The natural immune response against H. pylori does not effectively clear infection and a combination of immune evasion techniques and bacterial factors leads to a persistent infection ( Figure 1 and Table 1 ). Initial Type 1 inflammatory responses in most infected persons become biased over time towards Th2 and Treg that acts in a kind of damage control measure, which does not clear infection but reduces damage to the host (59–61).

An important feature of H. pylori is its ability to overcome natural barriers preventing most microbes from colonising the human stomach (62, 63). Early studies concluded that H. pylori flagellin evades TLR5 recognition and recombinant H. pylori flagellin was considerably less stimulatory than Salmonella flagellin for example (64). Furthermore, a series of residues within the CD0 domain of H. pylori flagella protein FlaA has recently been found to enable evasion of TLR5 ( Figure 1 ), these are speculated to be a result of point mutations not present in flagella proteins from other bacterium (65). However, a number of studies have now reported that H. pylori activates both TLR2 and 5 on epithelial cells, but not TLR4 (66). A recent report revealed that TLR5 activation is via interaction with the CagY protein which forms part of the TIVSS (67).

CagA is one of the most studied and most important virulence factors possessed by H. pylori. Contact and adherence of H. pylori to host gastric epithelial cells upregulates the expression of both CagA and VacA, as well as inducing the synthesis and assembly of the TIVSS ( Figure 2 ) (71). Both CagA and the TIVSS are encoded by the cag PAI which is an approximately 40 kb chromosomal DNA region present in the most virulent strains of H. pylori. The TIVSS is a pilus-based “molecular syringe” structure that translocates CagA protein into host epithelial cells. The Type IV pilus binds epithelial cells via interaction with β1 integrin on the basolateral side (72). The bacteria gain access to the basolateral side through secretion of the HtrA enzyme which breaks down tight junctions (33).

Once in the cytoplasm, translocated CagA associates with Src and c-Abl family kinases which phosphorylate the CagA carboxyl-terminal EPIYA motif in the inner host cell membrane. Phosphorylated CagA then interferes with host cell signaling pathways, disrupting cell-cell junctions (73). In addition, CagA targets and causes the ubiquitination of the upstream kinase TAK1 (74). The downstream effects of CagA induce signaling cascades that activate NF-κB and downstream proinflammatory signaling (IL-8 secretion), cytoskeletal rearrangements, and changes to tight junctions (31, 33, 75) and inducing apoptosis (70, 76). Higher levels of CagA were shown to cause nuclear translocation of the AP-1 transcription factor and nuclear factor of activated T-cells (NFAT) (77). Furthermore, CagA is directly related to H. pylori carcinogenesis, enhancing DNA double strand breaks, and disabling homologous recombination-mediated DNA repair as well as stimulating Hippo signaling via the inhibition of PAR1b-mediated BRCA1 phosphorylation (78). Additionally, recent work has indicated that CagA initiates the oncogenic YAP pathway, further contributing to the association between CagA and gastric cancer risk (79, 80).

Infection with CagA positive H. pylori strains is one of the strongest risk factors for the development of gastric cancer (75). In addition, multiple H. pylori outer membrane proteins are associated with proinflammatory cytokine induction ( Figure 2 ), these include OipA and BabA, which can induce IL-6, IL-8 and IL-11 (44, 45, 81). Like many bacteria, H. pylori have also been shown to release outer membrane vesicles (OMV) which contain various components, notably including portions of the H. pylori LPS as well as Hop family proteins including BabA and SabA along with other major virulence factors such as VacA and CagA ( Figure 2 ) (74, 82). Interestingly, it appears that H. pylori is also able to regulate cytokine release from host epithelial cells via small noncoding RNAs that have been detected in OMV. These OMV have been shown to suppress IL-8 stimulation from AGS cells in cell culture, presumably by targeting host cell mRNAs, effectively reducing the overall immunostimulatory response to H. pylori LPS and outer membrane proteins (83). Consequently, OMV may play a role in immune modulation and disease progression. Regulation of cytokine release by H. pylori has also been investigated in relation to inflammasomes. H. pylori can modulate NLRP3, a regulator important for inflammasome activation, by inducing secretion of IL-10 in THP-1 monocytic leukemia cell lines, and through the action of miRNAs such as has-miR-223-3p (84). Furthermore, H. pylori infection was reported to suppress the secretion of mature IL-1β from human THP1 monocytes, in spite of upregulation of pro-IL-1β, potentially promoting bacterial persistence (85). Interestingly, cytokine expression has been seen to differ between children and adults, with Il-1β, IL17A, IL-23, IL2, IL-12p70 and IFN-γ upregulated in adults and downregulated in children, and IL-6, TGF-β1, IL-10, TNF-α and IL-1a downregulated in adults and upregulated in children. As a result of these differences, children display a Treg primary response as opposed to the Th1 and Th17 response seen in adults (86–88).

