Potential conflict of interest: Nothing to report.
Sepsis is an infection-induced syndrome with systemic inflammatory response leading to multiorgan failure and occasionally death. During this process, signal transducer and activator of transcription 3 (STAT3) is activated in the liver, but the significance of this molecule has not been established. We generated hepatocyte-specific STAT3-deficient mice (L-STAT3 KO) and examined the susceptibility of these mice to cecal ligation and puncture–induced peritonitis, a well-established septic model. L-STAT3 KO mice showed significantly higher mortality and produced lesser amounts of various acute phase proteins than control littermates. Although blood bacterial infection did not differ between L-STAT3 KO mice and control mice, the former showed deterioration of the systemic inflammatory response as evidenced by a significant increase in various cytokines such as tumor necrosis factor α, IFN-γ, IL-6, IL-10, monocyte chemoattractant protein 1, and macrophage inflammatory protein 1β. A similar hyperinflammatory response was observed in another septic model caused by lipopolysaccharide (LPS) injection. In vitro analysis revealed that soluble substances derived from hepatocytes and dependent on STAT3 were critical for suppression of cytokine production from LPS-stimulated macrophage and splenocytes. Conclusion: STAT3 activation in hepatocytes can attenuate a systemic hyperinflammatory response and lethality in sepsis, in part by suppressing immune cell overactivation, implying a critical role of hepatocyte STAT3 signaling in maintaining host homeostasis. (HEPATOLOGY 2007.)
Signal transducer and activator of transcription 3 (STAT3) mediates a signal from the IL-6 family of cytokines such as IL-6, oncostatin M, leukemia inhibitory factor, and ciliary neutrophic factor, and activates transcription of various target genes.1 Although a STAT3 is now known to be ubiquitously expressed in variety of cells and has pleiotropic functions, it was formerly termed acute phase response factor and was first identified in the liver as an inducible DNA binding protein binding to type 2 IL-6–responsive elements within the promoter of hepatic acute phase protein (APP) genes.2, 3 Because deletion of STAT3 leads to embryonic lethality in mice, the significance of STAT3 in adult organs has been investigated using conditional knockout animals generated by the Cre/loxP recombination system.4 Research has shown that STAT3 signaling within hepatocytes controls a variety of physiological or pathological processes, including hepatocyte proliferation after partial hepatectomy,5 apoptosis resistance of hepatocytes during Fas-mediated liver injury,6 and regulation of hepatic gluconeogenic genes.7 Although STAT3 is activated in response to a rise of circulating cytokines, the significance of hepatic STAT3 has not been elucidated under systemic inflammatory conditions.
Sepsis is an infection-induced systemic syndrome, the incidence of which is estimated at 750,000 cases annually in North America with overall mortality being approximately 30%, but rising to 40% in the elderly.8 Sepsis develops when the initial, appropriate host response to an infection becomes amplified and then dysregulated.9 Among those harmful or damaging responses is the rise of a variety of circulating cytokines such as IL-6, tumor necrosis factor α (TNF-α), IL-10, and IFN-γ. These cytokines lead directly to the development of systemic inflammatory response syndrome. During this process, an increasing proportion of patients will develop adult respiratory distress syndrome, disseminated intravascular coagulation, and/or acute renal failure, leading to the multiple organ dysfunction syndrome.10 The liver is also one of the target organs of multiple organ dysfunction syndrome, although liver dysfunction may cause patient death less frequently than cardiovascular dysfunction.11 Conversely, sepsis is a serious complication of severe liver diseases such as fulminant hepatitis12 and decompensated cirrhosis.13 Thus, research on the relevance of signal transduction in liver cells in the septic condition would not only satisfy basic scientific interest but would also have clinical implications.
In the present study, we used hepatocyte-specific STAT3-deficient (L-STAT3 KO) mice and examined the significance of STAT3 signaling within hepatocytes in a well-established murine model of sepsis. We found that STAT3 deficiency in hepatocytes causes exacerbation of the hyperinflammatory response by attenuating hepatic production of soluble substances that can suppress immune cell activation and also increases mortality in septic mice. This study identified an anti-inflammatory function of hepatic STAT3 signaling and its protective role against systemic inflammation, providing genetic evidence for a close link between hepatocytes and the immune system.
