immersion55fang.com

ORIGINALARTICLECpGislandshoremethylationregulatescaveolin-1expressioninbreastcancerXRao1,2,12,JEvans3,13,HChae3,14,JPilrose2,SKim3,14,PYan4,R-LHuang5,H-CLai5,HLin6,YLiu6,DMiller2,J-KRhee7,Y-WHuang8,FGu9,JWGray10,TH-MHuang9andKPNephew1,2,11Caveolin-1(Cav1)isanintegralmembrane,scaffoldingproteinfoundinplasmamembraneinvaginations(caveolae).
Cav1regulatesmultiplecancer-associatedprocesses.
Inbreastcancer,atumorsuppressiveroleforCav1hasbeensuggested;however,Cav1isfrequentlyoverexpressedinaggressivebreastcancersubtypes,suggestinganoncogenicfunctioninadvanced-stagedisease.
TofurtherdelineateCav1functioninbreastcancerprogression,weevaluateditsexpressionlevelsamongapanelofcelllinesrepresentingaspectrumofbreastcancerphenotypes.
Inbasal-like(themostaggressiveBCsubtype)breastcancercells,Cav1wasconsistentlyupregulated,andpositivelycorrelatedwithincreasedcellproliferation,anchorage-independentgrowth,andmigrationandinvasion.
ToidentifymechanismsofCav1generegulation,wecomparedDNAmethylationlevelswithinpromoter'CpGislands'(CGIs)with'CGIshores',recentlydescribedregionsthatankCGIswithlessCG-density.
Integrationofgenome-wideDNAmethylationproles('methylomes')withCav1expressionin30breastcancercelllinesshowedthatdifferentialmethylationofCGIshores,butnotCGIs,signicantlyregulatedCav1expression.
InbreastcancercelllineshavinglowCav1expression(despitepromoterCGIhypomethylation),wefoundthattreatmentwithaDNAmethyltransferaseinhibitorinducedCav1expressionviaCGIshoredemethylation.
Inaddition,furthermethylomeassessmentsrevealedthatbreastcanceraggressivenessassociatedwithCav1CGIshoremethylationlevels,withshorehypermethylationinminimallyaggressive,luminalbreastcancercellsandshorehypomethylationinhighlyaggressive,basal-likecells.
Cav1CGIshoremethylationwasalsoobservedinhumanbreasttumors,andoverallsurvivalratesofbreastcancerpatientslackingestrogenreceptora(ERa)negativelycorrelatedwithCav1expression.
BasedonthisrststudyofCav1(apotentialoncogene)CGIshoremethylation,wesuggestthisphenomenonmayrepresentanewprognosticmarkerforERa-negative,basal-likebreastcancer.
Oncogene(2013)32,4519–4528;doi:10.
1038/onc.
2012.
474;publishedonline5November2012Keywords:Cav1;CpGislandshore;DNAmethylation;breastcancerINTRODUCTIONCaveolin-1(Cav1)isaubiquitousscaffoldingproteinthatcoatsplasmamembraneinvaginationstermedcaveolaeinvariouscelltypes.
1AvarietyofproteinshavebeenidentiedtointeractwithCav1,2suggestingthatCav1functionsasa'molecularhub'tointegratetheactivityofmultiplesignalingmolecules,includingSrc-familytyrosinekinases,growthfactorreceptors(epidermalgrowthfactorreceptor),GproteinandG-protein-coupledreceptors,andH-Ras.
3–6InteractionswiththeCav1-scaffoldingdomainanchortheseproteinsinarestrainedconformation,negativelyregulatingtheiractivities.
7Inaddition,locusD7S522ofhumanchromosome7q31.
1,thelocationoftheCav1gene,isfrequentlydeletedinhumancancers,8furtherimplicatingCav1asatumorsuppressor.
However,Cav1upregulationhasbeenobservedinavarietyofhumancancers,9andCav1expressionisapredictivemarkerofpoorprognosisincancerpatients.
10Furthermore,Cav1upregulationhasbeencorrelatedwithmetastaticpotential11–13andmultidrugresistance.
14,15Thus,dependingonthecellularcontext,Cav1mayalsofunctionasanoncogene.
10Inbreastcancer,Cav1downregulation(comparedwithnormaltissue)wasobserved,demonstratinganinversecorrelationbetweenCav1expressionandtumorsize,16,17andlossofCav1expressionwasassociatedwithtamoxifenresistance.
18Conversely,Cav1wasoverexpressedinasubsetofaggressivebreastcarcinomas,19includingsubsetsofbasal-likeandmetaplastictumorsandinammatorybreastcancers.
20Inbreastcancercellculturemodels,Cav1downregulationwascharacteristicofluminalbreast1InterdisciplinaryBiochemistryGraduateProgram,DepartmentofMolecularandCellularBiochemistry,IndianaUniversity,Bloomington,IN,USA;2MedicalSciencesProgram,SchoolofMedicine,IndianaUniversity,Bloomington,IN,USA;3BioinformaticsProgram,SchoolofInformaticsandComputing,IndianaUniversity,Bloomington,IN,USA;4NASRIlluminaSequencingCore,ComprehensiveCancerCenter,TheOhioStateUniversity,Columbus,OH,USA;5DepartmentofObstetricsandGynecology,InstituteofBiomedicalInformatics,NationalYang-MingUniversity,TaipeiCity,Taiwan;6DepartmentofMedicalandMolecularGenetics,IndianaUniversitySchoolofMedicine,Indianapolis,IN,USA;7InterdisciplinaryPrograminBioinformatics,SeoulNationalUniversity,Seoul,Korea;8DepartmentofObstetricsandGynecology,MedicalCollegeofWisconsin,Milwaukee,WI,USA;9DepartmentofMolecularMedicine,CancerTherapyandResearchCenter,UniversityofTexasHealthScienceCenter,SanAntonio,TX,USA;10DepartmentofBiomedicalEngineering,OregonHealthandScienceUniversity,Portland,OR,USAand11IUSimonCancerCenterandDepartmentsofCellularandIntegrativePhysiology,IndianaUniversitySchoolofMedicine,Indianapolis,IN,USA.
Correspondence:ProfessorKPNephew,CellularandIntegrativePhysiology,IndianaUniversitySchoolofMedicine,302JordanHall,1001EastThirdStreet,Bloomington,IN47405-4401,USA.
E-mail:knephew@indiana.
edu12Currentaddress:DepartmentofRadiology,SchoolofMedicine,StanfordUniversity,Stanford,CA94305,USA.
13Currentaddress:MayoClinic,Rochester,MN55905,USA.
14Currentaddress:SchoolofComputerScienceandEngineering,SeoulNationalUniversity,Seoul151-742,Korea.
Received26January2012;revised10August2012;accepted29August2012;publishedonline5November2012Oncogene(2013)32,4519–4528&2013MacmillanPublishersLimitedAllrightsreserved0950-9232/13www.
nature.
com/onccancercells;incontrast,basal-likecellsdisplayedoverexpressionofCav1.
21–23Despitethesenumerousobservations,theroleofCav1inbreastcancer,andthemechanism(s)thatregulatesitsdiversepatternsofexpression,remaintobefullyestablished.
ToidentifyCav1generegulatorymechanisms,weexaminedepigeneticchangesassociatedwithCav1expressioninbreastcancer.
Epigeneticalterations,includingDNAmethylation,histonemodicationsandnucleosomeremodeling,arenowconsideredhallmarksofallstagesofcancerdevelopment.
24DNAmethylation,inthecontextofCpGdinucleotides,hasprofoundeffectsongeneexpressionbyinuencingtheaccessibilityoftranscriptionfactorstoDNA,alteringgeneticstabilityandmodifyinggenomicstructure.
25,26Specically,DNAmethylationofpromoterCpGislands(CGIs)resultsintranscriptionalsilencing,anditsdysregulationhasanimportantroleinoncogenesisandtumorprogression.
