Effects of AMPK Activation on Insulin Sensitivity and Metabolism in Leptin-Deficient ob/ob Mice

AMP-activated protein kinase (AMPK) is a heterotrimeric complex, composed of a catalytic subunit (α₁ or α₂) and two regulatory subunits (β₁ or β₂ and γ₁, γ₂, or γ₃), which acts as a metabolic sensor to regulate glucose and lipid metabolism (1). AMPK is activated in response to changes in the intracellular AMP/ATP and ADP/ATP ratios in response to cellular stress or nutrient deprivation, changes in calcium concentration, or alterations in circulating levels of various hormones including leptin, cytokines, or adiponectin (2,3). Several lines of evidence highlight AMPK as an intertissue signal integrator among peripheral tissues and the hypothalamus to control whole-body energy and glucose homeostasis (4–10).

Acute activation of AMPK in peripheral tissues stimulates glucose uptake and lipid oxidation to produce energy, while turning off energy-consuming processes including synthesis of glycogen, lipids, and proteins (2). Pharmacological activation of AMPK in rodents or humans with insulin resistance or type 2 diabetes increases skeletal muscle glucose uptake (5,7,11,12) by an insulin-independent mechanism (13). Although AMPK can form up to 12 unique heterotrimeric complexes based on the expression of different α/β/γ subunits, only three complexes (α₁/β₂/γ₁, α₂/β₂/γ₁, and α₂/β₂/γ₃) are found in human skeletal muscle (14). Expression profiling of the AMPK γ-subunits in human and rodent skeletal muscle highlights a specialized role for the γ₃ isoform in glycolytic fibers (15). Expression of a naturally occurring mutant (R225Q) form of the AMPK γ₃-subunit (16) in COS7 cells increases basal AMPK phosphorylation, while concomitantly reducing sensitivity to AMP (17). Moreover, AMPKγ₃^R225Q (γ₃^R225Q) transgenic mice have increased glycogen content and enhanced mitochondrial biogenesis, and are protected against dietary-induced triglyceride accumulation and insulin resistance in glycolytic skeletal muscle (17,18). This phenotype is copied in humans expressing the rare AMPKγ₃^R225W mutation, which increases basal AMPK activity and muscle glycogen content, and decreases intramuscular triglyceride levels (19). Thus, lifelong expression of activated forms of AMPKγ₃ may prevent disturbances in glucose and lipid homeostasis that are characteristic of obese people with insulin resistance or type 2 diabetes (20,21).

The hypothalamus is a master regulator of food intake and energy balance, and coordinates glucose and energy homeostasis in response to the adipose-derived peripheral hormone leptin (22). Hypothalamic AMPK signaling plays an important role in the regulation of food intake (4,8,9). Activation of hypothalamic AMPK by counter-regulatory hormones involved in appetite control as well as by pharmacological AMPK activators increases food intake (4,9). Conversely, leptin inhibits hypothalamic AMPK signaling to reduce food intake and body weight, and increases AMPK signaling in peripheral tissues to promote lipid oxidation (4) and deplete triglyceride stores (23,24). Thus, an appropriate balance between AMPK signaling in central and peripheral tissues appears to be important for glucose and energy homeostasis. Previously, we have provided evidence that muscle-specific γ₃^R225Q transgenic mice rendered obese and hyperleptinemic by a high-fat diet are protected against excessive intramuscular triglyceride accumulation and the development of insulin resistance, presumably due to increased AMPK activation and lipid oxidation in skeletal muscle (17). This raises the question of whether AMPK activation can improve defects in insulin action and metabolism arising from severe obesity from either leptin deficiency or impaired leptin signaling. For example, in severely obese diabetic leptin receptor-deficient db/db mice, GLUT4 overexpression can improve glucose tolerance (25) as well as skeletal muscle insulin sensitivity (26). Furthermore, AMPK activation by AICAR treatment improves glucose homeostasis in rodents (5,7). Given the important role of leptin on energy balance, we determined whether glucose and energy homeostasis are improved by skeletal muscle overexpression of the mutant AMPKγ₃^R225Q subunit in leptin-deficient ob/ob mice and whether the effects of AMPK are leptin-dependent.

