Metabolic reprogramming induced by ketone bodies diminishes pancreatic cancer cachexia - PDF document

Acorrection to thisarticlehas beenpublished:http://www.cancerandmetabolism/content/2/1/22 RESEARCHOpenAccessMetabolicreprogramminginducedbyketonebodiesdiminishespancreaticcancercachexiaSurendraKShukla,TeklabGebregiworgis,VineePurohit,NinaVChaika,VenugopalGundaPrakashRadhakrishnan,KamiyaMehla,IraklisIPipinos,RobertPowers,FangYuandPankajKSinghAberrantenergymetabolismisahallmarkofcancer.Tofulfilltheincreasedenergyrequirements,tumorcellssecretecytokines/factorsinducingmuscleandfatdegradationincancerpatients,aconditionknownascancercachexia.Itaccountsfornearly20%ofallcancer-relateddeaths.However,themechanisticbasisofcancercachexiaandtherapiestargetingcancercachexiathusfarremainelusive.Aketogenicdiet,ahigh-fatandlow-carbohydratedietthatelevatescirculatinglevelsofketonebodies(i.e. *Correspondence:pankaj.singh@unmc.eduTheEppleyInstituteforResearchinCancerandAlliedDiseases,UniversityofNebraskaMedicalCenter,Omaha,NE68198,USADepartmentofPathologyandMicrobiology,UniversityofNebraskaMedicalCenter,Omaha,NE68198,USAFulllistofauthorinformationisavailableattheendofthearticle Cancer & Metabolism ©2014Shuklaetal.;licenseeBioMedCentralLtd.ThisisanOpenAccessarticledistributedunderthetermsoftheCreativeCommonsAttributionLicense(http://creativecommons.org/licenses/by/2.0),whichpermitsunrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycredited.TheCreativeCommonsPublicDomainDedicationwaiver(http://creativecommons.org/publicdomain/zero/1.0/)appliestothedatamadeavailableinthisarticle,unlessotherwisestated.etal.Cancer&Metabolismhttp://www.cancerandmetabolism.com/content/2/1/18 accountforthecachexiasyndromeandresultinapoorresponsetochemotherapy,fatigue,andareducedqualityoflifeforcancerpatients[5].Cancercellsexhibitreprogrammingofseveralmetabolicpathwaysalongwithmultiplegenetic,epigenetic,andgrowthsignalingalterations[6,7].Mostcancercellsdem-onstrateanincreaseinglucoseuptake,ahigherrateofglycolysis,andanincreaseinlactatesecretiondespitethepresenceofoxygen,aphenomenonknownastheWar-burgeffect[8].Aerobicglycolysisplaysanimportantroleinrapidcellulargrowthasitprovidesseveralintermedi-atesrequiredforbiomasssynthesisbyroutingthecarbonfluxthroughthepentosephosphatepathway[9].Thein-creasedconversionofpyruvateintolactatebyaerobicgly-colysisleadstoacidosisintumormicroenvironmentsthatfacilitatesinvasionandmetastasisofcancercells[10].Aer-obicglycolysisisalsoanenergy-inefficientprocessrequir-inglargeamountsofglucose.Correspondingly,tumorcellsserveasaglucosesink[11].Additionally,lactatepro-ducedfromtumorcellspassestotheliverandgetscon-vertedtoglucosebymeansoftheCoricycle,anotherenergy-inefficientprocess[9].Alongwithglucoseuptakeandenhancedaerobicglycolysis,cancerpatientsalsopresentglucoseintoleranceandincreasedhepaticglucoseproduction[12].Anincreasedrequirementforglucosemightbethecriticalstimulusneededforenhancedhepaticglucoseproduction.Tumorcellsalsohavealterationsinthemetabolismofglutamine,anitrogensourceandargu-ablythemostsignificantmetaboliteprecursorfortumorcellsafterglucose[13].Aketogenicdietisahigh-fatandlow-carbohydratedietthatleadstoelevatedcirculatinglevelsofketonebodies(i.e.,acetoacetate,-hydroxybutyrate,andacet-one)andanalternativeenergysource[14].Ketogenicdietspossessanticonvulsantandantiinflammatoryac-tivities[15,16].Ithasalsobeenproposedthataketo-genicdiettreatmentresultsinsystemicmetabolicchangeslikeincreasedglucosetolerance,reducedfattyacidsynthesis,andweightloss[17].Keepinginviewthesignificantroleofinflammationandmetabolicalter-ationsincancer,aketogenicdietmayprovideaneffi-cienttherapeuticstrategy.Furthermore,mostcancercellslackkeymitochondrialenzymestometabolizeke-tonebodiesandgenerateATP,whilemyocytesandothertissues,includingthebrain,stillretainthisability[18].Hence,aketogenicdietmayactagainstthecancer-inducedcachexiawhilecausingminimalsideef-fectsaspreviouslyithasbeenshownthata27-mMke-tonebodyconcentrationcanbeachievedsafelywithoutgivingrisetoclinicalacidosis[19,20].Inthepresentstudy,wehaveevaluatedanticancerousandanticachec-ticpropertiesofketonebodiesincellcultureconditions,aswellastheeffectofaketogenicdietontumorburdenandcachexiainanimalmodelsFurthermore,ourstudiesestablishaketonebody-inducedmetabolomicreprogrammingasthemechanismofactionofaketo-genicdietagainstcancerandcancer-inducedcachexia.CellsandreagentsThehumanpancreaticcancercelllineCapan1,mousemyoblastC2C12,andmouseembryofibroblast(preadi-pocyte)3T3L1wereobtainedfromAmericanTypeCul-tureCollection(Manassas,VA,USA).S2-013isaclonedsublineofahumanpancreatictumorcellline(SUIT-2)derivedfromalivermetastasis[21].AllthecelllineswereculturedinDulbeccosmodifiedEaglesmedium(DMEM)supplementedwith10%fetalbovineserum,penicillin(100mg/mL),andstreptomycin(100mg/mL)andincubatedat37°Cinahumidifiedchamberwith5%.Sodium-3-hydroxybutyrate,lithiumacetoacetate,dihydroethidium(DHE),3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazoliumbromide(MTT),BCPCF,andhydroxybutyricacidwerepurchasedfromSigmaChe-micals(Sigma-Aldrich,St.Louis,MO,USA).Cellviabilityandcaspase3/7activityassayCellviabilitywasdeterminedbyperformingMTTassay.Capan1andS2-013cells(5×10cellsperwell)wereseededin96-wellplatesfor12handthentreatedwithdifferentconcentrationsofsodium-3-hydroxybutyrateorlithiumacetoacetatefor72h.Aftertreatment,cellswereincubatedwithMTTreagentfor2h;theresultantfor-mazancrystalsweredissolvedindimethylsulfoxideandtheabsorbancewasrecordedat590nm.Untreatedcellswereutilizedasacontrolfortheviabilityassays.Caspase3/7activitywasdeterminedbyutilizingaPromegaCaspase-Glokit(Madison,WI,USA).Capan1andS2-013cells(0.6×10cellsperwell)wereseededin6-wellplatesfor12handthentreatedwithdifferentconcen-trationsofsodium-3-hydroxybutyrateandlithiumacet-oacetatefor48h.Caspase3/7activitywasthendeterminedasperthemanufacturersprotocol.GlucoseandglutamineuptakeassayTodetermineglucoseuptake,Capan1andS2-013cells(5×10cellsperwell)wereseededin24-wellplates.After12h,cellsweretreatedwithmultipleconcentrationsofsodium-3-hydroxybutyrateandlithiumacetoacetatefor24h.Aftertreatment,cellswerestarvedforglucosefor2handthenincubatedfor20minwith1Ci[H]-2-deoxyglucose(DG)foraglucoseuptakeassay.Cellswerewashedwithphosphate-bufferedsaline(PBS)andlysedwith1%sodiumdodecylsulfate(SDS).Thelysateswerethensubjectedto[H]countingbyutilizingascintillationcounter.Scintillationcountsfromcellstreatedwithlabeledandexcessunlabeled2-DGwereutilizedascon-trolsforbaselinecorrection.Theresultswerenormalizedetal.Cancer&MetabolismPage2of19http://www.cancerandmetabolism.com/content/2/1/18 tothecellcounts.Fordeterminingglutamineuptake,Capan1andS2-013cells(5×10cellsperwell)wereseededin24-wellplates.After12h,cellsweretreatedwithsolventcontrol,multipleconcentrationsofsodium-3-hydroxybutyrate,orlithiumacetoacetatefor24h.Posttreatment,cellswerestarvedforglutaminefor2handthenincubatedfor3minwith1CitritiatedGlutamine,L-[3,4-H(N)].CellswerewashedwithPBSandlysedin1%SDS.Thelysateswereusedfor[H]countingbyutilizingascin-tillationcounter.Scintillationcountsfromcellstreatedwithlabeledandexcessunlabeledglutaminewereutilizedascontrolsforbaselinecorrection.Theresultswerenormal-izedtothecellcounts.LactatereleaseassayCapan1andS2-013cells(5×10cellsperwell)wereseededin24-wellplates.After12h,cellsweretreatedwithindicatedconcentrationsofsodium-3-hydroxybutyrateorlithiumacetoacetatefor24h.Theculturesupernatantswerethenutilizedfordetermininglactaterelease.TheassaywasperformedbyutilizingaLactateAssayKit(EtonBioscienceInc.,SanDiego,CA,USA),asperthemanufac-turersprotocol.ATPassayTotalATPlevelincellswasdeterminedbyusinganATPassaykit(Roche,Indianapolis,IN,USA).After12h,cellsweretreatedwithdifferentconcentrationsofsodium-3-hydroxybutyrateandlithiumacetoacetatefor24handATPlevelwasdeterminedasperthemanufacturerprotocol.ATPlevelwasnormalizedwithtotalproteinconcentration.ReactiveoxygenspeciesassayReactiveoxygenspecieslevelwasdeterminedbyusingoxidation-sensitivefluorescentdyeDHE.Capan1andS2-013cells(0.1×10cellsperwell)wereseededin12-wellplatesonglasscoverslips.After12h,cellsweretreatedwithsolventcontrolorindicateddosesofsodium-3-hydroxybutyrateandlithiumacetoacetatefor24h.Controlandtreatedcellswereincubatedat37°Cin2.5mMDHEcontainingDMEM.Afterincubation,cellswerewashedwithcoldPBSandfixedwithHistoChoice®fixative(Sigma-Aldrich)for15minatroomtemperature.CellswerewashedwithPBSandthecoverslipsweremountedontoglassslidesusingreal-mount.FluorescenceintensitypercellwasdeterminedbyscanningwithZeissAxiovert200Mmicroscope(Oberkochen,Germany)andanalyzingtheimageswithSlideBook5.5software(Intelli-gentImagingInnovations,Inc.,Denver,CO,USA).GeneexpressionanalysisbyqRT-PCRTotalRNAwasisolatedbyutilizingRNeasycolumns(Qiagen,Venlo,TheNetherlands)asperthemanufacturerprotocol.TotalRNA(5g)wasreversetranscribedbyutilizingVerso-cDNAsynthesiskit(ThermoScientific,Pittsburgh,PA,USA)accordingtothemanufacturerguidelines.Quantitativereversetranscriptionpolymer-asechainreaction(qRT-PCR)wasperformedwithgene-specificprimersat95°Cfor10sand60°Cfor60s(40cycles)in10Lreactionmixcontaining3cDNA,2Lprimers,and5LSYBRGreenMasterMix(AppliedBiosystems,GrandIsland,NY,USA)usinganABI7500thermocycler.Beta-actinwasutilizedasanin-ternalcontrol.ThesequenceofdifferentsetsofprimersusedinthestudyisgiveninAdditionalfile1.Quantifi-cationwasperformedwiththeCtmethod[22].ImmunoblottingForimmunoblotting,cellswerewashedtwicewithcoldPBSandlysedinRIPAlysisbufferbyincubatingat4°Crotatoryshakerfor30min.Celldebriswasremovedbycentrifugationat13,000rpmfor10minandthesuper-natantwascollected.ProteincontentwasmeasuredbyperformingBradfordassay.Westernblottingwasper-formedasdescribedpreviously[23].ThemembraneswereprobedwithprimaryantibodyagainstGLUT1(Abcam,Cambridge,UK),c-Myc-9E10(SantaCruzBio-technology,Dallas,TX,USA),HKII(CellSignalingTechnology,Beverly,MA,USA),andHSP90(SantaCruzBiotechnology).Luciferaseassayc-Myc-promoter(del1)-luciferasereporterconstructwasobtainedfromAddgene(Cambridge,MA,USA)[24].Cellsweretransfectedwith1gofplasmid,and16hposttransfection,cellsweretreatedwithdifferentketonebodiesfor24h.AsyntheticRenillaluciferasereporterpRL-TKwasutilizedasatransfectioncontrol.LuciferaseactivitywasdeterminedbyutilizingDual-LuciferaseRe-porterAssaySystem(Promega).ChromatinimmunoprecipitationForchromatinimmunoprecipitation,cellsweretreatedwith20mMsodium-3-hydroxybutyrateandlithiumacetoacetatefor24halongwithsolventcontrol.Chro-matinimmunoprecipitationwasperformedbyutilizingc-Mycantibody(9E10)asdescribedpreviously[25].MouseIgGwasutilizedasacontrol.qPCRdatawerenormalizedtoagenomicregionlocatedwithinGUSBgeneandrepresentedasfoldenrichmentrelativetotheIgGcontrol.PrimersequencesusedforqPCRamplifica-tionaredescribedinAdditionalfile1.TumorgrowthmeasurementCongenitallyathymicfemalenudemice(NCr-nu/nu)werepurchasedfromtheNationalCancerInstitute.Miceweretreatedaspertheguidelinesofourinstitutionalanimalcareetal.Cancer&MetabolismPage3of19http://www.cancerandmetabolism.com/content/2/1/18 