24However,asonlyabout70%ofhumangenescontainapromoterCGI27andonly6.
8%ofCpGsresidewithinCGIs,28manypotentiallyinformativeCpGsitesremaintobeexamined.
RecentworkhasshownthatDNAmethylationcandirectlysilencegeneswithnon-CGIpromotersandcontributetotheestablishmentoftissue-specicmethylationpatterns.
29Furthermore,tissue-andcancer-specicdifferentiallymethylatedregionsoccurmorefrequentlywithinCGIshores,regionsofrelativelylowCpGdensitythatanktraditionalCGIs(upto2kbdistant),thanwithinCGIsthemselves,30,31suggestingtheinvolvementofCGIshoremethylationintissuedifferentiation,epigeneticreprogrammingandcancer.
Inthisstudy,usingapanelofcelllinesrepresentingaspectrumofbreastcancerphenotypes,wedemonstrateddramaticupregu-lationandanoncogenicroleforCav1inestrogenreceptora(ERa)-negative,basal-likecells,inwhichCav1supportedcellproliferation,anchorage-independentgrowth,migrationandinvasion.
ToidentifygeneregulatorymechanismforCav1,weinvestigatedDNAmethylationlevelswithintheCav1promoterCGIandCGIshores.
DifferentialCGIshoremethylationstronglyassociatedwithCav1expression,andingfurtherconrmedinapanelof30breastcancercelllinesusingmethyl-CpGbindingdomainproteinsequencing.
32ThemethylomeanalysisfurtherindicatedanassociationbetweenmoreaggressivebreastcancersubtypesandCav1CGIshorehypermethylation.
Inaddition,variableCav1CGIshoremethylationwasalsoobservedinhumanbreasttumorsandoverallsurvivalratesofbreastcancerpatientslackingERanegativelycorrelatedwithCav1expression.
BasedonthisrstreportofCGIshoremethylationofapotentialoncogene,wesuggestthatCav1CGIshoremethylationmayrepresentanewprognosticmarkerforERa-negative,basal-likebreastcancer.
RESULTSOncogenicroleofCav1inERa-negative,basal-likebreastcancercellsCav1expressionlevelswereexaminedinapanelofbreastcancercelllinesrepresentingseveraldiseasesubtypes(Figure1a),includingluminalantiestrogen-sensitiveMCF7andBT-474,basal-likeantiestrogen-resistantMDA-MB-231andtwoantiestrogen-resistantMCF7-sublines,MCF7-FandMCF7-T.
33MCF7-F,derivedfromMCF7,isresistanttobothfulvestrant(aselectiveestrogenreceptordownregulator)andtamoxifen(aselectiveestrogenreceptormodulator)andhaveanERa-negativeandbasal-likephenotype.
MCF7-Tisresistanttotamoxifen,butnottofulvestrant,andmaintainsanERa-positiveandluminalphenotype.
Consistentwithapreviousreport,21Cav1mRNAlevelswerehigher(Po0.
01)inMDA-MB-231cellsthaninMCF7andBT-474(Figure1a).
MCF7-Tcellsdisplayeddecreased(2.
5-fold,Po0.
01)Cav1expression,relativetoitsMCF7parentalcells,agreeingwithapreviousreport.
18However,Cav1expressioninMCF7-Fcellswasincreased4-fold(Po0.
01)ascomparedwithitsparentalMCF7cells.
AnalysisofwholetranscriptomeRNA-seqdataofMCF7,MCF7-TandMCF7-FcellsconrmedthispatternofCav1expression(SupplementaryFigureS1).
Cav1proteinexpressionpatterns,assessedbyimmunouorescencestaining,weresimilartothemRNAexpressionpatternsofCav1inthesecelllines(Figure1b).
ToexaminewhetherthedistinctexpressionpatternsofCav1inMCF7-TandMCF7-Fwereduetodrugtreatment,weinvestigatedtheeffectofthesedrugsonendogenousCav1expression.
TheparentalMCF7cellsweretreatedwith4-hydroxytamoxifenorfulvestrantfor8daysandCav1mRNAlevelwasexaminedattheindicatedtimepoints.
4-hydroxytamoxifenrepressedCav1expres-sion(Po0.
01,afterday2);however,noeffectoffulvestrantonCav1levelswasobserved(Figure1c).
TheseresultssuggestthattamoxifentreatmentmayinitiallyinducelossofCav1expressionduringthedevelopmentoftamoxifenresistance,however,elevatedCav1levelsinfulvestrantresistantMCF7-Fcellsmayresultfromlong-termgeneticandepigeneticalterationsduringthedrugtreatmentandcellsubtypetransition(fromluminaltobasal).
ToexaminewhetherCav1expressiondirectlyassociateswithfulvestrantresistance,weectopicallyoverexpressedCav1infulvestrant-sensitiveMCF7andMCF7-Tcells,followedbyfulves-tranttreatmentfor7days.
Asexpected,fulvestrantinhibitedthegrowthofcontrolMCF7andMCF7-Tcells(vector,emptyvector-transfected)(Figure1dandSupplementaryFigureS2).
AlthoughthegrowthrateofCav1-overexpressingcellswasslightlyreduced,comparedwithcontrolcellswhentreatedwithDMSO,thesecellsweresimilarlygrowth-inhibitedbyfulvestrant,indicatingthatCav1alonecouldnotconferresistancetothisselectiveestrogenreceptordownregulator(Figure1dandSupplementaryFigureS2).
Furthermore,weexaminedcellresponseto4-hydroxytamoxifenandconcludedthatCav1overexpressiondidnotchangecellsensitivitytotheselectiveestrogenreceptormodulatoreither(SupplementaryFigureS3).
AsMCF7-Fcellsshowedincreasedclonogenicity,33andmigration/invasionactivity(SupplementaryFigureS4)ascom-paredwithMCF7andMCF7-T,wenextassessedwhetherCav1expressioncontributedtothemoreaggressivebasal-likepheno-type.
EctopicCav1overexpressiondidnotenhanceclonogenicactivityofMCF7cells(SupplementaryFigureS5a)andinfactdecreasedclonogenicityofMCF7-Tcells(SupplementaryFigureS5b).
ElevatedCav1expressionsignicantlyinhibitedmigrationandinvasionactivityofMCF7cells(SupplementaryFigureS6a),whileslightlyincreasedinvasivenessofMCF7-Tcells(SupplementaryFigureS6b).
Therefore,overexpressionofCav1wasnotsufcienttodrivethemoreaggressivephenotype.
Howeversmallinterfering-mediatedknockdownofCav1inMCF7-Fcellssignicantlydecreasedcellproliferation(Figure1e)andclonogenicactivity(Figure1f),aswellascellmigrationandinvasion(Figure1g).
SimilarresultswerealsoobservedinMDA-MB-231cells(SupplementaryFigureS7).
TheseresultssuggestthatCav1expressionisassociatedwithbreastcancersubtypeandthatthisscaffoldingproteinhasanoncogenicroleinERa-negative,basal-likebreastcancercells.
DirectroleforCGIshoremethylationinCav1geneexpressionAlthoughCav1hasbeenreportedtobeupregulatedbytheDNAmethyltransferaseinhibitor5-aza-CdRinprostate,lungandovariancancers,34–36hypermethylationoftheCav1promoterCGIinhumancancerappearstoberare,atonly6%ofcervical37and3.
8%ofcolorectalcancers,38andentirelyabsentinprimaryovariantumors.
39TofurtherinvestigatetheroleofDNAmethylationinCav1expression,wetreatedbreastcancercelllinesdisplayinglowCav1expression(MCF7,MCF7-TandBT-474)with5-aza-CdRfor6days.
AlthoughCav1mRNAlevelsincreased(Po0.
