Pathogenesis of Helicobacter pylori Infection

Authors


Reprint requests to: Daniela Basso, Department of Laboratory Medicine, University-Hospital of Padova, Via Giustiniani 2, 35128 Padova, Italy. E-mail: daniela.basso@sanita.padova.it

Abstract

Helicobacter pylori infections are thought to eventually lead to symptoms as a result of the long-lasting interactions between the bacterium and its host. Mechanisms that allow this bacterium to cause a life-long infection involve modulation of both the immune response and host cellular processes. Last year many novel findings that improve our knowledge on how H. pylori virulence factors interact with the host were reported, but because of space limitations we can only discuss a limited number of these studies. Among those are studies on the genetic variation of genes encoding outer membrane proteins and the mimicry of host antigens, factors that alter host-cell metabolism and factors that modulate the host’s immune response.

While chronic Helicobacter pylori infection is usually without any symptoms, disease ranges from peptic ulcer, gastric adenocarcinoma to gastric MALT lymphoma. Although the clinical outcome of the infection is thought to be determined by host, bacterial and environmental factors, the focus of this review is on recent findings relevant to H. pylori adaptation and virulence factors.

Helicobacter pylori Adaptation

H. pylori has developed several strategies that allow it to perfectly adapt to the gastric mucosa of its human host, its only known natural niche. Bacterial mimicry and genetic diversification represent successful strategies employed by the bacterium to evade the immune response and to survive in the human stomach throughout the life of its host. Genetic diversification may help H. pylori to adapt to a new host after transmission, and to different micro-niches within a single host and to changing conditions in the host over time. Genetic diversity arises from within-genome diversification and from integration of DNA from other H. pylori strains. Central to this is the ability of H. pylori to take up exogenous DNA and incorporate it into its genome. The H. pylori machinery for exogenous dsDNA uptake is composed of the type-IV secretion system ComB, which transports dsDNA across the outer membrane at the cell poles, and by ComEC, which mediates the subsequent transport into the cytoplasm through the inner membrane with higher specificity for DNA structure [1]. Adaptation to varying gastric conditions is enabled by several H. pylori genes that display phase variation; these include genes encoding outer membrane proteins (OMPs), like BabA which is a Lewis b ligand, and genes involved in lipopolysaccharide (LPS) biosynthesis. In animal model systems of H. pylori infection, Styer et al. [2] provided evidence that BabA expression is lost during persistent infection by phase variation and nonreciprocal gene conversion of babA with a duplicate copy of babB, a paralog of babA with unknown function. H. pylori not only binds to human Lewis antigens but also expresses Lewis antigens (H. pylori is a fucose expressing pathogen). The variable O-antigen chain part of the H. pylori LPS is uniquely composed of host-related Lewis antigens and this host-cell surface mimicry is thought to facilitate immune escape. Two studies [3,4] explored phenotype variation of H. pylori Lewis antigen expression. Both studies employed a mouse infection model to demonstrate that bacterial subpopulations expressing both Lewis x and Lewis y coexist and are stable during persistent infection. New subpopulations expressing Lewis b inevitably appear when Lewis b transgenic mice are infected. This finding supports the hypothesis of an increased fitness of H. pylori variants that match the Lewis phenotype of their host [4]. Changes in Lewis phenotypes could be linked to phase variation of the metastable poly-C tracts of the galactosyltransferase gene encoding β-(1,3)galT and the fucosyltransferase-encoding genes futA, futB and futC that are all involved in Lewis antigen biosynthesis. Skoglund et al. [3] demonstrated that a neutral pH favors Lewis y expression, while a more acidic pH favors a switch from solely Lewis y to both Lewis x and Lewis y glycosylation. In agreement with the above findings, Lehours et al. [5] demonstrated an increased prevalence of Lewis x negative/Lewis y positive strains among the cagPAI negative isolates form patients with MALT lymphoma versus patients with gastritis, possibly representing the result of H. pylori adaptation through futA and futB phase variation. Next to Lewis antigens, H. pylori can also bind to host syndecans, a family of transmembrane heparin sulfate proteoglycans. Magalhães et al. [6] found that H. pylori infection results in an increased expression of syndecan-4 when infections are sustained by CagA positive strains and provided in vitro evidence that CagA, but mainly CagE, is required to induce membrane-bound syndecan-4 expression. A global overview of the complexity of H. pylori OMPs gene expression was carried out by Odenbreit et al. [7] who studied the expression of eight OMPs (AlpA, AlpB, SabA, BabA, BabB, BabC, HP0227, OipA) in H. pylori isolates obtained from the antrum of infected children. The hypothesis that, to adapt to a changing niche, H. pylori genome co-evolves with host response has been pursued by Giannakis et al. [8]. In a genome-wide analysis, H. pylori isolates obtained from the gastric corpus of multiple patients with variable gastric pathology the authors found that isolates differed markedly between patients, but the H. pylori population within an individual was largely clonal and remained stable over a period of at least 4 years. By analyzing the transcriptome of infected gastric epithelial progenitors, the authors identified Serpin-1, several protein tyrosine phosphatases and superoxide dismutase 2 among the highest upregulated genes, and Cdkn2c, a tumor suppressor gene, among the most strongly downregulated genes in H. pylori-infected patients with chronic atrophic gastritis and with gastric cancer, but failed to find any new disease-associated gene. A genome-wide map of H. pylori transcriptional start sites (TSS) and operons was provided by Sharma et al. [9], which complemented genomic sequence and global protein–protein interaction map of the H. pylori strain 26695. Uncoupling of polycistrons and genome-wide antisense transcription (27% of the primary TSS are also antisense TSS) contribute to the high complexity of H. pylori gene expression. Antisense TSS for 22/34 putative phase variable genes involved in LPS biosynthesis, surface structure and DNA restriction/modification were identified and this might well represent a new mechanism of controlling surface structure variations and host interactions.

