Evaluation of Pharmaceuticals With a Novel 50-Hour Animal Model of Bone Loss

Authors


  • Drs. Tomimori, Mori, and Yasuda are employees of Oriental Yeast Co. All other authors state that they have no conflicts of interest

  • Published online on February 16, 2009

Abstract

Osteoporosis remains a major public health problem through its associated fragility fractures. Several animal models for the study of osteoporotic bone loss, such as ovariectomy (OVX) and denervation, require surgical skills and several weeks to establish. Osteoclast differentiation and activation is mediated by RANKL. Here we report the establishment of a novel and rapid bone loss model by the administration of soluble RANKL (sRANKL) to mice. Mice were injected intraperitoneally with sRANKL and used to evaluate existing anti-osteoporosis drugs. sRANKL decreased BMD within 50 h in a dose-dependent manner. The marked decrease in femoral trabecular BMD shown by pQCT and the 3D images obtained by μCT were indistinguishable from those observed in the OVX model. Histomorphometry showed that osteoclastic activity was significantly increased in the sRANKL-injected mice. In addition, serum biochemical markers of bone turnover such as Ca, C-telopeptide of type 1 collagen (CTX), and TRACP5b were also significantly increased in the sRANKL-injected mice in a dose-dependent manner. Bisphosphonates (BPs), selective estrogen receptor modulators (SERMs), and PTH are commonly used for the treatment of osteoporosis. We successfully evaluated the effects of anti–bone-resorbing agents such as BPs, a SERM, and anti–RANKL-neutralizing antibody on bone resorption in a couple of weeks. We also evaluated the effects of PTH on bone formation in 2 wk. A combination of sRANKL injections and OVX made it possible to evaluate a SERM. The sRANKL model is the simplest, fastest, and easiest of all osteoporosis models and could be useful in the evaluation of drug candidates for osteoporosis.

INTRODUCTION

Osteoporosis is a common bone disease characterized by reduced bone and increased risk of fracture. In postmenopausal women, osteoporosis results from bone loss attributable to estrogen deficiency.(1)

Inhibitors of bone resorption and bone anabolic drugs such as bisphosphonates (BPs) and PTH are used for the treatment of osteoporosis(2,3). BPs are structural analogs of pyrophosphoric acid and induce apoptosis in osteoclasts, thereby preventing bone resorption. PTH regulates calcium and phosphate homeostasis by acting primarily on target cells in bone(2,3). PTH also plays other important roles, such as in the regulation of calcium and magnesium excretion in the kidney.(4) PTH and BPs increase BMD and bone strength in humans and are considered effective treatments for osteoporosis and osteopenia.(2,3)

Many researchers use experimental animal models of osteoporosis, such as produced using a low-calcium diet, sciatic denervation, or tail suspension,(5–11) and the ovariectomized (OVX) rat is the standard model for osteoporosis in postmenopausal women.

In the past decade, we and others have identified three key factors, RANK (also called TNFRSF11A), RANKL (also called TRANCE, ODF, OPGL, and TNFSF11), and osteoprotegerin (OPG, also called OCIF and TNFRSF11B), that play critical roles in osteoclast differentiation and function.(12–18) RANKL, a member of the TNF family, is expressed as a membrane-associated protein in osteoblasts/stromal cells in response to bone-resorbing factors.(14,16) Osteoclast progenitors that express RANK, a member of the TNF receptor family, recognize RANKL through either cell-to-cell interactions with osteoblasts/stromal cells or interactions with sRANKL secreted from these cells.(18–20) Concurrent stimulation with sRANKL and macrophage colony-stimulating factor (M-CSF) induces the differentiation of progenitors into osteoclasts.(14,16) OPG, a secreted member of the TNF receptor family produced primarily by osteoblasts/stromal cells, acts as a decoy receptor for RANKL to inhibit osteoclastogenesis.(12,13,15) The ratio of RANKL to OPG is important in determining osteoclast formation.(21) RANKL also plays an important role in the immune system.(17,18) We, and others have shown the suppressive effects of RANKL on the immune response.(22,23) However, the importance of RANKL has been much more shown in bone metabolism than in the immune system.(24)

We focused on sRANKL and have established a novel and rapid bone loss model in mice through sRANKL injections in a procedure that is completed within 50 h. This approach makes use of the fact that exogenous sRANKL mediates osteoclast differentiation and activates osteoclasts to resorb bone. The speed and simplicity of this model could facilitate preclinical work in the discovery of drug candidates for the treatment of osteoporosis.

