Altered glial marker expression in autistic post-mortem prefrontal cortex and cerebellum

PAGE 1

RESEARCHOpenAccessAlteredglialmarkerexpressioninautistic post-mortemprefrontalcortexandcerebellumCatherineEdmonson1,2 †,MarkNZiats1,3,4* †andOwenMRennert1AbstractBackground: Thecellularmechanism(s)underlyingautismspectrumdisorders(ASDs)arenotcompletely understood,butASDsarethoughttoultimatelyresultfromdisruptedsynaptogenesis.However,studieshavealso shownthatglialcellnumbersandfunctionareabnormalinpost-mortembraintissuefromautisticpatients.Direct assessmentofglialcellsinpost-mortemhumanbraintissueistechnicallychallenging,limitingglialresearchin humanASDstudies.Therefore,weattemptedtodetermineifglialcell-typespecificmarkersmaybealteredin autisticbraintissueinamannerthatisconsistentwithknowncellularfindings,suchthattheycouldserveasa proxyforglialcellnumbersand/oractivationpatterns. Methods: Weassessedtherelativeexpressionoffiveglial-specificmarkersandtwoneuron-specificmarkersvia qRT-PCR.Westudiedtissuesamplesfromtheprefrontalcortex(PFC)andcerebellumofninepost-mortemautistic brainsamplesandnineneurologically-normalcontrols.Relativefold-changeingeneexpressionwasdetermined usingthe Ctmethodnormalizedtohousekeepinggene -actin,withatwo-tailedStudent ’ s t -test P <0.05 betweengroupsconsideredassignificant. Results: Bothastrocyte-andmicroglial-specificmarkersweresignificantlymorehighlyexpressedinautisticPFCas comparedtomatchedcontrols,whileinthecerebellumonlyastrocytemarkerswereelevatedinautisticsamples.In contrast,neuron-specificmarkersshowedsignificantlylowerexpressioninboththePFCandcerebellumofautistic patientsascomparedtocontrols. Conclusions: Theseresultsareinlinewithpreviousfindingsshowingincreasedglialcellnumbersand up-regulationofglialcellgeneexpressioninautisticpost-mortembraintissue,particularlyinthePFC,aswellas decreasednumberofneuronsinboththePFCandcerebellumofautisticpatients.Theconcordanceofthese resultswithcell-levelstudiesinpost-mortemautisticbraintissuesuggeststhatexpressionofglialcell-type specificmarkersmayserveasausefulalternativetotraditionalcellularcharacterizationmethods,especially whenappropriately-preservedpost-mortemtissueislacking.Additionally,theseresultsdemonstrateabnormal glial-specificgeneexpressioninautisticbrains,supportingpreviousstudiesthathaveobservedalteredglialcell numbersoractivationpatternsinASDs.Futureworkshoulddirectlyassessthecorrelationbetweencell-type specificmarkerlevelsandcellnumberandactivationpatterns. Keywords: Astrocyte,Autisticdisorder,Geneexpression,Glia,Interneuron,Microglia,Neuron *Correspondence: ziatsm@mail.nih.gov†Equalcontributors1LaboratoryofClinicalandDevelopmentalGenomics,NationalInstituteof ChildHealthandHumanDevelopment,NationalInstitutesofHealth,49 ConventDrive,Building49,Room2C078,Bethesda,MD20814,USA3UniversityofCambridge,RobinsonCollege,GrangeRd,CambridgeshireCB3 9AN,UK Fulllistofauthorinformationisavailableattheendofthearticle 2014Edmonsonetal.;licenseeBioMedCentralLtd.ThisisanOpenAccessarticledistributedunderthetermsofthe CreativeCommonsAttributionLicense(http://creativecommons.org/licenses/by/2.0),whichpermitsunrestricteduse, distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycited.TheCreativeCommonsPublic DomainDedicationwaiver(http://creativecommons.org/publicdomain/zero/1.0/)appliestothedatamadeavailableinthis article,unlessotherwisestated.Edmonson etal.MolecularAutism 2014, 5 :3 http://www.molecularautism.com/content/5/1/3

