Potential conflict of interest: Nothing to report.
Cholestasis occurs in a variety of clinical settings and often results in liver injury and secondary biliary fibrosis. Several matrix metalloproteinases (MMPs) are upregulated in the liver during cholestasis. The function of the major interstitial collagenase, MMP-13, in the initial phase of liver fibrosis is unknown. The aim of this study was to evaluate the role of MMP-13 during the development of cholestasis-induced liver fibrosis by comparing wild-type and MMP-13-deficient mice. Cholestasis was induced by bile duct ligation (BDL) for 5 days or 3 weeks. Activation and proliferation of hepatic stellate cells (HSCs) were detected by immunohistochemistry. Expression of MMP-13 mRNA increased significantly in BDL livers of WT mice. After BDL for 3 weeks liver fibrosis was suppressed in MMP-13-deficient mice versus WT animals. Activation and proliferation of HSCs were also suppressed in livers of MMP-13-deficient mice after BDL. To clarify the mechanism of this suppression, samples from 5-day BDL mice were used for evaluation of liver injury. Compared with those in WT animals, serum ALT and the number of hepatic neutrophils were reduced in MMP-13-deficient mice. Increased expression of the mRNA of inflammatory mediators such as tumor necrosis factor-alpha (TNF-α) was significantly suppressed in livers of MMP-13-deficient mice. Upregulation of fibrogenic markers, for example, transforming growth factor beta1 (TGF-β1), was also significantly suppressed in livers of MMP-13-deficient mice versus in WT mice. In conclusion, distinct from the known function of interstitial collagenase to reduce liver fibrosis by degrading the extracellular matrix, MMP-13 contributes to accelerating fibrogenesis in cholestatic livers by mediating the initial inflammation of the liver. (HEPATOLOGY 2006;44:420–429.)
Liver fibrosis represents the final common pathway for most chronic liver diseases.1 During hepatic fibrogenesis, extensive remodeling of the extracellular matrix (ECM) occurs, and matrix composition changes from the normal basement membranelike matrix to fibrillar collagen.2 Type I collagen is the most prevalent ECM protein deposited in cirrhosis.3 Activated hepatic stellate cells (HSCs) are the major source of ECM in the liver.4 In response to liver injury, the normally quiescent HSCs undergo progressive transdifferentiation into proliferating myofibroblast-like activated HSCs.4, 5 Activated HSCs also are able to degrade ECM, as they express matrix metalloproteinases (MMPs).6
MMPs are a family of extracellular zinc- and calcium-dependent proteases that degrade the ECM and other extracellular proteins.7 MMPs are classified structurally according to substrate specificity. The activity of MMPs is tightly regulated at several levels including gene transcription, proenzyme activation, and inhibition of activated enzymes by the tissue inhibitor of matrix metalloproteinases (TIMPs).8, 9 MMPs are essential for normal remodeling of the extracellular matrix, tissue morphogenesis, and wound healing.10 When dysregulated, they participate in various pathological processes such as pulmonary emphysema,11 atherosclerotic plaque remodeling,12 and acute hepatitis.13 Although expression of several MMPs increases during hepatic fibrogenesis, their activity is inhibited by high levels of TIMPs.14–16 The importance of MMPs during the resolution of liver fibrosis is well documented,17, 18 but the exact role of MMPs during the progression of liver fibrosis is yet to be explored.
MMP-13, the interstitial collagenase of rodents, is a highly specific protease capable of degrading insoluble fibrillar collagens, especially type I collagen, suggesting MMP-13 could play an important role in liver fibrogenesis. Although previous studies focused on MMP-13 expression during liver injury, alteration in expression of MMP-13 and the cellular source of MMP-13 during liver fibrogenesis are still in debate.16, 19
This study analyzed the expression pattern of MMP-13 in the liver after bile duct ligation (BDL) in mice. Moreover, we used the BDL model to examine the role of MMP-13 during liver fibrosis in MMP-13-deficient mice and their wild-type littermates. Our results indicated loss of MMP-13 reduces cholestasis-induced liver inflammation and fibrogenesis.