MATERIALS AND METHODS

Mice and materials

C57BL/6 mice were obtained from Charles River Japan. They were maintained under a 12-h light/dark cycle at 20°C and provided with water and food ad libitum. Mice were treated according to the institutional ethical guidelines for animal experimentation and safety guidelines. Recombinant human sRANKL, a fusion protein of glutathione S-transferase (GST) and the extracellular domain of human RANKL, was prepared as follows. Human RANKL (NM 003701) cDNA encoding amino acid residues 140–317 was obtained by PCR with the following primers: RANKL forward, 5′-ATCGGGAATTCCCATCAGAGCAGAGAAAGCGAT-3′ (574–593); RANKL reverse, 5′-TCAGGCCGTCGACATCCAGGAAATACATAACA-3′ (1127–1146). The resulting PCR products were subcloned into pGEX-4T-2(GE Healthcare) to add a GST tag. GST-fusion protein was expressed in the E. coli BL21 strain and purified using glutathione-Sepharose. Contaminating lipopolysaccharide (LPS) was <0.1 EU/mg as measured by limulus amebocyte lysate assay (LAL assay). Anti-human RANKL monoclonal antibodies were produced by hybridomas prepared by fusing mouse myeloma cells with B cells obtained from mice immunized with purified recombinant human sRANKL and purified by affinity chromatography. Serum C-telopeptide of type 1 collagen (CTX), Ca, and TRACP-5b were measured according to the manufacturer's instructions (Nordic Bioscience, Wako, and Immunodiagnostic Systems, respectively). Risedronate [1-hydroxy-2-(3–pyridinyl)ethylidene bisphosphonate] was purchased from Ajinomoto (Tokyo, Japan) and dissolved in PBS.(25) Human PTH(1-34) was purchased from Sigma-Aldrich (St. Louis, MO, USA) and dissolved in PBS.

TRACP assay

Osteoclasts were induced from bone marrow macrophages as previously described.(26) Bone marrow cells from ddY mice (Clea Japan, Tokyo, Japan) were cultured in the presence of recombinant human M-CSF (10,000 U/ml, Leukoprol; Kyowa Hakko Kirin, Tokyo, Japan). After 24 h, nonadherent cells were collected and cultured for 48 h with M-CSF. Thereafter, the cells were cultured in the presence of M-CSF and 5 nM sRANKL. After 3 days the cells were fixed and stained for TRACP. Murine sRANKL was purchased from Peprotech (London, UK). RAW264 (murine macrophage cell line) cells obtained from RIKEN Cell Bank were cultured in the presence of 5 nM human or murine sRANKL with or without anti-human RANKL-neutralizing antibody. After 3 days, the cells were fixed, and TRACP activity was measured as described elsewhere.(12,15) After 5 days, the cells were fixed and stained for TRACP.

Bone loss model mice

To establish the bone loss model, sRANKL (0.5, 1, and 2 mg/kg) or PBS was injected intraperitoneally at 24-h intervals for 3 days into 7-wk-old female mice (n = 10). The mice were killed 90 min after the last injection, and blood samples, tibias, and femora were harvested for analysis.(27) For longer treatment, sRANKL (2 mg/kg) was injected intraperitoneally at 24-h intervals for 7 days into 7-wk-old female mice (n = 5) that were killed 24 h after the last injection. To examine recovery from the established bone loss model, sRANKL (1 mg/kg) was injected intraperitoneally at 24-h intervals for 3 days into 7-wk-old female mice (n = 5) that were killed 90 min (0 wk) and 1, 4, 6, and 8 wk after the last injection.

Comparison of OVX and sRANKL model mice was carried out in four groups as follows.(10,11,28) Seven-week-old female mice (n = 6) were OVX or underwent sham operation under anesthesia and were kept for 28 days. In two other groups, 7-wk-old female mice (n = 6) were kept for 24 days, and sRANKL (1 mg/kg) or PBS was injected intraperitoneally on days 25 and 26. All mice were killed on day 28, and blood samples and femora were harvested for measurement of bone resorption biomarkers and pQCT analysis, respectively.

Evaluation of a BP using sRANKL-injected bone loss model mice was performed as follows. Risedronate (10 μg/kg) was injected subcutaneously at 24-h intervals for 3 days before the first sRANKL injection, and the mice (n = 5) were then injected with risedronate (subcutaneously) at 24-h intervals for 2 days and sRANKL (1 mg/kg, IP) at 24-h intervals for 3 days. To shorten the experimental period, risedronate (30 μg/kg) was injected subcutaneously 24 h before the first sRANKL injection, and the mice (n = 6) were given simultaneous injections of risedronate (subcutaneously) and sRANKL (1 mg/kg, IP) at 24-h intervals for 3 days. Mice were killed 90 min after the last injection of sRANKL, and blood samples and femora were collected for measurement of serum TRACP5b and for pQCT and μCT analyses.(27,29,30)

Evaluation of raloxifene using sRANKL-injected bone loss model mice was performed as follows. sRANKL (1 mg/kg) or PBS was injected intraperitoneally at 24-h intervals for 2 days into 7-wk-old female mice (n = 6). Twenty-four hours after the second sRANKL injection, mice were OVX or underwent sham operation under anesthesia. Twenty-four hours after OVX or sham operation, raloxifene (1, 10 mg/kg) or PBS was administered orally for 14 consecutive days. Mice were killed on day 18, 24 h after the last administration of raloxifene.

Evaluation of PTH using sRANKL-injected bone loss model mice was performed as follows. sRANKL (1 mg/kg) or PBS was injected intraperitoneally at 24-h intervals for 2 days into 7-wk-old female mice (n = 7). Forty-eight hours after the last sRANKL injection, PTH (160 μg/kg) or PBS was injected subcutaneously for 10 consecutive days.(2,3) Mice were killed on day 14, 24 h after the last injection of PTH.