01)inallthreecelllines(Figure2a),thatincreasewasnotduetoalteredmethylationoftheCav1promoterCGI,whichisCav1inbreastcancerXRaoetal4520Oncogene(2013)4519–4528&2013MacmillanPublishersLimitedconstitutivelyhypomethylated,determinedbybothmethylation-specicPCRandbisultesequencing(Figures2bandc)OnlysporadicmethylationwasobservedinBT-474cells,whichhadthelowestCav1expression.
Incontrast,methylationoftheCpGsiteslocatedupstreamoftheCav1CGI(a'50-CGIshore'),30wasapparent(Figures2bandd).
Moreover,after5-aza-CdRtreatment,decreasedCav1CGIshoremethylationwasobserved(Figure2d),indicatinganassociationbetweenshoremethylationandexpression.
TofurtherinvestigatetherelationshipbetweenCav1methyla-tionandexpression,wequantiedCav1CGIandCGIshoreDNAmethylationlevelsbypyrosequencing,afullyquantitativemethylationassessment,inthepanelofantiestrogen-sensitiveand-resistantcelllinesdisplayingdifferentialCav1expressionlevels(Figure1a).
SevenpairsofprimersweredesignedtoamplifydifferentregionsofCav1,covering10and18CpGsitesintheCGIshoreandCGI,respectively(SupplementaryFigureS8andSupplementaryTableS1).
TheanalysisofindividualCpGsitesconrmedthattheCav1CGIwashypomethylatedinallcelllines(SupplementaryFigureS9;summarizedinFigure3a).
However,50-CGIshoremethylationwasdecreasedintheaggressiveMDA-MB-231andMCF7-Fcells(median2.
7%,Po0.
05andmedian9.
6%,Po0.
01,respectively),butnotinlessaggressiveMCF7,MCF7-TandBT-474(median21.
1%,38.
6%and54.
6%,respectively).
Furthermore,30-CGIshoremethylationinverselycorrelatedwithCav1expressioninallthesecelllines,withthelowestmethylationlevelobservedinMDA-MB-231cells(median2.
9%,Po0.
01comparedwithallothercelllines),decreased(Po0.
01)30-CGIshoremethylationinMCF7-Fcell(median24.
3%)andincreased(Po0.
01)30-CGIshoremethylationinMCF7-Tcells(median64.
2%)comparedwithMCF7(median42.
1%).
Thehighest(Po0.
01)levelof30-CGIshoremethylationwasobservedinBT-474cells(median69.
5%vsothercelllinesexceptMCF7-T).
TodeterminewhethershoremethylationdirectlyorindirectlyinuencedCav1geneexpression,Cav1-lowexpressingcelllines(MCF7,MCF7-TandBT-474)treatedwith5-aza-CdRwereFigure1.
Cav1isupregulatedandhasanoncogenicroleinfulvestrant-resistantbreastcancercelllines.
(a)Cav1expressioninapanelofantiestrogen-sensitiveand-resistantbreastcancercelllinesasmeasuredbyquantitativereversetranscription–PCRanalysisandrelativetoitsexpressionlevelinMCF7cells(mean±s.
e.
,n3).
**Po0.
01.
Molecularfeaturesoftheselinesaregiven:ER,estrogenreceptor;PR,progesteronereceptor;Ba,basal-like;Lu,luminal.
(b)ImmunouorescencestainingofCav1inMCF7,MCF7-TandMCF7-Fcells.
Cellswereserum/E2starvedfor3days,followingbyuorescencemicroscopy(magnication60oilimmersionobjective.
Bar,15mM).
(c)EffectoftamoxifenandfulvestrantonendogenousCav1expression.
MCF7wereserum/E2starvedfor3daysandthentreatedwith1mM4-hydroxitamoxifen(OHT)or100nMfulvestrantfor8days.
Cav1expressionwasdetectedattheindicatedtimepointbyreversetranscription–PCR,relativetoitindimethylsulfoxide(DMSO)-treatedcells(mean±s.
e.
,n3).
**Po0.
01.
(d)FulvestrantsensitivityofCav1-overexpressingMCF7cells.
CellsweretransfectedwithpCMV6-XL5(vector)orCav1overexpressionplasmid(pCAV1)andthentreatedwiththeindicateddosesoffulvestrantfor7days.
CellviabilitywasdeterminedbyMTTassay(absorbanceat600nmlinearlycorrelatedwithcellnumber;mean±s.
e.
,n6).
Thedifferencebetweenvector-transfectedcellsandpCAV1-transfectedcellswasmeasuredbyStudent'st-test.
*Po0.
05.
(e)KnockdownofCav1inhibitsgrowthofMCF7-Fcells.
Cellsweretransfectedwith50nMcontrolsmallinterferingRNA(siCTR),orsmallinterferingRNAsthattargetCav1(siCav1andsiPool)andculturedfor6days.
CellviabilitywasdeterminedbyMTTassay(mean±s.
e.
,n6).
**Po0.
01.
(f)KnockdownofCav1inhibitsclonogenicityofMCF7-Fcells.
SmallinterferingRNA-transfectedcellswereculturedingrowthmediumfor2weeksandcoloniescontaining450cellswerescored(mean±s.
e.
,n3).
**Po0.
01.
(g)Migration/invasionactivityofMCF7-Fcells(mean±s.
e.
,n3).
**Po0.
01.
Cav1inbreastcancerXRaoetal4521&2013MacmillanPublishersLimitedOncogene(2013)4519–4528subjectedtopyrosequencinganalysis.
5-aza-CdRdecreased50-and30-CGIshoremethylationlevelsinMCF7-Tcellsbyanaverageof56%(changefrom39%to17.
1%)and45%(changefrom63to35%)(Po0.
01),respectively(SupplementaryFigureS10;summarizedinFigure3b).
SimilarresultswereobservedinMCF7andBT-474cells(SupplementaryFigureS11).
However,whenMCF7-Fcells,whichhavelowCav1CGIshoremethylation,weretreatedwith5-aza-CdR,Cav1expressionwasonlyslightlyincreased(B1.
8-fold,SupplementaryFigureS12),comparedwithCav1-lowexpressingcelllines(Figure2a).
Takentogether,theseresultssupportadirectregulatoryroleofCGIshoremethylationinCav1expressioninbreastcancercells.
Inaddition,accordingtoourbisultesequencingandpyrosequencingresults,wedidnotobserveanymutationsintheCav1promoterCGIandCGIshores.
TogetherwitharecentstudyreportinglackofCav1genemutationsinhumanbreastcancer,40thedifferentialCav1expressionandCGIshoremethylationseemsunlikelytobeassociatedwithDNAmutation.
Subtype-specic,Cav1CpGislandshoremethylationpatternsinbreastcancercelllinesToinvestigatewhetherCGIshoremethylationisacommonmechanismthatregulatesCav1expression,weperformedgenome-wideprolingofDNAmethylationusingmethyl-CpGbindingdomainproteinsequencing(seemethods)on30breastcancercelllines.
Thebreastcancercelllinesweredividedintotwogroups,'Cav1-low'(12luminalsubtypeand3basalAsubtype)and'Cav1-high'(15basalsubtype),basedonourpreviousndings.
21TheDNAmethylationlandscapesofthesecelllinesindicatedthatCav1promoterCGIwashypomethylatedinalmostallcelllines,regardlessofCav1-expressionlevel.
However,hypermethylationofCav1CGIshoreswasobservedmostlyinCav1-lowlines(Figure4).
AsummaryofDNAmethylationaroundtheCav1promoterCGIforluminalandbasalsubtypecellsisshowninFigure5a.
DNAmethylationinluminal/Cav1-lowcellsdemonstratedtwomajorpeaksattheCGIshoreregionsandasub-peakattheCGI;incontrast,thetwomajorDNAmethylationpeaksattheCGIshoreweredramaticallylowerinbasal/Cav1-highlines.
ByintegratingthemethylomedatawithCav1mRNAexpressionproles,wefoundthatCGIshoremethylationwithin500bpofCGIshowedastrongnegativecorrelationwithCav1expression(ro0.
