Abstract
We examined the effects of 60 days of co‐treatment of PTH with either OPG or alendronate in oophorectomized mice. Compared with PTH alone, co‐treatment of PTH with either of these two mechanistically distinct anti‐catabolics improved bone volume, mechanical strength, and appendicular and axial mineralization and prolonged the beneficial effect of PTH on BMD.
Introduction: Conflicting evidence exists as to whether the anabolic effect of PTH is inhibited by the action of anti‐catabolics. To examine this issue, we assessed the effects of alendronate and osteoprotegerin (OPG), two anti‐catabolics with different modes of action, on the anabolic activity of PTH(1‐34) in the skeleton of 4‐month‐old oophorectomized mice.
Materials and Methods: Mice treated with vehicle alone (PBS), alendronate alone (100 μg/kg/week), OPG alone (10 mg/kg twice a week), or PTH alone (80 μg/kg/day) were compared with each other and with animals administered PTH plus alendronate or PTH plus OPG. We assessed lumbar spine and femoral BMD at 0, 30, and 60 days. Contact radiography, histology, and histomorphometry, three‐point bending assay of the femur, and serum osteocalcin and TRACP5b assays were performed at 2 months.
Results: Although alendronate and OPG each suppressed bone turnover, at the doses used, this was more profound with OPG. Increases in lumbar spine and femoral BMD and in trabecular bone volume were at least as great with OPG as with alendronate, and mechanical indices of femoral bone strength improved only with OPG. Both produced a plateau in spine and femoral BMD increases by 30 days. Co‐treatment of PTH with each anti‐catabolic produced additive increases in BMD in the femur and supra‐additive increases in the lumbar spine with no plateau effects. Neither anti‐catabolic impeded the PTH‐induced increase in bone volume or the increase in mechanical strength of the femur.
Conclusions: These studies show that the highly potent anti‐catabolic OPG can produce dramatic increases in BMD and bone strength; that the temporal pattern of activity of bone formation and resorption modulators may have major influence on net skeletal accrual; and that, depending on timing, inhibition of osteoclastic activity may markedly augment the anabolic action of PTH.
INTRODUCTION
Anti‐catabolic agents have been the mainstay of treatment of osteoporosis, and nitrogen‐containing bisphosphonates, notably alendronate, are the most frequently used of this class of compounds. In view of the fact that bisphosphonates bind avidly to bone mineral and are released from bone during the resorptive process, the most important action of these agents seems to be a direct effect on mature bone‐resorbing osteoclasts. Nitrogen‐containing bisphosphonates are believed to act primarily by inhibiting farnesyl pyrophosphate (FPP) synthase, an enzyme in the mevalonate pathway that is critical for the signaling function of small GTPases; loss of this function results in loss of bone‐resorptive activity and osteoclast apoptosis.(1) With the discovery of the cytokine RANKL, its receptor RANK, and osteoprotegerin (OPG), a decoy receptor for RANKL, a novel pathway mediating osteoclast biology was discovered.(2,3) Modulators of this pathway are capable of regulating the entire sequence of osteoclast differentiation, activation, and survival and seem highly likely to provide new approaches to clinical disorders in which excess bone resorption occurs.(4)
Currently, PTH is the only bone‐forming or anabolic agent that is approved for treatment of osteoporosis. In view of the widespread use of anti‐catabolics, it is essential to understand the nature of the interaction between these agents and anabolic compounds to assure efficacious use of these materials. Some studies have indicated that bisphosphonates blunt the anabolic actions of PTH,(5–8) and others have shown that bisphosphonates have no effect.(9,10) Furthermore, beneficial effects of OPG in combination with PTH has been reported in ovariectomized rats.(11,12) In this study, we directly compared, for the first time, the effect of OPG, alendronate, and PTH on BMD and mechanical strength and the effect of co‐treatment with alendronate versus co‐treatment with OPG on the bone‐forming action of PTH in oophorectomized mice. We wished to determine whether bone resorption, in this paradigm, is essential to permit PTH to act as a bone‐forming agent and whether the different mechanisms of action of alendronate and OPG on bone resorption would result in discrete effects on the anabolic actions of PTH.
