ScandinavianJournalofImmunologyVolume87,Issue4EXPERIMENTALIMMUNOLOGYFreeAccessEndothelialextracellularvesiclesmodulatethemacrophagephenotype:PotentialimplicationsinatherosclerosisS.He

DepartmentofCardiology,UnionHospital,TongjiMedicalCollege,HuazhongUniversityofScienceandTechnology,Wuhan,China

SearchformorepapersbythisauthorC.Wu

DepartmentofCardiology,UnionHospital,TongjiMedicalCollege,HuazhongUniversityofScienceandTechnology,Wuhan,China

DepartmentofUltrasound,UnionHospital,TongjiMedicalCollege,HuazhongUniversityofScienceandTechnology,Wuhan,China

SearchformorepapersbythisauthorJ.Xiao

DepartmentofCardiology,UnionHospital,TongjiMedicalCollege,HuazhongUniversityofScienceandTechnology,Wuhan,China

SearchformorepapersbythisauthorD.Li

DepartmentofCardiology,UnionHospital,TongjiMedicalCollege,HuazhongUniversityofScienceandTechnology,Wuhan,China

SearchformorepapersbythisauthorZ.Sun

DepartmentofCardiology,UnionHospital,TongjiMedicalCollege,HuazhongUniversityofScienceandTechnology,Wuhan,China

SearchformorepapersbythisauthorM.Li

CorrespondingAuthor

E-mailaddress:liming518888cn@163.com

http://orcid.org/0000-0002-8934-1084

DepartmentofCardiology,UnionHospital,TongjiMedicalCollege,HuazhongUniversityofScienceandTechnology,Wuhan,China

Correspondence

M.Li,DepartmentofCardiology,UnionHospital,TongjiMedicalCollege,HuazhongUniversityofScienceandTechnology,Wuhan,China.

Email:liming518888cn@163.com

Searchformorepapersbythisauthor
S.He

DepartmentofCardiology,UnionHospital,TongjiMedicalCollege,HuazhongUniversityofScienceandTechnology,Wuhan,China

SearchformorepapersbythisauthorC.Wu

DepartmentofCardiology,UnionHospital,TongjiMedicalCollege,HuazhongUniversityofScienceandTechnology,Wuhan,China

DepartmentofUltrasound,UnionHospital,TongjiMedicalCollege,HuazhongUniversityofScienceandTechnology,Wuhan,China

SearchformorepapersbythisauthorJ.Xiao

DepartmentofCardiology,UnionHospital,TongjiMedicalCollege,HuazhongUniversityofScienceandTechnology,Wuhan,China

SearchformorepapersbythisauthorD.Li

DepartmentofCardiology,UnionHospital,TongjiMedicalCollege,HuazhongUniversityofScienceandTechnology,Wuhan,China

SearchformorepapersbythisauthorZ.Sun

DepartmentofCardiology,UnionHospital,TongjiMedicalCollege,HuazhongUniversityofScienceandTechnology,Wuhan,China

SearchformorepapersbythisauthorM.Li

CorrespondingAuthor

E-mailaddress:liming518888cn@163.com

http://orcid.org/0000-0002-8934-1084

DepartmentofCardiology,UnionHospital,TongjiMedicalCollege,HuazhongUniversityofScienceandTechnology,Wuhan,China

Correspondence

M.Li,DepartmentofCardiology,UnionHospital,TongjiMedicalCollege,HuazhongUniversityofScienceandTechnology,Wuhan,China.

Email:liming518888cn@163.com

Searchformorepapersbythisauthor

Giveaccess

Sharefulltextaccess

Sharefull-textaccess

PleasereviewourTermsandConditionsofUseandcheckboxbelowtosharefull-textversionofarticle.IhavereadandaccepttheWileyOnlineLibraryTermsandConditionsofUseShareableLink

Usethelinkbelowtoshareafull-textversionofthisarticlewithyourfriendsandcolleagues.Learnmore.

CopyURL

Sharealink

ShareonEmailFacebookTwitterLinkedInRedditWechatAbstract

Endothelialcells(ECs)andmacrophagesengageintightandspecificinteractionsthatplaycriticalrolesincardiovascularhomeostasisandthepathogenesisofatherosclerosis.Extracellularvesicles(EVs)arecircularmembranefragmentsreleasedfromtheendosomalcompartmentasexosomesorshedfromthesurfacesofthemembranesofmostcelltypes.IncreasingevidenceindicatesthatEVsplayapivotalroleincell‐to‐cellcommunication.However,thecontributionofEVs,asdeterminebyoxidizedlow‐densitylipoprotein(ox‐LDL)‐exposedand/orKruppel‐likefactor2(KLF2)‐transducedECsintheinteractionbetweenvascularECsandmonocytes/macrophages,whichisakeyeventinatheroscleroticplaquedevelopment,hasremainedelusive.ThisstudydemonstratesthecharacteristicimpactofEVsfromox‐LDL‐treatedand/orKLF2‐transducedECsonthemonocyte/macrophagephenotypeinvitroandinvivo.Q‐PCRshowedthatboththeatherosclerosisinducerox‐LDLandatheroprotectivefactorKLF2regulatedinflammation‐associatedmicroRNA‐155(miR‐155)expressioninhumanumbilicalveinendothelialcells(HUVECs).Moreover,coculture,immunofluorescenceandflowcytometryrevealedthatmiR‐155wasenrichedinox‐LDL‐inducedECs‐EVsandsubsequentlytransferredtohumanmonocyticTHP1cells,inwhichthesevesiclesenhancemonocyteactivationbyshiftingthemonocytes/macrophagesbalancefromanti‐inflammatoryM2macrophagestowardsproinflammatoryM1macrophages;EVsfromKLF2‐expressingECssuppressedmonocyteactivationbyenhancingimmunomodulatoryresponsesanddiminishingproinflammatoryresponses,whichindicatethepotentanti‐inflammatoryactivitiesofthesecells.Furthermore,oilredstainingshowedthatatheroscleroticlesionswerereducedinmicethatreceivedEVsfromKLF2‐transducedECswithdecreasedproinflammatoryM1macrophagesandincreasedanti‐inflammatoryM2macrophages,andthiseffectisatleastpartlyduetothedecreasedexpressionofinflammation‐associatedmiR‐155,confirmingourinvitrofindings.Insummary,thisstudyprovidesnovelinsightsintothepathophysiologicaleffectsofalteredEVsecretionand/ormicroRNAcontentandtheirinfluenceonmodulatingmonocyteactivationdependingontheenvironmentsurroundingEVs‐releasingECs.

1INTRODUCTION

Atherosclerosisisaconditionthatiscausedbylipid‐inducedinflammationofthevesselwallorchestratedbyacomplexinterplayofvariouscelltypes,suchasendothelialcells(ECs),smoothmusclecellsandmacrophages.SignallingandcommunicationbetweenECsandmonocytes/macrophagesplayacriticalroleincardiovascularhomeostasisandthepathogenesisofatherosclerosis.1-5

Classicalcommunicationinvolvescelljunctions,adhesioncontactsandsolublefactors,whichcanactuponthesamecellinwhichtheyareproducedoruponneighbouringcellsormayevenactoverlongdistancesinanendocrinemanner.6Recently,anadditionalmechanismofcellcommunicationinvolvingextracellularvesicles(EVs)hasbeenidentified.7EVsaresmall(0.05‐1μm)membrane‐boundparticlesreleasedbyhealthyanddiseasedcells.8EVsarealsoknownasexosomes,microparticlesorsheddingvesicles.9TheconceptthatEVsmayactasparacrine/endocrineeffectorsisbasedonevidencethatthesevesiclesareabletotransferbioactivemolecules,includingmembranereceptors,proteins,mRNAs,miRNAsandorganelles,totargetcells,thusmediatingcellactivationandphenotypicmodifications,aswellasreprogrammingofcellfunction.10-12

WhetherEVscanacceleratediseaseprogressionorinducevascularprotectionisstillamatterofdebate.NumerousevidencehasshownthatEVscanplayamajorroleininflammation,thrombosisandcoagulation‐allofwhichareconditionsthatareinvolvedinatherogenesis.13-16However,recentdatachallengedthepresumeddeleteriousroleofEVsandsuggestedthatEVspromotecellsurvival,exertanti‐inflammatoryeffects,counteractcoagulationprocessesorinduceendothelialregeneration.17,18Takentogether,theroleofEVsinthemaintenanceofvascularhomeostasisismorecomplexthaninitiallyconsidered.

