SEARCH

SEARCH BY CITATION

Keywords:

  • RESVERATROL;
  • SIRT1;
  • FOXO3A;
  • RUNX2;
  • OSTEOGENESIS;
  • MESENCHYMAL STEM CELLS

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Reports of the bone-protective effects of resveratrol, a naturally occurring phytoestrogen and agonist for the longevity gene SIRT1, have highlighted this compound as a candidate for therapy of osteoporosis. Moreover, SIRT1 antagonism enhances adipogenesis. There has been speculation that resveratrol can promote osteogenesis through SIRT1, but the mechanism remains unclear. In this study we investigated the molecular mechanism of how resveratrol can modulate the lineage commitment of human mesenchymal stem cells to osteogenesis other than adipogenesis. We found that resveratrol promoted spontaneous osteogenesis but prevented adipogenesis in human embryonic stem cell–derived mesenchymal progenitors. Resveratrol upregulated the expression of osteo-lineage genes RUNX2 and osteocalcin while suppressing adipo-lineage genes PPARγ2 and LEPTIN in adipogenic medium. Furthermore, we found that the osteogenic effect of resveratrol was mediated mainly through SIRT1/FOXO3A with a smaller contribution from the estrogenic pathway. Resveratrol activated SIRT1 activity and enhanced FOXO3A protein expression, a known target of SIRT1, in an independent manner. As a result, resveratrol increased the amount of the SIRT1-FOXO3A complex and enhanced FOXO3A-dependent transcriptional activity. Ectopic overexpression or silencing of SIRT1/FOXO3A expression regulated RUNX2 promoter activity, suggesting an important role for SIRT1-FOXO3A complex in regulating resveratrol-induced RUNX2 gene transcription. Further mutational RUNX2 promoter analysis and chromatin immunoprecipitation assay revealed that resveratrol-induced SIRT1-FOXO3A complex bound to a distal FOXO response element (−1269/−1263), an action that transactivated RUNX2 promoter activity in vivo. Taken together, our results describe a novel mechanism of resveratrol in promoting osteogenesis of human mesenchymal stem cells by upregulating RUNX2 gene expression via the SIRT1/FOXO3A axis. © 2011 American Society for Bone and Mineral Research


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Hormone-replacement therapy (HRT) is known to be an effective therapy for postmenopausal osteoporosis, but recent findings show that its use can be accompanied an increased risk of breast cancer and cardiovascular disease.1 Increasingly, natural phytoestrogens have been focused as alternative therapeutic agents for HRT. Resveratrol is a naturally occurring polyphenolic compound in red wine and numerous plants that has been regarded as a phytoestrogen, selectively activating estrogen receptor (ER) signaling.2, 3 Several biologic effects of resveratrol have been reported, including cardiovascular protection4 and anticancer,5 anti-inflammation, antioxidant,6 antiaging,7, 8 and bone-protective activities.5, 9 These multiple beneficial effects of resveratrol have been postulated to involve pathways other than the ER pathways, but data remain limited. We have been interested in the bone-protective activity of resveratrol in osteoblasts that is induced through ER-dependent bone morphogenetic protein 2 (BMP-2) upregulation,5 resulting in reducing loss of bone mass in an ovarietomized rat model in vivo.10 Recently, a few studies have suggested that in addition to the bone-protective effect in osteoblasts, resveratrol could enhance osteogenesis and inhibit adipogenesis in murine pluripotent mesenchymal cell lines or bone marrow–derived mesenchymal stem cells (MSCs).9, 11–13 However, the underlying mechanism of resveratrol-induced osteoblast differentiation on MSCs remains obscure except the well-characterized ER-dependent and Wnt/β-catenin pathways.

One of the most interesting and novel molecular targets of resveratrol is SIRT1, a member of the sirtuin family of nicotinamide adenine dinucleotide (NAD+)–dependent deacetylases, which is found to be an antiaging gene.8, 14, 15 SIRT1 deacetylates various transcription factors in the nucleus such as p53,16, 17 class O subfamily of forkhead box (FOXO),18 Ku70,19 and NF-κB,20 thus playing an important role in cell differentiation, cell survival, tumorigenesis, and metabolism.21, 22 An important target of SIRT1 is FOXO3A, a member of the multifunctional FOXO family of transcription factors. When SIRT1 deacetylates FOXO3A, transcription of FOXO3A-dependent stress-resistance genes is triggered and contributes to increased longevity.18, 23–25

The lineage commitment of MSCs between osteoblasts and adipocytes is competitively balanced and tightly controlled at the transcriptional level, and the two transcription factors RUNX2 and PPARγ2 are regarded as critical in initiating osteogenesis and adipogenesis, respectively.26–28 Recently, PPARγ2 has been shown to play a dominant role in repressing osteoblast differentiation by numerous gene-gene interactions.29 Interestingly, SIRT1 is presumed to enhance osteogenic differentiation indirectly through repressing PPARγ2 and inhibiting adipocyte formation.11, 30 A recent study demonstrated that PPARγ2 inhibits RUNX2 mRNA transcription and interferes with RUNX2-mediated transactivation activity in osteoblast cell lines,31 thereby raising the possibility that a regulatory effect of SIRT1 on RUNX2 transcription also exists. However, despite numerous studies on RUNX2 promoter regulation, there have been no reports confirming direct interactions between these two genes.27, 32 We report in this study the role of SIRT1 in RUNX2 transcriptional regulation and resveratrol-mediated osteogenesis in human MSCs, which we found to be mediated through SIRT1 interaction with FOXO3A. Our data provide a new insight into the molecular mechanism underlying the important bone-protective effect of resveratrol.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Cell culture and differentiation studies