APP, acute phase protein; CLP, cecal ligation and puncture; LPS, lipopolysaccharide; L-STAT3 KO, hepatocyte-specific STAT3-deficient mice; STAT3, signal transducer and activator of transcription 3; TNF-α, tumor necrosis factor α; TUNEL, terminal deoxynucleotidyl transferase-mediated dUTP nick end-labeling.
Materials and Methods
Mice carrying a STAT3 gene with 2 loxP sequences flanking exon 22 and a STAT3 null allele (STAT3 fl/−) have been described previously.14 To generate mice with hepatocyte-specific STAT3 deficiency, we crossed STAT3 fl/− mice and Alb-Cre transgenic mice,15 which express the Cre recombinase gene under the regulation of the albumin gene promoter. We crossed Alb-Cre STAT3 fl/fl mice and STAT3 fl/− mice. The resulting Alb-Cre STAT3 fl/− mice were used as L-STAT3 KO mice. Sex-matched STAT3 fl/− mice obtained from the same litter were used as control mice. All mice were used at the age of 12-15 weeks. All animals were housed under specific pathogen-free conditions and were treated with humane care under approval from the Animal Care and Use Committee of Osaka University Medical School.
Cecal Ligation and Puncture and Lipopolysaccharide Injection.
Cecal ligation and puncture (CLP) is a well-established murine model of septic shock. The mice underwent CLP surgery as described previously.16 In brief, the mice were anesthetized via intraperitoneal injection of sodium pentobarbital. Under sterile condition, the cecum was assessed via a 1-cm midline incision of the lower abdomen, ligated with a suture below the ileocecal valve, and punctured once with a 23-gauge needle. The cecum was replaced in the peritoneum, and the abdomen was closed with sutures. The mice were injected with 1 mL of lactate Ringer's solution subcutaneously for fluid resuscitation. As another septic model, lipopolysaccharide (LPS) (form Escherichia coli 055: B5; Sigma, St. Louis, MO) was injected intraperitoneally at a dose of 4 mg/kg body weight.
Preparation of Peritoneal Macrophage.
To isolate peritoneal macrophages, we injected mice intraperitoneally with 2 mL of 4% thioglycollate. Peritoneal exudates cells were isolated from the peritoneal cavity 4 days after injection. The cells were incubated for 4 hours in 96-well plates and washed 3 times with phosphate-buffered saline. We used the adherent cells as peritoneal macrophages for further experiments.
Determination of the Bacterial Load.
Mice were sacrificed 24 hours after CLP surgery. Samples of blood were obtained in sterile condition. Fifty microliters of the blood were then plated on heart-infusion plates. The heart-infusion plates were incubated at 37°C overnight, and the number of bacteria colonies was counted. Results were expressed as log10 of CFU.
Blood samples were obtained 24 hours after CLP or LPS injection. Acute phase proteins, cytokines, and chemokines in plasma were determined via MultiAnalyte Profile testing (Rules Based Medicine, Austin, TX). Levels of serum ALT and creatinine were measured with a standard UV method using a Hitachi type 7170 automatic analyzer (Tokyo, Japan).
Measurement of Culture Supernatant.
Levels of cytokines (TNF-α, IL-6, IL-10, and IFN-γ) in the culture supernatants were measured using commercially available ELISA kits in accordance with the manufacturer's instructions (BD Biosciences-Pharmingen, San Diego, CA). Haptoglobin was determined in cell-free supernatants by using a commercially available ELISA kit (Immunology Constants Laboratory, Newberg, OR).
Western Blot Analysis.
