Impaired Activation of Phosphatidylinositol 3-Kinase by Leptin Is a Novel Mechanism of Hepatic Leptin Resistance in Diet-induced Obesity*

Obesity is associated with the development of leptin resistance. However, the effects of leptin resistance on leptin-regulated metabolic processes and the biochemical defects that cause leptin resistance are poorly understood. We have addressed in rats the effect of dietinduced obesity (DIO), a situation of elevated tissue lipid levels, on the well described lipid-lowering effect of leptin in liver, an action that is proposed to be important for the prevention of tissue lipotoxicity and insulin resistance. In addition, we have addressed the role of phosphatidylinositol 3-kinase (PI 3-kinase) in mediating the acute effects of leptin on hepatic lipid levels in lean and DIO animals. A 90-min leptin (∼10 ng/ml) perfusion of isolated livers from lean animals decreased triglyceride levels by 42 ± 5% (p = 0.006). However, leptin concentrations ranging from ∼10 to ∼90 ng/ml had no effect on triglyceride levels in livers from DIO animals. The acute lipid-lowering effect of leptin on livers from lean animals was mediated by a PI 3-kinase-dependent mechanism, because wortmannin and LY294002, the PI 3-kinase inhibitors, blocked the effects of leptin on hepatic triglyceride levels and leptin increased liver PI 3-kinase activity by 183 ± 6% (p = 0.003) and insulin receptor substrate 1 tyrosine phosphorylation by 185 ± 30% (p = 0.02) in the absence of PI 3-kinase inhibitors. Contrary to the effects of leptin in lean livers, leptin did not activate PI 3-kinase in livers from DIO rats. These data present evidence for a role for 1) leptin resistance in contributing to the excessive accumulation of tissue lipid in obesity, 2) PI 3-kinase in mediating the acute lipid-lowering effects of leptin in liver, and 3) defective leptin activation of PI 3-kinase as a novel mechanism of leptin resistance.

Human obesity and diet-induced obesity in rodents are associated with the development of leptin resistance (1)(2)(3). However, despite substantial progress, our understanding of the consequences of leptin resistance for the metabolic actions of leptin and the molecular defects causing leptin resistance remains vague. Metabolic abnormalities arising from defective leptin action are most clear in leptin-deficient states caused by leptin gene mutations or lipodystrophy in humans or rodents. These states are characterized by hyperlipidemia, excessive storage of lipid in tissues such as liver and skeletal muscle, and insulin resistance, defects that are markedly improved by administration of leptin (4 -8). Similar effects on lipid metabolism and/or insulin sensitivity are obtained in normal rodents with acute (9,10) or prolonged (11)(12)(13) hyperleptinemia. In the ZDF rat characterized by non-functional leptin receptors and obesity, elevated hepatic lipid levels are reduced by adenovirusmediated expression of functional wild-type leptin receptors in liver (14), whereas administration of leptin to gold-thioglucosetreated mice that lack hypothalamic leptin function decreases hepatic lipogenesis (15). Taken together, these studies demonstrate potent lipid-lowering effects of leptin that are at least partially mediated by direct action of leptin on liver.
Unlike situations of leptin deficiency and leptin receptor mutations, human obesity and diet-induced obesity (DIO) 1 in rodents are characterized by hyperleptinemia and functional leptin receptors, but leptin resistance is also present (1)(2)(3). Abnormalities in lipid metabolism, including excessive storage of lipid in skeletal muscle and liver, are also associated with obesity (16 -19). However, it is unclear whether the lipid-depleting effects of leptin are impaired in obesity, although studies in skeletal muscle from DIO rats demonstrating impaired leptin-stimulated exogenous fatty acid oxidation and lipid hydrolysis are suggestive (20,21). Furthermore, post-receptor signaling events responsible for the effects of leptin on tissue lipid metabolism are poorly understood. Specifically, JAK/ STAT3 (22,23), PI 3-kinase (24 -27), and phosphodiesterase 3B (PDE3B) (25,26) have been implicated in mediating the metabolic actions of leptin, but the role each plays in the lipidlowering effects of leptin and the putative role of defects in these pathways in leptin resistance are unknown.
