Rosiglitazone suppresses RANKL‐induced NFATc1 autoamplification by disrupting the physical interaction between NFATc1 and PPARγ

Receptor activator of nuclear factor‐κB ligand (RANKL) is required for initiation of osteoclastogenesis, with the signaling pathway including the NF‐kB, c‐Fos, and nuclear factor of activated T cells, cytoplasmic 1 (NFATc1) transcription factors. Because NFATc1 expression is autoamplified, we investigated the molecular mechanism by which peroxisome proliferator‐activated receptor gamma (PPARγ) activation by the thiazolidinedione drug rosiglitazone decreases NFATc1 expression during RANKL stimulation. Western blotting demonstrated that rosiglitazone attenuated the increase in NFATc1 protein level induced by RANKL without affecting that of PPARγ. Immunofluorescence data indicated that rosiglitazone tended to suppress RANKL‐induced NFATc1 nuclear translocation, partly by reducing calcineurin activity, as reflected by the observed decrease in nuclear NFATc1 abundance. On coimmunoprecipitation, the intensity of the physical interaction between NFATc1 and PPARγ was unexpectedly higher in the RANKL‐stimulated group than in the control, but rosiglitazone reduced this to basal levels. Furthermore, RANKL failed to elevate mRNA expression of NFATc1 after PPARγ knockdown. ChIP assay indicated that rosiglitazone significantly reduced the binding of NFATc1 to its own promoter despite RANKL stimulation. These findings suggest that PPARγ activation by rosiglitazone blocks NFATc1 from binding to its own promoter, thereby reducing RANKL‐induced NFATc1 autoamplification.

Receptor activator of nuclear factor-jB ligand (RANKL) is required for initiation of osteoclastogenesis, with the signaling pathway including the NF-kB, c-Fos, and nuclear factor of activated T cells, cytoplasmic 1 (NFATc1) transcription factors. Because NFATc1 expression is autoamplified, we investigated the molecular mechanism by which peroxisome proliferator-activated receptor gamma (PPARc) activation by the thiazolidinedione drug rosiglitazone decreases NFATc1 expression during RANKL stimulation. Western blotting demonstrated that rosiglitazone attenuated the increase in NFATc1 protein level induced by RANKL without affecting that of PPARc. Immunofluorescence data indicated that rosiglitazone tended to suppress RANKL-induced NFATc1 nuclear translocation, partly by reducing calcineurin activity, as reflected by the observed decrease in nuclear NFATc1 abundance. On coimmunoprecipitation, the intensity of the physical interaction between NFATc1 and PPARc was unexpectedly higher in the RANKLstimulated group than in the control, but rosiglitazone reduced this to basal levels. Furthermore, RANKL failed to elevate mRNA expression of NFATc1 after PPARc knockdown. ChIP assay indicated that rosiglitazone significantly reduced the binding of NFATc1 to its own promoter despite RANKL stimulation. These findings suggest that PPARc activation by rosiglitazone blocks NFATc1 from binding to its own promoter, thereby reducing RANKL-induced NFATc1 autoamplification.
The formation and differentiation of osteoclasts occur through tightly regulated processes. In general, initiation of osteoclastogenesis requires receptor activator of nuclear factor-jB ligand (RANKL). RANKL signaling recruits multiple intracellular signaling molecules, of which the NF-kB, c-Fos, and nuclear factor of activated T cells, cytoplasmic 1 (NFATc1) transcription factors are important for osteoclastogenesis both in vivo and in vitro [1,2]. Previous studies have clearly shown that NFATc1-deficient stem cells or preosteoclasts fail to differentiate into osteoclasts [3][4][5], and conditional or rescued NFATc1-deleted mice develop severe osteopetrotic phenotypes [5,6]. Furthermore, NFATc1 directly regulates osteoclast-specific genes such as tartrate-resistant acid phosphatase (TRAP), calcitonin receptor, cathepsin K, and b3-integrin [7][8][9], thereby underscoring that NFATc1 is a master regulator of osteoclastogenesis. Because NFATc1 expression undergoes positive autoregulatory feedback [4,[10][11][12], NFATc1 nuclear translocation is necessary to initiate this step. Within the nucleus, NFATc1 binds to several sets of NFAT-binding motifs in its own promoter. When it binds to the promoter, NFATc1 forms heterodimers with other transcription factors, including peroxisome proliferator-activated receptor gamma (PPARc), activator protein-1 (AP-1), and NF-jB [13]. Accordingly, these transcription factors can play specific roles in modulating the transcriptional activity of NFATc1.
Peroxisome proliferator-activated receptor gamma is a nuclear receptor transcription factor critical for adipocyte differentiation and is also involved in bone homeostasis and remodeling. Activation of PPARc by thiazolidinedione drugs such as rosiglitazone and pioglitazone increases bone loss in normal mice and fracture incidence in diabetic humans [14,15], implying that PPARc may regulate the function or formation of both osteoblasts and osteoclasts. Although several reports suggested that PPARc activation by thiazolidinedione activates RANKL-induced osteoclastogenesis [16,17], many in vitro studies have demonstrated that PPARc activation inhibits it [18][19][20][21][22].
The molecular mechanisms by which PPARc activation inhibits osteoclastogenesis vary from reduction in NF-jB expression, suppression of NF-jB nuclear translocation and its DNA binding, and downregulated expression of c-Fos, to inhibition of mitogenactivated protein kinase activity [19,21,23]. Although PPARc activation also reduces NFATc1 expression at the transcriptional and protein levels during RANKLinduced osteoclastogenesis, the mechanism by which PPARc activation decreases NFATc1 expression has not been investigated. We explored the molecular mechanisms by which PPARc activation by rosiglitazone suppresses NFATc1 expression during the early period of RANKL-induced osteoclastogenesis using RAW264.7 monocytes.

