Protection of weaned rabbits against experimental Escherichia coli O103 intestinal infection by oral formalin-killed vaccine

Veterinary Microbiology, 21 (1990) 353-362 Elsevier Science Publishers B.V., Amsterdam - - Printed in The Netherlands

353

Protection of Weaned Rabbits against Experimental E s c h e r i c h i a coli 0 1 0 3 Intestinal Infection by Oral Formalin-Killed Vaccine R. CAMGUILHEM 1 and A. MILON 2

' Laboratoire de Pathologie du Bdtail et des Animaux de Basse-Cour, et 2Laboratoire de Microbiologie-Immunologie, Ecole Nationale Vdtdrinaire, I.N.R.A., 23 chemin des CapeUes, 31076 Toulouse cedex (France) (Accepted for publication 8 August 1989)

ABSTRACT Camguilhem, R. and Milon, A., 1990. Protection of weaned rabbits against experimental Escherichia coli 0103 intestinal infection by oral formalin-killed vaccine. Vet. Microbiol., 21: 353362. In an attempt to vaccinate young weaned rabbits against life-threatening enterocolitis caused by Escherichia coli of the 0103 serogroup, 32 New Zealand male rabbits were divided into three groups. One group remained unvaccinated as a control (Group C), and each of the other groups received one of two types of vaccine prepared with E. coli strain 0103/10 cultured either in trypticase-soy broth (Group A) or in Minca agar (Group B). Bacteria were killed by formalin and administered per os for 10 consecutive days after weaning at a daily dose of 4 X 109 organisms. Six days after the last administration, all the animals were challenged with 1 X 104 virulent E. coli O 103/10 and the experimental infection was monitored for 26 days. All rabbits in Group A were protected from symptoms of disease and remained alive, whereas two rabbits in Group B developed clinical signs and one died. Protection did not correlate with local or general responses to lipopolysaccharide (LPS) 0103, as judged by measurement of anti-LPS O103 IgA in faeces or serum, or by serum agglutinating antibodies. Numbers of E. coli and E. coli 0103 were significantly lower in vaccinated animals of Group A as compared with animals of the control group. The differences between both vaccine regimens may be partially explained by a different expression of the adhesins of strain 0103/10, depending on the medium used to prepare the vaccine.

INTRODUCTION

In recent years, rabbit-fattening farms in Western Europe have experienced epizootics of enteric disease in weaned animals. The death rate (20-30%), as well as the growth retardation of diarrhoeic animals, imply substantial economic losses. The disease is associated with proliferation in the ileum and caecum of Escherichia coli apparently similar to human enteropathogenic strains (E.P.E.C.) or to the attaching-effacing group of E. coli strains

0378-1135/90/$03.50

© 1990 Elsevier Science Publishers B.V.

354

R. CAMGUILHEM AND A. MILON

(A.E.E.C.) (Takeuchi et al., 1978; Peeters et al., 1985; Okerman, 1987 ). E. coli strains enteropathogenic for weaned rabbits belong to particular serogroups (mainly O15 and O103) and biotypes (Peeters et al., 1988; Camguilhem and Milon, 1989 ). In France, the disease is mainly associated with strains of serogroup O103 (Camguilhem et al., 1986b; Camguilhem and Milon, 1989). The pathogenic mechanisms by which they induce diarrhoea include adhesion to the intestinal mucosa and probably the secretion of cytotoxins (Boedecker and Sherman, 1986; Levine, 1987; Okerman, 1987). Several authors have reproduced the disease by feeding field isolates to healthy weaned rabbits (Renault et al., 1983; Peeters et al., 1984; Camguilhem, 1985; Camguilhem et al., 1986a). Vaccination trials in weaned rabbits with formalin-killed O103 whole bacteria, by the intradermal route, did not protect the animals against oral challenge with pathogenic E. coli O103 (Camguilhem, 1986). We showed that parenteral or oral vaccination of females does not protect the weaned young (Milon and Camguilhem, 1989 ). We present here some evidence that successful protection of weaned rabbits against E. coli enterocolitis may be achieved by oral vaccination with formalin-killed whole cells, provided that the bacteria have the appropriate adhesins on their surface. MATERIALSAND METHODS E. coli strains E. coli O103/10 is a field isolate which has been used to reproduce enterocolitis in weaned rabbits (Camguilhem et al., 1986a). It belongs to a characterized rhamnose-negative biotype (Camguilhem and Milon, 1989). E. coli C124 was isolated from a diarrhoeic weaned rabbit, but gives no clinical signs after experimental inoculation. It belongs to the O103 serogroup and is rhamnose positive. E. coli strains 92B (O8:K99:Cs31A) and 217B (O?:Fy) are bovine isolates kindly provided by Dr. Girardeau, I.N.R.A., Theix, France. Animals