The virulence factor HP-NAP also plays a major role in inducing and modulating host inflammation, recruiting neutrophils to the site of infection and causing a large increase in the neutrophil secretion of reactive nitrogen species, as well as stimulating the release of IL-12 and IL-23 from both neutrophils and monocytes (89). In vitro studies using cloned CD4⁺ T cells from allergen induced T cell lines have shown HP-NAP acts as a Th2 agonist, redirecting Th2 responses into Th1 to drive further inflammation through cytotoxic Th1 responses characterised by increased TNF-α and interferon-γ ( Figure 2 ) (56).

H. pylori is able to adhere to both the secreted gastric mucin MUC5AC (90) and gastric epithelial cells via several surface proteins, such as the blood group binding adhesin (BabA) and the sialic acid binding adhesin (SabA). These surface proteins interact with carbohydrate determinants such as the host Lewis system antigens, again providing a mechanism by which H. pylori prevents its removal by peristalsis as well as gastric shedding ( Figure 2 ) (46).

H. pylori also expresses Lewis system antigens within the LPS, that is thought to represent a form of molecular mimicry which subverts the host immune response. Indeed, there is also evidence suggesting the expression of Lewis determinants is involved in H. pylori interacting with the DC calcium dependent (C-type) lectin DC-SIGN ( Table 1 ) (91). Upon phagocytosis by an innate immune cell, H. pylori can also avoid destruction by interfering with phagosome maturation ( Figure 2 ) (92). Another method of immune evasion used by H. pylori involves the interaction with DC ( Figure 1 ). While H. pylori inhabits a unique niche in the stomach where few cells have access, there has been strong evidence of DC interaction, with DC maturation and function heavily affected in mouse models (93). Furthermore, there is evidence of H. pylori having the capability to infect DC in cell culture (48). H. pylori is capable of modulating IL-12 secretion by DC affecting the maturation and differentiation of CD4⁺ T cells and causing Th1 bias in mice (94). In-vitro studies have supported this, with H. pylori demonstrating the ability to modulate expression of IL-12 and TNF-α in cultured DC (95).

In addition to manipulating DC responses, other virulence factors including VacA (96) along with γ-glutamyltranspeptidase (GGT) (97) and CagA (98) have been reported to interfere with the T cell response to H. pylori, causing a characteristic hyposensitivity of CD4⁺ T cells and suppressing mucosal effector T cells (99). Notably, VacA has been found to have an influence on both DC cytokine expression and T cell differentiation, with VacA positive H. pylori strains showing the ability to supress IL-23 secretion by DC resulting in increased proliferation of Th17 cells. VacA plays additional roles in immune mediation including showing the ability to impair lysosomes and autophagosomes, a function with which VacA shares a synergistic effect with CagA, impairing the ability of both macrophages and gastric epithelial cells to remove internalised or invading bacteria. Specifically, H. pylori is able to inhibit the functions of lysosomal enzymes acid phosphatase, N-acetyl-β-d-glucosaminidase, and cathepsin D, prevent autolysosomal acidification, interfere with retrograde trafficking of mannose-6-phosphate receptors and inhibit autophagosome formation to promote intracellular survival of the bacterium (100).

It is likely H. pylori is highly resistant to destruction by phagocytes, although this has only been seen in-vitro (101). Even though H. pylori is frequently engulfed by phagocytes in vitro, it is proficient at neutralising and resisting the reactive oxygen and nitrate species released by monocytes and macrophages. This allows the bacteria to continue to replicate inside the phagosome, eventually inducing apoptosis of the host cell (49, 101). H. pylori possessing the cag pathogenicity island (cagPI) are even more resistant to phagocytosis than other strains (102), with isogenic mutants of H. pylori lacking cagPI far more readily taken up by phagocytes (49). The mechanism for this appears to be the ability of H. pylori to activate the reverse transsulfuration pathway to induce cystathionine γ-lyase (CTH) in host macrophages, promoting bacterial growth and enhancing bacterial survival in macrophages. H. pylori does this via the CagA-dependant induction of the PI3K/AKT1 pathway, the CagA-independent induction of the MTOR pathway and activation of SP1, increasing macrophage production of CTH (103) Furthermore, it has been shown that H. pylori is able to upregulate the metabolism of macrophage-associated polyamines, impairing M1 macrophage function (104).

The mechanisms of humoral immune evasion by H. pylori are somewhat less clear. Most immunocompetent individuals who are infected with H. pylori develop a specific IgG and IgA response, in many cases high antibody titres to the pathogen, however, this often is not enough to clear infection (53, 105). Furthermore, it has been reported that the type of antibody produced correlates with the outcome of infection, with patients experiencing gastritis and duodenal ulcers having greater titres of IgG, and patients suffering from gastric cancers often having greater titres of IgA (106). Other studies have suggested that a weaker overall antibody response is linked to the development of gastric cancers, and patients who had a weak but still detectable antibody response to H. pylori had a higher rate of gastric cancer than those with a stronger antibody response (107–109).

