Exercise effects on γ3-AMPK activity, phosphorylation of Akt2 and AS160, and insulin-stimulated glucose uptake in insulin-resistant rat skeletal muscle

Abstract

One exercise session can increase subsequent insulin-stimulated glucose uptake (ISGU) by skeletal muscle. Prior research on healthy muscle suggests that enhanced postexercise ISGU depends on elevated γ3-AMPK activity leading to greater phosphorylation of Akt substrate of 160 kDa (pAS160) on an AMPK-phosphomotif (Ser⁷⁰⁴). Phosphorylation of AS160^Ser704, in turn, may favor greater insulin-stimulated pAS160 on an Akt-phosphomotif (Thr⁶⁴²) that regulates ISGU. Accordingly, we tested if exercise-induced increases in γ3-AMPK activity and pAS160 on key regulatory sites accompany improved ISGU at 3 h postexercise (3hPEX) in insulin-resistant muscle. Rats fed a high-fat diet (HFD; 2-wk) that induces insulin resistance either performed acute swim-exercise (2 h) or were sedentary (SED). SED rats fed a low-fat diet (LFD; 2 wk) served as healthy controls. Isolated epitrochlearis muscles from 3hPEX and SED rats were analyzed for ISGU, pAS160, pAkt2 (Akt-isoform that phosphorylates pAS160^Thr642), and γ1-AMPK and γ3-AMPK activity. ISGU was lower in HFD-SED muscles versus LFD-SED, but this decrement was eliminated in the HFD-3hPEX group. γ3-AMPK activity, but not γ1-AMPK activity, was elevated in HFD-3hPEX muscles versus both SED controls. Furthermore, insulin-stimulated pAS160^Thr642, pAS160^Ser704, and pAkt2^Ser474 in HFD-3hPEX muscles were elevated above HFD-SED and equal to values in LFD-SED muscles, but insulin-independent pAS160^Ser704 was unaltered at 3hPEX. These results demonstrated, for the first time in an insulin-resistant model, that the postexercise increase in ISGU was accompanied by sustained enhancement of γ3-AMPK activation and greater pAkt2^Ser474. Our working hypothesis is that these changes along with enhanced insulin-stimulated pAS160 increase ISGU of insulin-resistant muscles to values equaling insulin-sensitive sedentary controls.

NEW & NOTEWORTHY Earlier research focusing on signaling events linked to increased insulin sensitivity in muscle has rarely evaluated insulin resistant muscle after exercise. We assessed insulin resistant muscle after an exercise protocol that improved insulin-stimulated glucose uptake. Prior exercise also amplified several signaling steps expected to favor enhanced insulin-stimulated glucose uptake: increased γ3-AMP-activated protein kinase activity, greater insulin-stimulated Akt2 phosphorylation on Ser474, and elevated insulin-stimulated Akt substrate of 160 kDa phosphorylation on Ser588, Thr642, and Ser704.

INTRODUCTION

Insulin resistance is a primary and essential defect in the process leading to type 2 diabetes (31). Skeletal muscle accounts for up to 85% of insulin-mediated glucose disposal (18). Because skeletal muscle is so quantitatively important for glucose disposal, it is a prime target for interventions aimed at combating insulin resistance. It is well established that exercise or muscle contractions can increase subsequent insulin-stimulated glucose uptake of skeletal muscle from either insulin-sensitive or insulin-resistant subjects (11, 13, 14, 50, 55–57, 78). Understanding how exercise regulates insulin-stimulated glucose uptake in insulin-sensitive subjects is important, but there is a more urgent need to identify the mechanisms responsible for enhanced insulin-stimulated glucose uptake after exercise during insulin resistance.

The postexercise enhancement in insulin-stimulated glucose uptake by skeletal muscle has been observed without increasing the level of insulin-stimulated signaling compared with sedentary controls at a number of insulin signaling steps in insulin-sensitive rodents and humans. A number of proximal insulin signaling steps ranging from insulin’s binding to its receptor to Akt phosphorylation have been reported to be similar in insulin-stimulated muscle from sedentary versus exercised, insulin-sensitive rodents or humans (10, 14, 19, 25, 26, 50, 67, 68, 83, 84). These results suggest that exercise might improve insulin sensitivity via regulation of a more distal insulin signaling step.

Akt substrate of 160 kDa (AS160; also known as TBC1D4) is the most distal signaling protein that has been clearly linked to insulin-stimulated glucose uptake in skeletal muscle (12). Multiple lines of evidence have suggested that altered AS160 phosphorylation is a strong candidate for mediating the improved insulin-stimulated glucose uptake by skeletal muscle following acute exercise (3, 22, 23, 71). Preventing phosphorylation of two Akt-phosphomotifs of AS160 (Thr642 to Ala642 and Ser588 to Ala588) markedly reduced insulin-stimulated GLUT4 translocation in adipocytes (59). Enhanced insulin-stimulated phosphorylation of AS160 on Thr642 and Ser588 in exercised compared with nonexercised skeletal muscle, concomitant with increased glucose uptake, has been reported in rat skeletal muscle with normal insulin sensitivity (4, 22, 23, 30, 61, 79). Acute exercise can enhance subsequent AS160 phosphorylation on Thr642 and/or Ser588 in insulin-stimulated skeletal muscles from insulin-resistant rats or humans independent of greater Akt activation (14, 50). Compelling new evidence was recently published supporting the idea that AS160 is essential for increased insulin-stimulated glucose uptake in skeletal muscle after acute contractile activity. Kjøbsted et al. (33) reported that insulin-stimulated glucose uptake by muscle several hours after in situ contraction was increased in wild-type mice, but not in AS160 muscle-specific knockout mice.

