Dissociation between liver inflammation and hepatocellular damage induced by carbon tetrachloride in myeloid cell–specific signal transducer and activator of transcription 3 gene knockout mice

Authors

  • Norio Horiguchi,

    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
  • Fouad Lafdil,

    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
  • Andrew M. Miller,

    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
  • Hua Wang,

    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
  • Mohanraj Rajesh,

    1. Section on Oxidative Stress and Tissue Injury, Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Partha Mukhopadhyay,

    1. Section on Oxidative Stress and Tissue Injury, Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Xin Yuan Fu,

    1. Department of Microbiology and Immunology, Indiana University School of Medicine, Indianapolis, IN
    Search for more papers by this author
  • Pal Pacher,

    1. Section on Oxidative Stress and Tissue Injury, 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, LPS, NIAAA/NIH, 5625 Fishers Lane, Rm 2S-33, Bethesda, MD 20892
    Search for more papers by this author
    • fax: 301-480-0257


  • Potential conflict of interest: Nothing to report.

Abstract

Liver injury is associated with inflammation, which is generally believed to accelerate the progression of liver diseases; however, clinical data show that inflammation does not always correlate with hepatocelluar damage in some patients. Investigating the cellular mechanisms underlying these events using an experimental animal model, we show that inflammation may attenuate liver necrosis induced by carbon tetrachloride (CCl4) in myeloid-specific signal transducer and activator of transcription 3 (STAT3) knockout mice. As an important anti-inflammatory signal, conditional deletion of STAT3 in myeloid cells results in markedly enhanced liver inflammation after CCl4 injection. However, these effects are also accompanied by reduced liver necrosis, correlating with elevated serum interleukin-6 (IL-6) and hepatic STAT3 activation. An additional deletion of STAT3 in hepatocytes in myeloid-specific STAT3 knockout mice restored hepatic necrosis but decreased liver inflammation. Conclusion: Inflammation-mediated STAT3 activation attenuates hepatocellular injury induced by CCl4 in myeloid-specific STAT3 knockout mice, suggesting that inflammation associated with a predominance of hepatoprotective cytokines that activate hepatic STAT3 may reduce rather than accelerate hepatocellular damage in patients with chronic liver diseases. Hepatology 2010

Inflammation is associated with various types of acute and chronic liver diseases and is considered a key contributor of hepatocellular damage and progression to liver fibrosis and hepatocellular carcinoma.1-7 However, clinical data show that hepatic inflammation does not always correlate with hepatocellular damage (serum alanine aminotransferase [ALT] elevations).8-13 To understand the cellular mechanism of this phenomena, a well-established experimental animal model14 to study hepatotoxin-induced liver inflammation and injury induced by carbon tetrachloride (CCl4) was used in our investigation. Once CCl4 is injected, the cytochrome p450 2E1 (CYP2E1) in hepatocytes metabolizes it into trichloromethyl radicals (CCl3*), which then causes lipid peroxidation and membrane damage.14 These damaged hepatocytes generate free radicals, thereby activating Kupffer cells/macrophages to produce proinflammatory and anti-inflammatory cytokines that control the progression of liver inflammation and injury.14 Among the cytokines produced, interleukin-6 (IL-6) has been shown to be a hepatoprotective cytokine in this model15 as well as in several other liver injury models, including ischemia/reperfusion, partial hepatectomy, alcoholic and nonalcoholic fatty liver, and concanavalin A (Con A)–induced T cell hepatitis.16-20 Additionally, the anti-inflammatory cytokine IL-10 also has been shown to play an important role in controlling inflammation in CCl4-induced chronic liver injury and fibrosis and therefore may participate in protecting the liver from disease progression.21, 22

