Chronic Gastrointestinal Nematode Infection Mutes Immune Responses to Mycobacterial Infection Distal to the Gut

Control of mycobacteria and other intracellular infections of macrophages are dependent on the generation of Th1 cells. Th1 cells produce IFN-γ, which is required to activate macrophages for killing the infecting organism (1). Development of such responses can be measured by a delayed-type hypersensitivity (DTH) skin test reaction in both mice and humans. Indeed, the Mantoux test for tuberculosis (TB) and the Montenegro test for leishmaniasis are still used to screen for infection with Mycobacterium and Leishmania, respectively. Skin test reactivity may suggest the generation of protective immune responses, but depending on the size of the induration, can also be indicative of disease and warrant further examination (http://www.cdc.gov/tb/) (2).

The only available vaccine against TB is infection with live attenuated Mycobacterium bovis bacille Calmette–Gúerin (BCG), normally given in the skin. This infection/vaccination regimen has limited and highly variable efficacy in different parts of the world (3).

Helminth infections evoke Th2 and regulatory immune responses. Both of these responses can counteract Th1 development. Accordingly, worm infection is proposed to impair immune responses that control mycobacteria (4–6). Infection with worms has also been associated with a reduced ability to respond to BCG vaccination (7, 8). Geographically, areas of high TB incidence and poor TB vaccine efficacy typically have a high prevalence of intestinal helminth infections (9). However, the impact helminths have on vaccine efficacy and other secondary infections remains an open question. Indeed, a number of studies report a lack of correlation between intestinal worms and secondary infections (10–13). In common for many of the studies describing an association between worms and increased susceptibility to secondary infection, or reduced inflammatory response in experimental autoimmune disease, is that the effects have been observed in tissue(s) in direct or close contact with the worm (14, 15). In contrast, the effects of gastrointestinal (GI) worms on infections distal to the worm itself remain poorly characterized.

The nematode Heligmosomoides polygyrus bakeri (in this paper referred to as H. polygyrus) causes an infection strictly confined to the gut. In resistant hosts, H. polygyrus infection stimulates a strong Th2-type response that drives the expulsion of the worm (16, 17). Despite the generation of a protective Th2 response, the worm can persist and establish long-lasting infection in most laboratory mouse strains (reviewed in Ref. 18). This is facilitated by the regulatory responses H. polygyrus evokes. In the chronic phase of H. polygyrus infection, there is an expansion of regulatory Foxp3⁺ T cells in the gut (17). These regulatory Foxp3⁺ T cells, driven in part by a TGF-β–like activity released from the parasite (19), dampen effector T cell responses aiding persistent worm infection.

Chronic infestation with worms is the norm in animals and humans. Thus, H. polygyrus provides a relevant model to study the effects a gastrointestinal nematode infection has on immune responses to secondary infections. Furthermore, H. polygyrus only causes moderate intestinal pathology and the infection is typically asymptomatic in wild-type mice. Thus, secondary infections can be delivered in animals that are seemingly healthy. We used this model to investigate the effect of H. polygyrus infection on the outcome of mycobacteria-triggered Th1 responses at distal sites. Our results show that H. polygyrus infection can inhibit priming and recall responses to BCG and promote mycobacterial growth in vivo. Our data reinforce TGF-β signaling as a key component of H. polygyrus–mediated immune suppression. On the basis of our findings, we also suggest caution in the use of a skin-test reactivity–based diagnostic when performed in worm-infected individuals.

C57BL/6, congenic CD45.1 (Ly5.1), and P25-TCRTg RAG-1^−/− (20) × RAG-1^−/− ECFP (provided by Dr. R. Germain, National Institute of Allergy and Infectious Diseases, Bethesda, MD) were bred and maintained under specific pathogen-free conditions (Department of Microbiology, Tumor, and Cell Biology, Karolinska Institutet, Stockholm, Sweden). Female mice were used if not otherwise mentioned. All experiments were conducted in accordance to ethical regulations following approval by Stockholm’s Norra Djurförsöketiskanämnd.

All infections were performed in wild-type (C57BL/6 or congenic Ly5.1/CD45.1) mice.

