The mannose receptor binds Trichuris muris excretory/secretory proteins but is not essential for protective immunity

Abstract

Trichuris muris is a natural mouse model of the human gastrointestinal nematode parasite Trichuris trichiura and it is well established that a T helper type 2-dominated immune response is required for worm expulsion. Macrophages accumulate in the large intestine of mice during infection and these cells are known to express the mannose receptor (MR), which may act as a pattern recognition receptor. The data presented here show for the first time that T. muris excretory/secretory products (E/S) induce bone-marrow-derived macrophages (BMDM) to produce several cytokines and have MR-binding activity. Using alternatively activated BMDM from MR knockout mice it is shown that the production of interleukin-6 partially depends on the MR. Infection of MR knockout mice with T. muris reveals that this receptor is not necessary for the expulsion of the parasite because MR knockout mice expel parasites with the same kinetics as wild-type animals and have similar cytokine responses in the mesenteric lymph nodes. Furthermore, despite acting to reduce serum levels of proinflammatory mediators, absence of the MR does not lead to increased gut inflammation after T. muris infection when assessed by macrophage influx, goblet cell hyperplasia and crypt depth. This work suggests that, despite binding components of T. muris E/S, the MR is not critically involved in the generation of the immune response to this parasite.

Introduction

Trichuris muris is a natural mouse model of the gastrointestinal nematode parasite Trichuris trichiura, one of the most prevalent human helminth infections. Strain variation in the mouse provides a spectrum of phenotypes varying from resistant strains such as BALB/c and C57BL/6, where the parasite is quickly expelled, to a completely susceptible strain such as AKR, where fecund adult worms survive in the intestine for many weeks or months.¹ It is well established that a T helper type 2 (Th2) -dominated immune response, characterized by the production of interleukin-4 (IL-4), IL-5, IL-9 and IL-13, is required for worm expulsion,^2–5 whereas the development of a Th1 response associated with high levels of interferon-γ (IFN-γ) and IL-12 leads to host susceptibility.^6,7

More recently, it has been established that large numbers of macrophages are found in the intestines of infected mice⁸ and that more macrophages accumulate in the lamina propria of resistant mice than susceptible mice;⁹ however, a functional role for these cells is yet to be established. Macrophages are pleiotropic cells capable of converting between a number of different phenotypes of which the classical and alternatively activated are the best described. Classically activated macrophages are important for the control of intracellular pathogens whereas alternatively activated macrophages (AAMφ) are proposed to play a role in type 2 cytokine-dominated immune responses, such as those to some parasitic infections.^10,11 Generation of AAMφ by exposure to IL-4 and/or IL-13 induces the expression of a distinct set of genes, including Fizz1, Ym1¹² and Arg1,¹³ some cytokines and chemokines including IL-10¹⁴ and CCL9,¹⁵ and surface markers such as CD23 (FcεRII)¹⁶ and the mannose receptor (CD206).^17,18 As resistant mice mount a type-2-dominated immune response, including the production of IL-4 and IL-13, it is possible that a proportion of the macrophages present in the large intestine of mice that expel the parasite will be of the alternatively activated phenotype.

The mannose receptor (MR) is a member of a family of endocytic receptors containing eight to ten lectin-like domains. Structurally, the MR comprises a cysteine-rich domain, a fibronectin type two repeat domain, eight C-type lectin-like carbohydrate recognition domains, a transmembrane region and a short cytoplasmic tail.¹⁹ The MR is detectable on macrophages throughout the body but expression is not limited to these cells because it is also found on dendritic cells and lymphatic endothelia as well as other cells.^19,20 Physiologically, the MR was first shown to be involved in the regulation of endogenous glycoprotein levels,²¹ an activity confirmed by the study of MR knockout (KO) mice.²² Subsequently, other immunological roles such as antigen capture leading to presentation²³ and possible function as a pattern recognition receptor²⁰ were described. In support of this last role, the MR binds numerous microorganisms and microbial products including Mycobacterium tuberculosis,²⁴Trypanosoma cruzi,²⁵Streptococcus pneumoniae and Klebsiella pneumoniae,²⁶ human immunodeficiency virus gp120²⁷ and Trichinella spiralis.²⁸ However, the importance of the MR in pathogen recognition is unclear because mice lacking this receptor show normal immune responses to Candida albicans,²⁹Pneumocystis carinii³⁰ and Leishmania sp.,³¹ despite all these pathogens possessing structures that bind to the MR.

