ContentslistsavailableatScienceDirect
JournalofProteomics
journalhomepage:www.elsevier.com/locate/jprot
TripleSILACquantitativeproteomicanalysisrevealsdifferentialabundanceofcellsignalingproteinsbetweennormalandlungcancer-derivedexosomes
DavidJ.Clarka,b,WilliamE.Fondrieb,c,AustinYangb,LiMaoa,b,⁎
abcDepartmentofOncologyandDiagnosticSciences,UniversityofMarylandSchoolofDentistry,Baltimore,MD,USAMarleneandStewartGreenebaumCancerCenter,UniversityofMaryland,Baltimore,MD,USA
CenterforVascularandInflammatoryDiseases,UniversityofMarylandSchoolofMedicine,Baltimore,MD,USA
articleinfoabstract
Exosomesare30–100nmsizedmembranevesiclesreleasedbycellsintotheextracellularspacethatmediateintercellularcommunicationviatransferofproteinsandotherbiologicalmolecules.Tobetterunderstandtheroleofthesemicrovesiclesinlungcarcinogenesis,weemployedaTripleSILACquantitativeproteomicstrategytoexaminethedifferentialproteinabundancebetweenexosomesderivedfromanimmortalizednormalbron-chialepithelialcelllineandtwonon-smallcelllungcancer(NSCLC)celllinesharboringdistinctactivatingmuta-tionsinthecellsignalingmolecules:Kirstenratsarcomaviraloncogenehomolog(KRAS)orepidermalgrowthfactorreceptor(EGFR).Intotal,wewereabletoquantify721exosomalproteinsderivedfromthethreecelllines.Proteinsassociatedwithsignaltransduction,includingEGFR,GRB2andSRC,wereenrichedinNSCLCexosomes,andcouldactivelyregulatecellproliferationinrecipientcells.Thisstudy'sinvestigationoftheNSCLCexosomalproteomehasidentifiedenrichedproteincargothatcancontributetolungcancerprogression,whichmayhavepotentialclinicalimplicationsinbiomarkerdevelopmentforpatientswithNSCLC.
Biologicalsignificance:Thehighmortalityassociatedwithlungcancerisaresultoflate-stagediagnosisofthedisease.Currentscreeningtechniquesusedforearlydetectionoflungcancerlackthespecificityforaccuratediagnosis.Exosomesarenano-sizedextracellularvesicles,andtheincreasedabundanceofselectproteincargoinexosomesderivedfromcancercellsmaybeusedfordiagnosticpurposes.Inthispaper,weappliedquantitativeproteomicanalysistoelucidateabundancedifferencesinexosomalproteincargobetweentwoNSCLCcelllineswithdistinctiveoncogenemutationsandanimmortalizednormalbronchialepithelialcellline.Thisstudyre-vealedproteinsassociatedwithcelladhesion,theextracellularmatrix,andavarietyofsignalingmoleculeswereenrichedinNSCLCexosomes.ThepresentdatarevealsaproteinprofileassociatedwithNSCLCexosomesthatsuggestsarolethesevesicleshaveintheprogressionoflungcarcinogenesis,aswellasidentifiesseveralpromisingcandidatesthatcouldbeutilizedasamulti-markerproteinpanelinadiagnosticplatformforNSCLC.
©2015ElsevierB.V.Allrightsreserved.
Articlehistory:
Received6October2015
Receivedinrevisedform7December2015Accepted17December2015
Availableonline29December2015Keywords:Exosomes
Non-smallcelllungcancerSILAC
Quantitativeproteomics
1.Introduction
Lungcanceristheleadingcauseofcancer-relatedmortality,oftenduetoalackofearlydetectionresultinginadvanceddiseasestageatdiagnosis[1].Non-smallcelllungcancer(NSCLC)isthemostcommonclassificationoflungcancer,consistingofvaryinghistologieswhichmakesdiagnosisandselectingeffectivetherapeuticinterventionchallenging[2].Althougharecentscreeningtrialusinglow-dosecomputed-tomography(LDCT)revealedthatthescreeningmethodreducedlungcancerassociatedmortalityby20%comparedtochestradiography[3],theassociatedhighfalsepositiverate(96.4%)hasmadeitnecessarytodevelopcomplementarydiagnosticapproaches.
⁎Correspondingauthorat:DepartmentofOncologyandDiagnosticSciences,UniversityofMarylandSchoolofDentistry,650W.BaltimoreSt.,Baltimore,MD21201,USA.
E-mailaddress:Lmao@umaryland.edu(L.Mao).
Exosomesaremicrovesicles(30–100nm)ofendocyticoriginconstitu-tivelyreleasedbymultiplecelltypesintotheextracellularenvironmentbyfusionofamultivesicularbodywiththeplasmamembrane[4].Exosomeshaveafunctionalroleinintercellularcommunication[5],andexaminationofthemolecularcargoofcancer-derivedexosomeshasrevealedmiRNA,mRNA,andproteinsassociatedwithoncogenesis[6].Previousstudiesexaminingexosomesfromavarietyofcancershaveshownthattheseoncogenicfactorsarefunctionallyactive;anduptakebyrecipientcellscaninduceactivationofcellsignaling[7],increasecellgrowth[8],increasecellinvasion[9],aswellaspromotionoftumorgrowth[10]andmetasta-sis[11]invivo.Additionally,themolecularcargoofexosomesisreflectiveoftheirparentaltissue/cell-type[12],andthishasmadethemattractivetargetsfordevelopmentasabiomarkervesicle.
Togaininsightintothepotentialroleofexosomesinlungcarcino-genesis,weemployedstableisotopelabelingwithaminoacidsincellculture(SILAC)quantitativeproteomicstorevealdifferentialexosomal
http://dx.doi.org/10.1016/j.jprot.2015.12.023
1874-3919/©2015ElsevierB.V.Allrightsreserved.