At 4–5 wk of age, mice were infected by oral gavage with 200 H. polygyrus L3 larvae, obtained as described previously (21, 22). The worm infections were considered chronic after 28 d. At the end of each experiment, the worm burden was estimated by counting viable worms that had migrated out of the opened intestine through a fine net into a tube containing RPMI 1640 medium at 37°C within 3–4 h.

M. bovis BCG strain SSI 1331 was obtained from Statens Serum Institute (Copenhagen, Denmark), expanded in 7H9 medium as previously described (23), and inoculated at 1 × 10⁶ CFU in the footpad, ear pinnae, or i.v. into the tail vein. For quantification of mycobacterial load in tissue, single-cell suspensions were plated onto 7H11 agar supplemented with OADC (BD Biosciences) and cultured at 37°C for 21 d.

Leishmania major, Freidlin (a gift from Dr. D. Sacks, National Institute of Allergy and Infectious Diseases), was maintained in M199 supplemented with 20% FCS, 2 mM l-glutamine, 100 U/ml penicillin, and 200 mM streptomycin. A total of 1 × 10⁵ metacyclic promastigotes, enriched form stationary cultures using Ficoll 400 gradient separation (24), were injected into the ear dermis. Control animals were either left untreated or injected with PBS/medium as indicated in the figure legends.

Detection of promastigote growth was done by microscopy. To determine the burden of Leishmania parasites, tissue homogenates were prepared as previously described (25) and cultured in limiting dilutions in 96-well plates using M199 medium supplemented with 20% FCS, 2 mM l-glutamine, 100 U/ml penicillin, and 200 mM streptomycin.

To enumerate and measure granulomas, livers and spleens from infected mice were fixed in 4% paraformaldehyde in PBS for ≥12 h, followed by dehydration and paraffin embedding. The tissue was cut into 6-μm-thick sections and stained with H&E. Processing and H&E staining of paraformaldehyde-fixed samples was performed at the Cancer Center Karolinska, Pathology Laboratory, Karolinska University Hospital (Stockholm, Sweden). Analysis was performed by microscopy on 10 fields/sample at a final magnification of ×100. To allow pooling of granuloma area, the individual experiment was normalized by dividing the sample value with the mean value of all samples in the experiment.

Adult worms were isolated from the small intestines by allowing the worms to migrate through a fine mesh into a 50-ml tube with RPMI 1640 medium with 100 U/ml penicillin, 200 mM streptomycin, and 0.5 mg/ml gentamicin at 37°C for 3–4 h. Collected worms were washed extensively, incubated overnight in RPMI 1640 medium containing antibiotics, followed by further washing in the same medium. The washed worms were used for preparation of whole soluble worm Ag (SWAg) by resuspending adult worms in 10 mM Tris-HCl and adding complete protease inhibitor (Roche), according to the manufacturer’s instructions. The worms were homogenized by repeated freeze-thaw cycles, followed by mechanical disruption and pulsed sonication. After centrifugation, protein concentration in the supernatant was determined by OD at 280 nm. For injection of larval Ags, L3 larvae were collected from cultures and washed extensively in water containing 200 U/ml penicillin and 400 mM streptomycin. The “PPD” footpad was conditioned with 200 L3 larvae, as previously described by others (26). Control mice were injected with the water used in the last wash of the L3 larvae preparation.

H. polygyrus excretory–secretory (HES) products were isolated from adult worm cultures as previously described (27). Purified protein derivative (PPD; Statens Serum Institute) was used to measure recall responses in vivo by injecting 10 μg PPD in the footpad. Whole freeze-thawed Leishmania Ag (LAg) (28) was prepared from stationary-phase promasitgote cultures by repeated freeze-thawing cycles. Protein concentration was measured at 280 nm, and 50 μg Ag was injected in the footpad.

All Ags injected in the footpad were delivered in a volume of 30 μl.

For in vivo assessment of Ag-specific T cells, total lymph nodes (LN) from P25-TCRTg RAG^−/− mice were collected, and single-cell suspensions were prepared. Isolated P25 TCRTg (P25) cells were labeled with 1 μM CFSE (Invitrogen) in PBS for 10 min at room temperature. The labeling reaction was stopped by adding FCS and cells were subsequently washed in RPMI 1640 medium supplemented with 10% FCS, 100 U/ml penicillin, and 200 mM streptomycin to remove any unbound dye. A total of 1 × 10⁵ CFSE-labeled P25-TCRTg T cells (CD45.2⁺) were injected i.v. into the tail vein of recipient (CD45.1⁺) congenic mice receiving BCG. Control animals received 1 × 10⁶ cells.

