Pathogenesis of Helicobacter pylori Infection

Authors


Reprint requests to: Steffen Backert, UCD School of Biomolecular and Biomedical Sciences, University College Dublin, Science Center West, Belfield Campus, Dublin 4, Ireland. E-mail: Steffen.Backert@ucd.ie

Abstract

Helicobacter pylori infections and clinical outcome are dependent on sophisticated interactions between the bacteria and its host. Crucial bacterial factors associated with pathogenicity comprise a type IV secretion system encoded by the cag pathogenicity island, the effector protein CagA, the vacuolating cytotoxin (VacA), peptidoglycan, lipopolysaccharide (LPS), γ-glutamyl transpeptidase (GGT), protease HtrA, and the adhesins BabA, SabA, and others. The high number of these factors and allelic variation of the involved genes generates a highly complex scenario and reveals the difficulties in testing the contribution of each individual factor. Much effort has been put into identifying the molecular mechanisms associated with H. pylori-associated pathogenesis using human primary tissues, Mongolian gerbils, transgenic, knockout, and other mice as well as in vitro cell model systems. Interactions between bacterial factors and host signal transduction pathways seem to be critical for mediating the induction of pathogenic downstream processes and disease development. In this review article, we discuss the most recent progress in this research field.

Host Factors that Play a Role in Mediating Infection and Disease Development

Role of Mucosal Surface Molecules

Previous work has shown that MUC1, a membrane bound mucin expressed on the surface of gastric epithelial cells, provides a protective barrier against Helicobacter pylori. New studies tested the hypothesis that MUC1 counter regulates gastric inflammation in infections [1]. Infected Muc1−/− mice displayed increased TNFα and KC mRNA levels compared with uninfected mice, and down-regulation of MUC1 in AGS cells increased transcription factor NF-κB and IL-8 induction. It was shown that MUC1 forms a protein complex with IKKγ but not with IKKβ, thus preventing IKKβ–IKKγ interactions resulting in the inhibition of NF-κB [1]. Further studies investigated glycosylated structures present on secreted mucins in the stomach. Infected Mongolian gerbils exhibited increased expression of sialylated structures which enabled SabA-expressing strains to interact and promote colonization [2], similar to the observations in infected humans and Rhesus monkeys. H. pylori also interacts with the Lewisb blood antigen. A study in children showed fucosylated blood group antigens playing a role in mediating mucosal innate defense against H. pylori [3]. Lewisb expression on gastric mucin resulted in decreased bacterial colonization compared to infection in Lewisb-negative children, indicating that Lewisb acts as a molecular decoy by binding the organism on the mucin and limiting the number of bacteria available to interact with the epithelium [3].

Methylation, Tumor Suppressors, and Gene Polymorphisms

The gastric epithelium undergoes extensive epigenetic alterations during the development of gastritis induced by infection. MGMT, the gene encoding the DNA repair protein O-6-methylguanine methyltransferase, was found to be hypermethylated in H. pylori-positive patients, and this effect was partially reversible following bacterial eradication [4]. H. pylori also reduced MGMT expression and induced MGMT-mediated CpG methylation in AGS cells in vitro. DNA repair is disrupted during H. pylori gastritis, thus increasing mutagenesis in infected gastric mucosa [4]. While there is increasing evidence emerging to indicate that global hypermethylation occurs in H. pylori-infected gastric tissue and promotes gastric cancer, the role of global hypomethylation is less well defined. Another study showed that H. pylori infection induced hypomethylation of the repetitive elements Alu and Satα, in gastric mucosa of infected humans, is an early event during gastric carcinogenesis, and hypomethylation of Alu but not Satα persisted after eradication [5].

A number of studies have looked at the role of H. pylori in promoting suppression of tumor suppressor genes (TSGs). Trefoil factor 1 (TFF1) in the antral stomach acts as TSG, and Tff1−/− mice are prone to the development of gastric adenocarcinomas [6]. Mice treated with N-methyl-N-nitrosurea (MNU) in the absence of H. pylori exhibited widespread TFF1 repression, and in mice with advanced tumors, DNA methylation at the TFF1 promoter was observed. TFF1 was also repressed by H. felis infection but the repression was more marked in mice fed MNU following H. felis infection [6]. In human tissue, TFF1 was also found to be epigenetically repressed by H. pylori. These results suggest that H. pylori infection alone may not always be sufficient to induce gastric cancer and underlines the importance of other factors including diet and environment. Interestingly, this epigenetic silencing of TFF1 could be suppressed by the hormone gastrin [6]. As gastrin is an important regulator of gastric acid secretion and cell growth, H. pylori regulation of this hormone has been implicated in pathogenesis. H. pylori-infected mice have increased gastrin mRNA levels, and studies with AGS cells showed that infection induces gastrin through MAP kinases, but not NF-κB. Direct contact of live H. pylori with human cells was sufficient to induce gastrin gene expression [7]. Thus, modulation of the production of gastrin may have potential as an epigenetic modifier.

