Resistance Exercise and Insulin Regulate AS160 and Interaction With 14-3-3 in Human Skeletal Muscle

Eight recreationally active males (age 23 ± 2 years; weight 76 ± 4 kg; height 183 ± 4 cm; BMI 22.8 ± 0.9 kg/m² [means ± SE]) volunteered as subjects. All subjects were healthy nonsmokers with no evidence of high blood pressure, cardiovascular disease, and insulin resistance or diabetes. None of the subjects had performed resistance exercise regularly (more than once per week) in the past 6 months. Experimental procedures and risks of the study were explained verbally and in writing. All subjects gave their informed written consent, and the experiment was approved by Deakin University Human Research Ethics Committee.

Subjects attended the laboratory on separate occasions for exercise familiarization and pre-experimental testing. An incremental workload test to exhaustion was performed on an electromagnetically braked cycle ergometer (LODE Instrument, Groningen, The Netherlands) (16) to determine peak pulmonary oxygen uptake (3.79 ± 0.22 l · min⁻¹ Vo_2peak). Subjects completed leg exercises (leg extension, inclined leg press, and leg curls), with incrementally greater weight application until one-repetition maximum was determined (17). The resistance exercise program was performed in a specialized research gymnasium under the supervision of a strength and conditioning–accredited individual.

Hyperinsulinemic-euglycemic clamps were performed 1) after a resting period (rest); 2) immediately after a single bout of resistance exercise (acute); and 3) 24 h after 1 week of repeated bouts of resistance exercise (trained). The acute trial was performed at least 1 week after the rest trial and consisted of a single resistance exercise bout (three sets of leg extension, inclined leg press, and leg curls at 80% one-repetition maximum with 10–12 repetitions). Subjects then completed three additional resistance exercise sessions as described above on alternate days for the next 7 days (trained).

Subjects presented to the laboratory in the morning after an overnight fast (10–12 h). For the day before (24 h) each trial, subjects consumed a standardized diet (∼14,500 kJ, ∼80% carbohydrate) and avoided consumption of alcohol and caffeine. Throughout the experiment, subjects refrained from any physical activity other than that associated with the experiment.

Hyperinsulinemic-euglycemic clamps were performed as previously described (16). Muscle samples were obtained from the vastus lateralis, by the percutaneous needle biopsy technique (18). In rest and trained trials, muscle was sampled before (without insulin) and after the clamp (with insulin). In the acute trial, muscle was sampled before and immediately after the single exercise bout (before clamp, without insulin) and after the clamp (with insulin). Biopsies were frozen in liquid nitrogen for later analysis. Blood was sampled at rest, after exercise, and during the clamp for analysis of blood glucose and lactate (EML105 Radiometer; Copenhagen, Denmark), plasma insulin (Human Insulin Specific RIA kit; Linco Research, St. Charles, MO), free fatty acids (FFAs) (Wako NEFA C test kit; Wako Chemicals, Osaka, Japan), and tumor necrosis factor-α (TNF-α) (Quantikine HS; R&D System, Minneapolis, MN).

Muscle was homogenized (Kinematica Polytron, Brinkmann, CT) in ice-cold buffer (20 mmol/l Tris-HCl, 5 mmol/l EDTA, 10 mmol/l Na₄P₂O₇, 100 mmol/l NaF, 2 mmol/l Na₃VO₄, 1% Nonident P-40, and protease inhibitor cocktail [Roche, Lewes, Sussex, U.K.]) and incubated for 30 min at 4°C. Homogenates were spun at 13,000 × g for 30 min, and the supernatant was frozen and stored at −80°C for later analysis. Total protein concentrations were determined by the Bradford method using BSA as standard.

Phospho-specific Akt (Ser⁴⁷³ and Thr³⁰⁸), phospho-(Ser/Thr) Akt substrate antibody, AMPK α2 (Thr¹⁷²), acetyl Co–carboxylase (ACC) (Ser⁷⁹) were from Cell Signaling Technology (Beverly, MA). Total insulin receptor substrate-1 (IRS-1) and AS160 (Rab-GAP) were from Upstate (Lake Placid, NY). Total 14-3-3 and anti-phosphotyrosine (pY99) were from Santa Cruz Biotechnology (Santa Cruz, CA). Total AS160 for immunoprecipitation was raised in sheep against the peptide CHPTNDKAKAGNKP (Cys + mouse residue 12951-307) and total AMPK-α2 were donated by Professor Graham Hardie (University of Dundee). ExtraAvidin peroxidase conjugate for total ACC was from Sigma. Secondary antibodies coupled to horseradish peroxidase were from Pierce (Rockford, IL). Anti-digoxigenin (DIG) coupled to horseradish peroxidase was from Roche.

