Gámez-Valero et al. Translational Neurodegeneration (2019) 8:31 https://doi.org/10.1186/s40035-019-0169-5 RESEARCH Open Access Exploratory study on microRNA profiles from plasma-derived extracellular vesicles in Alzheimer’s disease and dementia with Lewy bodies Ana Gámez-Valero1,2, Jaume Campdelacreu3, Dolores Vilas4, Lourdes Ispierto4, Ramón Reñé3, Ramiro Álvarez4, M. Pilar Armengol5, Francesc E. Borràs2,6*† and Katrin Beyer1*† Abstract Background: Because of the increasing life expectancy in our society, aging-related neurodegenerative disorders are one of the main issues in global health. Most of these diseases are characterized by the deposition of misfolded proteins and a progressive cognitive decline. Among these diseases, Alzheimer’s disease (AD) and dementia with Lewy bodies (DLB) are the most common types of degenerative dementia. Although both show specific features, an important neuropathological and clinical overlap between them hampers their correct diagnosis. In this work, we identified molecular biomarkers aiming to improve the misdiagnosis between both diseases. Methods: Plasma extracellular vesicles (EVs) -from DLB, AD and healthy controls- were isolated using size-exclusion chromatography (SEC) and characterized by flow cytometry, Nanoparticle Tracking Analysis (NTA) and cryo-electron microscopy. Next Generation Sequencing (NGS) and related bibliographic search was performed and a selected group of EV-associated microRNAs (miRNAs) was analysed by qPCR. Results: Results uncovered two miRNAs (hsa-miR-451a and hsa-miR-21-5p) significantly down-regulated in AD samples respect to DLB patients, and a set of four miRNAs (hsa-miR-23a-3p, hsa-miR-126-3p, hsa-let-7i-5p, and hsa-miR-151a-3p) significantly decreased in AD respect to controls. The two miRNAs showing decreased expression in AD in comparison to DLB provided area under the curve (AUC) values of 0.9 in ROC curve analysis, thus suggesting their possible use as biomarkers to discriminate between both diseases. Target gene analysis of these miRNAs using prediction online tools showed accumulation of phosphorylation enzymes, presence of proteasome-related proteins and genes involved in cell death among others. Conclusion: Our data suggest that plasma-EV associated miRNAs may reflect a differential profile for a given dementia-related disorder which, once validated in larger cohorts of patients, could help to improve the differential diagnosis of DLB versus AD. Keywords: Neurodegenerative disorders, Biomarker, Exosomes, Next generation sequencing, Extracellular vesicles * Correspondence: feborras@igtp.cat; katrinbeyer@hotmail.com † Francesc E Borràs and Katrin Beyer are senior authors contributed equally to this study. 2 REMAR-IVECAT group, Health Sciences Research Institute Germans Trias i Pujol (IGTP), 08916 Badalona, Spain 1 Department of Pathology, Health Sciences Research Institute Germans Trias i Pujol (IGTP), Universitat Autònoma de Barcelona, Badalona 08916, Spain Full list of author information is available at the end of the article © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Gámez-Valero et al. Translational Neurodegeneration (2019) 8:31 Page 2 of 17 Background Materials and methods Neurodegenerative disorders with largest economic burden Blood collection and sample processing in our society include dementia syndromes such as Alzhei- The Clinical Research Ethics Committee of our institu- mer’s disease (AD), frontotemporal dementia, dementia tion approved the following protocol and from each sub- with Lewy bodies (DLB), and movement disorders namely ject, written informed consent was obtained according Parkinson’s disease (PD), in which about 50% of patients to the Declaration of Helsinki Principles [24]. DLB pa- may develop dementia 15 years after PD onset [1–3]. tients (n = 18; age range 62–84 years; mean 72.5 years; Clinically, along the course of the three diseases, a progres- male: female ratio 1:1.75) were recruited by neurologists sive cognitive decline affecting the normal and social specialized in Lewy body disorders at the Dementia Unit, functions of the patients is observed [4]. Moreover, neuro- Department of Neurology from the University Hospital pathologically, a common neurodegeneration-related fea- Bellvitge, (L’Hospitalet de Llobregat, Barcelona). Diagnosis ture is the deposition of misfolded proteins. While in DLB of DLB patients was established according to the 2005 and PD α-synuclein accumulations, named Lewy bodies are DLB Consortium Criteria [25], and the age at onset was found [5], in AD –and also in about half of all DLB cases- defined as the age when parkinsonism or memory loss ß-amyloid senile plaques and hyperphosphorylated tau are was first reported by patients’ relatives. A group of age- accumulated in the brain [6, 7]. Correspondingly, DLB and gender-matched healthy individuals (n = 15; age-range shows an important neuropathological and therefore clin- 61–85; mean 70.5 years; male: female 1:2) from the same ical overlap with both, PD and AD, hampering its correct hospital were also recruited. Finally, a group of AD pa- identification. Currently, a high proportion of DLB cases tients (n = 10; age range 62–80; mean 71; male: female ra- are missed or misdiagnosed as AD, resulting in an incorrect tio 1:1.5) was enrolled by the Neurology Department of treatment of the patient, which leads to the development of the University Hospital Germans Trias i Pujol. AD diagno- adverse reactions in 50% of these treated patients [7]. sis was assessed following the 2011 revised criteria from Over the past decade, the role of extracellular the National Institute on Aging and the Alzheimer’s vesicles (EVs) in the development and functioning of Association [26]. These patients presented a Global the central nervous system has been deeply explored Deterioration Scale of 4.3 ± 1.2 degrees. Clinical data of all [8–10]. EVs, such as exosomes (EXs) and microvesi- patients and healthy controls enrolled in this study are cles (MVs), seem to be important effectors in the shown in Table 1. development of cognitive and neurodegenerative Blood samples from all participants were obtained disorders. They have been shown to mediate the following ISEV- International Society for Extracellular transport of prions and misfolded pro-aggregating Vesicles- guidelines [27] and applying the same col- proteins from cell to cell [11–14]. Because EVs are lection protocol in both hospitals. In short, 15 mL of produced by each individual cell and provide a pro- peripheral blood were collected by venous puncture tected environment to shuttle not only proteins but using a 21-gauge needle coupled to a butterfly device also microRNAs (miRNAs) to the intercellular space and using sodium citrate pre-treated tubes (BD Vacu- and to different body fluids, the study of their con- tainer, New Jersey, USA) as described previously [28]. tent has emerged as an area of interest in the bio- After discarding the first 2–3 ml, blood was collected marker field, also in neurodegenerative disorders and mixed with the anticoagulant by gently inverting [15–18]. To date, several EV-associated miRNAs tubes gently. Samples were processed within the first have been identified as altered and related to neuro- 2 h after collection. Plasma was clarified of platelets degenerative disorders [19–21]. Many of these stud- and cells by consecutive centrifugation steps at 500 x ied EVs were obtained from cerebrospinal fluid g for 10 min, 2500 x g for 15 min and a last step at (CSF) samples. However, due to several issues on 13,000 x g for 10 min. Plasma samples were frozen in CSF-collection (difficulty, invasiveness, morbidity, a freezing container with freezing rate of - 1 °C/min risk), alternative sources such as plasma-derived EVs and kept at − 80 °C until EV purification. Samples did have been lately reported [22, 23]. not suffer from more than 2 freeze-thaw cycles. In this study, we analysed the miRNA profile associated to plasma-derived EVs from DLB, AD and healthy controls. EV isolation by size exclusion chromatography (SEC) We further investigated whether differences in the plasma- Size Exclusion Chromatography (SEC) was used for isolat- EV-miRNA content could be of help to better characterize ing plasma-EVs as previously described [28]. Briefly, Seph- neurodegenerative disorders, specifically DLB and AD. Our arose-CL2B (Sigma Aldrich, St Louis, MO, USA) was data suggest that some plasma-EV associated miRNAs are stacked in a Puriflash column Dry Load Empty 12 g (20/pk) differentially found in DLB and AD patients, and thus could from