Abstract
PTH is a potent bone anabolic factor, and its combination with antiresorptive agents has been proposed as a therapy for osteoporosis. We tested the effects of PTH, alone and in combination with the novel antiresorptive agent OPG, in a rat model of severe osteopenia. Sprague Dawley rats were sham-operated or ovariectomized at 3 months of age. Rats were untreated for 15 months, at which time ovariectomy had caused significant decreases in bone mineral density in the lumbar vertebrae and femur. Rats were then treated for 5.5 months with vehicle (PBS), human PTH-(1–34) (80 μg/kg), rat OPG (10 mg/kg), or OPG plus PTH (all three times per wk, sc). Treatment of ovariectomized rats with OPG or PTH alone increased bone mineral density in the lumbar vertebrae and femur, whereas PTH plus OPG caused significantly greater and more rapid increases than either therapy alone (P < 0.05). OPG significantly reduced osteoclast surface in the lumbar vertebrae and femur (P < 0.05 vs. sham or ovariectomized), but had no effect on osteoblast surface at either site. Ovariectomy significantly decreased the mechanical strength of the lumbar vertebrae and femur. In the lumbar vertebrae, OPG plus PTH was significantly more effective than PTH alone at reversing ovariectomy-induced deficits in stiffness and elastic modulus. These data suggest that OPG plus PTH represent a potentially useful therapeutic option for patients with severe osteoporosis.
POSTMENOPAUSAL OSTEOPOROSIS IS a major cause of skeletal morbidity, leading to more than 1.3 million pathological fractures per yr (1). Approximately 10 million postmenopausal women in the United States have osteoporosis[ bone mineral density (BMD), >2.5 sd below the young normal mean], and another 17 million have osteopenia (BMD, >1 sd below the mean) (2). Antiresorptive agents, including estrogen, calcitonin, and bisphosphonates, are beneficial for many of these patients. These resorption inhibitors protect existing bone and cause modest (2–9%) increases in bone mass, but for patients with established osteoporosis, antiresorptives fail to fully restore bone mass (1). For these patients, the stimulation of bone formation might be required to restore bone mass. Intermittent PTH therapy stimulates bone formation in excess of bone resorption and causes greater increases in bone mass and strength compared with antiresorptives (3–7). Recent clinical data demonstrated dose-dependent effects of intermittent PTH on bone density, but hypercalcemia may be a dose-limiting side-effect (8). A theoretically ideal therapy for severe osteoporosis would combine the potent osteoblast activation of intermittent PTH with effective inhibition of osteoclastic bone resorption (9). An effective antiresorptive agent might also reduce PTH-related hypercalcemia and thereby increase the safety and tolerability of this potent anabolic agent.
Numerous studies have demonstrated that the anabolic effects of PTH can be realized in animals cotreated with antiresorptives, including bisphosphonates (3, 6, 7, 9–13), calcitonin (5, 7, 10, 12, 13), E (3, 5–7, 9, 11, 12–15), and selective E receptor modulators (16). Recent clinical trials have also indicated that PTH can exert positive effects on the skeleton of patients who were pre- or cotreated with antiresorptives (17–20). These studies strongly suggest that bone resorption per se is not a requirement for the anabolic response to PTH. However, nearly all of these studies failed to demonstrate a statistically significant benefit of combination therapy over PTH therapy alone. In the vast majority of preclinical combination therapy studies, antiresorptives failed to significantly increase the anabolic effect of PTH on bone mass or BMD (3–7, 10–13, 16, 21–25). In some animal studies antiresorptives appeared to blunt the anabolic response of osteoblasts to PTH (3, 5, 7, 11, 14, 22–24, 26). In a clinical trial calcitonin slowed the rate at which PTH increased bone mass (27). It is possible that different dosing regimens from those used previously are necessary to reveal an additive effect of antiresorptives on bone mass in the PTH-treated skeleton. To date, however, the therapeutic benefit of adding an antiresorptive to intermittent PTH therapy has not been demonstrated by direct comparison in clinical trials or in a suitable preclinical model of osteoporosis such as the aged ovariectomized (OVX) rat.
Recently, a potent and naturally occurring bone resorption inhibitor was discovered that has a unique mechanism of action. OPG is a member of the TNF receptor family that acts by preventing the association of OPG ligand [OPGL (28–30), also known as RANKL (31), TRANCE (32), or ODF (33)] with the RANK receptor on osteoclasts and osteoclast precursors (34). By blocking OPGL/RANKL-induced RANK activation, OPG inhibits osteoclast differentiation, activation, and survival (35). We recently demonstrated that OPG causes the rapid disappearance of osteoclasts in mice treated with intermittent PTHrP (36). The ability of OPG to eliminate PTHrP-induced osteoclasts was also associated with significant inhibition of PTHrP-related hypercalcemia. Interestingly, intermittent PTHrP treatment caused a dramatic increase in osteoblast surface that was not inhibited by OPG cotreatment (36). It is well established that PTHrP and PTH act on the same receptor (37, 38) to initiate increases in osteoclast and osteoblast surfaces, and that both factors increase bone mass when administered intermittently (39, 40). The ability of OPG to completely reverse PTHrP-induced osteoclasts and control hypercalcemia while preserving PTHrP-induced osteoblasts suggested a unique therapeutic opportunity for osteoporosis. We therefore tested in aged OVX rats whether the combination of OPG plus PTH would cause greater increases in bone mass, density, and strength compared with PTH alone.
