Endothelial cell-derived interleukin-6 regulates tumor growth

MISSING IMAGE

Material Information

Title:
Endothelial cell-derived interleukin-6 regulates tumor growth
Physical Description:
Mixed Material
Language:
English
Creator:
Neiva, Kathleen G.
Warner, Kristy A.
Campos, Marcia S.
Zhang, Zhaocheng
Moren, Juliana
Danciu, Theodora E.
Nor, Jacques E.
Publisher:
BioMed Central (BMC Cancer)
Publication Date:

Notes

Abstract:
Background: Endothelial cells play a complex role in the pathobiology of cancer. This role is not limited to the making of blood vessels to allow for influx of oxygen and nutrients required for the high metabolic demands of tumor cells. Indeed, it has been recently shown that tumor-associated endothelial cells secrete molecules that enhance tumor cell survival and cancer stem cell self-renewal. The hypothesis underlying this work is that specific disruption of endothelial cell-initiated signaling inhibits tumor growth. Methods: Conditioned medium from primary human dermal microvascular endothelial cells (HDMEC) stably transduced with silencing RNA for IL-6 (or controls) was used to evaluate the role of endothelial-derived IL-6 on the activation of key signaling pathways in tumor cells. In addition, these endothelial cells were co-transplanted with tumor cells into immunodefficient mice to determine the impact of endothelial cell-derived IL-6 on tumor growth and angiogenesis. Results: We observed that tumor cells adjacent to blood vessels show strong phosphorylation of STAT3, a key mediator of tumor progression. In search for a possible mechanism for the activation of the STAT3 signaling pathway, we observed that silencing interleukin (IL)-6 in tumor-associated endothelial cells inhibited STAT3 phosphorylation in tumor cells. Notably, tumors vascularized with IL-6-silenced endothelial cells showed lower intratumoral microvessel density, lower tumor cell proliferation, and slower growth than tumors vascularized with control endothelial cells. Conclusions: Collectively, these results demonstrate that IL-6 secreted by endothelial cells enhance tumor growth, and suggest that cancer patients might benefit from targeted approaches that block signaling events initiated by endothelial cells. Keywords: Cervical Cancer, Signaling pathways, Molecular targeted therapy, STAT3
General Note:
Neiva et al. BMC Cancer 2014, 14:99 http://www.biomedcentral.com/1471-2407/14/99; Pages 1-11
General Note:
doi:10.1186/1471-2407-14-99 Cite this article as: Neiva et al.: Endothelial cell-derived interleukin-6 regulates tumor growth. BMC Cancer 2014 14:99.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
© 2014 Neiva et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Resource Identifier:
System ID:
AA00021509:00001

Full Text

PAGE 1

Supplementary Figure Legends Supplementary Figure S1. Blockade of endothelial cell induced STAT3 phosphorylation in tumor cells does no t affect Akt and ERK pathways, whereas inhibition of Akt or ERK ha s a compensatory mechanism HeLa cells were serum starved overnight and exposed to A HDMEC conditioned medium (CM) or unconditioned medium (EBM) for the indicated time points. In addition, H eLa cells were pre incubated for 1 to 2 hours with B 20 M Stattic, C 20 M LY294002, or D 20 M U0126, and then exposed to HDMEC CM or EBM for the indicated time points. Phosphorylated and total STAT3, Akt, and ERK levels were determined by Western blot. Supplementary Figure S2 IL 6 potently activates STAT3 signaling in cervical adeno carcinoma cells. HeLa cells were serum starved overnight and exposed to 20 ng/ml rhIL 6 for the indicated time points. Phosphorylated and total levels of STAT3, Akt, and ERK were determined by Western blots. A HeLa cells exposed to rhIL 6. HeLa cells pre incubated for 1 to 2 hours with B, 20 M Stattic; C 20 M LY294002; or D, 20 M U0126, and then exposed to rhIL 6 for the indicated time points.



PAGE 1

RESEARCHARTICLEOpenAccessEndothelialcell-derivedinterleukin-6regulates tumorgrowthKathleenGNeiva1,KristyAWarner1,MarciaSCampos1,ZhaochengZhang1,JulianaMoren1, TheodoraEDanciu2andJacquesENr1,3,4,5,6*AbstractBackground: Endothelialcellsplayacomplexroleinthepathobiologyofcancer.Thisroleisnotlimitedtothe makingofbloodvesselstoallowforinfluxofoxygenandnutrientsrequiredforthehighmetabolicdemandsof tumorcells.Indeed,ithasbeenrecentlyshownthattumor-associatedendothelialcellssecretemoleculesthat enhancetumorcellsurvivalandcancerstemcellself-renewal.Thehypothesisunderlyingthisworkisthatspecific disruptionofendothelialcell-initiatedsignalinginhibitstumorgrowth. Methods: Conditionedmediumfromprimaryhumandermalmicrovascularendothelialcells(HDMEC)stably transducedwithsilencingRNAforIL-6(orcontrols)wasusedtoevaluatetheroleofendothelial-derivedIL-6onthe activationofkeysignalingpathwaysintumorcells.Inaddition,theseendothelialcellswereco-transplantedwith tumorcellsintoimmunodefficientmicetodeterminetheimpactofendothelialcell-derivedIL-6ontumorgrowth andangiogenesis. Results: WeobservedthattumorcellsadjacenttobloodvesselsshowstrongphosphorylationofSTAT3,akey mediatoroftumorprogression.InsearchforapossiblemechanismfortheactivationoftheSTAT3signaling pathway,weobservedthatsilencinginterleukin(IL)-6intumor-associatedendothelialcellsinhibitedSTAT3 phosphorylationintumorcells.Notably,tumorsvascularizedwithIL-6-silencedendothelialcellsshowedlower intratumoralmicrovesseldensity,lowertumorcellproliferation,andslowergrowththantumorsvascularizedwith controlendothelialcells. Conclusions: Collectively,theseresultsdemonstratethatIL-6secretedbyendothelialcellsenhancetumorgrowth, andsuggestthatcancerpatientsmightbenefitfromtargetedapproachesthatblocksignalingeventsinitiatedby endothelialcells. Keywords: CervicalCancer,Signalingpathways,Moleculartargetedtherapy,STAT3BackgroundUterinecervixcarcinoma(UCC)includesmalignantlesionsarisingfromthetissuesofthecervix,andrepresentsthe3rdmostcommoncancerinwomenworldwide withapproximately529,800newcasesdiagnosedevery year[1].Thethreemajorhistologicaltypesofinvasive cervicalcanceraresquamouscellcarcinoma(SCC), adenocarcinomas(AC)andadenosquamouscarcinoma (ASC).SCCcomprise80%ofcases,andadenocarcinomasandASCcompriseapproximately20%[1,2].Indevelopedcountries,itsincidencehasshowedamarked declineoverthepast40yearsbecauseofwidespread screeningwithcervicalcytology.Thisdeclineismainly attributabletoadecreaseintheincidenceofsquamous cellcarcinoma[3-10].Ontheotherhand,therehasbeen arelativeincreaseintheincidenceofadenocarcinomas andadenosquamouscarcinomaofthecervixoverthe sameperiod.Notably,thepathobiologyofadenocarcinomasremainsunclear,particularlytheimpactofthe crosstalkbetweenendothelialcellsandtumorcellsto cancergrowthandprogression.Betterunderstandingof signalingeventsthatmediateendothelialcell-tumorcell *Correspondence: jenor@umich.edu1AngiogenesisResearchLaboratory,DepartmentofCariology,Restorative Sciences,andEndodontics,UniversityofMichiganSchoolofDentistry,Ann Arbor,Michigan48109-1078,USA3DepartmentofBiomedicalEngineering,UniversityofMichiganCollegeof Engineering,AnnArbor,Michigan48109-1078,USA Fulllistofauthorinformationisavailableattheendofthearticle 2014Neivaetal.;licenseeBioMedCentralLtd.ThisisanOpenAccessarticledistributedunderthetermsoftheCreative CommonsAttributionLicense(http://creativecommons.org/licenses/by/2.0),whichpermitsunrestricteduse,distribution,and reproductioninanymedium,providedtheoriginalworkisproperlycredited.TheCreativeCommonsPublicDomain Dedicationwaiver(http://creativecommons.org/publicdomain/zero/1.0/)appliestothedatamadeavailableinthisarticle, unlessotherwisestated.Neiva etal.BMCCancer 2014, 14 :99 http://www.biomedcentral.com/1471-2407/14/99