162D.J.Clarketal./JournalofProteomics133(2016)161–169
proteinabundancebetweenanimmortalizednormalhumanbronchialepithelial(HBE)celllineandtwoNSCLCcelllines(A549andHCC827).OuranalysisrevealedexosomesderivedfromNSCLCcelllinestobeenrichedincellsignalingtransductionmolecules,andcouldinducecellproliferationinrecipientlungepithelialcells.TheresultsofthisstudyhaverevealedaproteinprofileassociatedwithNSCLCexosomesthatsuggestsexosomescanfacilitatelungcancerprogression,andiden-tifiedseveralcandidatesthatcanbefurtherexploredinadiagnosticplatform.2.Methods2.1.Materials
NSCLCcelllinesA549andHCC827werepurchasedfromAmericanTypeCultureCollection(ATCC,Manassas,VA).CDK4(cyclin-dependentkinase4)/hTERT(humantelomerasereversetranscriptase)-immortal-izedhumanbronchialepithelialcells(HBEC)wereagiftfromDr.JohnMinna(UniversityofSouthwesternMedicalCenter,Dallas,TX).Keratinocyte-SFMmedia,Keratinocyte-SFMsupplements,DMEMmedia,SILACDMEMmedia,fetalbovineserum,dialyzedfetalbovineserum,100U/mLpenicillin/streptomycinwerefromLifeTechnologies(GrandIsland,NY).SILACisotopes:4,4,5,5-D4-L-lysine(K4)and13C146N4-L-arginine(R6)or13C156N2-L-lysine(K8)and
13C156N4-L-arginine(R10)werefromSigmaAldrich(St.Louis,MO).Optiprep™wasfromSigma-Aldrich(St.Louis,MO).Mouseanti-ALIX(1A12),rabbitanti-CD63(H193),andrabbitanti-EGFR(1005)werefromSantaCruzBiotechnology(Dallas,TX),rabbitanti-SRC(36D10)andrabbitanti-NRP1(D62C6)werefromCellSignalingTechnology(Danvers,MA),mouseanti-TSG101(51/TSG101)wasfromBDBiosciences(SanJose,CA),andrabbitanti-KRAS(PA5-27234)wasfromThermoScientific(Waltham,MA).TrypsinwasobtainedfromThermoScientific(Waltham,MA).BCAassaykitPierce(Rockford,IL);SAXandC18SolidPhaseExtractionDiskwerefrom3M(StPaul,MN),andC18resinwasfromProspereonLifeScience(ArlingtonHeight,IL).2.2.Cellculture
HBE3andHBE4cellsweregrowninK-SFMmediumsupplementedwith50μg/mLbovinepituitaryextract(BPE)and5ng/mLrecombinantEGFaspreviouslydescribed[13].A549andHCC827celllinesweregrownandmaintainedinDMEMmediawith10%fetalbovineserumand100U/mLpenicillin/streptomycin.ForSILAClabelingA549andHCC827cellsweregrowninSILACmediawith10%dialyzedfetalbovineserumand100U/mLpenicillin/streptomycinsupplementedwith4,4,5,5-D4-L-lysine(K4)and13C146N4-L-arginine(R6)(medium)or13C156N2-L-lysine(K8)and13C156N4-L-arginine(R10)(heavy),respective-ly.Aftertencelldoublings,theincorporationofthestableisotopeswasN98%asdeterminedbyLC–MS/MSanalysisoftrypticpeptidesisolatedfrommedium-Lys4/Arg6-andheavy-Lys8/Arg10-labeledcells.Allcelllinesweremaintainedat37°Cand5%CO2inahumidifiedincubator.Forexosome-relatedexperiments,HBE4cellsweregrowninK-SFMmediumsupplementedwith50μg/mLbovinepituitaryextract(BPE)and5ng/mLrecombinantEGFthatwasexosome-conditioned(media-derivedexosomesweredepletedbyultracentrifugationfor16hat100,000×gtwice,priortofiltrationthrougha0.22μmfilter(Millipore,Billerica,MA)),andA549andHCC827cellswerefurthermaintainedinDMEMmediawith10%dFBS(whichhadbeendepletedofbovine-derivedexosomesbyultracentrifugationfor16hat100,00×gtwice,priortofiltrationthrougha0.22μmfilter)inthesameincubationconditions.
2.3.Exosomeisolation
A549andHCC827cellsweregrowntoapproximately90%confluencyintheirrespectiveSILACmediawith10%dFBS.Themedium
wasremovedandthecellswererinsedthreetimeswithphosphate-bufferedsaline(PBS)andgrownin10%dFBSsupplementedSILACmedia(20mLforeachdish).HBE4cellsweregrowntoapproximately90%confluencyintheirstandardSFMmedia.Themediumwasremovedandthecellswererinsedthreetimeswithphosphate-bufferedsaline(PBS)andgrowninexosome-conditionedSFMmedia(20mLforeachdish).After48h,theconditionedmediawascollectedandcentrifugedat2000×gfor20minat4°Ctoeliminatewholecellsandlargercelldebris.Thesupernatantwasfurthercentrifugedat20,000×gfor30minat4°Ctoremovecellfragmentsandlargermicrovesicles[14].Toprecipitateexosomes,supernatantswereincubatedovernightat4°Cinone-thirdvolumeofpoly-ethyleneglycolbuffer(33.4%PEG8000,50mMHEPESpH7.4,1MNaCl),followedbyultracentrifugationat12,000×gfor30min[15].Theprecipitatewasresuspendedin.25Msucrose,10mMTris–HCl,pH7.5andlayeredontopofanOptiprep™densitygradient.Adiscontinuousiodixanolgradientinconcentrationsof40%w/v,20%w/v,10%w/v,and5%w/vaspreparedbydilutingtheOptiprep™stocksolution(60%w/vaqueousiodixanolsolution)with0.25Msucrose,10mMTris–HCl,pH7.5aspreviouslydescribed[16].Theformedgradientwascentrifugedat100,000×gfor18hat4°C.Afterthecentrifugationwascompleted,twelve,1mLaliquotswerecollected(fromtoptothebottom),anddilutedwith5mLofPBS.Thefractionswerecentrifugedat100,000×gfor70min,thesupernatantswerediscarded,andtheresultingpelletswerewashedinPBS.Fractionswerecentrifugedat100,000×gfor70minoncemore,thesupernatantswerediscarded,andtheresultingpelletswereresuspendedinPBSbuff-er.TodeterminewhichOptiprepfractionscontainedexosomes,eachfractionwasimmunoblottedforthepresenceoftheexosomalmarkersALIXandCD63.Fractionspositiveforthemarkerswerepooledandusedinsubsequentanalyses(Optipreppellet).DensityoftheindividualOptiprepfractionswasdeterminedbyrunningacontrolgradientinpar-allel.Fractionswerecollectedaspreviouslydescribed,diluted1:10,000withdH2Oandabsorbancereadat244nmusingaNanodrop1000Spec-trometer(Nanodrop,Wilmington,DL)[17].Astandardcurvewascon-structedtodetermineourrespectivefractiondensities.2.4.Electronmicroscopy
Fornegativestainingmicroscopy,10μLofexosomeswerelayeredandabsorbedonaformvarcoated400meshcoppergridandstainedwith1%PTA.SamplewasimagedusingaFEItecnaiT12transmissionelectronmicroscope(FEI,Hillsboro,OR)equippedwithathermionicLab6filamentandoperatedatanaccelerationof80kV.Imagesweretakenwithapixelsizeof0.46nm,andadirectmagnificationof21000×usinganAMTCCDcamera.2.5.Westernblotting
Forimmunedetection,exosomeswerelysedbyaddingSDSinthesampletoafinalconcentrationof4%,sonicatedat90%amplitude,0.5scycleandboiledat95°C.LysatewasmixedwithLaemmlibufferwith50mMDTTandseparateusingprecast12%BioRadMini-PROTEANTGXgelswithTGSrunningbuffer.Followingelectrophore-sis,proteinswereelectrotransferredontoapolyvinylidenefluoride(PVDF)membrane.Afterbeingblockedin5%nonfatmilkinTBSTfor1hatRT,themembranewasprobedwiththeprimaryantibody(1:1000dilution)overnightat4°C.HRP-conjugatedsecondaryantibod-ies(goatanti-mouse,orgoatanti-rabbit)werediluted1:5000anddetectedusingSuperSignalWestPicoChemiluminescentSubstrate(Pierce,Rockford,IL).
