In fibroblasts, thrombin induces collagen deposition through activation of a G-protein–coupled receptor, proteinase-activated receptor 1 (PAR1). In the current study, we examined whether PAR1 antagonism inhibits hepatic stellate cell (HSC) activation in vitro and whether it protects against fibrosis development in a rodent model of cirrhosis. A rat HSC line was used for in vitro studies whereas cirrhosis was induced by bile duct ligation (BDL). The current results demonstrated that HSCs express PAR1, as well as proteinase-activated receptors 2 (PAR2) and 4 (PAR4), and that all three PARs were up-regulated in response to exposure to growth factor in vitro. Exposure to thrombin and to SFLLRN-(SF)-NH2, a PAR1 agonist, and GYPGKF (GY)-NH2, a PAR4 agonist, triggered HSC proliferation and contraction, as well as monocyte chemotactic protein-1 (MCP-1) production and collagen I synthesis and release. These effects were inhibited by the PAR1 antagonist. Administration of this antagonist, 1.5 mg/kg/d, to BDL rats reduced liver type I collagen messenger RNA (mRNA) expression and surface collagen by 63%, as measured by quantitative morphometric analysis. Similarly, hepatic and urinary excretion of hydroxyproline was reduced significantly by the PAR1 antagonist. In conclusion, PARs regulates HSC activity; development of PAR antagonists might be a feasible therapeutic strategy for protecting against fibrosis in patients with chronic liver diseases. (HEPATOLOGY 2004;39:365–375.)
Thrombin is a pluripotent serine protease that plays a central role in hemostasis after tissue injury by converting soluble plasma fibrinogen to an insoluble fibrin clot and by promoting platelet aggregation.1 In addition to these procoagulant effects, thrombin influences the recruitment and trafficking of inflammatory cells and is a potent mitogen for a number of cell types, including endothelial cells, fibroblasts, and smooth muscle cells.1–4 Thrombin also promotes the production and secretion of extracellular matrix proteins and influences connective tissue remodeling in normal tissues, as well as in a number of pathologic conditions including liver fibrosis.5–7
As with other serine proteases, most of the cellular effects elicited by thrombin are mediated via a family of widely expressed G-protein–coupled receptors, termed protease-activated receptors (PARs).1–4, 8–14 Proteases activate PARs by cleaving the NH2-terminal sequence of the extracellular exodomain.9, 10 This cleavage event unmasks an amino terminal sequence, which in turn serves as a tethered ligand, binding intramolecularly to the body of the receptor to trigger transmembrane signaling.9, 10 Molecular cloning has identified four PARs: PAR1 and PAR3, which are both preferentially activated by thrombin11, 13; PAR2, which is selectively activated by trypsin (and tryptase)12; and PAR4, which is activated by both thrombin and trypsin.8, 10, 14–16 PAR3 is a PAR4 cofactor and its activity is indistinguishable from that of PAR4.8–10 In addition to endogenous proteases, PARs can be selectively activated by short agonistic peptides (PAR-AP) that correspond to the tethered ligand exposed after PAR1and PAR2 cleavage.9, 10 Studies with these agonists, as well as with PAR1-deficient mice, have led to the conclusion that PAR1 is the major receptor responsible for mediating most of the proinflammatory and profibrotic effects17 of thrombin.
Cirrhosis is a scarring process that includes components of both increased fibrogenesis and wound contraction.6, 7 Hepatic stellate cells (HSCs) are increasingly being recognized as the key players in liver fibrogenesis.6, 7, 18, 19 In chronic liver disease, HSCs acquire an “activated” phenotype, which includes increased proliferation, contractility, and the ability to release collagen along with chemotaxis and cytokine.7, 8, 18–24 The current paradigm postulates that the activated state of HSCs is achieved through the transformed microenvironment, which is contributed to by the growth factors platelet-derived growth factor (PDGF) and transforming growth factor (TGF)-β, reactive oxygen intermediates which are released by hepatocytes and by the fibrillar matrix generated by previously activated HSCs.18–24
Thrombin might be mechanistically involved in regulating hepatic fibrogenesis.5, 17 Even though thrombin expression is increased in cirrhotic patients5 and exposure of HSCs to thrombin triggers a number of effector functions, including proliferation, contraction, and collagen synthesis,21, 22 the functional relevance of PAR1 in modulating HSC activity and collagen deposition in vivo is unknown.