PAGE 2

interactionswillleadtothedevelopmentofimproved therapiesforuterinecervixadenocarcinomas. Tumorprogressionrequirestheformationofnewblood vessels[11].Therefore,severalangiogenesisinhibitors havebeendevelopedtotargetendothelialcellsandblock tumorgrowth[12].Targetingcellsthatsupporttumor growth,ratherthanthecancercellsthemselves,isanattractiveapproachforcancertherapy.Thevascularendotheliumisdirectlyaccessibletodrugsinjectedinthe circulation,andiscomposedofcellsthataremorestable geneticallywhencomparedtocancercells[13-15].Notably,studieshavesuggestedthatbothtumorandnontumorcellsmaybeinvolvedinreducedresponsivenessto therapybydevelopingacquiredresistance[16].Despite significantadvancesintherapiestargetingangiogenicmolecules,thesurvivalbenefitsofthesetreatmentsarerelativelymodest[13],thetreatmentsarecostly[17],andhave significantsideeffects[18,19].Inaddition,single-agent therapythatiseffectiveinitiallymayultimatelyleadto drugresistance[20]andtumorrecurrence. Thedevelopmentofmoleculartargetedtherapiesmay leadtotherationalselectionoftreatmentforadenocarcinomapatientsbasedonspecificmolecularmechanismswhosederegulatedactivitycontributestothe initiation,development,andmetastaticspread[21-24]. Thederegulationofsignalingcascadesincludingthe transcriptionfactorsignaltransducerandactivatortranscription3(STAT3)pathwayhasbeenimplicatedinthe pathogenesisofcervicalcancer[21].Notably,theoverexpressionofactivatedSTAT3isaccompaniedbypoor prognosisinthissub-groupoftumors[22].Itiswell knownthatrecombinantinterleukin-6(IL-6)induces STAT3activation[23].However,theeffectofendothelial cell-secretedIL-6ontumorcellSTAT3andoverall tumorgrowthisnotknown.Thecharacterizationofthe functionalimpactofthecrosstalkbetweenendothelial cellsandtumorcellsontumorgrowthandprogression mayunveilendothelialcell-secretedmoleculesasanew conceptualtargetforcervicalcancertherapy. Theprevalentparadigmintumorbiologyisthat tumorcellssecretefactorsthatdrivetumorgrowthand thatendothelialcellssimplyrespondbygeneratingnew bloodvesselsthatsupportthehighmetabolicdemands oftumorcells.Here,wechallengedthisparadigmand observedthatendothelialcellIL-6levelshaveadirect impactontumorcellphenotypeandtumorgrowth invivo .OurresultsdemonstratethatendothelialcellsecretedIL-6definesthegrowthofadenocarcinomasin preclinicalmodels.MethodsCellcultureCervicaladenocarcinomacells(HeLaCells)wereculturedinDulbecco ’ sModifiedEagleMedium(DMEM; Invitrogen,Carlsbad,CA)supplementedwith10%fetal bovineserum(FBS),100U/mlpenicillin,and100 g/ml streptomycin(Invitrogen).Tumorcellswereserumstarvedovernightbeforeaddingtreatment.Animmortalizedhumanoralkeratinocytecellline(HOK-16B,gift ofNo-HeePark,UniversityofCalifornia,LosAngeles) wasculturedinserumfreemedium(OKM;ScienCell, Carlsbad,CA)containing1%penicillin/streptomycin, andsupplementedwith5 g/mlBSA,5 g/mltransferring,50 g/mlbovinepituitaryextract,2.5 g/mlinsulin,1ng/mlFGF,500ng/mlepinephrine,1 g/ml hydrocortisone,30nMprostaglandin,and40 g/ml plantextract(OKGS,BulletKit,ScienCell).Primaryhumandermalmicrovascularendothelialcells(HDMEC; Cambrex,Walkersville,MD)wereculturedinendothelialgrowthmedium-2(EGM2-MV;Cambrex).Conditionedmedium(CM)fromHDMECorHeLawere preparedinendothelialcellmedium(EBM)without supplementationwithgrowthfactorsorserumfrom 24-hourcultures.StableshorthairpinRNA(shRNA)transductionLentivirusesexpressingashorthairpinRNA(shRNA) constructforsilencingIL-6(VectorCore,Universityof Michigan)weregeneratedinhumanembryonickidney cells(293T)transfectedbythecalciumphosphate method,asdescribed[25].Ascrambledoligonucleotide sequence(shRNA-C)wasusedascontrol.Supernatants werecollected48hoursaftertransfectionandusedtoinfectHDMECin1:1dilutionmediumcontaining4 g/ml polybrene(Sigma-Aldrich,St.Louis,MO).CellswereselectedinEGM2-MVsupplementedwith1 g/mlpuromycin(InvivoGen,SanDiego,CA).Downregulationof IL-6wasconfirmedbyELISA.Westernblots8105HeLawereplatedin60mmdishes,starvedovernight,andexposedtoEBM,orconditionedmedium (CM)collectedfromHDMECorHeLafortheindicated timepoints.HDMECCMandHeLaCMwerenormalizedbytotalproteinconcentration.Inaddition,HOK16BwereexposedtoHDMECCM.Alternatively,tumor cellswereexposedtorhIL-6(BDP,NCI,Frederick,MD) fortheindicatedtimepoints.Signalingpathwayswere blockedbypre-incubatingtumorcellsfor1-2hourswith 20 MStattic(STAT3inhibitorV,Calbiochem,San Diego,CA),20 MLY294002(PI3kinaseinhibitor,Cell SignalingTechnology,Danvers,MA),or20 MU0126 (MEK1/2inhibitor,CellSignaling),asdescribed[26], andexposedtoHDMECCMorrhIL-6fortheindicated timepoints.Lysates(30 g)wereelectrophoresedin SDS-polyacrylamidegelsandtransferredtonitrocellulosemembranes.Primaryantibodieswere:mouseantihumanphospho-STAT3,rabbitanti-humanSTAT3,Neiva etal.BMCCancer 2014, 14 :99 Page2of11 http://www.biomedcentral.com/1471-2407/14/99

