Up-regulation of Adiponectin by Resveratrol

The natural polyphenol resveratrol (RSV) displays a wide spectrum of health beneficial activities, yet the precise mechanisms remain to be fully elucidated. Here we show that RSV promotes the multimerization and cellular levels of adiponectin in 3T3-L1 adipocytes. The stimulatory effect of RSV was not affected by knocking out Sirt1, but was diminished by suppressing the expression levels of DsbA-L, a recently identified adiponectin-interactive protein that promotes adiponectin multimerization. Suppression of the Akt signaling pathway resulted in an increase in the expression levels of DsbA-L and adiponectin. On the other hand, knocking out FOXO1 or suppressing the activity or expression levels of the AMP-activated protein kinase (AMPK) down-regulated DsbA-L and adiponectin. The stimulatory effect of RSV on adiponectin and DsbA-L expression was completely diminished in FOXO1-suppressed and AMPK-inactivated 3T3-L1 adipocytes. Taken together, our results demonstrate that RSV promotes adiponectin multimerization in 3T3-L1 adipocytes via a Sirt1-independent mechanism. In addition, we show that the stimulatory effect of RSV is regulated by both the Akt/FOXO1 and the AMPK signaling pathways. Last, we show that DsbA-L plays a critical role in the promoting effect of RSV on adiponectin multimerization and cellular levels.

The natural polyphenol resveratrol (RSV) displays a wide spectrum of health beneficial activities, yet the precise mechanisms remain to be fully elucidated. Here we show that RSV promotes the multimerization and cellular levels of adiponectin in 3T3-L1 adipocytes. The stimulatory effect of RSV was not affected by knocking out Sirt1, but was diminished by suppressing the expression levels of DsbA-L, a recently identified adiponectin-interactive protein that promotes adiponectin multimerization. Suppression of the Akt signaling pathway resulted in an increase in the expression levels of DsbA-L and adiponectin. On the other hand, knocking out FOXO1 or suppressing the activity or expression levels of the AMP-activated protein kinase (AMPK) down-regulated DsbA-L and adiponectin. The stimulatory effect of RSV on adiponectin and DsbA-L expression was completely diminished in FOXO1-suppressed and AMPK-inactivated 3T3-L1 adipocytes. Taken together, our results demonstrate that RSV promotes adiponectin multimerization in 3T3-L1 adipocytes via a Sirt1-independent mechanism. In addition, we show that the stimulatory effect of RSV is regulated by both the Akt/FOXO1 and the AMPK signaling pathways. Last, we show that DsbA-L plays a critical role in the promoting effect of RSV on adiponectin multimerization and cellular levels.
Adiponectin is an adipocyte-derived hormone that plays an important role in the regulation of insulin sensitivity and energy homeostasis. Adiponectin exists in cells and the plasma in three major forms: trimers, hexamers, and the high-molecular weight (HMW) 3 forms, and the HMW form of adiponectin has been shown to be the most bioactive with respect to insulin action (1)(2)(3). Reduction of the HMW form, rather than the total levels of adiponectin, has been shown to be associated with various metabolic disease states (1,(3)(4)(5).
How adiponectin multimerization is regulated is not yet completely understood. We have recently identified an adiponectin interactive protein DsbA-L (disulfide bond-A oxidoreductase-like protein) that promotes adiponectin multimerization in 3T3-L1 adipocytes (6,7). The expression levels of DsbA-L are significantly reduced in obese human subjects and mice (6). On the other hand, the expression levels of DsbA-L are stimulated by the peroxisome proliferatoractivated receptor ␥ agonist rosiglitazone (6). However, it is unknown whether DsbA-L expression could be stimulated by other insulin-sensitizing factors.
Resveratrol (RSV) is a polyphenol with antioxidant, antiinflammation, and anti-insulin resistance properties (8 -11). Some studies suggest that the beneficial health effects of RSV are mainly mediated by the deacetylase Sirt1 (9 -13). However, a number of recent studies demonstrate that RSV could exert its function via Sirt1-independent mechanisms (9, 14 -16). There are some studies showing that RSV has a stimulatory effect on the expression levels (17,18) or the secretion (19) of adiponectin. However, the precise underlying mechanisms remain largely unknown.
