Helicobacter pylori induces in-vivo expansion of human regulatory T cells through stimulating interleukin-1β production by dendritic cells

Abstract

Helicobacter pylori is one of the most common infections in the world. Despite inciting inflammation, immunological clearance of the pathogen is often incomplete. CD4⁺CD25^hiforkhead box protein 3 (FoxP3⁺) regulatory T cells (T_regs) are potent suppressors of different types of immune responses and have been implicated in limiting inflammatory responses to H. pylori. Investigating the influence of H. pylori on T_reg function and proliferation, we found that H. pylori-stimulated dendritic cells (DCs) induced proliferation in T_regs and impaired their suppressive capability. This effect was mediated by interleukin (IL)-1β produced by H. pylori-stimulated DCs. These data correlated with in-vivo observations in which H. pylori⁺ gastric mucosa contained more T_regs in active cell division than uninfected stomachs. Inciting local proliferation of T_regs and inhibiting their suppressive function may represent a mechanism for the chronic gastritis and carcinogenesis attributable to H. pylori.

Introduction

Helicobacter pylori, a prevalent Gram-negative bacterium, is considered to be one of the most common infective organisms in the world. H. pylori predominantly colonizes the gastric antrum and establishes life-long chronic infection. While the majority of infections are asymptomatic, H. pylori infection has significant public health and economic implications as it is an important risk factor for gastritis, peptic ulcer disease, malignant transformation in the upper gastrointestinal (GI) tract and elevated cardiovascular risk [1–3]. As a result, antibiotic therapy to eradicate this bacterium is a key treatment of chronic gastritis and peptic ulceration occurring in the context of H. pylori [4].

H. pylori elicits an inflammatory response recruiting neutrophils, lymphocytes and dendritic cells (DCs) to the gastric mucosa [5]. The initial interaction between H. pylori and the innate host immune response is mediated through pattern recognition receptors, such as Toll-like receptors (TLR), expressed on gastric epithelial cells and through the H. pylori virulence factor cag pathogenicity island (cagPAI) [6,7]. The recruitment of DCs to the gastric lamina propria allows for antigen sampling by the extension of their dendrites through the epithelial cell layer [8,9]. We have shown previously that DC activation by H. pylori leads to the production of interleukin (IL)-10, IL-23 and limited amounts of IL-12 [10], and these H. pylori-treated DCs stimulate interferon (IFN)-γ production in naive T cells in vitro [10]. Biopsy material from H. pylori-infected individuals confirms both local infiltration of T helper type 1 (Th1) [11,12] as well as Th17 cells [13,14], suggesting that H. pylori has more than one effect on immunological cells.

CD4⁺CD25^hiforkhead box protein 3 (FoxP3⁺) regulatory T cells (T_reg) are naturally occurring T cells capable of suppressing CD4⁺CD25⁻ effector T cell (T_eff) proliferation and cytokine production [15]. These cells play a critical role in maintaining peripheral tolerance, with their absence resulting in severe multi-organ autoimmune diseases [16]. T_regs also moderate the immune response to pathogens by regulating the balance between immunity and inflammation – while T_reg suppression needs to be overcome for effective anti-pathogen responses, excessive inflammation could result in disproportionate injury to healthy tissues [17]. Evidence has emerged to show a key role for T_regs in maintaining this balance, in some circumstances resulting in pathogen persistence in order to limit tissue injury [18,19]. For example, lesional sites in Leishmania major infection are characterized by the presence of both L. major and large numbers of T_regs that prevent the clearance of infection [18]. Similarly, T_regs limit the inflammatory response to H. hepaticus, thus limiting subsequent tissue damage [19]. In the case of H. pylori, infected individuals have H. pylori-specific circulating T_regs, impairing the memory response to H. pylori [20], and an elevated number of FoxP3⁺ cells in gastric biopsies [21]. This evidence suggests that H. pylori infection results in expansion of the T_reg population and their recruitment to the site of infection in order to limit the inflammatory response.