MATERIALS AND METHODS
Mice and materials
Four‐month‐old oophorectomized and sham‐operated C57/Bl6 mice were obtained from Harlan Animal laboratories (Indianapolis, IN, USA). Recombinant human OPG, recombinant human PTH(1‐34), and alendronate were generously provided by Amgen (Amgen, Thousand Oaks, CA, USA). All the compounds were freshly prepared in sterile PBS, and PBS alone was administered to animals as the vehicle control group.
In vivo experiments
All animal experiments were carried out in compliance with and approval by the Institutional Animal Care and Use Committee. Mice were maintained in a virus‐ and parasite‐free barrier facility and exposed to a 12‐h/12‐h light/dark cycle. Four‐month‐old oophorectomized C57/Bl6 mice were assigned to six different groups with 10 mice per group and each received subcutaneously for 60 days either PBS, OPG (10 mg/kg twice a week), alendronate (100 μg/kg/week), PTH (80 μg/kg/day), PTH plus alendronate, or PTH plus OPG. The animals were killed, and blood samples were collected for serum analysis. Tibias, femurs, and vertebrae were also obtained for histology and faxitron analysis. BMD was measured at 0, 1, and 2 months after initiating the experiment.
BMD analysis
Mice were anesthetized with a mixture of ketamine (100 mg/kg) and xylazine (0.1 mg/kg) in PBS and placed prone on the platform of a PIXImus densitometer (software version 1.46.007; Lunar Corp., Madison, WI, USA) for BMD and BMC measurements of the whole specimen according to the manufacturer's instructions. In some experiments, the variability in measurements was examined by repeating scans after repositioning the animals. CV of BMD for the repeated scans was 1–3% at all skeletal sites examined.
Three‐point bending mechanical test
A three‐point bending test was performed on the mid‐shaft of the femur using a Mach‐1 Micromechanical System (Motion A300.5, Version 3.0.2; Bio Syntech Canada, Laval, Canada). The extrinsic parameters (ultimate force [Fult], ultimate displacement [dult], stiffness [K or S], and work to failure [W or U]) were extracted from a force‐displacement curve. The span of two support points was 7 mm. The deformation rate was 50 μm/s.
Skeletal radiography
After removal of soft tissue, radiographs were taken using a Faxitron model 805 radiographic inspection system (Faxitron Contact; Faxitron; 22‐kV voltage and 4‐minute exposure time). X‐Omat TL film (Eastman Kodak, Rochester, NY, USA) was used and processed routinely.
Histology
Lumbar spine was removed and fixed in PLP fixative (2% paraformaldehyde containing 0.075 M lysine and 0.01 M sodium periodate) overnight at 4°C and processed histologically as previously described.(13) The samples were decalcified in EDTA glycerol solution for 5–7 days at 4°C. Decalcified tibias and other tissues were dehydrated and embedded in paraffin, after which 5‐μm sections were cut on a rotary microtome. The sections were stained with H&E or histochemically for TRACP activity(14) or immunohistochemically as described below. Alternatively, undecalcified lumbar spine was embedded in LR white acrylic resin (London Resin Co., London, UK), and 1‐μm sections were cut on an ultramicrotome. These sections were stained for mineral with the von Kossa staining procedure and counterstained with toluidine blue. Histomorphometric indices were determined as suggested by the ASBMR Histomorphometry Nomenclature Committee.(15)
Immunohistochemical staining
Mouse monoclonal antibody against the type 1 PTH receptor (Upstate Cell Signaling Solutions, Charlottesville, VA, USA) was applied to dewaxed paraffin sections overnight at room temperature. As a negative control, preimmune serum was substituted for the primary antibody. After washing with high salt buffer (50 mM Tris‐HCl, 2.5% NaCl, 0.05% Tween 20, pH 7.6) for 10 minutes at room temperature followed by two 10‐minute washes with tris‐buffered saline (TBS), the sections were incubated with secondary antibody (biotinylated goat anti‐mouse IgG; Sigma), washed as before, and incubated with the Vectastain ABC‐AP kit (Vector Laboratories, Burlington, Canada) for 45 minutes. After washing as before, red pigmentation was produced by incubating with a substrate of alkaline phosphatase (ALP; i.e., fast red TR/Naphthol AS‐MX phosphate; Sigma Chemical, St Louis, MO, USA), containing 1 mM levamisole as endogenous ALP inhibitor, for 10–15 minutes. After washing with distilled water, the sections were counterstained with methyl green and mounted with Kaiser's glycerol jelly.
