Mdr2 (Abcb4)-/- mice spontaneously develop severe biliary fibrosis

via massive dysregulation of pro- and antifibrogenic genes

Yury Popov1,3,†, Eleonora Patsenker1,†, Peter Fickert2, Michael Trauner2, Detlef Schuppan1,3,*

1Laboratory of Liver Research, Department of Medicine I, University of Erlangen-Nuremberg, Germany2Laboratory of Experimental and Molecular Hepatology, Division of Gastroenterology and Hepatology,

Department of Medicine, Medical University Graz, Austria3Division of Gastroenterology and Hepatology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA 02215, USA

Background/Aims: Mdr2 (Abcb4)-/- mice develop hepatic lesions resembling primary sclerosing cholangitis. Our aim

was to characterize the evolution of fibrosis in Mdr2-/- mice.

Methods: Mdr2-/-mice and their wild-type littermates were sacrificed at 2, 4 and 8 weeks after birth. Hepatic collagen

was determined biochemically. Fibrosis related transcript levels were quantified from livers by real-time RT-PCR, and

MMP activities determined by substrate assays. Liver histology was assessed by connective tissue staining and

immunohistochemistry for a-smooth muscle actin (a-SMA).

Results: Mdr2-/- mice demonstrated a time-dependent increase of relative and total hepatic collagen (fivefold at 8

weeks, compared to wildtype controls), and maximal a-SMA immunoreactivity at 4 weeks. Compared to wildtype

controls profibrogenic mRNA levels for procollagen a1(I), TGFb1, TGFb2, MMP-2 and -13, TIMP-1, PDGFb

receptor, and PAI-1 were upregulated up to 27-fold. Most transcripts peaked at 4 weeks, but procollagen a1(I) mRNA

increased steadily, TIMP-1 mRNA was constantly elevated (20-fold), MMP-13 mRNA was suppressed and interstitial

collagenase and gelatinase activities were downregulated.

Conclusions: Mdr2-/- mice spontaneously progress to severe biliary fibrosis. This is due to a characteristic temporal

pattern of upregulated profibrogenic and downregulated fibrolytic genes and activities. These mice are an attractive

model to test potential antifibrotics for the treatment of (biliary) liver fibrosis.

q 2005 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved.

Keywords: Animal model; Antifibrotics; Bile duct; Collagen; Liver fibrosis; Knockout mice; PAI-1; PDGF receptor;

Procollagen; PSC; TGF-b; TIMP-1

0168-8278/$30.00 q 2005 European Association for the Study of the Liver. Pub

doi:10.1016/j.jhep.2005.06.025

Received 22 April 2005; received in revised form 23 June 2005; accepted

27 June 2005; available online 15 August 2005* Corresponding author. Address: Division of Gastroenterology and

Hepatology, Beth Israel Deaconess Medical Center, Harvard Medical

School, Dana 501, 330 Brookline Ave, Boston, MA 02215, USA. Tel.: C1

617 6678377/9755041; fax: C1 617 6672767.

E-mail address: dschuppa@bidmc.harvard.edu (D. Schuppan).

Abbreviations: HSC, hepatic stellate cell; HYP, hydroxyproline; Mdr2,

multidrug resistance protein 2/canalicular phospholipids export pump; MF,

myofibroblast; MMP, matrix metalloproteinase; PAI-1, plasminogen

activator inhibitor-1; PDGFR-b, receptor for platelet-derived growth

factor-b; PSC, primary sclerosing cholangitis; a-SMA, a-smooth muscle

actin; TGFb, transforming growth factor-b; TIMP, tissue inhibitor of

matrix metalloproteinases.

† Both authors equally contributed to the work.

1. Introduction

Liver cirrhosis, the frequent consequence of chronic liver

disease, is associated with a high morbidity and mortality. In

chronic cholangiopathies such as PBC and PSC develop-

ment of cirrhosis is frequent and often requires transplan-

tation as the only effective treatment modality. Therapies

that halt disease progression are of limited efficacy and

pharmacological reversion of advanced fibrosis and

cirrhosis is unavailable. In view of the growing demand

for liver transplantation confronted with an increasing donor

shortage, effective antifibrotic therapies are urgently

needed.

Potential antifibrotic agents have to be tested in vivo.

Several experimental fibrosis models have been developed

Journal of Hepatology 43 (2005) 1045–1054

www.elsevier.com/locate/jhep

lished by Elsevier B.V. All rights reserved.

Y. Popov et al. / Journal of Hepatology 43 (2005) 1045–10541046

in rats and mice. So far there is no suitable animal model for

biliary fibrosis, since ligation of the common bile duct

results in a rapidly progressive interstitial (biliary) fibrosis,

associated with massive proliferation of bile ducts which is

only rarely observed in man [1]. Thus, there is a need for a

biliary fibrosis model, which resembles human pathology

and is characterized by a progressive, homogeneous and

reproducible liver fibrosis, allowing for testing of large

numbers of potential drug candidates and their

combinations.

Mice deficient in the canalicular phospholipid flippase

(Mdr2/Abcb4-/- mice) spontaneously develop liver injury, due

to the absence of phospholipid from bile [2], which leads to

morphological features of primary sclerosing cholangitis

(PSC) [3]. Moreover, these animals can be considered an

animal model for human MDR3 deficiency ranging from

progressive familial intrahepaic cholestasis type 3 to adult

liver cirrhosis [4]. The hepatic pathogenesis in Mdr2-/- mice

was recently characterized as a sequence of events that include

disruption of tight junctions and basement membranes of bile

ducts, and bile leakage to the portal tract. These features

trigger a multistep process that ultimately leads to the

formation of periportal biliary fibrosis [5]. However, neither

the extent nor the dynamics of fibrogenesis were investigated

in these mice in detail. Here we present an analysis of the

evolution of fibrosis and the temporal expression of fibrogenic

and fibrolytic genes in Mdr2-/- mice. Our results show that

these mice may also serve as a valuable in vivo tool to test

potential antifibrotic agents.

