SEARCH

SEARCH BY CITATION

Keywords:

  • BONE LOSS;
  • OVARIECTOMIZED RAT;
  • ROSIGLITAZONE;
  • THIAZOLIDINEDIONE;
  • BONE DENSITY;
  • BONE STRENGTH;
  • BONE HISTOMORPHOMETRY

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

Rosiglitazone (RSG) is an antidiabetic drug that has been associated with increased peripheral fractures, primarily in postmenopausal women. In this report, we investigated the underlying mechanisms of RSG-associated bone loss in ovariectomized (OVX) rats and determined whether changes in bone parameters associated with RSG administration are reversible on treatment cessation or preventable by coadministration with an antiresorptive agent. Nine-month-old Sprague-Dawley rats underwent OVX or sham operation. Sham-operated rats received oral vehicle only; OVX animals were randomized to receive vehicle, RSG, alendronate (ALN), or RSG plus ALN for 12 weeks. All treatment started the day after ovariectomy. After the 12-week treatment period, the OVX and RSG groups also underwent an 8-week treatment-free recovery period. Bone densitometry measurements, bone turnover markers, biomechanical testing, and histomorphometric analysis were conducted. Microcomputed tomography was also used to investigate changes in microarchitecture. RSG significantly increased deoxypyridinoline levels compared with OVX. Significant exacerbation of OVX-induced loss of bone mass, strength, and microarchitectural deterioration was observed in RSG-treated OVX animals compared with OVX controls. These effects were observed predominantly at sites rich in trabecular bone, with less pronounced effects in cortical bone. Coadministration of RSG and ALN prevented the bone loss associated with RSG treatment. Following cessation of RSG treatment, effects on bone mass and strength showed evidence of reversal. Thus, treatment of OVX rats with RSG results in loss of bone mass and strength, primarily at sites rich in trabecular bone, mainly due to increased bone resorption. These effects can be prevented by concomitant treatment with ALN and may be reversed following discontinuation of RSG.


Introduction

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

The thiazolidinediones (TZDs), rosiglitazone (RSG) and pioglitazone (PIO), are used as glucose-lowering agents for the treatment of type 2 diabetes mellitus (T2DM). Evidence from clinical trials and case control studies indicates that RSG[1] and PIO[2, 3] increase fracture risk in patients with T2DM.[4] The fractures were noted in both upper and lower limbs as opposed to classical osteoporotic fractures that typically occur in the spine or hip, pointing to possible mechanistic differences. In addition, the gender-specific effects of RSG and PIO on fracture rates in women[1-3, 5] suggest that hormonal status may be an important factor in susceptibility to TZD-induced bone loss. However, the cellular and molecular mechanisms underpinning TZD-induced effects on bone, or its hormone dependency, remain poorly understood. A number of mechanisms have been proposed and are currently under investigation.

TZDs are primarily peroxisome proliferator-activated receptor (PPAR)-γ agonists that play a significant role in mesenchymal stem cell differentiation into osteoblasts, adipocytes, and chondrocytes.[6-9] Studies in rodent models have suggested effects on both bone resorption and formation via an impact on osteoclasts and osteoblasts. Treatment of 6-month-old mice with RSG decreased bone mineral density (BMD) at 7 weeks but not at 4 weeks,[10] with similar observations seen with pioglitazone in the rat.[11] In these and other studies, loss of BMD appeared to result from a reduction in osteoblast number and/or activity, leading to changes in bone marrow composition and bone microarchitecture.[10, 12, 13] These effects on bone formation are supported by in vitro studies in which PPAR-γ agonists affected mesenchymal stem cell differentiation into adipocytes, often at the expense of osteoblasts.[6-9]

Other in vitro studies demonstrate that treatment with TZDs also inhibits osteoclast formation,[14, 15] and increased resorptive activity has been reported in bone organ culture[16] and in vivo.[17, 18] TZDs increased osteoclast number without impacting bone formation,[17, 18] and genetic studies with deletion of the PPAR-γ gene have strongly suggested its role in promoting osteoclast differentiation and bone resorption.[19]

Yet some studies with TZDs in rodents report no effects on bone,[20, 21] whereas others clearly show a negative effect on bone mass and, in some studies, strength.[10, 12, 22] In an isolated study, a PPARγ agonist (troglitazone) and a PPARα and γ agonist (JTT501) were even shown to increase bone density in Zucker diabetic fatty rats.[23] The gender, age, and hormonal status also appears to be important, because negative effects of TZDs in bone have been reported in ovariectomized (estrogen-deficient; OVX) rats but not in ovary-intact female rats.[18] Similarly, RSG-induced changes in bone microarchitecture, such as decreased bone volume fraction, trabecular number, and increased trabecular spacing, has been observed in adult and elderly but not in young male mice.[24]

Most in vivo rodent studies do not report actual bone strength data, but instead show the effects of TZDs on bone mass by dual-energy X-ray absorptiometry (DXA) or peripheral quantitative computed tomography (pQCT), and on histomorphometric parameters of bone formation and resorption. Almost all rodent studies are also devoid of drug exposure data, which makes extrapolation of TZD dose to the clinical situation challenging. Thus, preclinical in vitro and in vivo studies have previously reported varying and contrasting effects of TZDs on stimulation of osteoclast formation and activity, as well as the impact on osteoblast–adipocyte balance. What is clear, however, is that TZDs exhibit a net negative effect on bone, although the exact mechanisms behind these effects remain to be determined.

Given the range of varying evidence reported in the literature to date, in this study we investigated the effect of RSG on bone loss in aged OVX rats (a model with similar hormonal status to the osteoporotic patient population) in a comprehensive manner (including pharmacokinetic, biochemical biomarker, bone mass, bone dynamics, bone microarchitecture, and bone strength parameters) to provide an insight into the possible mechanisms. In addition to the typical bone sites such as lumbar spine, tibia, and femur, data from the humerus, a peripheral site with greater fracture incidence in this RSG-treated patient population, was also included. Furthermore, RSG administered concurrently with ALN was assessed to determine whether RSG-induced bone loss can be prevented by a bisphosphonate because this class of agents is commonly used as antiresorptive therapy.[25] We also evaluated whether the effect of RSG on bone was reversible after cessation of treatment.

In this work, we present the in vivo pharmacokinetics, assessment of biochemical markers of bone turnover, bone densitometry, biomechanical, micro–computed tomography (μCT), and histomorphometric parameters of both trabecular and cortical compartments, to elucidate the effects of RSG and antiresorptive treatments on bone strength parameters and bone structure, and discuss the possible mechanisms involved.

Subjects and Methods

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

Animals and experimental design

Nine-month-old female Sprague-Dawley rats (254–353 g) were obtained from Charles River Laboratories (Kingston, NY, USA). They were caged individually with the temperature maintained within a range of 19°C to 25°C, with an approximate 12-hour light/12-hour dark cycle. All cages had automatic watering valves, and the animals were offered at least 16 g daily of Certified Rodent Diet 4.2 EXT 5L35 (PMI Nutrition International, St. Louis, MO, USA), increased to 21 g in week 2. The animals were treated in accordance with the Standard Operating Procedures of Charles River Laboratories Preclinical Services, Montreal, Canada, and by the Institutional Animal Care and Use Committee. Animals were either sham-operated or ovariectomized and divided into various groups as detailed in Table 1. Sham-operated rats received oral vehicle only; OVX animals were randomized to receive vehicle, RSG, ALN, or RSG plus ALN for 12 weeks. All treatment started the day after ovariectomy. After the 12 week treatment period, the OVX and RSG groups underwent an 8-week treatment-free recovery period.

Table 1. Study Design
GroupTreatmentaMain study, nRecovery study, n
  1. Sham = sham operated; PO = per os, orally; OVX = ovariectomized; RSG = rosiglitazone; ALN = alendronate, SC = subcutaneously.

