Investigation of CCL18 and A1AT as potential urinary biomarkers for bladder cancer detection

MISSING IMAGE

Material Information

Title:
Investigation of CCL18 and A1AT as potential urinary biomarkers for bladder cancer detection
Physical Description:
Mixed Material
Creator:
Miyake, Makito
Ross, Shanti
Lawton, Adrienne
Chang, Myron
Dai, Yunfeng
Mengual, Lourdes
Publisher:
Bio-Med Central (BMC Urology)
Publication Date:

Record Information

Rights Management:
All applicable rights reserved by the source institution and holding location.
System ID:
AA00017955:00001

Full Text

Supplemental Figure 1


a


UROtsa Cancer
cell cell


A
*


J


1,250-


6 1,200-

o 1,150.
C

i~inn


Zo OP
UROtsa Cancer
cell cell


.... ..


I1


II


- ----- -


UROtsa Cancer
cell cell


UROtsa Cancer
cell cell


80-

600
.40-
r
U
0
.5 20-
3-


Eu" -


NO
qo


II


i


B


.I.-- i
oJp*
qo


II




Full Text

PAGE 1

RESEARCHARTICLEOpenAccessInvestigationofCCL18andA1ATaspotential urinarybiomarkersforbladdercancerdetectionMakitoMiyake1,ShantiRoss1,AdrienneLawton2,MyronChang3,YunfengDai3,LourdesMengual4, AntonioAlcaraz4,EvanGomesGiacoia1,SteveGoodison1,5andCharlesJRosser1,5,6*AbstractBackground: Inthisstudy,wefurtherinvestigatedtheassociationoftwobiomarkers,CCL18andA1AT,with bladdercancer(BCa)andevaluatedtheinfluenceofpotentiallyconfoundingfactorsinanexperimentalmodel. Methods: Inacohortof308subjects(102withBCa),urinaryconcentrationsofCCL18andA1ATwereassessedby enzyme-linkedimmunosorbentassay(ELISA).Inanexperimentalmodel,benignorcancerouscells,inadditionto blood,wereaddedtourinesfromhealthycontrolsandanalyzedbyELISA.Lastly,immunohistochemicalstainingfor CCL18andA1ATinhumanbladdertumorswasperformed. Results: MedianurinaryproteinconcentrationsofCCL18(52.84pg/ml vs. 11.13pg/ml, p <0.0001)andA1AT (606.4ng/ml vs. 120.0ng/ml, p <0.0001)weresignificantlyelevatedinBCasubjectscomparedtocontrols. Furthermore,theadditionofwholebloodtopoolednormalurineresultedinasignificantincreaseinbothCCL18 andA1AT.IHCstainingofbladdertumorsrevealedCCL18immunoreactivityininflammatorycellsonly,andthere wasnosignificantincreaseintheseimmunoreactivecellswithinbenignandcanceroustissueandnoassociation withBCagradenorstagewasnoted.A1ATimmunoreactivitywasobservedinthecytoplasmofepitheliacellsand intensityofimmunostainingincreasedwithtumorgrade,butnottumorstage. Conclusions: FurtherdevelopmentofA1ATasadiagnosticbiomarkerforBCaiswarranted. Keywords: Biomarkers,Bladdercancer,Specificity,UrineBackgroundNon-invasiveurinetestsfortheearlydetectionorpostsurgicalsurveillanceofbladdercancer(BCa)arehighly desirableforboththepatientandthehealthcaresystem. Currently,voidedurinarycytology(VUC)isthemost widelyusednon-invasiveurinetest,withreportedspecificitiesrangingfrom85-100%andsensitivitiesranging from13-75%[1,2].Twosingle-proteinbiomarkerurinebasedassays,bladdertumorantigen(BTA)testand nuclearmatrixprotein-22(NMP-22)test,havebeen developedandFDAapprovedforuseinthiscontext. However,theseassayshavesignificantlimitations.The BTAtests(BTAstat ™ andBTATRAK ™ (PolymedcoInc. CortlandtManor,NY,USA)havediagnosticsensitivities rangingfrom29-83%andspeci ficitiesrangingfrom56-86% [3,4].Inaddition,weandothershavedemonstratedinan experimentalmodelthathem aturiaadverselyaffectsthe accuracyoftheBTAassay[5,6].TheNMP-22tests (NMP22BladderCancerELISATestKitandtheNMP22 BladderChekpoint-of-caretest,AlereScarborough,Inc. Waltham,MA)havediagnosticsensitivitiesrangingfrom 47-100%andspecificitiesrangingfrom55-98%[7,8]. Atsu etal. andothershaverecentlydemonstratedinanexperimentalmodelthatNMP-22assaysmeasurethecellularityoramountofcellturnoverthatmaybeintroducedinto theurinebyavarietyofconditions,includinghematuria, infectionandinstrumentation[9,10].Thus,thesearchfor moreaccurateurine-basedb iomarkerscontinues. Throughgenomicandproteomicprofilingofurine components,wehavepreviouslyidentifiedapanelof biomarkersthatcanoutperformcurrenturine-based biomarkersforthenon-invasivedetectionofBCa [11-14].Inacase-controlledvalidationstudy,theurinary concentrationsofourpanelof14biomarkers(IL-8, *Correspondence: deacdoc@aol.com1CancerResearchInstitute,OrlandoHealth,Orlando,FL32827,USA5NonagenBioscienceCorporation,Orlando,FL32827,USA Fulllistofauthorinformationisavailableattheendofthearticle 2013Miyakeetal.;licenseeBioMedCentralLtd.ThisisanOpenAccessarticledistributedunderthetermsoftheCreative CommonsAttributionLicense(http://creativecommons.org/licenses/by/2.0),whichpermitsunrestricteduse,distribution,and reproductioninanymedium,providedtheoriginalworkisproperlycited.TheCreativeCommonsPublicDomainDedication waiver(http://creativecommons.org/publicdomain/zero/1.0/)appliestothedatamadeavailableinthisarticle,unlessotherwise stated.Miyake etal.BMCUrology 2013, 13 :42 http://www.biomedcentral.com/1471-2490/13/42

