Piceatannol, Natural Polyphenolic Stilbene, Inhibits Adipogenesis via Modulation of Mitotic Clonal Expansion and Insulin Receptor-dependent Insulin Signaling in Early Phase of Differentiation*

Background: Adipogenesis contributes to the increase in adipose tissue mass. Results: Preadipocytes treated with piceatannol showed reduced adipogenesis with impairment of the early cell cycle progress and insulin-signaling pathway. Conclusion: The anti-adipogenic function of piceatannol is through inhibition of mitotic clonal expansion and insulin receptor activity in the early phase of adipogenesis. Significance: Piceatannol is a novel anti-adipogenic compound that could modulate development of adipose tissue. Piceatannol, a natural stilbene, is an analog and a metabolite of resveratrol. Despite a well documented health benefit of resveratrol in intervention of the development of obesity, the role of piceatannol in the development of adipose tissue and related diseases is unknown. Here, we sought to determine the function of piceatannol in adipogenesis and elucidate the underlying mechanism. We show that piceatannol inhibits adipogenesis of 3T3-L1 preadipocytes in a dose-dependent manner at noncytotoxic concentrations. This anti-adipogenic property of piceatannol was largely limited to the early event of adipogenesis. In the early phase of adipogenesis, piceatannol-treated preadipocytes displayed a delayed cell cycle entry into G2/M phase at 24 h after initiation of adipogenesis. Furthermore, the piceatannol-suppressed mitotic clonal expansion was accompanied by reduced activation of the insulin-signaling pathway. Piceatannol dose-dependently inhibited differentiation mixture-induced phosphorylation of insulin receptor (IR)/insulin receptor substrate-1 (IRS-1)/Akt pathway in the early phase of adipogenesis. Moreover, we showed that piceatannol is an inhibitor of IR kinase activity and phosphatidylinositol 3-kinase (PI3K). Our kinetics study of IR further identified a Km value for ATP of 57.8 μm and a Ki value for piceatannol of 28.9 μm. We also showed that piceatannol directly binds to IR and inhibits IR kinase activity in a mixed noncompetitive manner to ATP, through which piceatannol appears to inhibit adipogenesis. Taken together, our study reveals an anti-adipogenic function of piceatannol and highlights IR and its downstream insulin signaling as novel targets for piceatannol in the early phase of adipogenesis.

Obesity is a global health concern, and it is caused by many factors, including overnutrition and lack of physical activity. Excess storage of lipids in adipose tissue is the key feature of obesity, which is associated with type 2 diabetes and cardiovascular disease (1). Both adipocytes hyperplasia and adipocytes hypertrophy are determinant factors for adipose mass increase. Adipogenesis is responsible for adipocyte hyperplasia, and it also contributes to a compensatory replacement of adipocyte loss due to an active adipocyte turnover in adipose tissue in adult humans (2). The molecular and cellular process of adipogenesis has been extensively characterized using preadipocyte clonal cell lines (e.g. 3T3-L1 and 3T3-F442A) (3,4). Adipogenesis requires a concerted transcriptional and cellular program, including growth arrest of confluent preadipocytes, reentry to the cell cycle for an additional two rounds of division, termed mitotic clonal expansion (MCE), 2 and the initiation of transcriptional events in the early and late phases of differentiation (3). Among factors promoting adipogenesis, adipogenic transcription factors such as CCAAT/enhancer-binding protein ␤ (C/EBP␤), peroxisome proliferator-activated receptor ␥ (PPAR␥), and C/EBP␣, and cellular signaling cascades involved in cell cycle and insulin-dependent signaling pathways in the early phase of adipogenesis are known to play critical roles (4,5). In particular, insulin-and insulin-like growth factor-1 (IGF-
Cell Culture and Differentiation-3T3-L1 preadipocytes were obtained from American Type Culture Collection and maintained in DMEM containing 10% (v/v) FCS in a humidified atmosphere of 5% CO 2 at 37°C. The cells were induced to differentiate on reaching 2-day postconfluency (designated as day 0) by supplementation of standard adipogenic mixture, including 5 M dexamethasone, 0.5 mM 3-isobutyl-1-methylxanthine, and 167 nM insulin (DMI) in 10% FBS/DMEM for 2 days. The medium was changed to 10% FBS/DMEM containing insulin on day 2 and to 10% FBS/DMEM on day 4. On day 6, cells were subjected to Oil Red O staining to visualize accumulated lipid droplets in the cells. Intracellular lipid content was quantified by extracting Oil Red O by isopropyl alcohol and measuring the absorbance at 490 nm by a spectrophotometer.