PAGE 3

rabbitanti-humanphosph o-Akt,rabbitanti-human Akt,rabbitanti-humanphospho-ERK1/2,mouseantihumanERK1/2(CellSignaling);andmouseantiglyceraldehyde-3-phosphatedehydrogenase(GAPDH; Chemicon,Millipore,Billerca,MA).Phosphorylation antibodiesdetectedendogenouslevelsofSTAT3,Akt, andERK1/2whenphosphor ylatedatTyrosine705, Serine473,andThreonine202/Tyrosine204,respectively. ImmunoreactiveproteinswerevisualizedbySuperSignalWestPicochemiluminescentsubstrate(Thermo Scientific,Rockford,IL).Enzyme-linkedimmunosorbantassay(ELISA)Supernatantsofendothelialortumorcellcultures(24hours) werecollectedandcentrifuged.IL-6expressionwasdeterminedusingELISAkits(Quantikine;R&DSystems, Minneapolis,MN)accordingtothemanufacturer ’ sinstructions.Datawerenormalizedbycellnumber.SCIDmousemodelofhumantumorangiogenesisXenografthumantumorsvascularizedwithhuman bloodvesselsweregeneratedunderanUCUCAapprovedprotocol,asdescribed[27-29].Briefly,highly porouspoly-L(lactic)acid(BoehringerIngelheim,Ingelheim,Germany)scaffoldswereseededwith9105HDMECand1105HeLaina1:1mixtureofgrowth factorreducedMatrigelandEGM2-MV.Inaddition, scaffoldswereseededwith9105HDMEC-shRNAcontrolorHDMEC-shRNA-IL-6and1105HeLa.Severecombinedimmunodeficient(SCID)mice(5-7week-oldmaleCB.17.SCID;CharlesRiver,Wilmington, MA)wereanesthetizedwithketamineandxylazine,and 2scaffoldswereimplantedinthesubcutaneousspaceof thedorsalregionofeachmouse, i.e. onescaffoldseeded withHDMEC-shRNA-control+HeLaandonescaffold seededwithHDMEC-shRNA-IL-6+HeLa.Tumorswere measuredwithacaliperevery2days,startingat14days afterimplantation.Micewereeuthanizedafter28days, implantswereretrieved,photographed,measured, weighed,fixedovernightin10%bufferedformalinat4C, andembeddedinparaffinfollowingstandardhistologicalprocedures.Thesestudieswereperformedtwo independenttimestoverifythereproducibilityofthe workunderaprotocolreviewedandapprovedbythe UniversityofMichiganCommitteeonUseandCareof Animals(UCUCA).Thetotal “ n ” ofeachexperimental conditionwasn=12tumors.ImmunohistochemistryoftissuesectionsImmunohistochemistrywasperformedinparaffinembeddedserialsectionsusingphospho-STAT3(Santa Cruz),STAT3,phospho-Akt,Akt,phospho-ERK,ERK (CellSignaling),andKi67(BiocareMedical,Concord, CA)antibodies,asdescribed[30].TumormicrovesseldensityTumormicrovesseldensitywasdeterminedfollowing identificationofbloodvesselsbyimmunohistochemistry withapolyclonalanti-humanfactorVIIIantibody(Lab Vision,Fremont,CA),aspreviouslydescribed[27,28]. Thenumberofstainedmicrovesselswascountedin10 randomfieldsperimplantinalightmicroscopeat100. Twelveimplantswereanalyzedpercondition.StatisticalanalysesT-testsorone-wayANOVAfollowedbyappropriate post-hoctestswereperformedusingSigmaStat2.0 (SPSS;Chicago,IL).StatisticalsignificancewasdeterminedatP<0.05.ResultsEndothelialcell-secretedfactorsactivatekeysignaling pathwaysintumorcellsWehavepreviouslydemonstratedthatacrosstalkinitiatedbyendothelialcellsenhancestumorcellsurvival andmigration invitro, andthatendothelialcell-derived IL-6inducesphosphorylationofSTAT3intumorcells [26].Theoverallhypothesisunderlyingthisstudyisthat theactivationofsignalingpathwaysintumorcells inducedbyendothelialcell-secretedfactorsenhances tumorgrowth.Tobegintoaddressthishypothesis, weexposedHeLacellstoserum-freeendothelialcell (HDMEC)conditionedmedium(CM)ortumorcell (HeLa)CMandanalyzedphosphorylationeventsover time(Figure1A).Weobservedthatphosphorylation levelsofSTAT3,Akt,andERKwerehigherintumor cellsexposedtoHDMECCMthanintumorcellsexposedtoHeLaCM,orunconditionedmedium(EBM). Theinductionofphosphorylationwasobservedprimarilyatearlytimepoints(15to30minutes),decreasingat 1hour(Figure1A).Notably,expressionlevelsofIL-6 werehigherinHDMECCMthaninHeLaCM,and silencingIL-6inendothelialcellsdidnothaveameasurableimpactinendothelialcellproliferation(data notshown).Inaddition,weanalyzedphosphorylation eventsonHeLacellsandonkeratinocytes(HOK-16B) exposedtoHDMECCMorunconditionedmedium (EBM)(Figure1B).Weobservedthatphosphorylation levelsofSTAT3,Akt,andERKwerehigherwhenboth tumorcellsandkeratinocyteswereexposedtoHDMEC CMthantoEBM.Similarly,phosphorylationwasobservedmainlyatearlytimepointsanddecreasedat 24hours(Figure1B). ToevaluatewhetherthetrendsofendothelialcellinducedphosphorylationofSTAT3,Akt,andERKin tumorcells invitro translateintoincreasedphosphorylationlevels invivo ,weusedtheSCIDmousemodel ofhumantumorangiogenesisinwhichweengineer cervicalcelladenocarcinomasvascularizedwithhumanNeiva etal.BMCCancer 2014, 14 :99 Page3of11 http://www.biomedcentral.com/1471-2407/14/99

