Genetic Disruption of Myostatin Reduces the Development of Proatherogenic Dyslipidemia and Atherogenic Lesions In Ldlr Null Mice

C57BL/6 mice with homozygous deletion of Ldlr (Jackson Laboratory, Bar Harbor, ME) were cross-bred with C57BL/6 mice with homozygous deletion of Mstn. Their offsprings were bred >10 times to generate Mstn^−/−/Ldlr^−/− double-knockout mice. All mice used in this study were males. Mice were provided food and water ad libitum and maintained on a 12-h light/dark cycle. Mice were fed normal rodent diet (TD#2018, Harland Teklad, Madison, WI) until 4 months of age, then fed with high-fat/high-cholesterol diet (HFD) (TD#94059, Harland Teklad). The HFD contained 15.8% w/w fat (primarily from cocoa butter), 1.25% w/w cholesterol, 41.2% w/w carbohydrate, and 19.7% w/w protein. Food consumption was measured using powdered HFD provided in glass jars with perforated metal tops to prevent scattering. Study protocols were approved by the Boston University's Institutional Animal Care and Use Committee.

Whole body fat and muscle mass were determined by nuclear magnetic resonance (NMR; EchoMRI-100, Echo Medical System, Houston, TX). Visceral and subcutaneous fat imaging were performed using low-energy X-ray micro–computed tomography (CT) scanner (LaTheta LCT-100A, Echo Medical System). Oxygen consumption and respiratory exchange ratio rates were measured using Oxymax system (Columbus Instruments, Columbus, OH) as previously described (11).

Mice were anesthetized using 1–2% (v/v) isoflurane and 1–2 L/min oxygen flow through an anesthesia monitoring system (Smiths Medical PM, Waukesha, WI). Respiration was monitored by respiration sensor pillow (SA Instruments, Stony Brook, NY). Imaging was performed using a Bruker Avance 500 vertical bore spectrometer with 11.7 T (Bruker, Billerica, MA). Data acquisition and reconstruction were performed with the ParaVision 3.0.2. 3D magnetic resonance angiography (MRA) and were obtained using the FLASH_3D_ANGIO pulse sequence (flip angle = 20°, TE/TR = 2.18 ms/20 ms, FOV = 15 mm × 15 mm × 15 mm). Four averages were obtained in each scan.

Preparation of aortas and quantification of atheromatous lesions were performed in animals after 12 weeks of HFD feeding, using procedures previously described (12). The heart with aorta was embedded in optimal cutting temperature (OCT) (Sakura Finetech USA, Torrance, CA), and serial 10 μm-thick cytosections of aortic root were stained with Oil Red O by the Rodent Histopathology Core at Harvard Medical School (Boston, MA).

Blood glucose and ketone concentrations were determined using Glucometer Elite XL (Bayer, Tarrytown, NY) and Precision Xtra (Abbott Laboratories, Bedford, MA), respectively. Plasma insulin and glucagon were measured using ultrasensitive ELISA (Alpco, Salem, NH). For lipid analyses, fasting (food removed for 12 h) and nonfasting (food removed for 3 h) blood samples were collected from the retro-orbital plexus under isoflurane anesthesia. Lipid analyses were performed using procedures previously described (13). Cholesterol and triglyceride distribution of the lipoproteins were performed by Cardiovascular Specialty Laboratories (Atlanta, GA) using fast-performance liquid chromatography (14). Fractions 2–6 contain VLDL; fractions 7–11 contain LDL; and fractions 14–17 contain HDL.

For glucose tolerance tests, fasting mice were given d-glucose (1 g/kg) by intraperitoneal injection. For insulin tolerance tests, nonfasting mice were given insulin (0.75 IU/kg) (Eli Lilly, Indianapolis, IN) by intraperitoneal injection. Blood glucose was measured using Glucometer Elite XL (Bayer).