Multimodal Coherent Anti-Stokes Raman Scattering (CARS) Microscope-A multimodal microscope capable of simultaneous CARS and two-photon excitation fluorescence imaging analysis was performed as described previously (26). Briefly, for CARS imaging of lipid droplets in adipocytes, the wave number difference between the pump laser and Stokes laser tuned to 2840 cm Ϫ1 , which matches the Raman shift of the symmetric CH 2 stretch vibration in lipid molecules. Combined beams were focused into the specimen through a 60ϫ water immersion objective with a 1.2 numerical aperture. The forward CARS signal was collected by an air condenser (numerical aperture ϭ 0.55), transmitted through a 600/65-nm bandpass filter and detected by a photomultiplier tube (H7422-40, Hamamatsu, Japan). For and two-photon excitation fluorescence imaging of intracellular fluorescent piceatannol, the backward signal was collected by the objective, spectrally separated from the excitation source, transmitted through a 520/ 40-nm bandpass filter, and detected by a photomultiplier tube mounted at the back port of the microscope. The combined pump and Stokes laser power at the specimen was kept constant at 55 milliwatts. Acquisition time for each image was 1.12 s. Images were analyzed using Fluo-View software (Olympus, PA).
Immunoblot Assay-The cells were cultured and treated as indicated in each experiment and harvested by scraping in cell lysis buffer containing Tris-HCl (100 mM (pH 8.0)), NaCl (100 mM), 0.5% Triton X-100, protease inhibitor mixture, sodium orthovanadate (1 mM), and sodium fluoride (10 mM). Protein concentration was determined protein assay kit (Bio-Rad). Pro-teins were separated on a SDS-polyacrylamide gel and transferred to nitrocellulose membrane. Immunoblot was performed with the respective primary antibodies and horseradish peroxidase-conjugated secondary antibodies (Santa Cruz Biotechnology). Signals were visualized by enhanced chemiluminescence kit (Pierce).
Cell Viability Assay-The cells were treated with DMI and various concentrations of piceatannol for 48 h. Subsequent incubation of the cells with MTT solution (0.5 mg/ml) for 1 h at 37°C allowed formation of a violet precipitate, formazan. It was dissolved in DMSO, and the absorbance was measured at 595 nm on a microplate reader (Beckman-Coulter).
Isolation of Total RNA and Real Time Reverse Transcription (RT)-PCR-Total RNA in the cells was extracted using TRIzol reagent as described in the manufacturer's instructions. 1 g/l isolated total RNA was used to synthesize cDNA by SuperScrip-tII kit, and real time PCR was performed to amplify cDNA. For real time PCR, cDNA was diluted and mixed with SYBR premixed Taq reaction mixture (Applied Biosystems) containing 100 ng/ml PCR primers. Reactions were performed in triplicate for each pair of primers using StepOne real time PCR system (Applied Biosystems). The sequences of primers used in the reactions are as follows: C/EBP␤ (forward, 5Ј-GCA AGA GCC GCG ACA AG-3Ј, and reverse, 5Ј-GGC TCG GGC AGC TGC TT-3Ј); PPAR␥ (forward, 5Ј-CCA CCA ACT TCG GAA TCA GCT-3Ј, and reverse, 5Ј-TTT GTG GAT CCG GCA GTT AAG A-3Ј); C/EBP␣ (forward, 5Ј-GCT GGA GTT GAC CAG TGA CA-3Ј, and reverse, 5Ј-AAA CCA TCC TCT GGG TCT CC-3Ј); ␤-actin (forward, 5Ј-TGA CGG GGT CAC CCA CAC TGT GCC CAT CTA-3Ј, and reverse, 5Ј-CTA GAA GCA TTT GCG GTG GAC GAT GGA GGG-3Ј).
Cell Cycle Analysis-The cells were collected at the indicated time points after stimulation with DMI in the absence or presence of piceatannol. Harvested cells were fixed with 70% ethanol for 2 h on ice, washed twice with PBS, and stained with 40 g/ml propidium iodine solution containing 500 g/ml of RNase A for 30 min at 37°C. Cell cycle analysis was performed using a Cell Lab Quanta SC flow cytometer (Beckman-Coulter) according to the manufacturer's manual.