Muscle lysate (40–60 μg) was heated in SDS sample buffer, separated by gel electrophoresis using precast SDS-polyacrylamide Bis-Tris gels (Invitrogen, Paisley, U.K.), and transferred to nitrocellulose membranes. Membranes were incubated for 1 h in blocking buffer (50 mmol/l Tris-HCl, pH 7.5, 0.15 mol/l NaCl, and 0.5% Tween [TBST]) containing 10% milk or 5% BSA. Membranes were incubated overnight at 4°C with antibodies (1 μg/ml or 1:1,000 dilution) in TBST with 5% milk or BSA. Protein was detected using secondary antibody in TBST with 5% milk for 60 min. Antibody binding was viewed by enhanced chemiluminescence. Film was scanned by ChemiGenius Bio-imaging system (SynGene, Cambridge, U.K.), and bands were quantified using Gene Tools analysis software (SynGene).

Muscle lysate (300–500 μg) was incubated at 4°C for 1 h with 5 μl protein G-Sepharose (Amersham Biosciences, Little Chalfont, United Kingdom) coupled to 3–4 μg AS160 or IRS-1 antibody. Immunoprecipitates were washed with buffer A (50 mmol/l Tris-HCl, pH 7.5, 1 mmol/l EGTA, 1 mmol/l EDTA, 1% Triton-X 100, 1 mmol/l Na₃VO₄, 50 mmol/l NaF, 5 mmol/l Na₄P₂O₇, 0.27 mmol/l sucrose, and 0.5 mol/l NaCl), and buffer B (50 mmol/l Tris-HCl, pH 7.5, and 0.1 mmol/l EGTA). Samples were heated in SDS sample buffer, separated by gel electrophoresis, and immunoblotted as described above.

14-3-3 Sepharose was prepared as previously described (19,20). 14-3-3 pulldown was performed by incubating 10 μl 14-3-3 Sepharose with muscle lysate (300 μg) at 4°C for 1 h. Samples were washed with buffers A and B, and after gel electrophoresis, membranes were incubated overnight with AS160 primary antibody, followed by secondary antibody as described above. For DIG–14-3-3 overlays, AS160 was immunoprecipitated from muscle lysate as described above. After separation of proteins by gel electrophoresis, DIG-labeled 14-3-3 (19,20) was used as primary antibody, and anti-DIG horseradish peroxidase was used as secondary antibody.

Muscle lysate (50 μg) was immunoprecipitated at 4°C for 1 h with protein G-Sepharose coupled to AMPK α2 antibody. Immunoprecipitates were washed with buffers A and B, and the assay was performed as previously described (21).

Standards were included in all assays and immunoblotting, and variation was accounted for by normalizing to control samples. All data are expressed as means ± SE. Statistical analysis was undertaken using a Student's t test and two-way ANOVA for repeated measures, with significance set at P < 0.05.

In response to a single bout of exercise (acute), there was no significant change in blood glucose (basal, 4.6 ± 0.2 mmol · l⁻¹; exercise, 4.7 ± 0.2 mmol · l⁻¹), plasma insulin (basal, 90 ± 30 pmol · l⁻¹; exercise, 134 ± 62 pmol · l⁻¹), and plasma TNF-α levels (basal, 2.28 ± 0.44 pg · ml⁻¹; exercise, 1.59 ± 0.21 pg · ml⁻¹), although blood lactate increased (basal, 1.3 ± 0.1 mmol · l⁻¹; exercise, 4.9 ± 0.5 mmol · l⁻¹, P < 0.05), and plasma FFAs decreased (basal, 0.38 ± 0.05 mmol · l⁻¹; exercise, 0.24 ± 0.03 mmol · l⁻¹, P < 0.05).

