STAT1 contributes to dsRNA inhibition of liver regeneration after partial hepatectomy in mice

Authors

  • Rui Sun,

    1. Section on Liver Biology, Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD 20892
    2. School of Life Sciences, University of Science and Technology of China, Hefei, Anhui 230027, China
    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 20892
    Search for more papers by this author
  • Norio Horiguchi,

    1. Section on Liver Biology, Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD 20892
    Search for more papers by this author
  • Shailin Kulkarni,

    1. Section on Liver Biology, Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD 20892
    Search for more papers by this author
  • 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 20892
    Search for more papers by this author
  • Hao-Yu Sun,

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

  • Potential conflict of interest: Nothing to report.

Abstract

Increasing evidence suggests that liver regeneration is suppressed in patients with chronic HCV infection; however, the underlying mechanisms remain unclear. Previously, we demonstrated that injection of the synthetic double-stranded RNA (dsRNA) poly I:C to mimic viral infection suppresses liver regeneration in the partial hepatectomy (PHx) model, whereby IFN-γ contributes to the inhibition. In this study, we examined the role of the IFN-γ-activated downstream signal (STAT1) and genes (IRF-1, p21cip1, and SOCS1) in liver regeneration and hepatocyte proliferation. Results show that disruption of the STAT1 gene abolished poly I:C suppression of liver regeneration and the inhibitory effect of poly I:C on liver regeneration was diminished in IRF-1−/− and p21cip1-/-mice. Treatment with IFN-γ in vitro inhibited cell proliferation of wild-type mouse hepatocytes, but not STAT1−/− hepatocytes. The inhibitory effect of IFN-γ on cell proliferation was also diminished in IRF-1−/− and p21cip1−/− hepatocytes, but enhanced in SOCS1−/− hepatocytes. Hepatocyte proliferation was unaffected by treatment with poly I:C alone, but when hepatocytes were co-cultured with liver lymphocytes, proliferation was inhibited by IFN-γ/STAT1-dependent mechanisms. Moreover, in HCV-infected livers with cirrhosis, activation of STAT1 was detected and correlated positively with liver injury (elevated serum levels of AST) but negatively with hepatocyte proliferation (hepatocyte PCNA and Ki-67 positive immunostaining). In conclusion, STAT1 is involved in dsRNA suppression of liver regeneration; not only does STAT1 activation contribute to liver injury, it may also block liver repair through inhibition of hepatocyte proliferation in HCV-infected patients, playing an important role in the pathogenesis of disease. (HEPATOLOGY 2006;44:955–966.)

The liver has the unique ability to regenerate after injury or loss of tissue. Liver regeneration is controlled by a wide array of signaling factors and plays a key role in recovery after acute and chronic liver injury1–5; however, emerging evidence suggests that liver regeneration is suppressed in patients with acute and chronic hepatitis C virus (HCV) infection6–12 or liver cirrhosis.13, 14 Moreover, infection with murine hepatitis virus15, 16 or cytomegalovirus (MCMV)17 has been shown to suppress liver regeneration in mice. At present, the cellular and molecular mechanisms impairing liver regeneration after viral infection remain obscure. Previously, we showed that injection of the synthetic double-stranded RNA (dsRNA) polyinosinic-polycytidylic acid (poly I:C) that mimics viral effects, inhibits liver regeneration in the murine partial hepatectomy (PHx) model.17 Further studies also suggest that poly I:C activates natural killer (NK) cells to produce interferon-γ (IFN-γ), leading to inhibition of liver regeneration.17 However, the downstream signals and genes activated by IFN-γ in poly I:C negative regulation of liver regeneration are still unclear.

Signaling mediated by IFN-γ is initiated upon its binding to the corresponding receptors IFNGR1 and IFNGR2, followed by activation of IFNGR-associated tyrosine kinases (JAK1 and JAK2) and subsequent activation of the cytoplasmic protein, signal transducer and activator of transcription 1 (STAT1). Tyrosine-phosphorylated STAT1 molecules then form dimers and translocate into the nucleus, inducing expression of a variety of genes, including members of the suppressors of cytokine signaling (SOCS) family of proteins, resulting in the termination of IFN-γ signaling.18 In the liver, activation of STAT1 by IFN-γ plays an important role in the development of hepatic inflammation and liver injury through induction of chemokines, adhesive molecules, and proapoptotic proteins;19 however, the role of STAT1 in liver regeneration remains unknown. Because we already demonstrated that IFN-γ contributes to poly I:C inhibition of liver regeneration,17 the regulatory effect of IFN-γ-activated STAT1 and its downstream genes in liver regeneration were examined in this investigation. Our results suggest that activation of STAT1 suppresses liver regeneration and hepatocyte proliferation in mice. Activation of STAT1 signaling and its positive correlation with liver injury and inverse correlation with hepatocyte proliferation in patients with chronic HCV infection are also demonstrated here.

