Extracellular vesicle microRNA-mediated transcriptional regulation may contribute to dementia with Lewy bodies molecular pathology

Abstract

Objective:

Dementia with Lewy bodies (DLB) is the second most common dementia. Advancing our limited understanding of its molecular pathogenesis is essential for identifying novel biomarkers and therapeutic targets for DLB. DLB is an α-synucleinopathy, and small extracellular vesicles (SEV) from people with DLB can transmit α-synuclein oligomerisation between cells. Post-mortem DLB brains and serum SEV from those with DLB share common miRNA signatures, and their functional implications are uncertain. Hence, we aimed to investigate potential targets of DLB-associated SEV miRNA and to analyse their functional implications.

Methods:

We identified potential targets of six previously reported differentially expressed miRNA genes in serum SEV of people with DLB (MIR26A1, MIR320C2, MIR320D2, MIR548BA, MIR556, and MIR4722) using miRBase and miRDB databases. We analysed functional implications of these targets using EnrichR gene set enrichment analysis and analysed their protein interactions using Reactome pathway analysis.

Results:

These SEV miRNA may regulate 4278 genes that were significantly enriched among the genes involved in neuronal development, cell-to-cell communication, vesicle-mediated transport, apoptosis, regulation of cell cycle, post-translational protein modifications, and autophagy lysosomal pathway, after Benjamini-Hochberg false discovery rate correction at 5%. The miRNA target genes and their protein interactions were significantly associated with several neuropsychiatric disorders and with multiple signal transduction, transcriptional regulation, and cytokine signalling pathways.

Conclusion:

Our findings provide in-silico evidence that potential targets of DLB-associated SEV miRNAs may contribute to Lewy pathology by transcriptional regulation. Experimental validation of these dysfunctional pathways is warranted and could lead to novel therapeutic avenues for DLB.

Significant outcomes

• Differentially expressed miRNA in serum SEV from people with DLB may contribute to the molecular pathogenesis of DLB by regulating expression of genes involved in neuronal development, cell-to-cell communication, vesicle-mediated transport, apoptosis, regulation of cell cycle, post-translational protein modifications, and autophagy lysosomal pathway.

• Potential target genes of the DLB-associated SEV miRNA were significantly enriched among the genes involved in axon guidance, focal adhesion, endocytosis, autophagy, mRNA surveillance, long-term potentiation, ubiquitin-mediated proteolysis, cholinergic and glutaminergic synapses, and neurotrophin, Wnt, mTOR, ErbB, and FoxO signalling pathways.

• Protein interactions of the potential target genes may contribute to DLB molecular pathogenesis by regulating cell cycle, protein phosphorylation, interleukin-6 signalling, transcriptional regulation, and post-transcriptional silencing by small RNA.

Limitations

• This study did not investigate the potential targets of other classes of SEV non-coding small RNA that were differentially expressed in people with DLB.

• miRNA target prediction is still an evolving scientific discipline.

• Further experimental validation of identified dysfunctional molecular processes and pathways is needed.

Introduction

Dementia with Lewy bodies (DLB) is the second most common neurodegenerative dementia (Rajkumar and Aarsland, 2020). Lewy bodies are intraneuronal eosinophilic cytoplasmic inclusion bodies. Post-mortem neuropathological examination of brains of people, who were clinically diagnosed to have DLB antemortem, often reveals Lewy bodies in brainstem, limbic system, and/or cerebral cortex (Garcia-Esparcia et al., 2017). DLB is a primary α-synucleinopathy (Rajkumar and Aarsland, 2020), and α-synuclein aggregation is widely recognised as the key initial step in the formation of Lewy bodies (Beyer et al., 2009). DLB causes earlier mortality (Oesterhus et al., 2014), earlier nursing home admissions (Rongve et al., 2014), poorer quality-of-life (Bostrom et al., 2007), more frequent falls, higher care costs (Vossius et al., 2014), and more caregivers’ burden (Svendsboe et al., 2016) than Alzheimer’s disease (AD). Accurate distinction between DLB and AD is essential because they differ in their prognosis, neuropsychiatric symptoms, and pharmacological management (Watts et al., 2023). Both AD and DLB are currently diagnosed by their clinical diagnostic criteria (McKeith et al., 2017; WHO, 2018), and DLB is often misdiagnosed as AD or other dementia in clinical settings (Freer, 2017). Failing to diagnose DLB early, and treating visual hallucinations and challenging behaviours, which are more frequent in DLB than AD, with any antipsychotic medication risk potentially fatal adverse events like neuroleptic malignant syndrome. Reliable biomarkers from clinically collectable biological fluids (Ashton et al., 2020) and safer effective treatments (Stinton et al., 2015; Velayudhan et al., 2017; Watts et al., 2023) are urgently needed for DLB. Improving our current limited understanding of DLB molecular pathology is prerequisite for addressing this need and for improving clinical diagnosis and management of DLB (Zhang et al., 2015).