Histochemical staining for TRACP and ALP
Enzyme histochemistry for TRACP was performed as previously described.(14) Dewaxed sections were preincubated for 20 minutes in buffer containing 50 mM sodium acetate and 40 mM sodium tartrate at pH 5.0. Sections were incubated for 15 minutes at room temperature in the same buffer containing 2.5 mg/ml naphthol AS‐MX phosphate (Sigma) in dimethylformamide as substrate and 0.5 mg/ml fast garnet GBC (Sigma) as a color indicator for the reaction product. After washing with distilled water, the sections were counterstained with methyl green and mounted in Kaiser's glycerol jelly. ALP staining was performed on methyl methacrylate (MMA)‐embedded undecalcified bone samples using a vector red alkaline phosphatase substrate kit (Vector Laboratories).
Serum osteocalcin and TRACP5b levels
A mouse osteocalcin 2‐site immunoradiometric assay (IRMA; Immutopics, San Clemente, CA, USA) was used for the measurement of serum osteocalcin levels according to the manufacturer's specifications. A mouse TRACP5b assay (IDS, Fountain Hills, AZ, USA) was used for the determination of osteoclast‐derived TRACP5b in mouse serum samples. The assay was performed according to the manufacturer's specifications.
Bone marrow cell cultures
Primary bone marrow cell cultures were performed as previously described.(13,14) Tibias and femurs of 3‐month‐old oophorectomized wildtype mice were removed under aseptic conditions, and bone marrow cells were flushed out with DMEM containing 10% FCS, 50 μg/ml ascorbic acid, 10 mM β‐glycerophosphate, and 10−8 M dexamethasone. Cells were dispersed by repeated pipetting, and a single cell suspension was achieved by forcefully expelling the cells through a 22‐gauge syringe needle; 106 bone marrow cells were cultured in 55‐cm2 petri dishes in 10 ml of the above medium. The medium was changed every 4 days. The nonadherent cells containing hematopoietic elements were removed by gently pipetting when the medium was changed for the first time. Cultures were maintained for 25 days. At the end of the culture period, cells were washed with PBS, fixed with PLP fixative, and stained. For determination of total colonies formed, cells were first washed in borate buffer (10 mM; pH 8.8) and stained with 1% methylene blue (wt/vol) in borate buffer for 30 minutes at room temperature. Cells were washed three times in borate buffer alone and left to dry before the number of colonies was quantitated by image analysis as described. For determination of differentiating colonies, cells were stained cytochemically for ALP as described previously.(13,14) After each staining, culture plates were photographed over a light box with a Sony charge‐coupled device camera. Images were analyzed using Northern Eclipse image analysis software. The data were imported to a spreadsheet program and processed as previously described.(13,14)
Computer‐assisted image analysis
After H&E staining or histochemical or immunohistochemical staining of sections from four mice of each group, images of fields were photographed with a Sony digital camera. Images of micrographs from single sections were digitally recorded using a rectangular template, and recordings were processed and analyzed using Northern Eclipse image analysis software.
Statistical analysis
Data from image analysis are presented as means ± SE. Statistical comparisons were made using two‐way ANOVA with Benferrroni adjustment or with a one‐way nonparametric ANOVA, and p < 0.05 was considered significant. To examine combinations of PTH, alendronate, and OPG, a Z‐test was used to calculate the p values, with p < 0.05 considered significant.
RESULTS
Both single therapy and combination therapy increased BMD
After 8 weeks of single treatment with alendronate, OPG, PTH, or of combined treatment with PTH plus alendronate or PTH plus OPG, BMD in the lumbar spine and femur was markedly increased compared with vehicle‐treated controls (Figs. 1A and 1B). In the lumbar spine, PTH plus OPG produced the greatest increase in BMD. In the femur, PTH produced a greater increase in BMD than either anti‐catabolic, but co‐treatment of PTH with OPG or with alendronate had the greatest effect on increasing BMD (Figs. 1A and 1B). To determine if the co‐treatments had an additive or synergistic effect, we compared the net increase in BMD produced by the treatment after subtracting initial BMD on day 0 (Figs. 1C and 1D). We determined the predicted net effects of the co‐treatments by adding the net effects of PTH alone with the net effect of alendronate alone and the net effect of PTH alone with the net effect of OPG alone (Figs. 1C and 1D, last two columns). The results indicated that in the lumbar spine the observed increases in co‐treatment of PTH with either alendronate or OPG exceeded the predicted increases (Fig. 1C), whereas in the femur, the observed increases were essentially the same as the predicted increases (Fig. 1D). These analyses suggest that in the femur the effect of co‐treatment was additive, whereas in the lumbar spine, the effect was synergistic.