The current study was undertaken to address each of these issues. Specifically, we determined the effects of DIO in rats on the acute triglyceride-lowering effects of leptin in liver. Additionally, we addressed the role of STAT3, PI 3-kinase, and PDE3B activation in the regulation of hepatic triglyceride levels by leptin and determined whether these mechanisms are defective in obesity. The data demonstrate that DIO rat liver is resistant to the acute lipid-lowering effects of leptin, that leptin acts directly on liver through a PI 3-kinase-dependent mechanism to decrease liver triglyceride levels, and that a mechanism of hepatic leptin resistance is defective leptin-stimulated PI 3-kinase activation. To our knowledge, this is the first study that demonstrates a role for impaired activation of PI 3-kinase by leptin in leptin resistance.

EXPERIMENTAL PROCEDURES
Animal Care and Maintenance-Male Wister rats were purchased from Charles River (Madison, WI) at a weight of 175-200 g. After arrival, rats were maintained on a constant 12-h light/12-h dark cycle with free access to water and ad libitum fed with standard rat chow diet (11% of calories from fat), a high fat diet (TD 96001, 45% of calories from fat, Harlan Teklad, Madison WI) for 3 days or 5-6 weeks or a highly palatable diet (Catalog 9389, 33% of calories from fat, Purina Mills Ltd.) (28) for 3 days. All of the experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Pittsburgh and were in accordance with the National Research Council's Guide for the Care and Use of Laboratory Animals.
Isolated Liver Perfusion System-The isolated liver perfusion system consisted of a peristaltic pump (model 66, Harvard Apparatus, Boston, MA), silicone tubing, a scaled flask serving as a perfusate reservoir in a heated bath, an air bubble trap, and a humidified and a thermostatically controlled (37°C) perfusion chamber (Microplate Incubator, Model 260700, Boekel Scientific, Feasterville, PA). Oxygenation of the perfusate was accomplished by directing 95% O 2 , 5% CO 2 through ϳ3.6 meters of gas-permeable silastic tubing coiled in the perfusate reservoir (0.76-mm inner diameter, 1.65-mm outer diameter, Dow Corning Corporation, Midland, MI). This modified membrane "lung" method provided sufficient oxygen for liver metabolism throughout the experiments (29).
Liver Isolation-Meijer et al. (30) describe the surgery procedures in detail. A rat was anesthetized with pentobarbital sodium (42 mg/kg) after an 18-h fast. The hepatic portal vein was isolated from the hepatic artery and the bile duct and was cannulated by inserting a 16-gauge needle (Protective Acuvance, Johnson & Johnson Medical Division/ Ethicon, Inc., Arlington, TX). The needle was secured by pre-placed sutures and then connected to the perfusion system. Perfusion of the liver with oxygenated buffer began immediately, ensuring that total time of ischemia was Ͻ10 s. The thoracic vena cava was then exposed and severed, and the abdominal aorta and vena cava were ligated above the kidneys. The liver was rapidly excised, and a PE-260 catheter was inserted and secured into the opening of the thoracic vena cava before the liver was transferred onto a liver platform inside the perfusion chamber. Total surgery time was ϳ5 min.
