Hindawi Publishing Corporation Stem Cells International Volume 2017, Article ID 2389753, 15 pages http://dx.doi.org/10.1155/2017/2389753 Research Article Allogeneic Adipose-Derived Mesenchymal Stromal Cells Ameliorate Experimental Autoimmune Encephalomyelitis by Regulating Self-Reactive T Cell Responses and Dendritic Cell Function PerAnderson,1,2ElenaGonzalez-Rey,1FranciscoO’Valle,3FranciscoMartin,2 F.JavierOliver,1andMarioDelgado1 1InstituteofParasitologyandBiomedicine“Lo´pez-Neyra”,CSIC,PTS,Granada,Spain 2GENYO,CentreofGenomicsandOncologicalResearch,Pfizer/UniversityofGranada/AndalusianRegionalGovernment, PTS,Granada,Spain 3SchoolofMedicine,DepartmentofPathology,IBIMER,CIBM,UniversityofGranada,Granada,Spain CorrespondenceshouldbeaddressedtoPerAnderson;[email protected];[email protected] Received17October2016;Accepted18December2016;Published30January2017 AcademicEditor:MarcellaFranquesa Copyright©2017PerAndersonetal.ThisisanopenaccessarticledistributedundertheCreativeCommonsAttributionLicense, whichpermitsunrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycited. Multipotentmesenchymalstromalcells(MSCs)haveemergedasapromisingtherapyforautoimmunediseases,includingmultiple sclerosis(MS).AdministrationofMSCstoMSpatientshasprovensafewithsignsofimmunomodulationbuttheirtherapeutic efficacyremainslow.Theaimofthecurrentstudyhasbeentofurthercharacterizetheimmunomodulatorymechanismsofadipose tissue-derivedMSCs(ASCs)invitroandinvivousingtheEAEmodelofchronicbraininflammationinmice.Wefoundthatmurine ASCs(mASCs)suppressTcellproliferationinvitroviainduciblenitricoxidesynthase(iNOS)andcyclooxygenase-(COX-)1/2 activities.mASCsalsopreventedthelipopolysaccharide-(LPS-)inducedmaturationofdendriticcells(DCs)invitro.Theaddition oftheCOX-1/2inhibitorindomethacin,butnottheiNOSinhibitorL-NAME,reversedtheblockinDCmaturationimplicating prostaglandin (PG) E2 in this process. In vivo, early administration of murine and human ASCs (hASCs) ameliorated myelin oligodendrocyteprotein-(MOG35−55-)inducedEAEinC57Bl/6mice.MechanisticstudiesshowedthatmASCssuppressedthe functionofautoantigen-specificTcellsandalsodecreasedthefrequencyofactivated(CD11c+CD40highandCD11c+TNF-𝛼+)DCs indraininglymphnodes(DLNs).Insummary,thesedatasuggestthatmASCsreduceEAEseverity,inpart,throughtheimpairment ofDCandTcellfunction. 1.Introduction clearly shown that in vitro-expanded MSCs possess potent immunomodulatoryproperties.Firstly,MSCscanmodulate Multipotent mesenchymal stromal cells (MSCs) are non- theactivationofcellsofboththeinnateandadaptiveimmune hematopoietic, perivascular cells which support hemato- system in vitro. Depending on species and immune cell poiesisandarethoughttoparticipateintissuerepairinvivo studied,thisimmunomodulatoryeffectisachievedthrough [1–3]. MSCs can be obtained from virtually all tissues and a combination of mechanisms including cytokines (IL-10, organsofthebodyandhavebeendefinedinvitroasplastic TGF-𝛽1, and IL-6), intracellular enzymes (inducible nitric adherent cells that express CD73, CD90, and CD105 while oxidesynthase(iNOS)inmurineMSCsandindoleamine2,3- lacking expression of CD45, CD34, and CD14, which can dioxygenase(IDO)inhumanMSCs),growthfactors(hepato- differentiate into osteoblasts, adipocytes, and chondroblasts cytegrowthfactor,vascularendothelialgrowthfactor),mem- invitro[4].Althoughtheproportionofbonafidestemcells braneboundmolecules(programmeddeathligand-1,FASL), withinMSCpreparationsislikelytobelow[5],ithasbeen and prostaglandin E2 (PGE2) [6–8]. Secondly, injection of 2 StemCellsInternational MSCpreparationshasshowndelayallograftrejection[9,10] ethical committee at the Institute for Parasitology and Bio- andamelioratingdiseaseinseveralanimalmodelsofinflam- medicine“Lo´pez-Neyra”andthecentralethicalcommitteeat mation/autoimmunitybyinhibitingthedeleteriousimmune theConsejoSuperiordeInvestigacionesCientificas(CSIC). responsesagainstself-antigens[11–14].Basedontheirsuccess Micewereanesthetizedbeforeimmunizationusingintraperi- in experimental models, the clinical potential of human tonealinjectionsofketamine-HCl(doseat100mg/kg)mixed autologousandallogeneicMSCsforthetreatmentofautoim- with xylazine-HCl (10mg/kg). Mice were sacrificed using mune/inflammatory diseases has been assessed in various CO2whenextractingthespinalcordforhistologicalanalyses. phaseI/IIandIIIclinicaltrials[15–19]. Inallothercases,miceweresacrificedusingcervicaldisloca- tion. Multiplesclerosis(MS)isachronicdiseaseofthecentral nervoussystem(CNS)whereautoreactiveTcellsandmacro- phagesattackanddisruptthecommunicationbetweenneu- 2.2. Isolation and Expansion of mASCs. Mesenchymal stro- rons which result in a multitude of neurological symptoms mal cells were isolated from adipose tissue as previously [20].SeveralclinicaltrialshaveutilizedMSCsforthetreat- described[28].CellswereresuspendedinMesenCult(Stem ment of MS and a few of these studies have analyzed the Cell,Grenoble,France)containing20%mousemesenchymal effects of injected MSCs on the immune status of patients supplements(StemCell)andpenicillin/streptomycin(Invit- withMS.Inaphase1/2open-safetyclinicaltrialforMS,injec- rogen,Carlsbad,CA),platedatadensityof2-3×104cells/cm2 ∘ tionsofautologousMSCsresultedinarapidinhibitionofthe andculturedat37 Cinhypoxiaat5%O2,5%CO2.Nonadher- immunesystem[17]whileanotherstudyobservedincreased entcellswereremovedafter24hoursinculture.Subsequent 4 2 foxp3transcriptlevelsinperipheralbloodmononuclearcells passages were plated at 10 cells/cm and maintained at in MSC-treated patients [21]. These data suggest that the 5% O2, 5% CO2. mASCs were used at passages 3–9. The phenotypiccharacterizationanddifferentiationcapacityinto therapeuticeffectofMSCsinMSdepends,atleastpartially, adipocytes,osteocytes,andchondrocyteswereperformedas upontheirimmunomodulatorycapacity.However,whilethe previously described ([28]; data not shown). Human ASCs administrationofMSCshasbeenprovensafe,theircapacity were obtained from Cellerix SA, Tres Cantos (Madrid), toameliorateMSremainspoor[16,17,19,22].Therefore,it Spain,andculturedinadvancedDMEMsupplementedwith is important to further study how MSCs can modulate the 10%FCS(Invitrogen),Glutamax(GIBCO,LifeTechnologies, immunesystembothinvitroandinvivoinordertoimprove CA), and 100U/mL penicillin/streptomycin (GIBCO, Life MSC-basedtherapiesforMS[23]. Technologies)aspreviouslydescribed[14]. Thus,theobjectiveofthecurrentstudyhasbeentochar- acterize the immunomodulatory properties of murine adi- 2.3. Assessment of mASC-Mediated Inhibition of T Cell Pro- posetissue-derivedMSCs(mASC)invitroandinvivousing liferation In Vitro. mASCs were treated with mitomycin C the experimental autoimmune encephalomyelitis (EAE) (50𝜇g/mL,SigmaAldrich,St.Louis,MO)for20minutesat modelofCNSautoimmunity.EAEisawell-acceptedanimal ∘ 37 Candwashed3timeswithcompleteRPMI1640(2mML- model of MS sharing both its immunological and patho- glutamine,100U/mLpenicillin/streptomycin,50𝜇M2-mer- logical features [24]. Injection of MSCs at distinct stages of captoethanol,and10%heat-inactivatedFCS,allfromInvitro- diseasehasbeenshowntoinhibitdiseaseseverityinvarious gen).MitomycinC-treatedmASCswereplatedinflatbotto- modelsofEAE,includingrelapsing/remittingEAE,chronic med96-wellplatesandallowedtoadherefor3-4hours.Mito- EAE,anddiseaseinducedbyadoptivetransferofencephal- mycinCtreatmentofmASCsdidnotimpactontheirability itogenic T cells [11, 25–27]. We show that mASC inhibited toinhibitTcellproliferationinvitro(seeFigureS1inSupple- TcellproliferationinvitroviaiNOSandCOX-1/2activities. mentary Material available online at https://doi.org/10.1155/ Bothallogeneicandxenogeneic(human)ASCsameliorated 2017/2389753).SpleensfromfemaleBalb/cmicewerehomo- MOG35−55-induced chronic EAE in C57Bl/6 mice. Murine genized and erythrocytes lysed using ACK buffer (0.15M ASCs reduced EAE severity through the inhibition of the autoimmuneTcellresponsewithnoincreaseinfoxp3Tregs. N2H×41C0l5,1c0elmlsMwKerHeCadOd3e,dantdo0w.1emllsMwNitah2EoDrTwAithatopuHt m7.4A)SaCnsd. Importantly, mASC inhibited the maturation of dendritic Concanavalin A (ConA, 2.5𝜇g/mL, Sigma Aldrich) was cells (DCs) in vitro via COX-1/2 activity and reduced the addedtoculturesasamitogenicstimulusofTcells.L-NAME percentageofactivatedDCsinthedrainingLNsofEAEmice. (1mM), indomethacin (20𝜇M), and L-norvaline (10mM) Our data suggests that MSC, through their modulation of (allfromSigmaAldrich)wereaddedattheinitiationofthe TcellandDCfunction,canameliorateinflammatory/auto- cocultures. After 3 days, cells were pulsed with 0.5𝜇Ci/well immuneCNSdisease. 3 [ H]-thymidine (Perkin Elmer, Waltham, MA) for 6 hours and harvested onto glass fiber filters using a FilterMate 96 2.MaterialsandMethods 3 well-harvester(PerkinElmer).Uptakeof[ H]-thymidinewas 2.1.Animals. MaleBalb/c(6–10weeks)andfemaleC57Bl/6 measured on a 1450 Microbeta Trilux scintillation counter (6–8 weeks) (Charles River, Barcelona, Spain) were used to (WallacOy,Turku,Finland). initiate cultures of mASCs and for the induction of EAE, respectively.Experimentalprotocolswereperformedaccord- 2.4. Measurement of Nitrite Production. To assess iNOS ing to the National/EU Guidelines for the Care and Use of activity in mASCs, supernatants from control/stimulated LaboratoryAnimalsinResearchwiththeapprovalofthelocal mASCs and mASC: DC cocultures were assayed for nitrite StemCellsInternational 3 contents using the Griess assay. In brief, 100𝜇L of Griess processed forparaffin inclusionandsectioning.Transversal reagent(a1/1mixtureof1%p-aminobenzene-sulfonamidein