Expression of TFF2, another member of the trefoil factor family in the stomach, has recently been shown to also suppress tumor development, and the expression is lost during the progression of human intestinal type gastric cancer. Indeed, experimental H. pylori infection in mice reduced antral expression of TFF2 by increased promoter methylation. In human tissue samples, DNA methylation at the TFF2 promoter increased throughout gastric tumor progression [8]. The TSGs p53 and p27 can also be negatively regulated by H. pylori [9,10]. Using the gerbil model and infection in vitro showed that H. pylori activates AKT1 kinase which leads to phosphorylation and activation of HDM2 resulting in the degradation of p53 in gastric epithelial cells [9].

Gene polymorphisms involved in the inflammatory response also increase the risk of developing gastric cancer [11]. For example, polymorphisms in the IL-1β and endogenous IL-1 receptor (IL-1R) antagonist genes are known examples. A novel study has established for the first time the involvement of IL-1RI and Rho kinase in H. pylori-mediated disruption of tight junctional proteins in gastric epithelial cells in vitro [12]. H. pylori disrupted claudin-4 in a Rho kinase-dependent manner, and IL-1β mediated a similar effect. Further experiments revealed that inhibition of IL-1R activation prevented H. pylori-induced Rho kinase activation and claudin-4 disruption.

Inflammatory and Other Responses

In a study aimed at elucidation of the differential susceptibility to H. pylori that is found both across and within populations, it was shown that 5–6-week-old infected mice developed gastritis, gastric atrophy, epithelial metaplasia, and hyperplasia, while 7-day-old neonatal mice were protected from preneoplastic lesions [13]. This occurred in the neonatal mice because of the development of a biased ratio of T-regulatory to effector cells promoted by prolonged exposure to a low dose of antigen, suggesting that the age at which acquisition of infection occurs may play a role in mediating disease. Another study using mouse models that promote development of benign preneoplastic lesions investigated the effect of COX-2 inhibition and prostaglandin E2 (PGE2) treatment on the induction of Helicobacter-specific immune responses and formation of precursor cancer lesions. It found that systemic administration of PGE2 during the early stages of Helicobacter-induced gastric carcinogenesis was protective against gastric cancer by modulating the effect of Th1 effector T-cells [14]. Inhibition of COX-2 enzyme activity promoted gastritis and premalignant lesions. PGE2 conferred a protective effect which could be attributed to its immunosuppressive activity on CD4+ CD25 T-cells. In contrast, another study showed that in combination with the normal flora, PGE2 promoted the development of gastric cancer [15]. Re-colonization of germfree K19-Wnt1/C2mE mice (Gan mice) with indigenous bacteria or infection with H. felis and signaling induced by PGE2 induced the expression of CCL2, resulting in macrophage recruitment and gastric cancer development. Another study also suggests that enteric microbiota exacerbate H. pylori-initiated disease [16]. Transgenic Insulin-Gastrin (Ins-Gas) mice overexpress human gastrin that is associated with an increased risk of glandular atrophy and gastric cancer in humans. Ins-Gas infection results in the development of gastrointestinal intraepithelial neoplasia (GIN) which in turn is associated with achlorhydria and predisposes the mice to bacterial overgrowth. H. pylori infection of germfree InS-GaS mice resulted in delayed gastric lesion development and onset of GIN compared with infected specific pathogen-free mice [16]. Interestingly, the gender of the mice also played a role in the development of pathology as only 2/26 female and 8/18 male mice developed GIN at 11 months postinfection. Some of the sex differences were lost in older age suggesting that female hormones play a role in protection against the development of GIN. The importance of taking gender differences in animal models into account is highlighted by another study [17]. They used a C57Bl/6 gpt delta mouse model to analyse deletion mutations induced upon infection in a whole body system. In this model, deletions in phage-λ DNA integrated into the chromosome can be selected and subjected to molecular analysis. Infection induced significant increases in the frequency of point mutations in the gastric mucosa of female compared with male mice [17].

