key: cord-007481-4mj5isyl
authors: Chanter, N.; Hall, G.A.; Bland, A.P.; Hayle, A.J.; Parsons, K.R.
title: Dysentery in calves caused by an atypical strain of Escherichia coli (S102-9)
date: 2002-11-13
journal: Vet Microbiol
DOI: 10.1016/0378-1135(86)90053-2
sha: 
doc_id: 7481
cord_uid: 4mj5isyl

Dysentery lasting 4–8 days was produced in five 4-day-old colostrum-fed calves, after inoculation with an atypical strain of Escherichia coli S102-9; peak excretion of S102-9 occurred during the period of dysentery. Two calves were killed when clinical signs were most severe and bacteria were seen attached to the surfaces of enterocytes in the large intestine; microscopic lesions were seen in these areas. The lesions were identical to those previously reported in a natural outbreak of dysentery in calves, from which E. coli S102-9 was isolated, and to those seen in gnotobiotic calves experimentally infected with S102-9. Reinfection of the three surviving calves 16–20 days later with S102-9 and primary infection of two calves aged 24 and 51 days did not cause dysentery. Four of 659 coliforms isolated from field outbreaks of calf diarrhoea resembled the atypical strain S102-9. These four isolates and S102-9 did not produce heat-stable enterotoxin, but all produced a toxin cytopathic for Vero and HeLa cells. Two of the four isolates were inoculated alone into 4-day-old gnotobiotic calves deprived of colostrum; neither calf developed dysentery but microscopic lesions identical to those produced by S102-9 were detected in the large intestines of both animals.

An atypical strain of Escherichia coli (designated $102-9) was isolated from the faeces and intestinal contents of farm calves aged 8--21 days old during an outbreak of dysentery (Hall et al., 1985) and reproduced the disease after experimental infection of gnotobiotic calves . The distinctive characteristics of $102-9 included an atypical colony morphology on MacConkey agar, production of urease and anaerogenicity, but it was identified as E. coli in particular by its abilities to produce acid in MacConkey broth at 44°C and indole at 44°C. $102-9 was shown by an immunoperoxidase method to adhere to the mucosae of the large bowel in gnotobiotic and farm calves with dysentery, and microscopic lesions were seen in the colonic and rectal mucosae (Hall et al., 1985) . These were identical to those seen in piglets (Moon et al., 1983) , man (Rothbaum et al., 1982) and rabbits (Takeuchi et al., 1978) infected with E. coli which do not produce the classical enterotoxins and which are not invasive. The objectives of the experiments reported here were firstly to establish the pathogenicity of E. coli S102-9 for conventional calves, secondly, to determine whether E. coli with atypical characteristics were an important cause of calf diarrhoea by examining a culture collection of 659 isolates from field outbreaks of diarrhoea, and thirdly, to investigate by experimental infection of gnotobiotic calves, the association between the atypical characteristics of these isolates and pathogenicity. Electron microscopy was used to examine more precisely lesions produced by atypical E. coli and to compare them with those described previously.

Experimental design Five Friesian cross calves were infected with E. coli $102-9 and compared with two uninfected calves for signs of dysentery. Two of the five infected calves were killed when clinical signs were maximum to investigate the presence of lesions. Immunity and age resistance to infection were investigated by allowing the other three calves to recover and then reinfecting them and comparing the results with two primarily infected calves, one age matched and one 27 days older.

The nine calves were left with their dams for 12 h, then moved to a loose box, separate from other calves, and fed on a milk replacer diet. On the fourth day, the calves were moved to rooms in an isolation unit; they were fed sterilised canned milk. Infected and control animals were housed in separate rooms. Calves were inoculated orally after the morning feed; 5 ml of sterilised canned milk (Carnation Ltd.) containing 5.0 X 109 --2.3 X 101° colony forming units (cfu) of E. coli grown on bovine blood agar at 37°C for 18 h, was given by syringe. Calves 1--5 were infected when 4 days old; Calves 4 and 5 were killed on days when clinical signs were judged to be most severe. Calves 1--3 were allowed to recover and then reinfected when 20, 22 and 24 days old, respectively. Calves 6 and 7 were infected for the first time at 24 and 51 days old respectively and were controls for the reinfection of Calves 1--3. Calves 8 and 9 were housed together until 42 and 41 days of age, respectively. Faeces were sampled daily. Blood samples were taken at 4 days, at reinfection and daily from calves with dysentery. Rectal temperatures were measured twice daily.

