HELIGMOSOMOIDES POLYGYRUS INFECTION CAN INHIBIT COLITIS THROUGH DIRECT INTERACTION WITH INNATE IMMUNITY

Abstract

Less developed countries have a low incidence of immunological diseases like inflammatory bowel disease (IBD), perhaps prevented by the high prevalence of helminth infections in their populations. In the Rag IL10−/− T cell transfer model of colitis, Heligmosomoides polygyrus (Hp), an intestinal helminth, prevents and reverses intestinal inflammation. This model of colitis was used to explore the importance of innate immunity in Hp protection from IBD. Rag mice briefly exposed to Hp before reconstitution with IL10−/− colitogenic T cells are protected from colitis. Exposure to Hp before introduction of IL10−/− and OT2 T cells reduced the capacity of the intestinal mucosa to make IFNγ and IL17 after either anti-CD3 mAb- or OVA-stimulation. This depressed cytokine response was evident even in the absence of colitis suggesting that the down-modulation in pro-inflammatory cytokine secretion was not just secondary to improvement in intestinal inflammation. Following Hp infection, dendritic cells (DCs) from the lamina propria (LP) of Rag mice displayed decreased expression of CD80 and CD86, and heightened expression of PDCA-1 and CD40. They were also less responsive to LPS, producing less IL12p40 and IL10. Also diminished was their capacity to present OVA to OT2 T cells. These experiments infer that Hp does not require direct interactions with T or B cells to render animals resistant to colitis. DCs have an important role in driving both murine and human IBD. Data suggest that phenotypic alternations in mucosal DC function are part of the regulatory process.

INTRODUCTION

Inflammatory bowel disease (IBD) and other immune-mediated illnesses are rare in tropical, less-developed countries. Helminth infections common in such countries may prevent IBD (1). Various helminth species used in several animal models of intestinal inflammation can avert or limit disease activity. For instance, rodents receiving non-viable schistosome ova (2) or intestinal helminths like Trichiura muris (3), Trichinella spiralis (4), Heligmosomoides polygyrus (Hp) or Hymenolepis diminuta (5) are protected from trinitrobenzene sulfonic acid (TNBS)-induced colitis. T. muris or H. polygyrus infection, or schistosome ova exposure prevents or reverses the chronic Th1-type colitis of IL10 deficient (IL10−/−) mice (3)(6)

At least part of the protective process involves induction of regulatory-type T cells and cytokines in the host. Hp is a murine intestinal helminth. IL10−/− mice develop colitis spontaneously. T cells from the mesenteric lymph nodes (MLN) of Hp-infected IL10−/− mice abrogate established colitis when transferred into IL10−/− recipients (6). Helminth colonization induces FoxP3 expression in MLN and LP T cells. In a Rag-transfer colitis model of IBD, Hp required CD8+ T cells in vivo to reverse the disease process (7). Hp infection also elicits a regulatory T cell population able to down-regulate allergen-induced lung pathology (8). Also, after Hp infection, lamina propria (LP) T cells from healthy wild-type (WT) mice make large amounts of regulatory cytokines like IL10 and TGFβ (9).

Interactions with cells of the innate immune system could be part of the protective process. For instance, schistosomes protect BALB/c mice from DSS enteritis via a macrophage-dependent mechanism not requiring regulatory T cells (10). Protection in animal models of asthma may involve alternatively activated macrophages (11).

We used a Rag IL10−/− T cell transfer murine model of IBD to further explore the importance of innate immunity in Hp protection from IBD. This investigation showed that direct interaction alone with the innate immune system is sufficient to allow Hp to impede intestinal antigen-specific responses and to protect mice from colitis. Data suggest that changes in dendritic cell (DC) function contribute to this regulatory process.