The wild-type (MMP-13+/+) and knockout (MMP-13−/−) mice used in this study were of a mixed 129/C57BL6 genetic background. The mice appear phenotypically normal into adulthood (J. D'Armiento, manuscript in preparation), similar to a previously described MMP-13 knockout mouse.20 Mice were anesthetized with ketamine and xylazine. After midline laparotomy (1 cm), the common bile duct was exposed and ligated three times with 6-0 silk sutures. Two ligatures were placed in the proximal portion of the bile duct and one ligature was located in the distal portion of the bile duct. The bile duct was then cut between the ligatures. The abdomen was closed in layers, and the animals were allowed to recover on a heat pad. After 5 days or 3 weeks, animals were anesthetized to collect blood and liver samples. Animals received humane care according to the criteria outlined in “Guide for the Care and Use of Laboratory Animals.”
Isolation and Culture of HSCs.
For HSC isolation, male Balb/C mice were used. HSCs were isolated from the livers of control mice and of mice 3 weeks after BDL as described previously.21 The liver was perfused via the inferior vena cava with pronase (EMD Chemicals, Gibbstown, NJ) and collagenase (Roche, Indianapolis, IN). After digestion, the cell suspension was filtered through nylon mesh and purified via 8.2% Nycodenz (Axis-Shield, Oslo, Norway) gradient centrifugation. The isolated HSCs were cultured in uncoated plastic dishes with DMEM (GIBCO-BRL, Life Technologies, Rockville, MD) supplemented with 10% fetal bovine serum and antibiotic solution at 37°C in 5% CO2. The purity of HSCs was always 95% as determined by their typical starlike shape and abundant lipid droplets.
From the liver tissue and HSCs, total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA) and was treated with RNase-free DNase (Promega, Madison, WI) for 30 min at 37°C. After DNase treatment, the RNA was cleaned up using an RNeasy kit (Qiagen, MD). RNA was reverse-transcribed using a first-strand cDNA kit with random hexamers (Amersham Pharmacia Biotechnology, Buckinghamshire, UK) according to the manufacturer's protocol. MMP-13, chemokine (C-C motif) ligand 2 (CCL2), TNF-α, macrophage inflammatory protein 1beta (MIP-1β), macrophage inflammatory protein 2 (MIP-2), α−SMA, tissue inhibitor of metalloproteinase-1 (TIMP-1), transforming growth factor beta1 (TGF-β1), and collagen α1(I) mRNA were quantitated using TaqMan PCR with glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA and 18s rRNA as internal controls. The PCR reaction and analysis were carried out using the ABI Prism 7000 Sequence Detector and software (Applied Biosystems, Foster City, CA). The relative abundance of the target genes was obtained by calculating against a standard curve and normalized to an internal control. All PCR primers and probes were purchased from ABI TaqMan Gene Expression Assays.
Frozen livers were homogenized in lysis buffer (20 mmol/L HEPES [pH 7.5], 150 mmol/L NaCl, 10 mmol/L CHAPS) on ice using polytron homogenizer. The homogenates were centrifuged at 20,000g for 20 min at 4°C, and the protein concentration of the supernatants was measured by bicinchoninic acid (BCA) protein assay reagent using bovine serum albumin as a reference standard (Pierce, Rockford, IL). Proteins (50 μg) were mixed with 4× Laemmli sample buffer lacking β-mercaptoethanol, and the proteins were separated in 7.5% polyacrylamide gel containing 1 mg/mL of porcine skin gelatin (Sigma, St Louis, MO). After electrophoresis, SDS was removed from the gels by two washes of 15 minutes each with 2.5% Triton X-100. Subsequently, the gels were equilibrated in developing buffer (50 mmol/L Tris [pH 7.4], 200 mmol/L NaCl, 10 mmol/L CaCl2, 0.02% sodium azaid) for 30 minutes, then incubated in fresh developing buffer at 37°C for 24 hours. The gel was stained with 0.5% Coomassie Blue R-250 (Bio-Rad, Hercules, CA) for 1 hour, followed by destaining (50% methanol:10% acetic acid). Gelatinolytic activity was detected as clear bands on a dark blue background, and the molecular weight marker was included in the gel for identification.
