Resveratrol Protects Mitochondria against Oxidative Stress through AMP-Activated Protein Kinase-Mediated Glycogen Synthase Kinase-3 Inhibition Downstream of Poly(ADP- ribose)polymerase-LKB1 Pathway

Arachidonic acid (AA, a proinflammatory fatty acid) in combination with iron promotes excess reactive oxygen species (ROS) production and exerts a deleterious effect on mitochondria. We have shown previously that activation of AMP-activated protein kinase (AMPK) protects hepatocytes from AA iron-induced apoptosis. Resveratrol, a polyphenol in grapes, has beneficial effects mediated through SIRT1, LKB1, and AMPK. This study investigated the potential of resveratrol to protect against the mitochondrial impairment induced by AA iron and the underlying mechanism for this cytoprotection. Resveratrol treatment inhibited apoptosis, ROS production, and glutathione depletion elicited by AA iron in HepG2 cells. In addition, resveratrol attenuated superoxide generation in mitochondria and inhibited mitochondrial dysfunction induced by AA iron. Overall, AMPK activation by resveratrol contributed to cell survival, as supported by the reversal of its restoration of mitochondrial membrane potential by either overexpression of a dominant-negative mutant of AMPK or compound C treatment. Resveratrol increased inhibitory phosphorylation of glycogen synthase kinase-3 (GSK3 ) downstream of AMPK, which contributed to mitochondrial protection and cell survival. Likewise, small interfering RNA knockdown of LKB1, an upstream kinase of AMPK, reduced the ability of resveratrol to protect cells from mitochondrial dysfunction. Furthermore, this LKB1-dependent mitochondrial protection resulted from resveratrol’s poly(ADP-ribose)polymerase activation, but not SIRT1 activation, as supported by the experiment using 3-aminobenzamide, a poly(ADP-ribose)polymerase inhibitor. Other polyphenols, such as apigenin, genistein, and daidzein, did not activate AMPK or protect mitochondria against AA iron. Thus, resveratrol protects cells from AA iron-induced ROS production and mitochondrial dysfunction through AMPKmediated inhibitory phosphorylation of GSK3 downstream of poly(ADP-ribose)polymerase-LKB1 pathway. Excess oxidative stress causes cell and tissue injury via the modification of membrane phospholipids (Browning and Horton, 2004). The oxidation of fatty acids in phospholipids by reactive oxygen species (ROS) or by the response to proinflammatory cytokines may activate phospholipases (Balboa and Balsinde, 2006). Oxidative modification of fatty acids and phospholipids triggers inflammatory processes and, thus, exerts detrimental effects on cell signaling. In particular, phospholipase A2 promotes the release of arachidonic acid (AA), a -6 polyunsaturated fatty acid, from membrane phospholipids (Balboa and Balsinde, 2006). AA then contributes to further oxidative stress. In addition, AA in the presThis work was supported by the World Class University project funded by the Korean government (Ministry of Education, Science and Technology Development) [Grant R32-2008-000-10098-0]. Article, publication date, and citation information can be found at http://molpharm.aspetjournals.org. doi:10.1124/mol.109.058479. ABBREVIATIONS: ROS, reactive oxygen species; AA, arachidonic acid; 3-AB, 3-aminobenzamide; ACC, acetyl-CoA carboxylase; AICAR, 5-aminoimidazole-4-carboxamide-1-D-ribofuranoside; AMPK, AMP-activated protein kinase; CaMKK, calcium/calmodulin-dependent kinase kinase; DCFH-DA, 2 ,7 -dichlorofluorescein diacetate; GSK3 , glycogen synthase kinase 3 ; MMP, mitochondrial membrane potential; MnTBAP, Mn(III) tetrakis 4-benzoic acid porphyrin; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide; PAR, poly(ADP-ribose)polymer; PARP, poly(ADP-ribose)polymerase; PI, propidium iodide; Rh123, rhodamine 123; SOD, superoxide dismutase; GSH, glutathione; siRNA, small interfering RNA; L-NAME, L-nitro-arginine-methyl-ester; TUNEL, terminal deoxynucleotidyl transferase dUTP nick-end labeling; DMEM, Dulbecco’s modified Eagle’s medium; FBS, fetal bovine serum; PBS, phosphate-buffered saline; FACS, fluorescence-activated cell sorting; DN, dominant negative; KM, kinase mutant; Ad, adenovirus; SB216763, 3-(2,4-dichlorophenyl)-4-(1-methyl-1H-indol-3-yl)-1H-pyrrole-2,5-dione; STO-609, 7-oxo-7H-benzimidazo[2,1-a]benz[de]isoquinoline-3-carboxylic acid; NOS, nitric-oxide synthase; MMP, mitochrondrial permeability. 