MATERIALS AND METHODS

Mice

This study used C57BL/6 Rag2 mice, OT2 and IL10−/− mice (Jackson Laboratory, Bar Harbor, ME). Breeding colonies were maintained in SPF facilities at Tufts University. Animals were housed and handled following national guidelines and as approved by our Animal Review Committee.

Colitis model

Rag mice of similar age were reconstituted with 10⁶ IL10−/− splenic T cells and 3×10 ⁵ OT2 splenic T cells given ip. One week later, the animals were administered piroxicam (Sigma) mixed into their feed for 2 weeks (piroxicam at 40mg/250g chow wk 1, and 60mg/250 g chow week 2) to induce colitis. The piroxicam then was stopped. Two weeks later, the mice were sacrificed and their colons were examined microscopically for colitis, and lamina propria mononuclear cells (LPMCs) were isolated for culture (Figure 1).

Figure 1.

Experimental Design

Hp infection

Infective, ensheathed Hp L3 (U.S. National Helminthological Collection no. 81930) were obtained from fecal cultures of eggs by the modified Baermann method and stored at 4°C. Mice were colonized with 125 Hp third stage larvae by oral gavage.

For some experiments, animals were infected for 2 wks with Hp after induction of colitis (end of piroxicam treatment) and then sacrificed to assess colitis severity (Figure 1). In other experiments, Rag mice first were infected with Hp or just received a sham infection for 2 wks and then both the infected and control groups received a single dose of pyrantel pamoate (0.5 mg/mouse, Sigma, St. Louis, MO) via oral gavage to eliminate Hp. One wk after de-worming, the mice were reconstituted with T cells, and some were treated with piroxicam as described above to induce colitis, whereas others were observed without prioxicam treatment. In separate control experiments, de-worming was confirmed by documenting the absence of adult Hp in the small bowel 1 wk after receiving the drug.

Dispersion of splenocytes, and splenic T cell enrichment

Single cell suspensions of splenocytes were prepared by gentle teasing in RPMI 1640 medium (GIBCO, Grand Island, NY). The cells were washed three times in RPMI. Splenic T cells or CD4+ T cells were isolated by negative selection using the EasySep mouse T cell Enrichment Kit as outlined by the manufacturer (Stemcell Technologies, #19751, Vancouver, Canada). Viability was determined using exclusion of trypan blue dye.

LPMC isolation and LP cell fractionation

Gut LPMCs were isolated from the terminal ileum as described (6). Cell viability was 90% as determined by trypan blue exclusion. LP dendritic cells (CD11c+), macrophages (F4/80+) and NK cells (DX5+) were isolated from dispersed LPMC using appropriate mAbs and FACS. Cells were surface stained with flurochrome-labeled anti-CD11c, -F4/80 and/or -DX5 mAb before sorting (see below).

Cell culture

LPMCs from T cell-reconstituted Rag mice were cultured (2×10⁵ cells per well) for 48h in 96-well round-bottomed plates. Cells were cultured along with OVA (50 μg/ml) (Sigma) or anti-CD3 mAb (2C11; American Type Culture Collection, Manassas, VA) and anti-CD28 mAbs (BD PharMingen, San Diego, CA); each at 0.5 μg/ml. The culture medium was RPMI 1640 containing 10% FCS, 25 mM HEPES buffer, 2 mM L-glutamine, 5 × 10⁻⁵ M 2-ME, 1 mM sodium pyruvate, 100 U/ml penicillin, 5 mg/ml gentamicin, and 100 mg/ml streptomycin (all Life Technologies, Gaithersburg, MD). After culture, the supernatants were assayed for IFNγ and IL17A using ELISA (described below).

In the Rag LPMC/OT2 T cell mix experiments, OT2 Thy1.2 splenic T cells were mixed with LPMCs from Rag mice at the ratio of (1:3). Cells (2×10⁵) were cultured in RPMI complete medium for 48 h. Some cultured contained OVA at up to 1000 μg/ml to stimulate cytokine release. Supernatants were assayed for IFNγ and IL17 after the incubation using ELISA.