Human embryonic mesenchymal progenitors (EMPs) were provided by Dr B Linju Yen. EMPs were derived from human embryonic stem cells, as reported previously, and cultured in bone marrow MSC medium consisting of low-glucose DMEM, 100 U/mL of penicillin/streptomycin (Gibco-Invitrogen, Carlsbad, CA, USA), and 10% fetal bovine serum (FBS; selected lots from Hyclone, Logan, UT, USA) in a humidified atmosphere of 95% air and 5% CO2 at 37 °C.33 Human kidney epithelial 293T cell line was obtained from American Type Culture Collection (ATCC, Manassas, VA, USA) and maintained in DMEM supplemented with 10% FBS, 100 U/mL of penicillin/streptomycin, 2 mM L-glutamine, and 0.1 mM nonessential amino acids (all from Gibco-Invitrogen). At confluence, cells were passaged by trypsinization (0.025% trypsin/EDTA; Gibco-Invitrogen) and expanded. For adipogenic differentiation, cells were cultured in complete medium with 0.5 mM isobutyl-methylxanthine, 1 µM dexamethasone, and 10 µM insulin (all from Sigma-Aldrich, St Louis, MO, USA). Thereafter, cells were fixed with 4% paraformaldehyde and stained with oil red O solution (Sigma-Aldrich) for 10 minutes to visualize lipid vacuoles. To quantify the oil droplet formation, three random fields with ×200 magnification were selected. The relative area of oil red O stain positivity within each individual cell was calculated and quantified by Image J analysis software (NIH, Bethesda, MD, USA). For osteogenic differentiation, cells were cultured in complete medium with 0.2 mM L-ascorbic acid 2-phosphate, 1 µM dexamethasone, and 0.05 mM β-glycerophosphate and replaced every 3 days. At the end of differentiation, alizarin red staining was performed to analyze calcium deposits. Briefly, cells were fixed with 100% methanol for 30 minutes, washed with boric acid buffer (0.1 M, pH 4.0), and stained with alizarin red solution (40 mM, pH 4.2; Sigma-Aldrich) for 30 minutes. After repeated washing with boric acid buffer and distilled water, calcium deposits were seen and visualized. Elution of alizarin red stain was performed and quantified by spectrophometric analysis by reading the absorbance at 520 nm.

Alkaline phosphatase (ALP) activity

After culture, cells were lysed by protein lysis buffer without protease inhibitor. Cellular ALP activity was assayed colorimetrically by incubating protein lysates with the substrate p-NPP (Sigma-Aldrich) at 37 °C for 30 minutes. The absorbance was read at 405 nm and normalized against the corresponding protein concentration. The relative fold change in ALP activity was shown.

Quantitative RT-PCR

Primer sets for human genes RUNX2 (sense, 5'-GGT TAA TCT CCG CAG GTC ACT-3'; antisense, 5'-CAC TGT GCT GAA GAG GCT GTT-3'), osteocalcin (sense, 5'-TCA CAC TCC TCG CCC TAT TG-3'; antisense, 5'-TCG CTG CCC TCC TGC TTG-3'), ALP (sense, 5'-AGC TGA ACA GGA ACA ACG TGA-3'; antisense, 5'-CTT CAT GGT GCC CGT GGT C-3'), COL1A1 (sense, 5'-ACC GCC CTC CTG ACG CAC -3'; antisense, 5'- GCA GAC GCA GAT CCG GCA G-3'), PPARγ2 (sense, 5'-GAA CGA CCA AGT AAC TCT CC-3'; antisense, 5'-CGC AGG CTC TTT AGA AAC TCC-3'), LEPTIN (sense, 5'-TTC ACA CAC GCA GTC AGT CTC C-3'; antisense, 5'-GGC ATA CTG GTG AGG ATC TG-3'), SOD2 (sense, 5'-TGG ACA AAC CTC AGC CCT AAC-3'; antisense, 5'-AAA CCA AGC CAA CCC CAA CC -3'), and β-actin (sense, 5'-GGC ACC CAG CAC AAT GAA G-3'; antisense, 5'-TGC GGT GGA CGA TGG AGG-3') were synthesized by Blossom Biotechnologies, Inc. (Taipei, Taiwan). Total RNA of cultured EMPs was extracted using REzol C&T reagent (PROtech Technologies Inc., Taipei, Taiwan), a new, improved isolation method based on the acid guanidinium-thiocyanate-phenol-chloroform extract procedure of Chromczynski and Sacchi. After DNase I digestion of the total extracted RNA, 3 µg of RNA was used to synthesize single-strand cDNA using Impron II reverse transcriptase (Promega, San Luis Obispo, CA, USA). Real-time quantitative RT-PCR was performed with the double-stranded DNA-binding dye SYBR Green I (Applied BioSystems, Inc., Foster City, CA, USA) on the ABI 7900 Real-Time PCR System (Applied BioSystems, Inc.), and relative gene expression was analyzed as indicated by the manufacturer.

