For a Commentary on this article, please see Khosla (J Bone Miner Res. 2012;27:506–508. DOI: 10.1002/jbmr.1541).
Odanacatib treatment increases hip bone mass and cortical thickness by preserving endocortical bone formation and stimulating periosteal bone formation in the ovariectomized adult rhesus monkey†
Article first published online: 21 FEB 2012
Copyright © 2012 American Society for Bone and Mineral Research
Journal of Bone and Mineral Research
Volume 27, Issue 3, pages 524–537, March 2012
How to Cite
Cusick, T., Chen, C. M., Pennypacker, B. L., Pickarski, M., Kimmel, D. B., Scott, B. B. and Duong, L. T. (2012), Odanacatib treatment increases hip bone mass and cortical thickness by preserving endocortical bone formation and stimulating periosteal bone formation in the ovariectomized adult rhesus monkey. J Bone Miner Res, 27: 524–537. doi: 10.1002/jbmr.1477
- Issue published online: 21 FEB 2012
- Article first published online: 21 FEB 2012
- Accepted manuscript online: 23 NOV 2011 09:53AM EST
- Manuscript Accepted: 8 AUG 2011
- Manuscript Revised: 27 JUL 2011
- Manuscript Received: 24 MAR 2011
- CATHEPSIN K INHIBITOR;
- NONHUMAN PRIMATE;
- Top of page
- Materials and Methods
Odanacatib (ODN) is a selective and reversible inhibitor of cathepsin K (CatK). Previously, ODN was shown to increase bone mineral density (BMD) and maintained normal bone strength at the spine in ovariectomized (OVX) rhesus monkeys. Here, we further characterize the effects of ODN on BMD, bone strength, and dynamic histomorphometric analyses of the hip from the same monkeys. Animals were treated for 21 months with vehicle, 6 or 30 mg/kg ODN (p.o., q.d.). ODN increased femoral neck (FN) BMD by 11% and 15% (p < 0.07) and ultimate load by 25% (p < 0.05) and 30% (p < 0.01) versus vehicle. Treatment-related increases in ultimate load positively correlated with the increased FN BMD, bone mineral content (BMC), and cortical thickness. Histomorphometry of FN and proximal femur (PF) revealed that ODN reduced trabecular and intracortical bone formation rate (BFR) but did not affect long-term endocortical BFR. Moreover, ODN stimulated long-term FN and PF periosteal BFR by 3.5-fold and 6-fold with the 30 mg/kg dose versus vehicle, respectively. Osteoclast surfaces were either unaffected or trended higher (∼twofold) in endocortical and trabecular surfaces in the ODN group. Lastly, ODN increased cortical thickness of FN by 21% (p = 0.08) and PF by 19% (p < 0.05) versus vehicle after 21 months of treatment. Together, both doses of ODN increased bone mass and improved bone strength at the hip. Unlike conventional antiresorptives, ODN displayed site-specific effects on trabecular versus cortical bone formation. The drug provided marked increases in periosteal bone formation and cortical thickness in OVX monkeys, suggesting that CatK inhibition may represent a novel therapeutic approach for the treatment of osteoporosis. © 2012 American Society for Bone and Mineral Research
- Top of page
- Materials and Methods
Remodeling continues throughout life to maintain structural bone integrity and to mediate responses to physiologic stimuli, leading to changes in density, shape, and strength of bone.1 Although cortical bone provides the correct architecture and stiffness to support loads, the trabecular bone contributes to both metabolic and strength functions.2, 3 In the mature skeleton, bone remodeling occurs in discrete basic multicellular units (BMUs), where osteoclastic bone resorption and osteoblast-mediated bone formation is coupled together.4, 5 BMUs can be found on trabecular, endocortical, and intracortical surfaces. Because trabecular bone has a larger surface-to-volume ratio than cortical bone, its remodeling rate is generally faster.6, 7 In adults, bone modeling continues to modify bone size and shape as an adaptive response to prevailing loading circumstances.8 It remains controversial whether periosteal bone formation in an adult should be classified as a modeling or remodeling process.9
Bone matrix contains organic components, primarily type I collagen, which gives it tensile strength and an inorganic component, primarily hydroxyapatite, which provides stiffness and resistance to compression. To degrade bone efficiently, osteoclasts (OC) must be able to solubilize both fractions. In addition to secreting acid to effectively demineralize hydroxyapatite, OCs also secrete the lysosomal cysteine protease cathepsin K (CatK) into the resorption lacunae to degrade demineralized bone matrix proteins.10 CatK is highly and relatively selectively expressed in OCs.11 Pycnodysostosis, a rare autosomal recessive skeletal disorder characterized by high bone mineral density (BMD), acroosteolysis of the distal phalanges, short stature, and skull deformities, is linked to a number of mutations in the CatK gene, leading to its loss of function.12–14 Several lines of CatK-deficient mice have been generated, and all show an osteopetrotic phenotype associated with high BMD at multiple bone sites and thicker cortices compared with wild-type mice.15–17 Although CatK-deficient OCs were demonstrated to resorb bone inefficiently, the surface-based bone formation rate in trabecular bone of the distal femur was twofold higher in CatK−/− mice compared with aged-matched wild-type mice regardless of gender.18 In contrast, transgenic mice overexpressing CatK display an increased bone turnover rate associated with low bone mass.19
Modulation of CatK activity by small molecule inhibitors reduces bone resorption activity of OCs in vitro and in vivo.20 In preclinical models, CatK inhibitors with limited potency toward the rodent enzymes prevented trabecular bone loss in ovariectomized (OVX) rats or mice at high doses.21–23 We demonstrated that two CatK inhibitors, L-006235 and odanacatib (ODN), dose-dependently blocked estrogen deficiency–induced bone loss in rabbits with efficacy similar to alendronate (ALN).16 However, unlike ALN, the CatK inhibitors did not suppress the bone formation rate in either the trabecular bone of the vertebral body or Haversian cortical bone of the femur.16 Relacatib, another potent human CatK inhibitor with limited selectivity against other cathepsins, decreased bone turnover markers in estrogen-deficient skeletally mature female and young intact male cynomolgus monkeys.24 Interestingly, in OVX monkeys, while preserving lumbar spine areal BMD and femoral neck (FN) volumetric BMD, relacatib reduced trabecular bone formation but did not affect the intracortical bone formation rate.25
ODN (MK-0822) is a highly selective, potent, and reversible inhibitor of human CatK and inhibits OC-mediated bone resorption in vitro.26 ODN is currently being developed as an orally bioavailable therapeutic for the treatment of postmenopausal osteoporosis. In a phase II clinical trial, 36 months of once-weekly ODN treatment dose-dependently increased lumbar spine (7.9%) and total hip (5.8%) BMD and reduced bone turnover markers in postmenopausal women with low BMD.27 In this study, patients treated with ODN had initially reduced levels of bone formation markers, BSAP and P1NP, reaching a plateau after ∼6 months and subsequently increased to above (BSAP, 18%) or near baseline (P1NP, −6%), significantly less than the reduction typically found with bisphosphonates treatment.27 The observation that bone formation is elevated in CatK-deficient mice18 and affected less than bone resorption in studies with pharmacological inhibition of CatK in the OVX preclinical models,16, 28 and in humans,27 has led to the hypothesis that CatK activity may play a key role in the coupling of bone formation to resorption during the remodeling process.
