Gut-derived commensal bacterial products inhibit liver dendritic cell maturation by stimulating hepatic interleukin-6/signal transducer and activator of transcription 3 activity†
Article first published online: 12 OCT 2007
Copyright © 2007 American Association for the Study of Liver Diseases
Volume 46, Issue 6, pages 1946–1959, December 2007
How to Cite
Lunz, J. G., Specht, S. M., Murase, N., Isse, K. and Demetris, A. J. (2007), Gut-derived commensal bacterial products inhibit liver dendritic cell maturation by stimulating hepatic interleukin-6/signal transducer and activator of transcription 3 activity. Hepatology, 46: 1946–1959. doi: 10.1002/hep.21906
Potential conflict of interest: Nothing to report.
- Issue published online: 28 NOV 2007
- Article first published online: 12 OCT 2007
- Manuscript Accepted: 18 JUN 2007
- Manuscript Received: 16 JAN 2007
- National Institutes of Health. Grant Number: DK49615
Intraorgan dendritic cells (DCs) monitor the environment and help translate triggers of innate immunity into adaptive immune responses. Liver-based DCs are continually exposed, via gut-derived portal venous blood, to potential antigens and bacterial products that can trigger innate immunity. However, somehow the liver avoids a state of perpetual inflammation and protects central immune organs from overstimulation. In this study, we tested the hypothesis that hepatic interleukin-6 (IL-6)/signal transducer and activator of transcription 3 (STAT3) activity increases the activation/maturation threshold of hepatic DCs toward innate immune signals. The results show that the liver nuclear STAT3 activity is significantly higher than that of other organs and is IL-6–dependent. Hepatic DCs in normal IL-6 wild-type (IL-6+/+) mice are phenotypically and functionally less mature than DCs from IL-6–deficient (IL-6−/−) or STAT3-inhibited IL-6+/+ mice, as determined by surface marker expression, proinflammatory cytokine secretion, and allogeneic T-cell stimulation. IL-6+/+ liver DCs produce IL-6 in response to exposure to lipopolysaccharide (LPS) and cytidine phosphate guanosine oligonucleotides (CpG) but are resistant to maturation compared with IL-6−/− liver DCs. Conversely, exogenous IL-6 inhibits LPS-induced IL-6−/− liver DC maturation. IL-6/STAT3 signaling influences the liver DC expression of toll-like receptor 9 and IL-1 receptor associated kinase-M. The depletion of gut commensal bacteria in IL-6+/+ mice with oral antibiotics decreased portal blood endotoxin levels, lowered the expression of IL-6 and phospho-STAT3, and significantly increased liver DC maturation. Conclusion: Gut-derived bacterial products, by stimulating hepatic IL-6/STAT3 signaling, inhibit hepatic DC activation/maturation and thereby elevate the threshold needed for translating triggers of innate immunity into adaptive immune responses. Manipulating gut bacteria may therefore be an effective strategy for altering intrahepatic immune responses. (HEPATOLOGY 2007.)
Intraorgan dendritic cells (DCs) develop from bone marrow precursors and are potentially the most potent antigen-presenting cells. In normal solid organs, resident DCs are maintained in a relatively immature state. They sample the microenvironment for pathogen-associated molecular pattern (PAMP) danger signals via pattern recognition receptors and phagocytotic sampling of potential antigens. The detection of PAMPs induces DC maturation, which is marked by reduced phagocytosis, increased costimulatory and major histocompatibility complex (MHC) class II molecule expression, and chemokine-mediated migration to the lymph nodes, where they stimulate naive T-lymphocytes. Through this process, DCs translate innate immune signals into adaptive immune responses.