The strain of atypical E. coli S102-9 (05:K-:H-) was used .

Rotavirus and coronavirus were detected by enzyme linked immunosorbent assays of faecal samples taken daily and calicivirus by electron microscopic examinations of faeces taken on the first day of illness (Bridger and Hall, 1979) .

Smears of faeces stained with Giemsa were examined for cryptosporidial oocysts on the first and last day of illness.

E. coli were enumerated in scrapings of mucosa from the ileum, caecum, colon and rectum, and in faeces and intestinal contents as described previously . The mucosae were washed, scraped off with a glass slide and ground to make 10 -1 suspensions in saline. One gram of the mucosal suspensions, intestinal contents or faeces were serially diluted 10fold in saline and 0.1 ml of appropriate dilutions spread on to MacConkey agar plates (Mackie and MacCartney, 1953} in triplicate before incubation at 37°C for 18 h. E. coli g-1 were enumerated according to the dilution of sample and the number of lactose fermenting colonies per dilution. Atypical E. coli were initially identified by their characteristic colonies on Mac-Conkey agar; these had a red centre, a clear outer zone and the surrounding medium was clear. Isolates from experimental infections were occasionally tested for anaerogenicity and production of urease. Faeces taken from calves aged 2 and 4 days and at onset of diarrhoea or dysentery were examined for salmonellae (Jones et al., 1983) .

Tissues were fixed, processed and examined for the presence of lesions as described previously (Hall et al., 1985) .

Paraffin sections of mercuric formol fixed tissue were stained by an indirect immunoperoxidase method with primary antisera to bovine rotavirus prepared in calves (Parsons et al., 1984) or by a peroxidase-antiperoxidase method (Sternberger et al., 1970) with primary rabbit antisera to live E. coli S102-9 produced by the method of Sojka (1965) .

Immunoglobulins were detected in calf serum taken at 4 days of age by the zinc sulphate turbidity test (McEwan et al., 1970) .

Serum samples were serially diluted in 0.1-ml volumes of isotonic saline in a U-bottomed microtitre tray. A suspension of E. coli S102°9 (0.1 ml equivalent to Browns tube No. 4) was added to each well and to wells con-taining saline alone. Tests were incubated at 37°C for 1 h followed by 18 h at 4°C. Rabbit antisera to live E. cpli S102-9 was used as a positive control.

During a survey of field cases of calf diarrhoea in southern England during the winters of 1981/82 and 1982/3 isolates that grew on MacConkey agar (Mackie and MacCartney, 1953) at 37°C and fermented lactose, were subinoculated into vials of ferrous metabisulphate pyruvate (FBP) semi-solid medium (Oxoid) and were frozen at -20°C. Stored cultures (total 659) were revived by scraping frozen medium from vials onto trypticase glucose extract (TGX) agar (Orskov et al., 1975) , subsequently incubated at 37°C for 24 h. E. coli strains K12 and B41 (0101 :K99) were used as controls in characterisation tests.

Tests described by Cowan and Steel {1974) were used. E. coli B41 was used as a typical control.

Serotyping of E. coli Rabbit antisera to live E. coli S102-9, produced by the method of Sojka (1965) , was used in a slide agglutination test. Further typing of somatic antigens was kindly undertaken by Dr. B. Rowe at the Central Public Health Laboratory, Colindale, London.

The microtitre method of Burrows et al. (1976) was used. Tests were incubated at 4°C or 37°C in the presence or absence of 1% D-mannose.