Plasmid, transient transfection, and luciferase assay

The human RUNX2 promoter–luciferase reporter plasmid (RUNX2-Luc) was constructed by subcloning the upstream promoter region of the human RUNX2 gene (−1557/ + 32) into pGL3-basic Luc (Promega). The wild-type FOXO response element (FRE) (−1269/−1263; ATA AAT A) of RUNX2-Luc was mutated into the mutant FRE version (AGA GAT A) by Quick Change XL Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA, USA). The plasmids, including FRE-luc, HA-FOXO3A, HA-FOXO3A WT-DBM (H212R),34 Flag-SIRT1, and Flag-SIRT1-H363Y,18 were purchased from Addgene (www.addgene.org). The wild-type and mutant versions of the RUNX2-Luc, together with a CMV-driven β-galactosidase construct (pCMV-β-gal), were transiently transfected at a 9:1 ratio into 80% subconfluent EMPs using the DNAFect LT transfection reagent (ATGCell, Edmonton, Alberta, Canada) according to the manufacturer's recommendations. After 24 hours, the transfected cells were changed into fresh complete medium with various treatments for 48 hours. Cell extracts were prepared, and luciferase activity was measured using the Promega Luciferase Assay System standardized against β-galactosidase activity. The values shown are the mean (± SD) of three replicates and at least three independent trials.

SIRT1 activity assay

SIRT1 activity was determined with a SIRT1 Fluorimetric Drug Discovery Kit (BML-AK555; Biomol International, Postfach, Lausen, Switzerland) according to the manufacturer's instructions.35 To determine SIRT1-dependent activity, cultured cells with various pretreatments were homogenized with protein lysis buffers, and then 25 µL of protein lysate was incubated with 25 uL of 2× substrates solution containing 200 µM Fluor de Lys–SIRT1, 1 mM NAD+, and 1 µM trichostatin A in SIRT1 assay buffer. After 40 minutes of incubation at 37 °C, the reaction was terminated by adding 50 µL of 1× Fluor de Lys Developer solution containing 2 mM nicotinamide. Plates were incubated at 37 °C for 20 minutes. Values were determined by reading fluorescence on a fluorometric plate reader (Beckman Coulter PARADIGM; Beckman Coulter Inc, Fullerton, CA, USA) with an excitation wavelength of 360 nm and an emission wavelength of 460 nm. Calculation of net fluorescence included the subtraction of a reaction time zero control of untreated control lysate. The net fluorescence was normalized against the corresponding protein concentration and shown as the mean ± SD of three replicates.

Nuclear extract preparation and Western blotting

Preparation of the nuclear extracts was carried out using ProteoJET Cytoplasmic and Nuclear Protein Extraction Kit (Fermentas, Glen Burnie, MD, USA) according to manufacture's recommendation. Then 50 to 100 µg of protein lysates was separated by 8.5% SDS-PAGE and blotted onto nitrocellulose membranes (protran NC). The membranes were blocked with 4% nonfat dry milk in Tris-buffered saline (TBS) for 1 hour and then probed with the appropriate primary antibody at 4 °C overnight. After washing five times with TBS, the NC membrane was incubated with secondary horseradish peroxidase (HRP)–conjugated antibodies (PerkinElemer Biosciences, Waltham, MA, USA) for 1 hour. The primary antibodies including anti-PPARγ1/2 were purchased from Abcam (A3409A; Cambridge, UK), anti-α-tubulin was from Sigma (T6074), and anti-SIRT1 (H-300), anti-FOXO3A (H-144), and anti-histone H1 (AE-4) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). The HRP signal was detected by a Western Lightning Chemiluminescence Kit (PerkinElemer Biosciences). Subsequent quantification was done by using the “quantity one” software (provided by BioRad, Hercules, CA, USA).

Immunoprecipitation (IP)

Protein lysates (300 µg) were preincubated with 3 µg of antibody of interest overnight on the rotator at 4 °C. Next, 50 µL of PureProteome Protein A magnetic beads (Millipore, St Charles, MO, USA) was added and incubated for additional 2 hours at 4 °C. The immunoprecipitated complex was pulled down by the magnetic rack, washed with PBS/0.1% Tween 20 surfactant three times, and eluted into 40 µL of electrophoresis buffer by heating at 90 °C for 10 minutes. The following Western blotting was performed to analyze the components of IP complex.

Chromatin immunoprecipitation (ChIP) assay

The detailed procedure was carried out using Magna ChIP A Kit (Millipore) according to the manufacturer's recommendations. Briefly, the protein-DNA complexes of EMPs were cross-linked by formaldehyde (1% final concentration) for 10 minutes, and then glycine (0.125 M final concentration) was added to terminate the reaction. Cells were scraped and resuspended in cell lysis buffer, and then the collected nuclei were resuspended in 0.5 mL of nuclear lysis buffer with protease inhibitors. The resuspended nuclei were subjected to sonication on ice with total 90-second pulses (5-second pulse and 5-second rest). The sonicated samples were centrifuged to spin down cell debris, and the soluble chromatin solution was immunoprecipitated using 5 µg of polyclonal antibody against SIRT1 or FOXO3A (Santa Cruz Biotechnology) and magnetic protein A beads provided within the kit. One part of the chromatin without immunoprecipitation was subjected to DNA purification and served as the input control. The protein-bound immunoprecipitated DNA was washed repeatedly with low-salt wash buffer, high-salt wash buffer, LiCL wash buffer, and TE buffer, and then immune complexes were eluted by adding elution buffer followed by incubation with proteinase K (final 0.1 µg/mL) for at least 6 hours at 65 °C to reverse cross-links and incubation for 1 hour at 37 °C with RNase A (final 0.1 µg/mL). Finally, the DNA was purified using a QIAquick PCR Purification Kit (Qiagen, Valencia, CA, USA). PCR then was performed to monitor the amount of immunoprecipitated chromatin bound to the FRE site (−1269) of RUNX2 promoter (sense, 5'-AAA TAC TCC ATC GCT CCC AAC T-3'; antisense, 5'-TAT CTC CTT TTT TGC CCC TTT G-3'), whereas the corresponding input DNA was checked by PCR of a random GAPDH promoter region (sense, 5'-TAC TAG CGG TTT TAC GGG CG -3'; antisense, 5'-TCG AAC AGG AGG AGC AGA GCG A-3').

