HA

AV*DH

BCcvHhsacfipsmcsficigcmi5Csbdioma

Ml1Crrca

GASTROENTEROLOGY 2005;129:246–258

epatitis C Virus–Replicating Hepatocytes Induce Fibrogenicctivation of Hepatic Stellate Cells

NJA SCHULZE–KREBS,* DOROTHEE PREIMEL,* YURY POPOV,*,‡ RALF BARTENSCHLAGER,§

OLKER LOHMANN,§ MASSIMO PINZANI,� and DETLEF SCHUPPAN*,‡

Department of Medicine I, University of Erlangen-Nuremberg, Erlangen-Nuremberg, Germany; ‡Division of Gastroenterology, Beth Israeleaconess Medical Center, Harvard Medical School, Boston, Massachusetts; §Department of Molecular Virology, University of Heidelberg,

eidelberg, Germany; and �Department of Internal Medicine, University of Florence, Florence, Italy

aCoct3mssnr

statovcvi

etamIcll

fmsmti

ackground & Aims: The mechanism by which hepatitisvirus induces liver fibrosis remains largely obscure. To

haracterize the profibrogenic potential of hepatitis Cirus, we used the hepatitis C virus replicon cell lineuh-7 5-15, which stably expresses the nonstructuralepatitis C virus genes NS3 through NS5B, and hepatictellate cells as fibrogenic effector cells. Methods: Ratnd human hepatic stellate cells were incubated withonditioned media from replicon cells, and expression ofbrosis-related genes was quantified by using real-timeolymerase chain reaction, protein, and functional as-ays. Transforming growth factor �1 activity was deter-ined by bioassay. Results: Hepatitis C virus replicon

ells release factors that differentially modulate hepatictellate cell expression of key genes involved in liverbrosis in a clearly profibrogenic way, up-regulating pro-ollagen �1(I) and procollagen �1(III) and down-regulat-ng fibrolytic matrix metalloproteinases. Transformingrowth factor �1 expression and bioactivity were in-reased severalfold in hepatitis C virus–replicating vsock-transfected hepatoma cells. However, transform-

ng growth factor �1 activity was responsible for only0% of the profibrogenic activity. Conclusions: Hepatitisvirus nonstructural genes induce an increased expres-

ion of transforming growth factor �1 and other profi-rogenic factors in infected hepatocytes. The direct in-uction of profibrogenic mediators by hepatitis C virus in

nfected hepatocytes explains the frequent observationf progressive liver fibrosis despite a low level of inflam-ation and suggests novel targets for antifibrotic ther-

pies in chronic hepatitis C.

ore than 200 million people worldwide are in-fected with the hepatitis C virus (HCV),1,2 which

eads to chronic liver disease in 50%–70% of patients,5%–20% of whom develop cirrhosis within 20 years.irrhosis is a prerequisite for the development of HCV-

elated hepatocellular carcinoma, which occurs at a yearlyate of 2%–3%.3 Progression of liver disease does notorrelate with serological or histological inflammation,

nd most infections are diagnosed by chance or in an

dvanced stage, when complications become apparent.urrently, the best available therapy is the combinationf pegylated interferon � and ribavirin.4–6 Althoughombination therapy can lead to viral elimination in upo 50% and 80% of patients with genotypes 1 and 2 or, respectively, less expensive and side effect–free treat-ents are desirable. For those patients who do not re-

pond to antiviral therapy or who are in an advancedtage of their disease, antifibrotic treatment is urgentlyeeded that can halt the progression of fibrosis or eveneverse it.

HCV is an enveloped flavivirus harboring a plus-tranded RNA with approximately 9600 nucleotideshat encode a polyprotein which is cleaved cotranslation-lly and posttranslationally by proteolysis into 10 struc-ural and nonstructural proteins. The liver is the mainrgan of HCV replication, although hepatic levels ofiral RNA and protein are usually low, with mononu-lear cells apart from hepatocytes as suggested sites ofiral replication. HCV exists as 6 genotypes, and type 1s the most prominent in Europe and the United States.7

Liver fibrosis results from excessive accumulation ofxtracellular matrix (ECM) components, a down-regula-ion of ECM-removing matrix metalloproteinases (MMPs),nd an increase of tissue inhibitors of MMPs (TIMPs),ainly TIMP-1. The fibril-forming interstitial collagensand III and the sheet-forming basement membrane

ollagen IV are the most abundant ECM molecules in theiver, with an up to 10-fold increase in cirrhosis.8 Col-agens, MMPs, and TIMPs are mainly produced by myo-

Abbreviations used in this paper: CTGF, connective tissue growthactor; DMEM, Dulbecco’s modified Eagle medium; ECM, extracellularatrix; ELISA, enzyme-linked immunosorbent assay; FCS, fetal calf

erum; HSC, hepatic stellate cell; MF, myofibroblastic cells; MMP,atrix metalloproteinase; PCR, polymerase chain reaction; ROS, reac-

ive oxygen species; TGF, transforming growth factor; TIMP, tissuenhibitor of matrix metalloproteinases.

© 2005 by the American Gastroenterological Association0016-5085/05/$30.00

doi:10.1053/j.gastro.2005.03.089

fivptmatbtcmToMstNmtrloTi

mdcsmshag

t(1ccuhHwp

ioW

twbsM

ibiwrw

�1rSs

scflhciphchtFtbr(

(piae

i

July 2005 HCV REPLICON CELLS DRIVE HEPATIC FIBROGENESIS 247

broblastic cells (MFs), which either derive from acti-ated hepatic stellate cells (HSCs) or activated portal orerivascular fibroblasts.9–11 Activated Kupffer or endo-helial cells, proliferating bile duct epithelia, and otherononuclear cells or even the myofibroblasts themselves

re sources of fibrogenic cytokines and growth factorshat can stimulate HSCs and portal or perivascular fibro-lasts to become MFs.11 A prominent profibrogenic cy-okine is transforming growth factor (TGF)-�1, whichan be released from almost any cell type during inflam-ation, tissue regeneration, and fibrogenesis.11–13 ActiveGF-�1 strongly up-regulates the production and dep-sition of ECM molecules and down-regulates mostMPs. Apart from causing an inflammatory response in

ome patients, HCV can significantly modulate the me-abolism of infected hepatocytes. Thus, core, NS3, andS5A can induce derangement of lipid metabolism,14–16

odulate hepatocyte apoptosis, or alter signal transduc-ion, potentially favoring hepatocarcinogenesis.17–19 De-angement of lipid compartmentalization and metabo-ism by HCV may lead to the production of reactivexygen species (ROS),20 and ROS were shown to induceGF-�1.21 However, a clear profibrogenic role of HCV-

nfected hepatocytes has not been shown.Therefore, we investigated whether and by whatechanisms HCV may induce liver fibrosis, indepen-

ent of inflammation, by using the human hepatomaell line Huh-7 5-15, which stably expresses the non-tructural viral genes NS3 through NS5B, and HSCs/yofibroblasts as fibrogenic effector cells. We could

how that HCV-replicating, but not mock-transfected,epatocytes express increased levels of TGF-�1, as wells other unidentified factors that induce a profibrogenicene-expression pattern in HSCs/MFs.

