The tyrosine kinase inhibitor dasatinib dysregulates bone remodeling through inhibition of osteoclasts in vivo

Authors

  • Kate Vandyke,

    1. Myeloma Research Laboratory, Department of Haematology, Centre for Cancer Biology, Institute of Medical and Veterinary Science, Adelaide, Australia
    2. School of Medicine, University of Adelaide, Adelaide, Australia
    Search for more papers by this author
  • Andrea L Dewar,

    1. Myeloma Research Laboratory, Department of Haematology, Centre for Cancer Biology, Institute of Medical and Veterinary Science, Adelaide, Australia
    Search for more papers by this author
  • Peter Diamond,

    1. Myeloma Research Laboratory, Department of Haematology, Centre for Cancer Biology, Institute of Medical and Veterinary Science, Adelaide, Australia
    Search for more papers by this author
  • Stephen Fitter,

    1. Myeloma Research Laboratory, Department of Haematology, Centre for Cancer Biology, Institute of Medical and Veterinary Science, Adelaide, Australia
    Search for more papers by this author
  • Christopher G Schultz,

    1. Department of Nuclear Medicine, Royal Adelaide Hospital, Adelaide, Australia
    Search for more papers by this author
  • Natalie A Sims,

    1. St Vincent's Institute, Fitzroy, Melbourne, Australia
    Search for more papers by this author
  • Andrew CW Zannettino

    Corresponding author
    1. Myeloma Research Laboratory, Department of Haematology, Centre for Cancer Biology, Institute of Medical and Veterinary Science, Adelaide, Australia
    2. School of Medicine, University of Adelaide, Adelaide, Australia
    • Myeloma Research Program, Bone and Cancer Research Laboratories, Department of Haematology, Institute of Medical and Veterinary Science, Hanson Institute, GPO Box 14, Adelaide, SA, Australia 5000.
    Search for more papers by this author

Abstract

Dasatinib is a potent tyrosine kinase inhibitor that is used to treat chronic myeloid leukemia in patients resistant or intolerant to imatinib mesylate. While designed to inhibit Abl and Src kinases, dasatinib shows multitarget effects, including inhibition of the macrophage colony-stimulating factor (M-CSF) receptor c-fms. We have shown previously that dasatinib abrogates osteoclast formation and activity in vitro owing, in part, to its specificity for c-fms. In this study we examined whether dasatinib could significantly alter bone volume in a model of physiologic bone turnover. Sprague-Dawley rats were administered dasatinib (5 mg/kg/day) or vehicle by gavage or zoledronic acid (ZOL; 100 µg/kg/6 weeks) subcutaneously. Following 4, 8, and 12 weeks of treatment, serum biochemical, bone morphometric, and histologic analyses were performed. Whole-body bone mineral density and tibial cortical thickness where unchanged in the dasatinib- or ZOL-treated animals relative to controls. However, micro–computed tomographic (µCT) analysis of cancellous bone at the proximal tibias showed that trabecular volume (BV/TV) and thickness (Tb.Th) were increased in dasatinib-treated animals at levels comparable with those of the ZOL-treated group. These changes were associated with a decrease in osteoclast numbers (N.Oc/B.Pm) and surface (Oc.S/BS) and decreased serum levels of the osteoclast marker c-terminal collagen crosslinks (CTX-1). Mineral apposition rate (MAR), bone-formation rate (BFR), and levels of the serum osteoblast markers osteocalcin and N-terminal propeptide of type I procollagen (P1NP) were not altered significantly in the dasatinib-treated animals relative to controls. These studies show that dasatinib increases trabecular bone volume at least in part by inhibiting osteoclast activity, suggesting that dasatinib therapy may result in dysregulated bone remodeling. © 2010 American Society for Bone and Mineral Research

Introduction

Dasatinib (BMS-354825, SPRYCEL; Bristol-Myers Squibb, New York, NY, USA) is an ATP-competitive protein tyrosine kinase inhibitor that at therapeutic concentrations inhibits the activity of Abl, Src-family kinases (c-Src, Lck, Hck, Yes, Fgr, Lyn, and Fyn), and the platelet-derived growth factor (PDGF) family members c-Kit and PDGF receptor α (PDGFR-α) and PDGFR-β.1–4 Dasatinib is clinically active against chronic myeloid leukemia (CML) in patients who are resistant or intolerant to the front-line chemotherapeutic agent imatinib mesylate (STI571,  Glivec/Gleevec; Novartis, Basel, Switzerland)3, 5 owing to its enhanced affinity for the CML oncoprotein BCR-Abl and its insensitivity to mutations in the BCR-Abl kinase domain.6

It has been demonstrated recently that dasatinib also inhibits the macrophage colony-stimulating factor (M-CSF) receptor c-fms at therapeutically achievable concentrations.2, 7, 8 Signaling through c-fms is crucial for the proliferation, survival, and activity of monocyte/macrophage lineage cells, including osteoclasts (OCs). A central role for M-CSF in OC formation and survival has been demonstrated in op/op mice (homozygous for the mutation osteopetrotic) and toothless homozygous (tl/tl) rats, which lack detectable c-fms and M-CSF, respectively.9–12 These animals are severely osteopetrotic owing to a profound decrease in OC numbers resulting from defective OC proliferation and differentiation.9–12

There is growing in vitro evidence that inhibition of c-fms by dasatinib can inhibit OC formation and activity. Treatment with dasatinib at clinically relevant concentrations inhibits the formation and activity of tartrate-resistant acid phosphatase (TRACP)–positive OCs from human peripheral blood mononuclear cells and primary mouse bone marrow cells in vitro.7, 8 It remains to be determined what effect this may have on bone remodeling in patients undergoing dasatinib treatment.

