Human Muscle Fiber Type–Specific Insulin Signaling: Impact of Obesity and Type 2 Diabetes

Skeletal muscle is important for whole-body insulin-stimulated glucose disposal (1), and skeletal muscle insulin resistance is a common phenotype of obesity and type 2 diabetes (T2D) (2). Skeletal muscle is a heterogeneous tissue composed of different fiber types, which can be divided according to myosin heavy chain (MHC) isoform expression. Studies in rodents show that insulin-stimulated glucose uptake in the oxidative type I fiber–dominant muscles is higher than in muscles with a high degree of glycolytic type II fibers (3–6). Whether this phenomenon is due to differences in locomotor activity of individual muscles or a direct consequence of the fiber-type composition is largely unknown. In incubated rat muscle, insulin-induced glucose uptake was higher (∼100%) in type IIa (oxidative/glycolytic) compared with IIx and IIb (glycolytic) fibers (7,8), suggesting that insulin-mediated glucose uptake is related to the oxidative capacity of the muscle fiber. In humans, a positive correlation between proportions of type I fibers in muscle and whole-body insulin sensitivity has been demonstrated (9–11). Furthermore, insulin-stimulated glucose transport in human muscle strips was associated with the relative type I fiber content (12). Thus, it is likely that human type I fibers are more important than type II fibers for maintaining glucose homeostasis in response to insulin. Indeed, a decreased proportion of type I fibers has been found in various insulin resistant states such as the metabolic syndrome (9), obesity (13,14), T2D in some (10,13,14) but not all (12,15) studies and following bedrest (16), as well as in tetraplegic patients (17), and subjects with an insulin receptor gene mutation (18).

Mechanisms for a fiber type–dependent regulation of glucose uptake could involve altered abundance/regulation of insulin-signaling elements and/or metabolic enzymes. In rats, insulin receptor content and Akt and GLUT4 protein abundance are higher in type I compared with type II fiber-dominated muscles (4,5,19–21). Furthermore, in rats, Akt phosphorylation under insulin stimulation is highest in type I compared with type II fiber-dominant muscles (20). In humans, GLUT4 protein levels are higher in type I compared with type IIa and IIx muscle fibers (14,22). Overall, these findings suggest that insulin signaling to and effect on glucose transport is highest in type I fibers. Thus, a shift toward reduced type I and hence higher type II fiber content in obesity and T2D (10,13,14) could negatively influence muscle insulin action on glucose metabolism. Insulin resistance in obesity and T2D is characterized by a decreased ability of insulin to induce signaling proteins proposed to mediate GLUT4 translocation by, for example, phosphorylation/activation of Akt (23–25) and/or TBC1 domain family member 4 (TBC1D4) (23,25). Whether this relates to differences in the response to insulin between fiber types is unknown.

Intracellular glucose metabolism could also be different between muscle fiber types. Glucose entering the muscle cell is initially phosphorylated by hexokinase (HK) and predominantly stored as glycogen or oxidized in the mitochondria through processes regulated by glycogen synthase (GS) and the pyruvate dehydrogenase complex, respectively. HKII content is higher in human soleus muscle (∼70% type I fibers) compared with gastrocnemius and vastus lateralis muscle (∼50% type I fibers) (26). Also, the content of the pyruvate dehydrogenase (PDH) complex subunit E1α (PDH-E1α) is decreased in muscle of proliferator-activated receptor γ-coactivator-1α knockout mice (27), concomitant with a switch toward reduced type I fiber abundance (28). Furthermore, mitochondrial density is higher in human type I compared with type II fibers (29). In contrast, no fiber type–specific expression pattern of GS has been shown (30). Altogether, these observations suggest that glucose phosphorylation and oxidation but not storage rate capacity are enhanced in type I compared with type II fibers. Whether HKII and PDH-E1α abundance as well as GS and PDH-E1α regulation by insulin is different between human muscle fiber types is unknown.

We investigated whether proteins involved in glucose metabolism were expressed and/or regulated by insulin in a fiber type–specific manner in human skeletal muscle. This was achieved by creating pools of single fibers expressing either MHC I (type I) or II (type II). These fibers were dissected from vastus lateralis muscle biopsies obtained from lean and obese normal glucose-tolerant subjects as well as T2D patients.

