Hepatitis B virus X antigen and aflatoxin B1 synergistically cause hepatitis, steatosis and liver hyperplasia in transgenic zebrafish

ContentslistsavailableatScienceDirect

Acta

Histochemica

jou rn a l h o me p a g e :w w w . e l s e v i e r . d e / a c t h i s

Hepatitis

B

virus

X

antigen

and

aflatoxin

B1

synergistically

cause

hepatitis,

steatosis

and

liver

hyperplasia

in

transgenic

zebrafish

Jeng-Wei

Lu

_,

_Wan-Yu

_Yang

_,

_Yueh-Min

_Lin

_,

_Shiow-Lian

_Catherine

_Jin

_,

_Chiou-Hwa

_Yuh

b_{DepartmentofLifeSciences,NationalCentralUniversity,JhongliCity,Taoyuan,Taiwan} c_{DepartmentofPathology,ChanghuaChristianHospital,ChanghuaCity,ChanghuaCounty,Taiwan} d_{InstituteofBioinformaticsandStructuralBiology,NationalTsingHuaUniversity,Hsinchu,Taiwan} e_{DepartmentofBiologicalScienceandTechnology,NationalChiaoTungUniversity,Hsinchu,Taiwan}

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received27December2012

Receivedinrevisedform14February2013 Accepted18February2013

Keywords:

HepatitisBvirusXantigen AflatoxinB1

Liverdiseases Hyperplasia Transgeniczebrafish

a

b

s

t

r

a

c

t

AflatoxinB1(AFB1)andthehepatitisBvirusXantigen(HBx)arelinkedtotheformationofliver dis-easesandhepatocellularcarcinoma(HCC).Theaimofthisstudywastoinvestigatethesynergisticeffects betweenHBxandAFB1incausingliverdisordersusingatransgeniczebrafishanimalmodel. Histo-pathology,Periodicacid-Schiff(PAS)staining,Siriusredstaining,TdT-mediateddUTPNickEndLabeling (TUNEL)assay,immunohistochemistry,andquantitativereversetranscriptase-polymerasechain reac-tion(Q-RT-PCR)wereusedtoexaminetheliversoftheHBxtransgenicfishinjectedwithAFB1.Wefound thatHBxandAFB1synergisticallypromotedsteatosisasindicatedbyhistopathologicalexaminationsand theincreasedexpressionoflipogenicfactors,enzymes,andgenesrelatedtolipidmetabolism.Moreover, treatmentofAFB1inHBxtransgenicfishacceleratedthedevelopmentofliverhyperplasiaandenhanced theexpressionofcellcyclerelatedgenes.PCNAwasco-localizedwithactivecaspase3proteinexpression inHBxzebrafishliversamplesandhumanHBVpositiveHCCsamplesbydoublefluorescence immuno-staining.Finally,wefoundthatinhumanpatientswithliverdisease,significantglycogenaccumulatedin theinflammation,cirrhosisstage,andallcasesofhepatocellularandcholangiocellularcarcinomashowed amoderatecytoplasmicaccumulationofglycogen.OurdatademonstratedasynergisticeffectofAFB1 andHBxontheregulationoflipidmetabolismrelatedgenesandcellcycle/division-relatedgeneswhich mightcontributetoenhancedsteatosisandhyperplasiaat5.75months.

© 2013 Elsevier GmbH. All rights reserved.

Introduction

Hepatocellularcarcinoma(HCC) is themostcommon malig-nancyinEastAsiaandAfrica.Hepatocarcinogenesisisamultistep processinvolvingsteatosis,fibrosis,cirrhosis,adenoma,and car-cinoma(Tarantinoetal.,2007).HCCiscommonlyassociatedwith chronichepatitisBvirus(HBV)orhepatitisC virus(HCV) infec-tionsorchronicexposuretoAFB1andisacomplicationofalcoholic cirrhosis(Morganetal.,2004;McGlynnandLondon,2005;Seeff andHoofnagle, 2006;El-SeragandRudolph, 2007;Marreroand Marrero,2007).Treatmentoptionsforlivercancerarecurrently limited,andtheprognosisisgenerallypoorandfiveyearsurvival rateaftersurgeryislimited(Luetal.,2011).Therefore,HCCremains oneofthedeadliestcancers,witha5-yearsurvivalrateofonly 5%(El-Seragetal.,2001).Manyriskfactors,suchastheHBxand AFB1-inducedgeneticandepigenetic alterationsinhepatocytes,

∗ Correspondingauthorat:InstituteofMolecularandGenomicMedicine,National HealthResearchInstitutes,Zhunan,Miaoli,Taiwan.

E-mailaddress:[email protected](C.-H.Yuh).

resultinthetransformationintoneoplasia(Roberts,2008;Siegel etal.,2010).However,thesynergiesbetweenthevariousrisk fac-torsthatresultinHCCareunclear.Theaimofpresentstudywasto identifythepotentialsynergisticeffectsbetweenHBxandAFB1in promotingliverdiseaseandHCCdevelopmentandtounderstand theunderlyingmechanismusingazebrafishmodel.

Chronic HBV infectionhas beenshown tobe strongly asso-ciatedwithHCC(Beasleyet al.,1981;Robinson, 1992;Brechot, 2004).OneoftheviralantigensofHBVthathasshownan onco-genicroleisHBx,whichispresentinthecytoplasmand,toalesser extent,thenucleusofhepatocytes(Kew,2011).HBxcommunicates withavarietyofhosttargetsandmediatesmanyopposing cellu-larfunctions,includingcellcycleregulation,apoptosis,signaling andtranscriptionalregulationandaffectsmanymolecules includ-ingcytoskeletalandcelladhesionmolecules,oncogenesandtumor suppressorgenes(Pengetal.,2005;Pangetal.,2006).AFB1isa mycotoxinproduced byAspergillus flavusand relatedfungithat growoncropsincludingcorn,riceandpeanut(BennettandKlich, 2003).Itismetabolizedpredominantlyintheliverto AFB1-8,9-exo-epoxide,whichformsapromutagenicAFB1-N7-guanineDNA adductthatresultsinG–Ttransversionmutations(WildandTurner, 0065-1281/$–seefrontmatter © 2013 Elsevier GmbH. All rights reserved.

J.-W.Luetal./ActaHistochemica115 (2013) 728–739 729 2002).AFB1hasgainedmuchattentionasoneofthemost

preva-lentcarcinogens(Akriviadisetal.,1998).Mutationalinactivation ofthep53genehasbeenwelldescribedtobeoneofthemolecular mechanismsinvolvedinHCCpathogenesis,especiallyin geograph-icalareaswheredietaryAFB1exposureismoreserve(Hsuetal., 1991;Aguilaretal.,1994;Pangetal.,2006).Differentanimal mod-elshavebeenusedtoexaminetheinteractionbetweenthehepatitis virusandAFB1exposure,but,ingeneral,thesehavesufferedfrom anumberoflimitations(Wildetal.,1993).IntheHBxandHCV transgenicmice,AFB1treatmentinducedsignificantlymoreliver tumorsthaninwild-type mice(Lianetal.,2006;Jeannotetal., 2011).Moreimportantly,hepaticovalcells(HOC)whichare con-sideredasliverstemcells,aftertransfectionwithHBxgeneand implantationintomice,producetumorsintheanimaltreatedwith aflatoxinB1(16.7%)intheliver(Lietal.,2011).Thisresultprovided evidenceforthecombinedeffectbetweenHBxandAFB1.