The 4-methoxy-N-1-[(2,6-dichlorophenyl) methyl]-3-(1-pyrrolidinylmethyl)-1H-indol-6-yl]amino]carbonyl]-L-phenylalanyl-N-(phenylmethyl)-L-argininamide (Fig. 1) was prepared using a modification of the method decribed by Andrade-Gordon et al.25 This peptide mimetic is a selective PAR1 inhibitor that inhibits platelet aggregation induced by SF-NH2 with an EC50 of 0.16 ± 0.06 μmol/L.25 The following APs were synthesized: SFLLRN-(SF)-NH2, a PAR1 agonist; SLIGRL (SL)-NH2, a PAR2 agonist; and GYPGKF (GY)-NH2, a PAR4 agonist.8–10
In vitro studies were performed on primary cultures of HSCs and on an immortalized cell line of rat HSCs (HSC-T6) kindly provided by S. Friedman (Mount Sinai Hospital, New York, NY).26 The HTC-T6 was cultured in P35 Petri dishes (Nunc International, Naperville, IL) at 37°C in an atmosphere of 5% CO2 in Dulbecco's modified minimal essential medium (Gibco BRL Life Technologies, Rockville, MD) containing 10% fetal calf serum (FCS), 2 mmol/L L-glutamine, and 5,000 IU/mL penicillin/ 5,000 μg/mL streptomycin for 1 to 2 days before starting the experiments. HSC-T6 had a pseudo-myofibroblastic phenotype, characteristic cell shape, and expression of α-smooth muscle actin (α-SMA), vimentin, and glial fibrillar acidic protein.26
Rat HSCs were isolated from control rats and rats that had undergone bile duct ligation (BDL) according to a modification of the technique of Knook et al.27 This entails liver perfusion in situ and digestion with solutions containing pronase E followed by collagenase Ia (Sigma Chemical, St. Louis, MO). The digested liver specimen was filtered and the resulting suspension was cleansed of hepatocytes and resuspended in a discontinuous Nycodenz gradient (Axis-Shield, Oslo, Norway). Centrifugation yields HSC in the top layer and a Kupffer/endothelial cell fraction in the lower layer. Isolated HSCs were more than 90% viable as assessed by trypan blue exclusion. HSCs were cultured on plastic plates in Dulbecco's modified minimal essential medium containing 20% FCS.
Expression of PARs on HSC-T6 and HSCs: Reverse-Transcription Polymerase Chain Reaction and SYBR Green.
HSCs and HSC-T6 were incubated for 1 to 5 days with 20% FCS. At indicated time points (Fig. 2), samples of total RNA were isolated using TRIzol reagent (Life Technologies, Milan, Italy).28 Reverse-transcription polymerase chain reaction (RT-PCR) was performed using the following sense and antisense primers (Sigma Genosys, Milan, Italy): β-actin 5′-TAC CAC TGG CAT TGT GAT GG-3′ and 5′-TTAATGTCACGCACGATT TC-3′; PAR1 5′-TAGCTGCATCGA CCCCTTGAT T-3′ and 5′-TTGACTCCGTTCCCATCACCTT-3′; PAR2 5′-GAACCA- GGCTTTTCC GTTGATG-3′ and 5′-ATGGCAGCACCGTGATGATG A-3′ ; PAR4 5′-GAACCAGGCTTTTCCGTT GAT G-3′ and 5′-ATGGACAGCACCGTG- ATGATG A-3′; procollagen-α1(I), 5′-ACAGCACGC-TTG TGGAT-3′ and 5′-GTCTTCAAGCAAGAGGACCA-3′; TGF-β1 5′-CGGACTACTACGCCAAAGAA-3′ and 5′-TGGTTTTGTCATAGATTGCCGTT-3′.