PAGE 2

BackgroundAutismspectrumdisorders(ASDs)areneurodevelopmentalsyndromesdefinedbyimpairmentsinlanguage, verbalandnon-verbalcommunication,andrestrictive/ repetitivepatternsofbehavior[1].ASDsymptomsmanifestwithinthefirsttwoyearsoflife,withtheirseverity andpresentationvaryingconsiderablybetweenindividuals,thusyieldingthe ‘ spectrum ’ classification.ASDsare estimatedtoaffect1in88childrenintheUnitedStates, andtheprevalenceofASDsisatleastfourtimesmore commoninmalesthanfemales[2]. Theetiologyofautismiscomplexandtheneurobiologicalmechanism(s)thatresultintheclinicalphenotyperemaintobefullyunderstood.However,thereis strongevidencethattheautismphenotypeultimately resultsfromaberrantsynapticwiringinthedeveloping brain[3].Inparticular,studieshaveshownthatlongdistancecommunicationbetweendisparateneocortical areasmaybedisruptedinASDs,causingdelaysininformationprocessingwithinthebrainthatmanifestas thecommunication,language,andsocialdevelopment problemsseeninchildrenwithautism[4].Additionally, parallelresearchhasshownthatneuronalmicrocircuitrywithinbrainareasmayalsobedisruptedin ASDs,andthatthismayresultinlocalprocessingdeficitswithinbrainregionsrelatedtohigherfunctioning, suchastheprefrontalcortex(PFC)[5].Underlyingthese circuitdisruptionsisalargebodyofevidencethathas demonstrateddecreasednumbersofneurons(andtheir varioussubtypes)throughouttheautisticbrainbyearly childhoodinpost-mortemstudies[6]. Inadditiontothebodyofevidenceimplicating aberrantlocalandlong-distancesynapticdysfunctionin ASDs,manystudieshavedemonstratedmicroglialand astrocytedysfunctioninASDbrains.Forinstance,postmortempathologicalstudiesofautisticbrainusing immunocytochemistry(IHC)and/orstereologyhave identifiedmicroglialactivationpatterns[7-9],andhave demonstratedincreasedmicroglialcelldensityinmultiplebrainregions[8,10].Glialactivationreferstothe wellcharacterizedcascadeofeventsthatoccursupon reactionofgliatostimuli,resultingintheirmobilization, releaseofcytokines,andabilitytophagocytize,among otherprocesses[11];glialactivationisassociatedwith specificgeneexpressionpatternsthataredistinctfrom ‘ resting ’ glialgeneexpression[12].Furthermore,positron emissiontomography(PET)usingamicroglial-specific radiotraceralsodemonstratedmicroglialactivationin multiplebrainregionsofautisticcases[13].Additionally, studiesinaRettsyndromemousemodel,asingle-gene deletiondisorderwithautismasacomponent,havealso demonstratedcellularmicroglialabnormalities[14],and aremarkablestudydemonstratedthatautistic-likephenotypescanbepartiallyreversedbyreplacingmutant Mecp2( / )microgliawiththeirrespectivewild-type cells[15]. Increasednumbersofastrocytes,withalteredcellsize andbranchingpatterns,havealsobeendemonstratedin post-mortemautisticbrains[16].Additionally,astrocytespecificcellmarkerproteinsareincreasedinmultiple autisticbrainregions[17,18].SimilartomicroglialstudiesinASDmousemodels,astrocyteshavebeenshown tobeabnormalinanumberofsingle-geneASDmodels, includingRett[19,20],FragileX[21],andTuberous Sclerosis[22].Inparalleltotheaforementionedmicroglial study,itwasalsoshownthatreplacingmutantastrocytes inMecp2( / )micecancorrectsomeaspectsofthe phenotype[23]. However,itisnotclearhowtheseseparatelinesof evidence — onedemonstratingimmune/glialdysfunction inASDsandtheotherimplicatingsynapticabnormalities — mayconvergeintoacommonmechanisminthe autisticbrainthatultimatelyresultsinthesharedclinical phenotype.Becauseseparatestudieshaveshownthat microgliaandastrocytesplaycriticalrolesinsculpting developingsynapsesduringnormalneurodevelopment [24,25],itisreasonabletohypothesizethatinherent defectsoraberrantnumbersofmicrogliaandastrocytes inthedevelopingautisticbrainmaybecausativeofthe synapticabnormalitiesbyaffectingtheproperwiringof developingneuronalconnections.However,because appropriately-preservedpost-mortemautisticbraintissueislacking[26],cellular-levelstudiesassessingglial numbersandactivationinhumanautisticbrainshave beenlimited.Moreover,quantificationofcellnumbers inpostmortemtissuebystereologyistechnicallychallenging,furtherlimitingtheabilityofresearchersto assessthefewappropriately-preservedtissuesamples thatareavailable.Finally,nostudieshaveconcurrently specificallyassessedformicroglia,astrocytes,andneurons inthesamesetofautisticbrainsamples.Asaconsequence,acomprehensiveunderstandingoftherelationshipbetweenglialandneuroncellsinautisticbrainsis