A total of 10 lean healthy, 11 obese nondiabetic, and 11 obese T2D subjects were randomly chosen from two studies conducted at Odense University Hospital (Odense, Denmark). One fraction (eight lean, seven obese, and six T2D) was from an already published study (31), while the remaining subjects were from an unpublished study, in which subjects were investigated with an identical experimental protocol as previously described (31). Both studies were approved by the regional ethics committee and carried out in accordance with the Declaration of Helsinki II. Subject medication is detailed in the Supplementary Material.

A detailed explanation of the in vivo study protocol has been published elsewhere (31). In short, all subjects were instructed to refrain from strenuous physical activity 48 h before the experimental day. After an overnight fast, subjects underwent a 2-h basal tracer equilibration period followed by a 4-h hyperinsulinemic-euglycemic clamp (Actrapid; Novo Nordisk) at an insulin infusion rate of 40 mU · m⁻² · min⁻¹ combined with tracer glucose and indirect calorimetry. A primed-constant [3-³H]glucose infusion was used throughout the 6-h study, and [3-³H]glucose was added to the glucose infusates to maintain plasma-specific activity constant at baseline levels during the 4-h clamp period as described in detail previously (32). Vastus lateralis muscle biopsies were obtained before and after the clamp under local anesthesia (1% lidocaine) using a modified Bergström needle with suction. Muscle biopsies were immediately frozen in liquid nitrogen and stored below −80°C.

Muscle fibers were prepared as previously described (33) but with minor modifications. A total of 20–60 mg of muscle tissue was freeze-dried for 48 h before dissection of individual muscle fibers in a climate-controlled room (20°C, <35% humidity) using a dissection microscope (in total, n = 5,384 fibers from 64 biopsies). The length of each fiber was estimated under the microscope (1.5 ± 0.4 mm [means ± SD]) before being carefully placed in a PCR tube and stored on dry ice. On the day of dissection, 5 µL of ice-cooled Laemmli sample buffer (125 mmol/L Tris-HCl [pH 6.8], 10% glycerol, 125 mmol/L SDS, 200 mmol/L dithiothreitol, and 0.004% bromophenol blue) was added to each tube. During method optimization, addition of protease and phosphatase inhibitors was found to be unnecessary for preservation of either protein content or protein phosphorylation for this type of sample preparation (data not shown). After thorough mixing at 4°C, each tube was inspected under a microscope to confirm that the fiber was properly dissolved (if not, the tube was discarded). Each sample was then heated for 10 min at 70°C and stored at −80°C.

A small fraction (1/5) of the solubilized fiber was used for identification of MHC expression using Western blotting and specific antibodies against MHC I or II (see Immunoblotting). Hybrid fibers (∼5%) expressing more than one MHC isoform were discarded. Pools of type I and II fibers from each biopsy were prepared (128 pools in total). The average number of type I and II fibers per muscle biopsy included in each pool was 20 (range 9–36) and 42 (range 22–147), respectively.

Protein content of the fiber-specific samples was estimated using 4–20% Mini-PROTEAN TGX stain-free gels (Bio-Rad), which allowed for gel-protein imaging following ultraviolet activation on a ChemiDoc MP Imaging System (Bio-Rad). The intensity of visualized protein bands (from 37–260 kDa) was compared with a standard curve from three different pools of human muscle homogenates with a known protein concentration (Supplementary Fig. 1). After gel imaging, the purity of each pooled sample was re-evaluated using Western blotting and MHC I– and II–specific antibodies (see Immunoblotting). All fiber-specific samples were diluted with Laemmli sample buffer to a protein concentration of 0.2 mg/mL.

Glycogen content in the fiber-specific pools was measured by dot blotting using a specific antibody against glycogen (34,35). Briefly, 150 ng of protein was spotted onto a polyvinylidene difluoride membrane. After air drying, the membrane was reactivated in ethanol before blocking, incubation in primary and secondary antibody, and visualization as described in Immunoblotting. The intensity of each dot was compared with a standard curve (Supplementary Fig. 2) obtained from a muscle homogenate with a glycogen content predetermined biochemically as previously described (31) and expressed accordingly.

