17. Enterotoxigenic Escherichia coli (ETEC).

(a) Travellers' Diarrhoea.

Travellers' diarrhoea has been known for many years to result when people from developed countries or temperate climates travel to developing or tropical climates. Whether it was the change in climate or move from a developed to developing situation was not clear. This condition has frequently gone under a number of colourful names such as Montezuma's Revenge, Delhi Belly, Bombay Two-Step or Gyppy Tummy to mention just a few.

It had often been speculated, that it might purely be the result of the change to an unexpected climate. However, if this were so then it might be expected also to occur, when there are sudden hot spells, as do happen in many temperate climates in the summer. While there may be a slight upsurge of intestinal infections in summer, it is certainly not at the level that occurs, when a whole tour party can suddenly go down with severe diarrhoea a few days after their holiday started.

With increasing jet travel, the effects on the circadian rhythm and also just the physiological stress of the travel situation have been considered likely candidates as causes of travellers' diarrhoea. If this were so then it should also occur when people travel from one part of the developed, temperate world to another, crossing a number of time zones. However, travellers' diarrhoea is much more likely to occur when travelling from a temperate to a tropical region, without crossing a time zone, than when moving from one temperate region to another, crossing many time zones.

Other suggestions as the cause of travellers' diarrhoea have been considered to be the irritation of the intestinal tract by such diverse agents as sand from the beaches; dust, which may contain unusual allergens, or unusual foods often containing unexpected spices. Also suggested was that the cold drinks, taken, when the outside weather was very hot, might 'chill' the abdomen causing irritation leading to diarrhoea. Perhaps, most far-fetched, was the theory that being chilled after a hot day at the beach or pool, by the fans, while drinking cool drinks in the bar may be a cause!

While all these factors discussed above may have some effects on the traveller, one should certainly not seriously consider them as contenders for the condition. It had been known for many years that diarrhoea is a major cause of morbidity and mortality, particularly among the young in the poorer sections of the community in the developed world. According to estimates a few years ago the average child in the rural tropical regions of the developing world might experience up to 25 life-threatening episodes of diarrhoea in its first two years of life, if it lived that long. Some six million children were believed to be dieing before they reach one year. These children and their families tended to live out-of-sight from the travellers and so were not considered.

Thus, it took a long time before an infectious agent was even considered. While intestinal parasites were never seriously considered, there had been some inconclusive studies attempting to link travellers' diarrhoea to a viral aetiolgy. Of the possible bacterial pathogens, which had been taken into consideration, studies to date had shown that only strains of Salmonella and Shigella might be of importance, but maximally they would account for 30-40% of cases, fequently many less. The conclusions reached by about the late 1960's were that an infectious agent was probably involved in travellers' diarrhoea, but was unlikely to be viral or one of the known enteric bacterial pathogens.

(b) Military aspects of Travellers' Diarrhoea.

In the 1960's there were many babies dieing in the developed world as a result of EPEC infections. In addition there were millions dieing from severe diarrhoeal diseases in the developing world. Of these certainly the pathogens Salmonella and Shigella as well as Vibrio cholerae were important causes, but it was certain that there were many others. Paticularly important as a cause of morbity and mortality was a condition known as non-Vibrio cholerae cholera-like diarrhoea, which as the name implies had all the characteristics of cholera, but the aetiological agent was not Vibrio cholerae and was not known. While a few dedicated scientists were trying to unravel these problems, they could not get very far for lack of research funding and equipment.

In the 1960's the cold war was at its height with the superpowers continuously trying to influence the policies of the so-called Third World countries. This frequently meant having to intervene in various disputes in these countries by being able to send and deploy troops rapidly and efficiently. As travellers' diarrhoea was likely to affect up to a third of the troops within 14 days of arrival, this seriously impaired the effectiveness of these forces. It was known from studies during the second world war that this was a serious problem in the deployment of troops in the Middle East, Burma, China and India. More recently United Nations observers in the Middle East, United States forces in Vietnam and British troops in the Arabian Penninsular, British Guiana and East Asia were succumbing in similar proportions of up to 50% with short-lived but incapacitating episodes of diarrhoea.

Now that the military saw travellers' diarrhoea as a problem, which had to be dealt with, money flowed into research into the problem. The crucial study, which solved the problem of the cause of travellers' diarrhoea was thus born and financed, in order to deal with a problem of military logistics. It was financed and supported by the British defence forces and involved 540 troops, who were about to be deployed in Aden, during the last period of British colonial rule there. These troops were studied for their first six weeks in Aden, and encouraged to report any attacks of diarrhoea.