Abbreviations

HCV, hepatitis C virus; dsRNA, double-stranded RNA; PHx, partial hepatectomy; poly I:C, polyinosinic-polycytidylic acid; JAK-STATs, Janus kinase-signal transducers and activators of transcription; NK cells, natural killer cells; NKT cells, natural killer T cells; IFN, interferon; TLR, toll-like receptor; AST, aspartate aminotransferase; BrdU, bromodeoxyuridine; MNC, mononuclear lymphocytes; LTPDS, Liver Tissue Procurement Distribution System; IRF-1, interferon responsible factor-1; PCNA, proliferating cell nuclear antigen.

Materials and Methods

Materials.

Polyinosinic:cytidylic acid (poly I:C) was purchased from Sigma-Aldrich Corp. (St. Louis, MO). Recombinant murine IFN-γ was obtained from Biosource International (Camarillo, CA). Anti-STAT1 and anti-phospho-STAT1 (Tyr701) antibodies were purchased from Cell Signaling Technology (Beverly, MA).

Mice.

Eight- to 10-week old male or female STAT1−/− mice (129/SvEv background) and control mice were purchased from Taconic Farms (Germantown, NY). IRF-1−/− (C57BL/6 background), IL-6−/− (C57BL/6 background), C57BL/6J control mice, p21cip1–/– mice (B6129SF2 background), and B6129SF2/J, were purchased from the Jackson Laboratory (Bar Harbor, ME). IFN-γ−/−SOCS1−/− mice were kindly provided by Dr. James Ihle, St. Jude Children's Research Hospital (Memphis, TN).

Partial Hepatectomy (PHx)+ Poly I:C Injection Model.

PHx and determination of liver regeneration by the bromodeoxyuridine (BrdU) incorporation assay were performed as described.17 For the PHx+poly I:C model, poly I:C was injected (iv) immediately after PHx. Numbers of BrdU-labeled hepatocytes were determined by counting positively stained hepatocyte nuclei in 3 to 6 low-power (10×) microscope fields and the mean calculated.

Primary Mouse Hepatocyte Culture and [3H]Thymidine Uptake Assay.

The isolation and culture of mouse hepatocytes were performed as described.20 Cell proliferation was determined by the [3 H]Thymidine uptake assay as described.20

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

RT-PCR analyses were performed as described.21 Murine SOCS1, SOCS3, and p21cip1 primers were described.20, 21 The IRF-1 primer sequences used were: 5′-GCA AAA CCA AGA GGA AGC TG-3′ (forward), 5′-CGG TGA CAG TGC TGG AGT TA-3′ (reverse). The expected PCR band size was 161 base pairs. The β-actin gene was amplified and used as an internal control. The PCR bands were scanned using the Storm PhosphoImager (Molecular Dynamics, Sunnyvale, CA).

TUNEL Assay.

Apoptosis in cultured cells were detected using in situ cell death detection kits (Roche, Indianapolis, IN) according to the manufacturer's instructions. Nuclei were stained with DAPI.

Aspartate Aminotransferase (AST).

Levels of AST in the supernatant of cultured hepatocytes were measured with a kit from Drew Scientific Ltd. (Barrow-in-Furness, UK).

Isolation of Liver Mononuclear Lymphocytes (MNCs) and Co-Culture With Hepatocytes.

Isolation of liver MNCs was performed as described.17 Mouse hepatocytes were co-cultured with liver MNCs using Transwell culture plate inserts (Millipore Corp., Bedford, MA) whereby cultured cells are separated by a semipermeable membrane. Briefly, hepatocytes were cultured in 24-well rat tail collagen-coated plates for 2 hours, then MNCs were added (MNC/hepatocyte ratio of 5:1) to the top Transwell and cultured for an additional 48 or 72 hours.

HCV-Infected Liver Tissues With Cirrhosis and Normal Healthy Liver Controls.

The Liver Tissue Procurement Distribution System (LTPDS) provided 34 liver specimens from patients with end-stage chronic HCV infection, including 24 samples from the University of Minnesota (Minneapolis, MN) and 10 samples from the Medical College of Virginia (Richmond, VA). Chronic HCV infection was confirmed histopathologically and serologically. Patient histories indicate that specimens were from patients who did not receive IFN-α therapy 2 years prior to collection of the specimens. Liver pathology of specimens showed bridging fibrosis and cirrhosis. LTPDS also provided normal healthy liver specimens obtained from human donor livers not used for transplantation. The LTPDS was funded by NIH Contract #N01-DK-9-2310.