Reagents were purchased from Sigma-Aldrich (St. Louis, MO), unless otherwise stated.

Wild-type (WT) and skeletal muscle–specific γ₃^R225Q transgenic mice were generated as described previously (17). The transgenic γ₃^R225Q mutant mice were crossed with heterozygous ob/+ mice to generate ob/+-γ₃^R225Q mice. These ob/+-γ₃^R225Q mice were bred with ob/+ mice to generate the four mouse models studied in this report: lean WT, lean γ₃^R225Q, ob/ob-WT (ob/ob), and ob/ob-γ₃^R225Q mice. The ob/+ mice (on a C57BL/6J background) were purchased from Charles River Germany. ob/ob mice are obese and insulin-resistant because of a mutation in the hormone leptin; the leptin receptor is, however, intact in ob/ob mice. All the animal experiments were approved by the regional ethical committee on animal research Stockholm North, Sweden. Animals had free access to water and standard rodent chow (Lantmännen, Stockholm, Sweden), and were maintained in a temperature- and light-controlled (12-h light/dark cycle) environment. Animals were cared for in accordance with regulations for the protection of laboratory animals. Female and male mice were studied at 12–16 weeks of age.

Mice fasted for 4 h were anesthetized with Avertin (2,2,2-tribromoethanol 99% and tertiary amyl alcohol [1:1 w/v], 500 mg/kg body weight) and gastrocnemius muscle and liver were removed, cleaned of fat and blood, and quickly frozen in liquid nitrogen. Glycogen content was determined fluorometrically on HCl extracts as described previously (17). Triglyceride content was determined using a triglyceride/glycerol blanking kit (Roche Diagnostics Scandinavia, Bromma, Sweden) using Seronorm Lipid as a standard (17).

Animals were fasted for 4 h, and baseline glucose levels were measured using a OneTouch Ultra glucose meter (LifeScan, Milpitas, CA). Blood samples were collected from the tip of the tail. Glucose (1 g/kg) was injected intraperitoneally, and blood glucose levels were measured at 15, 30, 60, and 120 min after the injection. Blood was collected at baseline and 15 min after the glucose injection to determine insulin levels. Plasma insulin concentration was determined using an Ultra Sensitive Insulin ELISA Kit (Crystal Chem, Downers Grove, IL).

The jugular vein was catheterized 5–7 days prior to the clamp under isoflurane anesthesia. On the day of experiment, animals were fasted for 4 h. Euglycemic-hyperinsulinemic clamps were performed on conscious lean WT and γ₃^R225Q mice (10 mU insulin/kg/min) and ob/ob and ob/ob-γ₃^R225Q mice (75 mU insulin/kg/min) and insulin-stimulated glucose uptake in glycolytic extensor digitorum longus (EDL) and gastrocnemius muscles was determined using 2-deoxy-d-[1-¹⁴C]-glucose (PerkinElmer, Waltham, MA) as a tracer (27). Results are reported in nanograms of glucose per milligram per minute.

Palmitate oxidation in isolated EDL muscle was analyzed as described previously (28). EDL muscles were incubated in a Krebs-Henseleit buffer. After recovery, muscles were incubated for 2 h in ³H-palmitate (PerkinElmer). Palmitate oxidation was determined by analyzing the ³H-labeled water content using liquid scintillation counting. Results are reported in picomoles per milligram per minute.

Mitochondrial respiration in freshly isolated EDL muscle was determined using high-resolution respirometry (Oxygraph-2k; Oroboros Instruments, Innsbruck, Austria) as described previously (29,30). EDL muscles were dissected out, and fibers were gently separated under a microscope. Following saponin permeabilization in ice-cold relaxing and biopsy preservation solution, tissues were equilibrated in ice-cold mitochondrial respiration medium (MiRO5) and 1–2 mg of tissue was added to the respirometry chamber containing MiRO5. Leak respiration was measured by adding malate and pyruvate in the absence of ADP. Thereafter, ADP was added to measure oxidative phosphorylation. Respiration through complex I (C I) was measured by the addition of glutamate followed by the addition of succinate to measure C I+II respiration. Maximum flux through the electron transfer system (ETS) was measured by the addition of exogenous uncoupler carbonylcyanide-4-(trifluoromethoxy)-phenyl-hydrazone (ETS I+II). Rotenone was used to inhibit electron transport through C I (ETS II). Absolute oxygen flux is expressed relative to tissue wet weight per second (picomoles of O₂ per milligram per second).

