Potentiating Effect of Bile on Enterotoxin-Induced Diarrhea

Transcription

1 INFECTION AND IMMUNITY, Feb. 1982, p /2391-5$2./ Vol. 35, No. 2 Potentiating Effect of Bile on Enterotoxin-Induced Diarrhea STEFAN LANGE'* AND IVAR LONNROTH2 Department of Clinical Bacteriology' and Institute of General Microbiology,2 University of Goteborg, Goteborg, Sweden Received 25 June 1981/Accepted 22 September 1981 The influence of bile acids on adenosine 3',5'-phosphate-induced intestinal secretion was studied in mice. Bile flow was stopped by ligation of the common bile duct, and secretion was induced in ligated loops of the small intestine. The decrease of bile led to inhibition of hypersecretion after challenge with heat-labile enterotoxins from Vibrio cholerae and Escherichia coli, as well as with prostaglandin E1. In contrast, the fluid response induced by dibutyryl-adenosine 3',5'- phosphate was unaffected by intestinal bile. Injection of bile or bile acids into intestinal loops before cholera toxin challenge restored the toxin-induced secretion in the bile-depleted intestine. At the subcellular level the decrease of intestinal bile led to inhibition of cholera toxin-activated adenylate cyclase, whereas the bile concentration did not influence the binding of 1251-labeled toxin to the intestinal epithelial cells. The results suggest that intestinal bile interacts with adenylate cyclase in the induction of fluid secretion by enterotoxins and prostaglandin E1. Diarrhea is frequently caused by bacterial enterotoxins which induce hypersecretion in the small intestine (13). Two of these toxins, the heat-labile enterotoxins from Vibrio cholerae and Escherichia coli, have been analyzed (14) and found to be neutral proteins causing intestinal hypersecretion of fluid and electrolytes by activation of the adenosine 3',5'-phosphate (cyclic AMP)-adenylate cyclase system in the mucosal epithelium (2, 7, 14). Enterotoxin present on the intestine does not ubiquitously induce diarrhea, however, as enterotoxigenic E. coli and V. cholerae are frequently found in healthy persons (13). This difference in propensity for enterotoxin in humans may reflect the absence of factors which participate in the toxin-mediated stimulation of fluid secretion. Prostaglandins have been suggested to play such a role on the basis of their capacity to mimic the effects produced by enterotoxins (6). No correlation, however, between the intestinal levels of these hormones and the propensity for enterotoxic diarrhea has been reported. Secretion with a concomitant increase of cyclic AMP stimulation has also been achieved in rats by increasing the level of certain bile acids in the rat colon (1, 12). The aim of the present study was to evaluate the role of bile on intestinal hypersecretion induced by the heat-labile enterotoxins from V. cholerae and E. coli in a mouse model. The level of intestinal bile was decreased by ligation of the common bile duct, whereafter the effects on toxin-induced fluid secretion were studied. MATERIALS AND METHODS Experimental design. Inbred 6- to 12-week-old C57 BI/6J (H-2b) male mice were used. The animals were deprived of food, but not water, for 24 h before use. Bile outflow in the duodenum was stopped by ligation of the common bile duct (CD). Intestinal challenge with the various secretagogues was performed on ether-anesthetized mice using an intestinal loop model. Secretion in the loop was determined together with the activity of adenylate cyclase in mucosal scrapings at the optimal time for fluid accumulation. In some experiments diluted bile or 2.5 mm bile acids were injected into the intestinal loop before or simultaneously with cholera toxin challenge in CD-ligated animals. "25I-labeled cholera toxin was used to test the toxin binding to the intestinal mucosal epithelium. Intestnal secretion. Intestinal fluid secretions were studied by using a ligated small intestine loop assay (9). After anesthesia the abdomen was opened, and one or two 4- to 6-cm-long intestinal loops were then prepared without prior rinsing of the intestine. When two loops were used, the proximal one was placed about 4 cm from the pylorus, with a 3- to 4- cm interval to the second loop, placed as far proximal on the ileum as possible. In the case of only one loop, it was positioned in the middle of the small intestine, i.e., on the distal part of the jejunum. A volume of.2 ml of the diarrhogenic agent, dissolved in phosphatebuffered isotonic saline (PBS), ph 7.2, was injected into the loop, and the abdomen was closed with agraffs. The animals were killed at the appropriate time for maximal fluid accumulation. This was 9 min after challenge with purified prostaglandin E1 (Sigma Chemical Co., St. Louis, Mo.) or N6-9dibutyrylcyclic AMP (DB-cAMP; Sigma) and 4 h after challenge with purified cholera toxin (Sigma) or heat-labile E. 391