Cytotoxin test E. coli isolates which were to be tested for production of cytotoxin were inoculated into 10 ml of Evans medium (Evans et al., 1973) which was incubated at 37°C for 24 h and shaken at 150 r.p.m. A 0.1-ml aliquot of this culture was subinoculated into a second Evans medium and treated in the same way as the first. Cultures were then centrifuged at 1000 X g for 20 min and the supernatant fractions filtered through an 0.45-pm Millipore membrane and the filtrate tested for the presence of toxin. Aliquots of 180 pl of l0 s Chinese hamster ovary (CHO) or HeLa cells m1-1 or 2 × l0 s Vero cells mY 1 were placed into all wells of a flat-bottomed microtitre tray (Intermed, Nunc); 20 pl of filtrate was placed in a duplicate set of wells and serially diluted. Microtitre trays were sealed and incubated in a CO2 cabinet at 37°C for 3 days. Cells were stained with a mixture of equal volumes of 0.2% (w/v) crystal violet and 12% (v/v) neutral buffered formalin for 1 h, washed, dried at 37°C and examined.

Toxin was prepared in Evans medium and tested in the suckling mouse test (Guinee et al., 1981) .

Bacteria were washed from an overnight culture on either 5% sheep blood agar or TGX medium with phosphate buffered saline (pH 7.2) and dropped onto Formvar/carbon-coated grids. After 30 s, excess fluid was removed with filter paper. The grids were rinsed in distilled water and then stained with 2% (w/v) phosphotungstate (pH 7.0), excess stain removed, air-dried and examined in a Philips 300 electron microscope.

Experimental infection of gnotobiotic calves with atypical E. coli isolates 37/1 or 6/193 Bacteria Two out of four atypical isolates from the culture collection were selected for experimental infections. E. coli 37/1 had been isolated from a calf with diarrhoea from which no other enteropathogen was detected and E. coli 6/193 had been isolated from a calf which did not have diarrhoea.

Gnotobiotic calves were derived by the methods of Dennis et al. (1976) . One was infected with E. coli 37/1 and another with E. coli 6/193. Infection of calves, pathology and immunoperoxidase staining methods These methods were as described for the experimental infection of conventional calves with E. coli S102-9.

The occurrence of dysentery in calves and the excretion of enteropathogens are summarised in Table I . All five calves (1--5) infected with E. coli S102-9, at 4 days of age, developed dysentery, whereas older calves (6--7) infected for the first time or reinfected (1--3), and control calves (8--9) did not. The dysentery was characterised by the presence in faeces of bright, fresh blood (sometimes copious) and large amounts of clear or faeces-stained mucus. The faecal consistency was soft or semi-liquid.

Peak excretion of S102-9 (>108 cfu g-i of faeces) occurred during signs of dysentery in all 5 calves (1--5) infected at 4 days of age. In contrast only 2 (1 and 2) of the 5 older calves {1--3, 6 and 7) excreted these numbers; this occurred on the first day after reinfection.

Rotavirus or coronavirus were excreted by Calves 1--3, but not at times Infection Occurrence of Excretion Excretion Excretion of dysentery of S102-9 when dysentery was seen. Calves 4 and 5, infected at 4 days of age and necropsied at the peak of dysentery (8 days), were excreting rotavirus whereas control calves (8--9) excreted rotavirus at times when mild diarrhoea was seen. Salmonellae, cryptosporidia and calicivirus were not detected. Pyrexia or bacteraemia did not occur consistently during the period of dysentery. Prior to infection, the serum of calves contained widely different amounts of immunoglobulin (2--33 zinc sulphate turbidity units). Prior to infection the amount of serum immunoglobulins of Calf 3 was the lowest recorded. Calf 3 alone responded to infection by the production of agglutinins to E. coli S102-9 (titre of 10), and this calf had the longest period of dysentery and the longest period of excretion of $102-9.