7),whileaweaknegativecorrelation(r40.
6)wasseenforCGImethylationorCGIshoremethylationwithin2kboftheCGI(Figure5b).
ThesendingsindicatethatCav1CGIshorehypermethylationisacommoneventinluminalbreastcancerandstronglyassociateswithCav1generepression.
AsCav1expressionstronglyassociateswiththebasal-likesubtype,weFigure2.
5-aza-CdRrestoresCav1expressionwithoutaffectingthehypomethylationofCav1CpGisland(CGI).
Cellsweretreatedwith5mM5-aza-CdRfor6days.
TotalRNAandgenomicDNAwereharvestedforreversetranscription–PCRandbisulteconversion.
(a)Cav1mRNAlevelmeasuredbyreversetranscription–PCR.
TheCav1levelinDMSOtreatedcellswasnormalizedto1(mean±s.
e.
,n3).
**Po0.
01.
(b)UCSCgenomebrowserviewofCav1anddistributionofCpGsites.
Thetwotwo-headedarrowsindicatedtheregionsthatareexaminedbymethylation-specicPCRin(c)and(d).
(c)CpGmethylationonCav1CGIofMCF7cell.
Left:methylation-specicPCRresult.
UD,unmethylatedDNAcontrol.
MD,methylatedDNAcontrol.
U,resultswithprimersspecicforunmethylatedsequence.
M,resultswithprimersspecicformethylatedsequence.
Right:resultsofbisultegenomicDNAsequencingcoveringtherst15CpGsitesonCav1CGI.
OpensquaresindicatethatCpGsitesarefullyunmethylated;partiallylledsquaresindicatevariousdegreesofCpGmethylation.
(d)CpGmethylationonCav150-CGIshore.
Left:methylation-specicPCRresultsofMCF7.
Right:quanticationofbanddensityonthegel(mean±s.
e.
,n3).
*Po0.
05,**Po0.
01.
Cav1inbreastcancerXRaoetal4522Oncogene(2013)4519–4528&2013MacmillanPublishersLimitedhypothesizethatCav1CGIshorehypomethylationlikelycon-tributestoenhancedaggressivenessofthesecells.
Cav1methylationandexpressioninbreasttumorsfrombreastcancerpatientsToinvestigatewhetherCav1CGIshoremethylationalsooccursinhumanbreasttumors,weperformedgenome-wideprolingofDNAmethylationon77breasttumors(50ERa-positiveand27ERa-negative,SupplementaryTableS2)and10normalbreasttissues.
Innormalbreasttissues,hypomethylationofbothCav1CGIandCGIshoreswasobserved(Figure7),consistentwithpreviouslyreportedhighCav1expressioninnormalbreasttissues.
16,17ThemajorityofERa-positivetumorsdisplayedhypermethylationoftheCav130CGIshore,variedmethylationof50-CGIshore,andsporadichypermethylationofCav1CGIFigure3.
DifferentialCav1CGIshoremethylationinantiestrogen-resistantbreastcancercelllines.
(a)Thebox-and-whiskerplotdisplayspyrosequencingresultsofvecelllines.
Themethylationlevelisindicatedaspercentage:0,nomethylation;100,100%methylationoftheCpGsites.
ThestatisticalanalysiswasperformedforMCF7vsallothercelllines.
**Po0.
01.
Formethylationof30-CGIshore,allthecelllinesweredifferentfromeachother(Po0.
01),exceptforMCF7-TvsBT474.
(b)Thebox-and-whiskerplotdisplayspyrosequencingresultsofMCF7-Tcellsaftertreatingwith5mM5-aza-CdRfor6days.
**Po0.
01.
Figure4.
IGV(IntegrativeGenomicsViewer)viewofsequencedDNAmethylationtracksof30breastcancercelllines.
Top,Cav1transcript,Cav1CGIandCGIshores;Right,Cav1expressionforeachcellline,accordingtopreviouslypublisheddata;21Left,subtypeandERaexpressionofeachcellline.
NumbersonrightareLog2expressionrationormalizedtothemeanofCav1expressioninallcelllines.
Cav1inbreastcancerXRaoetal4523&2013MacmillanPublishersLimitedOncogene(2013)4519–4528(Figure6).
IntheERa-negativetumors,Cav1promoterCGIhypomethylationwasapparent,butshoremethylationvariedamongtheindividualsamples(Figure6).
StatisticalanalysisconrmedhypomethylationofCav1CGIinbothERa-positiveand-negativebreasttumors,althoughERa-positivetumorshadsignicantlyhighermethylationlevels(median0.
1vs0,P0.
009,Figure7).
Cav150-CGIshoremethylationlevelswerevariable,withnodifferencebetweenERa-positiveand-negativetumors(P0.
64,Figure7).
However,Cav130-CGIshorewasheavilymethylatedandthemethylationlevelsweresignicantlyhigherinERa-positivetumors(P0.
045,Figure7).
Owingtolackofgeneexpressiondata,wewerenotabletoassociateCav1CGIshoremethylationwithCav1expressioninthiscohortofpatients.
However,usingtheKaplan–MeierPlottertool,41wefoundthatCav1expressionwasassociatedwithoverallsurvivalratesforbreastcancerpatientswithERa-negativetumors(totalpatientnumber63,P-value0.
028)(Figure8a).
ThisassociationwasalsoobservedinERa-negative,grade3patients(totalpatientnumber47,P-value0.
037)(Figure8b).
DISCUSSIONCav1expressionhasbeenassociatedwithbasal-likebreastcancer.
22Inthecurrentstudy,wedemonstrate,forthersttime,alterationofCav1expressionwhenabreastcancercelllinechangesfromluminalsubtype(MCF7)tobasal-like(MCF7-F),usingapreviouslyestablishedbreastcancermodel.
33WefurtherdemonstrateanoncogenicroleforCav1inbasal-likecelllines,byenhancingcellproliferation,anchorage-independentgrowth,migrationandinvasion(Figure1).
InagreementwithanewunderstandingoftheroleofCGIshoremethylationincancer,30previouslyconsideredamereextensionofCGImethylation,wereport,forthersttime,astrongassociationbetweenCGIshoremethylationandCav1expressioninbreastcancerandconrmthisuniquemethylationpatterninapanelof30breastcancercelllines(Figures2–5).
WefurthershowthatCav1CGIshoresarehypomethylatedinthebasalsubtypeofbreastcancercelllines(Figure5),supportingitsassociationwithtumorprogression.
Inhumanbreasttumors,Cav130-CGIshoreismoreoftenhyper-methylatedinERa-positivetumorsthaninERa-negativetumors(Figures6and7).
ThecorrelationofCav1expressionandclinicaloutcome(Figure8)suggeststhatCav1mayrepresentanovelprognosticfactorforERa-negative,basal-likebreastcancer.
ThefunctionofCav1incanceriscellcontextdependent.
10InERa-positivebreastcancercells,Cav1hasatumor-suppressiverole.
EstradioltreatmentreducesCav1expressiontopromotecellproliferation,18,42andMCF7cellsstablyoverexpressingCav1exhibitreducedcellgrowth,colonyformationandinvasiveness.
43However,inERa-negativebasal-likebreastcancercells,Cav1switchestoanoncogenicrole.
Itcanelevateinsulingrowthfactor-I(IGF-I)receptortranscription44topromotecellproliferationorbecomephosphorylatedattyrosine14toenhanceanchorage-independentgrowthandpromotestumorcellmigration.
45Here,weillustratedthattheswitchoftheroleofCav1isaccompaniedwithCav1upregulation,aswellasCav1CGIshoredemethylation.
Figure5.
NegativecorrelationofCav1CGIshoremethylationandCav1expression.
(a)Integratedmethylationresultsof30breastcancercelllines.
