ARTICLE OPEN ACCESS

Immune profiling of plasma-derived extracellularvesicles identifies Parkinson diseaseElena Vacchi MS Jacopo Burrello MD Dario Di Silvestre MS Alessio Burrello MS Sara Bolis MS

Pierluigi Mauri PhD Giuseppe Vassalli MD Carlo W Cereda MD Cinthia Farina PhD Lucio Barile PhD

Alain Kaelin-Lang MD PhD and Giorgia Melli MD PhD

Neurol Neuroimmunol Neuroinflamm 20207e866 doi101212NXI0000000000000866

Correspondence

Dr Melli

giorgiamellieocch

AbstractObjectiveTo develop a diagnostic model based on plasma-derived extracellular vesicle (EV) subpopu-lations in Parkinson disease (PD) and atypical parkinsonism (AP) we applied an innovativeflow cytometric multiplex bead-based platform

MethodsPlasma-derived EVs were isolated from PD matched healthy controls multiple system atrophy(MSA) and AP with tauopathies (AP-Tau) The expression levels of 37 EV surface markerswere measured by flow cytometry and correlated with clinical scales A diagnostic model basedon EV surface markers expression was built via supervised machine learning algorithms andvalidated in an external cohort

ResultsDistinctive pools of EV surface markers related to inflammatory and immune cells stratifiedpatients according to the clinical diagnosis PD and MSA displayed a greater pool of overex-pressed immune markers suggesting a different immune dysregulation in PD and MSA vsAP-Tau The receiver operating characteristic curve analysis of a compound EVmarker showedoptimal diagnostic performance for PD (area under the curve [AUC] 0908 sensitivity 963specificity 789) and MSA (AUC 0974 sensitivity 100 specificity 947) and good ac-curacy for AP-Tau (AUC 0718 sensitivity 778 specificity 895) A diagnostic model basedon EV marker expression correctly classified 889 of patients with reliable diagnostic per-formance after internal and external validations

ConclusionsImmune profiling of plasmatic EVs represents a crucial step toward the identification of bio-markers of disease for PD and AP

Condashfirst authors

From the Laboratory for Biomedical Neurosciences (EV AK-L GM) Neurocenter of Southern Switzerland Ente Ospedaliero Cantonale Faculty of Biomedical Sciences (EV GVLB AK-L GM) Universita della Svizzera Italiana Cellular and Molecular Cardiology Laboratory (JB GV) Cardiocentro Ticino Foundation Lugano Switzerland Proteomic andMetabolomic Laboratory (DDS PM) Institute for Biomedical TechnologiesndashNational Research Council (ITB-CNR) Segrate (Milan) Italy Department of Electrical (AB) Electronic andInformation Engineering ldquoGuglielmo Marconirdquo (DEI) University of Bologna Italy Laboratory for Cardiovascular Theranostics (SB LB) Cardiocentro Ticino Foundation LuganoSwitzerland Neurology Department (CWC AK-L GM) Neurocenter of Southern Switzerland Ente Ospedaliero Cantonale Lugano and Immunobiology of Neurological DisordersLab (CF) Institute of Experimental Neurology (INSpe) and Division of Neuroscience IRCCS San Raffaele Scientific Institute Milan Italy

Go to NeurologyorgNN for full disclosures Funding information is provided at the end of the article

The Article Processing Charge was funded by the authors

This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 40 (CC BY-NC-ND) which permits downloadingand sharing the work provided it is properly cited The work cannot be changed in any way or used commercially without permission from the journal

Copyright copy 2020 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the American Academy of Neurology 1

To date an effective causal treatment for Parkinson disease(PD) is missing and the diagnosis still relies exclusively onmotor symptoms that appear too late for a disease modifyingintervention1 Hence there is urgent need for biomarkers thatcan stratify patients with PD for clinical trials Furthermorethe differential diagnosis between PD and atypical parkin-sonisms (APs) like multiple system atrophy (MSA) is chal-lenging2 According to the misfolded protein aggregatespresent in the brain PD and MSA are collectively termed asalpha-synucleinopathies and are distinct from AP with tauo-pathies (AP-Tau)

Extracellular vesicles (EVs) are a heterogeneous population ofsecreted membrane particles involved into physiologic cell-to-cell communication and transmission of biological signalsEVs are subdivided based on physical characteristics such assize into small (30ndash150 nm) and large (150ndash500 nm) vesi-cles members of the tetraspanin protein family (CD9 CD63and CD81) are considered specific markers of EVs3 CNSneurons release EVs4 able to cross the blood-brain barrier andreach the peripheral blood5 EVs express surface antigenswhich affect the cellular uptake and allow their tracking to thecell of origin6

So far most of the studies on EVs in neurodegenerativediseases focused on their possible role on transmission ofpathologic misfolded proteins and fewer on their functionsin cell-to-cell signaling Indeed immune system is involvedin PD as demonstrated by neuroinflammatory changes inbrain histopathology as well as by elevated immune mark-ers in peripheral blood suggesting that immune systemmay play a primary pathogenic role in PD78 Therefore wehypothesized that circulating EVs carry important in-formation on brain inflammatory immune response andthat their characterization can be exploited for diagnosticpurposes

MethodsStudy designThis was a cross-sectional case-control study aiming (1) tocharacterize distinctive EV subpopulations in plasma of pa-tients with PD MSA and AP-Tau healthy controls (HCs) byimmunophenotyping 37 different membrane proteins using

an innovative flow cytometry multiplex bead-basedplatform910 (2) to correlate the differential expression ofEV surface antigens to clinical scales of gravity and (3) tobuild diagnostic models based on distinctive EV surfaceproteins through supervised machine learning algorithmsFinally because EVs are taken up by surrounding and distantcells we performed a functional evaluation of their proteininteractors with the purpose to highlight protein targets bi-ological pathways and molecular functions potentially af-fected in PD MSA and AP-Tau

SubjectsTwenty-seven patients with idiopathic PD 8 with probableMSA 9 with probable AP-Tau and 19 age-matched HCs forthe PD group were consecutively enrolled from July 2015 toJanuary 2019 These subjects served as the training cohort forthe diagnostic model

Patients were recruited from the movement disorders out-patient clinic at Neurocenter of Southern Switzerland inLugano HCs were recruited among patientsrsquo partners Theinclusion criteria for PD were (1) a definite clinical diagnosisaccording to the UK Parkinsonrsquos Disease Society Brain Bankcriteria for diagnosis1 and (2) no family history and no majorcognitive impairment or major dysautonomic symptoms inthe history The inclusion criteria for AP were based onpublished diagnostic criteria for MSA11 progressive supra-nuclear palsy (PSP)12 and corticobasal degeneration(CBD)13 Each subject underwent blood collection andclinical evaluation Disease gravity was assessed by the Hoehnand Yahr scale (HampY) and Movement Disorder SocietyndashUnified Parkinsonrsquos Disease Rating Scale (MDS-UPDRS)during the off stage cognitive profile by the Mini-MentalState Examination (MMSE) and Montreal Cognitive As-sessment (MoCA) mood disorder by the Beck DepressionInventory II (BDI-II) scale REM sleep behavior disorder(RBD) by the RBD screening questionnaire and olfactoryfunction by Burghart Messtechnik GmbH (olfactory test)Levodopa equivalent daily dose (LEDD) was calculated forpatients with PD and AP14

