Skeletal Muscle–Specific Deletion of MKP-1 Reveals a p38 MAPK/JNK/Akt Signaling Node That Regulates Obesity-Induced Insulin Resistance

Skeletal muscle is considered to be a major tissue for the disposal of glucose and free fatty acids and as such plays a critical role in the regulation of whole-body glucose metabolism and energy homeostasis (1–4). Skeletal muscle consists of myofibers that are classified as slow-twitch/oxidative myofibers that express type I myosin heavy chain (MHC) and fast-twitch/glycolytic myofibers that express type IIa, IIx, and IIb MHCs (5). A reduction in the proportion of type I myofibers, which are mitochondria rich, is observed in obesity, and the proportion of type I myofibers positively correlates with overall metabolic health (2,5–9).

Stress-responsive mitogen-activated protein kinases (MAPKs), p38 MAPK, and the c-Jun NH₂-terminal kinase (JNK) are important regulators of skeletal muscle metabolic function (10–12). The stress-responsive MAPKs control processes such as insulin signaling, glucose homeostasis, fatty acid metabolism, and energy expenditure (11,13–16). Although it is realized that metabolic stressors such as inflammation and nutrient excess activate both p38 MAPK and JNK (12,17), the results of these studies reflect the individual actions of these MAPKs on metabolism. It remains unknown as to what the integrated actions are of skeletal muscle stress-responsive MAPKs, such as p38 MAPK and JNK, on the progression of obesity and insulin resistance.

Inactivation of the stress-responsive MAPKs is achieved by the specific dephosphorylation of the MAPKs by dual-specificity MAPK phosphatases (MKPs) (18). The MKPs have been implicated in the regulation of metabolic homeostasis (19). MKP-1, which dephosphorylates JNK and p38 MAPK, is an important regulator of MAPK-dependent regulation of lipid homeostasis, energy metabolism, and mitochondrial biogenesis (20,21). In skeletal muscle of obese mice, MKP-1 is upregulated by free fatty acids concomitant with downregulation of p38 MAPK (22). Whole-body MKP-1–deficient mice are resistant to diet-induced obesity, exhibit increased peroxisome proliferator–activated receptor γ coactivator 1α (PGC-1α) activity in skeletal muscle, and are protected from the loss of oxidative myofibers (22). Although these observations suggest a role for MKP-1 in skeletal muscle metabolism, the limitations of whole-body MKP-1 deletion (20) preclude definitive assessment of the contribution of MKP-1 in skeletal muscle toward systemic energy homeostasis and glycemic control. Therefore, the goal of this study was to address the contribution of skeletal muscle MKP-1 toward glucose disposal and energy metabolism.

In this study, we found that in skeletal muscle, MKP-1 is overexpressed in obese humans. Our studies imply that when overexpressed, skeletal muscle MKP-1 contributes to obesity by antagonizing p38 MAPK and JNK, resulting in dysfunctional mitochondrial biogenesis and skewing of type I myofiber composition. These results demonstrate that it is the opposing action of skeletal muscle MKP-1 on the p38 MAPK/JNK module, which reflects a nodal point of interference that leads to metabolic imbalance and susceptibility to obesity.

All reagents were procured from standard chemical vendors. Phospho-p38 MAPK (#9215S), phospho-JNK1/2 (#4668S), phospho–extracellular signal–regulated kinase 1/2 (ERK1/2; #9101S), phospho-Akt (#9271S), phospho-p70 S6 kinase (#9234S), Akt1 (#2938S), p70 S6 kinase (2708S), PTEN (#9559S), and α-tubulin (#3873S) were obtained from Cell Signaling Technology (Danvers, MA). ERK1/2 (#sc-94), p38 MAPK (#sc-535), JNK (#sc-571), and MKP-1 (#sc-1199) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Skeletal muscle tissue from chow- and high-fat diet (HFD)–fed male Mkp-1^fl/fl and mice lacking skeletal muscle expression of MKP-1 (MKP-MKO) was isolated as described (23).