New Zealand male rabbits (I.N.R.A. strain A 1077) from eight different litters were used in this experiment. At 29 days of age (mean weight 528 + 70 g), they were weaned and randomly divided into three groups. Animals in two groups (A and B, eight animals each) were vaccinated with two different vaccines. Group C (16 animals) remained unvaccinated as a control group. During the vaccination period, vaccinated animals were housed in individual cages, while the 16 control animals were kept in four cages. After challenge, all the rabbits were redistributed in cages of two animals. They were fed throughout the experiment with a coccidiostatic-supplemented feed (Robenidine ® ) and watered ad libitum.

ORAL VACCINATION OF RABBITS AGAINST E. COLI O103

355

Vaccination and challenge protocols Two different vaccine preparations were used. Group A received Vaccine A: E. coli O103/10 strain was cultured for 24 h on trypticase-soy agar (Bio-Mdrieux, France), resuspended in Ringer buffer, inoculated into trypticase-soy broth (TSB) and incubated for 24 h at 37°C. The culture was then treated with 0.4% formalin and the cell concentration adjusted at 2 X l0 s colony-forming units (CFU) m l - 1by centrifugation and resuspension in the medium. Group B received Vaccine B: E. coli 0103/10 strain was cultured on Minca agar + Polyvitex ® (Bio-M~rieux, France) for 24 h at 37 ° C and resuspended in phosphate-buffered saline (PBS, pH 7.4) and formalin was added up to 0.4%. The cell concentration was 2 × 109 CFU m l - 1 Animals in Groups A or B received 2 ml of Vaccine A or B, respectively, daily per os for 10 consecutive days. Vaccine doses were given by oral cannula. Vaccination began 1 day after weaning and lasted from Day - 15 to Day - 6 before challenge. On Day 0 (45 days of age), animals of all three groups were challenged per os with 1 X 104 E. coli 0103/10 in 2 ml of saline. After challenge, animals were observed daily for diarrhoea and mortality. Individual weights were checked three times a week. Post-mortems were performed on animals that died. Caecal contents were examined and the number of E. coli present quantified by a dilution technique on MacConkey agar. On Day +13 and Day +26 (i.e., at slaughter), E. coli were enumerated in the faeces or caecal contents of the survivors by the same technique. On each occasion, three colonies were subcultured and tested with anti-O103 antiserum by slide agglutination. Coccidia were also counted each time.

Antibody monitoring Blood was sampled by cardiac puncture. Fresh faeces (2 g) were diluted in 8 ml of PBS containing sodium azide 0.02% (1/5 initial dilution, w/v). After trituration with glass beads and centrifugation for 20 min at 3500 × g and + 4 ° C, the supernatant was kept at - 20 ° C until assayed. The enzyme-linked immunosorbent assay (ELISA) technique used to measure the faecal or serum anti-lipopolysaccharide (LPS) O103 IgA has been described elsewhere (Milon and Camguilhem, 1989). Faeces were tested at a final dilution of 1/20 and sera at a dilution of 1/1000. Results are expressed as milliunits (mU) of specific optical density (SOD). Agglutinating antibodies were titrated in sera using a suspension of 5 × l0 s CFU ofE. coli O103/10 per ml of saline plus formalin 0.4% as antigen. Serial 2-fold dilutions of sera, starting at 1/20, were tested. Agglutinations were read after overnight incubation in tubes at 56 ° C.