BMD in lumbar spine and whole femur after 2 months of treatment with single or combined therapy. (A) BMD in lumbar spine at day 60, (B) BMD in femur at day 60, (C) net increase in lumbar spine calculated as BMD at day 60 minus BMD at day 0, and (D) net increases in femur BMD. The last two columns on the right in C and D show the sum of the net increases in alendronate (ALN) and PTH (columns 1 and 3), and in OPG and PTH (columns 2 and 3), respectively. Data are shown as mean ± SE (N = 10). Two‐way ANOVA with Bonferroni adjustment was used in A and B. A Z‐test was used in the analyses of C and D to calculate the p values. *p < 0.05, **p < 0.01, and ***p < 0.001 compared with vehicle (PBS) in A and B. ωp < 0.05 compared with PTH + ALN. ∧p < 0.05 compared with OPG and ALN. ∇p < 0.05 compared with ALN.
The lumbar spine and the femur samples obtained from all groups were also analyzed ex vivo by contact radiography. The results showed an increase in BMD in all treated groups compared with vehicle (Figs. 2A and 2B). Furthermore, the increase in BMD of the lumbar spine and the femur of animals treated with PTH alone was greater than with anti‐catabolic therapy, and PTH in combination with alendronate or OPG produced a greater increase in BMD than any of the single therapies. These results seemed similar with von Kossa stains of vertebral sections (Fig. 2C). Histomorphometric analysis of mineralized trabecular bone volume (BV/TV) at the proximal tibial metaphysis showed that all treatments produced increases compared with PBS but that alendronate alone produced a weaker effect than the other treatments (Fig. 1D).
Radiologic, histologic, and histomorphometric analysis of bone specimens. (A) Faxitron analysis of lumbar spine. (B) Faxitron analysis of femur. (C) Representative sections of von Kossa staining of lumbar spine showing mineralized bone. (D) Trabecular volume relative to tissue volume (BV/TV) in lumbar spine. Data are shown as mean ± SE (N = 3). One‐way ANOVA was used for analyses. *p < 0.05 compared with PTH‐only treated animals. ∧p < 0.05 and ∧∧p < 0.01 compared with ALN. #p < 0.05, ##p < 0.01, and ###p < 0.001 compared with PBS. PBS is the vehicle‐treated control.
Temporal effects of single and combination therapies on BMD
To determine the timing of the effects of single and combination therapies, we examined the BMD at the lumbar spine and femur on days 0, 30, and 60 of the experiment (Figs. 3A and 3B, respectively). In the lumbar spine, there was no significant difference between the BMD on days 30 and 60 for any of the single therapies (alendronate, OPG, or PTH). In contrast, in the femur, although the effect of the antiresorptives had plateaued by 30 days, there was still a significant difference between days 30 and 60 after treatment with PTH. Interestingly, there were no plateau effects observed with either of the combination therapies in the lumbar spine or the femur.
Temporal effects of single and combination therapies on BMD in (A) lumbar spine and (B) femur. In each group of columns, the first, second, and third columns represent days 0, 30, and 60, respectively, of the treatment shown on the horizontal axis. Data represent mean ± SE (N = 10). NS, not significant. Two‐way ANOVA with Bonferroni adjustment was used for analyses. *p < 0.05 and **p < 0.01. PBS is the vehicle‐treated control.
Effect of single and combination therapies on the mechanical strength of bone
We next examined the effect of the single and combination therapies on the mechanical strength of the femur using a three‐point bending test. PTH alone, and PTH plus alendronate or OPG increased the stiffness (K, N/mm, resistance of the bone to a displacement; Fig. 4A), the ultimate force (Fult, N, the maximum force that the bone can resist; Fig. 4B), and work to failure (U, N*m, the energy that is required to break the bone; Fig. 4C) of cortical bone compared with vehicle‐treated controls. OPG had no effect on the stiffness of cortical bone; however, it significantly increased the ultimate force and the energy to fracture in comparison with ALN or vehicle (Fig. 4C). Alendronate had no effect on any of the parameters of mechanical strength (Fig. 4) at the dose used.