needed. Therefore,thepurposeofthisstudywastwo-fold. First,wesoughttodetermineifmicroglia,astrocyte,and neuron-specificmarkerswerealteredinpost-mortem autisticbraintissue,inordertofurtherinvestigatethe roleofgliainASDs.Then,wedeterminedifglialand neuronalcell-typespecificmarkerexpressionpatterns areconsistentwithknowncellular-levelfindings,becausegeneexpressionstudiesofpost-mortemhuman brainareofteneasiertoperformthancell-levelstudies, andthereforethisapproachmayserveasavaluable ‘ screening ’ assaytoinferrelativecellproportions. Inthisstudy,wecomparedinternally-normalized mRNAexpressionlevelsofmicroglial,astrocyte,and neuronalcell-typespecificmarkergenesinpost-mortemEdmonson etal.MolecularAutism 2014, 5 :3 Page2of9 http://www.molecularautism.com/content/5/1/3

PAGE 3

braintissuefrompatientswithautismandhealthycontrols.Ourresultsprovidefurtherevidenceforaroleof gliainautismpathology,andsuggestthatassessmentof glialcell-typespecificmarkersmayserveasaproxyfor relativecellularnumbersoractivationpatterns.MethodsPost-mortembrainsamplesPost-mortembraintissuewasobtainedfromtheNational InstituteofChildHealthandHumanDevelopmentBrain andTissueBank,MD,USA.Thissourceobtainedconsent tousebraintissueforresearchfromeachpatientortheir guardianpriortohis/herdeath,andtheirprotocolwas approvedbytheirInstitutionalReviewBoard.Nopatientspecificidentifiableinformationwasobtained.Because multiplebrainregionshavebeenimplicatedinASDs,we performedouranalysisintwoseparateareasthathave beenconsistentlydemonstratedasabnormalinautism — thePFCandthecerebellum.Weobtainedpost-mortem PFCbraintissuefromfiveindividualswithautismand fromfivehealthycontrols(Table1).Weobtainedpostmortemcerebellumbraintissuefromfourindividuals withautismandfourhealthycontrols.Themajorityof samplepairs(PFCandcerebellum)werederivedfromthe samedonorbrain.AllcaseswereCaucasianmales,and casecontrolswerematchedbyageascloselyaspossible.RNAisolationandqualitycontrolRNAisolationandqualitycontrolanalysiswere performedaspreviouslydescribed[27].Briefly,total RNAwasextractedusingTRIZOLReagent(Life Technologies,Carlsbad,CA,USA)accordingtothe manufacturer ’ sprotocol.QuantificationofRNAwas performedusingaNanoDropND-1000,andRNA integritywasassessedusinganAgilentBioanalyzer 2100(Table2).ReversetranscriptasereactionTotalRNA(1 g)wasusedina20 LreversetranscriptasereactiontosynthesizecDNAwithSuperScript3 ReverseTranscriptase(LifeTechnologies)accordingto themanufacturer ’ sprotocol.Briefly,1 goftotalRNA wasaddedtoanaqueoussolutioncontaining250ng/ L ofrandomhexamerand10mMdeoxyribonucleotide triphosphate.TheRNAwasdenaturedfor5minutesat 65Candthensnapcooledonicefor2minutes.After which0.1MDTT,5First-StrandBuffer(250mM Tris – HCl,375mMKCl,15mMMgCl2),RNaseOUT RecombinantRibonucleaseInhibitor(40Units/ L),and SuperScript3ReverseTranscriptase(200Units/ L)were addedintoeachsamplemixture.Thereactionwas carriedoutunderthefollowingconditions:25Cfor5 minutes,50Cfor60minutes,and70Cfor15minutes. Table1ClinicalcharacteristicsandRNAqualityofautisticandcontrolbrainsamplesSample#UMB#DiagnosisBrain area Age(yr)Causeof death PMI(h)RNAquality A260/280A260/230RIN 15308*AutismPFC4.5Skullfracture212.0512.2664.9 21349AutismPFC5.6Drowning392.0442.2324.3 35144AutismPFC7.2Rhabdomyosarcoma32.0582.2715.4 45302*AutismPFC16.3DKA202.0312.2384 54999*AutismPFC20.8Cardiacarrhythmia142.0392.2326 64670*ControlPFC4.6CommotioCordis172.0482.2765.2 71185ControlPFC4.7Drowning172.0262.2434.7 84898ControlPFC7.7Drowning122.0562.1835.9 94848*ControlPFC16.7Drowning152.0442.1856.7 104727*ControlPFC20.5Multipleinjuries(MVA)52.0662.2096.5 115308*AutismCere4.5Skullfracture212.0871.7817.3 124899AutismCere14.3Drowning92.0772.3149.3 135302*AutismCere16.3DKA202.0831.6462.2 144999*AutismCere20.8Cardiacarrhythmia142.0812.1149.2 154670*ControlCere4.6CommotioCordis172.0882.1616.1 164722ControlCere14.5Multipleinjuries(ATV)162.0732.8286.5 174848*ControlCere16.7Drowning152.0872.3276.8 184727*ControlCere20.5MultipleInjuries(MVA)52.0672.3077.2PFC:Prefrontalcortex;Cere:Cerebellum;UMB:UniversityofMarylandBrainBanksamplenumber;PMI:Post-morteminterval;DKA:Diabeticketoacid osis;ATV: All-terrainvehicle;MVA:Motorvehicleaccident;RIN:RNAintegritynumber.*Indicatesbothprefrontalcortexandcerebellumsampleswereassess edfromthe samedonorbrain.Edmonson etal.MolecularAutism 2014, 5 :3 Page3of9 http://www.molecularautism.com/content/5/1/3

