HALLMARK_PANCREAS_BETA_CELLSios11

ARTICLEDiscoveryofnaturallyoccurringESR1mutationsinbreastcancercelllinesmodellingendocrineresistanceLesley-AnnMartin1,RicardoRibas1,NikianaSimigdala1,EugeneSchuster1,SunilPancholi1,TenchoTenev1,PascalGellert1,LakiBuluwela2,AlisonHarrod2,AllanThornhill3,JoannaNikitorowicz-Buniak1,AmandeepBhamra4,Marc-OlivierTurgeon5,GeorgePoulogiannis5,6,QiongGao1,VeraMartins7,MargaretHills7,IsaacGarcia-Murillas1,CharlotteFribbens1,NeillPatani1,ZheqiLi8,MatthewJ.
Sikora8,NicholasTurner1,WilbertZwart9,StefOesterreich8,JasonCarroll10,SimakAli2&MitchDowsett1,7Resistancetoendocrinetherapyremainsamajorclinicalprobleminbreastcancer.
Geneticstudieshighlightthepotentialroleofestrogenreceptor-α(ESR1)mutations,whichshowincreasedprevalenceinthemetastatic,endocrine-resistantsetting.
NonaturallyoccurringESR1mutationshavebeenreportedininvitromodelsofBCeitherbeforeoraftertheacquisitionofendocrineresistancemakingfunctionalconsequencesdifculttostudy.
WereporttherstdiscoveryofnaturallyoccurringESR1Y537CandESR1Y537SmutationsinMCF7andSUM44ESR1-positivecelllinesafteracquisitionofresistancetolong-term-estrogen-deprivation(LTED)andsubsequentresistancetofulvestrant(ICIR).
Mutationswereenrichedwithtime,impactedonESR1bindingtothegenomeandalteredtheESR1interactome.
Theresultshighlighttheimportanceandfunctionalconsequenceofthesemutationsandprovideanimportantresourceforstudyingendocrineresistance.
DOI:10.
1038/s41467-017-01864-yOPEN1BreastCancerNowTobyRobinsResearchCentre,InstituteofCancerResearch,LondonSW73RP,UK.
2DivisionofCancer,CRUKLabs,UniversityofLondonImperialCollege,LondonW120NN,UK.
3CentreforCancerImaging,InstituteofCancerResearch,SuttonSM25NG,UK.
4ProteomicUnit,InstituteofCancerResearch,LondonSW73RP,UK.
5DivisionofCancerBiology,TheInstituteofCancerResearch,LondonSW36JB,UK.
6DivisionofComputationalandSystemsMedicine,DepartmentofSurgeryandCancer,ImperialCollegeLondon,LondonSW72AZ,UK.
7RalphLaurenCentreforBreastCancerResearch,RoyalMarsdenHospital,LondonSW36JB,UK.
8DepartmentofPharmacologyandChemicalbiology,UniversityofPittsburgh,Pittsburgh,PA15213,USA.
9DepartmentofMolecularPathology,NetherlandsCancerInstitute,1066CXAmsterdam,Netherlands.
10CancerResearchUKCambridgeInstitute,UniversityofCambridge,CambridgeCB20RE,UK.
Lesley-AnnMartin,RicardoRibas,NikianaSimigdalaandEugeneSchustercontributedequallytothiswork.
CorrespondenceandrequestsformaterialsshouldbeaddressedtoL.
-A.
M.
(email:Lesley-ann.
martin@icr.
ac.
uk)NATURECOMMUNICATIONS|8:1865|DOI:10.
1038/s41467-017-01864-y|www.
nature.
com/naturecommunications11234567890Over70%ofbreastcancers(BC)areestrogenreceptor-α(ESR1)positiveatdiagnosis.
EstrogenmediatesitseffectsbybindingtoESR1leadingtoexpressionofgenescon-trollingproliferationandcellsurvival.
ESR1hastwodistinctactivationdomains,AF-1andAF-2.
AF-1isregulatedbyphos-phorylationwhileAF-2isintegraltotheligand-bindingdomain(LBD)andassociateswithcoactivators,controllingtheESR1transcriptionalcomplex(reviewedbyGreenandCarroll1).
Clas-sically,patientswithESR1-positiveBCaretreatedwithendocrineagentssuchastamoxifen,aromataseinhibitors(AIs),orfulves-trant,whichimpedeESR1-signaling(reviewedbyMaetal.
2).
Althoughover50%ofESR1-positivepatientsshowresponsetoendocrinetherapyandestrogendeprivationtherapyreducesBCmortalityby40%3,alargeproportionrelapsewithdenovooracquiredresistantdisease,makingitoneofthegreatestchallengesforBCresearchandtreatment.
Multiplemechanismsofresistancehavebeenproposed,mostofwhichhavebeenidentiedusingalimitednumberofESR1-positiveBCcelllines.
Theseincludeaberrantcross-talkbetweenESR1andgrowthfactorsignalingpathwaysoralterationsinthebalanceofcoactivatorsandcorepressors(reviewedbyMaetal.
2,Osborneetal.
4,andMilleretal.
5).
IthasbeenknownformanyyearsthatsomemutationsinESR1canleadtoligand-independentactivation,butuntilrecently,suchmutationsappearedtohavelittleclinicalsignicance6,astheirpresenceinprimarydiseaseisrare.
However,theprevalenceofESR1mutationsinmetastatictumorsthathaverecurredorprogressedafterendocrinetherapyisfarhigher7–9.
Wehaverecentlyreportedthatthedetectionofthesemutationsincircu-latingtumorDNA(ctDNA)of39.
1%ofmetastaticpatientsappearstocorrelatewithclinicalresistancetoAIs10.
ThemajorityofESR1mutationsarelocatedattwoaminoacidsintheLBDY537N/C/SandD538G.
FunctionalstudiesusingectopicexpressionofthesemutationsledtoconstitutiveactivityofESR1andconferredpartialresistancetoestablishedclinicaldosesoftamoxifenandfulvestrant11,12.
However,asthesemutationswereengineered,theroleofcellularcontextduringacquisitionofresistancewithtimewasnotexplored.
Inthismanuscript,wereportforthersttime,theidenti-cationofnaturallyoccurringESR1mutationsinBCcellmodelsandtheirenrichmentduringacquisitionofresistancetoendo-crinetherapy.
WeshowthatthemutatedESR1controlsacis-tromesimilartotheligand-dependentwtESR1andassociateswithanalteredproteininteractomeenablingligand-independentproliferation.
Furthermore,thesenaturallyoccurringESR1mutantsaresensitivetofulvestrant,suggestingthatthisandsimilaragentsmayhaveapplicabilityinpatientswithtumorsharboringthesemutationssupportingourrecentclinicaldata13.
ResultsDiscoveryofESR1mutationsinmodelsofendocrineresis-tance.
Previously,wereportedthedevelopmentoflong-term-estrogen-deprived(LTED)derivativesfromanumberofESR1-positiveBCcelllines(includingMCF7,HCC1428,T47D,ZR75.
1,andSUM44)14,15.
Ingeneral,estrogendeprivationleadstoaninitialquiescentpopulationaccompaniedbycelldeathandaftermanyweekstooutgrowthofacellpopulationthatthenpro-liferatesindependentlyofexogenousestrogen(SupplementaryFig.
1a–d).
ThephenotypeoftheLTEDcelllinesvariesleadingtoacontext-specicsensitivityorresistancetoadditionalagents14.
AsESR1mutationshavebeenassociatedwithresistancetoendocrinetherapy,weexploredwhetherthesemutationsorthoseofothergeneswereeitherenrichedoracquiredintheinvitromodelsdescribed.
Whole-exomesequencingfromwt-MCF7andMCF7-LTEDshowedanESR1Y537CmutationintheMCF7-LTEDatanestimatedvariantallelefrequency(VAF)of30%,whileitwasundetectableinthewt-MCF7.
Themutationwasvalidatedusingdigitaldroplet(dd)PCR(Fig.
1a,b).
ESR1mutationsoccurinLTEDbutnottamoxifen-resistantcells.
Asaresultofthisunexpectednding,wesequencedknownhotspotregionsforESR116byIonTorrentinwtandLTEDderivativesofMCF7,SUM44,HCC1428,andZR75.
1,togetherwithtamoxifen-resistant(TAMR)derivativesofMCF7andHCC1428andfulvestrant-resistant(ICIR)derivativesofwt-MCF7,MCF7-LTED,andZR75.
1-LTED(Table1;SupplementaryFig.
2).
TheESR1Y537CmutantwasdetectedintheMCF7-LTED-ICIRcellsataVAFof48%thatwasconrmedbyddPCR(49.
8%)(SupplementaryFig.
3a)butwasnotdetectedinthewt-MCF7-ICIRcells.
Comparisonofthetwoisogenicmodelsshowedthatfulvestrantresistance(SupplementaryFig.
3b)occurredirre-spectiveofthemutation.
Furthermore,bothICIRderivativesshowedamarkedreductioninESR1(SupplementaryFig.
3c)andaconcomitantdropinexpressionofestrogen-regulatedgenes(GREB1,PDZK1,PGR,andTFF1)butequivalentexpressionofgenesassociatedwithproliferationwhencomparedtotheirrespectivewt(SupplementaryFig.
