Microbiota Modulation With Synbiotic Decreases Liver Fibrosis in a High Fat Choline Deficient Diet Mice Model of Non-Alcoholic Steatohepatitis (NASH)

(1)

www.elsevier.pt/ge

ORIGINAL

ARTICLE

Microbiota

Modulation

With

Synbiotic

Decreases

Liver

Fibrosis

in

a

High

Fat

Choline

Deﬁcient

Diet

Mice

Model

of

Non-Alcoholic

Steatohepatitis

(NASH)

Helena

Cortez-Pinto

a,b,∗

_,

_Paula

_Borralho

c

_,

_Jorge

_Machado

d

_,

_Maria

_T.

_Lopes

e

_,

Inês

V. Gato

d

_,

_António

_M.

_Santos

e

_,

_António

_S.

_Guerreiro

e

a_{Gastroenterology}_and_Hepatology_Department,_Centro_Hospitalar_Lisboa_Norte,_Lisbon,_Portugal b_Nutrition_Laboratory,_Faculdade_de_Medicina_de_Lisboa,_Lisbon,_Portugal

c_Pathology_Institut,_Faculdade_de_Medicina_de_Lisboa,_Lisbon,_Portugal

d_Infectious_Diseases_Department,_Instituto_Nacional_de_Saúde_Dr._Ricardo_Jorge,_Lisbon,_Portugal

e_CEDOC_{--- Chronic}_Diseases_Research_Center,_NOVA_Medical_School,_Faculdade_de_Ciências_Médicas_de_Lisboa,_Lisbon,_Portugal

Received29December2015;accepted18January2016 Availableonline31March2016

KEYWORDS Microbiota; Mice; Non-alcoholicFatty LiverDisease; Prebiotics; Probiotics Abstract

Background: Gutmicrobiotamayplayaroleinnon-alcoholicsteatohepatitis(NASH).Previous studiesshowedthatprebioticsandprobioticsmighthalttheprogressionofsteatohepatitis.

Aim:ToclarifythepotentialeffectofSynbiotic2000®_Forte_(Synb)_in_preventing_or

ameliorat-ingdietinducedsteatohepatitis,particularlyinﬁbrosisprogressionandhowthisintervention correlateswithgutmicrobiotacompositionandendotoxinemia.

Methods:Twenty-sevenC57BL/6miceweredividedintothreegroups:chowdiet(CD,n=7); high-fat choline deﬁcient diet (HFCD, n=10) and HFCD diet supplemented with Synbiotic 2000®_Forte_(four_probiotic_strains_and_four_prebiotics_mixture)_(HFCD₊_Synb,_n₌_10)._At₆_and

18weeks,bloodsamples(lipopolysaccharidesassay---LPS),cecalfeaces(gutmicrobiota)and livertissue(histology)werecollectedforanalysis.

Results:BothHCFDdietmice developedsteatohepatitiswithballooningat6and18weeks, oppositetoCD.ComparisonofhistologicalscoresinHFCDandHFCD+Synb,at6and18weeks showednosignificantdifferenceregardingsteatosis,inflammation,orballooning.Evaluating fibrosiswithSiriusRed,anddegreeofsmooth-musclecellactivation,HFCDmicehad signifi-cantlymorefibrosis;additionofSynbsignificantlyreducedfibrosisat6weeksand18weeks. SerumendotoxinlevelsweresimilarlyincreasedinHFCDandHFCD+Synbatweek6;however atweek18 HFCD+Synbhadsignificantly lowerendotoxinlevelsthanHFCD. Gutmicrobiota

Abbreviations: NASH,non-alcoholicsteatohepatits;Synb,Synbiotic2000®_Forte;_CD,_control_diet;_HFCD,_high_fat_choline_deﬁcient_diet; LPS,lipopolysaccharide;NAFLD,non-alcoholicfattyliverdisease;TLR4,toll-likereceptor-4.

∗_{Corresponding}_author.

E-mailaddress:hlcortezpinto@netcabo.pt(H.Cortez-Pinto).

http://dx.doi.org/10.1016/j.jpge.2016.01.004

(2)

of HFCD vs CD, showed nosigniﬁcant differences regardingthe phyla Firmicutesand Bac-teroidetes, eitherat6or18weeks;Proteobacteriaincreasedat6week(3.3)and18 week (7.5), whilethe addition ofSynb resulted inadecrease atweek 18 (−3.90). Fusobacteria

markedlyincreaseatweek18(10.0),butlesssowiththeadditionofSynb(5.2).

Conclusion: Synbiotic2000®_Forte_is_able_to_modulate_the_mouse_gut_microbiota_reducing_the

degreeofﬁbrosiswhilesimultaneouslydecreasingendotoxemia.

© 2016 Sociedade Portuguesa de Gastrenterologia. Published by Elsevier España, S.L.U. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). PALAVRAS-CHAVE Microbiota; Ratos; EsteatohepatiteNão Alcoólica; Prebióticos; Próbioticos

AModulac¸ãodaMicrobiotaComSimbióticoReduzaFibroseHepáticanumModelode

RatinhoComEsteatohepatite(EHNA)NãoAlcoólicaInduzidaporDietaGordae

DeﬁcienteemColina

Resumo

Introduc¸ão: Amicrobiotaintestinalpodeterumpapelnaesteatohepatitenãoalcoólica(EHNA). Estudospréviosmostraramqueosprebióticosepróbioticospodeminterromperaprogressão daesteatohepatite.

Objetivo: ClariﬁcaroefeitopotencialdoSynbiotic2000®_Forte_(Synb)_na_prevenc_¸ão_ou_melhoria

daesteatohepatiteinduzidapeladieta,particularmentenaprogressãodaﬁbroseecomoessa intervenc¸ãosecorrelacionacomamicrobiotaintestinaleendotoxinemia.

Métodos: Vinte esete ratinhos C57BL/6 foram divididos em 3 grupos:dieta derac¸ão (CD, n=7);dietagordadeﬁcienteemcolina(HFCD,n=10)eHFCDdietasuplementadacomSynbiotic 2000®_Forte₍₄_cepas_de_probioticos_e_mistura_de₄_{prebióticos),}_(HFCD₊_Synb,_n₌_10)._Às₆_e

18semanas,foicolhidosangueparadosearolipopolisacárido(LPS);foramtambémanalisadas fezesdocego(microbiotaintestinal)etecidohepático(histologia).