Generation of MKP-1 “floxed” mice has been described (23). MKP-1 flox/flox (Mkp-1^fl/fl) mice were bred with mice expressing Cre recombinase under the control of human α-skeletal actin (HSA-Cre) (24) to obtain HSA-Cre-MKP-1flox/+ that were backcrossed to Mkp-1^fl/fl mice to generate skeletal muscle–specific deletion of MKP-1, HSA-Cre-MKP-1^fl/fl; referred to as MKP-1 skeletal muscle knockout mice (MKP1-MKO). Muscle biopsy samples from healthy, sedentary lean and obese individuals between the ages of 18 and 35 years were used. Lean individuals had a BMI <25 kg/m², and obese individuals had a BMI ≥30 kg/m².

Glucose tolerance tests (GTTs) were carried out in 8–12-week-old male Mkp-1^fl/fl and MKP-MKO littermates. Mice were fasted overnight for 16 h followed by an i.p. injection of glucose (2 g/kg). Blood glucose was measured at 0, 15, 30, 60, 90, and 120 min. For insulin tolerance tests (ITTs), mice were fasted for 4 h and injected (i.p.) with 0.5 mU/g human insulin (Humulin R; Eli Lilly and Company, Indianapolis, IN). Hyperinsulinemic-euglycemic clamp studies were carried out on male chow-fed Mkp-1^fl/fl and MKP1-MKO mice at 12 weeks old at the Yale Metabolic Phenotyping Center as described (23). Conscious male Mkp-1^fl/fl and MKP1-MKO mice were used to determine total body fat and lean mass using ¹H-magnetic resonance spectroscopy (Bruker mini-spec Analyzer; Echo Medical Systems, Houston, TX). Mice were adapted for 2 weeks, kept in metabolic cages (TSE Systems, Bad Homburg, Germany) for 1 week, and food and water intake, energy expenditure, respiratory exchange ratio (RER), and physical activity were measured (23).

C2C12 myoblasts were cultured as described (22) and transfected with constitutively active mutants of MKK6 (EE), MKK4 (EE), and MEK1 (23) in the absence or presence of 5 μmol/L MAPK inhibitors of SB203580 (p38 MAPK), SP600125 (JNK), and U0126 (MEK). After 48 h, myoblasts were lysed using RIPA buffer (23).

For mitochondrial oxygen consumption, samples of muscle from chow- and HFD-fed male Mkp-1^fl/fl and MKP1-MKO mice were analyzed in a respiration chamber (23). Total DNA was isolated from soleus muscle using QIAamp DNA mini kit (Qiagen, Hilden, Germany). Mitochondrial DNA (mtDNA) was quantified by quantitative RT-PCR using primers amplifying the D-loop region on mtDNA.

Real-time quantitative PCR was carried out in triplicate with the Applied Biosystems 7500 Fast RT-PCR system (Applied Biosystems, Foster City, CA) and TaqMan and SYBR Green gene expression master mix for the following genes: MKP-1, sterol regulatory element-binding protein 1c (SREBP1C), PPARG, PPARGC1B, PRC, c-Myc, p53, Spot14, and ME. The following TaqMan primers (Applied Biosystems) were used: PGC-1α, Mn01208835_m1; NDUFS1, Mn00523631_m1; NRF1, Mn00447996_m1; GABPA, Mn00484598_m1; NDUFB5, Mn00452592_m1; MFN2, Mn01255785; and TFAM, Mn00447485_m1. TaqMan primers and gene expression master mix from Applied Biosystems were used for miR-21 and MKP-1 quantitation.

Animals in the exercise-trained groups were habituated to running on a treadmill (Columbus Instruments, Columbus, OH) by increasing the duration and speed of running for 5 consecutive days. After habituation, animals were rested for 2 days and then performed 5 consecutive days of treadmill running for 60 min/day at 15 m/min, 0% grade. After 5 days of training, we performed a progressive exercise stress test. Maximum exercise capacity was determined by graded increase in treadmill speed (2 m/min to 6 m/min every 5 min) to exhaustion and then VO_2max and running distance were measured.

