Regulation of human osteoclast development by dendritic cell-specific transmembrane protein (DC-STAMP)

Abstract

Osteoclasts (OC) are bone-resorbing, multinucleated cells that are generated via fusion of OC precursors (OCP). The frequency of OCP is elevated in patients with erosive inflammatory arthritis and metabolic bone diseases. Although many cytokines and cell surface receptors are known to participate in osteoclastogenesis, the molecular mechanisms underlying the regulation of this cellular transformation are poorly understood. Herein, we focused our studies on the dendritic cell-specific transmembrane protein (DC-STAMP), a seven-pass transmembrane receptor-like protein known to be essential for cell-to-cell fusion during osteoclastogenesis. We identified an immunoreceptor tyrosine-based inhibitory motif (ITIM) in the cytoplasmic tail of DC-STAMP, and developed an anti-DC-STAMP monoclonal antibody 1A2 that detected DC-STAMP expression on human tumor giant cells, blocked OC formation in vitro, and distinguished four patterns of human PBMC with a positive correlation to OC potential. In freshly isolated monocytes, DC-STAMP(high) cells produced a higher number of OC in culture than DC-STAMP(low) cells and the surface expression of DC-STAMP gradually declined during osteoclastogenesis. Importantly, we showed that DC-STAMP is phosphorylated on its tyrosine residues and physically interacts with SHP-1 and CD16, an SH2-domain-containing tyrosine phosphatase and an ITAM-associated protein, respectively. Taken together, these data show that DC-STAMP is a potential OCP biomarker in inflammatory arthritis. Moreover, in addition to its effect on cell fusion, DC-STAMP dynamically regulates cell signaling during osteoclastogenesis.

Figure 1. Functional characterization of anti-DC-STAMP mAb 1A2

(A) A positive association was noted between DC-STAMP and CD16 expression. CD14+CD16+ monocytes demonstrated a higher surface expression of DC-STAMP than CD14+CD16-cells. Human PBMC were purified by Ficoll gradient and stained with an antibody cocktail composed of 7-AAD, anti-DC-STAMP and anti-CD16 antibodies. The expression of DC-STAMP on unstained, isotype control, CD14+CD16- and CD14+CD16+ cells are labeled in red, blue, orange and green, respectively. Commercially available anti-DC-STAMP polyclonal antibody KR104 was used for this analysis. (B) Lane 2: total proteins isolated from the RAW cell line; lane 3: proteins isolated from a RAW cell line expressing the PTHR-DC-STAMP fusion protein; lane 4: immunoprecipitated human monocyte proteins by 1A2; lane 5: immunoprecipitated human monocyte proteins by mouse IgG2a. All proteins were denatured, separated by 10% gradient protein gel, and probed with the anti-DC-STAMP mAb 1A2 on western blots. Pink and blue asterisks label the PTHR-DC-STAMP fusion protein and DC-STAMP native protein, respectively. (C) DC-STAMP expression on human PBMC and multinucleated giant cells from giant cell tumor of bone was detected by immunohistochemical (IHC) staining using 1A2, (a) & (b). Human PBMC were purified by Ficoll gradient, embedded in paraffin for section, and stained with (a) mouse IgG2a isotype control, or (b) 1A2. (c) & (d). Human biopsy samples collected from giant cell tumor were sectioned, and stained with (c) mouse IgG2a isotype control, or (d) 1A2. Both 1A2 and mouse IgG2a isotype control were diluted at 1:1500 for staining. The polarized expression of DC-STAMP in the multinucleated giant cells from a giant cell tumor was labeled by arrows. (D) The anti-DC-STAMP mAb 1A2 was able to block OC formation in vitro. Enriched human monocytes were cultured in the absence (a) or presence (b) of 1A2 for 8 days and TRAP-stained for visualization and enumeration of OC. (c) Enriched human monocytes and PBMC isolated from different subjects were cultured in the absence (open bar), presence of 1A2 (solid bar) or with IgG2a isotype control (slash bar) for 8 days. The concentrations of 1A2 used for monocytes and PBMC were 15 μg/ml and 150 μg/ml, respectively.

