Secretory Immunoglobulin A Response to Shiga ...

INFECTION AND IMMUNITY. JUIY 1989. p. 1885-1889 0019-9567/89/071885-05$02.00/0 Copyright © 1989. American Society for Microbiology

Vol. 57, No. 7

Secretory Immunoglobulin A Response to Shiga Toxin in Rabbits: Kinetics of the Initial Mucosal Immune Response and Inhibition of Toxicity In Vitro and In Vivo DAVID F. KEREN,l* J. EDWARD BROWN,- RODERICK A. McDONALD,1 AND JOSEPH S. WASSEF'

Department

ofJPathology,

Unih'ersit of Michigan Medi(al Sc1hool, Ann Arbor, Michigan 48109,1 and U.S. Armed For(ces Research Instititte of Medical Science, Bangkok, Thailand2 Received 3 January 1989/Accepted 15 March 1989

Although the role of Shiga toxin in dysentery is unknown, the toxin is cytotoxic to HeLa cells, causes fluid secretion in rabbit intestine, and is lethal to rabbits and mice when injected parenterally. In the present study, rabbits received three weekly doses of Shiga toxin directly into chronically isolated ileal loops. Within a week, secretions from these loops contained immunoglobulin A (IgA) anti-Shiga toxin. The titer of IgA anti-Shiga toxin increased after weekly doses 2 and 3. Little IgG anti-Shiga toxin was present in loop secretions, although high titers of IgG anti-Shiga toxin were found in the sera. These loop secretions were able to neutralize the cytotoxic effects of Shiga toxin in the HeLa cell assay. The capacity to neutralize the cytotoxicity of the toxin correlated strongly with the IgA anti-Shiga toxin titer in these same secretions. Pooled immune loop secretions were also able to significantly reduce fluid accumulation in acutely ligated loops in rabbits, while loop secretions from control rabbits could not. Shiga toxin elicited a strong secretory IgA response upon application to the intestine. Further, the mucosal antibodies produced functioned to prevent the toxic effects of Shiga toxin both in vitro and in vivo.

Shigella dysenteriae causes the most severe forms of dysentery. The major mechanism of virulence involves invasion of the colon or terminal ileum. These bacteria preferably invade surface epithelium over lymphoid follicles in the acutely ligated intestinal-loop model in rabbits (30). Following invasion, the shigellae replicate within the surface epithelium, damaging these cells and eventually causing ulcerations at these sites (21). This process is thought to be responsible for the bloody and frequent stools which are characteristic of dysentery. It has also been known for some time that S. dysenteriae produces a potent toxin (14, 15, 17). Although the exact role that Shiga toxin plays in human dysentery is not known, this toxin is lethal when given parenterally to rabbits and mice (18, 26, 29). In acutely ligated rabbit ileal loops, considerable fluid accumulates in 18 to 24 h (16). In tissue culture cell lines, Shiga toxin inhibits protein synthesis and is cytotoxic (6, 25). Lastly, Shiga toxin may cause damage to the vasculature in colonic lamina propria, thereby intensifying the severity of bacillary dysentery (5). Shiga toxin, purified from culture lysates of S. dysenteriae, is a protein composed of one A subunit (30,500 molecular weight) and five B subunits (5,000 molecular weight each) (3, 27). There has been some controversy about the binding receptors for Shiga toxin. Work by Keusch et al. suggests that the B subunit of Shiga toxin binds to two receptors: a 1-4-linked n-acetyl-D-glucosamine oligomer on the surface epithelium (19, 20) and the glycolipid globotriaosylceramide Gb3 (8). Lindberg et al. used inhibition studies to demonstrate that the HeLa cell receptor is a Galoa1-4-Galr3 (galabiose) linked to a glycoprotein (23). The A subunit is thought to be responsible for the toxic activity after its entry into the cytosol. It inhibits protein synthesis by inactivating aminoacyl-tRNA binding (2). Although Shiga toxin is a strong immunogen with high *