Single-cell suspension of the tissue of interest was prepared as described previously (23). Isolated cells were diluted in RPMI 1640 medium supplemented with 10% FCS, 100 U/ml penicillin, and 200 mM streptomycin to a concentration of 2 × 10⁶ cells/ml and stimulated in 96-well tissue culture plates with 10 μg/ml PPD, 2 μg/ml purified anti-CD3 Ab (BD Biosciences), or worm Ags as indicated in figures for 3 d at 37°C in 5% CO₂. Irradiated (30 Gy) splenocytes from naive mice were used as APCs at a ratio of 1:5. Supernatants were collected and stored at −80°C until cytokines were measured by ELISA.

Cytokine levels in culture supernatants were measured by sandwich ELISA. Immulon 2B plates (Nunc) were coated overnight, 4°C, with capture Ab. For IFN-γ, the capture Ab was diluted in carbonate buffer (0.1 M Na₂CO₃, 0.1 M NaHCO₃, and 1 mM NaN₃ [pH 9.6]), all other capture Abs were diluted in PBS. The plates were then blocked for 2 h at 37°C with 5% milk in diluent solution (1% BSA and 0.05% Tween 20 in PBS) and subsequently incubated overnight, 4°C, with the sample supernatants or the respective standards. Bound protein was detected with biotin-labeled detection Abs for 2 h at 37°C, followed by incubation with peroxidase-labeled streptavidin (2 h, 37°C) and development with ABTS peroxidase substrate (both KPL). Plates were read at 405 nm. The following Ab clone pairs were used: IFN-γ (R4-6A2/XMG1.2), IL-5 (TRFK5/TRFK4), and IL-10 (JES5-2A5/JES5-16E3) all from BD Biosciences and Jackson ImmunoResearch Laboratories. For detection of IL-17 and TGF-β, mouse Quantikine kits (R&D Systems) were used according to manufacturer’s instructions.

P25-TCRTg cells were isolated and in some experiments CFSE labeled, as described above. Splenic dendritic cells (DCs) were isolated from collagenase IV and DNAse I–treated spleens as previously described (23), followed by magnetic enrichment of CD11c⁺ cells using CD11c microbeads (Miltenyi Biotec), according to manufacturer’s instructions. Splenic DCs were incubated with HES and SWAg at various concentrations for 2 h and subsequently cocultured with P25-TCRTg at a ratio of 1:5 and stimulated with multiplicity of infection 1 of BCG 1331. In some experiments, HES and SWAg were given to DCs in the presence of 5 μM ALK5 inhibitor SB-431542 (Torics), treatment with 5 ng recombinant human TGF-β (R&D Systems) ± SB-431542 was used as control. Positive control stimulations were with 2 μg/ml Ag85B_240–254 peptide and 100 ng/ml LPS (Sigma-Aldrich).

Single-cell suspensions from tissues were incubated with various combinations of fluorochrome-conjugated rat anti-mouse mAbs specific for CD4 (RM4-5), CD11b (M1/70), CD11c (HL3), MHC class II (MHC-II) I A/I E (M5/114.15.2), CD44 (IM7), CD45.2 (104), CD69 (H1.2F3), (BD Biosciences), CD326/EpCAM (G8.8), CD103 (2E7), (BioLegend), CD4 (RM4-5), B220 (RA3-6B2), and latency-associated peptide (TW7-16B4) (eBioscience), for 45 min in FACS buffer (2% FCS in 5 mM EDTA and 0.1% azide) containing 0.5 mg/ml anti-mouse FcγIII/IIR (2.4G2) (BD Biosciences). Live-dead staining was done using live-dead dye (Life Technologies). For analysis of intracellular cytokine production, cells were stimulated ex vivo for 6 h with 10 μM Ag85B_240–254 peptide in the presence of 10 μg/ml brefeldin A (Sigma-Aldrich) prior to surface staining, followed by fixation in 2% paraformaldehyde (Electron Microscopy Sciences) and permeabilization with 1% saponin (Sigma-Aldrich) and staining with anti–IFN-γ (XMG1.2) (BD Biosciences). For staining of Foxp3, we first surface stained cells and then prepared cells for intranuclear/-cellular staining using eBioscience Foxp3 staining set (FJK-16S), according to the manufacturer’s instructions. Irrelevant isotype-matched Abs were used to determine levels of nonspecific binding. Cell proliferation was measured by CFSE dilution.

