osteopontin is induced by hedgehog pathway activation and

Osteopontin is Induced by Hedgehog PathwayActivation and Promotes Fibrosis Progression in

Nonalcoholic SteatohepatitisWing-Kin Syn,1,2 Steve S. Choi,1,3 Evaggelia Liaskou,2 Gamze F. Karaca,1 Kolade M. Agboola,1 Ye Htun Oo,2

Zhiyong Mi,4 Thiago A. Pereira,1,5 Marzena Zdanowicz,1 Padmini Malladi,6 Yuping Chen,1 Cynthia Moylan,1,3

Youngmi Jung,1,7 Syamal D. Bhattacharya,4 Vanessa Teaberry,1 Alessia Omenetti,1 Manal F. Abdelmalek,1

Cynthia D. Guy,8 David H. Adams,2 Paul C. Kuo,4 Gregory A. Michelotti,1

Peter F. Whitington,6 and Anna Mae Diehl1

Nonalcoholic steatohepatitis (NASH) is a leading cause of cirrhosis. Recently, we showed thatNASH-related cirrhosis is associated with Hedgehog (Hh) pathway activation. The gene encod-ing osteopontin (OPN), a profibrogenic extracellular matrix protein and cytokine, is a directtranscriptional target of the Hh pathway. Thus, we hypothesize that Hh signaling induces OPNto promote liver fibrosis in NASH. Hepatic OPN expression and liver fibrosis were analyzed inwild-type (WT) mice, Patched-deficient (Ptc1/2) (overly active Hh signaling) mice, and OPN-deficient mice before and after feeding methionine and choline–deficient (MCD) diets to induceNASH-related fibrosis. Hepatic OPN was also quantified in human NASH and nondiseased liv-ers. Hh signaling was manipulated in cultured liver cells to assess direct effects on OPN expres-sion, and hepatic stellate cells (HSCs) were cultured in medium with different OPN activities todetermine effects on HSC phenotype. When fed MCD diets, Ptc1/2 mice expressed more OPNand developed worse liver fibrosis (P < 0.05) than WTmice, whereas OPN-deficient mice exhib-ited reduced fibrosis (P < 0.05). In NASH patients, OPN was significantly up-regulated and cor-related with Hh pathway activity and fibrosis stage. During NASH, ductular cells stronglyexpressed OPN. In cultured HSCs, SAG (an Hh agonist) up-regulated, whereas cyclopamine (anHh antagonist) repressed OPN expression (P < 0.005). Cholangiocyte-derived OPN andrecombinant OPN promoted fibrogenic responses in HSCs (P < 0.05); neutralizing OPN withRNA aptamers attenuated this (P < 0.05). Conclusion: OPN is Hh-regulated and directly pro-motes profibrogenic responses. OPN induction correlates with Hh pathway activity and fibrosisstage. Therefore, OPN inhibition may be beneficial in NASH (HEPATOLOGY 2011;53:106-115)

Nonalcoholic steatohepatitis (NASH) is apotentially serious form of chronic liverinjury because it increases the risk of develop-

ing cirrhosis and primary liver cancer. The mechanismsthat lead to these outcomes have not been fully eluci-dated, but they appear to involve responses triggered

Abbreviations: aSMA, a-smooth muscle actin; AIH, autoimmune hepatitis; ALD, alcoholic liver disease; Gli, glioblastoma; Hh, Hedgehog; HSC, hepatic stellatecell; MCD, methionine and choline–deficient; MF, myofibroblasts; MF-HSC, myofibroblastic hepatic stellate cell; mRNA, messenger RNA; NAFLD, nonalcoholicfatty liver disease; NASH, nonalcoholic steatohepatitis; NKT, natural killer T; OPN, osteopontin; PBC, primary biliary cirrhosis; PSC, primary sclerosingcholangitis; Ptc, Patched; rOPN, recombinant osteopontin; Q-HSC, quiescent hepatic stellate cell; QRT-PCR, quantitative reverse-transcription polymerase chainreaction; WT, wild-type.From the 1Division of Gastroenterology, Department of Medicine, the 4Department of Surgery, and the 8Department of Pathology, Duke University Medical Center,

Durham, NC; the 2Centre for Liver Research and National Institute of Health Research Biomedical Research Unit, University of Birmingham, Edgbaston, Birmingham,UK; the 3Section of Gastroenterology, Department of Medicine, Durham Veteran Affairs Medical Center, Durham, NC; 5Nucleo de Deoncas Infecciosas, Centro deCiencias da Caude, Universidade Federal do Espirito Santo, Espirito Santo, Brazil; the 6Department of Pediatrics, The Feinberg School of Medicine, NorthwesternUniversity, Children’s Memorial Institute for Education and Research, Chicago, IL; and the7Department of Biological Sciences, Pusan National University, Pusan, Korea.Received August 11, 2010; accepted September 13, 2010.Supported by National Institutes of Health Grants RO1 DK053792 and RO1 DK077794 (to A. M. D.) with contributions from the Medical Research

Council, Wellcome Trust, the European Union, and National Institutes of Health Grant 5R01AA14257-5 (to D. H. A.) and a European Association for the Studyof the Liver Fellowship (to W. K. S.).Address reprint requests to: Anna Mae Diehl, M.D, Florence McAlister Professor & Chief, Division of Gastroenterology, Duke University, Snyderman Building,

Suite 1073, 595 LaSalle Street, Durham, NC 27710. E-mail: [email protected]; fax: 919-684-4183; or Wing-Kin Syn, M.D., Clinician-Scientist andHonorary Consultant Hepatolagist, University of Birmingham, 5th Floor IBR, BIS 2TT, UK. E-mail: [email protected].