Mice were acclimatized in individual cages for a period of 2–3 days. They were injected with either saline or leptin (1 mg/kg, reconstituted in saline solution from PeproTech, Rocky Hill, NJ) at 1600 h for a period of 5 days. Food consumption and body weights were recorded daily. On the sixth day, after a 4-h fast, blood glucose level was determined, and the gastrocnemius muscle was dissected and immediately frozen for determination of triglyceride, glycogen, diacylglycerol (DAG), and ceramide content, and Western blot analysis.

Body composition (lean and fat mass) before and after saline and leptin treatment was determined in conscious mice using quantitative magnetic resonance imaging (EchoMRI, Houston, TX).

DAG and ceramide content were determined in gastrocnemius muscle from 4-h fasted mice by conversion of DAG and ceramides to phosphorylation products by externally added DAG kinase from Escherichia coli (Enzo Life Sciences, Farmingdale, NY) in the presence of [γ-³²P]ATP as previously described (31).

Plasma free fatty acids were determined by a commercially available kit (Wako Chemicals, Dusseldorf, Germany) in ob/ob and ob/ob-γ₃^R225Q mice fasted for 4 h.

Gastrocnemius muscle was homogenized in ice-cold homogenization buffer (NaCl 137 mmol/L, KCl 2.7 mmol/L, MgCl₂ 1 mmol/L, Na₄O₇P₂ 5 mmol/L, NaF 10 mmol/L, Triton X-100 1%, glycerol 10%, Tris pH 7.8, 20 mmol/L, EDTA 1 mmol/L, phenylmethylsulfonyl fluoride 0.2 mmol/L, Na₃VO₄ 0.5 mmol/L, and protease inhibitor cocktail ×1) (Calbiochem; Merck Millipore, Billerica, MA) using the TissueLyser (Qiagen, Hamburg, Germany). Protein content in the supernatant was determined using the Pierce BCA protein assay kit (Thermo Scientific, Rockford, IL). Proteins were separated on a 4–12% Criterion XT Bis-Tris Precast Gel (Bio-Rad, Hercules, CA) and transferred to nitrocellulose membrane (100 V, 80 min), then blocked in Tris-buffered saline with 0.02% Tween-20 containing 7.5% nonfat dry milk for 1 h at room temperature. Membranes were incubated with primary antibodies overnight at 4°C. MitoProfile total oxidative phosphorylation antibody cocktail was from Abcam (Cambridge, U.K.). Abundance of the following complex markers was determined; C I, NADH dehydrogenase (ubiquinone) 1 β subcomplex 8 (NDUFB8); C II, succinate dehydrogenase complex, subunit B, iron sulfur (SDHB); C III, ubiquinol cytochrome c reductase core protein 2 (UQCRC2); C IV, cytochrome c oxidase I, mitochondrial (MTCO1); and C V, ATP synthase, H⁺ transporting, mitochondrial F1 complex, α subunit 1 (ATP5A). GLUT4 antibody was from Millipore (Temecula, CA). Phospho-AMPKα^Thr172, AMPKα, phospho-acetyl-CoA carboxylase (ACC) α^Ser79/β^Ser212, and ACCα/β antibodies were from Cell Signaling Technology (Danvers, MA). Phospho-ACCβ^Ser219/221 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA). The AMPKγ³ antibody was a gift from Dr. Grahame Hardie (University of Dundee, Dundee, U.K.). Membranes were incubated with appropriate secondary antibody conjugated with horseradish peroxidase (Bio-Rad). The immunoreactive proteins were detected by enhanced chemiluminescence (Amersham, Arlington Heights, IL) and quantified by calibrated densitometry using Quantity One image analysis software (Bio-Rad). GAPDH was used as a loading control.