In the present report, we show that RSV stimulates adiponectin expression and multimerization in 3T3-L1 adipocytes via a Sirt1-independent mechanism. In addition, we demonstrate that the stimulatory effect of RSV is regulated by both the Akt/FOXO1 and the AMPK signaling pathways. The promoting effect of RSV is greatly diminished in 3T3-L1 adipocytes in which the expression levels of DsbA-L are suppressed by RNAi. Our study thus uncovers a novel mechanism by which RSV exerts its health beneficial effect.
(University of California at San Diego). Sirt1 and PDK1 knock-out MEF cells and their wild-type control cells were described previously (20). To generate AMPK ␣2-suppressed C2C12 myoblasts, a plasmid encoding the AMPK ␣2 short hairpin RNA construct cloned in pSM2 vector (catalogue number RMM1766-96744125, Open Biosystems) was cotransfected with a plasmid encoding the puromycin resistance gene (pBSpac⌬p) into C2C12 cells. Cells containing the AMPK ␣2 shRNA construct were selected with puromycin as described previously (21). A stable cell line expressing the pSM2 vector alone was used as the control. The FOXO1 adenoviruses encoding Foxo1-specific RNAi or scramble RNAi were a generous gift of Dr. Henry Dong (University of Pittsburgh School of Medicine). The adenovirus encoding GFP or mouse DsbA-L and DsbA-L-suppressed cells were generated as described previously (6). The culture and differentiation of 3T3-L1 cells, 3T3 CAR cells, and the culture of C2C12 cells were described in our previous studies (6,22,23). The differentiation of wild-type and Sirt1 knock-out MEF cells into adipocytes were performed as described previously (24).
Size Exclusion Chromatography-A size exclusion chromatographic studies were performed as described in our recent studies (6). The distribution of the adiponectin complex was determined by separating cell lysates using a Superdex 200 10/30 column (GE Healthcare). 200-l fractions were collected and separated by SDS-PAGE followed with Western blot.
Western Blot and Immunoprecipitation-The expression and phosphorylation levels of various proteins in cell lysates or immunoprecipitates were detected by Western blot with specific antibodies as described (6,21). Quantification of the relative change in protein levels (arbitrary unit; expressed as percentage of control protein levels) was performed by analyzing Western blots using the Scion Image Alpha 4.0.3.2 program (Scion Corp., Frederick, MD) and normalized for the amount of protein expression in each experiment. The immunoprecipitation of PGC-1␣ were performed by using an anti-PGC-1␣ specific antibody as described previously (21).
Quantitative Real Time PCR-Adiponectin mRNA was purified from 3T3-L1 adipocytes using the TRIzol reagent (Invitrogen). Total RNAs were reverse transcribed using the SuperScript III First-strand synthesis supermix (Invitrogen). PCR were carried out in duplicates in 96-microtiter plates (Applied Biosystems) in an ABI PRISM 7700 real time PCR apparatus. Each PCR (20 l total volume) contained: 10 l of SYBR Green PCR Master Mix (Applied Biosystems), 2.5 pmol of each primer, and 250 ng of cDNA. The PCR primers for adiponectin were: 5Ј-CCC AAG GGA ACT TGT GCA GGT TGG ATG-3Ј and 5Ј-GTT GGT ATC ATG GTA GAG AAG AAA GCC-3Ј. The PCR primers for DsbA-L were: 5Ј-GCG CCG CGC ATC CTG GAA CTC TTC TAC-3Ј and 5Ј-GTC TGG CAT TCG CAG TTG GGG GTA CAG-3Ј. The mRNA of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was quantified as an endogenous control, using primers: 5Ј-ACC ACA GTC CAT GCC ATC AC-3Ј and 5Ј-TCC ACC ACC CTG TTG CTG TA-3Ј. Reactions without a template or without the reverse transcriptase were used as negative controls.