3d).
Strikingly,analysisbyIontorrentalsorevealedanESR1Y537SheterozygousmutationinSUM44-LTED(VAF47%).
ESR1mutationswereconrmedbySangersequencing,RNAsequen-cing,massspectrometry,andwhole-exomesequencing(Supple-mentaryFig.
4a–g).
ExomesequencingdidnotrevealanyadditionalmutatedgenesinvolvedinAIresistancebeyondthemutationinESR1nordiditrevealmutationsingenesknowntobedriversofBC17thatmightpromotegrowthbyothermechanisms(SupplementaryData1).
InordertodetermineiftheESR1Y537CVAFof30%intheMCF7-LTEDcellswasindicativeofamixedpopulationofcellsharboringeitherESR1wtorESR1Y537C,weassessedESR1copynumberbyuorescentinsituhybridization(FISH)andexomesequencing.
Thisrevealedanallelicimbalance,whichonaverageidentiedtwoormorewtcopiesofESR1andonemutantcopypercellintheMCF7-LTED,indicating100%ofthecellpopulationharboredthemutation.
Incontrast,theMCF7-LTED-ICIRcellswereenrichedfortwocopiesofESR1percellsimilartotheSUM44-LTED,accountingfortheVAFof50%againindicatingeverycellinthegivenpopulationcontainedamutation(SupplementaryFig.
5).
TemporalenrichmentofESR1mutationsduringestrogendeprivation.
AnalysisbyddPCRoveratimecourseshowedthattheESR1Y537Smutationwasdetectablewithin12weeksfollowingtransferofSUM44cellstoestrogen-freemedium(Fig.
1c).
Thereafter,theVAFincreasedprogressivelyupto50%.
Inordertodetermineifthemutationwaspresentintheparentalpopu-lationorwasacquiredasaresultoftheselectivepressureofestrogenwithdrawal,wescreenedover6*106matchedparentalSUM44copies.
Interestingly,theESR1Y537Smutationwaspresentinwt-SUM44atanapparentfrequencyof~1:1.
000.
000(Fig.
1d),implyingthattheESR1Y537Smutationpre-existsinaverysmallproportionofSUM44cells.
WefurtherscreenedasecondbatchofSUM44-LTEDandtheircorrespondingparentcellline18butnomutationwasidentied,suggestingthisisnottheonlyadaptivemechanism.
Inordertocontrolfurtherthepotentialofcontamination,wescreenedanequivalentnumberofESR1-negativeSKBR3cellsandnomutationwasevident(Fig.
1d).
Finally,toaddressthepossibilitythattheY537CmutationwasalsoresidentatlowfrequencyinMCF7cells,wescreenedthreeindependentbatches,coveringover6*106copies,howeverwewereunabletoidentifytheY537Cmutation.
ARTICLENATURECOMMUNICATIONS|DOI:10.
1038/s41467-017-01864-y2NATURECOMMUNICATIONS|8:1865|DOI:10.
1038/s41467-017-01864-y|www.
nature.
com/naturecommunicationsChannel1amplitude6000500040003000200010000MCF7-LTEDY537Cwt-SUM44Y537SSKBR3SignificantpeaksinSUM44-LTEDwt-SUM44wt-SUM44SUM44-LTEDSUM44-LTEDInSUM44-LTEDInbothInwt-SUM4418,592470210,055Significantpeaksinwt-SUM44654321065432106543210CommonpeaksAveragereadcount(permillionmappedreads)Averagereadcount(permillionmappedreads)Averagereadcount(permillionmappedreads)250020001500Channel1amplitude1000500029.
8%WT300020001000VAF(%)6050403020100Channel2amplitude010002000300040005000010002000300040005000Channel2amplitudeChannel2amplitudeChannel1amplitude02500200015001000500DigitaldropletPCRWT-SUM44SUM44wk1SUM44wk2SUM44wk10SUM44wk12SUM44wk16SUM44wk20SUM44wk24SUM44wk24Y537SWT150020002500300035004000Channel1amplitude030002500200015001000500Channel2amplitude–5kb+5kbSummit–5kbwt-SUM44SUM44-LTEDChIPseqmRNAchangesNES:1.
495Nominalp-value:0Upregulatedinthewt-SUM44UpregulatedintheSUM44-LTED+5kbSummit–5kb+5kbSummit–5kb–5kb5kb5kb00Sequencelogop-valueNamewt-SUM44SUM44-LTEDCommon10–31310–15110–4010–34>0.
0510–2010–13210–7310–3310–1310–610–410–72ESR1RARAFOXA1PAX2GATA3ANDR10–4810–2110–410–5810–5Zerocrossat2647Rankinordereddataset05001000200030004000500015002500350045005500EnrichmentprofileRankingmetricscoresRankedlistmetric(PreRanked)Enrichmentscore(ES)Hits0.
200.
150.
100.
050.
00–0.
05–0.
10–0.
157.
55.
02.
50.
0–2.
5–5.
0Enrichmentplot:InducedbindingeventsinLTEDbadcefghwt-MCF7MCF7-LTEDSequenceRefSeqgenesFig.
1IdenticationandcharacterizationofESR1mutationsinmodelsofendocrineresistance.
aVisualizationofESR1Y537Cidentiedduringexomesequencing.
bDigitaldropletPCR(ddPCR)showingthepresenceoftheESR1Y537CmutationinMCF7-LTED.
cddPCRshowingthepresenceoftheESR1Y537SmutationinSUM44-LTED.
Temporalanalysisshowingenrichmentofthemutationfromwk12post-estrogendeprivation.
dddPCRshowingthepresenceofESR1Y537Satlowvariantallelefrequency(VAF)inwt-SUM44butnotinSKBR3.
eOverlapbetweenwt-SUM44andSUM44-LTEDESR1bindingsitesandcorrespondingheatmap.
Theheatmapdepictsbindingpeakintensities,whicharecommonordifferentbetweenthetwocelllines.
Thewindowrepresents±5kbregionsfromthecenterofthebindingevent.
fComparisonoftheaveragereadcountbetweenwt-SUM44andSUM44-LTEDshowingpeakafnityforthecommonanddifferentbindingeventsbetweenthetwocelllines.
gMotifanalysisofcommonandaugmentedESR1peaksfromwt-SUM44vs.
SUM44-LTED.
p-valueof"commonpeaks"basedonaverageofthreerandomselectionsof2150peakstoapproximatelymatchthenumberofpeakswithinthe"augmentedpeak"comparisons.
hGSEAwasconductedcomparingRNA-seqwithESR1-inducedbindingeventsinSUM44-LTED.
ChIP-seqanalysiswascarriedoutusingdatafromtwobiologicalreplicatesandRNA-seqfromthreebiologicalreplicatesNATURECOMMUNICATIONS|DOI:10.
1038/s41467-017-01864-yARTICLENATURECOMMUNICATIONS|8:1865|DOI:10.
1038/s41467-017-01864-y|www.
nature.
com/naturecommunications3ESR1Y537Sdrivesligand-independenttranscription.
Todeter-minethefunctionofESR1Y537S,weperformedChIP-seqwithantibodiesforESR1inasynchronouswt-SUM44inthepresenceofestrogenandSUM44-LTEDintheabsenceofestrogen.
Overlapoftworeplicateexperimentscalled28,647and23,294ESR1bindingeventsinwt-SUM44andSUM44-LTEDcells,respectively.
Thevastmajority(80%)oftheESR1Y537SbindingsitesinSUM44-LTEDcellswerecommontoESR1wtbindingsitesinestrogen-treatedwt-SUM44(Fig.
1e).
Although4702differ-entialbindingsiteswerecalledintheSUM44-LTEDcells,thesewerenotunique,butrepresentedenrichedESR1binding,i.
e.
,theyalsoappearedinwt-SUM44andthiswassimilarlythecaseforthe10,055differentialbindingsitesinwt-SUM44thatoccurredintheSUM44-LTEDbutwerenotenrichedtothesamelevel(Fig.
1f).
Peakstrengthwasevaluatedatanumberoftargetgenes(SupplementaryFig.
6a),whereaugmentedESR1Y537SbindingwasevidentinSUM44-LTEDcomparedtowt-SUM44.
Further-more,ChIP-qPCRvalidationassessingrecruitmentofESR1Y537StogetherwithFOXA1,amajorpioneerfactorforESR119andCBPrequiredforanauthenticESR1transcriptionalcomplex20,showedenhancedbindingtothepromotersofTFF1andGREB1intheSUM44-LTEDcomparedtowtcellline(SupplementaryFig.
6b).
ESR1bindingsitesinbothcelllinesshowedasimilarpatternofgenomicdistribution(SupplementaryFig.
6c).
Furthermore,thevastmajorityofbindingmotifsweresimilarforESR1wtandESR1Y537S,however,signicantenrichmentformotifsrepresent-ingthetranscriptionfactorsESR1,RARA,PAX2,ANDR,andFOXA1wereevidentinrelationtotheenrichedESR1peaksfoundinSUM44-LTED,comparedtowt-SUM44,whichconverselyshowedincreasedGATA3(Fig.
1g).
ToidentifythetranscriptiontargetsofESR1Y537S,weintegratedChIP-seqandRNA-seqdatafromtherespectivecelllines.
