Original Article

Effects of Denosumab, Alendronate, or Denosumab Following Alendronate on Bone Turnover, Calcium Homeostasis, Bone Mass and Bone Strength in Ovariectomized Cynomolgus Monkeys† Paul J. Kostenuik, PhD1*; Susan Y. Smith, PhD2; Rana Samadfam, PhD2, Jacquelin Jolette, DVM2; Lei Zhou, PhD3; and Michael S. Ominsky, PhD1

1

Metabolic Disorders Research, and 3Biostatistics, Amgen Inc., 1 Amgen Center Drive, Thousand Oaks, CA

91320, USA 2

Musculoskeletal Research, Charles River Laboratories, Preclinical Services Montreal,

22022 Transcanadienne, Senneville, Quebec, Canada H9X 3R3

* Corresponding Author: Paul J. Kostenuik One Amgen Center Drive M/S 15-2-A Thousand Oaks, CA 91320 Phone: +1 (805) 447-5585 Email: [email protected] Keywords: Bone histomorphometry < ANALYSIS/QUANTITATION OF BONE, DXA < ANALYSIS/QUANTITATION OF BONE, Preclinical Studies < ANIMAL MODELS, Osteoporosis < DISEASES AND DISORDERS OF/RELATED TO BONE, Biomechanics < ORTHOPAEDICS

†

This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: [10.1002/jbmr.2401]

Additional Supporting Information may be found in the online version of this article. Initial Date Submitted August 11, 2014; Date Revision Submitted October 28, 2014; Date Final Disposition Set October 30, 2014

Journal of Bone and Mineral Research © 2014 American Society for Bone and Mineral Research DOI 10.1002/jbmr.2401

 

 

ABSTRACT Postmenopausal osteoporosis is a chronic disease wherein increased bone remodeling reduces bone mass and bone strength. Antiresorptive agents including bisphosphonates are commonly used to mitigate bone loss and fracture risk. Osteoclast inhibition via denosumab (DMAb), a RANKL inhibitor, is a newer approach for reducing fracture risk in patients at increased risk for fracture. The safety of transitioning from bisphosphonate therapy (alendronate; ALN) to DMAb was examined in mature ovariectomized (OVX) cynomolgus monkeys (cynos). One day after OVX, cynos (7-10/group) were treated with vehicle (VEH, s.c.), ALN (50 µg/kg, i.v., twice monthly) or DMAb (25 mg/kg/month, s.c.) for 12 months. Other animals received VEH or ALN for 6 months and then transitioned to 6 months of DMAb. DMAb caused significantly greater reductions in serum CTx than ALN, and transition from ALN to DMAb caused further reductions relative to continued ALN. DMAb and ALN decreased serum calcium (Ca), and transition from ALN to DMAb resulted in a lesser decline in Ca relative to DMAb or VEH-DMAb transition. Bone histomorphometry indicated significantly reduced trabecular and cortical remodeling with DMAb or ALN. Compared with ALN, DMAb caused greater reductions in osteoclast surface, eroded surface, cortical porosity and fluorochrome labeling, and transition from ALN to DMAb reduced these parameters relative to continued ALN. Bone mineral density increased in all active treatment groups relative to VEH controls. Destructive biomechanical testing revealed significantly greater vertebral strength in all three groups receiving DMAb, including those receiving DMAb after ALN, relative to VEH controls. Bone mass and strength remained highly correlated in all groups at all tested skeletal sites, consistent with normal bone quality. These data indicate that cynos transitioned from ALN to DMAb exhibited reduced bone resorption and cortical porosity, and increased BMD and bone strength, without deleterious effects on Ca homeostasis or bone quality.

 