Hyperinsulinemic-euglycemic clamps were performed by a modification of a described procedure (15). Briefly, right jugular vein of killed mice (80–10 mg/kg ketamine-xylazine i.p.) was catheterized with heparin-coated MRE-025 tubing (Becton Dickinson, Franklin Lakes, NJ). Animals were allowed to recover for 1–2 days. Insulin (10 mU · kg⁻¹ · min⁻¹) containing 0.3% BSA was infused at the rate of 1 μl/min using a syringe pump (HA11D, Harvard apparatus, Holliston, MA). At 5-min intervals, ∼2 μl of tail vein blood was drawn to measure blood glucose. Dextrose (25%) was infused at a rate sufficient to maintain blood glucose of 140 mg/dl. Average glucose infusion rate (GIR) was measured over the final 30 min of the 2-h clamp. Insulin was measured in blood samples from the last three collections.

Muscle uptake of the metabolically inert glucose analog, [³H]-2-deoxyglucose (DOG), was performed in mice at 30 min before the end of 2-h hyperinsulinemic-euglycemic clamp experiments, with modification to methods described previously (15). Briefly, DOG (30 μCi) was administered intraperitoneally, and mouse hind-limb muscles were harvested, denatured, and the eluates were quantified using liquid scintillation analyzer (TriCarb 3100TR, PerkinElmer, Shelton, CT). The rate of glucose uptake was calculated by dividing the muscle DOG uptake over a 30-min interval (cpm · kg⁻¹ · min⁻¹) by the mean blood glucose–specific activity (glucose cpm/mg).

Liver tissues were embedded in OCT. Serial 10 μm-thick cytosections were prepared at the Rodent Histopathology Core at Harvard Medical School (Boston, MA) and were stained with Oil Red O and hematoxylin-eosin.

Livers were lyophilized using Freeze Dry System (Freezone, Labconco, Fort Scott, KS) to obtain liver dry weight. Liver lipids were extracted by the Folch method (16). Liver cholesterol and triglycerides were measured using Infinity Cholesterol (Thermo Electron, Pittsburgh, PA) and T2449 Triglyceride Reagent (Sigma Aldrich, St. Louis, MO), respectively.

Fresh liver tissue was homogenized in 250 mmol/l sucrose, 0.1 mmol/l EDTA, 50 mmol/l KCl, and 10 mmol/l HEPES (pH 7.4) and centrifuged at 1,000 × g for 5 min. The supernatant was centrifuged at 10,000 × g for 15 min, and pellets were dissolved in 50 mmol/l KCl, 70 mmol/l sucrose, 3.6 mmol/l MgCl₂, 7.2 mmol/l K₂HPO₄, and 36 mmol/l Tris-HCl (pH 7.4). Protein concentrations were measured using Bradford reagent (Bio-Rad, Hercules, CA). In addition, 5 μCi/ml [³H]-palmitate was incubated with 1 mg of mitochondria in preincubation buffer with 2 mmol/l KCN at room temperature for 30 min. Reaction was subsequently incubated for 60 min in 5% perchloric acid, and supernatants were counted in a liquid scintillation counter.

VLDL secretion rate was measured by modification of a previously described method (17). Briefly, fasted mice were injected with Triton WR 1339 (500 mg/kg; Sigma Chemical, St. Louis, MO) via tail vein to block lipolysis. Blood samples were collected at baseline, 90 min and 180 min after injection. Plasma apoB100 was determined by Western blot analysis using anti-apoB antibody (Santa Cruz Biotechnology, Santa Cruz, CA). Hepatic VLDL secretion was calculated by the percentage increase in plasma apoB100 from baseline.

One microliter of plasma per lane was analyzed after treatment with sample buffer containing SDS and β-mercaptoethanol, followed by separation in 4–15% Tris-HCl gradient gel (Bio-Rad, Hercules, CA). Liver and muscle tissue samples were extracted and processed as previously described (12). All primary, except anti–fatty acid synthase antibodies (Cell Signaling Technology, Danvers, MA), and secondary antibodies were purchased from Santa Cruz Biotechnnology.