106

by hepatocyte apoptosis1,2 and myofibroblast accumu-lation.3 Certain apoptotic stimuli have been reportedto induce hepatocyte production of Hedgehog (Hh)ligands.4 Hh ligands, in turn, elicit several fibrogenicactions by engaging their receptors on Hh-responsiveliver cells, such as ductular type cells, hepatic stellatecells (HSCs), and natural killer T (NKT) cells. InHSCs, for example, Hh pathway activation functionsin a cell-autonomous fashion to promote transition ofquiescent HSCs (Q-HSCs) to myofibroblastic HSCs(MF-HSCs), enhance MF-HSC proliferation, and in-hibit MF-HSC apoptosis.5 Activating Hh signaling inother types of liver cells (such as ductular cells andNKT cells) also causes these cells to generatefactors that promote MF-HSC accumulation throughparacrine mechanisms.6,7

Osteopontin (OPN), a proinflammatory cytokineand integrin-binding ligand,8-10 is highly expressed inmany inflamed tissues and plays a critical role inwound healing.11-13 Recently, glioblastoma (Gli) bind-ing sites were demonstrated in the OPN promoter,prompting speculation that Hh signaling might regu-late OPN transcription.14 This concept is potentiallyrelevant to NASH-related liver fibrosis, because Hhpathway activity increases in parallel with fibrosis stagein NASH. Moreover, in other tissues, OPN is secretedby cells that mediate fibrogenic repair in NASH (suchas NKT cells and fibroblasts).15,16 Evidence that OPNmessenger RNAs (mRNAs) increase during culture-related activation of Q-HSCs to MF-HSCs16 and cor-relate with fibrosis severity in biliary atresia17 furthersupport a potential role for OPN in the pathogenesisof cirrhosis. Therefore, we manipulated Hh pathwayactivity in mice and cultured cells to determine effectson OPN production, and examined whether reductionof OPN affected Hh signaling or fibrogenesis. Theresults support and advance the concept that OPN isan Hh target gene and reveal a previously unsuspectedrole for OPN as a proximal mediator of the fibrogenicactions of Hh in NASH.

Materials and MethodsAnimals. C57BL/6 Patched-deficient (Ptcþ/�) mice

were obtained from R. J. Wechsler-Reya (Duke Uni-versity, NC), and wild-type (WT) mice were obtained

from The Jackson Laboratory (Bar Harbor, ME).Ptcþ/� mice have only one copy of Ptc, an Hh path-way repressor. Therefore, these mice are unable tosilence Hh signaling and exhibit excessive Hh pathwayactivity. WT and Ptcþ/� mice were fed a methionineand choline–deficient (MCD) diet to induce nonalco-holic steatohepatitis (NASH) and liver fibrosis, or con-trol chow (n ¼ 8/group) for 8 weeks. Additionally,129/SvJ Black-Swiss OPN-deficient (OPN�/�) miceand littermate controls were fed the MCD diet or con-trol chow, respectively (n ¼ 6/group). Because 129/SvJmice were reported to be more sensitive to an MCDdiet than C57Bl/6 mice,18 OPN�/� mice and litter-mate controls were fed the diets for 4 weeks ratherthan 8 weeks.Animal care and procedures were approved by the

Duke University and Northwestern University Institu-tional Animal Care and Use Committees as set forthin the National Institutes of Health Guide for the Careand Use of Laboratory Animals.Histopathologic Analysis. Serial sections were

stained with hematoxylin and eosin. Nonalcoholic fattyliver disease (NAFLD) severity was assessed using crite-ria described by Brunt et al.19 (Supporting Informa-tion Materials and Methods). To quantify liver fibrosis,5-lm sections were stained with Picrosirus red (Sigma,St. Louis, MO) and counterstained with fast green(Sigma, St. Louis, MO). Collagen stained with Siriusred was quantitated in the sections that were randomlychosen (under �20 magnification, 10 fields each fromsample) as described.20

Immunohistochemical Analysis. To localize andcharacterize cells that produce and/or respond to Hhligands and OPN, formalin-fixed, paraffin-embeddedlivers were prepared for immunohistochemical analysisas described.6,7 Protocols and antibodies used are listedin Supporting Information Materials and Methods andSupporting Information Table 1.Molecular Techniques. Real-time quantitative

reverse-transcription polymerase chain reaction (QRT-PCR) and western immunoblot analysis were per-formed using established protocols.7 Details are pro-vided in the Supporting Information Materials andMethods and Supporting Information Table 2.Isolation and Culture of Primary Rodent Hepatic

Stellate Cells. Hepatic stellate cells (HSCs) were iso-lated from normal Sprague-Dawley rats as described21

CopyrightVC 2010 by the American Association for the Study of Liver Diseases.View this article online at wileyonlinelibrary.com.DOI 10.1002/hep.23998Potential conflict of interest: Nothing to report.Additional Supporting Information may be found in the online version of this article.