Pathogen-stimulated DCs have been implicated in the expansion of T_regs. Yamazaki et al. demonstrated that while splenic APCs are poor promoters of T_reg proliferation, bone marrow-derived DCs are capable of inducing T_regs to proliferate to a degree comparable with T_eff during the first 3 days of culture [22]. The underlying mechanisms are thought to be through both contact-dependent (e.g. CD86/80 co-stimulation [23]) and non-contact-dependent [cytokine production, in particular the inflammatory cytokines IL-1, IL-6 and tumour necrosis factor (TNF)-α] processes [24–28].

Based on reports of elevated T_reg numbers in H. pylori-infected sites, we hypothesized that H. pylori instructs DCs to stimulate proliferation of T_regs locally. Furthermore, the presence of chronic inflammation despite the existence of elevated numbers of T_regs suggests that these T_regs have impaired ability to suppress local inflammation. We have investigated the direct and indirect effect of H. pylori on T_reg proliferation and function in vitro as well as investigating T_regs in gastric tissue biopsies infected with H. pylori. We have found that H. pylori-stimulated DCs drive T_reg proliferation, and impair their suppressive function through the production of IL-1β. This is corroborated by in-vivo data showing active division of T_regs in biopsy samples from infected individuals. Dissection of the long-term impact of T_reg modulation and dysregulated immunpathology in the context of H. pylori may provide new insights into the mechanisms underlying the development of H. pylori-associated complications and/or potential targets for the local treatment of inflammation associated with H. pylori in the 15–20% of individuals unresponsive to eradication therapy.

Material and methods

Cells

Peripheral blood mononuclear cells (PBMCs) were separated from buffy coats provided by the National Blood Transfusion Centre (South Thames, London, UK). CD14⁺ and CD14⁻ cells were then separated using CD14-Beads (Miltenyi Biotec, Woking, UK), according to the manufacturer's instructions. The CD14⁺ cells were then cultured in RPMI-1640 (Invitrogen, Paisley, UK) with 10% fetal calf serum (FCS; SeraQ, East Grinstead, UK), 50 IU/ml penicillin, 50 μg/ml streptomycin and 2 mM L-glutamine (PSG) (PAA Laboratories GmbH, Pasching, Austria). To develop DCs, IL-4 (10 ng/ml) (First Link, Birmingham, UK) and granulocyte–macrophage colony-stimulating factor (GM-CSF) (20 ng/ml) (kindly donated by Dr S. Brett, GlaxoSmithKline, Stevenage, UK) were added every 2 days before the cells were harvested at day 5. T cells were enriched from PBMCs derived from buffy coats by negative selection. CD4⁺ T cells were purified using a cocktail of antibodies against CD8, CD33, CD14, CD16, CD19, CD56 and γδ-T cell receptor (TCR). The CD4⁺ T cells were then divided into CD25⁺ and CD25⁻ cells using anti-CD25 beads (Dynal Biotech, Oslo, Norway). For the CD25^hi separation, CD4⁺ T cells were stained for CD4 and CD25 using anti-CD4-allophycocyanin (APC) (S3·5; Caltag, Buckingham, UK) and anti-CD25-phycoerythrin (PE) (3G10; Caltag). The CD4⁺CD25^hi (top 2% for expression of CD25) were then separated from the CD4⁺CD25⁻ T cell population by fluorescence-activated cell sorting (FACS) using a MoFlo high speed multi-laser cell sorter (Cytomation, Fort Collins, CO, USA) running Summit version 3·1 software (Cytomation).