Serum and hepatic triglycerides were measured at the Yale Mouse Metabolic Phenotyping Center. Male chow-fed Mkp-1^fl/fl and MKP1-MKO mice between 8 and 12 weeks old were used for the measurement of plasma glucose concentrations by a glucose oxidase method using a Beckman Glucose Analyzer II (Beckman Coulter, Sharon Hill, PA). The measurement of plasma insulin was performed by radioimmunoassay (Linco Research, St. Louis, MO). Oil Red O staining was performed as described (23).

All data represent the mean ± SEM. Differences between groups were calculated using a Student t test or ANOVA with Bonferroni posttest for multiple comparisons using Prism 6 statistical software (GraphPad Software, La Jolla, CA).

Aberrant skeletal muscle expression of MKP-1 could play a role in the progression of obesity and insulin resistance in humans. We determined the expression levels of MKP-1 in skeletal muscle of lean and obese human subjects. In skeletal muscle of obese subjects, we found that the expression levels of MKP-1 protein were significantly elevated (∼40%; P < 0.05; N = 7) as compared with lean subjects (Fig. 1A). Concomitantly, p38 MAPK was significantly downregulated in obese subjects (Fig. 1A). These results suggest that MKP-1 overexpression in skeletal muscle of humans may contribute to the development of obesity and/or insulin resistance.

To examine the role of MKP-1 in skeletal muscle, we bred MKP-1 “floxed” mice (23) with the HSA-Cre mice (24) to produce offspring in which MKP-1 is deleted specifically in skeletal muscle (MKP1-MKO) (Fig. 1B). MKP1-MKO mice displayed increased skeletal muscle p38 MAPK and JNK, but not ERK1/2, phosphorylation (Fig. 1C). Therefore, deletion of skeletal muscle MKP-1 results in the upregulation of both p38 MAPK and JNK phosphorylation, indicating functional loss of MKP-1.

Male MKP1-MKO mice fed a chow diet exhibited comparable weight gain and total body lean mass, but significantly reduced fat mass, as compared with Mkp-1^fl/fl littermate controls (Fig. 1D–F). Skeletal muscle sections and weights showed no differences between chow-fed MKP1-MKO and Mkp-1^fl/fl littermate controls (Fig. 1G and H). No differences either in weight or histological presentation were observed in the liver or adipose tissue of chow-fed MKP1-MKO mice (Supplementary Fig. 1). MKP1-MKO mice fed an HFD had a significantly (P < 0.01) reduced rate of weight gain as compared with their Mkp-1^fl/fl littermates (Fig. 2A). By 16 weeks of high-fat feeding, MKP1-MKO mice weighed ∼20% less (P < 0.01) as compared with their Mkp-1^fl/fl littermate controls (Fig. 2A). This was due to reduced adiposity rather than lean mass (Fig. 2B and C). Skeletal muscles from HFD-fed MKP1-MKO mice were unremarkable (Fig. 2D). Resistance to diet-induced obesity in MKP1-MKO mice was associated with reduced serum triglyceride accumulation (Fig. 2E) that occurred independently of changes in skeletal muscle lipogenic gene expression of PPARG and SREBP1C (Fig. 2F and G). These results demonstrate that loss of MKP-1 in skeletal muscle protects from diet-induced obesity.

Chow-fed male MKP1-MKO mice (8–12 weeks) exhibit significantly enhanced energy expenditure levels, oxygen consumption, and carbon dioxide output as compared with Mkp-1^fl/fl littermate controls (Fig. 3A–C). No differences in RER, food intake, or activity between the two genotypes were observed (Fig. 3D–F). These results demonstrate that loss of MKP-1 in skeletal muscle is protective from diet-induced obesity because of increased energy expenditure.

Upon HFD feeding for 18 weeks, the livers of male MKP1-MKO mice were resistant to the acquisition of hepatic steatosis (Fig. 4A), as evidenced by a marked decrease in liver weight and hepatic triglyceride accumulation (Fig. 4B and C). Consistent with this, MKP1-MKO livers exhibited significantly decreased expression levels of PPARG and SREBP1C mRNA (Fig. 4D and E). MKP1-MKO and Mkp-1^fl/fl controls displayed comparable levels of p38 MAPK, JNK (Fig. 4F), and Akt activities (Fig. 4G). Together, these results show that skeletal muscle MKP1-MKO mice, when fed an HFD, exhibit impaired hepatic de novo lipogenesis independently of p38 MAPK, JNK, Erk, and Akt activities.