Evaluation of anti-human RANKL antibody using sRANKL-injected bone loss model mice was performed as follows. Anti-human RANKL antibody (100 μg/mouse) was injected subcutaneously into 8-wk-old female mice (n = 6) on day −7 and day −4 before the first sRANKL injection. sRANKL (1 mg/kg) or PBS was injected intraperitoneally at 24-h intervals for 2 days. Mice were killed 24 h after the second injection of sRANKL, and femora were collected for pQCT analysis.

Binding assay

Ninety-six-well plates were precoated with 100 ng/ml of human or murine sRANKLs at 4°C overnight and were blocked with 1% FCS for 1 h. After washing twice with 50 mM Tris, 0.14 M NaCl, and 0.05% Tween 20 (pH 8.0), the plates were incubated with 1 μg/ml of the anti-human monoclonal antibody. After washing three times, the plates were incubated with 1 μg/ml horseradish peroxidase (HRP)-labeled anti-murine IgG. The enzyme activity was measured with the subtraction of absorbance 450–595 nm using 3,3′,5,5′-tetramethylbenzidine (TMB).

Measurement of BMD

Tomographic measurements of trabecular BMD were performed by pQCT (XCT Research SA+; Stratec Medizintechnik) using a voxel size of 0.08 × 0.08 × 0.46 mm. Image analysis was carried out using integrated XCT 2000 software version 6.00B. Three slices of distal femoral metaphysis starting at 0.6 mm from the growth plate were scanned at 0.2-mm intervals for the measurement of trabecular BMD.(29,30) 3D images of distal femoral metaphysis were reconstructed by μCT analysis at 0.2–1.2 mm from the growth plate using a ScanXmate-A080 (Comscan Tecno).(30)

Histomorphometry

Tibias were fixed with 70% ethanol, and the undecalcified bones were embedded in glycol methacrylate, after which 3-μm sections were cut longitudinally in the proximal region of the tibia and stained with toluidine blue O. Some sections were also stained for TRACP. Histomorphometry was performed with an image analyzing system (Osteoplan II; Carl Zeiss, Thornwood, NY, USA) linked to a light microscope. Histomorphometric measurements were made at ×400 magnification at 0.3–1.2 mm in the secondary spongiosa area from the growth plate–metaphysis junction. The bone volume/tissue volume (BV/TV, %), trabecular thickness (Tb.Th, μm), trabecular separation (Tb.Sp), total osteoid volume/bone volume (OV/BV), osteoid surface/bone surface (OS/BS), osteoblast surface/bone surface (Ob.S/BS), trabecular number (Tb.N, μm), osteoid thickness (O.Th, μm), eroded surface/bone surface (ES/BS, %), osteoclast number/bone perimeter (N.Oc/B.Pm/100 mm), and osteoclast surface/bone surface (Oc.S/BS, %) were calculated and expressed according to standard formulas and nomenclatures.(31)

Statistical analysis

Statistical analyses were performed using ANOVA with Dunnett's test or Student's t-test. p < 0.05 was considered significant.

RESULTS

Osteoporotic bone loss induced by sRANKL injections

We prepared a GST fusion protein containing the extracellular domain of human RANKL. GST-RANKL had a strong effect on osteoclast morphology and TRACP (an osteoclast enzyme marker) activity in an in vitro osteoclastogenesis assay (Figs. 1A and 1B). GST-RANKL is hereafter referred to as sRANKL.

Figure Figure 1.

Characterization of the sRANKL-injected bone loss model. GST-RANKL–induced osteoclasts from bone marrow macrophages (A) and RAW 264 cells (B) in the presence and absence, respectively, of M-CSF. Cells were fixed and stained for TRACP. Bars = 250 μm. (C) Experimental design. PBS or sRANKL (0.5, 1, and 2 mg/kg) was injected intraperitoneally at 24-h intervals for 3 days into 7-wk-old female mice (n = 10). Mice were killed 90 min after the last injection. (D) Serum biochemical markers of bone turnover (Ca, CTX, and TRACP5b) were measured. Data are mean ± SD. ap < 0.05 and bp < 0.01 (ANOVA) compared with the PBS-injected group. (E) Trabecular BMD was measured by pQCT at sites 0.6, 0.8, and 1 mm proximal to the distal femoral growth plate. Data are mean ± SD. bp < 0.01 (ANOVA) compared with the PBS-injected group. (F) μCT 3D images at 0.2–1.2 mm from the growth plate of femora of mice treated with sRANKL (0, 0.5, 1, 2 mg/kg). (G) Experimental design. PBS or sRANKL (2 mg/kg) was injected intraperitoneally at 24-h intervals for 7 days into 7-wk-old female mice (n = 6). Mice were killed 24 h after the last injection. (H) Trabecular BMD was measured by pQCT 1.0 mm proximal to the distal femoral growth plate in mice shown in G. Data are mean ± SD. bp < 0.01 (Student's t-test) compared with the PBS-injected group. (I) Serum ALP levels were measured in mice shown in G. Data are mean ± SD. bp < 0.01 (Student's t-test) compared with the PBS-injected group. (J) 3D images of femora from each of the groups shown in G.

