Cell Metabolism

Article
Targeted Ablation of Osteocytes Induces Osteoporosis with Defective Mechanotransduction
Sawako Tatsumi,1,2 Kiyoaki Ishii,1 Norio Amizuka,3 Minqi Li,3 Toshihiro Kobayashi,1 Kenji Kohno,4 Masako Ito,5 Sunao Takeshita,1 and Kyoji Ikeda1,*
1

Department of Bone and Joint Disease, Research Institute, National Center for Geriatrics and Gerontology, Obu, Aichi 474-8522, Japan 2 National Institute of Biomedical Innovation, Osaka 567-0085, Japan 3 Center for Transdisciplinary Research, Niigata University, Niigata 951-8514, Japan 4 Laboratory of Molecular and Cell Genetics, Division of Cell Biology, Nara Institute of Science and Technology, Nara 630-0192, Japan 5 Department of Radiology, Nagasaki University School of Medicine, Nagasaki 852-8501, Japan *Correspondence: kikeda@nils.go.jp DOI 10.1016/j.cmet.2007.05.001

SUMMARY

Bone remodeling is performed by osteoclasts and osteoblasts at the bone surface. Inside of bone is a network of numerous osteocytes, whose specic function has remained an enigma. Here we describe a transgenic mouse model in which inducible and specic ablation of osteocytes is achieved in vivo through targeted expression of diphtheria toxin (DT) receptor. Following a single injection of DT, approximately 70%80% of the osteocytes, but apparently no osteoblasts, were killed. Osteocyte-ablated mice exhibited fragile bone with intracortical porosity and microfractures, osteoblastic dysfunction, and trabecular bone loss with microstructural deterioration and adipose tissue proliferation in the marrow space, all of which are hallmarks of the aging skeleton. Strikingly, these osteocyte-less mice were resistant to unloading-induced bone loss, providing evidence for the role of osteocytes in mechanotransduction. Thus, osteocytes represent an attractive target for the development of diagnostics and therapeutics for bone diseases, such as osteoporosis.
INTRODUCTION The integrity of the skeletal system is maintained by a continuous remodeling process, i.e., bone resorption by osteoclasts followed by bone formation by osteoblasts (Karsenty and Wagner, 2002; Partt, 1994). The recruitment and activity of osteoclasts and osteoblasts are under hormonal (Manolagas and Jilka, 1995; Pacici, 1996), neuronal (Ducy et al., 2000), immunological (Takayanagi et al., 2000), and mechanical control (Burger and Klein-Nulend, 1999; Ehrlich and Lanyon, 2002). With aging, especially after menopause in females, the balancing of activities
464 Cell Metabolism 5, 464475, June 2007 2007 Elsevier Inc.

of these two cell types fails, leading to a decline in both the quantity and quality of bone (Raisz, 2005; Seeman and Delmas, 2006), culminating in fragility fractures (Raisz, 2005; Riggs and Melton, 1986; Seeman, 2002). Osteoporotic fractures are a serious public health issue in the aging population and impose a major economic burden worldwide (Sambrook and Cooper, 2006). Osteoclasts differentiate from hematopoietic progenitor cells in bone marrow, especially of the monocyte/macrophage lineage, under the stimulation of two essential cytokines, macrophage colony stimulating factor (M-CSF) and receptor activator of NF-kB ligand (RANKL), and through the coordinated actions of nuclear transcription factors, such as c-Fos, nuclear factor of activated T cells c1 (NFATc1), and NF-kB (Teitelbaum and Ross, 2003). The generation and death of osteoclasts are also kept in check by local and systemic negative regulators, most notably osteoprotegerin (OPG), a decoy receptor of RANKL, and estrogen, respectively (Boyle et al., 2003). It is hypothesized from the perspective of bone remodeling that osteoclasts recognize and are targeted to skeletal sites of compromised mechanical integrity and initiate the bone remodeling process for the purpose of generating replacement bone with mechanical competence, although the exact cues and underlying mechanisms that target osteoclasts to specic sites of prospective basic multicellular units remain elusive (Partt, 2002). Osteoclastic bone resorption is followed by the recruitment of osteoblasts, which derive from mesenchymal stem cells (Harada and Rodan, 2003). Osteoblasts actively synthesize extracellular matrix, represented by type I collagen, and mineralize it (Karsenty and Wagner, 2002). Bone formation and resorption are not independent processes but are in fact mutually and intimately linked in that osteoblastic/stromal cells provide an osteoclastogenic microenvironment through presentation of RANKL and in that the bone formation by osteoblasts depends on the preceding resorption by osteoclasts through the action of a putative coupling factor (Martin and Sims, 2005). Osteocytes, the third type of bone cell, are terminally differentiated from osteoblasts and exist embedded in

Cell Metabolism
Osteocytes in Mechanotransduction

mineralized matrix, residing individually in caves called lacunae (Aarden et al., 1994). Osteocytes are stellateshaped or dendritic cells that communicate with each other and with osteoblasts and lining cells on the bone surface via their prominent cell processes within tunnels called canaliculi, thereby forming a neuron-like network throughout the skeleton (Burger and Klein-Nulend, 1999). Although osteocytes represent the most abundant cell type in bone, their peculiar location and inaccessibility in mineralized matrix have hampered the cellular and molecular analysis of osteocytes, and there have been few osteocyte-specic markers (Franz-Odendaal et al., 2006). A lack of established methods for isolation and cultivation of osteocytes has imposed another obstacle, and there have been very few cell lines available to study osteocyte biology (Kato et al., 1997). Furthermore, even if available, isolated osteocytes on a dish may not behave as osteocytes do in the complex three-dimensional structure and environment in vivo. Our current knowledge of this particular cell type is based mainly on observations in vitro, and the in vivo function of osteocytes remains to be determined. In the present study, we generated a mouse model of osteocyte ablation that allowed elucidation of the physiological function of osteocytes in bone remodeling and mechanotransduction in an intact animal. RESULTS Generation of a Transgenic Mouse Model for Inducible Ablation of Osteocytes In order to delineate the function of osteocytes in vivo, we attempted to establish a transgenic mouse model based on the diphtheria toxin receptor-mediated cell knockout (TRECK) system (Saito et al., 2001). In order to direct the expression of diphtheria toxin receptor (DT-R) specically in osteocytes, we used the promoter of dentin matrix protein 1 (DMP1), which was originally identied in dentin (George et al., 1993). In situ hybridization and immunohistochemical analysis with a specic antibody demonstrated that DMP1 is highly expressed in osteocytes as well as in dentin, but not in osteoblasts (Toyosawa et al., 2001). A transgenic cassette (pDTR-9.6kb) carrying a genomic fragment comprised of a 9.6 kb 50 -anking region, exon 1, intron 1, and a part of exon 2 of the mouse DMP1 gene fused to human HB-EGF (DT-R) cDNA with a polyadenylation signal was constructed (Figure 1A). Transgenic DTR-9.6kb lines were produced using a standard pronuclear injection protocol and were identied by genomic Southern hybridization and PCR (data not shown) using the primers shown in Figure 1A. Four independent transgenic founders were mated to wild-type C57BL/6 mice, and the F1 offspring were analyzed. All four lines produced pups at the expected ratio. These pups appeared normal, grew indistinguishably from wild-type mice, and were fertile. Two lines (#2 and #4) showed a very similar DT sensitivity and phenotype (data not shown), and the results presented here are from DTR-9.6kb line #2. RT-PCR with RNAs extracted from various tissues revealed that DT-R mRNA was ex-

