Pharmacological inhibition of PPARγ increases osteoblastogenesis and bone mass in male C57BL/6 mice



Infiltration of bone marrow with fat is a prevalent feature in people with age-related bone loss and osteoporosis, which correlates inversely with bone formation and positively with high expression levels of peroxisomal proliferator-activated receptor gamma (PPARγ). Inhibition of PPARγ thus represents a potential therapeutic approach for age-related bone loss. In this study, we examined the effect of PPARγ inhibition on bone in skeletally mature C57BL/6 male mice. Nine-month-old mice were treated with a PPARγ antagonist, bisphenol-A-diglycidyl ether (BADGE), alone or in combination with active vitamin D (1,25[OH]2D3) for 6 weeks. Micro-computed tomography and bone histomorphometry indicated that mice treated with either BADGE or BADGE + 1,25(OH)2D3 had significantly increased bone volume and improved bone quality compared with vehicle-treated mice. This phenotype occurred in the absence of alterations in osteoclast number. Furthermore, the BADGE + 1,25(OH)2D3-treated mice exhibited higher levels of unmineralized osteoid. All of the treated groups showed a significant increase in circulating levels of bone formation markers without changes in bone resorption markers, while blood glucose, parathyroid hormone, and Ca+ remained normal. Furthermore, treatment with BADGE induced higher levels of expression of vitamin D receptor within the bone marrow. Overall, treated mice showed higher levels of osteoblastogenesis and bone formation concomitant with decreased marrow adiposity and ex vivo adipogenesis. Taken together, these observations demonstrate that pharmacological inhibition of PPARγ may represent an effective anabolic therapy for osteoporosis in the near future. © 2013 American Society for Bone and Mineral Research.


Age-related bone loss starts after the third decade of life in men and women, predisposing them to develop osteoporosis. The mechanisms of age-related bone loss and osteoporosis in women are primarily associated with estrogen deficiency, which increases bone resorption, whereas those in men remain poorly understood but are mostly linked to low bone formation.1, 2 The vast majority of therapeutic interventions for osteoporosis inhibit bone resorption and favor secondary mineralization,1 which makes them highly effective in the context of high levels of bone resorption. However, there is also a significant reduction in bone formation with age. Therefore, the stimulation of bone formation would be a useful alternative in the treatment of osteoporosis in both male and female older persons.1, 3

One of the approaches to increase bone formation is to stimulate the differentiation of mesenchymal stem cells (MSC) into osteoblasts.1, 4 The progressive accumulation of marrow fat is one of the invariable features of age-related bone loss and osteoporosis in older persons,5 which is the result of a shift in the differentiation program of MSC from osteogenic in young bone into adipogenic in older bone.6 This switch in the differentiation program is associated with age-related changes in the expression of transcription factors including the widely described reduction in expression of the osteogenic runt-related transcription factor 2 (Runx2) and increased expression of the adipogenic transcription factor peroxisome proliferator-activated receptor gamma (PPARγ).7

Correction of the imbalance between osteoblastogenesis and adipogenesis in aging bone has been suggested as one of the potential therapeutic approaches to osteoporosis.6 It follows that either increasing Runx2 or decreasing PPARγ would have a beneficial effect on osteoblastogenesis, which would improve bone quality by increasing bone formation and decreasing marrow fat. Interestingly, the induction of high levels of Runx2 using transgenic technology resulted in a paradoxical low bone mass because of high bone turnover and a decrease in mineralization.8 In contrast, Akune and colleagues9 investigated the effect of inhibiting PPARγ in bone using a mouse model of PPARγ haploinsufficiency. Heterozygous PPARγ-deficient mice exhibited high bone mass with increased osteoblastogenesis, associated with normal osteoblast and osteoclast function. This osteogenic effect of PPARγ haploinsufficiency became prominent with advancing age, was not changed by ovariectomy, and was not mediated by insulin or leptin.

In 2000, Wright and colleagues10 reported that bisphenol-A-diglycidyl ether (BADGE), a synthetic inhibitor of PPARγ, reduced adipocyte differentiation. BADGE is a ligand for PPARγ with a Kd of approximately 100 mM. It can antagonize the ability of ligands such as rosiglitazone, used in the treatment of diabetes, to activate the transcriptional and adipogenic action of this receptor.