andusecommittee(IACUC).S2-013cells(5×10)wereusedfororthotopicinjectionsintothepancreasofnudemice.After7daysofimplantation,miceweredividedingroupsofnineanimalseachandfedadlibitumwithanor-maldietoraketogenicdiet(compositiongiveninTableS2inAdditionalfile1).After3weeksoftreatment,miceweresacrificedandtumorweight,tumorvolume,muscleweight,carcassweight,etc.wererecorded.Tumortissueandotherorganswereflashfrozeninliquidnitrogenforfurtherana-lysis.AnimalprotocolswereinaccordancewiththeNIHGuidefortheCareandUseofLaboratoryAnimalsandwereapprovedbytheUniversityofNebraskaMedicalCen-terAnimalCareandUseCommittee.ImmunohistochemistryImmunohistochemistrywasperformedasdescribedprevi-ously[26].Ki67(ThermoFisherScientific,Waltham,MA,USA),c-Myc(Epitomics,Burlingame,CA,USA),andCleavedCaspase3(CellSignalingTechnology)primaryantibodieswereutilized.Thestainedsectionswereimagedat×20underanuprightmicroscopeandrepresentativeim-ageswerecapturedandpresented.MetaboliteextractionandNMRsamplepreparationAfterconfirmingtheconfluenceofthecells,themediawasaspiratedandthecellswerewashedtwicewith1×phosphatebuffertoremoveremnantsofthemediabeforelysingthecells.Thecellswerethencoldshockedwith1mLofcryogenicallycold80%methanol/watermixture.Theplateswiththe80%methanol/waterwereincubatedina80°Cfreezerforatleast15min.ThecellsfromthecoldplateswerescrapedwithacellscraperandpipettedintoanEppendorftubeandcentrifugedat13,000rpmfor5min.Thesupernatantwascollectedand250LofMilli-Qwater(Millipore,Billerica,MA,USA)wasaddedtotheremainingcelldebrisforre-extraction.Aftermixingthecelldebriswiththewaterbypipetting,thesamplewasagaincentrifugedat13,000rpmfor5min.Thenewsuper-natantwascombinedwiththepreviouslycollectedsuper-natant.Finally,thesamplewasdriedusingspeedvacuumevaporator(SpeedVac®Plus,Savant,ThermoScientific,Waltham,MA)toevaporatethemethanolandsubjectedtofreezedrying(Labconco,KansasCity,MO)tolyophilizethewaterconsecutively.ThedriedsamplewasmadereadyforanNMRexperimentbydissolvingin600Lof50mMphosphatebufferin99.8%DO(Isotec,St.Louis,MO)atpH7.2(uncorrected)with50M3-(tetramethysilane)propionicacid-2,2,3,3-d(TMSP)(500Mfor2DHSQC)forspectralreferencing.NMRexperimentanddataanalysisTheNMRspectrawereacquiredonaBrukerAVANCEDRX500MHzspectrometerequippedwith5mmtriple-resonancecryogenicprobe(C,andN)withaZ-axisgradient.TheNMRdatacollectionwasautomatedusingaBACS-120samplechanger,ATM(automatictun-ingandmatching),andBrukerIconNMRsoftware.Theone-dimensional(1D)protonnuclearmagneticresonanceHNMR)datawasacquiredusinganexcitationsculptingpulsesequencetoremovethesolventpeakandmaintainaflatbaseline[27].Thespectrawerecollectedat300Kwith32Kdatapoints,128scans,16dummyscans,andaspec-tralwidthof5,483Hz.OurMVAPACKsoftware(http://bionmr.unl.edu/mvapack.php)[28]wasusedtoprocessthe1DHNMRspectra.TherawNMRdatawasonlyFouriertransformedandautomaticallyphased.Theresult-ingNMRspectrumwasbinnedusinganadaptiveintelli-gentbinningalgorithmthatautomaticallyadjustsbinsizestoavoidsplittingNMRresonancesbetweenmultiplebins[29].ThespectralregionbeforetheTMSPwasusedasatrainingsettoremovethenoisefromthedatausingthemethodstatedbyHalouskaetal.[30].Thespectrawerethennormalizedusingstandardnormalvariate(SNV)andscaledusingParetoscaling.Theprocesseddatawasuti-lizedtogenerateplotsofprincipalcomponentanalysis(PCA)andorthogonalprojectionstolatentstructuresdiscriminantanalysis(OPLS-DA)scoresandbackscaledloadings(Additionalfile2)usingourMVAPACKsoft-ware[28].Metaboliteidentificationfromthe1DNMRspectrawasaccomplishedusingtheChenomxNMRSuite7.6(http://www.chenomx.com/)andthebackscaledloadings.The2DChetero-nuclearsinglequantumcoher-ence(HSQC)NMRspectrawerecollectedat300Kwith64scans,16dummyscans,anda1.5-srelaxationdelay.Thespectrawerecollectedwith2Kdatapointsandaspectrumwidthof4,735Hzinthedirectdimensionand64datapointsandaspectrumwidthof17,607Hzintheindirectdimension.The2DCHSQCNMRspectrawereprocessedusingNMRPipe(NIH,Bethesda,Mary-land)[31]andanalyzedusingNMRViewJVersion8.0.3.Peakintensitieswerenormalizedbytheaveragepeakin-tensityforagivenspectrumandthenassignedtoame-taboliteusingchemicalshiftreferencesfromtheHumanMetabolomicsDatabase[32],MadisonMetabolomicsConsortiumDatabase[33],andPlatformforRIKENMetabolomics[34].Chemicalshifterrorsof0.08and0.25ppmfortheHandCchemicalshifts,respect-ively,wereusedtomatchtheexperimentalchemicalshiftswiththedatabases.Inadditiontochemicalshifts,peaksplittingpatternsandpeakshapeswerealsousedtoverifymetaboliteassignments.The2DCHSQCNMRexperimentisamorereli-ableapproachformetaboliteidentificationbecauseofthesignificantlyhighersignaldispersion,andthecorrelationbetweenHandCchemicalshiftsforeachC-Hpairinamolecule[35].Moreimportantly,the2DCHSQCexperimentsimplifiestheanalysisofthemetabolomeetal.Cancer&MetabolismPage4of19http://www.cancerandmetabolism.com/content/2/1/18 