Bacterial Factors and Induction of Intracellular Signaling Pathways

The cag Pathogenicity Island

The H. pylori cag pathogenicity island (cagPAI) encodes a type IV secretion system (T4SS) which mediates the injection of CagA effector protein into host target cells [18,19]. In a new study, the cagPAI of 38 isolates from various geographic populations was sequenced [20]. Overall, the cagPAI gene content and order were conserved. Interestingly, several gene products predicted to be under Darwinian selection have been proposed to be novel injected effectors like CagA and include HP0522/Cagδ and HP0535/CagQ proteins [20]. A yeast two-hybrid approach was applied to identify interacting T4SS proteins. Novel interactions between the coupling protein HP0524/Cagβ and HP0526/CagZ were found [21]. CagZ seems to stabilize Cagβ, and interactions between Cagβ and CagA and between CagZ and Cagβ were also described, thus giving fresh insights into T4SS assembly [21].

Further studies demonstrated that purified HP0539/CagL mimics the host extracellular matrix protein fibronectin in vitro [22]. Upon contact with integrin α5β1 receptor of various human and mouse cell lines, purified CagL (like fibronectin) triggers cell spreading, focal adhesion formation, and activation of several tyrosine kinases including focal adhesion kinase (FAK), Src, and epidermal growth factor receptors EGFR and Her3/ErbB3. These findings suggest that CagL exhibits functional mimicry with fibronectin [22]. Investigation of how CagL activates EGFR revealed that docking dissociates metalloprotease ADAM17 from integrin α5β1, which activated HB-EGF production, and also repressed HKα promoter activity important in hypochlorhydria [23,24] Interestingly, CagL polymorphisms (Y58/E59) were described in gastric cancer patients from Taiwan to correlate with a corpus shift of integrin α5β1 leading to severe corpus gastritis and carcinogenesis [25]. Thus, CagL is a profound T4SS-factor with important roles in pathogenesis. Furthermore, new studies investigated the effects of H. pylori infection on histone modifications in AGS cells. Infection induced the dephosphorylation of histone H3 at serine residue 10 and other modifications [26]. The results demonstrate that histone alterations occur via cagPAI-dependent but cagA-independent mechanisms, which may contribute to transcriptional changes and pathogenesis [26].

CagA-dependent Effects

Studies reporting the effects of injected CagA on gp130-receptor-mediated signaling were evaluated. CagA, phosphatase SHP2, and gp130 were in complex, and phospho-CagA showed enhanced SHP2-binding activity and ERK1/2 phosphorylation, whereas nonphospho-CagA showed preferential STAT3 activation. These findings indicate that the phosphorylation status of CagA affects a switch between the SHP2/ERK and JAK/STAT3 pathways through gp130 [27]. In nonpolarized epithelial cells, ERK activation results in oncogenic stress, up-regulation of the p21 (Waf1/Cip1) cyclin-dependent kinase (Cdk) inhibitor, and induction of senescence [28]. In polarized epithelial cells, CagA-driven ERK signals prevent p21 (Waf1/Cip1) expression by activating a guanine nucleotide exchange factor-H1-RhoA-RhoA-associated kinase-c-Myc pathway. The microRNAs miR-17 and miR-20a, induced by c-Myc, are needed to suppress p21 (Waf1/Cip1) expression. CagA also drives an epithelial-mesenchymal transition in polarized epithelial cells which may be important in oncogenesis [28]. Another study identified the actin-binding protein cortactin as a novel downstream target of H. pylori-activated ERK kinase [29]. Upon infection, serine-phosphorylated cortactin was found to interact with and stimulate the kinase activity of FAK, suggesting that H. pylori may target cortactin to trap active FAK which results in disturbed cell adhesion turnover [29].

The mechanism of CagA delivery into host cells was also further investigated. Exposed CagA interacts with phosphatidylserine to initiate its entry into cells [30]. In addition, a novel CagA inhibitory domain at the N-terminus (amino acids 1–200) was identified using transfection constructs in epithelial cells [31]. This domain localizes to cell-cell contacts and increases cell-cell adhesion in epithelial cells [31]. Other new work showed that CagA can also be injected into dendritic cells (DCs) [32] and human B lymphoid cells [33]. While injected CagA suppresses host immune responses in DCs, it induces activation of ERK and p38 kinases in B cells and upregulates the expression of Bcl-2 and Bcl-X(L), which prevents apoptosis. Thus, CagA is directly delivered into B cells which may be associated with MALT lymphoma development [33]. Finally, another article highlighted that administration of d,l-α-difluoromethylornithine (DFMO) to mice reduces gastritis and bacterial colonization by inhibiting ornithine decarboxylase in macrophages and enhanced immune responses [34]. DFMO also inhibited the expression of CagA, and its translocation into AGS cells, which was associated with the reduced levels of IL-8, suggesting suppressive effects on the bacteria which may be useful in future therapies [34].