Hydroxyproline content of the liver was measured by a spectrophotometric assay as an assessment of liver collagen content. Liver tissue was homogenized in ice-cold distilled water (1 mL) using a polytron homogenizer. Subsequently, 125 μL of 50% trichloroacetic acid was added, and the homogenates were further incubated on ice for 30 minutes. Precipitated pellets were hydrolyzed for 24 hours at 110°C in 6N HCL. After hydrolysis, the samples were filtered and neutralized with 10N NaOH, and the hydrolysates were oxidized with Chloramine-T (Sigma) for 25 minutes at room temperature. The reaction mixture was then incubated in Ehrich's perchloric acid solution at 65°C for 20 minutes and cooled to room temperature. Sample absorbance was measured at 560 nm. Purified hydroxyproline (Sigma) was used to set a standard. Hydroxyproline content was expressed as micrograms of hydroxyproline per gram liver.
Histological Analysis and Immunohistochemistry.
Liver tissues were fixed in PBS containing 4% paraformaldehyde for 24 hours at 4°C and embedded in paraffin. Sections (5 μm thick) were stained with Sirius Red (saturated picric acid containing 0.1% Direct Red 80 and 0.1% Fast Green FCF) for collagen visualization. Sections were also stained with chloroacetate esterase using a Naphthol-ASD Chloroacetate Esterase Kit (Sigma) to evaluate neutrophil accumulation in the liver. Immunohistochemistry was performed using anti-α-SMA (clone 1A4, Sigma), anti-desmin (clone D33, Dako, Carpinteria, CA), or anti-MMP-13 (Chemicon.com , Temecula, CA) with a Mouse on Mouse kit (Vector, Burlingame, CA). Detection of the primary antibody was carried out by using a biotinylated antibody and peroxidase DAB kit (Vector). Hematoxylin was used for counterstaining. For desmin staining, heat antigen retrieval was performed after rehydration.
Western Blot Analysis.
Liver tissue was homogenized in lysis buffer (20 mmol/L Tris HCl [pH 7.4], 150 mmol/L NaCl, 5 mmol/L EDTA, 10% glycerol, 0.1% Triton X-100) containing protease inhibitor cocktail (Roche) and phosphatase inhibitor. Protein lysate was separated on SDS-PAGE and transferred to nitrocellulose membrane (Schleicher & Schuell BioScience, Keene, NH). Equal loading of protein was determined by Ponceau S staining (Sigma) and by Western blot using a mouse anti-β-actin antibody (1:5,000 dilution, Sigma). Blotting was performed as described,22 using rabbit anti–heme oxygenase antibody (1:2,000 dilution, Stressgen, Victoria, Canada), rabbit anti-phospho-JNK (c-Jun N-terminal kinase) antibody (1:1,000 dilution, Cell Signaling, Beverly, MA) and an appropriate secondary antibody (1:2,000 dilution, Santa Cruz Biotechnology, Santa Cruz, CA). Protein expression was detected by enhanced chemiluminescence system (Amersham Pharmacia Biotechnology) and visualized using image station 2000R (Kodak, New Haven, CT).
Data are expressed as means ± SDs. Comparisons between two groups were performed with the Student t test. If the data were not normally distributed, the Mann-Whitney U test was applied for comparison of two groups. P less than .05 was considered statistically significant.
MMP-13 mRNA Expression Was Upregulated in HSCs and Livers after BDL.
To confirm that HSCs are a source of MMP-13 in the liver, HSCs were isolated from control and BDL-treated mice. One-day cultured HSCs from BDL mice exhibited a myofibroblast phenotype even though they still had lipid droplets in the cytoplasm (Fig. 1A). This phenotype was very similar to the in vitro–activated HSCs cultured on a plastic dish for 4 days. Nine-day-cultured HSCs exhibited a myofibroblast phenotype without lipid droplets (data not shown). Expression of MMP-13 mRNA did not change in quiescent and in vitro–activated HSCs (Fig. 1B). However, MMP-13 mRNA expression was increased in HSCs from the BDL livers.