Resultados: Todos os ratinhos sujeitos a dieta HCFD, desenvolveram esteatohepatite com balonizaçãoàs 6e18 semanas,em oposição aosCD.A comparação dosscores histológicos nosHFCDeHFCD+Synb,às6e18semanasnãoreveloudiferençassignificativasemrelaçãoà esteatose,inflamação,oubalonização.AavaliaçãodafibroseatravésdoSiriusRed,eograu de ativaçãodascélulas domúsculo liso,revelouque osratinhosHFCD tinham significativa-mentemaisfibrose;aadiçãodoSynbreduziusignificativamenteafibroseàs6e18semanas. Osníveis deendotoxinaséricaestavamaumentadosdeformasemelhanteàs6semanasnos HFCDeHFCD+Synb;contudoàsemana18,osHFCD+Synbtinhamsignificativamentemenores níveisdeendotoxinaemrelaçãoaosHFCD.AmicrobiotaintestinaldosHFCDvsCD,não rev-eloudiferençassignificativasnosfilosFirmicuteseBacteroidetes,queràs6ouàs18semanas;

Proteobacteriaaumentouàs6semanas(3.3)e18semanas(7.5),enquantoaadiçãodeSynb resultounumareduçãoàsemana18(-3.9).Fusobacteriaaumentousignificativamenteàsemana 18(10.2),masmenoscomaadiçãodeSynb(5.8).

Conclusão:OSynbiotic 2000®_Forte _foi_capaz _de_modular _a_microbiota _intestinal_do_ratinho

reduzindoograudeﬁbrose,enquantosimultaneamentereduziuaendotoxemia.

1. Introduction

Non-alcoholicsteatohepatitis(NASH)isametabolicliver dis-easeofepidemicimportanceinmanycountries,andstrongly relatedwiththeincreasedprevalence ofobesity and fea-tures of the metabolic syndrome in western countries.1 It is characterized by the association of hepatic steato-sis withliver cell injury,ballooning, lobular inﬂammation andvariabledegreesofﬁbrosis,thatcan progressto end-stagecirrhosis.Clinicalandexperimentalresearchsuggests that the pathogenesis of NASH involves a multifactorial

mechanismbywhich,inthesettingofafattyliver,altered lipid metabolism, dysregulated cytokine production and oxidative stress cause injury and provoke hepatic inﬂam-mationandﬁbrosis.2

In the last years, there has been a great interest in the role that microbiota may play in liver diseases, par-ticularly in non-alcoholic fatty liver disease.3---8 _In _fact, several studies have highlighted the association between smallintestinalbacterial overgrowth, increasedintestinal permeabilityandnonalcoholicsteatohepatitis(NASH), prob-ablyduetoincreasedendotoxinaemiaandactivationofthe

(3)

Toll-likereceptor-4(TLR4)signallingcascade;studieshave focused onthe possibilitythat activation of liver ﬁbroge-nesis maydepend onthe stimulationof stellatecells and Kupffer cells by TLR4.9---13 _An _additional _possible _role _for microbiota in NAFLD, is related with the pathogenesis of obesity,sincethereisevidencethatsomeGram-anaerobes, suchasBacteroidesthetaiotaomicronareabletodegrade plantpolysaccharides,14 _thus_increasing_the_caloric extrac-tionfromdiet;also,colonizationofthegutinexperimental models of germ-free mice, induces obesity,15 _and _in _the opposite,microbiotadepletionpromotesbrowningofwhite adipose tissue and reduces obesity.16 _Another _potential contributionofmicrobiotatoNAFLDrelatestosuppression ofthelipoproteinlipase(LPL)inhibitor,angiopoietin-like4, henceallowingcontinuedexpressionofLPLthatregulates fattyacidreleasefromtriglyceridesintheliver.17

Our aim was to clarify the potential effects of the four probiotic strains and four prebiotics mixture (Synbi-otic2000®_Forte)_in_preventing_or_ameliorating_diet_induced steatohepatitis, and particularly fibrosis progression. In order to get a model more similar to human NASH we havechosenamoderatehigh-fatdiet(35%)simultaneously choline deficient (HFCD), since it has been shown that addingcholinedeficiencytoahigh-fatdiet,amplifiesliver fat accumulation.18 _We _also _aimed _to _evaluate _how _this intervention correlates with microbiota composition and endotoxinemia.

2. Methods

2.1. Animalhousinganddiet

Atotalof27pathogen-freemaleC57BL/6sixweek-oldmice (Harlan Laboratories, Castellar, Spain) weighing 15---20g werehousedinindividualcagesandkeptatatemperature of 22◦C, on a 12:12h light/dark cycle, in the Portuguese NationalInstituteofHealth.

Uponarrival,themicewereacclimatedfor14daysprior tobeing used in experiments, and had ad libitum access to tap water and feeding. This study was reviewed and approved by the National Animal Care Committee from Direcc¸ãoGeraldeVeterinária.Allprocedureswithanimals werein compliancewith the Careand Use of Laboratory Animals.

Steatohepatitiswasinducedbyahigh-fatdiet(35%total fat,54%trans-fattyacidenriched)choline-deﬁcient---HFCD diet, ---Harlan Laboratories, Castellar, Spain and Standard chowdietwasobtainedfrom(Mmucedolasrl---4RF21 cer-tiﬁcatebatch:250202).

2.2. Studydesign

Afteracclimatizationperiod,miceweredividedintothree different groups according the type of diet and product administration:(1)regularchowdiet(CD)n=7;(2)highfat cholinedeﬁcientdiet(HFCD)n=10;(3)HFCDplusSynbiotic 2000®_Forte_(HFCD₊_Synb)_n₌_10.

Usinga20-gaugeballpointmetalfeedingtube(Harvard Apparatus,Inc.,Holliston,MA),micewereinoculated intra-gastric,threetimes/weekwitheither phosphatebuffered

saline(PBS),CDandHFCDgroupsor Synbiotic2000®_Forte: 50mg/mice(4×108_{bacteria/mouse),}_HFCD₊_Synb_group.

Atweek6,ﬁvemiceofeachgroupwereeuthanized.The remainingmiceofeachgroupwereeuthanizedatweek18.

2.3. BacterialviabilityofSynbiotic2000®Forte

Eachsachetconsistedofacombinationof1011_CFU_of_each_of fourprobiotics:Lactobacillusparacasei:25%,Lactobacillus plantarum: 25%, Leuconostoc mesenteroides: 25%, Pedio-coccuspentosaceus:25%withbioactiveﬁbres,totalweight: 10g, oatbran: 2.5g,pectin:2.5g,resistantstarch: 2.5g, inulin:2.5g.