Materials and Methods
Animals and treatments
Female Sprague Dawley rats, obtained from Harlan Sprague Dawley, Inc. (Indianapolis, IN), were either OVX or sham-operated at 3 months of age. Success of the surgery was confirmed in all rats at death (20.5 months later) by the lack of ovarian tissue and by uterine atrophy (data not shown). Fifteen months after surgery, bone mineral density (BMD) was determined at multiple skeletal sites by dual energy x-ray absorptiometry (DEXA; Hologic, Inc., Bedford, MA). OVX rats were randomly divided into treatment groups, with 14 animals/group. Thirteen sham-operated rats served as controls. OVX rats were treated for 5.5 months with PBS (vehicle), human PTH-(1–34) (80 μg/kg; Bachem, Torrance, CA), OPG (10 mg/kg), or with OPG plus PTH (given as separate injections). The dose of PTH-(1–34) chosen was previously reported to be an optimal anabolic dose in rats (41). The recombinant OPG used included amino acids 22–194 of native rat OPG, expressed in Escherichia coli, covalently linked to 20,000 mol wt polyethylene glycol (Shearwater Polymers, Inc., Huntsville, AL). This monomeric form of OPG, which lacks the native heparin-binding domain, binds to OPGL/RANKL and inhibits osteoclast differentiation and activity in vitro (data not shown). The dose of OPG was chosen based on its maximal inhibition of osteoclast activity in young growing mice and in rats (data not shown). Sham-operated rats were treated with PBS. All treatments were by sc injection, 3 times/wk. Blood was obtained from the tail vein at baseline (before treatment) and then monthly for serum chemistry analysis. Serum calcium and serum creatinine were measured with a Hitachi 717 Automatic Chemistry Analyzer (Roche, Indianapolis, IN). Serum osteocalcin was measured in duplicate with a rat osteocalcin immunoradiometric assay kit (Immunotopics, San Clemente, CA). DEXA scanning was performed monthly for the 5.5-month treatment period. At the end of the study all rats were killed by CO2 inhalation. The fifth lumbar vertebrae (L5) and one tibia were processed for histology. One femur and the third lumbar vertebrae (L3) were wrapped in saline-soaked gauze and stored at −20 C for mechanical testing. All studies were performed in accordance with the policies of the institutional animal care and use committee of Amgen, Inc.
BMD by DEXA
BMD was measured monthly in anesthetized rats (87 mg/kg ketamine and 13 mg/kg xylazine) starting at the beginning of the treatment period. BMD was determined using DEXA (QDR 4500a, Hologic, Inc.). Small animal software (Hologic, Inc.) was used to obtain BMD in the lumbar vertebrae (L1–L5) and in the femur/tibia. The femur/tibia site consisted of the proximal half of the tibia and the entire femur.
Histomorphometry
The tibia and the lumbar vertebrae were decalcified in formic acid, embedded in paraffin, and longitudinally sectioned. Histomorphometric analyses were made by tracing the section image onto a digitizing platen with the aid of a camera lucida attachment on the microscope and Osteomeasure (Osteometrics, Inc., Decatur, GA) bone analysis software. For each section, 10–15 fields of cancellous bone were measured at ×10–20 magnification. The tibial analysis was performed in the proximal metaphysis starting adjacent to the epiphyseal growth plate, an area that encompassed 1.32 mm2 of the section. In the second lumbar vertebra, a 1.54-mm2 area was analyzed in the center of the bone from growth plate to growth plate. To reveal osteoclasts, sections were stained for immunoreactivity to cathepsin K, an osteoclast marker (42). Cathepsin K staining was accomplished with a biotinylated rabbit polyclonal antibody to cathepsin K, and sections were counterstained with hematoxylin. Cancellous bone volume was assessed as a percentage of the total bone tissue volume (BV/TV), and the length of the perimeter of cancellous bone surfaces was measured. Osteoblast perimeter was determined by scoring osteoblasts in direct contact with cancellous bone surfaces. Osteoclast perimeter was determined as the perimeter of multinucleated cathepsin K-stained osteoclasts in direct contact with cancellous bone surfaces.
Mechanical testing
A Material Testing System (model 5501R, Instron Corp., Canton, MA) was used to perform mechanical testing in the femur and the third lumbar vertebral body (L3). The load and extension (deformation) curves were collected with the accompanied software (Merlin II, Instron Corp.). All tests were conducted using a 5-kN load cell at a constant loading rate of 6 mm/min, and data were collected every 6 msec. A compression test was used to determine the mechanical properties of L3, as previously described (5). Briefly, the L3 body was separated from the ephiphyseal ends, the posterior pedicle, and the spinous process using a low speed saw. An electronic caliper was used to determine the dimensions of the L3 body. The L3 body was then compressed to failure. A three-point bending test was used to determine the mechanical properties of the femoral midshaft. The moment of inertia along the load axis was determined using a pQCT scan of the femur and accompanying software (XCT-RM, Stratech, Norland Corp., Fort Atkinson, WI). The diameter of the loading axis was measured with an electronic caliper. The midshaft of the femur was then subjected to three-point bending to failure with a support span of 14 mm at the bottom and load applied at the midpoint of the posterior aspect of the femur. A cantilever compression test was used to determine the mechanical properties of the femoral neck, as previously described (4). The proximal end of the femur was anchored in a hole made in an aluminum block with a notch that holds the greater trochanter in place. The femoral neck was compressed to failure, perpendicular to the shaft. For each skeletal site measured, the maximal load and stiffness were obtained directly from the load and extension (deformation) curve. Ultimate strength and elastic modulus were calculated using the deformation curve and caliper measurements of cross-sectional area (CSA). CSA was calculated based on the formula CSA = π × a × b, where a is the average dorsal to ventral diameter, and b is the side to side diameter.
Statistical analysis
All statistical analyses were performed by one-way ANOVA using an α value of 0.05. For DEXA data, groups were compared at each time point. Where significant overall differences were observed by one-way ANOVA, the Tukey Kramer test was applied for the comparison of multiple pairs. The comparisons reported include each group vs. the vehicle-treated OVX groups, as well as the PTH-treated OVX group vs. the PTH- plus OPG-treated OVX group. Analyses were performed using SAS software version 6.0 (SAS Institute, Inc., Cary, NC). All data are expressed as the mean ± sem.
Results
The long duration of the study resulted in the natural deaths of several animals. A higher proportion of sham-operated animals died compared with OVX rats, whereas the deaths within the OVX groups were similarly distributed. As these rats were approximately 2 yr of age at the end of the study, these results were not unexpected (Table 1). The effectiveness of ovariectomy was indicated by uterine atrophy and a lack of ovarian tissue compared with shams (data not shown). Combining all treatment groups, OVX rats had a final mean body weight of 409 ± 6 vs. 317 ± 17 g for shams (P < 0.05). There was no significant effect of any of the treatments on body weight in the OVX rats (Table 1). Serum calcium levels were not significantly different between groups at the end of the study (Table 2). At baseline, PTH-treated OVX rats had slightly lower calcium levels compared with vehicle-treated OVX rats, which may be attributed to chance randomization. OPG alone had no significant effect on serum calcium after 1, 2, 4, or 5 months of treatment, but a transient increase in serum calcium was observed at 3 months compared with vehicle-treated OVX rats (P < 0.05). Treatment with PTH alone caused significant hypercalcemia after 1, 3, and 4 months of treatment (P < 0.05), and cotreatment with OPG significantly inhibited PTH-induced hypercalcemia at 3 and 4 months (P < 0.05). Serum creatinine levels were essentially similar in all groups throughout the study, although creatinine was significantly elevated in shams at 2 months and in PTH-treated OVX rats at 4 months compared with vehicle-treated OVX rats (data not shown).
