MitochondrialAktRegulationofHypoxicTumorReprogramming
Highlights
d
ApoolofactiveAktisrecruitedtotumormitochondriaduringhypoxia
d
MitochondrialAktphosphorylatesPDK-1inhypoxiaonaT346site
d
Akt-PDK1activationmaintainstumorcellproliferationinhypoxia
d
PDK1phosphorylationbyAktisanegativeprognosticfactoringliomas
Chaeetal.,2016,CancerCell30,257–272August8,2016ª2016ElsevierInc.
http://dx.doi.org/10.1016/j.ccell.2016.07.004
Authors
YoungChanChae,ValentinaVaira,M.CeciliaCaino,...,LuciaR.Languino,DavidW.Speicher,DarioC.Altieri
Correspondence
daltieri@wistar.org
InBrief
Chaeetal.identifyapoolof
phosphorylatedAktthattranslocatestothemitochondriaduringhypoxia.
MitochondrialAktphosphorylatesPDK1,whichsupportscellsurvivaland
proliferationduringhypoxia,andelevatedAkt-dependentphosphorylationofPDK1correlateswithreducedsurvivalofgliomapatients.
AccessionNumbers
PXD004024
CancerCell
Article
MitochondrialAktRegulation
ofHypoxicTumorReprogramming
YoungChanChae,1ValentinaVaira,2,3,4M.CeciliaCaino,1Hsin-YaoTang,4JaeHoSeo,1AndrewV.Kossenkov,5LuisaOttobrini,4,6CristinaMartelli,4GiovanniLucignani,7,8IreneBertolini,3,4MarcoLocatelli,9KellyG.Bryant,1JagadishC.Ghosh,1SofiaLisanti,1BonsuKu,10SilvanoBosari,3,4LuciaR.Languino,11DavidW.Speicher,5,12andDarioC.Altieri1,*
1ProstateCancerDiscoveryandDevelopmentProgram,TumorMicroenvironmentandMetastasisProgram,TheWistarInstitute,3601Spruce
Street,Philadelphia,PA19104,USA
2IstitutoNazionaleGeneticaMolecolare‘‘RomeoandEnricaInvernizzi’’,Milan20122,Italy
3DivisionofPathology,FondazioneIRCCSCa’GrandaOspedaleMaggiorePoliclinico,Milan20122,Italy4DepartmentofPathophysiologyandTransplantation,UniversityofMilan,Milan20122,Italy
5CenterforSystemsandComputationalBiology,TheWistarInstitute,Philadelphia,PA19104,USA
6InstituteforMolecularBioimagingandPhysiology(IBFM),NationalResearchCouncil(CNR),Milan20090,Italy7DepartmentofHealthSciences,UniversityofMilan,Milan20142,Italy
8DepartmentofDiagnosticServices,UnitofNuclearMedicine,SanPaoloHospital,Milan20142,Italy
9DivisionofNeurosurgery,FondazioneIRCCSCa’GrandaOspedaleMaggiorePoliclinico,Milan20122,Italy
10FunctionalGenomicsResearchCenter,KoreaResearchInstituteofBioscienceandBiotechnology,125Gwahak-ro,Yuseong-gu,Daejeon305-806,RepublicofKorea
11DepartmentofCancerBiology,KimmelCancerCenter,ThomasJeffersonUniversity,Philadelphia,PA19107,USA12MolecularandCellularOncogenesisProgram,TheWistarInstitute,Philadelphia,PA19104,USA*Correspondence:daltieri@wistar.org
http://dx.doi.org/10.1016/j.ccell.2016.07.004
SUMMARY
Hypoxiaisauniversaldriverofaggressivetumorbehavior,buttheunderlyingmechanismsarenotcompletelyunderstood.Usingaphosphoproteomicsscreen,wenowshowthatactiveAktaccumulatesinthemitochondriaduringhypoxiaandphosphorylatespyruvatedehydrogenasekinase1(PDK1)onThr346toinactivatethepyruvatedehydrogenasecomplex.Inturn,thispathwayswitchestumormetabolismtowardglycolysis,antagonizesapoptosisandautophagy,dampensoxidativestress,andmaintainstumorcellpro-liferationinthefaceofseverehypoxia.MitochondrialAkt-PDK1signalingcorrelateswithunfavorableprog-nosticmarkersandshortersurvivalingliomapatientsandmayprovidean‘‘actionable’’therapeutictargetincancer.
INTRODUCTION
Hypoxiaisanearlyuniversalfeatureoftumorgrowth(HockelandVaupel,2001),conferringworsediseaseoutcomeviaprotectionfromapoptosis(Graeberetal.,1996),resistancetotherapy(Tre-danetal.,2007),andenhancedmetastaticcompetence(Coxetal.,2015).Thispathwayrequiresthetranscriptionalactivityofhypoxia-induciblefactor1(HIF1),amasterregulatorofoxygenhomeostasis(Keithetal.,2012)thatbecomesstabilizedupon
dropsinoxygenpressurebyescapingprolylhydroxylationandproteasome-dependentdestructionbythevonHippel-Lindautumorsuppressor(Semenza,2013).Inturn,nuclearlocalizedHIF1contributestooncogenesignaling(Mazumdaretal.,2010),angiogenesis(Ravietal.,2000),cellinvasion(Gilkesetal.,2014),andtumormetabolicreprogramming.
Inthiscontext,mitochondriaaretheprimarysitesofhypoxia-inducedmetabolicreprogrammingintumors(Denko,2008).ThisresponseinvolvesHIF1-dependenttranscriptionofmitochondrial
SignificanceTheabilitytoflexiblyadapttoanunfavorablemicroenvironmentisadistinctivefeatureoftumorcells,engenderingtreatmentresistanceandunfavorablediseaseoutcome.Lowoxygenpressure,orhypoxia,isapowerfuldriveroftumoradaptation,but‘‘druggable’’therapeutictarget(s)inthisresponsehaveremainedelusive.Here,weshowthathypoxictumorsrecruitapoolofactiveAkttomitochondria,culminatingwithAktphosphorylationofthemetabolicgatekeeper,pyruvatedehydro-genasekinase1.Thisphosphorylationstepimprovestumorfitness,preservestumorcellproliferationinthefaceofseverehypoxia,andisanegativeprognosticfactoringliomapatients.Repurposingsmall-moleculeAktinhibitorscurrentlyintheclinicmayprovideanapproachtopreventhypoxicreprogrammingandimproveanticancertherapy.CancerCell30,257–272,August8,2016ª2016ElsevierInc.257
Figure1.MitochondrialPhosphoproteomeinHypoxia
(A)PhosphoproteomeofprostateadenocarcinomaPC3cellsinhypoxiaversusnormoxia.IdentifiedphosphositesmetaminimumMaxQuantlocalizationprobabilityof0.75andascoredifferenceof5.Foldchangeswerecalculatedfromthenormalizedheavy/lightSILAC(stableisotopelabelingbyaminoacidsincellculture)ratio.SixAkttargetproteinsshowingincreasedphosphorylationinhypoxiaareindicated.Gray,notsignificant;red,upregulated;blue,downregulated;yellowsquares,Akttargets.
(B)Ingenuitypathwayanalysisofmitochondrialphosphoproteomeandglobalproteomeinhypoxia.
(C)Kinasesforwhichatleastfiveknowntargetsshowedsignificantchangesinphosphorylationinhypoxicversusnormoxicconditionsasin(A).Up,upregulation;Dn,downregulation.*Themodulatedgenesare:ARID1A;HIST1H1E;HMGA1;LARP1;LIG1;LIG3;LMNB2;LRCH3;LRWD1;MARCKS;MED1;MKI67;NCL;
(legendcontinuedonnextpage)
258CancerCell30,257–272,August8,2016
pyruvatedehydrogenasekinase(PDK)(Kimetal.,2006;Papan-dreouetal.,2006),whichinturnphosphorylatesthepyruvatede-hydrogenasecomplex(PDC)onthreeseparatesites(Pateletal.,2014).Bysuppressingtheoxidativedecarboxylationofpyruvateintoacetyl-CoA(Pateletal.,2014),anactivePDKshutsoffoxida-tivephosphorylation,lowerstheproductionoftoxicreactiveoxy-genspecies(ROS),andswitchestumorbioenergeticstowardglycolysis(Denko,2008),adriverofmoreaggressivediseasetraits(GatenbyandGillies,2004).Whathasremainedunclear,however,iswhetherHIF1-dependenttranscriptionisthesolemechanismforPDKactivationinhypoxia(Kimetal.,2006;Papandreouetal.,2006),andtheexistenceofotherpotentialregulatorsofthisresponsehasnotbeenwidelyinvestigated.Inthisstudy,weexaminedmechanismsofthetumorresponsetohypoxia.RESULTS
AMitochondrialAktPhosphoproteomeinHypoxia
Webeganthisstudybyprofilingthemitochondrialphosphopro-teomeofprostateadenocarcinomaPC3cellsexposedtoseverehypoxia(<0.5%oxygenfor48hr)versusnormoxia.Atotalof4,236phosphositeswereidentifiedinthephosphopeptide-en-richedsamples,withalargenumberofchangesinphosphoryla-tionlevelinhypoxia/normoxiasamples(Figures1AandS1A).Intotal,1,329phosphositesshowedasignificantchange(mini-mumfoldchangeof1.6)inatleastonesampleanalyzed(Fig-ure1A).Bybioinformaticsanalysis,themitochondrialphospho-proteomeinhypoxiacontainedregulatorsoforganelleintegrity,bioenergetics,generegulation,andproteostasis(Figure1B),whicharefunctionallyimplicatedintumorcellproliferation,motility,invasion,andapoptosis(FigureS1B).Tocomplementthesedata,wealsoexaminedchangesintheglobalmitochon-drialproteomeinhypoxiaversusnormoxia.Atotalof5,583pro-teinswereidentifiedinthisanalysis,and267oftheseproteinsshowedasignificantchangeinhypoxia/normoxiasamples(FigureS1C).Manyofthephosphositeswerenotmodulatedattheproteinlevel,suggestingthatthesephosphorylationeventswereindependentofproteinexpression.Inaddition,thehypoxia-regulatedmitochondrialphosphoproteomecontainedadiscrete‘‘Aktsignature’’(Figure1A),characterizedbyincreasedphosphorylationofsixAkttargetproteinsinhypoxiaversusnormoxia(Figure1C).