Experimental Design-Livers were perfused with sterile-filtered Krebs-Henseleit buffer (25 mM NaHCO 3 , 118 mM NaCl, 4.7 mM KCl, 1.2 mM MgSO 4 , 1.2 mM NaH 2 PO 4 , 1.2 mM CaCl 2 ) containing 4 mM D-(ϩ)glucose, 10 mM L-lactate, and 0.2% BSA in a re-circulating system at the rate of ϳ2.7 ml min Ϫ1 g Ϫ1 liver. After a 30-min equilibration period, recombinant rat leptin (R&D systems, Inc., Minneapolis, MN) in Krebs-Henseleit buffer or Krebs-Henseleit buffer alone (vehicle) was admixed into the portal vein perfusate using a syringe pump (Model 11, Harvard Apparatus) at a rate of 270 ng/min for 12 min and thereafter at a rate of 35 ng/min for the remainder of the 90-min perfusion period. This administration regime was based on pilot studies with a 400-ml recirculating volume and ensured that the leptin concentration in the perfusate reached a steady state rapidly and remained constant throughout the 90-min perfusion (Fig. 1). In experiments requiring higher perfusate leptin concentrations, the rate of leptin delivery was increased accordingly (3-or 9-fold). For experiments requiring the PI 3-kinase inhibitors wortmannin (final concentration 100 nM, Sigma) or LY294002 (final concentration 10 M, Cell Signaling, Beverly, MA) or the PDE3B inhibitor, cilostamide (final concentration 1 M, Biomol, Plymouth Meeting, PA), livers were perfused for 30 min with the inhibitors prior to a 90-min co-perfusion of the inhibitors with either vehicle or leptin. Liver viability during the experiment was assessed by tissue color, perfusate flow, perfusate pH and potassium concentration, and liver oxygen consumption (31). The pH (ϳ7.4) and potassium (ϳ4.8 mM) values of the perfusate were unchanged throughout the experiment, indicating an absence of hepatocyte damage (31). Oxygen consumption was above 1.2 mol/min/g tissue throughout the experiments, indicating adequate oxygen delivery and extraction for normal hepatic function (31,32). The calculation of the oxygen consumption was based on the difference in oxygen concentration of the perfusate before and after the liver.
Perfusate and Tissue Measurements-Perfusate leptin concentration was measured using a rat-specific RIA kit (Linco Research, St. Charles, MO). Oxygen and potassium concentrations and pH of the perfusate were measured by a gas analyzer (Rapidlab 348, Bayer Corporation, Norwood, MA). Intrahepatic triglycerides were determined as described previously (33). ϳ50 mg of frozen liver tissue was extracted in 1 ml of a chloroform-methanol mixture (2:1). After re-dissolving the lipid pellet with 60 l of tert-butanol and 40 l of Triton X-114-methanol (2:1) mixture, the triglycerides were measured spectrophotometrically (Beckman DU 530) using the GPO-triglyceride kit (Sigma) and Lintrol lipids as standard (Sigma). PI 3-kinase activity was assayed according to a standard protocol. Immunoprecipitation of active PI 3-kinase was achieved by incubating 250 g of liver protein from homogenized tissue with 3 g of ␣-IRS-1 antibody (Upstate Biotechnology, New York, NY) for 2 h followed by the addition of protein A-Sepharose for 1.5 h (Amersham Biosciences). The immune complexes were incubated for 10 min at 22°C with phosphatidylinositol (10 g, Avanti Polar Lipids Inc, Alabaster, AL) in the presence of 50 M [␥-32 P]ATP (5 Ci, PerkinElmer Life Sciences). 32 P-Containing PI(3)P was separated by TLC and was quantitated by scraping the PI(3)P spot from the TLC plate followed by scintillation counting. For quantification of phosphorylation of IRS-1, 500 g of liver protein was immunoprecipitated overnight with ␣-IRS-1 and the immunocomplexes were separated by 7.5% SDS-PAGE and probed with ␣-phosphotyrosine (PY20, 1:1000, CRP Inc., Denver, PA) or ␣-phosphoserine-IRS-1 (Ser-612, 1:1000, Cell Signaling Beverly, MA) antibodies. Leptin receptor protein level was determined by immunoblotting using an ␣-leptin receptor antibody (K-20, 1:3000, Santa Cruz Biotechnology, Santa Cruz, CA) and ␣-goat IgG-HRP (1:5000, Santa Cruz Biotechnology). Leptin receptor mRNA was measured using the ribonuclease protection assay (RPA). Radiolabeled ([␣-32 P]UTP) riboprobes were synthesized from rat leptin receptor cDNA fragments that were specific for the rat long form receptor mRNA (ObRb mRNA, GenBank TM accession number U52966) or that recognized all of the rat leptin receptor isoform mRNAs (ObRtot mRNA, GenBank TM accession number U53144) as described previously (34). A rat-specific ␤-actin riboprobe served as an internal control. RPAs were performed as described previously (34). The antisense riboprobes (1 ϫ 10 5 cpm) were incubated with 10 -20 g of total RNA for 18 h at 45°C in 20 l of hybridization buffer (Ambion, Austin, TX). After hybridization, 200 l of digestion buffer containing a 1:100 dilution of a RNase A/RNase T1 mixture (Ambion) was added and incubated at 37°C for 30 -60 min. RNase digestion was stopped, and the reaction products were precipitated and size-fractionated on a 6% polyacrylamide-denaturing gel. The reaction products were visualized by autoradiography. Autoradiographs were scanned (Hewlett Packard Scanjet 6300C), and the image was quantitated using ImageJ software (National Institutes of Health, Bethesda, MD). Values for ObRb mRNA and ObRtot mRNA levels were normalized to ␤-actin mRNA levels. 5Ј-AMP-activated protein kinase activity (AMPK) was measured by quantifying the incorporation of 32 P into a synthetic substrate peptide-HMRSAMSGLHLVKRR (SAMS) as described previously (35). AMPK was extracted from a liver homogenate using the polyethylene glycol precipitation method (35). Protein concentration was measured colorimetrically. Subsequently, AMPK activity was measured in a 25-l total volume in the presence of a buffer containing 40 mM HEPES-NaOH, pH 7.0, 80 mM NaCl, 8% glycerol, 0.8 mM EDTA, 0.8 mM dithiothreitol, and [␥-32 P]ATP/Mg at 30°C for 5 min.
Statistical Methods-All of the results are expressed as means Ϯ S.E. Statistical significance was determined by Student's unpaired t test using the Systat statistical program (Evanston, IL). Statistical significance was assumed at p Ͻ 0.05.

RESULTS
The Acute Lipid-lowering Effects of Leptin on Liver Are Absent in Diet-induced Obesity-We first compared the effects of leptin on liver triglyceride levels in perfused isolated livers from lean and diet-induced obese animals. In initial experiments, recombinant leptin was admixed into the perfusate at a rate that resulted in a steady-state leptin concentration in the perfusate of ϳ10 ng/ml (Fig. 1), consistent with physiological leptin concentrations in rodents. A 90-min leptin perfusion decreased liver triglyceride levels by 42 Ϯ 5% (p ϭ 0.006) in livers from lean animals compared with vehicle-perfused control livers (Fig. 2). Leptin also increased liver oxygen consumption by 292 Ϯ 21% (p ϭ 0.02) compared with vehicle-perfused controls, demonstrating a leptin-induced increase in oxidative processes (Fig. 3). However, the activity of AMPK, a proposed mediator of leptin effects on fatty acid oxidation, was similar in leptin-perfused compared with saline-perfused livers (736 Ϯ 12 versus 699 Ϯ 30 pmol/min/mg protein, respectively, n ϭ 4). In livers from rats fed for 5-6 weeks on a high fat diet to induce obesity (DIO), hepatic triglycerides levels were increased by 98 Ϯ 3% (p Ͻ 0.0001) compared with livers from lean animals (Fig. 2). Contrary to the triglyceride-lowering effects of leptin on livers from lean animals, leptin had no effects on triglyceride levels in livers from DIO rats (Fig. 2). Furthermore, liver oxygen consumption was similar in vehicle-perfused and leptin-perfused DIO livers (Fig. 3), demonstrating that acute leptin effects on oxidative processes were absent in DIO. Because plasma leptin concentration is elevated in DIO, we hypothesized that triglyceride levels in livers from DIO rats would be reduced by increasing leptin in the perfusate to concentrations associated with obesity. Surprisingly, this was not the case. Livers from DIO animals perfused with a leptin concentration of ϳ30 or ϳ90 ng/ml had similar levels of triglycerides as vehicle-perfused livers (data not shown). These data demonstrate that livers from DIO animals were completely refractory to the acute lipid-lowering effects of leptin. Recently, Wang et al. (28) reported that 3 days of feeding of a highly palatable diet that increased caloric intake by ϳ2-fold resulted in the development of resistance to the effects of leptin on hepatic carbohydrate metabolism. After 3 days of the high fat diet, leptin at a concentration of ϳ10 ng/ml was sufficient to decrease liver triglyceride levels by 37 Ϯ 6% (p ϭ 0.008) (Fig. 4). However, unlike the study of Wang et al. (28), rats on the high fat diet used in this study do not increase caloric intake compared with chow-fed controls (33). In contrast, with 3 days feeding of the highly palatable diet (28), rats increased caloric intake (from 72 Ϯ 3 to 112 Ϯ 4 kcal/day, p ϭ 0.001) and triglyceride levels were not significantly reduced by leptin, indicating the presence of leptin resistance (Fig. 4).