Materials and Methods

Cells and Cell Culture

The human hepatoma cell line Huh-7 was grown inhe presence of 5% CO2 in Dulbecco’s modified Eagle mediumDMEM; Biochrom, Berlin, Germany) supplemented with0% fetal calf serum (FCS) and antibiotics (100 IU/mL peni-illin and 100 �g/mL streptomycin). For the stably transfectedell lines, neomycin (Invitrogen, Karlsruhe, Germany) wassed at a concentration of 1 mg/mL. The Huh-7 clone 5-15arbors the nonstructural viral genes NS3 through NS5A ofCV genotype 1b, which are responsible for viral replication,hereas the mock-transfected Huh-7 clone contains only thearent vector pcDNA3.22,23

Human HSC/MF and rat CFSC-2G cells, an HSC line withntermediate activation (a gift of Dr M. Rojkind, Departmentsf Biochemistry and Molecular Biology and Pathology, George

ashington University, Washington, DC), were grown under f

he same conditions as the human hepatoma cell line, butithout neomycin. Human HSCs were isolated from wedgeiopsy samples by outgrowth of normal human livers notuitable for transplantation24 and show the characteristics of

Fs.25

Preparation of Conditioned Media

Huh-7 5-15 or Huh-7 pcDNA3 was grown to approx-mately 70% confluency and washed 3 times with phosphate-uffered saline (PBS), followed by culture in DMEM contain-ng 0.2% FCS without neomycin for another 72 hours. Mediaere harvested, centrifuged at 1000 rpm for 10 minutes to

emove cell debris, and stored at 4°C for no longer than 3eeks.

Inhibition of Bioactive Transforming GrowthFactor �

Conditioned medium was supplemented with 250g/mL of rabbit anti–TGF-�1–neutralizing antibody (AB-00-NA; R&D Systems, Wiesbaden, Germany) or unspecificabbit immunoglobulin G of the same concentration (Sigma,t Louis, MO) and incubated for 60 minutes on a rotationhaker at room temperature before addition to cells.

Expression of Extracellular Matrix andFibrosis-Related Genes

A total of 60,000 CFSC-2G cells or human HSCs wereeeded into 9.6 cm2-well plates and incubated in DMEMontaining 10% FCS for 24 hours. After reaching 50% con-uency, cells were washed 3 times with PBS and starved for 24ours with DMEM without FCS, followed by the addition ofonditioned medium of mock-transfected or replicon-contain-ng Huh-7 cells for another 24–72 hours. For coculture ex-eriments, equal cell numbers of stably transfected humanepatoma cells and CFSC-2G rat HSCs were seeded in 25-cm2

ell-culture flasks in DMEM containing 10% FCS. After 24ours, the medium was removed, and cells were washed 3imes in PBS, followed by culture in DMEM containing 0.2%CS for up to 96 hours. After RNA isolation at predefinedime points and reverse transcription, transcript levels of fi-rosis-related messenger RNAs (mRNAs) were quantified byeal-time polymerase chain reaction (PCR) on a LightCyclerRoche, Penzberg, Germany).

RNA Isolation and Preparation ofComplementary DNA

RNA was extracted from the cells by using RNAPurePeqLab, Erlangen, Germany) according to the manufacturer’srotocol. RNA concentration, purity, and integrity were ver-fied by spectrophotometry at 260 and 280 nm and by visu-lization of the 18S and 28S ribosomal RNA bands after gellectrophoresis and ethidium bromide staining.

Transcription of mRNA derived from 0.5 �g of total RNAnto corresponding complementary DNA (cDNA) was per-

ormed by SuperScript II reverse transcriptase (Invitrogen)

wpwui(epFt

T2epam0usfE(etlrt(wmdp

Rbol

bipwiwmTse

wDg9obaowwpp

s

T

HHHVRRRRRRRRRR

H

248 SCHULZE–KREBS ET AL GASTROENTEROLOGY Vol. 129, No. 1

ith 50 pmol of random hexamer and 100 pmol of oligo(dT)rimers (Promega, Mannheim, Germany). In brief, primersere added to the RNA, followed by incubation for 10 min-tes at 70°C in a thermocycler, and were subsequently put once for 5 minutes. After the addition of transcription bufferInvitrogen), RNAsin ribonuclease inhibitor (Promega), andnzyme, annealing was allowed for 10 minutes at room tem-erature, followed by 1 hour of cDNA synthesis at 42°C.inally, the enzyme was denatured at 95°C for 5 minutes, andhe reaction mixture was put on ice.

Quantification of Genes Involved in LiverFibrogenesis by Real-Time PolymeraseChain Reaction

Fluorescence- and quencher-labeled, sequence-specificaqMan probes (MWG, Ebersberg, Germany) were flanked byexternal primers. The probes generate their signal after their

nzymatic cleavage by the 5= to 3= exonuclease activity of Taqolymerase. This cleavage separates the dye from the quenchernd results in an increased fluorescein signal that can beonitored by real-time reverse-transcription PCR. A total of

.5–1 �g of cDNA derived from 0.5–1 �g of total RNA wassed per reaction. A negative water control and 5 internaltandard solutions were added to each LightCycler run. Probesor fibrosis-related mRNAs were designed by using the Primerxpress software according to published sequences

