STAT1 inhibits liver fibrosis in mice by inhibiting stellate cell proliferation and stimulating NK cell cytotoxicity

Authors

  • Won-Il Jeong,

    1. Section on Liver Biology, Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Ogyi Park,

    1. Section on Liver Biology, Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Svetlana Radaeva,

    1. Section on Liver Biology, Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Bin Gao

    Corresponding author
    1. Section on Liver Biology, Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD
    • Section on Liver Biology, NIAAA/NIH, 5625 Fishers Lane, Room 2S-33, Bethesda, MD 20892
    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

Abstract

Liver fibrosis, a common scarring response to chronic liver injury, is a precursor to cirrhosis and liver cancer. Here, we identified signal transducer and activator of transcription 1 (STAT1) as an important negative regulator in liver fibrosis. Our findings show that disruption of the STAT1 gene accelerated liver fibrosis and hepatic stellate cell (HSC) proliferation in an in vivo model of carbon tetrachloride (CCl4)-induced liver fibrosis. In vitro treatment with IFN-γ inhibited proliferation and activation of wild-type HSCs, but not STAT1−/− HSCs. Moreover, compared to wild-type cells, cellular proliferation stimulated by serum or platelet-derived growth factor (PDGF) was enhanced and accelerated in STAT1−/− HSCs, which was partially mediated via elevated PDGF receptor β expression on such cells. Polyinosinic-polycytidylic acid (poly I:C) or IFN-γ treatment inhibited liver fibrosis in wild-type mice but not in STAT1−/− mice. Induction of NK cell killing of activated HSCs by poly I:C was attenuated in STAT1−/− mice compared to wild-type mice, which was likely due to reduced NKG2D and TRAIL expression on STAT1−/− NK cells. Finally, activation of TGF-β/Smad3 signaling pathway was accelerated, whereas induction of Smad7 was diminished in the liver of STAT1−/− mice after CCl4 administration compared to wild-type mice. In conclusion, activation of STAT1 attenuates liver fibrosis through inhibition of HSC proliferation, attenuation of TGF-β signaling, and stimulation of NK cell killing of activated HSCs. STAT1 could be a new therapeutic target for treating liver fibrosis. (HEPATOLOGY 2006;44:1441–1451.)

Alcohol consumption, viral hepatitis, and nonalcoholic steatohepatitis are 3 major causes of chronic liver injury leading to liver fibrosis, cirrhosis, and liver cancer. Regardless of etiology, liver injury elevates a variety of cytokines and growth factors, which can activate hepatic stellate cells (HSC) to produce collagen and cause liver fibrosis.1–5 Emerging evidence suggests that both transforming growth factor-β (TGF-β) and platelet-derived growth factor (PDGF) are the 2 most important cytokines responsible for HSC activation and proliferation.1–7 It is believed that TGF-β plays a key role in transforming quiescent HSCs into fibrogenic myofibroblasts by stimulating synthesis of ECM proteins and inhibiting their degradation,2, 6, 8 whereas PDGF is a potent mitogen for HSC proliferation and activates a variety of signaling molecules including phosphatidylinositol 3 (PI3) kinase, Ras/Raf kinase, and MAP kinase.7, 9–14 Additionally, PDGF also activates the signal transducer and activator of transcription 1 (STAT1) and STAT315; however, the role of these STATs in HSC activation and proliferation remain undefined.

In contrast to TGF-β and PDGF which activate HSCs, interferon-γ (IFN-γ) and IFN-α have been identified as important negative regulators in liver fibrosis. In rodents, disruption of the IFN-γ gene enhances liver fibrosis,16, 17 whereas treatment with IFN-α or IFN-γ has been shown to ameliorate liver fibrosis induced by carbon tetrachloride (CCl4) or dimethylnitrosamine.18–20 Clinical data have shown that IFN-α therapy improves serum levels of fibrotic markers and liver histology in both virologic responder and nonresponder patients with chronic hepatitis C virus (HCV) infection,21–24 and that IFN-γ treatment improves liver fibrosis in patients with chronic HBV or HCV infection;25, 26 however, a recent double-blind, controlled clinical trial showed that IFN-γ therapy has no beneficial effect in HCV patients with advanced liver cirrhosis.27 The antifibrotic effects of IFN-α and IFN-γ are believed to be achieved by directly inhibiting HSC proliferation and activation,18–20, 28, 29 blocking TGF-β signaling,30 and stimulating natural killer (NK) cell killing of activated HSCs.17 Although the antifibrotic effects of IFNs in the liver are well documented, the downstream signals involved in direct attenuation of liver fibrosis remain unclear. The actions of IFN-γ are mediated through binding to IFN-γ receptor 1 (IFNGR1) and IFNGR2, followed by phosphorylation of receptor-associated Janus kinases (JAKs), IFNGRs, and STAT1.31 Similarly, IFN-α exerts its functions through binding to IFN-αR1 and IFN-αR2, followed by activation of STAT1, STAT2, and STAT3.32 Activation of STAT1 by IFN-α and IFN-γ has been shown to play an important role in liver injury, inflammation, antiviral effects, antitumor effects, and inhibition of liver regeneration.33 In this investigation, we explored the role of STAT1 in liver fibrosis using STAT1-deficient mice and demonstrated that STAT1 plays an inhibitory role in liver fibrosis through inhibition of HSC activation, attenuation of TGF-β signaling, and stimulation of NK cell cytotoxicity against activated HSCs.