pressed only in bone and dentin of transgenic mice (Figure 1B). Immunohistochemical analysis with a specic antibody against DT-R (human HB-EGF) showed that DT-R was expressed specically in osteocytes of trabecular (Figure 1C) as well as cortical bone (Figure 1D), but not in osteoblasts on the bone surface (Figure 1E). DT-R mRNA expression in bone cells was further examined by RT-PCR. DT-R mRNA was detected only in RNA extracted from the bone of DTR-9.6kb transgenic mice, and not in bone of wild-type mice (see right panel of Figure S1A in the Supplemental Data available with this article online). On the other hand, when osteoblasts were isolated from the calvaria of transgenic mice, DT-R mRNA was not detected (Figure S1A, left panel), and primary osteoblasts isolated from either transgenic or wild-type mice did not respond to DT with cell death (Figure S1B). Forced expression of DT-R in these osteoblasts by retroviral infection rendered them sensitive to DT; prompt and massive apoptosis was observed (Figure S1B), and all cells died within 24 hr. Collectively, the data indicate that the 9.6 kb DMP1 promoter with intron 1 conferred preferential expression of DT-R in osteocytes, but not in osteoblasts. Specic Ablation of Osteocytes in Transgenic Mice Osteocytes normally reside individually in their lacunae (Figures 2A and 2B), and administration of DT (up to 50 mg/kg body weight) to wild-type mice caused no change in osteocyte morphology (Figure 2A, upper right panel). DTR-9.6kb transgenic mice exhibited no overt differences in skeletal phenotype compared with wild-type mice. Bone histology of the transgenic mice injected with PBS as vehicle was indistinguishable from that of wild-type mice, and the number and morphology of osteocytes were normal (Figure 2A, lower left panel). There was no signicant difference in cortical width (151 4 mm in wild-type versus 156 3 mm in transgenic mice). In contrast, a number of empty lacunae that did not contain osteocytes were observed in the cortical bone of 10-week-old male DTR-9.6kb transgenic mice at 8 days following a single i.p. administration of DT (Figures 2A and 2E). Empty lacunae appeared within 48 hr after DT administration (data not shown). The counting of lacunae in bone sections at 8 days indicated that the number of empty lacunae increased markedly (up to 50%60%) in femoral cortical bone, specically in those sections prepared from transgenic mice injected with DT (Figure 2F, red bars). Even in the lacunae that were not empty, electron microscopic examination revealed that $40% of the osteocytes (Figure 2F, yellow bars) had undergone apoptotic or necrotic changes (Figures 2C and 2D). Thus, it was estimated that more than 70% of the osteocytes in the cortical bone were ablated in the current transgenic model (Figure 2F, red plus yellow bars). Successful ablation of osteocytes was also conrmed in the trabecular bone of DTR-9.6kb transgenic mice injected with DT (Figures S2A and S2B) and was also observed in vertebrae and in both male and female mice at 824 weeks of age (data not shown).
Cell Metabolism 5, 464475, June 2007 2007 Elsevier Inc. 465

Cell Metabolism
Osteocytes in Mechanotransduction

Figure 1. Generation of Transgenic Mice Expressing DT-R Specically in Osteocytes

(A) Schematic representation of the transgenic construct (pDTR-9.6kb). Diphtheria toxin receptor (DT-R, dotted box) with poly(A) signals (dark box) was placed under the control of the 9.6 kb mouse DMP1 promoter (thick line) plus exons 1 and 2 (white bars) and intron 1 between them. The arrows indicate the set of primers used for conrming integration of the transgene (one in exon 2 of the DMP1 gene, and the other in the DT-R cDNA). (B) RT-PCR analysis of transgene (DT-R) expression in DTR-9.6kb transgenic line #2. RNA for bone and dentin was prepared from tibiae/femurs and incisors, respectively. (CE) Immunohistochemical detection of the transgene product, DT-R, by anti-DT-R (hHB-EGF) antibody in osteocytes of trabecular (C) and cortical bone (D) of the femur of an 8-week-old transgenic mouse. (GP, growth plate; TB, trabecular bone; CB, cortical bone.) Representative osteocytes with prominent staining are indicated by black arrows. A high magnication of a trabecula reveals that osteoblasts on the surface do not express DT-R (red arrows in [E]).

In order to determine whether blockade of protein synthesis by DT induces apoptosis of osteocytes, we examined bone tissues by the TUNEL method. Transgenic mice injected with DT exhibited a markedly increased number of TUNEL-positive osteocytes, but not in osteoblasts, at 36 hr (Figure 2G). Osteoblasts on the trabecular bone surface appeared morphologically normal following DT administration (Figure S2A, black arrows). Lack of DT
466 Cell Metabolism 5, 464475, June 2007 2007 Elsevier Inc.

sensitivity was further conrmed in primary osteoblasts from the calvaria of transgenic mice. As shown in Figures S1C and S1D, proliferation and differentiation of primary osteoblasts from transgenic mice as assessed by BrdU incorporation and alkaline phosphatase (ALP) activity, respectively, were indistinguishable from those of wildtype osteoblasts and were not inuenced by DT treatment ex vivo. Thus, it was concluded that 70%80% of

Cell Metabolism
Osteocytes in Mechanotransduction

Figure 2. Targeted Ablation of Osteocytes in Transgenic Mice

(A) Ten-week-old male wild-type (WT, upper panels) and transgenic (Tg, lower panels) mice were injected i.p. with PBS (left panels) or 50 mg/kg body weight DT (right panels), and femoral cortical bone was stained with hematoxylin and eosin at 8 days post DT treatment. Note that empty lacunae are observed only in cortical bone of transgenic mice injected with DT (arrows). Scale bars = 50 mm. (BE) Osteocyte morphology by electron microscopy showing a normal osteocyte (B), osteocytes with nuclear condensation (C) or fragmentation (D), and an empty lacuna (E) from wild-type mice (B) or transgenic mice injected with DT (CE) at 8 days. (F) Numbers of empty lacunae (red), lacunae with apoptotic osteocytes (yellow), and lacunae with normal osteocyte morphology (blue) were counted in femoral cortical bone from WT and Tg mice at 8 days following injection of PBS or DT. Mean values are shown as a percentage of all lacunae (n = 36). (G) TUNEL staining (arrows) of femoral sections of WT and Tg mice 36 hr after DT administration.