In this study, the long-term treatment effects of ODN at two dose levels on BMD and biomechanical properties at the hip of newly OVX rhesus monkeys were evaluated. Efficacy of ODN on cortical dimension and histomorphometric parameters at trabecular, periosteal, endocortical, and intracortical envelopes in the femurs of these monkeys after 21 months of treatment were also evaluated.
Materials and Methods
- Top of page
- Materials and Methods
The in-life portion of this study was conducted at the New Iberia Research Center (New Iberia, LA, USA) with the approval of Institutional Animal Care and Use Committees (IACUCs) from both the University of Louisiana at Lafayette and Merck. A detailed study design including baseline characteristics of the animals, diets, surgeries, and housing conditions were previously described.28 Briefly, 42 female rhesus monkeys (Macaca mulatta, aged 13 to 23 years) were maintained on a high-protein diet (19.8%) containing 1.17% Ca, 0.7% P, 8 IU/g vitamin D3 (Harlan-Teklad [Indianapolis, IN, USA] 8773 NIB primate diet; ∼200 g/d). The animals were randomized into four groups by weight and baseline lumbar spine (L1 to L4) bone mineral density (LsBMD). At baseline, body weights of the monkeys ranged from 5.89 to 11.13 kg, with a mean of 7.89 ± 0.21 kg (mean ± SEM). After bilateral ovariectomy, the monkeys were dosed p.o., daily as follows: vehicle (n = 11), ODN 6 mg/kg (n = 10), or ODN 30 mg/kg (n = 10). Dosing started at ∼11 days postsurgery and continued for 21 months. Intact animals (n = 11) were also dosed daily with vehicle (0.5% methocel plus 0.1% SDS; Sigma-Aldrich, St. Louis, MO, USA). Plasma for pharmacokinetic (PK) studies was collected from the ODN-treated groups. Both serum and urine were collected from all groups for analyses of bone turnover markers, as reported elsewhere.28
In vivo dual-energy X-ray absorptiometry (DXA)
DXA scans of animals were performed at baseline and approximately at 3, 6, 9, 12, 16, and 20 months after study start, using a Hologic QDR Discovery A bone densitometer (Hologic, Inc., Waltham, MA, USA). The right total hip and FN as well as other skeletal sites including whole body, L1 to L4, distal radius, and 1/3 distal radius were scanned for areal bone mineral content (BMC) and BMD. Non-hip BMD data are reported elsewhere.28 A right total hip DXA scan analysis (version 12.3:3) was performed in high-definition mode. A manual analysis was performed to determine BMD of the femoral neck. Data were generated as an average from two separate scans.
The intact group (n = 11) was recruited to another study, thus no ex vivo results from the intact group are available. One monkey from the intact group was euthanized because of severe pulmonary edema after 16 months on study, and one monkey from the ODN 6 mg/kg group was euthanized because of lung pathology, abscesses, and adhesions after 10 months on study. The causes of death of these animals were determined to be non-drug related. In addition, one monkey from the ODN 6 mg/kg group developed pneumothorax, post rib biopsy at 12 months.
At necropsy, the left femora were collected and fixed in 70% ethanol for histomorphometric analyses. The right femora were wrapped in saline-soaked gauze and stored at −20°C until analyzed for bone strength. Lumbar vertebrae were collected, and data are reported elsewhere.28
Ex vivo analyses
Histomorphometric analyses of cortical bone
For histomorphometric evaluation of the left FN, two parallel cuts were made with a water-cooled Exakt 300 CP Bandsaw with a diamond blade (Exakt Technologies, Inc., Oklahoma City, OK, USA) perpendicular to the FN axis, 1 mm proximal to the base of the femoral head and through the intertrochanteric line, creating a ∼10-mm segment for embedding. The FN axis was defined in the coronal plane as the line extending from the fovea to the intertrochanteric line. To analyze proximal femur (PF) cortical bone, an additional cross-sectional segment located 4 cm distal to the greater trochanter and extending 1 cm distally was collected. The remaining PF specimen was subjected to frontal sectioning for histomorphometric analysis of trabecular bone.29
Cortical bone specimens isolated from the FN and PF shaft were dehydrated sequentially in ascending concentrations of ethanol (70%, 95%, 100%) and then infiltrated and embedded without decalcification in 90% methyl-methacrylate (Polysciences, Warrington, PA, USA)/10% dibutyl phthalate (Sigma, Atlanta, GA, USA). Serial thick ∼120-mm cross sections of the FN commenced at the femoral head and continued distally throughout its length. Serial sections of ∼100 µm were also collected from the distal PF shaft. Sections were obtained utilizing a Leica SP1600 saw microtome (Leica Biosystems, Heidelberg, Germany). Bone slices were mounted on glass slides (Fisher Scientific, Pittsburgh, PA, USA), and coverslips were applied using Eukitt's mounting media (EMS, Ft. Washington, PA, USA). All measurements were made using a light/epifluorescent microscope, Nikon Eclipse 80i (Nikon Instruments, Melville, NY, USA) equipped with an Optonics DEI-750 CE (Tuttlingen, Germany) video camera interfaced to an image analysis system (Bioquant R&M Biometrics, Nashville, TN, USA).