T cell assay

Suppression assays were all carried out in complete medium (RPMI with PSG) containing 10% human serum (Biosera, Ringmer, UK) using 2 × 10⁴ T cells with the following conditions: CD25⁻ alone, CD25⁻ : CD25⁺ at a 1:1 ratio and CD25⁺ alone. These cells were stimulated with CD3/CD28 expander beads (Dynal Biotech) in the presence of H. pylori. Alternatively, the T cells were stimulated with 2 × 10³ allogeneic DCs treated previously with H. pylori, or medium alone for 8 h. Media were supplemented, or not, as described by IL-1 receptor antagonist (IL-1RA) (10 μg/ml, kindly donated by Dr Keith P. Ray, GlaxoSmithKline, Stevenage, UK), anti-IL-6 (10 μg/ml; R&D Systems, Abingdon, UK) or anti-TNFRII (0·2 μg per well; R&D Systems). All conditions were set up in triplicate with appropriate controls for cell density and incubated for 5 days before removal of supernatant for cytokine analysis and addition of 1 μCi/well [³H]-thymidine (GE Healthcare, Little Chalfont, UK). Twenty hours later the cells were harvested and the amount of incorporated thymidine was measured using a 1205 Betaplate liquid scintillation counter (LKB Wallac, Turku, Finland).

Cytokine production

CD40L-transfected L cells (CD40Ltx) [29] were grown in RPMI with 10% FCS and PSG. Once growth was confluent, cells were harvested by incubating them with Versene [ethylenediamine tetraacetic acid (EDTA)] (Cambrex, Verviers, Belgium) for 15 min. They were then removed from the flask, washed in phosphate-buffered saline (PBS) and resuspended in RPMI with 10% FCS and PSG. DCs (6 × 10⁴/well) were incubated in 500 μl RPMI-1640 with 10% FCS and PSG, alone or in the presence of CD40L transfectants (1·5 × 10⁵/cells). These DCs were then incubated with H. pylori [10⁶ colony-forming units (cfu)/ml] or medium alone and supernatants collected 24 h later. Supernatants were analysed using a proinflammatory I 4-plex for IFN-γ, IL-1β, TNF-α and IL-6 (Meso Scale Discovery, Gaithersburg, MD, USA) in accordance with the manufacturer's protocol using a SECTOR™ Imager 2400 (Meso Scale Discovery).

Biopsy material and sectioning

Human gastric biopsy specimens were obtained from the gastric antra of subjects referred to the gastroenterology clinic for endoscopy. CLOtest® (Kimberly-Clark, West Malling, UK) was used to determine H. pylori status. Biopsies were snap-frozen in octreotide (OCT) (Lab-Tek Products, Miles Laboratories, Naperville, IL, USA), sectioned at 6–8 μm on a cryostat and fixed in 4% paraformaldehyde solution. Immunohistochemical analysis was performed in these sections double-stained with the primary antibodies mouse anti-human FoxP3 (259D/C7; BD Pharmingen, Oxford, UK) and rabbit polyclonal against the marker Ki-67 (MM1; Leica Microsystems, Germany). Secondary antibodies were goat anti-mouse IgG antibody conjugated with AlexaFluor 555 and goat anti-rabbit IgG conjugated with AlexaFluor 488 (both from Invitrogen, Paisley, UK). Prolong Gold AntiFade Reagent with 4′,6-diamidino-2-phenylindole (DAPI) (Invitrogen) staining was used to counterstain nuclei. Serial images were obtained with a fluorescence microscope.

Statistical analysis

Statistical analyses were carried out on Microsoft Excel for Windows 2003 (Microsoft Corporation, Redmond, WA, USA). Percentage suppression was calculated as the reduction in proliferation in the presence of T_regs expressed as a percentage of T_eff proliferation in the absence of T_regs. Parametric and non-parametric data were calculated as the mean ± standard deviation (s.d.) and median with associated interquartile range, respectively. For comparison of parametric data, paired and unpaired t-tests were used (for paired and unpaired data sets). Comparisons of central tendency for non-parametric data sets were made by using the Wilcoxon's signed-rank test and Mann–Whitney U-tests for paired and unpaired data, respectively. One-sided tests were used for comparison of small sample sizes (n < 5). A P-value of < 0·05 was considered significant in call cases.