sections(4𝜇mthickness)werestainedwithLuxolfastblue, 5%H3PO4and0.1%naphthylethylenediaminedihydrochlo- cresylviolet,andhematoxylinfollowingtheKlu¨ver-Barrera ride in distilled H2O; Sigma Aldrich) was added to 100𝜇L technique[31]andwereanalyzedforthepresenceofareasof culturesupernatantsandstandard(NaNO2)in96-wellplates. demyelination and cell infiltration using a light microscope Plates were incubated at room temperature for 10min, and (Olympus,Tokyo,Japan).Forimmunohistochemistry,spinal absorbancewasmeasuredat550nm. cord sections were obtained as described for paraffin pro- cessingfollowedbyblockingstepswithperoxidaseblocking ∘ reagents, heat-treated in 1mM EDTA buffer pH 8 at 95 C 2.5.MeasurementofArginaseActivity. Arginaseactivitywas during 20min for antigenic unmasking, and incubated for measured as previously described [29]. Urea was used for 30min at room temperature with polyclonal anti-myelin thestandardcurve.Theproteinconcentrationsofcelllysates basicproteinAb(MasterDiagnostica,Granada,Spain).The were measured using the bicinchoninic acid (BCA) assay immunohistochemical study was done on an Autostainer [30]. One unit of enzyme activity is defined as the amount 480 (Thermo Fisher Scientific Inc., Waltham, MA) using ofenzymethatcatalyzestheformationof1mmolofureaper the polymer-peroxidase-based method and developed with min. diaminobenzidine. 2.6. Induction of EAE. Female C57Bl/6 mice (6–8 weeks- 2.8.AssessmentofAutoreactiveTCellResponsesinEAEMice. old) were immunized s.c. in the flanks with 150𝜇L of an Spleenanddraininglymphnode(DLN)cellsfromEAEmice emulsion (75𝜇L/flank) containing 150𝜇g MOG35−55 (Gen- treatedi.p.withPBS(control,𝑛=4)orallogeneicmASC(𝑛= script, Hong Kong, China) in PBS mixed with an equal 4) at the onset of disease (clinical scores: 0-1) were isolated volumeofcompleteFreund’sadjuvant(CFA)supplemented 7daysaftermASCinjectionandstimulatedwithMOG35−55 with 4mg/mL Mycobacterium tuberculosis H37Ra (Difco, (50𝜇g/mL)oranti-CD3(1𝜇g/mL;145-2C11;BDPharmingen, Detroit, MI). Mice were injected i.p. with 200ng pertussis San Diego, CA) at 1.5 × 106 cells/mL in flat bottomed 96- toxin(SigmaAldrich)inPBSonthedayofimmunizationand 6 wellplatesforproliferationandat10 cells/mLin24-wellfor 2dayslater.Immunizedmicewererandomlydistributedin differentgroups.Group1:controlmice(𝑛=8)wereinjected cytokinedetermination.Supernatantswerecollectedafter48 hoursandassayedforcytokinelevelsbyELISA.After3days, i.p. with PBS at the onset of disease (clinical scores: 0-1). 3 Group2:controlmice(𝑛 = 13)wereinjectedi.p.withPBS cellproliferationwasdeterminedby[ H]-thymidineuptake at the acute phase of disease (clinical scores: 1–3). Group 3: asdescribedabove.Forintracellularcytokinestaining,spleen Mice (𝑛 = 9) were treated i.p. with allogeneic mASCs (106 andDLNcellsuspensions(1×106cells/mL)werestimulated cellsobtainedfromBalb/cmiceandexpandedinhypoxia)at withPMA(50ng/mL)andionomycin(0.5𝜇g/mL;bothfrom theonset(clinicalscores:0-1).Group4:mice(𝑛 = 7)were SigmaAldrich)for6hoursinthepresenceof3𝜇Mmonensin treated i.p. with allogeneic mASCs (106 cells obtained from (eBioscience, San Diego, CA) during the last three hours. Balb/cmiceandexpandedinhypoxia)attheacutephaseof CellswereprocessedforFACSanalysisasdescribedbelow. thedisease(clinicalscores:2-3).Group5:mice(𝑛 = 7)were 6 treatedi.p.withhASCs(10 cells)attheacutephaseofdisease 2.9. Isolation and Characterization of DCs from EAE Mice. (clinicalscores:1-2).ClinicalsymptomsofEAEwerescored To determine the effect of mASCs on DC phenotype and daily using a 0–8 scale as follows: 0, no detectable signs of function in vivo, EAE mice were treated i.p. with PBS EAE; 1, affected tail tonus; 2, tail paralysis; 3, mild hind leg (control,𝑛 = 4)orwithallogeneicmASC(𝑛 = 4)afterthe paresis;4,severehindlegparesis;5,onehindlegparalysis;6, onsetofdisease(clinicalscores:1-2),and7dayslater,DLNs completehindlegparalysis;7,completehindlegparalysisand wereisolatedanddigestedwith1.6mg/mLcollagenasetype forelegparesis;and8,death.Fortheacquisitionofcellsand IVand0.1%DNAseI(SigmaAldrich)inRPMI1640medium ∘ tissues, anotherset of mice were used and sacrificed 7 days withoutsupplementsat37 Cfor30minutes.Forintracellular aftertreatmentwithPBSormASCsasdescribedbelow.Mice TNF-𝛼staining,DLNcellswerewashedtwicewithcomplete werescoreddailyfordiseasesymptoms.Watergelproducts RPMI1640 and 2 × 106 cells/mouse were plated in 12-well providingwaterandmoistenedfoodpelletswereplacedon platesinthepresenceof3𝜇Mmonensinfor4h.Cellswere the cage floor in Petri dishes which were changed daily gently harvested using cell scrapers, washed using FACS to prevent dehydration. Mice were euthanized if exhibiting buffer, and processed for intracellular staining as