In some experiments, OT2 Thy1.2 T cells (1.5×10⁵) were mixed with isolated CD11c+, F4/80+ and/or DX5+ LPMCs (about 4×10⁴ of each) from Rag mice cultured 48 hours as above with up to 50 ug OVA before the supernatants were assayed for IFNγs and IL17.

In other experiments, isolated CD11c+ LP DCs were cultured for 48 hours alone with or without 100 ng/ml highly purified LPS derived from E. coli 026:B6 (Sigma, St. Louis, MO), or mixed with OT2 T cells and stimulated with OVA (50 ug/ml). Supernatants were assayed for IL12p40, IL10 and TGFβ using ELISA. Cells were again cultured in RPMI complete medium except for TGFβ determinations in which the cultures contained 1%, rather than 10%, FCS with 1 mg/ml albumin (Amresco, Solon, OH).

Sandwich ELISAs

ELISAs were performed using paired antibodies (R & D Systems, Minneapolis, MN) according to manufacturer’s instructions. IL17 ELISA was done using primary capture antibody from (R&D Systems, Inc) and biotinylated anti-IL17A antibody (R&D Systems). IL-12 (p40) ELISA was performed using primary capture antibody (BD Pharmingen, Franklin Lakes, NJ) and biotinylated secondary antibody (BD Pharmingen). IL10 was captured with anti-IL10 mAb (R&D Systems) and detected with biotinylated mAb (R&D Systems). To measure IFNγ, plates were coated with a mAb to IFNγ (HB170, ATCC) and incubated with supernatant. IFNγ was detected with polyclonal rabbit anti-IFNγ (gift from Dr. Mary Wilson, University of Iowa) followed by biotinylated goat anti-rabbit IgG (AXcell, Westbury, NY). Total TGFβ was measured using acid-treated supernatant and mAb240 for capture and biotinylated chicken IgY BAF240 for detection (both R&D Systems)

Flow cytometry analysis

LPMCs were washed twice and adjusted to 10⁷ cells/ml in FACS buffer (LGM) and stained with saturating amounts of conjugated mAb for 30 min at 4°C. Following staining, cells were washed twice and re-suspended in LGM for analysis on a FACSCalibur using Cell Quest software (BD Biosciences, Mountain View, CA). Before adding labeled mAb, each tube received 1 μg of anti-Fc mAb (eBioscience, San Diego, CA) to block nonspecific binding of conjugated Abs to FcRs. The mAbs used for staining or cell sorting were anti-Thy1.2-FITC, or - PECy5, or -APC; anti-CD11c-FITC; anti-F4/80-PE-Cy5; anti-CD4-PE or -PE-Cy5; anti-DX5-PE (all from eBioscience).

Statistical analysis

Data are means ±SE of multiple determinations. Difference between two groups was compared using Student’s t-test. Multiple group comparisons used analysis of variation and Dunnett’s t-test. P values <0.05 were considered significant.

Results

Development of an IBD model in which gut LPMC respond to OVA antigen

Rag mice reconstituted with IL10−/− T cells develop severe colitis after piroxicam exposure (12). Intestinal luminal antigens that drive LP T cells are poorly defined. To permit the study of antigen-specific responses in the intestine LP, we reconstitute Rag mice with IL10−/− T cells mixed with OT2 T cells bearing MHC Class II dependent TCR that recognize OVA. These mice were highly susceptible to colitis upon piroxicam administration. LPMC isolated from such mice cultured with OVA in vitro released large amounts of IFNγ and IL17, which were secreted minimally in the absence of OVA (Figure 2). LPMCs from mice reconstituted only with IL10−/− T cells did not respond to OVA suggesting that the cytokine response was OT2 T cell-dependent.

Figure 2.

