Rosiglitazone Improves Downstream Insulin Receptor Signaling in Type 2 Diabetic Patients

A total of 20 type 2 diabetic patients (11 men and 9 women, age 54 ± 3 years, BMI 30.5 ± 1.1 kg/m², 6 Caucasian and 14 Mexican-American, duration of diabetes 6 ± 1 years, HbA_1c 8.5 ± 0.3%, FPG 185 ± 8 mg/dl) (Table 1) were recruited from the outpatient clinic of the Texas Diabetes Institute. Eleven subjects had been taking a stable dose of sulfonylurea drugs for at least 3 months before the study and remained on the same dose of sulfonylurea throughout the study period. Nine subjects were treated with diet alone. Patients who had previously received insulin, metformin, or a TZD were excluded. Entry criteria included age from 30 to 70 years, BMI <38 kg/m², and an FPG concentration between 140 and 260 mg/dl. All subjects were in good general health without evidence of cardiac, hepatic, renal, or other chronic diseases, as determined by medical history, physical examination, and routine blood chemistries. Subjects were not consuming any medications known to affect glucose metabolism, and none were performing any regular physical activity. In all subjects, body weight was stable for at least 3 months before study. Six healthy individuals (Table 1) served as control subjects. All subjects gave signed voluntary informed consent before their participation. The protocol was approved by the Institutional Review Board of the University of Texas Health Science Center at San Antonio.

Four weeks before the study, all subjects met with a dietitian and were instructed to consume a weight-maintaining diet containing 50% carbohydrate, 30% fat, and 20% protein. During this period, FPG was measured at weekly intervals and varied by <5% in each subject. Blood pressure was measured during each visit, and HbA_1c and fasting plasma lipids were measured twice during this period. During the week before the start of rosiglitazone treatment, all subjects 1) received a 75-g oral glucose tolerance test (OGTT), 2) were measured for fat mass and fat-free mass (FFM) using an intravenous bolus of ³H₂O, and 3) received a euglycemic insulin clamp in combination with [3-³H]glucose and indirect calorimetry to examine peripheral (muscle) tissue sensitivity to insulin. The bolus injection of ³H₂O was performed on the same day as the OGTT. All studies were carried out in the postabsorptive state at 8:00 a.m. after a 10- to 12-h overnight fast. Sulfonylurea-treated subjects did not take their sulfonylurea on the day of the study. After completion of these studies, subjects were started on rosiglitazone (8 mg/day) for 16 weeks. During the rosiglitazone treatment period, subjects returned to the Clinical Research Center of Texas Diabetes Institute at 8:00 a.m. every 2 weeks for measurement of FPG and body weight and to check compliance with the treatment regimen. Fasting plasma lipids (total cholesterol, triglyceride, HDL cholesterol, and LDL cholesterol) and HbA_1c were measured monthly. During the last week of the treatment period, the OGTT, euglycemic insulin clamp, and body fat measurements were repeated.

Baseline blood samples for determination of plasma glucose, FFA, insulin, and C-peptide concentrations were drawn at −30, −15, and 0 min. At time 0, subjects ingested 75 g glucose in 300 ml orange-flavored water, and plasma glucose, FFA, insulin, and C-peptide concentrations were measured at 15-min intervals for 2 h. At time 0, a 100-μCi bolus of ³H₂O was given and the plasma tritiated water radioactivity was determined at 90, 105, and 120 min for calculation of FFM and fat mass as previously described (30).

