Troglitazone Treatment Increases Protein Kinase B Phosphorylation in Skeletal Muscle of Normoglycemic Subjects at Risk for the Development of Type 2 Diabetes

A total of 20 FDRs (with at least 1 first-degree relative [parent] and 1 second-degree relative with type 2 diabetes) were studied (Table 1). All were Caucasians and had a sedentary lifestyle. The subjects were instructed not to engage in excessive physical activity for at least 2 days before each metabolic investigation. None had impaired glucose tolerance as determined by an oral glucose tolerance test or any other disease as evaluated by clinical investigation and standard laboratory assessments. Before entry into the study, the purpose and risks of the experimental protocol were carefully explained to all of the volunteers and informed written consent was obtained. The protocol was approved by the local ethics committee and was in accordance with the Helsinki II declaration.

Subjects were randomized into two groups in a double-blind manner and treated for 3 months with troglitazone (200 mg/day) or placebo (containing excipients similar to the troglitazone tablets). Before and after these treatment periods, 3-h euglycemic-hyperinsulinemic glucose clamps were performed. For the clamp studies, subjects were admitted to the clinical research center at 8:00 a.m. after an overnight fast. Intravenous catheters were inserted for infusion of insulin and glucose into the antecubital vein, and blood samples for plasma glucose analyses were drawn from a catheter of the dorsal wrist vein of the opposite arm. During the study, the subjects remained in a supine position with the hand designated for collection of blood samples placed and maintained in a heated box for arterialization of venous blood (34). After a 30-min relaxation period, a surface-adjusted primed-constant dose of [ 3-³H]glucose was given (14 μCi Ci/m²) (DuPont-New England Nuclear, Boston, MA). This was followed by a continuous infusion of [3-³H]glucose (0.14 μCi · min⁻¹ · m⁻²), and, after a 120-min basal tracer equilibration period, insulin (Actrapid; Novo-Nordisk, Bagsvaerd, Denmark) was infused at a rate of 40 mU · min⁻¹ · m⁻² for 180 min. Euglycemia (5.9 ± 0.1 mmol/l) was maintained using a variable infusion of 18% glucose with added [ 3-³H]glucose (34). Steady-state periods were defined as the last 30 min during basal and insulin-stimulated measurements, respectively, and the glucose disposal rate was calculated from the plasma concentrations of tritiated glucose and plasma glucose (34). Body composition was estimated with the bioimpedance method using the formula of Kushner and Schoeller (35). Plasma glucose, insulin, triglycerides, and free fatty acid (FFA) were analyzed as described (34).

At the end of the basal and insulin-stimulated steady-state periods, percutaneous vastus lateralis muscle biopsies were obtained under local anesthesia (2% lidocaine without epinephrine) 20 cm above the knee (34). This was done at separate incision sites to avoid local effects of the prior biopsy and blood contamination. The biopsies were immediately frozen in liquid nitrogen. For the subsequent analytical procedures, ∼50 mg frozen tissue was homogenized at 4°C in 500 μl solubilization buffer (8 mmol/l EDTA, 20 mmol/l HEPES, 1% IGEPAL, 530 mmol/l NaF, 30 mmol/l sodium pyrophosphate, 7 mmol/l sodium vanadate, 2 mmol/l dichloroacetic acid, 9 mmol/l phenylmethylsulfonyl fluoride, 21 mmol/l benzamidine, 5 μmol/l leupeptin, 4 μmol/l pepstatin, and 10 μg/ml aprotinin; pH 7.4). After 20 min, detergent insoluble material was removed by centrifugation (12,000g for 20 min at 4°C), and the protein concentration was measured using a protein-dye binding assay (BioRad, Munich).

These were measured as previously described (34, 36, 37). Briefly, 40 μl muscle sample lysates were added to microwells coated with anti-insulin receptor antibody for 16 h at 4°C. The wells were washed, and receptor-mediated ³²P incorporation into recombinant IRS-1 (17 nmol/l) (Upstate Biotechnology, New York, NY) was measured at 120 nmol/l ³²P-ATP. [¹²⁵I-Tyr-A¹⁴]monoiodoinsulin (Amersham-Pharmacia, Freiburg, Germany) binding to immobilized insulin receptors was also measured in the wells (36). Insulin binding capacity was defined as the amount of specifically bound insulin at a concentration of 8.7 nmol/l (34, 37).