In addition, dasatinib may affect osteoblast (OB) activity through inhibition of PDGFR. PDGF is a potent mitogen for OB precursors that also inhibits OB differentiation in vitro.13–16 Studies from our laboratory17 and those of others18–21 have shown that imatinib activates OB activity while inhibiting cell proliferation at least in part through the inhibition of PDGFR. In cultures of human and murine stromal cells and cell lines, treatment with therapeutically achievable concentrations of imatinib significantly increased mineral deposition17, 19, 21 and decreased cell proliferation.17–20 In addition, imatinib treatment prevented the inhibitory effects of PDGF on mineralized matrix formation in vitro.17, 19 These results suggest that tyrosine kinase inhibitors could inhibit OB proliferation and activate their differentiation in vitro. However, it remains to be seen whether dasatinib has any effects on OB activity.

Since dasatinib is a potent inhibitor of tyrosine kinases that are essential for normal bone metabolism, this study examined whether dasatinib could affect OB and OC activity and thereby modulate bone remodeling in normal mature Sprague-Dawley rats in vivo.

Materials and Methods

Reagents

Dasatinib was provided by Bristol-Myers Squibb . Zoledronic acid was obtained from Novartis. Tissue culture medium, fetal bovine serum, L-glutamine, and N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES) were obtained from SAFC Biosciences (Lenexa, KS, USA). All other reagents were purchased from Sigma-Aldrich (Castle Hill, NSW, Australia) unless otherwise specified.

Establishment of rodent bone marrow cultures

Rodent bone marrow monocyte (rBM) cultures were established as described previously.22 Briefly, 10-week-old female Sprague-Dawley rats (Veterinary Services Division, Institute of Medical and Veterinary Science) were humanely euthanized by CO2 overdose. Using a 21-gauge needle, bone marrow was flushed from femurs using ice-cold α-minimal essential medium (α-MEM).

Cells were washed once in Hank's buffered saline solution (HBSS), triturated thoroughly to obtain a single-cell suspension, and seeded at 9 × 105 cells/cm2. Cells were incubated overnight at 37°C with 5% CO2 to allow stromal cells to adhere. The nonadherent rat bone marrow (rBM) cells were aspirated and used as monocyte/macrophage OC precursor cells in OC formation and activity assays.

OC formation and activity assays

For TRACP staining or calcium phosphate resorption assays, rBM cells were cultured as described previously22 on 96-well plates or on 16-well calcium phosphate–coated quartz slides (BioCoat Osteologic MultiTest Slides, BD Biosciences, North Ryde, NSW, Australia). Briefly, rBM cells were cultured at 3.1 × 105 cells/cm2 in 200 µL α-MEM supplemented with 10% FBS, additives (2 mM L-glutamine, 50 IU/mL of penicillin, 50 µg/mL of streptomycin sulfate, 1 mM sodium pyruvate, 15 mM HEPES buffer), and 75 ng/mL of recombinant human (rh)M-CSF (Chemicon International, Melbourne, VIC, Australia). Following overnight incubation and every 3 days thereafter, the medium was replaced with complete growth medium containing 75 ng/mL of rhM-CSF and 75 ng/mL of soluble rhRANKL (Roche Applied Science, Basel, Switzerland). The formation of OC-like cells and the resorption of calcium phosphate were assessed after a period of 6 and 9 days, respectively. Where indicated, dasatinib was added on day 0 and included for the duration of the experiment.

TRACP stain

Cells were stained for TRACP activity with the Leukocyte Acid Phosphatase (TRACP) Kit (Sigma) as per the manufacturer's recommendations. The number of OCs (TRACP+ cells with three or more nuclei) per well were enumerated in quadruplicate wells. Cells were visualized and photographed with an Olympus CKX41 inverted microscope and an Olympus DP11 digital camera (Olympus, Tokyo, Japan) at ×200 magnification.

von Kossa silver stain

Calcium phosphate–coated slides were stained for mineral using von Kossa silver stain. The slides were washed once in reverse osmosis (RO) water and subsequently bleached in 6% sodium hypochlorite (ACE Chemical Company, Camden Park, SA, Australia) for 5 minutes and washed three times with RO water to remove cell debris. The calcium phosphate layer then was stained with 5% silver nitrate (Rhone Poulenc Chemicals, Clayton South, VIC, Australia) for 30 minutes at room temperature. Slides were washed three times in RO water, and the brown stain was developed in a solution of 4% sodium carbonate (BDH, Kilsyth, VIC, Australia) and 9.25% to 10% formaldehyde (BDH) for 30 to 60 seconds. The slides were washed three times in RO water and then fixed in 5% sodium thiosulphate (Sigma) for 2 minutes, followed by a final wash in RO water. Slides then were photographed with an Olympus CKX41 inverted microscope with an Olympus DP11 digital camera at ×40 final magnification. Resorption area (ie, cleared, unstained foci in the mineral layer) was quantitated with Adobe Photoshop CS3 (San Jose, CA, USA).

Establishment of rodent bone marrow stromal cell cultures

Rodent bone marrow stromal cells (rBMSCs) were isolated from flushed femurs and tibias by crushing the bones and treating the bone chips with collagenase and dispase. Briefly, the bone chips were washed three times with ice-cold HBSS and then incubated with collagenase (1 mg/mL) and dispase (1 mg/mL) in 25 mL of HBSS at 37°C for 10 minutes. The cells recovered were discarded, and the rBMSCs then were isolated from the bone chips by three sequential digestions with collagenase/dispase at 37°C for 20 minutes. The cells were pooled, washed with HBSS, and seeded with the bone chips in α-MEM supplemented with 100 µM ascorbate in a 75-cm2 flask. rBMSCs were passaged twice before use in mineralization, proliferation, and Western blot experiments.