Of these troops, 100 were selected for more detailed study This involved examining a faecal specimen from each man as soon as possible after arrival, and then twice weekly, irrespective of the development of any symptoms of diarrhoea. In addition, faecal specimens were collected from the remaining 440 troops as soon as they reported diarrhoea and in the convalescent state afterwards. None of the troops were given antimicrobial agents unless strains of Salmonella and/or Shigella were isolated.

All these specimens were rapidly studied in a microbiological laboratory. They were examined for parasites, and cultured for all the then known enteric pathogens including Salmonella and Shigella. In addition all the different E. coli-like colonies were noted and representative colonies of each subjected to further analysis. After confirming that most were indeed E. coli by the standard biochemical tests, they were serotyped with all the then available internationally accepted 'O', 'K' and 'H' antisera.

Of the 540 troops in this study, 96 (17.8%) were found to have had at least one episode of acute diarrhoea. As some had more than one episode there were actually 107 episodes of diarrhoea during these six weeks of investigation. The episodes of diarrhoea began four days after arrival in Aden and peaked at ten days. The detailed microbiological study was possible from 74 of these episodes from 64 troops. With over 200 specimens also being processed from 43 healthy troops, it will be realized that this was a major undertaking.

There were some isolations of Salmonella and/or Shigella, but the most noteworthy finding was that a certain serotype of E. coli belonging to a new Ogroup, O148 and carrying the Hantigen H28 was isolated from the majority of the diarrhoeal faecal specimens during the peak period of the cases of travellers' diarrhoea. It was not isolated from any healthy individuals. The isolations of the Salmonella and/or Shigella occurred later and apart from a small outbreak of Salm. muenchen were sporadic. In their concluding remarks the main authors of this study note that this serotype O148:K?:H28, which was associated with 54.3% of all cases of diarrhoea within the first two weeks of the troops' arrival in Aden, should be considered as having been "a major cause of travellers' diarrhoea in Aden".

(c) Escherichia coli and Travellers' Diarrhoea.

Shortly after this investigation involving British troops travelling to Aden, and the discovery of the new serotype of E. coli, O148.H28, and its association with travellers' diarrhoea, the same serotype as well as another O6.H16 were subsequently isolated from American soldiers in Vietnam. Soon afterwrads more serotypes were found to be associated with these conditions. These strains isolated from cases of adult diarrhoea were subjected to a number of pathogenicity tests. It was found that a few were invasive like Shigella. These will be discussed separately under the heading Enteroinvasive E. coli. Most of the strains, however, including representatives of the serotypes O6.H16 and O148.H28 were shown to produce a substance similar in properties and characteristics to the then recently described cholera toxin. This became known as an enterotoxin. Many more serotypes of E. coli have become associated with travellers' diarrhoea. These enterotoxigenic E.coli have also been shown to be associated with the condition known as 'non-Vibrio cholerae cholera-like diarrhoea' A culture filtrate of a strain of E. coli O15.H11 isolated from the small intestine of a patient with a condition indistinguishable clinically from cholera was found to produce effects in both animal experiments and biochemical studies to be identical to those of cholera toxin. These results were found to be consistant with other descriptions of E. coli enterotoxin preparations being investigated at that time. Immunization of rabbits with filtrates of an enterotoxigenic strain of E. coli O78.H12, which had been isolated from an adult with severe cholera-like diarrhoea in Calcutta was shown to give some protection to rabbits against challenge with either live organisms or with cell-free preparations.

(d) The Toxins of Enterotoxigenic Escherichia coli.

These studies above and many others over the years have clearly demonstrated that these organisms produce toxins. It has further been shown that they produce one or both of two enterotoxins, which had been initially discriminated by their stability to increased temperature and been called: the heat stable enterotoxin (ST) and the heat labile enterotoxin (LT). Both these enterotoxins are actually families of variants with serological differences and differences in host range (see below).

As a result of examining two strains from patients with severe cholera-like non-V. cholerae diarrhoea it was found that the strains of serotype O78.H11 and the other O15.H11 were enterotoxigenic and using an animal model two types of enterotoxins were identified. The ST survives boiling for 30 minutes, while the LT is destroyed by such treatment. In addition the LT response was sensitive to acid, while the ST was not. LT preparations from these two strains as well as a porcine strain were antigenically related. The LT was also shown to be closely related to choleragen (CT) the enterotoxin of V.cholerae. The strains of E. coli, which produce these enterotoxins, are now known as enterotoxigenic E. coli or ETEC.