Intracellular Pathways Activated by Helicobacter pylori

Helicobacter pylori and Apoptosis

An increased proliferation not balanced by an increase in apoptosis has been postulated as a putative cause of H. pylori-associated gastric carcinogenesis. Yan et al. [10] demonstrated both in vitro and in an in vivo mouse model that a rodent adapted H. pylori cag-positive strain activates the epidermal growth factor receptor (EGFR) through the ADAM-17 mediated release of heparin-binding-EGF. EGFR activation in epithelial cells resulted in activation of Akt, decreased Bax expression and increased Bcl-2 expression, the downstream targets that promote an anti-apoptotic response in H. pylori-infected epithelial cells. The sonic hedgehog (Shh) expression in gastric adenocarcinoma samples from both mice and humans was induced by H. pylori in a time-dependent manner. Intriguingly, Shh overexpression in AGS gastric cancer cells was shown to protect these cells from H. pylori-induced apoptosis [11]. By contrast, a pro-apoptotic in vitro effect was obtained using a human CagA+ VacA+ strain, which induced Bax, decreased Bcl-2 and activated NF-kB [12].

Helicobacter pylori and STAT Signaling

Sox2 represents a crucial transcription factor for the maintenance of embryonic stem cell pluripotency and organ development and differentiation of e.g. lung and stomach. Asonuma et al. [13] provided both experimental and clinical evidence that the H. pylori induced IFN-γ results in downregulation of Sox2 on IL-4/STAT6 signaling. This interferes with the formation of oxyntic and pyloric glands, which might lead to precancerous gastric atrophy and intestinal metaplasia.

Helicobacter pylori and c-Met

Upon H. pylori infection, the hepatocyte growth factor receptor c-Met sheds from the surface of epithelial cells [14]. In addition to shedding, c-Met undergoes phosphorylation and associates with non-T-cell activation linker, lymphocyte-specific protein tyrosine kinase-interacting membrane protein and the SH2 domain of growth factor receptor-bound protein 2 (Grb2), thus activating the ERK signaling cascade [15].

