Downregulation of the metalloproteinases ADAM10 or ADAM17 promotes osteoclast differentiation

Abstract

Bone resorption is driven through osteoclast differentiation by macrophage colony-stimulating factor (M-CSF) and receptor activator of nuclear factor kappa-Β ligand (RANKL). We noted that a disintegrin and metalloproteinase (ADAM) 10 and ADAM17 are downregulated at the expression level during osteoclast differentiation of the murine monocytic cell line RAW264.7 in response to RANKL. Both proteinases are well known to shed a variety of single-pass transmembrane molecules from the cell surface. We further showed that inhibitors of ADAM10 or ADAM17 promote osteoclastic differentiation and furthermore enhance the surface expression of receptors for RANKL and M-CSF on RAW264.7 cells. Using murine bone marrow-derived monocytic cells (BMDMCs), we demonstrated that a genetic deficiency of ADAM17 or its required regulator iRhom2 leads to increased osteoclast development in response to M-CSF and RANKL stimulation. Moreover, ADAM17-deficient osteoclast precursor cells express increased levels of the receptors for RANKL and M-CSF. Thus, ADAM17 negatively regulates osteoclast differentiation, most likely through shedding of these receptors. To assess the time-dependent contribution of ADAM10, we blocked this proteinase by adding a specific inhibitor on day 0 of BMDMC stimulation with M-CSF or on day 7 of subsequent stimulation with RANKL. Only ADAM10 inhibition beginning on day 7 increased the size of developing osteoclasts indicating that ADAM10 suppresses osteoclast differentiation at a later stage. Finally, we could confirm our findings in human peripheral blood mononuclear cells (PBMCs). Thus, downregulation of either ADAM10 or ADAM17 during osteoclast differentiation may represent a novel regulatory mechanism to enhance their differentiation process. Enhanced bone resorption is a critical issue in osteoporosis and is driven through osteoclast differentiation by specific osteogenic mediators. The present study demonstrated that the metalloproteinases ADAM17 and ADAM10 critically suppress osteoclast development. This was observed for a murine cell line, for isolated murine bone marrow cells and for human blood cells by either preferential inhibition of the proteinases or by gene knockout. As a possible mechanism, we studied the surface expression of critical receptors for osteogenic mediators on developing osteoclasts. Our findings revealed that the suppressive effects of ADAM17 and ADAM10 on osteoclastogenesis can be explained in part by the proteolytic cleavage of surface receptors by ADAM10 and ADAM17, which reduces the sensitivity of these cells to osteogenic mediators. We also observed that osteoclast differentiation was associated with the downregulation of ADAM10 and ADAM17, which reduced their suppressive effects. We therefore propose that this downregulation serves as a feedback loop for enhancing osteoclast development.

Keywords: Colony stimulating factor 1 receptor; Knockout mice; Metalloproteinase; Monocyte; Osteoclast; Receptor activator of nuclear factor kappa-Β ligand; Shedding.

Fig. 1

Osteoclastic differentiation and regulation of mRNA expression in RAW 264.7 cells in response to RANKL. RAW 264.7 cells were stimulated with RANKL for 5 days or left unstimulated. (A–C) After fixation and permeabilization, the cells were stained for TRAP and with DAPI and phalloidin. Representative images of TRAP staining of (A) unstimulated cells, (B) stimulated cells and (C) the corresponding fluorescence images of stimulated cells are shown. (D–I) After 4 d of stimulation of RAW 264.7 cells with and without RANKL, the mRNA expression of the osteoclast-associated marker genes Acp5/TRAP (D), Mmp9 (E), Ctsk (F) and Nfatc1 (G) as well as Adam10 (H), Adam17 (I) and Rhbdf2 (J) was determined by qPCR. The expression of these genes was related to that of the reference genes Eef2 and Gapdh. Quantitative data points are shown for each independent experiment and are summarized as the mean + SD. Significant differences are indicated by asterisks (* p < 0.05, ** p < 0.01 and *** p < 0.001)