In Calves 4 and 5, necropsied at the peak of dysentery, E. coli $102-9 colonised the colonic and rectal mucosae in approximately equal or greater numbers than in the contents. In contrast E. coli of other types colonised the mucosae in fewer numbers than in the contents (Table II) . The mean number of S102-9 in these mucosae was log10 7.9 g-1 whereas for other E. coli it was log~0 6.6 g-1. The same distribution of E. coli was seen in the caecum of Calf 4. E. coli of all types colonised the mucosae of the small intestine in lower numbers than in the contents.

Petechial haemorrhages were seen in the rectal mucosae of Calves 4 and 5; all of the rectal mucosa of Calf 4 and the longitudinal folds of Calf 5 were markedly reddened. The same changes were seen in the colon of Calf 4 and were most noticeable over the gut-associated lymphoid tissue of the anterior colon and along the tops of longitudinal folds. In the caecal mucosa of Calf 4 there were areas of reddening 1--2 cm in diameter.

In the caecum, colon and rectum of Calves 4 and 5 foci of bacteria were seen adherent to clumps of irregularly arranged and exfoliated enterocytes and there were neutrophils in the lamina propria. In the caecum, neutrophils were in foci in the lamina propria; in the colon and rectum, neutrophils were more numerous and were also present in foci in an exudate on the mucosal surface with mucus and exfoliated enterocytes. Lesions were most severe in the colon.

Immunoperoxidase staining of sections of tissue with antiserum to E. coli S102-9 revealed microcolonies of bacteria adherent to the colonic and rectal rugae of Calves 4 and 5. All adherent bacteria that were seen were immunostained. Rotavirus was detected in the ileal enterocytes by immunoperoxidase staining and in the small intestine there were lesions of a viral enteropathy (stunted and fused villi covered by cuboidal enterocytes).

Of 659 isolates, including 373 from calves with diarrhoea, four had an atypical colony morphology on MacConkey agar indistinguishable from that of E. coli S102-9, were anaerogenic and produced urease. On the basis of other tests these isolates were identified as E. coli (Table III) . The four isolates originated from different farms and three were isolated from calves with diarrhoea. Coronavirus was isolated from one of these calves and Sa/monella typhimurium from another; the third isolate from a normal calf was The results of further characterisation of the four atypical isolates and S102-9 are in Table IV . Preliminary serotyping of E. coli 37/1 and 6/193 revealed that they produced the same somatic antigen (05) as S102-9. In addition to the four atypical isolates, another isolate, which was aerogenic but otherwise indistinguishable from E. coli S102-9, produced a toxin active on HeLa and Veto cells but not CHO cells. Fifty-five isolates were detected with either 1 or 2 of the atypical characteristics of E. coli S102-9; 16 of these were tested for the production of cytotoxins. They comprised four isolates which produced urease, five anaerogenic isolates, five which agglutinated in antiserum to E. coli S102-9, one with an atypical colony morphology and one with an atypical colony morphology which agglutinated in antiserum to E. coli $102-9. One isolate produced a cytotoxin active on Veto cells; it also produced urease.

In a preliminary investigation, two gnotobiotic calves were infected with either E. coli 37/1 or 6/193; they did not develop dysentery, although the calf infected with E. coli 37/1 produced mucoid liquid faeces from 5 to 9 days after infection, when it was killed. E. coli 6/193 was isolated from a normal calf and the faeces of the gnotobiotic calf infected with 6/193 were unchanged until 10 days after infection, when the animal was killed. This Absorption of antibody to S102-9 a Fimbriae (F) or haemagglutinins (H) produced on suggested that isolate 6/193 was avirulent. Calves excreted between 1.2 X 109 and 2.2 X 10 l° E. coli g-1 of faeces until they were killed. Three days after infection, faeces of the gnotobiotic calf infected with E. coli 37/1 inoculated onto MacConkey agar revealed E. coli with two different colony morphologies in approximately equal numbers. One was atypical and of the kind produced by E. coli $102-9, and the other was typical of most isolates of E. coli. E. coli of both types, however, were anaerogenic, produced urease and agglutinated in antiserum to E. coli S102-9. At necropsy of the two gnotobiotic calves there were more E. coli in the caecal, colonic and rectal contents (~3.9 X 109 cfu g-l) than in the respective mucosae (8.3 X 106 --4.8 X 108 cfu g-l) which indicated that colonisation of the large intestinal mucosae was less than that seen with E. coli S102-9 infections of gnotobiotic calves . In the mucosae of the colon and rectum of both calves there was marked congestion and occasional small haemorrhages. Neutrophils were numerous in the lamina and they were present as foci on the mucosal surface forming an exudate, together with mucous and exfoliated enterocytes. These lesions were apparently as severe as those seen with E. coli S102-9 infections of gnotobiotic calves {Hall et al., 1985) .