To promote bone loss, sRANKL was injected intraperitoneally into 7-wk-old female mice at 24-h intervals for 3 days at doses of 0.5, 1, and 2 mg/kg (Fig. 1C). The mice were killed 1.5 h after the third injection.(27) We measured serum Ca, CTX, and TRACP5b and found that each of these parameters was increased by sRANKL injections in a dose-dependent manner (Fig. 1D). Tomographic measurements of femoral trabecular BMD were performed using pQCT at 0.6, 0.8, and 1 mm from the distal growth plate.(29,30) sRANKL injections markedly decreased BMD at each position in a dose-dependent manner in 50 h (Fig. 1E). For example, sRANKL injections (0.5, 1, and 2 mg/kg) produced decreases in BMD of 23%, 31%, and 44%, respectively, 1 mm proximal to the growth plate.

Thereafter, distal femoral metaphyses were analyzed by μCT.(30) Consistent with the results of pQCT analysis, 3D images of distal femoral metaphyses reconstructed by μCT showed a drastic loss of trabecular bone in a dose-dependent manner (Fig. 1F).

The ultimate effect of sRANKL treatment was determined by intraperitoneally injection of 2 mg/kg sRANKL to mice for 7 consecutive days (Fig. 1G). The sRANKL injections markedly decreased BMD 1 mm proximal to the growth plate by 47% (Fig. 1H), and 3D images of distal femoral metaphyses showed the marked loss of trabecular bone (Fig. 1J). However, the effect was comparable to that observed after only three injections of sRANKL at the same dose (Figs. 1E, 1F, 1H, and 1J). Two or three injections of sRANKL were sufficient to induce osteoporotic bone loss. Serum ALP levels were robustly increased in the sRANKL-treated mice (Fig. 1I).

Histological analysis of tibias stained with toluidine blue and stained for TRACP confirmed the marked bone loss and showed a remarkable increase in numbers of TRACP+ multinucleated giant cells at resorption sites in the sRANKL-injected mice (Figs. 2A–2C).

Figure Figure 2.

Histomorphometric analysis of the sRANKL-induced bone loss model. Toluidine blue (A) and TRACP (B) staining of undecalcified sections of tibia in the PBS-injected and sRANKL (2 mg/kg)-injected groups shown in Figs. 1C–1F. Bars = 125 μm. (C) High-power field of TRACP staining shown in B. Bars = 500 μm. (D) Tibias of sRANKL-injected mice (2 mg/kg) were analyzed by histomorphometry. BV/TV, Tb.Th, Tb.N, Tb.Sp, OV/BV, OS/BS, Ob.S/BS, O.Th, ES/BS, N.Oc/B.Pm, and Oc.S/BS were measured. Data are mean ± SD. ap < 0.05 and bp < 0.01 (Student's t-test) compared with the PBS-injected group.

To analyze changes in bone structure in more detail, tibias were analyzed by histomorphometry.(31) Histomorphometric analysis of sRANKL-treated mice showed that BV/TV, Tb.Th, and Tb.N decreased by 30%, 10%, and 20%, respectively, compared with control mice. In contrast, O.Th, ES/BS, N.Oc/B.Pm, and Oc.S/BS increased 1.2-, 1.8-, 1.9-, and 1.8-fold, respectively. Tb.Sp, OV/BV, OS/BS, and Ob.S/BS were unchanged (Fig. 2D).

The sRANKL model was also established in male animals (data not shown). Furthermore, the model could be established in other strains of mice, such as ICR, and other species, such as rats (data not shown). In control experiments, we injected GST alone into mice and noted no significant changes in BMD or any histomorphometric parameter described above (data not shown). We also examined histological sections of cerebrum, lung, heart, liver, thymus, spleen, kidney, and skin from sRANKL-injected mice and found no abnormalities in any of these tissues (data not shown).

To further characterize the sRANKL model, the systemic clearance of sRANKL was determined. Serum levels of exogenous sRANKL peaked at 4 h after injection and subsequently decreased rapidly, no longer being apparent at 24 h (Fig. 3A). Serum TRACP5b levels were significantly increased 12 h after a single injection of sRANKL and had returned to normal 24 h after injection (Fig. 3B). Mild bone loss (12% decrease in BMD) was confirmed only at the site 1 mm proximal to the growth plate 24 h after a single injection, and no further decrease in BMD was observed 48 h after the injection (Figs. 3C and 3D). The observation that three injections of sRANKL induced significant and robust decreases in BMD at sites 0.6, 0.8, and 1 mm proximal to the growth plate (Fig. 1D) means that the second injection further decreased BMD. A single injection of sRANKL also failed to maintain an elevated level of TRACP5b for 24 h (Fig. 3B), indicating that daily injections of sRANKL are required for effective bone resorption in this model.

Figure Figure 3.