HEPATOLOGY, Vol. 53, No. 1, 2011 SYN ET AL. 107

(Supporting Information Materials and Methods). Asimilar isolation/culture protocol was used for studiesinvolving mouse primary HSCs.6 Day 4 HSCs wereused in all experiments.Human HSCs. The human HSC line LX-2 was

cultured in serum-supplemented Dulbecco’s modifiedEagle’s medium. Primary human HSCs were isolatedas described.21

Studies of Hh and OPN in Cultured HSCs. Toevaluate the effects of Hh signaling on HSCs, day 4HSC cultures were grown for an additional 24 hoursin medium containing either exogenous Hh agonist(SAG) at a concentration of 0.3 lM, 5 lM cyclop-amine (Toronto Research Chemicals, Inc., Toronto,ON, Canada), an inhibitor of Hh signaling, or 5 lMtomatidine (Calbiochem, San Diego, CA), a catalyti-cally inactive analog of cyclopamine5,21 (tomatidineserves as a control for cyclopamine).In separate experiments, recombinant OPN (rOPN)

or vehicle was added to cultures to assess their effects onHSC activation. A 100-ng/mL dose was used in thisstudy because it stimulated the greatest effects in vitro.22

The effects of inactivating OPN were subsequentlyassessed by treating HSCs with the human OPN RNAaptamer OPN-R3 or its biologically inactive mutant,OPN-R3-2 (both synthesized by Dharmacon, Lafayette,CO).23 Aptamers (100 nmol/L) were added to mediumfor 48 hours before harvest. This concentration of OPNaptamer has been shown to inhibit adhesion, migration,and invasion in the breast cancer cell line MDA-MB-231 (which highly expresses OPN and is a standard toolfor evaluating OPN actions).23

All cell experiments were performed at least in dupli-cate. Total RNA and protein were harvested beforetreatment and at the end of treatment and were analyzedby way of QRT-PCR and immunoblotting, respectively.Effect of Cholangiocyte-Conditioned Media on

Primary HSC Activation. The immortalized but non-transformed murine immature cholangiocyte cell line603B was maintained in six-well plates (Costar 3516,Corning Incorporated) in standard culture media asdescribed.24,25 At 90% confluence, cholangiocyte-con-ditioned media were harvested and added to primaryHSC cultures together with OPN-targeted RNAaptamers or null aptamers. HSCs were harvested 2days later, and mRNA expression was analyzed by wayof QRT-PCR. Experiments were performed in dupli-cate and were repeated twice.Human Liver Studies. Formalin-fixed, paraffin-em-

bedded liver sections from deidentified controls andsubjects with biopsy-proven NASH and explanted livertissues from individuals undergoing liver transplanta-

tion for NASH or ASH-related cirrhosis (n ¼ 6/group) from the Liver and Hepatobiliary Unit, Bir-mingham, UK, and Department of Pathology at DukeUniversity were used. Normal tissues were obtainedfrom nondiseased livers removed during resection forcolorectal hepatic metastases or from split-liver grafts.Freshly explanted livers with cirrhosis related to auto-

immune hepatitis (AIH), primary sclerosing cholangitis(PSC), and primary biliary cirrhosis (PBC) were alsosnap-frozen and used for total liver RNA analyses.All studies using material from Duke University

Hospital were conducted in accordance with NationalInstitutes of Health and institutional guidelines forhuman subject research. Studies of samples acquiredfrom the Hepatobiliary Unit in Birmingham were per-formed in accordance with local ethical approval 04/Q2708/41 and REC 2003/242 from the South-Bir-mingham Research Ethics Committee, UK (Support-ing Information Materials and Methods and Support-ing Information Table 1).

Statistical Analysis. The results are expressed as themean 6 SEM. Statistical significance was determinedusing Student t test and was set at P < 0.05.

ResultsUp-regulation of OPN Parallels Hh Pathway

Activation During MCD Diet–Induced NASH. WTmice (n ¼ 8/group) develop hepatic necro-inflamma-tion, accumulate markers of MF-HSCs, and exhibitliver fibrosis after 8 weeks of an MCD diet. Ptcþ/�

mice (with haplo-insufficiency of Ptc, a factor thatconstrains Hh signaling) exhibit worse liver fibrosisthan WT mice after MCD diet exposure (Fig. 1A;Supporting Information Fig. 1A-D).6,7 These findingssuggests that Hh pathway activation promotes fibrosisprogression in this model of NASH, but the exactmechanisms involved have not yet been determined.Hh pathway activation results in nuclear accumula-

tion of Gli proteins (downstream targets of Hh signal-ing),26 and Gli proteins may regulate transcription ofOPN, a potential profibrogenic factor.27-29 Therefore,we compared OPN expression in WT mice that werefed either control or MCD diets for 8 weeks. In WTmice, MCD diets caused a significant induction ofOPN mRNA (>10-fold) and protein (approximatelytwo-fold) expression (Fig. 1B,C; Supporting Informa-tion Fig. 1E). Ptcþ/� mice exhibited even greater up-regulation of OPN expression when fed an MCD diet(Fig. 1B,C; Supporting Information Fig. 2A,B). Thelatter finding supports the concept that Hh signalingincreases OPN expression.