Results

H. pylori expands T_regs in the gastric mucosa

Elevated T_reg numbers have been observed in response to H. pylori infection, both at the site of infection and circulating in the periphery [20,21]. To determine whether the elevated number of T_regs was due to active proliferation at the site of infection, we stained gastric biopsy specimens from patients with and without confirmed H. pylori infection for FoxP3 and the proliferation marker Ki67 (four sections from each patient and four patients). As expected from previous publications, H. pylori-positive biopsy specimens had greater numbers of FoxP3⁺ cells than H. pylori-negative specimens (Fig. 1a). In the presence of H. pylori, a greater percentage of T_regs stained positively for Ki67 (10·2 ± 1·5% versus 2·4 ± 2·0% of FoxP3⁺ cells, P < 0·05; Fig. 1a,b), suggesting that T_regs proliferate in vivo in the presence of H. pylori.

Fig. 1.

Regulatory T cells (T_regs) in Helicobacter pylori-infected gastric mucosa proliferate in vivo. Gastric antrum biopsy samples from patients with (a, right panel) and without (a, left panel) H. pylori infection were stained for forkhead box protein 3 (FoxP3) (green), Ki67 (red) and 4′,6-diamidino-2-phenylindole (DAPI) (blue). FoxP3⁺Ki67⁻ (non-proliferating T_regs, green cells) and FoxP3⁺Ki67⁺ (proliferating T_regs, brown cells, arrowed) were counted and the percentage of proliferating T_regs per section calculated (b). (a) Representative examples; (b) pooled data from four sections. *P < 0·05.

H. pylori-stimulated DCs induce expansion of human T_regs

DCs play a critical role in presenting pathogens to the adaptive immune response. Murine models have indicated that pathogen-stimulated DCs can alter T_reg function [22,26] and their presence in the gastric mucosa indicates that they have the opportunity to influence T_reg function [13]. To determine whether H. pylori-stimulated DCs (HpDCs) can influence T_reg proliferation and can, at least in part, explain the expansion of T_regs seen at biopsy sites of H. pylori-infected individuals [10,13], DCs were generated from peripheral blood monocytes using GM-CSF and IL-4, and incubated with H. pylori [10⁶⁻⁴ cfu/ml corresponding to multiplicity of infection (MOI) of 0·75] for 8 h before being washed and placed in co-culture with allogeneic T_regs for 5 days (Fig. 2), as described previously by us [10]. Allogeneic T_regs were used, as published previously [10], to ensure that the frequency of responding T_regs was not dependent on previous H. pylori exposure and relied purely on the high frequency of alloreactive T_regs [30]. HpDC-induced T_reg proliferation was assessed by [³H]-thymidine incorporation; an example is shown in Fig. 2a. This was confirmed through cumulative experiments with HpDCs (10⁶ cfu/ml), in which the differences between T_reg proliferation in the presence and absence of H. pylori were found to be statistically significant (Fig. 2b). T_regs were enriched using magnetic beads and, although the purity reached 90%, to ensure further that proliferation measured was not due to non-T_reg (e.g. CD4⁺CD25^int T cells) ‘contamination’ of T_reg preparations, T_regs were purified to >98% purity by FACS sorting (to ensure that only the CD25^hi cells were selected) and cultured with DCs as before. HpDCs expanded allogeneic CD25^hi cells, confirming that the proliferation observed was not due to impurities (Fig. 2c). We also ruled out the possibility that H. pylori acts directly on T_regs by stimulating T_regs with anti-CD3/CD28 microbeads in the presence of H. pylori and observing no change in T_reg proliferation under these conditions (data not shown), concluding that enhanced T_reg proliferation was DC-dependent.

Fig. 2.

Helicobacter pylori-stimulated dendritic cells (DCs) induce regulatory T cell (T_reg) proliferation: (a) H. pylori (10^4–6 colony-forming units/ml)-activated allogeneic DCs were co-cultured with T_regs for 5 days and their proliferation determined by [³H]-thymidine incorporation. The plot shows a representative example from 12 independent experiments. (b) Cumulative day 5 proliferation from 12 independent experiments in which immature DCs (ImmDCs) or H. pylori-stimulated DCs (HpDCs) were used to stimulate allogeneic T_regs. Note that (b) is plotted on a log₁₀ scale. (c) Proliferation of fluorescence activated cell sorted (FACS) CD25^hi cells were determined in response to HpDCs. Plot shows a representative example from two independent experiments. *P < 0·05.

