Kupffer cell–derived interleukin 10 is responsible for impaired bacterial clearance in bile duct–ligated mice

Authors

  • Tetsuya Abe,

    Corresponding author
    1. Laboratory of Host Defense and Germfree Life, Research Institute for Disease Mechanism and Control, Nagoya University School of Medicine, Nagoya, Japan
    2. Division of Surgical Oncology, Department of Surgery, Nagoya University Graduate School of Medicine, Nagoya, Japan
    • Laboratory of Host Defense and Germfree Life, Research Institute for Disease Mechanism and Control, Nagoya University School of Medicine, 65 Tsurumai-cho, Showa-ku, Nagoya 466-8550, Japan
    Search for more papers by this author
    • fax: +81 52-744-2230

  • Toshiyuki Arai,

    1. Division of Surgical Oncology, Department of Surgery, Nagoya University Graduate School of Medicine, Nagoya, Japan
    Search for more papers by this author
  • Atsushi Ogawa,

    1. Division of Surgical Oncology, Department of Surgery, Nagoya University Graduate School of Medicine, Nagoya, Japan
    Search for more papers by this author
  • Takashi Hiromatsu,

    1. Division of Surgical Oncology, Department of Surgery, Nagoya University Graduate School of Medicine, Nagoya, Japan
    Search for more papers by this author
  • Akio Masuda,

    1. Laboratory of Host Defense and Germfree Life, Research Institute for Disease Mechanism and Control, Nagoya University School of Medicine, Nagoya, Japan
    Search for more papers by this author
  • Tetsuya Matsuguchi,

    1. Laboratory of Host Defense and Germfree Life, Research Institute for Disease Mechanism and Control, Nagoya University School of Medicine, Nagoya, Japan
    Search for more papers by this author
  • Yuji Nimura,

    1. Division of Surgical Oncology, Department of Surgery, Nagoya University Graduate School of Medicine, Nagoya, Japan
    Search for more papers by this author
  • Yasunobu Yoshikai

    1. Division of Host Defense, Research Center of Prevention of Infectious Diseases, Medical Institute of Bioregulation, Kyushu University, Fukuoka, Japan
    Search for more papers by this author

Abstract

Extrahepatic cholestasis often evokes liver injury with hepatocyte apoptosis, aberrant cytokine production, and—most importantly—postoperative septic complications. To clarify the involvement of aberrant cytokine production and hepatocyte apoptosis in impaired resistance to bacterial infection in obstructive cholestasis, C57BL/6 mice or Fas-mutated lpr mice were inoculated intraperitoneally with 107 colony-forming units of Escherichia coli 5 days after bile duct ligation (BDL) or sham celiotomy. Cytokine levels in sera, liver, and immune cells were assessed via enzyme-linked immunosorbent assay or real-time reverse-transcriptase polymerase chain reaction. BDL mice showed delayed clearance of E. coli in peritoneal cavity, liver, and spleen. Significantly higher levels of serum interleukin (IL) 10 with lower levels of IL-12p40 were observed in BDL mice following E. coli infection. Interferon γ production from liver lymphocytes in BDL mice was not increased after E. coli infection either at the transcriptional or protein level. Kupffer cells from BDL mice produced low levels of IL-12p40 and high levels of IL-10 in vitro in response to lipopolysaccharide derived from E. coli. In vivo administration of anti–IL-10 monoclonal antibody ameliorated the course of E. coli infection in BDL mice. Furthermore, BDL-lpr mice did not exhibit impairment in E. coli killing in association with little hepatic injury and a small amount of IL-10 production. In conclusion, increased IL-10 and reciprocally suppressed IL-12 production by Kupffer cells are responsible for deteriorated resistance to bacterial infection in BDL mice. Fas-mediated hepatocyte apoptosis in cholestasis may be involved in the predominant IL-10 production by Kupffer cells. (HEPATOLOGY 2004;40:414–423.)

The high incidence of perioperative infectious complications in patients with cholestasis is well documented.1–4 Dysfunction of phagocytes and bacterial translocation from the gut due to loss of mucosal integrity are believed to be responsible for septic complication in patients with obstructive jaundice.1, 5–8 It has been reported that proinflammatory cytokines, such as tumor necrosis factor α (TNF-α) and interleukin (IL) 6, are increased in sera without any exogenous stimuli in cholestatic conditions, suggesting that cholestasis evokes inflammatory reaction in the host.9, 10 It is also demonstrated that rats and mice with experimental obstructive jaundice produce higher levels of proinflammatory cytokines including TNF-α, IL-1, and IL-6 after lipopolysaccharide (LPS) injection compared with those without obstructive jaundice, and that cholestatic animals are susceptible to LPS-induced organ failure and mortality.11–13 However, involvement of anti-inflammatory cytokines in host resistance to bacterial infection in cholestasis or cytokine profile in response to exogenously administered viable bacteria in cholestatic animals remains to be elucidated.