Several growth factors and cytokines expressed in BDL liver have been reported to upregulate MMP-13 mRNA. To test whether MMP-13 expression is stimulated by cytokines, culture day 4 HSCs were treated with TNF-α for 3 hours, and then MMP-13 mRNA levels were measured. TNF-α treatment resulted in a fivefold increase in induction of MMP-13 in mHSCs (Fig. 1C). Expression of MMP-13 mRNA in the liver was also measured 5 days and 3 weeks after BDL. Expression of MMP-13 mRNA in the whole liver gradually increased over time after BDL in MMP-13 wild-type mice (Fig. 1D). Next, to identify the cell types that expressed MMP-13, we employed immunohistochemistry against MMP-13. MMP-13 was not expressed in sham-operated liver (Fig. 1E). In contrast, 5 days after BDL, MMP-13 was detected in infiltrated nonparenchymal cells and some hepatocytes around the portal area in the liver (Fig. 1E). These results demonstrate that MMP-13 is upregulated in the liver during BDL, and hepatocytes and nonparenchymal cells, including HSCs, are a major source producing MMP-13.
Liver Injury After BDL Was Reduced in MMP-13−/− Mice.
To investigate the effect of MMP-13 on BDL-induced liver injury and hepatic fibrosis, we used MMP-13−/− and MMP-13+/+ littermates. As expected, MMP-13 mRNA was undetectable in the MMP-13−/− mice before or after BDL (data not shown). Liver injury after BDL was assessed by measuring serum ALT level, which was 40% lower in MMP-13−/− mice than in MMP-13+/+ mice (Fig. 2A). Similar effects of bile duct obstruction were confirmed by histology, demonstrating the same degree of portal edema, biliary proliferation, and portal infiltrates in MMP-13+/+ and MP-13−/− liver (data not shown). To check whether MMP family members implicated in liver fibrosis were affected by the loss of MMP-13 in this model, we measured the activity of MMP-2 and MMP-9 using gelatin zymography. No differences in activity were seen between WT and KO sham-operated mice (Fig. 2B). Consistent with the literature, gelatinolytic activity of both MMP-2 and MMP-9 increased in the liver of MMP-13+/+ mice after BDL. On the other hand, the activity of MMP-2 and MMP-9 were lower in the livers of MMP-13−/− mice after BDL (Fig. 2B). Because neutrophil infiltration in the liver during BDL is a pathological feature of liver injury23 and neutrophils are a source of MMP-9,24 neutrophil accumulation in the sinusoids was evaluated by staining for chloroacetate esterase. Severe neutrophil infiltration in the sinusoids was observed in the BDL livers of MMP-13+/+ mice. On the other hand, only mild neutrophil infiltration was seen in the BDL livers of MMP-13−/− mice (Fig. 2C). To confirm whether neutrophil-derived reactive oxygen species were also reduced in the MMP-13−/− livers, expression of heme oxygenase-1 (HO-1), a marker of oxidative stress, and phosphorylation of JNK, a stress-activated protein kinase, were evaluated by Western blot analysis (Fig. 2D). The analysis showed reduced expression of HO-1 and reduced phosphorylation of JNK in the livers of MMP-13−/− mice after BDL. These data implicate MMP-13 in the pathogenesis of cholestatic liver injury.
Inflammatory Response Was Attenuated in Livers of MMP-13−/− Mice After BDL.
mRNA expression of proinflammatory mediators in the livers of BDL mice was measured by TaqMan real-time PCR (Fig. 3). Expression of mRNA of TNF-α, MCP-1, and MIP-1β increased in the livers of MMP-13+/+ mice after BDL. In contrast, expression of the mRNA of these proteins was significantly suppressed in MMP-13−/− livers after BDL. MIP-2 mRNA expression was greater in the MMP-13+/+ BDL mice than in the MMP-13−/− BDL mice, but the difference was not significant.
Fibrogenesis Was Suppressed in Livers of MMP-13−/− Mice After BDL.