108 SYN ET AL. HEPATOLOGY, January 2011

In both strains, the OPN-immunoreactive cells weremostly ductular in appearance (Fig. 1D). Double im-munostaining for Gli2 and OPN in liver samples fromNASH patients demonstrated that Gli2 (þ) ductularcells that coexpressed OPN localized within fibrousseptae (Fig. 1E).MCD Diet–Fed OPN-Deficient Mice Develop Less

Fibrosis. To evaluate whether OPN directly contributedto the fibrogenic response evoked by MCD diets and gainfurther insight into the relationship between OPN andthe Hh pathway, OPN�/� mice (n ¼ 12) and littermatecontrols (n ¼ 12) were fed MCD or control diets for 4weeks. OPN deficiency had no obvious effect on expres-sion of the Hh target gene Gli2, because both OPN�/�

mice and littermate controls showed similar induction ofGli2 mRNA (data not shown) and protein (Fig. 2A) after4 weeks of an MCD diet. Despite apparent similarities inHh pathway activity, however, the fibrogenic responses ofOPN�/� mice were markedly attenuated when comparedwith their littermate controls. After MCD diet feeding,for example, OPN�/� mice accumulated 50% fewer a-smooth muscle actin (aSMA)–positive cells (Fig. 2B) andsignificantly fewer Sirius red–stained fibrils (Fig. 2C) thancomparably treated littermates. These results are consist-ent with an earlier report of reduced collagen gene expres-sion in OPN�/� mice18 and suggest that the Hh pathwaymediates its fibrogenic effects, at least in part, by inducingexpression of OPN.

Fig. 1. OPN expression parallelsHh pathway activation and fibrosisduring diet-induced NASH. WT andPtcþ/� mice (with an overly activeHh pathway) were fed control chow(n ¼ 8/group) or an MCD diet (n¼ 8/group) for 8 weeks and thensacrificed. (A) Sirius red quantifica-tion by morphometric analysis. Sec-tions from four animals were usedper group, and 10 randomlyselected, 200� fields chosen foranalysis with Metaview software.(B) Total hepatic OPN expressionby way of western blot and densi-tometry analysis. (C) RepresentativeOPN immunostaining after MCDtreatment in WT and Ptcþ/� mice(magnification �400), and quanti-fication by morphometry. Sectionsfrom four animals were used pergroup and eight randomly selected,200� fields chosen for analysisusing Metaview software. Resultsare expressed as fold change (%positive staining) relative to chow-fed control mice and are presentedas the mean 6 SEM. *P < 0.05versus control mice. (D) Represen-tative OPN (þ) biliary ductular cellsin murine NASH-related fibrosis(�630). (E) Representative OPN(blue) and Gli2 (brown) double-positive ductular cells in humanNASH-related cirrhosis (magnifica-tion �400). [Color figure can beviewed in the online issue, which isavailable at wileyonlinelibrary.com.]


Paracrine/Autocrine OPN Stimulates HSCExpression of Fibrogenic Phenotype. Hh-responsivebile ductular cells are major sources of OPN (Fig. 1E).Therefore, we treated primary cultures of rodent HSCswith conditioned medium from monocultures of acholangiocyte cell line, and assessed effects on HSCgene expression. To determine whether HSC responseswere mediated by OPN, studies were repeated usingcholangiocyte-conditioned medium plus OPN-targetedaptamers. Cholangiocyte-conditioned media aug-mented HSC expression of aSMA (Fig. 3A) and colla-gen (Fig. 3B); RNA aptamer treatment repressedaSMA induction by 50% and returned collagen

expression to basal values, proving that paracrine sig-naling involving OPN promoted fibrogenic geneexpression in HSCs. In separate studies, other primaryHSC cultures were treated with rOPN (100 ng/mL)or vehicle for 24 hours, and RNA was analyzed byway of QRT-PCR (Fig. 3C,D). rOPN also augmentedexpression of aSMA (Fig. 3C) and collagen Ia1 expres-sion (Fig. 3D). These findings support the conceptthat exogenous OPN can function as a paracrine factorto promote fibrogenic gene expression in HSCs.Because it has been reported that MF-HSCs them-

selves also express OPN,16 we next investigatedchanges in endogenous OPN gene expression during

Fig. 3. Paracrine OPN stimulates HSC acti-vation and collagen expression. (A,B) Isolatedmouse primary HSCs were cultured for 4 daysand treated with conditioned media (CM)from cholangiocytes for 48 hours, with orwithout OPN-targeted RNA aptamers. (C,D) Inseparate experiments, primary HSCs weredirectly treated with rOPN (0 or 100 ng/mL)for 24 hours and harvested for QRT-PCR anal-ysis. (A,C) aSMA. (B,D) Collagen mRNA.Results of duplicate experiments are pre-sented as the mean 6 SEM. *P < 0.05 ver-sus vehicle-treated stellate cells.

Fig. 2. OPN-deficient mice develop less fibrosis after MCD diet treatment. 129/SvJ OPN-deficient mice and littermate controls were fed anMCD diet or control chow (n ¼ 6 mice/group/dietary treatment) for 4 weeks. At the end of treatment, mice were sacrificed. (A) Accumulation ofGli2 (þ) cells in OPN�/� and littermates. Sections from six animals were used, and eight randomly selected, 400� fields chosen for cell count-ing. (B) aSMA morphometry. Sections from three animals were used at each time point, and 10 randomly selected, 400� fields were chosenfor analysis with Metaview software. (C) Sirius red quantification by way of morphometric analysis. Sections from six animals were used, and 10randomly selected 400� fields were chosen for analysis. Results are expressed as fold change relative to chow-fed littermates and are pre-sented as the mean 6 SEM. *P < 0.05 versus control mice.


spontaneous culture-related activation of Q-HSCs toMF-HSCs. We confirmed that HSC expression ofOPN mRNA and protein increased significantly as Q-HSCs transitioned to becoming MF-HSCs (Fig. 4A;Supporting Information Fig. 3A). Addition of OPNaptamers to day 4 cultures significantly repressedaSMA and collagen gene expression, providing novelevidence that HSC-derived OPN may help to main-tain the myofibroblastic phenotype of cultured HSCs.As HSCs become MF-HSCs, they repress expressionof the Hh inhibitor Hhip, induce expression of Hhligands, and up-regulate various myofibroblastic geneswhile down-regulating markers of quiescence.5 To clar-ify the relationship between Hh pathway activationand OPN expression, day 4 culture-activated MF-HSCs were treated with cyclopamine to selectively in-hibit Hh signaling. Cultures were harvested 24 hourslater, RNA was isolated, gene expression was assessedby way of QRT-PCR, and results were compared withparallel cultures that had been treated with tomatidine,an inactive cyclopamine analog. Inhibiting Hh signalingwith cyclopamine attenuated induction of OPN geneexpression (Fig. 4B; Supporting Information Fig. 4B).