2. Materials and methods

2.1. Animal studies

Mdr2-/- knockout and Mdr2C/C wildtype mice were obtained fromJackson Laboratory (Jackson Laboratory, Bar Harbor, ME) and housed witha 12-h light-dark cycle with water and standard mouse pellet chow adlibitum. At age 2-, 4-, and 8-weeks mice were sacrificed by cervicaldislocation under general anesthesia (400 mg avertin/kg i/p), livers andspleens were excised and weighed. Liver specimens from two lobes wereeither fixed in 4% buffered formalin or snap-frozen in liquid nitrogen forfurther analysis. Immunohistochemistry for a-SMA was performed asdescribed previously [5] on microwave treated (0.01 M citrate buffer pH6.0) paraffin sections (4 mm thick) using the monoclonal mouse antia-SMA (dilution 1:2500, Sigma, St Louis, MO). Ten highpower fields wereinvestigated for each animal. The experimental protocols were approved bythe local Animal Care and Use Committee according to criteria outlined inthe Guide for the Care and Use of Laboratory Animals.

2.2. Hepatic hydroxyproline determination

Hydroxyproline (HYP) was determined biochemically as describedbefore [6]. Briefly, two snap-frozen liver pieces from the left and right liverlobe (50–60 mg each) were hydrolyzed in 5 ml 6 N HCl at 110 8C for 16 h.Based on relative hepatic HYP (per 100 mg of wet liver), total hepatic HYPwas calculated (per total liver, as obtained by multiplying liver weights withrelative hepatic HYP).

2.3. Isolation of mRNA, reverse transcription

and real-time PCR

pieces of the right and left liver lobes (150–200 mg in total) werehomogenized in 1 ml RNApure solution (PeqLab, Erlangen, Germany), and50 ml aliquots were used for total RNA isolation according to themanufacturer’s recommendations. cDNA was obtained by reversetranscription of 1 mg of total RNA using Superscript II ReverseTranscriptase (Invitrogen, Karlsruhe, Germany), using 50 pmol randomhexamer and 100 pmol oligo-dT primers (Promega, Mannheim, Germany).Relative mRNA transcript levels were quantified using LightCyclerFastStart DNA Master Hybridization Probes kit on a LightCycler (allfrom Roche, Penzberg, Germany) applying the TaqMan methodology. Thehousekeeping gene beta2-microglobulin (b2MG) was amplified in aparallel reaction for normalization. TaqMan probes and primer sets weredesigned using the Primer Express software (Perkin Elmer, Foster City,CA) based on published sequences as summarized in Table 1. All TaqManprobes are positioned on exon–exon boundaries of corresponding genes toexclude co-amplification of genomic DNA. Sense and antisense primer(each at 0.5 mM) and 0.125 mM 5 0-phoshorylated probe, labeled at its 5 0-endwith the reporter dye 6-carboxyfluorescein (6-FAMe) and at the 3 0end withthe quencher dye 6-carboxy-tetramethyl-rhodamine (TAMRA), weresynthesized at MWG Biotech (Ebersberg, Germany). Quantification ofTGFb2, PAI-1 and PDGFR-b transcripts was done using the LightCyclerFastStart DNA Master SYBR Green I kit (Roche, Penzberg, Germany) andthe primers outlined in Table 1. Great efforts were made to ensure thespecificity of the products amplified, including melting curve analysis aftereach amplification and visualization of PCR-products on agarose gel.

2.4. Determination of MMP-activities

Tissue extraction. 50–100 mg of snap frozen liver was homogenized incold 50 mM Tris–HCl, 150 mM NaCl, 5 mM CaCl2, 0,025% Brij 35e, pH7.5, supplemented with EDTA-free protease inhibitor cocktail (Com-pletee, Roche Applied Science, Mannheim, Germany). Supernatants werecollected by centrifugation and assayed immediately, or stored in aliquotsat K80 8C until further processing. Measurements were performed induplicates and in three individual animals per time point. Results werenormalized to total protein determined with the Bradford reagent (BioRad,Munich, Germany).

MMP-2 activity was measured using the MMP-2 Biotrack ActivityAssay (Amersham, Buckinghamshire, UK) which is based on capture ofMMP-2 by immobilized anti-MMP-2 antibody, followed by incubationwith modified urokinase that is specifically cleaved by captured MMP-2.Human recombinant MMP-2 provided with the kit was used as a referenceand for assay calibration. Since assays were performed with proteaseinhibitors that prevent activation of pro-MMPs but do not block MMP-activity, results represent endogenously active MMP-2 levels but not pro-MMP-2 or the MMP-2/TIMP-2 complex.

Determination of interstitial collagenase and gelatinase activities wasperformed with assays which are based on degradation of biotinylatednative collagen type I and biotinylated gelatin, respectively (Chemicon,Temecula, CA). Ten microlitre of liver extracts were incubated with thesubstrates in 96 well plates for 2 h and solubilized biotinylated substratefragments transferred into biotin-binding plate, followed by addition ofstreptavidin-peroxidase, peroxidase substrate (diaminobenzidine) anddetection at 405 nm using a microplate reader according to themanufacturer’s recommendations. Human recombinant activated MMP-1and MMP-2 were used as standards. Since extraction procedures wereperformed in the presence of protease inhibitors, data obtained represent theendogenous collagenolytic and gelatinolytic activities.

