Article
TheRab2AGTPasePromotesBreastCancerStem
ErkSignalingActivation
Highlights
dPin1increasesRab2Atranscription,promotingBCSC
expansionandtumorigenesis
dRab2AblocksErk1/2inactivationbyMKP3,leadingtoZeb1
andb-cateninactivation
dRab2Aamplificationormutationinhumancanceractivates
Erk1/2andincreasesBCSCs
dRab2Aoverexpressionistightlylinkedtohighmortalityin
Authors
Man-LiLuo,ChangGong,...,ErweiSong,
KunPingLu
Correspondence
klu@bidmc.harvard.edu
InBrief
Rab2Atranscriptiontopromote
expansionofbreastcancerstem-like
cellsandtumorigenesisviaErk1/2
pathwayactivation.Thefindingssuggest
possibledrugtargetsforbreastcancer
Luoetal.,2015,CellReports11,111–124
April7,2015a2015TheAuthors
http://dx.doi.org/10.1016/j.celrep.2015.03.002
humanbreastcancer
GraphicalAbstract
CellsandTumorigenesisvia
therapy.
AccessionNumbers
GSE49971
thattheprolylisomerasePin1increases
Rab2AisasmallGTPaseinvolvedin
vesicletrafficking.Luoetal.nowshow
CellReports
Deaconess
Univers
overexpressioncorrelateswithpoorclinicaloutcome
Promotifsexistintwodistinctconformations,cisandtrans,
andidentifiedPin1,auniqueprolylisomerasethatbindstoand
regulatorymechanismsandtheidentificationoftargetstoerad-
icatetheCSCcompartmentinatumormaybeessentialto
achievelong-termremissionofcancer(LiuandWicha,2010).
downstreamtargetofPin1inBCSCsislargelyunknown.
InsearchingforPin1downstreamtargetsinBCSCsusing
genome-wideexpressionprofiling,weidentifiedRab2A,asmall
Anincreasingnumberofregulatorsofbreastcancerstem-like
cells(BCSCs),notablytranscriptionfactorsincludingZeb1and
GTPasemainlylocalizedtotheER-Golgiintermediatecompart-
ment(ERGIC)thatisessentialformembranetraffickingbetween
inbreastcancerpatients.Thus,Pin1/Rab2A/Erk
drivesBCSCexpansionandtumorigenicity,suggest-
ingpotentialdrugtargets.
INTRODUCTION
Cancerstemcells(CSCs)ortumor-initiatingcells(TICs)have
beenhypothesizedtoretainthecapacityofself-renewalto
regeneratethebulkofaheterogeneoustumor.TheCSCconcept
hasimportantimplicationsforunderstandingthemolecular
mechanismsofcancerprogressionandidentifyingtargetsfor
cancertherapeutics,sinceCSCsarethoughttoberesponsible
fortumorinitiation,progression,metastasis,relapse,anddrug
resistance(LiuandWicha,2010).Thus,theelucidationofCSC
catalyzescis/transisomerizationofspecificpSer/Thr-Promotifs
(Luetal.,1996;LuandZhou,2007;Yaffeetal.,1997).Pin1in-
ducesconformationalchangesoftheseSer/Thr-Promotifsafter
phosphorylation,whichnowcanbevisualizedbyprolineisomer-
specificantibodies(Nakamuraetal.,2012).Importantly,Pin1is
overexpressedand/oractivatedinhumancancersandplaysa
criticalroleinbreastcancerdevelopmentinvitroandinvivo
(Chenetal.,2013;Leeetal.,2011;LuandZhou,2007;Luand
Hunter,2014).Recently,weandothershavefoundthatPin1is
increasedinhumanBCSCsandplaysafundamentalrolein
drivingBCSCsandtumorigenesis(Luoetal.,2014;Rustighi
etal.,2014).AlthoughPin1hasbeenreportedtoactivateand
inactivatealargesubsetofkeyoncogenesandtumorsuppres-
sors,respectively(LuandZhou,2007;LuandHunter,2014),the
Rab2Apotentlyinhibitstheexpansionandtumori-
genesisoffreshlyisolatedBCSCs.Finally,Rab2A
senandHunter,2001).WehaveshownthatcertainpSer/Thr-
Article
TheRab2AGTPasePromotes
andTumorigenesisviaErkSignaling
Man-LiLuo,
1,2
ChangGong,
2
Chun-HauChen,
1
HaiHu,
1
Pengyu
GerburgWulf,
1
JudyLieberman,
4
XiaoZhenZhou,
1
ErweiSong,
1
DepartmentofMedicineandCancerResearchInstitute,BethIsrael
MA02215,USA
2
BreastTumorCenter,SunYat-SenMemorialHospital,SunYat-Sen
3
InstituteforTranslationalMedicine,FujianMedicalUniversity,Fuzhou
4
PrograminCellularandMolecularMedicine,BostonChildren’sHospital,
Correspondence:klu@bidmc.harvard.edu
http://dx.doi.org/10.1016/j.celrep.2015.03.002
ThisisanopenaccessarticleundertheCCBY-NC-NDlicense(http://creativeco
SUMMARY
Proline-directedphosphorylationisregulatedbythe
prolylisomerasePin1,whichplaysafundamental
roleindrivingbreastcancerstem-likecells(BCSCs).
Rab2AisasmallGTPasecriticalforvesicletraf-
ficking.Here,weshowthatPin1increasesRab2A
transcriptiontopromoteBCSCexpansionand
tumorigenesisinvitroandinvivo.Mechanistically,
Rab2Adirectlyinteractswithandpreventsdephos-
phorylation/inactivationofErk1/2bytheMKP3
phosphatase,resultinginZeb1upregulationand
b-cateninnucleartranslocation.Incancercells,
Rab2Aisactivatedviageneamplification,mutation
orPin1overexpression.Rab2Aoverexpressionor
mutationendowsBCSCtraitstoprimarynormal
humanbreastepithelialcells,whereassilencing
BreastCancerStemCells
Activation
Huang,
1
MinZheng,
1,3
YandanYao,
1,2
ShuoWei,
1
2
andKunPingLu
1,3,
MedicalCenter,HarvardMedicalSchool,Boston,
ity,Guangzhou510120,China
350108,China
HarvardMedicalSchool,Boston,MA02115,USA
mmons.org/licenses/by-nc-nd/3.0/).
b-catenin,havebeenidentified(ReyaandClevers,2005;Wellner
etal.,2009).Thesetranscriptionmodulatorsarefurtherregu-
latedbyupstreamsignalingpathways.Forexample,Erk
signalinghasbeenshowntoregulateBCSCsbyincreasingtran-
scriptionofZeb1andnuclearaccumulationofunphosphorylated
(active)b-catenin(Changetal.,2011;Shinetal.,2010).However,
regulatorypathwaysupstreamofErksignalingthatregulate
BCSCsarestillnotfullyunderstood.
AmongthesmallGTPasesuperfamily,Rashasbeenshownto
inducetheepithelial-mesenchymaltransition(EMT)andconfer
CSCtraitstobreastcellsinvitroandinvivo(Liuetal.,2009;
Shinetal.,2010).Rac1isinvolvedinCSCmaintenanceinnon-
smallcelllungadenocarcinomaandglioma,aswellasinintestinal
progenitorandstemcellexpansion(Akunuruetal.,2011;Myant
etal.,2013;Yoonetal.,2011).However,therolesofotherGTPase
familymembersinCSCsinsolidtumorsareyettobeelucidated.
Proteinphosphorylationoncertainserineorthreonineresi-
duesprecedingaproline(pSer/Thr-Pro)isacentralsignaling
mechanismincellproliferationandtransformation(Blume-Jen-
CellReports11,111–124,April7,2015a2015TheAuthors111
theERandGolgiapparatus,butwithnoknownfunctionincan-
cerorCSCs(Stenmark,2009;TisdaleandBalch,1996).We
showthatasaPin1transcriptionaltarget,Rab2AisaBCSC-
promotinggenethatenhancestumorigenesisviaactivatingErk
signaling.Thus,thePin1/Rab2A/ErkaxisdrivesBCSCexpan-
sionandtumorigenicity,providingattractivetargetsinBCSCs
forcancertherapy.
RESULTS
GenomicProfilingAnalysesIdentifiesRab2AasaPin1
DownstreamGene
WehavepreviouslydemonstratedafundamentalroleofPin1in
regulatinghumanBCSCsandmousemammarystemcells
(MaSCs)(Luoetal.,2014).Toelucidatetheunderlyingmolecular
mechanisms,weanalyzedtheeffectsofPin1knockout(KO)on
geneexpressioninmousemammaryepithelialcells(MECs).
GlobalexpressionprofilingofLin
C0
MECsfromPin1KOand
wild-type(WT)littermatesidentified1,723genesthatwere
downregulatedinbothPin1KOmice(Figures1Aand1B;Tables
S1andS2).TonarrowdownthelistofPin1-regulatedgenes,we
comparedMECgeneexpressionwiththatofneurospherespre-
paredfromthesamemice(Figure1B;TableS3).Although
comparingexpressionprofilesofstemcellsfromWTandPin1
KOmicemaybeabetterapproach,theMaSC-enriched
Lin
C0
CD24
+
CD29
+
orLin
C0
CD24
med
CD49f
hi
populationsarevery
smallinPin1KOmice(Luoetal.,2014),whichmadeitdifficult
togetenoughRNAfromeachmouse.Asanalternative
approach,were-analyzedtwopublishedexpressionprofiling
datasetsofmouseMaSCsandBCSCs(Stingletal.,2006;Zhang
etal.,2008)andcomparedthemwithourexpressionprofiling
ofLin
C0
MECsandneuronsfromPin1KOandWTmice.Weiden-
tified14genesdownregulatedinbothPin1KOMECsand
neuronsandupregulatedineitherMaSCsorBCSCs(Figure1C;
TableS4).
