Combined inhibition of rho-associated protein kinase and EGFR suppresses the invasive phenotype in EGFR-dependent lung cancer cells.

LungCancer90(2015)167–174

ContentslistsavailableatScienceDirect

Lung

Cancer

j o ur na l h o me p a g e :w w w . e l s e v i e r . c o m / l o c a t e / l u n g c a n

Combined

inhibition

of

rho-associated

protein

kinase

and

EGFR

suppresses

the

invasive

phenotype

in

EGFR-dependent

lung

cancer

cells

Ijeoma

Adaku

Umelo

_,

_Olivier

_De

_Wever

_,

_Peter

_Kronenberger

_,

_Alfiah

_Noor

_,

Erik

Teugels

_,

_Gang

_Chen

_,

_Marc

_Bracke

_,

_Jacques

_De

_Grève

a_{LaboratoryofMolecularOncologyandDepartmentofMedicalOncology,OncologischCentrum,UniversitairZiekenhuisBrussel,Belgium}

b_{LaboratoryofExperimentalCancerResearchandDepartmentofRadiationTherapyandExperimentalCancerResearch,UniversitairZiekenhuisGhent,}

Belgium

c_{LaboratoryforBiotechnology,DepartmentofHealthcare,ErasmushogeschoolBrussel,Brussels,Belgium}

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received3June2015

Receivedinrevisedform11August2015

Accepted15August2015

Keywords:

Non-smallcelllungcancer

Invasion ROCK EGFR Y-27632 Afatinib

a

b

s

t

r

a

c

t

Introduction: Lung cancer remains the leading cause of cancer-relatedmortality worldwide, with metastaticdiseasefrequentlyaprominentfeatureatthetimeofdiagnosis.TheroleofNSCLC-derived EGFRmutationsincancercellproliferationandsurvivalhasbeenwidelyreported,butlittleisknown aboutthefunctionofthesemutationsininvasivegrowthandmetastasis.Inthisstudy,wesoughtto evaluatetheintrinsicinvasivepropertiesofNSCLCcellswithdifferingEGFRstatusandexaminepossible therapeutictargetsthatcanabrogateinvasivegrowth.

Materialsandmethods:Collagen-basedassaysand3Dcellcultureswereusedtoassessmorphological fea-tures,actincytoskeletondynamicsandtheinvasivecapacityofNSCLCcelllineswithdifferingEGFRstatus. TheroleoftheRhoA/ROCK/MYPT1andEGFR/HERpathwaysinNSCLC-relatedinvasionwasinvestigated bypharmacologicalinhibitionandRNAinterferencetechniques.

Results:WedemonstrateapositivecorrelationbetweenEGFRmutational/amplificationstatusand inva-sivecapacity.Knockdownofwild-typeandmutantEGFRleadstodepletionofactiveandtotalMYPT1 levels.CombinedpharmacologicalinhibitionorgeneticablationofROCK/EGFRsuppressesthehallmarks ofcancercellsandabrogatestheinvasivephenotypeinEGFR-dependentNSCLCcells.

Conclusions:TheseobservationssuggestthatcombinedtargetingoftheROCKandEGFR/HERpathways maybeapotentialtherapeuticapproachinlimitinginvasivegrowthinNSCLC.

©2015ElsevierIrelandLtd.Allrightsreserved.

1. Introduction

Lungcanceris theleading causeof cancer-related mortality worldwide,withlocalinvasionanddistalmetastasesasa promi-nentfeatureattimeofpresentation[1].Theinvasiveprocessisthe resultofacomplexinteractionbetweencancercellsandthehost tissueenvironment,actincytoskeletondynamics,degradationof matrixbarriersandtheformationofmembraneprotrusionsand

∗ Corresponding authorat: LaboratoryofMedical andMolecular Oncology,

DepartmentofMedicalOncology,OncologyCenter,UZBrussel,VrijeUniversiteit

Brussel(VUB),Laarbeeklaan101,1090Brussels,Belgium.Fax:+3224776210.

E-mailaddresses:iumelo@vub.ac.be(I.A.Umelo),Olivier.DeWever@UGent.be

(O.D.Wever),Peter.kronenberger@vub.ac.be(P.Kronenberger),alfinoor@vub.ac.be

(A.Noor),Erik.Teugels@uzbrussel.be(E.Teugels),gangchen@vub.ac.be(G.Chen),

Marc1.Bracke@UGent.be(M.Bracke),Jacques.Degreve@uzbrussel.be(J.D.Grève).

othermigratoryfeaturesnecessaryforinvasivegrowth[2].Inthe sameperspective,advancesincancerbiologyhavealsoshownthat genomicalterationsofthehumanepidermalgrowthfactorreceptor (EGFR/HER)playacriticalroleincancerprogressionand tumori-genesisinabroadspectrumofhumancancersincludingnon-small celllungcancer(NSCLC)[3].TheroleoftheseNSCLC-derivedEGFR mutationsincancercellproliferation,andsurvivalhasbeenwidely studied[4–6];howeverstudiesdemonstratingtheirspecificrolein thecontextoftheinvasivephenotypehavesofarbeenlimited.