Subjects returned to the Clinical Research Center of the Texas Diabetes Institute at 8:00 a.m. 3–7 days after the OGTT for the hyperinsulinemic-euglycemic clamp with vastus lateralis muscle biopsies. A catheter was placed in an antecubital vein for the infusion of all test substances. A second catheter was placed retrogradely into a hand vein, and the hand was placed in a heated box (60°C) for sampling of arterialized blood. A primed (25 μCi × FPG/100)-continuous (0.25 μCi/min) infusion of [3-³H]glucose was begun 180 min before the start of insulin infusion in the diabetic subjects (at −120 min in control subjects) to allow for isotopic equilibration. At time −60 min, a percutaneous muscle biopsy was obtained with a Bergstrom cannula from the vastus lateralis muscle under local anesthesia (31). Muscle biopsy specimens were immediately blotted free of blood, frozen in liquid nitrogen, and stored under liquid nitrogen until processing. Baseline arterialized venous blood samples for determination of plasma [3-³H]glucose radioactivity and plasma glucose, FFA, and insulin concentrations were drawn at −30, −20, −10, −5, and 0 min. At time 0 (11:00 a.m.), a prime-continuous infusion of human regular insulin (Novolin; Novo Nordisk Pharmaceuticals, Princeton, NJ) was started at a rate of 80 mU · min⁻¹ · m⁻² body surface area and continued for 4 h to allow insulin to more fully exert its stimulatory effect on tissue glucose disposal. A second percutaneous muscle biopsy was obtained from a site about 4 cm distal to the first site 30 min after the start of insulin infusion (i.e., 90 min after the initial biopsy). In preliminary studies, we have shown that insulin stimulation of insulin receptor and insulin receptor substrate (IRS)-1 tyrosine phosphorylation is maximum at 30 min (4). After initiation of the insulin infusion, the plasma glucose concentration in diabetic subjects was allowed to drop until it reached 100 mg/dl, at which level it was maintained by appropriately adjusting a variable infusion of 20% dextrose. In control subjects, the plasma glucose concentration was clamped at each subject’s basal level. The insulin infusion was continued for 240 min to obtain a measure of insulin-mediated glucose disposal during the last 30 min of the insulin clamp (210- to 240-min time period). Continuous indirect calorimetry was performed with a ventilated hood system (Deltatrac II; SensorMedics, Yorba Linda, CA) during the last 40 min of the basal period and during the last 30 min of insulin infusion. Throughout the insulin clamp, blood samples for determination of plasma glucose concentration were drawn every 5 min. Blood samples for determination of plasma insulin and [3-³H]glucose radioactivity were collected every 10–15 min, and blood samples for the determination of plasma FFA concentration were determined every 60 min.

Plasma glucose concentration was measured at bedside using the glucose oxidase method (Glucose Analyzer 2; Beckman Instruments, Fullerton, CA). Plasma insulin (Diagnostic Products, Los Angeles, CA) and C-peptide (Diagnostic Systems Laboratories, Webster, TX) concentrations were measured by radioimmunoassay. HbA_1c was measured by affinity chromatography (Biochemical Methodology, Drower 4350; Isolab, Akron, OH). Plasma FFA concentration was measured by an enzymatic colorimetric method (Wako Chemicals, Neuss, Germany). Plasma total cholesterol, HDL cholesterol, and triglyceride levels were measured enzymatically (Boehringer Mannheim, Indianapolis, IN) on a Hitachi 704 autoanalyzer. LDL cholesterol was calculated from the Friedewald equation. Tritiated glucose specific activity was determined in deproteinized plasma samples. Insulin receptor and IRS-1 tyrosine phosphorylation and the amount of p85 associated with IRS-1 were assayed using immunoprecipitation and immunoblot analysis, as previously described (4). The association of PI 3-kinase activity with IRS-1 was assayed by determining the ability of anti-IRS-1 immunoprecipitates to incorporate [³²P]ATP into PI (4).

