Skeletal muscle fiber type-selective effects of acute exercise on insulin-stimulated glucose uptake in insulin-resistant, high-fat-fed rats

Abstract

Insulin-stimulated glucose uptake (GU) by skeletal muscle is enhanced several hours after acute exercise in rats with normal or reduced insulin sensitivity. Skeletal muscle is composed of multiple fiber types, but exercise’s effect on fiber type-specific insulin-stimulated GU in insulin-resistant muscle was previously unknown. Male rats were fed a high-fat diet (HFD; 2 wk) and were either sedentary (SED) or exercised (2-h exercise). Other, low-fat diet-fed (LFD) rats remained SED. Rats were studied immediately postexercise (IPEX) or 3 h postexercise (3hPEX). Epitrochlearis muscles from IPEX rats were incubated in 2-deoxy-[³H]glucose (2-[³H]DG) without insulin. Epitrochlearis muscles from 3hPEX rats were incubated with 2-[³H]DG ± 100 µU/ml insulin. After single fiber isolation, GU and fiber type were determined. Glycogen and lipid droplets (LDs) were assessed histochemically. GLUT4 abundance was determined by immunoblotting. In HFD-SED vs. LFD-SED rats, insulin-stimulated GU was decreased in type IIB, IIX, IIAX, and IIBX fibers. Insulin-independent GU IPEX was increased and glycogen content was decreased in all fiber types (types I, IIA, IIB, IIX, IIAX, and IIBX). Exercise by HFD-fed rats enhanced insulin-stimulated GU in all fiber types except type I. Single fiber analyses enabled discovery of striking fiber type-specific differences in HFD and exercise effects on insulin-stimulated GU. The fiber type-specific differences in insulin-stimulated GU postexercise in insulin-resistant muscle were not attributable to a lack of fiber recruitment, as indirectly evidenced by insulin-independent GU and glycogen IPEX, differences in multiple LD indexes, or altered GLUT4 abundance, implicating fiber type-selective differences in the cellular processes responsible for postexercise enhancement of insulin-mediated GLUT4 translocation.

INTRODUCTION

In 2015, an estimated 30.3 million individuals had diabetes in the U.S. and another 84.1 million had prediabetes (15). Skeletal muscle insulin resistance is an essential defect leading to type 2 diabetes, and skeletal muscle is responsible for 60–80% of insulin-mediated glucose disposal (17, 44), making muscle a preeminent target to combat insulin resistance. Independently of diabetes, insulin resistance is linked to many prevalent and devastating pathologies, including atherogenesis, hypertension, cognitive dysfunction, and some cancers (20, 25, 36). Exercise by either humans or rodents leads to increased insulin-stimulated glucose uptake by skeletal muscle that is evident ~1–3 h postexercise and can persist for up to 48 h (4, 6, 9, 11, 22, 51, 53, 55, 62, 67). Both insulin-sensitive (1, 53, 54, 55) and insulin-resistant (12, 18, 19, 51) rats and humans can experience this exercise benefit. However, acute exercise does not completely bring glucose uptake of insulin-resistant muscle to the level observed in healthy muscle after the same exercise protocol (8, 12). The present study focuses on the enhanced insulin-stimulated glucose uptake by insulin-resistant muscle after exercise, because attenuating or eliminating muscle insulin resistance would be expected to have major implications for improving health.

By definition, glucose uptake is a cellular process, so it is essential to understand exercise effects on muscle glucose uptake at the cellular level. Moreover, because skeletal muscle is composed of multiple fiber types based on myosin heavy chain (MHC) expression (I, IIA, IIX, and IIB) with different metabolic properties (52), exercise may exert different fiber type-specific effects during insulin-resistant versus insulin-sensitive conditions. Conventionally, researchers compare whole muscles or muscle portions with markedly different fiber type proportion to delineate fiber type differences. However, tissue analysis cannot reveal fiber type differences at the cellular level because of several caveats. 1) No muscle is composed entirely of a single fiber type. 2) No rat muscle has been identified that is mostly type IIX fibers. 3) It is impossible to evaluate hybrid fibers (fibers expressing more than one MHC) in whole muscles. The physiological relevance of evaluating hybrid fibers is undeniable, given that, although they can account for a substantial percentage of myofibers (>10–50%) in various species, including rats, mice, and humans (7, 60, 64, 68), almost nothing is known about their metabolic properties in response to in vivo exercise. 4) Muscle tissue contains cell types other than muscle fibers (vascular, neural, connective, and adipose cells), which contribute to measurements in tissues. We developed and validated the only method that enables glucose uptake [based on 2-deoxy-[³H]glucose (2-[³H]DG) accumulation] and MHC measurement (using SDS polyacrylamide electrophoresis) in individual muscle fibers that are isolated by microdissection after collagenase treatment of the intact rat epitrochlearis muscle (39). The specificity of glucose uptake by single fibers was established based on cytochalasin B’s inhibition of 2-[³H]DG accumulation.