Immunohistochemical Staining, Western Blotting, and Hydrodynamic Gene Delivery.

Details about these methods are described in supplementary materials and methods (http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html).

Statistical Analysis.

Data are expressed as means ± SEM. To compare values obtained from three or more groups, 1-factor analysis of variance (ANOVA) was used, followed by Tukey's post hoc test. To compare values obtained from 2 groups, the Student t test was performed. The correlations between variables were assessed by the Spearman rank order test. Statistical significance was taken at the P < .05 level.

Results

Poly I:C-Mediated Suppression of Liver Regeneration Is Diminished in STAT1−/− Mice, Which Can Be Restored by STAT1 Overexpression.

Previously, we showed that STAT1 activation was detected at low levels after PHx alone, but after PHx+poly I:C treatment, STAT1 was highly activated.17 The observed activation was determined to be IFN-γ-dependent, as STAT1 activation was minimal in IFN-γ−/− mice compared with wild-type mice (Fig. 1A). The activation of STAT3 after PHx+poly I:C was similar between these 2 groups (Fig. 1A). The role of STAT1 in liver regeneration was determined using male STAT1−/− mice. Results shown in Fig. 1B-C show that peak BrdU incorporation occurred 32 hours after PHx in the majority of male STAT1−/− mice, while peak BrdU incorporation was detected 40 hours after surgery in wild-type controls. A small percentage of male STAT1−/− mice had peak BrdU incorporation 40 hours after PHx. Hence, in order to see the true effect of STAT1 on liver regeneration in mice, a larger number of independent experiments were performed. A summary of 10 independent experiments is shown in Fig. 1C, which demonstrates that liver regeneration is accelerated in male STAT1−/− mice. The rate and amount of liver mass restoration were comparable between male STAT1−/− and their control mice (Supplementary Fig. 1). Interestingly, rate of BrdU incorporation after PHx was similar in female STAT1−/− mice and their wild-type controls (Fig. 1D).

Figure 1.

Disruption of the STAT1 gene accelerates liver regeneration and abolishes poly I:C suppression of liver regeneration. (A) Western blot analyses of STAT1 and STAT3 from wild-type and IFN-γ−/− mice treated with PHx+polyI:C (1 μg/g). (B-D) STAT1−/− and their control mice were subjected to PHx, and liver regeneration was determined by BrdU incorporation at various time points. A graphic representation of BrdU staining from male STAT1−/− and wild-type mice is shown in (B) (original magnification: 100×). A graphic representation of the percentage of BrdU+ hepatocytes is shown in male mice (C) and female mice (D). Values in (C) and (D) are shown as means ± SEM from 30 to 40 mice at 32 hours, 36 hours, 40 hours, 48 hours, and 5 to 10 mice at other time points. **P < .01, and ***P < .001 versus the corresponding wild-type groups. (E) STAT1−/− and wild-type mice were subjected to PHx+poly I:C. Liver regeneration was determined 32 and 40 hours later. A graphic representation of the percentage of BrdU+ hepatocytes is shown as means ± SEM from 8 to 12 mice. (F-H). STAT1−/− mice were rapidly injected in the tail vein with pcDNA3 vector DNA or with STAT1 expression vector as described in “Materials and Methods.” Nine hours later, PHx was performed and treated with poly I:C (1 μg/g). The removed livers were subjected to Western blot analysis for STAT1 protein (F). Liver regeneration was measured 40 hours after PHx. A representative photograph of BrdU staining is shown in (G) and a graphic representation of the percentage of BrdU is shown in (H) as means ± SEM from 4 to 6 mice. ***P < .001 vs. the vector DNA+poly I:C group.

Next we examined whether STAT1 contributes to poly I:C- and MCMV-mediated suppression of liver regeneration. As shown in Fig. 1E, injection of poly I:C suppressed liver regeneration by 90% and 95% at 32 hours and 40 hours after PHx in wild-type mice, respectively, but only 50% and 10% in STAT1−/− mice. Similarly, infection with MCMV markedly inhibited BrdU incorporation 40 hours after PHx in wild-type mice, but not in STAT1−/− mice (data not shown). Furthermore, Fig. 1F-H shows that overexpression of STAT1 restored poly I:C inhibition of liver regeneration in STAT1−/− mice. The expression of STAT1 protein in the liver of STAT1−/− mice after hydrodynamic injection of STAT1 cDNA was confirmed in Fig. 1F. In Fig. 1G-H, the results show that treatment with poly I:C did not inhibit BrdU incorporation 40 hours after PHx in STAT1−/− mice injected with vector DNA, but was markedly attenuated in STAT1−/− mice injected with STAT1 cDNA.

IRF-1 and p21cip1 Are Partially Involved in Poly I:C Suppression of Liver Regeneration.