PerkinElmer, Foster City, CA) and positioned to span exon–xon boundaries of corresponding genes to avoid coamplifica-ion of genomic DNA (Table 1). Primers and TaqMan probes,abeled at their 5= end with the reporter dye 6-carboxyfluo-escein and at their 3= end with the quencher 6-carboxy-etramethyl-rhodamine, were synthesized at MWG BiotechEbersberg, Germany). Quantitative analysis was achievedith LightCycler software (version 3.5). The copy number ofRNA in each reaction was computed by using the second

erivative maximum method. Normalization of mRNA ex-

able 1. Specific Primers and TaqMan Probes for Genes Invo

Target molecule 5= Primer Taq

u �2-microglobulin TGACTTTGTCACAGCCCAAGATA TGAu GAPDH GAAGGTGAAGGTCGGAGT CAAu TGF-�1 TGCGTCTGCTGAGGAGGCTCAA ACGiral NS3 CCTTGATGTATCCGTCATACCAACTAG CATTat �2-microglobulin CCGATGTATATGCTTGCAGAGTTAA AACat GAPDH CCTGCCAAGTATGATGACATCAAGA TGGat TGF-�1 AGAAGTCACCCGCGTGCCTAA ACCat procollagen �1(I) TCCGGCTCCTGCTCCTCTTA TTCTat procollagen �1(III) AATGGTGGCTTTCAGTTCAGCT TGGat MMP-2 CCGAGGACTATGACCGGGATAA TCTGat MMP-3 CCGTTTCCATCTCTCTCAAGATGA AGAat MMP-13 GGAAGACCCTCTTCTTCTCA TCTGat TIMP-1 TCCTCTTGTTGCTATCQATTGATAGCTT TTCTat CTGF ATCCCTGCGACCCACACAAG CTC

u, human; FAM, carboxyfluorescein; TAMRA, carboxy-tetramethyl-rho

ression data for sample-to-sample variability in RNA input, c

NA quality, and reverse-transcription efficiency was achievedy dividing copy numbers of target mRNAs by copy numbersf glyceraldehyde phosphate dehydrogenase or �2 microglobu-in mRNA for each run.

Quantification of Total and BioactiveTransforming Growth Factor �1

Total TGF-�1 protein of conditioned media from sta-ly transfected Huh-7 cells was determined by enzyme-linkedmmunosorbent assay (ELISA) according to the manufacturer’srotocol (R&D Systems). For activation, 0.1 mL of 1N HClas added to 0.5 mL of cell culture supernatant, mixed, and

ncubated for 10 minutes at room temperature. The sampleas neutralized by adding 0.1 mL of 1.2/N NaOH in 0.5ol/L HEPES, mixed, and used immediately for the sandwichGF-�1 ELISA. A standard curve was constructed by using

erial dilutions of human recombinant TGF-�1. TGF-�1 lev-ls were measured in duplicate determinations.

Bioactive TGF-�1 was quantified by growth inhibitionith mink lung epithelial cells. Cells were suspended inMEM containing penicillin (100 IU/mL), streptomycin (100/mL), and 10% FCS and placed in a humidified atmosphere of5% air/5% CO2 in an incubator maintained at 37°C. A totalf 50,000 cells per well in 0.1 mL of DMEM containing 0.5%ovine serum albumin were seeded on a 96-well culture dishnd allowed to attach for 3 hours. Next, conditioned mediumr defined concentrations of recombinant TGF-�1 (standard)ere added. After 21 hours of incubation, bromodeoxyuridineas added for 3 hours, followed by a bromodeoxyuridineroliferation ELISA performed according to the manufacturer’srotocol (Roche).

Quantification of Matrix Metalloproteinases1 and 2 and Their Inhibitor TIMP-1

To quantify the content of MMP-1, MMP-2, and theirpecific inhibitor TIMP-1 in the supernatants of human HSCs,

in Liver Fibrogenesis

probe with 5=-FAM � 3=-TAMRA 3= Primer

GCTTACATGTCTCGATCCCA AATCCAAATGCGGCATCTTCCCCGTTCTCAGCC GAAGATGGTGATGGGATTTCGCTGTACCAGAAATACAGCAACA TCCTGGCGATACCTCAGCAATAGCAACGGACGCTCTAATGAC TGAGTCGAAATCGCCGGTAA

ACCTGGGACCGAGACATGTA CAGATGATTCAGAGCTCCATAGAGCAGGCGGCCGAG GTAGCCCAGGATGCCCTTTAGTCAACGCAATCTATGACAAAACCA TCCCGAATGTCTGACGTATTGACATGCGTCAGGAGGG GTATGCAGCTGACTTCAGGGATGT

AAGTCTGAGGAAGGCCAGCTG TGTAATGTTCTGGGAGGCCCCGAGACCGCTATGTCCA CTTGTTGCCCAGGAAAGTGAAGATTCAATCCCTCTATGGACCTCC CAGAGAGTTAGATTGGTGGGTACCAGCATCATCATAACTCCACACGT TCATAGACAGCATCTACTTTGTCCTCGGACCTGGTTATAAGG CGCTGGTATAAGGTGGTCTCGAT

GCCAACCGCAAGAT CAACTGCTTTGGAAGGACTCGC

ne; GAPDH, glyceraldehyde phosphate dehydrogenase.

lved

Man

TGCTGCTTTGGAGTCG

CGTCTGAAGCAATGGC

AAAGCCC

TGGTGTTAGCAA

CCCC

dami

ommercially available human TIMP-1, MMP-1, and MMP-2

EFmBosapdMadMl

d

wpH3msg-c

F5nnmTo

July 2005 HCV REPLICON CELLS DRIVE HEPATIC FIBROGENESIS 249

LISAs (TIMP-1 and active MMP-2: Amersham Bioscience,reiburg, Germany; active MMP-1: Chemicon, Hofheim, Ger-any) were used according to the manufacturer’s protocols.riefly, human HSCs were incubated with conditioned mediaf replicon or mock-transfected cells for up to 72 hours. Theupernatant was harvested after 24, 48, and 72 hours, andliquots were stored at �80°C until use. A total of 20 mmol/L-aminophenylmercuric acid (Sigma, Deisenhofen, Germany)issolved in 0.1N NaOH were used to activate MMP-1 andMP-2 for the measurement of total MMP-1 and MMP-2

ctivity. Standard curves were constructed by using serialilutions of recombinant human TIMP-1, MMP-1, andMP-2, respectively. Measurements were performed in trip-

icate.

Statistics

All statistical analyses were performed with the Stu-

igure 1. HCV-replicating hepatoma cells express enhanced TGF-�1-15 and mock-transfected (pcDNA3) control cells were determined by� 6. (B and C) Time course of TGF-�1 and NS3 mRNA expression

ormalized to glyceraldehyde phosphate dehydrogenase (GAPDH) oreans � SD (n � 4). (D) Total TGF-�1 was assayed in the supernatan

GF-�1 protein concentration was measured by using the mink lung epf mock-transfected cells.

ent t test, and results are given as mean � SD. d

Results

Transforming Growth Factor �1 MessengerRNA and Protein Expression AreUp-regulated in Huh-7 5-15 Replicon Cells

Messenger RNA levels of fibrosis-related genesere analyzed by real-time PCR from the replicon-ex-ressing cell line Huh-7 5-15 and mock-transfecteduh-7 pcDNA cells. Transcript levels of TGF-�1 were