Using this method, we recently reported that, in rats with normal insulin sensitivity, acute exercise induced fiber type-specific effects on insulin-stimulated glucose uptake (10). We discovered increased insulin-stimulated glucose uptake in all fiber types, except type IIX fibers, postexercise. Subsequently, we found that a 2-wk high-fat diet (HFD) induced a fiber type-selective decrease in insulin-stimulated glucose uptake (48). There was no insulin resistance in type I fibers, slight insulin resistance in IIA fibers, and substantial insulin resistance in the other fibers (IIB, IIBX, IIX, and IIAX) (48). In whole epitrochlearis muscles from HFD-fed rats, exercise elevated the insulin-stimulated glucose uptake rate to values that were equal to those found in muscles from unexercised rats on a control, low-fat diet (12). However, the exercise effect on fiber type-specific insulin-stimulated glucose uptake during insulin resistance has not been reported, and it is impossible to predict whether this whole muscle outcome in insulin-resistant muscle was attributable to a uniform effect in all fiber types or was a fiber type-selective benefit. Therefore, our primary aim was to determine the exercise effect on insulin-stimulated glucose uptake in different fiber types from insulin-resistant rats. We evaluated fiber type-specific effects of a HFD and exercise (3 h postexercise, 3hPEX) on insulin-stimulated glucose uptake by using muscles from rats that consumed a 2-wk HFD and were sedentary (HFD-SED and low-fat diet, LFD-SED) controls. This brief dietary intervention produces muscle insulin resistance before large changes in body mass, body composition, or other outcomes that may obscure the primary mechanisms responsible for muscle insulin resistance (61). We also evaluated the effects of the exercise protocol on two separate indirect markers of fiber recruitment (insulin-independent glucose uptake and muscle glycogen immediately postexercise, IPEX).

Intramyocellular lipid (IMCL) content correlates with insulin resistance under some circumstances (35), but this relationship can be disrupted [e.g., endurance exercise training can increase IMCL content and also improve insulin sensitivity (24)]. Recent evidence has suggested that lipid droplet (LD) size and/or subcellular location may be more predictive of muscle insulin resistance than total IMCL content (13, 45, 46). However, previous studies have not evaluated LD size or location in different fiber types for which insulin-stimulated glucose uptake was also known. Therefore, another aim was to identify whether LD size and/or location are related to insulin-stimulated glucose uptake in different fiber types from HFD-fed rats postexercise along with HFD-SED and LFD-SED controls.

Skeletal muscle insulin-stimulated glucose uptake depends on GLUT4 protein expression (30, 37). We (12) previously found unaltered GLUT4 abundance in whole muscles 3hPEX in rats. However, increased muscle GLUT4 abundance was reported 3hPEX in humans (34). It seemed possible that exercise might induce a fiber type-specific increase in GLUT4 abundance that would be obscured by tissue analysis. Therefore, our final aim was to assess possible differences in fiber type-specific GLUT4 abundance postexercise from HFD-3hPEX, HFD-SED, and LFD-SED rats.

METHODS

Materials.

Reagents and apparatus for SDS-PAGE and nonfat-dry milk were from Bio-Rad (Hercules, CA). 2-[³H]DG and [¹⁴C]mannitol were from PerkinElmer (Waltham, MA). Tissue Protein Extraction Reagent (T-PER), and SimplyBlue SafeStain were from ThermoFisher (Pittsburgh, PA). Collagenase type 2 (305 U/mg) was from Worthington Biochemical (Lakewood, NJ). Periodic Acid-Schiff (PAS) kit and glycogen (type IX, bovine liver) were from Sigma-Aldrich (St. Louis, MO). Antibodies and fluorescent probes are listed in Table 1.

Table 1.

Antibodies and fluorescent probes

Antibody/Probe	Catalog No.	Source	Primary or Secondary	Dilution	Use
Anti-GLUT4	CBL-243	EMD Millipore (Billerica, MA)	Primary	1:100	Immunoblotting
Anti-rabbit IgG	7074	Cell Signaling Technology (Danvers, MA)	Secondary	1:20,000	Immunoblotting
Anti-MHC type IIB (IgM)	BF-F3	DSHB Univ. of Iowa (Iowa City, IA)	Primary	1:100	Immunohistochemistry
Anti-MHC type IIA (IgG1)	SC-71	DSHB Univ. of Iowa (Iowa City, IA)	Primary	1:500	Immunohistochemistry
Anti-MHC type I (IgG2b)	BA-D5	DSHB Univ. of Iowa (Iowa City, IA)	Primary	1:50	Immunohistochemistry
Alexa Fluor 555 IgM	A-21043	ThermoFisher (Pittsburgh, PA)	Secondary	1:500	Immunohistochemistry
Alexa Fluor 647 IgG1	A-21240	ThermoFisher (Pittsburgh, PA)	Secondary	1:500	Immunohistochemistry
Alexa Fluor 350 IgG2b	A-21140	ThermoFisher (Pittsburgh, PA)	Secondary	1:500	Immunohistochemistry
BODIPY 493/503, 2 mg/ml	D3922	ThermoFisher (Pittsburgh, PA)	N/A	1:500	Fluorescent Probe
WGA-AF488, 1 mg/ml	W11261	ThermoFisher (Pittsburgh, PA)	N/A	1:200	Fluorescent Probe
WGA-AF555, 1 mg/ml	W32464	ThermoFisher (Pittsburgh, PA)	N/A	1:200	Fluorescent Probe

Summary of antibodies and fluorescent probes used for immunoblotting and immunohistochemistry. All antibodies were validated by the vendor. MHC, myosin heavy chain; WGA-AF488/555, wheat germ agglutinin-Alexa Fluor 488/555.

Animal treatment and muscle preparation.