PI3K Assay-In vitro kinase assays were performed as described previously (23). Briefly, active PI3K protein (100 ng) was incubated with different concentrations of piceatannol or LY294002, a PI3K inhibitor, at 30°C for 10 min. Following the incubation with 20 l of 0.5 mg/ml phosphatidylinositol (Avanti Polar Lipids), the mixture was incubated with the reaction buffer (100 mM HEPES (pH 7.6), 50 mM MgCl 2 , and 250 M ATP containing 10 Ci of [␥-32 P]ATP) for an additional 10 min and finally with the stopping buffer (15 l of 4 N HCl and 130 l of chloroform/methanol, 1:1). A 1% potassium oxalate-coated silica gel plate (Merck), which was previously activated for 1 h at 110°C, was used to spot the lower chloroform phase in the mixture. Thin layer chromatography was used to separate the resulting 32 P-labeled phosphatidylinositol 3-phosphate, and autoradiography was used to visualize the radiolabeled spots.
IR Kinase Assay-Briefly, purified active recombinant human IR with a final concentration of 1.3 nM was incubated with reaction buffer (8 mM MOPS (pH 7.0), 0.2 mM EDTA, 1 mM Na 3 VO 4 , 5 mM sodium-␤-glycerophosphate, 500 M syn-thetic peptide (KKKSPGEYVNIEFG), 10 mM magnesium acetate, and [␥-33 P]ATP). Mg-ATP mixture was added to initiate the reaction. Following a 40-min incubation at room temperature, 3% phosphoric acid solution was added to stop the reaction. 10 l of reaction mixture was spotted onto a P30 filter paper. After washing three times in 75 mM phosphoric acid and once in methanol, it was dried and subjected for scintillation counting. The ATP concentrations used in the assay were 5, 25,45,90,200, and 500 M, and the activity of IR at each ATP concentration was determined at different concentrations of piceatannol. All data were performed in duplicate. Michaelis-Menten constant (K m ) and maximum activity (V max ), the halfmaximal inhibitory concentration (IC 50 ), and the inhibitory constant (K i ) values were determined using GraphPad Prism software.
In Vitro Pulldown Assay-Generation of piceatannol-Sepharose 4B was performed as described previously (23). Briefly, 0.3 g of Sepharose 4B was activated in 1 mM HCl and mixed with 2 mg piceatannol in the coupled solution of 0.1 M NaHCO 3 and 0.5 M NaCl. Following incubation at 4°C overnight, the mixture was transferred to 0.1 M Tris-HCl buffer (pH 8.0) and further incubated at 4°C overnight. The mixture was washed three times with 0.1 M acetated (pH 4.0) buffer and 0.1 M Tris buffer (pH 8.0), respectively, and suspended in PBS. In vitro pulldown assay was performed as described previously (27). Briefly, Histagged active IR protein (0.2 g) was incubated with piceatannol-Sepharose 4B or control Sepharose 4B beads. Following the incubation at 4°C overnight, the beads were washed five times with wash buffer, and protein bound to the beads was pulled down and analyzed by immunoblotting.
ATP and Piceatannol Competition Assay-Active IR protein (0.2 g) was incubated with 100 l of piceatannol-Sepharose 4B or control Sepharose 4B beads in the presence or absence of 10 or 100 M ATP in reaction buffer. After overnight incubation at 4°C, the protein was pulled down and analyzed by immunoblotting.
Molecular Modeling-Insight II (Accelrys Inc, CA) was used for the docking study and structure analysis with the crystal coordinates of IR in complex with ATP (Protein Data Bank accession code 1IR3) and IGF-1 receptor (IGF-1R) in complex with MSC160919A (Protein Data Bank accession code 3LW0).
Statistical Analysis-Data are shown as means Ϯ S.E. Statistical analysis was performed using SAS9.2 software. Significance of treatment effect and interactions were determined with one-way analysis of variance. Dunnett's multiple comparison analysis was used to determine the significant differences between group means. p values lower than 0.05 were regarded as statistically significant.

Piceatannol Inhibits Adipogenesis in 3T3-L1 Preadipocytes-
We first investigated the effect of piceatannol on the differentiation of 3T3-L1 preadipocytes to adipocytes. Post-confluent 3T3-L1 preadipocytes were treated with DMI to initiate differentiation. The medium was supplemented with various concentrations of piceatannol (0 -50 M) for 6 days. Adipogenesis of these cells was assessed by Oil Red O staining of intracellular lipid droplets, CARS microscopy, and immunoblot assay of adi-pogenic transcription factors. Morphologic and quantitative analysis of intracellular lipids by Oil Red O staining showed that piceatannol-treated 3T3-L1 cells displayed a dose-dependent inhibition of adipogenesis (Fig. 1A). The inhibitory effect of piceatannol on adipogenesis was first noted at 25 M, and ϳ80% decrease of lipid accumulation was observed at 50 M piceatannol (Fig. 1B). By adopting CARS microscopy, a labelfree and noninvasive microscopic imaging technique sensitive to lipid-rich molecules both in vitro and in vivo (28 -30), individual lipid droplets accumulated in these cells were visually analyzed. The number and size of lipid droplets accumulated in the piceatannol-treated cells were markedly decreased compared with those in differentiated control adipocytes (Fig. 1C). Consistent with these results, protein levels of adipogenic transcription factors such as PPAR␥ and C/EBP␣ in 3T3-L1 cells treated with piceatannol at concentrations of 25 and 50 M were lower than those in control cells differentiated for 6 days (Fig. 1D).