IL-10, an anti-inflammatory cytokine, was first described as having an ability to protect mice from LPS-induced fatal shock by suppressing proinflammatory cytokine production, including TNF-α and interferon (IFN) γ.14 On the other hand, it has also been shown that IL-10 hampers host defense mechanisms against microbial infection by suppressing macrophage functions.15 These contrary findings suggest that homeostasis during bacterial infection is maintained through a delicate balance between pro- and anti-inflammatory cytokines. It would thus appear that IL-10 might be involved in the immune dysfunction in cholestasis.

It has been demonstrated that macrophages are made capable of producing IL-10 after engulfing apoptotic cells in general.16, 17 Because IL-10 is shown to inhibit apoptosis pathways in a variety of cells, including hepatocytes, IL-10 production may play an important role in terminating cell death, including apoptosis, thereby suppressing excessive inflammatory reaction.18 Bile duct ligation (BDL) evokes liver injury with hepatocyte apoptosis in mice.19 It has been demonstrated that liver sinusoidal cells such as Kupffer cells and endothelial cells remove apoptotic hepatocytes induced by various stimuli, including lead nitrate, cycloheximide, and ultraviolet radiation.20–23 From these findings, it is possible to speculate that Kupffer cells may become capable of producing IL-10 predominantly as a result of ingesting increased apoptotic hepatocytes in cholestatic liver.

The overall objectives of this study were to elucidate the underlying mechanisms for impaired bacterial clearance in cholestasis, focusing on pro- and anti-inflammatory cytokines and to determine if aberrant cytokine production in BDL mice after Escherichia coli infection is dependent on Fas-mediated apoptosis of hepatocytes. We found that Kupffer cells but not peritoneal macrophages produced a large amount of IL-10 after E. coli infection in mice with cholestasis and that predominant IL-10 production by Kupffer cells was associated with hepatocyte apoptosis. Our data may provide new insight into the pathogenesis of bacterial infection in cholestasis.

Abbreviations

BDL, bile duct ligation/bile duct–ligated; IL, interleukin; TNF-α, tumor necrosis factor α; LPS, lipopolysaccharide; IFN-γ, interferon γ; mAb, monoclonal antibody; TUNEL, terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling; HBSS, Hanks' balanced salt solution; ELISA, enzyme-linked immunosorbent assay; mRNA, messenger RNA.

Materials and Methods

Mice and Microorganisms.

Eight- to 10-week-old female C57BL/6 mice and lpr/lpr mice with nonfunctional Fas expression with C57BL/6 background were purchased from Japan SLC (Hamamatsu, Japan). All mouse experiments were approved by the University Committee on Animal Research and received humane care in accordance with National Institutes of Health publication 86-23 (Guide for the Care and Use of Laboratory Animals).

E. coli (ATCC No. 26) grown in Trypto-soya broth (Nissui, Tokyo, Japan) was washed repeatedly, resuspended in phosphate-buffered saline, and stored at −80°C. The concentration of bacteria was quantitated by plate counts.

Reagents.

LPS from E. coli (serotype B6: O26) was obtained from Sigma Aldrich (St. Louis, MO). 2.4G2 (anti–FcRII/III-specific monoclonal antibody [mAb], rat immunoglobulin G1, producing hybridoma) was obtained from American Type Culture Collection (Manassas, VA). Phycoerythrin-conjugated anti-B220 and anti-CD11b mAb, biotin-conjugated anti-Gr.1 and NK1.1 mAb, fluorescein isothiocyanate–conjugated anti-CD3 mAb, and Cy-chrome-conjugated streptavidin were purchased from PharMingen (San Diego, CA). Rat immunoglobulin G anti–mouse IL-10 mAb was purchased from R&D Systems, Inc. (Minneapolis, MN). Control isotype rat immunoglobulin G was purchased from Sigma.

Surgical Procedure.

After 7 days of acclimation, surgery was performed under sterile conditions. Mice were anesthetized via intraperitoneal pentobarbital injection (50 mg/kg). An abdominal midline incision was made, and the common bile duct was ligated and divided as described previously.2 Control animals underwent a sham procedure in which the common bile duct was exposed but not ligated.