Statistical analysis. Statistical analyses were performed using MicrosoftEXCEL software. Data are expressed as meansGSEM. The statisticalsignificance of differences was evaluated using the unpaired, non-parametric Student’s t-test.

3. Results

Mdr2-/- mice and their wildtype littermates were

sacrificed at 2, 4 and 8 weeks of age, previously established

Table 1

Primers and probes used in quantitative RT-PCR

Target gene 5 0-Primer TaqMan probe 3 0-Primer

b2MG CTGATACATACGCCTGCAGAGTTAA GACCGTCTACTGGGATCGAGA-

CATGTG

ATGAATCTTCAGAGCATCATGAT

Procollagen

a1(I)

TCCGGCTCCTGCTCCTCTTA TTCTTGGCCATGCGTCAGGAGGG GTATGCAGCTGACTTCAGG-

GATGT

TGFb1 AGAGGTCACCCGCGTGCTAA ACCGCAACAACGCCATCTATGA-

GAAAACCA

TCCCGAATGTCTGACGTATTGA

MMP-2 CCGAGGACTATGACCGGGATAA TCTGCCCCGAGACCGCTATGTCCA CTTGTTGCCCAGGAAAGTGAAG

MMP-3 GATGAACGATGGACAGAGGATG TGGTACCAACC-

TATTCCTGGTTGCTGC

AGGGAGTGGCCAAGTTCATG

MMP-13 GGAAGACCCTCTTCTTCTCT TCTGGTTAACATCATCATAACTC-

CACACGT

TCATAGACAGCATCTACTTTGTT

TIMP-1 TCCTCTTGTTGCTATCACTGA-

TAGCTT

TTCTGCAACTCGGACCTGGTCA-

TAAGG

CGCTGGTATAAGGTGGTCTCGTT

TIMP-2 CCAGAAGAAGAGCCTGAACCA ACTCGCTGTCCCATGATCCCTTGC GTCCATCCAGAGGCACTCATC

a-SMA ACAGCCCTCGCACCCA CAAGATCATTGCCCCTCCA-

GAACGC

GCCACCGATCCAGACAGAGT

TGFb2 TCGTCCGCTTTGATGTCTCA – AAATCTCGCCTCGAGCTCTTC

PAI-1 TGGCTCAGAGCAACAAGTTCA – TTTGCAGTGCCTGTGCTACAG

PDGFR-B TCCCACATTCCTTGCCCTT – TCGCTACTTCTGGCTGTCGAT

Y. Popov et al. / Journal of Hepatology 43 (2005) 1045–1054 1047

to reflect important time points for the pathogenesis of liver

injury [5].

3.1. Mdr2-/- mice spontaneously develop progressive

biliary fibrosis with a characteristic biphasic pattern

of myofibroblast activation

Mdr2-deficient mice developed significant hepatome-

galy already at 2 weeks of age (liver weight 0.26G0.02

vs. 0.15G0.01), which further progressed until week 8,

the end point of our study (1.56G0.01 vs. 0.82G0.03 at

4w, 1.89G0.09 vs. 1.15G0.17 at 8w). Upon histological

examination the livers of Mdr2-/- mice displayed age-

dependent enlargement of bile ducts, periductular inflam-

mation and a broad rim of periductular extracellular

matrix as described [5]. At 2 weeks strong a-smooth

muscle actin (SMA) expression was associated with

vascular smooth muscle cells both in Mdr2-/- and their

wildtype controls. However, a highly increased number of

periductular a-SMA positive myofibroblastic cells were

found at week 4 in Mdr2-/-, while their number decreased

again at week 8 (Fig. 1(A) and (B)). Levels of hepatic

a-SMA mRNA, as quantified by real time PCR, followed

the similar pattern—there was no difference at age 2w,

twofold increase at 4w and decline at late time-point in

Mdr2-/- mice to non-significantly elevated level as

compared to their WT-controls (Fig. 1(C)).

As quantified by hepatic hydroxyproline (HYP), Mdr2-

/- mice demonstrated a steady increase of relative hepatic

collagen accumulation (mcg per 100 mg of tissue),

reaching significance at 4 weeks and being 2.8-fold

above that of their wildtype controls at 8 weeks

(Fig. 2(A)). Liver collagen deposition in Mdr2-/- mice

was significantly elevated at all time points when

expressed as total liver HYP, with a 1.8-, 2.8- and 5.1-

fold increase at weeks 2, 4 and 8 (Fig. 2(B)), comparable

to the progressive collagen accumulation observed in

rodent secondary biliary fibrosis subsequent to bile duct

obstruction [6].

3.2. Temporal expression patterns of profibrogenic genes

in livers of Mdr2-/- mice

To obtain further information about the key matrix

molecules involved in the rapid fibrosis progression of

Mdr2-/- mice, a spectrum of profibrogenic as well as

putative fibrolytic mRNA transcripts and the time course

of their expression were quantified by quantitative real

time PCR. Although all of the investigated target genes

were upregulated in Mdr2-/- mice, distinct time-depen-

dent patterns emerged. The central profibrogenic cytokine

TGFb1, its isoform TGFb2 which is mainly expressed by

proliferating bile duct epithelia [7], and the receptor for

PDGF-BB (PDGFR-b) which is considered as an

activation marker of hepatic stellate cells and myofibro-

blasts demonstrated similar expression profiles: they were

increased at week 2, reached a maximum at week 4, and

dropped to levels comparable to those at week 2 at week

8 after birth (Fig. 3(A)–(C)). Interestingly, the peak

levels of these transcripts coincided with the peak of

hepatic bile duct-associated a-SMA expression (Fig. 1).