0026-895X/09/7604-884–895$20.00 MOLECULAR PHARMACOLOGY Vol. 76, No. 4 Copyright © 2009 The American Society for Pharmacology and Experimental Therapeutics 58479/3516989 Mol Pharmacol 76:884–895, 2009 Printed in U.S.A. 884 at A PE T Jornals on N ovem er 7, 2017 m oharm .aspeurnals.org D ow nladed from ence of iron (a catalyst of auto-oxidation) leads cells to produce excess ROS and elicits mitochondrial dysfunction (Shin and Kim, 2009). Therefore, the combinatorial treatment of AA and iron exerts a negative effect on cell viability, and this treatment may be used to study potential cytoprotective agents targeting mitochondria against severe oxidative stress. Several lines of evidence indicate that resveratrol (3,4 ,5trihydroxystilbene), a polyphenolic component in grapes and red wine, has diverse beneficial actions, such as protecting cells and tissues against neurodegeneration, cardiovascular disease, cancer, diabetes, and obesity-related disorders (Baur and Sinclair, 2006), and extending lifespan of organisms (Howitz et al., 2003). This wide range of biological effects might be explained in part by resveratrol’s antioxidant properties, including increases in catalase and superoxide dismutase (SOD) activity (Rubiolo et al., 2008). In addition, the antioxidant capacity of resveratrol might be mediated by the induction of phase II enzymes via nuclear factor-E2-related factor-2 (e.g., NADPH-quinone oxidoreductase) (Rubiolo et al., 2008). The antioxidant effect of resveratrol has been claimed to be due to the presence of hydroxyl phenolic groups in its chemical structure (Leonard et al., 2003); however, direct scavenging activity cannot account for all cytoprotective efficacy because of its low bioavailability and weak ability to scavenge ROS (Leonard et al., 2003; Sale et al., 2004). It is more likely that resveratrol may act through specific cell signaling pathways that lead to activation of the defensive and cytoprotective systems. AMP-activated protein kinase (AMPK, an intracellular sensor of energy status) is activated to reserve cellular energy content and serves as a key regulator of cell survival or death in response to pathological stress (e.g., oxidative stress, endoplasmic reticulum stress, hypoxia, and osmotic stress) (Hayashi et al., 2000; Terai et al., 2005; Shin and Kim, 2009). This regulatory role is supported by increases in cell viability with treatment by the AMPK activators, including 5-aminoimidazole-4-carboxamide-1-D-ribofuranoside (AICAR) (Ido et al., 2002). Previous work showed that, through AMPK activation, dithiolethiones protect hepatocytes from AA iron-induced apoptosis (Shin and Kim, 2009). Although the biological effects of resveratrol have been diversely studied, it is not yet clear whether the cytoprotective effect of resveratrol against ROS is mediated by AMPK activation. Moreover, the potential cytoprotective effect of resveratrol against mitochondrial impairment has not been explored. Studies from several laboratories have shown that the beneficial effects of resveratrol result from the activation of SIRT1, LKB1, and/or AMPK (Howitz et al., 2003; Hou et al., 2008). In mammalian cells, the upstream kinases of AMPK include LKB1, calcium/calmodulin-dependent kinase kinase (CaMKK), and transforming growth factor -activated kinase-1 (Lage et al., 