Role of childhood infection in the sequelae of H. pylori disease

Abstract

The persistence of Helicobacter pylori infection plays a fundamental role in the development of H. pylori-associated complications. Since the majority of infected persons acquire the bacteria during early childhood, an examination of the immunobiology of H. pylori infection in children compared with that of adults may help identify host factors that contribute to persistent infection. Therefore, we begin our review of the role of persistence in H. pylori disease with an assessment of the clinical features of H. pylori infection in children. We next review the bacterial factors that promote colonization and evasion of host defense mechanisms. We then focus our attention on the early host immunological factors that promote persistence of the infection and its complications in humans and mouse models. We also highlight topics in which further research is needed. An examination of how immunological factors cause divergent manifestations of H. pylori infection in children compared with adults may provide new insight for therapeutic modification or prevention of persistent H. pylori infection and its complications.

Keywords: :

• Helicobacter pylori
• cancer
• infection
• children

Introduction

Helicobacter pylori colonizes the human stomach, typically during early childhood, and persists for decades, if not for the lifetime of the host.¹ Virtually all colonized persons develop gastritis, but only a small proportion develop clinical manifestations of infection, including gastric and duodenal ulceration, gastric non-cardia adenocarcinoma and gastric mucosa-associated lymphoid tissue (MALT) lymphoma.² Persistent H. pylori colonization is critical for the development of these complications³ and is the strongest risk factor for gastric adenocarcinoma and MALT lymphoma.⁴ Consequently, the clinical manifestations of H. pylori infection occur predominantly in adults, rather than in children.

During acute and chronic H. pylori infection, the bacteria induce both innate and adaptive immune responses.⁵ The innate inflammatory response in the gastric mucosa and the adaptive mucosal and systemic antibody responses, however, are insufficient to provide sterilizing immunity, underscoring the complicated biology of the host-microbe relationship in H. pylori infection.⁶ Impacting the balance in this relationship are bacterial factors that facilitate H. pylori evasion of host defense mechanisms and host factors that regulate inflammatory responses. Emerging evidence suggests that environmental factors also may impact the host-microbe relationship in H. pylori infection. The balance between bacterial, host and environmental factors varies over time. During childhood H. pylori infection, downregulation of the immune-mediated inflammation is strongest, substantially modifying the host response. Therefore, we begin with an assessment of the clinical features of H. pylori infection in children. We next review the bacterial factors that promote colonization and evasion of host defense mechanisms. We then focus our attention on the early immunological factors that appear to promote persistence of the infection and its complications in humans and mouse models of H. pylori infection. An examination of these factors in H. pylori-infected children compared with adults with established infection of many decades will provide new insight into the mechanisms by which H. pylori establishes its long-term relationship with the human gastric mucosa.

H. pylori Colonization in Childhood

H. pylori is typically transmitted orally to children from their mother or siblings.^7-9 In resource-limited settings, particularly in families of low socioeconomic status, exposed subjects acquire the bacteria at a very young age, with a prevalence that exceeds 60% in infants 1‒3 mo of age and over 80% in children 8 y of age in some countries.¹⁰ In developed countries as well, H. pylori is acquired predominately in early childhood, with the majority of children infected by age 5 y.⁸ Similar to the presence of gastritis in virtually all H. pylori-colonized adults,¹¹H. pylori colonization in children is always accompanied by gastric mucosal inflammation.¹² The degree of gastric inflammation in children, however, is significantly less compared with that of adults, even when the children and adults have similar levels of colonization, the same prevalence of CagA and VacA, and are infected by the same H. pylori genotype (predominantly HpEurope), as we have reported for subjects living in Santiago, Chile.^13,14 Reduced gastric inflammation in H. pylori-infected children compared with infected adults also has been reported in Ghana, where residents are likely colonized by the HpAfrica strain.¹⁵

Gastric inflammation in children colonized by H. pylori is characterized by a reduction in polymorphonuclear and mononuclear cell infiltration, decreased incidence of gastroduodenal ulceration, and intact epithelium compared with that of adults.^13,14 A consequence of the reduced gastric response in infected children is the absence or near absence of gastric ulceration in this age group.^16-19 Regarding duodenal ulceration, in one of the largest prospective studies to date of children with proven H. pylori infection,¹⁹ we detected duodenal ulcer in 3.1% of the children, a prevalence considerably lower than the reported 14–15% prevalence in infected adults.²⁰ In a separate study comparing children and adults referred to our tertiary medical center in Santiago, Chile for severe gastrointestinal symptoms, ulceration (predominantly duodenal) was present in 22% of H. pylori-infected children under 13 y of age compared with 47% of adults (P < 0.05).¹³

In addition to the discordant levels of gastric inflammation and gastroduodenal ulceration in H. pylori-infected children and adults, the precancerous mucosal lesions of atrophy and metaplasia are absent or markedly reduced in H. pylori-infected children in the US,^12,21 Europe,²² and Chile.²³ Since the lesions that predispose to gastric cancer, including atrophy, metaplasia, as well as dysplasia,^24,25 are absent or markedly reduced in pediatric H. pylori infection, it is not surprising that gastric adenocarcinoma has not been reported in infected children. Rather, the association between precancerous lesions in the gastric mucosa and gastric cancer applies to adults with persistent H. pylori infection, usually for decades. Thus, in countries with a high prevalence of H. pylori and gastric cancer, non-cardia adenocarcinoma of the stomach is associated with the acquisition of the bacteria during childhood, thereby permitting prolonged infection, rather than acquisition in adulthood.²⁶ Understanding the immunobiological basis for the reduced inflammatory response in H. pylori-infected children, particularly in comparison with that of infected adults, is critical for identifying mechanisms by which the host suppresses the neoplastic potential of H. pylori infection in children.