The effect of a single injection of sRANKL in mice. PBS or sRANKL (1 mg/kg) was injected intraperitoneally once into 7-wk-old female mice (n = 5). Serum was collected at the indicated time points after the injection of sRANKL and 48 h after the injection of PBS. (A) Kinetics of sRANKL in serum after injection. (B) Serum TRACP5b levels were measured at the indicated times after the sRANKL injection. Data are mean ± SD. ap < 0.05 (ANOVA) compared with the PBS-injected group. (C) Trabecular BMD was measured by pQCT at sites 0.6, 0.8, and 1 mm proximal to the distal femoral growth plate in mice injected with PBS or sRANKL. Data are mean ± SD. ap < 0.05 and bp < 0.01 (ANOVA) compared with the PBS-injected group. (D) 3D images of femora of mice injected with PBS or sRANKL (24 and 48 h after injection). (E) PBS or sRANKL (1 mg/kg) was injected intraperitoneally at 24-h intervals for 3 days into 7-wk-old female mice (n = 10). Mice were killed at 90 min (0 wk) and 1, 4, 6, and 8 wk after the last injection. Trabecular BMD (top panel) was measured by pQCT 1 mm proximal to the distal femoral growth plate at the indicated time points. Serum TRACP5b (middle panel) and ALP (bottom panel) levels were measured at the indicated time points. Data are mean ± SD. ap < 0.05 and bp < 0.01 (ANOVA) compared with the PBS-injected group. (F) 3D images of femora of mice shown in E were shown at the indicated time points.

The time course of recovery after the establishment of bone loss in the sRANKL model showed that the decreases in BMD produced by sRANKL injections persisted for 4 wk and that BMD significantly increased to the value seen in the PBS injection group by 6 wk (Fig. 3E). Serum TRACP5b and ALP levels were significantly increased at the establishment of bone loss (0 wk) but were equivalent to those seen in PBS-injected mice at 1, 4, 6, and 8 wk after the establishment of bone loss. Consistent with the results of pQCT analysis, 3D images of distal femoral metaphyses showed that the bone loss was recovered in a time-dependent manner (Fig. 3F).

Comparison of the sRANKL model with the OVX model

The gold standard osteoporosis model uses OVX, which mimics postmenopausal osteoporosis(10,11) because estrogen loss induces osteoclastic bone resorption. We compared the sRANKL model with the OVX model. Whereas OVX-induced bone loss was typically observed 28 days after surgery, bone loss was observed in the sRANKL model in female mice within 3 days of two injections of sRANKL at an interval of 24 h (Fig. 4A). The femoral trabecular BMD was 12% lower in OVX mice than in sham-operated mice. Similarly, BMD was 16% lower in sRANKL-injected mice than in PBS-injected mice (Fig. 4B). Consistent with the marked decreases in BMD shown by pQCT analysis, 3D images showed robust trabecular bone loss in both the OVX and the sRANKL models (Fig. 4C). Bone structure in the two models was indistinguishable. Only 3 days was required to induce bone loss in the sRANKL model equivalent to that generally requiring 4 wk to achieve in the OVX model.

Figure Figure 4.

Comparison of the sRANKL model with the OVX model. (A) Experimental design. Seven-week-old female mice (n = 6) were OVX or subjected to sham operation under anesthesia and kept for 28 days. sRANKL (1 mg/kg) or PBS was injected intraperitoneally at 24-h intervals for 2 days into 11-wk-old female mice (n = 6). All mice were killed at 11 wk of age. (B) Femoral trabecular BMD was measured 1 mm proximal to the distal growth plate by pQCT. Four weeks were required for the establishment of osteopenia in the OVX mice compared with 3 days in the sRANKL-injected mice. Data are mean ± SD. ap < 0.05 (ANOVA). (C) 3D images of femora from each group shown in B.

Evaluation of BPs in sRANKL model mice

As a proof of concept, we asked whether sRANKL model mice could be used for the evaluation of pharmaceuticals for the treatment of osteoporosis. BPs are known to induce osteoclast apoptosis and to inhibit osteoclastogenesis and are widely used for the treatment of osteoporosis. Etidronate, alendronate, and risedronate are first-, second-, and third-generation BPs, respectively.

We first evaluated risedronate in the sRANKL model. We injected subcutaneous risedronate (10 μg/kg) at 24-h intervals for 3 days before the first sRANKL injection, and the mice were subsequently injected with risedronate (subcutaneously) at 24-h intervals for 2 days and sRANKL (1 mg/kg, IP) at 24-h intervals for 3 days (Fig. 5A). The mice were killed 1.5 h after the third injection of sRANKL.(27) The total duration required for the experimental procedures was 121.5 h. The serum TRACP5b level was increased by sRANKL and the increase was markedly inhibited by risedronate (Fig. 5B). Risedronate significantly prevented the decrease in BMD in the sRANKL model (Fig. 5C). 3D images showed a complete prevention of bone loss by risedronate (Fig. 5D).

Figure Figure 5.