Thex-axisrepresentsnucleotidepositionaccordingtohumanreferencegenome(hg18).
They-axisrepresentsreaddensityofCpGmethylationlevel:thegreenlineshowsthemethylationresultsof18basal-likebreastcancercelllinesandtheredlineindicatesthemethylationresultsof12luminalbreastcancercelllines.
TheCav1transcriptionstartsiteisindicatedbytheorangetriangle.
ThelocationoftheCav1CGIisindicatedbytheblackline.
Thebox-and-whiskerplotdisplaysCav1expressionlevelinallbasal-likecelllines(greenbar)andluminalcelllines(redbar).
**Po0.
01.
(b)BarplotofthePearsoncorrelationcoefcient(r)showingthestrengthofthenegativecorrelationbetweenCav1expressionandmethylationonCGI,CGIshorewithin500bp,andCGIshorewithin2kb.
Allthreecorrelationsweresignicantlynegative(Po0.
05),butonlymethylationofCGIshorewithin500bpshowedastrongnegativeassociationwithCav1expression(ro0.
7).
Cav1inbreastcancerXRaoetal4524Oncogene(2013)4519–4528&2013MacmillanPublishersLimitedOurconclusionthatCav1expressionisnotstronglyregulatedbypromoterCGImethylationisfurtherconrmedbyarecentlyreportedbreastcancermethylomestudy,46whichusedtheInniumplatformcoveringCpGsitesonlyintheCav1promoterCGI.
AsshowninSupplementaryTableS3veofthesixCpGsiteswerehypomethylatedandnoassociationofmethylationwithCav1expressionorclinicaloutcomewasobserved.
Ourobserva-tionofCav1CGIshoremethylationisalsoconsistentwithseveralpreviousstudies.
Engelmanetal.
47reportedCav1CGImethylationinbreastcancercelllines.
However,theinterrogatedregioninthatstudy(359to330bp)wasactuallyupstreamofthecurrentlydenedCav1CGI.
MethylationofsevenCpGsites,locatedfrom404to149bpupstreamoftheCGI,withinthe50-CGIshoreofCav1westudied,weresufcienttoabolishCav1expression.
35,36,48Cav130-CGIshoremethylationhasnotbeenreportedpreviously,perhapsbecausethisregionsurroundsthesecondexonofCav1,whilepreviousstudiesexaminedwhetherpromoter/rstexonmethylationcorrelatedwithCav1transcriptionalsilencing.
However,becauseofalternativesplicing,Cav1hasfourtranscriptvariantsencodingaandbisoforms,transcribedfromthe50UTR(variant1,isoforma)orupstreamofthesecondexon(variant2/3/4,isoformb).
Wearecurrentlyinvestigatingwhetherthe30-CGIshoremethylationmayalsoregulateisoformexpression.
ToexaminewhethertheCav1CGIchromatinenvironment(andshoreregions)associateswithCav1expression,weperformedChIP-seqanalysisforhistonetailmodicationsinMCF7andMCF7-T(bothCav1-lowcelllines;Figure1).
InMCF7,highlevelofH3K4me2enrichmentwasobservedattheCav1CGIregion,withminimalenrichmentofthisactivemarkintheCGIshore(SupplementaryFigureS13).
EnrichmentofH3K9me2wasonlyobservedatCav130-CGIshore,whileH3K27me3occupiedbothCav150-and30-CGIshores.
SimilarenrichmentofH3K4me2attheCav1CGIregioninMCF7-Tcellswasseen;however,H3K9me2andH3K27me3repressivemarksaccruedatthecenteroftheCGIregionand30-CGIshore(SupplementaryFigureS13).
Combinedwiththeaboveresults,H3K4me2enrichmentintheCav1CGIdoesnotappearsufcienttoactivateCav1transcriptioninMCF7andMCF7-Tcells,thussuggestinganeedforotherfactors.
Ontheotherhand,H3K4me2occupancyonCav1CGImaycontributetotheCGIhypomethylation,asH3K4methylationinhibitsdenovoDNAmethylation.
49Figure6.
TheCancerMethylomeSystem(CMS)viewofsequencedDNAmethylationtracksof77breasttumorsand10normalbreasttissues.
ThegraysquaresrepresentmethylationlevelofCpGsites,asindicatedbythescale(dark,high;light,low).
Figure7.
DotplotsofCav1promotermethylationinhumanbreasttumors.
AsCav1CGIshoremethylationextendsintopartoftheCpGisland(CGI),the50-CGIshoreanalyzedherecovers500bpupstreamofCGIand250bpinthe50CGI.
The30-CGIshorecovers500bpdownstreamofCGIand250bpinthe30CGI.
TheCGIcoversthecenter1.
5kbCGIregion.
Theblacklinesindicatemedium(middle)andinterquartilerange(twoends).
Two-sidedMann–WhitneytestwasusedtoanalyzestatisticsignicanceandP-valuesareasindicated.
Cav1inbreastcancerXRaoetal4525&2013MacmillanPublishersLimitedOncogene(2013)4519–4528Inbreastcancer,Cav1expressionisstronglynegativelycorrelatedwithERaexpression,50andERahasbeenreportedtosilenceCav1.
48Ithasbeenreportedthattreatmentwithestradiol18,42decreasesCav1expression.
Wealsoobservedthattamoxifen(Figure1c),whichstabilizesERaproteinlevel,51reducedCav1expressioninMCF7cells.
ERaproteinlevelsareincreasedintamoxifen-resistantbreastcancercelllines,33,52andCav1expressionisfurtherrepressedinMCF7-TcomparedwithMCF7(Figure1a),alongwithanincreaseinCGIshoremethylation(Figure3a).
Takentogether,theseobservationsindicatethatERamaybetheprimaryfactorcontributingtoCav1CGIshoremethylationandrepressedCav1expressioninERa-positivebreastcancer.
However,CGIshoremethylationappearstopermanentlysilenceCav1,asshort-termtreatmentofMCF7cellswithfulvestrant,whichinducesrapidERadegradation,51hasnoeffectonCav1levels48(Figure1c).
DuringthedevelopmentofstablefulvestrantresistanceandtheERa-negative,basal-likephenotype,Cav1CGIshorehypomethylationappearstobeinducedbyanunknownmechanism(s)thatmaywarrantfurtherinvestigation.
NoassociationbetweenCav1expressionandoverallsurvivalwasobservedwhentheentirebreastcancercohortwasusedintheanalysis(SupplementaryFigureS14),consistentwithapreviousreport50andthefactthatover70%ofbreastcancerinthecohortareERapositiveanddisplaylowCav1expression.
However,oursubgroupanalysisrevealedthatelevatedCav1levelsinmoreaggressiveERa-negativepatientsassociatedwithshorteroverallsurvival(Figure7a).
Furthersubgroupingofthecohorttograde3breastcancerresultedinasimilarnegativeassociation(Figure7b).
TheseresultsimplythatCav1expressionmayserveasaprognosticfactorforERa-negativeandhigh-gradebreastcancerpatients.
However,itshouldbenotedthatthenumberofpatientswithERa-negativebreastcancerandCav1expressionwaslimited.
Inaddition,cohortdifferencesmayaffecttheresults,asarecentstudyidentiedCav1asanindependentprognosticfactorforinvasivebreastcarcinoma.
53InthistherstreportofanoncogenicroleforCav1inbasal-likebreastcancer,andrstcomprehensiveanalysisofCav1methyla-tion,werevealanegativerelationshipbetweenCav1CGIshoremethylationandCav1expressioninbreastcancer.
OurndingssupportapreviouslydescribedroleforCGIshoremethylationincancer,30demonstratingtheimportanceofincludingCGIshoresinfutureDNAmethylomeanalyses.
WespecicallydemonstrateanassociationbetweenCav1CGIshoremethylationandbreasttumorprogression.