The Vacuolating Cytotoxin

Studies also continued focusing on vacuolating cytotoxin (VacA). A global phosphoproteome analysis of strain 26 695 was performed by mass spectrometry [35]. Eighty-two phosphopeptides from 67 proteins with 126 sites for serine/threonine/tyrosine phosphorylation were identified. Most interestingly, VacA was phosphorylated at serine-1244 and threonine-1245 residues. An interaction network was constructed centering on VacA, indicating that phosphorylation may regulate multiple aspects of metabolism and virulence [35]. It is well established that VacA p34 and p55 subunits enter host target cells by endocytosis. In a new study, p34 was shown to carry a unique import sequence for mitochondria. By forming an anion channel in the mitochondrial inner membrane, the toxin highjacks organellar functions [36]. Surprisingly, it was then shown that p55 is also involved. The colocalization of p34 and p55 subunits suggests that they could reassemble and form a pore in the inner mitochondrial membrane [37]. Another novel study showed that incubation of AZ-521 cells with purified VacA results in cell swelling, poly (ADP-ribose) polymerase (PARP) activation, decreased intracellular ATP concentration, and lactate dehydrogenase release. These features are consistent with the occurrence of cell death through a programmed necrosis pathway [38]. Investigation of gastric endoscopic biopsies from dyspeptic patients by immunocytochemistry showed that VacA and other factors accumulated in discrete novel 13-nm-thick cytoplasmic organelles [39]. Inside this structure, bacterial proteins colocalized with NOD1, ubiquitin-activating enzyme E1, polyubiquitinated proteins, phosphatase SHP2, and ERK kinases. These compartments differed from VacA-induced vacuoles, phagosomes, aggresomes, or related bodies and may be of potential pathologic relevance [39].

Peptidoglycan

The mechanism by which H. pylori generates helical shape is unknown. A novel study identified four genes being involved: peptidoglycan endopeptidases (csd1-3) and ccmA [40]. The findings suggest that the coordinated action of multiple proteins relaxes peptidoglycan crosslinking, enabling helical cell curvature and twist, which is required for robust bacterial colonization in the stomach [40]. Besides CagA that is delivered into host cells by an integrin-dependent pathway, the T4SS also injects peptidoglycan involved in innate immune responses. As α5β1 integrin is found in cholesterol-rich microdomains (lipid rafts), it was hypothesized that they may also induce pro-inflammatory responses mediated by NOD1 recognition of injected peptidoglycan [41]. Indeed, depletion of cholesterol and α5β1 integrin had significant inhibitory effects on peptidoglycan injection and NF-κB and IL-8 responses. These data implicate lipid rafts as novel platforms for T4SS-dependent delivery of bacterial products into target cells [41]. Further investigation defined the NOD1-dependent regulation of human beta-defensins in two epithelial cell lines [42]. This study also demonstrates the involvement of NOD1 and beta-defensin-2 in direct killing of H. pylori by epithelial cells, thus confirming the importance of NOD1 in host defense [42]. An unexpected novel NOD1-dependent cascade in human epithelial cell lines is activated and results in a RICK→TRAF3→TBK1→IKK→IRF7 pathway leading to the synthesis of type-I interferon, but not NF-κB activation [43]. Thus, more studies are necessary to solve some discrepancies between previously published articles [44].

Lipopolysaccharide

Two new enzymes typically involved in lipopolysaccharide (LPS) biosynthesis were discovered in H. pylori, glycosyltransferase WecA and O-antigen ligase WaaL, but a flippase enzyme normally involved in O-antigen synthesis could not be detected [45]. Instead, H. pylori uses a translocase named Wzk in a novel LPS biosynthetic pathway, evolutionarily connected to protein N-glycosylation [45]. In addition, 3-deoxy-d-manno-octulosonic acid (Kdo) hydrolase genes (HP0579 and HP0580) were shown to affect the expression of O-antigen Lewis epitopes [46]. Finally, an ADP-l-glycero-d-manno-heptose-6-epimerase ortholog (HP0859) was characterized [47]. ΔHP0859 mutants exhibited severe truncation of LPS, decreased growth rate, higher susceptibility to novobiocin, and failed to induce the AGS elongation phenotype. Genetic complementation restored these phenotypes, revealing HP0859 essential for LPS biosynthesis and virulence [47].

