Effects of anti-inflammatory drugs on fever and neutrophilia induced by Clostridium difficile toxin B

This study investigated the ability of Clostridium difficile toxin B, isolated from the VPI 10463 strain, to induce fever and neutrophilia in rats. Intravenous injection of toxin B (0.005–0.5 μg/kg) evoked a dose-dependent increase in body temperature. The febrile response to 0.5 μg/kg of the toxin started in 2.5 h, peaked at 5 h, and subsided fully within 24 h. Toxin B also induced a dosedependent neutrophilia. Pretreatment with indomethacin (2 mg/kg, i.p.) did not affect the neutrophilia induced by toxin B, but significantly reduced the febrile response measured 4 to 8 h after toxin B injection. Dexamethasone (0.5 mg/ kg) also markedly diminished the febrile response induced by toxin B. These results show that Clostridium difficile toxin B induced a febrile response susceptible to inhibition by dexamethasone and indomethacin. Furthermore, they suggest that prostaglandins are not involved in the neutrophilia caused by this toxin.


Introduction
Fever is an important component of the acutephase response (APR), which is induced by exogenous pyrogens (bacterial products), and is related to their ability to stimulate cytokine production by host cells, e.g. macrophages, lymphocytes and endothelial cells. Many of these cytokines, including interleukin (IL)-I, IL-6, tumour necrosis factor-cz (TNF-z), interferons (IFNs), IL-8 and macrophage inflammatory protein-1 (MIP-1), generally known as endogenous pyrogens (EPs), can induce fever when injected into experimental animals and humans. [1][2][3][4][5] It is likely that IL-1, IL-6, TNF-x and IFNs induce fever by stimulating synthesis of prostaglandins (PGs). On the other hand, fevers induced by MIP-1 in rabbits and rats and IL-8 in rats are independent of prostaglandin synthesis. [2][3][4][5] Another important component of APR is neutrophilia, which is also dependent on synthesis and release of cytokines, including IL-Iz and [3, TNF-cz and IL-6, IL-8, as well other inflammatory mediators 6 11 such as C5a and PGs (El, Ei, Fzcz). Clostridium difficile, a Gram-positive anaerobic bacillus, is a common cause of diarrhoea associated with antibiotic therapy in elderly and 12 13 debilitated hospitalized patients.
The linical presentation of this infection is broad and ranges from an asymptomatic carriage to an acute abdomen and fulminant, life-threatening colitis.14,15 Clinical symptoms of C. difficile colitis are profuse and debilitating diarrhoea, abdominal pain and distension. Furthermore, systemic features may also be present, such as polymorphonuclear leukocytosis, nausea, anorexia, malaise, dehydration and fever. 14 This last sign is clinically relevant in pseudomembranous colitis associated with the infection by C. difficile. [16][17][18] Also, it has been suggested that, in postoperative patients who develop diarrhoea, the presence of an unexplained fever or high whim blood-cell count is an important clue to this. 9 Although C. difficile is non-invasive, its toxins seem to permeate the mucosal barrier since they can induce an antibody response by the host. i To date, two C. difficile exotoxins have been described, toxin A (TXA) and toxin B (TXB), each encoded by distinct genes. 21-24 TXA causes fluid accumulation associated with mucosal damage in several animal models. 25 In contrast, 26 28 TXB has no enterotoxic activity, but it is 1 000-fold more potent as a cytotoxin than is 26 27" TXA in tissue culture lines.
However, both toxins are potent proinflammatory agents that cause erythematous and haemorrhagic lesions in rabbit and guinea-pig skin. 26 '27 In addition, both TXA and TXB induce release of cytokines from human monocytes, polymorphonuclear neutrophils and cultured epithelial cells. [29][30][31] l A Cardoso et al.
To our knowledge there is no experimental the appropriate vehicle only. The injection of demonstration of the ability of toxins from C TXB was carried out between 09.00 and 10.00 h. di2ficile to induce fever and neutrophilia, although these signals occur with this infection. Leucocyte counts.. Six h after TXB injection, Hence, the aim of this study was to investigate animals were anaesthetized with pentobarbitone the pyrogenic activity and the ability to induce sodium (40 mg/kg, i.p.) and blood samples were neutrophilia of the C. difficile toxin B, in rats. In taken from the abdominal aorta for determinaaddition, we also investigated the effect of antition of total and differential cell counts. These inflammatory drugs in TXB-induced fever and are reported as the number of neutrophils per neutrophilia, ml of blood.