To investigate the role of MMP-13 in liver fibrogenesis caused by BDL, expression of αSMA mRNA, a marker of HSC activation, was quantified by TaqMan real-time PCR (Fig. 4). In the MMP-13+/+ mice, expression of mRNA of αSMA was markedly increased compared with that in sham-operated mice. This increase was significantly reduced in the BDL livers of the MMP-13−/− mice. To ascertain if the suppression of HSC activation was also associated with reduced fibrogenesis, the expression of genes implicated in fibrogenesis was quantified (Fig. 4). Expression of collagen α1(I) mRNA, the major form of collagen in cirrhosis, was markedly increased in the BDL livers of MMP-13+/+ mice. In contrast, this increase was significantly suppressed in the BDL livers of MMP-13−/− mice. The mRNA of TGF-β, a key fibrogenic cytokine, was greater in the BDL livers of MMP-13+/+ mice than in MMP-13−/− mice. The mRNA of TIMP-1, which inhibits collagen degradation by MMPs and protects HSCs from apoptosis, was significantly higher in MMP-13+/+ mice than in MMP-13−/− mice. These data suggest the reduction of fibrogenesis correlated with the suppression of the inflammatory reaction during cholestasis in the livers of the MMP-13−/− mice.
Hepatic Fibrosis Was Reduced in Livers of MMP-13−/− Mice After BDL.
Three weeks after BDL, immunohistochemistory for αSMA was performed to evaluate HSC activation. Consistent with the mRNA data, the number of αSMA-positive cells was reduced in the livers from MMP-13−/− mice compared with livers from MMP-13+/+ mice (Fig. 5A). Immunostaining for desmin, which detects quiescent and activated HSCs, also exhibited a reduction in the total number of HSCs in the MMP-13−/− mice compared with in the MMP-13+/+ mice (Fig. 5B). Finally, hepatic collagen deposition was evaluated by morphometric analysis of Sirius Red staining (Fig. 6A) and quantified by digital image analysis (Fig. 6B). In the livers of the MMP-13+/+ mice, significant collagen deposition around the portal tracts with the formation of bridging fibrosis was seen after BDL. In contrast, these changes were markedly attenuated in the MMP-13−/− mice. We also measured total liver hydroxyproline levels, a biochemical marker of hepatic collagen content (Fig. 6C). In sham-operated livers of MMP-13−/− animals, there was a modest increase in hydroxyproline levels compared with that in MMP-13+/+ animals. After BDL, hydroxyproline levels were increased in both MMP-13+/+ and MMP-13−/− mice. The increase in the hydroxyproline level per gram liver after BDL was significant in the MMP-13+/+ mice but not in the MMP-13−/− mice (P < .05).
The beneficial effect of collagenase during the resolution of liver fibrosis has been well documented in recent studies.17, 25 Adenoviral gene delivery of human MMP-1 or MMP-8 in a rat liver with cirrhosis induced by thioacetamide, CCl4, or BDL resulted in the regression of fibrosis. In all cases, adenovirus delivery of collagenase to the liver was performed during the spontaneous recovery phase, when the inducing agent was removed. However, the role of collagenase during the development of fibrosis has not been clarified. In the present study, acute liver injury after BDL was reduced in mice deficient in MMP-13. These findings suggest that interstitial collagenase is required to induce fibrosis during the initial injury phase in BDL. Reduced liver injury is directly related to the decreased inflammation seen in the MMP-13-null mice, suggesting MMP-13 plays an important role in this initial inflammatory response, resulting in less fibrosis.
In the BDL-induced fibrosis model, MMP-13 levels were upregulated, and hepatocytes and nonparenchymal cells, including HSCs, around the portal area were a principal cellular source of MMP-13 production. Our results strongly support previous studies that demonstrated expression of MMP-13 in scar-associated macrophages and hepatocytes in CCl4-injured liver.26–28 Other studies have shown that activated HSCs express MMP-13 in vivo19 and in culture.29, 30 Cytokines, such as IL-1 and TGF-β, stimulate MMP-13 production in cultured-HSCs,30, 31 which was also found in our study. It is likely the expression and activity of MMPs are regulated by the ECM environment via the integrin and/or DDR2 pathway,31–34 and therefore upregulation of MMP-13 is found only in the context of a specific matrix environment, which was recapitulated in the BDL model system. These results indicate the pattern of expression of MMPs in HSCs may vary with the etiology and progression of the liver disease.