Abbreviations

HSC, hepatic stellate cells; α-SMA, alpha-smooth muscle actin; TGF-β, transforming growth factor-beta; PI3, phosphatidylinositol 3; PDGF, platelet-derived growth factor; PDGFR, PDGF-receptor; CCl4, carbon tetrachloride; poly I:C, polyinosinic-polycytidylic acid; JAK-STATs, Janus kinase-signal transducers and activators of transcription; IFN-γ, interferon-gamma; NK cells, natural killer cells; RT-PCR, reverse-transcription polymerase chain reaction; MNC, mononuclear cells.

Materials and Methods

Materials.

Polyinosinic-polycytidylic acid (poly I:C), Gey's balanced salt solution (GBSS), OptiPrep, and carbon tetrachloride (CCl4) were purchased from Sigma (St. Louis, MO). Collagenase type I and DNase I were obtained from Roche (Indianapolis, IN). IFN-γ was obtained from R&D Systems Inc. (Minneapolis, MN) (the activity of IFN-γ is 8.4 IU/ng)

Animals.

STAT1−/− mice on 129S6/SvEv background were originally purchased from Taconic (Germantown, NY) and backcrossed with C57BL/6J mice for at least 9 generations. C57BL/6J mice were purchased from the Jackson Laboratory (Bar Harbor, ME).

Liver Fibrosis Induced by CCl4.

Hepatic fibrosis in mice was induced by intraperitoneal (i.p.) injection with 2 mL/kg body weight of 10% CCl4 (Sigma) dissolved in olive oil (Sigma), 3 times a week for up to 12 weeks.

Co-treatment of Mice With CCl4 Plus Poly I:C or CCl4 Plus IFN-γ.

Mice were injected with CCl4 (10% in olive oil, 2 mL/kg body weight, 3 times a week) and poly I:C (5 μg/g body weight i.p., 3 times a week, administered 12 hours before CCl4 injection.) or IFN-γ (50,000 IU/mouse of recombinant murine IFN-γ, subcutaneous injection, 7 times a week) for 3 weeks. Control mice received CCl4 plus saline injection. For acute poly I:C treatment, mice were injected (i.p.) with poly I:C once. After 16 hours, mononuclear cells (MNCs) were isolated and used for in vitro analyses.

Serum Biochemical Measurements, Histology, Immunohistochemistry, TUNEL Assay, Determination of Hepatic Hydroxyproline Content, Cytotoxicity Assay, and Cell Proliferation.

Details about these methods are described in the Supplementary Materials and Methods (available at: http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html).

Flow Cytometric Analysis of NK Cells, TRAIL, and NKG2D Expression.

TRAIL and NKG2D expression on NK cells were determined using anti-NK1.1, anti-CD3 (BD Pharmingen, San Diego, CA), anti-TRAIL, and anti-NKG2D antibodies (eBioscience, San Diego, CA) and measured by FACSCalibur fluorescence-activated cell sorting (FACS) equipment (BD Biosciences, Mountain View, CA).

Reverse Transcription-Polymerase Chain Reaction (RT-PCR).

The RT-PCR was carried out using primers as described.17

Isolation and Culture of Hepatic Stellate Cells.

Mouse liver HSCs were isolated by in situ collagenase perfusion and differential centrifugation on OptiPrep (Sigma) density gradients, as described with some modifications.34 The details about this method are described in the Supplementary Materials and Methods.

Western Blot.

