently highlighted candidate microRNAs in Parkinson’s and Alzheimer’s disease brains, to date. Our results implicate hsa-miR-132-3p/-5p and hsa-miR-129-5p to be differentially expressed in both Parkinson’s and Alzheimer’s disease, pinpointing shared pathogenic mechanisms across these neurodegenerative diseases. Intriguingly, based on publicly available high-throughput sequencing of RNA isolated by cross-linking immunoprecipitation data, hsa-miR-132 may interact with SNCA messenger RNA in the human brain, possibly pinpointing novel therapeutic approaches in fighting Parkinson’s disease.


Supplementary Table 1.
Overview of the Parkinson's disease case-control post-mortem brain samples analyzed in this study Legend. Post-mortem brain samples from the superior temporal gyrus of Parkinson's (PD) patients and corresponding controls were provided by the Parkinson's UK Brain Bank at Imperial College London. SD = standard deviation; IQR = interquartile range; PMI = post-mortem interval; RIN = RNA integrity number; a = Pearson's Chi-squared test with Yates' Supplementary note. Summary of previous studies on hsa-miR-129-5p and hsa-miR-132-3p/5p in the context of Parkinson's disease following a systematic literature review

hsa-miR-129-5p
a) Differential expression of hsa-miR-129-5p in post-mortem Parkinson's disease and control brains: Sample sizes in previous studies were small with 11, 62, and 32 combined Parkinson's disease cases and controls, and none of the studies had reported hsa-miR-129-5p as differentially expressed. [2][3][4] Our previous meta-analysis combining these small studies was based on a total of 58 Parkinson's disease and 47 control brains (total n=105), and showed nominal significant differential downregulation in Parkinson's disease vs control brains (p=7.74E-04), which did not survive multiple testing correction. 5 After the data freeze of our previous meta-analysis, one more study including 8 Parkinson's disease and 8 control brains was published. 6 This study was included in the updated metaanalysis performed in the current study.

b) Association with Parkinson's disease neuropathology as marker of disease progression:
We did not identify any studies reporting on the quantification of hsa-miR-129-5p levels in any tissue of Parkinson's disease patients (brain/blood/CSF) with Parkinson's disease neuropathology (Parkinson's disease Braak staging) in brain.

c) Quantifications of hsa-miR-129-5p in other human Parkinson's disease studies:
Hsa-miR-129-5p was reported to be increased in peripheral blood lymphocytes in 18 treated vs 20 untreated Parkinson's disease patients. 7 A recent NGS-based miRNA screen in plasma samples of Parkinson's disease and control blood samples assessed association with clinical disease progression but did not report on individual miRNAs. 8

d) Evidence from other in vitro and in vivo functional studies:
We did not identify any in vitro nor in vivo studies assessing the role of hsa-miR-129-5p in Parkinson's disease cellular or animal models.

=> Interpretation:
Our study is the first to report a significant downregulation of hsa-miR-129-5p in Parkinson's disease brain vs. controls and to assess the association of hsa-miR-129-5p expression in brain with neuropathological Parkinson's disease Braak staging. Functional studies investigating the molecular mechanisms of hsa-miR-129-5p are currently missing.
b) Association with Parkinson's disease neuropathology as marker of disease progression: We did not identify any study reporting association analyses of hsa-miR-132-3p/-5p expression in brain and Parkinson's disease neuropathology (Parkinson's disease Braak staging). Burgos et al. reported that hsa-miR-132-5p expression in CSF decreased with advanced Lewy body pathology in the brain, although this study was not limited to Parkinson's disease but also analyzed Alzheimer's disease cases and controls. 11 Furthermore, in serum samples of Parkinson's disease patients, expression of hsa-miR-132-3p (and hsa-miR-146-5p) was reported to be negatively correlated with Braak staging. 12

c) Quantifications of hsa-miR-132-3p/5p in other human Parkinson's disease studies:
As for hsa-miR-129-5p (see above), hsa-miR-132-3p was reported to be increased in peripheral blood lymphocytes of 18 treated vs 20 untreated Parkinson's disease patients. 7 A recent NGS-based miRNA screen in plasma samples of Parkinson's disease and control blood samples assessed association with clinical disease progression but did not report on individual miRNAs 8 .

d) Evidence from other in vitro and in vivo functional studies:
Hsa-miR-132-3p has been implicated in a variety of functional domains relevant for the human brain, including neuroprotection, memory, neural growth, synapse function, apoptosis, and inflammation (e.g., reviewed in ref. [13][14][15] ). For Parkinson's disease, several different molecular mechanisms were suggested as a possible link between hsa-miR-132-3p/5p dysregulation and Parkinson's disease: In an cell model, hsa-mir-132-3p was reported to activate the SIRT1/p53 pathway resulting in neuronal death 16 and to downregulate GLRX with subsequent promotion of neuroinflammation 17 . Inhibition of hsa-miR-132-5p in both cell and mouse models reduced apoptosis and suppressed autophagy, presumably by targeting ULK1. 18 Finally, simultaneous downregulation of miR-212-3p and miR-132-3p was reported in alpha-synuclein (A30P)-transgenic mice. 19 It is important to emphasize that these reports still need to be validated independently.
=> Interpretation: Downregulation of hsa-miR-132-3p was reported in Parkinson's disease brains vs controls previously, but sample sizes were very small. Our study now confirms this finding using a much larger independent dataset. To the best of our knowledge, no previous study has investigated the association of hsa-miR-132-3p/5p gene expression in brain and neuropathological Parkinson's disease phenotypes, such as alpha-synuclein staging. Thus, our study is the first to report such an association for hsa-miR-132-3p/5p.
Functional evidence of hsa-miR-132-3p/5p points towards an important role of this miRNA in various neuronal processes; additional studies in Parkinson's disease are needed to determine the exact underlying mechanisms.