The total cellular protein was extracted with the RIPA buffer containing 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 50 μg/mL aprotinin, 100 μg/mL phenylmethylsulfonyl fluoride, and 50 mM sodium fluoride in phosphate-buffered saline (pH 7.4). Twenty micrograms of protein were separated via 7.5% SDS-PAGE and blotted onto a polyvinylidene difluoride membrane. After blocking with Tris-buffered saline 0.1% Tween 20 containing 5% skim milk or Blocking One-P (Nacalai Tesque, Kyoto, Japan) for 1 hour at room temperature, the membrane was incubated overnight at 4°C with antibodies to STAT3 or tyrosine705-phosphorylated STAT3 (Cell Signaling Technology, Danvers, MA), respectively. After washing with Tris-buffered saline 0.1% Tween 20, the membrane was incubated with anti–horseradish peroxidase–linked antibody for 1 hour at room temperature. The immune complex was detected by an enhanced chemiluminescent assay. In some experiments, tyrosine701-phosphorylated STAT1 antibody (Cell Signaling Technology) was also used. This antibody recognizes the phosphorylated form of both STAT1α and STAT1β.
Histology and Terminal Deoxynucleotidyl Transferase-Mediated dUTP Nick End-Labeling.
The formalin-fixed livers were paraffin-embedded, and liver sections were analyzed by hematoxylin-eosin staining. Terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick-end labeling (TUNEL) was performed using an ApopTag kit according to the manufacturer's instructions (Serological Corporation, Norcross, GA).
Primary Culture of Hepatocytes.
Livers were digested using a standard in situ 2-step collagenase perfusion procedure (Gibco BRL, Rockville, MD). Hepatocytes were isolated from nonparenchymal cells via subsequent centrifugation at 50g for 1 minute. In a selected experiment, nonparenchymal cells in the supernatants were pelleted at 1,500 rpm for 5 minutes and subjected to western blot analysis. Isolated hepatocytes with >90% viability were cultured in Williams' medium E containing 10% fetal bovine serum overnight. On the next day, the cells were stimulated with recombinant IL-6 (PeproTech, London, UK). The cells were harvested after 2 hours for the analysis of STAT3 activation. In another experiment, supernatants were harvested after 48 hours.
Cytokine Production by Macrophage and Splenocytes.
The murine macrophage cell line RAW 264.7 was obtained from the American Type Culture Collection (Manassas, VA). RAW cells were plated at a density of 5 × 105/well in a 96-well plate and were incubated at 37°C in culture supernatants of hepatocyte from L-STAT3 KO mice or control mice. As a control, RAW cells were also cultured in Williams' medium E. After 24 hours, LPS was added to achieve a final concentration of 100 ng/mL. After 24 hours of incubation at 37°C in an atmosphere of 5% CO2, the supernatant was collected and stored at −80°C for measurement of TNF-α, IL-6, and IL-10. Splenocytes were isolated by way of a standard procedure for wild-type mice17 and incubated with hepatocyte culture supernatant. Twenty-four hours after incubation, the cells were stimulated with LPS (1,000 ng/mL) for 24 hours. The resultant culture supernatant was subjected to IFN-γ ELISA.
Kaplan-Meier curves were used to show survival over time. Data are expressed as interquartile range and median and compared using the Mann-Whitney U test. Statistical significance was set at P < 0.05.
Mice with hepatocyte-specific STAT3 deficiency were produced by crossing floxed STAT3 mice and Alb-Cre transgenic mice carrying the Cre recombinase gene under the regulation of the albumin gene promoter. L-STAT3 KO mice were born and grew without any gross abnormality. Western blot analysis revealed that STAT3 expression was substantially decreased in the liver but not in other organs (Fig. 1A). Isolation of hepatocytes from nonparenchymal cells by liver perfusion followed by centrifugation confirmed that STAT3 deficiency is specific in hepatocytes (Fig. 1B). In addition, STAT3 expression did not differ in peritoneal macrophages between L-STAT3 KO mice and control littermates (Fig. 1C). Those cells isolated from L-STAT3 KO mice produced similar levels of TNF-α in response to LPS compared with those from control littermates (Fig. 1D).
L-STAT3 KO Mice Are More Vulnerable to Septic Shock.
