RESEARCHOpenAccess - PDF document

RESEARCHOpenAccess Plasminogenincerebrospinalfluidoriginates fromcirculatingblood AnnaMezzapesa 1,2 ,CyrilleOrset 2 ,LaurentPlawinski 3 ,LoicDoeuvre 2 ,SaraMartinezdeLizarrondo 2 , GuglielminaChimienti 1 ,DenisVivien 2 ,AlexandreMansour 2 ,SabrinaMatà 4 ,GabriellaPepe 1 5* Abstract Background: Plasminogenactivationisaubiquitoussourceoffibrinolyticandproteolyticactivity.Besidesitsrolein preventionofthrombosis,plasminogenisinvolvedininflammatoryreactionsinthecentralnervoussystem. Plasminogenhasbeendetectedinthecerebrospinalfluid(CSF)ofpatientswithinflammatorydiseases;however,its originremainscontroversial,astheblood – CSFbarriermayrestrictitsdiffusionfromblood. Methods: WeinvestigatedtheoriginofplasminogeninCSFusingAlexaFluor488 – labelledratplasminogen injectedintoratswithsystemicinflammationandblood – CSFbarrierdysfunctionprovokedbylipopolysaccharide (LPS).Near-infraredfluorescenceimagingandimmunohistochemistryfluorescencemicroscopywereusedtoidentify plasminogeninbrainstructures,itsconcentrationandfunctionalityweredeterminedbyWesternblottinganda chromogenicsubstrateassay,respectively.Inparallel,plasminogenwasinvestigatedinCSFfrompatientswith Guillain-Barrésyndrome( n =15),multiplesclerosis( n =19)andnoninflammatoryneurologicaldiseases( n =8). Results: EndogenousratplasminogenwasdetectedinhigheramountsintheCSFandurineofLPS-treatedanimals ascomparedtocontrols.InLPS-primedrats,circulatingAlexaFluor488 – localizedinthechoroidplexus,CSFandurine.PlasminogeninhumanCSFwashigherinGuillain-Barrésyndrome (median=1.28ng/
l(interquartilerange(IQR)=0.66to1.59))ascomparedtomultiplesclerosis(median=0.3ng/
l (IQR=0.16to0.61))andtononinflammatoryneurologicaldiseases(median=0.27ng/
l(IQR=0.18to0.35)). Conclusions: OurfindingsdemonstratethatplasminogenistransportedfromcirculatingbloodintotheCSFofrats viathechoroidplexusduringinflammation.OurdatasuggestthatasimilarmechanismmayexplainthehighCSF concentrationsofplasminogendetectedinpatientswithinflammation-derivedCSFbarrierimpairment. Keywords: Blood – cerebrospinalfluidbarrier,Inflammation,LPS,Plasminogen Introduction Vesselwallfibrinolyticactivityandpericellularproteoly- sisintissuesrequiresplasminogenbindingandtrans- formationintoplasminatthesurfaceoffibrin,cellsor theextracellularmatrixbyeitherthetissue-typeplas- minogenactivator(tPA)ortheurokinase-typeplasmino- genactivator(uPA)[1].Thebindingofplasminogento cellsorfibrinisalysine-dependentmechanismthatcan beinhibitedbylysineanaloguessuchastranexamicacid bloodclotdissolution[2],whereasintissues,plasminis responsibleforpericellularproteolysisaccompanying cellmigration,angiogenesis,woundhealing , tissuere- modellingandinflammation[3-5] . Pericellularproteoly- sisinthecentralnervoussystem(CNS)isinvolvedin development,regenerationofnervoustissues,neuronal andsynapticplasticityandtheinflammatoryresponse [6-9].Inhumans,plasminogendeficiencyisassociated withneuraldisordersincludingcongenitalhydroceph- alus,periventricularnodularheterotopiasandDandy-Walker malformation[10].Traceamountsofplasminogenhavebeen reportedinnormalhumancerebro spinalfluid(CSF),whereas raisedconcentrationshavebeendetectedinpatientswith *Correspondence: eduardo.angles-cano@inserm.fr 5 InsermUMRS1140,FacultyofPharmaceuticalandBiologicalSciences,Paris DescartesUniversity,4Avenuedel ’ Observatoire,75270Paris,cedex06, Fulllistofauthorinformationisavailableattheendofthearticle JOURNAL OF NEUROINFLAMMATION ©2014Mezzapesaetal.;licenseeBioMedCentralLtd.ThisisanOpenAccessarticledistributedunderthetermsofthe CreativeCommonsAttributionLicense(http://creativecommons.org/licenses/by/4.0),whichpermitsunrestricteduse, distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycredited.TheCreativeCommonsPublic DomainDedicationwaiver(http://creativecommons.org/publicdomain/zero/1.0/)appliestothedatamadeavailableinthis article,unlessotherwisestated. Mezzapesa etal.JournalofNeuroinflammation 2014, 11 :154 http://www.jneuroinflammation.com/content/11/1/154 