titers in sera from patients with shigellosis, no information is available concerning the mucosal immunoglobulin A (IgA) response or its ability to protect against the effects of Shiga toxin (24). Our laboratory has been investigating the mucosal secretory IgA response of the intestine to shigella bacilli. In previous work, we demonstrated that after oral immunization with live nonpathogenic strains, a mucosal secretory IgA memory response was elicited in rabbits (11, 12). However, as with most other nonreplicating antigens, heatkilled shigellae given orally to rabbits elicited much weaker mucosal immune responses than did live shigellae and were not able to elicit a mucosal memory response (11). In the present study, we have used a chronically isolated ileal (Thiry-Vella)-loop model to evaluate the immunogenicity of Shiga toxin for stimulating a secretory IgA response. We demonstrate that Shiga toxin can elicit a vigorous secretory IgA response and that these antibodies can protect against the effects of Shiga toxin both in vitro with HeLa cells and in vivo with acutely ligated rabbit intestine. (Parts of this research were presented at the American Society for Microbiology Annual Meeting [D. F. Keren, J. E. Brown, and R. A. McDonald, Abstr. Annu. Meet. Am. Soc. Microbiol. 1988, E-23, p. 112] and at the United States-Japan Conference on Cholera [November 1988].) MATERIALS AND METHODS Preparation of chronically isolated ileal loops. Chronically isolated ileal (Thiry-Vella) loops were prepared in each of five rabbits by a previously described method (10). Briefly, 3- to 4-kg New Zealand White rabbits were anesthetized with xylazine and ketamine. A midline abdominal incision was made, and the terminal ileum was identified. A 20-cm segment of ileum containing a grossly identifiable Peyer's patch was isolated while keeping the vascular supply intact. The remaining intestine was reanastomosed to provide continuity. Silastic tubing (Dow-Corning Corp., Midland, Mich.) was sewn into each end of the isolated loop, which

Corresponding author. 1885

1886

KEREN ET AL.

was left in the abdominal cavity. The silastic tubes were brought out through the midline abdominal incision and tunneled subcutaneously to the nape of the neck, where they were exteriorized and secured. The midline abdominal incision was closed in two layers. Daily care of the isolated loop required flushing with air and saline. About 2 ml of intestinal secretions were collected daily from these loops. Secretions were centrifuged, and supernatants were stored at -20°C until time of assay. Shiga toxin preparations. Two preparations of Shiga toxin were used in the present study. Partially purified toxin obtained from S. dysenteriae 1 strain 3818-0 as described previously was used for the immunization (1). This preparation had 104 to 105 50% cytotoxic doses per ml. The second preparation was purified Shiga toxin which was used for the enzyme-linked immunosorbent assay (ELISA) system. This was purified as described previously by using a combination of ion-exchange chromatography, gel filtration chromatography, and preparative isoelectric focusing (1). This preparation of pure toxin contained 108 50% cytotoxic doses per mg. Immunization schedule. Five rabbits were inoculated directly intraloop on the day of surgery (antigen day 0) and on days 7 and 14 postsurgery. They were immunized with 0.5 ml of crude toxin preparation in 4 ml of saline. Intestinal secretions were collected daily, and blood samples were collected weekly. Antibody titers were measured for 30 days after immunization. ELISA. A previously described assay for detecting rabbit IgG and IgA antibodies to shigella products was used to detect specific antibody directed against Shiga toxin in intestinal-loop secretions and serum (9). Briefly, polystyrene microdilution wells were coated with pure Shiga toxin (50 lI of a solution containing 0.5 p.g/ml in carbonate buffer, pH 9.6). These plates were covered with Parafilm to prevent evaporation and were stored at 4°C until the time of assay. Immediately prior to addition of serum or intestinal secretions, the Shiga toxin solution was removed and the wells were washed with phosphate-buffered saline, pH 7.4, containing 0.1% Tween 20. Samples to be assayed were serially diluted in this buffer and incubated in both Shiga toxincoated wells and uncoated wells (to control for nonspecific adsorption) for 4 h. Standard solutions of IgA and IgG anti-Shiga toxin were pooled from positive loop secretions and positive sera, respectively. Replicate analyses of these solutions were performed to establish their titers. A fourpoint standard curve was assayed on each plate with the unknown samples. Following this incubation, the wells were washed with the buffer and incubated overnight at room temperature with solutions containing either alkaline phosphatase-conjugated sheep anti-rabbit IgA or alkaline phosphatase-conjugated sheep anti-rabbit IgG (affinity column purified and shown to be monospecific by ELISA) (9). After an additional wash with buffer, the substrate reaction was carried out with p-nitrophenyl phosphate in carbonate buffer (1 mg/ml). After the kinetics of the enzyme-substrate reaction were extrapolated to 100 min, as measured on a Titertek Multiscan MicroELISA Reader (Flow Laboratories, Inc., McLean, Va.), the optical density at 405 nm (OD405) of uncoated wells was subtracted from the OD405 of the coated wells. To ensure consistency, a four-point standard curve for both IgG and IgA was run on each plate. The reciprocal of the dilution giving an OD reading between the two lowest values on the standard curve was defined as the titer. Cytotoxicity assay. Shiga toxin activity was determined by examining the extent of HeLa cell damage by a previously described assay (1, 6). Briefly, HeLa cell monolayers were