The mechanisms accounting for a sustained, postexercise increase in pAS160^Thr642 and/or pAS160^Ser588 of insulin-stimulated skeletal muscle are unknown. However, the phosphorylation of AS160 on Ser704, an AMPK-phosphomotif, has recently emerged as a potentially important step for enhanced insulin-stimulated glucose uptake after exercise. In human (36, 50) and rat (80) skeletal muscle with normal insulin sensitivity, the insulin-independent phosphorylation of AS160^Ser704 was reported to be increased immediately postexercise and remained elevated for hours after exercise, when insulin-stimulated glucose uptake was enhanced (54, 77). Additionally, multiple studies have reported enhanced insulin-stimulated pAS160^Ser704 in skeletal muscle after exercise or contraction (35, 50, 64, 76, 80). Preventing the phosphorylation of AS160^Ser704 in mouse muscle by mutating Ser704 to Ala704 attenuates the insulin-stimulated phosphorylation of AS160^Thr642, an Akt phosphomotif which regulates insulin-stimulated glucose uptake (37). These data suggest that the enhanced phosphorylation of AS160^Ser704 after exercise may prime Thr642 to be more easily phosphorylated, a potentially critical event for enhancing insulin sensitivity. It will be important to identify exercise effects on pAS160^Ser704 during insulin resistance to better understand the mechanisms which govern enhanced postexercise insulin sensitivity in insulin-resistant subjects. Previous research has not assessed prior exercise effects on Ser704 phosphorylation in insulin-resistant rat skeletal muscle. Therefore, our first major aim was to investigate the effect of acute exercise on key AS160 phosphorylation sites (Ser588, Thr642, and Ser704) in muscles from insulin-resistant rats fed a high-fat diet (HFD) after exercise, versus sedentary HFD- and low-fat diet-fed (LFD) controls.

AS160^Ser704 is an AMPK phosphosite, and AMPK is a heterotrimeric complex composed of a catalytic α subunit (α1 or α2 isoform) and two regulatory subunits (β1 or β2 isoform; and γ1, γ2, or γ3 isoform). γ3-AMPK has been shown to be activated in muscle by exercise in humans (36, 72) and rats (80) or by electrically stimulated contractions in mice (35, 72). Prior incubation of isolated muscles from wild-type mice with 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR) resulted in enhanced insulin-stimulated glucose uptake, but this AICAR effect on insulin sensitivity was absent in muscles from γ3-AMPK knockout mice (37). Moreover, electrically stimulated ex vivo contractile activity which enhances pAS160^Ser704 in muscles from wild-type (WT) mice did not increase pAS160^Ser704 in γ3-AMPK knockout mice (35). These findings suggest γ3-AMPK activation and pAS160^Ser704 may be critical steps for enhancing insulin-stimulated glucose uptake after AICAR or contraction in insulin-sensitive animals. However, neither prior AICAR treatment nor in situ contractions completely recapitulate in vivo exercise. Furthermore, no previously published study have evaluated either γ1- or γ3-AMPK activity in insulin-resistant rat skeletal muscle after acute exercise. Therefore, the second major aim of this experiment was to investigate γ3-AMPK activity in muscles from insulin-resistant rats fed a high-fat diet (HFD) after exercise, compared with sedentary HFD- and low-fat diet-fed (LFD) controls.

Earlier research has reported that acute exercise can lead to subsequently elevated insulin-stimulated AS160 phosphorylation on key regulatory sites and glucose uptake in skeletal muscle independent of enhanced Akt phosphorylation (3, 14, 30, 50, 61). There are three isoforms of Akt (Akt1, Akt2 and Akt3) that are expressed in mammals (75), but only Akt1 and Akt2 are highly expressed in skeletal muscle (45, 60, 87). Akt2, the predominant Akt isoform expressed in human skeletal muscle (77), regulates insulin-stimulated AS160 phosphorylation (38) and insulin-stimulated glucose uptake (16, 42). Previous research using methods that cannot selectively distinguish between the phosphorylation of Akt2 and other Akt isoforms has indicated that a single exercise session does not increase subsequent insulin-stimulated pan-Akt phosphorylation in human skeletal muscle (36, 65). Because Akt2 has been reported to account for almost all of the Akt phosphorylation in insulin-stimulated muscles from sedentary humans (77), these previous results provide indirect evidence suggesting that prior exercise may not substantially increase Akt2 phosphorylation in human muscle. However, earlier studies have not used methods that exclusively assess Akt2 phosphorylation to directly test if prior exercise alters Akt2 phosphorylation in either healthy or insulin-resistant skeletal muscle of humans or rodents. Accordingly, the third major aim of the current study was to use appropriate methods to directly determine in insulin-resistant skeletal muscle the effects of acute exercise on the phosphorylation of Akt2 on its key regulatory sites, Ser474 and Thr309.

METHODS

Materials.

The reagents and apparatus for SDS-PAGE and nonfat dry milk (#170–6404) were from Bio-Rad (Hercules, CA). [³H]-2-deoxyglucose (NET328001MC), [¹⁴C]-mannitol (NEC314250UC), and [γ-³²P]-ATP were from PerkinElmer (Waltham, MA). Tissue Protein Extraction Reagent, (T-PER; #PI78510), Bicinchoninic Acid Protein Assay Kit (#PI23223), MemCode Reversible Protein Stain Kit (#PI24585), and Protein G magnetic beads (#10004D), DynaMag^TM-2 magnet (#12321D) were from ThermoFisher (Pittsburgh, PA). Anti-rabbit IgG horseradish peroxidase conjugate (#7074), anti-phospho Akt Ser⁴⁷³ (pAkt^Ser473; #4060), anti-phospho Akt Thr³⁰⁸ (pAkt^Thr308; #13038), anti-Akt (#4691), anti-phospho Akt2 Ser⁴⁷⁴ (pAkt2^Ser474; #8599), anti-Akt2 (#3063; used for immunoblotting), anti-Akt3 (#3788), anti-phospho AS160 Thr⁶⁴² (pAS160^Thr642; #4288), anti-phospho AS160 Ser⁵⁸⁸ (pAS160^Ser588; #8730), anti-phospho AMPKα Thr¹⁷² (pAMPKα^Thr172; #2535), anti-AMP-activated protein kinase-α (AMPKα; #2532), anti-AMP-activated protein kinase-β1 (AMPK-β1; #12063), anti-AMP-activated protein kinase-β2 (AMPK-β2; #4148), anti-acetyl CoA carboxylase (ACC; #3676), and anti-phospho ACC Ser⁷⁹ (pACC^Ser79; #3661) were from Cell Signaling Technology (Danvers, MA). Anti-Akt1 (#AF1775) and anti-Akt2 (#AF23151; used for immunoprecipitation) were from R&D Systems (Minneapolis, MN). Skeletal muscle expresses two isoforms of ACC (ACC1 and ACC2). ACC1 has a relatively low expression in skeletal muscle and is phosphorylated by AMPK on Ser⁷⁹. ACC2 has a relatively high expression in skeletal muscle and is phosphorylated by AMPK on Ser²¹² (1). Since the pACC^Ser79 antibody (#3661) detects both pACC1^Ser79 and pACC2^Ser212 (46), we hereafter refer to the results with this antibody as pACC^Ser79/212. Anti-phospho AS160^Ser704 was provided by Dr. Jonas Treebak (University of Copenhagen, Denmark). Anti-AMP-activated protein kinase γ1 (γ1-AMPK; #32508) was from Abcam. Anti-AMP-activated protein kinase γ3 (γ3-AMPK) was provided by Dr. David Thomson (Brigham Young University) (27). Liquid scintillation cocktail (#111195-CS) was from Research Products International (Mount Prospect, IL). Anti-Akt substrate of 160kDa (AS160; #ABS54), anti-AMP-activated protein kinase α1 (AMPK-α1; #07–350), anti-AMP-activated protein kinase α2 (AMPK-α2; #07–363), P81 phosphocellulose squares (#20–134), and enhanced chemiluminescence Luminata Forte Western HRP Substrate (#WBLUF0100) were from EMD Millipore (Billerica, MA).

