Generation of novel anti-human-βGR mAb
To explore the role of the human Dectin-1 (βGR) on primary cells, we generated novel mAb from C57BL/6 mice immunized with NIH3T3 cells expressing the full-length βGR isoform, βGR-A, as described in the “Materials and methods” section. Two mAb were identified, GE2 (IgG1), which recognizes both of the functional βGR isoforms – βGR-A and βGR-B – and BD6 (IgG2b), which recognizes βGR-A only (Fig. 1A). The specificity of these antibodies was demonstrated by specific staining of the respective βGR isoforms on the surface of live transfected cells, demonstrating that they recognized extracellular epitopes. Given their specificities, and the structure of the receptor, it is likely that GE2 recognizes the carbohydrate-recognition domain whereas BD6 recognizes either the stalk region or a membrane-proximal epitope on the CRD.
Figure 1. Fig. 1. Characterization of novel anti-human-βGR antibodies. (A) Flow-cytometric analysis of NIH3T3 fibroblasts expressing βGR-A (solid histograms), βGR-B (dotted histograms) or untransfected control cells (grey histograms) stained with GE2 or BD6, as indicated. GE2 recognized both βGR-A and βGR-B, whereas BD6 recognized βGR-A only. (B) Inhibition of zymosan recognition by transfected NIH3T3 fibroblasts using GE2, BD6 or the soluble β-glucan (GluP), as indicated. GE2, but not an isotype-matched control or BD6, inhibits zymosan recognition to a level comparable with GluP, demonstrating that it can block human Dectin-1 function. Shown are mean ± SD of data pooled from at least three independent experiments.
Download figure to PowerPoint
As we had previously been successful in using an inhibitory mAb to examine the expression and function of Dectin-1 in primary murine cells 6, 7, we assessed the ability of GE2 and BD6 to inhibit recognition of the β-glucan-rich particle, zymosan, by NIH3T3 cells expressing βGR-A or βGR-B (Fig. 1B). Cells expressing βGR-A or βGR-B were able to recognize zymosan and intact yeast, in a β-glucan-dependent fashion 10 and, as expected, zymosan binding by cells expressing these receptors was inhibited by the soluble β-glucan, glucan phosphate (GluP). Furthermore, GE2, but not an isotype control, was able to inhibit zymosan binding to a similar degree to that obtained by GluP, indicating that the mAb could inhibit receptor function. BD6 had no effect on zymosan binding. Thus we have generated isoform-specific mAb that can be used to define the expression and function of the human βGR in primary cells.
Distribution of the βGR in human peripheral blood cells
The activity of βGR has been described on a variety of human leukocytes including monocytes 17, macrophages 18, eosinophils 19, neutrophils 20 and NK cells 21, and we had previously detected βGR mRNA in some of these cell populations 10. To demonstrate the presence of this receptor on these cells, we examined surface expression of βGR on human PBL by flow cytometry, using GE2 to detect both major functional βGR isoforms. Although the levels of the βGR were relatively low, a number of cell populations expressing this receptor were detected, mostly consistent with the previous data described above (Fig. 2A). Cells were initially separated into three major populations – granulocytes (gate R1), monocytes (gate R2) and lymphocytes (gate R3) – based on size and granularity, and then further subdivided using cell-specific markers.
Figure 2. Expression of βGR in peripheral blood, detected with GE2. (A) Flow-cytometric analysis of peripheral blood, after erythrocyte lysis and blocking, as detailed in the ”Materials and methods“ section, showing the identification and gating of the three major blood populations – granulocytes (R1), monocytes (R2) and lymphocytes (R3) – as well as whole-blood staining showing that all three populations contain βGR-expressing cells. (B) Demonstration of expression of βGR on both major granulocyte populations – eosinophils and neutrophils. Although βGR expression was always detected on eosinophils, the level of expression was donor variable. (C) Expression of βGR on defined monocyte populations, including recruited and inflammatory cells, and DC. (D) Expression of βGR was also detectable on B lymphocytes and CD4+ T lymphocytes. These data are representative of at least three independent donors. The gray histograms represent the isotype controls and the filled dark histograms represent GE2 staining, as indicated.