WST-1 assay

The effect of dasatinib on rBMSC survival/proliferation was assessed using WST-1. Briefly, rBMSCs (2.4 × 103 cells per well, cultured in α-MEM + 100 µM ascorbate) were plated in 96-well plates and allowed to adhere overnight. Cells then were treated with dasatinib (2.5 to 80 nM, all containing 0.05% DMSO vehicle) or vehicle alone. At the indicated time points, the relative number of viable, metabolically active cells per well was detected by the quantification of mitochondrial dehydrogenase activity via the formation of formazan from 4-[3-(4-iodophenyl)-2-(4-nitrophenyl)-2H-5-tetrazolio)-1,3-benzene disulphonate (WST-1) following the manufacturer's instructions (Tanaka, Madison, WI, USA).

OB activity assays

For mineralization assays, rBMSCs were cultured on 96-well plates in 200 µL of α-MEM with 10% fetal calf serum (FCS), additives, and ascorbate. Cells were allowed to reach 80% confluence, and then the medium was replaced with mineralization medium containing 10 mM β-glycerophosphate and dasatinib [in 0.05% (v/v) DMSO vehicle] or vehicle alone. Treatment medium was changed twice weekly. After 4 weeks, the mineralized matrix was dissolved in 0.6 M HCl, and calcium levels were quantitated by the cresolphthalein complexone assay (Thermo Electron Corporation, Melbourne, VIC, Australia). Calcium levels were normalized to cell number by quantitation of DNA content per well using Hoescht, as described previously.23

Western blot analysis

For Western blotting experiments, rBM cells were cultured at 3.1 × 105 cells/cm2 on 60-mm dishes with 75 ng/mL of rhM-CSF for 7 days to form committed OC precursors. The cells were starved for 120 minutes at 37°C in serum-free medium supplemented with dasatinib or vehicle (0.05% DMSO). Where indicated, following starvation, cells were stimulated with rhM-CSF for 5 minutes at 37°C.

Similarly, rBMSCs were cultured at 8 × 103 cells/cm2 in 6-well plates and allowed to adhere overnight. The cells then were starved in serum-free medium for 22 hours prior to pretreatment with dasatinib or 0.05% DMSO vehicle for 2 hours. Where indicated, the cells were stimulated with 10 ng/mL of PDGF-BB (PeproTech, Rocky Hill, NJ, USA) for 5 minutes prior to harvest.

Cell lysates were prepared in 100 µL of nonreducing lysis buffer [1% (v/v) Nonidet P40, 20 mM tris(hydroxymethyl)aminomethane, 150 mM sodium chloride, amd 1 mM ethylenediaminetetraacetic acid (pH 8.0) supplemented with 25 mM sodium fluoride, 10 mM sodium pyrophosphate, 2 mM sodium vanadate, 0.5 mM phenylmethylsulfonyl fluoride, and complete protease inhibitors; Roche].

Equivalent amounts of protein, as determined using a RC/DC Protein Assay Kit (Pierce, Rockford, IL, USA), were used for Western blotting, as described previously.24 Membranes were probed with antibodies against phosphorylated c-fms (Tyr723), total c-fms, phosphorylated c-Src (Tyr418), phosphorylated ERK1/2 (Tyr204) (all from Cell Signaling, Danvers, MA, USA), or heat shock protein 90 (Santa Cruz Biotechnology, Santa Cruz, CA, USA) diluted in 1% bovine serum albumin/tris buffered saline with 1% tween (BSA/TBST). The immunoreactive proteins were detected subsequently with alkaline phosphatase–conjugated antibodies against rabbit immunoglobulin (1:2000 in 1% BSA/TBST; Chemicon) and developed using enhanced chemifluorescence (ECF) substrate (GE Healthcare, Little Chalfont, Buckinghamshire, UK). The membrane was imaged using a Typhoon 9410 fluorimager (Amersham) at 488-nm excitation, and quantitation was performed using ImageQuant software (Amersham). Where required, membranes were stripped using a commercial stripping buffer following the instructions of the manufacturer (Alpha Diagnostics, San Antonio, TX, USA).

Animal care

Female Sprague-Dawley rats (n = 36) were obtained from the Veterinary Services Division of the Institute of Medical and Veterinary Science (IMVS, Adelaide, Australia). Since dasatinib is used predominantly in adult CML patients with a median age of 53 years at presentation,25 9-month-old rats were chosen to examine the effects of this compound on mature, nonmodeling skeletons. Rats were provided with standard commercial rat chow and tap water ad libitum. All procedures were approved by the Animal Ethics Committees of the University of Adelaide and of the IMVS.

Experimental design

Prior to treatment, the proximal tibias of all animals were scanned using micro–computed tomography (µCT) while anesthetized with ketamine (45 mg/kg; Ilium Veterinary Products, Smithfield, NSW, Australia) and xylazine (5.6 mg/kg; Ilium), injected intraperitoneally (i.p.). Animals were randomly assigned to three groups: Group 1 received treatment by daily gavage with 1.6 mg dasatinib (equivalent to approximately 5 mg/kg; n = 18) in 10% DMSO/90% polyethylene glycol (PEG) 300 vehicle (n = 18), a dose previously shown to inhibit tumor growth in vivo1; group 2 received vehicle alone by daily gavage (n = 18); and group 3 received a subcutaneous injection with zoledronic acid (100 µg/kg) on day 0 and at 6 weeks (n = 4).26

After 4, 8, and 12 weeks of treatment, six animals each from the vehicle and dasatinib groups were culled for sera, µCT, and histologic analysis. The zoledronic acid–treated group was culled after 12 weeks of treatment. At 2 weeks and at 3 days prior to cull, the rats were injected i.p. with calcein (30 mg/kg) and alizarin (30 mg/kg), respectively, in sodium carbonate buffer (pH 7.4).

