Presence and extent of estrogen receptor-alpha expression in patients with simple steatosis and NASH

(1)

Pathology–ResearchandPractice209 (2013) 429–432

ContentslistsavailableatSciVerseScienceDirect

Pathology

–

Research

and

Practice

jo u r n al ho me p a g e :w w w . e l s e v i e r . c o m / l o c a t e / p r p

Original

article

Presence

and

extent

of

estrogen

receptor-alpha

expression

in

patients

with

simple

steatosis

and

NASH

Gulbanu

Erkan

a,∗

_,

_Guldal

_Yilmaz

b

_,

_Ceyla

_Konca

_Degertekin

c

_,

_Gulen

_Akyol

b

_,

_Seren

_Ozenirler

d

a_Ufuk_University_Hospital,_Department_of_{Gastroenterology,}_Faculty_of_Medicine,₀₆₅₂₀_Balgat,_Ankara,_Turkey

b_Gazi_University_Hospital,_Department_of_Pathology,_Faculty_of_Medicine,₀₆₅₀₀_Besevler,_Ankara,_Turkey

c_Gazi_University_Hospital,_Department_of_{Endocrinology,}_Faculty_of_Medicine,₀₆₅₀₀_Besevler,_Ankara,_Turkey

d_Gazi_University_Hospital,_Department_of_{Gastroenterology,}_Faculty_of_Medicine,₀₆₅₀₀_Besevler,_Ankara,_Turkey

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received4January2013

Receivedinrevisedform15March2013

Accepted16April2013 Keywords: Estrogenreceptor-alpha Nonalcoholicsteatohepatitis Simplesteatosis

a

b

s

t

r

a

c

t

Lossofestrogenreceptor-alpha(ER-␣)intheliverisassociatedwithhepaticsteatosisandinflammation. WeconductedastudyinordertoinvestigatethepresenceandextentofER-␣expressioninNASH,and itsrelationshipwithhistologicalfindings.Fifty-fourpatientswithhistologicallyconfirmedNASH,12 patientswithsimplesteatosis(SS),and6patientswithnormallivertissue(NLT)wereincluded.NASH activityscoreandfibrosisscorewerecalculatedaccordingtobiopsyfindings.Liverbiopsyspecimens wereimmunohistochemicallystainedforER-␣expression.NuclearER-␣expressionpercentage(staining index)wascalculated.MeanstainingindexwassignificantlydifferentacrosstheNASH,SS,andNLTgroups (6.3±9.9vs.22.1±26.4vs.44.2±24.8,respectively,p<0.001forallcomparisons).Stainingindexwas significantlyhigherinwomenthaninmen(19.4±22.2vs.7.9±15.3,respectively,p=0.003).Staining indexnegativelycorrelatedwithserumALT(r=−0.240;p=0.04),fastingplasmaglucose(r=−0.261; p=0.027),andfibrosisscore(r=−0.312;p=0.011).Asaconclusion,hepaticnuclearER-␣expression percentage(stainingindex)islowerinpatientswithNASHwhencomparedtoSSandNLTgroups.Staining indexisnegativelycorrelatedwithserumALTlevels,plasmaglucose,andfibrosisscore.Furtherstudies arerequiredtoclarifythesignificanceofER-␣expressioninNASH.

Introduction

Theterm“nonalcoholicsteatohepatitis”(NASH)hasbeenused

todescribeadistinctdiseaseprocesscharacterizedbyliverbiopsy

ﬁndingsidenticaltoalcoholichepatitis,withoutahistoryofnotable

alcoholintake[7].Itisapartofthenon-alcoholicfattyliver

dis-eases(NAFLD)spectrum,inwhichhepatocyteinjury,inﬂammation,

andﬁbrosisareobservedinadditiontomacrovesicularhepaticfat

accumulation(simplehepaticsteatosis)[8].

RecentstudieshavedemonstratedthatNAFLDismorecommon

inmenthaninwomen[15].Yatsujietal.reportedthatwomenolder

than55yearsofagehaveahigherincidenceofNAFLDthanmenin

thesameagegroup[17],suggestingthatmenopausemayhavean

impactontheinitiationofNAFLD.Breastcancerpatientstreated

withestrogenreceptorantagonisttamoxifendevelopprominent

hepaticsteatosis,andinsomecases,full-blownNASH[13].These

observations suggest that estrogens retard thedevelopment of

NAFLDandNASH.