For the quantitative SYBR Green RT-PCR,29, 30 250 ng of complementary DNA was used per reaction. Each 25-μL SYBR Green reaction consisted of 5 μL complementary DNA (50 ng/μL), 12.5 μL of 2× Universal SYBR Green PCR Master Mix (PE Biosystems, Foster City, CA), and 4.00 μL of 50 nmol/L forward and reverse primers. Optimization was performed for each gene-specific primer before the experiment to confirm that 50 nmol/L primer concentrations did not produce nonspecific primer-dimer amplification signals in no-template control tubes. Quantitative RT-PCR was performed on a BioRad I cycler (BioRad, Hercules, CA).27, 28 Values are expressed as a ratio between threshold cycles (CT).29, 30 The CT for each PAR in each sample was normalized to the CT of β-actin and the resulting ratio was compared with a similar ratio observed in control cells.
DNA Synthesis and Cell Growth Assay.
To assess the effect of PAR1 antagonist on HSC-T6 proliferation, cells were plated in 12-well dishes at a density of 5 × 104 cells per well and allowed to adhere for 24 hours. Cells were then washed twice and incubated for 1 to 5 days with or without 50 μmol/L of PAR1 antagonist without replacing the medium, but supplementing it daily with 50 μL of 10% FCS. In the experiment, 2 mCi/mL of [3H]-thymidine was added for the last 18 hours of incubation. Labeled cells were washed twice with phosphate-buffered saline and lysed in 1 mol/L NaOH. Lysates were counted in a liquid scintillation counter. For cell counting, HSCs were plated in 12-well dishes at a density of 5 × 104 cells per well and allowed to adhere for 24 hours. After incubation with appropriate agents, the cells were detached by gentle trypsinization and counted daily using a cell counter (Beckman-Coulter SPA, Rome, Italy).
Contraction of HSC-T6 on collagen lattices was performed in 24-well flat-bottom tissue culture plates as described previously.28 Cell contraction was induced by adding 10 U thrombin or 1 to 100 μmol/L SF-NH2, SL-NH2, and GY-NH2 to the HSC monolayers with or without 0.1 to 100 μmol/L anti-PAR1. The lattice was detached by gentle circumferential dislodgment with a 200-μL micropipette tip and contraction was measured by monitoring the change in lattice area over 24 hours.
Monocyte Chemotactic Protein-1 Release by HSC-T6.
HSC-T6 (4 × 105/mL) was incubated with 100 ng/mL of tumor necrosis factor-α with or without PAR1 antagonist (50 μmol/L) for 24 hours and the monocyte chemotactic protein-1 (MCP-1) released in cell supernatants assayed by a specific enzyme-linked immunosorbent assay (Endogen, Woburn, MA).22, 28
ERK-1 and ERK-2.
ERK-1 and ERK-2 phosphorylation was assessed in HSCs and HSC-T6 lysates as described previously using a phosphorylated ERK-1 and ERK-2 (phospho-p44/42 MAPK [Thr202/Tyr204]; 1:2,000 dilution) monoclonal antibody (Cell Signalling Technologies, Beverly, MA).31, 32 ERK-1 and ERK-2 activity was measured as described previously28, 31 using a commercial assay (Cell Signalling Technology).
Immunoprecipitation and Shc/Grb2 Immunoblotting.
HSCs and HSC-T6 were serum starved overnight, treated with PAR-AP or thrombin in the presence of anti-PAR1, and lysed in ice-cold modified radioimmuno precipitation assay (RIPA) buffer. Proteins were incubated with 5 μg of rabbit polyclonal anti-Shc antibody (Transduction Laboratories, San Jose, CA) for 4 hours at 4°C and then with protein A-conjugated agarose beads overnight with constant shaking at 4°C.31, 32 Immune complexes were washed three times with ice-cold RIPA buffer, denatured in Laemmli sample buffer, and resolved by 7.5% sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Shc/Grb2 complexes were detected by incubating immunoprecipitates with polyclonal anti-Grb2 immunoglobulin G (Santa Cruz Biotechnology, Santa Cruz, CA) as described previously.30, 31 Immune complexes on nitrocellulose were visualized by enzyme-linked chemiluminescence (Amersham Biosciences, Arlington Heights, IL) and quantified by densitometry analysis (Kodak Digital Science, Rochester, NY).