PAGE 2

MMP-9,MMP-10,SDC1,CCL18,PAI-1,CD44,VEGF, ANG,CA9,A1AT,SPP1,PTX3,andAPOE)weremeasuredbyenzyme-linkedimmunosorbentassay(ELISA) invoidedurinesfrom127patients(64tumorbearing subjects)[15-18].Ofthese14biomarkers,twobiomarkers(CCL18andA1AT)hadhighcorrelationcoefficients(Spearmancorrelationcoefficient>0.76)with urinarybloodcontentandtherefore,ratherthanmeasuringavalidtumorantigenthebiomarkermaybemerelya surrogateforhematuria.Subsequently,thesetwobiomarkershavebeenexcludedfromongoingmultiplexstudies[19]untilwecanclarifythesourceoftheseprotein biomarkers.Herein,wereporttheurinaryconcentrations ofCCL18andA1ATinanindependentlargercase – controlstudy,andillustrateinanexperimentalmodelthe influenceofcellularproteinsandwholebloodontheperformanceofthesepotentialurine-basedbiomarkers.MethodsEthicsstatementUnderInstitutionalReviewBoardapprovalbythecommitteesatMDAndersonCancerCenterOrlandoand HospitalClinicofBarcelona,writteninformedconsent wasobtainedpriortocollectionandstorageofbiological specimens(voidedurinesamplesandblood)ingenitourinarybiorepositories.FurthermoreunderInstitutional ReviewBoardapprovalbythecommitteeatMDAnderson CancerCenterOrlandowithawaiverofwritteninformed consent,archivedbladdertissuesfromtheDepartmentof PathologyatOrlandoHealthwasidentifiedforimmunohistochemicalanalysis.Theabovereviewboardsmonitoredstudyrecruitmentandstudycompliance.PatientsanddatacollectionFortheurinaryELISAvalidationstudy,308nonconsecutivesubjects(102withBCa)fromMDAnderson CancerCenterOrlandoandHospitalClnicofBarcelona wereavailableforanalysis.Thecontrolcohortconsistedof 206individuals(47withvoidingsymptoms,44withurolithiasis,9withgrosshematuria,14withurinarytractinfectionand92withoutanydiagnosedcondition).Patients withahistoryofrenaldysfunctionwereexcluded.The cohortof308subjectsservedasourphaseII(validation study)accordingtotheInternationalConsensusPanelon BladderTumorMarkersandfindingswerereported accordingtotheSTARDcriteria[20].Fortheexperimentalmodel,threehealthyvolunteers(2males,1female, meanage36years)providedurineandbloodsamples.For theimmunohistochemicalstudy,formalin-fixedparaffin embeddedblockscontaining165bladdertumortissue specimensand8benigntissuespecimenswereretrieved fromtheOrlandoHealthDepartmentofPathology.SpecimenprocessingFiftymillilitersofvoidedurinefromeachsubjectwas assignedauniqueidentifyingnumberbeforedeliveryto thelaboratoryforprocessing.Eachurinesamplewas centrifugedat1000 g 4Cfor10min.Thesupernatant wasdecantedandaliquoted,andtheurinarypelletwas snapfrozen.Boththesupernatantandpelletwerestored at-80Cpriortoanalysis.UrinesupernatantproteinconcentrationwasdeterminedusingPierce660-nmProtein AssayKit(ThermoFisherScientificInc.,Waltham,MA, USA).Patientswithsignificantproteinuriawereexcluded.Enzyme-linkedimmunosorbentassaysforurinaryCCL18 andA1ATThelevelsofhumanCCL18(Cat#ab100620,Abcam, Cambridge,MA)andhumanA1AT(Cat#ab108799, Abcam)inurinesamplesweremonitoredusingELISA. Theassayswereconductedaccordingtothemanufacturer ’ sinstructions.Laboratorypersonnelwereblindedto finaldiagnosis.Calibrationcurveswerepreparedusing purifiedstandardsforeachproteinassessed.Curvefitting wasaccomplishedbyeitherlinearorfour-parameterlogisticregressionfollowingmanufacturer ’ sinstructions.Urinarycreatininelevelsweremonitoredwithacommercial ELISAassay(Cat#KGE005R&DSystemsInc.,Minneapolis,MN,USA)aspreviouslydescribed[21].CelllinesandcultureHumanbladdercancercelllinesT24(ATCC,Manassas, VA)andUM-UC-14(agenerousgiftfromDr.H.Bart Grossman,TheUniversityofTexasM.D.AndersonCancerCenter,Houston,TX)[22]wereavailableforanalysis. Thebenignhumanbladdercellline,UROtsa,wasagenerousgiftfromDr.DonaldSensattheUniversityof NorthDakotaSchoolOfMedicine(GrandForks,ND) [23].T24andUM-UC-14celllinesweremaintainedin RPMI1640media.UROtsacellsweremaintainedin McCoy ’ s5Amedium(LifeTechnologies,Inc.,Gaithersburg,MD).Allmediaweresupplementedwith10%fetal bovineserum,100units/mlofpenicillinand100 g/ml ofstreptomycin.Allcellswereincubatedat37Cina humidifiedatmosphereof5%CO2.ExperimentalmodelTheexperimentalmodelwasessentiallyaspreviously published[6,10].Figure1illustratestheexperimental modelcomponentsanddilutions.Briefly,10milliliters ofwholebloodinheparinizedtubeand200mloffreshly voidedurinesamplesinsterilecontainerswereobtained fromthreehealthycontrols.Theurinesamplesfromthe threehealthysubjectswerepooled,mixedanddistributedinto10mlaliquotsin15mlcentrifugetubes.The humanbladdercelllineswerewashed,trypsinizedand counted.ForUROtsa,1104cells(lowconcentration),Miyake etal.BMCUrology 2013, 13 :42 Page2of10 http://www.biomedcentral.com/1471-2490/13/42

