Dr S. Raghavan, Department of Medical Microbiology and Immunology, Göteborg University, P.O. Box 435, 405 30 Göteborg, Sweden. E-mail: email@example.com
A Helicobacter pylori-specific in vitro coculture system was established and used to study the role of CD4+CD25+ regulatory T cells (Treg) in gastritis development in mice with H. pylori infection. Effects of therapeutic immunization against H. pylori infection on the Treg function were also studied to better understand the mechanisms leading to postimmunization gastritis in these mice. Depletion of Treg led to extensive proliferation to H. pylori antigens of CD4+ T cells isolated from either naïve, H. pylori-infected or H. pylori-immunized mice. Using the Treg-depleted CD4+ T cells from immunized mice as effector cells, we compared the suppressive efficacy of Treg isolated from naïve, infected or immunized mice and found that Treg from naïve mice, and slightly less efficiently from infected mice, suppressed the CD25– effector T-cell response and in most cases were distinctly more efficacious than Treg isolated from immunized mice. The suppressive efficacy of Treg isolated from the differently treated mice correlated closely with production of interleukin-5 (IL-5) by the Treg and suppression of interferon-γ and IL-2 production by the CD25– effector T cells. Our study is the first to demonstrate in H. pylori-induced chronic infection, antigen-specific Treg with differential efficacy in suppressing H. pylori proinflammatory T effector cells.
A unique population of suppressor T cells coexpressing CD4 and CD25 [interleukin-2Rα (IL-2Rα)] antigens was first identified in mice and is now referred to as natural CD4+CD25+ regulatory T cells (Treg) . Mice thymectomized at day 3 of life are depleted of Treg and develop autoimmune diseases , suggesting that these T cells originate in the thymus and migrate to the periphery . Recently Treg (CD4+CD25high) have also been purified and characterized in peripheral blood [4–7], thymus [6, 7] and cord blood  of humans. In vitro, Treg from both mice and humans are anergic (but not unreactive) to stimulation via their T-cell receptor (TCR) and further inhibit CD4+CD25– T-cell responses to anti-CD3 [8, 9]. Suppression occurs only when the Treg are activated through their TCR , and both anergy and suppression can be overcome by addition of IL-2 into the coculture or by enhancing endogenous IL-2 production by the addition of anti-CD28 antibody [8, 10]. In vivo, Treg have been shown to suppress autoimmune and inflammatory responses in mice which in some, but not all systems, involves the secretion of the suppressor cytokines IL-10 and transforming growth factor-β.
Data from several animal models of infection have further shown that the removal of Treg cells can result in an excessive immune response to microbial antigens causing subsequent immunopathology. For example, specific depletion of Treg elicits inflammatory bowel disease in mice due to the remaining T cells reacting to commensal bacteria in the intestine [11, 12]. Enhanced stimulation of the immune system in the absence of Treg has also been documented in the development of lung inflammation caused by Pneumocyctis carinii and cutaneous inflammation caused by Leishmania major. In both these models, removal of the Treg population resulted in severe pathology but also led to complete clearance of the infection and resolution of the inflammation within a short period of time. Whether this is a beneficial immunological intervention to infection is still a matter of debate.
Helicobacter pylori are spiral-shaped gram-negative bacteria that colonize the human stomach and duodenum and may cause chronic inflammation, gastric atrophy and peptic ulcers, and H. pylori infection has also been identified as an important risk factor for the development of gastric cancer . More than half the world's population is chronically infected with this bacterium with a higher incidence of infection in developing than in the industrialized countries. The chronic infection results in symptomatic disease in 10–15% of those colonized, while the remainder are designated as asymptomatic carriers, in spite of an ongoing chronic active inflammation in the stomach . In H. pylori-infected mice, absence of Treg has been shown to be associated with a loss of immune regulation leading to increased pathology . In humans with asymptomatic H. pylori infection, Treg-suppressing T-cell responses have also been described, suggesting that the activation of Treg helps maintain a balance between chronicity of infection and development of symptoms .