INTRODUCTION Denosumab (DMAb), a fully human monoclonal antibody that inhibits RANKL, is a potent antiresorptive agent shown in clinical trials to reduce fracture risk in men and women with osteoporosis (1;2). DMAb inhibits the formation, activation and survival of osteoclasts, (3) which rapidly inhibits the activity of mature osteoclasts while reducing the creation of new resorption cavities (4). These effects, combined with continued refilling of existing resorption spaces, leads to improvements in bone mass and architecture that can explain much of the increased bone strength achieved with DMAb (5;6). Bisphosphonates (BPs) also increase bone mass and reduce fracture risk in osteoporosis patients, (7;8) primarily by inhibiting mature osteoclasts. The general mechanism of action of BPs involves their adsorption to bone surfaces, followed by their uptake by bone-resorbing osteoclasts that then become inactivated through the inhibition of certain enzymes essential for cellular metabolism (9;10). Clinical trials of postmenopausal women with low bone mass showed that DMAb caused greater reductions in bone resorption and greater increases in bone mineral density (BMD) relative to the BPs alendronate (ALN), ibandronate or risedronate (11-13). Greater effects of DMAb on cortical bone density and cortical porosity versus BPs may contribute to these differences (14). Recent clinical trial data in osteoporosis patients indicated that DMAb caused a greater reduction in fracture risk relative to ALN (15). Clinical trials also showed further increases in BMD and decreases in bone resorption after transitioning postmenopausal women from BPs to DMAb (11;16). It has not been shown that transition from BPs to DMAb reduces fracture risk relative to continued BP therapy, but due to factors such as poor compliance, inadequate BMD responses, tolerability or other reasons, some osteoporosis patients discontinue BP therapy despite remaining at increased risk of fracture (17). For those patients who switch from a BP to DMAb, it is relevant to understand potential safety issues involved in this therapeutic transition. The current study examined the safety and pharmacodynamic profiles of adult OVX cynos transitioned from ALN to DMAb. In the interest of safety assessments, DMAb was administered at a ~25-fold multiple of the weight-based dose approved for osteoporosis patients, on a monthly rather than twice-yearly schedule. One safety concern with potent antiresorptives is hypocalcemia. Osteoclast inhibition with DMAb leads to reduced skeletal calcium (Ca) mobilization (3), which usually has no clinically meaningful effect on serum Ca concentration, presumably due to a parathyroid hormone (PTH) response (6) that may lead to increases in intestinal  

Ca absorption and renal Ca reabsorption. DMAb is associated with an increased risk of hypocalcemia in individuals for whom Ca demands cannot be met via these adaptive mechanisms, including patients with severe renal impairment (18). BPs including ALN also reduce Ca mobilization from bone as they inhibit resorption, which can lead to hypocalcemia in individuals with additional risk factors (19). The current study provides the first detailed description of Ca responses after DMAb initiation in BP-treated versus treatment-naïve animals. These data provided insights into skeletal attributes that may influence reductions in serum Ca after DMAb. Another potential safety concern with potent antiresorptives is that remodeling inhibition may impair bone matrix quality, leading to compromised bone strength and fractures. Rigorous assessments of such outcomes upon transitioning from a BP to DMAb may be impractical in a clinical setting, whereas animal studies allow destructive bone strength testing that could identify impaired bone quality before fractures occur. The current study evaluated bone strength and bone quality after transitioning from ALN to DMAb, with a study design that allowed the first direct comparison of bone biomechanical properties in OVX animals treated with DMAb versus a BP.

MATERIALS AND METHODS All animal procedures and activities were approved by and performed in an AAALAC-accredited facility. The study was performed under Good Laboratory Practice (GLP) conditions according to the protocol and consistent with standard operating procedures. Animals and experimental design Fifty-two adult female cynomolgus monkeys (Macaca fascicularis), aged 9-14 years on arrival, were received from Mauritius Island and were cared for in accordance with established guidelines (20). Animals were housed two per cage at an indoor AAALAC-accredited facility in stainless steel cages equipped with automatic watering provided ad libitum. Animals were provided Certified Hi-Fiber Primate Diet 5K91 containing 0.91% Ca, 0.55% phosphorus, and 6.6 IU vitamin D/g (PMI Nutrition International Inc., Shoreville, MN, USA), twice daily except during designated procedures. Environmental enrichment included food supplements (PrimaTreat, Prima-Foraging Crumbles and fruits or vegetables). After 6 weeks of acclimation, including health examinations and baseline densitometry, animals (10-11/group) were allocated to five treatment groups balanced for body weight, whole body bone mineral content (BMC) and lumbar spine BMD. All animals underwent OVX at the end of acclimation as previously described (5), and treatments commenced the next day. One group (VEH; n=10) received the DMAb  