DiversestimuliandhemodynamicforcemodulatetheendothelialphenotypeandtherebyimpacttheeffectsofEVs.Oxidizedlow‐densitylipoprotein(ox‐LDL)canelicitanarrayofatherogenicresponsesbyinducingaproinflammatoryECsphenotype.EVsproductionisenhancedbyenvironmentalstimuli,includingoxidativestress,hypoxiaandinjury.19Thelevelsofendothelialcell‐derivedEVsaresignificantlyincreasedindiseaseswithsystemicendothelialdamageasobservedinhypertension,diabetesmellitusandcoronaryarterydisease.18,20Kruppel‐likefactor2(KLF2)hasemergedasacriticalintegratorofflow‐mediatedchangesintheECsphenotype,includinganti‐inflammatoryandvasoprotectivestates.21Notably,recentevidenceindicatesthatmicroRNAs(miR)canbesecretedandintercellularytransportedbymicrovesicles(MVs)derivedfromKLF2‐expressingECs.KLF2inducedupregulationofthemiR‐143/145clusterinECs,whichcontrolsvascularsmoothmusclecellphenotypesthroughaMVs‐mediatedmechanism.22Furthermore,ECs‐specificmiR‐126signalstheneedforendothelialrepairthroughitstransferinMVs.23miR‐126alsosuppressesendothelialvascularcelladhesionmolecule‐1expressionandleucocyteadhesionbyactivatingECs,whichmaycontributetoitsatheroprotectiverole.24Theabove‐mentioneddataindicatethatthedescribedparadoxicaleffectsofEVsmayvarydependingontheenvironmentsurroundingEVs‐releasingECs.

Monocyteandmacrophagephenotypesarehighlyheterogeneousandcanbedynamicallymodulatedbythemicroenvironment.25-27Classicalactivationpromotesaproinflammatory(M1‐like)response,whichincludesthesecretionofproinflammatorycytokinesandreactiveoxygenandnitrogenspecies,andisdrivenbyexposuretobacteriallipopolysaccharide(LPS)orTh1cytokines,suchasinterferon‐ɣ28.Conversely,exposuretoTh2cytokines,suchasinterleukin‐4(IL‐4),IL‐10,orIL‐13,supportsalternativeactivation,whichisanimmunomodulatory,proangiogenicandtissue‐reparative(M2‐like)responsethatincludesthesecretionofIL‐10andtransforminggrowthfactorb(TGF‐b).26Thebalanceofproinflammatoryvs.immunomodulatoryresponsesappearstoplayanimportantbutpoorlyunderstoodroleincardiovascularpathologies.26,28-30Here,weaimtoinvestigatetheeffectsofthealterationsofEVsonmonocyte/macrophagepolarization.

Primaryhumanumbilicalveinendothelialcells(HUVECs)wereisolatedaspreviouslyreported.31HUVECswereisolatedfromcannulatedhumanumbilicalveins,perfusedwithHanks’solutiontoremovebloodandthenincubatedwith1%Collagenasefor15minutesat37°C.Afterremovalofcollagenase(Sigma,USA),cellswereculturedinMl99medium(Gibco,USA)supplementedwith20%foetalcalfserum(Gibco),100μg/mLheparin(Sigma),50μg/mlendothelialcellgrowthsupplement(Gibco),25mmol/LHepesbuffer,2mmol/Ll‐glutamine,100U/mLpenicillinand100μg/mLstreptomycinandgrownat37°Contissuecultureplatescoatedwith0.1%gelatine.Cellswerepassagedatconfluencebysplittingataratioof1:4andwereusedwithinthefirst8passages.TheuseofhumanumbilicalveinsfromnormaldonorswasapprovedbytheEthicsCommitteeofTongjiMedicalCollege,HuazhongUniversityofScienceandTechnology.

AhumanmonocyticTHP1celllinederivedfromanacutemonocyticleukaemiawaspurchasedfromATCC(ATCC®TIB‐202™)andmaintainedinthelaboratoryusingRPMI1640mediumsupplementedwith10%FBS,2mmol/Ll‐glutamine,1mmol/Lsodiumpyruvate,100U/mLpenicillinand100μg/mLstreptomycinat37°CinaCO2incubator.

ApreviousstudyidentifiedthattreatingnaïveTHP‐1cellswithphorbol12‐myristate13‐acetate(PMA)significantlyincreasedthesurfaceexpressionofmacrophagedifferentiationMarkerscomparedtovehicle‐treatedsamples;therefore,inthisstudy,THP‐1cellsweretreatedwith60nmol/LPMAtoevolveintocellswithamacrophage‐likephenotype.32

Mouseprimaryaorticendothelialcells(mouseECs)werepurchasedfromCellBiologics(C57‐6052;Chicago,USA).ThesemouseECswereculturedinDMEM(highglucose,noHEPES)supplementedwith10%FCS,1%sodiumpyruvate,1%glutamineand1%penicillinandstreptomycin.

Mouseperitonealmacrophageswereharvestedafterintraperitonealinjection(5mLpermouse)ofphosphate‐bufferedsaline(PBS).Aftercentrifugationat1000gfor5minutes,thecellswereresuspendedincompleteRPMI1640mediumcontaining10%foetalbovineserum(FBS)and100U/mLofpenicillin/streptomycinandthenadjustedtoaconcentrationof1×106/mLandusedforfurtherexperiments.

BloodforlipoproteinisolationwascollectedinEDTA(1mg/mL)fromnormallipidaemicdonorsafter12hoursoffasting.LDL(density=1.03‐1.063g/L)wasisolatedfromtheplasmaafteradensityadjustmentwithKBrbypreparativeultracentrifugationat300000g/minfor22hours,usingatype50rotor.ThesamplesweredialysedagainstPBScontaining0.3mmol/LEDTA,sterilizedbyfiltrationthrougha0.22‐mmfilterandstoredundernitrogengasat4°C.TheproteincontentwasdeterminedbythemethodofLowryetal31CopperoxidationofLDLwasperformedbyincubationofpost‐dialysedLDL(1mgofprotein/mLinEDTA‐freePBS)withcoppersulphate(10mmol/L)for24hoursat37°C.Lipoproteinoxidationwasconfirmedbyanalysisofthiobarbituricacid‐reactivesubstances.

Inthisstudy,wefirstconstructedavectorcarryingKLF2cDNAfortheproductionofadenovirus.Briefly,KLF2cDNAwasamplifiedusingRNAsamplesfrom293TcellsaccordingtoGenBanksequences.TheprimersforamplifyingKLF2cDNAwere5′‐GAGGATCCCCGGGTACCGGTCGCCACCATGGCGCTGAGTGAACCCATC‐3′and5′‐TCACCATGGCGACCGGCATGTGCCGTTTCATGTGCAGC‐3′.TheseoligonucleotidesweresubsequentlysynthesizedbyQiagen(Valencia,USA),andKLF2cDNAwasamplifiedusingPCR.Subsequently,theseKLF2cDNAfragmentsweretheninsertedintotheCMV‐MCS‐EGFPGV135vectorandligatedintoashuttleplasmid;theshuttleplasmidandadenoviralbackboneplasmidwerecotransfectedintoHEK‐293AcellstoproducetherecombinantadenoviralvectorAd‐KLF2.AnotherwiseidenticalvectorwithoutKLF2cDNAwasusedtogenerateemptyvirusesascontrols(emptyvectors).

ThetitresofthenewlyproducedadenoviruseswereassessedusingtheTCID50methodandpreparedattitresof1×1010PFU/mL.Particularly,HUVECswereseededatadensityof1×106cell/mLonto6‐wellplatesovernightandinfectedwith20μLofadenovirus(1×1010PFU/mL)forupto48hours.Beforeoraftervirusinfection,HUVECsweretreatedwith50μg/mLofox‐LDLfor24hours,andthen,EVswereisolatedfromtheconditionedmediumfordifferentexperiments.

WealsoconstructedanEGFPvectorbyPCRamplificationofEGFPcDNA.Theprimers(thesequenceswere5′‐TTTATGGTGAGCAAGGGCGAG‐3′and5′‐TTTTGGTGCAGATGAACTTCAG‐3′)wereinsertedintoCMV‐MCS‐EGFPGV135vector.AfteramplificationandDNAsequenceconfirmation,thisplasmidwasusedtoproduceadenoviruscarryingEGFPcDNA.Next,thisadenoviruscarryingEGFPcDNAwastheninfectedintoHUVECsfor48hours,andtheseHUVECsweresubsequentlycoculturedwithTHP1inaTranswellsystem(seebelow).

