Impairments in Site-Specific AS160 Phosphorylation and Effects of Exercise Training

Study 1 used a cross-sectional design encompassing younger to older individuals to determine if insulin action is associated with a decrement in insulin signaling at the level of AS160. Study 2 determined if a specific exercise training modality (endurance or strength training) could effectively ameliorate the insulin resistance evident in sedentary, insulin-resistant aged individuals by enhancing the most distal component of insulin signal transduction. Figure 1 provides an overview of the experimental design.

Specifically recruited for this study were 73 participants (41 women, 32 men) comprising a wide age range (18–84 years). Physical characteristics of the subjects are provided in Table 1. All participants were nonsmokers and participated in less than 1 h/week of organized physical activity, as assessed by a standardized questionnaire. In an attempt to study a representative population, inclusion required that a participant’s BMI be between the 25th and 75th percentile for his or her decade of age (32). Excluded were individuals with heart disease, diabetes, endocrine, and/or metabolic disorders, and those taking lipid-altering medication. Premenopausal women were tested during the early follicular phase of the menstrual cycle (days 1–6). Written informed consent was obtained, and the protocol was in accordance of the Declaration of Helsinki and approved by the East Carolina University Policy and Review Committee on Human Research.

Cardiorespiratory fitness was measured with an incremental, maximal treadmill test (33), with expired gases analyzed continuously (TrueMax 2400; ParvoMedics, Sandy, UT) to determine Vo₂peak. Body composition was measured by dual X-ray absorptiometry, and circumference measurements of the waist, hip, and thigh were obtained with a spring-loaded measuring tape.

Subjects reported to the laboratory at 0700 after a 12-h overnight fast. A 2-h euglycemic-hyperinsulinemic clamp was used to determine insulin action and elicit activation of insulin signaling, as previously described (18,34). Briefly, a primed insulin (Humulin; Eli Lilly, Indianapolis, IN) infusion was performed for 10 min (starting at 313 mU · m⁻² · min⁻¹), followed by a continuous infusion of insulin at a submaximal dosage of 100 mU · m⁻² · min⁻¹. Blood samples were obtained every 5 min, centrifuged, and autoanalyzed for serum glucose (YSI 2300 STAT Plus Glucose and Lactate Analyzer; YSI Inc., Yellow Springs, OH), and the glucose infusion rate was adjusted as needed to maintain euglycemia. Blood plasma was obtained every 10 min and stored at −80°C for the subsequent analysis of plasma insulin (Access Immunoassay System; Beckman Coulter, Fullerton, CA). A steady-state M-value was determined from the final 20 min of the clamp (35).

A biopsy specimen was obtained from the vastus lateralis with the percutaneous muscle biopsy technique at baseline and at 60 min of the clamp. The 60-min time point was selected because we have previously reported that components of insulin signal transduction (PI3-kinase activation and Akt Ser-473 phosphorylation) appeared to be maximally activated at this time (36–38). Tissue samples were immediately frozen in liquid nitrogen for subsequent analyses.

Skeletal muscle was homogenized and protein content determined as previously described (39,40). For Western blot analyses, muscle lysate (30–100 μg cellular protein) was separated by SDS-PAGE, electrotransferred onto polyvinylidene difluoride membranes (Millipore, Billerica, MA), and probed overnight with Cell Signaling (Beverly, MA) antibodies for PAS, phospho (p)Akt-Ser473 (recognizes Akt-Ser472/473/474), AS160, Akt2, and cyclooxygenase IV (COXIV). Membranes were also probed for pAS160-Thr642 (Millipore), pAS160-Ser666 (Millipore), pAS160-Ser588 (Symansis NZ Ltd, Timaru, New Zealand), GLUT4 (Affinity BioReagents, Golden, CO), RIP140 (Santa Cruz Biotechnology, Santa Cruz, CA), AS160 (Abcam, Cambridge, MA), and phospho-specific antibodies for AS160 at sites Ser-318 and Ser-751, as previously described (23,41). Proteins were visualized by horseradish peroxidase-conjugated IgG antibodies and ECL SuperSignal (Pierce Biotechnology, Rockford, IL) exposed to X-ray film. All samples were normalized to a control sample on each gel, and phosphorylation levels were additionally normalized to total protein after membranes were stripped, as previously reported (41), and reprobed with the corresponding antibody for total protein. For immunoprecipitation, lysates (200 μg) were incubated at 4°C overnight with Cell Signaling Technology antibodies for AS160 or Akt2 and for 3 h with protein A Sepharose beads (GE Healthcare Biosciences Corp., Piscataway, NJ). Supernatant portions from samples were removed and immunocomplexes analyzed with Western blotting.