Extracellular vesicles (EV) are cell-derived vesicles with lipid bilayer membrane that cannot replicate (Thery et al., 2018). They are secreted into extracellular environment (Yanez-Mo et al., 2015; Ibrahim and Khan, 2022), and they play an important role in communication between cells (Saint-Pol et al., 2020). EV carry biologically active proteins, lipids and nucleic acids including messenger RNA (mRNA), microRNA (miRNA), other small non-coding RNA, mitochondrial DNA, and genomic DNA fragments from their cells of origin (Trotta et al., 2018). Because of their potential for transferring biological information, several drug delivery systems using EV are being developed (Vader et al., 2016; Mustajab et al., 2022). EV are usually classified on the basis of their physical or biochemical characteristics. The Minimal information for studies of extracellular vesicles (MISEV2018) guidelines (Thery et al., 2018) encourage using the term ’small extracellular vesicles’ (SEV) for the EV that are smaller than 200 nm in size. SEV consist of heterogeneous population of EV including those called as exosomes (Kalluri and LeBleu, 2020).

SEV RNA cargo includes miRNA that can regulate expression of thousands of genes (Lu and Rothenberg, 2018). miRNA are approximately 22 nucleotides-long single-stranded non-coding RNA (Liu et al., 2014). Several hundred nucleotides-long miRNA precursors, called pri-miRNA, are transcribed from miRNA genes usually by RNA polymerase-II. The pri-miRNA are processed first in nuclei and then in cytoplasm to generate small mature miRNA that regulate transcription of other genes (Vishnoi and Rani, 2023). miRNA exert important regulatory control over neuronal development and homeostasis (Liu et al., 2014). Differentially expressed EV miRNA have been associated with neurodegenerative diseases such as Alzheimer’s disease (Gamez-Valero et al., 2019; Wang et al., 2022), demyelinating diseases like multiple sclerosis (Ovchinnikova et al., 2022), and psychiatric disorders such as schizophrenia and bipolar disorder (Du et al., 2019; Gruzdev et al., 2019).

SEV, separated from cerebrospinal fluid (CSF) of people with DLB, has been demonstrated to induce alpha-synuclein oligomerisation in H4 neuro-glioma cells (Stuendl et al., 2016). Moreover, it has been shown that injection of brain-derived SEV from people with DLB into the brains of mice could induce α-synuclein oligomerisation (Ngolab et al., 2017). As SEV may transmit disease-associated nucleic acids and proteins between cells (Candelario and Steindler, 2014), we hypothesise that differentially expressed SEV miRNA from people with DLB may contribute to DLB pathology by regulating expression of genes involved in the molecular pathways that are known to be associated with DLB. We have already reported an RNA sequencing (RNA-Seq) study that investigated differentially expressed genes (DEG) in post-mortem anterior cingulate and dorsolateral prefrontal cortices of people with DLB (Rajkumar et al., 2020) and another RNA-Seq study that investigated serum SEV DEG in people living with DLB (Rajkumar et al., 2021). Four miRNA genes, MIR320C2, MIR320D2, MIR548BA and MIR4722, were found significantly (p < 0.05) differentially expressed in both post-mortem DLB brains and serum SEV from people living with DLB (Rajkumar et al., 2020; Rajkumar et al., 2021). Differential expression levels of two more serum SEV miRNA DEG in DLB, MIR26A1 and MIR556, could be replicated by high-throughput quantitative polymerase chain reaction (qPCR) (Rajkumar et al., 2021). MIR26A1 and MIR556 were significantly upregulated, and other four miRNA genes were significantly downregulated in serum SEV from people living with DLB. In this study, we aimed to investigate the potential targets of the miRNA, transcribed by these six miRNA genes, and to analyse their downstream functional implications.

