Ciliary Neurotrophic Factor Stimulates Muscle Glucose Uptake by a PI3-Kinase–Dependent Pathway That Is Impaired With Obesity

C57Bl6/J, ob/ob, AMPKα2 kinase-dead (KD) and gp130^ΔSTAT mice aged 8 to 18 weeks with corresponding littermate controls were used in experiments. Transgenic mice overexpressing a KD form of the AMPKα2 protein under control of the muscle creatine kinase promoter have been described previously (34). Briefly, the KD mutation was generated by Lys45Arg mutagenesis to encode a KD α2 protein that displaces all detectable endogenous α2 and 50–70% of endogenous α1 protein from αβγ heterotrimer complexes (34,35). The knockin mutant gp130^ΔSTAT mice express a COOH-terminal truncation mutation of gp130 that prevents activation of STAT1/3 as described previously (36). This mutation has a deletion of the distal intracellular domain responsible for AMPK and STAT activation. All mice were kept at a 12:12 h light:dark cycle at 20–21°C and provided ad libitum access to food and water. Mice were maintained on a standard rodent chow diet/low-fat diet (5% calories fat; Harlan Teklad) except in high-fat diet (HFD) experiments, in which the diet was composed of 45% calories from fat (SF04–027; Specialty Feeds). All procedures were approved by the St. Vincent's Hospital Animal Ethics Committee.

C57Bl6/J mice were fasted for 6 h before being injected with either saline or 0.3 mg/kg CNTF, a concentration of CNTF that activates AMPK and induces weight loss without eliciting an inflammatory response (20). Blood glucose was monitored over 150 min.

Soleus (oxidative and glycolytic fibers) and extensor digitorum longus (EDL; primarily glycolytic fibers) muscles were dissected from anesthetized mice (6 mg of pentobarbital per 100 g⁻¹ body weight) and transferred to incubation flasks containing 2 ml of essential buffer (Krebs-Henseleit buffer, pH 7.4, with 2 mmol/l pyruvate, 8 mmol/l mannitol, and 0.1% BSA), gassed with 95% O₂ + 5% CO₂, and maintained at 30°C as previously described (15). For all experiments, muscles were preincubated for 15 min in this buffer before it was replaced with buffer containing AICAR (2 mmol/l for 40 min; Toronto Research Chemicals Inc, Ontario, Canada), CNTF (at concentrations and durations indicated above, Axokine; Regeneron Pharmaceuticals Inc, New York), or insulin (30 nmol/l [additivity study] or 2.8 nmol/l [HFD study] for 40 min, Actrapid; Novo Nordisk, Bagsvaerd, Denmark). In separate experiments, muscles were incubated in LY-294002 hydrochloride (37) (60 nmol/l for 30 min; Sigma-Aldrich Corp, St. Louis, MO) before treatment with CNTF or insulin as described.

2-deoxy-d-glucose (2DG) uptake was measured over 10 min by replacing existing incubation buffer with the buffer described above but with the addition of 0.5 μCi/ml⁻¹2-[2,6-³H]-deoxy-d-glucose, 1 mmol/l 2-deoxy-d-glucose, and 0.2 μCi [1-¹⁴C]-mannitol/ml. After preparing muscles as described below, radioactivity was measured in muscle lysates by liquid scintillation counting (Tri-Carb 2000; Packard Instrument Co).

Muscles were homogenized in ice-cold buffer (50 mmol/l Hepes, pH 7.4, 150 mmol/l NaCl, 10 mmol/l NaF, 1 mmol/l sodium pyrophosphate, 0.5 mmol/l EDTA, 250 mmol/l sucrose, 1 mmol/l dithiothreitol, 1% TritonX-100, 1 mmol/l Na₃VO₄, and one Roche protease inhibitor tablet per 50 ml buffer) using an electrical homogenizer. Lysates were prepared as previously described (38) and stored at −80°C until analysis. Protein content in lysates was measured by the bicinchoninic acid method (Pierce).

Expression or phosphorylation of investigated proteins was determined in muscle lysates by SDS-PAGE and immunoblotting using the following primary antibodies: pan-αAMPK, phospho-AMPK T172, and phospho–acetyl CoA carboxylase (ACC) S222 (as previously described [(39)]), STAT3, phospho-STAT3 Y705, Akt, and phospho-Akt S473 (Cell Signaling Technology Inc). Secondary antibodies were horseradish-conjugated protein G (Bio-Rad Laboratories, Richmond, CA). Bands were visualized using an enhanced chemoluminescence system and quantified using ImageQuant TL 05 software (Amersham Biosciences, U.K.). Values obtained using phospho-specific antibodies are expressed as the ratio to the total content of the protein measured after stripping the membrane and reprobing.

