Troglitazone but not Metformin Restores Insulin-Stimulated Phosphoinositide 3-Kinase Activity and Increases p110β Protein Levels in Skeletal Muscle of Type 2 Diabetic Subjects

A total of 14 type 2 diabetic subjects (11 men and 3 women between the ages of 30 and 70 years) who were poorly controlled (HbA_1c >8.5% and fasting plasma glucose >140 mg/dl) on at least half-maximal doses of any sulfonylurea agent were recruited for these studies. Except for diabetes, the subjects were healthy and on no other medications known to influence glucose metabolism. After screening, their current sulfonylurea medication was discontinued and all subjects were uniformly started on glyburide at 10 mg b.i.d. for at least 4 weeks. Baseline studies were subsequently performed and subjects randomized to the addition of either troglitazone or metformin to glyburide therapy. Subjects were counseled at 2-week intervals to consume a weight-maintenance diet for the duration of the study. Over 4–6 weeks the troglitazone treatment was titrated up to 600 mg/day or metformin up to 2,550 mg/day as required to achieve glycemic goals. After 3–4 months of this additional therapy, patients were readmitted for repeat studies. These subjects were part of a larger study group comparing the effects of troglitazone and metformin therapy on multiple clinical parameters (21). With regard to clinical parameters and response to treatment, the current subjects were indistinguishable from the larger study group. The experimental protocol was approved by the Committee on Human Investigation, University of California, San Diego. Informed written consent was obtained from all subjects after explanation of the protocol.

All subjects were admitted to the Special Diagnostic and Treatment Unit at the VA San Diego Healthcare System. To quantitate peripheral insulin action, all subjects underwent a 3-h hyperinsulinemic (300 mU · m⁻² · min⁻¹) euglycemic (5.0–5.5 mmol/l) clamp after a 12-h overnight fast, as previously described (22). The glucose disposal rate (GDR) was determined during the last 30 min of the clamp. Percutaneous needle biopsies of vastus lateralis muscle were performed before and after 3 h of insulin infusion, as previously described (22), and muscle tissue was immediately frozen in liquid nitrogen. Plasma glucose and insulin levels were determined before each biopsy.

A total of 50 mg of muscle was homogenized using a polytron at half-maximum speed for 1 min on ice in 500 μl buffer A (20 mmol/l Tris [pH 7.5], 5 mmol/l EDTA, 10 mmol/l Na₄P₂O₇, 100 mmol/l NaF, and 2 mmol/l Na₃VO₄) containing 1% NP-40, 1 mmol/l phenylmethylsulfonyl fluoride, 10 mg/ml aprotinin, and 10 mg/ml leupeptin. Tissue lysates were solubilized by continuous stirring for 1 h at 4°C and centrifuged for 10 min at 14,000g. The supernatants were stored at −80°C until analysis.

Muscle lysates (150 μg protein) were subjected to serial immunoprecipitations with 5 μl of a polyclonal IRS-1 antibody (1:100 dilution; gift from Dr. Morris White, Joslin Diabetes Center) coupled to protein A-sepharose beads (Sigma, St. Louis, MO) overnight, followed by 2 μg of a polyclonal Akt antibody that recognizes both Akt1 and Akt2 (Upstate Biotechnology, Lake Placid, NY) coupled to protein G-sepharose beads (Pharmacia Biotech, Piscataway, NJ) for 4 h. The immune complex was washed, and PI 3-kinase and Akt activities were determined as previously described (7). The effect of insulin to stimulate Akt activity tended to be lower than in our previous report (7), probably because of a different immunoprecipitation methodology. In our previous report, we performed immunoprecipitation for 4 h and obtained approximately a twofold insulin stimulation (300 mU insulin · m⁻² · min⁻¹ for 3 h) of Akt activity over the basal value. Because of limitations in the size of muscle samples in the present study, we carried out serial immunoprecipitations (see above), which could result in less activation of Akt in response to insulin.

