18. Enterohaemorrhagic Escherichia coli (EHEC).

(a) Vero or Shiga and Shiga-like Toxins.

In the 1970's there were extensive studies, whose initial object was to determine whether Enteropathogenic E. coli (EPEC) produced some sort of a toxin, similar to the enterotoxins of the enterotoxigenic E. coli (ETEC). These studies demonstrated that some strains of E.coli, which belonged to serogroups associated with EPEC were able to produce a cytotoxin which was active on Vero cells. This toxin was found to be distinct from both the Heat labile (LT) and Heat stable (ST) enterotoxins associated with ETEC. Vero cells are tissue culture cells, which have been derived from the kidney of an African green monkey (Cercopithecus aethiops). This Verocell cytotoxin was given the abbreviation VT. It was thus shown that it was indeed produced by a number of EPEC but not by all. This seemed to raise the possibility that there may be two types of EPEC, those producing VT and those, which did not.

Some twenty years earlier a condition, known as haemolytic uraemic syndrome (HUS) had been described. Although it affects all age groups, it has been found more frequently in young children. It is considered as a major cause of renal failure in children. This condition was linked to infection with Shigella dysenteriae Type 1. This member of the Shigella genus, which had originally been considered a separate species under the name Shigella shiga had been shown to be unique among the genus in producing a toxin, which was named Shiga toxin. It was considered that the development of HUS following Sh. dysenteriae Type 1 infection may be associated with the production of this toxin. However, although Shiga toxin had been described as early as the beginning of this century, it was always considered as a little of an orphan toxin, because it did not to seem to play a role in the course of the dysentery.

Many studies on HUS have now also linked this condition with VT+ E. coli. In addition prodromal bloody diarrhoea, which occurs mainly in children particularly in the summer months both as outbreaks of haemorrhagic colitis and as sporadic cases, has been associated especially with strains of a serotype of E.coli O157.H7 which produce VT but do not produce LT or ST. This serotype, of which much more will be told later, had not been seen before 1982.

When it was realized, that infection with Shigella dysenteriae Type 1, producing Shiga toxin can also lead to HUS as well as VT+ E. coli, neutralization studies were instigated. These showed that there were two serologically distinct types of VT. One of the serological types of VT was found to be closely antigenically related to Shiga toxin, while the other was not. As a result some authors call this group of toxins Shiga-like toxins (SLT). More recently, as a result of the realization that the Shigella toxin and the related E. coli toxin are really members of a family of related related toxins, the E. coli toxins have been termed Shiga Toxins (ST) and the E. coli producing them as Shiga Toxigenic E. coli (STEC), rather than Shiga-like Toxigenic E. coli (SLTEC). Unfortunately, there can be confusion between ST, standing for Shiga Toxin or for the Heat Stable Enterotoxin. The term VT for Vero Toxin does not produce this confusion and it and the term Verotoxigenic E. coli (VTEC) will be used throughout the proceeding sections, unless otherwise specifically designated.

Bacteria belonging to the species of E. coli and the genus Shigella have always been seen to be very closely related. In many ways strains of Shigella, could be seen as non-lactose fermenting, non-gas producing, non-motile variants of E. coli. In addition the O antigens of virtually all strains of Shigella are either very closely related to or even identical with some of the O antigenis of E. coli. Thus, the O antigen of Sh. dysenteriae Type 1 is the same as the E. coli O1 antigen. Sh. dysenteriae Type 1 could quite easily be designated as a non-lactose fermenting, non-gas producing strain of E. coli O1:H-.

To return to the VT's, it is now considered that the VT which is neutralised by antibody to Shiga toxin has been termed VT 1 (SLT I; ST I). The other cytotoxin produced by many strains of VTEC either alone or with VT1, and which has the same biological effect on Vero cells was called VT2 (SLT II; ST II). Some strains produce only one of these cytotoxins, while others produce both. It has been demonstrated further that the production of these VT's is phage mediated. Specific DNA probes have been introduced for use in DNA hybridization studies and these have shown that under stringent conditions no cross hybridization could be detected between VT1 and VT2.

(b) Mode of Action of the Verotoxins.

The Verotoxins (VT) act on eukaryotic ribosomes. Their mode of action is similar to that of ricin, the toxin of the castor oil plant. The result of their action is irreversible inhibition of eukaryotic protein synthesis. It has been shown that at the molecular level the toxins act as enzymes, specifically on the N-glycosidic bond at A-4324 of 28S ribosomal RNA thereby inactivating the 60S ribosome. A single adenine residue is released resulting in the failure of the amino-tRNA binding to the eukaryotic ribosome as mediated by elongation factor 1. The active enzymic site of VT, Shiga toxin and ricin has been shown to be similar.