AMPKα1 and -α2 activities were measured from 100 μg of muscle lysate protein using rabbit polyclonal AMPK antibodies for immunoprecipitation as previously described (40).

Lipids were extracted from freeze-dried powdered gastrocnemius muscle tissues using chloroform:methanol:PBS (1:2:0.8) and 0.2% SDS. Triacylglycerol were saponified in an ethanol-KOH solution at 60°C, and glycerol content was measured fluorometrically. Diacylglycerols and ceramides were extracted and quantified according to the methods of Preiss et al. (41).

Muscle cells were differentiated and serum starved overnight. The next morning, 20 nmol/l C2-dihydroceramide (Toronto Research Chemicals Inc, Ontario, Canada) was added for 6 h before cells were treated for 20 min with either CNTF (10 ng/ml) or insulin (10 nmol/l) before collecting protein lysates or measuring 2DG uptake as previously described (20). C2C12 cell were used for signaling experiments and L6 muscle cells for 2DG uptake because of the limited capacity of C2C12 to increase glucose uptake in response to stimuli.

Data are expressed as means ± SE. Statistical evaluations were performed by either Student's t test or two-way ANOVA using the Student-Newman-Keuls method as a post hoc test when appropriate. Differences between groups were considered statistically significant if P < 0.05.

Injection of mice with CNTF was associated with a reduction in blood glucose after 40 min compared with saline-injected controls. Blood glucose remained reduced in the CNTF group throughout the remainder of the test (Fig. 1,A). The AUC was 20% lower in CNTF-treated mice than in saline-injected controls (Fig. 1 B).

The effects of CNTF to reduce blood glucose in vivo may have been mediated by the suppression of hepatic glucose production or increased glucose uptake into adipose tissue and skeletal muscle. Therefore, to directly test the effects of CNTF on skeletal muscle glucose uptake, we incubated isolated muscles with CNTF at concentrations ranging from 2 to 500 ng/ml. CNTF increased muscle glucose uptake over 30 min in a dose-dependent manner with a significant increase at 2 ng/ml in both soleus and EDL muscles (Fig. 2,A–B). Glucose uptake was elevated with increasing CNTF concentrations up to 100 ng/ml with no further increases observed at 500 ng/ml (Fig. 2,A and B), consistent with dose-dependent effects on fatty acid oxidation observed previously (20). We next investigated if the effect of CNTF on glucose uptake was additive to the effects of insulin and AICAR. In soleus, CNTF, insulin, and AICAR independently increased glucose uptake to a similar degree, and combining AICAR and insulin with CNTF had an additive effect (Fig. 2,C). In EDL, insulin and AICAR increased glucose uptake more strongly than CNTF and no further increase was observed when combining these treatments with CNTF (Fig. 2 D).

CNTF increased glucose uptake in soleus and EDL after 30 min—an effect that diminished thereafter in both muscle types (Fig. 3,A). CNTF has been reported to increase the activity/phosphorylation of AMPK in muscle (20) and Akt in neurons (42), and because both kinases regulate muscle glucose uptake, we examined their activating phosphorylation. CNTF increased Akt S473-P nearly 200% in soleus and 450% in EDL muscle, and these effects diminished after 90 min (Fig. 3,B). AMPK T172-P was increased 90% in the soleus with CNTF after 30 min, and this was maintained for 60 min (Fig. 3,C). Although CNTF induced an apparent rise in ACCβ S218-P in soleus, this did not achieve statistical significance (P = 0.09) (Fig. 3,D). CNTF had no detectable effect on AMPK- or ACCβ-P in EDL (Fig. 3 C–D), which is consistent with previous observations in glycolytic muscle (20).

We next examined the relative importance of AMPK and PI3K/Akt signaling for CNTF-mediated glucose uptake. Muscles from wild-type and AMPK KD mice were preincubated with or without the PI3-kinase inhibitor, LY-294002, followed by stimulation with CNTF. Glucose uptake was increased to a similar degree with CNTF in wild-type and AMPK KD muscles, indicating that this response was AMPK independent. CNTF-stimulated glucose uptake was abolished by LY-294002 in both muscle types irrespective of mouse genotype (Fig. 4 A), which is consistent with PI3-kinase/Akt signaling being essential for this response.