Portal blood accounts for 75% of the total hepatic blood flow, and it is rich in nutrients, hormones, potential antigens, and bacterial products, including stimulators of innate immunity.1 The liver filters portal blood of potentially immunogenic materials before it enters the general circulation and thereby protects the central immune organs from overstimulation. Resident sinusoidal macrophages (Kupffer cells) and specialized liver sinusoidal endothelial cells clear macromolecules.2 Hepatic DCs are less responsive to endotoxin exposure and have reduced T-cell stimulatory capacity compared with DCs from other organs.3–5
Potential mechanisms of liver DC resistance to maturation include lower expression of lipopolysaccharide (LPS)-binding toll-like receptor 4 (TLR4) in comparison with splenic DCs.4 The liver might also influence DC maturation by producing cytokines that inhibit DC maturation. For example, in vitro, liver sinusoidal endothelial cells and Kupffer cells secrete anti-inflammatory cytokines, including transforming growth factor β (TGFβ)6 and interleukin-10 (IL-10).7In vivo, liver nonparenchymal cells (NPCs), including Kupffer cells, constituently secrete IL-6.8, 9 However, the influence of these cytokines on liver DCs has not been directly tested.
IL-6/glycoprotein 130 (gp130) signaling is one of the most potent activators of signal transducer and activator of transcription 3 (STAT3),10 and this signaling pathway contributes importantly to hepatic pathophysiology.11 IL-6 can also inhibit DC maturation and function through the activation of STAT3,12–16 but the effect of gp130/STAT3 signaling on liver DCs has not been investigated. The microenvironment of tumors with elevated STAT3 activity suppresses DC responsiveness, resulting in immune ignorance.14, 17 This suggests that DCs in normal tissues with higher endogenous STAT3 activity, such as the liver, might be less responsive to maturational stimuli.
To test this hypothesis, hepatic DC maturation and function were studied in IL-6 wild-type (IL-6+/+), IL-6–deficient (IL-6−/−), and STAT3-inhibited IL-6+/+ mice. We show that hepatic STAT3 signaling is significantly dependent on IL-6 and is significantly higher than STAT3 signaling in other organs. Liver DCs from IL-6−/− and STAT3-inhibited IL-6+/+ are phenotypically and functionally more mature than hepatic DCs from IL-6+/+ mice. DC maturation in response to bacterial components is also significantly reduced in IL-6+/+ liver DCs. Conversely, the IL-6 treatment of IL-6−/− hepatic DCs can inhibit their maturation. An antibiotic treatment reduced portal blood endotoxin and hepatic phopho-STAT3 in IL-6+/+ mice, resulting in elevated liver DC maturation marker expression. These results implicate gut commensal bacteria and liver IL-6/STAT3 signaling as important mechanisms that raise the threshold for hepatic DC maturation and migration signals.
Materials and Methods
Male C57Bl/6J, Balb/cJ, or C57Bl/6J IL-6−/− mice (Jackson Laboratories, Bar Harbor, ME), 6-8 weeks old, were used for all studies. The STAT3 activity was blocked by a JSI-124 (EMD Biosciences, San Diego, CA) treatment at 1 μg/g of body weight intraperitoneally for 3 days.13 Antibiotic-treated mice received 250 mg/kg neomycin sulfate, 9 mg/kg polymyxin B, and 50 mg/kg metronidazole by gavage for 5 days. Control-treated mice received water only by gavage. Portal blood endotoxin was quantified with a limulus amebocyte lysate endochrome (Charles River, Charleston, SC). Animal experiments were performed under the guidelines of the University of Pittsburgh Institutional Animal Care and Use Committee.
STAT3 DNA Binding Activity.
Fifty micrograms of liver, lung, spleen, intestine, and heart nuclear protein from IL-6+/+ and IL-6−/− mice was assayed for the STAT3 DNA binding activity with a BD TransFactor STAT3 kit (Clonetech, Mountain View, CA) according to the manufacturer's protocol.
Cell Isolation and Culture.
Liver DCs were isolated from NPCs collected from collagenase perfused livers.18 Liver homogenates were passed through a 100-μm nylon filter, and hepatocytes were removed by centrifugation. NPCs were pelleted, resuspended in 4 mL of Roswell Park Memorial Institute 1640 medium supplemented with 10% fetal bovine serum, L-glutamine, sodium pyruvate, 2-mercaptoethanol, nonessential amino acids, and antibiotics (Sigma-Aldrich, St. Louis, MO) per milliliter of cells, and mixed with 7 mL of 30% Histodenz (Sigma-Aldrich) in phosphate-buffered saline. The suspension was overlaid with ice-cold phosphate-buffered saline and centrifuged at 1500g for 20 minutes. Anti-CD11c conjugated magnetic beads (Miltenyi Biotec Inc., Auburn, CA) were used to select DCs from the NPCs. Three to six livers were pooled for each experiment. More than 85% of the freshly isolated liver DCs expressed CD11c, as assessed by flow cytometry.