2.6.Proteindigestionandnano-ESI–LC–MS/MSanalysis
Exosomepelletsfromeachcelllineweresubjectedtolysis(4%SDS,50mMtriethylammoniumbicarbonate(TEAB))andproteinconcentra-tionoftheexosomelysatewasmeasuredviaBCAProteinAssay.50μgof
D.J.Clarketal./JournalofProteomics133(2016)161–169163
eachexosomelysatefromHBE4-K0R0,A549-K4R6,andHCC827-K8R10weremixedina1:1:1ratioandsubjectedtotrypticdigestionusingtheEnhancedFASP(eFASP)protocolwithminormodifications[18].Samplewasmixedwith200μLof8MUrea,100mMTEABpH8.5,0.2%deoxycholicacid(DCA),andloadedontoaMicroconUFUnit(YM-3030kDacutofflimit;Millipore).Sampleswerereducedwith10mMTris2-carboxyethylphosphine(TCEP)andalkylatedwith50mMiodoacetamide.Trypticdigestionwasperformedovernightat37°C.Peptideswererecovered,lyophilizedandsubjectedtoStageTipSAXfractionationanddesaltedpriortoLC–MS/MSanalysis[19].Peptideswereseparatedbynanoscalereversed-phaseliquidchromatographyusinganXtremeSimplenanoLCsystem(CVC/Micro-Tech).Theanalyt-icalcolumnwaspreparedbypacking1.7μm200ÅC18resin(ProspereonLifeSciences)intoalaser-pulledfusedsilicacapillary(75μminnerdiameter,10.5cmlength,10μmtip;SutterInstruments)usingapressureinjectioncell(NextAdvance).PeptideswereinjectedintothesampleloopusinganEnduranceautosampler(SparkHolland)andwereloadedontothecolumnwith95%solventA(0.5%aceticacidinwater).A180minLCgradientmethodfrom5to70%SolventB(60%acetonitrile,0.5%aceticacid)withapostsplitflowof0.3mL/minwasusedtoelutethepeptidesintothemassspectrometer.TheLTQ-OrbitrapXLmassspectrometer(ThermoElectron)wasequippedwithananosprayionizationsource.Thesprayvoltagewas1.7kVandtheheatedcapillarytemperaturewas180°C.MS1datawereacquiredinprofilemodeintheOrbitrapwitharesolutionof60,000at400m/z,andthetop10mostintenseionsineachMS1scanwereselectedforcollisioninduceddissociationinthelineariontrap.Dynamicexclusionwasenabledwitharepeatcount2,repeatduration30s,andexclusionduration180s.Othermassspectrometrydatagenerationparameterswereasfollows:collisionenergy35%,ionselectionthresholdforMS/MS500counts,isolationwidth3m/z,defaultchargestate3,andchargestatescreeningenabled.Atotalofthreereplicateanaly-seswereperformed.
2.7.Proteinidentification,quantification,andanalysis
MSConvert(v3.06736)wasusedtoconverttheacquiredRAWfilestomzML.DatabasesearchesandPSMvalidationwereperformedintheTrans-ProteomicPipeline(TPP)[20].Comet(v2014.1)wasusedtoperformthedatabasesearchesagainstaUniProtKBhumanproteindatabase(versionJuly,2014;40,826reviewedsequences;48,167unre-viewedsequences)withaconcatenatedreversesequencedecoydatabase[21].Searchparametersinclude50ppmpeptidemasstolerance,1.0Dafragmenttolerance,variablemodification:methionine+15.99492,lysine+4.02511(D4),lysine+8.01420(13C156N2),arginine
+6.02013(13C6),arginine+10.00827(13C156N4),thefollowingfixedmodifications:Cysteine+57.02510(carbamidomethylation),andsemi-trypticpeptideswithuptothreemissedcleavages.PeptideProphetwasusedtovalidatethepeptidesearchresultsfromComet,andonlypeptideidentificationswithgreaterthan0.8probabil-ityandacalculatedfalsediscoveryrate(FDR)b1%werekeptforquan-tification.AllpeptideswereusedinProteinProphettoinferproteinsfromtheidentifiedpeptides,andproteinswithaProteinProphetprobabilitylessthan0.8werediscarded.QuantiMORE(formerlyIsoQuant)wasemployedtocalculatetheSILACratiosforthevalidatedpeptidesineachreplicate[22].Thesoftwareisavailablefordownloadatthefollowingwebaddress:http://www.proteomeumb.org/MZw.html.ProteinratioswerecomputedintheRprogramminglanguage(v3.1.1)fromthepeptideratiosgeneratedbyQuantiMORE.TheproteinratioistheaverageofthepeptideratiosassignedbyProteinProphetandweightedbythenumberofquantifiedPSMsforeachpeptide.Thereportedproteinratiosaretheaveragebetweenlog2mediannormal-izedreplicates.Proteinsquantifiedinatleasttworeplicateswereconsideredforstatisticalanalysis.Afoldchange(FC,log2ofM/LandH/Lproteinratios)of±0.5wasestablishedtodetermineproteinabundancedifferenceNSCLCandHBE4exosomepopulations.AStudent
t-testwasperformedtodistinguishsignificantproteinabundancedif-ferences.TheStoreyandTibshiraniq-valuemethodforfalsediscoveryrateestimationwasemployedthroughtheqvalueRpackage(v2.2.0)tocorrectformultiplehypothesistesting[23],andcanbefoundatthefollowingwebaddress:http://qvalue.princeton.edu/.Themassspec-trometryproteomicsdata,includingrawMSdata,COMETsearchresults,andQuantiMOREresultfileshavebeendepositedtotheProteomeXchangeConsortium[24]viathePRIDEpartnerrepositorywiththedatasetidentifierPXD003001.FunctionalannotationanalysisbyUniprotaccessionemployedPANTHER(ProteinAnalysisThroughEvolutionaryRelationships)[25]toassignProteinClasstoquantifiedproteinsandemployedDAVID(DatabaseforAnnotationandIntegratedDiscovery)(v6.7)[26,27]forFunctionalKEGGPathwayanalysis.
2.8.Invitroproliferationassay
HBE3cells(1000cellsperwell)wereseededina96-wellplateandculturedin100μLofK-SFMmediumsupplementedwith50μg/mLbovinepituitaryextract(BPE)and5ng/mLrecombinantEGF.At24hmediawasremovedandreplacedwithK-SFMmedium(withoutsupplements)andexosomesofvaryingconcentrations(0μg/mL,10μg/mL,50μg/mL,100μg/mL,and150μg/mL)inPBSwereadded.At48h,mediawasexchangedwithfreshK-SFMmedium(withoutsup-plements).At72hanMTTassaywasperformedasdescribedpreviously[28].Following2hincubationinMTTassayreagent,mediawasre-moved,DMSOaddedandtheabsorbanceofthesolutionwasmeasuredat570nm.
2.9.Statisticalanalysis
Proteinsconsideredtobesignificantlyincreasedordecreasedinabundanceincludedthecriteriaofaveragelog2mediannormalizedfoldchangeN±0.5andat-testp-valueb0.05.Thecellproliferationre-sultsarereportedasmean±s.d.,andareindicatedinthefigurelegends.DatasetswereanalyzedforstatisticalsignificanceusingStudent'spairedt-test.Statisticalsignificancewasreportedasp-valuesb0.01(*).
3.Results
3.1.Exosomeenrichmentandcharacterization
Togeneratetheexosomesamplesusedinthisanalysis(Fig.1A),weemployeddifferentialcentrifugationofourconditionedmedia(CM)tofirstremovewholecellsandcelldebris.Toreduceheterogeneityofthesecretedvesiclepopulation,largermicrovesicleswereisolatedbyhigher-speedcentrifugation[14,29],priortoprecipitationofexosomesfromtheCMwithapoly-ethyleneglycol(PEG)buffer.ThecrudeexosomeprecipitatewasloadedonanOptiprep™densitygradienttofurtherenrichforexosomesandreduceproteincontaminants(i.e.secretedproteinsandproteinaggregates).Ultracentrifugationofthedensitygradientwasperformedandtwelve1-mLfractionswerecol-lected.TheindividualfractionswereimmunoblottedforknownexosomalmarkerproteinsCD63andALIX.Fractions6and7,whichcorrespondedtoadensityof1.15and1.17andfallwithintheaccepteddensityofexosomes[30],stainedpositiveforCD63andALIXforbothNSCLCcelllines(A549andHCC827),whereasonlyFraction6stainedpositiveinourHBE4cellline(Fig.1B,C).Weusedelectronmicroscopytoexaminetheexosomal-positivefractionsfromeachcellline,whichrevealedahomogenousvesiclesizerangingfrom40to100nmindiam-eterandmorphologyconsistentwithexosomes(Fig.1D).Theyieldofexosomesobtainedusingthismethodologywere~36.6μg/10[8]cellsfromtheHBE4cellline,~55.7μg/10[8]cellsfortheA549cellline,and~55.6μg/10[8]cellsfortheHCC827cellline.