  2. a

    All treatment started the day after ovariectomy.

1Sham + vehicle PO12 
2OVX + vehicle PO1212
3OVX + RSG 10 mg/kg/d PO1212
4OVX + vehicle PO + ALN 0.03 mg/kg SC twice/week12 
5OVX + RSG 10 mg/kg/d PO + ALN 0.03 mg/kg SC twice/week12 

Pharmacokinetic analysis

Blood samples were collected at weeks 1 and 12 and plasma concentrations of RSG were determined by the GlaxoSmithKline bioanalysis group. A noncompartmental method was used to obtain estimates of pharmacokinetic parameters, including maximum observed plasma concentration (Cmax), time to Cmax (Tmax), and the area under the plasma concentration–time curve (AUC).

Bone mass measurements by DXA and pQCT

The in vivo DXA scans (performed using a DXA Hologic QDR Discovery A, Denver, CO, USA) were used to measure BMD, bone mineral content (BMC), and area of the lumbar spine (L1–L4, dorso-ventral view), right humerus, and whole body. pQCT scans were performed using an XCT Research SA, SA + bone scanner (Stratec Medizintechnik GmbH, Pforzheim, Germany) with software version 5.50D. A scan slice was obtained in the proximal tibia metaphysis and an additional scan at the diaphysis. The scan at the metaphysis (a site rich in trabecular bone) was assessed for total slice and trabecular and cortical/subcortical bone area, BMC, and BMD (ContMode 2, PeelMode 20, 40% trabecular area). The scan at the diaphysis site (cortical bone) was assessed for total slice and cortical bone area, cortical BMC and BMD, periosteal circumference, endosteal circumference, and cortical thickness (CortMode 2, threshold 0.930 1/cm). DXA and pQCT analyses were performed pretreatment and at weeks 4, 8, 12, and 20.

Ex vivo bone mineral density and geometry

The bone densitometry and bone geometry parameters of excised bones were evaluated prior to strength testing using pQCT. Excised bone specimens, L3 vertebrae and the whole right femur, were wrapped in saline-soaked gauze and plastic wrap, and placed in a plastic bag before being frozen at approximately −20°C. Specimens were thawed overnight in the refrigerator at approximately −4°C before preparation for scanning. The parameters obtained by pQCT were used to normalize bone strength data to bone size.

For L3 compression, pQCT scans were performed midsection of the lumbar spine vertebral body and were analyzed using Contmode 2 and PeelMode 20 (trabecular area 50%). pQCT scans were obtained for the femur at the expected breaking point in three-point bending and were analyzed using Cortmode 2 (threshold 0.930 1/cm).

Biomechanical testing

The L3 vertebral body was tested in compression to failure. The loading rate was set at 20 mm/min. Load and displacement data were collected and apparent strength and modulus were calculated using the peak load, stiffness, individual vertebral body height, and cross-sectional area from the pQCT scans. Work to failure was determined from AUC and toughness was calculated using cross-sectional area from the pQCT scans and vertebral body height. The whole right femur was tested to failure in three-point bending to determine bone strength at a site rich in cortical bone. The actuator was set at a rate of 1 mm/s until failure occurred. Load and displacement data were collected. Peak load was determined from the resulting load versus displacement curve and converted to ultimate stress using the radius, cross-sectional moment of inertia, and the span. The work to failure was determined from AUC and toughness was calculated. Stiffness was recorded (defined as the slope of the linear/elastic region of the load versus displacement curve) and modulus was calculated using the moment of inertia obtained from the pQCT data and span. Biomechanical tests were performed using an 858 Mini Bionix Servohydraulic Test System, Model 242.03 (MTS System Corporation, Eden Prairie, MN, USA) and data were evaluated using TestWorks V 3.8A for TestStar V4.0C (MTS System Corporation).

µCT analysis

Three-dimensional µCT analysis was carried out using a Micro-CT 40 scanner (software version 6.0; Scanco Medical AG, Brüttisellen, Switzerland) to determine trabecular (left proximal tibia) and cortical bone (right mid-femur) parameters. Although the nomenclature and analyses were performed in accordance with the guidelines put forward by Bouxsein and colleagues,[26] that article was not published at the time of these studies.

Histomorphometry

At 7 days and again at 2 days before necropsy, rats were subcutaneously injected with 8 mg/kg of calcein green. Bones were preserved in 10% neutral buffered formalin followed by subsequent storage in 70% alcohol and processed undecalcified through steps of dehydration using graded alcohol and xylene, and embedding in methylmethacrylate. For trabecular bone evaluation, sections from two levels were cut from the right proximal tibia (frontal plane). These sections were stained with Goldner's trichrome or left unstained. For cortical bone assessment, two levels of transversally cut ground and unstained sections of diaphysis were prepared from the tibiofibular junction of the right tibia.

Histomorphometric data were gathered using a Bioquant image analyzer (Bioquant Nova Prime and/or Bioquant Osteo II; Bioquant Image Analysis Corporation, Nashville, TN, USA). Trabecular bone measurements excluded the primary spongiosa and were gathered from the secondary spongiosa of the tibial metaphysis.

The nomenclature recommended by the Histomorphometry Nomenclature Committee of the American Society of Bone and Mineral Research was used,[27] but no correction factor was applied for section obliquity during calculation of trabecular thickness and mineral apposition rate (MAR). Variables related to trabecular bone microarchitecture were calculated according to the parallel plate model. Mineralizing surface (MS/BS), endocortical labeled surface (Ec.L.Pm/Ec.Pm; Pm = perimeter) and periosteal labeled surface (Ps.L.Pm/Ps.Pm) were calculated as half-single plus double-label surfaces. When it was not possible to calculate the MAR, the variable was reported as 0.6 µm/d; this value represented the smallest measurable interlabel width based on our exclusion criteria and reflected the assumption that a minimal bone turnover was continually ongoing.

Statistics

Statistical analysis was performed on the least squares means. Group differences were examined by one-way analysis of variance (ANOVA) during the treatment period using the MIXED procedure of the SAS/STAT module (SAS Institute, Inc., Cary, NC, USA). If the overall ANOVA F-test was significant (p ≤ 0.05), then the Dunnett's t test was used to perform the comparisons between the OVX controls and each of the other groups. For the recovery phase, a two-sample t test or a Satterthwaite approximation was used to compare the OVX control to the RSG group whenever group variances were homogeneous or heterogeneous, respectively. To compare the end of the recovery to end of the treatment, a paired t test with a two-tailed distribution was used. For in vivo densitometry data (DXA and pQCT), each treatment result was calculated as percentage change from baseline. These variables were submitted to the analysis above.

Results

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

RSG pharmacokinetics

Nine-month-old Sprague-Dawley rats underwent OVX or sham operation and randomized to receive vehicle, RSG, ALN, or RSG plus ALN for 12 weeks as detailed in Table 1. All treatment started the day after ovariectomy. The OVX and RSG groups also underwent an 8-week treatment-free recovery period. Oral administration of RSG at 10 mg/kg/d to OVX rats for 12 weeks resulted in a Cmax and AUC of 21.8 µg/mL and 156 µg.h/mL, respectively. No differences in systemic exposure were observed among dose groups receiving RSG with or without ALN.

Biochemical markers of bone turnover

Procollagen type 1 N-propeptide

Consistent with the effect of OVX, Procollagen type 1 N-propeptide (P1NP) levels were significantly increased for OVX controls compared with sham controls over the treatment period (means at week 12: 9.85 ± 0.86 versus 6.15 ± 0.37 ng/mL, p ≤ 0.05). Treatment of OVX rats with RSG resulted in a nonsignificant decrease in P1NP levels compared with OVX controls at weeks 4 and 8, but not at week 12. ALN had no significant effect on P1NP levels. For RSG/ALN-treated animals, the mean P1NP values were significantly lower at week 4 (p ≤ 0.05) and comparable to OVX controls at weeks 8 and 12 (Supplementary Fig. 1A). A statistically significant (p ≤ 0.05) increase in P1NP level was noted for RSG-treated animals compared with OVX controls 4 weeks after the last dose (week 16, Supplementary Fig. 1B). However, at the end of the recovery period, the mean P1NP value was similar to OVX controls (Supplementary Fig. 1B).