Tovalidatethesecandidategenes,weusedqRT-PCRtodeter-
minetheeffectsofPin1knockdown(KD)ontheirexpressioninsix
humanbreastcelllines.Rab2A,Lamp2,andMagi3weredown-
regulatedinmorethanfiveKDcelllines(Figures1DandS1A).
Pin1KDalsoreducedRab2Aproteininallsixdifferentcelllines
(FigureS1B).TotesttheeffectsofthesegenesinBCSCs,we
silencedtheirexpressionusingtwodifferentsmallhairpinRNAs
(shRNAs)inMCF10AcellsandexaminedtheCD24
C0
CD44
+
pop-
ulation,whichenrichedhumanBCSCs(Al-Hajjetal.,2003).
Rab2AKDconsistentlydecreasedtheCD24
C0
CD44
+
population
(FigureS1C),suggestingarequirementofRab2AforBCSCmain-
tenance.Thus,wefocusedonRab2AasaPin1target.
InthepromoterregionofRab2A,therearetwoputativeAP-1
bindingsites(C01,293andC0890)(FigureS1D).Notably,Pin1ac-
tivatestranscriptionfactorsc-Junandc-FostoincreaseAP-1
activity(Monjeetal.,2005;Wulfetal.,2001).Wethereforetested
whetherPin1regulatedRab2Atranscription.Inthereporter
assay,wefoundthatPin1overexpressionenhancedRab2A
transcriptioninadose-dependentmanner(Figure1E).Deletion
analysissuggestedthatPin1appearedtoactonthedistal
AP-1site,butnottheproximalsite(Figure1F).Toconfirmthat
Pin1regulatesRab2AtranscriptionthroughAP-1,wefirstexam-
inedwhetherPin1boundtoRab2Apromoterbychromatin
112CellReports11,111–124,April7,2015a2015TheAuthors
immunoprecipitation(ChIP)usingcellstransfectedwithPin1
expressionplasmid.ComparedtocontrolimmunoglobulinG,
anti-Pin1antibodiesshowedappreciablebindingtotheC01,293
locus,asassayedbyqPCR(Figure1G).Next,weusedac-Jun
antibodytoperformtheChIPassayinhumanmammarylarge-T
andhTERTtransformedepithelialcells(HMLE)-Rascells,
becausePin1bindstoc-Jun,whichisphosphorylatedbyJNK
andcooperateswithRastoincreasethetranscriptionalactivity
ofc-Juntowarditstargetgenes(Wulfetal.,2001).Indeed,
c-JunspecificallyassociatedwiththeC01,293locusinthe
Rab2Apromoter(Figure1H).Moreover,asequentialChIP(re-
ChIP)analysisusingc-JunantibodyfollowedbyPin1antibody
demonstratedthatbothproteinswerepresentinthesamecom-
plexontheC01,293locus(Figure1I).Thus,Pin1activatesRab2A
transcriptionthroughAP-1andincreasesitsproteinlevelsin
breastcancercells.
Rab2AIsaMajorMediatorofPin1inRegulatingBCSC
Function
ToinvestigatewhetherRab2Aisafunctionaldownstreamtarget
ofPin1,weknockeddownRab2AincontrolorPin1-overex-
pressingHMLEcellstoexaminewhetherRab2Amediatesthe
actionofPin1inBCSCs(Figure1J).Asshownpreviously(Luo
etal.,2014),Pin1overexpressiondrasticallyincreasedthepop-
ulationofBCSC-enrichedCD24
C0
CD44
+
cellsby8-to9-fold
abovethatofthevectorcontrol-infectedHMLEcells(Figures
1Kand1L).Rab2AKDgreatlyreducedthesizeofthe
CD24
C0
CD44
+
populationinvectorcontrolHMLEcells(Figures
1Kand1L),asdidPin1KD(Luoetal.,2014).InPin1-overex-
pressingcells,Rab2AKDpartiallydecreasedtheabundance
ofCD24
C0
CD44
+
cells(Figures1Kand1L).Wethenperformed
amammosphere-formingassay,whichmeasuresthefrequency
ofearlyprogenitor/stemcellsandBCSCs(DontuandWicha,
2005).Rab2AKDdecreasedmammosphereformationbyboth
vectorcontrolandPin1-overexpressingHMLEcells(Figure1M).
Thus,Rab2AwasrequiredtosustaintheBCSC-enrichedpopu-
lationincontrolcellsandPin1-overexpressingcells.
Recently,weshowedthatPin1overexpressioninducesEMT
inHMLEcells(Luoetal.,2014).Strikingly,Rab2AKDinPin1-
overexpressingcellsrevertedtheEMTphenotype.AfterRab2A
KD,Pin1-overexpressingHMLEcellschangedtotheepithelial
morphology(Figure1N),withincreasedE-cadherinand
decreasedN-cadherin,vimentin,andfibronectinlevels(Fig-
ure1O).Cellmigration,apropertyassociatedwiththeEMT,
wasalsogreatlyattenuatedbyRab2AKDinPin1-overexpress-
ingcellsinwound-healing(FigureS2A)andtranswellmigration
assays(FigureS2B).Thus,Rab2AisamajormediatorofPin1
inregulatingBCSCfunction.
Rab2AIsAmplifiedinHumanCancersandIts
OverexpressionIncreasestheBCSC-Enriched
PopulationandTumorigenicity
TodeterminemoredirectlytheroleofRab2Ainbreastcancer,
wefirstcheckedRab2AgenealterationsincancersinthecBio
CancerGenomicsPortal(Ceramietal.,2012).Significantly,
Rab2Ageneamplificationoccursinawiderangeofhumancan-
cers,withthehighestfrequencyofC249.5%(72of760)ininvasive
breastcarcinomapatients(Figure2A).Importantly,Rab2A
mRNAlevelsincreasesignificantlywithincreasinggenecopy
numberininvasivebreastcarcinomas(p=1.56EC084)andothers
(FigureS3A).Moreover,Rab2Aamplificationisindependentof
MYCon8q,accordingtoTumorscapesoftware(Beroukhim
etal.,2010).
WethenoverexpressedRab2AincontrolshRNAorPin1KD
HMLEcells(Figure2B).ModerateRab2Aoverexpression(two
tothreetimestheendogenouslevel)notonlystronglyincreased
Figure1.Genome-wideExpressionProfilingIdentifiesRab2AasaPin1
(A)HierarchicalclusterofthemicroarraydataoftheLin
C0
populationofmammary
(B)Genomicprofilingidentified14potentialtargetgenesthatweredownregulated
orBCSCs.
(C)Heatmapdepictingthefoldchangesof14candidategenes,whichweredownregulat
mouseMaSCsorBCSCs(presentedbySC/non-SCratio).
(D)Pin1KDreducedRab2AmRNAinhumanbreastcancerlines,asassayedby
(EandF)Pin1activatedtheRab2Apromoterinadose-dependentmanner.
(G–I)BothPin1andc-JunboundtotheRab2Apromoter,asshownbyChIPandRe-ChIP
controlandnormalizedwithsampleinputsofchromatinharvestedpriortoimmunoprecipit
(J)Rab2AwasknockeddowninvectorcontrolandPin1-overexpressingHMLE
(KandL)Rab2AKDinHMLEcellsreducedtheCD24
C0
CD44
+
populationand
population.
(M)Rab2AKDinHMLEcellsreducedmammosphere-formingactivityandimpaired
(NandO)Rab2AKDimpairedtheabilityofPin1overexpressiontoinducetheEMT
anddownregulationofN-cadherin,fibronectin,andvimentinbyqRT-PCR(O).Scale
Inallpanels,errorbarsrepresentSDofthreeindependentexperiments.Seealso
theCD24
C0
CD44
+
population(Figures2CandS3B)andmammo-
sphereformation(Figure2D)incontrolHMLEcellsbutalso
significantlyrescuedthedefectsinBCSC-associatedproperties
inPin1KDcells(Figures2Cand2D).LikePin1overexpression
(Luoetal.,2014),ectopicRab2AexpressionalsoinducedEMT
inHMLEcells,whichdevelopedanelongatedfibroblast-like
morphologywithdecreasedcell-cellcontact(Figure2E).The
EMTphenotypewasfurtherconfirmedbydecreasedE-cadherin
TargetthatPromotesBCSCExpansionandEMT
epithelialcellsintwopairsofWTandPin1KOlittermates.
inPin1KOMECsandneuroncells(NCs)butupregulatedinmouseMaSCs
edinPin1KOcells(presentedbyKO/WTratio)butupregulatedineither
real-timePCR.
analyses.Real-timePCRdatawerecalibratedtoimmunoglobulinG(IgG)
ation.
cells,asconfirmedbyimmunoblot.
suppressedtheabilityofPin1overexpressiontoincreasetheCD24
C0
CD44
+
theabilityofPin1overexpressiontoincreasemammosphere-formingactivity.
inHMLEcells,asshownbycellmorphology(N)orupregulationofE-cadherin
barrepresents100mm.
FiguresS1andS2.
CellReports11,111–124,April7,2015a2015TheAuthors113
andincreasedN-cadherin,vimentin,andfibronectinexpression
inRab2A-overexpressingcells(Figure2F).Rab2Aoverexpres-
sionalsoenhancedcellmigrationinwound-healing(Figures
S3CandS3D)andtranswell(FiguresS3EandS3F)assays.To
furtherinvestigatewhetherRab2Aissufficienttoinducetrans-
formation,weperformedsoft-agarcolonyformationassayon
Rab2A-overexpressingandcontrolvectorcells.Whereascon-
trolcellscouldhardlyformcolonies,Rab2A-overexpressing
Figure2.Rab2AIsAmplifiedinHumanCancers,ItsOverexpressionExpands
Q58HMutantHasReducedGTPaseActivity
(A)Rab2AgeneamplificationinawiderangeofhumancancersreportedincBioPort
(B)StableoverexpressionofRab2AinPin1KDorcontrolHMLEcellsusingretrovi
(C)OverexpressionofRab2AinHMLEcellspotentlyinducedtheCD24
C0
CD44
+
population
(D)OverexpressionofRab2Aincreasedthemammosphereformationincontrol
(EandF)OverexpressionofRab2ApotentlyinducedtheEMTinHMLEcells,
expressions(F).