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 provided with standard rodent chow (Laboratory Diet no. 5L0D; LabDiet, St. Louis, MO) or high-fat chow (Laboratory Diet no. D12492; ResearchDiets, New Brunswick, NJ) and water ad libitum for 2 wk until they were fasted the night before the experiment at ∼1700. Caloric intake for each rat during the 2-wk diet period was estimated based on the difference between the food provided on day 1 and the food remaining at ~1700 on the night before the experiment. Beginning at 0700 on the day of the experiment rats either remained sedentary or swam in a barrel filled with water (35°C) for four 30-min bouts with a 5-min rest between bouts. Immediately postexercise (IPEX), some rats (IPEX and Sedentary) were anesthetized (intraperitoneal sodium pentobarbital, 50 mg/kg weight), weighed, and their epitrochlearis muscles were dissected. At 3 h postexercise (3hPEX) other rats (3hPEX and Sedentary) were anesthetized, weighed, and their epitrochlearis muscle were dissected. After muscle dissections, the epididymal fat pads were dissected and weighed.

We used the same diet protocols for the LFD and HFD groups that were used in an earlier study which reported that the HFD resulted in insulin resistance compared with LFD controls (14). In the current study, sedentary LFD rats were included to confirm that the HFD protocol induced insulin resistance in this cohort of rats. In the earlier study, we used the same exercise protocol as the current study. The previous study evaluated both LFD and HFD rats. The results of the earlier study demonstrated that insulin-stimulated glucose uptake was increased at 3hPEX in each diet group compared with diet-matched controls. Furthermore, insulin-stimulated glucose uptake for the LFD-3hPEX group exceeded the values for the HFD-3hPEX group. The focus of the current study was to probe potential mechanisms for the exercise-induced increase in insulin-stimulated glucose uptake only in the HFD-fed rats. The goal was not to assess the potential mechanisms for the greater insulin-stimulated glucose uptake in the LFD-3hPEX versus HFD-3hPEX rats. Accordingly, in the current study, only HFD-fed rats were studied under exercised conditions (IPEX and 3hPEX).

Ex vivo incubations of muscles for glucose uptake.

Muscles were incubated, as previously described (47), in glass vials gassed (95% O₂, 5% CO₂) in a temperature controlled bath by a two-step incubation process (35°C during both steps). For the IPEX experiments, muscles were placed in vials for 10 min containing 2 ml of incubation step 1 media (Krebs Henseleit Buffer, KHB, supplemented with 0.1% bovine serum albumin, BSA, 2 mM sodium pyruvate and 6 mM mannitol). For incubation step 2 (15 min) these muscles were transferred to vials containing 2 ml of incubation step 2 media (KHB supplemented with 0.1% BSA, 0.1 mM 2-DG (2.25 mCi/mmol [³H]-2-DG), 2 mM sodium pyruvate and 6 mM mannitol (2 mCi/mmol [¹⁴C] mannitol)). For the 3hPEX experiment, paired muscles were placed in vials containing 2 ml of incubation step 1 media for 30 min with or without 100 µU/ml insulin. These muscles were then transferred to vials containing 2 ml of incubation step 2 media for 20 min with or without 100 µU/ml insulin. After step 2, whole muscles were blotted, freeze clamped, and stored at −80°C until further processing.

Whole muscle glucose uptake.

Frozen muscles used for 2-DG uptake were weighed and homogenized in ice-cold lysis buffer (T-PER supplemented with 1 mM Na₃VO₄, 1 mM EDTA, 1 mM EGTA, 2.5 mM sodium pyrophosphate tetrabasic decahydrate, 1 mM β-glycerophosphate, 1 µg/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride). Homogenates were then rotated at 4°C for 1 h before centrifugation at 15,000 g for 15 min at 4°C. Aliquots (200 µl) of supernatant were added to vials along with 8 ml of scintillation cocktail. 2-[³H]-DG and 2-[¹⁴C]-mannitol disintegrations per minute were measured by scintillation counter, and then 2-DG uptake was calculated as previously described (9).

Immunoblotting.