PAGE 3

1105cells(mediumconcentration)and1106cells (highconcentration)cellswereeachaddedto10ml pooledurinesamplesintriplicate.Equalnumbersof T24andUM-UC-14werepooled,and1104pooled cells(lowconcentration),1105pooledcells(medium concentration)and1106pooledcells(highconcentration)cellswereaddedto10mlofpooledurinesamples intriplicate.Forcelllysateanalyses,1106cellsfrom eachcelllinewerelysedwithRIPAbuffer(Pierce,Rockford,IL)andtotalproteinconcentrationmeasured.The totalproteinextractedfrom1106cellsofUROtsa,T24 andUM-UC-14were431 g,471 g,280 gand 420 g,respectively,withameantotalproteinextractof 400 g.Inthespikingexperiments,4 g,40 gand 400 gofcellularproteinsfromeitherUROtsaorthe pooledBCacelllineswereused,correspondingto ~1104cells(lowconcentration),~1105cells(medium concentration)and~1106cells(highconcentration). UROtsalysatesandpooledcancercelllysateswere addedtopooledurinesamplesintriplicate.Tomonitor theinfluenceofhematuria,pooledwholebloodfrom threehealthysubjectswasaddedintriplicateto10mlof pooledurinesamplesinthefollowingamounts;1 l,1/ 10000finaldilution;5 l,1/2000finaldilution;20 l,1/ 500finaldilution;50 l,1/200finaldilutionandzero control.Thenumberofredbloodcells(RBC)ineach urinesamplewasdeterminedbymicroscopicexaminationbeforeandafteraddingwholeblood.Standard urinalysiswasperformedwithMULTISTIXPROReagentStrips(BayerHealthCare,Elkhart,IN).ImmunohistochemistryAtotalof173paraffinblockswer everifiedhistologicallyby H&Estaining.Forimmunochemicalstaining,blockswere cutin5 msectionsandplacedonaSuperfrostPlus Miscroslide.Sectionsweredeparaffinized,followedbyantigenretrievalusingcitricacidbuffer(pH6.0,95Cfor20 min).Theslidesweretreatedwith1%hydrogenperoxidein methanoltoblockendogenousp eroxidaseactivity.After20 minblockingin1%bovineserumalbumin(BSA),theslides wereincubatedovernightat 4Cwithanti-humanCCL18 antibody(MAB394;mousemonoclonal,dilution1/500in 1%BSA)fromR&DSystemsInc.,oranti-humanA1AT antibody(NBP1-90309;rabbitpolyclonal,dilution1/2500 in1%BSA)fromNovusBiologicalsInc.(Littleton,CO). Next,theslideswereincubatedwith2 g/mLof biotinylatedanti-mouseoran ti-rabbitIgGsecondaryantibody(VectorLaboratories,Burlingame,CA)for30minat roomtemperature.Subsequently,thesectionswerestained usingStandardUltra-Sensiti veABCPeroxidaseStainingkit (Pierce/ThermoFisherScientific,SanJose,CA)and3,3'diaminobenzidine(DAB;Vecto rLaboratories),counterstainedbyhematoxylin,deh ydrated,andmountedwitha coverslide.Humanliver,knowntostainstronglyfor CCL18andA1AT,wasusedasapositivecontrol,andnegativecontrolswereperformedbyomittingtheprimaryantibodies.Usinglightmicroscopy ,twoinvestigators(MMand AL)interpretedimmunostainingresultsblindedtospecimenandpatientdata.Athirdinvestigator(CJR)reviewed discrepanciesandrenderedafinalscore.Thelocationof immunoreactivity( e.g., nuclear,cytoplasm,cellmembrane, andstroma)wasnoted.CCL18immunostainingwaspositiveonlyininflammatorycellsinthestroma.Threerandomlychosenhighpowerfields(1HPF=0.237mm2)were analyzedforCCL18-positivecellsinthestromalareaand averagedineachcase.A1ATimmunostainingwaspositive onlyinthecytoplasmofepit helialcells.Immunostaining intensitywasreportedasweak,moderateorstrong.DataanalysisTheWilcoxonranksumtestwasusedonELISAdatato determinetheassociationbetweenurinaryCCL18, Figure1 Schematicoftheexperimentalmodel. Low concentration(1104),mediumconcentration(1105)andhigh concentrations(1106)ofintactUROtsabenignhumanbladdercells, oramixtureofhumanbladdercancerlines,T24andUM-UC-14 wereaddedto10mlofpooledurinefromthreehealthycontrols. Lowproteinconcentrationofcellularlysate(4 g),medium concentration(40 g)andhighconcentrations(400 g)ofUROtsa benignhumanbladdercells,orfromamixtureofhumanbladder cancerlines,T24andUM-UC-14wereaddedto10mlofpooled urinefromthreehealthycontrols.Wholeblood(1,5,20and50 l) wasalsoaddedto10mlofpooledurinefromhealthycontrols. Miyake etal.BMCUrology 2013, 13 :42 Page3of10 http://www.biomedcentral.com/1471-2490/13/42