PAGE 4

ThecDNAproducedfromthereactionwasdilutedto 0.25withnucleasefreewater.RealtimequantitativePCRSYBRGreenExpressionAssaySystem(AppliedBiosystems,FosterCity,CA,USA)wasusedtomeasurerelative, normalized,mRNAexpressionlevels.Weassessedfour separatemicroglial-specificcellsurfacegenes:Triggering receptorexpressedonmyeloidcells2( TREM2 ), DAP12 CX3Cchemokinereceptor1( CX3CR1 ),andallograft inflammatoryfactor1( AIF1 )[28].Twocelltypespecificintermediatefilaments,glialfibrillaryacidicprotein( GFAP ),whichisastrocyte-specific[29],andthe pan-neuronalcellmarker NEFL [30],wereusedto assessforastrocytesandneurons,respectively.Additionally,weassessedforGABAergicinterneurons specificallywithparvalbumin( PVL )[31].Theintermediatefilamenthousekeepinggenebeta-actin( ACTB )was usedasacontrol.Forwardandreverseprimersequences weregeneratedusingPrimer3softwareandsynthesizedby EurofinsMWGOperon(Huntsville,AL,USA)(Table2). Quantitativereversetranscriptasepolymerasereaction (qRT-PCR)wasperformedusinganABIPrism7900SequenceDetectionSystem(LifeTechnologies)witha 96-wellformat.EachqRT-PCRreactioncontained6.5 L water,12.5 LSYBRGreenmastermix(AppliedBiosystems),1 Lforwardprimer(10 M),1 Lreverseprimer (10 M),and4 LofcDNA(0.25).Datawascollected usingtheSDS2.3Program(AppliedBiosystems)underthe followingrunparameters:48Cfor30minutes,95Cfor 10minutes,40cyclesof95Cfor15seconds,60Cfor1 minute,andafinaldissociationstage.DataanalysisThetargetgenesandtheendogenouscontrolswere measuredwithtechnicaltriplicatesineachqRT-PCR reaction,andallgeneswereassessedinthreeseparate, independentqRT-PCRruns.Thecyclethresholdnumber(Ct)wascalculatedusingRQManager1.2Software (AppliedBiosystems).Relativeexpressionofeachtarget genewasnormalizedto ACTB usingthe Ctmethod. All P valuesreportedarebasedonatwo-tailedStudent ’ s t -test.Onlyresultswitha P valuelessthan0.05were consideredsignificant.ResultsTheaveragepost-morteminterval(PMI)wasnotsignificantlydifferentbetweenautisticandcontroltissue samples(Table1;ASD=17.9h,ctrl=13.2h, P =0.16). Thisremainedtrueaftersub-stratifyingbybrainregion (ASDPFC=19.4h,ctrlPFC=13.2h, P =0.28;andASD cerebellum=16.0h,ctrlcerebellum=13.25h, P =0.48). RNAisolatedfrompost-mortembraintissuewasgenerallyofhighquality,andtheRNAIntegrityNumber (RIN)wasnotsignificantlydifferentbetweenautismand controls(ASD=5.84,ctrl=6.18, P =0.67).TheRINwas alsonotsignificantlydifferentaftersub-stratifyingby brainregion(ASDPFC=4.92,ctrlPFC=5.80, P =0.13; andASDcerebellum=7.00,Ctrlcerebellum=6.65, P =0.85). Table2PrimersusedforqRT-PCRPrimernamePrimersequence(5 to3 )ODMW%GCcontentTm(C) ActinB-FAGAAAATCTGGCACCACACC4.160645060.4 ActinB-RAGAGGCGTACAGGGATAGCA4.36240.15562.4 Trem2-FCCGGCTGCTCATCTTACTCT3.359955562.4 Trem2-RAGTCATAGGGGCAAGACACC4.261605562.4 Dap12-FGAGACCGAGTCGCCTTATCA3.861025562.4 Dap12-RGTCATGATTCGGGCTCATTT3.76114.14558.4 Cx3cr1-FGCAGATCCAGAGGTTCCCTT3.760935562.4 Cx3cr1-RTAACAGGCCTCAGCCAAATC3.960555060.4 Gfap-FCTGCGGCTCGATCAACTCA3.55748.857.962.3 Gfap-RTCCAGCGACTCAATCTTCCTC3.66277.152.462.7 Nefl-FAGCTGGAGGACAAGCAGAAC4.46209.15562.4 Nefl-RTGCCATTTCACTCTTTGTGG3.560654558.4 Parvalbumin-FCTGGAGACAAAGATGGGGAC4.36240.15562.4 Parvalbumin-RCAGAGAGGTGGAAGACCAGG4.46265.16064.5 Aif1-FAGCAGTGATGAGGATCTGCC4.06182.15562.4 Aif1-RAGCATTCGTTTCAGGGACAT3.96132.14558.4F:Forward;R:Reverse;OD:Opticaldensity;MW:Molecularweight;Tm:Meltingtemperature.Edmonson etal.MolecularAutism 2014, 5 :3 Page4of9 http://www.molecularautism.com/content/5/1/3