The hepatoprotective effects of IL-6 are mediated mainly via activation of signal transducer and activator of transcription 3 (STAT3) in hepatocytes.16 In contrast, the anti-inflammatory effects of IL-10 in the liver are likely mediated via activation of STAT3 in Kupffer cells/macrophages. In general, IL-6 binds to its corresponding gp80 receptor located on hepatocytes and induces dimerization of the gp130 signal chain. Consequently, the Janus kinases associated with the gp130 chains also dimerize and autophosphorylate. Once activated, the Janus kinases then phosphorylate gp130 to activate STAT3 monomers. Phosphorylated STAT3 then forms dimers and translocates into the nuclei to induce expression of a wide array of genes, including anti-apoptotic and anti-oxidative genes that protect against hepatocyte damage. The protective role of hepatic STAT3 has been well documented in studies with hepatocyte-specific gp130 or STAT3 knockout mice; these mice have shown to be more susceptible to hepatocellular damage induced by various toxins.23-25 Moreover, a conditional ablation of the STAT3 gene in myeloid linage cells (for example, macrophages) have shown markedly enhanced systemic and liver inflammation,26-28 which clearly suggests the anti-inflammatory functions of STAT3 in myeloid cells. In this investigation, we found that myeloid-specific STAT3 knockout (STAT3math image) mice are more susceptible to CCl4-induced liver inflammation, but are surprisingly resistant to CCl4-induced necrosis. Further study revealed that the enhanced inflammation observed is associated with elevated hepatic STAT3 activation, which may explain the reduced necrosis observed in these mice.

Abbreviations

ALT, alanine aminotransferase; CCl4, carbon tetrachloride; CCR2, CC chemokine receptor 2; ConA, concanavalin A; CYP2E1, p450 cytochrome p450 2E1; GSH, glutathione; IFN, interferon; IL, interleukin; KO, knockout; MPO, myeloperoxidase; OSM, oncostatin M; STAT, signal transducer and activator of transcription; TNF-α, tumor necrosis factor alpha.

Materials and Methods

Mice.

Eight-week-old to ten-week-old male mice were used in all experiments. Hepatocyte-specific STAT3 knockout (KO) mice (STAT3math image) and myeloid-specific STAT3 KO mice (STAT3math image)were generated as described previously.27 The STAT3math image mice were described previously as M/N-STAT3KO mice.27 The corresponding littermates of the wild-type mice were used as controls. For deletion of STAT3 in both hepatocytes and myeloid cells, STAT3math image and STAT3math image were bred to generate four lines of mice, including double KO (STAT3math image), STAT3math image, STAT3math image, and littermate wild-type controls. All mice were fed regular chow unless specified. All animal experiments were conducted in accordance with National Institutes of Health guidelines and approved by the Institutional Animal Care and Use Committee of the National Institute on Alcohol Abuse and Alcoholism.

Mouse Model for CCl4-Induced Liver Injury.

Liver injury was induced by intraperitoneal injection with 2 mL/kg body weight 10% CCl4 (Sigma) dissolved in olive oil (Sigma).

Statistical Analysis.

Data are expressed as mean ± SD. To compare values obtained from two groups, the Student t test was performed. To compare values obtained from three or more groups, one-factor analysis of variance was used, followed by Tukey's post hoc test. Statistical significance was taken at the P < 0.05 level.

Other Methods.

All other methods are described in the supporting documents.

Results

STAT3math image Mice Are More Susceptible to Liver Inflammation But Are Resistant to Liver Necrosis Induced by CCl4.

As shown in Fig. 1A-C, treatment of wild-type littermate control mice with CCl4 induced serum ALT elevation, liver necrosis, and inflammation, with peak effect occurring 24 hours after injection. Compared with wild-type mice, STAT3math image mice had greater inflammatory cell infiltration around the hepatic central vein, but surprisingly, these mice had lower serum ALT levels and less liver necrosis (Fig. 1A-C). Terminal deoxynucleotidyl transferase-mediated 2′-deoxyuridine 5′-triphosphate nick-end labeling assay results revealed that the number of apoptotic hepatocytes was significantly lower in STAT3math image mice than in wild-type mice 24 hours after CCl4 injection (Fig. 1D).

Figure 1.

STAT3math image mice are more susceptible to liver inflammation, but resistant to necrosis/apoptosis induced by CCl4. Wild-type and STAT3math image mice were injected with CCl4 and then sacrificed at various time points. (A) Serum ALT levels. (B) Hematoxylin-eosin staining of liver tissues from the mice treated 24 hours after CCl4 injection. (C) Necrotic area was quantified. (D) Liver tissues were also stained with terminal deoxynucleotidyl transferase-mediated 2′-deoxyuridine 5′-triphosphate nick-end labeling to determine hepatocyte apoptosis, and the number of terminal deoxynucleotidyl transferase-mediated 2′-deoxyuridine 5′-triphosphate nick-end labeling+ hepatocytes was quantified. Values represent means ± SD (n = 4-10). *P < 0.05, **P < 0.01, ***P < 0.001 in comparison with corresponding WT groups. WT, littermate wild-type control mice (LyzMCre-STAT3flox/flox); STAT3math image (LyzMCre+STAT3flox/flox).