We found that activity of MMP-2 and MMP-9 in the MMP-13−/− liver after BDL was lower than that in the MMP-13+/+ mice. In normal liver, a small amount of type I collagen is distributed in the portal tract and space of Disse, where hepatocytes are separated from the sinusoidal endothelium.35 During the initial phase, it is likely the first cleavage of collagen is a result of MMP-13, and partially denatured collagen subsequently activates MMP-2 and MMP-9 as downstream effectors to further degrade the collagen. This positive feedback loop has been implicated in previous studies and suggests a role for MMPs in HSC proliferation.36, 37 Such a mechanism could explain why HSC proliferation was suppressed in the BDL livers of the MMP-13−/− mice, as assessed by desmin immunostaining.
Using Sirius Red staining and hydroxyproline quantification, we demonstrated BDL-induced hepatic fibrosis was attenuated in the MMP-13−/− mice. Compared with the results with Sirius Red staining, the decrease in hydroxyproline content was less impressive. The same phenomenon has been reported in previous studies17, 38, 39 and may be a result of the hydroxyproline assay detecting not only intact collagen accumulation but also degraded collagen, unlike Sirius Red, which detected only intact collagen.
Excessive or inappropriate expression of MMPs likely contributes to the pathogenesis of BDL-induced liver injury. Consistent with our result, several investigators reported that MMP-2 and MMP-9 were upregulated in liver diseases such as liver fibrosis and acute hepatitis, and inhibition of these MMPs suppressed liver damage.13, 40, 41 In addition, expression of human MMP-1 in the rat liver using an adenoviral gene delivery system resolved the fibrosis but also caused liver damage.17, 42 These results strongly suggest that breakdown of the ECM structure leads to cellular damage and results in an inflammatory reaction. The normal ECM is essential for maintaining the homeostasis of all resident liver cells. The importance of the normal ECM in the liver has been well documented in recent attempts to develop artificial liver support. It is recognized that all cellular elements and supporting structures must be reconstituted to maintain the differentiated function of the liver.43 Therefore, it is likely that degradation of the ECM by MMPs alters cell-matrix and cell-cell interactions and enhances hepatocyte susceptibility to necrosis and/or apoptosis caused by cholestasis. Loss of MMP-13 possibly attenuated this process, which is fundamental for the initiation of fibrotic repair post–liver injury in the BDL model.
The ECM serves as a binding reservoir for several key cytokines such as TGF-β, TNF-α, PDGF, and basic FGF.44–46 MMPs release soluble bioactive factors through ECM degradation and regulate macrophage chemoattractant and leukocyte infiltration during injury. MMP-13 may activate MMP-2, which in turn is able to activate chemokines CCl-7 and CXCl-12.47, 48 In fact, colonic myofibroblasts secrete metalloproteinases that activate the neutrophil chemokine CXCl-7.49 This function of MMPs may explain the detrimental effect of MMPs on hepatic inflammation and hepatic fibrogenesis. Inflammation is an important feature of cholestatic liver disease and leads to fibrogenic changes in the liver such as HSC activation. Cytokines released from activate Kupffer cells stimulate proliferation and activation of HSCs.2 Reactive oxygen species (ROS) derived from neutrophils stimulate collagen synthesis in HSCs.50 In addition, selective depletion of macrophages during CCl4-induced liver injury results in suppression of liver fibrosis.51 We demonstrated that the inflammatory reaction was reduced in the BDL livers of the MMP-13−/− mice. Induction of HO-1, an indicator of oxidative stress,23 was also reduced in the livers of MMP-13−/− mice after BDL. Finally, we demonstrated that expression of mRNA for markers of HSC activation, collagen α1(I), α-SMA (alpha smooth muscle actin), and TIMP-1, was suppressed in the BDL livers of MMP-13−/− mice. These findings indicate that absence of MMP-13 attenuates the inflammatory reaction during BDL and suppresses HSC activation and fibrogenesis. Further studies will investigate the mechanism of this protective role for MMP-13.
In summary, our findings suggest MMP-13 mediates liver injury, inflammation, HSC activation, and hepatic fibrogenesis during cholestasis post-BDL in the mouse. Although previous studies have shown that interstitial collagenase exhibits a beneficial function for the resolution of stable liver fibrosis, at the onset of cholestasis-induced fibrosis, the role of MMP-13 is distinct and opposite.