After appropriate treatment, HSC and HSC-T6 lysates were sonicated and boiled for 3 to 5 minutes. An aliquot of 10 μg cell protein was resolved on a 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel and electroblotted onto a nitrocellulose membrane. The membrane was incubated with a monoclonal anti–α-SMA (clone 1A4; Sigma Chemical) as the primary antibody, followed by alkaline-conjugated goat anti-mouse immunoglobulin G as the secondary antibody. Immunoreactive proteins were visualized and quantified as described earlier in the text.
In Vivo Studies.
In this protocol, we evaluated whether PAR1 antagonism could reverse fibrosis induced by BDL in 8 to 9-week-old male Wistar rats (Charles River, Monza, Italy).32 The protocol was approved by the animal study committee of the University of Perugia. Three weeks after BDL, rats were randomized to receive placebo (subcutaneous injection of 100 μL phosphate-buffered saline) or 0.5, 1.5, and 3 mg/kg/d PAR1 antagonist subcutaneously starting 21 days after surgery. Animals were followed for 10 days. At the end of the study, surviving rats were anesthethesized with pentobarbital sodium (50 mg/kg intraperitoneally), killed, and bled via cardiac puncture. The blood was centrifuged at 7,250g for 20 minutes at 4°C. The resultant serum sample was stored at −20°C until analysis (a maximum of 2 weeks). After confirming that the bile duct ligature was intact with proximal dilatation of the common bile duct, liver specimens were removed and snap frozen at −70°C in liquid nitrogen.
Sections of the right and left liver lobes (10-15 mg each) from each animal were fixed in 10% formalin, embedded in paraffin, sectioned, and stained with hematoxylin-eosin or Sirius red. For Sirius red staining,33, 34 the sections were incubated for 30 minutes in 0.1% Sirius red F3B (Sigma Chemical) containing saturated picric acid and 0.1% Fast Green. After rinsing twice with distilled water, sections were briefly dehydrated with 70% ethanol and coverslipped. The collagen surface density of the liver samples was quantified using a computerized image analysis system (Image Acquisition System version 005, Delta Sistemi, Rome, Italy).28 The surface density of collagen in blinded specimens was measured on a video screen display magnification and expressed as a percent (the ratio of collagen surface area per total analyzed field surface). The average of the score taken from 10 random fields was used to generate a single score for each animal's liver specimen.
Identification of Proliferating and Activated HSCs.
Analysis of HSC activation was performed by simultaneous staining of α-SMA (as a marker of activated HSCs) and proliferating cell nuclear antigen (as a marker of S-phase nuclei). A sequential double immunoenzymatic reaction was performed using methanol-fixed sections.
Hepatic and Urinary Hydroxyproline Determination.
Hydroxyproline content in liver and urine samples was measured by high-performance liquid chromatography with the LC Varian Prostar 330 (Varian Instruments, Walnut Creek, CA) equipped wth a UV-Vis 310 Varian detector.35
ANOVA with Bonferroni correction was used to make statistical comparisons. Data are means ± SEM.
Expression of PAR1, PAR2, and PAR4 on HSCs.
As illustrated in Fig. 2, rat HSCs express PAR1, PAR2, and PAR4 messenger RNA (mRNA). The expression of PARs increases over time following the same time course of α-SMA expression (Fig. 2A). Quantitive RT-PCR analysis (Fig. 2C) demonstrates that the a 5-day culture of HSCs results in a seven to eightfold increase of PAR1, PAR2, and PAR4 mRNA expression in comparison to a 1-day culture of HSCs (n = 6, P < .01). Similarly, HSC-T6 express PAR1, PAR2, and PAR4 mRNA (data not shown). HSC-6 were used for most of the functional experiments described hereafter.
To investigate whether the PAR1 agonist activates effector function, HSC-T6 were cultured with 1 to 100 μmol/L SF-NH2, a short mimetic peptide that selectively binds PAR1. Exposure to SF-NH2 resulted in a concentration-dependent stimulation of HSC-T6 proliferation as measured by [3H]-thymidine incorporation and cell counts (Fig. 3A,B ; n = 6, P < .05), MCP-1 release (Fig. 3C; n = 6, P < .05), and contraction (Fig. 3D; n = 6-8, P < .01). The relative EC50 for each of these effects was 62.5 ± 5.3, 57.5 ± 11.3, 37 .5 ± 9.0, and 28.0 ± 6.1 μmol/L, respectively. Incubating the HSC-T6 with 50 μmol/L PAR1 antagonist significantly attenuated the stimulatory effect of SF-NH2 (n = 6-8, P < .01).