The VT's all have a similar structure, of a single enzymically active A subunit surrounded by five B subunits. In this general structure the VT's resemble the heat-labile enterotoxin (LT) or cholera toxin (CT), although their mode of action is quite different. The A subunit of VT's is about 33 kDa and the B subunits about 7.5 kDa each. Although the two toxins VT1 and VT2 differ substantially in their amino acid sequences in both subunits, hybrid toxins of an A subunit from VT1 and B subunits from VT2 or vice versa are fully active.

It is the B subunits, which attach to the surfaces of the eukaryotic cells, and on the specificity of their binding depends the first stage of the action of the toxins. Not all VT's have identical binding specificities. The original Shiga toxin of Sh. dysentery Type 1 as well as the E. coli VT1 and VT2 bind specifically to a eukaryotic cell surface structure known as globotriosyl ceramide (Gb3). Of this structure, the terminal disaccharide of two galactose residues linked as: gal-a(1Æ4)-gal-b(1Æ4)-glcCeramide is the minimum requirement for binding. For the porcine VT, VT2e, globotetraosyl ceramide (Gb4) is the binding site and its minimum binding structure has a terminal N-acetyl linked to the galactose as: gal-NAcb-(1Æ3)gal-a(1Æ4)-gal-b(1Æ4)-glcCeramide. There have also been suggestions that other binding sites for the toxins may exist on some eukaryotic cell surfaces.

The Gb3 and related structures are also present on erythocytes. Differences in the amounts of Gb3 have been associated with differences among the blood group P types. It has also been shown that cytokines such tumour necrosis factor-a (TNFa) and interleukin-1 will increase the number of Gb3 sites on the glomerular endothelial cells of the kidneys and it is noteworthy that the VT's as well as endotoxin will cause the production of these factors to be induced in cells. It has been particularly shown that TNFa synthesis is induced in the kidney by VT and the increase in Gb3 sites is achieved by the TNFa inducing galactosyl transferase activity. Thus, the VT's induce cells to make factors, which increase the numbers of binding sites for VT's on neighbouring cells. It has also been shown that Gb3 is more common on paediatric glomerular cells, explaining the higher incidence of HUS in children.

The internalisation of the toxin is probably through the association of the glycolipid in the membrane with some other transmembrane proteins. They may then be intracellularly cross-linked into clathrin-coated pits through the action of cytosolic transglutaminase. The proteins with which Gb3 may be associated could be the human a interferon receptor or CD19. It is still uncertain, to what extent both the A and B subunits enter the target cells. Cleavage of the A chain into two disulphide bonded fragments A1 and A2, appears to be an essential prerequisite for the induction of toxicity. It was shown that by using Brefeldin A, a drug which inhibits Golgi functions, it also prevents the effects of VT on the protein synthetic activity of target HeLa cells. This confirms earlier suggestions that both the Golgi apparatus and the endoplasmic reticulum are involved in the action of the toxins.

A further action of the VT molecules has been shown to be their ability to induce programmed cell death or apoptosis in susceptible cells. This was demonstrated in Burkitt's Lymphona cells, which carry the Gb3 receptor on their surface. The apoptosis is induced not by the enzymically active A subunit but by the B subunits. Further it appears that the VT may be able to cause neurological disorders particularly in conjunction with lipopolysaccharide (LPS). The hypothesis has been suggested that the induced apoptosis by VT is a protective response of the organism to the infection. By programming the death of the infected cell, this cell cannot induce the increase of the pathological effects, such as perhaps the induction of TNFa, which in turn induce further VT receptors and cause more cells to be destroyed. Thus, this programmed cellular self sacrifice should be seen a means of the host protecting itself.

VT2 also inhibits 60S ribosomal subunits of rabbit reticulocytes, while not inhibiting those of wheat germ or E.coli. These studies confirm that the inhibition was at the level of elongation 1-dependent aminoacyl-tRNA-binding to ribosomes. It did not affect Met-tRNA binding to ribosomes, non-enzymic binding of aminoacyl-tRNA to ribosomes, peptide bond formation or translocation.