In contrast to this pattern but in line with the steady

progression of fibrosis until week 8, transcript levels of

procollagen a1(I), which represents the major extracellu-

lar matrix protein in fibrosis, steadily increased from 3.5-

to 12.9-fold compared to the wild-type controls

(Fig. 3(D)).

Fig. 1. Sirius Red staining of the livers of Mdr2-/- mice shows a progressive accumulation of peribiliar extracellular matrix around enlarged bile ducts

(A) compared to their wildtype controls (B). Sequential sections demonstrate a peak of peribiliary a-SMA immunoreactivity in portal fibrotic lesions

at 4 weeks (A), while wildtype mice demonstrate a-SMA expression restricted to smooth muscle cells around blood vessels (a: hepatic artery; b: bile

duct; v: portal vein). Magnification !20. (C). Relative hepatic a-SMA mRNA determined by real-time RT-PCR in Mdr2-/- mice (closed columns) and

their wildtype controls (open columns) at weeks 2, 4 and 8 of age, expressed as meansGSEM (nZ4 per group) in arbitrary units relative to b2-

microglobulin mRNA. Data are presented as an x-fold increase vs. the corresponding wildtype controls. *P!0.05 compared to Mdr2C/C controls of

the corresponding age. [This figure appears in colour on the web.]

Fig. 2. Mdr2-/- mice demonstrate progressive hepatic collagen accumulation. (A): hepatic collagen deposition expressed as relative (mg per 100 mg liver,

derived from the right and left lobes) HYP content at 2, 4 and 8 weeks age in Mdr2-/- (black columns) compared to wildtype mice (white columns). (B):

total hepatic HYP expressed as mg per liver, as calculated by multiplication of individual liver weight with relative HYP content. Results are expressed as

meansGSEM (nZ4 each bar). *P!0.05 compared to Mdr2C/C mice of the corresponding age group.

Y. Popov et al. / Journal of Hepatology 43 (2005) 1045–10541048

Fig. 3. Profibrogenic gene expression in Mdr2-/- mice. Relative hepatic mRNA transcript levels of procollagen a1(I) (A), platelet-derived growth factor

receptor PDGFR-b (B), and transforming growth factor-b (TGFb) isoforms 1 (C) and 2 (D) were determined by realtime RT-PCR in Mdr2-/- mice

(black columns) and their wildtype controls (white columns) at weeks 2, 4 and 8 of age. Results are expressed as meansGSEM (nZ4 per group), and in

arbitrary units relative to b2-microglobulin mRNA. Data are presented as an x-fold increase vs. the corresponding wildtype controls. *P!0.05

compared to Mdr2C/C controls of the corresponding age.

Y. Popov et al. / Journal of Hepatology 43 (2005) 1045–1054 1049

3.3. Temporal expression patterns of fibrolytic MMPs

and their inhibitors in livers of Mdr2-/- mice

We next assessed the expression patterns of main

components of fibrolysis, i.e. putatively fibrolytic MMPs and

their major inhibitors at the transcript level. When compared to

wildtype controls, MMP-13 mRNA, encoding the major

interstitial collagenase in rodents, was maximally increased at

week 2 (fivefold), declining to twofold and rising to fourfold at

weeks 4 and 8, respectively (Fig. 4(A)). MMP-3 mRNA

followed the pattern of MMP-13 expression, although

differences did not reach statistical significance due to a high

interindividual variability in Mdr2-/- mice (Fig. 4(B)). TIMP-

1 mRNA which encodes the major physiological inhibitor of

most MMPs was upregulated about 20-fold at all time points

(Fig. 4(C)), and transcripts of PAI-1, a potent inhibitor of the

proteolytic activation of pro-MMPs by plasmin, was up-

regulated 27-fold only at 2 weeks, rapidly dropping to almost

normal levels at 4 and 8 weeks (Fig. 4(D)).

To determine if and how far hepatic collagen-degrading

activity in Mdr2-/- mice is altered compared to their

wildtype controls, naturally occurring interstitial collage-

nolytic activity in liver homogenates was measured based

on degradation of biotinylated native collagen type I.

Elevated interstitial collagenase activity in liver was

suppressed fivefold at 2 weeks of age in Mdr2-/- mice,

and reached comparably low levels at weeks 4 and 8 in both

Mdr2-/- and wildtype mice (Fig. 4(E)).

3.4. MMP-2 demonstrate a regulation in Mdr2-/- mice

distinct from other MMPs

MMP-2 (gelatinase A) mRNA rose from twofold at week

2 to 8- fold at week 4 (Fig. 2), following the expression

pattern of the TGFb and the HSC activation markers

(Fig. 3). Hepatic mRNA expression of the physiological

inhibitor of MMP-2, TIMP-2 was increased by 70% at week

8 in Mdr2-/- vs. control mice, but comparable to the controls

at weeks 2 and 4 (Fig. 5(C)).

To our surprise and despite of high up-regulation of pro-

MMP-2 mRNA, endogenously active MMP-2 at the protein

level was even down-regulated at the peak of its mRNA

expression (4w) in liver extracts of Mdr2-/- and not altered

at week 2 and 8 as compared to WT-littermates, as

determined by ELISA-based activity assay (Fig. 5(C)).