Total RNA was extracted from frozen liver using RNeasy mini kit (Qiagen, Valencia, CA). For reverse transcription, 1 μg of the total RNA was converted to first-strand complementary DNA in 20 μl reactions using AffinityScript Q-PCR cDNA Synthesis Kit (Agilent Technologies, La Jolla, CA), which was subsequently diluted five times. For PCR sample preparation, 5 μl of cDNA was mixed in 20 μl reaction volume with 10 μmol/l primer and SYBR master enzyme mix (SABiosciences, Frederick, MD). The reaction was initiated at 94°C for 10 min, followed by 40 cycles through 94°C × 15 s and 60°C × 1 min. All reactions were performed in duplicate. All CT values were in the range of 20–30 cycles. Amplification curves were analyzed using SDS 1.9.1 software (Applied Biosystems, Foster City, CA). Hprt1 controls provided relative gene expression levels. Refer to supplementary Table 1 (available in an online appendix at http://diabetes.diabetesjournals.org/cgi/content/full/db09-0349/DC1) for primers.

Experimental results are shown as means ± SE. The mean values from two independent groups were compared by using Student's t test for independent samples. Differences among multiple groups were assessed using ANOVA models. If ANOVA revealed a significant overall effect, then individual groups were compared by using Tukey's procedure. Longitudinal data from multiple groups of animals were compared by using repeated measures ANOVA, with a time-in-treatment factor and a treatment group factor. All statistical tests were one-tailed and P < 0.05 were considered significant.

Ldlr^−/− mice with varying Mstn genotypes, wild type (Mstn^+/+/Ldlr^−/−), heterozygous deletion (Mstn^+/−/Ldlr^−/−), and homozygous deletion (Mstn^−/−/Ldlr^−/−), were fed HFD for 11–12 weeks. Mstn^+/+/Ldlr^−/− controls demonstrated substantial gains in body weight, whole body fat mass (measured by NMR), and visceral and subcutaneous fat mass (measured by micro-CT scan). In contrast, Mstn^−/−/Ldlr^−/− mice resisted weight gain and accumulation of whole body and visceral fat mass in response to HFD (Fig. 1,A and B, supplementary Fig. 1A and B, available in an online appendix). In agreement with the in vivo NMR data, the wet weights of the major fat depots (inguinal, epididymal, perirenal, as well as the interscapular brown fat) were significantly lower in the Mstn^−/−/Ldlr^−/− mice, as compared with wild-type, Mstn^+/+/Ldlr^−/−, and Mstn^+/−/Ldlr^−/− mice (Fig. 1 C). Thus, Mstn^−/−/Ldlr^−/− mice are resistant to fat accumulation in all fat depots including the visceral and subcutaneous fat depot.

As expected, the Mstn^−/−/Ldlr^−/− mice had significantly greater skeletal muscle mass, as indicated by the twofold higher mass of the gastrocnemius and the quadriceps femoris muscle groups (Fig. 1,D; supplementary Fig. 1C and D, available in an online appendix) than in the Mstn ^+/+/Ldlr^−/− controls. Throughout the course of HFD, Mstn^−/−/Ldlr^−/− mice maintained significantly greater lean mass (assessed by NMR) compared with other genotype groups (Fig. 1 D; supplementary Fig. 1C, available in an online appendix). The soleus muscle, a primarily oxidative muscle type, was not statistically significantly affected by Mstn disruption (supplementary Fig. 1E, available in an online appendix). Ldlr^−/− mice with heterozygous deletion of Mstn only displayed modest increase in muscle mass but have similar fat mass accumulation compared with Mstn^+/+/Ldlr^−/− controls when fed HFD.