Rag mice were reconstituted with IL10−/− T cells (10⁶) or IL10−/− T cells (10⁶) and OT2 T cells (3×10 ⁵) given i.p. One week later, piroxicam was mixed in their chow and administered for 2 weeks to induce colitis. Two weeks after stopping the piroxicam, LPMCs were isolated for culture. Cultures contained 2×10⁵ LPMCs/well maintained for 48 hours in vitro with or without OVA (50 ng/ml). ELISAs measured IL17 and IFNγ. Data are mean ±SE from 3 independent experiments. OT2 + OVA vs. all other comparisons, p<0.01.

Hp infection stops colitis and reduces the LPMC cytokine response to OVA

Hp can reverse colitis in Rag mice reconstituted with IL10−/− T cells (6)(13). Using the Rag IL10 KO/OT2 T cell reconstitution model of IBD, we determined if infection with Hp after induction of colitis could suppress intestinal inflammation and decrease cytokine release from LPMCs isolated from these mice. Figure 3 shows that Hp markedly suppressed the intestinal inflammation and that isolate LPMC from Hp-infected animals produced much less IFNγ and IL17 after OVA stimulation compared to the sham-infected controls. No IL4 was detected.

Figure 3.

Hp infection (125 larvae) reversed colitis, and LPMCs isolated from the TI of Hp-infected mice released less IFNγ and IL17 upon OVA stimulation (50 μg/ml). Rag mice were reconstituted with 10⁶ IL−/− KO T cells and 3×10⁵ OT2 T cells given i.p. One week later, piroxicam was mixed in their chow and administered for two weeks to induce colitis. After stopping the piroxicam, the treatment group was infected with Hp. The control group received no infection (Figure 1). Two weeks later, the colons were examined microscopically for colitis, and LPMCs were isolated for culture. Cultures contained 2×10⁵ LPMCs/well maintained for 48 hours in vitro with or without OVA (50 ng/ml). ELISAs measured IL17 and IFNγ. LPMC made little or no IFNγ or IL17 without OVA stimulation (<50 pg). Histology was scored blinded on a 4-point scale. All data are mean ±SE from 3 independent experiments. Control vs Hp p<0.01.

Additional experiments reconstituted Rag mice with IL10 KO and OT2 T cells as above, but one week later the animals were infected with Hp and never given piroxicam to induce colitis. Two wks after infection, LPMC isolated from the TI and cultured as in figure 3 also made much less IFNg and IL17 upon OVA stimulation (10 ug/ml) compared to the LPMC isolated from uninfected control animals. (IFNg 470+/−40 vs. 51+/−10 pg/ml; IL17 220+/−21 vs. 105+/−15 pg/ml; uninfected vs. infected, N=3, mean+/−SE).

Rag mice exposed to Hp only before T cell reconstitution are protected from colitis

Immune effector and regulatory pathways of both innate and adaptive immunity help drive colitis. Experiments explored if Hp interacting exclusively with cells of the innate immune system would suffice to control disease. Rag mice devoid of functional T and B cells were exposed to Hp for two weeks and then treated with a single dose of pyrantel pamoate to eliminate the worms. Control mice received a sham infection and drug treatment. One week after drug treatment, mice were reconstituted with IL10−/− and OT2 T cells, exposed to piroxicam for 2 wks and then sacrificed 2 wks after stopping the piroxicam (Figure 1). Figure 4 shows that Hp exposure before introduction of T cells was sufficient to block inflammation and to suppress both anti-CD3/CD28 mAb- and OVA-induced cytokine responses from the LPMC. Once more, there was no detectable IL4.

Figure 4.