becauseonlycompoundscontainingaC-carbonderivedfromthe-glucoseaddedtothemediawillbedetected.Thus,using-glucosewillhighlightmetabolitechangesassociatedwiththeglycolyticfluxintumorcells.Thisavoidsthechallengewiththe1DHNMRexperimentswherethespectraweredominatedbycatabolicproductsofketonebodies.Thus,theidentificationofmetabolitesfromthe2DCHSQCexperimentsismorereliableandpertinenttotheanalysisofmetabolicchangesresult-ingfromketonebodyeffectsonpancreaticcancercachexia.Eachmetabolitepeakfromthetwosets(controlandketonebody-treated)oftriplicate2DCHSQCNMRspectrawerefurthernormalizedbyusingthemaximumpeakintensityforthemetaboliteandthenscaledfrom0to100.ThepeakintensitiesforeachmetabolitewerethenaveragedandcomparedbetweenthecontrolandtreatedgroupsusingaStudenttest.Metaboliteswithvalue0.1wereusedtogenerateaheatmapusingtheRstatisticalpackage[36].Arelativechangeinpeakintensityimpliesacorrespondingmetaboliteconcentra-tionchange.Absoluteconcentrationsarenotmeasur-ablefromthe2DCHSQCbecauseotherfactors,suchascouplingconstants,relaxation,anddynamics,alsocontributetopeakintensities.Measurementofbloodglucoseand-hydroxybutyrateconcentrationBloodglucoselevelofmicewasmeasuredafter16hofstarvationbyutilizingContourUSBbloodglucosemeter(BayerHealthCare,Mishawaka,Japan),asperthemanu-facturersprotocol.Bloodketonelevelwasmeasuredbyutilizingabloodglucoseandketonemonitor(NovaBiomedical,Waltham,MA,USA)asperthemanufac-turersprotocol.Theconcentrationof-hydroxybutyratewasmeasuredbycomparingtheTMSP-normalizedme-thylpeaksofHNMRcollectedforsixanimals.C2C12and3T3L1differentiationandconditionedmediumpreparationC2C12mousemyoblastsweregrowninDMEMwith10%FBS.Toinducedifferentiation,cellswereswitchedto2%horseserumand10g/mLinsulin-containingDMEMandgrownfor72h.3T3L1mouseembryofibroblastswereculturedinDMEMwith10%FBS.Fordifferentiationof3T3L1preadipocytes,after2daysatconfluency,cellsweretreatedwithdifferentiationmedium:DMEMcon-taining10%FBS,1Mdexamethasone,0.5mMmethyli-sobutylxanthine(IBMX),and1g/mLinsulinfor2days.After2days,differentiationmediumwasreplacedwithDMEMcontaining10%FBS.Mediumwaschangedregu-larlyafter48h.Forconditionedmediumpreparation,cellswereplatedatadensityof50,000cells/cm,andafter12hofseeding,cellswerewashedtwicewithPBSandculturedinserum-freeDMEMforthenext24h.Conditionedmediumwascentrifugedat1,200for10minandfilteredwitha0.2-msyringefilterandusedimmediatelyorstored80°C.MeasurementofintracellularpHCytosolicpHwasmeasuredbyusingfluorescencespec-troscopyusingBCPCF-AMasdescribedbyMarinoetal.al.StatisticalanalysisComparisonsbetweendifferentgroupswereperformedbyusingANOVA(one-way;GraphPadPrismversion4.03)withDunnettsposthoctest.Studenttestwasusedforinvivostudies.Avalueof.05wasconsideredtobesignificant.Ketonebodiesdiminishpancreaticcancercellgrowthandinduceapoptosisinadose-dependentmannerWeinvestigatedtheeffectofketonebodies(sodiumhydro-xybutyrateandlithiumacetoacetate)oncellsurvivalinmultiplepancreaticcancercelllines.Initially,weevaluatedtheeffectofmultipledosesofsodiumhydroxybutyrateandlithiumacetoacetate(1to20mM)onthesurvivalofCapan1andS2-013pancreaticcancercellsbyperformingMTTassays.Ketonebodieswereobservedtoinhibitcellsurvivalinadose-dependentmanner.Ketonebodiessignifi-cantlyinhibitedcellgrowthatconcentrationsof10and20mMaftera72-htreatment(Figure1A,B).Toevaluateiftheketonebody-inducedsurvivaldiminutionisspecifictocancercells,wesubjectedimmortalized,non-transformedpancreaticepithelialcelllinesHPNEandRAPANtoincu-bationwithsodiumhydroxybutyrateandlithiumacetoace-tate.Weobservednosignificanteffectonsurvivalofthesecellsundertreatmentwithketonebodies(Additionalfile3).Similarly,toruleoutthepossibilityoforganicacidandlith-iumionbeingresponsibleforreductionincancercellsur-vival,ratherthanketonebodiesthemselves,weevaluatedthesurvivalofS2-013andCapan1undertreatmentwithhydroxybutyricacidandlithiumchloride.WeobservednosignificanteffectonS2-013andCapan1cellsurvivalafter72hoftreatment(Additionalfile4).Furthermore,weinvestigatedtheeffectofsodiumhydroxybutyrate,lithiumacetoacetate,and-hydroxybutyrateonintracellularpH.After6hoftreatmentwithdifferentdosesofsodiumhydroxybutyrate,lithiumacetoacetate,and-hydroxybuty-rateonS2-013,Capan1,HPNE,andRAPANcells,weobservedsignificantdecreaseinintracellularpH(0.2to0.4units)inallthecelllines(Additionalfile5).Althoughintra-cellularacidificationupontreatmentwithketonebodiesmightcontributetotheantiproliferativeeffectsofketonebodies,itisnottheprimarycauseofcellsdeathaswedidnotobservesignificantcelldeathupontreatmentwithetal.Cancer&MetabolismPage5of19http://www.cancerandmetabolism.com/content/2/1/18 isomerofhydroxybutyratethat causedsimilarintracellular pHchangeasdidtheotherketonebodies.Furthermore,we observedasimilareffectonintracellularpHinnon- transformedpancreaticepithelialcelllinesHPNEand RAPANaftertreatmentwithsodiumhydroxybutyrate,lith- iumacetoacetate,and S -isomerofhydroxybutyratebutno significantcelldeath,asmentionedearlier.Bright-fieldmi- croscopyimagesofthecellstreatedwith10and20mMof NaHBandLiAcAcarepresentedinFigure1C,D.Further- more,weinvestigatedtheeffectofketonebodiesonfive pancreaticcancercelllinesandobservedthatcellgrowth issignificantlyinhibitedinadose-dependentfashion (Figure1E).Wealsoinvestigatedtheeffectofketonebody treatmentoncaspaseactivityinCapan1andS2-013cell lines.Caspase3/7activityincreasedupontreatmentofthe pancreaticcancercellswithketonebodiesinadose- dependentmanner(Figure1F). Ketonebodiescausemetabolicalterationsinpancreatic