osteocytes had been successfully ablated in cortical as well as trabecular bone (Figure 2F and Figure S2B, red plus yellow bars) following a single DT administration. Short-Term Effects of Osteocyte Ablation The roles that osteocytes play in bone metabolism were then investigated using the current transgenic model. At 8 days following a single i.p. administration of DT to DTR-9.6kb mice, regions of enlarged empty lacunae with vascular invasion (Figure 3A and inset) and invasion of tartrate-resistant acid phosphatase (TRAP)-positive osteoclasts (Figure 3B) were observed in the cortical bone. ALP staining revealed the presence of ALP-positive osteoblasts on the wall of the intracortical cavities excavated by osteoclasts (Figure 3C), indicating that osteoblasts had been recruited into the cavities along with the invading osteoclasts. von Kossa staining counterstained with toluidine blue revealed areas that stained positively with toluidine blue on the wall of enlarged lacunae (blue staining indicated by white arrows in Figure 3D) and also on the endosteal

surface of the osteocyte-ablated cortical bone (blue staining indicated by red arrows in Figure 3D) of DTinjected transgenic mice, suggesting the development of unmineralized matrix. The osteoid surface in the trabecular bone was also increased (Figure 3F). Electron microscopy revealed the presence of microfractures in the cortical bone following osteocyte ablation (Figure 3E, yellow arrows). These results suggest that osteocytes play a critical role in keeping bone resorption from undergoing aberrant acceleration and that, in the absence of osteocytes, intra- as well as endocortical mineralization is impaired. Osteocyte ablation had no short-term impact on trabecular bone mass; 3D bone volume at the tibial metaphysis, as determined by microcomputed tomography (micro-CT) scanning, did not differ between osteocyte-ablated mice and controls at 12 weeks after DT administration (Figure 3G and data not shown). In the trabecular bone of osteocyte-ablated mice, however, increased TRAP staining was observed compared with the DT-injected wild-type mice used as a control (Figure 3H).
Cell Metabolism 5, 464475, June 2007 2007 Elsevier Inc. 467

Cell Metabolism
Osteocytes in Mechanotransduction

Figure 3. Short-Term Effects of Osteocyte Ablation

(A) Histological analysis of hematoxylin-andeosin-stained sections of femurs from 10week-old WT (left) and Tg (right) mice at 8 days after a single injection of DT. Note the empty lacunae (arrows) and intracortical cavities with vascular invasion. (BV in inset, blood vessel.) (BD) TRAP (B), ALP (C), and von Kossa (D) staining of femurs of WT (left) and Tg (right) mice injected with DT. Arrows and the asterisk in (B) indicate empty lacunae and intracortical cavities, respectively. The blue staining shown by the red arrows and white arrows in (D) indicates osteoid on the endosteal surface and on the wall of intracortical cavities, respectively. (E) The presence of microfractures (yellow arrows) following osteocyte ablation as revealed by electron microscopy. Red arrows indicate red blood cells in the cracks. The asterisk indicates an empty lacuna. Magnication 10003. (F) Osteoid surface fraction per bone surface (OS/BS) measured at the trabecular bone of the proximal tibia. In this and all other gures, error bars indicate SEM. *p < 0.05 (n = 5). (G) Trabecular bone volume/tissue volume (BV/TV) in the proximal tibia as determined by micro-CT was unchanged at 8 days following osteocyte ablation. (H) TRAP staining of the trabecular bone of WT (left) and Tg (right) mice at 8 days after DT administration. (I) Histomorphometric indices of bone formation in the femur at 7 days after injection of DT or PBS in 10-week-old WT or Tg mice. Osteoblast surface (Ob.S) and bone formation rate (BFR) were corrected for bone surface (BS). MAR, mineral apposition rate; Mlt, mineralization lag time. **p < 0.01 (n = 3 per group).

Histomorphometric analysis of the trabecular compartment of femurs at 8 days of osteocyte ablation revealed no change in the number of osteoblasts (Figure 3I), suggesting that differentiation/recruitment of osteoblasts was not affected shortly after DT administration. The mineralizing function of mature osteoblasts on trabecular bone as assessed by the mineral apposition rate (MAR) and mineralization lag time (Mlt) was signicantly reduced following osteocyte ablation (Figure 3I). Taken together with the presence of unmineralized bone matrix on the endosteal surface and within the cortical bone of osteocyte-ablated mice (Figure 3D), this suggests that osteocytes play an important role in maintaining the mineralizing function of mature osteoblasts. Bone-marrow cells were isolated from femurs and tibiae of wild-type as well as transgenic mice treated in vivo with DT or PBS, and ex vivo colony-formation assays were performed (Figure S3). FACS analysis revealed that the
468 Cell Metabolism 5, 464475, June 2007 2007 Elsevier Inc.

number of total bone-marrow cells and major cell populations did not differ among the four experimental groups, except for a modest decrease in B220-positive B lineage cells in the DT-injected transgenic mice (data not shown). The results of the colony-formation assays revealed that the number of cfu-F, cfu-ALP, and cfu-OB did not differ between the wild-type and transgenic-derived cultures or between cultures of PBS-injected and DT-injected transgenic mice (Figure S3), suggesting that osteoblast differentiation was not affected by osteocyte ablation. Osteocyte Ablation Increases the RANKL/OPG Ratio Quantitative RT-PCR using RNA samples extracted from bone of transgenic mice treated with DT revealed that the expression of DT-R (human HB-EGF) mRNA as well as DMP1 mRNA decreased markedly at 2 days following DT injection in DTR-9.6kb transgenic mice (Figure 4), which documents successful osteocyte ablation at the

Cell Metabolism
Osteocytes in Mechanotransduction

Figure 4. Quantitative RT-PCR Analysis of Molecular Markers of Bone Cells

After the bone marrow had been ushed out, RNA was extracted from the femurs and tibiae of 10-week-old DTR-9.6kb transgenic mice injected with PBS (white bars) or DT (black bars) at 2 days. Gene expression was assessed by real-time PCR using a LightCycler system and normalized to EF-1a mRNA. MEPE, matrix extracellular phosphoglycoprotein; Phex, phosphate-regulating gene with homologies to endopeptidases on the X chromosome; RANKL, receptor activator of NF-kB ligand; OPG, osteoprotegerin; ALP, alkaline phosphatase; OC, osteocalcin. **p < 0.01; *p < 0.05 (n = 3 mice per group).

molecular level. Reduction in DT-R mRNA reached nearly 100%, pointing to nearly complete ablation of osteocytes that expressed DT-R. The DMP1 mRNA level was reduced by approximately 70%, consistent with the histological observations that 70%80% of osteocytes were ablated, provided that all osteocytes expressed DMP1 mRNA. mRNA expression of FGF-23, sclerostin, MEPE, Phex, and E11/gp38, known products of osteocytes (FranzOdendaal et al., 2006), also decreased signicantly following osteocyte ablation (Figure 4). Among the molecules involved in bone resorption, mRNA for RANKL increased; mRNA for OPG was unchanged at 2 days (Figure 4) but decreased modestly at 4 days (data not shown). Interestingly, ex vivo cultures of bone-marrow cells from osteocyte-ablated mice under osteogenic conditions gave rise to TRAP-positive osteoclast-like cells in association with ALP-positive osteoblastic colonies even without the addition of any stimulator of osteoclastogenesis, such as RANKL, whereas the formation of osteoclasts on osteoblastic colonies occurred much less frequently in cultures from control mice (Figure S4). With regard to the molecular markers of osteoblast differentiation, osteocalcin mRNA level decreased, whereas levels of Runx2 (Karsenty and Wagner, 2002) and Osterix (Nakashima et al., 2002) as well as ALP mRNA remained unaltered (Figure 4 and data not shown). Thus, following the ablation of osteocytes, the impaired transcription function of Runx2 without a change in gene expression per se is thought to lead to the reduced expression of the osteocalcin gene. Longer-Term Consequences of Osteocyte Deciency Histological examination of tibiae at 40 days following a single injection of DT revealed that empty lacunae still

