Clostridium difficile Toxins Influence Hepatocyte Protein Synthesis Through the Interleukin 1 Receptor

Hypothesis  Clostridium difficile toxins require interleukin 1 (IL-1) production or a functioning IL-1 receptor to elicit acute-phase protein production by murine hepatocytes.

Design  Experimental study.

Setting  Research laboratory at the DVA Medical Center, St Louis, Mo.

Cells Studied  Hepatocytes prepared from normal mice, from knockout mice deficient in IL-1 production due to loss of IL-1 converting enzyme, or from knockout mice deficient in the IL-1 p80 receptor.

Interventions  Cells were treated with lipopolysaccharide, a crude C difficile toxin extract, or purified C difficile toxins A or B for 24 hours in vitro, then radiolabeled with ³⁵ S methionine. Newly synthesized acute-phase proteins were identified by electrophoresis and autoradiography.

Main Outcome Measures  Synthesis of a 23-kd acute-phase protein in response to the various stimuli.

Results  Lipopolysaccharide, C difficile culture extract, and purified toxins A and B stimulated the synthesis of the 23-kd acute-phase protein by hepatocytes from normal mice and by hepatocytes from knockout mice deficient in the IL-1 converting enzyme. However, hepatocytes from knockout mice deficient in the IL-1 p80 receptor failed to produce this acute-phase protein when treated with the C difficile toxins, although they responded fully to lipopolysaccharide.

Conclusions  Stimulation of acute-phase protein synthesis by C difficile toxins does not require IL-1 production, but does require a functioning IL-1 p80 receptor. This suggests that some of the actions of these toxins are mediated by this receptor.

CLOSTRIDIUM DIFFICILE may produce fulminant infections through its 2 exotoxins, toxin A and toxin B.^1-10 Even though these infections are commonly localized to the colon, a significant systemic inflammatory response is usually apparent, characterized by high fever, neutrophilia, and the elaboration of acute-phase proteins.

In previous studies of the hepatic acute-phase response, we observed that many agents stimulated the production of a 23-kd acute-phase protein by murine hepatocytes. Synthesis of this protein was stimulated directly by lipopolysaccharide (LPS), by glucocorticoids, and by certain cytokines, particularly interleukin 1 (IL-1).^11-13 This protein, which is a member of the lipocalin family, is described here as the LPS-induced protein (LIP).¹⁴

In recent studies, we found that C difficile toxins also stimulated the synthesis of this protein. However, C difficile toxin–stimulated synthesis was inhibited by simultaneous treatment of the cultures with the IL-1 receptor antagonist (RA).¹⁵ This was markedly different from LPS-stimulated synthesis of LIP, which was not affected by IL-1 RA.¹³

The inhibitory effects of IL-1 RA on C difficile toxin–stimulated synthesis of this protein could be explained by at least 2 mechanisms. It was possible that IL-1 was produced by contaminating nonparenchymal cells within the culture system in response to the toxins, and that IL-1 RA simply inhibited the effects of this cytokine. This model does not indicate a direct interaction between the C difficile toxins and the hepatocytes. Alternatively, the IL-1 receptor might mediate the interaction of the C difficile toxins with the hepatocytes in some way. The inhibitory effects of IL-1 RA would then be due to its blockade of the IL-1 receptor. Previous experimental data indicated that the amounts of IL-1 produced by the C difficile toxin–stimulated hepatocyte cultures were essentially undetectable and that dexamethasone, a known inhibitor of cytokine production by proinflammatory cells, actually enhanced the effects of these toxins in the production of LIP.¹⁵ These data, while not definitive, were more consistent with the latter explanation of the inhibitory effects of IL-1 RA.