(e) The Heat-Labile Enterotoxin (LT) of Escherichia coli.

The LTs generally have a molecular weight of 73,000. They consist of five identical (B) subunits surrounding a central (A) subunit. They attach to the brush border of the small intestine by the binding of the five B subunits of the LT to the galactose residues of the glycolipid ganglioside GM1. This is specific chemical component of the cell surface. The specificity, with which this binding occurs is similiar to the specificity of an enzyme-substrate interaction. There are distinct differences in binding to different types of ganglioside of the different variants of LT. The A subunit, which is an enzyme, then enters through the 'pore' created in the enterocyte membrane by the ring of five B subunits and interacts with the enzyme adenylate cyclase, such that the adenyl cyclase is permanently active or out of control and 'swiched on'.

This causes the level of cyclic adenosine monophsphate (cAMP) to rise with the result that there is a loss of electrolytes including Na+, Cl- and HCO3- as well as water into the intestinal lumen. cAMP is known as a secondary messenger, which is activated in cells following their reaction with the primary messenger, which is generally a hormone. The hormone (or other primary messenger) determines the specificity of the interaction. By stimulating the secondary or general messenger within the cell, the specific stimulus ellicits a general response. Similarly the LT can be considered as subverting a specific response of the cell (or fooling the cell) into a general response, which acts to the disadvantage of the cell and the organism as a whole.

Further studies on the structures and properties of the LTs have shown that there are two families, which are related in structure and function. They are classified as type I (LTI) and type II (LTII). Antibodies raised in animals against one type do not neutralise the other type, while neutralising members of the inoculum type. The enterotoxins of type LTI have been further subdivided into two types known as LTh-I and LTp-I. These are respectively the enterotoxins produced by ETEC affecting humans (LTh-I) and pigs (LTp-I). Two variants of LTII have also been reported they are designated LTIIa and LTIIb. ETEC producing these LTII variants have been isolated from water-buffalo, cattle, pigs and foods, but they are rarely isolated from humans.

The genes coding for the production of LTI are on a plasmid, a stretch of DNA separate from the main genome of E. coli, which can be passed from organism to another. LTI is excreted into the space between the inner membrane of the E. coli, which surrounds the main cytoplasm of the organism, and the cell wall. This region is known as the periplasmic space.

As was mentioned earlier LT consists of two types of protein, A and B, with the general structure AB5, implying five B subunits surrounding one enzymic A subunit. The A subunit consists of 240 amino acids and each B subunit consists of 103 amino acids. The A subunit consis of two separate polypeptides A1 and A2, which are linked together by a disuphide connection. When originally formed on the ribosome, the two parts of the A subunit are one long polypeptide. They are subsequently split into the A1 and A2 components. Of these, the catalytic action resides on the A1 peptide, while the A2 peptide is currently believed to be required to facilitate the entery into the target cell.

The A subunit has been shown to be wedge shaped, with the pointed carboxyl-terminal end entering the target cell through the doughnut-shaped pore created by the five B subunits.

The main differences between LTh-I and LTp-I are based on the physicochemical and immunological properties of the proteins. The similarities between LTh-I and Cholera toxin (CT) are very great. The amino acid sequences of the peptides are very similar with only scattered amino acids differing. The greatest difference between TTh-I and CT is in the area where the A1 and A2 peptides split from the originally synthesized A polypeptide. It is noteworthy that the B subunits, which are the ones making the contact with the target cells, of LTh-I more closely resemble the B subunits of CT than they do of LTp-I.

The structures of LTIIa and LTIIb are basically very similar to those of the LTI types. The differences are mainly in their physicochemical characteristics and immunological properties. LTIIa and LTIIb also differ both from each other as well as from LTh-I in the specificity of their binding to the the components of the target cell surfaces. Also currently neither LTIIa nor LTIIb have been specifically shown to have a role in disease processes. It is therefore also curious that the genes coding for LTIIa and LTIIb are on the chromosome not on a separate plasmid as was shown for LTI.

The genes which code for these various polypeptides making up the different LT proteins have been sequenced. It has been shown that A subunit genes of both LTI and LTII form branches of an evoloutionary tree, but a similar relationship could not shown for the B subunits of LTI and LTII.