Statistical analysis was performed by unpaired two-tailed Student t test or two-way ANOVA, where applicable. The γ₃^R225Q mice were compared with WT mice, and the ob/ob-γ₃^R225Q mice were compared with the ob/ob, unless otherwise stated. The effect of leptin treatment was compared with saline treatment in the lean and obese mouse models. Results were considered statistically significant at P < 0.05.

Glycogen content was determined in the gastrocnemius muscle from 4-h fasted WT, lean γ₃^R225Q transgenic, ob/ob, and ob/ob-γ₃^R225Q transgenic mice. As previously reported (17), the AMPKγ₃^R225Q mutation increased glycogen content in lean mice (Fig. 1). WT and ob/ob mice have similar levels of glycogen content in the gastrocnemius muscle. Consistent with the lean γ₃^R225Q transgenic mice, glycogen content in the gastrocnemius muscle was increased in ob/ob-γ₃^R225Q transgenic mice compared with ob/ob mice (Fig. 1).

Glucose tolerance was similar between ob/ob and ob/ob-γ₃^R225Q transgenic mice (Fig. 2A), consistent with our previously observation that glucose tolerance is unaltered in lean WT and γ₃^R225Q transgenic mice (17). Plasma insulin concentrations determined at baseline and 15 min after the glucose injection were similar between the ob/ob and ob/ob-γ₃^R225Q transgenic mice (Fig. 2B).

We performed a euglycemic-hyperinsulinemic clamp to assess in vivo glucose uptake in EDL and gastrocnemius muscles. Lean WT and γ₃^R225Q transgenic mice were clamped using an insulin infusion of 10 mU/kg/min (baseline plasma insulin levels were 0.5 ± 0.1 and 0.6 ±0.2 ng/mL, respectively, in WT and γ₃^R225Q mice, and levels achieved during the euglycemic-hyperinsulinemic clamp were 9.4 ± 0.9 and 11.3 ± 1.2 ng/mL, respectively, in WT and γ₃^R225Q mice (n = 6–9). ob/ob and ob/ob-γ₃^R225Q transgenic mice were clamped using an insulin infusion of 75 mU/kg/min because of their extreme insulin-resistant state (baseline plasma insulin levels in ob/ob and ob/ob-γ₃^R225Q mice were 15.1 ± 5.6 and 9.3 ±1.6 ng/mL, respectively; levels achieved in ob/ob and ob/ob-γ₃^R225Q mice during the euglycemic-hyperinsulinemic clamp were 281.5 ± 60.3 and 338.1 ± 17.9 ng/mL, respectively; n = 4). The body weight of lean mice (WT mice 31.6 ± 0.8 g; γ₃^R225Q mice 32.4 ± 1.1 g; n = 10–12) and ob/ob mice (ob/ob 43.6 ± 1.3 g; ob/ob-γ₃^R225Q 41.3 ± 1.7 g; n = 6–9) without or with the γ₃^R225Q transgene was unaltered. Four-hour fasted plasma glucose levels were similar between WT and γ₃^R225Q mice (8.4 ± 0.4 and 9.8 ± 0.6 mmol/L, respectively; n = 10–12) and ob/ob and ob/ob-γ₃^R225Q mice (9.3 ± 0.7 and 11.4 ± 1.8 mmol/L, respectively; n = 6–9). Insulin-stimulated glucose uptake was increased in glycolytic EDL muscle (Fig. 3A), but not in gastrocnemius muscle (Fig. 3B), from lean γ₃^R225Q transgenic mice under in vivo conditions. Conversely, insulin-stimulated glucose uptake in EDL and gastrocnemius muscles was unchanged between ob/ob and ob/ob-γ₃^R225Q transgenic mice (Fig. 3A and B). Glucose concentration as well as the glucose infusion rate were similar at the steady state of the euglycemic-hyperinsulinemic clamp and after the 2-deoxy-d-[1-¹⁴C]-glucose injection in lean (Fig. 3C and D) and ob/ob mice (Fig. 3E and F).