Histological Studies.

Liver specimens were removed and fixed with 10% buffered formalin, paraffin-embedded, and stained with hematoxylin-eosin for light microscopic examination. In situ terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) assay was performed using an in situ apoptosis detection kit (Apoptag, Intergen, Purchase, NY). All steps were performed according to the instructions of the manufacturer.

Assay for Serum Bilirubin Levels and Alanine Aminotransferase Activity.

Serum total bilirubin levels were measured using a commercially available kit following the manufacturer's instructions (Sigma Dianostics Kit no. 550 for bilirubin). Serum alanine aminotransferase activity was determined using the aminotransferase test kit (Wako, Osaka, Japan).

Preparation of Peritoneal Macrophages.

Peritoneal exudate cells were obtained from peritoneal cavity via lavage with 3 mL of Hanks' balanced salt solution (HBSS). Peritoneal exudates were centrifuged at 110g for 5 minutes, washed twice, and resuspended at optimal concentrations in Dulbecco's Modified Eagle Medium (Gibco, Grand Island, NY) supplement with 10% fetal bovine serum. Peritoneal exudate cells were spread on plastic plates and incubated for 1 hour in a CO2 incubator at 37°C to obtain adherent cells.

Preparation of Kupffer Cells.

Kupffer cells were isolated from sham and BDL mice by collagenase digestion and differential centrifugation using Percoll (Pharmacia, Uppsala, Sweden) as described elsewhere24, 25 with slight modifications. Briefly, the liver was perfused in situ through the portal vein with Ca2+- and Mg2+-free phosphate-buffered saline containing 10 mM ethylenediaminetetraacetic acid at 37°C for 5 minutes. Subsequently perfusion was performed with HBSS containing 0.1% collagenase IV (Sigma) at 37°C for 5 minutes. After digestion, the liver was excised and the suspension was filtered. The filtrate was centrifuged twice at 50g at 4°C for 1 minute. The supernatant was collected and centrifuged at 300g for 5 minutes, and the pellet was resuspended with buffer. The cell suspension was then layered on top of a density cushion of 25%/50% discontinuous Percoll (Pharmacia) and centrifuged at 900g for 20 minutes to obtain the Kupffer cell fraction, followed by washing with the buffer again. Cells were plated in 6-cm plastic culture dishes (FALCON, Becton Dickinson, NJ) and cultured in RPMI 1640 medium (Gibco) supplemented with 10% fetal bovine serum and 10 mM hydroxyethylpiperazine-N-2 ethanesulfonic acid. After incubation for 30 minutes, nonadherent cells were removed, cold Ca24- and Mg2+- phosphate-buffered saline with 10 mM ethylenediaminetetraacetic acid was added, and the cells were put on ice for 40 minutes. After tapping the dish gently, the supernatant was collected and centrifuged at 300g for 5 minutes. The pellet was resuspended with 1 × 106 cells/mL in RPMI and immediately used. The purity and cell viability of Kupffer cells isoplated were more than 91% and 95% as assessed by phagocytotosis of latex beads and trypan blue exclusion, respectively (data not shown).

Preparation of Liver Lymphocytes.

Fresh liver was immediately perfused with sterile HBSS through the portal vein and then meshed with stainless steel mesh. After the coarse pieces were removed by centrifugation at 50g for 1 minute, the cell suspensions were again centrifuged, resuspended in 8 mL of 45% Percoll (Pharmacia), and layered on 5 mL of 66.6% Percoll. The gradients were centrifuged at 600g at 20°C for 20 minutes. Lymphocytes at the interface were harvested and washed twice with HBSS.

Bacterial Growth in Organs.

After infection, peritoneal exudates were obtained from the peritoneal cavity by lavage with 3 mL of HBSS. Serial dilutions of the exudate samples were plated to determine the number of viable bacteria. For the enumeration of viable bacteria in the liver, the liver was perfused with 8 mL of sterile HBSS to wash out bacteria in the blood vessels immediately after mice were bled. The liver and spleen were removed and separated into sterile Teflon-coated homogenizers (Asahi Techno Glass Co., Tokyo, Japan) containing 2 mL of cold phosphate-buffered saline. After each organ was homogenized thoroughly, the bacterial counts in the homogenates were established by plating serial 10-fold dilutions in sterile distilled water on tryptic soy agar (Nissui). Colonies were counted 24 hours after incubation at 37°C.

Serum and Peritoneal Lavage Fluid Cytokine Assays.