INFECT. IMMUN.

14 04

0

12

0 -J

10 I-

IC,)

8

I-

z

6

a

O

4

10

20

30

DAY AFTER IMMUNIZATION

FIG. 1. Mean IgG anti-Shiga toxin titers in serum (0) and in secretions from the isolated loops (0). Standard errors of the means are indicated. grown in 96-well microdilution plates. For the assay, a standard crude toxin lysate of S. dysenteriae was incubated with serial dilutions of loop fluids for 30 min at room temperature. This mixture was placed onto the HeLa cell monolayer and allowed to incubate overnight at room temperature. The monolayers were then stained with crystal violet dissolved in 50% ethanol-1% sodium dodecyl sulfate, and the OD620 was determined for each well. The dye remaining in each well correlated with the percentage of cells remaining adherent to the microdilution dishes (1, 6). The OD620s of wells containing the standard toxin alone were averaged, and that value plus 2 standard deviations was defined as the endpoint titer of loop fluids for neutralization of the cytotoxicity of the toxin preparation. All dilutions of loop fluid which gave an OD620 in the assay greater than this value were scored as positive. Acutely ligated loop protection studies. Pooled loop secretions from rabbits with high-titer IgA anti-Shiga toxin activity as determined by ELISA were mixed 1:2 in saline and then mixed with an equal volume of a 1:256 dilution of crude toxin and injected into 5-cm-long isolated segments of ileum in unimmunized rabbits. This dose of toxin was chosen because it consistently elicited fluid accumulation when given to acutely ligated loops. As controls, toxin was mixed with secretions from nonimmune animals or saline and injected into other loops in the same rabbit. After 18 h, the animals were sacrificed and the volume of fluid in each segment was measured. Statistical methods. Data were statistically analyzed with the RS1 interactive data analysis system. Differences between groups on specific days were tested for significance by the Mann-Whitney test for unpaired samples with equal dispersions.

RESULTS IgG anti-Shiga toxin in serum and in loop secretions. A significant (P < 0.01) increase in the mean IgG anti-Shiga toxin titer over the day 0 value was detectable in serum by day 7 after the first intraloop immunization. This level continued to rise after dose 3 on day 14 (Fig. 1). By day 21, the IgG anti-Shiga toxin titer reached a plateau and did not change significantly through the end of the study on day 30.