Exclusion criteria were significant comorbidities diabetesrenal failure thyroid pathology vitamin B12 deficiency HIVinfection syphilis coagulopathy fever acute or chronic in-flammatory diseases and tumors

GlossaryAP = atypical parkinsonism AP-Tau = atypical parkinsonism with tauopathies AUC = area under the curve BDI-II = BeckDepression Inventory II CBD = corticobasal degeneration EV = extracellular vesicle HC = healthy control HampY = Hoehnand Yahr scale KEGG = Kyoto Encyclopedia of Genes and Genomes LEDD = levodopa equivalent daily dose MCSP =melanoma-associated chondroitin sulfate proteoglycan MDS-UPDRS = Movement Disorder SocietyndashUnified ParkinsonrsquosDisease Rating ScaleMFI = median fluorescence intensityMoCA =Montreal Cognitive AssessmentMSA = multiple systematrophy nMFI = normalized median fluorescence intensity NTA = nanoparticle tracking analysis PD = Parkinson diseasePPI = protein-protein interaction PSP = progressive supranuclear palsy RF = random forest ROC = receiver operatingcharacteristic TSG101 = tumor susceptibility gene 101

2 Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 6 | November 2020 NeurologyorgNN

A separate cohort of 40 subjects (20 HC 10 PD 5MSA and 5AP-Tau) served as the validation cohort for the diagnosticmodel (see below the paragraph ldquoDiagnostic modeling andvalidation in an external cohortrdquo)

Standard protocol approvals registrationsand patient consentsSubjects were consecutively included in the NSIPD001 studyaccording to the study protocol that was approved by theCantonal Ethics Committee All enrolled subjects gave writ-ten informed consent to the study in accordance with theDeclaration of Helsinki

Blood collection and plasma preparationTen milliliters of blood were collected into anticoagulantethylenediamine tetraacetic acid (EDTA) tubes in themorning after 4-hour fasting and the following protocol wasperformed to obtain plasma enriched in EVs15 fresh wholeblood was centrifuged for 15 minutes at 1600g at 10degC toeliminate cellular components To further deplete plateletsand cellular debris the supernatant was centrifuged 15 mi-nutes at 3000g at 4degC then 2 consecutive centrifuges wereperformed at 10000g for 15 minutes and 20000g for 30 mi-nutes at 4degC allowing the elimination of apoptotic bodies andlarger EVs (figure 1A) The obtained plasma was aliquotedand stored at minus80degC The storage period varied among sam-ples according to the consecutive enrollment of subjects in thestudy between July 2015 and January 2019

Nanoparticle tracking analysisNanoparticle concentration and diameter were measured byNanoSight LM10 (Malvern Instruments Malvern UK)equipped with a 405-nm laser and nanoparticle trackinganalysis (NTA) 23 software One microliter of plasma wasdiluted 11000 in particle-free phosphate buffered salineThree consequent videos of 60 seconds each were acquiredMinimum expected particle size minimum track length andblur setting were set to automatic and the detection thresholdwas set to 4 to reveal all particles as previously described16

The particle concentration and the distribution graph of theparticle size were determined per each sample by averagingthe results from the analysis of 3 independent videos

MACSPlex exosome assay and flowcytometry analysisThe screening approach (MACSPlex Human Exosome KitMiltenyi Bergisch Gladbach Germany) was previouslydescribed910 Briefly it is based on 48-μm diameter poly-styrene beads labeled with different amounts of 2 dyes (phy-coerythrin and fluorescein isothiocyanate) to generate 39different bead subsets discriminable by flow cytometry analysisEach bead subset is conjugated with a different capture anti-body that recognizes EVs carrying the respective antigen (37EV surface epitopes plus 2 isotype controls) The list of 37antigens is reported in table e-1 (linkslwwcomNXIA293)After beads + sample overnight incubation EVs bound to beadsare detected by allophycocyanin-conjugated anti-CD9 anti-

CD63 and anti-CD81 antibodies (figure 1A) Plasma samples(60 μL) diluted 12 in buffer solution were analyzed with theMACSQuant Analyzer-10 flow cytometer (Miltenyi) Triggersfor the side scatter and the forward scatter were selected toconfine the measurement on the multiplex beads A blankcontrol composed only by MACSPlex Buffer and incubatedwith beads and detection antibodies was used to measure thebackground signal Each EV markerrsquos median fluorescence in-tensity (MFI) was normalized to the mean MFI for specific EVmarkers (CD9 CD63 and CD81) obtaining normalized MFI(nMFI) All analyses were based on nMFI values Samples wereanalyzed blindly to the clinical diagnosis

To test the reliabilityspecificity of MACSPlex Human Exo-some Kit for EVs we compared the procedure described abovewith andwithout EV enrichment by ultracentrifugation and wefound no differences between procedures (figure e-1 linkslwwcomNXIA293) Therefore plasma samples were directlyprocessed without EV enrichment by ultracentrifugation

Technical consistency and reproducibility of the assay wereconfirmed by analyzing repeatedly the same sample and byassessing plasma from the same subject at different timepoints (figure e-2 linkslwwcomNXIA293)

Western blot analysisWestern blot analysis was performed on 100 μL of plasmasamples incubated overnight with 5 μL of MACSPlex de-tection beads at 10degC at 800 rpm The next day the un-bounded fraction was discarded and samples were lysed withradioimmunoprecipitation assay buffer Total proteins wereseparated on a gradient sodium dodecyl sulphate-polyacrylamide gel electrophoresis 4ndash12 gel and trans-ferred onto polyvinylidene difluoride membrane The blotwas incubated with the following primary antibodies anti-Alix(rabbit polyclonal Abcam Cambridge UK 11000) antindashtumor susceptibility gene 101 (TSG101) (rabbit polyclonalAbcam 11000) anti-CD81 (mouse monoclonal ThermoFisher Scientific Waltham MA 1300) antindashapolipoproteinA1 (APOA1) (rabbit polyclonal Abcam 1300) and anti-GRP94 (rabbit polyclonal Abcam 1500)