Hepatic Leptin Resistance in Diet-induced Obesity Is Associated with Impaired Activation of PI 3-kinase, a Required Mediator of the Lipid-lowering Effects of Leptin in Liver-A number of post-receptor signaling pathways have been proposed to mediate leptin signal transduction and ultimately the metabolic effects of leptin. Of these, leptin activation of PI 3-kinase (24 -27), PDE3B (25,26), and STAT3 (22,23) are best described. We hypothesized first that the activation of one or a subset of these signaling elements would be required for the lipid-depleting effects of leptin on liver and second that defects in leptin activation of these pathways may contribute to hepatic leptin resistance in DIO. As a first step in addressing these hypotheses, we determined the effects of wortmannin (100 nM) and LY294002 (10 M), two inhibitors of PI 3-kinase activity, and cilostamide (1 M), a PDE3B-specific inhibitor, on the lipid-lowering effects of leptin (Fig. 5). Co-perfusion of wortmannin or LY294002 with leptin completely blocked the lipid-depleting effects of leptin on livers from lean rats, dem- Livers were isolated and perfused as described in Fig. 1. Subsequently, livers were flash-frozen in liquid nitrogen and pulverized and triglyceride content was measured. The asterisk indicates a significant difference between the vehicle (Leptin Ϫ) and the leptin (Leptin ϩ) groups or (**) between OBESE (Leptin Ϫ or ϩ) and LEAN (Leptin Ϫ). n ϭ minimum of 6 livers/group. Results are presented as mean Ϯ S.E.

FIG. 3. Effects of 90 min of leptin perfusion on oxygen consumption in livers from lean and diet-induced obese rats.
Livers were isolated and perfused as described in Fig. 1. Oxygen consumption was measured as described under "Experimental Procedures." Data are presented as percentage change in oxygen consumption (0 versus 90 min) in leptin-perfused (Leptin ϩ) livers compared with corresponding vehicle-perfused livers (Leptin Ϫ) at the end of the 90-min perfusion period. The asterisk indicates a significant difference between Leptin Ϫ and Leptin ϩ groups. n ϭ minimum of 6 livers/group. Results are presented as mean Ϯ S.E.

FIG. 4. Effects of 3 days of a high fat diet or a highly palatable diet on leptin-induced decreases in liver triglyceride levels.
