The effects of PPARγ inhibitor on bones and bone marrow fat in aged glucocorticoid-treated female rats

Progressive bone marrow (BM) fat accumulation is a common bone loss characteristic in older populations and glucocorticoid (GC)-induced skeletal destruction that is inversely associated with bone synthesis and directly associated with increased peroxisomal proliferator-activated receptor gamma (PPARγ) expression. PPARγ inhibition is an efficient therapeutic strategy for aged- and GC-related skeletal disorders. This study aimed to evaluate the effect of PPARγ inhibition on aged GC-treated female rats. It was hypothesised that bisphenol A diglycidyl ether (BADGE) could inhibit marrow adiposity and improve osteogenesis by inhibiting PPARγ, thereby preventing GC-induced osteoporosis (GIO). Female Sprague-Dawley rats (n = 32, age = 18 months) were randomly allocated to one of the following groups: (1) control, (2) BADGE (30 mg/kg/day, intraperitoneal), (3) methylprednisolone (MP; 30 mg/kg/day, subcutaneous), and (4) MP + BADGE. After eight weeks of treatment, bone density (BD) and trabecular bone microarchitectures were quantified by micro-computed tomography (CT), and BM adipocytes were quantified by histopathology. Additionally, mRNA and protein expression of adipogenic and osteogenic markers were quantified by reverse transcription-quantitative polymerase chain reaction. Furthermore, serum bone turnover biomarker levels were quantified by enzyme-linked immunosorbent assay. MP treatment led to marrow adipogenesis and bone deterioration. However, rats treated with MP + BADGE showed lower marrow adipogenesis, as indicated by smaller marrow adipocyte diameter, decreased density and area percentages, reduced expression of marrow adipogenic genes and proteins, improved BD and trabecular microarchitectures, increased expression of osteogenic genes and proteins, and higher levels of serum bone formation markers. These results were consistent with the differences observed between control and BADGE mono-treated rats. In conclusion, BADGE treatment attenuates BM adiposity and improves bone formation in aged GC-treated female rats by inhibiting PPARγ. Therefore, PPARγ might be a potential target for treating GIO in older populations.


Introduction
Glucocorticoid (GC)-induced osteoporosis (GIO) is the most frequent secondary osteoporosis and is characterised by fractures, the most common adverse event associated with GC treatment.The incidence of bone fractures associated with GIO is increasing in the ageing population globally (Xi et al., 2020;Zhang et al., 2020;Diez-Perez et al., 2011), possibly because GCs are prescribed by many clinicians for various inflammatory disorders.Most older adults are not adequately and timely diagnosed for their increased fracture risk, and only some are offered treatment (Adler, 2019).In addition, older women receiving GC therapy are more prone to osteoporotic fractures than older men due to the suppressive effect of GCs on their already decreased estrogen levels (Cheng et al., 2022;Paggiosi et al., 2015;Weinstein, 2011).The GIOassociated pain and bone fracture significantly affect older adults' quality of life.The GIO mechanism is primarily associated with an early and transient increase in osteoclast-induced bone resorption and a prolonged decrease in osteoblast-induced bone formation (Weinstein, 2011).Recent clinical data from rodent studies have highlighted that osteoblast lineage cells mainly affect GIO and the GC-induced increase in fracture risk (Hardy et al., 2018;Gado et al., 2022).Therefore, based on acquired bone formation data, new preventive and treatment strategies for GIO in high-risk populations are urgently required.
Bone formation can be increased by stimulating the differentiation of bone marrow (BM) mesenchymal stem cells (BMSCs).Progressive BM fat accumulation is a constant characteristic of age-associated bone loss and osteoporosis in aged populations caused by a change in the BMSC's differentiation mechanism from osteogenic in young bone into adipogenic in old bone (Hu et al., 2018;Chandra et al., 2022;Schwartz et al., 2013).BMSCs are the precursors of adipocytes and osteoblasts, and an inverse relationship exists between BMSC differentiation into osteoblasts and adipocytes.Much evidence suggests a reciprocal relationship between bone integrity and marrow adiposity in numerous physiological and pathological conditions, such as advanced age and GC treatment (Sadie- Van et al., 2013).Supraphysiological levels of GC-induced BMSCs divert from the osteoblast to the adipocyte lineage, ultimately decreasing the osteoblast progenitor pool and limiting bone synthesis (Martel et al., 2019;Han et al., 2019;Li et al., 2013;Sui et al., 2016;Bensreti et al., 2023).Therefore, targeting marrow adipocytes may be an alternative therapeutic strategy for age-and GC-related bone loss.
Bisphenol A diglycidyl ether (BADGE) is a synthetic PPARγ antagonist that inhibits marrow adipocyte formation ex vivo and in vivo (Yuan et al., 2018;Wang et al., 2019;Duque et al., 2013;Marciano et al., 2015;Beekman et al., 2019a).Duque et al. pharmacologically suppressed PPARγ using BADGE in an age-related bone loss model, showing a substantial decrease in marrow fat content concomitant with enhanced bone formation and osteoblastogenesis (Duque et al., 2013).Moreover, a recent study showed that a BADGE regimen improved bone synthesis and reduced marrow adiposity in mice with GIO (Wang et al., 2019).However, these studies used young animal models, and whether BADGE has the same effect in aged animals remains unknown.This study aimed to evaluate the preventive efficiency of BADGE in GC-treated aged female rats.It was hypothesised that BADGE could inhibit marrow adiposity and improve osteogenesis by inhibiting PPARγ, thereby inhibiting GIO.