Double fluorochrome labels with a 15-day interval were administered: calcein (12 mg/kg, sc) at 12 months and tetracycline (250 mg/kg, p.o.) at study end to label bone-forming surfaces. Dynamic histomorphometric endpoints measured at the periosteal, endocortical, and intracortical surfaces were based upon the “classical” short-term labeling interval of 15 days. At the periosteal and endocortical surfaces, the long-term (LT) labeling interval between calcein and tetracycline of approximately 292 days was also assessed. As previously described,30 mineralizing surface (MS/BS, %) was calculated as double plus half the single-labeled surface divided by total bone surface. Mineral apposition rate (MAR, µm/d) was calculated as the mean distance between the first and second label at equidistant points divided by the interlabel time period of 15 days. Bone formation rate (BFR/BS, µm3/µm2/year) was expressed per unit of bone surface and was calculated by multiplying MS/BS and MAR. Before statistical analysis, bone specimens without double-label were excluded from the calculation of mean MAR, but MS/BS and BFR were assigned a value of zero.31 For LT histomorphometric measurements, LT MS/BS was calculated as the length of overlapping regions of tetracycline- and calcein-labeling plus half of the length of either single tetracycline or calcein alone. LT MAR was calculated as the mean distance between tetracycline and calcein labels at equidistant points divided by the long-term labeling interval (292 days). LT BFR was calculated as LT MS/BS multiplied by LT MAR. It is difficult to distinguish the dense interfaces between trabecular and endocortical bone in the FN, and as such, endocortical bone formation in the FN was not quantified.
To measure intracortical or Haversian (H) remodeling at the PF, each ∼100-µm cortical cross-sectioned specimen was divided into anterior, posterior, medial, and lateral, utilizing the insertion of the gluteal tuberosity as a landmark. Each subregion of interest measured ∼2.5 mm2. Total mineralized surface per Haversian bone surface (H.MS/BS), the mean interlabel distance (H.MAR) between double-labeled osteons, and surface-based bone formation rate (H.BFR/BS) were determined. Total Haversian bone surface included both active and resting osteons, as well as resorption cavities.
Static histomorphometric endpoints measured utilizing the 1× objective were total tissue area (T.Ar; mm2), the periosteal and endocortical perimeter (Ps.Pm; Ec.Pm; mm), the cortical area and marrow area (Ct.Ar; Ma.Ar; mm2), and cortical thickness (Ct.Th; mm). Ct.Th at the PF was derived by subtracting the mean radius of the Ct.Ar and Ma.Ar. For quantification of total FN cortical area, the endocortical surface was defined as that surface between connections of trabeculae to the cortex. Evaluation of FN cortical thickness was calculated by averaging approximately 30 individual lengths between the endocortical and periosteal surfaces that were evenly spaced around the cortex.30
Histomorphometric analyses of trabecular bone
Measurements of FN trabecular bone were quantified from a ∼4 × 5-mm region of interest (ROI) from the thick section obtained ∼220 µm distal to the section with the smallest total tissue area between the articular surface of the femoral head and the intertrochanteric region utilizing a 20× objective. The ROI was defined by the cortical bone on all sides, thereby avoiding the endocortical surfaces. To maximize visualization of trabecular bone in the intertrochanteric region of the PF, 6-µm longitudinal sections in the frontal plane were prepared on a Reichert-Jung Polycut Sledge microtome (Leica Microsystems, Wetzlar, Germany). Measurements were obtained from a region 1 mm below the point on the distal lip of the trochanteric fossa and in ∼3 × 4-mm subregion. Endocortical surface evaluation was conducted at 6.5 mm below the point on the distal lip of the trochanteric fossa along a 5.5-mm length on the left side. Sections were either stained by a modified toluidine blue method32 for analysis of static parameters or left unstained for analysis of fluorochrome-based data. Static morphometric endpoints measured include: osteoid area (O.Ar/B.Ar., %), osteoblast surface (Ob.S/BS, %), osteoclast surface (Oc.S/BS, %), and osteoclast number (N.Oc/BS, #/mm).
Specimens were stored at −20°C in saline-soaked gauze and thawed overnight in a refrigerator at 4°C before ex vivo DXA and bone-strength testing. The right femur was cut with a diamond saw through the diaphysis, approximately 24 mm distal to the lesser trochanter. The PF was secured in a fixture that held the trochanteric region and then subjected to an FN shear test using an MTS 858 Mini Bionix servohydraulic test system (MTS, Minneapolis, MN, USA) and Testworks (version 3.8A) for Teststar II (version 4.0c) to collect load to failure data in simulated single-legged stance. Force was applied to the femoral head at a rate of 1 mm/s. After testing, the broken bone was examined to ascertain the direction and location of the fracture. Ultimate load, stiffness, area under the curve (AUC), and ductility were determined from the load-displacement curve. Ultimate load is the maximum value on the load-displacement curve before fracture. Stiffness is the maximum slope of the elastic portion of the load-displacement curve. AUC, or the total energy required to break the bone, is calculated as the area under the load-displacement curve. Ductility is the extent to which bone can be deformed without fracture.31
All statistical comparisons of areal BMD, biomechanics of bone, and histomorphometric parameters were analyzed using Statview software (SAS Institute Inc., Cary, NC, USA). Numerical data obtained were subjected to calculation of group mean values and standard deviations. Data points outside of two standard deviations of the mean were excluded for each parameter of interest with the exception of those used to derive parameters. All data in tables and figures are shown as means ± the standard error (SEM). Data were tested for normal distribution and equal variance, and differences among three or more groups were analyzed by one-way analysis of variance (ANOVA) and followed by Fisher's least significant difference (Fisher's PLSD), except where indicated. If group variances were found to be heterogeneous, data were log transformed and reanalyzed using Dunnett's test. Although differences were considered significant only when p ≤ 0.05, p ≤ 0.1 are also reported. Correlation analysis of ex vivo densitometry- and geometrically-derived parameters and biomechanical strength parameters from the femoral neck were performed using linear regression.
- Top of page
- Materials and Methods
Odanacatib protects bone mineral density of the hip
The baseline characteristics of the animals were previously reported.28 Briefly, the monkeys were randomized primarily by lumbar spine areal BMD and secondarily by body weight. As previously discussed, three animals were excluded from the analyses for reasons determined to be non-drug related. Pharmacokinetic studies showed that ODN at a dose of 6 mg/kg resulted in an exposure of 2.2 µM/24 h, and at 30 mg/kg provided an exposure of 4.2 µM/24 h, indicating that oral bioavailability of ODN is not linear using the methocel-based vehicle in monkeys.