Livers were isolated from 18-h fast rats after 3 days of a high fat diet (3D-HF) or a highly palatable diet (3D-HC) and then perfused with either vehicle (Leptin Ϫ) or leptin (Leptin ϩ) for 90 min. Subsequently, livers were flash-frozen in liquid nitrogen and pulverized, and triglyceride content was measured. The asterisk indicates a significant difference between Leptin Ϫ and Leptin ϩ groups. n ϭ 6 livers/group for 3D-HF and n ϭ 4 livers/group for 3D-HC. Results are presented as mean Ϯ S.E. onstrating that PI 3-kinase activation is necessary for the effects of leptin on hepatic lipid levels. Conversely, cilostamide was unable to prevent the leptin-induced decrease in liver triglycerides (24 Ϯ 5%, p ϭ 0.01), demonstrating that PDE3B activation is not required for the lipid-reducing effects of leptin. Furthermore, in the absence of inhibitors, leptin increased hepatic PI 3-kinase activity by 183 Ϯ 6% (p ϭ 0.003) and increased tyrosine phosphorylation of IRS-1 by 185 Ϯ 30% (p ϭ 0.02) compared with vehicle-perfused control livers (Fig. 6, panels A, C, and D). In agreement with previous reports (36,37), leptin had no effects on STAT3 phosphorylation (data not shown). Because PI 3-kinase activation is required for the lipid-reducing effects of leptin in livers from lean rats, we next determined the effects of leptin on PI 3-kinase activity in livers from DIO rats (Fig. 6, panels B, C, and D). Unlike in livers from lean animals, leptin had no effects on PI 3-kinase activity or on tyrosine phosphorylation of IRS-1 in DIO livers, demonstrating a mechanism for hepatic leptin resistance. DIO did not alter serine phosphorylation of IRS-1 compared with lean controls (data not shown). Importantly, the inability of leptin to activate PI 3-kinase in DIO was not associated with a down-regulation of leptin receptor expression (Fig. 7, demonstrating in liver that leptin resistance is a post-receptor defect. Furthermore, although the short form leptin receptor (ObRa) was easily detectable by Western blot and RPA, we were unable to detect ObRb in liver by either method (data not shown). These data suggest that ObRa mediates the lipid-depleting effects of leptin in liver. In combination, these data demonstrate that leptin acutely decreases hepatic triglyceride levels via a PI 3-kinase mechanism that is independent of PDE3B and STAT3 activation and that leptin activation of PI 3-kinase is defective in DIO. DISCUSSION The current study makes a number of important and novel observations on the mechanisms of the lipid-depleting action of leptin in liver and hepatic leptin resistance in diet-induced obese rats. First, we demonstrate that excessive hepatic storage of lipid in obesity is associated with the development of resistance to the acute lipid-lowering effects of leptin. Second, we demonstrate that PI 3-kinase activation is required for the acute lipid-lowering effects of leptin on liver. Third, we demonstrate that a mechanism of hepatic leptin resistance is the impaired leptin activation of PI 3-kinase, and finally, we dem-onstrate that hepatic leptin resistance is at least partially due to the loss of direct actions of leptin on liver as opposed to an indirect hypothalamic mechanism.
Few studies have addressed the effects of polygenic environment-influenced obesity such as diet-induced obesity and human obesity on leptin-regulated metabolic processes. In DIO rats, leptin stimulation of exogenous fatty acid oxidation and lipid hydrolysis in skeletal muscle is decreased compared with lean controls (20), and in humans, leptin increases fatty acid oxidation in lean but not in obese human skeletal muscle (21). In liver, 3 days of excess (ϳ2-fold greater than normal) caloric intake impairs leptin effects on hepatic glucose metabolism (28). The current study extends these findings to demonstrate that the direct acute effects of leptin on hepatic lipid levels are profoundly impaired in diet-induced obesity, a situation of elevated hepatic storage of lipid. Of particular note is the observation that a perfusate leptin concentration of ϳ90 ng/ml (ϳ4fold greater then fed plasma leptin concentrations in a 6-week high fat fed animal) 2 was not sufficient to reduce the elevated hepatic triglyceride levels in obese livers, whereas ϳ10 ng/ml was sufficient in lean livers. Taken together, these studies suggest that leptin resistance may contribute to abnormalities 2 L. Nguyen and R. M. O'Doherty, unpublished data. of lipid metabolism associated with obesity, particularly elevated lipid storage in non-adipose tissues. Furthermore, given the close association between dyslipidemia and insulin resistance (19) and observations that leptin-induced increases in insulin sensitivity are associated with decreases in plasma and tissue lipid levels (33,38), it is also plausible to speculate that leptin resistance may indirectly contribute to the development of insulin resistance in obesity.