Network analysis of EV surface markersrsquoprotein interactorsProtein interactors of differentially expressed EV surfacemarkers were retrieved by Cytoscape PESCA plugin17 and aglobal Homo sapiens protein-protein interaction (PPI) networkof 1588 nodes and 36984 edges was reconstructed For eachquantitative comparison (PD vs HC MSA vs HC AP-Tau vsHC) a specific PPI subnetwork was built considering the firstneighbors of each EV surface protein Each subnetwork wasanalyzed at a topological level by Cytoscape Centiscape plu-gin18 to select putative hubs and bottlenecks we took intoaccount the network size and only nodes with all BetweennessBridging and Centroid values above the average calculated onthe corresponding whole network were retained as previouslyreported1920 At the same time nodes belonging to each

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 6 | November 2020 3

Figure 1 EV enrichment MACSPlex exosome assay and EV characterization

(A) Protocol for EV enrichment and MACSPlex exosome assay Blood collected into anticoagulant EDTA tubes underwent serial centrifugation to eliminatecellular components and larger EVs Plasma sampleswere incubated overnight with dye-labeled capture beads coatedwith antibodies against 37 different EVsurface antigens Detection antibodies against CD9 CD63 and CD81 were then added and incubated for 1 hour After washing steps samples were analyzedby flow cytometry (B) Nanoparticle concentration (NmL plasma) by nanoparticle tracking analysis (NTA) stratified for diameter (smaller nanoparticles30ndash150 nm larger nanoparticles 151ndash500 nm) (C) Mean median fluorescence intensity (MFI) for CD9 CD63 and CD81 at flow cytometry analysis (D)Correlation betweenmeanMFI of CD9minusCD63minusCD81 andNmL by NTA the regression line is reported in red with 95 CI (E)Western blot of samples fromHCPD MSA and AP-Tau subjects after immunocapturing comparedwith whole plasma (dilution 1100) showing the presence of specific EVmarkers (CD81 Alixtumor susceptibility gene 101) and the absence of plasma contaminants (apolipoprotein A1 GRP94) Data are expressed asmedian and interquartile range pvalues lt 005were considered significant (p lt 005 p lt 001 p lt 0001) AP-Tau = atypical parkinsonismwith tauopathies EV = extracellular vesicle HC =healthy control MSA = multiple system atrophy PD = Parkinson disease

4 Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 6 | November 2020 NeurologyorgNN

subnetwork were evaluated at a functional level by DAVID21

and the most enriched Kyoto Encyclopedia of Genes and Ge-nomes (KEGG) pathway databases Molecular functions wereextracted specificallyH sapiens set as background count gt 5and p lt 0001 corrected by the Bonferroni test

Statistical analysisStatistical analyses were performed with IBM SPSS Statistics220 PYTHON 27 and GraphPad PRISM 70a Variabledistribution was assessed by the Kolmogorov-Smirnov testNormally distributed variables (age) were expressed as meanplusmn SD and analyzed by the 1-way analysis of variance test withthe post hoc Bonferroni test for multiple comparisons Non-normally distributed variables (disease duration HampY MDS-UPDRS BDI-II MMSE MoCA olfactory test RBD LEDDNTA and MACSPlex analysis) were expressed as mediansand interquartile range and analyzed using the Kruskal-Wallistest Categorical variables (sex) were expressed as absolutenumber and percentage () and analyzed by χ2 or Fisherexact tests Univariate logistic regression analysis was per-formed to assess the ORs Receiver operating characteristic(ROC) curve analysis was used to evaluate the area under thecurve (AUC) and to compare diagnostic performances ofselected variables The Youden index (J = Sensitivity +Specificity minus 1) was calculated to determine the cutoffwith thegreater accuracy Correlations were evaluated by the PearsonR test and regression curve analysis correlations were con-sidered strong for R between |10| and |05| moderate be-tween |05| and |03| and weak between |03 and |01| Ap value less than 005 was considered significant

Diagnostic modeling and validationMachine learning supervised algorithms are exploited in clinicalpractice to formulate predictions of selected outcomes basedon a given set of labeled paired input-output training sampledata2223 The linear discriminant analysis was used to build the3D canonical plot (figure 2B) canonical components 1 2 and3 were calculated from weighted linear combinations of vari-ables to maximize separation between the 4 groups (HC PDMSA and AP-Tau) in the plot each patient is represented by apoint the center of the spheres indicates the mean of (ca-nonical 1 canonical 2 canonical 3) for each diagnosis andspheres include patients with a linear combination coefficientthat falls within the mean plusmn SD (canonical 1 plusmn SD canonical 2plusmn SD canonical 3 plusmn SD) A diagnostic model was built througha random forest (RF) classification algorithm on the trainingcohort (n = 63) the algorithm created 20 different classifica-tion trees with a maximumnumber of 8 splits for each tree Thediagnosis derives from the outcome of each classification tree ofthe RF for example if at least 11 of 20 trees of the RF predictPD the patient will be classified as PD The model was vali-dated by a leave-one-out algorithm (internal validation) and ina different cohort (n = 40) (external validation) The leave-one-out validation was used to exclude overfitting bias and toevaluate generalizability of the model briefly the algorithm istrained on nminus1 patients (where ldquonrdquo is the total number ofpatients) and the remaining patient is used to test the model

The test patient is then changed and accordingly the trainingsubgroup The process is repeated a total of n times with thetest patient rotating at each round and the remaining subgroupused for model training The external validation was performedwith the same RF model trained on the training cohort

Data availabilityThe raw data that support the findings of this article areavailable on request to the corresponding author

ResultsDemographic and clinical characteristics ofstudy groupsDemographic data and clinical assessments for each group aresummarized in table 1 TheMSAgroup included 4MSA-C and 4MSA-P the AP-Tau group included 6 patients with probablePSP and 3 with possible CBD (table e-2 linkslwwcomNXIA293) Subjects with AP-Tau were significantly older than HCSex ratio and disease duration did not differ across groups It isknown that AP is characterized by a more aggressive diseasecourse than PD indeed MSA and AP-Tau had a more severedisease gravity measured by the HampY and by theMDS-UPDRSin addition they displayed a higher cognitive impairment mea-sured by the MMSE and MoCA Finally subjects with AP-Tauresulted more depressed than PD as measured by the BDI-IILEDD was not different between groups of patients

The PD group shows an increased numberof EVsNTA showed that the PD group had the highest number ofnanoparticlesmL compared with HC and AP-Tau (p =0001) not with MSA whereas no differences in diameterwere found between groups (figure 1B table e-3 linkslwwcomNXIA293) Because NTA is not specific for EVs weused the MFI of CD9CD63CD81 (specific markers ofEVs) by flow cytometry analysis as a measure of EV con-centration MeanMFI of CD9CD63CD81 was significantlyhigher in PD compared with HC (p = 0023) and AP-Tau (p =0037) not with MSA (figure 1C) Importantly meanMFI forCD9CD63CD81 correlated with nanoparticle concentra-tion obtained with NTA analysis (figure 1D)

EVs were furtherly characterized according to current stan-dard guidelines3 After EV immunocapture by MACSPlex kitcapture beads we performed a Western blot analysis showingthe presence of EV-specific luminal markers (TSG101 Alix)EV-specific tetraspanin (CD81) and the absence of con-taminants (APOA1 and GPR94) (figure 1E) These resultsconfirm the presence of EVs and the absence of relevantcontamination in samples analyzed by flow cytometry