VOL 57 1989

IgA RESPONSE TO SHIGA TOXIN

,

1887

12 -

10 -

C, a

0 -J

CM4

8-

I--

C3

0

I-

M

6-

1

4-

LU

z

4c

4c CO

Cl) Iz

20

1 0

20

30

DAY AFTER IMMUNIZATION 10

FIG. 2. Mean IgA anti-Shiga toxin titers in serum (0) and in secretions (0), with standard errors of the means indicated.

In contrast to the high level of IgG anti-Shiga toxin in the serum, only trivial amounts of IgG anti-Shiga toxin were detectable in the loop secretions (Fig. 1). By day 15 after intraloop immunization, the content of IgG anti-Shiga toxin in loop fluids was significantly greater than the day 0 values (P < 0.01). There was no significant change in the content of IgG anti-Shiga toxin from day 15 through the end of the experiment on day 30. These findings indicate that only a small fraction of IgG from the systemic circulation finds its way, in an active form, into the intestinal lumen. IgA anti-Shiga toxin in serum and in loop secretions. The IgA anti-Shiga toxin titer in the serum of these animals was lower than the IgA titer in secretions (Fig. 2). A significant (P < 0.01) increase over day 0 values was not seen until day 14 after the first intraloop immunization. Between day 14 and the end of the study on day 30, there was no significant change in the IgA anti-Shiga toxin titer. In contrast to the IgG anti-Shiga toxin content in loop secretions, a strong IgA anti-Shiga toxin response was noted (Fig. 2). As early as day 2, a weak but significant (P < 0.05) increase in the IgA anti-Shiga toxin titer was seen (Fig. 2). The content of IgA anti-Shiga toxin declined on the day after intraloop dose 3 (day 14) but had another striking increase 3 days later (Fig. 2). After this peak on day 18, the mean IgA anti-Shiga toxin titer slowly declined, though it never dropped below the level of activity seen after intraloop dose 2 on day 7. It is possible that the slight decline in IgA level the day after each booster immunization (days 8 and 15) reflects the presence of free toxin in the loop which binds to the specific IgA being secreted. This binding either would precipitate some of the available pool of IgA anti-Shiga toxin in the secretions or would compete in the ELISA. Alternatively, Shiga toxin may interfere with local antibody synthesis or secretion of IgA into the gut lumen. In vitro anti-Shiga toxin-neutralizing activity of intestinalloop secretions. The mean Shiga toxin-neutralizing activity of the intestinal-loop secretions is depicted in Fig. 3. The loop fluids from immunized rabbits neutralized the toxin added to HeLa cell monolayers, thereby inhibiting lysis of the cells. Further, the curve in Fig. 3 shows the same basic triphasic response as occurred with the IgA anti-Shiga toxin in loop secretions used for Fig. 2. The correlation coefficient of the mean IgA activity in secretions with the mean toxin neutralization titer was 0.928, while the correlation of IgG level in

20

30

DAY AFTER IMMUNIZATION FIG. 3. Mean Shiga toxin neutralizing titer in secretions in the HeLa cell cytotoxicity assay. Standard errors of the means are indicated. secretions with the mean toxin neutralization titer was only 0.116. In vivo anti-Shiga toxin-neutralizing activity of loop secretions. Pooled loop secretions from animals immunized with Shiga toxin reduced toxin-induced fluid accumulation in acutely ligated loops of rabbit intestine (Fig. 4). Secretions with no detectable IgA or IgG anti-Shiga toxin by ELISA had no significant inhibitory effect on Shiga toxin-induced fluid production by rabbit intestine. The heterogeneity shown by the standard errors of the means reflects the differential response of the genetically diverse outbred rabbits used in these studies. Even with this heterogeneity, the difference between the fluid production in loops protected with immune secretions and those given nonimmune secretions was highly significant (P < 0.01). No significant difference in fluid production was observed between acutely ligated loops protected with nonimmune secretions and those given only saline, although the former had a lower mean fluid secretion. 108

6a

-

4.