Solubilized muscle (500 μg protein) was added to protein G Sepharose beads (Pierce, Rockford, IL) to which anti-IRS-1 antibody (2 μg) (Upstate Biotechnology) had been bound. After 4 h at 4°C, the beads were washed three times with PBS containing 1% IGEPAL and 0.1 mmol/l sodium orthovanadate (pH 7.5), three times with 500 mmol/l LiCl, 100 mmol/l Tris, and 0.1 mmol/l sodium orthovanadate (pH 7.5), and twice in reaction buffer (10 mmol/l Tris, 100 mmol/l NaCl, 1 mmol/l EDTA, and 0.1 mmol/l sodium orthovanadate; pH 7.5). The pellets were then suspended in 60 μl of reaction buffer supplemented with 15 mmol/l MgCl₂ and phosphatidylinositol (20 μg) (Sigma, Deisenhofen, Germany). The phosphorylation reaction was started by the addition of 10 μl of a solution that contained 720 μmol/l ATP (25 μCi/mmol) (NEN, Dreieich, Germany). After 20 min at room temperature, the reaction was stopped by the addition of 40 μl 25% HCl and 190 μl methanol/chloroform (1:1). The organic phase was extracted and applied to a silica gel thin-layer chromatography plate. Phosphatidylinositol-3 phosphorylation was determined using a phosphorimager.

Tissue lysates (100 μg protein for the detection of IRS-1, IRS-2, phosphoinositol-dependent kinase-1 [PDK-1], PKB, PKB Ser⁴⁷³, and PKB Thr³⁰⁸ phosphorylation and 50 μg protein for the detection of GLUT-4) were separated by SDS-PAGE and transferred to nitrocellulose (Schleicher & Schuell, Dassel, Germany). Proteins were detected with antibodies against IRS-1 or -2, p85, PDK-1 (all from Upstate Biotechnology), PKB, Ser⁴⁷³- or Thr³⁰⁸-phosphorylated PKB (all from BioLabs, Beverly, MA) or GLUT-4 (Chemicon International, Temecula, CA), horseradish peroxidase-labeled secondary antibodies (Dako Diagnostics, Hamburg, Germany), and enhanced chemiluminescence (ECL; Amersham-Pharmacia, Freiburg, Germany). The antibodies against PKB and phosphorylated PKB were known to detect the α, β, and γ isoforms. Bands were quantified densitometrically with Molecular Analyst software (BioRad).

To exclude distortion of the results stemming from interassay variation between PI3K assays or immunoblots, material from two troglitazone-treated and two placebo-treated subjects was analyzed in parallel per experiment (total of 16 samples: basal and insulin-stimulated biopsies from two clamps in four subjects). For most parameters, at least two internal standard samples (pooled human muscle samples from clamp biopsies) were also analyzed per experiment, and the densitometric (Western blots) or phosphorimager-derived value (PI3K activity) obtained for the individual samples was expressed as a percentage of the value obtained for these standards. For some parameters, internal standard data were not available and arbitrary units were calculated. This was done by dividing the densitometric- or phosphorimager-derived value obtained for the individual sample by the average value of all 16 samples that had been studied in parallel within the same experiment. Statistical comparisons were carried out by t tests for paired or unpaired data where appropriate. Differences were considered significant at P < 0.05.

The subjects in the troglitazone and placebo group did not differ in BMI or percent fat-free mass (Table 1), and there were no significant changes in these parameters during the treatment period (data not shown). Basal glucose disposal was not different after treatment with troglitazone (Fig. 1); however, the plasma insulin concentration was decreased (Table 2), indicating increased insulin sensitivity. The insulin effect on glucose disposal was 50% higher after treatment with troglitazone, an effect not observed with placebo (Fig. 1). Troglitazone treatment also resulted in lower fasting plasma triglyceride levels (Table 2). Fasting FFA levels were not different after troglitazone treatment, although there was a significant difference of the suppressed FFA levels at the end of the clamp (Table 2).

The increase of plasma insulin concentration from basal to steady-state clamp levels resulted in two- to threefold increases of IRK and PI3K activities, and no significant effects of the troglitazone treatment were observed (Fig. 2A and B). PKB Ser⁴⁷³ phosphorylation was also increased three- to fourfold by the clamp insulin infusion, and, in contrast to IRK and PI3K activities, in which no troglitazone effects were found, insulin-stimulated PKB Ser⁴⁷³ phosphorylation was significantly higher after troglitazone treatment (94 or 152% compared with the respective pretreatment or placebo treatment level, respectively) (Fig. 2C). PKB Ser⁴⁷³ phosphorylation in the biopsies obtained at the basal insulin level was also significantly higher after troglitazone treatment compared with pretreatment. There was also a trend for an enhanced insulin action on the phosphorylation of the other activating site of PKB, Thr³⁰⁸, after troglitazone treatment (Fig. 2D). However, most likely due to the higher variation of the results, there was no statistically significant difference.