Experimental design
The Ethics Committee of Tianjin Medical University General Hospital approved this study (approval no.: IRB-2023-DW-108).Female Sprague-Dawley rats (n = 32, age = 18 months, and weight = 491-527 g) were housed in a controlled environment at 22 • C ± 2 • C, 45 %-60 % relative humidity, and 12/12-h light/dark cycle, with lights switched on at 8 am and off at 8 pm.Ad libitum standard rat chow and water were provided to all animals, and they were acclimatised for seven days before being randomly allocated to one of four groups (n = 8/group): (1) control (CON), (2) BADGE, (3) methylprednisolone (MP), and (4) MP + BADGE.Rats in the MP group subcutaneously received 30 mg/kg of MP (Pfizer, USA) daily.Rats in the BADGE group were intraperitoneally administered 30 mg/kg of BADGE (Sigma, St. Louis, MO, USA) dissolved in 10 % dimethyl sulfoxide (DMSO) daily.We selected 18-month-old rats for model generation since they have sufficient BM fat at relatively older ages, which is preferred for simultaneously studying bone and BM fat (Beekman et al., 2019b).Moreover, female rats were used because they are more prone to GIO due to the GC inhibitory effect on their already decreased estrogen levels (Cheng et al., 2022;Paggiosi et al., 2015).The doses were selected based on published studies, which suggested they were potent for alleviating BM adiposity in skeletally mature mice (Duque et al., 2013), ovariectomised rats (Li et al., 2016a), type 1 diabetic mice (Botolin and McCabe, 2006), and immune-induced aplastic anaemic mice (Sato et al., 2016).Rats in the MP + BADGE group received 30 mg/kg of both MP and BADGE.Rats in the CON group received the same volume of vehicle (phosphate-buffered saline with % DMSO [v/v]).All groups were treated for eight weeks.The rats were weighed weekly, and the administered dose was adjusted accordingly.
After treatment was completed, blood specimens were collected by cardiac puncture under anaesthesia and centrifuged for 15 min at rpm to collect serum for biochemical analyses.Furthermore, the fifth lumbar vertebrae (L5) and right femurs were dissected out and stored at − 20 • C for bone microparameter assessment, and the right tibias (for bone histomorphometry) and left femurs (for reverse-transcription polymerase chain reaction [RT-qPCR] and Western blot assays) were also dissected out.

Micro-computed tomography (micro-CT)
A Hiscan XM Micro CT (Suzhou Hiscan Information Technology Co., Ltd) was used to scan the right femur and lumbar vertebrae with 80 kV and 100 uA X-ray tube settings.The images were obtained at 25 μm resolution, 0.5 • rotation step through a 360 • angular range, and 50 ms exposure/step and were assessed using Hiscan Analyser software (version 3.0).The following parameters were assessed: trabecular separation (Tb.Sp), percentage bone volume (BV/TV), trabecular thickness (Tb.Th), bone surface density (BS/TV), trabecular bone number (Tb.N), connectivity density (Conn.Dn), and bone density (BD).

RT-qPCR
Total RNA was isolated from the left femur using TRIzol reagent (Invitrogen), according to the manufacturer's instructions.The isolated RNA was reverse transcribed to create cDNA using the RevertAid First Strand cDNA Synthesis Kit (Thermo) and examined by RT-qPCR with the SYBR Premix Ex Taq (Takara) on an ABI 7900 system.β-actin (ACTB) was used for data normalization.The data were compared using the cycle threshold (C t ) method.The C t values were computed at the beginning of the RT-qPCR's logarithmic amplification.The fold change was calculated using the 2 − ΔΔCt method, where ΔC t = C t, target gene − C t, ACTB and ΔΔC t = ΔC t, treatment − ΔC t, control (Ko et al., 2012).The sequences of the primers used are listed in Table 1.

