Effects of Vibrio cholerae recombinant strains on rabbit ileum in vivo

GASTROENTEROLOGY

1991;101:319-324

Effects of Vibrio cholerae Recombinant Strains on Rabbit Ileum In Vivo Enterotoxin

Production

and Myoelectric

Activity

CHRISTOPHER D. LIND, RICHARD H. DAVIS, RICHARD L. GUERRANT, JAMES B. KAPER, and JOHN R. MATHIAS Department of Internal Medicine, Virginia; and Center for Vaccine Baltimore, Maryland

Previous

School of Medicine, Development, School

studies

have identified the effects of Vibrio choleragen (CT A+B+), on the myoelectric activity of rabbit ileal loops in vivo. The response was defined as the migrating action potential complex, the single ring contraction that propels luminal contents aborad. In this study the same rabbit model is used to assess whether migrating action potential complex activity or fluid output is induced by recombinant strains of V. cholerae that produce no subunit of cholera toxin (CT A-B-) or only by the inactive binding subunit (CT A-B+). Three live strains were studied: El Tor N16961 (CT A+B+) and recombinant wild-type strains CVD106 (CT A-B+) and JBK70 (CT A-B-). Controls received sterile culture broth. Migrating action potential complex frequency in animals inoculated with CT A+B+ was significantly increased compared with that in all other experimental groups (P < 0.01). Fluid output was also increased in animals inoculated with CTA+B+ compared with fluid output in all other groups (P < 0.05). Migrating action potential complex frequency and fluid output in rabbits given CTA-B+ or CTA-Bdid not differ from activity in controls. How these recombinant strains induce diarrhea is unknown, but the mechanism may involve bacterial colonization or production of an unknown toxin.

cholerae and its enterotoxin,

V ibrio

cholerae and its heat-labile enterotoxin, choleragen (A,A,B5), induce distinct effects on the myoelectric activity in ligated intestinal loops of anesthetized rabbits. This activity is characterized by the migrating action potential complex (MAPC), defined as spike potential discharge lasting > 2.5 seconds on a single recording site and migrating over at

University of Virginia, Charlottesville, of Medicine, University of Maryland,

least three consecutive recording sites. The MAPC represents the myoelectric characteristics of a single moving ring contraction, which propels intraluminal contents in an aborad direction (1). The mechanisms of action of choleragen, which also induces secretion from the intestinal mucosa, have been extensively studied (z-5). Choleragen is composed of three separate peptides, A,, A,, and B, which has five identical subunits. The five B subunits of choleragenoid (6) serve specifically to bind to the GM, ganglioside receptor on the enterocyte membrane (3). Once binding occurs, the A, subunit is passed into the cell, ultimately leading to activation of the adenylate cyclase system, cyclic adenosine monophosphate (AMP) production, and cellular chloride secretion (5). How choleragen induces the MAPC and the relationship of the MAPC to choleragen-induced secretion and diarrhea are not understood. Sinar et al. (7)showed that choleragenoid (B5), the inactive binding subunit of choleragen, can induce the MAPC without activating the adenylate cyclase system and without stimulating fluid production. Their study suggests that MAPC activity is not linked to fluid secretory states and may independently contribute to the pathophysiology of certain diarrhea1 diseases. Since that study, recombinant strains of V. cholerae that produce no subunits of cholera toxin (CT A-B-) or the inactive binding subunits only (CT A-B+) have been developed for a live oral cholera

Abbreviations used in this paper: CT A+B+, choleragen; CT A-B+, cholera toxin with inactive (-) A and active (+) B subunits; CT A-B-, cholera toxin with inactive A and B subunits: MAPC, migrating action potential complex. o 1991by the American Gastroenterological Association

00X-5085/91/$3.00

320

LIND ET AL.

vaccine (g-10). These live vaccines induce immunity to V. cholerae infection in 90% of patients, but both vaccines also induce a mild diarrhea in half of the patients. A previously described attenuated V, cholerae strain, Texas-Star-SR (CT A-B+) was also found to induce a mild diarrhea in human volunteers (9). We studied the effects of both types of recombinant strains on the myoelectric activity of rabbit ileum in vivo to assess if MAPC activity contributes to the diarrhea caused by one or both of these live vaccines.