T_regs stimulated with HpDC have similar rates of proliferation to T_eff in the first 3 days of stimulation

The efficiency of T_reg suppression of T_effs is dependent on their relative ratio within the same environment. Thus, proliferation of T_regs induced by HpDCs has the potential to favour T_reg suppression by altering this ratio. To gauge the relative ratio of T_regs to T_effs, we therefore compared the kinetics of T_reg proliferation against that of T_effs, starting with the same number of cells. T_regs and T_effs were stimulated by HpDCs for 1, 2, 3, 4, 5 and 8 days and their proliferation determined by [³H]-thymidine incorporation. We found that T_reg proliferation was enhanced by HpDCs as early as day 2, and was comparable to T_eff proliferation. However, after day 4, T_eff proliferation continued to increase whereas the proliferation of T_regs plateaued and then declined (Fig. 3). This suggests that while T_eff have a greater capacity for expansion, T_reg expansion in response to HpDCs is short-lived, this follows similar observations in mouse models [22] that showed a short-lived burst of expansion in T_regs in response to activated DCs, and that the efficiency of T_reg-mediated suppression might be expected to decline after day 3 due to significant changes in relative numbers altering the T_reg : T_eff ratio.

Fig. 3.

Regulatory T cell (T_reg) proliferation induced by Helicobacter pylori-stimulated DCs (HpDCs) is similar to the effector T cell (T_eff) population during the first 3 days: (a) T_effs and T_regs were stimulated with HpDCs and immature DCs (ImmDCs) and proliferation determined at days 1, 2, 3, 4, 5 and 8. The plot is represented on a log₁₀ scale. (b) T_reg proliferation was normalized at each time-point to corresponding T_eff proliferation (where proliferation of the latter = 1). The dotted line shows the point of equivalence, where proliferation of T_effs and T_regs is identical. (a,b) Cumulative mean data from three experiments. Error bars show standard deviation.

H. pylori-treated DCs produce IL-1β, IL-6 and TNF-α and enhance T_reg proliferation via IL-1β

We have demonstrated previously that H. pylori induces DCs to produce IL-23 but only small amounts of IL-12 [10,13]. Because inflammatory cytokines, in particular IL-1, IL-6 and TNF-α, have been implicated in the modulation of T_reg function [24–28], we sought to determine whether T_reg proliferation induced by DCs treated with H. pylori could be caused by production of inflammatory cytokines. To investigate the cytokines produced by DCs in response to H. pylori, DCs were treated for 24 h with H. pylori (10⁶ cfu/ml) and supernatant concentrations of IL-1β, IL-6 and TNF-α determined. H. pylori stimulated IL-1β, IL-6 and TNF-α release by DCs (Fig. 4). As it has been demonstrated previously that ligation of CD40 on DCs further enhanced cytokine release mediated by TLR engagement [31], DCs were cultured with H. pylori in the presence or absence of murine L cells transfected with human CD40L (CD40Ltx cells) [29]. The cytokine production was amplified by the presence of CD40Ltx cells (Fig. 4). Altogether, IL-6 and TNF-α were produced in higher quantities than IL-1β in response to H. pylori, with an interquartile range of 14–20, 1800–8800 and 130–1400 pg/ml for IL-1β, IL-6 and TNF-α, respectively, in the absence of CD40L and 120–250, 12 000–42 000, 8900–19 000 pg/ml for IL-1β, IL-6 and TNF-α, respectively, with CD40Ltx (Fig. 4).

Fig. 4.

Helicobacter pylori-treated dendritic cells (DCs) produce interleukin (IL)-1β, IL-6 and tumour necrosis factor (TNF)-α. DCs were cultured alone or stimulated with H. pylori (10⁶ colony-forming units/ml) for 24 h in the presence or absence of CD40L transfected L (CD40LTx) cells. Supernatants were taken and concentrations of IL-1β, IL-6 and TNF-α (a,b,c, respectively) determined. Graphs show median and interquartile ranges of cumulative data from seven independent experiments. *P < 0·05.