The Effects of RSV on Adiponectin and DsbA-L Expression in Mouse Adipose Tissue-C57BL6 mice (2 months old) were fed ad libitum custom-prepared laboratory chows (AIN-93G), which were either a normal, maintenance diet, or a diet supplemented with 0.1% resveratrol (w/w), for 14 weeks. Resveratrol (batch number 7044J; MP Biomedical, Solon, OH) was of natural origin, extracted from Canadian Polygonum root, and was 98.01% pure by HPLC analysis. Both chows were prepared by Bio-Serv (Frenchtown, NJ). By analysis, the maintenance diet had the following composition (the nominal content is shown in parentheses): 17.20% (18.00%) protein, 6.99% (7.00%) fat, 4.18% (5.00%) fiber, 2.41% (2.80%) ash, 6.56% (10.00%) moisture, and 62.66% (59.00%) carbohydrate. The vitamin and mineral content was as specified by the AIN-93G diet. The chows were supplied as pellets, which were placed in food hoppers in the cages. Tissues were isolated from overnight fasted mice, homogenized, and used for Western blot to determine the expression levels of DsbA-L and adiponectin.
Data Analysis-Statistical analysis of the data were performed using analysis of variance (for cellular studies) or unpaired t test (for animal studies). Data are presented as mean Ϯ S.E. Statistical significance was set at p values of Ͻ0.05 (*), Ͻ0.01 (**), and p Ͻ 0.05 was considered statistically significant.

RSV Up-regulates DsbA-L and Adiponectin and Promotes Adiponectin Multimerization in 3T3-L1 Adipocytes and in
Vivo-Treating 3T3-L1 adipocytes with RSV led to a timeand dose-dependent increase in the cellular levels of DsbA-L ( Fig. 1, A and B), a recently identified adiponectin interactive protein involved in promoting adiponectin multimerization (6). Concurrently with the increase in DsbA-L expression levels, RSV treatment also led to a great increase in the expression levels of adiponectin ( Fig. 1, A and B). Gel filtration experiments revealed that RSV treatment significantly increased the HMW form of adiponectin (Fig. 1C). Consistent with these findings, RSV treatment significantly enhanced the protein levels of DsbA-L and adiponectin in mouse adipose tissue (Fig. 1D). Because overexpression of DsbA-L has been shown to increase the stability and thus the cellular levels of adiponectin (6,25), these results suggest that enhanced expression levels of DsbA-L may provide a mechanism by which RSV up-regulates adiponectin expression in 3T3-L1 adipocytes.
DsbA-L Is Required for the Regulation of RSV on Adiponectin-To determine whether RSV-promoted multimerization and up-regulation of adiponectin is mediated by DsbA-L, we examined the effect of RSV in 3T3-L1 adipocytes in which the expression levels of DsbA-L are increased by overexpression or suppressed by RNAi. RSV treatment greatly stimulated the cellular levels of DsbA-L and adiponectin in the scramble 3T3-L1 adipocytes ( Fig. 2A). The stimulatory effect of RSV on adiponectin expression, however, was markedly diminished in the DsbA-L-suppressed cells ( Fig. 2A). On the other hand, overexpression of DsbA-L greatly increased adiponectin expression in 3T3-L1 adipocytes and RSV had no further stimulatory effect of RSV in the DsbA-L-overex-pressed cells (Fig. 2B). Gel filtration experiments revealed that RSV promoted adiponectin multimerization in the scramble 3T3-L1 adipocytes and the stimulatory effect of RSV was blocked in DsbA-L-suppressed cells (Fig. 2C). Taken together, these results suggest that DsbA-L plays a critical role in RSVinduced adiponectin multimerization and up-regulation.
RSV Increases the mRNA Level of DsbA-L in 3T3-L1 Adipocytes-To elucidate the mechanism by which RSV promotes DsbA-L and adiponectin up-regulation, we examined the effects of RSV on mRNA levels of DsbA-L and adiponectin in 3T3-L1 adipocytes by real time PCR. We found that RSV significantly increased the mRNA levels of DsbA-L, but had no significant effect on the mRNA levels of adiponectin (Fig. 2D). These results suggest that the promoting effect of RSV on adiponectin up-regulation is predominantly mediated by enhancing the expression levels of DsbA-L.

