Myostatin-deficient mice exhibit reduced insulin resistance through activating the AMP-activated protein kinase signalling pathway
- First Online:
- Cite this article as:
- Zhang, C., McFarlane, C., Lokireddy, S. et al. Diabetologia (2011) 54: 1491. doi:10.1007/s00125-011-2079-7
- 1.7k Downloads
Myostatin-null mice (Mstn−/−) have reduced body fat and increased tolerance to glucose. To date the molecular mechanisms through which myostatin regulates body fat content and insulin sensitivity are not known. Therefore, the aim of the current study was to identify signalling pathways through which myostatin regulates insulin sensitivity.
Wild-type (WT) mice and Mstn−/− mice were fed either a control chow diet or a high fat diet (HFD) for 12 weeks. Glucose tolerance testing and insulin stimulated glucose uptake by M. extensor digitorum longus (EDL) were used as variables to determine insulin sensitivity. Quantitative PCR, Western blotting and enzyme assays were used to monitor AMP-activated protein kinase (AMPK) levels and activity.
Mstn−/− mice exhibited reduced fat accumulation and peripheral insulin resistance when compared with WT mice, even when they were fed an HFD. Furthermore, treatment with a myostatin antagonist also increased insulin sensitivity during HFD. Consistent with increased insulin sensitivity, we also detected elevated levels of GLUT4, AKT, p-AKT and insulin receptor substrate-1 in Mstn−/− muscles. Molecular analysis showed that there is increased expression and activity of AMPK in Mstn−/− muscles. Furthermore, we also observed an increase in the AMPK downstream target genes, Sirt1 and Pgc-1α (also known as Ppargc1a), in skeletal muscle of Mstn−/− mice.
We conclude that myostatin inactivation leads to increased AMPK levels and activity resulting in increased insulin sensitivity of skeletal muscle. We propose that, by regulating AMPK in skeletal muscle and adipose tissues, myostatin plays a major role in regulating insulin signalling.
Activin type IIB receptor
Soluble activin type IIB receptor
AMP-activated protein kinase
Carnitine palmitoyl transferase-1
M. extensor digitorum longus
Glucose tolerance test
High fat diet
Institutional Animal Care and Use Committee
Insulin tolerance test
Quantitative real-time PCR
Peroxisome proliferator-activated receptor-gamma coactivator-1alpha
Solute carrier family 2 (facilitated glucose transporter)
Tris buffered saline Tween 20
White adipose tissue
Myostatin (MSTN) is a growth and differentiating factor that belongs to the TGF-β superfamily. While a high level of Mstn gene expression is detected in skeletal muscle , a low level of Mstn expression is also detected in other tissues including adipose tissue  and heart . Inactivation of the Mstn gene in mice  or mutations in the bovine [3, 4], ovine  and human  Mstn gene result in a similar phenotype of increased muscle growth. Therefore, it is believed that MSTN is a negative regulator of skeletal myogenesis and its function is highly conserved during evolution. Functionally, MSTN not only regulates the proliferation and differentiation phases of myoblast growth [7, 8], but also controls the activation and further proliferation of muscle stem cells, the satellite cells . MSTN has been shown to bind to the activin type IIB receptor (ActIIBR) and, therefore, either overproduction of a dominant negative form of ActIIBR in muscle  or treatment of mice with a soluble form of ActIIBR  (sActIIBR) have been shown to result in increased musculature.