Whereas areal BMD scanned from the right PF of the intact group increased by 12%, from 0.695 ± 0.012 g/cm2 at baseline to 0.765 ± 0.026 g/cm2 at 20 months, hip BMD in the vehicle-treated group (OVX-vehicle) increased by 4%, from 0.638 ± 0.018 g/cm2 to 0.653 ± 0.020 g/cm2 (Fig. 1A,B). After excluding the three monkeys that died, the intact group had significantly higher BMD at baseline compared with OVX-vehicle. Treatment with ODN 6 mg/kg increased hip BMD by 17%, from 0.656 ± 0.014 g/cm2 to 0.745 ± 0.028 g/cm2, significantly higher than that in the OVX-vehicle (p < 0.05) and not different from intact (Fig. 1A,B). Hip BMD in the ODN 30 mg/kg group increased by 19%, from 0.651 ± 0.014 g/cm2 to 0.789 ± 0.030 g/cm2 (p < 0.05 versus OVX-vehicle) as shown in Fig. 1A,B.
Femoral neck BMD in the intact group increased by 14%, from 0.464 ± 0.014 g/cm2 at baseline to 0.535 ± 0.027 g/cm2 after 20 months of dosing, and in the OVX-vehicle by 8.6%, from 0.430 ± 0.014 g/cm2 to 0.470 ± 0.028 g/cm2 (Fig. 1C,D). Treatment with ODN 6 mg/kg increased FN BMD in OVX monkeys by 14%, from 0.461 ± 0.025 g/cm2 at baseline to 0.523 ± 0.027 g/cm2 at 20 months, and with ODN 30 mg/kg by 24% from 0.453 ± 0.016 g/cm2 to 0.539 ± 0.025 g/cm2 (p < 0.05 versus OVX-vehicle) as shown in Fig. 1C,D.
Odanacatib treatment increased bone strength of the femoral neck
Femoral neck bone strength endpoints were assessed (Table 1). ODN treatment for 21 months dose-dependently and significantly increased ultimate load, which achieved significance with both doses of ODN at 6 mg/kg (p < 0.05) and 30 mg/kg (p < 0.01) compared with OVX-vehicle (Table 1). There were nonsignificant but treatment-related trends to increase stiffness and AUC (up to 17%) for ODN-treated animals versus OVX-vehicle (Table 1). Ductility was unchanged in all groups. The increases in biomechanical parameters from the right femurs were consistent with the dose-dependent increases in in-life DXA-based measurements of BMC and BMD of the right femurs (up to 11% and 15%, with 6 and 30 mg/kg, respectively) compared with OVX-vehicle. Regression analyses to assess the relationship between densitometric and geometric properties with the biomechanical properties at the FN were performed. Correlation of ultimate load to FN BMD (R = 0.777; p < 0.0001) and BMC (R = 0.515; p < 0.007) showed a highly positive and linear relationship (Fig. 2). Positive correlations were also observed between ultimate load and FN cortical thickness (R = 0.492, p < 0.013; Fig. 2) and between stiffness and FN BMD (R = 0.458, p < 0.016; data not shown).
|Group||OVX + vehicle||OVX + ODN (6 mg/kg)||OVX + ODN (30 mg/kg)|
|aBMD (g/cm2)||0.47 ± 0.03||0.52 ± 0.03||0.54 ± 0.03|
|BMC (mg)||576 ± 75||557 ± 28||621 ± 25|
|Ultimate load (N)||1924 ± 95||2406 ± 109a||2502 ± 192b|
|Stiffness (N-mm)||1649 ± 115||1803 ± 133||1920 ± 117|
|AUC (N-mm)||3246 ± 222||3791 ± 283||3675 ± 382|
|Ductility (mm)||1.9 ± 0.1||1.8 ± 0.2||1.7 ± 0.2|
Odanacatib reduced trabecular remodeling at the hip
Tetracycline double labels were faintly detectable in the 6-µm thin sections of the PF; therefore, dynamic histomorphometric parameters at this site were determined based upon calcein double labeling. Typical fluorescent images of PF trabeculae from the vehicle-treated group and ODN 30 mg/kg–treated group are shown in Fig. 3A-a,b. Although both calcein (in green) and faint tetracycline (in yellow) were detected in the same regions of the PF of the vehicle-treated OVX monkey (Fig. 3A-a), sporadic fluorescent labels of either agents were occasionally found in the ODN-treated specimens (Fig. 3A-b). Compared with OVX-vehicle, ODN at 6 mg/kg and 30 mg/kg dose-dependently reduced trabecular mineralized surface of the PF (PF.Tb.MS/BS) by 41% and 91% (p < 0.001), respectively (Fig. 3B-a). Although ODN 6 mg/kg decreased the trabecular mineral apposition rate (PF.Tb.MAR, p < 0.05), ODN 30 mg/kg had no significant effect on Tb.MAR (Fig. 3B-b). Thus, treatment with ODN at 6 and 30 mg/kg compared with OVX-vehicle reduced bone formation rate by 51% (p < 0.05) and 93% (p < 0.001), respectively (Fig. 3B-c).
Static histomorphometric endpoints measured from the trabecular PF demonstrated that ODN treatment trended to increase osteoclast surface (Tb.Oc.S/BS) and osteoclast number (Tb.N.Oc./BS) compared with the vehicle control (Table 2). Consistent with dynamic histomorphometric data, ODN treatment dose-dependently reduced osteoblast surface (Tb.Ob.S/BS; p < 0.05) with a trend to decrease osteoid area (Tb.O.Ar/BV) compared with vehicle (Table 2).
|Group||OVX + vehicle||OVX + ODN (6 mg/kg)||OVX + ODN (30 mg/kg)|
|Tb.Oc.N/BS (#/mm)||0.53 ± 0.05||0.84 ± 0.13||0.84 ± 0.17|
|Tb.Oc.S/BS (%)||4.09 ± 0.45||7.11 ± 1.04a||6.08 ± 1.23|
|Tb.Ob.S/BS (%)||1.84 ± 0.48||0.71 ± 0.27a||0.21 ± 0.11b|
|Tb.OS Ar/B Ar (%)||0.91 ± 0.52||0.15 ± 0.05||0.03 ± 0.02|
|Ec.Oc.N/BS (#/mm)||0.27 ± 0.10||0.25 ± 0.10||0.08 ± 0.04|
|Ec.Oc.S/BS (%)||1.69 ± 0.61||2.02 ± 0.92||0.54 ± 0.31|
|Ec.Ob.S/BS (%)||1.79 ± 0.86||0.24 ± 0.16||1.39 ± 1.21|
Effects of ODN on short-term trabecular bone turnover parameters at the FN were quantified utilizing both fluorescent labels, calcein double labeling at 12 months (Fig. 4A), and tetracycline double labeling at 21 months (Fig. 4B). From the calcein labeling at 12 months, whereas ODN at 6 mg/kg did not exert any significant effects on the dynamic histomorphometric parameters in the FN compared with OVX-vehicle (Fig. 4A-a–c), ODN 30 mg/kg significantly reduced Tb.MS/BS (p < 0.005) by 93% and Tb.MAR (NS) by 34% (Fig. 4A-a,b), leading to significant reduction in FN Tb.BFR/BS (p < 0.005) compared with OVX-vehicle (Fig. 4A-c). As determined from tetracycline labeling at 21 months, ODN at 6 mg/kg also did not display significant impact on the bone formation parameters in the FN (Fig. 4B-a–c). ODN 30 mg/kg numerically reduced Tb.MS/BS by 62% (Fig. 4B-a), decreased Tb.MAR (p < 0.05) by 34% (Fig. 4B-b), and thus significantly reduced FN Tb.BFR/BS (p < 0.01) 58% compared with OVX-vehicle (Fig. 4B-c). The results at 21 months showed similar findings to that at 12 months (Fig. 4A,B), suggesting that the effects of ODN on bone remodeling remains consistent during treatment.