To track cell migration from the skin to the draining LN 20 μl 0.5 mM CFSE in PBS were injected into the same footpad in which BCG vaccination had been delivered 48 h earlier, as described previously (29). The popliteal lymph node (pLN) was collected 24 h after CFSE injection. Single-cell suspensions of pLN were stained for expression of CD11c, CD11b, MHC-II, CD103 and CD326. Detection and phenotypic characterization of CFSE^hi cells were done by FACS. FACS acquisition was performed using CyAn (Beckman Coulter), LSRII or FACSCanto (BD Biosciences). Analysis was done on single cells gated as lymphocytes by forward-side scatter using FlowJo software (Tree Star).

Total mRNA was isolated from tissue using TRIzol (Sigma-Aldrich) according to manufacturer’s instructions. RNA concentration was determined by spectrometry and first strand cDNA generation and real-time PCR was performed as previously described (29, 30), using the T100 and CFX 384 Real-time System (Bio-Rad), respectively. Expression of HRPT was used as baseline and the relative expression of gene expression was determined as follows: ΔCT between gene of interest and HRPT in the sample/ΔCT between gene of interest and HRPT in the in assigned unstimulated (control) sample.

Unpaired Student t test was used for comparison between two groups using Prism (version 5.0a; GraphPad Software). Outliers were excluded from analysis of following Grubb’s test for one outlier, α = 0.05 (GraphPad, QuickCalcs). A p value < 0.05 was considered to indicate significant differences between the groups.

To test the effect of an underlying H. polygyrus infection on induction of mycobacteria-specific CD4 T cells, we infected mice i.v. with BCG and assessed activation of P25-TCRTg cells in animals with or without worms introduced 28 d previously. P25-TCRTg cells recognize the major CD4 T cell epitope of Ag 85B, present in both BCG and M. tuberculosis (20).

Ag-specific T cell priming was clearly impaired in BCG-infected mice with chronic worm infection compared with animals free of worms. The expansion and proliferative capacity of mycobacteria-specific P25-TCRTg cells was substantially reduced in H. polygyrus–coinfected animals, as measured by total number (Fig. 1A) and percentages (Fig. 1B) of P25-TCRTg cells in the spleen. This was accompanied by a reduced ability of P25-TCRTg cells to produce IFN-γ upon restimulation with Ag85B (Fig. 1C, 1D). Although separated, the spleen and the gut are still in relative proximity, and Th2 responses to H. polygyrus can be detected in the spleen (28). To increase the distance between the site of worm infestation and the secondary infection, we delivered BCG in the footpad. Similar to the observations described above, fewer P25-TCRTg T cells and fewer IFN-γ⁺ P25-TCRTg T cells were found in the skin-draining pLN 6 d after BCG footpad infection in mice with chronic H. polygyrus infection compared with worm-free mice (Fig. 1E–G). Furthermore, the cellularity of the BCG-draining pLN was notably lower in H. polygyrus–infected mice compared with worm-free mice (lymphocytes per pLN 6 d after BCG infection; H. polygyrus/BCG: 4.56 × 10⁶ ± 8.68 × 10⁵; BCG: 13.15 × 10⁶ ± 2.48 × 10⁶, n = 5/group; p = 0.0115, one of three or more experiment with similar results). The viability of CD4 T cells in the BCG-draining LN was high (95%) and similar between H. polygyrus–infected and worm-free mice. This indicates that the impaired T cell response observed in worm-infected mice is not due to increased T cell death. These observations demonstrate that an intestinal worm infection can affect T cell priming to infection/vaccination at sites distal to the gut.