During the synthesis of LTI, the individual A and B subunits, including those part of the amino acid sequence, which later removed, are transported into the periplasmic space, where they are processed and assembled into the AB5 compound protein. Detailed analysis has shown that certain amino acids are essential for this assembly. Most important is the amino acid alanine in position 64 on the B subunit for it to be able to the AB5 compound protein. The close similarity between CT and LTI has be donstrated by the ability to create functionally intact hybrid toxins between the B subunits of one type and the A subunit of the other.

The main mode of action for the B subunits of the LT, is to bind to a specific receptor on the target cell, known as ganglioside, which is given the symbol GM1. LT can also bind to a chemically related ganglioside known as GM2. The tryptophan-88 and glycine-33 aminoacids on the B subunits are required for this binding. Further minor binding differences have been noted with the two LTII proteins.

The most important factor of the mode of action of the LT and CT proteins is that they cause fluid accumulation in the intestinal lumen as a result of the activation of the adenyl cyclase system. Following the entry of the A subunit into the target cell, it's disulphide is reduced forming the two separate A1 and A2 fragment polypeptides. The A1 fragment now acts as an enzyme activating the membrane adenylate cyclse by catalysing the ADP-ribosylation, i.e. binding the moiety ADP-ribose to the regulatory protein Gsa. This means that the synthesis of cAMP is continuously switched on and not under celluar or host control. The ions including Na+, Cl- and HCO3- are continuously excreted from the cell with water molecules following them, the cell tries to replace them from the blood stream and thus diarrhoea and dehydration follow an untreated infection with an LT-producing E. coli.

(f) The Heat-Stable Enterotoxin (ST) of Escherichia coli.

The Heat-Stable Enterotoxin (ST) molecules are remarkable in consisting of only around 18 amino acids. They are a group of closely related peptides resposnible for causing diarrhoea in humans and animals. There are currently recognized among the ETEC two groups. STa (also known as STI) and STb (also known as STII), which are structurally, immunologically and functionally unrelated. While STa is resistant to the action of trypsin, STb is inactivated by this enzyme.

As with many such toxins STI and STII, in fact, represent two families of closely related toxins, each of similar structure and antigenic cross-reactivity. The STa family of toxins includes the toxins STh from human strains of ETEC and STp from porcine strains. Other representative members of this STa family are the ST's from other enterobacteriaceae such as Yersinia enterocolitica, Citrobacter freundii, Klebsiella sp. as well as Vibrio cholerae, and V. mimicus.

They maintain a very strong structural integrity due to six cystine molecules, which hold them in position with their three disuphide bonds. All STa molecules have a conserved region of ten amino acids located in a similar situation with respect to the molecule as a whole. This strongly suggests that their tertiary structure is very similar. Of the two different STa types, STp consists of 18 and STh 19 amino acid residues respectively, which share 13 common amino acids in the active centre of the molecule. The toxins result from a 72-amino-acid precurser, known as the pre-pro-STp or pre-pro-STh respectively. This is first cut by a proteolytic enzyme into a 53-amino-acid pro-STp or STh within the periplasmic space of the E. coli. The fully effective mature toxins are processed outside the cell from this 53-amino-acid pro-STp or STh.

Both the STp or STh toxins in general retain their biological effects even after heating at 100°C for 30 minutes. However, when STh at a very low concentration of 30ng/ml is heated to 100°C for only 10 minutes it is inactivated, while at double that concentration it is not. The toxins are inactivated by reducing agents, which reduce the disulphide bonds.

The mode of action of all these molecules is similar in that following attachment of the ETEC to the intestinal cell, known as the enterocyte, the STI is bound to a specific brush border membrane receptor. This binding then activates the guanylate cyclase-cyclic guanosine monophosphate system, leading to an increase in the intracellular concentration of guanine 3',5'-cyclic monophosphate (cGMP). This is followed by the activation of a cGMP dependent protein kinase. The result is an impairment of the coupled sodium and chloride influx and stimulating active chloride secretion, resulting in the blockage of neutral NaCl absorption.

Detailed electron micrographic studies of rat jejunal mucosa showed that the changes induced by STI could be observed in both vilous as well as crypt cells. First there was a swelling of the cells. The intracellular ionic concentrations of Na+, Cl- and Ca2+ increased and those of K+ and Mg2+ decreased. About 3 hours later there was observed an increase in extracellular space and a decrease in the cell volume also the concentration of all above mentioned inorganic ions is augmented.