164D.J.Clarketal./JournalofProteomics133(2016)161–169
Fig.1.Anoverviewofexosomeisolationandcharacterization(A)differentialultracentrifugation,PEGprecipitation,followedbyOptiprepdensitygradientseparationwasemployedtoisolateexosomevesicles.(B)ImmunoblotforknownexosomemarkersALIXandCD63intherecovered1mLOptiprepfractions.Fractions6and7werepositivefortheexosomemarkerswithacorrespondingdensityof1.15–1.17g/mL.(C)ElectronmicroscopyimageofnegativelystainedA549,HCC827,andHBE4exosomevesicles.Bar=100nm.
3.2.TripleSILACproteomicanalysisoflungepithelialcelllinesHBE4,A549,andHCC827
ToevaluatedifferencesintheproteinprofileofexosomesderivedfromNSCLCwithdifferentgeneticbackgrounds,weemployedaTripleSILACquantitationstrategy.Forisotopiclabeling,wechosetwoNSCLCcelllines:A549cells,whichharboraG12SKRASmutation,werelabeledwith“medium”(M—K4R6)isotopewhereasHCC827cellswhichharboranE745-A750EGFRdeletionwerelabeledwith“heavy”(H—K8R10)isotope.Thethirdcellline,animmortalizednormalbronchialepithelialcelllineHBE4,was“light”(L—K0R0)containingendogenouslysineandarginine.AschematicofourTripleSILACworkflowisillus-tratedinFig.2.ThetwoNSCLCcelllinesselectedarerepresentativeoftwopredominant,butdistinct,non-overlappingmutationsthatoccurinlungadenocarcinomasinthesignalingproteinsEGFRandKRAS.
Exosomeswereisolatedfromsupernatantofeachcellline,mixedina1:1:1ratioandsubjectedtotrypsindigestion,followedbyLC–MS/MSanalysis.TheMS/MSspectraweresearchedandvalidatedusingtheTransProteomicPipeline(TPP).UsingaProteinProphetprobability0.8filter,1400proteingroupswereidentifiedbetweenallthreerepli-cates(proteinFDRb0.015forallreplicates).UsingtheSILACquantita-tionsoftwareQuantiMORE,wequantifiedatotalof1116proteinsbetweenthethreeexosomepopulations,with721exosomalproteinsquantifiedinatleasttwoofthethreereplicates.Apartiallistofquanti-fiedproteinswiththeirrespectivelog2mediannormalizedSILACratiosispresentedinTable1,withanexpandedlistofproteinsquantifiedinatleasttworeplicatesinTableS1.Quantifiedproteinsinourdatasetin-cludedESCRTproteins(STAM1,STAM2,VPS37B,VPS37C,VPS4B,andVTA1)proteinsassociatedwithexosomebiogenesis(ALIX,TSG101,Syntenin-1,andSyndecan-1),proteinsinvolvedinvesicletransport(VAMP2,VAMP8,SNX2,andSyntaxin-3),andtetraspanins(CD9,CD63,andCD81).Theanalysisrevealed233proteinsthatweremoreabundantinexosomesderivedfromthecelllineA549(M/L;log2FCN0.5)relativetoexosomesderivedfromHBE4,with97proteinssignificantlyincreasedinabundance(p-valueb0.05).WhencomparingexosomesderivedfromHCC827relativetoHBE4-derivedexosomes,240proteinswereincreasedinabundance(H/L;log2FCN0.5),with158proteinssignificantlyincreasedinabundance(p-valueb0.05).Overall,152proteinswerecommonlyincreasedinabundanceintheNSCLCexosomescomparedtoHBE4exosomes,with61proteinssignificantlyenrichedinbothNSCLCexosomepopulations.WethenanalyzedproteinsfoundtobemoreabundantinexosomesderivedfromHBE4relativetoNSCLCexosomes(M/L,H/L;log2FCb−0.5),iden-tifying212proteinsreducedinexosomesderivedfromtheA549cellline,butonly68proteinsreducedinabundancewhencomparedtoexosomesderivedfromHCC827.Whenincludingtheadditionalcriteria
D.J.Clarketal./JournalofProteomics133(2016)161–169165
Fig.3.ComparisonofproteinabundancedistributionbetweenHBE4,A549,andHCC827exosomes.Westernblottinganalysisofselectexosomalproteins(left),withcorrespondingreportedaveragemediannormalizedQuantiMOREproteinratiosfromtheTripleSILACproteomicanalysis(right).
Fig.2.SchematicofourTripleSILACquantitativeanalysisstrategy.Threecelllines(HCC827,A549,andHBE4)aredifferentiallyisotopicallylabeled,exosomesisolated,andexosomelysatesmixedpriortoenzymedigestion.SAXfractionationwasemployedpriortoLC–MS/MSanalysis.PeptidespectraareidentifiedattheMS/MSlevel,andquantifiedattheMSlevelusingQuantiMORE.
ofap-valueb0.05,78and29proteinswherefoundtobesignificantlyincreasedinHBE4exosomesrelativetoA549andHCC827exosomes,respectively.Intotal,35proteinswerecommonlyreducedintheNSCLCexosomescomparedtoHBE4exosomes,however,only8pro-teinswerefoundtobesignificantlyreduced(p-valueb0.05).WeusedWesternblotanalysistoverifyseveralselectedproteinsandobservedsimilarabundancepatternsoftheexosomalcargoproteinsacrossthethreeexosomepopulationsasquantifiedinourTripleSILACanalysis(Fig.3).
Table1
QuantifiedproteinsidentifiedinTripleSILACexosomeanalysis.Reportedproteinsvaluesaretheaveragelog2mediannormalizedproteinratiosreportedineachreplicate.Proteinaccession#
Genesymbol
Proteinname
HCC827/HBE4log2FC
ExosomemarkersandassociatedproteinsQ8WUM4PDCD6IPQ99816TSG101P08962CD63P21926CD9P60033CD81O00560SDCBPP18827SDC1CellAdhesionProteinsP16422P12830Q14126O95297O60487P48509Q12860
Programmedcelldeath6-interactingprotein(ALIX)Tumorsusceptibilitygene101proteinCD63antigenCD9antigenCD81antigenSyntenin-1Syndecan-1
0.480.681.321.81−0.571.770.65
p-value5.30E-031.28E-055.41E-033.90E-04ns
1.87E-042.48E-02
A549/HBE4log2FC0.900.701.720.941.182.181.37
p-value4.58E-031.09E-021.79E-022.94E-022.96E-034.10E-03ns
EPCAMCDH1DSG2MPZL1MPZL2CD151CNTN1EpithelialcelladhesionmoleculeCadherin-1Desmoglein-2
Myelinproteinzero-likeprotein1Myelinproteinzero-likeprotein2CD151antigenContactin-13.472.994.001.341.911.040.642.13E-02Ns
3.13E-031.59E-023.40E-021.80E-023.95E-020.871.042.421.462.080.924.73nsns
3.66E-03nsns
4.01E-028.55E-03
ReceptorsandsignalingproteinsP00553EGFRP01116KRASP08581METP12931SRCP35222CTNNB1O14786NRP1O60462NRP2O43854EDIL3Q08431MFGE8Q15582TGFBINSCLC-relatedproteinsP05556P35613P08195P13473P40199
ns—non-significant.