Furthermore, intrinsic gelatinolytic activity was down-

regulated by 50% at week 4 in Mdr2-/- mice, with no

significant difference observed vs. their wildtype controls at

weeks 2 and 8 (Fig. 5(D)).

4. Discussion

Mdr2/Abcb4 (mouse orthologue of human MDR3/

ABCB4) knockout mice (Mdr2-/-) were recently shown to

develop biliary fibrosis shortly after birth [3] as a result of

leakage of potentially toxic bile acids (e.g. cholic acid) into

Fig. 4. Expression of fibrolytic MMPs, their inhibitors and interstitial collagenase activity in Mdr2-/- mice. Relative hepatic transcript levels of MMP-

13 (A) and MMP-3 (B), tissue inhibitor of metalloproteinase (TIMP-1) (C), and plasminogen activator inhibitor (PAI-1) (D) were determined by real-

time RT-PCR at 2, 4 and 8 weeks of age in Mdr2-/- mice (black columns) and their wildtype littermates (white columns). Results are expressed in

arbitrary units relative to b2-microglobulin mRNA and as an x-fold increase vs. the corresponding wildtype controls (meansGSEM, nZ4 per group).

(E) Interstitial collagenase activity was assessed from liver extracts by degradation of biotinylated native type I collagen. Data represent meansGSEM

(nZ3 each column) assayed in duplicates and expressed as ng of human recombinant MMP-1 used as reference. *P!0.05 compared to Mdr2C/C

controls of the corresponding age.

Y. Popov et al. / Journal of Hepatology 43 (2005) 1045–10541050

the periductal area [5]. The model resembles biliary fibrosis

in humans, such as PSC or congenital MDR3 deficiency

ranging from progressive familial intrahepaic cholestasis

type 3 to adult liver cirrhosis [4], and should therefore be

well suited to better understand the pathogenesis of liver

fibrosis in these diseases. We therefore, characterized the

extent and the temporal expression patterns of fibrosis

related genes (Fig. 6(A)) and of matrix degrading activities

in Mdr2-/- compared to wildtype mice.

Activated hepatic stellate cells (HSC) and perivascular/

periportal/peribiliary myofibroblasts (MF) drive hepatic

fibrosis [8–10]. In Mdr2-/-, we observed a massive increase

of activated MF around bile ducts at 4 weeks, as assessed by

a-SMA immunohistochemistry, a surrogate marker of their

activation [11]. The number of activated (a-SMA positive)

MFs decreased with the formation of mature dense septa at

week 8. This was paralleled by a prominent upregulation of

PDGFR-b, TGFb1, and TGFb2 transcripts at 4 weeks

(Fig. 6(B)), all associated with fibrogenic activation of HSC/

MF [12]. Therefore, the time point around week 4 likely

represents the most active phase of fibrogenesis, which is at

least in part driven by TGFb production. At 8 weeks, when

TGFb expression and the HSC/MF activation marker

a-SMA declined, procollagen a1(I) expression still

increased and hepatic collagen accumulation was progress-

ive. In this line, a recent study using transgenic mice with a

dual reporter transgene for a-SMA and procollagen I has

shown that co-expression of a-SMA is observed in only

30% of HSC/MF isolated from fibrotic liver [15]. Thus

upregulation of a-SMA as (classical) activation marker

might only be required at some, possibly early stages of

fibrogenesis in vivo. A phenomenological explanation

would be that once the periductal fibrous ring has been

formed there is no longer need for contractile elements as

represented by a-SMA. As regards TGFb, fully activated vs.

early activated HSC/MF were shown to have a decreased

responsiveness to this cytokine [13,14]. This possibly

underlies the observed downregulation of TGFb at 8

weeks due to a lower autoinduction, while a high

constitutive procollagen expression is maintained.

Fig. 5. Expression of MMP-2, TIMP-2, and gelatinolytic activity in Mdr2-/- mice. Transcript levels in Mdr2-/- mice (black columns) and their wildtype

controls (white columns) were determined by realtime RT-PCR. Results are expressed as meansGSEM (nZ4 per group), and in arbitrary units

relative to b2-microglobulin mRNA (x-fold increase vs. the corresponding wildtype controls). Intrinsically active MMP-2 protein was determined by

Biotrack activity assay (C) and gelatinolytic activity was assessed by degradation of biotinylated gelatin (D). Data represent meansGSEM (nZ3 each

column) assayed in duplicates and expressed as ng of human recombinant MMP-2 used as a reference. *P!0.05 compared to Mdr2C/C controls of

the corresponding age.

