PPARγ silencing enhances osteogenic differentiation of human adipose-derived mesenchymal stem cells

Peroxisome proliferator-activated receptor gamma (PPARγ) is the master regulator of adipogenesis, and has been indicated as a potential therapeutic target to promote osteoblast differentiation. However, recent studies suggest that suppression of PPARγ inhibits adipogenesis, but does not promote osteogenic differentiation in human bone marrow-derived mesenchymal stem cells (hBMSCs). It was reasoned that the osteogenic effect of PPARγ suppression may be masked by the strong osteogenesis-inducing condition commonly used, resulting in a high degree of matrix mineralization in both control and experimental groups. This study investigates the role of PPARγ in the lineage commitment of human adipose-derived mesenchymal stem cells (hADSCs) by interfering with the function of PPARγ mRNA through small interfering RNAs (siRNAs) specific for PPARγ2. By applying an osteogenic induction condition less potent than that used conventionally, we found that PPARγ silencing led to retardation of adipogenesis and stimulated a higher level of matrix mineralization. The mRNA level of PPARγ decreased to 47% of control 2 days after treatment with 50 nmol/l PPARγ2 siRNA, while its protein expression was 60% of mock control. In the meantime, osteogenic marker genes, including bone morphogenic protein 2 (BMP2), runt-related transcription factor 2 (Runx2), alkaline phosphatase (ALP) and osteocalcin (OC), were up-regulated under PPARγ silencing. Our results suggest that transient suppression of PPARγ promotes the onset of osteogenesis, and may be considered a new strategy to stimulate bone formation in bone tissue engineering using hADSCs.


Introduction
Peroxisome proliferator-activated receptor gamma (PPARc) is the master regulator of adipogenesis, and has been indicated as a critical switch in the commitment of progenitor cells to either the adipogenic or osteogenic pathway [1][2][3]. Down-regulation of PPARc is considered a novel strategy to promote bone regeneration, which is critical in the treatment of bone defects caused by trauma or bone resorption [4]. Transient suppression of PPARc in mouse embryonic stem (ES) cells by PPARc-specific siRNAs directs ES cells to differentiate into an osteoblastic lineage [5]. Peroxisome proliferator-activated receptor gamma siRNAs have also been reported to enhance bone formation in human preadipocytes and foetal-femur-derived mesenchymal cells, and reverse osteogenic repression of alcohol on human bone marrow-derived mesenchymal stem cells (hBMSCs) [6,7]. In vivo studies also demonstrate that mice with impaired expression of PPARc exhibit increased bone mass [8,9].
Recently, it was reported that suppression of PPARc through either PPARc antagonists or RNA interference inhibits adipogenesis, but does not promote osteogenic differentiation of hBMSCs, suggesting that PPARc may not be the master regulator of lineage determination in human bone marrow [10]. In this study, we provide a closely related but controversial observation in human adipose-derived mesenchymal stem cells (hADSCs), in which PPARc silencing resulted in higher level of osteogenic gene expression and matrix mineralization. The discrepancy between our study and the previous one was discussed.

Isolation and cell culture conditions of hADSCs
The protocol for this study was approved by the institutional review board of Kaohsiung Medical University Hospital. The detailed procedures of the isolation and characterization of hADSCs have been reported previously [11,12]. Human ADSCs were maintained in K-NAC medium, which consists of Keratinocyte-SFM (Gibco-BRL, Grand Island, NY, USA) supplemented with 2 mmol/l N-acetyl-L-cysteine (NAC) and 0.2 mmol/l Lascorbate 2-phosphate (Asc 2-P) [13]. Cells were seeded at 10 6 cells per 10-cm dish for protein extraction and western blot analysis, and at 5 9 10 5 cells per well in 6-well plates for RNA extraction in preparation for reverse transcription polymerase chain reaction (RT-PCR) or real-time PCR. For alizarin red S or oil red O staining, hADSCs were seeded at a density of 10 5 cells per well in 12-well plates.