We examined the roles of IRF-1 and p21cip1, downstream genes of STAT1, in liver regeneration. Figure 2A shows that poly I:C injection significantly induced expression of IRF-1 in wild-type mice, but not in STAT1−/− mice. Expression of p21cip1 was also induced in wild-type mice, but only slightly reduced in STAT1−/− mice. Figure 2B reveals that the number of BrdU+ hepatocytes after PHx was similar in IRF-1−/− and wild-type control mice. Treatment with low doses of poly I:C (0.5 and 1 μg/g) completely inhibited liver regeneration in wild-type mice, but was only partially suppressed in IRF-1−/− mice, while treatment with high doses of poly I:C (5 μg/g) completely blunted liver regeneration in both groups (Fig. 2C-D). Consistent with previous reports,22 BrdU incorporation after PHx was much greater in p21cip1–/– mice compared with wild-type mice (Fig. 2E). Similar to IRF-1−/− mice, poly I:C suppression of liver regeneration was also diminished in p21cip1–/– mice compared to wild-type mice (Fig. 2E).

Figure 2.

IRF-1 and p21cip1 contribute to poly I:C suppression of liver regeneration. (A) RT-PCR analyses of hepatic IRF-1 and p21cip1 mRNA from wild-type and STAT1−/− mice treated with poly I:C (1 μg/g iv). (B) IRF-1−/− and wild-type mice were subjected to PHx, and liver regeneration was determined by BrdU immunostaining. (C-E) IRF-1−/− and p21cip1–/– mice and their wild-type control mice were subjected to PHx+poly I:C treatment. Liver regeneration was measured 40 hours later. A representative photograph for BrdU staining is shown (C). A graphic representation of the percentage of BrdU+ hepatocytes is shown in (B), (D), and (E). Values are shown as means ± SEM from 8 to 10 mice. *P < .05, and **P < .01 in comparison with corresponding poly I:C-treated wild-type groups.

Poly I:C Inhibits Proliferation of Hepatocytes Co-cultured With MNCs via IFN-γ/STAT1-Dependent Mechanisms.

Treatment with poly I:C has been shown to inhibit liver regeneration via IFN-γ/STAT1-dependent mechanisms in vivo17 (see above), so we examined whether poly I:C directly inhibits hepatocyte proliferation in vitro. In vitro treatment with poly I:C only marginally affected proliferation of hepatocytes isolated from wild-type, STAT1−/−, IRF-1−/−, and p21cip–/– mice (Fig. 3A), suggesting that poly I:C does not directly inhibit hepatocyte proliferation. However, poly I:C treatment significantly inhibited wild-type hepatocyte proliferation when co-cultured with wild-type liver MNCs, but not IFN-γ−/− liver MNCs. Inhibition was not observed in STAT1−/− mouse hepatocytes (Fig. 3B).

Figure 3.

Effects of poly I:C on hepatocyte proliferation. (A) Primary mouse hepatocytes from various strains of knock-out mice and their control mice were incubated with poly I:C for 48 or 72 hours. (B) Primary mouse hepatocytes from STAT1−/− and their control mice were co-cultured with or without liver MNCs from wild-type mice or IFN-γ−/− mice, and stimulated with poly I:C for 48 hours. In (A) and (B), [3 H]thymidine was added 24 hours before harvesting. Cells were then harvested and [3 H]thymidine incorporation was measured. Values are shown as means SEM from 5 to 7 independent experiments. *P < .05 in comparison with corresponding poly I:C-untreated group.

IFN-γ Induces Cell Cycle Arrest and Apoptosis in Hepatocytes via a STAT1-Dependent Mechanism: Involvement of IRF-1 and p21cip1.

The above findings suggest that poly I:C inhibits liver regeneration in vivo through stimulating immune cells indirectly to produce IFN-γ, which may directly inhibit hepatocyte proliferation. Therefore, the effect of IFN-γ on hepatocyte proliferation and the underlying mechanisms were examined. IFN-γ treatment induced IRF-1 and p21cip1 mRNA expression in wild-type mouse hepatocytes (Fig. 4A), which was markedly attenuated in STAT1−/− and IRF-1−/− hepatocytes, suggesting that IRF-1 and p21cip1 are 2 genes downstream of STAT1 activated by IFN-γ.

Figure 4.