Inthepresentstudy,weusedtheliverfattyacid-binding pro-tein(l-fabp)promotertodriveHBxexpressionspecificallyinthe liver(Andreetal.,2000;Denovan-Wrightetal.,2000).The pro-teinproductofl-fabpbindstolongchainfattyacidsinhepatocytes andisresponsiblefortheirtrafficking.Thezebrafishl-fabp pro-moter faithfully drives GFP expression in the liver (Her et al., 2003).ThehepatotoxinAFB1wasadministeredviaintraperitoneal injection,asshownpreviously.Tothebestofourknowledge,the collaborativeeffectofHBxandAFB1incausingliverdiseaseusing a transgeniczebrafish model hasnot beenreported. Therefore, weusedhistopathology,PASstaining,Siriusredstaining,TUNEL assay,immunohistochemistry,andQ-RT-PCRinthepresentstudy toexaminethesynergisticeffectofHBxandAFB1inpromoting liverdiseaseandexplorethepossiblemechanismsthatunderlie thosedisorders.

Materialsandmethods

Zebrafishmaintenance

Zebrafish were maintained in the Zebrafish core facility at NTHU-NHRI(ZeTH)accordingtoestablishedprotocols(Tsengetal., 2011). TheTg(l-fabp:GFP-mCherry) and Tg(l-fabp:HBx-mCherry) transgenicembryos,larvae,andadultfishweremaintainedat28◦C undercontinuousflowintheZebrafishCoreFacilitywithautomatic controlfora14-hlightand10-hdarkcycle.Allexperiments involv-ingzebrafishwereapprovedbytheInstitutionalAnimalCareand UseCommittee(IACUC)oftheNHRI(NHRI-IACUC-101005-A).The liver-specifictransgenicfishweregeneratedasdescribed previ-ously(Kwanetal.,2007).

AFB1treatment,livertissuecollection,andparaffin sectioning

Accordingtoa previous study,AFB1 canbe dissolvedin 8% DMSO,and thevehicleitselfdoesnothaveanytoxicitytoliver oranimal survival (Luyendyk et al.,2003).AFB1 (Acros Organ-ics,ThermoFisherScientific,Geel,Belgium)wasdissolvedin8% DMSOto a final concentration of 0.15␮g/␮l and injected with 10␮l.AFB1inducesHCCinamousemodelwithasingledoseof 3mg/kginjectedintraperitoneally(Hirumaetal.,2001).Because eachzebrafishweighsapproximately0.5g,weinjected1.5␮gof AFB1per fish.Transgenic zebrafish of2.75monthsof agewere injectedintraperitoneallywiththeAFB1solution(1.5␮g/10␮l). Transgenicfishweresacrificedat1,3,6,9and12weekspost-AFB1 injection.Thetissueswerefrozeninliquidnitrogenimmediately afterdissectionandstoredat₋₇₀◦CforlaterRNAextraction.For histochemicalanalysis,livertissueswerefixedina10%formalin solution(Sigma–Aldrich,St.Louis,MO,USA).Thefixedtissuewas

embeddedinparaffin,sectionedat5-␮mthickness,andmounted onpoly-l-lysinecoatedslides.Followingdeparaffinization,the sec-tionswere stainedwith hematoxylin and eosin. Diagnosis was performedby using a single-blind evaluation ofall samplesby trainedpathologists(Dr.Yueh-MinLin,ChanghuaChristian Hos-pital,Changhua,Taiwan).Thereare57samplesrepresentingfive differentstagesforfourgroups(noAFB1+noHBx,HBxalone,AFB1 aloneandAFB1+HBx).Thenumbersofbiologicalreplicateswere variableforeachstageandgroup,andthedetailsarelistedin Sup-plementaryTable1.

Histochemicalanalysis

Paraffin-embeddedtissuesectionsslidesweredeparaffinized. Theslidesweretreatedwith3%H2O2toblockendogenous per-oxidase activity, and heated in 10mM citrate buffer at 100◦C for 30minto retrieve antigen. The sectionswere incubated at 4◦Covernightwithdifferentprimaryantibodiesincludingrabbit anti-active caspase 3 (1:100dilution; BD Biosciences, Le Pont-de-Claix, France), mouseanti-PCNA (1:100 dilution;SantaCruz Biotechnology,SantaCruz,CA,USA),mouseanti-HBx(1:100 dilu-tion; Abcam, Cambridge, MA, USA), and mouse anti-AFB1-DNA adduct(1:100dilution;NovusBiologicals,LLC,Littleton,CO,USA). Afterwashing with1× PBS,thesectionswereincubatedwitha secondaryantibodyfor30minatroomtemperature,which was followedbydevelopmentusingtheLiquidDABSubstrateKit (Invi-trogen,Carlsbad,CA,USA). Tissuesectionswerecounterstained withhematoxylin,dehydrated,clear,mountedandexaminedby lightmicroscopy.Fordoublefluorescentimmunohistochemistry, sectionswereincubatedat4◦Covernightwithprimary antibod-ies, including rabbit anti-active caspase 3 (1:100 dilution; BD Biosciences,LePont-de-Claix, France), mouse anti-PCNA(1:100 dilution;Santa Cruz Biotechnology,SantaCruz, CA, USA). After washingwith1xPBS,thesectionswereincubatedwithgoat anti-rabbitIgGorgoatanti-mouseIgGrhodamine(FRITC) secondary antibodies(1:100dilution;both fromJackson ImmunoResearch (West Grove, PA, USA) for 60min atroom temperature.Tissue sectionsweremountedandexaminedbyfluorescentmicroscopy. SiriusredstainingwasperformedusingthePicrosiriusRedStain Kit(Polysciences,Warrington,PA,USA)todetectcollagenfibers. TheglycogenaccumulationwasdetectedusingthePeriodic acid-Schiff(PAS)StainKit(Polysciences).TUNELassaywasperformed using the In Situ Cell Death Detection Kit, Fluorescein (Roche, Indianapolis, IN, USA) according to themanufacturer’s instruc-tions.Thetissuesectionsthenweredehydrated,cleared,mounted and examined by light or fluorescent microscopy. We scored the Sirius red staining and TUNEL assay results by intensity, which ranged from negative (0) to minimal (1), moderate (2), and maximal (3).The stainingintensity forPAS, caspase3 and PCNA were scoredas 0 (0–15%), 1 (16–25%), 2 (26–50%), or 3 (76–100%)accordingtothepercentageofpositivelystainedcells (Fig.S1).

Humanliverdiseasespectrumtissuemicroarray

Toinvestigatetherelationshipofglycogenaccumulationwith humanliverdisease,wepurchasedthehumanliverdiseasetissue arrayforPASstaining.Theliverdiseasespectrumtissue microar-ray(catalog#:BC03002)waspurchasedfromUSBiomax(Rockville, MD,USA).Eacharraycontains15primaryHCCs,8 cholangiocellu-larcarcinomas,8samplesoflivercirrhosis,5liversampleswiththe hepatitisvirusand2normaladjacentliversamples,duplicatecores percase.Periodicacid-Schiff(PAS)stainingforglycogen accumu-lationwasperformedasdescribedintheprevioussection.