2 392 LANGE AND LONNROTH coli enterotoxin (prepared from strain as described in reference 15). The dose used for enterotoxin challenge, as well as challenge with prostaglandin E1 (15,ug) and DBcAMP (2.5 mg), was selected to give 9% of the fluid response which was maximal for the agent in question. The animal was killed, the loop with accumulated fluid was free dissected, and the length and weight were measured. The fluid accumulation was calculated by subtracting the weight (milligrams per centimeter) of the non-manipulated intestine from that of the test loop. Initial experiments showed that the use of a PBSinjected control loop was unnecessary, as the weights of the PBS-injected loop and the non-manipulated intestine were the same. Operation procedure. In some experiments bile or bile acids were injected before or simultaneously with cholera toxin into the intestinal loops of CD-ligated mice. These experiments required opening and closing of the abdominal wall more than once, and subsequently agraffs were used. Control mice were sham operated. The sham operation was performed at the same time as the CD ligation. The abdominal wall was opened, and a ligature was placed in the omentum without affecting blood or lymphatic vessels. The ligature was placed at a distance of about 1 mm from the entrance of the bile duct into the intestine. When bile was injected into loops of experimental mice, loops were also prepared in the control groups and injected with PBS. Intestinal bile. Decrease of intestinal bile was induced by ligation of the CD. This operation was performed via a small laparatomy, and a ligature was placed a few millimeters proximal to the CD entrance into the duodenum. Bile was collected from mice by aspirating from the gall bladder. Cholic acid and deoxycholic acid were obtained from Sigma. Adenylate cyclase. The above-described free-dissected intestinal loops were cut open, and the mucosa was carefully scraped off the muscularis with a scalpel. After suspension of 5 mg (wet weight) of tissue per ml in a 75 mm Tris-hydrochloride buffer (ph 7.5, containing 12 mm MgCl2 and 1 mm dithiothreitol), the mucosal preparation was homogenized at C in a loosely fitting homogenizer and centrifuged (85 x g, 2 min). The assay of adenylate cyclase activity was performed by the procedure of Krishna et al. (8), with a previously described modification (1). Evaluation of 1"'I-labeled cholera toxin binding to the intestinal mucosa. Binding of cholera toxin to the intestinal mucosa was tested by using "M'I-labeled cholera toxin as described previously (5). An intestinal loop was injected with 2.5.ag of toxin (specific activity,.1,ci/,ug). After the animal was sacrificed, the loop was carefully free dissected, cut open, and washed in three successive solutions with PBS-.5% Tween 2. The quality of the labeled toxin was tested by mixing it with various proportions of the unlabeled toxin, and the fraction of radioactivity bound to mouse thymus cells in suspensions was measured (11). No difference in the binding or the labeled and unlabeled toxin was observed. Statistics. Values given are the mean ± standard error of the mean of 5 to 1 animals in each group. The significance of the difference in intestinal loop fluid INFECT. IMMUN. accumulation between experimental and control groups was calculated with Student's t test. RESULTS Intestinal secretion induced by enterotoxins. Cholera toxin induced a fluid secretion which was more pronounced in the jejunum than in the ileum (Table 1). Ligation of the CD for 24 h inhibited secretion in all portions of the small intestine. A slight inhibition of the secretion was observed 2 h after CD ligation, indicating a rapid turnover of bile in the small intestine. A toxic effect on the mice was not observed until after more than 48 h of CD ligation. Inhibited secretion after CD ligation was also observed when mice were challenged with E. coli enterotoxin (Table 1). The secretory response to the E. coli toxin was even more prone to inhibition than that after challenge with cholera toxin. The effect of CD ligation on the fluid response to various doses of cholera toxin was also tested. The results illustrated in Fig. 1 show that the degree of inhibition was highly dependent on the dose of toxin; 9% inhibition was noted with a challenge dose of 1.5,ug of cholera toxin, whereas no difference was seen with a toxin dose of 1,ug. Addition