TNF-α, IL-12, IL-10, and IFN-γ levels in serum, peritoneal lavage fluid, and culture supernatants were determined via enzyme-linked immunosorbent assay (ELISA). ELISAs were performed using Genzyme mAb according to the manufacturer's instructions (Genzyme, Cambrigde, MA).

Flow Cytometry Analysis.

Peritoneal exudate cells were preincubated with a culture supernatant from 2.4 G2 to prevent nonspecific staining. For the identification of macrophages and polymorphonuclear cells, the cells were then stained with phycoerythrin-conjugated anti-CD11b mAb and biotinylated anti-Gr.1 mAb. For the identification of lymphocytes, the cells were stained with fluorescein isothiocyanate–conjugated anti-CD3 mAb, phycoerythrin-conjugated B220 mAb, and biotinylated anti-NK1.1 mAb. To detect biotin-conjugated mAb, cells were stained with Cy-Chrome–conjugated streptavidin. All incubation steps were performed at 4°C for 30 minutes. The stained cells were analyzed with a FACSCalibur flow cytometer (Becton Dickinson, San Jose, CA). The data were analyzed using FACSCalibur research software (Becton Dickinson).

Expression of IL-10, IL-12, TNF-α, and IFN-γ Genes in Liver Homogenates, Liver Lymphocytes, Kupffer Cells, and Peritoneal Macrophages.

Total RNA was extracted from liver homogenates, liver lymphocytes, Kupffer cells, and peritoneal macrophages using TRIzol reagent (Life Technologies, Rockville, MD). Complementary DNA was synthesized from 2 μg of total RNA by reverse transcription.26 Real-time polymerase chain reaction was performed with the SYBR Green PCR Master Mix and ABI PRISM 7700 Sequence Detection Systems (Applied Biosystems, Foster City, CA) according to the manufacturer's suggested protocol. The specific primers were as follows: IL-12p40 sense, 5′-CGTGCTCATGGCTGGTGCAAAG- 3′; IL-12p40 antisense, 5′-CTTCATCTGCAAGTTCTTGGGC- 3′; IL-10 sense, 5′-CCAGTTTTACCTGGTAGAAGTGATG- 3′; IL- 10 antisense, 5′-AACTCAGACGACCTGAGGTCCTGGATCTGT-3′; IFN-γ sense, 5′-AGCGGCTGACTGAACTCAGATTGTAG-3′; IFN-γ antisense, 5′-GTCACAGTTTTCAGCTGTATAGGG-3′; TNF-α sense, 5′-GGCAGGTCTACTTTGGAGTCATTGCCCC-3′; TNF-α antisense, 5′-ACATTCGAGGCTCCAGTGAATTCGG-3′; β-actin sense, 5′-TGGAATCCTGTGGCATCCATGAAAC-3′; and β-actin antisense, 5′-TAAAACGCAGCTCAGAACAGTCCG-3′.

In Vitro Cytokine Production of Peritoneal Macrophages and Kupffer Cells.

Purified peritoneal macrophages in Dulbecco's Modified Eagle Medium (2 × 106 /mL) or Kupffer cells in RPMI 1640 (1 × 106/mL) were harvested in 96-well culture plates and stimulated with LPS (0.1 μg/mL) for indicated times. Supernatants were harvested at 0, 3, 6, 10, and 16 hours. TNF-α, IL-12p40, and IL-10 concentrations of supernatants were measured using ELISA.

In vitro IFN-γ Production of Cultured Liver Lymphocytes.

Freshly isolated liver lymphocytes were harvested 6 hours after E. coli infection. Liver lymphocytes in RPMI (5 × 106/mL) were cultured in vitro for 24 hours without additional stimulation. Culture supernatants were harvested and analyzed for IFN-γ content using ELISA.

Statistical Analysis.

All data are presented as means ± SD. Data were analyzed for significance using Student's t test, and a Bonferroni correction was applied for multiple comparison. A value of P < .05 was considered statistically significant.

Results

Fas-Dependent Hepatocyte Apoptosis in Obstructive Jaundice.

Five days after BDL, serum bilirubin levels reached 13.2 ± 4.6 mg/dL and remained at the plateau thereafter in C57BL/6 mice. Histological examination revealed that liver in BDL mice exhibited infrequent focal necrosis and cellular infiltration with marked bile duct proliferation, whereas liver of sham-operated mice showed almost normal appearance (Fig. 1A and 1B). For all subsequent experiments, mice undergoing BDL for 5 days were used. Apoptotic cells were next identified using the TUNEL technique. After 5 days of BDL, hepatocytes undergoing apoptosis could be observed (Fig. 1B and 1E). In contrast, the number of TUNEL-positive cells remained low in control mice and BDL-lpr mice (Fig. 1D and 1F). These results suggest that BDL induces hepatocyte injury, including Fas-dependent apoptosis.