Insulin infusion during the clamp increased (P < 0.05, main effect) plasma insulin from basal (rest, 84 ± 18 pmol · l⁻¹; acute, 134 ± 62 pmol · l⁻¹; and trained, 54 ± 8 pmol · l⁻¹) to high physiological levels, with the average insulin concentration during the final 30 min of the clamp similar in all trials (rest, 808 ± 68 pmol · l⁻¹; acute, 766 ± 57 pmol · l⁻¹; and trained, 808 ± 42 pmol · l⁻¹). In each trial, exogenous glucose was variably infused such that blood glucose levels remained at ∼5 mmol · l⁻¹ for the duration of the clamp (rest, 4.9 ± 0.0 mmol · l⁻¹; acute, 5.0 ± 0.1 mmol · l⁻¹; and trained, 4.9 ± 0.1 mmol · l⁻¹, average final 30 min of clamp). Whole-body insulin action or glucose disposal measured by the average glucose infusion rate during the final 30 min of the clamp was significantly decreased after a single bout of resistance exercise (acute) compared with the rest trial (Fig. 1), which indicates reduced insulin sensitivity. In response to repeated bouts of resistance exercise over a 7-day period (trained), the glucose infusion rate was similar to the rest trial (Fig. 1). Blood lactate levels were elevated (P < 0.05) during the first 30 min of the clamp in the acute trial (Table 1). Before the clamp, plasma FFAs were higher (P < 0.05) in the trained trial, but infusion of insulin significantly reduced FFAs to a similar level in all trials (Table 1). TNF-α levels were not different before or during the clamp (Table 1).

A single resistance exercise bout significantly decreased AS160 phosphorylation at phospho-Akt substrate motifs by ∼40% immediately after exercise (Fig. 2 and 3), whereas AS160 phosphorylation measured 24 h after repeated bouts of resistance exercise was similar to basal levels in the rest trial. Insulin significantly increased phosphorylation of AS160 by two- to threefold (P < 0.05, main effect), and this response was similar after rest, a single resistance exercise bout, and repeated bouts of resistance exercise. Total AS160 protein levels were not different in response to exercise and/or insulin (data not shown; representative blot in Fig. 4).

AS160 is a downstream target of Akt, and after a single resistance exercise bout, there was a tendency, although not statistically significant, for Akt Ser⁴⁷³ (Figs. 2 and 5) and Akt Thr³⁰⁸ (data not shown; representative blot in Fig. 2) phosphorylation to decrease. Repeated bouts of exercise did not alter basal Akt phosphorylation. In all trials, insulin resulted in a similar increase (P < 0.05, main effect) in Akt Ser⁴⁷³ and Thr³⁰⁸ phosphorylation of ∼four- to fivefold. For both Akt Ser⁴⁷³ (r = 0.71; P < 0.05) and Akt Thr³⁰⁸ (r = 0.70; P < 0.05), the changes in phosphorylation after exercise and/or insulin correlated significantly with AS160 phosphorylation. Total Akt protein levels were similar in response to exercise and/or insulin (data not shown; representative blot in Fig. 4). IRS-1 tyrosine phosphorylation decreased (P < 0.05) in the basal state after a single resistance exercise bout. After repeated bouts of resistance exercise, IRS-1 tyrosine phosphorylation was similar to basal levels in the rest trial (Figs. 2 and 5). Insulin infusion resulted in a similar increase (P < 0.05, main effect) in IRS-1 tyrosine phosphorylation in all trials. Total IRS-1 protein levels were not different in response to exercise and/or insulin (data not shown; representative blot in Fig. 4).

The interaction between AS160 and 14-3-3 was examined with two different approaches involving affinity purification of AS160 with 14-3-3 Sepharose (pulldown) and far Western overlay with DIG–14-3-3 (Fig. 2). Similar results were found for each technique. The 14-3-3 pulldown method was used for quantification (Fig. 3). Similar to results for AS160 phosphorylation, the interaction between AS160 and 14-3-3 was significantly decreased after a single bout of exercise. Compared with basal levels in the rest trial, repeated bouts of resistance exercise did not affect the interaction between AS160 and 14-3-3. In response to insulin, there was a significant increase (P < 0.05, main effect) in the interaction of AS160 with 14-3-3, although this was not influenced by a single bout or repeated bouts of resistance exercise. Total 14-3-3 protein levels were similar in response to resistance exercise and/or insulin (data not shown; representative blot in Fig. 4).