Splenic DCs were prepared from spleens injected with a collagenase solution, minced, and passed through a 70-μm cell strainer. CD11c+ DCs were selected as previously described.
Liver-derived DCs and bone marrow–derived dendritic cells (BMDCs) were generated under the influence of granulocyte-macrophage colony-stimulating factor (GM-CSF) or GM-CSF and IL-4, respectively, as previously reported.5 The nonadherent DCs were greater than 90% CD11c+ by flow cytometry (data not shown). DCs were incubated for 48 hours in the presence of 1 μM cytidine phosphate guanosine (CpG) oligodeoxynucleotide 1826 (Invivogen, San Diego, CA), 100 ng/mL LPS (Sigma-Aldrich), and 50 ng/mL mutein IL-6 (Imclone, New York, NY).
Messenger RNA (mRNA) Quantification.
The total RNA was isolated with the Trizol reagent (Invitrogen, Carlsbad, CA). Reverse transcription was performed on 1 μg of DNaseI-treated RNA by Superscript III (Invitrogen). Complementary DNA was amplified with the primers listed in Table 1 by a quantitative, real-time reverse-transcription polymerase chain reaction (RT-PCR) with Power SYBR green master mix (Applied Biosystems, Foster City, CA) and an ABI7500 fast real-time polymerase chain reaction system (Applied Biosystems). The gene expression is reported as a ratio of the target gene to glyceraldehyde 3-phosphate dehydrogenase (GAPDH).
|IL-10||CCA AGC CTT ATC GGA AAT GA||TTT TCA CAG GGG AGA AAT CG|
|TGFβ||AAA CGG AAG CGC ATC GAA||GGG ACT GGC GAG CCT TAG TT|
|TLR4||GCT TTC ACC TCT GCC TTC AC||AGC CTT CCT GGA TGA TGT TG|
|TLR9||ACT TCG TCC ACC TGT CCA AC||TCA TGT GGC AAG AGA AGT GC|
|IRAK-M||TGA GCA ACG GGA CGC TT||GAT TCG AAC GTG CCA GGA A|
|GAPDH||TGG CAA AGT GGA GAT TGT TGC C||AAG ATG GTG ATG GGC TTC CCG|
A flow cytometric analysis was performed on an LSR II cytometer (BD Biosciences, San Jose, CA). Cells were blocked with FcBlock (BD Biosciences) and stained with the antibodies listed in Table 2.
|Anti-IA/IE (MHC class II)||PE/AlexaFluor488||BD Biosciences||FC/IF|
|Anti–phospho-STAT3 (Tyr705)||Unconjugated||Cell Signaling||WB|
Mixed Lymphocyte Culture.
Freshly isolated liver and spleen DCs were analyzed for their T-cell stimulatory capacity by a mixed lymphocyte reaction (MLR). Nylon wool–purified T-cells from Balb/cJ mice were used as allogeneic responders. The T-cells (5 × 104) were incubated with γ-irradiated (2000 rad) liver or spleen DCs in triplicate for 3 days. 3H thymidine (1 μCi) was added in the final 18 hours, and the thymidine incorporation was analyzed.
Cell culture supernatants were analyzed with a ProteoPlex murine cytokine array (EMD Biosciences) according to the manufacturer's protocol.
Five-micrometer frozen liver sections were fixed in 96% ethanol and probed overnight with the antibodies listed in Table 2.
A statistical comparison of the groups was performed through a t test and an analysis of variance with SPSS version 11 software (SPSS Inc., Chicago, IL). A P value of 0.05 was considered significant.
Constitutive IL-6/STAT3 Signaling Is Higher in the Normal Liver Than in Other Organs.
The baseline hepatic STAT3 activity is dependent on IL-6 and contributes to liver pathophysiology.11 Because there are no comparisons of IL-6/nuclear STAT3 DNA binding activity in the liver and other solid organs, we conducted such an analysis for normal IL-6+/+ and IL-6−/− mice. In normal IL-6+/+ mice, the liver STAT3 activity was significantly greater than the STAT3 activity in the spleen, lungs, intestines, and heart (Fig. 1). The IL-6 status also influenced the STAT3 activity in all other organs, it being significantly higher in the lungs, intestines, and liver of IL-6+/+ mice in comparison with IL-6−/− mice. The absolute level of the STAT3 activity and the difference in the STAT3 activity between IL-6+/+ and IL-6−/− mice, however, were most dramatic in the liver.