Toll-like receptors (TLRs) recognize various microbial products, resulting in the initiation of innate immunity. Accumulating data suggest that H. pylori LPS is not recognized by classical TLR-4, but TLR-2. Experimental evidence demonstrated that LPS functions as a TLR-2 ligand by signaling through pathways involving MyD88, IRAK1, IRAK4, TNFR-associated factor 6, IκB kinase-β, and IκBα [48]. Infection of gastric epithelial cells was associated with the decreased expression of signaling factor tribbles-3 (TRIB3), and knockdown of TRIB3 and C/EBP homologous protein enhanced TLR2-mediated NF-κB activation and chemokine induction by LPS. Thus, modulation of TRIB3 may be an important mechanism during H. pylori-associated pathogenesis downstream of TLR2 [48]. In addition, using two colon carcinoma cell lines, it was observed that LPS upregulates the expression of inducible nitric oxide (NO), demonstrating its ability to interfere with the DNA repair machinery and increasing risk of genotoxic effects [49]. Finally, LPS from H. pylori increased the paracellular permeability of cultured gastric cells [50]. Such an effect in vivo would have an important impact on epithelial barrier functions and pathology.

Outer Membrane Vesicles and Other Bacterial Factors

H. pylori continuously buds-off outer membrane vesicles (OMVs) from its surface. Purified OMVs revealed their major protein and phospholipid components and some virulence factors [51]. Additional functional and biochemical analyses focused on BabA and SabA adhesins and their respective interactions with the gastric epithelium. Thus, OMVs carry effector-promoting properties which may be important for disease development [51]. However, the mechanism of OMV uptake in host cells is poorly understood. Using inhibitors and mutants, a new report has shown that VacA enhances the association of OMVs with cells and that clathrin-mediated endocytosis is involved, while vesicle internalization did not require cholesterol in this study [52].

γ-Glutamyl transpeptidase (GGT) has been reported as a pathogenicity factor associated with H. pylori colonization and cell apoptosis. A new study showed that purified GGT inhibits the growth of AGS cells and that caspase-3 inhibitors effectively blocked GGT-induced apoptosis [53]. Cell cycle analysis showed G1 phase arrest and apoptosis following GGT treatment, and this was associated with down-regulation of cyclin-E, cyclin-A, Cdk-4, and Cdk-6 and the upregulation of the Cdk inhibitors p27 and p21 [53]. In addition, recombinant GGT, infection with wild-type but not isogenic GGT mutants generated H2O2 in primary gastric epithelial and AGS cells, resulting in the activation of NF-κB and up-regulation of IL-8 [54]. The clinical importance was shown by significantly higher GGT activity in strains obtained from patients with peptic ulcer disease (PUD) than isolates from nonulcer dyspepsia [54]. Another pathogenicity-associated factor is the duodenal ulcer-promoting gene A (dupA). The dupA locus of 34 strains was sequenced. Most dupA alleles were longer (1884 bp; dupA1) than previously described, although some had truncated versions (dupA2) [55]. Interestingly, dupA1 (but not dupA2 or the cagPAI) substantially increased H. pylori-induced IL-12p40 and IL-12p70 production from CD14+ mononuclear cells. This indicates that while virulent strains cause inflammation by stimulating epithelial cells through cagPAI-delivered products [44], mononuclear inflammatory cells are stimulated through dupA1 products [55]. Finally, an entirely new function of protease HtrA as a secreted virulence factor was found. HtrA cleaves-off the ectodomain of the TSG and cell adhesion protein E-cadherin in vivo and in vitro [56]. E-cadherin shedding disrupts epithelial barrier functions and possibly pathogenesis allowing access of bacteria to the intercellular space. A small-molecule inhibitor that efficiently blocks HtrA activity was generated, and blocked E-cadherin cleavage and intercellular invasion of H. pylori [56]. Future studies will show whether HtrA is of general importance to H. pylori or only present/active in a subset of disease-specific strains.

Conflicts of Interest

The authors have declared no conflicts of interest.

Ancillary