Procedures for animal care were approved by the University of Michigan Committee on Use and Care of Animals. Male Wistar rats (6–7 wk old; Charles River Laboratories, Boston, MA) were individually housed and randomly provided standard rodent chow (LFD: 14% kcal fat, 58% kcal carbohydrate, 5L0D; Laboratory Diet, St. Louis, MO) or HFD (60% kcal fat, 20% kcal carbohydrate, D12492; ResearchDiets, New Brunswick, NJ) for 2 wk ad libitum until fasted at ~1700 on the night before the experiment. Caloric intake was estimated based on food provided on day 1 and food remaining at ~1700 on the night before the experiment. Beginning at 0700 on the experimental day, rats remained sedentary or swam in a barrel filled with water (35°C) for four 30-min bouts (5-min rest between bouts), as previously described (10, 12, 63). Some rats were anesthetized (pentobarbital sodium, 50 mg/kg body wt ip) immediately postexercise (IPEX and sedentary controls) and weighed, and epitrochlearis muscles were dissected and used either for immunohistochemical analyses or for measuring single fiber insulin-independent glucose uptake. Other rats were anesthetized ~3 h postexercise (3hPEX and sedentary controls) and weighed, and epitrochlearis muscles were dissected and used for measuring insulin-stimulated glucose uptake and fiber type in single fibers. Epididymal fat pads were dissected and weighed. Muscles used for immunohistochemical analyses were embedded in Tissue-Tek (Sakura, Torrance, CA), snap-frozen in liquid nitrogen-cooled isopentane, and stored (−80°C) until analyzed.

Ex vivo muscle incubations.

Dissected muscles used for single-fiber glucose uptake were incubated in glass vials gassed (95% O₂-5% CO₂) in a temperature-controlled bath for a four-step process (35°C during steps 1, 2, and 4; step 3 was on ice) essentially as previously described (10, 48). For muscles from the IPEX experiment, in the first step, isolated muscles were placed for 10 min in vials containing 2 ml of medium 1 [Krebs-Henseleit buffer (KHB) supplemented with 0.1% bovine serum albumin (BSA), 2 mM sodium pyruvate, and 6 mM mannitol]. In the second step, the muscles were incubated for 30 min in a vial containing 2 ml of step 2 medium [KHB supplemented with 0.1% BSA, 0.1 mM 2-DG (13.5 mCi/mmol 2-[³H]DG), 2 mM sodium pyruvate, and 6 mM mannitol]. In the third step, muscles were rinsed three times (5 min/rinse with shaking at 115 rpm) in ice-cold rinse medium (Ca²⁺-free KHB with 0.1% BSA and 8 mM glucose) to wash 2-[³H]DG from the extracellular space. During the fourth step, muscles were incubated for 60 min in vials that included collagenase medium (rinse medium with 8 mM glucose and 2.5% type 2 collagenase) to enzymatically digest muscle collagen. For the muscles from the 3hPEX experiment, during the first step, the medium included KHB plus 2 mM sodium pyruvate and 6 mM mannitol with either 0 or 100 µU/ml insulin. During the second step, muscles were incubated for 60 min in medium that was the same as the in first step but supplemented with 0.1 mM 2-deoxy-d-glucose and 13.5 mCi/mmol 2-[³H]DG. The third and fourth steps in the 3hPEX experiment were identical to the third and fourth steps for the IPEX experiment.

Isolation and processing of single fibers.

Under a dissecting microscope (EZ4D; Leica, Buffalo Grove, IL), intact single fibers (~55 fibers/muscle) were gently isolated using fine forceps (47). After isolation, each fiber was imaged using a camera-enabled microscope with Leica Application Suite EZ software. After imaging, fibers were transferred into individual tubes and processed for glucose uptake, MHC, and immunoblotting as previously described (47, 48).

Glucose uptake and MHC.

Each lysed fiber was processed for glucose uptake and MHC isoform identification as previously described (10, 47).

Glycogen.

External standards of known glycogen concentrations (2, 10, 25, 50, 100, and 150 mmol/l) were created as previously described (57). PAS staining of 8-µm sections of these external glycogen standards was performed on all slides to reliably compare staining of muscle sections on separate slides.

Muscles were serially sectioned at 8 µm for PAS staining and MHC detection. Slides used for PAS staining were fixed in 10% formalin (1 h, 4°C). Following fixation, slides were treated with 1% periodic acid (room temperature, 5 min) and washed (deionized water, 1 min) before Schiff’s reagent was applied (10 min). Slides were then quickly washed (deionized water, ~5 s) and gently rinsed in tap water (10 min) before being mounted with Dako mounting medium. The muscle sections used for MHC identification were processed using appropriate primary and secondary antibodies. Antibodies against MHCI, MHCIIa, and MHCIIb were used for the immunofluorescent detection of fiber type. MHCIIx was identified by the absence of signal among the other three MHC isoforms, as previously described (5). After exposure to secondary antibodies, slides were incubated in WGA-AF488 (10 min), washed in PBS, and mounted (Dako mounting medium).

Lipid droplet analysis.

Muscles were serially sectioned (8 µm) and thaw-mounted on glass slides (one section from each treatment on each slide to minimize staining variability). One slide was used for MHC identification and processed as described above. The other slide was used for BODIPY staining. BODIPY is a lipophilic dye that is widely used to stain neutral lipid droplets (LDs) (16, 59, 65). The slides used for BODIPY staining were immediately fixed in 10% formalin (1 h, room temperature). Slides were then washed (PBS), incubated in a 1:500 dilution of 2 mg/ml BODIPY 493/503 (30 min), and washed again (PBS). Finally, slides then were stained with wheat germ agglutinin-Alexa Fluor 555 (WGA-AF555), washed (PBS), and mounted (Dako mounting medium).

Image capture and processing.