Genesetenrichmentanalysis(GSEA)showedthatincreasedESR1Y537Sgenomicbindingcorrelatedwithincreasedtranscription,whereaslossofbindingcorrelatedwithdown-regulationofgenesinSUM44-LTED(Fig.
1h;SupplementaryFig.
6d).
WenextusedK-meansclusteringtocomparetheESR1bindingpatternswithexpressionofgenesinwt-SUM44,wt-SUM44after1weekofestrogendeprivationandtheSUM44-LTED(20weeksofestrogendeprivation).
Weidentiedfourdistinctgenesets17(Fig.
2a–c):GS1consistedofclassicalestrogen-regulatedgenessuchasTFF1,GREB1,PGR,andCCND1,whichdecreasedinexpressionafter1weekofdeprivationbutwereelevatedintheSUM44-LTED.
GS4containedgenessuchasFOXA1thatwereenrichedaftertherstweekofestrogendeprivationandremainedactiveintheLTED.
GS2and3includedgenes,suchasMYCandJUN,whichweredownregulatedintheSUM44-LTEDcomparedtowt-SUM44.
PathwayanalysisofthefourclustersshowedenrichmentofESR1signaling,epithelial-to-mesenchymaltransition(EMT),mTORC1complexactivation,andcholesterolhomeostasisintheSUM44-LTED.
Toaddressthisfurther,weassessedthemetaboliccapabilityofthewt-SUM44andSUM44-LTEDusingSeahorse(Fig.
2d).
Nosignicantchangeinglutaminedependencywasevidentbetweenthetwocelllines;however,theSUM44-LTEDshowedasignicantlyhigherglutaminecapacityandfattyaciddependencycomparedtothewt-SUM44.
TheSUM44-LTEDalsoshowedaslightbutsignicantdecreaseinglucosedependency.
Finally,weassessedthemigratoryabilityofthecelllines(Fig.
2e).
TheSUM44-LTEDshowedatwo-foldincrease(p2–101Averagelog2differenceAveragelog2differenceAveragelog2differenceAveragelog2difference0.
40.
20.
0–0.
2–0.
4–0.
6–0.
80.
0–0.
2–0.
4–0.
6–0.
8–1.
0HALLMARK_P53_PATHWAYHALLMARK_EPITHELIAL_MESENCHYMAL_TRANSITIONHALLMARK_HYPOXIAHALLMARK_TNFA_SIGNALING_VIA_NFKBHALLMARK_GLYCOLYSISHALLMARK_KRAS_SIGNALING_UPHALLMARK_APOPTOSISHALLMARK_INTERFERON_GAMMA_RESPONSEHALLMARK_MTORC1_SIGNALINGHALLMARK_MYOGENESISHALLMARK_PANCREAS_BETA_CELLSHALLMARK_FATTY_ACID_METABOLISMHALLMARK_TNFA_SIGNALING_VIA-NFKBHALLMARK_ESTRIGEN_RESPONSE_EARLYHALLMARK_MTORC1_SIGNALINGHALLMARK_UNFOLDED_PROTEIN_RESPONSEHALLMARK_ESTROGEN_RESPONSE_LATEHALLMARK_HYPOXIAHALLMARK_P53_PATHWAYHALLMARK_UV_RESPONSE_DNHALLMARK_EPITHELIAL_MESENCHYMAL_TRANSITIONHALLMARK_GLYCOLYSISHALLMARK_IL2_STAT5_SIGNALINGHALLMARK_XENOBIOTIC_METABOLISMHALLMARK_UV_RESPONSE_UPHALLMARK_INTERFERON_GAMMA_RESPONSEHALLMARK_CHOLESTEROL_HOMEOSTASISHALLMARK_MITOTIC_SPINDLEHALLMARK_G2M_CHECKPOINTHALLMARK_P53_PATHWAYHALLMARK_E2F_TARGETSHALLMARK_HYPOXIAHALLMARK_IL2_STAT5_SIGNALINGHALLMARK_TNFA_SIGNALING_VIA_NFKBHALLMARK_ADIPOGENESISHALLMARK_APICAL_JUNCTIONHALLMARK_HEME_METABOLISMHALLMARK_ANDROGEN_RESPONSEHALLMARK_APOPTOSISHALLMARK_GLYCOLYSISHALLMARK_KRAS_SIGNALING_UPHALLMARK_UV_RESPONSE_DNHALLMARK_MTORC1_SIGNALINGHALLMARK_XENOBIOTIC_METABOLISMHALLMARK_E2F_TARGETSHALLMARK_HYPOXIAHALLMARK_PEROXISOMEHALLMARK_UV_RESPONSE_DNHALLMARK_G2M_CHECKPOINTHALLMARK_BILE_AICD_METABOLISMHALLMARK_INFLAMMATORY_RESPONSEHALLMARK_KRAS_SIGNALING_UPHALLMARK_MYOGENESISHALLMARK_CHOLESTEROL_HOMEOSTATISHALLMARK_APICAL_JUNCTIONHALLMARK_EPITHELIAL_MESENCHYMAL_TRANSITIONHALLMARK_KRAS_SIGNALING_DNSUM44-LTEDwt-SUM440100,000200,000300,000400,000123Geneset4+Ewk0–Ewk1–Ewk20+Ewk0–Ewk1–Ewk20+Ewk0–Ewk1–Ewk20Glutaminedependency-SUM44Fattyaciddependency-SUM44Fattyacidcapacity-SUM44Glutaminecapacity-SUM44wt-SUM44SUM44-LTEDwt-SUM44SUM44-LTEDwt-SUM44SUM44-LTEDwt-SUM44SUM44-LTEDwt-SUM44SUM44-LTEDwt-SUM44SUM44-LTED***50403020100504030201005040302010010080604020010080604020050403020100%ΔOCR(pmol/min)%ΔOCR(pmol/min)%ΔOCR(pmol/min)%ΔOCR(pmol/min)%ΔOCR(pmol/min)%ΔOCR(pmol/min)1weekafterremovalofE(+E_wk0vs–E_wk1)Additional19weeks–E(–E_wk1vs–E_wk20)+Ewk0–Ewk1–Ewk200.
40.
20.
0–0.
21.
21.
00.
80.
60.
40.
20.
0Glucosecapacity-SUM44Glucosedependency-SUM44***Fluorescenceunits***acbdeNATURECOMMUNICATIONS|DOI:10.
1038/s41467-017-01864-yARTICLENATURECOMMUNICATIONS|8:1865|DOI:10.
1038/s41467-017-01864-y|www.
nature.
com/naturecommunications5AsFOXA1isanimportantpioneerfactorregulatingESR1-driventranscription23,andFOXA1siteswereenrichedinourChIP-seqanalysisofSUM44-LTEDcells,wehypothesizedthatitplayedapivotalroleintranscriptionalregulationofESR1Y537S.
SmallinterferingRNA(siRNA)knockdownofFOXA1signicantlyreducedproliferationofbothwt-SUM44(42%,p201123Geneset–15–log10q-value0.
00010.
0010.
010.
11–1515–55log2foldchangeGENESETNAMEMCF7-LTEDwtvs.
wt-MCF7MCF7-LTEDY537Cvs.
wt-MCF7PercentoverlapMCF7-LTEDwtvs.
MCF7-LTEDY537C15–55log2foldchangeWT-LTEDwtLTEDWT-LTEDY537CWT-LTEDY537C5060403020100Percentoverlap5060403020100Percentoverlap50604030201000100,000200,000300,000400,000Fluorescenceunits******DistalintergenicPromoter(10–20kb)Promoter(5–10kb)Promoter(1–5kb)Promoter(≤1kb)5′UTR3′UTR1stexonOtherexon1stintronOtherintronDownstream(≤3kb)DistalintergenicPromoter(10–20kb)Promoter(5–10kb)Promoter(1–5kb)Promoter(≤1kb)5′UTR3′UTR1stexonOtherexon1stintronOtherintronDownstream(≤3kb)DistalintergenicPromoter(10–20kb)Promoter(5–10kb)Promoter(1–5kb)Promoter(≤1kb)5′UTR3′UTR1stexonOtherexon1stintronOtherintronDownstream(≤3kb)HALLMARK_ESTROGEN_RESPONSE_EARLYFDRq-valueFDRq-valueFDRq-valueFDRq-value4.
70E–305.
90E–283.
68E–121.
04E–084.
75E–081.
51E–071.
08E–061.
08E–061.
23E–052.
22E–052.
22E–053.
27E–175.
52E–131.
28E–121.
28E–126.
60E–122.
00E–105.
24E–091.
16E–072.
06E–072.
53E–071.
92E–062.
40E–201.
48E–167.
56E–141.
33E–101.
70E–106.
42E–092.
18E–082.
18E–082.
18E–082.
18E–082.
18E–089.
35E–413.
25E–351.
36E–126.
63E–123.
30E–111.
64E–101.
52E–094.
01E–087.
96E–081.
68E–064.