The Rho-associated coiled-coil forming protein (ROCK) is a known regulatorof actin cytoskeletal organization, involved in Rho-inducedformationofactinstressfibersandfocaladhesions [7].Theseactomyosinstructuresplayanimportantroleincellular contractilityandasaconsequencealsoincelladhesion,migration and morphogenesis.Thereisemerging evidenceindicating that thecloselyrelatedROCK1andROCK2kinasessupporttheinvasive

http://dx.doi.org/10.1016/j.lungcan.2015.08.008

168 I.A.Umeloetal./LungCancer90(2015)167–174

Fig.1.CorrelationbetweenEGFRmutationalstatusandinvasivecapacityincellularmodelsofNSCLC.(a)PhasecontrastimagesandinvasiveindexquantificationofNSCLC

cellsculturedonatypeIcollagenmatrixfor24h.Arrowsindicateinvasiveextensions.(b)Confocalimagesandfactorshapequantification(boxplot)off-actin(red)/DAPI

(purple)stainedcellsculturedfor24honatypeIcollagenmatrix.(c)TransverselightsectionsofH&EstainedcellsculturedonatypeIcollagenmatrixfor14days.

Arrowsindicatecellsinvadingtheunderlyingmatrix.(d)NSCLCcellsweregrownfor72hinathree-dimensionallaminin-richmicroenvironmentandimagedunderalight

microscope.Scalebars,50␮m.

andmetastaticphenotypeinavarietyofcancersincludinglung cancer[8].However,thebiologicalrole oftheROCKproteinsin NSCLCinvasionandmetastasisisunclear.Here,weinvestigated theintrinsic invasivecapacity ofNSCLCcelllineswithdiffering EGFRmutational/amplificationstatusand examinetheeffect of combinedEGFR/ROCKinhibition ontheinvasionand survivalof NSCLCcellswithdifferentialinvasivecapacities.

2. Results

2.1. CorrelationbetweenEGFRmutationalstatusandinvasive capacity

TodetermineapossiblecorrelationbetweenEGFRmutational statusandinvasion,wefirstcomparedtheintrinsicinvasive capac-ityof three NSCLCcell linesharboring differentialEGFR status. Aftera24hcultureonanativetypeIcollagensubstrate,H1975 NSCLCcells carryingboth aprimaryEGFRL858R_activating

muta-tionand a secondary EGFRT790M _{tyrosine kinaseinhibitor (TKI)}

resistance-conferringmutation[8],displayedamorespread cel-lularmorphologywiththeformationofcellularextensionsinside thematrix(invasiveindex:75±1.7%).Incontrast,H1650(EGFR

E746–A750_{)andH358(EGFR}wt_{)NSCLCcellsattachedasrounded}

cells with a significantly lower invasive index (invasiveindex: 8.3_±0.7%and 2.3_±0.3%respectively;Fig.1a).Inaddition,PC-9 NSCLCcellsbearingthesameexon19deletionmutationasH1650 cells(EGFRE746–A750),displayedacomparableinvasiveindexof

12.2±0.9%; whileH1781 cells harboring wild-type EGFRand a HER2insertion mutationdidnotforminvasiveextensionsona collagenmatrix(SupplementalFig.1a).Interestingly,A549NSCLC cellsharboringawild-typeEGFRamplificationwerecategorizedas highlyinvasive,withaninvasiveindexof72±2.5%(Supplemental Fig.1b).

Toexaminedifferencesincorticalcytoskeletonandcellsurface morphology,thepanelofNSCLCcelllineswerestainedforf-actin (bythetoxinphalloidin)aftera24hcultureontype Icollagen. AsshowninFig.1b,H1975cellsdisplayedamorepolarized mor-photypewithmultiplelongprotrusionswhereasH1650andH358 cellsassumedaroundedmorphologywithcorticalF-actinandthe appearanceofmembraneblebs.AsshowninSupplementalFig.1c, A549cellsdisplayedasimilarmorphologicalphenotypeasH1975 cells,withadisruptedandpolarizedcorticalf-actincytoskeleton.