To better delineate the mechanism by which the C difficile toxins stimulate hepatocyte acute-phase protein synthesis, the present studies were undertaken. For these studies, we utilized hepatocytes from knockout mice deficient in either the IL-1 converting enzyme or in the IL-1 p80 receptor. The IL-1 converting enzyme knockout (ICE KO) mice can produce only small amounts of biologically active mouse interleukin (mIL-1α) and no mIL-1β, although their response to exogenous IL-1 is preserved.^16,17 In contrast, the IL-1 p80 receptor knockout (p80 KO) mice cannot respond to IL-1 or other agents requiring an intact IL-1 receptor.¹⁸ Thus, investigations using the hepatocytes from these knockout mice would allow us to determine whether C difficile– induced acute-phase protein synthesis depended on IL-1 production within the culture system or was mediated in part by the IL-1 receptor.

Female C57BL/6 mice were obtained from Harlan SD, Indianapolis, Ind. Female C57BL/6 ICE KO mice and p80 KO mice were supplied by James G. Norman, MD, University of South Florida, Tampa. The knockout event in these mice was confirmed by polymerase chain reaction analysis of tail DNA. All protocols for the use of these mice were approved by the Animal Care Committee at the DVA Medical Center, St Louis, Mo.

Hepatocytes were isolated from the livers of control and knockout mice using a modification of the method of Seglen¹⁹ as described previously,¹¹ except that the cells were not purified through a polyvinylpyrrolidone-colloidal silica (Percoll) cushion (Pharmacin, Piscataway, NJ). Primary cultures were established and maintained as described previously,¹¹ except that each culture was initially established with 30,000 live cells per microwell. After 48 hours, cultures were treated with a control medium or a medium containing various concentrations of LPS, murine mIL-1α, C difficile culture extract, or purified C difficile toxins A or B (TechLab, Blacksburg, Va). After 24 hours of treatment, the cultures were exposed to a medium containing ³⁵ S methionine to allow radiolabeling of proteins. The proteins secreted into the culture supernatants were separated by electrophoresis on 13.5% sodium dodecyl sulfate–polyacrylamide gels. After staining with Coomasie blue, the gels were destained, dried, and exposed to Kodak XAR-70 film (Eastman Kodak, Rochester, NY) for autoradiography.

Autoradiograms were scanned into a computer and digitized using NIH Image software (National Insitutes of Health, Bethesda, Md). The density of the band corresponding to LIP was then quantified and expressed as a percentage of the total density of the entire electrophoretic lane. Mean values obtained for LIP synthesis under different conditions were compared using the t test for unpaired data with significance set at P<.05. The Bonferroni adjustment was used to correct for multiple comparisons with the same control groups.

As had been observed in previous experiments using hepatocytes from Balb/C mice, both LPS and IL-1 stimulated LIP synthesis by hepatocytes obtained from normal female C57BL/6 mice (Table 1, Figure 1). The crude C difficile culture extract also stimulated LIP synthesis, with the stimulation being statistically significant for dilutions of 1:200,000, 1:20,000, and 1:2000 (Table 1). Very low concentrations of purified C difficile toxins A and B also appeared to induce hepatocyte LIP synthesis in the 2 experiments in which these purified toxins were used (Table 1). Thus, the response of hepatocytes from normal C57BL/6 mice was qualitatively and quantitatively quite similar to that previously observed using Balb/C hepatocytes.¹⁵

Hepatocytes obtained from the ICE KO mice, which cannot make biologically active IL-1 but can respond to exogenous IL-1, manifested similar responses to these mediators. Lipopolysaccharide, IL-1, the crude C difficile culture extract, and purified C difficile toxins A and B reproducibly stimulated LIP synthesis by hepatocytes from ICE KO mice (Table 2). Thus, production of IL-1 within the culture system did not appear to be required for C difficile toxin–induced stimulation of hepatocyte LIP synthesis.

Results were strikingly different when using hepatocytes obtained from the p80 KO mice, which lack the physiological receptor for IL-1. As expected, hepatocytes from these mice did not synthesize LIP when stimulated with IL-1 (Figure 2, Table 3). In addition, however, there was little or no stimulation of LIP synthesis observed when hepatocytes from these mice were treated with either the crude C difficile culture extract or purified toxins A or B, even at the highest concentrations of the toxins (Figure 2, Table 3). In fact, even basal levels of LIP synthesis appeared lower in hepatocytes from the p80 KO mice than in hepatocytes from control mice. However, hepatocytes from these mice were not refractory to all agents, since LPS stimulated LIP synthesis by these cells (Figure 2, Table 3). Thus the IL-1 p80 receptor seemed essential in mediating the effects of the C difficile toxins on LIP synthesis, but was not required to mediate the effect of LPS.