We have previously reported the AMPKγ₃^R225Q mutation increases oleate oxidation in EDL muscle from high-fat–fed, but not chow-fed mice (17). Basal palmitate oxidation was similar in isolated EDL muscle from WT and ob/ob mice (Fig. 4). There was an effect of the γ₃^R225Q transgene to increase palmitate oxidation in EDL muscle from WT and ob/ob mice (Fig. 4).

Mitochondrial respiration in freshly dissected EDL muscles was analyzed using high-resolution respirometry. The leak respiration, which denotes endogenous uncoupling, was similar between the lean and obese nontransgenic and transgenic mice (Fig. 5A). Mitochondrial respiration at the level of C I, ETS I+II, and ETS II was increased in ob/ob versus WT mice (Fig. 5A). The AMPKγ₃^R225Q mutation did not alter the mitochondrial respiration in skeletal muscle from either WT or ob/ob mice. Protein abundance of markers of mitochondrial complexes (NDUFB8, SDHB, UQCRC2, MTCO1, and ATP5A) in gastrocnemius muscle was unaltered between WT and ob/ob mice with or without the AMPKγ₃^R225Q mutation, except for the C II marker UQCRC2, which was increased in ob/ob mice compared with WT mice (Fig. 5B–F).

Five-day intraperitoneal leptin treatment reduced food intake in lean and obese nontransgenic and transgenic mice, compared with the respective saline-treated mice (Fig. 6A). Leptin treatment decreased body weight in WT, ob/ob, and ob/ob-γ₃^R225Q mice compared with the respective saline-treated mice (Fig. 6B). In contrast, leptin treatment did not significantly decrease body weight in the lean γ₃^R225Q transgenic mice (Fig. 6B).

Body composition was assessed using magnetic resonance imaging. Total fat mass was reduced in leptin-treated ob/ob and ob/ob-γ₃^R225Q transgenic mice, compared with the respective saline-treated mice (Fig. 6C). Lean mass was unaltered after saline or leptin treatment in the lean and obese nontransgenic and transgenic mice (Fig. 6D). Leptin treatment reduced 4-h fasting blood glucose levels in ob/ob-γ₃^R225Q transgenic mice, with a similar trend in ob/ob mice (Fig. 6E).

Gastrocnemius muscle was used for biochemical analysis. Glycogen content was increased in the saline-treated lean γ₃^R225Q transgenic mice, compared with the WT mice (Fig. 7A). Similarly, glycogen content was increased in ob/ob-γ₃^R225Q transgenic mice, compared with the ob/ob mice. Glycogen content was not altered in WT and γ₃^R225Q transgenic mice after leptin treatment. However, glycogen content was reduced in ob/ob mice compared with saline-treated ob/ob mice after leptin treatment, but not in ob/ob-γ₃^R225Q transgenic mice (Fig. 7A). GLUT4 protein abundance was unaltered in gastrocnemius muscle from lean and ob/ob mice γ₃^R225Q transgenic mice (Fig. 7B).

Triglyceride content in gastrocnemius muscle was unaltered between saline- and leptin-treated WT and γ₃^R225Q transgenic mice (Fig. 7C). While triglyceride content was elevated in ob/ob versus WT mice, the presence of the AMPKγ₃^R225Q transgene in the ob/ob mice (ob/ob-γ₃^R225Q mice) reduced triglyceride content. Furthermore, leptin treatment reduced triglyceride content in ob/ob mice. However, leptin treatment did not further reduce triglyceride content in gastrocnemius muscle from the ob/ob-γ₃^R225Q transgenic mice (Fig. 7C). DAG and ceramide content were assessed in gastrocnemius muscle from saline- and leptin-treated mice. DAG content was increased in ob/ob mice compared with WT mice (Fig. 7D). DAG content was unaltered in saline- or leptin-treated γ₃^R225Q transgenic mice (Fig. 7D). Ceramide content was unaltered in gastrocnemius muscle from saline- or leptin-treated WT and ob/ob nontransgenic and transgenic mice (Fig. 7E). Furthermore, liver triglyceride content was unaltered between ob/ob and ob/ob-γ₃^R225Q mice (52.6 ± 5.3 and 52.6 ± 4.3 mg/g, respectively, for ob/ob and ob/ob-γ₃^R225Q mice; n = 8–11). Circulating free fatty acids were similar between ob/ob and ob/ob-γ₃^R225Q mice (0.39 ± 0.07 and 0.37 ± 0.05 mmol/L, respectively, for ob/ob and ob/ob-γ₃^R225Q mice; n = 7).