RNAisolationandquantitativeRT-PCR

Total RNA wasisolated by theRNAspin MiniRNA Isolation Kit(GEHealthcare,Pittsburgh,PA,USA).RNA(1␮g)wasreverse transcribedintocDNAusingtheHighCapacityRNA-to-cDNAKit (AppliedBiosystems,Carlsbad,CA,USA).Theresultingfirst-strand cDNA was used as a template for qualitative PCR in triplicate using the SYBR Green Q-PCR Master Mix Kit (Applied Biosys-tems).OligonucleotidesPCRprimerpairsweredesignedtocross intron-exonboundariesfrompublishedsequencesintheGeneBank databaseusingPrimer3.PrimerswerepurchasedfromaTaiwanese company(MissionBiotech,Taipei,Taiwan,ROC)andthe oligonu-cleotideswerecolumnpurifiedandsaltfree.

Wechosesomemarkersforgenesinvolvedinthehepaticlipid accumulation,fibrosis,cellcycleandtumorfromliteraturebased onapreviousstudyonHBxtransgenicfish(Rekhaetal.,2008; Shiehetal.,2010).ThesequenceofprimersforQ-PCRisprovided inSupplementaryTable2.Thespecificityoftheamplification prod-uctswasconfirmedbysizeestimationsona1.2%agarosegeland byanalyzingtheirmeltingcurves.Aprimerdimerwasruledout byperforminga“notemplate”controlandanalyzingthe dissoci-ationcurve,andthewithouttemplate,nocycle(Cq)valuescould bedetermined.Afternormalizationtointernalcontrolactinor18S, theexpressionratiobetweentheexperimentalandcontrolgroups wascalculatedusingthecomparativeCtmethod.Allexperiments wereperformedintriplicate,andthemeansofthreevaluesare pre-sented.AtleastthreeindependentsampleswereusedforQ-PCR, andthestandarderrorwascalculatedandincorporatedintothe presenteddataasmedians±standarderror.Forabsolutemolecules analysisshowninFig.1,theRNAmoleculesperembryowas cal-culatedusingthestandardcurveasdescribedearlier(Chanetal., 2009;Tsengetal.,2011).Briefly,aseriesofknownamountofGFP fragmentwasusedforQ-PCRandstandardcurveofRNAmolecules verseCtwasgenerated,theCtofHBxfromQ-PCRanalysiswas convertedintoRNAmoleculesusingthestandardcurve.

Statistics

StatisticalanalysisofhumanglycogenaccumulationandQ-PCR resultswasperformedusingtheunpairedStudent’st-test. Cumu-lativefrequencyofthepathologychangesintransgeniczebrafish wascalculatedbyKaplan–Meieranalysis(Zhuetal.,2012).Inall statisticalanalyses,ap-valueoflessthan0.05wasconsideredtobe statisticallysignificant.

Results

ExpressionoftheHBx-mCherryfusionproteinand formationofAFB1-DNAadductsintransgenicfish

Toinvestigate theinterplay betweenHBx and AFB1in liver disease, we used the liver-specific l-fabp promoter to drive expressionof the GFP-mCherry orHBx-mCherry fusion protein in zebrafish. Two different sizes of l-fabp were used: 2264bp (attB4-l-fabp-F1+attB1r-l-fabp-R for GFP-mCherry) and 2870bp (attB4-l-fabp-F+attB1r-l-fabp-R for HBx-mCherry), and both of themcandrivethefusionproteinintheliverasdescribedbelow. We first examined the expression of the fusion protein using fluorescence microscopy. The livers of three-month-old adults exhibitedgreenandredfluorescencefortheF1foundercarrying l-fabp:GFP-mCherryorredfluorescencefortheF1foundercarrying l-fabp:HBx-mCherry(Fig.1A).UsingQ-PCRforabsolute quantifica-tion,wefoundthattheHBxRNAexpressionlevelintheliversof HBxtransgenicfishdecreasedsteadilyovertime(Fig.1B),whichis similartotheresultsshownintheHBxtransgenicmouse(Luetal., 2012).TheexpressionoftheHBxproteinwasfurtherverifiedby

immunohistochemistryusingananti-HBxantibody.Incontrastto theEGFP-mCherrytransgenicfishthatshowednoHBxexpression, hepatocytesfromHBxtransgenicfishshowedastrongexpression oftheHBxproteinfrom3to5.75monthsofage(Fig.1C).Inatotal of57liversamples,asuccessfulAFB1treatmentwasverifiedusing anantibodyspecificfortheAFB1-DNAadduct,whichwas continu-ouslypresentinthehepatocytesateven12weekspost-treatment, butnotintheHBxtransgenicfishthatwerenottreatedwithAFB1 (Fig.1D).

HistopathologicalexaminationofsynergisticeffectofHBx andAFB1inthetransgeniczebrafishmodel

Previous studies showed that HBx overexpression in the zebrafish liver caused hepatic steatosis and liver degeneration, butnosignsofHCC(Shiehetal.,2010).Asynergisticinteraction betweenchronicHBVinfectionand AFB1in hepatocarcinogene-siswaspreviouslyreported(Kew,2003).Toincreasethechance of HCC formation in zebrafish, we intraperitoneally injected 2.75-month-oldtransgenicfish(expressingGFP-mCherryor HBx-mCherry) with AFB1 and collectedthe livers at 1, 3, 6, 9 and 12weekspost-injection(wpi)for analysis.A totalof 57 samples were collected from different treatment groups, and detailed informationregardingthehistopathologicalexaminationofthese specimensislistedinSupplementaryTable2.

Compared to the control Tg(l-fabp:GFP-mCherry) fish (Fig. 2A1–A5), hepatocytes from the Tg(l-fabp:HBx-mCherry) fishdisplayedsteatosis,inflammation,andballoondegeneration from3to5.75monthsofage(Fig.2B1–B5).Theseobservationsare consistentwithpreviousreports(Shiehetal.,2010).AFB1alone promotedhepatitisorsteatosisdevelopmentstartingfrom6wpi (or4.25monthsof age)(Fig.2C1–C5).Moreover,AFB1and HBx synergistically promotedhepatocyte steatosisas early as1wpi (Fig.2D1),andthefishdevelopedhepatitisandsteatosisfrom3 to 9wpi(Fig. 2D2–D4),while finally hyperplasia wasobserved at5.75monthsofage,which correspondsto12weeks afterthe AFB1treatment(Fig.2D5).Thecumulativefrequencyanalysisof theH&EstainingresultsfrommultipleHBxtransgenicfishlinesis summarizedinFig.2E.AlthoughHBxandAFB1alonecaninduce hepatitis (Fig.2E1)and steatosis, an increasingnumber of fish developedsteatosiswiththepresenceofbothfactors(Fig.2E2). Approximately20%oftheHBxtransgenicfishtreatedwithAFB1 developedhyperplasia(Fig.2E3)at5.75monthsofage.

For additionalpathological changes,theliversectionsof the AFB1-treatedHBxtransgenicfishwerestainedwithSiriusredand periodicacid-Schiff(PAS)staintodetectthedepositionof extracel-lularmatrixandglycogenaccumulation,respectively.Theterminal deoxynucleotidyltransferasedUTPnickendlabeling(TUNEL)assay wasperformedtoexamineapoptosis,stainingwithanantibodyto activecaspase3wasusedtodetectactivatedapoptosis,and stain-ingwithanantibodytotheproliferatingcellnuclearantigen(PCNA) wasusedtoanalyzecellproliferation.Multiplesamplesfromeach stageofthedifferentzebrafishlineswereusedfortheseanalyses, andthescoringwasstandardizedbycomparingthepercentageof cellswiththestainingasshowninFig.S1.