CagA and VacA

The best described H. pylori virulence factors with respect to intracellular interaction are CagA and VacA. Their known [16–18] and recently discovered effects are summarized in Table 1. East Asian CagA was confirmed to be more oncogenic than Western CagA in transgenic mice models [19] and the number of EPIYA-C motifs of Western type CagA was confirmed to enhance premalignant lesions and gastric cancer risk in vivo, and to correlate with the degree of CagA phosphorylation and with the magnitude of cellular morphological alterations in vitro [20,21]. In an elegant study, Umeda et al. [22] provided experimental evidence for the direct role of CagA in chromosomal instability. They showed that CagA binds to and inhibits the partitioning-defective 1 (PAR1)/microtubule affinity-regulating kinase (MARK), a master regulator of cell polarity. This results in a delayed progression from prophase to anaphase. During mitosis, cells exposed for 12 hours to CagA showed spindle misorientation and perturbed cell division axis, while prolonged CagA exposure (up to 5 days) caused a reduction of the number of cells in G1 phase, an enhancement of cells in G2/M phase and a dramatic increase in polyploidy cells. CagA binds and inhibits other PAR1 isoforms that are involved in the maintenance of tight junctions [23]; this leads to a stabilization of the microtubules and contributes to the hummingbird phenotype. The CagA–PAR1 interaction is mediated by the C-terminal 16 amino acid stretch of CagA, termed CagA-multimerization sequence and by the 27 amino acid stretch present in the C-terminal of the PAR1 domain. CagA–PAR1 complex formation causes PAR1 kinase inhibition, but it also increases CagA stability within epithelial cells [24]. CagA residues 951–956, critically involved in PAR1/MARK interaction, were shown to possess a main chain conformation that mimics host substrates inhibitors of PAR1/MARK family kinases [25]. Intriguingly the presence of Gly955 (found in East Asian strains) allows a more potent inhibition of PAR1/MARK than the presence of Lys955, found in the lesser carcinogenetic Western CagA strains. Lamb et al. [26] demonstrated that CagA associates intracellularly with both TRAF6 and TAK1 (transforming growth factor β activated kinase 1), enhances TRAF6-mediated Lys63-linked ubiquitination of TAK1 which, in turn, promotes IKK activation followed by IkBα phosphorylation and degradation, nuclear translocation of NF-kB and activation of its target genes IL-8 and TNF-α. By targeting NF-kB with the inhibitor caffeic acid phenetyl ester (CAPE), a suppression of inflammatory infiltration and of inflammatory mediators release in the gastric mucosa of H. pylori infected gerbils was shown [27]. Using isogenic strains with identical CagA proteins differing only by the presence of a single EPIYA-C motif, Lee et al. [28] demonstrated that both EPIYA-C− and EPIYA-C+ CagA preferentially activate gp130/STAT3 and SHP2/ERK signaling pathways respectively, indicating a novel role for EPIYA-C negative CagA. Phosphorylated CagA blocks the tyrosine kinase activity of the Src kinase family (SKF) proteins that are critically involved in the control of intracellular VacA trafficking. As a result, VacA remains confined in the inner cell periphery and it does not move to its target intracellular compartments [29]. Unphosphorylated forms of CagA were suggested to counteract VacA-induced apoptosis by directly blocking the intrinsic apoptotic pathway [29]. The VacA pro-apoptotic action was demonstrated to be enhanced by ammonia [30]. Data regarding other known and new potential virulence factors are summarized in Table 1, which reports also data on urease activation and metal ion responsive proteins [31–44] that because of space limitations cannot be discussed in detail.