Western blots
Briefly, the left femur was crushed in liquid nitrogen.Total protein was isolated from this crushed bone using RIPA protein extraction reagents (Solarbio, Beijing, China) according to the manufacturer's protocol.The supernatant was collected after centrifugation at 12000 rpm and 4 • C 10 min, and its protein content was quantified using the bicinchoninic acid assay.Next, the protein samples were size-separated by 10 % sodium dodecyl sulphate-polyacrylamide gel electrophoresis and transferred onto a polyvinylidene difluoride membrane.Then, the membrane was blocked with 5 % non-fat milk in Tris-buffered saline for 1 h and labelled with primary antibodies against PPARγ (ab272718; 1:1000; Abcam, Cambridge, MA, USA), RUNX2 (ab76956; 1:1000; Abcam), osteocalcin (OCN; DF12303; 1:1000; Affinity, China), and ACTB (AF7018; 1:1000; Affinity) at 4 • C overnight.Next, the membrane was incubated with the corresponding secondary antibodies.The experiments were repeated thrice.Band densities were analyzed using ImageJ software (version 1.49P), with ACTB used for normalization.

Serum biochemical analysis
Blood glucose, total cholesterol (CHO), and triacylglycerol (TG) were quantified at the clinical laboratory of the Tianjin Center for Disease Control and Prevention using a TBA-40FR Automatic Analyser (Tokyo, Japan) using corresponding commercial kits (BioSinoBio-Technology & Sciencelnc, Beijing, China).Furthermore, the levels of serum bone turnover biomarkers, including bone formation marker OCN and bone resorption marker C-terminal telopeptide of type I collagen (CTX -I; Rigorbio, Beijing, China), were assessed using an enzyme-linked immunosorbent assay (ELISA) kit according to the manufacturer's instructions.

Statistical analyses
Statistical analyses were conducted using SPSS software (version 26.0).The acquired data are presented as the mean ± standard deviation (SD).The normality of the data distribution was assessed using the Shapiro-Wilk test.Two-way repeated-measures analyses of variance (ANOVAs) followed by Bonferroni post hoc tests were used to compare the repeated weight measurements between the four groups.The other studied variables were compared using two-way ANOVAs with multiple comparison corrections with Tukey's post hoc test.Model and drug interventions were considered the analysis variables.A P < 0.05 was considered statistically significant.

Changes in body weight and serum lipid and glucose levels
Fig. 1 shows how body weights changed with treatment time within the four groups.The body weights of rats were similar in the four groups at the start of the study.However, they gradually increased with treatment time in the CON and BADGE groups.In contrast, they gradually decreased for the first four weeks before gradually increasing in the last four weeks in the MP and MP + BADGE groups.At the end of the eighth week, the body weights of rats in the MP group were 5.1 % (499 ± 4 vs 524 ± 6 g; P < 0.001) and 4.0 % (499 ± 4 vs 520 ± 3 g; P < 0.001) lower than rats in the CON and BADGE groups, respectively.Similarly, the body weights of rats in the MP + BADGE group were 6.4 % (494 ± 7 vs 524 ± 6 g; P < 0.001) and 5.1 % (494 ± 7 vs 520 ± 3 g; P < 0.001) lower than rats in the CON and BADGE groups, respectively.Body weights did not differ significantly between rats in the MP + BADGE and MP (494 ± 7 vs 499 ± 4 g; P = 0.064) and BADGE and CON (520 ± 3 vs 524 ± 6 g; P = 0.065) groups.No rat died, and no severe adverse reactions were identified during the experiment.
Table 2 shows the blood glucose, TG, and CHO concentrations in the CON, BADGE, MP, and MP + BADGE groups at the end of the eighth week.Blood glucose was significantly higher in the MP than in the CON group (6.00 ± 0.50 vs 5.10 ± 0.69 mmol/L; P = 0.032) but did not differ significantly between the MP and MP + BADGE groups (6.00 ± 0.50 vs 6.11 ± 0.47 mmol/L; P = 0.983) or between the BADGE and CON groups (5.36 ± 0.73 vs 5.10 ± 0.69 mmol/L; P = 0.841).Interestingly, TG was significantly higher in the MP than in the CON group (2.65 ± 0.38 vs 1.56 ± 0.27 mmol/L; P < 0.001) but significantly lower in the MP + BADGE than in the MP group (1.96 ± 0.35 vs 2.65 ± 0.38 mmol/ L; P < 0.001).CHO did not differ significantly among groups.