PAGE 5

InthePFC,quantificationofmicroglialmarkersdemonstratedsignificantlyincreasedexpressioninautistic samplesof TREM2 DAP12 ,and CX3CR1 ,butnot AIF1 (Figure1).Theexpressionof TREM2 washighestofall microglialmarkers,approximately1.75-foldhigherin autismbraintissuethancontrols( P =0.0016).Thelevels of CX3CR1 and DAP12 were1.50-fold( P =0.0092)and 1.34-fold( P =0.0086)higherinautisticsamplesrelative tocontrols,respectively.Similarly,theexpressionof astrocytemarker GFAP wassignificantlyhigherinautisticbrains(1.70-fold, P =0.0049).Conversely,however, boththepan-neuronalmarker NEFL ,andtheGABAergicinterneuron-specificmarker PVA ,weresignificantly lowerinautisticsamplescomparedtocontrols(0.68fold, P =0.0034;and0.52-fold, P =0.0020,respectively). Inpost-mortemcerebellum,theexpressionofastrocytemarker GFAP wasalsosignificantlyhigherinautism samplesthaninhealthycontrols(2.63-fold, P =0.0022; Figure2).Incontrast,theexpressionofmicroglial markers TREM2 DAP12 CX3CR1 ,and AIF1 werelower inautismtissuethanincontroltissue,withfoldchangesof 0.780( P =0.0056),0.797( P =0.0083),0.659( P =0.0029), and0.808( P =0.0052),respectively.Expressionofneuronal markers PVA and NEFL werealsolowerinautismsamples thanincontrolsamples(0.862-fold, P =0.033;and0.798fold, P =0.013,respectively),aswasfoundinthePFC.DiscussionWhiletherehavebeenmultiplestudiesassessingRNA expressionlevelsinautistictissue,herewereportonthe expressionofmicroglial,astrocyte,andneuron-specific cellmarkersconcurrentlyintworegionsofautistic brains.Ourresultsshowthatglial-specificmarkers demonstratealteredexpressioninautisticbrains.Moreover,theexpressionpatternofthesecell-typespecific markersparallelspreviousfindingsofglialcellnumber andactivationpatternsinautisticbrainassessedvia cell-leveltechniques(discussedbelow).Therefore,this approachmaybeausefulalternativeforassessing activationand/orcellnumbersinpost-mortembrain tissuestudiesofASDpatients. Microgliacell-markerresearchisstillarelativelynew area,andthusthemarkersusedtoquantifymicroglial cellnumberandactivationarestilldebated.Toaddress thisissue,weusedfourdifferentmarkersthatare putativelymicroglial-specific.Ourresultsdemonstrate thatinthePFC,thereisincreasedexpressionofall microglialmarkersassessed,although AIF1 didnot reachstatisticalsignificance.However,previousreports haveshownthat AIF1 expressioninthebrainislow [32],potentiallycontributingtothisresult.Thefinding ofincreasedmicroglialcellmarkersinautisticPFCisin agreementwithanumberofstudiesthathavefoundincreasednumbersandactivationofmicrogliainautistic brains.Forinstance,Morganetal showedincreased microglialdensityindorsolateralPFCgreymatterof ASDbrainsviaIHCandstereology[8],andtheyalso demonstratedthatmicrogliaaremorecloselyassociated withneuronsinautisticdorsolateralPFCthanincontrols[9].Similarly,Tetreaultetal alsodemonstrated increasedmicroglialdensityinthefrontoinsularand visualcortexofautisticbrainsascomparedtocontrols [10].Additionally,anumberofstudieshavespecifically identifiedmicroglialactivationinautisticfrontalcortex, throughbothPETradiotracerimaging[13]andIHC/ cytokineprofilingapproaches[7].Ourresultsprovide furthersupporttothegrowingbodyofevidencedemonstratingincreasedmicroglialnumbersandactivationin autisticPFC,andourcellmarkergeneexpressionresults arelargelyinconcordancewiththesecelllevelstudies. Incontrast,ourcerebellarresultsshowsignificantly lowerexpressionofallfourmicroglialcellspecific Figure1 Expressionofcell-typespecificmarkersinprefrontalcortexsamplesofautisticcasesrelativetocontrols. Errorbarsrepresent 95%confidenceintervals.* P <0.05,** P <0.005,n.s.=notsignificant. Edmonson etal.MolecularAutism 2014, 5 :3 Page5of9 http://www.molecularautism.com/content/5/1/3

