Myostatin Induces Insulin Resistance via Casitas B-Lineage Lymphoma b (Cblb)-mediated Degradation of Insulin Receptor Substrate 1 (IRS1) Protein in Response to High Calorie Diet Intake*

Background: Excess nutrient intake and elevated levels of Mstn are both associated with the development of insulin resistance. Results: High calorie diet increases Mstn levels. Mstn induces insulin resistance through Cblb. Conclusion: Mstn promotes insulin resistance via Cblb-mediated degradation of IRS1 in response to energy dense diets. Significance: Inhibition of Mstn is a potential therapeutic to combat insulin resistance and T2D. To date a plethora of evidence has clearly demonstrated that continued high calorie intake leads to insulin resistance and type-2 diabetes with or without obesity. However, the necessary signals that initiate insulin resistance during high calorie intake remain largely unknown. Our results here show that in response to a regimen of high fat or high glucose diets, Mstn levels were induced in muscle and liver of mice. High glucose- or fat- mediated induction of Mstn was controlled at the level of transcription, as highly conserved carbohydrate response and sterol-responsive (E-box) elements were present in the Mstn promoter and were revealed to be critical for ChREBP (carbohydrate-responsive element-binding protein) or SREBP1c (sterol regulatory element-binding protein 1c) regulation of Mstn expression. Further molecular analysis suggested that the increased Mstn levels (due to high glucose or fatty acid loading) resulted in increased expression of Cblb in a Smad3-dependent manner. Casitas B-lineage lymphoma b (Cblb) is an ubiquitin E3 ligase that has been shown to specifically degrade insulin receptor substrate 1 (IRS1) protein. Consistent with this, our results revealed that elevated Mstn levels specifically up-regulated Cblb, resulting in enhanced ubiquitin proteasome-mediated degradation of IRS1. In addition, over expression or knock down of Cblb had a major impact on IRS1 and pAkt levels in the presence or absence of insulin. Collectively, these observations strongly suggest that increased glucose levels and high fat diet, both, result in increased circulatory Mstn levels. The increased Mstn in turn is a potent inducer of insulin resistance by degrading IRS1 protein via the E3 ligase, Cblb, in a Smad3-dependent manner.

To date a plethora of evidence has clearly demonstrated that continued high calorie intake leads to insulin resistance and type-2 diabetes with or without obesity. However, the necessary signals that initiate insulin resistance during high calorie intake remain largely unknown. Our results here show that in response to a regimen of high fat or high glucose diets, Mstn levels were induced in muscle and liver of mice. High glucose-or fat-mediated induction of Mstn was controlled at the level of transcription, as highly conserved carbohydrate response and sterol-responsive (E-box) elements were present in the Mstn promoter and were revealed to be critical for ChREBP (carbohydrate-responsive element-binding protein) or SREBP1c (sterol regulatory element-binding protein 1c) regulation of Mstn expression. Further molecular analysis suggested that the increased Mstn levels (due to high glucose or fatty acid loading) resulted in increased expression of Cblb in a Smad3dependent manner. Casitas B-lineage lymphoma b (Cblb) is an ubiquitin E3 ligase that has been shown to specifically degrade insulin receptor substrate 1 (IRS1) protein. Consistent with this, our results revealed that elevated Mstn levels specifically up-regulated Cblb, resulting in enhanced ubiquitin proteasome-mediated degradation of IRS1. In addition, over expression or knock down of Cblb had a major impact on IRS1 and pAkt levels in the presence or absence of insulin. Collectively, these observations strongly suggest that increased glucose levels and high fat diet, both, result in increased circulatory Mstn levels. The increased Mstn in turn is a potent inducer of insulin resistance by degrading IRS1 protein via the E3 ligase, Cblb, in a Smad3-dependent manner.
A number of studies on mice have been performed to understand the effect of dietary fats, sugars, and proteins on the etiology of obesity and type-2 diabetes (T2D) 4 (1)(2)(3). These findings indicate that high nutrient intake increases the risk of obesity and T2D. Insulin resistance in skeletal muscle is a key phenotype associated with obesity and T2D for which the molecular mediators remain unclear. Insulin resistance is seen independent of obesity and as a strong predictor of development of T2D. To date very little information is known about the factors and signaling events that are activated in response to high calorie diets, which in turn promote the development of insulin resistance in skeletal muscle.