2008). Several signaling pathways including SIRT1, nitric oxide, protein kinase A, and poly(ADPribose)polymerase (PARP) were identified as upstream pathways that regulate LKB1 activity (Alessi et al., 2006; Hou et al., 2008; Huang et al., 2009; Vázquez-Chantada et al., 2009). However, it remains to be elucidated which upstream signaling pathway regulates the beneficial effect of resveratrol under conditions of oxidative stress. In addition, it is unclear what molecule or component downstream of AMPK contributes to mitochondrial protection against oxidative stress. In view of the importance of AMPK in the action of resveratrol, this study investigated whether resveratrol is capable of protecting mitochondria against the severe oxidative stress induced by AA iron and, if so, whether this compound has the ability to prevent apoptosis. Our work demonstrates, for the first time, that resveratrol protects against AA iron-induced oxidative stress through the inhibition of mitochondrial impairment and ROS production: this cytoprotective effect is mediated by AMPK-dependent GSK3 serine phosphorylation downstream from LKB1. Moreover, we revealed that this particular cytoprotective effect of resveratrol against oxidative stress is dependent on PARP activation, which leads to LKB1 activation. Additional work compared the effect of other polyphenolic compounds on AMPK and their efficacy on the mitochondrial dysfunction induced by AA and iron. Materials and Methods Materials. MitoSOX was supplied by Invitrogen (Carlsbad, CA). Anti-procaspase-3, anti-phospho-acetyl-CoA carboxylase (ACC), anti-phospho-AMPK, anti-AMPK, anti-phospho-LKB1, anti-phosphoGSK3 , and anti-GSK3 antibodies were obtained from Cell Signaling Technology (Danvers, MA). Antibodies directed against PARP, Bcl-xl, Bcl-2, and LKB1 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-poly(ADP-ribose)polymer (PAR) antibody was supplied by BD Bioscience (San Jose, CA). Horseradish peroxidase-conjugated goat anti-rabbit and goat anti-mouse IgGs were purchased from Zymed Laboratories (San Francisco, CA). Mn(III) tetrakis 4-benzoic acid porphyrin (MnTBAP), SB216763, L-nitro-arginine-methyl-ester (L-NAME), and compound C were obtained from Calbiochem (Darmstadt, Germany). DeadEnd Colorimetric TUNEL System was supplied from Promega (Madison, WI). Resveratrol, AA, ferric nitrate, nitrilotriacetic acid, AICAR, 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT), rhodamine 123 (Rh123), propidium iodide (PI), 2 ,7 -dichlorofluorescein diacetate (DCFH-DA), anti-actin antibody, trolox, 3-aminobenzamide (3-AB), and other reagents were purchased from Sigma (St. Louis, MO). The solution of iron-nitrilotriacetic acid complex was prepared as described previously (Shin and Kim, 2009). Cell Culture and Treatment. HepG2 cells, a human hepatocytederived cell line, were obtained from American Type Culture Collection (Manassas, VA) and maintained in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal bovine serum (FBS), 50 U/ml penicillin, and 50 g/ml streptomycin at 37°C in humidified atmosphere with 5% CO2. For all experiments, cells (1 10 ) were plated in 10-cm (diameter) plastic dishes for 2 to 3 days (i.e., 80% confluence) and serum-starved for 24 h. Cells were incubated with 10 M AA for 12 h, followed by additional exposure to 5 M iron for the time period indicated under Results or in the figure legends. To assess the effects of resveratrol, the cells were treated with 3 to 60 M resveratrol for 1 h before incubation with AA iron. MTT Assay. To measure cytotoxicity, HepG2 cells were plated at a density of 5 10 cells/well in a 96-well plate. After treatment, viable cells were stained with MTT (0.25 mg/ml, 4 h). The media were then removed, and formazan crystals produced in the wells were dissolved with the addition of 200 l of dimethyl sulfoxide. Absorbance at 540 nm was measured using an enzyme-linked immunosorbent assay microplate reader (Tecan, Research Triangle Park, NC). Cell viability was defined relative to untreated control [i.e., viability (% control) 100 (absorbance of treated sample)/ (absorbance of control)]. TUNEL Assay. TUNEL assay was performed using the DeadEnd Colorimetric TUNEL System, according to the manufacturer’s inPARP-LKB1-Dependent Mitochondrial Protection 885 at A PE T Jornals on N ovem er 7, 2017 m oharm .aspeurnals.org D ow nladed from struction. HepG2 cells were fixed with 10% buffered formalin in PBS at room temperature for 30 min and were permeabilized with 0.2% Triton X-100 for 5 min. After washing with PBS, each sample was incubated with biotinylated nucleotide and terminal deoxynucleotidyltransferase in 100 l of equilibration buffer at 37°C for 1 h. The reaction was stopped by immersing the samples in 2 saline sodium citrate buffer for 15 min. Endogenous peroxidases were blocked by immersing the samples in 0.3% H2O2 for 5 min. The samples were treated with 100 l of horseradish peroxidase-labeled streptavidin solution (1:500) and incubated for 30 min. Finally, the samples were developed using the diaminobenzidine substrate, chromogen, H2O2, and diaminobenzidine for 10 min. The samples were washed and examined under light microscope (200 ). The counting of TUNELpositive cells was repeated three times, and the percentage from each counting was calculated. Immunoblot Analysis. Cell lysates were prepared according to methods published previously (Shin and Kim, 2009). In brief, the cells were centrifuged at 3000g for 3 min and allowed to expand osmotically to the point of lysis after the addition of lysis buffer. Lysates were centrifuged at 10,000g for 10 min to obtain supernatants and were stored at 70°C until use. Immunoblot analysis was performed according to procedures published previously (Shin and Kim, 2009). Protein bands of interest were developed using an enhanced chemiluminescence system (Amersham, Chalfont St. Giles, Buckinghamshire, UK). Equal protein loading was verified by immunoblotting for -actin. Determination of Reduced GSH and Iron. Reduced GSH in the cells was quantified using a commercial GSH determination kit (Oxis International, Portland, OR). In brief, the GSH-400 method was a two-step chemical reaction. The first step led to the formation of substitution products (thioethers) between 4-chloro-1-methyl-7trifluromethyl-quinolinum methylsulfate and all mercaptans present in the sample. The second step included -elimination reaction under alkaline conditions. This reaction was mediated by 30% NaOH, which specifically transformed the substituted product (thioether) obtained with GSH into a chromophoric thione. Analyses of total iron in the cells were performed on an ICP-AES Optima 4300DV (PerkinElmer Life and Analytical Sciences, Waltham, MA). Flow Cytometric Analysis of Mitochondrial Membrane Potential. Mitochondrial membrane potential (MMP) was measured with Rh123, a membrane-permeable cationic fluorescent dye. Cells were treated according to the individual experiment, were stained with 0.05 g/ml Rh123 for 1 h, and were harvested by trypsinization. After washing with PBS containing 1% FBS, cells were stained with 0.25 g of PI. The change in MMP was monitored using a BD FACSCalibur flow cytometer (San Jose, CA). In each analysis, 15,000 events were recorded. Measurement of H2O2 Production. DCFH-DA is a cell-permeable nonfluorescent probe that is cleaved by intracellular esterases and is turned into the fluorescent dichlorofluorescein upon reaction with H2O2 and reactive nitrogen species. H2O2 generation was determined by the concomitant increase in dichlorofluorescein fluorescence. After treatment, cells were stained with 10 M DCFH-DA for 1 h at 37°C. Fluorescence intensity in the cells was measured using FACS. In each analysis, 10,000 events were recorded. Measurement of Mitochondrial ROS. MitoSOX is a live cellpermeable and mitochondrial localizing superoxide indicator. After treatment of HepG2 cells with AA iron, the cells were stained with Fig. 1. Inhibition of AA iron-induced cell death by resveratrol. A, the effect of resveratrol on cell viability. Light micrographs show the morphology of the cells incubated with 3 to 60 M resveratrol for 1 h and continuously treated with 10 M AA for 12 h, followed by exposure to 5 M iron for 6 h (magnification, 200 ). The doseresponse effect of resveratrol on cell viability was assessed using MTT assays. Data represent the mean S.E. of four separate experiments. For graphs in A and B, the statistical significance of differences between treatments and either the vehicle-treated control ( , p 0.01) or cells treated with AA iron (##, p 0.01) was determined. B, TUNEL assay. Cells were treated with 30 M resveratrol for 1 h, followed by the addition of 10 M AA for 12 h, and finally treated with 5 M iron for 6 h. The percentage of TUNEL-positive cells (dark-brown staining) was quantified. Data represent the mean S.E. of four separate experiments. C, immunoblots for apoptotic proteins. Proteins were immunoblotted from the lysates of cells incubated with 30 M resveratrol for 1 h, continuously treated with 10 M AA for 12 h, and then exposed to 5 M iron for 1 h. Equal protein loading was verified by -actin immunoblotting. Results were confirmed by four separate experiments. 886 Shin et al. at A PE T Jornals on N ovem er 7, 2017 m oharm .aspeurnals.org D ow nladed from 5 M MitoSOX for 10 min at 37°C. Fluorescence intensity in the cells was measured using FACS. In each analysis, 10,000 events were recorded. Recombinant Adenoviral DN-AMPK Construct and Plasmid Transfection. A plasmid encoding a dominant-negative mutant of AMPK (D157A; DN-AMPK ) was kindly provided by Dr. J. Ha (Kyunghee University, Seoul, Korea). To generate a recombinant adenovirus expressing DN-AMPK , the construct was subcloned into the attL-containing shuttle plasmid, pENTR-BHRNX (Newgex, Seoul, Korea). Recombinant adenoviral DN-AMPK was constructed and generated by using the pAd/CMV/V5-DEST gateway plasmid. HepG2 cells were infected with adenovirus diluted in DMEM containing 10% FBS at a multiplicity of infection of 50 and incubated for 12 h. After removal of the viral suspension, cells were further incubated with DMEM containing 10% FBS for 2 days and then were treated with the indicated reagent. Adenovirus that expresses LacZ (Ad-LacZ) was used as an infection control. Efficiency of infection was consistently 90% with this method. The construct encoding for a kinase mutant (KM) form of GSK3 was kindly provided by Dr. J. R. Woodgett (Samuel Lunenfeld Research Institute, Toronto, ON, Canada). Cells were transfected with the plasmids by using Lipofectamine 2000 (Invitrogen). The empty plasmid, pCDNA3.1, was used for the mock transfection. siRNA Knockdown. To knock down LKB1, cells were transfected with either an siRNA directed against human LKB1 (Santa Cruz Biotechnology) or a nontargeting control siRNA (100 pmol/ml) by using Lipofectamine 2000 according to the manufacturer’s instructions. After transfection for 24 h, cells were exposed to AA with or without resveratrol for 12 h, followed by treatment with 5 M iron. The resultant samples were analyzed by FACS. LKB1 knockdown was confirmed by immunoblot analysis. Data Analysis. Scanning densitometry was performed with an Image Scan and Analysis System (Alpha Innotech, San Leandro, CA). One-way analysis of variance procedures were used to assess significant differences among treatment groups. For each treatment showing a statistically significant effect, the Newman-Keuls test was used for comparisons of multiple group means. The criterion for statistical significance was set at p 0.05 or 0.01.