Receptor-mimic probiotics: potential therapeutics for bacterial toxin-mediated enteric diseases

Many important bacterial pathogens cause gastrointestinal disease in humans by colonizing the gut and elaborating potent exotoxins. These toxins recognize specific oligosaccharide receptors displayed on host glycolipids or glycoproteins, and then enter the cells and corrupt essential functions. The presence and distribution of the target receptors affects host susceptibility, tissue tropism of the toxin and clinical manifestations of the resultant disease. By way of example, Vibrio cholerae produces cholera toxin (Ctx), which is one of the AB5 family toxins, so called because they comprise an enzymatic A subunit, noncovalently linked to a pentameric B subunit that is responsible for cell binding. Enterotoxigenic Escherichia coli (ETEC) produce a very closely related AB5 toxin called heat-labile enterotoxin (LT), which has an identical mode of action. The B pentamers bind to the oligosaccharide components of certain gangliosides, particularly GM1, that are present on the gut epithelial cell surface. This triggers endocytosis of the holotoxin and retrograde trafficking via the Golgi to the endoplasmic reticulum. Here, the A subunits dissociate and are retro-translocated into the cytosol, where they ADP-ribosylate the host cell G_s protein, resulting in uncontrolled stimulation of adenylate cyclase. This, in turn, interferes with ion transport, thereby causing copious watery diarrhea and electrolyte imbalance, which is the hallmark of cholera and ETEC disease. Shigatoxigenic E. coli (STEC) produce Shiga toxin (Stx), a distinct AB5 toxin, whose B pentamer binds to globoseries glycosphingolipids, such as Gb₃, while its subunit is an RNA-N-glycosidase that modifies 28S rRNA, thereby inhibiting host protein synthesis. Interestingly, human intestinal epithelium expresses very little Gb₃, so Stx has minimal direct enterotoxic effects. However, microvascular endothelial cells, particularly in the gut, kidney and brain, are rich in this glycolipid and are highly susceptible to Stx that has been absorbed systemically from the gut lumen. Stx-mediated endothelial injury leads to the formation of microthrombi, and the resultant ischemic damage accounts for the gastrointestinal symptoms of STEC disease (abdominal pain, diarrhea and hemorrhagic colitis), as well as the potentially fatal hemolytic uremic syndrome. Clearly, blockade of the initial interaction between AB5 toxins and their host glycan receptors has the capacity to prevent the aforementioned diseases from developing.

Toxin–receptor blockade as an anti-infective strategy

The design of therapeutics capable of blocking toxin–receptor interactions has been facilitated by access to crystal structures of AB5 toxin B pentamers in complex with their cognate glycans [1]. Earlier studies had shown that simple oligosaccharides mimicking the receptor structure were not particularly potent toxin inhibitors. However, vastly improved toxin binding kinetics were achieved when analogous oligosaccharides were displayed on complex molecular scaffolds that optimized multivalent docking with the B pentamer [2–5]. At least in some cases, protection in animal models has been demonstrated [4,5]. However, notwithstanding the elegance of the underlying chemistry, such multivalent synthetic ligands are likely to be expensive to produce on a large scale.

We have developed an alternative probiotic approach by engineering expression of host oligosaccharide receptor mimics on the surface of a harmless bacterium (reviewed in [6]). This involved manipulation of the outer core region of the lipopolysaccharide (LPS), the dominant surface antigen of Gram-negative bacteria. In the prototypic example targeted at STEC disease, we expressed two Neisseria galactosyltransferase genes (lgtC and lgtE) in a derivative of E. coli R1 with a truncated LPS core terminating in glucose (Glc). The exogenous transferases directed the addition of two galactose (Gal) residues to the Glc acceptor, generating a chimeric LPS terminating in Galα1–4Galβ1–4Glc-, an exact mimic of the Stx receptor Gb₃[7]. This presented a high-density array of receptor mimics on the bacterial surface, each capable of lateral diffusion in the fluid outer membrane to optimize docking with the Stx B pentamer. A dose of 1 mg dry weight of this recombinant probiotic could neutralize over 100 µg of purified toxin in vitro. Moreover, oral administration to mice infected with highly virulent STEC strains was 100% effective at preventing otherwise fatal disease [7].