RNA interference experiments (siRNA)

Dharmacon siRNA ON-TARGETplus SMART pool reagents (Thermo Scientific, Waltham, MA, USA) were transfected into EMPs (100 nM per 5 × 104 cells within 24-well plates) by using DharmaFECT 1 reagent (Thermo Scientific) to silence human SIRT1 and FOXO3A gene expression according to manufacturer's recommendations. After 24 hours of transfection, cells were transfected again with RUNX2-luc by DNAFect LT transfection reagent (ATGCell), and luciferase activities were measured after 48 hours.

Statistical analysis

All experimental results and measurements are triplicates and expressed as the means ± SD. To confirm reproducibility, all experiments were repeated at least three times. Statistical analysis was performed using Student's paired t test.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Resveratrol enhances osteogenesis and inhibits adipogenesis in EMPs

Human embryonic stem cell–derived mesenchymal progenitors (EMPs) have been reported as a new source of multipotent MSCs possessing similar cell surface marker profile, differentiation capacity, and immunomodulatory properties as their bone marrow–derived counterpart.36–39 In this study, we used EMPs as a model for evaluating the molecular mechanisms of reveratrol-induced osteogenesis. Multipotency of EMPs was demonstrated first. On stimulation in the appropriate differentiation medium, EMPs readily differentiated into osteogenic and adipogenic lineages, with the corresponding lineage-specific gene expression and phenotypes confirmed by alizarin red and oil red stains, respectively. As shown in Fig. 1A, under osteogenic medium (OM) induction for 7 to 14 days, expression of osteogenic genes including RUNX2, ALP, collagen type 1α1 (COLIA1), and osteocalcin (OCN) was increased, and osteogenic marker ALP enzyme (activity) and calcium deposition were increased by five- and twofold, respectively. Under adipogenic medium (AM) induction, adipogenic genes including PPARγ2 and leptin were increased within 3 days, and the percentage of intracellular oil droplet formation was dramatically elevated within 6 days (Fig. 1B). The effects of resveratrol on lineage commitment of EMPs were evaluated subsequently. ALP activity and calcium deposition were both elevated 1.5-fold by resveratrol (5 µM) compared with untreated controls in complete medium (Fig. 1C). On the other hand, treatment with resveratrol (10 µM) for 6 days prevented AM-induced adipocyte formation by 50% (Fig. 1D). Overall, we observed that resveratrol in EMPs modulated commitment toward an osteogenic rather than adipogenic lineage.

thumbnail image

Figure 1. Resveratrol promotes osteogenic but inhibits adipogenic differentiation in EMPs. (A) The osteogenic differentiation potential of EMPs is shown by RT-PCR analysis for expression of osteogenic lineage–specific genes, including RUNX2, ALP, COL1A1, and OCN, under complete medium (CM) or osteogenic medium (OM) induction for 6 to 9 days. ALP activity was determined after OM induction for 6 days (ap < 0.005 by Student's t test), and alizarin red staining was performed to monitor mineralization of EMPs after OM induction for 14 days (phase-contrast microscopy, magnification ×200), with quantification after elution of stain (ap < 0.005 by Student's t test). (B) The adipogenic differentiation potential of EMPs by RT-PCR analysis for expression of adipocytic lineage–specific genes including PPARγ2 and leptin after adipogenic medium (AM) induction for 1 day. Oil red staining was performed to monitor adipocyte formation after AM induction for 6 days (phase-contrast microscopy, magnification ×200), with quantification for droplet formation by software ImageJ analysis (ap < 0.0005 by Student's t test). (C) ALP activity and alizarin red staining of EMPs treated with 5 µM of resveratrol for 6 and 14 days was determined (ap < 0.005 and bp < 0.05 by Student's t test). (D) Oil red staining of AM-induced EMPs in the presence of resveratrol (10 µM) for 6 days was performed to analyze oil droplet formation with quantification by software ImageJ analysis (ap < 0.05 by Student's t test).

Download figure to PowerPoint

Regulation of resveratrol on the master transcription factors RUNX2 and PPARγ2

The molecular mechanism underlying osteogenesis induction in EMPs by resveratrol stimulation was investigated further. We first evaluated whether resveratrol induced osteogenic commitment of EMPs by controlling the transcriptional program. Using quantitative RT-PCR analysis, the master osteo-lineage transcription factor RUNX2 and its downstream target gene OCN were analyzed. After treatment with 5 µM of resveratrol for 9 days, expression of RUNX2 and OCN in EMPs was elevated by 2.3- and 10-fold respectively (Fig. 2A). We further investigated the role of resveratrol in regulating RUNX2 gene transcription by subcloning upstream of the human RUNX2 promoter region (−1557/ + 32) into a luciferase reporter construct (named as hRUNX2-luc). As shown in Fig. 2B, resveratrol dose-dependently stimulated the promoter activity of hRUNX2-luc approximately threefold.