Total protein concentrations for whole muscle lysates were determined by bicinchoninic acid assay, and equal amounts of protein for each sample were separated via SDS-PAGE, and transferred to polyvinyl difluoride membranes. After electrotransfer, the MemCode protein stain was used to confirm equal loading (2). Membranes were then blocked with 5% BSA or nonfat milk in TBST (Tris-buffered saline, pH 7.5 plus 0.1% Tween-20) for 1 h, incubated with appropriate concentrations of primary (1:1000; overnight) and secondary (1:20,000; 1 h) antibodies, and subjected to enhanced chemiluminescence to quantify protein bands by densitometry (FluorChem E Imager, AlphaView software; ProteinSimple, San Jose, CA). Individual values were normalized to the mean value of all samples on the membrane. The vendor validated that the pAkt2^Ser474 antibody used in this experiment (Cell Signaling Technology #8599) is selective for this Ak2 phosphosite without cross reactivity to either Akt1 or Akt3. Therefore, it was unnecessary to isolate Akt2 by immunoprecipitation (IP) before immunoblotting with the selective pAkt2^Ser474 antibody. In contrast, no commercially available phospho-antibody is currently available for the selective detection of pAkt2^Thr309. Therefore, we evaluated pAkt2^Thr309 using methods very similar to those that we previously described in detail (62). In brief, muscle lysates were first immunoprecipitated using an antibody (R&D Systems #AF23151) that the vendor has documented to selectively recognize recombinant Akt2, but not recombinant Akt1 or Akt3. After Akt2-IP, we immunoblotted using an antibody that has been validated by the vendor to recognize pThr309 on Akt2 (Cell Signaling Technology #13038).

We performed additional analyses using rat epitrochlearis samples to confirm that our Akt2-IP protocol effectively isolated Akt2 (results shown in Supplemental Fig. S1, available at https://doi.org/10.6084/m9.figshare.11402694). Immunoblotting was performed for both the immunoprecipitated fraction and the post-IP supernatant. The results demonstrated that the Akt2-IP protocol efficiently pelleted essentially all of the Akt2. Immunoblotting both the Akt2-IP pellet and the post-IP supernatant fractions revealed that, in contrast to Akt2, almost all of the Akt1 was in the post-IP supernatant. Akt1 was barely detectable in the pellet. Earlier research reported that both human and mouse skeletal muscle has very low Akt3 expression (40, 74). However, previous studies had not evaluated Akt3 protein abundance in rat epitrochlearis muscle. Accordingly, we analyzed rat epitrochlearis muscle. Akt3 protein was undetectable in rat epitrochlearis samples. Taken together, the results of these analyses demonstrated the efficacy of the Akt2-IP protocol for isolating Akt2 found in rat epitrochlearis muscle.

AMPK-γ1 and -γ3 isoform activity.

The specificity of γ1-AMPK and γ3-AMPK antibodies used for immunoprecipitation has been previously confirmed (80). AMPK isoform-specific activity was determined as previously described (37). Briefly, 300µg (for γ3-AMPK assay) or 600 µg (for γ1-AMPK assay) of protein from each sample was rotated at 4°C overnight with appropriate AMPK γ isoform antibody (1:1,000). Then 50 μl of protein G-magnetic beads was added to the muscle lysate/antibody mixture and rotated for 2 h at 4°C. A DynaMag-2 magnet was used to precipitate the protein G-immunocomplex. Each pellet was washed one time in buffer A [50 mmol/L NaCl, 1% Triton X-100, 50 mmol/L sodium fluoride, 5 mmol/L sodium-pyrophosphate, 20 mmol/L Tris-base (pH 7.5), 500 μmol/L PMSF, 2 mmol/L dithiothreitol (DTT), 4 μg/mL leupeptin, 4 μg/mL aprotinin, and 250 mmol/L sucrose], once in 6X assay buffer (240 mmol/L HEPES, 480 mmol/L NaCl, pH 7.0), and two times in 3X assay buffer. The activity assay was then performed in 30 μl of kinase reaction buffer [40 mmol/L HEPES, pH 7.5, 80 mmol/L NaCl, 800 μmol/L DTT, 200 μmol/L AMP, 100 μmol/L AMARA peptide, 5 mmol/L magnesium chloride (MgCl₂), 200 μmol/L ATP, and 10 µCi of [γ- ³²P]-ATP] for 30 min at 30°C. The reaction was stopped by the addition of 10 μl of 1% phosphoric acid and then 40 µl of supernatant was transferred to phosphocellulose paper. After 4 × 15 min washes with 1% phosphoric acid, the phosphocellulose paper was dried for 5 min and placed in the vials containing scintillation cocktail for scintillation counting. It was not feasible to assay all of the samples in a single batch. Therefore, every batch included an equal number of samples from each treatment group. The results are expressed relative to the normalized mean value for all of the samples from each batch.

Statistics.

Data are expressed as means ± standard error (SEM), with two-tailed significance levels of α < 0.05. Two-tailed t-tests were performed to compare LFD versus HFD means for body mass, fat pad mass, epitrochlearis mass, and estimated caloric intake. One-way ANOVAs were used to determine the treatment group effect of IPEX experiments. Two-way ANOVAs were used to compare means among more than two groups from 3hPEX experiments. The treatment design was not a complete factorial of diet and exercise; therefore main effect comparisons of those treatments could not be made. Instead, comparisons and conclusions were restricted to the planned comparisons of the three treatment groups (LFD-SED, HFD-SED, and either HFD-IPEX or HFD-3hPEX). The analysis evaluated the effect of exercise on rats with high-fat diets, the effect of diet in sedentary rats, and compared HFD-fed exercised rats to the LFD-SED rats. Bonferroni post hoc tests were performed to identify the source of significant differences.

RESULTS

Body mass, muscle mass, epididymal fat mass, and estimated caloric intake.

Following the 2-wk diet intervention, the HFD rats versus LFD rats had a significantly greater body mass (299 ± 4 versus 290 ± 6 g; P < 0.05), epididymal fat mass (2318 ± 128 versus 1297 ± 81 mg; P < 0.001), and estimated caloric intake (97 ± 2 versus 82 ± 2 kcal/day; P < 0.001). Epitrochlearis muscle mass did not significantly differ between LFD and HFD (59 ± 5 versus 61 ± 5 mg, respectively).

Glucose uptake.