described severehindlegparalysisandforelegparesis(aclinicalscore below. For TNF-𝛼 ELISA, CD11c+ DCs were immunomag- of7). neticallypurifiedusingCD11c-microbeads(MiltenyiBiotech, Bergisch Gladbach, Germany) from collagenase type IV- 2.7.HistologicalAnalysisofCellInfiltrationandDemyeliniza- digestedDLNsandplatedat2.5×105cells/mLinthepresence tion. SpinalcordsfromEAEmicetreatedi.p.withPBS(𝑛 = ofLPS(1𝜇g/mL).Supernatantswerecollectedafter48hours 4)orallogeneicmASC(𝑛=4)attheonsetofdisease(clinical forcytokinedeterminations. scores: 0-1) were removed 7 days after treatment and pro- cessedforimmunohistochemistryandKlu¨ver-Barrerastain- 2.10.GenerationofBoneMarrow-DerivedDCs(BM-DCs)and ing. For light microscopy, cervical and lumbar spinal cord mASC Cocultures. BM-DCs were generated as previously segmentswerefixedwithbuffered10%formalinfor48hand described[32].Briefly,0.4×106 BMcells/mLfromC57Bl/6 4 StemCellsInternational micewereculturedincompleteRPMI1640containing20ng/ (BD Pharmingen), IL-10, IL-17, and CXCL10 (eBioscience). mLGM-CSF(Peprotech,London,UK).After6–8days,non- Platesweredevelopedusingperoxidase-labelledstreptavidin + adherent cells were harvested and CD11c immature DCs (Sigma) and 2,2 -azino-bis(3-ethylbenzthiazoline-6-sulpho- wereimmunomagneticallypurifiedusingCD11c-microbeads nicacid)(ABTS)andreadat405nm.PGE2 levelsinsuper- (Miltenyi Biotech) according to the manufacturer’s instruc- natantsweremeasuredbyELISA(CaymanChemicals,Ann tions. Cell preparations consisted of >95% CD11c+ DCs. Arbor,MI)accordingtothemanufacturer’sinstruction. To assess the effect of mASCs on the maturation of DCs, different numbers of mASCs were plated in 12-well plates 2.14.GeneExpressionAnalysis. TotalRNAwaspurifiedusing + and allowed to adhere for at least 6 hours. CD11c DCs Ultraspec (Biotecx, Huston, TX) from spinal cords isolated (0.4×106 cells/well)wereseededintowellswithorwithout from PBS-treated (control, 𝑛 = 4) or allogeneic mASC- mASCsandLPS(1𝜇g/mL)wasaddedfor48hourstoinduce treated (𝑛 = 4) EAE mice (7 days after treatment) or from DC maturation/activation. In some experiments L-NAME mASCs stimulated with LPS (1𝜇g/mL, Sigma) or TNF-𝛼 (1mM) or indomethacin (20𝜇M) was added to the DC: (10ng/mL,Peprotech)andIFN-𝛾(10ng/mL,BDBiosciences) mASC cocultures during maturation. DCs were harvested for6,12,and24hours.TotalRNA(1𝜇g/sample)wasreverse from the cocultures by gently collecting the nonadherent transcribed using M-MuLV RT (Roche Diagnostic, Basel, DCsfromtheadherentmASCmonolayer.Theacquiredcells Switzerland)andrandomhexamerprimers.Semiquantitative consistedof>95%CD11c+ DC.Supernatantswerecollected PCR and qPCR were performed using Taq polymerase and assayed for cytokine levels by ELISA or NO2 levels (Biotools, Madrid, Spain) or Supermix (Bio-Rad, Hercules, by Griess assay. Cell suspensions were processed for flow CA), respectively. Primer pairs include the following: iNOS cytometryanalysisofcellsurfaceantigensorusedforMLR FW: 5-GTTCTCAGCCCAACAATACAAGA-3; iNOS RV: asdescribedbelow. 5-GTGGACGGGTCGATGTCAC-3; MCP-1 FW: 5-TTA- AAAACCTGGATCGGAACCAA-3 ; MCP-1 RV: 5 -GCA- 2.11. Mixed Leukocyte Reaction. Spleens from BALB/c (H- TTAGCTTCAGATTTACGGGT-3; TGF-𝛽1 FW: 5-TGC- 2d) were homogenized and erythrocytes lysed using the GCTTGCAGAGATTAAAA-3; TGF-𝛽1 RV: 5-AGCCCT- Ammonium-Chloride-Potassium (ACK) lysis buffer. Cells GTATTCCGTCTCCT-3 ;IL-10FW:5 -GGTTGCCAAGCC- (4 × 106 cells/mL) were plated in Petri dishes and allowed TTATCGGA-3;IL-10RV:ACCTGCTCCACTGCCTTGCT; ∘ to adhere at 37 C for 90 minutes. Nonadherent cells were IDO FW: 5 -GGCTAGAAATCTGCCTGTGC-3 ; IDO RV: collectedandusedasresponders.C57Bl/6(H-2b)DCsfrom 5-AGAGCTCGCAGTAGGGAACA-3; COX-2 FW: 5- + theDC:mASCcoculturesorpurifiedCD11c DCsfromthe GGGTTGCTGGGGGAAGAAATGTG-3 , COX-2 RV: 5 - DLNsofuntreatedormASC-treatedEAEmicewereaddedin GGTGGCTGTTTTGGTAGGCTGTG-3 ;ArginaseIFW:5 - flatbottomed96-wellplates(15000DCs/well)togetherwith CAGAAGAATGGAAGAGTCAG-3; Arginase I RV: 5- 3 × 105 responder cells. After 4 days, cell proliferation was CAGATATGCAGGGAGTCACC-3; Foxp3 FW: 5-TTC- 3 determinedby H-thymidineuptakeasdescribedabove. ATGCATCAGCTCTCCAC-3; Foxp3 RV: 5 -CTGGAC- ACCCATTCCAGACT-3; IFN-𝛾 FW: 5-ACACTGCAT- 2.12.FlowCytometry. Collagenase/DNAse-treatedDLNcell CTTGGTTTGC-3;IFN-𝛾RV:5-TTGCTGATGGCCTGA- suspensionsormASC-culturedDCswerepreincubatedwith TTGTC-3; 𝛽-actin FW: 5-AATCGTGCGTGACATCAA- anti-Fc𝛾RII/Fc𝛾RIIImAbat2.5𝜇g/mL(2.4G2;BDPharmin- AG-3;𝛽-actinRV:5-ATGCCACAGGATTCCATACC-3. gen) and 7-aminoactinomycin D (2𝜇g/mL; Sigma Aldrich) andthenstainedwiththefollowingantibodies:CD11c-APC 2.15. Statistical Analysis. All results are expressed as mean (N481; eBioscience), CD40-PE (2/23), CD80-PE (16-10A1), (SEM)ofatleast3independentexperimentsunlessotherwise and/orCD86-PE(GL-1)(BDPharmingen).Forintracellular statedinthefigurelegends.TheMann–Whitney𝑈-testwas cytokinestainingcellswerefixed,permeabilized,andstained appliedonallinvivoresultsandcell-cultureexperimentsto for IFN-𝛾-FITC (XMG1.2), IL-17-PE (TC11-18H10), TNF-𝛼- comparenonparametricdataforstatisticalsignificance.A𝑝 PE(MP6-XT22),IL-4-PE(BVD4-1D11),andIL-10-PE(JES5- value<0.05wasconsideredsignificant. 