PAGE 4

functionalbloodvesselsthatanastomizewiththemouse vasculature[27-29].Weimplantedhighlyporousbiodegradablescaffoldscontainingprimaryhumanendothelial cells(HDMEC)togetherwithcervicaladenocarcinoma cells(HeLa)inthesubcutaneousofSCIDmiceandanalyzedthetissuesbyimmunohistochemistry28daysafter Figure1 Endothelialcell-derivedfactorsphosphorylateSTAT3,Akt,andERKintumorcells invitro and invivo .A ,westernblotfor phosphorylatedandtotalSTAT3,Akt,andERKinHeLaserum-starvedovernightandexposedtoHDMECconditionedmedium(CM),HeLaCM,or controlunconditionedmedium(EBM)fortheindicatedtimepoints. B ,westernblotforphosphorylatedandtotalSTAT3,Akt,andERKinHeLa orHOK-16Bserum-starvedovernightandexposedtoHDMECCMorEBMfortheindicatedtimepoints. C ,immunohistochemicalanalysisforphosphorylatedSTAT3,Akt,andERK(withnuclearlocalization)inrepresentativespecimensfromxenografthumansquamouscellcarcinomas. Toppanelsrepresent100andlowerpanelsrepresent200magnification.Redarrowspointtobloodvessels. Neiva etal.BMCCancer 2014, 14 :99 Page4of11 http://www.biomedcentral.com/1471-2407/14/99

PAGE 5

transplantation.Weobservedthattumorcellsadjacentto bloodvesselsshowedphosphorylationofSTAT3,Akt,and ERK(Figure1C).Incontrast,theexpressionoftotal STAT3,Akt,andERKwasrelativelyuniformthroughout thetissues(datanotshown).Endothelialcell-inducedSTAT3phosphorylationis independentofAktandERKToexploretheinterdependenceofmolecularsignaling eventsinitiatedbyendothelialcellsontumorcells,we exposedHeLatoHDMECCMinthepresenceofchemicalinhibitorsofSTAT3,Akt,orERKpathwaysandanalyzedtheinterdependencyofthephosphorylation events.Toestablishthebaselinefortheseexperiments, weexposedHeLatoHDMECCMandanalyzedphosphorylationofSTAT3,Akt,andERKwithadetailed timecourseupto1hour(Figure2A).Weobservedthat HDMECCMinducesfirstERKphosphorylation(with strongactivationasearlyas1minute,persistinguntil 15minutes,anddecreasingat30minutes),followedby STAT3andAkt(increasinguntil15minutes,andmaintainingactivationforupto1hour)(Figure2A).When weinhibitedSTAT3phosphorylationusingthechemical inhibitorStattic,wedidnotobservesignificantchanges inphosphorylationofAktorERK(Figure2B).However, whenweinhibitedAktphosphorylationusingthePI3K inhibitorLY294002weobservedanincreaseinERK phosphorylationlevels(maintainingstrongphosphorylationforupto1hour),whilephosphorylationlevels ofSTAT3didnotchange(Figure2C).Similarly,when weinhibitedERKphosphorylationusingtheMEK1/2 inhibitorU0126weobservedincreasedAktphosphorylation(maintainingstrongphosphorylationforupto 1hour),whereasphosphorylationlevelsofSTAT3 remainedunchanged(Figure2D). Then,weextendedthetimecourseexperimentsto 24hours,andobservedthesamerelationshipbetween STAT3,Akt,andERKphosphorylationintumorcellsinducedbyendothelialcell-secretedfactors(Additionalfile1: FigureS1).STAT3,Akt,andERKphosphorylationwere strongeratearlytimepoints(15to30minutes),anddecreasedovertime.STAT3phosphorylationdecreasedat 1hourandwasmaintainedforupto24hours,phosphorylationofAktdecreasedat2hoursanddisappeared at4to24hours,whilephosphorylationofERKdecreasedsignificantlyat1hourandwasabsentat3to 24hours(Additionalfile1:FigureS1A).Inhibitionof STAT3phosphorylationdidnotaffectAktorERKphosphorylationlevels(Additionalfile1:FigureS1B).Onthe otherhand,inhibitionofAktphosphorylationincreased activationofERK(Additionalfile1:FigureS1C),andinhibitionofERKphosphorylationincreasedAktactivation(Additionalfile1:FigureS1D).Nomajoreffectwas observedinSTAT3phosphorylationlevelsusingAktor ERKinhibitors.Collectively,thesestudiesdemonstrated thatendothelialcell-inducedAktandERKphosphorylationintumorcellsinduceamutuallycompensatoryeffect,whiletheSTAT3pathwayisactivatedindependently.IL-6inducestheSTAT3signalingpathwayintumorcellsConsideringtheclinicalrelevanceoftheSTAT3signalingpathwayincervicalcarcinoma[21,22]wefocused theremainingstudiesofthisworkontheeffectofendothelialcell-secretedIL-6inthebiologyofadenocarcinomacells.TounderstandtheCervicalAdenocarcinoma responsetoIL-6stimulation,weperformedadetailed timecourseanalyzingthephosphorylationeventsin HeLacells(Figure3A).Weobservedthatwhentumor cellswereexposedtorhIL-6,thephosphorylationof STAT3,Akt,andERKfollowedsimilarpatternsaswhen tumorcellswereexposedtoHDMECCM(Figure1A; Additionalfile1:Figure1A).Wethenexposedtumor cellstoIL-6inthepresenceofchemicalinhibitorsof STAT3,Akt,orERKpathwaysandanalyzedthephosphorylationresponses(Additionalfile1:FigureS2).IL-6 stronglyactivatedSTAT3pathwayinHeLa,andslightly activatedAktorERK(Additionalfile1:FigureS2A). BlockadeofSTAT3phosphorylationhadnomajoreffect onAktbutincreasedERKphosphorylation(Additional file1:FigureS2B).InhibitionofAkthadnoeffecton STAT3,whileincreasedERKphosphorylation(Additional file1:FigureS2C).Lastly,inhibitionofERKphosphorylationhadnosignificanteffectonSTAT3orAktphosphorylation(Additionalfile1:FigureS2D).Collectively, theseresultsshowedthatIL-6isapotentinducerSTAT3 signaling,whileithasaweakereffectonthephosphorylationofAktandERKinCervicalAdenocarcinoma. TheseresultsledustofurtherexploretheIL-6/STAT3 signaling invivo. WeusedtheSCIDmousemodelofhumantumorangiogenesistogeneratehumanadenocarcinomas.WeobservedthatwhiletotalSTAT3waspresent diffuselythroughtheentiretissue(Figure3B,a),phosphorylatedSTAT3showedatendencytolocalizeadjacent tobloodvessels(Figure3B,b).Interestingly,immunostainingforthecellproliferationmarkerKi67showedthe samepatternasphosphorylatedSTAT3(Figure3B,c). TheseresultssuggestedthatphosphorylationofSTAT3in xenograftcarcinomascorrelateswithtumorcellproliferationandtheproximitytobloodvessels.Silencingofendothelialcell-IL-6issufficienttoinhibit tumorgrowthToinvestigatewhetherthese invitro trendshaveabiologicaleffect invivo ,wegeneratedxenografttumors vascularizedwithendothelialcellssecretinglowlevels ofIL-6(HDMEC-shRNA-IL-6)oremptyvectorcontrol endothelialcells(Figure4A).Tumorspopulatedwith HDMEC-shRNA-controlgrewsignificantlyfasterandNeiva etal.BMCCancer 2014, 14 :99 Page5of11 http://www.biomedcentral.com/1471-2407/14/99