(GandH)Rab2AoverexpressionincreasedtumorigenicityofBCSCs,whileits
BCSCs,asmeasuredbyalimiting-dilutiontumor-initiationassayinnudemice.Two
tumorincidence(H).
(I)Q58inRab2Aisevolutionallyconservedacrossspecies.
(JandK)TheQ58HmutantdisplayeddecreasedGTPhydrolysisactivity,compared
labeledGTPhydrolysis(J),andquantifiedbydensitometryofthreeindependent
(L)HMLE-RascellsinfectedwithRab2AQ58Hweremorepotentinforming
endogenouslevels.
Inallpanels,errorbarsrepresentSDofthreeindependentexperiments.Seealso
114CellReports11,111–124,April7,2015a2015TheAuthors
cellsrobustlyformedcolonies(FiguresS3GandS3H),further
supportingtheoncogenicactivityofRab2A.
ToevaluatetheimpactofRab2Aontumorinitiation,weas-
sessedtheeffectsofRab2Aoverexpressionontumorformation
bylimitingdilutiontransplantationassaysinnudemice.Weused
HMLERcells,HMLEcellstransformedwithV12H-Rastoformtu-
morsinnudemice(Elenbaasetal.,2001).Nomiceinoculated
with1310
4
controlHMLERcellsdevelopedtumors,while
BCSCsandTumorigenicity,andtheCancer-DerivedRab2A
al.
rus-mediatedgenetransfer,asdeterminedbyimmunoblot.
andrescuedthephenotypesinhibitedbyPin1KD.
shRNA(shCtrl)HMLEcellsandrescuedthephenotypesinhibitedbyPin1KD.
asassayedbycellmorphology(E)andreal-timeRT-PCRofthemarker
KDimpairedtheabilityofPin1overexpressiontoincreasetumorigenicityof
monthslater,miceweresacrificedandevaluatedfortumorweight(G)and
totheWTRab2AproteinintheinvitroGTPaseassay,asmonitoredbya-
32
P-
experiments(K).
tumorsthanthoseinfectedwithWTRab2Awhenoverexpressedatthe
FigureS3.
tumorsdevelopedinonlytwoofeightmiceinoculatedwith10
5
controlcellsandthreeofsixmiceinjectedwith10
6
controlcells
(Figures2Gand2H).ToexaminewhetherendogenousRab2Ais
necessaryforPin1topromotetumorigenicityofBCSCs,we
knockeddownRab2AinPin1-overexpressingHMLERcells.
Notumorsarosewhen1310
4
Pin1-expressingcellsinfected
withRab2AshRNA(shRab2A)wereinjectedintomice(Figures
2Gand2H).Althoughfourofeightmiceinoculatedwith10
5
Pin1-shRab2AHMLERcellsformedtumors,sevenmiceinjected
withanequalnumberofPin1cellsdevelopedtumors.Similarly,
withPin1overexpression,10
6
Rab2AKDcellsformedfewer
tumorsthancontrolPin1-overexpressingHMLERcells.Thus,
Rab2AinhibitionpotentlyimpairstheabilityofPin1topromote
BCSCtumorigenicity.
Rab2AIsMutatedinHumanCancersandtheQ58H
MutationActivatesRab2A
Inadditiontogeneamplification,theRab2AQ58Hmutationhas
beenrepeatedlyidentifiedinlungcancerpatientsinthecBio
CancerGenomicsPortal(Ceramietal.,2012).GiventhatQ58
ishighlyconservedinRab2Agenesacrossspecies(Figure2I)
andmostoftheoncogenicmutantsintheRassuperfamilyaffect
theenzyme’sabilitytohydrolyzeGTP(Schubbertetal.,2007),
weexaminedwhetherthismutationmightaffecttheintrinsic
abilityofRab2AtohydrolyzeGTP.Indeed,Rab2AQ58Hhydro-
lyzed[a-
32
P]GTPto[a-
32
P]GDPmoreslowlythantheWTprotein
(Figures2Jand2K),resultinginmoreproteinintheGTP-bound
state.
WethenaskedwhethertheQ58Hmutationmightincreasethe
potencyofRab2AtoexpandtheBCSC-enrichedpopulation.We
stablyexpressedFlag-Rab2Aanditsmutantusinglentiviruses
withalessoptimalKozaksequence,resultinginproteinsex-
pressedsimilartotheendogenouslevel(FigureS3I).Interest-
ingly,Rab2AincreasedtheCD24
C0
CD44
+
percentageto59%,
butRab2AQ58Hincreasedthispopulationto79%(FigureS3J).
ToexaminewhethertheQ58Hmutationincreasedtumorige-
nicity,weexaminedtumorformationbyinjecting1310
6
HMLER
cellsinfectedwithvectorcontrolorFlag-Rab2AandQ58H
mutantexpressingattheendogenouslevelintonudemice.
AlthoughcellsexpressingWTRab2AoritsQ58Hmutantformed
tumorsinallmice,theQ58Hmutanttumorsgrewsignificantly
fasterthanWTcontrols(Figures2LandS3K),suggestingthat
theRab2AQ58Hmutantismoreactiveinpromotingtumor
growththanWTprotein.
Erk1/2ActivationIsEssentialforRab2AtoRegulate
BCSCExpansion
TounderstandhowRab2AexpandstheBCSC-enrichedpopula-
tion,weexaminedwhetherRab2AactivatesErk1/2,whichis
crucialforRastoinduceEMTandincreasetheCD24
C0
CD44
+
population(Shinetal.,2010).Afterserumstarvationand
EGFstimulation,Rab2Aoverexpressionsignificantlyincreased
Erk1/2activationmonitoredbyp-Erk1/2inatime-dependent
mannerandalsoincreasedexpressionofZeb1(Figures3A
and3B),atranscriptionfactorcriticalforinducingtheEMTand
CD24
C0
CD44
+
population(Shinetal.,2010;Wellneretal.,
2009).Incontrast,Rab2AKDsubstantiallyimpairedErk1/2acti-
vation(Figures3Aand3B).WethenaskedwhethertheQ58H
mutationmightincreaseErk1/2phosphorylation.Whenex-
pressedattheendogenouslevel,theQ58Hmutantinduced
Erk1/2activationevenfasterthanWTRab2AafterEGFstimula-
tion(Figures3Cand3D).Thus,Rab2AanditsQ58Hmutant
promoteErk1/2activation.Next,wesilencedErk1or2in
Rab2A-overexpressingHMLEcells(Figure3E).Erk1KDonly
partiallyinhibited,butErk2KDsubstantiallydecreased,theabil-
ityofRab2Aoverexpressiontoinducemammosphereformation
(Figure3F)andtheCD24
C0
CD44
+
population(Figures3Gand
3H),indicatingthatRab2AactsthroughErk1/2tomaintain
BCSC-associatedproperties.
Rab2ADirectlyInteractswithandPreventsErk1/2
InactivationbyMKP3
ToelucidatehowRab2AoverexpressionoritsQ58Hmutation
activatesErk1/2,wefirstexaminedwhetherRab2Aco-localized
withErk1/2.OverexpressingWTRab2Anotonlyactivated
Erk1/2butalsosurprisinglycolocalizedwithactivatedErk1/2
attheperinuclearregionat5min(Figure4A)and1hr(FigureS4A)
afterEGFstimulation.OverexpressingRab2AQ58Hatlevels
similartotheendogenouslevelalsoactivatedandcolocalized
withErk1/2(Figures4AandS4A).Importantly,Rab2Aandits
Q58HmutantcolocalizedwithErk1/2attheERGIC,asassayed
byERGIC53staining(Figure4B).ToexaminewhetherRab2A’s
vesiculartraffickingfunctionisassociatedwithErkactivation,
weusedbrefeldinA(BFA)toblockthetraffickingfromtheERGIC
toER.Asshownpreviously(Haurietal.,2000),BFAtreatment
damagedtheERGICstructure(FigureS4B)butdidnotobviously
affectErkphosphorylation(FigureS4C),suggestingthatErk1/2
activationislikelytobeindependentofRab2A’strafficking
function.
TheunexpectedfindingthatRab2Acolocalizeswithactivated
Erk1/2attheERGICsuggestedthatRab2Amightdirectly
interactwithErk1/2.Indeed,wedetectedco-immunoprecipita-
tionoftheendogenousRab2AwithtotalErk1/2inHMLEcells
byreciprocalco-immunoprecipitation(coIP)(Figure4C).To
seewhetherRab2Ainteractedwithp-Erk1/2,wetransfected
constitutivelyactiveMEK1(AcMEK1)toincreasep-Erk1/2levels
andfoundthatRab2Ainteractedwithp-Erk1/2inthecoIPassay
(Figure4D).Theseresultswerefurthersupportedbyourfindings
thatRab2Acontainsaconservedcommondockingmotiffor
bindingErk(Tanoueetal.,2000)(Figure4E)andthatGST-
Rab2ApulleddownErk1/2incells(Figure4F),aswellasrecom-
binantErk1orErk2invitro(FigureS5A).Toexaminewhetherthe
integrityofthisdockingmotifisrequiredforRab2AtobindtoErk,
wesubstitutedtheknowncriticalresiduesKR(mut1)orLXI
(mut2)orbothresidues(mut1/2)withAlaresidues.Compared
toWTRab2A,whileeithermut1ormut2markedlyreducedbind-
ingtoErk,mutatingbothsequencescompletelyabolishedthe
abilityofRab2AtobindtoErk(Figure4F).Thus,Rab2Adirectly
interactswithErkthroughthespecificErkdockingsequencein
Rab2A.