Download figure to PowerPoint
In the granulocyte population, both neutrophils (CD15+CD16+) and eosinophils (CD15+CD16–) expressed the βGR on their cell surface, although the level of expression on eosinophils was more variable (Fig. 2B). The presence of the βGR on eosinophils is of particular interest, as a number of studies have implicated these cells in allergic pulmonary disease following inhalation of β-glucans 22, 23. Indeed, β-glucans have been shown to induce activation of these cells in vitro19. However, Dectin-1 is not expressed on these cells in the mouse 24, which have similar pulmonary responses to these carbohydrates 25, suggesting that expression of this receptor on human eosinophils may not be linked with disease.
Receptors for β-glucan were initially defined on human monocytes as receptors for unopsonised zymosan 26, 27, and the presence of the βGR on these cells is consistent with these early findings (Fig. 2C). Monocytes have been divided into two main subsets, now termed the “recruited” (CD14lowCD16+) and “inflammatory” (CD14+CD16–) cells 28, 29, both of which have been shown to recognize zymosan 30. We detected the βGR on the surface of both these cell types as well as on a third, intermediate population (CD14+CD16low), whose function is unclear. Expression of βGR was also detected on peripheral blood myeloid DC (CD1c+CD19–), consistent with previous reports 13, but not on BDCA2+ plasmacytoid DC.
A fraction of the lymphocytes [R3: low forward scatter (FSC) and side scatter (SSC)] expressed the βGR, which upon further staining were identified as (CD19+) B cells and (CD3+CD4+) T cells (Fig. 2D). In some donors, a potentially activated CD3lowCD4low T cell population was observed that had higher levels of βGR surface expression (data not shown). The presence of βGR has not previously been defined on lymphocytes, although we had detected Dectin-1 mRNA expression in B and T cell lines 10, and there are some reports that their activities can be influenced by these carbohydrates 31, 32. In addition to the recognition of β-glucans, the βGR on these cells may also be involved in mediating interactions with other lymphocytes through recognition of the unidentified endogenous ligand 9, 10, 13.
It has been shown that β-glucans modulate NK cell function through a receptor distinct from CR3 21, 33. However, we did not detect the βGR on these cells (CD56+CD16+) (Fig. 2D), suggesting other receptor(s) were mediating these activities. Although βGR is classified as an NK-like C-type lectin, the lack of expression on NK cells is not surprising, given that Dectin-1 forms part of a subgroup of these receptors that is predominantly expressed on myeloid and other cells 34. All other lymphocyte populations, including NKT cells (CD56+CD3+) and CD8+ T cells, were essentially negative for βGR expression.
The human βGR is alternatively spliced into two major isoforms – βGR-A and βGR-B – which differ by the presence and absence of a stalk region, respectively 10. By Northern blotting, these two molecules appeared to be differentially expressed in granulocytes (expressing βGR-A and βGR-B) and monocytes/macrophages (expressing βGR-B only) 10. To explore this in more detail, we examined the distribution of βGR-A on peripheral blood cells using the mAb BD6 (Fig. 3A). Although BD6 staining was as efficient as GE2 on transfected cells (Fig. 1), it stained primary cells poorly. This may be due to low levels of this isoform on the cell surface, but as human Dectin-1 is thought to form complexes with at least one other receptor, CD63 35, the poor staining may also be due to steric hindrance of the epitope recognized by this mAb. Nevertheless, βGR-A was detected on the various cell populations that had been identified as GE2+ (Fig. 3B), although some donor variability in the levels of this receptor was observed on monocytes and CD4+ T cells (data not shown). Thus, differential expression of the βGR isoforms, particularly on monocytes, was not apparent on human peripheral blood cells.