PAGE 6

reached2,000mm 3 at28daysafterimplantation,whereas tumorsvascularizedwithIL-6-silencedendothelialcells presentedapproximatelyhalfofthissize(Figure4B andC).Indeed,IL-6silencingspecificallyinthevascularendothelialcellswassufficienttosignificantlyslow downxenografttumorgrowth(Figure4BandC).Tumorspopulatedwithcontrolendothelialcellsalsopresentedsignificantlyhighervolume(Figure4D)and weight(Figure4E)thantumorspopulatedwithIL-6downregulatedendothelialcells. Toexplorethemechanismsinvolvedintheinhibition oftumorgrowthmediatedbythesilencingofendothelial cell-IL-6,weanalyzedtumorcellproliferationandintratumoralmicrovesseldensitybyimmunohistochemistry. Weobservedthatexpressionoftheproliferationmarker Ki67waslowerintumorscellswhenxenograftswere vascularizedwithIL-6-silencedendothelialcells(Figure5A andB).Wealsoobservedadecreaseinmicrovessel densityintumorsvascularizedwithendothelialcellswith downregulatedIL-6expression,ascomparedtoxenografts vascularizedwithcontrolendothelialcells(Figure5C andD).Takentogether,theseresultsdemonstratedthat downregulationofIL-6intumor-associatedendothelial cellsissufficienttoinhibittumorgrowth. Discussion Abetterunderstandingofthemolecularmechanisms underlyingthedevelopmentandprogressionofthecervicaladenocarcinomamayhelptoidentifynoveltargets forpharmacologicalinterventioninthisdevastatingdisease.Wehaveshownthatfactorssecretedbyendothelial cellsincreasetumorcellsurvivalandmigration invitro Figure2 STAT3phosphorylationinducedbyendothelialcell-secretedfactorsisindependentofAktandERKphosphorylation. Western blotforphosphorylatedandtotalSTAT3,Akt,andERKinHeLaserum-starvedovernightandexposedto A ,HDMECconditionedmedium(CM) orunconditionedmedium(EBM)fortheindicatedtimepoints.Inaddition,HeLawerepre-incubatedfor1to2hourswith B ,20 MStattic; C ,20 MLY294002;or D ,20 MU0126,andthenexposedtoHDMECCMorEBMinpresenceofthespecificinhibitorfortheindicatedtimepoints. Neiva etal.BMCCancer 2014, 14 :99 Page6of11 http://www.biomedcentral.com/1471-2407/14/99

PAGE 7

[26].Here,weinvestigatedtheimpactofendothelial cell-initiatedsignalingeventstothepathobiologyofcervicaladenocarcinomas invivo. Ithasbeenshownthatconditionedmediumcollected fromendothelialcellsstimulatephosphorylationof STAT3,Akt,andERKinheadandnecksquamouscell carcinomas[26].However,itisnotknownwhetherthe abilitytoactivatethesepathwayswasuniquetoendothelialcells,oriftumorcellsthemselvescouldalsoinduce thesesignalingevents.Severalstudiesdescribeanautocrineeffectoftumorcell-secretedfactorsoncancerprogression[31-33].Here,wedemonstratedthattumor Figure3 STAT3phosphorylationinxenografthumanoralsquamouscellcarcinomascorrelateswithtumorcellproliferationand presenceofbloodvessels.A ,HeLacellswereserum-starvedovernightandexposedto20n g/mlrhIL-6fortheindicatedtimepoints.Phosphorylated andtotallevelsofSTAT3,Akt,andERKweredeterminedbyWesternblots. B-D ,xenografthumantumorsweregener atedinSCIDmicebyco-implanting HeLaandHDMEC.Tumorswereretrievedafter28days,andtissueswereanalyzedbyimmunohistochemistry: B ,totalSTAT3withcytoplasmiclocalization, diffusedthroughthetissue; C ,phosphorylatedSTAT3withnuclearlocalizatio n,concentratedintheproximityofbloodvessels; D ,Ki67withnuclear translocation,localizedprimarilyaround bloodvessels.Photomicrographsat200. Figure4 DownregulationofIL-6intumor-associatedendothelialcellsinhibitstumorgrowth.A ,ELISAforIL-6expressioninHDMEC transfectedwithshRNA-IL-6orwithacontrolscrambledoligonucleotidesequence(shRNA-C). B ,xenografthumantumorsweregeneratedin SCIDmicebyco-implantingHeLaandHDMEC-shRNA-IL-6orcontrolHDMEC-shRNA-C.Tumorsgrowthwasanalyzedwithcalipersevery2days forthedurationoftheexperiment. C ,macroscopicviewoftworepresentativexenografttumorspergroup. D-E ,graphsdepictingtumorvolume (D) andtumorweight (E) afterretrieval(28dayspost-implantation).Asteriskdepictsp<0.05. Neiva etal.BMCCancer 2014, 14 :99 Page7of11 http://www.biomedcentral.com/1471-2407/14/99

PAGE 8

cellsexposedtoendothelialcellconditionedmedium showedsignificantlyhigherlevelsofSTAT3,Akt,and ERKphosphorylationthantumorcellsexposedtoconditionedmediumcollectedfromtumorcells.SeveralstudieshaveshownthatderegulationofSTAT3,Akt,and ERKsignalingisimplicatedintumorigenesis[34-39], suggestingthataberrantactivityofanetworkofinterrelatedsignalingpathways,ratherthanasinglederegulatedroute,contributestocarcinogenesis.Weobserved thatwhilelevelsoftotalSTAT3,Akt,andERKwereuniformlydistributedthroughoutthexenografttumors, theexpressionofphosphorylatedSTAT3,Akt,andERK wasmoreclusteredaroundbloodvessels.Theseresults providefurtherevidencethatendothelialcell-secreted factorsmayplayaroleintheactivationofthesepathwayswithinthetumormicroenvironment. Toourknowledge,thecrosstalkbetweenSTAT3,Akt, andERKpathwayshasnotbeenstudiedincervicalcancer.Tryingtounderstandtherelationshipbetweenthese endothelialcell-initiatedsignalingeventsontumorcells, weexposedtumorcellstoendothelialcellconditioned mediuminthepresenceofchemicalinhibitorsof STAT3,Akt,andERKpathways.Ourresultsshowed thatendothelialcell-inducedAktandERKsignalinghave amutuallycompensatoryeffect,whileSTAT3pathway appearstobeactivatedindependently.Theseresultsare inaccordancewithaccumulatingevidencethatAktand ERKpathwaysmaycooperatetopromotethesurvivalof Figure5 DownregulationofIL-6intumor-associatedendothelialcellsreducestumorcellproliferationanddecreasesintratumoral microvesseldensity.A ,immunohistochemicalanalysisforKi67,indicatingtumorcellproliferationinimplantscontainingHeLaandHDMECshRNA-C,orHeLaandHDMEC-shRNA-IL-6.Toppanelsrepresent100andbottompanelsrepresent200. B ,quantificationoftumorcellproliferation determinedbyscoringKi67immunostaining.Datarepresentmeanvaluesobtainedinrandommicroscopicfields(100)from12tumorspercondition. Asteriskdepictsp<0.05. C ,immunohistochemistryforFactorVIIIdepictingbloodvessels(100).Leftpanelshowsarepresentativetumorpopulated withHeLaandHDMEC-shRNA-C,andrightpanelshowsatumorcontainingHeLaandHDMEC-shRNA-IL-6. D ,quantificationofthemicrovesseldensity inthesetumors.Datarepresentmeanvaluesobtainedin10randommicroscopicfieldsperimplant(100)from12tumorspercondition.Asterisk depictsp<0.05.Redarrowspointtobloodvessels. Neiva etal.BMCCancer 2014, 14 :99 Page8of11 http://www.biomedcentral.com/1471-2407/14/99