Materials and Methods Surgical Preparation Male New Zealand White rabbits weighing 1.5-2.5 kg were used for all studies. The investigation was approved by the University of Virginia Animal Research Committee, protocol number 1561-02-87. After an overnight fast, the rabbits were anesthetized with pentobarbital sodium (30 mg/kg) through an ear vein. Additional anesthesia was administered through a catheter placed in the external jugular vein, and each animal received a tracheostomy. A midline laparotomy was performed, and the distal ileum located. A 12-cm ligated ileal loop, modified slightly from that used previously by Mathias et al. (l), was prepared. After the loop was isolated with ligatures, seven monopolar silver-silver chloride electrodes were sewn onto the serosal surface of the ileum at 2.5-cm intervals, three electrodes orad to the loop and four on the loop (11). Large-bore polyethylene catheters (OD, 7.9 mm; ID, 4.7 mm) were placed orad to the loop and in the loop to measure fluid output. Each electrode was connected to a rectilinear recorder (model R612; SensorMedics, Anaheim, CA) through SensorMedics model 9878 AC couplers. An indifferent electrode was placed in the SC tissue of a hind limb. All myoelectric activity was recorded at a sensitivity of 0.5 mV/cm, a high-frequency filter of 30 Hz, a time constant of 1.0 seconds, and a paper speed of 2.5 mm/s. Respiration was recorded by a pneumograph placed around the chest and attached to a pressure transducer. Temperature was monitored by a thermometer placed IP and maintained at 39°C by a hydrothermal heating pad. After the surgical preparation, all animals were allowed to stabilize. Subsequently, 1.0 mL of test material was injected into the proximal end of the ligated loop through a 19-gauge sterile needle. The abdominal incision was closed with 3.0 silk suture, and myoelectric activity and fluid output were recorded for at least 8 hours. During the entire surgical and recording periods, the animals were kept anesthetized with pentobarbital sodium, 2.2 mg/kg IV every 15 minutes or as needed.

Test Materials Three separate V. cholerae strains were used for the study: strain N16961, El Tor wild-type (CT A+B+); two derivatives of N16961, strain CVD106 (CT A-B+); and

GASTROENTEROLOGY

Vol. 101, No. 2

strain JBK70 (CT A-B-). Sterile culture broth was used as a control test material. All live V. cholerae organisms were incubated for 18 hours in Casamino acids yeast extract broth at 37°C with aeration, which resulted in colony counts of > 10’ organisms/ mL. The V. cholerae strains were provided by J. B. Kaper (9,lO). During the development of these strains, bacterial cultures were assayed for cholera holotoxin with the Y-l adrenal cell assay and were assayed for B subunit antigen with a ganglioside-dependent, enzyme-linked immunosorbent assay (9).

Data Analysis Studies were performed on five rabbits in each experimental group (total, 20 rabbits); myoelectric recordings and fluid outputs were measured for at least 8 hours after inoculation. Fluid output from the ligated loop was measured hourly. Myoelectric recordings were analyzed for slow-wave frequency, number of MAPCs per hour, time of onset of MAPC activity, and MAPC propagation velocity. The MAPC results were evaluated by the Kruskal-Wallis analysis of variance (ANOVA) test (nonparametric), and fluid output results were evaluated by the independent t test; P value < 0.05 was required for significance in both tests.

Results Figure 1 illustrates a representative myoelectric pattern that occurred in an ileal loop inoculated with wild-type El Tor V. cholerae (CT A+B+). This pattern fulfilled the criteria for the MAPC and was seen in all animals given CT A+B+. Fluid movement from the ligated loop was seen in the drainage catheter during each MAPC. Of note, none of the animals in the CT A+B+ experimental group had MAPCs above the ligated loop. Table 1 and Figure 2A summarize the MAPC and slow-wave characteristics for each group. The MAPC activity occurred in all animals given CT A+B+ , whereas minimal MAPC activity was seen in only one rabbit in each of the other groups (CT A-B+, CT A-B-, and control). The MAPC frequency was significantly greater in animals inoculated with CT A+B+ compared with the frequency in all other groups (P < 0.01). However, the MAPC onset time, MAPC propagation velocity, and slow-wave frequency did not significantly differ among the groups. The mean onset time for MAPC activity after inoculation was 369 k 38 minutes, and mean MAPC propagation velocity was 0.99 + 0.08 cm/s for all groups. Slowwave frequency did not differ among groups, nor did it differ before and after MAPC activity began. Table 1 and Figure 2B summarize the mean fluid output per 8 hours per rabbit for each group. In animals given CT A+B+, fluid output was signifi-