Accordingto theresult of PASstaining, HBxoverexpression combinedwiththeAFB1treatmentcausedanincreaseinglycogen accumulation,althoughHBxandAFB1alonecouldalsoenhance glycogenaccumulation(Fig.3A).AsindicatedbySiriusredstaining, HBxandAFB1alonecanpromotefibrosis,andtheredoesnotappear tobeanysynergisticeffectbetweenthem(Fig.3B).Cell prolifera-tion,asindicatedbynuclearPCNAstaining,graduallyincreasedin theHBxandAFB1alonefish,however,AFB1treatedHBxtransgenic fishhadthestrongestPCNAstainingatscore3(Fig.3C).FromTUNEL assay,zebrafishoverexpressingHBxorthatwastreatedwithAFB1 hadsignificantly moreapoptoticcells thanthecontrolfish,and

J.-W.Luetal./ActaHistochemica115 (2013) 728–739 731

Fig.1.ExpressionoftheHBx-mCherryfusionproteinandHBxandtheformationofAFB1-DNAadductinthetransgenicfish.(A)TheEGFP-mCherryfusionproteinwas expressedintheliver,asindicatedbyboththegreenandredfluorescencedetectedintheliverofawild-typefishcontainingthel-fabp:GFP-mCherrytransgene.The HBx-mCherryfusionproteinwasexpressedintheliverofthel-fabp:HBx-mCherrytransgenicfish,asindicatedbytheredfluorescence.(B)Q-RT-PCRanalysisoftheexpression ofHBxRNAinthelivertissueofHBx-mCherrytransgenicfishwithout(redcolumn)orwithAFB1treatment(bluecolumn)at3,3.5,4.25,5and5.75monthsofage.(C) ImmunohistochemicalanalysisoftheexpressionoftheHBxproteininhepatocytesfromthe3-,3.5-,4.25-,5-and5.75-month-oldHBx-mCherrytransgenicfishandcontrol EGFP-mCherrytransgenicfish(400×).(D)ImmunohistochemistryforthepresenceoftheAFB1-DNAadductinthehepatocytesofHBx-mCherrytransgenicfishat1,3,6,9 and12weekspost-AFB1injection(wpi),whichwerecomparedtothefishwithouttheAFB1treatment(400×).Scalebars=50␮m.(Forinterpretationofthereferencesto colorinthisfigurelegend,thereaderisreferredtothewebversionofthearticle.)

Fig.2.HistopathologyofhepatocytesaffectedbyHBxandAFB1atfivedifferentstages.Liversectionswerepreparedfrom(A)GFP-mCherrytransgenicfish,(B)HBxtransgenic fish,(C)AFB1-treatedGFP-mCherryfish,and(D)AFB1-treatedHBxtransgenicfish(400×or200×).Thefishwere3,3.5,4.25,5or5.75monthsoldor1,3,6,9,or12weeks post-AFB1injection(wpi).Scalebars:50␮m.(E)Cumulativefrequencyofhepatitis(E1),steatosis(E2)orhyperplasia(E3)frommultiplelinesoftransgenicfishbyKaplan–Meier analysis.GreenlineindicatedAFB1treatedGFP-mCherrytransgenicfish,bluelinedenotedHBxtransgenicfishandredlinesrepresentedAFB1treatedHBxtransgenicfish. (Forinterpretationofthereferencestocolorinthisfigurelegend,thereaderisreferredtothewebversionofthearticle.)

thereappearedtobenosynergisticeffectbetweenthesetwo con-ditions(Fig.3D).Thecaspase3stainingresultssimilartotheTUNEL assay(datanotshow).Takentogether,thesedatasuggestthatHBx andAFB1haveasynergisticeffectonglycogenaccumulationand cellproliferation.However,whileeachconditioncaninduce fibro-sisandapoptosis,therewasnoadditiveeffectwhenthetwofactors werecombined.

Glycogenaccumulationoccurredintheinflammationstage andreappearedinthehyperplasiastage

Wefoundsignificantglycogenaccumulationinthehepatocytes fromboththeearlystage(1,3,6weekspost-injection)ofHBx trans-genicfishandthelatestage(9and12weekspost-injection)ofHBx transgenicfishtreatedwithAFB1.Toidentifythecorrelationof glycogenaccumulationwithhumanliverdisease,weusedthe tis-suearrayforPASstaining.PASstainresultsshowedthatglycogen accumulationatdifferentlevelsindifferentsamplesthatcanbe scoredintofourclassesaccordingtothepercentageofpositively

stainedcells(Fig.4A).Inhumanpatientswithliverdisease,the correlationbetweenglycogenaccumulationandliverdiseasewas calculatedbyt-testand thefollowing stagesshowedsignificant differences(P<0.001):inflammation,cirrhosis,stagesIIandIIIof hepatocellularandcholangiocellularcarcinoma.However,stageI ofhepatocellularandcholangiocellularcarcinomashoweda mod-erate cytoplasmic accumulation of glycogen. By comparing the PASstainingfromhumansandzebrafishwithliverdiseases,we observeddramaticallysimilarphenomenafortheglycogen accu-mulationthatoccurredintheinflammationstageandreappeared inthehighlyproliferativestage(Fig.4B).

SynergisticeffectofHBxandAFB1ontheexpressionof lipogenicfactorsandcellcycle-relatedgenes

WefurtherusedQ-PCRtoquantifytheexpressionofmarker genesthat representdifferentliverdiseases(Fig.5).The choice oflipogenicgenesinvolvedinthehepaticlipidaccumulationwas followedpreviousstudyonHBxtransgenicfish(Rekhaetal.,2008;

J.-W.Luetal./ActaHistochemica115 (2013) 728–739 733

Fig.3.CumulativefrequencyofhistochemistryanalysisfrommultiplelinesoftransgenicfishbyKaplan–Meieranalysis.(A)Statisticalanalysisofperiodicacid-Schiff(PAS) stainingforglycogenaccumulation.(B)StatisticalanalysisofthefibrosisanalysisbySiriusredstaining.(C)IHCstainingfornuclearPCNA.(D)Statisticalanalysisofthe TUNELassayforapoptosis.GreenlineindicatedAFB1treatedGFP-mCherrytransgenicfish,bluelinedenotedHBxtransgenicfishandredlinesrepresentedAFB1treated HBxtransgenicfish.Theintensityofeachstainingwasshownontopasscore1–4,respectively.Thefivestagesare3,3.5,4.25,5and5.75-month-oldfishor1,3,6,9,and 12weekspost-AFB1injection(wpi).(Forinterpretationofthereferencestocolorinthisfigurelegend,thereaderisreferredtothewebversionofthearticle.)

Shiehetal.,2010).Correspondingtothepathologicalfeaturesin whichsteatosisdevelopedatbetween5and5.75months,thePCR resultsshowedthatmostofthelipidmetabolismrelatedgenes areup-regulatedmoresignificantlyinHBxandAFB1togetherthan eachfactoralone(Fig.5A–D).

Important transcription factors involved in fatty acid syn-thesisincludes cebp␣ (CCAAT enhancer binding protein-alpha) whichisabZIPtranscriptionfactor,whichcanbindasa homod-imertocertain promoters and enhancers,and transcriptionally activated during adipocytes differentiation, ppar␥ (peroxisome proliferator-activatedreceptorgamma)whichisnuclearreceptor

proteinthatfunctionastranscriptionfactorregulatinggenesfor fattyacidstorageandglucosemetabolism,srebp1(sterol regula-torybindingprotein-1)whichencodesatranscriptionfactorthat bindstothesterolregulatoryelement-1(SRE1)ofgenesinvolvedin sterolbiosynthesis,andchrebp(carbohydrate-response element-bindingprotein)which is a transcriptionalregulatorof glucose and lipid metabolism. The lipogenic factors (ppar␥, chrebp and srebp1areup-regulatedat5M,exceptthecebp␣whichshowed muchlowerexpressioninallthreegroup(Fig.5A),andtheHBx togetherwith AFB1 exhibit thesignificant overexpression than eachfactoralone.