ThecorrelationofCav1levelandclinicaloutcomefurthersuggeststhatCav1expressionandCav1CGIshoremethylationmayrepresentnovelprognosticfactorsforERa-negative,basal-likebreastcancer.
MATERIALSANDMETHODSCelllinesandreagentsBreastcancercelllineswereobtainedfromAmericanTypeCultureCollection(Manassas,VA,USA)andculturedaccordingtotheprotocolsprovidedbythecompany.
MCF7,MCF7-T,andMCF7-Fcellswereculturedaspreviouslydescribed.
335-Aza-20-deoxycytidine(5-aza-CdR)waspurchsedfromSigma-AldrichCo.
LLC(StLouis,MO,USA).
SmallinterferingRNAsthattargetCav1werefromSantaCruzBiotechnologyInc.
(SantaCruz,CA,USA).
pCMV6-XL5(vector)andpCav1(plasmidthatoverexpressesCav1)werepurchasedfromOriGeneTechnologiesInc.
(Rockville,MD,USA).
Anti-Cav1(clone2297)antibodywasfromBDBiosciences(SanDiego,CA,USA).
Transfection,cellviabilityassayandclonogenicityassaysPlasmidsweretransfectedwithEugeneHDTransfectionReagent(RocheAppliedScience,Indianapolis,IN,USA),andLipofectamine2000(Invitro-gen,Carlsbad,CA,USA)wasusedforsmallinterferingRNAtransfections.
Cellviabilitywasdeterminedbythe3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazoliumbromide(MTT,Sigma-Aldrich)assay,asdescribedpreviously.
33Toexamineclonogenicactivity,cellswereseeded(300erwell)insix-wellplates,culturedfor10–14days,andthenstainedwith0.
5%methylenebluein50%methanol.
ColoniescontainingX50cellswerescored.
Migration/invasionassaysMigration/invasionassayswereperformedaccordingtopreviousdescrip-tion.
54Transwellchambers(24-well;BDBioCoatControlInsertsfromBDBiosciences)with8.
0-mmporesizepolycarbonatemembranewereusedformigrationassay,andBDBioCoatMatrigelInvasionChamber(BDBiosciences)wasusedforinvasionassays.
Migratedandinvadedcellswerestainedandobservedundertheopticalmicroscopeatamagnicationof100.
Cellswerecountedin5eldsoftriplicatemembranes.
SeeSupplementaryMaterialsfordetails.
DNAextractionandbisulphiteconversionGenomicDNAwasisolatedusingQIAmpDNABloodMiniKit(Qiagen,Valencia,CA,USA),asdescribedpreviously.
55SodiumbisulteconversionandcleanupwereperformedusingEZDNAMethylationkit(ZymoResearch,Orange,CA,USA),accordingtothemanufacturer'sinstruction.
MethylationspecicPCRandbisultegenomicDNAsequencingDNAmethylationontheCav1CGIand50-CGIshorewasdeterminedusingmethylation-specicPCRaccordingtopreviousdescription.
35,38MethylatedandunmethylatedcontrolDNAswerepurchasedfromQiagen.
BisultegenomicDNAsequencingwasusedtodeterminemethylationofCav1CGI,asdescribedpreviously.
35SeeSupplementaryMaterialsfordetails.
Figure8.
RelationshipofCav1expressionandoverallsurvivalofbreastcancerpatients.
(a)SurvivalanalysisindicatingsignicantlyworseoverallsurvivalofpatientswithERa-negativetumorsexpressinghigherlevelofCav1(totalpatientnumber63,P-value0.
028).
(b)SurvivalanalysisindicatingsignicantlyworseoverallsurvivalofERa-negative,grade3patients(totalpatientnumber47,P-value0.
037).
Cav1inbreastcancerXRaoetal4526Oncogene(2013)4519–4528&2013MacmillanPublishersLimitedDNAmethylationanalysisbypyrosequencingThePyroMarkAssayDesignprogram(Qiagen)wasusedtodesignprimersforamplicationofspecicregionsofCav1CGIandCGIshoresfrombisulte-convertedgenomicDNA(SupplementaryTableS1).
Themethyla-tionlevelofindividualCpGsitesineachampliconwasdetectedusingPyrosequencingsystem56andanalyzedbythePyroQ-CpGsoftware.
Box-and-whiskerplots,generatedinGraphPadPrismversion4.
0(GraphPadSoftware,SanDiego,CA,USA)usingdefaultsettings,wereusedtodisplaythemethylationresults.
Theboxesshowtheinterquartilerange(IQR)aroundthemedian;thewhiskersextendfromtheminimumvaluetothemaximumvalue.
Methyl-CpGbindingdomain-basedcapturecoupledwithmassivelyparallelsequencingandidenticationofdifferentiallymethylatedregionsMethyl-CpGbindingdomainproteinsequencingwasperformedfor30breastcancercelllines,77breasttumors(SupplementaryTableS2)and10normalbreasttissuesaspreviouslydescribed.
32,57,58Thegenome-widemethylationdatawereprocessedusingmCpG-SNP-EXPRESS(Chaeetal.
unpublished).
MACS59wasusedtoidentifydifferentiallymethylatedregionsinbreastcancercelllines.
Thebi-asymmetric-Lapilacemodel58wasusedtoidentifydifferentiallymethylatedregionsinbreasttumors.
SeeSupplemen-taryMaterialsfordetails.
Themethylomedatasetisavailableat''TheCancerMethylomeSystem''website:http://cbbiweb.
uthscsa.
edu/KMethylomes/.
CorrelationanalysisAPearsoncorrelationone-tailedt-testwasperformedtomeasuretheassociationbetweenthegeneexpressionandmethylation.
AP-valueo0.
05wasconsideredsignicant.
Correlationcoefcients(rvalues)from1.
0to0.
7representedastrongnegativeassociation,withrvaluefrom0.
7to0.
3consideredweaknegativeassociationsandrvaluesfrom0.
3to0.
3indicatinglittleornoassociation.
RNA-seqForwholetranscriptomeanalysis,RNA-seqlibrariesweregeneratedutilizingamodiedversionoftheIlluminadirectionalmRNA-seqlibraryprotocolwithduplexspecicnuclease(DSN;Evrogen,Moscow,Russia)normalization(Nephewandco-workers,manuscriptinpreparation).
NextgenerationsequencingwasperformedwiththeIlluminaGAIIanalyzer.
Sequencereads(51bp)weremappedtothehumangenome(NCBI36/hg18)usingtheSolexaAnalysisPipelinewithBFASTalignmentsandaTopHat-likestrategytodeterminesplicingjunctionsfollowedbyexpres-sionlevel(RPKM)analyseswithCufinks.
60SeeSupplementaryMaterialsfordetails.
Kaplan–MeiersurvivalanalysisTheassociationofCav1expressionandoverallsurvivalrateinbreastcancerpatientswasanalyzedusinganonlinesurvivalanalysistool,Kaplan–MeierPlotter(http://kmplot.
com/backup/breast).
Itassessestheeffectofgeneexpressiononbreastcancerprognosisusingmicroarraydatafrom1809patients.
41ThepatientdataarefromGEO,withavailablerawdataandclinicalsurvivalinformation),StatisticalanalysisForMTTcellviability,clonogenicity,quantitativereversetranscription–PCR,andmigration/invasionassays,statisticalsignicancewasanalyzedbyunpairedStudent'st-test.
Forthemethylationresults(box-and-whiskerplots),pairedStudent'st-testwasperformed.
P-valueso0.
05wereconsideredstatisticallysignicant.
ForCav1promotermethylationinbreasttumors,statisticalsignicancewasanalyzedbytwo-sidedMann–Whitneytest.
CONFLICTOFINTERESTTheauthorsdeclarenoconictofinterest.
ACKNOWLEDGEMENTSTheauthorswishtothankDrFFangforpyrosequencinganalysis,DrCBalchformanuscriptpreparation,andtheIUBlightMicroscopyImagingCenterformicroscopyresources.