cancercells Previousstudiesindicatethataketogenicdietinduces metabolicalterationsinmice[17].Hence,wenextinvesti- gatedtheeffectofketonebodiesonpancreaticcancercell metabolism.Becausecancercellsdemonstrateanincrease inglycolysis[38],weexaminedtheeffectofketonebodies onglucoseuptakeandlactatereleaseinCapan1andS2- 013cells.TreatmentofCapan1andS2-013cellswith ketonebodiesresultedinadecreaseinglucoseuptake (Figure2A,B)andreleaseoflactate(Figure2C,D)ina dose-dependentmanner.Sinceglutaminealsosupports pancreaticcancercellgrowth[7],wealsoevaluatedtheef- fectofketonebodiesonglutamineuptake.Ourresults indicateareduceduptakeofglutaminebyCapan1and S2-013pancreaticcancercellsundertreatmentwith ketonebodies(Figure2E,F).Furthermore,weobserveda Figure1 Ketonebodiesinhibitgrowthandinduceapoptosisinpancreaticcancercelllines. Capan1 (A) andS2-013 (B) cellsweretreated withdifferentconcentrationsofsodium-3-hydroxybutyrate(NaHB)andlithiumacetoacetate(LiAcAc)for72h,andcellviabilitywasdetermined byMTTassay. Bar representspercentviabilityunderindicatedtreatmentsrelativetotreatmentwithsolventcontrol.Representativebright-field imagesofCapan1 (C) andS2-013 (D) cellsundertreatmentwith10-and20-mMconcentrationsofNaHBandLiAcAcfor72h. (E) Multiple pancreaticcancercelllinesweretreatedwith10-and20-mMconcentrationsofNaHBandLiAcAcfor72h,andrelativecellviabilitydetermined byMTTassayisplottedinthe barcharts . (F) Capan1andS2-013cellstreatedwith10-and20-mMconcentrationsofsodium-3-hydroxybutyrate andlithiumacetoacetatefor48handtherelativecaspase3/7activityareplotted.Valuesrepresentedaremean±SEM.* P 0.05;** P 0.01. Shukla etal.Cancer&Metabolism 2014, 2 :18 Page6of19 http://www.cancerandmetabolism.com/content/2/1/18 Figure2(Seelegendonnextpage.)etal.Cancer&MetabolismPage7of19http://www.cancerandmetabolism.com/content/2/1/18 reductioninintracellularATPlevelsupontreatmentwithketonebodies(Figure2G,H).Reactiveoxygenspecies(ROS;anaturalby-productofoxidativemetabolism)levelswerealsodownregulatedupontreatmentwithketonebodies(Figure2I,J).Overall,ourresultsindicatethatke-tonebodiesdiminishtheoverallenergetichealthofpan-creaticcancercellsbyreducingglucoseuptake,lactaterelease,glutamineuptake,cellularATPcontent,andROSlevels.KetonebodiesdiminishtheexpressionofglycolyticTodeterminethemechanismofreducedglucoseuptakeandlactatereleaseupontreatmentofpancreaticcancercellswithketonebodies,wenextexaminedifketonebody-mediatedchangeswereduetoalterationsingeneexpressionlevels.Weevaluatedtheexpressionofgenesencodingglucosetransporter-1(GLUT1)andglycolyticenzymeshexokinaseII(HKII)andlactatedehydrogen-aseA(LDHA)inketonebody-treatedandcontrolcells.InbothCapan1andS2-013cells,weobservedareducedexpressionofGLUT1andLDHAbutnosignificantchangeinexpression(Figure3A,B)upontreat-mentwithketonebodies.Furthermore,weanalyzedtheproteinexpressionlevelsofGLUT1andHKIIinS2-013andCapan1cellsundertreatmentwithketonebodiesorcontrol.OurresultsindicateareducedexpressionofGLUT1butalmostnochangeinHKIIlevelsincellstreatedwithketonebodies(Figure3C,D).Toeliminatethepossibilityofnon-specificeffectsoforganicacidandlithiumionongeneexpressionandmetabolism,wede-terminedGLUT1,HKII,andLDHAgeneexpressionandglucoseuptakeinS2-013cellsaftertreatmentwith-hydroxybutyricacidandlithiumchloride.Wefoundanegligibleeffectoftheseagentsonmetabolicgeneex-pressionandglucoseuptake(Additionalfile6).Ketonebodiesdiminishc-MycexpressionandactivitySinceweobservedadownregulationofglucoseandglu-tamineuptakeupontreatmentwithketonebodiesandacorrespondingdecreaseinglycolyticgeneexpression,wenextinvestigatedthemolecularmechanismbehindthisde-crease.Specifically,weevaluatedtheeffectofketonebodiesontheactivityandexpressionofc-Myc,acommonregula-torofbothglucoseandglutaminemetabolism[39].UsingtheTFsearchprogram,thepromotersequencesofGLUT1LDHAwerebothfoundtocontainc-Mycbindingsites.Chromatinimmunoprecipitationassayswerethenper-formedtoevaluatethec-MycoccupancyattheGLUT1LDHApromoters.Ourresultsindicateareducedoccu-pancyofc-MyconGLUT1LDHApromotersforS2-013cellstreatedwithketonebodies(Figure4A,B).Fur-thermore,weperformedreal-timePCRtodeterminethec-MyctranscriptlevelsinS2-013andCapan1cellsaftertreatmentwithketonebodiesorcontrol.Weobservedsig-nificantreductioninc-Mycexpressionaftertreatmentwithketonebodies(Figure4C,D).Wealsoevaluatedc-Mycpro-teinexpressionaftertreatmentwithketonebodiesandob-servedareductioninc-Myclevelsinadose-dependentmanner(Figure4E,F).Toconfirmifthealterationsinc-Mycexpressionwereduetoalteredc-Myctranscription,weevaluatedthetranscriptionalactivityofc-Mycpromoterbyutilizingc-Mycpromoter-luciferasereporterconstructs[24].Weobservedsignificantlyreducedc-Mycpromoteractivityundertreatmentwithketonebodies(Figure4G).Altogether,ourresultsdemonstratethatthetreatmentofpancreaticcellswithketonebodiesleadstoareducedMycexpressionatmRNAlevelandareducedc-Mycoccu-pancyontheGLUT1LDHApromoters.Ketonebodiesinhibittumorcell-inducedmusclefiberandadipocytedegradationincell-basedassaysMalignantcellshavebeenshowntolackkeymitochon-drialenzymesrequiredformetabolizingketonebodiestoproduceATP,whilemusclecellsretainthiscapacity[18].Ofnote,-hydroxybutyrateimprovesbodyweight,whilereducingproteolysisinmusclecells[40].Hence,wedevelopedacellculture-basedsystemtoevaluatetheeffectofketonebodiesonmusclefibersandadipocyteadiposedepositsundercoculturewithcancercells.ThiswasachievedbydifferentiatingC2C12premyocytesintomyotubesand3T3L1preadipocytesintodifferentiatedadipocytes.TreatmentwithCapan1orS2-013cancercell-conditionedmedium(CCCM)induceddegradationofmyotubesanddepletionofadiposedepositsin3T3L1adipocytes.Treatmentwithketonebodiesdemonstrated (Seefigureonpreviouspage.)Figure2Ketonebodiesinducemetabolicalterationsinpancreaticcancercelllines.andCapan1cellsweretreatedwithdifferentdosesofketonebodiesfor24h,andglucoseuptakewasdeterminedbyperformingH-2DGuptakeassay.representcountsnormalizedwithcellnumberandplottedrelativetocontrol.LactatereleasewasdeterminedbycolorimetricassayusingculturemediumofandCapan1cellstreatedwithdifferentconcentrationsofNaHBandLiAcAcfor24h.Valueswerenormalizedwithtotalcellnumberandrepresentedrelativetocontrols.S2-013andCapan1cellsweretreatedwithindicatedconcentrationsofketonebodiesfor24h,andglutamineuptakewasdeterminedbyperformingtritiatedGlutamine,L-[3,4-H(N)]uptakeassays.Countswerenormalizedwithcellnumberandplottedrelativetocontrol.ATPlevelsinS2-013andCapan1cellspost24-htreatmentwithketonebodiesweredeterminedbyperformingATPbioluminescenceassays.Valueswerenormalizedtototalproteinconcentrationandrepresentedrelativetocontrol.ReactiveoxygenspecieslevelofS2-013andCapan1cellsundertreatmentwithketonebodieswasdeterminedbyutilizingafluorescenceprobe,dihydroethidium(DHE),andfluorescenceintensitynormalizedtocellcountwasplotted.Valuesrepresentedaremean±SEM.*0.05;**0.01.etal.Cancer&MetabolismPage8of19http://www.cancerandmetabolism.com/content/2/1/18 asignificantprotectionofmyotubes(Figure5A,B)and adipocytes(Figure5D,E)againstCCCM.Furthermore, weanalyzedthegeneexpressionlevelsof muscle-specific ringfingerprotein1 ( MuRF1 or Trim63 )and Atrogin ( Fbxo32 ;alsoknownas muscleatrophyF-boxprotein or MAFbx )inmyotubes,and zincalpha-2-glycoprotein1 ( Zag or Azgp1 )andhormone-sensitivelipase( HSL or Lipe )inadipocytestreatedwithCCCMundertreatment withketonebodiesorcontrol.Theseproteinsareupreg- ulatedduringcancer-inducedcachexia[41].Ketone bodieswereshowntosignificantlylower MuRF1 and Atrogin mRNAlevelseveninthepresenceofCCCM (Figure5C).Wealsoobservedasignificantdecreasein Zag and HSL expressionupontreatmentwithketone bodies(Figure5F).Hence,ourresultsindicatetheutility ofketonebodiesininhibitingcancer-inducedcachexia. Ketonebodiesaltercentralcarbonmetabolismin pancreaticcancercells Toevaluatetheeffectofketonebodiesonmetabolite levelsinpancreaticadenocarcinomacells,weperformeda seriesofNMR-basedmetabolomicsstudies.S2-013cells weretreatedwith20mMsodium-3-hydroxybutyratefor 24h,andthenasetof1D 1 HNMRspectrawereacquired forcellextractsofsodium-3-hydroxybutyrate-treatedor controlS2-013cells.OPLS-DAoftheNMRspectraindi- catesthatthemetabolomesfromtheketonebody-treated cellsandthecontrolcellsclusteredintotwoseparate groupsinaOPLS-DAscoreplot,indicatingthatthe treatedcancercellsaremetabolicallydifferentiatedfrom thecontrols(Figure6A).TheOPLS-DAmodelwasvali- datedusingCV-ANOVAyieldinga p valueof8.74×10  6 (Additionalfile2).Thebackscaledloadings(Additional file7)ofthe1D 1 HNMRdataindicateareductionin cellularglutamineandglutamatelevel.Manycancersare dependentonglutaminemetabolismandhaveenhanced glutamineuptake.Asignificantreductionofcholine- containingcompoundsisalsoobservedinketone body-treatedcells.Totalcholine-containingcompound concentrationistypically increasedinmultiplecancer phenotypesandreferredasametabolichallmarkofcan- cer.Itisparticularlynot eworthythatketonebody Figure3 Ketonebodiesrepresstheexpressionofkeyglycolyticenzymes. RelativemRNAexpressionlevelsof GLUT1 , HKII ,and LDHA inCapan1 (A) andS2-013 (B) cellstreatedwith10-and20-mMconcentrationsofNaHBandLiAcAcfor24h.TotalRNAwasisolatedfromNaHB-and LiAcAc-treatedaswellascontrolcells,andrelativemRNAlevelsofdifferentgenesweredeterminedbyperformingqRT-PCR.
-Actinwasutilizedasan internalcontrol.ProteinexpressionofGLUT1andHKIIwasdeterminedbyimmunoblottingthetotalcelllysatesfromS2-013 (C) andCapan1 (D) cells treatedwith10and20mMNaHBandLiAcAcfor48h.
-Tubulinwasutilizedasaninternalcontrol.Valuesshownaremean±SEM.* P .05; ** P .01. Shukla etal.Cancer&Metabolism 2014, 2 :18 Page9of19 http://www.cancerandmetabolism.com/content/2/1/18 treatmentappearstoreversethesetrends[42].Ketone bodytreatmentalsoelevatesmyo-inositol,taurine,and alanine.Myo-inositolanditsderivativeareknownto suppresspancreaticcancer[43]. Tofurtherexploretheimpactofketonebodiesonother metabolicprocess,wethenlabeledtheS2-013metabolome with 13 C 6 -glucoseandcomparedthecellextractsfrom sodium-3-hydroxybutyrate-tr eatedS2-013cellswithcontrol cellsusing2D 1 H- 13 CHSQCNMRexperiments.The2D 1 H- 13 CHSQCspectraidentifiedthe 13 C-labeledmetabolites inthecellextractsderivedfrom 13 C 6 -glucose.Importantly, 13 C 6 -glucosehighlightedmetabol itechangesassociatedwith Figure4 Ketonebodiesreducec-Mycexpressionanditsrecruitmenttoglycolyticgenepromoters. Recruitmentofc-Myconto GLUT1 (A) and LDHA (B) promotersinS2-013cellsundertreatmentwith20mMNaHB,LiAcAc,orcontrolwasconfirmedbyperformingChIPusing anti-c-MycAbandIgGcontrol,followedbyqRT-PCRanalysis.Relativec-MycmRNAlevelsinCapan1 (C) andS2-013 (D) cellstreatedwith10and 20mMNaHB,LiAcAc,orcontrolfor24h.TotalRNAwasisolatedandrelativemRNAlevelof c-Myc wasdeterminedbyqRT-PCR.