MYOELECTRIC

August 1991

ACTIVITY

FROM V. CHOLERAE STRAINS

321

01 2.5 cm

IO.2

mvlcm

Figure 1. The MAF’C, defined as spike potential discharge lasting >2.5 seconds on a single recording site and migrating over at least three consecutive recording sites.

cantly increased compared with that in the other groups (P < 0.05). Fluid output in animals inoculated with CT A-B+ or CT A-B- did not differ from that in controls. Figure 3 illustrates the time relationship between the fluid output and MAPC activity in all experimen-

Table 1. MigratingAction

Potential

Complex and Slow-Wave CT A+B+ (N16961)

No. of MAPCi8 h MAPC onset time (min) MACP propagation velocity (cm/s] Slow-wave frequency (cycles/s) Fluid output (mLl8 h)

42.6 + 393 2 0.96 2 17.1 t 6.0 ?

17.7' 55 0.08 0.4 1.9"

tal groups. No animal had fluid movement from the ileal loop before the 4th hour of recording. In all rabbits that had net fluid output from the ligated loop, the fluid movement per hour increased steadily from the 4th to the 8th hour of recording. Similarly, no animal had MAPC activity in the ligated ileal loop Characteristics

After V. Cholerae

Strains

CT A-B+” (CVD106)

CT A-B-”

0.4 -c 0.4 377 + 0 0.90 -+ 0.14 16.1 2 0.3 1.2 2 0.4

0.2 " 242 k 0.81 k 15.6 2 2.0 2

NOTE. Values are expressed as means k SEM; n = 5 for each group. “Only one animal in each of these groups had MAPC activity, and that was minimal. *Significantly greater than in all other groups (P < 0.01). ‘P < 0.05 compared with fluid output in other groups.

tJBK7aJ 0.2 0 0.00 0.2 1.6

Controls” (broth) 2.0 2 2.0 366 2 0 1.50 f 1.10 16.6 f 0.4 1.8 I? 1.2

322

LIND ET AL.

GASTROENTEROLOGY Vol. 101, No. 2

60

CT A+B+

CT A-B+

CT

A-B-

Control

CT A+B+

CT A-B+

CT

A-B-

Control

Figure 2. A. Mean number of MAPCs per 8 hours per rabbit is represented by the vertical bar for each treatment group. The line above each bar represents the SEM; n = 5 for each group. *P < 0.01 by the Kruskal-Wallis test. B. Mean fluid output from the ligated loop per 8 hours per rabbit is represented each bar represents the SEM. * *P < 0.05 by the independent t test.

before the 5th hour of recording. In those with MAPC activity, the frequency increased from the 5th to the 8th hour of recording.

animals steadily

Discussion The secretory diarrhea that live V. cholerae and its enterotoxin, choleragen, induce remains a major cause of mortality and morbidity on a worldwide

-

FLUID OUTPUT

. .._ ??. . .

M*p(..

0123456769

HOURS

AFTER ONSET OF RECORDING

Figure 3. Cumulative MAPC frequency and fluid output for all experimental groups (n = 20). Vertical axes represent cumulative number of MAPCs (right) and cumulative volume (in milliliters) of fluid output (left) over the &hour recording period. The horizontal axis represents hours after onset of recording of the ligated ileal loop.

by the vertical bar for each treatment group. The line above

scale (12). The specific mechanism of action of choleragen in producing cellular secretion has been fairly well elucidated (2-5). Choleragen is a heat-labile enterotoxin composed of three separate peptides: A, and A,, which are linked together by a disulfide bond, and five identical B subunits, which are linked to each other and the A,, A, subunits by noncovalent forces (6). The five B subunits that constitute choleragenoid serve specifically to bind to the GM, ganglioside receptor on the enterocyte membrane (3). After binding, A, is cleared from the cholera toxin molecule and enters the enterocyte cytosol to act on the adenylate cyclase system located on the basolateral membranes. The A, subunit acts by inducing adenosine diphosphate (ADP) ribosylation of the IYcomponent of G, in the adenylate cyclase system, ultimately leading to adenylate cyclase activation, cyclic AMP production, and cellular chloride secretion (5). The B5 subunits, or choleragenoid, will not activate this system in intact or disrupted cells, whereas the A, subunit alone will activate the system in disrupted cells, but not in intact cells (3). Hence, the B5 subunits serve only to bind cholera toxin to intact enterocytes, allowing A, to enter the cell and activate the adenylate cyclase system. In 1976, Mathias et al. (1) first described the specific pattern of myoelectric activity, the MAPC, induced by choleragen in ligated intestinal loops of anesthetized rabbits. The relationship of this single rapid ring contraction to choleragen-induced secretion and diarrhea is not understood, and the mechanism by which choleragen induces the MAPC has not been elucidated. Evidence exists that the MAPC may not result