∗ Correspondingauthorat:UfukUniversitesiTıpFakultesiMevlanaBulvarı

No:86-88,06520Balgat,Ankara,Turkey.Tel.:+903122044172;fax:+903122044055.

E-mailaddress:gcanbaloglu@gmail.com(G.Erkan).

Theeffectsofestrogenaremediatedviaitsspeciﬁcreceptor,the

estrogenreceptor(ER).ERisanuclearreceptorthatfunctionsas

atranscriptionfactorandmodiﬁestheexpressionofsomecertain

genes.Estrogenactionsarebasicallycarriedoutby2distinct

recep-tors:estrogenreceptor-alpha(ER-␣)and estrogenreceptor-beta

(ER-␤)[10].

Chowetal.demonstratedthatonlyER-␣isexpressedinthe

mouseliverandthatestrogensactviaER-␣toregulatetriglyceride

homeostasis.AselectiveER-␣agonistalleviateshepaticsteatosisin

themalearomataseknockoutmice[2].Ribasetal.reportedthatloss

ofER-␣bringsaboutlipiddeposition,inﬂammation,andinsulin

resistance,andimpairedglucosetoleranceinfemalemice[12].

WhileestrogenispresumedtoretardthedevelopmentofNAFLD

and NASH, thereis nopublisheddata concerninghepatic ER-_␣

expressioninhumansubjectswithNASH.Inthisstudy,weaimed

toinvestigatetherelationshipbetweenthepresenceandextentof

ER-␣expressionanditsrelationshiptohistopathologicﬁndingsin

patientswithnormallivertissue(NLT),simplesteatosis(SS),and

NASH.

Materialandmethods

Fifty-four patients with histologically conﬁrmed NASH, 12

patients with histologically conﬁrmed SS, and 6 patients with

(2)

430 G.Erkanetal./Pathology–ResearchandPractice209 (2013) 429–432 Table1

Demographicandbiochemicalﬁndingsofthestudygroup.

NASH(n:54) SS(n:12) NLT(n:6) p-value Age(years) 45.5±9.6 42.8±12.2 42.8±15.1 0.654 Gender(female/male) 18/36 5/7 3/3 0.663 AST(IU/L) 50.0±33.1 43.1±40.6 31±6.6 0.033 ALT(IU/L) 86.9±43.0 63.4±46.8 36.2±12.2 0.001 ALP(IU/L) 94.9±34.2 81.3±29.4 106.7±18.1 0.213 GGT(IU/L) 83.5±77.3 89.2±50.2 70.2±32.3 0.279 Totalbilirubin(mg/dl) 0.73±0.27 0.7±0.45 0.91±0.64 0.343 Directbilirubin(mg/dl) 0.26±0.11 0.24±0.07 0.35±0.14 0.241

Fastingplasmaglucose(mg/dl) 107.4±19.8 89.8±15.5 91.5±12 0.003

ERpositivity 28(51.8%) 9(75%) 6(100%) 0.012

ERexpressionpercentage(%) 6.3±9.9 22.1±26.4 44.2±24.8 <0.001

Resultsexpressedasmean±SD.Numbersinparenthesesdenotepercentages.

histologicallyconﬁrmed NLTwere includedinthe study.NASH

and SS groups consisted of individuals with persistently

ele-vated liver enzymes (>6 months) and ultrasound-proven fatty

liver. NLT group consisted of patients with unexplained,

per-sistent liverenzyme elevation (>6 months) whose liver biopsy

resultsrevealednormalliverhistology.Allpatientsunderwenta

completelaboratoryexaminationinordertoexcludeother

pos-sible etiologies of liver disease (viral hepatitis, primary biliary

cirrhosis,sclerosingcholangitis,autoimmunehepatitis,

hemochro-matosis, ␣1-antitrypsin deﬁciency, Wilson’s disease). Exclusion

criteriaalsoincludedhistoryofexposuretodrugsknowntocause

hepaticsteatosis(corticosteroids,methotrexate,tetracycline,

cal-ciumchannelblockers,oramiodarone)intheprevious6months,

clinicalevidenceofhepaticdecompensation,andcurrentorpast

consumptionofmorethan20gofalcoholperday.Afterthe

exclu-sion of other causes, the patients underwent liver biopsy, the

techniqueofwhichisdescribedbelow.