An examination of surgical specimens from the small intestine and 20 from the colon derived from children ranging in age from 1day to 16 years found that, while all sections bound STI, there were more ST receptors in both small intestinal and colonic fractions in the youngest children. The number of receptors rapidly fell off with age. The children's sex, race, blood group, nutritional status or underlying diagnosis was unrelated to binding of STI. A close relationship between STI stimulated guanyl cyclase activity and STI-receptor density was also noted. These increased numbers of STI-receptors in the youngest children may be a contributor to the severity of their diarrhoea.

STI also seems to have a modulating action on the humoral immune response. Injecting ST into mice during their development of antibodies to sheep erythrocytes, both immunosuppression and immunostimulation were observed at varying stages during the innoculation course.

There is far less known about STII than about STI. The mature toxin consists of 48 amino acids, which is derived from a 71 amino acid polypeptide, whose 23 amino-terminal amino acids are cleaved off in a single event in the periplasmic space. Four cysteine residues forming intramolecular disulphide bonds are present in each STII molecule. Like LT and STI, STII is plasmid encoded, but the codes for STII and the other two enterotoxins are unrelated.

STII has been mainly associated with ETEC isolated from pigs with diarrhoea and initially it appeared that STII only acted in pigs, but when applied in conjunction with a trypsin inhibitor, responses to STII could also be shown to act in mice, rats, rabbits and calves. However, it remains that the most common ETEC isolated from swine, carry genes for STII. STII gene carrying ETEC have also been isolated from humans with travellers' diarrhoea.

STII has no effect on either the levels of cGMP or cAMP unlike STI and LT respectively. It causes an increase in the levels of the hormone protaglandin E2.

(g)Colonization Factors of ETEC.

Adherence is a very important prerequisit for a bacterium's ability to colonise a host and a number of adherence factors or colonization factors have been described for E. coli. There is a strong specificity among the ETEC and their ability to bind to different animal host cells. Some of these adhesins are fimbriae such as the type 1 fimbriae, which are present in nearly all strains of E. coli. Other fimbriae were intially described as "K" antigens but were also shown to be fimbrial in structure. This include the K88 which is an adherence factor for piglets the K99 which is an adherence factor for a number of animals. The first human factor identified was the colonization factor antigen (CFA) I to be closely followed by a second designated CFA II. Since then other factors including a CFA III for humans and factors currently only named after their strain names have been identified. These include F41 for calves, 987P for piglets and E8775 for humans. It is noteworthy that different adhesins, specific for different hosts are present on the same E. coli.

Most of these fimbriae are rigid, hair-like filamentous structures. Their diameter varies from 5-7 nm, but that of some including K88 and K99 is thinner. Due to this, they are sometimes refferred to as fibrillae. All the fimbriae are composed of protein elements. These are assembled into helical structures. There is also a specificity between a strain of E. coli's ability to produce one of these types of fimbriae and its O group. Thus, the K88 fimbriae are only formed by strains belonging to O groups O8, O45, O138, O141, O147, O149 and O157, while CFA/I by O groups O4, O7, O20, O25, O63, O78, O110, O126, O128, O136, O153 and O159. The genes which code for these adhesins may be on the E. coli chromosome or on plasmids. These plasmids may also code for some of the enterotoxins.

Two families of adhesin fimbriae have been identified on the basis of the amino squences of the protein subunits of the fimbriae. The first group includes the fimbriae K88, F41, CS31A from animal ETEC and CS13 from human ETEC. They are composed of relatively large subunits, which share limited homology in amino acid sequence. The operons controlling their production are also very similarly organised. This suggests very strongly that they may have originated from a common ancestor. Serological variants amongst some of these fimbriae have also been observed especially K88, for which three main groups notated as K88ab, K88ac and K88ad have been identified.

The molecular weight of the subunits belonging to the second group is smaller and the amino acid sequences are even more similar amongst each other than in the first group. This group comprises the fimbriae CFA/I, CS1, CS2, CS4, PCFO166 and CS17. All are from ETEC affecting humans and again a common ancestor has been postulated.

The general view is that fimbriae interact with glycolipids and/or glycoproteins in a way similar to the lectin-interractions. It is believed that there are a variety of glyco-conjugates present both within the mucus and/or the brush border cells, which are recognized by the fimbrial adhesin. An oligosaccharide projecting such that a specific epitope is exposed fits with the part of the fimbria, which acts as receptor binding site. A fixed spacial arrangement of carbohydrate moieties on these receptors has been shown.