There was a tendency to slightly higher protein expression of PDK-1 and PKB in the biopsies that were obtained after troglitazone treatment; however, statistically significant effects on protein expression were not observed for any of the investigated proteins (Table 3).

In the present placebo-controlled study, we report that the increase in insulin-stimulated glucose disposal after 3 months of troglitazone treatment was associated with an increased Ser⁴⁷³ phosphorylation of PKB at basal and clamp insulin concentrations. Because our study population consisted of insulin-resistant subjects with a genetic background for type 2 diabetes but normal glucose tolerance, the treatment did not alter the plasma glucose concentration. The observed increases of PKB phosphorylation and glucose disposal were therefore independent of a normalization of plasma glucose levels that by itself could have lead to a normalization of insulin signaling and action (29–31).

PKB has been proposed to be an intermediate in the signaling pathway by which insulin, among other things, controls glucose uptake (17, 19, 20). Although there remains some controversy with respect to the exact role of PKB in the insulin signaling toward glucose uptake (38, 39), strong lines of evidence indicate that Thr³⁰⁸ and Ser⁴⁷³ phoshorylation and concomitant activation of the enzyme are important. Thus, overexpression of constitutively active forms of PKB induced glucose uptake, GLUT-4 translocation, and glycogen synthesis (16, 40), and recently it was demonstrated that mice deficient in PKB-β, due to targeted disruption of the PKB-β locus, were insulin resistant in muscle and liver (21). Finally, in IRS-1-deficient mice, adenovirus-mediated gene therapy with an IRS-1 mutant lacking the binding site for p85 enhanced PKB activity and normalized insulin sensitivity without increasing PI3K activity, suggesting that PKB activation might be sufficient for reducing insulin resistance even without increasing PI3K activity (18). Therefore, the increased PKB Ser⁴⁷³ phosphorylation and the tendency to increased Thr³⁰⁸ phosphorylation after long-term troglitazone treatment observed in our study could be part of a mechanism that leads to the improved insulin sensitivity of glucose uptake and glucose storage (41) in human skeletal muscle.

The notion that the insulin-sensitizing effect of PPARγ agonists in skeletal muscle involves increased signaling at the level of PKB is consistent with previous reports. An increased insulin-stimulated PKB phosphorylation has also been observed in skeletal muscle of obese Zucker fatty rats that had been treated with troglitazone for 3 weeks (22), in the membrane fraction of cultured human skeletal muscle cells that had been treated with troglitazone for 5 days (10), and in isolated human adipocytes that had been treated for 12 h with the PPARγ agonist BRL 49563 (24). Moreover, increased insulin-stimulated adipocyte PKB serine phosphorylation (4) and enhanced insulin-stimulated skeletal muscle PI3K and PKB activity (32) have most recently been reported in type 2 diabetic subjects after troglitazone treatment for 3–4 months. Because these two latter studies, in contrast to our study, were performed with type 2 diabetic subjects, treatment-induced changes in plasma glucose levels were inevitable. Therefore, to demonstrate that the troglitazone effects on insulin signaling were not simply due to the changes in glycemia, a similar group of subjects was treated with metformin. Since both compounds led to comparable improvements in metabolic control, but effects on PI3K and PKB were only observed with troglitazone, the authors concluded that it was unlikely that a generalized reduction in glycemia was responsible for the observed troglitazone effects on insulin signaling (4,32).