-fold higher in the replicon cells compared with theock-transfected controls (Figure 1A), whereas expres-

ion of other fibrosis-related genes, ie, connective tissuerowth factor (CTGF), procollagen �1(I), and MMP-1,2, and -3, was undetectable (data not shown). Repliconell TGF-�1 and viral NS3 mRNA increased in a time-

A and protein. (A) TGF-�1 transcript levels of HCV-expressing Huh-7titative PCR. Data are means � SD of 2 experiments performed withuh-7 5-15 replicon cells. RNA was extracted from 60,000 cells andcroglobulin mRNA as described in Materials and Methods. Data areplicon and mock-transfected Huh-7 cell lines by ELISA. The bioactivel cell bioassay. *P � .05 vs cells incubated with conditioned medium

mRNquanof H

�2-mit of reithelia

ependent manner, with a 2.5-fold and 2-fold increase,

rohrt

l

dmlmeb

250 SCHULZE–KREBS ET AL GASTROENTEROLOGY Vol. 129, No. 1

espectively, after 96 hours (Figure 1B and C). Thebserved decrease of TGF-�1 transcript levels after 24ours vs 48 hours paralleled a temporary decrease ineplicon (NS3 mRNA) expression as assessed by real-ime PCR.

In line with mRNA levels, TGF-�1 protein (both

atent and activated) was increased 6-fold in the con- t

itioned media of the replicon compared with theock-transfected cells (data not shown). The mink

ung epithelial cell bioassay, a sensitive method ofeasuring bioactive TGF-�1 via inhibition of prolif-

ration of these cells, showed a 3-fold increase ofioactive TGF-�1 in the media of replicon vs mock-

Figure 2. Conditioned media ofreplicon cells induce CTGF butnot TGF-�1 expression in HSC.(A) TGF-�1 and (B) CTGF tran-script levels of CFSC-2G cellswere determined after incuba-tion with conditioned media fordifferent time periods. Mea-surements were performed byquantitative PCR and representmeans � SD (n � 4). *P � .05vs cells incubated with condi-tioned media of mock-trans-fected cells. As a positive con-trol, cells were stimulated with250 pg/mL TGF-�1 (Calbio-chem, Schwalbach, Germany).CM, conditioned medium; CMreplicon, conditioned mediumof HCV replicon–bearing Huh-75-15 cells; CM mock, condi-tioned medium of mock-trans-fected Huh-7 pcDNA3 cells;GAPDH, glyceraldehyde phos-phate dehydrogenase.

ransfected cells (Figure 1D).

etwd

wccaoCw

FrpeaHbfatCro5tfGpmvtf

July 2005 HCV REPLICON CELLS DRIVE HEPATIC FIBROGENESIS 251

Hepatic Stellate Cells Show a ProfibrogenicGene-Expression Profile When Exposed toConditioned Media of Replicon Cells

Moderately activated rat HSCs (CFSC-2G) werexposed to conditioned media of HCV-infected vs mock-ransfected hepatocytes, and TGF-�1 mRNA expressionas measured at different time points. There was no

igure 3. Conditioned media ofeplicon cells increase the ex-ression of the fibrogenic genesncoding (A) procollagen �1(I)nd (B) procollagen �1(II) in ratSCs. CFSC-2G cells were incu-ated with conditioned mediaor the indicated time points,nd gene expression was quan-ified by using real-time PCR.M, conditioned medium; CMeplicon, conditioned mediumf HCV replicon–bearing Huh-7-15 cells; CM mock, condi-ioned medium of mock-trans-ected Huh-7 pcDNA3 cells;APDH, glyceraldehyde phos-hate dehydrogenase. Data areeans � SD (n � 4). *P � .05

s cells incubated with condi-ioned medium of mock-trans-ected cells.

etectable increase in HSC TGF-�1 mRNA, and there c

as no difference between CFSC-2G cells exposed toonditioned medium of mock-transfected vs repliconells, whereas the expected autoinduction of TGF-�1fter the addition of 250 pg/mL of bioactive TGF-�1 wasbserved (Figure 2A). However, a 5-fold increase ofTGF mRNA was found after 48 hours of incubationith medium derived from replicon vs mock-transfected

ells (Figure 2B); this indicates that CTGF might play a

ralwAc2ib

wc7Mdscb

252 SCHULZE–KREBS ET AL GASTROENTEROLOGY Vol. 129, No. 1

ole during HCV-induced activation of these moderatelyctivated HSCs. A 2-fold increase in the CTGF mRNAevel was also detectable through the addition of DMEMithout FCS during the studied time course (Figure 2B).total of 24–72 hours after addition to CFSC-2G HSCs,

onditioned media of HCV replicon cells induced a-fold increase of procollagen �1(I) mRNA and a 5-foldncrease in procollagen �1(III) mRNA vs control vector-

earing cells (Figure 3). Furthermore, MMP-2 mRNA D

as up-regulated 4-fold 48 hours after the addition ofonditioned replicon medium and remained increased at2 hours (Figure 4A). In contrast, transcript levels ofMP-3 (data not shown) and MMP-13 genes were

own-regulated up to 10-fold (Figure 4B). Therefore,ecretion products of replicon-infected hepatoma cellshanged the equilibrium between fibrogenesis and fi-rolysis in exposed HSCs strongly toward fibrogenesis.

Figure 4. Conditioned media ofreplicon cells reciprocally modu-late the expression of (A) fibro-genic MMP-2 mRNA and (B) fi-brolytic MMP-13 mRNA genes inrat HSCs. CFSC-2G cells wereincubated with conditioned me-dia for the indicated timepoints, and gene expressionwas quantified by using real-time PCR. CM, conditioned me-dium; CM replicon, conditionedmedium of HCV replicon–bear-ing Huh-7 5-15 cells; CM mock,conditioned medium of mock-transfected Huh-7 pcDNA3cells; GAPDH, glyceraldehydephosphate dehydrogenase.Data are means � SD (n � 4).*P � .05 vs cells incubatedwith conditioned medium ofmock-transfected cells.

own-regulation of MMP-13 mRNA was accompanied

bTs

Hsl

mwttaM

Fgbamaccmpc

July 2005 HCV REPLICON CELLS DRIVE HEPATIC FIBROGENESIS 253

y a 2-fold increase of the major inhibitor of MMPs,IMP-1, again reflecting a shift to a more fibrogenic

tate (data not shown).Like the moderately activated CFSC-2G cells, humanSCs/MFs also showed a profibrogenic mRNA expres-

ion profile when exposed to conditioned media of rep-

igure 5. Conditioned media of replicon cells induce a profibrogenicene-expression pattern in human hepatic stellate cells/myofibro-last-like cells: (A) CTGF, (B) procollagen �1(I), (C) TIMP-1, (D) MMP-1,nd (E) MMP-2 mRNA. Transcript levels of relevant matrix geneseasured 24 and 48 hours after the addition of conditioned mediare shown. Gene expression was measured by real-time PCR. CM,onditioned medium; CM replicon, conditioned medium of HCV repli-on–bearing Huh-7 5-15 cells; CM mock, conditioned medium ofock-transfected Huh-7 pcDNA3 cells; GAPDH, glyceraldehyde phos-hate dehydrogenase. Data are means � SD (n � 4). *P � .05 vsells incubated with conditioned medium of mock-transfected cells.