siRNA transfection
Silencer â Select Validated siRNA specific for the human PPARc2 gene (5′-AAGAAATGACCATGGTTGACACAGAGAT-3′) was designed and synthesized by Life Technologies, Carlsbad, CA, USA. Silencer â Select Negative Control #1 siRNA (Life Technologies) or non-specific FITColigonucleotides (Life Technologies) were applied as negative controls to assess transfection efficiency. Transfection of siRNA was carried out 24 hrs after hADSCs were plated, using Lipofectamine RNAiMAX transfection reagent (Life Technologies) by following the manufacturer's instructions. The final siRNA concentration was 50 nmol/l.

Western blot analysis
Cell lysates obtained from hADSCs were separated by SDS-PAGE and transferred to polyvinylidene difluoride (PVDF) membrane. After blocking with 1% BSA in PBS-Triton X-100 (PBST) for 1 hr at room temperature, the membrane was washed with PBST and reacted overnight with mouse anti-human PPARc antibody (sc-7273, Santa Cruz, Dallas, TX, USA) or mouse anti-human b-actin antibody (sc-47778; Santa Cruz). After washing with PBST, the membrane was incubated with horseradish peroxidase (HRP)-conjugated goat antimouse secondary antibody (AP-124P, Millipore, Billerica, MA, USA) for 1 hr at room temperature. Bound antibodies were visualized with Amersham ECL Western Blotting Detection Reagents (GE Healthcare Life Sciences, Uppsala, Sweden).

Osteogenic differentiation
The concentrations of osteogenic inducing agents in the osteogenic induction medium prepared in this study were half of those in conventional ones. After treatment with PPARc2 siRNA for 48 hrs, hAD-SCs were cultured in osteogenic induction medium consisting of DMEM supplemented with 5 mmol/l b-glycerol phosphate, 50 nmol/l dexamethasone and 25 lmol/l L-ascorbate 2-phosphate, for another 2 weeks. Cells were fixed with 4% paraformaldehyde and stained with alizarin red S as described in our previous studies [16].

Adipogenic differentiation
After treatment with PPARc2 siRNA for 48 hrs, hADSCs were cultured in adipogenic induction medium consisting of DMEM supplemented with 1 lmol/l dexamethasone, 0.5 mmol/l methyl-isobutylxanthine, 10 lg/ml insulin and 100 lmol/l indomethacin for another 10 or 14 days (corresponding to 12 or 16 days after siRNA transfection). Cells were fixed with 4% paraformaldehyde at room temperature for 10 min., and incubated in oil red O staining solution (0.36% in 60% isopropanol) for 50 min. at room temperature. Cell-bound oil red O was extracted by adding 500 ll of isopropanol per well [17]. The amount of oil red O released was determined using a spectrophotometer at a wavelength of 540 nm.

Statistical analysis
Statistical analysis was performed on data from at least three independent experiments. Significant difference relative to the control was tested using one-way ANOVA. Difference in the mean was determined using Duncan's new multiple range test. Levels of significance of P < 0.05 and 0.01 were accepted as significant and highly significant respectively.

Suppression of PPARc expression and adipogenic differentiation by PPARc2 siRNA
Human ADSCs were transiently transfected with either PPARc2specific siRNA or a non-specific control oligonucleotide (mock siRNA), and PPARc gene expression was analysed by RT-PCR at various PPARc2 siRNA concentrations (10-70 nmol/l). Two days after siRNA transfection, the mRNA level of PPARc was lowered in a dose-dependent manner in the presence of 10-70 nmol/l PPARc2 siRNA (Fig. 1A, upper panel), whereas mock siRNA had no effect on the expression of PPARc (Fig. 1A, lower panel). In addition, cell morphology of hADSCs was not affected by PPARc2 siRNA within the concentration range tested (Fig. 1B). As determined by real-time PCR, treatment of hADSCs with 50 nmol/l PPARc2 siRNA for 2 days resulted in 53-59% suppression of PPARc gene expression, compared with cells without siRNA treatment or those treated with mock control (Fig. 1C). PPARc protein expression decreased with time in both control and siRNA-treated hADSCs, because of incubation in osteogenic induction medium after siRNA treatment. Nevertheless, treatment with PPARc2 siR-NA resulted in further reduction in intracellular PPARc level compared to the controls (Fig. 1D). In PPARc2 siRNA-treated cells, the amount of PPARc protein was~60% of the controls 2 days after transfection (Fig. 1E). When hADSCs were treated with PPARc2 siRNA followed by adipogenic induction for 2 weeks, lipid accumulation was significantly reduced compared to the control, indicating that adipogenic differentiation was blocked by interfering with the function of PPARc (Fig. 2).