Clostridium difficile toxins produce a variety of effects on mammalian tissues and cells. Toxin A is generally described as an enterotoxin because of its capacity to induce secretion into isolated intestinal loops. Both toxins A and B are cytotoxic, but for most cell lines, toxin B is more potent.²⁰ Characteristically, both toxins cause cells to round up due to disruption of the actin cytoskeleton. This effect is probably mediated by various intracellular Rho proteins.^21-26 However, although cell rounding is used for clinical assays of C difficile toxins, it is unclear how much this phenomenon directly contributes to the disease processes elicited by these substances.

There has been some clinical and experimental evidence that IL-1 is involved in the pathogenesis of disease due to C difficile. High concentrations of mIL-1β have been observed in stool samples obtained from patients with C difficile colitis.²⁷ In addition, in experimental animal models, antibodies directed against mIL-1β have been shown to reduce C difficile toxin–induced migration of neutrophils into the peritoneal cavity and fluid secretion from intestinal mucosal preparations.^28,29 Finally, C difficile toxins have been found to stimulate IL-1 production by human monocytes.³⁰ However, we have not uncovered any reports implicating the IL-1 receptor in mediating the effects of the C difficile toxins.

There have actually been 2 IL-1 receptors identified. The type I (p80) receptor is an 80-kd glycoprotein originally identified on T cells and fibroblasts,³¹ whereas the type II (p60) receptor is a 60-kd glycoprotein originally identified on B lymphocytes.³² The p80 receptor has a 217–amino acid cytoplasmic domain, whereas the p60 receptor has only a very short, 29–amino acid cytoplasmic domain.^33,34 Most evidence indicates that the p80 receptor mediates the physiological responses of cells to mIL-1α and mIL-1β.^35-37 In contrast, the p60 receptor may serve as a decoy molecule for IL-1, and thereby modulate the inflammatory reaction.³⁸

Our previous investigations using murine hepatocytes have suggested that IL-1 or the IL-1 receptor was involved in mediating the effects of the C difficile toxins. Specifically, we found that the C difficile toxins were similar to LPS in stimulating synthesis of the 23-kd acute-phase protein, LIP.¹⁵ However, C difficile toxin–triggered synthesis of LIP was completely inhibited by IL-1 RA,¹⁵ which antagonizes the effects of IL-1 by binding to the IL-1 p80 receptor. In contrast, IL-1 RA did not interfere with LPS-stimulated LIP synthesis.^13,15

The capacity of IL-1 RA to interfere with the C difficile toxin–stimulated production of LIP could be due to several mechanisms. The simplest explanation was that the toxins triggered production of IL-1 by contaminating nonparenchymal cells within the culture system, and that it was this IL-1 that then induced the production of the LIP. Under this hypothesis, the C difficile toxins would alter hepatocyte function in an indirect manner only. However, an alternative explanation was that the C difficile toxins directly interacted with the hepatocyte. It was even possible that they acted through the IL-1 receptor itself.

Earlier experimental results provided some evidence against the hypothesis that IL-1 was required to mediate the effects of C difficile toxins on hepatocytes. We found that there was little IL-1 produced by cultured hepatocytes, even when using concentrations of the C difficile toxins that fully stimulated LIP synthesis.¹⁵ Further, it was observed that dexamethasone, which inhibits cytokine production by macrophages and other proinflammatory cells, did not interfere with C difficile toxin–stimulated synthesis of LIP.¹⁵ Nonetheless, these data only provided indirect support for the hypothesis that the hepatocyte IL-1 receptor was necessary for C difficile toxin–mediated induction of LIP synthesis.