Several gastrointestinal nematodes have been reported to express ligands for the MR on their surface^32–34 and Trichinella spiralis larval components bind the MR leading to the activation of nitric oxide production from peritoneal macrophages.²⁸Trichuris muris excretory/secretory products (E/S) form a heterogeneous solution of worm proteins containing substances that have been shown to express structures bearing mannose and N-acetylglucosamine residues,³⁵ which makes them potential ligands for the MR. Therefore it was assessed whether T. muris E/S contains MR-binding activity and if the absence of this receptor altered the immune response to T. muris, particularly with regard to the pathology that occurs in the large intestine during infection with this parasite.

Data presented here reveal that T. muris E/S contains one or more components capable of binding to the mannose-recognizing domain of the MR. However, E/S either does not stimulate macrophages via this receptor or compensatory pathways exist. In vitro studies demonstrated MR KO-derived bone-marrow-derived macrophages (BMDM) expressed similar levels of several cytokines when exposed to T. muris E/S. The only difference seen was a reduction in the production of IL-6 by alternatively activated BMDM in the absence of the MR. Infection of MR KO mice revealed expulsion of T. muris with the same kinetics as wild-type animals and a similar cytokine response in the draining mesenteric lymph nodes. Moreover, there were no differences in macrophage recruitment, the ability of macrophages to become alternatively activated, goblet cell hyperplasia or gross crypt pathology during infection. In summary the MR is not required for the development of an immune response leading to the expulsion of T. muris.

Materials and methods

Lectin-binding assays

Enzyme-linked immunosorbent assay plates (Maxisorb; Fisher Scientific, Loughborough, UK) were coated with mannan (Sigma, Poole, Dorset, UK) or SO₄-3-β-d-Gal-PAA (Lectinity Holding, Moscow, Russia), or T. muris E/S in phosphate-buffered saline. Wells were washed with TTBS [10 mm Tris–HCl (pH 7·5)/10 mm CaCl₂/154 mm NaCl/0·05% Tween-20]. CD206 constructs (CTLD4–7-Fc and CR-FNII-CTLD1–3-Fc)^26,36 were then incubated at 1 μg/ml in TTBS, followed by incubation with a goat anti-human immunoglobulin Fc-specific antibody coupled to alkaline phosphatase diluted 1 : 1000 (Sigma). Binding was measured at 405 nm after development with a p-nitrophenyl phosphate substrate (Sigma). For inhibition studies, incubation of Fc proteins with carbohydrates was performed in the presence of 1 m NaCl as described elsewhere.²⁶

BMDM preparation and culture

Bone marrow was recovered from the femurs of wild-type and KO mice by flushing the bone with Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen, Paisley, UK) containing 10% fetal calf serum, 2 mm l-glutamine, 100 units/ml penicillin and 100 μg/ml streptomycin (complete DMEM). After passing the cell suspension through a 21 gauge needle, cells were washed and cultured at 1 × 10⁶/ml in complete DMEM containing 30% L929-conditioned media for 7–10 days with the media replaced after 4–5 days. Macrophage purity was assessed by staining with phycoerythrin-conjugated anti-CD11b-PE and anti-F4/80-APC (BD Biosciences, Oxford, UK). More than 80% of live cells [measured by 7-amino-actinomycin d (Sigma) exclusion] were positive for either or both markers. Fewer than 1% of cells were positive for CD11c. The BMDM (1 × 10⁶) were adhered to six-well plates and cultured for 24 hr in media alone or in the presence of 20 ng/ml IL-4 (R&D Systems, Abingdon, UK) and 25 μg/ml anti-IFN-γ (clone XMG1·6, produced in house) to induce an alternative phenotype. Cells were then washed and cultured for a further 24 hr in the presence or absence of T. muris E/S. Supernatants were collected for analysis by enzyme-linked immunosorbent assay (ELISA) and cells were lysed in TRIzol (Invitrogen) for a quantitative polymerase chain reaction (PCR) analysis.