A recently proposed cytotoxic effect of piceatannol in other cell lines (31) prompted us to test a possibility that piceatannol-inhibited adipogenesis could result from its cytotoxic effect in 3T3-L1 cells. To test this possibility, we examined the effect of piceatannol on the viability of differentiating 3T3-L1 cells for 48 h by performing an MTT assay. Piceatannol (0 -100 M) displayed little effect on the viability of differentiating 3T3-L1 cells with a maximum 20% decrease in cell viability at 100 M (Fig. 1E). Collectively, these results suggest an anti-adipogenic property of piceatannol in vitro without any cytotoxic effects.
Piceatannol Suppresses MCE in the Early Phase of Adipogenesis-In an effort to understand the molecular basis underlying piceatannol-inhibited adipogenesis, we first attempted to identify the key phase during the adipogenic program that is most sensitive to anti-adipogenic function of piceatannol. As Fig. 2A illustrates, we divided the adipogenesis process into early (days 0 -2), intermediate (days 2-4), and late  (days 4 -6) phases (3). Fifty M piceatannol was then given to the differentiating cells at the indicated times of adipogenesis. After 6 days of differentiation, these cells were subjected to a quantification of Oil Red O staining of intracellular lipids. Consistent with the result shown in Fig. 1, treatment 5 resulted in 80% inhibition of adipogenesis (Fig. 2, B and C). Moreover, the cells in treatments 3 and 4 exhibited 65-70% decrease in lipid accumulation to a similar extent as those in treatment 5. However, treatment 6 showed a 20% decrease in adipogenesis (Fig.  2B). In addition, the presence of piceatannol only in days 0 -1 showed ϳ50% inhibition of adipogenesis (Fig. 2D). These results implicate that the inhibitory effect of piceatannol on adipogenesis is primarily attributable to its inhibitory effect during the early phase of adipogenesis. In addition, cells under the treatments 7 and 8, which included late phase (days 4 -6) of adipogenesis, also displayed significantly reduced levels of lipid accumulation compared with control cells, suggesting an additional role of piceatannol in regulating the terminal adipogenesis program presumably through modulation of lipogenesis and/or fat mobilization. To further assess the effect of piceatannol on transcriptional levels of adipogenic transcription factors during adipogenesis, mRNA levels of C/EBP␤, PPAR␥, and C/EBP␣ were measured at day 2 (the early phase) and day 6 (the late phase) of 3T3-L1 adipocytes differentiated in the presence or absence of 50 M piceatannol. Real time PCR analysis revealed dramatically decreased mRNA levels of C/EBP␤, PPAR␥, and C/EBP␣ in piceatannol-treated cells both at days 2 and 6 compared with nontreated control cells (Fig. 2E). The most striking inhibition was in the expression of C/EBP␤, an early adipogenic transcription factor linking the cell cycle progression in the MCE process and the subsequent initiation of transcription of PPAR␥ and C/EBP␣ genes in the early phase of adipogenesis. Accumulating evidence suggests a regulatory role of piceatannol in cell cycle progression in various cell types (20,32,33). Because induction of additional rounds of the cell cycle is a key feature of MCE progress in adipogenesis, we next examined the effect of piceatannol on cell cycle progression of 3T3-L1 preadipocytes during the MCE process. Flow cytometry assay results showed that piceatannol-treated cells displayed a delayed cell cycle progression at both 16 and 24 h after induction of differentiation with DMI (Fig. 3A). Approximately 50% of differentiating 3T3-L1 cells was in the S phase of the cell cycle after 16 h of DMI treatment, whereas only 15% of piceatannol-treated cells was in the S phase, and more than 80% of piceatannol-treated cells was in the G 1 phase of the cell cycle (Fig. 3B). This in turn resulted in 35% of DMI-treated control cells found in G 2 /M phase, whereas only 10% of piceatannol-treated cells were found to be in G 2 /M phase after 24 h of treatment (Fig. 3C). Thus, our results indicate that piceatannol impairs induction of early adipogenic transcription factors and cell cycle programs in the early phase of adipogenesis of 3T3-L1 cells.