EVswereisolatedfromconditionedmediumofculturedHUVECsinfectedwithKLF2cDNAcontainingadenovirusesand/ortreatedwithox‐LDL(50μg/mL)for24hoursaccordingtopreviousstudies.21Briefly,culturesupernatantswerecollectedandpreclearedbycentrifugationat400gfor5minutesandthen2000gfor20minutestoeliminatedeadcellsandcellulardebris.Thesupernatantwasthenultracentrifugedat120000gfor120minutesat4°C,followedbywashingoftheEVspelletwithPBSat120000gfor120minutesat4°C.PelletedEVswereresuspendedinculturemediaandusedimmediately.SurfaceexpressionofEVsmarkerCD63wasquantifiedusingaflowcytometerandanalysedwithCellQuestsoftwareonaFACSCantoII(BDBiosciences).

TheEVssizeandnumberwereassessedwiththeNanoSightNS300system(Nanosight,UK)followingthemanufacturer"sprotocol.Thedetailsofthismethodologyhavepreviouslybeendescribed.33FinalpelletedEVswereresuspendedanddilutedwithPBSataconcentrationrangeof4‐8×108particles/mL,and1mLwasusedforNanoSightanalysis.TheNanoSightsystemusesalaserlightsourcetoilluminatenanoscaleparticlesanddetectthemindividuallyaslight‐scatteredpointsmovingviaBrownianmotion.Polydispersitywasquantified,andweusednanoparticletrackinganalysis(NTA)softwaretotrackandsizenanoparticlesonanindividualbasis.Theresultsaredisplayedasafrequencysizedistributiongraphdescribingthenumberofparticlespermillilitre.TheconcentrationofreleasedEVswascalculatedtodeterminetheaveragenumberofEVs.EVswereusedataconcentrationof2500particles/ulforallexperiments,unlessindicatedotherwise.

EVsfromKLF2‐transducedHUVECswereisolatedandincubatedwithRNase,proteinaseKorTritonX‐100for45minutesat37°CbeforeisolatingRNAandmeasuringthelevelsoftheRNU‐6ormiR‐155byreal‐timePCR.

Theabsenceofresidualox‐LDLinEVssampleswasconfirmedusingELISA.EVspreparationswerecheckedforendotoxincontaminationusingtheLimulusamebocytelysateassay.Theendotoxincontentwasalways0.05ng/mL.Insomeexperiments,thesupernatantsresultingfromthelastwashwereusedasacontrol.

Briefly,EVswerepelletedbycentrifugationandthenfixedwitha2.5%glutaraldehydesolution(Sigma)at4°Covernight.Onthenextday,thesampleswerepelletedbycentrifugationandwashedwith0.1mol/LPBSandthenpost‐fixedwith1%(w/v)osmiumtetroxide(Sigma)for2hoursatroomtemperature.Afterdehydrationwithagradedseriesofethanol(70%‐100%,allbycentrifugationateachstep),theseEVswereembeddedinpropyleneoxideandEpon(Sigma)andsolidifiedat60°Cfor48hoursin100%epoxyresinforpreparingultrathinsections,andthesesectionsweresubsequentlystainedwithuranylacetateandleadcitrateandreviewedunderatransmissionelectronmicroscope(JEOL,Japan).

ToassessHUVECscommunicationwithTHP1,weutilizedtranswellchambers(0.4μmporesize,Corning,USA)tocultureTHP1withHUVECs,EVsisolatedfromdifferentgenesmanipulatedortreatedHUVECs.Briefly,THP1cellswereseededontothebottomwellatadensityof6×104cellsincompleteRPMI1640medium,andthen,differentamountsofEVsorHUVECswereaddedintotheupperchamber;theplateswerethenculturedfor24hours.Attheendofeachexperiment,THP1cellswereobservedunderafluorescencemicroscope(OlympusMicroscopeBX‐51,Japan).

Twenty‐fourhoursaftercellseeding,PMA‐treatedTHP1cellswereincubatedwithavaryingnumberofEVsfromthe3plantsfor24hoursat37°C.FollowingremovaloftheEVsfromeachwell,THP1cellswerewashedinphosphate‐bufferedsaline.THP1cellswerethenincubatedinserum‐freeDMEMtowhichMTT(0.5mg/mL)wasaddedtoeachwell(100mL)andincubatedforanadditional4hours.Then,themediumwasremovedandtheTHP1cellswereincubatedfor15minuteswith100mLofacidicisopropanol(0.08NHCl)todissolvetheformazancrystals.TheabsorbanceoftheMTTformazanwasdeterminedat570nminanenzyme‐linkedimmunosorbentassay(ELISA)reader.Viabilitywasdefinedastheratio(expressedasapercentage)ofabsorbanceoftreatedcellstountreatedcells.

TotalcellularRNAwasisolatedfromcellsusingtheRNAsimpleTotalRNAKit(Tiangen,China)andreversetranscribedintocDNAusingthePrimeScriptRT‐PCRkit(TaKaRaBioInc.,Japan)accordingtothemanufacturers’protocols.ThepelletedEVswereresuspendedin700μLofQiazol(Qiagen)forRNAisolationusingthemiRNeasykit(Qiagen)accordingtothemanufacturer"sinstructions.Next,thesecDNAsampleswereamplifiedinanABI7500fastReal‐TimePCRsystem(ABI,USA),andU6snRNAwasusedasaninternalcontrolformicroRNAexpressionandβ‐actinforothergeneexpression.ThePCRamplificationconditionsweresettoaninitial95°Cfor15minutesandthen40cyclesof94°Cfor15seconds,60°Cfor30seconds,and70°Cfor35seconds.Primersused:KFL2(5′‐GCACGCACACAGGTGAGAAG‐3′and5′‐ACCAGTCACAGTTTGGGAGGG‐3′);CD80(5′‐TGACAACCAACCACAGCTTC‐3′and5′‐AGCAGTAGGTCAGGCAGCAT‐3′);CD86(5′‐ACACGGTTACCCAGAACCTA‐3′and5′‐GGTGAAGATAAAAGCCGCGT‐3′);CD206(5′‐AACGAGGCTACATATGCCAGA‐3′and5′‐GCGATCCACACACGTTCATTA‐3′)andβ‐actin(5′‐CACGATGGAGGGGCCGGACTCATC‐3′and5′‐TAAAGACCTCTATGCCAACACAGT‐3′).ToamplifymaturemiRNAsequences,TaqManmiRNAassayswereusedtomeasurethelevelsofmaturemiRNAs.RNU6‐2wasusedasanormalizationcontrolinallmiRNAmeasurements.Real‐timePCRwasperformedintriplicate,andtherelativeexpressionlevelsofmiRNAormRNAwereanalysedusingthe2−ΔΔCtmethod.

Expressionofdifferentcytokinesandox‐LDLwasassessedusingELISA.Briefly,conditionedmediumfromdifferentHUVECsculturesorEVssampleswascollectedforthedetectionofTNF‐a,IL‐6,IL‐10orox‐LDLusingcommerciallyavailableELISAkitsfromR&Dsystems(Minneapolis,USA)orBio‐SwampLifeScienceLab(Wuhan,China)accordingtothemanufacturers’instructions.Eachsamplewasassayedintriplicateandrepeatedatleastonce.

PMA‐treatedTHP1cellswereisolatedfromtheconditionedTHP1‐EVscoculturedsystem.Next,THP1cellsweretreatedwith0.25%trypsinandaliquotedatadensityof1×106cells/100μLintofluorescence‐activatedcellsorting(FACS)tubesandthenincubatedwithananti‐CD80,CD86orCD206antibody(allfromR&Dsystems,USA)for1hour.Thecellswerewashedwithaflowcytometrystainingbufferandthenfurtherincubatedwithahorseanti‐mousesecondaryantibodyconjugatedwithfluoresceinfor30minutesinthedark.Afterwashingwiththeflowcytometrystainingbuffer,thecellsweresubjecttoflowcytometricanalysisusingaflowcytometerandanalysedwithCellQuestsoftwareonaFACSCantoII(BDBiosciences).

ThisstudywasapprovedbyTongjiMedicalCollegeofHuazhongUniversityofScienceandTechnologyInstitutionalAnimalCareandUseCommitteeandperformedinaccordancewiththeprotocolguidelines.Briefly,maleapolipoproteinE‐deficient(ApoE−/−)miceat6weeksofagewerepurchasedfromExperimentalAnimalDepartmentofMedicalCentre,PekingUniversity(Beijing,China).Themicewerefedahigh‐fatdiet(HFD,WesternDiet,TD.88137,HarlanTeklad,Madison,WI)containing17.3%protein,48.5%carbohydrate,21.2%fatand0.2%cholesterolbyweight,with42%ofkcalfromfat,whereasApoE−/−controlmicewerefedanormalstandarddiet(OrientalYeast)for8weeks.Micewererandomlydividedintothefollowing5groups,ApoE−/−micefedwithstandarddiet,ApoE−/−micefedwithHFD(ApoE−/−)andApoE−/−micethatwereinjectedintravenouslyviathetailveintwiceperweekwithEVs(1×107EVsdilutedin200μLofsterilizedPBS,derivedfromemptyvectorormiR‐155‐treatedmouseECsorfromKLF2‐transducedmouseECstreatedwithorwithoutmiR‐155(10ng/mLfinalconcentration,Exiqon).Attheendofexperiments,micewereanaesthetizedat14weeksofagewithavertin(Sigma).Thearchportionsoftheaortawereseparatedandfixedwith4%paraformaldehydeinDEPC‐treatedPBSandsubsequentlyembeddedwithparaffinforOilredstaining.

