Osteoclast-specific cathepsin K deletion stimulates S1P-dependent bone formation

Cathepsin K (CTSK) is secreted by osteoclasts to degrade collagen and other matrix proteins during bone resorption. Global deletion of Ctsk in mice decreases bone resorption, leading to osteopetrosis, but also increases the bone formation rate (BFR). To understand how Ctsk deletion increases the BFR, we generated osteoclast- and osteoblast-targeted Ctsk knockout mice using floxed Ctsk alleles. Targeted ablation of Ctsk in hematopoietic cells, or specifically in osteoclasts and cells of the monocyte-osteoclast lineage, resulted in increased bone volume and BFR as well as osteoclast and osteoblast numbers. In contrast, targeted deletion of Ctsk in osteoblasts had no effect on bone resorption or BFR, demonstrating that the increased BFR is osteoclast dependent. Deletion of Ctsk in osteoclasts increased their sphingosine kinase 1 (Sphk1) expression. Conditioned media from Ctsk-deficient osteoclasts, which contained elevated levels of sphingosine-1-phosphate (S1P), increased alkaline phosphatase and mineralized nodules in osteoblast cultures. An S1P_1,3 receptor antagonist inhibited these responses. Osteoblasts derived from mice with Ctsk-deficient osteoclasts had an increased RANKL/OPG ratio, providing a positive feedback loop that increased the number of osteoclasts. Our data provide genetic evidence that deletion of CTSK in osteoclasts enhances bone formation in vivo by increasing the generation of osteoclast-derived S1P.

One of the major issues with the use of currently available antiresorptive drugs for the treatment of osteoporosis is that while they are very efficient at reducing osteoclast numbers and activity, leading to markedly reduced bone resorption, they also profoundly reduce bone formation. The reduction in activation frequency of bone remodeling units along bone surfaces and the coupling between bone resorption and formation result in a decrease in bone formation following treatment (1). Furthermore, it has been argued that such profound inhibition of bone turnover could be contributing to rare, but significant, clinical complications such as osteonecrosis of the jaw (ONJ) and atypical subtrochanteric femoral fractures (2–5). The profound decrease in bone remodeling associated with these drugs has also been suggested to decrease the responsiveness of the skeleton to parathyroid hormone (PTH), the only anabolic drug currently available in the clinic (6). In this context, it became important to identify novel therapeutic pathways through which bone resorption could be inhibited efficiently while maintaining bone turnover and bone formation at clinically acceptable levels.

Information about what these novel therapeutic approaches could actually accomplish came from the identification of several genes whose mutation or deletion caused osteopetrosis, a phenotype in which defects in osteoclasts lead to reduced bone resorption and high bone density. Careful analysis of the changes in bone remodeling activity in these human and/or murine mutants showed that, as with currently available antiresorptive drugs, bone formation was always decreased in mutations leading to deficient osteoclast differentiation and numbers, such as those affecting RANK or RANKL (7, 8). In contrast, bone formation was maintained or even increased in mutations leading to strictly functional defects in osteoclasts, such as mutations of the chloride channel CLC-7, VATPase subunits, or carbonic anhydrase (9). One such mutation affects cathepsin K, a cysteine protease secreted by osteoclasts that is essential for the degradation of matrix collagen and the activation of tartrate-resistant acid phosphatase (TRAP) (10, 11).

In humans, mutation of cathepsin K leads to pycnodysostosis, a rare autosomal recessive skeletal dysplasia in which osteoclast function is defective. It is characterized by short stature, osteopetrosis, acroosteolysis (resorption of distal phalanges), spondylolysis, delayed cranial suture closure, and bone fragility (12–14). Several mouse models of germline cathepsin K deficiency have also been generated by deleting Ctsk, the gene encoding cathepsin K (15–19). These animal models showed reduced bone resorption, together with normal or increased bone formation, contributing to the osteopetrotic phenotype. These studies therefore confirmed that it is possible to decrease bone resorption without the inevitable reduction in bone formation. Mechanistically, however, Ctsk was globally deleted in all these studies, making it impossible to determine whether the increased bone formation resulted from cell-autonomous effects in cells other than osteoclasts, including cells of the osteoblast lineage, or was secondary to changes in osteoclast-produced osteoanabolic coupling factors.

To address this important mechanistic question, we generated mice in which the Ctsk allele was specifically deleted in cells of the hematopoietic lineage, in the osteoclast lineage, or in the osteoblast lineage using the Cre-loxP system, and analyzed their skeletal phenotypes. We determined the effects of these cell-specific deletions on bone resorption and bone formation, and characterized the molecular mechanism underlying these changes. Targeted deletion of Ctsk in hematopoietic cells, or more specifically, in the osteoclast lineage, resulted in mild osteopetrosis with an increase in the number of poorly functional osteoclasts, as well as an increase in osteoblasts and bone formation. In vitro assays of primary osteoblasts derived from long bones of these conditional knockout mice showed an increase in osteoprogenitors, alkaline phosphatase (ALP), and mineralized bone nodules. In contrast, ablation of Ctsk in cells of the osteoblast lineage had no effect on either bone formation or bone resorption, suggesting that the increased bone formation seen in the mice with germline deletion of Ctsk results from increased coupling between osteoclasts and osteoblasts. A search for molecules that could mediate this increase in the coupling mechanism showed a significant increase in sphingosine kinase 1 (Sphk1) in Ctsk-deleted osteoclasts. Sphk1 catalyzes the phosphorylation of sphingosine to sphingosine-1-phosphate (S1P), which promotes osteoblast differentiation and bone-forming activity (20). Consistent with the increased Sphk1, conditioned medium derived from Ctsk-deleted osteoclasts contained more S1P and ALP in vitro. Furthermore, the RANKL/OPG ratio was increased in osteoblasts derived from long bones of these mice, suggesting a positive feedback loop to increase the number of osteoclasts in the absence of cathepsin K. These results demonstrate that the increased bone formation rate (BFR) observed after deletion of Ctsk is osteoclast mediated, and establish the fact that osteoclasts can indeed promote bone formation in an in vivo setting. Furthermore, these findings provide additional evidence that S1P is an important osteoclast-derived anabolic factor, a clastokine, that couples bone resorption to bone formation.

Generation of a cathepsin K conditional knockout mouse. We generated mice carrying the floxed Ctsk allele by inserting loxP sites flanking exon 5 that encodes for one of the active site cysteine residues of cathepsin K (Ctsk^fl)(Figure 1A). The point mutations in this area are sufficient to cause pycnodysostosis disease in humans. Southern analysis showed a wild-type band of 11 kb and a mutant band of 9.5 kb (Figure 1B). Normal Ctsk mRNA expression was observed in Ctsk^fl/fl mice using Northern analysis (Figure 1C). To ensure that insertion of the loxP sites did not affect bone mass, we examined the long bones of the Ctsk^fl/fl mice by peripheral quantitative CT (pQCT). The analysis showed that cancellous, cortical, and total bone mineral density (BMD) in Ctsk^fl/fl mice was similar to that in Ctsk^+/+ mice (Figure 1D).