EV surface antigen expression differsbetween groupsAmong the 37 EV surface markers 17 resulted differentiallyexpressed between groups Sixteen markers differed betweenPD and HC T-cell (CD4) B-cell (CD19) leukocyte (CD45)

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 6 | November 2020 5

and antigen-presenting cell (CD1c) related markers 8 involvedin cell adhesion (CD2 CD11c CD31 CD41 CD42a CD62CD146 and melanoma-associated chondroitin sulfate pro-teoglycan [MCSP]) 3 in immune cell activation (CD25 CD40and CD209) and the molecules of major histocompatibilitycomplex class I human leukocyte antigen-ABC (HLA-ABC)

Twelve markers were different between MSA and HC T cellrelated (CD4) B cell related (CD19) involved in cell adhesion(CD29 CD2 CD11c CD31 CD42a CD62 CD146 and

CD209) immune cell activation (CD40) and HLA-ABC TheAP-Tau and HC groups differ for only 4 markers CD25 CD31CD40 and CD42a No significant difference was found betweenthe PD and MSA groups whereas CD2 resulted different be-tween PD and AP-Tau CD2 and CD19 discriminated betweenpatients withMSA andAP-Tau The nMFI of each EVmarker inall groups is reported in table e-4 (linkslwwcomNXIA293)

The heat map (figure 2A) built on the differentially expressedEV markers highlighted HC as a homogenous group with

Figure 2 Differential expression of extracellular vesicle (EV)-surface markers

Patientsrsquo stratification for diagnosisand EV surface marker expression(expressed as normalized medianfluorescence intensity [nMFI]) (A)Heatmap representation of the 17 EVsurface markers differentiallyexpressed between patients with PDMSA and AP-Tau and HCs (purple =low nMFI yellow = high nMFI) (B)Canonical plot showing patientsaccording to the diagnosis PD red vsMSA orange vs AP-Tau gray vs HCblue themodel was built consideringthe 37 EV surface markers analyzedby flow cytometry The axes of theplot (canonical 1 canonical 2 andcanonical 3) were calculated fromweighted linear combinations of var-iables to maximize separation be-tween the 4 groups Each subject isrepresented by a point and spheresinclude patients with a linear combi-nation coefficient that falls within themean plusmn SD (canonical 1 plusmn SD canon-ical 2 plusmn SD canonical plusmn SD) AP-Tau =atypical parkinsonism with tauo-pathies HC = healthy control MSA =multiple system atrophy PD = Par-kinson disease

6 Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 6 | November 2020 NeurologyorgNN

Table 1 Demographic data and clinical scores

Variable HC [n = 19] PD [n = 27]

AP

Overall pvalue

Pairwise comparisons

MSA [n = 8] AP-Tau [n = 9] HCvs PDHC vsMSA

HC vs AP-Tau

PD vsMSA

PD vs AP-Tau

MSA vs AP-Tau

Age (y) 61 plusmn 82 66 plusmn 118 68 plusmn 86 74 plusmn 52 0013 0556 0729 0008 1000 0184 0924

Sex (ref male) 10 (526) 17 (630) 2 (250) 4 (444) 0279 mdash mdash mdash mdash mdash mdash

Disease duration (y) mdash 40 [20ndash80] 55 [13ndash78] 40 [25ndash55] 0863 mdash mdash mdash mdash mdash mdash

HampY mdash 20 [10ndash30] 50 [40ndash50] 40 [30ndash55] lt0001 mdash mdash mdash lt0001 0001 1000

MDS-UPDRS mdash 230 [130ndash345] 425 [380ndash430] 405 [288ndash815] 0043 mdash mdash mdash 0509 0277 1000

BDI-II mdash 50 [30ndash85] 80 [28ndash158] 145 [118ndash190] 0008 mdash mdash mdash 1000 0012 0682

MMSE mdash 300 [290ndash300] 260 [240ndash290] 260 [220ndash280] lt0001 mdash mdash mdash 0049 0002 1000

MoCA mdash 270 [238ndash290] 245 [173ndash270] 200 [145ndash230] 0016 mdash mdash mdash 0955 0041 1000

Olfactory test mdash 70 [40ndash90] 90 [75ndash103] 70 [48ndash88] 0131 mdash mdash mdash mdash mdash mdash

RBD mdash 30 [18ndash50] 30 [10ndash58] 30 [05ndash45] 0875 mdash mdash mdash mdash mdash mdash

LEDD mdash 5625[2025ndash7375]

3750[1080ndash3750]

2500[1000ndash4520]

0448 mdash mdash mdash mdash mdash mdash

Abbreviations AP = atypical parkinsonism AP-Tau = atypical parkinsonism with tauopathies BDI-II = Beck Depression Inventory II HC = healthy control HampY = Hoehn and Yahr scale LEDD = levodopa equivalent daily doseMDS-UPDRS =Movement Disorder SocietyndashUnified Parkinsonrsquos Disease Rating Scale MMSE =Mini-Mental State Examination MoCA =Montreal Cognitive Assessment MSA =multiple system atrophy PD = Parkinson diseaseRBD = REM sleep behavior disorder screening questionnaireClinical characteristics of patients with PD MSA and AP-Tau compared with HCs p Values lt005 were considered significant and shown in bold

Neurolo

gyorgN

NNeu

rologyN

euroim

munology

ampNeuroinflam

mation

|Volum

e7N

umber

6|

Novem

ber2020

7

relatively low expression of EV markers in analogy to the AP-Tau group whereas PD and MSA were characterized byhigher levels of expression

Furthermore a linear discriminant analysis model based ondifferential expression of all EV markers allowed the separa-tion of subjects according to their diagnosis as shown in thecanonical plot (figure 2B)

Protein network hubs and functional pathwayanalysis of EV surface antigensThe most relevant interactors of differentially expressed EVmarkers were selected by PPI network topological analysis interms of hubs Hubs refer to proteins with the greater numberof connections within the cell or occupying crucial networkpositions suggesting therefore a critical role on the control ofinformation flow over the network2024 Analysis of hubs takesaccount of networks size and only nodes with values above theaverage normalization for the level of connections in the totalnetwork are selected1920 Hubs for PD vs HC andMSA vs HCcomparisons were coincident (SP1 MSN ITGB2 EZRC1QBP and CARL) whereas the network on AP-Tau vs HCshowed a different set of proteins (FLNA FN1 GP1BBHSPA4 NFKB1 STAT3 VIM VWF and YWHAZ) (figure 3AndashC table e-5 linkslwwcomNXIA293) Similarities be-tween PD and MSA were observed also in terms of pathwaysand molecular functions (figure 3D table e-6) Most repre-sented KEGG categories included immune system signaltransduction endocrine system and signaling molecules andinteraction Except for the endocrine system they were moreenriched in PD and MSA suggesting potential stronger acti-vation of immune response in these groups Of note FoxOsignaling pathway was higher in AP-Tau