We next measured α-isoform–specific AMPK activity to verify that CNTF had activated AMPK signaling in wild-type mice and that the activation was impaired in AMPK KD muscles (Fig. 4,B). As anticipated, overexpression of the AMPK KD construct lowered basal AMPK activity of both α-AMPK isoforms in both muscle types, with the most pronounced effect on α2-AMPK activity (34). In wild-type soleus, CNTF increased α1-AMPK activity by 180% and α2-AMPK activity by 35%, whereas CNTF had no effect on either AMPK isoform in AMPK KD muscles (Fig. 4,B). CNTF did not increase AMPK activity in the EDL (Fig. 4,B), which is consistent with CNTF not increasing AMPK phosphorylation in the EDL muscle (Fig. 3 C).

Lastly, to establish whether inhibition of PI3-kinase or AMPK signaling altered proximal CNTF signaling through the gp130-LIF receptor complex, we examined STAT3 Y705-P in the EDL muscle of wild-type and AMPK KD mice in the absence or presence of LY-294002. Incubation with CNTF increased STAT3 Y705-P by 140% in wild-type muscle, and importantly, the increase was unaffected in AMPK-KD muscles and unaffected by coincubation with LY-294002 (Fig. 4,C). We also verified that increases in Akt-P with CNTF were not affected by overexpression of the AMPK KD construct and that LY 294002 inhibited Akt activation (Fig. 4 D). Similar findings were observed in the soleus muscle (data not shown).

Signaling from the gp130-LIF receptor complex activates two distinct intracellular signaling pathways. The first depends on phosphorylation of four tyrosine residues (Y765, Y812, Y904, and Y914) on the distal cytoplasmic tail of gp130, which is required and sufficient for activating STAT1/3 (23) and AMPK (20). The second arm requires phosphorylation of the more proximal Y757 residue in gp130, which induces ERK and PI3-kinase/Akt signaling (23). We incubated muscles from gp130^ΔSTAT mice, which have a deletion of the cytosolic distal portion of the gp130 receptor required for induction of STAT1/3 and AMPK signaling but retain the Y757 residue that is required for activation of Erk and PI3-kinase/Akt signaling. Incubation with CNTF increased glucose uptake in soleus and EDL of wild-type mice, and this increase was maintained in gp130^ΔSTAT knockin mutant mice (Fig. 5,A). CNTF increased Akt S473-P in both genotypes, whereas STAT3 Y705-P was increased in wild-type but not in gp130^ΔSTAT muscles as expected (Fig. 5 B). Thus, CNTF increased glucose uptake normally despite impaired gp130 signaling and activation of AMPK.

It is well established that obesity is associated with impaired insulin sensitivity in skeletal muscle, and impaired activation of IRS-1 has been attributed a primary role (28,43). Because CNTF stimulates glucose uptake downstream of IRS-1, we tested whether CNTF-stimulated glucose uptake is maintained in muscles from mice fed a HFD for 12 weeks and obese mice deficient in leptin (ob/ob). As dictated by design, the high-fat feeding increased body weight by ∼20% (chow, 28.4 ± 0.5 g; HFD, 34.0 ± 0.7 g), fasting insulin levels (836 ± 76 pg/ml to 1,670 ± 377 pg/ml), and the AUC during a GTT (chow, 539 ± 23; HFD, 870 ± 94). Similarly, ob/ob mice had dramatically increased body mass (wild-type, 29.6 ± 0.7 g; ob/ob, 57.3 ± 2.8 g) and AUC during a GTT (WT, 574 ± 26; ob/ob, 1,347 ± 98). CNTF increased glucose uptake by 80 to 100% in the soleus and EDL in chow-fed mice, whereas CNTF-stimulated glucose uptake was completely blunted in the soleus and reduced by ∼50% in EDL of mice fed a HFD (Fig. 6,A). Similarly, CNTF increased muscle glucose uptake in both muscle types of lean littermates of ob/ob mice; however, this increase was abolished in muscles from ob/ob mice (Fig. 6 B).