Conversely, addition of the Hh agonist SAG augmentedOPN gene expression significantly (Fig. 4C). Therefore,endogenous OPN gene expression in MF-HSCs is regu-lated, at least in part, by the Hh pathway.To determine the relative importance of OPN as a

downstream target of the Hh pathway in HSCs, day 4primary MF-HSCs from Ptcþ/� mice were incubatedwith OPN aptamers for 48 hours. At baseline, HSCsfrom Ptcþ/� mice expressed three-fold more gli2mRNA than WT HSCs, confirming that Ptc defi-ciency enhances Hh signaling (Fig. 5A). Consistentwith our in vivo findings (Fig. 1), Ptcþ/� HSCs alsoexpressed more OPN mRNA than WT HSCs (Fig.5B). Neutralizing OPN significantly reduced collagenand aSMA mRNA levels in Ptc-deficient HSCs (Fig.5C,D) but had little effect on gli2 mRNA expression(Fig. 5E). These findings suggest that the Hh pathwaymediates induction of certain fibrogenic genes indi-rectly through up-regulation of the Hh-responsivegene OPN.

Evidence of OPN Overexpression in Humans withProgressive NAFLD. Because the mouse model ofMCD diet–induced NASH differs from human NASH

Fig. 4. Hh-regulated OPN overexpression promotes HSC activation in an autocrine fashion. Rat primary HSCs were isolated and cultured for 0,4, and 7 days. At the end of treatment, HSCs were washed and cell lysates were obtained for western blot analysis. (A) OPN western blot anddensitometry analysis. (B,C) In separate experiments, day 4 HSC cultures were treated with OPN aptamers or null aptamers (control) for 48 hoursand RNA isolated for QRT-PCR. (D,E) To assess whether OPN expression was Hh-regulated, day 4 HSCs were treated with cyclopamine (Cyc), to-matidine (Tom), or SAG for 24 hours and harvested for RNA analysis. Results of duplicate experiments are presented as the mean 6 SEM. *P< 0.05 versus vehicle-treated or Q-HSCs.


in several aspects,30 it was important to determinewhether OPN overexpression occurred in human patientswith NAFLD. Coded liver sections from 11 patients withwell-characterized NAFL (n ¼ 3), NASH (n ¼ 3), andNASH-related cirrhosis (n ¼ 5) were stained to demon-strate OPN and analyzed using computer-assisted mor-phometry. Expression of OPN was lowest in patients withNAFL and highest in patients with NASH-related cirrho-sis (Fig. 6A,B). To further validate the association betweenNAFLD-related fibrosis stage and OPN expression, totalliver RNA was isolated from a separate cohort of 36patients with early (stage F0-F1) or advanced (stage F3-F4) NASH-related fibrosis (n ¼ 18/group) and analyzedby way of QRT-PCR. In livers with advanced fibrosis,OPN gene expression was double that of livers with earlyfibrosis (Fig. 6C). Additional analysis was conductedusing RNA and protein harvested from explanted liverswith NASH-related cirrhosis and residual tissue fromnondiseased donor livers (n ¼ 6/disease). Livers withNASH-related cirrhosis contained over 10 times moreOPN protein (Fig. 6D; Supporting Information Fig. 4)and 5 times more OPN mRNA compared with nondi-seased control livers (Fig. 7A). Interestingly, OPN wasalso significantly increased in livers from individuals withalcoholic liver disease (ALD)-related cirrhosis (Fig. 7A,B;Supporting Information Fig. 4), PBC (Fig. 7A,C), AIH,and PSC (Fig. 7A), suggesting that OPN induction is aconserved response to chronic liver injury.

Discussion

Cirrhosis rarely occurs unless NAFLD progresses toNASH, a condition that is characterized by inflamma-tion and hepatocyte death.31 However, only a minorityof individuals with NASH actually develop cirrhosis.Fibrosis stage in NASH has been shown to correlatewith the level of hepatic apoptotic activity.1,32 Indexliver biopsy specimens of NASH patients who eventu-ally progress to cirrhosis also harbor more myofibro-blasts (MFs) than specimens of patients who do notprogress to cirrhosis.3 These observations link hepato-cyte apoptosis with myofibroblast accumulation and fi-brosis progression in NASH.There is further evidence that phagocytosis of apo-

ptotic debris stimulates HSCs to become MFs.33

Recently, we identified a potentially related mechanismby which hepatocyte apoptosis promotes MF accumu-lation and liver fibrosis by demonstrating that dyinghepatocytes produce Hh ligands.4 Biologically activeHh ligands are detected in apoptotic fragmentsreleased by ligand-producing cells.34 Hh ligands, inturn, engage various types of Hh-responsive cells,including HSCs, ductular-type cells, and NKT cells, totrigger fibrogenic responses.7,35,36 Consistent withthese findings, we observed that Hh activity correlatedwith MF accumulation and fibrosis stage in NAFLDpatients and in rodent models of NAFLD, suggestingthat interindividual differences in Hh signaling