Myostatin (Mstn) is a TGF-␤ superfamily member and a lack of Mstn increases lean muscle mass. Recent studies have reported increased Mstn expression in both muscle and adipose tissues derived from obese leptin-deficient (ob/ob) mice (4). In addition, increased secretion and expression of Mstn has been observed in plasma and skeletal muscle of obese women (5). Furthermore, higher levels of Mstn mRNA expression were observed in T2D individuals (6). Consistent with this, studies have reported increased Mstn expression during streptozotocin-induced type-1 diabetes (7). Taken together, these studies clearly demonstrate that during pathological conditions such as obesity, type-1 and type-2 diabetes Mstn expression is high in both circulation and in peripheral tissues including skeletal muscle and adipose tissues. However, to date it is not well identified what signaling events trigger the up-regulation of Mstn. Previous studies have revealed that MyoD and glucocorticoids can up-regulate Mstn at the transcriptional level in skeletal muscle (8,9). Moreover, peroxisome proliferator-activated receptor-␥, CCAAT/enhancer-binding protein-␣, and sterol regulatory element-binding protein 1c (SREBP1c) have been shown to increase Mstn promoter activity in 3T3L1 adipocytes (4). Furthermore nutrients like high protein diet or high glucose treatment have been reported to increase Mstn expression (10,11). However, the mechanism(s) behind up-regulation of Mstn in response to high nutrients and, for that matter, the downstream targets of Mstn that promote the development of insulin resistance remain poorly focused.
Casitas B-lineage lymphoma b (Cblb) is a RING-type E3 ubiquitin ligase that belongs to the Cbl family of proteins, consisting of Cblb, c-Cbl, and Cbl-c. Cbl family proteins share a conserved N-terminal region containing a tyrosine kinase binding domain and a RING-finger domain to facilitate E3 ubiquitin ligase activity. In addition, the C-terminal regions of Cblb and c-Cbl have another domain termed a ubiquitin-associated (UBA) domain (12). Studies have reported that Cblb, Cbl-c, and c-Cbl proteins share commonalties in their mode of action and target selection. The three Cbl family proteins have been shown to down-regulate EGFR-mediated signaling (13)(14)(15). Interestingly, Cbl-and c-Cbl-deficient mice are protected from high fat diet-induced adiposity and insulin resistance with improved energy expenditure and improved insulin sensitivity (16 -18). Moreover, Cblb has been shown to be a highly correlated susceptibility gene for the development of type-1 diabetes both in humans and rodents (19 -21). In addition, it has been reported that Cblb can target insulin receptor substrate 1 (IRS1) and reduce pAkt levels during skeletal muscle atrophy (22). However, to date the role of Cblb in energy metabolism as well as the development of skeletal muscle insulin resistance is not fully understood.
Our results here show that in response to a regimen of high fat or high glucose, Mstn levels were induced in muscle and liver of mice. This induction is mediated at the level of transcription through carbohydrate-responsive element-binding protein (ChREBP) and SREBP1c binding to carbohydrate response (ChoRE) and sterol-responsive (E-box) elements, respectively, on the Mstn gene promoter. Furthermore, we present data to support that increased Mstn levels during high calorie intake promotes the development of insulin resistance via Smad3-mediated up-regulation of Cblb and subsequent degradation of IRS1.

EXPERIMENTAL PROCEDURES
Animals-All wild type (WT) mice (C57BL/6) were purchased from Center for Animal Resources, National University of Singapore (NUS-CARE) Singapore. Mstn Ϫ/Ϫ mice, Cblb Ϫ/Ϫ , and WT mice were maintained at 20°C with a 12-h light-dark cycle. All animal procedures were reviewed and approved by the Institutional Animal Ethics Committee (IACUC), Singapore. 8 -10-week-old mice were used for all animal experiments. WT and Mstn Ϫ/Ϫ mice (n ϭ 8) were fed either a high fat diet (58V8, Test Diet, IN) or chow diet (58Y2, Test Diet, IN) for 12 weeks. After the 12-week feeding regimen, mice were euthanized with CO 2 , and tissues were harvested for further analysis. The supplementation of glucose in WT and Mstn Ϫ/Ϫ mice (n ϭ 8) was performed as previously published (23). To study the effect of excess Mstn on insulin resistance, C57BL/6J mice were randomly grouped into two groups (n ϭ 8) and either injected with 5 g/kg body weight (BW) recombinant Mstn protein or an identical volume of saline subcutaneously 3 times a week for 12 weeks. The recombinant Mstn protein (both human and mouse) was expressed and purified from Escherichia coli as described previously (24). Gastrocnemius muscle and liver tissue were collected from all trial mice for subsequent molecular analysis.
In Vitro Glucose and Palmitate Treatment-To generate the in vitro model of high glucose-induced insulin resistance, C2C12 myoblasts, 96-h-differentiated C2C12 myotubes, or HepG2 cells were treated with 2.5 mM (control), 10 or 25 mM glucose as previously described (33,34). To generate an in vitro model of high fat-induced insulin resistance, C2C12 myoblasts, 96-h differentiated C2C12 myotubes or HepG2 cells were subjected to palmitate loading as described previously (35,36). Cells were treated with either glucose or palmitate for a period of 24 h. Palmitate was purchased from Sigma (catalog #P9767).