Piceatannol Regulates Insulin Signaling Pathway in the Early Phase of Adipogenesis-An immediate induction of the insulinsignaling pathway is one of the key events in the early phase of adipogenesis, and this occurs within several hours after treatment of preadipocytes with differentiating medium containing DMI. Moreover, a recent study suggested a potential inhibitory role of piceatannol on PI3K activity in human aortic smooth muscle cells (23). Thus, we questioned whether the piceatannol effect in adipocytes involved the regulation of the insulin-signaling pathway in the early phase of adipogenesis. We first tested the effect of piceatannol on the insulin-dependent PI3K/ Akt signaling pathway in the early phase of adipogenesis. Adipogenic stimuli strongly induced phosphorylation of Akt after 30 min of stimulation with a sustained level of phosphorylation up to 120 min (Fig. 4A). However, 50 M piceatannol markedly inhibited DMI treatment-induced Akt phosphorylation in 3T3-L1 cells (Fig. 4A). Moreover, almost complete inhibition of Akt phosphorylation in DMI-treated differentiating cells was observed by treatment with 25 M piceatannol (Fig. 4B). PI3K is an upstream target of insulin-induced Akt phosphorylation, and piceatannol has been reported to modulate PI3K activity through its direct binding to PI3K in human aortic smooth muscle cells (23). To investigate if piceatannol-inhibited Akt phosphorylation was due to suppression of PI3K activity and PI3K upstream regulators such as IRS-1 and IR in the early phase of adipogenesis, we next examined the effect of piceatannol on PI3K activity by performing an in vitro kinase assay in the presence of different concentrations of piceatannol and LY294002 (10 M), a PI3K inhibitor. Piceatannol at both 25 and 50 M completely inhibited PI3K kinase activity, whereas 10 M LY294002 resulted in a moderate inhibition of PI3K activity (Fig. 4C). Although DMI treatment resulted in phosphorylation of both IRS-1 and IR in differentiating 3T3-L1 cells, piceatannol at 25 and 50 M showed significant inhibition of phosphorylation of IRS-1 and IR (Fig. 4D). In addition, we also observed that piceatannol inhibited DMI-induced phosphorylation of ERK1/2, an extracellular serine/threonine kinase necessary for initiating adipogenesis (34,35), in the early phase of adipogenesis but to a much lesser extent than phosphorylation of IRS-1 and IR (Fig. 4D). These results suggest that the piceatannol plays an inhibitory role in insulin-signaling pathway in the early phase of adipogenesis largely through modulation of insulininduced IR-dependent phosphorylation signaling.
Piceatannol Inhibits IR Kinase Activity and Physically Interacts with IR in a Mixed Noncompetitive Manner to ATP-To elucidate whether piceatannol modulates IR function, we next examined the direct effect of various concentrations of piceatannol on IR kinase activity. Piceatannol strongly suppressed IR kinase activity in a dose-dependent manner with 75 and 100% inhibition at 50 and 100 M, respectively (Fig. 5A). To further understand how piceatannol inhibits IR activity, we studied the kinetics of IR kinase. IR kinase activity followed saturation kinetics when ATP concentration was varied from 5 to 500 M (Fig. 5B) with estimated K m and V max values for ATP of 57.8 M and 1699 units/mg, respectively. We next measured IR kinase activity at each ATP concentration with respect to various piceatannol concentrations to analyze the binding kinetics of piceatannol to IR. We observed an increase in K m values for ATP with a decrease in V max at higher piceatannol concentrations (Fig. 5C), indicating that piceatannol exhibits mixed inhibition kinetics to IR kinase activity. Piceatannol inhibited IR kinase activity with IC 50 values of 23-64 M at different ATP concentrations (Fig. 5D). Moreover, replotting the IC 50 values obtained from Fig. 5C against ATP concentrations allowed us to estimate the K i value for piceatannol of 28.9 M (Fig. 5E).