The experiment was conducted similar to that of Figure 3 except that mice were exposed to Hp only before T cell reconstitution. Rag mice (5-6 weeks of age) were infected with Hp (125 larvae) for 2 weeks and then de-wormed by giving pyrantel pamoate. Control mice received the same drug. Mice were reconstituted with T cells 1 week after pyrantel pamoate treatment. Colitis was induced with 2 weeks of piroxicam treatment, and the colons were examined 2 weeks after stopping the piroxicam. At the end of the experiment, LPMCs isolated from the TI of mice with Hp pre-exposure released much less IFNγ and IL17 upon OVA (50 ng/ml) or anti-CD3/CD28 mAb stimulation (0.5 μg/ml) compared to the control group, and the mice were resistant to colitis. Data are mean± SE of 5 experiments (Control vs. Hp p<0.01)

Additional experiments determined if the reduced LPMC cytokine response to OVA was only secondary to improvement in inflammation. Following a brief infection with Hp and reconstitution of adaptive immunity as described above, mice were maintained for 3 weeks, but they were not given piroxicam to induce colitis. LPMCs isolated from these animals made substantially less IFNγ and IL17 and no IL4 after either OVA or anti-CD3/CD28 mAb stimulation in vitro compared to the uninfected controls (Figure 5).

Figure 5.

After Hp exposure, LPMCs isolated from mice without colitis released little IFNμ and IL17 upon OVA (50 ng/ml) or anti-CD/CD28 mAb stimulation compared to uninfected controls. The experiment was conducted as described in Figure 4 except that the animals did not receive piroxicam to induce colitis. There was no colitis in the colonic sections. Data are mean ± SE of 5 independent experiments. Control vs. Hp p<0.01.

In Rag mice without T cell reconstitution, Hp exposure alters intestinal DC function, cell surface protein expression and cytokine secretion

It was assumed that Hp interactions with cellular components of innate immunity caused the above described protection from colitis and alterations in cytokine production. To test this hypothesis, Rag mice were infected with Hp for two weeks. After two weeks of infection, LPMCs isolated from Rag mice were mixed with splenic OT2 T cells and then cultured in vitro with or with various concentrations of OVA to stimulate cytokine release. After incubation for 48 hours, the supernatants were assayed for IFNγ and IL17 content. LPMCs isolated from age-matched Rag mice that never had Hp infection served as controls. Figure 6 shows that LPMCs from Rag mice with previous Hp infection supported OVA-induced, IFNγ production poorly. IL17 secretion was minimally affected. LPMC cultured without OT2 T cells and OT2 T cells cultured without LPMCs released no detectable IFNγ or IL17 upon OVA or LPS stimulation. IL4 was not detected in any cultures.

Figure 6.

Rag mice were infected with Hp for 2 weeks. LPMCs were isolated from Hp-infected Rag mice (Hp LP) and from age-matched uninfected control Rag mice (no Hp LP). OT2 splenic T cells were mixed with Rag LP cells (ratio 1:3) (either No Hp LP or Hp LP) (a total of 2×10⁵ cells/well) and cultured with the indicated concentration of OVA for 48 hours. Supernatants were assayed for IFNγ and IL17 after the incubation. Data are mean ± SE of 3 experiments. LP cells cultured without OT2 T cells made no IFNγ or IL17 after OVA (100 μg/ml), anti-CD3/CD28 (each at 0.5 μg/ml) or LPS (100 ng/ml). *No Hp vs Hp p<0.01

The alteration noted in the supportive role of isolated LPMC in the OT2 T cell IFNγ response could have reflected Hp-induced changes in the cellular composition of the isolated, dispersed LPMCs. Flow analysis of the isolated LPMCs from Hp-infected and control mice revealed no significant alterations in the composition of dendritic cells (DC) (CD11c+), macrophages (F4/80+) or NK cells (DX5+) (Figure 7).

Figure 7.