icon cells. Thus, a 9-fold increased expression of CTGF m

RNA could be detected after 72 hours (Figure 5A),hereas procollagen �1(I) mRNA increased 4.5-fold af-

er 48 hours (Figure 5B). A trend of down-regulation ofhe collagen I–degrading MMP-1 was accompanied byn increased expression of its inhibitor TIMP-1 and ofMP-2 (Figure 5C–E). Similar to CFSC-2G HSC, hu-

an HSCs did not alter their expression of TGF-�1

wsal

ttsa

emmtgcucmss

dgcamitcgatrgtrg

4Fp(talDchcmd

254 SCHULZE–KREBS ET AL GASTROENTEROLOGY Vol. 129, No. 1

hen exposed to replicon cell– conditioned media buthowed autoinduction after the addition of 250 pg/mLctive TGF-�1. Whereas MMP-2 and TIMP-1 proteinevels secreted into the media of human HSCs showed a

c

rend to increase after the addition of replicon-condi-ioned vs control media after 24–72 hours, total inter-titial collagenase (MMP-1) activity was reduced 5-foldfter 24 hours (Figure 6).

Short-Lived or Short-Range Mediators HaveNo Major Influence on the ProfibrogenicEffects

To characterize the potential profibrogenic influ-nce of short-lived or short-range-acting molecules thatight have been induced in replicon cells, such as lipidediators or ROS, direct cocultures of replicon or con-

rol cells with rat HSCs were performed. The profibro-enic effects observed in these cocultures (Figure 7) wereomparable to those with conditioned media alone (Fig-res 2–4). Therefore, a major influence of such mediatorsould be excluded. Again, no increase in HSC TGF-�1RNA could be detected, even after 96 hours, but a

ignificant decrease in MMP-13 mRNA expression waseen after 96 hours (data not shown).

Profibrogenic Gene Expression Is MainlyDue to Transforming Growth Factor �1

To show that it is indeed TGF-�1 that is pre-ominantly involved in the induction of the profibro-enic gene-expression profile induced in HSCs by repli-on-conditioned media, we used neutralizing TGF-�1ntibodies. After incubation of the cells with conditionededia (72 hours) in the presence of TGF-�1–neutraliz-

ng antibody, 60%–70% of the profibrogenic activity ofhese media could be blocked (Figure 8), but the use ofontrol immunoglobulin G did not influence profibro-enic gene expression (data not shown). Even after theddition of recombinant TGF-�1 protein (250 pg/mL),he blocking activity of the neutralizing antibodyeached only 50%–70% on the basis of HSC profibro-enic gene expression (data not shown). Taken together,hese results indicate that TGF-�1 is the major cytokineeleased from replicon cells that drives profibrogenicene expression in HSCs.

™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™igure 6. Profibrogenic modulation of fibrosis-associated protein ex-ression by conditioned media of replicon cells. Levels of (A) TIMP-1,B) MMP-2, and (C) p-aminophenylmercuric acid–activated (total) in-erstitial collagenase (MMP-1) were measured 24, 48, and 72 hoursfter the addition of conditioned media of mock-transfected and rep-icon cells by using ELISA or a collagen substrate degradation assay.ata are means � SD (n � 3). *P � .05 vs cells incubated withonditioned medium of mock-transfected cells. White bars indicateuman HSCs incubated with conditioned media of mock-transfectedells; gray bars indicate human HSCs incubated with conditionededia of replicon cells. CM, conditioned medium; CM replicon, con-itioned medium of HCV replicon–bearing Huh-7 5-15 cells; CM mock,

onditioned medium of mock-transfected Huh-7 pcDNA3 cells.

wtdctsglHrd

HgwKjbs

iaHs�vaeeeahhgircae

July 2005 HCV REPLICON CELLS DRIVE HEPATIC FIBROGENESIS 255

Discussion

Roughly one third of patients chronically infectedith HCV develop significant liver fibrosis, and many of

hem develop cirrhosis, with a high risk for hepaticecompensation or the development of hepatocellulararcinoma.26 However, little information is available onhe mechanisms by which the virus causes hepatic fibro-is. TGF-�1 plays an important role during liver fibro-enesis by acting as a potent inducer of ECM accumu-ation.25,26 Furthermore, TGF-�1 activates quiescentSCs to smooth muscle �-actin–expressing MFs, up-

egulating their synthesis of interstitial collagens andown-regulating most MMPs.13

We investigated the effects of conditioned medium ofCV replicon cells on the synthesis of ECM-related

enes in HSCs, hypothesizing that infected hepatocytesould release profibrogenic factors. It is established thatupffer cells and activated HSCs/myofibroblasts are ma-

or sources of TGF-�1 in the liver, and hepatocytes andile duct epithelia from injured livers have also been

hown to secrete TGF-�1 and TGF-�2, respectively.27–32 s

We could show that TGF-�1 mRNA is up-regulatedn HCV replicon–bearing hepatocytes; it apparently actss an important profibrogenic mediator released fromCV-infected hepatocytes. No expression of other fibro-

is-related genes, ie, CTGF, procollagen �1(I) and1(III), and MMP-1, -2, and -3, was found. TGF-�1 andiral NS3 mRNA concentrations in replicon cells showedparallel, time-dependent increase, with the highest

xpression levels after 96 hours. This pattern of mRNAxpression was consistent with total TGF-�1 proteinxpression (in its latent and bioactive form), which waslso increased 4-fold compared with mock-transfectedepatocytes. These results strongly suggest that an en-anced hepatocyte expression of the prominent profibro-enic cytokine TGF-�1 at least in part drives HCV-nduced liver fibrosis. Our results are in contrast to thoseeported by Taniguchi et al33 who showed that only theore protein, but not any other of the HCV proteins, wasble to up-regulate TGF-�1 transcription. This could bexplained by different experimental conditions and cell

igure 7. Short-lived or short-range mediators have no influence onrofibrogenic gene expression. CFSC-2G cells were incubated in co-ulture with equal numbers of replicon or mock-transfected cells forhe indicated time points, and expression of (A) procollagen �1(I), (B)MP-2, and (C) procollagen �1(III) mRNA was quantified by using

eal-time PCR. CM, conditioned medium. Data are means � SD (n �). *P � .05 vs cells incubated with conditioned medium of mock-ransfected cells.