Effect of single and combination therapies on the mechanical strength of femur. (A) Stiffness (K, resistance of bone to displacement). (B) Fult (maximum force bone can resist). (C) Work to failure (U, energy required to break bone). Data are shown as mean ± SE (N = 6–7). One‐way ANOVA was used for analyses. *p < 0.05, **p < 0.01, and ***p < 0.001 compared with vehicle (PBS) or ALN.
Effect of single and combination therapies on osteoclastic bone resorption
Both anti‐catabolic drugs, alendronate and OPG, decreased the number of osteoclasts in bone samples (Figs. 5A and 5B); however, this decrease was highly remarkable with OPG, and very few TRACP+ cells were observed in bone. PTH injection alone markedly increased the number of osteoclasts, whereas PTH in combination with OPG produced minimal stimulation, and, as with OPG treatment alone, very few TRACP+ cells were detected. On the other hand, PTH was able to significantly increase the number of osteoclasts when combined with alendronate. We also measured serum TRACP5b levels at the end of the experiment as a marker of bone resorption (Fig. 5C). Serum TRACP5b levels were correlated overall with the TRACP staining of bone specimens. Thus, PTH alone markedly increased the levels of TRACP5b, whereas alendronate alone, and especially OPG alone, significantly decreased levels (Fig. 5C). When PTH was given in combination with alendronate, it reversed the suppressive effect of alendronate on the levels of TRACP5b (Fig. 5C). PTH only minimally increased TRACP5b when administered with OPG compared with OPG alone (Fig. 5C).
Effect of single and combination therapies on the number and activity of osteoclasts. (A) Representative sections of osteoclasts, denoted by TRACP staining. (B) Quantitation of the number of TRACP+ osteoclasts per bone perimeter (Oc.N/B.Pm). (C) Serum levels of TRACP5b. Data are mean ± SE (N = 10). One‐way ANOVA was used for analyses. *p < 0.05 and **p < 0.01 compared with vehicle‐treated controls (PBS). ▾p < 0.001 compared with PTH + ALN. ∧p < 0.001 compared with ALN. ##p < 0.01 compared with PTH.
Effect of single and combination therapies on osteoblastic activity
We assessed osteoblasts by immunostaining for the PTH receptor (Fig. 6A) and by alkaline phosphatase staining (Fig. 6B). Osteoblasts were significantly decreased by OPG treatment, and co‐treatment of PTH with OPG produced only a minimal increase relative to OPG alone. Alendronate alone decreased osteoblasts only slightly, and co‐treatment of PTH with alendronate markedly increased osteoblasts. Osteoblastic activity was determined by serum levels of the bone marker osteocalcin and generally correlated with the histological assessment. Thus, the levels of serum osteocalcin were markedly decreased by OPG, and less so by alendronate (Fig. 6C). PTH co‐treatment with OPG only minimally increased osteocalcin levels compared with OPG alone, although PTH co‐treatment with alendronate significantly increased osteocalcin compared with alendronate alone (Fig. 6C).
Effects of single and combination therapies on osteoblasts. (A) Representative sections of osteoblasts, denoted by immunohistochemical staining for PTHR. (B) Quantitation of ALP+ osteoblastic perimeter per bone perimeter (Ob.Pm/B.Pm). (C) Serum levels of osteocalcin. Data are shown as mean ± SE (N = 10). One‐way ANOVA was used for analyses. *p < 0.05 and **p < 0.01 compared with vehicle‐treated control (PBS). Δp < 0.05 compared with ALN. ###p < 0.001 compared with PTH + OPG.
Effect of the anti‐catabolic agents on osteoblastic colony formation ex vivo
To determine if the anti‐catabolic agents had a direct influence on osteogenesis, we examined the effect of these agents on osteoblastic colony formation in bone marrow cultures (BMCs) in vitro. Alendronate (10−6 M) significantly decreased the number of colonies but increased ALP expressing osteoblastic colonies in BMC culture (Figs. 7A and 7B). OPG (100 ng/ml) produced a less marked reduction in proliferation but more markedly increased the number of differentiating colonies (Figs. 7C and 7D). Furthermore, the number of colonies and also the number of ALP+ colonies were significantly greater with OPG compared with ALN.