A shorter length of time on the high fat diet (3 days) that did not increase caloric intake compared with chow-fed animals was not sufficient to induce hepatic leptin resistance, whereas the same length of time on a highly palatable cafeteria diet (28) that did increase caloric intake induced hepatic leptin resistance. These results are in agreement with the observation that overfeeding is required for hepatic leptin resistance to develop on a short term diet (28). However, the high fat diet also differs from the cafeteria diet in nutrient composition. Thus, the high fat diet has a higher percentage of calories derived from fat (45 versus 33% for the highly palatable diet), a lower percentage of calories delivered from carbohydrate (40 versus 45%), a higher saturated fatty acid content (19 versus 6%), and a higher sucrose content (29 versus 3.5%). It is unclear if or how these nutrient differences may contribute to the capacity of the highly palatable diet to induce rapid leptin resistance.
Unlike the acute effects of leptin, prolonged hyperleptinemia in diet-induced obese rats decreases food intake, increases insulin sensitivity, and decreases skeletal muscle triglyceride levels (33,38). Thus, leptin resistance is not absolute in DIO and may be related to differences between acute and chronic leptin treatment, variability in the responsiveness of different tissues to leptin, hypothalamic versus peripheral effects of leptin, or indirect effects of decreased food intake on metabolic processes.
The metabolic effects of leptin are dependent on the activation of post-receptor signaling cascades. In recent studies, leptin activation of PI 3-kinase in liver (24,26,36,39) and in other tissues (25,27,40) has been demonstrated and a role for PI 3-kinase in mediating hypothalamic leptin action has been established (25,27). Thus, it has been demonstrated that inhibition of leptin-stimulated PI 3-kinase activity in the hypothalamus abrogates the effects of leptin on food intake (27). In the current study, we demonstrate that the lipid-lowering effect of leptin on the liver is dependent on PI 3-kinase activation. In support of this conclusion are our observations that leptin stimulates PI 3-kinase activity and increases tyrosine phosphorylation of IRS-1 in liver, that inhibition of PI 3-kinase activity by wortmannin and LY294002 is sufficient to block the lipiddepleting effects of leptin, and that resistance to the lipiddepleting effects of leptin in diet-induced obesity are associated with an absence of leptin-stimulated PI 3-kinase activity and tyrosine phosphorylation of IRS-1. Taken together, these data suggest an important role for PI 3-kinase in mediating leptin action and suggest a mechanism for mediating leptin effects on lipid metabolism in other tissues such as skeletal muscle and adipose tissue.
Neither PDE3B nor STAT3 (data not shown) appear to be required for the lipid-lowering effects of leptin in liver. In support of this possibility, cilostamide at a dose that is ϳ200fold greater than is required to inhibit PDE3B activity (IC 50 ϭ 5 nM) (41) did not block the lipid-reducing effects of leptin and STAT3 phosphorylation was not increased by leptin (data not shown), in agreement with previous in vivo reports (36,37). Despite the observations of Zhao et al. (26) that leptin increases PDE3B activity in primary hepatocytes, the role of PDE3B in the metabolic effects of leptin in liver remains unclear. Thus, O'Doherty et al. (42) demonstrated that an acute infusion of leptin does not increase PDE3B activity in rat liver. Furthermore, Cases et al. (43) determined that leptin suppression of insulin secretion in vivo is not blocked by milrinone, a PDE3B inhibitor. However, it should be noted that inhibition of PDE3B in the hypothalamus suppresses the effects of leptin on food intake (25), suggesting that a requirement for PDE3B activity for leptin action may be process-and tissue-specific.