Fig. 2

Effect of ADAM inhibitors on the osteoclastic differentiation of RAW 264.7 cells. (A–D) RAW 264.7 cells were stimulated with RANKL in the presence of the vehicle control DMSO (A), 10 µM GI (B) or 20 µM TAPI (C). TRAP staining was performed after 5 d. Representative images are shown. Osteoclast differentiation was quantified as the area of TRAP-positive cells (D). (E–G) After 4 d of stimulation with RANKL, the mRNA expression of the osteoclast-associated marker genes Acp5/TRAP (E), Mmp9 (F) and Ctsk (G) was determined by qPCR. (H, I) After 2 days, the surface levels of RANK (H) and CSF1R (I) were assessed by flow cytometry. Representative histograms of fluorescence intensities are shown. To quantify surface expression of RANK and CSF1R, geometric mean values of fluorescence intensities were determined and nonspecific binding of an isotype control antibody was subtracted. Quantitative data are presented as the mean + SD of independent experiments, and significant differences are indicated by asterisks (* p < 0.05, ** p < 0.01 and *** p < 0.001)

Fig. 3

Osteoclastic differentiation of BMDMCs and regulation of RANK and CSF1R surface expression levels. (A–C) Isolated BMDMC from mice were differentiated into osteoclasts by stimulation for 7 d with M-CSF and subsequently for 5 d with additional RANKL. Cells that were only treated with M-CSF served as a control. After 12 d, the cells were stained for TRAP together with DAPI and phalloidin. Representative images of TRAP stained murine BMDMCs treated with M-CSF (A) or with both M-CSF and RANKL (B) and the corresponding fluorescence image for B (C) are shown. (D–I) The mRNA expression of the osteoclast-associated marker genes Acp5/TRAP (D), Mmp9 (E) and Ctsk (F) as well as that of Adam10 (G), Adam17 (H) and Rhbdf2 (I) was determined by qPCR. The expression of these genes was related to that of the reference genes Gapdh and Rps29. (J, K) BMDMCs were stimulated with M-CSF in the presence or absence of GI or TAPI. After 7 d, the surface levels of RANK (J) and CSF1R (K) were assessed by flow cytometry. The quantitative data are presented as the means + SD of independent experiments, and significant differences are indicated by asterisks (* p < 0.05, ** p < 0.01 and *** p < 0.001)

Fig. 4

Effect of ADAM17 and iRhom2 knockout on osteoclastic differentiation and surface levels of RANK and CSF1R on BMDMCs. (A–H) ADAM17- (A–E) or iRhom2-deficient (F–H) BMDMCs (KO) and BMDMCs from corresponding wild-type (WT) mice were stimulated with M-CSF on day 0 and subsequently with additional RANKL on day 7. After day 11, the cells were fixed and stained for TRAP. Representative images of the osteoclastic differentiation of WT (A) and ADAM17-deficient BMDMCs (B) are shown. For quantitative analysis, the number and total area of TRAP-positive cells were determined per image and are displayed as data points for each independent experiment (C, D, F, G), and the average osteoclast size was calculated from the total osteoclast area divided by the number of osteoclasts (E, H). The surface levels of RANK and CSF1R on WT or ADAM17-deficient BMDMCs were assessed at day 7 of differentiation by flow cytometry (I, J). The quantitative data are presented as the means + SD of independent experiments, and significant differences are indicated by asterisks (* p < 0.05, ** p < 0.01 and *** p < 0.001)

Fig. 5

Effect of ADAM10 inhibition on the osteoclastic differentiation of BMDMCs with and without ADAM17 knockout. (A–I) WT or ADAM17-deficient BMDMCs were treated with GI (10µM) or DMSO together with M-CSF on day 0 or together with RANKL on day 7, and the effect on osteoclast differentiation was analyzed on day 12. Representative images of TRAP-stained WT BMDMCs are shown (A–C). The number, total area and average size of TRAP-positive cells were determined per image and are displayed as data points for each independent experiment (D–I). The quantitative data are presented as the means + SD of independent experiments, and significant differences are indicated by asterisks (* p < 0.05, ** p < 0.01 and *** p < 0.001)

Fig. 6

Effect of ADAM10 or ADAM17 inhibition on the osteoclastic differentiation of human PBMCs. (A–I) PBMCs from healthy donors were treated with 10 µM GI (A, D, G) or 20 µM TAPI (B, E, H) or DMSO (C, F, I) and were stimulated with M-CSF beginning on day 0. On day 7, half of the cells additionally received RANKL (G–I). On day 17, the effect on osteoclast cell number and area was analyzed. The data are presented as representative images (A–C) and as the mean + SD (D–I) of independent experiments. Significant differences are indicated by asterisks (* p < 0.05, ** p < 0.01 and *** p < 0.001)