Under steady-state postabsorptive conditions, the rate of endogenous glucose appearance (R_a) is calculated as the [3-³H]glucose infusion rate (dpm/min) divided by the steady-state plasma [3-³H]glucose specific activity (dpm/mg). During the insulin clamp, nonsteady conditions prevailed, and R_a was calculated from Steele’s equation (32). Endogenous glucose production (EGP) was calculated as the R_a minus the exogenous glucose infusion rate. Total-body glucose disposal (TGD) equals the sum of residual EGP plus the exogenous glucose infusion rate. Total-body glucose metabolic clearance rate (MCR) equals the TGD rate divided by the steady-state plasma glucose concentration, where TGD is expressed as milligrams per kilogram FFM per minute and plasma glucose is expressed as milligrams per milliliter. Rates of glucose and lipid oxidation were calculated from oxygen consumption and carbon dioxide production data obtained from indirect calorimetry using formulas described previously (33). Nonoxidative glucose disposal, an index of glycogen formation, was calculated by subtracting the rate of glucose oxidation from the rate of TGD. Total-body water was calculated from the mean of the plasma 3-³H-water radioactivity measurements at 90, 105, and 120 min after the intravenous bolus of ³H₂O. Plasma tritiated water specific activity was calculated assuming that plasma volume represents 93% of total volume. FFM was calculated by dividing total-body water by 0.73 (34).

The area under the glucose, insulin, C-peptide, and FFA curves during the OGTT were determined using the trapezoidal rule. The mean plasma glucose, insulin, C-peptide, and FFA concentrations during the OGTT were calculated by dividing the area under the curve by the duration of the OGTT (120 min). The insulinogenic index was calculated as the increment in plasma insulin concentration divided by the increment in plasma glucose concentration during the 0- to 30- and 0- to 120-min time periods during the OGTT (20).

Statistics were performed with StatView for Windows, version 5.0 (SAS Institute, Cary, NC). Comparison of values before and after insulin infusion were made using the paired Student’s t test. Because rosiglitazone increased insulin-stimulated glucose disposal and IRS-1 tyrosine phosphorylation similarly in the diet- and sulfonylurea-treated groups, all 20 diabetic subjects were analyzed collectively. Comparison of variables during the OGTT and during the insulin clamp between before and after rosiglitazone treatment was made using repeated-measures ANOVA with Bonferroni/Dunn post hoc testing. Comparison of variables between control and diabetic subjects were made using ANOVA. Linear regression analysis was used to examine the relationship between measured variables. All data are presented as means ± SE. A P value <0.05 was considered statistically significant.

After 16 weeks of rosiglitazone treatment, body weight, BMI, and fat mass increased, whereas plasma HbA_1c and FPG, insulin, and FFA concentrations decreased significantly. Rosiglitazone therapy was associated with significant increases in both the fasting plasma HDL and LDL cholesterol concentrations. After rosiglitazone treatment, the mean plasma glucose (P < 0.0001) and FFA (P < 0.0001) concentrations during the OGTT decreased significantly without change in the mean plasma insulin or C-peptide concentrations. The decrement in mean plasma glucose concentration (65 ± 11) during the OGTT was significantly greater (P < 0.01) than the decrement in FPG concentration (46 ± 9), indicating that rosiglitazone had a specific effect to reduce postprandial hyperglycemia. Although there was no difference in the plasma insulin excursion during the OGTT performed before and after rosiglitazone treatment, the insulinogenic index (0–120 min) increased significantly (P = 0.04) after rosiglitazone treatment (Table 1 and Fig. 1).

During the euglycemic insulin clamp studies performed before and after rosiglitazone, the mean steady-state plasma glucose concentrations were 99 ± 1 and 99 ± 1 mg/dl, with coefficients of variation of 3.1 ± 0.5 and 3.0 ± 0.4%, respectively. In all diabetic subjects, the desired glycemic goal (100 mg/dl) was reached within 90 min after the start of insulin infusion. The steady-state plasma insulin concentrations during the euglycemic clamp were similar before and after rosiglitazone treatment (123 ± 9 vs. 114 ± 9 μU/ml, respectively). At all time points during the euglycemic insulin clamp, the suppression of the plasma FFA concentration was significantly greater during the study performed after rosiglitazone treatment (Fig. 2). In control subjects, the steady-state plasma glucose and insulin concentrations during the insulin clamp were 99 ± 1 mg/dl and 103 ± 5 μU/ml, respectively. The suppression of plasma FFAs during the insulin clamp in control subjects was greater than in diabetic subjects before rosiglitazone at all time points.