For MHC determination in muscle biopsies, lysates were prepared, and protein content was measured as previously described (31). Muscle lysates were diluted 1:3 with 100% glycerol/Laemmli sample buffer (50/50) and run on 8% self-cast stain-free gels containing 0.5% 2,2,2-trichloroethanol (36). A total of 3 µg of lysate protein was separated for ∼16 h at 140 V as previously described (37). Protein bands were visualized by ultraviolet activation of the stain-free gel on a ChemiDoc MP Imaging System (Bio-Rad) and quantified as stated below. Coomassie staining of the gel and the use of muscle homogenates provided similar results as stain-free gel imaging and muscle lysates, respectively (data not shown).

For MHC determination of single muscle fibers and evaluation of total and phosphorylated levels of relevant proteins, equal amounts of sample volume (for MHC determination) or protein amount were separated using either precast (Bio-Rad) or self-cast 7.5% gels. On each gel, an internal control (muscle lysate) was loaded two times per gel in order to minimize assay variation. Muscle fiber pool values were divided by the average of the internal control sample from the corresponding gel. Furthermore, on one gel, a standard curve of muscle homogenate was loaded to ensure that quantification of each protein probed for was within the linear range. Following separation, proteins were transferred (semidry) from multiple gels to a single polyvinylidene difluoride membrane that was incubated with blocking agent (0.05% Tween 20 and 2% skimmed milk in Tris-buffered saline) for 45 min at room temperature, followed by incubation in primary antibody solution overnight at 4°C (for antibody details, see Supplementary Table 1). Membranes were incubated with appropriate secondary antibodies (Jackson ImmunoResearch Laboratories) that were conjugated to either horseradish peroxidase or biotin for 1 h at room temperature. Membranes incubated with biotin-conjugated antibody were further treated with horseradish peroxidase–conjugated streptavidin. Protein bands were visualized using a ChemiDoc MP imaging system (Bio-Rad) and enhanced chemiluminescence (SuperSignal West Femto; Pierce). Band densitometry was performed using Image Laboratory (version 4.0). Membranes were reprobed with an alternate antibody according to the scheme given in Supplementary Table 2.

Subject characteristics and blood parameters were evaluated by a one-way ANOVA. To compare fiber type, insulin, and group effects, a three-way ANOVA with repeated measures for fiber type and insulin was used. If no triple interaction was present, a two-way ANOVA on the increment with insulin (Δinsulin-basal values) was performed for fiber type and group effects with repeated measures for fiber type. Main effects of group and significant interactions were evaluated by Tukey post hoc testing. Statistical analyses were performed in SigmaPlot (version 12.5, Systat Software; one- and two-way ANOVA) and in SAS statistical software (version 9.2, SAS Institute; three-way ANOVA). Unless otherwise stated, n equals number of subjects as indicated in Table 1. Differences were considered significant at P < 0.05.

BMI and fat mass were higher in the obese and T2D groups compared with the lean group (Table 1). Patients with T2D compared with lean and obese subjects had elevated HbA_1c levels, increased fasting plasma glucose, insulin, and triglyceride (vs. lean only) concentrations (Tables 1 and 2). During the hyperinsulinemic-euglycemic clamp, the glucose disposal rate (GDR) was decreased in T2D versus lean and obese subjects (Table 2). The decrease in GDR resulted from both lower glucose oxidation rates and reduced nonoxidative glucose metabolism (Table 2).

In muscle biopsies from lean and obese subjects, MHC I, IIa, and IIx constituted 45, 46, and 9% (total 55% MHC II), respectively (Fig. 1A). This fiber type composition is in accordance with previous observations using (immuno)histochemistry (9–11,13–15,26) and biochemical methods (18,22). In the T2D group, MHC I, IIa, and IIx constituted 35, 45, and 20% (total 65% MHC II), respectively. In the T2D group compared with the lean and obese group, the relative number of type I muscle fibers was lower, and the relative number of type IIx muscle fibers was higher. MHC IIa expression was similar among all three groups.