Images of cross-sectionally oriented muscles stained with PAS or MHC were imaged (×20) using a Zeiss Apotome microscope capable of wide-field fluorescent and bright-field illumination coupled to a high-resolution axiocam black and white camera system with DAPI, GFP, TRITC, and Cy5 filters. For BODIPY-stained muscle sections (and corresponding MHC sections), a Nikon Confocal A1 microscope was used at ×60 magnification (×20 for MHC imaging) to image fluorescently labeled LDs and MHC expression. Image capture settings and conditions were kept constant to minimize variability. At least three images from different regions of each muscle section were taken for fiber analysis. Two researchers independently identified MHC expression of each fiber to confirm accurate fiber typing. After the fiber type was identified for PAS-stained images, up to 10 fibers of each fiber type were manually traced and quantified for PAS stain intensity from each image. For BODIPY-stained images, regions of interest representing individual fibers were identified by the WGA lectin extracellular matrix marker using ImageJ. Fiber borders were automatically identified using the “Analyze Particles” feature, and up to eight fibers of each fiber type were used for LD analysis from a given image. An automated threshold representing positive LD signal was set using ImageJ and applied uniformly to all images. LD density was quantified as the percentage covered by BODIPY stain. The region 1 µm below the cell border (subsarcolemmal) was quantified for lipid density using the “Enlarge” and “Make Band” tools in ImageJ. LD size was determined using the “Analyze Particles” tool in ImageJ. Fiber cross-sectional area (CSA) was quantified using BODIPY-stained images.

Immunoblotting.

To measure fiber type-specific GLUT4 abundance, all of the single-fiber lysates (nonheated) expressing the same MHC from a muscle from each rat from each group (LFD-SED, HFD-SED, and HFD-3hPEX) were pooled together for GLUT4 immunoblotting. Aliquots of pooled fiber lysates were separated by SDS-PAGE using 10% gels and then transferred to polyvinylidene difluoride membranes. After electrotransfer, gels were stained (SimplyBlue SafeStain,1 h, room temperature), then destained (deionized water, 2 h). SimplyBlue-stained MHC bands quantified by densitometry (AlphaView; ProteinSimple, San Leandro, CA) were the loading controls for immunoblotted proteins (40, 41). Membranes were incubated with primary and secondary antibodies and subjected to enhanced chemiluminescence to quantify protein bands by densitometry. Individual values were normalized to the mean value of all samples on the membrane and divided by the corresponding MHC loading control.

Statistics.

Data are expressed as means ± 95% confidence interval (95% CI), with two-tailed significance levels of α < 0.05. Two-tailed t-tests were used to determine the diet effect on daily caloric intake, epididymal fat pad mass, and body mass. One-way ANOVA was used to determine the treatment group effect for GLUT4 abundance in pooled fibers of each fiber type from each muscle. Because single-fiber glucose uptake, glycogen, LD measurements, and CSA were collected from multiple individual fibers per rat, we analyzed these data using mixed-effects linear regression models, incorporating fixed parameters evaluating the contributions of treatment group (LFD-SED, HFD-SED, HFD-IPEX, or HFD-3hPEX) and insulin (insulin, no insulin; for glucose uptake only) and interaction effects, random Y-intercepts to account for multiple observations within each rat. The analysis revealed main effects of insulin dose or treatment (diet or exercise) group [i.e., the effect of one independent variable (insulin or treatment group) on the dependent variable (e.g., glucose uptake) distinct from the other independent variables (insulin or treatment group)]. The analysis also revealed whether there were significant insulin × group (diet or exercise) interaction effects (i.e., if the magnitude of the insulin effect on the dependent variable (e.g., glucose uptake) was different between the treatment groups). Analyses were performed using StataIC 14.2 (College Station, TX).

RESULTS

Caloric intake, body mass, and fat pad mass.

Estimated daily caloric intake (kcal/day) was 22% greater (P < 0.001) in HFD (105.8 ± 4.8) vs. LFD (86.4 ± 3.8) rats. Body mass (g) was not significantly different (P = 0.116) for the HFD (313 ± 12) vs. LFD (299 ± 10) animals. Epididymal fat pad mass (g) was greater (P < 0.001) for HFD (6.1 ± 0.7) vs. LFD rats (4.0 ± 0.4).

MHC isoform expression.

MHC isoform expression in each of the isolated single fibers was determined using SDS-PAGE followed by protein staining (Fig. 1).

Fig. 1.

Representative image of SDS-PAGE gel of single fibers for the identification of fiber type. Lanes 1 and 8: E+S [pooled extensor digitorum longus (EDL) + soleus standard expressing myosin heavy chain (MHC) types I, IIA, IIX, and IIB]. Lane 2: type I fiber. Lane 3: type IIA fiber. Lane 4: type IIAX fiber. Lane 5: type IIX fiber. Lane 6: type IIBX fiber. Lane 7: type IIB fiber.

Glucose uptake.

For the HFD-IPEX group vs. HFD-SED controls, insulin-independent glucose uptake was increased (P < 0.05) in all fiber types (Fig. 2). For the 3hPEX experiment, there was a significant main effect of insulin (P < 0.001) on glucose uptake in each fiber type (Fig. 3). In LFD-SED vs. HFD-SED muscles, there was a significant insulin × treatment group interaction on glucose uptake in type IIAX, IIX, IIBX, and IIB fibers, where HFD reduced insulin-stimulated glucose uptake in these fiber types. In HFD-SED vs. HFD-3hPEX muscles, there was a significant insulin × treatment group interaction in type IIA, IIAX, IIX, IIBX, and IIB fibers, where insulin-stimulated glucose uptake was greater for HFD-3hPEX. There was a significant insulin × treatment group interaction for LFD-SED vs. HFD-3hPEX in type IIA and IIBX fibers, where insulin-stimulated glucose uptake was greater for HFD-3hPEX rats.

Fig. 2.