Bacteria such as E. coli have been shown to grow better and faster in mucus than they are sloughed off. They can even grow on phophatidylserine a component of the mucus as sole source of carbon and nitrogen.

In some cases the specific arrangements of the sugar moieties acting as receptors have been identified. Those glycolipids, which carry Gal-(1Æ3)-Gal residues are probably the receptors for the K88 fimbriae. K99 fimbriae bind to acid glycolipids N-glycolyl-GM3 and CFA/I bind sialic-acid-containing glyco-conjugates, which include ganglioside GM2.

There appear age differences in animals in the number of receptors carried on their intestinal cells. In seven week old pigs there is 16 times more K88 receptor in the ileal mucus than in neonatal pigs. Therefore, K88 carrying ETEC are generally prevented from binding to the older piglets' intestinal epithelial cells, being held in the mucus. In the neonatal piglets they are not held back, bind to the intestinal epithelial cells, release enterotoxin and cause diarrhoea.

Like the synthesis of all cellular components, there are gene clusters involved in the biosynthesis of fimbriae in E. coli. These clusters contain genes coding for regulatory proteins, for proteins required to transport the fimbrial subunits accross the E. coli cell enevelope, as well as for minor and major fimbrial subunits. Generally the clusters encode a pariplasmic chaperone and high molecular weight protein, which is associated with the outer membrane. The CFA/I and CS1 gene cluster is exceptional in lacking this.

The biosynthesis of the fimbriae is repressed at temperatures below 37°C. It has also been shown that the production of K99, fimbriae is subject to glucose repression. Similar catabolite repression also controls the production of CFA/I and CFA/II fimbriae. There are also effects on fimbrial expression by other metabolites.

(h)ETEC as Human Pathogens.

It has been discussed in general terms at the beginning of this section that ETEC are commonly associated with travelers' diarrhoea. A more detailed discussion of cases and the general human ecological aspects of ETEC will be considered here.

During an investigation into the enteric pathogens among travelers and foreign residents with diarrhoea in Nepal, it was found that the largest number of patients (24%) yielded ETEC. To demonstrate the variety of pathogenic phenotypes found it was noted that of the 79 strains examined, there were 32 LT+ST+, 27 ST+ and 20 LT+.

Another study invloving a two-year survey of returning travelers to the Sheffield area in England found that 35 (5.8%) of 600 travelers from the more developed countries such as the Mediterraniean region carried LT+ ETEC, while 36 (11.3%) of 320 travelers to less developed countries yielded LT+ ETEC.

ETEC are a relatively common cause of diarrhoea in developing countries but the number of reports of their occurrence in developed countries is comparatively small. A study of faecal isolates of E. coli from 106 consecutive cases of watery diarrhoea in infants and children in Valencia, Spain between January and August 1984 found that five neonates over a period of three weeks yielded strains of E. coli O153.H45 ST+ all belonging to the same biotype. Another unconnected case of diarrhoea yielded a strain of E. coli O128.H49 LT+. A study of 112 strains of the same serogroup of O153 isolated in other parts of the world showed that the types had been previously found in Spain, and also in Italy, Chile and Liberia, but not in Asia, from where many ETEC strains have been collected and serotyped. The ETEC serotype O128.H49 LT+ has been isolated in other parts of Europe. This study shows that ETEC may be more widespread in the developed world than current reports suggest.

However, general examination for ETEC in Europe and other developed countries is not recommended, but in outbreak situations, studies for the possible presence of ETEC should be undertaken.

More extensive studies on strains of E. coli O153.H45 ST+, which has been shown to be widely disseminated in Chile, revealed that production of the CFA/I was a common feature of many strains. The ST genes were contained in non-congugable plasmids. This suggests that ST production is a very stable characteristic of these strains.

A recent study on strains of ETEC from Italy and Somalia also showed a variety of serotypes and phenotypes, with eight different serotypes being represented among the nine Italian strains and eight serotypes among the twelve Somalian strains. The two Italian strains which belonged to the same serotype were both O6.H16 and also had the same phenotype with respect to the other characteristics. They were the only strains in this collection to produce both LT and ST, the CFA of type II and mannose resistant haemagglutination (MRHA) with bovine but not human erythrocytes. In addition they were resistant to streptomycin and sulphametoxazole. These were also the only Italian strains to belong to serotypes considered to be of the global group of classical ETEC serotypes. It is therefore noteworthy that one of these strains was isolated from a recent traveller to Tunisia and the other from an asymptomatic infant. Of all the other Italian strains five were LT+ and two ST+. None of them produced MRHA or CFA. Three were sensitive to all antimicrobials tested and four resistant to a variety of agents. Four of the Somalian strains belonged to serogroup O128.H49, were LT+, and did not produce MRHA or CFA and were fully sensitive. Of the remainder all were ST+ and were O166 H27. O166.HR and O153.H45. This study again indicates how ETEC can be transferred with travellers from one geographical region to another.