Group sizes, body weights, and raw bone mineral density (BMD) data
| Group | n | Final BW (g) | Raw BMD (LV) start/end of Rx (mg/cm3) | Raw BMD (Fem/Tib) start/end of Rx (mg/cm3) | |
|---|---|---|---|---|---|
| Pre | Post | ||||
| Sham | 13 | 8 | 317 ± 20 | 0.270 ± 0.007 | 0.324 ± 0.005 |
| 0.259 ± 0.004 | 0.319 ± 0.005 | ||||
| OVX+ PBS | 14 | 12 | 392 ± 61a | 0.217 ± 0.004a | 0.294 ± 0.004a |
| 0.203 ± 0.006a | 0.292 ± 0.005 | ||||
| OVX OPG | 14 | 12 | 429 ± 48a | 0.215 ± 0.004a | 0.293 ± 0.004a |
| 0.221 ± 0.004a | 0.306 ± 0.003 | ||||
| OVX+ PTH | 14 | 13 | 407 ± 62a | 0.218 ± 0.003a | 0.295 ± 0.003a |
| 0.260 ± 0.003 | 0.352 ± 0.003a | ||||
| OVX+ OPG+ PTH | 14 | 11 | 407 ± 40a | 0.212 ± 0.003a | 0.288 ± 0.003a |
| 0.275 ± 0.003 | 0.358 ± 0.004a | ||||
| All OVX | 56 | 48 | 409 ± 54a | 0.216 ± 0.003a | 0.293 ± 0.003a |
| Group | n | Final BW (g) | Raw BMD (LV) start/end of Rx (mg/cm3) | Raw BMD (Fem/Tib) start/end of Rx (mg/cm3) | |
|---|---|---|---|---|---|
| Pre | Post | ||||
| Sham | 13 | 8 | 317 ± 20 | 0.270 ± 0.007 | 0.324 ± 0.005 |
| 0.259 ± 0.004 | 0.319 ± 0.005 | ||||
| OVX+ PBS | 14 | 12 | 392 ± 61a | 0.217 ± 0.004a | 0.294 ± 0.004a |
| 0.203 ± 0.006a | 0.292 ± 0.005 | ||||
| OVX OPG | 14 | 12 | 429 ± 48a | 0.215 ± 0.004a | 0.293 ± 0.004a |
| 0.221 ± 0.004a | 0.306 ± 0.003 | ||||
| OVX+ PTH | 14 | 13 | 407 ± 62a | 0.218 ± 0.003a | 0.295 ± 0.003a |
| 0.260 ± 0.003 | 0.352 ± 0.003a | ||||
| OVX+ OPG+ PTH | 14 | 11 | 407 ± 40a | 0.212 ± 0.003a | 0.288 ± 0.003a |
| 0.275 ± 0.003 | 0.358 ± 0.004a | ||||
| All OVX | 56 | 48 | 409 ± 54a | 0.216 ± 0.003a | 0.293 ± 0.003a |
BMD values represent measurements taken at the beginning (not underlined) and end (underlined) of a 5.5-month treatment (Rx) period.
Significantly different from sham group, by two-way ANOVA (P < 0.05).
Group sizes, body weights, and raw bone mineral density (BMD) data
| Group | n | Final BW (g) | Raw BMD (LV) start/end of Rx (mg/cm3) | Raw BMD (Fem/Tib) start/end of Rx (mg/cm3) | |
|---|---|---|---|---|---|
| Pre | Post | ||||
| Sham | 13 | 8 | 317 ± 20 | 0.270 ± 0.007 | 0.324 ± 0.005 |
| 0.259 ± 0.004 | 0.319 ± 0.005 | ||||
| OVX+ PBS | 14 | 12 | 392 ± 61a | 0.217 ± 0.004a | 0.294 ± 0.004a |
| 0.203 ± 0.006a | 0.292 ± 0.005 | ||||
| OVX OPG | 14 | 12 | 429 ± 48a | 0.215 ± 0.004a | 0.293 ± 0.004a |
| 0.221 ± 0.004a | 0.306 ± 0.003 | ||||
| OVX+ PTH | 14 | 13 | 407 ± 62a | 0.218 ± 0.003a | 0.295 ± 0.003a |
| 0.260 ± 0.003 | 0.352 ± 0.003a | ||||
| OVX+ OPG+ PTH | 14 | 11 | 407 ± 40a | 0.212 ± 0.003a | 0.288 ± 0.003a |
| 0.275 ± 0.003 | 0.358 ± 0.004a | ||||
| All OVX | 56 | 48 | 409 ± 54a | 0.216 ± 0.003a | 0.293 ± 0.003a |
| Group | n | Final BW (g) | Raw BMD (LV) start/end of Rx (mg/cm3) | Raw BMD (Fem/Tib) start/end of Rx (mg/cm3) | |
|---|---|---|---|---|---|
| Pre | Post | ||||
| Sham | 13 | 8 | 317 ± 20 | 0.270 ± 0.007 | 0.324 ± 0.005 |
| 0.259 ± 0.004 | 0.319 ± 0.005 | ||||
| OVX+ PBS | 14 | 12 | 392 ± 61a | 0.217 ± 0.004a | 0.294 ± 0.004a |
| 0.203 ± 0.006a | 0.292 ± 0.005 | ||||
| OVX OPG | 14 | 12 | 429 ± 48a | 0.215 ± 0.004a | 0.293 ± 0.004a |
| 0.221 ± 0.004a | 0.306 ± 0.003 | ||||
| OVX+ PTH | 14 | 13 | 407 ± 62a | 0.218 ± 0.003a | 0.295 ± 0.003a |
| 0.260 ± 0.003 | 0.352 ± 0.003a | ||||
| OVX+ OPG+ PTH | 14 | 11 | 407 ± 40a | 0.212 ± 0.003a | 0.288 ± 0.003a |
| 0.275 ± 0.003 | 0.358 ± 0.004a | ||||
| All OVX | 56 | 48 | 409 ± 54a | 0.216 ± 0.003a | 0.293 ± 0.003a |
BMD values represent measurements taken at the beginning (not underlined) and end (underlined) of a 5.5-month treatment (Rx) period.