11E–06HALLMARK_ESTROGEN_RESPONSE_LATEHALLMARK_UV_RESPONSE_DNHALLMARK_TGF_BETA_SIGNALINGHALLMARK_P53-PATHWAYHALLMARK_WNT-BETA-CATENIN_SIGNALINGHALLMARK_EPITHELIALMESENCHYMAL_TRANSITIONHALLMARK_TNFA_SIGNALING_VIA-NFKBHALLMARK_APOPTOSISHALLMARK_HYPOXIAHALLMARK_IL2-STAT5-SIGNALINGGENESETNAMEGENESETNAMEGENESETNAMEHALLMARK_ESTROGEN_RESPONSE_EARLYHALLMARK_UNFOLDED_PROTEIN_RESPONSEHALLMARK_P53_PATHWAYHALLMARK_TNFA_SIGNALING_VIA_NFKBHALLMARK_HYPOXIAHALLMARK_ESTROGEN_REPONSE_LATEHALLMARK_MTORC1_SIGNALINGHALLMARK_GLYCOLYSISHALLMARK_INTERFERON_ALPHA_RESPONSEHALLMARK_UV_RESPONSE-DNHALLMARK_ANDROGEN_RESPONSEHALLMARK_TNFA_SIGNALING_VIA_NFKBHALLMARK_ESTROGEN_RESPONSE_LATEHALLMARK_ESTROGEN_RESPONSE_EARLYHALLMARK_UV_RESPONSE_UPHALLMARK_GLYCOLYSISHALLMARK_APICAL_JUNCTIONHALLMARK_COAGULATIONHALLMARK_EPITHELIAL_MESENCHYMAL_TRANSITIONHALLMARK_HYPOXIAHALLMARK_E2F_TARGETSHALLMARK_G2M_CHECKPOINTHALLMARK_MITOTIC_SPINDLEHALLMARK_MTORC1_SIGNALINGHALLMARK_GLYCOLYSISHALLMARK_MYC_TARGETS_V1HALLMARK_PEROXISOMEHALLMARK_UV-RESPONSE_UPHALLMARK_P13K_AKT_MTOR_SIGNALINGHALLMARK_P53_PATHWAYHALLMARK_FATTY_ACID_METABOLISMHALLMARK_IL2_STAT5_SIGNALINGHALLMARK_P53_PATHWAY4abcdefg*ARTICLENATURECOMMUNICATIONS|DOI:10.
1038/s41467-017-01864-y10NATURECOMMUNICATIONS|8:1865|DOI:10.
1038/s41467-017-01864-y|www.
nature.
com/naturecommunicationspartialresistanceofESR1Y537SbutnotESR1Y537C.
Nonetheless,atthepredictedclinicallyachievableconcentrationsoffulves-trant32,33,ESR1Y537SwasasequallysensitiveastheESR1wt.
Thisisinkeepingwithourpreviousclinicaldata,whichsuggestspatientsharboringanESR1mutationshowlongerprogression-freesurvivalwhentreatedwithfulvestrantvs.
exemestane13.
However,incontrasttoY537C,Y537Sshowedreducedsensitivityto4-OHT.
Oneexplanationfortheseobservationsisthat,4-OHTcausesY537StostabilizeH12bytheformationofahydrogenbondbetween537SandE380,effectivelyreducingthepotencyofthedrug.
Incontrast,bindingoffulvestrantdisordersH12.
Assuch,someofthenewSERM/SERDagentswithenhancedpharmacokineticscapableofincreasingthedynamicsofH12mayshowincreasedpotencyagainstthismutation25.
Interestingly,MCF7-LTEDwtshowevidenceofreducedESR1activity,withlowerexpressionofestrogen-regulatedgenessuchasPGRandincreasedexpressionofgenesassociatedwithanti-apoptoticactivity34.
Unexpectedly,LTEDcellsexpressingESR1wtwerealsolesssensitivetofulvestrantcomparedtoESR1Y537C.
OneexplanationisthatthesecellsalreadyhaveelevatedkinaseactivitiesandarethuslessdependentonESR1,highlightingonceagainthecomplexityofcellularcontextaswellasmutationstatusonresponsetoendocrinetherapy.
RecentgeneticstudiesthathaveidentiedESR1mutationsinmetastatic,endocrine-resistantBCindicatethatthesemutationsresultfromtheselectivepressureimposedbyinhibitionofESR1activitybyhormonaltherapies.
Theresultspresentedherepro-videsupportforthishypothesis.
TheindependentBCcelllinemodelsidentiedherealsoprovideanimportantresourceforstudyingtherelativecontributionofESR1mutationsandalterationsinothersignalingpathways,thatleadtoendocrineresistance.
Indeed,thegenomicstudiesdescribedhereinprovidesupportfortheimportanceofkinasesignalingcascadesthathavealreadybeenimplicatedinendocrineresistancebyourstudies,aswellasthoseofotherinvestigators.
OurndingsdemonstratethatESR1mutationsprovideanimportant,albeitnottheonlydriverofacquiredendocrineresistance,concordantwiththeclinicalobservationthat~20%ofmetastatictumorsharbormutantESR1.
UsingresistancemodelsfeaturingESR1mutationsandthosethatdonotinvolveESR1mutationsshouldprovetobevaluableinaidingpatientmanagement,andforassessingnewtreatmentapproachesforendocrine-resistantBC.
Weandotherswillneedtoconsiderthepresenceandanyphenotypiceffectsoftheseandpossiblyotheracquired/selectedmutationswhenusingthesederivedcelllinesformechanisticorpharmacologicalstudiesandinterpretingdatafromthem.
MethodsReagents.
Followingantibodieswereusedforimmunoblotting:pESR1ser167(CSTcat-5587,1:1000),pESR1ser118(CSTcat-2511,1:1000),total-ESR1(SantaCruzsc8002,1:800orNovacastra(NCL-ER-6F11),1:1000),total-FOXA1(AbcamAb23738,1:1000)total-PGR(NovocastraNCL-L-PGR,1:500orSantaCruzsc-538,1:200),pERBB2(CST-2243,1:1000),total-ERBB2(CST-4290,1:1000),pEGFR(CST-3777,1:1000),total-EGFR(CST-2232,1:1000),pAKTser437(CST-9271,1:1000),total-AKT(CST-9272,1:1000),pERK1/2(Sigma-Aldrich,1:2000),total-ERK1/2(CST-9102,1:1000),TFF1(SantaCruzsc28925,1:200),RARA(AbcamAb39971,1:1000),cathepsinD(CTSD)(AbcamAb6313,1:2000),actin(AbcamAb6276,1:10,000),andtubulin(SigmaT-9026,1:2000).
Secondaryantibodies(horseradishperoxidase-linked,1:2000)wereobtainedfromDako.
ForChIP,thefollowingantibodieswereused:ESR1(SantaCruzsc543),CBP(SantaCruzsc369),andFOXA1(AbcamAb23738).
17-β-estradiol(E)and4-hydroxytamoxifen(4-OHT)werepurchasedfromSigma-Aldrichandfulvestrant(ICI182780)fromTocrisBioscience.
Cellculture.
Wt-MCF7,wt-HCC1428,wt-ZR75.
1,andwt-SUM44werepurchasedfromtheATCCandAsterand.
Celllineswerebankedinmultiplealiquotsuponreceipttoreduceriskofphenotypicdriftandidentityconrmedbyshorttandemrepeats(STR)proling.
Allcelllineswereroutinelyscreenedformycoplasmacontamination.
Wtcelllineswereculturedinphenolred-freeRPMIsupplementedwith10%fetalbovineserum(FBS)andexogenousestradiol(1nM).
TherespectiveLTEDderivativeswerecultured,aspreviouslydescribed14,15inphenolred-freeRPMIsupplementedwith10%dextrancharcoalstrippedFBS(DCCmedium).
ICI-RandTAMRcelllineswereculturedintheirrespectivebasalmediumsupple-mentedwith100nMfulvestrant(ICI182780)or100nM4-OHT.
Allexperimentswereperformedunderbasalconditionsunlessotherwisestated.
Proliferationassays.
ProliferationassayswereperformedaspreviouslydescribedforexperimentsinvolvingdrugsandsiRNAstudies14,15.
Insummary,cellsweredeprivedofestrogenfor48–72hpriortotreatmentwithOn-targetplussiRNAforhuman-siFOXA1ornon-targetingpool(sicontrol)(Thermoscientic,Dharma-con).
KnockdownefcacywasdeterminedbyqRT-PCR.
Fordrugstudies,cellsweretreatedfor6dayswithamediumchangeatday3,aspreviouslydescribed14.
Toanalyzegrowthovertime,cellswereculturedasdetailedaboveinDCCmediumwithorwithoutestradiolanddatarecordedusinganIncuCyteZOOMlivecellanalyzer(EssenBiomedics).
Threeimagesperwellweretakenevery12hovera6-dayperiod.
qRT-PCR.
RNAwasextractedusingtheRNeasykit(Qiagen),quantiedandreverse-transcribedwithSuperScriptIIIFirstStrandSynthesisSystem(Invitrogen).
Taqmangeneexpressionassays(AppliedBiosystems)wereusedtoquantifyTFF1(Hs00907239_m1andHs00170216_m1),PGR(Hs00172183_m1),GREB1(Hs00536409_m1),CTSD(Hs00157201_m1),ESR1(Hs00174860_m1),CCND1(Hs00765553_m1),andthehouse-keepinggenesFKBP15(Hs00391480_m1)andGAPDH(Hs99999905_m1).