Effects of ALN and OPG on proliferation and differentiation of BMCs in culture. BMCs were cultured in osteogenic differentiation medium with PBS (control), ALN, or OPG and examined for total number of colony forming units, fibroblastic (Total CFU‐f) by methylene blue staining (A) and for osteoblastic colony‐forming units (CFU‐fob) by ALP staining (C). Quantitation of total number of CFU‐f (B) and of CFU‐fob (D) was performed by image analysis. Values are mean ± SE of triplicate determinations from three replicate experiments. One‐way ANOVA was used for analyses. *p < 0.05 compared with control. |P%p < 0.05 compared with ALN.
DISCUSSION
Using an oophorectomized mouse model, we specifically examined the question of whether concomitant treatment of PTH with an anti‐catabolic might interfere with the bone‐forming actions of PTH and compared the effect of two ant‐catabolics that use different molecular mechanisms: a bisphosphonate and an inhibitor of RANKL. A mouse model is limited by the fact that, even at 4 months of age, both modeling and remodeling activity are occurring. In addition, vertebrae are not weight bearing as in humans, and the size of the vertebrae precludes determination of crush strength, which is why bending strength of femoral bone, which determines primarily cortical bone, was determined. Size considerations also preclude measuring femoral neck BMD rather than total femoral BMD. Furthermore, calcium and phosphorus kinetics and physiology may be different than in higher animals. With these limitations in mind, our results showed that neither alendronate nor OPG inhibited the capacity of PTH to increase trabecular bone volume above the levels seen with either agent alone. Increases observed were, in fact, above the levels seen with PTH alone. Furthermore, PTH, when coupled with either anti‐catabolic, produced BMD increases that seemed additive in the femur and supra‐additive in the lumbar spine. Mechanical loading tests of the femur revealed that OPG alone but not alendronate alone, at the doses tested, increased both the maximum load and the energy absorbed, whereas these parameters and also stiffness were increased by PTH alone. Alendronate may therefore have a more profound effect on the mechanical strength of trabecular rather than cortical bone leading to its known efficacy in fracture prevention. The results with PTH and OPG alone are consistent with results recently reported in aged oophorectomized rats.(11,12) Importantly, in our studies, neither OPG nor alendronate impeded the strengthening effects of PTH on femoral bone when either was used in association with PTH. Similar results with PTH and OPG have been reported in rats,(11,12) in which improvements in the mechanical strength of femoral diaphysis in PTH‐treated or PTH plus OPG‐treated animals was thought most likely to be caused by increased bone mass at approximately equal material quality.
Overall, these results from our studies indicate that combination therapy of PTH with either anti‐catabolic compound is in fact beneficial for BMD, bone volume, and the mechanical strength of appendicular bone.
Each anti‐catabolic decreased osteoclastic bone resorption and osteoblastic bone formation as indicated by both histological and biomarker analysis. This inhibition of bone turnover was especially marked with OPG, yet OPG increased bone volume, bone strength, and BMD to a greater extent than alendronate. Consequently, although cellular instruments of both resorption and formation were diminished, the greater inhibition of bone resorption relative to formation achieved by OPG clearly was still sufficient to produce a substantial net bone gain.
The effect of increasing BMD was observed to plateau with either anti‐catabolic by 30 days of therapy in both the spine and long bones. Such an effect has previously been reported to occur with alendronate.(16) In view of the fact that histologic and biomarker indices of bone resorption were still decreased at 60 days with both agents despite the plateau, it seems likely that osteoblastic bone deposition had now stabilized to match the low bone resorption that was occurring. The relatively rapid plateau effect achieved may reflect the rapidity with which the lost bone of the oophorectomized animal is restored by bone formation processes, after which no net bone is accrued but bone turnover reaches a new equilibrium at a reduced rate.
PTH alone produced the familiar increase in coupled resorption and formation; however, formation clearly exceeded resorption. The anabolic effects of PTH may involve initial stimulation of proliferation of pre‐osteoblastic cells,(17) and subsequent stimulation of differentiation, including differentiation of bone lining cells,(18–21) and inhibition of osteoblast apoptosis.(22)
Although the effect of PTH also plateaued in the lumbar spine by 30 days, BMD increases in the femur continued. Such divergent responses to PTH in different skeletal sites, including more profound effects on long bones than on the axial skeleton, has previously been noted.(23,24)
Despite the plateau in BMD observed after 30 days of PTH treatment, bone turnover in trabecular bone, as determined histologically, remained high at 60 days. Consequently, net PTH‐stimulated bone resorption must have increased to match the increased bone formation that still persisted. This effect may occur more rapidly in the more metabolically active trabecular versus cortical bone, causing the temporal differences in response observed.