Of the 73 subjects recruited for the cross-sectional study, 45 volunteered for the experiment examining the effects of exercise training on insulin signal transduction. Inclusion criteria required participants to be ≤35 years of age (young) or ≥55 years of age (aged). These individuals were then randomized into a 12-week endurance- (n = 12 young, n = 11 aged) or strength-training (n = 11 young, n = 11 aged) program. Two individuals were subsequently excluded from the young, endurance group due to noncompliance (Fig. 1).

All subjects performed preliminary cardiovascular, body composition, muscular strength, blood chemistry measurements, and a 3-day diet record, which were repeated during the final week of training (Tables 2 and 3). The euglycemic-hyperinsulinemic clamp with muscle biopsies was performed before exercise training and ∼40 h after the final exercise training session (Fig. 1). All subjects completed a 24-h diet record the day before their pretraining euglycemic-hyperinsulinemic clamp and were then asked to duplicate this diet the day before their post-training test.

The endurance-training program consisted of exercising on a graded treadmill, stationary cycle, or elliptical trainer within a target heart rate zone equivalent to 70–75% Vo₂peak for a total of 180 min/week (three to four sessions per week). To ensure the appropriate workload during training, a Vo₂peak test was performed after 6 weeks of training and workload adjusted appropriately.

Strength training consisted of upper and lower body exercises, performed three times per week (∼45 min/session). The upper body exercises included chest press, latissimus pull down, seated row, triceps pull down, and biceps curl. The lower body exercises included leg press, leg extension, and leg curl. Participants alternated between upper and lower body exercises to minimize fatigue, with 60–90 s rest between sets, and performed 10–12 repetitions to failure. When a subject could complete 12 repetitions on two consecutive occasions, resistance was increased by ∼5%. Two sets were completed for upper and lower body exercises during weeks 1–6, and during weeks 7–12, a third set was added for lower body exercises.

Analyses were performed using SPSS version 18.0 software (SPSS Inc., Chicago, IL). Pearson correlation coefficients and stepwise regression were used to determine associations. Comparisons between young and aged individuals before and after exercise training and under basal and insulin-stimulated conditions were performed with repeated-measures ANOVA. Significant main effects and interactions were further analyzed using unpaired (age group) and paired (pre- vs. post-training and baseline vs. insulin-stimulated) contrast-contrast comparisons. Data are presented as means ± SEM. Statistical significance was defined as P ≤ 0.05.

The women exhibited a lower body and fat-free mass, waist-to-hip ratio, Vo₂peak, and elevated body and trunk fat percentage compared with the men (Table 1). Univariate correlations indicated that mass (r = 0.26, P < 0.05), BMI (r = 0.45, P < 0.001), body fat percentage (r = 0.49, P < 0.001), trunk fat (r = 0.53, P < 0.001), and waist-to-hip ratio (r = 0.50, P < 0.001) increased with age. Relative Vo₂peak (mL/kg/min) declined with age (r = −0.78, P < 0.001), whereas fasting blood glucose (r = 0.40, P < 0.001) increased with age.

The euglycemic-hyperinsulinemic clamp increased (P < 0.001) plasma insulin concentration from fasting (5.0 ± 0.4 μIU/mL) to 160 ± 6 μIU/mL. Whole-body insulin action (M-value) declined with age (r = −0.52, P < 0.001; Supplementary Fig. 1). Sex had no effect on this relationship; therefore, data from men and women were combined for subsequent analyses.

Stepwise linear regression analysis was used to determine variables (age, body weight, BMI, percentage body fat, percentage trunk fat, waist-to-hip ratio, Vo₂peak, and fasting insulin) that independently predicted insulin action. Age and BMI were both independent predictors of insulin action (total adjusted R² = 0.44, P < 0.001).