ToaccesstheviabilityoftheSynbiotic2000®_Forte mix-ture,assaysinduplicatewereperformed.Afterrecoveryof thelyophilized,themixturewasculturedonSchaedleragar andincubatedat35◦Cfor48hunderanaerobicconditions. Colonies weretested for biochemicalactivities and moni-toredmicroscopicallyby Gramstain.It wasobserved that the viability of Lactobacillus casei and L. plantarum was 83%,thePediococus100%andtheLeuconostoc72%.

2.4. Sampling

At the timeof sacriﬁceby cervical dislocation, after14h fasting, total body and liver of each mice was weighed. BloodwascollectedtoperformLPSassay.Livertissuewas processedforhistologyandcecalfaecesweretakento eval-uategutmicrobiota.Serumandfaeceswerestoredat−80◦C untilused.

2.5. Bloodanalysis

Bloodwascollectedandcentrifugedfor10minat3000rpm. Serumsamples were usedtoperform LPSassayusing the Cusabio Biotech CO.,LTD, Wuhan, Chinaaccording tothe manufacturer’sinstruction.

2.6. Histology

In each time point, liver mice was obtained and divided intotwofragments:onewasflash-frozeninliquidnitrogen andkeptat −80◦Cuntilanalysis.Thetissuewas homoge-nized andtotalproteinswereextracted andisolated.The otherfragmentwasplacedina10%bufferedformalin solu-tion.Thetissuewasroutinelyprocessed,paraffinembedded andsectionedat athicknessof 4␮m.Haematoxylin---eosin and Sirius Red staining were performed. For the detec-tionofactivatedhepaticstellatecells,␣-SMApositivecells were immunohistochemicallyassessed by ␣-SMA antibody. Immunostainingwasperformedmanuallybythe peroxidase-indirect-polymermethod(EnvisionDako---HRP,CodeK4011) for 30min.Sections 2␮m thickwere cut, unto Superfrost plus slides,fromparaffin-embeddedroutinetissueblocks. The sections were de-waxed, rehydrated and subjected to epitope antigen retrieval (20min, 94◦C) with Target RetrievalSolutionLowpH50xDakoEnvisionTM_Flex_in_a pre-treatmentmodulePTlink(Dako,ModelPT10130).Primary polyclonal rabbit antibody Pan-Actin (Cell Signalling ref.: 4968)wasincubatedovernightatdilution1:200.Appendix

(4)

wasusedaspositivecontrol.Fornegativecontrol,primary antibodywasomittedduringthestaining.Endogenous per-oxidisewasblockedwith2%H2O2inabsolutemethanolfor 10min.Sectionswerecounterstained withMayer’s haema-toxylin.

The haematoxylin---eosin slides were evaluated blindly byanexperiencedpathologistandgradedforsteatosisand necro-inflammationas0(absent),1(sparseormild,focal), 2 (noticeable) and 3 (severe) and hepatocyte ballooning was graded as 0 (absent), 1 (few ballooned cells), and 2 (prominent).Fibrosiswasquantifiedbyhistomorphometric analysisofSiriusRedstainedslides,usingadigitalcamera (DFC320,Leica,Germany)coupledtoabrightfield micro-scope(DM4000,Leica,Germany).Threerandomlyselected fields(10×magnification)ofSiriusRed-stainedslides(5-mm sections)wereevaluatedatafinalmagnification of100×. ImageswereanalyzedusingtheLASImageAnalysisfromLAS --- LeicaApplicationSuite software.Using the‘Image’tool window,imageswerepreparedforanalysisusingthe‘Adjust contrast’tooltooptimizecontrastandthe‘Threshold’ func-tiontoselectspecificsignal.Resultsweremeasuredinterms of%areaoccupiedbyspecificstaining.19

The same method of histomorphometric analysis was usedtoquantifyactivatedstellatecells.SMA-positivecells wereevaluatedinsixﬁeldsperspecimen,whichwere ran-domlyselectedat200×magniﬁcation.

2.7. Gutmicrobiota

Constituentsofcecalmicrobiotawerequantiﬁedusing Fluo-rescenceinsituhybridization(FISH),targetingthephyla Fir-micutes(Staphylococcus,Enterococcus,L.casei/paracasei, L. plantarum, Clostridium, Leuconostoc, Pediococcus, BacilleaceaandStreptococcus),Proteobacteria (Enterobac-teriacea), Fusobacteria (Fusobacterium), Actinobacteria (PropionibacteriumandBiﬁdobacterium)andBacteroidetes (Bacteroides).

2.8. Choosingandpreparingprobes

Fourteen probes of 16S rRNA-targeted oligonucleotide probes were designed, validated and used to quantify predominant bacterial groups (SABiosciences Corporation, USA).

Samplepreparation andﬁxation:Faecessamples were weighed and diluted in distillate water (1:10,000). One microlitreof each dilution wasplaced in slidespreviously washed with 70% ethanol. After that, all samples were ﬁxedbyadding1␮lof4%paraformaldehyde(Sigma---Aldrich, Sintra-Portugal)atroomtemperaturefor20minandwashed in 3× PBS solution by 2min, then repeated two more times. Permeabilization: Treatment with 0.05% (p/v) of SDS(sodiumdodecylsulphate)for 20minandproteinaseK (Sigma---Aldrich,Sintra-Portugal)(Cf=10␮g/mL)for10min. Denaturationandhybridization:Theprobes(200ng/␮L)and samplesweresimultaneousdenaturatedinabath at80◦C for3minandthehybridizationwasat42◦Cfor4h.Washes: After the hybridization, slides were immersed in 0.01× SSC (saline-sodium citrate) and BT/BSA (Sigma---Aldrich, Sintra-Portugal) solutions for 10min each according the temperature (Table 1). DNA colouration and observation:

DNA waslabelled with DAPI (4,6-Diamino-2 Phenylindole; C16H15N5.2HCl-VectorLaboratories)(1mg/ml).The slides wereobserved in a fluorescent microscopy Zeiss Axiovert 200M.Thismicroscopyhadsimplefiltersthatallowedtosee separatedtheCY3, TexasRed, CY5,FITCandPacific Blue and DAPI fluorochromes, through the Plan-Neofluar 100× immersionobjective.

Afterthebacteria’s identiﬁcation, picturesweretaken foreachsample.UsingaMATLABprogram,andenteringinto accounttheweight,dilutionofeachfaecessampleusedin theassay,aswell thepicturearea,totalof eachbacteria wasquantiﬁed.20,21

2.9. Statisticalanalysis

General descriptive statistics used all available samples. Bacterialvariationswereconsideredonlyforvaluesgreater than2.0(2×SD/mean)inatleastoneofthegroups.