Patients were diagnosed with diabetes mellitus in case of

documentedantidiabeticmedicationuseorafastingplasma

glu-cose≥126mg/dL or a random plasma glucose≥200mg/dL on

2separatemeasurements.The remainingpatientsunderwent a

standard75-goral glucosetolerancetestandwereclassiﬁedas

“normal”, “impaired glucose tolerance”, or “diabetes mellitus”

basedontheAmericanDiabetesAssociationguidelines[1].

Histopathologicalanalysis

Percutaneousliverbiopsywasperformedinallpatientsusing

adisposableMenghinitypeneedle(Hepaﬁx1.6mm;Braun

Mel-sungenAG,Melsungen,Germany).Thebiopsyspecimensofliver

tissuewerestainedwithhematoxylin–eosinandMassontrichrome

stain,andallbiopsyspecimenswereanalyzedbytwoexperienced

pathologistsblindedtoclinicaldata.

Presenceofmacrovesicularsteatosis,lobularinﬂammation,and

acinar zone 3 hepatocellular injury or ballooning degeneration

wererequiredfordiagnosisofsteatohepatitisonliverbiopsy[5].

Thedegreeofsteatosiswasgradedasgrade0if<5%hepatocytes

containmacrovesicularfatdroplets;grade1if5–33%hepatocytes

containmacrovesicularfatdroplets;grade2if33–66%ofthe

hepa-tocytescontainmacrovesicularfatdroplets;andgrade3if>66%

ofthehepatocytescontainmacrovesicularfatdroplets.Thestages

ofﬁbrosiswerescoredasstage0,ifnoﬁbrosiswaspresent;stage

1,ifperivenularand/orperisinusoidalﬁbrosiswasfoundinzone

3;stage2, ifcombinedpericellular portal ﬁbrosiswasevident;

stage3,ifseptal/bridgingﬁbrosiswerepresent;andstage4,ifthe

patienthadcirrhosis.TheNASHactivityscore(NAS)wascalculated

asproposedbyKleineretal.[5].

Liverbiopsyspecimenswereimmunohistochemicallystained

forthepresenceofER-␣expression.Allspecimenswereﬁxated

ina10%formalinsolution.Cross-sectionswereobtainedfromthe

parafﬁnblocksinordertoassesstheestrogenreceptorexpression.

Immunohistochemical staining was performed using an

auto-matedsystemandapplyingthestreptavidin–biotintripleindirect

immunoperoxidasemethod.TheantibodyemployedforER-␣was

IgGtypeandofmonoclonalcharacter(SantaCruzBiotechnology,

CA,USA).ER-positivebreastcancertissuewasusedasapositive

control.NuclearERexpressionpercentagewascalculatedinall

ER-positivecases.NuclearERexpressionpercentagewascomputedby

countingatleast1000cellsatrandomlyselected10ﬁeldsunder

highmagniﬁcation,andthus“stainingindex”wascalculated.

Thestudywasapprovedbythelocalethicscommittee,and

writ-teninformedconsentwasobtainedfromallpatientspriortostudy

entry.Thestudywasconductedinaccordancewiththedeclaration

ofHelsinki.

Statisticalanalysis

DatawasanalyzedusingtheSPSS15.0softwarepackage(SPSS

Inc.,Chicago,IL,USA).Frequenciesandpercentagedistributionsof

thevariablesweregiven.Normaldistributionwasveriﬁed

accord-ingtotheKolmogorov–Smirnovtest.Forvariableswhichdidnot

displaynormaldistribution,comparisonsbetween2groupswere

madeusingtheMann–WhitneyUtest,andcomparisonsbetween

multiplegroupsweremadeusingtheKruskal–WallisHtestwith

Bonferronicorrection.