H. pylori Structural Components Impact Colonization and Disease Outcome

Multiple bacterial factors facilitate H. pylori colonization, adaptation and survival in the highly inhospitable environment of the stomach (Table 1). Indeed, more than 100 H. pylori genes are required for colonization.^27-29 Foremost among the products of these genes is urease, a 550 kD surface protein that hydrolyzes urea into ammonia and carbamate. Carbamate spontaneously hydrolyzes to ammonia and carbonic acid, and the ammonia equilibrates with water to form ammonia hydroxide, rapidly increasing the pH of the stomach to permit colonization³⁰ and promote less viscous mucus through which H. pylori flagella propel the bacteria.^31,32 We have detected the 61-kD subunit of H. pylori urease in the gastric lamina propria of infected adults, noteworthy because urease is chemotactic for both mononuclear and polymorphonuclear inflammatory cells³³ and activates monocyte-macrophage pro-inflammatory responses.^34-37 The contribution of the pro-inflammatory activity of urease to the discordant H. pylori-associated inflammatory response of children vs. adults is not known. In addition to the secretion of acid, the stomach protects itself against luminal microorganisms through peristaltic activity. Resistance to peristaltic activity is facilitated by the bacteria’s spiral shape and flagella, which together enhance motility and thus counteract the “cleansing” effect of gastric peristalsis, thereby promoting colonization and persistence.^38,39

Table 1. Bacterial factors associated with colonization and persistence of Helicobacter pylori infection

Factor	Effect
Colonization
Flagella	Motility in mucus layer
Urease	Evasion of acid
Persistence
Adhesins	Attachment to epithelium
LPS	Evasion of the innate immune response
Flagella	Evasion of the innate immune response
VacA	Modulation of the T-cell response
Gamma glutamyl transpeptidase	Modulation of the T-cell response
NOD-like receptor (NLR) ligands	Modulation of the Treg response
Fucosylated ligands	Modulation of DC-SIGN signaling

After penetrating through the mucus, a small proportion of H. pylori adhere to the underlying gastric epithelium, a process facilitated by the bacteria’s outer membrane proteins (OMPs). These proteins promote bacterial attachment to gastric epithelial cell components: H. pylori OMP blood group antigen-binding adhesion (BabA) binds the fucosylated blood-group antigen Lewis^b,^40,41 and sialic acid-binding adhesion (SabA) binds sialyl-Lewis^x antigen.⁴² Related OMPs include outer inflammatory protein (OipA) and duodenal ulcer promoting protein (DupA), which promote inflammation and duodenal ulceration. Additional OMPs identified by comparative genomics indicate an expansive network of molecules involved in H. pylori adhesion to epithelial cells⁴³; together, the OMPs are encoded by an unusually high proportion (4%) of the bacterial genome.⁴⁴ Such a large number of adhesion proteins indicates the critical and overlapping role that these molecules play in initiating and maintaining H. pylori colonization.

Reflecting the remarkable plasticity of the H. pylori genome, DNA sequences in the OMP genes can undergo nucleotide replacement, frameshift events and slipped-strand misparing in both coding and promotor regions, permitting in vitro and in vivo variation in the expression of adhesion molecules, as shown in a series of elegant experiments.^45,46 The variation in OMPs equips H. pylori with the capacity to persist in the setting of fluctuating environmental challenges and evolving host immune responses. Thus, H. pylori OMPs promote bacterial colonization, intracellular delivery of virulence factors, and persistence of infection,⁴ and the variable expression of these proteins in population-based studies may contribute to the variable level of inflammation, ulceration and possibly adenocarcinoma among different populations. Whether expression of the receptors for the OMP adhesion molecules differ in pediatric vs. adult gastric epithelium is not known.