Insulin-independent glucose uptake was significantly greater in the HFD-IPEX group compared with both sedentary control groups (P < 0.001; Fig. 1A). In the 3hPEX experiment, muscles incubated with insulin had a greater glucose uptake than paired muscles incubated without insulin in all treatment groups (P < 0.001; Fig. 1B). Within the insulin-treated muscles, glucose uptake was lower in the HFD-SED group compared with either the LFD-SED or HFD-3hPEX groups (P < 0.01), but there was not a significant difference between insulin-treated LFD-SED and HFD-3hPEX muscles. There were significant main effects of diet and exercise treatment (P < 0.01) and insulin (P < 0.001), and a significant treatment × insulin interaction effect (P < 0.05) for glucose uptake.

Fig. 1.

A: epitrochlearis 2-DG uptake immediately postexercise. Data were analyzed by one-way ANOVA. *P < 0.001, HFD-IPEX vs. both LFD-SED and HFD-SED. Values are mean ± SEM; n = 14/group. B: 2-DG uptake from paired epitrochlearis muscles incubated ± 100 µU/mL insulin at 3 h postexercise. Data were analyzed by two-way ANOVA. *P < 0.001, insulin vs no insulin. †P < 0.005, within insulin-treated muscles HFD-SED vs. both LFD-SED and HFD-3hPEX. Values are means ± SEM; n = 15/group. 2-DG, 2-deoxy-d-glucose; HFD, high-fat diet; LFD, low-fat diet; SED, sedentary; IPEX, immediately postexercise; 3hPEX, 3 h postexercise.

γ1-AMPK and γ3-AMPK activity.

There was not a significant effect of exercise on γ1-AMPK activity in HFD-IPEX or HFD-3hPEX groups compared with either sedentary control group (Fig. 2, A and C). Increased γ3-AMPK activity was observed in both HFD-IPEX and HFD-3hPEX rats compared with both HFD-SED and LFD-SED controls (P < 0.05; Fig. 2, B and D). AMPK activity assays were performed in separate batches for the IPEX and 3hPEX groups along with their respective sedentary controls. Activity assays for γ1-AMPK and γ3-AMPK were also performed separately from each other. Therefore, it would be inappropriate to statistically compare the AMPK activity values determined either for IPEX versus 3hPEX rats or for γ1-AMPK versus γ3-AMPK.

Fig. 2.

A: γ1-AMPK activity immediately postexercise. B: γ3-AMPK activity immediately postexercise. C: γ1-AMPK activity 3 h postexercise. D: γ3-AMPK activity 3 h postexercise. For γ1-AMPK activity assays n = 5/group. For γ3-AMPK activity assays n = 8/group. Data were analyzed by one-way ANOVA. *P < 0.01, significantly greater than both other treatment groups. Values are expressed as means ± SEM. HFD, high-fat diet; LFD, low-fat diet; SED, sedentary; IPEX, immediately postexercise; 3hPEX, 3 h postexercise.

AMPK isoform abundance.

There was no significant effect of treatment group on the abundance of α1-AMPK, α2-AMPK, β1-AMPK, β2-AMPK, γ1-AMPK, or γ3-AMPK (results not shown). Because γ2-containing AMPK heterotrimers have been reported to be undetectable in skeletal muscle (69), γ2-AMPK was not measured.

AMPKα and ACC phosphorylation.

For the IPEX experiment, pAMPKα^Thr172 expressed relative to total AMPKα (pAMPKα^Thr172/ AMPKα) was significantly greater in the HFD-IPEX group compared with both SED controls (P < 0.001; Fig. 3A). At 3hPEX there was a significant main effect of exercise for increased pAMPKα^Thr172/ AMPKα (P < 0.001; Fig. 3C). pACC^Ser79/212 was significantly increased in the HFD-IPEX group compared with SED controls (P < 0.005; Fig. 3B), but there was no effect of insulin or treatment group on pACC^Ser79/212 at 3hPEX (Fig. 3D).

Fig. 3.

A: phosphorylated AMPKα^Thr172/AMPKα immediately postexercise. B: phosphorylated ACC^Ser79/212/ACC immediately postexercise. C: phosphorylated AMPKα^Thr172/AMPKα 3 h postexercise. D: phosphorylated ACC^Ser79/212/ACC 3 h postexercise. IPEX data were analyzed by one-way ANOVA and 3hPEX data were analyzed by two-way ANOVA. *P < 0.005, HFD-IPEX greater than both other groups. †P < 0.01, main effect of 3hPEX vs. both LFD-SED and HFD-SED. Values are expressed as means ± SEM; n = 8–17/group. HFD, high-fat diet; LFD, low-fat diet; SED, sedentary; IPEX, immediately postexercise; 3hPEX, 3 h postexercise.

AS160 phosphorylation.

For the IPEX experiment, pAS160^Ser704 and pAS160^Thr642 expressed relative to total AS160 were significantly greater in the HFD-IPEX group compared with both SED controls (P < 0.005; Fig. 4, A and B). pAS160^Ser588 expressed relative to total AS160 (pAS160^Ser588/AS160) was significantly greater in the HFD-IPEX group compared with the HFD-SED control (P < 0.005; Fig. 4C), but not compared with the LFD-SED control. For the 3hPEX experiment, insulin-stimulated pAS160^Ser704 and pAS160^Thr642 were lower in the HFD-SED group compared with either LFD-SED and HFD-3hPEX (P < 0.01; Fig. 4, D and E). Insulin-stimulated pAS160^Ser588 was significantly greater in the HFD-3hPEX group compared with HFD-SED control (P < 0.01; Fig. 4F), but not compared with LFD-SED control. There were significant main effects of diet and exercise treatment (P < 0.01) and insulin (P < 0.001) for pAS160^Ser588, pAS160^Thr642, and pAS160^Ser704. There was a significant treatment × insulin interaction effect (P < 0.05) for pAS160^Thr642.

Fig. 4.