We next examined the possibility that the inhibitory effect of piceatannol on IR kinase activity is through a physical interaction between this small molecule and an IR protein, and we conducted a pulldown assay using piceatannol-conjugated Sepharose beads. A pulldown assay using small molecule-conjugated beads has been successfully employed to various studies to elucidate direct binding between small molecules and proteins (36 -38). By adopting this assay, we demonstrated that IR protein was pulled down by piceatannol-Sepharose 4B beads but not by the Sepharose 4B beads alone. The presence of His-tagged IR in the precipitated complex with piceatannol-Sepharose 4B beads was judged by immunoblot analysis using anti-His antibody (Fig. 6A). This result demonstrates the presence of a direct regula-tion of IR by piceatannol through physical interaction between the two molecules. Furthermore, the binding between piceatannol and IR was not altered by the presence of an excess amount of ATP (Fig. 6B). Collectively, our results suggest that piceatannol inhibits the acute activation of the insulin-signaling pathway in the early phase of adipogenesis. This is through inhibition of tyrosine kinase activity of IR via a physical interaction with this protein in a mixed noncompetitive manner to ATP.

DISCUSSION
Piceatannol is a natural polyphenolic stilbene and an analog of resveratrol. Piceatannol is synthesized in plants largely in response to environmental stress and fungal invasion. A rela- tively limited amount of piceatannol (Ͻ0.5 g/g) rather than resveratrol has been found in fruits such as blueberry, deerberry, and grapes. However, a recent study reported that a high concentration of naturally occurring piceatannol can be seen in fruit like passion fruit (Passiflora edulis) (39). Up to 2.2 mg/g of piceatannol exists in freeze-dried passion fruit seed, whereas resveratrol is detected at the concentration of 0.22 mg/g. Moreover, UV irradiation has been demonstrated to produce large amounts of piceatannol in callus of peanut (40). Piceatannol is also found in resveratrol-treated cells when resveratrol is metabolized by cytochrome P450 enzyme CYP1B1 (41). Despite the well documented beneficial function of resveratrol in chemoprevention and lowering the risk of chronic diseases, including obesity (24,25), the role of piceatannol in the aforementioned diseases has not yet been resolved.
In this study, we demonstrate that piceatannol dose-dependently inhibits adipogenesis of 3T3-L1 cells (Fig. 1). We further pinpointed that the early phase of adipogenesis, particularly in the first 24 h of adipogenesis, is largely responsible for the antiadipogenic function of piceatannol (Fig. 2, A-D). The inhibitory role of piceatannol in the early phase of adipogenesis was reflected by the reduced mRNA levels of C/EBP␤, PPAR␥, and C/EBP␣ (Fig. 2D). Piceatannol has been shown to inhibit cell viability and to induce apoptosis in various cell types such as macrophages (42), prostate cancer cells (43), adrenal pheochromocytoma cells (44), lymphoblasts (45), and neutrophils (46). However, our result showed that the viability of the differentiating 3T3-L1 cells was not significantly influenced by piceatannol treatment up to concentration of 100 M (Fig. 1E). This result allowed us to eliminate the possibility that piceatannolinduced cellular cytotoxicity is responsible for its inhibition of adipogenesis. The pro-apoptotic property of piceatannol in other cell types has been shown to be correlated with an induction of cell cycle arrest at the G 2 /M phase (18). Indeed, our cell cycle analysis result mirrors this phenomenon with delayed cell cycle progression of differentiating 3T3-L1 cells by piceatannol treatment after 24 h of DMI treatment. During the first 24 h of adipogenesis, transient activation of insulin-and IGF-1-signaling pathways coincides with expression of C/EBP␤ protein and its phosphorylational activation, which are thought to be important for further induction of adipogenic transcriptional events (11,12). Within several hours of insulin treatment in differentiating 3T3-L1 preadipocytes, an immediate stimulation of IR phosphorylation and its subsequent activation of the PI3K/Akt-signaling pathway were observed in differentiating 3T3-L1 cells (7,47). Insulin-activated MAPK and glycogen syn-  MARCH 30, 2012 • VOLUME 287 • NUMBER 14 thase kinase 3␤ during this time are further known to promote phosphorylation of C/EBP␤ and its transcriptional activity within 16 h of induction of adipogenesis (11,12). Because the critical role of IR in adipose development has been demonstrated in IR-deficient mice (15), it would be reasonable to assume that transient control of IR activity in the early phase of adipogenesis could impact adipocyte differentiation. We then investigated the effect of piceatannol on the IR-dependent PI3K/Akt pathway in the early phase of adipogenesis. Our