Interestingly,theaboveconserveddockingmotifisalsofound
inMKP3andMEK1.ToexaminewhetherRab2AandMKP3or
MEK1competetointeractwithErk,HEK293cellswereco-trans-
fectedwithdecreasingdosesofmyc-MKP3ortheconstitutively
activehemagglutinin-MEK1andaconstantdoseofFlag-Rab2A.
WithdecreasingamountsofMKP3expressed,moreErk1/2were
CellReports11,111–124,April7,2015a2015TheAuthors115
Figure3.Rab2AandItsQ58HMutantDriveBCSCExpansionviaActivatingErk1/2
(AandB)Rab2AregulatedErk1/2phosphorylationanddownstreamZeb1expression.HMLEcellsstablyexpressingRab2AorshRNAorcontrolvectorswere
treatedwithEGFafterserumstarvationfortheindicatedtimepointstoactivateErk1/2andanalyzedbyimmunoblot.
(CandD)Rab2AQ58HmutantactivatedErk1/2fasterthanWTRab2AwhenoverexpressedattheendogenouslevelsafterEGFtreatment.Arrowhead,
exogenousFlag-Rab2A;arrow,endogenousRab2A.
(legendcontinuedonnextpage)
116CellReports11,111–124,April7,2015a2015TheAuthors
immunoprecipitatedbyRab2Ainadose-dependentmanner
(Figure4G),suggestingthatRab2AcompetedwithMKP3to
bindErk1/2invivo.However,unliketheMKP3competitionre-
sults,similaramountsofErk1/2wereimmunoprecipitatedby
Flag-Rab2AeventhoughdecreasingamountsofMEK1wereex-
pressed(FigureS5B).Theseresultsmaybeexpected,because
althoughthedockingmotifofMEKisimportantfortheERK-MEK
(E)Erk1orErk2wasknockeddownbytwoindependentlentivirus-mediatedshRNAs
(F)KDofErk1/2,especiallyErk2,preventedRab2Afromincreasingthemammos
(GandH)KDofErk1/2,especiallyErk2,preventedRab2AfromincreasingtheCD24
Inallpanels,errorbarsrepresentSDofthreeindependentexperiments.
Figure4.Rab2AInteractswithandPreventsInactivationofErk1/2by
(A)OverexpressedRab2AanditsQ58Hmutantco-localizedwithp-Erk1/2.Stable
Scalebarrepresents10mm.
(B)Rab2AanditsQ58Hmutantco-localizedwithERGIC53.Scalebarrepresents
(C)ReciprocalcoIPofendogenousRab2AwithErk1/2.LysatesofHMLEcellswere
blotforRab2AandErk1/2,respectively.
(D)Rab2AimmunoprecipitatedwithtotalErk1/2andp-Erk1/2inHEK293cellsco-transf
(E)TheconsensusErkdockingmotifswerefoundinRab2AandseveralotherErk
respectively.Xrepresentsanyaminoacids.
(F)MutationsintheErkdockingmotifinRab2AimpaireditsbindingtoErk1/2.
(G)Rab2AandMKP3competedtobindErk1/2.Lysatesof293Tcellstransfected
immunoprecipitatedwithM2(Flag)antibody,followedbywesternblotforErk1/2
(H)Rab2AcompetedwithMKP3andkeptErk1/2inthephosphorylatedstatus.
inducedp-Erk1/2inserum-starvedcellsandwaslargelyreversedbyMyc-MKP3expres
phosphorylation.
SeealsoFiguresS4andS5.
interaction,thereareothermechanismstoensuretheactivation
ofErkbyMEK1,suchasscaffoldproteinfacilitationandtheMEK
catalyticsiteinteractingwiththeErkactivationloop(Roskoski,
2012).
TheaboveresultssuggestthatRab2AmightpreventErk1/2
dephosphorylationbycompetingwithMKP3forErk1/2binding.
Totestthispossibility,wetransfectedHEK293cellswithMKP3
inRab2A-overexpressingcells.
phere-formingcapability.
C0
CD44
+
population.
MKP3
HMLEcellswereserumstarvedandthentreatedwith10ng/mlEGFfor5min.
20mm.
immunoprecipitatedwithRab2AorErk1/2antibodies,followedbywestern
ectedwithFlag-Rab2AandconstitutiveactivatedMEK1(AcMEK1).
bindingpartners.+and4representbasicandhydrophobicaminoacids,
withdecreasingdosesofmyc-MKP3andaconstantdoseofFlag-Rab2Awere
andFlag-Rab2A.
293TcellsweretransfectedtoexpressRab2A,MKP3,andAcMEK1,which
sion,whereasFlag-Rab2Aexpressiondose-dependentlyrestoredErk1/2
CellReports11,111–124,April7,2015a2015TheAuthors117
andtheconstitutivelyactiveMEK1aswellasdifferentamounts
ofepitope-taggedRab2A.ExpressionoftheactiveMEK1
inducedErk1/2phosphorylationeveninserum-starvedcells,
andthiswaslargelyreversedbymyc-MKP3expression(Fig-
ure4H).However,Flag-Rab2AexpressionrestoredErk1/2phos-
phorylationinadose-dependentmanner(Figure4H).Thus,
Rab2AdirectlybindstoErk1/2andkeepsitinanactiveform
bycompetingwithMKP3.
TofurtherdemonstratewhetherthisRab2A-Erkinteractionis
functionallyimportantforRab2AtoregulateBCSC,weinfected
HMLEcellswithFlag-taggedRab2Aoritsmutantsdefectivein
bindingtoErk(FigureS4C).Rab2Amarkedlyincreasedthe
CD24
C0
CD44
+
population,butnoneofitsRab2Amutantsaltered
thispopulation(FigureS5D).Inaddition,overexpressionof
Rab1A,thesmallGTPasethatishighlyrelatedtoRab2Awith
over70%similarityandalsolocalizedtotheERGICbutwithout
aconserveddockingmotifforbindingtoErk,hadnoeffecton
eitherErkactivationortheBCSCphenotype(FiguresS5G–
S5K).Thus,Rab2AdirectlyinteractswithandpreventsErk1/2
inactivationbyMKP3toregulateBCSCs.
Rab2APromotestheNuclearTranslocationofErk1/2
Downstreamb-Catenin
AsErk1/2signalingincreasesthenuclearaccumulationofun-
phosphorylated(active)b-catenin(Changetal.,2011),and
becausePin1alsohasasimilareffectonb-catenininbreastcan-
cercells(Ryoetal.,2001),weexaminedwhetherPin1/Rab2A/
p-Erksignalingregulatesnuclearb-cateninlevels.Confocal
analysisshowedthatmostunphosphorylatedb-cateninlocal-
izedattheplasmamembraneinstarvedHMLEcellsbuttranslo-
catedintothenucleus,alongwithincreasedp-Erk1/26hrafter
EGFstimulation(Figure5A).However,inRab2A-overexpressing
andPin1-overexpressingcells,notonlywasp-Erk1/2obviously
increasedbutalsounphosphorylatedb-cateninwasreadilyde-
tectedinthenucleusasearlyas2hrandaccumulatedfurther
withtimeafterEGFstimulation(Figures5Band5C).Incontrast,
inRab2AorPin1KDcells,notonlywasp-Erk1/2notincreased
butalsonuclearunphosphorylatedb-cateninwashardlydetect-
ableeven6hrafterstimulation(Figures5Dand5G).Notably,
overexpressionofRab2AinPin1KDcellscausedErk1/2
activationandnucleartranslocationand,importantly,unphos-
phorylatedb-cateninlocalizationtothenucleus(Figure5E).
Conversely,Rab2AKDinPin1-overexpressingcellsprevented
Erk1/2activationandnucleartranslocationofunphosphorylated
b-catenin(Figure5F).Westernblotanalysiswithnuclearfraction
furtherconfirmedtheseresults(Figure5H).Thesedatatogether
supportamodelinwhichthePin1/Rab2A/Erk1/2pathwayacti-
vatesb-catenin.
Rab2AEndowsBCSCTraitstoNormalHumanMECsand
IsNecessaryforTumorigenesisofHumanPrimary
BCSCs
Toextendourfindingstoprimaryhumancells,wesortedLin
C0
MECsisolatedfromreductionmammoplastytissuefromtwo
humandonorsandinfectedthemwithlentivirusesexpressing
Flag-Rab2A,Flag-Rab2AQ58Hatlevelssimilartoorthreetimes
theendogenouslevel(Figures6AandS6A).Rab2Aoverexpres-
sionledtoadose-dependentincreaseintheCD24
C0
CD44
+
pop-
118CellReports11,111–124,April7,2015a2015TheAuthors
ulation(Figure6B).OverexpressingRab2AQ58Hsimilartothe
endogenouslevelincreasedtheCD24
C0
CD44
+
populationmore
thanoverexpressingRab2Aatthree-times-higherlevels(Fig-
ure6B).Thus,increasingRab2Aactivitybyeitheroverexpres-
sionorusingnaturallyoccurringcancer-derivedmutation
endowsBCSCtraitstonormalhumanMECs.