Extraction of total RNA, reverse transcription and quantitative PCR

The BMDM were lysed in TRIzol and total RNA extracted according to the manufacturer’s instructions. Integrity of RNA was confirmed by visualization of the 18S and 28S ribosomal bands under ultraviolet light following separation on a 1·5% agarose gel. The concentration of total RNA was measured by absorbance at 260 nm on a Nanodrop ND-1000 spectrophotometer (Labtech International, Lewes, East Sussex, UK). 1·0 μg of total RNA was reverse transcribed using SuperScript 2 (Invitrogen) in a final volume of 40 μl according to the manufacturer’s instructions and stored at −20° until used. Quantitative PCR was performed using SYBR green (New England Biolabs, Hitchin, UK) on an OPTICON DNA engine with OPTICON monitor software version 2·03 (Real-Time systems; MJ Research, Hemel Hempstead, UK). Amplification of mRNA encoding Hprt1, 18S and β-actin was performed to control for the starting amount of cDNA. Expression levels of genes of interest are shown as fold change over that seen in naïve animals after normalization to housekeeping gene levels using the ΔΔC^t method. Primers used were: GTAATGATCAGTCAACGGGGGAC and CCAGCAAGCTTGCAACCTTAACCA for Hprt1, AGTCCCTGCCTTTGTACACA and GATCCGAGGGCCTCACTAAC for 18S, GTGGGCCGCTCTAGGCACCAA and CTCTTTGATGTCACGCACGATTTC for β-actin, TCCCAGTGAATACTGATGAGA and CCACTCTGGATCTCCCAAGA for Fizz1, CAGAAGAATGGAAGAGTCAG and CAGATATGCAGGGAGTCACC for Arg1; and GGGCATACCTTTATCCTGAG and CCACTGAAGTCATCCATGTC for Ym1. All sequences are 5′–3′ with the sense primer given first.

Animals, T. muris and E/S protein

The MR KO mice were a kind gift from Michel C. Nussenweig (The Rockefeller University, New York, NY) and C57BL/6 mice were purchased from Harlan UK. (Bicester, UK). Male mice were used in all experiments. Mice were infected with T. muris when 8–10 weeks old and killed at various time-points after infection with approximately 150 infective eggs. The worm burden in the large intestine was assessed as previously described.² Animal experiments were performed under the regulation of the Home Office Scientific Procedures Act (1986). The maintenance of T. muris, the method of infection and production of E/S protein were as previously described.³⁷ Batches of E/S used for the stimulation of BMDM were tested for endotoxin content by the chromogenic Limulus amoebocyte lysate assay (Charles River, Charleston, SC). The level of contamination was found to be less than 3 pg/ml, over 100-fold less than the concentration of lipopolysaccharide required to stimulate the production of cytokines or chemokines from BMDM (data not shown).

Culture of mesenteric lymph node (MLN) cells for in vitro cytokine measurement

Suspensions of MLN cells from infected or naive mice were prepared as described previously.³⁸ Briefly, 5 ×10⁶ total MLN cells were resuspended in RPMI-1640 medium supplemented with 5% fetal calf serum, 2 mm l-glutamine, 100 units/ml penicillin, 100 μg/ml streptomycin (all from Invitrogen) and 60 μm monothioglycerol (Sigma). The MLN cells were stimulated with 50 μg/ml T. muris E/S antigen in 48-well plates (Helena Biosciences, Sunderland, UK) with supernatants harvested after 24 hr and stored at −20° until use.

Cytokine and chemokine ELISAs

Interleukin-5, IFN-γ and IL-12p40 were analysed by sandwich ELISA as previously described.³⁸ The antibody clones used were: IL-12p40; C15.6 (G. Trinchieri, Schering-Plough, Dardilly, France) and C17.8, IFN-γ; R4-6A2 and XMG1.2, IL-5; TRFK5 and TRFK4 (all from BD Biosciences). The ELISAs for CCL2, CCL3, CCL9 (R&D Systems), IL-6, IL-10 and tumour necrosis factor (TNF) (BD Biosciences) were performed according to the manufacturer’s instructions. All detection antibodies were biotinylated and used in conjunction with streptavidin–peroxidase (Roche Diagnostics, Sussex, UK). Colour development following the addition of tetramethylbenzidine (BD Biosciences) was monitored and the reaction was stopped with 4·1 m sulphuric acid. Plates were read at 405 nm.