After 16 weeks of rosiglitazone treatment the basal rate of EGP decreased significantly from 3.5 ± 0.2 to 3.0 ± 0.1 mg · kg FFM⁻¹ · min⁻¹ (P < 0.01), and the decrement in basal EGP was correlated (r = 0.45, P < 0.05) with the decrement in FPG concentration. During the hyperinsulinemic-euglycemic clamp performed before rosiglitazone, EGP was suppressed to 17% (0.6 ± 0.2 mg · kg FFM⁻¹ · min⁻¹). After rosiglitazone treatment, the suppression of EGP during the insulin clamp was more complete: 0.2 ± 0.1 mg · kg FFM⁻¹ · min⁻¹ (P < 0.05 vs. pre-rosiglitazone). In control subjects, the basal EGP was 2.8 ± 0.2 mg · kg FFM⁻¹ · min⁻¹ and declined to 0.2 ± 0.1 mg · kg FFM⁻¹ · min⁻¹ during the insulin clamp (Fig. 3).

After rosiglitazone treatment, both the TGD (6.9 ± 0.7 to 9.2 ± 0.8 mg · kg FFM⁻¹ · min⁻¹, P < 0.001) and MCR of glucose (7.0 ± 0.7 to 9.2 ± 0.8 mg · kg FFM⁻¹ · min⁻¹, P < 0.001) increased by 33 and 31%, respectively, but still remained less than in control subjects (11.1 ± 0.9 mg · kg FFM⁻¹ · min⁻¹ and 11.2 ± 0.9 mg · kg FFM⁻¹ · min⁻¹, respectively). In 2 of the 20 diabetic subjects, insulin-mediated MCR of glucose did not increase, and in these two subjects, there was no decrease in HbA_1c. The increase in TGD in diabetic subjects after rosiglitazone treatment was accounted for by an increase in nonoxidative glucose disposal (NOGD) (3.6 ± 0.7 to 5.6 ± 0.7, P < 0.001). The glucose oxidation rate did not change significantly (3.3 ± 0.2 to 3.6 ± 0.2 mg · kg FFM⁻¹ · min⁻¹, NS). In diabetic subjects treated with rosiglitazone, NOGD still remained less than in control subjects (7.6 ± 0.9 mg · kg FFM⁻¹ · min⁻¹). The increase in insulin-mediated TGD after rosiglitazone treatment was strongly correlated with the decline in postprandial glucose excursion (ΔTGD vs. Δmean plasma glucose during the OGTT, r = −0.74, P = 0.0002).

Basal and insulin-stimulated values for insulin receptor and IRS-1 tyrosine phosphorylation for the amount of p85 associated with IRS-1 and for IRS-1-associated PI 3-kinase activity are shown in Figs. 4 and 5. All values are expressed as the percent of the mean insulin-stimulated value in control subjects. Under postabsorptive conditions, insulin receptor tyrosine phosphorylation tended to be higher in diabetic versus control subjects, consistent with previous observations from our laboratory (4). During the insulin clamp performed before rosiglitazone, hyperinsulinemia did not significantly increase insulin receptor tyrosine phosphorylation, but IRS-1 tyrosine phosphorylation and the amount of p85 and PI 3-kinase activity associated with IRS-1 rose slightly (P < 0.05). Compared with the nondiabetic control group, the insulin-stimulated increases in insulin receptor and IRS-1 tyrosine phosphorylation and the amount of p85 and PI 3-kinase activity associated with IRS-1 were significantly lower (P < 0.01) in diabetic subjects. During the insulin clamp performed after 4 months of rosiglitazone treatment, IRS-1 tyrosine phosphorylation (P < 0.01) and the amount of p85 (P < 0.05) and PI 3-kinase activity (P < 0.05) associated with IRS-1 increased significantly without improvement in insulin-stimulated insulin receptor tyrosine phosphorylation. After rosiglitazone, the incremental response in insulin-stimulated IRS-1 tyrosine phosphorylation and PI 3-kinase activity associated with IRS-1 was similar to the responses in control subjects.