series of immunoblot analyses showed that piceatannol inhibited insulin-stimulated phosphorylation of IR, IRS-1, and Akt. It should be noted that a substantial inhibition of insulin-induced Akt phosphorylation and PI3K activity was observed when differentiating preadipocytes were exposed to even low concentrations of piceatannol (e.g. 25 M) (Fig. 4, B and C). However, 25 M piceatannol inhibited ϳ50% of lipid accumulation (Fig.  2D) and IR kinase activity (Fig. 5, A and C). In addition, we also observed that piceatannol inhibited DMI-induced phosphorylation of ERK1/2, which is necessary for initiating the early phase of adipogenesis, but to a much lesser extent than phosphorylation of IR, IRS-1, and Akt (Fig. 4D). Although the significance of our study is to identify IR as a novel target for the anti-adipogenic function of piceatannol, it would be possible that the piceatannol-inhibited insulin-signaling pathway in the early phase of adipogenesis is a combined result from the direct inhibitory effect of piceatannol on IR as well as PI3K activities. Supporting this notion, a low concentration of piceatannol (0 -20 M) is able to effectively inhibit Akt phosphorylation and PI3K activity in platelet-derived growth factor-treated human aortic smooth muscle cells. Furthermore, a study of physical binding between piceatannol and PI3K by which piceatannol inhibits PI3K activity has been reported (23). Although piceatannol is suggested to be poorly absorbed in cells, we observed intracellular piceatannol absorbed in 3T3-L1 preadipocytes after 24 h of treatment (data not shown). Thus, we speculate that the piceatannol-inhibited insulin-signaling pathway in differentiating 3T3-L1 cells is a combined result from a direct binding of piceatannol to IR on the cell surface and possibly to intracellular PI3K in preadipocytes, which in turn results in effective suppression of Akt phosphorylation by piceatannol even at lower concentrations. More studies are needed to determine how intracellular piceatannol modulates the insulin-signaling pathway in the early phase of adipogenesis.

Inhibition of Adipogenesis by Piceatannol
To have insights into the IR inhibition mechanism by piceatannol, we carried out kinetics and modeling studies of interac-tion between piceatannol and IR. The K m value of IR for ATP is 57.8 M, and this is in agreement with a previously reported K m value of 58 M for ATP of IR kinase in rat adipocytes (48). Our data clearly demonstrate that piceatannol is a mixed noncompetitive inhibitor to ATP, by which it results in a marked increase in K m values and a decrease in the V max value of IR in relation to increasing concentrations of piceatannol. This result indicates that piceatannol and ATP could interact with IR at different binding sites. Recently, it was reported that the compounds of the indole alkylamine scaffold bind to IGF-1R, sharing a high sequence identity of 84% in the tyrosine kinase domain with IR, in an ATP-noncompetitive manner (49). Based on this finding, we simulated whether piceatannol could be docked to the pocket separate from the ATP-binding site of the kinase domain of IR crystal structure in its inactive conformation (49,50). Fig. 7A shows that piceatannol and ATP-binding sites in IR are likely to be separate but adjacent to each other. The predicted binding mode of piceatannol is somewhat similar to that of MSC160910A (49). The hydroxyl group of piceatannol at the 4Ј position can make a hydrogen bond with the backbone carbonyl group of Val-1060 and the hydroxyl groups at the 3 and 5 positions interact with the side chains of Glu-1047 and Asp-1132 via hydrogen bonds. In addition, several van der Waals interactions exist with the hydrophobic surface formed by Met-1051, Phe-1054, Met-1076, His-1130, and Phe-1128 (Fig. 7A). The putative piceatannol-binding site partially overlaps the activation loop site of the kinase domain of IR in the active conformation (Fig. 7B). Collectively, these results elucidate IR as an important target for the anti-adipogenic property of piceatannol. It should be also noted that our study also provides evidence of a potential inhibitory role of piceatannol in the late phase of adipogenesis (Fig. 2C). The presence of piceatannol from days 4 to 6 of adipogenesis resulted in a significant suppression of lipid droplet accumulation in 3T3-L1 cells. Presumably, its inhibitory role in the late phase of adipogenesis is likely to be due to modulation of either lipogenesis or lipolysis. Although little is known about the role of piceatannol in lipid metabolism, resveratrol has been shown to inhibit total fatty acid and triglyceride synthesis in adipocytes (51,52). Current evidence also suggests a role for resveratrol in stimulating lipolysis and fatty acid oxidation in adipocytes (53)(54)(55). Thus, it will be of interest to determine the role of piceatannol in lipid synthesis and hydrolysis in adipose tissue.