IF

IL

2-

SALINE

IMMUNE

-S CONTROL

FIG. 4. Mean fluid accumulations after acutely ligated segments of bowel where challenged with a standard dose of Shiga toxin mixed with an equal volume of saline, pooled immune loop secretions diluted 1:2, or pooled control secretions diluted 1:2.

1888

KEREN ET AL.

DISCUSSION

While the precise role of Shiga toxin in human dysentery is not known, evidence in experimental animals indicates that it has the potential to damage the surface epithelium, a condition which results in fluid secretion (16). Further, recent studies implicate Shiga toxin in damage to the vasculature in the colonic lamina propria, which may result in bloody diarrhea (5). The present study demonstrates that a strong secretory IgA mucosal immune response to Shiga toxin can be elicited and suggests that such antibodies could interfere with the toxic effects on epithelial cells. The mucosal immune response in the intestine is known to be elicited against antigens present within the gut lumen, particularly against live microorganisms (11, 12). However, attempts to elicit mucosal immune responses against proteins and haptens conjugated to carrier proteins usually have yielded weak secretory IgA responses. The major exception to this has been cholera toxin. This molecule, which like Shiga toxin consists of one A subunit and five B subunits (which attach to the surface epithelium), consistently has been a strong mucosal immunogen (28). Rabbit Thiry-Vella loops immunized intraluminally three times with cholera toxin produced a strong secretory IgA response which correlated well with protection against fluid accumulation in acutely ligated segments of intestine challenged with toxin (31). Other proteins such as keyhole limpet hemocyanin and bovine serum albumin have been weak or ineffective at eliciting secretory IgA responses following intraluminal immunization (7). The ability of Shiga toxin to elicit high titers of secretory IgA antibodies suggests that it may be able to serve as an effective carrier protein for other molecules. Cholera toxin has been used as carrier protein to enhance the secretory IgA response against other proteins and microorganisms (4, 22). While there has been speculation that the adjuvant effect is due to the binding by the B subunit to GM1 on the surface epithelium and subsequently to lymphocytes, the B subunit alone is not as effective as intact cholera toxin for eliciting a vigorous secretory IgA response to the attached antigen (22). The immune responses against Shiga toxin in the present study also corroborate the dichotomy which exists between the systemic and the mucosal immune systems. While a high titer of IgG anti-Shiga toxin was detected in the serum following direct intraintestinal immunization with Shiga toxin, only trivial amounts of IgG anti-Shiga toxin were present in the intestinal loop secretions. Previous studies have demonstrated that this is not due to degradation of the IgG in the isolated ileal-loop secretions (13). In the environment of the isolated ileal loop, IgG has stability similar to that of secretory IgA (bile, digestive enzymes, and gastric acid are absent). This reflects the fact that IgG is not a secretory immunoglobulin, does not bind to secretory component, and is not transported by this mechanism into the gut lumen. In contrast, high titers of IgA anti-Shiga toxin were detectable in the intestinal secretions, while lower levels of IgA anti-Shiga toxin were measured within the serum. This substantiates the possibility that the preferential formation of IgA along mucosal surfaces after intestinal priming with antigen also happens with Shiga toxin. The anti-Shiga toxin activity of ileal-loop secretions as reflected by the tissue culture cell assays paralleled the IgA anti-Shiga toxin titer within the intestinal-loop secretions, while there was no correlation with the IgG titer in these secretions. These findings show that a strong IgA anti-Shiga toxin response

can

be elicited in the intestinal secretions and

INFECT. IMMUN.