As represented in Fig. 1B, all fiber pools contained one MHC isoform only. Actin was used as a reference protein, and actin abundance was equal between fiber pools (Supplementary Fig. 3). Higher protein levels of insulin receptor β (16%), HKII (470%), GLUT4 (29%), and electron transport chain complex II (35%) was found in type I versus II fibers (Fig. 1C–F). No differences between groups were observed except for a reduced (−24%) insulin receptor β level in the T2D compared with the lean and obese groups (Fig. 1C–F).

Akt2 protein content was lower (−27%) in type I versus II fibers (Fig. 2C). In the three groups, the average increases under insulin stimulation of phosphorylated (p-)Akt^Thr308 and p-Akt^Ser473 were 5.8- and 3.5-fold in type I fibers and 6.1- and 3.7-fold in type II fibers, respectively (Fig. 2A and B). In lean and obese groups, levels of insulin-stimulated p-Akt^Thr308 were lower (−25%) in type I versus II fibers. In the T2D group, the insulin-stimulated p-Akt^Thr308 and p-Akt^Ser473 were lower in both fiber types compared with lean and obese groups. In response to insulin, phosphorylation of Akt^Ser473/Akt2 but not Akt^Thr308/Akt2 was fiber type dependent, although the relative response to insulin was similar between fiber types (Supplementary Fig. 4A and B). In type I fibers, a higher protein level of mammalian target of rapamycin (mTOR) (20%) and its downstream target N-myc downstream-regulated gene (NDRG) 1 (68%) compared with type II fibers was evident (Fig. 3B and D). Insulin had no effect on p-mTOR²⁴⁸¹ but increased p-NDRG1^Thr346 only in type I fibers from obese (86%) and T2D (100%) groups (Fig. 3A and C). No fiber-type differences were evident when p-NDRG1^Thr346 was adjusted for NDRG1 protein abundance (Supplementary Fig. 4C).

TBC1D1 and TBC1D4 protein levels were lower (−45% and −16%) in type I versus II fibers, respectively (Fig. 4B and G). Irrespective of fiber type, insulin stimulation increased p-TBC1D1^Thr596 (36%) and p-TBC1D4 at all sites investigated (Ser³¹⁸ [122%], Ser⁵⁸⁸ [59%], Thr⁶⁴² [103%], and Ser⁷⁰⁴ [113%]) (Fig. 4A and C–F). Statistically significant main effects of fiber type were evident for the level of phosphorylation of both TBC1D1 and TBC1D4. More specifically, p-TBC1D1^Thr596 (−62%), p-TBC1D4^Ser318 (−21%), p-TBC1D4^Ser588 (−21%), p-TBC1D4^Thr642 (−24%), and p-TBC1D4^Ser704 (−24%) were lower in type I compared with type II fibers. No significant group differences in protein abundance or protein phosphorylation of TBC1D1 and TBC1D4 were evident, although the response to insulin of p-TBC1D4^Ser588 tended (P = 0.07) to be group dependent.

In the basal state, glycogen content was lower (−29%) in type I versus II fibers in the lean (P < 0.001), obese (P = 0.09), and T2D (P = 0.09) groups (Fig. 5A). Insulin induced no significant changes in glycogen content in either of the fiber types. The protein levels of GS kinase (GSK) 3β were 14% less in type I versus type II, whereas GS protein was 53% higher in type I compared with type II fibers (Fig. 5C and F). In all three groups and in both fiber types, insulin induced a similar change in phosphorylation of GSK3β^Ser9 (62%), GS^2+2a (−36%) and GS^3a+b (−38%) (Fig. 5B, D, and E). Phosphorylation of GSK3β^Ser9 was lower (−31%), whereas phosphorylation of GS^site2+2a and GS^site3a+b was, respectively, 68 and 51% higher in type I versus II fibers. No significant differences were evident between individual groups in protein abundance and protein phosphorylation of GSK3β and GS.