Figure 1

Generation of Ctsk^fl/fl mouse line. (A) To generate the targeting construct pG3NEOBLoxP-2, the expression cassette for the neomycin resistance gene (NEO) flanked by loxP sites (large arrowheads) was inserted in reverse orientation between the EcoRI (ERI) and the BamHI (BHI) sites in intron 5 of the Ctsk gene. Successful targeting of the Ctsk locus resulted in insertion of the 3 loxP sites and the NEO gene into the Ctsk gene. Cre recombinase activity resulted in sequential removal of the sequences between the loxP sites including exon 5 and the NEO cassette, leaving only one loxP site in the locus. (B) The results of the Southern hybridization of tail DNA from control, Ctsk^fl/+, and Ctsk^fl/fl mice indicating a band of 11 kb for the wild-type allele, and 9.5 kb for the mutant allele after a digestion with BHI restriction enzyme and hybridization with probe pMCK-5′. (C) Northern analysis of humerus total RNA demonstrating normal Ctsk mRNA expression in Ctsk^fl/fl mice. (D) pQCT analysis shows normal femoral BMD in Ctsk^fl/fl mice.

Hematopoietic cell- and osteoclast-specific deletion of Ctsk. To examine the role of cell-specific targeted deletion of Ctsk in bone remodeling, we first determined the effects of Ctsk deletion in hematopoietic cells, including osteoclasts, using the Mx1-Cre mice (21), or more specifically, in the monocyte-osteoclast lineage using the CD11b-Cre mice (22–24). Ctsk^fl/fl mice were crossed with those expressing Cre recombinase under the control of a type I interferon-inducible Mx1 promoter to generate Ctsk^fl/fl (control) and Mx1;Ctsk^fl/fl (conditional knockout) mice and those expressing Cre recombinase under the control of CD11b promoter to generate CD11b;Ctsk^+/+ (control) and CD11b;Ctsk^fl/fl (conditional knockout) mice.

Induction of Mx1-Cre expression by intraperitoneal injection of polyinosinic-polycytidylic acid (polyI:C) at 2 weeks of age led to the complete deletion of the Ctsk exon 5 in spleen DNA (Figure 2A), reduced Ctsk mRNA levels in femur and osteoclasts by 92% and 94%, respectively (Figure 2B), and markedly diminished cathepsin K protein levels at 9 weeks of age (Figure 2C). It was previously suggested that the Cre recombinase in Mx1-Cre transgenic mice is expressed not only in osteoclasts, but also in osteoblasts (25). However, quantitative real-time PCR (qRT-PCR) analysis of the Mx1;Ctsk^fl/fl mice indicated that the Ctsk gene, which is expressed at much lower levels (0.5%) than in osteoclasts, was not deleted in osteoblasts (Figure 2B). CD11b-driven Cre recombinase expression resulted in deletion of exon 5 in spleen genomic DNA (Figure 2D). Western blotting analysis showed a reduction of more than 90% in cathepsin K protein levels in CD11b;Ctsk^fl/fl osteoclasts (Figure 2F). However, qRT-PCR showed that the efficiency of CD11b-Cre–mediated ablation of Ctsk expression was only 60% and 61% in femur and osteoclasts, respectively (Figure 2E). As expected, CD11b-Cre deletion also did not affect expression of Ctsk in osteoblasts (data not shown).

Figure 2

Ablation of Ctsk in hematopoietic cells (Mx1;Ctsk^fl/fl) and osteoclasts (CD11b;Ctsk^fl/fl) suppresses Ctsk expression. (A and D) PCR of genomic DNA from spleen showed the deletion of Ctsk exon (Δ). (B and E) qRT-PCR analysis of Ctsk showed a decrease in Ctsk mRNA expression in bone and osteoclasts (OC), but not osteoblasts (OB). (C and F) Western blot analysis of cathepsin K in protein lysates from osteoclasts derived from Mx1;Ctsk^fl/fl or CD11b;Ctsk^fl/fl mice and controls. The gels in A, C, D, and F show samples from multiple animals per condition. Results are mean ± SEM. *P < 0.05 versus controls.

Targeted deletion of Ctsk in hematopoietic cells and osteoclasts increases cancellous bone mass and expression of osteoblast-specific marker genes. We analyzed the skeletal phenotype of 9-week-old Ctsk^fl/fl and Mx1;Ctsk^fl/fl mice. No difference in body weight and bone length was observed between Ctsk^fl/fl and Mx1;Ctsk^fl/fl mice in the C57BL/6 background, indicating that targeted Ctsk deletion did not affect skeletal development (data not shown). X-ray and micro-CT (μCT) analysis of Cre^– and Cre⁺ mice revealed that targeted deletion of Ctsk in hematopoietic cells increased bone mass in femurs (Figure 3A). μCT showed an increase in cancellous bone volume, trabecular number, and connectivity density, with a concomitant decrease in trabecular separation (Figure 3B). Histomorphometric analysis of distal femurs confirmed a significant increase in cancellous bone volume in both males and females (Figure 4 and Supplemental Table 1; supplemental material available online with this article; doi: 10.1172/JCI64840DS1). Analysis of trabecular architecture in Mx1;Ctsk^fl/fl mice indicated an increase in trabecular number in males and females, and thickness in females, with a concomitant decrease in trabecular separation in both sexes. While the mineralizing surface was not affected, the mineral apposition rate (MAR) was higher in Mx1;Ctsk^fl/fl mice, leading to an increase in the BFR. Moreover, osteoblast and osteoclast numbers per tissue area were both markedly increased. As shown in Figure 5A, qRT-PCR indicated that ablation of Ctsk in hematopoietic cells increased the expression of several osteoblast marker genes in femurs, including Runx2, Osterix, ALP, and osteopontin. These data suggest that the high bone mass observed in Mx1;Ctsk^fl/fl mice is likely the consequence of increased bone formation and decreased bone resorption in vivo.

Figure 3

Deleting Ctsk in hematopoietic cells and osteoclasts increases cancellous bone mass. (A) μCT and x-ray images of femurs from controls (Ctrl) and conditional knockouts (cKO) at 9 weeks of age. (B) μCT analysis of distal femur. Results are mean ± SEM. *P < 0.05 versus controls.

Figure 4

Deleting Ctsk in hematopoietic cells and osteoclasts increases cancellous bone mass, bone formation, and numbers of osteoblasts and osteoclasts. Histomorphometric analysis of cancellous bone in femurs from 9-week-old mice. Results are mean ± SEM. *P < 0.05 versus controls. Original magnification, ×2.

Figure 5

Osteoblast marker genes are upregulated in distal femur metaphysis of Mx1;Ctsk^fl/fl and CD11b;Ctsk^fl/fl mice. (A) qRT-PCR analysis of mRNA expression in distal femur metaphysis from 9-week-old Ctsk^fl/fl and Mx1;Ctsk^fl/fl mice. (B) qRT-PCR analysis of mRNA expression in distal femur metaphysis from 9-week-old CD11b;Ctsk^+/+ and CD11b;Ctsk^fl/fl mice. Data are expressed as percentage of mRNA level in controls. Results are mean ± SEM. *P < 0.05 versus controls.

Like Mx1;Ctsk^fl/fl mice, CD11b;Ctsk^fl/fl mice exhibited a significant (P < 0.05) increase in bone mass in the femur at 9 weeks of age, as determined by an increase in cancellous bone volume (Figure 3A) and trabecular thickness using μCT (data not shown). However, CD11b;Ctsk^fl/fl mice had a milder skeletal phenotype than Mx1;Ctsk^fl/fl mice. Histomorphometric analysis indicated that CD11b-Cre–induced cathepsin K deficiency increased osteoclast numbers, but also resulted in increases in several indices of bone formation (MAR, BFR, and osteoblast numbers), demonstrating that the increased BFR is osteoclast mediated (Figure 4 and Supplemental Table 2). Femoral cancellous bone volume was increased, but to a lesser extent (29%) compared with that in Mx1;Ctsk^fl/fl mice (78%). Transcriptional profiling of osteoblast target genes in distal femur metaphysis also showed increased expression of osteoblast marker genes, in particular, Osterix and osteocalcin (Figure 5B).