Deoxypyridinoline

Consistent with the effect of OVX, significant increases in deoxypyridinoline (DPD) levels were observed for OVX animals compared with sham controls over the treatment period (week 12: 33.39 ± 1.20 versus 12.58 ± 1.88 nM/mM creatinine, p ≤ 0.05). Treatment of OVX rats with RSG resulted in further significant increases in DPD levels compared with OVX controls (p ≤ 0.05) at week 12. Treatment with ALN/RSG prevented the OVX/RSG-induced increases in DPD levels (p ≤ 0.05) (Supplementary Fig. 1C). Similar to the treatment with ALN alone, the mean DPD values for ALN/RSG-treated animals were significantly lower compared with OVX controls over the treatment period (p ≤ 0.05). At the end of the recovery period, the mean DPD values for RSG-treated animals were comparable to the OVX controls consistent with recovery (Supplementary Fig. 1D).

In vivo densitometry by DXA (lumbar spine and humerus BMC and BMD)

In the OVX group, significant reductions in lumbar spine and humerus BMC and BMD (p ≤ 0.001) were observed compared to sham controls (Fig. 1A, B, D, E). Treatment with RSG significantly reduced (up to 16%) BMD and BMC in the lumbar spine over the treatment period (p ≤ 0.001) and humerus at week 12 (p ≤ 0.05 for both BMC and BMD) compared with OVX controls. As expected, ALN prevented OVX-induced bone loss, with significant increases in BMD and BMC in the lumbar spine (p ≤ 0.001) and humerus (p ≤ 0.05) compared with OVX. Cotreatment with ALN prevented RSG-induced loss of bone, with BMD values comparable to, or higher than, sham controls at week 12 in the lumbar spine and humerus (p ≤ 0.05 for both) (Fig. 1A, B, D, E).

image

Figure 1. Changes over time (as percent change from baseline) in bone mineral content (BMC; A, D) and density (BMD; B, E) for lumbar spine and humerus as measured by DXA during the treatment phase. The recovery phase data is also shown for the lumbar spine (C) and the humerus (F). Values are mean ± SEM. DXA = dual-energy X-ray absorptiometry; OVX = ovariectomized; RSG = rosiglitazone; ALN = alendronate.

Download figure to PowerPoint

The effects of RSG on bone mass reversed during the recovery period relative to the end of treatment (Fig. 1C, F). At the lumbar spine, whereas the OVX group continued to lose bone (p <0.05 between weeks 12 and 20, for BMD), approximately 80% to 100% recovery in densitometry parameters was observed, indicating at least stabilization of RSG-induced bone loss following treatment cessation and even significant increases in BMD (p <0.05 between weeks 12 and 20) to levels comparable with OVX controls (Fig. 1C). For the humerus, although evidence of recovery was noted, the rate of recovery was slower relative to the lumbar spine (Fig. 1F).

In vivo densitometry by pQCT (tibia metaphysis)

In the OVX group, significant reductions in proximal tibia trabecular BMD (p ≤ 0.05), cortical/subcortical BMD (p ≤ 0.05), and total BMD (p ≤ 0.001) were noted compared with sham controls (Fig. 2A–C). Treatment with RSG significantly decreased both total and trabecular and BMD at the proximal tibia compared with OVX controls (p ≤ 0.05). ALN prevented OVX-induced bone loss, with significant differences in total slice BMD and BMC (p ≤ 0.001) compared with OVX. Significant differences were also observed in subcortical/cortical BMD (p ≤ 0.001) and trabecular BMD (p ≤ 0.001) for ALN-treated rats versus OVX controls. Concurrent treatment with ALN significantly reduced the RSG-induced loss of total and trabecular BMD at the proximal tibia (p ≤ 0.001 for all). Similar effects were noted for BMC for all treatments (data not shown). Following cessation of RSG treatment, recovery of cortical/subcortical BMD (data not shown) was relatively slow compared to almost complete recovery of total BMD (p < 0.05 between weeks 12 and 20) (Fig. 2D) and trabecular BMC and BMD (data not shown).

image

Figure 2. Changes over time (percent change from baseline) in trabecular (A), cortical/subcortical (B), and total (C) bone mineral density (BMD) at tibial metaphysis as measured by pQCT during the treatment phase. Total BMD data is also shown for the recovery phase (D). Values are mean ± SEM. OVX = ovariectomized; RSG = rosiglitazone; ALN = alendronate.

Download figure to PowerPoint

Ex vivo densitometry (by pQCT) and bone strength evaluation at the lumbar spine and femur diaphysis

Biomechanical parameters including peak load, apparent strength, yield load, yield stress, stiffness, modulus, AUC, and toughness following the 12-week treatment were measured and selected parameters for the lumbar spine are shown in Fig. 3AC. The effects of RSG on bone densitometry and bone geometry parameters were also measured ex vivo by pQCT (Table 2). Consistent with the effects of OVX, the mean values for bone strength parameters at the lumbar spine for OVX animals were lower relative to sham controls at the end of the treatment period, except for stiffness and modulus. These effects are consistent with slight but nonsignificant decreases in pQCT BMC and BMD and a statistically significant decrease in total slice BMD (p < 0.05). Treatment of OVX rats with RSG decreased bone mass (trabecular and total BMC and BMD, up to 17%) compared to OVX vehicle controls, attaining statistical significance for total slice BMD (Table 2). These decreases were associated with decreases in bone strength parameters for RSG-treated animals (up to 35%), attaining statistical significance for apparent strength and toughness (p < 0.05 for both). ALN increased bone strength parameters at the lumbar spine (except for toughness), attaining statistical significance for peak load, apparent strength, yield load, and yield stress (p < 0.05 for all). These increases were associated with statistically significant increases in BMC and BMD (total and trabecular; p < 0.05). ALN in combination with RSG completely prevented the negative effects of OVX and/or RSG on bone density and bone strength, with values generally comparable to sham vehicle controls (Fig. 3AC, Table 2).

image

Figure 3. Bone biomechanical values for the lumbar spine (apparent strength [A], toughness [B], modulus [C]) as measured by compression testing and for femur (ultimate stress [D], toughness [E], modulus [F]) as measured by three-point bending. Values are mean ± SEM. OVX = ovariectomized; RSG = rosiglitazone; ALN = alendronate.

Download figure to PowerPoint

Table 2. Ex Vivo pQCT Analysis Data for Trabecular (Lumbar Spine) and Cortical (Femur) Bone
 Main studyRecovery study
Surgical status + treatmentSurgical status + treatment
Sham + VehOVX + VehOVX + RSGOVX + ALNOVX + (RSG + ALN)OVX + VehOVX + RSG
  • Values are mean ± SEM. Values in bold are significantly different from respective OVX + Veh group in the main or recovery study.

  • Sham = sham operated; OVX = ovariectomized; Veh = vehicle control; RSG = rosiglitazone; ALN = alendronate; BMC = bone mineral content; BMD = bone mineral density.

  • *

    p ≤ 0.05.

  • **

    p ≤ 0.01.

  • ***

    p ≤ 0.001.