PAGE 6

markersinautisticbrains.Whileotherstudieshave identifiedmicroglialactivationinthecerebellumof autistictissue[33],nostudyhasattemptedtospecifically quantifymicroglialcellsinthecerebellumusingthe markersassessedhere,andthereforehistopathologic studiesinthecerebellumareneededtoconfirmthese findings.Onereportdescribedincreasedmicroglialcell activationinthecerebellum,assessedviaHLA-DRstaininginthewhitematterandgranularcelllayerofthe cerebellum[7],andanothershowedincreasedmicroglial activationthroughoutthebrain(althoughmostprominentlyinthecerebellum)usingan invivo PETmetabolic radiotracer[13].However,HLA-DRexpressionin humanmicroglialcellshasbeenshowntobehighly variablebetweenindividuals,anditsexpressionactually decreasesuponcytokinestimulation[34].Moreover,as wepreviouslydiscussed[35],thecerebellumisanatomicallyandphysiologicallyunique;thusmetabolicand pathologicalfindingsinthecerebellummustbeinterpretedwithcaution.Furthermore,ourtissuesamples fromcerebellumcontainedallthreelayersofthecerebellarcortex,asopposedtothemolecularlayeronly. Untildirecthistologicassessmentofthesemicroglial markersareperformedinpost-mortemautisticcerebellum,ourresultsmustbeinterpretedcautiously.However,ourresultsdosuggestthattherearesignificant differencesinmicroglialcellmarkersbetweenASDPFC andcerebellum,perhapsreflectingdifferencesinmicroglialactivationand/orcellnumbersbetweentheseareas inASDbrain. InboththePFCandthecerebellum,therewassignificantlyincreasedexpressionoftheastrocyte-specific marker GFAP inautisticbrains.Thistrendwasmost prominentinthecerebellum,where GFAP expression wasovertwo-foldhigherinASDbrainsthaninhealthy controls.Ourfindingsparallelthoseofpreviousstudies, whichhaveshownincreasedexpressionofGFAPprotein inthecerebellumandcortexofpatientsofautism throughIHCstaining,westernblotting,andmRNAexpression[7,17,36,37].Whilestudieshavenotbeendone toquantifyastrocytenumbersintheautisticcerebellum, ourresultsandthoseofpreviousstudiesprovideevidenceforastroglialreactioninautism. Interestingly,wealsofoundsignificantlydecreased expressionofthepan-neuronalmarker NEFL inboth thePFCandthecerebellumofautisticbrains.This resultisalsosupportedbypreviousstudies,whichhave showndecreased NEFL mRNAexpressionintheanteriorcingulategyrus,motorcortex,andthalamusofautisticbrains[38].However,cell-levelstudiesinautistic brainhaveproducedconflictingresultsaboutneuron numbers.Whilealargebodyofevidencehassuggested thereisalossofneuronsinmanyareasofautistic brains,asrecentlyreviewedin[39],otherstudieshave shownthatyoungautisticbrainsmayhave70%more neuronsinthePFC[40].Importantly,though,istheage ofthepatientattimeofdeath,aslongitudinalstudies havesuggestedthatearlybrainovergrowthinASDs quicklyreversestoaphenotypeofneuronalloss[41]. Consequently,theolderageofpatientsinthisstudymay biasourfindingstowardstheneuronallossspectrumof thedisease. Similarly,wefoundsignificantdecreasesintheGABAergicinterneuron-specificmarker PVA inboththePFCand cerebellumofautisticsamples Despitemanystudies demonstratingdecreasedGABAergiccomponentsacross differentareasoftheautisticbrain,asreviewedin[42],the onepathologicalanalysisofparvalbumin-positiveinterneuronsinASDdidnotidentifydifferencesintheautistic cerebellum[43].However,thisstudyonlyassessedthe molecularlayerofthecerebellarcortex,whereasourtissue samplescontainedallthreelayers.Additionally,while Figure2 Expressionofcell-typespecificmarkersincerebellumsamplesofautisticcasesrelativetocontrols. Errorbarsrepresent 95%confidenceintervals.* P <0.05,** P <0.005. Edmonson etal.MolecularAutism 2014, 5 :3 Page6of9 http://www.molecularautism.com/content/5/1/3

PAGE 7

parvalbumininterneuronshavebeenshowntobe unchangedintheautisticposteriorcingulatecortexand fusiformgyrus[44],andincreasedintheautistichippocampus[45],theyhavenotbeendirectlyassessedinthe autisticPFC. Inadditiontopotentiallyservingasmarkersofglial cellnumbersand/oractivation,manyofthegenes assessedhavespecificputativebiologicalrelevanceto ASDsthemselves.Forinstance,inbrain, Dap12 (also knownas TYROBP )encodesamicroglial-specifictransmembranesignalingpolypeptide[46].Theencoded proteinactsasanactivatingsignaltransductionelement withknownrolesinbrainmyelinationandinflammation [47].Itsreceptor, TREM2 ,encodesamembraneprotein thatfunctionsinmodulatingthebrain ’ simmuneresponseviaproductionofconstitutivecytokines,andis criticalforactivatingmicroglialphagocytosis[48].Rare mutationswithinthesetwogeneshavebeenassociated withNasu-Hakoladisease[49].Nasu-Hakoladiseaseis characterizedneurologicallybynewonsetpsychiatric andcognitivesymptomsinthefourthdecadeoflife, evolvingtomemorylossandcognitivedeclineresemblingAlzheimer ’ sdisease[50].Interestingly,singlenucleotidepolymorphismsinthisreceptorpathwaywere alsorecentlylinkedtoasignificantlyincreasedriskof Alzheimer ’ sdisease,andarethoughttorelatetothe inabilityofmicrogliatoproperlyremoveneurodegenerativedebrissuchasbeta-amyloid[51].Itisintriguingto speculatethatdefectsinthissamepathwayinneurodevelopmentaldisorderssuchasASDsmayalsoresultin defectsinmicroglialphagocytosis,butinthedevelopmentalcontextitistheinabilitytopruneoverabundant synapses,asopposedtoneurodegenerativedebris,that resultsinthebehavioralphenotype. Similarly, CX3CR1 ,alsoknownasthefractalkine receptor,encodesaproteinreceptoronthesurfaceof microgliathatbindstothechemokineCX3CL1(also calledfractalkine);fractalkinefunctionstoinducemicroglialmigrationandadhesionduringphagocytosis[52]. Thispathwaywasrecentlydemonstratedtoplaya criticalroleinthedevelopingbrainofmiceallowingfor migrationofmicrogliatotheirsynaptictargets,where phagocytosisandsynapticrefinementoccur. CX3CR1 knockoutmicehadmoresynapsesoncorticalneurons thanwild-typemice,anddisplayedsubtleneurological deficits[25]. AIF1 ,alsoknownas IBA1 ,encodesaproteinthatis consistentlyup-regulatedinexpressionduringmicroglial activation,andthereforehasbeenusedtodiscriminate betweenrestingandactivatedmicroglia[53].The AIF1 geneislocatedwithinthehighlyvariablelocuscontainingpartsofthemajorhistocompatibilitycomplex,which itselfhasbeenlinkedtoASDsanddysfunctionalmicrogliaphagocytosis[54]. Glialfibrillaryacidicprotein(GFAP)isanintermediate filamentproteinthatisexpressedbyastrocytes,andas discussedabove,hasbeenpreviouslyshowntobeupregulatedinautisticbrainsaswasalsoshownhere. Interestingly,decreasedexpressionofGFAPhasbeen reportedinotherneurodevelopmentaldisorders,suchas Schizophreniaandbipolardisorder[55].GFAP,like allintermediatedfilamentproteins,isimportantin maintainingthecellularcytoskeletoninastrocytes. Thecytoskeletonplaysanumberofkeyfunctional rolesinadditiontomaintainingcellshape,including inter-cellularcommunication,mitosis,andcellularmigrationduringphagocytosis.Mutationsin GFAP are responsibleforAlexander ’ sdisease,araredisorder characterizedbyseveredevelopmentaldelay,increased headsize,andseizures[56]. Overall,ourfindingsdemonstratethattheautistic brainbymid-childhoodhasmolecularchangesconsistentwithincreasedastrocyteexpressionanddecreased neuronalandinterneuronexpressioninboththePFC andcerebellum,withPFC-specificincreasedmicroglial markerexpression.Furthermore,thespecificglial moleculesfoundtobeabnormalinautisticbrainsare intimatelyinvolvedinglialmobilizationandphagocytosispathways,theyhavepreviouslybeenshowntobe criticalfornormalneurodevelopment,andareknown tobecausativeofotherrareneurodevelopmental phenotypes.Thesefindingssupportthenotionthata complexinterplaybetweenglialdysfunctionand neurogenesismayunderlietheclinicalmanifestations ofASDs. Thisstudyhasanumberoflimitationsofnote.Foremostistherelativelymodestsamplesize.Unfortunately, post-mortemhumanbrainresearchingeneralishamperedbythelackofaccessibilitytotissuesamples,and pediatricsamplesinparticulararescarce[26].Therefore, replicationwithalargenumberofsampleswillbe important.However,wechoseqRT-PCRtechniquesin thispilotstudybecauseoftheincreasedsensitivity comparedtowhole-genomemicroarrayorsequencing approaches,andthereforesomeaspectsofthesmall samplesizelimitationareaddressed.Secondly,dueto theinter-individualheterogeneityofthebrain,andin thebrain-bankingmethodologiesusedindistinguishing areasofpost-mortembraintissue,itcannotbeassumed thatallsampleswillderivefromtheexactsame anatomicsitewithinthePFCorcerebellum;thislimitationislargelyunavoidable.Lastly,theapproachofusing cell-typespecificmarkerexpressionasaproxyforcell numberand/oractivationstillneedsverification,by assessingthemconcurrentlywithtraditionalhistopathologic/stereologyanalysis.However,theconcordanceofourresultswithpreviouslypublishedstudies,andthe scarcityofappropriateASDbraintissueandtechnicalEdmonson etal.MolecularAutism 2014, 5 :3 Page7of9 http://www.molecularautism.com/content/5/1/3