The role of AMPK in mediating the effects of leptin on fatty acid oxidation remains unclear. Thus, Minokoshi et al. (44) demonstrated that leptin stimulation of fatty acid oxidation in skeletal muscle is associated with increased AMPK activity. However, Atkinson et al. (45) demonstrated that AMPK activity is not increased by leptin in the perfused heart, despite increases in fatty acid oxidation. In this study, the lipid lowering effect of 90-min leptin perfusion in liver was not associated with increases in AMPK activity. Clearly, more work is required to determine the exact nature of leptin effects on AMPK activity.
To our knowledge, this study is the first to demonstrate a role for defective leptin-stimulated PI 3-kinase activity in leptin resistance. It has been proposed that defects in leptin signaling are a potential mechanism of leptin resistance in obesity. Decreased STAT3 and JAK2 activity and increased SOCS3 expression and PTPB1 activity have been implicated in the development of hypothalamic leptin resistance (46 -49). More recently, it has been established that pharmacological inhibition of leptin-stimulated PI 3-kinase activity in the hypothalamus inhibits the effects of leptin on food intake (25,27), suggesting that a decreased ability of leptin to activate PI 3-kinase could play a role in leptin resistance. The current study offers direct evidence to support this hypothesis. Thus, hepatic leptin resistance is associated with both a loss of leptin effects on hepatic lipid levels and a loss of leptin-stimulated PI 3-kinase activity. Furthermore, leptin was unable to increase tyrosine phosphorylation of IRS-1 in obese livers, whereas serine phosphorylation was unchanged by obesity (data not show). Importantly, the hepatic leptin resistance cannot be explained by a down-regulation of leptin receptors, because we show that ObR mRNA and protein levels are not decreased in DIO. Because we demonstrate that leptin-stimulated activation of PI 3-kinase is required for the effects of leptin on liver lipid levels, it is plausible to suggest that the inability of leptin to activate PI 3-kinase in livers from obese animals is a primary mechanism underlying hepatic leptin resistance. Important future studies will be to determine the effects of DIO on leptin-activated PI 3-kinase activity in the hypothalamus and the mechanism of impaired PI 3-kinase activation by leptin in the liver.
A number of studies have demonstrated metabolic effects of leptin on peripheral tissues that do not require leptin action at the hypothalamus, demonstrating that peripheral leptin receptors play a role in mediating whole body leptin action (50). The current study demonstrates that leptin effects on liver lipid metabolism can be mediated by hepatic leptin receptors at physiological leptin concentrations, confirming observations from previous studies (14,15,51). Because we were unable to detect ObRb protein or mRNA (data not shown), similar to previous studies (26,52), it is most likely the predominant ObRa that mediates the lipid-lowering effects of leptin. However, we cannot rule out the possibility that extremely low levels of ObRb may be present and functional in liver. There are few in vivo studies that directly address the role of peripheral leptin receptors in mediating the metabolic actions of leptin. Expression of wild-type leptin receptors in ZDF liver decreases liver triglyceride levels (14), denervated adipose tissue is depleted by hyperleptinemia in vivo (53), and body fat is decreased by hyperleptinemia without increasing hypotha-lamic leptin action in high fat fed rats (54). Minokoshi et al. (44) conclude that leptin activation of AMPK in skeletal muscle is partially a result of direct actions of leptin on muscle. Conversely, Cohen et al. (55) conclude that a 70% reduction in hepatic leptin receptors has no effect on hepatic triglyceride levels. However, this observation is difficult to interpret because a complete knock-out of liver leptin receptors was not achieved, comparisons of triglyceride levels were made with obese animals displaying liver steatosis, and effects of leptin receptor down-regulation on intrahepatic metabolism were not studied.
In conclusion, our study demonstrates the development of hepatic leptin resistance defined as the inability of leptin to acutely decrease hepatic triglyceride levels in DIO. We demonstrate that leptin activation of PI 3-kinase is required for the lipid-lowering effect of leptin in liver and is absent in DIO, demonstrating a novel mechanism of leptin resistance in obesity.