Liver tissues were homogenized in lysis buffer (30 mmol/L Tris, pH 7.5, 150 mmol/L sodium chloride, 1 mmol/L phenylmethylsulfonyl fluoride, 1 mmol/L sodium orthovanadate, 1% Nonidet P-40, 10% glycerol, and phosphatase and protease inhibitors). Western blot analyses were performed with 80 μg protein from liver homogenates using anti-phospho-Smad3 (1:1000 dilution, Cell Signaling Technology) and anti-Smad 7 antibodies (1:1000 dilution, Santa Cruz Biotechnology).

Statistical Analysis.

Data are expressed as means ± SEM. To compare values obtained from 2 groups, the Student t test was performed. A value of P < .05 was considered significant.

Results

CCl4-Induced Liver Fibrosis and HSC Proliferation Are Accelerated, While HSC Apoptosis Is Reduced in STAT1−/− Mice.

To investigate the role of STAT1 in liver fibrosis, CCl4 was injected into wild-type and STAT1−/− mice. After CCl4 injection, there were no differences in alanine aminotransferase or aspartate aminotransferase levels for both groups (data not shown); however, α-SMA staining (Fig. 1A) and the hydroxyproline assay (Fig. 1B) revealed that livers from STAT1−/− mice after CCl4 injection possessed higher grades of fibrosis than wild-type mice at week 1 to 3, whereas similar grades of liver fibrosis between the 2 groups were observed at week 6 to 12, suggesting that STAT1 is involved in the inhibition of liver fibrosis at early times. Moreover, Fig. 1C shows that serum levels of IFN-γ, a major activator for STAT1, were elevated in wild-type mice after CCl4 injection with peak effect occurring 1 week after injection, with the higher levels sustained for up to 12 weeks. Serum IFN-γ levels were also elevated in STAT1−/− mice after CCl4 injection, but much lower compared with wild-type mice at week 1 and 2.

Figure 1.

Disruption of the STAT1 gene accelerates liver fibrosis and HSC proliferation, but reduces HSC apoptosis. STAT1−/− mice and wild-type controls were injected with CCl4 as described in “Materials and Methods.” (A) α-SMA positive HSCs were detected by immunohistochemistry (original magnification ×60). (B-C) Hepatic hydroxyproline contents were determined colorimetrically (B) and serum IFN-γ levels were measured (C). (D-E) Liver sections were double stained with Ki67 and α-SMA antibodies (D) or with TUNEL and α-SMA antibodies (E). Ki67+ α-SMA+ or TUNEL+α-SMA+ cell numbers were counted from 4 different fields (×200 magnification) and the number/field is shown. Values in (B) through (E) are means ± SEM from 4–5 mice in each group, *P < .05, **P < .01 compared with corresponding wild-type controls.

To explore the mechanism underlying accelerated liver fibrosis in STAT1−/− mice, we used double immunostaining to examine HSC proliferation and apoptosis after CCl4 injection. The number of Ki67+ α-SMA+ cells (proliferating HSCs) in STAT1−/− mice was greater than from wild-type controls 1 week after CCl4 treatment, but was lower than corresponding wild-type controls 6 weeks after CCl4 treatment (Fig. 1D). In contrast, TUNEL assay results showed that the number of apoptotic HSCs (TUNEL+ α-SMA+ cells) was lower in STAT1−/− mice compared with wild-type mice 3 weeks after CCl4 injection (Fig. 1E).

IFN-γ Induces HSC Cell Cycle Arrest and Apoptosis Via a STAT1-Dependent Mechanism.

Treatment with IFN-γ activated STAT1 and STAT3 in wild-type HSCs, as evidenced by positive staining for pSTAT1 and pSTAT3 in the nuclei (Fig. 2A). As expected, STAT1 activation was not detected in STAT1−/− HSCs, while significant STAT3 activation was observed in these cells. Treatment with IFN-γ significantly suppressed cell proliferation of wild-type HSCs; however, the same treatment did not inhibit, but rather slightly stimulated STAT1−/− HSC proliferation (Fig. 2B). Furthermore, in vitro treatment with IFN-γ inhibited activation of wild-type HSCs as detected by α-SMA staining (Fig. 3C) and induced apoptosis of wild-type HSCs (Fig. 2D). In contrast, the same treatment affected neither cell activation (α-SMA expression) (Fig. 2C) nor cell viability of STAT1−/− HSCs (Fig. 2D).

Figure 2.