Bacteriainwhichthevariationsinallgroupswerelower than this value, were not considered. Results from dif-ferent groups were compared using the Student’s t test orKruskal---Wallisnon-parametricANOVA.Valuesofp<0.05 wereconsideredstatisticallysigniﬁcant.Allstatistical anal-ysiswasperformedusingSPSSstatistics(version21).

3. Results

3.1. Bodyandliverweight

All mice had a body weight increase at each time-point duringthestudy,butnosigniﬁcantdifferenceswerefound between the study groups. However, both HFCD groups showedat 18weeksan increasedliver weightwhen com-paredwithCDgroup,asshown inTable2.The additionof Synbdidnotabrogatethisincreaseinliverweightobserved inmiceundergoingHFCDdiet.

3.2. Liverhistology

Control mice that were given chow diet, had normal liv-ersthroughouttheentirestudy(Fig.1).Ontheopposite,all miceundergoingtheHFCDdietdevelopedaspectsof steato-hepatitis,includingsteatosis,inflammation,ballooningand fibrosis. Consequently, the degree of steatosis, inflamma-tion, ballooning and pericellular fibrosis was significantly higher(p<0.05)inboth HFCDgroupsthan CD(whereonly mildlobularinflammationwasseen);theuseofSynbdidnot abrogatetheseverityofthesehistologicalparametersatany timepoint,asobservedandscoredusingconventionalH&E stainedslides(Fig.1)(Table3).

However,whenevaluatingfibrosisusingSiriusRed,itwas foundthatthepercentageoffibrosisat6and18weekswas significantlydecreasedwiththeadditionofSynb,asseenin Table4,p=0.001at6weeks,andp=0.04at18weeks.

Activation of stellate cells was evaluated using ␣-SMA antibody.AsseeninFig.2,signiﬁcantincreaseswerefound atweek18inmiceundergoingHFCDdietwhencomparing withCD,p=0.007.Thisincreasewasmarkedlyabrogatedby theadditionofSynbiotic2000®_Forte_to_HFCD_diet,_p₌_0.003.

(5)

Table1 Probessequence,ﬂurochomesandhybridizationtemperature.

Identiﬁcation Sequence Fluorochome Hybridizationtemp.(◦C)

Pediococcuspentosaceus 5-GCCATCTTTTAAAAGAAAACCATGC-3 Cy3 48◦C

Leuconostocmesenteroides 5-TTTGTGTCATGCGACACTAAGTTTT-3 Cy5

Lactobacilluscasei/paracasei 5-CGTTCCATGTTGAATCTCGGTG-3 PaciﬁcBlue

Bacillaceae 5-CGTCCATTRNKGAAGATTCCCTA-3 FITC

Fusobacterium 5-CGTCCATTRNKGAAGATTCCCTA-3 TaxasRed 53◦C

Staphylococcus 5-CNCCGAAGRGGDARNYTCTATCTCTAGA-3 Cy3

Enterococcus 5-TTTCCAAGTGTTATCCCYYNCTGANG-3 Cy5

Bacteroides 5-CCTTCACGCTACTTGGCTGGTTCAG-3 PaciﬁcBlue

Enterobacteriacea 5-GACNTYATGCGGTATTAGCYACCG-3 FITC

Propionibacterium 5-TTGACCCCGGCGGTCTCCACTGAGTCC-3 TexasRed 58◦C

Clostridium 5-GTGGCTTCCTCCNHNGGTACCGTCATTATC-3 Cy3

Lactobacillusplantarum 5-TGTTATCCCCCGCTTCTGGGCAGGTT-3 Cy5

Biﬁdobacterium 5-CTGATAGGACGCGACCCCATCNCA-3 PaciﬁcBlue

Streptococcus 5-ACCAACTAGCTAATACANCGCAGGTNCATCT-3 FITC

Table2 Bodyweightandliverweightatdifferenttimepoints.

Group Week Bodyweight(g) Meanweightgain Liverweight(g) Liver/totalbodyweightratio(%)

CD 0 18.6 6 23.4±1.6 5.0 0.9±0.1 3.8 18 25.6_±1.7 7.2 1.2_±0.1 4.7 HFCD 0 18.8 6 21.6±0.7 2.8 1±0.1 4.6 18 27.9±3.0 9.1 2±0.2* _7.1* HFCD+Synb 0 18.9 6 23.4_±1.6 4.5 1.6_±1.3 6.8 18 25.1±0.4 6.2 1.7±0.1* _6.7*

CD:chowdiet;HFCD:high-fatcholinedeﬁcientdiet;Synb:synbiotic.

* _{Statistically}_{signiﬁcant,}_p_<_0.01.

At18weeks,liverweightaswell asliver/bodyweightratiowassigniﬁcanthigherineitherHFCDorHFCD+SynbdietthanCDdiet,

p<0.01.

CD

HFCD

6 W

eeks

18 W

eeks

HFCD + Synb

Figure1 H&Estainedliversections20×fromrepresentativemicefedwithCDdiet(CD)andmicefedhigh-fatcholine-deﬁcient

diet(HFCD)dietorhigh-fatcholine-deﬁcientdietplusSynbiotic,asdetailedinMethods.HFCDdietandHFCDdiet+Synbshowed aspectsofsteatohepatitsat6and18weeksthatwerenotpresentinmiceundergoingCDdiet.TheadditionofSynbdidnotabrogate theseaspects.

(6)

Table3 Meanhistologicalscoresforeachtimepointaccordingtogroup(H&E).

Group Week Steatosis Intralobularinﬂammation Ballooning Pericelularﬁbrosis

CD 6 0 0.75 0 0 18 0 0.6 0 0 HFCD 6 3 3 1 1.6 18 2.4 2.6 2 3 HFCD+Synb 6 3 2 2 1 18 3 2.2 2 1.6

Significantdifferencesinmeanscores(p<0.05),ateachtimepointforsteatosis,intralobularinflammation,ballooningandpericellular fibrosisbetweenCDvsHFCDandHFCD+Synb.NodifferencesbetweenHFCDandHFCD+Synb.

CD HFCD Weeks 18 Weeks 6 0 2 4 6 8 10 12 HFCD+Synb

Figure2 Percentageof_␣-SMApositivecells(activated stel-latecells).Signiﬁcantincreaseswerefoundatweek18inmice undergoingHFCDdietwhencomparingwithCD,p=0.007.This increase was markedlyabrogatedby the addition ofSynb to HFCDdiet,p=0.003.