Immunohistochemical analyses with anti-myeloperoxidase (MPO) staining showed that most cells infiltrating the liver after CCl4 injection were neutrophils (Fig. 2A). These neutrophils are either located within sinusoids or infiltrated into liver parenchyma (Fig. 2A and Supporting Fig. S1a). The number of MPO+ neutrophils was much higher in STAT3math image mice compared with wild-type mice after CCl4 injection. Flow cytometry analyses of liver inflammatory cells showed that the percentage and total number of neutrophils (CD11b+Gr-1bright cells) in the liver were significantly higher in STAT3math image mice than in wild-type mice before or 24 hours after CCl4 injection (Fig. 2B-D). Lee et al.29 previously demonstrated that STAT3-deficient neutrophils matured normally and were functional.29 Here we also confirmed that neutrophils from wild-type and STAT3math image mice had similar respiratory burst (Supporting Fig. S1c).

Figure 2.

STAT3math image mice have more neutrophil infiltration in the liver than wild-type mice after CCl4 treatment. Wild-type (WT) and STAT3math image mice were injected with CCl4 and sacrificed 24 hours after injection. (A) Liver tissues were stained with anti-MPO (myeloperoxidase) antibody for neutrophils (MPO-positive brown staining). (B) Liver immune cells were isolated and analyzed by flow cytometry with anti-CD11b and anti-Gr-1 antibodies. Gr-1brightCD11b+ cells in the R3 area represent neutrophils. (C, D) The percentage and the total number of liver lymphocytes were quantified. Values represent means ± SD (n = 4-5). *P < 0.05, in comparison with corresponding littermate WT groups.

Figure 3A shows that basal levels of various hepatic inflammatory cytokines and chemokines were higher in STAT3math image mice compared with wild-type mice. In wild-type mice, treatment with CCl4 increased the expression of these cytokines and chemokines; however, this increase was much more profound in STAT3math image mice. Serum levels of several inflammatory cytokines also were found to be elevated after CCl4 injection, and again, these elevations were higher in STAT3math image mice compared with wild-type mice (Fig. 3B). Serum IL-12p70 and IL-10 were below detection levels in both groups (data not shown).

Figure 3.

Up-regulation of cytokines and chemokines in STAT3math image mice. Wild-type littermate controls and STAT3math image mice were injected with CCl4 and then sacrificed 12 hours and 24 hours after injection. (A) Liver tissues were collected for real-time polymerase chain reaction analyses for expression of proinflammatory cytokines and chemokines. (B) Serum levels of cytokines and chemokines. Values represent means ± SD (n = 4-10) *P < 0.05, **P < 0.01 in comparison with corresponding WT groups.

STAT3math image Mice Show Higher Hepatic STAT3 Activation and Are Resistant to Hepatic Oxidative Stress After CCl4 Injection Compared with Wild-Type Mice.

Because p450 CYP2E1-mediated CCl4 metabolism is essential for CCl4-induced liver injury,14 we examined whether alterations in CCl4 metabolism are responsible for reduced liver injury in STAT3math image mice. As shown in Fig. 4A and Supporting Fig. S1b, the basal levels of CYP2E1 expression were comparable in livers from wild-type and STAT3math image mice. After CCl4 administration, CYP2E1 expression was down-regulated in both groups of STAT3math image and wild-type mice. The down-regulation seemed to be more profound in the former group compared with the latter group, suggesting that CCl4 metabolism is not reduced in STAT3math image mice compared with wild-type mice, because the down-regulation of CYP2E1 is caused by CCl4 metabolism.

Figure 4.

Elevated hepatic STAT3 activation in STAT3math image mice. (A, B) Mice were injected with CCl4 and then sacrificed at various time points. Liver tissues were collected for western blot analyses (A) or for measurement of GSH levels and HNE-His adduct (B). (C, D) Kupffer cells were isolated from wild-type and STAT3math image mice and treated with lipopolysaccharide (1 ng/mL) for 24 hours. The supernatants were collected for measurement of various cytokines (C), and cells were collected for real-time polymerase chain reaction analyses (D). Values represent means ± SD (n = 5). *P < 0.05 in comparison with corresponding WT littermate groups. (E) STAT3math image mice were treated with control antibody or anti-IL-6 antibody for 2 days. Liver tissues were then collected for western blot analyses.