Histology

Immunohistochemical staining of F4/80 was performed as previously described.⁸ The same protocol was used for Fizz1 and MR staining except tissue was fixed at 4° in 2% paraformaldehyde containing 154 mm NaCl, 5 mm CaCl and 25 mm HEPES and non-specific binding was blocked with 7% goat serum (Vector Laboratories, Peterborough, UK). The primary antibodies were anti-mouse Fizz1 (5 μg/ml, BD Biosciences) and anti-mouse MR (5 μg/ml, AbD Serotec, Oxford, UK) and the secondary antibody was biotinylated goat anti-rat F(ab′)₂ (1 in 4000, Millipore, Watford, UK).

For dual immunofluoresent visualization of F4/80 and MR, sections were fixed and blocked as above before sequential incubation with an avidin/biotin blocking kit (used according to the manufacturer’s instructions, Vector Laboratories), anti-mouse MR, biotinylated goat anti-mouse F(ab′)₂, Texas red-avidin D (30 μg/ml, Vector Laboratories), avidin/biotin blocking kit, biotinylated anti-mouse F4/80 (7·5 μg/ml, AbD Serotec) and Fluorescein-avidin D (30 μg/ml, Vector Laboratories). Sections were washed twice in phosphate-buffered saline between each step. Finally slides were mounted in Vector Shield containing DAPI (Vector Laboratories) and images were collected on an Olympus BX51 upright microscope using a 40 ×/0·75 Plan Fln objective and captured using a Coolsnap ES camera (Photometrics, Tucson, AZ) through metavue Software (Molecular Devices, Wokingham, UK). Specific band pass filter sets for DAPI, FITC and Texas red were used to prevent bleed-through from one channel to the next. Images were then processed and analysed using ImageJ (http://rsb.info.nih.gov/ij).

To enumerate goblet cells, tissue was fixed in 10% neutral-buffered formalin, wax processed and 5-μm sections were cut. Sections were dewaxed with citroclear and taken to water through decreasing concentrations of ethanol. Mucins in goblet cells were stained with 1% alcian blue (Sigma) in 3% acetic acid, washed and treated with 1% periodic acid (Sigma) followed by counterstaining in Mayer’s haematoxylin (Sigma). Slides were dehydrated and mounted in aquamount (BDH Laboratory Supplies, Poole, UK). For enumeration of immunohistochemistry and goblet cell staining, the average number of cells from 20 crypts was taken from three different sections per mouse.

The measurement of crypt depth was performed on digital photomicrographs of sections stained for goblet cells taken at 100 × magnification. A minimum of 10 measurements of crypt depth were taken, evenly distributed across the section, using image-pro plus (version 4·5, Media Cybernetics, Bethesda, MD).

Statistical analysis

Statistical analysis was performed using the Student’s t-test or analysis of variance as appropriate with the statistical package graphadprism (GraphPad Software, San Diego, CA). A probability value of < 0·05 was considered significant.

Results

T. muris E/S contains MR-binding activity

Trichuris muris E/S is a heterogeneous solution of worm proteins containing substances that have been shown to express structures bearing mannose and N-acetylglucosamine residues.³⁵ To demonstrate binding to the MR, an in vitro ELISA-based method using constructs of the MR with mannose or SO₄-3-galactose-binding activity was used. Figure 1(a) shows mannan binding to CTLD4-7-Fc and SO₄-3-galactose to CR-FN-CTLD1-3-Fc as positive controls. The T. muris E/S at 10 and 5 μg/ml showed strong binding to the mannose recognition domain, CRD-4-7-Fc, but not the SO₄-3-galactose recognition domain (Fig. 1a). To confirm the specificity of binding, an inhibition assay was performed. Mannose (10 mm) completely blocked the binding of both mannan and E/S to CRD-4-7-Fc, whereas galactose at the same concentration had no effect (Fig. 1b).

Figure 1.