A total of 10 μg tissue lysate protein per lane was resolved by SDS-PAGE (7.5 and 10% gels) and transferred to nitrocellulose membranes (Schleicher & Schuell). The membranes were blocked with 5% nonfat dry milk overnight at 4°C and incubated with the following antibodies for 3 h at room temperature: a polyclonal IRS-1 antibody (Upstate Biotechnology), a polyclonal antibody against the p85α subunit or p110β subunit of PI 3-kinase (Upstate Biotechnology), and a polyclonal Akt antibody (Upstate Biotechnology) in 1% nonfat dry milk. The membranes were washed and bands visualized as previously described (23).

Data are presented as the mean ± SEM. Statistical analyses were performed using the Stat View program (Abacus Concepts, Berkeley, CA). Statistical significance was tested with the paired Student’s t test when appropriate.

The subjects in the troglitazone and metformin groups were matched for age and obesity. After 3 months of troglitazone treatment, BMI was slightly increased compared with before treatment (P < 0.05). Subjects on metformin for 3 months had no change in BMI. HbA_1c, glucose, and insulin concentrations were significantly reduced by troglitazone therapy (Table 1). Triglyceride and free fatty acid concentrations tended to be reduced after troglitazone treatment. Similar to the effects of troglitazone, metformin treatment significantly decreased HbA_1c, glucose, and insulin concentrations (Table 1). Triglycerides were lower (P < 0.05) after metformin treatment (Table 1). The GDR, as determined by a hyperinsulinemic-euglycemic clamp, was increased by 35% (P < 0.01) after troglitazone treatment and by 22% (P < 0.05) after metformin (Table 1). The change in GDR after troglitazone treatment was greater than that seen after metformin treatment (P < 0.05).

Figure 1A presents IRS-1-associated PI 3-kinase activity in muscle from obese type 2 diabetic subjects before and after treatment with troglitazone. The basal activity of PI 3-kinase was not altered by treatment. Insulin stimulated PI 3-kinase activity by 1.5-fold over basal levels (P < 0.01) before treatment. After troglitazone treatment, insulin stimulation was increased nearly threefold. The absolute insulin-stimulated PI 3-kinase activity in muscle after troglitazone treatment was 67% greater than the pretreatment value (P < 0.05) (Fig. 1A).

Metformin treatment tended to increase basal PI 3-kinase activity, but this effect did not reach statistical significance (P < 0.09) (Fig. 1B). Insulin stimulated PI 3-kinase activity 2.6-fold before metformin treatment but unlike troglitazone, metformin therapy resulted in no further increase in IRS-1-associated PI 3-kinase in muscle (Fig. 1B). Insulin stimulated PI 3-kinase activity 1.9-fold after metformin treatment compared with 3-fold after troglitazone treatment. The magnitude of the increase in PI 3-kinase activity after troglitazone treatment was greater than the increase of PI 3-kinase activation with metformin treatment (P < 0.05).

Troglitazone treatment also significantly increased insulin-stimulated Akt activity compared with pretreatment levels (Fig. 2A). Before treatment, 3 h of insulin infusion resulted in a 1.3-fold stimulation of Akt activity. After 3 months of troglitazone therapy, basal Akt activity was unchanged, but insulin now stimulated Akt activity in skeletal muscle by 1.9-fold. Insulin-stimulated Akt activity was increased by 53% (P < 0.05) after treatment with troglitazone compared with pretreatment (Fig. 2A).

Metformin treatment did not alter Akt activity in the basal or insulin-stimulated state (Fig. 2B). The small effect of insulin to stimulate Akt activity before metformin treatment was similar to that in the troglitazone group before treatment (NS). However, in contrast to troglitazone treatment, there was no enhancement of Akt activation in response to insulin after metformin treatment (Fig. 2B).

The effects of treatment on the expression of key insulin-signaling molecules in skeletal muscle were determined by Western blotting (Fig. 3). Figure 3 shows representative results from two subjects. Table 2 displays the quantitation of the blots from each group. Neither troglitazone nor metformin treatment had any effect on the expression of IRS-1, p85, and Akt1/2. However, with troglitazone, but not metformin treatment, the amount of the p110β catalytic subunit of PI 3-kinase was increased by 46 ± 22% compared with pretreatment (P < 0.05) (Table 2).