thumbnail image

Figure 2. Resveratrol activates the osteogenic RUNX2 gene and inhibits the adipogenic PPARγ2 gene expression in EMPs. (A) Osteogenic gene expression in EMPs after reveratrol treatment (5 µM), assessed by real-time RT-PCR, with fold change of RUNX2 and OCN shown as mean ± SD (ap < 0.05; bp < 0.001 by Student's t test). (B) RUNX2 promoter activity in EMPs after resveratrol treatment (0 to 10 µM) for 48 hours, with transient cotransfection of RUNX2-luc and pCMV-β-gal. Luciferase activity (mean ± SD) was determined and normalized to the corresponding β-galactosidase activity (ap < 0.01; bp < 0.001 by Student's t test). (C) PPARγ2 and leptin gene expression of EMPs after reveratrol treatment (10 µM) under AM stimulation for 1 day, with further quantification by real-time RT-PCR (ap < 0.01; bp < 0.05; cp < 0.0001 by Student's t test), gel images of RT-PCR of PPARγ2 and leptin genes are shown; with fold change to β-actin determined by band densities; PPARγ2 fold change in CM, 1.0; AM, 1.8 ± 0.07 (ap < 0.005); AM + RSV, 1.4 ± 0.12 (bp < 0.05); leptin fold-change in CM, 1.0; AM, 1.7 ± 0.05 (ap < 0.005); AM + RSV, 1.2 ± 0.01 (cp < 0.01); and (D) inhibitory effect of resveratrol on PPARγ2 protein expression of EMPs for 3 days was assessed by Western blotting, with fold change relative to α-tubulin levels determined by band densities; fold change in CM, 1.0; AM, 2.0 ± 0.14 (ap < 0.01); and AM + RSV, 0.9 ± 0.07 (ap < 0.01).

Download figure to PowerPoint

The effect of resveratrol on gene expression of the master adipocytic lineage transcription factor PPARγ2 also was examined by quantitative RT-PCR. Stimulating EMPs with AM for 1 day increased PPARγ2 mRNA expression by 1.7-fold, which was significantly inhibited by resveratrol treatment (10 µM; Fig. 2C, upper panel, bp < 0.05). Under the same treatment, AM-induced PPARγ2 downstream target gene leptin also was suppressed with resveratrol treatment (Fig. 2C, upper panel, bp < 0.05). These results also were demonstrated by gel image analysis of RT-PCR (Fig. 2C, lower panel). Furthermore, we demonstrated that treatment with resveratrol (10 µM) for 3 days significantly inhibited AM-induced PPARγ2 protein expression by Western blotting (Fig. 2D, ap < 0.01). Thus we found that resveratrol modulates the commitment of EMPs to an osteogenic lineage by increasing RUNX2 expression while inhibiting AM-induced PPARγ2 expression.

Resveratrol-induced osteogenesis is mediated mainly through SIRT1 rather than ER signaling

To determine the possible signaling pathways involved in the osteogenic effect of resveratrol, specific inhibitors of SIRT1 and ER signaling were used. ALP activity and RUNX2 promoter activity were chosen as indicators of osteogenic induction. As shown in Fig. 3A, the specific SIRT1 inhibitor sirtinol dose-dependently suppressed resveratrol-induced ALP activity, completely blocking resveratrol's effects when sirtinol (≥30 µM) is simultaneously added but without increasing cytotoxicity. However, the ER antagonist ICI182780 showed only moderate inhibitory effects even up to a concentration of 20 µM. Furthermore, the effects of sirtinol and ICI182780 on RUNX2 promoter activity also were examined. Only sirtinol (30 µM) exhibited inhibitory effect on resveratrol-induced RUNX2 promoter activity (ap < 0.05), whereas ICI182780 (10 µM) had no effect at all (Fig. 3B). These results suggested that resveratrol-induced osteogenesis in EMPs was mediated mainly through SIRT1 rather than ER signaling.

thumbnail image

Figure 3. SIRT1 signaling is involved mainly in resveratrol-induced osteogenesis in EMPs. (A) ALP activity of EMPs after treatment with either the SIRT1 inhibitor sirtinol (20 to 40 µM) or the ER inhibitor ICI182780 (5 to 20 µM) in the presence of resveratrol (5 µM); fold activity was expressed as mean ± SD (ap < 0.05; bp < 0.01 by Student's t test). (B) RUNX2 promoter activity in EMPs after treatment with either sirtinol (30 µM) or ICI182780 (10 µM) in the presence of resveratrol (5 µM); fold-activity was expressed as mean ± SD (ap < 0.05; NS= not significant [p > 0.05] by Student's t-test.

Download figure to PowerPoint

Resveratrol induces FOXO3A-mediated transcriptional activity through SIRT1 activation and enhanced FOXO3A expression