AMPKα^Thr172 phosphorylation in gastrocnemius muscle from saline-treated lean and obese nontransgenic and transgenic mice was similar (Fig. 8A). Skeletal muscle AMPKα^Thr172 phosphorylation was increased after the 5-day leptin administration in ob/ob-γ₃^R225Q transgenic mice (Fig. 8A). AMPKα and AMPKγ₃ subunit protein abundance was increased in lean γ₃^R225Q transgenic mice independent of treatment compared with WT mice (Fig. 8B and C), but not in ob/ob and ob/ob-γ₃^R225Q transgenic mice. ACCα^Ser79/β^Ser212 and ACCβ^Ser219/221 phosphorylation was increased in gastrocnemius muscle from lean and ob/ob-γ₃^R225Q transgenic mice (Fig. 8D and E). ACCα/β protein abundance was increased in lean γ₃^R225Q transgenic mice (Fig. 8F). ACC phosphorylation or abundance was unaltered in leptin-treated lean and obese nontransgenic and transgenic mice compared with the respective saline-treated mice.

Overexpression of key signaling proteins regulating energy metabolism in skeletal muscle can improve metabolic disturbances associated with insulin resistance. Improvements in glucose homeostasis have been achieved by overexpression of either GLUT4 or uncoupling protein 3 in skeletal muscle of mice (25,32). AMPK activation by AICAR treatment can also improve glucose homeostasis in rodents (5,7). Here we tested the hypothesis that expression of a single missense mutation (R225Q) in the AMPKγ₃ isoform in skeletal muscle of ob/ob-γ₃^R225Q mice improves glucose and energy homeostasis. We also determined whether leptin treatment would further normalize metabolic disturbances in obese individuals and whether the effects of AMPK activation on glucose homeostasis are leptin-dependent.

Impairments in whole-body glucose metabolism in type 2 diabetes patients are mainly attributed to defects in skeletal muscle glucose uptake and glycogen synthesis (33). Despite severe insulin resistance, AMPK activation promotes GLUT4 translocation and increases glucose uptake directly in skeletal muscle from insulin-resistant type 2 diabetes patients (11). A role for the AMPKγ-subunit in glucose metabolism was first appreciated from studies of Hampshire pigs expressing a naturally occurring dominant mutation (denoted RN−) in the gene encoding the AMPKγ₃-isoform (34). This mutation results in an excessive amount of glycogen storage in glycolytic skeletal muscle. In vitro studies in COS7 cells reveal that AMPKγ₃^R225Q complexes have higher basal AMPK activity and lack AMP dependence (17). Thus, the R225Q mutation is a gain-of-function mutation that abolishes allosteric regulation by AMP/ATP, which thereby increases basal AMPK activity.

γ₃^R225Q transgenic mice fed a high-fat diet are protected against the development of skeletal muscle insulin resistance (17). Here we report that in vivo glucose uptake in glycolytic skeletal muscle is increased during a euglycemic-hyperinsulinemic clamp in lean γ₃^R225Q transgenic mice. Consistent with our earlier findings in fat-fed γ₃^R225Q transgenic mice (17), we found that glycogen content was increased and triglyceride content was decreased in gastrocnemius muscle from ob/ob-γ₃^R225Q transgenic mice. These changes in ob/ob-γ₃^R225Q mice indicate that leptin is not required for AMPK-mediated glycogen synthesis. Moreover, the effect of the AMPKγ₃^R225Q mutation on glycogen content is independent of adiposity, consistent with clinical studies in obese humans harboring a similar mutation in the AMPKγ₃ subunit who have higher skeletal muscle glycogen content compared with WT carriers (19). However, in contrast to our earlier study in fat-fed mice (17), improvements in muscle biochemistry arising from the AMPKγ₃^R225Q mutation were insufficient to ameliorate whole-body insulin resistance in leptin-deficient mice. Thus, a permissive amount of leptin may be required to fully confer the AMPKγ₃^R225Q-dependent improvements in skeletal muscle insulin sensitivity.