EpidermalgrowthfactorreceptorGTPaseKRAS
Hepatocytegrowthfactorreceptor
Proto-oncogenetyrosine-proteinkinaseSrcCateninbeta-1Neuropilin-1Neuropilin-2
EGF-likerepeatanddiscoidinI-likedomain-containingprotein3Lactadherin
Transforminggrowthfactor-beta-inducedproteinig-h33.880.802.781.302.091.191.233.370.33−1.041.28E-03nsns
4.07E-031.31E-02nsns
1.06E-03ns
3.63E-021.44−0.171.031.141.162.152.212.864.57−1.622.39E-02nsns
2.70E-03nsnsns
3.07E-032.55E-041.65E-02
ITGB1BSGSLC3A2LAMP2CEACAM6Integrinbeta-1Basigin
4F2cell-surfaceantigenheavychain
Lysosome-associatedmembraneglycoprotein2
Carcinoembryonicantigen-relatedcelladhesionmolecule62.241.530.851.622.492.65E-056.66E-041.23E-035.48E-021.54E-021.520.540.982.864.352.97E-034.57E-026.82E-031.79E-022.84E-03
166D.J.Clarketal./JournalofProteomics133(2016)161–169
3.3.ProteinannotationofNSCLCexosomes
TogaininsightintothefunctionalrolesofthequantifiedproteinsintheTripleSILACdataset(TableS1),weemployedontologyanalysisusingPANTHER(ProteinAnalysisThroughEvolutionaryRelationships)andpathwayanalysisusingDAVID(DatabaseforAnnotation,Visualiza-tion,andIntegratedDiscovery)(v6.7).Wefoundcytoskeletal,nucleicacidbinding,enzymemodulators,hydrolasesandreceptorproteinpopulationstobethemostabundantproteinclassesidentifiedinourdataset(Fig.4A).Next,weinvestigatedwhichproteinclassesweresig-nificantlyincreasedinabundance—IA(FCN0.5andp-valueb0.05;M/L,H/L)orsignificantlydecreasedinabundance—DA(FCb−0.5andp-valueb0.05,M/L,H/L)betweenNSCLCexosomesandexosomesderivedfromtheimmortalizednormalbronchialepithelialcellline.
Fig.4.FunctionalannotationofproteinsidentifiedintheTripleSILACanalysisusingPANTHERbioinformaticstool.(A)Distributionofproteinannotatedbyproteinclassesforallquantifiedproteins.(B)Distributionofannotatedproteinsbyproteinclassesforproteinsfoundtosignificantlyincreasedinabundance(IA;FCN0.5andp-valueb0.05,M/L,H/L)ordecreasedinabundance(DA;FCb−0.5andp-valueb0.05,M/L,H/L)inA549exosomes,and(C)inHCC827exosomes.
D.J.Clarketal./JournalofProteomics133(2016)161–169167
Extracellularmatrixproteins(EDIL3,MFGE8,HSPG2),proteases(PAPPA,CTSA,GGT3P),celladhesionproteins(DSG2,CD151,CNTN1),andreceptors(EGFR,GPRC5A,ITGB1,andITGB6)wereenrichedinA549exosomescomparedtoHBE4exosomes(Fig.4B).ProteinsfoundtobedecreasedinabundanceinA549exosomesrelativetoHBE4exosomesincludedcytoskeletalandnucleicacidbindingproteins.Similarly,proteinsincreasedinabundancebetweenHCC827exosomesandHBE4exosomesincludedproteases(PAPPA,CTSA,GGT3P),celladhesionproteins(DSG2,CD151,EPCAM,MPZL1,andMPZL2),andreceptors(EGFR,TACSTD2,ITGB1,andITGB6),aswellasproteinsclas-sifiedasenzymemodulators(GNB1,TOM1L1)(Fig.4C).Onlyproteinsclassifiedasnucleicacidbindingwerefoundtobedecreasedinabun-danceinHCC827exosomesrelativetoHBE4exosomes.
WhenweemployedKEGGPathwayAnalysisviaDAVID,wefoundproteinsassociatedwithcellproliferationtobeincreasedinNSCLCexosomes(Fig.5).InexosomesderivedfromHCC827,apanelofproteinswithincreasedabundanceincludingJAK1,CTNNB1,EGFR,GRB2,andRALAweremappedinthe“PathwaysofCancer”byKEGGanalysis(p-value4.7e−3).ExosomesderivedfromA549sharedasimilarmappingasHCC827withtheexceptionofCTNNB1andRALA(p-value4.0e−2).Takentogether,theseresultsindicatethatNSCLCexosomesareenrichedinproteinsclassifiedasreceptorsandsignaltransductionmoleculesrelativetoexosomesfromourimmortalizednormalbronchialepithelialcellline,withanassignedfunctionalroleincellproliferation.
3.4.FunctionalroleofNSCLCexosomes
Theidentificationofseveralcellproliferationassociatedsignaltrans-ductionmoleculesenrichedinNSCLCexosomespromptedustoassessthefunctionalroleoftheseextracellularvesicles.TodemonstratethatexosomesfromA549andHCC827cellscouldtransportfunctionallyac-tiveproteincargo,weinvestigatedtheimpactoftheNSCLCexosomesoncellproliferationinanotherimmortalizedbronchialepithelialcellline,HBE3.HBE3cellswereseededandculturedovernightbeforeaddingexosomesfromA549orHCC827cellsinconcentrationsof10μg/mL,50μg/mL,100μg/mLand150μg/mL.AsseeninFig.6,relativetocontrol(0μg/mL),NSCLCexosomescouldsignificantlyincreaseproliferationofHBE3cells(N20%)inadose-dependentmanneratconcentrationof50μg/mLandgreater(pb0.01).
Next,weassessedtheimpactofexosomesderivedfromHBE4cellsontheproliferationratesofHBE3cells.IncontrasttoNSCLCexosomes,
Fig.6.Exosomesimpactcellproliferationinrecipientcells.MTTassaymeasurementofHBE3cellgrowthafterincubationwithvaryingconcentrations(10–150μg/mL)ofexosomesderivedfromA549,HCC827,andHBE4cellsrelativetocontrol(0μg/mL).*pb0.01,n=3biologicalreplicates.
HBE4exosomesreducedHBE3cellproliferation(Fig.6),resultinginanapproximatedecreaseof20%and30%atconcentrationsof100μg/mLand150μg/mL,respectively(pb0.01).TheseresultsindicatethatNSCLCexosomestransportfunctionallyactivemoleculesthatcouldin-creasetheproliferationratesofrecipientcells,whereasHBE4exosomesdecreasedtheproliferativeratesinthesamerecipientcells.4.Discussion
Exosomesaresecretedvesiclesthatcontainvariousbiologicalmole-culesincludingsmallncRNA,mRNA,lipids,andproteins[31].Exosomes
Fig.5.ProteinsassociatedwithcellproliferationandcellsurvivalareenrichedinNSCLCexosomes.ProteinsmappedusingKEGGPathwayAnalysisviatheDAVIDbioinformaticstool.Legend:proteinssignificantlyincreasedinabundanceinA549exosomesonly(blue),HCC827exosomesonly(yellow),orbothNSCLCexosomepopulations(green).ProteinsquantifiedbutnotsignificantlyincreasedinabundanceinNSCLCexosomes(orange),andproteinsnotquantified(white).