16E3) (all from BD Biosciences, San Diego, CA) using the BD cytofix/cytoperm kit (BD Biosciences) according to 3.Results the manufacturer’s instructions. Each sample was stained with appropriate isotype controls. For the analysis of foxp3 3.1. Immunomodulatory Mechanisms of mASCs In Vitro. expression cells were stained with CD4-FITC (GK1.5; BD Acquiring high numbers of low passage MSCs with potent Biosciences) and subsequently processed for foxp3 staining immunosuppressive capacity is crucial for their successful using the foxp3 staining kit (eBioscience) according to the useasatherapyforinflammatory/autoimmunediseases[33]. manufacturer’s instructions. Stained cells were analyzed on In agreement with previous studies [34, 35], we found that aFACSCaliburflowcytometer(BDBiosciences). mASCexpandedatlowoxygentension(5%O2)proliferated atahigherratecomparedtomASCsexpandedatnormoxia 2.13. Cytokine and PGE2 ELISAs. The cytokine content in (Figure 1(a)). Thus, we decided to culture the ASCs at 5% supernatants from the mASC: splenocyte and DC cultures O2 and use these cells for the subsequent characterization weredeterminedbysandwichELISAsusingcapture/biotiny- of their immunomodulatory functions in vitro and in vivo. lated detection antibody pairs for IFN-𝛾, TNF-𝛼, IL-12p40 Whereas mASCs constitutively expressed TGF-𝛽1, COX-2, StemCellsInternational 5 mber 125 ∗ − LPS TIFNN+F--𝛾𝛼O nu 100 ∗ Time (h): 0 612246122461224H2 cell 40) 75 ∗ TGF-𝛽1 ative ×1( 50 IL-10 ul 25 IDO m Cu 0 COX-2 0 5 10 15 20 iNOS Time (days) Arginase I 21%O2 𝛽-Actin 5%O2 (a) (b) 5 5 120 PGE(ng/mL)2 01234 2 S 01234 𝜇(M)NO2 Arginase activity (U/mg protein)Co123456n0000000A − − + ∗*+ Proliferation(% of control) 124680000000 0 10 20 40 OX-1/ iNO SplenmocAytSeCs −+ −+ −+ ++ Number of ASCs (×103) C mASC mASC + L-norvaline (c) (d) (e) 4 ∗ L) 4 L) ∗ 100 m m 3 ol) 75 ng/ 3 ng/ of contr 2550 𝛾FN-] ( 12 ∗ ]PGE(2 12 ^ n (% 0 ∗ L)[I 0 ∗ [ 102 ∗ Proliferatio 1570050 ∗ ∗ L10] (ng/m 12 ]𝜇(M)NO2 48 25 XC [ ^ 0 0 10 20 40 C[Con 0A − + + + + Con 0A − + + + + Number of ASCs (×103) mASC − − + + + mASC − − + + + L-NAME − − − + − L-NAME − − − + − mASC mASC + INDO INDO − − − − + INDO − − − − + mASC + L-NAME (f) (g) (h) Figure 1: mASCs expanded in hypoxia maintain their immunosuppressive activities. (a) Proliferative response of mASCs expanded in 2 normoxia(21%O2)orinhypoxia(5%O2).mASCswereseededat5000cells/cm inT25flasksandcountedatdifferenttimepointswhen reaching70–80%ofconfluence.Dataareshownasmean(SD)ofonerepresentativeexperimentoutofthree.∗𝑝 < 0.05versus21%O2. ((b)and(c))ImmunephenotypiccharacterizationofmASCs.mASCswereexpandedunderhypoxicconditionsandthenculturedwith medium(unstimulated)orstimulatedwithLPS(1𝜇g/mL)orTNF-𝛼(10ng/mL)andIFN-𝛾(10ng/mL)fordifferenttimepoints(24hoursfor NO2determination).GeneexpressionofdifferentimmunemarkerswasanalyzedusingsemiquantitativePCR.iNOSactivitywasmeasured inTNF-𝛼/IFN-𝛾-stimulatedmASCsbythedetectionofNO2inculturesupernatantsusingGriessassay.COX-1/2activitywasdetermined bymeasuringthePGE2 levelsinculturesupernatantsfromunstimulatedmASCs.(d)ArginaseactivityisinducedinmASC:splenocyte cocultures.SplenocyteswereculturedwithorwithoutmASCsandstimulatedwithConAfor3days.Definedquantitiesofcelllysateswerethen subjectedtoaninvitroarginaseactivityassay.Dataareshownasmean(SEM)of3independentexperiments.∗𝑝<0.05versussplenocytes 6 +ConA.((e),(f),and(g))mASCssuppressTcellproliferationinvitroviaiNOSandCOX-1/2butnotarginaseactivity.Splenocytes(10 cells/mL) were stimulated with ConA (2.5𝜇g/mL) in the presence of different numbers of mitomycin C-treated mASCs (ratio 20:1 for IFN-𝛾andCXCL10determination).((e)and(f))TheiNOSinhibitorL-NAME(1mM),theCOX-1/2inhibitorindomethacin(20𝜇M),and 3 thearginaseinhibitorL-norvaline(10mM)wereaddedtocultureswhenindicated.After3days,cellproliferationwasmeasuredby[ H]- thymidineincorporationand(g)IFN-𝛾andCXCL10contentsinthesupernatantsweredeterminedbyELISA.Dataareshownasmean (SEM)of3independentexperiments.∗𝑝 < 0.05versusmASCin(f)and∗𝑝 < 0.05versusConAin(g).(h)iNOSandCOX-2activities 6 areincreasedinmASC:splenocytecocultures.ConA-stimulatedsplenocytes(10 cells/mL)wereculturedwithmitomycinC-treatedmASCs (ratio20:1)inthepresenceofL-NAMEorindomethacinfor3daysandthePGE2andNO2contentsmeasuredintheculturesupernatants. Resultsareshownasmean(SEM)of3separateexperiments.