Despite the remarkable effects of Mstn deletion on body composition, neither the energy intake nor energy expenditure differed significantly among the various groups on normal rodent diet (data not shown) or on HFD (supplementary Fig. 2A, available in an online appendix). Ambulatory activity levels on HFD were assessed by infrared beam interruption, and notably, the horizontal ambulatory activity of Mstn^−/−/Ldlr^−/− double-knockout group was 38% lower than that of controls (supplementary Fig. 2B, available in an online appendix). However, despite this reduction in activity, analysis of whole body O₂ consumption (Vo₂) showed that energy expenditure was similar between the two groups (supplementary Fig. 2C, available in an online appendix). If the data were expressed as a function of lean mass weight, Mstn^−/−/Ldlr^−/− actually had lower Vo₂ rate (13%) compared with controls. Furthermore, the respiratory exchange ratios were ∼0.85 in Ldlr^−/− mice with or without Mstn disruption, reflecting mixed utilization of carbohydrate and fatty acid oxidation in both animal groups (supplementary Fig. 2D, available in an online appendix). Thus, the resistance to fat accumulation in Mstn^−/−/Ldlr^−/− mice cannot be explained based on reduced food intake, increased whole-body O₂ consumption, or preference for fuel disposition.

To determine whether deletion of Mstn mitigates the development of vascular lesions, we first used in vivo MRA to evaluate lumen occlusion of the brachiocephalic artery, one of the most lesion-prone sites (Fig. 2,A). After 8 weeks of HFD, Mstn^+/+/Ldlr^−/− mice showed reduced lumen size and increased asymmetry (Fig. 2,B–D); in contrast, brachiocephalic arterial lumen of Mstn^−/−/Ldlr^−/− maintained patency without significant occlusion, as indicated by greater symmetry and significantly larger (1.7-fold) lumen area (Fig. 2 C and D).

After 12 weeks of HFD, animals were killed and aortas were dissected. Mstn^+/+/Ldlr^−/− mice developed large plaques that were eccentric (partial circumference of arterial wall) and dispersed along the arch and in the descending aorta (Fig. 2,E). Staining with Oil Red O revealed that the lesions were lipid rich (Fig. 2,E). Analysis of the en face aorta demonstrated that atheromatous lesion area in Mstn^−/−/Ldlr^−/− mice was 41% lower than in Mstn^+/+/Ldlr^−/− controls (Fig. 2,F and G). In correlation with the similar fat mass accumulation (Fig. 1,B and C), Mstn^+/−/Ldlr^−/− mice had similar lesion area compared with Mstn^+/+/Ldlr^−/− controls (Fig. 2 G), suggesting that heterozygous deletion of Mstn is not sufficient in providing protection against either fat accumulation or atherogenesis. The negative control, Mstn^+/+/Ldlr^+/+ mice group, had no detectable lesions. Thus, Ldlr^−/− mice with Mstn deletion were protected against the development/progression of atheromatous lesions.

We next investigated the mechanisms by which disruption of Mstn retards atherogenesis progression. As circulating lipids are important contributors to atherogenesis, we determined the changes in plasma lipids and lipoproteins.

Even before HFD induction, Mstn^−/−/Ldlr^−/− mice had significantly lower plasma triglycerides, free fatty acids (FFA), and total cholesterol levels compared with Mstn^+/+/Ldlr^−/− mice (Fig. 3,A). After 10 weeks of HFD, Mstn^+/+/Ldlr^−/− controls displayed marked increase in plasma triglycerides (3.0-fold), FFA (2.3-fold), and cholesterol (3.8-fold) levels compared with baseline; the changes in plasma lipids were significantly greater than those observed in Mstn^−/−/Ldlr^−/− mice after HFD. Plasma triglycerides, FFA, and cholesterol levels after HFD were significantly lower in Mstn^−/−/Ldlr^−/− than in Mstn^+/+/Ldlr^−/− controls (Fig. 3 B), suggesting that Mstn deletion protects against the development of proatherogenic dyslipidemia in Ldlr^−/− mice fed HFD.

