The association of verotoxigenic Escherichia coli
(VTEC) with human disease goes back over 30 years [1
]. The occurrence of outbreaks due to VTEC in the USA in 1982 [4
] focused the world's attention onto these pathogens. Since the discovery of verocytotoxin [1
], and the paper by Karmali et al.
] of cases of post-diarrheal hemolytic uremic syndrome (D+HUS) caused by VTEC, otherwise known as Shiga-toxigenic Escherichia coli
(STEC), a large body of knowledge has accumulated, yet despite this information, successful treatment of these infections has remained elusive.
Sources and pathogenesis of VTEC infection
Sources and spread of VTEC
Gut colonization of farm animals, especially ruminants such as cattle, sheep and goats is the likely origin of VTEC/STEC. From these sources derive a variety of vehicles of transmission to humans, including many different foods of animal or plant origin, and water used for swimming and drinking and for growing edible plants. Human fecal contamination of food and seeds could also play a role, especially in developing countries [6
The potential for VTEC spread is further compounded by globalization of food, which presents a great opportunity for VTEC to spread quickly to large sections of the population. Global food distribution carries an inherent risk and presents great difficulties in controlling food-borne pathogens and in identifying sources of outbreaks, as was recently witnessed in Europe. This is further discussed in the commentary by Werber et al.
Various strains of VTEC exist, and, as discussed in the linked commentary, O157 clones, although less prevalent than non-O157 strains, tend to be more virulent. Thus, although non-O157 VTEC strains had originally been reported and continued to be reported, albeit only by dedicated microbiologists, most researchers in the field largely ignored them. No attention appears to have been given to the generally observed fact that there is a widespread diversity of E. coli
serotypes in the human intestine at any one time [8
] and this has also been found in animals, especially cattle [9
]. Most ruminant feces contain a variety of VTEC serotypes, but some, such as O157 and also O111, though rarely present and then in only small numbers, are particularly virulent. Importantly, an increasing number of other serotypes can be involved and one study of an outbreak has shown that the more VTEC serotypes with which a patient is infected, the worse the clinical condition [10
] (though the main VTEC serotype was O111). While isolations of VTEC O111 from cattle are rare, non-VTEC strains, which are otherwise indistinguishable from the VTEC strains, appear abundant, especially in the feces of sick cattle and patients [11
Detailed studies [12
] have shown that the Shiga toxins can be subdivided into a series of subtypes and that these are also host specific. Thus there is a 'double host specificity' among VTEC strains. Some clones are specific to cattle, while others are specific to sheep. The toxin subtypes these strains carry are specific to the VTEC types found in these mammalian hosts. Therefore, by not looking for the presence of all VTEC serotypes during an outbreak, a great deal of epidemiological information is lost and indication of the source animal is not identified.
Pathogenesis of post-diarrheal hemolytic uremic syndrome
VTEC/STEC/enterohemorrhagic E. coli
(EHEC) belong to clones of zoonotic E. coli
of different O serogroups. These serogroups have evolved and acquired specific virulence factors that enable the bacteria to colonize and infect the human colon, usually without invasion of the blood stream [13
]. Once they have been ingested, STEC/VTEC/EHEC cause bloody diarrhea (BD), severe colitis and HUS. These bacteria are known as EHEC when infection is associated with severe colonic and/or renal disease. The production of Vero/Shiga toxins have been considered the basis for their pathogenicity, however, other toxins such as subtilase cytotoxin (SubAB) [14
] and cytolethal distending toxin [15
] and secreted protease of C1 esterase inhibitor from EHEC (StcE) probably play a role [16
The recent outbreak of foodborne E. coli