PAGE 4

A1ATandBCastatus.Nonparametricreceiveroperatingcharacteristic(ROC)curveswereplottedandthe abilityofthebiomarkertopredictthepresenceof BCawasestimatedbycalculatingtheareaunderthe ROCcurves(AUROC).Thesensitivityandspecificity ofthebiomarkerattheoptimalcutoffvaluewas definedbycalculatingtheYoudenindex[24].Comparisonofimmunohistochemicaldistributiondatawas performedusingChisquaretest.Spearmanrank correlationcoefficientswereusedtoexaminethecorrelationbetweenurinaryCCL18andA1ATconcentrationsandurinaryhemoglobinconcentration.The associationbetweenCCL18andA1ATlevelsandBCa wastestedusingtheMannW hitneytest.Statistical significanceinthisstudywassetat p <0.05and allreported p valueswere2-sided.Allanalyses wereperformedusingPRISMsoftwareversion5.00 (SanDiego,CA).ResultsCCL18andA1ATinvoidedurinesamplesTable1depictsdemographicsandclinicalcharacteristics ofthestudycohorts.NinetypercentoftheBCasubjects sampledwereCaucasian(medianage69years)with60% notedtohavenon-muscleinvasivebladdercancer (NMIBC)and37%withlow-gradedisease.Inthecancer cohort,urinarycytologyachievedadiagnosticsensitivity of39%.Medianurinaryproteinconcentrationsof CCL18(52.84pg/ml vs. 11.13pg/ml, p <0.0001)and A1AT(606.4ng/ml vs. 120ng/ml, p <0.0001)weresignificantlyelevatedinBCasubjectscomparedtocontrols. Furthermore,medianurinaryCCL18wassignificantly elevatedinmuscleinvasivebladdercancer(MIBC) comparedtoNMIBC(90.65pg/ml vs .44.72pg/ml, p =0.044),andapproachedsignificance(Figure2a)in high-gradecomparedtolow-gradedisease(79.57pg/ml vs. 38.10pg/ml, p =0.073).Similarly,medianurinary Table1Demographicandclinicopathologiccharacteristicsof308subjectscomprisingELISAstudycohortand173 subjectscomprisingIHCstudycohortELISACohortIHCCohort BCa(%)n=102Controls(%)n=206BCa(%)n=165Controls(%)n=8 MedianAge(range,y)69(20 – 93)56(18 – 89)67(47 – 91)59(42 – 83) Male:Femaleratio84:18152:54132:337:1 Race White91(90%)135(66%)137(83%)7(88%) AfricanAmerican5(5.9%)20(10%)8(5%)0(0%) Other6(6.1%)51(24%)20(12%)1(12%) PositiveFISH40/74(54%)2/22(9%)N/AN/A Suspicious/positivecytology37/94(39%)2/22(9%)N/AN/A Medianfollow-up(months)144113 Clinicalstage Tis6(6%)15(9%) Ta41(40%)15(9%) T114(14%)15(9%) T241(40%)120(73%) Tumorgrade Low38(37%)36(22%) High64(63%)129(78%) Mediantumorsize(cm)3.03.5 MedianurinaryCCL18pg/ml(range)52.8411.13 (2 – 6949)(0 – 662) MedianurinaryA1ATng/ml(range)606.4120.0 (26 – 8830)(1 – 1997) Medianurinaryprotein g/ml(range)24.238.9 (24 – 789)(24 – 561) Miyake etal.BMCUrology 2013, 13 :42 Page4of10 http://www.biomedcentral.com/1471-2490/13/42

PAGE 5

A1ATwassignificantlyelevatedinMIBCcomparedto NMIBC(978.1ng/ml vs. 414.5ng/ml, p =0.0042)and approachedsignificance(Figure2b)inhigh-grade comparedtolow-grade(917.8ng/ml vs. 414.5ng/ml, p =0.073).TheabilityofthetestedbiomarkerstopredictthepresenceofBCawasanalyzedusingnonparametricROCanalyses,accordingtoNationalCancer Instituteguidelines[25].BasedontheAUROC,wedeterminedYoudenindexcutoffvaluestomaximizethe sumofsensitivityandspecificity.UrinaryCCL18hadan areaunderthecurveof0.768(95%CI:0.713-0.824) (Figure3),achievedasensitivityof70.4%,specificityof 67.7%,positivepredictivevalueof53.1%andanegative predictivevalueof81.5%.UrinaryA1AThadanarea underthecurveof0.775(95%CI:0.721-0.829)(Figure3), achievedasensitivityof70.6%,specificityof71.8%,positivepredictivevalueof55.4%andnegativepredictive valueof83.2%.ExperimentalmodelUrineandbloodsampleswereobtainedfromthree healthyvolunteercontrolsforanalysisintheexperimentalmodel.Therewasnoevidenceofgrosshematuria, urinarytractinfectionoranybiochemicalabnormalities inanyvolunteerurinesamples.Bothurinarydipstick andurinarymicroscopywerenegativeforhematuria, however,urinaryhemoglobinmeasuredbyELISAassay revealedtraceamountsinallsamples(3.630.59ng/ ml).Inthesehealthycontrols,themeanurinaryCCL18 levelwas5.967.73pg/ml,andthemeanurinaryA1AT levelwas117071.94ng/ml(Table2). Urinesampleswerepooled,andwholecellsorcell lysatesofthecancercelllinepoolandUROtsacellswere addedtotheurinesample(asdepictedinFigure1)and re-analyzedforCCL18andA1ATusingELISA.The additionofahighconcentration(400 g)ofproteinlysatefromUROtsacells,ormediumtohighconcentration (40 gto400 g)ofproteinlysatefrompooledcancerouscellsresultedinasignificantincreaseintestsample CCL18(Additionalfile1).Furthermore,theadditionof wholeblood(50 L)resultedinasignificantincreasein CCL18( p <0.05)(Figure4).AsforA1AT,theaddition ofhighconcentrationofwholeblood(50 L)resulted inasignificantincreaseintestsampleA1ATlevels ( p <0.05)(Figure4). Withtheadditionofonly1 lofwholebloodto 10mLoftesturines(1/10,000dilution),themeanurinary hemoglobinlevelwas174.2815.69ng/ml,andmicroscopyrevealedamedian1RBC/hpf.Thislevelwouldbe termed ‘ Negative ’ or ‘ Trace ’ inclinicaltestssuchasthe MultistixProdipsticktest(negativebloodis<100ng/ml). Atthislevel,CCL18andA1ATconcentrationswere Figure2 Urinarybiomarkerlevels.a) ComparisonofurineconcentrationsofCCL-18betweenthecancerandnon-cancergroups(leftpanel), low-gradeandhigh-grade(middlepanel)andnon-muscleinvasivebladdercancer(NMIBC)andmuscle-invasivebladdercancer(MIBC). b) ComparisonofurineconcentrationsofA1ATbetweenthesamegroups.Medianlevelsaredepictedbyhorizontallineswithinboxes,standard deviationsaredepictedbybars.Significance( p<0.05 )wasassessedbytheWilcoxonranksumtest. Miyake etal.BMCUrology 2013, 13 :42 Page5of10 http://www.biomedcentral.com/1471-2490/13/42