Transfection and Luciferase Reporter Assays-Co-transfection of plasmids into myoblasts using LF2000 has been previously described (38). After transfection, myoblasts were subsequently differentiated into myotubes for further analysis. Assessment of Mstn promoter-reporter (1.6 and 0.9 kb) activity was performed as previously described (38).
Quantitative Real Time PCR and Primer Sequences-Tissue and whole cell RNA was extracted using TRIzol reagent as per the manufacturer's protocol (Invitrogen). Synthesis of cDNA, quantitative real time PCR, and subsequent data analysis was performed as previously described (31). The gene-specific primers used in this manuscript are available upon request.
Protein Isolation, Immunoblotting (IB), and Immunoprecipitation (IP)-Protein isolation, quantification, gel electrophoresis, and target protein detection were performed as previously published (31). Co-immunoprecipitation analysis of Cblb and IRS1 interaction was performed as previously described (40). For the detection of ChREBP and SREBP-1c, nuclear and cytoplasmic extracts were prepared using the NE-PER Nuclear and Cytoplasmic Extraction kit from Thermo Fisher Scientific, Rockford, IL (catalog # 78833). The nuclear (50 g) and cytoplasmic (50 g) proteins prepared from cells and tissues were subjected to 10% SDS-PAGE. The details of the antibodies used in this manuscript are available upon request.
Electrophoretic Mobility Shift Assay (EMSA)-Nuclear extracts were prepared as described above. The 3Ј biotin-labeled double-stranded oligonucleotides, which contained the wild type/mutated ChoRE sequences or putative SREBP1c interacting E-box sequences specific for the Mstn promoter, were commercially synthesized by Sigma. The sequences of the Mstn ChoRE probes used are as follows: wt-ChoRE (5Ј-AGA TCC CTG Cca ggt gTC TGC cct ctg GTC AAA ATG A (biotin)-3Ј and mut-ChoRE (5Ј-AGA TCC CTG CTC TGC cct ctg GTC AAA ATG A (biotin)-3Ј). The E-box sequences, which make up the ChoRE in the Mstn promoter, are indicated in lowercase. The mut-ChoRE probe lacks the 5Ј most E-box sequence (CAGGTG). The sequences of Mstn E-box #1, #2, and #3 used as probes are as follows: E-box #1, 5Ј-ATC CTG ACG Aca ctt gTC TCC TCT AAG T-3Ј; E-box #2, 5Ј-TAT GAA GTA GTc aaa tgA ATC AGC TTG C-3Ј; E-box #3, 5Ј-AGA TCC CTG Cca ggt gTC TGC CCT CTG G-3Ј. The Mstn promoter-specific E-box sequences located within each probe are indicated in lowercase. All EMSA were performed using the Lightshift Chemiluminescent EMSA kit (catalog #20148, Thermo Fisher Scientific) as per the manufacturer's protocol. Briefly, myoblasts were plated at a density of 15,000 cells/cm 2 . Cells were 90% confluent at the time of extraction for nuclear extract preparation. Nuclear extracts were prepared, and final protein concentration was measured and adjusted to 1 g/l with extraction buffer for use in the gel shift assay. The ChoRE and E-box containing oligos both unlabeled and labeled with 3Ј biotin were diluted to 1 pmol/l concentration before use. The samples were prepared by adding 10ϫ binding buffer, nuclear extract, and oligos (labeled and unlabeled). Following 30 min of incubation at room temperature, loading dye was added to a final concentration of 1ϫ, and the samples were subjected to PAGE, after which they were transferred to nylon membrane. The detection of specific interactions was performed using the LightShift Chemiluminescent EMSA kit as per the manufacturer's protocol.
Chromatin Immunoprecipitation (ChIP)-ChIP was performed using myoblasts transfected with the 1.6-kb bovine Mstn promoter alone or together with FLAG-ChREBP and HA-Mlx␥ in the absence (2.5 mM) or presence of glucose (25 mM). ChIP was also performed on C2C12 myoblasts transfected with the FLAG-SREBP1c construct and treated with or without 0.25 mM palmitate (PA). ChIP was performed as described previously (38). The sequences of the primers used to detect the Insulin Stimulation-All in vitro insulin stimulation was carried out as follows. Cells were plated at a density 15,000 cell/ cm 2 and grown in DMEM. After 24 h, medium was removed and replaced with 2 ml of serum-free MEM-␣, and cells were incubated for 15 min at 37°C, 5% CO 2 . After this the medium was removed, the fresh serum-free MEM-␣ was added (2 ml), and the cells were further incubated for 16 h at 37°C, 5% CO 2 . After 16 h fasting in serum-free MEM-␣, the cells were incubated in 1 ml of fresh serum-free-MEM-␣ without or with increasing concentrations (0.001, 0.1, and 1 M) of either bovine (Sigma; catalog #10516) or porcine (Sigma; catalog #I6634) insulin for 15 min at 37°C, 5% CO 2 . After stimulation, cells lysates were collected for subsequent IB analysis.