PAGE 8

expertise,suggestthismaybeavaluableandsimple alternative ‘ screening ’ approach.ConclusionsInsummary,assessmentofglialnumbersandactivation inautismpost-mortembrainresearchishamperedbythe scarcityofappropriately-preservedtissue,andthetechnicalchallengeoftraditionalstereotacticmethods.We showthatglialandneuroncell-typespecificmarkershave mRNAexpressionpatternsthatparallelknowncellular aberrationsinASDs.Ourresultsprovidefurtherevidence thatglialcellsmayplayaroleinthepathogenesisof ASDs,andsuggestthatassessingforglialcell-typespecific markerexpressionmayrepresentaviableapproachto relativelyquantifyglialcellpatternsinASDpost-mortem research.Abbreviations AIF1 : Allograftinflammatoryfactor1;ASD:Autismspectrumdisorder; CX3CR1 :CX3Cchemokinereceptor1;IHC:Immunocytochemistry; PET:Positronemissiontomography;PFC:Prefrontalcortex;PMI:Post-mortem interval;qRT-PCR:Quantitativereal-timepolymerasechainreaction;RIN:RNA integritynumber; TREM2 :Triggeringreceptorexpressedonmyeloidcells2. Competinginterests Theauthorsdeclarethattheyhavenocompetinginterests. Authors  contributions CE,MNZ,andOMRconceivedofanddesignedtheanalysis.CEperformed theexperiments.CE,MNZ,andOMRanalyzedresultsandwrotethe manuscript.Allauthorsreadandapprovedthefinalmanuscript. Acknowledgements TheIntramuralResearchProgramattheNationalInstituteofChildHealth andHumanDevelopmentsupportedthiswork.MNZwasalsosupportedby BaylorCollegeofMedicineMSTPandtheNIH-UniversityofCambridgeBiomedicalScholarsProgram.CEwasadditionallysupportedbytheUniversity ofFloridaCollegeofMedicine.Thefundershadnoroleinstudydesign,data collectionandanalysis,decisiontopublish,orpreparationofthemanuscript. Authordetails1LaboratoryofClinicalandDevelopmentalGenomics,NationalInstituteof ChildHealthandHumanDevelopment,NationalInstitutesofHealth,49 ConventDrive,Building49,Room2C078,Bethesda,MD20814,USA.2UniversityofFloridaCollegeofMedicine,1600SWArcherRd,Gainesville,FL 32603,USA.3UniversityofCambridge,RobinsonCollege,GrangeRd, CambridgeshireCB39AN,UK.4BaylorCollegeofMedicineMSTP,OneBaylor Plaza,Houston,TX77030,USA. Received:14September2013Accepted:23December2013 Published:10January2014 References1.AmericanPsychiatricAssociation: DiagnosticandStatisticalManualofMental Disorders. 5thedition.Arlington:AmericanPsychiatricPublishing;2013. 2.CentersforDiseaseControlandPrevention: Prevalenceofautism spectrumdisorders — autismanddevelopmentaldisabilitiesmonitoring network,14sites,UnitedStates,2008. MMWR 2012, 3: 1 – 24. 3.ZoghbiHY: Postnatalneurodevelopmentaldisorders:meetingatthe synapse? Science 2003, 5646: 826 – 830. 4.JustMA,CherkasskyVL,KellerTA,KanaRK,MinshewNJ: Functionaland anatomicalcorticalunderconnectivityinautism:evidencefromanFMRI studyofanexecutivefunctiontaskandcorpuscallosummorphometry. CerebCortex 2007, 17 (4):951 – 961. 5.RubensteinJL,MerzenichMM: Modelofautism:increasedratioof excitation/inhibitioninkeyneuralsystems. GenesBrainBehav 2003, 2 (5):255 – 267. 6.CourchesneE,PierceK,SchumannCM,RedcayE,BuckwalterJA,Kennedy DP,MorganJ: Mappingearlybraindevelopmentinautism. Neuron 2007, 56 (2):399 – 413. 7.VargasDL,NascimbeneC,KrishnanC,ZimmermanAW,PardoCA: Neuroglialactivationandneuroinflammationinthebrainofpatients withautism. AnnNeurol 2005, 57 (1):67 – 81. 8.MorganJT,ChanaG,PardoCA,AchimC,SemendeferiK,BuckwalterJ, CourchesneE,EverallIP: Microglialactivationandincreasedmicroglial densityobservedinthedorsolateralprefrontalcortexinautism. BiolPsychiatry 2010, 68 (4):368 – 376. 9.MorganJT,ChanaG,AbramsonI,SemendeferiK,CourchesneE,EverallIP: Abnormalmicroglial-neuronalspatialorganizationinthedorsolateral prefrontalcortexinautism. BrainRes 2012, 1456: 72 – 81. 10.TetreaultNA,HakeemAY,JiangS,WilliamsBA,AllmanE,WoldBJ,Allman JM: Microgliainthecerebralcortexinautism. JAutismDevDisord 2012, 42 (12):2569 – 2584. 11.WatkinsLR,MilliganED,MaierSF: Glialactivation:adrivingforcefor pathologicalpain. TrendsNeurosci 2001, 24 (8):450 – 455. 12.HickmanSE,KingeryND,OhsumiTK,BorowskyML,WangLC,MeansTK,El KhouryJ: ThemicroglialsensomerevealedbydirectRNAsequencing. NatNeurosci2013, 16 (12):1896 – 1905. 13.SuzukiK,SugiharaG,OuchiY,NakamuraK,FutatsubashiM,TakebayashiK, YoshiharaY,OmataK,MatsumotoK,TsuchiyaKJ,IwataY,TsujiiM,Sugiyama T,MoriN: Microglialactivationinyoungadultswithautismspectrum disorder. JAMAPsychiatry 2013, 70 (1):49 – 58. 14.MaezawaI,JinLW: Rettsyndromemicrogliadamagedendritesand synapsesbytheelevatedreleaseofglutamate. JNeurosci 2010, 30 (15):5346 – 5356. 15.DereckiNC,CronkJC,LuZ,XuE,AbbottSB,GuyenetPG,KipnisJ: Wild typemicrogliaarrestpathologyinamousemodelofRettsyndrome. Nature 2012, 484 (7392):105 – 109. 16.CaoF,YinA,WenG,SheikhAM,TauqeerZ,MalikM,NagoriA,SchirripaM, SchirripaF,MerzG,BrownWT,LiX: Alterationofastrocytesand Wnt/ -cateninsignalinginthefrontalcortexofautisticsubjects. JNeuroinflammation 2012, 9 (1):223. 17.LaurenceJA,FatemiSH: Glialfibrillaryacidicproteiniselevatedinthe superior,frontal,parietalandcerebellarcorticesofpatientswithautism. Cerebellum 2005, 4: 206 – 210. 18.FatemiSH,FolsomTD,ReutimanTJ,LeeS: Expressionofastrocytic markersaquaporin4andconnexin43isalteredinbrainsofsubjects withautism. Synapse 2008, 62 (7):501 – 507. 19.MaezawaI,SwanbergS,HarveyD,LaSalleJM,JinLW: Rettsyndrome astrocytesareabnormalandspreadMeCP2deficiencythroughgap junctions. JNeurosci 2009, 29 (16):5051 – 5061. 20.YasuiDH,XuH,DunawayKW,LasalleJM,JinLW,MaezawaI: MeCP2 modulatesgeneexpressionpathwaysinastrocytes. MolAutism 2013, 4 (1):3. 21.YuskaitisCJ,BeurelE,JopeRS: Evidenceofreactiveastrocytes butnotperipheralimmunesystemactivationinamouse modelofFragileXsyndrome. BiochimBiophysActa 2010, 1802 (11):1006 – 1012. 22.UhlmannEJ,ApicelliAJ,BaldwinRL,BurkeSP,BajenaruML,OndaH, KwiatkowskiD,GutmannDH: Heterozygosityforthetuberoussclerosis complex(TSC)geneproductsresultsinincreasedastrocytenumbers anddecreasedp27-Kip1expressioninTSC2+/ cells. Oncogene 2002, 21 (25):4050 – 4059.23.LioyDT,GargSK,MonaghanCE,RaberJ,FoustKD,KasparBK,HirrlingerPG, KirchhoffF,BissonnetteJM,BallasN,MandelG: Aroleforgliainthe progressionofRett ’ ssyndrome. Nature 2011, 475 (7357):497 – 500. 24.ErogluC,BarresBA: Regulationofsynapticconnectivitybyglia. Nature 2010, 468: 223 – 231. 25.PaolicelliRC,BolascoG,PaganiF,MaggiL,ScianniM,PanzanelliP,Giustetto M,FerreiraTA,GuiducciE,DumasL,RagozzinoD,GrossCT: Synaptic pruningbymicrogliaisnecessaryfornormalbraindevelopment. Science 2011, 6048: 1456 – 1458. 26.AbbottA: Tissue-bankshortage:brainchild. Nature 2011, 478 (7370):442 – 443. 27.ZiatsMN,RennertOM: AberrantexpressionoflongnoncodingRNAsin autisticbrain. JMolNeurosci 2013, 49 (3):589 – 593.Edmonson etal.MolecularAutism 2014, 5 :3 Page8of9 http://www.molecularautism.com/content/5/1/3