Evaluation of risedronate in the sRANKL model mice. (A) Experimental design. Risedronate (Ris, 10 μg/kg) was injected subcutaneously at 24-h intervals for 3 days before the first sRANKL injection, and the mice (n = 5) subsequently received injections of Ris (subcutaneously) at 24-h intervals for 2 days and sRANKL (1 mg/kg, IP) at 24-h intervals for 3 days. The total duration required for these procedures was 121.5 h. (B) Serum TRACP5b levels and (C) femoral trabecular BMD were measured. Data are mean ± SD. bp < 0.01 (ANOVA) compared with the sRANKL-injected group. (D) 3D images of femora showing the protection of bone mass by Ris in sRANKL-injected mice shown in A and C. (E) Experimental design. Ris (30 μg/kg) was injected subcutaneously 24 h before the first sRANKL injection, and the mice (n = 6) subsequently received simultaneous injections of Ris (subcutaneously) and sRANKL (1 mg/kg, IP) at 24-h intervals for 3 days. The total duration required for these procedures was 73.5 h. Serum TRACP5b levels (F) and femoral trabecular BMD (G) were measured. Data are mean ± SD. ap < 0.05 and bp < 0.01 (ANOVA) compared with the sRANKL-injected group. (H) 3D images of femora from each group shown in E and G.

To shorten the experimental period, risedronate (30 μg/kg) was injected subcutaneously 24 h before the first sRANKL injection and the mice subsequently received simultaneous injections of risedronate (subcutaneously) and sRANKL (1 mg/kg, IP) at 24-h intervals for 3 days (Fig. 5E). The duration required to perform these experimental procedures was 73.5 h. The 4-day treatment with risedronate significantly prevented any increase in the serum TRACP5b level and decreased BMD in sRANKL model mice (Figs. 5F–5H). Similar results were obtained when we tested etidronate (30 mg/kg) and alendronate (300 μg/kg; data not shown).

Evaluation of a selective estrogen receptor modulator in sRANKL model mice

Two major categories of bone resorption inhibitors have been used for the treatment of osteoporosis: BPs and selective estrogen receptor modulators (SERMs). To test the latter category, we used the sRANKL model to evaluate raloxifene,(32) a second-generation SERM. Because SERMs need to be evaluated in mice with estrogen loss, we performed OVX after establishment of the sRANKL model (Fig. 6A). Briefly, 24 h after the last injection of sRANKL, the mice were OVX. Twenty-four hours after OVX, oral treatment with raloxifene (1, 10 mg/kg) for 2 wk increased BMD to a level significantly higher than that in the control group (Fig. 6B). 3D μCT images confirmed the results of pQCT analysis (Fig. 6C).

Figure Figure 6.

Evaluation of raloxifene and PTH in sRANKL model mice. (A) Experimental design. sRANKL (1 mg/kg) or PBS was injected intraperitoneally at 24-h intervals for 2 days into 7-wk-old female mice (n = 6). Twenty-four hours after the second sRANKL injection, mice were OVX or underwent sham operation under anesthesia. Twenty-four hours after the OVX or sham operation, raloxifene (1, 10 mg/kg) or PBS was administered orally for 14 consecutive days. Mice were killed on day 18, 24 h after the last administration of raloxifene. (B) Total femoral BMD was measured at sites 0.6 (top panel), 0.8 (middle panel), and 1 mm (bottom panel) proximal to the distal growth plate by pQCT. Data are mean ± SD. ap < 0.05 and bp < 0.01 (ANOVA) compared with the sRANKL (intraperitoneal)/PBS(oral)/OVX group. (C) 3D femoral images showing increased bone mass in sRANKL-injected and OVX mice after treatment with raloxifene (1, 10 mg/kg). (D) Experimental design. sRANKL (1 mg/kg) or PBS was injected intraperitoneally at 24-h intervals for 2 days into 7-wk-old female mice (n = 7). PTH (160 μg/kg) or PBS was injected subcutaneously for 10 days, 48 h after the second injection of sRANKL. Mice were killed on day 14, 24 h after the last injection of PTH. (E) Femoral trabecular BMD was measured 0.6 mm proximal to the distal growth plate by pQCT. Data are mean ± SD. ap < 0.05 and bp < 0.01 (ANOVA). Relative BMDs are shown as PBS (intraperitoneally)/PBS (subcutaneously)-injected group 100%. (F) 3D images of the femora from each group shown in D and E.

Evaluation of PTH in sRANKL model mice

We next tested this model for the evaluation of PTH(1-34), a commercially available bone anabolic drug.(2,3) PTH (160 μg/kg) was injected subcutaneously for 10 days after establishment of the sRANKL model (Fig. 6D). Bone loss was such that BMD in the sRANKL (intraperitoneally)/PBS (subcutaneously) injected group was 18% lower than in the PBS (intraperitoneally)/PBS (subcutaneously) injected group (Fig. 6E). PTH treatment increased BMD by 33% and 49% in the PBS (intraperitoneally) injected and sRANKL(intraperitoneally) injected groups, respectively (Fig. 6E). 3D μCT images confirmed the results of pQCT analysis (Fig. 6F). Similar results were obtained when the PTH doses were reduced to 40 and 80 μg/kg (data not shown).