A–C: phosphorylated AS160^Ser704, AS160^Thr642, and AS160^Ser588/AS160 immediately postexercise. Data were analyzed by one-way ANOVA. *P < 0.005, HFD-IPEX vs. both LFD-SED and HFD-SED. †P < 0.005, HFD-IPEX vs. HFD-SED. Values are expressed as means ± SEM; n = 17/group. D–F: phosphorylated AS160^Ser704, AS160^Thr642, and AS160^Ser588/AS160 at 3 h postexercise. Data were analyzed by two-way ANOVA. *P < 0.001, insulin vs. no insulin. †P < 0.001, HFD-SED vs. both LFD-SED and HFD-3hPEX within insulin-treated muscles. ‡P < 0.01, HFD-3hPEX vs. HFD-SED within insulin-treated muscles. Values are expressed as means ± SEM; n = 10–17/group. HFD, high-fat diet; LFD, low-fat diet; SED, sedentary; IPEX, immediately postexercise; 3hPEX, 3 h postexercise.

Akt phosphorylation.

For the IPEX experiment, neither pAkt^Thr308 nor pAkt^Ser473 relative to total Akt (pAkt^Thr308/Akt and pAkt^Ser473/Akt) significantly differed among the LFD-SED, HFD-SED and HFD-IPEX rats (results not shown). Insulin-stimulated pAkt^Thr308 and pAkt^Ser473 relative to total Akt (pAkt^Thr308/Akt and pAkt^Ser473/Akt) were significantly lower in the HFD-SED group compared with the LFD-SED group and the HFD-3hPEX group (Fig. 5, A and C). Insulin-stimulated pAkt2^Thr309 relative to total Akt2 (pAkt2^Thr309/Akt2) was lower in HFD-SED versus LFD-SED (Fig. 5B). Insulin-stimulated pAkt2^Ser474 relative to total Akt2 (pAkt2^Ser474/Akt2) was greater in the HFD-3hPEX group compared with both the LFD-SED and HFD-SED control groups (Fig. 5D). There was a significant main effect of insulin (P < 0.001) for pAkt^Thr308, pAkt2^Thr309, pAkt^Ser473, and pAkt2^Ser474. For rats fed a HFD there was a significant main effect of exercise (P < 0.005) for pAkt^Thr308, pAkt^Ser473, pAkt2^Ser474. For SED rats there was a significant main effect of diet (P < 0.005) for pAkt^Thr308, pAkt^Ser473, pAkt2^Ser474. There was a significant treatment × insulin interaction effect (P < 0.005) for pAkt^Thr308, pAkt^Ser473, and pAkt2^Ser474. The discovery of significant effects on pAkt2 in the current study spurred the question: Can acute exercise enhance pAkt2 in insulin-stimulated muscles from rats with normal insulin sensitivity? To answer this question, we evaluated pAkt2 in epitrochlearis muscles from LFD-SED and LFD-3hPEX rats. The muscles analyzed were from rats described in our recently published study (80) that used the same exercise protocol as the current study with LFD-fed male Wistar rats. This earlier study used the same epitrochlearis incubation protocol (±100 µU/ml insulin) as the current study. There were no differences in either basal or insulin-stimulated values between sedentary and 3hPEX muscles for pAkt2^Ser474/Akt2 or pAkt2^Thr309/Akt2 of LFD-fed rats (results not shown).

Fig. 5.

A–D: phosphorylated Akt^Thr308/Akt, Akt2^Thr309/Akt2, Akt^Ser473/Akt, and Akt2^Ser474/Akt2 at 3 h postexercise. Data were analyzed by two-way ANOVA. *P < 0.001, insulin vs. no insulin. †P < 0.01, significantly different from both other treatment groups within insulin-treated muscles. ‡P < 0.05, significantly different from LFD-SED within insulin-treated muscles. Values are expressed as means ± SEM; n = 10/group. HFD, high-fat diet; LFD, low-fat diet; SED, sedentary; 3hPEX, 3 h postexercise.

DISCUSSION

Because insulin resistance is a major health concern that affects millions of people, it is important to better understand the biological processes that underlie the exercise-induced improvement in insulin sensitivity of insulin-resistant skeletal muscle, the main tissue for insulin-mediated glucose disposal. The current study used a 2-wk HFD protocol that has been shown to induce skeletal muscle insulin resistance in rats (14, 48, 49). Having confirmed muscle insulin resistance based on reduced insulin-stimulated glucose uptake in the epitrochlearis in the current study (Fig. 1), we assessed key postexercise signaling events that have been proposed to be critical for enhancing insulin-stimulated glucose uptake in insulin-sensitive muscle. The main new findings of the study were: 1) γ3-AMPK activity, but not γ1-AMPK activity, in skeletal muscle was elevated both IPEX and 3hPEX compared with sedentary controls, 2) at 3hPEX insulin-stimulated pAS160^Ser704 was increased, but insulin-independent pAS160^Ser704 was unaltered compared with HFD-SED controls, 3) insulin-stimulated pAkt2^Ser474 was elevated at 3hPEX compared with HFD-SED control, and 4) insulin-stimulated pAkt2^Thr309 was decreased in sedentary muscle following a 2 wk HFD versus LFD-SED control. In addition, insulin-independent glucose uptake, pAS160^Thr642, and pAS160^Ser704 were elevated above sedentary (LFD and HFD) values IPEX. Furthermore, the abundance of α1, α2, β1, β2, γ1, and γ3 AMPK-subunits were unaltered by either HFD or exercise. Although a causal relationship remains to be established, these results support the idea that γ3-AMPK activity and AS160 phosphorylation are potentially part of the process leading to enhanced insulin-stimulated glucose uptake in insulin-resistant skeletal muscle following exercise.

We recently proposed a model to help elucidate the processes underlying increased insulin-stimulated glucose uptake by skeletal muscle postexercise (11). In this model, triggers are initiating events that activate downstream memory elements, which store the information that can be passed on to the mediators, which convert the memory into action that ultimately results in greater insulin-stimulated glucose uptake.