Deletion of Ctsk in the osteoclast lineage also affects cortical cross-sectional volume and cortical bone volume. In contrast to Mx1;Ctsk^fl/fl mice, which showed only a similar trend (Supplemental Table 3 and Supplemental Figure 1), deletion of Ctsk in osteoclasts using the CD11b promoter led to an increase in femoral cross-sectional bone volume and cortical bone volume (Supplemental Table 4 and Supplemental Figure 1), implying an increase in periosteal bone formation after deletion of Ctsk in osteoclasts.

Lack of cathepsin K in hematopoietic cells and osteoclasts impairs bone resorption. To determine whether the increased number of osteoclasts in Mx1;Ctsk^fl/fl and CD11b;Ctsk^fl/fl mice was cell autonomous or secondary to changes occurring in other cells, we analyzed osteoclast differentiation in vitro. Mx1-Cre–induced deletion of Ctsk in hematopoietic cells and osteoclasts increased osteoclast differentiation from BM macrophages (BMMs) (Figure 6, A and B). We plated osteoclasts on dentin slices and compared their resorbing ability to examine whether defective osteoclast function contributed to the increased bone volume in Mx1;Ctsk^fl/fl and CD11b;Ctsk^fl/fl mice. We found that resorption pit depth was reduced, such that the pits formed by Mx1;Ctsk^fl/fl osteoclasts were shallow (Figure 6C). Immunofluorescence labeling of phalloidin confirmed that Mx1;Ctsk^fl/fl osteoclasts formed a flat, actin-rich sealing zone (Supplemental Figure 2). It has been shown that deletion of Ctsk and treatment with cathepsin K inhibitors interrupts intracellular vesicle trafficking, leading to an accumulation of type I collagen–containing vesicles and reduced bone resorption efficiency (17, 26). Immunofluorescence staining of synaptotagmin 7, a marker of “secretory” lysosomes (27), showed an accumulation of labeled vesicles in Ctsk-deleted osteoclasts (Figure 7A). Similarly, vesicles labeled for the lysosome-associated membrane protein 2 (LAMP2), a late endosomal-lysosomal marker, also accumulated in Ctsk-deleted osteoclasts (Figure 7B). This is consistent with the reports indicating that deletion or inhibition of cathepsin K alters normal intracellular vesicle trafficking (17, 26), possibly further altering osteoclast function. Thus, the absence of cathepsin K in osteoclasts increased their numbers, but decreased their bone-resorbing activity, leading to an increase in bone mass.

Figure 6

Targeted deletion of Ctsk in hematopoietic cells and osteoclasts increases osteoclast differentiation and decreases bone resorption. (A) Control (Ctrl) and conditional knockout (cKO) osteoclasts were generated on glass coverslips in the presence of M-CSF and RANKL. Original magnification, ×10. (B) TRAP-positive osteoclast numbers per area (/mm²) were quantified by OsteoMeasure software (OsteoMetrics). Results are mean ± SEM. *P < 0.05 versus controls. (C) Osteoclasts generated by coculture with osteoblasts on collagen gel in the presence of vitamin D₃ and PGE₂ were replated onto dentin slices and cultured for 48 hours. Images of resorption pits on dentin slices resorbed by control and conditional knockout osteoclasts were analyzed by Fuji software.

Figure 7

Deletion of Ctsk in hematopoietic cells and osteoclasts increases accumulation of intracellular lysosomes. Osteoclasts generated by coculture with osteoblasts on collagen gel in the presence of vitamin D₃ and PGE₂ were replated onto dentin slices and cultured for 48 hours. (A) Fluorescence micrographs of osteoclasts showing synaptotagmin 7 (green) and actin (red). (B) The localization of LAMP2 (green) and actin (red) was analyzed in control and conditional knockout osteoclasts. Scale bar: 10 μm.

Targeted deletion of Ctsk in the osteoblast lineage has no effect on bone formation. Although the level of expression of Ctsk is much lower in osteoblasts than in cells of the osteoclast lineage, this enzyme could still play an intrinsic role in osteoblasts, possibly affecting bone formation. To examine whether the increased BFR in global or targeted Ctsk-deficient mice could be due to the absence of cathepsin K in osteoblasts, we generated mice with Ctsk conditionally deleted in cells of the osteoblast lineage by crossing Ctsk^fl/+ mice with Osx-Cre mice, thereby targeting expression of Cre to the osteoblast lineage at early stages of differentiation. Since Osx-Cre mice are mildly osteopenic (28), we compared Osx;Ctsk^+/+ (controls) with Osx;Ctsk^fl/fl mice (conditional knockouts). We found no changes in bone density at 5, 9, or 13 weeks of age by x-ray; data from 13-week-old mice are presented. μCT analysis showed no difference in cancellous and cortical bone volume at 13 weeks of age (Figure 8, A and B, Supplemental Table 5, and Supplemental Figure 1). Furthermore, histomorphometric analysis demonstrated no alterations of the skeletal phenotype in osteoblast-targeted Ctsk-deficient mice (Figure 8C and Supplemental Table 6). In contrast to Mx1-Cre– or CD11b-Cre–driven deletion of Ctsk, deletion of Ctsk in the osteoblast lineage had no effect on bone volume, BFR, osteoblast numbers, or osteoclast numbers. In vitro assays of primary osteoblasts derived from the long bones showed that ALP and mineralized bone nodules were similar in cells derived from Osx;Ctsk^fl/fl mice and their control littermates (Figure 8D). These data demonstrate that the bone volume and BFR changes observed in Mx1;Ctsk^fl/fl or CD11b;Ctsk^fl/fl mice were not due to the deletion of Ctsk in osteoblasts. Therefore, the observed increases in BFR in Mx1-Cre and CD11b-Cre–deleted cathepsin K mice are secondary to the deletion of Ctsk in cells of the osteoclast lineage.

Figure 8

Deletion of Ctsk in cells of osteoblast lineage has no effect on bone formation and bone resorption. (A) μCT and x-ray images of femurs from Osx;Ctsk^+/+ and Osx;Ctsk^fl/fl mice at 13 weeks of age. (B) μCT analysis of distal femur metaphysis. (C) Histomorphometric analysis of cancellous bone in femurs from 13-week-old Osx;Ctsk^fl/fl mice and controls. Original magnification, ×2. (D) ALP and mineralized nodule in osteoblasts derived from long bones of Osx;Ctsk^+/+ and Osx;Ctsk^fl/fl mice. Results are mean ± SEM.