Animals were anesthetized with isoflurane, and blood was collected by cardiac puncture prior to euthanization by CO2 overdose. The left tibias were fixed in 10% neutral buffered formalin (NBF; Fronine Laboratory Supplies, Riverstone, NSW, Australia) at 4°C for 3 to 5 days prior to scanning by µCT. Following scanning, the undecalcified bones were embedded in methyl methacrylate as described below.

Bone densitometry

Animals were anesthetised with ketamine/xylazine i.p. prior to undergoing dual-energy X-ray absorptiometry (DXA). Whole-body bone mineral content, bone mineral density (BMD), adipose tissue content, and lean tissue content were determined using a GE-Lunar Prodigy DXA system (GE Healthcare). The interanimal variation for measured BMD was 2.75%; the intraanimal variation was 1.70%.

µCT analysis

3D trabecular microarchitecture at the proximal tibia was evaluated using µCT (Skyscan 1076 X-ray Microtomograph, SkyScan, Kontich, Belgium). All bone samples were scanned at 74 kV/100 mA with an isometric resolution of 8.7 µm per pixel using a 1-mm aluminium filter and two-frame averaging. Reconstruction of the original scan data was performed using NRecon (SkyScan).

Analysis of microarchitectural parameters was performed using CTAn (SkyScan). For the proximal tibia, a trabecular region of interest was defined manually to exclude the cortex. A total of 500 slices (4.351 mm) were analyzed for each tibia, commencing 250 slices (2.351 mm) distal to the growth plate. A 1.3-mm-long diaphyseal region was selected that extended 13.05 mm (1500 slices) distal to the growth plate. Digital segmentation of the bone from air/tissues was performed by adaptive (median-C) thresholding. For trabecular bone regions, bone volume fraction (BV/TV), bone surface fraction (BS/BV), trabecular thickness (Tb.Th.), trabecular number (Tb.N), structure model index (SMI), and trabecular pattern factor (Tb.Pf) were calculated using CTAn. Measurements of trabecular thickness (Tb.Th) and separation were calibrated by scanning and analyzing four aluminum foils with thicknesses of 20, 50, 125, and 250 µm (SkyScan).27 The cortical thickness (Ct.Th) and cortical bone volume fraction (Ct.BV/TV) were calculated for the diaphyseal region.

Bone mineral density (BMD) was calculated with CTAn using phantoms of known density (0.25 g/cm3 and 0.75 g/cm2) and a region corresponding to water as a reference. The volumes of interest (VOIs) were reconstructed in 3D using ANT software (SkyScan).

Methacrylate embedding

Left tibias were embedded in methyl methacrylate as described previously.28 Briefly, tissues were fixed in 20 mL of 10% NBF at 4°C for 4 days. The cortex was shaved from the medial facet of the tibia using a Buehler Isomet (Lake Bluff, IL, USA) low-speed saw. The samples were then dehydrated in sequential changes of graded acetone (70% acetone for 1 hour, 90% acetone for 1 hour, 100% acetone for 1 hour, 100% acetone for 1 hour) at 4°C and subsequently infiltrated in 91% methyl methacrylate (Merck, Kilsyth, VIC, Australia) and 9% PEG 400 under a vacuum for 1 week at room temperature. After infiltration, the samples were embedded in 10 mL of methyl methacrylate and PEG 400 (10:1) with Perkadox 16 (0.4% w/v; AkzoNobel, Tullamarine, VIC, Australia) at 37°C for 48 hours. Blocks were attached to aluminium block holders with Araldite epoxy resin (Selleys, Padstow, NSW, Australia) prior to sectioning.

Using a D-profile blade, 5-µm-thick sections were taken from each block using a Leica SM2500 motorized sledge microtome (Leica Microsystems, Wetzlar, Germany). Sections were attached to gelatine-coated slides (SuperFrost, Menzel-Gläser, Braunshweig, Germany) using spreading solution [30% 2-ethylene glycol monoethyl ether (Merck) and 49% ethanol] at 60 to 70°C, and slides were incubated at 42°C for 18 to 48 hours.

Histologic analysis

Sections were deplasticized by submersion in 100% acetone for 15 minutes prior to staining with TRACP using napthol AS-BI as a substrate and pararosaniline as a coupler.29 Briefly, slides were incubated at 37°C for 30 minutes in acetate-tartate buffer (200 mM sodium acetate, 100 mM potassium sodium tartrate, pH 5.2) containing 0.4 mg/mL of napthol AS-BI phosphate. The slides then were transferred to 1 mg/mL of hexaazotized pararosaniline solution in acetate-tartrate buffer and were incubated at 37°C for a further 30 minutes. Methyl green (0.5 mg/mL) was used as a counterstain.

Unstained, deplasticized sections were cover-slipped and used for dynamic measurements of bone formation. Owing to the slow rate of bone formation in these aged rats, the alizarin label, administered 3 days prior to cull, was not visible. Mineral apposition rate (MAR*) therefore was evaluated as the mean distance between the calcein-labeled mineralized surface (MS) and the edge of the bone (BS) divided by 14 days (the interval between labeling and death of the animals). Bone-formation rate (BFR*) was derived using the formula: BFR* = MAR* × MS/BS* × 365/100.

Histomorphometric analyses were performed on duplicate slides using OsteoMeasure XP (OsteoMetrics, Decatur, GA, USA) in a 2.85-mm2 region of secondary spongiosa distanced 0.78 mm from the growth plate.

Serum biochemistry

Cardiac blood samples were collected and spun at 800 × g for 15 minutes, and the collected sera were stored at –20°C until analysis. Serum levels of phosphate and total calcium were measured using an Olympus AU5400 Chemistry Analyzer. Levels of C-terminal collagen cross-links (CTX-1) were measured in cleared sera by ELISA, as described by the manufacturer (Nordic Biosciences, Herlev, Denmark). The inter- and intraassay variations of this ELISA were both 6%. The serum levels of osteocalcin were measured by ELISA in accordance with the manufacturer's instructions (Biomedical Technologies, Stoughton, MA, USA). The inter- and intraassay variations were 7% and 4%, respectively. Serum levels of N-terminal propeptide of type I procollagen (P1NP) were measured by ELISA, as described by the manufacturer (Immuno Diagnostic Systems, Boldon, Tyne & Wear, UK). The intraassay variation was 6.4%; the interassay variation was 9.2%.