Table4 Percentageﬁbrosis(SiriusRED).

Group 6weeks 18weeks CD 1.1±0.2 0.6±0.06 HFCD 7.4±2.9 10.7±2.6 HFCD+Synb 3.2±1.0 6.3±2.3

6weeks:CDvsHFCD,p=0.001;HFCDvsHFCD+Synb,p=0.001. 18weeks:CDvsHFCD,p=0.000;HFCDvsHFCD+Synb,p=0.04.

3.3. LPSandGram-negativebacteria

Atweek6,therewasasigniﬁcantincreaseinGram-negative bacteriaandLPSlevelsinthetwoHFCDdietgroups,when comparingwithCDdiet,not abrogatedbythe additionof Synbiotic2000®_Forte._However,_at_week_18,_although_levels

ofLPSandGram-negativebacteriacontinuedtoincreasein HFCDgroup,theadditionofSynbiotic2000®_Forte_abrogated

thiseffect,asshowninFig.3.

3.4. Gutmicrobiota

Datafrommicrobiotawillbepresentedatthephylumand specieslevels.

3.5. Phylumlevel(Table5)

InHFCDversusCD,therewasanincreaseofProteobacteria atweek6(3.3)andatweek18(7.5)andFusobacteria(2.8) and (10.2) respectively; concerning Actinobacteria there wasadecreaseatweek18(−6.7).RegardingFirmicutesand Bacteroidetesnomajorchangeswereobservedatthesame timecourse.

In HFCD+Synb vs CD, there was an increase of Pro-teobacteria at week 6 (5.3) and a decrease at week 18 (−3.9).Fusobacteriahadahugeincreaseatweek6(41.7), andatweek18(5.8).RegardingBacteroidetestherewasa decreaseatweek6(−4.6).Actinobacteriahadadecreaseat week18(−7.6).NomajorchangesatFirmicuteswerefound both times. InHFCD+Synb vs HFCD, nochanges found at ProteobacteriaandActinobacteria,anincreaseof Fusobac-teriaatweek6(15.1),adecreaseofBacteroidetesatweek 6(−5.0),andanincreaseofFirmicutesatweek6(4.2)and decreaseatweek18(−2.3).

3.6. Specieslevel(Table6)

Inthespecieslevel,comparingHFCDversusCD,thereisan increaseofEnterobacteriaceae (2.5and7.5)and Fusobac-terium (2.8 and 10.2) at 6 and 18 weeks, while there is a decreaseof Biﬁdobacterium (−47.7)and Streptococcus (−5.0)atweek18.AdecreaseofLeuconostocwasfoundat bothtimes(−14.9and−4.6),adecreaseofPediococcusat week6(−10.0)andanincreaseatweek18(4.2),adecrease ofL.caseiatbothtimes(−8.7and−2.7)andadecreaseof L.plantarumonlyatweek18(−6.6).

RegardingHFCD+SynbversusCDthereisanincreaseof Enterobacteriaceae(5.3and3.9)andFusobacterium(41.7 and5.8) adecrease ofBiﬁdobacterium (−8.8and −14.3) atbothtimes,StreptococcusandLeuconostocadecreaseat week18(−21.8)and(−56.4).ConcerningL.caseithereisa decreaseatbothtimes(−8.6and−2.9)andadecreaseof L.plantarumonlyatweek18(−6.5).

WhencomparingHFCD+SynbversusHFCD,nochangesin Enterobacteriaceaewerefound atbothtimes,anincrease ofFusobacteriumatweek6(15.1)adecreaseof Biﬁdobac-teriumatweek6(−14.6)andanincreaseatweek18(3.3), adecreaseofStreptococcusatbothtime(−2.2and−4.4). Regarding Leuconostoc and Pediococcus an increase at week6(23.6)and(14.7)andadecreaseatweek18(−2.3) and(−2.8) respectively. Concerning L. casei andL. plan-tarum,nochangesfoundatbothtimes.

(7)

Table5 Phylumvalues.

Firmicutes Proteobacteria Actinobacteria Bacteroides Fusobacteria HFCDvsCD(6wk) −1.7 3.3 1.1 1.1 2.8 HFCDvsCD(18wk) 1.6 7.5 −6.7 1 10.2 HFCD+SynbvsCD(6wk) 2.4 5.3 1.1 −4.6 41.7 HFCD+SynbvsCD(18wk) _−1.4 _−3.9 _−7.6 _−1.2 5.8 HFCD+SynbvsHFCD(6wk) 4.2 2.1 1 −5 15.1 HFCD+SynbvsHFCD(18wk) _−2.3 _−1.9 _−1.1 _−1.2 _−1.8

Bacterialvariationswereconsideredonlyforvaluesgreaterthan2.0(2×SD/mean).

Therewasnochangeinspeciesnotmentioned

(Entero-coccus,ClostridiumandPropionibacterium).

4. Discussion

Themainfindingofthepresentstudyisthatasymbioticwas abletoreduceHFCDdiet-inducedfibrosis,whenevaluated eitherbythemeasurementoffibrosiswithSiriusRedorthe percentageofactivatedstellatecells.Simultaneously,there wasevidencethattheadditionoftheSynbiotic2000®_Forte

abrogatedtheHFCD-associatedincrease intheabundance ofGram-bacteriain thestools,thiseffect,simultaneously reducingserum levelsof LPS. These results reinforce the concept that increased endotoxemia is one of the main mechanisms through which microbiota composition asso-ciateswiththepresenceandseverityofNAFLD,aspreviously shown.22---24 _In_fact, _human_studies_were _able_to_show ele-vatedserumendotoxinlevelsinpatientswithNAFLD,when comparedto controls,25 _furthermore_elevated _in _patients withNASH.26

Thedietmicemodelusedinthepresentstudyresultedin afull-blownpictureofsteatohepatitis.Thehigh-fatcholine deficientdietmodelwaschoseninordertohavemore simi-laritieswiththehumanphenotypeofNASH.Themoreclassic methioninecholinedeficientdietmodelwasavoided,since thosemiceinfactloseweight,inoppositiontohumanNASH. Also,we did not use high-fatdiet model, since although theygetsteatosis, NASHdoes not systematicallydevelop, thuscreatingdifficulties in theevaluation ofan interven-tion. As intervention, the use of a Synbiotic was chosen insteadofaprobioticalone,duetothepotentialadvantage

ofimprovingtheviabilityoftheprobiotic.27_In_human stud-ies, this particular Synbiotic (Forte 2000) has shown to induce lower intestinalpermeability and fewerinfections intrauma patients,28 _but_had_no_signiﬁcant _impact_on_the incidenceofVAPincriticallyillpatients.29