As described above, the O antigen portion of the H. pylori LPS is associated with the expression of various Lewis antigens in addition to playing a role in the adhesion and immune evasion of the bacterium (116). The O antigen chain is comprised of a glucan group (saccharide composed of glucose), a D-glycerol-D-manno-heptan (DD-heptan) group and a highly conserved trisaccharide (trio) (115). Variability in the O antigen is derived from phase variation, whereby H. pylori uses an on/off system to regulate its biosynthetic genes, including the fucosyltransferase genes FutA, FutB, and FutC (117), allowing the bacterium to adjust the carbohydrate expression of the O antigen with the changing environment of the stomach and gastric mucosa (16). Additionally, phase variation of the H. pylori LPS allows for a greater range of phenotypes and gives H. pylori the ability to modify the expression of Lewis system antigens in its LPS. Additionally, it has been indicated that the H. pylori LPS induces lower biological responses when compared to other gram-negative organisms such as E. coli and Salmonella spp., with in vivo mouse and rabbit studies showing 500 to 1000 fold lower mitogenic and pyrogenic responses (118).

The genes encoding the glycotransferases responsible for the construction of this LPS vary between different regional strains, with the common European strain of H. pylori G27 possessing the trisaccharide fucosyltransferase (HP0102), heptan transferase (HP1283) and GlcNAc transferase (HP1578), the last of which is responsible for initiating synthesis of Lewis system antigens onto a heptan motif (111). In addition to this, a comparison between the European model strain G27 and East Asian strains found that the East Asian strains lacked the genes encoding for heptan transferase and GlcNAc transferase (119). Furthermore, East Asian strains instead express additional copies of other genes, HP1105 and JHP0562, which may act as GlcNAc transferases as well as Gal transferase in place of HP1578 (119). This was identified as an area of interest due to the higher rates of gastric cancer in East Asia and the potential to further characterise the role of the H. pylori LPS heptan in pathogenesis (119). Figure 3 illustrates an abbreviated version of the established structure of the O antigen structure from the mouse-adapted H. pylori SS1 strain including the trio, attachment site and the Lewis antigenic determinants.

The H. pylori LPS is generally considered less immunogenic in comparison to enterobacterial LPS, with early studies showing significant reductions in pyrogenicity, mitogenicity and toxicity, as well as a reduced cytokine and chemokine response (118, 120, 121). Previous studies have shown that synthesised partial H. pylori LPS structures, specifically lipid A compounds and Kdo-lipid A compounds, can modulate cytokine production by host cells. All lipid A and Kdo-lipid A H. pylori LPS structures synthesised either failed to induce, or induced very low levels of IL-1β, IL-6 and IL-8, and conversely stimulated high levels of cytokines IL-12 and IL-18 in heparinised human peripheral whole blood (122, 123). IL-12 induction by H. pylori is generally linked to MyD88 expression in macrophages (124). Furthermore, H. pylori modifies the lipid A portion of the LPS by via dephosphorylation of the 1- and 4-’positions of the lipid A backbone. Mutations to the lpxE/F machinery required for this modification have shown an increased susceptibility of H. pylori to the antimicrobial peptide polymyxin B, as well as reducing the ability of H. pylori to colonise mice (125). Knockout studies on H. pylori genes HP0044 and HP1275 demonstrated the production of a truncated H. pylori LPS missing fucose residues found in both the trio and Lewis antigen portions of the H. pylori LPS O antigen, resulting in the loss of a significant portion of the H. pylori LPS. This in turn affected the growth of the bacterium, increased its susceptibility to both the detergent SDS, promoted bacterial autoaggregation, increased surface hydrophobicity and affected bacterial virulence and OMV protein sorting (126). Notably, phase variation of H. pylori fucosyltransferases, and consequently the glycosylation pattern of the LPS has previously been identified to occur in high frequencies, with in-vitro studies demonstrating the expression of Le^x can vary at a frequency of 0.2-0.5% resulting in differing LPS variants in the same population. Variation of Lewis antigen expression, particularly between Le^x and Le^y has been suggested to be influenced by relative pH in liquid medium. The function for this has not been fully identified, but is hypothesised to aid in bacterial persistence (127). On a molecular level the variation in H. pylori fucosylation has previously been attributed to slipped-strand mispairing as a result of differing numbers of polyC repeats in fucosyltransferase genes futA, futB and futC. However, a recent study has described a role for small RNAs (sRNA) in the regulation of the H. pylori LPS biosynthesis. The sRNA RepG modulate HP0102 and TlpB, providing post-transcriptional regulation to LPS biosynthesis potentially indicating further means by which H. pylori is able to adapt to host immune response (128).