To investigate the effect of H. polygyrus on recall responses we measured DTH to PPD in BCG-infected/vaccinated animals with or without worm infection. In these experiments, BCG infection was delivered in the right footpad, and the magnitude of the DTH response was measured 2 wk later by recording swelling upon PPD delivery in the other footpad. This measurement can be seen as an equivalent of the PPD skin tests done in humans to assess BCG vaccination or TB infection. In addition, we also studied the effect of depositing worm Ags proximal to the site of PPD challenge. To that end, footpads were preconditioned by injecting H. polygyrus larvae 4 wk prior to PPD restimulation.

Interestingly, footpad swelling was significantly reduced when PPD was given in footpads preconditioned with L3 larvae (Fig. 2A). However, no significant effect on footpad swelling was observed in mice tested for PPD reactivity 4 wk after worm infection and 2 wk after BCG infection, compared with BCG-infected, worm-free mice (Fig. 2A). Because most helminth infections are chronic in nature, we prolonged the time to BCG infection, allowing the worm to become chronic before infection with BCG was given. Interestingly, we now found that the DTH responses were significantly diminished in worm-infected animals (Fig. 2B), implying that it takes time before the impact of worm-mediated inhibition can be observed on immune responses at distal sites.

When the PPD injection site was preconditioned with L3 larvae, in vitro IFN-γ recall responses to PPD in the draining LN were reduced by ∼65% compared with control LN (data not shown). However, this was not seen in LN from worm-infected animals. There were also no significant effects of H. polygyrus infection on in vitro production of IL-5, IL-12, or IL-17 in LN cultures (data not shown).

To determine whether fewer total T cells and Ag-specific CD4⁺ T cells accumulated at the site of recall response in worm-infected mice compared with worm-free mice, we measured the infiltration of total and P25-TCRTg CD4⁺ T cells in the skin where PPD was delivered. To facilitate the isolation of cells from skin, PPD was in these experiments injected in the ear pinnae. We found that worm-infected mice had fewer T cells and fewer Ag-specific P25-TCRTg cells in the skin (Fig. 2C, 2D), suggesting that the impaired DTH response is linked to a reduction of Ag-specific T cells in the PPD-challenged tissue. There was, however, no difference in the number or frequency of P25-TCRTg cells in PPD-draining auricular LNs (Supplemental Fig. 1A, 1B).

In line with the observations that both T cell priming and recall responses were reduced in H. polygyrus–infected mice, we found that these animals had a 2.3-fold higher bacterial load in the liver when given a systemic BCG infection (Fig. 3A). Interestingly, BCG-induced granulomas were both fewer and smaller in H. polygyrus–infected mice compared with worm-free mice (Fig. 3B, 3C). This indicates that worm-infection can interfere with granuloma formation, leading to an increased susceptibility to mycobacterial infection. The impaired control of mycobacteria was coupled to reduced infiltration of both DCs and macrophages and lower expression of inducible NO synthase, IFN-γ, and TNF-α in livers of coinfected animals (Fig. 3D–H). Assessment of CFU in spleen also indicated an increase in bacterial load in worm-infested animals, although this did not reach statistical significance (data not shown).

The impaired immune response and control of BCG in worm-infected mice was not associated with an increase in Foxp3 expression or more Foxp3⁺ CD4 T cells (Supplemental Fig. 2A–C).

The numbers of viable worms were similar in mice infected with H. polygyrus alone and H. polygyrus–infected mice given BCG i.v. (Fig. 3I). Thus, the established worm infection was not affected by the superimposed BCG infection in our model.

To investigate whether another Th1-controlled infection was affected by an underlying, chronic H. polygyrus infection, animals were coinfected with L. major in the ear. We found that H. polygyrus–infected mice had double the amount of parasites in the ear (Fig. 4A) and a reduced cellularity in the ear-draining LN (Fig. 4B) compared with worm-free mice. In vitro recall response to LAg by an equal number of LN cells were similar in worm-free and worm-infected mice (data not shown), indicating that a more sensitive model (e.g., one involving TCRTg cells) is needed to detect such differences. In line with the observations shown in Fig. 2, H. polygyrus– and L. major–coinfected mice had diminished DTH responses to LAg in the footpad compared with worm-free mice (Fig. 4C). This shows that a pre-existing intestinal helminth infection can influence host control of secondary infections occurring distal to the worm.