Based on the preceding observations, we then searched for the mechanism behind resveratrol-induced SIRT1-mediated osteogenesis in EMPs. First, we confirmed whether resveratrol indeed augmented SIRT1 activity in vivo. When EMPs were treated with 2 µM of resveratrol for 6 hours, the intracellular deacetylation activity of SIRT1 was increased significantly to 140% compared with untreated control (Fig. 4A, ap < 0.05). However, resveratrol-induced SIRT1 activity was totally abolished on cotreatment with sirtinol at 30 µM and above. Subsequently, we analyzed the effects of resveratrol on the protein levels of SIRT1 and FOXO3A, a SIRT1-regulated transcription factor. By nuclear and cytosolic fractionation, we found that SIRT1 and FOXO3A predominantly localized in the nucleus (Fig. 4B). Treatment of EMPs with 5 µM of resveratrol for 24 hours enhanced nuclear FOXO3A protein expression by 1.4-fold independent of augmented SIRT1 activity because this inductive effect was not inhibited by sirtinol; meanwhile, the nuclear protein level of SIRT1 was not altered under resveratrol treatment (Fig. 4B). This inductive effect of resveratrol on FOXO3A protein level was confirmed again in unfractionated cellular lysates (Fig. 4C). Furthermore, we investigated the nature of resveratrol-induced SIRT1-FOXO3A interaction by a coimmunoprecipitation assay. We found that SIRT1 formed immunocomplex with FOXO3A endogenously (Supplemental Fig. S1). After resveratrol treatment for 24 hours, this SIRT1-FOXO3A complex was increased 1.7-fold, but it was not abolished by sirtinol (Fig. 4D). To assess whether resveratrol-induced SIRT1-FOXO3A complex was transcriptionally active, a FRE reporter assay was performed. Resveratrol enhanced FRE reporter activity threefold, and this enhancement was inhibited by sirtinol cotreatment (Fig. 4E). In addition, the FOXO3A-responsive gene superoxide dismutase 2 (SOD2) was induced by resveratrol using quantitative RT-PCR analysis (Fig. 4F). These results together showed that resveratrol enhanced SIRT1/FOXO3A-mediated transcription activity by increasing FOXO3A protein expression and SIRT1 activation, respectively.

thumbnail image

Figure 4. Resveratrol triggers FOXO3A-mediated transcriptional activity via SIRT1 activation and enhanced FOXO3A expression. (A) Intracellular SIRT1 activity in EMPs after treatment with sirtinol (30 to 100 µM) in the presence of 2 µM of resveratrol for 6 hours (ap < 0.05 by Student's t test). (B) Nuclear protein levels of SIRT1 and FOXO3A in EMPs as determined by Western blotting after treatment with 5 µM of resveratrol alone or combined with 30 µM of sirtinol for 24 hours, with ratio of densities of SIRT1 and FOXO3A bands quantified relative to histone H1 band; fold change of FOXO3A in lane 1, 1.0; lane 2, 1.4 ± 0.12 (ap < 0.05); lane 3, 1.5 ± 0.13 (NS, p > 0.05); and lane 4, 1.1 ± 0.08. (C) Effect of resveratrol alone (5 µM) or combined with sirtinol (30 µM) for 24 hours on FOXO3A protein expression level in EMPs, with ratio of densities of FOXO3A bands quantified relative to α-tubulin; fold change in lane 1, 1.0; lane 2, 1.5 ± 0.09 (ap < 0.05); lane 3, 1.4 ± 0.04 (NS, p > 0.05); and lane 4, 1.1 ± 0.07. (D) Interaction of SIRT1 and FOXO3A after treatment with resveratrol alone (5 µM) or combined with sirtinol (30 µM) for 24 hours, as determined by immunoprecipitation (IP) with anti-SIRT1 antibody overnight. The SIRT1 immunocomplexes were subjected to Western blotting against anti-FOXO3A antibody, with ratio of densities of FOXO3A band quantified relative to SIRT1 band; fold change in lane 1, 1.0; lane 2, 1.7 ± 0.11 (ap < 0.05); lane 3, 1.5 ± 0.5 (NS, p > 0.05); and lane 4, 1.4 ± 0.008. (E) Relative FRE activity with cells cotransfected with FRE-luc and pCMV-β-gal plasmids for 48 hours and then treated with sirtinol (30 µM) in the presence of resveratrol (5 µM) for another 48 hours. Relative FRE-luc activity was expressed as mean ± SD (ap < 0.005; bp < 0.05 by Student's t test). (F) Effect of resveratrol (10 µM for 9 days) on SOD2 gene expression, assessed by real-time RT-PCR with fold change shown as mean ± SD (ap < 0.001 by Student's t test).

Download figure to PowerPoint

Resveratrol-induced SIRT1-FOXO3A complex transactivates RUNX2 promoter activity through binding to the distal FRE site

To further investigate the mechanism of SIRT1/FOXO3A action on osteogenesis, we performed sequence analysis of the upstream 1500 bp of the RUNX2 promoter region to identify FRE binding sites, and one putative FRE site (−1269/−1263) was found. We then assessed whether the resveratrol-induced SIRT1-FOXO3A transcriptional complex can control RUNX2 mRNA expression. Transient overexpression of either SIRT1 or FOXO3A alone induced FRE reporter activity in 293T cells, revealing that SIRT1 and FOXO3A were activators of FRE binding site (Fig. 5A). Similarly, RUNX2 promoter activity also was induced by ectopic overexpression of either SIRT1 or FOXO3A alone (Fig. 5A). When SIRT1 and FOXO3A were overexpressed together, FRE reporter activity was synergistically enhanced 12-fold. However, overexpression of dominant-negative forms of SIRT1-H363Y and FOXO3A-DBM decreased 61% of this maximum activation (Fig. 5B, dp < 0.01). Similar patterns were observed in the case of RUNX2-luc (Fig. 5B, bp < 0.05). These results indicated that SIRT1 and FOXO3A cooperated to activate FOXO3A-mediated RUNX2 gene transcription.

