Enterotoxigenic escherichia coli of animal origin as a possible source of infection for man
Acute diarrhoeal diseases account for the highest infant and childhood morbidity and mortality in tropical developing countries. In recent years, attention has been focused on enterotoxigenic Escherichia coli (ETEC) which are among the most commonly encountered enteropathogens. The occurrence of ETEC in both animals and man has posed the questions whether animals could serve .as sources of infection for man and to what extent transmission between different hosts takes place. This study deals with these issues. It also aims at determining the prevalance of ETEC in various animal populations, either as a cause of disease, or as normal inhabitants of the intestines of healthy animals. Some samples of human origin have been examined for comparison. One prerequisite for the causation of disease is the production of enterotoxins, inherent in the term "ETEC" itself. Another is the possession of fimbriae which enable the organism· to adhere to tissuess ETEC elaborate a heat-labile enterotoxin (LT) and at least two heat-stable enterotoxins (ST) responsible for intestinal fluid hypersecretion. Since the discovery of the functional, structural and immunological similarities between LT and the enterotoxin of Vibrio cholerae, assays which were based on these concepts were developed concurrently~ A major problem in the diagnosis of ETEC disease is the inherent difficulty in identifytng enterotoxigenic strains among the many non-toxigenic E. coli present in a faecal sample as a part of the normal intestinal flora. There is also no consensus of opinion on how many colonies should be tested for toxin production in order to obtain reliable results. The bioassays, tissue culture techniques and immunoassays which are available for the detection of enterotoxins are time consuming, expensive and require skilled technical assistanceo The application of the staphylococcal coagglutination technique (CAG) which allows the examination of a large number of isolated colonies of E. coli from a primary culture of a stool specimen would . _present a major advantage. An evaluation of this test to determine its suitability for routine laboratory detection of LT was one of the objectives of the present study. The sensitivity and specificity of the coagglutination test were compared with those of the enzyme immunoassay (ErA) for the detection of LT. E. coli isolates were also tested for the heat-stable enterotoxin, STA, by means of the suckling mouse assay. A total of 10,709 colonies of E. coli from man and animals w~re examined for LT by means of the ErA and the CAG test. The CAG test had a sensitivity of 93% and a specificity of 100% compared with the EIA, This finding is in conformity with the sensitivity and specificity reported for the previously described tube coagglutination test for the detection of LT. An average of 20 colonies per stool sample were tested for LT, and in most cases about 50% of these were positive. In a few cases, only 1 to 3 colonies were posiLive per culture. The CAG test can be performed rapidly and easily for the screening of a large nwnber of colonies from a primary culture, a feature which may more than compensate for its lower sensitivity compared with EIA. It was also found necessary to effect the release of LT from cultures of E. coli to achieve uniform, reliable results in the CAG test. Triton X-l00 at a dilution of 0.05% was found to serve as an adequate substitute for the more expensive antibiotic polymyxin B. Enterotoxigenic ,&.. coli, defined as LT positive, ST positive or LT/ST positive, were isolated from 41% (135/329) of children with diarrhoea from Kenyatta National Hospital, and 53% (9/17) of children with diarrhoea from Muhimbili Medical Centre, Dar es Salaam, Tanzania. ETEC was also isolated from 19% (la/54) of healthy pigs, 33% (18/55) of diarrhoeic pigs 22% (12/55) of healthy cattle, 7% (8/110) of healthy sheep and 1.4% (1/73) of healthy goats. A calf with diarrhoea also harboured ETEC. Most pigs with diarrhoeal disease were less than 12 weeks old. The ages of other domestic animals were unavailable. ETEC was most prevalent in children within the 0-2 year age group. This is in agreement with findings in Ethiopian children. ST-producing strains of E. coli were more frequent than LT positive E. coli in both children and animals. This finding is in conformity with results of studies on diarrhoeic