Under conditions of chow and HFD feeding, MKP1-MKO mice and Mkp-1^fl/fl controls were comparable in their basal glucose levels (Fig. 5A). However, MKP1-MKO mice exhibited significantly reduced plasma insulin levels as compared with Mkp-1^fl/fl controls under both fasted and fed conditions (Fig. 5B). These results suggested that MKP1-MKO mice were more insulin sensitive than Mkp-1^fl/fl controls and required less insulin to maintain euglycemia. On a chow diet, GTTs and ITTs in MKP1-MKO mice were comparable (Fig. 5C and D), and this was confirmed by hyperinsulinemic-euglycemic clamp analyses (Supplementary Fig. 2). GTTs showed that MKP1-MKO mice managed an acute glucose load to levels that were comparable to Mkp-1^fl/fl controls under HFD feeding conditions (Fig. 5E). However, during the GTT, MKP1-MKO mice produced lower levels of circulating insulin as compared with Mkp-1^fl/fl controls (Fig. 5F). Furthermore, ITTs demonstrated that HFD-fed MKP1-MKO mice were insulin sensitive (Fig. 5G). Collectively, these data demonstrate that MKP1-MKO mice are protected from the development of insulin resistance when fed an HFD.

To investigate the molecular basis for the enhanced insulin sensitivity in HFD-fed MKP1-MKO mice (Fig. 5), we measured the phosphorylation status of p38 MAPK, JNK, and ERK in skeletal muscles of HFD-fed MKP1-MKO mice. MKP1-MKO mice displayed increased levels of skeletal muscle p38 MAPK and JNK, but not ERK1/2, phosphorylation as compared with Mkp-1^fl/fl controls (Fig. 6A). Remarkably, in skeletal muscles of MKP1-MKO mice, Akt Ser⁴⁷³ phosphorylation was significantly increased as compared with Mkp-1^fl/fl controls (Fig. 6B). The phosphorylation of S6 kinase (S6K) Thr³⁸⁹ was also increased compared with that of Mkp-1^fl/fl littermates (Fig. 6B). However, no significant differences in either Akt or S6K phosphorylation were observed between chow-fed MKP1-MKO and Mkp-1^fl/fl littermate controls (Supplementary Fig. 3). These data demonstrate that MKP-1 in skeletal muscle plays a role in negatively regulating both p38 MAPK and JNK but additionally the Akt/S6K pathway in HFD-fed mice.

The observation that skeletal muscle MKP-1 negatively regulates Akt prompted us to examine the mechanism of MKP-1/Akt signaling cross talk. We first tested the idea that Akt signaling could be indirectly affected by alterations in the expression levels of PTEN, which negatively regulates Akt. When we assessed the expression levels of PTEN in skeletal muscles of HFD-fed MKP1-MKO mice, we found significantly reduced levels as compared with Mkp-1^fl/fl controls (Fig. 7A). Consistent with the observation that the expression levels of Akt Ser⁴⁷³ were unaltered in the liver of MKP1-MKO mice, PTEN expression was unchanged between MKP1-MKO and Mkp-1^fl/fl controls (Supplementary Fig. 4). These data suggest that in skeletal muscle, MKP-1 negatively regulates Akt activity by controlling PTEN expression.

It has been reported that one of the miR-21 targets is PTEN mRNA, which causes a decrease in the expression of PTEN (25,26). The expression levels of miR-21 were significantly increased in skeletal muscle of MKP1-MKO mice as compared with Mkp-1^fl/fl littermates but not in the liver (Fig. 7B), where Akt phosphorylation is unaffected (Fig. 4G). The synthesis of mature microRNA (miRNA) starts with the formation of primary (pri-)miRNAs in the nucleus (27). Pri-miRNAs are further cleaved into precursor miRNAs in the cytoplasm into mature miRNAs (27). Therefore, we measured pri-miR-21 in skeletal muscles of MKP1-MKO mice. In addition to the enhanced levels of mature skeletal muscle miR-21 (Fig. 7B), MKP1-MKO mice exhibited increased expression of pri-miR-21 as compared with Mkp-1^fl/fl littermates (Fig. 7B). However, pri-miR-21 levels were unaltered in its level of expression in the liver between genotypes (Fig. 7B). To substantiate the results of the pri-miR-21/miR-21 expression in the liver, we measured hepatic PTEN expression in MKP1-MKO mice. As expected, the expression of hepatic PTEN is equivalent between MKP1-MKO mice and Mkp-1^fl/fl controls (Supplementary Fig. 4). These results support the notion that MKP-1 negatively regulates the miR-21/PTEN/Akt pathway in skeletal muscle.