Fig.4. Periodicacid-Schiff(PAS)staininganalysisofhumanliverdiseasearrayatvariousstages.(A)Theglycogenaccumulationfromdifferenthumanliverdiseasesranged fromnormal,inflammation,cirrhosis,andstagesI–IIIofhepatocellularcarcinoma(HCC)andcholangiocarcinoma(CC)(200×).Scalebars=50␮m.(B)Eachstainresultwas evaluatedandgivenascoreof0–4,andtheaverageofthescoreswasobtainedfrommultiplesamplesofthesamestageforthePASstaining.Eightyhumanlivertissueswere purchasedasahumantissuearrayformat.Thesignificancewascalculatedusingat-testandshownas“**forP<0.001”.

Lipogenicenzymeincludekey enzymeforfattyaciddenovo synthesis-fasn(fattyacidsynthase),whichisamulti-enzymethat playsakeyroleinfattyacidsynthesis;andacacb(ACACBacetyl-CoA carboxylasebeta)whichconverts acetyl-CoAtomalonyl-CoA,is thecommittedstepofthefattyacidsynthesispathway.Lipogenic enzymeinclude agapt(acyl-CoA: 1-acylglycerol-sn-3-phosphate acyl-transferase),pap(phosphatidicacidphosphatase)whichisa keyenzymeintriglyceridebiosynthesis,anddgat2(diacylglycerol O-acyltransferase2)whichcatalyzestheformationoftriglycerides

fromdiacylglycerolandAcyl-CoA.Amongthem,acacb,agapt,and papareallup-regulatedinHBx+AFB1morethaneachfactoralone, butthedgat2doesnotshowsynergisticeffect,andthefasnwasnot highlyexpressedinallthreegroups(Fig.5B).

ThePPAR-␥targetgenesincludeCD36,whichplaysacritical roleinfattyaciduptake;ucp2(mitochondrialuncouplingprotein 2)whichcanprotectagainstoxidativestress-associatedneurogenic hypertension; caveolin belongs to a family of integral mem-braneproteinswhich aretheprincipalcomponentsof caveolae

J.-W.Luetal./ActaHistochemica115 (2013) 728–739 735

Fig.5. QuantitativeRT-PCRanalysisofselectedmarkergenesrepresentingdifferentliverdiseasesintheHBx,AFB1andHBx+AFB1fish.(A)Lipogenicfactor,(B)lipogenic enzyme,(C)PPAR-gammatarget,(D)lipidbeta-oxidation,(E)fibrosismarkers,(F)cellcycle/divisionfromfishat9wpi/5Mand(G)tumormarkersand(H)cellcycle/division fromfishat12wpi/5.75M.Therearethreegroupsoffish,asfollows:HBxtransgenicfish(firstgroup),AFB1-treatedGFP-mCherrytransgenicfish(secondgroup)and AFB1-treatedHBxtransgenicfish(thirdgroup).EachCtwasfirstnormalizedagainst␤-actinandthencomparedtotheCtfromtheEGFP-mCherryfish.Thedelta–deltaCtvalues werethenconvertedtofolddifferences.Multiplereplicateswereperformed,andthemeansareshownwithstandarddeviations.Thesignificancewascalculatedusinga t-testandshownas“*”whencomparedtoGFP-mCherrytransgenicfish,as“+”whenHBx+AFB1wascomparedtoHBxonly,oras“#”whenHBx+AFB1wascomparedto AFB1only.Thesignificanceisindicatedas*,+,or#for0.01<P≤0.05;**,++,or##for0.001<P≤0.01;or***,+++,or###forP≤0.001.

membranesand involved in receptor-independent endocytosis; cfdl(complementfactorDoradipisin)istheadipocyteserine pro-tease;adipoql2(adiponectin)isaproteinhormonethatmodulates anumberofmetabolicprocesses,includingglucoseregulationand fattyacidcatabolism.Amongthem,ucp2,carveolinandadipoql2

showedsignificantsynergisticeffectbetweenHBxandAFB1.The increasedofcd36isnotsignificant,andtheexpressionofadipisin ishighinallthreegroups(Fig.5C).

Thelipidbeta-oxidationprotein(ppar-␤,cpt1(carnitine palmi-toyltransferase1)whichmediatesthetransportoflong-chainfatty

acidsacrossthemembranebybindingthem tocarnitine;l-pbe (l-peroxisomalbifunctionalenzyme)whichisthesecondenzyme of thel-hydroxy-specific classical ␤-oxidation system,cyp4a10 (cytochromeP450,family 4,subfamilya,polypeptide10)which isthemajorenzymesinvolvedindrugmetabolismand bioactiva-tion,andacox3(fattyacyl-CoAoxidase(AOX),thefirstenzymeof thel-hydroxy-specificclassical␤-oxidationsystem.Allofthelipid beta-oxidationgeneswereupregulatedat5monthsofageinthe HBxorAFB1alone,andseemsnosynergisticeffectbetweenHBx andAFB1(Fig.5D).

The choiceof fibrosis markersand tumormarkers wasalso based on a previous study using transgenic fish for liver dis-easestudy(Rekhaetal.,2008).Fibrosismarkerswhichincluded col1a1(collagen,typeI,alpha1),atgfa(connectivetissuegrowth factor(CTGF)which mirrorssomeoftheeffectsof TGFbetaon skinfibroblasts,suchasstimulationofextracellularmatrix pro-duction,chemotaxis,proliferationandintegrinexpression;hpse (heparanase)isanenzymethatactsbothatthecell-surfaceand withintheextracellularmatrixtodegradepolymericheparan sul-fatemoleculesintoshorterchainlengtholigosaccharides.Other genes related to metastasisincluded lepr (leptin receptor)is a single-transmembrane-domainreceptorofthecytokinereceptor family functionsasa receptorfor the fatcell-specific hormone leptin; mmp2 is theprotease responsible for metastasis,tgfb1a (transforminggrowthfactorbeta1alpha)actssynergisticallywith transforminggrowthfactoralphaininducingtransformation,and timp2a(tissueinhibitorofmetalloproteinase2a)maintains con-nectivetissueintegritybymodulatingMMPactivity.Amongthose genes,mostofthemwereupregulatedat5months(Fig.5E),and decreasedat5.75month.Formostofthegenes,AFB1hasstronger effectonincreasingtheirexpressionthanHBxalone,andthe syn-ergisticeffectismostobviousforcol1a1andhpse.

Cellcycleanddivisionrelatedgenesincludeccna1,ccnb1,ccne1, ccng1,cdk1,cdk2andpcna.Theeukaryoticcellcycleisgoverned bycyclin-dependent proteinkinases(cdks) whoseactivitiesare regulatedbycyclinsand cdkinhibitors.Cyclin-A1(ccna1) func-tionsasactivatingsubunitsofenzymaticcomplextogetherwith cyclin-dependentkinases.G2/mitotic-specificcyclin-B1(ccnb1)is aregulatoryproteininvolvedinmitosis.CyclinE(ccne)bindstoG1 phasecdk2,whichisrequiredforthetransitionfromG1toSphase. Cyclindependentkinase1(cdk1)isahighlyconservedproteinthat functionsasaserine/threoninekinase,andisakeyplayerincell cycleregulation.Theactivityofcyclin-dependentkinase2(cdk2)is restrictedtotheG1-Sphaseofthecellcycle,andisessentialforthe G1/Stransition.Proliferatingcellnuclearantigen(pcna)isan anti-genthatisexpressedinthenucleiofcellsduringtheDNAsynthesis phaseofthecellcycle.Cellcycle/division-relatedgeneswerenot upregulatedat5monthsofagebutweredramaticallyupregulated at5.75monthsofage(Fig.5FandH).