Interchim (France)-Cromlab, S.L. (Barcelona, Spain). help to improve the differential diagnosis of these overlap- Once the column was completely stacked, two mL of ping neurodegenerative disorders. plasma (previously thawed on ice) were loaded onto the Gámez-Valero et al. Translational Neurodegeneration (2019) 8:31 Page 3 of 17 Table 1 Data of patients and control individuals included in the study Sample Clinical Diagnosis DatScan Gender Age (blood coll)a Age at onset MMSEb APOE Ex1 DLB abnormal F 85 83 20 33 Ex2 DLB abnormal M 79 73 24 23 Ex3 DLB positive F 82 78 15 34 Ex4 DLB positive F 73 68 16 33 c Ex5 DLB no F 90 84 5 34 Ex6 DLB positive F 80 64 10 33 Ex7 DLB positive F 86 79 28 33 Ex8 DLB no F 79 74 19 34 Ex9 DLB positive M 74 67 6 33 Ex10 DLB positive F 79 74 12 33 Ex11 DLB positive M 65 59 5 33 Ex12 DLB positive M 77 67 24 33 Ex13 DLB positive F 83 80 16 33 Ex14 DLB positive M 70 62 12 33 Ex15 DLB positive F 77 73 18 34 Ex16 DLB positive M 63 59 15 33 Ex17 DLB positive F 64 62 11 nad Ex18 DLB normal M 73 73 22 34 Ex19 AD – F 75 74 18 34 Ex20 AD – M 75 75 22 23 Ex21 AD – F 70 70 23 33 Ex22 AD – F 80 80 20 33 Ex23 AD – F 70 63 12 33 Ex24 AD – M 62 60 16 34 Ex25 AD – M 72 65 15 34 Ex26 AD – F 74 70 22 34 Ex27 AD – M 64 na 18 34 Ex28 AD – F 68 na 20 34 C-Ex1 CTRL – F 71 – 28 33 C-Ex2 CTRL – F 67 – 29 33 C-Ex3 CTRL – F 66 – 27 33 C-Ex4 CTRL – F 75 – 28 34 C-Ex5 CTRL – F 74 – 26 33 C-Ex6 CTRL – M 69 – 30 33 C-Ex7 CTRL – M 72 – 27 23 C-Ex8 CTRL – F 69 – 28 33 C-Ex9 CTRL – F 67 – 26 34 C-Ex10 CTRL – M 67 – 28 23 C-Ex11 CTRL – F 72 – 27 33 C-Ex12 CTRL – F 69 – 29 33 C-Ex13 CTRL – F 61 – 28 33 C-Ex14 CTRL – M 73 – 27 33 C-Ex15 CTRL – M 85 – 26 33 a age at blood collection; bMMSE: The Mini-Mental State Examination; cno DaTSCAN evaluation available; dnot available Gámez-Valero et al. Translational Neurodegeneration (2019) 8:31 Page 4 of 17 column and eluted with filtered PBS. A total of 20 fractions was subjected to a last 2 min centrifugation to com- (0.5 mL each) were immediately collected. pletely dry the membrane. MiRNA elution was per- formed by adding 100 μL of RNase-free H2O directly EV characterization onto the membrane and centrifuging it for 1 min at 11, EV fractions were characterized for their protein con- 000 x g. The obtained material was kept at − 80 °C until centration, presence of specific EV-markers, size and further analysis. morphology. In samples utilized for RT-qPCR validation experi- As published before, protein concentration for each SEC ments, 2 artificial RNAs (UniSp4 and UniSp5 from the collected fraction was measured by absorbance at 280 nm RNA spike-in kit, Exiqon, Vedbaek, Denmark) were in a Thermo Scientific Nanodrop® ND-100 (Thermo Fisher spiked into the lysis buffer before miRNA purification, Scientific, Waltham, MA). Also, SEC-fractions were ana- according to the manufacturer’s protocol enabling the lysed for the presence of CD9, CD81, CD63 and CD5L as assessment of miRNA purification and amplification specific EV-markers by bead-based flow cytometry assay efficiency. [28, 29]. Briefly, 50 μL of each fraction were incubated with 0.5 μL aldehyde/sulphate-latex beads-4 μm (Invitrogen, MicroRNA discovery by next generation sequencing (NGS) Carlsbad, CA) for 15 min at room temperature, re-sus- and raw data analysis pended in coupling buffer and incubated overnight. After The total volume of the obtained miRNAs from 7 DLB and two washing steps with the same buffer, EV-coated beads 7 control samples was precipitated overnight at − 20 °C were incubated with anti-CD9 (Clone VJ1/20), anti-CD63 with 1 μL of glycogen (20 μg/ μL), 10% 3 M AcNa (ph 4.8) (Clone TEA 3/18), anti-CD5L (ab45408 from Abcam) and and 2 volumes of absolute ETOH. miRNAs were re-sus- anti-CD81 (clone 5A6, from Santa Cruz Biotech) or poly- pended in RNase free H2O and heated at 65 °C for 2–3 clonal IgG isotype (Abcam, Cambridge, UK) for 30 min at min. Quality control and size distribution of the purified 4 °C. After a washing step, samples were subjected to 30 small RNA was assessed by Bioanalyzer 2100, (Agilent min incubation at 4 °C with a FITC-conjugated secondary Technologies, Santa Clara, USA). goat anti-mouse antibody (Southern Biotech, Birmingham, The whole precipitated volume of each sample (10 μL) AL) for CD9, CD81 and CD63, and secondary Ab anti- was used for library preparation by NEBNext Multiplex rabbit AlexaFLuor 488 (Invitrogen, Carlsbad, CA) for Small RNA Sample Preparation Set for Illumina (New CD5L. Two final washing steps were performed before ana- England Biolabs, Massachusetts, USA) following kit in- lysing the samples by flow cytometry in a FacsVerse cyt- structions. Individual libraries were subjected to the ometer (BD Biosciences, New Jersey, USA). Mean quality analysis using a D1000 ScreenTape (TapeStation, fluorescence intensity (MFI) values were plotted (Flow Jo Agilent Technologies), quantified by fluorimetry and software, Tree Star, Ashland, OR) and tetraspanin-positive pooled. Clustering and sequencing were done in an Illu- fractions with the highest MFI (fractions 10–12 from our mina Sequencer (MiSeq, Illumina, San Diego, USA) at 1 SEC column) were considered as EV-containing fractions x 50c single read mode and 200,000 reads were obtained and pooled for the forthcoming analysis. for each sample. Obtained FastQ raw data were analysed Aiming to check EV morphology and size, EV- as follow: (1) Trimmomatic was used to remove the enriched pools were also subjected to cryo-electron mi- adapter sequences from the reads [30]; (2) Reads were croscopy (cryo-EM) and to Nanoparticle Tracking Ana- mapped to the human genome using Bowtie2 algorithm lysis (NTA) (n = 6), as reported earlier [28]. and individual miRNAs were identified [31]; (3) For each sample, the number of reads mapped to a particular Isolation of microRNA miRNA sequence was counted; and (4) the total count A volume of 750 μL of pooled EVs from each sample of reads was normalized applying the weighted trimmed was lyophilized at − 23 °C /overnight and used for means of M-values (TMM) [32]. Before differential ex- miRNA extraction using the miRCURYTM RNA Isola- pression analysis, Lilliefors’ composite goodness-of-fit tion Kit-biofluids (Exiqon Vedbaek, Denmark) at room test, Jarque-Bera hypothesis test and Shapiro-Wilk test temperature as described by the manufacturer. Briefly, were applied to check the normality of our samples. lyses solution and protein precipitation solution were For NGS expression analysis the following criteria added to each sample. After incubation at room were followed: at least minimum of 5 reads per sample temperature for 1 min, samples were centrifuged at 11, when present to considered a miRNA as present; present 000 x g for 3 min. Isopropanol was added to the super- in all patient samples (minimum of 5 reads) and absent natants and the mix was transferred to the provided col- (less than 5 reads) in more than half of the control co- umns. After a first centrifugation at 11,000 x g for 30 s, hort; present in all control samples and absent in more several washing steps at the same centrifugation speed than half of the patients for presence/absence consider- with 1BF and 2BF solutions were applied. The column ation; present in most samples from both cohorts but Gámez-Valero et al. Translational Neurodegeneration (2019) 8:31 Page 5 of 17 differentially expressed between both groups. Differential was corrected by the Cq values for UniSp3 as interplate expression analysis for DLB versus control cohorts was calibrator. As stable genes, RNA-isolation control spike-in performed by Wilconson-rank sum test (p-value < 0.05) UniSp4 and the retro-transcription control spike-in [33] and validated by the Leave-One-Out (LOO) cross- UniSp6 were considered as reference genes. The average validation methodology. of the two Cq values was used as reference value to calcu- late miRNA expression. Relative expression in DLB and Bibliographic search and miRNA selection AD was estimated compared to healthy controls and rep- Literature search based on PubMed and The Nervous resented as fold expression changes obtained by 2-ΔΔCt. System Disease NcRNAome Atlas (NSDNA) [34] data- Considering the low number of samples, multiple compar- bases was performed aiming to identify miRNAs already isons between the three groups (DLB, AD and controls) described as deregulated in dementia and neurodegener- were performed using the Kruskal-Wallis non-parametric ative disorders. We then combined this bibliographic in- test and Dunn’s test was used for multiple corrections. A formation with our data from NGS results in