The Stimulatory Effect of RSV on DsbA-L and Adiponectin
Expression Is Independent of Sirt1-To elucidate the signaling mechanism by which RSV promotes adiponectin multimerization and cellular levels, we asked whether the stimulatory effect of RSV on DsbA-L and adiponectin up-regulation is dependent on Sirt1, given that Sirt1 has been shown to mediate numerous functions of RSV (9,10,13). Inhibition of Sirt1 by sirtinol, as demonstrated by a marked increase in the acetylation levels of PGC-1␣ (Fig. 3A), had no significant effect on RSV-stimulated DsbA-L and adiponectin expression in 3T3-L1 adipocytes (Fig. 3A, the first and second panels). To further confirm the role of Sirt1 in regulation of RSV on DsbA-L and adiponectin expression, we examined the effect of RSV on the expression of DsbA-L and adiponectin in Sirt1deficient adipocytes. RSV treatment enhanced the expression levels of DsbA-L and adiponectin in both wild-type and the   A and B, the expression levels of adiponectin in cell lysates were determined by Western blot using specific antibodies as indicated. C, the cell lysates from DsbA-L-suppressed or scramble cells treated with or without 50 M RSV were separated by gel-filtration chromatography and the distribution of adiponectin complex was determined by Western blot using an antibody to adiponectin. Quantification of the relative abundance of adiponectin oligomers in cell lysates was performed by analyzing Western blots by using National Institutes of Health Scion Image software. The data were shown as changes in the ratio of HMW/total adiponectin. D, the mRNA levels of DsbA-L and adiponectin in 3T3-L1 adipocytes treated with or without 50 M RSV for 24 h were determined by quantitative real time PCR. All data (A, C, and D) are representatives of three independent experiments with similar results and were quantified by the NIH Scion Image program. Data are presented as mean Ϯ S.E. Differences between groups were analyzed using analysis of variance. **, p Ͻ 0.01; *, p Ͻ 0.05. Sirt1-deficient adipocytes (Fig. 3B). Taken together, these results indicate that Sirt1 is dispensable for RSV-stimulated upregulation of DsbA-L and adiponectin in adipocytes.

The Stimulatory Effect of RSV on DsbA-L and Adiponectin Expression Is Partially Mediated by Suppression of the PDK1/ Akt
Signaling Pathway-We have recently found that RSV treatment inhibits both mTORC1 and mTORC2 activity by promoting the interaction between mTOR and DEPTOR in C2C12 cells (20). Because activation of mTORC2 is critical for Akt phosphorylation at Ser 473 , which is critical for full activation of Akt, it is possible that RSV up-regulates DsbA-L and adiponectin by down-regulation of the Akt signaling pathway. To test this possibility, we examined whether RSV inhibits Akt phosphorylation in 3T3-L1 adipocytes. Treating 3T3-L1 adipocytes with RSV remarkably reduced insulin-stimulated Akt phosphorylation at Thr 308 and Ser 473 , concurrently with a decrease in FOXO1 phosphorylation at Ser 256 , an Akt-phosphorylation site (Fig. 4A). To further determine the role of the Akt signaling pathway in the stimulatory effect of RSV, we examined DsbA-L expression levels in wild-type and PDK1-KO MEFs (we were unable to differentiate the PDK1-KO MEFs into adipocytes and thus only examined the effect of RSV on DsbA-L but not adiponectin expression). The expression levels of DsbA-L are greatly enhanced in the PDK1-KO MEFs compared with wild-type cells (Fig. 4B), suggesting that the PDK1/Akt signaling pathway plays a negative role in regulation of DsbA-L expression. Interestingly, RSV treatment led to a further increase in the expression levels of DsbA-L (Fig. 4B), suggesting that, in addition to inhibition of  To further determine the role of the Akt signaling pathway in regulation of DsbA-L and adiponectin expression, we treated 3T3-L1 adipocytes with an Akt-specific inhibitor, AKTi VIII. As expected, AKTi VIII treatment inhibited the phosphorylation of Akt and FOXO1 concurrently with an increase in the expression levels of DsbA-L and adiponectin (Fig. 4C). The expression levels of DsbA-L and adiponectin were further increased in cells co-treated with both AKTi VIII and RSV (Fig. 4C), further demonstrating the presence of an Akt-independent mechanism underlying the stimulatory effect of RSV on DsbA-L and adiponectin expression.