Statistical Analysis-All variations were compared using one-way analysis of variance and two-tailed Student's t-tests. Comparisons between four different groups during high fat diet feeding and in response to high glucose injection were calculated using two-way analysis of variance with multiple comparisons. Results were deemed statistically significant at p Ͻ 0.05. Data are presented as the mean Ϯ S.E. For statistical analysis GraphPad Prism Version 4 software was used.

Mstn Is Up-regulated upon High Glucose Treatment and
High Fat Diet Feeding-Quantitative real time PCR and IB analysis revealed that Mstn expression was significantly up-regulated in response to high glucose and palmitate treatment in C2C12 myotubes and HepG2 human hepatocytes (Fig. 1, A-D) when compared with control-treated cells. The increase in Mstn due to high glucose was both time-and dose-dependent (Fig. 1A). In addition to elevated Mstn, high glucose and palmi-tate treatment led to reduced levels of pAkt, which is consistent with impaired insulin signaling ( Fig. 1, C and D). We next validated these results in vivo. Upon 12 weeks of injections with high glucose, WT mice gained BW ( Fig. 2A), exhibited increased fat pad weights ( Fig. 2B) and increased liver weights (Fig. 2C), and developed insulin resistance, as measured by glucose tolerance testing (GTT) (Fig. 2G) and insulin tolerance testing (ITT) (Fig. 2H). Similarly, WT mice gained significant BW, fat mass, and liver weight upon high fat diet (HFD) feeding ( Fig. 2, D-F) and developed insulin resistance, as measured by GTT (Fig. 2I) and ITT (Fig. 2J). A list of the full biochemical data relating to high glucose and HFD trials is provided in Table 1. ChREBP and SREBP1c Induce Mstn Expression in Response to High Glucose and Fatty Acid Loading, Respectively-Analysis of the Mstn gene promoter revealed the presence of a putative ChoRE (Fig. 3A, upper panel). Subsequent promoter-reporter analysis revealed that the 1.6-kb Mstn upstream element was sufficient for robust activation by high glucose (Fig. 3A, lower panel). Partial deletion of the ChoRE element, however, Shown is IB analysis of Mstn, pAkt, and total Akt in glucose-treated (C) or in PA-treated (D) cells. Tubulin levels were assessed to ensure equal loading. Shown is IB analysis of Mstn, pAkt, and total Akt protein levels in muscle (M) and liver (L) tissue from WT mice injected with either saline or glucose (E) or in muscle (M) and liver (L) tissue from WT mice fed on CD or HFD (F). Tubulin levels were assessed to ensure equal loading (n ϭ 4, for each group). All graphs display the mean Ϯ S.E. *, p Ͻ 0.05; **, p Ͻ 0.01, and ***, p Ͻ 0.001. G, ELISA of Mstn levels in serum from WT mice injected with either saline or glucose (n ϭ 8, for each group). H, ELISA of Mstn levels in serum from WT mice fed either CD or HFD (n ϭ 8 for each group). resulted in a complete loss of Mstn promoter-reporter activity in response to high glucose (Fig. 3A, lower panel). ChREBP is a well characterized transcription factor that in response to glucose binds to ChoREs in target genes to regulate expression (41)(42)(43)(44). Consistent with this, we found increased ChREBP translocation into the nucleus upon high glucose treatment in vitro (Fig. 3B, upper panel) and in muscle and liver tissues of WT mice injected with high glucose (Fig. 3B, lower panel). Furthermore, EMSAs revealed enhanced dose-dependent interaction between nuclear extracts (NE) from high glucose treated myoblasts and the wild type ChoRE (wt-ChoRE) sequence found in the Mstn promoter region (Fig. 3C, upper panel), which was ablated upon incubation with a probe containing a mutated ChoRE sequence (mut-ChoRE) (Fig. 3C, lower panel). This shifted band could be competed out with unlabeled oligo (wt-un ChoRE), confirming the specificity of the protein-DNA complex (Fig. 3C, lower panel). In addition, overexpression of ChREBP and its obligatory partner Mlx␥ increased Mstn expression (Fig. 3D) and Mstn promoter-reporter activity (Fig.  3E), which was further enhanced upon treatment with high glucose. Consistent with this, knockdown of ChREBP expression (Fig. 3, F and G) abolished glucose-mediated induction of Mstn promoter-reporter activity (Fig. 3H). Three independent ChIP experiments further confirmed that ChREBP does indeed bind to the identified ChoRE within the Mstn promoter and that this interaction was enhanced in response to high glucose treatment (Fig. 3I).