plasma-EV p-value below 0.05 was considered statistically significant. from DLB and healthy control samples. Thus, a group of One-variable area under the ROC curve (AUC) was calcu- 15 highly represented miRNAs (all of them producing lated considering the expression change in order to assess more than 5000 reads in the NGS assay and belonging the possible diagnostic potential of each differentially to the top most abundant miRNAs identified by NGS) expressed miRNA by the Wilson/Brown method (Graph- were further considered for expression studies to com- Pad Prism v7; 95% C.I., AUC > 0.750 was assessed as mini- pare AD and DLB by qPCR. mum value to be considered as a good-potential biomarker). Reverse transcription and qPCR analysis We have submitted all relevant data of our experi- MiRCURY LNATM Universal cDNA synthesis Kit II ments to the EV-TRACK knowledgebase (EV-TRACK and miRNA PCR system Pick & Mix (Exiqon, Ved- ID: EV180020) [35]. baek, Denmark) were used for cDNA synthesis and qPCR validation analysis of the selected miRNAs microRNA target prediction according to the manufacturer’s instructions (see A list of possible affected target genes for the differentially Additional file 1: Table S1 for further purchasing in- expressed miRNAs by qPCR was obtained by miRGate formation). Due to the low concentration of genetic [36] database. Filtering of the software output data was material associated to vesicles, 2 μL of undiluted miR- performed following the instructions of the developers. NAs were used for retrotranscription at 42 °C for 60 We first considered validated targets, named “confirmed min. A third spike-in miRNA (UniSp6) from the same predictions” and showing the number of confirmed data- kit (Exiqon) was used as retrotranscription control. bases of gene-miRNA biding supporting this prediction After enzyme inhibition at 95 °C for 5 min, cDNA was group in the output data. For the remaining suggested tar- diluted 1:80 and 4 μL were used for qPCR reaction gets, we took into account the column named “computa- with ExiLENT SYBR Green Master Mix (Exiqon, tional predictions”, defined by the number of Vedbaek, Denmark) on a LightCycler 480 (Roche, computational methods supporting this prediction group. Basel, Switzerland), following kit’s instructions. For These referred to genes bioinformatically predicted to be each sample, miRNAs were analysed in duplicate and regulated by the given miRNA, although no validation has the mean value was used in next data analyses. Amp- been reported. Target with more than one computational lification of the used spike-ins (UniSp4, UniSp5, and prediction were also considered [36]. UniSp6) was performed in the same PCR pre-de- The relationship between the miRGate validated targets signed panels (Exiqon, Vedbaek, Denmark) with an (those predicted with more than 1 confirmed predictions) interplate calibrator control miRNA (UniSp3). were analysed with String [37] and Panther Gene Ontol- ogy [38] Databases, obtaining an integrated clustered net- Data analysis and statistical testing work based on biological processes, cellular components All statistical analyses were performed using Prism 7 and KEEG (Kyoto Encyclopaedia of genes and genomes) (GraphPad Software, Inc., CA, USA). Pathways. In both cases, default settings were used. In Regarding EV characterization, MFI values for EV- Panther GO-analysis, Fisher’s exact test was applied and markers, EV-concentrations, and EV-size are given as Bonferroni correction for multiple testing was used. mean ± SD. Two-tailed unpaired T-test was applied (p < 0.05 was considered as statistically significant). Results For qPCR analysis, Cq (quantification cycle) values were EV isolation and characterization determined for each qPCR and the average of duplicates EVs were isolated from 2 mL of platelet-depleted plasma by was obtained. The variability between the different plates SEC from three different cohorts (healthy controls, DLB Gámez-Valero et al. Translational Neurodegeneration (2019) 8:31 Page 6 of 17 Fig. 1 (See legend on next page.) Gámez-Valero et al. Translational Neurodegeneration (2019) 8:31 Page 7 of 17 (See figure on previous page.) Fig. 1 Isolation and characterization of plasma-EVs. a EVs are eluted in low protein SEC fractions and EV-markers could be detected by bead- based flow cytometry. An example graph for one of the processed samples is shown (i); MFI values of the tetraspanins CD9, CD63 and CD81, and marker CD5L in the SEC-pooled fractions are represented. Bars represent mean +/−SD of 15 independent experiments in the case of tetraspanins and 5 in the case of CD5L (ii); (b) Cryo-EM images with characteristic shape and size of isolated vesicles; (c) above: representative NTA profiles for one control-EV and one DLB-EV sample, below: concentration and modal distribution for 6 control and 6 DLB samples (mean ± SD) and AD), obtaining 20 fractions of 0.5 mL from each sam- 307,581 ± 2,076,192 reads per sample. Although 311 ple. EV-fractions were identified by the presence of tetra- mature miRNAs were identified among all samples, spanins CD9, CD81, and CD63. Alternatively, some only those showing at least 5 reads (n = 238) were samples were also profiled for the presence of CD5L, re- further considered. More than 95% of these miRNAs cently described as plasma-EV marker [39]. In all cases, were already included in vesicular databases EVpedia EV-enriched fractions –positive marked for the 4 EV [40] and ExoCarta [41] as related to EV or exosomes markers used- were eluted, preceding the bulk of soluble from human samples (Fig. 2a). The hsa-let-7 family protein (Fig. 1A i). Three fractions showing the highest appeared as one of the most representative miRNA MFI values of each sample were pooled for further experi- groups among our data set (Fig. 2b). The whole list ments. No differences in MFI and expression of EV- of all identified miRNA has been submitted to the markers were found between the three groups (Fig. 1A ii). EVpedia database and entitled as this manuscript. Some samples were submitted to cryo-electron microscopy Statistical analyses of the 238 miRNAs identified by analysis confirming the presence of vesicles with the ex- NGS showed no differences between DLB and controls pected size and morphology (Fig. 1b). NTA data did not (Wilcoxon p-value > 0.05 for all). show any difference in particle concentration or particle size (Fig. 1c) between DLB and control groups. Selection of dementia related miRNAs for further qPCR analysis Discovery phase: miRNA profiles of DLB and control Despite the lack of differences and statistical significance plasma-EVs between DLB and healthy control cohorts, we carried out As first approach (discovery phase), 7 samples of a bibliographic search for those miRNAs that yielded DLB patients and 7 samples of age-matched controls more than 5000 reads in the NGS experiment. Their asso- were included for library construction and NGS ana- ciation to dementia, DLB, AD and PD was examined and lysis. An average of 4,364,157 ± 647,775 raw reads a group of 15 miRNAs (Table 2), mainly associated to de- per sample was obtained for control samples mentia and AD, was further analysed by qPCR in three in- (Additional file 2: Table S2); for DLB, we obtained 4, dependent cohorts of samples. Fig. 2 Profiling and characterization of NGS data. a Mature miRNAs identified were compared to the already described as human-EV content in ExoCarta and EVpedia; b Hsa-let-7 family was the most representative miRNA family in our data set Gámez-Valero et al. Translational Neurodegeneration (2019) 8:31 Page 8 of 17 Table 2 Selected microRNAs for further qPCR validation analysis miRNA Previously reported in the literature Ref NGS Mean Fold change (2-ΔΔCt) Counts (95% C.I.) CTRL DLB AD hsa-miR-21- Down-regulated in serum-EVs from AD patients compared [20] 14,337 1.49 (0.91–2.06) 1.81 (0.67–2.96) 0.29 (0.03–0.56) 5p to controls Up-regulated in serum-EVs from PD patients in comparison to AD Down-regulated in CSF from AD patients compared to [42] control individuals Down-regulated in plasma from PD patients compared to [43] normal controls hsa-miR- Deregulated in AD blood (different results in NGS and qPCR) [44] 27,896 1.88 (0.96–2.79) 1.04 (0.54–1.53) 0.8 (0.09–1.5) 26a-5p miR-26a is up-regulated in CSF from PD patients compared [21] to controls Up-regulated in blood from AD patients compared to [45, controls 46] Down-regulated in CSF from AD patients compared to [19] controls Down-regulated in serum from AD patients compared to [47] healthy controls Down-regulated in ALS blood compared to controls [48] hsa-let-7i- Down-regulated in PD brains compared to controls [49] 15,170 1.43 (0.78–2.09) 1.23 (0.05–1.75) 0.41 (0.06–0.75) 5p Increased expression in AD patients’ hippocampus [50] Up-regulated in CSF from AD compared to controls. [42] Down-regulated in ALS compared to controls [48] hsa-miR- miR-126 is down-regulated in CSF-EXs from AD and PD [21, 37,418 2.23 (1.41–3.04) 1.88 (0.66–3.1) 0.89 (0.03–1.74) 126-3p patients vs controls 51] Increased expression in the hippocampus of AD mouse [52] model vs WT controls hsa-miR- Up-regulated in serum-EXs from MS patients compared to [53] 10,058 2.05 (1.21–2.88) 1.84 (0.76–2.92) 0.19 (0.05–0.33) 451a controls Increased in plasma from vascular dementia patients [54] compared to healthy controls Decreased expression in CSF-EXs from AD compared to [55] controls Down-regulated in ALS compared to controls [48] hsa-miR- Up-regulated in brain tissue from AD patients [50] 6834 1.85 (1.24–2.45) 1.17 (0.65–1.68) 0.52 (0.06–0.97) 23a-3p miR-23a is down-regulated in serum samples from AD [56] patients’ vs FTD and controls Down-regulated in blood from MS patients compared to [57] controls Down-regulated in CSF from AD patients compared to [42] control individuals Reflect MS disease status in serum-EXs [53] Increased expression in brain tissue from AD patients [51] compared to controls Down-regulated in ALS blood compared to controls [48] hsa-let-7f- Up-regulated in AD hippocampus compared to healthy [50] 191,299 1.29 (0.59–1.99) 1.28 (0.23–2.34) 1.05 (0.33–1.76) 5p controls Up-regulated in AD serum compared to healthy controls [58] Down-regulated in blood from AD patients compared to [45] controls Gámez-Valero et al. Translational Neurodegeneration (2019) 8:31 Page 9 of 17 Table 2 Selected microRNAs for further qPCR validation analysis (Continued) miRNA Previously reported in the literature Ref NGS Mean Fold change (2-ΔΔCt) Counts (95% C.I.) CTRL DLB AD Down-regulated in ALS blood /plasma compared to controls [48, 59] hsa-miR- Down-regulated in the prefrontal cortex of AD patients [50] 5236 1.48 (0.05–3.01) 1.46 (0.12–2.79) 1.35 (0.19–2.52) 409-3p Down-regulated in CSF from PD patients compared to [20] controls Up-regulated in CSF-EXs from PD patients compared to AD [21, and control EXs 60] Up-regulated in serum-EXs from MS patients compared to [53] controls Down-regulated in plasma from vascular dementia patients [54] compared to controls hsa-miR- Down-regulated expression in serum from PD patients [61] 30,066 1.89 (0.42–3.35) 1.22 (0.54–1.89) 1.38 (− 0.29–3.06) 92a-3p compared to controls Differentially expressed in PD and Huntington patients’ brain [62] Up-regulated in CSF from AD patients compared to control [42] individuals Down-regulated in the serum samples of AD patients’ vs [63] MCI subjects Differentially expressed in AD and MCI [64] hsa-let-7b- Let-7b miRNA is up-regulated in AD patients’ brain [51] 107,394 1.98 (0.07–3.89) 0.81 (0.39–1.23) 2.097 (0.58–3.61) 5p Let-7b is down-regulated in the white matter of AD patients [65] Increased amounts of let-7b in CSF from AD patients [66] Differentially expressed in AD in comparison to controls [64] hsa-miR- Up-regulated in AD blood compared to controls [44] 5798 2.04 (− 0.13–4.22) 0.98 (0.34–1.63) 0.70 (− 0.09–1.5) 151a-3p Up-regulated in blood from AD patients compared to [45] controls Differentially expressed in AD in comparison to controls [64] hsa-miR-24- miR-24 is up-regulated in serum and plasma of MSA [67] 10,896 1.5 (0.68–2.33) 1.49 (0.28–2.69) 25.21 (3.11–47.31) 3p compared to PD patients Dow-regulated in plasma-EXs from AD patients compared to [22] controls miR-24 is deregulated in CSF from AD patients [68] Decreased expression in AD-CSF compared to controls [42, 69] miR-24 expression is decreased in CSF from PD patients [16] compared to controls Differently expressed in blood and CSF in AD and FTD [70] patients hsa-miR- miR-143 is up-regulated in AD brain patients [51] 17,380 5.85 (− 2.88–14.6) 1.22 (− 0.11–2.56) 2.42 (− 1.28–6.11) 143-3p Down-regulated in CSF from ALS patients compared to [71] controls Increased expression in serum-EXs of AD patients vs healthy [23] controls Down-regulated in serum from AD patients [72] Up-regulated in CSF from AD and dementia patients [42] compared to controls Up-regulated in brain of PD mouse model [73] Increased expression in serum ALS compared to controls [74] hsa-miR- Under-represented in the cortex of AD patients [50, 8928 1.99 (0.52–3.46) 1.23 (0.43–2.04) 8.04 (2.02–14.06) Gámez-Valero et al. Translational Neurodegeneration (2019) 8:31 Page 10 of 17 Table 2 Selected microRNAs for further qPCR validation analysis (Continued) miRNA Previously reported in the literature Ref NGS Mean Fold change (2-ΔΔCt) Counts (95% C.I.) CTRL DLB AD 423-5p 65] Increased expression in CSF from AD and dementia patients [42] vs controls Down-regulated expression in PD putamen tissue [73] Differentially expressed in AD blood compared to healthy [64] controls Down-regulated in plasma from PD patients compared to [43] normal controls Up-regulated in CSF from AD patients compared to controls [51] Low expression in CSF from PD patients [75] hsa-miR- In serum, associated to neurofibrillary tangles score in AD [20] 3820 1.18 (0.63–1.73) 0.79 (0.15–1.44) 15.77 (0.84–30.7) 183-5p patients Differentially expressed in AD and other types of dementia [42] patients vs controls Down-regulated in peripheral blood from ALS patients [48] Decreased expression is associated to PD [76] Literature search was performed in different databases for the top most abundant miRNAs found by the exploratory Next Generation Sequencing (NGS) study. PubMed, and the Nervous System Disease NcRNAome Atlas (NSDNA) (30) were used to explore their relation with neurodegeneation-related processes. Total read count for each miRNA by NGS is shown; For qPCR analysis, mean and 95% C.I. is shown Key: EVs Extracellular vesicles, AD Alzheimer’s disease, DLB Dementia with Lewy bodies, CSF Cerebrospinal fluid, PD Parkinson’s disease, MS Multiple Sclerosis, ALS Amyotrophic lateral sclerosis, EXs Exosomes, FTD Frontotemporal dementia, MCI Mild cognitive impairment Validation phase: qPCR analysis of selected microRNAs in microRNA target prediction and affected pathways DLB, AD and control cohorts The six significantly and differentially down-regulated Three groups of samples including 11 DLB patients, 11 miRNAs in AD compared to controls and/or DLB were age-matched controls and 10 AD patients were used in screened for their possible target genes using miRGate the expression analysis by qPCR of the 15 selected miR- software. Up to 217 genes were predicted to be poten- NAs. Confirming our previous analyses by NGS, no dif- tially regulated by these miRNAs and already validated ferences were found for any of these miRNAs between as targets by 2 or more tools. These genes were further DLB and healthy controls (Fig. 3). analysed and networked using String and Panther data- In contrast, two miRNAs were differentially down- bases. Most of the analysed genes were related to meta- regulated in AD patients compared to DLB and bolic processes (p = 1.6•10E-9), specifically protein aged-matched controls (Fig. 3) – specifically hsa- metabolic processes (p = 5.1•10E-7). Among them, 37 miR-451a (p = 0.0003 and p = 0.0031, vs controls and were involved in the regulation of phosphorylation DLB respectively) and hsa-miR-21-5p (p = 0.0075 and processes (p = 1.8•10E-6); specifically related to MAPK p = 0.0064). Moreover, four miRNAs were also sig- cascades and regulation of protein kinase activity (p = nificantly down-regulated also in AD compared to 0.009 and p = 1.2•10E-6, respectively). GO analysis for the control cohort (Fig. 3) - hsa-miR-23a-3p (p = 0.0016), biological process also revealed the enrichment of re- hsa-miR-126-3p (p = 0.004), hsa-let-7i-5p (p = 0.0154), and sponse to stress (p = 4.3•10E-4), aging-related genes hsa-miR-151a-3p (p = 0.0335)-. The remaining nine (p = 0.01), genes involved in neuronal morphogenesis miRNAs (out of 15) did not show any statistical and differentiation (p = 0.017 and p = 0.04, respectively). difference between the three groups, though three of Particularly, several genes were related to negative them showed a trend to be over-expressed in AD regulation of neurogenesis and cell death (p = 0.003 and compared to DLB and controls –hsa-miR-183-5p, p = 0.01, respectively). Focusing in the analysis of the 2 hsa-miR-24-3p and hsa-miR-423-5p (Fig. 3). A pre- miRNAs differentially expressed in AD vs DLB and/or dictive diagnostic value -based on ROC curves- to controls, hsa-miR-451a and