The treatment of IL-6+/+ mice with JSI-124, a chemical inhibitor of STAT3,13 for 3 days significantly reduced STAT3 activity in the liver (Supplementary Fig. 1).
Liver DCs Are More Mature in IL-6−/− Mice Than IL-6+/+ Mice.
Elevated IL-6/STAT3 signaling inhibits DC maturation,12–16 whether the STAT3 activity is in the tissue microenvironment or in the DCs themselves. Therefore, we tested the hypothesis that IL-6/STAT3 signaling might be responsible for the reduced maturation of hepatic DCs observed in a normal liver by comparing IL-6+/+ and IL-6−/− mice. Maturation was assessed, in situ, by immunofluorescence staining and by flow cytometry of freshly isolated hepatic DCs.
As shown previously by many groups, the dendritic-shaped cells expressing CD11c, CD86, and MHC class II by immunofluorescence staining localized primarily in the portal tracts (Fig. 2A). Far fewer DCs were also present around the central veins and in the subcapsular regions (data not shown). In an IL-6+/+ liver, CD11c+ DCs coexpressed significantly less CD86 than IL-6−/− or STAT3-inhibited IL-6+/+ liver DCs (Fig. 2). MHC class II coexpression in CD11c+ DC was elevated in IL-6−/− and STAT3-inhibited IL-6+/+ livers versus IL-6+/+ livers, but this did not reach statistical significance (Fig. 2).
Flow cytometry studies of freshly isolated CD11c+ DCs from IL-6+/+ and IL-6−/− livers confirmed that IL-6+/+ liver DCs were less mature than IL-6−/− and STAT3-inhibited IL-6+/+ hepatic DCs. An analysis of costimulatory molecule expression in CD11c+ DCs confirmed that IL-6+/+ liver DCs had reduced maturation marker expression [CD80, CD86, and chemokine (C-C motif) receptor 7 (CCR7)] in comparison with liver DCs from IL-6−/− mice (Fig. 3A,B). MHC class II was also elevated in IL-6−/− liver DCs, but not significantly so. The reduced expression of CCR7 in IL-6+/+ liver DCs, the chemokine receptor facilitating CCL19-mediated and CCL21-mediated migration in mature DCs,16 suggested that IL-6/STAT3 signaling might inhibit the migration of liver DCs to the lymph nodes.
Liver CD11c+ DC subtypes, based on the expression of CD11b and CD8α (myeloid versus lymphoid) and plasmacytoid (CD11c+B220+), were also examined in IL-6+/+ and IL-6−/− mice. This analysis showed that most liver DCs were CD11b+CD8α− (myeloid), with minor populations of CD11b−CD8α+ (lymphoid) and CD11b−CD8α− (Supplementary Fig. 2A,B). The predominant liver DC subtype (CD11b+CD8α−; myeloid) showed the largest difference in maturation marker expression (CD80, CD86, and MHC class II) between IL-6+/+ and IL-6−/− liver DCs, but differences were detected across all subtypes examined (Supplementary Fig. 2C-E). Liver DCs isolated from IL-6−/− mice also showed a significantly higher percentage of CD11c+B220+ plasmacytoid DCs than IL-6+/+ liver DCs, and maturation marker expression was significantly higher in IL-6−/− plasmacytoid liver DCs (Supplementary Fig. 3).
IL-6/STAT3 signaling had a more pronounced effect on DC maturation marker expression in the liver DCs than in the spleen DCs. This is because STAT3 activity is significantly higher in the liver than in the spleen, and the difference in the STAT3 activity in the spleen between IL-6+/+ and IL-6−/− mice is not significant (Fig. 1). Consequently, splenic DCs showed minimal, nonsignificant differences in CD80, CD86, and MHC class II expression between IL-6+/+ and IL-6−/− mice (Fig. 3A,C). The CCR7 expression was marginally increased in IL-6−/− and STAT3-inhibited IL-6+/+ splenic DCs compared with IL-6+/+ splenic DCs.