vehicle (10 mM sodium acetate, 5% sorbitol, pH 5.2) by s.c. injection (0.71 mL/kg) once monthly for 12 months. A second group received s.c. 25 mg/kg DMAb (n=11) once monthly for 12 months, based on evidence that this regimen potently inhibited bone resorption for this duration in this species (5). A third group received 50 µg/kg of ALN (Sigma; n=10) twice monthly for 12 months by i.v. injection (1 mL/kg), a regimen that inhibited bone resorption and increased BMD in OVX cynos and baboons (21;22). A fourth group (VEH-DMAb; n=10) received 6 months of VEH followed by 6 months of DMAb, and the fifth group (ALN-DMAb; n=11) received 6 months of ALN followed by 6 months of DMAb, at the aforementioned doses. Animals not receiving ALN received ALN vehicle (Ca- and magnesium-free PBS; Mediatech Inc., Cellgro, Manassas, VA, USA) by i.v. injection twice monthly. Animals were sacrificed at month 12 and necropsied under a pathologist’s supervision, which included an external and detailed internal examination. Blood and serum analyses Animals were fasted overnight and administered glycopyrrolate and ketamine HCl (i.m.) prior to blood collections. Blood was collected 1 day prior to dosing at months 0, 1, 2, 3, 6, 9 and 12. Blood was also collected 1, 3, 7 and 14 days after the first dose (month 0) and 1, 3, 7, 14 and 42 days after the transition (month 6). Ctelopeptide of type I collagen (CTx; Serum CrosslapsTM ELISA, Nordic Biosciences, Denmark) and bone-specific alkaline phosphatase (BSAP; MetraTM BAP ELISA, Quidel, San Diego CA) were measured from prepared serum at certain time points. Declines in estrogen after OVX were confirmed by measuring serum estradiol (Ultrasensitive Estradiol RIA, Diagnostics Systems Laboratories Inc, Webster, TX). Serum was also assayed for intact PTH (Coat-A-Count Intact PTH IRMA, Diagnostic Products Corp, Siemens, Los Angeles CA), and for total Ca and phosphorus (Modular Analytics, Roche Diagnostics, Mannheim Germany). Serum concentrations of DMAb and anti-DMAb binding and neutralizing antibodies were monitored at Amgen Inc. (Thousand Oaks, CA), as previously described (5). Bone histopathology and histomorphometry Fluorochrome labels were injected i.v. 15 and 5 days prior to bone collections; tetracycline (30 mg/kg) prior to month 6 bone biopsies, and bicarbonate-buffered calcein solution (8 mg/kg) prior to necropsy. Ilium and rib biopsies were collected after month 6. The contralateral ilium and rib, the second lumbar vertebral body (L2), and the right tibial proximal end and diaphysis were collected at necropsy. All tissues were fixed in 10% neutralbuffered formalin and transferred to 70% ethanol, and trimmed samples were embedded in methyl-methacrylate.  

Cancellous bone sections from the proximal tibia (frontal), L2 (sagittal), and ilia were left unstained (7 µm thick) or were stained with toluidine blue or Goldner’s Trichrome (5 µm thick) for histomorphometric evaluation. Unstained transverse cortical sections (ground to 20-40 µm thickness) were evaluated from the tibial diaphysis at the mid-shaft and ribs. Histomorphometry analyses were conducted using a Bioquant/TCW image analyzer (R and M Biometrics; Nashville, TN) linked to a microscope with bright and epifluorescence illumination. Regions of interest (ROIs) for cancellous bone at L2 and ilia were as previously described (6). The ROI in the proximal tibial metaphysis was a 6 x 5 mm rectangle starting 2 mm distal from the physeal plate. Standard nomenclature was used as previously recommended (23), and other details for specific parameters were previously described (6). When mineral apposition in cortical or cancellous compartments could not be measured, the variables were reported as 0.30 µm/day and 0.24 µm/day, respectively. These values represent the smallest measurable linear distance, with a correction factor of 4/pi applied to cancellous bone to account for obliquity. Bone densitometry by DXA and pQCT Areal BMD (aBMD) was measured in vivo at the lumbar spine (L1-L4; A/P), right proximal femur, right distal radius and whole body by DXA (Hologic Discovery A Bone Densitometer, Hologic, Inc. Bedford, MA) prior to OVX and once following doses 3, 6 and 12. Appropriate clinically-based scan modes were used for each site, with scanner settings reported elsewhere (5). Ex vivo DXA was used to determine bone area, BMC and BMD of trimmed L3 and L4 vertebral bodies, L5 and L6 vertebral cancellous cores, and the left femur diaphysis (point resolutions of 0.0311, 0.0311 and 0.0901 cm, respectively). Peripheral quantitative computed tomography (pQCT) was used for in vivo assessments of the right distal radius metaphysis (XCT Research SA+ Bone Scanner; software version 5.50D, Stratech). Data were generated as an average of three sequential scans of 0.5 mm thickness, taken 0.5 mm apart, using contour mode 2, peelmode 2 and cortmode 2 (threshold 0.930 cm-1), with a nominal in-plane voxel size of 0.2 mm. A single scan was also taken ex vivo of the L3-L4 vertebral bodies and of the left femur diaphysis at the site of loads applied during biomechanical testing. Bone strength Lumbar vertebrae (L3-L6) and left femurs were stored at –20°C prior to biomechanical testing. Vertebrae were trimmed to produce L3-L4 vertebral bodies and L5-L6 cancellous cores as previously described (5). Destructive strength testing was performed with an MTS 858 Mini Bionix Servohydraulic test system, model 242.03 (MTS, Eden Prairie, MN), using TestWorks® version 3.8A for TestStarII™ software version 4.0C. The left femur  