Materials and methods
Cell culture and drug treatment RAW264.7 cells were routinely cultured in a-MEM (Gibco, Waltham, MA, USA) supplemented with 10% FBS (Gibco) and 1% penicillin-streptomycin at 37°C in a 5% CO 2 condition. Rosiglitazone (Cayman Chemical, Ann Arbor, MI, USA) was added to the medium 30 min before RANKL (R&D Systems, Minneapolis, MN, USA) administration, and cells were maintained for an additional 24 h. For siRNA experiments, complete DMEM/F12 (Gibco) without antibiotics was used as the culture medium.

Osteoclastogenesis and cell viability
To observe osteoclast differentiation, RAW264.7 cells were cultured in 96-well plates in medium containing various concentrations of rosiglitazone in the presence 50 lgÁmL À1 of RANKL. After 5 days, cells were stained for TRAP with a commercial kit (TRACP & ALP double-stain kit, Takara, Shiga, Japan). Cytotoxicity of rosiglitazone was assayed using an MTT reagent (Amresco, Inc., Solon, OH, USA). In this case, cells were treated for 1 or 5 days without RANKL addition.

Conventional RT-PCR
Total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA), and 1 lg of RNA was used as the template to synthesize cDNA using AccuPower

Calcineurin activity
Cellular calcineurin (PP2B) activity was measured using a kit according to the manufacturer's instructions (Enzo Life Sciences, Farmingdale, NY, USA). Cells incubated in six-well plates were lysed in lysis buffer. Cell lysate (20 lg protein) from each group or a positive calcineurin (40 U) control was added to the tube with or without EGTA buffer, and the reaction with phosphopeptide substrate was conducted at 30°C for 30 min and terminated by adding stop buffer. The phosphate standard curve was prepared by serial dilution of phosphate solution. Because calcineurin is a Ca ++ /calmodulin-dependent phosphatase, its activity was determined by subtracting Ca ++ -independent activity in the presence of EGTA from total activity in the absence of EGTA. Final calcineurin activity was presented as the amount of inorganic phosphate released per lg of protein.
In our experiment, the presence of EGTA quenched total calcineurin activity by approximately 60-70%.