TGD and NOGD during the insulin clamp correlated positively with the insulin-stimulated increment in IRS-1 tyrosine phosphorylation (r = 0.52, P = 0.0006 and r = 0.49, P = 0.001, respectively) and inversely with the plasma FFA concentration (r = −0.50, P = 0.0006 and r = −0.48, P = 0.001, respectively) in control and diabetic subjects before and after rosiglitazone treatment (Fig. 6). TGD and NOGD also correlated with the insulin-stimulated increment in the association of p85 with IRS-1 (r = 0.38, P = 0.01 and r = 0.41, P = 0.006, respectively). Neither the fasting plasma FFA concentration nor the plasma FFA concentration during the insulin clamp correlated with the insulin-stimulated increment in IRS-1 tyrosine phosphorylation or the association of p85 with IRS-1.

In the present study, 16 weeks of rosiglitazone treatment in type 2 diabetic patients improved glycemic control (ΔHbA_1c = 1.4%) by suppressing basal hepatic glucose production, improving insulin-mediated glucose disposal in muscle (Fig. 3), and enhancing β-cell function (insulinogenic index, 0–120 min) (Table 1). The improvement in β-cell function could represent a direct effect of the TZD on the β-cell (35) or could occur secondary to amelioration of glucose toxicity (2,36) or lipotoxicity (2,37,38), i.e., decreased fasting and/or postprandial FFA levels (Table 1 and Fig. 1). The clinical relevance of these physiological findings are underscored by the following: 1) the decline in basal EGP was correlated with the reduction in FPG concentration and 2) the improvement in insulin-mediated TGD (which primarily represents muscle [2]) was strongly correlated with the decline in postprandial glucose excursion (ΔTGD vs. Δmean plasma glucose during the OGTT, r = −0.74, P = 0.0002).

Consistent with the improvement in peripheral tissue (muscle) sensitivity to insulin, 16 weeks of rosiglitazone treatment enhanced insulin-stimulated IRS-1 tyrosine phosphorylation and increased the association of the p85 regulatory subunit of PI 3-kinase, as well as PI 3-kinase activity, with IRS-1 (Figs. 4 and 5). Moreover, insulin-stimulated TGD/NOGD rates correlated well with the insulin-stimulated increments in IRS-1 tyrosine phosphorylation and the association of p85 of PI 3-kinase with IRS-1 before and after rosiglitazone therapy (Fig. 6). The biochemical and molecular mechanism(s) via which rosiglitazone improves muscle sensitivity to insulin have yet to be defined, but some insights are gained from the present results. To initiate their action, TZDs must first bind to PPAR-γ (11–13). PPAR-γ is predominantly expressed in adipose tissue, and its expression in skeletal muscle is low—only ∼10% of the expression observed in adipose tissue (14,39). This tissue distribution raises some questions about how rosiglitazone (Fig. 3) and other TZDs (17,19,20) improve insulin-mediated glucose metabolism in skeletal muscle, the tissue responsible for ∼70–80% of insulin-stimulated glucose disposal.