-Actinwas utilizedasaninternalcontrol.Capan1 (E) andS2-013 (F) cellsweretreatedwithindicateddosesofketonebodiesfor48h,andc-Mycprotein levelwasdeterminedbyimmunoblottingthewholecelllysates.HSP90wasusedasaninternalcontrol. (G) c-Myc-promoter-fireflyluciferase reporterand Renilla luciferasereporterplasmidsweretransientlytransfectedintoS2-013cells.After16hoftransfection,cellsweretreatedwith solventcontrolorketonebodiesfor24h.Normalizedfireflyto Renilla luciferaseactivityratioisplottedinthe barchart .Valuesrepresentedare mean±SEM.* P 0.05;** P 0.01. Shukla etal.Cancer&Metabolism 2014, 2 :18 Page10of19 http://www.cancerandmetabolism.com/content/2/1/18 glycolyticfluxintumorcells.TheNMRmetabolomics studiesidentifiedmultiplemet abolitesexhibitingstatisti- callysignificantconcentrationchangesduetoketonebody treatment(Figure6B).Theseidentifiedmetaboliteswere thenincorporatedintoanetwo rkusingCytoscape[44]and Metscape[45]bylinkingnearest-neighbormetabolites. Overall,glucose-derivedmetabolitesinvolvedinglycolysis, aminoacidmetabolism,andTCAcyclewerealteredupon treatmentwithketonebodies(Figure6C).The1D 1 Hload- ingsarealsoconsistentwiththisgeneralobservation.Spe- cifically,bothexperimentsidentifychangesinaminoacid metabolism.Again,theketonebodiesarebeingusedasa Figure5 Ketonebodiesinhibittumorcell-conditionedmedium-induceddegradationofmyofibersandadipolysis. DifferentiatedC2C12 cellsweretreatedwithS2-013 (A) andCapan1 (B) cell-conditionedmediumwithorwithoutsolventcontroland10and20mMNaHBandLiAcAc for72h,andbright-fieldimageswererepresentedforindividualtreatments. (C) DifferentiatedC2C12cellswereculturedinCapan1andS2-013 cell-conditionedmediumwithorwithoutketonebodytreatmentfor24h.TotalRNAwasisolatedandrelativemRNAlevelsof MuRF1 and Atrogin weredeterminedbyperformingqRT-PCR.
-Actinwasutilizedasaninternalcontrol.Differentiated3T3L1cellswereculturedinS2-013 (D) and Capan1 (E) cell-conditionedmediumwithorwithoutketonebodytreatmentfor72handstainedwithnilered.Fluorescentandbright-field imagesforindividualtreatmentsarepresented. (F) Differentiated3T3L1cellswereculturedinCapan1andS2-013cell-conditionedmediumwith orwithoutketonebodytreatmentfor24h.TotalRNAwasisolatedandrelativemRNAlevelsof Zag and HSL weredeterminedbyqRT-PCR.
-Actinwasutilizedasaninternalcontrol.Valuesrepresentedaremean±SEM.Allstatisticalanalyseswereconductedwithone-wayANOVAwith Dunnett ’ sposthoctestandCMasthereferencegroup.* P 0.05;** P 0.01. Shukla etal.Cancer&Metabolism 2014, 2 :18 Page11of19 http://www.cancerandmetabolism.com/content/2/1/18 metabolicsubstrateandarereplacingthecellularneed forglucoseandglycolysis.Theimportanceofadetailed analysisofmetabolicchangesishighlightedbyakeyex- ample.The2D 1 H- 13 CHSQCexperimentsindicatethat theconcentrationsofglutamateandglutaminederived fromglucosehavedecreasedsignificantly.Similarly,the 1D 1 Hloadingsindicatethatthe overall concentrations ofglutamateandglutaminehavealsodecreasedupon ketonebodytreatment.Themetabolitesderivedfrom ketonebodiesdonotcontaina 13 C-carbonandarenot detectedinthe2D 1 H- 13 CHSQCexperiment.Corres- pondingly,the1D 1 Hloadingsidentifymajorchanges forunlabeledmetabolitesthatarenotvisibleina 2D 1 H- 13 CHSQCspectrum.Thus,the2D 1 H- 13 C HSQCexperimentsarecomplimentarytothe1D 1 Hex- perimentsandprovideamoredetailedanalysisofaspe- cificsubsetofthemetabolome. Tostudyifketonebodiesaregettingmetabolizedby cancercells,wetreatedS2-013cellswith 13 C 4 -labeledhy- droxylbutyrateandidentifieditscatabolicproductsby usinga2D 1 H- 13 CHSQCNMRexperiment(Additional file8).Becauseofthelownaturalabundanceof 13 C (1.1%),theonlypeaksobservableina2D 1 H- 13 CHSQC spectrummustoriginatefromthe 13 C 4 -labeledhydroxyl butyrate.Asareferencepoint,thecellularextractwas spikedwith500  MofTMSP.Thesinglepeakoriginating fromthenaturalabundantTMSP 13 C-methylgroupsis barelydetectableinthe2D 1 H- 13 CHSQCspectrum.A negativecontrol( datanotshown ),whereS2-013cells arenottreatedwitha 13 C-labeledmetabolite,yieldeda nullspectrum.Thus,thefactthatmultipleintense peaksareobservableinthe2D 1 H- 13 CHSQCspectrum isaclearevidencethattheS2-013cellsuptakehydroxyl butyrate.The 1 Hand 13 Cchemicalshiftsmeasured fromthe2D 1 H- 13 CHSQCspectrumwerethencom- paredagainstreferenceNMRspectraavailablefromthe HumanMetabolomicsDatabase[32],MadisonMetabo- lomicsConsortiumDatabase[33],andPlatformfor RIKENMetabolomics[34]toidentifyother 13 C-labeled metabolitespresentinthecellextract.Weobserved chemicalshiftsconsistentwith3-hydroxybutyrate,the original 13 C-carbonsource,andbetaine, N -acetylgluco- samine,homocarnosine,andsuccinate,whichallmust becatabolicproductsof3-hydroxybutyrate. Inhibitingglycolyticfluxintumorcellspreventscachexia phenotypeincelllinemodels Metabolicalterationsincancercells[46]maycontribute tothesecretionofcachecticcytokinesandmetabolites involvedincancer-inducedcachexia.Basedonourfind- ingsthatketonebodiesmediatedadecreaseintheex- pressionof GLUT1 resultinginametabolicflux,we evaluatedifinhibitingtheglycolyticfluxaltersthecach- ecticpotentialincancercells.WepretreatedS2-013 Figure6 Ketonebodiesmodulatemetabolitelevelsinpancreaticcancercells.