168D.J.Clarketal./JournalofProteomics133(2016)161–169
havebeenidentifiedinavarietyofbiologicalfluids[32–34],andelevat-edlevelshavebeenobservedincancerpatients[35–38].Incancer,thesemicrovesiclesareinvolvedinintercellularcommunicationandcantransferoncogenicmaterialstorecipientcellstomodulateareasproxi-maltothetumor(cancermicroenvironment)[39],aswellasdistal(metastaticniche)[40].Inlungcancer,previousstudieshaveexploredtheroleofexosomesindrugresistance[41,42],andangiogenesis[43],inadditiontotheirpotentialapplicationindiagnostics[38,44].
Inthisstudy,weemployedTripleSILACquantitativeproteomicstocharacterizetheexosomalproteinprofilederivedfromtwoNSCLCcelllinesandanimmortalizednormalbronchialepithelialcellline.Usingthisstrategy,wequantified721exosomalcargoproteinsderivedfromthethreecelllines.Weshowedthatcelladhesionproteins,extracellularmatrixproteins,proteases,andcellsignalingmoleculesareenrichedinNSCLCexosomes.SeveralcellsignalingmoleculesassociatedwithcellproliferationweresignificantlyenrichedinNSCLCexosomes(Table1),includingSRCandEGFR,aswellasdownstreameffectorssuchasGRB2andRALA.Additionally,ouranalysisrevealedthattheMETrecep-tor,RAC1,andKRASproteinswereincreasedinabundanceinNSCLCexosomes,buttheseabundancedifferenceswerenotstatisticallysignif-icant(p-valueN0.05).Previously,examinationofthemicrovesiclepopulation(exosomesandotherextracellularvesicles)derivedfromNSCLCpleuraleffusionsrevealedasimilaroverlapofthesesamecellsig-nalingproteins[45].Thelatterstudyexaminedamoreheterogeneouspopulationofmicrovesicles,therefore,itisdifficulttoascertaintheproportionofthesecellsignalingproteinsthatarederivedexclusivelyfromexosomes.Regardless,theresultsofourstudyclearlyindicatethatthesesameproteinsareenrichedinNSCLCexosomes.
Ofnote,weidentifiedmutantsignaltransductionproteinsinourNSCLCexosomepopulations.ThecelllinesA549andHCC827harbormutationsinKRASandEGFR,respectively,andwefoundbothoftheseproteinstolocalizeinexosomesderivedfromtheirrespectivecelllines.Thelocalizationofthesemutantsignaltransductionproteins,aswellastheproteinsidentifiedinourKEGGPathwayAnalysis,intoNSCLCexosomescouldaffectsignalinginrecipientcells.Toexaminethislatterpointfurther,weco-incubatedA549andHCC827exosomeswithanimmortalizednormalbronchialepithelialcellline,HBE3,andfoundthatthevesiclescouldimpactcellproliferation.Thisisconsistentwithstudiesexaminingothertypesofcancerderivedexosomes,where-incancerexosomescouldstimulateproliferationinrecipientcellsbothinvitro[16,46]andinvivo[10,47].Interestingly,whenexosomesderivedfromHBE4wereaddedtoacultureofHBE3cells,weobservedsignificantreductionincellproliferation,suggestingapotentialroleofexosomesfromnormalcellsinregulatingorinhibitinggrowthofrecipientcells.Thus,weexaminedproteinsthatwereincreasedinabundanceinHBE4exosomescomparedtoNSCLCexosomesthatmightexplainthisphenomenon.TGFBIwasobservedtohavehigherabundanceinHBE4derivedexosomesrelativetobothNSCLCexosomepopulations.IthasbeenpreviouslyshownthatTGFBI/Betaig-h3proteinlevelsarereducedinlungtumorsrelativetomatchednormallungsasaresultofincreasedmethylationoftheTGFBIpromoterintumorcells[48].Furthermore,whenectopicallyexpressed,TGFBIcouldreducecellproliferationinvitroandreduceorevenabrogatetumorgrowthinvivo[49].TheincreasedabundanceofthisproteininHBE4exosomesmay,therefore,explaintheimpactofHBE4exosomesonthecellgrowthinhibition.
WealsoidentifiedseveralproteinsenrichedinNSCLCexosomespreviouslyimplicatedincellinvasionandangiogenesis(Table1).Contactin-1/CNTN1hasbeenshowntobeessentialfortumorcellinvasionandmetastasisinlungadenocarcinoma[50],andadditionalanalysisrevealedContactin-1couldfacilitatedownregulationofE-cadherininanAKT-mediatedmechanismtoinduceaninvasivephenotype[51].Lactadherin/MFGE8hasbeenassociatedwithinductionofanangiogeneticswitchinamousemodel[52],andEDIL3/DEL1hasawell-describedroleinangiogenesis[53].Increasedexpressionoftheneuropilins,NRP1andNRP2,hasbeenreportedinNSCLC[54],and
studieshaveindicatedapotentialroleinmetastasisandangiogenesis[55].Althoughweobservedmorethanatwo-foldincreaseinabun-danceofNRP1andNRP2inNSCLCexosomes,theyweren'tfoundtobestatisticallysignificant.Wedidnotexaminethisadditionalaspectofexosome-mediatedintercellularcommunicationinourstudy,where-inexosomesmayincreasetheinvasivenessofrecipientcells,orresultinapro-angiogenicphenotypeofrecipientendothelialcells.However,theidentificationofamultitudeofproteinswithknownrolesininvasionandangiogenesisenrichedinNSCLCexosomessupportsthehypothesisthatexosomescaninfluencethearchitectureofthetumormicroenvironmenttofacilitatelungcarcinogenesis.
Ofparticularinteresttomanyinthecancerfieldisthediagnosticpotentialofexosomes,especiallyinthecontextoftheirutilityinliquidbiopsies[56].OnepromisingapplicationofexosomesinadiagnosticssettingisTheExtracellularVesicleArray(EVArray),whichwasusedtoassessthediscriminatoryvalueofexosomalproteinsindistinguishingNSCLCpatientsfromcontrolindividuals[57].Theproteinarrayconsistedofbothknownexosomemarkers(ALIX,CD9,CD81)aswellasseveralcancercellmarkers(MUC1,CD151,CEA,EGFR),andresultsindicatedthataselect30markerpanelhadasensitivityof0.75andspecificityof0.76forlungcancerdiagnosis.Severalproteinswithasignificantp-valueintheEVarrayidentifiedasenrichedNSCLCexosomalcargoinouranalysis(FCN1.0)includedEGFR,MET,CD151,EPCAM,andCD9,providingadditionalevidencesupportingthisap-proach.However,theauthorsacknowledgedthatsomeoftheselectedmarkersusedinthearrayweredeemed“weak”intheNSCLCmodel,andthesubstitutionofstrongercandidatesmightallowforanincreaseinbothsensitivityandspecificity.Promisingcandidatesidentifiedinouranalysis,suchasBasigin/CD147andSLC3A2/CD98[58],CEACAM6[45],ITGB1[59],andLAMP2[60],warrantfurtherexamination.Itiswidelyacknowledgedthatamultivariatepanelofproteinmarkerswouldbemorereliablethanasinglemarkerinadiagnosticsetting[61],andtheexosomeproteomeprovidesamultitudeofcandidatetargets.
5.Conclusion
OuranalysisincorporatedSILAC-basedquantitativeproteomicstoexaminedifferentialproteinabundancebetweenimmortalizednormalepithelialcellderivedexosomesandNSCLCexosomes.ProteinsenrichedintheNSCLCexosomalproteomeincludeproteinsassociatedwithcellinvasion,angiogenesis,andcellproliferation,andNSCLCexosomescanactivelyregulatetheproliferativecapacityofrecipientcells.Inaddition,weidentifiedseveralproteinsthathavebeenprevi-ouslyimplicatedhavingaroleinlungcarcinogenesisthatmayserveasdiagnosticmarkersoflungcancer.Together,thisstudyprovidesasolidfoundationforthefuturedevelopmentofaplatformthatutilizesexosomesasmulti-markerphenotypingtoolinclinicalstudies.