There were no significant effects of ODN at either dose on the trabecular bone structural endpoints including bone volume (BV/TV), trabecular thickness (Tb.Th), trabecular number (Tb.N), or trabecular separation (Tb.Sp) (data not shown).
Odanacatib reduced intracortical remodeling in the proximal femur
As shown in Fig. 5, dynamic labeling data indicated a treatment-related reduction in the bone formation parameters at the intracortical surface of the proximal femur after 12 and 21 months in the ODN groups versus OVX-vehicle. At 12 months, ODN at 6 and 30 mg/kg doses significantly reduced total Haversian mineralized surface (H.MS/BS) by 41% (p < 0.07) and 66% (p < 0.01), respectively (Fig. 5A-a). Both doses of ODN trended to decrease total H.MAR versus OVX-vehicle (Fig. 5A-b). Hence, compared with OVX-vehicle, ODN at 6 and 30 mg/kg dose-dependently reduced total H.BFR by 49% (p = 0.05) and 80% (p < 0.01), respectively (Fig. 5A-c). At 21 months, ODN treatment only produced nonsignificant trends in decreasing H.MS/BS and H.BFR/BS versus OVX-vehicle. As described in Materials and Methods, the subregion analysis of H.BFR at both 12 and 21 months of ODN treatment showed significant reduction with the 30 mg/kg dose in the PF posterior region compared with OVX-vehicle (data not shown). No significant treatment-related suppression of H.BFR was detected in other subregions of the proximal femur (data not shown).
Odanacatib spares endocortical bone formation in the proximal femur
Dynamic histomorphometric measurements of PF were performed just below the lesser trochanter, a site where minimal trabecular bone was present at the endocortical (Ec) interface. Representative low- and higher-magnification images of PF cross sections illustrate relatively equal amounts of both calcein (in green) and tetracycline (in yellow) labels visible on the endocortical bone surface of the vehicle-treated animal (Fig. 6A-a,b) compared with that in the ODN 30 mg/kg–treated animal (Fig. 6A-c,d).
Short-term endocortical histomorphometric measurements based upon a 15-day tetracycline labeling interval indicated that ODN treatment had no effects on endocortical bone formation parameters (data not shown). LT Ec bone formation was measured between the calcein and tetracycline labels (292 days) as described in Materials and Methods. Whereas ODN treatment significantly reduced trabecular bone formation as observed in the intertrochanteric region, ODN 30 mg/kg had no effects on the LT Ec bone formation rate in the proximal femur (Fig. 6C-c). Interestingly, treatment with ODN 6 mg/kg for 21 months increased LT Ec.MS/BS by ∼twofold (p = 0.05) higher than in OVX-vehicle animals (Fig. 6C-a).
Table 2 summarizes the static morphometric data as determined from an endocortical segment within the intertrochanteric region of the PF using the 6-µm thin sections. There were no significant differences in parameters of Ec.Oc.N/BS, Ec.OcS/BS, and Ec.Ob.S/BS in the endocortical surfaces of the ODN-treated groups compared with that of the OVX-vehicle group (Table 2).
Odanacatib stimulates periosteal bone formation at the hip
Representative images of periosteal bone formation in the subtrochanteric region are shown for the vehicle-treated animal (Fig. 6B-a,b) compared with that in the ODN 30 mg/kg–treated animal (Fig. 6B-c,d). Limited mineralized surface was detected by either calcein labeling (in green) or tetracycline labeling (in yellow) at the PF periosteal surface of the vehicle-treated OVX monkeys (Fig. 6B-a,b). In the ODN 30 mg/kg–treated group, significant increases in mineralized surface as well as mineral apposition rate are detected by either fluorescent-labeled agent on the periosteal surface (Fig. 6B-c,d).
Measurements of tetracycline utilizing the short-term labeling interval (15 days) showed that ODN at 30 mg/kg trended to increase periosteal MS/BS (Ps.MS/BS) by ∼1.7-fold (p < 0.07) and Ps.BFR/BS by ∼threefold (p < 0.06) compared with OVX-vehicle animals (data not shown). There were no differences with the ODN 6 mg/kg dose (data not shown). Proximal femoral Ps.MAR was not affected by either dose of ODN (data not shown). On the other hand, long-term measurements of calcein-tetracycline labeling for 292 days clearly showed that ODN treatment for 21 months increased periosteal bone formation in the PF cortical shaft in a dose-dependent manner (Fig. 6D-a–c). ODN treatment trended to increase LT Ps.MS/BS by up to 49% in the PF (Fig. 6D-a). Although not significant, ODN at 6 mg/kg increased LT Ps.MAR by 1.5-fold and Ps.BFR/BS by ∼twofold versus OVX-vehicle (Fig. 6D-b,c). ODN 30 mg/kg significantly stimulated long-term Ps.MAR by ∼3.5-fold (p < 0.05) and Ps.BFR/BS by ∼sixfold (p < 0.05) compared with OVX-vehicle (Fig. 6D-c).
Fig. 7A shows the incorporation of both calcein and tetracycline into newly mineralizing bone on the periosteal surface of the FN. Vehicle-treated OVX monkeys exhibit very little overlapping regions of calcein (in green) and tetracycline (in yellow) labels on the periosteal surface of the FN (Fig. 7A-a). In monkeys treated with ODN 30 mg/kg, visible increases in both calcein as well as tetracycline-labeled surfaces and overlapping regions of both labels were detected at the periosteal surface of the FN (Fig. 7A-b).