Immunofluorescence
Immunofluorescent detection of NFATc1 was performed according to the method described by Hasega et al. [25] with minor modifications. Briefly, cells were plated at less than 50% confluence (a density of 5 9 10 4 cells/chamber) in a four-chamber slide (Nunc, Rochester, NY, USA), grown overnight, and treated with vehicle or drugs for 24 h in complete a-MEM. After fixation with 4% paraformaldehyde for 10 min at room temperature, cells were permeabilized with 0.2% Triton X-100 for 15 min and then blocked with 5% bovine serum albumin at room temperature for 1 h. After washing and blocking, anti-NFATc1 antibody was added at a dilution of 1 : 100 and the fixed cells were incubated at 4°C overnight. NFATc1 signals were detected with Alexa Fluor 488-conjugated secondary antibody (Invitrogen) at a dilution of 1 : 500 for 1 h, after which the fixed cells were mounted in DAPI-containing medium (Sigma).

Coimmunoprecipitation
Peroxisome proliferator-activated receptor gamma or NFATc1 was immunoprecipitated by incubating whole cell lysates (200-300 lg of protein) with anti-PPARc or anti-NFATc1 antibody overnight at 4°C. NFATc1 immunocomplexes were captured with protein G beads (Invitrogen) for 1 h at 4°C. PPARc immunocomplexes were captured using the ImmunoCruz Optima C System (Santa Cruz) to avoid cross-reaction with immunoglobulin heavy chain, which has a molecular weight similar to that of PPARc. The resulting immunocomplexes were collected by centrifugation. After washing, the pellets were eluted and immunoblotted as described above.
siRNA Cells were seeded at less than 80% confluence in six-well plates and cultured in antibiotic-free DMEM/F12 with 10% FBS overnight. Transfection was carried out according to the manufacturer's protocol. SmartPool PPARc siRNA or control siRNA (Dharmacon, Lafayette, CO, USA) was used at a final concentration of 100 nM using DharmaFECT 4 transfection reagent (Dharmacon). At 24 h after transfection, cells were treated with relevant drugs and cultured for an additional 24 h in DMEM/F12.

ChIP assay
Cells were cultured in 100-mm dishes and transfected with control siRNA or PPARc siRNA. ChIP assays were performed with monoclonal anti-NFATc1 or mouse IgG. ChIP procedures were performed according to the protocol provided by the supplier (Upstate Biotechnology, Lake Placid, NY, USA). Primer sequences for amplifying the region including the NFAT-binding sites of the mouse NFATc1 P1 promoter were published previously [11]. The precipitated DNA was evaluated by quantitative PCR using Universal SYBR Green Master Mix (Takara, Shiga, Japan). C t values were normalized to input using the comparative delta-delta method [26].

Statistical analysis
Data are expressed as mean AE SD and were analyzed by either one-way or two-way ANOVA with the LSD test using IBM SPSS 23 (Armonk, NY, USA). P-values < 0.05 were considered statistically significant.

Effect of rosiglitazone on osteoclastogenesis and NFATc1 expression
Previous reports demonstrated that PPARc activation by pioglitazone or rosiglitazone inhibits RANKLinduced osteoclastogenesis. As shown in Fig. 1A,B, our results also indicated that rosiglitazone treatment attenuated RANKL-stimulated osteoclast differentiation in a dose-dependent manner. Cytotoxicity of rosiglitazone was not observed on days 1 or 5 with the range of concentrations used (Fig. 1C). However, treatment with 5 lM of rosiglitazone reduced the level of RANKL-induced NFATc1 abundance by about 40% without reducing PPARc abundance (Fig. 1D-F). Rosiglitazone alone did not affect the level of either PPARc or NFATc1.