H. pylori Virulence Factors Modulate Infection and the Host Response

Cytotoxin-associated gene A protein (CagA) and vacuolating cytotoxin (VacA) are the most studied virulence factors of H. pylori. CagA is a 120-kD protein encoded by the 40-kb cag pathogenicity island (cag PAI) that contains approximately 30 genes. Among the cag PAI genes, 18 encode components of the type IV secretory apparatus through which the bacteria translocates CagA into gastric epithelial cells, where the molecule undergoes tyrosine phosphorylation, leading to deregulation of SHP-2 phosphatase that in turn triggers Erk-dependent mitogenesis.^47-49 CagA also interacts with the junctional proteins ZO-1 and JAM, causing disruption of intercellular tight junctions and preventing formation of new tight junctions in model epithelial cells.^50,51 The CagA-dependent molecular events elucidated in these highly refined in vitro cell culture systems cause morphologic changes similar to the dysplastic features induced by H. pylori infection in vivo. The cagA PAI-encoded type IV secretory system also allows inoculation of H. pylori peptidoglycan into gastric epithelial cells where it encounters the intracellular pattern-recognition molecule Nod1, leading to induction of the NF-κB-mediated proinflammatory signaling cascade.⁵² CagA⁺ isolates are strongly associated with gastric carcinoma, as well as more intense gastritis and peptic ulceration.^53-57 However, not all H. pylori are CagA⁺,⁵⁸ and CagA^- isolates, like CagA⁺ isolates, induce persistent infection. Thus, CagA is a critical bacterial virulence factor that promotes H. pylori-gastritis and disease complications, but it appears not to contribute specifically to persistence.

VacA is a prototypic exotoxin that forms anion-selective channels in the membrane of endocytic compartments in epithelial cells. These channels allow the entry of chloride anions, which is compensated by the entry of protons and ammonia, followed by water influx and compartment swelling (vacuolization).^59,60 As shown in epithelial cell studies, VacA also permeabilizes the plasma membrane, leading to release of Fe³⁺, Ni²⁺, sugars, and amino acids, a possible source of nutrients to support bacterial colonization.⁶¹ H. pylori strains with different combinations of signal and mid region nucleotide sequences in VacA alleles differ in cytotoxin production,⁶² geographic distribution,⁶³ and association in vivo with peptic ulcer and gastric cancer.^60,64 As a multifunctional toxin, VacA is capable of interacting with immune cells. VacA binds to T cells through CD18 (integrin β2) and inhibits antigen-induced proliferation of primary human CD4⁺ T cells by suppressing cell cycle progression,⁶⁵ conceivably downregulating expansion of T cells activated by bacterial antigens. H. pylori can be taken up by macrophages^33,66,67 and after internalization may disrupt phagosome maturation, promoting the survival and accumulation of bacteria in megasomes.⁶⁸ VacA can impair antigen presentation in a B cell model,⁶⁹ but whether antigen presentation by macrophages is altered is not clear. Although mouse dendritic cells (DCs) pulsed with VacA express a tolerogenic phenotype, as discussed below, the effect of VacA on the phenotype of human DCs and the specific contribution of VacA to the persistence of H. pylori infection in children have not been elucidated.

H. pylori Evasion of Host Recognition at the Gastric Epithelium

H. pylori has evolved elaborate strategies to evade host immune defenses. These strategies impact the microbe-host relationship at several stages, beginning with the initial interaction between the bacteria and gastric epithelial cells. Epithelial and immune cells express an array of molecules that recognize highly conserved pathogen-associated molecular patterns. Pertinent to H. pylori recognition, gastric epithelial cell lines express toll-like receptors (TLRs) TLR2, TLR4, and TLR5,^70,71 which recognize bacterial lipoteichoic acid, lipopolysaccharide (LPS) and flagellin, respectively, to initiate a pro-inflammatory response. These receptors, as well as TLR9, which recognizes bacterial CpG motifs, are upregulated in gastric epithelium during H. pylori infection in children.⁷² However, whether the level of epithelial TLR cell expression is less than that of gastric epithelium in infected adults, thereby contributing to the reduced gastric inflammatory response in children relative to adults, is not known.

H. pylori can evade TLR recognition by several routes. H. pylori LPS is markedly less biologically active than the LPS of other Gram-negative enteric bacteria,^73,74 preventing induction of TLR4-mediated responses. Nevertheless, the bacteria have retained the ability to induce macrophage pro-inflammatory responses through an LPS-independent mechanism.^34,35 Also, modification of the N-terminal portion of the flagellin protein of H. pylori substantially diminishes TLR5 responses to the bacterium.⁷⁵ In addition, H. pylori genomic DNA induces the TLR9 signaling pathway, eliciting an anti-inflammatory response, at least during early infection in mice.^76,77

The ability of H. pylori to induce a pro-inflammatory response by macrophages, albeit in an LPS-independent manner, also is noteworthy because the released TNF-α downregulates macrophage clearance of apoptotic epithelial cells. As we have reported,⁷⁸ mononuclear phagocytes in human gastric mucosa contain TUNEL- and cytokeratin-positive material, indicating the uptake of apoptotic epithelial cells. Importantly, H. pylori causes apoptosis of primary human gastric epithelial cells while decreasing macrophage phagocytosis of the apoptotic cells.⁷⁸ Thus, the higher numbers of apoptotic epithelial cells and levels of apoptotic material in H. pylori-infected gastric mucosa in vivo is consistent with reduced macrophage clearance. Whether ineffective clearance contributes to mucosal inflammation in atrophic gastritis when the bacteria may no longer be present is unclear. The contribution of these evasive and regulatory pathways to the discordant inflammatory response associated with pediatric vs. adult H. pylori infection is under investigation.