To examine the role of hepatic STAT3 during septic shock, we used a well-examined clinically relevant murine model of sepsis performed by CLP.16 CLP clearly activated liver STAT3, which was determined via phosphorylation of STAT3 in control mice (Fig. 2A), in agreement with a previous report.18 Liver STAT3 activation during sepsis is mostly due to the activation of STAT3 in hepatocytes, because liver STAT3 was only marginally activated in L-STAT3 KO mice. CLP activated liver STAT1 both in L-STAT3 KO mice and wild-type mice, suggesting that the absence of STAT3 does not affect the activation of other STATs. Given that STAT3 is a well-known mediator for APP,19 we measured APPs such as fibrinogen and haptoglobin in plasma after CLP (Fig. 2B). The levels of fibrinogen and haptoglobin clearly increased after CLP in wild-type mice. In contrast, induction of fibrinogen was completely diminished in L-STAT3 KO mice, whereas that of haptoglobin was partially inhibited. This is consistent with the previous notion that fibrinogen is a class 2 gene and haptoglobin is a class 1 gene; the class 2 gene is predominantly regulated by type 2 IL-6 responsive elements binding to STAT20 and the class 1 gene by both type 1 IL-6 responsive elements binding to CCAAT enhancer-binding protein (C/EBP) and type 2 IL-6 responsive elements.21
To address the issue of whether hepatic STAT3 is involved in the outcome of CLP-induced lethality, we performed CLP blinded to the genetic background and checked the survival of the mice every 6 hours. L-STAT3 KO mice were significantly more vulnerable to CLP-induced lethality than wild-type littermates (Fig. 2C). To examine the possible difference in bacterial infection after CLP, we measured colony forming unit of blood bacteria 24 hours after CLP. Because there was no significant difference in bacterial amount between L-STAT3 KO mice and wild-type mice (Fig. 2D), we considered hepatic STAT3 to have had a beneficial effect on the outcome of septic shock without affecting bacterial infection.
Hepatic STAT3-Deficient Mice Show Exacerbated Liver Injury.
To examine liver injury and renal dysfunction in CLP-induced sepsis, we measured ALT and creatinine levels. L-STAT3 KO mice showed increased levels of serum ALT and creatinine compared with wild-type littermates, although the difference in creatinine did not reach a significant level (Fig. 3A). TUNEL of the liver revealed that the number of apoptotic hepatocytes was significantly higher in L-STAT3 KO mice than in wild-type littermates (Fig. 3B,C). However, the liver injury itself presumably is not a direct cause of animal death, because histologic abnormality was modest. Furthermore, LPS injection, which is another model of septic shock, induced more hepatocyte apoptosis than CLP but did not kill any mice tested (Fig. 3A-C and data not shown), supporting the idea that increased liver injury could not explain the increased lethality in L-STAT3 KO mice.
Exacerbated Systemic Inflammatory Response in L-STAT3 KO Mice.
Hypercytokinemia underlying systemic inflammatory response syndrome may play an important role in the development of multiple organ dysfunction syndrome and lethality.9 We measured several circulating cytokines and chemokines in septic mice and found that TNF-α, IFN-γ, IL-6, IL-10, monocyte chemoattractant protein-1 (MCP-1) and macrophage inflammatory protein-1β (MIP-1β) had clearly increased 24 hours after CLP in L-STAT3 KO mice. Of importance is the finding that the plasma levels of these cytokines and chemokines were significantly higher in L-STAT3 KO mice than in wild-type mice, although they did not differ before CLP. This result indicates that the increased lethality found in L-STAT3 KO mice is associated with hypercytokinemia (Fig. 4A). Although plasma insulin levels significantly increased 24 hours after CLP, there was no significant difference between L-STAT3 KO mice and wild-type mice, suggesting that insulin levels do not affect the difference in animal lethality (Supplementary Fig. 1).
Given that bacterial infection did not differ between L-STAT3 KO mice and wild-type mice, we examined the response of cytokine production upon endotoxin stimulation. To this end, we injected the same amount of LPS to L-STAT3 KO mice and control mice and measured circulating cytokines. LPS injection into L-STAT3 KO mice upregulated those cytokines to a lesser extent than CLP. In agreement with the finding on the CLP model, the levels of TNF-α, IL-10, MCP-1, and MIP-1β were significantly higher in L-STAT3 KO mice than in wild-type mice after LPS injection (Fig. 4B), indicating that L-STAT3 KO mice were highly sensitive to endotoxin and prone to show hypercytokinemia.