August

1991

simply from fluid production in the ligated ileal loop. The MAPC activity is not induced by ileal perfusion with saline at rates above the secretory rate induced by choleragen (1). Further, fluid secretion produced by other mechanisms such as Escherichia coli heatstable enterotoxin or 8-bromo-5’-cyclic guanosine monophosphate does not induce the MAPC (13). Conversely, Sinar et al. (7) showed that choleragenoid (B5) alone did induce the MAPC in the absence of fluid production or adenylate cyclase activation. Their report suggests that MAPC activity is not directly linked to fluid secretory states. Recombinant strains of V. cholerae that produce only the B subunit and not the A subunit (CT A-B+) have been developed as live oral human vaccines (8-10). In clinical testing, these strains induce a mild diarrhea in - 50% of those vaccinated. We postulated that the mild diarrhea caused by the V. cholerae strain that produces CT A-B+ may reflect MAPC activity in the absence of significant fluid secretion. Our results do not support this hypothesis. Neither strain CT A-B+ nor strain CT A-Binduced MAPC activity or fluid output that differed from that of controls (sterile culture broth). In contrast, CT A+B+ did induce significantly more MAPC activity and net fluid output than all other groups, as would be expected from a live wild-type K cholerae strain. What are the possible reasons why CT A-B+ did not induce the MAPC when choleragenoid (B5) has been shown to induce the MAPC in this same experimental model? CT A-B+ might produce inadequate quantities of B subunit to induce the MAPC in vivo, or it might produce adequate quantities of B but with inadequate delivery of this subunit to the effector cell to induce the MAPC. Neither possibility was specifically measured and, hence, cannot be ruled out. However, CT A+B+ clearly produced adequate quantities of cholera holotoxin to induce both the MAPC and an increased volume of output. A third possible reason why CT A-B+ did not induce the MAPC is that pure choleragenoid (B5) may not induce the MAPC. Although the MAPC does not seem to be linked to fluid production alone, as summarized here, all other recognized inducers of MAPC activity, including E. coli heat-labile enterotoxin (14), vasoactive intestinal polypeptide (15), and ricinoleic acid (161,may activate the adenylate cyclase system. Consequently, we think there might be a link between MAPC induction and adenylate cyclase activation in the enterocyte. If so, choleragenoid alone should not induce the MAPC. To test this possibility, we have used high-performance liquid chromatography to show that purified choleragenoid does not induce the MAPC in this experimental model (17). These results, in conflict with those of Sinar et al. (i’),

MYOELECTRIC

ACTIVITY

FROM

V. CHOLERAE STRAINS

323

need to be confirmed because of major implications in understanding the mechanism of MAPC induction. The cause of the diarrhea induced in human volunteers by V. cholerae strains CT A-B+ and CT A-Bremains unknown. Either or both strains may produce an unknown enterotoxin that induces secretion or abnormal intestinal motility and diarrhea. In this acute rabbit model, neither abnormal intestinal motility nor net increased fluid output was present with either strain. Similarly, small bowel bacterial colonization alone may be the cause of diarrhea by these V. cholerae strains. For example, a nontoxigenic E. coli strain that colonizes the small bowel mucosa but produces no recognized toxin has been shown to cause fluid accumulation and diarrhea in a 72-hour rabbit model (18). We have demonstrated in the current study that recombinant strains of V. cholerae, which do not produce choleragen or produce only the binding subunit (B5), do not induce MAPC activity or increased fluid output in the ligated rabbit ileal loop in vivo. The mechanism by which these V. cholerae strains induce diarrhea in humans is unknown but may involve bacterial colonization or production of an unknown toxin. This study raises further questions about the mechanisms by which MAPC activity is induced. If MAPC activity cannot be induced by choleragenoid alone, it may be linked to the mechanisms of the cyclic nucleotide-dependent secretory action of cholera holotoxin. Further studies are clearly needed to delineate the relationships between enterocyte adenylate cyclase activation, cellular secretion, and intestinal myoelectric activity such as the MAPC.