Dendritic Cell Induction of the Host Response to H. pylori

H. pylori attachment to, and penetration of, the gastric epithelium provides local DCs close proximity for direct or indirect contact with the bacteria. We,⁷⁹ and others^80,81 have established that CD11c⁺ DCs are present in the human gastric epithelium and sub-epithelial lamina propria and are more prevalent and activated in H. pylori-infected mucosa compared with uninfected mucosa. The sub-epithelial location of gastric DCs, as well as their rapid accumulation in and subsequent exit from the gastric mucosa, have been shown in a mouse model of H. pylori infection.^82,83 DCs express the C-type lectin Dendritic Cell-Specific Intercellular adhesion molecule-3-grabbing Non-integrin (DC-SIGN) that serves as a cell adhesion receptor,⁸⁴ participates in antigen presentation,⁸⁵ and binds mannose-containing and fucose-expressing pathogens such as H. pylori.⁸⁶H. pylori positive for the LPS epitope Lewis (Le) antigen (Le⁺ phase variants) bind to the carbohydrate recognition domain of DC-SIGN and block Th1 responses, whereas Le^- phase variants escape binding to DCs and induce a strong Th1 response.⁸⁷H. pylori Le sugars, specifically fucose-containing DC-SIGN ligands, modulate DC-directed T cell responses through the DC-SIGN signaling complex (signalosome), causing downregulation of IL-12 and increased expression of IL-10.⁸⁸ Thus, H. pylori influencescross-talk between gastric DCs and T helper cells, potentially modulating the local inflammatory response.

After capturing H. pylori antigens, the DCs likely migrate to gastric lymph nodes to present antigen to naïve CD4⁺ T cells, consistent with the identification of H. pylori-containing DCs in the gastric lymph nodes of infected human subjects.⁶⁸ In mouse studies, increased numbers of DCs and H. pylori-specific T cells have been detected in the gastric lymph nodes of infected mice,⁸⁹ supporting the notion that human gastric lymph nodes are sites of T cell priming during H. pylori infection. Importantly, we also have shown that freshly isolated human gastric DCs pulsed with H. pylori induce naïve T cell proliferation and IFN-γ secretion in vitro,⁷⁹ consistent with the Th1-mediated response characteristic of human⁹⁰ and murine H. pylori infection.⁹¹ In a related mouse study, a DC-based H. pylori vaccine induced an H. pylori-specific Th1 response that caused a reduction in bacterial colonization.⁹² H. pylori infection often induces the accumulation of lymphoid follicles in the lamina propria of human gastric mucosa, reaching its highest expression in children as macroscopically evident nodularity.⁹³ The ability of DCs in gastric lymphoid follicles to present H. pylori antigens to follicle T cells seems likely, given the ability of DCs in intestinal Peyer’s patches to present luminal antigens,⁹⁴ including H. pylori antigens,⁹⁵ to local T cells.

Gastric DC presentation of H. pylori antigens to T cells may be influenced by the microenvironment of the gastric mucosa. Factors associated with, and released by, the extracellular matrix (stroma) of human gastric mucosa potently downregulate the ability of gastric DCs pulsed with H. pylori to induce a Th1 response by naïve T cells, as we have reported,⁹⁶ even though antigen presentation likely occurs in the gastric lymph nodes. Gastric stromal suppression of DC-induced T cell responses is mediated through inhibition of DC release of IL-12, which directs T cell polarization toward the Th1 pathway, and not through a TGF-β-, PGE₂-, IL-10- and TSLP-dependent mechanism.⁹⁶ Whether the gastric stroma down-modulates DC induction of Th17 responses as well has not been reported. However, intestinal stromal products from Crohn disease tissue are able to modulate Th17 responses, as we have reported,⁹⁷ suggesting the possibility that gastric stromal regulation of DC-induced responses extend to the modulation of the Th17 response in H. pylori infection.

In addition to interacting with DCs, H. pylori appear capable of modulating DC function. Using mouse bone marrow-derived DCs and a mouse model of H. pylori infection, investigators have shown that H. pylori secrete a heat-stable factor that inhibits murine DC secretion of IL-12 but not IL-10, thereby limiting the induction of a Th1 response while promoting immune suppression.⁸² H. pylori CagA, which activates the negative regulator SHP-2, also inhibits DC secretion of IL-12, at least in mice.⁹⁸ Interestingly, H. pylori DNA suppresses the production of IL-12 and type 1 IFN by mouse bone marrow-derived DCs and human plasmacytoid DCs and attenuates the severity of the inflammatory response in dextran sodium sulfate-induced colitis,⁹⁹ suggesting a partial explanation for the inverse association between H. pylori infection and inflammatory bowel disease.^100-102 To dissect the interaction between DCs and Tregs, investigators have elegantly shown that H. pylori reprogram DCs to a tolerance-promoting phenotype in an IL-18-dependent manner, thereby promoting Tregs and suppressing effector T cell function.¹⁰³ Subsequently, H. pylori γ-glutamyl transpeptidase and VacA were shown to drive the DC-induced Treg skewing.¹⁰⁴ However, other investigators have provided evidence that tolerance to H. pylori in mice is mediated by DC-induced Treg skewing that is independent of CagA and VacA but dependent on IL-10 and TGF-β.⁸³ Together, the weight of evidence from murine studies indicates that H. pylori induce Treg skewing and that such skewing likely involves DC reprogramming to a more tolerant phenotype and function. The precise mechanism of DC reprogramming and its role in the enhanced Treg response that characterizes H. pylori infection in children is an important topic for future study.