of bile and bile acids. Table 2 shows the effect of adding bile or bile acids into the intestine after CD ligation. A complete normalization of the secretory capacity to cholera toxin was noted when diluted bile or 2.5 mm bile acids were injected into the loop 4 h before toxin challenge. No effect was observed when these substances were injected simultaneously with cholera toxin. Challenge with 1-fold-diluted bile or 2.5 mm bile acids in loops of control mice did not induce intestinal secretion per se, whereas undiluted bile and 1 mm deoxycholate caused fluid secretion (data not shown). When cholate was tested in a concentration of 2.5 to 15. mm, no secretory response was induced. Binding of 125I-labeled cholera toxin to intestinal mucosa. Binding of 125I-labeled cholera toxin to the intestinal mucosa of control and CDligated mice is shown in Table 3. It can be seen that the decreased level of intestinal bile after 24 h of CD ligation did not diminish, but rather increased the attachment of cholera toxin to the intestinal mucosa. Intestinal secretion induced by prostaglandin El and DB-cAMP. It is generally assumed that prostaglandin E1 induces intestinal secretion by activating of the adenylate cyclase system. In contrast, DB-cAMP bypasses this enzyme and causes secretion through a rise of the intracellular cyclic AMP itself. To test the effect of low bile concentration on intestinal adenylate cyclase by decreased bile, the secretory responses to prostaglandin E1 and DB-cAMP challenge

3 VOL. V39FFECT 35, 1982 OF BILE ON ENTEROTOXIN-INDUCED DIARRHEA 393 TABLE 1. Inhibitory effect of ligation of the common bile duct on the intestinal secretory response to enterotoxin from V. cholerae and E. coli Challenge Bile duct ligation Weight of intestinal loops (mg/cm) time (h) Proximal jejunum Distal jejunum Ileum Cholera toxina 168 ± ± ± ± ± 6 68 ± ± ± t ± 5 52± 4 41±8 E. coli enterotoxinb 164 ± ± ± ± 14 a Cholera toxin (2.5,ug) in.2 ml of PBS was injected in a 4- to 5-cm-long loop. Each animal had either two loops as far proximally and distally as possible in the small intestine or one loop on the middle pan of the intestine. The values are the means ± standard errors of the mean of the net fluid secretion of 6 animals from three experiments. b The E. coli heat-labile toxin was prepared as described in the text, diluted 1:5 in PBS, and injected at a volume of.2 ml. One loop was prepared in the middle of the small intestine in each animal. The values represent the means ± standard errors of the mean of 4 animals from two experiments. were evaluated after 24 and 48 h of CD ligation. The results illustrated in Fig. 2 show that there was a significant inhibition of the secretion induced by prostaglandin E1, whereas the fluid response to DB-cAMP challenge remained unaffected. Effects on the activity of intestinal adenylate cyclase. The interaction of bile and cholera toxin with the intestinal adenylate cyclase was tested. The activity of mucosal adenylate cyclase was tested 24 h after CD ligation in a toxin-stimulated loop and a control loop of the small intestine. Table 4 shows that bile depletion reduced the stimulatory effect of cholera toxin on the enzyme activity more than 5%o as compared with control mice. DISCUSSION Ligation of the CD inhibited intestinal secretion induced by cholera toxin, E. coli heat-labile toxin, or prostaglandin E1. This effect was probably caused by the decreased levels of bile acids in the intestinal mucosa, as inhibition was reversed by the addition of bile or bile acids to the intestine. The influence of CD ligation on the kinetics of the cyclic AMP-mediated secretion needs to be further investigated. However, a lag period preceded the normalization of the fluid response to toxin challenge after addition to the intestine of bile or bile acids. Cholera toxin-induced activation of intestinal adenylate cyclase is also preceded by a lag period (7, 11). This delay in action of the toxin and bile acids may reflect the time for penetration through the brush border membrane and diffusion to the adenylate cyclase in the basal-lateral membrane (7). U U -j 15- z~~~~ I pg CT CHALLENGE FIG. 1. Inhibition of intestinal secretory response to various amounts of cholera toxin (CT) induced by ligation of the common bile duct. Symbols: A, mice after 24 h of bile duct occlusion;, untreated animals. Each value is the mean + standard error of the mean of 5 to 2 animals from one to three experiments. One loop was prepared on the middle part of the small intestine in each animal. The net fluid accumulation in the intestinal loop is given in milligams per centimeter.