Injection of CCl4 induced increased oxidative stress, which was higher in STAT3math image mice than in wild-type mice as determined by measuring the 4-hydroxynonenal–histidine (HNE-His) adduct (Fig. 4B). Although STAT3math image mice had higher levels of oxidative stress, CCl4 treatment–induced glutathione (GSH) depletion, which was observed in wild-type mice, was not observed in STAT3math image mice (Fig. 4B). Why STAT3math image mice had higher levels of oxidative stress without GSH depletion after CCl4 treatment compared with wild-type mice is not clear. Elevated inflammation may trigger some compensatory effects to prevent GSH depletion in STAT3math image mice, which should be explored in future studies.

To understand the mechanism by which STAT3math image mice are resistant to CCl4-induced liver injury, we measured activation of hepatic STAT3, a signaling molecule that has been shown to promote hepatocyte survival in the liver.23-25 Basal STAT3 activation (pSTAT3) was higher in STAT3math image mice than in wild-type mice (Fig. 4A). Injection with CCl4 induced much higher and prolonged STAT3 activation in STAT3math image mice compared with wild-type mice. Expression of STAT3 protein was also slightly higher in STAT3math image mice than in wild-type mice, whereas expression of STAT1 protein was comparable between these groups. Figure 4A shows that the basal levels (0 hour time point) of hepatic pSTAT3 are higher in STAT3math image mice than wild-type mice. Our previous study showed that STAT3math image mice had similar basal levels of hepatic pSTAT3 compared with wild-type mice (Fig. 2C in Lafdil et al.28). The discrepancy between our current and previous studies was likely attributable to the mice being fed regular chow in the current study and a medicated diet in our previous study. Supporting Fig. S2a confirmed that feeding with a medicated diet abolished the basal levels of hepatic pSTAT3 in STAT3math image mice. Despite the diminished basal levels of hepatic pSTAT3 activation after feeding with a medicated diet, STAT3math image mice were resistant to CCl4-induced liver injury (elevation of serum ALT/aspartate aminotransferase) (Supporting Fig. S2b). In addition, the basal levels of p38 MAPK were higher in the livers of STAT3math image mice compared with wild-type mice, whereas activation of extracellular signal-regulated kinase in the liver was lower in STAT3math image mice than in wild-type mice (Supporting Fig. S3). Activation of phospho-nuclear factor kappaB p65 was higher in the liver of STAT3math image mice compared with wild-type mice after CCl4 injection (Supporting Fig. S3).

To understand the mechanisms underlying elevated hepatic STAT3 activation in STAT3math image mice, the production and expression of several cytokines (IL-6, IL-22, and oncostatin M [OSM]) and growth factors (hepatocyte growth factor, epidermal growth factor), which stimulate STAT3 activation in hepatocytes, were examined in Kupffer cells. Production and expression of IL-6 were markedly higher in Kupffer cells from STAT3math image mice than from wild-type mice with or without lipopolysaccharide stimulation (Fig. 4C, D). Expression of OSM was also higher, to a lesser extent, in STAT3math image Kupffer cells than in wild-type cells. Expressions of IL-22 and hepatocyte growth factor were comparable between these two types of cells (Fig. 4D), whereas expression of epidermal growth factor was undetectable in Kupffer cells (data not shown). In addition, STAT3math image Kupffer cells produced higher levels of IL-10, tumor necrosis factor alpha (TNF-α), and interferon-gamma (IFN-γ) compared with wild-type Kupffer cells with or without lipopolysaccharide stimulation. Finally, IL-6 neutralizing antibody significantly diminished pSTAT3 levels in the liver of STAT3math image mice (Fig. 4E).

STAT3math image Mice Are More Susceptible to Liver Necrosis/Apoptosis But Less Sensitive to Inflammation Induced by CCl4 Injection.