Insulin infusion increased (P < 0.001) phosphorylation of all five AS160 phospho-specific sites and PAS by ∼2- to 8-fold and of Akt2 Ser-473 by ∼30-fold. As presented in Fig. 2, whole-body insulin action was positively related with AS160 phosphorylation when determined by PAS (r = 0.33, P < 0.01) and on sites Ser-588 (r = 0.34, P < 0.05), Thr-642 (r = 0.33, P < 0.05), and Ser-666 (r = 0.32, P < 0.05). Insulin action was not related to AS160 phosphorylation of Ser-751 or Ser-318 sites or with Akt2 Ser-473 phosphorylation (n = 61, r = 0.15, P > 0.05).

Stepwise linear regression analysis was used to determine variables (age, BMI, percentage body fat, percentage trunk fat, waist-to-hip ratio, Vo₂peak, fasting glucose, and insulin) that independently predicted insulin-induced phosphorylation of AS160. Chronological age was the sole predictor of insulin-induced phosphorylation of AS160 Ser-666 (R² = 0.12, P < 0.05). Body fat was determined to be the best predictor of insulin-induced phosphorylation of AS160 PAS (R² = 0.12, P < 0.01) and AS160 Ser-588 (R² = 0.10, P < 0.05), whereas fasting insulin was determined to be the best predictor of insulin-induced phosphorylation of AS160 Thr-642 (R² = 0.10, P < 0.05).

Age was not related to the ability of insulin to phosphorylate Akt2 Ser-473 (r = −0.18, P > 0.05; data not shown). Chronological age was negatively related to the insulin-stimulated phosphorylation of AS160 determined by PAS (r = −0.25, P < 0.05) and sites Ser-588 (r = −0.28, P < 0.05), Thr-642 (r = −0.28, P < 0.05), and Ser-666 (r = −0.30, P < 0.05; Supplementary Fig. 2). Basal phosphorylation levels of AS160 Thr-642 increased with age (r = 0.30, P < 0.05; data not shown). Age had no other effect on AS160 basal phosphorylation levels or protein content. GLUT4 protein content demonstrated a trend for a negative relationship with age (r = −0.24, P = 0.06; Supplementary Fig. 2).

Total RIP140 protein content was determined in 19 young (range 18–35 years) and 20 insulin-resistant aged (range 57–84 years) individuals, and the amount of RIP140 complexed with AS160 was determined in a subset of 11 young (range 18–26 years) and 9 insulin-resistant aged (range 56–82 years) individuals. There were no differences in total RIP140 protein content between young and aged individuals (Fig. 3A), but insulin-resistant aged individuals had a higher amount of RIP140 complexed with AS160 (P < 0.03, Fig. 3B) compared with their young counterparts. As presented in Fig. 3C, the amount of RIP140 complexed with AS160 exhibited a trend (r = −0.42, P = 0.06) to be negatively related to insulin action.

Differences in whole-body insulin action (M-value), body composition, Vo₂peak, and fasting blood chemistries were evident between the young and aged groups before initiating the 12 weeks of exercise training (Tables 2 and 3). With endurance training, there was a reduction in body fat and an increase in lean body mass, which resulted in no change in overall body mass (Table 2). Endurance training also increased insulin action (M-value), Vo₂peak, and 1-repetition maximum (RM) chest press and decreased fasting plasma glucose and insulin (Table 2). Strength training increased 1-RM leg press, 1-RM chest press, and peak isokinetic leg flexion strength in young and insulin-resistant aged participants and also increased lean and total body mass and Vo₂peak (Table 3). Insulin action also increased irrespective of group, whereas fasting plasma glucose and insulin concentrations decreased in response to strength training (Table 3). Skeletal muscle COXIV protein, a marker of mitochondrial content, increased in response to endurance training (∼30%) but not with strength training (Supplementary Fig. 3). Analysis of 3-day diet records revealed no significant changes in caloric intake (total, protein, carbohydrate, and fat) during the training period.