Level of signiﬁcance was assumed as 0.05 for comparisons

betweengroups.Apvaluelessthan0.05denotesthatthedifference

betweengroupsissigniﬁcant.

Correlationbetweencategoricalvariableswasanalyzedusing

theChi-squaretest.Relationshipbetweennumericalvariableswas

assessedusingtheSpearmancorrelationanalysis.Dependencyof

thevariableswasconsideredassigniﬁcantifthepvaluewasless

than0.05.

Results

Fifty-four patients with histologically conﬁrmed NASH, 12

patientswithhistologicallyconﬁrmedSS,and6patientswith

nor-malliverhistologywereincludedinthestudy.

Demographic and biochemical ﬁndings are summarized in

Table1.Therewasnosigniﬁcantdifferencewithregardtoageand

genderbetweenthethreegroups(p>0.05).

AllofthepatientsintheNLTgroup(n:6),75%ofthepatients

intheSSgroup(n:9),and51.8%ofthepatientsintheNASHgroup

(n:28)werepositiveforER-␣expression.Therewasasigniﬁcant

differencebetweenthegroupswithregardtopresenceof ER-␣

expression(100%vs.75%vs.51.8%,p=0.012).Therewasno

statis-ticallysigniﬁcantdifferencebetweentheER-_␣positiveandER-_␣

negativepatientsintermsofage,diabeticstatus,NASscore,degree

(3)

G.Erkanetal./Pathology–ResearchandPractice209 (2013) 429–432 431

Fig.1. Estrogenreceptorexpressioninpatientswithnonalcoholicsteatohepatitis.

andbilirubin(p>0.05).Nevertheless,therewasasigniﬁcant

dif-ferencebetweentheER-␣positiveandnegativegroupsintermsof

gender:ER-␣positivitywassigniﬁcantlymoreprevalentinfemale

subjects(76.9%vs.50%,infemaleandmalesubjects,respectively,

p=0.04).FibrosisscorewassigniﬁcantlylowerinER-␣positive

sub-jectswhencomparedtotheER-␣negativesubjects(1.0±0.8vs.

1.4±0.9,respectively,p=0.047).

MeanER-␣expressionpercentage(stainingindex)was

signif-icantlydifferentacross theNASH,SS and NLT groups (6.3±9.9

vs.22.1±26.4vs.44.2±24.8,respectively,p<0.001forall

com-parisons)(Figs.1 and 2).Meanstaining indexwassigniﬁcantly

higherinwomenwhencomparedtomen(19.4±22.2vs.7.9±15.3,

respectively,p=0.003).Genderspeciﬁcanalysisrevealedthatthere

wasasigniﬁcantdifferenceintermsofER-␣expression

percent-agebetweenNASH,SS,andNLTgroupsinmen.Nevertheless,in

women,ER-␣expressionpercentagewasnotsigniﬁcantly

differ-entacrossNASH,SS,andNLTgroups(Table2).Inourstudygroup,

36patientswerenon-diabetic,19patientshadimpairedglucose

tolerance,and17patientswerediabetic.Therewasnostatistical

differenceintermsofmeanstainingindexaccordingtothe

dia-beticstatus(p>0.05).Meanstainingindexnegativelycorrelated

withserumALTlevel(r=−0.240;p=0.04),fastingplasmaglucose

(r=−0.261;p=0.027),andﬁbrosisscore(r=−0.312;p=0.011).No

statisticallysigniﬁcantrelationshipwasfoundbetweenthemean

Fig.2. Estrogenreceptorexpressioninpatientswithsimplesteatosis.

stainingindexandtheNASscore,steatosis,inﬂammation,or

bal-looning.

Discussion

NAFLDrepresentsaspectrumofhistopathologicalﬁndings

vary-ingfromsimplesteatosis(SS)tosteatosisplusnecroinﬂammation

(non-alcoholic steatohepatitis, NASH), with or without ﬁbrosis.

Liverbiopsyshouldbeperformedtodiscriminatetheseentities.

SSisexpectedtohaveafavorableprognosis;nevertheless,NASH

mayproceedtocirrhosisin20–25%ofthepatientsinaperiodof10

years[11].