Recent studies (22–24,40,41) have underscored the original observations of Randle et al. (42) that elevated circulating FFA levels lead to insulin resistance. This may be of particular importance with respect to the TZDs, because numerous studies have demonstrated that treatment of type 2 diabetic patients with this class of drugs causes a 20–30% reduction in fasting plasma FFA concentration and plasma FFA turnover, which is paralleled by a 20–30% reduction in FPG concentration and HbA_1c (17–20,43). The decrease in circulating plasma FFA concentration results both from an inhibition of lipolysis and an increased uptake of FFA in adipose tissue after activation of PPAR-γ (11,20,44). Our results demonstrate that rosiglitazone treatment decreases the fasting and post-OGTT FFA concentrations and improves the suppression of plasma FFA concentration in response to physiological hyperinsulinemia during the euglycemic insulin clamp (Fig. 2). Furthermore, the fasting plasma FFA concentration, as well as the plasma FFA concentration during the OGTT and during the insulin clamp, correlated inversely with peripheral tissue (muscle) insulin sensitivity. However, these FFA concentrations (fasting, post-OGTT, and insulin clamp) were not correlated with insulin-stimulated increments in insulin receptor tyrosine phosphorylation, IRS-1 tyrosine phosphorylation, or the association of p85 or PI 3-kinase activity with IRS-1 after rosiglitazone treatment. These results can be interpreted in several ways. 1) The improvement in insulin signaling after rosiglitazone treatment is not causally related to a decline in plasma FFA concentration and that rosiglitazone exerts parallel but separate effects to reduce plasma FFA levels and to enhance insulin signaling. It could be argued that the small number of PPAR-γ receptors in muscle are sufficient to activate the insulin signal transduction cascade, whereas the stimulatory effect of rosiglitazone on fat cell PPAR-γ is responsible for the decline in plasma FFA concentration. 2) Alternatively, it is possible that the rosiglitazone-induced decline in basal and post-OGTT FFA levels is causally related to the improvement in insulin signaling but that the actual cellular mediator responsible for the increased insulin signaling is a metabolic product of FFA metabolism, such as fatty acyl-CoA (45) or diacylglycerol (26) or some intracellular molecule whose formation is stimulated by FFA or fatty acyl-CoA, such as ceramide (46). In support of this hypothesis is the strong correlation between the reductions in basal/post-OGTT/insulin clamp FFA levels and improved muscle insulin sensitivity during the insulin clamp and improved glucose tolerance/insulin sensitivity index during the OGTT. 3) The improvement in insulin signal transduction could result from the reduction in plasma glucose concentration and amelioration of glucose toxicity (36), or 4) enhanced insulin signaling could reflect both direct (mediated via PPAR-γ in muscle) and indirect (mediated via the reduction in plasma glucose and FFA concentration/intracellular metabolites of FFA) effects of rosiglitazone. Although several studies have found that mRNA levels of PPAR-γ in skeletal muscle are much less than in adipose tissue (14,39), a recent study demonstrated that PPAR-γ expression is similar in skeletal muscle and adipose tissue (47), supporting a possible direct effect of rosiglitazone on skeletal muscle.

Under basal conditions, insulin receptor tyrosine phosphorylation was increased in type 2 diabetic subjects compared with control subjects, and insulin-stimulated (insulin clamp) tyrosine phosphorylation of the insulin receptor was impaired in type 2 diabetic subjects. A high basal and decreased insulin-stimulated tyrosine phosphorylation of the insulin receptor in skeletal muscle has previously been described by us in insulin-resistant type 2 diabetic subjects (4) and in insulin-resistant nondiabetic humans (48). It is noteworthy that, in the present study, rosiglitazone treatment enhanced downstream insulin receptor signaling (IRS-1 tyrosine phosphorylation and amount of p85 and PI 3-kinase activity associated with IRS-1) without any improvement in insulin-stimulated tyrosine phosphorylation of the insulin receptor in diabetic subjects. These results suggest that rosiglitazone might enhance the amplification of a small signal from the insulin receptor to IRS-1 or might have a direct effect on IRS-1 tyrosine phosphorylation. Further studies will be needed to determine the precise mechanism/site via which rosiglitazone improves insulin signal transduction.