356

R. CAMGUILHEM AND A. MILON

In vitro adhesion tests Adhesion to rabbit intestinal villi of strain O103/10 after different culture conditions was tested by a modification of the technique described by Girardeau (1980). Eight-day-old rabbit jejuno-ileal villi were prepared and preserved as described. For the adhesion test, suspensions of 5 X l0 s CFU ml-1 were prepared in 0.01 M PBS (pH 6.4). Bacteria grown in Minca agar + Polyvitex ® were gently flushed with buffer and resuspended by slow agitation on a rocking platform, at 6 revolutions min-1 during 30 min. Bacteria grown in broth (TSB or Penassay broth, Difco Laboratories, Detroit, MI) were incompletely pelleted by slow centrifugation at 600 Xg and 20 ° C during 20 min. Broth was discarded, PBS at pH 6.4 was added and the cells were gently resuspended by slow agitation as above. Cell concentrations were adjusted by spectrophotometry. Intestinal villi and bacterial suspensions were mixed in 24-well plates and agitated on a Kline apparatus for 20 min at 37°C. Adhesion was then checked qualitatively by phase-contrast microscopy at a magnification of X 400. When positive, confluent bacterial sheets could easily be seen sticking to the brush borders of the villi or on their upper face, depending on the focus of the microscope.

Electron microscopy To detect pili, standard electron microscopic techniques were used. The strains were grown for 5 h in appropriate media, adsorbed on Parlodion ®- and carbon-treated grids, negatively stained with 3% phosphotungstic acid (pH 6.6) and examined in a Philips electron microscope, at magnifications of X 50 000- X 100 000. RESULTS

Mortality and morbidity following challenge No symptoms were noted in rabbits in Group A. In the control, Group C, 10 of 16 animals (62.5%) experienced periods of weight loss, six (37.5%) showed profuse watery diarrhoea beginning 9-19 days after inoculation; four (25%) died 2-6 days after the onset of diarrhoea. In the vaccinated Group B, two animals suffered growth retardation and only one had fatal diarrhoea.

Growth parameters of survivors [oUowing challenge Growth in the vaccinated Group A was excellent, with a mean weight gain of 1046 g recorded at slaughter, 26 days after challenge, and a mean weight gain

357

ORAL VACCINATION OF RABBITS AGAINST E. COLI O103

of 40.2 g per day. Growth of survivors in the control Group C was significantly less (776 g at Day +26; 29.8 g per day) and survivors in Group B gave an intermediate value (908 g at D a y + 26) which did not differ significantly from the control group. Enumeration of caecal or faecal E. coli

The caecal contents of the five dead animals of Groups B and C were very liquid, b u t did not contain any blood. There was no evidence of haemorrhagic lesions on the mucosa. N u m b e r s of E. coli reached 7 × 10s-1 × 101° CFU g - ' . All colonies tested belonged to the O103 serogroup. Two of five animals had a few coccidia in their caecal contents (respectively, 4 × 102 and 5 × 103 oocysts g-l). E. coli numbers in the faeces (Day + 13) and caecal contents (Day + 26) of survivors are shown in Table 1. In all groups, the mean numbers were lowered by ~ 2 logs between Day + 13 and Day +26. This mean was 2 logs higher in the control Group C than in Group A. It is worth noting that the O103 strain had disappeared as early as Day + 13 in Group A, while it was still present at slaughter in Group C only. Two animals in Group B and three in Group C had low rates of coccidia in their faeces (2 × 102-5 × 103 oocysts g-1 ). Antibody responses