Analysisofatheroscleroticplaqueswasperformedaspreviouslydescribed.25TheheartanddescendingaortawereexcisedandfixedinambientPBS/4%formalin30%sucroseovernightbeforemountinginOCTmediumandfreezingat−70°C.Aorticsinusconsequentcryosections(6μm)werestainedwithOilRedO.Forquantitativeanalysisofatherosclerosis,thepercentlesionareaineachof5sectionsfromeachmousewasobtained.Enfacepreparationsofthedescendingaortawerewashedindistilledwater;enfaceanalysisofthedescendingaortawasperformedafterstainingthedescendingaortawithOilRedO.TheareaoftheplaquewasmeasuredusingImage‐ProPlus6.0.

HUVECsweregrownovernightin6‐wellplatesandlabelledwiththeDil‐C16immunofluorescencedyeortransfectedwithFITC‐labelledmiR‐155for6hoursat37°Candwashed3timeswithPBS.Subsequently,EVswereisolatedfromtheseHUVECs.THP1cellswerecoculturedwiththeseEVsfor24hoursandsubsequentlyfixedwith4%paraformaldehydeatroomtemperaturefor10minutes;stainedwithDAPI;reviewedunderanOlympusmicroscope(IX71;OlympusCorporation,Japan)andafluorescentmicroscope(OlympusMicroscopeBX‐51,Japan)oraconfocalmicroscope(Alsi,Nikon,Japan);andquantifiedusingdesignedsoftware.

Theaortatissuesampleswerefixedin4%paraformaldehydeatroomtemperaturefor2hoursandembeddedinparaffinforthepreparationofsections.FrozenconsecutivesectionsoftheseaortatissueswereimmunostainedwithF4/80orCD206accordingtothemanufacturer"sprotocol.Briefly,thesectionswerefixedin4%paraformaldehydeatroomtemperaturefor10minutesandincubatedwithnormalratserumata1:5dilutionfor30minutesfollowedbythemonoclonalanti‐F4/80(thebestknownmousemicrogliaandmacrophagemarker)oranti‐CD206(amarkeroftype2transmembranecellsurfaceglycoproteinexpressedonmonocytesormacrophages)antibody(allfromAbDSerotec,USA)atadilutionof1:100for2hoursattheroomtemperature.Subsequently,thesectionswerewashedwithPBS,counterstainedwithDAPI,reviewedandscoredunderaconfocalmicroscope(Alsi,Nikon).

Bloodsampleswerecollectedbycardiacpunctureatthetimeofeuthanasia,andtheserumwasseparated.Serumlipidsandlipoproteinswereassayedusingcommerciallyavailablekitsandaclinicalchemistryanalyser.ThetriglyceridelevelsweredeterminedusingtheTriglycerideQuantificationColorimetric/FluorometricKit(BioVision,MountainView,CA)asdescribedbytheprotocol.TotalcholesterolwasmeasuredusingtheQuickDetect™TotalcholesterolELISAKit(BioVision,MountainView,CA),HDLandLDLcholesterolweremeasuredbyacolorimetricassayusingaHDLandLDL/VLDLQuantificationColorimetric/FluorometricKit(BioVision,MountainView,CA).

Dataareexpressedasthemeans±SDorasotherwiseindicatedfromatleast3experiments.Student"sttestswereusedtocompare2conditions,andone‐wayANOVAwithBonferroni"scorrectionwasusedformultiplecomparisons.AllstatisticalanalyseswereperformedusingSPSS19.0software(Chicago,IL,USA).Probabilityvaluesof.05wereconsideredsignificant.

ProinflammatoryactivationofECsinducedbyox‐LDLisaninitialeventinthedevelopmentofatherosclerosis.34AsNF‐κBactivationisessentialfortheinflammatoryeffectsofox‐LDL,35-37weassayedtheexpressionofmicroRNAsdescribedasbeingNF‐κB‐dependent‐miR‐155inHUVECsstimulatedwithox‐LDL.Wefoundthatox‐LDLmarkedlyupregulatedmiR‐155expressioninHUVECs,whichwasobservedat3hours,withthemostchangesobservedat6hours.Subsequently,themiR‐155levelsstartedtoreboundat12hoursand,finally,showednodifferencepast24hourscomparedwithHUVECsculturedalone(Figure1A)

Figure1OpeninfigureviewerPowerPointKLF2andox‐LDLmediatedmiR‐155expressioninHUVECs.A,HUVECsweretreatedwithox‐LDLforupto24handtotalcellularRNAwasthenisolatedforqRT‐PCRanalysisofmiR‐155fortheindicatedtimeperiods(n=3foreachtimepoint);B,HUVECswereinfectedwithadenovirusescarryingemptyvector,KLF2cDNA(Ad‐KLF2)for24handtotalcellularRNAwasthenisolatedforqRT‐PCRanalysisofKLF2levels,n=3independentexperiments;C,HUVECswereinfectedwithadenovirusescarryingKLF2cDNAoremptyvectorfor24handtotalcellularRNAwasthenisolatedforqRT‐PCRanalysisofmiR‐126andmiR‐10a,miR‐155,miR‐150,miR‐145andmiR‐143,n=3independentexperiments;HUVECsweretreatedwithox‐LDLforindicatedtimeperiods(D),orinfectedwithadenovirusescarryingKLF2cDNAoremptyvectorfor48handtheninducedwithox‐LDLfor6h(E)orforupto24h(F),totalcellularRNAwasthenisolatedforqRT‐PCRanalysisofKLF2,miR‐155(n=3foreachtimepoint).*and#indicatesvs0horemptyvector;#P.05,*P.05,**P.01,***P.001;cellsinfectedwithKLF2cDNAindicatesAd‐KLF2

AsKLF2exertsanti‐inflammatoryeffectsandinhibitstheproinflammatoryactivationofECsthroughinhibitionofNF‐κBfunctionbycompetingwithCBP/p300,akeycofactorrequiredforoptimalNF‐κBactivity,37wewereinterestedinidentifyinghowinflammation‐associatedmiR‐155wasregulatedbyKLF2inHUVECs.First,wetransducedHUVECswithanadenoviralvectorencodingKLF2(Ad‐KLF2),resultinginaremarkableincrease(5.9‐fold)ofKLF2expression(Figure1B).Then,wemeasuredtheexpressionofanumberofmicroRNAswithanestablishedroleinvascularbiologybyreal‐timePCR.Consistentwithapreviousstudy,21KLF2overexpressionregulatedtheexpressionofmicroRNAs,particularlytheupregulationoftheexpressionofatheroprotectivemiR‐145andmiR‐143(6.2‐foldand7.8‐fold,respectively)anddownregulationoftheexpressionofproinflammationmiR‐155(0.24‐fold,Figure1C)inquiescentHUVECs.

TofurtherdeterminehowendothelialmiR‐155isregulatedbyKLF2intheinflammatorysetting,HUVECsweretreatedwithox‐LDLforupto24hours.KLF2showedthelowestexpressionatthe6‐hourstimepoint(Figure1D).Then,weinfectedHUVECswithadenovirusescarryingKLF2cDNAoranemptyvectorfor48hoursandinducedthemwithox‐LDLfor6hours,resultinginincreasedexpression(4.6‐fold)ofKLF2(Figure1E).Mostimportantly,KLF2overexpressionsignificantlyattenuatedox‐LDL‐mediatedinductionofmiR‐155expressioninHUVECs(Figure1F).

Tostudytheinfluenceofox‐LDLandKLF2ontheregulationofmiR‐155inEVs,weisolatedEVsfromthesupernatantsofanemptyvectorandKLF2‐transducedHUVECswithorwithoutox‐LDL.RNAwasisolatedfromthevesicles,andtheremainingconcentratedsupernatantsandmicroRNAexpressionlevelsweremeasuredbyreal‐timePCR.Indeed,miR‐155wasmarkedlyupregulated(~10‐foldenrichment)inEVsderivedfromox‐LDL‐treatedHUVECsinatime‐dependmannercomparedtothosefromuntreatedHUVECs(Figure2A).KLF2transductionresultedinthedownregulationofmiR‐155inEVsfromquiescentandox‐LDL‐treatedHUVECs(Figure2B,C),whereastheremainingsupernatantsdidnotshowthedifferentiationofmiR‐155levels(Figure2D).Ourexperimentsindicatedthatalthoughox‐LDLandKLF2affectedtheexpressionpatternandlevelofmiR‐155inbothcellsandEVs,EVsshowedremarkablygreaterchangesthanoverallcells.