Effect of rosiglitazone on NFATc1 nuclear translocation
Nuclear factor of activated T cells, cytoplasmic 1 exhibits strong phosphorylation of serine residues in the resting state of cells. NFATc1 is initially dephosphorylated by the protein phosphatase calcineurin, which is a serine/threonine protein phosphatase that is composed of a catalytic A subunit and regulatory B subunit. Dephosphorylated NFATc1 is then translocated into the nucleus. Thus, we examined whether suppression of RANKL-induced NFATc1 expression was due to blockage of NFATc1 nuclear translocation. As shown in Fig. 2A, RANKL stimulation moderately increased the activity of calcineurin compared with the control, and rosiglitazone reduced this increase. However, protein levels of calcineurin Aa and calcineurin B2 were not altered regardless of treatment with RANKL and/or rosiglitazone (Fig. 2B). Because changes in calcineurin activity can be accompanied by alterations in NFATc1 localization, both immunofluorescence microscopy and western blotting of the nuclear fraction were carried out.
Immunofluorescence images revealed that rosiglitazone inhibited RANKL-induced NFATc1 nuclear translocation (Fig. 2C). NFATc1 levels in the nuclear fraction showed similar alterations (Fig. 2D), although these findings were not statistically significant (Fig. 2E, P = 0.11). Meanwhile, most PPARc remained in the nuclear fraction and its protein level was unchanged despite rosiglitazone treatment (Fig. 2D). Collectively, these findings imply that rosiglitazone inhibits NFATc1 nuclear translocation by suppressing the action of calcineurin activated by RANKL.

Physical interaction between PPARc and NFATc1
Stimulation of steroid receptors by ligands disturbs NFAT binding to NFAT sites on DNA by complexing NFAT with steroid receptors [27]. To test whether PPARc activation has similar mechanisms, we analyzed the protein-protein interaction between PPARc and NFATc1 in situ. As shown in Fig. 3, PPARc-NFATc1 binding was barely detectable under unstimulated conditions, and rosiglitazone alone did not affect this complex formation. However, RANKL strongly induced formation of the PPARc-NFATc1 complex, which was found in immunoprecipitation using either anti-NFATc1 or anti-PPARc, and this induction was greatly attenuated by rosiglitazone. This indicates that ligand-bound PPARc disrupts the physical interaction between NFATc1 and PPARc.

Effects of PPARc knockdown on NFATc1 mRNA expression and NFATc1 autoregulation
To further explore the effects of PPARc-NFATc1 complex formation on positive autoregulatory NFATc1 expression, we performed a PPARc knockdown experiment. When cells were transfected with PPARc siRNA, PPARc protein was barely expressed (Fig. 4A). As shown in Fig. 4B, knockdown of PPARc significantly abolished mRNA expression of NFATc1 compared to levels seen with control siRNA, but knockdown barely affected the mRNA transcription of RANK, an NFATc1-independent gene. It should be noted that the effect on PPARc-dependent expression of TRAP mRNA (TRAP, an NFATc1-dependent osteoclast-specific gene), although visible, was less profound than the effect on NFATc1 mRNA expression. TRAP transcription can be also subject to regulation by RANKL-mediated molecules, such as reactive oxygen species (ROS) and MITF transcription factor, which are physically unassociated with PPARc [28,29]. These findings indicate that PPARc-NFATc1 complex formation is critical for autoregulatory transcription of the NFATc1 gene and expression of NFATc1-dependent genes. To test whether transcriptional autoregulation of NFATc1 is followed by NFATc1 binding to its own promoter, we performed a ChIP assay. As depicted in Fig. 4C, quantitative PCR from the ChIP assay showed that rosiglitazone significantly blocked NFATc1 binding to its promoter despite RANKL stimulation, whereas in the absence of PPARc, NFATc1 binding was completely prevented, regardless of RANKL stimulation.