To determine the MAPK dependency for the increase in miR-21 expression, C2C12 myoblasts were transfected with activating mutants of MKK6, MKK4, and MEK1 in order to constitutively activate p38 MAPK, JNK, and ERK, respectively (Fig. 7C and Supplementary Fig. 5A). Both activating mutants of MKK6 and MKK4 enhanced miR-21 expression (Fig. 7C). In contrast, we did not observe any significance change in miR-21 expression with an MEK1-activating mutant (Fig. 7C). Next, C2C12 myoblasts were transiently transfected with activating mutants of MKK6, MKK4, and MEK1 in the absence or presence of MAPK inhibitors of SB203580 (p38 MAPK), SP600125 (JNK), and U0126 (MEK) and miR-21 expression assessed. Both activating mutants of MKK6 and MKK4 enhanced miR-21 expression, which was inhibited in the presence of SB203580 and SP600125, respectively (Supplementary Fig. 5B). Whereas pharmacological inhibition of MEK1 in the presence of an activated MEK1 had no effect on miR-21 expression (Supplementary Fig. 5B). These data establish that MKP-1 negatively regulates Akt signaling by opposing p38 MAPK and/or JNK-mediated expression of miR-21 in skeletal muscle.

Clinical studies from human skeletal muscles have demonstrated that obesity and insulin resistance are linked to a shift toward glycolytic myofibers concomitant with a reduction in oxidative myofiber fiber content (2,5–9). To evaluate the composition of myofiber types in MKP1-MKO mice, we measured the expression levels of MHC isoforms. In soleus muscle under both chow and HFD feeding conditions, the percentage of MHCI expression was significantly increased, whereas MHCIIB expression was reduced in MKP1-MKO mice as compared with Mkp-1^fl/fl controls (Fig. 8A). The expression of MHCIIA and MHCIIX was comparable between the two genotypes (Fig. 8A). Similarly, tibialis anterior muscle revealed that MHCI expression was significantly increased concomitant with decreased levels of MHCIIB in MKP1-MKO mice compared with Mkp-1^fl/fl controls fed chow and HFD (Fig. 8B). No differences were observed in the tibialis anterior of MHCIIA and MHCIIX between the genotypes (Fig. 8B). Collectively, these data demonstrated increased oxidative capacity in skeletal muscle of MKP1-MKO mice.

Next, we asked whether skeletal muscle mitochondrial genes were affected in MKP1-MKO mice. Under both chow and high-fat feeding, skeletal muscle PGC-1α, a master regulator of mitochondrial biogenesis (28), was enhanced in expression in MKP1-MKO mice (Fig. 8C and D). A significant upregulation in the expression of PGC-1β, TFAM, GABPα and NDUFS1 genes was detected in skeletal muscle of either chow- or HFD-fed MKP1-MKO mice (Fig. 8C and D). The expression of thyroid hormone target genes (29) SPOT14 and ME1 (malic enzyme) were also significantly increased in expression from skeletal muscle of chow-fed MKP1-MKO mice (Fig. 8C). The expression of MYC, which positively regulates mitochondrial biogenesis (30,31), was significantly increased in MKP1-MKO mice compared with Mkp-1^fl/fl chow-fed mice (Fig. 8C). These results demonstrate that skeletal muscle MKP-1 plays a key role in negatively regulating the expression of genes involved in mitochondrial biogenesis.