existed and that cortical bone became thin as compared with bone of DT-injected wild-type mice (Figure 5A). The cortical area of the femur was decreased (data not shown), and intracortical porosity was markedly increased in the osteocyte-ablated mice (Figures 5B and 5C). In fact, bone strength, as determined by the four-point bending test, was signicantly compromised 40 days following osteocyte ablation (Figure 5H). These results suggest that the presence of osteocytes is critical for the maintenance of structural integrity and mechanical competence of cortical bone. Histological examination of the tibial metaphysis 40 days after a single injection of DT revealed that the trabecular bone had become thin and sparse, with an expansion of adipose tissue in the marrow space, as compared with bone of DT-injected wild-type mice (Figures 5A and 5D). Not only appendicular bone but also axial bone displayed substantial structural changes 40 days after osteocyte ablation. Micro-CT scanning of the lumbar vertebrae revealed a substantial decrease in the 3D trabecular bone volume per tissue volume (BV/TV) (Figures 5E and 5F). This decrease was associated with microstructural changes in the trabeculae, represented by decreased trabecular connectivity, reduced trabecular thickness, and an increased structure model index (SMI), implying transformation of the trabeculae from plate-like structures to more fragile rod-like ones (Figure 5F). These results indicate that the long-term consequences of osteocyte ablation result in a loss of trabecular bone mass with ensuing microarchitectural deterioration. In order to follow up on the recovery process after a transient ablation of osteocytes, a cohort of transgenic as well as wild-type mice were treated with DT, and tibial trabecular bone was analyzed by micro-CT at 40 days and then 90 days post DT administration, when normal osteocytes
Cell Metabolism 5, 464475, June 2007 2007 Elsevier Inc. 469

Cell Metabolism
Osteocytes in Mechanotransduction

that the quantity and quality of bone are normalized as new osteocytes differentiate from osteoblasts and become embedded in the skeleton. Osteocytes in Mechanotransduction Transgenic mice with inducible osteocyte ablation provide an ideal model to determine the role of osteocytes in mechanotransduction in an intact animal. DTR-9.6kb transgenic mice as well as wild-type controls were subjected to unloading by tail suspension for 7 days (Bikle and Halloran, 1999; Globus et al., 1984); DT was administered 1 day prior to the initiation of tail suspension to deplete osteocytes (Figure 6A). Mice on the ground under normal loading conditions during the same period served as controls. Note that during a short period of 1 week following DT administration, trabecular bone volume did not differ between the wild-type and transgenic mice on the ground (Figures 6A and 6B), as previously observed in separate experiments (Figure 3G). As expected, wild-type mice injected with DT as well as transgenic mice injected with PBS lost a substantial portion of trabecular bone during the 1 week tail suspension (Figures 6A and 6B). Strikingly, transgenic mice injected with DT and therefore rendered osteocyte decient were resistant to unloading-induced bone loss (Figures 6A and 6B), which provides in vivo evidence that the presence of osteocytes is crucial for mechanotransduction in bone. The microarchitectural deterioration induced by unloading, as demonstrated by micro-CT scanning, was also completely inhibited in the osteocyte-ablated mice (Figure S5). Histomorphometric analysis revealed that the increase in osteoclast number observed in control mice during mechanical unloading as well as suppressed bone formation was abrogated in osteocyte-ablated mice (Figure 6C). When RNA extracted from tibiae and femurs at 1, 2, and 3 days after the initiation of tail suspension was analyzed by quantitative RT-PCR, RANKL mRNA increased within 12 days compared to control and increased progressively thereafter at 3 days, with a slightly delayed increase in OPG mRNA at 3 days in the wild-type mice (data not shown). In osteocyte-ablated mice, the increases in RANKL and OPG mRNA during tail suspension were blocked (Figure 6D), suggesting that the increase in RANKL expression in osteoblasts in response to mechanical unloading does not take place in the absence of osteocytes, which may underlie the resistance to disuseinduced bone loss. sclerostin mRNA increased during tail suspension, a result that was not observed in osteocyteablated mice (Figure 6D). In order to examine the role of osteocytes in reloading, following the completion of tail suspension for 7 days, we injected either DT or PBS into DTR-9.6kb transgenic mice and left them ambulatory for the subsequent 14 days (Figure 7A, groups 4 and 3, respectively). The limiting of osteocyte ablation to the reloading period did not affect the robust recovery of bone mass, and osteocytedecient mice gained an amount of bone similar to PBS-injected control mice (Figures 7A and 7B). Bone histomorphometry revealed that the increased osteoclast

Figure 5. Osteoporotic Changes as Long-Term Consequences of Osteocyte Ablation

(A and B) Representative tibial sections of 18-week-old WT and Tg mice at 40 days following a single DT administration. Note the thinning of the cortical bone and accumulation of adipose tissue in the marrow space following osteocyte ablation. Scale bars = 200 mm. Part of the cortical bone in (A) is highlighted in (B) to illustrate intracortical porosity (arrows). (C and D) Intracortical porosity area per cortical area (Po.Ar/Ct.Ar) (C) and marrow adiposity (% fat volume/marrow volume) (D) of femurs at 40 days after DT administration. **p < 0.01 (n = 56 per group). (E and F) Representative micro-CT images of lumbar vertebrae of WT and Tg mice at 40 days following DT administration (E) with microstructural parameters (BV/TV, 3D bone volume fraction per tissue volume; Conn-Dens., connectivity density; SMI, structure model index; Tb.Th, trabecular thickness) derived from micro-CT analysis (F). **p < 0.01 (n = 5 per group). (G and H) Recovery of bone volume (G) and bone strength (H) along with osteocyte regeneration. Trabecular bone volume and bone strength were determined by micro-CT and the four-point bending test, respectively, on the femurs of transgenic mice injected with PBS (white bars) or DT (black bars) at 40 and 90 days. *p < 0.05; **p < 0.01 (n = 5 per group).

resided in most lacunae (data not shown). Trabecular bone volume was restored to a level indistinguishable from the control group at 90 days following DT administration (Figure 5G). Accordingly, bone strength was also restored to the control level at 90 days (Figure 5H). Thus, it is suggested that the bone loss and compromised strength following osteocyte ablation are reversible and
470 Cell Metabolism 5, 464475, June 2007 2007 Elsevier Inc.