that these antibodies are likely to confer protection against the cytotoxic effects of Shiga toxin. This notion was substantiated by the in vivo protection seen in the acutely ligated loop studies. It was also notable that by the end of the study on day 30, the titer of secretory IgA anti-Shiga toxin in the HeLa cell system had not fallen significantly from its peak. This, together with the booster effect after dose 2 of Shiga toxin (Fig. 2), suggests that a long-term immunity may be expected from this immunization regimen. Studies are under way in our laboratory to examine whether a mucosal memory response which will be effective in preventing damage by this toxin can be elicited. ACKNOWLEDGMENTS We thank Terri Throne for her excellent assistance in preparing the manuscript. This research was supported in part by contract DAMD1787-C-7143 from the U.S. Army Medical Research and Development Command. LITERATURE CITED 1. Brown, J. E., D. E. Griffin, S. W. Rothman, and B. P. Doctor. 1982. Purification and biological characterization of shiga toxin from Shigella dysenteriace 1. Infect. Immun. 36:996-1005. 2. Brown, J. E., T. G. Obrig, M. A. Ussery, and T. P. Moran. 1986. Shiga toxin from Shigella dysenteriae 1 inhibits protein synthesis in reticulocyte lysates by inactivation of aminoacyl-tRNA binding. Microb. Pathog. 1:325-334. 3. Donohue-Rolfe, A., G. T. Keusch, C. Edson, D. Thorley-Lawson, and M. Jacewicz. 1984. Pathogenesis of shigella diarrhea. IX. Simplified high yield purification of shigella toxin and characterization of subunit composition and function by the use of subunit-specific monoclonal and polyclonal antibodies. J. Exp. Med. 160:1767-1781. 4. Elson, C. O., and W. Ealding. 1984. Generalized systemic and mucosal immunity in mice after mucosal stimulation with cholera toxin. J. Immunol. 132:2736-2741. 5. Fontaine, A., J. Arondel, and P. J. Sansonetti. 1988. Role of Shiga toxin in the pathogenesis of bacillary dysentery, studied by using a tox- mutant of Shigella dvsenteriae 1. Infect. Immun. 56:3099-3109. 6. Gentry, M. K., and J. M. Dalrymple. 1980. Quantitative microtiter cytotoxicity assay for Shigella toxin. J. Clin. Microbiol. 12: 361-366. 7. Hamilton, S. R., D. F. Keren, J. H. Yardley, and G. Brown. 1981. No impairment of local intestinal immune response to keyhole limpet haemocyanin in the absence of Peyer's patches. Immunology 42:431-435. 8. Jacewicz, M., H. Clausen, E. Nudelman, A. Donohue-Rolfe, and G. T. Keusch. 1986. Pathogenesis of shigella diarrhea. XI. Isolation of a shigella toxin-binding glycolipid from rabbit jejunum and HeLa cells and its identification as globotriaosylceramide. J. Exp. Med. 163:1391-1404. 9. Keren, D. F. 1979. Enzyme-linked immunosorbent assay for immunoglobulin G and immunoglobulin A antibodies to Shigella flexneri antigens. Infect. Immun. 24:441-448. 10. Keren, D. F., J. L. Elliiott, G. D. Brown, and J. H. Yardley. 1975. Atrophy of villi with hypertrophy and hyperplasia of Paneth cells in isolated (Thiry-Villa) ileal loops in rabbits. Gastroenterology 68:83-93. 11. Keren, D. F., S. E. Kern, D. H. Bauer, P. J. Scott, and P. Porter. 1982. Direct demonstration in intestinal secretions of an IgA memory response to orally administered Shigella flexneri antigens. J. Immunol. 128:475-479. 12. Keren, D. F., R. A. McDonald, and S. B. Formal. 1986. Secretory immunoglobulin A response following peroral priming and challenge with Shigella flexneri lacking the 140-megadalton virulence plasmid. Infect. Immun. 54:920-923. 13. Keren, D. F., P. J. Scott, and D. Bauer. 1980. Variables affecting the local immune response in Thiry-Vella loops. II. Stability of antigen-specific IgG and secretory IgA in acute and chronic