meningitis[11],subarachnoidhaemorrhage[12]andmultiplesclerosis[13]However,thequestionhasnotbeensettledastowhetherplasminogeninCSForiginatesfromcirculatingbloodorisexpressedintheCNS,assuggestedbythepres-enceofmRNAinmousebrain[14,15]orbyitssynthesisbyratmicroglialcellsinculture[16].TheCNSisprotectedfromthebloodstreambythebrainbarrier(BBB)andthebloodCSFbarrier[17,18].TheBBBisthebarrierseparatingthebraininterstitialfluidcompartmentfromthegeneralcircula-tion(endothelialtightjunctions,basallaminaofendo-thelialcellsandastrocytepedicles).Itselectivelylimitspenetrationofavarietyofnoxioussubstancesandsup-pliesthebrainwithnutrients.ThebloodCSFbarrierseparatestheCSFcompartmentfrombloodthroughchoroidplexusepithelialcells,tightjunctions,abasalmembraneandtheendothelium.ItrestrictsthepassageofharmfulsubstancesfromthebloodintotheCSFse-cretedacrossthechoroidplexusepithelialcellsintothesventricularsystem[18].BecausethebloodbarrierismorepermeablethantheBBB,manyplasmaproteinsenterthecerebrospinalliquid(throughpinocyt-osisoractivetransport).AnimpairmentofthebloodbarrierthusleadstoanincreaseintheconcentrationofproteinsintheCSF.BecausediffusionofplasmaproteinsintotheCNSisaselectiveprocess,wesoughttodemonstratethatcircu-latingplasminogenmaycrossthebloodCSFbarrierandcanbedetectedintheCSF.Forthatpurpose,westudiedcontrolratsandratswithlipopolysaccharide(LPS)-inducedsystemicinflammation.OurresultsshowthatcirculatingplasminogenenterstheCSFspacedur-ingbloodCSFbarrierdysfunctioninducedbysystemicinflammation,thussuggestingthatplasminogenfoundintheCSFofpatientswithinflammatoryneurologicaldisordersoriginatesfromcirculatingblood.MaterialsandmethodsReagentsandproteinsTheenhancedchemiluminescencereagentkitwasobtainedfromBio-RadLaboratories(Hercules,CA,USA).Thechromogenicsubstrateselectiveforplasmin(methylmalonyl)-hydroxyprolylarginine-para-nitroaniline(CBS0065)waspurchasedfromStago(Asnières-sur-Seine,France).EscherichiacoliLPSserotype0111:B4,trans-4-(aminomethyl)cyclohexane-1-carboxylicacid(ortranexamicacid,TXA),amilorideand4,6-diamidino-2-phenylindole(DAPI)andEvansbluewereobtainedfromSigmaChemicalCo(StLouis,MO,USA).AlexaFluor488succinimidylesterwaspurchasedfromLifeTechnologies(Carlsbad,CA,USA).Goatanti-collagentypeIV(ColIV)usedforimmunohistochemicalanalyseswasobtainedfromSouthernBiotech(Birmingham,AL,USA),anddonkeyanti-goatantibodyF(ab)2fragmentslinkedtotetramethylrhodamineisothiocyanatewerepurchasedfromJacksonImmunoResearch(WestGrove,PA,USA).Glu-plasminogen,plasminandperoxidase-conjugatedmonoclonalantibodyagainstplasminogenkringle1(CPL15-PO)werepreparedandcharacterizedasdescribedpreviously[19-21].Arabbitanti-mouseplasminogenpolyclonalantibodywaskindlyprovidedbyHRLijnen(UniversityofLeuven,Belgium).PatientsamplingAtotalof42patientswithneurologicaldiseasesattend-ingtheNeurologicalClinicoftheCareggiUniversityHospital,Florence,Italy,wereadmittedinthisstudy.In-formedconsentwasobtainedaccordingtotheDeclar-ationofHelsinki.TheCareggiUniversityHospitalReviewBoardapprovedtheprotocol.Diagnoseswerebasedonclinical,laboratoryandmagneticresonanceimagingdataaccordingtotheInternationalClassificationofDiseases,TenthRevision,andtheDiagnosticandStatisticalMan-ualofMentalDisorders,FourthEditionTextRevision[22].PatientswithParkinsonsdisease,braintumour,epilepsyandalcoholorothersubstancedependencewereexcluded.AllCSFsampleswereobtainedbylum-barpunctureandwereimmediatelyusedforroutinelaboratoryanalyses,whichincludederythrocyteandleucocytedifferentialcellcounts,totalproteinconcen-tration,albuminandimmunoglobulinG(IgG)levelsandagaroseisoelectricfocusingforIgGoligoclonalbands.CSFsampleswithincreasedlymphocytecounts�(5mm)orwithbloodcontamination(erythrocyteconcentration�50cells/mm)wereexcluded.TheCSFtoserumalbuminquotient(Qalb)wascalculatedandusedtoevaluatebloodCSFbarrierintegrity.Qalb�0.007wasconsideredamarkerofbloodCSFbarrierdysfunction[23,24].Theremainderofeachsamplewasstoredinaliquotsat80°Cforfurtheranalysis.AnimalexperimentalmodelExperimentswereperformedusingmaleWistarrats(280to350g)inaccordancewiththedirectivesoftheCounciloftheEuropeanCommunities(86/609/EEC)andtheFrenchAgricultureandForestryMinistryforhandlinganimals(decree87-848).PurificationandlabellingofratplasminogenTomonitorplasminogenuptakebycellsandtissuesinvivointheratmodelofinflammation,weusedfluorescence-labellednativeplasminogenpurifiedfromratplasma.Bloodwasobtainedbycardiacpunctureofanaesthetizedanimalsandcollectedin0.129Msodiumcitrate.Theplasmawasseparatedfrombloodbycentri-fugation,andplasminogenwasisolatedbylysineaffinityandsievingchromatographyasdescribedpreviously[19].Thepurifiedplasminogenhadarelativemolecularetal.JournalofNeuroinflammation:154Page2of10http://www.jneuroinflammation.com/content/11/1/154 