IL-6/STAT3 Signaling Impairs Liver DC Function.
DCs initiate adaptive immune responses by directly stimulating T-cell activation in the liver or in regional lymph nodes after migration. DCs also indirectly influence T-cell activation through the secretion of proinflammatory and anti-inflammatory cytokines. We examined, therefore, liver DC cytokine secretion and T-cell stimulation to gauge their functional characteristics. Cytokine secretion was measured in isolated liver DCs cultured in the presence of GM-CSF. IL-6−/− and STAT3-inhibited IL-6+/+ liver DCs showed significantly more secretion of the proinflammatory cytokines IL-1α and tumor necrosis factor α (TNFα; Fig. 4 C,D). Conversely, IL-6+/+ liver DCs secreted significantly more IL-10 than IL-6−/− and STAT3-inhibited IL-6+/+ liver DCs (Fig. 4A). Even though STAT3-inhibited IL-6+/+ liver DCs are more mature than untreated IL-6+/+ liver DCs (Fig. 3), we observed no significant difference in the amount of IL-6 secreted (Fig. 4B). No detectable IL-4, IL-12, or interferon γ (IFNγ) was observed (data not shown), and this was consistent with previous reports of cytokine production from liver DCs kept in the presence of GM-CSF.20
The ability of DCs to stimulate T-cell activation was measured in an allogeneic mixed leukocyte reaction with CD11c+ liver DCs isolated from IL-6+/+, IL-6−/−, or STAT3-inhibited IL-6+/+ mice as stimulators and allogeneic splenic T-cells as responders. For all ratios of the stimulator to the responders tested, IL-6+/+ liver DCs showed significantly reduced T-cell stimulation compared with IL-6−/− or STAT3-inhibited IL-6+/+ liver DCs (Fig. 5A). The IL-6/STAT3 status had little effect when syngeneic T-cells were used as responders (data not shown).
For comparison, splenic DCs were also used as stimulators (Fig. 5B), and they were better T-cell stimulators than liver DCs, as expected. However, IL-6+/+ splenic DCs remained less potent T-cell stimulators than IL-6−/− or STAT3-inhibited IL-6+/+ splenic DCs.
IL-6 Status Does Not Influence Other Anti-Inflammatory Cytokines.
The aforementioned data suggest that the IL-6/STAT3 activity–rich milieu of the liver inhibits liver DC maturation. Other cytokines, however, including IL-10 and TGFβ, contribute to liver pathophysiology and are known to prevent DC maturation.21 We considered the possibility that the suppressive effect of IL-6 on DC maturation might be mediated through these cytokines and therefore measured their hepatic expression. No significant differences in whole liver IL-10 (Fig. 6A,B) or TGFβ (Fig. 6C,D) mRNA or protein expression were detected between IL-6+/+ and IL-6−/− mice. It is likely, therefore, that among these cytokines, IL-6 plays a major role inhibiting hepatic DC maturation in the normal liver.
Increased Threshold to TLR Stimulation in IL-6+/+ Liver DCs.
Because IL-6/STAT3 signaling significantly contributes to the physiological inhibition of hepatic DC maturation, we next determined whether the IL-6 status affects the sensitivity of liver DC maturation to exogenous maturation stimuli. Because the liver is continually bathed in bacterial components, we stimulated liver DCs with the bacterially derived TLR agonists LPS and CpG oligonucleotides.
Liver DCs were prepared from hepatic NPCs, cultured with GM-CSF, and compared to similarly prepared BMDCs.5 In agreement with previous reports,5 unstimulated liver DCs expressed less CD86 than BMDCs (Fig. 7A). IL-6−/− liver DC and BMDC cultures expressed more CD86 than IL-6+/+liver DC and BMDC cultures. The maturation of liver IL-6−/− DCs was less than that of IL-6−/− BMDCs, possibly reflecting the elevated basal STAT3 in the IL-6−/− liver (Fig. 1).