Cell Metabolism
Osteocytes in Mechanotransduction

Figure 6. Osteocyte-Ablated Mice Are Resistant to Unloading-Induced Bone Loss

(A) Experimental schedule with representative micro-CT images. Eighteen-week-old WT and Tg mice were subjected to skeletal unloading by tail suspension (TS) for 7 days. Mice on the ground (Gr) served as controls. DT was administered 1 day prior to the initiation of TS (1). (B) Changes in 3D bone volume at the tibial metaphysis following 7 day unloading by tail suspension (black bars). Mice on the ground (white bars) served as controls. **p < 0.01 (n = 5 per group). (C) Number of osteoclasts (N.Oc/BS) and bone formation rate (BFR/BS) in TS (black bars) and ground control (white bars) mice. *p < 0.05 (n = 5 per group). (D) Quantitative RT-PCR analysis. RNA was extracted from tibiae and femurs of WT and DTR-9.6kb Tg mice injected with DT (red bars) or PBS (pale green bars) at 3 days following the initiation of TS, after bone marrow had been ushed out. Transgenic mice on the ground (Gr) served as controls (white and black bars). *p < 0.05; **p < 0.01 (n = 3 per group).

number following tail suspension as well as suppressed bone formation was normalized during the reloading period in osteocyte-ablated mice to the same extent as in the control mice (Figure 7C). These results indicate that osteocytes are indispensable for unloading-induced bone loss, while recovery by reloading can bypass osteocytes (Figure 7D). DISCUSSION In this study, we have established a mouse model for the inducible ablation of osteocytes, which has allowed delineation of crucial functions in skeletal homeostasis in vivo. We have demonstrated that osteocytes play an important role in the regulation of osteoblastic and osteoclastic activities on the bone surface and have provided proof for the essential function of osteocytes in mechanotrans-

duction in vivo. Despite the fact that 70%80%, not 100%, of osteocytes were ablated, transgenic mice injected with DT exhibited marked alterations in bone metabolism, which may be explained by osteocytes functioning as a network, such that 70%80% ablation is sufcient to cause a serious malfunction of the network as a whole. We have achieved specic ablation of osteocytes in a transgenic model by targeting DT-R to osteocytes under the control of the DMP1 promoter. DMP1 was originally identied in dentin and encodes an extracellular-matrixassociated phosphoprotein (George et al., 1993). While the reported phenotypes of DMP1 knockout mice in bone mineralization (Feng et al., 2006; Ling et al., 2005) and chondrogenesis (Ye et al., 2005) suggest that DMP1 serves physiological functions in osteoblasts and chondrocytes, the transgenic approach that we and others
Cell Metabolism 5, 464475, June 2007 2007 Elsevier Inc. 471

Cell Metabolism
Osteocytes in Mechanotransduction

Figure 7. Absence of Osteocytes Does Not Affect Bone Gain during Reloading
(A) Experimental schedule for reloading after tail suspension (TS/Gr) and representative micro-CT images. Eighteen-week-old WT and DTR-9.6kb Tg mice were subjected to TS for 7 days and were then left ambulatory for the following 14 days. DT (4) or PBS as vehicle (3) was injected before initiation of reloading so that osteocytes were ablated specically during the reloading period (4). Mice on the ground throughout the experimental period (1) and mice subjected to TS only (2) served as controls. (B and C) Changes in 3D bone volume at the tibial metaphysis of WT and Tg mice with or without DT administration (B) and histomorphometric indices for osteoclast number (N.Oc) and bone formation rate (BFR) corrected for bone surface (BS) (C) after reloading (TS/Gr). **p < 0.01; *p < 0.05 (n = 5 per group). (1)(4) correspond to the four experimental groups shown in (A). (D) Proposed roles of osteocytes under different mechanical conditions. Based on the current results, under normal loading conditions (Mechanical input +), osteocytes function to keep osteoclastic (Oc) bone resorption in check and to maintain mineralization by osteoblasts (Ob). Thus, when osteocytes are ablated, aberrantly elevated bone resorption with impaired mineralization takes place. In response to unloading (Mechanical input ), osteocytes execute the stimulation of bone resorption and suppression of bone formation, resulting in marked bone loss and microstructural deterioration in a short period. Thus, when osteocytes are ablated specically during tail suspension, those changes do not take place, and bone is resistant to disuse-induced atrophy. However, based on the results of the reloading experiments following unloading (Mechanical input / +), it is suggested that osteocytes are dispensable for this recovery phase and that osteoclasts (Oc) and osteoblasts (Ob) respond to the reloading stimulus, bypassing osteocytes, with reversal of elevated bone resorption and release from suppressed bone formation, respectively.

(Kalajzic et al., 2004; Yang et al., 2005) employed with a relatively long promoter conferred selective expression of DT-R in osteocytes. In addition to immunohistochemical and RT-PCR analyses, we conrmed specic expression of DT-R in osteocytes in several functional assays. First, treatment ex vivo of transgenically derived osteoblasts with DT did not
472 Cell Metabolism 5, 464475, June 2007 2007 Elsevier Inc.

induce cell death, whereas osteoblasts isolated in the same manner and transduced ex vivo with DT-R by retroviral infection responded promptly with cell death. Second, bone histomorphometric analysis shortly after DT administration to transgenic mice revealed no change in the number of osteoblasts compared with control mice. Third, bone-marrow cultures from transgenic mice treated

Cell Metabolism
Osteocytes in Mechanotransduction

in vivo with DT gave rise to the same number of cfu-F, cfuALP, and cfu-OB as cultures from wild-type mice in colony-formation assays. Finally, when primary osteoblasts isolated from transgenic mice were treated with DT, their proliferation and differentiation as determined by BrdU incorporation and ALP activity, respectively, were not affected and were indistinguishable from wild-type osteoblasts. Taken together, these results indicate that DT-R was targeted specically to osteocytes, and not to osteoblasts. Bone remodeling is performed by effector cells at the bone surface, i.e., osteoclasts and osteoblasts. We have demonstrated that osteocytes deep inside the bone serve an important function in regulating the activities of osteoclasts and osteoblasts at the bone surface and that depletion of osteocytes impairs bone homeostasis through altered activities of osteoclasts and osteoblasts. The results of gene expression analysis suggest that osteocytes control osteoclast differentiation indirectly through modulation of RANKL expression in osteoblasts. Since osteocytes make direct contact with osteoblasts and lining cells on the bone surface through gap junctions (Aarden et al., 1994), it is plausible to assume that cell-cell contact between osteocytes and osteoblasts normally restrains RANKL gene expression from aberrant activation, while osteocyte ablation or disruption of the osteocyte network releases this negative regulation, resulting in aberrantly elevated RANKL expression and osteoclastogenesis. Small molecules passing from osteocytes to osteoblasts through gap junctions may be involved in the regulation of RANKL expression in osteoblasts. Alternatively, as osteocytes extend cell processes into the marrow space (Kamioka et al., 2001), humoral factors produced by osteocytes and released through canaliculi into the bone marrow may regulate the differentiation and activity of osteoclasts. The data suggest that osteocytes are also involved in the mineralizing function of mature osteoblasts in association with reduced expression of the osteocalcin gene. Sclerostin, the product of the SOST gene, is secreted from osteocytes (Poole et al., 2005) and negatively regulates osteoblastic bone formation by antagonizing BMP and Wnt signals (Li et al., 2005; van Bezooijen et al., 2004; Winkler et al., 2003), making it a likely candidate as an osteocyte-derived regulator of osteoblastic function. Since the expression of sclerostin mRNA decreased markedly following osteocyte ablation in the current model, which would cause stimulation of bone formation, it is conceivable that in addition to negative regulators such as sclerostin, one or more positive regulators are produced by osteocytes to maintain the mineralizing function of osteoblasts. The expression of DMP1 and FGF-23, osteocyte-derived factors that regulate mineralization and phosphate homeostasis (Feng et al., 2006; Liu et al., 2006; Shimada et al., 2004), was reduced together with that of other markers of osteocytes, such as MEPE, Phex, and E11/gp38 (Franz-Odendaal et al., 2006). Following osteocyte depletion, however, there were no alterations in the systemic calcium or phosphate concentrations (data not shown). Further studies are required to identify osteocyte