PAGE 9

transformedcells,andarealternativelyand/orcoordinatelyexpressedinseveralcancers,raisingthepossibility thatafeedbackloopmightexistinthisnetwork[40-44]. ItiswellestablishedthatactivationoftheSTAT3signalingpathwaypromotestumorgrowthandexpression ofpro-angiogenicfactors[45].Weobservedthatblockadeofendothelialcell-derivedIL-6inhibitedSTAT3 phosphorylationincancercells[26]andexpressionof CXCL8(IL-8),apotentpro-angiogenicfactorthatis stronglycorrelatedwithtumormicrovesseldensity[46]. Indeed,despitethefactthatendothelialcellssecrete manycytokinesandgrowthfactors,silencingofIL-6 withshRNA(oruseofanetutralizingantibody)completelyabrogatedinducedphosphorylationofSTAT3in tumorcells[26].Notably,expressionofIL-6ishigherin endothelialcellsthaninthetumorcellsthemselves(data notshow).Here,wereportedthatxenografttumors engineeredwithendothelialcellsstablytransducedwith shRNA-IL-6exhibitlowermicrovesseldensity.TheseresultscorroboratethehypothesisthatIL-6mediatesa pro-angiogenicparacrineloopthatplaysanimportant roleintumorgrowthandangiogenesis.Inotherwords, downregulationofIL-6secretedbyendothelialcellsinhibitsphosphorylationofSTAT3intumorcells,which willthensecretelessangiogenicfactors( e.g. CXCL8) causingadecreaseintumormicrovesseldensityand tumorgrowth. Notably,tumorcellsexpressingphosphorylatedSTAT3 localizedprimarilyadjacenttobloodvesselsandcorrelatedwithexpressionoftheproliferationmarkerKi67.We onlyanalyzedKi67positivityadjacenttobloodvesselsin bothgroupstoeliminatepossibledifferencesduetohypoxia.ExpressionofKi67intumorcellsandtumormicrovesseldensitywerelowerintumorsvascularizedwith IL-6-silencedendothelialcells.Earlystudieshaveshown thatBcl-2isupregulatedintumor-associatedendothelial cells,thatupregulationofBcl-2inmicrovascularendothelialcellsacceleratestumorgrowth,andthatendothelial cellsoverexpressingBcl-2secretehigherlevelsofIL-6 thanvectorcontrolcells[25-28]Thesefindings,along withtheresultspresentedhere,begintoprovideapossible mechanismfortheimpactofendothelialcell-derivedIL-6 ontumorgrowth.ConclusionTargeteddisruptionofthevascularendotheliumhas beenproposedbyDr.Folkmanfourdecadesagoandhas shownefficacyinsometumortypes[11-13,47].However,thisapproachresultsinhypoxic,nutrient-deprived tumormicroenvironmentsthatcanbeassociatedwith enhancedmotilityoftumorcellsanddevelopmentof evasiveresistancetotherapy[48].Here,weshowedthat specificblockadeoftheendothelialcell-tumorcellcrosstalk( e.g. IL-6)issufficienttoinhibittumorgrowth. Theseresultssuggestthatcervicalcancerpatientsmight benefitfromthetherapeuticblockadeofkeysignaling eventsthatregulatethecrosstalkbetweenendothelial cellsandtumorcells.AdditionalfileAdditionalfile1:FigureS1. Blockadeofendothelialcell-induced STAT3phosphorylationintumorcellsdoesnotaffectAktandERK pathways,whereasinhibitionofAktorERKhasacompensatory mechanism.HeLacellswereserumstarvedovernightandexposedtoA, HDMECconditionedmedium(CM)orunconditionedmedium(EBM)for theindicatedtimepoints.Inaddition,HeLacellswerepre-incubatedfor 1to2hourswithB,20 MStattic,C,20 MLY294002,orD,20 M U0126,andthenexposedtoHDMECCMorEBMfortheindicatedtime points.PhosphorylatedandtotalSTAT3,Akt,andERKlevelswere determinedbyWesternblot.FigureS2.IL-6potentlyactivatesSTAT3 signalingincervicaladenocarcinomacells.HeLacellswereserum-starved overnightandexposedto20ng/mlrhIL-6fortheindicatedtime points.PhosphorylatedandtotallevelsofSTAT3,Akt,andERKwere determinedbyWesternblots.A,HeLacellsexposedtorhIL-6.HeLa cellspre-incubatedfor1to2hourswithB,20 MStattic;C,20 M LY294002;orD,20 MU0126,andthenexposedtorhIL-6forthe indicatedtimepoints. Competinginterests Theauthorshavenocompetingofinteresttodeclare. Authors ’ contributions KGNparticipatedinthedesignofthestudy,carriedoutthe invitro and invivo experimentsanddraftedthemanuscript.KAWandMSCparticipated inthemouseexperiments,andZZparticipatedinthegenerationofthe stablecelllines.JMandTEDhelpedtodraftthemanuscriptandprovided clinic/pathologicexpertiseforthiswork.JENconceivedthestudy, participatedinitsdesignandcoordinationandhelpedtodraftthe manuscript.Allauthorsreadandapprovedthefinalmanuscript. Acknowledgments TheauthorsthankNo-HeePark(UniversityofCaliforniaLosAngeles)forthe HOK-16Bcellsusedhere.WealsothankKitrinaCordell,ValerieCastle,Cun-Yu Wang,andPeterPolverinifortheirthoughtfulinputandstrongsupportto thisproject.ThisworkwasfundedbygrantP50-CA-97248(Universityof MichiganHeadandNeckSPORE)fromtheNIH/NCI;andgrantsR21-DE19279 andR01-DE21139fromtheNIH/NIDCR(JEN). Funding FundedbygrantP50-CA-97248(UniversityofMichiganHeadandNeck SPORE)fromtheNIH/NCI;andgrantsR21-DE19279andR01-DE21139from theNIH/NIDCR(JEN). Authordetails1AngiogenesisResearchLaboratory,DepartmentofCariology,Restorative Sciences,andEndodontics,UniversityofMichiganSchoolofDentistry,Ann Arbor,Michigan48109-1078,USA.2DepartmentofPeriodonticsandOral Medicine,UniversityofMichiganSchoolofDentistry,AnnArbor,Michigan 48109-1078,USA.3DepartmentofBiomedicalEngineering,Universityof MichiganCollegeofEngineering,AnnArbor,Michigan48109-1078,USA.4DepartmentofOtolaryngology,UniversityofMichiganSchoolofMedicine, AnnArbor,Michigan48109-1078,USA.5ComprehensiveCancerCenter, UniversityofMichigan,AnnArbor,Michigan48109-1078,USA.6Angiogenesis ResearchLaboratory,UniversityofMichiganSchoolofDentistry,1011N. UniversityRm.2309,AnnArbor,MI48109-1078,USA. Received:29October2013Accepted:12February2014 Published:17February2014Neiva etal.BMCCancer 2014, 14 :99 Page9of11 http://www.biomedcentral.com/1471-2407/14/99