Statistical analysis

Analysis of data was performed using GraphPad PRISM (GraphPad Software, San Diego, CA, USA). For in vitro dose-response experiments, treatments were compared using one-way analysis of variance (ANOVA) with Dunnett's posttests. For in vivo experiments, treated groups were compared with vehicle controls using unpaired Student's t tests. For serum CTX–1 analysis, since some values were below the detection limit of the ELISA used, the nonparametric Mann-Whitney test was used. Differences were considered to be statistically significant when the p value was less than 0.05.

The calculation of 50% inhibitory concentration (IC50) values was performed using the Hill equation: y = 100/(1 + 10math image), where y is the level of inhibition and x is the logarithmic drug concentration.

Results

In vitro OC formation and activity

Previous studies from our laboratory and those of others have shown that dasatinib inhibits human and murine OC formation and activity in vitro.7, 8 To confirm that rat OCs were sensitive to dasatinib, we examined the effect of dasatinib on OC formation and activity in rat bone marrow cultures in vitro. A significant decrease in the number of TRACP+ multinucleated cells was observed at 1.25 nM dasatinib (p < .05; Fig. 1A, C), with complete abrogation of OC formation at 10 nM (IC50 = 1.59 nM). The effect of dasatinib on OC activity also was assessed in rBM cell cultures using calcium phosphate–coated slides. A significant inhibition of resorption was observed at 1.25 nM dasatinib or greater (p < .05; Fig. 1B).

Figure 1.

Dasatinib treatment abolishes OC formation and activity in rBM cell cultures in vitro. (A) rBM cells were cultured on 96-well plates for 6 days in medium containing rhM-CSF and rhRANKL with dasatinib (containing 0.05% DMSO vehicle) or vehicle alone. Cultures then were stained for TRACP, and TRACP+ cells with three or more nuclei were counted. (B) rBM cells were cultured on calcium phosphate–coated slides in rhM-CSF and rhRANKL and the indicated concentrations of dasatinib or 0.05% DMSO vehicle. Following 9 days of treatment, calcium phosphate layer was silver stained, and the area of substrate resorbed (lacunae area) was calculated. Graphs depict mean ± SEM of three independent experiments (A) and mean ± range of teo independent experiments (B). ap < .05 relative to vehicle control. (C) Representative photos of TRACP+ OC cultures (scale bars = 200 µm). rBM cells were starved in serum-free α-MEM containing dasatinib or vehicle for 120 minutes. Where indicated, cells were stimulated with M-CSF (75 ng/mL) for 5 minutes prior to harvest of cell lysates. Cell lysates (75 µg/lane, D; 30 µg/lane, F) were resolved on SDS-PAGE gels (8%, D; 10%, F), transferred to PVDF and were probed with a specific antibody to phosphorylated c-fms, total c-fms, and HSP 90 (D) or phosphorylated c-Src and HSP 90 (F). I = ubiquitinated c-fms; II = mature c-fms. A representative of two independent experiments is shown. A sigmoidal curve fit of the normalized band intensity is shown for phosphorylated c-fms (E) and for phosphorylated c-Src (G).

Since dasatinib may affect OC formation and activity through inhibition of c-fms and c-Src, we next examined the effects of dasatinib on c-fms and c-Src phosphorylation in rBM cell cultures. Pretreatment of cultures with dasatinib inhibited M-CSF-induced phosphorylation of c-fms at concentrations of 1.25 nM and greater (IC50 = 4.1 nM; Fig. 1D, E). Basal levels of c-Src phosphorylation were inhibited at concentrations in excess of 10 nM dasatinib (IC50 = 35.6 nM; Fig. 1F, G).

In vitro OB activity

It has been demonstrated previously that the tyrosine kinase inhibitor imatinib increases OB activity while inhibiting OB proliferation in vitro, at least in part owing to the inhibition of PDGFR.17–20 To determine whether OBs are similarly sensitive to dasatinib, we examined the effect of dasatinib on cell proliferation and mineral formation in rodent stromal cell cultures in vitro.

Treatment of rBMSC cultures with dasatinib significantly decreased cell proliferation (Fig. 2A). At all time points examined, rBMSC numbers were significantly lower than controls at concentrations in excess of 10 nM dasatinib (p < .05; IC50 = 10.7 nM at day 6; Fig. 2A).

Figure 2.

Dasatinib treatment increases OB activity while inhibiting OB proliferation in vitro. (A) rBMSCs were cultured on 96-well plates for up to 14 days in the indicated concentrations of dasatinib or 0.05% DMSO vehicle. Relative cell numbers per well were quantitated using the WST-1 assay. Graphs depict mean ± range of two independent experiments. At each time point examined, 5, 10, 20, and 40 nM dasatinib were significantly different from vehicle controls (ap < .05). (B) rBMSCs were cultured on 96-well plates for 21 days in mineralization medium containing dasatinib (containing 0.05% DMSO vehicle) or vehicle alone. Deposited mineralized matrix was quantitated using the cresolpthalein method and was normalized to DNA content per well. Graphs depict mean ± SEM of three independent experiments. ap < .05 relative to vehicle control. (C) rBMSCs were starved in serum-free α-MEM containing dasatinib or vehicle overnight. Where indicated, cells were stimulated with PDGF-BB (10 ng/mL) for 10 minutes prior to harvest of cell lysates. Cell lysates (30 µg/lane) were resolved on 10% SDS-PAGE gels, transferred to PVDF, and probed with a specific antibody to phosphorylated ERK1/2, phosphorylated Akt, or HSP 90. A representative of two independent experiments is shown. (D) A sigmoidal curve fit of the normalized band intensity is shown for Akt (circles) and ERK1/2 (squares).