Intestine was dispersed from either uninfected control (no Hp) or Hp-infected (Hp) Rag mice, and the cellular composition was analyzed with flow cytometry. Cells were stained for DCs (CD11c), macrophages (F480) and NK cells (DX5). A) Shows a representative FACS analysis. B) Depicts the cellular composition of the lamina propria showing only small differences between Hp-infected and control mice. Data are mean ± SD, N=4. No Hp vs Hp p NS

Additional experiments explored the potential role of these three cell types in modulating the OT2 IFNγ response. Using FACS, CD11c+ DCs were isolated from the dispersed LPMCs from Rag mice previously exposed to Hp or from mice that never received this infection. CD11c+ DCs were mixed with OT2 T cells and then cultured with OVA for 48 hours. CD11c+ intestinal DCs from mice previously infected with Hp were less proficient at supporting the IFNγ response compared to DCs from the controls, whereas there was no change in IL17 secretion (Figure 8). F4/80+ and DX5+ LP cells from either Hp-infected or control Rag mice mixed with OT2 T cells individually or in combination did not support OVA-induced IFNg or IL17 secretion. Adding F4/80+ macrophages or DX5+ NK cells to the CD11c+ LP DC/OT2 T cell mix did not effect OVA-induced, IFNγ or IL17 secretion either.

Figure 8.

Rag mice were colonized with Hp for 2 weeks, whereas control mice received only a sham infection. Isolated DCs (4×10⁴) were mixed with OT2 T cells (1.5×10⁵) and cultured in vitro for 2 days with OVA (50 ng/ml) to stimulate IFNγ and IL17 production. Data are mean± SE, N=3. *p<0.01

DCs support or modulate T cell responses by cell-to-cell contact and through secretion of cytokines. Flow cytometric analysis was used to study cell surface proteins associated with DC function. We examined CD11b, MHCII, CD8a, CD103, CD80, CD86, PDCA-1 and CD40. DCs from the TI of Rag mice exposed to Hp, compared to DCs from control Rag animals, displayed a relative decrease in surface expression of CD80 and CD86 with enhanced expression of PDCA-1 and CD40 (Figure 9). We also examined CD11c+ DCs in LPMC isolated from the colon. These cells showed a similar pattern of changes in cell surface protein expression.

Figure 9.

Rag mice were exposed to Hp for 2 weeks (Hp), whereas age-matched control animals received no infection (no Hp). Isolated LPMCs were analyzed by FACS for expression of the indicated cell surface proteins on the DC population. Each data set is from four independent experiments. Data are mean ± SD. *p<0.01

DCs can produce various cytokines like IL12, IL10 and TGFβ that can stimulate, modulate and/or inhibit T cell development and secretion. To see if Hp exposure altered the capacity of intestinal CD11c+ DCs to produce cytokines, DCs from Rag mice previously infected with Hp or DCs from control animals that never experienced this infection were cultured alone or with LPS for 48 hours to stimulate cytokine release. DC isolated from mice previously exposed to Hp released less IL12p40 and IL10 in response to LPS than did comparable DCs isolated from the control animals (Table 1). In other cultures, the DCs from either Hp-exposed or unexposed Rag mice were mixed with OT2 cells and then cultured with OVA (100 μg/ml). After OVA stimulation, supernatants from cultures containing Hp-exposed DCs also contained less IL10 and IL12p40. In all experiments, cellular supernatants did not contain measurable amounts of TGFβ.

Table I.

LP DCs from Hp-infected Mice Make Less IL12p40 and IL10

IL12 p40
	No Hp	Hp	p
CD11c	27+/−0.3	16+/−0.7	<0.01
CD11c + LPS	259+/−21	147+/−3.3	<0.01

IL10
	No Hp	Hp	p
CD11c	16+/−0.7	13+/−0.2	<0.05
CD11c + LPS	643+/−33	355+/−43	<0.01

Rag mice were colonized with Hp (Hp) for 2 weeks, whereas control mice received only a sham infection (no Hp). Isolated intestinal DCs (5×10⁴) were cultured in vitro for 2 days with or without LPS (100 ng/ml) to stimulate IL12p40 and IL10 production. Data are mean ± SE, N=3.