Findingsof previousstudieshaveconsistently demonstrated

thatNAFLDismorecommoninmenthaninwomen;andismore

likelytobeencounteredinpostmenopausalwomenthanitisin

premenopausalwomen,suggestingthatestrogensmayhavea

pro-tectiveeffectagainstthedevelopmentNAFLD[3].

Inananimalmodel,itwasdemonstratedthathepatic

steato-sisspontaneouslydevelopsinaromatase-deﬁcientmice,whichare

notcapableofproducingestrogenandhaveimpaired

hepatocellu-larfattyacidbeta-oxidation. Replacementofestradiolalleviates

hepatic steatosis and improvesmitochondrial and peroxisomal

fattyacidbetaoxidationtothelevelsobservedinwild-typemice

[9].

Chowetal.demonstratedthatonlyER-␣isexpressedinthe

mouseliver.TheirﬁndingssuggestthatestrogensactviaER-␣to

regulatetriglyceridehomeostasisinthemouseliver.Aselective

ER-␣agonistalleviates hepaticsteatosisin themale aromatase

knockoutmice[2].

Ribasetal.reportedthatlossofER-␣bringsaboutdiminished

oxygenutilization,lipid deposition,inﬂammation,insulin

resis-tance,andimpairedglucosetolerance.Moreover,itwasshownthat

ER-_␣playsacrucialroleagainstthedeleteriouseffectsofhigh-fat

dietinfemalemice[12].

ThesetwostudiesdemonstratedthatlossofER-␣intheliveris

associatedwithhepaticsteatosisandinﬂammation[2,12].

Estradiolisapotentendogenousantioxidantthatcounteracts

liverﬁbrosisinanimalmodels[14].Moreover,neutralizing

anti-bodiesagainstestradiol inmaleratsandovariectomyinfemale

ratsresultsinincreasedﬁbrogenesis[16].

Inourstudy,therewasasigniﬁcantdifferencewithregardto

presenceofER-␣expressionbetweenNASH,SS,andNLTgroups.

MeanhepaticER-␣expressionpercentage(stainingindex)was

sig-niﬁcantlylowerinNASHwhencomparedtotheSSandNLTgroups

(6.3±9.9 vs. 22.1±26.4 vs. 44.2±24.8, respectively, p<0.001)

Stainingindexwassigniﬁcantlyhigherinwomenwhencompared

tomen(19.4±22.2vs.7.9±15.3,respectively,p=0.003).As

previ-ouslyreportedintheliterature,estrogenshaveaprotectiveeffect

againsthepatic steatosisandﬁbrosis.Moreover,NAFLDismore

prevalentinmen.Therefore,hepaticER-␣expressionpercentage

(stainingindex)mightbeexpectedtobelowerinmencomparedto

womenandintheNASHpatientscomparedtoSSandNLTgroups.

Indeed,theﬁndingsofourstudysupportthishypothesis.However,

gender-speciﬁcanalysisrevealedthattherewasasigniﬁcant

differ-encewithregardtoER-␣expressionpercentageacrosstheNASH,

SS,andNLTgroupsinmenbutnotinwomen.Thisﬁndingmay

beattributedtosmallsamplesizeandshouldbeveriﬁedinlarge

scalestudies.Theoretically,steatosisandﬁbrosisareexpectedto

occurlessintheER-positivegroupcomparedtotheER-negative

group,nevertheless,therewasnosigniﬁcantdifferencebetween

theER-positiveandER-negativegroupswithregardtoNASscore,

steatosis,andinﬂammation.However,therewasasigniﬁcant

dif-ferencebetweenERpositiveandnegativegroupintermsofﬁbrosis

(4)

432 G.Erkanetal./Pathology–ResearchandPractice209 (2013) 429–432 Table2

ER-␣expressionpercentageacrosstheNASH,SSandNLTgroupswithregardtogender.

Female(n:26) Male(n:46)

NASH(n:18) SS(n:5) NLT(n:3) p NASH(n:36) SS(n:7) NLT(n:3) p

ERexpressionpercentage(%) 11.9±13.6 37±34.9 35±22.9 0.063 3.5±5.7 11.4±12.2 53.3±27.5 0.004

Resultsexpressedasmean±SD.