Insulin-independent 2-deoxyglucose (2-DG) uptake measured immediately postexercise in single fibers of each fiber type from rats. Bars represent means of all fibers within a given treatment group [high-fat diet + sedentary (HFD-SED) or HFD + immediately postexercise (HFD-IPEX)]. Error bars are means ± 95% confidence interval. *P < 0.05, HFD-IPEX vs. HFD-SED. Numbers of rats used for the groups in this experiment were HFD-SED (n = 8) and HFD-IPEX (n = 8). All fibers isolated from each muscle (~40–56 fibers/muscle) were used to determine both 2-DG uptake and MHC expression. However, not every fiber type was included in the fibers sampled from every muscle. Therefore, the number of rats (HFD-SED/HFD-IPEX) from which fibers were isolated of each fiber type were: type I (3/7), type IIA (7/7), type IIAX (7/7), type IIX (8/8), type IIBX (8/7), and type IIB (8/8). Numbers of fibers (HFD-SED/HFD-IPEX) isolated for each fiber type were: type I (8/17), type IIA (43/59), type IIAX (15/27), type IIX (94/94), type IIBX (72/75), and type IIB (144/120).

Fig. 3.

2-Deoxyglucose (2-DG) uptake measured in single fibers of each fiber type isolated from incubated paired rat muscles without or with insulin. Paired muscles were used within each treatment group [low-fat diet + sedentary (LFD-SED), high-fat diet SED (HFD-SED) or HFD + 3 h postexercise (HFD-3hPEX)] for incubation without or with 100 µU/ml insulin. Bars represent means of all fibers within a given insulin treatment and group. Error bars are means ± 95% confidence interval. P < 0.05 was considered statistically significant. A main effect of insulin was detected for all fiber types. There was no significant main effect of group detected for any fiber type. Insulin × Group interaction effects are displayed above each fiber type, and P values are displayed. Insulin × Group interaction effects between LFD-SED vs. HFD-SED, HFD-SED vs. HFD-3hPEX, and LFD-SED vs. HFD-3hPEX are indicated by symbols ^a, ^b, and ^c, respectively. Numbers of rats used for the groups in this experiment were: LFD-SED (n = 8), HFD-SED (n = 12), and HFD-3hPEX (n = 12). All fibers isolated from each muscle (~20–56 fibers/muscle) were used to determine 2-DG uptake and MHC expression. However, not every fiber type was included in the fibers sampled from every muscle. Therefore, the number of rats (LFD-SED no insulin/insulin, HFD-SED no insulin/insulin, HFD-3hPEX no insulin/insulin) from which fibers were isolated of each fiber type were: type I (5/5, 7/7, 7/7), type IIA (6/7, 12/12, 12/12), type IIAX (8/6, 10/9, 6/8), type IIX (8/8, 12/12, 12/12), type IIBX (8/8, 11/10, 12/10), and type IIB (8/8, 8/9, 12/8). Numbers of fibers (LFD-SED no insulin/insulin, HFD-SED no insulin/insulin, HFD-3hPEX no insulin/insulin) isolated for each fiber type were: type I (30/37, 16/12, 16/26), type IIA (114/103, 162/166, 142/165), type IIAX (38/33, 75/66, 38/35), type IIX (79/104, 156/183, 131/127), type IIBX (44/47, 59/53, 91/74), and type IIB (74/69, 48/62, 102/109).

Glycogen.

All fiber types had lower (P < 0.05) glycogen in the HFD-IPEX group vs. LFD-SED and HFD-SED groups (Figs. 4 and 5). The greater insulin-independent glucose uptake and lower glycogen IPEX provided strong evidence that the exercise protocol caused recruitment of all fiber types.

Fig. 4.

Glycogen content in single fibers of each fiber type measured by periodic acid-Schiff (PAS) staining of histologically sectioned rat muscle. Bars represent means of all fibers within a given treatment group [low-fat diet + sedentary (LFD-SED), high-fat diet + Sed (HFD-SED), or HFD + immedately postexercise (HFD-IPEX)]. Error bars are means ± 95% confidence interval. *P < 0.05, HFD-IPEX vs. LFD-SED and HFD-SED. Numbers of rats used for the groups in this experiment were: LFD-SED (n = 4), HFD-SED (n = 4), and HFD-IPEX (n = 4); 39–116 fibers/muscle were used to determine glycogen content. There was one HFD-SED muscle and one LFD-SED muscle for which no type IIAX fibers were identified. Additionally, there was one HFD-SED muscle for which no type IIBX fibers were identified. Numbers of fibers (LFD-SED/HFD-SED/HFD-IPEX) used for glycogen measurement of each fiber type were: type I (56/22/54), type IIA (66/74/60), type IIAX (39/19/33), type IIX (25/50/72), type IIBX (28/18/25), and type IIB (72/79/107).

Fig. 5.

Representative images of serially cross-sectioned rat muscle for identification of glycogen content and fiber type. Serially sectioned muscle from low-fat diet + sedentary (LFD-SED; A and D), high-fat diet + Sed (HFD-SED; B and E), and HFD + immedately postexercise (HFD-IPEX; C and F) are shown. Fiber type I myosin heavy chain (MHC) is shown in blue, type IIA MHC is shown in red, type IIB MHC is shown in green, and type IIX MHC is represented by absence of signal (black). Glycogen content was determined by stain intensity (by grayscale) within each fiber.

Lipid droplet analysis.

Representative images of BODIPY-stained cross sections used for LD density and size analyses are displayed in Fig. 6. The LD density, indicated by the percentage of fiber CSA covered by BODIPY-stained LDs, was lower (P < 0.05) in LFD-SED vs. either HFD-SED or HFD-IPEX in type IIA and IIAX fibers (Fig. 7). In type I fibers, the LD density was lower (P < 0.01) in LFD-SED vs. HFD-SED fibers. In type IIX fibers, the LD density was lower (P < 0.05) in LFD-SED vs. HFD-IPEX fibers. Subsarcolemmal LD density was greater (P < 0.01) in HFD-SED vs. other groups for type I fibers (Fig. 8). Additionally, subsarcolemmal LD density was greater (P < 0.05) in HFD-IPEX vs. other groups in type IIX fibers. LD size was greater (P < 0.05) in HFD-SED vs. other groups in type I fibers and was lower (P < 0.05) in LFD-SED vs. HFD-SED in type IIA fibers (Fig. 9). Subsarcolemmal LD size was not different among the groups for any fiber type (results not shown).