Fig. 5. OPN is a downstream target of the Hh pathway. WT and Ptcþ/� primary HSCs were placed in culture for 4 days and then harvestedfor RNA analysis by way of QRT-PCR. (A) gli2 RNA. (B) OPN mRNA. Results of duplicate experiments are presented as the mean 6 SEM. *P <0.05 versus WT HSCs. In separate experiments, Ptcþ/� HSCs were treated with OPN aptamers or null aptamers (control) for 48 hours, andeffects on fibrogenesis were assessed by way of QRT-PCR. (C) Collagen mRNA. (D) aSMA mRNA. (E) gli2 mRNA. Results of triplicate experimentsare presented as the mean 6 SEM. *P < 0.05 versus null aptamer (control)-treated HSCs.


influenced intensity of fibrogenic responses duringNASH.7 The present study provides further supportfor this concept, because it identifies OPN as a proxi-mal effector of Hh-mediated fibrogenesis, and demon-strate that livers of OPN-deficient mice are signifi-cantly protected from NASH-related fibrosis.A recent analysis of the OPN gene revealed binding

sites for Gli transcription factors, suggesting that OPNtranscription is likely to be regulated by Hh signal-ing.14 By demonstrating colocalization of Gli andOPN in liver cells, and proving that expression ofOPN mRNA is increased by a Smoothened agonistbut decreased by a Smoothened antagonist, our resultssupport and advance this concept. In addition, ourdata demonstrate that changes in OPN gene expressionare paralleled by changes in OPN protein content andbiological activity, because OPN aptamers reverse the

profibrogenic actions of OPN. The latter findings alsoverify that OPN is a significant downstream target ofHh signaling (rather than vice versa), because neutral-izing OPN had no effect on cellular expression of theHh target gene Gli2 but significantly diminished fibro-genic gene expression, even in Ptc-deficient cells withsupranormal Hh pathway activity. Coupled with theevidence that OPN�/� mice are significantly protectedfrom NASH-related fibrosis, the data explain why wenoted that hepatic content of Gli2þ cells and OPNþ

cells increased in parallel as liver fibrosis advanced inpatients with NAFLD.Although the present studies focused on the roles of

OPN as a paracrine factor for cholangiocyte-stellatecell fibrogenic interactions, and an autocrine mediatorof fibrogenic gene expression in MF-HSCs, other celltypes might also contribute to the fibrogenic actions of

Fig. 6. OPN overexpression in NAFLDpatients. Coded liver sections from patientswith NAFL, NASH, and NASH-related cirrhosiswere stained for OPN and analyzed by way ofcomputer-assisted morphometry. (A) Repre-sentative photomicrographs of liver sectionsfrom patients with NAFLD and normal liverfrom excess donor tissues (magnification�400). (B) Quantitative analysis of OPN inall patients. Amount of OPN is expressed aspercentage of stained cells per high-poweredfield. *P < 0.05 versus normal. Total liverRNA was isolated from individuals (n ¼ 18/group) with early or late NASH-related fibrosisand analyzed by way of QRT-PCR. (C) OPNmRNA. Results are presented as the mean 6SEM. *P < 0.05 versus early fibrosis. (D) He-patic OPN protein from normal individualsand individuals with ALD-related cirrhosis andNASH-related cirrhosis. [Color figure can beviewed in the online issue, which is availableat wileyonlinelibrary.com.]


OPN in NASH. NKT cells are particularly noteworthyin this regard. These liver-enriched immune cells arecapable of producing and responding to Hh ligands35

and are also known to secrete OPN.15 To our knowl-edge, the possibility that Hh signaling might regulateOPN expression in NKT cells has not been evaluated.However, we have demonstrated that Hh pathway acti-vation enhances hepatic accumulation of NKT cells.6

We and others6,37 have also shown that hepatic NKTcell content is significantly increased in patients withNASH-related cirrhosis. Moreover, activated liverNKT cells generate soluble factors that evoke expres-sion of fibrogenic genes in cultured HSCs, and micethat are genetically deficient in NKT cells are relativelyprotected from NASH-related fibrosis,6 similar toOPN-deficient mice.Therefore, OPN induction may represent a con-

served profibrogenic mechanism among several distincttypes of Hh-responsive liver cells, including ductularcells, MF-HSCs, and NKT cells. Such reasoning sug-gests that interindividual differences in OPN produc-tion may contribute to differences in the outcomes ofNASH. Indeed, OPN may also dictate the fibrogenicresponse in other chronic liver diseases, because it is sig-nificantly overexpressed in livers with cirrhosis relatedto ALD, AIH, PBC, and PSC, and a recent studyreported that plasma OPN levels correlate with hepaticinflammation and fibrosis in chronic hepatitis C.38

Although more work is needed to delineate the interac-tions between OPN and other putative profibrogenicfactors,39 this concept suggests that OPN levels mayprovide a useful biomarker for liver fibrosis in NASH,and that OPN neutralization might be useful for pre-venting progressive hepatic fibrosis in NASH patients.

Acknowledgment: We thank Robert J. Wechsler-Reya (Duke University Medical Center, Durham, NC)for providing Ptcþ/� mice, Gregory J. Gores (MayoClinic, Rochester, MN) and Yoshiyuki Ueno (TohokuUniversity, Sendai, Japan) for providing the 603B cellline, Marcus Rojkind (George Washington University,Washington, DC) for providing the 8B cell line, andScott L. Friedman (Mount Sinai School of Medicine,New York, NY) for providing the LX-2 cell line. Weare grateful to Mari Shinohara (Duke University Medi-cal Center) and Toshimitsu Uede (Hokkaido Univer-sity, Sapporo, Japan) for helpful discussions. Finally,we thank Jiawen Huang for assistance with animalcare and Carl Stone for administrative support.