IFN-γ inhibits HSC proliferation and induces HSC apoptosis via a STAT1-dependent mechanism. HSCs from wild-type and STAT1−/− mice were cultured in serum-containing medium for 3 days, then cultured in serum-free medium for 24 hours. (A) Serum-starved HSCs were stimulated with IFN-γ for 30 minutes, followed by fixation and staining with anti-pSTAT1 and anti-pSTAT3 antibodies. The arrows indicate representative pSTAT1 or pSTAT3 staining in the nuclei. (B) Serum-starved HSCs were stimulated with IFN-γ for 24 hours, and then [3 H]thymidine was added and cultured for an additional 24 hours, followed by measuring [3 H]thymidine uptake. (C-D) Serum-starved HSCs were stimulated with IFN-γ for 24 hours, followed by immunostaining with anti-α-SMA antibodies (C), or with TUNEL for apoptosis (D). Values in (B) and (D) are shown as means ± SEM from 3 independent experiments. *P < .01; **P < .001 compared with corresponding controls without IFN-γ treatment.

Figure 3.

Disruption of the STAT1 gene accelerates HSC proliferation and activation induced by serum. (A) The same number of HSCs from wild-type and STAT1−/− mice was cultured in 6-well uncoated plastic plates in serum-containing medium for up to 7 days. HSCs on day 7 were stained with the anti-α-SMA antibody. Representative pictures are shown. On day 1, wild-type and STAT1−/− HSCs showed normal stellate shape with many lipid droplets and little proliferation. On day 4–7, STAT1−/− HSCs grew faster compared with wild-type HSCs. (B) Total RNA was isolated from cultured wild-type (WT) and STAT1−/− (ST1−/−) HSCs, and subjected to RT-PCR analyses with α-SMA and β-actin primers. (C) After 3 days of culture, HSCs were cultured in serum-free medium for 24 hours, and then [3 H]thymidine was added for an additional 24 hours, followed by measurement of [3 H]thymidine uptake. (D) After 1 and 4 days of culture, the number of HSCs per well was counted. Values in (C) and (D) are shown as means ± SEM from 3 independent experiments. **P < .01 compared with corresponding wild-type controls.

Disruption of the STAT1 Gene Accelerates HSC Activation and Proliferation Induced by Serum.

Figure 3A shows that STAT1−/− HSCs cultured in serum-containing medium grew faster compared with wild-type HSCs. On day 7, there were no differences in size, but more HSCs were observed in the STAT1−/− groups compared with wild-type HSCs (Fig. 3A). Immunocytochemical analyses revealed that similar α-SMA staining was detected in wild-type and STAT1−/− HSCs cultured for 7 days (Fig. 3A). RT-PCR analyses show that α-SMA expression was much greater in STAT1−/− HSCs than wild-type HSCs on postculture day 4 (Fig. 3B), suggesting that STAT1−/− HSCs are activated faster than wild-type HSCs.

Moreover, [3 H]thymidine uptake and cell counting further confirmed that cell proliferation was enhanced in STAT1−/− HSCs compared with wild-type HSCs. The [3 H]thymidine uptake on day 4 was ∼2.5-fold greater in STAT1−/− HSCs compared with wild-type HSCs (Fig. 3C). Cell count showed that the number of STAT1−/− HSCs on day 4 was significantly greater than wild-type HSCs (Fig. 3D).

Disruption of the STAT1 Gene Enhances PDGFR-β Expression and Accelerates PDGF-BB Stimulation of HSC Proliferation.

Cell counting revealed that the number of wild-type HSCs increased on day 2, peaked on day 4, and declined on day 7, and the number of STAT1−/− HSCs was much greater than that of wild-type HSCs at the same time after culture in medium containing PDGF (Fig. 4A). Furthermore, [3 H]thymidine uptake studies showed that PDGF-BB treatment induced much greater cell proliferation in STAT1−/− HSCs than in wild-type HSCs (Fig. 4B). Next, we examined the expression level of PDGFR-β mRNA in wild-type and STAT1−/− HSCs. Levels of PDGFR-β mRNA expression were upregulated in wild-type cells until postculture day 4 (Fig. 4C). In contrast, significant induction of PDGFR-β mRNA in STAT1−/− HSCs was observed 2 days after culture, and levels of PDGFR-β mRNA in STAT1−/− HSCs were still higher than in wild-type cells on day 4. On day 7, similar levels of PDGFR-β mRNA were detected in wild-type and STAT1−/− HSCs. Similarly, TGF-β mRNA expression was also enhanced in STAT1−/− HSCs compared to wild-type HSCs at postculture day 1 to day 4.

Figure 4.