Y. Popov et al. / Journal of Hepatology 43 (2005) 1045–1054 1051

Expression and activities of key MMPs that are

instrumental in hepatic collagen turnover showed character-

istic, but also complex patterns in Mdr2-/- mice. As described

before for fibroblasts [16] and HSC [17], both MMP-3 and -

13 mRNAs are regulated in an opposite manner compared to

the profibrogenic transcripts. Thus mRNA expression of

MMP-2, a protease that is considered pro-fibrogenic due to

its ability to degrade basement membrane collagen [18], a

structure that can induce quiescence and inhibit migration of

HSC/MF [19], was upregulated in the Mdr2-/- mice at week

4, paralleling their peak profibrogenic activation. The

observed MMP-expression patterns can be explained by

the effects of TGFb which upregulates MMP-2 [20] and

downregulates MMP-3 and -13 mRNAs in HSC ([17] and

own unpublished observations). However, active MMP-2

and gelatinase activities in the livers of Mdr2-/- mice were

minimal at this time point. An excess of endogenous

inhibition by TIMP-2 is unlikely, since TIMP-2 mRNA

was upregulated only at week 8. Activation of pro-MMP-2 is

complex and not well understood. It occurs on the cell

surface, requires complex formation with TIMP-2 and

membrane-anchored MMP-14 [21], and the interaction of

HSC/MF with hepatocytes [22]. In contrast to the in vitro

studies that suggested MMP-2 as profibrogenic factor acting

via promoting HSC migration and proliferation [23,24], our

observation that MMP-2 and gelatinolytic activities in

Mdr2-/- mice were downregulated at the peak of HSC

activation at 4 weeks despite its maximal upregulation at the

mRNA level suggests that active MMP-2 generation in vivo

is also governed by local factors such as adhesion to ECM

and its compartmentalization in certain membrane domains

[25].

MMPs are subject to multiple levels of regulation [18].

Apart from their transcriptional control, MMPs are secreted

as proenzymes and must undergo activation via proteolytic

cleavage. Furthermore, MMP activity is inhibited by their

specific tissue inhibitors (TIMP-2 for membrane-type

MMP-14 and MMP-2 [26,27], and TIMP-1 for virtually

all other MMPs ([28]). Similar to the observed divergence

between MMP-2 mRNA expression and gelatinase activity,

hepatic interstitial collagenase activity did not parallel the

MMP-13 mRNA expression pattern, pointing to activity

regulation at the posttranscriptional level. Accordingly, the

temporal patterns of PAI-1 and TIMP-1, two protease

inhibitors that are centrally involved in the cascade that

leads to MMP activation via urokinase plasminogen

activator (PAI-1) or to irreversible inhibition of most active

MMPs themselves (TIMP-1) [18] appear to better reflect

MMP activities in Mdr2-/- mice. TIMP-1 mRNA was

upregulated 20-fold in Mdr2-/- vs. wildtype mice through-

out the observation period, while PAI-1 mRNA peaked at 2

weeks (27-fold), a time point when fibrosis was still

minimal. Since interstitial collagenase activity was mark-

edly reduced at 2 weeks, coinciding with the maximal PAI-1

expression, inhibition of plasmin by PAI-1 appears to be

instrumental for the reduced MMP-13 activity at this early

stage. The massive and continuous hepatic overexpression

of TIMP-1 is most likely a central determinant of fibrosis

progression in Mdr2-/- mice, as found in other studies

[29,30]. TIMP-1 is an almost universal MMP inhibitor [28]

Fig. 6. (A) Summary of measured profibrogenic and profibrolytic genes and their functional role in progession of fibrosis. (B) Scheme of ECM-related

gene expression profiles during liver fibrosis progression in MDR2-/- mice at weeks 2, 4 and 8 of age, fold to MDR2C/C controls (on the left). HSC,

hepatic stellate cell; MF, myofibroblast.

Y. Popov et al. / Journal of Hepatology 43 (2005) 1045–10541052

that also has an antiapoptotic and proproliferative effect on

activated HSC/MF [31]. TIMP-1-trangenic mice show

accelerated fibrogenesis in the carbon tetrachloride model

of liver fibrosis [32] and do not reverse as well as their

wildtype controls after cessation of injury [33].

The most critical features of good animal models to test

potential antifibrotics is significant net collagen accumu-

lation that is accompanied by only minor interindividual

variability of this parameter, major weaknesses of existing

rat or mouse models of liver fibrosis [34]. So far,

biochemical determination of hepatic HYP content remains

the ‘gold standard’ to determine collagen accumulation and

thus the exact stage of fibrosis, allowing to detect even

minor antifibrotic drug effects in animal models. This is

more difficult in humans, since histological and biochemical

assessments are prone to sampling error and histological

staging is subject to investigator’s bias [35–37]. In Mdr2-/-

mice, quantification of relative and total collagen as HYP

content demonstrated 2,8 and fivefold increases at 8 weeks

of age, respectively, with low standard deviations. This

degree of liver collagen deposition is comparable to that of

thioacetamide-induced fibrosis after 12 weeks [38], and own

Y. Popov et al. / Journal of Hepatology 43 (2005) 1045–1054 1053

unpublished data) or biliary fibrosis secondary to bile duct

ligation after 4 weeks [30], both accepted models of

advanced liver fibrosis/cirrhosis. Furthermore, fibrosis in

Mdr2-/- mice bears more resemblance to human biliary

fibrosis than any other rat or mouse model described so far.

In addition, due to larger individual differences in fibrosis

evolution these models require a higher number of animals

per group (10 or more), for obtaining reliable results. In

contrast, only 4-5 Mdr2-/- mice are needed at weeks 4 or 8

for this purpose. These are important considerations in a

number of issues related to antifibrotic drug development—

such as facilitating testing of many drugs and drug

combinations in mice instead of rats which will speed up

antifibrotic drug screening and lower costs. Moreover, the

present model should facilitate longitudinal studies demon-

strating the durability and impact of potential antifibrotic

effects. Importantly, it appears to be possible to identify

antifibrotic drug candidates already at week 2 when

surrogate markers of fibrogenesis, such as PAI-1, procolla-

gen a1(I), and TIMP-1 mRNA are already highly increased.

Such studies will have to determine whether ‘pure’

antifibrotic approaches halt or reverse biliary fibrosis even

when the primary trigger (i.e. leakage of potentially toxic

bile acid into the periductal area) is not neutralized.