To avoid host immune responses and to establish chronic infection without causing detrimental pathology, helminths produce and secrete molecules with immunomodulatory capacity (27). Such molecules could, directly or indirectly, be the cause of the diminished T cell priming in response to BCG observed in worm-infected mice. To test whether worm-derived products interfere with T-cell priming, splenic DCs were treated with SWAg or the HES prior to stimulation with BCG and coculture with P25-TCRTg cells. SWAg was a relatively poor inhibitor of BCG-induced P25-TCRTg-cell IFN-γ production, and significant inhibition was only seen when a high concentrations of SWAg concentration was used (Fig. 5A). SWAg also inhibited BCG-induced IL-5 production in a dose-dependent manner (Fig. 5B), reinforcing the general downmodulation of effector–T cell responses by worms. HES strongly inhibited IFN-γ production by P25-TCRTg cells, whereas pre-exposure to BSA or LAg did not significantly affect IFN-γ production by P25-TCRTg CD4⁺ T cells (Fig. 5C), indicating that the effect observed was specific to HES. IL-10 was not detected in BCG-stimulated cultures and was not induced by HES or SWAg (data not shown).

The activity of HES was sensitive to heat treatment (Fig. 5D). HES has been shown to have a heat-sensitive TGF-β–like activity (19), suggesting that the TGF-βR activating molecule of HES-mediated inhibition of P25-TCRTg-cell IFN-γ production. In support of this, the effect of HES was reverted by inhibition of TGF-βR signaling, using the TGF-βR1/ALK5 inhibitor SB431542 (Fig. 5E). The SWAg-mediated effect also appeared to involve TGF-βR signaling because the inhibition of BCG-induced responses was reversed by blockade of TGF-βR signaling (Fig. 5E). These results thus propose a role for TGF-βR signaling in worm-mediated inhibition of T cell priming to BCG. The inhibitory effect of worm Ags in these in vitro cultures were likely through a direct inhibition on the T cells because both HES and SWAg inhibited anti-CD3/anti-CD28–induced T cell activation in absence of DCs (data not shown). Furthermore, it was not enough to simply preincubate (4 h) DCs with HES prior to coculture with T cells to observe the inhibitory effect on IFN-γ production. That said, direct effects of worm Ag on DCs have been shown by others (31) and cannot be excluded.

Given the above, we decided to investigate if there were effects of H. polygyrus on APC in vivo. In this regard, migration of APC from the site of infection to the draining LN is fundamental for the initiation of a primary T cell response. The number of APC reaching the LN will accordingly influence the magnitude of that response (32). Given that BCG-specific T cell expansion was reduced in worm-infected animals, we tested the effect of chronic H. polygyrus infection on migration of APC from the site of BCG infection to the draining LN. To track migratory cells, we injected CFSE into the skin of the footpad and quantified the number of CFSE⁺ cells detected in the draining pLN. Following BCG footpad infection we found that the majority of migrating (CFSE^hi) cells were MHC-II^hiCD11c^+/int, consistent with the phenotype of migratory DC (Fig. 6A). The majority of these cells were also CD11b⁺ (Fig. 6A). To address whether H. polygyrus affected such cell migration, we measured the number of CFSE^hiMHC-II^hiCD11c^+/int cells in the pLN postinfection with BCG. Mice with chronic H. polygyrus infection had fewer CFSE^hiMHC-II^+/hiCD11c^+/int cells in the pLN after BCG injection compared with worm-free mice (Fig. 6B, 6C). This may in part explain the reduction in Ag-specific T cell response found in the worm-infected mice (as shown in Fig. 1). TGF-β is a key cytokine in H. polygyrus–induced immune regulation (33), with the capacity to block Ag-specific CD4–T cell activation, as shown above, and with the potential to downmodulate the migratory capacity of DCs (34). To test whether worm-derived molecules and TGF-β could influence BCG-induced DC migration, we conditioned the footpad with TGF-β or HES 10 d prior to BCG inoculation. As a control, we heat-inactivated HES, thereby eliminating its TGF-β–like activity (19). In line with studies of skin cancer (34), our data suggest that TGF-β lowers the migratory capacity of DCs. Fewer CFSE^hiMHC-II^hiCD11c^int/+ cells were found in the draining pLN after BCG infection when the site of BCG injection was preconditioned with TGF-β (Fig. 6D). HES conditioning of the footpad also reduced the migratory response of MHC-II^hiCD11c^int/+ cells to BCG, whereas heat-inactivated HES was found to be a less potent inhibitor of the same (Fig. 6D). This finding supports worm-driven TGF-β signaling as a potential mechanism underlying impaired responses to BCG.