Material and methods

Differentially expressed serum SEV miRNA in DLB

Our prior RNA-Seq study investigated serum SEV RNA profiles in people with DLB (n = 10) and age- and gender-matched comparisons (n = 10). That study obtained the 20 serum samples from the biobanks of three Norwegian cohorts that have obtained generic ethical approval for further studies using their serum samples (Selnes et al., 2013; Rongve et al., 2016; Fladby et al., 2017). Serum samples from 10 people living with DLB and three gender- and age (±3 years)-matched comparisons without cognitive impairment or Parkinson’s disease were obtained from the dementia study of western Norway (DemWest) (Rongve et al., 2016). Another seven gender- and age-matched comparison serum samples were obtained from two cohorts at Akershus University Hospital dementia research centres (Selnes et al., 2013; Fladby et al., 2017). 70% (n = 14) of the samples were from men. The mean age of the people with DLB was 77.3 (SD = 3.7) years, and the mean age of the people in comparison group was 76.8 (SD = 4.1) years. Further details of the cohorts have been published elsewhere (Selnes et al., 2013; Rongve et al., 2016; Fladby et al., 2017).

Our prior study separated SEV using an ultracentrifugation and OptiPrep^TM (Sigma-Aldrich, UK) density gradient approach based on their buoyant density. Size distribution and concentration of separated EVs were verified using the Malvern NanoSight LM10 nanoparticle analyser (Malvern Instruments Ltd., UK). Separation of SEVs was confirmed by Western blotting using antibodies against Flotillin-1 and CD63. Total RNA was isolated from SEV, and RNA samples were sequenced using the Illumina Hiseq-2500. We identified DEG using a previously experimentally validated edgeR algorithm (Anders et al., 2013; Rajkumar et al., 2015). Further details of this study have been published elsewhere (Rajkumar et al., 2021). Those RNA-Seq raw data files are available at https://www.ncbi.nlm.nih.gov/bioproject/PRJNA530121.

MiRNA target discovery

We first identified the mature miRNA sequences, transcribed by the six differentially expressed miRNA genes (MIR26A1, MIR320C2, MIR320D2, MIR548BA, MIR556, and MIR4722) in DLB serum SEV, using the miRbase database (Liu et al., 2014). ‘MiRbase’ is a comprehensive database for miRNA sequences and annotations (Kozomara et al., 2019). Expression of these six miRNA genes leads to nine mature miRNA sequences (hsa-miR-26a-5p, hsa-miR-26a-1-3p, hsa-miR-320c, hsa-miR-320d, hsa-mir-548ba, hsa-miR-556-5p, hsa-miR-556-3p, hsa-miR-4722-5p, and hsa-miR-4722-3p). We next identified potential targets of these miRNA using the MiRDB miRNA target prediction database (Liu and Wang, 2019). MiRDB analyses miRNA-target interactions using MirTarget2 bioinformatics tool and data derived from high-throughput RNA sequencing experiments (Liu and Wang, 2019; Chen and Wang, 2020). miRTarget2 can perform genome-wide miRNA target prediction using support vector machine framework (Benson et al., 2007; Maglott et al., 2007; Wang and El Naqa, 2008; Chen et al., 2013). We combined all potential target genes of these miRNA and removed duplicates (Supplementary information-1).