We fractionated plasma lipoproteins by using fast-phase liquid chromatography. Plasma total cholesterol, VLDL-cholesterol, and LDL cholesterol levels were lower in Mstn^−/−/Ldlr^−/− mice than in Mstn^+/+/Ldlr^−/− controls (Fig. 3,C). Triglyceride contents of VLDL (VLDL-TG) and LDL fractions (LDL-TG) were also significantly lower in Mstn^−/−/Ldlr^−/− mice than controls (Fig. 3 D).

ApoB is an important predictor of cardiovascular risk; some patients with elevated apoB100 levels are at increased risk of cardiovascular disease despite having target cholesterol levels (18,19). We analyzed the plasma apoB100 contents in our mice. Western blot analysis of plasma samples of Mstn^+/+/Ldlr^−/− mice showed a significant increase in apoB100 after 11 weeks of HFD, and this diet effect was markedly attenuated in Mstn^−/−/Ldlr^−/− mice (Fig. 3,E). We also measured apoA1, the scaffold protein for HDL. The ratio of apoB100 to apoA1, which is stoichiometrically related to the ratio of non-HDL to HDL particles, was lower in Mstn^−/−/Ldlr^−/− mice compared with controls (Fig. 3 E), suggesting a more favorable lipoprotein profile in Mstn^−/−/Ldlr^−/− mice.

Diet-induced obesity is strongly associated with deterioration of insulin sensitivity, clinically evidenced by increased circulating insulin, glucose, triglycerides, and FFA (20). Insulin also serves as an important regulator of adipose lipolysis and hepatic VLDL production. Dysregulation of these metabolic functions of insulin could contribute to hyperlipidemia and accelerate atherosclerosis in mice fed with HFD (21,22). Accordingly, we assessed the effect of Mstn deletion on insulin sensitivity in Ldlr^−/− mice.

In addition to having lower plasma triglycerides, FFA, VLDL, and apoB concentrations, Mstn^−/−/Ldlr^−/− mice also had significantly lower fasting blood glucose and insulin levels. These data suggested that Mstn^−/−/Ldlr^−/− mice were more insulin sensitive than Mstn^+/+/Ldlr^−/− controls even when fed HFD (Fig. 4,A–C). Plasma insulin- to-glucagon ratio was also significantly lower (30% in Mstn^−/−/Ldlr^−/− mice). Glucose tolerance test did not reveal significant differences between rate of glucose disposal (supplementary Fig. 3A, available in an online appendix), but this could be because of the compensatory increase in fasting insulin in Mstn^+/+/Ldlr^−/− (Fig. 4 B). When mice were injected with a fixed dose of insulin during insulin tolerance tests, Mstn^−/−/Ldlr^−/− mice maintained significantly lower blood glucose than their Mstn^+/+/Ldlr^−/− littermates (supplementary Fig. 3B, available in an online appendix). Furthermore, Ldlr^−/− mice with Mstn deletion had significantly lower phosphoenolpyruvate carboxykinase one (Pepck) expression (supplementary Fig. 3C, available in an online appendix), the rate-limiting hepatic gluconeogenic enzyme (23), consistent with improved insulin sensitivity.

To confirm our findings of improved insulin sensitivity in Mstn^−/−/Ldlr^−/− mice, we performed hyperinsulinemia-euglycemic clamp studies, the accepted gold standard for assessing insulin sensitivity (24). Mstn^−/−/Ldlr^−/− required significantly higher GIR to maintain euglycemia (Fig. 4,D and E) than in controls. These data demonstrate that Mstn deletion improves whole body insulin sensitivity through the enhancement of insulin-stimulated glucose disposal in peripheral tissues. It is possible that the increased GIR is simply because of greater muscle mass and the resulting increase in metabolic demand of the double-knockout mice. To explore this further, we measured the rates of [³H]-2-deoxyglucose uptake in skeletal muscle during the clamp procedure. In the Mstn^−/−/Ldlr^−/− mice, glucose uptake in quadriceps muscle was significantly higher compared with that of the Mstn^+/+/Ldlr^−/− controls (Fig. 4,F) in a similar hyperinsulinemic state (Fig. 4,G). Furthermore, phosphorylation of serine-473 in Akt, the pleiotropic kinase essential for many metabolic actions of insulin, was significantly higher in muscle tissues of Mstn^−/−/Ldlr^−/− mice compared with that of the controls. Phosphorylation of serine-21/9 in GSK3, one of Akt substrates, was also significantly higher, providing further evidence of improved insulin action in the skeletal muscle (Fig. 4 H). Together, these data demonstrate that HFD induces an insulin-resistant state in Mstn^+/+/Ldlr^−/− mice and that disruption of Mstn prevents the development of insulin resistance, as consistent with other reports (9,10).