Loombaetal.havereportedthatdiabetesisstronglycorrelated

withtheriskofNASHandﬁbrosis[6].Inourstudy,therewasno

signiﬁcantdifferencebetweendiabetic,non-diabetic,andimpaired

glucosetolerancegroupsintermsofﬁbrosis,NASscore,steatosis,

inﬂammation,ER-␣presenceandpercentage.Nevertheless,ER-␣

expressionpercentage(stainingindex)wasnegativelycorrelated

withplasmaglucoselevel.Further,large-scalestudiesareneeded

toelucidatethisissue.

Kamada et al. studied the effects of estrogen deﬁciency in

ovariectomized mice which were fed on a high-fat and

high-cholesterol diet for 6 weeks, which resulted in enhanced liver

injurywithincreasedlivermacrophageinﬁltrationandelevated

serumcholesterollevels.Hepatocytemonocytechemoattractant

protein-1 (MCP-1) expression, and hepatic inﬂammatory gene

expressionsweresigniﬁcantlyelevatedinovariectomizedmiceon

ahigh-fatdiet.Estrogentreatmentreducedserumcholesterol

lev-els,liverinjury, macrophageinﬁltration,MCP-1 expression,and

inﬂammatorygeneexpressionsinovariectomizedmice.Thisstudy

demonstratedthatestrogendeﬁciencyacceleratedNASH

progres-sioninovariectomizedmicefedonahigh-fatdiet,andthatthis

effectwasreversedbyestrogentherapy[4].

TheﬁndingsofKamadaandRibas[4,12]suggestthatestrogen

deﬁciencyisassociatedwithhepaticinﬂammation.Nevertheless,

wedidnotobserveasigniﬁcantcorrelationbetweenER-_␣

expres-sionpercentage(stainingindex)andthedegreeofinﬂammationin

liverbiopsyspecimens.Thismaybeduetotherelativelysmall

sam-plesize.Itmayalsobeassertedthattheanti-inﬂammatoryeffects

andtheprotectionfromsteatosisconferred byestrogenmaybe

mediatedbyotherpathwaysratherthanER-␣.Thisassertionalso

warrantsfurtherresearch.

Themainlimitationofourstudyistherelativelysmallsample

size.

Asa conclusion, in ourstudy, hepatic ER-␣expression

per-centage(stainingindex)isdecreasedinpatientswithNASH,and

this expression percentage is negatively correlated with serum

ALT, plasma glucose levels, and ﬁbrosis score. On the other

hand, ER-␣ expression percentage is not correlated with NAS

score, steatosis, inﬂammation, and ballooning. Further studies

are required to clarify the signiﬁcance of ER-␣ expression in

NASH.

Authors’contributions

G.Erkan,S.Ozenirler,G.Yilmaz,G.Akyoldesignedthestudy;

G.ErkanandC.KoncaDegertekinrecruitedthepatients,G.Yilmaz

andG.Akyolperformedthehistopathologicexaminations,andG.

Erkanwrotethemanuscript.

Funding

None.

Conﬂictofinterest

None.

References

[1]AmericanDiabetesAssociation,Diagnosisandclassiﬁcationofdiabetes melli-tus,DiabetesCare32(Suppl.1)(2009)62–67.

[2]J.D.Chow,M.E.Jones,K.Prelle,E.R.Simpson,W.C.Boon,Aselective estro-genreceptor␣agonistameliorateshepaticsteatosisinthemalearomatase knockoutmouse,J.Endocrinol.210(3)(2011)323–334.

[3]Y.Gutierrez-Grobe,G.Ponciano-Rodríguez,M.H.Ramos,M.Uribe,N. Méndez-Sánchez,Prevalenceofnonalcoholicfattyliverdiseaseinpremenopausal, posmenopausalandpolycysticovarysyndromewomen.Theroleofestrogens, Ann.Hepatol.9(4)(2010)402–409.

[4]Y.Kamada,S.Kiso,Y.Yoshida,N.Chatani,T.Kizu,M.Hamano,M.Tsubakio, T.Takemura,H.Ezaki,N.Hayashi,T.Takehara,Estrogendeﬁciencyworsens steatohepatitisinmicefedhigh-fatandhigh-cholesteroldiet,Am.J.Physiol. Gastrointest.LiverPhysiol.301(6)(2011)1031–1043.