ThisworkwassupportedbyNIHgrantsCA085289andCA113001,pilotprojectfundingfromtheIntegratedCancerBiologyProgramandtheWaltherCancerFoundation(Indianapolis,IN,USA).
WeacknowledgetheuseoftheICBP45Kitinthisstudy.
REFERENCES1RazaniB,WoodmanSE,LisantiMP.
Caveolae:fromcellbiologytoanimalphy-siology.
PharmacolRev2002;54:431–467.
2LiuP,RudickM,AndersonRG.
Multiplefunctionsofcaveolin-1.
JBiolChem2002;277:41295–41298.
3LiS,CouetJ,LisantiMP.
Srctyrosinekinases,Galphasubunits,andH-Rasshareacommonmembrane-anchoredscaffoldingprotein,caveolin.
Caveolinbindingnegativelyregulatestheauto-activationofSrctyrosinekinases.
JBiolChem1996;271:29182–29190.
4CouetJ,SargiacomoM,LisantiMP.
Interactionofareceptortyrosinekinase,EGF-R,withcaveolins.
Caveolinbindingnegativelyregulatestyrosineandserine/threoninekinaseactivities.
JBiolChem1997;272:30429–30438.
5OstromRS,InselPA.
TheevolvingroleoflipidraftsandcaveolaeinGprotein-coupledreceptorsignaling:implicationsformolecularpharmacology.
BrJPhar-macol2004;143:235–245.
6EngelmanJA,WykoffCC,YasuharaS,SongKS,OkamotoT,LisantiMP.
Recom-binantexpressionofcaveolin-1inoncogenicallytransformedcellsabrogatesanchorage-independentgrowth.
JBiolChem1997;272:16374–16381.
7PatelHH,MurrayF,InselPA.
Caveolaeasorganizersofpharmacologicallyrelevantsignaltransductionmolecules.
AnnuRevPharmacolToxicol2008;48:359–391.
8EngelmanJA,ZhangXL,GalbiatiF,LisantiMP.
Chromosomallocalization,geno-micorganization,anddevelopmentalexpressionofthemurinecaveolingenefamily(Cav-1,-2,and-3).
Cav-1andCav-2genesmaptoaknowntumorsup-pressorlocus(6-A2/7q31).
FEBSLett1998;429:330–336.
9ShatzM,LiscovitchM.
Caveolin-1:atumor-promotingroleinhumancancer.
IntJRadiatBiol2008;84:177–189.
10GoetzJG,LajoieP,WisemanSM,NabiIR.
Caveolin-1intumorprogression:thegood,thebadandtheugly.
CancerMetastasisRev2008;27:715–735.
11NestlA,VonSteinOD,ZatloukalK,ThiesWG,HerrlichP,HofmannMetal.
Geneexpressionpatternsassociatedwiththemetastaticphenotypeinrodentandhumantumors.
CancerRes2001;61:1569–1577.
12TseEY,KoFC,TungEK,ChanLK,LeeTK,NganESetal.
Caveolin-1overexpressionisassociatedwithhepatocellularcarcinomatumorigenesisandmetastasis.
JPathol2011;226:645–653.
13ArpaiaE,BlaserH,Quintela-FandinoM,DuncanG,LeongHS,AblackAetal.
Theinteractionbetweencaveolin-1andRho-GTPasespromotesmetastasisbycon-trollingtheexpressionofalpha5-integrinandtheactivationofSrc,RasandErk.
Oncogene2011;31:884–896.
14RavidD,MaorS,WernerH,LiscovitchM.
Caveolin-1inhibitscelldetachment-inducedp53activationandanoikisbyupregulationofinsulin-likegrowthfactor-Ireceptorsandsignaling.
Oncogene2005;24:1338–1347.
15BelangerMM,RousselE,CouetJ.
Up-regulationofcaveolinexpressionbycyto-toxicagentsindrug-sensitivecancercells.
AnticancerDrugs2003;14:281–287.
16ParkSS,KimJE,KimYA,KimYC,KimSW.
Caveolin-1isdown-regulatedandinverselycorrelatedwithHER2andEGFRexpressionstatusininvasiveductalcarcinomaofthebreast.
Histopathology2005;47:625–630.
17SagaraY,MimoriK,YoshinagaK,TanakaF,NishidaK,OhnoSetal.
ClinicalsignicanceofCaveolin-1,Caveolin-2andHER2/neumRNAexpressioninhumanbreastcancer.
BrJCancer2004;91:959–965.
18ThomasNB,HutchesonIR,CampbellL,GeeJ,TaylorKM,NicholsonRIetal.
Growthofhormone-dependentMCF-7breastcancercellsispromotedbycon-stitutivecaveolin-1whoseexpressionislostinanEGF-R-mediatedmannerduringdevelopmentoftamoxifenresistance.
BreastCancerResTreat2010;119:575–591.
19SavageK,LambrosMB,RobertsonD,JonesRL,JonesC,MackayAetal.
Caveolin1isoverexpressedandampliedinasubsetofbasal-likeandmetaplasticbreastcarcinomas:amorphologic,ultrastructural,immunohistochemical,andinsituhybridizationanalysis.
ClinCancerRes2007;13:90–101.
20VandenEyndenGG,VanLaereSJ,VanderAuweraI,MerajverSD,VanMarckEA,vanDamPetal.
Overexpressionofcaveolin-1and-2incelllinesandinhumansamplesofinammatorybreastcancer.
BreastCancerResTreat2006;95:219–228.
21NeveRM,ChinK,FridlyandJ,YehJ,BaehnerFL,FevrTetal.
Acollectionofbreastcancercelllinesforthestudyoffunctionallydistinctcancersubtypes.
CancerCell2006;10:515–527.
22PinillaSM,HonradoE,HardissonD,BenitezJ,PalaciosJ.
Caveolin-1expressionisassociatedwithabasal-likephenotypeinsporadicandhereditarybreastcancer.
BreastCancerResTreat2006;99:85–90.
23KaoJ,SalariK,BocanegraM,ChoiYL,GirardL,GandhiJetal.
Molecularprolingofbreastcancercelllinesdenesrelevanttumormodelsandprovidesaresourceforcancergenediscovery.
PLoSOne2009;4:e6146.
24JonesPA,BaylinSB.
Theepigenomicsofcancer.
Cell2007;128:683–692.
Cav1inbreastcancerXRaoetal4527&2013MacmillanPublishersLimitedOncogene(2013)4519–452825JonesPA,LiangG.
RethinkinghowDNAmethylationpatternsaremaintained.
NatRevGenet2009;10:805–811.
26HermanJG,BaylinSB.
Genesilencingincancerinassociationwithpromoterhypermethylation.
NEnglJMed2003;349:2042–2054.
27DeatonAM,BirdA.
CpGislandsandtheregulationoftranscription.
GenesDev2011;25:1010–1022.
28RollinsRA,HaghighiF,EdwardsJR,DasR,ZhangMQ,JuJetal.
Large-scalestructureofgenomicmethylationpatterns.
GenomeRes2006;16:157–163.
29HanH,CortezCC,YangX,NicholsPW,JonesPA,LiangG.
DNAmethylationdirectlysilencesgeneswithnon-CpGislandpromotersandestablishesanucleosomeoccupiedpromoter.
HumMolGenet2011;20:4299–4310.
30IrizarryRA,Ladd-AcostaC,WenB,WuZ,MontanoC,OnyangoPetal.
Thehumancoloncancermethylomeshowssimilarhypo-andhypermethylationatconservedtissue-specicCpGislandshores.
NatGenet2009;41:178–186.
31DoiA,ParkIH,WenB,MurakamiP,AryeeMJ,IrizarryRetal.
Differentialmethylationoftissue-andcancer-specicCpGislandshoresdistinguisheshumaninducedpluripotentstemcells,embryonicstemcellsandbroblasts.