Similar to high glucose, treatment with palmitate also increased Mstn promoter-reporter activity (Fig. 4A). SREBP1c has been previously shown to regulate Mstn promoter activity in 3T3L1 cells (4). In agreement with this, we observed strong nuclear translocation of SREBP1c in cells treated with palmitate (Fig. 4B) and in muscle and liver tissues of WT mice fed with HFD (Fig. 4C). Furthermore, overexpression of SREBP1c dramatically enhanced Mstn expression (Fig. 4D) and Mstnpromoter reporter activity (Fig. 4E), which was further enhanced upon treatment with palmitate. Consistent with this, knock down of SREBP1c expression (Fig. 4, F and G) blocked palmitate-mediated induction of Mstn-promoter reporter activity in muscle cells (Fig. 4H). It is well documented that SREBP1c can bind to E-box motifs to regulate target gene activation (45,46). Analysis of the Mstn gene promoter revealed the presence of five E-box motifs consisting of three different consensus E-box sequences (Fig. 4I). With this in mind we performed EMSA using the 3 different E-box sequences contained within the Mstn 1.6-kb promoter region on nuclear extracts from palmitate-treated hMb15 myoblasts. Results revealed that a specific band shift was seen upon incubation with labeled oligo containing Mstn E-box #1 sequence (5Ј-CACTTG-3Ј) (Fig. 4J), whereas no band shift was observed upon incubation FIGURE 2. Physiological changes in WT mice during high glucose and high fat diet regimens. A, BW changes of mice injected with either high glucose or saline for 12 weeks (n ϭ 8 per group). B, epididymal (Epi), retroperitoneal (Retro), and inguinal (Ingu) white adipose tissue (WAT) weight changes, normalized as percentage BW in mice injected with either high glucose or saline (n ϭ 8 per group). C, liver weight changes of mice injected with either high glucose or saline for 12 weeks (n ϭ 8 per group). D, BW changes of mice fed either CD or HFD for 12 weeks (n ϭ 8 per group). E, epididymal, retroperitoneal, and inguinal WAT weight changes, normalized as percentage BW mice fed either CD or HFD (n ϭ 8 per group). F, liver weight changes of mice fed either CD or HFD for 12 weeks (n ϭ 8 per group). G, GTT on mice injected with high glucose or saline (n ϭ 8 per group). H, ITT on mice injected with high glucose or saline (n ϭ 8 per group). I, GTT performed on mice fed either CD or HFD (n ϭ 8 per group). J, ITT performed on mice fed either CD or HFD (n ϭ 8 per group). All graphs display the mean Ϯ S.E. *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001.
Mstn Induces Degradation of IRS1 Protein by Activating the E3 Ligase Cblb-To find out the possible mechanisms through which Mstn can induce insulin resistance, microarray was performed on Mstn-treated myotubes (Fig. 5A). The differential gene expression changes observed in response to Mstn treatment are listed in Table 2 (up-regulated) and Table 3 (downregulated). We identified that Cblb, a ubiquitin E3 ligase, was up-regulated in response to Mstn treatment ( Table 2). Interestingly, Cblb has been shown to degrade IRS1 protein during muscle atrophy and is linked with the development of type-1  diabetes in rats and in humans (19 -22, 47). However, to date the role of Cblb in the development of insulin resistance remains unknown. Hence we considered Cblb as a valid candidate gene through which Mstn may regulate insulin sensitivity. Cblb expression was validated through quantitative real time PCR and IB analysis, and our results revealed that Mstn treatment significantly increased Cblb mRNA levels in C2C12 myotubes and HepG2 cells (Fig. 5B, upper panel). The increased levels of Cblb were associated with reduced IRS1 and pAkt protein levels in the presence of Mstn in C2C12 myotubes and HepG2 cells (Fig. 5B, lower panel). Consistent with this, siRNAmediated knock down of Mstn in C2C12 myotubes and HepG2 cells resulted in a significant reduction in Mstn expression concomitant with low levels of Cblb and increased IRS1 and pAkt levels (Fig. 5C, upper panel). Furthermore, muscle and liver tissues isolated from Mstn Ϫ/Ϫ mice also had low Cblb expression as well as increased levels of IRS1 and pAkt (Fig. 5C, lower   panel), suggesting that Mstn is a potent regulator of Cblb activation. Importantly, co-IP studies revealed enhanced association of IRS1 with Cblb (Fig. 5D) as well as increased ubiquitination (Ub) of IRS1 upon Mstn treatment of myotubes (Fig. 5D). In agreement with this, blockade of the ubiquitin-proteasome pathway through treatment with the proteasome inhibitor MG132 prevented Mstn-mediated loss of IRS1 (Fig. 5E). Taken together these data confirm that loss of IRS1 in response to increased Mstn is ubiquitin proteasome pathway-dependent.