Having found that HpDCs produce IL-1β, IL-6 and TNF-α, we investigated whether these cytokines influenced T_reg proliferation. T_regs were stimulated initially by allogeneic immature DCs (ImmDCs) in the presence of each of these cytokines at 1 ng/ml and 10 ng/ml. We found that only IL-1β was capable of inducing significant T_reg proliferation (Fig. 5a). In addition, IL-1β was capable of mediating its affect in the absence of DCs and could amplify anti-CD3/CD28-mediated T_reg proliferation at concentrations as low as 100 pg/ml, lower than the amount of IL-1β produced naturally by H. pylori-treated DCs (Fig. 5b). We confirmed the role of IL-1β in HpDC-induced T_reg proliferation by stimulating T_regs with HpDCs in the presence of a neutralizing IL-1RA. The addition of IL-1RA inhibited T_reg proliferation, while anti-IL-6 and anti-TNFRII antibodies had no effect (Fig. 5c). These results suggest that IL-1β is the key inflammatory cytokine produced by DCs in response to H. pylori that is responsible for T_reg expansion.

Fig. 5.

Helicobacter pylori dendritic cells (HpDCs) enhance regulatory T cell (T_reg) proliferation through the production of interleukin (IL)-1β. (a) Immature DCs (ImmDCs) were co-cultured with allogeneic T_regs for 5 days in the presence of either 1 or 10 ng/ml of IL-1β, IL-6 or tumour necrosis factor (TNF)-α and their proliferation was determined by [³H]-thymidine incorporation. (a) T_regs were stimulated with CD3/CD28 beads for 5 days in the presence of 10–10 000 or pg/ml of IL-1β and their proliferation was determined by [³H]-thymidine incorporation. (c) T_regs were stimulated by HpDCs in the presence of either IL-RA, anti-IL-6 or anti-TNFRII, and their proliferation was determined after 5 days. Graphs show mean ± standard deviation of cumulative data from three experiments. *P < 0·05.

HpDCs inhibit T_reg suppression of T_effs via IL-1β

Suppression of pathogen-responsive T_effs by T_regs at a site of infection is key to determining pathogen persistence/clearance and the degree of tissue injury caused by local inflammation. To determine, therefore, whether H. pylori affects the suppressive capacity of T_regs, ImmDcs and HpDCs were used to stimulate allogeneic T_eff in the presence and absence of 1:1 T_regs for 5 days and suppression of proliferation calculated. HpDCs impaired suppression by T_regs when compared to co-cultures stimulated with ImmDCs (Fig. 6a). To rule out the possibility that proliferation of T_eff impurities in the T_reg population caused an apparent loss of suppression, we repeated the experiments with CD25^hi T_regs and CD4⁺CD25⁻ T_eff FACS-sorted to >98% purity. As before, suppression of T_effs was still impaired significantly by HpDCs (Fig. 6b). To determine whether the loss of suppression was mediated by IL-1β, T_regs and T_effs were co-cultured at a 1:1 ratio and activated with HpDCs in the presence of IL-1RA. Antagonism of IL-1β resulted in partial restoration of suppression (Fig. 6c), suggesting that suppression of T_effs by T_regs is abrogated by IL-1β produced by HpDC. To determine the capacity of T_regs to inhibit the effector function of T_effs, we measured proinflammatory cytokine concentrations in supernatants of T_effs, T_regs and 1:1 T_reg : T_eff co-cultures stimulated by immDCs or HpDCs. IL-17 production was not detectable in this system, and IFN-γ production was not inhibited by T_regs in co-cultures stimulated with HpDCs, whereas ImmDC-stimulated T_regs could suppress IFN-γ production. (Fig. 6d). Taken together, these data demonstrate that the presence of H. pylori instructs DCs to inhibit T_reg-mediated suppression of T_effs in an IL-1β-mediated manner.