Figure2OpeninfigureviewerPowerPointKLF2andox‐LDLmediatedmiR‐155expressioninEVs.A,HUVECsweretreatedwithox‐LDLforupto24h,andthentotalcellularRNAfromextracellularvesicels(EVs)wereisolatedforqRT‐PCRanalysisofmiR‐155levels(n=3);B,HUVECswereinfectedwithadenovirusescarryingKLF2cDNAoremptyvectorfor48handtotalcellularRNAwasthenisolatedfromEVsforqRT‐PCRanalysisofKLF2andmiR‐155levels(n=3);C,HUVECswereinfectedwithKLF2cDNAoremptyvectorfor48handthentreatedwithox‐LDLfor6h.Afterthat,totalcellularRNAfromEVswasisolatedforqRT‐PCRanalysisofmiR‐155levels;D,RNAwasisolatedfromtheremainingsupernatantafterultracentrifugation;miR‐155wasmeasuredbyreal‐timePCR.Dataarepresentedasmean±SEMofatleast3independentexperiments.#P.05,***P.001whencomparedwithuntreatedcells

TofurthercharacterizethemicroRNA‐containingEVs,weperformedelectronmicroscopyonisolatedvesicles.Thesevesicleswereidentifiedbytheirmorphologicallyuniformvesicularstructureinelectronmicroscopy(Figure3A).AnalysisofHUVEC‐conditionedmediumbynanoparticletrackinganalysisrevealedthatHUVECssecretedvesicleswithapeakdistributionatapproximately100nm(Figure3B).TheseEVswerepositivefortheexosomalmarkerproteinCD63(Figure3C).Todistinguishbetweenmembrane‐enclosedmicroRNAs,microRNAsthatmightbeattachedtotheoutsideofthesevesiclesandmicroRNAsinproteincomplexes,wedemonstratedthatthemicroRNAswereindeedpresentwithintheEVsbutnotsimplyreleasedasacontaminantduringtheEVspurificationprocess,asthemicroRNAsexpressionlevelinEVstreatedwithRNasewereonlyslightlychangedcomparedtothoseinuntreatedEVs,whilethemicroRNAslevelinEVstreatedwithbothRNaseandTritonX‐100wassignificantlyreduced(Figure3D).

Figure3OpeninfigureviewerPowerPointCharacterizationofEVsfromHUVECs.A,HUVECswereinfectedwithKLF2cDNAoremptyvectorfor48handthentreatedwithox‐LDLfor6h;then,thesupernatantswerecollectedandsubjectedtoanOptiPrep‐baseddensitygradientcentrifugationtoisolateextracellularvesicels(EVs)andtransmissionelectronmicroscopicanalysisofEVs;B,NanoparticletrackinganalysissoftwareanalysisofisolatedEVssize;C,FlowcytometryanalysisofEVspurifiedfromHUVECsmediumandlabelledwithCD63;D,miR‐155wasprotectedfromdegradationbyRNasebyaphosphobilayer‐basedmembrane.VesiclesofKLF2‐transducedHUVECswereisolatedandincubatedwiththeindicatedreagentsbeforeisolatingRNAandmeasuringlevelsofmiR‐155byreal‐timePCR(n=3independentexperiments).Dataarepresentedasmean±SEMofatleast3independentexperiments.***P.001whencomparedwithuntreatedcells

TodeterminewhethertheEVsactsasamessengertodelivermicroRNAsfromECstoTHP1intheabsenceofcell‐cellcontact,weemployedacoculturesystemofHUVECswithTHP1inwhichthecellswereseparatedbyamembraneof0.4‐μmporesizetopreventdirectcellcontactortransferoflargervesicles(Figure4A).WefirstinfectedHUVECswithadenoviruscarryingeGFPcDNAandthencoculturedtheseHUVECswithTHP1cellsinaTranswellsystem.Subsequently,eGFPexpressioninTHP1cellswasdetected,revealingthatTHP1cellswerepositiveforeGFPstaining(Figure4A).Next,welabelledisolatedHUVECswithDil‐C16immunofluorescencedyeandsubsequentlyisolatedEVsfromtheseHUVECs.TheseEVswerethencoculturedwithTHP1cellsintheTranswellsystem.Attheendofeachexperiment,THP1cellswerereviewedunderafluorescencemicroscope,andthedatashowedpositiveDil‐C6fluorescenceintheseTHP1cells(Figure4B).WedirectlyfluorescentlytaggedasyntheticmicroRNAoligonucleotide(FITC‐taggedmiR‐155)andtransfectedthisconstructintoHUVECs.EVsweresubsequentlyisolatedfromculturemediumHUVECsandaddedtoTHP‐1cells.FluorescentlylabelledoligonucleotidesweredetectedinTHP‐1cells.Apreviousstudydemonstratedthatlabellingofhumanmicrovascularendothelialcell‐1cellsisnotduetoDiI‐C16carryoverbuttotheinternalizationofmicrovesiclesintotargetedcells,whichisanactiveprocessthatdependsonthetemperature,incubationtimeandamountofmicrovesicles.38DiI‐C16isweaklyfluorescentinwater,buthighlyfluorescentandphotostablewhenincorporatedintomembranes.Onceappliedtocells,thedyediffuseslaterallywithintheplasmamembrane.Figure4BshowstheinternalizationofDil‐C16‐labelledEVsintotheTHP‐1cellmembrane.WealsoobservedDil‐C16‐labelledEVsinthecytoplasmofTHP‐1cells.Figure4CshowedthatTHP‐1cellstreatedwithEVsbearingFITC‐taggedmiR‐155werefluorescentlylabelledandobservedunderfluorescencemicroscopy.TheresultsclearlydemonstratedthatmiR‐155canbesecretedanddeliveredfromHUVECstoTHP‐1cellsviaEVs.

Figure4OpeninfigureviewerPowerPointHUVECscommunicationwithTHP1throughEVstodelivermiR‐155.A,HUVECsweregrownandinfectedwitheGFPcDNAorrAAVadenovirusfor48handthensubculturedtheseEVswithTHP1inaTranswellsystemfor24h,THP1in24‐wellplateswerestainedwitheGFPdye(green)andDAPI(blue),andthenreviewedandphotographedunderafluorescentmicroscope;B,IsolatedHUVECswerelabelledwithDil‐C16immunofluorescencedye,andthenisolatedEVsfromtheseHUVECstococulturedthemwithTHP1intheTranswellsystem,Dil‐C16(red)andDAPI(blue)expressioninTHP1weredetected;C,HUVECsweregrownandinfectedwithadenoviruscarrying‐FITCormiR‐155‐FITCfor6h,andthenisolatedEVsfromtheseHUVECstococulturedthemwithTHP1intheTranswellsystem,FITC(green)andDAPI(blue)expressioninTHP1weredetected.Scalebars:10um

Asmonocyte/macrophageactivationandECsdamageassociatedwithendothelial‐derivedEVsproductionarekeyeventsinatheroscleroticplaquedevelopment,weexploredtheseinteractions,aswellasthecapacityofendothelial‐derivedEVstomodulatemonocyte/macrophagedifferentiation.39-42WeaddedpurifiedEC‐EVsisolatedunderdifferentconditionstoPMA‐treatedTHP‐1monocyticcellsfor24hours.First,cellsurvivalofTHP1treatedunderdifferentconditionswastestedbytheMTTassay,asshowninFigureS1,andnosignofanynegativeeffectonTHP1cellstestedinthecoculturesystemwasobservedafterexposuretoadifferentnumberofEVsunderallconditions(fromHUVECstreatedwithPBS,ox‐LDL,emptyvector,orAd‐KLF2).Then,RT‐PCRwasperformedtoquantifythelevelsofmiR‐155inTHP1cellstreatedwithEC‐EVsisolatedfromdifferentconditionsandtocomparetheseeffectstothoseofuntreatedcells.AsshowninFigure5A,BandD,miR‐155waselevatedinrecipientTHP‐1cellstreatedwithox‐LDL‐inducedEC‐EVs,whereasnosignificantdifferencewasobservedinTHP‐1cellstreatedwithEVsfromemptyvectorrAdVorKLF2‐cDNA‐transducedHUVECscomparedwithEVsfromPBStreatedHUVECs.However,miR‐155wasreducedinrecipientTHP‐1cellsinfectedwithox‐LD‐inducedEC‐EVsplusKLF2‐cDNArAdVwhencomparedwithox‐LDL‐inducedEC‐EVsplusemptyvectorrAdV.Mostimportantly,ox‐LDL‐inducedEC‐EVssignificantlyenhancedthetypicalM1macrophagesmarkers(CD80andCD86)andactivationofproinflammatorycytokines/chemokines(TNF‐ɑandIL‐6),whiletheysimultaneouslydecreasedthetypicalM2markerCD206oranti‐inflammatoryresponsesofIL‐10inTHP1cells.Bycontrast,THP1cellscoculturedwithEVsfromKLF2‐transducedHUVECswithox‐LDLdisplayedareducedexpressionoftheM1markers(CD80andCD86)andsuppressionoftheircapacitytoproduceproinflammatorycytokines/chemokines(TNF‐ɑandIL‐6),butdisplayincreasedexpressionofthetypicalM2markerCD206andproductionoftheanti‐inflammatorycytokinesIL‐10(Figure5C).