Following stimulation with either 100 ng/mL LPS or 1 μM CpG, liver DC and BMDC CD86 expression increased but was significantly less in IL-6+/+ liver DCs (Fig. 7A-C). Almost all IL-6−/− liver DCs expressed CD86 after LPS and CpG stimulation, whereas a significant population of IL-6+/+ liver DCs resisted maturation. Interestingly, IL-6−/− liver DCs were dramatically more sensitive to stimulation with CpG than IL-6+/+ liver DCs (Fig. 7A,B). A similar trend was also apparent between IL-6+/+ and IL-6−/− BMDCs after the CpG treatment, although the differences were not as striking as those in the liver DCs (Fig. 7A,C).
We speculated that elevated IL-6 secretion might account for the reduced maturation of IL-6+/+ liver DCs to CpG stimulation. The measurement of the IL-6 protein revealed that CpG-treated IL-6+/+ liver DCs secreted nearly 3 times more IL-6 protein than control-treated cells (Fig. 7D). In contrast, although LPS-treated IL-6+/+ liver DCs secreted significantly more IL-6 than control-treated cells, this response was significantly less than that of CpG-treated liver DCs. This shows that IL-6+/+ liver DCs in fact are responsive to the CpG treatment, as evident from increased IL-6 secretion, and this, in turn, prevents liver DC maturation.
The ability of IL-6 to prevent liver DC maturation was confirmed by the treatment of IL-6−/− liver–derived DCs with exogenous IL-6 prior to LPS exposure. IL-6−/− liver DCs exposed to IL-6 for 24 hours had a slight effect on CD86 expression in comparison with untreated cells (Fig. 7E). However, the IL-6 pretreatment of IL-6−/− liver DCs inhibited LPS-induced CD86 expression in comparison with DCs treated with LPS alone (Fig. 7E). Collectively, these results show that either endogenous or exogenous IL-6 can prevent DC maturation by TLR stimulation.
We also investigated whether the IL-6 status influenced the expression of LPS and CpG recognizing TLR4 and TLR9. No significant differences were seen in TLR4 and TLR9 mRNA between IL-6+/+ and IL-6−/− liver DCs (Fig. 8A,B). Although TLR4 protein expression was equal, TLR9 protein expression was significantly lower in IL-6+/+ liver DCs versus IL-6−/− liver DCs (Fig. 8C,D). Thus, the reduced TLR9 protein expression in IL-6+/+ liver DCs may also contribute to the reduced maturation.
Following LPS and/or CpG exposure, TLR signaling is negatively regulated by IL-1 receptor associated kinase-M (IRAK-M).22 Because of the continual hepatic exposure to endotoxin and the blunted maturation response of IL-6+/+ liver DCs to LPS and CpG, we investigated whether IL-6 influences IRAK-M expression in liver DCs. IL-6+/+ liver DCs (Fig. 8E) and whole liver tissue (Fig. 8F) expressed significantly more IRAK-M mRNA than IL-6−/− liver DCs and liver tissue. However, the expression of the IRAK-M protein revealed an opposite trend: IL-6−/− liver and liver DCs had equal or greater IRAK-M expression (data not shown). The discrepancy between the mRNA and protein results suggests a complex posttranslational regulation of IRAK-M, which we are currently investigating.
Gut-Derived Commensal Bacterial Endotoxin Inhibits Liver DC Maturation.
If gut-derived bacterial PAMPs are responsible for the IL-6/STAT3 signaling–mediated immaturity of liver DCs, the depletion of commensal bacteria by oral antibiotics should increase DC maturation in IL-6+/+ livers. IL-6+/+ and IL-6−/− mice were fed antibiotics or control water by gavage for 5 days. The antibiotic treatment significantly reduced the endotoxin levels in the portal blood (Fig. 9A), resulting in a significant decrease in the liver tissue IL-6 protein (Fig. 9B) and nuclear phospho-STAT3 (Fig. 9C) expression in IL-6+/+ liver tissue compared with control-treated control mice. Control-treated IL-6−/− liver tissue expressed little phosphorylated nuclear STAT3, as expected. Antibiotic-treated IL-6−/− mice, however, showed a significant increase in baseline phosphorylated STAT3 to levels comparable to those seen in an antibiotic-treated IL-6+/+ liver. Preliminary observations suggest that an alternative compensation mechanism may be operative in the IL-6−/− mice because after the antibiotic treatment, they also showed increased glycogen storage, which was absent in all other treatment groups (data not shown).