Basedontheseresults,wenextlookedataroleofAktinthetumorresponsetohypoxia.ExposureofPC3cellstohypoxiare-sultedinincreasedrecruitmentofAkttomitochondria,whereasthecytosoliclevelsofAktwereunchangedbetweenhypoxiaandnormoxia(Figures1DandS1D).Thehypoxia-regulatedpoolof
mitochondrialAktwas‘‘active’’asitwasphosphorylatedonSer473(Figure1D)andpersistedforupto24hrafterre-oxygen-ation(Figures1EandS1E).Consistentwiththeseresults,hypox-iawasaccompaniedbyincreasedphosphorylationofasetofmitochondrialproteinsrecognizedbyanantibodytotheAktconsensusphosphorylationsequence,RxRxxS/T(AktconsAb)(Figure1F).PreincubationofmitochondrialextractswithAktconsAb(Figure1F),orsilencingAkt2bysmallinterferingRNA(siRNA)(FigureS1F),removedthemitochondrialproteinsrecog-nizedbyAktconsAbinhypoxia,confirmingthespecificityofthisresponseandAkt-directedphosphorylationactivityinmitochon-driainhypoxia.SilencingAkt1hadminimaleffect(FigureS1F).siRNAsilencingofHIF1adidnotaffectAktrecruitmenttomitochondriainhypoxia(FigureS1G),suggestingthatthispathwaydidnotrequireHIF1-dependenttranscription.Inaddi-tion,depletionofHIF1adidnotaffectAktlevelsinthecytosolormitochondriaundernormoxicconditions,whereasphosphor-ylatedAkt2levelswereincreasedinthecytosolinresponsetohypoxia(FigureS1G).AsdetectedbyAktconsAb,theexpres-sionlevelsofdownstreamAkt-phosphorylatedtargetmoleculeswereunchangedinnormoxicorhypoxicconditions(Figure1F).Inresponsetohypoxia,activeAktwasfoundpredominantlyinthemitochondrialinnermembrane,and,toalesserextent,thematrix(FigureS1H).
Themechanism(s)ofhowAktisrecruitedtomitochondriainhypoxiawasfurtherinvestigated.Wefoundthatblockingthechaperoneactivityofheatshockprotein-90(Hsp90)with17-al-lylaminogeldanamycin(17-AAG)preventedtheaccumulationofmitochondrialAktinhypoxia(Figure1G).Also,scavengingmitochondrialROSwithMito-TempoinhibitedAktrecruitmenttomitochondria(Figure1H).TheantioxidantN-acetylcysteine(NAC)hadnoeffect(Figure1H),identifyingmitochondria-derivedROSasacriticalstimulusformitochondrialaccumulationofAktinhypoxia.
PDK1IsaPhosphorylationTargetofMitochondrialAktinHypoxia
Wenextsetupa1DproteomicsscreentoidentifymitochondrialproteinsphosphorylatedbyAktinhypoxia(Figure2A).ImmunecomplexesprecipitatedwithAktconsAbfromnormoxicorhypoxicPC3cellscontainedbandswithmolecularweightsof$35to$120kDathatweremoreabundantinhypoxia(Fig-ureS2A).PreclearingmitochondrialextractswithAktconsAbremovedmostoftheseproteins,validatingthespecificityoftheimmunoprecipitationstep.Fromtheseexperiments,weiden-tified84high-confidenceAktsubstratesdifferentiallyexpressedinhypoxiausingmassspectrometry(TableS1).Sixteenofthese
NPM1;NUCKS1;PDS5B;PTPN2;RB1;RBL1;RBL2;SAMHD1;SETDB1;TERF2;VIM.**Themodulatedgenesare:DUT;EEF1D;HIST1H1E;HMGA1;IRS2;LIG1;LIG3;LMNA;LMNB1;MAP4;NOLC1;NPM1;NUCKS1;PDS5B;PTPN2;RB1;SAMHD1;TCOF1;TOP2A;TPX2;VIM.
(D)PC3cellsinnormoxia(N)orhypoxia(H)werefractionatedincytosol(Cyto)ormitochondrial(Mito)extractsandanalyzedbywesternblotting.pAkt,phos-phorylatedAkt(Ser473);TCE,totalcellextracts.
(E)PC3cellsinhypoxia(H)wereexposedtore-oxygenation(O2)fortheindicatedtimeintervalsandanalyzedbywesternblotting.
(F)Theindicatedsubcellularfractionsisolatedfromnormoxic(N)orhypoxic(H)PC3cellswereanalyzedwithanantibodytotheAktconsensusphosphorylationsiteRxRxxS/T(AktconsAb)bywesternblotting.MitoSup,supernatantofmitochondrialextractsafterpreclearingwithAktconsAb.
(G)PC3cellsinnormoxia(N)orhypoxia(H)weretreatedwithvehicle(Veh)orHsp90small-moleculeinhibitor17-AAG(5mMfor6hr),andcytosolic(Cyto)ormitochondrial(Mito)extractswereanalyzedbywesternblotting.
(H)PC3cellsinnormoxia(N)orhypoxia(H)weretreatedwithvehicle(Veh),theantioxidantN-acetylcysteine(NAC,1mM)ormitochondria-specificROSscavenger,Mito-Tempo(MT,25mM),andsubcellularfractionswereanalyzedbywesternblotting.SeealsoFigureS1.
CancerCell30,257–272,August8,2016259
ABC
D
EFG
H
I
J
Figure2.MitochondrialAktPhosphorylationofPDK1
(A)SchematicdiagramfortheidentificationofamitochondrialAktphosphoproteomeinhypoxicversusnormoxicPC3cells.(B)MitochondrialproteinsreactingwithAktconsAbshowingdifferentialexpressioninhypoxicversusnormoxicPC3cells.
(C)RecombinantPDK1orGSK3bwasmixedinakinaseassaywithactiveAkt1orAkt2,andphosphorylatedbandsweredetectedwithAktconsAbbywesternblotting.
(D)TheindicatedPDKisoformsweremixedinthepresenceorabsenceofactiveAkt2inakinaseassay,andphosphorylatedbandsweredetectedwithAktconsAbbywesternblotting.
(E)PC3cellsinnormoxia(N)orhypoxia(H)wereimmunoprecipitated(IP)withanantibodytoPDK1followedbywesternblotting.HIF1areactivity(bottom)wasusedasamarkerofhypoxia.Bottom,densitometricquantificationofphosphorylated(p)PDK1bands.U,a.u.
(F)ExtractedionchromatogramofthePDK1-phosphorylatedT346chymotrypticpeptide(STAPRPRVEpTSRAVPL,m/z=908.9751)resultingfromincubationwithorwithoutactiveAkt1orAkt2inakinaseassay.
(G)PC3cellsweretransfectedwithvectororFlag-taggedwild-type(WT)PDK1ortheT346APDK1mutant,IPwithanantibodytoFlag,andimmunecomplexesweremixedwithactiveAkt2inakinaseassayfollowedbywesternblottingwithAktconsAb.Bottom,densitometricquantificationofphosphorylated(p)PDK1bands.U,a.u.
(H)MoleculardynamicssimulationofthestructureofPDK1(ribbon)withstickrepresentationofresidues336–356comprisingthe‘‘ATPlid.’’TheATPmoleculeisderivedfromthestructureofPDK3-L2-ATP(PDB:1Y8P)superimposedontothestructureofPDK1.Theproteinispresentedasribbondrawings.Theloopis
(legendcontinuedonnextpage)
260CancerCell30,257–272,August8,2016
moleculeswereknownmitochondrialproteins(Figure2B),includinghypoxia-andHIF1-regulatedeffectorsofbioenergetics(UGP2,SLC2A1,PDK1,HK2),extracellularmatrixremodeling(P4HA1),Ca2+homeostasisattheER-mitochondriainterface(Ero1L),oxidativephosphorylation(LonP1,IBA57),andmeta-bolism(Acot9).DuetopreviouslypublishedworksuggestingtheimportanceofPDK1inthetumorhypoxicresponse,wenextfocusedonPDK1asapotentialsubstrateofmitochondrialAktinhypoxia.Inkinaseassays,activeAkt1orAkt2readilyphosphorylatedPDK1,aswellascontrolGSK3b,asdeterminedbywesternblottingwithAktconsAb(Figure2C).Thisphosphor-ylationeventwasselectiveforPDK1,asrelatedPDK2,PDK3,orPDK4isoformswereunreactive(Figure2D).Inaddition,PDK1immunecomplexesreactedwithAktconsAbpreferen-tiallyinhypoxia(Figure2E)and,reciprocally,immunecomplexesprecipitatedwithAktconsAbinhypoxiacontainedPDK1(Fig-ureS2B),consistentwiththemodelofAktphosphorylationofPDK1inhypoxia.
WenextlookedforpotentialAktphosphorylationsitesinPDK1byliquidchromatography-tandemmassspectrometryanalysisofchymotrypsindigestsofAkt-phosphorylatedPDK1inakinaseassayseparatedbySDS-PAGE(FigureS2C).WeidentifiedThr346(T346)inanumberofPDK1chymotrypticpep-tides,includingthesequenceSTAPRPRVEpTSRAVPL(m/z=908.9751)asthesolephospho-aminoacidmodifiedbyAkt1orAkt2,comparedwithcontrol(Figure2F).ThePDK1sequencesurroundingT346matchedanAktconsensusphosphorylationsite,RxRxxS/T(FigureS2D),whichwasnotpresentinPDK2,PDK3,orPDK4(FigureS2E).Consistentwiththesedata,activeAkt2phosphorylatedwild-type(WT)PDK1butnotaphosphory-lation-defectiveThr346/Ala(T346A)PDK1mutantintrans-fectedPC3cells(Figure2G).InthePDK1crystalstructure,T346ispredictedtolocalizetoaflexible,‘‘ATPlid’’hingeregion(Figure2H),positionedtoaffectATPloadingandkinaseactivation.