Rosiglitazone reduced the mean day-long plasma glucose concentration by decreasing both the fasting and postprandial glucose levels (Fig. 1). Chronic hyperglycemia has been shown to induce insulin resistance in muscle, a phenomenon referred to as glucose toxicity (36). Conversely, a reduction in plasma glucose concentration by sulfonylurea drugs (49) and phlorizin (50) improves insulin sensitivity in muscle. The deleterious effect of hyperglycemia on muscle insulin sensitivity has been shown to be closely related to a downregulation of GLUT4 transport system (50). We believe that amelioration of glucose toxicity is unlikely to explain the rosiglitazone-enhanced insulin-signaling activity observed in the present study, because the improvement in muscle sensitivity to insulin after the reduction in glucose toxicity occurs in the absence of any effect on the insulin signal transduction system. Kim et al. (29) demonstrated that 16 weeks of troglitazone, but not metformin, therapy in type 2 diabetic patients enhanced insulin-stimulated PI 3-kinase activity in skeletal muscle biopsies, although both troglitazone and metformin improved insulin-stimulated TGD and glycemic control similarly in type 2 diabetic patients. These results suggest that TZDs enhance the insulin-signaling cascade in muscle and that this effect cannot be explained by removal of glucose toxicity, because improved insulin signaling was not observed in the metformin-treated group. However, this study (29) cannot distinguish between a direct effect of troglitazone on the insulin-signaling cascade versus an indirect effect mediated via a reduction in plasma FFA levels (25,26). Moreover, the infusion rate of insulin (300 mU · m⁻² · min⁻¹) used by these investigators (29) produced markedly pharmacological elevations in the plasma insulin concentration, making the physiological meaning of these results difficult to interpret. Our results, which used physiological plasma insulin concentrations and a different TZD (rosiglitazone), are consistent with those of Kim et al. (29) and indicate that, throughout the physiological and pharmacological range of plasma insulin concentrations, TZDs improve insulin-stimulated insulin signal transduction and tissue glucose disposal in vivo.

In the present study, 16 weeks of rosiglitazone treatment was associated with a significant increase in insulin-mediated NOGD, which primarily represents glycogen synthesis (51). Glucose oxidation was not enhanced by rosiglitazone treatment. Our results suggest that rosiglitazone improves NOGD via two distinct mechanisms: enhanced insulin-signaling transduction (IRS-1 tyrosine phosphorylation and the association of p85 and PI 3-kinase activity with IRS-1) and the reduction in circulating FFA levels. Both of these metabolic parameters were significantly correlated with the NOGD rate in the present study. A number of studies have demonstrated that high plasma FFA levels inhibit insulin-stimulated glycogen synthesis, in association with decreased glycogen synthase activity (22,23,28,52). Increased fatty acyl-CoAs also have been shown to directly inhibit glycogen synthase activity in vitro (53).

In summary, 16 weeks of rosiglitazone treatment improved glycemic control (ΔHbA_1c = 1.4%) in type 2 diabetic patients by augmenting peripheral (muscle) and hepatic insulin sensitivity and by enhancing β-cell function. The improvement in peripheral (muscle) insulin sensitivity after rosiglitazone treatment is accounted for entirely by an increase in NOGD (glycogen synthesis). Our results suggest that the improvement in insulin-stimulated muscle glucose disposal after rosiglitazone therapy results from two separate effects of TZDs: 1) enhanced insulin-stimulated insulin signal transduction (increased IRS-1 tyrosine phosphorylation and the association of p85/PI 3-kinase activity with IRS-1) and 2) chronic reduction in circulating plasma FFA levels. Alternatively, it is possible that the reduction in plasma FFA and intracellular FFA metabolites is responsible for both the improvements in insulin signal transduction and insulin-mediated glucose disposal and that significant correlations between the reduction in plasma FFA concentration and insulin-signaling events could not be detected because of individual patient variability and because intracellular fatty acyl-CoAs/other FFA metabolites were not measured.

This work was supported in part by National Institutes of Health Grant DK-24092, a VA Merit Award, and a grant from GlaxoSmithKline.

We thank our nurses, Diane Franz and Magda Ortiz, for their assistance in performing the insulin clamp and OGTTs and for the care of the patients throughout the study. Elva G. Chapa and Lorrie Albarado provided expert secretarial assistance in preparation of the manuscript.