PDH-E1α protein content was 34% higher in type I versus II fibers (Fig. 6C). Basal levels of PDH-E1α site 1 phosphorylation were similar between fiber types in all three groups (Fig. 6A). After insulin, the degree of phosphorylation was significantly lower in type II versus I fibers in the obese and T2D groups only, indicating dephosphorylation by insulin in type II but not in type I fibers. In line, PDH-E1α site 2 phosphorylation was decreased by insulin, and this effect was dependent on fiber type toward a greater effect of insulin in type II versus I fibers (Fig. 6B). Fiber-type differences were not evident when p-PDH^site1 and p-PDH^site2 was adjusted for PDH-E1α content (Supplementary Fig. 4D and E).

The current study is the first to evaluate changes in signaling events in response to insulin in fiber type–specific pools from human muscle. Based on our findings, we propose a model in which human type I fibers have a greater abundance of proteins to transport (29% GLUT4), phosphorylate (470% HKII), and oxidize (35% electron transport chain complex II and 34% PDH) glucose and to synthesize glycogen (35% GS) compared with type II fibers. These observations are supported by significant positive correlations between the MHC I content in whole muscle lysates and insulin-stimulated GDR (r = 0.53; P = 0.002), glucose oxidation rate (r = 0.52; P = 0.003), and nonoxidative glucose metabolism (r = 0.44; P = 0.01) (Supplementary Fig. 5). Interestingly, even though insulin receptor content was higher (16%) in type I fibers, phosphoregulation of TBC1D1, TBC1D4, and GS by insulin was similar between fiber types (all normalized to actin). The apparent fiber-type differences in insulin-stimulated phosphorylation of Akt, NDRG1, and PDH-E1α (when related to actin) were eliminated when adjusted for Akt2, NDRG1, and PDH-E1α protein abundance. These findings suggest a similar sensitivity of type I and II muscle fibers for regulation by insulin of the proteins investigated.

Insulin-stimulated GDR, glucose oxidation rates, and nonoxidative glucose metabolism were decreased in T2D compared with the lean and obese groups. This was accompanied by lower insulin receptor content and altered response to insulin of p-Akt³⁰⁸, p-Akt⁴⁷³, p-TBC1D4^Ser588 (P = 0.07), and p-NDRG1^Thr346 in the muscle fiber–specific pools from the T2D compared with the lean and obese groups. In cells, NDRG1 phosphorylation has been suggested to be a readout of mTOR complex (mTORC) 2 activities (38). mTORC2 is also a widely accepted upstream kinase for Akt^Ser473 (39). Since the response to insulin of p-NDRG1^Thr346/NDRG1 was similar between groups, these data could imply a specific dysfunctional link between mTORC2 and p-Akt^Ser473, as the latter was decreased in response to insulin in both type I and II fibers in T2D compared with the lean and obese groups. In rat muscle, abundance and insulin-stimulated phosphorylation of Akt were higher (660 and 160–180%, respectively) in soleus muscle primarily containing type I fibers, as opposed to epitrochlearis and extensor digitorum longus muscles primarily consisting of type II fibers (20). In contrast, in human muscle, we report a decreased Akt phosphorylation after insulin in type I versus II fibers, due to higher Akt2 levels in type II fibers. Thus, findings in rat muscles with a diverse fiber-type composition could simply result from differences in locomotor activity, although species-related differences cannot be excluded. For instance, TBC1D4 and TBC1D1 protein abundance in the current study are only modestly lower (−16% and −45%) in human type I versus II fibers. In mice, a high (>10-fold) TBC1D4 and a low (<20%) TBC1D1 content are evident in the type I fiber–abundant soleus compared with the type II fiber–abundant extensor digitorum longus muscle (40). In rats, no significant correlations between MHC isoform abundance in various muscles and either TBC1D1 or TBC1D4 protein content were found (21). These findings indicate that fiber-type differences in TBC1D4 and TBC1D1 protein levels are highly dependent on the species investigated.