In agreement with the correlation analyses (Supplementary Fig. 1), reductions in insulin-stimulated Ser-588, Thr-642, Ser-666 (P < 0.05), and PAS (P = 0.06) phosphorylation were evident in insulin-resistant aged individuals in the pretraining, sedentary state (Fig. 4). Endurance training increased insulin-stimulated AS160 phosphorylation of PAS by ∼60% in young and ∼75% in insulin-resistant aged individuals (P < 0.05, Fig. 5A), whereas AS160 Ser-588 phosphorylation increased ∼25% in both groups (P < 0.05, Fig. 5B). There was a significant interaction (P < 0.05) for AS160 Thr-642 in response to endurance training as the aged individuals increased AS160 Thr-642 phosphorylation by ∼57% (P < 0.05, Fig. 5C), whereas no significant changes were observed in the young individuals. There was a tendency (P = 0.07) for insulin-stimulated Ser-666 phosphorylation to increase with endurance training in the aged subjects (Fig. 5D). No changes with endurance training were evident in insulin-stimulated Ser-751 and Ser-318 (data not shown). Strength training increased insulin-stimulated AS160 phosphorylation of PAS by ∼75% (P < 0.01, Fig. 6A ) in both groups, whereas insulin-stimulated Thr-642 phosphorylation increased by ∼33% and ∼73% in young and aged individuals (P < 0.05, Fig. 6C,), respectively. In addition, strength training increased insulin-stimulated Ser-666 phosphorylation of AS160 by ∼100% (Fig. 6D) in the insulin-resistant aged group. There were no changes in Ser-588 (Fig. 6B), Ser-751 (data not shown), or Ser-318 (data not shown) with strength training.

Exercise training had no effect on AS160 protein content (data not shown). Basal phosphorylation levels of PAS were significantly reduced (∼50%, P < 0.01) in response to strength training, which contributed to the significant training effect (Fig. 6A). Exercise training had no other effect on basal AS160 phosphorylation levels. Endurance training increased GLUT4 (∼10% for young and ∼15% for aged, P < 0.05; Supplementary Fig. 4), whereas strength training had no effect. Multiple regression analysis indicated that exercise-induced changes in phosphorylation of PAS, Ser-588, Thr-642, and Ser-666 accounted for 28% of the variance in the improvement in insulin action (P < 0.05).

Before exercise training, Akt2 protein levels did not differ between the groups (data not shown). Endurance training increased Akt2 protein levels by ∼60% in the young group (P < 0.01), with a similar trend in the aged individuals (P = 0.08). In response to strength training, Akt2 protein levels demonstrated a trend for an increase (P = 0.08). Neither age nor exercise training had an effect on Akt2 Ser-473 phosphorylation when normalized to protein content (Supplementary Fig. 5).

In the current study, we show for the first time that insulin-stimulated AS160 phosphorylation, measured by the PAS antibody, and specific phosphorylation at sites Ser-588, Thr-642, and Ser-666 are impaired in human skeletal muscle in conjunction with the decrement in insulin action typical with advancing age and a sedentary lifestyle (Figs. 2 and 4). Impaired insulin-mediated AS160 phosphorylation has been reported in other insulin-resistant conditions, including type 2 diabetes (12) and polycystic ovary syndrome (14), using the PAS antibody. The PAS antibody may recognize multiple phosphorylation sites on AS160; however, current research suggests this antibody is limited to only recognizing AS160 phosphorylation on Thr-642 (9,23). More recently, site-specific impairments were identified in patients with type 2 diabetes (Ser-318, Ser-588, and Ser-751) (13) and in healthy individuals after fasting-induced insulin resistance (Ser-588 and Ser-751) (4). The current data (Figs. 2 and 4), in combination with other data (4,13), provide the important information that an impairment in Ser-588 phosphorylation appears to be consistent in human skeletal muscle across conditions inducing insulin resistance. In contrast, other AS160 sites demonstrate differential phosphorylation patterns, possibly as a product of kinases and phosphatases being regulated by the severity or the pathology of insulin resistance. Collectively, these findings show that conditions of whole-body insulin resistance are linked with site-specific impairments in AS160 and provide novel insight into a signaling impairment located distally in the insulin-signaling cascade.