FpctMr4t

ystems. Thus, these authors investigated TGF-�1 pro-

mrra

inhotpHt

mHpmwwgnmrplq

ibTicaliepm

tmlCmrChetpnpHptDmiaaEnpm

FtamPtdtg�n

256 SCHULZE–KREBS ET AL GASTROENTEROLOGY Vol. 129, No. 1

otor activity in HepG2 cells after transfection with theespective vectors, but neither TGF-�1 transcriptionates, nor protein levels or bioactivity were characterized,s was done in our work.

Using a monoclonal antibody for immunohistochem-cal studies, Verslype et al34 showed that almost 96% ofaturally occurring or in vitro HCV-infected primaryepatocytes were immunoreactive for HCV-E2 after in-culation with HCV RNA–positive serum. An interac-ion of HCV with liver cells other than hepatocytes isossible. Recently, Bataller et al35 transfected HSCs withCV proteins by using an adenoviral vector. HCV pro-

igure 8. Neutralizing TGF-�1 antibody abrogates most of the induc-ion of profibrogenic genes in rat HSCs. Transcript levels of TGF-�1nd CTGF were measured 72 hours after the addition of conditionededia with or without TGF-�1-neutralizing antibody (250 �g/mL). (A)rocollagen �1(I) and (B) CTGF mRNA levels were measured by real-ime PCR. CM, conditioned medium; CM replicon, conditioned me-ium of HCV replicon–bearing Huh-7 5-15 cells; CM mock, condi-ioned medium of mock-transfected Huh-7 pcDNA3 cells; GAPDH,lyceraldehyde phosphate dehydrogenase. Data are means � SD (n

4). *P � .05 vs cells incubated with conditioned medium withouteutralizing antibody.

eins thus expressed in HSCs induced procollagen I t

RNA expression. The authors concluded that uptake ofCV proteins by HSC via HCV-specific receptors could

lay a role in HCV-induced liver fibrosis through theodulation of profibrogenic signal transduction path-ays by HCV proteins. However, it remains to be shownhether and how far a vector system expressing viralenes in primary HSCs or exposure of HSCs to 1–20g/mL of recombinant HCV proteins, a concentrationuch higher than is found in serum of infected patients,

eflects the in vivo situation. Accordingly, no HCVroteins have been detected in HSCs or MFs of infectedivers by sensitive detection techniques such as PCRuantification of HCV in microdissected livers.36

The up-regulation of TGF-�1 by NS3 through NS5Bn infected hepatocytes that we observed can be explainedy the proven induction of ROS by NS3 and NS5A.17–19

hese HCV proteins interact with mitochondria andnduce both lipid accumulation and degradation, withonsequent derangement of lipid compartmentalizationnd metabolism, thus favoring ROS production. ROSead to the induction of TGF-�1, and TGF-�1 cannduce ROS.18,20,37,38 Conversely, TGF-�1 induces thexpression of procollagen �1(I) mRNA by a hydrogeneroxide–CAAT/enhancer binding protein �–dependentechanism in rat HSCs.21

When we analyzed the response of moderately ac-ivated HSCs to conditioned media of replicon vsock-transfected Huh-7 cells, a pronounced up-regu-

ation of CTGF could be observed. The time course ofTGF expression in HSCs exposed to the conditionededia supports the hypothesis that this factor plays a

ole in HSC activation and progression of fibrosis.39,40

TGF is highly overexpressed at the sites of activeepatic fibrogenesis in fibrotic human liver and inxperimental models of liver disease.39 – 41 It is impor-ant to note that CTGF is produced by HSCs inroportion to their activation status and that exoge-ous CTGF is able to modulate HSC function in arofibrogenic manner, including the promotion ofSC adhesion, proliferation, locomotion, and collagen

roduction.39,40 We also observed a minor up-regula-ion of CTGF mRNA after the addition of pureMEM, and this might indicate a repression of CTGFRNA transcription exerted by other factors present

n FCS. Collectively, our data support a role for CTGFs a sensitive downstream mediator of the fibrogenicctions of TGF-�1, particularly the promotion ofCM production. In contrast to the well-known phe-omenon that TGF-�1 can induce its own gene ex-ression,41 we did not detect induction of TGF-�1RNA transcription through the addition of condi-

ioned media. This can be explained by the lower

(mw�t

rvgclhrdfiisMstfcavMiiTs

mesmtPesfltcbp

uorpweoi

casdr

1

1

1

1

1

1

1

1

July 2005 HCV REPLICON CELLS DRIVE HEPATIC FIBROGENESIS 257

bioactive) TGF-�1 concentration in the conditionededia, ie, at best, 160 pg/mL total TGF-�1 comparedith 250 pg/mL of externally added bioactive TGF-1, a dose also used by others to show autoinduc-

ion.42

We could show that conditioned media of HCV-eplicating hepatocytes can stimulate the expression ofarious key ECM-related genes in a clearly profibro-enic way. Transcript levels of the most abundantollagens in fibrosis, eg, procollagen �1(I) and procol-agen �1(III), showed a significant increase after 48ours. MMP-2, which was also up-regulated by theeplicon-conditioned media, is involved in the degra-ation of the basal lamina, which is partly replaced bybrillar collagens and other interstitial proteins dur-ng fibrosis8,11,43,44 and is therefore considered a de-tructive and profibrogenic MMP. In contrast, levels ofMP-3 (data not shown) and MMP-13 (MMP-1) tran-

cripts were down-regulated up to 10-fold. MMP-13,he rodent equivalent of human MMP-1, is responsibleor the degradation of fibril-forming collagens, mainlyollagens I and III, whereas MMP-3, next to its role asmatrix-degrading proteinase, is regarded as an acti-

ator of other MMPs—mainly MMP-1,45 but alsoMP-13.46 Down-regulation of MMP-13 mRNA and

nterstitial collagenase activity was accompanied by anncrease of the physiological inhibitor of most MMPs,IMP-1, again reflecting a shift to a more fibrogenic

tate.By performing coculture experiments of replicon orock-transfected hepatoma cells with HSCs, we could

xclude a significant contribution of short-range orhort-lived (membrane-associated) mediators thatight be produced in HCV-infected hepatocytes to

he induction of the profibrogenic response in HSCs.rocollagen �1(I) mRNA was transcribed 24 hoursarlier in coculture experiments, as compared with theetting of conditioned medium addition, possibly re-ecting a minor contribution of ROS. Thus, in addi-ion to smad-mediated signal transduction, TGF-�1an also induce procollagen �1(I) mRNA expressiony a faster, hydrogen peroxide–C/EBP�-dependentathway in rat HSCs.21