^*p<0.01

DISCUSSION

The most important observation of this study is that Rag mice briefly exposed to Hp before reconstitution with IL10−/− colitogenic T cells are protected from colitis. Moreover, the protection is as profound as that seen in Rag mice exposed to Hp after T cell reconstitution. T cells drive the pathology in this animal model of IBD (12). The data presented here suggest that Hp regulates the pathogenic T cell response that causes IBD via direct interaction with cellular components of the innate immune system requiring no direct interface with T cells.

The intestinal mucosal immune system responds to many molecules within the gut. It is difficult to study the control of T cell responses to intraluminal antigens in the lamina propria because a multitude of poorly defined antigens shape mucosal immunity. To overcome this limitation, we injected OT2 transgenic T cells, which recognize OVA in a Class II-dependent fashion, into the IL10−/− murine model of IBD. When introduced at the time of IL10−/− T cell transfer, OT2 T cells appear in the LP, and isolated LPMC cultured in vitro with OVA produce large amounts of IFNγ and IL17. IFNγ and IL17 are pro-inflammatory cytokines incriminated in driving colitis in both human and many murine models of IBD (13).

Exposure to Hp before introduction of IL10−/− and OT2 T cells reduced the capacity of the intestinal mucosa to make IFNγ and IL17 after either anti-CD3 mAb- or OVA-stimulation. This depressed cytokine response was evident even in the absence of colitis, suggesting that the down-modulation in pro-inflammatory cytokine secretion was not just secondary to improvement in intestinal inflammation.

It was assuming that innate immunity, influenced by Hp exposure, rendered mice resistant to colitis and the LP less prone to secrete colitogenic cytokines like IFNγ and IL17. The dispersed LPMCs from these mice were comprised mostly of DCs, macrophages and NK cells, which were the focus of this study. Hp infection did not alter the relative proportion of these three cell subsets in the dispersed LPMC preparations. Our in vitro studies using various combinations of these 3 cell subsets revealed that DCs poorly supported OVA-specific, IFNγ production after Hp exposure. Neither the macrophages nor the NK cells had a substantial effect on this antigen-specific response.

DCs within the intestines function to limit local immune responses to the mostly harmless luminal antigens (14). Disruption in DC function may be one of the factors promoting IBD. The aim of this study was not to categorize all the various DC subsets expressed in the gut after Hp infection, but to determine if Hp exposure substantially affected the phenotypic and functional expression of intestinal DC in Rag mice expressing no functional T or B cells. We examined the state of DC activation after two weeks of infection because the Rag mice were de-wormed at that time point.

Hp infection had a prominent effect on the capacity of DCs to display various surface proteins and to secrete cytokines. There was decreased expression of the co-stimulatory molecules CD80 and CD86 associated with a decrease in IL12p40 secretion. These alterations in DC function could explain why the intestinal DCs isolated from mice with Hp infection were less efficient at presenting OVA to OT2 T cells, stimulating IFNγ production. IL10 suppresses IL12 secretion, depresses Th1 cell function and has other immunoregulatory properties important for maintaining mucosal immune homeostasis (15). Hp exposure actually decreased rather than increased the capacity of the DC population to produce IL10, suggesting that IL10 did not participate in the down-regulation of DC CD80/86 expression, IL12 secretion or OT2 stimulation. Soluble TGFβ, another important mucosal immunoregulatory cytokine (15), and IL4 were not detected within culture supernatants, suggesting that these cytokines were not critical for this regulation. Our study did not rule out the possible participation of cell-surface bound TGFβ in the regulatory process.

PDCA-1 is a marker of plasmacytoid DCs and was more widely expressed on the intestinal DC after Hp infection. Activation of T cells by plasmacytoid vs. conventional DCs appears to favor generation of regulatory-type T cells (16). Thus, the appearance of more intestinal DC expressing PDCA-1 after Hp colonization could be of regulatory significance.