A single resistance exercise bout did not statistically increase AMPK Thr¹⁷² phosphorylation (basal, 3.34 ± 0.67 arbitrary units; exercise, 3.34 ± 0.67 arbitrary units) or AMPK α2 activity significantly (Fig. 6), although phosphorylation of ACC Ser⁷⁹ was elevated (P < 0.05) more than threefold (Figs. 2 and 6). Insulin had no effect on AMPK Thr¹⁷² phosphorylation in the rest trial (basal, 2.99 ± 0.70 arbitrary units; insulin, 2.64 ± 0.68 arbitrary units) or after a single exercise bout (basal, 3.34 ± 0.67 arbitrary units; insulin, 2.85 ± 0.70 arbitrary units) or repeated exercise bouts (basal, 3.40 ± 0.77 arbitrary units; insulin, 3.00 ± 0.89 arbitrary units). There was no effect of insulin on AMPK α2 activity, but insulin decreased (P = 0.052, main effect) ACC Ser⁷⁹ phosphorylation similarly in all trials (Fig. 6). Total AMPK-α2 and ACC protein levels were similar in response to resistance exercise and/or insulin (data not shown; representative blot in Fig. 4).

In the present study, whole-body insulin action was decreased in the period immediately after a single bout of resistance exercise in humans, which is in contrast to studies where insulin action is enhanced during a similar period after aerobic exercise (1,3). The inhibitory effect of a single bout of resistance exercise on insulin action is likely to be temporary because insulin sensitivity or glucose tolerance is enhanced 12–24 h after a single bout of resistance exercise in some (22–24) but not all studies (25). Extrapolation between this and previous studies (22–24) is difficult because of differences in the type of subjects and methods used to measure insulin sensitivity. In the present study, insulin action was also measured 24 h after the completion of repeated bouts of resistance exercise over a 7-day period. Unlike aerobic exercise, in which repeated exercise bouts over 7 days further enhance insulin action (26,27), repeated bouts of resistance exercise did not translate into improved insulin action compared with the resting control trial. It is not clear from this study whether repeated bouts of exercise ameliorate the inhibitory effect of a single bout of resistance exercise. From these findings, it is likely that the beneficial effects of resistance exercise on insulin action may only occur after a much longer period of training, as demonstrated by previous studies (4–6).

To explain how resistance exercise may mediate changes in whole-body insulin action, we examined skeletal muscle insulin signaling, in particular the regulation of AS160, and interaction with the novel regulatory protein 14-3-3 (12), because this appears to play a key role in insulin and contraction regulated skeletal muscle glucose uptake (14). A single bout of resistance exercise decreased AS160 phosphorylation by ∼40%, whereas AS160 phosphorylation was not altered by repeated bouts of resistance exercise. Prior exercise training may influence how AS160 is regulated as a previous study in trained cyclists failed to detect changes in AS160 phosphorylation after a single bout of resistance exercise (28). Interestingly, in the same study (28), AS160 phosphorylation increased in response to a single bout of aerobic exercise, suggesting that AS160 regulation may also be influenced by mode, and perhaps intensity, of exercise.

In response to insulin stimulation, and as previously measured in human (29,30) and rodent skeletal muscle (14,15), AS160 phosphorylation increased two- to threefold above basal levels. However, single or repeated bouts of resistance exercise did not alter insulin-stimulated AS160 phosphorylation. It is possible that direct inhibitory effects of a single resistance exercise bout on AS160 phosphorylation may only be short lasting. Alternatively, it could be due to the complex regulation or phosphorylation pattern of AS160, which is currently not well understood. AS160 appears to integrate different signals from the insulin-mediated/Akt pathway, muscle contraction/AMPK, and potentially other unknown kinases (14). Furthermore, the phospho-Akt substrate antibody detects a number of putative phosphorylation sites on AS160, which make it difficult to detect changes occurring on single, multiple, or other unidentified phosphorylation sites.