Subsequently, Botolin and colleagues11 used BADGE in diabetic mice to determine if PPARγ inhibition would decrease marrow fat while increasing osteoblast differentiation and bone formation. In their study, they used control and insulin-deficient diabetic BALB/c mice, which were treated with BADGE to block adipocyte differentiation. In their study, treatment with 30 mg/Kg of BADGE did not prevent diabetes-associated hyperglycemia or weight loss but did prevent diabetes-induced hyperlipidemia and effectively blocked diabetes type I-induced bone adiposity. In addition, although BADGE treatment decreased marrow fat, it did not prevent the suppression of osteoblast markers (Runx2 and osteocalcin [OCN]) and bone loss usually associated with diabetes type I. These negative results were probably because of the use of a diabetic mouse model in which the inhibitory effect of diabetes type I on bone formation could have exceeded the potential beneficial effect of BADGE on osteoblast differentiation and function.

The objective of the current study was to determine if BADGE alone, or in combination with the bone anabolic agent 1,25 dihydroxy-vitamin D3 (1,25[OH]2D3),12 could promote bone formation in skeletally mature mice by binding to and inhibiting the activity of PPARγ. The results demonstrate that pharmacological inhibition of PPARγ could become a useful anabolic treatment for osteoporosis.

Materials and Methods


We purchased 9-month-old C57BL/6 mice (n = 8 per group) from Jackson Laboratory (Bar Harbor, ME, USA). After acclimation, mice received treatment for a total of 6 weeks with either BADGE alone (daily intra-peritoneal [IP] injection of 30 mg/kg in 15% DMSO) or in combination with 1,25(OH)2D3 (subcutaneous osmotic minipump at a dose of 18 pm/d) (Alzet, Cupertino, CA, USA) implanted as previously described.12 Mice treated with daily ip injections of normal saline and implanted with PBS filled pumps were used as controls. Mice were housed in cages in a limited access room. Animal husbandry adhered to Canadian Council on Animal Care Standards, and all protocols were approved by the McGill University Health Center Animal Care Utilization Committee.

Reagents, antibodies, and media

BADGE and 1,25(OH)2D3 were purchased from Sigma-Aldrich (Montreal, Canada). Cell culture reagents were purchased from Sigma-Aldrich unless otherwise specified. Antibodies for Western blotting were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA).

Radiography and micro-computed tomography (µCT) analysis

Radiography and µCT were performed essentially as described previously.12, 13 Mice were administered a lethal dose of anesthetic at the indicated times, exsanguinated, and imaged using a Faxitron MX20 equipped with an FPX-2 Imaging system (Dalsa Medoptics, Waterloo, Canada). µCT was performed on the left femur after removal of soft tissues and overnight fixation in 4% paraformaldehyde. The distal metaphysis was scanned with a Skyscan 1072 µCT instrument (Skyscan, Antwerp, Belgium). Image acquisition was performed at 100 kV and 98 µA, with a 0.9° rotation between frames. The two-dimensional images were used to generate three-dimensional reconstructions to obtain quantitative data with the 3D Creator software supplied with the instrument. Nomenclature and abbreviations of 3D-µCT parameters follow the recommendations of the American Society of Bone and Mineral Research.14

Histological and histomorphometrical analysis of bone

The details of these methods were described previously.12, 13 Detection of alkaline phosphatase (ALP) and tartrate-resistant acid phosphatase (TRAP) activity was carried out as previously described.13 Naphtol-AS-TR (Sigma-Aldrich Canada Ltd., Oakville, Canada) was used as substrate for both enzymes; Fast Blue BB salt (Sigma-Aldrich Canada Ltd.) was used as a coupler for ALP.

For dynamic histomorphometry, tetracycline labeling was achieved through the intraperitoneal injection of demeclocycline (20 mg/kg) (Sigma Chemicals, St. Louis, MO, USA) to the three treated groups at 5 and 2 days before sacrifice. One femur from each animal in each group was removed at the time of euthanization, fixed in 70% ethanol, dehydrated, and embedded undecalcified in methylmethacrylate (J-T Baker, Phillipsburg, NJ, USA). At 50-µm intervals, longitudinal sections of 5 and 8 µm thick were cut using a polycut-E microtome (Reichert-Jung Leica, Heerbrugg, Switzerland), placed on gelatin-coated glass slides, deplastified, and stained with Goldner's trichrome. Histomorphometry was done with a semi-automatic image analyzing system combining a microscope equipped with a camera lucida and digitizing tablet linked to a computer using the OsteoMeasure Software (Osteometrics Inc., Decatur, GA, USA). Nomenclature and abbreviations of histomorphometric parameters follow the recommendations of the American Society of Bone and Mineral Research.15