Significantly different from sham group, by two-way ANOVA (P < 0.05).
Serum calcium data
| Baseline | 1 month | 2 months | 3 months | 4 months | 5.5 months | |
|---|---|---|---|---|---|---|
| Sham | 9.29 ± 0.07 | 9.65 ± 0.10 | 10.14 ± 0.11 | 10.19 ± 0.15 | 10.19 ± 0.14 | 10.64 ± 0.10 |
| OVX-Veh | 9.41 ± 0.09 | 9.77 ± 0.13 | 9.85 ± 0.09 | 10.28 ± 0.19 | 10.21 ± 0.14 | 10.96 ± 0.15 |
| OVX-OPG | 9.38 ± 0.03 | 9.80 ± 0.06 | 9.84 ± 0.11 | 11.02 ± 0.15a | 9.98 ± 0.10 | 11.14 ± 0.25 |
| OVX-PTH | 9.19 ± 0.05a | 10.1 ± 0.10a | 10.04 ± 0.09 | 11.31 ± 0.24a | 11.96 ± 0.16a | 11.05 ± 0.07 |
| OVX-OPG+ PTH | 9.32 ± 0.07 | 10.08 ± 0.08 | 10.28 ± 0.21 | 10.28 ± 0.19b | 11.27 ± 0.23ab | 10.75 ± 0.10 |
| Baseline | 1 month | 2 months | 3 months | 4 months | 5.5 months | |
|---|---|---|---|---|---|---|
| Sham | 9.29 ± 0.07 | 9.65 ± 0.10 | 10.14 ± 0.11 | 10.19 ± 0.15 | 10.19 ± 0.14 | 10.64 ± 0.10 |
| OVX-Veh | 9.41 ± 0.09 | 9.77 ± 0.13 | 9.85 ± 0.09 | 10.28 ± 0.19 | 10.21 ± 0.14 | 10.96 ± 0.15 |
| OVX-OPG | 9.38 ± 0.03 | 9.80 ± 0.06 | 9.84 ± 0.11 | 11.02 ± 0.15a | 9.98 ± 0.10 | 11.14 ± 0.25 |
| OVX-PTH | 9.19 ± 0.05a | 10.1 ± 0.10a | 10.04 ± 0.09 | 11.31 ± 0.24a | 11.96 ± 0.16a | 11.05 ± 0.07 |
| OVX-OPG+ PTH | 9.32 ± 0.07 | 10.08 ± 0.08 | 10.28 ± 0.21 | 10.28 ± 0.19b | 11.27 ± 0.23ab | 10.75 ± 0.10 |
Blood was drawn from the tail vein on a monthly basis before the injection of treatment agents. Data are expressed as the mean ± sem, and units are in milligrams per dl.
Significant difference from OVX-Veh group (by Dunnett’s test, P < 0.05).
Significant difference from OVX-PTH group (by Dunnett’s test, P < 0.05).
Serum calcium data
| Baseline | 1 month | 2 months | 3 months | 4 months | 5.5 months | |
|---|---|---|---|---|---|---|
| Sham | 9.29 ± 0.07 | 9.65 ± 0.10 | 10.14 ± 0.11 | 10.19 ± 0.15 | 10.19 ± 0.14 | 10.64 ± 0.10 |
| OVX-Veh | 9.41 ± 0.09 | 9.77 ± 0.13 | 9.85 ± 0.09 | 10.28 ± 0.19 | 10.21 ± 0.14 | 10.96 ± 0.15 |
| OVX-OPG | 9.38 ± 0.03 | 9.80 ± 0.06 | 9.84 ± 0.11 | 11.02 ± 0.15a | 9.98 ± 0.10 | 11.14 ± 0.25 |
| OVX-PTH | 9.19 ± 0.05a | 10.1 ± 0.10a | 10.04 ± 0.09 | 11.31 ± 0.24a | 11.96 ± 0.16a | 11.05 ± 0.07 |
| OVX-OPG+ PTH | 9.32 ± 0.07 | 10.08 ± 0.08 | 10.28 ± 0.21 | 10.28 ± 0.19b | 11.27 ± 0.23ab | 10.75 ± 0.10 |
| Baseline | 1 month | 2 months | 3 months | 4 months | 5.5 months | |
|---|---|---|---|---|---|---|
| Sham | 9.29 ± 0.07 | 9.65 ± 0.10 | 10.14 ± 0.11 | 10.19 ± 0.15 | 10.19 ± 0.14 | 10.64 ± 0.10 |
| OVX-Veh | 9.41 ± 0.09 | 9.77 ± 0.13 | 9.85 ± 0.09 | 10.28 ± 0.19 | 10.21 ± 0.14 | 10.96 ± 0.15 |
| OVX-OPG | 9.38 ± 0.03 | 9.80 ± 0.06 | 9.84 ± 0.11 | 11.02 ± 0.15a | 9.98 ± 0.10 | 11.14 ± 0.25 |
| OVX-PTH | 9.19 ± 0.05a | 10.1 ± 0.10a | 10.04 ± 0.09 | 11.31 ± 0.24a | 11.96 ± 0.16a | 11.05 ± 0.07 |
| OVX-OPG+ PTH | 9.32 ± 0.07 | 10.08 ± 0.08 | 10.28 ± 0.21 | 10.28 ± 0.19b | 11.27 ± 0.23ab | 10.75 ± 0.10 |
Blood was drawn from the tail vein on a monthly basis before the injection of treatment agents. Data are expressed as the mean ± sem, and units are in milligrams per dl.
Significant difference from OVX-Veh group (by Dunnett’s test, P < 0.05).
Significant difference from OVX-PTH group (by Dunnett’s test, P < 0.05).