Lack of Ctsk in hematopoietic cells increases osteoblast precursors. Since Cre-induced deletion of Ctsk using the CD11b promoter was less efficient than the deletion obtained with the Mx1-Cre promoter, generating a qualitatively similar but milder skeletal phenotype, we used Mx1;Ctsk^fl/fl mice to further characterize the mechanism by which Ctsk deficiency in osteoclasts led to an increase in BFR. To determine whether the increased bone formation observed in Mx1;Ctsk^fl/fl mice was due to an increased commitment of osteoblast progenitors, we examined their proliferation and differentiation in long bone–derived primary osteoblasts in vitro. Osteoblasts derived from long bones of Mx1;Ctsk^fl/fl mice had more ALP activity and formed more mineralized bone nodules than did control cells (Figure 9A). In addition, ablation of Ctsk increased CFU-ALP and CFU osteoblasts (CFU-OBs), but not CFU fibroblasts (CFU-Fs) (Supplemental Figure 3). Consistent with the CFU-F assay, BrdU labeling showed no change in osteoblast proliferation between Mx1;Ctsk^fl/fl osteoblasts and controls (Supplemental Figure 3). These observations suggest that the lack of cathepsin K in hematopoietic cells increased the number of committed osteoblast progenitors in bone, contributing to increased bone formation in vivo and in vitro. Consistent with these in vitro findings, FACS analysis confirmed a 17% increase in ALP-positive cells derived from Mx1;Ctsk^fl/fl long bone osteoblasts compared with controls (Figure 9B). To analyze the differentiation of Mx1;Ctsk^fl/fl osteoblasts, we isolated total RNA from osteoblasts derived from long bone of Mx1;Ctsk^fl/fl mice and their control littermates, and the expression of osteoblast marker genes was analyzed by qRT-PCR (Figure 9C). Expression of several markers of osteoblast differentiation (ALP, type I collagen, Runx2, and Dmp1) was significantly upregulated in Mx1;Ctsk^fl/fl cells. Osteopontin mRNA expression (P = 0.064) was slightly increased in Mx1;Ctsk^fl/fl osteoblasts, but did not reach significance. In addition, RANKL mRNA was markedly increased, whereas OPG mRNA levels decreased, leading to a significant increase in the ratio of RANKL and OPG mRNA levels. Consistent with the increase in RANKL/OPG ratio, we observed greater numbers of TRAP-positive osteoclasts formed in coculture of Mx1;Ctsk^fl/fl osteoblasts with Ctsk^fl/fl BMMs (Figure 9D). Taken together, these in vitro results show that ablation of Ctsk in the osteoclast lineage: (a) increases the number and differentiation of cells of the osteoblast lineage, and (b) increases the RANKL/OPG ratio in these cells, leading to an increase in the number of osteoclasts, as shown in vivo and in vitro.

Figure 9

Genetic deletion of Ctsk in hematopoietic cells increases committed osteoprogenitors. (A) ALP and mineralized nodule in osteoblasts derived from long bones of Mx1;Ctsk^fl/fl mice and their control littermates. (B) FACS analysis of the percentage of ALP-positive cells from long bone–derived osteoblasts of Mx1;Ctsk^fl/fl mice (yellow; n = 4) and Ctsk^fl/fl (blue; n = 4) mice. Cells were stained with either PE-labeled anti-ALP or PE-labeled isotype control (red). (C) qRT-PCR analysis of mRNA expression in culture osteoblasts derived from long bones of Ctsk^fl/fl mice (n = 3) and Mx1;Ctsk^fl/fl mice (n = 3). Data are expressed as percentage of mRNA level of controls. (D) TRAP-positive osteoclast number per area (/mm²) from coculture of either Ctsk^fl/fl or Mx1;Ctsk^fl/fl osteoblasts with Ctsk^fl/fl BMMs. Results are mean ± SEM. *P < 0.05 versus controls.

Deletion of Ctsk increases Sphk1 expression in osteoclasts, and S1P in osteoclast-conditioned medium increases ALP in osteoblasts. Since Osx;Ctsk^fl/fl mice do not have a skeletal phenotype, we hypothesized that reducing Ctsk expression in osteoclasts resulted in more anabolic coupling from osteoclasts to osteoblasts, and thus increased remodeling-based bone formation. To test this hypothesis, we first determined whether conditioned medium from Mx1;Ctsk^fl/fl and CD11b;Ctsk^fl/fl osteoclasts increased ALP in osteoblasts. CD1 calvarial osteoblasts were cultured in conditioned medium from either knockout or control osteoclasts. As shown in Figure 10, calvarial osteoblasts treated with conditioned medium derived from Mx1;Ctsk^fl/fl and CD11b;Ctsk^fl/fl osteoclast cultures showed increased ALP staining, indicating that the conditioned medium contained more osteoanabolic factors.

Figure 10

Conditional knockout osteoclast-conditioned medium increases ALP staining. CD1 calvarial osteoblasts were treated with either control or conditional knockout osteoclast-conditioned medium for 7 days and stained with ALP. Original magnification, ×2.

To identify coupling factors that might be increased in the Ctsk-null osteoclasts, we measured in vitro gene expression of the known osteoclast-derived coupling factors Efnb2, Wnt10b, Bmp6, and SPHKs (which generates S1P) (20, 29–31) in osteoclasts generated from the Mx1;Ctsk^fl/fl and CD11b;Ctsk^fl/fl mice. Ablation of Ctsk in hematopoietic cells and osteoclasts increased Sphk1 mRNA expression, but not Sphk2, Efnb2, Wnt10b, or Bmp6 expression (Figure 11A). Immunostaining confirmed the increased SPHK1 in Mx1;Ctsk^fl/fl and CD11b;Ctsk^fl/fl osteoclasts plated on dentin slices (Figure 11B). We also determined SPHK1 protein levels by Western blotting (Figure 11C). Consistent with the mRNA analysis, Ctsk-deficient osteoclasts had higher SPHK1 protein levels. Sphk1 catalyzes the phosphorylation of sphingosine to S1P, which has been reported to target osteoblasts and increase bone formation (20). We therefore examined the levels of S1P in Mx1;Ctsk^fl/fl osteoclast–conditioned medium using mass spectrometry. The conditioned medium from Mx1;Ctsk^fl/fl osteoclasts contained significantly more S1P than Ctsk^fl/fl osteoclast–conditioned medium (Figure 12A). Interestingly, immunofluorescence staining of Sphk1 was not increased in Src^–/– osteoclasts (Supplemental Figure 4), indicating that it was not the result of osteoclast functional impairment, but rather was a specific effect of Ctsk deletion.

Figure 11

Deletion of Ctsk in hematopoietic cells and osteoclasts increases Sphk1 mRNA expression and protein level in conditional knockout osteoclasts. (A) qRT-PCR analysis of mRNA expression in control and conditional knockout osteoclasts. Data are expressed as percentage of mRNA level of controls. Results are mean ± SEM. *P < 0.05 versus controls. (B) Osteoclasts generated by coculture with osteoblasts on collagen gel in the presence of vitamin D₃ and PGE₂ were replated onto dentin slices and cultured for 48 hours. The localization of TO-PRO-3 (blue), Sphk1 (green), and actin (red) were analyzed by confocal microscopy. Original magnification, ×63. (C) Lysates from osteoclasts were subjected to immunoblotting with anti-Sphk1 antibodies.

Figure 12

S1P is increased in the absence of cathepsin K, and blocking S1P activity attenuates ALP activity and mineralization induced by Ctsk-deficient and control osteoclast–conditioned medium. (A) Mx1;Ctsk^fl/fl osteoclast–conditioned medium contained elevated levels of S1P. (B) CD1 calvarial osteoblasts were treated with either Mx1;Ctsk^fl/fl or Ctsk^fl/fl osteoclast–conditioned medium for 7 days and stained with ALP, or for 20 days and stained with alizarin red in the presence or absence of VPC23019 (VPC). Original magnification, ×1. (C) ALP activity (n = 5 per group) and mineralization (n = 6 per group) were quantified. Results are mean ± SEM. *P < 0.05 versus controls.