References 1. Mathias JR, Carlson Cohen S. Intestinal

GM, DiMarino AJ, Bertiger G, Morton HE, myoelectric activity in response to live

Vibrio cholerae and cholera enterotoxin. 91-96. 2. Lai CY, Mendez E, Chang D. Chemistry 3.

4. 5.

6. 7.

a. 9.

J Clin Invest of cholera

1976;58: toxin:

the

subunit structure. J Infect Dis 1976;133(Suppl):S23-530. Gill MD. Mechanism of action of cholera toxin. In: Greengard P, Robison GA, eds. Advances in cyclic nucleotide research. Volume 8. New York: Raven, 1977:65-118. Moss J, Vaughan M. Activation of adenylate cyclase by choleragen (review). Annu Rev Biochem 1979;48:581-600. Moss J, Burns DL, Hsia JA, Hewlett EL, Guerrant RL, Vaughan M. Cyclic nucleotides: mediators of bacterial toxin action in disease. Ann Intern Med 1984;101:653-666. Finkelstein RA, LoSpalluto JJ. Production of highly purified choleragenoid. JInfect Dis 1970;121(Suppl):563-570. Sinar DR, Charles LG, Burns TW. Migrating action-potential complex activity in absence of fluid production is produced by B subunit of cholera enterotoxin. Am J Physiol 1982;242:G47G51. Kaper JB, Lockman H, Baldini MM, Levine MM. A recombinant live oral cholera vaccine. Biotechnology 1984;2:345-349. Kaper JB. Lockman H, Baldini MM, Levine MM. Recombinant

324

10.

11.

12.

13.

14.

15.

LIND ET AL.

GASTROENTEROLOGY Vol. 101, No. 2

nontoxigenic Vibrio cholerue strains as attenuated cholera vaccine candidates. Nature 1984;308:655-658. Kaper JB, Levine MM, Lockman HA, Baldini MM, Black RE, Clements ML, Morris JG. Development and testing of a recombinant live oral cholera vaccine. Vaccines 1985;85:107-111. Davis RH, Mathias JR. Phorbol esters induce retrograde myoelectric activity in rabbit ileum in vivo. Am J Physiol 1989;257: G578-G583. Guerrant RL. Microbial toxins and diarrhoeal diseases: introduction and overview. In: Evered D, Whelan J, eds. Microbial toxins and diarrhoeal disease (CIBA Foundation symposium 112). London: Pitman, 1985:1-13. Mathias JR, Nogueira J, Martin JL, Carlson GM, Giannella RA. Escherichia coli heat-stable toxin: its effect on motility of the small intestine. Am J Physiol 1982;242:G360-363. Burns TW, Mathias JR, Carlson GM, Martin JL, Shields RP. Effect of toxigenic Escherichia coli on myoelectric activity of small intestine. Am J Physiol 1978;235:E311-E315. Sninsky CA, Wolfe MM, Martin JL, Howe BA, O’Dorisio TM, McGuigan JE, Mathias JR. Myoelectric effects of vasoactive intestinal peptide 1983;244:G46-G51.

on rabbit

small

intestine.

Am J Physiol

16. Mathias JR, Martin JL, Burns TW, Carlson GM, Shields RP. Ricinoleic acid effect on the electrical activity of the small intestine in rabbits. J Clin Invest 1978;61:640-644. 17. Lind CD, Guerrant RL, Kurosky A, Mathias JR. Purified choleragenoid does not induce migrating action potential complex (MAPC) activity in rabbit ileum in vivo. J Pharmacol Exp Ther (in press). 18. Wanke CA, Guerrant RL. Small-bowel colonization alone is a cause of diarrhea. Infect Immun 1987;55:1924-1926.

Received March 7,199O. Accepted December 11,199O. Address requests for reprints to: John R. Mathias, M.D., Division of Gastroenterology G-64, Department of Internal Medicine, The University of Texas Medical Branch, Galveston, Texas 77550. Dr. Lind’s present address is: Division of Gastroenterology, Department of Medicine, Vanderbilt University, Nashville, Tennessee 37232. Mr. Davis’s present address is: Center for Swallowing Disorders, University of South Florida Medical Clinics, Tampa, Florida 33612. The authors thank Mary H. Clench and Alice W. Cullu for reviewing the manuscript.