Figure 3. Expression of βGR-A detected with BD6 on peripheral blood cells. Although βGR-A could only be poorly detected with BD6 on the surface of live peripheral cells, as shown in the whole-blood staining (A), the presence of this isoform was detected on all populations gated as shown in Fig. 2A (B). The data are representative of at least three different donors, although some donor variability in receptor levels on monocytes and T lymphocytes was noted (data not shown). The gray histograms represent the isotype controls and the filled dark histograms represent BD6 staining.
Download figure to PowerPoint
βGR expression during monocyte maturation
In tissues, monocytes are recruited from the blood and they differentiate into macrophages and DC. As we had demonstrated that Dectin-1 was highly expressed on inflammatory murine macrophages and some murine tissue macrophages, and DC 24, 36, we next examined the changes in human βGR expression during monocyte differentiation in vitro (Fig. 4). For this analysis, peripheral blood monocytes were cultured as described in the “Materials and methods” section, and examined for βGR expression at various time-points with other known surface markers, to monitor cellular differentiation. During monocyte differentiation into macrophages, characterized by the down-regulation of CD14 and the up-regulation of HLA-DR 37, 38, βGR expression tended to decrease, but remained detectable by day 7. The decrease in expression was not due to internalization of the receptor, as no intracellular pools of Dectin-1 were detected (data not shown). Expression of βGR-A, as detected by BD6, decreased during maturation into macrophages and was not detectable on any donor by day 7.
Figure 4. Expression of βGR-A and βGR-B on monocyte-derived macrophages and DC. The expression of βGR, as detected by GE2, was observed to decrease during differentiation into macrophages but to increase during differentiation into immature DC. The expression of βGR-A, detected with BD6, was low/absent on both cell types. Maturation of DC with Salmonella LPS greatly reduced βGR expression. HLA-DR, CD14, DC-SIGN and CD86 were included as markers of macrophage and DC maturation, as described in the text. The gray histograms represent the isotype controls and the filled dark histograms represent specific antibody staining, as indicated. The data are representative of at least three different donors.
Download figure to PowerPoint
The expression of the βGR was also examined with GE2 on monocytes cultured in GM-CSF and IL-4, to induce a DC-like phenotype 39 (Fig. 4). Elevated levels of the βGR were detected on the immature DC, which express DC-SIGN and reduced levels of CD14, at day 4 but maturation of these cells with LPS, characterized by the lack of CD14 and up-regulation of CD86 40, 41, resulted in down-regulation of the βGR. These changes in βGR expression may reflect similar regulatory mechanisms to Dectin-1 on murine macrophages, where has been shown to be up-regulated by IL-4 and GM-CSF and down-regulated by LPS 42. Little or no βGR-A was detected with BD6 on any DC population.
The lack of detectable expression of βGR-A on DC and mature macrophages indicates that this receptor does undergo cell-specific control of alternative splicing during monocyte maturation, as previously mentioned 10. This further suggests that the various isoforms may serve specific roles in each cell type. As βGR-A and βGR-B are equally functional for β-glucan recognition 10, there may be differences in their ability to interact with T cells. Indeed, the contact site between antigen-presenting cells and T cells (the immunological synapse) is known to require particular molecular spacing and redistribution of receptors 43, and the presence or absence of a stalk region, as found in the βGR isoforms, is perhaps suggestive of a role in this process.
βGR expression in an in vivo inflammatory model
We next examined the expression of the βGR in an inflammatory setting, using a skin-window model to study expression of this receptor on recruited myeloid cells ex vivo. In this model, cells recruited onto filter-paper discs, which have been placed onto skin abrasions, are harvested and analyzed for surface markers by flow cytometry and compared with PBL isolated at the same time (A. S. J. Marshall et al., manuscript in preparation) (Fig. 5A). The recruited myeloid populations could not be easily separated as they overlapped on FSC and SSC, but were identified by CD86 and CD15 staining into granulocytes (CD86–CD15+) and macrophages (CD86+CD15–) (data not shown; A. S. J. Marshall et al., manuscript in preparation). In comparison with the peripheral blood cells, βGR expression was maintained on recruited macrophages, but was slightly, but consistently, down-regulated on the recruited granulocytes (Fig. 5B). The decreased expression on recruited granulocytes may reflect βGR released on microvesicles, which have been shown to be induced upon granulocyte activation and which are capable of blocking cellular responses to zymosan 44, 45.