hsa-miR-21-5p, we found discriminate between AD and DLB patients rendered an enrichment of genes related to SMAD protein phos- high specificity and sensitivity (AUC over 0.9) for phorylation (p = 3.8•10E-5). the miRNAs hsa-miR-451a and hsa-miR-21-5p Additionally, we screened the whole list of predicted (Fig. 4). target looking for genes related to “neurodegeneration” Gámez-Valero et al. Translational Neurodegeneration (2019) 8:31 Page 11 of 17 Fig. 3 MiRNA expression levels in DLB, AD and controls. MiRNAs with highest expression differences are shown first. From the 15 analysed miRNAs, a 6-miRNA group showed significantly down-regulated expression in AD samples when compared to the two other cohorts. In all cases, mean and range for fold change are plotted; (*p < 0.05, **p < 0.005, ***p < 0.0005) Gámez-Valero et al. Translational Neurodegeneration (2019) 8:31 Page 12 of 17 Fig. 4 ROC curves for the differentiation between DLB and AD. Highest AUC ROC curves were obtained for hsa-miR-21-5p and hsa-miR-451a and potentially regulated by these six miRNAs. The neurologically unaffected individuals. Of notice, 10 identified genes are listed in Table 3 and include of these miRNAs have been described among the ADAM10 (ADAM Metallopeptidase Domain 10), APP most abundant miRNAs in the human brain [78]. (β-amyloid precursor) and APPBP2 (amyloid protein- No difference was observed in the expression of any binding protein 2). of the analysed miRNAs between DLB and controls, as predicted by NGS results. However, 6 miRNAs Discussion (hsa-miR-451a, hsa-miR-21-5p, hsa-miR-23a-3p, hsa- DLB and AD show an important neuropathological, neuro- miR-126-3p, hsa-let-7i-5p, and has-miR-151a-3p) chemical and neuropsychiatric overlap, hampering correct were significantly down-regulated in AD patients DLB diagnosis, treatment and clinical management. Gen- compared to DLB or controls. Noteworthy, all these etic and molecular characterization of these heterogeneous six miRNAs are also described as extracellular space- and complex disorders will lead to a better handling and or exosome- associated in NCBI/Gene database. Our diagnosis, although the definition of specific, differential results are in line with previous observations on the and early biomarkers is still required. Plasma-EVs may be- reduced expression of hsa-miR-23a-3p in AD serum come a promising reservoir of biomarkers also for neurode- compared to controls [56] and the down-regulation generative disorders. Besides their specific cell-derived of hsa-miR-126 previously described in CSF derived content and the RNase-protected environment [15], EVs vesicles from AD patients [21, 51]. Besides confirm- have been proved to cross the blood-brain barrier [10]. ing these previous studies, our expression results Therefore, although CSF would be the ideal source for spe- further indicate that DLB and AD patients may be cific biomarkers of central nervous system disorders, the distinguished by determining some of these miRNAs, difficulty, invasiveness, morbidity, and risk related to CSF- as specifically shown by the expression profile of collection have paved the way for the analysis of plasma-de- miRNAs hsa-miR-451a and hsa-miR-21-5p. rived EVs also in these pathologic scenarios. In comparison to healthy controls, hsa-miR-451a was In the current study, we analysed for the first time the previously described as down-regulated in CSF-derived miRNA content associated to plasma-EVs from DLB AD-exosomes [55], and hsa-miR-21-5p was down-regu- and AD patients compared to healthy controls. No dif- lated in serum and CSF from AD patients [20, 42]. There- ferences in EV-markers, EV size, and morphology, or fore, both miRNAs were reported as putative biomarkers particle concentration were observed between the differ- for AD. Here we confirmed the down-regulation of these ent cohorts. Our sequencing analysis focused on 238 two miRNAs also in plasma-derived EVs of AD patients miRNAs, most of them previously related to vesicles compared to controls and, additionally to DLB. Finally, in- from human samples [40, 41]. Accordingly, the let-7 creased levels of hsa-let-7i-5p have been described in brain family accounted for around 54% of the identified miR- and CSF of AD patients [42, 50]. Also, up-regulation of NAs, as previously reported [77]. Overall, the NGS ana- hsa-miR-151a-3p has been reported in blood from AD pa- lyses of miRNAs did not reveal significant differences tients in comparison to controls [44, 45]. These results are between DLB and aged-control samples. contradictory to our expression data that showed a reduced Despite this lack of NGS significant results, a set expression in AD versus controls in both cases. Having no of 15 miRNAs previously described in the literature reason to explain this later observation, it has been pro- as associated to neurodegenerative diseases and de- posed that plasma and/or plasma-EV concentration of a mentia were analysed by qPCR in an independent given molecule including miRNA do not always show the group of DLB and an additional group of AD pa- same tendency [79], and the expression levels in peripheral tients compared to a different control cohort of circulation can also differ from those observed in CSF [42]. Gámez-Valero et al. Translational Neurodegeneration (2019) 8:31 Page 13 of 17 Table 3 Neurodegenerative processes-related genes among the target output from the differentially expressed miRNAs Input Gene Input miRNA Start Stop Computational Predictions Confirmed Predictions ADAM10 hsa-let-7i-5p 783 801 Miranda 0 hsa-miR-23a-3p 1477 1499 Miranda 0 hsa-miR-451a 3640 3660 Miranda 0 AKT1 hsa-miR-451a 60 79 Miranda Mirtarbase APP hsa-let-7i-5p 532 553 Miranda 0 APPBP2 hsa-let-7i-5p 3533 3553 Miranda 0 hsa-miR-151a-3p 1868 1888 Miranda 0 hsa-miR-21-5p 4099 4120 Miranda 0 BCL2 hsa-miR-126-3p 4169 4192 Miranda 0 hsa-miR-23a-3p 4400 4418 Miranda 0 CASP2 hsa-let-7i-5p 2265 2285 Miranda 0 hsa-miR-151a-3p 2025 2045 Miranda 0 CASP3 hsa-let-7i-5p 141 160 Miranda 0 hsa-miR-23a-3p 808 828 Miranda 0 CASP8 hsa-miR-21-5p 867 888 Miranda 0 hsa-let-7i-5p 118 139 Miranda 0 CASP10 hsa-miR-23a-3p 2104 2124 Miranda 0 CASP14 hsa-let-7i-5p 56 77 Miranda 0 CCNE2 hsa-miR-126-3p 1212 1233 Miranda Mirtarbase hsa-miR-151a-3p 249 270 Miranda 0 hsa-let-7i-5p 350 370 Miranda 0 COX6B2 hsa-let-7i-5p 69 91 Miranda 0 COX6C hsa-miR-21-5p 312 332 Miranda 0 COX7A1 hsa-let-7i-5p 49 54 Rnahybrid 0 COX7B hsa-let-7i-5p 301 321 Miranda 0 hsa-miR-23a-3p 832 852 Miranda 0 COX11 hsa-miR-151a-3p 1584 1604 Miranda 0 GSK3B hsa-miR-21-5p 912 934 Miranda 0 hsa-let-7i-5p 320 342 Miranda 0 hsa-miR-21-5p 5125 5147 Miranda 0 hsa-miR-23a-3p 988 1008 Miranda 0 MIF hsa-miR-451a 90 109 Miranda Mirtarbase|OncomiRDB| UBE2J1 hsa-miR-23a-3p 1920 1942 Miranda 0 UBE2G2 hsa-let-7i-5p 251 272 Miranda 0 UBE2L3 hsa-miR-451a 322 343 Miranda 0 hsa-miR-23a-3p 1188 1209 Miranda 0 MirGate results for Computational and (when possible) Confirmed target genes are shown (in alphabetical order). The name of the prediction tool reporting each target gene is indicated. “0” means no confirmation was found in the analysis for that specific target. Start and Stop indicate the miRNA binding site (beginning and ending nucleotide) in the targeted gene sequence. Key: ADAM10 ADAM Metallopeptidase Domain 10), APP, β-amyloid precursor), APPBP2 Amyloid protein- binding protein 2, GSK3B Glycogen synthase kinase 3, AKT1 RAC-alpha serine/ threonine-protein kinase, CAB39 Calcium binding protein 39, CASP Caspase, CCNE Cyclin E, COX Cytochrome c oxidase, MIF Macrophage migration inhibitory factor, UBE2 Ubiquitin Conjugating enzyme E2 As a preliminary approach to evaluate the discrimin- discrimination. Nevertheless, given the small number of ation power of these miRNAs, ROC curves for the AD samples analysed in this study, these results must be taken down-regulated miRNAs were calculated. Hsa-miR-21-5p as tentative and as a first proof-of-concept, needing fur- and hsa-miR-451a rendered high AUC values to be ther validation in larger cohorts of patients. Likewise, it considered a putative bio-signature for AD-DLB would be neither accurate to establish a specific Gámez-Valero et al. Translational Neurodegeneration (2019) 8:31 Page 14 of 17 correlation between miRNA levels and the clinical charac- presence of smad2 within amyloid plaques and neurofib- teristics of the studied groups. Moreover, as plasma-EV rillary tangles [89]. associated miRNAs have been analysed, diverse causes, Altogether our data point to a specific reduced expres- such as age, patient characteristics or the existence of con- sion of AD-related miRNAs