Fig. 6.

Representative images of a cross section of rat muscle used to identify fiber-specific neutral lipids. Images were captured for identification of neutral lipids (A) and extracellular matrix (ECM; B). C: merged lipid and ECM images are shown. By use of ImageJ, an automated threshold (D) was applied, and the border of each fiber was identified (E). F: a region 1 µm from the border of each fiber was used for quantification of subsarcolemmal lipids.

Fig. 7.

Quantification of lipid droplet (LD) density in single fibers of each fiber type measured by BODIPY staining of histologically sectioned rat muscle. Bars represent means of all fibers within a given treatment group [low-fat diet + sedentary (LFD-SED), high-fat diet + Sed (HFD-SED), and HFD + immedately postexercise (HFD-IPEX]. Error bars are means ± 95% confidence interval. *P < 0.05, different from both other groups; †P < 0.05, LFD-SED vs. HFD-SED; ‡P < 0.05, LFD-SED vs. HFD-IPEX. Numbers of rats used for the groups in this experiment were: LFD-SED (n = 4), HFD-SED (n = 4), and HFD-IPEX (n = 4); 24–77 fibers/muscle were used to determine LD density. There was one HFD-SED muscle for which no type IIAX fibers were identified. Numbers of fibers (LFD-SED/HFD-SED/HFD-IPEX) used for LD density of each fiber type were: type I (31/23/32), type IIA (61/43/54), type IIAX (12/9/17), type IIX (43/34/53), type IIBX (14/19/27), and type IIB (40/33/59).

Fig. 8.

Quantification of the density of lipid droplets (LD) located <1 µm from the sarcolemma in single fibers of each fiber type. Bars represent means of all fibers within a given treatment group [low-fat diet + sedentary (LFD-SED), high-fat diet + Sed (HFD-SED), and HFD + immedately postexercise (HFD-IPEX]. Error bars are means ± 95% confidence interval. *P < 0.05, different from both other groups. Subsarcolemmal LD density was quantified from the same images as total LD density; therefore, rat numbers and fiber numbers are identical to those displayed in the Fig. 7 legend.

Fig. 9.

Quantification of lipid droplet (LD) size in single fibers of each fiber type measured by BODIPY staining of histologically sectioned rat muscle. Bars represent the mean of all fibers within a given treatment group [low-fat diet + sedentary (LFD-SED), high-fat diet + Sed (HFD-SED), and HFD + immedately postexercise (HFD-IPEX]. Error bars are means ± 95% confidence interval. *P < 0.05, different from both other groups; †P < 0.05, LFD-SED vs. HFD-SED. LD size was quantified from the same images as LD density; therefore, rat numbers and fiber numbers are identical to those displayed in the Fig. 7 legend.

Muscle fiber CSA.

Fiber CSA was determined from the MHC-stained cross sections used for lipid analysis. There were no significant treatment group differences within any fiber type (results not shown).

GLUT4 protein.

GLUT4 abundance was not significantly different among the treatment groups (LFD-SED, HFD-SED, and HFD-3hPEX) for any fiber type (Fig. 10).

Fig. 10.

Glucose transporter protein GLUT4 abundance in pooled single fibers of each fiber type. Bars represent means of fiber pools within a given treatment group [low-fat diet + sedentary (LFD-SED), high-fat diet + Sed (HFD-SED), and HFD + 3 h postexercise (HFD-3hPEX]. Error bars are means ± 95% confidence interval; n = 6–8/group. There were no significant differences among treatment groups for any fiber type.

DISCUSSION

The results revealed a number of novel insights with regard to muscle fiber type-selective effects of acute exercise on glucose uptake in insulin-resistant muscle. The most important results were: 1) in SED rats, the HFD resulted in fiber type-selective insulin resistance (no HFD-related insulin resistance in type I fibers, a nonsignificant trend for slightly lower values for IIA fibers from HFD-fed rats and significantly lower insulin-stimulated glucose uptake for all other fiber types from the HFD group); 2) the HFD-related insulin resistance was not accompanied by greater values for any of the LD indexes (total LD size or density or subsarcolemmal LD size or density) or lower GLUT4 abundance; 3) the recruitment of each fiber type was indirectly evidenced by both greater insulin-independent glucose uptake IPEX and lower glycogen-measured IPEX; and 4) prior exercise resulted in fiber type-selective improvements in insulin-stimulated glucose uptake (no significant increase in type I fibers vs. either SED control group, greater values vs. HFD-SED controls for IIAX, IIX, and IIB fibers, and greater values vs. both HFD-SED and LFD-SED controls for type IIA and IIBX fibers) without an exercise-related increase in GLUT4 abundance in any of the fiber types.