WefurtherevaluatedtheintrinsicinvasivecapacityofH1975 NSCLCcells,byexaminingtheirinvasivedepthinatypeI colla-genmatrix.NSCLCcellswereculturedonacollagenfor14days, wherebytransversesectionswereanalyzedforthepresenceofcells invadingthecollagensubstrate.OurresultsshowH1975cells infil-tratingintothecollagenmatrixwhereasH1650cellsareconfined toasinglecellularlayerandH358cellsformcoloniesatthetopof thecollagen(Fig.1c).

To establish a cell culture that more closely reflects an in vivolungcancermicroenvironment,wecharacterizedtheNSCLC celllinesinlaminin-richthree-dimensionalcultures.Asshownin Fig.1d,H1975cellsappearedascoloniesshowingextensive

con-I.A.Umeloetal./LungCancer90(2015)167–174 169

Fig.2. Theroleoftherho-associatedproteinkinase(ROCK)anditseffectorsinhighlyinvasiveandEGFR-dependentNSCLCcells.(aandb)NSCLCwereculturedfor3days

andcelllysateswerecollectedforimmunoblotanalysisusingtheindicatedantibodies.(candd)H1975cellsweretransientlytransfectedwiththeindicatedsiRNAfor72h

andassessedforgrowthbyMTSanalysis.Totalcelllysateswerecollectedforimmunoblotanalysisusingtheindicatedantibodies.

nectingextensions,suggestiveofdirectadhesionbetweenthecells andtheextracellularmatrix(ECM).Incontrast,H358andH1650 cells appeared ascolonies thatwere situated apartin theECM (Fig.1d).

2.2. Roleofrho-associatedcoiledcoilformingprotein(ROCK)and itseffectorsinhighlyinvasiveEGFRmutantNSCLCcells

OverexpressionoftheROCKproteinshasbeenimplicatedinthe invasiveandmetastaticpotentialofvarioustumortypes[9–11]. BaselineproteinexpressionofROCK1andROCK2washigherin H1975cellscomparedtoH1650andH358NSCLCcells(Fig.2a), indicatingthat theROCKproteinmaybea crucialfactor inthe invasivephenotypeofH1975cells.EGFRactivationstatuswasalso greatlyenhancedinH1975cellscomparedtotheothercelllines; possiblyduetothepresenceoftheEGFRT790M_{mutationinthiscell}

line(Fig.2b).TheEGFRT790M_{mutationisassociatedwithacquired}

resistancetoEGFRTKItherapyandhasbeenfoundtoincreaseEGFR activityaswellasoncogenicityinNSCLC[12,13].

Accumulatinglines ofevidencehave demonstratedthat AKT activationcorrelateswithenhancedtumorinvasion[14,15].AKT phosphorylationwasalsosignificantlyhigherinH1975cells com-paredtoH358cells,butH1650cells exhibitedevenhigherAKT phosphorylation compared to the other two tested cell lines (Fig.2b).ThisobservedhighAKTsignalinginH1650cellsmaybe theresultofthecombinedeffectoftheEGFRmutationandaPTEN nullstatus,whichhasbeenimplicatedinderegulatedAKTactivity [16].

ToelucidatetheroleoftheEGFRT790M_{mutationintheinvasive}

propertiesofH1975cells,weperformedRNAinterference exper-iments.EGFRT790M2600 _{-specificmediated-knockdown effecteda}

∼25%reductionincellgrowth72hpost-transfectionin concord-ance withour previous findings [17], while EGFRWT2600 _siRNA

reducedcellgrowthby∼15%atthesametimepoint(Fig.2c).By 72hpost-transfection,weobservedamoredistinctreductionin bothphosphorylatedandtotalEGFRlevelsaswellasactiveAKT afterEGFRT790M2600_{siRNA-mediatedknockdowncomparedtoboth}

EGFRWT2600 _{and scrambled}T790M2600_{treatment(Fig.2d).Further}

analysisontheeffectofwild-typeand mutantEGFR-mediated knockdownonROCKexpressiondidnotrevealareductionintheir proteinlevels(datanotshown).Interestingly,however,wild-type andmutant EGFRsiRNA treatmentclearlyreduced thelevelsof myosinphosphastasetargetsubunit1(MYPT1);adirecttargetof ROCK(Fig.2d).ThesedatathussuggestthatEGFRinhibitioncould playaroleinsuppressingtheinvasivephenotypeofH1975cells.