STAT3-Regulated Soluble Factors Produced by Hepatocytes Suppress Cytokine Production From Immune Cells.
To examine the underlying mechanisms of the hyperimmune response in L-STAT3 KO mice, we hypothesized that STAT3-mediated soluble factors from hepatocytes repress cytokine production from immune cells. We isolated hepatocytes from L-STAT3 KO mice and control mice and stimulated them with or without IL-6, collecting the conditional medium of hepatocytes. Wild-type hepatocytes displayed STAT3 activation in primary culture without stimulation, but the levels increased upon IL-6 exposure, whereas KO hepatocytes did not show any STAT3 activation (Fig. 5A). Consistent with this was the finding that the wild-type hepatocytes produced more haptoglobin than KO hepatocytes, even in the absence of IL-6 (Fig. 5B).
Next, we cultured RAW cells, a murine macrophage cell line, in the presence or absence of culture supernatant of hepatocytes. RAW cells produced TNF-α, IL-6, and IL-10 but not IFN-γ upon stimulation of LPS, and hepatocyte culture supernatant suppressed the production of these cytokines (Fig. 5C). Importantly, the suppression was significantly weaker in the presence of conditional medium of KO hepatocytes than in the presence of conditional medium of wild-type hepatocytes. Furthermore, murine primary splenocytes produced IFN-γ upon LPS stimulation, and the production was also suppressed in the presence of conditional medium of hepatocytes. Again, IFN-γ production was significantly higher in splenocytes cultured with KO hepatocyte supernatant than in those with wild-type hepatocyte supernatant (Fig. 5D). These data indicate that soluble substances from hepatocytes suppressed activation of immune cells, which was critically dependent on STAT3.
The present study clearly demonstrated that the absence of STAT3 in hepatocytes leads to high levels of circulating cytokines and increased mortality of CLP-induced septic mice without affecting bacterial infection. L-STAT3 KO mice produced high levels of cytokines when injected with LPS, confirming that the absence of STAT3 signaling within hepatocytes induces a hyperinflammatory response even if the extent of the input stimuli remains constant. This phenomenon is similar to a previous report of macrophage-specific disruption of STAT3 in which serum cytokines such as TNF-α, IL-6, and IL-10 increased upon LPS stimulation.22 In those mice, immune cells could not respond to IL-10, which potentially inhibit TNF-α production via STAT3 signaling, and thus produced high levels of TNF-α. Further study revealed those mice to be vulnerable to CLP-induced sepsis.23, 24 However, in our L-STAT3 KO mice, the levels of STAT3 in macrophage did not differ from control mice and produced the same amount of TNF-α in response to LPS (Fig. 1C-D). Thus, suppression of the inflammatory response in wild-type mice was critically dependent on hepatic STAT3 signaling. Indeed, in vitro analysis revealed that soluble factors from hepatocytes repress cytokine production from activated macrophage and splenocytes in a hepatic STAT3-dependent manner. Whereas research has established that STAT3 mediates a variety of effects on hepatocytes, including proliferation,5 apoptosis protection,6 and glucose metabolism,7 the present study reveals that hepatic STAT3 has an important extrahepatic effect. This effect is activated by a variety of cytokines produced from immune cells such as IL-6 but, in turn, suppresses immune cell activation via production of soluble factors, providing a negative feedback loop. Thus, the present study describes a role of hepatic STAT3 in maintaining host homeostasis by negatively regulating the immune system.