Functional enrichment analyses

We employed the EnrichR functional enrichment analysis tool and identified the biological processes, molecular pathways, and disease processes that were significantly enriched among the potential target genes of these miRNA (Chen et al., 2013; Kuleshov et al., 2016). We repeated the EnrichR analyses for understanding the functional implications of the potential target genes of each miRNA gene of interest individually. EnrichR considers at least 204 gene set libraries that are constructed from prior published studies and major biological and biomedical online databases including BioCarta, the database of Genotypes and Phenotypes (dbGaP), DisGeNET (Pinero et al., 2021), GeneSigDB (Culhane et al., 2012), Online Mendelian Inheritance in Man (OMIM), and UK Biobank genome-wide association study EnrichR includes Kyoto Encyclopaedia of Genes and Genomes (KEGG) pathway information (Kanehisa and Goto, 2000; Kanehisa et al., 2021). EnrichR calculates its p-values using Fisher’s exact test and employs appropriate corrections for multiple testing using the Benjamini-Hochberg false discovery rate (FDR) correction at 5% (q-values) (Benjamini and Hochberg, 1995).

Protein interactions of potential target genes

Interactome implies the total molecular interactions that occur within a cell (Yan et al., 2018), and interactome analyses map these networks. Interactome analyses help improving our understanding of molecular pathways involved in neurodegenerative diseases (Haenig et al., 2020). First, we identified significantly enriched molecular interaction pathways of the potential miRNA target genes using the Reactome pathway database (Fabregat et al., 2017). We then expanded the Reactome enrichment analyses including all relevant protein-protein interactions from the IntAct protein interaction analysis database (Hermjakob et al., 2004). Reactome calculates its p-values using over-representation analysis and hypergeometric distribution. It employs the Benjamini-Hochberg FDR correction at 5% (Benjamini and Hochberg, 1995).

Results

Biological processes

We identified 4278 potential target genes of these miRNA (Supplementary information-1). These potential target genes significantly (Fisher’s exact test p < 0.0001) overlapped with the previously reported post-mortem DLB brain DEG (Rajkumar et al., 2020) and with serum SEV DEG in people living with DLB (Rajkumar et al., 2021). The potential target genes were significantly (q < 0.05) enriched among the genes associated with 62 biological processes after Benjamini-Hochberg FDR correction at 5% (Table 1; Supplementary information-2). The enriched biological processes clustered around communication between cells, post-translational protein modification, apoptosis, neuronal development, transcriptional regulation, regulation of cell cycle, autophagy lysosomal pathway, vesicle-mediated transport, plasma membrane organisation, and cellular macromolecule biosynthesis. The potential target genes were significantly enriched among the genes associated with several transcriptional regulation processes including regulation of transcription from RNA polymerase-II promoter (q = 7.31 × 10⁻⁸), negative regulation of gene expression (q = 1.50 × 10⁻⁶), regulation of DNA-templated transcription (q = 1.55 × 10⁻⁶), and regulation of gene expression (q = 9.80 × 10⁻⁵). Significantly enriched biological processes included apoptosis-associated regulation of apoptotic process (q = 1.59 × 10⁻⁵), negative regulation of programmed cell death (q = 5.30 × 10⁻⁴) and negative regulation of apoptotic process (q = 0.0017), as well as autophagy lysosomal pathway associated lysosomal transport (q = 0.0196) and negative regulation of autophagy (q = 0.0065) processes. Supplementary information-2 provides further details of all 62 significantly enriched biological processes with their gene ontology (GO) numbers, p-values, FDR-adjusted q-values, odds ratios, and overlapping genes.

Table 1. Biological processes that were enriched among the potential target genes of differentially expressed small extracellular vesicle miRNA in people with dementia with Lewy bodies

GO: Gene Ontology.

* Benjamini-Hochberg false discovery rate at 5%.