weight(M=92,000Da)similartothatofhumanplas-minogenandwaslabelledwithAlexaFluor488dye(A488-Pg)accordingtotheinstructionsofthemanufacturer.RatmodelofsystemicinflammationRatswereassignedtoeitherLPStreatment(L)orthesa-linecontrolgroup(C)(=3ratspergroup(Figure1).GroupsL1,L2andL3receivedanintraperitonealinjec-tionofLPS(1mg/kg)dilutedinsalinesolution,andcontrolgroupsC1,C2andC3receivedonlythesalinesolution.ThisdosageofLPShasalreadybeendemon-stratedtobeabletoinducebraininflammationintherat[25].GroupsL1andC1receivednoothertreatment.GroupsL2andC2receivedatailveininjectionofA488-Pg(1mg)18hoursafterLPSorsalineadministration.GroupL3receivedatailveininjectionof10mg/kgTXA30minutesafterLPSadministration,followedbyasub-cutaneousdoseof100mg/kgTXA.After18hours,eachratreceivedatailveininjectionofA488-Pg(1mg)pre-incubatedfor15minuteswith0.1MTXA,followedbyasubcutaneousinjectionof100mg/kgTXA.GroupC3receivedidenticaldosesofTXA,butnotofA488-Pg.ThisdosagescheduleofTXAwasnecessarytoensureneutralizationoflysine-bindingsitesinboththeen-dogenousandinjectedplasminogen.Attheendofthescheduledtreatments(24hours),sampleswerecollectedandanimalskilledforfurtherstudies.SamplecollectionTheanaesthetizedratswereplacedinastereotaxicframeandsecuredwithearbars,andamidlineincisionintheskinwasmadeuptotheheadareatopermiteasyaccesstothecisternamagna.Theneedle,whichwasconnectedtoadrawsyringe,wasinsertedhorizontallyandcentrallyintothecisternamagnaforCSFcollection.TheCSFsamplewasslowlydrawnintothesyringe,andthecolouroftheCSFwascloselyobservedtoavoidanypossiblebloodcontamination.Approximately100lofCSFwascollectedfromeachanimal.Afterashortcen-trifugationstep(3minutesat5,000,4°C)allCSFsam-pleswereimmediatelyflash-frozeninliquidnitrogenandstoredat80°Cuntiluse.Bloodwasdrawnbycardiacpunctureusinga1.2×mmgaugeneedlesyringecontaininga0.1volumeof0.129Msodiumcitrate.Urinewascollectedbypunctur-ingthebladder.Plasmawasseparatedfrombloodbycen-trifugationat1,500for15minutes.Allthesampleswerekeptoniceduringcollectionandkeptfrozenat80°Cuntiluse.EvansbluepermeabilityassayRats(=3pergroup)treatedwithsaline(groupC1)orLPS(groupL1)aloneorsupplementedwithTXA(groupsC3andL3)receivedatailveininjectionofEv-ansbluedye(4%insaline)4hoursbeforebeingkilled.Aftertranscardiacperfusionwithheparinizedsaline,thebrainsandkidneyswereharvestedandplungedintocoldsalineuntilfluorescencemeasurementswereperformed.Evansbluedyepermeabilitywasmonitoredbyexvivonear-infraredfluorescence(NIRF)imagingusingtheIVIS200imagingsystem(CaliperLifeSciences,Hopkinton,MA,USA). TimeLPS challenged pSaline challengedpi.v. & s.c. TXA i.v. A488-Pg i.v. A488-Pg* + s.c TXAi.v. & s.c. TXA i.v. A488-Pg s.c. TXA Sampling of CSF, plasma and urineSampling of CSF, plasma and urine Figure1Diagramofratinflammatorymodelandstudydesign.Ratswereseparatedinsixgroups(threeratspergroup).GroupsL1,L2andL3werechallengedwith1mg/kglipopolysaccharide(LPS),andthecontrolgroups(C1,C2andC3)werechallengedwithsaline.BothLPSandsalinewereinjectedintraperitoneally.*Preincubatedwithtranexamicacid(TXA);i.v.,IntravenousinjectionofTXA(10mg/kg)orplasminogenlabelledwithAlexaFluor488dye(A488-Pg)(1mg);s.c.,SubcutaneousinjectionofTXA(100mg/kg).etal.JournalofNeuroinflammation:154Page3of10http://www.jneuroinflammation.com/content/11/1/154 FluorescencemicroscopyAftertheratswerekilled,partsofthebrainswerepost-fixedwith4%paraformaldehydein0.1Mphosphate-bufferedsaline,pH7.4,at4°Cfor18hours,followedby24hoursin20%sucroseandthenfrozeninisopentane.Forimmunocytochemistry,cryostatbrainsections(8werecollectedonpolylysine-coatedslidesandincubatedovernightwithagoatanti-ColIVprimaryantibody(1:1,500),whichwasrevealedwithdonkeyanti-goatantibodyF(ab)2fragmentslinkedtotetramethylrho-damineisothiocyanate(1:500).Thecellswerethencoun-terstainedwithDAPI,mountedwithFluoprep(Dako,Glostrup,Denmark)andobservedunderanepifluores-cencemicroscope.ImagesweredigitallycapturedwithaLeicaDM6000microscope-coupledCoolSnapcam-era(LeicaMicrosystems,Wetzlar,Germany)andvisual-izedwithMETAVUE5.0software(MolecularDevices,Sunnyvale,CA,USA).Thedetectionofexogenouslyad-ministeredA488-Pginthosesampleswasalsostudied.DetectionoffunctionalplasminogenincerebrospinalTransformationofCSFplasminogenintoplasminwasmeasuredusinguPA(5IU/ml)andaplasmin-selectivechromogenicsubstrate(0.75mMCBS0065),asprevi-ouslydescribed[26].Inthissystem,theinitialvelocityof-nitroanilinereleasedfromCBS0065isproportionaltotheamountofplasmin(ogen).TheabilityofTXAtoim-pairplasminogenbindingtothecellmembranewasin-vestigatedasdescribedpreviously[27].WesternblotanalysisHumanCSFsamples(7goftotalprotein)andratCSF,urineandplasma(diluted1:50)samples(10l)wereelectrophoresedin8%SDS-PAGEgelundernonreduc-ingconditions.Proteinsweretransferredontoapolyvi-nylidenefluoridemembrane,andplasminogenwasrevealedusingamonoclonalantibodydirectedagainsthumanplasminogen(CPL15-PO,humanCSFsamples)orarabbitanti-mouseplasminogenpolyclonalantibody(ratsamples).TheamountofplasminogenwasexpressedasthenumberofpixelsdirectlydetectedusingImage-QuantTL7.0imageanalysissoftwarewiththeImage-QuantLAS4000imagingsystem(GEHealthcareLifeSciences,Pittsburgh,PA,USA)andexpressedasnano-gramspermicrolitrebyreferencetoaknownamountofplasminogen(15ng,10l)electrophoresedundersimilarconditions.Inratsamples,theamountofplasminogenwasexpressedasthenumberofpixelsdirectlydetectedusingtheimageanalysissoftware.StatisticalanalysisDataarerepresentativeofatleastthreeindependentex-perimentsandareexpressedasmedian(25thto75thinterquartilerange(IQR)).TheMannWhitneywasusedtocomparevaluesobtainedintreatedversuscontrolrats.Forhumanexperiments,one-wayanalysisofvariance(ANOVA)wasperformedtocomparethepa-tientgroups.Bonferronismultiple-comparisonsposttestwasrunforthepairwisecomparisonofgroups.Statis-ticalsignificancewassetat0.05.Aspecificstatisticalpackageforexactnonparametricinference(StataStatis-ticalSoftwarerelease9(2005);StataCorp,CollegeSta-tion,TX,USA)wasused.Lipopolysaccharide-inducedincreaseinplasminogeninratcerebrospinalfluidEndogenousplasminogenwasdetectedbyWesternim-munoblottinginplasma,CSFandurineofbothLPS-andsaline-treatedcontrolrats(groupsL1andC1,respectively)(Figure2A).TheLPSinjectionproducednosignificantvariationintheamountofcirculatingplasminogenascomparedtocontrols(0.1-foldincreaseininjectedversuscontrolrats)(Figure2B).Incontrast,wedetectedhigherplasminogenlevelsinCSFandurinefromLPS-treatedrats(fourfoldincreaseinCSF,0.0002;twofoldincreaseinurine,=0.0071(bothbyMannWhitneytest))(Figure2B).RatCSFplasminogenwasefficientlytrans-formedintoplasmin;itsactiveconcentrationappearedtobedirectlyrelatedtotheamountofplasminogendetectedinCSFsamplesbyWesternimmunoblotting=0.0045,MannWhitneytest)(Figure2C).Lipopolysaccharide-inducedincreaseinbarrierpermeabilityTheaforementioneddatasuggestthatvascularperme-abilitywasmodifiedfollowingtheLPSinjection.ThishypothesiswasinvestigatedbytestingpermeabilitytoEvansbluedye.ExvivofluorescentandNIRFimagesshowedhigherbrainfluorescenceintheLPS-treatedrats(groupL)ascomparedtocontrols(groupC),indicatinganincreaseinbarrierpermeabilitytothedyeduetoCSFbarrierimpairment(Figure3A).Modifica-tionsintheglomerularfiltrationbarrierwerealsode-tectedwiththeEvansbluetest(Figure3B).HigherkidneyfluorescencewasobservedinLPS-treatedrats(groupL)ascomparedtocontrols(groupC).Thead-ministrationofTXA(groupsC3andL3)waswithoutef-fectonbarrierpermeability.InvivoevidenceofthecirculatingoriginofplasminogenincerebrospinalfluidToobtaininvivoevidenceoftheextravasationofplas-minogenfromblood,animalswerechallengedwithex-ogenousA488-Pg18hoursaftertheLPS(groupL2)orsalineinjection(groupC2).A488-PgappearedintheCSFandurineofallLPS-treatedrats,indicatingthatetal.JournalofNeuroinflammation:154Page4of10http://www.jneuroinflammation.com/content/11/1/154 Figure2 (Seelegendonnextpage.) Mezzapesa etal.JournalofNeuroinflammation 2014, 11 :154Page5of10 http://www.jneuroinflammation.com/content/11/1/154 bloodplasminogencrossesthebarrierduringitspres- enceinthecirculation(withahalf-lifeofabout2days) (Figure4A)andintegratestheCSFexchangedatleast threetimesdaily[28].Actually,thepresenceofcirculat- ingA488-Pgwasclearlyvisualizedbyfluorescencemi- croscopyonlyinthechoroidplexusofLPS-treatedrats (Figure4B).Figure4Cisamagnifiedviewofthechoroid plexusshowingthepresenceofA488-Pgandthevessel walldetectedbyimmunocytochemistryofColIV.Of note,A488-Pgisabsentinthesaline-treatedcontrol. Typically,bindingofplasminogenbycellsismediated byalysine-dependentmechanismvialysineresidues encompassedincellularreceptorsandlysine-binding siteslocatedinplasminogenkringledomains[1].To explorewhetherthepassageofplasminogenthroughthe blood – CSFbarrierimplicatesalysine-dependentmech- anism,wemeasuredcellularuptakeofA488-Pginthe presenceofthelysineanalogueTXA.Figure5shows thatplasminogenremainedincreasedintheCSFofrats treatedwithTXA,thusexcludinguptakeviaknown plasminogenreceptorsharbouringcarboxy-terminally- sineresidues[1]. Tosubstantiatetherelevanceofthese invivo ratstudies tohumanpathology,weinvestigatedthepresenceofplas- minogenintheCSFofpatientswithinflammatoryand noninflammatorydisorders.Plasminogenantigenwasde- tectedinallCSFhumansamples(Figure6).Highercon- centrationsofplasminogenwerefoundinGuillain-Barré (Seefigureonpreviouspage.) Figure2 Determinationofplasminogeninratplasma,cerebrospinalfluidandurine. Sampleswereobtainedfromratschallengedwith saline(groupC1)versusratschallengedwithlipopolysaccharide(LPS)(groupL1)(seeflowchartinFigure1). (A) Representativeimmunoblot obtainedforplasminogendetectionusingarabbitantibodytomouseplasminogen.Electrophoresisofreferenceratplasminogen(Pg)inan equalvolume(10