Several molecular mechanisms account for skeletal muscle glycogen accumulation. Skeletal muscle GLUT4 protein abundance directly influences the rate of insulin-stimulated glucose uptake and metabolism (35). Moreover, GLUT4 overexpression prevents insulin resistance (26) and glucose intolerance (25) in severely obese diabetic leptin receptor–deficient db/db mice, highlighting the importance of skeletal muscle glucose transport in maintaining whole-body glucose homeostasis. Skeletal muscle–specific overexpression of glycogen synthase increases glycogen accumulation by an insulin-independent mechanism not involving glucose transport (36), reinforcing the importance of glucose metabolism. Despite the profound increase in glycogen content in chow-fed lean γ₃^R225Q mice, basal and insulin-stimulated glucose uptake in isolated skeletal muscle was similar to WT mice (17). Furthermore, GLUT4 protein abundance was unaltered in gastrocnemius muscle from lean and ob/ob nontransgenic and transgenic mice. Thus, the increase in glycogen content in γ₃^R225Q transgenic mice is unlikely to arise from constitutive increases in glucose uptake. Consequently, γ₃^R225Q mice resemble glycogen synthase transgenic mice, rather than GLUT4 transgenic mice. Glycogen content negatively regulates both AMPK activity (37) and insulin-stimulated glucose uptake (38). However, glucose uptake in glycolytic muscle was increased despite increased glycogen content. Thus, improvements in glucose uptake may require permissive levels of leptin, since the AMPKγ₃^R225Q mutation did not increase glucose uptake in leptin-deficient ob/ob mice. Alternatively, increased GLUT4 abundance or translocation may be required to achieve improvements in glucose tolerance and insulin sensitivity in individuals with severe obesity due to leptin deficiency.

Although lipids serve as an important fuel source for skeletal muscle, excessive levels may trigger insulin resistance (39). Strategies to reduce excess triglyceride levels in skeletal muscle improve insulin sensitivity (17,40). Intramuscular triglyceride content is reduced and insulin sensitivity is increased in γ₃^R225Q transgenic mice fed a high-fat diet concomitant with increased lipid oxidation (17). Likewise, intramuscular triglyceride content was reduced in ob/ob-γ₃^R225Q transgenic mice, presumably due to decreased triglyceride synthesis and a modest increase in lipid oxidation. DAG and ceramide content, intermediates in lipid metabolism, have been linked to the development of insulin resistance in skeletal muscle (41). Here we confirm that DAG content is increased in gastrocnemius muscle from ob/ob mice (42), yet it was unaltered in γ₃^R225Q transgenic mice. Ceramide content in gastrocnemius muscle was similar between transgenic and nontransgenic WT and ob/ob mice, confirming previous reports (42,43). However, phosphorylation of ACC, a downstream target of AMPK (44), was increased in γ₃^R225Q transgenic mice. Increased ACC phosphorylation leads to inactivation of the enzyme, which thereby increases lipid oxidation and decreases triglyceride levels (45). The increase in ACC phosphorylation suggests that the cellular energy charge driven by the AMPK mutation is altered (17), consistent with increased lipid oxidation in the γ₃^R225Q transgenic mice.

Leptin influences AMPK signaling in central and peripheral tissues. Leptin inhibits AMPK activity in the brain and reduces food intake (4), while enhancing lipid metabolism in skeletal muscle (46). Thus, we explored whether leptin treatment of ob/ob-γ₃^R225Q transgenic mice would lead to a further metabolic improvement imposed by the AMPKγ₃^R225Q mutation. The AMPKγ₃^R225Q mutation did not alter the body weight response to leptin treatment in ob/ob mice. Thus, the metabolic differences observed between the nontransgenic and ob/ob-γ₃^R225Q transgenic mice are directly related to the mutation, rather than to changes in food intake. Leptin treatment improved the fasting glucose level in ob/ob-γ₃^R225Q transgenic mice, concomitant with a normalization of food intake and a reduction in adiposity. Moreover, we found that skeletal muscle glycogen content was decreased in leptin-treated ob/ob mice, consistent with the effects of leptin in decreasing glycogen synthesis in ob/ob mice (47). However, leptin treatment did not decrease skeletal muscle glycogen content in γ₃^R225Q transgenic lean and ob/ob mice, indicating that the AMPKγ₃^R225Q mutation has a dominant influence on fuel partitioning within skeletal muscle, which may be overcome by hyperleptinemia. Although the concentration of leptin used in this study was sufficient to improve blood glucose and body weight, higher doses trigger a shift in substrate use such as that observed in fat-fed γ₃^R225Q mice (17).