The decreased portal blood endotoxin, and hepatic IL-6 protein, activated STAT3 in antibiotic-treated IL-6+/+ mice resulted in a significant increase in CD80, CD86, and MHC class II expression in comparison with control-treated IL-6+/+ liver DCs (Fig. 9D). No differences were observed in antibiotic-treated or control-treated IL-6−/− liver DCs, even though STAT3 was elevated in an antibiotic-treated IL-6−/− liver. The antibiotic treatment did not alter the distribution or maturation in the CD11c+ liver DC subtype populations (data not shown). These data show that gut-derived bacterial products stimulate STAT3 in an IL-6–dependent manner, thereby inhibiting liver DC maturation.
In this study, we show that, under normal physiological circumstances, constitutive hepatic IL-6/STAT3 signaling stimulated by commensal gut bacteria contributes significantly toward the maintenance of hepatic DCs in a relatively immature state in comparison with DCs from other organs. Because the liver is continually bathed in bacterial components and potential antigens, a mechanism preventing liver DC maturation in response to these stimuli would be beneficial. Ironically, these studies show that the bacteria themselves, through endotoxin and CPGs, contribute significantly to hepatic DC hyporesponsiveness. The interdependence between gut bacteria and liver DCs raises the threshold needed for PAMP molecules to stimulate liver DC activation/maturation and thereby translate innate signaling into adaptive immunity.
IL-6 is generally considered to be a proinflammatory cytokine because it is elevated along with other proinflammatory cytokines, including TNFα, IL-1, and IFNγ, during early stages of inflammatory responses.23 IL-6 also inhibits the suppressive ability of regulatory T-cells,24 stimulates B-cell proliferation and the generation of plasma cells,10 and, in conjunction with TGFβ, drives the polarization of CD4+ T-cells to a Th17 lineage, which has been linked to autoimmunity.25 Transgenic mice overexpressing IL-6 are also more prone to develop plasmacytosis and exhibit autoimmune characteristics.26
In contrast, several lines of evidence suggest that IL-6 might also be considered an anti-inflammatory cytokine. For example, IL-6/STAT3 signaling inhibits macrophage and DC maturation.12–17 Also, IL-6−/− mice treated with an intraperitoneal injection of Escherichia coli and/or LPS have increased mortality,27 a deficient fever response,28 and a significantly increased TNFα and IFNγ cytokine response23 in comparison with IL-6+/+ mice. Additionally, granulomas formed after an intravenous Rhodococcus aurantiacus injection were larger and were associated with significant elevations of proinflammatory cytokines in IL-6−/− mice in comparison with IL-6+/+ controls.29 Collectively, these studies suggest that IL-6 can both facilitate the activation of adaptive cellular immunity and suppress innate immune responsiveness by acting in a spatiotemporal manner on each arm of the immune response. In this role, IL-6 appears to be involved in setting the threshold for communication between the 2 arms of the immune system.
IL-6 is a potent activator of hepatic STAT3 activity19 and is widely recognized as a participant in various aspects of hepatic pathophysiology, including liver regeneration,30 induction of the acute phase response,10 sepsis,31 and biliary tract barrier function and wound repair.11 The current results show that the baseline STAT3 activity in the liver is significantly higher than that in the spleen, intestines, lungs, and heart. Although the hepatic STAT3 activity is significantly lower in IL-6−/− mouse livers versus IL-6+/+ mouse livers, the liver STAT3 levels in IL-6−/− mice are still significantly higher than those of all other organs from both IL-6−/− and IL-6+/+ mice. Cytokines responsible for the maintenance of STAT3 activity in IL-6−/− livers are currently unknown.
Previous studies have shown that liver DCs are more immature than DCs from other solid organs,3 but aside from reduced TLR4 expression,4 the underlying mechanisms responsible for this immaturity are not fully understood. Diverse subtypes of liver DCs, including potentially hyporesponsive populations, may also contribute to the overall liver DC phenotype.20, 32 Our results, however, show that hepatic IL-6 contributes significantly to liver DC immaturity and hyporesponsiveness in all subtypes analyzed.