Exposure of HSC-T6 with 50 μmol/L SL-NH2, a PAR2 agonist, and GY-NH2, a PAR4 agonist, also resulted in cell activation as measured by cell proliferation (Fig. 4A ; n = 8, P < .01 vs. control), MCP-1 release (Fig. 4B; n = 8, P < .01 vs. control), and contraction (Fig. 4C; n = 8, P < .01 vs. control). Treating the cells with anti-PAR1 reverted in a concentration-dependent manner HSC-T6 activation induced by SF-NH2, whereas the antagonist had no effect on activation induced by SL-NH2 (n = 6-8). As shown in Fig. 4D-F, the PAR1 antagonist was also effective in inhibiting HSC-T6 activation induced by QY-NH2, that is, PAR4. However, inhibition of HSC-T6 induced by QY-NH2 was partial and required 50 to 100 μmol/L of PAR1 antagonist, a concentration significantly higher than that required for PAR1 inhibition. HSC-T6 proliferation induced by 25 ng/mL PDGF and 100 ng/mL TGF-β was insensitive to the effect of the PAR1 antagonist (data not shown), confirming the selectivity of this agent (n = 6, P > .05 vs. PDGF and TGF-β).
Anti-PAR1 Reverses HSC-T6 Activation Induced by Thrombin.
Figure 5 illustrates that similar to SF-NH2, exposure to 10 U/mL thrombin triggers HSC-T6 proliferation, contraction, and MCP-1 release (Fig. 5). These effects were attenuated by 50 μmol/L of the PAR1 antagonist (n = 6, P < .01).
PAR1 Antagonism Inhibits Collagen Synthesis by HSCs.
Exposure of HSC-T6 to thrombin and SF-NH2 is associated with a significant increase in collagen synthesis as measured by hydroxyproline release in cell supernatants (Fig. 6A ; n = 6-8, P < .01) and α1 collagen mRNA accumulation (Fig. 6B; n = 8, P < .05). These effects were significantly attenuated by exposing the cells to the PAR1 antagonist (n = 6-8, P < .01 vs. thrombin or SL-NH2).
The PAR1 Antagonist Inhibits Shc/Grb2 Complex Formation and ERK Activation Induced by SF-NH2 and Thrombin.
To examine whether PARs activate Shc/Grb2/ERK-1 and ERK-2 in HSCs, and to compare the activity of this pathway in HSCs obtained from intact and BDL rats, primary cultures of HSCs were exposed to thrombin and SF-NH2 and Shc expression/activation was examined. As illustrated in Fig. 7A, by using a Shc polyclonal antibody three Shc isoforms (p46Shc, p52Shc, and p66Shc) were found in the immunoprecipitates of HSCs in a 5-day culture.35, 36 Exposure of HSCs to 10 U of thrombin (Fig. 7B) caused p46Shc and p52Shc, but not p66Shc, phosphorylation. Thrombin-induced tyrosine phosphorylation of Shc proteins occurred after 1 minute, peaked at 15 minutes, and remained stable for approximately 2 hours (data not shown). As with thrombin, exposure to 50 μmol/L SF-NH2 caused p46Shc and p52Shc phosphorylation (Fig. 7B). Incubating HSCs with 50 μmol/L of PAR1 antagonist abrogated p46Shc and p52Shc phosphorylation induced by thrombin and SF-NH2 (Fig. 7B).
Western blot analysis with anti-Grb2 polyclonal antibody of Shc immunoprecipitates demonstrates that after thrombin and SF-NH2 stimulation of HSCs, Grb2 is recruited into a Shc/Grb2 complex (Fig. 7C). Exposure to anti-PAR1 abolished the Shc/Grb2 complex formation triggered by thrombin and SF-NH2 (Fig. 7C). As shown in Fig. 7D and E, thrombin and SF-NH2 caused ERK-1 and ERK-2 phosphorylation/activation. This effect was also abrogated by exposure to 50 μmol/L of the PAR1 antagonist (Fig. 7D,E).
In Vivo Study.