Anti-LPS O103 IgA was found in the faeces of some animals of all groups before vaccination, i.e., 1 day after weaning. These antibodies were probably transferred from the mother by the milk. We showed that such antibodies could be found in the lactoserum of several non-vaccinated does (Milon and Camguilhem, 1989). Eight days after the beginning of oral vaccination, these antibodies had almost completely disappeared from the faeces of all groups and, TABLE 1 Mean numbers ofE. colig- 1and proportion of 0103 strains recoveredfrom the faeces (Day + 13 ) or caecal contents (Day + 26) of survivor rabbits Group

A B C

Day + 13 after challenge (faeces)

At slaughter (Day ÷ 26) (caecal contents)

Mean E. coli (CFU g-l)

No. of 0103 strains/ No. of tested strains

Mean E. coli (CFU g-l)

No. of 0103 strains/ No. oftested strains

1.3 × 104 5.4× 105 1.8×10 ~

0/12"* 9/12 20/24

8.7× 102 1.2× 103 2 ×104

0/18" 0/15" 9/33

*AorB
358

R. CAMGUILHEMANDA. MILON

TABLE 2

Mean anti-LPS 0103 IgA antibody levels in the faeces of animals before (Day - 15 ), during (Day - 7 ) and after vaccination (Day - 1 ) , and after challenge (Day + 16) with E. coli 0103/10 as measured by the ELISA technique. The results are expressed in mU SOD given by a final dilution of 1/20 (w/v) of faeces Group

A B C

Anti-LPS 0103 I g A J ~ ( S E M ) (mU SOD) Day - 15

Day - 7

Day - 1

Day + 16

198 (128) 107 (69) 257 (162)

2 (2) 2.3 (1) 0

24 (20) 153 (86) 17 (17)

76 (54) 228 (184) 281 (145)

TABLE 3

Antibody levels in serum after vaccination and after challenge as measured by the ELISA technique (anti-LPS O 103 IgA in mU SOD) or by tube agglutination (first dilution tested: 1/20 ) Group

A B C

Anti-LPS O103 IgA )~ (SEM)

log2 agglutinin titre/10/~ (SEM)

Day - 1

Day +26

Day - 1

Day +26

59 (38) 111 (32) 4 (1.6)

22 (7) 89 (69) 34 (40)

2.875 (0.89) 5.000 (0.84) 0.187 (0.2)

0.75 (0.49) 2.14 (0.59) 2.25 (0.67)

just before challenge, they were irregularly detected, but at very low levels in the control group (one positive reaction: SOD = 69) and Group A (range: 0167 mU SOD) (Table 2). Responses were higher in Group B (range: 2-728 mU SOD). After challenge, the faeces from most cages showed anti-0103 IgA, except those of Group A, in which four of eight animals were negative. Oral vaccination elicited moderate levels of serum antibodies (Table 3), which could be detected by the agglutination technique using the homologous strain as antigen. These antibodies were not found in control animals before challenge (Table 3). The mean titre elicited by vaccination was higher in Group B than in Group A. Anti-O103 antibodies of the IgA class were scarcely detected in serum (Table 3 ). After challenge, most of the control animals became serologicaUy positive, whereas the titres from animals in Group A decreased.

Adhesion and electron microscopy of the 0103/10 strain after different culture conditions Vaccination results in Group A showed that Vaccine A is effective. The difference between Vaccine A and the less efficient Vaccine B can only be ex-