There were no significant changes in periosteal dynamic histomorphometric parameters detected by the short-term tetracycline labeling of the FN from ODN-treated groups versus OVX-vehicle (data not shown). On the other hand, both doses of ODN numerically elevated LT Ps.MAR by ∼1.5-fold (Fig. 7B-b). ODN at 6 mg/kg trended to increase LT Ps.MS/BS and Ps.BFR/BS versus OVX-vehicle by 58% and 57%, respectively (Fig. 7B-a,c). ODN at 30 mg/kg numerically increased LT Ps.MS/BS (p < 0.07) by ∼twofold and significantly increased Ps.BFR/BS (p < 0.01) by 3.5-fold compared with OVX-vehicle (Fig. 7B-a,c).
Odanacatib increases cortical dimension
ODN treatment of OVX-NHP for 21 months did not reduce endocortical bone formation and stimulated long-term periosteal bone formation in the PF and FN. Thus, we evaluated the treatment effects of this drug on hip cortical dimension, including cortical thickness and cortical area versus bone marrow area. Measurements of cortical thickness were determined from thick cross sections of PF shaft and FN as described in Materials and Methods. Cross-sectional analysis of the absolute measurements of the ex vivo femoral cortical dimensions were highly varied amongst the study groups because of the broad range in size of the animals recruited to this study.
Treatment with ODN resulted in dose-related increases in cortical thickness of the PF shaft (Fig. 8A) and FN cortex (Fig. 8B). Significant increases in cortical thickness of the PF by 19% (p < 0.05) were detected with ODN 30 mg/kg compared with OVX-vehicle. ODN dose-dependently increased PF cortical area (Ct.Ar) relative to total tissue area (T.Ar) by 12% (p < 0.01) with the ODN 30 mg/kg dose compared with OVX-vehicle. Conversely, ODN also dose-dependently decreased marrow area (Ma.Ar/T.Ar) by 31% (p < 0.01) with the high dose of ODN versus vehicle control. In the FN, increases in cortical thickness of 21% (p = 0.08) were observed with the ODN 30 mg/kg dose compared with OVX-vehicle. However, only trends of treatment-related increases in cortical area compared with vehicle control were observed (data not shown). We also made an attempt to determine periosteal and endosteal perimeters from the PF and FN regions of the ODN-treated groups versus vehicle control. However, because of the great limitation of sample numbers (n = 8 to 11/group), there were no significant treatment effects on these parameters (data not shown).
- Top of page
- Materials and Methods
The primary objective of this study was to assess the effects of 21 months of treatment with the CatK inhibitor odanacatib on bone turnover, BMD, and bone strength in multiple bone sites in newly OVX adult rhesus monkeys. The efficacy of ODN on the lumbar spine in these monkeys has been previously reported.28 Here, we report our findings in the hip and the segmented proximal femur and femoral neck region from the same animals. The study subjects received a daily dosing regimen of either vehicle (0.5% methocel/0.1% SDS) or ODN at 6 mg/kg (2.2 µM/24 h) or 30 mg/kg (4.2 µM/24 h), as compared with the estimated daily exposure in patients receiving the 50-mg weekly dose (6 to 7 µM/24 h) in the phase III clinical trial. In OVX monkeys, ODN 6 and 30 mg/kg ODN significantly increased total hip BMD compared with the vehicle-treated group to levels comparable to or higher than in the intact group. Although a significant difference in BMD was found between intact and OVX-vehicle 21 months postsurgery, there was no significant absolute loss of BMD in the OVX-vehicle compared with baseline. A similar observation on the failure to lose lumbar spine BMD after ovariectomy in these monkeys was previously reported.28 A prestudy alteration in diet and physical activity may explain the BMD gain at the femoral sites of intact and OVX monkeys. These doses of ODN were demonstrated to effectively suppress bone turnover markers and increased the lumbar spine bone mass of newly OVX monkeys.28 Because of substantial anatomic variability in primate FN geometry combined with a complex architecture,33, 34 it is therefore not surprising that the FN BMD values were more variable than the larger total hip region. Nonetheless, FN BMD in the ODN-treated groups was consistently higher than that in the vehicle-treated group, confirming that daily exposure of ODN even at 2.2 µM/24 h prevented the development of relative osteopenia in the hip region of the newly OVX-NHP.
Although there was a dose-dependent increase in hip aBMD in the ODN-treated groups, there was only a numerical treatment-related effect on FN aBMD (p = 0.07). In aged rhesus monkeys, the FN proved to be a difficult site for consistent positioning for in-life DXA scans, greatly limiting a treatment-related increase in BMD at this bone site. In addition, study groups were randomized based on lumbar spine BMD of monkeys varying broadly in size at baseline. Furthermore, the small group size (n = 8 to 11/group) used for biomechanical strength testing at termination of the study may be inadequately powered for this cross-sectional analysis. Taken together, these limitations may be responsible for the apparent lack of treatment effects on FN BMD versus bone strength endpoints. Hence, we combined data from all groups for regression analysis and concluded that although fully effective in increasing hip bone mass in OVX monkeys, ODN maintained normal hip bone biomechanical properties post-treatment for 21 months. Consistent with the treatment effects on lumbar vertebral body strength as previously reported from this study,28 ODN-treated animals had higher ultimate load in the FN as determined by the shear biomechanical test.35 The relationship of ultimate load at the termination of the study to BMD and BMC measured by in-life DXA at 20 months was positive and highly correlated, indicating that normal bone quality was maintained by ODN during the 21 months of dosing. In addition, good correlation between stiffness and aBMD indicated that continuous treatment of ODN did not decrease the biomechanical strength of the femoral neck. Treatment effects of ODN on microarchitecture and strength of the distal radius from the ODN-treated OVX monkeys were also quantified using high-resolution peripheral QCT and compression loading configuration finite element analysis (FEA).36 The results from this study on measurable differences observed in FEA-estimated ultra distal radial bone strength between the ODN-treated and vehicle-treated animals will be reported elsewhere.
In skeletally mature newly OVX monkeys, we provide evidence that ODN treatment increased cortical thickness in the FN and PF by blocking bone resorption without reducing bone formation. ODN also stimulated periosteal bone formation in the hip at a higher rate than that in the vehicle-treated OVX monkeys. As a result, the treatment-related increases of cortical thickness in FN and PF may be responsible for the improved hip strength. Despite the broad body weight range of the monkeys recruited at baseline and the relatively small study size, treatment effects on FN strength are substantiated by the positive correlation of FN ultimate load with cortical thickness at this bone site.