It has been suggested that to cause the symptoms characteristic of EHEC these toxins must adhere to epithelial cells and possibly cause direct damage to the vascular endothelium. Dividing cells are particularly sensitive to VT1. The incorporation of [3H]leucine into protein was more severely reduced in non-confluent cells than in confluent ones. In addition the cerebral effects which have been associated for some time with Shiga-toxin and are also believed to be caused by E. coli VT's, may be due to vascular damage and small infarcts rather than by a direct effect on the neurones.

(c) Adherence Factors of the EHEC.

It has previously been noted that there seem to be some similarities between the Enteropathogenic and Enterohaemorrhagic E. coli (EPEC and EHEC). It was first thought that the VT's were produced by EPEC and in fact the realization is now appearing that some outbreaks in the past, which were ascribed to EPEC may actually have been due to EHEC. This is particularly so for outbreaks due to strains of serotype O26:H11 and O111:H-, which can be both EPEC and EHEC.

It also seems that many EPEC and EHEC share a common attaching and effacing (A/E) mechanism in order to colonise the intestine. A model has been proposed for the adherence of EPEC, which may also applicable to EHEC. According to this model, there are three stages. Firstly the bacteria loosely adhere to the intestinal surface epithelial cells by means of fimbrial adhesins. Secondly, as a result of signal transduction events caused by the E. coli eaeB gene product and the cfm locus, specific host cell proteins are phosphorylated. The eae and cfm loci are quite distinct regions mapping on different parts of the E. coli chromosome. This phosphorylation mainly affects tyrosine residues and the major substrate is a 90-kDa protein.

However, this tyrosine phosphorylation does not appear to be the mechanism for EHEC. Using infection with EHEC of cultured epithelial cells, activation of the phosphatidylinositol cascade has been demonstrated. Similar to EPEC infection of epithelial cells, with EHEC infection there is an increase in second messenger molecules, cytosolic free calcium and inositol 1,4,5-triphosphate (IP3). As the buffering of the intracellular calcium inhibits these A/E lesions, a strong relationship between them should be considered established. The activated IP3 mobilizes the intracellular calcium stores elevating the internal calcium ion concentration [Ca++]i. This triggers the breakdown of the host microvilli.

Thus in EHEC the effacement of the A/E lesion is an independent A/E lesion. It has also been found that strains lacking the eaeA gene are still able to ellicit the increased [Ca++]i response, suggesting that this gene's products are not involved in this response. If the situation with EHEC is similar to that with EPEC, it is probable that the eaeB gene products are responsible for the soluble factors which mediate the [Ca++]i response. Despite EHEC having the eaeB and/or cfm genes the tyrosine phosphorylation does not appear to be a factor in EHEC infections. It is possible that this lack probably also explains why EHEC are not internalised in cell lines such as HEp-2 cells, unlike EPEC which are. On the basis of these results it has thus been suggested that EHEC are not invasive due to this lack.

Further unlike EPEC, it appears that EHEC induce a-actinin aggregation at the sites of bacterial attachment as indicated by a positive fluorescent actin staining (FAS) response. The phosphorylation may be at serine or threonine residues, which are also phosphorylated in EPEC infection of HEp-2 cells. Therefore, while many EHEC have the eae system, it seems to operate in a different way. It should also be noted that there have been isolated strains of EHEC, which produce VT's, from clinical cases of HUS, bloody diarrhoea and/or diarrhoea, which do not have the eae system. It is believed that these EHEC must have other adherence mechanisms to gain entry into the enterocytes, however, this has not yet been found. These EHEC include strains belonging especially to serotypes O5:H- and O113:H21.

(d) The Haemolysin of EHEC.

A number of types of haemolysin have been described as being produced by E. coli. The best known and most studied of these is a-haemolysin (aH). This has been known for many years and extensively studied. It is often associated with strains causing urinary tract infections. More recently there was described another haemolysin, which has been named Enterohaemolysin (EHly). The genes coding for EHly appear to reside on the around 60 MDa plasmid of most VTEC, but not all.

There are two genetic elements, which share a strong homology of about 60% with two genes of the aH operon, suggesting an evolutionary link between the two haemolysins. The actual haemolytic effect on erythrocytes is weaker than that of aH and it is suggested that the target cells for EHly are different to those for aH. An earlier study on aH has shown that it can cause the release of interleukin-1b from some cultured cells. It has been pointed out in relation to the modes of action of the VT's that the action of cytokines like interleukin can stimulate cells to increase the number of VT receptors on their surfaces, thus this may be the role of EHly. It should never be forgotten that it may still play a role in releasing haemolglobin and with it iron from red blood cells and this could also be important in giving EHEC a selective advantage over other E. coli.