l)ofplasmadiluted1:50,cerebrospinalfluid(CSF)andurine. (B) Box-and-whiskerplotofplasminogeninplasma,CSFandurine afterimmunoblotanddensitometricanalyses.Resultsrepresentmedian(25thto75thinterquartilerange(IQR)).* P =0.0002and§ P =0.0071in CSFandurine(Mann – Whitney U test),respectively. (C) Box-and-whiskerplotofmeasurementoftheactivationofplasminogeninCSFsamples (chromogenicsubstrateassay).Resultsrepresentmedian(IQR)ofthevelocityofplasminformation.# P =0.0045(Mann – Whitney U test).mOD:milli opticaldensity. Figure3 Blood – cerebrospinalfluidbarrierandglomerularbarrierpermeabilitytoEvansbluedye. Exvivo fluorescent(F)andnear-infrared fluorescent(NIRF)imagesofbrainsandkidneysfromratsinjectedwithEvansbluedye24hoursafterlipopolysaccharide(LPS)treatmentaccording totheoutlineinFigure1.Themeasurementwasperformedinperfusedbrains24hoursafterEvansbluedyeinjection.C1,Saline-challengedrats;L1, LPS-challengedrats;C3,Saline/tranexamicacid(TXA);L3,LPS /TXA;Negcontrol,RatwithoutEvansbluedye;BF,Brightfield. (A) BrainNIRF images. (B) KidneyNIRFimages.Thehighfluorescencein tensityobservedinLPS-treatedrats(L1andL3)comparedtosaline-treatedrats(C1 andC3)indicatesleakageoftheblood – cerebrospinalfluidbarrier(A)ortheglomerularfiltrationbarrier(B).Nodifferenceswereobserved betweensaline-treated(C1)andTXA-treated(C3)animals,in dicatingnoeffectofTXAonextravasationofEvansbluedye. Mezzapesa etal.JournalofNeuroinflammation 2014, 11 :154Page6of10 http://www.jneuroinflammation.com/content/11/1/154 syndrome(GBS)(median=1.28ng/
l(25thto75thinter- quartilerange(IQR)=0.66to1.59), n =15)ascomparedto multiplesclerosis(MS)(median=0.3ng/
l(IQR=0.16to 0.61), n =19)andnoninflammatoryneurologicaldiseases (NINDs)(median=0.27ng/
l(IQR=0.18to0.35), n =8) ( P 0.0001,one-wayANOVA).Thepairwisecomparison yieldedsignificantdifferencesinGBSversusMS( P 0.05) andGBSversusNINDs( P 0.05),bothanalysedwith Bonferroni ’ sposttest.GBSpatientshadsignificantly highervaluesofCSFtotalprotein,albuminandIgG,as wellashigherQalb,whereasMSpatientshadhigher indexvaluesofIgG(QIgG/Qalb)(seeAdditionalfile1: TableS1). Discussion Plasminogenissynthesizedmainlybytheliverandthen isdistributedtotissuesviathesystemiccirculation[29]. Theconcentrationofplasminogenincirculatingblood isrelativelyhigh(1.5to2
M)andconstant,asplas- minogendoesnotbehaveasanacutephasereactant suchasfibrinogen.Therefore,inflammatoryconditions ortheinjectionofaninflammatoryagentsuchasLPS donotincreasethecirculatingconcentrationofplas- minogen.However,amechanisticlinkbetweeninflam- mationandtheblood – CSFbarrierdysfunctionhasbeen established[30],whichmayexplainanincreasedtransfer ofplasminogenfromcirculatingbloodtoCSF.Indeed, traceamountsofplasminogenhavebeenfoundinthe CSFofpatientswithnoninflammatorydiseases[31,32].By usinganLPS-inducedmodelofsystemicinflammationin rats,wefoundthatplasminogenispresentatsignificantly higherconcentrationsintheCSFofLPS-treatedratsas comparedtocontrols.Accordingly,LPS-treatedratshad increasedblood – CSFbarrierpermeability,asdemon- stratedwithEvansbluedye.CSFplasminogenisefficiently transformedintoplasmin,thussuggestingthattransferof plasminogenfromcirculatingbloodtotheCSFcouldbea sourceforplasminformationinthenervoussystem.Fur- thermore,usingA488-Pg,wefoundthatthislabelledex- ogenousratplasminogencrossestheblood – CSFbarrier. Thehalf-lifeofplasminogeninthecirculationisabout2 days,whereastheCSFisproducedandexchangedbycells ofthechoroidplexusatleastthreetimesdaily[28].Thus, circulatingA488-PgcouldbedetectedintheCSFandthe Figure4 Invivo evidenceofthecirculatingoriginofplasminogenincerebrospinalfluid. Sampleswereobtainedfromratsinjectedwith plasminogenlabelledwithAlexaFluor488dye(A488-Pg).GroupL2ratswerechallengedwithlipopolysaccharide(LPS),andgroupC2was challengedwithsaline(seeflowchartinFigure1). (A) .Anequalvolume(10