Fig. 6.

Helicobacter pylori dendritic cells (HpDCs) impair regulatory T cell (T_reg) suppression through the production of interleukin (IL)-1β. (a) Effector T cells (T_eff) : T_reg (1:1) co-cultures were stimulated with HpDCs or immature DCs (ImmDCs) and their suppression at day 5 determined. Graph shows data from 12 independent experiments. (b) Suppression of T_effs by fluorescence activated cell sorted (FACS) CD25^hi cells was determined at day 5 following stimulation by ImmDCs or HpDCs. The graph shows one experiment as a representative from two. (c) Suppression of T_eff by T_regs in response to HpDCs was determined in the presence of either IL-RA, anti-IL-6 or anti-TNFRII. Graphs show data from a representative example from three independent experiments. (d) IFN-γ production was determined for day 5 T_eff : T_regs (1:1) co-cultures stimulated with either HpDCs or ImmDCs. *P < 0·05.

Discussion

Persistence of H. pylori is the result of both resistance against the local gastric microenvironment and immunological evasion [32]. Despite making physical contact with immune cells in the lamina propria [33], H. pylori evades immune clearance through a variety of mechanisms including its unique site of colonization, modulation of adhesion and alteration of the host immune response [34]. H. pylori colonization results in recruitment of inflammatory immune cells and chronic increase in proinflammatory cytokines locally, including IL-1β, IL-6, IL-8, IL-17 and TNF-α [35,36]. Among the others, IL-1 has been shown to be a key cytokine in initiating and amplifying the inflammatory responses against H. pylori [37–39]. Very recently, IL-1β present in the gastric mucosa has been shown to play an important role in H. pylori-induced epigenetic changes linking inflammation to carcinogenesis [40]. Finally, H. pylori virulence and IL-1B genes contribute to peptic ulcers and intestinal metaplasia [41].

Elevation of T_regs at the site of infection and H. pylori-specific T_regs in the circulation [20,21] has been suggested as a mechanism of pathogen persistence, on the assumption that T_regs are differentiated cells with professional suppressive function. In this study we show for the first time that H. pylori interacts with human T_regs indirectly via DCs and modifies their function. Our data show that H. pylori-treated DCs stimulate T_reg proliferation, diminish their suppressive function and that DC-derived IL-1β drives this process. Biopsy data from in-vivo H. pylori-infected antrum corroborated these findings, showing that a significant portion of T_regs found in infected gastric biopsies are actively undergoing mitosis. The persistence of H. pylori in the gastric mucosa may allow continual restimulation of the T_reg population. This restimulation may allow for expansion of the T_reg population beyond the 3-day peak observed in vitro. In this model it is not the presence of T_regs that promote the persistence of infection, but rather the persistence of infection that expands the T_reg population in an attempt to limit the damage caused by a prolonged and excessive inflammatory response.

Demonstrations that suppressive function of T_regs can be undermined by pathogens have been shown previously in the context of L. major and H. hepaticus infections, limiting inflammation while hindering pathogen clearance [18,19]. Although pathogens can influence T_reg function directly, such as through engagement of TLR-2, -4 and -8 [42–44], we found that H. pylori had no direct effect on T_regs and that the changes induced in T_reg behaviour could be explained by cytokine production from DCs. We have found that IL-1β plays a central role in mediating the effects of H. pylori on T_regs. This is of particular interest, as virulent strains of H. pylori expressing cagPAI are associated with elevated levels of IL-1β [13,45]. As a result, the influence of H. pylori DCs on T_regs may be enhanced by the local microenvironment. In addition, IL-1β has a significant inhibitory effect on gastric acid production [46], which encourages H. pylori colonization to spread and downstream pathological events (gastritis and gastric cancer). As IL-1β appears to have a central role in H. pylori biology and its mechanisms of immune evasion and chronic inflammation, it may be revealing to study the relationship between polymorphisms in IL-1β and interactions between H. pylori and the host [47,48].