Figure 1.

Representative histological sections of the liver on the 5th postoperative day. (A–C) Hemotoxylin-eosin stain. (D–F) TUNEL stain. (A, D) Sham-operated mouse. (B, E) BDL mouse. (C, F) BDL-lpr mouse. Bile duct proliferation and cellular infiltration were seen in both wild type and lpr mice after BDL, whereas focal spotty necroses were observed only in wild type mice with BDL. Increased numbers of TUNEL-positive cells (arrowheads) were observed in the liver of BDL mice in contrast to sham-operated or BDL-lpr mice. Data shown are representative of 3 independent experiments. HE, hemotoxylin-eosin; TUNEL, terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling; BDL, bile duct ligation.

Increased Susceptibility of BDL Mice to E. coli Infection.

To evaluate bactericidal activity of BDL mice, we examined the kinetics of bacterial growth in peritoneal cavity, liver, and spleen after intraperitoneal inoculation with E. coli (1 × 107 colony-forming units/mouse). As shown in Fig. 2, the bacterial counts were significantly higher at any time point after E. coli infection in BDL mice than in sham-operated mice (P < .05). Because these results were consistent with previous reports,5, 27, 28 we concluded that bacterial killing is severely impaired in BDL mice.

Figure 2.

Delayed clearance of E. coli in liver, spleen, and peritoneal cavity in BDL mice. Data are representative of 4 separate experiments and are expressed as the mean ± SD for 5 mice in an experiment. (▪) BDL mice. (□) Sham-operated mice. CFU, colony-forming units.

Emergence of Peritoneal Exudate Cells After E. coli Infection in BDL Mice.

A prominent increase in polymorphonuclear cells, which are thought to be responsible for rapid elimination of the bacteria, was observed in the peritoneal cavity after an intraperitoneal infection with E. coli. To elucidate the cause for deteriorated exclusion of bacteria in BDL mice, we examined the influx of phagocytes in the peritoneal cavity after E. coli inoculation. There was no substantial difference in numbers of polymorphonuclear cells, lymphocytes, or macrophages in the peritoneal cavity between BDL and sham-operated mice at 6 hours after E. coli infection when a large difference in number of viable bacteria was seen (Fig. 3). The numbers of polymorphonuclear cells and macrophages were significantly larger at 24 hours in BDL mice after E. coli infection, presumably due to increased bacterial burden at this stage. These results indicate that accumulation of phagocytes is not impaired in BDL mice after E. coli infection.

Figure 3.

Kinetics of peritoneal exudate cells of sham-operated (□) and BDL (▪) mice after intraperitoneal inoculation with 107 colony-forming units of E. coli were analyzed with flow cytometry. Accumulation of immune cells after E. coli infection was not impaired in BDL mice. Data are representative of 3 separate experiments and are expressed as the mean ± SD for 4 mice in an experiment.

Aberrant Cytokine Production in BDL Mice After E. coli Challenge.

Cytokine production was examined in the sera of sham-operated and BDL mice after E. coli infection. As shown in Fig. 4A, serum TNF-α and IL-10 levels were maximal at 1 hour after E. coli infection, while the IL-12 level reached a peak at 3 hours after infection in both BDL and sham-operated mice. Serum IL-10 levels were significantly higher in BDL mice than in sham-operated mice, while increases in IL-12 levels were significantly suppressed in BDL mice (P < .05). There was no significant difference in serum TNF-α level between sham-operated and BDL mice. The patterns of cytokine profile of peritoneal lavage fluid were similar to those of the serum (Fig. 4B). Neither IL-4 nor IFN-γ was detected in the sera or peritoneal lavage fluids of either BDL or sham-operated mice at any stage after E. coli infection (data not shown). These results clearly indicated that mice with obstructive jaundice present predominant IL-10 production over IL-12 after E. coli infection.

Figure 4.

Kinetics of cytokine levels in (A) serum and (B) peritoneal lavage fluid after E. coli challenge. Plasma and peritoneal lavage fluid samples were obtained at the indicated time points after intraperitoneal inoculation with 107 colony-forming units of E. coli. TNF-α, IL-12, and IL-10 concentrations were determined via ELISA. Data are representative of 3 separate experiments and are expressed as the mean ± SD for 5 mice in an experiment. *P < .05 versus the sham-operated group. (▪) BDL mice. (□) Sham-operated mice. TNF-α, tumor necrosis factor α; IL, interleukin.