l)ofeitherplasmadiluted1:50orofcerebrospinalfluid(CSF)or urinewaselectrophoresed,andfluorescence(F)inthegelwasdirectlyrevealedusingImageQuantTL7.0imageanalysissoftware(upperpanel). ThegelwasthentransferredontoapolyvinylidenefluoridemembraneanddetectedbyWesternblottingwitharabbitantibodytomouse plasminogen(WB,lowerpanel).Representativesamplesareshown. (B) MicrographshowingthepresenceofcirculatingA 488-Pg(indicatedby arrows)inthechoroidplexusofanLPS-treatedratsdetectedbydirectfluorescencemicroscopy.4  ,6-diamidino-2-phenylindole(DAPI)staining (blue)indicatescellnuclei. (C) Magnifiedimagesofchoroidplexusofsaline-andLPS-treatedratsshowingthepresenceofcirculatingA488-Pg (indicatedbyyellowarrows )onlyintheLPScondition,asdetectedbydirectfluorescencemicroscopy.DAPIstaining(blue)indicatescell nuclei,andcollagentypeIV(ColIV,red)isusedasavesselmarker. Mezzapesa etal.JournalofNeuroinflammation 2014, 11 :154Page7of10 http://www.jneuroinflammation.com/content/11/1/154 choroidplexusofLPS-treatedrats.These invivo datasug- gestthatcirculatingplasminogenenterstheCSFatthe choroidplexus.Thisenormousepithelial/endothelialsur- faceareaispotentiallyavailableforuntowardleakageof plasmaproteinsintoCSF[33,34].BecauseLPSisalso knowntoinduceinterferencewiththeintegrityofthe glomerularfiltrationbarrier[35],wealsoinvestigatedthe passageofplasminogenthroughtherenalglomerulus.We detectedmodificationsintheglomerularfiltrationbarrier withtheEvansbluetestandhighconcentrationsofboth endogenousplasminogenandA488-Pgintheurineof LPS-treatedratsascomparedtotraceamountsfoundin controlrats. Bindingofplasminogentocarboxy-terminallysineresi- duesofmembranereceptorsisawell-knownmechanism governingitscaptureandconcentrationontothecellsur- face,whereitcanbeeithertransformedintoplasminor internalized[1].Becausethesereceptorscouldpotentially beinvolvedinthetransferofplasminogenfrombloodto CSF,wetestedthepossibilitythatTXA,alysineanalogue, mightpreventplasminogenfromenteringtheCSF. Occupationofthelysine-bi ndingsiteofplasminogen byTXApreventsbindingtolysineresiduesofmem- braneglycoproteinreceptors.However,TXAdidnot inhibitA488-Pgtransferthroughtheblood – CSFbar- rier,thussuggestingthatotherbindingsitesmightbe involvedinplasminogenuptake.Forinstance,apep- tidederivedfromtheplasminogen-bindingsitedomain 1ofM6PIGF2R,peptidefragment18to36,induces plasminogeninternalization[36]. Figure5 Effectoftranexamicacidonplasminogentransfer throughtheblood – cerebrospinalfluidbarrierandtherenal glomerulus. Ratswithlipopolysaccharide(LPS)-inducedsystemic inflammationreceivedanintravenousinjectionofplasminogen labelledwithAlexaFluor488dye(A488-Pg)alone(groupL2)orin thepresenceofsustainedconcentrationsoftranexamicacid(TXA) (groupL3)(seeflowchartinFigure1).Plasminogenwasdetectedby immunoblotting. (A) Representativeimmunoblotofratplasminogen (Pg),plasmadiluted1:50cerebrospinalfluid(CSF)andurinesamples usingarabbitantibodytomouseplasminogen.Experimentswere performedasindicatedinFigure2withanequalvolume(10
l)of eachsample. (B) Representativecolumnbarsofplasminogenin plasma,CSFandurineafterimmunoblotanddensitometricanalysis. Figure6 Quantificationofplasminogeninhuman cerebrospinalfluid. Cerebrospinalfluid(CSF)sampleswere collectedfrompatientswithGuillain-Barrésyndrome(GBS, n =15), multiplesclerosis(MS, n =19)andnoninflammatoryneurological disease(NIND, n =8). (A) RepresentativehumanCSFimmunoblot showingperoxidase-conjugated monoclonalantibodydirected againstplasminogenkringle1(CPL15-PO).Electrophoresisofhuman plasminogen(Pg)(M r =92,000Da),10
lofhumanplasmadiluted1:50 and7
goftotalproteininCSF. (B) Box-and-whiskerplotofhuman CSFplasminogenafterimmunoblotanalysisanddensitometricanalysis. Resultsrepresentmedian(interquartilerange)( P 0.0001,one-way analysisofvariance).* P 0.05forGBSversusMSand§ P 0.05forGBS versusNIND(Bonferroni ’ smultiple-comparisonsposttest). Mezzapesa etal.JournalofNeuroinflammation 2014, 11 :154Page8of10 http://www.jneuroinflammation.com/content/11/1/154 Collectively,thesefindingsdemonstratethat,during inflammation,plasminogenistransportedfromcirculat- ingbloodintotheCSFofratsviathechoroidplexus. Interestingly,thehighplasminogenCSFcontentde- tectedinLPS-treatedversuscontrolratsparallelsthe significantlyhigherconcentrationofplasminogenfound intheCSFofpatientswithblood – CSFbarrierdysfunction determinedonthebasisofahighQalb.Inagreementwith previouslypublisheddata[31,32],wefoundthatplas- minogenispresentintraceamountsinnoninflammatory CSFsamples.However,wefoundsignificantlyhighercon- centrationsintheCSFofpatientswithblood – CSFbarrier dysfunction.Ourpresentdataareinaccordwithourpre- viousdemonstrationofthepresenceofapolipoprotein(a), aglycoproteinhomologoustoplasminogenintheCSFof patientswithblood – CSFbarrierdysfunction[37]. Althoughtherearelimitationstotranslatingresults obtainedinexperimentalanimalstohumandiseases, ourdataaresuggestiveofasimilarmechanismforplas- minogentransferfrombloodtoCSFandmaybeofrele- vancetopatientswithinflammation-derivedCSFbarrier impairment. Conclusions Wehavedemonstratedthatplasminogenisincreasedin theCSFofpatientswithadysfunctionalblood – CSFbarrier. Wehavereproduced,inan invivo ratmodelofsystemic inflammation,similarmodificationsofbarrierfunction,and