Cblb Expression Is Essential for Mstn-induced Insulin Resistance-We next treated mice with Mstn, and GTT and ITT confirmed that Mstn-treated WT mice developed insulin resistance (Fig. 6, A and B). Mstn treatment resulted in increased BW (Fig. 6C) with significantly increased epididymal fat mass (Fig. 6D) and reduced tibialis anterior and quadriceps muscle weights (Fig. 6E) despite normal comparable food intake between the groups (Fig. 6F). A list of the full biochem- ical data relating to circulatory triglycerides, adiponectin, leptin, total cholesterol, and insulin levels in saline-and Mstntreated mice is provided in Table 4. Treatment with excess Mstn resulted in elevated Cblb levels in skeletal muscle and liver tissues (Fig. 6, G and H), which was associated with reduced IRS1 and pAkt protein levels (Fig. 6, G and H). Primary myoblasts cultures established from Mstn-injected mice also expressed higher levels of Cblb with reduced IRS1 and pAkt (Fig. 6I). Furthermore, insulin-mediated phosphorylation of Akt was impaired in these myoblasts. To further confirm the role of Cblb in Mstn-induced insulin resistance, primary myotubes from Cblb Ϫ/Ϫ mice were challenged with Mstn. The absence of Cblb resulted in a rescue of Mstn-mediated degra-dation of IRS1 and loss of pAkt (Fig. 6J). Moreover, the absence of Cblb prevented the enhanced ubiquitination of IRS1 observed after treatment with Mstn (Fig. 6K). Collectively, these data reveal that increased Mstn induces insulin resistance through a mechanism involving Cblb-mediated loss of IRS1 protein.
Mstn Signal through Smad3 to Regulate Cblb Expression-Mstn has been previously shown to signal through Smad3 to elicit biological function (24). Consistent with this, in silico analysis revealed the presence of several putative Smad3 binding motifs in the Cblb promoter region (Fig. 7A). To test whether Smad3 is important in Mstn regulation of Cblb, we initially assessed the levels of active phosphorylated Smad3 in

List of genes up-regulated in Mstn-treated myotubes compared to control treated myotubes at all time points in the microarray
Gene accession numbers, gene symbols, complete gene names, and -fold change are given. response to high glucose (Fig. 7B) and HFD regimen (Fig. 7C). Consistent with elevated Mstn levels (Fig. 1, A-H) the abundance of phosphorylated Smad3 was increased in response to both high glucose (Fig. 7B) and HFD regimen in vivo (Fig. 7C). Furthermore, treatment with Mstn failed to up-regulate Cblb mRNA expression (Fig. 7, D and F) or protein levels (Fig. 7, E and G) in both Smad3 knock down myotubes (Fig. 7, D and E) and in primary myoblasts isolated from Smad3 Ϫ/Ϫ mice (Fig. 7, F and G) when compared with respective controls. In addition, Mstn-mediated loss of IRS1 and pAkt levels was also rescued in both Smad3 knock down myotubes and in primary myoblasts isolated from Smad3 Ϫ/Ϫ mice when compared with respective controls (Fig. 7, E and G). Taken together these data suggest that Smad3 has an indispensible role in Mstn-mediated activation of Cblb at the transcript level.

GenBank TM accession
Glucose and Fatty Acids Require Mstn-Cblb Signaling to Promote Insulin Resistance-To ascertain whether or not the Mstn-Cblb pathway plays a role in the development of insulin resistance in response to high glucose and HFD regimen, we next assessed the expression of Cblb, IRS1, and pAkt in high glucose-and palmitate-treated C2C12 myotubes and HepG2 cells in the presence or absence of Mstn. In agreement with the elevated Mstn levels detected in response to in vitro treatment with high glucose or palmitate (Fig. 1, A-D), we also noted increased levels of Cblb and reduced IRS1 and pAkt levels in C2C12 myotubes and HepG2 cells after treatment with high glucose (Fig. 8A) and palmitate (Fig. 8C). However, siRNA-mediated knock down of Mstn in C2C12 myotubes and HepG2 cells treated with high glucose (Fig. 8B) or palmitate (Fig. 8D) led to reduced Cblb levels, concomitant with a

List of genes down-regulated in Mstn-treated myotubes compared to control treated myotubes at all time points in the microarray
Gene accession numbers, gene symbols, complete gene names, and -fold change are given. rescue in the levels of IRS1 and pAkt, which is consistent with improved insulin sensitivity. In agreement with the in vitro analysis above, whereas high glucose and HFD treatment resulted in increased Cblb levels in both skeletal muscle and liver isolated from WT mice (Fig. 8, E and F), when compared with respective controls (saline or chow diet (CD)-fed), no increase in Cblb levels was noted in Mstn Ϫ/Ϫ mice fed either high glucose or HFD (Fig. 8, E and F). Moreover, the levels of IRS1 and pAkt in both skeletal muscle and liver were comparable between CD-fed controls and high glucose or HFD-treated Mstn Ϫ/Ϫ mice (Fig. 8, E and F). However, in contrast, a dramatic reduction in the levels of both IRS1 and pAkt was observed in WT mice during high glucose and HFD treatment (Fig. 8, E and  F), which was consistent with both the high levels of Cblb protein (Fig. 8, E and F) and the development of insulin resistance observed in WT mice (Fig. 2, G-J). Furthermore, overexpression of Cblb through lentiviral-mediated transduction in hMb15 myoblasts reduced both basal and insulin-stimulated pAkt levels together with IRS1 protein (Fig. 8G). On the other  hand, enhanced basal and insulin stimulated pAkt levels, and IRS1 abundance was observed in myoblasts derived from Cblb Ϫ/Ϫ mice when compared with myoblasts isolated from control Cblb ϩ/ϩ mice (Fig. 8H). Taken together these data con-firm that the Mstn-Cblb pathway appears to play a critical role during the induction of insulin resistance in response to both high glucose and HFD regimen and that Cblb plays a major role in inhibiting insulin signaling.