Toindependentlyvalidatethesefindings,wenextgeneratedaphospho-specificantibodytophosphorylatedT346(pT346Ab)inPDK1.ThepT346Abdose-dependentlyreactedwiththephosphorylatedPDK1peptideCAPRPRVE(pT)SRAVPLA,butnotthenon-phosphorylatedsequence(FigureS2F).Asecondantibodyraisedagainstthenon-phosphorylatedsequencerecognizedthenon-phosphorylatedPDK1peptide(FigureS2G).Undertheseconditions,pT346AbreactedwithAkt2-phosphor-ylatedWTPDK1,butnottheT346APDK1mutant(Figure2I).ConsistentwiththemodelthatT346phosphorylationishypoxiasensitive,WTPDK1,butnotT346APDK1,precipitatedfromhyp-oxicPC3cellsreactedwithpT346Ab(Figure2J).pT346AbonlyweaklyreactedwithWTorT346APDK1precipitatedfromnor-moxiccells(Figure2J).
Finally,wegeneratedclonesofPC3cellsstablysilencedforendogenousPDK1byshorthairpinRNA(shRNA).pT346Abdidnotreactwiththesecellsinnormoxia(FigureS2H).In
contrast,pLKOtransfectantsreactedwithpT346Abinhypoxia,andthisresponsewasabolishedbyshRNAsilencingofPDK1(FigureS2H).
TheAkt-PDK1PhosphorylationAxisinHypoxia
ExpressionofWTPDK1inhypoxicPC3cellsincreasedthephosphorylationoftheE1acatalyticsubunit(PDHE1)ofthePDC(Figure3A)onsite1(Ser264inthematureprotein),oneofthreeregulatoryphosphorylationsites(Pateletal.,2014).Conversely,expressionoftheT346APDK1mutantreducedPDHE1phosphorylationinhypoxia(Figure3A),andnoPDHE1phosphorylationwasdetectedinnormoxia(Figure3A).ImmunecomplexesoftheWTorT346APDK1mutantcontainedcompa-rableamountsofthePDCcomponent,PDHE1a,suggestingthatT346doesnotcontributetoaPDK1-PDCcomplex(Fig-ureS3A).Inakinaseassay,activeAkt2increasedPDK1phos-phorylationofPDHE1(Figure3B).WhileWTPDK1phosphory-latedPDHE1inthepresenceofAkt2(Figure3C),theT346APDK1mutantwasineffective(Figure3C).Consistentwiththesedata,increasedPDHE1phosphorylationwasdetectedonlyinthepresenceofAkt2andPDK1,butnotPDK2,PDK3,orPDK4(FigureS3B).SilencingofAkt2inhibitedPDHE1phos-phorylationinhypoxia,whereasAkt1knockdownonlyhadapartialeffect(Figure3D).Asacomplementaryapproach,weusedapan-Aktsmall-moleculeinhibitor,MK2206,whichindis-tinguishablysuppressedAktphosphorylationinhypoxiaandnormoxia(FigureS3C).IncubationofPC3cellswithMK2206suppressedPDHE1phosphorylationinhypoxia(Figure3E).ThisresponsewasspecificbecausePDK1immunoprecipitatedfromMK2206-treatedcellsalsofailedtophosphorylatePDHE1inakinaseassayinhypoxia(FigureS3D).Innormoxia,MK2206hadnoeffectonPDHE1phosphorylationincellextracts(Figure3E)orinakinaseassaywithimmunoprecipitatedPDK1(FigureS3D),validatingthespecificityofAkt-directedphosphor-ylationinhypoxicconditions.
Asanindependentapproach,wenextgeneratedWTorkinasedead(KD)Akt2constructstargetedtothemitochondriabythecytochromecoxidasesubunit8mitochondrialimportsequence.Similartotheendogenousprotein,mitochondrial-targetedAkt2accumulatedinvarioussubmitochondrialfractions(FigureS3E).Functionally,mitochondrial-targetedAkt2-KDinhibitedPDHE1phosphorylationinhypoxicPC3cells(Figure3F),whereasnon-mitochondrial-targetedAkt2-KDhadnoeffect.TherewasnoPDHE1phosphorylationinthecytosolofhypoxicornormoxictu-morcells,andAkt2-KDormitochondrial-targetedAkt2-KDhadnoeffectinthesesettings(FigureS3F).Reciprocally,forcedexpressionofmitochondrial-targetedWTAkt2wassufficienttoincreasePDHE1phosphorylationevenintheabsenceofhypoxia(FigureS3G).
Finally,wereconstitutedPDK1-depletedcellswithvariousPDK1cDNAs.ExpressionofWTPDK1inthesesettingsrestoredPDHE1phosphorylationinhypoxia,whereasT346APDK1
coloredinyellowandtherestoftheproteinisinnavy.Thesidechainsoftheloopandtheligandareshownassticks,coloredredforoxygen,bluefornitrogen,andorangeforphosphate.ThepredictedlocationofThr346aswellasArg343andArg348isshown.
(I)Theexperimentalconditionsareasin(G)exceptthatFlag-PDK1immunecomplexesmixedwithactiveAkt2inakinaseassaywereanalyzedwithphospho-specificpT346Abbywesternblotting.Exp.,exposure.Bottom,densitometricquantificationofphosphorylated(p)PDK1bands.U,a.u.
(J)Flag-PDK1immunecomplexesasin(G)wereprecipitatedfromPC3cellsinnormoxia(N)orhypoxia(H)andanalyzedwithpT346Abbywesternblotting.Bottom,densitometricquantificationofpPDK1bands.U,a.u.SeealsoFigureS2andTableS1.
CancerCell30,257–272,August8,2016261
ACFDGBEHI
JKLMFigure3.AMitochondrialAkt-PDK1-PDHE1PhosphorylationAxisinHypoxia
(A)PC3cellsinnormoxia(N)orhypoxia(H)weretransfectedwithvector,WTPDK1,ortheT346APDK1mutantandanalyzedbywesternblotting.Bottom,densitometricquantificationofphosphorylated(p)PDHE1bands.U,a.u.
(B)Theindicatedrecombinantproteinsweremixedinakinaseassayandanalyzedbywesternblotting.
(C)PC3cellstransfectedwithvectorortheindicatedFlag-taggedWTPDK1ortheT346APDK1mutantwereIPwithanantibodytoFlag,andimmunecomplexesweremixedinakinaseassaywithrecombinantAkt2andPDHE1followedbywesternblotting.
(D)PC3cellsinnormoxia(N)orhypoxia(H)weretransfectedwithcontrolsiRNA(Ctrl)orsiRNAtoAkt1orAkt2,andanalyzedbywesternblotting.
(E)PC3cellsinnormoxia(N)orhypoxia(H)weretreatedwithvehiclecontrol(Veh)orasmall-moleculeAktinhibitor,MK2206(1mM),andanalyzedbywesternblotting.
(F)PC3cellsinnormoxia(N)orhypoxia(H)weretransfectedwithvector,Akt-kinasedead(Akt-KD),orthemitochondrial-targetedAkt-KD(mtAkt-KD)mutant,andmitochondrialextracts(Mito)wereanalyzedbywesternblotting.
(G)PC3cellsinnormoxia(N)orhypoxia(H)weretransducedwithpLKOorPDK1-directedshRNA,reconstitutedwithvector,WTPDK1,ortheT346APDK1mutantandanalyzedbywesternblotting.Bottom,densitometricquantificationofphosphorylated(p)PDHE1bands.U,a.u.
(H)PC3cellstransducedwithpLKOorPDK1-directedshRNAwereanalyzedforPDHactivityinnormoxia(N)orhypoxia(H).Left,representativetracings(n=4).Right,quantificationofPDHactivity.ns,notsignificant.Mean±SEM.*p=0.03.
(legendcontinuedonnextpage)
262CancerCell30,257–272,August8,2016
mutanthadnoeffect(Figure3G).Withrespecttoitsenzymaticfunction,PDK1knockdownincreasedPDHactivityinnormoxicPC3cells(Figure3H).HypoxiccellsshowedreducedPDHactiv-ity,andthisresponsewaspartiallyrescuedbyshRNAsilencingofPDK1(Figure3H).ReconstitutionofthesecellswithWTPDK1,butnottheT346APDK1mutant,suppressedPDHactivityinhyp-oxia(Figure3I).Inaddition,expressionofAkt-KDormitochon-drial-targetedAkt-KDinPC3cellshadnoeffectonPDHactivityinnormoxia,butmodestlyelevatedPDHfunctioninhypoxia(FigureS3H),consistentwithlossofanAkt-regulatedinhibitoryfunctionofPDK1inthesesettings.Vectorornon-mitochon-drial-targetedAkt2-KDhadnoeffect(FigureS3H).
MitochondrialAkt-PDK1PhosphorylationControlsTumorMetabolicReprogramming
TounderstandhowmitochondrialAkt-PDK1signalingaffectstumorbehavior,wefirstlookedatpotentialchangesincancermetabolism.Consistentwithpreviousstudies,hypoxiastimu-latedglycolyticmetabolismintumorcells,characterizedbyincreasedglucoseconsumption(Figure3J)andlactateproduc-tion(Figure3K).MitochondrialAkt-PDK1signalingwasimpor-tantforthisresponse,asPDK1knockdownreducedglucoseconsumptioninhypoxia,whereasreconstitutionoftargetedcellswithWTPDK1,butnottheT346APDK1mutant,restoredglycolysis(Figure3J).Similarly,AktinhibitionwithMK2206(Figure3K)orsilencingofAkt2(Figure3L)impairedmetabolicreprogramming,reducinglactateproductioninhypoxia.Nor-moxiccultureswerenotaffected(Figures3Kand3L),andAkt1knockdownhadonlypartialeffect(Figure3L).Consistentwithametabolicswitchtowardglycolysis(Kimetal.,2006;Pa-pandreouetal.,2006),PC3cellsreconstitutedwithWTPDK1exhibitedreducedoxygenconsumption,amarkerofoxidativephosphorylation,whereasexpressionoftheT346APDK1mutantrestoredoxygenconsumption(Figure3M),furthersup-portingaroleofmitochondrialAkt-PDK1signalinginhypoxicmetabolicreprogramming.