Figure5OpeninfigureviewerPowerPointEffectsofmiR‐155‐containingEVsontheactivationofTHP1.HUVECswereinfectedwithadenovirusescarryingKLF2cDNAoremptyvectorfor48handthentreatedwithorwithout50μg/mLox‐LDLfor6handsubjectedtoisolateEVsandsubculturedtheseEVswithPMA‐treatedTHP1for24h,andTHP1in24‐wellplatesweresubjectedtoanalysistheexpressionofCD80,CD86andCD206byqRT‐PCR(A)orFACS(B);C,HUVECswereinfectedwithadenovirusescarryingKLF2cDNAoremptyvectorfor48handthentreatedwithorwithout50μg/mLox‐LDLfor6handsubjectedtoisolateEVsandthensubculturedtheseEVswithPMA‐treatedTHP1infor24h,andTHP1in24‐wellplatesweresubjectedtoanalysistheexpressionofTNF‐α,IL‐6andIL‐10byELISA;Mean±CV;D,thegraphshowsthequantitativedataofB.Mean±SEM,*indicatesvsemptyvector;blankindicatesHUVECstreatedwithPBS.emptyvectorindicatesAd‐EV.KLF2cDNAindicatesAd‐KLF2,extracellularvesiclesindicateEVs.*P.05;**P.01;***P.001
3.5TheeffectsofEVsfromKLF2‐transducedECssignallingonthemacrophagephenotypeinatherosclerosisformation

TodeterminewhetherEVsthatarereleasedbyKLF2‐overexpressingECscanalsomodulatethemacrophagephenotypeinatherosclerosisformation,wefirstculturedmousemacrophageswithEVsisolatedfrommouseECstreatedunderdifferentconditions.AsshowninFigureS2,theexpressiontrendsofmiR‐155,CD80,CD86andCD206examinedbyreal‐timePCRwereconsistentwiththatofTHP1cellstreatedwithHUVECs‐EVs.Second,weinjectedEVsisolatedfromdifferentpreparationsofmiceECsintravenouslytwiceweeklytoApoE−/−micefedahigh‐fatdiet(HFD)overaperiodof8weeks.After8weeksonaHFD,therewerenodifferencesobservedinthebodyweight,totalcholesterol,low‐densitylipoprotein,high‐densitylipoproteinandtriglyceridelevelsbetweenApoE−/−miceinjectedwithEVsisolatedfromdifferentpreparationsofmiceECs(Table1).Micewerethensacrificedandtheiraorticarterieswereharvestedforanalysis.RT‐PCRanalysisshowedthattherewerenodifferencesintheKLF2expressioninHFD‐fedApoE−/−miceinjectedwithdifferentpreparationsofEVs,butKLF2expressionwassignificantlylowerthanthatinstandarddiet‐fedmice(Figure6A).Moreover,miR‐155expressionwasincreasedinHFD‐fedApoE−/−micecomparedwiththatinstandarddiet‐fedmice,andalso,remarkablyincreasedmiR‐155expressionwasobservedinmiceinjectedwithEVstransfectedwithmiR‐155(HFD+EVs[miR‐155]group)comparedwithemptyvector‐transducedcells(HFD+EVs[emptyvector]group).However,asignificantreductionofexpressionofmiR‐155wasfoundinHFD‐fedApoE−/−micethatreceivedEVsisolatedfromKLF2cDNAandmiR‐155transfectedECs(HFD+EVs[miR‐155+Ad‐KLF2]group)comparedwiththatinHFD+EVs(miR‐155)group(Figure6B),but,therewasnodifferenceinmiR‐155expressionbetweenHFD‐fedApoE−/−micethatreceivedEVsisolatedfromKLF2cDNAtransfectedECsHFD+EVs(Ad‐KLF2)groupandmicefromtheHFD+EVs(miR‐155+Ad‐KLF2)group.Analysesofatherosclerosislesionformationrevealedanincreasedlesionareainthedescendingthoracic,abdominalaortaandaorticsinusfrommiceinjectedwithEVspreparedfromtheHFD+EVs(miR‐155)groupwhencomparedwiththatfrommiceinjectedwithEVspreparedfromtheHFD+EVs(emptyvector)group.AtheroscleroticlesionswerereducedintheHFD+EVs(Ad‐KLF2)groupcomparedwithfrommiceinjectedwithEVspreparedfromtheHFD+EVs(emptyvector)group.However,thedifferenceofatherosclerosislesionsbetweenheHFD+EVs(Ad‐KLF2)groupandHFD+EVs(miR‐155+Ad‐KLF2)grouphasnostatisticalsignificance.Nevertheless,EVspreparedfromtheHFD+EVs(miR‐155+Ad‐KLF2)groupabrogatedtheatheropromotiveeffectofthevesiclesfrommiR‐155overexpressingcellsbydecreasingatherosclerosislesionformationinthedescendingthoracicandabdominalaorta(Figure6C‐F).

AllvaluesvsStandarddiet,P=NS,Mean±SD,n=8‐11miceeachgroup.

Figure6OpeninfigureviewerPowerPointEffectsofmiR‐155‐containingEVsonregulationofatherosclerosisinApoE−/−mice.ApoE‐/‐mice(6weeksold)werefedahigh‐fatdietoveraperiodof8weeksandinjectedwithPBS,EVsfromemptyvector‐ormiR‐155‐transducedmouseECs,orEVsfromKLF2‐transducedcellsthatweretreatedwithorwithoutmiR‐155twiceaweek.MicewerethenkilledandfattylesionsintheaortawerestainedenfacewithOilRedOandquantified.ExpressionofKLF2A,miR‐155(B)inaorticarterieswasdetectedbyq‐PCR.LesionareasshownwerequantifiedusingOilredO(ORO)stainingofthethoracoabdominalaorta(C).DatarepresentORO‐stainingareasasapercentageofaortaareas(D).LesionareasshownwerequantifiedasareasbetweenthelumenandtunicamediaonORO‐stainedaorticsinussections(E).ThegraphisthequantitativedataoftheoilredstainingofE(F).Datashownaremean±SEM(n=8‐11micepergroup).EVsfromKLF2‐transducedmouseECsindicateEVs(Ad‐KLF2).EVsfromKLF2‐transducedandthentreatedwithmiR‐155mouseECsindicateEVs(Ad‐KLF2+miR‐155).*P.05;**P.01;***P.001;#P.05

TofurtherinvestigatewhetherEVsgeneratedfromKLF2‐transducedECsinfluencethebalanceofthemacrophagephenotypebetweenproinflammatorymacrophagesM1andanti‐inflammationmacrophagesM2,thedistributionofM1macrophagesandM2macrophagesinaortatissuesectionsfromthesemicewasassayedbydouble‐immunolabellingforF4/80antigen(red)andCD206antigen(green).F4/80isamacrophage‐specificmarkerandCD206isexpressedbyM2macrophages.35TheobservationsshowedthatF4/80+/CD206‐M1macrophagesandF4/80+/CD206+M2macrophageswerecodistributedinaortatissuesfromApoE−/−mice(Figure7A).However,HFD‐fedApoE−/−miceexhibitedhighermacrophageinfiltrationinaortatissuescomparedwithstandarddiet‐fedmice(Figure7B).Amongthesemacrophages,thepercentageofM1(F4/80+/CD206‐)inaortatissuesfromApoE−/−miceplusinjectionofEVsfromtheHFD+EVs(miR‐155)groupwasmuchhigherthanthatplusinjectionofEVsfromtheHFD+EVs(emptyvector)group(P.05)(Figure7B),whilethepercentageofM1decreasedwhenmicewereinjectedwithEVsfromtheHFD+EVs(Ad‐KLF2)group.However,EVspreparedfromtheHFD+EVs(miR‐155+Ad‐KLF2)groupabrogatedthehighpercentageofM1intheHFD+EVs(miR‐155)group.TheproportionofM2(F4/80+/CD206+)increasedinaortatissuesfromtheHFD+EVs(Ad‐KLF2)groupanddecreasedfromtheHFD+EVs(miR‐155)groupcomparedwiththatfromtheHFD+EVs(emptyvector)group(P.05)(Figure7C),whiletheproportionofM2inaortatissuesfromtheHFD+EVs(miR‐155+Ad‐KLF2)groupwasmuchhigherthanthatfromtheHFD+EVs(miR‐155)group(P.001)(Figure7C).