In the current study, no differences in the response to insulin were observed between fiber types in phosphorylation of TBC1D4 and TBC1D1. We previously reported a decreased response to insulin of p-TBC1D4^Ser318 and p-TBC1D4^Ser588 in skeletal muscle from obese T2D subjects compared with weight-matched control subjects (23). In the current study, insulin-induced (delta values [insulin minus basal]) p-TBC1D4^Ser588 was borderline (P = 0.07) group dependent. The average response to insulin was 62, 96, and 19% in the lean, obese, and T2D groups, respectively. It has been shown that exercise training normalizes defects in insulin action on TBC1D4 regulation in T2D versus control subjects (23). Thus, in the current study, the lack of significant defects in TBC1D4 regulation by insulin in the T2D group compared with control groups could be due to the physical fitness level of the groups studied. We found that p-TBC1D1^Thr596 was increased by insulin in agreement with another study (41) and that the relative increase was irrespective of fiber type and group. We conclude that the relative response to insulin of Akt, TBC1D4, and TBC1D1 is independent of fiber type, while the absolute amount of phosphorylated protein is lower in type I versus II fibers. Whether a higher total amount of phosphorylated protein is important for the regulation of glucose uptake is unknown. To investigate the impact of the present findings on glucose uptake in different human muscle fiber types, future studies need to examine the membrane-bound fraction of GLUT4 in different fiber types or even measure single muscle fiber glucose transport as performed in rat muscle (7).

Interestingly, Gaster et al. (14) previously reported that GLUT4 abundance was significantly lower in type I fibers only in muscle from T2D patients compared with lean and obese control subjects. This was not evident in the current study. However, we found a nonsignificantly lower GLUT4 content of the same magnitude (10–20%) as previously reported (14) in both type I and II fibers from the T2D compared with the lean and obese groups. Also, GLUT4 levels were generally higher in type I versus II fibers. Thus, fewer type I fibers in the T2D compared with the lean and obese groups possibly lowers the glucose uptake capacity in diabetic skeletal muscle. In support, HKII content was higher in type I compared with type II fibers. The influence of HKII protein levels on glucose uptake is controversial and has recently been estimated to control ∼10% of human skeletal muscle glucose metabolism during insulin-stimulated conditions (42). In the current study, fiber type–specific HKII levels were not different between groups investigated. Thus, it is likely that decreased HKII levels reported in muscles from T2D subjects (43) are at least partly influenced by a lower number of type I fibers in T2D versus control subjects as also shown in the current study. Interestingly, in contrast to HKII, HKI protein abundance was lower (−19%) among the three groups in type I versus II fibers (Supplementary Fig. 6). This observation could indicate a different role of HK isoforms in type I and II muscle fibers.

A close correlation between the insulin-stimulated increase in nonoxidative glucose metabolism and GS activity has been reported (44). In the current study, insulin-stimulated nonoxidative glucose metabolism was decreased in the T2D compared with the lean and obese groups as shown by others (23,31,41,45). Thus, we investigated the fiber type–specific regulation of GS by insulin. We were unable to detect any differences in the response to insulin between fiber types, although the absolute amount of phosphorylated GS was highest in type I fibers. Increased phosphorylation of GS in type I fibers could be accounted for by a higher GS protein level in type I versus II fibers. Previously, a similar GS content in type I, IIa, and IIx fiber pools was reported in muscle from young (23 years) subjects (30). Thus, the present findings of a higher GS content in type I versus II fibers in muscle from middle-aged (∼55 years) subjects indicates an age-dependent fiber type–specific regulation of GS abundance. The functional consequence of a differentiated GS content between fiber types is unknown, since we were unable to detect any differences in basal and insulin-stimulated glycogen content in both fiber types. This is likely due to the relatively small (<6%) increase in glycogen content during a clamp procedure (46). If glycogen levels were solely dependent on GS, the activity of this enzyme would be expected to be lower in type I versus II fibers. However, our data cannot support this because the higher expression and phosphorylation of GS indicates that total GS activity is in fact higher in type I versus II fibers. Thus, other factors than GS activity per se determines glycogen levels.

In a recent study, Nellemann et al. (47) did not find any changes in phosphorylation of PDH-E1α in human skeletal muscle in response to insulin. Interestingly, in the current study, PDH-E1α phosphorylation was decreased by insulin in type II fibers only. Thus, results by Nellemann et al. (47) could have been influenced by a muscle fiber type–dependent regulation not detected in their whole-muscle biopsy preparation. An inverse relationship between PDH-E1α phosphorylation and PDHa activity has been shown in human skeletal muscle during exercise (48). Thus, findings in the current study suggest an increased PDHa activity in response to insulin in type II fibers only.