ToassesswhetherRab2Aisalsoimportantfortumorigenesis
ofBCSCsinhumanprimarybreastcancer,wesorted
Lin
C0
CD24
C0
CD44
+
cellsfromfreshlyisolatedbreastcancercells
ofeightpatients(TableS5)andanalyzedRab2Aexpressionand
itsimpactonBCSCsinvitroandinvivo(FigureS6B).As
comparedwiththoseinLin
C0
non-CD24
C0
CD44
+
cancercells,
Rab2AmRNAlevelswereapproximatelyseventimeshigherin
Lin
C0
CD24
C0
CD44
+
cells,andfivetoseventimeslowerinnormal
breastepithelialcells(Figure6C).Consistentwiththeseresults,
Rab2Aproteinandunphosphorylatedb-cateninwerehigherin
theLin
C0
CD24
C0
CD44
+
cellsthanthoseinLin
C0
non-CD24
C0
CD44
+
cancercellsornormalMECs(Figure6D).Wethenknocked
downRab2AinLin
C0
CD24
C0
CD44
+
primarybreastcancercells
(Figure6E).TheCD24
C0
CD44
+
populationwassignificantly
reducedinRab2AKDcells,beingonlyone-ninthofthatincon-
trolcells(Figure6F).Rab2AKDalsosignificantlydecreasedthe
mammosphere-formingactivityoftheCD24
C0
CD44
+
cells(Fig-
ures6Gand6H).Thus,Rab2Aisrequiredforsustainingthe
BCSC-associatedpropertiesofhumanprimarybreastcancer
cells.
WefinallyinvestigatedwhetherRab2Awasrequiredforthe
tumorigenicityoftheLin
C0
CD24
C0
CD44
+
population.Weinjected
2,000controlorRab2AshRNA-transducedLin
C0
CD24
C0
CD44
+
cells,orLin
C0
non-CD24
C0
CD44
+
cellsisolatedfromeightbreast
cancerpatients,intonudemice.Whilenotumorsdevelopedin
miceinjectedwiththecellsthatwerenotCD24
C0
CD44
+
,2,000
controlLin
C0
CD24
C0
CD44
+
cellsgeneratedsixtumorsineightin-
jectedmice(Figures6I–6K).Lentivirus-mediatedKDofRab2A
notonlydrasticallyreducedtumorincidence(Figure6K)but
alsopotentlyreducedtumorgrowth,asmeasuredbytumorvol-
umesandweights(Figures6Iand6J).Wethendissociatedthe
tumorsandsortedagainforCD24
C0
CD44
+
cellsfortheserial
transplantation.Whencontroltumorswerepassagedinnude
mice,theywereseriallytransplantedatleastfortwomorepas-
sageswithoutreducedtumorigenicity(Figure6K),butRab2A
KDcellshadsubstantiallydecreasedfrequencyoftumorforma-
tionandreducedtumorgrowth(Figures6I–6K).Thus,Rab2Ais
highlyexpressedinBCSC-enrichedpopulationinhumanbreast
cancer,andsilencingRab2ApotentlyinhibitsBCSCexpansion
andtumorigenesis.
Rab2AOverexpressionCorrelateswithPoorClinical
OutcomesinBreastCancerPatients
WeanalyzedexpressionofRab2A,Pin1,andALDH1,amarker
forstemandprogenitorcellsaswellasBCSCs(Ginestieretal.,
2007),innormalandcancerousbreasttissuearraysusing
immunohistochemistry.Pin1andRab2Awereundetectable
orlowinall24humannormalbreasttissue,buttheirexpres-
sionwasdramaticallyincreasedinmanyof65humanbreast
cancertissue(Figures7Aand7B).Remarkably,Rab2Aexpres-
sionwashighlycorrelatedwithPin1andALDH1expressionin
normalandcancerousbreasttissue(Figures7A–7C).Wenext
analyzedthecorrelationofRab2Aexpressionandclinical
outcomeinthesubsetof52breastcancerpatients,forwhich
clinicaldatawereavailable.HigherRab2Aexpressionwas
significantlyassociatedwithhighermortalityinbreastcancer
patients,asshownbyKaplan-Meiersurvivalcurves(p=
0.012)(Figure7D).
Toexpandourimmunohistochemistryfindingsonlimitedsam-
ples,westudiedmultipleindependentbreastcancerdatasets
fromOncomine(Rhodesetal.,2007),whichcollectivelylinkclin-
icaldatawithRab2AmRNAexpressioninC243,000patients.
Rab2Aoverexpressionwascloselyassociatedwithadvanced
stageintheBittnerdataset,withmetastasisintheSchmidtdata-
set,andwithdeathat3or5yearsintheBildandKaodatasets
Figure5.Rab2A/p-ErkSignalingPromotestheNuclearTranslocation
(AandB)Rab2Apromotedthenucleartranslocationofunphosphorylatedb-catenin
theindicatedtimepoints.
(C–E)Pin1alsopromotedthenucleartranslocationofunphosphorylatedb-catenin
b-catenintranslocationfromthecellmembranetothenucleus.
(FandG)Rab2AKDinPin1-overexpressingorvectorcontrolcellsinhibitedp-Erk1/2
(H)Rab2Apromotedthenuclearaccumulationofp-Erk1/2andunphosphorylated
followingserumstarvation.
(Bildetal.,2006;Kaoetal.,2011;Schmidtetal.,2008)(Figures
S7A–S7D).
Givingthatthemicroarrayexperimentsandthemethodsto
normalizedatavaryamongdifferentdatasets,makingitdifficult
topoolthedatafromdifferentdatasets,wechosetofurther
analyzetheCurtisdataset,whichhasover2,000patients(Curtis
etal.,2012).TheRab2AmRNAlevelwasastrongprognostic
factorforsurvivalbyunivariateCoxregressionanalysis(Fig-
ure7E).AhighRab2Alevelwasstillindependentlyassociated
withhighmortality,evenusingmultivariateanalysisadjusted
forproliferationmarkers(MKI67andPCNA),ortumorgrade
andstage,orthestatusofHER2,ER,andPR(Figure7E).
WenextanalyzedRab2AexpressioninthePAM50intrinsic
ofActiveb-Catenin
(activeform).HMLEcellswereserumstarvedandthenstimulatedbyEGFfor
andRab2AoverexpressioninPin1KDcellsrescuedErk1/2activationand
activationandb-cateninnucleartranslocation.Scalebarsrepresent10mm.
b-catenin.NuclearandtotalproteinswereextractedafterEGFstimulation
CellReports11,111–124,April7,2015a2015TheAuthors119
subtypes(Parkeretal.,2009)andintegrativesubgroups(Curtis
etal.,2012).Strikingly,highRab2Alevelswerefoundinthe
poor-prognosissubtypes,definedasPAM50intrinsicsubtypes,
luminalB,HER2-enriched,andbasal-like,andintheIntClust5,
IntClust6,IntClust9,andIntClust10integrativesubgroups(Fig-
ures7Fand7G),whereaslowRab2Alevelsweremostly
observedinthebetter-prognosissubtypes(normal-likePAM50
intrinsicsubtypeandintegrativesubgroupsIntClust3and
IntClust4)(Figures7Fand7G).Notably,ahighRab2Alevel
wastightlylinkedtohighmortalityinthemostcommonsub-
groupsofbreastcancerpatients,definedasHER2-negativeor
non-TNBC(triple-negative)patients(Figure7H),whichaccount
for87.5%and87.3%ofallcases,respectively.Thus,Rab2A
playsakeyoncogenicroleinpromotingBCSCsandaggravating
breastcancermalignancy.
Figure6.Rab2AandItsQ58HMutantEndowBCSCTraitstoNormal
HumanBCSCExpansionandTumorigenesis
(A)Westernblotshowedlentivirus-mediatedoverexpressionofRab2AandQ58H
taggedprotein;arrow,endogenousprotein.
(B)Rab2AorRab2AQ58HmutantincreasedtheCD24
C0
CD44
+
populationinprimar
(C)Real-timePCRshowedthatexpressionofRab2AmRNAwasmarkedlyincreased
normalepithelialcells.
(D)ExpressionofRab2Aandunphosphorylatedb-cateninproteinwasmarkedlyincreased
humanbreastcancertissueandthoseinnormalbreasttissuefromthesamepatient.
(E)Rab2AwasknockeddowninLin
C0
CD24
C0
CD44
+
cellssortedfromhumanbreast
(F)Rab2AKDinLin
C0
CD24
C0
CD44
+
breastcancercellsdecreasedtheCD24
C0
CD44
(GandH)Rab2AKDinLin
C0
CD24
C0
CD44
+
breastcancercellsdecreasedmammos
(I–K)Rab2AKDinterferedwithbothtumorinitiationandgrowthofprimaryBCSC
incidence(K).2,000lentivirus-transducedLin
C0
CD24
C0
CD44
+
cellsisolatedfromeight
nudemice.P0,freshlyisolatedprimarycells;P1,passage1;P2,passage2.
In(C)and(H),errorbarsrepresentSDofthreeindependentexperiments.In(I)and
120CellReports11,111–124,April7,2015a2015TheAuthors
DISCUSSION
OurprofilingofPin1-regulatedgenesledtothediscoverythat
Pin1promotesBCSCsbyincreasingRab2Atranscription.We
suggestthatinbreastcancer,Pin1overexpression,Rab2A
geneamplification,orgeneticmutationactivatesRab2A,which
drivesBCSCsandtumorigenesisinlargepartbypreventing
Erk1/2inactivationviaMKP3,leadingtoactivationofknown
BCSCregulatorsandcontributingtohighmortalityinbreastcan-
cerpatients(Figure7I).ThispathwaymayofferattractiveBCSC
targetsforcancertherapy.Notably,ATRAhasbeenidentifiedto
bindandultimatelydegradeactivePin1selectivelyincancer
cells(datanotshown).
EmergingevidencehasshownthatRabfamilymembers
playimportantrolesinepithelialneoplasia(Goldenring,2013).