Regulatory T Cell (Treg) Modulation of H. pylori Gastritis in Children

Studies of the immunopathogenesis of H. pylori infection in humans have focused largely on adults. In contrast, H. pylori-induced inflammation in children has received little investigative attention, although elucidating the biology of the pediatric response to infection may uncover cellular events that promote tolerance to the bacteria and persistence of the associated inflammation. To begin to understand the mechanism(s) responsible for the reduced gastric inflammatory response to H. pylori in children compared with the response of infected adults (see above), we have studied two separate child-adult cohorts in Chile, where H. pylori is endemic. In the first study,¹³ infected children had reduced levels of gastric inflammation and Th1 responses with reciprocal increases in the number of regulatory T cells (Tregs) (Fig. 1A and B) and the level of Treg cytokines (TGF-β and IL-10) (Fig. 1C). Gastric TGF-β in the infected children localized predominantly to mucosal CD25⁺Foxp3⁺ cells, indicating Tregs were the primary source of the TGF-β. In the second cohort, the reduced gastritis in H. pylori-infected children was accompanied by reductions in neutrophil infiltration, Th17 cell numbers, and IL-17-specific mRNA and protein levels compared with infected adults¹⁴ (Fig. 2). The reduced Th17 responses in the children were accompanied by reciprocal increases in the number of Treg cells and the level of IL-10 in the gastric mucosa,¹⁴ similar to the reciprocal relationship between Th1 and Treg responses in the children.¹³ The increased gastric Treg response in infected children exceeded that of uninfected children.^13,14,105 These findings confirm the identification of Treg cells in human gastric mucosa by other investigators^106-108 and indicate that the reduced gastric inflammation of H. pylori-infected children is strongly associated with enhanced mucosal Treg activity.

Figure 1. CD25⁺Foxp3⁺ cell respones in the gastric mucosa of children and adults infected with H. pylori. (A, left panels) For control staining, serial sections of ileal mucosa from a patient with Crohn disease were stained with mouse anti-CD25 antibodies followed by green fluorescence-labeled goat anti-mouse IgG antibodies, or mouse anti-Foxp3 antibodies followed by red fluorescence-labeled goat anti-mouse IgG antibodies, and examined by confocal microscopy. Merge of the upper and middle panels shows lamina propria CD25⁺ cells were also Foxp3⁺. (A, right panels) Gastric mucosa from an H. pylori-infected adult and child stained similarly with anti-CD25 and then anti-Foxp3 antibodies and examined by confocal microscopy shows a CD25⁺Foxp3⁺ cell in the adult gastric mucosa and multiple CD25⁺Foxp3⁺ cells in the gastric mucosa of the child. (B) Foxp3⁺ cells in coded gastric biopsies from nine adults and nine children with H. pylori gastritis were enumerated. The number of Foxp3⁺ cells in the gastric mucosa of infected children was 4-fold more prevalent than in the gastric mucosa of the adults (P < 0.03). Inset shows typical Foxp3⁺ cells with nuclear location of the transcription factor. (C) Gastric biopsies from the subjects were homogenized, centrifuged and the supernatants analyzed by immunosorbent and bicinchonimic acid assays for cytokine and protein levels, respectively. Gastric levels of TGF-β1 and IL-10 were significantly higher in the infected children compared with the adults (c = P < 0.01). (Modified from Harris, et al., ref. 13)

Figure 2. IL-17-specific mRNA and protein in gastric mucosa of non-infected and H. pylori-infected children and adults. IL-17 (A) mRNA and (B) protein in H. pylori-infected children were significantly lower compared with that of infected adults. mRNA expression was determined by real-time PCR and normalized to the housekeeping gene GAPDH. IL-17 pg protein was determined by ELISA and normalized to total mg protein. Data are presented as fold change (+/−SEM) in H. pylori-infected children and infected adults compared with non-infected cohort subjects. (Modified from Serrano et al., ref. 14)

Tregs play a fundamental role in the maintenance of immunological tolerance.^109,110 The contribution of gastric Tregs to establishing tolerance to H. pylori begins during early childhood infection. Similar to the role of intestinal DCs in initiating the Treg response in the intestinal mucosa,⁹⁴ gastric DCs likely initiate the accumulation of Tregs in the gastric mucosa through a complex series of cellular events. Elegant studies in mice^111-116 have established that antigen-primed, retinoic acid-producing DCs induce CD4⁺CD25^- naive T cells in the mesenteric lymph nodes to differentiate into CD4⁺CD25⁺Foxp3⁺ Tregs in the presence of TGF-β, which induces the Foxp3 transcription factor. After induction in mesenteric lymph nodes, Tregs acquire the homing receptors CCR9 and α4β7, which direct them to migrate to the intestinal mucosa.^117,118 Although similar receptors have not yet been identified in T cells in gastric lymph nodes, a network of homing receptors, possibly including CCR9 and α4β7, likely direct the recruitment of “educated” Tregs induced in gastric lymph nodes to the gastric mucosa in a manner similar to that of the intestine.