Perhaps the the most noteworthy result of this investigation is the variety of serotypes encountered and the rarity of the so-called common or classical ETEC types. This confirms the earlier suggestion, that the ability to carry the enterotoxin genes can be present in many strains of E. coli , but may perhaps not be as stable in those. Thus one Italian strain of serotype O89.H45 lost its ability to produce LT after two years' storage.

A recent survey for the presence of ETEC among 187 faecal specimens from 146 Australian aboriginal children with diarrhoea in Alice Springs used DNA probes to identify strains producing LT or ST, found that 50 patients had ETEC. Sixteen specimens from control patients did not yield any ETEC. The ETEC were also serotyped and again no particular pattern was observed, though O115 and O126 appeared to predominate among the ST producing strains. E. coli belonging to serogroup O126 producing ST have also been found associated with an outbreak of diarrhoea in adults and children in New Zealand a few years ago.

An investigation into the causes of diarrhoea among Mexican children during the first year and six months of life, from a village approximately 180 km SW of Mexico City, showed that 19% yielded serologically identified EPEC and 33% ETEC of which nearly half produced both LT and ST, while only 4% of control children yielded EPEC and 20% ETEC. This suggests that ETEC are widespread in the community. Although other enteric pathogens were also found they were in much lower proportions. The isolation rate of ETEC from cases of diarrhoea stayed in the 30% range throughout the first two years of life. The presence of these organisms in such cases in the developing countries has been considered the main reason why these organisms are particularly associated with travelers' diarrhoea.

This brief survey of a few of the many reports from around the world should demonstrate the importance of ETEC as causes of human mortality and morbidity. Until the hygienic standards of water availability and sewage disposal are significantly improved, the ETEC will be major problems in the developing world and for travelers to these countries.

(i) ETEC as Animal Pathogens.

(I) Pigs

Post weaning diarrhoea (PWD) of piglets is still considered as one of the major causes of economic loss to the the pork rearing industry. Apart causing significant production losses by reduced growth rates of the piglets, there is also an average of 1.5-2.0% mortality of weaned pigs. Although PWD has a number of linked causes the main pathogens involved are ETEC.

PWD manifests itself by the piglets producing for about a week a fluid, yellowish to gray diarrhoea. In addition the animals have reduced appetite and appear depressed, developing a loss of condition such as rough hair coat, becoming pot-bellied and shivering. Untreated animals can reach a mortality of up to 25%. Occasionally during an outbreak animals in apparently good condition suddenly die. Occasionally outbreaks are also linked to neurological signs among the piglets.

ETEC producing either STa, STb, LT or a combination of these are the main organisms involved in PWD, which is therefore also known as post weaning colibacillosis. The genes coding for these toxins are plasmid mediated. The ETEC involved in PWD can also carry an additional toxin known as Shiga-like Toxin IIv (SLT IIv). This toxin goes under a number of names including SLT IIe, Verotoxin 2v, and Verotoxin 2e. The suffix v was initially added to distinguish it as a variant from the human Shiga-like Toxin II but now is more commonly referred to as SLT IIe indicating its role in causing oedema. These toxins will be discussed at length in the section on Enterohaemorrhagic E. coli (EHEC).

The full role of SLT IIe in PWD has not been fully clarified. In addition there have been isolated strains of E. coli exclusively, which do not produce any of these toxins. The serotypes of E. coli involved in PWD tend to include types not normally associated with human infection, including strains belonging to serogroups O8, O45, O116, O138, O139, O141, O147, O149 and O157. Also the adhesive factors associated with these porcine ETEC are different to those found in human strains. These include the pili initially believed to be K antigens, known as K81, K82, K85, K87, K88 and K91. Other pili have more recently been described. Differences between the course of the diseases of different ETEC phenotypes have been noted. K88+ strains are more likely to cause PWD some four days after weaning while wth K88- strains the onset of PWD is delayed to 7-10 days post-weaning.