Disruption of the STAT1 gene enhances PDGF-BB stimulation of HSC proliferation and PDGFR-β expression on HSCs. (A) HSCs were cultured in 6-well plates in serum-free medium containing PDGF-BB (20 ng/mL) for up to 7 days, and cell number was counted in each well. (B) HSCs were cultured in serum-free medium containing various concentrations of PDGF-BB for 4 days, after which [3H]thymidine was added, cultured for an additional 24 hours, and [3H]thymidine uptake was then measured. Values in (A) and (B) are shown as means ± SEM from 3 independent experiments. *P < .05; **P < .01 compared with the corresponding controls. (C) Wild-type and STAT1−/− HSCs were cultured in serum-containing medium for various time periods, followed by RT-PCR analyses for PDGFR-β, TGF-β, and β-actin expression.

STAT1 Plays an Important Role in Poly I:C Suppression of Liver Fibrosis and Poly I:C Activation of NK Killing of HSCs.

Results from α-SMA staining (Fig. 5A) show that poly I:C treatment inhibited liver fibrosis in wild-type mice, but not in STAT1−/− mice, suggesting that poly I:C inhibition of liver fibrosis is STAT1-dependent. Liver MNCs from poly I:C-treated wild-type mice had approximately 9% and 35% cytotoxicity against HSCs cultured for 1 and 4 days, respectively (Fig. 5B), suggesting that liver MNCs kill activated HSCs (4-day culture), but not quiescent HSCs (1-day culture), which is consistent with our previous report.17 Liver MNCs from poly I:C-treated STAT1−/− mice demonstrated significantly lower cytotoxicity compared with wild-type mice. Basal levels of NKG2D and IFN-γ in STAT1−/− liver MNCs were lower than that from wild-type liver MNCs (Fig. 5C). Treatment with poly I:C induced expression of NKG2D, TRAIL, perforin, Fas L, and IFN-γ in wild-type mouse liver MNCs. Such induction was diminished in STAT1−/− mouse liver MNCs. Analysis using FACS revealed that about 8% and 2% of NK cells (NK1.1+CD3) were detectable in liver MNCs from wild-type mice and STAT1−/− mice, respectively (Fig. 5D). Treatment with poly I:C induced marked accumulation of NK cells in the livers of wild-type mice, but not in STAT1−/− mice (Fig. 5D). Furthermore, expression of NKG2D and TRAIL was detected at high basal levels in wild-type mouse NK1.1+ cells, which increased significantly after poly I:C injection (NK1.1+NKG2D+ cells: 11.0% in WT saline versus 17.9% in WT poly I:C; NK1.1+TRAIL+ cells: 11.5% WT saline versus 20.9% in WT poly I:C). Such expression was detected at very low basal levels on STAT1−/− mouse NK1.1+ cells, and was not induced after poly I:C treatment (NK1.1+NKG2D+ cells: 2.6% in STAT1−/− saline versus 3.1% in STAT1−/− poly I:C; NK1.1+TRAIL+ cells: 2.2% STAT1−/− saline versus 2.9% in STAT1−/− poly I:C) (Fig. 5E).

Figure 5.

STAT1 plays an important role in poly I:C suppression of liver fibrosis and poly I:C activation of NK cell cytotoxicity against activated HSCs. (A) Mice were injected with CCl4 and with or without poly I:C injection for 3 weeks as described in “Materials and Methods”. Liver tissues were stained with anti-α-SMA antibody for activated HSCs (original magnification, ×40). (B-E) Wild-type and STAT1−/− mice were injected (i.p.) with poly I:C (5 μg/g body weight) and killed 16 hours later. Liver MNCs were then isolated and used in cytotoxicity assays (B), RT-PCR analyses using various primers (C), and FACS analyses using NK1.1, CD3, NKG2D, and TRAIL antibodies (D) and (E). In (B), liver MNCs were used as effector cells. Quiescent HSCs (1-day culture) and activated HSCs (4-day culture) were used as target cells and cytotoxicity was measured (target cells/effector cells = 1/50). Values in (B) are means ± SEM from 3 independent experiments, **P < .01 compared with corresponding wild-type controls.

Antifibrotic Effect of IFN-γ In Vivo Is STAT1-Dependent.