TherelativequantitywasdeterminedusingΔΔCt,accordingtothemanufacturer'sinstructions(AppliedBiosystems).
Exomesequencing.
ExomelibrariesweregeneratedwithSureSelectHumanAllExonV5kitandsequenced(paired-end100bp)onanIlluminaHiSeq2500.
ReadswerealignedtoGRCh37-lite-build37usingBWAmem(v0.
7.
12-r1039)35,sortedwithsamtools(v1.
2)36andfurtherprocessedusingpicardtools(http://picard.
sourceforge.
net)(v1.
128)withdefaultparameters.
Singlenucleotidevariants(SNVs)weredetectedusingVarScanv2.
3.
537withdefaultparameters(except--mpileup1,--output-vcf)andwtcellsamplesasbaseline.
Multi-mappedreadswereexcludedandbasealignmentquality(BAQ)wasturnedoffforpileupwithsamtools.
Togethighcondencesomaticmutations,SNVswerelteredbyusing:(i)processSomaticofVarScanwithempirically-derivedcriteria:minimumVAFinLTEDcells:0.
10,maximumVAFinwt:0.
05,p-value=0.
07;(ii)fplter.
plfromVarScantogetherwithbam-readcount(--min-base-quality15,--min-mapping-quality1)toreducenumberoffalsepositives.
VariantswereannotatedusingSnpEffv4.
1B(http://snpeff.
sourceforge.
net/SnpEff_manual.
html).
MutationswereannotatedwithTierlevels38usingBedToolsv2.
22.
139.
ascatNGS(https://github.
com/cancerit/ascatNgs)wasusedtogenerateLogRandBAFvalues.
Datahavebeendepositedinthesequencereadarchive:BioProjectIDPRJNA390496.
Iontorrent.
DNAwasampliedusingIonAmpliSeqLibraryKit2.
0(LifeTech-nologies),digested,IonXpress.
BarcodeadaptersligatedandpuriedwithAgencourtAMPureXPmagneticbeads(BeckmanCoulter).
LibrarieswerequantiedbyqPCRusinganIonLibraryQuanticationKit(LifeTechnologies),templatedontheIonOneTouch2SystemFig.
5ESR1wtandESR1Y537CregulatealteredESR1cistrome.
aBarchartshowingthegenomicdistributionofESR1bindingsitesacrossthegenomeinwt-MCF7,MCF7-LTEDwt,andMCF7-LTEDY537Cshowingalteredpromoter(≤1kb)anddistalintergenicoccupancy.
bVolcanoplotsshowingchangesingeneexpressionbyRNA-seqinMCF7-LTEDY537C,MCF7-LTEDwt,andwt-MCF7.
cHeatmapdepictingthechangesingeneexpressionofthefourclusterscomparingwt-MCF7toMCF7-LTEDY537C,wt-MCF7toMCF7-LTEDwt,andMCF7-LTEDY537CtoMCF7-LTEDwt.
dAveragelog2differencesforallgeneswithineachsetforwt-MCF7,MCF7-LTEDwt,andMCF7-LTEDY537C.
ePathwayanalysisofthefourclustersusingGSEA.
Datawerederivedfromn=2biologicalreplicatesforChIP-seqandn=3biologicalreplicatesforRNA-seq.
fMetabolicdependencyandcapacityofwt-MCF7,MCF7-LTEDwt,andMCF7-LTEDY537Conglutamine,fattyacid,andglucoseusingaSeahorseXFe96analyzer(n=4technicalreplicates).
Signicancewasassessedbyone-wayANOVAandTukey'stest.
*p<0.
05,**p<0.
01,***p<0.
001.
gComparisonofthemigratoryabilityofwt-MCF7,MCF7-LTEDwt,andMCF7-LTEDY537C(n=8technicalreplicates).
Datashownaremean±SEM.
SignicancewasassessedbyStudent'sttest.
*p<0.
05,**p<0.
01,***p<0.
001NATURECOMMUNICATIONS|DOI:10.
1038/s41467-017-01864-yARTICLENATURECOMMUNICATIONS|8:1865|DOI:10.
1038/s41467-017-01864-y|www.
nature.
com/naturecommunications11(LifeTechnologies)andsequencedontheIonPGMSystem(LifeTechnologies).
ReadswerealignedbythePGMserverwithstandardsettingstothereferencegenomehg19,samtoolsv1.
2wasusedtocalculatetheon-targetcoverage.
IonReporter(v4.
4)wasusedformutationcalling(parameters:dataqualitystringency=12,downsampletocoverage=4000,SNP/InDel/MNPmincoveachstrand=50,SNP/InDel/MNPminvariantscore=15,SNP/InDel/MNPmincoverage=250,hotspotminvariantscore=6,hotspotmincoverage=150).
AllmutationscalledweremanuallyreviewedinIntegrativeGenomicsViewer(IGV)andincludedintheanalysisiftheyhadaVAF≥1%.
ddPCR.
ddPCRassaysfortheESR1mutationsY537SandY537CusingTaqmanprobeswereusedaspreviouslydescribed10.
Very-low-frequencymutationswereonlyconsideredtobepresentiftwoormoreFAM-positivedropletsweredetectedinthetotalofthewtsample.
Cyclesequencingforvalidation.
ESR1mutationswerevalidatedbycyclesequencingbyeuronsgenomics(Eurons).
DNAwasampliedusingforwardprimer5-AAGTGGCTGCAGGGAGAGT-3andreverseprimer5-TGGTGCATGATGAGGGTAAA-3.
Fluorescenceinsituhybridization.
FISHprobeshybridizingat6q25(ESR1)andchromosome-6(CEN6)werepurchasedfromEmpireGenomics.
Cellpelletswerexedin4%paraformaldehydeandparafn-embedded.
Five-micronsectionsweresubjectedtotheSwiftFISHrapidhybridizationprotocol(EmpireGenomics),accordingtothemanufacturer'sinstructions.
SectionsweremountedinDAPI-containingVectashield(Vector).
FISHprobessignalswereanalyzedusinguor-escentmicroscope(Leica).
RNA-seq.
LibrarieswerecreatedafterRibo-zerorRNARemovalKit(Illumina)usingNEBNextUltraDirectionalRNA(NEB)orTruseqStrandedTotalRNA(Illumina)LibraryPrepKitandsequencedusingtheHiSeq2500(paired-end100bpv4chemistry).
Tophat(v2.
1)andCuffdiff(v2.
2.
1)40usingdefaultparameters(GSE100075).
K-meansclusteringwasperformedusingthek-meansfunctioninthestatspackageinR.
Thenumberofclustersusedwasdeterminedbythenumber1.
21.
00.
80.
60.
40.
20.
01.
21.
00.
80.
60.
40.
20.
0wt-SUM44SUM44-LTEDwt-MCF7MCF7-LTEDwtE(0.
01nM)ICI(10nM)E(0.
01nM)–––++++++++–+++–++++76kDa76kDa52kDa52kDa76kDawt-SUM44wt-MCF7SUM44-LTEDY537SMCF7-LTEDY537CMCF7-LTEDwt76kDa52kDa52kDaICI(10nM)ESR1ESR1TubulinTubulinMCF7-LTEDY537C1.
21.
00.
80.
60.
40.
20.
0Cellviability(foldchange)1.
21.
00.
80.
60.
40.
20.
06*1054*1052*1050Cellviability(foldchange)1.
21.
00.
80.
60.
40.
20.
0Cellviability(foldchange)1.
21.
00.
80.
60.
40.
20.
0Cellviability(foldchange)Cellviability(Absluminescence)Cellviability(foldchange)Cellviability(foldchange)Fulvestrant-DCC(nM)Fulvestrant-DCC(nM)Daysafterstartoftreatmentp<0.
0011.
51.
00.
50.
0Medianrelativetumorvolume047111721283235Fulvestrant+0.
01nME(nM)Fulvestrant+0.
01nME(nM)4-OHT+0.
01nME(nM)00.
0100.
10011010000.
0100.
10011010000.
0100.
1001101004-OHT+0.
01nME(nM)E(nM)00.
0110.
110100100000.
0110.
11010000.
0110.
11010000.
00010.
010.
0010.
11VehicleFulvestrantabcdFig.
6AntiproliferativeeffectofendocrinetherapyinESR1mutantandwtcelllines.
aProliferationassaysassessingresponseofwt-SUM44andSUM44-LTEDandbwt-MCF7,MCF7-LTEDwt,andMCF7-LTEDY537Ctoescalatingconcentrationoffulvestrant±E(estradiol)and4-OHTplusE(estradiol).
cTreatmentofwt-SUM44,SUM44-LTEDY537S,wt-MCF7,MCF7-LTEDY537C,andMCF7-LTEDwtwithfulvestrant(10nM)resultsinlossofESR1expressionirrespectiveofmutationstatus(n=3biologicalreplicatesconsistingofn=8technicalreplicates).
Datarepresentmean±SEM.
dXenograftmodelsofMCF7-LTEDY537Cinresponsetovehicleorfulvestrant.
Datarepresentmedianfoldchangeintumorvolume.
Signicancewasassessedusinganunpairedt-testARTICLENATURECOMMUNICATIONS|DOI:10.