In response to exercise and/or insulin, AS160 phosphorylation correlated significantly with IRS-1 and Akt phosphorylation. However, the mechanism(s) that may account for changes in upstream insulin signaling proteins after resistance exercise are unknown. Elevated levels of TNF-α as a result of either muscle damage after eccentric exercise (31) or exogenous infusion (32) have been associated with insulin resistance, impaired IRS-1 signaling (33,34), and decreased AS160 phosphorylation (32). In the present study, there was no significant change in the plasma levels of TNF-α, although it is possible that localized changes in muscle TNF-α could have influenced insulin signaling. Despite these findings, changes in skeletal muscle insulin signaling do not fully explain the inhibitory effect of a single bout of resistance exercise on whole-body insulin action. It is possible that resistance exercise may decrease GLUT4 protein content (34,35), or alternatively, immune responses (other than TNF-α) related to resistance exercise or muscle damage may impair insulin action in skeletal muscle or other insulin-sensitive tissues (34). It also cannot be completely ruled out that changes in hormones and/or metabolites during resistance exercise may influence whole-body insulin action.

AS160 phosphorylation is increased in rodent skeletal muscle by contraction (13) and AICAR (5-aminoimidazole-4-carboxamide riboside, a pharmacological activator of AMPK), and from AMPK transgenic animal studies, this effect is regulated, at least in part, through AMPK (14,15). A single bout of resistance exercise can increase AMPK Thr¹⁷² phosphorylation (36) and AMPK α2 activity (37), but using similar resistance exercise to that used in the latter studies (36,37), AS160 phosphorylation was decreased in the present study. Although a significant increase in AMPK phosphorylation and α2-activity with exercise was not detected in the present study, ACC Ser⁷⁹ phosphorylation, an in vivo measure of AMPK activity, increased more than threefold. Because ACC Ser⁷⁹ phosphorylation is dominantly regulated by AMPK in muscle (38), this suggests that the effect of resistance exercise on AMPK activation was transient and not detected within the time course of muscle biopsy sampling. Taken together, these findings suggest that after a single bout of resistance exercise AS160 phosphorylation may be regulated primarily by factors other than AMPK. Whether AMPK plays a direct role in regulating the increase in AS160 phosphorylation after aerobic exercise in human skeletal muscle (28) remains to be determined.

Recent evidence in adipocytes suggests that the function of AS160 in insulin-mediated GLUT4 translocation is closely linked to its interaction with a novel binding partner 14-3-3 (12). Given the significance of this interaction to further understanding the regulation of glucose transport, we examined this in human skeletal muscle in response to insulin stimulation and exercise. The results demonstrate an increase in the interaction between AS160 and 14-3-3 with insulin, which is the first evidence, to our knowledge, in support of a physiological role in human skeletal muscle. The interaction between AS160 and 14-3-3 is also influenced by a single bout of resistance exercise and is likely mediated indirectly through changes in upstream insulin signaling components that regulate AS160 phosphorylation. However, it cannot be completely ruled out that exercise may have direct effects, given that exercise or muscle contraction can influence GLUT4 and glucose uptake via insulin-independent processes. Future studies will be required to establish whether the interaction between AS160 and 14-3-3 is influenced by exercise mode and intensity. Furthermore, given that insulin-stimulated phosphorylation of AS160 is impaired in skeletal muscle of type 2 diabetes (29), additional studies examining the interaction of AS160 with 14-3-3 in response to insulin and/or exercise in type 2 diabetes may give greater insight into the pathogenesis of this disease.

In conclusion, in the immediate period after a single bout of resistance exercise, whole-body insulin action is decreased. Repeated bouts of resistance exercise over a 7-day period do not enhance insulin action, suggesting that potential beneficial effects of resistance exercise on insulin action may only occur after a longer period of training. Regulation of AS160 and its interaction with the novel binding protein, 14-3-3, are influenced by both resistance exercise and insulin in human skeletal muscle. However, the changes in insulin signaling do not fully explain the effect of resistance exercise on whole-body insulin action.

This study was supported by grants from the Diabetes Australia Research Trust (DART) and Australian Research Council DP-0450338 (to K.H. and M.H.).

We are also grateful to the Medical Research Council and the pharmaceutical companies supporting the Division of Signal Transduction Therapy Unit (AstraZeneca, Boehringer-Ingelheim, GlaxoSmithKline, Merck, Merck KgaA, and Pfizer) for financial support.

We acknowledge Benjamin Gleeson, Sam Wright, and Chris Wilson (Deakin University) for excellent technical assistance; Prof. Carol MacKintosh, Dr. Jane Murphy, and Dr. Shuai Chen (University of Dundee) for reagents and discussion about AS160 and 14-3-3; Prof. Graham Hardie (University of Dundee) for reagents; and Greg Stewart (University of Dundee) for AMPK antibody purification.