ORAL VACCINATION OF RABBITS AGAINST E. COLI 0103

359

plained by the different methods of culture. To explain the discrepancy in these results, we tried to find which strain characteristics might be impaired in the strain cultured on Minca medium compared with that on TSB. As adhesion to intestinal enterocytes has been shown to be important for the pathogenicity of E.P.E.C. and A.E.E.C., we tested the adhesion properties of the O 103 / 10 strain after culture in Minca medium, TSB, and Penassay broth, the latter being a medium already described which favours the expression of adhesins in the rabbit E.P.E.C. RDEC-1 strain {Cheney et al., 1983). Concordant results were given by six different manipulations, using villi from three different rabbits. These results clearly show that the O103/10 strain is adherent only if grown in Penassay broth or TSB, but not after culture in Minca medium, in contrast with the non-pathogenic C124 strain (non-adherent whichever medium was used) or with strains 92B or 217B, which showed the best adhesion after Minca culture. It must be stressed here that adhesion tests were carried out at pH 6.4, which has been shown to be intermediate between the pH of the normal rabbit caecum ( ~ p H 6.0) and the caecal pH of diarrhoeic experimentally infected rabbits (pH 6.6) (Camguilhem et al., 1986a). At pH 7.2, the 0103/10 strain did not adhere to the villi. Concurrently, O 103/10 bacterial cells were shown to be piliated after culture in Penassay, but not after culture in Minca, in good correlation with the results of adhesion tests. Therefore, it may be concluded that the expression of timbrial adhesins by the enteropathogenic strain O103/10 seems to be linked to culture conditions. DISCUSSION When applied to strict mucosal infections, such as the enteritis caused by enterotoxigenic or enteropathogenic E. coli, vaccination must be effective in inducing local immune effectors which may prevent the gut colonization a n d / or toxic effects of the bacteria on the gut epithelium (Mestecky, 1987). We have already shown that oral administration of killed E. coli O103/10 induced both local and systemic responses when administered per os to adult does, as judged by anti-LPS O103 IgA responses in faeces, milk and serum, and by serum agglutinins as well (Milon and Camguilhem, 1989). Yet, transfer of these effectors to young animals through milk was ineffective in inducing post-weaning protection. It was therefore of interest to see whether these weaned animals could be protected by an orally administered vaccine, given after weaning. In this experiment, we undoubtedly reproduced E. coli 0103 enteritis in the control Group C. Diarrhoea, mortality of 25% of challenged control animals, as we!l as colonization of the caecum by E. coli O103 allow us to conclude that the disease was due to the challenge dose, despite the presence of a few coccidia. The observed mortality was similar to that seen on French industrial farms with E. coli diarrhoea problems.

360

R. CAMGUILHEM AND A. MILON

The results observed in Group A showed a good efficacy of Vaccine A, in the described schedule, in preventing illness. No symptoms were observed and growth was excellent during the observation period. Furthermore, no E. coli O 103 could be recovered from these animals 13 days after challenge, suggesting that the challenge organisms could not colonize the intestinal tract. In Group B, results were not significantly different from those of the control group, although diarrhoea and mortality appeared somewhat lower and the growth rate slightly higher. It may be considered that Vaccine B has brought about partial protection. Strain O103 was still excreted in the faeces of the Group B animals 13 days after challenge, but was no longer found 26 days after challenge, in contrast to control animals. Protection against challenge seems to be poorly correlated with local or general antibody responses against LPS. In faeces, the levels of anti-LPS O103 IgA were irregular during the experiment. Initial levels detected at weaning were probably of maternal origin. The amounts quickly decreased to very low levels during vaccination (Day - 7, Table 2), probably through normal catabolism of passively transferred maternal IgA (the half-life of IgA does not exceed 3 days) or through complexation of the local antibodies with the large amounts of O103 bacteria introduced in the gut. Levels recovered 4 days after the end of the vaccination protocol were also very low (Day - 1 , Table 2) and were somewhat elevated after challenge. In both these cases, the levels recorded in protected Group A were lower than those of unprotected Group B. The absence of colonization of the gut by the challenge strain may explain this result. It is well known that large doses of antigen or repeated/continuous contacts are required for an effective and long lasting secretory IgA response in orally immunized individuals (Mestecky, 1987 ). Nevertheless, we must conclude that protection is not correlated with IgA responses against bacterial LPS, in contrast with the results obtained in rabbit enterotoxigenic E. coli infection (Sack et al., 1988). The apparent discrepancies in the efficacy of the two vaccines may only be explained by the different methods used to prepare the vaccines. Vaccine B contained only bacterial cells cultured on Minca medium, whereas Vaccine A was formed by the whole TSB culture of the strain, i.e., it contained bacterial cells and the liquid medium. Thus, Vaccine A may contain antigenic substances excreted during the growth of bacteria which may be of importance for subsequent protection. As in all E.P.E.C. strains found in humans, strain O103/ 10 may produce, for instance, cytotoxins (or shiga-like toxins ) involved in the haemorrhagic lesions which are found in the acute forms of the illness (Okerman, 1987) and local antibodies against these toxins may be implicated in the protective mechanism. Vaccine A may include these toxins, in an excreted form, whereas Vaccine B may not. Work is in progress in our laboratory to test this hypothesis. Another important pathogenic mechanism of enteritic E. coli, especially of