Culture of rBMSCs with β-glycerophosphate induced the formation of mineralized extracellular matrix. When cultured in the presence of 40 nM dasatinib, there was a significant increase in mineral formation when corrected for cell number per well (p < .05; Fig. 2B).

We next determined the effects of dasatinib on PDGFR signaling pathways. Treatment of rBMSCs with PDGF-BB induced the phosphorylation of Akt and ERK1/2 (Fig. 2C). The induction of Akt and ERK1/2 phosphorylation was inhibited by treatment with 10 nM dasatinib and higher (IC50 = 14.8 and 12.2 nM, respectively). Activation of ERK1/2 and Akt was reduced to unstimulated levels at concentrations in excess of 20 nM dasatinib.

DXA measurements

Nine-month-old Sprague-Dawley rats were treated with dasatinib (5 mg/kg) or vehicle (10% DMSO/90% PEG 300) by daily oral gavage for up to 12 weeks. Four additional animals received zoledronic acid (100 µg/kg for 6 weeks) subcutaneously as a positive control for inhibition of OCs. Four animals per group were monitored throughout the experiment for whole-body tissue composition by DXA.

There were no statistically significant changes in body mass, lean tissue mass, or adipose tissue mass in dasatinib-, zoledronic acid–, or vehicle-treated animals at any time point (data not shown). Whole-body BMD was not significantly different in the dasatinib-treated groups compared with the vehicle-treated groups at any time point (Fig. 3). Similarly, zoledronic acid treatment did not result in a detectable difference in whole-body BMD relative to vehicle controls (Fig. 3).

Figure 3.

Dasatinib treatment does not result in detectable changes in whole-body BMD in normal mature rats. Nine-month-old female Sprague-Dawley rats were treated for 12 weeks with dasatinib (5 mg/kg/day), vehicle (10% DMSO/90% PEG 300), or zoledronic acid (100 µg/kg for 6 weeks). Whole-body BMD was monitored by DXA at the indicated time points (n = 4 per group). Graphs depict mean ± SEM.

Bone morphometry

The effect of dasatinib on trabecular bone morphometry was assessed using µCT (Fig. 4A). There were no significant differences in trabecular bone parameters at baseline in the dasatinib group relative to the control group (data not shown). In contrast, after 8 and 12 weeks of treatment, bone volume (BV/TV) was increased 21% and 62%, respectively, in dasatinib-treated animals relative to vehicle controls (p < .05; Fig. 4B). At 12 weeks of treatment, dasatinib induced a similar increase in BV/TV as that induced by zoledronic acid (Fig. 4B). No significant change in Tb.BMD was detected at 4, 8, or 12 weeks (p = .465, .107, and .083, respectively; data not shown).

Figure 4.

Trabecular morphology at the proximal tibia is altered in dasatinib-treated animals. At 4, 8, and 12 weeks, tibias from dasatinib-, vehicle-, and zoledronic acid–treated animals were collected for µCT analysis of a 4.351-mm-long region of trabecular bone 2.351 mm distal to the growth plate. Cross-sectional 3D images of representative tibias from animals treated for 12 weeks are shown (A): BV/TV (B), Tb.Th (C), BS/BV (D), BS/TV (E), Tb.N (F), SMI (G), and Tb.Pf (H). Graphs depict mean ± SEM (n = 4 to 6). ap < .05 relative to vehicle control at each time point. bp < .05 relative to week 4 and week 8 vehicle control.

In the vehicle-treated group, there was a significant decrease in trabecular thickness (Tb.Th) after 12 weeks of treatment relative to the 4- and 8-week cohorts, suggesting a loss of bone over time with aging (p < .05; Fig. 4C). In contrast, in the dasatinib-treated animals, there was no change in Tb.Th over time, resulting in a significantly higher Tb.Th in the dasatinib and zoledronic acid groups than in the vehicle-treated groups at 12 weeks (p < .05; Fig. 4C). This translated into a decrease in bone-surface-to-bone-volume ratio (BS/BV) in dasatinib- and zoledronic acid–treated animals at 12 weeks compared with controls that did not reach significance (dasatinib, p = .059; zoledronic acid, p = .077; Fig. 4D). Similarly, there were no significant changes in bone-surface-to-total-volume ratio (BS/TV) in dasatinib- and zoledronic acid–treated animals (Fig. 4E). Dasatinib had no significant effect on trabecular number (Tb.N) at 4, 8, or 12 weeks of treatment (p = .092, .117, and .161, respectively; Fig. 4F). Histomorphometric analysis of trabecular architecture revealed similar changes to those detected by µCT (data not shown).

Structure model index (SMI) was decreased significantly in dasatinib-treated animals at 8 and 12 weeks relative to vehicle-treated controls, indicating a shift to a more platelike trabecular architecture (p < .05; Fig. 4G). Similarly, trabecular pattern factor (Tb.Pf) was decreased significantly by 12 weeks of dasatinib treatment relative to the control group, indicating an increase in trabecular connectivity (p < .05; Fig. 4H). There were no statistically significant changes in SMI and Tb.Pf in the zoledronic acid–treated group relative to controls (p = .054 and .114, respectively; Fig. 4G, H).

Cortical bone volume (Ct.BV/TV) and cortical thickness (Ct.Th) were not altered in the dasatinib- or the zoledronic acid–treated animals relative to controls at any time point examined (data not shown).