The use of the Synbiotic 2000®_Forte _had _no _effect _in abrogatingsteatosis ornecro-inflammationinduced bythe HFCD diet,but wasabletoreduce thedegree offibrosis, withagoodcorrelationwithadecreaseinGram- bacteria and LPS.A previous study in rats had shown that a high-fat diet(45% fat) wasassociatedwitha decreasein total bacterialdensityandanincreaseintherelativeproportion of Bacteroidales and Clostridiales, which mayinduce LPS elevation,gutinflammationandTLR4activation.30,31_In_the present study, usinga less fat diet (35%), but 54% trans-fattyacidenriched,wehavenotobservedahugeincrease inFirmicutesandadecreaseofBacteroidetes,32_but_only_a smalleffect(increaseofFirmicutesonlyatweek18(1.6)in populationratioHFCDversus CDgroup,andnochangesin Bacteroidetes). Regarding Proteobacteriawe observed an increase,that latterbeingreducedbythe additionof the synbiotic. The reduction in Gram-bacteria may bedue to increasedproductionofshort-chainfattyacids(SCFA), dur-ing the anaerobic metabolism of the prebiotic associated carbohydrates. In fact,SCFA arenonspecific antimicrobial substances that decrease pH, inhibitinggrowth of a wide rangeofGram-bacteria.33

It is of note, that in one recent study where human microbiota wasstudied, it wasfound that the abundance of Proteobacteria increased from the healthy group to the obese group and then tothe NASH group, being the only abundant phylum exhibiting a signiﬁcant difference

12 9.00E+07 8.00E+07 7.00E+07 6.00E+07 5.00E+07 4.00E+07 3.00E+07 2.00E+07 1.00E+07 0.00E+00 CD 6 CD 18 LPS (ng/mL) Nº Bacter ia g ram – HFCD 6 HFCD 18 HFCD+Synb 6 LPS Gram – HFCD+Synb 18 10 8 6 4 2 0

Figure3 LPSlevels(ng/ml),andGram-negativebacteria.Atweek6,therewasasigniﬁcantincreaseinLPSlevelsandnumber

ofGram-negativebacteria,inthetwoHFCDdietgroups,whencomparingwithCDdiet,notabrogatedbytheadditionofSynb.At week18,althoughlevelsofLPSandnumberofGram-negativebacteriacontinuedtoincreaseinHFCDgroup,theadditionofSynb partiallyabrogatedthiseffect.

(8)

T able 6 Species values. L. casei L. plantarum Leuconostoc Pediococcus Streptococcus Enterobacteriacea Fusobacterium Biﬁdobacterium HFCD vs CD 6 − 8.7 − 1.1 − 14.9 − 10 1 2.5 2.8 1.7 HFCD vs CD 18 − 2.7 − 6.6 − 4.6 4.2 − 5 7.5 10.2 − 47.7 HFCD + Synb vs CD 6 − 8.6 1.3 1.6 1.5 2.5 5.3 41.7 − 8.8 HFCD + Synb vs CD 18 − 2.9 − 6.5 − 56.4 1.5 − 21.8 3.9 5.8 − 14.3 HFCD + Synb vs HFCD 6 1 1.4 23.6 14.7 2.4 2.1 15.1 − 14.6 HFCD + Synb vs HFCD 18 − 1.1 1 − 12.3 − 2.8 − 4.4 − 1.9 − 1.8 3.3 Bacterial variations were considered only for values greater than 2.0 (2 × SD/mean).

betweenthelattertwogroups.34_It_was_furthermore_shown that theProteobacteria were mostly Enterobacteria,and amongthose,mostlyEscherichiacolithatareethanol pro-ducers;theauthorsalsofoundNASHpatientstohavemore elevated blood ethanolconcentrations,34 _similar_to previ-ousstudies.35,36_The_increase_of_{Bifidobacterium}_at_week_18, mayberesponsibleforthebeneficialeffectofthesymbiotic; infactithasbeenpreviouslyshownthatanincreasein Bifi-dobacteriumassociateswithananti-inflammatoryeffect.37 Previous studies showed that obesity associates with a low-gradechronic inflammation that contributes tothe developmentofinsulinresistance,type2diabetesand car-diovasculardiseases.38,39 _In_our_HFCD_diet_model,_mice_did notin fact get obese,but had allthe features of steato-hepatitis and an increase in LPS suggesting an on-going low-grade inflammation. The mechanisms explaining the high-fatinduced, increasein endotoxinlevels,mayrelate to the increase in Gram-bacteria, but it is also possible thathigh-fatfeedingmightinduceanincreaseinintestinal permeability.Infact,Stenmanandcol,foundthathigh-fat feedinginmicesignificantlyincreasedintestinal permeabil-ityofthejejunumandcolon,associatingwithadecreased proportionoffaecalUDCAandincreasedFXRexpression.40 There is also evidence, that high-fat diet induces hyper reactivityagainst low-dose LPS, throughan up-regulation ofCD14byleptin-mediatedsignalling.41

Theobservationofincreasedactivationofhepatic stel-late cells (HSC), and increased fibrosis, is probably due to LPS-mediated signalling through toll-like receptor 4 (TLR4),42 _which _has _been _identified_as _a _fibrogenic _signal inHSC.43_It_is_however_intriguing_that_we_were_not_able_to findanyreductionin thedegree ofnecro-inflammation in micesupplemented with the synbiotic. It is possible that conventionalhistologywasnotsensitive enough todetect minorchangesin necro-inflammation,sincealsoregarding fibrosisdegree, nosignificant changes werechangeswere foundin conventional histology, althoughusing either Sir-iusRed,or alphasmooth-muscleimmunohistochemistry to detectfibrosisorstellatecellactivation,asignificant reduc-tionwasfound.Itispossiblethatamoresensitiveapproach, eventuallyusingimmunohistochemistrytoquantify inflam-matorycells(macrophages,polymorphonuclearleukocytes, andlymphocytesubpopulations) couldclarify theseverity ofcell-deathandassociatedinflammation.