Evaluation of anti-human RANKL-neutralizing antibody in sRANKL model mice

To determine whether the sRANKL model can be used for the evaluation of a novel agent, we prepared and tested anti-human RANKL monoclonal antibodies. One of these antibodies showed a strong sRANKL-neutralizing activity and specificity for human RANKL (Figs. 7A and 7B). The antibody did not bind murine RANKL but did bind human RANKL (Fig. 7A). Addition of the antibody to RAW 264 cell culture with human or murine sRANKL only inhibited the formation of TRACP+ multinucleated cells (MNCs) by human sRANKL (Fig. 7B). The antibody did not neutralize the activity of murine RANKL but did neutralize that of human RANKL (Fig. 7B). The anti-human RANKL neutralizing antibody (100 μg/mouse) or PBS was injected subcutaneously 7 and 4 days before the first sRANKL injection (Fig. 7C). Antibody treatment of the human sRANKL–injected mice completely inhibited the decrease in BMD, whereas BMD in PBS-injected control mice was unaffected by the antibody treatment (Fig. 7D).

Figure Figure 7.

Evaluation of anti-human RANKL neutralizing antibody in sRANKL model mice. (A) Specific binding to human sRANKL. Anti- human RANKL antibody did not bind to immobilized murine sRANKL but did bind to immobilized human sRANKL (B) Neutralization of sRANKL activity by anti-human RANKL antibody. RAW 264 cells were cultured in the presence of 5 nM human or murine sRANKL and/or 0.5 μg/ml anti-human RANKL antibody (Ab). After 4 days the cells were fixed, and TRACP+ MNCs were counted. (C) Experimental design. Anti-human RANKL antibody (100 μg/mouse) or PBS was injected subcutaneously into 8-wk-old female mice (n = 6) 7 and 4 days before the first sRANKL injection. sRANKL (1 mg/kg) or PBS was injected intraperitoneally at 24-h intervals for 2 days. Mice were killed 24 h after the second sRANKL injection. (D) Femoral trabecular BMD was measured 1.0 mm proximal to the distal growth plate by pQCT. Data are mean ± SD. bp < 0.01 (ANOVA).

DISCUSSION

We established a novel model of bone loss by injecting mice with sRANKL. First, the model can be established very rapidly, requiring only 50 h for completion of the experimental procedures. Although there are several useful models of bone loss such as OVX, denervation, tail suspension, and a low Ca diet, each of these requires at least several weeks(10,11) to establish. In the case of the OVX model, the gold standard for the evaluation of pharmaceuticals for postmenopausal osteoporosis, a minimum of 4 wk is generally needed.(10,11) Pharmaceutical candidates can be tested immediately after OVX or >4 wk after OVX, once the initial bone loss has taken place. In general, it takes several months to evaluate a new drug using the OVX model. The sRANKL model that we have described is comparable to the OVX model in terms of the decrease in BMD that occurs and the structural changes that are evident in 3D images. Longitudinal μCT 2D images showed that trabecular bone was predominantly lost (data not shown). Evaluation with the 50-h sRANKL model takes no longer than the in vitro osteoclastogenesis assay.(26)

Second, the model is very easy to establish, requiring only two or three injections of sRANKL without any need for the animal anesthesia or technical skills required for OVX, denervation, or tail suspension. The dose–response relationship between sRANKL and BMD indicates that the severity of the bone loss could easily be controlled by varying the amount of sRANKL injected.

Third, the mechanism of bone loss is simple, in that excessive sRANKL directly mediates the differentiation and activation of osteoclasts. TRACP staining and histomorphometry confirmed that the bone loss in the sRANKL model was caused by increased osteoclast differentiation and activation. The rapid decrease in BMD in this model seems not only to be caused by stimulation of the final differentiation of osteoclast progenitors but also to the activation of a preexisting pool of osteoclasts. Although a single injection of sRANKL was sufficient to activate osteoclasts, the observation that a second injection was necessary to reduce BMD strongly suggests that activated osteoclasts require continuous stimulation by exogenous sRANKL to exert their activity. The activation of osteoclasts by sRANKL may be different from normal osteoclast activation by membrane-bound RANKL produced by osteoblasts. Osteoblast-bound RANKL would likely continue to stimulate osteoclasts by cell-to-cell interaction for longer than exogenous sRANKL.

Fourth, the sRANKL model can be used for the evaluation of pharmaceuticals for the treatment of osteoporosis (see below). Finally, the sRANKL model is more protective of laboratory animal welfare because of the shorter experimental periods required, the lack of any requirement for anesthesia or surgery, and the lower numbers of treatments with test materials required compared with existing approaches.

The application of the sRANKL model to in vivo studies is promising, with proof of concept provided for all three major categories of pharmaceuticals used for the treatment of osteoporosis: BPs, PTH, and SERMs. It could be used for screening inhibitors of bone resorption in the period up to 50 h after the first sRANKL injection (Fig. 8). We also showed that a novel anti-human RANKL neutralizing antibody completely prevented bone loss in the sRANKL model without affecting endogenous murine RANKL because it was not cross-reactive with murine sRANKL. Namely, the anti-human RANKL antibody cannot bind osteoblastic/stromal cells that are making murine RANKL and can bind human sRANKL in serum in the treated mice. Because the antibody in the blood stream can neutralize exogenous sRANKL, the addition of exogenous sRANKL after administration of the antibody cannot induce bone resorption. Recently Shinohara et al.(33) used our sRANKL model for the evaluation of an inhibitor of tyrosine kinase Tec that is important in osteoclast differentiation. The application of the sRANKL model for novel antiresorptive agents, anti-human RANKL antibody, and an inhibitor of tyrosine kinase Tec, shows that this model can be used for the evaluation of novel pharmaceuticals.