A growing body of evidence implicates early and intermediate steps in the insulin signaling pathway, including the insulin receptor, IRS-1, and PI 3-kinase as candidates for defects contributing to insulin resistance in skeletal muscle of diabetic rodents and humans (7,8,24,25). The present studies were carried out to determine whether the defect in PI 3-kinase activation observed in obese type 2 diabetic people could be reversed by therapy with the insulin-sensitizing agents troglitazone or metformin. The major finding of this study is that treatment with the TZD troglitazone ameliorates the impairment of insulin-stimulated PI 3-kinase activity and enhances insulin-stimulated Akt activity in skeletal muscle of obese type 2 diabetic subjects. Unlike troglitazone, the biguanide metformin does not have these effects. These data suggest that the insulin-sensitizing action of troglitazone, but not metformin, on whole body glucose disposal in obese type 2 diabetic subjects may be caused, in part, by enhanced PI 3-kinase activation in skeletal muscle.

In the present study, both troglitazone and metformin treatment were efficacious in reducing plasma glucose levels and increasing insulin-stimulated GDR (Table 1), consistent with improved insulin sensitivity. Similar results for whole-body glucose homeostasis have been demonstrated in other studies of type 2 diabetic subjects and obese nondiabetic subjects (26,27), as well as in rodent models of diabetes and insulin resistance (28,29). The antidiabetic actions of troglitazone and metformin appear, however, to be mediated by different mechanisms. The principal action of troglitazone is to increase insulin-mediated peripheral glucose disposal in skeletal muscle, whereas metformin acts primarily by decreasing hepatic glucose output (11,16). Although many recent reports have provided new insight into the mechanism of action of the insulin-sensitizing agents, only a few studies report the effects of TZDs or metformin on insulin action (12,13,21), and there are no data assessing the effects of TZDs or metformin on insulin signaling in skeletal muscle of humans in vivo.

A possible explanation for the enhancement of PI 3-kinase and Akt by troglitazone is that TZD activation of PPAR-γ directly or indirectly stimulates the activity of these two kinases in skeletal muscle through effects on gene expression. In cultured muscle cells from obese type 2 diabetic individuals, troglitazone exerts direct effects, increasing PPAR-γ mRNA and protein levels (30), together with improved insulin-stimulated glucose uptake and glycogen synthase activity (31). In addition, in vitro studies of human adipocytes demonstrate that PPAR-γ activation through TZD directly induces the expression of genes whose products are involved in insulin signaling (32,33). In this study, we found that the amount of the p110β catalytic subunit of PI 3-kinase was increased with troglitazone treatment, suggesting the improvement of PI 3-kinase activity could be caused, at least in part, by increased p110β expression in skeletal muscle. Akt activity but not protein level in muscle is altered by TZD therapy, suggesting that the modulators of Akt activity may be under transcriptional control of PPAR-γ or that the increased PI3-kinase activity enhances insulin-stimulated Akt activity. Further data supporting the direct effect of TZDs on insulin signaling include the fact that TZD treatment of human skeletal muscle cells in vitro induces the activation of PI 3-kinase and Akt (12). Such effects could play an important role in the increased insulin-stimulated glucose uptake and glycogen synthesis that occurs in these TZD-treated cells (12,31).

In vivo and in vitro studies have demonstrated that chronic hyperglycemia can lead to the development of insulin resistance, which results from downregulation of glucose transport as well as alteration of insulin signaling in peripheral tissues (34–37). Restoration of euglycemia improves the impaired insulin stimulation of glucose transport in skeletal muscle of type 2 diabetic subjects (36) and the reduced Akt activity in skeletal muscle of diabetic rats (38). Based on these results, normalization of plasma glucose levels could be expected to lead to an improvement in the insulin-signaling cascade in type 2 diabetic subjects. Interestingly, we found that troglitazone but not metformin treatment significantly enhances insulin action at the level of PI 3-kinase and Akt in skeletal muscle. These differences occurred despite similar improvements in fasting glucose and HbA_1c concentrations with the two agents, thus making it unlikely that a generalized reduction in glycemia was responsible for changes in insulin signaling. Whether the degree of enhancement in insulin activation of PI 3-kinase and Akt in muscle after troglitazone treatment is sufficient to account for the increased whole-body glucose disposal beyond that attained with metformin treatment is not known.