PAGE 6

unaffectedat7.067.78ng/mland1227.311.22ng/ml, respectively.Withtheadditionof50 lofwholebloodto 10mLofurine,allurineshadvisiblygrosshematuria,the meanurinaryhemoglobinlevelwas7,898.05184.67ng/ mlandmedianof17RBC/hpfwasnoted.Atthislevel, CCL18concentrationwasraisedto31.9043.19ng/ml (~4.5foldincrease),buttheA1ATconcentrationwas similartocontrolsat1,273.165.53ng/ml(Table2).As theconcentrationofwholebloodaddedtotheurinesamplesincreased,themeanurinaryhemoglobinlevel,theextentofhematuriaassessedbymicroscopy,andthemean urinaryconcentrationsofCCL18andA1ATincreasedaccordingly(Table2).TherewerehighcorrelationcoefficientsbetweenhemoglobinandCCL18(Spearman correlationcoefficient=0.90)andhemoglobinandA1AT (Spearmancorrelationcoefficient=1.00).ImmunohistochemicalanalysisofbladdertumorsThestudycohortconsistedof8subjectswithoutcancer and165non-consecutivesubjectswithBCa(37subjects withlow-gradeBCaand128subjectswithhigh-grade BCa,45subjectswithNMIBCand120MIBC).ImmunohistochemicalstainingpatternsforCCL18andA1AT wereassessedinbothmalignantandnormalbladdertissue.NoepithelialstainingwasevidentforCCL18,however,inflammatorycellsinthestromalwerepositive. ThenumberofCCL18-positiveinflammatorycellsper highpowerfieldwasnotincreasedinbladdertumors comparedtocontrols(1.01.2 vs. 4.56.9, p =0.57). Inaddition,anincreaseinCCL18-positiveinflammatory cellswasnotassociatedwithhighergradeorhigher stagedisease(Figure5a).ImmunostainingforA1AT revealedapredominantlyepithelialandcytoplasmic localization.Intensityofstainingrangedfromweak andfocaltostronganddiffuse.Nodifferencein stainingintensitywasseenbetweenbenignandcancer( p =0.99).Stainingintensityincreasedwithanincreaseintumorgrade( p =0.05),however,staining patternwasnotsignificantlyassociatedwithtumor stage( p =0.79)(Figure5b).DiscussionandconclusionsWehavepreviouslyidentifiedCCL18andA1ATas potentialbiomarkersforthedetectionofBCainvoided urinesamples[15,18].CCL18isamemberoftheserumbasedcytokinefamilyofsecretedproteinsinvolvedin immunoregulatoryandinflammatoryprocesses.CCL18 isthoughttopromotetheinvasivenessofcancercells bytriggeringintegrinclusteringandenhancingtheir adherencetotheextracellularmatrix,andareceptor (PITPNM3)forthiscytokinehasbeenrecentlyidentified [26].CCL18hasbeenidentifiedingynecologicaltumors Figure3 DiagnosticperformanceofurinaryCCL18andA1AT. Receiveroperatorcharacteristic(ROC)curveswerecalculatedfromtheanalysisof urinesamplesobtainedfromacohortof308subjects(102withconfirmedbladdercancer)forCCL18andA1AT.AUROC,areaundertheROCcurve. Table2ResultsofCCL18andA1ATinexperimentalhematuriamodelDilution(volume ofwholeblood) Redblood cells(/hpf*) Concentration(meanSD)Urinaryprotein (dipsticktest) Hemoglobin(ng/ml)CCL18(pg/ml)A1AT(ng/ml) 0003.630.595.967.731170.7071.94negative 1/10,000(1 l)0.61.0174.2815.697.067.781227.311.22negative 1/2,000(5 l)2.41.22289.34479.076.809.611252.492.9230mg/dL 1/500(20 l)10.04.43633.392006.8716.6222.171258.4313.48100mg/dL 1/200(50 l)17.33.07898.05184.6731.9043.191273.165.53300mg/dL* hpf highpowerfield.Miyake etal.BMCUrology 2013, 13 :42 Page6of10 http://www.biomedcentral.com/1471-2490/13/42