Skeletal muscle AMPK activation is a hallmark response to exercise (15, 21, 82, 85) and a potential trigger for increased insulin sensitivity. Exercise results in the reversible phosphorylation of AMPKα^Thr172 which markedly increases AMPK’s enzymatic activity (34). Increased pAMPKα^Thr172 and phosphorylation of AMPK’s substrate, acetyl CoA carboxylase (ACC) are widely used as indicators of AMPK activity (7, 28, 37, 39, 66). As expected, we observed increased pAMPKα^Thr172 and pACC^Ser79 IPEX. However, AMPK is a heterotrimeric protein complex that contains one catalytic (α) and two regulatory (β and γ) subunits, and values for pAMPK^Thr172 and pACC^Ser79 do not provide information regarding the isoform-specific activity of AMPK. The activation of various heterotrimeric combinations of AMPK in response to exercise can differ depending on the intensity and duration of the exercise bout (7, 70). In humans and rats, the effect of in vivo exercise has consistently been shown to increase γ3-containing AMPK activity (7, 36, 39, 70, 80). However, exercise-induced enhancement of γ1-AMPK activity in skeletal muscle has been less consistent. The current results, which are the first assessment of isoform-selective effects of exercise on AMPK activity in insulin-resistant rat skeletal muscle, show elevated γ3-AMPK activity, but unaltered γ1-AMPK activity, both IPEX and 3hPEX.

Earlier research provides several lines of evidence implicating γ3-AMPK activation as being potentially important for increased insulin sensitivity. Results using γ3-AMPK knockout mice revealed that γ3-AMPK is essential for the effect of prior exposure of skeletal muscle to AICAR, an AMPK-activator, on insulin-stimulated glucose uptake (37). Furthermore, γ3-AMPK activity was elevated 3 h after in situ muscle contraction along with greater insulin-stimulated glucose uptake in mouse skeletal muscle (35). Similar to the results for insulin-resistant rats in the current study, we recently reported in muscles from rats with normal insulin sensitivity that γ3-AMPK activity was increased (both IPEX and 3hPEX), and insulin-stimulated glucose uptake was enhanced 3hPEX (80).

What are possible cellular events that may serve as the memory elements that connect exercise-induced γ3-AMPK activation and insulin-stimulated glucose uptake? Increased postexercise pAS160^Ser704 has emerged as a candidate memory element that is linked to AMPK because Ser704 is an AMPK phosphomotif. Furthermore, mutating AS160 Ser704 to Ala704, which prevents its phosphorylation, results in decreased insulin-stimulated pAS160^Thr642 (37). Although insulin-independent pAS160^Ser704 remained elevated for several hours after exercise in some studies (36, 50, 80), in the current study, pAS160^Ser704 was no longer significantly increased above sedentary values at 3hPEX. Therefore, the current results do not support the idea that a sustained elevation in insulin-independent pAS160^Ser704 is a likely memory element for greater insulin-stimulated glucose uptake in insulin-resistant rat skeletal muscle. It has been convincingly documented that Ser704 can be phosphorylated by AMPK (70). Phosphorylation on this site can also be increased by an insulin-dependent process (35, 50, 80) that is not mediated by AMPK (70). The insulin-regulated kinase that phosphorylates Ser704 remains to be identified. However, it has been demonstrated that insulin-stimulated Ser704 phosphorylation is inhibited by wortmannin, and is independent of either Akt or mTOR (73). It will be important for future research to identify this currently unknown, insulin-regulated kinase and to determine if it is modulated by exercise.

Given that greater activity of γ3-AMPK 3hPEX in insulin-resistant muscle was not accompanied by a sustained increase in insulin-independent pAS160^Ser704, what are other possible memory elements that might rely on elevated γ3-AMPK activity? Mass spectrometry analysis identified several other sites that became phosphorylated in skeletal muscles from mice that had been injected with the AMPK-activator AICAR (73). Validated antibodies do not appear to be available for these sites, and it is unknown if they are responsive to in vivo exercise or if they have any effect on insulin-stimulated glucose uptake. However, it seems possible that AMPK-regulated phosphomotifs on AS160 other than Ser704 might influence insulin-stimulated glucose uptake in insulin-resistant muscle after exercise. Because various proteins can bind to AS160 (e.g., 14–3-3, RUVBL2, RIP140, and ClipR-59) (17, 24, 53, 86), an alternative possibility is that γ3-AMPK might phosphorylate these or other proteins, and thus indirectly regulate AS160s function. Alternatively, γ3-AMPK may lead to inhibition of protein phosphatase-1α or other phosphatases that regulate AS160 dephosphorylation (4, 63). Finally, γ3-AMPK activity remained elevated 3hPEX in insulin-resistant muscle, and we previously also observed a similar, long-lasting increase in γ3-AMPK activity in muscles from normal rats (80). Others have also reported that γ3-AMPK activity in skeletal muscle can remain elevated several hours after exercise by humans (29, 36, 65), in situ contraction (35), or ex vivo incubation with AICAR (37). Thus, it is possible that a sustained increase in muscle γ3-AMPK activity serves as a memory element that can lead to enhanced insulin sensitivity, and the mechanism may not necessarily involve AS160-dependent processes.

Insulin-stimulated phosphorylation of Akt is often reported to be unchanged following acute exercise (14, 20, 25, 30, 50, 80, 83). However, some studies have reported elevated insulin-stimulated Akt phosphorylation after exercise (3, 23, 61). Similarly, insulin-stimulated pAkt^Ser473 and pAkt^Thr308 were greater in insulin-stimulated muscle postexercise compared with SED controls. It is important to note that Akt2 is the isoform of Akt that is primarily responsible for insulin stimulation of both pAS160^Thr642 and glucose uptake (6, 42, 58). No earlier research has evaluated acute exercise effects on the insulin-stimulated phosphorylation of Akt2 in skeletal muscle with a physiological insulin dose. Accordingly, we also measured insulin-stimulated pAkt2 and found that exercise increased insulin-stimulated pAkt2^Ser474 and pAkt2^Thr309 in HFD-fed rats. Increased pAkt2^Ser474 is required for insulin’s full effect on Akt2 activity, pAS160^Thr642, GLUT4 translocation and glucose uptake (32). In this context, our novel results suggest that the exercise-induced enhancement in insulin-stimulated pAS160^Thr642 and glucose uptake may be related, at least in part, to greater phosphorylation on one or both of Akt2’s key regulatory sites. It would be valuable for future research to determine if the elevated insulin-stimulated phosphorylation on Akt2’s key regulatory sites after exercise occurs concomitant with greater Akt2 activity.