In conclusion, we performed a detailed analysis of

spontaneous progressive liver fibrosis in Mdr2-/- mice,

occurring via massive up-regulation of profibrogenic

mRNAs and downregulation of collagenolytic activity.

These features coupled with a similarity to human liver

disease (PSC and MDR3 deficiency) and a highly reproducible

fibrosis progression qualify these mice as a promising in vivo

model both to study the mechanisms of biliary fibrosis and for

the testing of potential antifibrotic agents.

Acknowledgements

This work was supported by the German research

Council (DFG) grant 646/14-1, and grants by the German

Network for Viral Hepatitis (Hepnet) and the Interdisci-

plinary Center for Clinical Research (IZKF) of the

University of Erlangen-Nuernberg to D.S. and grant

P-15502 from the Austrian Science Foundation and a

GEN-AU project grant from the Austrian Ministry for

Science to M.T. Y.P. was a recipient of EASL/Yamanouchi(2002) and EASL/Sheila Sherlock (2003) Fellowships, and

E.P. of a DFG-Graduate College (GRK 750) Scholarship.

The excellent technical assistance of Edith Niedobitek and

Andrea Fuchsbichler is gratefully acknowledged.

References

[1] Sedlaczek N, Jia JD, Bauer M, Herbst H, Ruehl M, Hahn EG, et al.

Proliferating bile duct epithelial cells are a major source of connective

tissue growth factor in rat biliary fibrosis. Am J Pathol 2001;158:

1239–1244.

[2] Smit JJ, Schinkel AH, Oude Elferink RP, Groen AK, Wagenaar E, van

Deemter L, et al. Homozygous disruption of the murine mdr2

P-glycoprotein gene leads to a complete absence of phospholipid from

bile and to liver disease. Cell 1993;75:451–462.

[3] Fickert P, Zollner G, Fuchsbichler A, Stumptner C, Weiglein AH,

Lammert F, et al. Ursodeoxycholic acid aggravates bile infarcts in bile

duct-ligated and Mdr2 knockout mice via disruption of cholangioles.

Gastroenterology 2002;123:1238–1251.

[4] Jacquemin E. Role of multidrug resistance 3 deficiency in pediatric

and adult liver disease: one gene for three diseases. Semin Liver Dis

2001;21:551–562.

[5] Fickert P, Fuchsbichler A, Wagner M, Zollner G, Kaser A, Tilg H,

et al. Regurgitation of bile acids from leaky bile ducts causes

sclerosing cholangitis in Mdr2 (Abcb4) knockout mice. Gastroenter-

ology 2004;127:261–274.

[6] Cho JJ, Hocher B, Herbst H, Jia JD, Ruehl M, Hahn EG, et al. An oral

endothelin-A receptor antagonist blocks collagen synthesis and

deposition in advanced rat liver fibrosis. Gastroenterology 2000;

118:1169–1178.

[7] Milani S, Herbst H, Schuppan D, Stein H, Surrenti C. Transforming

growth factors beta 1 and beta 2 are differentially expressed in fibrotic

liver disease. Am J Pathol 1991;139:1221–1229.

[8] Reeves HL, Friedman SL. Activation of hepatic stellate cells—a key

issue in liver fibrosis. Front Biosci 2002;7:d808–d826.

[9] Knittel T, Kobold D, Piscaglia F, Saile B, Neubauer K, Mehde M,

et al. Localization of liver myofibroblasts and hepatic stellate cells in

normal and diseased rat livers: distinct roles of (myo-)fibroblast

subpopulations in hepatic tissue repair. Histochem Cell Biol 1999;

112:387–401.

[10] Cassiman D, Libbrecht L, Desmet V, Denef C, Roskams T. Hepatic

stellate cell/myofibroblast subpopulations in fibrotic human and rat

livers. J Hepatol 2002;36:200–209.

[11] Geerts A. History, heterogeneity, developmental biology, and

functions of quiescent hepatic stellate cells. Semin Liver Dis 2001;

21:311–335.

[12] Pinzani M, Gentilini A, Caligiuri A, De Franco R, Pellegrini G,

Milani S, et al. Transforming growth factor-beta 1 regulates platelet-

derived growth factor receptor beta subunit in human liver fat-storing

cells. Hepatology 1995;21:232–239.

[13] Dooley S, Delvoux B, Lahme B, Mangasser-Stephan K, Gressner AM.

Modulation of transforming growth factor beta response and signaling

during transdifferentiation of rat hepatic stellate cells to myofibro-

blasts. Hepatology 2000;31:1094–1106.

[14] Dooley S, Delvoux B, Streckert M, Bonzel L, Stopa M, ten Dijke P,

et al. Transforming growth factor beta signal transduction in hepatic

stellate cells via Smad2/3 phosphorylation, a pathway that is

abrogated during in vitro progression to myofibroblasts. TGFbeta

signal transduction during transdifferentiation of hepatic stellate cells.

FEBS Lett 2001;502:4–10.

[15] Magness ST, Bataller R, Yang L, Brenner DA. A dual reporter gene

transgenic mouse demonstrates heterogeneity in hepatic fibrogenic

cell populations. Hepatology 2004;40:1151–1159.

[16] Lee KS, Ryoo YW, Song JY. Interferon-gamma upregulates the

stromelysin-1 gene expression by human skin fibroblasts in culture.

Exp Mol Med 1998;30:59–64.