Thepresentfindingsreinforcetherationalfor modulat-ingmicrobiotainthesettingofobesityandNAFLD.Infact, therearealargenumberofstudiesevaluatingtheeffectof probioticsinmurinemodelsandinhumanswithNAFLD,as reviewed.44,45 _However,_we _are_still _missing_{well-designed} trials,toevaluate the effectof mixtures of pre and pro-bioticsinNAFLDthatareverydifficulttoaccomplish.The difficulty results from the slow and rather unpredictable courseofNAFLD,but alsotothe difficultyofassuring the stabilityandtheamountofthepreandprobiotics prepara-tions.

5. Conclusions

So,in summarywe found thatmiceundergoinga high-fat cholinedeﬁcientdietdevelopsteatohepatitis,while simul-taneously leading to microbial dysbiosis, with increased

(9)

endotoxemia and liver ﬁbrosis, this effect being at least partiallycorrected withtheuse aofamixtureofpreand probiotics.Theprospectforusingthesesubstancesin clin-icalpractice,althoughattractive,isnotyetdemonstrated ingoodclinicaltrialsinhumans.Also,weneedtobeaware thatinthisstudy,theSynbiotic2000®_Forte_intervention_was donesimultaneouslywiththediet.Inpractice,weusually intervenewhen thesituation isalready established, what canmakealargedifference.

Ethical

disclosures

Protection of human and animal subjects.The authors declarethat the proceduresfollowed were in accordance withtheregulationsoftherelevantclinicalresearchethics committeeandwiththoseoftheCodeofEthicsoftheWorld MedicalAssociation(DeclarationofHelsinki).

Conﬁdentialityofdata.Theauthorsdeclarethatnopatient dataappearinthisarticle.

Right to privacy and informed consent.The authors declarethatnopatientdataappearinthisarticle.

Funding

This work is supported by a grant from Fundac¸ão para a CiênciaeaTecnologia(PTDC-SAU-OSM/66323/2006).

Conﬂicts

of

interest

Theauthorshavenoconﬂictsofinteresttodeclare.

References

1.LonardoA,BellentaniS,ArgoCK,BallestriS,ByrneCD, Cald-well SH, et al. Epidemiological modiﬁers of non-alcoholic fatty liver disease: focus on high-risk groups. Dig Liver Dis. 2015;47:997---1006.

2.Machado MV, Cortez-Pinto H. Non-alcoholic fatty liver dis-ease:whattheclinicianneedstoknow.WorldJGastroenterol. 2014;20:12956---80.

3.QuigleyEM,MonsourHP.Thegutmicrobiotaandnonalcoholic fattyliverdisease.SeminLiverDis.2015;35:262---9.

4.MachadoMV,Cortez-PintoH.Gutmicrobiotaandnonalcoholic fattyliverdisease.AnnHepatol.2012;11:440---9.

5.LeRoyT,LlopisM,LepageP,BruneauA,RabotS,Bevilacqua C,etal.Intestinalmicrobiotadeterminesdevelopmentof non-alcoholicfattyliverdiseaseinmice.Gut.2013;62:1787---94. 6.MouzakiM,ComelliEM,ArendtBM,BonengelJ,FungSK,Fischer

SE,et al.Intestinal microbiotainpatientswithnonalcoholic fattyliverdisease.Hepatology.2013;58:120---7.

7.Henao-MejiaJ, ElinavE,JinC,HaoL, MehalWZ, StrowigT, etal.Inﬂammasome-mediateddysbiosisregulatesprogression ofNAFLDandobesity.Nature.2012;482:179---85.

8.BoursierJ,MuellerO,BarretM,MachadoM,FizanneL, Araujo-Perez F,et al. Theseverity of NAFLD isassociated withgut dysbiosisandshiftinthemetabolicfunctionofthegut micro-biota.Hepatology.2015.

9.Csak T,Velayudham A, HritzI,PetrasekJ,Levin I,Lippai D, etal.Deﬁciencyinmyeloiddifferentiationfactor-2andtoll-like receptor4expressionattenuatesnonalcoholicsteatohepatitis

andﬁbrosisin mice.AmJPhysiolGastrointestLiver Physiol. 2011;300:G433---41.

10.Miura K, Seki E, Ohnishi H, Brenner DA. Role of toll-like receptorsandtheirdownstreammoleculesinthedevelopment ofnonalcoholic fatty liver disease. Gastroenterol Res Pract. 2010;2010:362847.

11.Soares JB, Pimentel-Nunes P, Roncon-Albuquerque R, Leite-MoreiraA.Theroleoflipopolysaccharide/toll-likereceptor4 signalinginchronicliverdiseases.HepatolInt.2010;4:659---72. 12.FrasinariuOE,CeccarelliS,AlisiA,MoraruE,NobiliV.Gut-liver axisandﬁbrosisinnonalcoholicfattyliverdisease:aninputfor noveltherapies.DigLiverDis.2013;45:543---51.

13.Portela-CidadeJ,Borges-CanhaM,Leite-MoreiraA, Pimentel-NunesP.Systematicreviewoftherelationbetweenintestinal microbiota and toll-like receptors in the metabolic syn-drome: what do we know so far? GE-Port J Gastroenterol. 2015;22:240---58.

14.MartensEC,LoweEC,ChiangH,PudloNA,WuM,McNultyNP, etal.Recognitionanddegradationofplantcellwall polysaccha-ridesbytwohumangutsymbionts.PLoSBiol.2011;9:e1001221. 15.TurnbaughPJ,LeyRE,MahowaldMA,MagriniV,MardisER, Gor-donJI.Anobesity-associatedgutmicrobiome withincreased capacityforenergyharvest.Nature.2006;444:1027---31. 16.Suarez-Zamorano N, Fabbiano S, Chevalier C,Stojanovic O,

ColinDJ,Stevanovic A,et al.Microbiota depletionpromotes browningofwhiteadiposetissueandreducesobesity.NatMed. 2015;21:1497---501.

17.BackhedF,DingH,WangT,HooperLV,KohGY,NagyA,etal. Thegutmicrobiotaasanenvironmentalfactorthatregulates fatstorage.ProcNatlAcadSciUSA.2004;101:15718---23. 18.RaubenheimerPJ,NyirendaMJ,WalkerBR.Acholine-deﬁcient

dietexacerbatesfatty liverbutattenuatesinsulinresistance andglucoseintoleranceinmicefedahigh-fatdiet.Diabetes. 2006;55:2015---20.

19.MalkuschW,RehnB,BruchJ.AdvantagesofSiriusRedstaining forquantitativemorphometriccollagenmeasurementsinlungs. ExpLungRes.1995;21:67---77.