The effects of ODN on bone formation are different depending on bone envelopes of the PF and FN. ODN reduced trabecular and intracortical bone formation, whereas bone formation was unchanged at the endocortical surface and increased at the periosteal surface in skeletally mature primates. The finding that ODN dose-dependently reduced trabecular bone formation in the FN and PF is consistent with its effect at the trabecular surfaces of transilial biopsies and lumbar vertebrae from the same animals.28 From the biopsies collected from the same treatment cohorts, we previously demonstrated that ODN at 6 mg/kg and 30 mg/kg doses indeed brought trabecular bone formation rate in OVX monkeys down to similar levels as found in the intact controls from the same study.28 More important, ODN at these doses also reduced trabecular bone formation rate of the FN and PF in OVX monkeys to levels comparable to those previously published for intact rhesus monkeys of a similar age range.31, 37 We, therefore, suggest that treatment with ODN normalized the trabecular bone formation rate in estrogen-deficient animals to the level of intact animals. Our finding is also consistent with an earlier study, in which another CatK inhibitor, relacatib, significantly reduced indices of bone resorption and formation at FN trabecular sites, as did alendronate, in OVX cynomolgus monkeys.25 The bone surface–specific action of ODN in reducing trabecular bone formation may be unique to NHP. In both CatK knockout mice18 and OVX rabbits treated with CatK inhibitors,16 the bone formation rate is elevated in both trabecular and cortical sites. Furthermore, from a few iliac crest biopsies collected in the ODN phase II clinical study, no treatment-related reduction in trabecular activation frequency was observed in the patients.38
In the same femoral bone sites of the OVX monkeys, histomorphometric evidence in the endocortical envelopes of the FN and PF clearly demonstrated that ODN did not affect endocortical mineralized surface with the 6 mg/kg ODN dose. Again, these findings are consistent with previous reports of other CatK inhibitors, balicatib and relacatib, in OVX cynomolgus monkeys,25, 39 suggesting that CatK may participate in the OC-derived coupling signals to promote bone formation. The effects of ODN on intracortical bone formation were also examined in the proximal femurs from these animals. In humans and other large animal species, cortical bone also undergoes Haversian remodeling.40 This process serves to continually rejuvenate bone mineral and optimize bone architecture to maintain normal skeletal competence.2 The amount of normal adult skeleton undergoing intracortical remodeling at any point in time was estimated to be about 5%.41 As has been demonstrated consistently in preclinical models, bisphosphonates reduce intracortical remodeling in multiple species.42 Similarly, in the proximal shaft of these skeletally mature newly OVX monkeys, we provide evidence that ODN treatment at the high dose reduced intracortical remodeling. Further detailed analysis on the potential impact of ODN treatment-related effects on cortical porosity in OVX nonhuman primates (NHP) is currently under way.
Because remodeling occurs in discrete BMUs in which OC bone resorption is in close proximity to formation,43 some have proposed that nonresorbing OCs generate the coupling signals for bone anabolic activity.44 A number of osteopetrotic mutations in mice, including the CatK−/− mice, result in decreased resorption but demonstrate elevated OC numbers and bone formation.44 However, other OC-poor osteopetrotic mutations, as in c-fms−/− mice, develop defective bone formation in correlation with the reduction of OC number.44 The genetic findings are supported by pharmacological evidence. Treatment with anti-receptor activator of NF-κB ligand (RANKL) antibody or bisphosphonates results in decreased bone formation, secondary to the reduction of bone resorption.45 These therapies lead to greatly diminished OC number, perturb OC metabolism, or induce apoptosis.45, 46 In this study, treatment with ODN did not reduce OC number at any bone sites. Interestingly, although the number of OCs was elevated ∼twofold at the trabecular surface of the PF in ODN groups, this treatment reduced histomorphometric indices of bone formation at this surface. Meanwhile, OC numbers were relatively unchanged compared with the vehicle group at the endocortical surface at the same bone site; the bone formation rate was elevated by ODN treatment. Thus with pharmacological inhibition of CatK, an increase in number of nonresorbing OCs may be in part responsible for the coupling-mediated bone formation with this mechanism.
Modeling occurs primarily on the periosteal surfaces and is responsible for the large radial increases in bone size that occur during growth. In adults, periosteal bone formation persists, albeit at an extremely low rate. Depending upon the cortical site, it has been suggested that the periosteal envelope experiences remodeling instead of modeling.9 Estrogen deficiency leads to cortical thinning resulting from increased endocortical resorption and loss of trabecular bone owing to increased remodeling activity that proceeds with a negative balance at the BMU level, leading to a compensatory increase in periosteal formation that may be an adaptive response to maintain resistance to bending.47 Indeed, by long-term fluorochrome labeling, we observed small but distinct increased periosteal bone formation in the FN and PF regions of the vehicle-treated estrogen-deficient monkeys. Similarly, others have reported a distinct increase (two- to threefold) in periosteal bone formation in OVX animals compared with intact skeletally mature rhesus monkeys.31, 37 However, in this study, treatment with ODN significantly and dose-dependently increased bone formation rate even above that in the vehicle-treated OVX group at the periosteal surfaces at these femoral sites. Recently, balicatib, another CatK inhibitor, was tested in OVX cynomolgus monkeys for 18 months.39 Most histomorphometric indices of bone turnover in vertebra and FN were reported to be significantly lower in the balicatib-treated groups than that in the vehicle controls. However, an unexpected increase in the periosteal bone formation rate in mid-femur was also reported in the treatment group. The above findings with two different CatK inhibitors indicate that the enhanced periosteal bone formation is mechanism based. Thus far, periosteal bone formation has not been reported in CatK knockout mice or in OVX rabbits treated with the CatK inhibitors; this mechanism may also be unique for the OVX-NHP model of osteopenia.
Curiously, in the NHP, periosteal bone turnover was also reported to correlate with an increase in the numbers of Trap(+) OC-like cells on the periosteal surface in the FN of sex steroid–deficient animals.37 Interestingly, periosteal osteoclastic activity did not appear to derive from intracortical tunneling at the bone surface but rather through recruitment from the circulation.37 In mice, these periosteal Trap(+) cells were thought to be bone-specific macrophages or osteomac.48 However, the nature and expression of CatK has not been confirmed for the recently identified periosteal macrophages.48 It could be speculated that the potential signals for CatK-mediated periosteal bone formation may derive from these cells because CatK expression has been documented in macrophages.49 CatK expression has also been reported in human osteoblasts, bone lining cells, and some osteocytes in situ.50 Because CatK expression in these cells is low relative to that in OCs and because of the expression of other lysosomal cathepsins, the role of CatK in osteoblasts or osteocytes remains unclear. More recently, OC-specific deletion of CatK in mice using the Mx1-cre was reported to recapitulate the high bone mass and high bone formation phenotype seen in the global CatK−/− mice.51 Furthermore, mice with osteoblast deletion of CatK using the Osx-cre had no phenotype.51 Although the potential direct effect of ODN on the osteocytes cannot be ruled out, the above genetic evidence indicates that CatK-mediated signaling leading to stimulation of bone formation derives from OCs during the remodeling process.