Antagonizing S1P reduces Mx1;Ctsk^fl/fl osteoclast–conditioned medium-induced ALP activity and mineralization. Mouse osteoblasts express three S1P receptors: S1P₁, S1P₂, and S1P₃ (20). To further verify whether biologically active S1P is responsible for the coupling-mediated increase in bone formation, VPC23019, an S1P_1,3 antagonist, was added to the cultures of CD1 calvarial osteoblasts in the presence of conditioned medium from either Mx1;Ctsk^fl/fl or Ctsk^fl/fl osteoclast cultures. VPC23019 at 2 μM did not have any measurable effect on ALP activity (data not shown), but 10 μM VPC23019 markedly inhibited the increase in ALP activity and mineralization that are induced by Mx1;Ctsk^fl/fl osteoclast–conditioned medium (Figure 12, B and C). These data confirm that S1P is a major contributor to the Ctsk-deficient osteoclast-mediated enhancement of bone formation in vitro.

Our results establish several important findings regarding bone remodeling. First, our genetic studies show that bone formation is regulated, in part, by cells of the osteoclast lineage, experimentally providing evidence for a coupling mechanism in an in vivo setting. Second, in vitro studies have established S1P as a key osteoclast-derived coupling messenger, a “clastokine” synthesized and secreted by osteoclasts that affects osteoblast differentiation and acts locally to promote the osteoblasts’ bone-forming activity. Third, our studies indicate that the presence and/or activity of cathepsin K within cells of the osteoclast lineage represses the expression of Sphk1, and thereby the synthesis of S1P, decreasing the ability of osteoclasts to stimulate bone formation. As a result, deleting or decreasing cathepsin K expression in osteoclasts favors S1P synthesis and bone formation. Accordingly, blocking S1P_1,3 in osteoblast cultures decreases osteoclast-conditioned medium-induced ALP activity and mineralization. Fourth, these studies suggest that cells of the osteoblast lineage attempt to compensate for the functional defect of individual osteoclasts by increasing their numbers through activation of the RANKL pathway, possibly also in response to increased S1P levels (20). Taken together, these results support the hypothesis that therapeutic manipulation of osteoclast activity via inhibition of cathepsin K can decrease bone resorption, while maintaining or increasing bone formation at remodeling sites. Finally, our finding of an increase in total cross-sectional volume and cortical bone volume, without any change in marrow volume after deletion of Ctsk in cells of the osteoclast lineage, suggest that it may also indirectly affect periosteal modeling–based bone formation, an observation also made after pharmacological inhibition of cathepsin K in nonhuman primates (32, 33).

The adult skeleton undergoes constant remodeling throughout life. The remodeling process begins with resorption by multinucleated osteoclasts, followed by osteoblast-mediated bone formation. To maintain skeletal homeostasis, bone resorption is coupled to bone formation, and imbalances in this process lead to a variety of human diseases including osteoporosis and osteopetrosis. Osteoporosis is the result of an excess of bone resorption over bone formation during the remodeling process. The most common approach for the treatment of this debilitating disease has been to inhibit bone resorption through antiresorptive drugs. As a consequence of the coupling process, inhibition of bone resorption generally leads to a parallel decrease in bone formation. This, in turn, leads to low bone turnover and may contribute to unwanted consequences such as ONJ or atypical femoral fractures. In contrast, PTH, which is the only approved anabolic drug for the treatment of osteoporosis, increases bone formation, but secondarily leads to increased bone resorption. The secondary increase in resorption induced by PTH is due to increased RANKL production by osteoblasts, the direct targets of PTH (34). These observations, made both in animal studies and human clinical trials, have led to the conclusion that bone mass could be gained only during limited periods of time (the so-called “anabolic windows”) until bone formation decreases after antiresorptive treatment, or until bone resorption increases after PTH treatment.

Global loss of cathepsin K leads to an osteopetrotic phenotype due to defective bone resorption that is characterized in mice by an increase in the number of poorly functional osteoclasts and increased bone formation (18, 19). The present study was aimed at understanding the molecular mechanisms by which such a potentially beneficial dissociation of bone resorption and bone formation was induced by the deletion of Ctsk. We initially demonstrated that the phenotype of the germline deletion of Ctsk was reproduced by deleting Ctsk in osteoclasts using Mx1-Cre. However, Mx1-Cre targets multiple hematopoietic cell types, including megakaryocytes, lymphocytes, mast cells, and cells of the monocyte-macrophage lineage, and we could not rule out possible cathepsin K–dependent contributions to regulation of bone homeostasis by other types of cathepsin K–expressing hematopoietic cells in this mouse model. To confirm that the altered bone homeostasis was a consequence of deleting Ctsk in osteoclasts, we used a more specific conditional knockout of CD11b-Cre for targeting Cre recombinase exclusively to the monocyte-macrophage lineage. The skeletal phenotype of CD11b;Ctsk^fl/fl mice was identical, albeit milder, and was probably due to the lower efficiency of Ctsk deletion in CD11b;Ctsk^fl/fl osteoclasts than that in Mx1;Ctsk^fl/fl mice. The smaller increase in bone volume in CD11b;Ctsk^fl/fl mice than that seen in Mx1;Ctsk^fl/fl mice (29% vs. 78%), notwithstanding the similar increases in MAR and BFR, is likely due to the greater residual bone-resorbing activity of the CD11b;Ctsk^fl/fl osteoclasts. Thus, targeted deletion of Ctsk in cells of the osteoclast lineage resulted in a skeletal phenotype comparable to the global Ctsk knockouts.

It has been reported that cathepsin K is expressed in mesenchymal stem cells, osteoblasts, osteocytes, and bone marrow stromal cells (BMSCs) in humans, albeit at much lower levels than in osteoclasts (35–37). The role of cathepsin K in human osteoblasts is unclear. Our data demonstrate that it does not appear to affect bone remodeling and homeostasis, since mice with osteoblast-targeted disruption of cathepsin K using Osx-Cre did not exhibit any skeletal phenotype, indicating that decreasing the already low levels of Ctsk expression in cells of the osteoblast lineage does not contribute to the increased bone formation observed in the globally Ctsk-deleted mice. Thus, it appears that inhibition of cathepsin K leads to increased cancellous bone formation through indirect osteoclast-mediated effects and not through mesenchymal or osteoblast lineage cell-autonomous effects in mice. The fact that osteoblasts isolated from the osteoclast-targeted conditional knockout mice form more bone nodules in vitro and the number of osteoblasts in bone increases, suggests that in vivo changes occurring in cells of the osteoblast lineage cells after deletion of Ctsk in osteoclasts result from an increase in the number of committed osteoblast precursors. This conclusion is supported by our observation of increased numbers of ALP-positive cells by FACS analysis, as well as increased numbers of CFU-ALP and CFU-OBs without any changes in total CFU-Fs.

Within bone remodeling sites, the transition from bone resorption to bone formation occurs as a result of coupling activity mediated by several factors. In recent years, several coupling factors have been identified that can be classified in 2 categories: (a) matrix-derived growth factors (MDGFs) that are released from bone matrix during the bone resorption process, and (b) osteoclast-derived cytokines (clastokines) that are soluble factors secreted by osteoclasts into the local microenvironment, or membrane-bound proteins that mediate bidirectional signaling between osteoblasts and osteoclasts (11, 20, 29–31, 38). Ctsk deletion appears to favor coupling through alteration in both categories of coupling factors (Figure 13, A and B). Fuller and colleagues have shown in vitro that the decreased degradation of collagen and other bone matrix proteins when cathepsin K is inhibited increases the levels of intact IGF-1 released during bone resorption by osteoclasts, potentially enhancing local bone formation (11). Similarly, inhibition of cathepsin K may locally increase the levels of activated TGF-β, which is released from bone matrix during osteoclastic bone resorption and has recently been shown to induce migration of osteoblast precursors to bone remodeling sites, thus favoring bone formation (38). Although all these studies were performed in vitro and no information is available in vivo, our studies cannot exclude the possibility that MDGFs contribute to the phenotype we observed here in vivo in osteoclast-targeted Ctsk deletions. Notwithstanding this possibility, the finding that conditioned medium derived from Mx1;Ctsk^fl/fl or CD11;Ctsk^fl/fl osteoclasts cultured in the absence of bone matrix increased ALP in osteoblasts demonstrates that osteoclast-secreted anabolic signals that are independent of resorptive activity are increased after deletion of Ctsk.