1038/s41467-017-01864-y12NATURECOMMUNICATIONS|8:1865|DOI:10.
1038/s41467-017-01864-y|www.
nature.
com/naturecommunicationsofclustersgeneratedinunsupervisedclusteringusinghclust(method=complete)functioninRwithofamatrixofcorrelation-baseddistancesusingtheSpearmanmethod.
ChIP-seq.
ChIP-qPCRandChIP-seqwereperformed,aspreviouslydescribed14,41.
Paired-end50bpChIP-seqdataweregeneratedbyrapid-modeHiSeq.
ReadswerealignedtotheHumanReferenceGenome(assemblyhg19)usingBWA35removingallreadswithaqualityscore<15.
PeakswerecalledusingMACS2(v2.
1.
0.
20150420)42withdefaultparameters.
OnlybindingeventsthatoccurredintwobiologicalreplicateswereconsidereddifferentialbindingsitesusingDiffbindv1.
14.
543andRv3.
2.
1.
Motifanalysiswasperformedusingcentrimo(500bpcenteredonsummitofpeak)(http://meme-suite.
org/)(GSE100074).
BarchartsweregeneratedwithChIPseekerpackageinR44.
GSEA.
IntegrationofRNA-seqandChIP-seqdiffBinddatawerecarriedoutusingGSEA,aspreviouslydescribed45.
Insummary,allgenesassessedusingRNA-seqwererankedandweightedbytheirmeanLog2foldchange.
Listsofgenesthatoverlappedwithregionsshowingsignicantdifferentialbindingwereidentied.
ThesedatawerethenanalyzedusingtheGSEAv2.
0.
13GSEAPre-rankedtool.
Thedefaultsettingwasapplied.
Finally,additionalanalysisofgenesets(e.
g.
,overlapsbetweensignicantbindingeventsandclosestgenesthataresignicantlydiffer-entiallyexpressed)wereperformedusingtheMolecularSignatureDatabase(http://software.
broadinstitute.
org/gsea/msigdb/annotate.
jsp)tocomputeoverlapswithHallmarkgenesetsthatrepresentwell-denedbiologicalstatesorprocesses.
Sig-nicanceofoverlapbetweengenesetswasdeterminedbyhypergeometrictest.
RIMEanddimethyllabeling.
RIME22andstableisotopedimethyllabeling21wereperformed,aspreviouslydescribed.
Thewt-SUM44andSUM44-LTEDwerelabeledwiththemediumandlightisotopereagent,respectively.
Labeledsampleswerepooledatanapproximate1:1ratio,drieddownandfractionatedusing12cmIPGstrippH3–10,aspreviouslydescribed46.
RIMEanddimethyllabelfractionsweredesalted(SUMSS18V,TheNestGroupInc)andrunthroughLC-MS/MSusingLTQVelosOrbitrapMS.
Thedataacquisitionmodewasset,aspreviouslydescribed46.
RawdataforRIMEanddimethyllabelingwereanalyzedusingMaxQuant1.
5.
1.
046,47.
Searchparameterswereaspreviouslydescribed46.
AllproteomicsdataaredepositedwithinthePRIDEdatabase(PXD004807).
IdenticationofmutationatproteinlevelusingddMS2/PRM.
ESR1-RIMEsamplesweresubjectedtoddMS2-PRManalysisinordertoverifythepresenceofwtandmutatedserineorcysteineintheSUM44-LTEDandMCF7-LTEDsamples,respectively(SupplementaryData3).
TheanalysiswasperformedusingaQ-ExactiveHFmassspectrometer(ThermoScientic,HemelHempstead,UK).
Foreachanalysis,threebiologicalreplicateswithtwotechnicalreplicateswererun.
HeavypeptideswerepurchasedfromThermoFischerScientic(PEPOTEC,grade3).
ReversedphasechromatographywasperformedonaDionexUltiMate3000RSLCnanosystem(ThermoFisherScientic,HemelHempstead,UK)usinganAcclaimPepMap100C18trapcartridge(0.
5mmi.
d.
*5mm,5mbeadsize,100poresize;loadedinabi-directionalmanner).
Peptideswereresolvedona75mI.
D.
50cmC18Easy-Spraypackedemittercolumn(2mparticlesize;PepMapRSLC,ThermoScientic,HemelHempstead,UK)over90minusingathree-stepgradientof96:4to50:50bufferA:B(t=0min4%B,0.
5min4%B,12.
0min10%B,43.
0min25%B,90.
0min50%B)(bufferA:2%acetonitrile/0.
1%formicacid;bufferB:80%acetonitrile/0.
1%formicacid)at250nlpermin.
Peptideswereiodizedbyelectrosprayionizationusing1.
8kVappliedusingtheEasy-SprayionSource.
Samplewasinfusedintothemassspectrometerdirectlyfromthepackedemitter(5mexitbore).
Theiontransfertubewasheatedto275°CandtheS-lenssetto50%.
MS/MSwereacquiredusingparallelreactionmonitoring(PRM)anddata-dependent(ddMS2)acquisitionsbasedonafullFT-MSscanfrom350to1850m/zat120,000resolution,withatargetautomaticgaincontrol(AGC)valueof3,000,000andamaximuminjectiontimeof50ms.
Nointernallockmasscalibrantwasused.
EightPRMscansweretriggered(FT-Orbitrapscansat30,000resolution,AGCtarget2e5,100msmaximuminjectiontime,normalizedcollisionenergy35)ifanionfromscheduledinclusionlistwaspresent.
Then,thetopvemostintenseionswerefragmentedbyhigherenergycollision-induceddissociationanddyna-micallyexcludedfor20s(FT-Orbitrapscansat30,000resolution,AGCtarget1e5,activationtime10ms,50msmaximuminjectiontime,normalizedcollisionenergy28,selectedrstmassat140m/z).
Precursorionswithunknownorsinglechargestateswereexcludedfromselection.
DataanalysisofrawMS/MSwascarriedoutusingMascotV2.
3viaProteomeDiscovererv1.
4.
PeaklistsweresearchedagainstthehumanUniprotFASTAdatabase(20,305sequences)containingthewtandmutantsequence.
Spectraweresearchedforamatchtofully-trypticpeptideswithuptotwomissedcleavagesites.
Searchparameterswerechosenasfollows:serine/threoninephosphorylation,proteinN-terminalacetylation,peptideN-terminalglutaminetopyroGlu,andoxidationofmethionineswereallconsideredasvariablemodications,whereascysteinecarbamidomethylationwasselectedasaxedmodication.
Precursorionmasstolerancewassetto15ppmfortherstsearch,fragmentionmasstoleranceforionanalyzedspectrawassetto0.
02Da.
ResultingpeptideandproteinlistsweregroupedandvalidatedusingScaffoldv4(ProteomeSoftwareInc.
,Portland,OR).
Proteinidenticationswereautomaticallyacceptediftheycontainedatleasttwouniquepeptidesassignedat1%FDR.
TherawdatahavebeendepositedinPassel(PASS01062).
Immunoblotting.
Whole-cellextractsweregeneratedfromcellsculturedunderbasalconditionsorDCCmediumwithorwithouttheadditionofestrogenforcomparativestudieswherenoted.
EqualamountsofproteinwereresolvedbySDS–PAGEandsubjectedtoimmunoblotanalysis.
Antigen–antibodyinteractionsweredetectedwithECLreagent(Amersham,UK)usingtheantibodiesreferredabove.
CRISPR-Cas9-mediatedgenerationoftheMCF7-LTEDΔ537Ccells.
GeneknockinsforamodiedESR1exon8,encodingawtopenreadingframewithsilentmutationstofacilitatePCRanalysis,weremadeusingCRISPR-Cas9-mediatedhomologousrecombinationinMCF7-LTEDY537Ccells.
ESR1genetargetingwascarriedoutusingCRISPR4834093(5-GAGTGCTGAAATCCCTAGAA-3)clonedintoaguide-RNAexpressionplasmid(agiftfromGeorgeChurch;Addgeneplasmid#41824),asdescribedpreviously24.
ThetargetsequenceforthisCRISPRislocatedinintron7,ontheantisensestrand,73ntfromthestartofESR1Exon8.
Formakingthegeneknockin,apreviouslydescribedESR1exon8Y537Sgenetargetingdonorconstruct24wasmodiedbysite-directedmutagenesistochangecodon537fromSerine(TCT)toTyrosine(TAT),asfoundinthewtsequence.
Additionalmutations,todestroythePAMforCRISPR4834093,weremadebychangingarunoffourCnucleotides,located77nt5tothestartofESR1Exon8,tofourGnucleotides.
Genomeediting,detectionofgenetargetingevents,andsequencecharacterizationofgenetargetedalleleswerecarriedoutasdescribedpreviously24,withtheexceptionthatfollowingtransfection,cellswererecoveredinfullmediumsupplementedwith10%FCS,andExon8knockinclonesidentiedthroughstochasticcloning.
EnergyphenotypeandMitoFuelFlexanalysis.
SUM44andMCF7cellswereplatedataconuencyof1.
0*104perwellina96-wellSeahorsecellculturemicroplatesandincubatedina5%CO2incubatorat37°Covernight.
Thenextmorning,culturemediawasreplacedwithpH-adjusted(pH=7.