Because STAT3 has been shown to play an important role in hepatoprotection in several murine models of liver injury,23-25 we hypothesized that increased STAT3 activation in the liver of STAT3math image mice may contribute to the reduced necrosis found in these mice after CCl4 injection. To test this hypothesis, we used hepatocyte-specific STAT3 knockout (STAT3math image) mice to first examine whether STAT3 in hepatocytes protects against CCl4-induced liver injury. As illustrated in Fig. 5A-C, CCl4 injection induced greater liver damage in STAT3math image mice than in wild-type mice, as evidenced by increased serum ALT levels and more severe necrosis and apoptosis. Additionally, injection of CCl4 induced more profound GSH depletion in STAT3math image mice compared with wild-type mice (Fig. 5D). Although STAT3math image mice had more liver necrosis, the expression of inflammatory cell markers (such as CC chemokine receptor 2 and F4/80) and proinflammatory cytokines (such as TNF-α, IL-6, macrophage inflammatory protein 2, and intracellular adhesion molecule 1) were lower in these mice compared with wild-type mice (Fig. 5E, F, Fig. 6A). Serum TNF-α and IL-6 levels were slightly higher in STAT3Hep−/− mice than in wild-type mice 24 hours post CCl4 injection (Fig. 6B).

Figure 5.

STAT3math image mice are more susceptible to liver necrosis/apoptosis but resistant to inflammation induced by CCl4. Mice were injected with CCl4 and then sacrificed 12 hours and 24 hours after injection. (A) Serum ALT levels. (B) Liver tissues collected for hematoxylin-eosin staining and terminal deoxynucleotidyl transferase-mediated 2′-deoxyuridine 5′-triphosphate nick-end labeling staining to determine hepatocyte apoptosis (original magnification ×200). (C) Necrotic area and terminal deoxynucleotidyl transferase-mediated 2′-deoxyuridine 5′-triphosphate nick-end labeling–positive hepatocytes in B were quantified. (D) Hepatic GSH levels. (E, F) Liver tissues were collected for real-time polymerase chain reaction analyses for MPO, CCR2, and F4/80 expression. Values represent means ± SD (n = 5-6 in each group). *P < 0.05, **P < 0.01, ***P < 0.001 in comparison with corresponding wild-type littermate groups.

Figure 6.

Reduced expression of hepatic cytokines and chemokines in STAT3math image mice after CCl4 injection. Mice were injected with CCl4 and then sacrificed at 12 hours and 24 hours. (A) Liver tissues were collected for real-time PCR analyses for cytokine and chemokine expression at 12 hours postinjection. (B) Serum cytokines and chemokines. Values represent means ± SD (n = 5-6 in each group) *P < 0.05, **P < 0.01, ***P < 0.001 in comparison with corresponding wild-type littermate groups.

Disruption of STAT3 in Hepatocytes Restores the Sensitivity of STAT3math image Mice to CCl4-Induced Necrosis But Reduces CCl4-Induced Inflammation.

To test the hypothesis that the resistance of STAT3math image mice to CCl4-induced liver injury is caused by enhanced STAT3 activation in hepatocytes, we deleted STAT3 from the hepatocytes of STAT3math image mice by generating hepatocyte and macrophage/neutrophil specific double KO (STAT3math image double KO). Double KO mice showed greater ALT elevation and more necrosis at 24 hours than STAT3math image mice (Fig. 7A-C). Regarding liver inflammation, double KO mice showed significantly lower expression of MPO and CC chemokine receptor 2 (CCR2) at, respectively, 24 and 12 hours after CCl4 injection, compared with STAT3math image mice (Fig. 7D, E). Expression of F4/80 was comparable between STAT3math image mice and double KO mice.

Figure 7.

Disruption of STAT3 in hepatocytes restores the sensitivity of STAT3math image mice to CCl4-induced liver injury. Four different lines of mice were injected with CCl4 and then sacrificed 12 hours and 24 hours after injection. (A) Hematoxylin-eosin staining of liver tissues 24 hours after CCl4 treatment. (B) Serum ALT levels. (C) Necrotic area in panel A was quantified. (D, E) Real-time PCR analyses for MPO, CCR2, and F4/80 expression 12 hours and 24 hours after CCl4 injection. Values represent mean ± SD (n = 7-8 in each group). *P < 0.05, **P < 0.01, ***P < 0.001 in comparison with groups as indicated.

Hepatic expression of several cytokines (TNF-α, IL-1β, IFN-γ) and chemokines (macrophage inflammatory protein 2, monocyte chemotactic protein-1, intercellular adhesion molecule 1) and serum levels of proinflammatory cytokines (TNF-α and IL-6) were lower in double KO than in STAT3math image mice 12 hours after CCl4 injection (Fig. 8).

Figure 8.