Table 1.   Known and novel facts on Helicobacter pylori virulence determinants and disease markers
Known actions [16–18]New findings
(I) Adhesion molecules
BabA and SabA. BabA recognizes Lewis b and ABH, SabA sialyl-Lewis x/a. Two babA alleles (babA1 and babA2) and one homologous babB. Only babA2 is functionally active and differs from babB mainly in the central region, which determines binding specificity. babA and sabA expression might be modulated (on-off switch) through recombination (gene conversion) or by slipped strand mispairing (phase variation) base on the number of CT repeats in 5′A BabA allele not binding to Lewis b differs from BabA in six amino acids residues in the highly polymorphic mid region suggesting that this region is critical for binding to the fucosylated ABH antigens [2]
(II) cag PAI AND VacA
CagE. Secretory protein that is required for the induction of IL-8 and for translocation and phosphorylation of the CagA proteinInduces syndecan-4 expression [6]
CagA. Polymorphic effector protein with four distinct EPIYA regions (EPIYA-A, -B, -C and -D). Western strains typically contains EPIYA-A, -B, and one to three repeated EPIYA-C. East-Asian strains contains EPIYA-A, -B and usually one -D region• Induces syndecan-4 expression, but less than CagE [6]
• Perturbs cell division axis [22]
• Initiates TAK1 ubiquitination by TRAF6 inducing IKK and NF-kB [26]
CagA Phosphorylation-dependent effects. EPIYA tyrosine phosphorylation is mediated by SRC kinases and continued by c-ABL kinases. Phospho-CagA interacts with SRC homology (SH2) domain containing host cell proteins, including tyrosine phosphatase SHP-2 (binds to phospho EPIYA-C and -D), the C-terminal Src tyrosine kinase (CSK, binds to phospho EPIYA-A and -B) and the adaptor protein CRK. Activation of CRK induces several downstream signaling pathways resulting in cell scattering and cell-cell dissociation. CagA activated SHP-2 dephosphorylating FAK causes downregulation of FAK kinase and activation of ERK MAP kinases. This results in cytoskeletal reorganization, cell elongation (hummingbird phenotype) and cell cycle progression• East Asian CagA more carcinogenetic than Western CagA in transgenic mice models [19]
• The number of EPIYA-C confirmed to enhance gastric cancer risk [20,21]
• Phospho CagA inhibits VacA trafficking by blocking SKF kinase [29]
CagA Phosphorylation–independent effects. Interaction with host cell proteins including tight junction scaffolding protein zonulin (ZO-1), E-cadherin, c-Met, b-catenin, adaptor protein GRB-2, PAR1 kinase with the consequent disruption of tight and adherent junctions, loss of cell polarity, pro-inflammatory and mitogenic effects. Interaction with GRB-2 and activation of RAS/MEK/ERK pathway with consequent cell scattering and proliferation• Inhibits PAR1 isoforms and stabilizes microtubules [23]
• Interaction with PAR1 confers stability to CagA within epithelial cells [24]
• EPIYA-C negative CagA activates gp130/STAT3 pathway and STAT3 activation facilitates gastric epithelial wound healing [28]
VacA. Made of two subunits p33 and p55, is encoded by the polymorphic vacA gene present in all strains. The vacA gene variability among strains seems to be predominantly limited to N-terminus and mid-region: the 5′ region encodes the signal peptide and N-terminus of the protein (s1 and s2 alleles), the intermediate region (i1 and i2 alleles) encodes part of the p33; the mid region (m1 and m2 alleles) encodes part of the p55 epithelial cell–binding subunit. The s1/i1/m1 VacA is fully active, the s2/i2/m2 is inactive; other forms have intermediate activity. VacA induces cytoplasmic vacuolization in cultured cells by disrupting endosomal maturation, increases permeability of polarized epithelial cells, induces cytochrome c release and apoptosis, interferes with phagocytosis and antigen presentation, inhibits T cell proliferation, inhibits NFAT• vacA s1m1 and cagA independently associates with elevated levels of promoter methylation of RPRM, a TP53-dependent mediator of the G2/M cell cycle check point, and of TWIST1, a basic-helix-loop-helix transcription factor that regulates metastases [21]
• Pro-apoptotic action of VacA is counteracted by CagA [29] and increased by ammonia [30]
• Amino acids 484–504, 511–536, 517–544 dispensable for vacuolating toxin activity. Amino acids 559–628 dispensable for protein folding and secretion [31]
(III) Disease markers
homA and homB. homB gene associates with increased risk of peptic ulcer disease. homA gene is a paralog of homB and associates with non ulcer dyspepsia. Both homA and homB can be found as a single or double copy in the H. pylori genomeDifferent distribution among East Asian and Western strains: in East Asian strains homB is more frequent than homA with no association with disease; in Western strains homB associates with cagA and vacA s1 [32]. Regulation by phase variation [33]
Plasticity zones (PZ). Genome region with a great genetic variability, with a G+C content lower (about 35%) than H. pylori chromosome overall (39%)PZ are transposable (Tn) elements which might affect bacterial phenotype and fitness. Full length TnPZ contains, among others, a cluster of type IV protein secretion genes (tfs3), distinct from the other two type IV secretion systems (cag PAI and comB locus) [34]
Heat and protease-sensitive H. pylori factor different from CagA and VacAActivates the intrinsic poly(ADP-ribosylation activity of the nuclear factor poly(ADP-ribose) polymerase-1 (PARP-1) which undergoes automodification with polymers of poly(ADP-ribose)(PAR). This modulates cell signaling and might enhance disease risk [35]
Coiled coil rich proteins (Ccrp) and actin like MreB proteinCcrp form filamentous structures and affect cell morphology; Ccrp genes shows high degree of sequence variation. MreB is essential for urease activity [36]
(IV) Urease
Crucial enzyme for acid resistance. The structural subunits UreA and UreB form a complex that is activated by Ni2+ insertion, aided by the protein pairs UreE/UreG and UreF/UreH. The pH-gated urea channel UreI regulates urease activityCytoplasmic histidine kinases HP0165 and HP0244 are acidic pH sensors. Active HP0244 allows assembly and activation of urease [37]. SlyD participates in Ni2+-dependent urease assembly [38]
UreE binds Ni2+ or Zn2+. Zn2+ stabilizes UreE/UreG [39]
(V) Metal ion responsive proteins
nikR. Ni2+ responsive transcription factor activator of urease expressionnikR binds tightly with genes involved in Ni2+ uptake and processing, including urease [40], and downregulates fecA3 [41]
HspA and HypA. Ni2+-binding chaperonesHspA binds Ni2+ in vivo, protects H. pylori against Ni2+ overload, allows maturation of hydrogenase [42]. HypA binds to both Zn2+ and Ni2+ [43]
Fur. Iron-dependent ferric uptake regulatorFur was biochemically characterized [44] and demonstrated to regulate transcription of fecA1 and fecA2 (putative outer membrane proteins involved in iron transport) [41]