4 394 LANGE AND LONNROTH TABLE 2.. Stimulating effect of bile and bile acids on the intestinal secretory response to cholera toxin challenge after 24 h of ligation of the CD Substances used for preincubation Preincubation Secretory response (mg/cm)a time (h) Proximal jejunum Ileum Controlb 167 ± 7 82± 5 Bile diluted 1-foldc 62 ± 8 (P <.1) 31 ± 6 (P <.1) Bile diluted 1-foldc ± 9 (P <.1) 62 ± 7 (P <.1) Bile diluted 1-foldc ± 6 NS 91 ± 7 NS Cholic acid, 2.5 mmc ± 1 NS 88 ± 5 NS Deoxycholic acid, 2.5 mmc ± 8 NS 93 ± 11 NS a The fluid content (milligrams per centimeter) of the intestinal loops was estimated 4 h after toxin challenge. Values represent the means ± standard errors of the mean of three experiments with five animals in each. The P values were calculated by comparing the values of the experimental animals with that of the sham-operated controls. NS, Not significant. b Control mice did not undergo CD ligation. c At 24 h after CD ligation, bile or bile acids (neutralized with NaOH) were injected before or simultaneously with cholera toxin challenge (1.5,ug in.2 ml) into intestinal loops. Depletion of intestinal bile inhibited the cholera toxin-induced stimulation of adenylate cyclase in the intestinal mucosa. The level of intestinal bile exerted no effect, however, on either the binding of cholera toxin to the intestinal epithelial cells or on DB-cAMP-induced secretion which bypasses adenylate cyclase. These results suggest that bile may interfere with the toxin activation of adenylate cyclase. By using rat colon, it was recently shown that challenge with a high concentration of deoxycholic acid and certain other bile acids induces secretion with a concomitant rise of cyclic AMP and damage of the mucosal epithelium (4). However, a much higher concentration of deoxycholic acid was needed for the induction of secretion in the colon than for its influence on enterotoxininduced secretion in the small intestine. Further- TABLE 3. Effect of CD ligation on the binding of 1"'I-labeled cholera toxin to the intestinal mucosa Bie duct Incubation Bound 25I-labeled cholera ligation Icbto toxinb time (h) time (min)' cpm/cm % ± ± ,248 ± ± ± ± ± ± 1. a The incubation time is the time that "25I-labeled cholera toxin (2.5,ug,.2,uCilg) was incubated in a loop in the distal jejunum. b Radioactivity refers to the total amount of radioactivity attached to the mucosal cells in a 5-cm-long jejunum intestinal loop. The injected 2.5 1Lg of 125Ilabeled cholera toxin contained 31,25 cpm. The values in the columns are expressed as radioactivity per centimeter of intestine and as a percentage of the total binding. 1' -J5 Oh 24h 48h INFECT. IMMUN. Oh 24h 48h PROSTAGLANDINE El DIBUTYRYL CYCLIC- AMP FIG. 2. Effect of ligation of the common bile duct on the capability of prostaglandin E1 and DB-cAMP to induce intestinal secretion. The figures in the staples represent hours after bile duct ligation. Prostaglandin E1 and DB-cAMP were diluted to a concentration of.75 and 1 mg per ml in PBS, respectively, and.2 ml was injected in a loop in the middle part of the small intestine. The animals were killed after 9 min. The columns represent the mean value of the net fluid accumulation (milligrams per centimeter) in loops in 1 animals. Vertical bars represent the standard errors of the mean.