Effects of BADGE on adipogenic and osteogenic marker expression
The MP treatment increased adipogenic marker (PPARγ and adipocyte fatty acid binding protein [aP2]) levels but decreased osteogenic marker (RUNX2 and OCN) levels more than the control treatment (all P < 0.001).Furthermore, the mRNA and protein levels were significantly higher for RUNX2 and OCN but lower for aP2 in the MP + BADGE group than in the MP group (all P < 0.001).Additionally, mRNA and protein levels were significantly lower for adipogenic markers but higher for enhanced osteogenic markers in the BADGE group than in the CON group (all P < 0.001, except Runx2 mRNA [P = 0.021]; Fig. 4).

Discussion
It has been suggested that GC treatment reduces bone mass and promotes osteoporosis.Six months of oral prednisolone administration reduced lumbar bone mass by 5-15 % in adults, increasing fracture risk (Amiche et al., 2016).However, studies treating rats with GC in vivo have provided contradictory data, with some reporting increased BD while others reported deceased BD (Ogoshi et al., 2008;Govindarajan et al., 2013).These contradictions might reflect the different ages of the animal models used.Our study first assessed the influence of GC on bone in normal-aged female rats, finding substantially reduced BD and trabecular bone microarchitectures in aged MP-treated rats than in aged vehicle-treated rats.This finding is consistent with a study on ovariectomized rabbits that found GC treatment after ovariectomy had a more pronounced effect on bone than ovariectomy alone (Govindarajan et al., 2013).
Previous ex vivo studies have shown that exogenous GC exposure in rodents and humans is associated with enhanced BM adiposity (Martel et al., 2019;Han et al., 2019;Li et al., 2013;Li et al., 2016b).Furthermore, gene expression analysis showed that GC upregulated adipogenesis-related genes but downregulated osteogenic genes in bone tissue (Yuan et al., 2018;Wang et al., 2019;Yao et al., 2008a).These results are consistent with those of our study, which is the first to report increased BM fat content (adipocyte size, density, and percentage area) and adipocyte transcription factor (PPARγ) expression but decreased osteogenic transcription factor (RUNX2) expression in the bone of aged MP-treated female rats.
As a synthetic PPARγ antagonist, BADGE suppresses adipogenesis and reduces BM fat.Yu et al. showed decreased in vitro adipocyte differentiation in BMSCs cultured with BADGE (Yao et al., 2008b).Similarly, Duque et al. reported that BADGE treatment significantly reduced BM adiposity in male C57BL/6 mice without affecting glucose metabolism (Duque et al., 2013).Furthermore, Li et al. found that BADGE could recover ovariectomy-mediated BM adiposity levels to normal in rats (Li et al., 2016a), consistent with our data.Our findings indicate that BADGE attenuated BM adiposity in aged vehicle-treated female rats.Furthermore, our study assessed the effect of BADGE on adipogenesis in aged GC-treated female rats, finding that BADGE treatment significantly reduced BM adiposity in MP-treated rats.Additionally, the levels of adipogenic factor aP2 were significantly decreased.

CON BADGE MP MP+BADGE
While BADGE has been shown to successfully reverse BM adipogenesis, its effects on skeletal integrity remain unknown.For example, Li et al. found that BADGE does not influence bone turnover during estrogen deficiency, although it decreased BM adiposity (Li et al., 2016a).Similarly, Botolin and McCabe found that BADGE reversed the effect of type 1 diabetes on BM adiposity but did not influence bone deterioration (Botolin and McCabe, 2006).However, Akune et al. showed that mice with heterozygous PPARγ-deficiency had reduced BM adiposity but increased bone mass, osteoblast numbers, and bone synthesis rates (Akune et al., 2004).Akune et al. and Duque et al. reported that pharmacological PPARγ inhibition with BADGE increased bone mass, osteoblastogenesis, and bone formation but decreased BM adiposity in male C57BL/6 mice (Duque et al., 2013).Recently, Wang et al. reported that a BADGE regimen improved bone synthesis and reduced BM adiposity in young mice with GIO (Wang et al., 2019).As mentioned above, current data on the effects of BADGE on bone are inconsistent.These discrepancies could reflect the different models used.For example, Li et al. used an ovariectomy model in which the effect of reduced estrogen on enhanced bone resorption and impaired osteoblast function could have exceeded the potential beneficial effects of BADGE on osteoblast differentiation and function.Moreover, Botolin and McCabe used a diabetic mouse model in which the inhibitory effect of diabetes type I on bone formation could have exceeded the potential effect of BADGE on osteoblast differentiation and function.Our study assessed the effects of BADGE on bone mass and osteogenic transcription factor expression in aged mice, finding increased bone mass after BADGE administration in both aged vehicle-and MP-treated female rats.Its effects were associated with enhanced levels of osteogenic transcription factors (OCN and RUNX2) in the BADGE-treated rats.
Furthermore, the effects of BADGE treatment on osteoblast formation and osteoclasts' absorption activities were assessed by measuring serum OCN and CTX-I levels.MP treatment was found to decrease serum OCN levels and increase serum CTX-I levels.BADGE treatment increased serum OCN levels but did not affect serum CTX-I levels compared to MP treatment.A similar pattern was observed between BADGE and control treatments.These data were consistent with previous studies indicating that PPARγ inhibitor treatment enhances bone synthesis without affecting its resorption (Li et al., 2016a).
Overall, our data support the hypothesis that PPARγ inhibition attenuates BM adiposity and improves bone formation in GC-treated aged female rats.Furthermore, the mechanisms underlying GC-induced bone deterioration involved PPARγ activation, which mediates BMSCs to differentiate into adipocytes and inhibits their differentiation into osteocytes.Therefore, targeting PPARγ is a possible strategy to alleviate GIO and BM fat.As a PPARγ inhibitor, BADGE could be used therapeutically in older adults.