Staining of α-SMA (Fig. 6A) shows that chronic treatment with IFN-γ attenuated CCl4-induced fibrosis in wild-type mice, but not in STAT1−/− mice, suggesting that IFN-γ inhibits liver fibrosis via a STAT1-dependent mechanism. In wild-type mice, chronic CCl4 treatment (CCl4+saline group) for 3 weeks decreased hepatic NKT cells (NK1.1+CD3+) but increased NK cells (NK1.1+CD3) (Fig. 6B). The latter was further increased after chronic IFN-γ treatment. In STAT1−/− mice, CCl4 treatment decreased hepatic NKT cells, but slightly increased NK cells. The latter was only slightly increased after chronic IFN-γ treatment. IFN-γ-induced accumulation of hepatic NK cells was much lower in STAT1−/− mice compared with wild-type mice (CCl4+IFN-γ group: 7.3% NK cells in STAT1−/− mice versus 26.4% in wild-type mice). Chronic IFN-γ treatment also increased expression of NKG2D, TRAIL, perforin, Fas L in hepatic MNCs from wild-type mice, but not from STAT1−/− mice (Fig. 6C).

Figure 6.

Chronic IFN-γ treatment inhibits liver fibrosis via a STAT1-dependent mechanism. (A) Mice were injected with CCl4 with or without IFN-γ injection for 3 weeks as described in “Materials and Methods.” Liver tissues were stained with anti-α-SMA antibody for activated HSCs (original magnification, ×40). (B-C) MNCs were isolated from CCl4+saline-treated and CCl4+IFN-γ-treated wild-type and STAT−/− mouse livers, and subjected to FACS analyses using NK1.1 and CD3 antibodies (B) or subjected to RT-PCR analyses using various primers as indicated (C).

Activation of TGF-β/Smad3 Signaling Pathway Is Accelerated in STAT1−/− Mice During Liver Fibrosis.

Western blot analyses (Fig. 7A) show that p-Smad3 expression was induced in wild-type mice with peak effect occurring 6 weeks after CCl4 injection, whereas such peak induction in STAT1−/− mice occurred at 3 weeks, suggesting that activation of the TGF-β/Smad3 signaling pathway is accelerated in STAT1−/− mice during liver injury. Significant induction of Smad7 was observed in the livers of wild-type mice 1 week after CCl4 injection, but not detected in STAT1−/− mice (Fig. 7A). Immunohistochemical analyses revealed that the number of pSmad3+α-SMA+ HSCs was higher in STAT1−/− mice compared with wild-type mice 1 week after CCl4 injection, whereas a similar number of pSmad3+α-SMA+ HSCs was observed between those 2 groups 6 weeks after CCl4 injection (Fig. 7B).

Figure 7.

Activation of TGF-β/Smad3 signaling is enhanced in STAT1−/− mice during liver fibrosis. (A) Western blot analyses of liver tissues from wild-type and STAT1−/− mice treated with CCl4 for up to 8 weeks. (B-C) Liver tissues from 1-week-treated or 6-week-treated wild-type and STAT1−/− mice were double stained with p-Smad3 and α-SMA antibodies. Representative photomicrographs are shown (original magnification, ×40) in (B). Red arrows indicate pSmad3+ α-SMA+ cells, black arrows indicate pSmad3+ hepatocytes. The pSmad3+ α-SMA+ cells were counted and are shown in (C). Values are means ± SEM from 5 mice per group, **P < .01 compared with corresponding wild-type controls.

Discussion

STAT1 is a key signaling component of IFNs and plays an important role in antiviral and antitumor defense, induction of hepatic inflammation and injury, and inhibition of liver regeneration.33, 35 In this report, we provide evidence suggesting that STAT1 also negatively regulates liver fibrosis via multiple mechanisms, including: (1) inhibition of HSC proliferation; (2) suppression of PDGFR-β expression; (3) inhibition of TGF-β/Smad3 signaling; and (4) stimulation of NK cell cytotoxicity against activated HSCs.

Activation of STAT1 Inhibits HSC Proliferation, Thereby Attenuating Liver Fibrosis.