Tregs potently downregulate effector T cells, as well as myeloid cells, to maintain mucosal homeostasis and limit tissue damage through the secretion of immunosuppressive cytokines.^119,120 At sites of infection and inflammation, the Treg cytokines IL-10 and TGF-β suppress effector T cell proliferation and cytokine release, limiting the tissue-destructive consequences of the inflammation. The reciprocal relationship between Treg responses and Th1 and Th17 responses in the gastric mucosa of H. pylori-infected children^79,96 is consistent with the ability of gastric Tregs to suppress H. pylori-induced T cell proliferation, IFN-γ and IL-17 production, and H. pylori-specific memory CD4⁺ T cells.^121-123 The consequence of a strong gastric Treg response in childhood H. pylori infection is a dampened effector cell response to the bacteria and diminished inflammatory damage, resulting in tolerance to the bacteria and persistence of the infection into adulthood.

Persistent H. pylori Infection in Murine Models

Mouse models have provided important insights into persistent H. pylori infection. Similar to the reduced level of inflammation in infected children compared with adults,^13-15 neonatal mice infected with H. pylori do not develop, or develop reduced, gastric inflammation and diminished levels of IFN-γ compared with mice infected as adults.¹²⁴ Moreover, when mice infected with H. pylori as neonates became adult mice, they continue to have less gastritis than mice infected as adults. Neonatally infected animals also have absent or reduced levels of gastric atrophy and metaplasia, similar to the reduced prevalence of precancerous lesions in children infected with H. pylori.^12,21-23 Notably, neonatal mice become colonized by larger numbers of bacteria than adult mice,¹²⁴ whereas children and adults are colonized by equivalent levels of bacteria.^13-15 The increased bacterial burden and absent (or reduced) inflammation and precancerous lesions in neonatal mice can be reversed after Treg depletion, confirming the crucial role of Tregs in mediating tolerance to H. pylori in young mice. Interestingly, neonatally infected mice are permanently protected against inflammation and precancerous lesions due to long-lived tolerogenic Tregs. In sharp contrast, infected children loose their Treg-mediated protection, reflected in the reduced Treg responses and increased Th1- and Th17-mediated inflammation (and precancerous lesions) in infected adults who acquire their infection in early childhood, as we have shown for H. pylori infected subjects in Chile.^13,14 Thus, although murine studies indicate that the neonatal immune system may default to Treg responses upon antigen stimulation, in human H. pylori infection the Treg-induced tolerance is similarly associated with reduced gastric inflammation and complications of infection, but these changes are not long-lived and the bacterial burden is not appreciably impacted by the Treg response. Despite these species differences in the host response to H. pylori infection, murine and human studies together confirm the pivotal role of DCs in directing a potent Treg-mediated response to H. pylori, leading to immune escape and persistent infection (Table 2).

Table 2. Differences in H. pylori infection between children vs. adults and young/neonatal mice vs. adult mice

	Children	Young/Neonatal mice
Inflammation
Polymorphonuclear and     mononuclear cell infiltration	Diminished*	Diminished**
Gastroduodenal ulceration
Gastric ulcer	Absent
Duodenal ulcer	Reduced
Epithelium	Intact*
Precancerous lesions***	Absent*	Absent**
Bacterial factor
Colonization level	Similar*	Increased**
Virulence factors	Similar*
Bacteria genotype	Similar*
Immune response
T reg responses	Increased but not maintained in adulthood	Increased and maintained in adulthood
Th1 responses	Decreased*	Decreased**
Th 17 responses	Decreased*	Decreased**

* Compared with adults; ** Compared with adult mice; *** Atrophy, metaplasia, dysplasia

Understanding the cellular events that regulate tolerance vs. immune-mediated inflammation is critical for devising effective vaccine strategies against H. pylori. Because young children are expected to be the primary target population of vaccine eradication programs that seek to reduce the worldwide burden of H. pylori infection and its complications, the following considerations are of paramount importance. H. pylori-associated inflammation is mediated in part by IL-17 producing cells, which play a crucial role in the induction of inflammation and tissue destruction.^125,126 In mice, H. pylori stimulates DCs in or near the gastric epithelium to induce Th17 cells to produce IL-17, which in turn promotes gastric pathology and neutrophil accumulation.^89,127-129 In neonatal mice, however, H. pylori infection induces tolerizing Tregs, leading to reduced levels of IL-17 in the gastric mucosa and the absence of, or marked reduction in, gastric inflammation in the infected infant mice.^83,124,130 In addition, effective vaccination of mice (adjuvant plus either urease or H. pylori lysate) against H. felis or H. pylori appears to be mediated by IL-17.^131-133 Thus, the absence or near absence of H. pylori-induced inflammation in neonatal mice is the consequence of Treg-mediated suppression of the Th17 (and Th1) response, which corresponds to the Treg-mediated reductions in inflammation and IL-17 levels in children,^13,14,122 yet vaccine clearance of H. pylori is strongly IL-17-dependent. The confounding issue of reduced Th17 responses in young H. pylori-infected hosts but IL-17-dependence of effective vaccine eradication of the bacteria needs to be further clarified, since the global reduction in the prevalence of H. pylori will require immunization of children in order to prevent the complications of persistent infection.