Figure7OpeninfigureviewerPowerPointEffectsofmiR‐155‐containingEVsonmodulatingmacrophagephenotypes.Aorticarteriesfromeachgroupofexperimentalmicewerecollected.Frozensectionsofaorticrootwerestainedforanti‐F4/80(red),whichbindsallmacrophagephenotypes;anti‐CD206(green),whichbindsM2macrophagesandsatellitecells,andDAPI(blue),whichbindsDNAtoshownthepositionofnuclei(A).M1macrophagesareF4/80+/CD206‐(red)(A‐enlarge).M2macrophagesareF4/80+/CD206+(orange).ThegraphBandCisthequantitativedataofimmunofluorescentstainingofA.Datashownaremean±SEM(n=8‐11micepergroup).*P.05;**P.01;***P.001;#P.05

ThedysfunctionofECsandrecruitmentofinfiltratingmonocytes/macrophagesorchestratecomplexsignallingandcommunicationintheinitiationanddevelopmentofatherosclerosisthatisgenerallyachievedbydirectcell‐cellcontactortransferofsecretedparacrinemolecules.2-4Here,wepresentevidencethatathirdmechanisminvolvestheexchangeofEVs‐secretedinflammation‐associatedmiR‐155betweenECsandmonocytes/macrophagesthatisreleasedunderbasalorox‐LDL‐mediatedstressconditionswithendothelialoverexpressionofKLF2invitroandvivo,whichprovidescompellingevidencethatmiR‐155‐containingEVsregulateinflammationinassociationwithatherosclerosis.Inconclusion,thepresentstudyshowsthat(1)boththeatherosclerosisinducerox‐LDLandatheroprotectivefactorKLF2areabletoregulatetheexpressionofmiR‐155inHUVECs;thismicroRNAisenrichedinox‐LDL‐inducedEC‐EVsandthentransferredtoTHP1,inwhichtheyindeedactivatemonocyteactivationbyshiftingthemonocyte/macrophagebalancefromanti‐inflammatoryM2macrophagestowardsproinflammatoryM1macrophages;EC‐EVsfromKLF2‐expressingHUVECsthathavepotentanti‐inflammatoryactivitiessuppressedmonocyteactivationbyenhancingtheimmunomodulatoryresponsesanddiminishingtheproinflammatoryresponses.(2)Furthermore,EVsfromKLF2‐transducedECsreduceatheroscleroticlesionformationbyinducingdecreasedlevelsofproinflammatoryM1macrophagesandincreasedlevelsofanti‐inflammatoryM2macrophagesinanexperimentalmousemodel,whichisatleastpartlyduetothedecreasedexpressionofinflammation‐associatedmiR‐155.

WhetherEVscanexertproinflammatoryand/oranti‐inflammatoryeffectsremainsamatterofdebate.Endothelialdysfunctionandactivationoccurinvariouschroniccardiovasculardiseasestates,whichlikelycontributetoalterationsinthelevelsandtypesofEVsincirculation,aswellasmodulationoftheircontents,ultimatelyinfluencingtheirfunction.Forexample,microparticlesecretionisenhancedfromECsandleukocytesexposedtoinflammatorymediators,andthesevesicleshavebeenshowntopromotemonocyteactivation,thrombosisandthedysfunction/activationofECsduringatherogenesis.43Interestingly,wefoundthatEVsisolatedfromox‐LDLactivatedECslosesomeoftheiranti‐inflammatoryproperties,suggestingthatanalteredEVsfunctionmaybeanearlyeventinvascularinflammatorydisease.KLF2playsacentralroleinmediatingtheatheroprotectiveendothelialphenotypegeneratedbyshearstress.30Consistentwitharecentlypublishedstudy,44theresultsofthepresentstudyalsosuggestthatKLF2‐expressingcellssecreteEVs,whichreducethemagnitudeofproinflammatoryresponsesandshiftmonocyteactivationtowardsanimmunomodulatoryresponse.Insummary,thepresentfindingsprovidenovelinsightsintothepathophysiologicaleffectsofEVsandtheirinfluenceontargetcellsdependingontheconditioninwhichtheyarereleased.

MiR‐155,atypicalmultifunctionalmicroRNA,playsimportantrolesinimmunityandinflammation,particularlyintheinflammatoryresponsesofmacrophages,implyingthatthismoleculemayalsobeinvolvedinatherogenesis.44TheKLF2transcriptionfactormodulatesmicroRNAexpressioninseveralcelltypes.Forexample,KLF2bindstothepromoterofthemiR‐143/145geneclustertoupregulatetheexpressionofvascularprotectivegenesinECs.21LingrelJBandcolleaguesrecentlyobservedreducedexpressionofmiR‐124aandmiR‐150inmacrophagesfrommyeKlf2−/−mice,indicatingthatKLF2directlymediatestheexpressionofthese2miRNAsinmacrophages.40OurpreviousstudyalsodemonstratedthatKLF2‐mediatedsuppressionofmicroRNA‐155reducestheproinflammatoryactivationofmacrophages.45ThedatapresentedheredemonstratethatmiR‐155isenrichedinox‐LDL‐inducedEC‐EVsandthentransferredtoTHP1,inwhichthesevesiclescanenhancemonocyteactivationbyshiftingthemonocyte/macrophagebalancefromanti‐inflammatoryM2macrophagestowardsproinflammatoryM1macrophages;endothelium‐derivedEVsfromKLF2‐expressingHUVECssuppressedmonocyteactivationbyenhancingimmunomodulatoryresponsesanddiminishingproinflammatoryresponses,andthiseffectwasaccompaniedbyreducedlevelsofmiR‐155.Furthermore,vesiclesfromKLF2‐transducedECs,butnotfrommockcontrols,reducedatheroscleroticlesionformationbyinducingdecreasedlevelsofproinflammatoryM1macrophagesandincreasedlevelsofanti‐inflammatoryM2macrophagesinanexperimentalmousemodel,whichwasatleastpartlyduetothedecreasingexpressionofinflammation‐associatedmiR‐155,confirmingtheinvitrofindingsofthepresentstudy.Thus,thepresentstudyextendsourunderstandingtheroleofKLF2inprotectingagainstatherosclerosisbyillustratingthatKLF2inhibitsinflammation‐associatedmiR‐155inECs,whichprovidescriticalsignalsfortheselectivegrowthanddifferentiationofmacrophagestowardsananti‐inflammatoryM2‐likephenotypethroughanMV‐mediatedmechanism.

ThisworkwasfinanciallysupportedthroughgrantsfromtheNationalKeyR&DProgramofChina(2017YFA0208000)toDr.ShaolinHe,andtheNationalNaturalScienceFoundationofChinato(No.81172781)Dr.MingLi,(No.81500338)Dr.ShaolinHeand(No.81370406)Dr.DazhuLi.

Pleasenote:Thepublisherisnotresponsibleforthecontentorfunctionalityofanysupportinginformationsuppliedbytheauthors.Anyqueries(otherthanmissingcontent)shouldbedirectedtothecorrespondingauthorforthearticle.

