The human stomach has long been regarded as a biological sanctuary, and although curved bacilli had occasionally been visualized adjacent to the gastric epithelium, their relationship to disease remained obscure until Marshall and Warren's seminal report1 relating the presence of these organisms, Helicobacter pylori, to peptic ulceration. Although H. pylori-induced gastritis clearly increases the risk for peptic ulceration, distal gastric adenocarcinoma and gastric mucosal lymphoproliferative disease, only a fraction of colonized persons ever develop clinical sequelae2. Pathologic outcomes likely involve choreographed interactions between H. pylori virulence determinants and host constituents. On page 375, Akihiro Fujikawa and colleagues3 describe a novel mechanism through which a specific H. pylori disease-related component, VacA, may induce ulcer formation by binding and activating the receptor protein tyrosine phophatase, Ptprz, expressed on gastric epithelial cells. Engagement of Ptprz by VacA leads to cellular detachment, which may heighten the risk for ulcerogenesis by exposing denuded mucosa to noxious gastric contents.

Varieties of VacA

H. pylori populations are extremely genetically diverse, which may engender differential host responses that influence clinical outcome. A specific locus of variability is vacA, which encodes a vacuolating cytotoxin. The gene vacA is present in virtually all H. pylori isolates4,5, but strains differ in vacuolating activity owing to variations in vacA gene structure. The regions of greatest diversity are localized near the N terminus of VacA (allele families s1a, s1b, s1c or s2) and the mid-region (allele families m1 or m2; ref. 5). Most s1 VacA toxins possess vacuolating activity in vitro, whereas s2 VacA proteins possess little if any cytotoxic activity. Strains containing s1 alleles are more commonly isolated from individuals with ulcer disease than from individuals with gastritis alone5, underscoring the importance of vacA as a bacterial locus related to high-grade host responses within the gastric niche.

After export, VacA oligomerizes into rosettes that have minimal vacuolating activity6. Acid treatment dissociates VacA oligomers into monomers, however, exposing critical hydrophobic regions, which then mediate its insertion into cellular membranes where it forms anion-selective channels (refs. 7,8; see figure). Cytoplasmic internalization of VacA is regulated by active cellular processes9, and as vacuoles contain membrane markers of late endosomes and lysosomes, vacuole biogenesis seems to occur at the level of late endosomes (see figure).

Gut check. Working model of cellular alterations induced by the H. pylori VacA toxin that lead to ulcer formation. a, VacA alters cellular tight junctions and enhances paracellular permeability to iron, nickel and other organic molecules, which may provide essential nutrients required for H. pylori growth in the gastric niche. VacA also inserts into the plasma membrane, where it forms an anion-selective channel that can transport urea from the cell cytosol, which, in turn, acts as a substrate for the generation of ammonia by H. pylori urease. VacA channels are then endocytosed and incorporated into endosomes, and then they form vacuoles. In a separate pathway, VacA binds to a specific receptor-type protein tyrosine phosphatase, Ptprz, and activates Git1 by an as yet unidentified mechanism. Git1 integrates multiple intracellular signals that regulate membrane trafficking, organelle structure, actin cytoskeletal changes and cellular adhesion. b, Engagement of Ptprz by VacA leads to epithelial cell detachment from the underlying basement membrane, thereby rendering the lamina propria vulnerable to the damaging effects of acid in the stomach. c, Prolonged exposure of denuded and inflamed gastric mucosa to acid ultimately leads to peptic ulceration.

VacA also exerts other biological effects that may influence clinical outcome. Inoculation of mice with either purified VacA or broth filtrates containing VacA leads to epithelial cell injury10,11, and in vitro, VacA induces gastric epithelial cell apoptosis12. VacA functions as a transmembrane pore, permeabilizing host cells to urea13, which may allow H. pylori to manipulate gastric pH by generating ammonia (see figure). When added to polarized monolayers, VacA increases paracellular permeability to organic molecules, iron, and nickel (ref. 14; see figure). Collectively, these data indicate that VacA can induce multiple physiologic consequences that may contribute to pathogenesis.

Targeting a phosphatase

VacA has been reported to bind to a variety of high-affinity cell-surface receptors, including a receptor-type protein tyrosine phosphatase, Ptprz (also known as PTPζ and RPTPβ; ref. 15). Protein tyrosine phosphatases constitute a diverse family of cytoplasmic and transmembrane receptor–like enzymes that regulate cellular proliferation, differentiation and adhesion. Specific ligands for Ptprz have been well characterized, and binding of one of these, pleiotrophin, mediates cellular attachment and migration.