Helicobacter pylori Induced Chromosomal Instability, Genetic and Epigenetic Alterations of the Host Cell

The interactions between H. pylori virulence determinants and host epithelial cells induce genetic, epigenetic and chromosomal alterations in the host genetic material. It results in a continuous patching of the genetic information of the host cells, which favors the development of gastric carcinoma. Genomic instability of the host genome following H. pylori infection was investigated by Machado et al. [45], who demonstrated in vitro that gastric adenocarcinoma AGS cells infected with H. pylori for 5 days have reduced levels of mismatch repair enzymes, the same finding being recorded in vivo in H. pylori-infected mice. In cancer cells, repetitive element lose their methylation and CpG islands of many promoters become aberrantly methylated. Promoter hypermethylation of tumor suppressor genes is considered an important factor in carcinogenesis and known to be present in H. pylori associated gastric tumors. Park et al. [46] showed that this aberrant methylation is already present in the premalignant stages of gastric cancer. Work by Schneider et al. [21] showed that the degree of aberrant methylation is associated with the presence/absence of known virulence factors in the infecting H. pylori strain. Of particular interest are the results regarding runx3 promoter methylation, which were described by Park et al. [46] in intestinal metaplasia and confirmed by Katayama et al. [47] who showed runx3 promoter methylation occurs in gastric epithelial cells co-cultured with macrophages exposed to live H. pylori. Among the epigenetic alterations following H. pylori infection, deregulation of microRNAs (miRs) expression might also be relevant for pathogenesis. miRs are non coding small RNAs which control mRNA translation and they frequently are deregulated in human cancers. Ando et al. [48] studied the methylation status of a series of miRs in a series of gastric cancer cell lines, in primary gastric cancers, and in gastric mucosa from patients with or without H. pylori infection, and provided evidence that H. pylori infection is associated with higher methylation of miR-124. Gao et al. [49] demonstrated a reduction of miR-218 in gastric cancer tissue, but also a putative amplification of this reduction by H. pylori infection. In vitro experiments with overexpression or silencing of miR-218 allowed the authors to demonstrate that miR-218 induces apoptosis and decreases cell proliferation by promoting ECOP (epidermal growth factor receptor coamplified and overexpressed protein) degradation, which decreases NF-kB activation. Interference with these miR methylations might provide novel options for fighting gastric cancer development in H. pylori-infected patients.

Helicobacter pylori and Inflammatory Response

The inflammatory response induced by H. pylori is a key event linked to pathogenesis. Significant insights, summarized in Fig. 1, have been made in the last year on the interactions between H. pylori, mucosal dendritic cells and IL17. The readers are referred to the article on the host response of this issue for more data regarding H. pylori and inflammation.

Figure 1.

Helicobacter pylori, dendritic cells and CD4+ T cells interactions. The direct contact between H. pylori and human gastric epithelial cells (EC) induces the latter to produce both TSLP (thymic stromal lymphopoietin), an activator of dendritic cells (DCs), and MIP-3α, a DCs-attractant chemokine [50], and, in a cag PAI-dependent manner IL-8 and IL-12 [51]. Independently from CagA or VacA, H. pylori alters DCs-polarized Th17/Treg balance toward a Treg-biased response by a stimulatory effect on IL-6 and IL-23 and an inhibitory effect on TGF-β expression by DC [52]. The H. pylori fucose–containing glycans, which can be switched on and off through phase variation expression, bind to the DC receptor SIGN, induce IL-10, not IL-12 or IL-6 expression, thus tailoring Th1 polarization [53]. Activated DCs are induced by CagA, but mainly by CagE, to release high amounts of IL-23 and IL-1 thus inducing the expansion of CD4+ Th17 T cells [54] which might exert an anti-inflammatory role [55] and a role in vaccine induced immunization [56]. Mast cells (MC) [57] and matrix metalloproteinase-7 (MMP-7) [58] were suggested to regulate Th17-dependent inflammatory response to H. pylori infection. Dotted lines represent inhibitory effects and continuous lines stimulatory effects.

In conclusion, in the last year an impressive number of papers have been published on H. pylori genetic variation of genes encoding OMPs, on microbe mimicry with host antigens, on factors that alter host-cell signaling and modulate the host’s immune response. These new insights allow us to improve our knowledge on the pathogenetic mechanism and the true nature of this pathogen, paving the way to better understanding its role in the human disease. In addition, this knowledge may lead to develop a more personalized diagnosis and tailored treatment of H. pylori-related gastrointestinal diseases.

Conflict of Interest

The authors declare no conflict of interest.

Ancillary