The physiological effects of Mstn deletion are not restricted to skeletal muscle. A number of studies have reported significant decreases in the amount of adipose tissue in association with loss of MSTN function [12–14]. A flurry of recent research also suggests that loss of, or reduction in, MSTN levels leads to increased insulin sensitivity. In fact, suppression of MSTN function in mice, through transgenic overproduction of the MSTN propeptide, results in resistance to high fat dietary-induced obesity and insulin resistance . This finding is further supported by studies on insulin resistant human participants, which show increased levels of MSTN in muscle and plasma . Interestingly, short-term administration of recombinant MSTN protein has also been shown to induce insulin resistance in mice . In extremely obese human participants it was further discovered that there were elevated levels of MSTN protein in both muscle and plasma samples . In addition, injection of MSTN-specific antibodies along with exercise results in significantly increased insulin sensitivity . Recently, Guo et al. showed that inhibition of MSTN signalling, specifically in adipose tissue, had no effect on glucose and insulin tolerance . Contrary to this, muscle-specific inactivation of MSTN resulted in increased lean mass and improved glucose metabolism on both standard and high fat diets (HFDs) . Although it is clear that lack of MSTN affects glucose uptake and metabolism, the specific mechanisms and signalling events in skeletal muscle and adipose tissue that help overcome insulin resistance in Mstn−/− mice are not clearly understood.
Recently, a series of investigations has delineated the role of AMPK (also known as PRKAB1) in insulin sensitivity and resistance. AMPK is a heterotrimeric serine/threonine kinase, which consists of a catalytic α-subunit and regulatory β- and γ-subunits [19–21]. In skeletal muscle, α2 [22, 23] is the major catalytic isoform produced, which forms a heterotrimeric complex with β2 and γ3 subunits. AMPK acts as a metabolic switch, which becomes activated by acute increases in the cellular AMP/ATP ratio, such as occurs following exercise, or during conditions such as hypoxia, ischaemia or osmotic stress . When intracellular ATP levels are low, AMPK switches off ATP-consuming processes, such as glycogen, fatty acid, and protein synthesis pathways, and activates alternative pathways for ATP regeneration, which include glucose transport, glycolysis and fatty acid oxidation. Activation of AMPK has also been shown to improve insulin sensitivity resulting in enhanced insulin-mediated glucose uptake in skeletal muscle [25–27].
In the current manuscript we have identified the molecular mechanisms that are responsible for the increased insulin sensitivity in Mstn−/− mice. Here for the first time we show that there is increased abundance of AMPK in skeletal muscle, adipose tissue and liver, leading to an overall increase in glucose uptake and reduced fat accumulation in Mstn−/− mice, despite being fed on an HFD. This in turn leads to the increased sensitivity to insulin in Mstn−/− mice. We further show that treatment with an MSTN antagonist also reduces insulin resistance in mice fed on an HFD.
Wild-type (WT) C57BL/6 mice were obtained from the Centre for Animal Resources (National University of Singapore, Singapore). Mstn−/− mice (C57BL/6 background) were gifted from Se-Jin Lee (Johns Hopkins University, Baltimore, MD). The Institutional Animal Care and Use Committee (IACUC), Singapore, approved all experiments involving animals. Seven-week-old male Mstn−/− mice and WT littermates were housed in groups and maintained on standard chow diet (CD) at a constant temperature (20°C) under a 12 h/12 h artificial light/dark cycle with unlimited access to water until commencement of the trial. Two groups, consisting of five WT and five Mstn−/− mice each, were either fed HFD (45% energy from fat, 21.3% energy from protein, 23.6% energy from fat and 41.2% energy from carbohydrates [58V8, TestDiet, Richmond, IN, USA]) or CD (10% energy from fat; 17.3% energy from protein, 4.3% energy from fat and 67.4% energy from carbohydrates [58Y2, TestDiet]) for 12 weeks. For MSTN antagonist studies, 7-week-old male WT mice were fed the same CD or HFD (see above) for a period of 12 weeks together with injection (i.p.) of either saline or sActIIBR MSTN antagonist (5 μg/g body weight) three times a week (see electronic supplementary material [ESM] for details of purification). Body weight and energy intake (energy intake is equal to food consumption multiplied by metabolisable energy content) were measured twice a week in all trial mice. Mice were killed 1 week after GTT and insulin tolerance test (ITT) analysis, and were not fasted prior to death.