(e) Ecology of EHEC.

The general view is that the problem with EHEC began with the description of two separate outbreaks in U.S.A. in 1982, which were at the time ascribed to a rare serotype of E. coli namely O157:H7. EHEC have been described several years before, as being associated with mainly intestinal diseases from the time that it was realized that there were E. coli which could make these VT's in the later 1970's. However, it should also be clear that these organisms would have been around for many years before that. Unfortunately those EHEC serotypes, which were similar to the EPEC serotypes would have been classed as such. Many others would have been missed and perhaps a few might have been misclassified. It should be noted that in the 1950's less than half the cases of intestinal infections could be ascribed to a specific pathogen.

An especially interesting epidemic occurred in the early 1950's in Virginia, U.S.A. associated with the U.S. Army hospital at Fort Belvoir. It is believed that two mothers admitted to the obstetrical service may have brought the infection into the area. It was noted that the disease affected children much more severely and that strains of E. coli belonging to serogroup O111 were isolated from most cases. This serogroup had been described about ten years ago in Scotland and was the first to be considered as the pathogens we now know as EPEC. It seemed eminently reasonable that the investigators of this U.S. outbreak should believe that this was an EPEC outbreak. There were a number of fatalities and the clinical pictures as well as post mortem examinations did make the investigators believe they were dealing with a particularly virulent form of EPEC. Clinicians reviewing these observations in the 1990's have concluded that this outbreak was most likely to have been caused by EHEC belonging to the O111 serogroup of E. coli. Unfortunately there has been no microbiological confirmation of the organisms.

In a study on a few strains belonging to EPEC serogroups isolated from cases of diarrhoea in that period, most conformed to the current deffinition of EPEC. However, as HUS had been described as long ago as 1955, albeit mainly in conjunction with Shigella dysenteriae type 1 infections, it seems reasonable that these EHEC would also have been present. As soon as the Vero cell assay was available in the latter 1980's there were reports from a few laboratories as far apart as England and New Zealand that indeed VT+ could be isolated from cases of diarrhoea. These were still mainly members of the known EPEC serogroups such as O26.

Into this scene of the occasional report of cases of infection due to strains of VT-producing E. coli, there came a report in 1983 of two apparently unrelated outbreaks the previous year in the United States due to VT-producing E. coli belonging to what then was described as a "rare" serotype, namely O157:H7. The source of both outbreaks was hamburger meat from a well-known restaurant chain. Since that time there have been many outbreaks reported around the world in which hamburgers, especially underdone are associated with infection due to this VT-producing E. coli serotype.

Until 1982 this serotype had not been reported from anywhere in the world. However, a retrospective study re-examined strains of E. coli beloging to the O157 group which had been submitted to the International Escherichia and Klebsiella Centre since the establishment of this O group. There were found only three strains, which also had the H antigen H7. These had been part of a batch of 39 strains isolated from calves with coli bacillosis in Argentina. All of these three strains had been isolated from the faeces of one of three animals from one farm. They were the only VT-producing E. coli strains among the 39. Being also biochemically similar to the O157.H7 strains, which had been isolated from a number of the outbreaks, it might be suggest that cattle may be the reservoir of these organisms. It also possible that they have originated in South America.

Thus, with most reported outbreaks being mainly associated with the consumption of hamburgers, particularly if undercooked, or raw milk, the infection reservoire was identified as cattle. This view was changed with the report of an outbreak of haemorrhagic colitis due to E.coli O157.H7 described from the United Kingdom. This outbreak was unusual because it appeared to show a particularly strong association with preparing raw vegetables, especially potatoes. A British survey of VT-producing E.coli in faecal specimens of children with HUS revealed faecal carriage of EHEC by 19 of 66 patients studied. While 15 of these, belonged to serogroup O157 and all but one of these were motile and had the H7 antigen many other EHEC serotypes were found including O26.H11, O104.H2, O153.H25 and O163.H19. The strains either produced VT 1, VT 2 or both. This re-established that serotypes other than O157.H7 can be EHEC. A study on gnotobiotic pigs used some of these strains from cases of haemorrhagic colitis and showed that they produced the same characteristic lesions as did typical strains of O157.H7.