DISCUSSION
Energy-dense diets that are high in fat, protein, and sugar are associated with a risk of obesity and T2D (48). Insulin resistance is key predictor of T2D and is associated with both non-obese and obese pathological conditions (49). However, the underlying molecular mechanism(s) that initiates the development of insulin resistance during continued high calorie intake is poorly characterized. Previous work by Hittel et al. (50) has revealed that injection of exogenous Mstn protein into mice leads to the development of insulin resistance. However, the role of Mstn in initiating insulin resistance in response to high calorie intake has not been studied. Here we show a conclusive mechanism through which energy-rich diets in mice induce high levels of Mstn, which subsequently results in the targeted degradation of the critical insulin-signaling molecule IRS1 by up-regulation of the ubiquitin E3 ligase Cblb.
The myokine, Mstn, belongs to the TGF-␤ super family and primarily functions to control muscle growth and development (51). However, recent reports have shown that Mstn also plays a role in regulating muscle metabolism; in fact, either inhibition of Mstn or lack of Mstn reduces fat accumulation and enhances insulin sensitivity (24,52). Our laboratory characterized increased AMP-activated protein kinase and peroxisome proliferator-activated receptor signaling in Mstn Ϫ/Ϫ muscle as mechanisms behind the increased fat oxidation observed in Mstn Ϫ/Ϫ mice (24). Moreover, recent evidence also reveals that elevated levels of Mstn in muscle in both human and mouse models are associated in obesity, type-1 and type-2 diabetes (5,7,53). Therefore, we considered Mstn to be an excellent candidate that could potentially respond to nutrient signals to further regulate muscle metabolism. Indeed we observed that in vitro treatment of C2C12 and HepG2 cells with high glucose or fatty acid (palmitate) as well as subjecting mice to high glucose or HFD feeding in vivo significantly increased Mstn levels, suggesting that Mstn expression is indeed under the control of major nutrient factors like glucose and fatty acids. In agreement with this, high glucose treatment of C2C12 has been previously shown to increase Mstn, block myogenesis, and promote myotube atrophy (10). Specifically, our results revealed that glucose and fatty acid/HFD treatment induce Mstn via independent transcription factors; although SREBP1c was shown to be sufficient to induce Mstn in response to fatty acids by binding to an E-box motif, ChREBP was found to be responsible for glucosemediated induction of Mstn through interaction with a ChoRE. Glucose regulation of gene transcription through ChREBP is not a new concept; in fact studies have reported that in response to high glucose treatment, ChREBP binds and activates genes involved in lipogenesis (54 -56). In addition, fatty acid regulation of SREBP1c has been previously reported whereby treatment with palmitate was shown to increase SREBP1c mRNA and protein levels (57). Moreover, SREBP1c-mediated up-regulation of Mstn promoter has been reported in adipocytes previously (4). Thus, these data reveal that ChREBP and SREBP1c are critical for Mstn gene regulation in response to energy-rich diets.