RNA extraction, quantitative real-time PCR (qPCR) and Western blotting
Insulin and glucose tolerance testing
Glucose transport in muscle
M. extensor digitorum longus (EDL) were rapidly removed from killed mice and were pre-incubated in Krebs–Ringer bicarbonate (KRB) buffer containing 2 mmol/l pyruvate for 20 min, followed by a further incubation in KRB buffer containing 3,129.3 pmol/l insulin for 30 min. The muscles were then incubated in KRB buffer containing the same concentration of insulin, 1 mmol/l 2-deoxy-d-[3H]glucose (55,500 Bq/ml) and 7 mmol/l d-[14C]mannitol (11,100 Bq/ml) at 30°C for a further 10 min. The level of glucose transport was subsequently measured as previously described [29, 30].
See ESM for details of the kits and reagents used for total serum insulin, total cholesterol, triacylglycerol and serum NEFA measurement.
A modified protocol from the Animal Models of Diabetic Complications Consortium  was used to measure AMPK activity in protein lysates from muscle tissue. Briefly, after muscle homogenisation, lysates containing 1 mg of muscle protein were centrifuged in a TLA 120.1 rotor (Beckman Coulter, Brea, CA, USA) at 37,000 rpm (48,000×g) for 30 min at 4°C. The supernatant fraction was then transferred to another centrifuge tube on ice and stirred together with 144 mg of ammonium sulphate, per ml solution, for an incubation period of 30 min. The sample was then centrifuged at 37,000 rpm for 30 min at 4°C, the supernatant fraction was discarded, and the pellet resuspended in 100 μl of homogenisation buffer.
The SAMS peptide (His-Met-Arg-Ser-Ala-Met-Ser-Gly-Leu-His-Leu-Val-Lys-Arg-Arg) utilised in this study was synthesised commercially. The reaction was started by incubating the resuspended ammonium sulphate pellet into a mixture of water, SAMS peptide and radioactive ATP (to 100 μl of required working assay mix add 2 μl of 370,000 Bq/ml P32-γATP) for 10 min at 37°C. The reaction was terminated by pipetting 15 μl of incubation mix onto a 15 × 15 mm piece of P81 Whatman filter paper, allowing at least 20 s for the solution to absorb into paper, then dropping into 150 ml of 1% phosphoric acid. Finally the filter paper was washed six times, 5 min each, in 150 ml of 1% phosphoric acid, followed by a final wash in 50 ml acetone. After drying, the filter paper was measured on a scintillation counter (1450 LSC and luminescence counter; PerkinElmer Life Sciences, Waltham, MA, USA).
The data were analysed using two-way ANOVA and robust regression analysis. Not all the data were normally distributed. Unlike the conventional regression model based on ordinary least square (OLS), the robust model estimates with an iterative reweighted least square (IRWLS) algorithm and generates more reliable estimates about the influence of mouse genotype and diet type on the outcome. The proposed robust model has several advantages over the conventional approach in data analysis. First, it analyses the skewed outcome directly without the need to perform back-transformation for interpretation. This is desirable because arithmetic transformation may not necessarily induce normality in the skewed outcome. Second, it does not discard any observations in the process even if they are confirmed outliers. Finally, it does not impose any strict distributional assumption on the outcome like the generalised linear model (GLM), as there is no precise information about the outcome’s probabilistic characteristics [32, 33].
Mstn−/− mice exhibit a lean phenotype
Decreased circulatory cholesterol levels in Mstn−/− mice
To determine whether the reduced fat pad weight in Mstn−/− mice occurred in conjunction with a decrease in lipid levels, we analysed the circulatory levels of cholesterol, triacylglycerol and NEFA. In WT mice fed HFD, there was a 110% increase in circulatory cholesterol levels when compared with WT CD fed controls, while in Mstn−/− mice, we observed an increase of 50% in serum cholesterol levels upon HFD feeding when compared with Mstn−/− CD fed controls (Fig. 1g). Therefore, absence of Mstn appears to result in reduced circulating cholesterol levels during HFD feeding. In contrast to the reduced cholesterol levels, we observed no significant change in triacylglycerol and NEFA levels between WT and Mstn−/− mice fed either CD or HFD (Fig. 1h, i).