Molecular functions

The potential target genes of these miRNA were significantly enriched among the genes associated with 35 molecular functions after Benjamini-Hochberg FDR correction at 5% (Supplementary information-2). These enriched molecular functions clustered around transcriptional regulation, protein phosphorylation, ubiquitin protease system (UPS), and histone acetylation. The potential target genes were significantly enriched among the genes associated with several molecular functions regulating transcription such as transcription coactivator activity (q = 3.34 × 10⁻⁵), transcription regulatory region DNA binding (q = 0.0024), RNA polymerase-II core promoter proximal region sequence-specific binding (q = 0.0240) transcription factor (q = 0.0124) and transcriptional activator (q = 0.0139) activities, activating transcription factor binding (q = 0.0134), RNA polymerase-II transcription cofactor activity (q = 0.0329), and transcription corepressor activity (q = 0.0439). The potential SEV miRNA target genes were significantly enriched among the genes involved in the UPS-related molecular functions, ubiquitin-protein transferase activity (q = 0.0035), and ubiquitin-protein ligase activity (q = 0.0342), that have been known to be associated with DLB pathology (Chowdhury and Rajkumar, 2020). Significantly enriched molecular functions included histone acetyltransferase activity (q = 0.0188) and several protein phosphorylation regulating functions such as protein serine/threonine kinase activity (q = 8.65 × 10⁻⁷), protein kinase activity (q = 3.17 × 10⁻⁶), protein kinase binding (q = 4.30 × 10⁻⁴), and protein kinase A regulatory subunit binding (q = 0.0044). Supplementary information-2 provides further details of all 35 significantly enriched molecular functions with their GO numbers, p-values, FDR-adjusted q-values, odds ratios, and overlapping genes.

Enriched molecular pathways

The potential target genes were significantly enriched for 59 KEGG pathways after Benjamini-Hochberg FDR correction at 5% (Table 2; Supplementary information-2). These enriched molecular pathways clustered around neuronal development, cell signalling, synaptic activity, programmed cell death, autophagy, UPS, and vesicle-mediated transport. They included axon guidance (q = 1.69 × 10⁻⁶), focal adhesion (q = 5.08 × 10⁻⁴), autophagy (q = 8.78 × 10⁻⁴), mRNA surveillance (q = 0.0016), long-term potentiation (q = 0.0053), ubiquitin-mediated proteolysis (q = 0.0283), cholinergic (q = 0.0493) and glutaminergic (q = 0.0283) synapses, ferroptosis (q = 0.0350), endocytosis (q = 0.0493), as well as Wnt (q = 3.22 × 10⁻⁴), mTOR (q = 5.08 × 10⁻⁴), ErbB (q = 0.0018), FoxO (q = 0.0025), and neurotrophin (q = 0.0076) signalling pathways. Further details of all 59 significantly enriched KEGG pathways with their pathway (hsa) numbers, p-values, FDR-adjusted q-values, odds ratios, and overlapping genes are presented in the supplementary information-2.

Table 2. Molecular pathways that were enriched among the potential target genes of differentially expressed small extracellular vesicle miRNA in people with dementia with Lewy bodies

KEGG: Kyoto Encyclopaedia of Genes and Genomes; hsa: Homo sapiens.

* Benjamini-Hochberg false discovery rate at 5%.

Disease processes

We assessed how the potential target genes were related to the genes that are known to be associated with specific disease processes in OMIM and DisGeNET (Pinero et al., 2021) databases. The potential target genes were significantly enriched among the genes that are associated with 74 disease processes in the DisGeNET after Benjamini-Hochberg FDR correction at 5% (Supplementary information-2). These enriched disease processes included several neurodevelopmental disorders such as intellectual disability (q = 7.58 × 10⁻⁸), delayed speech and language development (q = 1.47 × 10⁻⁵), and autism (q = 1.47 × 10⁻⁵), as well as neuropsychiatric disorders such as schizophrenia (q = 7.23 × 10⁻⁴), speech impairment (q = 6.37 × 10⁻⁵), epilepsy (q = 7.49 × 10⁻⁴), cognitive disorders (q = 0.0192), bipolar disorder (q = 0.0353), and impaired cognition (q = 0.0446). Besides, the potential target genes were significantly (p < 0.05) enriched only among the genes associated with encephalopathy (p = 0.0037) in OMIM database, and this enrichment was not statistically significant after FDR correction (Supplementary information-2).