ORALVACCINATIONOF RABBITSAGAINSTE. COLIO103

361

E.T.E.C. and E.P.E.C., is their ability to adhere to enterocytes, frequently by means of fimbrial adhesins, which allows their colonization of the gut and the delivery of toxins in close contact with the epithelium (Levine, 1987). Local anti-adhesin responses have been shown in many models to be protective against infection. In rabbits, E.P.E.C.-like E. coli have been studied, especially through the RDEC-1 strain described by Cantey and Blake (1977). This strain is pathogenic for weaned rabbits and induces diarrhoea with clinical signs similar to those induced by O103 strains. It adheres to rabbit intestinal brush borders through a plasmid-encoded fimbrial adhesin called AF/R1, which is specific for a receptor present on the enterocytes of rabbits > 3 weeks of age and is expressed in vitro under defined growth conditions only (Cheney et al., 1983; Boedecker and Sherman, 1986). Thus, the differential expression of fimbrial adhesins in our vaccine preparations could also explain the discrepancies in their efficiency. As shown here, strain O103/10 seemed to express adhesive properties, as judged by in vitro adhesion to rabbit intestinal villi, that depend on the culture conditions of the strain: the adhesion test was positive after TSB or Penassay culture only, but negative after Minca culture. Furthermore, the strain appeared piliated only when cultured in conditions where they were shown to be adhesive. Phenotypic suppression of fimbriae and adhesion have already been described in the RDEC-1 strain: the expression of AF/R1 fimbrial adhesin was suppressed when the strain was cultured in brain-heart infusion broth (Cheney et al., 1983). Therefore, the lack of expression of the strain adhesin may explain the relative inefficiency of the vaccine prepared by Minca culture (Vaccine B) as compared with Vaccine A. It is worth noting here that the fimbrial adhesin described, although needing more study, already seems different from the fimbrial AF/R1 since it allows adherence to 8-dayold rabbit villi as well as to weaned rabbit villi (data not shown). ACKNOWLEDGEMENT

This work was partly supported by the French Institut National de la Recherche Agronomique.

REFERENCES Boedecker, E.C. and Sherman, P.M., 1986. Mechanisms ofEscherichia coli enteritis. Front. Gastrointest. Res., 13: 309-330. Camguilhem, R., 1985. Isolement d'une souche d'Escherichia coli (serogroupe 0103 ) responsable d'enterite colibacillaire du lapin en engraissement. Mise en 6vidence de son pouvoir pathog~ne. Rev. Med. Vet., 136: 61-68. Camguilhem, R., 1986. Essai de vaccination des lapins par voie intradermique contre l'ent6rite colibacillaire ~ E. coli O103.4~mes Journ6es de la recherche cunicole INRA-ITAVI, Commun. No. 35.