PAGE 7

butnoturologictumors[27,28].A1AT,alsoknownas SERPINA1,isamemberofafamilyofserineproteases inhibitors.SpecificallyA1ATirreversiblyinhibitstrypsin, chymotrypsinandplasminogenactivator.Serpinsare knowntohavediversebutcriticalrolesinthecell,includingregulationofhomeostasis,cellularsurvivaland bloodclotting[29].Withintheoncologyliterature,reportsdescribegeneticaberrationsincancers,elevated levelsintheseraofcancerpatients,andsurvivaldisadvantageintumorsexpressingA1AT[30,31]. Inourearlystudies,wenotedthatthesebiomarkers hadarelativelyhighcorrelation(Spearmancorrelation coefficient>0.76)withurinaryhemoglobin.Giventhe confoundingeffectsofhematuriathatweandothers havedescribedfortheurine-basedBCadetectionassays BTAandNMP-22[6,10],wesetouttomoremethodicallyanalyzetheassociationofCCL18andA1ATwith BCabyanalyzingcohortsofurineandtissuesamples.In thisstudy,ELISAanalysisofurinesamplesfromacohortof308subjectsconfirmedourpreviousfindings [19]thatbothCCL18andA1ATaresignificantlyelevatedintheurinesofsubjectswithBCa. Toinvestigatethepotentialinfluenceofhematuria andotherfactorsontheperformanceofthesebiomarkersweemployedanexperimentalmodel.Although themodeldoesnotmimictheactualphysiologicalsituationexactly,itdoesenabletheidentificationofpotentialsourcesofspecificanalytesandtowhatextent incursionofbloodcomponentsintotheurinemay influencethedata.Wepreviouslyusedasimilarmodel approachtodemonstratethatBTAurinetestsprimarily detectaserum-basedprotein[6],andthatNMP-22 urinetestsmonitorcellularturnover,ratheraspecific bladdertumorantigen[10]. AnalysesfromthemodelandtheELISAassaysrevealeddifferentcharacteristicsforCCL18andA1AT withrespecttothembeingpotentiallyreliableBCadiagnosticbiomarkers.ForCCL18,thefirstobservationis thatthreehealthycontrolsampleshadverylowurinary CCL18levels(5.96pg/ml).ThemedianlevelofCCL18 inthenon-cancersamplesfromthe308subjectcohort wasalsoverylow(23.4pg/ml).Alowbaselinelevelisan advantagethatcanenableacleardistinctionbetween healthyanddiseasestateforagivenassay.Conversely,if thedifferentialbetweenurineandbloodislarge,thena smallamountofhematuriamayhaveasignificantimpact.Inthespikingexperiment,weobservedthat50 l ofbloodin10mlofurine,alevelthatwouldbetermed ‘ grosshematuria ’ inclinicaltests,raisedtheCCL18level inhealthycontrolsto31.9pg/ml,anincreaseof~5.3 fold.Inthe308subjectcohort,themedianlevelof CCL18was10-foldhigherinBCasubjects(230.5pg/ml vs. 23.4pg/ml).Theadditionofbenignandtumorcell lysatestotheurinesampleisdesignedtoindicate whetherthereleaseofubiquitouscellularfactorsmaybe thesourceofthebiomarker.Increasedcellularturnover istobeexpectedinamalignantcondition,andsosuch factors,forexampleNMP22,mayincreaseeventhough theyarenotactuallycancer-specificbiomarkers.Dueto thelowlevelsofCCL18inthehealthyurinesamples, theadditionofcelllysatesfrombenignandtumorcells didsignificantlyimpacttheCCL18levels.Finally,immunohistochemicalanalysisofbladdertumortissuesrevealedthatCCL18waspresentonlyintheinflammatory cellslocatedinthestroma.Nodifferenceinthenumber oftheimmunoreactivecellswasobservedinbenignversuscanceroustissue,oramongthevariousgradesor stagesofbladdercancer.Together,thesefindingssuggestthatCCL18monitoringisunlikelytobeareliable biomarkerforthenon-invasivedetectionofBCa. TheanalysisoftheA1ATbiomarkerrevealedopposite characteristicsforthemostpart.ELISAdataandtheexperimentalmodelconfirmedthatA1ATispresentat highlevelsinhealthyandnon-cancersubjecturinesamples.Themedianlevelinthe308subjectcohortwas120 ng/mlinnon-cancercases,rising5.5-foldto606.4 ng/mlinsubjectswithconfirmedBCa.Themedianlevel inthethreehealthyvolunteersampleswasintermediate (1,170ng/ml).Lowerlevelsinthenon-cancersubjects fromtheELISAdataismostlikelyduetodegradation withfreezingandstorageinthesesamplescomparedto thefreshurinesobtainedfromthevolunteers.Thehigh baselinelevelofA1ATinhealthyurinesamplesmaynot beidealfordiagnosticevaluation,buttheimpactof hematuriaonA1ATassayswouldbeexpectedtobeless pronounced.Accordingly,grosshematuriainthe Figure4 AnalysisofCCL18andA1ATbiomarkerperformance inanexperimentalmodel. Usingtheexperimentalmodel depictedinFigure1,urinarylevelsofCCL18andA1ATwere analyzedbyELISA.Theadditionofwholebloodresultedinan increaseinCCL18aswellasanincreaseinA1AT.Errorbarsindicate standarddeviations.*,significance( p <0.05)comparedtopooled urinesfromhealthysubjects.^,significance( p <0.05)comparedto correspondinglowerconcentration. Miyake etal.BMCUrology 2013, 13 :42 Page7of10 http://www.biomedcentral.com/1471-2490/13/42

PAGE 8

experimentalmodel(50 lofbloodinto10mlofurine) raisedA1ATto1,273ng/ml,anincreaseofonly9%. Comparethistothe>400%riseinCCL18inthesame experimentalmodelconditions.Theadditionofbenign orcancercelllysatesintheexperimentalmodelhadno impactontheA1ATlevels.PreviouspreliminaryresearchhaslinkedthepresenceofurinaryA1ATasa biomarkerforrenaldysfunction[32].Howeverwetook precautionsinourstudytominimizerenaldysfunction asaconfounderbyexcludingsubjectswithahistoryof renaldysfunctionaswellasexcludesubjectswithgrossly elevatedurinaryproteinlevels.Recently,researchers havereportedthatimpairedrenalfunction( i.e. ,reduced glomerularfiltrationrate)mayadverselyaffecturinary Figure5 AssessmentofCCL18andA1ATinbladdertissue.a) RepresentativeimmunostainingofbenignbladderforCCL18(top-left),lowgradenon-muscleinvasivebladdercancerforCCL18(top-middle),andhigh-grademuscleinvasivebladdercancerforCCL18(top-right).Red arrowsindicateCCL18-positivecellsinthestroma.CCL18stainingwaspresentonlyininflammatorycellsassociatedwiththestroma.Lower panelsareboxplotsofCCL18immunohistochemicalstainingintensityofbenignbladder vs. bladdercancer,low-grade vs. high-grade,non-muscle invasivebladdercancer(NMIBC) vs .muscleinvasivebladdercancer(MIBC).Errorbarsindicatestandarddeviations.Wilcoxonranksumtestwas usedtoassesssignificance. b) RepresentativeimmunostainingofbenignbladderforA1AT(top-left),low-gradenon-muscleinvasiveBCaforA1AT (top-middle),andhigh-grademuscleinvasiveBCaforA1AT(top-right).A1ATstainingwaspresentinthecytoplasmandstroma.A1ATstaining variedfromweakandfocaltostronganddiffuse.LowerpanelshowscolumnbargraphsofA1ATimmunohistochemicalstainingintensityof benignbladder vs. bladdercancer,low-grade vs. high-grade,NMIBCvs.MIBC. Miyake etal.BMCUrology 2013, 13 :42 Page8of10 http://www.biomedcentral.com/1471-2490/13/42