To establish whether the enhanced expression of mitochondrial genes in skeletal muscles of MKP1-MKO mice improves oxidative mitochondrial capacity, we measured mitochondrial respiration in permeabilized myofibers. The respiratory control ratio (state 3, ADP-stimulated maximal respiration; state 4, basal mitochondria respiration) was found to be increased in MKP1-MKO mice in both chow- and HFD-fed mice as compared with HFD- and chow-fed Mkp-1^fl/fl controls (Fig. 8E). These results suggest that skeletal muscle mitochondria from MKP1-MKO mice show improved ability to couple oxygen consumption to ATP synthesis (Fig. 8E). Although the expression of UCP3 in chow-fed MKP1-MKO and Mkp-1^fl/fl control mice was comparable, a trend toward increased UCP3 expression was observed in HFD-fed MKP1-MKO mice as compared with Mkp-1^fl/fl controls (Supplementary Fig. 6). Furthermore, mtDNA content is increased in skeletal muscle of both chow- and HFD-fed MKP1-MKO mice (Fig. 8F).

Finally, we examined endurance capacity in MKP1-MKO mice. The maximal amount of oxygen used at exhaustion (VO_2max) was significantly higher in MKP1-MKO mice (Fig. 8G). We also observed an increased running distance in the MKP1-MKO chow-fed mice as compared with Mkp-1^fl/fl controls (Fig. 8H), and this correlated with reduced total body fat mass (Fig. 8I). Thus, MKP1-MKO mice have increased endurance capacity consistent with the notion that skeletal muscle MKP-1 acts to negatively regulate mitochondrial oxidative capacity and myofiber composition.

In this study, we show that MKP-1 is increased in its levels of expression in skeletal muscle of obese humans. MKP-1 is induced in many cases, if not all, by the same stressors that activate p38 MAPK and JNK to function in a negative-feedback pathway (18). Consistent with this, mice fed an HFD exhibit increased skeletal muscle MKP-1 expression and concomitant downregulation of p38 MAPK (22). Interestingly, p38 MAPK has also been reported to be downregulated in the livers of HFD-fed mice, which is in line with the observation that hepatic MKP-1 is upregulated (23,32). Additionally, in obese human subjects, MKP-1 has been shown to be overexpressed in adipose tissue and macrophages with concomitant downregulation of p38 MAPK and PGC-1α (33). It is well established that the expression levels of adiponectin are negatively correlated with obesity (34). In skeletal muscle, adiponectin negatively regulates the expression of MKP-1 (35). Consistent with this, HFD-fed mice that have low levels of adiponectin exhibit reduced p38 MAPK-mediated activation of PGC-1α and reduced skeletal muscle mitochondrial biogenesis (35). Collectively, these observations are consistent with our findings and support a mechanism by which in obesity increased free fatty acids (22) and hypoadiponectin states promote reduced skeletal muscle oxidative metabolism by upregulating MKP-1.

By examining the composition of skeletal muscle myofibers, we found that there was an increase in the proportion of slow oxidative myofibers and a reduction in the proportion of fast glycolytic myofibers in skeletal muscle derived from MKP1-MKO mice. Thus, our results imply that changes in myofiber type composition could contribute to the enhanced oxidative and reduced glycolytic capacity of skeletal muscles of MKP1-MKO mice. Mitochondrial oxidation, ATP synthesis, and myofiber type composition contribute to the control of muscle endurance (36,37). MKP1-MKO mice were observed to exhibit significantly increased levels of endurance, which further supports the interpretation that the oxidative characteristics of these skeletal muscles is improved. Together, these observations provide an explanation for the increased levels of mitochondrial function and increased whole-body energy expenditure in MKP1-MKO mice and thus resistance to diet-induced obesity. It has been suggested that slow oxidative myofibers, which confer enhanced mitochondrial capacity, are more insulin sensitive than fast glycolytic myofibers (8,38). Moreover, studies have shown that decreased PGC-1α and NRF-1 levels correlate with insulin resistance and type 2 diabetes (39). Several studies have linked decreased mitochondrial biogenesis to the pathogenesis of insulin resistance and type 2 diabetes (9,40). Consistent with this, MKP1-MKO mice are insulin sensitive when fed an HFD, which is consistent with that the increased percentage of slow oxidative myofibers and increased Akt in MKP1-MKO mice. In contrast, under chow-fed conditions, in which p38 MAPK and JNK activity were also increased in skeletal muscles of MKP1-MKO mice, we were unable to detect changes in the expression levels of the PTEN/Akt/miR-21 axis and subsequently differences in insulin-stimulated glucose homeostasis. These results indicate that loss of MKP-1 only under HFD conditions is sufficient to induce changes in the expression of the PTEN/Akt/miR-21 pathway. This could be due to the magnitude of p38 MAPK/JNK hyperactivation in the absence of MKP-1 in skeletal muscles of mice fed a chow diet as compared with those fed an HFD. Indeed, we observed that the increased levels of p38 MAPK/JNK activity are greater in MKP1-MKO mice fed an HFD as compared with those mice fed a chow diet. Loss of MKP-1 in skeletal muscle either in chow- or HFD-fed MKP1-MKO mice was sufficient to induce the expression of oxidative myofibers, suggesting that oxidative myofiber regulation may be engaged at a lower threshold of p38 MAPK/JNK hyperactivation than that required to influence the PTEN/Akt/miR-21 axis.