Figure 13

Schematic representation of bone remodeling and the coupling process in controls and Ctsk-deficient mice. (A) Under normal circumstance, osteoclasts secrete clastokines, including osteoanabolic factors such as S1P, for the regulation of bone formation. Additionally, osteoclasts resorb bone, thus releasing MDGFs to enhance bone formation. (B) Absence of cathepsin K increases S1P released by osteoclasts and may increase intact MDGFs (11). Furthermore, S1P induces an increase in the production of RANKL by osteoblasts (20), leading to increased osteoclast differentiation. An overall increase in the level of S1P stimulates bone formation. OC-pre, osteoclast precursor.

We have identified S1P, a known clastokine (20, 29–31), as a factor derived from osteoclasts and as being responsible for the observed phenotype. S1P is produced by sphingosine kinase in osteoclasts and interacts with the S1P receptors expressed on osteoblasts, enhancing migration and survival of osteoblasts as well as RANKL expression (20). Others have shown by microarray analysis that increased expression of Sphk1 by osteoclasts correlates with enhanced osteoblast differentiation in vitro (29). We found that Sphk1 expression was upregulated in Mx1;Ctsk^fl/fl and CD11b-Ctsk^fl/fl osteoclasts, and the S1P content of Mx1;Ctsk^fl/fl osteoclast–conditioned medium was increased. In addition, an S1P_1,3 antagonist blunted the anabolic effects of osteoclast-conditioned medium in vitro. The upregulated Sphk1 in individual Ctsk-deficient osteoclasts and the consequent increase in S1P production, together with the increased osteoblast differentiation induced by the Mx1;Ctsk^fl/fl osteoclast–conditioned medium and the increased bone formation in mice in vivo, confirm that osteoclasts can affect BFR via secreted molecules acting on cells of the osteoblast lineage. Thus, S1P appears to be an important coupling factor, and our data indicate that cathepsin K acts to decrease its production.

Others have shown that extracellular S1P stimulates osteoclastogenesis in coculture of BMMs and osteoblasts by increasing RANKL production, but has no effect on osteoclast differentiation in BMM cultures treated with RANKL (20). Consistent with this mechanism, we observed a significant increase in RANKL expression in osteoblasts derived from long bones in Mx1;Ctsk^fl/fl mice. Since cells of the osteoblast lineage, including osteocytes, are the major source of RANKL in mice (39, 40), this implies that changes occurring in cathepsin K–deficient osteoclasts led to this osteoblast-mediated response, forming a positive feedback loop to promote bone resorption by increasing osteoclastogenesis. It is therefore likely that osteoclast-derived S1P is responsible not only for the increased BFR observed in vivo and ALP activity observed in vitro, but also for the stimulation of RANKL production by cells of the osteoblast lineage that leads to the increased osteoclastogenesis observed in both osteoclast-targeted and global Ctsk deletions.

Our observation that in vitro osteoclastogenesis is increased in BMMs derived from conditional knockout mice in response to M-CSF and RANKL raises an interesting additional question, and suggests that the increased osteoblast-derived RANKL/OPG ratio may not be the only explanation for the greater number of osteoclasts observed in vivo. As the cells in this assay are exposed to equal levels of M-CSF and RANKL, the increased number of osteoclasts generated in vitro reflects either an increase in sensitivity to RANKL, or an increased number of committed osteoclast precursors in the BMMs isolated from the conditional knockout mice, or both. Further studies will be required to address this question.

The mechanism by which the deletion of Ctsk regulates the intracellular expression of Sphk1 and production of S1P remains to be determined. Secreted cathepsin K, or a factor released by cathepsin K as it degrades the bone matrix, could act in a paracrine fashion on osteoclasts to regulate Sphk1 expression. Alternatively, evidence is accumulating that cathepsin K acts intracellularly in addition to its extracellular function of degrading bone matrix proteins. This protease acts within lysosomes to activate TRAP (10), and it is conceivable that a lack of activated TRAP also affects osteoclast biology after Ctsk deletion. Osteoclasts transport protons and chloride ions into the resorbing compartment to dissolve the bone mineral phase and secrete lysosomal acid hydrolases such as cathepsin K to digest bone matrix proteins. Cathepsin K is sorted intracellularly into secretory lysosomes, which translocate to and fuse with the ruffled border membrane, releasing their content into the resorption compartment (26). The molecules controlling the trafficking and fusion of lysosomes with the ruffled border membrane therefore play a critical role in bone resorption. Synaptotagmin 7, a Ca²⁺-sensitive synaptotagmin isoform, regulates exocytosis of lysosomes in osteoclasts, where it colocalizes with LAMP2 and cathepsin K (27). We observed intracellular accumulation of lysosomal vesicles that are positive for synaptotagmin 7 and LAMP2 after Ctsk deletion, similar to what has been reported after pharmacological inhibition of cathepsin K (26). Although it is not possible to determine whether this is the result of impaired vesicular fusion with the plasma membrane, or of increased generation of lysosomes in an attempt to compensate for the defective activity resulting from Ctsk deletion, these changes seem to be specific to Ctsk deletion. In addition, we did not observe an accumulation of secretory lysosomes in Src-deleted osteoclasts (Supplemental Figure 4), which favors the hypothesis of a specific intracellular role of cathepsin K in affecting both vesicular traffic and S1P synthesis.

From a practical point of view, this study suggests that it should be possible to inhibit bone resorption without impairing bone formation in patients with osteoporosis by inhibiting cathepsin K activity. Indeed, it has recently been reported that inhibition of cathepsin K activity with orally administered small molecules not only inhibits the production of markers of bone resorption, but also maintains or enhances bone formation in some animal models (32, 33, 41, 42), and maintains bone formation markers close to control values in human studies (43, 44). Our mouse genetic studies would suggest that the effect of inhibiting cathepsin K on bone formation is indirect and is due to the effect of the compounds on osteoclasts rather than on cells of the osteoblast lineage. It is, however, more difficult to understand how such a mechanism of action could explain the reported increase in periosteal bone formation in nonhuman primates (32), since osteoclasts are rarely seen at the periosteum. Our finding that total cortical cross-sectional volume and cortical bone volume were increased after deletion of Ctsk in cells of the osteoclast lineage also implies an increase in periosteal bone formation. Thus, periosteal changes could be secondary to osteoclast-mediated events, which are possibly transduced by the cortical osteocytic network. Alternatively, deletion of Ctsk in monocytes-macrophages residing in the periosteum could be responsible for this increased activity. Further work will be required to determine whether this also occurs in humans and to explain this phenomenon.

In conclusion, this study demonstrates that events occurring within osteoclasts have a strong influence on the number of osteoblast progenitors and bone formation in vivo. It also demonstrates in an in vivo setting that the bone resorbing activity of osteoclasts can be dissociated from their ability to stimulate bone formation, such that the coupling process is dependent not only on putative MDGFs, but also on osteoclast-derived molecules — “clastokines”— among which S1P appears to be an important player.

