Background
Dendritic cells (DC), APC specialised for the efficient stimulation of naïve T cells, are of fundamental importance in the control of antigen-specific immune responses (reviewed in [
1]). Immature DC are sparsely distributed in peripheral organs, where they act as sentinels, continuously sampling the antigenic environment. They undergo maturation in response to stimuli that include microbial components and tissue damage, and migrate to T dependent areas of lymphoid organs. Here, they upregulate expression of costimulatory molecules and peptide-loaded surface MHC class I and II molecules and develop the capacity to stimulate antigen-specific T cells restricted by MHC class I and II.
Immature DC capture antigens by receptor-mediated endocytosis, in addition to macropinocytosis and phagocytosis (reviewed in [
2]). DC are phenotypically and functionally heterogenous (reviewed in [
3]), so the ability to target antigens via specific receptors to different subsets of DC in vivo may help to reveal distinctive features of their roles. One candidate receptor for endocytosis in DC is the mannose receptor (MR), or CD206, which recognises glycoconjugates bearing terminal mannose, fucose and N-acetylglucosamine by interaction with its carbohydrate recognition domains (CRDs). Natural ligands include microbial polysaccharides, glycoproteins and glycolipids, and mammalian glycoproteins with N-linked high-mannose (reviewed in [
4]). MR is expressed mainly by subsets of Mφ and endothelial cells in vivo [
5,
6], but it is also expressed by immature cultured DC, where it endocytoses mannosylated ligands for processing and presentation to T cells by MHC class II [
7]. Mannosylation of antigen confers a greatly enhanced efficiency of presentation to T cells of the order of 100 [
8], and 200 to 10 000-fold [
9]. However, MR is not expressed by immature splenic DC or epidermal Langerhans cells in situ in naïve mice [
6], and its contribution to T cell immunity remains unknown.
Notwithstanding their role in stimulating immune responses, it is becoming increasingly apparent that DC play a role in maintaining T cell tolerance to self antigens in the periphery (reviewed in [
10,
11]). Direct evidence that DC can induce T cell unresponsiveness under non-inflammatory conditions came from an elegant study in which a model MHC class II peptide was targeted to DC in situ [
12]. The peptide was engineered into a mAb specific for DEC-205 [
12], an endocytic receptor of DC that is structurally related to MR [
13]. Although T cells initially proliferated in vivo in response to DC targeting, the response was short-lived, and mice were rendered unresponsive to subsequent challenge with antigen in adjuvant [
12]. Antigen targeting via DEC-205 also led to CD8+ T cell tolerance in the steady state [
14], and the generation of regulatory T cells [
15]. The latter suppressed proliferation of conventional CD4+ T cells in vitro, and appeared to exert immunosuppressive effects in both CD4+ and CD8+ T cell driven immunopathologies [
15].
We recently found MR positive interstitial cells in the thyroid, pancreas and salivary gland: secretory organs which bear endogenous ligands of the CRDs of MR [
16]. Thyroglobulin was identified as the major MR ligand in the thyroid. Since this glycoprotein is an autoantigen, we postulated that APC in the thyroid may use MR for antigen capture. As a basis for experiments to determine the role of MR in antigen handling in vivo, especially in secretory organs, we surveyed naïve murine non-lymphoid organs for MR positive APC and report here on their phenotypic characterisation. We also provide further evidence for the existence of distinct ligands of MR in pancreas and salivary gland. In light of the immunosuppressive function of immature DC in non-inflammatory conditions, a role for MR in antigen capture by APC for the purposes of maintenance of T cell tolerance to its ligands is credible.
Discussion
The MR was the first non-opsonic C-type lectin receptor shown to be expressed and mediate adsorptive endocytosis in DC [
7,
20], properties it shares with a growing family of such receptors (reviewed in [
21]). Expression of MR by DC appears to be tightly regulated. In human peripheral blood monocyte-derived DC, MR expression is down-regulated by inflammatory stimuli, and for this reason it has been considered a prototype marker of immature DC [
7]. However, human epidermal Langerhans cells which are functionally immature, do not express MR in situ [
22‐
25], although have been found to be positive for MR after a short isolation procedure [
26]. In fact, many DC subsets that have differentiated in vivo do not express MR; a summary of MR expression by APC is shown in table
1. Here, the expression of MR by APC in non-lymphoid organs of the naive mouse is described. Cells were analysed in situ, circumventing the problem that isolation or culture of cells for ex vivo analysis may affect expression of MR.
In contrast to our earlier observation that MR is not expressed by well-characterised subsets of DC in lymphoid organs [
6], many APC in peripheral tissues expressed MR, including a proportion of MHC class II positive cells in skin, liver, tongue and heart. Whilst a limited phenotypic analysis is not sufficient to prove that these cells are DC, in the case of MR positive cells in the thyroid, pancreas and salivary glands, this designation appears to be a strong possibility. DC are normally distinguished from monocytes and macrophages by phenotype, morphology and function, properties which are most amenable to study after cell isolation. MR is expressed by interstitial cells in situ in pig thyroids, and MR positive cells have been successfully isolated from this source, and maintained in co-culture with thyrocytes in the presence of thyroid stimulating hormone [
27]. Isolated cells expressed MHC class II, S100 protein and had a high endocytic capacity, characteristic of DC. They also exhibited morphological features of DC including a plasma membrane with numerous processes, a reniform nucleus and intracellular annular structures. In response to a stimulator of DC maturation, TNFα, the cells detached from the culture substratum and lost the ability to endocytose the MR ligand, FITC-dextran. By extension, the MR positive APC detected in the murine thyroid may also be immature DC. MR positive APC in the pancreas and salivary gland were phenotypically indistinguishable from the thyroid cells, according to the markers used in this study. APC in all three secretory organs exhibited a strongly myeloid phenotype, expressing macrosialin and sialoadhesin, but were unlikely to be activated Mφ since they were also present in IFNγ -/- mice. MR positive immature DC or other APC at peripheral sites are likely to down-regulate MR when they mature and migrate to lymph nodes, in accordance with in vitro models [
7,
27], since the mature interdigitating cells of tissue-draining lymph nodes do not express MR [
6].