For immunofluorescence, sections were incubated overnight at 4°C with both a rabitt monoclonal antibody IgG against PPARγ (anti-rabitt PPARγ, sc-28001-R, Santa Cruz Biotechnology) and a mouse monoclonal antibody IgG against vitamin D receptor (VDR) (anti-mouse VDR, sc-13133, Santa Cruz Biotechnology). Primary antibodies were detected by incubation with a fluorescein-conjugated donkey anti-rabitt IgG secondary antibody (1:500 in BSA 1%, Alexa Fluor, Life Technologies, Gaithersburg, MD, USA) for PPARγ and a red conjugated donkey anti-mouse IgG secondary antibody (1:500 in BSA 1%, Alexa Fluor) for VDR. After washing with PBS, the sections were mounted in 90% glycerol and covered with glass cover-slips. Photographs were taken under an Olympus fluorescence microscope (x10 magnification) controlled by an IPLab system. Brightness, overlap, and contrast adjustments were performed in Adobe Photoshop (San Jose, CA, USA). Image-J image analysis software was used to analyze fluorescent images. Threshold intensities for each picture were determined, and from this the number of pixels within the picture that were above the threshold intensity were calculated and considered positively stained. The positively stained pixels were expressed as a percentage of the total pixels in the picture. The ratio between levels of expression of VDR/PPARγ was then calculated, with higher ratio indicating increasing levels of VDR expression in the same bone section.

Biochemical analysis

Mice were euthanized at week 6 of treatment, and blood was removed by cardiac puncture. Calciotropic hormones were measured using specific kits for PTH (Immunotopics, San Clemente, CA, USA) and 25(OH)D (Immunodiagnostic Systems Ltd., Boldon, UK). Osteocalcin (OCN) was measured in 20 µL of serum using the mouse OCN immunoradiometric assay kit (Immunotopics). TRAP was measured in 20 µL of serum to assess osteoclastic activity using the Mouse TRAP assay kit (Immunodiagnostic Systems Ltd., Scottsdale, AZ, USA).

Western blot analysis

Marrow cells were obtained from the left femora from BADGE, BADGE + 1,25(OH)3, and vehicle-treated animals (n = 8 per group) by flushing with DMEM. Red blood cells in marrow were hemolyzed in 0.017 M Tris-HCl, pH 7.5, buffer containing 0.8% ammonium chloride. Hemolyzed bone marrow suspensions were rinsed twice with PBS. Protein extracts were obtained after suspending the cells in two volumes of buffer containing 10 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, and one Complete protease inhibitor mixture tablet (Boehringer Mannheim, Laval, Canada). Solutions were then centrifuged at 25,000g for 20 minutes at 4°C; resuspended in 20 mM HEPES, pH 7.9, 25% glycerol, 0.42 M NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, and 0.5 mM dithiothreitol. Protein content was determined with a protein assay kit (Bio-Rad, Mississauga, Canada), and samples were then aliquoted and stored at –80°C. For Western blot analysis, lysates were dissolved in SDS electrophoresis buffer (Bio-Rad, Hercules, CA, USA) and proteins separated on SDS-polyacrylamide gels and subsequently electrotransfered to polyvinylidene difluoride membranes. After blocking with PBS containing 0.1% Tween 20 and 10% nonfat dry milk, membranes were incubated overnight at 4°C using mouse monoclonal antibodies directed against Runx2, OCN, osteopontin (OPN), PPARγ, CCAAT/enhancer-binding protein alpha (CEBPα), and sterol regulatory element-binding protein-1 (SREBP-1) (1:1000, Santa Cruz Biotechnology). Secondary antibodies conjugated to horseradish peroxidase were from Sigma (1:5000). Antigen-antibody complexes were detected by chemiluminescence using a kit of reagents from ECL (Amersham, Buckinghamshire, UK), and blots were exposed to high-performance chemiluminescence film (Amersham). Films were scanned and the optical density of each specific band analyzed using the ImageMaster program and expressed as OD/mm2/100 µg of total protein. Relative intensity of the samples was determined comparing the protein of interest in the treated mice using the values of vehicle-treated mice as controls (100%). Values are reported as the average of samples obtained from eight mice.

Ex vivo cultures of bone marrow cells

Adherent cells were isolated from whole bone marrow by adherence to plastic and induced to differentiate into osteoblast as previously described.13 Briefly, both tibias from BADGE, BADGE + 1,25(OH)3, and vehicle-treated animals (n = 8 per group) were flushed using a 21-gauge needle attached to a 10-mL syringe filled with DMEM (GIBCO BRL, Gaithersburg, MD, USA). Cells from both tibias were filtered through a cell strainer with 70-micron nylon mesh (BD Bioscience, Bedford, MA, USA) and then combined to produce a volume of 2 mL containing ∼107 cells/mL. Six-well plate cultures were then established in triplicate, with each well containing a 100-µL aliquot of cell suspension combined with 4 mL of fresh DMEM medium. The cells were incubated in MSC growth media at 37°C with 5% humidified CO2 and isolated by their adherence to tissue culture plastic. Medium was aspirated and replaced with fresh medium to remove nonadherent cells every 2 to 3 days. The adherent MSC were grown to ∼80% confluence for about 7 days defined as MSC at passage 0, harvested with 0.25% trypsin and 1 mM EDTA for 5 minutes at 37°C, diluted 1:3 in MSC growth media, plated, and grown to confluence for further expansion. After second and third passages, MSC were used for subsequent experiments.