Interestingly, despite the capacity of OPG to produce very profound inhibition of bone turnover, and the reduced capacity of PTH to stimulate bone turnover when used in combination with OPG relative to co‐treatment with alendronate, the net effects of PTH on bone volume and on BMD were approximately the same when PTH was combined with OPG or with alendronate. Consequently, sufficient net osteoblastic activity occurred in both cases. This indicates that, although the magnitude of increases in bone turnover as denoted by biomarkers may predict anabolic responses to PTH when used as a sole therapeutic,(25,26) such increases may not accurately reflect the efficacy of PTH when used in conjunction with anti‐catabolic agents.
Importantly both anti‐catabolics prevented the plateau effect observed with PTH in the lumbar spine. Consequently, the increased magnitude of the PTH response observed with anti‐catabolic therapy both in the lumbar spine and also in the femur may be caused by inhibition of PTH‐stimulated bone resorption, which may contribute to the plateau, and the increased efficacy of combination therapy may lie at least in part in retarding the achievement of equilibrium bone turnover. Therefore, rather than impeding the action of PTH,(27) inhibition of bone osteoclastic activity may have been beneficial for the efficacy of PTH.
More complex mechanisms may explain the interactions between PTH and the ant‐catabolics, however, as suggested by our studies examining the effect of the anti‐catabolics on osteogenic colony formation in vitro. Thus, alendronate markedly reduced proliferation of colonies, but osteogenic differentiation proceeded in these reduced colonies. This may in part have occurred by a direct effect, because direct effects of bisphosphonates on cells of the osteoblastic lineage have previously been described,(28,29) at least in vitro. On the other hand, OPG produced a less profound decrease in colony numbers but resulted in a greater increase in differentiating osteoblastic colonies. This effect was almost certainly indirect, likely on responsive cells of the osteoclast lineage in the marrow cultures, because the RANK/RANKL signaling system is not known to be expressed in osteoblastic cells. It is therefore possible that suppression of osteoclastic release of a negative regulator of osteoblastic differentiation,(30) or suppression of positive regulators of osteoblast proliferation may have occurred, and indeed, direct release of osteoblast modulators from osteoclasts has been suggested.(27,31) Nevertheless, with both anti‐catabolics, the resulting increase in differentiating osteogenic colonies we observed should have provided an increased osteoblastic pool for the anabolic effects of PTH. In fact, however, osteoblasts were decreased in vivo after co‐treatment of PTH with the anti‐catabolics relative to PTH alone. Consequently, the decreased pool of differentiated osteoblasts observed in vivo and that was related to the anti‐catabolics was likely caused by a reduction in liberation of growth factors from resorbed matrix,(32) a mechanism that would not have been captured by the in vitro studies. These studies therefore emphasize the important regulatory role that osteoclasts exert on osteoblast pools, even in the presence of PTH, and suggest that such a coupling mechanism, at least under the current experimental paradigm, is largely indirect and requires bone matrix resorption.
Overall, our results, using a mouse model, indicate that even an anti‐catabolic as potent as OPG may produce dramatic increases in bone; that the temporal pattern of activity of bone formation and resorption modulators may have major influence on net skeletal accrual; and that, depending on timing, inhibition of osteoclastic activity may not impede the anabolic action of PTH but that PTH‐like anabolic agents may be complemented in their bone‐forming action by concomitant addition of anti‐catabolic compounds.
Acknowledgements
RS is a Bone Scholar of the Skeletal Health Training Program of the Canadian Institutes for Health Research (CIHR). This work was supported by grants to DG from the CIHR. We thank Dr Paul Kostenuik of Amgen for providing important materials and assistance with the study and for very thoughtful review of the manuscript, Stephen Adamu for kindly performing the biomarker assays, and Dr Masoud Asgharian of the Department of Mathematics and Statistics at McGill University for excellent assistance with statistical calculations.
REFERENCES
Author notes
Dr Goltzman has served as a consultant to Amgen and Merck. All other authors state that they have no conflicts of interest.