EV surface antigens correlate with cognitiveimpairment and disease gravity in PD andMSAIn PD there was a negative correlation between CD25 andMMSE and MoCa scores a negative correlation betweenCD146 and MMSE score whereas CD62P directly correlatedwith the BDI-II (figure 4 AndashD) No significant correlationswere found between EV antigenrsquos expression and LEDD inthe PD and AP groups (table e-7 linkslwwcomNXIA293)In the MSA group mean MFI for CD9 CD63 and CD81inversely correlated with the MoCA whereas nanoparticleconcentration directly correlated with the disease durationand CD31 inversely correlated with the HampY (figure 4 EndashG)No correlations were observed in the AP-Tau group(table e-7)

Differential EV surface antigen expression anddiagnostic outcomeUnivariate logistic regression analysis allowing the assess-ment of associations between each EV marker and the di-agnosis confirmed 11 EV surface antigens as potentialdiscriminants for PD diagnosis 6 for MSA and 3 for AP-Tau(figure e-3 linkslwwcomNXIA293) Among markers sig-nificantly associated with the different diagnoses CD1c

CD11c CD19 CD41b CD45 and CD146 were exclusive ofthe PD group CD29 was exclusive for the MSA group CD2displayed the strongest association with the diagnosis of PD(OR = 1191) and MSA (OR = 1256) whereas CD40 withAP-Tau (OR = 1131)

ROC curve analysis of EV surface antigensshows best performance for MSA and PDROC curve analysis for all pathologic groups (PD MSA andAP-Tau) vs HC confirmed a reliable diagnostic performanceof each single differentially expressed EV markers (figure 5table e-8 linkslwwcomNXIA293) The linear weightedcombination of the 3 markers with the highest AUC showedbetter diagnostic performance respect to single markers in allgroups The combination of all EV surface markers in 1 singlecompound marker showed a further diagnostic improvementin the PD and MSA groups

Random forest model discriminates thedifferent groupsAn RF diagnostic model was built using the 17 surface anti-gens differentially expressed in plasma-derived EVs Theforest was composed by 20 different classification trees (arepresentative tree is reported in figure 6A) The model dis-criminated patients of the 4 different groups (HC PD MSAand AP-Tau) with high accuracy (889) all subjects with PDwere correctly diagnosed and 1 MSA and 1 HC were re-spectively misdiagnosed as HC and PD whereas among 9patients with AP-Tau 2 were predicted as HC and 3 as PD(figure 6B) Subsequently pairwise comparisons were per-formed (figure 6 CndashH) The RF model was validated by theleave-one-out algorithm which confirmed the generalizabilityof the model and excluded overfitting bias (accuracy of in-ternal validation 638 with a 722ndash915 range for pair-wise comparisons) Finally we validated our model in anexternal cohort of 40 subjects the overall accuracy was 775resulting in the correct diagnosis of 31 of 40 subjects (figure 6I and J) The accuracy after external validation was consistentwith the one resulting from the internal validation supportingthe reliability of the diagnostic model Demographic data ofthe external cohort were similar to those of the training cohortand are shown in table e-9 (linkslwwcomNXIA293)

DiscussionThe major finding of this study consists in the setup of adiagnostic model for the stratification of patients with PD andAP based on immunologic profiling of plasmatic EV subpop-ulations obtained from minimally invasive peripheral bloodsampling We systematically evaluated the diagnostic perfor-mance of differentially expressed EV antigens and a diagnosticmodel was built using supervised machine learning algorithmsThe model showed an overall reliable accuracy correctly pre-dicting patient diagnosis with the best performance for thediagnosis of PD (978) and MSA (100) vs HC Theseresults were supported by ROC curve analysis on the com-poundmarker originated from the linear combination of all the

8 Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 6 | November 2020 NeurologyorgNN

Figure 3 Extracellular vesicle (EV) surface proteins upregulated in PD MSA and AP-Tau and functional evaluation of theirprotein interactors

PPI network showing the first neighbors of each differentially expressed EV surfacemarker in (A) PD (B) MSA and (C) AP-Tau vs HC (D) Kyoto Encyclopedia ofGenes and Genomes pathways enriched by considering the first neighbors of each EV surface protein in PD MSA and AP-Tau vs HC DAVID databasebackground Homo sapiens gene count gt5 and p lt 0001 AP-Tau = atypical parkinsonism with tauopathies HC = healthy control MSA = multiple systematrophy PD = Parkinson disease PPI = protein-protein interaction

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 6 | November 2020 9

differentially expressed EV markers showing very high sensi-tivity and specificity for PD and MSA (AUC 0908 and 0974respectively) Previous works have explored the utility of EVs asbiomarkers for PD by quantifying brain-derived exosomes(AUC 075ndash082)25 or by measuring specific target proteinslike alpha-synuclein (αSyn) orDJ-1 in plasma neuronal-derivedexosomes (AUC 0654 0724)26 The combination of multiplemarkers improved the diagnostic accuracy of neuronal-derivedexosomes as shown by a recent work on quantification of bothαSyn and clusterin differentiating PD from other proteino-pathies and fromMSAwith high accuracy (AUC 098 and 094respectively)27

This analysis of multiple immune surface markers of circu-lating EVs in PD and AP shows a high diagnostic perfor-mance likely due to the advantage of simultaneously profilingseveral EV subpopulations First of all we demonstrated thatplasma EV concentration was higher in patients with PDPrevious reports have shown that the total number and size ofEVs were not augmented in serum of PD28 whereas a morerecent study demonstrated an increased number of plasmaticbrain-derived EVs in PD25 Methodological factors such asisolationextraction and quantification of EVs explain thesedifferences However at the molecular level it is recognizedthat endosomelysosome pathway is a common defectivepathway in sporadic and genetic PD29 and EVs are generatedand secreted by the endosomal compartment called multi-vesicular bodies by fusion with plasma membrane The pro-cess of EV secretion may be enhanced when there is an

inhibition of fusion of multivesicular bodies with lysosomesas expected in PD30 so that an increased production of EVs inPD is likely