2.3. EffectofROCKandEGFRinhibitiononthegrowthofNSCLCs withdifferingEGFRmutationalstatus

AsbothROCK1andROCK2arehighlyexpressedinH1975cells (seeFig.2a),wesoughttodeterminetheeffectsongrowthof Y-27632,aninvestigationalinhibitoroftheROCKfamilyofkinases thatcompeteswithATPforbindingtoitscatalyticcleft[18].In addition,wealsoinvestigatedtheinhibitoryeffectsofafatinib,an EGFR/HERfamilyinhibitorthatcovalentlyandirreversiblybinds totheATPbindingcleftofthetyrosinekinaseandalsohassome activityinT790Mcelllines[19].Dose-dependentanalysisofthe growthinhibitoryeffectof Y-27632in thepanelofNSCLCcells (H358,H1650,H1975andA549)demonstratesthatY-27632was unabletoachieveanIC50valueevenataconcentrationof10␮M

(SupplementalFig.2a).Ontheotherhand,weobservedastrikingto modestgrowthinhibitoryeffectwithafatinibinthemajorityofthe

170 I.A.Umeloetal./LungCancer90(2015)167–174

Fig.3.EffectofcombinedinhibitionofROCKandEGFR/HERinNSCLCcellswithdifferinginvasivecapacity.(a)NSCLCcellsweretreatedwithY-27632(5␮M),afatinib

(50nM)ortheircombinationfor5days.(b)NSCLCcellsweretransientlytransfectedwith100nMoftheindicatedsiRNAortheircombinationsfor72h.Growthinhibition

wasassessedbyMTSanalysis.

celllinestested (SupplementalFig.2b).Treatmentwithafatinib didnotsignificantlyreducethegrowthofA549cellsatclinically relevantdoses(SupplementalFig.1b).

ToinvestigatethecombinedeffectofEGFRandROCKkinase inhibitionwetreatedtherespectiveNSCLCcelllines(H358,H1650 andH1975)for5dayswithY-27632andafatinibaseither sin-gle agents or combined. Our results display modest effects of theY-27632/afatinib combination onthe growthof both H358 andH1975 cells when compared tosingleinhibitor treatments (Fig.3a),whileweobservednoadditionaleffectwhenY-27632 wasaddedtoafatinibinH1650 cells(Fig.3a).While combined treatmentwithY-27632/afatinibdidnotreducethegrowthofA549 cellscomparedtovehiclecontrol,fasudil(aclinicallyactiveROCK inhibitor)combinedwithafatinibachievedmodesteffectsinthis cellline(SupplementalFig.1d).Interestingly,fasudil monother-apyalsopotentlyreducedtheabilityofA549cellstoformcolonies insoftagarcomparedtocombinedtreatment(SupplementalFig. 1e).

We subsequently performed siRNA transfection experi-ments to further verify our pharmacological findings. We show that the addition of ROCK siRNA treatment potently reduces cell growth in the H1975 and H358 cells 72h post-transfection when compared to combined EGFR/HER2 or single EGFR or HER2 siRNA treatments (Fig. 3b). Consistent with the pharmacological experiments, no significant addi-tional effect was obtained when ROCK-specific siRNA was added to the EGFR/HER2 siRNA combination in H1650 cells (Fig.3b)

2.4. EffectofROCKandEGFRinhibitionontheinvasive phenotypeofEGFRmutantNSCLCcells

TofurtherassesstheefficacyoftheY-27632/afatinib combina-tion,weevaluatedtheireffectinabrogatingtheinvasivephenotype inH1975cells.OurresultsshowthatY-27632combinedwith afa-tinibwassuperiortoeitheragentaloneinthereductionofboth invasiveindexandfactorshapeaftera24htreatmentonatype Icollagensubstrate(Fig.4aand b).Moreover,this combination significantlyreducedtheamountofH1975cellsinfiltratinginto thecollagenaftera14 dayinhibitor treatment(Fig.4c). Impor-tantly,wealsoobservedamarkeddecreaseinthenumberofH1975 cellsinvadingthecollagensubstratewithsingleY-27632treatment comparedtotheeffectsseenwithsingletreatmentwithafatinib (Fig.4c).Theeffectof Y-27632onH358non-invasive cells was lessstrikingafterasimilar14daytreatmentonacollagen sub-strate.Afatinibtreatmentcoincidedwithrobustsuppressioninthe amountofcoloniesformedbyH358cells,whiletreatmentwith Y-27632didnotyieldsignificanteffects(SupplementalFig.3a).