Discussion
Evidence indicates that PPARc activation by natural ligands or synthetic drugs suppresses RANKL-induced osteoclastogenesis by inhibiting NFATc1 expression [20,21]. We showed that rosiglitazone significantly attenuated osteoclast differentiation and the induction of NFATc1 expression by RANKL (Fig. 1). Because the NFATc1 transcription factor is a critical regulator in RANKL-stimulated osteoclastogenesis, we further explored the molecular mechanism by which NFATc1 expression was inhibited by PPARc activation. NFATc1 is autoamplified during the early phase of osteoclastogenesis [4,11,12]. Prior to NFATc1 autoamplification, NFATc1 nuclear translocation is conventionally held to initially occur via calcineurin-mediated dephosphorylation of NFAT. In our experiments, because protein levels of calcineurin subunits Aa and B2 were unchanged in treated groups compared with the control, RANKLinduced increases in calcineurin activity are presumed to play certain roles in NFATc1 entry into the nucleus, which was to some degree inhibited by rosiglitazone. This was reflected by the reduced degree of NFATc1 nuclear translocation observed in immunofluorescence images and the lower NFATc1 protein level in the nuclear fraction after western blotting (Fig. 2C,D). Similar to our findings, Bao et al. found that rosiglitazone prevents endothelin-1-induced increases in both the activity and protein level of calcineurin in cardiomyocytes [30]. Sander et al. also showed that rosiglitazone antagonized vascular endothelial growth factor (VEGF)-mediated NFATc1 nuclear translocation in cardiac valve endothelial cells [31], but their study did not examine the action of calcineurin. On the other hand, Wagner et al. observed that PPARb stimulation in neonatal cardiomyocytes directly induced calcineurin expression [32]. It remains unclear whether PPARc affects either the activity or protein level of calcineurin in a cell type-specific or PPAR isoform-specific manner.
Aside from the action of PPARc activation on calcineurin, the reduction in RANKL-induced NFATc1 expression by rosiglitazone could be influenced by changes in signaling pathways upstream of NFATc1 or modulatory actions of other transcriptional factors on NFATc1. Although a predicted PPAR response element (PPRE) is found in the NFATc1 promoter, direct transcriptional repression of PPARc on NFATc1 expression is not possible because rosiglitazone alone did not affect NFATc1 abundance (Fig. 1). PPARc negatively regulates gene expression through several transrepression mechanisms. One such mechanism is through the interaction of PPARc with transcription factors such as NF-jB, AP-1, and NFAT [33,34]. In this regard, Yang et al. reported that the PPARc agonist troglitazone increases formation of a complex between PPARc and NFATc1 in PMA/ionomycin-activated T cells [35]. Similarly, Bao et al. demonstrated that rosiglitazone greatly enhances association of PPARc with NFATc4 in endothelin-1stimulated cardiomyocytes [30]. We unexpectedly found that formation of an NFATc1-PPARc complex was strongly induced by RANKL stimulation and was reduced to basal levels by rosiglitazone treatment (Fig. 3). This implies that once bound or activated by ligands, PPARc can be separated from the NFATc1-PPARc complex rather than remaining associated with NFATc1. Therefore, if PPARc was absent or liganded within cells, NFATc1 would either remain alone or be associated with other transcriptional partners even upon RANKL stimulation. This hypothesis was partly confirmed by our PPARc knockdown experiment. Knockdown of PPARc resulted in suppression of mRNA expression of NFATc1 and NFATc1-dependent TRAP, an osteoclast-specific gene, without affecting mRNA expression of NFATc1-independent RANK (Fig. 4B). In addition, our ChIP experiment showed that PPARc knockdown completely inhibited NFATc1 from binding to the NFAT site in its own promoter, regardless of RANKL stimulation (Fig. 4C). Taken together, when the NFATc1-PPARc complex is dissociated by PPARc activation, NFATc1 seems to lose its binding affinity to NFAT sites, thereby lessening the transcriptional activity of NFATc1 that regulates its autoamplification and the mRNA expression of NFATc1-dependent genes. Because both NFATc1 and NF-kB share a core DNA binding motif (GGA), our findings are consistent with the molecular mechanical explanation suggested by Wen et al. [36]. According to their hypothesis, PPARc activation decreases the protein level of tumor necrosis factor-astimulated RANTES in kidney mesangial cells because liganded PPARc dissociates from the NF-jB-PPARc complex, inactivating the ability of NF-jB to bind to the jB site in the RANTES promoter, thereby blocking RANTES expression. Therefore, we presume that dissociation rather than association of the NFATc1-PPARc complex plays an important role in suppression of NFATc1 expression.
In conclusion, PPARc activation inhibits RANKLinduced formation of an NFATc1-PPARc complex. Once dissociated from the complex, NFATc1 cannot bind to NFAT response elements in its own promoter, which suppresses the autoregulatory expression of NFATc1 during the early period of RANKL-stimulated osteoclastogenesis.