IFN-γ induces cell cycle arrest and apoptosis of hepatocytes via a STAT1-dependent mechanism involving IRF-1 and p21cip1. (A) RT-PCR analyses of IRF-1 and p21cip1 from hepatocytes treated with IFN-γ (5 ng/mL) for various time periods. (B) and (C) Primary mouse hepatocytes from various strains of knockout mice and their control mice were incubated with IFN-γ for 48 hours (B) or 72 hours (C). [3 H]Thymidine was added 24 hours before harvesting. Cells were then harvested and [3 H]thymidine incorporation was measured. Values are shown as means ± SEM from 5 to 7 independent experiments. (D-F) Primary mouse hepatocytes from STAT1−/−, IRF-1−/−, and wild-type mice were treated with IFN-γ for various time periods. Apoptosis was determined by TUNEL and nuclei were stained with DAPI (D). A graphic representation of 3 independent experiments is shown in (E). The levels of AST were also determined from the supernatant of the cultured hepatocytes (F).

The effects of IFN-γ on hepatocyte proliferation are shown in Fig. 4B-C. Treatment with IFN-γ for 48 or 72 hours inhibited hepatocyte proliferation in a dose-dependent manner, with greater suppression observed in groups treated for 72 hours (Fig. 4C) than those for 48 hours (Fig. 4B). As expected, IFN-γ suppression of hepatocyte proliferation was completely abolished in IFNGR−/− mouse hepatocytes. Treatment with a low dose of IFN-γ (0.1 ng/mL) markedly inhibited cell proliferation of wild-type mouse hepatocytes, but enhanced cell proliferation of STAT1−/− mouse hepatocytes. Treatment with higher doses (1-10 ng/mL) inhibited STAT1−/− mouse hepatocytes by 10%, which was much lower than the 40% to 75% inhibition in wild-type mouse hepatocytes. Similarly, IFN-γ-mediated inhibition of cell proliferation was diminished in both IRF-1−/− and p21cip1–/– mouse hepatocytes compared with corresponding wild-type controls.

The effects of IFN-γ on hepatocyte apoptosis are shown in Fig. 4D-E. Treatment with IFN-γ for 2 or 3 days increased the number of apoptotic hepatocytes (TUNEL+) in wild-type hepatocytes, but not in STAT1−/− and IRF-1−/− hepatocytes (Fig. 4D-E). Figure 4F shows that IFN-γ treatment induced wild-type hepatocyte cell death, as evidenced by increased AST levels in the supernatant. Such increase was not observed in STAT1−/− and IRF-1−/− hepatocytes after IFN-γ treatment.

Disruption of the SOCS1 Gene Enhances IFN-γ Activation of STAT1 and Inhibition of Hepatocyte Proliferation.

It has been reported that expression of SOCS1, a gene downstream of STAT1, was induced in the liver after PHx,23 but others reported that SOCS3, not SOCS1, was induced after PHx.24 We confirmed the latter report in this study. SOCS3, not SOCS1, was elevated after PHx and was attenuated in IL-6−/− mice (Fig. 5A). Interestingly, in our poly I:C treatment model, both SOCS1 and SOCS3 were induced with peak effect occurring 3 and 6 hours after treatment with poly I:C, respectively (Fig. 5B). Poly I:C induction of SOCS1 was completely abolished in STAT1−/− mice while induction of SOCS3 remained unchanged, suggesting that poly I:C induction of SOCS1 is STAT1-dependent (Fig. 5B). Next, we examined the role of SOCS1 in hepatocyte proliferation in vitro. As shown in Fig. 5C, IFN-γ induced stronger and prolonged STAT1 activation in SOCS1−/− mouse hepatocytes compared with wild-type cells. Treatment with IFN-γ for 48 or 72 hours inhibited cell proliferation in wild-type mouse hepatocytes in a dose-dependent manner (Fig. 5D). This inhibition was slightly enhanced in SOCS1+/– mouse hepatocytes, but was significantly augmented in SOCS1−/− mouse hepatocytes (Fig. 5D), suggesting that activation of STAT1 correlates inversely with hepatocyte proliferation.

Figure 5.

IFN-γ induces more profound inhibition of hepatocyte proliferation in SOCS1−/− hepatocytes. (A) and (B) RT-PCR analyses of SOCS1 and SOCS3 from IL-6−/− and wild-type mice after PHx (A), or (B) from STAT1−/− and wild-type mice treated with poly I:C (1 μg/g, iv). (C) Western blot analyses of pSTAT1 from primary IFN-γ−/−SOCS1+/+ and IFN-γ−/−SOCS1−/− mouse hepatocytes treated with IFN-γ (5 ng/mL) for various time periods. (D) Primary mouse hepatocytes from IFN-γ−/−SOCS1+/+, IFN-γ−/−SOCS1+/−, and IFN-γ−/−SOCS1−/− mice were incubated with IFN-γ for 48 or 72 hours. [3 H]Thymidine was added 24 hours before harvesting. Cells were then harvested and [3 H]thymidine incorporation was measured. Values are shown as means ± SEM from 5 to 7 independent experiments.