PrimaryHumanMECs,whereasSilencingRab2AInhibitsPrimary
mutantintwocasesofhumannormalLin
C0
MECs.Arrowhead,exogenousFlag
yhumanMECs.
intheLin
C0
CD24
C0
CD44
+
population,comparedtoLin
C0
non-CD24
C0
CD44
+
or
inLin
C0
CD24
C0
CD44
+
cellscomparedtoLin
C0
non-CD24
C0
CD44
+
cellsin
cancertissue.
+
population.
phereformation.Scalebarrepresents100mm.
sinvivo,asshownbytumorgrowthcurve(I),tumorweights(J),andtumor
breastcancerpatientswereseriallytransplantedasxenograftsintoeight
(J),errorbarsrepresentSDofeightmice.SeealsoFigureS6.
Figure7.Rab2AExpressionCorrelateswithPin1andALDH1ExpressionandwithPoorClinicalOutcomeinBreastCancerPatients
(A–C)Rab2AexpressioncorrelatedwithPin1andALDH1expressioninthetissuearrayofnormalandcancerousbreasttissue.
(D)HighRab2Aexpressioncorrelatedwithpooroverallsurvivalinthetissuearraydatasetofbreastcancerpatients.
(E)Rab2AwasastrongandindependentbiomarkertopredictbreastcancerspecificsurvivalintheCurtisbreastcancerdatasetbyCoxregressionanalyses.
(F)BoxplotsofRab2AexpressionstratifiedbythePAM50classifierintheCurtisbreastcancerdataset.
(G)BoxplotsofRab2AexpressionstratifiedbytheIntClustsubtypesintheCurtisbreastcancerdataset.
(legendcontinuedonnextpage)
CellReports11,111–124,April7,2015a2015TheAuthors121
AlthoughRab25hasopposingfunctionsincancer,promotingtu-
moraggressiveness(Chengetal.,2004)andsuppressingtumor
development(Nametal.,2010;Tongetal.,2012),mostRab
memberswithincreasedexpressionincancerarepositively
associatedwithcancerprogression,especiallywithinvasion
andmetastasis(Houetal.,2008;Jacobetal.,2013).Wereport
herethatRab2Aplaysamajoroncogenicroleinhumanbreast
cancerviapromotingBCSCexpansionandtumorigenicity.
Furthermore,wehaveprovidedstronggeneticevidencethat
Rab2Aisabnormallyactivatedinhumancancerbyeithergene
amplificationoractivatingRab2Apointmutation,bothofwhich
enhanceitsabilitytoactivateErk1/2signaling.Finally,gene
expressionprofilingdatainfivelargedatasetscontaining
C243,000breastcancerpatientsstronglylinkRab2Atopoorpa-
tientoutcome.Notably,inthe>2,000-patientCurtisdataset,
highRab2Alevelswerefoundinpoor-prognosissubtypes,
whereaslowRab2Alevelsweremostlyobservedinbetter-prog-
nosissubtypes.Significantly,ahighRab2Alevelwasstrongly
associatedwithhighmortality,eveninmultivariateanalysis.It
isworthnotingthattheassociationbetweenRab2Aandcancer
mortalitywasevenmorestrikinginthemostcommongroups
(C2487%)ofbreastcancerpatients,definedasHER2-negative
ornon-TNBCpatients.Sinceitischallengingtopredictpoor
prognosisinthesecommonsubgroupsofbreastcancerpatients
withoutprofilingmanygenes,Rab2Aexpressionmightrepresent
ausefulprognosisbiomarker.
Rab2Ahaspreviouslybeenimplicatedinintercellularvesicle
traffickingregulation(Stenmark,2009;TisdaleandBalch,
1996).OurfindingshighlighttheroleofRabGTPasesinaltering
cellsignaling.Previously,Rab5wasshowntotransducecell-sur-
vivalsignalsbyfacilitatingvesiculartranslocationtothenucleus
ofmultiplesignalingkinasesinneurons,includingAktandErk
(Delcroixetal.,2003).TheendosomalRab23isanessential
negativeregulatorthatattenuatesthemousesonichedgehog
signalingpathwaydownstreamofthesonichedgehogreceptor
anditseffector(Eggenschwileretal.,2001).Ourresultsthat
ErksignalingiscriticalforRab2AtoregulateBCSCsarereminis-
centofpreviousreportsshowingthattheErkpathwayisrequired
forBCSCmaintenance.Forexample,Rassignalsthroughthe
Erkpathway(Chiuetal.,2002),andErk2activationisessential
forRastoinducetheEMTandtopromotetheCD24
C0
CD44
+
population(Shinetal.,2010).SHP2influencesBCSCsand
enhancesbreasttumormaintenanceandprogressionthrough
activationoftheErkpathway(Acetoetal.,2012).
ItisinterestingtonotethatthedockingmotifinErk1/2iscom-
montoitsactivators,substrates,andregulators,suggestingthat
theymaycompeteforbindingtoErk1/2invivo(Bardwelletal.,
2003;Tanoueetal.,2000).Indeed,thesedockinginteractions
arecrucialforthespecificityofErk1/2signalingandtheduration
oftheactivation-inactivationcyclesofErk1/2.Intheabsenceof
activation,Erk1/2ismainlylocalizedtothecytoplasm.Activated
Erk1/2translocatestothenucleusandphosphorylatesnuclear
targets.MKP3shuttlesbetweenthecytoplasmandnucleus
(H)UnivariateCoxregressionanalysisshowedthatHER2-negative,non-triple-negative,
mRNAlevelshadahigherriskofbreastcancermortality.
(I)AschematicmodelforhowthePin1/Rab2A/Erksignalpathwayregulatestumor
SeealsoFigureS7.
122CellReports11,111–124,April7,2015a2015TheAuthors
butislargelylocalizedinthecytoplasm(Karlssonetal.,2004).
ItispossiblethatErk1/2boundbyRab2Aistemporarilypro-
tectedfromtheactionofthephosphataseMKP3.Thus,the
competitivedockingofRab2AandphosphatasetoErk1/2may
playanimportantroleinenhancingErksignaling.Moreover,
wefoundthatErk1/2andRab2AcolocalizedattheERGIC.
AlthoughErk3isprincipallylocalizedtotheGolgiandERGICin
severalcelllines(Bindetal.,2004),theroleofErksattheERGIC
islargelyunknown.Itwouldbeinterestingtodefinethemolecu-
lardetailsoftheErk1/2andRab2Ainteractionsandtheirregula-
tionattheERGICforErksignaling.
SmallGTPasesactthroughcyclingbetweenanactive,GTP-
boundandaninactive,GDP-boundstate.ActivatingRasmuta-
tionsoccurinC2430%ofhumancancers.Mostoftheactivating
mutationsimpairintrinsicGTPaseactivityorconferresistance
toGTPaseactivatingproteins(GAPs),causingmutantRasto
accumulateintheactiveGTP-boundform(Schubbertetal.,
2007).TheRab2AQ58Hmutation,anaturallyoccurringgenetic
mutationfoundinhumancancers,notonlyreducedGTPhy-
drolysisbutalsowasmorepotentinactivatingErksignaling
anddrivingBCSCsthantheWTprotein.Therefore,theQ58H
proteinmightbeahyperactiveformofRab2Ainhuman
cancers.
EXPERIMENTALPROCEDURES
MicroarrayAnalysis
RNAfromLin
C0
MECsandneuroncellsofPin1KOandWTmicewasextracted
withthetotalRNAisolationkit(Agilent).Microarrayexpressionprofileswere
collectedusingAffymetrixGeneChipMouseExpressionArray430A.
SerialTransplantationAssay
Lin
C0
CD24
C0
CD44
+
cellsweresortedfromeightbreastcancerspecimensand
culturedasasingle-cellsuspensioninultra-low-attachmentdishes,thenin-
fectedwithlentivirusexpressingcontrolvectororRab2AshRNA.After
1weekofpuromycinselection,2,000transducedcellsfromeachpatient
wereinjectedintothemammaryfatpadsof5-week-oldnudemice.For
serialpassaging,cellsfromtheprimarytumorsweresortedagainfor
Lin
C0
CD24
C0
CD44
+
cells.Amongthesixprimarytumorsformedinthecontrol
shRNAgroup,fourtumorswererandomlyselectedandpassagedintoeight
mice(twomicepertumor).Fortheonetumorformedfrom2,000shRab2A
cells,thistumor’scellswereinjectedintoeightmiceforserialpassaging.
Thesameprocedurewasappliedtothesecondpassageofxenograftcells,
asdescribedpreviously(Yuetal.,2007).Allstudiesinvolvinghumansubjects
wereapprovedbytheinstitutionalreviewboardatBethIsraelDeaconess
MedicalCenterorSunYat-SenMemorialHospital.Allstudiesinvolvingmice
wereapprovedbytheinstitutionalanimalcareandusecommitteeatBeth
IsraelDeaconessMedicalCenterandperformedinaccordancewiththerele-
vantprotocols.
StatisticalAnalysis
Alldataarepresentedasthemeans±SD,followedbydeterminingsignificant
differencesusingthetwo-tailedttestorANOVAtest.Limiting-dilutiondata
wereanalyzedbythesingle-hitPoissonmodelusingacomplementarylog-
loggeneralizedlinearmodel(Bonnefoixetal.,2001)withL-CalcSoftware
(STEMCELLTechnologies).CorrelationsofRab2Aexpressionwithother
geneexpressionwereanalyzedwiththePearsoncorrelationtest.Forsurvival
orPAM50-normalsubtypesofbreastcancerpatientswithhigherRab2A
initiationviaCSCregulators,contributingtohighmortalityinbreastcancer.
analysis,Kaplan-MeieranalysisandunivariateandmultivariateCoxregression
analysiswereused.Alltestsofsignificanceweresetatp<0.05.
ACCESSIONNUMBERS
MicroarraydatahavebeendepositedtotheNCBIGEOandareavailableunder
accessionnumberGSE49971.