thumbnail image

Figure 5. Resveratrol enhances RUNX2 promoter activity through inducing SIRT1-FOXO3A transcriptional complex recruitment to a novel FRE site (–1269/–1263) in vivo. (A) Inductive effects of SIRT1 and FOXO3A individually on RUNX2-luc and FRE-luc. Mock, SIRT1, or FOXO3A expressing plasmids were cotransfected with either RUNX2-luc or FRE-luc along with pCMV-β-gal into 293T cells; relative RUNX2-luc and FRE-luc activities were expressed as mean ± SD (ap < 0.005; bp < 0.001 by Student's t test). (B) Synergistic effects of SIRT1 and FOXO3A on RUNX2-luc and FRE-luc. SIRT1 and FOXO3A or SIRT1-H363Y and FOXO3A-DBM expressing plasmids were cotransfected with either RUNX2-luc or FRE-luc, along with pCMV-β-gal, into 293T cells; relative RUNX2-luc and FRE-luc activities were expressed as mean ± SD (ap < 0.0005; bp < 0.05; cp < 0.005; dp < 0.01 by Student's t test). (C) Effects of resveratrol on FRE site of the RUNX2 promotor. Wild-type RUNX2-luc (WILD) and FRE-site mutated RUNX2-luc (mFRE [–1269]) were transiently cotransfected with pCMV-β-gal into EMPs and treated with indicated concentrations of resveratrol for 48 hours, with fold activity expressed as mean ± SD (ap < 0.01 by Student's t test). (D) ChIP assay for binding of FOXO3A to the RUNX2 promoter FRE site (−1269/−1263) in EMPs after resveratrol (5 µM) treatment alone or combined with 30 µM of sirtinol for 48 hours. The amount of associated DNA on the FRE site and input DNA were monitored by PCR, with the ratio of densities of the FRE band quantified relative to input DNA control; fold change of control, 1.0; RSV, 2.9 ± 0.33 (ap < 0.05); RSV + sirtinol, 1.1 ± 0.24 (ap < 0.05). (E) Effects of silencing SIRT1 and FOXO3A expression on resveratrol-induced RUNX2 promoter activity. siNT (nontarget siRNA control), siSIRT1, and siFOXO3A were transfected into EMPs to interfere with corresponding gene expression. After 24 hours, RUNX2-luc was transfected and treated with resveratrol (2 µM) for 48 hours; relative RUNX2-luc activities were expressed as mean ± SD (ap < 0.0001; bp < 0.001; cp < 0.01 by Student's t test).

Download figure to PowerPoint

To assess the critical role of FRE in regulation of resveratrol-induced RUNX2 promoter activity, the FRE site 5'-ATAAATA-3' was mutated to 5'-AgAgATA-3' within an mFRE (−1269) RUNX2-luc construct. As shown in Fig. 5C, resveratrol-induced RUNX2 promoter activity was partially and significantly inhibited in the mFRE (−1269) RUNX2-luc (Fig. 5C, ap < 0.01), whereas the basal RUNX2 promoter activity was decreased only 10%. This result demonstrated that the distal FRE site was responsive to resveratrol-induced RUNX2 promoter activation. Also, chromatin immunoprecipitation (ChIP) assay revealed that resveratrol increased the DNA-binding activity of the SIRT1-FOXO3A complex to the FRE site (−1269/−1263) of RUNX2 promoter in vivo, which was blocked by sirtinol treatment (Fig. 5D). To further validate the specificity of the interaction, SIRT1 and FOXO3A gene expression were silenced with siRNA specific for the genes (Supplemental Fig. S2), which led to abolishing resveratrol-induced RUNX2 promoter activity (Fig. 5E). In summary, resveratrol increased SIRT1-FOXO3A complex recruitment to the FRE site so as to transactivate RUNX2 promoter activity in vivo.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

In this study we explored the molecular mechanism of resveratrol on human MSC commitment to the osteogenic lineage rather than the adipogenic lineage. Our results demonstrate a novel action of resveratrol in which osteogenesis of EMPs, one kind of human MSC, was promoted through activating the master osteogenic transcription factor RUNX2 gene transcription via SIRT1 and its interactions with FOXO3A; however, adipogenesis was prevented by resveratrol via reducing the master adipogenic transcription factor PPARγ2 gene expression. This is also the first report to directly link FOXO3A to osteogenesis owing to binding of the SIRT1-FOXO3A complex to the novel FRE site we identified on the RUNX2 promotor.

We first observed that resveratrol could induce osteogenesis of EMPs by enhancing activity of the osteogenic marker ALP and osteogenic phenotype of calcium deposition (Fig. 1C). At the same time, adipogenesis of EMPs was inhibited by resveratrol treatment (Fig. 1D). These results suggest that resveratrol-induced osteogenesis may be due to controlling the lineage-specific transcriptional program on the basis of inducing RUNX2 and suppressing PPARγ2 gene expression (Fig. 2). RUNX2 is the early and master transcription factor initiating the osteogenic lineage transcriptional program, which leads to upregulation of various bone-related genes and markers such as OCN, a late bone-formation marker responsive to construct bone matrix.40, 41 It is possible that inhibition of PPARγ2 gene expression by resveratrol may contribute to augmenting RUNX2-mediated osteogenic transcription indirectly because PPARγ2 has been shown to inhibit RUNX2 transcription activity and osteogenic signaling pathways, including Wnt and transforming growth factor β (TGF-β)/BMP.29, 31 However, this interaction remains to be further investigated in our model. Our current data support the hypothesis that resveratrol-induced osteogenesis is mainly due to enhancing RUNX2 gene expression.