MitochondrialAkt-PDK1PhosphorylationInVivo
Whenanalyzedintimecourseexperiments,hypoxiaincreasedphosphorylationofAkt1andAkt2,aswellasPDHE1,startingat3and6hr,respectively(FigureS4A).Theoverallhypoxicresponseundertheseconditionswascell-typespecific.Aktinhi-bitionstronglyreducedPDHE1phosphorylationinprostateadenocarcinoma(DU145),lungadenocarcinoma(A549),andglioblastoma(GBM)(LN229),buthadnoeffectonPDHE1phos-phorylationinbreastadenocarcinomacellsMCF-7(ER+)orMDA-231(ERÀ)(FigureS4B).Knockdownofphosphataseandtensinhomolog(PTEN)inMCF-7cellsincreasedPDHE1phos-phorylationinnormoxiaand,toalesserextent,hypoxia,whereas
LN229cellswereunaffected(FigureS4C),suggestingthatPTENstatusmaydifferentiallyaffecthypoxia-stimulatedmitochondrialAkt-PDK1signalingdependingonthetumorcelltype.
Toexamineamore‘‘physiologic’’modeloftumorhypoxia,wenextlookedat3Dculturesofpatient-derived,stemcell-enrichedGBMneurospheres(DiCristoforietal.,2015).Theseculturesbecomehypoxicintheir‘‘core,’’asdeterminedbyexpressionofahypoxiaprobe(Figures4A,S4D,andS4E).Underthesecon-ditions,GBMneurospheresexhibitedstrongphosphorylationofPDK1,asdeterminedbyimmunofluorescencewithpT346Ab(Figure4A).Conversely,differentiatedGBMcellsdepletedofstemcellsandgrowingasmonolayerswerenormoxic,containedcytosolicHIF1a,anddidnotreactwithpT346Ab(Figure4A).Pre-absorptionofpT346Abwiththeimmunizingpeptideabol-ishedreactivitywithGBM(FigureS4D).
Next,welookedatAktphosphorylationofPDK1inprimary,patient-derivedGBMtissuesamples(TableS2).GBMswithahighscore(R2)fornuclearHIF1ashowedincreasedphosphor-ylationofPDK1byAkt(pT346Ab),aswellasphosphorylationofPDHE1andSrc,amajordeterminantofgliomainvasiveness(Duetal.,2009),inhypoxicareas(Figures4BandS4F).Incontrast,GBMswithundetectablenuclearHIF1a(score=0)showedlowtoundetectablelevelsofPDK1-PDHE1phosphorylation(Figures4CandS4F).Inthesepatients,phosphorylationofPDK1(pT346Ab)(Figure4D)orPDHE1phosphorylation(Figure4E)correlatedwithexpressionofnuclearHIF1a.Reciprocally,PDHE1phos-phorylationcorrelatedwiththeexpressionofAkt-phosphory-latedPDK1(pT346Ab)(Figure4F),reinforcingalinkbetweenhypoxiaandmitochondrialAkt-PDK1phosphorylationinprimarypatientsamples.
MitochondrialAkt-PDK1RegulationofTumorCellProliferationinHypoxia
TotestaroleofamitochondrialAkt-PDK1signalingintumorgrowthinvivo,wefirstutilizedhumanU251GBMcellsexpress-ingaluciferasereporterunderthecontrolofaHIF1-responsiveelement(HRE)andamCherryreporterunderaconstitutivephos-phoglyceratekinasepromotertoquantifycellviability.Stereo-tacticintracranialinjectionofthesecellsinimmunocompro-misedmicegaverisetoGBMscharacterizedbyHIF1-directedluciferaseactivityandreactivitywithahypoxia-sensitivemarker(Figures5Aand5B).Despitelowoxygenation,theseorthotopicGBMsremainedviable,asdeterminedbyhighmCherryexpres-sion(Figures5Aand5B),andexhibitedatime-dependentin-creaseinthenumberofmitotictumorcells(FigureS5A).TheseproliferatingcellsstainedintenselypositiveforAkt-phosphory-latedPDK1(Figures5C,S5B,andS5C),correlatingwithincreasedHIF1activity(FigureS5D).PDHE1wasalsohighlyphosphorylatedinintracranialGBMs(Figure5C).
(I)PC3cellsinhypoxiaweretransducedwithPDK1-directedshRNA,reconstitutedwithvector,WTPDK1,ortheT346APDK1mutantcDNAandanalyzedforPDHactivity.Left,representativetracings(n=3).Right,quantificationofPDHactivity.Mean±SEM.**p=0.009.
(J)PC3cellstransducedwithpLKOorPDK1-directedshRNAwerereconstitutedwithvector,WTPDK1,ortheT346APDK1cDNAandanalyzedforglucoseconsumption(n=4).Mean±SEM.***p<0.0002.
(K)PC3cellsinnormoxia(N)orhypoxia(H)weretreatedwithvehiclecontrol(Veh)orAktinhibitor,MK2206(1mM),andanalyzedforlactateproduction(n=3).Mean±SEM.**p=0.001–0.004,***p=0.0005–0.0009.
(L)PC3cellsinnormoxia(N)orhypoxia(H)weretransfectedwithcontrolsiRNA(Ctrl)orsiRNAtoAkt1orAkt2andanalyzedforlactateproduction(n=2).Mean±SD.**p=0.004,***p=0.0005.
(M)PC3cellsstablysilencedforPDK1weretransfectedwithvector(Vec),WTPDK1orT346APDK1mutant,andanalyzedforoxygen(O2)consumption(n=3).Mean±SEM.Forallpanels,datawereanalyzedusingthetwo-sidedunpairedStudent’sttests.SeealsoFigureS3.
CancerCell30,257–272,August8,2016263
A
BC
DEF
No.of patientsNo.of patientsFigure4.MitochondrialAkt-PDK1PhosphorylationInVivo
(A)GBMneurospheres(top)ordifferentiatedGBMcultures(bottom)werestainedforDNA(DAPI),HIF1a,pT346-phosphorylatedPDK1,orhypoxia(hypoxia-sensitiveprobe).Mergedimagesofnuclear-localizedHIF1ainhypoxicneurospheres(byvelocitymask)areindicated(Merge).Yellowbox,VolocityanalysistoidentifycellswithnuclearHIF1aineachsinglezstack.Scalebar,20mm.
(BandC)Immunohistochemicalstainingofprimary,patient-derivedGBMsampleswithhigh(R2)(B)orlow(0)(C)scoreforHIF1aandphosphorylatedprotein(pProt)expression.Scalebars,100mm.
(D–F)Quantitativeimmunohistochemicalcorrelationofpatient-derivedGBMsamples(n=24)orgradeIIgliomas(n=2)forHIF1aexpressionandpPDK1(D)orpPDHE1(E),orbetweenpPDK1andpPDHE1(F).Fourtissuemicroarray(TMA)cores/patient.Thescoringisasfollows:0,nostaining;1,staininginatleastoneTMAcore;2,staininginR2TMAcores.Theindividualpvaluespereachanalysisareindicated(chi-squaretest).SeealsoFigureS4andTableS2.
WenexttestedtherequirementofmitochondrialAkt-PDK1signalinginregulatingproliferationunderhypoxicconditions.siRNAknockdownofAkt1orAkt2(Figure5D)orstableshRNAknockdownofPDK1(Figure5E)suppressedtumorcellprolifer-ationinhypoxia.Normoxiccultureswerepartiallyaffected(Fig-264CancerCell30,257–272,August8,2016
ures5Dand5E).Whencellswereanalyzedforcellcycletransi-tions,MK2206orthePDK1inhibitordichloroacetatesuppressedS-phaseprogressioninhypoxiaandincreasedthepopulationoftumorcellsinG1/subG1phase(FigureS5E).Finally,stablesilencingofPDK1abolishedPC3colonyformationinhypoxia,
No.of patientsA
BC
DEF
GFigure5.RequirementofMitochondrialAktforTumorCellProliferationinHypoxia
(AandB)BioluminescenceimagingofimmunocompromisedmicecarryingU251intracranialGBMs(threeanimals/group)expressingluciferaseunderthecontrolofHIF1-responsiveelements(Luc)andmCherry(cellviability)andexposedtoahypoxia-sensitiveprobe(Hypox).Scanswereobtainedatdays20and34(A)andfluorescencesignalswerequantified(B).*p=0.016–0.057byMann-Whitneytest.
(C)TissuesamplesfromintracranialGBMsasin(A)wereharvestedatday34andanalyzedforexpressionofHIF1a,phosphorylated(p)PDK1(pT346Ab),orpPDHE1,byimmunohistochemistry.Yellowlineswereusedtodelineatethetumormasswithinmousebrains.Scalebar,100mm.Asterisks,mitoticcells;insets(H&EandpPDK1panels),high-powermagnificationofmitoticcells.Scalebars,25mm.
(DandE)PC3cellstransfectedwithcontrolsiRNA(Ctrl),Akt1-orAkt2-directedsiRNA(D),orstablytransducedwithpLKOorPDK1-directedshRNA(E),wereanalyzedforcellproliferationinnormoxiaorhypoxiabydirectcellcounting(n=5).Mean±SEM.***p<0.001,**p=0.002.
(FandG)PC3cellsstablytransducedwithpLKOorPDK1-directedshRNAwereanalyzedinnormoxiaorhypoxiaforcolonyformationbycrystalvioletstainingafter10days(F)andquantified(n=3)(G).Mean±SEM.ns,notsignificant.**p=0.003.Forallpanels,datawereanalyzedusingthetwo-sidedunpairedStudent’sttest.SeealsoFigureS5.