We then examined the mechanism by which obesity inhibited CNTF-stimulated glucose uptake. CNTF increased STAT3 Y705-P in skeletal muscle of chow- and HFD-fed animals (Fig. 6,C), indicating no impairment of CNTF action at the receptor level. CNTF-induced Akt S473-P was increased by 4- and 19-fold in soleus and EDL of chow-fed mice, respectively, and this increase was markedly impaired with obesity in both muscle types (Fig. 6,D). In contrast, AMPK T172-P in soleus was also unaffected by diet (Fig. 6 E).

High-fat feeding increased muscle contents of triacylglycerol, diacylglycerol, and ceramide by ∼100%, ∼50%, and ∼50%, respectively, compared with chow-fed controls (Fig. 7,A). Because ceramides have been shown to directly inhibit insulin-stimulated Akt phosphorylation (33), we tested whether exposure of cultured myotubes to a short-chain ceramide analog also impaired CNTF-stimulated glucose uptake and Akt phosphorylation. Consistent with our findings in obese skeletal muscle, myotubes treated with C2-ceramide had impaired CNTF-stimulated glucose uptake (Fig. 7,B). Reductions in glucose uptake were not the result of inhibition of the gp130-LIF receptor complex because CNTF increased STAT3-P by ∼350% in both vehicle and C2-ceramide–treated cells (Fig. 7,C). However, C2-ceramide reduced basal and CNTF-stimulated Akt S473-P (Fig. 7,D). AMPK T172-P was increased (P < 0.05) by 60% with CNTF, and this increase was unaffected by C2-ceramide treatment (Fig. 7 E).

CNTF can regulate metabolic and growth signaling pathways in several tissue types. The present study shows that CNTF regulates glucose uptake and delineates the proximal signaling events mediating this response. We show that acute exposure of CNTF reduces blood glucose in vivo and increases glucose uptake in soleus and EDL muscle in vitro through the PI3-kinase/Akt signaling pathway. It was expected that AMPK would also exert a role in this process, but we found that AMPK was not required for CNTF to stimulate glucose uptake. Skeletal muscle is the most important tissue for insulin-stimulated glucose disposal (44), and skeletal muscle insulin resistance is a major defect in most obese phenotypes. Accordingly, we investigated whether CNTF-stimulated glucose uptake was maintained in muscle from obese, insulin-resistant mice. The rationale for these studies was to establish whether gp130-LIF signaling could constitute an alternative pathway to substitute for insufficient insulin signaling to skeletal muscle glucose uptake. However, obesity was associated with impaired CNTF-stimulated glucose uptake, which, at least in part, may have been attributed to increased muscle lipids and impaired PI3-kinase/Akt signaling with CNTF.

CNTF is a peptide hormone of the IL-6 family, which is highly expressed in peripheral nerves and other tissues (24,45,46). CNTF levels are low in serum of healthy individual as a result of the absence of an exocytosis targeting sequence (47,48). Recent reports have shown significant promise for the use of CNTF as an antiobesity therapeutic because it suppresses food intake acutely (49) and induces hypothalamic neurogenesis (50), leading to a new set point in body mass. CNTF also has direct antidiabetic effects in peripheral tissues. Chronic treatment of obese diabetic mice with CNTF increases metabolic rate (51) and reduces liver steatosis by enhancing fat oxidation and reducing synthesis of complex lipids (52). We have recently shown that chronic CNTF treatment of obese diabetic mice reverses obesity-induced insulin resistance by activating AMPK and reducing muscle lipid accumulation (20). In line with this study, CNTF acutely prevents muscle insulin resistance in response to a 2-h lipid infusion by preventing lipid accumulation (27). The clinical efficacy and safety of CNTF as a possible therapeutic for obesity is supported by findings in humans demonstrating weight loss and improved glycemic control (53).