An outbreak in a nursing home in Ontario, Canada not only demostrated the severity of outbreaks due to EHEC O157.H7, but also the way this strain especially can spread. It affected 55 out of 169 residents and 18 out of 137 staff members. It was believed to have probably started by contaminated turkey sandwiches, which had been prepared on a marred wooden surface and then not kept refrigerated till they were served 2-3 hours later. Most noteworthy was that few days after the first set of cases, there developed a second wave of cases. This was presumed to be due to person to person spread, which has also been reported elsewhere. The severity of the outbreak can be guaged by noting that of the affected residents, HUS occurred in 12 (22%), and of the 19 (35%) residents who died it was considered that this specific infection with EHEC O157.H7 contributed in 17 (31%) of cases.

It was noted that VT-producing strains of E.coli O157.H7 wherever they may be found were unusual in not fermenting the carbohydrate, sorbitol. This led to the development of a medium, on which strains and especially E. coli, which did not ferment sorbitol could be found. Thus there became no reason why a laboratory should not be able to isolate these VT-producing strains of E.coli O157.H7. However, there remain the problems of identifying other EHEC but also VT-producing strains of E.coli O157.H7 which have recently been found in Europe, which ferment sorbitol. This has meant that unfortunately many still think the only EHEC that matter are VT-producing strains of E.coli O157.H7.

In one large survey in the 1980's in U.S.A. it was found that VT-producing strains of E.coli O157.H7 comprised 0.4% of the enteric pathogens. The same survey yielded in 3% Campylobacter, in 1.7%) Salmonella, and in 0.6% Shigella. From these total isolation figures it was estimated that the population incidence rates per 100,000 person-years, were 8 for E. coli O157.H7. Of the 24 patients affected with E. coli O157.H7, 14 had to be hospitalised for a period of 2 - 8 days. None demonstrated any complications but both ulceration and bleeding were noted in some patients. The majority of the 24 patients had their onset in the summer months. Only two of these appeared epidemiologically linked. Consumption of rare ground beef or raw milk were risk factors for acquiring E. coli O157.H7. This study showed that cases of bloody diarrhoea due to E. coli O157.H7 are more common than had been anticipated, being at the same level of frequency as those due to Shigella.

With the realization that HUS is an important cause of acute renal failure in children, an investigation was carried out into its frequency in one defined area in the United States. Cases were defined both clinically and on the basis of characteristic pathology over the periods 1971-5, 1976-80 and 1981-6. During these periods there were 59 cases that fitted the fully detailed definition with 11 in the first period, 22 in the second and 26 in the third.period. These figures on the basis of population statistics show that the three periods demonstrated an incidence rate per 100,000 children of 0.69, 1.77 and 1.74 cases respectively. This was considered to be a true increase in incidence which was being observed, rather than purely an increase in diagnostic awareness. The possibility of the increasing isolations of E.coli O157.H7 from cases of diarrhoea in this area, where this serotype has become the second most common cause of diarrhoea after Campylobacter may be linked to this increase in HUS as strains of E.coli O157.H7 are also associated with HUS.

Similarly a two year prospective study in the Calgary area of Canada showed that EHEC are the most frequent causes of bacterial diarrhoea, being particularly prevalent in the summer months. Most of the EHEC isolated were VT-producing strains of E.coli O157.H7, but a number of other serotypes were also identified including the EPEC serotypes O26.K60.H11; O111.K58.H- as well as O5.H-; O6.H-; O38.H21; O91.H-; O103.H2; O145.H- and O?.H21. The illnesses due to E. coli O157.H7 included in the majority of cases the typical symptoms of haemorrhagic colitis. In the patients with non-O157 EHEC infections the diarrhoea was generally more prolonged but bloody diarrhoea was less frequent. Two patients with non-O157 EHEC developed HUS.

In an investigation on the role of EHEC as possible causes of diarrhoea among children in Thailand stool specimens from children under five years old with visibly bloody diarrhoea and also from age-matched controls without symptoms of diarrhoea were examined. EHEC were present in 7% of with bloody diarrhoea and in 6% of age matched controls. Thus, as VT-producing E.coli appear to be present as fequently in children with diarrhoea as in well children, other virulence factors may have to be invoked before the VT-producing E.coli can be considered as causes of diarrhoea among children in Thailand.