362

R. CAMGUILHEMANDA. MILON

Camguilhem, R. and Milon, A., 1989. Biotypes and O serogroups of Escherichia coli involved in intestinal infections of weaned rabbits: clues to diagnosis of pathogenic strains. J. Clin. Microbiol., 27: 743-747. Camguilhem, R., Lebas, F. and Labie, C,, 1986a. Reproduction experimentale chez le lapin en engraissement d'une diarrh6e provoqu6e par une souche d'Escherichia coli de s6rogroupe 0103. Ann. Rech. VeL, 17: 409-424. Camguilhem, R., Mureau, G., Nicolas, J.A., Brocas, J, and Tournut, J., 1986b. Groupage s6rologique O et antibiosensibilit6 des souches d'Escherichia coli isol6es en France sur les lapins diarrh6iques apr~s le sevrage. Rev. Med. Vet., 137: 205-212. Cantey, J.R. and Blake, R.K, 1977. Diarrhea due to Escherichia coli in the rabbit: a novel mechanism. J. Infect. Dis., 135: 454-462. Cheney, C.P., Formal, S.B., Schad, P.A. and Boedecker, E.C., 1983. Genetic transfer of a mucosal adherence factor ( R1 ) from an enteropathogenic Escherichia coli strain into a ShigeUaflexne,'i strain and the phenot~2oic suppression of this adherence factor. J. Infect. Dis., 147: 711-723. Girardeau, J.P., 1980. A new in vitro technique for attachment to intestinal villi using enteropathogenic Eschcrichia coli. Ann. Microbiol. (Inst. Pasteur), 131B: 31 37. Levine, M.M., 1987. Eseherichia coli that cause diarrhea: enterotoxigenic, enteropathogenic, enteroinvasive, enterohemorragic and enteroadherent. J. Infect. Dis., 155: 377-389. Mestecky, J., 1987. The common mucosal immune system and current strategies for induction of immune responses in external secretions. J. Clin. Immunot., 7:265 274. Milon, A. and Camguilhem, R., 1989. Essais de protection des lapereaux sevr6s contre l'ent&ite 5 Escherichia eoli O-103: vaccination des m~res avec un vaccin inactive. Rev. Med. Vet., 140: 389 395. Okerman, L., 1987. Enteric infections caused by non-enterotoxigenicEseherichia coli in animals: occurrence and pathogenicity mechanisms. A review. Vet. Microbiol., 14: 33-46. Petters, J.E.. Pohl, P., Okerman, L. and Devriese, L.A., 1984. Pathogenic properties of Esche richia coli strains isolated from diarrheic commercial rabbits. J. Clin. Microbiol., 20: 34-39. Peeters, J.E., Charlier, G.J. and Raeymaekers, R., 1985. Scanning and transmission electron microscopy of attaching effacing Escherichia coli in weaning rabbits. Vet. Pathol., 22: 54-59. Peeters, J.E., Geeroms, R. and Orskov, F., 1988, Biotype, serotype and pathogenicity of attaching and effacing enteropathogenic Escherichia coil strains isolated from diarrheic commercial rabbits. Infect. Immun., 56:1442 1448. Renault, L., Roux, J., Le Bourhis, E., Coudert, P., Licois, D. and Guillot, J.F., 1983. Description d'un s~rogroupe (O 103) d'Escherichia coli enteropathog~ne chez le lapin au sevrage. Bull. Acad. V&. Ft., 56: 387-400. Sack, R.B., Kline, R.L. and Spira, W.M., 1988. Oral immunization of rabbits with enterotoxigenic Escherichia coli protects against intraintestinal challenge. Infect. Immun., 56: 387-394. Takeuchi, A., Inman, L.R., O'Hanley, P.D., Cantey, J.R. and Lushbaugh, W.B., 1978. Scanning and transmission electron microscopic study of Escherichia coli 015 (RDEC-1 ) enteric infection in rabbits. Infect. Immun., 19:686 694.