The fiber type-selective insulin resistance corresponded closely with the results of our earlier study using the same dietary protocol only in sedentary rats (48). Pearson correlation analysis performed between the mean sedentary values for glucose uptake of each fiber type from the previous experiment vs. the corresponding sedentary control values in the current experiment revealed an R² = 0.926 (P < 0.001). The mechanisms underlying skeletal muscle fiber type-selective insulin resistance are uncertain, but the current study provided new insights in this regard. We did not detect a significant HFD-induced change in GLUT4 abundance in any fiber type. These results are consistent with the findings of our earlier study (48) for type I, IIA, IIAX, IIX, and IIBX fibers. Although GLUT4 abundance was also not significantly different between diet groups in IIB fibers in the current study, we previously that reported GLUT4 abundance was significantly lower for IIB fibers from HFD vs. LFD rats, using the same dietary protocol. In the earlier study, we evaluated GLUT4 abundance in individual fibers of each fiber type, whereas in the current study we determined GLUT4 in a pooled sample of all of the fibers collected for each fiber type in each diet group. In the earlier study, a total of 19 individual fibers from each diet group were analyzed for GLUT4; in the current study, GLUT4 abundance was analyzed using a total of 143 pooled fibers from LFD rats and 99 pooled fibers from HFD rats. The results from the larger number of fibers in the current study suggest that the HFD-induced insulin resistance in the current study was not attributable to significantly lower GLUT4 abundance in any of the fiber types. It is notable that earlier studies had reported that HFD-induced insulin resistance in whole muscles could occur without a decrement in total GLUT4 abundance (12, 27, 28, 69). The localization of GLUT4 at the cell surface membranes is crucial for insulin-stimulated glucose uptake (23), and in whole muscles, the insulin resistance was secondary to reduced insulin-stimulated cell surface GLUT4 localization (28, 69). Similarly, the current data suggest that the observed fiber type-specific insulin resistance is attributable to fiber type-selective effects of the HFD on GLUT4 translocation.

IMCLs have been reported to be positively associated with insulin resistance (35), but this relationship is not always observed (24, 42). Some studies have suggested that LDs localized near the sarcolemma (13, 46) and/or LD size (45) are more closely linked to insulin resistance than total IMCL content. Therefore, we assessed multiple indicators of IMCL content and subcellular localization in each fiber type. The current study is apparently the first to report fiber type-specific values determined in single muscle fibers for both insulin-stimulated glucose uptake and indicators of LD density, size, and subcellular localization. We anticipated that there might be greater LD size and/or localization near the sarcolemma in the fiber types that became insulin resistant during a HFD. However, there was no clear relationship between insulin-stimulated glucose uptake and any of the LD measurements. Although total LD density was greater in HFD-SED vs. LFD-SED rats for some of the fiber types that became insulin resistant (IIAX and IIX), none of the LD indexes differed between diet groups in other fiber types with substantial insulin resistance (IIBX and IIB). Furthermore, although neither type I nor type IIA fibers had significant HFD-induced insulin resistance, both fiber types had greater HFD-related values for several LD indexes. These results argue that these LD indexes were likely not responsible for the HFD-induced insulin resistance. Accumulation of specific lipid metabolites, such as ceramides and/or diacylglycerols, has been implicated in the processes leading to skeletal muscle insulin resistance (14, 29, 56). Furthermore, recent evidence has suggested that the subcellular location of diacylglycerols and ceramides, rather than their total concentrations, may modulate whole body insulin sensitivity (50). In this context, it would be valuable to develop the novel methods that will be necessary to assess HFD-induced subcellular changes in these lipid metabolites at a fiber type-specific level.

To interpret the fiber type-selective consequences of exercise on insulin-stimulated glucose uptake, it was important to evaluate the effects of the exercise protocol on the recruitment of each fiber type that was studied. Electromyography (EMG) can directly assess muscle recruitment, but determination of single fiber type-selective EMG during swim exercise was not feasible. Glycogen depletion is widely used as an indirect indicator of fiber recruitment during exercise (3, 31, 33). Accumulation of 2-[³H]DG is another useful indirect indicator of fiber recruitment at the whole muscle, motor unit, and single fiber level (10, 32, 43). The current study is apparently the first to use both glycogen depletion and 2-[³H]DG accumulation as indirect evidence for recruitment of individual fibers after in vivo exercise. There was a strong indication of fiber recruitment in every fiber type, based on the increased insulin-independent glucose uptake and decreased glycogen in the IPEX group compared with SED controls.

Although there was evidence that each fiber type was recruited by the exercise, there were striking fiber type-related differences in exercise effects on insulin-stimulated glucose uptake determined in the 3hPEX experiment. The results revealed several distinct patterns with regard to exercise effects on insulin-stimulated glucose uptake. In types IIB and IIX fibers, which were both characterized by HFD-induced insulin resistance, prior exercise increased insulin-stimulated glucose uptake to values that exceeded their HFD-SED controls, but exercise failed to increase insulin-stimulated glucose uptake by either fiber type above their respective LFD-SED controls. In types IIA and IIBX fibers, prior exercise increased insulin-stimulated glucose uptake to levels that not only exceeded their HFD-SED controls but also exceeded their LFD-SED controls. Insulin-stimulated glucose uptake of type IIAX fibers was also greater than HFD-SED controls, but they did not significantly exceed their LFD-SED controls. Prior exercise failed to induce significantly greater insulin-stimulated glucose uptake vs. HFD-SED controls only in type I fibers. The explanation for the lack of an increase in insulin-stimulated glucose uptake in HFD-SED compared with HFD-3hPEX type I fibers is uncertain, but it is apparently not attributable to a lack of type I fiber recruitment, as evidenced by clear exercise effects on both insulin-independent glucose uptake and glycogen-determined IPEX.