In an effort to investigate cellular mechanisms that could contribute to the impaired AS160 phosphorylation, we examined RIP140 expression and its association with AS160. In adipocytes, the binding of RIP140 to AS160 results in reduced glucose uptake, likely as a result of RIP140 impeding the ability of Akt to inactivate GAP activity on AS160 (15). The current finding that insulin-resistant aged individuals had a higher amount of RIP140 complexed with AS160 (Fig. 3B) provides novel evidence that the impairment in AS160 phosphorylation may be linked to the association of AS160 with RIP140, which in turn induces insulin resistance (Fig. 3C).

In an attempt to gain an understanding of factors that may regulate site-specific phosphorylation on AS160, we performed regression analyses using variables linked with insulin action. Body fat percentage was the best predictor of Ser-588 phosphorylation, whereas basal plasma insulin levels proved to be the best predictors of Thr-642 phosphorylation, suggesting these phosphorylation sites may be differentially regulated. In agreement with these findings, in vitro experiments in adipocytes demonstrated that insulin-stimulated phosphorylation of Thr-642 occurs much more rapidly than Ser-588, and hierarchical clustering analysis revealed that Thr-642 did not cluster with Ser-588 (42). Taken together, this information provides insight into potential regulatory mechanisms of AS160 phosphorylation; however, we acknowledge that regression analyses only imply relationships and that additional variables not measured in the current study may also play a role in the regulation of site-specific AS160 phosphorylation.

Of the two Akt isoforms expressed in skeletal muscle (Akt1 and Akt2), Akt2 is considered crucial for glucose uptake in skeletal muscle (43). In relation to upstream signaling of AS160, we recognize that phosphorylation of both Akt Ser-473 and Thr-308 is required for the full activation of Akt; however, the current study was limited to Akt2 Ser-473 based on results in human skeletal muscle indicating that insulin-stimulated Akt2-Ser473 phosphorylation (as opposed to Akt-Thr308) was closely related to AS160-PAS phosphorylation and glucose uptake (43). Insulin-stimulated Akt2-Ser473 phosphorylation was not associated with insulin-resistant aged individuals in the current study; however, Sharma et al. (44), recently reported reduced insulin-stimulated Akt2-Thr308 phosphorylation in the soleus of aged (25-month) compared with adult (9-month) rats; therefore, we cannot conclusively state that all Akt sites were preserved with insulin resistance.

Exercise training has long been recognized as a method to improve insulin action (16–19) (Tables 2 and 3). The effect of exercise training on insulin-stimulated AS160 phosphorylation has been sparsely addressed, particularly in regards to phospho-specific sites. Previous research reported that insulin-mediated AS160 phosphorylation increased in healthy young men after 3 weeks of one-legged endurance-oriented exercise training; however, these effects were negated when phosphorylation was normalized to AS160 protein content (21). In addition, O’Gorman et al. (22) reported that short-term endurance training (7 days) was not sufficient to increase insulin-stimulated AS160 phosphorylation in obese, nondiabetic, or diabetic individuals. However, conclusions from these studies (21,22) are limited based on the use of the PAS antibody.

A key finding in the current study was that decrements in specific insulin-stimulated AS160 phosphorylation sites were improved with exercise training (Figs. 5 and 6), with the exception of AS160 Ser-588 (Fig. 6B), which did not appear to be responsive to strength training (Fig. 6B). Vind et al. (13) previously reported increased insulin-induced AS160 phosphorylation on Ser-588 in type 2 diabetic patients, but not in nondiabetic control subjects, in response to 10 weeks of endurance training, suggesting that this site may be particularly responsive to endurance exercise in insulin-resistant populations. From our cross-sectional data, we determined that body fat percentage was the best predictor of AS160 Ser-588 phosphorylation. Body fat percentage was reduced in response to endurance, but not strength training, which could explain why improvements in Ser-588 phosphorylation were only evident with this training modality. Protein kinase Cζ (PKCζ) activity has been hypothesized to regulate AS160 Ser-588 phosphorylation (42), and endurance training has been reported to increase skeletal muscle PKCζ activity (45). Although speculative, it is plausible to suggest that our endurance-training program improved PKCζ activity, which could in part explain improvements in insulin-stimulated AS160 Ser-588 phosphorylation.