Lack of MSTN reduces insulin resistance
Increased glucose uptake in the absence of MSTN
To determine whether or not disruption of MSTN has a cell autonomous effect on glucose uptake by muscle, we measured glucose uptake in EDL muscle from both WT and Mstn−/− mice. The results show that even at the basal level (unstimulated by insulin), there is a 2.7-fold increase in glucose uptake by EDL muscle from Mstn−/− mice compared with WT mice. As expected, insulin was able to stimulate glucose uptake twofold in WT EDL muscle, and only to a minor extent in Mstn−/− EDL muscle (Fig. 2c). To determine the molecular basis for the improved glucose uptake in Mstn−/− muscle, we next estimated insulin levels using ELISA. The results show that there was no significant difference in systemic insulin levels between WT and Mstn−/− on CD. When challenged with HFD, we observed an increase in circulatory insulin levels in the WT mice; however, in contrast, no significant change in systemic insulin levels was observed in Mstn−/− mice fed on HFD (Fig. 2d).
Lack of MSTN increases insulin signalling
Interfering with MSTN signalling increases tolerance to glucose
To ascertain whether or not interfering with MSTN signalling postnatally has a similar effect to that observed in Mstn−/− mice, WT mice fed on HFD were treated with sActIIBR, which has been previously shown to act as a potent antagonist to MSTN activity [34, 35]. Treatment with sActIIBR reduced the insulin resistance seen in mice fed on a HFD. In a GTT test, we observed a significant elevation in blood glucose levels in mice treated with HFD; however, treatment with sActIIBR increased insulin sensitivity and enhanced glucose disposal, despite the fact that the mice were fed HFD. Thus, as can be seen in Fig. 2e, glucose levels were similar between mice fed on regular CD and HFD fed mice treated with sActIIBR (Fig. 2e).
Increased levels of AMPK in Mstn−/− mice
In contrast to WT mice, Mstn−/− mice had reduced adipose tissue and failed to accumulate fat in adipose tissue even on an HFD (Fig. 1f). Prolonged high caloric intake leads to metabolic overload and increased triacylglycerol levels, resulting in adipocyte hypertrophy and increased inflammatory response . Macrophage accumulation in turn leads to secretion of pro-inflammatory cytokines such as TNF-α, resulting in further requirement of macrophages and impaired triacylglycerol deposition . It is noteworthy to mention that in Mstn−/− mice there are reduced levels of circulating TNF-α , raising the possibility that this may be partly responsible for the reduced adipose tissue observed in the absence of Mstn during HFD feeding (Fig. 1f).
Absence of Mstn results in increased Ampk mRNA expression in skeletal muscle, adipose tissue and liver (Fig. 4). Furthermore, we observed increased protein levels of the AMPKγ subunit in skeletal muscle, and increased phosphorylation of the catalytic AMPKα subunit in skeletal muscle, WAT and liver from Mstn−/− mice (Fig. 5a, b). Consistent with enhanced AMPK production we also observed enhanced activity of AMPK in skeletal muscle (Fig. 5c). AMPK plays a critical role in energy metabolism. Increased AMPK activation leads to phosphorylation and thus inhibition of ACC enzyme activity (Fig. 5a). Reduced ACC enzyme activity leads to reduced malonyl-CoA expression, which further limits fatty acid synthesis and increases fatty acid oxidation. In addition to ACC, it has recently been demonstrated that AMPK controls the expression of genes involved in energy metabolism by modulating another metabolic sensor, the NAD+-dependent type III deacetylase, SIRT1 . AMPK enhances SIRT1 activity by increasing NAD+ levels resulting in deacetylation and further activation of SIRT1 targets including PGC-1α, leading to increased expression of mitochondrial and lipid metabolism genes . Thus, it is quite possible that the acute actions of AMPK in lipid metabolism in Mstn−/− mice could also be signalled via the decreased acetylation of PGC-1α observed in Mstn−/− mice (Fig. 5h). However, there is an anomaly; as others and we have shown [39–41], in Mstn−/− mice there are a greater proportion of fast twitch glycolytic muscle