(II) Cattle

Enteric Colibacillosis is an important condition affecting calf production and development. It is characterised by the production of profuse amounts of foul-smelling, pale yellow to white watery faeces, which may occasionally have specks of blood in them. The animals get increasingly dehydrated due to fluid and electrolyte loss. Their condition rapidly deteriorates specially marked are loss of skin elasticity, sinking of the eyes into their sockets, lassitude and loss of appetite. In extreme cases body weight can decrease by up to 10% only about eight hours after onset of diarrhoea. The disease can also take on a chronic form. In both cases the disease may overwhelm the animal leading to death. In herds morbidity rates of up to 50% have been noted, even reaching 75% in confined dairy herds. Depending on the virulence of causative ETEC, mortality rates of 10-50% have been described. These bovine ETEC can thus cause severe economic losses to a beef or dairy industry.

The bovine ETEC only tend to produce STa, with LT-producing ETEC being only very rarely isolated from calves. The pili, originally known as the K antigen K99, are the main factor mediating the adherence of these bovine ETEC to the bovine ileum. Attachment of K99 strains predominantly occurs to the posterior small intestine.

Apart from the K99 pili, there are other adherence factors, which also play a role in the attachment of ETEC to the bovine intestine. These include the F41 and F6 fimbriae. While these latter types mainly occur together with the K99 adhesins, there are ETEC, which only carry one of them. The O serogroups most commonly associated with bovine disease are O8, O9, O20 and O101. Often these have certain K antigens as well such as O8:K25, O8:K85, O9:K35, O101:K28 and O101:K30.

(III) Sheep

Lambs have been known to develop a scouring disease from about the first few days after birth. The characteristic of this often fatal disease caused by ETEC in lambs is its occurrence so soon after birth. They often carry as many as 108 bacteria per gram of intestinal contents. The ETEC predominate in the intestinal lumen. The main cause of death is dehydration, acidosis and shock due to fluid and electrolyte loss.

The ovine ETEC belong to the same O groups as the bovine ETEC, being predominantly O8, O9, O20 and O101. Similarly, most produce the fimbrial adhesin K99 but some strains also produce in addition F41. The enterotoxin STa is the main toxin associated with ovine ETEC. In many ways the ovine and bovine ETEC are very similar.

(IV) Goats

Newborn goat kids are prone to diarrhoea within the first few weeks of life. A profuse watery diarrhoea, often followed by dehydration and death, can have a significant economic effects. The K99 adhesin is associated with most strains involved in these conditions.

(V) Horses

There have been a few reports of diarrhoea in foals associated with ETEC. Haemolytic E. coli of serogroups O101, O8 and O147 were most commonly associated with these infections. However, while these serogroups are similar to ones found among ETEC from other animals, most notably pigs and cattle, they did not produce the typical enterotoxins or adhesins. However, foals appear to have a K88 receptor in their intestine and deliberate infection with a K88+LT+ ETEC did lead to profuse watery diarrhoea in a colostrum-deprived foal. Under natural conditions ETEC do not seem to cause disease in horses.

(VI) Poultry

There have been a few reports of ETEC causing enteritis in poultry. There seem to be pathogenic types of E. coli present in nearly all the intestinal tracts of poultry, which have been examined, without them causing disease. Septicaemic infections in poultry often associated with watery, yellowinsh faecal droppings and significant reductions in body weight have been noted. ETEC seem very rarely to be associated with such diseases. There has been a report from the Philippines of ETEC producing a heat-labile enterotoxin similar to the human LT isolated from from cases of poultry enteritis. Some of the strains also produced STa. The serogroups of ETEC isolated from from cases of poultry eneteritis seem to be the same types associated with porcine strains, namely O8 and O149.

(VII) Dogs

ETEC have been isolated from dogs, but they do not alsways appear to be associated with disease in these animals. Isolation rates from dogs with diarrhoea have been reported between 3 and 30% and it has been suggested that in dogs, additional aetiological agents have to be present to cause for ETEC to be involved in disease.

Most canine ETEC produce STa, while LT producing ETEC are only occasionally recovered. There have been a number of serogroups associated with canine ETEC, of which O4, O6, O8 and O42 were the most common. Less commonly isolated were strains belonging to serogroups O5, O17, O20, O23, O25, O70 and O105. In some strains, most notably strains of O42:H37, the presence of K99 adhesin was observed, but generally the human and porcine adhesins are not found associated with canine ETEC.

(VIII) Cats

There have not been sufficient studies on enteritis in cats to determine whether ETEC play a role in these animals.

(IX) Rabbits

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

© 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.