CombinedtreatmentofY-27632andafatinibreduced phospho-rylatedlevelsofROCKproteinactivatorsandeffectors(RhoAand MYPT1)andpotentlysuppressedAKTphosphorylationinH1975 cells(Fig.4dandSupplementalFig.3b).Inaddition,whileafatinib treatmentreducedactivelevelsofEGFRandHER2inH1975cells, activeERKlevelswerecooperativelyreducedwithcombination treatment(SupplementalFig.3c).BothY-27632andafatinibhad aneffectoncellsurvival,aftera5daytreatment,buttheeffect wasnotenhancedwhenY-27632wasaddedtoafatinib(Fig.4e). To support our pharmacological data, we examined the effect EGFR/HER2/ROCKsiRNAtreatmentonactincytoskeleton

morphol-I.A.Umeloetal./LungCancer90(2015)167–174 171

Fig.4.ROCKinhibitionsynergiseswithEGFRtosuppresssurvivalandinvasioninhighlyinvasiveNSCLCcells.(a)Phasecontrastimagesandinvasiveindexquantification

ofH1975cellstreatedwithY-27632(5␮M),afatinib(50nM)ortheircombinationfor24h.(b)Confocalimagesandfactorshapequantificationoff-actin(red)/DAPI(blue)

stainedcellstreatedwiththeindicatedinhibitorsortheircombinationfor24h.(c)TranverselightsectionsofH1975cellsculturedonatypeIcollagenmatrixfor14days

andtreatedwiththeindicatedinhibitorsortheircombination.Arrowsindicatecellsinvadingtheunderlyingmatrix.(dande)H1975cellsweretreatedwiththeindicated

inhibitorsortheircombinationfor5days.Totalcelllysateswerecollectedforimmunoblotanalysiswiththeindicatedantibodies(valuesrepresentnormalizedratiosof

phospho/totalproteinlevelsrelativetovehicle)andcellularapoptosiswasquantifiedbyhoechst33342(blue)andpropidiumiodide(red)doublefluorescentchromatin

staining.Bars,50␮m.(Forinterpretationofthereferencestocolorinthisfigurelegend,thereaderisreferredtothewebversionofthisarticle.).

ogyofH1975cells.H1975cellsremainedwithapolarizedF-actin morphologyaftersiRNA-mediateddepletionofEGFR (Supplemen-talFig.3d).However,depletionofHER2promotedamorerounded morphologywithcombineddepletionofEGFR,HER2andROCK fur-therpromotinganorganizedF-actincytoskeleton(Supplemental Fig.3d).

3. Discussion

Inthisstudyweevaluatedtheintrinsicinvasivepropertiesof NSCLCcellswithdifferingEGFRstatus,andexaminedthebiological roleoftheRho/ROCKpathwayinregulatingthefunctionofhighly invasivecells.Ourdatademonstrates thatbothEGFRandROCK playaroleinthegrowthandinvasionofNSCLCcellswithdifferent EGFRgenotypes.ThecombinedeffectsofROCKandEGFR inhibi-tiononthegrowthandinvasivecapacityofhighlyinvasiveH1975 cells(EGFRL858R+T790M_{)suggeststhatdualtargetingoftheROCKand}

EGFR/HERproteinsmayofferapotentialtherapeuticbenefit. OurfindingsrevealthatNSCLCcellscarryingawild-typeEGFR statusdonotpromoteinvasionandpointtoapositivecorrelation betweenaninvasivephenotypeandaconstitutivelyactivemutant EGFRorawild-typeEGFRamplificationstatus.Arecentstudyby Talasilaetal.,supporttheseobservations,whereamplificationor activationofwild-typeEGFRwereconsideredtobeimportant fac-torsintheinvasivephenotypeofmousemodelsofglioblastoma [20].

OurfindingsalsoindicatethatinhighlyinvasiveNSCLCs,the ROCKpathwayisactivatedandthatmutantEGFRplaysarolein this(Fig.5).TheactivationlevelofthehighlyexpressedROCK pro-teininthesecellscausesastrikingresponsetobothEGFR/HER2and RhoA/ROCK/MYPT1inhibitionresultinginthesuppressionofboth survival andinvasion (Fig.5).Inaddition,RhoA/ROCK signaling plays a pivotalrole in actin cytoskeleton dynamics and tumor cellmotility[21] andourfindingsalsohighlightthatcombined EGFR/ROCKinhibitioncorrespondswithchangesincell morphol-ogyandthereorganizationoffilamentousactininhighlyinvasive andEGFR-dependentNSCLCcells.AKTactivationcorrelateswith enhanced tumor invasion [14,15], and based onthe presented results dual ROCK/EGFR treatment antagonizes AKT signaling; whichsimilarlyleadstotherobustsuppressionofMAPK/ERK path-wayactivity.It must benoted,however, thatwhile theROCK1 and ROCK2 proteinsshare significantsequence homology, they havebeendescribedtoexertdifferentfunctionsinthecell[22,23]. ROCK1is reportedtobeinvolvedin actincytoskeleton destabi-lizationwhileROCK2hasbeenshowntobeanessentialfactorin stabilizingtheactincytoskeleton.Theobservedsuppressiveeffects oninvasionmaythusresultfrominhibitionoftheROCK1protein, aspharmacologicalinhibitorsofROCKequallyinhibittheactivity ofbothROCK1andROCK2[24].