To induce differentiation, a total of 104 cells were diluted in osteogenic medium (prepared with DMEM, 10% FCS, 0.2 mM dexamethasone, 10 mmol/L β glycerol phosphate, and 50 µg/mL ascorbic acid) and plated in six-well plates. Media was aspirated and replaced with fresh osteogenic medium every 3 days. After 14 days in culture, cells were washed with PBS, ethanol fixed, stained for alkaline phosphatase, and counterstained with haematoxylin (Sigma). The colonies with more than 10% of cells staining positive for ALP were considered as colony-forming units–osteoblasts (CFU-OB).

Runx2 and PPARγ activity measurement

Active Runx2 and PPARγ binding to DNA was determined using the ELISA-based Runx2 (cat. no. 47396) and PPARγ (cat. no. 40196 activation Trans-AM kit (Active Motif, Rixensart, Belgium) as previously described.16 The Trans-AM Runx2 kit contains a 96-well plate on which an oligonucleotide containing either a Runx2 or a PPARγ consensus-binding site has been immobilized. The active form of either Runx2 or PPARγ contained in nuclear extract specifically binds to this oligonucleotide. The primary antibody used in the Trans-AM kit recognizes an accessible epitope on either Runx2 or PPARγ protein upon DNA binding. Addition of a secondary horseradish peroxidase (HRP)-conjugated antibody provides a sensitive colorimetric readout easily quantified by spectrophotometry (450 nm). To quantify active Runx2 or PPARγ binding, 15 to 20 mg of nuclear extract was measured using the corresponding Trans-AM kit according to the manufacturer's instructions (Active Motif, Carlsbad, CA, USA).

Quantification of marrow fat

For marrow fat analysis, the right femur was cleaned of soft tissue, fixed for 36 hours in 4% paraformaldehyde, rinsed thoroughly in PBS, decalcified in 10% EDTA, and processed for paraffin embedding. Serial 4-µm sections were cut on a modified Leica RM 2155 rotary microtome (Leica Microsystems, Richmond Hill, Canada). Percentage of marrow fat content was calculated as fat volume/total volume (FV/TV) using haematoxylin/eosin-stained sections as previously described.16, 17

Statistical methods

All results were expressed as the mean ± standard error (SE). Differences of the structural and static parameters of bone histomorphometry between different groups of mice were determined using Levene's test for homogeneity of variances and the unpaired t test for equality of means. In all experiments, a value of p < 0.05 was considered significant.


Treatment of adult male C57Bl6 mice with PPARγ inhibitors increases bone mass and stimulates bone formation

Nine-month-old C57BL/6 male mice were treated for 6 weeks with a daily ip injection of 30 mg/kg BADGE in 15% DMSO alone or in combination with a previously validated12, 17 anabolic dose of 1,25(OH)2D3 administered by subcutaneous pump at a dose of 18 pm/d, using vehicle-treated mice as control. Faxitron X-ray (Fig. 1A), quantitative µCT (Fig. 1B, C), and undecalcified histology (Fig. 1D, E) showed evidence of increased cortical and trabecular bone formation in mice treated with the PPARγ inhibitors BADGE alone or in combination with 1,25(OH)2D3. Quantitative µCT data (Fig. 1C) indicated a significant increase in bone relative to tissue volume (BV/TV) in the region of interest, which was reflected in both increased numbers (Tb.N) and thickness (Tb.Th) of trabeculae. Quantitative histomorphometry performed on tetracycline-labeled sections (Fig. 1D, E) revealed increased mineral apposition rates (MAR) in both trabecular and cortical bone in the mice receiving PPARγ inhibitors compared with vehicle-treated control. Furthermore, an anomalous increase in osteoid was seen in the mice that received BADGE and 1,25(OH)2D3 (Fig. 1D, red arrows) in the absence of vitamin D deficiency or hyperparathyroidism (Table 1).

Figure 1.