wealsohavedemonstratedthatplasminogenfromcirculat- ingbloodenterstheCSFspace.Altogetherourdatastrongly suggestthattheincreaseinCSFplasminogenconcentrations detectedinpatientsobeysamechanismbywhichcirculating plasminogencrosses thealteredblood – CSFbarrier. Additionalfile Additionalfile1:TableS1. StudygroupsandCSFbiochemical parameters. Abbreviations A488-Pg: AlexaFluor488plasminogen;CBS0065:(methylmalonyl)- hydroxyprolylarginine- para -nitroaniline;CNS:Centralnervoussystem; CSF:Cerebrospinalfluid;DAPI:4  ,6-diamidino-2-phenylindole;GBS:Guillain- Barrésyndrome;MS:Multiplesclerosis;NIND:Noninflammatoryneurological disease;NIRF:Near-infraredfluorescence;TXA:Tranexamicacid; uPA:Urokinase-typeplasminogenactivator. Competinginterests Theauthorsdeclarethattheyhavenocompetinginterests. Authors ’ contributions AnaMezzapesaperformedtheresearch;collected,analysedandinterpreted data;andparticipatedinmanuscriptdrafting.GPandGCanalysedand interpreteddataandparticipatedinmanuscriptdrafting.SMrecruitedpatients, collectedcerebrospinalfluid,analysedandinterpreteddataandparticipatedin manuscriptdrafting.COandSMperformedanimalexperimentsandanalysed data.DVrevisedandcommentedonthemanuscriptandmanagedfunding.LP, LDandAlexandreMansourperformedchromogenicandimmunoblotassays andparticipatedintheinterpretationanddiscussionoftheresults.EAC conceivedanddesignedthestudy,analysedandinterpretedthedata,and wrotethemanuscript.Allauthorsreadandapprovedthefinalmanuscript. Acknowledgments WeacknowledgefinancialsupportfromADISUPuglia,Consorzio InteruniversitàperleBiotecnologie,Italy,andSocietàItalianadiBiochimicae BiologiaMolecolareforthemobilitytoAM.ThisstudywasfundedbyInserm (NationalInstitutesforHealthandBiomedicalResearch,France)andthe LowerNormandyRegionalCouncil. Authordetails 1 DepartmentofBiosciences,Biotechnologies,andBiopharmaceutics, UniversityofBari,ViaAmendola165/A,70125Bari,Italy. 2 InsermU919,GIP Cyceron,BP5229,BoulevardHenriBecquerel,14074Caencedex,Caen, France. 3 CNRSUMR5248CBMN,InstitutEuropéendeChimieetBiologie,2 rueRobertEscarpit,33607Pessac,France. 4 DepartmentofNeurology, UniversityofFlorence,CareggiHospital,VialeMorgagni85,50134Florence, Italy. 5 InsermUMRS1140,FacultyofPharmaceuticalandBiologicalSciences, ParisDescartesUniversity,4Avenuedel ’ Observatoire,75270Paris,cedex06, France. Received:28March2014Accepted:14August2014 References 1.DasR,PluskotaE,PlowEF: Plasminogenanditsreceptorsasregulatorsof cardiovascularinflammatoryresponses. TrendsCardiovascMed 2010, 20: 120 – 124. 2.RijkenDC,LijnenHR: Newinsightsintothemolecularmechanismsofthe fibrinolyticsystem. JThrombHaemost 2009, 7: 4 – 13. 3.PlowEF,Hoover-PlowJ: Thefunctionsofplasminogenincardiovascular disease. TrendsCardiovascMed 2004, 14: 180 – 186. 4.LiWY,ChongSS,HuangEY,TuanTL: Plasminogenactivator/plasminsystem: amajorplayerinwoundhealing? WoundRepairRegen 2003, 11: 239 – 247. 5.Garcia-TouchardA,HenryTD,SangiorgiG,SpagnoliLG,MaurielloA, ConoverC,SchwartzRS: Extracellularproteasesinatherosclerosisand restenosis. ArteriosclerThrombVascBiol 2005, 25: 1119 – 1127. 6.SappinoAP,MadaniR,HuarteJ,BelinD,KissJZ,WohlwendA,VassalliJD: Extracellularproteolysisintheadultmurinebrain. JClinInvest 1993, 92: 679 – 685. 7.ZhangY,PothakosK,TsirkaSA: Extracellularproteases:biologicaland behavioralrolesinthemammaliancentralnervoussystem. CurrTopDev Biol 2005, 66: 161 – 188. 8.MonardD: Cell-derivedproteasesandproteaseinhibitorsasregulatorsof neuriteoutgrowth. TrendsNeurosci 1988, 11: 541 – 544. 9.Gur-WahnonD,MizrachiT,Maaravi-PintoFY,LourbopoulosA,GrigoriadisN, HigaziAA,BrennerT: Theplasminogenactivatorsystem:involvementin centralnervoussysteminflammationandapotentialsitefortherapeutic intervention. JNeuroinflammation 2013, 10: 124. 10.KlammtJ,KobeltL,AktasD,DurakI,GokbugetA,HughesQ,IrkecM, KurtulusI,LapiE,MechoulamH,Mendoza-LondonoR,PalumboJS,SteitzerH, TabbaraKF,OzbekZ,PucciN,SotomayorT,SturmM,DrogiesT,ZieglerM, SchusterV: Identificationofthreenovelplasminogen(PLG)genemutations inaseriesof23patientswithlowPLGactivity. ThrombHaemost 2011, 105: 454 – 460. 11.NewmanRL: Amethodfordetectingandestimatingplasminogenin cerebrospinalfluid. JClinPathol 1964, 17: 313 – 315. 12.VermeulenM,vanVlietHH,LindsayKW,HijdraA,vanGijnJ: Sourceof fibrin/fibrinogendegradationproductsintheCSFaftersubarachnoid hemorrhage. JNeurosurg 1985, 63: 573 – 577. 13.HammackBN,FungKY,HunsuckerSW,DuncanMW,BurgoonMP,OwensGP, GildenDH: Proteomicanalysisofmultiplesclerosiscerebrospinalfluid. Mult Scler 2004, 10: 245 – 260. 