that would target genes pri- comitant pathologies in these aged patients, could also marily involved in protein phosphorylation cascades and alter and modify these miRNAs, which seem to be the neuropathology of AD. The differential expression of deregulated. these miRNAs in AD versus DLB patients could be con- It is known that different miRNAs can converge on sidered as a putative biomarker for the identification/dis- the same function or be involved in the same molecular crimination between both disorders, which could help pathway. The deregulation and differential expression of neurologist to overcome the clinical and pathological these miRNAs in AD and DLB could be not only a cause overlap between AD and DLB. Nevertheless, as a first or specific result of each disease, but a consequence of a exploratory study, these data have to be further con- common neurodegenerative process. Nevertheless, a pre- firmed in larger cohorts of patients. liminary target gene prediction for the 6 down-regulated miRNAs in our study defined a group of 217 most con- Conclusions firmed target genes (with more than 2 confirmed predic- To our knowledge, this is the first study on the compari- tions), mostly implicated in metabolic processes and son of the miRNA profile associated to plasma-EVs from protein phosphorylation. The role of protein-phosphor- DLB and AD patients. Despite the limited number of ylation during neurodegeneration has been widely de- samples, our study provides preliminary evidence for dif- scribed as important in the spread and accumulation of ferent miRNA expression levels between the two most α-synuclein in DLB or PD brains [80, 81]. Also, the in- common types of degenerative dementia, with changes creased levels of phosphorylated tau in AD have been related to target genes and pathways involved in the considered as an important marker of AD pathogenesis pathogenesis of AD. [82]. Specifically, genes involved in the Pl3K-AKt path- way, such as CAB39 [83], are among the predicted genes Additional files in our analysis. Together with AKT1, CAB39 would be possibly regulated by hsa-miR-451a. Therefore, our data Additional file 1: Purchasing information about miRCURY LNA™ suggest an impairment of this pathway, involved in Universal RT microRNA PCR, Pick-&-Mix, Ready-to-use Panels used in the study. (XLSX 53 kb) neuron differentiation-proliferation and death [84]. Additional file 2: Raw data from the Next Generation Sequencing Other genes related to neurodegeneration were also experiment for DLB and control samples. (XLSX 84 kb) found, although with a lower prediction rate (Table 3). For instance, ADAM10, APPBP2 and APP, identified as Abbreviations hsa-let-7i-5p targets, are directly involved in the intracel- AD: Alzheimer’s disease; ADAM10: ADAM Metallopeptidase Domain 10); lular transport and deposition of β-amyloid peptides [85]. AKT1: RAC-alpha serine/ threonine-protein kinase; ALS: Amyotrophic lateral sclerosis; APP: β-amyloid precursor; APPBP2: Amyloid protein-binding protein As inferred from our results, the reduction of hsa-miR-let- 2; AUC: Area under the curve; CAB39: Calcium binding protein 39; 7i-5p, together with hsa-miR-21-5p and hsa-miR-151a-3p, CASP: Caspase; CCNE: Cyclin E; COX: Cytochrome c oxidase; could alter APPBP2 and APP expression resulting in an Cq: Quantification cycle; Cryo-EM: Cryo-electron microscopy; CSF: Cerebrospinal fluid; DLB: Dementia with Lewy bodies; EVs: Extracellular incorrect APP cleavage and promoting β-amyloid accu- vesicles; EXs: Exosomes; FTD: Frontotemporal dementia; GSK3B: Glycogen mulation. Other genes like GSK3B (Glycogen synthase synthase kinase 3; ISEV: International Society for Extracellular Vesicles; kinase 3) may play a role in neuroinflammation [86], neur- KEEG: Kyoto Encyclopedia of Genes and Genomes; LOO: Leave-one-out; MCI: Mild cognitive impairment; MFI: Mean fluorescence intensity; onal apoptosis and accumulation of phosphorylated tau - MIF: Macrophage migration inhibitory factor; miRNA: Micro RNA; MS: Multiple in AD [87]. In our study, several miRNAs targeting sclerosis; MVs: Microvesicles; NGS: Next Generation Sequencing; NSDNA: The GSK3B were down-regulated in AD samples. Further- Nervous System Disease NcRNAome Atlas; PD: Parkinson’s disease; ROC: Receiver Operating Characteristic; SD: Standard desviation; SEC: Size- more, cell-death related genes, such as BCL2, cyclins in- Exclusion chromatography; TMM: Trimmed means of M-values; volved in cell cycle or genes involved in the degradation of UBE2: Ubiquitin Conjugating enzyme E2 selective proteins (including β-amyloid peptide) such as Acknowledgements proteasomal proteins were also present in the predicted We would like to express our profound thanks to Dr. María Yáñez-Mó network. The impairment of the proteasomal pathway (Unidad de Investigación, Hospital Sta Cristina, IIS-IP; Departamento Biología would increase β-amyloid deposition and promote AD Molecular/CBM-SO, UAM) and Dr. Francisco Sánchez-Madrid (Servicio de Inmunología, Hospital Universitario de la Princesa, IIS-IP, UAM; Cell-cell pathology [88]. Of notice, among the miRNAs and pos- Communication Laboratory, CNIC) for antibodies anti-CD9, anti-CD63 and sible target genes differentially regulated in DLB vs AD anti-CD81. The authors especially express gratitude to Marco A. Fernández (hsa-miR-21-5p and hsa-miR-451a) we found genes re- (Flow Cytometry Unit, IGTP) and Pablo Castro Hartmann (Electron Microscopy Unit, UAB). Also, to Dr. Hernando Del Portillo’s group (ICREA Research lated to SMAD protein phosphorylation. The role of Professor at ISGLOBALIGTP) for kindly provide us antibody anti-CD5L. We also SMAD proteins in AD has been described, by the thank Anna Oliveira from the Genomics Unit (Health Sciences Research Gámez-Valero et al. Translational Neurodegeneration (2019) 8:31 Page 15 of 17 Institute Germans Trias i Pujol), Dr. Mireia Coma from ANAXOMICS Biotech 8. Court FA, Midha R, Cisterna BA, Grochmal J, Shakhbazau A, Hendriks WT, et S.L. (Barcelona) and Dr. Sonia Jansa (BioNova científica, S.L.) for their support al. Morphological evidence for a transport of ribosomes from Schwann cells and guiding for genomic data processing and analysis. to regenerating axons. Glia. 2011;59:1529–39. 9. Lachenal G, Pernet-Gallay K, Chivet M, Hemming FJ, Belly A, Bodon G, et al. Authors’ contributions Release of exosomes from differentiated neurons and its regulation by FBS and KB Conceived and designed the project. AG-V, FBS and KB synaptic glutamatergic activity. Mol Cell Neurosci. 2011;46:409–18. conceived and decided the experiments, and revised the manuscript. AG-V 10. García-Romero N, Carrión-Navarro J, Esteban-Rubio S, Lázaro-Ibáñez E, Peris- performed the experiments and wrote the manuscript. JC, DV, LI, RR, RA Celda M, Alonso MM, et al. DNA sequences within glioma-derived recruited patients and collected samples from patients and healthy controls. extracellular vesicles can cross the intact blood-brain barrier and be RA critically revised the discussion. MPA provided assistance analysing detected in peripheral blood of patients. Oncotarget. 2017;8:1416–28. genomic data and discussing the results. All authors read and approved the 11. Kalani A, Tyagi A, Tyagi N. Exosomes: mediators of neurodegeneration, final manuscript. neuroprotection and therapeutics. Mol Neurobiol. 2014;49(1):590–600. 12. Grey M, Dunning CJ, Gaspar R, Grey C, Brundin P, Sparr E, et al. Acceleration Funding of α-synuclein aggregation by exosomes. J Biol Chem. 2015;290:2969–82. This work was supported by Spain’s Ministry of Health, project PI15/00216, 13. Polanco JC, Scicluna BJ, Hill AF, Götz J. Extracellular vesicles isolated from integrated in the National R + D + I and funded by the ISCIII and the the brains of rTg4510 mice seed tau protein aggregation in a threshold- European Regional Development Fund. This work was also supported by the dependent manner. J Biol Chem. 2016;291(24):12445–66. MaratóTV3 grant 201405/10. 14. Stuendl A, Kunadt M, Kruse N, Bartels C, Moebius W, Danzer KM, et al. Induction of α-synuclein aggregate formation by CSF exosomes from patients with Parkinson’s disease and dementia with Lewy bodies. Brain. 2016;139:481–94. Availability of data and materials 15. Bellingham SA, Hill AF. Analysis of miRNA signatures in neurodegenerative The datasets used and/or analysed during the current study are available prion disease. Methods Mol Biol. 2017;1658:67–80. from the corresponding author on reasonable request. 