The function of MR in relation to antigen handling in vivo is poorly understood, in part because most studies have used glycosylated Ag that are now known to be ligands of other receptors in addition to MR (reviewed in [
28]). However, a specific role for MR in enhancing uptake and presentation to both CD4
+ and CD8
+ T cells by DC has recently been shown using Ag fused to mAb specific for MR [
29,
30]. Surprisingly, reports of any ability of MR to enhance primary immune responses in vivo have not been forthcoming, but since MR is not expressed by splenic DC or epidermal Langerhans cells (figure
1; [
6]), MR may not play a role in DC capture of antigen administered intravenously or by skin absorption. Isolated human lung DC exhibit a high endocytic capacity for the MR ligand, FITC-dextran [
31]. Therefore, MR may play a role in handling of inhaled antigens, such as the house dust mite allergen, Der p 1, whose endocytosis by DC is dependent on MR [
32]. A link between MR and allergy has been suggested based on the finding that monocyte-derived DC from allergic patients expressed more MR and endocytosed Der p 1 more efficiently than did DC derived from healthy donors [
32].
Antigen uptake by MR on cultured DC promotes presentation to T cells of antigen-derived peptides on MHC class II [
7]. The colocalisation of MR and MHC class II in confocal microscope images of APC in thyroid, pancreas and salivary gland described here implies a role in antigen uptake for presentation. Interestingly, in mouse bone marrow-derived DC and a MHC class II-expressing fibroblast cell line, MR did not colocalise with MHC class II [
33]. However, the structurally similar receptor, DEC-205, did target MHC class II compartments, and this was found to depend on the presence of an EDE triad within the receptor's cytoplasmic tail, which MR lacks. When fused to the extracellular domain of the Fc receptor, the cytoplasmic tail of DEC-205 directed endocytosed human IgG to a compartment from which MHC class II loading and subsequent presentation of IgG-derived peptides was 100-fold more efficient than that driven by the cytoplasmic tail of MR. In another study, efficiencies of processing and presentation of ribonucleases A and B by a fibroblast cell line expressing MR and MHC class II IA
K were compared [
34]. These model Ag differ only by the presence of a high-mannose oligosaccharide on RNAse B. Although RNAse B, a ligand of MR, was endocytosed more efficiently than the non-ligand RNAse A, efficiencies of presentation of the two Ag were indistinguishable [
34]. Therefore, MR does not constitutively enhance Ag presentation of its ligands. To do so, it may need to target MHC class II compartments. The apparent colocalisation of MR and MHC class II in secretory organ APC implies that the intracellular locations of MR or MHC class II molecules in these cells may be distinct from those in bone marrow-derived DC and transfected fibroblasts, perhaps permitting MR to target Ag to these compartments with greater efficiency. Unusually for immature DC, those cultured from pig thyroid were characterised by plasma membrane expression of MHC class II [
35].
If MR is poised to direct Ag to the MHC class II compartments of secretory organ APC, the question of which Ag are captured in these environments must be raised. We previously showed that thyroglobulin, a major auto-antigen in murine autoimmune thyroid diseases, is a ligand of MR, and suggested that MR expressed by interstitial cells in the thyroid may be involved in maintaining tolerance to this Ag in normal mice [
16]. Here, further evidence is presented that the pancreas and salivary glands also bear abundant MR ligands. A high molecular weight fraction in salivary gland lysate (>250 kD) may correspond to secretory IgA. This Ab inhibits adhesion of microbes to the gut wall, but is a non-inflammatory isotype, to avoid unwanted immune reactions against indigenous microflora and dietary Ags (reviewed in [
36]). Uptake of secretory IgA by cultured human DC could be partially blocked by mAb against MR and this pathway has been suggested to allow modulation of mucosal immune responses [
37]. The MR ligands in pancreas lysate were of different Mr to those in salivary gland, and remain undefined. A role for MR in maintenance of immune tolerance to its ligands in this organ is credible since pancreatic DC of naïve mice are believed to be tolerogenic [
38]. DC isolated from the lymph node draining the pancreas, but not from other lymph nodes, were able to confer protection against diabetes when transferred to the nonobese diabetic (NOD) mouse strain [
38].
In light of the location of MR positive APC in secretory organs, the abundance of endogenous ligands, and the positive identification of one ligand as thyroglobulin, the available data could suggest a tolerogenic role of MR is more likely than an immunostimulatory role in these organs in the steady-state. Such a mechanism may represent a physiological correlate of the experimentally induced immune tolerance achieved by targeting Ag to DEC-205 in naïve mice [
12,
14,
15]. Further studies will be required to address the significance of MR to immune homeostasis.