IFN-γ Production By Liver Lymphocytes of BDL Mice After E. coli Infection.

To further investigate the cytokine profiles in BDL mice, we next examined messenger RNA (mRNA) expression in the whole liver homogenates of BDL mice or sham-operated mice after E. coli infection. Consistent with serum cytokine levels, high expression of IL-10 mRNA together with low expression of IL-12p40 mRNA was seen in BDL mice. Notably, expression of IFN-γ mRNA was not increased in liver of BDL mice after E. coli infection (Fig. 5A).

Figure 5.

Cytokine gene expressions in (A) liver homogenates and (B) liver lymphocytes after E. coli infection and in (D) Kupffer cells and (E) peritoneal macrophages after LPS stimulation. Cytokine gene expressions were quantitated using real-time polymerase chain reaction. Cytokine mRNA expressions of liver homogenates and liver lymphocytes were pooled from 4 mice in each group after E. coli infection. Cytokine gene expressions of cultured Kupffer cells and peritoneal macrophages were pooled from 5 mice in each group before and after stimulation with LPS (0.1 μg/mL). The expression of mRNA levels for each cytokines were normalized as a ratio using β-actin mRNA as a housekeeping gene. (C) In vitro IFN-γ production of liver lymphocytes after E. coli infection in sham-operated and BDL mice. Liver lymphocytes were isolated from each group before and after E. coli infection and cultured in vitro for 24 hours without additional stimulation. Culture supernatants were analyzed for IFN-γ content via ELISA. IFN-γ was not detected in culture supernatants of sham-operated or BDL mice before infection. Results were derived from 6–8 mice per group. *P < .05. IL, interleukin; IFN-γ, interferon γ; mRNA, messenger RNA; BDL, bile duct ligation; TNF-α, tumor necrosis factor α;

Because IFN-γ is necessary for macrophages to be activated so that they may kill microorganisms, we next determine the quantitative difference in IFN-γ production by liver lymphocytes between BDL and sham-operated mice. The expression of IFN-γ mRNA remained suppressed in liver lymphocytes isolated from BDL mice at any time point after E. coli infection compared with those from sham-operated mice (Fig. 5B). Moreover, whereas liver lymphocytes isolated from sham-operated mice 6 hours after E. coli infection produced a high level of IFN-γ in vitro without additional stimulation, IFN-γ production was barely detectable in lymphocytes from BDL mice (Fig. 5C).

Differences in Cytokine Production Between Kupffer Cells and Peritoneal Macrophages of BDL Mice.

To seek the source of serum cytokines produced after E. coli inoculation, we compared cytokine production by Kupffer cells with that by peritoneal macrophages in response to LPS derived from E. coli in vitro. The peritoneal macrophages and Kupffer cells isolated from BDL or sham-operated mice were stimulated in vitro with LPS, and the concentrations of TNF-α, IL-12, and IL-10 in the culture supernatants were determined via ELISA. As shown in Fig. 6A, the peritoneal macrophages of BDL mice produced considerably higher levels of TNF-α but significantly lower levels of IL-10 3 hours after LPS stimulation than those of sham-operated mice. The level of IL-12p40 production was higher in the peritoneal macrophages of BDL mice at 10 hours and 16 hours after LPS stimulation compared with those of sham-operated mice. On the other hand, Kupffer cells from BDL mice, in response to LPS, produced a larger amount of IL-10 and a smaller amount of IL-12 compared with those from sham-operated mice (Fig. 6B). There was no difference in TNF-α production by Kupffer cells between BDL and sham-operated mice.

Figure 6.

Difference in cytokine profile between (A) peritoneal macrophages and (B) Kupffer cells in response to LPS in vitro. Kinetics of TNF-α, IL-12p40, and IL-10 production by peritoneal macrophages (2 × 106/mL) and Kupffer cells (1 × 106/mL) after stimulation with LPS (0.1 μg/mL) are shown. Culture supernatants were collected at the indicated time points. TNF-α, IL-12p40, and IL-10 concentrations were determined via ELISA. Each value represents the mean ± SD of 4 experiments. *P < .01 versus the sham-operated group. (▪) BDL mice. (□) Sham-operated mice. TNF-α, tumor necrosis factor α; IL, interleukin.