OC activity

We next examined whether dasatinib treatment affected bone resorption in vivo. OC number (N.Oc/B.Pm) and OC surface (Oc.S/BS) at the proximal tibia were decreased in the dasatinib-treated animals compared with controls (p < .05; Fig. 5AC). Similar decreases in N.Oc/B.Pm and Oc.S/BS were observed in the zoledronic acid–treated animals (p < .05; Fig. 5AC). Serum levels of the bone resorption marker CTX-1 were decreased by 45% in the dasatinib-treated group relative to controls after 12 weeks of treatment (p < .05; Fig. 5D). A significant decrease in serum phosphate levels was observed in the dasatinib-treated animals at 12 weeks (vehicle: 1.56 ± 0.05 mM; dasatinib: 1.43 ± 0.02 mM; p < .05), whereas total calcium levels were unchanged (p = .976). In zoledronic acid–treated animals, CTX-1 levels were undetectable.

Figure 5.

Dasatinib treatment decreases OC activity in vivo. Tibias were collected and embedded in methyl methacrylate, and 5-µm sections were stained for TRACP and counterstained with methyl green. TRACP+ multinucleated OC numbers (N.Oc/B.Pm) (A) and OC surface (Oc.S/BS) (B) were quantitated in a defined region of secondary spongiosa at the proximal tibia on duplicate slides. Graphs depicts mean ± SEM (n = 3 to 6). ap < .05. Representative images are shown (C). Scale bar = 50 µm. Arrows indicate OC. Blood was collected by cardiac puncture after treatment with dasatinib or vehicle (4, 8, or 12 weeks) or zoledronic acid (12 weeks only), and serum levels of C-terminal collagen cross-links (CTX-1) (D) were quantitated by commercial ELISA. Graphs depicts mean ± SEM (n = 3 to 6). ap < .05 relative to vehicle control.

OB activity

The dynamic measurements of OB activity MS/BS*, MAR*, and BFR* were not significantly different in the dasatinib-treated group compared with controls (Fig. 6AC). In the zoledronic acid–treated group, there was a significant decrease in MS/BS* and MAR* relative to controls (p < .05). Serum levels of the bone-formation markers osteocalcin and P1NP were not significantly altered by treatment with dasatinib or zoledronic acid compared with vehicle-treated controls (Fig. 6D, E).

Figure 6.

Dasatinib treatment does not affect OB activity in vivo. Tibias were collected and embedded in methyl methacrylate, and dynamic histomorphometric measurements were conducted in duplicate on unstained 5-µm sections. Mineralized bone surface (MS/BS*) (A), bone-formation rate (BFR*) (B), and mineral apposition rate (MAR*) (C) were quantitated in a defined region of secondary spongiosa at the proximal tibia. Graphs depicts mean ± SEM (n = 3 to 6). Blood was collected by cardiac puncture after 4, 8, or 12 weeks of treatment with dasatinib, vehicle, or zoledronic acid, and serum levels of osteocalcin (D) and P1NP (E) were quantitated by commercial ELISA. Graphs depicts mean ± SEM (n = 3 to 6).

Discussion

Therapeutic concentrations of the tyrosine kinase inhibitor dasatinib inhibit the M-CSF receptor c-fms.2, 7, 8 Since c-fms plays a crucial role in regulating bone remodeling by OCs, we examined whether dasatinib could modulate bone remodeling in normal mature Sprague-Dawley rats in vivo. While changes in whole-body BMD were not observed using DXA, treatment of rats with dasatinib substantially increased trabecular bone volume. This is attributable, at least in part, to an increase in trabecular thickness (Fig. 4). The change in bone morphology in the dasatinib-treated animals was strikingly similar to that induced by the potent OC inhibitor zoledronic acid, suggesting that dasatinib and zoledronic acid may have similar mechanisms of action. In support of this hypothesis, data presented here suggest that dasatinib increased bone volume by inhibiting OC activity, as evidenced by a decrease in OC numbers and OC-occupied bone surface, as well as decreased serum CTX-1 levels (Fig. 5). This is further supported by in vitro studies that demonstrate that dasatinib abrogates OC formation (Fig. 1).7, 8

c-fms and c-Src are two known dasatinib targets that play a crucial role in bone remodeling by OCs. Binding of M-CSF to its receptor, c-fms, regulates the proliferation and survival of OC precursors and mature OCs30 and induces cell migration and cytoplasmic spreading.31–33 In in vitro OC cultures and animal models, such as tl/tl rats and op/op mice, an absence of signaling through c-fms results in a severe depletion of OCs and OC precursors.10–12, 30 c-fms signaling also functions by inducing the expression of a number of OC-associated genes, including RANK.34 c-Src is also known to play an important role in OC activity in vivo, as demonstrated by c-Src–/– mice, which are severely osteopetrotic owing to an absence of OC activity.35–38 While the precise mechanism(s) remain(s) to be elucidated, data shown here (Fig. 1) and previously7, 8 suggest that dasatinib inhibits in vitro osteoclastogenesis at concentrations at which c-fms but not c-Src is affected. This suggests that inhibition of c-fms, but not c-Src, contributes to the antiosteoclastic effects of dasatinib treatment in vitro. However, it is not clear to what extent inhibition of c-Src and c-fms contributes to the inhibitory effects of dasatinib on OCs in vivo.