Protein interactions

We assessed the protein interaction pathways of the potential miRNA target genes using the Reactome pathway database (Fabregat et al., 2017). The potential target genes were significantly (p < 0.05) overrepresented among 13 protein interaction pathways, and none of these associations were significant after FDR correction (Table 3; Supplementary information-2). These 13 protein interaction pathways indicated that the protein interactions of the potential target genes may contribute to DLB pathology by regulating transcription, cell cycle, signal transduction, and interleukin-6 signalling. Moreover, the potential target genes were significantly (p < 0.05) overrepresented among four protein interaction pathways after considering protein-protein interactions, and none of these associations were significant after FDR correction (Table 3; Supplementary information-2). These four protein interaction pathways indicated the possibility of the potential target genes influencing cell cycle, post-transcriptional silencing by small RNA, and phosphorylation of β-catenin that is one of the core molecules in the Wnt signalling pathway.

Table 3. Protein interaction pathways* that were enriched among the potential target genes of differentially expressed small extracellular vesicle miRNA in people with dementia with Lewy bodies

* Derived from Reactome pathway database (https://reactome.org/).

Secondary analyses of target genes of individual miRNA

We repeated the functional enrichment and protein interaction analyses for the potential target genes of each of the six miRNA genes individually. These secondary analyses identified additional significantly (q < 0.05) enriched biological processes, molecular functions, and KEGG pathways, as well as nominally significant (p < 0.05) protein interaction pathways. Supplementary information-3 presents these additional findings. The potential target genes of hsa-miR-26a-5p and hsa-miR-26a-1-3p were significantly enriched among the genes associated with signal transduction, post-translational protein modifications, nucleotide binding, regulation of cellular catabolic process (q = 0.0335), thiol-dependent ubiquitinyl hydrolase activity (q = 0.0085), ubiquitin-like protein ligase activity (q = 0.0461), and RNA polymerase-II distal enhancer sequence-specific binding transcription factor activity (q = 0.0461). The potential target genes of hsa-miR-320d were significantly enriched among the genes involved in longevity regulating pathway (q = 0.0019) and transcriptional misregulation in cancer (q = 0.0451). They were significantly (p < 0.05) overrepresented among nuclear receptor transcription (p = 8.99 × 10⁻⁵) and Cohesin loading onto chromatin (p = 0.0464) protein interaction pathways. Moreover, the potential target genes of hsa-mir-548ba were significantly enriched for ubiquitin-like protein ligase activity (q = 0.019) and protein carboxyl O-methyltransferase activity (q = 0.0359). Their protein interactions were significantly (p < 0.05) overrepresented among 19 protein interaction pathways including activation of G-protein gated potassium channels (p = 0.0034), inhibition of voltage-gated Ca²⁺ channels via Gbeta/gamma subunits (p = 0.0034), GABA receptor activation (p = 0.0132), GABA-B receptor activation (p = 0.0267), RNA Polymerase-III transcription initiation from type-1 (p = 0.0410) and type-2 (p = 0.0367) promoters, and GRB2 events in ERBB2 signalling (p = 0.0397). Furthermore, functional enrichment and protein interaction analyses for the potential target genes of hsa-miR-556-5p and hsa-miR-556-3p did not reveal any additional significantly (q < 0.05) enriched biological processes, molecular functions, and KEGG pathways. Their protein interactions were significantly (p < 0.05) overrepresented among seven protein interaction pathways including FGFR2 ligand binding and activation (p = 0.0068) and voltage-gated potassium channels (p = 0.0094) (Supplementary information-3).