16. Marques TM, Kuiperij HB, Bruinsma IB, van Rumund A, Aerts MB, Esselink RAJ, et al. MicroRNAs in cerebrospinal fluid as potential biomarkers for Ethics approval and consent to participate Parkinson's disease and multiple system atrophy. Mol Neurobiol. 2017; The Clinical Research Ethics Committee of our institution (Research Institute 54(10):7736–45. Germans Trias i Pujol) approved the applied protocol and from each subject, 17. Yáñez-Mó M, Siljander PR, Andreu Z, Zavec AB, Borràs FE, Buzas EI, et al. a written informed consent was obtained according to the Declaration of Biological properties of extracellular vesicles and their physiological functions. J Helsinki Principles in order to use samples and results for publication. Extracell Vesicles. 2015;4:27066. https://doi.org/10.3402/jev.v4.27066. 18. Cheng L, Sharples RA, Scicluna BJ, Hill AF. Exosomes provide a protective Consent for publication and enriched source of miRNA for biomarker profiling compared to The manuscript has been read and approved for submission by all authors intracellular and cell-free blood. J Extracell Vesicles. 2014;26:3. who contributed to the study. The Ethics Committee also approved the use 19. Riancho J, Vázquez-Higuera JL, Pozueta A, Lage C, Kazimierczak M, Bravo M, of the results from human samples for publication. et al. MicroRNA profile in patients with Alzheimer’s disease: analysis of miR- 9-5p and miR-598 in raw and exosome enriched cerebrospinal fluid Competing interests samples. J Alzheimers Dis. 2017;57(2):483–91. The authors declare that they have no competing interests. 20. Burgos K, Malenica I, Metpally R, Courtright A, Rakela B, Beach T, et al. Profiles of extracellular miRNA in cerebrospinal fluid and serum from Author details patients with Alzheimer's and Parkinson's diseases correlate with disease 1 Department of Pathology, Health Sciences Research Institute Germans Trias i status and features of pathology. PLoS One. 2014;9(5):e94839. Pujol (IGTP), Universitat Autònoma de Barcelona, Badalona 08916, Spain. 21. Gui Y, Liu H, Zhang L, Lv W, Hu X. Altered microRNA profiles in 2 REMAR-IVECAT group, Health Sciences Research Institute Germans Trias i cerebrospinal fluid exosome in Parkinson disease and Alzheimer’s disease. Pujol (IGTP), 08916 Badalona, Spain. 3Department of Neurology, Hospital Oncotarget. 2015;6(35):37043–53. Universitari de Bellvitge, L’Hospitalet de Llobregat 08907, Spain. 4Department 22. Lugli G, Cohen AM, Bennet DA, Shah RC, Fields CJ, Hernandez AG, et al. of Neurology, Hospital Universitari Germans Trias i Pujol, 08916 Badalona, Plasma exosomal miRNAs in persons with and without Alzheimer disease: Spain. 5Genomic and Microscopy facilities, Health Sciences Research Institute altered expression and prospects for biomarkers. PlosOne. 2015;10(10): Germans Trias i Pujol (IGTP), 08916 Badalona, Spain. 6Department of Cell e0139233. Biology, Physiology and Immunology, Universitat Autònoma de Barcelona 23. Cheng L, Doecke JD, Sharple RA. Prognostic serum miRNA biomarkers associated (UAB), Barcelona, Spain. with Alzheimer’s disease shows concordance with neuropsychological and neuroimaging assessment. Mol Psychiatry. 2015;20:1188–96. Received: 4 April 2019 Accepted: 19 August 2019 24. Lynöe N, Sandlund M, Dahlqvist G, Jacobsson L. Informed consent: study of quality of information given to participants in a clinical trial. BMJ. 1991;303:610–3. References 25. McKeith IG, Dickson DW, Lowe J, Emre M, O'Brien JT, Feldman H, et al. 1. Harada CN, Natelson-Love MC, Triebel K. Normal cognitive aging. Clin Diagnosis and management of dementia with Lewy bodies: third report of Geriatr Med. 2013;29(4):737–52. the DLB consortium. Neurology. 2005;65:1863–72. 2. Prince MJ. The global impact of dementia. An analysis of prevalence, 26. Khachaturian ZS. Revised criteria for diagnosis of Alzheimer's disease: incidence, costs and trends. World Alzheimer report; 2016. National Institute on Aging-Alzheimer's Association diagnostic guidelines for 3. Ahmed RM, Paterson RW, Warren JD, Zetterberg H, O'Brien JT, Fox NC, et al. Alzheimer's disease. Alzh Dement. 2011;7(3):253–6. Biomarkers in dementia: clinical utility and new directions. J Neurol 27. Théry C, Witwer KW, Aikawa E, Alcaraz MJ, Anderson JD, Andriantsitohaina Neurosurg Psychiatry. 2014;85(12):1426–34. R, et al. Minimal information for studies of extracellular vesicles 2018 4. Przedborski S, Vila M, Jackson-Lewis V. Neurodegeneration: what is it and (MISEV2018): a position statement of the international society for where are we? J Clin Invest. 2003;111:3–10. extracellular vesicles and update of the MISEV2014 guidelines. J Extracell 5. Jellinger KA. Dementia with Lewy bodies and Parkinson’s disease-dementia: Vesicles. 2019;7:1535750. current concepts and controversies. J Neural Transm. 2018;125(4):615–50. 28. Gámez-Valero A, Monguió-Tortajada M, Carreras-Planella L, Franquesa M, 6. Breitve MH, Chwiszczuk LJ, Hynninen MJ, Rongve A, Brønnick K, Janvin C, et Beyer K, Borràs FE. Size-exclusion chromatography-based isolation minimally al. A systematic review of cognitive decline in dementia with Lewy bodies alters extracellular Vesicles' characteristics compared to precipitating agents. versus Alzheimer’s disease. Alzheimers Res Ther. 2014;6:53. Sci Rep. 2016;6:33641. 7. Mark RE, Griffin WST. Dementia with Lewy bodies: definition, diagnosis, and 29. Lozano-Ramos I, Bancu I, Oliveira-Tercero A, Armengol MP, Menezes-Neto A, pathogenic relationship to Alzheimer’s disease. Neuropsychiatr Dis Treat. Del Portillo HA, et al. Size-exclusion chromatography-based enrichment of 2007;3(5):619–25. extracellular vesicles from urine samples. J Extracell vesicles. 2015;4:27369. Gámez-Valero et al. Translational Neurodegeneration (2019) 8:31 Page 16 of 17 30. Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illumina 54. Prabhakar P, Chandra SR, Christopher R. Circulating microRNAs as potential sequence data. Bioinformatics. 2014;30:2114–20. biomarkers for the identification of vascular dementia due to cerebral small 31. Langmead B, Trapnell C, Pop M, et al. Ultrafast and memory-efficient vessel disease. Age Ageing. 2017;46:861–4. alignment of short DNA sequences to the human genome. Genome Biol. 55. McKeever PM, Schneider R, Taghdiri F, Weichert A, Multani N, Brown 2009;10(3):R25. RA, et al. MicroRNA expression levels are altered in the cerebrospinal 32. Robinson MD, Oshlack A. A scaling normalization method for differential fluid of patients with Young-onset Alzheimer's disease. Mol Neurobiol. expression analysis of RNA-seq data. Genome Biol. 2010;11(3):R25. 2018;55(12):8826–41. 33. Lowry R. Concepts & Applications of inferential statisticsRetrieved March; 2011. 56. Galimberti D, Villa C, Fenoglio C, Serpente M, Ghezzi L, Cioffi SM, et al. 34. Wang J, Cao Y, Zhang H, Wang T, Tian Q, Lu X, et al. NSDNA: a Circulating miRNAs as potential biomarkers in Alzheimer's disease. J manually curated database of experimentally supported ncRNAs Alzheimers Dis. 2014;42(4):1261–7. associated with nervous system diseases. Nucleic Acids Res. 2017; 57. Sanders KA, Benton MC, Lea RA, Maltby VE, Agland S, Griffin N, et al. 45(D1):D902–7. Next-generation sequencing reveals broad down-regulation of 35. EV-TRACK Consortium, Van Deun J, Mestdagh P, Agostinis P, Akay Ö, Anand microRNAs in secondary progressive multiple sclerosis CD4+ T cells. S, et al. EV-TRACK: transparent reporting and centralizing knowledge in Clin Epigenetics. 2016;8(1):87. extracellular vesicle research. Nat Methods. 2017;14(3):228–32. 58. Hara N, Kikuchi M, Miyashita A, Hatsuta H, Saito Y, Kasuga K, et al. Serum 36. Andrés-León E, González Peña D, Gómez-López G, Pisano DG. miRGate: a microRNA miR-501-3p as a potential biomarker related to the progression curated database of human, mouse and rat miRNA-mRNA targets. Database of Alzheimer's disease. Acta Neuropathol Commun. 2017;5(1):10. (Oxford). 2015;8:bav035. 59. Takahashi I, Hama Y, Matsushima M, Hirotani M, Kano T, Hohzen H, et al. 37. Szklarczyk D, Morris JH, Cook H, Kuhn M, Wyder S, Simonovic M, et al. Identification of plasma microRNAs as a biomarker of sporadic amyotrophic The STRING database in 2017: quality-controlled protein-protein lateral sclerosis. Mol Brain. 2015;8(1):67. association networks, made broadly accessible. Nucleic Acids Res. 2017; 60. Leggio L, Vivarelli S, L'Episcopo F, Tirolo C, Caniglia S, Testa N, et al. 45:D362–8. microRNAs in Parkinson's disease: from pathogenesis to novel diagnostic 38. Mi H, Huang X, Muruganujan A, Tang H, Mills C, Kang D, et al. PANTHER and therapeutic approaches. Int J Mol Sci. 2017;18(12):2698. version 11: expanded annotation data from gene ontology and reactome 61. Botta-Orfila T, Morató X, Compta Y, Lozano JJ, Falgàs N, Valldeoriola F, et al. pathways, and data analysis tool enhancements. Nucleic Acids Res. 2017; Identification of blood serum micro-RNAs associated with idiopathic and 45(D1):D183–9. LRRK2 Parkinson's disease. J Neurosci Res. 2014;92(8):1071–7. 