Figure 5. Expression of βGR in an in vivo inflammatory model. (A) Histograms showing that βGR expression is maintained on monocyte/macrophages (MΦ; gated as CD86+CD15–) but is decreased on granulocytes (Grn; gated as CD86–CD15+). CD11b and CD86 are also shown as markers of activation for granulocytes and monocyte/macrophages, respectively. The gray histograms represent the isotype controls and the filled dark histograms represent specific antibody staining, as indicated, and the data are representative of three different donors. (B) Quantitation of GE2 expression levels, shown in (A) and represented as mean ± SD of data pooled from two independent donors.
Download figure to PowerPoint
Overall, it is notable that there are differences in expression between human and mouse Dectin-1. In the mouse, expression of Dectin-1 appears to be more restricted and has not been detected on eosinophils or on B cells 24. Furthermore, in contrast to the murine receptor, the human βGR was not up-regulated on macrophages in culture or highly expressed on recruited inflammatory cells. Although the significance of these differences is unclear, they should be considered when utilizing the mouse as a model system for this receptor, once a knockout becomes available.
Function of the βGR on primary human macrophages
A number of receptors, including CR3 14, scavenger receptors 15 and lactosylceramide 16, have been proposed to recognize β-glucans on human leukocytes. As we had shown that Dectin-1 was the major β-glucan receptor on murine macrophages 7, we next examined the contribution of this receptor to zymosan recognition in human cells (Fig. 6A). Zymosan binding by monocyte-derived macrophages was significantly inhibited in the presence of GluP, indicating that the recognition of these particles is mainly mediated through β-glucan-dependent mechanisms. The ability of GE2, but not an isotype control, to inhibit zymosan recognition in an equivalent way to GluP demonstrated that the human βGR was responsible for this activity.
Figure 6. Human Dectin-1 is the major βGR on monocyte-derived macrophages and contributes to the proinflammatory response to fungal particles. (A) Zymosan binding by day-7 matured monocyte-derived macrophages is β-glucan-dependent, as demonstrated by inhibition with GluP. Equivalent inhibition with GE2 demonstrates that this activity is mediated by the βGR. (B) TNF-α production in response to zymosan is β-glucan-dependent and requires the βGR. Shown is mean ± SD of data pooled and normalized from three independent donors in triplicate whose absolute TNF-α levels (pg/ml) were 231.00±8.95, 2236.00±70.44 and 1065.00±92.38. Macrophage TNF-α production was not detectable in the absence of zymosan.
Download figure to PowerPoint
In murine macrophages, we have also shown that Dectin-1 collaborates with the TLR to induce proinflammatory cytokine production 4. To assess the role of the human βGR in this process, we incubated monocyte-derived macrophages with zymosan in the presence or absence of GluP, GE2 or an isotype control and assayed for the production of TNF-α. Although the absolute amount of TNF-α produced was donor variable, the presence of GluP or GE2, but not an isotype control, significantly inhibited TNF-α production in response to zymosan in all donors. Thus, like its murine homologue, the human βGR acts as the major receptor for β-glucans on macrophages and contributes to the inflammatory response to these particles.
It has also been reported that βGR are involved in the recognition of nontypeable Haemophilus influenzae in a variety of human cells, including monocytes and eosinophils 46, 47. The presence of Dectin-1 on both these cell types was suggestive that this receptor may be involved in the recognition of these organisms. However, we have been unable to demonstrate any interaction between this bacterium and either βGR-A or βGR-B (data not shown), suggesting that receptor(s) other than Dectin-1 are involved in this activity.
In summary, we have shown that human Dectin-1 is widely expressed on leukocytes and is not myeloid restricted. Although the expression patterns and regulation of the human βGR differs from those of the murine receptor, it appears to serve the same immune function. Future studies should address the identity of the endogenous ligand and the significance and roles of the alternatively spliced βGR isoforms.