(A) OPLS-DAscoreplotgeneratedfrom1D 1 HNMRspectra collectedfromcelllysatesofS2-013cells( redsquare )andS2-013cellstreatedwith20mMNaHB( greendiamond );each point intheOPLS-DAscoreplot representsasingle1D 1 HNMRspectrum. Ellipses enclosethe95%confidenceintervalsestimatedbythesamplemeansandcovariancesofeachclass. Theleave-n-outcrossvalidationyieldedaqualityassessment( Q 2 )valueof0.959and R 2 valueof0.997.TheOPLS-DAmodelwasvalidatedusing CV-ANOVAyieldinga p valueof8.74×10  6 . (B) Heatmapgeneratedfrom2D 1 H- 13 CHSQCNMRspectraldataforS2-013cells.The heatmap representstriplicatemeasurementsofmetaboliteintensitiesrecorded(* P .1;** P 0.05;*** P .001). (C) Metabolicpathwaydepicts 13 Ccarbonflow fromglucosetotheintermediatesofglycolyticpathway,citricacidcycle,aminoacidmetabolism,andnucleotideanalogues.The arrows represent relativeincrease( greenarrowup )ordecrease( redarrowdown )inmetaboliteconcentrationsduetoketonebodytreatment. Shukla etal.Cancer&Metabolism 2014, 2 :18 Page12of19 http://www.cancerandmetabolism.com/content/2/1/18 cellswith3-sodiumhydroxyb utyrate,lithiumacetoace- tate,orglycolyticinhibitor3-bromopyruvicacid(BPA) andthencollectedtheconditionedmediumandevalu- atedthecachecticpotentialbyutilizingC2C12myotube and3T3L1adipocytesystems.Wealsoassayedcondi- tionedmediumfrom GLUT1 knockdownS2-013cells (S2-013-sh GLUT1 )forcachecticpotential.Treatment withconditionedmediumgeneratedfromketonebodies orglycolyticinhibitor-pretreatedS2-013cells,orGLUT1 knockdownS2-013cells,demonstratedasignificantpro- tectionagainstmyotubedegradation(Figure7A,B)and adipocytefatdepletion(Figure7C,D)incomparisonto controls.Furthermore,weanalyzedthegeneexpression levelsof MuRF1 and Atrogin inmyotubesand Zag and HSL inadipocytestreatedwiththeabove-described CCCM.Weobservedasignificantreductionof MuRF1 and Atrogin geneexpressioninC2C12myotubesunder treatmentwithCCCMfromketonebodyorinhibitor- pretreatedcells,orGLUT1knockdowncell-derived conditionedmedia,incomparisontocontrolCCCM- treatedmyotubes(Figure7E).Similarly,weobserved areducedexpressionof Zag and HSL in3T3L1adi- pocytesinthepresenceofpretreatedCCCMincom- parisontothecontrols(Figure7F).Toensurethe inhibitionofglycolysisbyBPAandGLUT1knock- down,weperformedglucoseuptakeassayinS2-013 cellstreatedwithBPAorsolventcontrolaswell asS2-013-sh GLUT1 andS2-013-shScrcells.Weob- servedasignificantreductioninglucoseuptakeunder theseconditions(Additionalfile9).Asweobserved thatketonebodiesactlikemetabolicinhibitors,we furtherexploredifglycolyticproductlactateand alanine,whichcouldbesecretedintothemedium,in CCCMcouldabrogatetheeffectofketonebodytreat- mentoftumorcellsontheexpressionofcachectic markersinadipocytesandmyotubes.Weobservedno significantalterationincachecticmarkerexpression afteraddinglactateoralanine(Additionalfile9), Figure7 Pretreatmentoftumorcellswithketonebodiesorglycolyticinhibitiondiminishestheircachecticpotential. S2-013cellswere treatedwithsolventcontrol,20mMNaHB(NaHB-S2-013),20mMLiAcAc(LiAcAc-S2-013),and10  M3-bromopyruvicacid(BPA-S2-013)for24h.The cellswerethenwashedtwicewithphosphate-bufferedsalineandculturedinserum-freeDMEM.After24h,theconditionedmediumwascollected. TheconditionedmediumwasalsopreparedfromGLUT1knockdownS2-013(S2-013-sh GLUT1 )andcontrolcells(S2-013-shScr).Differentiated myotubesfromC2C12cellswereculturedin (A) control,S2-013-CM,NaHB-S2-013-CM,LiAcAc-S2-013-CM,andBPA-S2-013-CMor (B) control, S2-013-shScr-CM,andS2-013-sh GLUT1 -CMfor72h,andbright-fieldimageswererepresentedforindividualtreatments.Differentiated3T3L1cellswere culturedin (C) control,S2-013-CM,NaHB-S2-013-CM,LiAcAc-S2-013-CM,andBPA-S2-013-CMor (D) control,S2-013-shScr-CM,andS2-013-sh GLUT1 -CM for72handstainedwithnilered,andimagesforindividualtreatmentsarerepresented. (E) DifferentiatedmyotubeformC2C12cellswereculturedin similarconditionsfor24h.TotalRNAwasisolatedandrelativemRNAlevelsof MuRF1 and Atrogin weredeterminedbyqRT-PCR.
-Actinwasutilized asaninternalcontrol. (F) Differentiated3T3L1cellswereculturedintheabove-mentionedconditionsfor24h.TotalRNAwasisolatedandrelative mRNAlevelsof Zag and HSL weredeterminedbyqRT-PCR.
-Actinwasutilizedasaninternalcontrol.Valuesrepresentedaremean±SEM.Allstatistical analyseswereconductedwithone-wayANOVAwithDunnett ’ sposthoctestandS2-013-CMasthereferencegroup.* P .05;** P .01. Shukla etal.Cancer&Metabolism 2014, 2 :18 Page13of19 http://www.cancerandmetabolism.com/content/2/1/18 whichindicatesthatthesemetabolicby-productsare notsolelyresponsibleforthecachecticphenotype. Aketogenicdietreducestumorgrowthandcachectic phenotypeinanimalmodels Todeterminetheeffectofaketogenicdietontumor growth,weimplantedS2-013pa ncreaticcancercellsortho- topicallyintothepancreasofathymicnudemice.Oneweek later,S2-013cell-implantedmicewerefed adlibitum ona ketogenicdietornormalchowfor3weeks.After3weeks onthediets,miceweresacrificedandtumorweight,tumor volume,muscleweight,andcarcassweightwererecorded. S2-013tumor-bearingmicefedontheketogenicdiet demonstratedreducedtumorweightandtumorvolume (Figure8A,B).Theketogenicdietalsoreduceddesmoplasia asobservedbyMasson ’ strichromestain(Figure8C).Fur- thermore,weinvestigatedtheeffectoftheketogenicdiet ontumorcellproliferationandapoptosisbyimmunohisto- chemicallystainingthetumor sectionsforKi67andcleaved caspase3,respectively.Weobservedthattheketogenicdiet reducedtumorcellproliferati onandincreasedapoptosisas indicatedbythedecreasedpercentageofcellswithKi67- positivestainingandincreasedstainingforcleavedcaspase 3,incomparisontothetumorsfrommicefedonacontrol Figure8 Aketogenicdietreducestumorgrowthandproliferationandrevertsthecachecticphenotype. Intothepancreasofathymic nudemice,0.5×10 6 S2-013cellswereorthotopicallyimplanted.After1weekofimplantation,miceweredividedintwogroupsandfedwith eitheracontroldietoraketogenicdiet.Threeweeksposttreatment,miceweresacrificedandtumorweight (A) andtumorvolume (B) were measured. (C) Masson ’ strichromestaining(bluestainindicatesdesmoplasticregion)andimmunohistochemistryimagesfromthecontroldiet andtheketogenicdiet-fedmicetumorsections. (D) Bloodglucoseandketonelevelsinthecontrolandtheketogenicdiet-fedmicebefore necropsy. (E) Muscleweightandcarcassweightofthecontrolandtheketogenicdiet-fedmice. (F) Immunohistochemistryofc-Mycinthe controlandtheketogenicdiet-fedmicetumorsections.Valuesshownaremean±SEM.* P 0.05;** P 0.01. Shukla etal.Cancer&Metabolism 2014, 2 :18 Page14of19 http://www.cancerandmetabolism.com/content/2/1/18