[5]D.E.Kleiner,E.M.Brunt,M.VanNatta,C.Behling,M.J.Contos,O.W.Cummings, L.D.Ferrell,Y.C.Liu,M.S.Torbenson,A.Unalp-Arida,M.Yeh,A.J.McCullough, A.J.Sanyal,Nonalcoholicsteatohepatitisclinicalresearchnetwork,designand validationofahistologicalscoringsystemfornonalcoholicfattyliverdisease, Hepatology41(6)(2005)1313–1321.

[6]R.Loomba,M.Abraham,A.Unalp,L.Wilson,J.Lavine,E.Doo,N.M.Bass,NASH ClinicalResearchNetwork,Associationbetweendiabetes,familyhistoryof dia-betesandriskofnonalcoholicsteatohepatitisandﬁbrosis,Hepatology56(3) (2012)943–951.

[7]J.Ludwig,T.R.Viggiano,D.B.McGill,B.J.Oh,Nonalcoholicsteatohepatitis:Mayo clinicexperienceswithahithertounnameddisease,MayoClin.Proc.55(1980) 434–438.

[8]G. Marchesini,E. Bugianesi,G. Forlani,F. Cerrelli,M.Lenzi, R. Manini,S. Natale,E.Vanni,N.Villanova,N.Melchionda,M.Rizzetto,Nonalcoholicfatty liver,steatohepatitis,andthemetabolicsyndrome,Hepatology37(4)(2003) 917–923.

[9]Y.Nemoto,K.Toda,M.Ono,K.Fujikawa-Adachi,T.Saibara,S.Onishi,H.Enzan, T.Okada,Y.Shizuta,Alteredexpressionoffattyacid-metabolizingenzymesin aromatase-deﬁcientmice,J.Clin.Invest.105(12)(2000)1819–1825.

[10]S.Nilsson,S.Mäkelä,E.Treuter,M.Tujague,J.Thomsen,G.Andersson,E. Enmark,K.Pettersson,M.Warner,J.A.Gustafsson,Mechanismsofestrogen action,Physiol.Rev.81(4)(2001)1535–1565.

[11]V.Ratziu,S.Bellentani,H.Cortez-Pinto,C.Day,G.Marchesini,Aposition state-mentonNAFLD/NASHbasedontheEASL2009specialconference,J.Hepatol. 53(2010)372–384.

[12]V.Ribas,M.T.Nguyen,D.C.Henstridge,A.K.Nguyen,S.W.Beaven,M.J.Watt, A.L.Hevener,Impairedoxidativemetabolismandinﬂammationareassociated withinsulinresistanceinERalpha-deﬁcientmice,Am.J.Physiol.Endocrinol. Metab.298(2)(2010)304–319.

[13]T.Saphner,S.Triest-Robertson,H.Li,P.Holzman,Theassociationof nonalco-holicsteatohepatitisandtamoxifeninpatientswithbreastcancer,Cancer115 (14)(2009)3189–3195.

[14]I.Shimizu,N.Kohno,K.Tamaki,M.Shono,H.W.Huang,J.H.He,D.F.Yao,Female hepatology:favorableroleofestrogeninchronicliverdiseasewithhepatitisB virusinfection,WorldJ.Gastroenterol.13(32)(2007)4295–4305.

[15]S.R.Weston,W.Leyden,R.Murphy,N.M.Bass,P.B.Bell,M.M.Manos,N.A. Terrault, Racialandethnicdistributionofnonalcoholicfattyliverin per-sons with newly diagnosed chronic liverdisease, Hepatology41 (2005) 372–379.

[16]M.Yasuda,I.Shimizu,M.Shiba,S.Ito,Suppressiveeffectsofestradiolon dimethylnitrosamine-inducedﬁbrosisoftheliverinrats,Hepatology29(3) (1999)719–727.

[17]S.Yatsuji,E.Hashimoto,M.Tobari,K.Tokushige,K.Shiratori,Inﬂuenceofage andgenderinJapanesepatientswithnonalcoholicsteatohepatitis,Hepatol. Res.37(2007)1034–1043.