Improved insulin sensitivity after acute exercise has been reported in multiple species, including humans, rats, dogs, and sheep (41, 43, 44, 51). Humans and rats are the species that have been most frequently studied with regard to postexercise insulin sensitivity. The results for these species are similar in several important respects. The improvement in insulin sensitivity has been reported at ~1 to 4 h postexercise in both rats and humans (8, 52). Furthermore, the effect has also been observed to be sustained up to 48 h postexercise in both species (8, 52). In addition, earlier research has demonstrated that increased insulin-sensitivity can occur after acute exercise that does not result in increased total GLUT4 protein abundance in both rats and humans (14, 83). There is also evidence in both species that exercise with insulin sensitizing effects can be accompanied by greater AS160 phosphorylation on key regulatory sites (14, 50). Although many similarities have been reported between rats and humans with regard to acute exercise effects on insulin sensitivity, it remains quite possible that the specific mechanisms and features of this outcome may not be identical for all species. Nonetheless, research using various preclinical models, including rats, has substantial value for gaining insights about the potential mechanisms that underlie exercise benefits in humans.

Several important limitations of the current study should be noted. One limitation is that a LFD exercised group was not included in this study. However, it is informative to consider together some of the key outcomes in LFD and HFD rats from this and earlier studies using the same exercise, diet, and muscle incubation protocols. In the current study, there was a 42% increase in insulin-stimulated glucose uptake of epitrochlearis muscles from HFD rats in the 3hPEX group versus HFD sedentary controls. This increase is similar to the 32% increase in HFD rats after the same exercise protocol in our earlier study (14). In both studies, exercise eliminated the HFD-induced insulin resistance for glucose uptake compared with LFD sedentary rats. In three earlier studies, we reported an ~65–70% increase in insulin-stimulated glucose uptake of epitrochlearis muscles following an identical exercise protocol in LFD-3hPEX rats versus LFD sedentary controls (4, 14, 80). When LFD-3hPEX and HFD-3hPEX rats were directly compared in a previous study, insulin-stimulated glucose uptake was significantly greater for LFD compared with HFD rats (14). Thus, the same exercise protocol elicited greater insulin-stimulated glucose uptake in LFD versus HFD rats at both relative and absolute levels. An earlier study from our laboratory evaluated the influence of the same exercise protocol on isoform-selective AMPK activity and pAS160^Ser704 in LFD rats (80). The pattern for exercise effects was similar for isoform-selective AMPK activity in each diet group (increased γ3-AMPK activity IPEX and 3hPEX; unaltered γ1-AMPK at both time points). In both diet groups, pAS160^Ser704 was greater in muscles sampled from rats IPEX versus sedentary controls. However, the response to exercise (3hPEX versus sedentary) diverged between LFD rats in the earlier study (80) and HFD rats in the current study with regard to pAS160^Ser704 in muscles incubated without insulin. The reason for the magnitude of the exercise-induced improvement in insulin-stimulated glucose uptake was greater for LFD versus HFD rats is uncertain. However, it is notable that the two diet groups did not differ with regard to the magnitude of the: increase in insulin-independent glucose uptake, increase in AMPK phosphorylation, or decrease in glycogen in muscle (14). Accordingly, other differences between the diet groups must be responsible for the lower insulin-stimulated glucose uptake after exercise by HFD rats. The muscle pAS160^Ser704 was increased at 3hPEX for the LFD-fed rats (80), but not for HFD-fed rats in the current study. It is uncertain if this difference played any role in the lesser exercise effect on insulin-stimulated glucose uptake in HFD rats. However, it is notable that for both diets, in muscles incubated with insulin, pAS160^Ser704 was increased for 3hPEX versus sedentary rats eating the same diet. Earlier research has demonstrated that in healthy rats consuming a LFD, prior exercise elevates the insulin-stimulated increase in GLUT4 translocation (26). The relative magnitude of this increase was similar to the magnitude of the increase in insulin-stimulated glucose uptake. Previous research has not reported the effect of acute exercise on subsequent insulin-stimulated GLUT4 translocation in muscle from insulin-resistant rats. Our working hypothesis is that prior exercise would elevate insulin-stimulated GLUT4 translocation to a lesser extent in insulin-resistant muscle versus insulin-sensitive muscle. Another limitation of the current study is that only relatively short-term HFD was evaluated. It would be informative to perform a study under conditions that induce more profound obesity and/or insulin resistance (e.g., longer durations of dietary interventions or genetically modified animals).

In conclusion, the current results provided novel evidence that acute exercise can result in elevated γ3-AMPK activity along with greater insulin-stimulated Akt2 phosphorylation and AS160 phosphorylation on Ser704 in insulin-resistant rat skeletal muscle. Appropriate genetic models will be invaluable to test the possibility that causal relationships may exist between each of these outcomes and improved insulin-stimulated glucose uptake in insulin-resistant muscle postexercise. A recent study reported that insulin-stimulated muscle glucose uptake several hours after in situ contraction was increased in wild-type mice, but not in AS160 muscle-specific knockout mice (33). It is essential to appreciate that, although electrically stimulated contraction can be a useful model for exercise, results with this model cannot be assumed to perfectly replicate in vivo exercise. It is also relevant to recognize that there is a much more extensive literature related to the processes responsible for in vivo exercise effects on insulin-stimulated glucose uptake by skeletal muscle in rats compared with mice (11). In this context, it is important to note the recent creation of an AS160-null rat model (5). This new genetic model will enable future research aimed at testing the idea that AS160 plays a key role in the exercise-induced improvement of insulin-stimulated glucose uptake in both normal and insulin-resistant rat skeletal muscle.

GRANTS

This research was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant 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., E.B.A., H.W., and X.Z. performed experiments; M.W.P. and G.D.C. 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., E.B.A., H.W., X.Z., and G.D.C. edited and revised manuscript; M.W.P., E.B.A., H.W., X.Z., and G.D.C. approved final version of manuscript.