14.BashamME,SeedsNW: Plasminogenexpressionintheneonataland adultmousebrain. JNeurochem 2001, 77: 318 – 325. 15.ZhangL,SeiffertD,FowlerBJ,JenkinsGR,ThinnesTC,LoskutoffDJ,ParmerRJ, MilesLA: Plasminogenhasabroadextrahepaticdistribution. Thromb Haemost 2002, 87: 493 – 501. 16.NakajimaK,TsuzakiN,NagataK,TakemotoN,KohsakaS: Productionand secretionofplasminogeninculturedratbrainmicroglia. FEBSLett 1992, 308: 179 – 182. Mezzapesa etal.JournalofNeuroinflammation 2014, 11 :154Page9of10 http://www.jneuroinflammation.com/content/11/1/154 17.WolburgH,PaulusW: Choroidplexus:biologyandpathology. Acta Neuropathol 2010, 119: 75 – 88. 18.SaundersNR,EkCJ,HabgoodMD,DziegielewskaKM: Barriersinthebrain: arenaissance? TrendsNeurosci 2008, 31: 279 – 286. 19.FleuryV,Anglés-CanoE: Characterizationofthebindingofplasminogen tofibrinsurfaces:theroleofcarboxy-terminallysines. Biochemistry 1991, 30: 7630 – 7638. 20.WimanB,WallénP: Activationofhumanplasminogenbyaninsoluble derivativeofurokinase:structuralchangesofplasminogeninthecourse ofactivationtoplasminanddemonstrationofapossibleintermediate compound. EurJBiochem 1973, 36: 25 – 31. 21.MontesR,PáramoJA,A-CanoE,RochaE: Developmentandclinical applicationofanewELISAassaytodetermineplasmin-
2-antiplasmin complexesinplasma. BrJHaematol 1996, 92: 979 – 985. 22.AmericanPsychiatricAssociation: DiagnosticandStatisticalManualofMental Disorders ,Volumetextrevision(DSM-IV-TR).4thedition.Washington,DC: Author;2000. 23.TibblingG,LinkH,OhmanS: PrinciplesofalbuminandIgGanalysesin neurologicaldisorders.I.Establishmentofreferencevalues. ScandJClin LabInvest 1977, 37: 385 – 390. 24.BrettschneiderJ,ClausA,KassubekJ,TumaniH: Isolatedblood – cerebrospinal fluidbarrierdysfunction:prevalenceandassociateddiseases. JNeurol 2005, 252: 1067 – 1073. 25.JeongHK,JouI,JoeEH: SystemicLPSadministrationinducesbrain inflammationbutnotdopaminergicneuronaldeathinthesubstantia nigra. ExpMolMed 2010, 42: 823 – 832. 26.BangertK,ThorsenS: Assayoffunctionalplasminogeninratplasma applicabletoexperimentalstudiesofthrombolysis. ThrombHaemost 2000, 84: 299 – 306. 27.Ho-Tin-NoéB,EnslenH,DoeuvreL,CorsiJM,LijnenHR,Anglés-CanoE: Role ofplasminogenactivationinneuronalorganizationandsurvival. MolCell Neurosci 2009, 42: 288 – 295. 28.BrownPD,DaviesSL,SpeakeT,MillarID: Molecularmechanismsof cerebrospinalfluidproduction. Neuroscience 2004, 129: 957 – 970. 29.BohmfalkJF,FullerGM: Plasminogenissynthesizedbyprimaryculturesof rathepatocytes. Science 1980, 209: 408 – 410. 30.BegleyDJ,CouraudPO,GreenwoodJ,PanW,PerryVH,PersidskyY, RansohoffR,SaundersNR,InternationalBrainBarriersSociety: Inflammation andthebrainbarriers. Availableathttp://www.ibbsoc.org/Reports.htm (accessed28August2014). 31.WuKK,JacobsenCD,HoakJC: Plasminogeninnormalandabnormal humancerebrospinalfluid. ArchNeurol 1973, 28: 64 – 66. 32.NewmanRL,StewartGT: Theuseoffibrinolyticactivatorsinmeningitis andsimilarconditions. ArchDisChild 1965, 40: 235 – 242. 33.DamkierHH,BrownPD,PraetoriusJ: Cerebrospinalfluidsecretionbythe choroidplexus. PhysiolRev 2013, 93: 1847 – 1892. 34.JohansonC,StopaE,McMillanP,RothD,FunkJ,KrinkeG: The distributionalnexusofchoroidplexustocerebrospinalfluid,ependyma andbrain:toxicologic/pathologicphenomena,periventricular destabilization,andlesionspread. ToxicolPathol 2011, 39: 186 – 212. 35.XuC,ChangA,HackBK,EadonMT,AlperSL,CunninghamPN: TNF-mediated damagetoglomerularendotheliumisanimportantdeterminantofacute kidneyinjuryinsepsis. KidneyInt 2014, 85: 72 – 81. 36.LeksaV,PfistererK,Ondrovi
ováG,BinderB,Lakato  ováS,DonnerC, SchillerHB,ZwirzitzA,MrvováK,PevalaV,KutejováE,StockingerH: Dissectingmannose6-phosphate-insulin-likegrowthfactor2receptor complexesthatcontrolactivationanduptakeofplasminogenincells. JBiolChem 2012, 287: 22450 – 22462. 37.PepeG,ChimientiG,LiuzziGM,LamanuzziBL,NardulliM,LolliF,Anglés-CanoE, MatàS: Lipoprotein(a)inthecerebrospinalfluidofneurological patientswithblood – cerebrospinalfluidbarrierdysfunction. ClinChem 2006, 52: 2043 – 2048. doi:10.1186/s12974-014-0154-y Citethisarticleas: Mezzapesa etal. : Plasminogenincerebrospinalfluid originatesfromcirculatingblood. JournalofNeuroinflammation 2014 11 :154. Submit your next manuscript to BioMed Central and take full advantage of: € Convenient online submission € Thorough peer review € No space constraints or color “gure charges € Immediate publication on acceptance € Inclusion in PubMed, CAS, Scopus and Google Scholar € Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Mezzapesa etal.JournalofNeuroinflammation 2014, 11 :154Page10of10 http://www.jneuroinflammation.com/content/11/1/154