O104:H4 in Europe has once again drawn attention to STEC or EHEC infections together with their devastating complications of renal failure (through HUS), and stroke from intravascular coagulopathy and vasculopathy or thrombotic microangiopathy. The unusual virulence and lethality of the O104:H4 strain is the result of genetic admixture of virulence factors, including enteroaggregative properties and multiple antibiotic resistance, and is a lesson in microbial evolution and the genomic plasticity of E. coli
]. The O104:H4 strain is now known as an enteroaggregative and enterohemorrhagic E. coli
We have recently observed the combined properties of enteroaggregative ability (providing strong attachment via fimbriae and colonization of the colonic epithelium) with Shiga toxin (Stx) production in the novel and highly lethal European E. coli
O104:H4 strain. It has since been shown that this strain belonged to an enteroaggregative E. coli
lineage that had acquired genes for Shiga toxin 2 and antibiotic resistance [18
The pathogenesis of HUS disease remains incompletely understood; remarkably, during HUS serum Stx is undetectable. It seems polymorphonuclear leukocytes (PMN) are key players in delivering Stx to critical sites such as the kidneys. The extent of renal damage in children with STEC-associated HUS may relate to the concentration of Stx present on circulating PMN [19
]. Paradoxically, patients with high amounts of Stx on PMN showed preserved or slightly impaired renal function (incomplete form of HUS), whereas cases with low amounts of PMN-Stx usually present with acute renal failure. Moreover, high amounts of PMN-Stx induce a reduced release of cytokines by the renal endothelium, with congruent lower degree of inflammation, while low toxin PMN amounts trigger a cytokine cascade, provoking inflammation with consequent tissue damage. The microvasculature plays an important role in pathogenesis: D+HUS is associated with platelet thrombi in the microvasculature of almost all vascular beds [20
]. Plasma from HUS patients induces apoptosis of cultured microvascular endothelial cells from most organs [21
]. Two key events are involved in the pathogenesis of D+HUS: altered Von Willebrand factor (VWF) activity (for example, as seen with 'a disintegrin and metalloproteinase with thrombospondin motif-13' (ADAMTS13) deficiency) and site-specific activation and/or apoptosis of microvascular endothelial cells. A deficiency in ADAMTS13, which mediates proteolytic processing of newly released proadhesive ultralarge VWF multimers from endothelial cells, is also thought to play a role in D+HUS coagulopathy [22
]. Targeting the interruption of these processes gives hope for potential novel treatment modalities.
Bacterial gut pathogens target the follicle-associated epithelium overlying Peyer's patches. The microorganisms breech the intestinal barrier via M cells and are captured by mucosal macrophages [23
]. STEC/EHEC are able to interact in vivo
with Peyer's patches and translocate through the mucosa. After being taken up by macrophages and M cells the bacteria produce Stx and induce apoptosis of these host cells and Stx release. These microbe/host cell interactions could represent new therapeutic targets [23
Current treatment strategies: a multitargeted approach
HUS comprises acute renal failure and its consequential perturbation of fluid and electrolyte balance, hemolysis, disruption of the clotting cascade with thrombocytopenia, with the risk of stroke. This syndrome, together with the further effects of toxin, and complement complex formation, must be managed and addressed urgently using a multitargeted approach. This involves the institution of general supportive measures, antiplatelet and thrombolytic agents and thrombin inhibitors, selective use of antimicrobials, probiotics, toxin neutralizers (synthetic and natural binders, antibodies, and so on); and antibodies against key pathogenetic pathway elements to interrupt pathological processes (for example, inhibition of terminal complement complex formation). Targeting PMNs carrying Stx could be a productive strategy for future research, as could possible gene therapy. The management of D+HUS is complex by virtue of the nature of the condition and the variety of pathways affected. Table summarizes the approach to management and lists trialed and experimental treatments.
Approach to management: summarizing trialed and experimental treatments.
General supportive measures
Fluid levels and electrolyte balance are extremely important in preventing and managing the development of HUS [24
] (See Table ).