Discussion

This is the first study that systematically investigated the functional implications of differentially expressed miRNA in serum SEV from people with DLB. It adds evidence supporting the hypothesis that differentially expressed SEV miRNA may contribute to DLB molecular pathogenesis by widespread transcriptional regulation (Gamez-Valero et al., 2019; Rajkumar et al., 2021). The potential target genes of the differentially expressed SEV miRNA were significantly enriched among the genes involved in post-translational protein modifications (Manzanza et al., 2021), transcriptional regulation (Chowdhury and Rajkumar, 2020; Feleke et al., 2021), autophagy lysosomal pathway (Crews et al., 2010; Arotcarena et al., 2019), UPS (Bedford et al., 2008; Zheng et al., 2016), apoptosis (Mahul-Mellier et al., 2020), regulation of cell cycle (Lee et al., 2003), histone acetylation (Urbizu and Beyer, 2020), cell signalling (Beyer et al., 2009; Garcia-Esparcia et al., 2017), synaptic dysfunction (Overk and Masliah, 2014), and vesicle-mediated transport (Kurzawa-Akanbi et al., 2021; Longobardi et al., 2022) that have been associated with DLB pathology. Strengths of this study include employing appropriate FDR corrections, focusing on differentially expressed SEV miRNA that were either differentially expressed in both post-mortem DLB brains and serum SEV or replicated by high-throughput qPCR, and considering protein interactions. However, there were limitations including the methodological difficulties in miRNA target prediction (Riolo et al., 2020), limited information regarding potential targets of other differentially expressed SEV small RNA in DLB, and investigating miRNA that were identified in total serum SEV population that was not enriched for neuronal origin (Rajkumar et al., 2021).

Several lines of evidence indicate that post-mortem brain and CSF-derived EV carry pathogenic cargo that are capable of inducing alpha-synuclein oligomerisation and propagating Lewy pathology (Danzer et al., 2012; Stuendl et al., 2016; Ngolab et al., 2017; Kurzawa-Akanbi et al., 2021; Herman et al., 2022). However, the molecular mechanisms by which the pathogenic EV propagate Lewy pathology remain uncertain (Choi et al., 2021). SEV miRNA, derived from people with AD or Parkinson’s disease (PD), have been shown to contribute towards neurodegeneration by transcriptional regulation of genes involved in neuroinflammation and autophagy (Vassileff et al., 2020; Mavroeidi et al., 2022). Current evidence supporting the uptake of SEV miRNA by cells and their transcriptional regulation in recipient cells (Hu et al., 2012; Li et al., 2021; Mustajab et al., 2022) outweigh the evidence to the contrary (Albanese et al., 2021). Moreover, several miRNA are known to regulate alpha-synuclein levels, functions and oligomerisation in people with PD (Zhao and Wang, 2019), and pertinent research involving people with DLB remains sparse. Such prior evidence and the findings of this study highlight the need for further experiments investigating the importance of SEV miRNA in the molecular pathogenesis of DLB.

MIR26A1, also known as MIR26A, is well known to regulate neuronal development and synaptic plasticity (Li and Sun, 2013). hsa-miR-26a-5p has been included in the blood-based 12 miRNA signature panel for AD (Leidinger et al., 2013), and its expression level has been found to be significantly higher in people with frontotemporal dementia, when compared to people without dementia (Martinez and Peplow, 2022). The diagnostic biomarker potential of hsa-miR-26a-5p in DLB (Rajkumar et al., 2021) warrants further investigation. Moreover, this study has identified PTGS2 encoding cyclooxygenase-2 as one of the potential target genes of hsa-miR-26a-5p. Dual-luciferase and qPCR experiments have confirmed that hsa-miR-26a-5p directly targets and regulates transcription of PTGS2 (Xie et al., 2022). Transcriptional regulation of PTGS2 by hsa-miR-26a-5p has been reported to lead to proliferation of AD pathology (Xie et al., 2022), and further investigation investigating the contribution of this miRNA-mRNA pair in DLB pathology is needed.