∗𝑝<0.05versusConA;∧𝑝<0.05versusmASC+ConA. 6 StemCellsInternational andarginaseImRNAsandPGE2,theexpressionofbothIDO foxp3 mRNAs in mASC-treated mice compared to control and iNOS mRNA and NO2 production was induced upon EAEmice(Figure3(c)). stimulation with TNF-𝛼 and IFN-𝛾, but not LPS (Figures 1(b)and1(c),FigureS2).Interestingly,wefoundasignificant 3.4. Allogeneic mASCs Inhibit MOG-Specific Immune Res- induction of arginase activity in the splenocyte: mASCs ponses but Do Not Induce foxp3+ Tregs. In both EAE and coculturesincomparisontomASCsorsplenocytescultured MS, autoreactive Th1 and Th17 cells producing IFN-𝛾 and alone (Figure 1(d)). Although arginase activity has been IL-17, respectively, infiltrate the CNS and promote disease, shown to inhibit the proliferation of T cells [36], we found whereas treatments that induce a skewing towards an IL- that the addition of the arginase inhibitor L-norvaline did 4 dominated Th2 response generally suppress EAE [39]. not significantly revert the suppressive activity of mASCs mASC treatment significantly reduced the number of IFN- in concanavalin A- (ConA-) induced splenocyte prolifera- 𝛾- and IL-17-secreting cells in both DLNs and spleen in tion (Figure 1(e)). In contrast, inhibition of iNOS activity EAEmice(Figure4(a)andrepresentativegatingstrategyin with L-NAME or of COX-1/2 activities with indomethacin Figure S3). Moreover, there was a significant reduction in significantly alleviated the mASC-mediated suppression of theMOG35−55-specificrecallresponse,withrespecttoboth T cell proliferation (Figure 1(f)). However, none of these proliferation and production of IFN-𝛾 and IL-17 compared inhibitorsaffectedthemASC-mediatedsuppressionofIFN-𝛾 to control EAE mice. The effect was antigen specific since ortheinductionofCXCL10inthesecocultures(Figure1(g)). activationoftotalTcellswithanti-CD3Abswhichresultedin Since inhibition of iNOS may decrease COX-2 expression similarproliferationandcytokineproductionbybothgroups and PGE2 secretion [37, 38], we measured the PGE2 levels (Figures 4(b) and 4(c)). However, we could not detect an in mASC: splenocyte cocultures treated with L-NAME. We increaseintheanti-inflammatorycytokinesIL-10andIL-4in found that L-NAME did not affect the productionof PGE2 mASC-treatedmicecomparedtocontrolmice(Figure4(a)), by mASCs and indomethacin did not modulate their iNOS suggestingthatthemASCsdidnotinduceaskewingtowards activity.Insummary,thesedatashowthatlowoxygentension a Th2 response or IL-10 producing Tregs. In addition, we + + increases the expansion of mASCs in vitro and that these could not detect any change in the size of the CD4 foxp3 cellsareeffectiveinsuppressingimmuneresponsesinvitro, TcellpopulationinDLNsorspleensofmASC-treatedEAE using both constitutive (COX-1/2) and inducible (iNOS) micecomparedtocontrolmice(Figure4(d)). mechanisms. 3.5. mASCs Inhibit the Maturation/Activation of Dendritic 3.2. Therapeutic Effect of Allogeneic and Xenogeneic ASC Cells In Vitro and In Vivo. Considering the importance of on Chronic EAE. We next wanted to assess the capacity of DCs in the initiation of the autoimmune response seen in allogeneicandxenogeneicASCstoinhibitimmuneresponses EAEandMS[40],wedecidedtoevaluatetheeffectofmASC in vivo. To this end, we induced chronic EAE in C57Bl/6 on the DC activation in vitro and in vivo. To this end, we mice and administered mASCs to the animals at different cocultured BM-derived immature DCs with different num- time points after immunization. A single injection of allo- bers of mASCs in the presence of LPS. We found that geneic mASC at the onset or at the acute phase of disease mASCsignificantlyinhibitedtheLPS-inducedexpressionof significantlyreducedthediseaseseveritycomparedtocontrol CD40andTNF-𝛼whilenotaffectingtheinductionofCD80 animals receiving PBS (Figure 2(a)). Both early and late and CD86 or the production of IL-10 or IL-12 by BM-DCs injections of mASCs significantly reduced the cumulative (Figure5(a)andrepresentativegatingstrategyinFigureS4). diseaseindex(Figure2(b)).XenogeneichASCsinjectedinto ThispartialblockofDCmaturationresultedinasignificant EAEmiceduringtheacutephaseofthedisease(animalswith decreaseintheirTcellactivatingcapacity(Figure5(b)).Since a mean score of 1.5) also reduced the mean clinical score both iNOS and PGE2 inhibited the proliferation of spleno- and significantly decreased the cumulative disease index cytes in vitro and can modulate DC activation [41–44], comparedtocontrolmice(Figures2(c)and2(d)). we set out to investigate their involvement in the mASC- mediatedinhibitionofDCmaturation.Wefoundthataddi- 3.3. Allogeneic mASCs Reduce Demyelinization and Inflam- tionofindomethacinbutnotL-NAMEsignificantlyreversed matoryInfiltrationintheCentralNervousSystemduringEAE