All fiber pools were prepared from vastus lateralis muscle, which expresses relatively small (<10%) amounts of type IIx fibers (26). No significant differences in the MHC IIx expression were observed between type II fiber pools among the three groups (Supplementary Fig. 7). Thus, differences between type I and II fiber pools observed in the current study are likely not influenced by differences in protein abundance/regulation between type IIa and IIx fibers. No measure of physical activity was performed. It has been shown that training-induced increases in GLUT4 content mainly occur in type I fibers (22). Thus, training status of the subjects in the current study could potentially influence differences between muscle fibers and/or groups. All measures were performed in muscle fibers from the vastus lateralis muscle. Whether fiber type–specific differences in protein expression can be extended to other muscles is unknown, but has been challenged by one study (30), in which GLUT4 expression was higher in type I versus IIa and IIx fibers from vastus lateralis muscles but similar between fiber types in soleus and triceps brachii muscles. The current study design did not allow exploration of this further. To evaluate the biological impact of fiber-specific signaling events further, the methods used in the current study could be combined with ex vivo incubation of human muscle strips (12) and the recently described method of single-fiber glucose uptake measurements (7). Such design demands open surgical biopsies and was therefore not applicable to the cohort of the current study.

In conclusion, based on protein level measures, the enzymatic capacities for glucose uptake, phosphorylation, and oxidation as well as for glycogen synthesis are higher in human type I compared with type II muscle fibers. In response to insulin, most differences in phosphorylation between fiber types were due to differences in protein levels. Thus, sensitivity for phosphoregulation by insulin of these proteins is similar between fiber types. Even though insulin-induced GDR was decreased in patients with type 2 diabetes compared with lean and obese subjects, few group differences in the muscle fiber–specific measurements were observed. However, our observations favor the idea that fewer type I fibers and a higher number of type IIx fibers in muscles from T2D patients contributes to the reduced GDR under insulin-stimulated conditions compared with lean and obese subjects.

Acknowledgments. The authors thank M. Kleinert (University of Copenhagen, Denmark) for sharing know-how on the mTOR/NDRG1 analyses. The authors also thank the following for the donation of material essential for this work: L.J. Goodyear (Joslin Diabetes Center and Harvard Medical School, Boston, MA), O.B. Pedersen (University of Copenhagen, Denmark), and J. Hastie and D.G. Hardie (University of Dundee, U.K.). The monoclonal antibodies against MHC I and II isoforms (A4.840 and A4.74) were developed by H.M. Blau, and antibody directed against MHC IIx (6H1) was developed by C. Lucas. All MHC antibodies were obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the National Institute of Child Health and Human Development and maintained by The University of Iowa, Department of Biology, Iowa City, IA.

Funding. This work was carried out as a part of the research programs “Physical activity and nutrition for improvement of health” funded by the University of Copenhagen Excellence Program for Interdisciplinary Research and the UNIK project Food, Fitness & Pharma for Health and Disease (see www.foodfitnesspharma.ku.dk) supported by the Danish Ministry of Science, Technology and Innovation. This study was funded by the Danish Council for Independent Research Medical Sciences, the Novo Nordisk Foundation, and a Clinical Research Grant from the European Foundation for the Study of Diabetes.

Duality of Interest. P.H.A. and J.N. are employees at Novo Nordisk A/S and own stocks in Novo Nordisk A/S. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. P.H.A. was responsible for conception and design of research, performed analysis, interpreted results, drafted the manuscript, edited and revised the manuscript, and approved the final version. A.J.T.P. performed in vivo experiments and analysis, edited and revised the manuscript, and approved the final version. J.B.B. performed analysis, interpreted results, edited and revised the manuscript, and approved the final version. D.E.K. performed analysis, edited and revised the manuscript, and approved the final version. B.F.V. performed in vivo experiments and analysis, edited and revised the manuscript, and approved the final version. O.B. and J.N. edited and revised the manuscript and approved the final version. K.H. interpreted results, edited and revised the manuscript, and approved the final version. J.F.P.W. was responsible for conception and design of research, interpreted results, drafted the manuscript, edited and revised the manuscript, and approved the final version. J.F.P.W. 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.