NatGenet2009;41:1350–1353.
32SerreD,LeeBH,TingAH.
MBD-isolatedGenomeSequencingprovidesahigh-throughputandcomprehensivesurveyofDNAmethylationinthehumangenome.
NucleicAcidsRes2010;38:391–399.
33FanM,YanPS,Hartman-FreyC,ChenL,PaikH,OyerSLetal.
DiversegeneexpressionandDNAmethylationprolescorrelatewithdifferentialadaptationofbreastcancercellstotheantiestrogenstamoxifenandfulvestrant.
CancerRes2006;66:11954–11966.
34WiechenK,DiatchenkoL,AgoulnikA,ScharffKM,SchoberH,ArltKetal.
Caveolin-1isdown-regulatedinhumanovariancarcinomaandactsasacandidatetumorsuppressorgene.
AmJPathol2001;159:1635–1643.
35SunagaN,MiyajimaK,SuzukiM,SatoM,WhiteMA,RamirezRDetal.
Differentrolesforcaveolin-1inthedevelopmentofnon-smallcelllungcancerversussmallcelllungcancer.
CancerRes2004;64:4277–4285.
36BachmannN,HaeuslerJ,LuedekeM,KueferR,PernerS,AssumGetal.
ExpressionchangesofCAV1andEZH2,locatedon7q31approximatelyq36,arerarelyrelatedtogenomicalterationsinprimaryprostatecarcinoma.
CancerGenetCytogenet2008;182:103–110.
37ChanTF,SuTH,YehKT,ChangJY,LinTH,ChenJCetal.
Mutational,epigeneticandexpressionalanalysesofCaveolin-1geneincervicalcancers.
IntJOncol2003;23:599–604.
38LinSY,YehKT,ChenWT,ChenHC,ChenST,ChangJG.
PromoterCpGmethylationofcaveolin-1insporadiccolorectalcancer.
AnticancerRes2004;24:1645–1650.
39HurlstoneAF,ReidG,ReevesJR,FraserJ,StrathdeeG,RahillyMetal.
AnalysisoftheCAVEOLIN-1geneathumanchromosome7q31.
1inprimarytumoursandtumour-derivedcelllines.
Oncogene1999;18:1881–1890.
40PataniN,LambrosMB,NatrajanR,DedesKJ,GeyerFC,WardEetal.
Non-existenceofcaveolin-1genemutationsinhumanbreastcancer.
BreastCancerResTreat2012;131:307–310.
41GyorffyB,LanczkyA,EklundAC,DenkertC,BudcziesJ,LiQetal.
Anonlinesurvivalanalysistooltorapidlyassesstheeffectof22,277genesonbreastcancerprognosisusingmicroarraydataof1809patients.
BreastCancerResTreat2010;123:725–731.
42RazandiM,OhP,PedramA,SchnitzerJ,LevinER.
ERsassociatewithandregulatetheproductionofcaveolin:implicationsforsignalingandcellularactions.
MolEndocrinol2002;16:100–115.
43FiucciG,RavidD,ReichR,LiscovitchM.
Caveolin-1inhibitsanchorage-indepen-dentgrowth,anoikisandinvasivenessinMCF-7humanbreastcancercells.
Oncogene2002;21:2365–2375.
44GlaitC,TencerL,RavidD,SarfsteinR,LiscovitchM,WernerH.
Caveolin-1up-regulatesIGF-IreceptorgenetranscriptioninbreastcancercellsviaSp1-andp53-dependentpathways.
ExpCellRes2006;312:3899–3908.
45WilliamsTM,LisantiMP.
Caveolin-1inoncogenictransformation,cancer,andmetastasis.
AmJPhysiolCellPhysiol2005;288:C494–C506.
46FacklerMJ,UmbrichtCB,WilliamsD,ArganiP,CruzLA,MerinoVFetal.
Genome-widemethylationanalysisidentiesgenesspecictobreastcancerhormonereceptorstatusandriskofrecurrence.
CancerRes2011;71:6195–6207.
47EngelmanJA,ZhangXL,LisantiMP.
Sequenceanddetailedorganizationofthehumancaveolin-1and-2geneslocatedneartheD7S522locus(7q31.
1).
Methy-lationofaCpGislandinthe5'promoterregionofthecaveolin-1geneinhumanbreastcancercelllines.
FEBSLett1999;448:221–230.
48ZschockeJ,MantheyD,BayattiN,vanderBurgB,GoodenoughS,BehlC.
Estrogenreceptoralpha-mediatedsilencingofcaveolingeneexpressioninneuronalcells.
JBiolChem2002;277:38772–38780.
49OoiSK,QiuC,BernsteinE,LiK,JiaD,YangZetal.
DNMT3Lconnectsunme-thylatedlysine4ofhistoneH3todenovomethylationofDNA.
Nature2007;448:714–717.
50ElsheikhSE,GreenAR,RakhaEA,SamakaRM,AmmarAA,PoweDetal.
Caveolin1andCaveolin2areassociatedwithbreastcancerbasal-likeandtriple-negativeimmunophenotype.
BrJCancer2008;99:327–334.
51WijayaratneAL,McDonnellDP.
Thehumanestrogenreceptor-alphaisaubiquitinatedproteinwhosestabilityisaffecteddifferentiallybyagonists,antagonists,andselectiveestrogenreceptormodulators.
JBiolChem2001;276:35684–35692.
52PinkJJ,JiangSY,FritschM,JordanVC.
Anestrogen-independentMCF-7breastcancercelllinewhichcontainsanovel80-kilodaltonestrogenreceptor-relatedprotein.
CancerRes1995;55:2583–2590.
53JoshiB,StrugnellSS,GoetzJG,KojicLD,CoxME,GrifthOLetal.
Phosphorylatedcaveolin-1regulatesRho/ROCK-dependentfocaladhesiondynamicsandtumorcellmigrationandinvasion.
CancerRes2008;68:8210–8220.
54RosmanDS,PhukanS,HuangCC,PascheB.
TGFBR1*6AenhancesthemigrationandinvasionofMCF-7breastcancercellsthroughRhoAactivation.
CancerRes2008;68:1319–1328.
55FangF,BalchC,SchilderJ,BreenT,ZhangS,ShenCetal.
Aphase1andpharmacodynamicstudyofdecitabineincombinationwithcarboplatininpatientswithrecurrent,platinum-resistant,epithelialovariancancer.
Cancer2010;116:4043–4053.
56TostJ,GutIG.
DNAmethylationanalysisbypyrosequencing.
NatProtoc2007;2:2265–2275.
57ZuoT,LiuTM,LanX,WengYI,ShenR,GuFetal.
EpigeneticsilencingmediatedthroughactivatedPI3K/AKTsignalinginbreastcancer.
CancerRes2011;71:1752–1762.
58LanX,AdamsC,LandersM,DudasM,KrissingerD,MarnellosGetal.
HighresolutiondetectionandanalysisofCpGdinucleotidesmethylationusingMBD-Seqtechnology.
PLoSOne2011;6:e22226.
59ZhangY,LiuT,MeyerCA,EeckhouteJ,JohnsonDS,BernsteinBEetal.
Model-basedanalysisofChIP-Seq(MACS).
GenomeBiol2008;9:R137.
60PataniN,MartinLA,Reis-FilhoJS,DowsettM.
Theroleofcaveolin-1inhumanbreastcancer.
BreastCancerResTreat2012;131:1–15.
ThisworkislicensedundertheCreativeCommonsAttribution-NonCommercial-NoDerivativeWorks3.
0UnportedLicense.
Toviewacopyofthislicense,visithttp://creativecommons.
org/licenses/by-nc-nd/3.
0/SupplementaryInformationaccompaniesthepaperontheOncogenewebsite(http://www.
nature.
com/onc)Cav1inbreastcancerXRaoetal4528Oncogene(2013)4519–4528&2013MacmillanPublishersLimited