products that regulate osteoblastic as well as osteoclastic activities. Although it is widely believed that the osteocyte network serves a mechanosensory function (Aarden et al., 1994; Burger and Klein-Nulend, 1999), direct proof for this hypothesis in the intact animal has not been obtained. Osteocyte-ablated mice exhibited almost complete resistance to bone loss induced by tail suspension, which was associated with an absence of any increase in osteoclastic bone resorption in response to unloading. These results indicate that osteocytes play a crucial role in sensing the local changes in mechanical strains evoked by unloading and/or play a role in transmitting pro-osteoclastogenic signals to stimulate bone resorption in response to mechanical unloading (Figure 7D). Taken together with the accelerated bone resorption and elevated RANKL expression following osteocyte ablation under ambulatory (i.e., normal loading) conditions, this suggests that osteocytes transduce opposite signals, an anti-osteoclastogenic signal (under normal loading) versus a pro-osteoclastogenic signal (in response to unloading), depending on the nature and direction of mechanical inputs (Figure 7D). Osteoporosis has been viewed as a disease encompassing osteoclasts and osteoblasts, in which the boneresorbing activity of the former exceeds the bone-forming ability of the latter, and its treatment as well as diagnosis has been directed toward these two major cell types. It has been reported that osteocyte density declines with aging (Qiu et al., 2002), and osteocyte apoptosis has been documented in estrogen deciency (Tomkinson et al., 1997) and glucocorticoid treatment (Weinstein et al., 2000). The ndings that the loss of osteocytes results in not only reduced bone mass but compromised bone quality with trabecular microstructural alterations, intracortical porosity, and microfractures suggest that osteocyte deciency/malfunction may underlie bone fragility under various conditions (Seeman and Delmas, 2006), and that osteocytes therefore comprise an important target for the development of diagnostics and therapeutics for bone disease, as exemplied by osteoporosis.
EXPERIMENTAL PROCEDURES Reagents Diphtheria toxin was purchased from Sigma and dissolved in PBS. Polyclonal goat anti-DT-R (hHB-EGF) antibody was obtained from Genzyme-Techne. Generation of Transgenic Mice and Animal Experiments A transgenic cassette carrying a genomic fragment comprised of the 9.6 kb 50 -anking region, exon 1, intron 1, and a part of exon 2 of the mouse DMP1 gene fused to human HB-EGF (DT-R) cDNA with a polyadenylation signal was constructed. Founder mice were generated by microinjection of the DNA into fertilized eggs of C57BL/6 mice. Integration of the transgene was conrmed by Southern and PCR analyses of genomic DNA extracted from the tail. Mice were raised under standard laboratory conditions at 24 C 2 C and 50%60% humidity and were allowed free access to tap water and commercial standard rodent chow (CE-2) containing 1.20% calcium, 1.08% phosphate, and 240 IU/100 g vitamin D3 (Clea Japan Inc.). Transgenic mice were injected i.p. with 250 mg/kg body weight

Cell Metabolism 5, 464475, June 2007 2007 Elsevier Inc. 473

Cell Metabolism
Osteocytes in Mechanotransduction

DT in PBS. Unloading of the hindlimbs was induced by tail suspension for 7 days, followed by reloading for 14 days, as described previously (Globus et al., 1984; Sessions et al., 1989). All experiments were performed in accordance with National Center for Geriatrics and Gerontologys ethical guidelines for animal care, and the experimental protocols were approved by the animal care committee. Cell Culture and Gene Expression Studies Bone-marrow cells were prepared from C57BL/6 mice and cultured in aMEM medium supplemented with 10% FCS as described (Takeshita et al., 2002). Retroviral vectors encoding DT-R were used to transfect Plat-E retrovirus packaging cells (a gift from T. Kitamura, University of Tokyo). Primary calvarial osteoblasts were infected with the retroviral vector pMX-puro expressing DT-R under the control of Mo-MLV LTR, cultured in the presence of polybrene (4 mg/ml) for 1 day, and then treated with DT in the presence of puromycin (1.6 mg/ml). Apoptosis of osteoblasts was assessed using a Cell Death Detection ELISA kit (Roche Diagnostics). Osteoclast differentiation was evaluated as described previously (Takeshita et al., 2002). Total RNA was isolated from bone with TRIzol Reagent (Invitrogen). Tissue distribution of DT-R was analyzed by RT-PCR with the primer set 50 -TCGAGAACTTCGCTGAGG-30 and 50 -CGCCAGTCAC CAGTGCCGAG-30 at 94 C for 30 s, 55 C for 1 min, 72 C for 1 min for 35 cycles. For quantitative RT-PCR, total RNA (1 mg) was reverse transcribed using SuperScript III (Invitrogen), and samples were analyzed using a LightCycler system (Roche Diagnostics). The primers included 50 GGCTGTCCTGTGCTCTCCCAG-30 and 50 -GGTCACTATTTGCCTGT GCCTC-30 for DMP1, 50 -ACTTGTCGCAGAAGCATC-30 and 50 -GTGG GCGAACAGTGTAGAA-30 for FGF-23, 50 -CTTCAGGAATGATGCCAC AGAGGT-30 and 50 -ATCTTTGGCGTCATAGGGATGGTG-30 for sclerostin, 50 -AGGACCCAAGGGCAAGAG-30 and 50 -TGTCTTCATTCGGCA TTGG-30 for MEPE, 50 -GTGCATCTACCAACCAGATACG-30 and 50 TCTGTTCCCCAAAAGAAAGG-30 for Phex, 50 -CAGTGTTGTTCTGGG TTTTGG-30 and 50 -TGGGGTCACAATATCATCTTCA-30 for E11/gp38, 50 -TGGAAGGCTCATGGTTGGAT-30 and 50 -CATTGATGGTGAGGTGT GCAA-30 for RANKL, 50 -CTTGCCTTGATGGAGAGCCT-30 and 50 TCGCTCGATTTGCAGGTCT-30 for OPG, 50 -CACTGCCACCTCTGA CTTCT-30 and 50 -GCTCTCAGTGAGGGATGAAA-30 for Runx2, 50 CAGGGTACACCATGATCTCACC-30 and 50 -CGCCCATACCATCTCC CAGG-30 for ALP, 50 GAGGACCATCTTTCTGCTCAC-30 and 50 -CCA AGGTAGCGCCGGAGTCTG-30 for osteocalcin, and 50 -TGCTGCCA TTGTTGATATGG-30 and 50 -TCCACAGCTTTGATGACACC-30 for EF-1a. The amount of target mRNA was normalized to EF-1a mRNA. Bone Analysis Eight- to eighteen-week-old transgenic mice and their wild-type littermates were perfused with 4% paraformaldehyde in 0.1 M cacodylate buffer (pH 7.4). Femurs and tibiae were immersed in the same xative for 12 hr prior to decalcication with 10% EDTA (pH 7.4) for 2 weeks. These specimens were embedded in parafn and subjected to histochemistry for TRAP and ALP, as described previously (Amizuka et al., 1999). The other specimens were postxed with a mixture of 1% osmium tetroxide and 1.5% potassium ferrocyanide for 4 hr and dehydrated and embedded in epoxy resin (Taab). Femoral and tibial sections were employed for von Kossa staining and electron microscopy examinations (Hitachi H-7000, Hitachi Co. Ltd.), respectively. Immunohistochemical analysis was performed on decalcied sections by using commercially available antibodies. Microcomputed tomography scanning was performed on proximal tibiae and lumbar vertebrae using mCT-40 (SCANCO Medical AG) with a resolution of 12 mm, and microstructure parameters were calculated three-dimensionally as described previously (Ito et al., 2005). Bone histomorphometry was performed on undecalcied sections with tetracycline and calcein double labeling. Histomorphometric parameters were measured at the Ito Bone Science Institute. Bone strength was evaluated at the femur by using the four-point bending test (Brodt et al., 1999).