Piceatannol is best known as a potent inhibitor of Syk. Syk is a 72-kDa nonreceptor tyrosine kinase with an N-terminal tandem pair of Src homology 2 domains and a C-terminal catalytic domain, which mediates immunoreceptor signaling events involved in proliferation, differentiation, and phagocytosis of hematopoietic cells (56). It is also reported to suppress interferon ␣-induced tyrosine phosphorylation of signal transducers and activator of transcription (STAT)3 and STAT5 (57), and other serine/threonine protein kinases such as cyclic AMP-dependent protein kinase, phospholipid-dependent protein kinase C, and Ca 2ϩ -calmodulin-dependent myosin light chain kinase (58). Interestingly, tyrosine-phosphorylated Syk was found in the early phase of adipogenesis with its maximum level at 18 -24 h after induction of differentiation with dexamethasone and 3-isobutyl-1-methylxanthine in the absence of insulin FIGURE 6. Piceatannol directly binds to IR. A, His-IR protein was incubated with control-Sepharose 4B or Sepharose 4B conjugated with piceatannol (Pic) at 4°C overnight. IR-piceatannol-Sepharose 4B complex was pulled down by centrifugation, and the precipitated IR was analyzed by immunoblot analysis using anti-His antibody. B, IR protein was incubated with control-Sepharose 4B or piceatannol-Sepharose 4B in the presence of indicated concentrations of ATP at 4°C overnight. The IR bound to the piceatannol-Sepharose 4B beads was pulled down by centrifugation and analyzed by immunoblot analysis using anti-His antibody.
in 3T3-L1 cells (59). Moreover, ectopic expression of Syk resulted in enhanced adipogenesis in vitro (59). Given the role of piceatannol in inhibiting diverse target kinase activities, including Syk, it will be of interest to study the requirement of Syk for adipogenesis and the role of piceatannol in Syk function in the early phase of adipogenesis.
Although many dietary small molecules are proposed to have health benefits, poor bioavailability with low solubility in the aqueous system generally limits the efficacy of these in physiological conditions. For instance, a single intravenous dose of 10 mg/kg piceatannol in rats resulted in a decline of plasma piceatannol concentrations from ϳ41 to 1 M in 6 h (60). Similarly, a single i.v. administration of resveratrol (20 mg/kg) to rats showed maximum plasma concentrations of resveratrol and its metabolites in the range from 2 to 13 M (61). However, a number of studies demonstrated that the poor bioavailability of resveratrol can be enhanced by various methods such as encapsulation with cyclodextrin, formulation with piperine, and incorporation of resveratrol into microparticles, liposomes, nanocapsules, and emulsion (62). For example, oral administra-tion of a single high dose of resveratrol (100 mg/kg) in mice with or without piperine (10 mg/kg) supplementation resulted in a dramatic increase in maximum plasma resveratrol concentration from 10 to 154 M (63). Furthermore, a repeated dose of resveratrol up to 5 g/day, which is equivalent to ϳ85 mg/kg for a 60-kg man, has been shown to be safe in humans (64). Because piceatannol is often compared with resveratrol due to their similar chemical structures, these studies underscore the importance of the development of a novel piceatannol formulation that would allow us to overcome its poor solubility and bioavailability in vivo, thereby achieving elevated plasma concentration of piceatannol for exerting its maximum health benefit. However, piceatannol was reported to inhibit its target kinases at K i values of 3-20 M (58, 65, 66). In line with this, our finding of a 50% inhibition of adipogenesis at ϳ30 M piceatannol and a K i value of piceatannol for IR activity at 28.9 M in vitro indicates that these concentrations are possibly achievable in a physiological condition when piceatannol is administrated with improved formulation and/or structural modification. Piceatannol (atomic color with green carbons) is overlaid on the active structure for comparison. C, speculated model for how piceatannol regulates adipogenesis of 3T3-L1 preadipocytes. Piceatannol inhibits DMI-induced early cellular processes during adipogenesis. Piceatannol attenuates mitotic clonal expansion of differentiating cells by delaying the cell cycle progression. Parallel to this, piceatannol suppresses IR-dependent signaling pathway in the early phase of adipogenesis by a direct interaction with IR and inhibition of its kinase activity. Simultaneously, piceatannol also inhibits PI3K activity. Collectively, these result in suppression of Akt phosphorylation and adipogenesis.
In summary, our study suggests a new role of piceatannol in adipogenesis through targeting the early biochemical and cellular events of adipogenesis such as insulin-dependent signaling pathway, MCE, and expression of early adipogenic transcription factors (a suggested mechanism of piceatannolinhibited adipogenesis is summarized as a schematic in Fig. 7C). Although we cannot rule out the possibility that piceatannol has additional effects on adipogenesis, our finding of a direct binding between piceatannol and IR and the inhibitory role of piceatannol in IR tyrosine kinase activity and its subsequent phosphorylation cascade of insulin-signaling pathway reveals a new mechanism by which a natural polyphenolic stilbene modulates adipogenesis.