It is useful to put the fiber type-specific results for glucose uptake into context based on earlier results for the whole muscles from rats using the same diet and exercise protocol. In whole epitrochlearis muscles, we previously found, 1) the HFD caused a moderate level of insulin resistance that was similar to the results in most of the individual fiber types; 2) insulin-independent glucose uptake was increased approximately twofold for IPEX vs. HFD-SED controls, comparable to the approximately two- to threefold increase in each of the fiber types; 3) exercise enhanced the insulin-stimulated glucose uptake of the HFD-3hPEX group to values ~30% greater than HFD-SED controls, which roughly approximates the midpoint for the range of exercise effects in the different fiber types; and 4) insulin-stimulated glucose uptake of the HFD-3hPEX rats was increased to values that were similar to their LFD-SED controls for each fiber type (i.e., prior exercise eliminated HFD-induced insulin resistance in whole epitrochlearis) (12). There was good correspondence between the glucose uptake by whole epitrochlearis and glucose uptake by single fibers with regard to the effects of diet and exercise (both IPEX and 3hPEX). The heterogeneity in diet effects and exercise effects among the fiber types for insulin-stimulated glucose uptake that was revealed by single fiber analysis would have been impossible to discern on the basis of only the conventional analysis of whole muscles.

Although the current study was the first to assess the fiber type-selective effects of exercise on glucose uptake by single fibers from insulin-resistant muscle, we (10) recently reported the effects of the same exercise protocol on insulin-stimulated glucose uptake by single fibers from rats that were not insulin resistant eating a LFD. As in the current study, insulin-independent glucose uptake was increased in each fiber type evaluated IPEX compared with SED controls. However, at ~3 h postexercise, insulin-stimulated glucose uptake was significantly increased in type I fibers and in each of the type II fiber types evaluated except for type IIX fibers. Given that exercise increased insulin-stimulated glucose uptake by type I fibers in LFD rats (10), and that the HFD did not cause insulin resistance in type I fibers in either the current study or an earlier study (48), it was surprising that exercise did not enhance insulin-stimulated glucose uptake in type I fibers of HFD-fed rats in the current study. The results for type IIX fibers in the current study indicated that prior exercise increased insulin-stimulated glucose uptake above values for HFD-SED controls but that exercise failed to elevate values above those for the LFD-SED controls. The available results do not reveal the reason that LFD and HFD rats had different effects of exercise on insulin-stimulated glucose uptake in type I fibers, or why exercise failed to increase insulin-stimulated glucose uptake of type IIX fibers above LFD-SED controls in either diet group. A speculative scenario is that the time course for an exercise-induced increase in insulin-stimulated glucose uptake may not be identical for all fiber types. For example, perhaps the insulin sensitivity was transiently increased in type I fibers from HFD rats at an earlier time point but had reversed by 3hPEX. The time course for postexercise increases in insulin-stimulated glucose uptake by the whole epitrochlearis muscle has been reported at multiple times between 0.5 and 48 h postexercise, with peak values evident at ~3hPEX (11, 21, 62). However, the current study is the first to evaluate postexercise insulin sensitivity at the single-fiber level in insulin-resistant muscle, and only one earlier study assessed single fiber insulin sensitivity after exercise in normal muscle (10). In both studies, insulin-stimulated glucose uptake was determined only at ~3hPEX, so the time course for postexercise effects on insulin sensitivity in individual fiber types has not been characterized.

An important next step will be to elucidate the specific cellular mechanisms that underlie the different postexercise effects that we have discovered for insulin-stimulated glucose uptake in various fiber types. Our working hypothesis is based on our results in whole epitrochlearis muscles from LFD and HFD rats after the same exercise protocol. In several studies using LFD rats, we found that prior exercise resulted in greater phosphorylation of the Rab-GAP protein known as Akt substrate of 160 kDa (AS160, also called TBC1D4) in insulin-stimulated muscles on sites that are important for GLUT4 translocation and glucose uptake (2, 12, 21, 58). We also found that HFD resulted in attenuated AS160 phosphorylation and that acute exercise by HFD-fed rats increased AS160 phosphorylation in insulin-stimulated muscles (12). Results from multiple studies using rats, mice, or humans have reported that increased insulin-stimulated glucose uptake can occur without alterations in proximal insulin signaling steps, from insulin receptor binding to Akt activation (9, 12, 26, 49, 66, 67). We hypothesize that in insulin-stimulated muscles from both LFD- and HFD-fed rats, prior exercise will lead to greater phosphorylation of AS160 on key regulatory sites in a fiber type-specific manner by exercise, resulting in greater cell surface GLUT4 localization, leading in turn to the fiber type-specific effects of exercise on insulin-stimulated glucose uptake.

In conclusion, glucose uptake occurs at the cellular level, and skeletal muscle is composed of fiber types with diverse metabolic properties (52), including their capacity for glucose uptake (10, 38, 39). Accordingly, it is impossible to fully understand the effect of interventions on skeletal muscle glucose uptake without evaluating the possibility of differences at the cellular and fiber type-specific level. By doing so, we made the unexpected discovery that exercise robustly enhanced insulin-stimulated glucose uptake in each of the insulin-resistant fiber types (including the fiber types with the greatest insulin resistance), but it did not significantly elevate insulin-stimulated glucose uptake above that of HFD-SED controls in the type I fibers (which did not become insulin resistant with the HFD). These unique results offer an opportunity for future research aimed at delineating the cellular mechanisms that were responsible for the fiber type-selective effects of diet and exercise in glucose uptake.

GRANTS

These experiments were supported by a grant from the National Institute of Diabetes and Digestive and Kidney Diseases (R01 DK-71771).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.W.P. and G.D.C. conceived and designed research; M.W.P., C.S.Y., Y.N., E.B.A., and M.S. performed experiments; M.W.P., M.S., and R.J.P.-S. analyzed data; M.W.P. and G.D.C. interpreted results of experiments; M.W.P. and G.D.C. prepared figures; M.W.P. and G.D.C. drafted manuscript; M.W.P., C.S.Y., Y.N., E.B.A., M.S., C.L.M., R.J.P.-S., and G.D.C. edited and revised manuscript; M.W.P., C.S.Y., Y.N., E.B.A., M.S., C.L.M., R.J.P.-S., and G.D.C. approved final version of manuscript.