Treatment of adult male C57Bl6 mice with PPARγ inhibitors increases bone mass and stimulates bone formation. Nine-month-old C57BL/6 male mice (n = 8 per group) were treated for 6 weeks with BADGE, BADGE + 1,25(OH)3, or vehicle as described in Materials and Methods. (A) Faxitron analysis demonstrated an increased bone density in the femurs of BADGE-treated mice as compared with their vehicle-treated counterparts. (B, C) Three-dimensional µCT analysis showed a significant positive anabolic effect of BADGE, alone or in combination with 1,25(OH)3, on bone structure including higher trabecular bone volume relative to tissue volume (BV/TV), trabecular number (Tb.N), and trabecular thickness (Tb.Th). (D, E) Dynamic histomorphometry showed higher levels of mineral apposition rate (MAR) both in trabecular (lower left panels) and cortical bones (lower right panels) of BADGE- and BADGE + 1,25(OH)3-treated mice. In addition, treatment with BADGE + 1,25(OH)3 induced a significantly higher area of unmineralized osteoid (D, red arrows, and E) as compared with BADGE- and vehicle-treated mice. Each bar represents the mean ± SE. *p < 0.01.

Table 1. Circulating Concentrations of Glucose, Calcium, Calciotropic Hormones PTH and 25(OH)D and Markers of Bone Formation (Osteocalcin) and Bone Resorption (TRAP) in 9-Month-Old C57BL/6 Mice Treated With BADGE, BADGE + 1,25(OH)2D3, or Vehicle
AssayBADGE (n = 8)BADGE + 1,25 (OH)2D3 (n = 8)Vehicle (n = 8)p Value
  • a

    BADGE versus BADGE + 1,25(OH)2D3-treated mice.

  • b

    Vehicle versus BADGE + 1,25(OH)2D3-treated mice.

Glucose (mg/dL)106 ± 12102 ± 1496 ± 15NS
Calcium (mg/dL)8.2 ± 0.28.5 ± 0.38 ± 0.3NS
25(OH)D (nmol/L)52 ± 1460 ± 1554 ± 12NS
PTH (pg/mL)36 ± 616 ± 546 ± 7< 0.01a,b
Osteocalcin (ng/mL)140 ± 10170 ± 2090 ± 20< 0.01a,b
TRAP (U/L)4.5 ± 16 ± 25 ± 1NS

Treatment with PPARγ inhibitors increased osteoblast number and activity in the absence of alterations in osteoclasts

The anabolic effect of PPARγ inhibitors on bone was further examined by quantifying the number of bone-forming cells (osteoblasts) and bone-resorbing cells (osteoclasts) in situ and measuring circulating concentrations of markers of their activity. After normalization to the bone surface, a significant increase in osteoblast numbers was seen in both BADGE and BADGE + 1,25(OH)2D3–treated mice (Fig. 2A, B) compared with those treated with vehicle alone (p < 0.01). The increase in osteoblast number correlated with an increase in serum OCN (Table 1), which is used as a biomarker for bone formation in the clinical setting. OCN was also significantly higher in BADGE + 1,25(OH)2D3–treated mice compared with those treated with BADGE alone (p < 0.01).

Figure 2.

Treatment with PPARγ inhibitors increased osteoblast number and activity in the absence of alterations in osteoclasts. Sections of plastic embedded tibias from BADGE-, BADGE + 1,25(OH)3-, and vehicle-treated mice were stained sequentially for alkaline phosphatase (ALP) for osteoblasts (A). A significant increase in the number of ALP-expressing cells (osteoblasts) (A, B) was observed in mice treated with BADGE and BADGE + 1,25(OH)3 compared with the vehicle-treated group. In contrast, staining of bone with tartrate-resistant acid phosphatase (TRAP) (C) showed no difference in the number of cells stained in red corresponding to osteoclasts (OC) (arrows) in either BADGE-, BADGE + 1,25(OH)3-, or vehicle-treated mice (D). Micrographs are representative of those from eight different mice of each treatment group. Arrows indicate representative cells/areas in each section. Magnification x20. Each bar represents the mean ± SE. *p < 0.01.

In striking contrast to the effect of PPARγ inhibition on osteoblasts, there was no change in osteoclast number, as evidenced by tartrate resistant acid phosphatase (TRAP) staining in situ in bone (Fig. 2C, D) or in serum TRAP (Table 1). These data suggested an uncoupling of bone formation from resorption by favoring osteoblastic over osteoclastic number and activity.