5 VOL. 35, 1982 EFFECT OF BILE ON ENTEROTOXIN-INDUCED DIARRHEA 395 TABLE 4. Inhibitory effect of ligation of the common bile duct on the activity of adenylate cyclase in the intestinal mucosaa Challenge Adenylate Weight of Bile duct with cyclase; cyclic intestinal ligation cholera AMP formed loops toxin (pmol/mg/min) (mg/cm) ±.57 NS a Ligation of the common bile duct was performed 24 h before challenge with cholera toxin (performed as described in Table 1). The enzymatic activity was assayed on washed mucosal homogenates. Values are means ± standard errors of the mean of triplicate determinations on pooled material from four mice. NS, Not significant compared with untreated control animals. more, cholic acid was effective in neither the rat colon (4) nor the mouse small intestine; nevertheless, it, together with cholera toxin, effectively restored the impaired secretory response to cholera toxin challenge after CD ligation. The level of intestinal bile acids may be a critical factor for the induction of hypersecretion during cholera. The concentration of bile acids in the stool might therefore provide information on the sensitivity for enterotoxic diarrhea. These studies should preferably be performed not only on cholera and normal stools, but also on material from subclinical and convalescent cases. ACKNOWLEDGMENT This work was supported by grants from the Swedish Medical Research Council (16X-625), A. B. Leo research foundation, and The Swedish Society of Medical Sciences (grant 234). LITERATURE CITED 1. Coyne, M. J., G. G. Bonorris, and A. Chug Propanolol inhibits bile acid and fatty acid stimulation of cyclic AMP in human colon. Gastroenterology 73: Field, M Intracellular mediators of secretion in the small intestine, p In H. J. Binder (ed.), Mechanisms of intestinal secretion. Alan R. Liss. Inc., New York. 3. Feld, M., H. E. Sheerin, A. Henderson, and. Smith Catecholamine effects on cyclic AMP levels and ion secretion in rabbit ileal mucosa. Am. J. Physiol. 229: Gordon, S. J., D. Kinsey, and J. S. Magen Structure of bile acids associated with secretion in the rat cecum. Gastroenterology 77: Holngren, J., I. L6nnroth, J. E. Mnason, and L. Svennerholm Interaction of cholera toxin and membrane GM1 ganglioside of small intestine. Proc. Natl. Acad. Sci. U.S.A. 72: Kanto, H. S., P. Tao, and H. C. Kiefer Kinetic evidence for the presence of two prostaglandin receptor sites regulating the activity of intestinal adenylate cyclase sensitive to E. coli enterotoxin. Proc. Natl. Acad. Sci. U.S.A. 71: Kimberg, D. V Cyclic nucleotides and their role in gastrointestinal secretion. Gastroenterology 67: Krishna, G., B. Weiss, and B. B. Brodie A simple sensitive method for the assay of adenyl cyclase. J. Pharmacol. Exp. Ther. 163: Lange, S., and J. Holngren. Protective antitoxic cholera immunity in mice: influence of route and number of immunizations and mode of action of protective antibodies. Acta Pathol. Microbiol. Scand. Sect. C 86: L6nnroth, I., B. Andrdn, S. Lange, K. Martnsson, and J. Hohlgren Chlorpromazine reverses Escherichia coli (LT') diarrhea in piglets. Infect. Immun. 24: L6nnroth, I., and C. L6nnroth Interaction of cholera toxin and its subunits with lymphocytes. The effect on intracellular cyclic AMP. Exp. Cell Res. 14: Mekhjlian, H. S., S. F. Philips, and A. F. Hohnann Colonic secretion of water and electrolytes induced bile acids: perfusion studies in man. J. Clin. Invest. 5: Merson, M. H., R. E. Black, M. Kahn, and I. Huq Epidemiology of cholera and enterotoxigenic Escherichia coli diarrhea, p In Cholera and related diarrheas, 43rd Nobel Symposium. Karger, Basel. 14. Moss, J., and M. Vaughan Mechanism of activation of adenylate cyclase by choleragen and E. coli heat-labile enterotoxin, p In M. Field, J. S. Fordtran, and S. G. Schultz (ed.), Secretory diarrhea. The Williams & Wilkins Co., Baltimore. 15. Svennerholm, A. M., and J. Holmgren Identification of E. coli heat-labile enterotoxin by means of a ganglioside immunosorbent assay (GM1-ELISA) procedure. Curr. Microbiol. 1:19-23.