Acute renal replacement therapy (ARRT); for example, peritoneal dialysis (PD) or hemodialysis) has been shown to improve outcomes. Children with D+HUS and acute kidney injury given early PD may have improved outcomes without risk of bleeding in patients with low platelet counts. Moreover, the procedure seems safe especially in cases with very low platelet counts with no bleeding episodes recorded [26
]. Alternatively, hemodialysis is often necessary. Antihypertensive therapy for hypertension when appropriate is also necessary. There seems to be a beneficial role for plasma infusion [27
] and plasma exchange [28
], however, benefit from apheresis remains uncertain [29
Managing hematological issues and coagulopathy
Monitoring of hemoglobin, hematocrit and platelet count is essential. Monitoring hemolysis with lactate dehydrogenase (LDH) and haptoglobin is also helpful. Anemia resulting from hemolysis may need correction with transfusions of whole blood or packed red cells. Platelet transfusion is rarely required and usually avoided [13
Preventing the further effects of toxin
Antimicrobials: to use or to avoid?
Due to the potential for undesirable release of verotoxin (VT)/Stx by dying and dead bacterial cells, antibiotics are usually avoided [31
]. In addition, the risk of endotoxin release could add to the patient's already potentially lethal burden. In vitro
subinhibitory concentrations of antibiotics may increase production and release of VT/Stx [32
] via bacteriophage induction [33
]. A mouse [34
] and piglet study [35
] suggested human trials of fosfomycin were warranted. However, pooled prospective data showed no benefit of antibiotics [36
]. Only one fosfomycin trial has been reported [37
]. However, fosfomycin data has been questioned [38
] (See Table ). While many doctors in Japan still use antibiotics including fosfomycin in patients with definite or possible enteric STEC infections the prevailing consensus elsewhere indicates antibiotics should be avoided [13
]. More recent evidence supports this especially in relation to β lactam and other bactericidal antibiotics [39
Lumenal toxin neutralizers (synthetic and natural binders, antibodies, and so on)
Strategies using ligand mimics of the receptor for Stx, globotriaosylceramide (Gb3), binding to Stx in the gastrointestinal tract with the intention of preventing the spread of toxin to extraintestinal sites have been proposed. However, in clinical practice the damage has already been done before these ligands could be of benefit. Only one clinical trial has been conducted (alas unsuccessfully) with one agent, Synsorb PK, which bore out this fact [40
]. Other agents are listed in Table [41
Intralumenal neutralizers might be effective in reducing systemic uptake of toxin but because the toxin is purportedly not found in serum, studies designed to examine the effect of neutralizers on the toxic effects of polymorphonuclear leukocyte-associated toxin would be a first step.
Neutralizing Shiga toxin-specific antibodies are potentially useful as therapeutic agents. The toxins are AB toxins with active and binding elements and are obvious targets for antibody neutralization. Monoclonal antibodies targeting the A subunit epitopes of Stx1 have been shown to be highly protective, when administered to lethally treated animals [43
]. Orally administered immunoglobulin has been used therapeutically for a number of gastrointestinal infections (for example, rotavirus; Gastrogard-R) [44
]. Patients with diarrhea caused by diarrheagenic E. coli
, specifically STEC and E. coli
-expressing intimin and HEC-hemolysin were treated by administration of pooled bovine colostrum, rich in antibodies to Shiga toxin and enterohemorrhagic E. coli
-hemolysin, in a placebo-controlled, double-blind study. Symptom resolution and fecal excretion of infecting strains were assessed. No effect of colostrum therapy on the carriage of the pathogens or on complications of the infection could be demonstrated, however, stool frequency was reduced [45
]. Antibody to E. coli
lipopolysaccharide (LPS) also has the potential of therapeutic use through its blocking effect on adherence of STEC to the human intestinal epithelial (Henle 407) cell line [46
]. Likewise, human trials would be needed to show clinical effectiveness.
Other toxin binders/neutralizers
Most of these agents bind to toxin directly and inhibit the binding to its receptor present on the target cells [47
]. Such novel Stx neutralizers offer a new therapeutic modality against STEC/EHEC infections [47
] and are detailed in Table .
Systemically-applied (intravenous) toxin binders
A cell-permeable peptide (TVP) that binds to Stx2 was shown to reduce disease severity and rescue juvenile baboons from a lethal Stx2 dose (50 ng/kg) [48
Blockers of endosome-to-Golgi trafficking of Stx
Recently it was shown that the metal manganese (Mn2+
) blocks endosome-to-Golgi trafficking of STx. [49
] This offers a possible cheap therapeutic approach. (Table ).