Serum osteocalcin, a marker of osteoblast differentiation, was similar in vehicle-treated shams and OVX rats throughout the treatment period (Table 3). PTH alone caused significant increases in serum osteocalcin at all time points compared with vehicle-treated OVX rats (P < 0.05). The maximum induction of osteocalcin, which occurred after 2 months of PTH treatment, was 78% greater than that in vehicle-treated OVX rats. OPG alone caused small, but significant, decreases in serum osteocalcin (9%–17%) at 3, 4, and 5.5 months of treatment (P < 0.05). OPG also partially attenuated the PTH-associated increases in osteocalcin observed after 2, 3, 4, and 5.5 months of treatment. In rats treated with OPG plus PTH, osteocalcin levels were significantly greater than in vehicle-treated OVX rats after 1 and 2 months of treatment (P < 0.05), after which time osteocalcin returned to levels similar to those in vehicle-treated OVX rats (Table 3).
Serum osteocalcin data
| Baseline | 1 month | 2 months | 3 months | 4 months | 5.5 months | |
|---|---|---|---|---|---|---|
| Sham | 12.08 ± 0.22 | 13.90 ± 0.28 | 16.35 ± 0.64 | 12.10 ± 0.43 | 13.74 ± 0.32 | 10.36 ± 0.09 |
| OVX-Veh | 11.94 ± 0.07 | 14.38 ± 0.48 | 16.31 ± 0.57 | 12.57 ± 0.40 | 14.13 ± 0.44 | 10.43 ± 0.16 |
| OVX-OPG | 12.25 ± 0.10 | 12.33 ± 0.26 | 14.76 ± 0.33 | 10.43 ± 0.13a | 12.07 ± 0.11a | 9.54 ± 0.07a |
| OVX-PTH | 12.19 ± 0.09 | 24.29 ± 0.54a | 28.96 ± 1.88a | 13.99 ± 0.29a | 18.16 ± 0.82a | 12.26 ± 0.22a |
| OVX-OPG+ PTH | 11.93 ± 0.13 | 22.72 ± 2.44a | 23.02 ± 1.36ab | 12.06 ± 0.24b | 13.19 ± 0.27b | 10.53 ± 0.20b |
| Baseline | 1 month | 2 months | 3 months | 4 months | 5.5 months | |
|---|---|---|---|---|---|---|
| Sham | 12.08 ± 0.22 | 13.90 ± 0.28 | 16.35 ± 0.64 | 12.10 ± 0.43 | 13.74 ± 0.32 | 10.36 ± 0.09 |
| OVX-Veh | 11.94 ± 0.07 | 14.38 ± 0.48 | 16.31 ± 0.57 | 12.57 ± 0.40 | 14.13 ± 0.44 | 10.43 ± 0.16 |
| OVX-OPG | 12.25 ± 0.10 | 12.33 ± 0.26 | 14.76 ± 0.33 | 10.43 ± 0.13a | 12.07 ± 0.11a | 9.54 ± 0.07a |
| OVX-PTH | 12.19 ± 0.09 | 24.29 ± 0.54a | 28.96 ± 1.88a | 13.99 ± 0.29a | 18.16 ± 0.82a | 12.26 ± 0.22a |
| OVX-OPG+ PTH | 11.93 ± 0.13 | 22.72 ± 2.44a | 23.02 ± 1.36ab | 12.06 ± 0.24b | 13.19 ± 0.27b | 10.53 ± 0.20b |
Blood was drawn from the tail vein on a monthly basis before the injection of treatment agents. Data are expressed as the mean ± sem, and units are in anograms per ml.
Significant difference from OVX-Veh group (by Dunnett’s test, P < 0.05).
Significant difference from OVX-PTH group (by Dunnett’s test, P < 0.05).
Serum osteocalcin data
| Baseline | 1 month | 2 months | 3 months | 4 months | 5.5 months | |
|---|---|---|---|---|---|---|
| Sham | 12.08 ± 0.22 | 13.90 ± 0.28 | 16.35 ± 0.64 | 12.10 ± 0.43 | 13.74 ± 0.32 | 10.36 ± 0.09 |
| OVX-Veh | 11.94 ± 0.07 | 14.38 ± 0.48 | 16.31 ± 0.57 | 12.57 ± 0.40 | 14.13 ± 0.44 | 10.43 ± 0.16 |
| OVX-OPG | 12.25 ± 0.10 | 12.33 ± 0.26 | 14.76 ± 0.33 | 10.43 ± 0.13a | 12.07 ± 0.11a | 9.54 ± 0.07a |
| OVX-PTH | 12.19 ± 0.09 | 24.29 ± 0.54a | 28.96 ± 1.88a | 13.99 ± 0.29a | 18.16 ± 0.82a | 12.26 ± 0.22a |
| OVX-OPG+ PTH | 11.93 ± 0.13 | 22.72 ± 2.44a | 23.02 ± 1.36ab | 12.06 ± 0.24b | 13.19 ± 0.27b | 10.53 ± 0.20b |
| Baseline | 1 month | 2 months | 3 months | 4 months | 5.5 months | |
|---|---|---|---|---|---|---|
| Sham | 12.08 ± 0.22 | 13.90 ± 0.28 | 16.35 ± 0.64 | 12.10 ± 0.43 | 13.74 ± 0.32 | 10.36 ± 0.09 |
| OVX-Veh | 11.94 ± 0.07 | 14.38 ± 0.48 | 16.31 ± 0.57 | 12.57 ± 0.40 | 14.13 ± 0.44 | 10.43 ± 0.16 |
| OVX-OPG | 12.25 ± 0.10 | 12.33 ± 0.26 | 14.76 ± 0.33 | 10.43 ± 0.13a | 12.07 ± 0.11a | 9.54 ± 0.07a |
| OVX-PTH | 12.19 ± 0.09 | 24.29 ± 0.54a | 28.96 ± 1.88a | 13.99 ± 0.29a | 18.16 ± 0.82a | 12.26 ± 0.22a |
| OVX-OPG+ PTH | 11.93 ± 0.13 | 22.72 ± 2.44a | 23.02 ± 1.36ab | 12.06 ± 0.24b | 13.19 ± 0.27b | 10.53 ± 0.20b |
Blood was drawn from the tail vein on a monthly basis before the injection of treatment agents. Data are expressed as the mean ± sem, and units are in anograms per ml.
Significant difference from OVX-Veh group (by Dunnett’s test, P < 0.05).
Significant difference from OVX-PTH group (by Dunnett’s test, P < 0.05).