Both modes increased PAS, Thr-642, and Ser-666 (P = 0.07 with endurance training) phosphorylation in the aged group, indicating the effectiveness of exercise in treating insulin resistance. The clinical relevance of our findings is that either endurance or strength training appears to improve insulin action through similar mechanisms in relation to insulin signaling at the level of AS160. This finding may provide relevant information in terms of therapeutic treatments for insulin-resistant conditions.

Consistent with other studies examining human subjects over a wide life span (1,24,26,30,46), our data demonstrated that whole-body insulin action declined with age (Supplementary Fig. 1). The nature of this age-related insulin resistance has been well-studied and likely involves a number of contributing factors, including increased abdominal adiposity (30), declining cardiorespiratory fitness (29,31), and chronological age itself (29), all of which were evident in our population (Table 1). Despite reports suggesting that the effect of chronological age is negated when adjusting for BMI (26), adiposity (27), or physical inactivity (28), our data indicate that chronological age was an independent predictor of whole-body insulin action, which is in agreement with the findings of at least one other study (29).

A limitation of the current study was that muscle fiber typing was not performed. Animal studies have reported greater age-related impairments in glucose uptake in slow-twitch compared with fast-twitch muscle (44,47), despite the apparent preservation of type I fiber cross-sectional area with aging (48). Likewise, 12 weeks of endurance or strength training has been associated with increases in type I (49) and type II (50) fiber area, respectively, independent of age (48,49). Therefore, the age- or exercise-related differences in AS160 phosphorylation in our current study could possibly have been influenced by changes in muscle fiber type. In addition, despite previous research suggesting maximal AS160 phosphorylation occurs at 60 min of an euglycemic-hyperinsulinemic clamp (36,38), we cannot exclude the possibility that the rate of site-specific AS160 phosphorylation was influenced by either age and/or exercise training.

In conclusion, the findings of the current study indicate for the first time that deficits in whole-body insulin action evident with the aging process and a sedentary lifestyle are associated with reduced insulin-stimulated phosphorylation of specific AS160 sites (Thr-642, Ser-588, Ser-666, and PAS). With respect to intervention/prevention, 12 weeks of endurance- or strength-oriented exercise training increased whole-body insulin action and rescued impairments in AS160 phosphorylation. Collectively, these findings suggest that decrements in the ability of insulin to phosphorylate specific sites on skeletal muscle AS160 contribute to insulin resistance and that exercise training is an effective treatment option to counteract these impairments.

This work was supported by National Institutes of Health grants AG-025205 and DK-56112 to J.A.H. and by the Danish Medical Research Council and The Novo Nordisk Research Foundation to J.F.P.W. J.T.T. was supported by a postdoctoral fellowship from The Danish Agency for Science, Technology and Innovation, Denmark. The Novo Nordisk Foundation Center for Basic Metabolic Research is an independent Research Center, based at the University of Copenhagen, Denmark, and partially funded by an unconditional donation from the Novo Nordisk Foundation (www.metabol.ku.dk). No other potential conflicts of interest relevant to this article were reported.

L.A.C. and J.A.H. conceived and developed the experiments. L.A.C. and J.V.M. collected and analyzed the exercise training data. L.A.C., C.A.N., D.N.C., M.S.D., C.J.T., and J.A.H. collected and analyzed euglycemic-hyperinsulinemic data. C.A.N., D.N.C., M.S.D., and J.A.H. obtained skeletal muscle biopsies. L.A.C. generated Western blot data. L.A.C., J.F.P.W., J.T.T., and J.A.H. analyzed Western blot data. L.A.C. and J.A.H. wrote the manuscript. All authors provided comments and approved the final version. L.A.C. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

The authors thank Rita Bowden, RN, and Angela Clark, RN, East Carolina University, for their clinical expertise during the euglycemic-hyperinsulinemic clamp procedure; and Dr. Kristen Boyle, Dr. Benjamin Bikman, Todd Weber, MS, and Gina Battaglia, MS, East Carolina University, for assisting with data collection during the euglycemic-hyperinsulinemic clamp procedure. The authors are grateful to the graduate students in the Kinesiology Department at East Carolina University for monitoring the exercise training programs of the participants and express their appreciation toward the study participants.