Lumbar spine
Total area (mm2)6.91 ± 0.216.87 ± 0.227.30 ± 0.217.32 ± 0.217.09 ± 0.217.21 ± .397.82 ± 0.31
Total BMC (mg/mm)5.15 ± 0.194.71 ± 0.204.26 ± 0.195.62 ± 0.195.45 ± 0.194.67 ± 0.424.77 ± 0.20
Total BMD (mg/cm3)764.40 ± 15.39*682.79 ± 16.07582.41 ± 15.39***768.14 ± 15.39**766.07 ± 15.39**639.53 ± 22.29610.90 ± 12.39
Femur diaphysis
Cortical BMC (mg/mm)9.45 ± 0.229.54 ± 0.239.26 ± 0.229.10 ± 0.229.95 ± 0.229.52 ± 0.169.36 ± 0.15
Cortical BMD (mg/cm3)1336.72 ± 6.571348.87 ± 6.141345.59 ± 6.971352.80 ± 4.551351.60 ± 7.451350.80 ± 3.121346.40 ± 2.93
Cortical thickness (mm)0.81 ± 0.010.81 ± 0.010.80 ± 0.010.87 ± 0.01*0.86 ± 0.01*0.80 ± 0.010.78 ± 0.00
Periosteal circumference (mm)11.20 ± 0.1711.19 ± 0.1811.01 ± 0.1711.19 ± 0.1711.21 ± 0.1711.30 ± 0.1211.30 ± 0.12
Endosteal circumference (mm)6.20 ± 0.196.22 ± 0.206.03 ± 0.195.79 ± 0.195.85 ± 0.196.34 ± 0.186.50 ± 0.16

The effects of RSG on bone mass and bone strength showed evidence of reversal at the end of the recovery period. The mean values for peak load, apparent strength, stiffness, modulus, AUC, toughness, and BMC (total slice and trabecular) were comparable to OVX vehicle controls (Supplementary Fig. 2A–C).

The effects of study treatments on bone strength were also investigated at the femur (peak load, ultimate stress, stiffness, modulus, AUC, and toughness; with selected data shown in Fig. 3DF). The ex vivo pQCT analysis was also performed at the expected fracture site, the femur diaphysis (Table 2). There were no statistically significant differences in bone densitometry (pQCT) or bone strength parameters for OVX vehicle controls versus sham vehicle controls or RSG-treated animals at the femur diaphysis. Concurrent treatment of OVX rats with RSG/ALN prevented the minimal effects of RSG/OVX on bone strength parameters with values comparable to sham vehicle controls (Fig. 3DF, Table 2).

No statistically significant differences were noted between RSG and OVX controls at the end of the recovery period (Supplementary Fig. 2D–F).

Ex vivo µCT analysis (tibial metaphysis and femur diaphysis)

µCT analysis data for trabecular (tibia metaphysis) and cortical (femur) bone for selected parameters are presented in Table 3. Expected OVX-induced loss of bone mass principally characterized by a statistically significant decrease in relative bone volume (BV/TV, 59%), trabecular number (Tb.N, 32%), and trabecular thickness (Tb.Th, 22%), with statistically significant increase in trabecular separation (Tb.Sp, >300%) was observed compared to sham controls (p ≤ 0.001) RSG exacerbated these effects reaching statistical significance for BV/TV (45%) and Tb.Sp (22%) compared with OVX controls (p ≤ 0.05), respectively. Treatment with ALN prevented the RSG/OVX-induced bone loss. The mean values were significantly increased for BV/TV, Tb.N, and decreased for Tb.Sp compared to OVX vehicle controls (p ≤ 0.001 for all). Co-treatment of RSG/OVX animals with ALN significantly attenuated the RSG effect with values similar to ALN alone relative to OVX vehicle controls. Representative μCT images from each group are shown in Fig. 4.

Table 3. Ex Vivo μCT Analysis Data for Trabecular (Proximal Tibia) and Cortical (Femur) Bone
 Main studyRecovery study
Surgical status + treatmentSurgical status + treatment
Sham + VehOVX + VehOVX + RSGOVX + ALNOVX + (RSG + ALN)OVX + VehOVX + RSG
  • Values are mean ± SEM. Values in bold are significantly different from respective OVX + Veh group in the main or recovery study.

  • Sham = sham operated; OVX = ovariectomized; Veh = vehicle control; RSG = rosiglitazone; ALN = alendronate; BV = bone volume; TV = tissue volume.

  • *

    p ≤ 0.05.

  • **

    p ≤ 0.01.

  • ***

    p ≤ 0.001.

Proximal tibia
Relative bone volume (BV/TV, %)32.74 ± 2.14***13.51 ± 1.637.47 ± 0.70*30.55 ± 2.74***25.55 ± 2.88**8.05 ± 1.318.07 ± 1.58
Trabecular number (Tb.N, 1/mm)5.06 ± 0.19***3.46 ± 0.202.81 ± 0.195.16 ± 0.19***4.88 ± 0.19***2.59 ± 0.162.61 ± 0.13
Trabecular thickness (Tb.Th, mm)0.09 ± 0.00***0.07 ± 0.000.07 ± 0.000.08 ± 0.000.08 ± 0.000.07 ± 0.000.08 ± 0.00
Trabecular separation (Tb.Sp, mm)0.08 ± 0.01***0.29 ± 0.020.37 ± 0.02*0.18 ± 0.02***0.19 ± 0.02***0.40 ± 0.030.40 ± 0.02
Femur diaphysis
Relative bone volume (BV/TV, %)64.95 ± 1.2665.14 ± 1.3265.75 ± 1.2668.12 ± 1.2668.74 ± 1.2664.39 ± 1.1063.76 ± 0.65
Cortical thickness (Ct.Th, mm)0.66 ± 0.010.67 ± 0.020.66 ± 0.010.70 ± 0.010.71 ± 0.010.66 ± 0.010.65 ± 0.01
Cortical porosity (Ct.Po, %)0.23 ± 0.010.22 ± 0.010.22 ± 0.010.23 ± 0.010.21 ± 0.010.17 ± 0.000.16 ± 0.00
image

Figure 4. Representative ex vivo μCT image of proximal tibia at 12 weeks. OVX = ovariectomized; RSG = rosiglitazone; ALN = alendronate.

Download figure to PowerPoint

Evidence of rapid recovery in the tibia was observed at the end of the RSG treatment-free period. The OVX group continued to lose bone during the treatment-free recovery period, whereas no further bone loss occurred in the RSG group. The mean values for μCT parameters for the RSG group were generally comparable to OVX vehicle controls at the end of the recovery period (Table 3).

Neither OVX nor any of the treatments had significant effects on bone parameters in the femoral cortex (femur diaphysis) compared with sham controls (Table 3). No significant or meaningful differences were noted between OVX vehicle controls and RSG-treated animals at the end of the recovery period at the femur diaphysis.

Histomorphometric analysis (tibial metaphysis and diaphysis)

Results from histomorphometric analysis of the tibia for selected parameters are reported in Table 4 for the trabecular and Table 5 for the cortical bone compartments. Consistent with μCT findings, OVX was associated with marked bone loss in the tibial metaphysis versus sham controls (Table 4), as measured by a decrease in BV/TV (p ≤ 0.001). Reduction in Tb. N (p ≤ 0.001) at this site was regarded as the main contributor to the OVX-related microarchitectural deterioration. Consequently, Tb.Sp was significantly increased after OVX whereas Tb.Th was relatively spared (data not shown). OVX was associated with significant increases in osteoblast surface (p≤0.01) and fat tissue volume (p≤0.001). Mineralizing surface (p ≤0.001) and bone formation rates (BFR/BS and BFR/BV, p≤0.001) also increased after OVX. The osteoclast-derived variables, Oc.S/BS and N.Oc/BS, were slightly increased after OVX without attaining significance. These effects are consistent with increased bone turnover and formation.

Table 4. Effect of Treatment on Trabecular Bone Histomorphometry in the Tibial Proximal Metaphysis
 Main studyRecovery
Surgical status + treatmentSurgical status + treatment
Sham + VehOVX + VehOVX + RSGOVX + ALNOVX + (RSG + ALN)OVX + VehOVX + RSG
  • Values are mean ± SEM. Values in bold are significantly different from respective OVX + Veh group in the main or recovery study. Rank-transformed data used for inferential analyses in the main study.