Figure 2. DC-STAMP is expressed on the surface of monocytes and a small subset of CD3+ cells on human PBMC

(A) Analysis of DC-STAMP expression on human PBMC. Human PBMC were purified from whole blood by Ficoll gradient and subjected to antibody staining and flow cytometry analysis. Human PBMC were stained with an antibody cocktail composed of 6 antibodies (see Materials and Methods for details). (a) Dead (7AAD+) cells were first excluded from our analysis; and (b) live PBMC were gated based on cell size by FSC and cell granularity by SSC into 3 cell subsets (P1, P2, and P3). The expression of CD14 and DC-STAMP on the P1, P2 and P3 subset is shown in (c), (d) and (e), respectively. (f) The surface expression of DC-STAMP in P1-(blue), P2-(green), P3-(purple) gated cells. The red line represents the isotype control. Data shown is representative of 8 subjects. (B) DC-STAMP is expressed on a small subset of CD3+ cells. Human PBMC were purified and stained with an antibody cocktail composed of 6 antibodies (see Materials and Methods for details). (a) Monocytes, T and B cells were identified by gating of CD14+, CD3+ and CD19+ cells, respectively. The histogram shows the overlay of DC-STAMP expression on CD14+ (green line), CD3+ (blue line), and CD19+ (red line) populations. A small percentage of CD3+ are DC-STAMP+ (indicated by arrow). (b) The relative expression of DC-STAMP and CD3 on human PBMC. Data shown here represent staining phenotypes from 3 healthy controls.

Figure 3. Human PBMC have four distinct DC-STAMP expression patterns that differ between Ps/PsA and HC subjects

Four distinct DC-STAMP expression patterns were observed on human PBMC. PBMC were isolated from 11 healthy control (HC) and 21 PsA subjects and stained with the 1A2-FITC antibody. Dead cells were excluded by 7-ADD. Green lines represent the isotype control. The number of subjects observed in each pattern. Fisher’s exact revealed significant difference among HC and PsA in 4 DC-STAMP patterns (p-value=0.01). Table 2 summarizes the definition of 4 DC-STAMP patterns and the distribution of HC and PsA patients in these 4 patterns.

Figure 4. DC-STAMP is down-regulated in human monocytes during ostoeclastogenesis

(A) Dynamic changes of DC-STAMP surface expression on human monocytes during osteoclastogenesis. Enriched human monocytes were cultured in media supplemented with RANKL and M-CSF, and the surface expression of DC-STAMP was examined at different time points (a:day0, b:day1, c:day2, d:day5, e:day7). Purple areas in each panel represent the original DC-STAMP expression level on fresh monocytes and open green lines show the expression of DC-STAMP at the various time points. Isotype control is shown in red (a). (B) Cellular localization of DC-STAMP. Human monocytes were cultured with RANKL+M-CSF for 8 days, fixed and immunostained with DAPI which binds to nuclei (blue), 1A2 anti-DC-STAMP-FITC (green), and rhodamine phalloidin for actin (red). Localization of DC-STAMP on a spindle-shaped pro-OC (a) and on mature OC (b). Cells shown in (a) and (b) were cultured on a single slide with the same magnification. The images are representative of ten cells with similar phenotypes (mononuclear vs. multi-nucleated, and spindle-shaped vs. large round shape). (C) Overexpression of DC-STAMP on RAW 264.7 cells. RAW 264.7 cells were left untransfected (a), transfected with the pCMV6-entry vector (b), or the pCMV6-DC-STAMP (c) plasmid. (a) to (c) showed the phenotypes of cells 16-hr post-transfection. A multinucleated cell which went thru cell-to-cell fusion more than once as evidenced by the presence of 4 nuclei was labeled by an arrow in (c). RANKL was added to the cultures 24-hr post-transfection to promote OC formation. The phenotypes of cells after 3 days in culture were shown from (d) to (f).