Although the antifibrotic effect of IFN-γ has been well documented,16–19, 28, 29 the downstream signals contributing to such effects remain obscure. In this study, we show that IFN-γ activates both STAT1 and STAT3 in cultured HSCs (Fig. 2) and provide in vivo and in vitro evidence suggesting that STAT1 is a negative regulator of HSC proliferation and liver fibrosis. By using an in vivo model of CCl4-induced liver fibrosis, we demonstrated that disruption of the STAT1 gene accelerated liver fibrosis and HSC proliferation (Fig. 1), and abolished the antifibrotic effect of poly I:C/IFN-γ (Fig. 5 and Fig. 6). Then we further explored the role of STAT1 in HSC proliferation in vitro. Our results show that compared with wild-type HSCs, STAT1−/− HSCs had accelerated cell proliferation induced by either serum or PDGF (Fig. 3 and Fig. 4) and were resistant to IFN-γ-induced cell cycle arrest and apoptosis (Fig. 2). Although the in vitro evidence clearly shows that STAT1 plays a key role in inducing HSC cell cycle arrest and apoptosis, disruption of the STAT1 gene only accelerated liver fibrosis at the earlier times, but not at later times (Fig. 1). The reasons why liver fibrosis was not enhanced at later times in STAT1−/− mice are not clear, but we can speculate on several possibilities. Firstly, serum levels of IFN-γ were elevated in wild-type mice after CCl4 injection with peak effect occurring between weeks 1 and 3 after injection (Fig. 1C), suggesting that IFN-γ may play a more important role in inhibiting liver fibrosis via activation of STAT1 at earlier times than at later times. Moreover, serum levels of IFN-γ were lower in STAT1−/− mice after CCl4 injection compared with wild-type mice (Fig. 1C), which correlates well with the accelerated liver fibrosis observed at the earlier times. Secondly, we previously reported that liver regeneration is accelerated in STAT1−/− mice in the partial hepatectomy model.35 Our unpublished data show that STAT1−/− mouse livers had higher percentages of Ki67+ hepatocytes than wild-type mouse livers after CCl4 treatment (Jeong W, Gao B, unpublished data). This suggests that STAT1−/− mouse livers regenerate faster than wild-type mice during CCl4-induced chronic liver injury, which may account partly for slow progression of liver fibrosis in STAT1−/− mice at the later times because increased hepatocyte proliferation correlated negatively with progression of liver fibrosis in animal models and in human liver cirrhosis.36, 37

STAT1 can be activated by many cytokines, including IFN-α/β and IFN-γ, which are the 2 major cytokines responsible for STAT1 activation in the liver.32, 33 Both IFN-α/β and IFN-γ have been shown to activate STAT1 in HSCs and inhibit liver fibrosis.16–29 Findings given here suggest that STAT1 is a key signaling component responsible for the antifibrotic effect of IFN-γ, therefore, it is plausible that STAT1 may also be responsible for IFN-α/β amelioration of liver fibrosis. Moreover, PDGF has been identified as the most potent growth factor to stimulate HSC proliferation and has been shown to activate multiple signaling pathways including STAT1, STAT3, MAP kinase, and PI3 kinase.7, 11 Disruption of the STAT1 gene enhanced HSC proliferation induced by PDGF (Fig. 4), suggesting that STAT1 is also a negative regulator for the proliferative effect of PDGF on HSCs. In addition to STAT1, both IFN-α/β and IFN-γ also activate other signaling pathways including STAT3, MAP kinase, and PI3 kinase, etc. In general, IFNs induce weak activation of MAP kinase and PI3 kinase, but strongly induce activation of STAT1 and STAT3 in a variety of cell types, including HSCs (Fig. 2A). Treatment with IFN-γ inhibited wild-type HSC proliferation; however, the same treatment did not inhibit, but rather slightly stimulated STAT1−/− HSC proliferation (Fig. 2B). These findings suggest that STAT1 plays a key role in IFN-γ inhibition of HSC proliferation; other signaling pathways (such as STAT3) may be responsible for IFN-γ stimulation of cell proliferation in STAT1−/− HSCs because activation of STAT3 has been implicated in stimulating cell proliferation in several cell types.38 Further studies are required to clarify the role of STAT3 in HSC proliferation and liver fibrosis.

PDGFR-β Expression Is Enhanced in STAT1−/− HSCs.

It has been well documented that the antiproliferative effects of STAT1 are mediated via induction of several antiproliferative genes including p21cip1 and IRF-1,39, 40 which are also likely to contribute to STAT1-mediated inhibition of HSC proliferation. In this report, we demonstrated that PDGFR-β induction was accelerated on STAT1−/− HSCs during activation (Fig. 4), which may be another important mechanism contributing to the antiproliferative effect of STAT1 on HSC proliferation. It has been shown that PDGF is the most potent stimulator of HSC proliferation through binding PDGFR, including PDGFR-α and PDGFR-β,7, 9–14 Expression of PDGFR-β, which is very low in quiescent HSCs and significantly induced in activated HSCs, was shown to be correlated with HSC proliferation in vivo and in vitro.9, 11–13 In contrast, PDGFR-α mRNA is expressed in quiescent HSCs and unchanged after liver injury.13 Here, we show that PDGFR-β expression was significantly higher on STAT1−/− HSCs compared with wild-type HSCs during activation (Fig. 4C), which likely contributes to increased proliferation induced by PDGF in STAT1−/− HSCs (Fig. 4A-B). At present, the mechanisms underlying upregulation of PDGFR-β expression on STAT1−/− HSCs remain unclear. It has been shown that PDGFR-β induction is mediated via a TGF-β autocrine loop in HSCs,10 and TGF-β expression is enhanced in STAT1−/− HSCs (Fig. 4C). Therefore, upregulation of PDGFR-β expression on STAT1−/− HSCs could be due to augmentation of the autocrine TGF-β loop on these HSCs.