Fifteen months after ovariectomy (OVX), and before the treatment phase, DEXA analysis revealed significant decreases in the raw BMD values for both the femur/tibia and the lumbar vertebrae compared with shams (P < 0.05; Table 1). OPG alone caused a gradual increase in lumbar vertebral BMD compared with vehicle-treated OVX controls that became significant after 4 months of treatment. PTH treatment, alone or in combination with OPG, rapidly increased lumbar vertebral BMD, with a significant increase evident after 1 month. After 3 months of treatment with OPG plus PTH, the gain in lumbar vertebral BMD was significantly greater than that in all other groups, including PTH alone (Fig. 1A). The raw BMD data for the lumbar vertebra revealed that by the end of the treatment period, PTH and PTH plus OPG had restored BMD to levels found in shams (Table 1). OPG alone caused a significant increase in raw BMD in OVX rats and restored 32% of the BMD relative to that in shams (Table 1).
Effect of OPG, PTH, or OPG plus PTH on BMD. DEXA scans of the lumbar vertebrae (A) and the femur/tibia (B) were performed as described in Materials and Methods. All data are expressed as the percent change from baseline (BL) DEXA (±sem). Baseline DEXA values were obtained 15 months after OVX and just before treatment (raw baseline DEXA data are provided in Table 1). ▪, Sham-operated PBS-treated controls; •, vehicle-treated OVX; ▴, OPG-treated OVX; ♦, PTH-treated OVX; □, OPG- plus PTH-treated OVX. *, Significant difference from vehicle-treated OVX rats; #, significant difference from PTH-treated OVX rats (by Tukey-Kramer test, P < 0.05).
OPG alone significantly increased BMD in the femur/tibia within 3 months, compared with PBS/OVX controls. PTH treatment, alone or in combination with OPG, increased femur/tibia BMD within 1 month compared with PBS/OVX controls. Within 2 months of treatment, OPG plus PTH caused a significantly larger increase in BMD of the femur/tibia compared with all other groups, including PTH alone (Fig. 1B). The effects of combination therapy on BMD in both the lumbar vertebrae and the femur/tibia appeared to be a purely additive effect of each agent. The raw BMD data for the femur/tibia indicated that at the end of the study, both PTH and PTH plus OPG increased BMD to levels significantly greater than those in shams, whereas OPG alone restored about 50% of the BMD relative to that in shams (Table 1).
These changes in BMD were concordant with histomorphometric analysis of cancellous bone volume (BV/TV). At the end of the study, BV/TV in the lumbar vertebrae of PBS-treated OVX rats was 45% reduced compared with that in shams (P < 0.05). PTH alone, but not OPG alone, significantly increased BV/TV in the lumbar vertebrae of OVX rats. OPG plus PTH caused a slightly greater increase in BV/TV compared with PTH alone, and this combination treatment restored BV/TV to levels similar to those in shams (Fig. 2A). In the proximal tibial metaphysis, OVX caused an 80% decrease in BV/TV compared with shams (P < 0.05). OPG treatment of OVX rats caused a nonsignificant 75% increase in tibial BV/TV, whereas PTH cause a significant 200% increase in BV/TV (P < 0.05). The combination of OPG plus PTH caused a significantly greater increase in tibial BV/TV compared with PTH treatment alone (P < 0.05; Fig. 2D).
Effects of OPG, PTH, or OPG plus PTH on bone histomorphometry. Decalcified, cathepsin K-stained sections of the fifth lumbar vertebra (L5; A–C) and the proximal tibial metaphysis (D–F) were analyzed with an Osteomeasure workstation. A and D, Percentage of total bone volume occupied by trabecular bone (BV/TV); B and E, percentage of trabecular bone perimeter occupied by osteoclasts (OcPm/BPm); C and F, percentage of trabecular bone perimeter occupied by osteoblasts (ObPm/BPm). Data are expressed as the mean ± sem. *, Significant difference from vehicle-treated OVX rats; #, significant difference from PTH-treated OVX rats (by Tukey-Kramer test, P < 0.05).
Histomorphometry also revealed interesting differences in the cellular response to therapies. At the end of the 5.5-month treatment period, OVX caused a nonsignificant 70% increase in vertebral osteoclast surface compared with shams. PTH treatment further increased osteoclast surface in the lumbar vertebra by a significant 46% compared with that in vehicle-treated OVX rats. OPG treatment reduced osteoclast surface in OVX rats by more than 98%, independent of PTH cotreatment (P < 0.05; Fig. 2B). Osteoblast surface in the lumbar vertebrae was not affected by OVX or any treatment (Fig. 2C). A different cellular response was observed in the proximal tibial metaphysis. Ovariectomy did not have a significant effect on osteoclast surface in the tibia at the time of death. OPG alone reduced osteoclast surface in OVX rats by 65% (P < 0.05), whereas PTH alone had no effect on osteoclast surface. Treatment with OPG plus PTH caused a nonsignificant 41% reduction in tibial osteoclast surface compared with PTH alone (Fig. 2E). PTH treatment caused a 3-fold increase in osteoblast surface compared with that in vehicle-treated OVX rats (P < 0.05), and the addition of OPG had no significant effect on osteoblast surface (Fig. 2F).
Mechanical compression testing of the third lumbar vertebra (L3) indicated that OVX alone caused significant reductions in stiffness, maximum load, ultimate strength, and elastic modulus compared with those in shams (P < 0.05; Fig. 3). OPG alone had no significant effect on these parameters. PTH treatment alone significantly increased maximum load and ultimate strength compared with PBS/OVX rats (P < 0.05). OPG plus PTH also increased maximum load and ultimate strength significantly compared with vehicle-treated OVX rats (P < 0.05). OPG plus PTH also significantly increased stiffness and elastic modulus, and these increases were significantly greater than those observed with PTH alone (P < 0.05; Fig. 3).
Effects of OPG, PTH, or OPG plus PTH on the mechanical strength of the third lumbar vertebra (L3). L3 was compressed to failure to provide data on stiffness (A), maximum load (B), ultimate strength (C), and elastic modulus (D). These parameters are defined in Materials and Methods. Data are expressed as the mean ± sem. *, Significant difference from vehicle-treated OVX; #, significant difference from PTH-treated OVX (by Tukey-Kramer test, P < 0.05).