The efficacy of BCG vaccination is highly variable in different parts of the world. Worms, through the immune responses they evoke, have been suggested as one factor that can impair BCG vaccine efficacy and increase susceptibility to mycobacterial infection (6, 35). In support of this, there is ample evidence that worms and worm products can counteract Th1 immunity and downmodulate inflammatory responses to secondary Ags (36). However, most of these observations have been made in systems where the worm or the worm Ag is either in direct contact with or in close proximity to the site of inflammation/infection. Yet, most parasitic worms live in the gut and are therefore not proximal to the site where injection-based vaccines are delivered. Although some findings indicate that intestinal worms have more systemic effects on immunity (7, 8, 37, 38), evidence that intestinal nematodes modulate immune response in tissue distal to the worms and thereby impair immune responses to vaccination and secondary infections is scarce and remains questioned (12, 26, 39, 40).

In this study, we have experimentally addressed how a nematode infection confined to the gut influences Th1 responses to secondary infections at sites separated from the worm infection. Our results support the view that intestinal worms diminish immune responses to secondary vaccinations and infections. We found that CD4⁺ T cell priming in response to BCG was reduced in mice chronically infected with H. polygyrus, compared with worm-free animals. Likewise, pre-exposure to H. polygyrus Ags in vitro dramatically decreased IFN-γ production by mycobacteria-specific CD4⁺ T cells.

Similar to observations made in humans (8), we found that DTH responses to PPD following footpad BCG infection/vaccination and to L. major Ag following experimental leishmanization were smaller in mice with chronic H. polygyrus infection compared with worm-free mice. This may reflect vaccine efficacy, but more importantly, it suggests that intestinal worms can influence skin test–based diagnostics and possibly reduce DTH-based detection of TB and leishmaniasis. The notion that worms decrease the sensitivity of recall response-based diagnosis is something that would need more careful investigation in clinical studies.

The proximity between worm and coinfections appear to influence the suppressive effects of H. polygyrus on responses to BCG. The inhibitory effect was most evident when the distance to the BCG effector site was small [e.g., when BCG was delivered systemically (i.v.) or when the effector site itself had been preconditioned with worm Ags (Fig. 2A)]. Time is another factor that may influence the inhibitory effects of H. polygyrus. The DTH responses to PPD were attenuated in mice with chronic (>4 wk) but not acute (2 wk) worm infection, at the time when infected with BCG.

Systemic dissemination of worm-induced Th2 or regulatory T cells could explain how an infection confined to the gut can modulate immunity at peripheral sites. During the first weeks of H. polygyrus infection, worm-induced Th2 cells can be found in the spleen (28, 37). An increase in Th2 cells could underlie the reduced priming of BCG-specific CD4⁺ T cells observed in the spleen. Moreover, using IL-4 reporter mice, Mohrs et al. (37) found that H. polygyrus–induced Th2 cells that spread systemically have a preference for nonlymphoid organs, such as the liver. This could in turn explain why the impact of H. polygyrus on BCG load was more evident in the liver compared with the spleen. We could, however, not find any evidence for dissemination of Th2 cells to pLN in mice with chronic H. polygyrus infection. There were no differences in mRNA expression of IL-4 or T cell transcription factors GATA-3, T-bet, and Foxp3 in pLN when comparing H. polygyrus–infected and worm-free mice (data not shown). However, FACS analysis did reveal a modest decrease in the percentage of T cells in skin-draining LN of mice with chronic H. polygyrus infection (percentage of lymphocytes gated as T cells in pLN; naive mice: 60.9 ± 0.9%; mice with chronic H. polygyrus infection: 54.8 ± 1.5%, n = 10, p = 0.0034). This indicates that an intestinal worm may affect lymphocyte composition in skin-draining LN, which in turn could influence the subsequent ability to respond to infection/vaccination.