Three weeks after BDL, rats showed significant weight loss compared with sham-operated animals and controls (Table 1). Analysis of PARs in HSCs obtained from BDL rats and cultured for 1 day (Fig. 8A) demonstrates that BDL resulted in a five to eightfold induction of liver expression of PAR1, PAR2, and PAR4 mRNA (n = 4-6, P < .01 vs. control and sham-operated rats). PARs mRNA accumulation correlated with increased expression of p46Shc and p52Shc and Shc/Grb2 complex formation in HSCs obtained from BDL rats (Fig. 8C,D). BDL significantly increased collagen content as demonstrated by the liver histology and liver fibrosis score (Table 2), liver hydroxyproline content (Fig. 9A), urinary excretion of hydroxyproline (Fig. 9B), and type I collagen mRNA expression in the liver specimens (Fig. 9C). Liver histopathology demonstrates that Sirius red-positive fibrils were not only associated with proliferating bile ducts in the portal area. We also observed a continuous meshwork of connective tissue infiltrating the hepatic parenchyma with central-central, central-portal, and portal-portal bridging (Fig. 9D,E). In comparison to control and sham-operated rats, BDL rats had a significant increase in α-SMA as examined by Western blot analysis and α-SMA–expressing cells (Fig. 9H,I).
Table 1. Effect of PAR1 Antagonist on Rat Body Weight, Liver Weight, and Biochemical Parameters in BDL Rats
Central and portal refer to the central vein and the portal tract areas, as well as the parenchymal area immediately surrounding these spaces. All refers to all hepatic areas, as visualized under low magnification.
In vivo delivery of anti-PAR1 resulted in a dose-dependent amelioration of liver injury as assessed by fibrinogen, alanine aminotransferase, bilirubin, and gamma glutamyl transpeptidse plasma levels (Table 1; n = 10-12, P < .05 vs. BDL). In dose-finding experiments, we demonstrate that maximal protection was observed in rats treated with 1.5 mg/kg/d and no statistically difference was found between 1.5 and 3 mg/kg/d. Analysis of PARs on freshly isolated HSCs demonstrated that treating rats with 1.5 mg/kg/d of the PAR1 antagonist reduced PAR1, PAR2, and PAR4 mRNA expression (Fig. 8B; n = 6, P < .01 vs. BDL alone), Shc expression/phosphorylation, and Shc/Grb complex formation (Fig. 8C,D). PAR1 antagonism also reduced liver collagen deposition, liver hydroxyproline content, hydroxyproline urinary excretion, type I collagen mRNA expression in the liver, hepatic fibrosis score, and α-SMA content (Tables 1, 2; Fig. 9). Liver histopathology (Fig. 8F,G) demonstrates significant attenuation of bridging by PAR1 antagonism. Quantitative analysis demonstrated a 62% reduction in liver collagen content as stained by Sirius red (Fig. 8F,G; Table 2).
PARs are involved in the tissue remodeling and repairing activity of thrombin.1–4 PAR1 activation on endothelial cells causes a range of responses (e.g., changes in cell shape and monolayer permeability, expression of interleukin-8 and other chemokines, as well as expression of ICAM-1, VCAM-1, and cyclo-oxygenase-2–derived prostanoids) that presumably serve to recruit platelets and leukocytes to sites of injury and to promote plasma protein extravasation.1, 8–10 Furthermore, thrombin-induced activation of PAR1 stimulates the synthesis and shedding of profibrogenic factors such as heparin-binding epidermal growth factor, PDGF, connective tissue growth factor, fibroblast migration and proliferation, and extracellular matrix production, providing a system that detects tissue injury and triggers a set of cellular responses that contribute to inflammation and tissue repair and remodeling.6, 7, 37
Thrombin has been involved in hepatic fibrogenesis in humans, as well in experimental models of liver injury.5, 38 Immunohistochemistry and in situ hybridization studies have demonstrated, that although PAR1 is expressed in the normal liver in the endothelial lining of the hepatic sinusoids, its expression is up-regulated in response to acute or chronic liver injury. This suggests that PAR1 plays a mechanistic role in liver remodeling and scarring.