PAGE 9

biomarkersperformance[33,34].Thisisanexcellent pointandshouldbetakenintoconsiderationinfuture studies.However,wemuststressthatweconfirmed theseA1ATELISAresultsbyperformingimmunohistochemistryandthusdemonstratedthatA1ATispresent withinurothelialcells.WereportedA1ATreactivitywas epithelialinnormalbladdertissue,andstronglypositive inalltumorcells,specificallyhigh-gradecells.Although A1ATIHCmaynotbeparticularlyusefulindifferential histologicalevaluations,itdoessuggestthatthesource oftheincreasedA1ATobservedinBCasamplesismost likelybladdertumorcells.Thereleaseintotheurine maybeviasecretionortheturnoveroftumorcellsat theurineinterface.Thus,eventhoughnormalurinary levelsofA1ATarerelativelyhigh,measurementofthis biomarkerinthecontextofcancerdetectionmaybe worthwhile.Thegoodsepara tionbetweennon-cancer andBCaurinarylevelsandthelimitedinfluenceby secondarysourcessuggestthatvaliddiagnosticcutoffthresholdsmaybepossibleforurinaryA1AT monitoring. Whennovelurinarybiomarkersareproposed,the investigationofreliabilityinthefaceofpotentially confoundingeffectsiswarranted,especiallythoseintroducedintotheurinethroughbleeding,acommonpresentingfactorinbladdertumor-bearingpatients. FurtherstudiesintotheutilityofA1ATasabiomarker forthenon-invasivedetectionofBCaarecurrently underway.AdditionalfileAdditionalfile1: AnalysisofCCL18andA1ATbiomarker performanceinanexperimentalmodel. Usingtheexperimental modeldepictedinFigure1,urinarylevelsofCCL18andA1ATwere analyzedbyELISA.Theadditionofhighconcentrationofbenigncelllysate ormediumtohighconcentrationofcancercelllysateresultedinan increaseinCCL18.TheadditionofcellsorcelllysatedidnotalterA1AT levels.Errorbarsindicatestandarddeviations.*,significance( p <0.05) comparedtopooledurinesfromhealthysubjects.^,significance( p <0.05) comparedtocorrespondinglowerconcentration. Competinginterests CharlesJ.RosserandSteveGoodisonareofficersforNonagenBioscience Corporation.MakitoMiyake,ShantiRoss,AdrienneLawton,MyronChang, YunfengDai,LourdesMengual,AntonioAlcaraz,EvanGomesGiacoiahave noCOI. Authors ’ contributions MM:Acquisitionofdata.SR:Acquisitionofdata.AL:Acquisitionofdata.MC: Statisticalanalysis.YD:Statisticalanalysis.LM:Clinicalsamples,draftingof manuscript.AA:Clinicalsamples,draftingofmanuscript.EGG:Acquisitionof data.SG:Studyconceptanddesign,draftingofmanuscript.CJ:conceptand design,draftingofmanuscript,funding.Allauthorsreadandapprovedthe finalmanuscript. Acknowledgement ThisworkwassupportedbyresearchgrantsfromFlightAttendantMedical ResearchInstitute(CJR),FloridaDepartmentofHealthJamesandEstherKing TeamScienceAward10KT-01(CJR),FloridaStateJamesandEstherKing BiomedicalResearchAwardTechnologyTransferFeasibility1KF06(SG)and NationalCancerInstituteRO1CA116161(SG). Authordetails1CancerResearchInstitute,OrlandoHealth,Orlando,FL32827,USA.2DepartmentofPathology,OrlandoHealth,Orlando,FL32806,USA.3DepartmentofBiostatistics,UniversityofFlorida,Gainesville,FL32601,USA.4LaboratoryandDepartmentofUrology,HospitalClnic,Universitatde Barcelona,Barcelona,Spain.5NonagenBioscienceCorporation,Orlando,FL 32827,USA.6SectionofUrologicOncology,1400S.OrangeAve.,Orlando,FL 32806,USA. Received:29May2013Accepted:3September2013 Published:5September2013 References1.RifeCC,FarrowGM,UtzDC: Urinecytologyoftransitionalcellneoplasms. UrolClinNorthAm 1979, 6: 599 – 612. 2.CajulisRS,HainesGK3rd,Frias-HidvegiD,McVaryK,BacusJW: Cytology, flowcytometry,imageanalysis,andinterphasecytogeneticsby fluorescenceinsituhybridizationinthediagnosisoftransitionalcell carcinomainbladderwashes:acomparativestudy. DiagnCytopathol 1995, 13: 214 – 23.discussion24. 3.HeicappellR,WettigIC,SchostakM, etal : Quantitativedetectionofhuman complementfactorH-relatedproteinintransitionalcellcarcinomaofthe urinarybladder. EurUrol 1999, 35 (1):81 – 7. 4.KindersR,JonesT,RootR, etal : ComplementfactorHorarelatedprotein isamarkerfortransitionalcellcancerofthebladder. ClinCancerRes 1998, 4 (10):2511 – 20. 5.OgeO,KozaciD,GemalmazH: TheBTAstattestisnonspecificfor hematuria:anexperimentalhematuriamodel. JUrol 2002, 167 (3):1318 – 9. discussion1319 – 20. 6.MiyakeM,GoodisonS,RizwaniW,RossS,BartGrossmanH,RosserCJ: UrinaryBTA:indicatorofbladdercancerorofhematuria. WorldJUrol 2012, 30 (6):869 – 73. 7.GrossmanHB,MessingE,SolowayM,TomeraK,KatzG,BergerY,ShenY: Detectionofbladdercancerusingapoint-of-careproteomicassay. JAMA 2005, 293 (7):810 – 6. 8.SzenS,BiriH,SinikZ,KpeliB,AlkibayT,BozkirliI: Comparisonofthe nuclearmatrixprotein22withvoidedurinecytologyandBTAstattest inthediagnosisoftransitionalcellcarcinomaofthebladder. EurUrol 1999, 36 (3):225 – 9. 9.AtsN,EkiciS,OgeOO,ErgenA,HaselikG,OzenH: False-positiveresults oftheNMP22testduetohematuria. JUrol 2002, 167 (2Pt1):555 – 8. 10.MiyakeM,GoodisonS,GiacoiaEG,RizwaniW,RossS,RosserCJ: Influencing factorsontheNMP-22urineassay:anexperimentalmodel. BMCUrol 2012, 12: 23.doi:10.1186/1471-2490-12-23. 11.RosserCJ,LiuL,SunY,VillicanaP,McCullersM, etal : Bladdercancerassociatedgeneexpressionsignaturesidentifiedbyprofilingofexfoliatedurothelia. CancerEpidemiolBiomarkersPrev 2009, 18: 444 – 453. 12.UriquidiV,GoodisonS,CaiY,SunY,RosserCJ: MolecularBiomarker SignaturefortheNon-InvasiveDetectionofBladderCancer. CEBP;2012. 13.KreuninP,ZhaoJ,RosserC,UrquidiV,LubmanDM,GoodisonS: Bladder cancerassociatedglycoproteinsignaturesrevealedbyurinaryproteomic profiling. JProteomeRes 2007, 6 (7):2631 – 2639.Epub2007May23. 14.YangN,FengS,SheddenK,XieX,LiuY, etal : Urinaryglycoprotein biomarkerdiscoveryforbladdercancerdetectionusingLC-MS/MSand label-freequantification. ClinCancerRes 2011, 17: 3349 – 3359. 15.UrquidiV,KimJ,ChangM,DaiY,RosserCJ, etal : CCL18Inamultiplexurinebasedassayforthedetectionofbladdercancer. PLoSOne 2012, 7: e37797. 16.UrquidiV,GoodisonS,KimJ,ChangM,DaiY, etal : VEGF,CA9and angiogeninasaurinarybiomarkerforbladdercancerdetection. Urology 2012, 79: 1185 – 1188. 17.UrquidiV,ChangM,DaiY,KimJ,WolfsonED, etal : L-8asaurinary biomarkerforthedetectionofbladdercancer. BMCUrol 2012, 12: 12 – 15. 18.UrquidiV,GoodisonS,RossChangM,DaiY, etal : Diagnosticpotentialof urinaryalpha1-antitrypsinandapolipoproteinEinthedetectionof bladdercancer. JUrol 2012, 188: 2211 – 2214. 19.GoodisonS,ChangM,DaiY,UrquidiV,RosserCJ: Amulti-analyteassay forthenon-invasivedetectionofbladdercancer. PLoSOne 2012, 7 (10):e47469.doi:10.1371/journal.pone.0047469.Epub2012Oct19.Miyake etal.BMCUrology 2013, 13 :42 Page9of10 http://www.biomedcentral.com/1471-2490/13/42