We further compared mRNA expression for IL-10, IL-12p40, and TNF-α in Kupffer cells and peritoneal macrophages between BDL and sham-operated mice after LPS stimulation. As shown in Fig. 5D and 5E, IL-10 mRNA was increased in Kupffer cells of BDL mice after LPS stimulation, whereas the peritoneal macrophages of BDL mice expressed less IL-10 mRNA than those of sham-operated mice. Increase in IL-12p40 mRNA was only marginal in Kupffer cells of BDL mice but prominent in the peritoneal macrophages of BDL mice after LPS stimulation. TNF-α mRNA increased more markedly in both Kupffer cells and peritoneal macrophages of BDL mice after LPS stimulation compared with those in sham-operated mice. These results indicate that peritoneal macrophages and Kupffer cells from BDL mice respond differently to LPS in terms of IL-10 and IL-12 production and that the cytokine profiles of peritoneal macrophages are similar to those of Kupffer cells in sham-operated mice.

Effect of IL-10 Neutralization on Resolution of E. coli Infection.

Because IL-10 is known to hamper the resolution of bacterial infection in mice, we next determined whether or not increased IL-10 production in BDL mice is responsible for the impaired host defense against E. coli infection. BDL mice were injected intraperitoneally with anti–IL-10 neutralizing mAb (200 μg/mouse) 2 hours before E. coli challenge, and the number of the bacteria in organs was examined 24 hours after infection. As shown in Fig. 7, impaired bactericidal activity was reversed by administration of anti–IL-10 mAb in BDL mice. These results suggest that early IL-10 production by Kupffer cells might be responsible for hampered resolution of E. coli infection in BDL mice.

Figure 7.

The effect of neutralization of IL-10 in the course of E. coli infection. Sham-operated and BDL mice were injected intraperitoneally with anti–IL-10 neutralizing mAb (200 μg/mouse) 2 hours before E. coli challenge, and the numbers of the bacteria in liver, spleen, and peritoneal cavity were examined 24 hours after infection. *P < .05; **P < .01. CFU, colony-forming units; IgG, immunoglobulin G; IL, interleukin; BDL, bile duct ligation.

IL-10 Production in Fas-Mutated BDL Mice After E. coli Infection.

As demonstrated above, we found that BDL induced predominant IL-10 production specifically by Kupffer cells in mice. We hypothesized that Kupffer cells in BDL mice have been changed to readily produce IL-10 as a result of ingesting apoptotic hepatocytes, because it is demonstrated that macrophages produce IL-10 after ingesting apoptotic cells to prevent unnecessary immune responses.16, 17 To investigate this possibility, Fas-mutated lpr/lpr (lpr) mice were used in the next experiment, because lpr mice are shown to be resistant to cholestatic liver injury, which partially involves Fas-dependent hepatocyte apoptosis. As shown in Fig. 1E, hepatocyte apoptosis was seen in liver of BDL mice, whereas few apoptotic hepatocytes were observed in liver of lpr mice undergoing BDL (Fig. 1F); this is consistent with previous reports.19, 29 There was no difference in serum total bilirubin levels between wild type and lpr mice after BDL, but serum alanine aminotransferase activity of lpr mice was significantly lower in lpr mice than in wild type mice (42.2 ± 12.4 IU/L vs. 85.1 ± 15.6 IU/L; P < .05). As expected, serum IL-10 level of BDL-lpr mice after E. coli infection was comparable to that of sham-operated mice (Fig. 8B). Surprisingly, the bacterial number in organs 24 hours after E. coli infection was significantly lower in BDL-lpr mice when compared with BDL wild type mice, whereas no significant difference in bacterial clearance was seen between sham and sham-lpr mice (Fig. 8A). These results indicate that Fas may be partially involved in increased IL-10 production by Kupffer cells and subsequent impairment in bacterial killing in BDL mice.

Figure 8.

Bacterial clearance in Fas-mutated lpr mice with BDL. (A) Bacterial clearances in organs 24 hours after E. coli challenge were counted in C57BL/6 and lpr mice. BDL-lpr mice did not show impaired bacterial clearance at the indicated time. Data are expressed as the mean ± SD for 5 mice. (B) Serum IL-10 production in lpr mice 1 hour after E. coli infection. Serum IL-10 level was lower in BDL-lpr mice compared with that in BDL mice. Results were derived from 5 mice per group. *P < .05. CFU, colony-forming units; BDL, bile duct ligation.