Scanning electron microscopy revealed many short, rod-shaped bacteria attached to the mucosal surface of the large intestines of both calves. Enterocytes to which bacteria were attached, and adjacent enterocytes, lacked microvilli or microvilli were abnormally orientated or were shortened or lengthened. Changes to microvilli were confirmed by transmission electron microscopy which also showed that the bacteria were closely associated with the enterocyte cell membrane. At the point of attachment, enterocyte cytoplasm was often cup-shaped or arranged as a pedestal.

Immunoperoxidase staining revealed microcolonies of E. coli adherent to the colonic and rectal rugae of both calves.

E. coli S102-9 was pathogenic for conventional calves; all animals infected at 4 days of age developed dysentery similar to that previously seen in gnotobiotic and farm calves Hall et al., 1985) . The calves excreted other enteropathogens, but in two these were not associated with dysentery, and in a third they were not excreted during the major episode of dysentery. In two calves, slaughtered at the estimated peak of clinical signs, rotavirus was detected in the ileum and was not associated with lesions that could have been a source of fresh blood in faeces. Lesions in the large bowel, which appeared to be the source of blood in faeces, were consistently associated with adherent E. coli and were identical in nature to those produced in gnotobiotic calves and those seen previously in naturallyaffected farm calves (Hall et al., 1985) .

Primary infection of two older calves did not reproduce dysentery, which suggested that the calves had acquired resistance to the effects of infection. In this respect, resistance might be similar to that acquired with age by calves to infection with K99 ÷ E. coli (Smith et al., 1967) . Detection of four atypical E. coli out of 659 isolates collected from natural outbreaks of calf diarrhoea suggested that they were not an important cause of enteric disease in calves in that survey. One isolate with these properties has previously been associated with diarrhoea in a calf in France (De Ryke et al., 1982) .

Two isolates of atypical E. coli (37/1 and 6/193) caused microscopic lesions indistinguishable from those produced by E. coli S102-9 (Hall et al., 1985) . However, one (6/193) did not cause disease and the other (37/1) caused mucoid diarrhoea. Failure of these strains to cause dysentery may have been related to the extent of bacterial colonisation and the lesions caused. There were approximately 10-fold fewer E. coli 37/1 or 6/193 in the mucosae of the large intestine than were previously reported for E. coli $102-9 in gnotobiotic calves . The extent of microscopic lesions in the gnotobiotic calves was difficult to assess, particularly because of their focal nature, which was apparently similar to that seen in calves infected with E. coli S102-9. Consequently, the virulence of these E. coli may vary not only with the ability to colonise but also the extent of the lesions they cause or, alternatively, it may differ with the production of an unknown pathogenic determinant. In man infected with different strains of verotoxic E. coli, which cause the same type of lesion, signs of disease vary from mild diarrhoea to dysentery (Anon, 1983 ).

An alternative explanation for the failure of E. coli 37/1 and 6/193 to cause dysentery may have been a variation of animals in their susceptibility. Moon et al. (1983) reported considerable animal to animal variation in the response of pigs to infection with E. coli enteropathogenic for humans; identical lesions were seen in the colons of pigs with or without diarrhoea.