PAGE 10

20.BossuytPM,ReitsmaJB,BrunsDE,GatsonisCA,GlasziouPP, etal : Towards completeandaccuratereportingofstudiesofdiagnosticaccuracy:the STARDinitiative. FamPract 2004, 21: 4 – 10. 21.PisitkunT,JohnstoneR,KnepperMA: Discoveryofurinarybiomarkers. MolCellProteomics 2006, 5: 1760 – 71. 22.WatanabeT,ShinoharaN,SazawaA,HarabayashiT,OgisoY,KoyanagiT, TakiguchiM,HashimotoA,KuzumakiN,YamashitaM,TanakaM,Grossman HB,BenedictWF: Animprovedintravesicalmodelusinghumanbladder cancercelllinestooptimizegeneandothertherapies. CancerGeneTher 2000, 7 (12):1575 – 80. 23.RossiMR,MastersJR,ParkS,ToddJH,GarrettSH,SensMA,SomjiS,NathJ, SensDA: TheimmortalizedUROtsacelllineasapotentialcellculture modelofhumanurothelium. EnvironHealthPerspect 2001, 109 (8):801 – 8. 24.FlussR,FaraggiD,ReiserB: EstimationoftheYoudenIndexandits associatedcutoffpoint. BiomJ 2005, 47: 458 – 472. 25.PepeMS,FengZ,JanesH,BossuytPM,PotterJD: Pivotalevaluationofthe accuracyofabiomarkerusedforclassificationorprediction:standards forstudydesign. JNatlCancerInst 2008, 100: 1432 – 8. 26.ChenJ,YaoY,GongC,YuF,SuS,LiuB, etal : CCL18Fromtumorassociatedmacrophagespromotesbreastcancermetastasisvia PITPNM3. CancerCell 2011, 19: 541 – 55. 27.ZohnySF,FayedST: Clinicalutilityofcirculatingmatrix metalloproteinase-7(MMP-7),CCchemokineligand18(CCL18)andCC chemokineligand11(CCL11)asmarkersfordiagnosisofepithelial ovariancancer. MedOncol 2010, 27: 1246 – 53. 28.WangQ,LiD,ZhangW,TangB,LiQQ,LiL: Evaluationofproteomics-identified CCL18andCXCL1ascirculatingtumormarkersfordifferentialdiagnosisbetween ovariancarcinomasandbenignpelvicmasses. Int:JBiolMarkers;2011. 29.NormandinK,PeantB,LePageC, etal : ProteaseinhibitorSERPINA1 expressioninepithelialovariancancer. ClinExpMetastasis 2010, 27: 55. 30.LindorNM,YangP,EvansI, etal : Alpha-1-antitrypsindeficiencyandsmoking asriskfactorsformismatchrepairdeficientcolorectalcancer: astudyfromthecoloncancerfamilyregistry. MolGenetMetab 2010, 99: 157.31.HamritaB,ChahedK,TrimecheM, etal : Proteomics-basedidentificationof alpha1-antitrypsinandhaptoglobinprecursorsasnovelserummarkers ininfiltratingductalbreastcarcinomas. ClinChimActa 2009, 404: 111. 32.Navarro-MuozM,IbernonM,BonetJ,PrezV,PastorMC,BaysB,CasadoVelaJ,NavarroM,AraJ,EspinalA,FluviL,SerraA,LpezD,RomeroR: Uromodulinand (1)-antitrypsinurinarypeptideanalysistodifferentiate glomerularkidneydiseases. KidneyBloodPressRes 2012, 35 (5):314 – 25. 33.TrojanB,TangA,ChandrapalJ,FilleurS,NeliusT: Theclinicalusefulnessof nuclearmatrixprotein-22inpatientswithend-stagerenaldiseaseand microscopichematuria. RenFail 2013, 35 (1):72 – 6. 34.TodenhferT,HennenlotterJ,WitstrukM,GakisG,AufderklammS,KuehsU, StenzlA,SchwentnerC: Influenceofrenalexcretoryfunctiononthe performanceofurinebasedmarkerstodetectbladdercancer. JUrol 2012, 187 (1):68 – 73.doi:10.1186/1471-2490-13-42 Citethisarticleas: Miyake etal. : InvestigationofCCL18andA1ATas potentialurinarybiomarkersforbladdercancerdetection. BMCUrology 2013 13 :42. 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 Miyake etal.BMCUrology 2013, 13 :42 Page10of10 http://www.biomedcentral.com/1471-2490/13/42