AbbreviationsCMconditionedmediaESCRTendosomalsortingcomplexesrequiredfortransportEGFRepidermalgrowthfactorreceptorHBEhumanbronchialepithelialKRASkirstenratsarcomaviraloncogenehomologNSCLCnon-smallcelllungcancerSILACstableisotopelabelingofaminoacidsincellcultureSupplementarydatatothisarticlecanbefoundonlineathttp://dx.doi.org/10.1016/j.jprot.2015.12.023.
Conflictsofinterest
Authorsdeclarethatthereisnoconflictofinteresttodisclose.
D.J.Clarketal./JournalofProteomics133(2016)161–169169
References
[1]M.Hassanein,J.C.Callison,C.Callaway-Lane,M.C.Aldrich,etal.,Thestateofmolec-ularbiomarkersfortheearlydetectionoflungcancer,CancerPrev.Res.5(2012)992–1006.
[2]A.F.Brown,D.Sirohi,J.Fukuoka,P.T.Cagle,etal.,Tissue-preservingantibody
cocktailstodifferentiateprimarysquamouscellcarcinoma,adenocarcinoma,andsmallcellcarcinomaoflung,Arch.Pathol.Lab.Med.137(2013)1274–1281.
[3]D.R.Aberle,A.M.Adams,C.D.Berg,W.C.Black,etal.,Reducedlung-cancermortality
withlow-dosecomputedtomographicscreening,N.Engl.J.Med.365(2011)395–409.
[4]B.Février,G.Raposo,Exosomes:endosomal-derivedvesiclesshippingextracellular
messages,Curr.Opin.CellBiol.16(2004)415–421.
[5]C.Théry,Exosomes:secretedvesiclesandintercellularcommunications,F1000Biol.
Rep.3(2011)15.
[6]M.C.Henderson,D.O.Azorsa,Thegenomicandproteomiccontentofcancercell-derivedexosomes,Front.Oncol.2(2012)1–9.
[7]D.G.Meckes,K.H.Y.Shair,A.R.Marquitz,C.-P.Kung,etal.,Humantumorvirus
utilizesexosomesforintercellularcommunication,Proc.Natl.Acad.Sci.U.S.A.107(2010)20370–20375.
[8]J.Skog,T.Würdinger,S.vanRijn,D.H.Meijer,etal.,Glioblastomamicrovesicles
transportRNAandproteinsthatpromotetumourgrowthandprovidediagnosticbiomarkers,Nat.CellBiol.10(2008)1470–1476.
[9]S.Atay,S.Banskota,J.Crow,G.Sethi,L.Rink,A.K.Godwin,OncogenicKIT-containing
exosomesincreasegastrointestinalstromaltumorcellinvasion,Proc.Natl.Acad.Sci.U.S.A.111(2014)711–716.
[10]W.Zhu,L.Huang,Y.Li,X.Zhang,etal.,Exosomesderivedfromhumanbonemarrow
mesenchymalstemcellspromotetumorgrowthinvivo,CancerLett.315(2012)28–37.
[11]H.Peinado,M.Alečković,S.Lavotshkin,I.Matei,etal.,Melanomaexosomeseducate
bonemarrowprogenitorcellstowardapro-metastaticphenotypethroughMET,Nat.Med.18(2012)883–891.
[12]R.J.Simpson,J.W.E.Lim,R.L.Moritz,S.Mathivana,Exosomes:proteomicinsightsand
diagnosticpotential,ExpertRev.Proteomics6(2009)267–283.
[13]R.D.Ramirez,Immortalizationofhumanbronchialepithelialcellsintheabsenceof
viraloncoproteins,CancerRes.64(2004)9027–9034.
[14]B.J.Tauro,D.W.Greening,R.A.Mathias,S.Mathivanan,etal.,Twodistinctpopula-tionsofexosomesarereleasedfromLIM1863coloncarcinomacell-derivedorganoids,Mol.Cell.Proteomics12(2012)587–598.
[15]C.Lee,S.A.Mitsialis,M.Aslam,S.H.Vitali,etal.,Exosomesmediatethe
cytoprotectiveactionofmesenchymalstromalcellsonhypoxia-inducedpulmonaryhypertension,Circulation126(2012)2601–2611.
[16]H.Ji,D.W.Greening,T.W.Barnes,J.W.Lim,etal.,Proteomeprofilingofexosomes
derivedfromhumanprimaryandmetastaticcolorectalcancercellsrevealdifferen-tialexpressionofkeymetastaticfactorsandsignaltransductioncomponents,Proteomics13(2013)1672–1686.
[17]M.Schröder,R.Schäfer,P.Friedl,Spectrophotometricdeterminationofiodixanolin
subcellularfractionsofmammaliancells,Anal.Biochem.244(1997)174–176.
[18]J.Erde,R.R.O.Loo,J.A.Loo,EnhancedFASP(eFASP)toincreaseproteomecoverage
andsamplerecoveryforquantitativeproteomicexperiments,J.ProteomeRes.13(2014)1885–1895.
[19]J.R.Wiśniewski,A.Zougman,M.Mann,CombinationofFASPandStageTip-based
fractionationallowsin-depthanalysisofthehippocampalmembraneproteome,J.ProteomeRes.8(2009)5674–5678.
[20]M.C.Chambers,B.Maclean,R.Burke,D.Amodei,etal.,Across-platformtoolkitfor
massspectrometryandproteomics,Nat.Biotechnol.30(2012)918–920.
[21]J.K.Eng,T.A.Jahan,M.R.Hoopmann,Comet:anopen-sourceMS/MSsequencedata-basesearchtool,Proteomics13(2013)22–24.
[22]Z.Liao,Y.Wan,S.N.Thomas,A.J.Yang,IsoQuant:asoftwaretoolforstableisotope
labelingbyaminoacidsincellculture-basedmassspectrometryquantitation,Anal.Chem.84(2012)4535–4543.
[23]J.D.Storey,R.Tibshirani,Statisticalsignificanceforgenomewidestudies,Proc.Natl.
Acad.Sci.U.S.A.100(2003)9440–9445.
[24]J.Vizcaíno,E.Deutsch,R.Wang,ProteomeXchangeprovidesgloballycoordinated
proteomicsdatasubmissionanddissemination,Nat.Biotechnol.32(2014)223–226.
[25]H.Mi,A.Muruganujan,J.T.Casagrande,P.D.Thomas,Large-scalegenefunctionanal-ysiswiththePANTHERclassificationsystem,Nat.Protoc.8(2013)1551–1566.[26]D.W.Huang,B.T.Sherman,R.A.Lempicki,Bioinformaticsenrichmenttools:paths
towardthecomprehensivefunctionalanalysisoflargegenelists,NucleicAcidsRes.37(2009)1–13.
[27]D.W.Huang,B.T.Sherman,R.A.Lempicki,Systematicandintegrativeanalysisof
largegenelistsusingDAVIDbioinformaticsresources,Nat.Protoc.4(2009)44–57.
[28]Y.Mei,Y.Wang,P.Kumari,A.C.Shetty,etal.,ApiRNA-likesmallRNAinteractswith
andmodulatesp-ERMproteinsinhumansomaticcells,Nat.Commun.6(2015)7316.
[29]M.Wysoczynski,M.Z.Ratajczak,Lungcancersecretedmicrovesicles:underappreci-atedmodulatorsofmicroenvironmentinexpandingtumors,Int.J.Cancer125(2009)1595–1603.