The repertoire of oligosaccharide structures that can be mimicked on the surface of recombinant probiotics is limited only by the current state of knowledge of bacterial glycosyltransferases capable of modifying the LPS core. Already, a large number of such enzymes have been functionally characterized, and the explosion of publicly available bacterial genome sequences is facilitating selection and cloning of requisite transferase genes. Similarly, characterization of oligosaccharide targets for diverse toxins is being facilitated by access to modern glycomic technologies, such as glycan array analysis. We have already constructed additional probiotics expressing mimics of Gb₄ and lacto-N-neotetraose (LNT), using transferases also derived from the LPS core region of Neisseria spp. [8,9]. The Gb₄ mimic was targeted at a variant Stx produced by STEC strains responsible for edema disease in piglets, which prefers this receptor over Gb₃. The LNT mimic was targeted at ETEC LT, which is capable of binding to this structure as well as to GM1. Mimics of gangliosides GM1 (the preferred target of Ctx and LT) and GM2 were constructed using a combination of transferases derived from Neisseria and Campylobacter jejuni[9,10]. The Gb₄ and ganglioside mimics also required a UDP-Glc-4-epimerase gene, because the E. coli host is unable to synthesise the activated precursor for N-acetylgalactosamine, a component of these structures. The LNT mimic significantly reduced LT-induced fluid accumulation in a rabbit ileal loop model [9], while the GM1 mimic was highly protective against V. cholerae in an infant mouse cholera model [10]. Of course, there are numerous additional toxin-producing enteric pathogens that could be targeted with appropriate receptor-mimic probiotics. Foremost amongst these is Clostridium difficile, the commonest cause of nosocomial diarrhea and pseudomembranous colitis, although the identities of the human receptors for its key cytotoxins (TcdA and TcdB) are not yet known with certainty.

Practical considerations

As with any anti-infective strategy, receptor-mimic probiotics are likely to be most effective when administered early in the course of infection, necessitating prompt diagnosis. This is certainly achievable with STEC disease, as rapid and sensitive diagnostic methods are available, and there is a substantial lag between initial onset of gastrointestinal symptoms and development of the life-threatening sequelae, such as hemolytic uremic syndrome. Of course, prophylactic administration would be ideal, and this is conceivable in some circumstances. For example, travelers visiting regions where ETEC disease is endemic or where there is an existing cholera outbreak could commence prophylaxis on arrival. Similarly, in the case of STEC disease, close contacts of known cases are at high risk of acquiring infection, and in the outbreak setting it may be possible to identify persons exposed to a suspected contaminated food source who have not yet developed symptoms. Importantly, for both STEC disease and cholera, studies in animal models have still shown significant protection when receptor-mimic probiotic therapy was commenced well after infection had been established [10,11]. Thus, for cholera, patients are likely to benefit from inclusion of the Ctx-binding probiotic in oral rehydration formulations.

Although receptor-mimic probiotics are genetically modified organisms, the associated regulatory issues can be addressed, or circumvented, by the use of killed receptor-mimic constructs, which are also protective in animal models [11]. Frustratingly, the commercial sector has been reluctant to embrace this technology owing to perceived marketplace resistance to genetically modified organisms and concerns that oral administration of noninvasive bacteria that display mimics of certain host molecules might somehow break immune tolerance and trigger an autoimmune response (discussed in detail previously [6]). Notwithstanding these issues, the ongoing threat posed by the aforementioned pathogens, combined with the lack of effective vaccines and increasing problems due to antibiotic resistance, necessitates an open-minded approach. In addition to its efficacy in vivo, the receptor-mimic probiotic strategy has further important strengths. Firstly, the bacteria can be produced cheaply using large-scale fermentation. More importantly, unlike conventional antimicrobials, it is refractory to evolution of resistance. Any spontaneous mutation in the toxin sequence that diminished binding to the mimic construct would also interfere with recognition of the natural host receptor, and thereby attenuate the pathogen.

Financial & competing interests disclosure

The University of Adelaide owns patents covering the use of the receptor-mimic probiotics discussed in this article. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.