APPs are liver plasma proteins whose levels of expression are either positively or negatively regulated by cytokines during inflammation. It has been established that STAT3 regulates the expression of most, if not all, APPs in the liver.19 Consistent with this, L-STAT3 KO mice displayed impaired production of APPs in response to CLP (Fig. 2B). Some APPs such as C-reactive protein,25 serum amyloid P,26 and α2-macroglobulin27 have been shown to bind bacteria and to positively or negatively affect their eradication. Several reports also suggest that APPs exert proinflammatory as well as anti-inflammatory effects.25, 28 C-reactive protein binds to the phosphocholine of some foreign pathogens as well as phospholipid constituents of damaged cells and can activate the complement system, whereas the antioxidants haptoglobin and hemopexin protect against reactive oxygen species. Thus, each APP has a unique role in the complex mechanism controlling infection-induced inflammation. The L-STAT3 KO mice used in the present study offer a unique model for identifying the net effect of STAT3-regulated APPs during the septic condition. Our work has revealed that the most prominent effect of STAT3-regulated APPs is suppression of the hyperinflammatory response and lethality without an effect on bacterial infection. The soluble factors from hepatocytes that suppress cytokine production from immune cells are still unknown. Although there may be several substances involved in this phenomenon, one candidate might be haptoglobin, which was recently demonstrated to suppress TNF-α, IL-12, and IL-10 from human peripheral blood mononuclear cells in vitro.29 We also obtained a similar finding that RAW cells produced a lesser amount of TNF-α upon LPS stimulation in the presence of haptoglobin (Supplementary Fig. 2). Identification of these substances may have important therapeutic implications for controlling the hyperinflammatory condition. Further study is needed to clarify this point.
The liver is one of the target organs of sepsis-induced multiple organ dysfunction syndrome. Evidence for this comes from the fact that CLP mice or LPS mice showed liver injury as evidenced by increases in serum ALT and TUNEL-positive hepatocytes scattered in the liver lobule. Furthermore, L-STAT3 KO mice displayed more hepatocyte apoptosis in mice subjected to CLP or LPS injection. Previous research has indicated that the absence of hepatic STAT3 renders hepatocytes more vulnerable to Fas-mediated apoptosis.6 It is possible that STAT3-null hepatocytes are more vulnerable to apoptosis in the septic model. However, at the same time, L-STAT3 KO mice showed higher levels of proinflammatory cytokines such as TNF-α, which is a direct inducer of hepatocyte apoptosis. In our model, it is difficult to differentiate which contributed more to increased liver injury: the decrease in apoptosis resistance or the increase in proinflammatory cytokine. It can be said that the increase of proinflammatory cytokines is presumably one of the causes, but not a result, of liver injury. In addition, as discussed in the Results section, liver injury was relatively modest and probably not a direct cause of animal death.
In the present study, the lack of hepatic STAT3 caused increased mortality in CLP mice. Although we did not address the direct link between hypercytokinemia and animal death, accumulating evidence suggests that an increase in a variety of cytokines is involved in lethality in CLP mice. For example, it was shown that IL-6 plays an important role in the increased expression of the C5a receptor in the lung, liver, kidney, and heart during the development of sepsis in CLP mice and that interception of IL-6 leads to reduced expression of the C5a receptor and improved survival.30 In addition, enforced expression of the IL-6 gene in wild-type mice led to high mortality (unpublished data). TNF-α and other cytokines increase expression of inducible nitric oxide synthase, and increased production of nitric oxide causes further vascular instability and may also contribute to the direct myocardial depression that occurs in sepsis.31 Thus, dysregulation of cytokines may be harmful for host organs and is probably linked to animal death.
The present study revealed an important role of hepatocytes in repressing the hyperinflammatory response in pathologic conditions. This raises the possibility that hyperinflammation may be ill-controlled when liver function is severely impaired. Although sepsis itself is not a frequent cause of liver failure, it is a serious complication of acute or chronic liver failure. Systemic inflammatory response syndrome is an important determinant of prognosis in fulminant hepatitis.12 Sepsis originating from spontaneous bacterial peritonitis or renal infection is one of the causes of patient death with decompensated cirrhosis.13 In patients with limited function of the liver, possible impairment of STAT3-regulated hepatocyte function may be involved in their poor prognosis when complicated with severe inflammation. Careful liver-supporting therapy or early liver transplantation should be considered not only for maintaining liver function but also from the aspect of controlling dysregulated hyperinflammatory responses.
In conclusion, hepatic STAT3 represses systemic hyperinflammatory response by stimulating hepatic production of soluble substances that can attenuate immune cell overactivation and also improves host survival during septic condition. This sheds light on hepatocytic STAT3 as a negative regulator for immune cell overactivation and its role in host defense during systemic severe inflammation.