We and others have shown that immunization of mice with H. pylori antigen together with a mucosal adjuvant results in protection against H. pylori infection . However, the reduction in bacterial load as a result of immunization is associated with intense inflammation in the gastric mucosa (postimmunization gastritis) due to an expansion of H. pylori-specific T cells that migrate to the gastric tissue to perform their effector functions [18–21]. Given the importance of Treg in suppressing effector T-cell responses in many inflammatory conditions and their proposed regulatory significance in H. pylori infection [16, 17], our aim in this study was to set up and use an in vitro H. pylori-specific T-cell coculture system that would allow us to [i] test the ability of highly purified Treg to suppress H. pylori-reactive effector T cells in an antigen-specific manner and [ii] study the effects of therapeutic immunization against H. pylori infection on the functional efficacy of Treg.
Materials and methods
Animals. Six- to eight-week-old C57BL/6 mice were obtained from Charles River (Sulzfeld, Germany). They were housed in microisolators at the Laboratory for Experimental Biomedicine, Göteborg University during the study. All experiments were approved by the ethics committee of the National Board for Laboratory animals.
Primary infection and immunization of mice. The mouse-adapted H. pylori strain SS1, stored at −70 °C kept in Luria Bertani medium containing 20% glycerol, was used as the stock culture for all experiments. H. pylori strain SS1 was cultured for infection as described previously . Mice were infected intragastrically with approximately 3 × 108 CFU of bacteria in Brucella broth. Two weeks after infection, mice were divided into two groups [i]: either left untreated (infection controls) or [ii] given four weekly doses of 400 µg of H. pylori lysate antigen perorally together with 10 µg of cholera toxin as previously described . Eight weeks after infection or 2 weeks after the last booster immunization, mice were killed and sera and mesenteric lymph nodes (MLNs) were collected.
Serum antibody titres. Serum antibody titres were determined by enzyme-linked immunosorbent assay (ELISA) , using a membrane antigen preparation of H. pylori SS1 for coating . Sera from individual mice in each group were tested at an initial dilution of 1/10 followed by serial threefold dilution, and the antibody titres were expressed as the reciprocal sample dilution giving an absorbance of 0.4 above the background.
H. pylori antigen-specific proliferation assay. Preparation of antigen-presenting cells (APC): CD19+ B cells were isolated by positive selection using MACS beads (Miltenyi Biotech, Bergish Gladbach, Germany) and plated in Complete Iscoves medium at 1 × 106 cells/well in a 24-well plate with or without H. pylori lysate antigen  (10 µg/ml) for 24 h. The cells were then harvested, washed twice and irradiated (2000 rad) before being counted and added to the T cells.
Preparation of T cells: CD4+ T cells were isolated from the MLN cells by negative selection using the MACS CD4+ T cell isolation kit (Miltenyi Biotech) according to the manufacturer's instructions. The CD4+ T cells were further purified into CD25+ and CD25– T-cell populations using anti-CD25 PE and anti-PE microbeads (Miltenyi Biotech). The resulting purity of the MACS-isolated CD4+CD25+ and CD4+CD25– T cells was confirmed by FACS analysis to be 90–95%.
Cell culture in vitro: APCs pre-cultured with or without antigen (2 × 105) were mixed with 4 × 105 total CD4+ or isolated CD4+CD25– T cells in triplicate. In some wells, CD4+CD25+ T cells (4 × 105) were added to the CD4+CD25– cells. The cells were cultured for 96 h, and proliferation was determined by addition of 1 µCi of [3H]-thymidine (3H-dThd, Amersham International, Bucks, UK) per well as previously described .
Transfer of Treg prevents immunopathology in a mouse model of H. pylori infection and inflammation
Our previous study  showed that transfer of LN cells depleted of Treg (CD25–LN) into nu/nµ mice results in severe pathology after H. pylori infection, compared to mice receiving LN cells containing Treg. Independent microscopic re-examination of the stomach sections from the differently treated mice of that study confirms the notion that Treg are able to inhibit effector T-cell responses against H. pylori infection in vivo. Thus, in mice receiving CD25– LN cells 6 weeks after infection with H. pylori, a higher incidence and increased severity of gastritis was seen compared to mice receiving total LN cells containing Treg (Table 1). Furthermore, as Treg need to be stimulated via their TCR in order to display regulatory function , these results also suggest that the anti-inflammatory Treg in this model specifically recognize and are functionally activated by H. pylori antigens. Our goal in the subsequent work presented here has been to define the presence and function of such H. pylori-specific Treg using in vitro coculture assays with effector T cells.