Tumormarkersincludep53(protein53ortumorprotein53) whichisimportantinmulticellularorganisms,whereitregulates thecellcycleand,thus,functionsasatumorsuppressorthatis involved in preventingcancerand upregulated in many cancer typeswithmutantforms;c-mybproto-oncogene(cmyb)playsan importantroleinthecontrolofproliferationanddifferentiationof hematopoieticprogenitorcells,andccnd1(cyclinD1:G1/S-specific cyclin-D1)formsacomplexwithandfunctionsasaregulatory sub-unitofCDK4orCDK6,whoseactivityisrequiredforcellcycleG1/S transition.Tumormarkerswerealsoupregulatedat5.75months ofage(Fig.5G),buttheeffectofAFB1onthosetumormarkersare strongerthanHBxaloneorHBxandAFB1together.Forcomparison, wealsoprovidedthelipidmetabolismmarkersandfibrosisgenesat 5.75monthsinSupplementaryFig.2.Thoseresultsdemonstrated asynergisticeffectofAFB1andHBxontheregulationoflipogenic factorsandlipidmetabolismrelatedgenesat9wpi/5Mandcell cycle/division-relatedgenesat12wpi/5.75M.

Co-localizationofPCNAwithcaspase3inHBxzebrafish liversamplesandhumanHBVpositiveHCCsamples

We foundthePCNApositiveproliferativefrequency and the TUNELpositiveapoptoticcellswereincreasingwiththe progres-sion of time, we could like to determine whether those two conditionshappenedinthesamelocationornot.Weperformed thedoublefluorescenceimmunostainingusingantibodyagainst PCNAand active caspase3.The resultsshowedthatPCNA was co-localizedwithaactivecaspase3ainHBxorAFB1-treatedHBx zebrafish liver (Fig. 6A)and humanHBV positive HCC samples ofstageI–II(Fig.6B).Thisresultindicatedtheproliferationand apoptoticeventsco-existedinbothhumanHCCandzebrafishliver hyperplasia.

Discussion

Liver cancerorhepatocellular carcinoma(HCC) ranksas the third leadingcauseof mortalityworldwide(Robertsand Gores, 2005;El-SeragandRudolph,2007).Themorbidityandmortality inthesurgicaltreatmentofHCCpatientsinrecentyearshasbeen reduced.However,theprognosisofHCCremainsunsatisfactory; the5-yearsurvivalrateof25–49%aftersurgeryislimited(Luetal., 2011).Therefore,itisimportanttostudycarcinogenesisinanimal modelsanddevelopdrugscreeningplatforms.HumanhepatitisB virus(HBV)isoneofthemajoretiologicalfactorsandriskfactors forHCC(Boschetal.,2004).TheHBVXprotein(HBx)isessentialfor virusreplicationinvivoandplayacrucialroleinthedevelopment ofHCC(Chengetal.,2007;NgandLee,2011).InHBxtransgenic mousemodel,80%incidenceofHCCwasobservedataround 16-monthsofage(Wuetal.,2006).Intransgenicfish,HCVtransgenic fishtreatedwithTAAdevelopedHCCtentimesfasterthanHCV transgenicmouse(Rekhaetal.,2008).However,morethan99%of HBxtransgenicfishdeveloponlyhepaticsteatosisandliver degen-erationwithin8–10monthswithoutanysignofHCC(Shiehetal., 2010).AflatoxinB(AFB1)isoneofthemostprevalentcarcinogens inducinglivercancer(Wooetal.,2011;Woganetal.,2012).By treatingtheHBxtransgenicfishwithAFB1,wetestthehypothesis ofthesynergisticeffectbetweenetiologicalfactors,andexamine thecarcinogenesisbyhistological,immunostainingandmolecular methods.

Thezebrafishisanexcellentmodelfordelineatingthe mech-anisms that underlie hepatocarcinogenesis and for use as a therapeuticdrug-screeningplatform(SpitsbergenandKent,2003; ZonandPeterson,2005;LieschkeandCurrie,2007).Many onco-genes have been shown to induce HCC in zebrafish, including krasV12 (Nguyen etal., 2011,2012), xmrk (Li et al.,2012), and humanRAF-1 (He et al., 2011), and thetumors induced in the zebrafish share many similar features with those occurring in humans(Lietal.,2012).Viralinfectionshavebeencorrelatedto hepatocarcinogenesis.TheexpressionoftheHCVcoreprotein com-binedwithatreatmentwiththehepatotoxinthioacetamide(TAA) hasbeenshowntoinduceHCCinzebrafish(Rekhaetal.,2008). However, the overexpression of the HBx protein alone caused hepatic steatosisandliverdegenerationbut nosignsof HCCin zebrafish(Shiehetal.,2010).Co-expressionofHBxandHCVcore antigencaninducetheformationofintrahepatic cholangiocarci-nomainzebrafishprovidingevidenceoftheusefulnessofstudying synergisticeffectbetweenoncogenicfactorsusingzebrafishmodel (Liuetal.,2012).Inthisstudy,wereportthecollaborativeeffectof HBxandAFB1incausingsteatosis,fibrosisandhyperplasiausinga transgeniczebrafishmodel.Toinvestigateproliferationor apopto-siswereassociatedwithenhancedcellproliferationorcelldeath, proliferationanddeathassaysusingproliferatingcellnuclear anti-gen(PCNA)andactivecaspase3doublefluorescencestainingwere

J.-W.Luetal./ActaHistochemica115 (2013) 728–739 737

Fig.6.DoubleimmunofluorescencestainingofPCNAandactivecaspase3intheliversectionsoftransgeniczebrafishandHumanHCCsample(A)detectionofproliferating andapoptoticcellsinliversof5.75-month-oldEGFP-mCherry,HBx,AFB1,HBX+AFB1fish(400×)and(B)humanHCCliversofstageI–II(200×)bydoubleimmunofluorescence staining.ThePCNA(green)positivecellsareproliferating,andactivecaspase3(red)immuno-positivecellsareconsideredapoptotic.MergeofbothPCNAandactivecaspase 3apositivecellsareshownasyellow.Scalebars=50␮m.(Forinterpretationofthereferencestocolorinthisfigurelegend,thereaderisreferredtothewebversionofthe article.)

performed.PCNAwasco-localizedwithcaspase3inHBxor AFB1-treatedHBxzebrafishliverandhumanHBVpositiveHCCsamples ofstageI–II,whichindicatedthattheymightberesponsibleforthe processoflivercancerformation,whichissimilarinzebrafishand human.