Elevated Levels of STAT1 Protein Correlate Positively With Liver Injury But Negatively With Hepatocyte Proliferation in HCV-Infected Livers With Cirrhosis.

We next examined whether STAT1 is also involved in regulating hepatocyte proliferation in human HCV-infected livers with cirrhosis. Western blot analyses revealed that STAT1 protein levels in HCV-infected livers with cirrhosis were greater compared with normal healthy livers (Fig. 6A). In contrast, expression of STAT3 protein was detected at similar levels in all HCV-infected livers with cirrhosis (data not shown). The levels of hepatic STAT1 protein show a strong positive correlation with serum AST levels in these HCV-infected specimens (P = .0013), suggesting that STAT1 is involved in liver injury in HCV-infected livers with cirrhosis.

Figure 6.

Elevated levels of STAT1 protein correlate positively with liver injury, but inversely with hepatocyte proliferation in HCV-infected livers with cirrhosis. (A) Liver samples from 34 HCV-infected patients were subjected to Western blot analyses with anti-STAT1 and anti-β-actin antibodies. The ratio of STAT1/β-actin was calculated. (B) A significant positive correlation was observed between STAT1/β-actin ratio and serum AST levels. (C) The percentage of STAT1+ hepatocyte nuclei/total hepatocytes was counted. (D) Representative double-immunohistochemical staining of STAT1 and PCNA from the livers of 34 HCV-infected livers with cirrhosis. Left panel is a representative area where high levels of STAT1 were detected in hepatocyte nuclei; middle panel is a representative area where high levels of PCNA were detected in hepatocyte nuclei; right panel is a representative area where high levels of both STAT1 and PCNA were detected, but not localized in the same cells. STAT1 proteins were stained blue, as indicated by blue arrows; PCNA was stained brown, as indicated by red arrows. (E) A negative correlation was observed between STAT1/β-actin ratio and the percentage of PCNA+ hepatocytes. (F) Representative immunohistochemical staining of Ki-67 from the livers of 34 HCV-infected livers with cirrhosis. Ki-67 was stained brown, as indicated by red arrows. (G) A negative correlation was observed between STAT1/β-actin ratio and the percentage of Ki-67+ hepatocytes.

We used immunostaining to further examine the activation of STAT1 in HCV-infected livers with cirrhosis. Because activated STAT1 proteins translocate into the nuclei, STAT1 activation is reflected as positive STAT1 staining in the nuclei by anti-STAT1 antibodies. In normal healthy human livers, STAT1 was found diffusely and at low levels in the cytosol of hepatocytes. Very few hepatocytes exhibited positive nuclei STAT1 staining. From HCV-infected livers with cirrhosis, the majority of the specimens displayed positive nuclei STAT1 staining in hepatocytes with varying degrees of staining. Figure 6C shows that the percentage of STAT1+ nuclei in hepatocytes from HCV-infected livers with cirrhosis was significantly higher than normal healthy controls. High levels of STAT1 staining in the cytosol of hepatocytes were also detected in about 40% to 60% of HCV-infected livers with cirrhosis (data not shown). We also performed double immunostaining with PCNA and STAT1, and 3 representative staining patterns are shown in Fig. 6D. In general, double immunostaining for PCNA and STAT1 revealed that cells which stained with high levels of STAT1 in either cytosol or nuclei were generally PCNA negative, and cells which stained positively with PCNA usually expressed STAT1 at low levels. Interestingly, hepatic levels of STAT1 proteins correlated inversely with the number of PCNA+ hepatocytes in these patients (P = .035) (Fig. 6E). Although PCNA has been widely used as an index of hepatocyte proliferation,25 recent studies suggest that PCNA is not specific for S phase.26 Thus, we also stained these samples with Ki-67, a marker widely used for cell proliferation in many organs including the liver.27 A representative immunostain of Ki-67 in HCV-infected liver tissues with cirrhosis is shown in Fig. 6F, while Fig. 6G shows an inverse correlation between the levels of hepatic STAT1 protein and Ki-67+ hepatocytes in these samples (P = .016).

Discussion

Previously, we demonstrated that the synthetic dsRNA poly I:C inhibits liver regeneration through activation of the innate immunity (NK/IFN-γ).17 In this report, we expand upon this work and provide in vivo and in vitro evidence suggesting: (1) poly I:C/IFN-γ suppression of liver regeneration and hepatocyte proliferation is mediated via a STAT1-dependent mechanism involving IRF-1 and p21cip1, and (2) STAT1 is activated and positively correlates with liver injury, but inversely correlates with hepatocyte proliferation in HCV-infected livers with cirrhosis.