4±0.
1)bicarbonate-freeDMEMwith10mMglucose,1mMsodiumpyruvate,and2mML-glutamine.
Theplatewasthenincubatedat37°Cfor1hinanon-CO2incubator.
FortheMitoFuelFlextest,oxygenconsumptionratesweremeasuredusingtheSeahorseXFMitoFuelFlexTestKit(Agilent,103260-100)onanXFe96Analyzer.
CellnumberswerenormalizedusingCyQuant(ThermoFisher,C35012).
Cellmigrationassay.
Cellsgrowinginbasalmediawerewashedseveraltimeswithphenolred-freeRPMI1640containing1%DCC-FBS.
Atotalof2.
5*104cellswereseededintotheupperchambersofCorningFluoroBlok96-multiwellinsertsystemplates(Corning,UK).
ThelowerchamberswerelledwithRPMI1640containing1%DCC-FBSplus100ng/mlhumanrecombinantEGF,aschemo-attractant,andplateswereincubatedat37°C.
After16h,themediumwasremovedfromthelowerchambersandwellswerewashedwithPBS.
PBScontaining1μMcalceinAM(Invitrogen)wasaddedtothelowerchambersandtheplateswereincubatedat37°Cfor30min.
Fluorescenceintensitywasmeasuredfromthebottomoftheplatesusinga490nmexcitationlteranda520nmemissionlterinaVictorX5platereader(PerkinElmer).
Dataareexpressedasthemeanofeighttechnicalreplicates.
HumantumorxenograftsmodelingrelapseonAItherapy.
Invivostudieswerecarriedoutinovariectomized8-to12-week-oldfemaleBALB/cnudemiceinaccordancewithHomeOfceGuidelinesandapprovedbytheInstituteofCancerResearchEthicsCommittee.
XenograftsmodelingpatientsresistanttoAIwereinitiatedbyinnoculatingMCF7-LTEDY537C(107)cellsinbasementmembranematrix(Matrigel;BDBiosciences)intotherightankofeachanimal.
Oncetumorsreached7mminsize,theyweresizematchedandmicetreatedwitheither5mgperkgfulvestrantonceperweekorvehiclecontrol.
Thestudyoperatorwasblindedtotreatment.
Tumorgrowthwasassessedtwiceweeklyinbotharmsbycalipermeasurementsofthetwolargestdiameters.
Volumeswerethencalculatedaccordingtotheformula:a*b2*π/6,whereaandbareorthogonaltumordia-meters.
Tumorvolumeswerethenexpressedasmedianrelativefoldchangeinvolumeatthestartoftreatment.
Attheendofstudy,datawereavailableforsevenanimalsinthecontrolarmandnineanimalsinthefulvestranttreatmentarm.
Overallstatisticaldifferencesbetweenthetreatmentandcontrolarmswerecal-culatedusinganunpairedt-test.
Statisticsanalysis.
Statisticalmethodologiespertinenttoeachmethodareheldwithinthesectionsabove.
Dataavailability.
Thedatasupportingthendingfromthismanuscripthavebeendepositedasfollows.
Whole-exomesequencinghasbeendepositedinthesequencereadarchiveBioProjectIDPRJNA390496.
RNA-seqandChIP-seqdatahavebeendepositedwiththeNCBIgeneexpressionomnibus(GEO)(http://ncbi.
nlm.
nih.
gov/geo/):ChIP-seqdataforwt-MCF7,MCF7-LTEDwt,MCF7-LTEDY537Cwt-SUM44andSUM44-LTED(GSE100074),RNA-seq(GSE100075)forwt-MCF7,NATURECOMMUNICATIONS|DOI:10.
1038/s41467-017-01864-yARTICLENATURECOMMUNICATIONS|8:1865|DOI:10.
1038/s41467-017-01864-y|www.
nature.
com/naturecommunications13MCF7-LTEDwt,MCF7-LTEDY537C,wt-SUM44andSUM44-LTED,CRISPR-cas9MCF7Y537SChIP-seqandRNA-seqdata(GSE78286)24.
AllproteomicsdatasetsaredepositedwithinthePRIDEdatabase(PXD004807)orPassel(PASS01062)fortargetedsequencing.
Received:13October2016Accepted:20October2017References1.
Green,K.
A.
&Carroll,J.
S.
Oestrogen-receptor-mediatedtranscriptionandtheinuenceofco-factorsandchromatinstate.
Nat.
Rev.
Cancer7,713–722(2007).
2.
Ma,C.
X.
,Reinert,T.
,Chmielewska,I.
&Ellis,M.
J.
Mechanismsofaromataseinhibitorresistance.
Nat.
Rev.
Cancer15,261–275(2015).
3.
EarlyBreastCancerTrialists'CollaborativeGroup.
Aromataseinhibitorsversustamoxifeninearlybreastcancer:patient-levelmeta-analysisoftherandomisedtrials.
Lancet386,1341–1352(2015).
4.
Osborne,C.
K.
etal.
RoleoftheestrogenreceptorcoactivatorAIB1(SRC-3)andHER-2/neuintamoxifenresistanceinbreastcancer.
J.
Natl.
CancerInst.
95,353–361(2003).
5.
Miller,T.
W.
,Balko,J.
M.
&Arteaga,C.
L.
Phosphatidylinositol3-kinaseandantiestrogenresistanceinbreastcancer.
J.
Clin.
Oncol.
29,4452–4461(2011).
6.
Fuqua,S.
A.
,Gu,G.
&Rechoum,Y.
Estrogenreceptor(ER)alphamutationsinbreastcancer:hiddeninplainsight.
BreastCancerRes.
Treat.
144,11–19(2014).
7.
Jeselsohn,R.
,Buchwalter,G.
,DeAngelis,C.
,Brown,M.
&Schiff,R.
ESR1mutations-amechanismforacquiredendocrineresistanceinbreastcancer.
Nat.
Rev.
Clin.
Oncol.
12,573–583(2015).
8.
Hortobagyi,G.
N.
etal.
Correlativeanalysisofgeneticalterationsandeverolimusbenetinhormonereceptor-positive,humanepidermalgrowthfactorreceptor2-negativeadvancedbreastcancer:resultsfromBOLERO-2.
J.
Clin.
Oncol.
34,419–426(2016).
9.
Wang,P.
etal.
Sensitivedetectionofmono-andpolyclonalESR1mutationsinprimarytumors,metastaticlesions,andcell-freeDNAofbreastcancerpatients.
Clin.
CancerRes.
22,1130–1137(2016).
10.
Schiavon,G.
etal.
AnalysisofESR1mutationincirculatingtumorDNAdemonstratesevolutionduringtherapyformetastaticbreastcancer.
Sci.
Transl.
Med.
7,313ra182(2015).
11.
Jeselsohn,R.
etal.
Emergenceofconstitutivelyactiveestrogenreceptor-alphamutationsinpretreatedadvancedestrogenreceptor-positivebreastcancer.
Clin.
CancerRes.
20,1757–1767(2014).
12.
Toy,W.
etal.
ESR1ligand-bindingdomainmutationsinhormone-resistantbreastcancer.
Nat.
Genet.
45,1439–1445(2013).
13.
Fribbens,C.
etal.
PlasmaESR1mutationsandthetreatmentofestrogenreceptor-positiveadvancedbreastcancer.
J.
Clin.
Oncol.
34,2961–2968(2016).
14.
Ribas,R.
etal.
AKTantagonistAZD5363inuencesestrogenreceptorfunctioninendocrineresistantbreastcancerandsynergiseswithfulvestrant(ICI182780)invivo.
Mol.
CancerTher.
14,2035–2048(2015).
15.
Ribas,R.
etal.
Identicationofchemokinereceptorsaspotentialmodulatorsofendocrineresistanceinoestrogenreceptor-positivebreastcancers.
BreastCancerRes.
16,447(2014).
16.
Ellis,M.
J.
etal.
Whole-genomeanalysisinformsbreastcancerresponsetoaromataseinhibition.
Nature486,353–360(2012).
17.
Nik-Zainal,S.
etal.
Landscapeofsomaticmutationsin560breastcancerwhole-genomesequences.
Nature534,47–54(2016).
18.
Sikora,M.
J.
etal.
WNT4mediatesestrogenreceptorsignalingandendocrineresistanceininvasivelobularcarcinomacelllines.
BreastCancerRes.
18,92(2016).
19.
Carroll,J.
S.
etal.
Chromosome-widemappingofestrogenreceptorbindingrevealslong-rangeregulationrequiringtheforkheadproteinFoxA1.
Cell122,33–43(2005).
20.
Shang,Y.
,Hu,X.
,DiRenzo,J.
,Lazar,M.
A.
&Brown,M.
Cofactordynamicsandsufciencyinestrogenreceptor-regulatedtranscription.
Cell103,843–852(2000).
21.
Boersema,P.
J.
,Raijmakers,R.
,Lemeer,S.
,Mohammed,S.
&Heck,A.
J.
Multiplexpeptidestableisotopedimethyllabelingforquantitativeproteomics.
Nat.
Protoc.
4,484–494(2009).
22.
Mohammed,H.
etal.
EndogenouspuricationrevealsGREB1asakeyestrogenreceptorregulatoryfactor.