We tested for the potential role of IL-1 or the IL-1 receptor in mediating the effects of the C difficile toxins. Different types of KO mice proved extremely valuable in studying the mechanism by which these toxins stimulate hepatocyte acute-phase protein synthesis. The ICE KO mice, due to the defect in ICE, produced only small amounts of biologically active mIL-1α and no mIL-1β.^16,17 Thus, if IL-1 were essential to mediate the effects of the C difficile toxins, it would be expected that the hepatocytes from these mice would not respond to the C difficile toxins. In contrast, the IL-1 p80 KO mice lack the receptor required for physiological responses to mIL-1α or mIL-1β. If this receptor was the one needed to mediate the effects of the C difficile toxins, hepatocytes from these mice would fail to respond to these toxins.

The results clearly demonstrated that it was the IL-1 receptor, and not mIL-1α or mIL-1β, that was required. The C difficile toxins stimulated LIP synthesis by hepatocytes from ICE KO mice, and the response was virtually identical to that observed using hepatocytes from normal mice. In contrast, hepatocytes from p80 KO mice proved incapable of synthesizing LIP in response to the C difficile toxins, although they could still respond to LPS. Thus, it would seem that the C difficile toxins act directly on the hepatocytes to induce LIP synthesis, and that the IL-1 p80 receptor is in some way required in this process.

The exact role of the IL-1 receptor cannot be determined directly from these experiments, however. It is possible that the C difficile toxins are recognized directly by the p80 receptor, and that this interaction initiates the cascade leading to LIP production. Thus, the C difficile toxins would be agonists for the p80 receptor, similar to mIL-1α and mIL-1β. There have been no previous reports, however, that the C difficile toxins can bind directly to the IL-1 receptor.

Alternatively, the IL-1 receptor may produce an intracellular message that serves as a cofactor for transduction of the signal associated with binding of the C difficile toxins to their membrane receptors. Inactivation of the IL-1 p80 receptor with IL-1 RA, or its loss, as seen in the IL-1 p80 KO mice, would thereby remove that necessary cofactor and prevent signal transduction. Further study is required to differentiate between these mechanisms.

Regardless of the actual mechanism, it is apparent that some activities of the C difficile toxins require the presence of a functioning IL-1 receptor. In addition to induction of the 23-kd acute-phase protein, we have observed that the IL-1 receptor is required for C difficile toxin–induced suppression of albumin synthesis in vitro (Mazuski et al, unpublished data, 1999), as well as for C difficile toxin–induced release of lactate dehydrogenase from hepatocytes.³⁹ However, we have also observed that cytoskeletal rearrangements induced by the C difficile toxins are not altered in hepatocytes from the p80 KO mice, indicating that some cytotoxic responses to these toxins are independent of the IL-1 receptor.³⁹

These observations lead us to speculate that the IL-1 receptor might play a protective role in the organism's response to the C difficile toxins. Our data suggest that the acute-phase responses to these toxins, which are mediated by the IL-1 receptor, are expressed at very low concentrations of the toxins (< 1 fmol). In contrast, the toxic effects of these substances require substantially greater concentrations. It is possible, then, that the IL-1 receptor enhances the capacity of very low concentrations of C difficile toxins to initiate the systemic acute-phase response, allowing host defenses to be mobilized before higher toxin concentrations, capable of producing overt toxicity, are actually produced. Whether such a speculation is correct, however, it seems clear that the mechanisms by which C difficile toxins produce alterations in cellular physiology are quite complex, and that for at least some cellular responses, the IL-1 p80 receptor plays a role in mediating the effects of the C difficile toxins.

Box Section Ref ID

Presented at the 19th Annual Meeting of the Surgical Infection Society, Seattle, Wash, April 30, 1999.

We would like to acknowledge the excellent technical assistance of Kim Tolman.

Corresponding author: John E. Mazuski, MD, PhD, Department of Surgery, Saint Louis University School of Medicine, 3635 Vista Ave at Grand Boulevard, PO Box 15250, St Louis, MO 63110-0250 (e-mail: mazuskim@slu.edu).