CancerCell30,257–272,August8,2016265
AB
C
D
EFGHI
Figure6.MitochondrialAktRegulationofStressSignalinginHypoxia
(AandB)PC3cellsinnormoxia(N)orhypoxia(H)weretreatedwithvehicle(Veh)orMK2206(1mM)(A)ortransducedwithpLKOorPDK1-directedshRNA(B)andanalyzedforROSproductionbyCellROXGreenstainingandflowcytometry.Upperpanels,representativetracings.Bottompanels,quantificationofROSproductionunderthevariousconditionstested(n=2).Mean±SDforbothdatasets.*p=0.01–0.02,**p=0.004;ns,notsignificant.
(C)Theexperimentalconditionsareasin(A)andtreatedcellswereanalyzedforcellviabilitybydirectcellcountingrelativetocontrol(n=3).Mean±SEM.***p<0.0001.
(D)PC3cellsinnormoxia(N)orhypoxia(H)wereincubatedwithvehicle(Veh)orsmall-moleculeinhibitorsofAkt(MK2206,1mM)orPI3K(PX-866,10mM)andanalyzedbywesternblotting.
(E)PC3cellsstablysilencedforPDK1werereconstitutedwithvector,WTPDK1,ortheT346APDK1mutantandanalyzedforcellviabilitybydirectcellcountingrelativetocontrol(n=3).Mean±SEM.***p=0.0002.
(FandG)PC3cellsinnormoxia(N)orhypoxia(H)weretransducedwithpLKOorPDK1-directedshRNA(F),controlsiRNA(Ctrl),orAkt1-orAkt2-directedsiRNA(G),andanalyzedbywesternblotting.
(HandI)PC3cellsasin(E)wereanalyzedforLC3reactivitybyfluorescencemicroscopy.Scalebars,10mm(H),andcellswithLC3puncta(>3)werequantified(n=250–860cells)(I).Mean±SEM.*p=0.014,***p=0.0005;ns,notsignificant.Forallpanels,datawereanalyzedusingthetwo-sidedunpairedStudent’sttest.SeealsoFigureS6.
amarkeroftumorigenicity(Figures5Fand5G),whereasnor-moxicgrowthwasnotsignificantlyaffected.Together,thesedatapointtoanimportantroleofmitochondrialAkt-PDK1signalinginmaintainingtumorcellproliferationinhypoxia.MitochondrialAktRegulationofStressSignalinginHypoxia
ThedownstreamimplicationsofdefectivemitochondrialAkt-PDK1signalingwerenextinvestigated.First,inhibitionofAkt
266CancerCell30,257–272,August8,2016
withMK2206(Figure6A)orstableshRNAsilencingofPDK1(Fig-ure6B)increasedaberrantROSproductionintumorcells,espe-ciallyinhypoxia.Thiswasassociatedwithdecreasedtumorcellviability(Figure6C),andappearanceofcleavedcaspase3(Fig-ure6D),amarkerofapoptosis.Innormoxia,cleavedcaspase3wasundetectable.Confirmingthespecificityofthisresponse,exposureoftumorcellstoasmall-moleculeinhibitorofphospha-tidylinositol3-kinase(PI3K),PX-866,didnotresultincaspaseacti-vation(Figure6D).ReconstitutionofthesecellswithWTPDK1,but
nottheT346APDK1mutant,partiallyrescuedtumorcellviabilityinhypoxia(Figure6E).Normoxiccultureswerenotaffected,furthersupportingaroleofPDK1signalingselectivelyinhypoxia.
Asaseconddownstreampathwayoftumormaintenancemodulatedbybioenergetics,wenextobservedthatstableknockdownofPDK1(Figure6F)orsiRNAsilencingofAkt1orAkt2(Figure6G)inhypoxiaincreasedthephosphorylationoftheenergystresssensor,AMP-regulatedkinase(AMPK).Thisresponsewasassociatedwithconcomitantactivationofauto-phagy,asdeterminedbyLC3conversiontoalipidatedform(Fig-ures6Fand6G)andpunctateLC3fluorescencestaining(Figures6Hand6I).Normoxicculturesshowedaminimallevelofauto-phagyinductionafterPDK1silencing(Figures6F,6G,andS6A).InPDK1-depletedcells,re-expressionofWTPDK1,butnottheT346APDK1mutant,attenuatedAMPKphosphorylationandreducedautophagyinhypoxia(Figures6H,6I,andS6B).RequirementofHypoxicMitochondrialReprogrammingforTumorGrowthInVivo
Next,weaskedifmitochondrialAkt-PDK1signalingwasimpor-tantfortumorgrowthinvivo.shRNAsilencingofPDK1signifi-cantlyimpairedthegrowthofPC3xenografttumorsimplantedinimmunocompromisedmice(Figure7A).Re-expressionofWTPDK1inthesecellsrestoredtumorgrowthinvivo(Figures7Band7C),whereastheT346APDK1mutantfurtherimpairedtumorgrowth(Figure7C).Byimmunohistochemistry,PC3tumorsharboringWTPDK1showedincreasedcellproliferation,reducedapoptosis,andlowerlevelsofautophagycomparedwithtumorsreconstitutedwiththeT346APDK1mutant(Figures7Dand7E).Inaddition,tumorswithlossofendogenousPDK1showedatrendtowardincreasedapoptosisandheightenedautophagyinvivo,whereastumorcellproliferationbyKi-67stainingwasunchanged(Figures7Fand7G).TheseresultssuggestthatmitochondrialAkt-PDK1signalingpromotestumoradaptationtohypoxia,andspecificallyenablescontinuedtumorcellproliferationdespiteanunfavorablemicroenvironment(Figure7H).
Totesttherelevanceofthismodelinhumancancer,wenextlookedattheprognosticimpactofAktphosphorylationofPDK1inaclinicallyannotatedcohortof116gliomapatients(TableS3).Undetectableinnormalbrainparenchyma,theexpressionofAkt-phosphorylatedPDK1onT346progressivelyincreasedingliomas,withthehighestreactivityobservedinGBM(Figures8Aand8B).PDK1phosphorylationonT346segregatedwithothermarkersofdiseaseprogression,includingHIF1aexpression(Figure8C),WTstatusofisocitratedehydrogenase-1(IDH1)(Figure8D),andunmethylatedO6-methylguanine-DNAmethyl-transferase(MGMT)promoter(Figure8E).Consistentwiththisprognosticprofile,elevatedexpressionofAkt-phosphorylatedPDK1,asdeterminedbyreceiveroperatingcharacteristics(ROC)curvesanalysis(FiguresS7AandS7B),wassignificantlyassociatedwithreducedoverallsurvivalinpatientswithgliomas(p=0.006;hazardratio[HR]=2.2;95%confidenceinterval[CI]:1.17–4.12;Figure8F)aswellaspatientswithGBM(p=0.032;HR=2.03;95%CI:0.95–4.32;Figure8G).DISCUSSION
Inthisstudy,wehaveshownthathypoxia,auniversaldriverofmalignancy,promotestherecruitmentofactiveAkttomitochon-driaoftumorcells.Inturn,mitochondrialAkt,inparticularAkt2,phosphorylatesthebioenergeticsregulatorPDK1onaThr346targetsite,resultinginincreasedkinaseactivityandphosphory-lationofitsdownstreamsubstrateinthePDC,PDHE1a.Thispathwayimprovestumorfitness,stimulatingglycolysis,coun-teringautophagyandapoptosis,dampeningoxidativestress,andenablingcontinuedcellproliferationinfaceofseverehypoxiainvivo.Accordingly,mitochondrialAkt-PDK1signalingemergedasapowerfulnegativeprognosticfactoringliomapa-tients,correlatingwithmarkersofunfavorableoutcomeandshortenedsurvival.
Aktisanessentialsignalingnodeexploitedinmostcancers,integratinggrowthfactorresponseswithmechanismsofcellproliferation,survival,andbioenergetics(ManningandCantley,2007).Theroleofthispathwayasaregulatoroftumoradaptationtohypoxiahasnotbeenpreviouslydescribed,anditsspatialarrangementinsubcellularcompartments,inparticularmito-chondria,hasonlyrecentlybeguntoemerge(Ghoshetal.,2015).Datapresentedheresuggestthattherecruitmentofpre-dominantlyactiveAkttomitochondriaduringhypoxia(SantiandLee,2010)maybepartofabroaderstressresponseintumors,enabledbythechaperonefunctionofcytosolicHsp90inmito-chondrialpreproteinimport(Youngetal.,2003)andmitochon-drialROSproduction,whichmayparticipateinsubcellulartraf-fickingofsignalingmolecules(Nakahiraetal.,2006),includingmitochondrialimport(FischerandRiemer,2013).
Inourphosphoproteomicsscreen,Aktrecruitmenttomito-chondriawasassociatedwithadiscreteAktphosphorylationsignaturethatincludedregulatorsoforganellehomeostasisandglycolyticreprogramminginhypoxia,suchas6-phospho-fructose-2-kinase/fructose-2,6-bisphosphatase3andPDK1(DeBocketal.,2013).InthecaseofPDK1,Aktphosphorylationtookplaceexclusivelyinhypoxia,didnotinvolveotherPDKiso-forms,andtargetedauniqueThr346siteinthe‘‘ATPlid’’(Zhangetal.,2015),ideallypositionedtoaffectATPloading,andkinaseactivation(Pateletal.,2014).Thr346didnotaffectPDK1bindingtothePDC,thusdifferentlyfromanotherproposedpost-transla-tionalmodificationofPDK1involvingTyrphosphorylationofthe‘‘ATPlid’’(Hitosugietal.,2011).
ExtensivelystudiedaspartofHIF1signaling(Semenza,2013),thetumorresponsetohypoxiahasbeenlinkedtometabolicre-programming(Kimetal.,2006;Papandreouetal.,2006),withsuppressionofmitochondrialrespirationinfavorofglycolysis(Denko,2008).However,thereisevidencethatthispathwaymayextendwellbeyondbioenergetics,asPDK1signalinghasbeenimplicatedinsenescence(Kaplonetal.,2013),metastatictropism(Dupuyetal.,2015),andmultiplemechanismsoftumormaintenance(McFateetal.,2008).Thefactthatthisresponseis‘‘druggable’’andthatsmall-moleculePDK1inhibitorshaveenteredclinicaltestingincancerpatients(Michelakisetal.,2010)addstotherelevanceofPDK1signalingasapotentialdriveroftumorprogression.