diaphysis was tested to failure in three-point bending (50 mm span length), and the proximal femur segment was then used to test the femur neck in shear to failure in a simulated single-legged stance. Femur tests used a displacement rate of 1 mm/s. L3-L4 bodies and L5-L6 cores were tested to failure in compression at a displacement rate of 20 mm/min. Averaged values were calculated for L3 and L4 bodies, and for L5 and L6 cores. Load-displacement curves for peak load, stiffness, and energy (to failure for the femur diaphysis, and to peak load for vertebra), and for estimated material properties (ultimate strength, elastic modulus and toughness), were analyzed as previously described (5). Apparent toughness of L3-L4 bodies and L5-L6 cores was calculated as energy to peak load divided by specimen height multiplied by cross-sectional area. Normalized vertebral toughness was calculated as apparent toughness divided by L2 trabecular bone volume fraction (BVF), as previously described (24). Statistical Methods Data from the first 6 months for the similarly-treated VEH and VEH-DMAb groups were combined into one VEH group, and data from the first 6 months for the ALN and ALN-DMAb groups were combined into one ALN group. Statistical analyses as described below were conducted using release 9.1 of SAS/STAT, and tests were conducted at the 0.05 level of significance. A statistical model using group as a fixed effect was applied to each endpoint at each time point, with no multiplicity correction applied across time points. Levene’s test was applied to assess the variance homogeneity among groups. If the equal variance assumption was violated at α=0.05 significance level for any endpoint at any sampling time point, the same group effect statistical model was applied to the ranks of the responses. Levene’s test was then applied to assess the variance homogeneity among groups for analysis on ranks. If the equal variance assumption was violated at α=0.05 significance level for any endpoint at any sampling point, the Wilcoxon exact test was applied to ascertain treatment effects. Sidak multiple comparison procedure was then used to correct for multiplicity among the inter-group comparisons. Linear regression analyses were performed in GraphPad Prism (v.6.03) to correlate ex vivo BMD or BMC to bone strength parameters at each site.

 

RESULTS General health No animal deaths occurred during the study, and overall there were no safety issues related to treatment administration based on the lack of clinical signs or effects on body weight or appetence. OVX was confirmed in all animals by gross and microscopic reproductive tract examinations and by reductions in serum estradiol (data not shown). Drug exposure Several animals (2 in the DMAb group, 3 in the VEH-DMAb group, 2 in the ALN-DMAB group) developed anti-DMAb antibodies detectable via immunoassays and bioassays. These animals had no detectable DMAb in predose serum samples and exhibited loss of pharmacodynamic effects, and their data were excluded, resulting in final sample sizes of 10, 7, 10, 9 and 9 for the VEH, VEH-DMAb, ALN, ALN-DMAb, and DMAb groups, respectively. Biochemical markers of bone turnover and mineral homeostasis The bone resorption marker serum CTx increased steadily in the VEH control group over the first 6 months after OVX (Fig. 1A). DMAb rapidly and substantially decreased serum CTx starting 1 day after the first dose, an effect that persisted through month 12. ALN also decreased serum CTx from day 1 through month 12, albeit to a lesser degree relative to DMAb. Animals transitioning from either VEH or ALN to DMAb at month 6 exhibited steep and persistent declines in CTx within 1 day of DMAb dosing. The bone formation marker BSAP showed responses to OVX and to treatments that were generally similar to those seen for CTx (Fig. 1B). A normal serum Ca range was calculated based on values from all animals prior to treatment (mean ± 2 SD = 9.0-11.3 mg/dL). Serum Ca increased gradually in the VEH group over the first 6 months (Fig. 2A), consistent with increased bone resorption after OVX. DMAb and ALN administration at month 0 led to transiently decreased serum Ca, with greater and more sustained reductions in the DMAb group. The group transitioned from VEH to DMAb exhibited the greatest reductions in serum Ca, with a nadir (group mean = 8.8 mg/dL) observed 7 days after transition to DMAb and recovery thereafter. One animal in this group exhibited a serum Ca value of 7.7 mg/dL 14 days after the first DMAb dose, with recovery into the normal range thereafter. No other animal had serum Ca