Statistical Analysis Data are expressed as the mean SEM. Statistical analysis was performed by unpaired Students t test or ANOVA followed by Dunnetts test or the Student-Newman-Keuls test. Values were considered statistically signicant at p < 0.05. Supplemental Data Supplemental Data include ve gures and can be found with this article online at http://www.cellmetabolism.org/cgi/content/full/5/6/464/ DC1/. ACKNOWLEDGMENTS We thank A. Matsuda for isolation of the mouse DMP1 gene, N. Motoyama for advice on transgenic construction, T. Kitamura (University of Tokyo) for pMX-puro, S. Niida for help in taking the microphotographs, K.-i. Miyamoto (Tokushima University) for encouragement and support, K. Tsutsumi for technical assistance, A. Ito (Ito Bone Science Institute) for valuable suggestions on bone histomorphometry, and members of the National Center for Geriatrics and Gerontology for stimulating discussions. This study was supported in part by a grant from the program Grants-in-Aid for Scientic Research (B) from the Ministry of Education, Culture, Sports, Science and Technology of Japan (#15390290 to K. Ikeda), a grant from the Promotion of Niigata University Research Project (to N.A.), a grant from the program Promotion of Fundamental Studies in Health Sciences of the National Institute of Biomedical Innovation of Japan (MF-14 and 06-31 to K. Ikeda and M.I.), and a grant from the Novartis Foundation for Gerontological Research (to K. Ikeda). This research was carried out as part of the Ground-based Research Program for Space Utilization promoted by the Japan Space Forum. Pacic Edit reviewed the manuscript prior to submission. Received: October 17, 2006 Revised: February 14, 2007 Accepted: May 1, 2007 Published: June 5, 2007 REFERENCES Aarden, E.M., Burger, E.H., and Nijweide, P.J. (1994). Function of osteocytes in bone. J. Cell. Biochem. 55, 287299. Amizuka, N., Kwan, M.Y., Goltzman, D., Ozawa, H., and White, J.H. (1999). Vitamin D3 differentially regulates parathyroid hormone/ parathyroid hormone-related peptide receptor expression in bone and cartilage. J. Clin. Invest. 103, 373381. Bikle, D.D., and Halloran, B.P. (1999). The response of bone to unloading. J. Bone Miner. Metab. 17, 233244. Boyle, W.J., Simonet, W.S., and Lacey, D.L. (2003). Osteoclast differentiation and activation. Nature 423, 337342. Brodt, M.D., Ellis, C.B., and Silva, M.J. (1999). Growing C57Bl/6 mice increase whole bone mechanical properties by increasing geometric and material properties. J. Bone Miner. Res. 14, 21592166. Burger, E.H., and Klein-Nulend, J. (1999). Mechanotransduction in bonerole of the lacuno-canalicular network. FASEB J. Suppl. 13, S101S112. Ducy, P., Schinke, T., and Karsenty, G. (2000). The osteoblast: a sophisticated broblast under central surveillance. Science 289, 15011504. Ehrlich, P.J., and Lanyon, L.E. (2002). Mechanical strain and bone cell function: a review. Osteoporos. Int. 13, 688700. Feng, J.Q., Ward, L.M., Liu, S., Lu, Y., Xie, Y., Yuan, B., Yu, X., Rauch, F., Davis, S.I., Zhang, S., et al. (2006). Loss of DMP1 causes rickets and osteomalacia and identies a role for osteocytes in mineral metabolism. Nat. Genet. 38, 13101315.