Limitations
The main strength of our study is that it evaluated the influence of BADGE on bone formation and adipogenesis in GC-treated aged female rats, mimicking older women receiving GC therapy in a clinical setting.Nevertheless, it also had several limitations.First, the BADGE dose was based on published studies, and only a single dose was assessed.However, a high BADGE dose (60 mg/kg) has been reported to be cytotoxic (Sato et al., 2016), and lower doses, such as that used in our and other studies, significantly reduced BM adiposity without any significant side effects on energy metabolism.Therefore, our study used a low dose (30 mg/kg) calculated according to previous in vitro and in vivo studies (Wang et al., 2019;Duque et al., 2013;Marciano et al., 2015;Li et al., 2016a).Second, while BADGE has a preventive effect on GC-treated female rats, our study lacked cardinal histological analyses, such as histomorphometric assays for bone synthesis, including bone formation, mineral apposition, and bone cell apoptosis rates which are commonly used to evaluate bone formation.As an alternative, our study used ELISA to measure serum OCN levels, directly reflecting the bone formation rate.In addition, it quantified bone mass and trabecular bone microarchitectures using micro-CT and quantified mRNA and protein levels of RUNX2 -which play a crucial role in osteoblast differentiation and bone formationusing RT-qPCR and Western blots.Therefore, the lack of cardinal histological analyses does not affect our conclusion.Third, our study used female-aged rats because they are more prone to GIO than males due to the suppressive effect of GC on their already decreased estrogen levels (Cheng et al., 2022;Paggiosi et al., 2015).While the effect of BADGE treatment must be explored in aged male rats, our study provides the first evidence demonstrating that PPARγ inhibition is efficient, does not affect systemic glucose metabolism, and has no side effects in aged rats.Nevertheless, studies of aged male rats must be conducted to confirm and extend our findings.

Conclusions
In summary, our study indicates that BADGE prevents bone deterioration in aged MP-treated female rats by inhibiting PPARγ and provides evidence that BADGE inhibits GIO by decreasing BM adiposity and adipogenic transcription factor levels and increasing osteogenic transcription factor levels.In the future, BADGE may become a potential agent for treating older adults with GIO.

Fig. 1 .
Fig. 1.Alteration in the body weight of control, BADGE, MP, and MP + BADGE mice throughout the experiment period.Data are expressed as mean ± SD (n = 8 rats per group).P values represent Bonferroni-corrected P values (2-way repeated measures ANOVA).*P < 0.001 in the MP vs. the control group; # P < 0.001 in the MP vs. the BADGE group; a P < 0.001 in the MP + BADGE vs. the control group; b P < 0.001 in the MP + BADGE vs. the BADGE group.

Fig. 2 .
Fig. 2. Representative micro-CT 2D sectional images of the femur (A) and lumbar vertebrae (B).The femur and lumbar vertebrae of the MP group indicated trabecular microarchitecture deterioration, and BADGE treatment improved bone trabecular microarchitecture.

Table 1
Primer sequences used for the quantitative reverse-transcription polymerase chain reaction.

Table 2
Effects of BADGE treatment on the biomarkers of blood glucose, lipid, bone microarchitecture, marrow adipocytes, and bone.