  • Sham = sham operated; OVX = ovariectomized; Veh = vehicle control; RSG = rosiglitazone; ALN = alendronate; BV = bone volume; TV = tissue volume; BS = bone surface; BFR = bone formation rate.

  • *

    p ≤ 0.05.

  • **

    p ≤ 0.01.

  • ***

    p ≤ 0.001.

Relative bone volume (BV/TV, %)25.16 ± 1.36***13.77 ± 1.506.27 ± 1.07**26.89 ± 1.19***23.12 ± 1.43***7.70 ± 1.186.70 ± 1.62
Trabecular number (Tb.N, mm−1)3.98 ± 0.13***2.50 ± 0.211.31 ± 0.21**4.28 ± 0.13***4.07 ± 0.20***1.51 ± 0.231.22 ± 0.25
Osteoblast surface (Ob.S/BS, %)0.55 ± 0.11**8.09 ± 1.778.04 ± 1.690.18 ± 0.04***0.17 ± 0.10***11.40 ± 3.075.30 ± 1.45
Osteoclast surface (Oc.S/BS, %)1.73 ± 0.362.43 ± 0.243.14 ± 0.381.07 ± 0.14*1.40 ± 0.252.93 ± 0.473.48 ± 0.60
Osteoclast number (N.Oc/BS, %)0.50 ± 0.100.72 ± 0.070.95 ± 0.110.29 ± 0.03***0.34 ± 0.06**0.81 ± 0.130.95 ± 0.16
Fat tissue volume (Fa.V/TV, %)2.38 ± 0.49***10.14 ± 1.319.02 ± 2.385.08 ± 0.82*2.47 ± 0.86***15.04 ± 1.9017.04 ± 1.91
Mineralizing surface (MS/BS, %)8.27 ± 1.30***19.56 ± 1.7923.97 ± 2.284.09 ± 0.43***6.28 ± 0.72***17.50 ± 1.6115.0 ± 1.60
Mineral apposition rate (MAR, µm/d)1.40 ± 0.101.44 ± 0.041.56 ± 0.040.67 ± 0.07***0.72 ± 0.12***1.28 ± 0.091.26 ± 0.06
BFR, surface referent (BFR/BS, µm3/µm2/year)42.21 ± 7.98***103.24 ± 10.27136.91 ± 14.4510.36 ± 1.91***16.13 ± 2.65***84.20 ± 10.8570.60 ± 9.69
BFR, volume referent (BFR/BV, %/year)137.93 ± 27.85***376.76 ± 32.49576.27 ± 52.13*32.31 ± 5.03***55.68 ± 7.62***332.90 ± 43.37280.0 ± 35.24
Table 5. Effect of Treatment on Cortical Bone Histomorphometry in the Tibial Diaphysis
 Main studyRecovery
Surgical status + treatmentSurgical status + treatment
Sham + VehOVX + VehOVX + RSGOVX + ALNOVX + (RSG + ALN)OVX + VehOVX + RSG
  • Values are mean ± SEM. Values in bold are significantly different from respective OVX + Veh group in the main or recovery study. Rank-transformed data used for inferential analyses in the main study.

  • Sham = sham operated; Veh = vehicle control; OVX = ovariectomized; RSG = rosiglitazone; ALN = alendronate; Pm = perimeter; MAR = mineral apposition rate; BFR = bone formation rate; BS = bone surface.

  • *

    p ≤ 0.05.

  • **

    p ≤ 0.01.

  • ***

    p ≤ 0.001.

Cortical area (Ct.Ar, mm2)4.45 ± 0.104.82 ± 0.104.36 ± 0.09**4.81 ± 0.094.74 ± 0.144.85 ± 0.094.56 ± 0.07*
Cortical width (Ct.Wi, µm)732.79 ± 15.96782.27 ± 16.05720.13 ± 14.86*776.92 ± 8192783.62 ± 22.40761.0 ± 17.12710.0 ± 14.58*
Endocortical eroded surface (Ec.E.Pm/Ec.Pm, %)3.41 ± 0.867.71 ± 1.2915.68 ± 2.774.43 ± 0.695.72 ± 1.208.90 ± 1.675.50 ± 0.93
Periosteal labeled surface (Ps.L.Pm/Ps.Pm, %)14.70 ± 4.97**47.71 ± 6.6832.67 ± 5.9251.55 ± 5.2736.94 ± 5.257.30 ± 2.3710.90 ± 3.38
Periosteal MAR (Ps.MAR, µm/d)0.96 ± 0.101.32 ± 0.111.14 ± 0.071.27 ± 0.041.09 ± 0.090.78 ± 0.081.03 ± 0.08*
Periosteal BFR, surface referent (Ps.BFR/BS, µm3/µm2/year)67.09 ± 23.70**248.76 ± 50.39146.53 ± 28.89243.31 ± 29.80149.43 ± 24.7624.00 ± 8.8947.40 ± 15.79
Endocortical labeled surface (Ec.L.Pm/Ec.Pm, %)2.42 ± 0.43***6.82 ± 1.296.29 ± 0.402.44 ± 0.47***2.99 ± 0.54**6.27 ± 0.937.84 ± 0.86
Endocortical MAR (Ec.MAR, µm/d)0.60 ± 3.70**1.20 ± 0.161.56 ± 0.210.66 ± 0.06*0.73 ± 0.090.77 ± 0.940.94 ± 0.09
Endocortical BFR, surface referent (Ec.BFR/BS, µm3/µm2/year)5.31 ± 0.95*32.33 ± 8.5035.26 ± 5.076.21 ± 1.54*9.05 ± 2.7219.60 ± 5.0028.36 ± 5.00

RSG exacerbated OVX-induced bone loss, resulting in microarchitectural deterioration associated with a further statistically significant reduction in relative bone volume and trabecular number (p ≤ 0.01 for both) compared to OVX controls. There were no evident or consistent variations in cellular or label-derived variables, such as increases in bone resorption variables and/or decreases in bone formation variables that explained the increased bone loss associated with RSG treatment in the OVX rat model. The extent of bone loss left little bone available for bone surface measurements and likely increased the individual variability. Treatment with ALN prevented OVX-induced cancellous bone loss in the tibial metaphysis, with relative bone volume and trabecular number significantly increased compared with OVX controls (p ≤ 0.001 for both) and similar to sham controls (Table 4). ALN prevented OVX-induced increases in bone formation variables (osteoblast surface, mineralizing surface, BFR/BS, and BFR/BV, p ≤ 0.001 for all). ALN also decreased bone resorption compared with OVX controls as shown by significant reductions in osteoclast surface and osteoclast number. These variables in the ALN group were either comparable or even lower than those of the sham vehicle controls, indicating that the compound had completely prevented the OVX-related increase in bone turnover. Concurrent treatment with RSG and ALN significantly prevented not only the BV/TV and Tb.N reductions due to RSG alone but also most of the OVX-induced bone loss in the tibial metaphysis. Compared with OVX rats, the combination therapy also significantly reduced many bone formation variables (Ob.S/BS, MS/BS, MAR, BFR/BS, and BFR/BV) as well as the bone resorption parameter N.Oc/BS. The amount of fat in the marrow was significantly decreased likely due to the relative increase in cancellous bone. Overall, these effects were very similar to those noted with the ALN treatment, suggesting that this compound could not only prevent the OVX-related bone loss but also the exacerbated bone loss associated with RSG (Table 4).

Recovery period in trabecular bone

The OVX group continued to lose bone during the treatment-free period, whereas no further bone loss occurred in the RSG-treated group based upon comparable bone volume at both time points. Therefore, the RSG-induced bone loss did not become worse after the treatment cessation but improved relative to the controls at the end of the recovery period. There were no meaningful or consistent effects upon the variables related to bone formation and turnover in the RSG treatment group compared to the OVX vehicle controls at the end of the recovery period (Table 4).