Discussion

Although bile duct ligation has been shown to induce impairment in bacterial clearance, only a few reports have addressed the bactericidal activity in cholestatic animals from a standpoint of pro- and anti-inflammatory cytokines.27, 28 We show here that increased IL-10 production with decreased IL-12 release in the serum following E. coli infection is characteristic of BDL mice as opposed to sham-operated mice. The early IL-10 production was potentially involved in impaired resolution of E. coli infection because in vivo administration of anti–IL-10 mAb significantly augmented bacterial clearance in BDL mice. We concluded that Kupffer cells were major sources of serum cytokines because these serum cytokine levels were well correlated with those produced by Kupffer cells but not peritoneal macrophages.

T cells and natural killer cells, in the presence of IL-12, initially produce IFN-γ after bacterial infection; later increase in IL-10 suppresses IFN-γ production by these lymphocytes, leading to subsidance of inflammatory reaction.14, 30 We have reported previously that IFN-γ is important for bacterial clearance after E. coli infection in mice.31 In fact, we have shown in the present study that liver lymphocytes of sham-operated mice expressed abundant levels of IFN-γ mRNA after E. coli infection, although the serum level of IFN-γ was not detected even in sham-operated mice. On the other hand, those from BDL mice expressed only a marginal level of IFN-γ mRNA. Furthermore, in vitro IFN-γ production was significantly higher in liver lymphocytes of sham-operated mice when compared with those of BDL mice. These data strongly suggest that the predominant IL-10 production and concomitant suppression of IL-12 production by Kupffer cells in BDL mice might be responsible for host defense dysfunction against bacterial infection.

We have demonstrated that there is no significant difference in serum TNF-α production between sham and BDL mice after E. coli infection. In contrast, previous papers reported that BDL mice produced a large amount of TNF-α after LPS stimulation.11, 32 This may reflect a difference of LPS versus whole bacterial challenge. In fact, we found in our experimental model that serum TNF-α levels of BDL mice were more than 10 times higher than those of sham-operated mice 1 hour after 4 mg/kg LPS injection (data not shown).

A notable finding in this study is a difference in ex vivo production of IL-10 and IL-12 in response to LPS between Kupffer cells and peritoneal macrophages from BDL mice. We found that peritoneal macrophages from BDL mice were able to produce a large amount of IL-12 and TNF-α, whereas IL-10 production was prominent in Kupffer cells from BDL mice. These results suggest that cytokine production by Kupffer cells and peritoneal macrophages are differentially regulated by their milieu. Although serum factors in cholestasis (e.g., increased concentrations of bile acids and bilirubin) are shown to affect functions of immune cells, it seems unlikely that such serum factors are involved in the differential regulation of Kupffer cells and peritoneal macrophages.1, 33–35

Kuppfer cells are shown to be activated by engulfing apoptotic hepatocytes induced by various stimuli.20–23 Because macrophages are demonstrated to become capable of producing IL-10 after engulfing apoptotic bodies,16, 17 it is possible to speculate that Kupffer cells in cholestatic mice are also activated to generate IL-10 predominantly after ingesting increased apoptotic hepatocytes induced by BDL. To evaluate this possibility, we conducted a series of experiments using Fas-mutated lpr mice, which are demonstrated to be resistant to Fas-mediated hepatocyte apoptosis. We have shown that lpr mice are resistant to BDL-induced hepatocyte apoptosis and that lpr mice with BDL are able to kill E. coli efficiently to a similar extent to that of sham-operated mice, with a small amount of IL-10 production. These results strongly support the scenario that BDL induces predominant IL-10 production by Kupffer cells through ingesting Fas-mediated apoptosis of hepatocytes. It has also been demonstrated that Fas signaling in macrophages induces IL-10 gene expression.36 We have recently reported that natural killer T cells in the liver express Fas-ligand directly through toll-like receptors induced by gram-negative bacteria such as Salmonella choleraesuis and E. coli.37, 38 Taken together, it also seems likely that Fas-expressing Kupffer cells in BDL mice are susceptible to signals via Fas-ligand expressed on natural killer T cells after E. coli infection and predominantly produce IL-10. Further studies are needed to clarify the mechanism of predominant IL-10 production by Kupffer cells in BDL mice.

In conclusion, increased IL-10 and reciprocally suppressed IL-12 production by Kupffer cells is responsible for deteriorated resistance to bacterial infection in BDL mice. Fas-mediated hepatocyte apoptosis may be involved in the predominant IL-10 production by Kupffer cells. These data support the clinical practice of biliary drainage before surgery to decrease perioperative septic complications. Moreover, our findings may provide a therapeutic approach to the control of cholestasis-associated bacterial infection.

Acknowledgements

The authors thank K. Itano, A. Nishikawa, and M. Yoshimura for their excellent technical assistance.

Ancillary