Againstthisbackdrop,thepathwayofmitochondrialAkt-PDK1signalingdescribedhereappearsideallypoisedtoaffectaplethoraofdownstreamsignalingmechanismsimportantfortumoradaptationandimprovedfitness.Thisinvolvessuppres-sionofapoptosisviamitochondrialAktphosphorylationofhexo-kinaseIIattheoutermembrane(Robertsetal.,2013)andcyclo-philinDinthepermeabilitytransitionpore(Ghoshetal.,2015),as
CancerCell30,257–272,August8,2016267
ADEB
CF
GH
Figure7.MitochondrialAkt-DirectedHypoxicReprogrammingSupportsTumorGrowthInVivo
(A)PC3cellstransducedwithpLKOorPDK1-directedshRNAwereinjectedsubcutaneously(s.c.)intheflanksofmaleNSGimmunocompromisedmice(threeanimals/group;twotumors/mouse),andsuperficialtumorgrowthwasquantifiedwithacaliperattheindicatedtimeintervalsfor20days.Datawereanalyzedusingthetwo-sidedunpairedStudent’sttest.Mean±SEM.***p<0.0001.
(B)PC3cellsstablytransducedwithpLKOorPDK1-directedshRNAwerereconstitutedwithWTPDK1ortheT346APDK1mutantandinjecteds.c.intheflanksofimmunocompromisedmice(fivemice/group;twotumors/mouse).Tumorgrowthinthevariousgroupswasquantifiedattheindicatedtimeintervalsfor20days.Datawereanalyzedusingthetwo-sidedunpairedStudent’sttest.Mean±SEM.*p=0.01–0.02,***p<0.0001.
(C)PC3cellsstablytransducedwithpLKOorPDK1-directedshRNAwerereconstitutedwithvector,WTPDK1,ortheT346APDK1mutantandinjecteds.c.inimmunocompromisedmicewithdeterminationoftumorgrowthafter18days.Eachpointcorrespondstoanindividualtumor.
(DandE)Tumorsharvestedfromtheanimalsin(C)wereanalyzedforhistology(D),andcellproliferation(top,Ki-67),autophagy(middle,LC3-II),orapoptosis(bottom,TUNEL)wasquantified(E).ThestatisticalanalysisofthevariousgroupsbyANOVAisasfollows:Ki-67,***p<0.0001;LC3,*p=0.024;TUNEL,*p=0.039.Scalebars,100mm.
(FandG)SuperficialflanktumorsofPC3cellstransducedwithcontrolpLKOorPDK1-directedshRNAwereharvestedafter18daysandprocessedforimmunohistochemistry(F)withquantificationofreactivityforKi-67(top),LC3(middle),orTUNEL(bottom)(H).Representativeimagespereachconditionareshown(n=3,10imagespermouse),Mean±SD.Scalebars,100mm.
(H)SchematicmodelofamitochondrialAkt-PDK1-PDHE1phosphorylationaxisinhypoxictumorreprogramming.
268CancerCell30,257–272,August8,2016
A
BCD
FG
E
Figure8.MitochondrialAktPhosphorylationofPDK1IsaNegativePrognosticMarkerinHumanGliomas
(A)Representativemicrographsofimmunohistochemicalstainingofnon-neoplastichumanbrainparenchyma(normal)orgradeII–IVgliomas(WHOclassifi-cation)withPDK1pT346Ab.OD,oligodendroglioma;AOD,anaplasticOD.Scalebar,100mm.
(B)QuantificationofpT346staininginaseriesofhumanbraintumors(n=116)and85non-neoplasticbrainparenchymausingatwo-factorscoringsystemthatconsidersthepercentageofpositivecellsandtheintensityofthestaining(pPDK1score).***p<0.0001byMann-WhitneyUtest.Eachsymbolrepresentsanindividualpatient.
(C–E)DifferencesinpPDK1scoreinhumanbraintumorsasin(B)(n=116)accordingtonuclearHIF1aexpression(C;**p=0.008byMann-WhitneyUtest),IDH1mutationstatus(D;*p=0.02byMann-WhitneyUtest),orMGMTpromotermethylation(E;*p=0.01byMann-WhitneyUtest).DataarepresentedasTukeybox-and-whiskerplots.Thebottomandtopoftheboxrepresentthefirstandthirdquartiles,andthebandinsidetheboxrepresentsthemedian(i.e.,thesecondquartile).Thebottomendofthewhiskerrepresentsthelowestdatumwithinthe1.5interquartilerange(IQR)ofthelowerquartile,andthetopendofthewhiskerrepresentsthehighestdatumwithin1.5IQRoftheupperquartile.Outlierdata,ifany,arerepresentedbysinglepoints.
(FandG)Kaplan-Meiercurvesweregeneratedwitheitherthecompleteseriesofgliomapatients(n=116;F)orwithGBMcasesonly(n=61;G)sortedinto‘‘Low’’or‘‘High’’groupsaccordingtopPDK1score.CutoffstorankpatientsinthesetwocategoriesweregeneratedusingROCcurvesandtheYoudenJstatistic.Overallsurvivalcurveswerecomparedusingthelogranktest.HR,hazardratio;CI,confidenceinterval.SeealsoFigureS7andTableS3.
wellasinhibitionofoxidativephosphorylationthroughAktphos-phorylationofPDK1andsubsequentinhibitionofthePDC(Pateletal.,2014).Inturn,thislowerstheproductionoftoxicROS,pre-ventsthephosphorylationofthestressenergysensor,AMPK(LiangandMills,2013),andinhibitsdownstreamactivationofautophagy(White,2012).Althoughthesepathwayshavebeenimplicatedinbothtumorsuppressionandoncogenesis(LiangandMills,2013;White,2012),thereisevidencethatAMPKinhi-bitionandsuppressionofautophagymaypromotemalignantexpansion(White,2012),andheightenedmetastaticcompe-tence(Cainoetal.,2013),includinginhypoxia(Liuetal.,2015),furthercompoundingthemoreaggressivetraitsofglycolytictu-mors(GatenbyandGillies,2004).
Here,apivotalfeatureofmitochondrialAkt-PDK1signalingwasitsactivationinmitoticcellsanditsrequirementtosupporttumorcellproliferationinthefaceofhypoxiainvivo.WhetherthisresponsecanbeentirelyattributedtothesuppressionofROSorinvolvesothermechanismsofmitochondria-to-nucleiretrogradesignaling(Jazwinski,2013)remainstobeelucidated.Ontheotherhand,hypoxicreprogrammingmayparticipateincellcycletransitionsviaregulationofp27expression(Gardneretal.,2001)orMyctranscriptionalactivity(Gordanetal.,
CancerCell30,257–272,August8,2016269
2007),andeffector(s)ofglycolysishavebeenlinkedselectivelytochromosomalsegregationandmitoticprogressioninhypoxia(Jiangetal.,2014).InlinewiththebroadimpactofmitochondrialAkt-PDK1signalingonmultiplemechanismsoftumoradapta-tion,thispathwaywasfoundtohavesignificantimplicationsfordiseaseprogressioninhumans.Accordingly,expressionofPDK1phosphorylatedonT346wasundetectableinnormalbrain,butincreasedsteadilyinthehypoxicenvironmentofgli-omas,includingGBMs,segregatedwithunfavorableprognosticmarkers,andwascorrelatedwithshortenedoverallsurvivalinthesepatients.Althoughtheseresultsawaitfurtherconfirmationinlargerpatientcohorts,detectionofAkt-phosphorylatedPDK1mayprovideaneasilyaccessiblebiomarkerforclinicaldecisionmakinginpatientswithgliomas,includingGBM(Wicketal.,2014).
Despiteexpectationsforpersonalized,orprecisionmedicine,small-moleculeantagonistsofAktanditsassociatedsignalingnodes,PI3Kandmammaliantargetofrapamycin(ManningandCantley,2007),haveproducedonlylimitedresponsesintheclinicalsetting,hamperedbydrugresistanceandsignificanttoxicity(FrumanandRommel,2014).Whiletheseresultsmayreflectmechanismsoftumoradaptation(Ghoshetal.,2015),includingmitochondrialreprogramming(Cainoetal.,2015),theidentificationofmitochondrialAkt(Ghoshetal.,2015)asapost-translationalregulatorofPDK1activityandtumorprogres-sionmayrationallyrepurposeAkt-directedmoleculartherapiesasanapproachtoimpairtumoradaptationtohypoxia.Inthiscontext,targetedinhibitionofthemitochondrialpoolofAktmayselectivelydisableahostofmetabolic,survival,andprolif-erativerequirementsfortumorfitness(McFateetal.,2008)andreawakenendogenoustumor-suppressormechanismsimpor-tantforanticanceractivityinpatients.
EXPERIMENTALPROCEDURES
Patients
Allpatient-relatedstudieswerereviewedandapprovedbyanInstitutionalRe-viewBoardatFondazioneIRCCS,Ca’GrandaOspedaleMaggiorePoliclinicoMilan,Italy.Afirstcohortof26patientsdiagnosedwithdenovogliomawereenrolledatFondazioneIRCCSCa’GrandaOspedaleMaggiorePoliclinico(Mi-lan,Italy)between2010and2011,anddescribedpreviously(DiCristoforietal.,2015).Allpatientsweretreatedwithsurgicalresectionwithcurativeintent.Gli-omaswerestagedaccordingtotheWHOclassification(Louisetal.,2007),andtheclinicopathologicandmolecularcharacteristicsofthepatientseriesusedinthisstudyaredescribedinTableS2.ThiscohortwasusedtoevaluatetheexpressionofphosphoT346-PDK1(pPDK1),phosphoPDHE1a(pPDHE1),phosphoT416-Src(pSrc),andnuclearHIF1areactivity,byimmunohistochem-istry.Tissuemicroarraysofgliomaornormalbraintissueswerecarriedoutasdescribedpreviously(DiCristoforietal.,2015).ImmunohistochemistryslidesweredigitalizedusinganAperioscannerat203magnification,andHIF1anu-clearstainingwasquantifiedusinganuclear-specificalgorithmimplementedinGenieHistologyPatternRecognitionsoftware(Aperio,LeicaMicrosystems).TospecificallyquantifynuclearHIF1aexpressioninprimaryGBMculture,theVolocityalgorithmthatcountsanddisplaysredsignals(HIF1a)withintheHoechstsignal(nuclei)ineachzstackwasused.Asecondseriesof116pa-tientswithdenovogliomawhounderwentsurgerywithcurativeintentbe-tween2008and2009,andforwhomcompleteclinicalandfollow-uprecordswereavailable(TableS3),wasretrievedfromthearchivesofthePathologyDivision.ThiscohortwasusedinthepresentstudytocorrelatetheexpressionofAkt-phosphorylatedPDK1onT346withprognosticmarkersofgliomapro-gression,includingnuclearHIF1a,MGMTpromotermethylationandIDH1mutationalstatus(WT/R132H),andpatients’overallsurvival.