Historically, the term enteropathogenic E. coli or EPEC was first applied to strains of E. coli implicated in epidemic infantile diarrhoea of humans. Some of these strains and others virulent for a number of species have been shown to cause highly characteristic microscopic lesions of the gut mucosa and to produce a cytotoxin detectable in vitro and have been called EPEC (Anon, 1983) . The microscopic lesions and production of cytotoxin may be unifying features of an emerging sub-group of enteropathogenic E. coli for which the term EPEC is not sufficiently descriptive. Takeuchi et al. (1978) suggested that these lesions in rabbits were induced by a cytotoxin which showed similarities to the shiga toxin produced by Shigella dysenteriae (O'Brien et al., 1982) . E. coli enteropathogenic for man have been shown to produce a toxin (Konowulchuk et al., 1978) which is biologically and serologically related to the shiga toxin (O'Brien et al., 1982) . The E. coli described here may form a distinctive sub-group in view of their isolation from cattle, their atypical characteristics and their distinctive properties in tests for fimbriae and haemagglutinins. These E. coli produced a mannose-resis-tant haemagglutinin in anaerobic but not aerobic conditions which was of particular interest; apparently the haemagglutinin was not associated with fimbriae. In aerobic conditions these bacteria produced fimbriae without detectable haemagglutinating activity which, although previously described (Isaacson, 1981) , are nonetheless unusual.

The unusual colony morphology produced on MacConkey agar might usefully serve as a primary selective characteristic for potentially pathogenic E. coli of the calf. However, strain 37/1 produced in vivo a variant with a colony morphology indistinguishable from that produced by most E. coli. Furthermore, if in vitro the common characteristic of the enteropathogenic E. coli is cytotoxigenicity then it is probable that E. coli which can cause enteric disease in the calf may be found with otherwise typical properties.

Mechanisms in enteropathogenic Escherichia coli diarrhoea

Effects of a calicivirus-like agent (The Newbury Agent) on gnotobiotic calves

Haemagglutinating and adhesive properties associated with K99 antigen of bovine strains of Escherichia coli

Dysentery in gnotobiotic calves caused by atypical Escherichia coli

Identification of Medical Bacteria

Evidence ofEscherichia coli with bacteraemic properties in mucoid enteritis of newborn calves

A simplified apparatus for the microbiological isolation of calves

Identification of enterotoxigenic Escherichia coli and serum antitoxic activity by the vascular permeability factor assay

Escherichia coli associated with neonatal diarrhoea in piglets and calves

Dysentery caused by Escherichia coli (S102-9) in calves: natural and experimental disease

Pili of enterotoxigenic Escherichia coli

Salmonella saint-paul infection in two dairy herds

Properties of an Escherichia coli cytotoxin

Handbook of Practical Bacteriology, Edinburgh

A turbidity test for the estimation of immune globulin levels in neonatal calf serum

Attaching and effacing activities of rabbit and human enteropathogenic Escherichia coli in pig and rabbit intestines

Production of Shigella dysenteriae type l-like cytotoxin by Escherichia coli

The establishment of K99, a thermolabile, transmissible Escherichia coli K antigen, previously called 'Kco', possessed by calf and lamb enteropathogenic strains

Localisation of enteropathogens in paraffin embedded tissue by immunoperoxidase

Evaluation of ELISA and electron microscopy for the detection of coronavirus and rotavirus in bovine faeces

A clinico-pathologic study of enterocyte-adherent Escherichia coli: a cause of protracted diarrhoea in infants

Observations by the ligated intestinal segment and oral inoculation methods on Escherichia coli infections in pigs, calves, lambs and rabbits

The unlabelled antibody enzyme method of immunohistochemistry : preparation and properties of soluble antigen-antibody complex (horseradish peroxidase-anti-horseradish peroxidase) and its use in the identification of spirochaetes

Scanning and transmission electron microscopic study of Escherichia coli OIS (RDEC-1) enteric infections in rabbits