Table 1. Incidence and severity of gastritis 6 weeks after infection with Helicobacter pylori in nu/nµ mice transferred with lymph node (LN) cells or CD25-depleted LN cells: increased frequency of severe gastritis after transfer of CD25-depleted cells
P < 0.05 compared to mice transferred with LN cells.
Development of an in vitro culture system allowing for the analysis of CD4+ T-cell responses to H. pylori infection and immunization
To study the specific T-cell response to H. pylori antigen, we first isolated CD4+ T cells and used irradiated T-depleted spleen cells as APC. However, addition of H. pylori antigen to the cell culture with APC and CD4+ T cells resulted in cell death of the T cells, most likely due to the toxicity of Vac A cytotoxin in the antigen preparation . To counteract the toxic effects of the antigen preparation on the T cells, we therefore developed an alternative system, wherein we used isolated CD19+ primary B cells from naïve mice as APC and exposed them to antigen for 24 h and subsequently washed off any antigen that had not been taken up by the B cells. FACS analysis showed that B cells pulsed with H. pylori antigen upregulated MHC II and CD69 activation markers compared to unpulsed B cells (data not shown).
Addition of antigen pulsed B cells to purified CD4+ T cells isolated from either naïve H. pylori-infected or immunized mice resulted in antigen-specific proliferation that was the highest in the immunized mice (Fig. 1A).
When sera from naïve, H. pylori-infected or therapeutically immunized mice, respectively, were analysed for H. pylori-specific antibody titres by ELISA, the results concurred with the immunization status inferred by the antigen-specific T-cell responses. Thus, naïve mice had low but detectable antibody levels to H. pylori antigens, which were elevated in the sera of infected mice, while immunized mice had the highest H. pylori-specific antibody titres (Fig. 1B).
Different efficacy of Treg isolated from naïve, infected or immunized mice in suppressing proliferative responses to H. pylori antigen
Most studies until now have examined the suppressive activity of Treg after polyclonal stimulation. We used an antigen-specific system to demonstrate the presence of Treg capable of suppressing H. pylori-reactive effector T cells and to address the question whether Treg efficacy is altered in mice infected with H. pylori or therapeutically immunized with H. pylori antigen. To this end, Treg were isolated from naïve, H. pylori-infected or H. pylori-infected and immunized mice, respectively, and their capacity to suppress H. pylori-specific proliferative response was evaluated in coculture with CD4+CD25– T cells. In all groups, removal of Treg from the CD4+ T-cell fraction resulted in a further increase in proliferative response to H. pylori antigens, indicating that the CD4+ response to H. pylori in vivo is under the control of Treg (data not shown). In initial experiments, we saw that the CD25– T cells isolated from H. pylori-immunized mice responded with the highest proliferation and IFN-γ production. We therefore used CD25– cells from immunized mice as effector cells (CD25– Teff) in subsequent experiments to compare the efficacy of Treg isolated from naïve, infected or immunized mice, respectively.
Treg from all groups of mice had demonstrable suppressive activity. To our surprise, Treg isolated from naïve mice were often the most potent ones in suppressing the proliferation of CD25– Teff cells, while Treg from immunized mice were the least effective (Fig. 2A,B). The lesser potency of Treg from immunized mice to suppress T-cell proliferation of CD25– Teff cells was most likely not due to a contamination with activated/effector CD25+ T cells. The FACS analysis profile showed a similar percentage of CD4+CD25+ T cells in CD4+ T cells isolated from the different groups of mice (mean ± SD percentage CD25 of CD4+ T cells, 15 ± 2, 15 ± 3 and 18 ± 2 from naïve, infected and immunized mice, respectively; three independent experiments) (Fig. 2C). In addition, CD25+ cells of similar purities were recovered from the different groups of mice, and Treg from all three groups of mice were anergic to stimulation with H. pylori antigen, and their anergic state was overcome by the addition of rIL-2 in culture (data not shown). Finally, suppression of polyclonal stimulation (αCD3) by Treg from naïve, infected and immunized mice was similar (Fig. 2D), indicating that the differential efficacy of Treg isolated from the different groups of mice to downregulate CD25– Teff cell responses is specific to H. pylori antigen.