270CancerCell30,257–272,August8,2016
XenograftTumorGrowthStudies
AllexperimentsinvolvinganimalswereapprovedbyanInstitutionalAnimalCareandUseCommitteeatTheWistarInstituteinaccordancewiththeGuidefortheCareandUseofLaboratoryAnimalsoftheNIH,or,alternatively,attheUniversityofMilan,incompliancewiththeItalianMinistryofHealth.Inafirstsetofexperiments,PC3cellsstablytransducedwithcontrolpLKOorPDK1-directedshRNAwerereconstitutedwithvector,WTPDK1orT346APDK1mutantcDNAat80%confluency,suspendedinPBS(pH7.4),andinjected(0.2mLcontaining23106cells)subcutaneouslyintotheflankof6–8-week-oldmalenon-obesediabetic(NOD)severecombinedimmunodeficiencygamma(NSG,NOD.Cg-PrkdcscidIl2rgtm1Wjl/SzJ)immunocompromisedmice(JacksonLaboratory,threemicepercondition/twotumorspermouse).Thewidthandlengthofsuperficialtumorsweremeasuredwithacaliperattheindi-catedtimeintervals,andtumorvolumewascalculatedaccordingtothefor-mulavolume=width23length/2.After21daysxenografttumorswerehar-vested,fixed,andprocessedforimmunohistochemistry.
AnorthotopicmurinemodelofGBMwasobtainedbystereotacticinjection(coordinates:1.5mmlateraltothebregma,0mmbehindand3.0mmventraltothedura)(Maesetal.,2009)of13105U251-HRE-mCherryGBMcellsin2mLofPBSinto7–8-week-oldfemalenudemice(HarlanLaboratories)atday0(LoDicoetal.,2014).Followingsurgery,miceweremonitoredforrecoveryuntilcompleteawakening.Sixanimalspertimepointwereusedandmicewerekilledafter20or34days.IntracranialGBMsampleswereharvestedfromthevariousgroupsandprocessedfordifferentialexpressionofphosphory-latedPDK1orPDHE1,HIF1a,orKi-67byimmunohistochemistryonserialsections.
StatisticalAnalysis
Datawereanalyzedusingthetwo-sidedunpairedtorchi-squaretestsusingaGraphPadsoftwarepackage(Prism6.0)forWindows.Correlationparametersbetweenimmunohistochemical(IHC)scoresingliomapatientsandclinico-pathologicvariableswerederivedusingMann-WhitneyUorchi-squaretestsforcontinuousordiscretevariables,respectively,usingGraphPadPrismorMedCalcstatisticalpackages.TheROCcurvesmethodwasusedtotesttheaccuracyofT346-phosphorylatedPDK1tocorrectlydiscriminatebetweengli-omapatientsaccordingtotheirsurvivalstatus(aliveordeadforthedisease)andtogeneratecutoffsforphosphorylatedPDK1IHCscoreusingthenon-arbitrarycriterionderivedfromtheYoudenJstatistic(MedCalcSoftware)asdescribedpreviously(DiCristoforietal.,2015).ThepPDK1IHCscorevaluethatmoreaccuratelydiscriminatedbetweenpatientswhowerealiveordeadwas>25and>40forgliomaorGBMpatients,respectively(Youdencriterion).Gliomapatientswerethensortedintolow-orhigh-expressorcategories,andKaplan-Meiersurvivalcurveswerecomparedusingthelogranktest(MedCalcSoftware).Dataareexpressedasmean±SDormean±SEMofreplicatesfromarepresentativeexperimentoutofatleasttwoorthreeindependentdetermi-nations.Apvalueof<0.05wasconsideredasstatisticallysignificant.ACCESSIONNUMBERS
ThemassspectrometryproteomicsdatahavebeendepositedintheMassIVEdatarepository(http://massive.ucsd.edu)andProteomeXchangedatabase(http://www.proteomexchange.org)withtheaccessionnumbers.MassIVE:MSV000079671andProteomeXchange:PXD004024.SUPPLEMENTALINFORMATION
SupplementalInformationincludesSupplementalExperimentalProcedures,sevenfigures,andthreetablesandcanbefoundwiththisarticleonlineathttp://dx.doi.org/10.1016/j.ccell.2016.07.004.AUTHORCONTRIBUTIONS
Y.C.C.andD.C.A.conceivedtheproject.Y.C.C.,K.G.B.,M.C.C.,J.C.G.,J.H.S.,andS.L.performedexperimentsofAkt-PDK-PDHE1phosphorylation,metabolicreprogramming,modulationofautophagy,mitochondriallocaliza-tion,andcellproliferationandsurvival.I.B.andV.V.performedexperimentswithGBMneurospheresandpatient-derivedGBMsamples.M.L.provided
primary,patient-derivedGBMsamples.L.O.performedexperimentswithanorthotopicmouseGBMmodel.H.Y.T.andD.W.S.performedproteomicsex-periments.B.K.performedmolecularmodelingofthePDK1phosphorylationsite.A.V.K.analyzedbioinformaticsdataofglobalphosphoproteomicsandproteomicsidentificationofPDK1phosphorylation.S.B.,L.R.L.,D.W.S.,andD.C.A.analyzeddata,andY.C.C.,V.V.,andD.C.A.wrotethepaper.ACKNOWLEDGMENTS
ThisworkwassupportedbytheNIHgrantsP01CA140043(D.C.A.andL.R.L.),R01CA78810andCA190027(D.C.A.),R01CA089720(L.R.L.),F32CA177018(M.C.C.),theOfficeoftheAssistantSecretaryofDefenseforHealthAffairsthroughtheProstateCancerResearchProgramunderAwardNo.W81XWH-13-1-0193(D.C.A.),andaChallengeAwardfromtheProstateCancerFounda-tion(PCF)toM.C.C.,L.R.L.,andD.C.A.V.V.issupportedbyanawardfromFondazioneCariplo(2014-1148)andL.O.issupportedbyFP7-INSERTproject(HEALTH-2012-INNOVATION-1,GA305311).I.B.issupportedbyafellowshipfromtheDoctorateSchoolofMolecularandTranslationalMedicineattheUni-versityofMilan,Italy.SupportforCoreFacilitiesutilizedinthisstudywaspro-videdbyCancerCenterSupportGrantCA010815toTheWistarInstitute.Received:August19,2015Revised:March4,2016Accepted:July1,2016Published:August8,2016REFERENCES
Caino,M.C.,Chae,Y.C.,Vaira,V.,Ferrero,S.,Nosotti,M.,Martin,N.M.,Weeraratna,A.,O’Connell,M.,Jernigan,D.,Fatatis,A.,etal.(2013).Metabolicstressregulatescytoskeletaldynamicsandmetastasisofcancercells.J.Clin.Invest.123,2907–2920.
Caino,M.C.,Ghosh,J.C.,Chae,Y.C.,Vaira,V.,Rivadeneira,D.B.,Faversani,A.,Rampini,P.,Kossenkov,A.V.,Aird,K.M.,Zhang,R.,etal.(2015).PI3Kther-apyreprogramsmitochondrialtraffickingtofueltumorcellinvasion.Proc.Natl.Acad.Sci.USA112,8638–8643.
Cox,T.R.,Rumney,R.M.,Schoof,E.M.,Perryman,L.,Hoye,A.M.,Agrawal,A.,Bird,D.,Latif,N.A.,Forrest,H.,Evans,H.R.,etal.(2015).Thehypoxiccancersecretomeinducespre-metastaticbonelesionsthroughlysyloxidase.Nature522,106–110.
DeBock,K.,Georgiadou,M.,Schoors,S.,Kuchnio,A.,Wong,B.W.,Cantelmo,A.R.,Quaegebeur,A.,Ghesquiere,B.,Cauwenberghs,S.,Eelen,G.,etal.(2013).RoleofPFKFB3-drivenglycolysisinvesselsprouting.Cell154,651–663.
Denko,N.C.(2008).Hypoxia,HIF1andglucosemetabolisminthesolidtumour.Nat.Rev.Cancer8,705–713.
DiCristofori,A.,Ferrero,S.,Bertolini,I.,Gaudioso,G.,Russo,M.V.,Berno,V.,Vanini,M.,Locatelli,M.,Zavanone,M.,Rampini,P.,etal.(2015).ThevacuolarH+ATPaseisanoveltherapeutictargetforglioblastoma.Oncotarget6,17514–17531.
Du,J.,Bernasconi,P.,Clauser,K.R.,Mani,D.R.,Finn,S.P.,Beroukhim,R.,Burns,M.,Julian,B.,Peng,X.P.,Hieronymus,H.,etal.(2009).Bead-basedprofilingoftyrosinekinasephosphorylationidentifiesSRCasapotentialtargetforglioblastomatherapy.Nat.Biotechnol.27,77–83.
Dupuy,F.,Tabaries,S.,Andrzejewski,S.,Dong,Z.,Blagih,J.,Annis,M.G.,Omeroglu,A.,Gao,D.,Leung,S.,Amir,E.,etal.(2015).PDK1-dependentmetabolicreprogrammingdictatesmetastaticpotentialinbreastcancer.CellMetab.22,577–589.