Severalpreclinicalstudieshavedemonstrated thatinhibition of Rho/ROCKin cells bearing oncogenicforms of KIT,FLT3and BCR-ABLTKreducesactivelevelsofMYPT1andsuppressestheir constitutivegrowth[25].ApublicationbyBurthemandcolleagues

172 I.A.Umeloetal./LungCancer90(2015)167–174

Fig.5.ProposedschematicmodelforROCK/EGFRinhibitioninNSCLCcellswithanEGFR-dependentandhighlyinvasivephenotype.(a)GTP-boundRhoAisactivatedby

theEGFRreceptor(wild-typeormutant)whichinturnactivatesthehighlyexpressedROCKprotein.Thiscouplingleadstothephosphorylationofdownstreamspecific

targetproteinssuchasMYPT1whichisincytoskeletonregulationandinvasivegrowth.EGFRalsoactivatesPI3KwhichphosphorylatesPIP2toPIP3,therebyactivatingAKT

andtriggeringthetranscriptionofcellsurvivalfactors.PTENisanotherdownstreamtargetofROCKwhichnegativelyregulatesAKTbydephosphorylatingPIP3.(b)afatinib

reducesEGFRandROCKactivationtherebysuppressinginvasionandsurvival,(c)Y-27632reducesROCKactivationmainlysuppressinginvasion,and(d)theircombination

cooperativelysuppressesbothsurvivalandinvasion.

shows that ROCK inhibitors suchas Y-27632 and fasudil com-binedwiththeTKIimatinibsynergisticallyinhibitstheactivityof humanleukemiccells[26].TheseresultsshowthatROCKmaybe apotentialadditional therapeutictargetin impairingthe onco-genicactivityofTK-drivenmalignanciesandvalidatetheuseof theseinhibitorsaseithersingleagentsorcombinedwithotherTKIs asapossibletherapeuticstrategy.SinceTKIssuchasafatiniband imatinibarecurrentlyindicatedforthetreatmentofsomecancers andROCKinhibitorssuchasfasudilhavebeeneffectivelyusedfor severalyearsinthetreatmentofvariousdiseases[27],itmaybe promisingtotestthesecombinationsinfutureclinicaltrials[28].

Inconclusion,ourfindingsdemonstratethattheROCKfamily of protein kinases is important for highly invasive NSCLC

can-cers.WhereasthefunctionalroleofthegeneticallyalteredEGFR receptorinthecontextofNSCLCpathogenicityhasbeenwidely described[29,30],ourfindingsfurtherdemonstratethatmutational activationofEGFRmayplayaroleintheinvasivephenotype.The combinedtargetingtheRho/ROCKandEGFR/HERpathwaysshould befurtherexplored.

4. Materialsandmethods

4.1. Celllinesandreagents

Y-27632(Selleckchem)andafatinib(BoehringerIngelheim)were of highest grade available. The human NSCLC cell lines H358,

I.A.Umeloetal./LungCancer90(2015)167–174 173

H1650, H1975 were obtained from the ATCC and cultured as described[6].TheA4549humanNSCLCcelllinewasobtainedfrom Sigmaandculturedaccordingtothesuppliersrecommendations. AllcelllineswereauthenticatedbySTR-analysis.

4.2. Invasionassays

The Collagen invasion assay was performed as previously described [7]. Collagen matrices were imaged under a phase-contrastlightmicroscopeandtheinvasiveindexwassubsequently determinedbymanualcountingofthenumberofinvadingcells andnon-invadingcellspresentin10–15randomlyselectedfields. Cellmorphologywasstudiedbydeterminationoffactorshape of minimal 25 single cells. Collagen matrices were fixedin 3% paraformaldehydeand stainedwithAlexaFluor594conjugated phallodinwithaDAPInucleicounterstain[31].Imagingwas sub-sequentlyperformedbyconfocalmicroscopy.

Forinvasivedepth,collagenmatriceswerefixedwithbuffered formalin, embeddedin paraffin and 6-_␮m sectionswere taken every 250␮m over a total distance of 1500␮m. All sections werestainedwithhematoxylinandeosin(H&E)andsubsequently imagedunderalightmicroscope.Onerepresentativefieldfromone sectionoutof6isshown.

4.3. Three-dimensionalcellcultures

TheNSCLCcells(1×104_{)wereseededinaMatrigel(Corning)}

suspensioninclear-bottomed96wellplates.TheMatrigel suspen-sionwaskeptat37◦Cuntilpolymerizationandatoplayerofgrowth mediumwasthenadded.CellsweremaintainedinMatrigelfor3 daysandimagingwasperformedwithastandardlight-microscope toexaminemorphologicaldifferences.