474 Cell Metabolism 5, 464475, June 2007 2007 Elsevier Inc.

Cell Metabolism
Osteocytes in Mechanotransduction

Franz-Odendaal, T.A., Hall, B.K., and Witten, P.E. (2006). Buried alive: how osteoblasts become osteocytes. Dev. Dyn. 235, 176190. George, A., Sabsay, B., Simonian, P.A., and Veis, A. (1993). Characterization of a novel dentin matrix acidic phosphoprotein. Implications for induction of biomineralization. J. Biol. Chem. 268, 1262412630. Globus, R.K., Bikle, D.D., and Morey-Holton, E. (1984). Effects of simulated weightlessness on bone mineral metabolism. Endocrinology 114, 22642270. Harada, S., and Rodan, G.A. (2003). Control of osteoblast function and regulation of bone mass. Nature 423, 349355. Ito, M., Ikeda, K., Nishiguchi, M., Shindo, H., Uetani, M., Hosoi, T., and Orimo, H. (2005). Multi-detector row CT imaging of vertebral microstructure for evaluation of fracture risk. J. Bone Miner. Res. 20, 18281836. Kalajzic, I., Braut, A., Guo, D., Jiang, X., Kronenberg, M.S., Mina, M., Harris, M.A., Harris, S.E., and Rowe, D.W. (2004). Dentin matrix protein 1 expression during osteoblastic differentiation, generation of an osteocyte GFP-transgene. Bone 35, 7482. Kamioka, H., Honjo, T., and Takano-Yamamoto, T. (2001). A threedimensional distribution of osteocyte processes revealed by the combination of confocal laser scanning microscopy and differential interference contrast microscopy. Bone 28, 145149. Karsenty, G., and Wagner, E.F. (2002). Reaching a genetic and molecular understanding of skeletal development. Dev. Cell 2, 389406. Kato, Y., Windle, J.J., Koop, B.A., Mundy, G.R., and Bonewald, L.F. (1997). Establishment of an osteocyte-like cell line, MLO-Y4. J. Bone Miner. Res. 12, 20142023. Li, X., Zhang, Y., Kang, H., Liu, W., Liu, P., Zhang, J., Harris, S.E., and Wu, D. (2005). Sclerostin binds to LRP5/6 and antagonizes canonical Wnt signaling. J. Biol. Chem. 280, 1988319887. Ling, Y., Rios, H.F., Myers, E.R., Lu, Y., Feng, J.Q., and Boskey, A.L. (2005). DMP1 depletion decreases bone mineralization in vivo: an FTIR imaging analysis. J. Bone Miner. Res. 20, 21692177. Liu, S., Zhou, J., Tang, W., Jiang, X., Rowe, D.W., and Quarles, L.D. (2006). Pathogenic role of Fgf23 in Hyp mice. Am. J. Physiol. Endocrinol. Metab. 291, E38E49. Manolagas, S.C., and Jilka, R.L. (1995). Bone marrow, cytokines, and bone remodeling. Emerging insights into the pathophysiology of osteoporosis. N. Engl. J. Med. 332, 305311. Martin, T.J., and Sims, N.A. (2005). Osteoclast-derived activity in the coupling of bone formation to resorption. Trends Mol. Med. 11, 7681. Nakashima, K., Zhou, X., Kunkel, G., Zhang, Z., Deng, J.M., Behringer, R.R., and de Crombrugghe, B. (2002). The novel zinc nger-containing transcription factor osterix is required for osteoblast differentiation and bone formation. Cell 108, 1729. Pacici, R. (1996). Estrogen, cytokines, and pathogenesis of postmenopausal osteoporosis. J. Bone Miner. Res. 11, 10431051. Partt, A.M. (1994). Osteonal and hemi-osteonal remodeling: the spatial and temporal framework for signal trafc in adult human bone. J. Cell. Biochem. 55, 273286. Partt, A.M. (2002). Targeted and nontargeted bone remodeling: relationship to basic multicellular unit origination and progression. Bone 30, 57. Poole, K.E., van Bezooijen, R.L., Loveridge, N., Hamersma, H., Papapoulos, S.E., Lowik, C.W., and Reeve, J. (2005). Sclerostin is a delayed secreted product of osteocytes that inhibits bone formation. FASEB J. 19, 18421844. Qiu, S., Rao, D.S., Palnitkar, S., and Partt, A.M. (2002). Age and distance from the surface but not menopause reduce osteocyte density in human cancellous bone. Bone 31, 313318. Raisz, L.G. (2005). Pathogenesis of osteoporosis: concepts, conicts, and prospects. J. Clin. Invest. 115, 33183325.

Riggs, B.L., and Melton, L.J., 3rd. (1986). Involutional osteoporosis. N. Engl. J. Med. 314, 16761686. Saito, M., Iwawaki, T., Taya, C., Yonekawa, H., Noda, M., Inui, Y., Mekada, E., Kimata, Y., Tsuru, A., and Kohno, K. (2001). Diphtheria toxin receptor-mediated conditional and targeted cell ablation in transgenic mice. Nat. Biotechnol. 19, 746750. Sambrook, P., and Cooper, C. (2006). Osteoporosis. Lancet 367, 20102018. Seeman, E. (2002). Pathogenesis of bone fragility in women and men. Lancet 359, 18411850. Seeman, E., and Delmas, P.D. (2006). Bone qualitythe material and structural basis of bone strength and fragility. N. Engl. J. Med. 354, 22502261. Sessions, N.D., Halloran, B.P., Bikle, D.D., Wronski, T.J., Cone, C.M., and Morey-Holton, E. (1989). Bone response to normal weight bearing after a period of skeletal unloading. Am. J. Physiol. 257, E606E610. Shimada, T., Kakitani, M., Yamazaki, Y., Hasegawa, H., Takeuchi, Y., Fujita, T., Fukumoto, S., Tomizuka, K., and Yamashita, T. (2004). Targeted ablation of Fgf23 demonstrates an essential physiological role of FGF23 in phosphate and vitamin D metabolism. J. Clin. Invest. 113, 561568. Takayanagi, H., Ogasawara, K., Hida, S., Chiba, T., Murata, S., Sato, K., Takaoka, A., Yokochi, T., Oda, H., Tanaka, K., et al. (2000). T-cell-mediated regulation of osteoclastogenesis by signalling crosstalk between RANKL and IFN-gamma. Nature 408, 600605. Takeshita, S., Namba, N., Zhao, J.J., Jiang, Y., Genant, H.K., Silva, M.J., Brodt, M.D., Helgason, C.D., Kalesnikoff, J., Rauh, M.J., et al. (2002). SHIP-decient mice are severely osteoporotic due to increased numbers of hyper-resorptive osteoclasts. Nat. Med. 8, 943949. Teitelbaum, S.L., and Ross, F.P. (2003). Genetic regulation of osteoclast development and function. Nat. Rev. Genet. 4, 638649. Tomkinson, A., Reeve, J., Shaw, R.W., and Noble, B.S. (1997). The death of osteocytes via apoptosis accompanies estrogen withdrawal in human bone. J. Clin. Endocrinol. Metab. 82, 31283135. Toyosawa, S., Shintani, S., Fujiwara, T., Ooshima, T., Sato, A., Ijuhin, N., and Komori, T. (2001). Dentin matrix protein 1 is predominantly expressed in chicken and rat osteocytes but not in osteoblasts. J. Bone Miner. Res. 16, 20172026. van Bezooijen, R.L., Roelen, B.A., Visser, A., van der Wee-Pals, L., de Wilt, E., Karperien, M., Hamersma, H., Papapoulos, S.E., ten Dijke, P., and Lowik, C.W. (2004). Sclerostin is an osteocyte-expressed negative regulator of bone formation, but not a classical BMP antagonist. J. Exp. Med. 199, 805814. Weinstein, R.S., Nicholas, R.W., and Manolagas, S.C. (2000). Apoptosis of osteocytes in glucocorticoid-induced osteonecrosis of the hip. J. Clin. Endocrinol. Metab. 85, 29072912. Winkler, D.G., Sutherland, M.K., Geoghegan, J.C., Yu, C., Hayes, T., Skonier, J.E., Shpektor, D., Jonas, M., Kovacevich, B.R., StaehlingHampton, K., et al. (2003). Osteocyte control of bone formation via sclerostin, a novel BMP antagonist. EMBO J. 22, 62676276. Yang, W., Lu, Y., Kalajzic, I., Guo, D., Harris, M.A., Gluhak-Heinrich, J., Kotha, S., Bonewald, L.F., Feng, J.Q., Rowe, D.W., et al. (2005). Dentin matrix protein 1 gene cis-regulation: use in osteocytes to characterize local responses to mechanical loading in vitro and in vivo. J. Biol. Chem. 280, 2068020690. Ye, L., Mishina, Y., Chen, D., Huang, H., Dallas, S.L., Dallas, M.R., Sivakumar, P., Kunieda, T., Tsutsui, T.W., Boskey, A., et al. (2005). Dmp1-decient mice display severe defects in cartilage formation responsible for a chondrodysplasia-like phenotype. J. Biol. Chem. 280, 61976203.

Cell Metabolism 5, 464475, June 2007 2007 Elsevier Inc. 475