Fischer,M.,andRiemer,J.(2013).Themitochondrialdisulfiderelaysystem:rolesinoxidativeproteinfoldingandbeyond.Int.J.CellBiol.2013,742923.Fruman,D.A.,andRommel,C.(2014).PI3Kandcancer:lessons,challengesandopportunities.Nat.Rev.DrugDiscov.13,140–156.
Gardner,L.B.,Li,Q.,Park,M.S.,Flanagan,W.M.,Semenza,G.L.,andDang,C.V.(2001).HypoxiainhibitsG1/Stransitionthroughregulationofp27expres-sion.J.Biol.Chem.276,7919–7926.
Gatenby,R.A.,andGillies,R.J.(2004).Whydocancershavehighaerobicglycolysis?Nat.Rev.Cancer4,891–899.
Ghosh,J.C.,Siegelin,M.D.,Vaira,V.,Faversani,A.,Tavecchio,M.,Chae,Y.C.,Lisanti,S.,Rampini,P.,Giroda,M.,Caino,M.C.,etal.(2015).Adaptivemito-chondrialreprogrammingandresistancetoPI3Ktherapy.J.Natl.CancerInst.107,http://dx.doi.org/10.1093/jnci/dju502.
Gilkes,D.M.,Semenza,G.L.,andWirtz,D.(2014).Hypoxiaandtheextracel-lularmatrix:driversoftumourmetastasis.Nat.Rev.Cancer14,430–439.Gordan,J.D.,Bertout,J.A.,Hu,C.J.,Diehl,J.A.,andSimon,M.C.(2007).HIF-2alphapromoteshypoxiccellproliferationbyenhancingc-myctranscriptionalactivity.CancerCell11,335–347.
Graeber,T.G.,Osmanian,C.,Jacks,T.,Housman,D.E.,Koch,C.J.,Lowe,S.W.,andGiaccia,A.J.(1996).Hypoxia-mediatedselectionofcellswithdimin-ishedapoptoticpotentialinsolidtumours.Nature379,88–91.
Hitosugi,T.,Fan,J.,Chung,T.W.,Lythgoe,K.,Wang,X.,Xie,J.,Ge,Q.,Gu,T.L.,Polakiewicz,R.D.,Roesel,J.L.,etal.(2011).Tyrosinephosphorylationofmitochondrialpyruvatedehydrogenasekinase1isimportantforcancermetabolism.Mol.Cell44,864–877.
Hockel,M.,andVaupel,P.(2001).Tumorhypoxia:definitionsandcurrentclin-ical,biologic,andmolecularaspects.J.Natl.CancerInst.93,266–276.Jazwinski,S.M.(2013).Theretrograderesponse:whenmitochondrialqualitycontrolisnotenough.Biochim.Biophys.Acta1833,400–409.
Jiang,Y.,Li,X.,Yang,W.,Hawke,D.H.,Zheng,Y.,Xia,Y.,Aldape,K.,Wei,C.,Guo,F.,Chen,Y.,etal.(2014).PKM2regulateschromosomesegregationandmitosisprogressionoftumorcells.Mol.Cell53,75–87.
Kaplon,J.,Zheng,L.,Meissl,K.,Chaneton,B.,Selivanov,V.A.,Mackay,G.,vanderBurg,S.H.,Verdegaal,E.M.,Cascante,M.,Shlomi,T.,etal.(2013).Akeyroleformitochondrialgatekeeperpyruvatedehydrogenaseinonco-gene-inducedsenescence.Nature498,109–112.
Keith,B.,Johnson,R.S.,andSimon,M.C.(2012).HIF1alphaandHIF2alpha:siblingrivalryinhypoxictumourgrowthandprogression.Nat.Rev.Cancer12,9–22.
Kim,J.W.,Tchernyshyov,I.,Semenza,G.L.,andDang,C.V.(2006).HIF-1-mediatedexpressionofpyruvatedehydrogenasekinase:ametabolicswitchrequiredforcellularadaptationtohypoxia.CellMetab.3,177–185.
Liang,J.,andMills,G.B.(2013).AMPK:acontextualoncogeneortumorsup-pressor?CancerRes.73,2929–2935.
Liu,X.D.,Yao,J.,Tripathi,D.N.,Ding,Z.,Xu,Y.,Sun,M.,Zhang,J.,Bai,S.,German,P.,Hoang,A.,etal.(2015).AutophagymediatesHIF2alphadegrada-tionandsuppressesrenaltumorigenesis.Oncogene34,2450–2460.LoDico,A.,Valtorta,S.,Martelli,C.,Belloli,S.,Gianelli,U.,Tosi,D.,Bosari,S.,Degrassi,A.,Russo,M.,Raccagni,I.,etal.(2014).ValidationofanengineeredcellmodelforinvitroandinvivoHIF-1alphaevaluationbydifferentimagingmodalities.Mol.ImagingBiol.16,210–223.
Louis,D.N.,Ohgaki,H.,Wiestler,O.D.,Cavenee,W.K.,Burger,P.C.,Jouvet,A.,Scheithauer,B.W.,andKleihues,P.(2007).The2007WHOclassificationoftumoursofthecentralnervoussystem.ActaNeuropathol.114,97–109.Maes,W.,Deroose,C.,Reumers,V.,Krylyshkina,O.,Gijsbers,R.,Baekelandt,V.,Ceuppens,J.,Debyser,Z.,andVanGool,S.W.(2009).Invivobiolumines-cenceimaginginanexperimentalmousemodelfordendriticcellbasedimmu-notherapyagainstmalignantglioma.J.Neurooncol.91,127–139.
Manning,B.D.,andCantley,L.C.(2007).AKT/PKBsignaling:navigatingdown-stream.Cell129,1261–1274.
Mazumdar,J.,O’Brien,W.T.,Johnson,R.S.,LaManna,J.C.,Chavez,J.C.,Klein,P.S.,andSimon,M.C.(2010).O2regulatesstemcellsthroughWnt/beta-cateninsignalling.Nat.CellBiol.12,1007–1013.
McFate,T.,Mohyeldin,A.,Lu,H.,Thakar,J.,Henriques,J.,Halim,N.D.,Wu,H.,Schell,M.J.,Tsang,T.M.,Teahan,O.,etal.(2008).Pyruvatedehydroge-nasecomplexactivitycontrolsmetabolicandmalignantphenotypeincancercells.J.Biol.Chem.283,22700–22708.
Michelakis,E.D.,Sutendra,G.,Dromparis,P.,Webster,L.,Haromy,A.,Niven,E.,Maguire,C.,Gammer,T.L.,Mackey,J.R.,Fulton,D.,etal.(2010).Metabolicmodulationofglioblastomawithdichloroacetate.Sci.Transl.Med.2,31ra34.
CancerCell30,257–272,August8,2016271
Nakahira,K.,Kim,H.P.,Geng,X.H.,Nakao,A.,Wang,X.,Murase,N.,Drain,P.F.,Wang,X.,Sasidhar,M.,Nabel,E.G.,etal.(2006).CarbonmonoxidedifferentiallyinhibitsTLRsignalingpathwaysbyregulatingROS-inducedtraf-fickingofTLRstolipidrafts.J.Exp.Med.203,2377–2389.
Papandreou,I.,Cairns,R.A.,Fontana,L.,Lim,A.L.,andDenko,N.C.(2006).HIF-1mediatesadaptationtohypoxiabyactivelydownregulatingmitochon-drialoxygenconsumption.CellMetab.3,187–197.
Patel,M.S.,Nemeria,N.S.,Furey,W.,andJordan,F.(2014).Thepyruvatede-hydrogenasecomplexes:structure-basedfunctionandregulation.J.Biol.Chem.289,16615–16623.
Ravi,R.,Mookerjee,B.,Bhujwalla,Z.M.,Sutter,C.H.,Artemov,D.,Zeng,Q.,Dillehay,L.E.,Madan,A.,Semenza,G.L.,andBedi,A.(2000).Regulationoftu-morangiogenesisbyp53-induceddegradationofhypoxia-induciblefactor1alpha.GenesDev.14,34–44.
Roberts,D.J.,Tan-Sah,V.P.,Smith,J.M.,andMiyamoto,S.(2013).Aktphos-phorylatesHK-IIatThr-473andincreasesmitochondrialHK-IIassociationtoprotectcardiomyocytes.J.Biol.Chem.288,23798–23806.
Santi,S.A.,andLee,H.(2010).TheAktisoformsarepresentatdistinctsubcel-lularlocations.Am.J.Physiol.CellPhysiol.298,C580–C591.
272CancerCell30,257–272,August8,2016
Semenza,G.L.(2013).HIF-1mediatesmetabolicresponsestointratumoralhypoxiaandoncogenicmutations.J.Clin.Invest.123,3664–3671.
Tredan,O.,Galmarini,C.M.,Patel,K.,andTannock,I.F.(2007).Drugresis-tanceandthesolidtumormicroenvironment.J.Natl.CancerInst.99,1441–1454.
White,E.(2012).Deconvolutingthecontext-dependentroleforautophagyincancer.Nat.Rev.Cancer12,401–410.
Wick,W.,Weller,M.,vandenBent,M.,Sanson,M.,Weiler,M.,vonDeimling,A.,Plass,C.,Hegi,M.,Platten,M.,andReifenberger,G.(2014).MGMTtesting–thechallengesforbiomarker-basedgliomatreatment.Nat.Rev.Neurol.10,372–385.
Young,J.C.,Hoogenraad,N.J.,andHartl,F.U.(2003).MolecularchaperonesHsp90andHsp70deliverpreproteinstothemitochondrialimportreceptorTom70.Cell112,41–50.
Zhang,S.L.,Hu,X.,Zhang,W.,Yao,H.,andTam,K.Y.(2015).Developmentofpyruvatedehydrogenasekinaseinhibitorsinmedicinalchemistrywithpartic-ularemphasisasanticanceragents.DrugDiscov.Today20,1112–1119.
因篇幅问题不能全部显示，请点此查看更多更全内容