PAGE 10

References1.JemalA,BrayF,CenterMM,FerlayJ,WardE,FormanD: Globalcancerstatistics. CACancerJClin 2011, 61: 69 – 90. 2.CommitteeonPracticeB-G: ACOGpracticebulletin.Diagnosisandtreatmentofcervicalcarcinomas,number35,May2002. ObstetGynaecol 2002, 99: 855 – 867. 3.EifelPJ,MorrisM,OswaldMJ,WhartonJT,DelclosL: Adenocarcinomaof theuterinecervix.Prognosisandpatternsoffailurein367cases. Cancer 1990, 65: 2507 – 2514. 4.LaiCH,HsuehS,HongJH,ChangTC,TsengCJ,ChouHH,HuangKG, LinJD: Areadenocarcinomasandadenosquamouscarcinomas differentfromsquamouscarcinomasinstageIBandIIcervical cancerpatientsundergoingprimaryradicalsurgery? IntJGynecol Cancer 1999, 9: 28 – 36. 5.SmithHO,TiffanyMF,QuallsCR,KeyCR: Therisingincidenceof adenocarcinomarelativetosquamouscellcarcinomaoftheuterine cervixintheUnitedStates – a24-yearpopulation-basedstudy. GynecolOncol 2000, 78: 97 – 105. 6.SasieniP,AdamsJ: Changingratesofadenocarcinomaandadenosquamous carcinomaofthecervixinEngland. Lancet 2001, 357: 1490 – 1493. 7.WangSS,ShermanME,HildesheimA,LaceyJVJr,DevesaS: Cervical adenocarcinomaandsquamouscellcarcinomaincidencetrendsamong whitewomenandblackwomenintheUnitedStatesfor1976-2000. Cancer 2004, 100: 1035 – 1044. 8.ShermanME,WangSS,CarreonJ,DevesaSS: Mortalitytrendsforcervical squamousandadenocarcinomaintheUnitedStates.Relationto incidenceandsurvival. Cancer 2005, 103: 1258 – 1264. 9.BrayF,CarstensenB,MollerH,ZappaM,ZakeljMP,LawrenceG,HakamaM, WeiderpassE: Incidencetrendsofadenocarcinomaofthecervixin13 Europeancountries. CancerEpidemBiomarkersPrev 2005, 14: 2191 – 2199. 10.CastellsagueX,DiazM,deSanjoseS,MunozN,HerreroR,FranceschiS, PeelingRW,AshleyR,SmithJS,SnijdersPJ,MeijerCJ,BoschFX: International AgencyforResearchonCancerMulticenterCervicalCancerStudyGroup: Worldwidehumanpapillomavirusetiol ogyofcervicaladenocarcinomaand itscofactors:implicationsforscreeningandprevention. JNatCancerInst 2006, 98: 303 – 315. 11.FolkmanJ: Tumorangiogenesis:therapeuticimplications.NewEngJMed 1971, 285: 1182 – 1186. 12.KerbelR,FolkmanJ: Clinicaltranslationofangiogenesisinhibitors. NatRev Cancer 2002, 2: 727 – 739. 13.KerbelRS: Tumorangiogenesis. NewEngJMed 2008, 358: 2039 – 2049. 14.StCroixB,RagoC,VelculescuV,TraversoG,RomansKE,MontgomeryE,Lal A,RigginsGJ,LengauerC,VogelsteinB,KinzlerKW: Genesexpressedin humantumorendothelium. Science 2000, 289: 1197 – 1202. 15.KoloninM,PasqualiniR,ArapW: Molecularaddressesinbloodvesselsas targetsfortherapy. CurrOpinionChemBiol 2001, 5: 308 – 313. 16.ShojaeiF,FerraraN: Roleofthemicroenvironmentintumorgrowthand inrefractoriness/resistancetoanti-angiogenictherapies. DrugResist Updat 2008, 11: 219 – 230. 17.SchragD: Thepricetagonprogress – chemotherapyforcolorectalcancer. NewEngJMed 2004, 351: 317 – 319. 18.EskensFA,VerweijJ: Theclinicaltoxicityprofileofvascularendothelial growthfactor(VEGF)andvascularendothelialgrowthfactorreceptor (VEGFR)targetingangiogenesisinhibitors;areview. EurJCancer 2006, 42: 3127 – 3139. 19.VerheulHM,PinedoHM: Possiblemolecularmechanismsinvolvedinthe toxicityofangiogenesisinhibition. NatRevCancer 2007, 7: 475 – 485. 20.LeTourneauC,SiuLL: Molecular-targetedtherapiesinthetreatmentof squamouscellcarcinomasoftheheadandneck. CurrOpinionOncol 2008, 20: 256 – 263. 21.WeiLH,KuoML,ChenCA,ChengWF,ChengSP,HsiehFJ,HsiehCY: Interleukin-6incervicalcancer:therelationshipwithvascularendothelial growthfactor. GynecolOncol 2001, 82: 49 – 56. 22.TakemotoS,UshijimaK,KawanoK,YamaguchiT,TeradaA,FujiyoshiN, NishioS,TsudaN,IjichiM,KakumaT: Expressionofactivatedsignaltransducerandactivatoroftranscription-3predictspoorprognosisin cervicalsquamous-cellcarcinoma. BritJCancer 2009, 101: 967 – 972. 23.WegenkaUM,BuschmannJ,LuttickenC,HeinrichPC,HornF: Acute-phase responsefactor,anuclearfactorbindingtoacute-phaseresponseelements,israpidlyactivatedbyinterleukin-6attheposttranslationallevel. MolCellBiol 1993, 13: 276 – 288. 24.WeiLH,KuoML,ChenCA,ChouCH,LaiKB,LeeCN,HsiehCY: Interleukin-6 promotescervicaltumorgrowthbyVEGF-dependentangiogenesisviaa STAT3pathway. Oncogene 2003, 22: 1517 – 1527. 25.KanekoT,ZhangZ,MantelliniMG,KarlE,ZeitlinB,VerhaegenM, SoengasMS,LingenM,StrieterRM,NunezG,NrJE: Bcl-2orchestrates across-talkbetweenendothelialandtumorcellsthatpromotes tumorgrowth. CancerRes 2007, 67: 9685 – 9693. 26.NeivaKG,ZhangZ,MiyazawaM,WarnerKA,KarlE,NorJE: Crosstalk initiatedbyendothelialcellsenhancesmigrationandinhibitsanoikisof squamouscellcarcinomacellsthroughSTAT3/Akt/ERKsignaling. Neoplasia 2009, 11: 583 – 593. 27.NorJE,PetersMC,ChristensenJB,SutorikMM,LinnS,KhanMK, AddisonCL,MooneyDJ,PolveriniPJ: Engineeringandcharacterization offunctionalhumanmicrovesselsinimmunodeficientmice. LabInvest 2001, 81: 453 – 463. 28.NorJE,ChristensenJ,LiuJ,PetersM,MooneyDJ,StrieterRM,PolveriniPJ: Up-RegulationofBcl-2inmicrovascularendothelialcellsenhances intratumoralangiogenesisandacceleratestumorgrowth. CancerRes 2001, 61: 2183 – 2188. 29.WarnerKA,MiyazawaM,CordeiroMM,LoveWJ,PinskyMS,NeivaKG, SpaldingAC,NorJE: Endothelialcellsenhancetumorcellinvasion throughacrosstalkmediatedbyCXCchemokinesignaling. Neoplasia 2008, 10: 131 – 139. 30.SquarizeCH,CastilhoRM,SriuranpongV,PintoDSJr,GutkindJS: Molecularcross-talkbetweentheNFkappaBandSTAT3signaling pathwaysinheadandnecksquamouscellcarcinoma. Neoplasia 2006, 8: 733 – 746. 31.GaoSP,MarkKG,LeslieK,PaoW,MotoiN,GeraldWL,TravisWD,BornmannW, VeachD,ClarksonB,BrombergJF: MutationsintheEGFRkinasedomain mediateSTAT3activationviaIL-6productioninhumanlungadenocarcinomas. JClinInvest 2007, 117: 3846 – 3856. 32.SansoneP,StorciG,TavolariS,GuarnieriT,GiovanniniC,TaffurelliM, CeccarelliC,SantiniD,PateriniP,MarcuKB,ChiecoP,BonafM: IL-6 triggersmalignantfeaturesinmammospheresfromhumanductal breastcarcinomaandnormalmammarygland. JClinInvest2007, 117: 3988 – 4002. 33.LeeS,ChenTT,BarberCL,JordanMC,MurdockJ,DesaiS,FerraraN, NagyA,RoosKP,Iruela-ArispeML: AutocrineVEGFsignalingisrequired forvascularhomeostasis. Cell 2007, 130: 691 – 703. 34.FletcherS,TurksonJ,GunningPT: Molecularapproachestowardsthe inhibitionofthesignaltransducerandactivatoroftranscription3(Stat3) protein. ChemMedChem 2008, 3: 1159 – 1168. 35.AggarwalBB,SethiG,AhnKS,SandurSK,PandeyMK,KunnumakkaraAB, SungB,IchikawaH: Targetingsignal-transducer-and-activator-oftranscription-3forpreventionandtherapyofcancer:moderntarget butancientsolution. AnnNYAcadSci 2006, 1091: 151 – 169. 36.RobertsPJ,DerCJ: TargetingtheRaf-MEK-ERKmitogen-activated proteinkinasecascadeforthetreatmentofcancer. Oncogene 2007, 26: 3291 – 3310. 37.FridayBB,AdjeiAA: AdvancesintargetingtheRas/Raf/MEK/ErkmitogenactivatedproteinkinasecascadewithMEKinhibitorsforcancertherapy. ClinCancerRes 2008, 14: 342 – 346. 38.LuoJ,ManningBD,CantleyLC: TargetingthePI3K-Aktpathwayinhuman cancer:rationaleandpromise. CancerCell 2003, 4: 257 – 262. 39.VivancoI,SawyersCL: Thephosphatidylinositol3-KinaseAKTpathwayin humancancer. NatureRevCancer 2002, 2: 489 – 501. 40.GrantS: Cotargetingsurvivalsignalingpathwaysincancer. JClinInvest 2008, 118: 3003 – 3006. 41.KinkadeCW,Castillo-MartinM,Puzio-KuterA,YanJ,FosterTH,GaoH,SunY, OuyangX,GeraldWL,Cordon-CardoC,Abate-ShenC: TargetingAKT/ mTORandERKMAPKsignalinginhibitshormone-refractoryprostate cancerinapreclinicalmousemodel. JClinInvest 2008, 118: 3051 – 3064. 42.CarracedoA,MaL,Teruya-FeldsteinJ,RojoF,SalmenaL,AlimontiA,EgiaA, SasakiAT,ThomasG,KozmaSC,PapaA,NardellaC,CantleyLC,BaselgaJ, PandolfiPP: InhibitionofmTORC1leadstoMAPKpathwayactivation throughaPI3K-dependentfeedbackloopinhumancancer. JClinInvest 2008, 118: 3065 – 3074. 43.GaoH,OuyangX,Banach-PetroskyWA,GeraldWL,ShenMM,Abate-ShenC: CombinatorialactivitiesofAktandB-Raf/Erksignalinginamousemodelofandrogen-independentprostatecancer. ProcNatlAcadSciUSA 2006, 103: 14477 – 14482.Neiva etal.BMCCancer 2014, 14 :99 Page10of11 http://www.biomedcentral.com/1471-2407/14/99