During the period of June 1985-July1986 a total of 896 fresh meats and poultry products, purchased in Madison (Wisc., USA) grocery stores were analysed for the presence of E. coli O157.H7. Isolation rates of 3.7% from beef, 1.5% from pork, 1.5% from poultry and 2.0% from lamb samples were observed in these investigations. The types of meat from which E. coli O157.H7 were isolated included ground beef; pork chop, loin and hock; ground pork; chicken leg; turkey drumstick; lamb riblet, loin chop, shoulder blade chop and leg. In addition, E. coli O157.H7 was isolated from food specimens which have been associated with outbreaks of haemorrhagic colitis and HUS including unpasteurized milk, frozen chicken nuggets and ground beef from a restaurant.

As extensive earlier studies have shown that in general E. coli found on meats is derived from the faeces and intestinal contents of the animals and survive the slaughtering and preparation processes in the abattoirs the probable source of these organisms is in the animal population. A study on the distribution of serotypes of E. coli in cow pats in various parts of England has shown that there is a correlation between them and those found in the human population.

By the early 1990's all these above facts were known, VT-producing strains of E.coli O157.H7 were being isolated from patients and foods, mainly meat derived foods in increasing numbers, there were being reported cases and outbreaks linked to the consumption of meat, but also other foods including dairy products, but also vegetables and even apple cider. Even paddling pools were found to be possible vehicles of infection.

Then there was an outbreak in a number of the Western States of the U.S.A. in 1992-1993. This has since been termed the multistate outbreak. A total of at least 583 persons were affected across four states. The majority of cases (477) were in Washington state. A strong connection was established with a chain of hamburger restaurants. The severity of the outbreak can be seen by noting that 171 patients required hospitalisation, 41 developed HUS and four children died. VT-producing strains of E.coli O157.H7, which were positive for the adherence factors associated with EHEC and characteristically did not ferment sorbitol were isolated from many patients and from frozen hamburger patties belonging to batches which had been incriminated epidemiologically with the outbreaks. These were shown to have had fewer than 100 E. coli O157 per frozen pattie, suggesting that the cooked patties would have very few VT-producing strains of E.coli O157.H7 per pattie. This major outbreak thus also highlighted the problem that there were very few organisms required to cause infection.

Since then there have been many more outbreaks due to VT-producing strains of E.coli O157.H7 with a particularly severe outbreak in Scotland affecting mainly elderly people with many deaths an one in Japan where around 10,000 mainly children were affected. In this case 11 deaths were ascribed to the outbreak. In The Scottish outbreak meat was again the main vehicle of infection, while in Japan a number of different sources were suggested but raddish sprouts were probably of greatest significance.

The other EHEC serotypes were also still present. Throughout the 1980's there were occasional reports of EHEC isolations in Australia and it appeared that VT-producing strains of E.coli O111 were particularly prominent. A report in 1994 from Italy described a small EHEC outbreak in 1992, which is the first reported Italian outbreak. There were a significant number of sporadic cases due to these E. coli O111.H- as far back as 1988. Cases due to O111 have also occurred throughout many parts of Europe particularly Italy, France and Germany. In Australia a great variety of serotypes have been associated with cases of HUS including especially strains belonging to Ogroup O111, which were particularly prominent. The first Australian outbreak of HUS was found to be associated mainly with VT-producing strains of E.coli O111 but VT-producing strains of E.coli O157 and other serotypes were also found. The VT-producing strains of E.coli O111 associated with the outbreak were identified by modern molecular biological techniques, which are not generally available. If the laboratory had purely relied on the cultural techniques in common use, it would have picked out the few VT-producing strains of E.coli O157 and ascribed the outbreak to this pathogen. Although VT-producing strains of E.coli O157 were only isolated from about 18% of patients this isolation rate is of the same order of magnitude as that reported from many outbreak attributed exclusively to VT-producing strains of E.coli O157. The question has to be asked, whether these may not also have been due to other EHEC serotypes.

In surveys conducted around the world many EHEC serotypes in addition to those belonging to O groups O26, O111 and O157 have been isolated from healthy animals and not associated with disease in humans or animals. The list of serotypes associated with human infection including HUS is steadily increasing.

It is intriguing that EHEC should be so widespread, but also that VT-producing strains of E.coli O157, which probably emerged only in the late 1970's should have spread so effectively around the globe, being now found on every continent. Ruminants appear particularly to harbour VTEC in their intestinal tract. More importantly it appears that under reduced feed availability the carriage rate in cattle of VTEC can be tripled. These cattle, sheep and other ruminants are apparently healthy despite carrying these VTEC. The role these VTEC play in these animals has to be explained before an understanding of the ecology of EHEC can be achieved.