It is noteworthy to mention that in addition to elevated Mstn levels we consistently observed concomitant reduction in the phosphorylation of Akt in both in vitro and in vivo models, suggesting that increased Mstn inhibited insulin signaling. To delineate the molecular mechanism through which Mstn promoted insulin resistance, microarray was employed, and notably we discovered that Mstn potently induces the expression of Cblb, an ubiquitin E3 ligase previously shown to target and degrade IRS1 protein in skeletal muscle (22). It is noteworthy to mention that the expression of other IRS1-specific E3 ligases, including Cul-7, SOCS1, and SOCS3, was not significantly altered in the microarray analysis (data not shown). Cblb has been previously reported to be associated with type-1 diabetes both in rodents and humans (19 -21). In addition, it has also FIGURE 8. High glucose injection or HFD increases Cblb protein levels in muscle and liver tissues. IB analysis of Cblb, IRS1, pAkt, and total Akt in cells treated with (ϩ) or without (Ϫ) glucose (Glu) (A) and in cells transiently transfected with either negative siRNA (Ϫ) or Mstn siRNA (ϩ) and treated with (ϩ) glucose (Glu) (B). IB analysis of Cblb, IRS1, pAkt, and total Akt cells treated with (ϩ) or without (Ϫ) PA (C) and in cells transfected with either negative siRNA (Ϫ) or Mstn siRNA (ϩ) and treated with (ϩ) PA (D) is shown. IB analysis of Cblb, IRS1, pAkt and total Akt protein levels in muscle and liver tissues collected from WT and Mstn Ϫ/Ϫ mice injected with either saline (-) or glucose (ϩ) (E) and in muscle and liver tissues collected from WT and Mstn Ϫ/Ϫ mice fed either CD (Ϫ) or HFD (ϩ) (F) is shown (n ϭ 2 for each group). G, IB analysis of Cblb, IRS1, pAkt, and total Akt in hMb15 myoblasts infected with either control (pLOC) or with Cblb overexpressing lentivirus (pLOC-Cblb) in the presence of increasing concentrations of insulin (0, 0.01, 0.1, and 1 M). IB analysis of IRS1, pAkt, and total Akt in myoblasts derived from Cblb ϩ/ϩ and Cblb Ϫ/Ϫ mice stimulated with increasing concentrations of insulin (0, 0.01, 0.1 and 1 M) (H) is shown. Tubulin levels were assessed in all IBs to ensure equal loading. All IB images are representative of at least two independent experiments. been shown that during skeletal muscle atrophy Cblb targets IRS1 protein for degradation via the ubiquitin proteasome pathway (22). However to date, the importance of Cblb during the development of insulin resistance and for that matter factors that regulate Cblb expression during obesity are poorly understood. Subsequent quantitative PCR and IB analysis independently confirmed that Mstn indeed induces Cblb expression not only in vitro but also in mice in vivo. Consistent with the E3 ligase function of Cblb, induction of Cblb by Mstn led to increased association of Cblb with IRS1 and subsequent ubiquitination of IRS1 protein. The Mstn-Cblb-IRS1 pathway was further confirmed when Mstn failed to reduce IRS1 protein levels either in the absence of Cblb or in the presence of the proteasome inhibitor MG132. IRS1 is a key molecule in the insulin-signaling pathway (58 -60) and is highly expressed in skeletal muscle and white adipose tissue (61). Consistent with this, deletion of IRS1 in mice induces severe insulin resistance (62)(63)(64). Moreover, reduced IRS1 mRNA and/or protein levels are detected in subjects with T2D (65,66). Therefore, Mstn/ Cblb-mediated loss of IRS1 would most certainly contribute to the insulin resistance phenotype observed in response to increased Mstn levels.
Importantly, studies have also reported that serine phosphorylation of IRS1, in contrast to the tyrosine phosphorylation of IRS1 normally observed after insulin treatment, is associated with T2D (67,68). In addition, serine phosphorylation of IRS1 has been shown to promote both enhanced degradation of IRS1 through the ubiquitin proteasome pathway (69) and development of insulin resistance. However, it is noteworthy to mention that in the current study we did not assess the serine phosphorylation status of IRS1; as such, development of insulin resistance, due to increased serine phosphorylation of IRS1 in response to high calorie diet and Mstn treatment cannot be ruled out. Nevertheless, through several independent investigations we have clearly shown that initiation of the Mstn-Cblb pathway leads to degradation of IRS1.
We hypothesized that Mstn-Cblb degradation of IRS1 would further result in hypophosphorylation of Akt and development of insulin resistance. In agreement with this, overexpression of Cblb resulted in loss of IRS1 and reduced pAkt levels, whereas the absence of Cblb blocked IRS1 degradation and increased the levels of pAkt under basal conditions and in response to insulin. These data strongly suggest increased insulin sensitivity in the absence of Cblb. In addition, we observed a greater increase in Mstn and Cblb protein levels in both skeletal muscle and liver tissues of mice treated with either high glucose or HFD, with an associated decrease in IRS1 protein, reduced phosphorylation of Akt, and impaired insulin sensitivity in WT mice but not in Mstn Ϫ/Ϫ mice. These data demonstrate that in the absence of Mstn, high glucose and HFD feeding failed to activate Cblb in both liver and muscle tissues. These observations also support a role for Mstn in promoting insulin resistance in liver in response to high glucose and HFD feeding, which is in fact quite consistent with a previously published report demonstrating that lack of Mstn protects the liver from diet-induced insulin resistance (39). Although Wilkes et al. (39) speculated that the improved insulin sensitivity observed in the liver may be due to reduced TNF-␣ levels, our data presented here suggests that loss of Mstn may also lead to reduced activation of Cblb in liver and improved insulin sensitivity.
In summary, we for the first time report that energy-rich diets, specifically high glucose and high fat, signal through ChREBP and SREBP1c, respectively, to induce high levels of Mstn, subsequently resulting in the targeted degradation of the critical insulin-signaling molecule IRS1 via Smad3-dependent up-regulation of the ubiquitin E3 ligase Cblb (Fig. 9).