In contrast to several previous reports, in which the troglitazone-induced increases in PKB phosphorylation or activity appeared to be related to an increased PI3K activity (10, 32) or increased IRS-1 phosphorylation (22), our data suggest that the increase in PKB phosphorylation occurred in the absence of a change in IRS-1-associated PI3K activity. We cannot exclude that small differences in IRS-associated PI3K activity might have been missed in our study. On the other hand, it is certainly conceivable that, as suggested by our data, the troglitazone effect on PKB phosphorylation was independent of changes in PI3K activity. For example, it has been observed that hyperglycemia reduced insulin-stimulated PKB phosphorylation despite PI3K activity being unchanged (30) or even increased (31). The current model for insulin stimulation of PKB phosphorylation and activity involves the PI3K-mediated generation of phosphatidylinositol-3,4,5-tris and 3,4-bisphosphate in the plasma membrane. This induces the translocation of PKB from the cytosol to the plasma membrane and its phosphorylation by PDK-1 on the Thr³⁰⁸ site. PDK-1 appears to also be involved in the phosphorylation of the Ser⁴⁷³ site (17, 19, 42), and overexpression of PDK-1 has been shown to increase GLUT-4 translocation in adipocytes (43). Thus, there are multiple locations where the signal between PI3K and PKB could be modulated, including PDK-1 and the phosphoinositide phosphatases (44).

There was a tendency to slightly higher protein expression of PDK-1 and PKB in the biopsies obtained after troglitazone treatment (Table 3), and although this was not statistically significant, a contribution of such changes in protein expression to the troglitazone effects on PKB phosphorylation and/or insulin sensitivity cannot be excluded. Our finding that the insulin binding capacity and the protein expression of IRS-1, IRS-2, and GLUT-4 were not changed after therapy is in partial contrast to findings in human adipocytes, in which increased IRS-2 (26) or GLUT-4 (4) expression has been observed, and may indicate tissue-specific modulation of protein expression. The observation that GLUT-4 protein expression was not increased by troglitazone is also in contrast to a previous report in rat skeletal muscle (22) but consistent with most data in muscle cell culture systems (8, 10, 28).

Evidently, due to the limited amount of tissue available, only a small spectrum of the proteins that might contribute to the effects of PPARγ agonists on insulin signaling has been studied, and it is not clear whether, or to what degree, the effects on signaling observed in the present study contributed to the increase in insulin sensitivity. Other effects of PPARγ -agonists on insulin signaling that have been described include, for example, increased cbl-associated protein expression in 3T3L1-adipocytes (45) or increased protein kinase C-ζ and -λ activation in rat adipocytes (25). Nevertheless, our data demonstrate that a 3-month treatment of nondiabetic insulin-resistant human subjects enhances insulin action on PKB phosphorylation in skeletal muscle, the most important tissue for peripheral insulin-stimulated glucose disposal (46).

Whereas some studies suggest that skeletal muscle PKB activity or phosphorylation in response to insulin may be impaired in type 2 diabetes (47), other studies (48), including our own (33), indicate that this is not the case. Moreover, the FDRs in our study were insulin resistant compared with non-FDR control subjects, but insulin-stimulated PKB phosphorylation was similar (33). The troglitazone effect observed in the present study therefore does not appear to reflect a normalization of a diabetes-associated impairment of insulin-stimulated PKB phosphorylation or activity, but rather an enhancement above normal. This does not exclude, however, that the increased PKB phosphorylation contributes to the observed normalization of glucose disposal. It might, for example, compensate for impairments that exist in FDRs and type 2 diabetic subjects at other locations, e.g., downstream of the investigated signaling cascade or in alternative signaling pathways (33).

The advantage of an investigation of FDRs rather than type 2 diabetic subjects was that treatment-induced metabolic changes, which by themselves could have influenced insulin signaling, were much smaller. Thus, there were no detectable troglitazone effects on fasting plasma glucose or FFA concentrations, both of which are known to influence insulin signaling (30, 31, 49). FFA levels at the clamp insulin level were very low; therefore, it appears rather unlikely that the troglitazone-induced difference in plasma FFA levels observed under this condition affected insulin signaling. This does not, of course, exclude the possibility that the troglitazone effects on skeletal muscle PKB phosphorylation and/or glucose disposal were through secondary mechanisms. For example, altered intramyocellular lipid levels (50), altered local metabolite concentrations, or altered adipocyte mediators of insulin resistance have contributed to the altered insulin signaling and/or action.

In summary, we have shown in normoglycemic FDRs that the increased glucose disposal after a 3-month treatment with troglitazone was associated with increased PKB phosphorylation in skeletal muscle. This increase in PKB phosphorylation may contribute to the insulin-sensitizing effects of PPARγ agonists.

This work was supported in part by grants from the Deutsche Forschungsgemeinschaft (KL 503/7-3 to H.H.K.), the German Diabetes Association (H.H.K.), the Danish Diabetes Association (K.L. and H.B.-N.), the Novo Nordisk Foundation (K.L.), and the Institute of Clinical Research, Odense University Hospital (H.B.-N.).