It is notable that Hp exposure did not affect the overall level of MHC Class II expression, which is centrally important for antigen-induced OT2 cell activation. Also notable was the failure of Hp infection to alter the number of DCs displaying CD8, CD11b and CD103, which are markers of intestinal DC subsets reported to influence Th and Treg polarization (14).

Expression of CD40 on DCs increased after Hp colonization. DCs express CD40 upon activation following interactions with some pathogens and microbial products (17). Some data suggest that the expression of CD40 on DC subsets is more dependent on the nature of the stimulus rather than on the phenotype of the DCs. Ligation of CD40 can result in enhanced expression of co-stimulatory molecules and pro-inflammatory molecules like IL12 and has been implicated in DC-mediated induction of effector T cell responses ((18)). The physiological significance of Hp-induction of CD40 on the intestinal DCs remains obscure.

It is assumed, although not yet proven, that changes in DC function lead to the protection. It still remains possible, although less likely, that that the intestinal DCs assimilated Hp antigens after the infection and that this retained antigen drove the protection.

Hp mostly inhabits the proximal small intestine in our Rag mice. We evaluated the function of DCs from the distal terminal ileum. It remains unknown how Hp communicated with the DCs in distal regions of the intestine. DCs can extend dendrites across the epithelial barrier to sample luminal molecules. Perhaps Hp antigens were sampled in the fecal stream. At least one secretory product of Hp, calreticulin, has been characterized (19). It binds to the scavenger receptor type A on DCs promoting Th2 responses. Crude supernatants containing excretory-secretory products from Hp blunt CpG-stimulated bone marrow-derived DC activation (20). These supernatants also impair DC-induced Ab responses, but enhance their ability to promote Treg development. These observations support the hypothesis that Hp modulated intestinal DCs function via release of immune modulatory molecules.

Hp infection induces a significant shift in the abundance and relative distribution of intestinal bacteria (in press). Among the various changes, there is a prominent increase in the Lactobacillaceae family of organisms, which contains bacterial species reported to decrease intestinal inflammation in murine models of IBD (21). We speculate that Hp also affects intestinal DC function through altering the complex intestinal bacterial flora.

As a group, a broad array of helminths and their products condition DCs through triggering some distinctly different intracellular signaling pathways to support Th2 rather than Th1 responses (22). The response of DCs to helminths is substantially different and much more blunted compared to their response to microbial pathogens. Helminths also impede DC maturation, which favors a tolerogenic response.

Helminth infection can antagonize the effects of bacterial products recognized via TLRs. This was demonstrated in our study showing that the LPS response of intestinal DCs was blunted after Hp infection. Helminths can signal via TLRs on DCs to promote a Th2/regulatory T cell response rather than the usual microbial-induced Th1 response (22). It is tempting to speculate that this could be one of their mechanisms of action.

The experiments reported here did not directly prove that Hp-induced changes in intestinal DC function protected the mice from colitis. However, there are ample data suggesting that DCs have an important role in driving both murine and human IBD (23).

Exposure to Hp impaired the capacity of the intestinal DCs to support an IFNγ response. Although IFNγ is important in colitis in many animal models of IBD (24), the profound blockade in intestinal inflammation observed here after Hp exposure likely is mediated through alteration of several important pro-inflammatory circuits. Other cells in the gut can have immune regulatory functions (25). In infection with Trichuris muris for instance, intestinal epithelial cells produce thymic stromal lymphopoietin, which impedes DCs from producing Th1- and Th17-promoting cytokines, favoring Th2 T cell development (26). Continuing studies are exploring several additional avenues of immune regulation potentially important for the control and prevention of colitis.

Abbreviations

APC

Antigen presenting cells

DC

Dendritic cells

Hp

Heligmosomoides polygyrus

IBD

Inflammatory bowel disease

LP

Lamina propria

LPMC

Lamina propria mononuclear cells

RPMI

RPMI 1640 medium

TI

Terminal ileum