Cortical bone effects

Compared to the sham group, OVX had no significant effect on any of the structural variables measured in the tibial diaphysis (Table 5), although dynamic responses to OVX were observed. These included an increased labeled surface and bone formation rate at the periosteal compartment (p ≤ 0.01), and increased labeled surface, mineral apposition and bone formation rates at the endocortical compartment (p ≤ 0.05 for all). Endocortical eroded surface, an indicator of bone resorption, was also increased (126%, nonsignificant) as a result of OVX. Whereas no further meaningful dynamic effects of RSG treatment at the periosteal and endocortical surfaces were observed, it significantly reduced cortical area and width (p≤0.05 for both) but minimally increased endocortical eroded surface (103%, nonsignificant) compared with OVX controls. ALN completely and significantly prevented OVX-related increases in endocortical bone turnover and formation, because Ec.L.Pm/Ec.Pm, Ec.MAR, and Ec.BFR/BS were comparable to those of the Sham controls. There were no meaningful effects of ALN treatment upon bone dynamics at the periosteal surface. Concomitant treatment with RSG and ALN significantly prevented OVX-related increases in endocortical bone turnover as shown by decreased endocortical labeled surface (56%, p ≤ 0.01). Overall, these effects of RSG plus ALN in the tibia were markedly similar to those of ALN alone, suggesting that ALN offset the negative effects of RSG on cortical bone structure (Table 5).

Recovery period in cortical bone

After the 8-week treatment-free period, the decreases in cortical bone structure caused by RSG persisted. The significant reductions in cortical area and width (Table 5) observed at the end of RSG treatment in the tibial diaphysis were still present and were consistent with in vivo pQCT findings. These effects were related to medullary expansion as indicated by significantly increased endocortical perimeter and relative medullary area (data not shown).

Discussion

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

The study reported here is a comprehensive evaluation of the effects of RSG, a PPARγ agonist, and ALN on bone turnover markers, mass, strength, and microarchitecture using the rat OVX model. Combination therapy with RSG plus ALN was also evaluated to determine whether concurrent treatment prevented the effects of RSG on bone health. In addition, an off-treatment recovery period was employed to examine the reversibility of the effect of RSG after cessation of RSG treatment. Previous studies have examined the effect of TZDs on bone mass in the context of bone turnover but all-inclusive assessment of bone quality, strength and histomorphometry were often not conducted.

To our knowledge this is the first study in which the pharmacokinetics and exposure of RSG was determined to allow comparison with clinical exposures. The systemic exposure obtained with 10 mg/kg/d of RSG in female Sprague-Dawley rats is predicted to be approximately 50 times more than the equivalent human exposure at 8 mg/d. The relatively high dose used in this study was intended to magnify the effects of RSG on bone and were consistent with dosages used in previous studies by other investigators.[18]

Treatment with RSG resulted in significant increases in DPD (49%) levels, compared to OVX vehicle controls. Although no significant effect was noted on bone formation marker P1NP, a strong rebound effect during early phase of the recovery period (4 weeks after treatment withdrawal) suggested some suppression of bone formation markers during the treatment period. At the end of the recovery, no differences were noted on P1NP levels. Consistent with the apparent lack of a significant effect on bone formation markers during the treatment period (week 12) and at the end of the recovery period (week 20), no RSG-related effects were observed on BFR, MAR, or number of osteoblasts at the end of the treatment or end of the recovery period as determined by histomorphometric analysis. The increase in bone resorption suggested by an increases in DPD levels were not confirmed by the histomorphometry findings (Oc.S/BS or N.Oc/BS), likely due to the high individual variability noted for histomorphometry analysis and/or the timing of the histomorphometric analysis that was conducted only at the end of the 12-week treatment. The substantial loss of trabecular bone induced by OVX and/or RSG resulted in little bone available for bone surface measurements and contributed to the individual variability. Previously reported in vitro findings demonstrate direct effects of RSG on both osteoblasts and osteoclasts, which are in agreement with the effects on both bone formation and bone resorption.[28]

The loss of bone mass caused by OVX was further exacerbated by RSG treatment (in vivo and/or ex vivo pQCT, DXA, µCT, histomorphometry) resulting in weaker bone, particularly for sites rich in trabecular bone. As expected, treatment of OVX rats with ALN increased bone mass and bone strength. Furthermore, ALN completely prevented RSG/OVX-induced losses in bone mass and bone strength. The effect of ALN was mediated by decreases in bone resorption and formation, as evidenced by decreases in osteoclast, osteoblast, and label-derived histomorphometry parameters, and in the bone resorption marker DPD. These findings are consistent with the established pharmacological action of ALN. The effectiveness of ALN, as an antiresorptive agent, in preventing RSG-induced bone loss, further confirms that the negative effect of RSG on bone is mediated mainly by increased bone resorption both in this study and another recently published study using the OVX Zucker fatty rat.[29] However, based on significant increase noted for P1NP during the early phase of recovery at week 16, a moderate effect to reduce bone formation cannot be completely excluded.

One of the key outstanding questions is whether the effects of RSG are reversible and this is the first study to address this issue. The effects of RSG on bone densitometry and bone strength showed evidence of reversal after drug withdrawal. At the tibial metaphysis and L3 vertebral body (sites rich in trabecular bone with red marrow); greater effects of RSG were noted on bone mass relative to other sites, with a faster recovery after drug withdrawal. Interestingly, at the calcaneus (data not shown), the magnitude of response was smaller with slower recovery (as analyzed by pQCT and µCT). The differences in the magnitude of response and the extent of recovery may be associated with a high proportion of adipocytes, the absence of hematopoietic cells, and high proportion of cortical bone. Dynamic histomorphometry parameters also indicated a lower bone turnover rate at the calcaneus (data not shown) relative to the tibial metaphysis or L3 vertebral body. Consistent with aging, bones were generally weaker for OVX vehicle controls or RSG-treated animals at the end of the recovery period compared to the end of the treatment period, most notably at sites rich in trabecular bone.

Histomorphometric data indicated that ovariectomy decreased bone volume fraction, trabecular number, and increased trabecular separation, mineralizing surface, and bone formation rate in trabecular bone. In cortical bone, ovariectomy increased periosteal and endocortical bone turnover. Overall, RSG exacerbated OVX-induced microarchitectural deteriorations in trabecular bone, with less pronounced effects in the periosteal and endocortical surface of cortical bone. These RSG-induced changes are consistent with the effect of RSG to significantly increase urinary DPD levels and decrease bone mass. These outcomes are supported by findings from a previous µCT analysis in aged normal mice that reported decreased bone volume, trabecular width and number, and increased trabecular spacing following RSG treatment.[24]

Effects of treatment on histomorphometric parameters did not appear to improve following treatment cessation, with reduced but stable values of trabecular and cortical bone measures persisting throughout the treatment-free period. It is worth noting that during the recovery period in this study, OVX controls continued to lose bone whereas relative bone volume in the RSG group remained comparatively unchanged. These data strongly suggest that cessation of treatment reversed or attenuated the bone loss that occurred with RSG treatment. In general, histomorphometry analyses were concordant with the bone densitometry (pQCT, µCT, and DXA) and bone strength parameters for OVX controls and animals treated with ALN alone or in combination with RSG. However, some inconsistencies were noted for animals treated with RSG. These inconsistencies were likely related to the smaller bone surface (as a result of OVX, RSG, and/or aging process) and expected higher variability for histomorphometry. Furthermore, the changes in bone turnover markers were greater at weeks 4 to 8 and at week 16 compared to the end of the treatment (week 12) and recovery periods (week 20), respectively. This suggests that the histomorphometry analysis performed postnecropsy may not have been an optimal time point to capture the subtle changes in the dynamic parameters that may have occurred earlier.