The balance between bone resorption and bone formation is essential for maintaining normal bone homeostasis. Under normal conditions, the activity of bone-resorptive OCs is closely coupled with bone formation by OBs. In this study we have shown for the first time that dasatinib inhibits rodent stromal cell proliferation and promotes the activity of mature OBs (Fig. 2). Previous studies have shown that treatment of human and rodent cells with the tyrosine kinase inhibitor imatinib inhibited cell proliferation and increased mineralized matrix formation via a mechanism that involved inhibition of the PDGFR.39 In cultures of human stromal cells isolated from bone biopsies, treatment with imatinib significantly increased mineral deposition17, 21 and decreased proliferation of stromal cells.17, 18 Similarly, imatinib increased mineralization19 and partially inhibited proliferation19, 20 in primary rat stromal cells and murine cell lines. The proliferation, differentiation, and maturation of OBs are influenced by signaling pathways, including PDGFR, that are inhibited by dasatinib. PDGF is an important regulator of OB formation because it enhances OB proliferation while inhibiting OB differentiation in vitro.13–16 Treatment with imatinib relieved the inhibitory effects of PDGF on mineralization in human bone explant cultures and MC3T3-E1 cells,17, 19 suggesting that inhibition of PDGFR signaling may be one mechanism whereby imatinib activates differentiation in vitro. It is likely that dasatinib also inhibits OB proliferation and stimulates differentiation owing to its specificity for the PDGFR, although this remains to be demonstrated. In this study, serum osteocalcin levels, P1NP levels, BFR*, and MAR* were not altered in dasatinib-treated animals, suggesting that dasatinib did not affect OBs in vivo at the dose administered. The concentration at which dasatinib activates OB activity in vitro is more than eightfold higher than that required to inhibit OC activity, which may suggest that the serum concentrations achieved in this study were not sufficient to modulate OB function.

Compelling evidence has been published recently indicating that dramatic changes in bone parameters may occur in patients treated with the 2-phenylpyrimidine-derived tyrosine kinase inhibitor imatinib mesylate. We have shown previously that CML patients on imatinib therapy for 17 to 62 months exhibited a significant increase in mean trabecular bone volume in iliac crest trephines, with 8 of 17 patients exhibiting a more than twofold increase in trabecular bone volume relative to baseline.17 Additionally, two subsequent studies showed that CML patients undergoing imatinib therapy exhibited small increases in regional BMD.40, 41 In support of these findings, several groups have shown that CML and gastrointestinal stromal tumor (GIST) patients undergoing imatinib treatment display altered serum biochemistry that is suggestive of changes in bone remodeling. In particular, a number of studies show that decreased levels of serum phosphate occur as early as 3 months following initiation of imatinib treatment in CML and GIST patients.17, 42–44 This decrease in serum phosphate is maintained for at least 12 months,44 and in at least 50% of patients, the phosphate levels decrease to below the normal range (hypophosphataemia) at least once during therapy.42, 45, 46 Decreased serum phosphate levels in imatinib-treated patients are associated with increased serum parathyroid hormone levels, decreased levels of the bone resorption marker CTX-1, and in at least some cases, a decrease in serum calcium levels.17, 40, 41, 43–45 To our knowledge, there have been no published reports suggesting that changes in bone remodeling occur in patients receiving dasatinib. However, grade 3 or 4 hypophosphatemia has been reported in 11% of dasatinib-treated patients.47 In keeping with this observation, this study found that serum phosphate levels were decreased significantly in dasatinib-treated rats. This may be indicative of decreased release of bone phosphate stores through inhibition of osteoclastogenesis, as has been suggested for patients undergoing imatinib therapy.17, 45

There is some concern that an inhibition of bone turnover may result in an increase in bone fragility through the accumulation of microfracture damage.48–50 In healthy bone, the accumulation of microfractures over time is controlled by targeted bone remodeling.48, 49 Abrogation of bone remodeling therefore may increase the level of skeletal microdamage and lead to bone with decreased mechanical strength.50 In support of this, several studies have suggested that long-term inhibition of bone turnover by bisphosphonates may result in increased microfracture accumulation and, in some cases, decreased bone strength.51–53 At the dosing schedule used here, dasatinib treatment of normal rats decreased OC numbers and serum CTX-1 levels by 50%. In contrast, there were no changes in OB activity in dasatinib-treated animals, as determined by the serum markers osteocalcin and P1NP and by the dynamic measurements MS/BS*, BFR*, and MAR*, suggesting that bone turnover still may have been occurring.

In this study, dasatinib increased trabecular bone volume and trabecular thickness, associated with an increase in trabecular connectivity (decreased Tb.Pf) and a change in trabecular architecture from a more rodlike to a more platelike form (decreased SMI; Fig. 4). Previous studies have demonstrated that cancellous bone strength is determined primarily by trabecular geometry; increased cancellous bone strength is strongly associated with increased bone volume and trabecular thickness and by decreased Tb.Pf and SMI.54–56 Therefore, our structural results indicate that dasatinib is able to positively affect trabecular bone architecture in normal rats, which is suggestive of an increase in bone strength. The potential effects of dasatinib on bone quality in patients undertaking long-term treatment need further elucidation.

The improvements in trabecular architecture induced by dasatinib mirrored those induced by zoledronic acid. Bisphosphonates, such as zoledronic acid, are in use as effective therapeutic agents for diseases characterized by aberrantly increased bone resorption, including osteoporosis, Paget disease, osteogenesis imperfecta, and bone metastatic tumors in breast cancer, prostate cancer, and multiple myeloma.57, 58 The similarities in the effects of dasatinib and zoledronic acid on trabecular bone architecture observed in this study suggest that dasatinib may be an effective treatment for diseases characterized by increased OC activity.

Importantly, this study highlights a need for further elucidation of the long-term effects of dasatinib on bone quality in patients undertaking long-term dasatinib therapy. Further studies to determine whether dasatinib can be used effectively for the treatment of diseases with increased bone resorption are also warranted.

Disclosures

All the authors state that they have no conflicts of interest.

Acknowledgements

We are grateful to Behzad Baradaran and the staff at Veterinary Services, IMVS, for their assistance with the animal experiments. We wish to thank Lee Anne Griffiths, Francis Lee, Richard Smykla, and Kate Church from Bristol-Myers Squibb for providing the dasatinib and helpful discussions. This work is supported by a grant from the Leukemia and Lymphoma Society Translational Research Program (awarded to ACWZ).

Ancillary