It has been speculated that anti-inflammatory cytokines such as TGFβ and IL-10 are important suppressors of hepatic DC maturation.21 We found, however, no differences in whole liver TGFβ and IL-10 mRNA or protein expression, suggesting that the anti-inflammatory effects of IL-6 are not due to the elevated expression of these cytokines. This is also consistent with reports showing that IL-10 does not compensate for the absence of IL-6 during LPS-mediated inflammation.23 Hepatic IL-6, therefore, is likely an important component of maintaining phenotypic immaturity in liver DCs under physiologic conditions.
We show that the IL-6 protein is detectable in IL-6+/+ liver tissue and is likely produced and used within the liver itself. Mouse blood sampled from the portal vein and inferior vena cava had similar IL-6 protein levels (data not shown). IL-6 mRNA is also readily detectable from freshly isolated liver mononuclear cells and microdissected Kupffer cells and is significantly higher than in monocytes from the spleen or Peyer's patches.8, 9 Bacterial LPS and CpG induce IL-6 expression in Kupffer cells33 and liver DCs.32 Antibiotic-treated IL-6+/+ mice with significantly reduced portal blood endotoxin have decreased hepatic IL-6 expression in comparison with control-treated mice (Fig. 9). This is consistent with data showing lower IL-6 mRNA in livers from gnotobiotic mice compared with specific-pathogen–free mice possessing normal gut flora.34
Lower portal blood endotoxin and hepatic IL-6 in antibiotic-treated IL-6+/+ mice resulted in a significant reduction of phospho-STAT3 expression and liver DCs that were more mature than control-treated IL-6+/+ liver DCs. Antibiotic-treated IL-6−/− mouse liver DC maturation did not differ from that of controls. Our results can be melded into a simple, yet elegant, physiological model in which elevated portal venous blood commensal gut bacterial products stimulate hepatic IL-6/STAT3 activity, which, in turn, inhibits liver DC maturation (Fig. 10). Elevated liver IL-6/STAT3 likely raises the threshold needed by a multitude of potential portal venous blood antigens and bacterial products to stimulate systemic adaptive immunity.
In vitro, the stimulation of IL-6+/+ liver DCs with either LPS or CpG resulted in dramatically less CD86 expression than the stimulation of IL-6−/− liver DCs. In contrast, a similar treatment of IL-6+/+ BMDCs resulted in significantly more maturation than the treatment of hepatic DCs. Although IL-6+/+ liver DCs resisted CpG-induced maturation marker expression, the CpG treatment did stimulate DCs to produce significantly more IL-6 than controls or LPS (Fig. 7D). Thus, liver DCs do respond to CpG stimulation with elevated IL-6 secretion, possibly through a TLR9-independent pathway, and this maintains DCs in a relatively immature state.
It will be important to determine whether qualitatively or quantitatively different molecular mechanisms are needed to counteract the inhibitory influence of IL-6/STAT3 signaling on DC maturation. This will help us to better understand how immune responses are initiated in the liver or after DCs have migrated to regional lymph nodes. IL-6/STAT3 appears to up-regulate intracellular negative regulators of TLR signaling, including IRAK-M. After the initial TLR agonist stimulation, IRAK-M inhibits TLR signaling by preventing the formation of IRAK-1/TNF receptor associated factor 6 complexes, leading to an endotoxin-tolerant state.22 Our results show that IRAK-M mRNA levels were significantly higher in IL-6+/+ liver DCs and liver tissue in comparison with IL-6−/− liver tissue. However, IRAK-M protein expression revealed an opposite trend. Because IL-6/STAT3 signaling has not been previously shown to influence the expression of IRAK-M, we are currently investigating the molecular mechanisms responsible for these observations.
Lastly, the influence of gut bacteria and IL-6/STAT3 signaling on DC allostimulation might help explain why long-surviving stable liver allografts recipients can be more readily weaned from immunosuppression35 and why septic allograft recipients that have elevated serum IL-627 can be withdrawn from all immunosuppression with a low risk of rejection.36 This study also raises the question of whether the manipulation of gut bacteria can be exploited therapeutically to either augment or inhibit intrahepatic immune responses.
Supplementary material for this article can be found on the H EPATOLOGY Web site ( http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html ).
|jws-hep.21906.pdf||1012K||Supporting Information file jws-hep.21906.pdf|
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.