PPAR-γ is generally accepted as a central determinant of adipocyte differentiation from bone marrow mesenchymal cells.[6, 9] In this respect, the PPAR-γ-agonist effects of RSG may have stimulated differentiation of cells of the osteoblast lineage to become adipocytes while simultaneously suppressing osteoblastogenesis, as reported previously.[9, 28, 30] Thus, reduced numbers of osteoblasts have the potential to decrease bone turnover and uncouple balanced remodeling, which may lead to a relative increase in osteoclast activity and bone resorption. Indeed, in trabecular bone, ovariectomy consistently increased fat tissue and fat marrow volume, an effect exacerbated by treatment with RSG. Interestingly, it has been recently reported that concomitant treatment with fenofibrate, a PPAR-α-agonist, can attenuate the negative effects of pioglitazone, a PPAR-γ-agonist.[11]

The findings of this study clearly demonstrate that treatment of OVX rats with RSG, results in loss of bone mass and bone strength. This is consistent with currently available clinical and preclinical data suggesting that treatment of diabetes with TZDs compounds targeting PPARγ are associated with bone loss and an increased risk of fracture.[1-3, 5, 31] Although a nonsignificant trend for decreases in bone strength and/or bone mass were noted at sites rich in cortical bone, these effects were more profound at sites rich in trabecular bone. The A Diabetes Outcome Progression Trial (ADOPT) study showed the largest increases in relative fracture risk for bones of the extremities such as foot, hand, and proximal humerus.[1, 32] Because the bones of the extremities have a relatively high proportion of cortical bone, it was suggested that RSG would have the largest impact only on cortical bone. However, the results of ADOPT were confounded by the small number of typical osteoporotic hip and spine fractures in the population participating in the study (average age 57 years) and the fact that the rate of these fractures tends to be relatively low until after age 65 years.[33] In addition, the negative effects of RSG on trabecular bone (hip and spine) were reported in another small clinical study with nondiabetic postmenopausal women.[34] Their findings showed treatment with RSG significantly decreased hip BMD relative to placebo group. Similar results were noted at the lumbar spine; however, the differences did not attain statistical significance.

Although long bones have a relatively high proportion of cortical bone (diaphysis), the extremities also contain significant amount of trabecular bone; therefore, RSG may induce loss of trabecular bone at the extremities of the long bones in humans. Alternatively, the effect of RSG in humans may be different, considering the differences in anatomy and bone biology between rats and humans. The continuous longitudinal growth of bone and the low incidence of Haversian systems in rats are clearly two important factors that likely influence the outcome of treatment on cortical bone and may explain the differences in the results. It is important to note that our study was conducted in OVX rats that were otherwise normal and not diabetic animals. Therefore, it is also possible that the effect of RSG may be different under diabetic conditions compared with normal OVX rats. The choice of normal OVX rats was based on our desire to mimic the postmenopausal condition in humans and rule out any confounding effect of diabetes on bone because diabetes is known to adversely impact bone.[35] Furthermore, in a previous study, PPARγ or PPARα and γ agonists were shown to counter osteopenia observed in a rat model of diabetes, the Zucker fatty rats.[23] However, in a recent study, treatment of OVX Zucker fatty rats with RSG resulted in findings similar to our study, including a decrease in vertebral BMD, bone volume, and trabecular thickness. These changes were also accompanied by a decrease in the mechanical properties of both trabecular and cortical bones. Interestingly, an increase in cortical porosity was also noted,[29] in contrast to a lack of profound effect on cortical bone observed in our study. It is important to note that cortical porosity was measured only by μCT and not confirmed by histomorphometry.[29] The pores appeared to be limited at, or near to, the endocortical surface and could be the result of actual endocortical resorption rather than true cortical porosity (generally due to Haversian remodeling in higher species). Nonetheless, the mechanism of bone loss was attributed to increased bone resorption as minimal effects on bone formation were observed,[29] similar to our study. More importantly, the authors concluded that diabetes itself was not a significant confounder of the RSG actions on bone, such that the results using normal OVX rats or OVX Zucker fatty rats are generally comparable.[29]

In summary, administration of RSG to OVX rats for 12 weeks significantly reduced bone mass compared with OVX controls, predominantly at sites rich in trabecular bone. These effects appear to be primarily due to increased bone resorption with marginal reductions in bone formation. Two of the most important and unique observations from this study are that the loss of bone mass following RSG administration was largely prevented by concurrent treatment with an antiresorptive agent and that the negative effects of RSG on bone were reversible or attenuated. Further clinical studies are required to determine whether TZD-induced bone loss can be wholly or partially prevented by antiresorptive therapies.

Disclosures

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

SK, SJH, and LAF are employees and stockholders of GlaxoSmithKline. RS and PM are employees of Charles River. SYS and JJ are employees and stockholders of Charles River. RG is an employee of Georgia Institute of Technology and consultant for Charles River.

Acknowledgments

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

We acknowledge the excellent technical support of the imaging, histomorphometry and biomechanics teams at Charles River, the bioanalysis group at GlaxoSmithKline, and Duncan Pennington at International Medical Press for editorial assistance in the preparation of this manuscript.

Authors' roles: SK, LAF, SJH, and SYS participated in the conception and design of the study. SK, LAF, RS, SYS, JJ, and RG carried out the analysis and interpretation of data. RS, JJ, and PM collected the data, with the Charles River central support group carrying out the statistical analysis. SK, SJH, LAF, RS, SYS, PM, JJ, and RG contributed to the writing and critical review of the manuscript, and all authors read and approved the final version of the manuscript.