It is difficult to track the origin of EVs because the majority ofmarkers are shared by several cell types and virtually any cell canrelease EVs in blood In blood normally a large number of EVsarises from platelets erythrocytes however leucocytes endo-thelial cells monocytes neutrophils and lymphocytes may re-lease EVs31 The flow cytometry analysis demonstrated that 16and 12 EV markers related to immune cells were upregulatedrespectively in PD and MSA only 4 in AP-Tau compared withhealthy condition In particular PD and MSA shared 11 EVsurface markers Considering functions and roles of EV surfacemarkers analyzed in this study this result favors the hypothesisof a major or at least different immune dysregulation in PD andMSA vs AP-Tau Despite sharing several overlapping clinicalfeatures synucleinopathies and tauopathies are distinguished bydistinctive neuropathologic hallmarks deposits of aggregatedαSyn (Lewy bodies) in neurons and in glial cells in the formergroup and neurofibrillary tangles of Tau in the latter as shownby immunohistological studies32 Although inflammatory fea-tures have been described in patients with both synucleino-pathies and tauopathies by PET studies33ndash36 the pathwaysactivated are probably different Animal studies have shown thatthe neurotoxic effects of beta-amyloid aggregates in a model oftauopathy (Alzheimer disease) are mediated via Toll-like re-ceptor 4ndashdependent glial cell activation while αSyn aggregatesin a model of PD activated Toll-like receptor 2 independently

Figure 4 Correlations between clinical scales and extracellular vesicle (EV) surface marker expression

Correlations between EV surface markers normalized MFI nanoparticle concentration (NmL plasma) and clinical parameters in patients with Parkinsondisease (circles A-D) and multiple system atrophy (triangles E-G) The regression line is reported together with its 95 CI (dashed line) BDI-II = BeckDepression Inventory II HampY = Hoehn and Yahr scale MFI = median fluorescence intensity MMSE = Mini-Mental State Examination MoCA = MontrealCognitive Assessment RBD = REM sleep behavior disorder screening questionnaire

10 Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 6 | November 2020 NeurologyorgNN

Figure 5 Receiver operating characteristic (ROC) curve analysis of extracellular vesicle (EV)-surface markers

ROC curves identifying the best cutoff for each EV surfacemarker discriminating pathologic groups fromHC The referral line is reported in gray (A) PD vs HC(B) MSA vs HC (C) AP-Tau vs HC In each plot ROC curves for the combination of the 3 EV surface markers with the highest AUCs and for a compound EVmarker (linearweighted combination of all EV surfacemarkers differentially expressed for each comparison) are shown (black and red lines respectively) Thetables provide asymptotic significance AUCwith 95 CI sensitivity and specificity on the compound EVmarkers p Values lt 005 were considered significantAP-Tau = atypical parkinsonismwith tauopathies AUC = area under the curve HC = healthy control MSA =multiple system atrophy PD = Parkinson disease

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 6 | November 2020 11

from Toll-like receptor 43738 Moreover a recent multicenterstudy has shown higher levels of CSF inflammatory biomarkersin PDwith dementia andMSA compared with controls and not

in AP-Tau vs controls plus thosemarkers correlated withmotorand cognitive impairment39 Likewise our analysis showed amoderate correlation between CD25 CD146 and cognitive

Figure 6 Random forest (RF) modeling to predict diagnosis and its validation in an external cohort of subjects

RFmodeling to diagnose patients based on the combination of the 17 differentially expressed extracellular vesicle surfacemarkers (A) Representation of 1 ofthe 20 different classification trees created by the algorithm to predict the diagnosis PD vs MSA vs AP-Tau vs HC (BndashH) Confusion matrix reporting real andpredicted diagnosis accuracy sensitivity specificity and internal validation by the leave-one-out algorithm for each comparison (see Methods) (I) Externalvalidation of the RF model 40 patients were included in the analysis (20 HC blue 10 PD red 5 MSA orange 5 AP-Tau gray) AP-Tau = atypical parkinsonismwith tauopathies HC = healthy control MSA = multiple system atrophy PD = Parkinson disease

12 Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 6 | November 2020 NeurologyorgNN

impairment in PD suggesting a link between inflammation and amajor cognitive decline CD25 is a costimulatory moleculesupporting immune cell activation40 and CD146 acts as anessential regulator of pericytendashendothelial cell communicationin the blood-brain barrier and it has been identified as a potentialkey therapeutic target for cerebrovascular disorders41 In MSAthe concentration of EVsmeasured byNTA and flow cytometryanalysis correlated with disease duration and cognitive impair-ment These findings favor the hypothesis of a perpetuation oftoxic effects by circulating EVs due to chronic immune activa-tion even if a compensatoryneuroprotective role of EVsin response to the progressive neurodegeneration cannot beexcluded

Among EV markers differentially expressed in PD CD146and MCSP are of interest because they have been associatedwith melanoma and used for detection of circulating tumorcells42 Consistently a link between PD and melanomahas been supported by many epidemiologic studies show-ing that patients with PD have a higher incidence of thistumor even if the underlying pathogenic mechanisms areunknown43

The network analysis of potential interactors of EV surfacemarkers demonstrated that functional pathways and net-work hubs in PD and MSA were coincident and differentfrom those of AP-Tau Of interest among hubs shared by PDand MSA we found SP1 a transcription factor playing a keyrole in regulating neuroinflammation in MS44 The mostrepresented KEGG pathways were immune system signaltransduction signaling molecules and folding sorting anddegradation in alpha-synucleinopathies whereas FoxO sig-naling pathway and some pathways of the endocrine systemwere higher in AP-Tau matching with the relation that hasbeen found by many authors between endocrine signalingtauopathies and FoxO4546 However this exploratory net-work analysis should be interpreted with caution becauseAP-Tau had less differentially expressed EV markers con-sequently the smaller network was a limiting factor to re-cover potential pathways and functions in tauopathiesAnyhow among the identified hubs it has been encouragingto find some of them described in the literature cytoplasmicprotein NCK2 was recently described as a PD-associatedgene47 Tyrosine-protein kinase Lyn (LYN) a specific hub ofMSA was related to enhanced microglial migration byαSyn48 Of note signal transducer and activator of tran-scription 3 (STAT3) a specific hub of AP-Tau has beenfound to be a direct target of C3 and C3a receptor signalingthat functionally mediates Tau pathogenesis49 Howeverthese network analyses are hypothetical and further vali-dation studies are required to assess their possible roles incausing PD and AP

Limitations of this study are the relatively low number ofsubjects especially in AP groups and the inclusion of patientsonly with long duration of disease larger studies and inclusionof different cohorts of patients especially at early stages of

disease are strongly recommended Moreover a customizedpanel of EV surface proteins including CNS and microgliamarkers would probably increase the diagnostic model Fi-nally this is an antemortem study and it lacks the diagnosticconfirmation of postmortem brain histopathologic analysis

In conclusion we systematically characterized circulating EVsin plasma of patients with PD or AP Several EV surfaceantigens were differentially expressed and correlated withdisease gravity and cognitive impairment suggesting EVs aspotential biomarkers of disease also in clinical trials fordisease-modifying drugs We propose a diagnostic model builtthrough supervised machine learning algorithms based onEV-specific signature which was able to discriminate patientswith PD and MSA with high accuracy Finally we providedinternal and external validations of our model confirmingreliable diagnostic performance