Because we previously showed that CNTF activates AMPK in muscle (20) and AMPK is well known to increase glucose uptake in skeletal muscle (15,54,55), we expected AMPK rather than PI3-kinase/Akt signaling to be required for CNTF-stimulated glucose uptake. However, multiple lines of evidence indicate that CNTF signaling to glucose uptake may be mediated by a PI3-kinase/Akt–dependent signaling pathway: 1) the kinetics in Akt S473 phosphorylation correlated with increases in glucose uptake in both soleus and EDL muscles; 2) incubation with the PI3-kinase inhibitor, LY-294002, completely blocked CNTF-stimulated Akt phosphorylation and glucose uptake; 3) CNTF increased glucose uptake was not impaired in muscles overexpressing a KD AMPKα2; 4) AMPK was not activated by CNTF in the EDL muscle despite a 100% increase in glucose uptake; 5) a knockin mutation of a truncated gp130 receptor, which impairs gp130 signaling toward STAT3 and AMPK, had no effect on CNTF-stimulated glucose uptake; and 6) obesity induced by an HFD or treatment of cells with C2-ceramide impaired CNTF stimulated glucose uptake and was associated with blunted Akt but not AMPK phosphorylation. However, our findings demonstrating that CNTF-stimulated glucose uptake was additive to a maximal dose of insulin in Soleus, but not EDL muscle suggest that under some conditions, AMPK activation may also play a role in the stimulation of glucose uptake in Soleus, albeit to a lesser extent than originally anticipated. A possible explanation for the limited role of AMPK in CNTF-stimulated glucose uptake is that CNTF predominantly activated the α1-AMPK isoform. Studies in AMPKα2-null mice have shown that AMPKα2, but not α1 is required for the regulation of muscle glucose uptake in response to stimuli such as AICAR (15), possibly explaining why CNTF activation of AMPK is not required to increase glucose transport.

Because we found CNTF to increase muscle glucose uptake, it was of interest to test the effects of CNTF in insulin-resistant muscle from obese mice. We found that the efficacy of CNTF to regulate glucose uptake was reduced by ∼50% or more in muscles from obese HFD-fed and ob/ob mice. CNTF resistance with obesity seems not to be the result of impaired activation of the gp130-LIF receptor complex because CNTF-induced phosphorylation of AMPK and STAT3 were normal in obese muscle. In contrast, CNTF-stimulated phosphorylation of Akt was suppressed in obese muscles and was associated with muscle DAG and ceramide accumulation. Ceramides downregulate Akt phosphorylation and membrane translocation by activating protein phosphatase 2A (56). Ceramides also increase the activity of protein kinase C ζ, leading to inhibitory Akt S34 phosphorylation (57). In addition to the finding that ceramides were elevated in muscle from obese mice, our studies in myotubes showed that the C2-ceramide not only reduced basal Akt phosphorylation but also completely prevented CNTF-induced Akt phosphorylation and CNTF-stimulated glucose uptake. These results agree with previous studies in myotubes (33) and isolated rodent (32) and human (58) skeletal muscle demonstrating impaired insulin-stimulated Akt activation and glucose uptake when ceramides are elevated. In the present studies, DAG and a reduction in PI-3 kinase activity may have also contributed to the reduced CNTF activation of Akt in muscles from obese animals. Importantly, the finding that neither the C2-ceramide nor obesity altered CNTF-induced STAT3 and AMPK phosphorylation is consistent with our previous findings that CNTF activation of AMPK and lipid metabolism is maintained in skeletal muscles from obese, insulin-resistant mouse models (20). Thus, although the signaling arm of CNTF to glucose uptake is compromised with obesity, the activation of AMPK and lipid oxidation is preserved.

In summary, our data show that CNTF-stimulated glucose uptake is associated with activation of the PI3-kinase/Akt signaling pathway. Furthermore, CNTF-stimulated Akt S473 phosphorylation and glucose uptake is impaired in muscles from obese insulin-resistant mice despite the maintenance of STAT3 and AMPK signaling. This mechanism may, in part, be attributable to ceramide accumulation with obesity, which in turn impairs CNTF signaling to the PI3-kinase/Akt signaling arm of the gp130 receptor. Because CNTF-induced activation of AMPK was maintained in obesity, CNTF or CNTF analogs provide the bases for a viable therapeutic to prevent muscle lipid accumulation and to restore insulin sensitivity in obesity.

This research was supported by grants from the National Health and Medical Research Council, Australia (G.R.S., M.J.W., and B.E.K.) and the National Institutes of Health (NIH), U.S. (M.J.B.; R01-DK56886). S.B.J. was supported by a Danish Research Council of Health and Diseases postdoctoral fellowship. M.J.W. is an R. Douglas Wright Fellow, and G.R.S. is a National Health and Medical Research Council of Australia Research Fellow and a Canadian Research Chair in Metabolism, Obesity and Type 2 Diabetes. B.E.K. is an Australian Research Council Federation Fellow.

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