References

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information
  • 1
    Kahn SE, Zinman B, Lachin JM, Haffner SM, Herman WH, Holman RR, Kravitz BG, Yu D, Heise MA, Aftring RP, Viberti G. Rosiglitazone-associated fractures in type 2 diabetes: an Analysis from A Diabetes Outcome Progression Trial (ADOPT). Diabetes Care. 2008; 31:84551.
  • 2
    Glintborg D, Andersen M, Hagen C, Heickendorff L, Hermann AP. Association of pioglitazone treatment with decreased bone mineral density in obese premenopausal patients with polycystic ovary syndrome: a randomized, placebo-controlled trial. J Clin Endocrinol Metab. 2008; 93:1696701.
  • 3
    Dormandy J, Bhattacharya M, van Troostenburg de Bruyn AR. Safety and tolerability of pioglitazone in high-risk patients with type 2 diabetes: an overview of data from PROactive. Drug Saf. 2009; 32:187202.
  • 4
    Douglas IJ, Evans SJ, Pocock S, Smeeth L. The risk of fractures associated with thiazolidinediones: a self-controlled case-series study. PLoS Med. 2009; 6:e1000154.
  • 5
    Schwartz A. TZDs and bone: a review of the recent clinical evidence. PPAR Res [Internet]. 2008 [cited 2013 Mar 6];2008:297893. Available from: http://www.hindawi.com/journals/ppar/2008/297893.
  • 6
    Beresford JN, Bennett JH, Devlin C, Leboy PS, Owen ME. Evidence for an inverse relationship between the differentiation of adipocytic and osteogenic cells in rat marrow stromal cell cultures. J Cell Sci. 1992; 102(Pt 2):34151.
  • 7
    Nuttall ME, Gimble JM. Is there a therapeutic opportunity to either prevent or treat osteopenic disorders by inhibiting marrow adipogenesis?. Bone. 2000; 27:17784.
  • 8
    Rosen ED. The transcriptional basis of adipocyte development. Prostaglandins Leukot Essent Fatty Acids. 2005; 73:314.
  • 9
    Gimble JM, Robinson CE, Wu X, Kelly KA, Rodriguez BR, Kliewer SA, Lehmann JM, Morris DC. Peroxisome proliferator-activated receptor-gamma activation by thiazolidinediones induces adipogenesis in bone marrow stromal cells. Mol Pharmacol. 1996; 50:108794.
  • 10
    Rzonca SO, Suva LJ, Gaddy D, Montague DC, Lecka-Czernik B. Bone is a target for the antidiabetic compound rosiglitazone. Endocrinology. 2004; 145:4016.
  • 11
    Samadfam R, Awori M, Benardeau A, Bauss F, Sebokova E, Wright M, Smith SY. Combination treatment with pioglitazone and fenofibrate attenuates pioglitazone-mediated acceleration of bone loss in ovariectomized rats. J Endocrinol. 2012; 212:17986.
  • 12
    Ali AA, Weinstein RS, Stewart SA, Parfitt AM, Manolagas SC, Jilka RL. Rosiglitazone causes bone loss in mice by suppressing osteoblast differentiation and bone formation. Endocrinology. 2005; 146:122635.
  • 13
    Soroceanu MA, Miao D, Bai XY, Su H, Goltzman D, Karaplis AC. Rosiglitazone impacts negatively on bone by promoting osteoblast/osteocyte apoptosis. J Endocrinol. 2004; 183:20316.
  • 14
    Mbalaviele G, Abu-Amer Y, Meng A, Jaiswal R, Beck S, Pittenger MF, Thiede MA, Marshak DR. Activation of peroxisome proliferator-activated receptor-gamma pathway inhibits osteoclast differentiation. J Biol Chem. 2000; 275:1438893.
  • 15
    Okazaki R, Toriumi M, Fukumoto S, Miyamoto M, Fujita T, Tanaka K, Takeuchi Y. Thiazolidinediones inhibit osteoclast-like cell formation and bone resorption in vitro. Endocrinology. 1999; 140:50605.
  • 16
    Schwab AM, Granholm S, Persson E, Wilkes B, Lerner UH, Conaway HH. Stimulation of resorption in cultured mouse calvarial bones by thiazolidinediones. Endocrinology. 2005; 146:434961.
  • 17
    Li M, Amizuka N, Takeuchi K, Maeda T. Histochemical examination on the osteoclastic function on the microenvironment of extracellular matrices in osteolytic metastasis. J Bone Miner Res. 2005; 20:S1678.
  • 18
    Sottile V, Seuwen K, Kneissel M. Enhanced marrow adipogenesis and bone resorption in estrogen-deprived rats treated with the PPARgamma agonist BRL49653 (rosiglitazone). Calcif Tissue Int. 2004; 75:32937.
  • 19
    Wan Y, Chong LW, Evans RM. PPAR-gamma regulates osteoclastogenesis in mice. Nat Med. 2007; 13:1496503.
  • 20
    Lazarenko OP, Rzonca SO, Suva LJ, Lecka-Czernik B. Netoglitazone is a PPAR-gamma ligand with selective effects on bone and fat. Bone. 2006; 38:7484.
  • 21
    Tornvig L, Mosekilde LI, Justesen J, Falk E, Kassem M. Troglitazone treatment increases bone marrow adipose tissue volume but does not affect trabecular bone volume in mice. Calcif Tissue Int. 2001; 69:4650.
  • 22
    Lecka-Czernik B, Lazarenko O, Teng K, Suva L. Activation of PPAR-gamma by rosiglitazone induces bone loss due to changes in the phenotype and function of bone mesenchymal cells. Second Joint Meeting of ECTS/IBMS, June 25–29, 2005, Geneva, Switzerland. MIS 2005 (2569163):P559-Su.
  • 23
    Shibata T, Takeuchi S, Yokota S, Kakimoto K, Yonemori F, Wakitani K. Effects of peroxisome proliferator-activated receptor-alpha and -gamma agonist, JTT-501, on diabetic complications in Zucker diabetic fatty rats. Br J Pharmacol. 2000; 130:495504.
  • 24
    Lazarenko OP, Rzonca SO, Hogue WR, Swain FL, Suva LJ, Lecka-Czernik B. Rosiglitazone induces decreases in bone mass and strength that are reminiscent of aged bone. Endocrinology. 2007; 148:266980.
  • 25
    Bilezikian JP. Efficacy of bisphosphonates in reducing fracture risk in postmenopausal osteoporosis. Am J Med. 2009; 122:S1421.
  • 26
    Bouxsein ML, Boyd SK, Christiansen BA, Guldberg RE, Jepsen KJ, Muller R. Guidelines for assessment of bone microstructure in rodents using micro-computed tomography. J Bone Miner Res. 2010; 25:146886.
  • 27
    Parfitt AM, Drezner MK, Glorieux FH, Kanis JA, Malluche H, Meunier PJ, Ott SM, Recker RR. Bone histomorphometry: standardization of nomenclature, symbols, and units. Report of the ASBMR Histomorphometry Nomenclature Committee. J Bone Miner Res. 1987 Dec; 2(6):595610.
  • 28
    Benvenuti S, Cellai I, Luciani P, Deledda C, Baglioni S, Giuliani C, Saccardi R, Mazzanti B, Dal Pozzo S, Mannucci E, Peri A, Serio M. Rosiglitazone stimulates adipogenesis and decreases osteoblastogenesis in human mesenchymal stem cells. J Endocrinol Invest. 2007; 30:RC2630.
  • 29
    Sardone LD, Renlund R, Willett TL, Fantus IG, Grynpas MD. Effect of rosiglitazone on bone quality in a rat model of insulin resistance and osteoporosis. Diabetes. 2011; 60:32718.
  • 30
    Lecka-Czernik B, Suva LJ. Resolving the two “bony” faces of PPAR-gamma. PPAR Res [Internet]. 2006 [cited 2013 Mar 6];2006:27489. Available from: http://www.hindawi.com/journals/ppar/2006/027489/abs/.
  • 31
    Home PD, Pocock SJ, Beck-Nielsen H, Curtis PS, Gomis R, Hanefeld M, Jones NP, Komajda M, McMurray JJ. Rosiglitazone evaluated for cardiovascular outcomes in oral agent combination therapy for type 2 diabetes (RECORD): a multicentre, randomised, open-label trial. Lancet. 2009; 373:212535.
  • 32
    Borges JL, Bilezikian JP, Jones-Leone AR, Acusta AP, Ambery PD, Nino AJ, Grosse M, Fitzpatrick LA, Cobitz AR. A randomized, parallel group, double-blind, multicentre study comparing the efficacy and safety of Avandamet (rosiglitazone/metformin) and metformin on long-term glycaemic control and bone mineral density after 80 weeks of treatment in drug-naive type 2 diabetes mellitus patients. Diabetes Obes Metab. 2011; 13:103646.
  • 33
    Viberti G, Lachin J, Holman R, Zinman B, Haffner S, Kravitz B, Heise MA, Jones NP, O'Neill MC, Freed MI, Kahn SE, Herman WH. A Diabetes Outcome Progression Trial (ADOPT): baseline characteristics of Type 2 diabetic patients in North America and Europe. Diabet Med. 2006; 23:128994.
  • 34
    Grey A, Bolland M, Gamble G, Wattie D, Horne A, Davidson J, Reid IR. The peroxisome proliferator-activated receptor-gamma agonist rosiglitazone decreases bone formation and bone mineral density in healthy postmenopausal women: a randomized, controlled trial. J Clin Endocrinol Metab. 2007; 92:130510.
  • 35
    Moseley KF. Type 2 diabetes and bone fractures. Curr Opin Endocrinol Diabetes Obes. 2012; 19:12835.

Supporting Information

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgments
  9. References
  10. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
jbmr1918-sm-0001-SupFig1.eps1047KSupplementary Figure 1
jbmr1918-sm-0002-SupFig2.eps986KSupplementary Figure 2
jbmr1918-sm-0003-SupFigLegend.rtf35KSupplementary Figure Legend

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.