In the femoral midshaft, three-point bending tests failed to show any significant effect of OVX on stiffness, maximum load, ultimate strength, or elastic modulus (Fig. 4). OPG alone caused a significant increase in maximum load compared with that in vehicle-treated OVX rats (P < 0.05; Fig. 4). PTH alone caused significant increases in stiffness, maximum load, and ultimate strength (P < 0.05; Fig. 4). OPG plus PTH was the only treatment that significantly increased elastic modulus in the femur (P < 0.05; Fig. 4). Cantilever compression testing of the femoral neck did not show any OVX-induced decreases in maximum load or stiffness (Fig. 5). OPG alone had no significant effect on these end points. PTH treatment, alone or in combination with OPG, caused significant increases in maximum load and stiffness compared with vehicle-treated OVX rats (P < 0.05; Fig. 4).
Effects of OPG, PTH, or OPG plus PTH on the mechanical strength of the femoral diaphysis. The diaphysis was subjected to three-point bending to failure, which provided data on stiffness (A), maximum load (B), ultimate strength (C), and elastic modulus (D). These parameters are defined in Materials and Methods. Data are expressed as the mean ± sem. *, Significant difference from vehicle-treated OVX; #, significant difference from PTH-treated OVX (by Tukey-Kramer test, P < 0.05).
Effects of OPG, PTH, or OPG plus PTH on the mechanical strength of the femoral neck. The femoral neck was subjected to cantilever compression to failure, which provided data on maximum load (A) and stiffness (B). These parameters are defined in Materials and Methods. Data are expressed as the mean ± sem. *, Significant difference from vehicle-treated OVX; #, significant difference from PTH-treated OVX (by Tukey-Kramer test, P < 0.05).
Discussion
PTH has an anabolic effect on the skeleton when given intermittently, a phenomenon that was first recognized in the 1930s. The ability to stimulate bone formation makes PTH an attractive alternative to antiresorptive therapy in patients with severe osteoporosis. Intermittent PTH also stimulates bone resorption (43) and can cause hypercalcemia (8), so it is conceivable that the effective inhibition of PTH-induced osteoclast activity with antiresorptives could augment the effects of PTH on bone mass while preventing PTH-associated hypercalcemia. We have explored whether OPG, a novel antiresorptive agent (reviewed in Ref. 44), could enhance the bone effects of PTH while inhibiting PTH-related hypercalcemia.
The present study adds to the growing body of literature indicating that the addition of PTH to antiresorptive therapy is clearly superior to antiresorptive therapy alone (3, 4, 6, 13, 13, 19). However, it has been much more difficult to demonstrate that antiresorptives can add significantly to the robust anabolic actions of PTH. Indirect evidence supporting a benefit of antiresorptive cotherapy was provided by a recent clinical trial of intermittent PTH therapy. Daily PTH-(1–34) treatment of postmenopausal patients receiving hormone replacement therapy (HRT) caused a 20.6% increase in lumbar BMD at 1 yr compared with placebo (20). The lack of a PTH monotherapy group in that study prevents any conclusions regarding the benefit of adding the antiresorptive HRT regimen. However, it may be noteworthy that in separate clinical trials, PTH-(1–84) monotherapy caused a 6.9% increase in lumbar BMD after 1 yr (45), whereas PTH-(1–34) monotherapy increased lumbar BMD by up to 13.7% after 17 months (8). The larger gain in BMD observed with PTH and HRT (20) compared with PTH monotherapy suggests an advantage of combination therapy. However, variations in the study designs and the use of different PTH fragments hinder conclusions on the benefits of added HRT. Clearly, randomized placebo-controlled studies will be required to directly compare the effects of PTH alone vs. those of PTH and antiresorptives.
The present preclinical study is among the first to directly demonstrate a significant additive effect of PTH and antiresorptive therapy on bone volume, BMD, and parameters of mechanical strength compared with PTH alone. The beneficial effects of combination therapy in this study might be attributed to the novel antiresorptive employed or to the nature of the study design. The current study design allowed for 15 months of OVX-induced bone loss, followed by 5.5 months of therapy, which are among the longest durations reported. However, the additive effects of OPG plus PTH observed here do not appear to be a direct function of animal age or the severity of osteopenia, as other long-term aged OVX rat studies have failed to show a benefit of combination therapy over PTH treatment alone. For example, 12 months of OVX-induced bone loss resulted in similarly large deficits in bone density, but there was no apparent benefit of adding IGF-I to intermittent PTH therapy (46). Other rat studies compared the effects of PTH treatment, with and without antiresorptives, on bone mass and mechanical strength starting 12 months after OVX. The combination therapies, which added E, risedronate, or calcitonin to PTH, were no more effective than PTH alone (7, 12, 13).
The rationale for using severely osteopenic aged OVX rats in the current study was based in part on the assumption that in the clinic, the severely osteoporotic elderly patient has the greatest clinical need for effective anabolic therapy to rapidly restore bone mass. Also, acute OVX models, particularly in young rats, may show therapeutic effects that are not realized in more clinically relevant, long-term, aged OVX rat models. For example, a TNF-binding protein was demonstrated to block OVX-induced bone loss in young rats, but not in old rats (47, 48). In another acute OVX study, E plus PTH increased bone mass better than did PTH alone, but only when combination therapy was initiated within 1 wk after OVX (15). There was no benefit of adding E to PTH when treatment was initiated 3 or 5 wk after OVX. In the clinical setting it is not clear that anabolic agents, alone or in combination with antiresorptives, are appropriate or necessary for the modest osteopenia that is typically associated with the early postmenopausal period. In addition to their clinical relevance, aged OVX rats are suitable for combination therapy studies because of their very slow skeletal growth rate. Treatment of young rapidly growing rats with antiresorptives produces growth-related increases in bone mass (35) that cannot be realized in skeletally mature humans. Severely osteopenic aged OVX rats are clearly responsive to the anabolic effects of PTH (Refs. 7, 12, 46 , and 49 and current study), and the effects of combination therapy in these animals can be interpreted with minimal confounding growth effects from the antiresorptive. This idea is highlighted in the current study, where OPG treatment was associated with a near-total lack of osteoclasts, and yet the BMD increases associated with OPG treatment alone were a modest 5%. Adding PTH to OPG led to 20–25% increases in BMD compared with vehicle-treated OVX rats. These observations highlight the challenge of reversing severe osteopenia in aged OVX rats with antiresorptives alone; even a 98% reduction in osteoclast number does not reverse osteopenia in these animals unless accompanied by an anabolic stimulus such as PTH.