SUPPLEMENTALINFORMATION
SupplementalInformationincludesSupplementalExperimentalProcedures,
sevenfigures,andfivetablesandcanbefoundwiththisarticleonlineat
http://dx.doi.org/10.1016/j.celrep.2015.03.002.
ACKNOWLEDGMENTS
WeareverygratefultoDrs.RobertWeinbergandWenjunGuoforproviding
expertadviceandcriticalreagents,andwethankDrs.JohnBlenisand
WilliamC.Hahnforreagentsand/oradviceandmembersofK.P.L.and
X.Z.Z.laboratoriesforconstructivediscussions.C.-H.C.isaDODBreastCan-
cerResearchProgramPostdoctoralFellowandaNIHT32traininggrant
awardee.ThisworkwassupportedbyaKomenforCuregrant(toX.Z.Z.),
theMinistryofScienceandTechnologyofChina973project(toE.S.),and
theNationalNaturalScienceFoundationofChina(U1205024)andNIH(grants
CA167677,AG039405,andDA031663toK.P.L.).K.P.L.andX.Z.Z.areinven-
torsofPin1technology,whichwaslicensedbyBIDMCtoPinteonTherapeu-
tics.BothownequityinandconsultforPinteon.K.P.L.alsoservesonitsboard
ofdirectors.TheirinterestswerereviewedandaremanagedbyBIDMCin
accordancewithitsconflictofinterestpolicy.
Received:August7,2014
Revised:January14,2015
Accepted:February26,2015
Published:March26,2015
REFERENCES
Aceto,N.,Sausgruber,N.,Brinkhaus,H.,Gaidatzis,D.,Martiny-Baron,G.,
Mazzarol,G.,Confalonieri,S.,Quarto,M.,Hu,G.,Balwierz,P.J.,etal.
(2012).TyrosinephosphataseSHP2promotesbreastcancerprogression
andmaintainstumor-initiatingcellsviaactivationofkeytranscriptionfactors
andapositivefeedbacksignalingloop.Nat.Med.18,529–537.
Akunuru,S.,Palumbo,J.,Zhai,Q.J.,andZheng,Y.(2011).Rac1targetingsup-
presseshumannon-smallcelllungadenocarcinomacancerstemcellactivity.
PLoSONE6,e16951.
Al-Hajj,M.,Wicha,M.S.,Benito-Hernandez,A.,Morrison,S.J.,andClarke,
M.F.(2003).Prospectiveidentificationoftumorigenicbreastcancercells.
Proc.Natl.Acad.Sci.USA100,3983–3988.
Bardwell,A.J.,Abdollahi,M.,andBardwell,L.(2003).Dockingsiteson
mitogen-activatedproteinkinase(MAPK)kinases,MAPKphosphatasesand
theElk-1transcriptionfactorcompeteforMAPKbindingandarecrucialfor
enzymicactivity.Biochem.J.370,1077–1085.
Beroukhim,R.,Mermel,C.H.,Porter,D.,Wei,G.,Raychaudhuri,S.,Donovan,
J.,Barretina,J.,Boehm,J.S.,Dobson,J.,Urashima,M.,etal.(2010).Theland-
scapeofsomaticcopy-numberalterationacrosshumancancers.Nature463,
899–905.
Bild,A.H.,Yao,G.,Chang,J.T.,Wang,Q.,Potti,A.,Chasse,D.,Joshi,M.B.,
Harpole,D.,Lancaster,J.M.,Berchuck,A.,etal.(2006).Oncogenicpathway
signaturesinhumancancersasaguidetotargetedtherapies.Nature439,
353–357.
Bind,E.,Kleyner,Y.,Skowronska-Krawczyk,D.,Bien,E.,Dynlacht,B.D.,and
Sa′nchez,I.(2004).Anovelmechanismformitogen-activatedproteinkinase
localization.Mol.Biol.Cell15,4457–4466.
Blume-Jensen,P.,andHunter,T.(2001).Oncogenickinasesignalling.Nature
411,355–365.
Bonnefoix,T.,Bonnefoix,P.,Callanan,M.,Verdiel,P.,andSotto,J.J.(2001).
Graphicalrepresentationofageneralizedlinearmodel-basedstatisticaltest
estimatingthefitofthesingle-hitPoissonmodeltolimitingdilutionassays.
J.Immunol.167,5725–5730.
Cerami,E.,Gao,J.,Dogrusoz,U.,Gross,B.E.,Sumer,S.O.,Aksoy,B.A.,
Jacobsen,A.,Byrne,C.J.,Heuer,M.L.,Larsson,E.,etal.(2012).ThecBiocan-
cergenomicsportal:anopenplatformforexploringmultidimensionalcancer
genomicsdata.CancerDiscov2,401–404.
Chang,C.J.,Yang,J.Y.,Xia,W.,Chen,C.T.,Xie,X.,Chao,C.H.,Woodward,
W.A.,Hsu,J.M.,Hortobagyi,G.N.,andHung,M.C.(2011).EZH2promotes
expansionofbreasttumorinitiatingcellsthroughactivationofRAF1-b-catenin
signaling.CancerCell19,86–100.
Chen,C.H.,Chang,C.C.,Lee,T.H.,Luo,M.,Huang,P.,Liao,P.H.,Wei,S.,Li,
F.A.,Chen,R.H.,Zhou,X.Z.,etal.(2013).SENP1deSUMOylatesandregulates
Pin1proteinactivityandcellularfunction.CancerRes.73,3951–3962.
Cheng,K.W.,Lahad,J.P.,Kuo,W.L.,Lapuk,A.,Yamada,K.,Auersperg,N.,
Liu,J.,Smith-McCune,K.,Lu,K.H.,Fishman,D.,etal.(2004).TheRAB25
smallGTPasedeterminesaggressivenessofovarianandbreastcancers.
Nat.Med.10,1251–1256.
Chiu,V.K.,Bivona,T.,Hach,A.,Sajous,J.B.,Silletti,J.,Wiener,H.,Johnson,
R.L.,2nd,Cox,A.D.,andPhilips,M.R.(2002).Rassignallingontheendo-
plasmicreticulumandtheGolgi.Nat.CellBiol.4,343–350.
Curtis,C.,Shah,S.P.,Chin,S.F.,Turashvili,G.,Rueda,O.M.,Dunning,M.J.,
Speed,D.,Lynch,A.G.,Samarajiwa,S.,Yuan,Y.,etal.;METABRICGroup
(2012).Thegenomicandtranscriptomicarchitectureof2,000breasttumours
revealsnovelsubgroups.Nature486,346–352.
Delcroix,J.D.,Valletta,J.S.,Wu,C.,Hunt,S.J.,Kowal,A.S.,andMobley,W.C.
(2003).NGFsignalinginsensoryneurons:evidencethatearlyendosomes
carryNGFretrogradesignals.Neuron39,69–84.
Dontu,G.,andWicha,M.S.(2005).Survivalofmammarystemcellsinsuspen-
sionculture:implicationsforstemcellbiologyandneoplasia.J.Mammary
GlandBiol.Neoplasia10,75–86.
Eggenschwiler,J.T.,Espinoza,E.,andAnderson,K.V.(2001).Rab23isan
essentialnegativeregulatorofthemouseSonichedgehogsignallingpathway.
Nature412,194–198.
Elenbaas,B.,Spirio,L.,Koerner,F.,Fleming,M.D.,Zimonjic,D.B.,Donaher,
J.L.,Popescu,N.C.,Hahn,W.C.,andWeinberg,R.A.(2001).Humanbreast
cancercellsgeneratedbyoncogenictransformationofprimarymammary
epithelialcells.GenesDev.15,50–65.
Ginestier,C.,Hur,M.H.,Charafe-Jauffret,E.,Monville,F.,Dutcher,J.,Brown,
M.,Jacquemier,J.,Viens,P.,Kleer,C.G.,Liu,S.,etal.(2007).ALDH1isa
markerofnormalandmalignanthumanmammarystemcellsandapredictor
ofpoorclinicaloutcome.CellStemCell1,555–567.
Goldenring,J.R.(2013).Acentralroleforvesicletraffickinginepithelial
neoplasia:intracellularhighwaystocarcinogenesis.Nat.Rev.Cancer13,
813–820.
Hauri,H.P.,Kappeler,F.,Andersson,H.,andAppenzeller,C.(2000).ERGIC-53
andtrafficinthesecretorypathway.J.CellSci.113,587–596.
Hou,Q.,Wu,Y.H.,Grabsch,H.,Zhu,Y.,Leong,S.H.,Ganesan,K.,Cross,D.,
Tan,L.K.,Tao,J.,Gopalakrishnan,V.,etal.(2008).Integrativegenomicsiden-
tifiesRAB23asaninvasionmediatorgeneindiffuse-typegastriccancer.Can-
cerRes.68,4623–4630.
Jacob,A.,Jing,J.,Lee,J.,Schedin,P.,Gilbert,S.M.,Peden,A.A.,Junutula,
J.R.,andPrekeris,R.(2013).Rab40bregulatestraffickingofMMP2and
MMP9duringinvadopodiaformationandinvasionofbreastcancercells.
J.CellSci.126,4647–4658.
Kao,K.J.,Chang,K.M.,Hsu,H.C.,andHuang,A.T.(2011).Correlationof
microarray-basedbreastcancermolecularsubtypesandclinicaloutcomes:
implicationsfortreatmentoptimization.BMCCancer11,143.
Karlsson,M.,Mathers,J.,Dickinson,R.J.,Mandl,M.,andKeyse,S.M.(2004).
Bothnuclear-cytoplasmicshuttlingofthedualspecificityphosphataseMKP-3
anditsabilitytoanchorMAPkinaseinthecytoplasmaremediatedbya
conservednuclearexportsignal.J.Biol.Chem.279,41882–41891.