themASC-mediatedsuppressionofTNF-𝛼productionand Progression. WethenwantedtoanalyzeCNStissuesections CD40 expression resulting in a significant increase in their fromcontrolandmASC-treatedEAEmiceforthepresence immunostimulatorycapacity(Figure5(c)).Theseeffectswere of inflammatory infiltrates and demyelinating plaques. To observeddespitethepresenceofiNOSactivityinthemASC: this end, mASCs were injected intraperitoneally into EAE DCcocultures(FigureS2). miceaftertheonsetoftheclinicalsymptoms(averagescore To study DC activation in vivo, we purified DCs from 1.9 (0.4)) and spinal cords were obtained 7 days later from DLNs of mASC-treated and control EAE mice 7 days after mASC-treatedmice(averagescore2.2(1.9))andcontrolmice mASCinjectionandanalyzedtheirexpressionofcostimula- (averagescore3.5(1.1))andprocessedforstaining.Wefound torymarkersandTNF-𝛼production.Consistentwiththein thatmASC-treatedmiceexhibitedlittlecellinfiltrationinto vitrodata,DCsfrommASC-treatedEAEmiceexpressedsig- the CNS parenchyma and subsequently few demyelinating nificantlylowerlevelsofCD40,butnotCD86,comparedto plaques compared to control EAE mice (Figures 3(a) and DCsfromcontrolEAEmice(Figure5(d)).Inaddition,intra- 3(b)).Consistentwiththereducedleukocyteinfiltration,we cellularstainingofTNF-𝛼onwholeLNsuspensionsrevealed detected lower levels of IFN-𝛾, but also TGF-𝛽1, IL-10, and thatmASC-treatedEAEmicehadfewerCD11c+TNF-𝛼+DCs StemCellsInternational 7 8 8 6 ore ore al sc al sc 6 nic 4 nic ∗ n cli n cli 4 mASC ∗ a a Me Me ∗ 2 2 ∗ mASC 0 0 0 5 10 15 20 25 30 0 5 10 15 20 25 30 Days after immunization Days after immunization Control mASC (a) 160 8 160 Cumulative disease index 1248000 ∗ ∗ Mean clinical score 246 hACSs ∗ Cumulative disease index 1482000 ∗ 0 0 0 5 10 15 20 25 30 0 Days after immunization Control Control mASCs at onset Control mASCs at acute phase hASC hASC (b) (c) (d) Figure2:AllogeneicandxenogeneicASCreduceEAEseverity.ChronicprogressiveEAEwasinducedinC57Bl/6micebyimmunization 6 withMOG35−55.AnimalswereinjectedintraperitoneallywithPBS(control)orwithhypoxia-expandedallogeneicmASCs(10 cells/mouse) eitherattheonsetofEAE(day11afterimmunization)whenthemeanclinicalscorewas0.25inthecontrolgroup(𝑛 = 8mice)and0.22in themASC-treatedgroup(𝑛=9mice)orattheacutephaseofEAE(day15afterimmunization)whenthemeanclinicalscorewas2.7inthe controlgroup(𝑛=6mice)and2.8inthemASC-treatedgroup(𝑛=7mice).(a)Clinicalsymptomswerescoreddailyinthegrouptreatedat onset(leftpanel)andattheacutephaseofEAE(rightpanel).(b)Cumulativediseaseindexisthesumofdailyclinicalscoresobservedbetween days20and40.Resultsareshownasmean(SD).∗𝑝<0.05versuscontrol.(c)HumanASCsameliorateEAE.MicewithMOG-inducedEAE wereintraperitoneallyinjectedwithPBS(control,𝑛=7)orwithhASCsexpandedinhypoxia(106cells/mouse,𝑛=6mice)duringtheacute phaseofthedisease(arrow)whenthemeanclinicalscorewas1.5inbothgroups.Clinicalsymptomswerescoreddaily.(d)Cumulativedisease indexisthesumofdailyclinicalscoresobservedbetweendays20and40.Resultsareshownasmean(SD).∗𝑝<0.05versuscontrol. compared to control EAE mice (Figure 5(d) and represen- 4.Discussion tative dot plots in Figure S5). In order to further analyze DC function we purified CD11c+ cells from the DLNs of MSCs have emerged as a promising therapy for MS with controlandmASC-treatedEAEmiceandassessedtheirTNF- some studies suggesting an MSC-mediated modulation of 𝛼 producing capacity upon LPS restimulation in vitro. We theaberrantimmuneresponse[17,21].However,theefficacy found that DCs from mASC-treated animals produced less hassofarbeenmodestwarrantingthesearchforimproved TNF-𝛼 upon LPS stimulation in vitro compared to control MSC culture protocols and a greater understanding of the DCs (Figure 5(d)). Taken together, these data suggest that mechanisms behind the immunomodulatory and trophic mASCcouldsuppressEAEpartlybypreventingDCfunction effectsofMSCsinvitroandinvivo[16,17,19].Ouranalysis invivo. oftheimmunomodulatorymechanismsemployedbymASCs 8 StemCellsInternational Klüver-Barrera staining MBP IHC Control mASC Control mASC (a) 40 Cervical SC 50 Lumbar SC U) 300 3 U) A. A. Plaque # 1230000 Control mASCs Plaque # 123400000 Control mASCs Relative gene expression ( 1200000 IDO Foxp∗3 TGF∗𝛽 IFN∗𝛾 IL-1∗0 MCP-1 012 Relative gene expression ( Control mASC (b) (c) Figure3:AllogeneicmASCsreducedemyelinizationandinflammatoryinfiltrationinthecentralnervoussystemduringEAEprogression. MicewithMOG-inducedEAEwereintraperitoneallyinjectedwithPBS(control,𝑛=4)orwithmASCsexpandedinhypoxia(106cells/mouse, 𝑛=4)aftertheonsetoftheclinicalsymptoms,andspinalcordswereobtainedatthepeakofthedisease7dayslater.