Gastric Microbiota in H. pylori Infection

The induction of gastric Tregs by H. pylori does not occur in a sterile environment. Indeed, genome sequence analyses indicate that distinct, previously undetected, bacterial communities have adapted to, and taken up, commensal residence in the human stomach. Among more than 50 phyla in the environment, only Proteobacteria, Firmicutes, Actinobacteria, Bacteriodetes, and Fusobacteria have been reported to dominate in the stomach.^134,135 During H. pylori infection, the relative abundance of bacterial communities within these phyla in the human stomach appears to change, reflected in the increase in non-Helicobacter Proteobacteria, Spirochetes, and Acidobacteria accompanied by a decrease in Actinobacteria, Bacteroidetes, and Firmicutes.¹³⁶ The impact of these changes on the pathogenesis of H. pylori-triggered inflammation in children and adults is not known. In related studies in the mouse, antibiotic treatment altered the microbiota of the lower gastrointestinal tract, including more cluster IV and XIVa Clostridium spp., resulting in less gastric inflammation and lowered H. pylori-associated IFN-γ mRNA, despite the apparent absence of a Treg response.¹³⁷ However, in a highly provocative study, colonization of gnotobiotic mice with Clostridia IV, XIVa, and XVIII strains from human fecal material did induce intestinal Tregs in the mice,¹³⁸ possibly through short-chain fatty acids, including acetic acid, propionic acid, and butyric acid, which are released during bacterial fermentation and may induce a Treg response.¹³⁹ The gastric environment and physiology, however, may not be as conducive to the generation of short-chain fatty acids as the lower gastrointestinal tract. Nevertheless, these intriguing studies raise the possibility that the microbiota of the stomach in humans acts in concert with H. pylori, at least in some subjects, to induce gastric Tregs.

The microbiota of the lower gastrointestinal tract appears to evolve over time. In infants, the predominant phyla that colonize the lower gut are Firmicutes, Bacteroidetes, and Proteobacteria.^140,141 By the end of infancy, the core microbiome is established.^141-143 By childhood and adolescence, the lower tract microbiota differs substantially from that of adults,¹⁴⁴ possibly contributing to development of specific illnesses such as pediatric irritable bowel syndrome.¹⁴⁵ The ability of a single bacterial species to influence the functional capacity of the enteric microbiome¹⁴⁶ raises the possibility that H. pylori similarly could influence the abundance and/or function of the microbiota of the stomach, or even the lower tract, including the induction of Treg responses. Whether microbial communities in the stomach evolve over time and lower the “set-point” for the induction of Treg responses in the gastric mucosa of children are important topics for future study.

Relationship between Allergic Diseases and H. pylori Infection in Children

H. pylori infection, especially in children, is associated with a reduction in allergic disease, including allergic rhinitis, asthma, and eczema. A relationship between H. pylori infection and atopic diseases was first suggested by the detection of a decrease in the seroprevalence of H. pylori and an increase in allergen-specific IgE in Finish subjects between 1973 and 1994.¹⁴⁷ Subsequently, analysis of the seroprevalence of 22 viral and bacterial pathogens and the presence of atopy in adults living in western Russia showed an inverse correlation between only H. pylori and atopy.¹⁴⁸ Shortly thereafter investigators firmly established an inverse relationship between H. pylori colonization and asthma and allergies in US adults^149,150 and children, especially those less than 13 y of age.¹⁵¹ The inverse association between H. pylori and asthma is especially strong in carriers of CagA⁺ strains.¹⁵⁰ The reciprocal relationship between H. pylori infection and allergic diseases in children has been confirmed in residents of other industrialized countries such as Japan and Germany,^152,153 as well non-industrialized countries such as Ethiopia.¹⁵⁴ In addition, we recently detected an inverse correlation between current H. pylori infection and positive skin tests to an array of allergens and the level of serum IgE in children, but not adults, with severe allergies living in Chile.¹⁵⁵ The H. pylori-infected children with allergies had lower levels of serum and gastric TGF-β compared with infected children without allergies, although the inverse correlation between allergies and TGF-β levels was not confirmed in another study.¹⁵⁶ The accumulating evidence for an association between H. pylori infection and reduced prevalence of allergies in children raises concern that broad H. pylori eradication could promote atopic disease.^135,155