This is a highly relevant result with a potential impact onclinical practice allowing with a noninvasive low-cost bloodtest to identify patients with PD and MSA Furthermore cir-culating EV surface protein analysis can shed light on the dif-ferential inflammationimmunity pathways involved inprotein aggregationndashrelated neurodegenerative disease to beconfirmed by functional analysis in experimental models ofdiseases

AcknowledgmentsThe authors are very grateful to all the patients and theirrelatives who participated in this study They are grateful toMrs Nicole Vago research nurse for her valuable work on theclinical database They thank the Scientific Research AdvisoryBoard of the Ente Ospedaliero Cantonale (ABREOC) for thefinancial support of this study

Study fundingThis study was funded by the Scientific Research AdvisoryBoard of the Ente Ospedaliero Cantonale (ABREOC)

DisclosureThe authors report no disclosures relevant to the manuscriptGo to NeurologyorgNN for full disclosures

Publication historyReceived by Neurology Neuroimmunology amp NeuroinflammationDecember 16 2019 Accepted in final form May 28 2020

Appendix Authors

Name Location Contribution

ElenaVacchi MS

Neurocenter ofSouthernSwitzerland Lugano

Acquisition analysis andinterpretation of data and draftedthe manuscript

JacopoBurrelloMD

Cardiocentro TicinoFoundation Lugano

Acquisition analysis andinterpretation of data and draftedthe manuscript

Continued

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 6 | November 2020 13

References1 Hughes AJ Daniel SE Kilford L Lees AJ Accuracy of clinical diagnosis of idiopathic

Parkinsonrsquos disease a clinico-pathological study of 100 cases J Neurol NeurosurgPsychiatry 199255181ndash184

2 Tolosa E Wenning G PoeweW The diagnosis of Parkinsonrsquos disease Lancet Neurol2006575ndash86

3 Thery C Witwer KW Aikawa E et al Minimal information for studies of extracellularvesicles 2018 (MISEV2018) a position statement of the International Society forExtracellular Vesicles and update of the MISEV2014 guidelines J Extracell Vesicles201871535750

4 Faure J Lachenal G Court M et al Exosomes are released by cultured corticalneurones Mol Cell Neurosci 200631642ndash648

5 Goetzl EJ Boxer A Schwartz JB et al Altered lysosomal proteins in neural-derivedplasma exosomes in preclinical Alzheimer disease Neurology 20158540ndash47

6 Thery C Zitvogel L Amigorena S Exosomes composition biogenesis and functionNat Rev Immunol 20022569ndash579

7 Sulzer D Alcalay RN Garretti F et al T cells from patients with Parkinsonrsquos diseaserecognize alpha-synuclein peptides Nature 2017546656ndash661

8 Tansey MG Romero-Ramos M Immune system responses in Parkinsonrsquos diseaseearly and dynamic Eur J Neurosci 201949364ndash383

9 Koliha N Heider U Ozimkowski T Wiemann M Bosio A Wild S Melanoma affectsthe composition of blood cell-derived extracellular vesicles Front Immunol 20167282

10 Koliha N Wiencek Y Heider U et al A novel multiplex bead-based platform high-lights the diversity of extracellular vesicles J Extracell Vesicles 2016529975

11 Gilman S Wenning GK Low PA et al Second consensus statement on the diagnosisof multiple system atrophy Neurology 200871670ndash676

12 Hoglinger GU Respondek G Stamelou M et al Clinical diagnosis of progressivesupranuclear palsy the Movement Disorder Society criteria Mov Disord 201732853ndash864

13 Armstrong MJ Litvan I Lang AE et al Criteria for the diagnosis of corticobasaldegeneration Neurology 201380496ndash503

14 Kipfer S Stephan MA Schupbach WM Ballinari P Kaelin-Lang A Resting tremor inParkinson disease a negative predictor of levodopa-induced dyskinesia Arch Neurol2011681037ndash1039

15 Thery C Amigorena S Raposo G Clayton A Isolation and characterization of exo-somes from cell culture supernatants and biological fluids Curr Protoc Cell Biol 2006Chapter 3Unit 322

16 Dragovic RA Gardiner C Brooks AS et al Sizing and phenotyping of cellular vesiclesusing Nanoparticle Tracking Analysis Nanomedicine 20117780ndash788

17 Scardoni G Tosadori G Pratap S Spoto F Laudanna C Finding the shortest pathwith PesCa a tool for network reconstruction F1000Res 20154484

18 Scardoni G Tosadori G FaizanM Spoto F Fabbri F Laudanna C Biological networkanalysis with CentiScaPe centralities and experimental dataset integration F1000Res20143139

19 Di Silvestre D Brambilla F Scardoni G et al Proteomics-based network analysischaracterizes biological processes and pathways activated by preconditioned mesen-chymal stem cells in cardiac repair mechanisms Biochim Biophys Acta Gen Subj201718611190ndash1199

20 Sereni L Castiello MC Di Silvestre D et al Lentiviral gene therapy corrects plateletphenotype and function in patients with Wiskott-Aldrich syndrome J Allergy ClinImmunol 2019144825ndash838

21 Huang DW Sherman BT Lempicki RA Systematic and integrative analysis of largegene lists using DAVID bioinformatics resources Nat Protoc 2009444ndash57

22 Burrello J Burrello A Stowasser M et al The primary aldosteronism surgical outcomescore for the prediction of clinical outcomes after adrenalectomy for unilateral primaryaldosteronism Ann Surg Epub 2019 Jan 18

23 Yang Y Burrello J Burrello A et al Classification of microadenomas in patients withprimary aldosteronism by steroid profiling J Steroid Biochem Mol Biol 2019189274ndash282

24 Vella D Zoppis I Mauri G Mauri P Di Silvestre D From protein-protein interactionsto protein co-expression networks a new perspective to evaluate large-scale proteo-mic data EURASIP J Bioinform Syst Biol 201720176

25 Ohmichi T Mitsuhashi M Tatebe H Kasai T El-Agnaf OMA Tokuda T Quanti-fication of brain-derived extracellular vesicles in plasma as a biomarker to diagnoseParkinsonrsquos and related diseases Parkinsonism Relat Disord 20196182ndash87

26 Zhao ZH Chen ZT Zhou RL Zhang X Ye QY Wang YZ Increased DJ-1 and alpha-synuclein in plasma neural-derived exosomes as potential markers for Parkinsonrsquosdisease Front Aging Neurosci 201810438

27 Jiang C Hopfner F Katsikoudi A et al Serum neuronal exosomes predict anddifferentiate Parkinsonrsquos disease from atypical parkinsonism J Neurol NeurosurgPsychiatry 202091720ndash729

28 Tomlinson PR Zheng Y Fischer R et al Identification of distinct circulating exo-somes in Parkinsonrsquos disease Ann Clin Transl Neurol 20152353ndash361