In the current study, we demonstrated that PARs are functionally involved in regulating liver collagen deposition. To support this view, we provided four lines of evidence. First, mRNA expression of PAR1, PAR2, and PAR4 was detected in HSCs and in the HSC-T6 cell line. Second, expression of PARs in HSCs increases with the appearance of a fibroblast-like phenotype (i.e., α-SMA). Third, exposure of HSCs to thrombin or SF-NH2 and GY-NH2 induces collagen synthesis and elicits a variety of functional responses (proliferation, MCP-1 production, and cell contraction) that are mechanistically linked to the development of hepatic inflammation and fibrogenesis.6, 7 Fourth, the finding that PAR1 antagonism inhibits HSC functions in vivo and in vitro and protects against fibrosis development in a well-characterized model of liver injury supports a mechanistic role for PAR1 in regulating liver collagen deposition.
Treating HSCs with anti-PAR1 inhibits a number of intracellular messengers involved in mediating downstream signaling induced by thrombin.36 Previous studies on fibroblasts have shown that thrombin-receptor activation causes Shc phosphorylation.36, 39, 40 In response to thrombin and other growth factors, Shc becomes tyrosine phosphorylated and forms a complex with Grb2,35 leading to translocation of the adaptor factor, Sos, to the plasma membrane where Ras is located.40 Ras recruitment triggers the activation of the MAP kinase cascade, as well as ERK-1 and ERK-2 phosphorylation/activation. We have shown that exposure of HSCs in a 5-day culture to thrombin and SF-NH2 causes Shc phosphorylation and Shc/Grb2 complex formation, ultimately leading to ERK-1 and ERK-2 phosphorylation and activation. Because this pathway is linked to cell proliferation and collagen production in fibroblasts,36 it seems reasonable to suggest that the antifibrotic effect we observed with the PAR1 antagonist in vivo and in vitro is likely due to the interference that anti-PAR1 exerts onERK-1 and ERK-2 activation.
The current study also demonstrated that all known members of the PAR family, including PAR2, stimulate HSC proliferation/activation in vitro. PAR2 is a receptor for trypsin and mast cell tryptase. However, its activation is also stimulated by coagulation proteases such as Factor VIIa and Factor Xa and by tissue factor.1, 8–10, 41 Factor VIIa activates only PAR2, whereas Factor Xa activates PAR1 and PAR2.1 A study by Riewald and Ruf41 suggested that a ternary complex of tissue factor-Factor VIIa-Xa is required for PAR1 and PAR2 activation.1, 3, 8–10 The observation that HSCs express PAR2 and that PAR2 mRNA expression is increased in the liver of BDL rats raises the possibility that PAR2 activation also contributes substantially to liver fibrosis in this model. Supporting this view, we also demonstrated that PAR1 antagonism reduces PAR2 expression in the liver of BDL rats, which suggests that not only is PAR expression coordinately modulated during liver fibrogenesis, but that inhibition of HSC activation with anti-PAR1 might regulate the expression of other PARs. Unfortunately, the current unavailability of selective PAR2 inhibitors prevents the testing of this hypothesis in established models of liver fibrosis.42
One limitation of the current study is that PAR1 is expressed on a variety of cell types. Therefore, we cannot exclude that PAR1 inhibition in cells other than HSCs might contribute to its antifibrotic effect. Inhibition of PAR1 reduces MCP-1 production, an event that might limit proinflammatory changes in the liver microenvironment. Because HSCs are activated in the presence of inflammatory mediators,19 inhibition of MCP-1, as well as inhibition of other cytokines and chemokines, might interrupt the autoamplifying loop involved in HSC activation.
In addition to its antifibrotic activity, PAR1 antagonism had a beneficial effect on serum liver enzymes. In this model, the biliary obstruction is mechanical. Yet, the parameters of cholestasis were improved during treatment with the PAR1 antagonist. Similar results were obtained with antifibrotic agents, including inhibitors of angiotensin-converting enzyme.43 Although the reason remains uncertain, these studies support the hypothesis that reduction of liver fibrosis limits cholestatic injury.44
In conclusion, our study demonstrates that PAR1 antagonism protects against fibrosis in a rodent model of liver injury. Our data also suggest that PAR1, PAR2, and PAR4 may play an important role in the development of liver fibrosis, supporting the hypothesis that the development of PAR antagonists might represent a new therapeutic approach to prevent fibrosis in chronic liver diseases.