patients in Bangladesh~ but in contrast to results in Ethiopian children where LT producers were more frequent. Other studies report equal numbers of LT positive and ST positive E. coli isolates. All ETEC isolates from sheep and goats~ which comprised 18% of the ETEC of animal origin, were positive for LT only. E. coli which produce both LT and ST are known to cause more severe diarrhoea than producers of LT only or ST only. The strains which are positive for both enterotoxins comprised 3% of ETEC from children with diarrhoea at Kenyatta National Hospital and ' 6% of children with diarrhoea from Muhimbili Medical Centre. These figures are similar to those reported for villagers from the South Nyanza District of Kenya. LT/ST positive E. coli strains comprised 4% of ETEC from healthy pigs, 16% of diarrhoeic pigs and 5% of healthy cattle. An LT/ST positive strain was also isolated from the calf with diarrhoea. Adhesion fimbriae of ETEC exhibit considerable host specificity in their attachment to the intestinal mucosal cells of man and animals. Nevertheless, a number of ETEC strains which do not possess the characterised host-specific fimbriae may still cause disease through as yet unknown attachment factors~ , F1, F2 and F3 fimbriae were detected orily in E. coli isolated from humans, while F4, F5 and F6 were identified only in isolates from domestic animals. F1 was present in ETEC as well as in non- ETEC isolates from children, while F2 and r3 were confined to only ETEC. Strains possessing F2 and F3 comprised 19% and 10% of al.L ETEC isolates fromchildren. F4 fimbriae were identified in 29%, F5 in 18% and F6 in 18% of ETEC isolates from pigs. The remaining isolates (36%) were found to be nonfimbriate. The only fimbrial type identified in ETEC isolates from ruminants was F5, which was present in 77% of ETEC from cattle and 38% of ETEC from sheep. The only isolate of ETEC from a goat was also positive for this fimbrial type. A number of ETEC strains from both children and animals were negative for fimbrial antigens. Due to a lack of sera,these were not serotyped, and the question of serologically identical strains of ETEC being present in both man and animals could therefore not be answered. It is also unclear whether non-fimbriate ETEC may acguire the appropriate fimbriae through plasmid transfer, which would confer the ability to adhere to the intestinal epithelium of a different species. Strains of E. coli which colonize an individual at infancy and become established as a part of the normal microflora are non-toxigenic. It is however conceivable that these non-toxigenic ~. coli, which ,may possess appropriate attachment factors, may acquire the ability to produce enterotoxins through plasmid transfer occurring in vivo following the ingestion of ETEC of animal origin. The extent to which this may occur will depend on whether the ingested ~. coli of animal origin are capable of transferring these plasmids, as well as the capability of non-toxigenic human strains for acquiring the plasmids which encode for enterotoxin production. ~fuile plasmids which encode for antibiotic resistance are known to be readily transferable within, as well as between species and genera of enteric bacteria, it is largely unknown to what extent transfer of plasmids that encode for the production of enterotoxins and host-specific fimbriae may occur in nature. Most ETEC strains are restricted to a few 0 serogroups. Therefore, the frequency of transfer tifplasmids encoding for fimbriae and enterotoxin production may to some extent depend on the presence~ of E. coli of the appropriate serogroups wh lch can acquire these plasmids. This study has confirmed that ETEC strains of human origin possess f'1 (somatic type 1) f'2 (Cf'A/1) and f'3 (Cf'A/II) fimbriae which are specific for man~ ETEC which possess f'4 (K88) and [6 (987P) were limited to pigs only, while ETEC strains with F5 (K99) fimbriae o.1erecommon to pigs, sheep, goats and cattle. Accordingly, the pattern which has emerged from this study indicates that ETEC strains exhibit strict host specificitY,the only exception being FS-fimbriated ETEC which were found in pigs, sheep, goats and cattle. It can therefore be concluded that cross-infections are unlikely to occur between man and animals. Nevertheless, the possibility of interchange of infective plasmids between strains of animal and human origin resulting in interspecies spread of ETEC disease cannot be ignored.