TherearemanyriskfactorsfordevelopingHCC,andwetested twoofthemhereforsynergisticeffectswithHBx.Previously,AFB1 wasused in rainbowtrout to inducea highincidence of liver neoplasia(Baileyetal.,1996).Here,wedemonstratedthatAFB1 workssynergisticallywithHBxtopromote steatosisasearlyas 1wpi/3M,andfinallydevelopedintohyperplasiaat12wpi/5.75M in zebrafish.However,HBx andAFB1 alone canpromote fibro-sis,glycogenaccumulation,andapoptosisandnosynergisticeffect betweenthemwasobservedinthosepathologicalsituations.Using quantitativeRT-PCRtoanalyzetheexpressionoflipogenicfactors, lipidmetabolismrelatedgenes,fibrosismarkers,tumormarkers andcellcycle/divisionrelatedgenes,wediscoveredthe synergis-ticeffectbetweenHBxandAFB1onincreasingtheexpressionof mostofthelipidmetabolismrelatedgenesat9wpi/5Mandcell cycle/divisionrelatedgenesat12wpi/5.75M.Ourdataprovide evi-denceforthesynergisticeffectsbetweendifferentriskfactorsin hepatocarcinogenesisespeciallyinpromotingsteatosisand hyper-plasia.

In a mouse model, the HBxtransgenic micedeveloped into tumornoduleswithallthecharacteristicsofhepatocellular ade-nomas at 22weeks. Thosenodules accumulated relatively high amountsofHBxproteinandcontainedincreasedlevelsofglycogen (Zhuetal.,2004).PASpositivityhadbeencorrelatedtolonger sur-vivalrateandtumorvolumedoublingtimeinhumans(Kitamura etal.,1993).Inthisstudy,bothHBxtransgenicfishtreatedwith AFB1andhumanpatientswithinflammation,cirrhosis,stagesII–III of HCC and CC, exhibited significantly stronger glycogen accu-mulationthanthenormalliverbyPASstaining.Wesuggestthat PASstainingcouldbeusefulinassessingtheprognosisofhuman patients.Ourzebrafishmodelprovidesarobustplatformfor under-standingtheinterplaybetweenenvironmentalsignalsandgenetic factorsin causing liverdisease and combined withtheclinical diagnosis.Thisanimal modelalsoprovidesnewpossibilitiesfor screeningpotentialtherapeuticsforfattyliver,hepatitis,cirrhosis, andlivercancer.

Acknowledgements

ThisresearchwassupportedbyNRPGM(NSC 99-3112-B-400-010)andNSCgrants(NSC101-2321-B-400-017)providedtoC.H.Y. FundingsupportfromtheNationalHealthResearchInstitute pro-videdtoC.H.Y.isgratefullyacknowledged.Wewouldliketothank theTaiwanZebrafishCoreFacilityatNTHU-NHRIsiteforproviding thefishlinesandresources,TZeTHissupportedbygrant 101-2321-B-400-014fromtheNationalScienceCouncil(NSC).Thefellowship toJeng-WeiLuandthefundingsupporttoDr.Chiou-HwaYuhfrom theLiverDisease Prevention &Treatment ResearchFoundation, Taiwanishighlyacknowledged.

AppendixA. Supplementarydata

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.acthis.2013.02.012.

References

AguilarF,HarrisCC,SunT,HollsteinM,CeruttiP.Geographic vari-ationofp53mutationalprofileinnonmalignanthumanliver. Science1994;264:1317–9.

AkriviadisEA,LlovetJM,EfremidisSC,ShouvalD,CaneloR,Ringe B,etal.Hepatocellularcarcinoma.BrJSurg1998;85:1319–31. AndreM,AndoS,BallagnyC,DurliatM,PoupardG,BrianconC,etal.

Intestinalfattyacidbindingproteingeneexpressionrevealsthe cephalocaudalpatterningduringzebrafishgutmorphogenesis. IntJDevBiol2000;44:249–52.

BaileyGS,WilliamsDE,HendricksJD.Fishmodelsfor environmen-talcarcinogenesis:therainbowtrout.EnvironHealthPerspect 1996;104(Suppl.1):5–21.

BeasleyRP,HwangLY,LinCC,ChienCS.Hepatocellularcarcinoma and hepatitis Bvirus.A prospectivestudy of22707menin Taiwan.Lancet1981;2:1129–33.

Bennett JW, Klich M. Mycotoxins Clin Microbiol Rev 2003;16:497–516.

BoschFX,RibesJ,DiazM,CleriesR.Primarylivercancer:worldwide incidenceandtrends.Gastroenterology2004;127:S5–16. Brechot C. Pathogenesis of hepatitis B virus-related

hepatocel-lular carcinoma: old and new paradigms. Gastroenterology 2004;127:S56–61.

ChanTM,ChaoCH,WangHD,YuYJ,YuhCH.Functionalanalysis oftheevolutionarilyconservedcis-regulatoryelementsonthe sox17geneinzebrafish.DevBiol2009;326:456–70.

Cheng AS,Wong N, Tse AM,Chan KY, Chan KK,Sung JJ,et al. RNAinterferencetargetingHBxsuppressestumorgrowthand enhancescisplatinchemosensitivityin humanhepatocellular carcinoma.CancerLett2007;253:43–52.

Denovan-WrightEM,PierceM,WrightJM.Nucleotidesequenceof cDNAclonescodingforabrain-typefattyacidbindingprotein anditstissue-specificexpressioninadultzebrafish(Daniorerio). BiochimBiophysActa2000;1492:221–6.

El-SeragHB,MasonAC,KeyC.Trendsinsurvivalofpatientswith hepatocellularcarcinomabetween1977and1996intheUnited States.Hepatology2001;33:62–5.

El-Serag HB, Rudolph KL. Hepatocellular carcinoma: epidemi-ology and molecular carcinogenesis. Gastroenterology 2007;132:2557–76.

He S, Krens SG, Zhan H, Gong Z, Hogendoorn PC, Spaink HP, etal. ADeltaRaf1-ER-inducibleoncogenic zebrafishlivercell modelidentifieshepatocellularcarcinomasignatures.JPathol 2011;225:19–28.

HerGM,ChiangCC,ChenWY,WuJL.Invivostudiesofliver-type fattyacidbindingprotein(L-FABP)geneexpressioninliverof transgeniczebrafish(Daniorerio).FEBSLett2003;538:125–33. HirumaS,Gopalan-KriczkyP,QinG,GaughanJP,LotlikarPD.

Dif-ferential effects of acetaminophen pretreatment on hepatic aflatoxin B(1)-DNA binding, cellular proliferation, and afla-toxinB1-inducedhepaticfociinratsandhamsters.CancerLett 2001;170:117–24.

HsuIC,MetcalfRA,SunT,WelshJA,WangNJ,HarrisCC.Mutational hotspotinthep53geneinhumanhepatocellularcarcinomas. Nature1991;350:427–8.

JeannotE,BoormanGA,KosykO,BradfordBU,ShymoniakS, Tumur-baatar B, et al. Increased incidence of aflatoxin B1-induced livertumorsinhepatitisvirusCtransgenicmice.IntJCancer 2011;130:1347–56.

KewMC.SynergisticinteractionbetweenaflatoxinB1andhepatitis Bvirusinhepatocarcinogenesis.LiverInt2003;23:405–9. KewMC.HepatitisBvirusxproteininthepathogenesisof

hep-atitisBvirus-inducedhepatocellularcarcinoma.JGastroenterol Hepatol2011;26(Suppl.1):144–52.

KitamuraS,TatsutaM,YamamotoR,IishiH,KajiI,KasugaiH,etal. PrognosticvalueofperiodicAcid-schiff(pas)stainingof fine-needle aspirates from patients withprimary hepatocellular-carcinoma.JGastroenterolHepatol1993;3:245–51.

KwanKM,FujimotoE,GrabherC,MangumBD,HardyME,Campbell DS,etal.TheTol2kit:amultisitegateway-basedconstruction

J.-W.Luetal./ActaHistochemica115 (2013) 728–739 739 kit for Tol2 transposon transgenesis constructs. Dev Dyn

2007;236:3088–99.