Disruption of hepatic STAT3 suppresses cytokine and chemokine production in STAT3math image mice. Mice were injected with CCl4 and then sacrificed 12 hours postinjection. (A) Real-time PCR analyses of cytokine and chemokine expression in the liver. (B) Serum levels of cytokines and chemokines. Values represent means ± SD (n = 7-8). *P < 0.05, **P < 0.01, ***P < 0.001 denotes significant differences between STAT3math image and STAT3math image mice. (C) Schematic of the interaction of hepatic and myeloid STAT3 in controlling liver necrosis and inflammation. Activation of Kupffer cells not only produces toxins that cause liver injury and inflammation, but also produces hepatoprotective cytokines (such as IL-6, OSM, and maybe other factors) that activate STAT3 in hepatocytes and ameliorate hepatocellular damage.

Discussion

In this paper, we reported a model of STAT3math image mice in which inflammation does not correlate with hepatocellular injury induced by CCl4 and whereby inflammation-associated hepatic STAT3 activation reduces liver necrosis (Fig. 8C). This may help explain a lack of correlation between inflammation and hepatocellular damage (serum ALT levels) observed in some patients with chronic liver diseases.8-13

In our study, injection with CCl4 induced much higher levels of systemic and hepatic inflammation in STAT3math image mice than in wild-type mice, suggesting that STAT3 in myeloid cells plays an important role in inhibiting inflammation in a model of CCl4-induced liver injury. This is consistent with previously well-documented studies showing the anti-inflammatory effects of STAT3 in myeloid cells using various models of organ injury.26-28 Surprisingly, despite higher levels of inflammation in STAT3math image mice, they had much lower serum ALT levels and less liver necrosis than wild-type mice after CCl4 administration. The resistance of STAT3math image mice to CCl4-induced liver necrosis may be attributable to either the impaired ability of STAT3-deficient neutrophils/macrophages to induce hepatocellular damage or the resistance of hepatocytes to CCl4-induced hepatocellular damage in STAT3math image mice. Several lines of evidence suggest that it is the second mechanism that contributes to reducing liver necrosis in STAT3math image mice because of enhanced STAT3 activation in the liver.

The hepatotoxicity of CCl4 is mediated by the direct induction of hepatocyte damage and indirect activation of Kupffer cells/macrophages and neutrophils.14 Activated Kupffer cells/macrophages produce free radicals and proinflammatory cytokines such as TNF-α that further trigger hepatocellular damage and induce neutrophil accumulation and activation.2, 5 Activated neutrophils can cause hepatocyte damage by releasing oxygen species and proteases.2, 5 We have previously shown that Kupffer cells from STAT3math image mice produce much higher levels of reactive oxygen species and TNF-α compared with those from wild-type mice.27 By using four different assays, Lee et al.29 previously demonstrated that STAT3-deficient neutrophils mature normally and are functional. In the current study we also confirmed that STAT3-deficient neutrophils from STAT3math image mice are functional in vitro because they produced a similar respiratory burst after phorbol 12-myristate 13-acetate stimulation compared with phorbol 12-myristate 13-acetate–stimulated wild-type mouse neutrophils (Supporting Fig. S1c). Moreover, an additional deletion of STAT3 in hepatocytes restored liver necrosis in STAT3math image mice after CCl4 administration, suggesting that neutrophils from STAT3math image mice are functional in vivo. Collectively, these findings suggest that STAT3-deficient macrophages and neutrophils in STAT3math image mice have normal ability to induce inflammatory responses and hepatocellular damage, and that the reduced liver necrosis observed in STAT3math image mice is not attributable to a defect in STAT3-deficient macrophages and neutrophils to induce hepatocellular damage.