Treatment with PPARγ inhibitors in vivo results in increased osteoblastogenesis in bone marrow stromal cells (BMSC) cultured ex vivo

To determine if the increased number of osteoblasts was related to differentiation of BMSC down the osteoblast lineage, stromal cells isolated from bone marrow of mice were treated for 6 weeks with vehicle, BADGE, or BADGE + 1,25(OH)2D3. BMSC were maintained for 2 weeks in osteogenic medium before quantifying the number of alkaline phosphatase (ALP)-positive colonies (Fig. 3A, B). The number of colony-forming units–osteoblasts (CFU-OB), an indicator of the capacity of BMSC to differentiate into osteoblasts, was significantly higher in cultures harvested from BADGE and BADGE + 1,25(OH)2D3–treated mice compared with those from vehicle-treated mice (Fig. 3B, p < 0.01). Immunoblot analysis (Fig. 3C, D) of protein extracted from bone marrow revealed a significant increase in expression of stromal cells isolated from OCN and the osteoblast transcription factor Runx2, in association with a reduction in osteopontin (OPN), which is thought to play a role in osteoclast adhesion to bone, in bone marrow from BADGE, and BADGE + 1,25(OH)2D3–treated mice. Furthermore, using a highly sensitive ELISA assay, it was shown that the increase in Runx2 expression led to a concomitant increase in Runx2 complex formation and DNA binding (Fig. 3E).

Figure 3.

Treatment with PPARγ inhibitors in vivo results in increased osteoblastogenesis in bone marrow stromal cells (BMSC) cultured ex vivo. (A, B) Formation of colony-forming units–osteoblasts (CFU-OB) in ex vivo cultures of bone marrow cells from 9-month-old C57BL/6 male mice treated for 6 weeks with either BADGE, BADGE + 1,25(OH)3, or vehicle alone. The number of CFU-OB per femur was significantly higher after 2 weeks of differentiation in both BADGE- and BADGE + 1,25(OH)3- treated mice as compared with mice treated with vehicle alone. *p < 0.01. (C, D) Treatment with BADGE alone or in combination with 1,25(OH)3 induced significant levels of protein expression for the osteogenic transcription factors osteocalcin (OCN) and Runx2. In contrast, levels of expression of osteopontin (OPN) were decreased after treatment with BADGE and BADGE + 1,25(OH)3. Data from scanning densitometric analyses is expressed as the ratio of the protein of interest in the BADGE and BADGE + 1,25(OH)3 mice using the values of vehicle-treated mice as controls representing the mean ± SD of triplicate determinations. *p < 0.01. (E) Runx2 DNA-binding activity was determined using ELISA-based Runx2 activation kit and quantified by colorimetry. The figure shows the levels of activity after treatment with either BADGE and BADGE + 1,25(OH)3 or vehicle alone. Treatment with BADGE alone or in combination with 1,25(OH)3 significantly increased the activity of the Runx2 complex in the nuclei. Values are mean ± SEM of protein extracts obtained from marrow fat of 8 mice per group. *p < 0.01.

Treatment with PPARγ inhibitors reduces marrow fat by a PPARγ-dependent mechanism

Enhanced osteoblastogenesis is commonly associated with inhibition of adipogenesis.6 As shown in Fig. 4A, B, marrow fat was significantly reduced in mice treated with BADGE, with an apparent additive effect in the BADGE + 1,25(OH)2D3–treated mice compared with vehicle-treated controls. In both cases, there was a reduction in PPARγ protein expression as well as that of the adipogenic transcription factor CEBPα but not SREBP-1 (Fig. 4C, D). In contrast to the increased osteogenic Runx2 complex formation and DNA-binding activity (Fig. 3E) that of the adipogenic PPARγ was reduced (Fig. 4E).

Figure 4.

Treatment with PPARγ inhibitors reduces marrow fat by a PPARγ-dependent mechanism. (A, B) Sections of decalcified bone were stained with H/E and images captured at original magnifications of x40 to evaluate fat infiltration (empty areas, arrows). Both BADGE- and BADGE + 1,25(OH)3-treated groups showed significantly lower volume of marrow fat (A, B) compared with their vehicle-treated controls. Micrographs are representative of four to six screened in each group of animals. Average of marrow fat content (fat volume/total volume) was quantified in 10 fields per section. *p < 0.01. (C, D) Treatment with BADGE alone or in combination with 1,25(OH)3 significantly decreased the levels of protein expression for the adipogenic transcription factors PPARγ and CEBPα. In contrast, levels of expression of SREBP-1 were not affected by treatment with BADGE, BADGE + 1,25(OH)3, or vehicle alone. Data from scanning densitometric analyses is expressed as the ratio of the protein of interest in the BADGE and BADGE + 1,25(OH)3 mice using the values of vehicle-treated mice as controls representing the mean ± SD of triplicate determinations. *p < 0.01. (E) PPARγ DNA-binding activity was determined using ELISA-based PPARγ activation kit and quantified by colorimetry. The figure shows the levels of activity after treatment with either BADGE and BADGE + 1,25(OH)3 or vehicle alone. Treatment with BADGE alone or in combination with 1,25(OH)3 significantly decreased the activity of the PPARγ complex in the nuclei. Values are mean ± SEM of protein extracts obtained from marrow fat of 8 mice per group. *p < 0.01.