PAGE 11

44.McCubreyJA,SteelmanLS,FranklinRA,AbramsSL,ChappellWH,WongEW, LehmannBD,TerrianDM,BaseckeJ,StivalaF,LibraM,EvangelistiC,Martelli AM: TargetingtheRAF/MEK/ERK,PI3K/AKTandp53pathwaysin hematopoieticdrugresistance. AdvEnzymRegul 2007, 47: 64 – 103. 45.KamranMZ,PatilP,GudeRP: RoleofSTAT3incancermetastasisand translationaladvances. BiomedResInt. 2013, 2013: 421821. 46.PiperiC,SamarasV,LevidouG,KavantzasN,BoviatsisE,PetrakiK,GrivasA, BarbatisC,VarsosV,PatsourisE,KorkolopoulouP: Prognosticsignificance ofIL-8-STAT-3pathwayinastrocytomas:correlationwithIL-6,VEGFand microvesselmorphometry. Cytokine 2011, 55: 387 – 395. 47.MonkBJ,WillmottLJ,SumnerDA: Anti-angiogenesisagentsinmetastatic orrecurrentcervicalcancer. GynecolOncol 2010, 116: 181 – 186. 48.BergersG,HanahanD: Modesofresistancetoanti-angiogenictherapy. NatureRevCancer 2008, 8: 592 – 603.doi:10.1186/1471-2407-14-99 Citethisarticleas: Neiva etal. : Endothelialcell-derivedinterleukin-6 regulatestumorgrowth. BMCCancer 2014 14 :99. Submit your next manuscript to BioMed Central and take full advantage of: € Convenient online submission € Thorough peer review € No space constraints or color “gure charges € Immediate publication on acceptance € Inclusion in PubMed, CAS, Scopus and Google Scholar € Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Neiva etal.BMCCancer 2014, 14 :99 Page11of11 http://www.biomedcentral.com/1471-2407/14/99