1HanssonGK.Inflammation,atherosclerosis,andcoronaryarterydisease.NEnglJMed.2005;352:1685‐1695.
3BinderCJ,ChangMK,ShawPX,etal.Innateandacquiredimmunityinatherogenesis.NatMed.2002;8:1218‐1226.
4MooreKJ,TabasI.Macrophagesinthepathogenesisofatherosclerosis.Cell.2011;145:341‐355.
5KassiteridiC,MonacoC.Macrophagesanddendriticcells:theusualsuspectsinatherogenesis.CurrDrugTargets.2015;16:373‐382.
6SchoberA,WeberC.Mechanismsofmonocyterecruitmentinvascularrepairafterinjury.AntioxidRedoxSignal.2005;7:1249‐1257.
7WeberC,NoelsH.Atherosclerosis:currentpathogenesisandtherapeuticoptions.NatMed.2011;17:1410‐1422.
8CamussiG,DeregibusMC,BrunoS,CantaluppiV,BianconeL.Exosomes/microvesiclesasamechanismofcell‐to‐cellcommunication.KidneyInt.2010;78:838‐848.
9RatajczakJ,WysoczynskiM,HayekF,Janowska‐WieczorekA,RatajczakMZ.Membrane‐derivedmicrovesicles:importantandunderappreciatedmediatorsofcell‐to‐cellcommunication.Leukemia.2006;20:1487‐1495.
10DeJongOG,VanBalkomBW,SchiffelersRM,BoutenCV,VerhaarMC.Extracellularvesicles:potentialrolesinregenerativemedicine.FrontImmunol.2014;5:608.
11LoyerX,VionAC,TedguiA,BoulangerCM.Microvesiclesascell‐cellmessengersincardiovasculardiseases.CircRes.2014Jan17;114:345‐353.
12RaposoG,StoorvogelW.EVs:exosomes,microvesicles,andfriends.JCellBiol.2013;200:373‐383.
13LawsonC,VicencioJM,YellonDM,DavidsonSM.Microvesiclesandexosomes:newplayersinmetabolicandcardiovasculardisease.JEndocrinol.2016;228:R57‐R71.
14KucharzewskaP,BeltingM.Emergingrolesofextracellularvesiclesintheadaptiveresponseoftumourcellstomicroenvironmentalstress.JExtracellVesicles.2013;2:3402.
15YamamotoS,AzumaE,MuramatsuM,HamashimaT,IshiiY,SasaharaM.Significanceofextracellularvesicles:pathobiologicalrolesindisease.CellStructFunct.2016;41:137‐143.
16HoyerFF,GiesenMK,FrancaCN,LutjohannD,NickenigG,WernerN.Monocyticmicroparticlespromoteatherogenesisbymodulatinginflammatorycellsinmice.JCellMolMed.2012;16:2777‐2788.WileyOnlineLibrary
17ChistiakovDA,OrekhovAN,BobryshevYV.Extracellularvesiclesandatheroscleroticdisease.CellMolLifeSci.2015;72:2697‐2708.
18LiuML,WilliamsKJ.Microvesicles:potentialmarkersandmediatorsofendothelialdysfunction.CurrOpinEndocrinolDiabetesObes.2012;19:121‐127.
19PirilloA,NorataGD,CatapanoAL.LOX‐1,OxLDL,andatherosclerosis.MediatorsInflamm.2013;2013:152786.
20AndersonHC,MulhallD,GarimellaR.Roleofextracellularmembranevesiclesinthepathogenesisofvariousdiseases,includingcancer,renaldiseases,atherosclerosis,andarthritis.LabInvest.2010;90:1549‐1557.
21AtkinsGB,JainMK.RoleofKrüppel‐liketranscriptionfactorsinendothelialbiology.CircRes.2007Jun22;100:1686‐1695.
22HergenreiderE,HeydtS,TréguerK,etal.AtheroprotectivecommunicationbetweenendothelialcellsandsmoothmusclecellsthroughmiRNAs.NatCellBiol.2012;14:249‐256.
23ZerneckeA,BidzhekovK,NoelsH,etal.DeliveryofmicroRNA‐126byapoptoticbodiesinducesCXCL12‐dependentvascularprotection.SciSignal.2009;2:ra81.
24HarrisTA,YamakuchiM,FerlitoM,MendellJT,LowensteinCJ.MicroRNA‐126regulatesendothelialexpressionofvascularcelladhesionmolecule1.ProcNatlAcadSciUSA.2008;105:1516‐1521.
25GordonS,TaylorPR.Monocyteandmacrophageheterogeneity.NatRevImmunol.2005;5:953‐964.
26MantovaniA,GarlandaC,LocatiM.Macrophagediversityandpolarizationinatherosclerosis:aquestionofbalance.ArteriosclerThrombVascBiol.2009;29:1419‐1423.
27MosserDM,EdwardsJP.Exploringthefullspectrumofmacrophageactivation.NatRevImmunol.2008;8:958‐969.
28CourtiesG,HeidtT,SebasM,etal.InvivosilencingofthetranscriptionfactorIRF5reprogramsthemacrophagephenotypeandimprovesinfarcthealing.JAmCollCardiol.2014;63:1556‐1566.
29Cardilo‐ReisL,GruberS,SchreierSM,etal.Interleukin‐13protectsfromatherosclerosisandmodulatesplaquecompositionbyskewingthemacrophagephenotype.EMBOMolMed.2012;4:1072‐1086.WileyOnlineLibrary
30SalagianniM,GalaniIE,LundbergAM,etal.Toll‐likereceptor7protectsfromatherosclerosisbyconstraining“inflammatory”macrophageactivation.Circulation.2012;126:952‐962.
31HeS,LiM,MaX,LinJ,LiD.CD4+CD25+Foxp3+regulatoryTcellsprotecttheproinflammatoryactivationofhumanumbilicalveinendothelialcells.ArteriosclerThrombVascBiol.2010;30:2621‐2630.
32IsmailN,WangY,DakhlallahD,etal.MacrophagemicrovesiclesinducemacrophagedifferentiationandmiR‐223transfer.Blood.2013;121:984‐995.
33WangR,DingQ,YaqoobU,etal.Exosomeadherenceandinternalizationbyhepaticstellatecellstriggerssphingosine1‐phosphate‐dependentmigration.JBiolChem.2015;290:30684‐30696.
34NatarelliL,SchoberA.Janus‐facedroleofKrüppel‐likefactor2‐dependentregulationofmicroRNAsinendothelialproliferation.ArteriosclerThrombVascBiol.2014;34:1605‐1606.
35HeS,YangL,LiD,LiM.Kruppel‐likefactor2‐mediatedsuppressionofmicroRNA‐155reducestheproinflammatoryactivationofmacrophages.PLoSONE.2015;10:e0139060.
36TaganovKD,BoldinMP,ChangKJ,BaltimoreD.NF‐kappaB‐dependentinductionofmicroRNAmiR‐146,aninhibitortargetedtosignalingproteinsofinnateimmuneresponses.ProcNatlAcadSci.2006;103:12481‐12486.
37SchulteLN,WestermannAJ,VogelJ.DifferentialactivationandfunctionalspecializationofmiR‐146andmiR‐155ininnateimmunesensing.NucleicAcidsRes.2013;41:542‐553.
38ZhangY,LiuD,ChenX,etal.SecretedmonocyticmiR‐150enhancestargetedendothelialcellmigration.MolCell.2010;39:133‐144.
39SteinM,KeshavS,HarrisN,GordonS.Interleukin4potentlyenhancesmurinemacrophagemannosereceptoractivity:amarkerofalternativeimmunologicmacrophageactivation.JExpMed.1992;176:287‐292.
40ManoharanP,BasfordJE,Pilcher‐RobertsR,NeumannJ,HuiDY,LingrelJB.ReducedlevelsofmicroRNAsmiR‐124aandmiR‐150areassociatedwithincreasedproinflammatorymediatorexpressioninKrüppel‐likefactor2(KLF2)‐deficientmacrophages.JBiolChem.2014;289:31638‐31646.
41MooreKJ,SheedyFJ,FisherEA.Macrophagesinatherosclerosis:adynamicbalance.NatRevImmunol.2013;13:709‐721.
42DiehlP,FrickeA,SanderL,etal.EVs:majortransportvehiclesfordistinctmicroRNAsincirculation.CardiovascRes.2012;93:633‐644.
43KarpmanD,StåhlAL,ArvidssonI,etal.Complementinteractionswithbloodcells,endothelialcellsandmicrovesiclesinthromboticandinflammatoryconditions.AdvExpMedBiol.2015;865:19‐42.
44NjockMS,ChengHS,DangLT,etal.EndothelialcellssuppressmonocyteactivationthroughsecretionofextracellularvesiclescontainingantiinflammatorymicroRNAs.Blood.2015;125:3202‐3212.
45Nazari‐JahantighM,WeiY,NoelsH,etal.MicroRNA‐155promotesatherosclerosisbyrepressingBcl6inmacrophages.JClinInvest.2012;122:4190‐4202.

Thefulltextofthisarticlehostedatiucr.orgisunavailableduetotechnicaldifficulties.

Pleasecheckyouremailforinstructionsonresettingyourpassword.Ifyoudonotreceiveanemailwithin10minutes,youremailaddressmaynotberegistered,andyoumayneedtocreateanewWileyOnlineLibraryaccount.

Can"tsignin?Forgotyourusername?

Enteryouremailaddressbelowandwewillsendyouyourusername

Iftheaddressmatchesanexistingaccountyouwillreceiveanemailwithinstructionstoretrieveyourusername