STAT1 Contributes to Poly I:C/IFN-γ Inhibition of Liver Regeneration and Hepatocyte Proliferation: Involvement of IRF-1/p21cip1 and Negative Regulation by SOCS1.

Disruption of the STAT1 gene accelerates liver regeneration in male but not female mice (Fig. 1), suggesting that STAT1 may play a role in suppressing liver regeneration in male mice, but not in female mice in the PHx model. The reason for this discrepancy is not clear. Furthermore, we provide evidence suggesting that activation of STAT1 is a negative regulator for liver regeneration in our model of poly I:C+PHx. First, although activation of hepatic STAT1 is detected at low levels in the PHx model alone, high levels are detected in the PHx+poly I:C model.17 Disruption of the STAT1 gene abolishes poly I:C inhibition of liver regeneration, which is restored by overexpression of STAT1 (Fig. 1). The levels of STAT1 protein after hydrodynamic injection of STAT1 cDNA are much greater than in wild-type mice. It is our belief that STAT1 overexpression is pathophysiologically relevant because STAT1 is significantly elevated in HCV-infected livers with cirrhosis (Fig. 6).28–31 Second, poly I:C indirectly inhibits hepatocyte proliferation in vitro when co-cultured with liver MNCs via IFN-γ/STAT1-dependent mechanisms (Fig. 3). Third, IFN-γ inhibition of hepatocyte proliferation in vitro is STAT1-dependent (Fig. 4). Taken together, these findings strongly suggest that STAT1 plays an important role in the inhibition of liver regeneration and hepatocyte proliferation when activated by poly I:C or IFN-γ.

Next, we demonstrate that IRF-1 and p21cip1, 2 genes downstream of STAT1, contribute to the negative effect of STAT1 in liver regeneration. First, results shown in Fig. 2A and Fig. 4A clearly demonstrate that poly I:C and IFN-γ induction of IRF-1 in vivo and in vitro, respectively, are STAT1-dependent. IFN-γ induction of p21cip1in vitro is completely abolished in STAT1−/− and IRF-1−/− hepatocytes, suggesting that p21cip1 is a downstream gene of STAT1/IRF-1 (Fig. 4A). However, poly I:C induction of p21cip1in vivo is only slightly reduced in STAT1−/− mice (Fig. 2A). In addition to IFN-α/β and IFN-γ, in vivo treatment with poly I:C also elevates expression of many other cytokines including IL-6 and TNF-α, followed by induction of p21cip1 expression via a STAT1-independent mechanism. This may be responsible for the significant induction of p21cip1 in STAT1−/− mice after poly I:C injection. Moreover, we also demonstrate that poly I:C inhibition of liver regeneration in vivo and IFN-γ inhibition of hepatocyte proliferation in vitro are only partially diminished in IRF-1−/− and p21cip1–/– mice (Figs. 2, 4), suggesting that poly I:C/IFN-γ inhibits liver regeneration and hepatocyte proliferation via IRF-1/p21cip1-dependent and -independent mechanisms. Recently, Brooling et al.32 reported that generation of nitric oxide also contributes to the inhibitory effect of IFN-γ on hepatocyte proliferation in conjunction with TNF-α or lipopolysaccharide, which could be another mechanism contributing to IFN-γ inhibition of hepatocyte proliferation in addition to IRF-1/p21cip1.

SOCS1, which is another downstream gene of IFN-γ/STAT1, plays a key role in terminating the IFN-γ signaling pathway.18 We demonstrate here that SOCS3, but not SOCS1, is induced after PHx (Fig. 5), which is consistent with data from the Fausto laboratory.24 Recently, the same group obtained results using liver-specific SOCS3−/− mice, suggesting that SOCS3 is a crucial negative regulator of liver regeneration after PHx.33 SOCS1 is not induced in the PHx model, suggesting that SOCS1 may play a minor role in liver regeneration in this model. However, SOCS1 is significantly elevated after poly I:C treatment in our model and is STAT1-dependent (Fig. 5). As SOCS1−/− mice typically do not survive after birth, liver regeneration cannot be studied in these mice. However, IFN-γ−/−SOCS1−/− mice do survive, and IFN-γ treatment induces greater inhibition of cell proliferation of SOCS1−/− hepatocytes isolated from these mice compared to IFN-γ −/−SOCS1+/+ controls (Fig. 5), suggesting that SOCS1 may stimulate liver regeneration via inhibition of the IFN-γ/STAT1 signaling pathway in our model of poly I:C+PHx. Conclusive evidence is dependent on further studies using liver-specific SOCS1−/− mice.