In the OVX-NHP, the reasons for the differential effect of ODN on bone remodeling and modeling of the trabecular, endocortical, intracortical, and periosteal envelopes remain unclear. There is conflicting evidence as to whether bone resorption and formation are coupled in a bone site-specific manner.52, 53 Traditionally, bone remodeling primarily occurs in the trabecular and Haversian envelopes, where resorption leaves an eroded, scalloped bone surface upon which bone formation ensues, thus producing cement lines with a scalloped appearance.41 Bone modeling occurs primarily in the cortical periosteal envelope, whereas both bone modeling and remodeling occur in the endocortical envelope.53 Because no bone resorption precedes bone formation in bone-forming modeling sites, the cement lines in bone modeling sites are typically smooth.41 Thus, in monkeys, a possible explanation for this observation is that decreased trabecular and intracortical bone formation by ODN is the result of suppressed bone remodeling via inhibiting osteoclast activity. Increased periosteal bone formation suggests that ODN could stimulate modeling-based bone formation. In the endocortical envelope, ODN treatment could have mixed effects on modeling and remodeling activities. To better our understandings of the cellular mechanisms of ODN on the coupling of bone resorption to bone formation, we are currently examining the bone site-specific effects of this CatK inhibitor on histomorphometric modeling versus remodeling cement lines in the trabecular and endocortical envelopes.
In summary, this study assessed the long-term effects of the CatK inhibitor ODN on BMD, strength, and bone formation of trabecular versus cortical surfaces in the hip of the OVX rhesus monkey, a relevant animal model for human osteoporosis. The results indicate that ODN increased bone mass in the total hip and FN. Furthermore, long-term treatment with ODN improved FN biomechanical properties, associated with a treatment-related increase of cortical thickness. This therapy appears not only to positively impact hip strength by increasing total hip BMD but also to stimulate bone formation at periosteal surfaces through a unique mechanism. ODN treatment effectively reduced remodeling in trabecular and intracortical bone surfaces and, at the same time, maintained endocortical bone formation and stimulated periosteal bone formation. Our findings demonstrate that the CatK inhibitor ODN both inhibits bone loss in trabecular bone regions and builds cortical bone in the hip of estrogen-deficient NHPs, via a molecular mechanism distinct from the bisphosphonates.
- Top of page
- Materials and Methods
TC, CMC, BLP, MP, BBS, and LTD are employees of Merck, the study's sponsor, and may own stock/stock options in the company. DK was an employee of Merck during the conduct of this study.
- Top of page
- Materials and Methods
We thank the excellent veterinary and technical staff at the New Iberia Research Center, Kenneth Lodge and Cory Merschman for assisting with specimen collection, and Pat Masarachia for her early efforts on the study. We also thank Susan Smith and Rana Samadfam at Charles River Laboratories, Preclinical Services, Montreal, Canada, for managing the bone-strength testing.
Authors' roles: Study design: TC, CNC, BLP, MP, DBK, LTD. Study conduct: TC, CNC, BLP, MP. Data collection: TC, CNC, BLP, MP. Data analysis: TC, CNC, BLP, MP, DBK, LTD. Data interpretation: TC, CNC, BLP, MP, DBK, BBS, LTD. Drafting manuscript: TC, CNC, BLP, MP, DBK, BBS, LTD. Revising manuscript content: TC, CNC, BLP, MP, DBK, BBS, LTD. Approving final version of manuscript: TC, CNC, BLP, MP, DBK, BBS, LTD. LTD takes responsibility for the integrity of the data analysis.
- Top of page
- Materials and Methods
- 2Skeletal heterogeneity and the purposes of bone remodeling. In: Marcus R, Feldman B, Kelsey J, editors. Osteoporosis. San Diego: Academic Press; 1996. p. 315– 9.
- 20Cathepsin K—a new molecular target for osteoporosis. IBMS BoneKey. 2008; 5: 16– 24.,
- 28Odanacatib reduces bone turnover and increases bone mass in the lumbar spine of skeletally mature ovariectomized rhesus monkeys. J Bone Miner Res. 2011 Nov 23 [Epub ahead of print]., , , , , , , , , ,
- 35Treatment of skeletally mature ovariectomized rhesus monkeys with PTH(1-84) for 16 months increases bone formation and density and improves trabecular architecture and biomechanical properties at the lumbar spine. J Bone Miner Res. 2007; 22: 260– 73., , , , ,
- 36HRpQCT based finite element analysis to estimate bone strength at distal radius in odanacatib treated rhesus monkeys. J Bone Miner Res. 2009; 24(Suppl 1). Available at http://www.asbmr.org/Meetings/AnnualMeeting/AbstractDetail.aspx?aid=3b43a0ef-2069-4772-8499-484fef7e6288. Accessed December 12th 2011., , , , , ,
- 38Odanacatib, a cathepsin-K inhibitor for osteoporosis: a two-year study in postmenopausal women with low bone density. J Bone Miner Res. 2010; 25: 937– 47., , , , , , , , ,
- 39Balicatib, a cathepsin K inhibitor, stimulates periosteal bone formation in monkeys. Osteoporos Int. 2011 Feb 10 [Epub ahead of print]., ,
- 40Overview: animal models of osteopenia and osteoporosis. J Musculoskelet Neuronal Interact. 2011; 1: 193– 207.,
- 41The physiological and clinical significance of bone histomorphometric data. In: Recker RR, editor. Bone histomorphometry. Techniques and interpretations. Boca Raton (FL): CRC; 1983. p. 143– 223.
- 42Skeletal microdamage: less about biomechanics and more about remodelling. Clinic Rev Bone Miner Metab. 2008; 6: 24– 30.,
- 51Disruption of cathepsin K in the osteoclast lineage increased bone formation through coupling-dependent mechanism. J Bone Miner Res. 2009; 24(Suppl 1). Available at http://www.asbmr.org/Meetings/AnnualMeeting/AbstractDetail.aspx?aid=94c9aecc-91f4-4549-8e7e-3bf7aa23dd91. Accessed December 12th 2011., , , , , , ,