Some of the effects of TZDs on insulin signaling may be through secondary mechanisms. Potential factors are plasma fatty acids and intramuscular lipids because elevations in these parameters are associated with insulin resistance (39,40). As previously reported (41,42), troglitazone treatment tended to lower free fatty acid concentrations (Table 1). TZDs can also reduce accumulation of muscle triglycerides and diacylglycerol (43). Because activation of protein kinase C by elevated diacylglycerol levels in muscle impairs insulin signaling (44), decreases in plasma lipid concentrations with troglitazone treatment could lead to improved insulin signaling by reducing diacylglycerol in muscle. Troglitazone may also reduce intramyocellular lipid content via PPAR-γ-activated redirection of free fatty acids from skeletal muscle to storage in adipocyte triglyceride (45). Thus, the ability of troglitazone to improve insulin action in skeletal muscle could involve multiple actions in both adipose and muscle tissue. Indeed, the increased BMI in our subjects after 3 months of troglitazone treatment may be caused by the potent adipogenic effect of TZDs (46) or the effect of these agents to promote fluid retention with increased plasma volume (47). Despite increased BMI, these subjects have improved insulin action greater than that in metformin-treated subjects.

Although the main mechanism by which metformin treatment in type 2 diabetic subjects ameliorates insulin resistance is reducing hepatic glucose production, metformin also increases insulin-stimulated GDRs into skeletal muscle (48,49). In vitro studies demonstrate that high doses of metformin increase insulin-stimulated glucose transport in isolated skeletal muscle from type 2 diabetic subjects (50). In the current study, despite an increase in the insulin-stimulated glucose disposal rate with metformin therapy, metformin was unable to enhance insulin activation of PI 3-kinase and Akt in skeletal muscle. In addition, other reports have shown that the defect of glycogen synthase activity in obese type 2 diabetic subjects was not restored by metformin treatment (51). It is therefore likely that the molecular mechanism whereby metformin exerts its effects on glycemic control is mediated through other signaling pathways that appear to include activation of AMP kinase (52). This pathway could mediate the effect of metformin to redistribute glucose transporters from an intracellular compartment to the plasma membrane (53).

In summary, this is the first demonstration that the activities of the insulin-stimulated kinases PI 3-kinase and Akt in human skeletal muscle are regulated by therapy with insulin-sensitizing agents. Our data demonstrate that troglitazone treatment improves the impairment of PI 3-kinase activation and enhances Akt activation by insulin in skeletal muscle of obese type 2 diabetic subjects. The change in insulin-stimulated PI 3-kinase activity with troglitazone treatment could be contributed to, in part, by increased expression of p110β. Although troglitazone is no longer available for the treatment of type 2 diabetes, the effects on insulin signaling in skeletal muscle are likely to be shared by the TZD class of compounds. In support of this, our preliminary data show that rosiglitazone treatment also increases insulin-stimulated PI 3-kinase and Akt in some insulin target tissues of insulin-resistant mice with adipose-specific GLUT4 knock out (data not shown). In contrast, treatment of diabetic humans with the biguanide metformin has no effect on PI 3-kinase or Akt activation in muscle. Thus, the insulin-sensitizing effect of the TZD troglitazone in obese type 2 diabetic subjects may be caused, at least in part, by enhancing the insulin-signaling cascade in skeletal muscle.

This work was supported by a grant from the Medical Research Service, Department of Veterans Affairs, VA San Diego HealthCare System (VASDHS); the American Diabetes Association; Pfizer Pharmaceuticals; Parke-Davis; grants from the National Institute of Diabetes and Digestive and Kidney Disease (DK43051 and DK56116); and Grant MO1 RR-00827 from the General Clinical Research Branch, Division of Research Resources, National Institutes of Health.

We thank Debra Armstrong and Leslie Carter at the VASDHS for assistance with the clamp, biopsy, and assay procedures and Dr. M.F. White at the Joslin Diabetes Center for the IRS-1 antibody.