FOXO1 is a direct substrate of Akt and this transcription factor has been shown to be involved in regulation of adiponectin expression (26 -28). To determine whether FOXO1 plays a role in regulating DsbA-L and adiponectin expression, we examined the effect of RSV in scramble and FOXO1-suppressed 3T3-L1 CAR adipocytes. The expression levels of DsbA-L and adiponectin were greatly reduced in FOXO1suppressed 3T3-L1 adipocytes compared with the scramble cells (Fig. 4D). Suppressing the cellular levels of FOXO1 greatly reduced basal and RSV-induced expression levels of DsbA-L and adiponectin (Fig. 4D). Consistent with the view that there is an Akt/FOXO1-independent mechanism by which RSV enhances DsbA-L and adiponectin cellular levels, RSV was still able to promote DsbA-L and adiponectin expression in the FOXO1-supppressed cells, although the stimulatory effect was much smaller compared with that observed in the scramble cells (Fig. 4D).
The AMPK Signaling Pathway Is Involved in RSV-stimulated Up-regulation of DsbA-L and Adiponectin-To elucidate the Akt-independent mechanism underlying RSV-induced DsbA-L and adiponectin up-regulation, we asked whether AMPK, which has been shown to be activated by RSV (14,29,30), plays a role in RSV-stimulated up-regulation of DsbA-L and adiponectin. Activation of AMPK by AICAR markedly enhanced the protein levels of DsbA-L and adiponectin in 3T3-L1 adipocytes (Fig. 5C). However, AICAR co-treatment had no further effect on RSV-stimulated DsbA-L and adiponectin up-regulation (Fig. 5C). Consistent with this data, treating 3T3-L1 adipocytes with Compound C, a specific inhibitor of AMPK (31), had little effect on the basal expression of DsbA-L and adiponectin (Fig. 5A). However, treating 3T3-L1 adipocytes with Compound C significantly suppressed RSV-stimulated expression of DsbA-L and adiponectin (Fig. 5A). Furthermore, RSV stimulated DsbA-L expression in scramble C2C12 cells and the stimulatory effect of RSV was significantly reduced in C2C12 cells in which the expression of AMPK is suppressed by RNAi (Fig. 5B).
To further confirm the necessity of the AMPK and FOXO1 signaling pathways in RSV-stimulated DsbA-L and adiponec- tin up-regulation, we treated scramble and FOXO1-RNAi 3T3-L1 CAR adipocytes with Compound C. Compound C treatment partially suppressed the stimulatory effect of RSV on the expression levels of DsbA-L and adiponectin in the scramble 3T3-L1 adipocytes (Fig. 5D). The promoting effect of RSV was completely diminished in the FOXO1-RNAi cells treated with Compound C (Fig. 5D), demonstrating that activation of FOXO1 and the AMPK signaling pathway are the major mechanisms by which RSV promotes DsbA-L and adiponectin expression.

DISCUSSION
RSV has various health beneficial roles, yet the molecular mechanisms by which RSV exerts its biological function remain to be fully elucidated. Some studies demonstrate that RSV exerts its roles by activation of the NAD ϩ -dependent deacetylase Sirt1 (8,9,25,32). However, a number of recent studies have shown that RSV regulates many of the biological events such cell growth, glucose homeostasis, and protection of the cardiovascular system via a Sirt1-independent mechanism (9,14,(33)(34)(35). In the present study, we show that RSV promotes adiponectin multimerization and up-regulation via a Sirt1-independent mechanism. Because adiponectin oligomer distribution is crucial for its biological functions (3,6) and adiponectin mutants with impaired multimerization are defective in both secretion and are associated with insulin resistance and hypoadiponectinaemia (1, 5), our results thus provide a new mechanism by which RSV exerts its beneficial roles in anti-inflammation, anti-insulin resistance, and cardioprotective functions.
We have shown that RSV stimulates the expression levels of DsbA-L, a recently identified protein that facilitates adiponectin multimerization and stability in cells (6,25). In addition, we found that the stimulatory effects of RSV on adiponectin cellular levels and multimerization are diminished in DsbA-L-suppressed 3T3-L1 adipocytes (Fig. 2, A and C). Taken together with the finding that RSV had no significant effect on the mRNA levels of adiponectin (Fig. 2D), it is conceivable that the effects of RSV on adiponectin multimerization and up-regulation are mainly mediated by up-regulation of DsbA-L. Consistent with this view, overexpresison of DsbA-L has been shown to enhance the cellular levels of adiponectin in 3T3-L1 adipocytes (6) and prevent endoplasmic reticulum stress-induced adiponectin down-regulation (25).
Several recent studies suggest that the PI3K/PDK1/Akt signaling pathway is involved in the action of RSV to regulate metabolism (36,37). However, it is unknown whether this pathway plays a role in RSV-stimulated adiponectin and DsbA-L expression. We have found that inhibition of this signaling pathway increased the cellular levels of DsbA-L and adiponectin in 3T3-L1 adipocytes. In addition, we show that up-regulation of DsbA-L and adiponectin is partially mediated by activation of FOXO1, a transcription factor that has been shown to play a role in regulating adiponectin gene expression (26 -28). However, we found that RSV treatment had no significant effect on the mRNA levels of adiponectin (Fig.  2D). This result is somewhat inconsistent with the finding of Qiao and Shao (38) who found that activation of FOXO1 pro-motes adiponectin transcription. The exact reason for this discrepancy remains unknown but FOXO1 has been found to suppress peroxisome proliferator-activated receptor ␥ gene expression (39). Because peroxisome proliferator-activated receptor ␥ positively regulates adiponectin gene expression and secretion, these findings suggest that the effects of FOXO1 on adiponectin biosynthesis may depend on cell content and upstream signaling events. Alternatively, because treating the cells with RSV significantly increased the mRNA levels of DsbA-L (Fig. 2D), it is possible that the major role of RSV on adiponectin up-regulation is mediated by DsbA-L. Consistent with this, increased cellular levels of DsbA-L have been shown to increase the multimerization and stability of adiponectin (6,25).
In addition to the Akt/FOXO1 signaling pathway, our work demonstrates that the AMPK signaling pathway also plays a key role in the stimulatory effect of RSV on DsbA-L and adiponectin expression. The stimulatory effect of RSV on adiponectin expression is completely suppressed in AMPK-inhibited and FOXO1-suppressed cells (Fig. 5D), implicating that the Akt/FOXO1 and the AMPK signaling pathways are the two major pathways mediating the stimulatory effect of RSV on DsbA-L and adiponectin expression. However, how RSV activates AMPK in 3T3-L1 adipocytes remains unknown, although some recent studies showed that RSV increase AMP levels by inhibition of the mitochondrial F1 ATPase in neuronal cells (35,40). Further studies will be needed to determine whether a similar mechanism is operative in adipocytes.
In summary, we have shown that RSV plays a positive role in regulating adiponectin expression and multimerization in adipocytes via a Sirt1-independent mechanism. In addition, we have demonstrated that the stimulatory effects of RSV are mediated mainly by suppressing the PDK1/Akt signaling pathway, which results in FOXO1 activation, and by activation of the AMPK signaling pathway. Our studies show that RSV treatment had little effect on the mRNA levels of adiponectin, but significantly enhanced the mRNA and protein levels of DsbA-L, suggesting that the promoting effects of RSV on adiponectin multimerization and expression are mainly mediated by up-regulating DsbA-L. Because enhancing adiponectin levels increases resistance to inflammation, insulin resistance, and cardiovascular disorders, the finding that RSV promotes adiponectin expression levels thus provide a novel mechanism by which RSV exerts its health beneficial functions.