Figure Figure 8.

Illustration of the change in BMD in the sRANKL bone loss model. sRANKL model mice can be used for the evaluation of bone resorption inhibitors (BPs, etc.) in the period up to 50 h after the first sRANKL injection. Osteoclasts (OCs) differentiate from their progenitors and are activated during the 50 h. The decrease in BMD is completely inhibited by BP injections. The combination of sRANKL injections and OVX makes it possible to evaluate SERMs in the period from 50 h to 4 wk after the first sRANKL injection. The model mice can also be used for the evaluation of bone anabolic drugs (PTH, etc.) in the period from 50 h to 4 wk after the first sRANKL injection. PTH administration markedly ameliorated bone loss in the sRANKL model and further elevated BMD to a value higher than that in a PBS-injected group. Osteoblasts (OBs) were activated by the PTH treatment and bone formation was robustly induced during the 10 days of treatment.

The model could also be useful for screening bone anabolic drugs such as PTH in the period from 50 h to 4 wk after sRANKL injection, during which time mild bone formation is proceeding (Fig. 8). The high bone turnover and remodeling that occurs in sRANKL model mice could facilitate the evaluation of bone anabolic drugs. Furthermore, the combination of sRANKL with OVX made possible the evaluation of a SERM under conditions of estrogen loss. The sRANKL model with some modifications could be applicable to any kind of drug candidate for osteoporosis and would be particularly suitable for the rapid screening of drug candidates in early development in preference to repeated in vitro experiments, which requires at least 5 days.(26) Positive candidates identified using the sRANKL model could be further evaluated using the standard OVX model.(10,11)

Although RANKL is involved in regulation of the immune system, the pretreatment of mice with BPs prevented bone loss in the sRANKL model, suggesting that their protective effect is not through an immune system altered by sRANKL because BPs do not generally show any effects on immune cells. In the sRANKL model, immune organs such as the spleen and thymus showed no histological abnormalities, suggesting that the effect of sRANKL on the immune system is less important than that on bone remodeling.

Sometimes transgenic or gene-deficient mice exhibit no significant phenotype under physiological conditions but do show abnormalities after OVX and/or ablation.(34) The sRANKL model is very useful for investigating defects in osteoclasts and their progenitors in vivo because sRANKL injections in mice cause a rapid increase in osteoclast differentiation and activation.

It is noteworthy that the coupling between bone formation and bone resorption,(35) was observed in the sRANKL model (Figs. 1, 3, and 8). The observations that O.Th was increased together with decreases in ES/BS, N.Oc/B.Pm, and Oc.S/BS in histomorphometry also supported the concept that elevated bone formation was induced after differentiation and activation of osteoclasts (Fig. 2D). We also recognized an increase in Ob.S/BS in the mice injected by sRANKL three times in histomorphometry in some experiments (data now shown). The increase in ALP level was accompanied with the increase in TRACP5b level (Figs. 1 and 3). We especially observed the marked increase in ALP activity after a week treatment of mice with sRANKL (Fig. 1). Together these results suggested that the elevated osteoclast differentiation and activation initiated coupling to stimulate bone formation. As shown in the previous report a continuous treatment of rats with sRANKL by an infusion pump induced a high turnover of bone remodeling and increased both of bone resorption marker (TRACP5b) and bone formation marker (osteocalcin).(36) Recently, we also showed that adenovirus-induced gene transfer of sRANKL in mice resulted in severe osteoporosis and hypercalcemia accompanied with the increase of bone resorption marker (TRACP5b) and bone formation marker (ALP).(37) Two or three times injections of sRANKL induced a weak coupling, whereas longer treatments induced it strongly (Figs. 1 and 3). Because OVX-induced bone loss was accompanied with a high turnover of bone remodeling, we suggested that the sRANKL model is similar in mechanism as in OVX-induced bone loss.

It is also interesting that the recovery of BMD after the establishment of sRANKL model is caused by the significant increase in BMD between 4 and 6 wk after the sRANKL injection. Because we have not detected any changes in serum TRACP5b and ALP activity between 4 and 6 wk after the sRANKL injection, it is uncertain whether the increase in BMD is caused by the increase of bone formation or the decrease of bone resorption. Identification of the mechanism of the increase in BMD will lead to a better understanding of coupling between bone resorption and bone formation. The factor(s) responsible for the coupling(35) could be activated by osteoclast differentiation and activation and it might be produced by the RANKL-activated osteoclasts. Application of the sRANKL model to basic studies with genetically manipulated mice would also facilitate the identification of novel mechanisms underlying coupling.

Acknowledgements

We thank T. J. Martin for critical review and Y. Marubayashi for the preparation of sRANKL.

Ancillary