PAGE 9

28.RansohoffR,PerryV: Microglialphysiology:uniquestimuli,stimuli, specializedresponses. AnnuRevImmunol 2009, 27: 119 – 145. 29.BabaH,NakahiraK,MoritaN,TanakaF,AkitaH,IkenakaK: GFAPgene expressionduringdevelopmentofastrocyte. DevNeurosci 1997, 19 (1):49 – 57. 30.Lpinoux-ChambaudC,EyerJ: Reviewonintermediatefilamentsofthe nervoussystemandtheirpathologicalalterations. HistochemCellBiol 2013, 140 (1):13 – 22. 31.CondeF,LundJS,JacobowitzDM,BaimbridgeKG,LewisDA: Localcircuit neuronsimmunoreactiveforcalretinin,calbindinD-28korparvalbumin inmonkeyprefrontalcortex:distributionandmorphology. JCompNeurol 1994, 341 (1):95 – 116. 32.ImaiY,IbataI,ItoD,OhsawaK,KohsakaS: AnovelgeneIba1inthemajor histocompatibilitycomplexclassIIIregionencodinganEFhandprotein expressedinamonocyticlineage. BiochemBiophysResCommun 1996, 224: 855 – 862. 33.CasanovaMF: Theneuropathologyofautism. BrainPathol 2007, 17 (4):422 – 433. 34.SmithAM,GibbonsHM,OldfieldRL,BerginPM,MeeEW,CurtisMA,Faull RL,DragunowM: M-CSFincreasesproliferationandphagocytosiswhile modulatingreceptorandtranscriptionfactorexpressioninadult microglia. JNeuroinflammation 2013, 10: 85. 35.ZiatsMN,RennertOM: Thecerebelluminautism:pathogenicoran anatomicalbeacon? Cerebellum 2013, 12 (5):776 – 777. 36.BaileyA,LuthertP,DeanA,HardingB,JanotaI,MontgomeryM, RutterM,LantosP: Aclinicopathologicalstudyofautism. Brain 1998, 121 (5):889 – 905. 37.PurcellAE,JeonOH,ZimmermanAW,BlueME,PevsnerJ: Postmortem brainabnormalitiesoftheglutamateneurotransmittersysteminautism. Neurology 2001, 57 (9):1618 – 1628. 38.AnithaA,NakamuraK,ThanseemI,YamadaK,IwayamaY,ToyotaT, MatsuzakiH,MiyachiT,YamadaS,TsujiiM,TsuchiyaKJ,MatsumotoK, IwataY,SuzukiK,IchikawaH,SugiyamaT,YoshikawaT,MoriN: Brain region-specificalteredexpressionandassociationofmitochondriarelatedgenesinautism. MolAutism 2012, 3 (1):12. 39.KernJK,GeierDA,SykesLK,GeierMR: Evidenceofneurodegenerationin autismspectrumdisorder.TranslNeurodegener 2013, 2 (1):17. 40.CourchesneE,MoutonPR,CalhounME,SemendeferiK,Ahrens-BarbeauC, HalletMJ,BarnesCC,PierceK: Neuronnumberandsizeinprefrontal cortexofchildrenwithautism. JAMA 2011, 306 (18):2001 – 2010. 41.CourchesneE,CampbellK,SolsoS: Braingrowthacrossthelifespanin autism:age-specificchangesinanatomicalpathology. BrainRes 2011, 1380: 138 – 145. 42.FatemiSH,BlattGJ: AlterationsinGABAergicbiomarkersintheautism brain:researchfindingsandclinicalimplications. AnatRec(Hoboken) 2011, 294 (10):1646 – 1652. 43.WhitneyER,KemperTL,RoseneDL,BaumanML,BlattGJ: Densityof cerebellarbasketandstellatecellsinautism:evidenceforalate developmentallossofPurkinjecells. JNeurosciRes 2009, 87 (10):2245 – 2254. 44.OblakAL,RoseneDL,KemperTL,BaumanML,BlattGJ: Alteredposterior cingulatecorticalcyctoarchitecture,butnormaldensityofneuronsand interneuronsintheposteriorcingulatecortexandfusiformgyrusin autism. AutismRes 2011, 4 (3):200 – 211. 45.LawrenceYA,KemperTL,BaumanML,BlattGJ: Parvalbumin-, calbindin-,andcalretinin-immunoreactivehippocampalinterneuron densityinautism. ActaNeurolScand 2010, 121 (2):99 – 108. 46.Paradowska-GoryckaA,JurkowskaM: Structure,expressionpatternand biologicalactivityofmolecularcomplexTREM-2/DAP12. HumImmunol 2013, 74 (6):730 – 737. 47.KaifuT,NakaharaJ,InuiM,MishimaK,MomiyamaT,KajiM,SugaharaA, KoitoH,Ujike-AsaiA,NakamuraA,KanazawaK,Tan-TakeuchiK,IwasakiK, YokoyamaWM,KudoA,FujiwaraM,AsouH,TakaiT: Osteopetrosisand thalamichypomyelinosiswithsynapticdegenerationinDAP12-deficient mice. JClinInvest 2003, 111 (3):323 – 332. 48.TakahashiK,RochfordCD,NeumannH: Clearanceofapoptoticneurons withoutinflammationbymicroglialtriggeringreceptorexpressedon myeloidcells-2. JExpMed 2005, 201 (4):647 – 657. 49.KanekoM,SanoK,NakayamaJ,AmanoN: Nasu-Hakoladisease:thefirst casereportedbyNasuandreview. Neuropathology 2010, 30 (5):463 – 470. 50.BianchinMM,MartinKC,deSouzaAC,deOliveiraMA,RiederCR: Nasu-Hakoladiseaseandprimarymicroglialdysfunction.NatRevNeurol 2010, 6 (9):2pfollowing523. 51.JonssonT,StefanssonH,SteinbergS,JonsdottirI,JonssonPV,SnaedalJ, BjornssonS,HuttenlocherJ,LeveyAI,LahJJ,RujescuD,HampelH,Giegling I,AndreassenOA,EngedalK,UlsteinI,DjurovicS,Ibrahim-VerbaasC,HofmanA,IkramMA,vanDuijnCM,ThorsteinsdottirU,KongA,StefanssonK: VariantofTREM2associatedwiththeriskofAlzheimer ’ sdisease. NEnglJ Med 2013, 368 (2):107 – 116. 52.WolfY,YonaS,KimKW,JungS: Microglia,seenfromtheCX3CR1angle. FrontCellNeurosci 2013, 18: 7 – 26. 53.SchwabJM,FreiE,KlusmanI,SchnellL,SchwabME,SchluesenerHJ: AIF-1 expressiondefinesaproliferatingandalertmicroglial/macrophage phenotypefollowingspinalcordinjuryinrats. JNeuroimmunol 2001, 119 (2):214 – 222. 54.NeedlemanLA,McAllisterAK: Themajorhistocompatibilitycomplexand autismspectrumdisorder. DevNeurobiol 2012, 72 (10):1288 – 1301. 55.Johnston-WilsonNL,SimsCD,HofmannJP,AndersonL,ShoreAD,Torrey EF,YolkenRH: Disease-specificalterationsinfrontalcortexbrainproteins inschizophrenia,bipolardisorder,andmajordepressivedisorder.The StanleyNeuropathologyConsortium. MolPsychiatry 2000, 5 (2):142 – 149. 56.LiR,JohnsonAB,SalomonsG,GoldmanJE,NaiduS,QuinlanR,CreeB, RuyleSZ,BanwellB,D ’ HoogheM,SiebertJR,RolfCM,CoxH,ReddyA, Gutirrez-SolanaLG,CollinsA,WellerRO,MessingA,vanderKnaapMS, BrennerM: Glialfibrillaryacidicproteinmutationsininfantile,juvenile, andadultformsofAlexanderdisease. AnnNeurol 2005, 57 (3):310 – 326.doi:10.1186/2040-2392-5-3 Citethisarticleas: Edmonson etal. : Alteredglialmarkerexpressionin autisticpost-mortemprefrontalcortexandcerebellum. MolecularAutism 2014 5 :3. 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 Edmonson etal.MolecularAutism 2014, 5 :3 Page9of9 http://www.molecularautism.com/content/5/1/3