The liver plays a leading role in systemic lipid homeostasis (25). In response to HFD feeding, the livers of Mstn^+/+/Ldlr^−/− mice displayed substantial fat accumulation (Fig. 5,A). In contrast, Mstn^−/−/Ldlr^−/− mice had significantly lower hepatic triglyceride content, as well as total wet and dry liver weights (Fig. 5,A; supplementary Fig. 4A, available in an online appendix). This protection against diet-induced hepatic fat infiltration in Mstn^−/−/Ldlr^−/− mice was associated with significant reduction in plasma adipose-derived FFA (Fig. 3,A), glucose, and insulin (Fig. 4,A and B), as well as with diminished expression of lipogenic master gene, sterol regulatory element–binding transcription factor 1 (Srebf1), and its key downstream target, fatty acid synthase mRNA and protein (Fig. 5 B–D).

Hepatic fatty acid β-oxidation has been shown to prevent diet-induced obesity, fatty liver, and hyperlipidemia (26). In our Mstn^−/−/Ldlr^−/− mice model, hepatic fatty acid β-oxidation rates in isolated hepatic mitochondrion were significantly greater (1.7-fold; supplementary Fig. 4B, available in an online appendix) than those from controls. In addition, we have previously shown that Mstn deletion resulted in a 25% increase of liver mitochondrial DNA (unpublished data), implying that these mice have greater mitochondrial oxidation capacity. Plasma level of the ketone body β-hydroxybutyrate, an indirect marker of hepatic fatty acid β-oxidation, was also significantly higher in Mstn^−/−/Ldlr^−/− (supplementary Fig. 4C, available in an online appendix) than in controls.

Decreased hepatic triglycerides in Mstn^−/−/Ldlr^−/− mice was associated with reduced secretion of hepatic triglycerides (23%) and VLDL (Fig. 5,E). Suppression of hepatic fat accumulation in Mstn^−/−/Ldlr^−/− mice was associated with enhanced insulin action in the liver. The mRNA expression levels of insulin receptor substrate (Irs)1 and Irs2, two key players in the regulation of hepatic insulin signaling (27), were significantly higher in Mstn^−/−/Ldlr^−/− mice compared with Mstn^+/+/Ldlr^−/− controls (supplementary Fig. 4D, available in an online appendix). Western blot analyses also showed significantly higher Akt and GSK phosphoryation in isolated liver tissues of Mstn^−/−/Ldlr^−/− mice (Fig. 5 F). As emerging data suggest a causative role of intracellular hepatic triglyceride accumulation in the pathogenesis of hepatic insulin resistance and hypersecretion of VLDL (23,28,–30), these findings provide a mechanistic link between the protective effects of Mstn deletion on obesity, insulin resistance, and atherosclerosis (supplementary Fig. 5, available in an online appendix).

In this report, we provide the evidence that Mstn disruption in Ldlr^−/− mice reduces the development of proatherogenic dysipidemia and atherogenesis progression. The Mstn^−/−/Ldlr^−/− mice had lower plasma triglycerides, FFA, and non-HDL cholesterol than controls. Mstn^−/−/Ldlr^−/− mice also had lower hepatic VLDL secretion rates. Although we cannot exclude direct actions of myostatin on the blood vessel wall, it is likely that the more favorable plasma lipid profile was the key contributor to reduced atherogenesis progression observed in Mstn^−/−/Ldlr^−/− mice.