Figure 5. Human DC-STAMP^high cells demonstrate a higher osteoclastogenesis potential

(A) Human monocytes were negatively selected and enriched by resetting (StemCell). (a) human PBMC before enrichment; (b) human PBMC after enrichment. (B) DC-STAMP expression levels on enriched monocytes ((A)-b), monocytes before enrichment, and lymphocyte ((A)-a). (C) Gating strategy of human monocytes based on the DC-STAMP expression. Human monocytes were first enriched by a monocyte enrichment kit (StemCell), stained with 1A2-FITC and 7-AAD. Live monocytes (7-AAD negative) were gated as shown in Figure 5(A)-b and sorted into DC-STAMP^high and DC-STAMP^low (1.9% and 1.8% of the highest and lowest). (D) Bone resorption activity of DC-STAMP^high and DC-STAMP^low cells. Sorted DC-STAMP^high and DC-STAMP^low cells shown in (A) were cultured with bone wafers for 14 days in the presence of RANKL and M-CSF. Numbers in parentheses represent the total number of TRAP+ OC per 10⁵ sorted cells. OC and erosion pits on bone wafers by DC-STAMP^low and DC-STAMP^high cells were shown in (a & b) and (c & d), respectively. Representative of three individual experiments performed on HC.

Figure 6. DC-STAMP proteins are phosphorylated on tyrosine residues, interact with SHP-1 and CD16, and may be involved in signal transduction

(A) Proteins were isolated from fresh human monocytes, immunoprecipitated (IP) with anti-DC-STAMP antibody 1A2 (lane 2) or with anti-CD16 antibody (lane 3). IP lysates were analyzed by SDS-PAGE followed by immunoblotting (IB) with 1A2. Arrow shows the DC-STAMP band (~53 kDa). (B) Proteins were isolated from fresh human monocytes, IP with anti-SHP1 antibody (lane 2) or with the control antibody mouse IgG (lane 3). IP lysates were analyzed by SDS-PAGE followed by IB with 1A2. Arrow shows the DC-STAMP band (~53 kDa). (C) Proteins were isolated from fresh human monocytes, IP with mouse IgG2a (lane 2), or anti-SHP1 antibody (lane3) or anti-DC-STAMP antibody 1A2 (lane 4). IP lysates were analyzed by SDS-PAGE followed by IB with anti-phosphotyrosine antibody 4G10. The bands of SHP-1 (~70 kDa) and DC-STAMP (~53 kDa) are marked by arrows. (D) Examination of 1A2 effects on DC-STAMP signaling. (I) Human monocytes were cultured in OC-promoting media (RANKL+M-CSF) in the absence (lanes 2 and 4) or presence (lanes 3 and 5) of 1A2 for 8 days. Proteins were isolated and IP with mouse IgG2a (lanes 2 and 3) or with anti-SHP-1 antibody (lanes 4 and 5), and subjected to SDS-PAGE, followed by IB with anti-phosphotyrosine antibody 4G10. Arrow indicates the location of the SHP-1 band. (II) Purified human monocytes were cultured in OC-promoting media (RANKL+M-CSF) in the absence (a to c) or the presence of 1A2 (d to f), or in the presence of mouse IgG2a isotype control (g to i). Cells were harvested and analyzed at 3 different time points (16hrs: a &d; 40 hrs: b & e; 64 hrs: c & f) for phosphorylated PLC-γ2 expression by intracellular staining with the anti- PLC- γ2 antibody (pY759) and flow cytometric analysis.

Figure 7. A model of signaling cascades induced by DC-STAMP, CD16, and RANK during osteoclast differentiation

This diagram is a modification of several models previously proposed by Humphrey et al.,⁽¹⁴⁾ Nimmerjahn & Ravetch,^(40,41) and Nakashima & Takayanagi.⁽¹²⁾ Briefly, DC-STAMP and CD16 are activated after ligand binding, which triggers the phosphorylation of ITIM and ITAM to recruit SHP-1 and Syk, respectively. The balance (net readout) between the activation and inhibitory signals delivered from ITAM and ITIM will determine if downstream osteoclastogenic genes will be activated through PLC-γ and Ca²⁺-NFATc1 regulation. CD16 and DC-STAMP were selected as the representatives of ITAM- and ITIM-bearing molecules in this figure. However, other ITAM- (OSCAR and TREM-2) and ITIM- (LILRB, PIRB and Ly49Q) bearing molecules also modulate calcium signaling through PLCγ. A more detailed explanation of this model can be found in the “Discussion”.