It appears that they occur predominantly in the rumen, which is a complex ecosystem, in which only modern molecular biological techniques may begin to characterise the many types of organisms seen in ruminal fluid. Many of them, most of which are strict anaerobes have not been cultured. Such a diversity of ruminal bacteria is required by the complexity of feed components likely to encountered. For the breakdown of the different substances, bacteria with different specialities may be required. A stable environment requires a diversity of bacteria to be present. It cannot be denied that the VTEC are successful in this ecological niche. It may be the production of the toxin(s) which gives these organisms a selective advantage.

Studies carried out since 1983 have shown, that phages of type l carry the gene sequences for the verocytotoxins. These l phages, most of which do not carry the VT gene sequences are quite widespread and can lysogenise the E. coli. It is just such lysogenised E. coli carrying the l derived DNA as a plasmid that are the VTEC associated with human disease, and which have apparently a selective advantage in the rumen. Organisms, which carry a plasmid whose function is not required, tend to lose the plasmid. Recently it was shown that under conditions of amino acid limitation these l plasmids replicate more readily in their host E. coli. This depends on the types of amino acids limited.

At times of high feed, there may be a small number of VTEC in the rumen. These can multiply as fast as any other bacteria due to excess nutrients available and are not under any stress. When feed to the animal is reduced then two things may happen. Those E. coli either carrying a l plasmid or being able to acquire one, the l plasmid will multiply. If it carries the verocytotoxin genes, the toxins will be synthesised. These probably kill the ruminal eukaryotes, providing the VTEC with nutrients and giving them their selective advantage over other bacteria in the rumen. The advantage cannot be too great or the destructuction of the ruminal biodiversity would destroy the environment of these VTEC. Thus the emergence of the EHEC has to be seen in a complex global context of complex ecology. Their spread may be linked to poor sewage disposal, extensive movement of people, bird migration and their ability to establish a niche in an important group of food animals.

(f) EHEC as Animal Pathogens.

(I) Pigs

Oedema disease of pigs has been known for many years. It has been reported in many countries around the world. It predomonantly involves recently weaned pigs about 7-10 days after weaning. It was originally described as long ago as 1950, when the decrease in susceptibility with age was commented upon. However, it does occur in older animals and has been seen in both brood sows and boars. The disease is marked by a rapid onset, and abrupt termination of the outbreak. Deaths usually occur 4-5 days after onset with case mortality rates of up to about 60% being reported. Sudden unexpected death without signs of previous illness are a feature of outbreaks of oedema disease. Loss of control of limb movements leading to paralysis are features of the disease. Swelling of eyelids, reddening of skin, drooping of ears, laryngeal oedema are often encountered. While constipation is the general rule, diarrhoea may also occur. This is dependent on the types of toxins produced by the causative E. coli. Terminal diarrhoea is a characteristic in some cases. The disease may also become chronic in some animals.

In the early studies in the 1950's, it was noted that the intravenous injection of pigs with centrifuged extract of the gut contents of affected pigs, mimmicked the disease. The toxin(s) involved were heat labile and antibody could be raised to them. Various names were given to these toxic factors including oedema disease toxin, oedema disease principle, neurotoxin, angiotoxin and vasotoxin. However, it was the observation that an isolate from oedema disease produces a factor that affecting Vero cells, that indicated that a Verotoxin (VT) may be involved.

The VT's from the porcine isoaltes of E. coli were found to be different from those isolated from human VTEC. These were shown to be closest to VT2 (or SLT II, ST II) and were initially desgnated VT2v (SLT IIv) as variants of VT2. The name has been changed to VT2e (SLT IIe, ST IIe) to signify the relation to oedema disease.

The different types of VT2e isolated from many different serotypes of E. coli from cases of oedema disease all seem to belong to one antigenic type. Some E. coli isolated from cases of oedema disease also produce one or more of the enterotoxins STIa, STII and LTp. It is the presence of these enterotoxins on the infecting E. coli, which will determine, whether the animals also get diarrhoea. Another feature of the porcine VTEC is that most of the strains isolated from outbreaks of oedema disease produce a-haemolysin. This is in contrast with the human VTEC isolates which predominantly produce Enterohaemolysin.