View Supplemental data

We thank Kei Yamana for providing PCR primer sequences. We also thank Nicholas Brady, Chantel Lester, and Antonios Aliprantis for μCT analysis. Confocal analysis was performed at the Harvard NeuroDiscovery Center (Boston, Massachusetts, USA). This work was supported in part by Procter & Gamble.

Address correspondence to: Roland Baron, Department of Oral Medicine, Infection and Immunity, Harvard School of Dental Medicine, 188 Longwood Avenue, REB 310, Boston, Massachusetts 02115, USA. Phone: 617.432.7320; Fax: 617.432.1897; E-mail: roland_baron@hsdm.harvard.edu.

Conflict of interest: Roland Baron is a consultant for Merck and receives research support from this company.

Reference information: J Clin Invest. 2013;123(2):666–681. doi:10.1172/JCI64840.

1. Nyman JS, Yeh OC, Hazelwood SJ, Martin RB. A theoretical analysis of long-term bisphosphonate effects on trabecular bone volume and microdamage. Bone. 2004;35(1):296–305.
   View this article via: PubMed CrossRef Google Scholar
2. Bone HG, Schurr W. Intravenous bisphosphonate therapy for osteoporosis: where do we stand? Curr Osteoporos Rep. 2004;2(1):24–30.
   View this article via: PubMed CrossRef Google Scholar
3. Khan AA, et al. Bisphosphonate associated osteonecrosis of the jaw. J Rheumatol. 2009;36(3):478–490.
   View this article via: PubMed CrossRef Google Scholar
4. Koh JS, Goh SK, Png MA, Kwek EB, Howe TS. Femoral cortical stress lesions in long-term bisphosphonate therapy: a herald of impending fracture? J Orthop Trauma. 2010;24(2):75–81.
   View this article via: PubMed CrossRef Google Scholar
5. Carvalho NN, Voss LA, Almeida MO, Salgado CL, Bandeira F. Atypical femoral fractures during prolonged use of bisphosphonates: short-term responses to strontium ranelate and teriparatide. J Clin Endocrinol Metab. 2011;96(9):2675–2680.
   View this article via: PubMed CrossRef Google Scholar
6. Black DM, et al. The effects of parathyroid hormone and alendronate alone or in combination in postmenopausal osteoporosis. N Engl J Med. 2003;349(13):1207–1215.
   View this article via: PubMed CrossRef Google Scholar
7. Dougall WC, et al. RANK is essential for osteoclast and lymph node development. Genes Dev. 1999;13(18):2412–2424.
   View this article via: PubMed CrossRef Google Scholar
8. Douni E, et al. A RANKL G278R mutation causing osteopetrosis identifies a functional amino acid essential for trimer assembly in RANKL and TNF. Hum Mol Genet. 2012;21(4):784–798.
   View this article via: PubMed CrossRef Google Scholar
9. Schinke T, et al. Impaired gastric acidification negatively affects calcium homeostasis and bone mass. Nat Med. 2009;15(6):674–681.
   View this article via: PubMed CrossRef Google Scholar
10. Zenger S, et al. Proteolytic processing and polarized secretion of tartrate-resistant acid phosphatase is altered in a subpopulation of metaphyseal osteoclasts in cathepsin K-deficient mice. Bone. 2007;41(5):820–832.
    View this article via: PubMed CrossRef Google Scholar
11. Fuller K, et al. Cathepsin K inhibitors prevent matrix-derived growth factor degradation by human osteoclasts. Bone. 2008;42(1):200–211.
    View this article via: PubMed CrossRef Google Scholar
12. Gelb BD, Shi GP, Chapman HA, Desnick RJ. Pycnodysostosis, a lysosomal disease caused by cathepsin K deficiency. Science. 1996;273(5279):1236–1238.
    View this article via: PubMed CrossRef Google Scholar
13. Gelb BD, et al. Pycnodysostosis: refined linkage and radiation hybrid analyses reduce the critical region to 2 cM at 1q21 and map two candidate genes. Hum Genet. 1996;98(2):141–144.
    View this article via: PubMed CrossRef Google Scholar
14. Johnson MR, Polymeropoulos MH, Vos HL, Ortiz de Luna RI, Francomano CA. A nonsense mutation in the cathepsin K gene observed in a family with pycnodysostosis. Genome Res. 1996;6(11):1050–1055.
    View this article via: PubMed CrossRef Google Scholar
15. Saftig P, et al. Impaired osteoclastic bone resorption leads to osteopetrosis in cathepsin-K-deficient mice. Proc Natl Acad Sci U S A. 1998;95(23):13453–13458.
    View this article via: PubMed CrossRef Google Scholar
16. Gowen M, et al. Cathepsin K knockout mice develop osteopetrosis due to a deficit in matrix degradation but not demineralization. J Bone Miner Res. 1999;14(10):1654–1663.
    View this article via: PubMed CrossRef Google Scholar
17. Kiviranta R, et al. Impaired bone resorption in cathepsin K-deficient mice is partially compensated for by enhanced osteoclastogenesis and increased expression of other proteases via an increased RANKL/OPG ratio. Bone. 2005;36(1):159–172.
    View this article via: PubMed CrossRef Google Scholar
18. Pennypacker B, et al. Bone density, strength, and formation in adult cathepsin K (–/–) mice. Bone. 2009;44(2):199–207.
    View this article via: PubMed CrossRef Google Scholar
19. Li CY, et al. Mice lacking cathepsin K maintain bone remodeling but develop bone fragility despite high bone mass. J Bone Miner Res. 2006;21(6):865–875.
    View this article via: PubMed CrossRef Google Scholar
20. Ryu J, Kim HJ, Chang EJ, Huang H, Banno Y, Kim HH. Sphingosine 1-phosphate as a regulator of osteoclast differentiation and osteoclast-osteoblast coupling. EMBO J. 2006;25(24):5840–5851.
    View this article via: PubMed CrossRef Google Scholar
21. Aliprantis AO, et al. NFATc1 in mice represses osteoprotegerin during osteoclastogenesis and dissociates systemic osteopenia from inflammation in cherubism. J Clin Invest. 2008;118(11):3775–3789.
    View this article via: JCI PubMed CrossRef Google Scholar
22. Beller DI, Springer TA, Schreiber RD. Anti-Mac-1 selectively inhibits the mouse and human type three complement receptor. J Exp Med. 1982;156(4):1000–1009.
    View this article via: PubMed CrossRef Google Scholar
23. Shalhoub V, et al. Characterization of osteoclast precursors in human blood. Br J Haematol. 2000;111(2):501–512.
    View this article via: PubMed Google Scholar
24. Ferron M, Vacher J. Targeted expression of Cre recombinase in macrophages and osteoclasts in transgenic mice. Genesis. 2005;41(3):138–145.
    View this article via: PubMed CrossRef Google Scholar
25. Ambrogini E, et al. FoxO-mediated defense against oxidative stress in osteoblasts is indispensable for skeletal homeostasis in mice. Cell Metab. 2010;11(2):136–146.
    View this article via: PubMed CrossRef Google Scholar
26. Leung P, Pickarski M, Zhuo Y, Masarachia PJ, Duong LT. The effects of the cathepsin K inhibitor odanacatib on osteoclastic bone resorption and vesicular trafficking. Bone. 2011;49(4):623–635.
    View this article via: PubMed CrossRef Google Scholar