fibres (type IIB). Therefore, one would expect muscles from Mstn−/− mice to be more glycolytic rather than oxidative in nature, as is described here. Interestingly, although predominately fast glycolytic in nature, unpublished data from our lab suggests that there is increased mitochondrial biogenesis, due to fission, in Mstn−/− fast twitch fibres from the EDL muscle (S. Lokireddy, C. McFarlane, M. Sharma and R. Kambadur, unpublished data). Furthermore, Mstn−/− muscles on average have higher numbers of muscle fibres. Therefore, we speculate that a combination of increased number of mitochondria due to augmented biogenesis and increased muscle fibre number due to hyperplasia are the main driving forces behind the fast glycolytic fibres of Mstn−/− muscles adopting a more oxidative nature, in at least as far as ATP generation is concerned. The benefits of increased fast glycolytic muscle fibre content to metabolic syndrome are clear. Specifically, Izumiya et al. have previously demonstrated that overexpression of a constitutively active form of AKT1 in muscle results in dramatic hypertrophy of type IIB fast glycolytic muscle fibres. Moreover, the increased type IIB muscle mass resulted in resistance to diet-induced obesity, improved sensitivity to insulin and enhanced hepatic fatty acid oxidation .
In addition to the reduced adiposity, Mstn−/− mice and, for that matter, mice treated with sActIIBR, the MSTN antagonist, have improved tolerance to glucose and increased sensitivity to insulin (Fig. 2). This increased sensitivity to insulin appears to be due to increased insulin signalling rather than increased insulin secretion, since gene expression and Western blot analysis confirms that there are increased GLUT4 levels and increased phosphorylation of AKT (Fig. 3). We propose that increased insulin signalling in Mstn−/− mice is due to enhanced AMPK levels and activity. Acute activation of AMPK has been previously shown to result in increased insulin-mediated glucose uptake by increasing GLUT4 levels. Also, prior incubation of epitrochlearis muscles with the AMPK agonist 5-aminoimidazole-4-carboxamide ribonucleoside (AICAR) results in a two-fold increase in insulin stimulated glucose transport . Furthermore, AICAR treatment of Zucker rats for 8 weeks improved insulin sensitivity . Moreover, it has been recently shown that ablation of AMPK activity in skeletal muscle exacerbates insulin resistance induced by HFD feeding . Therefore, these prior studies support the hypothesis that enhanced AMPK activity leads to increased insulin signalling in Mstn−/− mice.
In contrast to our results, Chen et al. recently showed that MSTN treatment increases AMPK abundance in vitro . The inconsistencies between our results and the observations made by Chen et al. could be due to the model system used. The addition of MSTN protein (Chen et al.) could result in a negative energy balance via MSTN-induced cachexia [28, 45], leading to a high AMP/ATP ratio resulting in AMPK activation. It is important to highlight that we have utilised an HFD model to study the regulation of glucose metabolism by MSTN and, much like previously published reports [12, 38], we find reduced insulin resistance and enhanced glucose uptake, which is consistent with the increase in AMPK activity shown in the present manuscript.
In summary, the results presented here demonstrate that lack of MSTN increases insulin signalling and enhances insulin sensitivity in skeletal muscle by increasing AMPK signalling. Furthermore, these results suggest that MSTN antagonists would have utility in resisting obesity and, moreover, in increasing insulin sensitivity during obesity (Fig. 6).
We would like to thank the Agency for Science, Technology and Research (A*STAR) for funding this project. We would also like to thank J. Swain, Singapore Institute for Clinical Sciences, Singapore, for her comments on the manuscript and R. Choo, Singapore Institute for Clinical Sciences, Singapore, for his invaluable help with statistical analysis of the data presented in this manuscript.
Duality of interest
The authors declare that there is no duality of interest associated with this manuscript.