Others have suggested that T_reg function can be modulated by the local cytokine microenvironment, in murine models inhibition of suppression by lipopolysaccharide (LPS)-treated DCs can be reversed by the addition of IL-6 neutralizing antibody [25]. We did not observe a role for IL-6 in the biological effects of H. pylori on T_regs, which is at variance with both the publication of Pasare and descriptions of IL-6R expression by T_regs in inflammatory environments [49]. This can be explained by suggestions that IL-6 is incapable of blocking suppression on its own and requires co-operative action with IL-1 to do so [26], whereas IL-1β has no obligate requirement for IL-6 and can break suppression of T cell proliferation on its own [24]. Alternatively, the variance could reflect differences between murine and human cells. Others have also suggested that IL-12 (but not IL-23) may also be capable of reversing suppression [28], but this result may not be of significance in H. pylori infections, as we have demonstrated previously that H. pylori-stimulated DCs are poor producers of IL-12 [10,13]. We also failed to find a role for TNF-α in the effect of H. pylori on T_regs. Although there is evidence in patients with rheumatoid arthritis that anti-TNF therapy reverses a defect in T_regs [27,50] we postulate that, in similar fashion to IL-6, this effect may be mediated through modification of other cytokines, such as IL-1, that may act in co-operation with TNF.

Finally, it has often been assumed that the presence of T_regs in inflamed sites indicates active T cell suppression. Our observations that H. pylori-stimulated DCs, as well as IL-1β, can subvert T_reg suppression suggests that we should be cautious in this assumption. Equally, emerging data suggest that T_regs, or a subset of T_regs, retain the capacity to convert to the Th17 lineage when stimulated appropriately in the context of inflammation, in particular (for human T_regs) by IL-1β [51]. Such IL-17-producing, or ‘plastic’, T_regs have been described previously in lesional sites of Crohn's disease [52]. We have shown previously that DCs infected with H. pylori stimulate autologous CD4⁺ T cells to produce IL-17 and that this cytokine is expressed in gastric biopsies of patients with H. pylori infection [13]. Infection with H. pylori might not only inhibit T_reg-mediated suppression but also differentiate subsets of T_regs to proinflammatory lineages, such as Th17. While, in this study, we looked for Th17 conversion of T_regs by HpDCs in vitro, we were unable to demonstrate Th17 conversion (data not shown), suggesting that Th17 conversion, if it occurs in response to H. pylori, is restricted to the in-vivo setting, where other components may be involved.

Very recently, a different role for H. pylori infection of DCs has been published. Oertli et al. have demonstrated in a murine model that H. pylori infection reprograms DCs towards a semi-mature phenotype, with the capacity to induce T_regs via the secretion of IL-18 [53]. This finding was supported partially in man by showing that DCs in H. pylori infected human gastric biopsies have a semi-mature phenotype and expressed DC-specific intercellular adhesion molecule-3-grabbing non-integrin (SIGN) [53]. In addition to this, the virulence factor vacuolating cytotoxin has also been shown to regulate DC maturation negatively [54], suggesting that the modulation of DC maturation plays an important role in H. pylori's subversion of the immune response. The study presented here has focused on the effect of H. pylori-infected DCs on naturally occurring T_regs, and whether or not infected DCs are able to produce IL-18 and induce de-novo T_regs has not been investigated. However, many reports published in the last few years have confirmed that H. pylori infection induced DC maturation and the release of IL-23 [10,13,55–57].

In conclusion, we have found that H. pylori expands T_regs in vitro and in vivo and subverts their suppressive function through the production of IL-1β from DCs. These findings question the role of T_regs at H. pylori-infected sites and provide mechanistic and therapeutic insights into the mechanisms of H. pylori-associated chronic gastritis and potential targets for the local treatment of inflammation associated with H. pylori in patients who do not respond to standard eradication therapy.

Funding

This work was funded by grants from the Medical Research Council (to B.A., P.M. and R.I.L.), the British Heart Foundation and Guy's and St Thomas' Charity Trust (R.I.L. and G.L.).

Disclosure

The authors of this manuscript have no conflicts of interest to disclose.