Another possible explanation for the beneficial effects of combination therapy in the present study is the nature of the antiresorptive, OPG. Like E, OPG appears to play an important physiological role in the regulation of bone remodeling in mice (35). It was recently demonstrated that E treatment of human osteoblasts increased OPG mRNA and protein (50), suggesting that the bone-protective effects of E may be mediated at least in part through increased OPG production. Providing recombinant OPG to the E-depleted skeleton may directly increase the ratio of OPG to OPGL/RANKL, thereby decreasing bone resorption. OPG also has several properties that appear to complement the effects of PTH. Intermittent PTH therapy, while having a net anabolic effect on bone, stimulates both bone formation and bone resorption (43). Recent data indicate that PTH stimulates bone resorption indirectly by increasing osteoblast production of OPGL/RANKL and by decreasing OPG (reviewed in Ref. 44). PTH inhibits OPG mRNA expression in cultured osteoblasts (51–55) and increases OPGL/RANKL mRNA (33). A single injection of PTH-(1–38) into rats was recently shown to rapidly decrease skeletal levels of OPG mRNA (54). These changes could decrease the ratio of OPG to OPGL/RANKL and thereby promote bone resorption. The administration of recombinant OPG might directly reverse this ratio and thereby prevent PTH-induced bone resorption. In support of this idea, recombinant OPG completely blocks PTH-induced resorption in bone organ cultures (56) and also blocks the calcemic response of mice to PTH (57). The vertebral histomorphometry data from the present study also suggest that OPG can block the PTH-induced bone resorption. In the vertebrae, PTH alone caused a significant increase in osteoclast surface, and cotreatment with OPG reduced osteoclast surface by 95%. OPG cotreatment had no effect on osteoblast surface in the vertebrae, suggesting a relatively selective inhibition of osteoclasts with OPG.
Despite the profound reduction in osteoclast surface observed at the end of the study, serum calcium was statistically elevated in OPG-treated rats at the 3 month point compared with that in vehicle-treated OVX rats. The significance and mechanism of this apparent calcemic response are currently unknown. Calcemic responses to OPG have not been previously reported in mice or rats, and in the current study serum calcium was normal in OPG-treated rats at all other time points. Furthermore, OPG cotherapy significantly inhibited the hypercalcemic responses associated with PTH treatment, so the paradoxical rise in serum calcium at a single time point is difficult to reconcile with the pharmacology of OPG.
In the tibia, after 5.5 months of treatment, combination therapy was associated with a more modest OPG-induced suppression of osteoclast surface combined with a PTH-induced increase in osteoblast surface that was unaffected by OPG cotreatment. Serum osteocalcin, a marker of osteoblast differentiation, was significantly elevated at all time points in PTH-treated rats. OPG cotherapy partially inhibited the osteocalcin response to PTH during the latter half of the treatment phase, presumably due to a coupled response of osteoblasts to the OPG-related decrease in osteoclast numbers. These responses predict that virtually all of the PTH-mediated osteoclast activity was eradicated by OPG, whereas OPG permitted some, but not all, of the PTH-mediated osteoblast activity. Validating this hypothesis would require assessment of bone formation and apposition rates at multiple skeletal sites throughout the treatment period. Regardless of the mechanism(s) involved, the net effect of these histological and biochemical changes was a significantly greater increase in vertebral and femoral BMD, tibial cancellous bone volume, and elastic modulus with OPG plus PTH compared with all other treatments. The histomorphometric data must be interpreted with caution, as the analysis was conducted at a single time point at the end of the study on 2-yr-old rats, 20.5 months after OVX. We have not assessed the temporal changes in bone histomorphometry, and it is likely that the osteoclast and osteoblast responses to OVX and treatments varied according to time and skeletal site.
OPG plus PTH was statistically superior to all other treatments at increasing lumbar vertebral stiffness, maximum load, ultimate strength, and elastic modulus as well as the elastic modulus of the femoral midshaft. For the remaining comparisons, OPG plus PTH was at least as effective as any other treatment, including PTH alone. It is difficult to compare these results to other combination therapy studies, all of which employed shorter periods of OVX-induced bone loss (<15 months) and/or shorter durations of treatment (≤5.5 months). In an OVX rat study of comparable treatment duration (24 wk), the addition of E or calcitonin to PTH treatment failed to improve the mechanical properties of the lumbar vertebrae compared with PTH alone (5). In studies with shorter treatment periods (5–15 wk), the combination of PTH with E or with risedronate did not increase vertebral strength (6) or femoral neck strength (4) compared with the increases observed with PTH alone. In a related study, risedronate (after 5 wk) and E (after 15 wk) actually blocked the beneficial effects of PTH on the strength of the femoral shaft (21). In an OVX rat study with a comparable duration of OVX-induced bone loss (12 months), the addition of E, risedronate, or calcitonin to PTH therapy provided no improvement in vertebral strength compared with PTH treatment alone (13). It remains to be determined whether other antiresorptive agents besides OPG could have additive effects on bone strength in PTH-treated animals under the same conditions as those in the present study. It is possible that dosing regimens different from those used previously would reveal additive effects of other antiresorptives with PTH.
In conclusion, we have demonstrated that PTH and OPG can each increase bone mass, BMD, and parameters mechanical strength in aged OVX rats. PTH as a monotherapy was clearly superior to OPG in these severely osteopenic rats. The combination of OPG plus PTH caused greater increases in BMD, cancellous bone volume, and parameters of mechanical strength compared with PTH alone. OPG also blocked much of the hypercalcemic effects of PTH, which suggests that OPG combination therapy may improve the therapeutic index of PTH by allowing more aggressive dosing of PTH with better control of hypercalcemia. These data provide an important proof of concept that combination therapies using appropriate antiresorptives and anabolic agents may represent a powerful approach to treating or reversing severe osteoporosis in humans.
Acknowledgments
The authors gratefully acknowledge the excellent histology support provided by Diane Duryea, Yan Cheng, Annie Luo, and Darlene Kratavil.
Abbreviations
- BMD
Bone mineral density
- BV/TV
cancellous bone volume assessed as a percentage of the total bone tissue volume
- DEXA
dual energy x-ray absorptiometry
- HRT
hormone replacement therapy
- OPGL
OPG ligand
- OVX
ovariectomized/ovariectomy