Expression level of MIR320C2 is reportedly significantly higher in plasma of people with AD (Nagaraj et al., 2017), and it has been found to be significantly downregulated in post-mortem PD brains (Briggs et al., 2015) as well as post-mortem DLB brains and serum SEV from people living with DLB (Rajkumar et al., 2021). This supports the need for investigating the biomarker potential of hsa-miR-320c for differentiating DLB from AD in large cohorts. Similarly, MIR320D2 has been found to be significantly upregulated in plasma of people with PD (Chen et al., 2021), and it was significantly downregulated in post-mortem DLB brains and serum SEV from people with DLB (Rajkumar et al., 2021). hsa-miR-320d may help differentiating DLB from PD dementia. Furthermore, hsa-miR-4722-5p is likely to regulate genes involved in signal transduction and autophagy (Liu et al., 2022). It has been reported to be significantly higher in serum of people with AD and has been identified as a potential blood-based biomarker for AD (Soleimani Zakeri et al., 2020; Liu et al., 2022). MIR4722 was significantly downregulated in post-mortem DLB brains (Rajkumar et al., 2020) as well as serum SEV from people with DLB (Rajkumar et al., 2021), and this may contribute to the impairment of autophagy lysosomal pathway in DLB (Crews et al., 2010; Arotcarena et al., 2019). Diagnostic biomarker and therapeutic potential of hsa-miR-4722-5p in DLB need further evaluation.

The potential target genes of the investigated SEV miRNA in DLB include UBE3A, UBE3B, and 11 ubiquitin-conjugating enzyme genes. The potential target genes were significantly enriched among the genes involved in ubiquitin-protein transferase activity, ubiquitin-protein ligase activity, ubiquitin-mediated proteolysis, lysosomal transport, and negative regulation of autophagy. Dysfunction of autophagy lysosomal pathway and/or UPS leads to accumulation of pathologically misfolded proteins and progression of Lewy pathology (Zheng et al., 2016). Recent transcriptomic studies (Pietrzak et al., 2016; Nelson et al., 2018; Santpere et al., 2018; Rajkumar et al., 2020) and a systematic review (Chowdhury and Rajkumar, 2020) have reported statistically significant downregulation of several autophagy lysosomal pathway and UPS-associated genes in post-mortem DLB brains. Statistically significant downregulation of UPS-associated genes, UBE3A, USP47, and PSMD4, and of the protein ubiquitination pathway in serum SEV from people with DLB have been reported (Rajkumar et al., 2021). The UPS closely interacts with the autophagy lysosomal pathway, and their dysfunction was shown to be sufficient for causing Lewy body-like inclusions in mice models (Zheng et al., 2009). Their dysfunction leads to cytoplasmic accumulation of alpha-synuclein and other misfolded proteins that can set off a vicious cycle by inhibiting neuronal lysosomal activity further (Mazzulli et al., 2011). Moreover, our findings showed that the potential target genes were likely to regulate transcription of genes involved in protein phosphorylation and post-translational protein modifications. Phosphorylation and other post-translational modifications of alpha-synuclein contribute substantially towards molecular pathogenesis of DLB (Outeiro et al., 2019; Manzanza et al., 2021). Additionally, the potential target genes may play a role in DLB pathogenesis by impacting Wnt signalling pathway and associated β-catenin phosphorylation cascade that regulate microglia-mediated neuroinflammation and synaptic plasticity (Rajkumar et al., 2020; Yang and Zhang, 2020).

Our findings support the biological plausibility of the differentially expressed SEV miRNA contributing towards DLB pathogenesis by transcriptional regulation of other genes. Therapeutic potential of these differentially expressed SEV miRNA and their target genes are warranted. Further experiments are required for verifying predicted miRNA targets as well as identified dysfunctional molecular processes and pathways. The potential regulation of post-translational modifications by SEV miRNA-mediated transcription, which directly act upon alpha-synuclein and other proteins and their functions, provides additional layers of regulatory cross-talk that could be targeted for therapeutic benefit. Moreover, SEV miRNA have opened a novel avenue for identifying blood-based diagnostic biomarkers for DLB. There is an urgent need for an adequately powered blood-based SEV RNA sequencing study that is not biased by post-transcriptional RNA modifications (Shi et al., 2021). Discovered potential diagnostic biomarker miRNA and other small RNA should be evaluated in large replication cohorts. As gene expression changes often differ with disease progression, measuring the expression levels of potential diagnostic biomarkers at various clinical stages of DLB is preferred. Besides, investigating SEV that are enriched for neuronal (Mustapic et al., 2017) and/or microglial (Winston et al., 2021) origin by immunoprecipitation may facilitate the discovery of novel diagnostic biomarkers and therapeutic targets for DLB.