IgA RESPONSE TO SHIGA TOXIN

VOL. 57, 1989

Thiry-Vella loops. J. lmmunol. 124:2620-2624. 14. Keusch, G. T., A. Donohue-Rolfe, and M. Jacewicz. 1982. Shigella toxin(s): description and role in diarrhea and dysentery. Pharmacol. Ther. 15:403-438. 15. Keusch, G. T., A. Donohue-Rolfe, and M. Jacewicz. 1985. Shigella toxin and the pathogenesis of shigellosis. CIBA Found. Symp. 112:193-214. 16. Keusch, G. T., G. F. Grady, L. J. Mata, and J. McIver. 1972. The pathogenesis of Shigella diarrhea. 1. Enterotoxin production by Shligella dysenteriae 1. J. Clin. Invest. 51:1212-1218. 17. Keusch, G. T., G. F. Grady, A. Takeuchi, and H. Spring. 1972. The pathogenesis of shigella diarrhea. II. Enterotoxin-induced acute enteritis in the rabbit ileum. J. Infect. Dis. 126:92-95. 18. Keusch, G. T., and M. Jacewicz. 1975. The pathogenesis of shiga enterotoxin, neurotoxin and cytotoxin. J. Infect. Dis. 131(Suppl.):S33-S39. 19. Keusch, G. T., and M. Jacewicz. 1977. Pathogenesis of shigella diarrhea. VII. Evidence for a cell membrane toxin receptor involving f1-4-linked N-acetyl-D-glucosamine oligomers. J. Exp. Med. 146:535-546. 20. Keusch, G. T., M. Jacewicz, and A. Donohue-Rolfe. 1986. Pathogenesis of shigella diarrhea. Evidence for an N-linked glycoprotein receptor and receptor modulation by P-galactosidase. J. Infect. Dis. 153:238-248. 21. LaBrec, E. H., H. Schneider, T. J. Magnani, and S. B. Formal. 1964. Epithelial cell penetration as an essential step in the pathogenesis of bacillary dysentery. J. Bacteriol. 88:1503-1518. 22. Liang, X., M. E. Lamm, and J. G. Nedrud. 1988. Oral administration of cholera toxin-Sendai virus conjugate potentiates gut

23.

24.

25. 26. 27. 28. 29.

30. 31.

1889

and respiratory immunity against Sendai virus. J. lmmunol. 141: 1495-1501. Lindberg, A. A., J. E. Brown, N. Stromberg, M. Westling-Ryd, J. E. Schultz, and K.-A. Karlsson. 1987. Identification of the carbohydrate receptor for Shiga toxin produced by Shigella dysenteriae type 1. J. Biol. Chem. 262:1779-1785. O'Brien, A. D., and R. K. Holmes. 1987. Shiga and Shiga-like toxin. Microbiol. Rev. 51:206-220. O'Brien, A. D., and G. D. Laveck. 1982. Immunochemical and cytotoxic activities of Shigella dysenteriae 1 (Shiga) and Shigalike toxins. Infect. Immun. 35:1151-1154. Olitsky, P. K., and I. J. Kliger. 1920. Toxins and anti-toxins of Bacillus dv.senteriae shiga. J. Exp. Med. 31:19-33. Olsnes, S., R. Reisbig, and K. Eiklid. 1981. Subunit structure of Slhigella cytotoxin. J. Biol. Chem. 256:8732-8738. Pierce, N. F. 1978. The role of antigen form and function in the primary and secondary intestinal immune responses to cholera toxin and toxoid in rats. J. Exp. Med. 148:195-206. Van Heyningen, W. E., and G. P. Gladstone. 1960. The neurotoxin of Shligella shiigace. 1. Production, purification and properties of the toxin. Br. J. Exp. Pathol. 34:202-216. Wassef, J. S., D. F. Keren, and J. L. Mailloux. 1989. Role of M cells in initial antigen uptake and in ulcer formation in the rabbit intestinal loop model of shigellosis. Infect. Immun. 57:858-863. Yardley, J. H., D. F. Keren, S. R. Hamilton, and G. D. Brown. 1978. Local (immunoglobulin A) immune response by the intestine to cholera toxin and its partial suppression with combined systemic and intra-intestinal immunization. Infect. Immun. 19: 589-597.