Blockers of bacterial and host cell interaction: probiotics
Gut pathogens display surface molecules enabling the organism to bind to host cell receptors. Similarly, bacterial toxins require host cell receptors for binding and cell entry. To block microbe and host cell interaction 'designer' probiotics have been developed. The harmless recombinant bacteria express molecules that mimic host cell receptors (for example, Gb3) on their surface, thereby deceiving the pathogen into attaching to the probiotic rather than the host cell receptor. Probiotic bacteria must survive the tube journey encountering digestive enzymes and other adverse conditions. Trial data is awaited.
A different approach has used the supernatant of cultures of Bifidobacterium longum
HY8001, designed to inhibit the effect of VT/Stx through interference of B subunit of VTs in binding to Gb3 [50
Inhibition of terminal complement complex formation
Based on evidence that Shiga toxin activates complement and binds factor H and evidence for an active role of complement via the alternative pathway in diarrhea-associated hemolytic uremic syndrome [51
], a few anecdotal reports of successful treatment of severe Stx-associated HUS with the monoclonal antibody eculizumab have been published [53
]. Neurologically, the three patients improved dramatically within 24 h after the first eculizumab infusion. Clinical improvement was associated with rapid normalization of markers of disease activity. These initial results are extremely promising and outcomes from large-scale randomized placebo-controlled trials are optimistically awaited.
Several vaccine strategies have been used with variable success in a number of animal models. The strategies have involved the use of recombinant virulence proteins such as Stx, intimin and E. coli
secreted protein A (EspA) [54
] or peptides [55
] or fusion proteins of A and B subunits of Stx2 and Stx1 such as Stx2Am-Stx1B [56
] or avirulent ghost cells of EHEC O157:H7 [57
]. The application of live attenuated bacteria such as Salmonella
as a carrier for vaccine proteins against mucosal pathogens including EHEC have obvious advantages [58
]. Other approaches are listed in Table [59
Antibodies produced in humans with HUS and in rabbits immunized with type III secreted proteins (T3SPs) from four STEC serotypes, and experimentally infected cattle revealed proteins common to several HUS serotypes [60
] (Table ). These were highly immunogenic in vaccinated and naturally infected subjects and represent future candidates for a STEC vaccine (Table ).
As well as protein-based vaccines, DNA vaccines are a recent development in EHEC prevention, providing encouraging results in a mouse model [61
] (Table ).
The mode of administration (intramuscular, intranasal, oral, intragastric, and so on) for a number of these vaccines not only affects immunogenicity but also protective effect under challenge. Vaccination with a plant-based oral vaccine protected mice against lethal systemic intoxication with Stx2 [62
]. This is seen as encouraging. Clearly there is some time to go before human trials are reported but the numerous and frequent outbreaks of EHEC disease constantly remind us of the urgent need to protect the population against these emerging and often devastating zoonoses.
Future directions and conclusions
There remain significant barriers to successful treatment of HUS given the complexity of the pathogenesis of HUS, which involves perturbation of key homeostatic pathways involving complex biochemical and physiological systems. It is unlikely that targeting a single pathway with a treatment modality will be sufficiently successful; a multitargeted approach would seem necessary. However, given the apparent success of eculizumab, albeit with tiny case numbers, it could offer a promising strategy for treatment. Treatment is designed to prevent the most serious complications of STEC infection (that is, renal failure and central nervous complications, for example, stroke, and shock), which remain far too common. It is clear that a better understanding of the pathogenesis of HUS will lead to additional and possibly better targets for treatment. The discovery that Mn2+ can block endosome-to-Golgi trafficking will no doubt lead to randomised controlled trials in humans. These will be awaited with keen interest. In terms of prevention, we should question the globalization of food distribution with its inherent dangers and its wasteful use of energy resources resulting in a giant carbon footprint.