[17] Schaefer B, Rivas-Estilla AM, Meraz-Cruz N, Reyes-Romero MA,

Hernandez-Nazara ZH, Dominguez-Rosales JA, et al. Reciprocal

modulation of matrix metalloproteinase-13 and type I collagen genes

in rat hepatic stellate cells. Am J Pathol 2003;162:1771–1780.

[18] Benyon RC, Arthur MJ. Extracellular matrix degradation and the role

of hepatic stellate cells. Semin Liver Dis 2001;21:373–384.

[19] Friedman SL, Roll FJ, Boyles J, Arenson DM, Bissell DM.

Maintenance of differentiated phenotype of cultured rat hepatic

lipocytes by basement membrane matrix. J Biol Chem 1989;264:

10756–10762.

Y. Popov et al. / Journal of Hepatology 43 (2005) 1045–10541054

[20] Herbst H, Wege T, Milani S, Pellegrini G, Orzechowski HD,

Bechstein WO, et al. Tissue inhibitor of metalloproteinase-1 and -2

RNA expression in rat and human liver fibrosis. Am J Pathol 1997;

150:1647–1659.

[21] Strongin AY, Collier I, Bannikov G, Marmer BL, Grant GA,

Goldberg GI. Mechanism of cell surface activation of 72-kDa type

IV collagenase. Isolation of the activated form of the membrane

metalloprotease. J Biol Chem 1995;270:5331–5338.

[22] Theret N, Musso O, L’Helgoualc’h A, Clement B. Activation of

matrix metalloproteinase-2 from hepatic stellate cells requires

interactions with hepatocytes. Am J Pathol 1997;150:51–58.

[23] Olaso E, Ikeda K, Eng FJ, Xu L, Wang LH, Lin HC, et al. DDR2

receptor promotes MMP-2-mediated proliferation and invasion by

hepatic stellate cells. J Clin Invest 2001;108:1369–1378.

[24] Yang C, Zeisberg M, Mosterman B, Sudhakar A, Yerramalla U,

Holthaus K, et al. Liver fibrosis: insights into migration of hepatic

stellate cells in response to extracellular matrix and growth factors.

Gastroenterology 2003;124:147–159.

[25] Yan L, Moses MA, Huang S, Ingber DE. Adhesion-dependent control

of matrix metalloproteinase-2 activation in human capillary endo-

thelial cells. J Cell Sci 2000;113:3979–3987.

[26] Itoh Y, Ito A, Iwata K, Tanzawa K, Mori Y, Nagase H. Plasma

membrane-bound tissue inhibitor of metalloproteinases (TIMP)-2

specifically inhibits matrix metalloproteinase 2 (gelatinase A)

activated on the cell surface. J Biol Chem 1998;273:24360–24367.

[27] Butler GS, Butler MJ, Atkinson SJ, Will H, Tamura T, van

Westrum SS, et al. The TIMP2 membrane type 1 metalloproteinase

‘receptor’ regulates the concentration and efficient activation of

progelatinase A. A kinetic study. J Biol Chem 1998;273:871–880.

[28] Brew K, Dinakarpandian D, Nagase H. Tissue inhibitors of

metalloproteinases: evolution, structure and function. Biochim

Biophys Acta 2000;1477:267–283.

[29] McCrudden R, Iredale JP. Liver fibrosis, the hepatic stellate cell and

tissue inhibitors of metalloproteinases. Histol Histopathol 2000;15:

1159–1168.

[30] Jia JD, Bauer M, Cho JJ, Ruehl M, Milani S, Boigk G, et al.

Antifibrotic effect of silymarin in rat secondary biliary fibrosis is

mediated by downregulation of procollagen alpha1(I) and TIMP-1.

J Hepatol 2001;35:392–398.

[31] Murphy FR, Issa R, Zhou X, Ratnarajah S, Nagase H, Arthur MJ, et al.

Inhibition of apoptosis of activated hepatic stellate cells by tissue

inhibitor of metalloproteinase-1 is mediated via effects on matrix

metalloproteinase inhibition: implications for reversibility of liver

fibrosis. J Biol Chem 2002;277:11069–11076.

[32] Yoshiji H, Kuriyama S, Miyamoto Y, Thorgeirsson UP, Gomez DE,

Kawata M, et al. Tissue inhibitor of metalloproteinases-1 promotes

liver fibrosis development in a transgenic mouse model. Hepatology

2000;32:1248–1254.

[33] Yoshiji H, Kuriyama S, Yoshii J, Ikenaka Y, Noguchi R, Nakatani T,

et al. Tissue inhibitor of metalloproteinases-1 attenuates spontaneous

liver fibrosis resolution in the transgenic mouse. Hepatology 2002;36:

850–860.

[34] Schuppan D, Popov Y. Hepatic fibrosis: from bench to bedside.

J Gastroenterol Hepatol 2002;17:S300–S305.

[35] Rosenberg WM, Voelker M, Thiel R, Becka M, Burt A, Schuppan D,

et al. Serum markers detect the presence of liver fibrosis: a cohort

study. Gastroenterology 2004;127:1704–1713.

[36] Bedossa P, Dargere D, Paradis V. Sampling variability of liver fibrosis

in chronic hepatitis C. Hepatology 2003;38:1449–1457.

[37] Regev A, Schiff ER. Drug therapy for hepatitis B. Adv Intern Med

2001;46:107–135.

[38] Bruck R, Genina O, Aeed H, Alexiev R, Nagler A, Avni Y, et al.

Halofuginone to prevent and treat thioacetamide-induced liver fibrosis

in rats. Hepatology 2001;33:379–386.