20.MoterA, GobelUB. Fluorescenceinsituhybridization (FISH) fordirectvisualizationofmicroorganisms.JMicrobiolMethods. 2000;41:85---112.

21.AmannR,FuchsBM,BehrensS.Theidentiﬁcationof microor-ganisms by ﬂuorescence in situ hybridisation. Curr Opin Biotechnol.2001;12:231---6.

22.FerollaSM, ArmiliatoGN, Couto CA,Ferrari TC. Therole of intestinalbacteriaovergrowthinobesity-relatednonalcoholic fattyliverdisease.Nutrients.2014;6:5583---99.

23.Miele L, Valenza V, La Torre G, Montalto M, CammarotaG, RicciR,etal.Increasedintestinalpermeabilityandtight junc-tionalterationsinnonalcoholicfattyliverdisease.Hepatology. 2009;49:1877---87.

24.Farhadi A, Gundlapalli S, ShaikhM, FrantzidesC, HarrellL, KwasnyMM, et al.Susceptibility togut leakiness:a possible mechanismforendotoxaemiainnon-alcoholicsteatohepatitis. LiverInt.2008;28:1026---33.

25.HarteAL,daSilvaNF,CreelySJ,McGeeKC,BillyardT, Youssef-ElabdEM,etal.Elevatedendotoxinlevelsinnon-alcoholicfatty liverdisease.JInﬂamm(Lond).2010;7:15.

26.RuizAG,CasafontF,CrespoJ,CayonA,MayorgaM,Estebanez A,etal.Lipopolysaccharide-bindingproteinplasmalevelsand liverTNF-alpha geneexpression inobese patients: evidence forthepotentialroleofendotoxininthepathogenesisof non-alcoholicsteatohepatitis.ObesSurg.2007;17:1374---80. 27.PandeyKR,NaikSR,VakilBV.Probiotics,prebioticsand

synbi-otics---areview.JFoodSciTechnol.2015;52:7577---87. 28.Spindler-VeselA, BengmarkS, VovkI,Cerovic O, KompanL.

Synbiotics,prebiotics,glutamine,or peptideinearlyenteral nutrition:arandomizedstudyintraumapatients.JPENJ Par-enterEnteralNutr.2007;31:119---26.

(10)

29.KnightDJ, GardinerD,Banks A,SnapeSE,WestonVC, Beng-markS,etal.Effectofsynbiotictherapyontheincidenceof ventilatorassociatedpneumoniaincriticallyillpatients:a ran-domised,double-blind,placebo-controlledtrial.IntensiveCare Med.2009;35:854---61.

30.deLaSerreCB,EllisCL,LeeJ,HartmanAL,RutledgeJC, Ray-bouldHE.Propensitytohigh-fatdiet-inducedobesityinratsis associatedwithchangesinthegutmicrobiotaandgut inﬂamma-tion.AmJPhysiolGastrointestLiverPhysiol.2010;299:G440---8. 31.LiDY,YangM,EdwardsS,YeSQ,Nonalcoholicfattyliverdisease: forbetterorworse,blamethegutmicrobiota?JPENJParenter EnteralNutr.2013.

32.LeyRE,BackhedF,TurnbaughP,LozuponeCA,KnightRD,Gordon JI.Obesityaltersgutmicrobialecology.ProcNatlAcadSciUS A.2005;102:11070---5.

33.CarrFJ,ChillD,MaidaN.Thelacticacidbacteria:aliterature survey.CritRevMicrobiol.2002;28:281---370.

34.ZhuL,BakerSS,GillC,LiuW,AlkhouriR,BakerRD,etal. Char-acterizationofgutmicrobiomesinnonalcoholicsteatohepatitis (NASH)patients:aconnectionbetweenendogenousalcoholand NASH.Hepatology.2013;57:601---9.

35.NairS,CopeK,RisbyTH,DiehlAM.Obesityandfemalegender increasebreathethanolconcentration:potentialimplications forthepathogenesisofnonalcoholicsteatohepatitis.AmJ Gas-troenterol.2001;96:1200---4.

36.CopeK,RisbyT,Diehl AM.Increasedgastrointestinalethanol productioninobesemice:implicationsforfattyliverdisease pathogenesis.Gastroenterology.2000;119:1340---7.

37.TurroniF,MarchesiJR, ForoniE,GueimondeM, ShanahanF, MargollesA,etal.Microbiomicanalysisofthebiﬁdobacterial populationinthehumandistalgut.ISMEJ.2009;3:745---51.

38.HotamisligilGS.Inﬂammationandendoplasmicreticulumstress in obesity and diabetes. Int J Obes (Lond). 2008;32 Suppl. 7:S52---4.

39.Shoelson SE, Goldfine AB. Getting away from glucose: fan-ning the flames of obesity-induced inflammation. Nat Med. 2009;15:373---4.

40.Stenman LK, Holma R, Korpela R. High-fat-induced intesti-nalpermeabilitydysfunctionassociatedwithalteredfecalbile acids.WorldJGastroenterol.2012;18:923---9.

41.Imajo K, Fujita K, Yoneda M, Nozaki Y, Ogawa Y, Shi-nohara Y, et al. Hyperresponsivity to low-dose endotoxin during progression to nonalcoholic steatohepatitis is reg-ulated by leptin-mediated signaling. Cell Metab. 2012;16: 44---54.

42.Feng C, Stamatos NM, Dragan AI, Medvedev A, Whitford M, Zhang L, et al. Sialyl residues modulate LPS-mediated signalingthroughtheToll-likereceptor4complex.PLoSONE. 2012;7:e32359.

43.Vespasiani-Gentilucci U, Carotti S, Perrone G, Mazzarelli C, Galati G, Onetti-Muda A, et al. Hepatic toll-like receptor 4 expressionisassociatedwithportalinﬂammationandﬁbrosis inpatientswithNAFLD.LiverInt.2015;35:569---81.

44.SepidehA,KarimP,HosseinA,LeilaR,HamdollahM,Mohammad EG,etal.Effectsofmultistrainprobioticsupplementationon glycemicandinﬂammatoryindicesinpatientswithnonalcoholic fattyliverdisease:adouble-blindrandomizedclinicaltrial.J AmCollNutr.2015:1---6.

45.Iacono A, Raso GM,Canani RB, Calignano A, MeliR. Probio-ticsasanemergingtherapeuticstrategytotreatNAFLD:focus on molecularand biochemical mechanisms. JNutrBiochem. 2011;22:699---711.