Materials and Methods
Drugs: Indomethacin was a gift from Merck, Sharp & Dohme, Brazil. Dexamethasone (Deca-Animals.. Wistar rats weighing 180-200g were dronal(R)) was purchased from Prodrome, Brazil. housed at 24 1C under a 12h light/dark cycle (lights on at 06.00 h), and had free access Statistical analysis: For analysis of the rectal temto water and food. perature data, the average baseline temperature before any injection was calculated for each Temperature measurements: Rectal temperature animal, and all subsequent temperatures were was measured by inserting a thermistor probe expressed as changes from this average value. All (Y.S.I. no. 402) 3 cm into the rectum. The values are reported as the mean S.E.M. The area animals were held manually in their cages during under the curve was calculated for each animal the temperature measurements. This procedure and expressed in arbitrary units as an index for was performed at least twice on the day before the magnitude of the febrile response over the the experiment to minimize stress-induced tem-period of measurements (fever index). The indiperature changes secondary to handling. On the vidual results were then combined to calculate the day of the experiment, the basal temperature of group mean _+ S.E.M. and were analysed for each animal was determined at least three times, statistical significance using one-way analysis of at 30 min intervals, before any injection. Only variance, followed by Tukey's test. The limit for animals with basal temperatures between 36.8 the level of significance was set at p < 0.05. and 37.6C were used. After the injection of the TXB or saline, the temperature of the rats was Results measured up to 6 or 8 h, at 30 min intervals. The experiments were conducted at the thermo- Figure 1 shows that i.v. injection of C. dijlficile neutral zone for rats ( Fig. 2). injection of TXB. Dexamethasone (0.5 mg/kg) Pretreatment of rats with dexamethasone (0.5 was administered subcutaneously (s.c.) I h mg/kg, s.c.) i h before TXB administration markbefore TXB. Control animals were treated with edly attenuated the febrile response induced by this toxin (0.5 btg/kg, i.v.; Fig. 3). In the whole time-course of that response there was no difference between this group and saline treated animals. Indomethacin (2 mg/kg, i.p.) injected 30 min before the TXB significantly reduced the febrile response measured between 4 and 8h after TXB injection (Fig. 4), but did not alter the neutrophilia (control, 5 The mortality in animals that received 0.5 l.tg/kg of TXB was 10.5% within 8 h, and was not modified by pretreatment with indomethacin or dexamethasone. The injection of 2.5 l.tg/kg of TXB (five times higher than the highest dose considered here) killed all animals within 4 h (n 4).

Discussion
This study shows that intravenous injection of TXB from Clostridium difficile induces dosedependent increases in rectal temperature and neutrophilia, thus mimicking two important components of APR. To our knowledge this study represents the first demonstration of the ability of C. difficile toxins to induce fever. Toxin B, at a dose of 2.5 btg/kg, induced 100% mortality in rats within 4h of injection. A significant level of mortality induced by TXB was also found by Lyerly et al., 27 where doses of 1 to 5 l.tg/animal of this toxin resulted in the death of about 50% of infant mice. This mortality may well be related to the ability of TXB to induce release of cytokines, such as IL-la, IL-I, IL-6, and TNF, 29 which are key mediators of APR 4 and septic shock. 5 Also, TXB is cytotoxic for a variety of cells, e.g. monocytes and colonic mucosal epithelial cells, 29'36 which also might contribute to the mortality caused by this toxin.
In the present study, dexamethasone markedly reduced TXB-induced fever. It is well known that dexamethasone inhibits febrile responses to a variety of pyrogenic stimuli, among them polyinosinic:polycytidilic acid, lipopolysaccharide from Gram-negative bacteria, as well as different cytokines. 4'37 This synthetic glucocorticoid has a wide range of activity and may exert its antipyretic effects by reducing the synthesis and secretion of pyrogenic cytokines and eicosanoids. It is well described that glucocorticoids suppress the cyto- TNF, 9 IL-6 and IL-8, 4 all of which are putative mediators of the febrile response. Furthermore, the inhibition of eicosanoid synthesis by glucocorticoids results from the inhibition of expression of the inducible cyclooxygenase isoenzyme and of mobilization of membrane phospholipids via synthesis of the phospholipase A2 inhibitory 41 42 lipocortins. Thus, the susceptibility of TXBinduced fever to inhibition by dexamethasone suggests that the mechanisms involved in this response might not be different from those triggered by other stimuli.
We have also demonstrated that indomethacin inhibits the febrile response, but not the neutrophilia, induced by TXB. The anti-inflammatory activities of non-steroidal anti-inflammatory drugs have been formerly ascribed to their ability to 186 Mediators of Inflammation Vol 5 1996 block PGs synthesis. 4 Nevertheless, besides their ability to block cyclooxygenase, new insights into their possible alternative mechanisms of action have emerged. In this regard, the antipyretic actions of indomethacin and sodium salicylate have been linked to their ability to induce release of arginine vasopressin (agP) 44-46 which is considered an endogenous cryogen. 47 6 In our hands indomethacin at the same dose significantly reduced the neutrophilia induced by endotoxin from E. coli (unpublished data). The true importance of each endogenous mediator to promote fever has been a point of intense debate. Since PGs do not seem to mediate TXB-induced neutrophilia, it is possible that the efficacy of indomethacin to inhibit the fever induced by this toxin may also involve mechanisms unrelated to cyclooxygenase blockade, such as AVP release. Further studies are needed to help clarify this point.
In conclusion, we report that the intravenous injection of TXB from C difficile induced fever in rats by a mechanism which was susceptible to inhibition by both dexamethasone and indomethacin. On the other hand, the neutrophilia induced by TXB was not inhibited by indomethacin. Therefore, TXB may be a useful tool for the understanding of the febrile response, as well as other events of APR.