29 Tofaris GK Lysosome-dependent pathways as a unifying theme in Parkinsonrsquos dis-ease Mov Disord 2012271364ndash1369

30 Tofaris GK A critical assessment of exosomes in the pathogenesis and stratification ofParkinsonrsquos disease J Parkinson Dis 20177569ndash576

31 Shah R Patel T Freedman JE Circulating extracellular vesicles in human diseaseN Engl J Med 2018379958ndash966

32 Kovacs GG Molecular pathological classification of neurodegenerative diseasesturning towards precision medicine Int J Mol Sci 201617189

33 Gerhard A Banati RB Goerres GB et al [11C](R)-PK11195 PET imagingof microglial activation in multiple system atrophy Neurology 200361686ndash689

34 Gerhard A Pavese N Hotton G et al In vivo imaging of microglial activation with[11C](R)-PK11195 PET in idiopathic Parkinsonrsquos disease Neurobiol Dis 200621404ndash412

35 Gerhard A Trender-Gerhard I Turkheimer F Quinn NP Bhatia KP Brooks DJ Invivo imaging of microglial activation with [11C](R)-PK11195 PET in progressivesupranuclear palsy Mov Disord 20062189ndash93

36 Gerhard A Watts J Trender-Gerhard I et al In vivo imaging of microglial activationwith [11C](R)-PK11195 PET in corticobasal degeneration Mov Disord 2004191221ndash1226

37 Balducci C Frasca A Zotti M et al Toll-like receptor 4-dependent glial cell activationmediates the impairment in memory establishment induced by beta-amyloid oligo-mers in an acute mouse model of Alzheimerrsquos disease Brain Behav Immun 201760188ndash197

38 La Vitola P Balducci C Cerovic M et al Alpha-synuclein oligomers impair memorythrough glial cell activation and via Toll-like receptor 2 Brain Behav Immun 201869591ndash602

39 Hall S Janelidze S Surova Y Widner H Zetterberg H Hansson O Cerebrospinalfluid concentrations of inflammatory markers in Parkinsonrsquos disease and atypicalparkinsonian disorders Sci Rep 2018813276

40 Elgueta R Benson MJ de Vries VC Wasiuk A Guo Y Noelle RJ Molecular mech-anism and function of CD40CD40L engagement in the immune system ImmunolRev 2009229152ndash172

41 Chen J Luo Y Hui H et al CD146 coordinates brain endothelial cell-pericytecommunication for blood-brain barrier development Proc Natl Acad Sci USA 2017114E7622ndashE7631

42 Rapanotti MC Campione E Spallone G Orlandi A Bernardini S Bianchi L Minimalresidual disease in melanoma circulating melanoma cells and predictive role ofMCAMMUC18MelCAMCD146 Cell Death Discov 20173177005

43 Bose A Petsko GA Eliezer D Parkinsonrsquos disease and melanoma co-occurrence andmechanisms J Parkinsons Dis 20188385ndash398

Appendix (continued)

Name Location Contribution

Dario diSilvestreMS

ITB-CNR Milan Analysis and interpretation ofdata

AlessioBurrello MS

University ofBologna

Analysis and interpretation ofdata

Sara BolisMS

Cardiocentro TicinoFoundation Lugano

Acquisition and interpretation ofdata

PierluigiMauri PhD

ITB-CNR Milan Interpreted the data and revisedthe manuscript for intellectualcontent

GiuseppeVassalli MD

Cardiocentro TicinoFoundation Lugano

Interpreted the data and revisedthe manuscript for intellectualcontent

Carlo WCereda MD

Neurocenter ofSouthernSwitzerland Lugano

Interpreted the data and revisedthe manuscript for intellectualcontent

CinthiaFarina PhD

San RaffaeleScientific InstituteMilan

Interpreted the data and revisedthe manuscript for intellectualcontent

Lucio BarilePhD

Cardiocentro TicinoFoundation Lugano

Study design interpretedthe data and revised themanuscript for intellectualcontent

AlainKaelin-Lang MDPhD

Neurocenter ofSouthernSwitzerland Lugano

Study design patientenrollment interpreted thedata and revised themanuscript for intellectualcontent

GiorgiaMelli MDPhD

Neurocenter ofSouthernSwitzerland Lugano

Study design patientenrollment interpreted thedata and manuscript writingand revision

14 Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 6 | November 2020 NeurologyorgNN

44 Menon R Di Dario M Cordiglieri C et al Gender-based blood transcriptomes andinteractomes in multiple sclerosis involvement of SP1 dependent gene transcriptionJ Autoimmun 201238J144ndashJ155

45 Gratuze M Joly-Amado A Vieau D Buee L Blum D Mutual relationship betweenTau and central insulin signalling consequences for AD and tauopathies Neuro-endocrinology 2018107181ndash195

46 Neri C Role and therapeutic potential of the pro-longevity factor FOXO and itsregulators in neurodegenerative disease Front Pharmacol 2012315

47 Sun Y Ye L Zheng Y Yang Z Identification of crucial genes associated with Par-kinsonrsquos disease using microarray data Mol Med Rep 2018173775ndash3782

48 Wang S Chu CH Stewart T et al alpha-Synuclein a chemoattractant directsmicroglial migration via H2O2-dependent Lyn phosphorylation Proc Natl Acad SciUSA 2015112E1926ndashE1935

49 Litvinchuk A Wan YW Swartzlander DB et al Complement C3aR inactivationattenuates Tau pathology and reverses an immune network deregulated in tauopathymodels and Alzheimerrsquos disease Neuron 20181001337ndash1353e5

NeurologyorgNN Neurology Neuroimmunology amp Neuroinflammation | Volume 7 Number 6 | November 2020 15

DOI 101212NXI000000000000086620207 Neurol Neuroimmunol Neuroinflamm

Elena Vacchi Jacopo Burrello Dario Di Silvestre et al Immune profiling of plasma-derived extracellular vesicles identifies Parkinson disease

This information is current as of August 12 2020

ServicesUpdated Information amp

httpnnneurologyorgcontent76e866fullhtmlincluding high resolution figures can be found at

References httpnnneurologyorgcontent76e866fullhtmlref-list-1

This article cites 47 articles 4 of which you can access for free at

Subspecialty Collections

mhttpnnneurologyorgcgicollectionparkinsons_disease_parkinsonisParkinsons diseaseParkinsonismfollowing collection(s) This article along with others on similar topics appears in the

Permissions amp Licensing

httpnnneurologyorgmiscaboutxhtmlpermissionsits entirety can be found online atInformation about reproducing this article in parts (figurestables) or in

Reprints

httpnnneurologyorgmiscaddirxhtmlreprintsusInformation about ordering reprints can be found online

Academy of Neurology All rights reserved Online ISSN 2332-7812Copyright copy 2020 The Author(s) Published by Wolters Kluwer Health Inc on behalf of the AmericanPublished since April 2014 it is an open-access online-only continuous publication journal Copyright

is an official journal of the American Academy of NeurologyNeurol Neuroimmunol Neuroinflamm