((a)and(b))Transverse sectionsofcervicalandlumbarregionsofspinalcordwererandomlyanalyzedforthepresenceofinflammatoryinfiltratesandplaquesof demyelinizationusingtheKlu¨ver-Barrerastaining(leftpanels)andanti-myelinbasicproteinimmunohistochemistry(rightpanels).Arrows pointtoareasofdemyelinization.(c)GeneexpressionofdifferentinflammatorymarkerswasdeterminedbyPCRinmRNAsisolatedfrom thespinalcords.Resultsareshownasmean(SD)of4micepergroup.∗𝑝<0.05versuscontrol. culturesat5%O2showedthatboththeiriNOSandCOX-1/2 microglia in controlling autoimmune-mediated inflamma- activitiesareinvolvedintheinhibitionofTcellproliferation tionintheCNSremainstobeelucidated[47,48]. invitro.AlthoughiNOSisusedonlybymurineMSCs,COX- In agreement with previous studies [11, 27], we found 1/2 activity has been shown to be important for the immu- that early injection of mASCs reduced the activation of nomodulatorycapacityofbothmurineandhumanMSCs[29, MOG35−55-specific T cells and significantly inhibited EAE 45,46].InadditiontoiNOSandCOX-1/2,wealsodetected progressionandseverity. asignificantinductionofarginaseactivityinthemASC:sple- Regulatory T cells (Tregs) control the severity of EAE nocyte cultures, which was found predominantly in the throughthepreventionofTcellexpansionintheDLNs[49] adherent splenocyte fraction and not in the mASC popula- and by inhibiting activated T cells in the CNS [50]. Several tion(FigureS6).Thisisinagreementwithourrecentfinding studieshavereportedanincreasedfrequencyofTregsinthe + thatMSCscaninducearginaseI regulatorymacrophagesin LNsofEAEanimalstreatedwithMSCs[26,51].Interestingly, vitro[29].Althougharginaseactivitydoesnotappeartobe two studies observed an increase in Treg number [17] or importantfortheMSC-mediatedinhibitionofTcellprolifer- foxp3 mRNA [21] in PBMCs from MS patients after MSC ationinvitro,theroleofMSC-educatedmacrophagesand/or administration.Incontrast,wedidnotobserveanincreased StemCellsInternational 9 DLN Spleen DLN Spleen Cytokine-producing cells (%) 01234 𝛾IFN∗ IL-4 IL-10 IL-17∗ 𝛾IFN∗ IL-4 IL-10 *IL-17∗ 01234 Cytokine-producing cells (%) 3×10Proliferation (CPM) 123400000 mulated OG3555–∗ nti-CD3 mulated OG3555– nti-CD3 12345060000003×10Proliferation (CPM) CmoAnStCrosl Unsti M A Unsti M A Control mASCs (a) (b) DLN Spleen 60 30 L) 𝛾[IFN] (ng/m 2400 ∗ 1200𝛾FN] (ng/mL) [I 0 0 5 ng/mL) 12 812 g/mL) cells (%) 43 [IL-17] ( 48 ∗ 4 [IL-17] (n ++43Dfoxp 12 ∗ C 0 0 0 mulated OG3555– nti-CD3 mulated OG3555– nti-CD3 DLCNontrol Spleen nsti M A nsti M A mASCs U U Control mASCs (c) (d) + Figure4:AllogeneicmASCsinhibitMOG-specificimmuneresponsesbutdonotinducefoxp3 Tregs.mASCswereinjectedintraperitoneally inMOG-inducedEAEmiceaftertheonsetoftheclinicalsymptomsandspleensandDLNswereisolatedatthepeakofthedisease.Untreated EAEmicewereusedascontrols.(a)Thepercentagesofcytokine-producingcellsinspleenandDLNsweredeterminedbyintracellularflow cytometryasdescribedinMaterialsandMethods.((b)and(c))DLNandspleencellswereculturedwithmedium(unstimulated)orstimulated withMOG35−55(50𝜇g/mL)oranti-CD3(1𝜇g/mL,usedasunspecificstimulation).Proliferationwasmeasuredafter3daysby[3H]-thymidine + + incorporation(b).Thecontentofcytokineswasdeterminedinculturesupernatantsafter48hours(c).(d)ThepercentagesofCD4 Foxp3 TcellsinspleenandDLNsweredeterminedbyflowcytometry.Resultsareshownasmean(SD)from4micepergroup.∗𝑝 < 0.05versus control. + + + frequencyofCD4 foxp3 TcellsorIL-10 cellsinspleenand relativetodiseaseprogressionandtothenumbersofMSCs lymph nodes nor increased foxp3 mRNA levels in the CNS reaching the DLNs [26, 51]. These differences could also aftermASCinjectionintoEAEmicewhichisinaccordance partiallyexplainwhytheclinicalefficacyofMSCsinMSisnot withotherstudiesonMSCsandEAE[27,52,53].Thesedis- quiteestablished.Importantly,inanotherMS-trial,Connick crepanciescouldbeduetodifferencesintheimmunomod- et al. [54] observed a cessation of the protective effects six ulatorymechanismsemployedbymurineandhumanMSCs, months after intravenous infusion of MSC, suggesting that underlyingdifferencesinthepathophysiologicalmechanisms a possible induction of Tregs by MSCs was not enough to between EAE and MS, and the timing of MSC injection restore immune homeostasisin MS patients. Future studies 10 StemCellsInternational CD40 TNF-𝛼 of control)17050 ∗ ∗ ∗ ∗ ng/mL) 2.0 pression (% 2550 𝛼[TNF-] ( 1.0 ∗ ∗ ∗ x E 0 0 CD80 IL-10 150 0.8 ol) r 125 cont 100 mL) 0.6 of g/ pression (% 257505 [IL-10] (n 00..24 x E 0 0 CD86 IL-12 150 30 ol) of contr110205 g/mL) 20 pression (% 257505 [IL-12] (n 10 x E 0 0 0 0.5 1 2 4 0 0.5 1 2 Number of mASCs (×105) Number of mASCs (×105) (a) 100 # ol) 80 ∗ r nt o 60 c % 0 ( 40 4 D 20 C 0 100 80 ol) ntr 60 o c % 40 𝛼F ( 20 # N T ∗ 0 % of control)1680000 % control)1680000 # Proliferation ( 24000 ∗ Proliferation ( 24000 ∗ Control mASC LPS − + + + + mASC − − + + + L-NAME − − − + − Indo − − − − + (b) (c) Figure5:Continued.
Description: