Reactive Nitrogen Species Is Required for the Activation of the AMP-activated Protein Kinase by Statin in Vivo^*

Coronary heart disease is a leading cause of mortality in diabetic patients. The increased susceptibility to cardiovascular diseases in type 2 diabetes is due to a constellation of risk factors, including hyperglycemia, insulin resistance, and dyslipidemia. One of the earliest phenomena for vascular injury in diabetes is endothelial dysfunction, characterized by impaired endothelium-dependent relaxation, oxidant stress, and accelerated endothelial cell apoptosis (1, 2).

Statins belong to the class of lipid-lowering drugs that target and inhibit 3-hydroxy-3-methylglutaryl-CoA reductase. Several clinical trials, including the Heart Protection Study (3) and the Collaborative Atorvastatin Diabetes Study (4), have shown significant benefits with low to moderate dose statin therapy in diabetic patients with and/or without overt cardiovascular diseases. A recent study from us (5) suggests that the protection conferred by statin therapy occurs not only as a consequence of its cholesterol-lowering activity but also as a direct result of its effects imparted on endothelium function. In addition, the anti-thrombotic and antiinflammatory effects of statins have been shown to contribute to its overall beneficial activity (6). In humans, improved endothelium functions are one of the earliest observed clinical effects following the initiation of statin treatment (7, 8). Most importantly, statin therapy improves endothelial function by virtue of its antioxidant (8, 9) and anti-inflammatory (8, 9) effects as well as its ability to up-regulate endothelial nitric-oxide synthase (eNOS)² (,10, 11). Thus, it is known that statin inhibits angiotensin II (Ang-II)-induced production of vascular reactive oxygen species (ROS) in spontaneously hypertensive and diabetic rats (12–14). However, the molecular target and biochemical mechanisms whereby statins lower ROS and exert their antiapoptotic effects remain poorly defined.

The AMP-activated protein kinase (AMPK) is a serine/threonine kinase and a member of the Snf1/AMPK protein kinase family found in all eukaryotes (15, 16). AMPK is considered to be a cellular energy sensor, which stimulates ATP-producing catabolic pathways and inhibits ATP-consuming anabolic pathways (17). Although the AMPK pathway is traditionally thought of as a regulator of metabolism, recent studies have demonstrated that AMPK may also act to maintain normal endothelial functions (18). AMPK exerts pleiotropic effects believed to be beneficial to endothelial functions and antiatherogenesis. These effects include, among others, induction of the eNOS/nitric oxide (NO) pathway to increase NO bioavailability; suppression of endothelial ROS production when stimulated by hyperglycemia or high FFA to improve endothelial FFA oxidation and limit lipid accumulation; inhibition of apoptosis and inflammation; and modulation of vascular tone (19–21).

Since many of the metabolic benefits and the endothelial protection conferred by statin are similar to those elicited by up-regulation of AMPK, we hypothesized that AMPK activation may mediate many of the pleiotropic and salutary activities exerted by statins on the cardiovascular system. Consistent with our hypothesis, several recent studies (22, 23) demonstrated that statin can rapidly activate AMPK via increased Thr¹⁷² phosphorylation in vitro, resulting in eNOS activation in cultured cells. Although statin is known to activate AMPK, the mechanism(s) by which statins activate AMPK was not yet defined. To this end, we examined the effects of statin on the kinases upstream of AMPK, specifically evaluating the actions on LKB1 and PKC-ζ. Here we report that statin treatment results in LKB1-dependent AMPK activation through reactive nitrogen species, including peroxynitrite (ONOO^-)-mediated but PKC-ζ-dependent mechanism. AMPK activation is required for the antiapoptosis effect of statin in endothelial cells exposed to high glucose.

EXPERIMENTAL PROCEDURES

Animals—Male eNOS knock-out (eNOS^-/-) and their littermates, C57BL6 mice, 10 weeks of age, were obtained from the Jackson Laboratory (Bar Harbor, ME). Mice were housed in temperature-controlled cages with a 12-h light-dark cycle and given free access to water and normal chows. These mice were randomly divided into control and statin-treated groups. Mice were abdominally injected with statin (5 mg/kg) for 4 h, and the control mice received a 0.9% physiological saline injection. The mice were euthanized with inhaled isoflurane. Mouse hearts, kidneys, livers, and aorta were removed and immediately frozen in liquid nitrogen. The animal protocol was reviewed and approved by the University of Oklahoma Health Science Center Institutional Animal Care and Use Committee.

Materials—Bovine aortic endothelial cells (BAEC), human umbilical vein endothelial cells (HUVEC), and cell culture media were obtained from Cambrex Bio Science Walkersville, Inc. (Walkersville, MD). A549 and HeLa S3 cells, both of which are deficient in LKB1, were from ATCC (Manassas, VA). Antibodies against phospho-AMPKα (Thr¹⁷²), AMPKα, phospho-LKB1 (Ser⁴²⁸), phospho-ACC (Ser⁷⁹), and phospho-PKC-ζ (Thr⁴¹⁰/Thr⁴⁰³) were obtained from Cell Signaling Technology, Inc. (Beverly, MA). Antibody against PKC-ζ was obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). LKB1 antibody was from BIOSOURCE International, Inc. (Camarillo, CA). Simvastatin (statin) was from Calbiochem. STO-609 and BAPTA-AM were from Sigma. LKB1 small interference RNA (siRNA) and PKC-ζ siRNA were from Santa Cruz Biotechnology, and control siRNAs were from Ambion (Austin, TX). 5-Aminoimidazole-4-carboxamide-1-d-ribo-furanoside was obtained from Toronto Chemicals (Toronto, Canada). Other chemicals and organic solvents of the highest grade were obtained from Sigma.

Cell Culture—All culture media were supplemented with penicillin (100 units/ml) and streptomycin (100 μg/ml). A549 and HeLa S3 cells were grown in F-12K medium supplemented with 10% serum. BAEC and HUVEC were maintained in endothelial basal medium with 2% serum and growth factors prior to use. BAEC were serum-deprived overnight prior to experiments. For adenoviral infection experiments, infected BAEC were treated with statin in the presence of normal or high glucose concentrations and for the indicated length of time.

Western Blot and Preparation of Subcellular Fractions—Western blots and preparation of nuclear, cytosol, and membrane fractions were performed as described previously (24). Western blot bands were qualified using the National Institutes of Health ImageJ program (version 1.37v).

Immunocytochemical Staining of LKB1 (25)—The HUVEC were cultured on coverslips and then fixed with 4% paraformaldehyde. After blocking, the HUVEC were incubated with a goat anti-LKB1 antibody (Santa Cruz Biotechnology) overnight. After three washes, the slides were incubated with a fluorescein isothiocyanate-conjugated donkey anti-rabbit and a fluorescein isothiocyanate-conjugated donkey anti-mouse antibody (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA), respectively, at a dilution of 1:150 for 1 h. The slides were then rinsed, counterstained with 4′-6-diamidino-2-phenylindole, mounted in Vectashield™ mounting medium (Vector Laboratories, Burlingame, CA), and viewed on an SLM 510 laser-scanning confocal microscope (Carl Zeiss Meditec, Inc., Jena, Germany).

Assay of Protein Kinase C-ζ Activity—PKC-ζ was immunoprecipitated from untreated (control) or treated cells with an antibody against PKC-ζ overnight at 4 °C in the presence of protein A/G-agarose. PKC-ζ activity present in the immunoprecipitates was determined by its ability to activate pseudosubstrate derivative (50 μm; ERMRPRKRQGSVRRRV) as described previously (24, 25).

AMPK Activity Assay—Total AMPK was immunoprecipitated from 500 μg of protein using an antibody against AMPKα, and AMPK activity was assessed by determining the incorporation of ³²P into the synthetic SAMS peptide as described (24, 25). The difference between the presence and absence of AMPK is calculated as the AMPK activity.

siRNA Gene Silencing of PKC-ζ (25)—Small interfering RNA (siRNA) duplex oligonucleotides used in this study are based on the human cDNAs encoding PKCζ. PKCζ siRNA as well as a nonsilencing control siRNA were obtained from Santa Cruz Biotechnology. The working concentration of siRNA duplexes applied was 100 nm. HUVEC were transfected with PKCζ siRNA or nonspecific control siRNA by using Lipofectamine™ 2000 (Invitrogen) according to the manufacturer's instructions. Transfected cells were starved in serum-free medium for 6 h and then exposed to the indicated concentrations of statin.

In Vivo Gene Silencing of PKC-ζ siRNA (27)—C57BL/6J mice were injected retroorbitally with either mouse-specific PKC-ζ siRNA or control siRNA (1 mg/kg, diluted in 200 μl) every other 3 days for 6 days using in vivo-jetPEI™ from Polyplus Transfection (Illkirch, France) according to the manufacturer's recommendations.

TUNEL Staining (27)—Apoptosis was assessed by TUNEL staining (TMR red) using a kit from Roche Applied Science and following the provided instruction manual. Percentage apoptosis was calculated from the number of TUNEL-positive cells divided by the total number of cells counted.

Detection of Reactive Oxygen Species—Intracellular superoxide anions () were assessed by the dihydroethidine (DHE) fluorescence/HPLC assay with minor modifications (26). Briefly, confluent HUVEC were incubated with DHE (0.5 μm) for 30 min before cell harvest, and then the cells were methanol-extracted. HPLC was used to separate and quantify oxyethidium (product of DHE and ) and ethidium (a product of DHE autoxidation) with a C-18 column (mobile phase: gradient of acetonitrile and 0.1% trifluoroacetic acid). production was determined by the conversion of DHE into oxyethidine.

Direct Mutagenesis of LKB1—Site-directed mutagenesis of LKB1 was performed as described previously (24, 25). Mutations were confirmed by DNA sequencing, and plasmid DNA was extracted on a large scale using Qiagen EndoFree plasmid maxikit (catalogue number 12362) and transfected into HeLa-S3 using the Lipofectamine 2000 kit from Invitrogen (catalogue number 11668-019), according to the instructions provided by the supplier. 24 h after transfection, the cells were treated with statin or vehicle for 1 h. Both LacZ expression vector and untreated cells were used as control.

Adenoviral Infection—BAEC were infected with adenovirus expressing a D194A mutant LKB1 (Ad-D194A), a S428A LKB1 (Ad-S428A), an AMPK dominant negative mutant (AMPK-DN), a PKC-ζ dominant negative mutant, or a constitutively active PKC-ζ mutant (PKC-ζ-WT), as described previously (25). A replication-defective adenoviral vector expressing GFP was used as control. Cells were infected with the adenovirus at a multiplicity of infection of 50 overnight and then washed and incubated in fresh medium for an additional 18–24 h prior to the experiment. Under these conditions, ∼85% of the cells showed positive for GFP.

Statistical Analysis—Results are expressed as mean ± S.E. of at least three independent experiments. Comparisons of the means were performed by the one-way or two-way analysis of variance. A value of p < 0.05 was considered statistically significant.

RESULTS

Statin Increases Phosphorylation of AMPK at Thr¹⁷² and ACC at Ser⁷⁹ in BAEC—Statin has been shown to activate AMPK in cultured endothelial cells (22, 23). To define the mechanism, we first examined whether the Ser⁴²⁸ phosphorylation of LKB1, a well characterized AMPK kinase in endothelial cells (24), was altered by statin. Consistent with previous studies, statin (1–50 μm) caused a dose-dependent increase in AMPK activity in cultured BEAC, as determined by examining phosphorylation of both AMPK (Thr¹⁷²) and ACC (Ser⁷⁹), the downstream AMPK substrate, by immunoblot analysis (Fig. 1) (25). Exposure of BAEC to statin (50 μm) resulted in a 2.6-fold increase in AMPK-Thr¹⁷² phosphorylation (Fig. 1A). AMPK activation by statin was further confirmed by enhanced phosphorylation of ACC at Ser⁷⁹. No change in the expression of endogenous AMPK α protein was observed by immunoblot analysis (Fig. 1A). To corroborate statin-induced AMPK activation, statin-enhanced AMPK activity was examined by [³²P]ATP incorporation into the SAMS peptide (substrate for AMP-activated protein kinase). Consistent with our previous findings, treatment with statin (5 μm, 72 h) significantly increased AMPK activity (6.5 ± 0.7 versus 4.1 ± 0.4 pmol/min·mg protein, statin versus control, n = 6, p < 0.05). AMPK activation by statin was also concentration-dependent. Although the maximal phosphorylation of AMPK by statin occurred at 50 μm (Fig. 1A), the level of AMPK activation by statin at 50 μm was comparable with that of 5-aminoimidazole-4-carboxamide-1-d-ribo-furanoside at 1 mm (data not shown).

Our initial studies suggest that AMPK activation by statin is biphasic. Acute exposure for 45 min was required for statin (50 μm) to activate AMPK. Increased phosphorylation of AMPK and ACC occurred 45 min following the initiation of statin treatment, with peak levels of activated protein occurring at 120 min (Fig. 1B). Chronic exposure of BAEC (>24 h) lowered the effective concentrations of statin to 5–10 μm (data not shown). Compared with control, statin (5 μm, 72 h) caused a drastic increase in phosphorylation of both AMPK (Thr¹⁷²) and ACC (Ser⁷⁹) (data not shown). Collectively, these data confirm that AMPK is activated by statin treatment.

Statin Increases LKB1 Phosphorylation at Ser⁴²⁸ in a Dose-dependent Manner—Previous studies have identified LKB1 as a kinase upstream of AMPK and determined that LKB1 phosphorylation at Ser⁴²⁸ is important for peroxynitrite (ONOO^-)-enhanced (24) or metformin-enhanced (25) AMPK activation in cultured endothelial cells. Since statin treatment activates AMPK, we next examined whether statin would affect LKB1 phosphorylation at Ser⁴²⁸. Statin (10–50 μm) for 2 h of treatment did not alter overall LKB1 levels but did result in significantly increased phosphorylation of LKB1-Ser⁴²⁸ compared with untreated cells (Fig. 1C). Consistent with statin-mediated AMPK phosphorylation, there was a dose-dependent increase in LKB1 phosphorylation at Ser⁴²⁸ in response to statin (Fig. 1C).

Inhibition of Ca²⁺/Calmodulin-dependent Kinase Kinase (CaMKK) Does Not Alter Statin-induced AMPK Activation—Recent studies (28, 29) have suggested that CaMKK-β functions as AMPK kinase under conditions in which intracellular Ca²⁺ increases. We first determined if STO-609, a selective CaMKK-β inhibitor, or BAPTA-AM (20 μm), an intracellular Ca²⁺ chelator, altered calcium-dependent AMPK activation in BAEC. As expected, exposure of BAEC to A23187 (1 μm) significantly enhanced the phosphorylation of AMPK-Thr¹⁷² in BAEC. Preincubation of BAEC with either STO-609 (1 μm) or BAPTA-AM (20 μm), significantly suppressed calcium inophore A23187-enhanced phosphorylation in BAEC (Fig. 2A).

To determine if Ca²⁺ or CaMKK-β is required for statin-enhanced AMPK activation, STO-609 (1 μm) or BAPTA were preincubated with BAEC prior to statin treatment. Neither STO-609 nor BAPTA-AM at the stated concentrations altered the basal or statin-enhanced phosphorylation levels of AMPK at Thr¹⁷² or ACC at Ser⁷⁹ in BAEC (Fig. 2, B–E). These results suggest that CaMKK-β was not required for statin-induced AMPK activation in BAEC.

Transfection of LKB1-expressing Plasmid Is Required for Statin-induced AMPK Activation in A549 Cells, Which Are Otherwise Deficient in LKB1—Since statin results in a parallel increase in both AMPK and LKB1 (Ser⁴²⁸) phosphorylation, we next evaluated whether LKB1 was required for statin-induced AMPK activation. We first examined if statin activated AMPK in A549 or HeLa-S3, two tumor cell lines that have been reported to be deficient in endogenous LKB1 (30, 31). Consistent with previous studies, we were unable to detect LKB1 protein in A549 cells, thereby confirming that A549 cells are deficient in endogenous LKB1 (Fig. 3A). Transfection of LKB1 wild type plasmids into A549 dramatically increased the detection of LKB1 protein in A549 cells (Fig. 3A), thus demonstrating the effective reconstitution of LKB1. In these LKB1-deficient A549 cells, we tested if statin would stimulate AMPK phosphorylation. Empirically, we found that statin did not alter the levels of AMPK-Thr¹⁷² or ACC-Ser⁷⁹ in A549 cells. In contrast, after these A549 cells were transfected with LKB1-expressing plasmid, there was a significant increase in both AMPK (Thr¹⁷²) and ACC (Ser⁷⁹) phosphorylation (Fig. 3A). Similar results were also obtained in LKB1-deficient HeLa-S3 cells (data not shown). Collectively, these data suggest that LKB1 is required for statin-dependent AMPK activation.

Genetic Inhibition of LKB1 Ablates Statin-induced AMPK Activation in Endothelial Cells—To further establish that LKB1 is required for statin-activated AMPK in BAEC, we suppressed LKB1 expression by applying siRNA in BAEC. LKB1 siRNA, but not control siRNA, suppressed the expression of LKB1 by 50% (Fig. 3B). Corroborating our previous results suggesting a requirement of LKB1 for statin-induced AMPK phosphorylation, we found that LKB1 siRNA, but not control siRNA, inhibited statin-dependent phosphorylation of both AMPK at Thr¹⁷² and ACC at Ser⁷⁹ (Fig. 3B). These experiments offer additional support for the notion that LKB1 is required for stain-induced AMPK activation in endothelial cells.

Phosphorylation of LKB1 at Serine 428 Is Required for Statin-induced AMPK Activation in Endothelial Cells—Since statin increased levels of LKB1 phosphorylation at Ser⁴²⁸ (Fig. 1, C and D), we next evaluated whether LKB1 phosphorylation at Ser⁴²⁸ was required for statin-induced AMPK activation. Using site-directed mutagenesis, we developed two LKB1 mutants in which an amino acid essential for LKB1 activation, aspartic acid 194 or serine 428, was mutated to alanine (LKB1-D194A or LKB1-S428A) (24, 25). Adenoviral overexpression of LKB1 dramatically increased LKB1 expression in BAEC (25). Since statin treatment resulted in increased phosphorylation of both AMPK and ACC in BAEC infected with adenovirus encoding GFP, we next investigated whether adenoviral overexpression of the kinase-defective LKB1 mutant (LKB1-D194A) would prevent statin-dependent phosphorylation of AMPK. Although overexpression of the phosphorylation-defective LKB1 mutant (LKB1-S428A) did not affect the basal level of AMPK-Thr¹⁷² in BAEC, it did abolish statin-enhanced phosphorylation of both AMPK-Thr¹⁷² and ACC-Ser⁷⁹ (Fig. 3C). These data suggest that Ser⁴²⁸ phosphorylation of LKB1 was essential for statin-induced AMPK activation in BAEC.

Statin Increases PKC-ζ Phosphorylation at Thr^410/403 in Both a Time- and Dose-dependent Manner—Previously we had determined that PKC-ζ regulates AMPK activity by increasing Ser⁴²⁸ phosphorylation of LKB1, which then stimulates AMPK Thr¹⁷² phosphorylation (24). Therefore, we next determined whether PKC-ζ activation was involved in statin-enhanced LKB1-dependent AMPK activation. Incubation of BAEC with statin (5–50 μm) did not detectably increase total levels of PKC-ζ protein, whereas it significantly increased PKC-ζ phosphorylation at Thr^410/403 within 15 min of treatment (Fig. 4B). The statin-induced phosphorylation of PKC-ζ-Thr^410/403 occurred in a dose-dependent manner (Fig. 4A). In addition, statin significantly (p < 0.01) increased PKC-ζ activity 2.1-fold relative to control groups (Fig. 4C), as determined by examining a PKC-ζ-specific substrate. Collectively, these results indicate that statin activated PKC-ζ prior to activating either LKB1 or AMPK in BAEC and suggest that PKC-ζ may be required for statin-induced activation of LKB1 and/or AMPK.

Inhibition of PKC-ζ Abolishes Statin-enhanced AMPK Activation—To define the role of PKC-ζ in statin-induced LKB1 and AMPK activation, PKC-ζ activity was suppressed by either pharmacological or genetic means. In order to suppress PKC-ζ activity, a dominant negative PKC-ζ mutant (Ad-PKC-ζ-DN) or an siRNA construct against PKC-ζ were adenovirally overexpressed in BAEC. Overexpression of PKC-ζ-DN, but not the control GFP, blocked statin-induced phosphorylation of AMPK at Thr¹⁷² and ACC at Ser⁷⁹ (Fig. 5A). In parallel, overexpression of PKC-ζ-DN, which did not alter basal levels of LKB1 Ser⁴²⁸ phosphorylation, abolished statin-enhanced LKB1 phosphorylation at Ser⁴²⁸ (Fig. 5B). Since overexpression of the dominant negative PKC-ζ mutant inhibited the statin-induced activation of both LKB1 and AMPK (Fig. 5, A and B) and since PKC-ζ activation occurred prior to LKB1 and AMPK (Fig. 4B), as we did in BAEC in response to metformin (25), these results collectively support the prevalent concept that PKC-ζ may function as an upstream kinase for both LKB1 and AMPK.

In order to further confirm the essential role of PKC-ζ, endogenous PKC-ζ in endothelial cells was suppressed by transfection of PKC-ζ-specific siRNA. As expected, BAEC transfected with control siRNA showed no significant change in PKC-ζ levels, whereas PKC-ζ was undetectable in BAEC transfected with PKC-ζ-specific siRNA (Fig. 5C), confirming the reduction of PKC-ζ by siRNA transfection. Confirming the results using cells overexpressing the dominant negative form of PKC-ζ, we found that expression of PKC-ζ-specific siRNA, but not control siRNA, abolished statin-induced phosphorylation of both AMPK at Thr¹⁷² and ACC at Ser⁷⁹, without altering the expression of AMPKα (Fig. 5D).

We next determined if PKC-ζ inhibition alters statin-enhanced AMPK activity. To investigate statin-induced AMPK activity, GFP, PKC-ζ-WT, or PKC-ζ-DN was transiently overexpressed in BAEC. Following mock or statin treatment, AMPK was immunoprecipitated with antibodies raised against AMPKα, which was then assayed for activity by the incorporation of [³²P]ATP into the SAMS peptides. Statin was found to significantly increase AMPK activity in BAEC infected with either GFP or PKC-ζ-WT (Fig. 6A). However, overexpression of PKC-ζ-DN, which did not alter the basal levels of AMPK activity, abolished statin-enhanced AMPK activity (Fig. 6A). These data provide further proof that PKC-ζ is essential for statin-induced AMPK activation in endothelial cells.

Statin-induced LKB1 Translocation Is PKC-ζ-dependent—Previous studies have shown that LKB1 is predominantly localized in the nucleus, whereas AMPK is mainly localized in the cytoplasm (32, 33). LKB1 must be exported from the nucleus to the cytosol to associate with STRAD and MO25 prior to becoming fully active (34, 35). Since commercially available antibodies could not be used for LKB1 immunohistochemical stainings in BAEC, therefore, we examined whether statin altered the subcellular localization of LKB1 in HUVEC by immunohistochemical staining. As expected, LKB1 was found to reside predominantly in the nucleus of nonstimulated HUVEC (Fig. 6B). Statin increased the export of LKB1 from the nucleus to the cytosol in stimulated HUVEC (Fig. 6B). We further found that the PKC-ζ pseudosubstrate, a selective PKC-ζ inhibitor, abolished statin-induced LKB1 translocation to cytosol (Fig. 6B). In order to corroborate the immunofluorescent microscopy data, statin-enhanced LKB1 nuclear export was further examined in subcellular fractions. Immunoblot analysis of cellular fractions confirmed that statin treatment significantly increased cytoplasmic LKB1 protein levels and concurrently and reciprocally decreased nuclear LKB1 protein levels (Fig. 6C). Together, these data suggest that statin triggers LKB1 translocation from nucleus into cytosol by a pathway dependent on PKC-ζ activity.

PKC-ζ-dependent AMPK Activation in Vivo—We next determined if PKC-ζ was required for AMPK activation in mice. C57BL/6J mice were injected retroorbitally with either mouse-specific PKC-ζ siRNA or control siRNA (1 mg/kg, diluted in 200 μl) every other 3 days for 6 days. As shown in Fig. 7A, in vivo transfection of PKC-ζ siRNA but not control siRNA significantly lowered the levels of PKC-ζ in isolated mouse aortas. In addition, statin significantly increased the phosphorylation of both AMPK-Thr¹⁷² and ACC-Ser⁷⁹ in C57BL/6J mice or C57BL/6J mice treated with control siRNA (Fig. 7, B and C). Compared with the levels of phosphorylated AMPK and ACC in mice treated with control siRNA, transfection of PKC-ζ siRNA significantly reduced statin-enhanced phosphorylation of AMPK-Thr¹⁷² (Fig. 7, B and C) and LKB1-Ser⁴²⁸ (Fig. 7D). These results suggest that PKC-ζ was required for statin-enhanced phosphorylation of LKB1 Ser⁴²⁸ and ACC-Ser⁷⁹ in the aortas in vivo.

Activation of AMPK by Statin Is - or ONOO^--dependent—Earlier studies from us and others found that AMPK is activated by an oxidant, such as H₂O₂ or ONOO^-, a potent oxidant formed by and NO. In addition, our earlier study found that ONOO^- is required for metformin-enhanced AMPK activation in vivo. Thus, it was interesting to determine if endogenous ONOO^- was involved in AMPK activation caused by statin. We first investigated if exposure to statin increased in BAEC. As shown in Fig. 8A, exposure of BAEC to statin (10 μm), a concentration in which AMPK was activated, significantly increased the DHE fluorescence. Mito-TEMPOL (10 μm), a mitochondria-targeting superoxide dismutase mimetic, markedly attenuated statin-enhanced (Fig. 8A).

To investigate if statin activated AMPK via ONOO^-, we monitored AMPK-Thr¹⁷² phosphorylation under conditions where ONOO^- was inhibited. Mito-TEMPOL (10 μm), which did not alter basal release in BAEC, markedly attenuated statin-enhanced phosphorylation of both AMPK-Thr¹⁷² and ACC-Ser⁷⁹ (Fig. 8B). Conversely, overexpression of catalase, which increased catalase activity by 2.8-fold, as measured by the reduction of 1% H₂O₂ absorption at 240 nm in BAEC overexpressing catalase, did not alter statin-enhanced phosphorylation of AMPK-Thr¹⁷² and ACC-Ser⁷⁹ (data not shown), indicating that hydrogen peroxide was not involved in statin-enhanced AMPK activation in BAEC.

ONOO^--dependent Activation of AMPK—To further establish if ONOO^- was involved in AMPK activation by statin in vivo, statin was given to eNOS^-/- mice (attenuate ONOO^- by lacking eNOS-derived NO) or to the wild type C57BL/6J mice. Mouse aortas were isolated for assaying 3-nitrotyrosine (3-NT), a footprint for reactive nitrogen species, including ONOO^-. The specificity of 3-NT staining was confirmed by the absence of staining when the antibody was omitted or was diluted in 10 mm 3-NT (data not shown). As shown in Fig. 8C, the proteins positive with the 3-NT antibody were only weakly visible in the aortic homogenates from sham-treated C57BL/6J mice. The levels of 3-NT-positive proteins in aortas of C57BL/6J were markedly increased by statin treatment (Fig. 8C). Compared with weak stainings of 3-NT in the aortas isolated from sham-treated eNOS^-/-, statin did not increase 3-NT-positive proteins in the aortas from eNOS^-/- mice (Fig. 8C), suggesting that NO from eNOS was required for statin-increased ONOO^- in vivo.

We next determined the effects of statin in C57BL/6J and eNOS^-/- mice. As shown in Fig. 8D, administration of statin significantly increased the phosphorylations of both AMPK-Thr¹⁷² and ACC-Ser⁷⁹ in C57BL/6J mice but not in eNOS^-/- mice (Fig. 8D). Consistently, statin increased the phosphorylation of PKC-ζ in the aortas of C57BL/6J wild type mice but not in eNOS^-/- mice (Fig. 8D). Taken together, these results suggested that the ONOO^--PKC-AMPK axis operates in statin-enhanced AMPK activation in vivo.

ONOO^--dependent and ONOO^--independent AMPK Activation in BAEC Exposed to High Glucose—We next determined if NO or ONOO^- derived from eNOS caused a feedback activation of AMPK in BAEC. Since our earlier studies (1, 2) had demonstrated that prolonged exposure of human aortic endothelial cells to high glucose (HG) resulted in ONOO^- formation, we used this model to dissect the contribution of ONOO^- in high glucose-induced AMPK activation. As depicted in Fig. 9A, exposure of BAEC to 30 mm d-glucose (HG) caused a biphasic increase of AMPK phosphorylation, which peaked at 2 and 48 h, respectively. In parallel, 3-nitrotyrosine, a footprint of ONOO^-, was increased in BAEC treated with HG for 48 h. Uric acid (50 μm), a potent scavenger for ONOO^-, significantly suppressed HG-enhanced 3-nitrotyrosine at 48 h (Fig. 9B) but did not alter 3-nitrotyrosine in cells treated with normal glucose or HG for 2 h. These results suggest that HG increased ONOO^- in BAEC at 48 h but not in BAEC exposed to normal glucose or HG for 2 h.

We further determined if uric acid altered high glucose-enhanced AMPK activity. As shown in Fig. 9C, uric acid did not alter AMPK phosphorylation in BAEC exposed to normal glucose but markedly attenuated HG-enhanced AMPK activity at 48 h of exposure. Since the suppression of AMPK activity by uric acid was co-related with its inhibition on ONOO^- formation in BAEC, these results strongly suggest that ONOO^- generated by HG exposure activated AMPK in BAEC at 48 h.

AMPK Activation by Statin Suppresses the Basal and Angiotensin-II-enhanced Superoxide Anions in BAEC—Previous studies have suggested that statin might exert its therapeutic effects by suppressing oxidant stress (5). Therefore, we next evaluated whether AMPK activation is required for the antioxidant effects of statin. Statin treatment was found to significantly lower the basal levels of , assayed by the oxidation of dihydroethidium in BAEC (Fig. 10A). Furthermore, we determined that compound C, a potent AMPK inhibitor, abolished the effect of statin on levels (Fig. 10A), suggesting that statin-dependent suppression of occurs through AMPK activation.

Since compound C might have off-target effects, we next determined if genetic inhibition of AMPK blocked the effects of statin on , levels. We determined that statin reduced the basal levels of in BAEC transfected with GFP; however, overexpression of AMPK-DN abolished statin-induced reduction of in BAEC (Fig. 10B), thus confirming that statin lowers by a mechanism through AMPK signaling.

Since statin is reported to inhibit Ang-II-induced in spontaneously hypertensive rats and diabetic rats (36), we tested if AMPK was required for statin to suppress Ang-II-induced . As expected, Ang-II (100 nm) significantly increased in BAEC (Fig. 10C). Statin significantly suppressed Ang-II-enhanced release of . Importantly, compound C abolished statin-induced reduction of in BAEC, implying that AMPK mediates the inhibition of Ang-II-induced by statins.

AMPK Activation Inhibits Endothelial Apoptosis in Cultured Cells—Previous studies demonstrated that AMPK inhibits apoptosis (18). Since statin is also known to inhibit endothelial apoptosis, we evaluated the role of AMPK in statin-mediated inhibition of apoptosis (37, 38). As determined by TUNEL staining, exposure of BAEC to high glucose (30 mm), but not osmotic controls, significantly increased endothelial cell apoptosis (Fig. 10D). Exposure of BAEC to statin (5 μm for 72 h) resulted in AMPK activation and significantly attentuated high glucose-induced apoptosis (Fig. 10D). Finally, inhibition of AMPK by adenoviral overexpression of AMPK-DN abolished the inhibitory effects of statin on apoptosis (Fig. 10D). These data indicate that AMPK activation by statin prevented high glucose-dependent endothelial cell apoptosis.

DISCUSSION

Recent studies have suggested that AMPK is a therapeutic target for treating diabetes. Oral hypoglycemic agents, such as metformin and rosiglitazone, have been reported to exert their therapeutic effects by activating AMPK (for a review, see Ref. 18). In the present study, we observe that LKB1 Ser⁴²⁸ phosphorylation by atypical PKC-ζ is required for statin-stimulated AMPK activation. The key evidence can be summarized as follows. First, inhibition of LKB1 with D194A and S428A mutants or LKB1 siRNA effectively blocked AMPK activation induced by statin (Fig. 3, B and C). In addition, statin could not activate AMPK in either A549 or HeLa S3 cells that lack LKB1, but wild type LKB1 expression via adenoviral transfection restored the stimulatory effects of statin on AMPK in these cells (Fig. 3A). Second, STO-609, a potent CaMKKβ inhibitor, failed to inhibit statin-induced AMPK activation in BAEC (Fig. 2). Third, statin increased PKC-ζ activity prior to the phosphorylation of both LKB1 and AMPK in BAEC (Fig. 4C). Genetic inhibition of PKC-ζ effectively blocked statin-induced AMPK phosphorylation (Fig. 5A) and AMPK activity (Fig. 6A). Fourth, statin increased export of LKB1 from the nucleus to the cytosol, previously reported to be an important and essential step for LKB1 to activate AMPK (25), and PKC-ζ pseudosubstrate abolished statin-induced LKB1 translocation (Fig. 6, B and C). It is noteworthy that statin-enhanced LKB1 phosphorylation in BAEC occurred within 30 min of treatment, which preceded the statin-induced increase in AMPK phosphorylation. These results suggest that LKB1 is an upstream activator of AMPK or a regulator of cellular signaling in concert with AMPK. Furthermore, the pleiotropic effects of statin thought to be beneficial to endothelial functions and antiatherogenesis might be mediated by PKC-ζ-dependent AMPK activation in endothelial cells.

We previously observed that inhibition of c-Src or phosphatidylinositol 3-kinase activity by pharmacologic agents (PKC-ζ siRNA) or genetic suppression (PKC-ζ-DN) of PKC-ζ abolishes AMPK stimulation by metformin or by ONOO^- in endothelial cells (25). In this study, we have extended these observation by showing that PKC-ζ is also required for statin-enhanced AMPK activation. Furthermore, we have for the first time demonstrated that in vivo inhibition of PKC-ζ was required for statin-enhanced phosphorylation of AMPK and LKB1. These important findings imply that PKC-ζ-LKB1-AMPK is a common pathway for AMPK activation in vivo.

The present study has, for the first time, demonstrated that statin released and the or its derived oxidant, such as ONOO^-, was required for statin-enhanced AMPK activation. The evidence supporting activation of AMPK by the increased formation of ROS is severalfold. First, exposure to statin significantly increased intracellular ROS. In addition, the concentrations of statin (10–50 μm) triggering ROS were similar to those required for the minimally effective concentrations required for phosphorylation and activation of AMPK-Thr¹⁷². This is collaborated by the fact that mito-TEMPOL markedly reduced statin-enhanced phosphorylation of AMPK-Thr¹⁷² and ACC-Ser⁷⁹. Finally, metformin significantly increased AMPK activity in the aortas and hearts of C57BL/6J mice but not those of eNOS^-/-, although eNOS^-/- mice expressed AMPK. Since eNOS^-/- mice did not generate ONOO^- in response to statin, the data strongly suggest that ONOO^- is required for AMPK activation by statin. These results strongly suggest that NO-derived oxidants, such as ONOO^-, might be required for AMPK activation by statin.

Numerous studies from us and others (1, 2, 5) have demonstrated antioxidant effects of statin and AMPK in vivo. This apparently contradictory observation might be similar to ROS in ischemic preconditioning, in which low levels of ROS precondition the tissues to prevent massive production of reactive species in index hypoxia. Our results suggest that statin, like ischemic preconditioning, via the generation of low levels of oxidative stress by statin, “preconditions” the cells or tissues to alleviate oxidant production by activating AMPK. In line with this conclusion, we found that statin suppressed ROS triggered by Ang-II and high glucose. Further, we found that AMPK activation by statin is required for the reduction of reactive nitrogen species (ONOO^-). Our earlier work (39, 40) has also demonstrated that metformin, one of the most used antidiabetic drugs, activates AMPK by increasing ONOO^-, and AMPK activation suppresses oxidant production. Thus, we consider that AMPK might function as a redox sensor, and AMPK activation might reduce oxidant stress by attenuating oxidant stress by other sources or by enhancing antioxidant potentials.

Overwhelming evidence suggests that in humans, improved endothelium functions are one of the earliest observed clinical effects, following the initiation of statin treatment (7, 8). Most importantly, statin therapy improves endothelial function by virtue of its antioxidant (8, 9) and anti-inflammatory (8, 9) effects as well as its ability to up-regulate eNOS (10, 11). In the present study, we have for the first time shown that AMPK might be implicated in the antiapoptosis effects of statin in diabetes. This finding is in line with the clinical studies. Several clinical trials, including the Heart Protection Study (3) and the Collaborative Atorvastatin Diabetes Study (4), have shown significant benefits with low to moderate dose statin therapy in diabetic patients with and/or without overt cardiovascular diseases. In addition, overwhelming evidence suggests a beneficial effect of AMPK in both endothelial functions and the cardiovascular system. Activation of AMPK could therefore explain the beneficial effects of statin on these systems independent of lipid reduction (for a review, see Ref. 18). In an earlier study from us (39), we demonstrated that activation of AMPK by metformin increases NO formation and NO bioactivity in BAEC when stimulated by high glucose. Indeed, exposure of cultured human aortic endothelial cells to 5-aminoimidazole-4-carboxamide-1-d-ribo-furanoside dramatically reduced the markers of oxidant stress and prostacyclin synthase nitration caused by 30 mm glucose.³ Thus, the beneficial effects of AMPK activation are mediated by two mechanisms: 1) increased NO; 2) decreased release promoted by diabetes. In support of our findings, a recent study by Cohen et al. (41) found that S17834, a polyphenol compound that strongly and persistently stimulates AMPK phosphorylation and activity in HepG2 cells, prevented the accelerated atherogenesis typically observed in streptozotocin-induced type 1 diabetic LDLR^-/- mice.

In summary, the results reported here constitute the first direct evidence showing that PKC-ζ is a critical regulator of AMPK activity. Statin stimulates PKC-ζ, and PKC-ζ can regulate AMPK activity by increasing ROS-dependent PKC-ζ activation, which results in LKB1 Ser⁴²⁸ phosphorylation, LKB1 nuclear export, and subsequent AMPK phosphorylation at Thr¹⁷² by LKB1.