In summary, we provide evidence that secretion prod-cts of HCV replicon cells strongly induce the expressionf a profibrogenic response and suppress the fibrolyticesponse in HSCs/myofibroblasts. TGF-�1 is the mainrofibrogenic factor released from these hepatocytes,hereas other factors, such as CTGF, act as downstream

ffectors of TGF-�1 in HSCs. Although further analysisf the involved signal transduction pathways operative in

nfected hepatocytes may allow the development of spe-

ific antifibrotic therapies for patients with chronic hep-titis C, our data also support the usefulness of virusuppression, eg, by low-dose interferon � in patients whoid not reach viral elimination after interferon-�/ribavi-in combination therapy.3–6,38

References1. Alter MJ. Epidemiology of hepatitis C. Hepatology 1997;26:62S–

65S.2. Wasley A, Alter MJ. Epidemiology of hepatitis C: geographic dif-

ferences and temporal trends. Semin Liver Dis 2000;20:1–16.3. Poynard T, Ratziu V, Charlotte F, Goodman Z, McHutchison J,

Albrecht J. Rates and risk factors of liver fibrosis progression inpatients with chronic hepatitis C. J Hepatol 2001;34:730–739.

4. McHutchison JG, Gordon SC, Schiff ER, Shiffman ML, Lee WM,Rustgi VK, Goodman ZD, Ling MH, Cort S, Albrecht JK. Interferonalfa-2b alone or in combination with ribavirin as initial treatmentfor chronic hepatitis C. Hepatitis Interventional Therapy Group.N Engl J Med 1998;339:1485–1492.

5. Poynard T, McHutchison J, Manns M, Trepo C, Lindsay K, Good-man Z, Ling MH, Albrecht J. Impact of pegylated interferon alfa-2band ribavirin on liver fibrosis in patients with chronic hepatitis C.Gastroenterology 2002;122:1303–1313.

6. Poynard T, Marcellin P, Lee SS, Niederau C, Minuk GS, Ideo G,Bain V, Heathcote J, Zeuzem S, Trepo C, Albrecht J. Randomisedtrial of interferon alpha2b plus ribavirin for 48 weeks or for 24weeks versus interferon alpha2b plus placebo for 48 weeks fortreatment of chronic infection with hepatitis C virus. InternationalHepatitis Interventional Therapy Group (IHIT). Lancet 1998;352:1426–1432.

7. Bartenschlager R, Lohmann V. Replication of hepatitis C virus.J Gen Virol 2000;81:1631–1648.

8. Schuppan D, Ruehl M, Somasundaram R, Hahn EG. Matrix as amodulator of hepatic fibrogenesis. Semin Liver Dis 2001;21:351–372.

9. McCrudden R, Iredale JP. Liver fibrosis, the hepatic stellate celland tissue inhibitors of metalloproteinases. Histol Histopathol2000;15:1159–1168.

0. Knittel T, Kobold D, Saile B, Grundmann A, Neubauer K, PiscagliaF, Ramadori G. Rat liver myofibroblasts and hepatic stellate cells:different cell populations of the fibroblast lineage with fibrogenicpotential. Gastroenterology 1999;117:1205–1221.

1. Friedman SL. Molecular regulation of hepatic fibrosis, an inte-grated cellular response to tissue injury. J Biol Chem 2000;275:2247–2250.

2. Bissell DM. Chronic liver injury, TGF-beta, and cancer. Exp MolMed 2001;33:179–190.

3. Gressner AM, Weiskirchen R, Breitkopf K, Dooley S. Roles ofTGF-beta in hepatic fibrosis. Front Biosci 2002;7:d793–d807.

4. Lerat H, Honda M, Beard MR, Loesch K, Sun J, Yang Y, Okuda M,Gosert R, Xiao SY, Weinman SA, Lemon SM. Steatosis and livercancer in transgenic mice expressing the structural and nonstruc-tural proteins of hepatitis C virus. Gastroenterology 2002;122:352–365.

5. Perlemuter G, Sabile A, Letteron P, Vona G, Topilco A, Chretien Y,Koike K, Pessayre D, Chapman J, Barba G, Brechot C. HepatitisC virus core protein inhibits microsomal triglyceride transfer pro-tein activity and very low density lipoprotein secretion: a model ofviral-related steatosis. FASEB J 2002;16:185–194.

6. Moriya K, Yotsuyanagi H, Shintani Y, Fujie H, Ishibashi K, Mat-suura Y, Miyamura T, Koike K. Hepatitis C virus core proteininduces hepatic steatosis in transgenic mice. J Gen Virol 1997;78:1527–1531.

7. Gong G, Waris G, Tanveer R, Siddiqui A. Human hepatitis C virus

NS5A protein alters intracellular calcium levels, induces oxidative

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

4

4

4

4

4

GHc2

AN

l

258 SCHULZE–KREBS ET AL GASTROENTEROLOGY Vol. 129, No. 1

stress, and activates STAT-3 and NF-kappa B. Proc Natl Acad SciU S A 2001;98:9599–9604.

8. Moriya K, Nakagawa K, Santa T, Shintani Y, Fujie H, Miyoshi H,Tsutsumi T, Miyazawa T, Ishibashi K, Horie T, Imai K, Todoroki T,Kimura S, Koike K. Oxidative stress in the absence of inflamma-tion in a mouse model for hepatitis C virus-associated hepato-carcinogenesis. Cancer Res 2001;11:4365–4370.

9. Arima N, Kao CY, Licht T, Padmanabhan R, Sasaguri Y, Pad-manabhan R. Modulation of cell growth by the hepatitis C virusnonstructural protein NS5A. J Biol Chem 2001;276:12675–12684.

0. De Bleser PJ, Xu G, Rombouts K, Rogiers V, Geerts A. Glutathionelevels discriminate between oxidative stress and transforminggrowth factor-beta signaling in activated rat hepatic stellate cells.J Biol Chem 1999;247:33881–33887.

1. Garcia-Trevijano ER, Iraburu MJ, Fontana L, Dominguez-RosalesJA, Auster A, Covarrubias-Pinedo A, Rojkind M. Transforminggrowth factor beta1 induces the expression of alpha1(I) procol-lagen mRNA by a hydrogen peroxide-C/EBPbeta-dependent mech-anism in rat hepatic stellate cells. Hepatology 1999;29:960–970.

2. Lohmann V, Korner F, Koch J, Herian U, Theilmann L, Barten-schlager R. Replication of subgenomic hepatitis C virus RNAs ina hepatoma cell line. Science 1999;285:110–113.