Treatment with PPARγ inhibitors increases VDR expression within the bone marrow

In an attempt to investigate the mechanism of action of the PPARγ inhibitor BADGE in bone, we quantified changes in both VDR and PPARγ within the same set of bone sections. We then calculated the VDR/PPARγ ratio. A positive ratio would indicate that VDR expression is facilitated by BADGE treatment. As shown in Fig. 5, PPARγ inhibition was associated with an increase in VDR/PPARγ ratio. This increase was significantly higher in mice treated with BADGE + 1,25(OH)2D3.

Figure 5.

Treatment with PPARγ inhibitor increases vitamin D receptor (VDR) expression within the bone marrow. (A) After decalcification, bone samples were embedded in low-melting paraffin and coronal sections were made for the epiphyseal parts and the shaft, respectively. Sections were treated as described in Materials and Methods. Sections were simultaneously incubated with PPARγ and VDR antibodies (red and green fluorescence, respectively). (B) Each section was analyzed using Image-J image analysis. Treatment with BADGE alone or in combination with 1,25(OH)3 significantly increased the VDR/PPARγ ratio, indicating a predominance of VDR expression in the bone marrow of the treated groups. Values are mean ± SEM of 8 mice per group. *p < 0.01.


In this study, we have characterized the bone phenotype of skeletally mature male mice after pharmacological inhibition of PPARγ. Our results indicate that PPARγ inhibition increased bone formation without affecting bone resorption and that the effect was in part owing to enhanced differentiation of bone marrow precursor cells into mature osteoblasts. The increase in bone mass was accompanied by a significant reduction in marrow fat infiltration and associated with decreased transcription of adipogenic and increased transcription of osteogenic genes.

Elucidation of the fundamental mechanisms that regulate bone biology during the last two decades has allowed the development of several effective treatments for osteoporosis. The majority of these new treatments have been directed at inhibition of bone resorption, whereas anabolic compounds in use or under investigation remain scarce.1, 3 An effective bone anabolic agent would induce bone formation by stimulating osteoblast activity, by protecting osteoblasts against premature death, and by promoting the differentiation of MSC down the osteoblast rather than the adipocyte lineage.1 In fact, several studies have demonstrated that it is possible to stimulate osteoblast differentiation at the expense of adipogenesis using compounds such as parathyroid hormone,18 interferon gamma,19, 20 and strontium ranelate.21 The active metabolite of vitamin D, 1,25(OH)2D3, has also been shown to stimulate bone formation by promoting osteoblastogenesis12 and inhibiting adipogenesis17 in a murine model of age-related bone loss. A common finding in these studies is the concomitant upregulation of Runx2 and downregulation of PPARγ, supporting the hypothesis that “fat loss is bone gain.”22 Overall, although all these compounds have demonstrated an inhibitory effect on PPARγ thus stimulating osteoblastogenesis, it was tempting to hypothesize that a direct pharmacological inhibitor of PPARγ would have a stronger stimulatory effect on bone formation while inhibiting marrow fat infiltration. PPARγ has been identified as the pivotal transcription factor involved in the differentiation of bone marrow MSC into adipocytes.7 Akune and colleagues9 showed that mice heterozygous for PPARγ deficiency had low marrow fat and high bone mass. The effect was shown to be independent of estrogens because no changes in bone mass were observed after oophorectomy. The haploinsufficient mice expressed high levels of the osteogenic factors OCN, OPN, and Runx2 and low levels of the adipogenic factor CEBPα. These changes in transcription were associated with increased osteoblast differentiation in ex vivo cultures of bone marrow–derived MSC. This work validated the hypothesis that targeted inactivation of the adipogenic transcription factor PPARγ resulted in increased bone formation in vivo.

In contrast to the results of Akune and colleagues9 using genetic modification of PPARγ, pharmacological inhibition of PPARγ with BADGE by Botolin and colleagues11 in diabetic mice resulted in a significant decrease in marrow fat in the absence of any effect on bone. Concomitantly, in vitro evidence from Yu and colleagues23 showed reduced adipocyte differentiation in cultures of MSC treated with BADGE, although these investigators failed to provide evidence of altered osteoblastogenesis. Both the in vitro and in vivo data indicated that BADGE is an effective inhibitor of adipocyte differentiation and adipogenesis. However, its effect on promoting osteoblast differentiation and bone formation remained undefined.