Li CH, Wang YJ, Dong W, Xiang S, Liang HF, Wang HY, et al. Hepaticovalcelllinesgeneratehepatocellularcarcinoma fol-lowingtransfectionwithHBxgeneandtreatmentwithaflatoxin B1invivo.CancerLett2011;311:1–10.

LiZ,HuangX,ZhanH,ZengZ,LiC,SpitsbergenJM,etal.Inducible andrepressableoncogene-addictedhepatocellularcarcinomain Tet-onxmrktransgeniczebrafish.JHepatol2012;56:419–25. LianM, Liu Y, Yu SZ, Qian GS, Wan SG, Dixon KR. HepatitisB

virusxgeneandcyanobacterialtoxinspromoteaflatoxin B1-inducedhepatotumorigenesisin mice.WorldJGastroenterol 2006;12:3065–72.

LieschkeGJ,CurriePD.Animalmodelsofhumandisease:zebrafish swimintoview.NatRevGenet2007;8:353–67.

LiuW,ChenJR,HsuCH,LiYH,ChenYM,LinCY,etal.Azebrafish modelofintrahepaticcholangiocarcinomabydualexpression ofhepatitisBvirusXandhepatitisCviruscoreproteininliver. Hepatology2012;56:2268–76.

LuJW,ChangJG,YehKT,ChenRM,TsaiJJ,HuRM.Overexpression of Thy1/CD90in humanhepatocellular carcinomais associ-atedwithHBVinfectionandpoorprognosis.ActaHistochem 2011;113:833–8.

LuJW,HsiaY,YangWY,LinYI,LiCC,TsaiTF,etal.Identification ofthecommonregulatorsforhepatocellularcarcinomainduced byhepatitisBvirusXantigeninamousemodel.Carcinogenesis 2012;33:209–19.

LuyendykJP,CoppleBL,BartonCC,GaneyPE,RothRA. Augmenta-tionofaflatoxinB1hepatotoxicitybyendotoxin:involvement of endothelium and the coagulation system. Toxicol Sci 2003;72:171–81.

MarreroCR,MarreroJA.Viralhepatitisandhepatocellular carci-noma.ArchMedRes2007;38:612–20.

McGlynnKA, LondonWT. Epidemiology and natural history of hepatocellular carcinoma. Best Pract Res Clin Gastroenterol 2005;19:3–23.

MorganTR,MandayamS,JamalMM.Alcoholandhepatocellular carcinoma.Gastroenterology2004;127:S87–96.

NgSA,LeeC.HepatitisBvirusXgeneandhepatocarcinogenesis.J Gastroenterol2011;46:974–90.

Nguyen AT, Emelyanov A, Koh CH, Spitsbergen JM, Lam SH, MathavanS,etal.Ahighlevelofliver-specificexpressionof oncogenicKras(V12)drivesrobustlivertumorigenesisin trans-geniczebrafish.DisModelMech2011;4:801–13.

NguyenAT,EmelyanovA,KohCH,SpitsbergenJM,ParinovS,Gong Z.Aninduciblekras(V12)transgeniczebrafishmodelforliver tumorigenesisandchemical drugscreening.DisModelMech 2012;5:63–72.

PangR,TseE,PoonRT.Molecularpathwaysinhepatocellular car-cinoma.CancerLett2006;240:157–69.

PengZ,ZhangY,GuW,WangZ,LiD,ZhangF,etal.Integration ofthehepatitisBvirusXfragmentinhepatocellularcarcinoma anditseffectsontheexpressionofmultiplemolecules:akeyto thecellcycleandapoptosis.IntJOncol2005;26:467–73. Rekha RD,AmaliAA, HerGM, Yeh YH, Gong HY,Hu SY, et al.

Thioacetamideacceleratessteatohepatitis,cirrhosisandHCCby

expressingHCVcoreproteinintransgeniczebrafishDaniorerio. Toxicology2008;243:11–22.

RobertsLR.Sorafenibinlivercancer–justthebeginning.NEnglJ Med2008;359:420–2.

Roberts LR, Gores GJ. Hepatocellular carcinoma: molecular pathways and new therapeutic targets. Semin Liver Dis 2005;25:212–25.

RobinsonWS.TheroleofhepatitisBvirusinthedevelopmentof pri-maryhepatocellularcarcinoma:partI.JGastroenterolHepatol 1992;7:622–38.

SeeffLB,HoofnagleJH.Epidemiologyofhepatocellularcarcinoma inareasoflowhepatitisBandhepatitisCendemicity.Oncogene 2006;25:3771–7.

ShiehYS,ChangYS,HongJR,ChenLJ,JouLK,HsuCC,etal.Increase ofhepaticfataccumulationbyliverspecificexpressionof Hep-atitis B virus X protein in zebrafish. Biochim Biophys Acta 2010;1801:721–30.

SiegelAB,OlsenSK,MagunA,BrownJrRS.Sorafenib:wheredowe gofromhere?Hepatology2010;52:360–9.

SpitsbergenJM,KentML.Thestateoftheartofthezebrafishmodel for toxicology and toxicologic pathology research – advan-tagesandcurrentlimitations.ToxicolPathol2003;31(Suppl.): 62–87.

TarantinoG,Saldalamacchia G,Conca P,ArenaA.Non-alcoholic fattyliver disease: furtherexpression of the metabolic syn-drome.JGastroenterolHepatol2007;22:293–303.

TsengWF,JangTH,HuangCB,YuhCH.Anevolutionarilyconserved kernelofgata5,gata6,otx2andprdm1aoperatesinthe forma-tionofendoderminzebrafish.DevBiol2011;357:541–57. WildCP,JansenLA,CovaL,MontesanoR.Moleculardosimetryof

aflatoxinexposure:contributiontounderstandingthe multi-factorialetiopathogenesisofprimaryhepatocellularcarcinoma withparticularreferencetohepatitisBvirus.EnvironHealth Perspect1993;99:115–22.

WildCP,TurnerPC.Thetoxicologyofaflatoxinsasabasisforpublic healthdecisions.Mutagenesis2002;17:471–81.

WoganGN,KenslerTW,GroopmanJD.Presentandfuturedirections oftranslationalresearchonaflatoxinandhepatocellular carci-noma.Areview.FoodAdditContamPartA2012;29:249–57. WooLL,EgnerPA,BelangerCL,WattanawarapornR,TrudelLJ,Croy

RG,etal.AflatoxinB1-DNAadductformationandmutagenicity inliversofneonatalmaleandfemaleB6C3F1mice.ToxicolSci 2011;122:38–44.

WuBK,LiCC,ChenHJ,ChangJL,JengKS,ChouCK,etal.Blockingof G1/Stransitionandcelldeathintheregeneratingliverof Hep-atitisBvirusXproteintransgenicmice.BiochemBiophysRes Commun2006;340:916–28.

ZhuH,WangY,ChenJ,ChengG,XueJ.Transgenicmiceexpressing hepatitis BvirusX protein are more susceptible to carcino-geninduced hepatocarcinogenesis. Exp Mol Pathol 2004;76: 44–50.

ZhuS,LeeJS,GuoF,ShinJ,Perez-AtaydeAR,KutokJL,etal.Activated ALKcollaborateswithMYCNinneuroblastomapathogenesis. CancerCell2012;21:362–73.

ZonLI,PetersonRT.Invivodrugdiscoveryinthezebrafish.NatRev DrugDiscov2005;4:35–44.