Trichuris muris excretory/secretory products (E/S) contains mannose receptor (MR) -binding activity. Binding of E/S to the MR constructs CTLD4–7-Fc and CR-FNII-CTLD1–3-Fc was measured by enzyme-linked immunosorbent assay (ELISA). (a) Plates were coated with E/S (5 or 10 μg/ml), mannan (man) (2 μg/ml) or SO₄-3-β-d-GalNAc-PAA (Gal) (5 μg/ml). CR-FNII-CTLD1–3-Fc (white bars), CTLD4–7-Fc (black bars) or secondary antibody alone (hatched bars) was added to the wells. Fc portions were detected with a mouse anti-human immunoglobulin Fc-specific antibody coupled to alkaline phosphatase. (b) Inhibition of the binding of CTLD4–7-Fc to E/S by mannose. E/S (10, 5, 2·5 or 1·25 μg/ml) or mannan was coated onto ELISA plates. CTLD4–7-Fc (1 μg/ml) was preincubated alone (white bars), with mannose (10 mm, black bars) or with galactose (10 mm, hatched bars) for 30 min in the presence of 1 m NaCl before loading onto the coated wells and the assay continued as before.

Response of MR KO BMDM to T. muris E/S in vitro

Treatment of peritoneal macrophages with IL-4 to induce alternative activation causes high levels of expression of the MR.^17,18 Expression of the MR was also increased on wild-type BMDM cultured with IL-4 and anti-IFN-γ when measured by quantitative PCR (15·9 ± 11·4 fold increase over untreated cells from three independent experiments, data not shown). Experiments were then performed to determine whether the MR plays any role in the alternative activation of macrophages in vitro. BMDM from wild-type and MR KO mice were cultured in media alone or with IL-4 and anti-IFN-γ to induce alternative activation. Alternatively activated BMDM expressed high levels of Fizz1, Arg1 and Ym1, shown as fold change in expression over cells cultured in medium alone (Fig. 2). These genes are associated with alternative activation and demonstrate that both wild-type and MR KO BMDM were stimulated to the alternative phenotype by culture with IL-4 and anti-IFN-γin vitro. Although the fold increase of Fizz1, Arg1 and Ym1 was higher in wild-type BMDM (Fig. 2), this increase between wild-type and KO BMDM was not significantly different when analysed across separate experiments.

Figure 2.

Alternative activation of bone-marrow-derived macrophages (BMDM) from wild-type (WT) and mannose receptor knockout (MR KO) mice. The BMDM were cultured in media alone (unpolarized) or with interleukin-4 (IL-4) and anti-interferon-γ (IFN-γ) (alternatively activated macrophages; AAMφ) for 24 hr and expression of genes associated with alternative activation was measured. (a) Fizz1, (b) Arg1 and (c) Ym1 expression levels for wild-type (black bars) and MR KO (white bars) AAMφ are shown as fold change over unpolarized BMDM. Results are representative of three independent experiments.

Next, the ability of wild-type and MR KO BMDM to respond to T. muris E/S was investigated. Following culture for 24 hr in either medium alone (unpolarized BMDM) or under AAMφ-inducing conditions, cells were washed and cultured for a further 24 hr with or without T. muris E/S and the supernatant was collected for analysis of proinflammatory and AAMφ-associated cytokines by ELISA. Differences were seen in the levels of cytokine production between unpolarized and AAMφ irrespective of MR status with both wild-type and MR KO AAMφ producing significantly more IL-6, TNF, IL-10, CCL3 and CCL9 as a result of culture with IL-4 and anti-IFN-γ (compare Fig. 3a white bars to Fig. 3b white bars). The effect of adding T. muris E/S to BMDM was then studied. Unpolarized BMDM from wild-type and MR KO mice stimulated with E/S secreted increased levels of IL-6, IL-10, CCL3 and CCL9 (Fig. 3a). However, E/S stimulation of AAMφ only led to higher expression of CCL2 from wild-type cells, possibly because of the increased secretion of cytokines from unstimulated BMDM as a result of the polarization treatment (Fig. 3b). However, these cells were not refractory to further stimulation because addition of lipopolysaccharide led to increased cytokine production (data not shown). Interleukin-6 was the only cytokine for which the measured expression varied between wild-type and MR KO BMDM, and then only when BMDM were alternatively activated. The expression of IL-6 from both unstimulated and E/S-treated MR KO AAMφ was reduced compared to the corresponding wild-type cells (Fig. 3b).

Figure 3.