CellReports11,111–124,April7,2015a2015TheAuthors123
Lee,T.H.,Chen,C.H.,Suizu,F.,Huang,P.,Schiene-Fischer,C.,Daum,S.,
Zhang,Y.J.,Goate,A.,Chen,R.H.,Zhou,X.Z.,andLu,K.P.(2011).Death-
associatedproteinkinase1phosphorylatesPin1andinhibitsitsprolylisom-
eraseactivityandcellularfunction.Mol.Cell42,147–159.
Liu,S.,andWicha,M.S.(2010).Targetingbreastcancerstemcells.J.Clin.
Oncol.28,4006–4012.
Liu,M.,Casimiro,M.C.,Wang,C.,Shirley,L.A.,Jiao,X.,Katiyar,S.,Ju,X.,Li,
Z.,Yu,Z.,Zhou,J.,etal.(2009).p21CIP1attenuatesRas-andc-Myc-depen-
Ryo,A.,Nakamura,M.,Wulf,G.,Liou,Y.C.,andLu,K.P.(2001).Pin1regulates
turnoverandsubcellularlocalizationofbeta-cateninbyinhibitingitsinteraction
withAPC.Nat.CellBiol.3,793–801.
Schmidt,M.,Bo¨hm,D.,vonTo¨rne,C.,Steiner,E.,Puhl,A.,Pilch,H.,Lehr,H.A.,
Hengstler,J.G.,Ko¨lbl,H.,andGehrmann,M.(2008).Thehumoralimmunesys-
temhasakeyprognosticimpactinnode-negativebreastcancer.CancerRes.
68,5405–5413.
dentbreasttumorepithelialmesenchymaltransitionandcancerstemcell-like
geneexpressioninvivo.Proc.Natl.Acad.Sci.USA106,19035–19039.
Lu,Z.,andHunter,T.(2014).ProlylisomerasePin1incancer.CellRes.24,
1033–1049.
Lu,K.P.,andZhou,X.Z.(2007).TheprolylisomerasePIN1:apivotalnewtwist
inphosphorylationsignallinganddisease.Nat.Rev.Mol.CellBiol.8,904–916.
Lu,K.P.,Hanes,S.D.,andHunter,T.(1996).Ahumanpeptidyl-prolylisom-
eraseessentialforregulationofmitosis.Nature380,544–547.
Luo,M.L.,Gong,C.,Chen,C.H.,Lee,D.Y.,Hu,H.,Huang,P.,Yao,Y.,Guo,W.,
Reinhardt,F.,Wulf,G.,etal.(2014).ProlylisomerasePin1actsdownstreamof
miR200ctopromotecancerstem-likecelltraitsinbreastcancer.CancerRes.
74,3603–3616.
Monje,P.,Herna′ndez-Losa,J.,Lyons,R.J.,Castellone,M.D.,andGutkind,
J.S.(2005).Regulationofthetranscriptionalactivityofc-FosbyERK.Anovel
rolefortheprolylisomerasePIN1.J.Biol.Chem.280,35081–35084.
Myant,K.B.,Cammareri,P.,McGhee,E.J.,Ridgway,R.A.,Huels,D.J.,Cor-
dero,J.B.,Schwitalla,S.,Kalna,G.,Ogg,E.L.,Athineos,D.,etal.(2013).
ROSproductionandNF-kBactivationtriggeredbyRAC1facilitateWNT-
drivenintestinalstemcellproliferationandcolorectalcancerinitiation.Cell
StemCell12,761–773.
Nakamura,K.,Greenwood,A.,Binder,L.,Bigio,E.H.,Denial,S.,Nicholson,L.,
Zhou,X.Z.,andLu,K.P.(2012).Prolineisomer-specificantibodiesrevealthe
earlypathogenictauconformationinAlzheimer’sdisease.Cell149,232–244.
Nam,K.T.,Lee,H.J.,Smith,J.J.,Lapierre,L.A.,Kamath,V.P.,Chen,X.,Aro-
now,B.J.,Yeatman,T.J.,Bhartur,S.G.,Calhoun,B.C.,etal.(2010).Lossof
Rab25promotesthedevelopmentofintestinalneoplasiainmiceandisasso-
ciatedwithhumancolorectaladenocarcinomas.J.Clin.Invest.120,840–849.
Parker,J.S.,Mullins,M.,Cheang,M.C.,Leung,S.,Voduc,D.,Vickery,T.,
Davies,S.,Fauron,C.,He,X.,Hu,Z.,etal.(2009).Supervisedriskpredictor
ofbreastcancerbasedonintrinsicsubtypes.J.Clin.Oncol.27,1160–1167.
Reya,T.,andClevers,H.(2005).Wntsignallinginstemcellsandcancer.
Nature434,843–850.
Rhodes,D.R.,Kalyana-Sundaram,S.,Mahavisno,V.,Varambally,R.,Yu,J.,
Briggs,B.B.,Barrette,T.R.,Anstet,M.J.,Kincead-Beal,C.,Kulkarni,P.,
etal.(2007).Oncomine3.0:genes,pathways,andnetworksinacollection
of18,000cancergeneexpressionprofiles.Neoplasia9,166–180.
Roskoski,R.,Jr.(2012).ERK1/2MAPkinases:structure,function,andregula-
tion.Pharmacol.Res.66,105–143.
Rustighi,A.,Zannini,A.,Tiberi,L.,Sommaggio,R.,Piazza,S.,Sorrentino,G.,
Nuzzo,S.,Tuscano,A.,Eterno,V.,Benvenuti,F.,etal.(2014).Prolyl-isom-
erasePin1controlsnormalandcancerstemcellsofthebreast.EMBOMol.
Med.6,99–119.
124CellReports11,111–124,April7,2015a2015TheAuthors
Schubbert,S.,Shannon,K.,andBollag,G.(2007).HyperactiveRasindevel-
opmentaldisordersandcancer.Nat.Rev.Cancer7,295–308.
Shin,S.,Dimitri,C.A.,Yoon,S.O.,Dowdle,W.,andBlenis,J.(2010).ERK2but
notERK1inducesepithelial-to-mesenchymaltransformationviaDEFmotif-
dependentsignalingevents.Mol.Cell38,114–127.
Stenmark,H.(2009).RabGTPasesascoordinatorsofvesicletraffic.Nat.Rev.
Mol.CellBiol.10,513–525.
Stingl,J.,Eirew,P.,Ricketson,I.,Shackleton,M.,Vaillant,F.,Choi,D.,Li,H.I.,
andEaves,C.J.(2006).Purificationanduniquepropertiesofmammaryepithe-
lialstemcells.Nature439,993–997.
Tanoue,T.,Adachi,M.,Moriguchi,T.,andNishida,E.(2000).Aconserved
dockingmotifinMAPkinasescommontosubstrates,activatorsandregula-
tors.Nat.CellBiol.2,110–116.
Tisdale,E.J.,andBalch,W.E.(1996).Rab2isessentialforthematurationof
pre-Golgiintermediates.J.Biol.Chem.271,29372–29379.
Tong,M.,Chan,K.W.,Bao,J.Y.,Wong,K.Y.,Chen,J.N.,Kwan,P.S.,Tang,
K.H.,Fu,L.,Qin,Y.R.,Lok,S.,etal.(2012).Rab25isatumorsuppressor
genewithantiangiogenicandanti-invasiveactivitiesinesophagealsquamous
cellcarcinoma.CancerRes.72,6024–6035.
Wellner,U.,Schubert,J.,Burk,U.C.,Schmalhofer,O.,Zhu,F.,Sonntag,A.,
Waldvogel,B.,Vannier,C.,Darling,D.,zurHausen,A.,etal.(2009).The
EMT-activatorZEB1promotestumorigenicitybyrepressingstemness-inhibit-
ingmicroRNAs.Nat.CellBiol.11,1487–1495.
Wulf,G.M.,Ryo,A.,Wulf,G.G.,Lee,S.W.,Niu,T.,Petkova,V.,andLu,K.P.
(2001).Pin1isoverexpressedinbreastcancerandcooperateswithRas
signalinginincreasingthetranscriptionalactivityofc-JuntowardscyclinD1.
EMBOJ.20,3459–3472.
Yaffe,M.B.,Schutkowski,M.,Shen,M.,Zhou,X.Z.,Stukenberg,P.T.,Rahfeld,
J.U.,Xu,J.,Kuang,J.,Kirschner,M.W.,Fischer,G.,etal.(1997).Sequence-
specificandphosphorylation-dependentprolineisomerization:apotential
mitoticregulatorymechanism.Science278,1957–1960.
Yoon,C.H.,Hyun,K.H.,Kim,R.K.,Lee,H.,Lim,E.J.,Chung,H.Y.,An,S.,Park,
M.J.,Suh,Y.,Kim,M.J.,andLee,S.J.(2011).ThesmallGTPaseRac1is
involvedinthemaintenanceofstemnessandmalignanciesingliomastem-
likecells.FEBSLett.585,2331–2338.
Yu,F.,Yao,H.,Zhu,P.,Zhang,X.,Pan,Q.,Gong,C.,Huang,Y.,Hu,X.,Su,F.,
Lieberman,J.,andSong,E.(2007).let-7regulatesselfrenewalandtumorige-
nicityofbreastcancercells.Cell131,1109–1123.
Zhang,M.,Behbod,F.,Atkinson,R.L.,Landis,M.D.,Kittrell,F.,Edwards,D.,
Medina,D.,Tsimelzon,A.,Hilsenbeck,S.,Green,J.E.,etal.(2008).Identifica-
tionoftumor-initiatingcellsinap53-nullmousemodelofbreastcancer.
CancerRes.68,4674–4682.
|
|