[30]C.Thery,S.Amigorena,G.Raposo,A.Clayton,Isolationandcharacterizationof
exosomesfromcellculturesupernatantsandbiologicalfluids,Curr.Protoc.CellBiol.(2006)(Chapter3).
[31]E.Pap,E.Pállinger,A.Falus,Theroleofmembranevesiclesintumorigenesis,Crit.
Rev.Oncol.Hematol.79(2011)213–223.
[32]T.Pisitkun,R.-F.Shen,M.A.Knepper,Identificationandproteomicprofilingof
exosomesinhumanurine,Proc.Natl.Acad.Sci.U.S.A.101(2004)13368–13373.[33]M.-P.Caby,D.Lankar,C.Vincendeau-Scherrer,G.Raposo,C.Bonnerot,Exosomal-likevesiclesarepresentinhumanbloodplasma,Int.Immunol.17(2005)879–887.[34]J.M.Street,P.E.Barran,C.L.Mackay,S.Weidt,etal.,Identificationandproteomic
profilingofexosomesinhumancerebronspinalfluid,J.Transl.Med.(2012)10.[35]F.Andre,N.E.C.Schartz,M.Movassagh,C.Flament,P.Pautier,P.Morice,C.Pornel,C.
Lhomme,B.Escudier,T.LeChevalier,T.Tursz,S.Amigorena,G.Raposo,E.Angevin,L.Zitvogel,Malignanteffusionsandimmunogenictumour-derivedexosomes,Lancet360(2002)295–305.
[36]D.D.Taylor,C.Gercel-Taylor,MicroRNAsignaturesoftumor-derivedexosomesas
diagnosticbiomarkersofovariancancer,Gynecol.Oncol.110(2008)13–21.
[37]M.Logozzi,A.DeMilito,L.Lugini,M.Borghi,etal.,Highlevelsofexosomesexpress-ingCD63andcaveolin-1inplasmaofmelanomapatients,PLoSOne(2009)4.
[38]G.Rabinowits,C.Gerçel-Taylor,J.M.Day,D.D.Taylor,G.H.Kloecker,Exosomal
microRNA:adiagnosticmarkerforlungcancer,Clin.LungCancer10(2009)42–46.[39]R.Ge,E.Tan,S.Sharghi-Namini,H.H.Asada,Exosomesincancermicroenvironment
andbeyond:haveweoverlookedtheseextracellularmessengers?CancerMicroenviron.1–10(2012).
[40]R.M.Hoffman,Stromal-cellandcancer-cellexosomesleadingthemetastaticexodus
forthepromisedniche,BreastCancerRes.15(2013)310.
[41]X.Xiao,S.Yu,S.Li,J.Wu,etal.,Exosomes:decreasedsensitivityoflungcancerA549
cellstocisplatin,PLoSOne9(2014)1–6.
[42]D.Y.Choi,S.You,J.H.Jung,J.C.Lee,etal.,Extracellularvesiclesshedfromgefitinib-resistantnonsmallcelllungcancerregulatethetumormicroenvironment,Proteo-mics14(2014)1845–1856.
[43]H.Cui,B.Seubert,E.Stahl,H.Dietz,etal.,Tissueinhibitorofmetalloproteinases-1
inducesapro-tumourigenicincreaseofmiR-210inlungadenocarcinomacellsandtheirexosomes,Oncogene(2014)1–11.
[44]Y.Li,Y.Zhang,F.Qiu,Z.Qiu,ProteomicidentificationofexosomalLRG1:apotential
urinarybiomarkerfordetectingNSCLC,Electrophoresis32(2011)1976–1983.
[45]J.O.Park,D.-Y.Choi,D.-S.Choi,H.J.Kim,etal.,Identificationandcharacterizationof
proteinsisolatedfrommicrovesiclesderivedfromhumanlungcancerpleuraleffu-sions,Proteomics13(2013)2125–2134.
[46]W.Mu,S.Rana,M.Zöller,Hostmatrixmodulationbytumorexosomespromotes
motilityandinvasiveness,Neoplasia15(2013)875–887.
[47]S.Keller,A.K.König,F.Marmé,S.Runz,etal.,Systemicpresenceandtumor-growth
promotingeffectofovariancarcinomareleasedexosomes,CancerLett.278(2009)73–81.
[48]G.Shao,J.Berenguer,A.C.Borczuk,C.A.Powell,etal.,Epigeneticinactivationof
betaig-h3geneinhumancancercells,CancerRes.66(2006)4566–4573.
[49]Y.Zhao,M.El-Gabry,T.K.Hei,Lossofbetaig-h3proteinisfrequentinprimarylung
carcinomaandrelatedtotumorigenicphenotypeinlungcancercells,Mol.Carcinog.45(2006)84–92.
[50]J.-L.Su,Knockdownofcontactin-1expressionsuppressesinvasionandmetastasisof
lungadenocarcinoma,CancerRes.66(2006)2553–2561.
[51]J.Yan,N.Wong,C.Hung,W.X.-Y.Chen,D.Tang,Contactin-1reducesE-cadherin
expressionviaactivatingAKTinlungcancer,PLoSOne8(2013)e65463.
[52]M.Neutzner,T.Lopez,X.Feng,E.S.Bergmann-Leitner,etal.,MFG-E8/lactadherin
promotestumorgrowthinanangiogenesis-dependenttransgenicmousemodelofmultistagecarcinogenesis,CancerRes.67(2007)6777–6785.
[53]Y.Aoka,F.L.Johnson,K.Penta,K.HirataKi,etal.,Theembryonicangiogenicfactor
Del1acceleratestumorgrowthbyenhancingvascularformation,Microvasc.Res.64(2002)148–161.
[54]S.Lantuéjoul,B.Constantin,H.Drabkin,C.Brambilla,etal.,ExpressionofVEGF,
semaphorinSEMA3F,andtheircommonreceptorsneuropilinsNP1andNP2inpreinvasivebronchiallesions,lungtumours,andcelllines,J.Pathol.200(2003)336–347.
[55]L.M.Ellis,Theroleofneuropilinsincancer,Mol.CancerTher.5(2006)1099–1107.[56]C.Rolfo,M.Castiglia,D.Hong,R.Alessandro,etal.,Liquidbiopsiesinlungcancer:the
newambrosiaofresearchers,Biochim.Biophys.ActaRev.Cancer2014(1846)539–546.
[57]K.R.Jakobsen,B.S.Paulsen,R.Baek,K.Varming,etal.,Exosomalproteinsaspotential
diagnosticmarkersinadvancednon-smallcelllungcarcinoma,J.Extracell.Vesicles4(2015)26659.
[58]F.Fei,X.Li,L.Xu,D.Li,etal.,CD147-CD98hccomplexcontributestopoorprognosis
ofNon-smallcelllungcancerpatientsthroughpromotingcellproliferationViathePI3K/Aktsignalingpathway,Ann.Surg.Oncol.21(2014)4359–4368.
[59]X.M.Wang,J.Li,M.X.Yan,L.Liu,etal.,I.A.I.Osteopontin,LAMB3andITGB1ascritical
pro-metastaticgenesforlungcancer,PLoSOne(2013)8.
[60]A.Soltermann,R.Ossola,S.Kilgus-Hawelski,A.VonEckardstein,etal.,N-glycoproteinprofilingoflungadenocarcinomapleuraleffusionsbyshotgunproteo-mics,Cancer114(2008)124–133.
[61]N.L.Anderson,N.G.Anderson,Thehumanplasmaproteome:history,character,and
diagnosticprospects,Mol.Cell.Proteomics1(2002)845–867.
因篇幅问题不能全部显示，请点此查看更多更全内容