Cytokine secretion by unpolarized and alternatively activated bone-marrow-derived macrophages (BMDM) from wild-type and mannose receptor knockout (MR KO) mice in response to excretory/secretory product (E/S) stimulation. The BMDM were cultured in medium alone (unpolarized; a) or with interleukin-4 (IL-4) and anti-interferon-γ (IFN-γ) (alternatively activated macrophages; AAMφ; b) for 24 hr, washed and cultured with E/S (E/S; black bars) or without E/S (unstimulated; white bars) for a further 24 hr. Expression of IL-6, tumour necrosis factor (TNF), IL-10, CCL3, CCL9 and CCL2 was measured by enzyme-linked immunosorbent assay. Data show the mean +1 SD and are representative of three independent experiments. Significant differences were calculated by analysis of variance; *P< 0·05 between the groups indicated.

Response of MR KO mice to infection with T. muris

Wild-type and MR KO mice were infected with approximately 150 embryonated T. muris eggs and the worm burden was assessed at days 14 and 21 postinfection. C57BL/6 mice are generally resistant to T. muris, normally expelling the worms between days 21 and 28 postinfection. There was no difference in the worm burdens of wild-type and MR KO mice at either time-point, demonstrating equal establishment of infection at day 14 as well as similar expulsion by day 21 postinfection (Fig. 4a). Both strains had almost completely expelled the worms by day 35 postinfection (data not shown). Expulsion of worms by C57BL/6 mice was dependent on the presence of a type 2 immune response, rather than the absence of a type 1 immune response. Figure 4(b–d) shows elevated levels of the type 2 cytokine IL-5 as well as increased amounts of IFN-γ and IL-12p40 produced by E/S-stimulated MLN cells at day 21 postinfection in wild-type and MR KO mice. However, there were no differences in E/S-induced MLN cytokine responses between wild-type and MR KO. The skewing of the immune response to a type 1 or type 2 response in vivo can also be inferred from the antibody isotype generated in response to infection as IL-4 induces immunoglobulin G1 (IgG1) switching in B cells and IFN-γ induces IgG2a. In agreement with the mixed type 1/2 cytokine data, measurement of parasite-specific antibody isotypes revealed equivalent levels of IgG1 and IgG2a in the sera of T. muris-infected wild-type and MR KO mice (data not shown).

Figure 4.

Worm burden and mesenteric lymph node (MLN) cytokine responses of wild-type and mannose receptor knockout (MR KO) mice following infection with Trichuris muris. (a) Worm burden was assessed at day 14 postinfection (p.i.) to confirm equal levels of establishment (wild-type; black bars, KO; white bars). At day 21 p.i. both wild-type and MR KO mice had expelled the parasite to the same degree. MLN expression of interleukin-5 (IL-5; b), interferon-γ (IFN-γ; c) and IL-12p40 (d) in response to excretory/secretory product (E/S) was measured by enzyme-linked immunosorbent assay in naïve and infected wild-type (black bars) and MR KO (white bars) mice. Data show the mean +1 SD of three to six mice per group and are representative of two independent experiments. Significant differences were calculated by analysis of variance;*P< 0·05 and **P< 0·01 compared to naïve mice.

Pathology of the large intestine following T. muris infection of MR KO and wild-type mice

To establish whether inflammation in the large intestine associated with T. muris infection was more severe in MR KO mice a number of parameters were measured. First, Fig. 5(a) confirms that MR expressing cells accumulate in the lamina propria of the large intestine of wild-type mice 21 days after T. muris infection. This correlated with a significant influx of F4/80⁺ macrophages in the large intestine of wild-type and MR KO mice postinfection (Fig. 5b,c).

Figure 5.

Macrophage phenotype in the large intestine of wild-type and mannose receptor knockout (MR KO) mice 21 days after infection with Trichuris muris. Immunohistochemistry was used to demonstrate MR (a) and F4/80 (b) expression in naïve and infected wild-type and MR KO mice. Original magnification ×200, scale bar 50 μm. (c) Quantification of macrophage influx in infected wild-type (black bars) and KO (white bars) mice detected by immunohistochemistry for F4/80. (d) Dual expression of the MR (red) and F4/80 (green) in wild-type but not KO mice was shown by immunofluorescence. White arrowheads MR⁺F4/80⁺ cells, black arrowheads MR⁺F4/80⁻ cells. Original magnification ×400, scale bar 50 μm. (e) Confirmation of alternatively activated macrophage (AAMφ) phenotype in vivo was demonstrated by Fizz1 staining (wild-type; black bars, KO; white bars). Data show the mean +1 SD of three to six mice per group and are representative of two independent experiments. Significant differences were calculated by analysis of variance; *P< 0·05 and **P< 0·01 compared to naïve mice.