Activation of TGF-β/Smad3 Signaling Pathway Is Accelerated in STAT1−/− Mice During Liver Fibrosis.

It has been shown that TGF-β is a major activator of HSCs by activating several signaling pathways including Smad3,30 and that IFN-γ inhibits the TGF-β signaling pathway via induction of Smad7 and YB-1 in a Jak1/STAT1-dependent manner.30, 41, 42 However, the effect of STAT1 on TGF-β signaling in vivo has not been reported. In this study, we demonstrated that TGF-β-mediated pSmad3 activation is significantly accelerated in the livers of STAT1−/− mice compared with wild-type mice following CCl4 administration (Fig. 7A). Expression of Smad7 protein, an inhibitor of TGF-β signaling, was induced in wild-type mice with peak effect occurring 1 week after CCl4 injection, which correlated well with peak elevation of serum IFN-γ levels observed at the same time (Fig. 1C). Induction of Smad7 protein was not observed in STAT1−/− mice. Moreover, double immunohistochemical analyses revealed that pSmad3 activation in HSCs was significantly higher in STAT1−/− mice than wild-type mice 1 week after CCl4 injection (Fig. 7B). Taken together, our findings suggest that STAT1 is a negative regulator for TGF-β signaling in vivo via induction of Smad7 during liver fibrosis.

STAT1 Plays an Important Role in the Cytotoxicity of NK Cells Against Activated HSCs.

Recently, we and others demonstrated that NK cells are able to kill activated HSCs but not quiescent HSCs, thereby inhibiting liver fibrosis.17, 43, 44 Clinical data also suggest that NK cells are implicated in negative regulation of liver fibrosis in patients with chronic HCV infection.45 The mechanisms underlying NK cell killing of activated HSCs but not quiescent HSCs are multiple, including the observation that compared to quiescent HSCs, activated HSCs expressed increased levels of NK cell activating ligand RAE-117 and TRAIL receptor-246 but decreased levels of NK cell inhibiting ligand MHC class-I molecule.43 In this report, we provide evidence suggesting that reduced NK cell cytotoxicity against HSCs may also contribute to accelerated liver fibrosis in STAT1−/− mice compared with wild-type mice. Firstly, STAT1−/− mouse livers contain far fewer NK cells (about 2% to 3%) compared with wild-type mice (8% to 10%), and poly I:C or IFN-γ treatment markedly induced elevation of hepatic NK cells in wild-type mice, but not in STAT1−/− mice (Fig. 5D and Fig. 6B). Secondly, liver MNCs from wild-type mice treated with poly I:C demonstrated approximately 35% cytotoxicity against HSCs, which is likely STAT1-dependent, because disruption of the STAT1 gene diminished poly I:C-induced NK cell cytotoxicity (Fig. 5B). Thirdly, poly I:C or IFN-γ treatment induced NKG2D/TRAIL in the MNCs of wild-type mice, but not in MNCs from STAT1−/− mice. Finally, less HSC apoptosis was observed in the liver of STAT1−/− mice compared with wild-type mice during CCl4-induced liver fibrosis (Fig. 1E). Taken together, these findings suggest that STAT1 plays an important role in stimulation of NK cell killing of activated HSCs, which is another important mechanism responsible for STAT1 inhibition of liver fibrosis.

In conclusion, STAT1 is an important negative regulator for liver fibrosis. We35, 47 and others48–50 demonstrated recently that STAT1 protein is highly induced and activated in the liver of patients with chronic HCV infection. Therefore, STAT1 activation may have a role in regulating the development and progression of liver fibrosis. Indeed, Smith et al.49 recently reported that STAT1 and its downstream genes were upregulated in HCV patients after liver transplantation. Such induction was much lower in patients who developed early fibrosis than those who did not develop fibrosis, suggesting that upregulation of STAT1 and its downstream genes may negatively regulate liver fibrosis. Hence, STAT1 could be a new therapeutic target for the treatment of liver fibrosis.

Ancillary