Treg produce IL-5 and inhibit production of Th1 cytokines by H. pylori-stimulated CD25– Teff cells
Using the cytometric bead array technique, we analysed cytokines IL-2, IFN-γ, TNF-α and IL-5 in the culture supernatants to study the Th1 and Th2 cytokine profile of CD25– Teff cells alone or in coculture with Treg. As shown in Fig. 3, while CD4+ and CD25– T cells secreted large amounts of the Th1 cytokine IFN-γ in response to H. pylori antigen, they produced little or no IL-5. This pattern was reversed in CD25– Teff cells cocultured together with Treg isolated from the different groups of mice. Thus, coculture with Treg isolated from naïve or infected mice resulted in complete suppression of IFN-γ production. Treg from immunized mice also suppressed IFN-γ production by the CD25– Teff cells but usually to a lesser extent than Treg isolated from naïve or infected mice (Fig. 3).
Effective suppression of H. pylori-specific proliferative responses was associated with production of high levels of IL-5. Naïve Treg that showed the highest fold suppression of CD25– Teff cells also produced the most IL-5 in coculture, and conversely, Treg from infected or immunized mice progressively produced less IL-5.
In addition to their suppression of IFN-γ, Treg from naïve mice and infected mice and to a lesser extent Treg from immunized mice also suppressed IL-2 and TNF-α responses of CD25– Teff cells in an antigen-specific manner (data not shown). Thus, the potency in suppressing T-cell-proliferative responses is associated with similar efficacy in inhibiting cytokine secretion by CD25– Teff cells in coculture.
In a mouse model of chronic inflammation induced by H. pylori infection, after confirming our results from a previous study  that infected animals lacking Treg develop more severe gastritis than in mice containing such cells, we could further demonstrate that Treg isolated from either näive or infected mice were potent in suppressing H. pylori-specific CD25– effector T-cell responses (proliferation and cytokine secretion) in vitro. Functionally suppressive Treg were also isolated from therapeutically immunized mice but were found to have a lesser potency than Treg from naïve mice. Suppression of proliferation was associated with increased levels of the cytokine IL-5 in the culture supernatants and decreased levels of IFN-γ, IL-2 and TNF-α. Thus, Treg activated by H. pylori antigen in vitro suppress CD25– effector T-cell responses in an antigen-specific manner.
Our finding of a reduced efficacy of Treg from immunized mice to suppress CD25– Teff cells suggests that this could be an important mechanism explaining postimmunization gastritis seen in these animals . The vaccine that was used in the study was comprised of a lysate antigen prepared from H. pylori bacteria administered together with cholera toxin as a mucosal adjuvant. Several studies have shown that the adjuvant effects of cholera toxin in vivo is mediated through the induction of IL-1 and IL-6 and upregulation of costimulatory molecules on the APCs , thus allowing for a more efficient immune response to coadministered antigens. Furthermore, IL-6 has been also implicated in directly suppressing Treg function . A mechanism for the development of postimmunization gastritis  could thus be both an enhanced activation and expansion of effector T cells and a concomitant reduction in Treg function. Our in vitro data support the latter hypothesis. As it can be assumed that reactivated effector T cells are more difficult to suppress than resting T cells, it would not have been surprising, if Treg from immunized animals were not able to suppress the effector T cells from the same immunized mice. Instead, we show to our surprise that Treg from naïve mice suppressed the effector cells from the immunized animals efficiently and in most cases better than those from the immunized animals. In naïve mice under non inflammatory conditions, a Th2 response to Helicobacter antigens most likely prevails in the periphery. Indeed, we saw a lower IFN-γ response of CD4+ T cells isolated from naïve mice to H. pylori antigens compared to the IFN-γ response of CD4+ T cells isolated from either infected or immunized mice (unpublished data). We also observed the highest amount of IL-5 in culture supernatants in coculture experiments using Treg isolated from naïve mice and progressively decreasing IL-5 production when Treg from infected or immunized mice were used. Suppression was consistently linked to the production of IL-5, and we believe that Treg are most likely the source of the cytokine, as Treg alone but not CD25– Teff cells cultured together with H. pylori antigen produced high levels of IL-5. To our knowledge, IL-5 production by Treg has not been described earlier. This could be due to the differences in the antigen-presenting conditions to the Treg. In this regard, it is noteworthy that, in contrast to others studying Treg function to either polyclonal or antigen-specific stimulation, we have used purified B cells as APC. The specific significance of the IL-5 produced for mediating Treg suppression of Teff cells remains to be determined; however, it is striking how well the levels of IL-5 correlated with the suppressive efficacy.