To determine if MR and F4/80 were expressed on the same cells, colocalization of these molecules was performed by immunofluorescence. Staining of MR expressing cells with Texas red and F4/80 with fluorescein avidin shows that most of the MR expressing cells are F4/80⁺ macrophages (yellow staining; white arrowheads, Fig. 5d). Cells were also present that express the MR but not F4/80 (red staining; black arrowheads, Fig. 5d). Immunohistochemical staining for the AAMφ marker Fizz1 revealed that a proportion of the macrophages in the large intestine expressed this molecule (32·4% and 42·0% of wild-type and MR KO, respectively; Fig. 5e). There were no differences in either the number of macrophages recruited to the large intestine of wild-type and MR KO mice or the proportion expressing Fizz1. Furthermore, there was only a very low level of expression of inducible nitric oxide synthase, a marker of classical macrophage activation, in MR KO and wild-type animals measured by immunohistochemistry (data not shown).

To establish the extent of inflammation caused by infection the degree of goblet cell hyperplasia and average crypt depth were measured. The number of goblet cells in the large intestine of infected mice was increased over that found in naïve animals, although this was only significant for MR KO mice (Fig. 6a,b). With regard to crypt depth, there was a small increase in infected mice but this was not significantly different to that in naïve animals (Fig. 6c), demonstrating that lack of MR does not result in excessive inflammation following infection with T. muris.

Figure 6.

Pathology of the large intestine in wild-type and mannose receptor knockout (MR KO) mice 21 days after infection with Trichuris muris. (a) Goblet cells were stained with alcian blue and periodic acid and photomicrographs were taken at an original magnification of ×400 (scale bar 20 μm). (b) The average number of goblet cells per crypt was assessed (wild-type; black bars, KO; white bars). (c) Photomicrographs were taken at 100 × magnification and the crypt depth was measured using image-pro plus software (wild-type; black bars, KO; white bars). Data show the mean +1 SD of three to six mice per group and are representative of two independent experiments. Significant differences were calculated by analysis of variance; *P< 0·05 compared to naïve mice.

Discussion

This report investigated the role of the MR in BMDM cytokine production, large intestinal inflammation and the expulsion of T. muris. Culture of BMDM isolated from wild-type and MR KO mice with parasite-derived proteins reveals for the first time that BMDM are capable of mounting a strong cytokine response to T. muris E/S with increased expression of several molecules. This suggests that in the absence of the MR, one or more pattern recognition receptors (PRR) exist for T. muris E/S and are expressed by macrophages. Candidate compensatory PRR which may bind mannose-containing glycoproteins include Toll-like receptor 2 (TLR-2), TLR-4 and TLR-6. Binding of mannan to these receptors induces TNF production by monocytes and macrophages.^39,40 TLR-4-deficient animals are, however, resistant to T. muris infection⁴¹ demonstrating that this PRR is not important in worm expulsion.

In terms of gene expression and cytokine production, wild-type and MR KO BMDM are shown here to be functionally very similar. In fact, the only difference found in this study between wild-type and MR KO BMDM was in the production of IL-6 by alternatively activated BMDM. The MR KO cells secreted lower levels of this cytokine both constitutively and following stimulation with E/S, despite adopting the alternatively activated phenotype to the same degree as wild-type BMDM. To be able to induce cytokine production and initiate an immune response, however, ligation of a receptor should initiate a signalling cascade leading to gene transcription. There is still some debate as to whether the MR initiates intracellular signalling as the cytoplasmic tail of the receptor is short and does not contain any known signalling motifs. However, there have been other reports of cytokine production^42–44 as well as nuclear translocation of nuclear factor-κB p50 and p65 subunits⁴⁵ following MR ligation. Interestingly, the MR is known to mediate the production of IL-6 by thioglycollate-elicited macrophages when exposed to C. albicans.⁴² The data presented here also suggest a role for the MR in the expression of IL-6 by macrophages.