AMPK activation is linked to mitochondrial biogenesis, providing a mechanism for the increased lipid oxidation observed in fat-fed γ₃^R225Q transgenic mice (17). The increase in mitochondrial respiration in ob/ob mice confirms our previous findings that obesity induces molecular adaptations in glycolytic skeletal muscle to enhance mitochondrial respiration (48). Nevertheless, these mice are severely insulin-resistant. Mitochondrial biogenesis is increased in glycolytic skeletal muscle from γ₃^R225Q transgenic mice, concomitant with increased expression of the coactivator peroxisome proliferator–activated receptor γ coactivator-1α and transcription factors that drive different mitochondrial proteins expression (18). However, mitochondrial respiration is unaltered between WT and γ₃^R225Q transgenic mice (18), as well as ob/ob-γ₃^R225Q transgenic mice. The increase in skeletal muscle mitochondrial markers in γ₃^R225Q transgenic mice (18) may account for the increase in lipid oxidation in the γ₃^R225Q transgenic mice. However, the regulation of insulin sensitivity is complex and not entirely coupled to increased skeletal muscle mitochondrial content. Nevertheless, cultured myotubes from probands carrying a homologous mutation (AMPKγ₃^R225W) reinforce the profound effect of this mutation on glucose uptake and metabolism, mitochondrial content, and oxidative capacity, and raise the clinical implications of mutations in the AMPKγ₃ subunit (49).

Defective leptin action leads to metabolic abnormalities associated with obesity. The effects of leptin are partly mediated via the AMPK pathway in central and peripheral sites. Here we show that the expression of a mutant form of the AMPKγ3 subunit in glycolytic skeletal muscle increases glycogen content and decreases intramuscular triglyceride levels. However, DAG and ceramide content were unaltered. The triglyceride depletion in ob/ob-γ₃^R225Q transgenic mice does not appear to improve glucose utilization and insulin sensitivity in ob/ob-γ₃^R225Q mice. Thus, the lack of central leptin signaling may override the favorable metabolic milieu conferred by peripheral overexpression of the AMPKγ₃^R225Q mutation to improve glucose and energy homeostasis. Further studies in hypothalamic-specific AMPK transgenic ob/ob mice may clarify the central role of this protein kinase in the control of glucose and energy homeostasis in leptin deficiency. Given our findings (Supplementary Table 1), targeting both peripheral and central AMPK actions may be required to improve glucose homeostasis.

Funding. This work was supported by grants from the European Foundation for the Study of Diabetes, Swedish Research Council, Swedish Diabetes Association, Swedish Foundation for Strategic Research (INGVAR II), the European Research Council, Novo Nordisk Research Foundation, the Strategic Research Programme in Diabetes at Karolinska Institutet, and Commission of the European Communities (Contract no. LSHM-CT-2004-005272 EXGENESIS).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. R.Z.T. researched the data, wrote the manuscript, and approved the final version of the manuscript. P.M.G.-R., A.V.C., and M.B. researched the data, reviewed and edited the manuscript, and approved the final version of the manuscript. R.J.O.S., L.Q.J., M.H.H., A.S.D., and E.V. researched the data and approved the final version of the manuscript. J.R.Z. wrote the manuscript and approved the final version of the manuscript. J.R.Z. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the Keystone Symposia Type 2 Diabetes, Insulin Resistance and Metabolic Dysfunction, Keystone, CO, 12–17 January 2011, and at the EMBO | EMBL Symposium Diabetes and Obesity, Heidelberg, Germany, 13–16 September 2012.