The hepatoprotective effect of hepatic STAT3 has been well documented in various models of liver injury,23-25 which is further confirmed in our study using an experimental model of CCl4-induced liver damage (Fig. 5). This suggests that elevated liver STAT3 activation in STAT3math image mice likely contributes to the resistance of these mice to CCl4-induced liver necrosis and oxidative stress. This concept is further supported by conclusive evidence showing that deletion of STAT3 in hepatocytes restores liver necrosis in STAT3math image mice after CCl4 administration (Fig. 7). Interestingly, STAT3math image mice are resistant to CCl4-induced liver necrosis as demonstrated in the current study but more susceptible to Con A–induced T cell hepatitis despite enhanced STAT3 activation in the liver (Fig. 2 in Lafdil et al.28). This discrepancy could be attributable to the deletion of STAT3 in myeloid cells preferentially enhancing the Th1 cytokine (IFN-γ) response during Con A–induced T cell hepatitis, dominating over the hepatoprotective effect of STAT3 and resulting in accelerated liver injury in this model.28 Such preferential induction of IFN-γ was not observed in STAT3math image mice in the CCl4-induced liver injury model (2500 pg/mL serum IFN-γ in Con A model28 versus 25 pg/mL IFN-γ in CCl4 model in STAT3math image mice in the current study). In addition, STAT3math image mice are also more susceptible to chronic ethanol feeding-induced liver inflammation and injury.27 It has been well documented that chronic ethanol consumption inhibits STAT3 activation in the liver.30 Therefore, it is probable that chronic ethanol feeding diminishes hepatic STAT3 activation and abolishes the hepatoprotective effect of STAT3 in STAT3math image mice, resulting in enhanced liver injury in this alcohol-induced liver injury model.27

We have previously shown that STAT3 activation in hepatocytes promotes liver inflammation in alcohol-induced liver injury.27 Our findings in current studies showed that STAT3 activation in hepatocytes also plays a proinflammatory role in CCl4-induced liver injury because deletion of STAT3 reduced systemic and hepatic inflammation although it enhanced CCl4-induced liver damage (Fig. 5). In addition, an additional deletion of STAT3 in hepatocytes enhanced CCl4-induced liver necrosis but partially counteracted the strong inflammation in STAT3math image mice after CCl4 administration (Fig. 8). This is probably because deletion of STAT3 in hepatocytes reduced the hepatic expression of several STAT3-controlled inflammatory mediators (such as complement 3/5, IL-1, macrophage inflammatory protein 2, monocyte chemotactic protein 1, and intercellular adhesion molecule 1) in STAT3math image mice (Fig. 7). Taken together, these findings suggest that enhanced hepatocellular STAT3 activation in STAT3math image mice may contribute to not only the reduced liver necrosis but also the enhanced inflammation after CCl4 administration in these mice. In addition, elevated nuclear factor kappaB activation (Supporting Fig. S3) also may contribute to the reduced necrosis in STAT3math image mice because of the hepatoprotective functions of nuclear factor kappaB.

With respect to what factors are responsible for enhanced STAT3 activation in the liver of STAT3math image mice (Fig. 4A), many cytokines and growth factors, including IL-6, IL-6 family cytokines (such as OSM, IL-11, cardiotrophin-1, ciliary neurotrophic factor, leukemia inhibitory factor), IL-22, epidermal growth factor, and hepatocyte growth factor, have been shown to stimulate STAT3 in the liver.16, 31-33 Here, we provided several lines of evidence suggesting that IL-6 is an important factor responsible for the higher levels of pSTAT3 in the liver of STAT3math image mice compared with wild-type mice. First, the basal levels of serum and hepatic IL-6 were higher in STAT3math image mice than in wild-type mice, which is consistent with previous reports.27 Second, Kupffer cells from STAT3math image mice produced much higher levels of IL-6 than wild-type Kupffer cells (200-500 pg/mL from STAT3math image versus 10 pg/mL from wild-type mice) (Fig. 4C). Finally, blockage of IL-6 with a neutralizing antibody diminished the basal levels of pSTAT3 in the liver of STAT3math image mice (Fig. 4E). In addition, OSM also may contribute, to a lesser extent, to the enhanced pSTAT3 in the liver of STAT3math image mice because Kupffer cells from these mice expressed higher levels of OSM compared with wild-type cells (Fig. 4D).

Clinical Implication of This Study.

It is believed that inflammation plays a key role in contributing to the progression of liver diseases1-6; however, many studies have reported that inflammation does not always correlate with hepatocellular damage in patients with chronic liver diseases.8-13 Based on the findings from this and other previous studies, we speculate that inflammation associated with a predominance of hepatoprotective cytokines such as IL-6, IL-6–related cytokines, and IL-22 may not correlate with hepatocellular damage, whereas inflammation with a predominant expression of Th1 cytokines (such as IFN-γ) may be closely associated with liver injury. Indeed, the downstream targets of IFN-γ, such as STAT1 and IP-10, have been shown to correlate with hepatocellular damage in patients with viral hepatitis C infection.34 Thus, understanding the effects of different types of liver inflammation on hepatocellular damage may help us design better strategies to treat patients with chronic liver diseases.

Ancillary