3. Pietschmann T, Lohmann V, Rutter G, Kurpanek K, Barten-schlager R. Characterization of cell lines carrying self-replicatinghepatitis C virus RNAs. J Virol 2001;75:1252–1264.

4. Casini A, Pinzani M, Milani S, Grappone C, Galli G, Jezequel AM,Schuppan D, Rotella CM, Surrenti C. Regulation of extracellularmatrix synthesis by transforming growth factor beta 1 in humanfat-storing cells. Gastroenterology 1993;105:245–253.

5. Cassiman D, Libbrecht L, Desmet V, Denef C, Roskams T. He-patic stellate cell/myofibroblast subpopulations in fibrotic humanand rat livers. J Hepatol 2002;36:200–209.

6. Alter HJ, Seeff LB. Recovery, persistence, and sequelae in hep-atitis C virus infection: a perspective on long-term outcome.Semin Liver Dis 2000;20:17–35.

7. Castilla A, Prieto J, Fausto N. Transforming growth factors beta 1and alpha in chronic liver disease. Effects of interferon alfatherapy. N Engl J Med 1991;324:933–940.

8. Bissell DM, Roulot D, George J. Transforming growth factor betaand the liver. Hepatology 2001;34:859–867.

9. Milani S, Herbst H, Schuppan D, Stein H, Surrenti C. Transform-ing growth factors beta 1 and beta 2 are differentially expressedin fibrotic liver disease. Am J Pathol 1991;139:1221–1229.

0. Armendariz-Borunda J, Seyer JM, Kang AH, Raghow R. Regulationof TGF beta gene expression in rat liver intoxicated with carbontetrachloride. FASEB J 1990;4:215–221.

1. Zhang LJ, Yu JP, Li D, Huang YH, Chen ZX, Wang XZ. Effects ofcytokines on carbon tetrachloride-induced hepatic fibrogenesis inrats. World J Gastroenterol 2004;10:77–81.

2. Westhoff JH, Sawitza I, Keski-Oja J, Gressner AM, Breitkopf K.PDGF-BB induces expression of LTBP-1 but not TGF-beta1 in a ratcirrhotic fat storing cell line. Growth Factors 2003;21:121–130.

3. Taniguchi H, Kato N, Otsuka M, Goto T, Yoshida H, Shiratori Y,Omata M. Hepatitis C virus core protein upregulates transforminggrowth factor-beta 1 transcription. J Med Virol 2004;72:52–59.

4. Verslype C, Nevens F, Sinelli N, Clarysse C, Pirenne J, Depla E,Maertens G, van Pelt J, Desmet V, Fevery J, Roskams T. Hepaticimmunohistochemical staining with a monoclonal antibody

against HCV-E2 to evaluate antiviral therapy and reinfection of D

liver grafts in hepatitis C viral infection. J Hepatol2003;38:208–214.

5. Bataller R, Paik YH, Lindquist JN, Lemasters JJ, Brenner DA.Hepatitis C virus core and nonstructural proteins induce fibro-genic effects in hepatic stellate cells. Gastroenterology 2004;126:529–540.

6. Vona G, Tuveri R, Delpuech O, Vallet A, Canioni D, Ballardini G,Baptiste Trabut J, Le Bail B, Nalpas B, Carnot F, Pol S, Brechot C,Thiers V. Intrahepatic hepatitis C virus RNA quantification inmicrodissected hepatocytes. J Hepatol 2004;40:682–688.

7. Pociask DA, Sime PJ, Brody AR. Asbestos-derived reactive oxygenspecies activate TGF-beta(1). Lab Invest 2004;8:1013–1023.

8. Schuppan D, Krebs A, Bauer M, Hahn EG. Hepatitis C and liverfibrosis. Cell Death Differ 2003;10(Suppl 1):59–67.

9. Paradis V, Dargere D, Bonvoust F, Vidand M, Segarini P, BedossaP. Effects and regulation of connective tissue tissue growthfactor on hepatic stellate cells. Lab Invest 2002;82:767–774.

0. Rachfal AW, Brigstock DR. Connective tissue growth factor(CTGF/CCN2) in hepatic fibrosis. Hepatol Res 2003;26:1–9.

1. Sedlaczek N, Jia JD, Bauer M, Herbst H, Ruehl M, Hahn EG,Schuppan D. Proliferating bile duct epithelial cells are a majorsource of connective tissue growth factor in rat biliary fibrosis.Am J Pathol 2001;158:1239–1244.

2. Choi ME, Kim EG, Huang Q, Ballermann BJ. Rat mesangial cellhypertrophy in response to transforming growth factor-beta 1.Kidney Int 1993;44:948–958.

3. Adhikary LP, Yamamoto T, Isome M, Nakano Y, Kawasaki K,Yaoita E, Kihara I. Expression profile of extracellular matrix andits regulatory proteins during the process of interstitial fibrosisafter anti-glomerular basement membrane antibody-induced glo-merular sclerosis in Sprague-Dawley rats. Pathol Int 1999;49:716–725.

4. Pauly RR, Passaniti A, Bilato C, Monticone R, Cheng L, Papado-poulos N, Gluzband YA, Smith L, Weinstein C, Lakatta EG, et al.Migration of cultured vascular smooth muscle cells through abasement membrane barrier requires type IV collagenase activityand is inhibited by cellular differentiation. Circ Res 1994;75:41–54.

5. Benyon RC, Arthur MJ. Extracellular matrix degradation and therole of hepatic stellate cells. Semin Liver Dis 2001;21:373–384.

6. Moilanen M, Sorsa T, Stenman M, Nyberg P, Lindy O, VesterinenJ, Paju A, Konttinen YT, Stenman UH, Salo T. Tumor-associatedtrypsinogen-2 (trypsinogen-2) activates procollagenases (MMP-1,-8, -13) and stromelysin-1 (MMP-3) and degrades type I collagen.Biochemistry 2003;42:5414–5420.

Received June 21, 2004. Accepted March 23, 2005.Address requests for reprints to: Detlef Schuppan, MD, PhD, Division of

astroenterology and Hepatology, Beth Israel Deaconess Medical Center,arvard Medical School, Dana 506, 330 Brookline Ave, Boston, Massa-husetts 02215. e-mail: dschuppa@bidmc.harvard.edu; fax: (617) 667-767.A.S.-K. and D.P. contributed equally to this article.Supported by the Interdisciplinary Center of Clinical Research (Grant

15) of the University of Erlangen-Nuremberg and by the Germanetwork of Hepatitis (Hepnet).The authors thank Dr. Margarete Goppelt-Struebe for providing mink

ung epithelial cells and Juergen Kressel, Margit Plommer, and Joerg

istler for excellent technical assistance.