39. De Menezes-Neto A, Sáez MJ, Lozano-Ramos I, Segui-Barber J, Martin- 62. Hoss A. The relationship of microRNAs to clinical features of Huntington's Jaular L, Ullate JM, et al. Size-exclusion chromatography as a stand- and Parkinson's disease. In: Boston University Theses & Dissertations alone methodology identifies novel markers in mass spectrometry database; 2016. https://open.bu.edu/bitstream/handle/2144/14604/Hoss_ analyses of plasma derived vesicles from healthy individuals. J Extracell bu_0017E_11703.pdf?sequence=17. Accessed 20 Oct 2018. Vesicles. 2015;4:1–14. 63. Kumar S, Vijayan M, Reddy PH. MicroRNA-455-3p as a potential peripheral 40. Kim DK, Lee J, Kim SR, et al. EVpedia: a community web portal for biomarker for Alzheimer's disease. Hum Mol Genet. 2017;26(19):3808–22. extracellular vesicles research. Bioinformatics. 2015;31(6):933–9. 64. Nagaraj S, Laskowska-Kaszub K, Dębski KJ, Wojsiat J, Dąbrowski M, 41. Keerthikumar S, Chisanga D, Ariyaratne D, et al. ExoCarta: a web-based Gabryelewicz T, et al. Profile of 6 microRNA in blood plasma distinguishes compendium of exosomal cargo. J Mol Biol. 2016;428(4):688–92. early stage Alzheimer's disease patients from non-demented subjects. 42. Sorensen SS, Nygaard AB, Christensen T. miRNA expression profiles in Oncotarget. 2017;8(10):16122–43. cerebrospinal fluid and blood of patients with Alzheimer’s disease 65. Wang WX, Huang Q, Hu Y, Stromberg AJ, Nelson PT. Patterns of microRNA and other types of dementia – an exploratory study. Transl expression in normal and early Alzheimer's disease human temporal cortex: Neurodegener. 2016;5:6. white matter versus gray matter. Acta Neuropathol. 2011;121(2):193–205. 43. Chen L, Yang J, Lü J, Cao S, Zhao Q, Yu Z. Identification of aberrant 66. Lehmann SM, Krüger C, Park B, Derkow K, Rosenberger K, Baumgart J, et al. circulating miRNAs in Parkinson's disease plasma samples. Brain Behav. An unconventional role for miRNA: let-7 activates toll-like receptor 7 and 2018;8(4):e00941. causes neurodegeneration. Nat Neurosci. 2012;15(6):827–35. 44. Leidinger P, Backes C, Deutscher S, Schmitt K, Mueller SC, Frese K, et al. A 67. Vallelunga A, Ragusa M, Di Mauro S, Iannitti T, Pilleri M, Biundo R, et al. blood based 12-miRNA signature of Alzheimer disease patients. Genome Identification of circulating microRNAs for the differential diagnosis of Parkinson's Biol. 2013;14:R78. disease and multiple system atrophy. Front Cell Neurosci. 2014;8:156. 45. Satoh J, Kino Y, Niida S. MicroRNA-Seq data analysis pipeline to identify 68. Hu YB, Li CB, Song N, Zou Y, Chen SD, Ren RJ, et al. Diagnostic value of blood biomarkers for Alzheimer’s disease from public data. Biomark Insights. microRNA for Alzheimer's disease: a systematic review and meta-analysis. 2015;10:21–31. Front Aging Neurosci. 2016;8:13. 46. Chang WS, Wang YH, Zhu XT, Wu CJ. Genome-wide profiling of miRNA and 69. Lusardi TA, Phillips JI, Wiedrick JT, Harrington CA, Lind B, Lapidus JA, et al. mRNA expression in Alzheimer's disease. Med Sci Monit. 2017;23:2721–31. MicroRNAs in human cerebrospinal fluid as biomarkers for Alzheimer's 47. Guo R, Fan G, Zhang J, Wu C, Du Y, Ye H, et al. A 9-microRNA signature in disease. J Alzheimers Dis. 2017;55(3):1223–33. serum serves as a noninvasive biomarker in early diagnosis of Alzheimer's 70. Alexandrov PN, Dua P, Hill JM, Bhattacharjee S, Zhao Y, Lukiw WJ. MicroRNA disease. J Alzheimers Dis. 2017;60(4):1365–77. (miRNA) speciation in Alzheimer's disease (AD) cerebrospinal fluid (CSF) and 48. Liguori M, Nuzziello N, Introna A, Consiglio A, Licciulli F, D'Errico E, et al. extracellular fluid (ECF). Int J Biochem Mol Biol. 2012;3:365–73. Dysregulation of MicroRNAs and target genes networks in peripheral blood 71. Freischmidt A, Müller K, Ludolph AC, Weishaupt JH. Systemic dysregulation of patients with sporadic amyotrophic lateral sclerosis. Front Mol Neurosci. of TDP-43 binding microRNAs in amyotrophic lateral sclerosis. Acta 2018;11:288. Neuropathol Commun. 2013;1:42. 49. Gehrke S, Imai Y, Sokol N, Lu B. Pathogenic LRRK2 negatively regulates 72. Dong H, Li J, Huang L, Chen X, Li D, Wang T, et al. Serum MicroRNA profiles microRNA-mediated translational repression. Nature. 2010;466(7306):637–41. serve as novel biomarkers for the diagnosis of Alzheimer's disease. Dis 50. Lau P, Bossers K, Janky R, Salta E, Frigerio CS, Barbash S, et al. Alteration of Markers. 2015;2015:625659. the microRNA network during the progression of Alzheimer's disease. EMBO 73. Martinez B, Peplow PV. MicroRNAs in Parkinson's disease and emerging Mol Med. 2013;5(10):1613–34. therapeutic targets. Neural Regen Res. 2017;12(12):1945–59. 51. Cogswell JP, Ward J, Taylor IA. Identification of miRNA changes in 74. Waller R, Goodall EF, Milo M, Cooper-Knock J, Da Costa M, Hobson E, et al. Alzheimer’s disease brain and CSF yields putative biomarkers and insights Serum miRNAs miR-206, 143-3p and 374b-5p as potential biomarkers for into disease pathways. J Alzh Disease. 2008;14:27–41. amyotrophic lateral sclerosis (ALS). Neurobiol Aging. 2017;55:123–31. 52. Gwon Y, Kam TI, Kim SH, Song S, Park H, Lim B, et al. TOM1 regulates 75. Dos Santos MCT, Barreto-Sanz MA, Correia BRS, Bell R, Widnall C, Perez LT, et neuronal accumulation of amyloid-β oligomers by FcγRIIb2 variant in al. miRNA-based signatures in cerebrospinal fluid as potential diagnostic Alzheimer's disease. J Neurosci. 2018;38(42):9001–18. tools for early stage Parkinson's disease. Oncotarget. 2018;9(25):17455–65. 53. Ebrahimkhani S, Vafaee F, Young PE, Hur SSJ, Hawke S, Devenney E, et al. 76. Roser AE, Caldi Gomes L, Halder R, Jain G, Maass F, Tönges L, et al. miR-182-5p Exosomal microRNA signatures in multiple sclerosis reflect disease status. Sci and miR-183-5p act as GDNF mimics in dopaminergic midbrain neurons. Mol Rep. 2017;7(1):14293. Ther Nucleic Acids. 2018;1(11):9–22. Gámez-Valero et al. Translational Neurodegeneration (2019) 8:31 Page 17 of 17 77. Quek C, Bellingham SA, Jung CH. Defining the purity of exosomes required for diagnostic profiling of small RNA suitable for biomarker discovery. RNA Biol. 2017;14(2):245–58. 78. Lukiw WJ. Evolution and complexity of micro RNA in the human brain. Front Genet. 2012;3:166. 79. Savelyeva A, Kuligina EV, Bariakin DN, Kozlov VV, Ryabchikova EI, Richter VA, et al. Variety of RNAs in peripheral blood cells, plasma, and plasma fractions. Biomed Res Int. 2017;2017:7404912. 80. Castillo-Gonzalez JA, Loera-Arias MJ, Saucedo-Cardenas O, Montes-de-Oca- Luna R, Garcia-Garcia A, Rodriguez-Rocha H. Phosphorylated α-Synuclein- copper complex formation in the pathogenesis of Parkinson's disease. Parkinsons Dis. 2017;2017:9164754. 81. Fujiwara H, Hasegawa M, Dohmae N, Kawashima A, Masliah E, Goldberg MS, et al. Alpha-synuclein is phosphorylated in synucleinopathy lesions. Nat Cell Biol. 2002;4:160–4. 82. Mondragón-Rodríguez S, Perry G, Luna-Muñoz J, Acevedo-Aquino MC, Williams S. Phosphorylation of tau protein at sites Ser 396-404 is one of the earliest events in Alzheimer’s disease and Down syndrome. Neuropathol Appl Neurobiol. 2014;40(2):121–35. 83. Tian Y, Nan Y, Han L, Zhang A, Wang G, Jia Z, et al. MicroRNA miR-451 downregulates the PI3K/AKT pathway through CAB39 in human glioma. Int J Oncol. 2012;40(4):1105–12. 84. Heras-Sandoval D, Ávila-Muñoz E, Arias C. The phosphatidylinositol 3-kinase/ mTor pathway as a therapeutic target for brain aging and neurodegeneration. Pharmaceuticals (Basel). 2011;4(8):1070–87. 85. Kuan YH, Gruebl T, Soba P, Eggert S, Nesic I, Back S, et al. PAT1a modulates intracellular transport and processing of amyloid precursor protein (APP), APLP1, and APLP2*. J Biol Chem. 2006;281(52):40114–23. 86. Golpich M, Amini E, Hemmati F, Ibrahim NM, Rahmani B, Mohamed Z, et al. Glycogen synthase kinase-3 beta (GSK-3β) signaling: implications for Parkinson's disease. Pharmacol Res. 2015;97:16–26. 87. Hernandez F, Lucas JJ, Avila J. GSK3 and tau: two convergence points in Alzheimer’s disease. J Alzheimers Dis. 2013;33(Suppl. 1):S141–4. 88. Hong L, Huang HC, Jiang ZF. Relationship between amyloid-beta and the ubiquitin-proteasome system in Alzheimer's disease. Neurol Res. 2014;36(3):276–82. 89. Lee HG, Ueda M, Zhu X, Perry G, Smith MA. Ectopic expression of phospho- Smad2 in Alzheimer's disease: uncoupling of the transforming growth factor-beta pathway? J Neurosci Res. 2006;84(8):1856–61.