In the current study, BADGE alone or combined with1,25(OH)2D3 was used as a pharmacological inhibitor of PPARγ in normal adult mice to determine its effect on osteoblastogenesis and bone formation. Male C57BL/6 mice were selected as an appropriate model to study endogenous, age-related osteopenia and senile osteoporosis,24 in which marrow fat infiltration is considered to be the primary pathophysiological mechanism. Our comprehensive bone phenotyping protocol revealed a significant anabolic effect of pharmacological inhibition of PPARγ on bone. Mice treated with BADGE had higher bone mass, increased trabecular number and thickness, and higher levels of mineral apposition both in trabecular and cortical bone. These effects were associated with increased osteoblastogenesis and elevated expression and activity of osteogenic transcription factors in the BADGE-treated groups. Taken together, these data indicate that the anabolic effect of inhibition of PPARγ on bone is explained by the induction of high levels of Runx2 and OCN expression, and by a significant increase in the nuclear binding capacity of the Runx2 complex.

Interestingly, when use in conjunction with 1,25(OH)2D3, wide seams of osteoid lined the surfaces of trabecular bone in mice treated with BADGE. A similar increase in osteoid was seen in mice homozygous for targeted disruption of the fibroblast growth factor receptor 3 (FGFR3) gene.25 In both cases, the absence of significant changes in circulating calcium and vitamin D suggested the increase in bone formation was greater than the capacity for matrix mineralization. In the current study, the accumulation of osteoid could have been the result of reduced transcription of OPN, which is an important regulator of bone mineralization. The relationship between PPARγ and OPN expression and activity during bone matrix mineralization warrants further investigation.

Another notable observation in our study was that PPARγ inhibition induced bone formation without affecting osteoclast number or activity. This supported Akune's observations and indicated that inhibition of PPARγ leads to uncoupling of osteoblast and osteoclast function in bone turnover to the end of favoring bone formation. Furthermore, our study supports others10, 11 in demonstrating reduced adipogenesis and marrow fat concomitant with reduced expression of adipogenic factors after treatment with BADGE. Despite the reduction in marrow adiposity, there were no significant systemic changes in blood glucose or body weight, which supported previous work11 but also suggested that at the dose used in these studies BADGE is a safe and effective therapeutic intervention to promote bone formation in osteopenic individuals without inducing any metabolic disorders. A potential pitfall for the use of BADGE in humans is that its precursors, which are used commercially in the manufacture of polycarbonate plastics, might have some unknown toxicity.

From the mechanistic point of view, we also assessed whether pharmacological inhibition of PPARγ would facilitate the expression of VDR in bone. Considering that both PPARγ and VDR form heterodimers with the same retinoid X receptor (RXR), it would be expected that inhibition of PPARγ would be followed by an increase in VDR expression, especially after treatment with 1,25(OH)2D3. The high levels of VDR expression observed in BADGE- and BADGE + 1,25(OH)2D3–treated mice allows us to hypothesize that inhibition of PPARγ facilitates the osteogenic response of bone to vitamin D.12 Further studies looking at changes in the interaction between RXR and VDR during PPARγ inhibition in vivo are still required.

Taken together, the data support the proposal that pharmacological inhibition of PPARγ could represent a new anabolic approach to the treatment of osteoporosis in the near future. Although this treatment should be tested in other classical models of osteoporosis such as the OVX mice and rats, ours is the first evidence demonstrating that PPARγ inhibition is feasible without affecting systemic glucose or calcium metabolism in mice and without any evident side effects. Because the combination of BADGE and 1,25(OH)2D3 induced higher levels of unmineralized osteoid, more studies are still required to test the effect of BADGE in combination with the nonactive forms of vitamin D.


All authors state that they have no conflicts of interest.


This study was supported by grants from the Australian National Health and Medical Research Council (NHMRC 632767) and the Nepean Medical Research Foundation. The authors thank Mrs. Miren Gratton for her collaboration in the biochemical analysis.

Authors' roles: Study design: GD. Study conduct: GD and JH. Data collection: GD, WL, CV, SB, and DR. Data analysis: GD, CV, and JH. Data interpretation: GD, CV, and JH. Drafting manuscript: GD, CV SB, and JH. Revising manuscript content: GD, CV, SB, and JH. Approving final version of manuscript: GD, WL, CV, SB, DR, and JH. GD takes responsibility for the integrity of the data analysis.