Resveratrol has a number of agonistic and antagonistic effects, one of the most important being that on SIRT1.7 However, the role of SIRT1 in mediating the bone-protective effects of resveratrol has been unclear until now. In this study we show that resveratrol can upregulate RUNX2 gene transcription by directly activating SIRT1/FOXO3A signaling in human MSCs (Fig. 6). On resveratrol stimulation, intracellular SIRT1 activity and FOXO3A protein expression were both elevated, increasing the amount of the active SIRT1-FOXO3A complex. Moreover, we identified in this report that a novel FRE (−1269/−1263) on the RUNX2 promotor was transactivated by SIRT1-FOXO3A complex on stimulation of resveratrol in vivo. Herein, our data demonstrated the novel role of a distal FRE site on the regulation of resveratrol-induced RUNX2 promoter activity. Notably, while resveratrol increased FOXO3A protein expression and complex formation between SIRT1 and FOXO3A, these effects were independent of resveratrol-induced SIRT1 activity because sirtinol failed to reverse them (Fig. 4). This result shows that SIRT1 binds to C-terminal domain of FOXO3A as described previously, which is a different region from the acetylation site and DNA-binding domain (forkhead domain) of FOXO3A,42 suggesting that the deacetylation activity of SIRT1 affected only the DNA-binding and transcription activity of FOXO3A but not formation of the SIRT1-FOXO3A complex.

thumbnail image

Figure 6. Schematic summary of molecular mechanism of resveratrol on human MSC commitment to the osteogenic lineage rather than the adipogenic lineage. Resveratrol committed EMPs to osteogenesis by increasing RUNX2 gene transcription, as well as inhibiting AM-induced PPARγ2 gene expression. On resveratrol stimulation, intracellular deacetylation activity of SIRT1 and FOXO3A protein expression was elevated, resulting in a further increase in the amount of the active SIRT1-FOXO3A transcriptional complex. As a consequence, increased active SIRT1-FOXO3A complex bound to and transactivated the distal FRE site within the RUNX2 promoter, which upregulated RUNX2 expression and initiated the downstream osteogenic program.

Download figure to PowerPoint

SIRT1 has been reported to function as a repressor of the fat regulator PPARγ2, inhibiting expression of PPARγ2 downstream adipogenic genes in adipocytes.30 We suspect that there are dual functions of SIRT1 activated by resveratrol, including directly upregulating RUNX2 gene and its downstream transcriptional programming and inhibiting PPARγ2-mediated adipogenic transcriptional programming, which prevents adipogenesis and indirectly favors hMSC commitment to osteogenic lineage. Further investigation is necessary to answer whether SIRT1 serves as a PPARγ2 repressor on resveratrol stimulation in hMSCs.

In our previous study, resveratrol exhibited bone-protective effects and decreased the risk of breast cancer in vitro and in vivo, and FOXO3A was found to be involved in the tumor-suppressive effect of resveratrol.5 However, whether FOXO3A also contributes to osteogenesis and bone mass homeostasis remains to be clarified. According to some reports, FOXO3A may control proapoptotic signaling in osteoblasts43 and regulate the antagonistic effect of oxidative stress on Wnt3a-induced osteoblast differentiation.44–46 Although these in vitro studies seem to conclude that FOXO3A is a negative regulator in osteoblast differentiation and survival, an in vivo animal study shows that FOXO3A-dependent oxidative defense in osteoblasts is indispensable for bone mass anabolism.47 One reason for the negative results in the in vitro studies may be the use of osteoblasts—which are lineage-committed cells—rather than MSCs. In our study, we demonstrate that FOXO3A functions as a direct and positive regulator of RUNX2 gene transcription, showing that it can contribute to osteogenesis in MSCs. It may be that our use of MSCs more closely reflects in vivo changes in terms of being at an earlier stage than osteoblasts developmentally. Further studies are needed to clarify whether the function of FOXO3A differs between MSCs versus committed osteoblasts and to correlate the in vitro findings with the more biologically relevant in vivo results.

In conclusion, we have found a novel molecular mechanism of resveratrol in mediating transcriptional lineage commitment change to osteogenesis rather than adipogenesis in human MSCs through the SIRT1-FOXO3A axis. In addition, we show for the first time a direct interaction between FOXO3A—a molecular target of SIRT1—and osteogenesis through binding of the SIRT1-FOXO3A complex on a novel distal FRE site of the RUNX2 promotor, which upregulated RUNX2 expression and its downstream gene targets. Overall, our findings contribute to unraveling the molecular mechanism underlying resveratrol-induced osteogenesis of human MSCs.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

We acknowledge Dr ME Greenberg for providing expression plasmids to Addgene, including flag-SIRT1, flag-SIRT1-H363Y, HA-FOXO3A, and HA-FOXO3A-DBM (H212R). This work was supported in part by NSC 99-3111-B-002-005 and NSC 97-2314-B-002-060-MY3 from the Taiwan National Science Council.

Authors' roles: Study design: Men-Luh Yen and Pei-Chi Tseng. Study conduct: Pei-Chi Tseng, Hsiao-Wen Peng, and Chi-Fen Hsieh. Data interpretation: Men-Luh Yen, Pei-Chi Tseng, Hsiao-Wen Peng, and Chi-Fen Hsieh. Data analysis: Men-Luh Yen, Pei-Chi Tseng, Sheng-Mou Hou, Min-Liang Kuo, and Ruey-Jien Chen. Drafting manuscript: Pei-Chi Tseng and Men-Luh Yen. Revising manuscript content: Men-Luh Yen, Pei-Chi Tseng, Sheng-Mou Hou, and Ruey-Jien Chen. Final manuscript approval: Pei-Chi Tseng and Men-Luh Yen.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. Acknowledgements
  9. References
  10. Supporting Information

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

FilenameFormatSizeDescription
jbmr_460_sm_SupplData.ppt509KSupplementary Data

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