4.4. Westernblotanalysisandantibodies

Cells were lysed in a Tris-buffer containing a protease and phosphatase inhibitorcocktail (Sigma).Lysates wereclearedby centrifugationandproteinconcentrationwasdeterminedbythe Bradfordprotein assaykit(Bio-Rad). Equivalentamountof pro-tein were loaded on a resolving acrylamide gel and blotted ona polyvinylidenefluoride membrane(PVDF).The membrane wasthensubjectedtoanimmunodetectionprocedureusingthe indicated antibodies. HRP-conjugated secondary antibodies (GE Healthcare)andachemiluminescentdetectionkit(Advansta)were usedtodetecttheindicatedproteins.

4.5. Growthassays

CellgrowthwasassessedbythecolorimetrictetrazoliumMTS assay (Promega). Cells were seeded in clear-bottomed 96-well plates at a density of 2×103 _{cells/wellin triplicates. For drug}

inhibitionstudies,cellsweretreatedafter24hwithvarious con-centrationsoftheindicateddrugsortheircombinations.Afterthe indicatedtimeperiods,thenumbersofviablecellswereanalyzed usinganELISAreader(Labsystems).

4.6. RNAinterference

Cells were seeded into 24-well or 96-well plates at a den-sityofapproximately1×105_{cells/wellor5×10}3_{cells/well.24h}

later, cells were transfected with siRNA against EGFR WT2600_,

EGFRT790M2600_{and scrambled EGFR}T790M2600 _{[17]; EGFR,HER2,}

GenomeNon-TargetingsiRNA(ThermoScientific);andROCK (Qia-gen) using Lipofectamine 2000 (Invitrogen) according to the manufacturer’sinstructions.

4.7. Hoechst3342/propidiumIodidestaining

Theeffectsoftheinhibitorsonapoptosisandnuclear morphol-ogyofH1975cellswasassessedbyHoechst33342andpropidium iodide(PI)doublefluorescentchromatinstainingasdescribed[32]. 4.8. Statisticalanalysis

Resultsarerepresentativeofthreeindependentexperiments. Allgraphical data ispresented as themean±standard errorof mean(SEM).TheChi-Squaretestwasutilizedforgroup compar-isonofinvasiveindexwhiletheMannWhitneytestwasutilized whencomparingfactorshape.Statisticalsignificanceisreported asfollows:*_,P<0.05,**_,P<0.01and***_,P<0.001.

Conflictofintereststatement

Theauthorsdeclarenoconflictofinterest.

Acknowledgments

ThisworkwassupportedbytheNationalCancerPlan(Belgium), theStichtingTegenKanker(Belgium)andaVrijeUniversiteit Brus-selPhDfellowship.

AppendixA. Supplementarydata

Supplementarydataassociatedwiththisarticlecanbefound,in theonlineversion,athttp://dx.doi.org/10.1016/j.lungcan.2015.08. 008.

References

[1]C.S.DelaCruz,L.T.Tanoue,R.A.Matthay,Lungcancer:epidemiology,etiology, andprevention,Clin.ChestMed.32(4)(2011)605–644.

[2]M.Yilmaz,G.Christofori,EMT,thecytoskeleton,andcancercellinvasion, CancerMetastasisRev.28(1–2)(2009)15–33.

[3]S.V.Sharma,J.Settleman,ErbBsinlungcancer,Exp.CellRes.315(4)(2009) 557–571.

[4]J.Jiang,etal.,Epidermalgrowthfactor-independenttransformationofBa/F3 cellswithcancer-derivedepidermalgrowthfactorreceptormutantsinduces gefitinib-sensitivecellcycleprogression,CancerRes.65(19)(2005) 8968–8974.

[5]R.Sordella,etal.,Gefitinib-sensitizingEGFRmutationsinlungcanceractivate anti-apoptoticpathways,Science305(5687)(2004)1163–1167.

[6]R.Mulloy,etal.,Epidermalgrowthfactorreceptormutantsfromhumanlung cancersexhibitenhancedcatalyticactivityandincreasedsensitivityto gefitinib,CancerRes.67(5)(2007)2325–2330.

[7]A.L.Bishop,A.Hall,RhoGTPasesandtheireffectorproteins,Biochem.J.2(Pt. (348))(2000)241–255.

[8]K.Riento,A.J.Ridley,Rocks:multifunctionalkinasesincellbehaviour,Nat. Rev.Mol.CellBiol.4(6)(2003)446–456.

[9]T.Kamai,etal.,SignificantassociationofRho/ROCKpathwaywithinvasion andmetastasisofbladdercancer,Clin.CancerRes.9(7)(2003)2632–2641.