Thus, our studies suggest that, in naïve mice, a predominant Th2 response favours the expansion of Treg, while in immunized mice, a Th1 response instead hinders the expansion of Treg, thus pointing to an important role of the Th1/Th2 balance to influence Treg function in H. pylori infection. Further studies are needed to confirm this finding with the possibility regulating Treg response for the control of inflammation caused by H. pylori infection or immunization.
One more trivial explanation for reduced function of the isolated CD25+ cell populations from immunized mice could also be the contamination of Treg with CD25+ expressing H. pylori-specific activated T cells. However, FACS analysis data showed that there was no quantitative difference in the CD25 intensity between naïve and immunized mice speaking against this possibility, although not formally disproving it. Furthermore, our finding that Treg from immunized mice show potent suppressive activity when stimulated with anti-CD3 and were completely anergic to stimulation with H. pylori in the absence of IL-2, which are hallmarks of Treg, also speaks against a dilution with effector T cells. Future quantitative studies of the specific marker Foxp3 in the Treg isolated from immunized mice compared to naïve or infected mice may resolve this question conclusively.
The reason for the highest efficacy of Treg suppression of CD25– effector responses by cells isolated from naïve mice in spite of not harbouring the infection in the stomach is not clear. The mice that were used for these studies were specific pathogen free (SPF) but not Helicobacter free and could thus have a low-grade infection in the intestine with naturally infecting Helicobacter species. Consistent with this, the naïve mice displayed low but significant antibody levels to H. pylori and also responded with T-cell proliferation to H. pylori antigen stimulation in vitro. It is possible that this low level of infection in the intestine also expands the Helicobacter-specific Treg which cross-recognize H. pylori antigen. A Helicobacter species that is known to naturally colonize mice in the intestine is H. hepaticus. Using the H. hepaticus model of intestinal inflammation in IL-10–/– mice, Kullberg et al.  have shown that immune regulation of intestinal inflammation is under the control of regulatory T cells and is driven specifically by H. hepaticus antigens. Thus, it reasonable to assume that, in the immunocompetent näive mice used in this study, Treg specific for Helicobacter antigens do exist that recognize H. pylori antigen and mediate suppression in our in vitro coculture experiments. Further studies using Helicobacter-free mice would clarify whether Treg in naïve mice are specific for Helicobacter or induced by cross-reacting enterobacterial antigens. In this context, it is interesting to note that Treg with the ability to suppress T-cell responses against enteric (fecal) antigens have also been isolated from germ-free mice .
In conclusion, our study is the first to demonstrate antigen-specific suppression of proliferative and proinflammatory responses of Th1 effector cells by Treg in relation to H. pylori infection and immunization and to show differences in suppressive efficacy of Treg isolated from naïve, H. pylori-infected or therapeutically immunized mice. These studies will help to better understand the immune mechanisms leading to inflammation as a result of H. pylori infection and should guide immunological interventions aiming at the control of inflammation through modulation of Treg function for future therapy in humans.