Mstn deletion prevented the development of diet-induced insulin resistance in Ldlr^−/− mice. Hyperinsulinemic-euglycemic clamp studies revealed that Mstn^−/−/Ldlr^−/− mice had significantly higher peripheral glucose disposal rates. Furthermore, decreased circulating insulin as well as decreased glucose and adipose-derived FFA influx to the liver of the double-knockout mice may contribute to the protection against hepatic lipogenesis and hepatic insulin resistance (23,25,28).

The mechanisms by which Mstn deletion improves insulin sensitivity are not fully understood. Mstn^−/−/Ldlr^−/− mice remained remarkably lean and resisted fat accumulation even when fed HFD for 12 weeks. Myostatin may exert complex effects on adipogenic differentiation (31,32). In addition, it is possible that acquisition of hypermuscularity by itself could lead to leanness (11). Indeed, mice with increased muscle oxidation, such as those overexpressing IGF-1 (33) or Akt-1 (11) in skeletal muscles, have little fat deposition. In our myostatin-deficient model, the marked hypertrophy of glycolytic skeletal muscle was associated with increased glucose uptake, lower blood glucose levels, and consequently lower circulating levels of insulin (11).

Increased hepatic fatty acid β-oxidation (FAO) and mitochondrial biogenesis with Mstn disruption may prevent the development of hepatic steatosis that was observed in Mstn^+/+/Ldlr^−/− controls (25,26). Consistent with the higher FAO rates, Mstn^−/−/Ldlr^−/− mice had higher ketone body production than controls. Improved insulin sensitivity might be expected to suppress FAO and lower ketone body production. However, insulin only effectively suppresses FAO in tissues, primarily muscle and fat, where glucose uptake is tightly regulated by insulin. Because hepatic glucose uptake is not regulated by insulin (34), improved insulin sensitivity in the Mstn^−/−/ Ldlr^−/− mice does not pose limitation on liver FAO as it does in the skeletal muscle. This phenomenon has also been observed by others (11,35). Our data suggest that the increased ketone body production in the double-knockout mice may be a result of physiological adaptation to limited adipose tissue storage for dietary fats from HFD.

Myostatin may also indirectly affect cardiometabolic risk by its effects on myokine secretion by the skeletal muscle. Similar to systemic loss of Mstn, inhibition of myostatin signaling exclusively in the skeletal muscle was sufficient to resist diet-induced obesity and insulin resistance (10). Myokines, such as interleukin-6, can alter metabolic states (36). Other myokines, such as fibroblast growth factor-21 (Fgf-21) and follistatin like-1 (Fstl-1), may also have salutary effects on fat accumulation and insulin sensitivity (37,38).

In summary, we have demonstrated here that Mstn deletion has beneficial cardiometabolic effects in an animal model of atherosclerosis. Mstn^−/−/Ldlr^−/− mice are partially protected against not only fat accumulation and insulin resistance but also the development of proatherogenic dyslipidemia and atherosclerosis. These proof-of-concept studies raise the possibility that myostatin inhibitors may be useful agents for the prevention or treatment of atherosclerosis. Importantly, administration of myostatin inhibitors to humans have demonstrated specificity, safety, and effectiveness in increasing muscle mass and decreasing fat mass (39). To the extent that loss of skeletal muscle mass, adiposity, and atherogenesis occur contemporaneously during aging and in many chronic illnesses, new classes of promyogenic drugs, such as myostatin inhibitors, that increase skeletal muscle mass and retard fat accumulation are particularly attractive against these cardiometabolic disorders.

We acknowledge the support of the U.S. National Institutes of Health grants RO1 DK49296, DK59261, DK64572, DK078512, DK72449, and T32 DK070201–31.

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

We thank S.-J. Lee for the generous gift of Mstn^−/− mice and Q. Yan for providing the training for the catheter placement surgery. We also thank L. Jiang, S. Wong, L. Kaoutzani, and E. Yang for valuable technical assistance.