H. pylori binds to the various Lewis system antigens expressed on the surface of gastric epithelial cells, including Le^b and sialyl-Le^x (131). Chronic inflammation also leads to upregulation of Le^x and Le^y in the host, further enhancing bacterial colonisation (131). Furthermore, H. pylori has a range of adhesive proteins that have been identified and examined. The first identified was BabA, a protein that commonly binds the Le^b determinant as well as mucin proteins and the H1 blood group antigen (132). Specifically, BabA binds the lacto series of glycans containing a terminal fucose molecule with an α1,2 linkage attached to a Galβ1-3GlcNAc core (133). This includes H1 blood group antigen as well as other H neoglycoconjugates and Le^b. However, BabA binds to Le^b with as much as a two-fold increase in affinity in comparison to H neoglycoconjugates (134). The BabA encoding gene(s) can be found in three separate loci on the H. pylori genome with the binding pattern of the protein differing based on the specific loci expressed (135). For example, strains containing locus A BabA (BabA2) bind the classically associated Le^b whereas strains with locus B BabA (BabA1) do not bind to Le^b (136). Furthermore, some strains may contain more than one BabA encoding gene allowing expression of proteins with varying binding characteristics (137). BabA in general is a highly polymorphic protein with variable expression, which may be lost during laboratory cultivation. Additionally about one quarter of strains isolated from chronically infected hosts tend to lose BabA expression (69, 138). This suggests that BabA is a specific adhesin that is advantageous in the colonisation of a human host that is unnecessary on solid media (137). There has been evidence of numerous binding molecules for BabA in addition to the Le^b molecules expressed by gastric epithelia. Several salivary proteins have also been identified to interact with BabA, such as the mucin MUC5B, the agglutinin glycoprotein-340 and the proline-rich glycoprotein containing the Fucα1-2Galβ motif. In addition to this, fucose-containing oligosaccharides present in secretory IgA also play a role in binding BabA, however this is not universal to all secretory IgA (132). It has been observed that BabA presence in a H. pylori strain is associated with expression of VacA, CagA and OipA among other common but variable virulence factors, particularly in the outcome of gastric metaplasia. An in-vitro study found that Le^b-positive AGS cells expressed increased levels of mRNA for the cytokines CCL5 and IL-8 as well as precancer related factors CDX2 and MUC2 in the presence of wild type H. pylori, however this was not the case for H. pylori mutants with either BabA genes or TIVSS deleted (139).

SabA is an additional adhesion molecule associated with the binding of Lewis system determinants by H. pylori. More specifically, SabA binds sialylated molecules such as sialyl-Le^x and some gangliosides characterised by Neu5Acα3-neolactohexaosylceramide and Neu5Acα3-neolactooctaosylceramide molecules (58). The gastric inflammation caused by H. pylori infection is essential for changes in glycosylation patterns within the gastric mucosa, promoting the expression of sialyl-Le^x as well as sialyl-Le^a, which in turn increases the adhesive properties of SabA (140). Similar to BabA, the expression of SabA is highly variable and is often subject to phase variation. It has been suggested that H. pylori may have multiple means of modulating the expression of SabA, with one such method being an acid-responsive ArsRS two-component signal transduction system, and an alternative involving the slipped-strand mispairing of SabA alleles during chromosomal replication in various H. pylori subpopulations (141). Additionally, SabA is commonly subject to homologous recombination and gene conversion due to changing environmental pressures (142). Deactivation of the SabA encoding gene may further assist in bacterial escape as cytotoxic activity by neutrophils is in part triggered by SabA binding to gangliosides. Neutrophil action has been shown to be defective against SabA deleted H. pylori cells (58).

OipA is also a member of the Hop family of outer membrane proteins possessed by H. pylori. There is currently no available crystal structure for OipA, limiting detailed studies of its binding (143). OipA is known to be closely associated with the cag PAI and CagA, with the oipA locus approximately 100 kb from the cag PAI. OipA is regulated by slipped-strand mispairing and displays on/off states which are closely associated with the expression of CagA. Alleles of oipA have been demonstrated in up to 96% of cag PAI positive H. pylori strains, solidifying the relationship of the two virulence factors (144). With the exception of CagA, OipA is typically independent of other H. pylori virulence factors. OipA is associated with the induction of IL-8 from host epithelial cells, inhibition of apoptosis and enhanced adhesion to gastric cells in vitro (144). Induction of IL-8 and its subsequent role in inducing gastritis has been associated with a synergistic effect from H. pylori strains possessing both the cag PAI and a functional oipA gene (145). In contrast oipA “off” strains can down-regulate anti-apoptotic processes in AGS cells, more regularly inducing apoptosis on infection than oipA “on” strains in-vitro (146).