CellRep.
3,342–349(2013).
23.
Hurtado,A.
,Holmes,K.
A.
,Ross-Innes,C.
S.
,Schmidt,D.
&Carroll,J.
S.
FOXA1isakeydeterminantofestrogenreceptorfunctionandendocrineresponse.
Nat.
Genet.
43,27–33(2011).
24.
Harrod,A.
etal.
GenomicmodellingoftheESR1Y537Smutationforevaluatingfunctionandnewtherapeuticapproachesformetastaticbreastcancer.
Oncogene36,2286–2296(2017).
25.
Fanning,S.
W.
etal.
EstrogenreceptoralphasomaticmutationsY537SandD538Gconferbreastcancerendocrineresistancebystabilizingtheactivatingfunction-2bindingconformation.
eLife5,e12792(2016).
26.
Zhang,Q.
X.
,Borg,A.
,Wolf,D.
M.
,Oesterreich,S.
&Fuqua,S.
A.
Anestrogenreceptormutantwithstronghormone-independentactivityfromametastaticbreastcancer.
CancerRes.
57,1244–1249(1997).
27.
Li,S.
etal.
Endocrine-therapy-resistantESR1variantsrevealedbygenomiccharacterizationofbreast-cancer-derivedxenografts.
CellRep.
4,1116–1130(2013).
28.
Robinson,D.
R.
etal.
ActivatingESR1mutationsinhormone-resistantmetastaticbreastcancer.
Nat.
Genet.
45,1446–1451(2013).
29.
Merenbakh-Lamin,K.
etal.
D538Gmutationinestrogenreceptor-alpha:anovelmechanismforacquiredendocrineresistanceinbreastcancer.
CancerRes.
73,6856–6864(2013).
30.
Bhat-Nakshatri,P.
etal.
AKTaltersgenome-wideestrogenreceptoralphabindingandimpactsestrogensignalinginbreastcancer.
Mol.
CellBiol.
28,7487–7503(2008).
31.
Lupien,M.
etal.
GrowthfactorstimulationinducesadistinctER(alpha)cistromeunderlyingbreastcancerendocrineresistance.
GenesDev.
24,2219–2227(2010).
32.
McCormack,P.
&Sapunar,F.
Pharmacokineticproleofthefulvestrantloadingdoseregimeninpostmenopausalwomenwithhormonereceptor-positiveadvancedbreastcancer.
Clin.
BreastCancer8,347–351(2008).
33.
Pritchard,K.
I.
etal.
ResultsofaphaseIIstudycomparingthreedosingregimensoffulvestrantinpostmenopausalwomenwithadvancedbreastcancer(FINDER2).
BreastCancerRes.
Treat.
123,453–461(2010).
34.
Gomez,B.
P.
etal.
HumanX-boxbindingprotein-1confersbothestrogenindependenceandantiestrogenresistanceinbreastcancercelllines.
FASEBJ.
21,4013–4027(2007).
35.
Li,H.
&Durbin,R.
FastandaccurateshortreadalignmentwithBurrows-Wheelertransform.
Bioinformatics25,1754–1760(2009).
36.
Li,H.
etal.
TheSequenceAlignment/MapformatandSAMtools.
Bioinformatics25,2078–2079(2009).
37.
Koboldt,D.
C.
etal.
VarScan:variantdetectioninmassivelyparallelsequencingofindividualandpooledsamples.
Bioinformatics25,2283–2285(2009).
38.
Mardis,E.
R.
etal.
Recurringmutationsfoundbysequencinganacutemyeloidleukemiagenome.
N.
Engl.
J.
Med.
361,1058–1066(2009).
39.
Quinlan,A.
R.
&Hall,I.
M.
BEDTools:aexiblesuiteofutilitiesforcomparinggenomicfeatures.
Bioinformatics26,841–842(2010).
40.
Trapnell,C.
etal.
TranscriptassemblyandquanticationbyRNA-Seqrevealsunannotatedtranscriptsandisoformswitchingduringcelldifferentiation.
Nat.
Biotechnol.
28,511–515(2010).
41.
Schmidt,D.
etal.
ChIP-seq:usinghigh-throughputsequencingtodiscoverprotein-DNAinteractions.
Methods48,240–248(2009).
42.
Zhang,Y.
etal.
Model-basedanalysisofChIP-Seq(MACS).
GenomeBiol.
9,R137(2008).
43.
Stark,R.
&Brown,G.
D.
DiffBind:DifferentialBindingAnalysisofChIP-SeqPeakDataBioconductorhttp://bioconductor.
org/packages/release/bioc/html/DiffBind.
html(2011).
44.
Yu,G.
,Wang,L.
G.
&He,Q.
Y.
ChIPseeker:anR/BioconductorpackageforChIPpeakannotation,comparisonandvisualization.
Bioinformatics31,2382–2383(2015).
45.
Mohammed,H.
etal.
ProgesteronereceptormodulatesERalphaactioninbreastcancer.
Nature523,313–317(2015).
46.
Simigdala,N.
etal.
Cholesterolbiosynthesispathwayasanovelmechanismofresistancetoestrogendeprivationinestrogenreceptor-positivebreastcancer.
BreastCancerRes.
18,58(2016).
47.
Cox,J.
&Mann,M.
MaxQuantenableshighpeptideidenticationrates,individualizedp.
p.
b.
-rangemassaccuraciesandproteome-wideproteinquantication.
Nat.
Biotechnol.
26,1367–1372(2008).
AcknowledgementsThisstudywassupportedbyBreastCancerNow.
WealsoacknowledgeNHSfundingtoTheRoyalMarsdenHospitalsNIHRBiomedicalResearchCentreandthegenerousfundingfromTheArthurFoundationinsupportofJ.
N.
-B.
S.
A.
andL.
B.
arefundedbyCRUK(C37/A18784),A.
H.
issupportedbyaPhDstudentshipfromCRUKImperialCentre.
WealsowouldliketothanktheHigh-ThroughputGenomicsGroupattheWellcomeTrustCentreforHumanGenetics(fundedbyWellcomeTrustgrantreference090532/Z/09/Z)forthegenerationoftheSequencingdata.
Finally,wewouldliketothanktheTumorProlingUnitwithintheInstituteofCancerResearch.
AuthorcontributionsTheauthorscontributedtothisworkindifferentcapacitiesdescribedasfollows.
Concept:L.
-A.
M.
;Generationofresistantmodels:L.
-A.
M.
,S.
P.
,N.
P.
,M.
J.
S;Experi-mentalwork:R.
R.
,N.
S.
,S.
P.
,J.
N.
-B.
;DropletdigitalPCR:I.
G.
-M.
,C.
F.
,Z.
L.
,S.
O.
;Bioinformatics:N.
T.
;RNA-seq,ChIP-seq:E.
S.
,R.
R.
,N.
S.
,J.
C.
,W.
Z.
;Exomesequencing:R.
R.
,Q.
G.
,P.
G.
;Proteomics:N.
S.
,A.
B.
;CRISPR:T.
T.
,A.
H.
,L.
B.
,S.
A.
;Fluorescentinsituhybridization:V.
M.
,M.
H.
;Metabolomics:M.
-O.
T.
,G.
P.
;Xenograft:A.
T.
;ManuscriptARTICLENATURECOMMUNICATIONS|DOI:10.
1038/s41467-017-01864-y14NATURECOMMUNICATIONS|8:1865|DOI:10.
1038/s41467-017-01864-y|www.
nature.
com/naturecommunicationsanddisplayitempreparation:L.
-A.
M.
,R.
R.
,N.
S.
,S.
A.
,E.
S.
,M.
D.
Allauthorsreviewedthepreparedmanuscript.
AdditionalinformationSupplementaryInformationaccompaniesthispaperatdoi:10.
1038/s41467-017-01864-y.
Competinginterests:Theauthorsdeclarenocompetingnancialinterests.
Reprintsandpermissioninformationisavailableonlineathttp://npg.
nature.
com/reprintsandpermissions/Publisher'snote:SpringerNatureremainsneutralwithregardtojurisdictionalclaimsinpublishedmapsandinstitutionalafliations.
OpenAccessThisarticleislicensedunderaCreativeCommonsAttribution4.
0InternationalLicense,whichpermitsuse,sharing,adaptation,distributionandreproductioninanymediumorformat,aslongasyougiveappropriatecredittotheoriginalauthor(s)andthesource,providealinktotheCreativeCommonslicense,andindicateifchangesweremade.
Theimagesorotherthirdpartymaterialinthisarticleareincludedinthearticle'sCreativeCommonslicense,unlessindicatedotherwiseinacreditlinetothematerial.
Ifmaterialisnotincludedinthearticle'sCreativeCommonslicenseandyourintendeduseisnotpermittedbystatutoryregulationorexceedsthepermitteduse,youwillneedtoobtainpermissiondirectlyfromthecopyrightholder.
Toviewacopyofthislicense,visithttp://creativecommons.
org/licenses/by/4.
0/.
TheAuthor(s)2017NATURECOMMUNICATIONS|DOI:10.
1038/s41467-017-01864-yARTICLENATURECOMMUNICATIONS|8:1865|DOI:10.
1038/s41467-017-01864-y|www.
nature.
com/naturecommunications15