In addition to inhibiting hepatocyte proliferation, IFN-γ also induces apoptosis in cultured hepatocytes via STAT1/IRF-1-dependent mechanisms (Fig. 4D-F). Although IFN-γ induces apoptosis in hepatocytes in vitro, apoptosis is not significantly enhanced in vivo after poly I:C+PHx compared to PHx alone, which may be due to activation of anti-apoptotic signals (such as STAT3 and NF-κB) in the in vivo model. Taken together, IFN-γ treatment induces both apoptosis and cell cycle arrest of hepatocytes via STAT1/IRF-1 dependent mechanisms; however, it is not clear how much apoptotic cell death contributes to reducing cell proliferation after IFN-γ treatment. Evidence suggests that different downstream genes controlled by IRF-1 are involved in IFN-γ-mediated cell cycle arrest and apoptosis. For example, IRF-1 induction of p53 and p21cip1 is involved in IFN-γ-mediated cell cycle arrest, while IRF induction of caspase 1 and caspase 7 is involved in IFN-γ-mediated cell apoptosis.34, 35 Further studies are required to explore the genes involved in both apoptosis and cell cycle arrest of hepatocytes induced by IFN-γ.

Elevated Levels of STAT1 Protein Correlate Positively With Liver Injury, But Inversely With Hepatocyte Proliferation in HCV-Infected Livers With Cirrhosis.

Increasing evidence suggests that STAT1 is activated and induced in HCV-infected livers. First, immunohistochemistry analyses show that positive STAT1 staining is found in hepatocyte nuclei in the HCV-infected livers with cirrhosis, but very few in normal healthy livers (Fig. 6). It has also been reported that levels of phospho-STAT1 protein are elevated in HCV-infected livers.30 Second, we28 and others29–31 reported that levels of STAT1 protein and mRNA are significantly induced in patients with chronic HCV infection. Because STAT1 activation by IFNs has been shown to be responsible for inducing STAT1 protein expression in hepatocytes,36 high levels of STAT1 protein in HCV-infected livers with cirrhosis may reflect prior activation of STAT1 in these specimens. Third, a variety of downstream STAT1 genes are upregulated in HCV-infected livers,37–42 further confirming that STAT1 is activated in these samples. It has been shown that both IFN-γ and IFN-α/β, two major cytokines responsible for STAT1 activation in the liver,19 are elevated in patients with HCV infection,38, 42, 43 while HCV itself has been shown to inhibit STAT1 activation.29, 44 This suggests that elevated IFNs may contribute to STAT1 activation in HCV-infected livers, while HCV itself is unlikely to be involved.

Additionally, we demonstrate that levels of STAT1 protein correlate positively with liver injury (elevated serum AST levels) but negatively with hepatocyte proliferation (PCNA- and Ki-67-positive hepatocytes) in HCV-infected patients with cirrhosis (Fig. 6). Accumulating evidence indicates that activation of STAT1 plays an important role in liver injury through at least two mechanisms, including induction of hepatocyte apoptosis via IRF-1-dependent mechanism (Fig. 4D-E)21, 36 and induction of hepatic inflammation via stimulating expression of a variety of chemokines and adhesion molecules.45 It has been reported that a variety of genes downstream of STAT1 are upregulated in HCV-infected livers,37–42 including chemokines (IP-10, I-TAC, CXCL, etc.)37–40 and proapoptotic genes (IRF-1),42 with the latter found to be positively correlated with liver injury (elevated serum ALT levels) in chronic HCV-infected patients.42 These findings suggest that activation of STAT1 contributes to liver injury in HCV patients through inducing a variety of downstream genes. In addition, STAT1 also plays an important role in inducing cell cycle arrest of hepatocytes in vivo and in vitro, and the downstream genes IRF-1 and p21cip1 are involved (Figs. 1, 3, and 4). Recently, Donato et al.46 reported that IFN-α therapy results in significant suppression of liver cell proliferation in both virologic responders and nonresponders, which is likely mediated via activation of the STAT1 signaling pathway. Overexpression of p21cip1 has been shown to correlate with inhibition of hepatocyte proliferation in patients with HCV infection12 and biliary cirrhosis.14 Moreover, IRF-1, an antiproliferative gene controlled by STAT1, is also detected at high levels in HCV-infected livers.42 Thus, it is likely that activation of STAT1 induces p21cip1 and IRF-1 protein expression, and subsequently inhibits hepatocyte proliferation in patients with HCV infection.

In conclusion, we found that in addition to well-documented antiviral, proapoptotic, and proinflammatory effects,19 STAT1 also plays an important role in dsRNA inhibition of liver regeneration via IRF-1/p21cip1-dependent mechanisms in a murine model of poly I:C+PHx treatment. STAT1 signaling is also activated in HCV-infected livers with cirrhosis and likely contributes to disease pathogenesis through induction of liver injury and inhibition of liver regeneration.

Ancillary