Effect of PPARc suppression on osteogenic differentiation
Under PPARc suppression, the mRNA level of BMP2 and Runx2 was significantly higher than the controls 2 days after siRNA transfection ( Fig. 3A and B), whereas ALP was significantly up-regulated at days 8 and 14 (Fig. 3C). The mRNA expression of OC, one of the late-onset genes associated with osteogenesis, was significantly higher than the controls 4 days after transfection of PPARc2 siRNA (Fig. 3D). When conventional osteogenic inducing conditions (DMEM supplemented with 10 mmol/l b-glycerol phosphate, 100 nmol/l dexamethasone and 50 lmol/l L-ascorbate 2-phosphate) were applied, no significant difference in matrix mineralization was observed between hADSCs treated with PPARc2 siRNA and the control (data not shown). However, when hADSCs were induced to differentiate under a relative mild condition (DMEM supplemented with 5 mmol/l b-glycerol phosphate, 50 nmol/l dexamethasone and 25 lmol/l L-ascorbate 2-phosphate), hADSCs transfected with PPARc2 siRNA had a significantly  higher level of calcium deposition than control cells (Fig. 3E). These results indicate that osteogenic differentiation of hADSCs was accelerated by PPARc silencing, which may promote terminal differentiation of hADSCs towards the osteoblast lineage by up-regulating the expression of osteogenic genes.

Discussion
The signalling cascades related to PPARc are potential therapeutic targets to promote osteoblast differentiation in osteoporosis and age-related osteopenia and therapeutic methods that act through suppression of PPARc or inhibition of its ligand synthesis have been proposed [3,18]. However, recent studies indicate that suppression of PPARc does not promote osteogenic differentiation of hBMSCs [10], which is in contrast to the results of this and several other studies [5][6][7]. We reasoned that in in vitro studies, the osteogenic effect of PPARc suppression may be masked by the strong osteogenesis-inducing condition as used in the literature, resulting in a high degree of matrix mineralization in both control and experimental groups [10]. When a less potent induction condition was applied, the difference in the level of matrix mineralization between PPARcsilenced and control cells can be discerned (Fig. 3E). It is also possible that there exist a difference in sensitivity to PPARc suppression between hADSCs used in this study and hBMSCs reported in the literature [10]. Further genome-wide gene expression profiling on stem cells and precursor cells of various stages of maturation or animal models with PPARc silencing [19] is necessary to elucidate the role of PPARc in determining the differentiation fate.
In this study, we demonstrated that transient suppression of PPARc prior to osteogenic induction resulted in up-regulation of osteogenic genes such as BMP2, Runx2, ALP and OC (Fig. 3A-D), and a higher level of matrix mineralization (Fig. 3E). Our results suggest that PPARc silencing may promote the activation of BMP2 and Runx2/Cbfa1, as well as the downstream transcription of ALP and OC to induce the onset of osteogenic differentiation in hADSCs (Fig. 3F).
In conclusion, our study indicates that osteogenesis of hADSCs can be enhanced by suppressing adipogenesis through PPARc silencing. To the best of our knowledge, this is the first report to demonstrate that transient suppression of PPARc prior to osteogenic induction is sufficient to stimulate differentiation of hADSCs towards the osteoblast lineage. This investigation highlights possible therapeutic interventions targeting PPARc-related pathways to promote skeletal tissue regeneration.