[10]J.Lane,etal.,TheexpressionandprognosticvalueofROCKIandROCKIIand theirroleinhumanbreastcancer,Int.J.Oncol.33(3)(2008)585–593.

[11]D.Rosel,etal.,Up-regulationofRho/ROCKsignalinginsarcomacellsdrives invasionandincreasedgenerationofprotrusiveforces,Mol.CancerRes.6(9) (2008)1410–1420.

[12]C.H.Yun,etal.,TheT790MmutationinEGFRkinasecausesdrugresistanceby increasingtheaffinityforATP,Proc.Natl.Acad.Sci.U.S.A.105(6)(2008) 2070–2075.

[13]W.Pao,etal.,Acquiredresistanceoflungadenocarcinomastogefitinibor erlotinibisassociatedwithasecondmutationintheEGFRkinasedomain, PLoSMed.2(3)(2005)e73.

[14]V.Vasko,etal.,Aktactivationandlocalisationcorrelatewithtumourinvasion andoncogeneexpressioninthyroidcancer,J.Med.Genet.41(3)(2004) 161–170.

[15]A.Dudley,etal.,DRRregulatesAKTactivationtodrivebraincancerinvasion, Oncogene33(41)(2014)4952–4960.

[16]M.L.Sos,etal.,PTENlosscontributestoerlotinibresistanceinEGFR-mutant lungcancerbyactivationofAktandEGFR,CancerRes.69(8)(2009) 3256–3261.

[17]G.Chen,etal.,EffectofsiRNAstargetingtheEGFRT790Mmutationina non-smallcelllungcancercelllineresistanttoEGFRtyrosinekinase

174 I.A.Umeloetal./LungCancer90(2015)167–174

inhibitorsandcombinationwithvariousagents,Biochem.Biophys.Res. Commun.431(3)(2013)623–629.

[18]T.Ishizaki,etal.,PharmacologicalpropertiesofY-27632:aspecificinhibitor ofrho-associatedkinases,Mol.Pharmacol.57(5)(2000)976–983.

[19]F.Solca,etal.,Targetbindingpropertiesandcellularactivityofafatinib(BIBW 2992),anirreversibleErbBfamilyblocker,J.Pharmacol.Exp.Ther.343(2) (2012)342–350.

[20]K.M.Talasila,etal.,EGFRwild-typeamplificationandactivationpromote invasionanddevelopmentofglioblastomaindependentofangiogenesis,Acta Neuropathol.125(5)(2013)683–698.

[21]N.Rath,M.F.Olson,Rho-associatedkinasesintumorigenesis:re-considering ROCKinhibitionforcancertherapy,EMBORep.13(10)(2012)900–908.

[22]J.Shi,etal.,DistinctrolesforROCK1andROCK2intheregulationofcell detachment,CellDeathDis.4(2013)e483.

[23]F.E.Lock,etal.,Differentialregulationofadhesioncomplexturnoverby ROCK1andROCK2,PLoSOne7(2)(2012)e31423.

[24]S.Vemula,etal.,ROCK1functionsasasuppressorofinflammatorycell migrationbyregulatingPTENphosphorylationandstability,Blood115(9) (2010)1785–1796.

[25]R.S.Mali,etal.,Rhokinaseregulatesthesurvivalandtransformationofcells bearingoncogenicformsofKIT:FLT3,andBCR-ABL,CancerCell20(3)(2011) 357–369.

[26]J.Burthem,etal.,Therho-kinaseinhibitorsY-27632andfasudilact synergisticallywithimatinibtoinhibittheexpansionofexvivoCD34(+)CML progenitorcells,Leukemia21(8)(2007)1708–1714.

[27]M.F.Olson,ApplicationsforROCKkinaseinhibiton,Curr.Opin.CellBiol.20(2) (2008)242–248.

[28]N.Rath,M.F.Olson,Rho-associatedkinasesintumorigenesis:re-considering ROCKinhibitionforcancertherapy,EMBORep.13(10)(2012)900–908.

[29]T.Mitsudomi,Y.Yatabe,Epidermalgrowthfactorreceptorinrelationto tumordevelopment:EGFRgeneandcancer,FEBSJ.277(2)(2010)301–308.

[30]S.V.Sharma,etal.,Epidermalgrowthfactorreceptormutationsinlung cancer,Nat.Rev.Cancer7(3)(2007)169–181.

[31]O.DeWever,etal.,Modelingandquantificationofcancercellinvasion throughcollagentypeImatrices,Int.J.Dev.Biol.54(5)(2010)887–896.

[32]G.Chen,etal.,Targetingtheepidermalgrowthfactorreceptorinnon-small celllungcancercells:theeffectofcombiningRNAinterferencewithtyrosine kinaseinhibitorsorcetuximab,BMCMed.10(2012)28.