27. Zhao H, Ito Y, Chappel J, Andrews NW, Teitelbaum SL, Ross FP. Synaptotagmin VII regulates bone remodeling by modulating osteoclast and osteoblast secretion. Dev Cell. 2008;14(6):914–925.
    View this article via: PubMed CrossRef Google Scholar
28. Davey RA, et al. Decreased body weight in young Osterix-Cre transgenic mice results in delayed cortical bone expansion and accrual. Transgenic Res. 2012;21(4):885–893.
    View this article via: PubMed CrossRef Google Scholar
29. Pederson L, Ruan M, Westendorf JJ, Khosla S, Oursler MJ. Regulation of bone formation by osteoclasts involves Wnt/BMP signaling and the chemokine sphingosine-1-phosphate. Proc Natl Acad Sci U S A. 2008;105(52):20764–20769.
    View this article via: PubMed CrossRef Google Scholar
30. Zhao C, et al. Bidirectional ephrinB2-EphB4 signaling controls bone homeostasis. Cell Metab. 2006;4(2):111–121.
    View this article via: PubMed CrossRef Google Scholar
31. Irie N, et al. Bidirectional signaling through ephrinA2-EphA2 enhances osteoclastogenesis and suppresses osteoblastogenesis. J Biol Chem. 2009;284(21):14637–14644.
    View this article via: PubMed CrossRef Google Scholar
32. Cusick T, et al. Odanacatib treatment increases hip bone mass and cortical thickness by preserving endocortical bone formation and stimulating periosteal bone formation in the ovariectomized adult rhesus monkey. J Bone Miner Res. 2012;27(3):524–537.
    View this article via: PubMed CrossRef Google Scholar
33. Jerome C, Missbach M, Gamse R. Balicatib, a cathepsin K inhibitor, stimulates periosteal bone formation in monkeys. Osteoporos Int. 2012;23(1):339–349.
    View this article via: PubMed CrossRef Google Scholar
34. Silva BC, Costa AG, Cusano NE, Kousteni S, Bilezikian JP. Catabolic and anabolic actions of parathyroid hormone on the skeleton. J Endocrinol Invest. 2011;34(10):801–810.
    View this article via: PubMed Google Scholar
35. Ilmer M, Karow M, Geissler C, Jochum M, Neth P. Human osteoblast-derived factors induce early osteogenic markers in human mesenchymal stem cells. Tissue Eng Part A. 2009;15(9):2397–2409.
    View this article via: PubMed CrossRef Google Scholar
36. Podgorski I, et al. Bone marrow-derived cathepsin K cleaves SPARC in bone metastasis. Am J Pathol. 2009;175(3):1255–1269.
    View this article via: PubMed CrossRef Google Scholar
37. Mandelin J, et al. Human osteoblasts produce cathepsin K. Bone. 2006;38(6):769–777.
    View this article via: PubMed CrossRef Google Scholar
38. Tang Y, et al. TGF-β1-induced migration of bone mesenchymal stem cells couples bone resorption with formation. Nat Med. 2009;15(7):757–765.
    View this article via: PubMed CrossRef Google Scholar
39. Xiong J, Onal M, Jilka RL, Weinstein RS, Manolagas SC, O’Brien CA. Matrix-embedded cells control osteoclast formation. Nat Med. 2011;17(10):1235–1241.
    View this article via: PubMed CrossRef Google Scholar
40. Nakashima T, et al. Evidence for osteocyte regulation of bone homeostasis through RANKL expression. Nat Med. 2011;17(10):1231–1234.
    View this article via: PubMed CrossRef Google Scholar
41. Masarachia PJ, et al. Odanacatib reduces bone turnover and increases bone mass in the lumbar spine of skeletally mature ovariectomized rhesus monkeys. J Bone Miner Res. 2012;27(3):509–523.
    View this article via: PubMed CrossRef Google Scholar
42. Pennypacker BL, et al. Cathepsin K inhibitors prevent bone loss in estrogen-deficient rabbits. J Bone Miner Res. 2011;26(2):252–262.
    View this article via: PubMed CrossRef Google Scholar
43. Vasiljeva O, Reinheckel T, Peters C, Turk D, Turk V, Turk B. Emerging roles of cysteine cathepsins in disease and their potential as drug targets. Curr Pharm Des. 2007;13(4):387–403.
    View this article via: PubMed CrossRef Google Scholar
44. Costa AG, Cusano NE, Silva BC, Cremers S, Bilezikian JP. Cathepsin K: its skeletal actions and role as a therapeutic target in osteoporosis. Nat Rev Rheumatol. 2011;7(8):447–456.
    View this article via: PubMed CrossRef Google Scholar
45. Rantakokko J, Kiviranta R, Eerola R, Aro HT, Vuorio E. Complete genomic structure of the mouse cathepsin K gene (Ctsk) and its localization next to the Arnt gene on mouse chromosome 3. Matrix Biol. 1999;18(2):155–161.
    View this article via: PubMed CrossRef Google Scholar
46. Rantakokko J, Aro HT, Savontaus M, Vuorio E. Mouse cathepsin K: cDNA cloning and predominant expression of the gene in osteoclasts, and in some hypertrophying chondrocytes during mouse development. FEBS Lett. 1996;393(2–3):307–313.
    View this article via: PubMed CrossRef Google Scholar
47. Ruocco MG, et al. I{kappa}B kinase (IKK){beta}, but not IKK{alpha}, is a critical mediator of osteoclast survival and is required for inflammation-induced bone loss. J Exp Med. 2005;201(10):1677–1687.
    View this article via: PubMed CrossRef Google Scholar
48. Parfitt AM, et al. Bone histomorphometry: standardization of nomenclature, symbols, and units. Report of the ASBMR Histomorphometry Nomenclature Committee. J Bone Miner Res. 1987;2:595–610.
    View this article via: PubMed CrossRef Google Scholar
49. Bouxsein ML, Boyd SK, Christiansen BA, Guldberg RE, Jepsen KJ, Muller R. Guidelines for assessment of bone microstructure in rodents using micro-computed tomography. J Bone Miner Res. 2010;25(7):1468–1486.
    View this article via: PubMed CrossRef Google Scholar
50. Bakker AD, Klein-Nulend J. Osteoblast isolation from murine calvaria and long bones. Methods Mol Biol. 2012;816:19–29.
    View this article via: PubMed Google Scholar
51. Lotinun S, Evans GL, Turner RT, Oursler MJ. Deletion of membrane-bound steel factor results in osteopenia in mice. J Bone Miner Res. 2005;20(4):644–652.
    View this article via: PubMed CrossRef Google Scholar
52. Purev E, Neff L, Horne WC, Baron R. c-Cbl and Cbl-b act redundantly to protect osteoclasts from apoptosis and to displace HDAC6 from beta-tubulin, stabilizing microtubules and podosomes. Mol Biol Cell. 2009;20(18):4021–4030.
    View this article via: PubMed CrossRef Google Scholar
53. Ruwisch L, Schafer-Korting M, Kleuser B. An improved high-performance liquid chromatographic method for the determination of sphingosine-1-phosphate in complex biological materials. Naunyn Schmiedebergs Arch Pharmacol. 2001;363(3):358–363.
    View this article via: PubMed CrossRef Google Scholar
54. Yatomi Y, Ruan F, Ohta J, Welch RJ, Hakomori S, Igarashi Y. Quantitative measurement of sphingosine 1-phosphate in biological samples by acylation with radioactive acetic anhydride. Anal Biochem. 1995;230(2):315–320.
    View this article via: PubMed CrossRef Google Scholar