Reduced immunogenicity of BCG in people chronically infected with worms has been associated with increased production of TGF-β by PBMC (7). Indeed, TGF-β is also induced by H. polygyrus infection. Although primarily found at the site of infection, increased levels of serum TGF-β have also been reported in H. polygyrus–infected mice (41). We found that splenic CD4 cells from mice with chronic H. polygyrus infection express more TGF-β latency-associated peptide (LAP) compared with naive mice (S2 D-F). TGF-β is a pluripotent, mainly anti-inflammatory, cytokine, which can limit both Th1 and Th2 responses. H. polygyrus and other intestinal worms exploit this cytokine to facilitate the chronic establishment in the host (18, 42). Experimentally, TGF-β has been found to be important for the control of worm-mediated inhibition of several inflammatory diseases (43). Many nematode species express TGF-β homologs (44) and some, including H. polygyrus, secrete products that can signal through the TGF-βR (19, 45). HES from H. polygyrus has previously been shown to drive regulatory (Foxp3) T cell responses and to increase production of both TGF-β and IL-10, another immune-modulatory cytokine (19). We found no evidence for an increase in Foxp3 expression in H. polygyrus coinfected animals. While we cannot exclude involvement of IL-10, using a IL-10 GFP reporter mice (46) we did not find more IL-10 expressing cells in the BCG-draining pLN in worm infected compared with worm free mice (not shown).

However, we found that HES can act directly on T cells and that the TGF-βR signaling capacity of HES was needed to inhibit mycobacteria-specific T cell priming in vitro. DCs typically acquire regulatory properties in the presence of TGF-β (47). Although the effect of HES in our in vitro cocultures was mainly on the T cells, H. polygyrus and the excretory–secretory products are also know to facilitate regulatory and Th2-promoting DC (31, 48–50). Interestingly, we found that preconditioning the site of BCG infection with TGF-β or HES significantly reduced BCG-triggered migration of DC (MHC-II^hiCD11c^+/int cells) to the draining LN. TGF-β can downmodulate the expression of CCR7 on DCs and inhibit their migratory capacity. In mesenteric LNs, Léon et al. (51) found that H. polygyrus infection alters the expression of CCR7 and CXCR5 on DCs, with implications on development of downstream Th responses. We did not find that HES affected the CCR7 or CXCR5 expression in the BCG-draining LN (data not shown). The inhibitory effect of HES on DC migration may involve more than TGF-β because HES is a complex mixture (27). Recently, we found that the levels of CCR7 do not correlate with the in vivo ability of BMDC to migrate in response to BCG (29). Interestingly, IL-12p40 was found to be important for the BCG-induced DC migration from footpad to the draining LN (29). Others have shown that exposure of DCs to HES products downmodulates their expression of IL-12p40 (31). Similar observations have been made following exposure of DCs to TGF-β (52). In line with a role for TGF-βR signaling in HES-mediated inhibition of DC migration, heat inactivation of HES, which destroys the TGF-β–like activity of HES, significantly diminished the inhibitory effect of HES on DC migration. Whether this inhibitory effect on migration involves downregulation of IL-12p40 by excretory–secretory products remains to be investigated.

Taken together, our data support worm-induced TGF-βR signaling as a mechanism behind helminth-mediated immune modulation of effector–T cell responses.

In summary, we show that a chronic worm infection confined to the gut impacts both primary and recall immune responses to secondary microbial challenge delivered in tissue distal to the gut. Compared with worm-free animals, mice with a chronic intestinal nematode infection had impaired T cell priming in responses to BCG, reduced DTH responses in the skin, and a higher bacterial/parasite load when infected with BCG and L. major, respectively. This implies that worms negatively can affect the diagnosis as well as the control of intracellular infections with Mycobacterium and Leishmania. We propose worm-evoked TGF-βR signaling as a part of the explanation as to why helminth-infected individuals are more susceptible to Th1-controlled infections and respond less well to immunizations dependent on such responses.

We thank Frank Heuts, Viviana Taylor, Zachary Darroch, Damïen Bierschenk, and Adrian Luscombe for contributions to the study and the Department of Microbiology, Tumor and Cell Biology animal facility for technical support and care of animals.

The authors have no financial conflicts of interest.