A number of serotypes of E. coli were shown to predominate among these porcine VTEC. They included O138:K81:H14; O139:K82(K12):H1 and O141:K85:H4. As with human VTEC, a number of other serotypes have also been occasionally reported. It was further shown that these VTEC serotypes produced a new type of fimbriae, which have been desgnated F107. The F107 fimbriae are long, slender and flexible structures, with a diameter of about 4.6 nm. They are not expressed at 18°C. No antigenic relationship with the fimbriae of human ETEC could be observed. They also do not agglutinate human or animal erythrocytes. Some serological variants of F107 have been observed with some strains of porcine VTEC.

Colonization of the intestine of the pigs by the VTEC ia an essential prerequisit for the disease. This colonization can be affected by significant changes in the intestinal ecology, such as happens at weaning or if there is a change in the animals' diet. Not all pigs are affected in the same way, even with outbreaks occurring with the same strain. There are probably some porcine genetic factors involved. The development of diarrhoea by the animals depended on whether the organisms also produced some enterotoxins. The VT2e, which is exclusively produced by the porcine E. coli is absorbed into the circulation as part of the disease process and can be neutralised by antibody.

(II) Cattle

As was discussed with human cases of VTEC, it appears that cattle as well as other ruminants, frequently carry these VTEC in their rumen, while remaining completely healthy. However, VTEC have also been shown to be able cause diseases in these animals. A dysentery-like syndrome was induced in gnotobiotic calves, which had been orally administered with a strain of VTEC O5:H-, which had been associated with an outbreak of calf dysentery. Other reports include an diarrhoea outbreak in calves due to an unusually urease producing strain of EHEC O5:H4. In an outbreak of dysentery-like symptoms among calves in Canada, an EHEC O111:H- was isolated from all affected animals. However, this strain when fed to a colostrum deprived calf, failed to cause diarrhoea. There have been other reports of similar cases or outbreaks generally involving calves around two weeks of age.

These reports suggest that under certain conditions EHEC can certainly cause disease in calves, however, these must be relatively rare, because carriage of VTEC by calves must be quite widespread. Also a recent study showed that nearly 90% of adult cattle had antibodies to VT 1, indicating extensive exposure. Only about 5% had antibody to VT 2. It is also noteworthy that EHEC serotype O157:H7 appears to be not pathogenic for cattle, although it has been isolated from healthy animals around the world, albeit with a low prevalence.

Unlike human isolates of EHEC, those from severely ill calves have been shown to be frequently multiply resistant to antibiotics. The reliance on the use of antibiotics rather then good agricultural practices may be ascribed to this. It may also be postulated, that the use of antibiotics may give these antibiotic resistant EHEC a selective advantage under certain circumstance, which is why they only occasionally cause bovine disease, despite being widespread amongst these animals.

(III) Sheep

There are no reports of EHEC causing disease in sheep or lambs. They appear to quite widespread among these animals. Reports of carriage rates by healthy sheep of up to 40% at any time have been noted.

(IV) Goats

There have not been many studies on the role of EHEC in diseases of goats. Strains of EHEC producing VT 1 and belonging to serotypes O157:H- and O103:H2 have been associated with diarrhoea in 1-2 moth old kids. While they did not appear to possess any other known virulence factors, the O103:H2 strains did appear able to cause typical attaching and effacing lesions.

(V) Horses

There are no reports of EHEC causing disease in horses. No studies or reports of carriage rates by horses or fowls have been noted.

(VI) Poultry

There are no reports of EHEC causing disease in poultry. No studies or reports of carriage rates by adult birds or chicks have been noted.

(VII) Dogs

There are no reports of EHEC causing disease in dogs. No studies or reports of carriage rates by dogs or pups have been noted.

(VIII) Cats

There have not been many studies on the role of EHEC in diseases of cats, but one Swedish study revealed that 3/4 E. coli isolated from cats with diarrhoea, produced VT 1, while only 1/8 of isolates from healthy animals did so. These VTEC belonged to serogroups O2, O4 and O6 were generally haemolytic.

(IX) Rabbits

There are no reports of EHEC causing disease in rabbits. No studies or reports of carriage rates by rabbits or pups have been noted.

41240 words

© Copyright Microbionet.

All literary matter in Microbionet is covered by copyright, and must not be reproduced, stored in a retrieval system, or transmitted in any form by electronic or mechanical means, photocopying, or recording, without written permission. This page is to be read in conjunction with the Disclaimer.

Use of this site signifes your agreement to the Legal Notices. Legal Notices © Copyright Microbionet. All rights reserved.