To determine whether the MR was necessary for protection against T. muris in vivo, wild-type and MR KO mice were infected. Expulsion from wild-type and MR KO mice began between day 14 and 21 postinfection and the worms were fully expelled by day 35, showing that either the MR is not involved in the generation of resistance to T. muris or that compensatory mechanisms exist in vivo. This result follows the pattern seen with other infectious organisms that bind the MR. The fungal pathogens C. albicans and P. carinii have both been shown to interact with the MR. The uptake of C. albicans by macrophages is inhibited by other MR ligands⁴⁶ and anti-sense MR oligonucleotides⁴² and macrophage invasion by P. carinii is believed to occur partially through the MR.⁴⁷ Despite this, infection of MR KO mice revealed no change in a number of immune parameters, with the mice reacting to infection as did wild-type animals.^29,30 It has since been shown that the β-glucan receptor, dectin-1, is more important than the MR for immunity to C. albicans and P. carinii^48,49 although there are differences in immunity to C. albicans in the two dectin-1 KO strains described in these reports that have yet to be resolved. In the case of P. carinii infection of MR KO mice, the reduced ability of macrophages to phagocytose the fungus was thought to be compensated for by increased recruitment of macrophages to the alveolar space and the use of other PRR, most likely dectin-1.³⁰ The data in Fig. 5(b) show that there was no difference in the accumulation of macrophages in the large intestine following infection with T. muris, which demonstrates that compensation through increased macrophage numbers expressing other mannose-binding receptors is not occurring in this model.

A further report using MR-deficient mice showed no role for this receptor during Leishmania infection. The entry of the obligate intracellular parasite Leishmania major into macrophages is thought to involve MR binding of surface mannose as it is partially blocked by the use of mannan as a competitive ligand. Infection of MR KO mice with L. major produced normal macrophage cytokine production, tissue healing and parasite clearance.³¹ Furthermore, the inflammatory infiltrate was comparable between wild-type and KO mice and phagocytosis of L. major metacyclic promastigotes by MR KO macrophages was as efficient as by wild-type cells, showing that despite bearing mannosylated ligands the MR is not essential for the uptake of this parasite. In combination with previous reports, the results presented here suggest that the MR is not an important receptor for the recognition of infectious organisms, that is, it does not function as a PRR. In fact, there have been no reports of MR KO mice showing increased susceptibility to any infectious organism, despite the strong in vitro evidence of numerous pathogen molecules binding to the MR.^24–31

One reported non-redundant role for the MR is as a clearance mechanism for endogenous lysosomal hydrolases.²² The removal from the serum of mannose and N-acetylglucosamine -bearing enzymes such as the known inflammatory proteins β-glucuronidase and β-N-acetylglucosaminidase, is important in the resolution phase of inflammation. However, it is demonstrated here that the large intestines of wild-type and MR KO mice were inflamed to a similar extent during infection with T. muris suggesting that compensatory mechanisms exist in vivo. Indeed, the infection models discussed above also showed no changes in inflammation or tissue repair. One explanation for this is the expression of other C-type lectin receptors on macrophages, such as Endo-180, that can bind mannose-bearing ligands and compensate for the absence of the MR.⁵⁰

An apparent lack of function for the MR in immune responses to mannosylated pathogens raises the possibility that intracellular infectious agents may use this receptor to silently infect macrophages. As a clearance mechanism for proinflammatory endogenous proteins, ligand binding by the MR should not result in macrophage activation, and consequently, pathogens invading cells via this mechanism would reach their desired niche without alerting the innate immune system to their presence. With regard to extracellular parasites such as T. muris, the MR has the potential to protect the host from excessive inflammation by acting in its normal role; the clearance of mannosylated, proinflammatory proteins from the extracellular milieu. During infection these would be of host as well as parasite origin to allow for regulation of the immune response and normal tissue repair. However, if this is the case, it is a redundant mechanism; with equivalent intestinal inflammation apparent in the T. muris-infected MR KO and wild-type mice, as discussed above.

The weight of evidence published to date leans toward the MR only being involved in the clearance of endogenous glycoproteins, despite many mannose-bearing and N-acetylglucosamine-bearing pathogens having been demonstrated to bind this receptor. The results presented here add the large intestinal-dwelling nematode T. muris to the list of organisms bearing such molecules in which the MR is redundant in the generation of an effective immune response leading to macrophage activation, parasite expulsion and tissue repair.