References1. Feldstein AE, Canbay A, Angulo P, Taniai M, Burgart LJ, Lindor KD,

et al. Hepatocyte apoptosis and fas expression are prominent features ofhuman nonalcoholic steatohepatitis. Gastroenterology 2003;125:437-443.

2. Wieckowska A, Zein NN, Yerian LM, Lopez AR, McCullough AJ,Feldstein AE. In vivo assessment of liver cell apoptosis as a novelbiomarker of disease severity in nonalcoholic fatty liver disease.HEPATOLOGY 2006;44:27-33.

Fig. 7. Enrichment of hepatic OPN inpatients with chronic liver disease. RNAs wereharvested from explanted livers with ALD-,NASH-, PBC-, PSC- and AIH-related cirrhosisand analyzed by way of QRT-PCR. Liver sec-tions from ALD- and PBC-related cirrhosiswere immunostained for OPN. (A) OPN mRNA(n ¼ 6/group). Results are expressed as foldchange relative to normal donor livers and arepresented as the mean 6 SEM. *P < 0.05versus normal. (B) Representative OPN immu-nostaining in ALD-related cirrhosis (magnifica-tion �100) and (C) PBC-related cirrhosis(magnification �100). [Color figure can beviewed in the online issue, which is availableat wileyonlinelibrary.com.]


3. Feldstein AE, Papouchado BG, Angulo P, Sanderson S, Adams L, GoresGJ. Hepatic stellate cells and fibrosis progression in patients with non-alcoholic fatty liver disease. Clin Gastroenterol Hepatol 2005;3:384-389.

4. Jung Y, Witek RP, Syn WK, Choi SS, Omenetti A, Premont R, et al.Signals from dying hepatocytes trigger growth of liver progenitors. Gut2010;59:655-665.

5. Choi SS,Omenetti A,Witek RP,Moylan CA,Syn WK,Jung Y, et al.Hedgehog pathway activation and epithelial-to-mesenchymal transitionsduring myofibroblastic transformation of rat hepatic cells in cultureand cirrhosis. Am J Physiol Gastrointest Liver Physiol 2009;297:G1093-G1106.

6. Syn WK, Oo YH, Pereira TA, Karaca GF, Jung Y, Omenetti A, et al.Accumulation of natural killer T cells in progressive nonalcoholic fattyliver disease. HEPATOLOGY 51:1998-2007.

7. Syn WK, Jung Y,Omenetti A, Abdelmalek M, Guy CD, Yang L, et al.Hedgehog-mediated epithelial-to-mesenchymal transition and fibrogenicrepair in nonalcoholic fatty liver disease. Gastroenterology 2009;137:1478-1488.

8. Ashkar S, Weber GF, Panoutsakopoulou V, Sanchirico ME, Jansson M,Zawaideh S, et al. Eta-1 (osteopontin): an early component of type-1(cell-mediated) immunity. Science 2000;287:860-864.

9. Chabas D, Baranzini SE, Mitchell D, Bernard CC, Rittling SR, DenhardtDT, et al. The influence of the proinflammatory cytokine, osteopontin, onautoimmune demyelinating disease. Science 2001;294:1731-1735.

10. Rangaswami H, Bulbule A, Kundu GC. Osteopontin: role in cell sig-naling and cancer progression. Trends Cell Biol 2006;16:79-87.

11. Miyazaki K, Okada Y, Yamanaka O, Kitano A, Ikeda K, Kon S, et al.Corneal wound healing in an osteopontin-deficient mouse. Invest Oph-thalmol Vis Sci 2008;49:1367-1375.

12. Fujita N, Fujita S, Okada Y, Fujita K, Kitano A, Yamanaka O, et al.Impaired angiogenic response in the corneas of mice lacking osteopon-tin. Invest Ophthalmol Vis Sci 2010;51:790-794.

13. Lorena D, Darby IA, Gadeau AP, Leen LL, Rittling S, Porto LC, et al.Osteopontin expression in normal and fibrotic liver. Altered liver heal-ing in osteopontin-deficient mice. J Hepatol 2006;44:383-390.

14. Das S, Harris LG, Metge BJ, Liu S, Riker AI, Samant RS, et al. Thehedgehog pathway transcription factor GLI1 promotes malignantbehavior of cancer cells by up-regulating osteopontin. J Biol Chem2009;284:22888-22897.

15. Diao H, Kon S, Iwabuchi K, Kimura C, Morimoto J, Ito D, et al.Osteopontin as a mediator of NKT cell function in T cell-mediatedliver diseases. Immunity 2004;21:539-550.

16. Lee SH, Seo GS, Park YN, Yoo TM, Sohn DH. Effects and regulationof osteopontin in rat hepatic stellate cells. Biochem Pharmacol 2004;68:2367-2378.

17. Whitington PF, Malladi P, Melin-Aldana H, Azzam R, Mack CL, SahaiA. Expression of osteopontin correlates with portal biliary proliferationand fibrosis in biliary atresia. Pediatr Res 2005;57:837-844.

18. Sahai A, Malladi P, Melin-Aldana H, Green RM, Whitington PF. Up-regulation of osteopontin expression is involved in the development ofnonalcoholic steatohepatitis in a dietary murine model. Am J PhysiolGastrointest Liver Physiol 2004;287:G264-G273.