The “hygiene hypothesis,” supported by studies in gnotobiotic mice, holds that early exposure to infectious microorganisms promotes the development of regulatory networks that limit immune-mediated responses to self and foreign antigens.^157,158 As a consequence of accelerating socioeconomic development and the attendant improvement in hygiene, residents of developed countries are less exposed to microbial infections that drive the development of these regulatory processes, particularly during early childhood. An alternative hypothesis suggests that environmental, socioeconomic and lifestyle changes, including less frequent breast feeding and more caesarean sections and antibiotic usage, have lead to alterations in the composition of the enteric microbiota with which humans have co-evolved,^135,159 thereby impacting immune regulatory mechanisms and the physiological basis of certain diseases such as obesity¹⁴³ and kwashiorkor.¹⁶⁰ A consequence of the recent decline in the prevalence of H. pylori,^20,161,162 a “pathobiont” constituent of the gastric microbiota that has colonized humans for at least 60 000 y, is reduced exposure to a major stimulus for the generation of Treg cells. Conceivably, reduction in H. pylori-triggered Treg differentiation diminishes the number of Tregs, resulting in a corresponding increase in the number and responses of tissue effector T cells, which promote inflammation. Thus, because H. pylori potently stimulate DC induction of Tregs in children and those Tregs may disseminate from gastric and possibly mesenteric lymph nodes, the declining prevalence of H. pylori in children diminishes in this population a major regulator of allergen-induced inflammation.

Atopic dermatitis, allergic rhinitis and asthma are mediated through Th2 pathway cytokines, including IL-4, -5, -9, and 13,^163-165 which regulate B cell production of allergen-specific IgE that triggers mast cell and basophil degranulation and release of mediators of airway reactivity.^166,167 In humans, CD4⁺CD25⁺ Tregs from atopic subjects are capable of suppressing allergen-driven Th2, as well as Th1, cell proliferation and cytokine cytokine production.^168,169 Mouse models have provided insight into the role of Tregs in the regulation and inhibition of Th2 cells, showing that Treg IL-10 plays a key role in suppressing Th2 cell differentiation, sensitization and IgE production.^170,171 Elegant studies in mice in which inducible Treg cells were depleted caused the spontaneous onset of Th2-type pathology in the gastrointestinal tract and lung, including allergic inflammation and asthma, in association with altered microbiota composition.¹⁷² These findings bear important relevance to the inverse correlation between H. pylori infection and allergic conditions. Addressing this issue in a series of highly novel studies,^130,173 investigators showed that H. pylori infection, especially in neonates, protected mice from airway hyperresponsiveness, tissue inflammation and goblet cell metaplasia, and prevented eosinophil, Th2 and Th17 cell infiltration into the animals’ lungs. Protection against allergen-induced airway disease was accomplished through H. pylori reprogramming DCs in an IL-18-dependent manner to a tolerogenic phenotype that induced Foxp3 expression in naïve CD4⁺ T cells. These important studies suggest that H. pylori-induced Treg cell differentiation and accumulation in the bronchial tissues and lung may, at least in part, contribute to the reduction in allergen pulmonary disease in H. pylori-infected persons, particularly in children in developing countries, similar to helminth-induced Treg protection against allergen-induced airway inflammation.^174,175 Further, the decline in H. pylori prevalence in industrialized countries with an accompanying reduction in Treg protection, offers a potential, albeit partial, explanation for the increasing prevalence of allergic diseases in developed countries.¹⁷⁶

Conclusion

The acquisition of H. pylori during early childhood elicits non-sterilizing innate and adaptive immune responses that set the stage for persistent infection. Important elements of the immune response that promote persistence are H. pylori stimulation of gastric DCs in or near the epithelium that likely migrate to gastric lymph nodes to induce Tregs that disseminate to the gastric mucosa and down-modulate H. pylori-elicited Th1 and Th17 cell-mediated gastritis and associated complications. For reasons that are unclear, the Treg response in children does not persist into adulthood, leading to reversal of the suppressed Th1 and Th17 responses and increased gastric inflammation with the attendant complications of gastroduodenal ulceration, adenocarcinoma, or lymphoma, at least in some subjects, despite similar levels of colonization in both age groups. In contrast to the enhanced Treg response in humans, which appears to be confined to childhood, H. pylori infection in neonatal mice induces a long-lived Treg response that is accompanied by increased numbers of colonizing bacteria and absent or near absent gastric inflammation and precancerous lesions. Murine models also indicate that vaccine eradication of H. pylori is IL-17-dependent, although the infection in children induces prominent suppression of gastric Th17 responses. Clearly, dissection of the key differences in the immunobiology of human and murine H. pylori infection will enhance our understanding of the pathogenesis of persistent infection and facilitate development of effective vaccine strategies to diminish the worldwide burden of H. pylori-associated disease.

Acknowledgments

Work supported in part by Fondecyt grant 1130387 (PRH); NIH grants DK 064000 and DK 084063 and the Research Service of the Veterans Administration (PDS); Broad Medical Research Institute grant (IBD-0361) (LES); and NIH grants DK 090989 and GM 063270 (GPP).

10.4161/gmic.26943

Disclosure of Potential Conflicts of Interest

No potential conflict of interest was disclosed.