Git1 (G protein–coupled receptor kinase-interactor 1, also called Cat-1, Cool-associated, tyrosine-phosphorylated 1) has recently been identified as a potential substrate of Ptprz, and stimulation of neuroblastoma cells with pleiotrophin paradoxically increases phosphorylation of Git1. As pleiotrophin also increases phosphorylation of β-catenin, Ptprz substrates are probably under constitutive negative regulation, and specific ligands may reduce the catalytic activity of Ptprz. Git1 regulates ADP-ribosylation factor GTPases that mediate membrane trafficking, actin cytoskeletal changes and the organization of focal adhesion complexes. As Ptprz regulates cellular phenotypes (for example, adhesion) that may contribute to mucosal damage and may represent a receptor for VacA, Fujikawa et al.3 investigated the role of VacA–Ptprz interactions in gastric injury using complementary in vivo and ex vivo genetic models of Ptprz deficiency.

Having previously shown that Ptprz is primarily expressed in the brain, the authors first sought to establish its presence in gastric tissue. Virtually all glands in the gastric corpus of wild-type mice contained detectable Ptprz and expression was localized to the glandular basal region, whereas no staining was present in gastric tissue harvested from Ptprz-deficient mice. The ability of VacA to bind to gastric tissue was then determined by passing gastric mucosal homogenates across a VacA-laden chip. The binding capacity of extracts from Ptprz-deficient mice was 30% less than that from wild-type mice, and these differences resolved in the presence of antagonistic Ptprz antibodies. The ability of VacA to associate with Ptprz was also shown by immunoprecipitation experiments in which Ptprz-containing beads were co-incubated with purified VacA. Resolution with antibodies against Ptprz and against VacA revealed immunoreactive bands, indicating that VacA binds to Ptprz in vitro and in vivo but that Ptprz only represents a fraction of the total VacA-binding sites in gastric epithelium.

How to cause an ulcer

Fueled by these results and previous data showing the ability of VacA per se to induce mucosal injury10,11, Fujikawa et al.3 next delivered purified VacA by gavage to wild-type and Ptprz-deficient mice. VacA was detected in the cytoplasm of a variety of epithelial cells from both genotypes and there were no differences in VacA distribution, supporting the conclusion that multiple VacA receptors are present in gastric tissue. In contrast, VacA induced a dose-dependent increase in gastric injury only in wild-type mice, with most developing severe gastric hemorrhage and ulcers when challenged with the highest dose of VacA (500 μg per kg of body weight).

To investigate the role of Ptprz in VacA-induced injury at a molecular level, primary gastric epithelial cells from wild-type and Ptprz-deficient mice were co-incubated with VacA in vitro. VacA was incorporated equally into wild-type and Ptprz-deficient cells, which mirrored the in vivo pattern of VacA distribution. When primary cells grown on a reconstituted basement membrane were exposed to VacA, however, only wild-type cells detached. The functional consequences of VacA–Ptprz binding were further explored by transfecting BHK-21 cells that lack Ptprz but contain its substrate Git1 with Ptprz or a phosphatase-inactive Ptprz construct. VacA induced phosphorylation of Git1 in cells transfected with functional Ptprz but not in cells transfected with defective Ptprz or with vector, recapitulating events that occur after treatment with pleiotrophin.

Fujikawa et al.3 then returned to their in vivo model and treated mice with oral doses of pleiotrophin. Wild-type, but not Ptprz-deficient, mice developed severe gastritis and ulcers after treatment with pleiotrophin, and there was no evidence of vacuolation. The authors then extended these findings by showing that Ptprz is expressed in human gastric tissue in a pattern similar to that observed in mice, which strengthens the biological relevance of this model to H. pylori-induced ulcerogenesis in humans.

Models and mechanisms

The results from this study invoke a model in which VacA binds specifically to Ptprz, leading to cellular detachment through modification of the phosphorylation patterns of cellular proteins, a pathway that is distinct from vacuole biogenesis (see figure). Although these experiments have provided fresh insights regarding determinants that may influence ulcer formation, there are new questions and hypotheses to be explored. As vacuolation and cellular responses induced by the binding of Ptprz seem to be independent events, do non- or minimally-toxigenic VacA molecules bind Ptprz or induce cellular detachment? What specific region(s) of VacA are required for Ptprz binding and activation? Do additional H. pylori constituents contribute to ulcer formation?

Despite these questions, mechanistic studies such as this are extremely valuable, not only because of the significance of H. pylori as a human pathogen, but also because they may facilitate translation of developments from the laboratory to the clinical setting. For example, understanding the role of specific H. pylori virulence determinants in the development of peptic ulceration could contribute to vaccine development. Understanding the mechanisms through which specific bacterial factors interact with host pathways may allow identification of infected individuals at high risk for clinical disease. Finally, understanding the pathogenesis of one well defined cause of microbially induced disease could lead to its use as a model system for other forms of inflammation and injury that develop within the context of bacterial infections.