Because MKP-1 is nuclear localized, its ability to affect JNK-mediated phosphorylation of the insulin substrate receptor-1, and subsequently Akt-mediated glucose uptake, was not likely operable (20). However, MKP1-MKO mice have increased levels of Akt, which led us to uncover a pathway by which skeletal muscle MKP-1 negatively regulates Akt through miR-21 antagonism of PTEN (26). Both p38 MAPK and JNK activities are capable of stimulating not only miR-21 expression but also we found that the precursor, pri-miR-21, was also increased in expression. In addition to transcriptional activation, processing of pri-miRNAs occurs in the nucleus (27), suggesting that a nuclear-localized p38 MAPK/JNK mechanism exists in the regulation of either miRNA transcription and/or processing. Our data are consistent with a role of MKP-1 in miR-21 transcriptional regulation. Whether MKP-1 affects miR-21 processing cannot be ruled out because p38 MAPK has been implicated in the regulation of miRNA processing (41). Given that MKP-1 is overexpressed in skeletal muscle in obesity, these data suggest that insulin resistance could be promoted, at least in part, by MKP-1–mediated suppression of Akt through an MAPK/miR-21/PTEN pathway. It is also conceivable that a similar pathway of MKP-1 overexpression in macrophages in obese humans (33) may also exist, and this, too, contributes to insulin resistance. Interestingly, our studies demonstrate that skeletal muscle MKP-1 impacts hepatic function. We found that the livers of MKP1-MKO mice were protected from the accumulation of triglycerides and development of hepatic steatosis. Skeletal muscle is one of the major tissues responsible for hydrolysis of triglycerides; it is conceivable that protection from hepatic steatosis in the MKP1-MKO mice could be due to enhanced triglyceride hydrolysis in skeletal muscle, leading to reduced triglycerides in the blood and hence the phenotype in the liver.

Taken together, we demonstrate that skeletal muscle MKP-1 serves as a critical signaling node for both p38 MAPK and JNK. In obese states in which skeletal muscle MKP-1 is overexpressed, the data suggest that the integrated output of the p38 MAPK/JNK module is disrupted. Thus, reduced rather than increased p38 MAPK/JNK module signaling in skeletal muscle promotes insulin resistance and metabolic dysfunction.

Acknowledgments. The authors thank members of the Bennett laboratory for helpful discussions during the course of this study.

Funding. This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases grant P01-DK-57751 (to A.M.B.), Yale University’s Mouse Metabolic Phenotyping Center (grant U24-DK-59635), and the Yale Liver Center (grant P30-DK-34989).

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

Author Contributions. A.L., K.M., L.Z., A.C.-D., M.J.J., J.P.G.C., and Y.N. performed experiments and analyzed data. A.L. and A.M.B. designed the experiments, reviewed all of the data, and wrote the manuscript. T.P.G., G.I.S., and C.H.-F. reviewed and edited the manuscript. A.M.B. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.