19. Brunt EM, Janney CG, Di Bisceglie AM, Neuschwander-Tetri BA, Ba-con BR. Nonalcoholic steatohepatitis: a proposal for grading and stag-ing the histological lesions. Am J Gastroenterol 1999;94:2467-2474.

20. Yata Y, Gotwals P, Koteliansky V, Rockey DC. Dose-dependent inhibi-tion of hepatic fibrosis in mice by a TGF-beta soluble receptor: impli-cations for antifibrotic therapy. HEPATOLOGY 2002;35:1022-1030.

21. Choi SS, Witek RP, Yang L, Omenetti A, Syn WK, Moylan CA, et al.Activation of Rac1 promotes hedgehog-mediated acquisition of themyofibroblastic phenotype in rat and human hepatic stellate cells.HEPATOLOGY 2010;52:278-290.

22. Heilmann K, Hoffmann U, Witte E, Loddenkemper C, Sina C,Schreiber S, et al. Osteopontin as two-sided mediator of intestinalinflammation. J Cell Mol Med 2009;13:1162-1174.

23. Mi Z, Guo H, Russell MB, Liu Y, Sullenger BA, Kuo PC. RNAaptamer blockade of osteopontin inhibits growth and metastasis ofMDA-MB231 breast cancer cells. Mol Ther 2009;17:153-161.

24. Yahagi K, Ishii M, Kobayashi K, Ueno Y, Mano Y, Niitsuma H, et al.Primary culture of cholangiocytes from normal mouse liver. In VitroCell Dev Biol Anim 1998;34:512-514.

25. Ishimura N, Bronk SF, Gores GJ. Inducible nitric oxide synthase up-regulates Notch-1 in mouse cholangiocytes: implications for carcino-genesis. Gastroenterology 2005;128:1354-1368.

26. Omenetti A, Diehl AM. The adventures of sonic hedgehog in develop-ment and repair. II. Sonic hedgehog and liver development, inflamma-tion, and cancer. Am J Physiol Gastrointest Liver Physiol 2008;294:G595-G598.

27. Mori R, Shaw TJ, Martin P. Molecular mechanisms linking woundinflammation and fibrosis: knockdown of osteopontin leads to rapidrepair and reduced scarring. J Exp Med 2008;205:43-51.

28. Morimoto J, Inobe M, Kimura C, Kon S, Diao H, Aoki M, et al.Osteopontin affects the persistence of beta-glucan-induced hepatic gran-uloma formation and tissue injury through two distinct mechanisms.Int Immunol 2004;16:477-488.

29. Berman JS, Serlin D, Li X, Whitley G, Hayes J, Rishikof DC, et al.Altered bleomycin-induced lung fibrosis in osteopontin-deficient mice.Am J Physiol Lung Cell Mol Physiol 2004;286:L1311-L1318.

30. Larter CZ, Yeh MM. Animal models of NASH: getting both pathologyand metabolic context right. J Gastroenterol Hepatol 2008;23:1635-1648.

31. Syn WK, Teaberry V, Choi SS, Diehl AM. Similarities and differencesin the pathogenesis of alcoholic and nonalcoholic steatohepatitis. SeminLiver Dis 2009;29:200-210.

32. Canbay A, Higuchi H, Bronk SF, Taniai M, Sebo TJ, Gores GJ. Fasenhances fibrogenesis in the bile duct ligated mouse: a link between ap-optosis and fibrosis. Gastroenterology 2002;123:1323-1330.

33. Canbay A, Feldstein AE, Higuchi H, Werneburg N, Grambihler A,Bronk SF, et al. Kupffer cell engulfment of apoptotic bodies stimu-lates death ligand and cytokine expression. HEPATOLOGY 2003;38:1188-1198.

34. Witek RP, Yang L, Liu R, Jung Y, Omenetti A, Syn WK, et al. Livercell-derived microparticles activate hedgehog signaling and alter geneexpression in hepatic endothelial cells. Gastroenterology 2009;136:320-330.

35. Syn WK, Witek RP, Curbishley SM, Jung Y, Choi SS, Enrich B, et al.Role for hedgehog pathway in regulating growth and function of invar-iant NKT cells. Eur J Immunol 2009;39:1879-1892.

36. Omenetti A, Porrello A, Jung Y, Yang L, Popov Y, Choi SS, et al.Hedgehog signaling regulates epithelial-mesenchymal transition duringbiliary fibrosis in rodents and humans. J Clin Invest 2008;118:3331-3342.

37. Tajiri K, Shimizu Y, Tsuneyama K, Sugiyama T. Role of liver-infiltrat-ing CD3þCD56þ natural killer T cells in the pathogenesis of nonal-coholic fatty liver disease. Eur J Gastroenterol Hepatol 2009;21:673-680.

38. Huang W, Zhu G, Huang M, Lou G, Liu Y, Wang S. Plasma osteo-pontin concentration correlates with the severity of hepatic fibrosis andinflammation in HCV-infected subjects. Clin Chim Acta 2010;411:675-678.

39. Fickert P, Thueringer A, Moustafa T, Silbert D, Gumhold J, Tsybrov-skyy O, et al. The role of osteopontin and tumor necrosis factor alphareceptor-1 in xenobiotic-induced cholangitis and biliary fibrosis inmice. Lab Invest 2010;90:844-852.


osteopontin is induced by hedgehog pathway activation and

Documents