Troglitazone Antagonizes Metabolic Effects of Glucocorticoids in Humans

RESEARCH DESIGN AND METHODS

We studied 10 healthy subjects (4 men, 6 women) with normal glucose tolerance. The clinical characteristics of the study group are listed in Table 1. Subjects were metabolically characterized as inpatients at the General Clinical Research Center, Medical University of South Carolina, Charleston, SC, on four sequential occasions: 1) baseline; 2) after dexamethasone at 2 mg b.i.d. per os for 4 days; 3) after the dexamethasone was discontinued, and the subjects were then restudied after troglitazone was given for 4–6 weeks at 400 mg/day (7 of 10 subjects were treated for 4 weeks, 2 for 5 weeks, and 1 for 6 weeks); and 4) after another period of dexamethasone administration (2 mg b.i.d. for 4 days) as troglitazone therapy was continued. Before the study, subjects were equilibrated on a weight maintenance diet (28–32 kcal · kg⁻¹ · day⁻¹) consisting of 50% carbohydrates, 30% fat, and 20% protein. The isocaloric diet was maintained throughout, and subjects were instructed to not make any changes in their usual levels of activity. None of the study subjects engaged in regular exercise. Percent body fat, regional percent body fat, and lean body mass were measured by dual-energy X-ray absorptiometry (DEXA; Lunar Radiation, Madison, WI) upon entry into the study, as previously described (39,40). None of the subjects had cardiovascular, renal, thyroid, or liver diseases and were not ingesting any medications that might affect substrate or energy metabolism. Liver function tests (alkaline phosphatase, total bilirubin, aspartate aminotransferase, and alanine aminotransferase) were checked every 1–2 weeks in all subjects. Protocols were approved by the Medical University of South Carolina institutional review board, and informed consent was obtained from every subject.

Glucose metabolism.

At each of the four study periods, standard 75-g oral glucose tolerance tests (OGTTs) were performed after a 12-h overnight fast. Glucose and insulin were measured at 0, 30, 60, 90, 120, and 180 min. In vivo insulin sensitivity was assessed using the euglycemic-hyperinsulinemic glucose clamp technique as previously described (39,41–43). After a 12-h fast, a catheter was inserted into the brachial vein to administer insulin, glucose, and KPO₄. A dorsal hand vein was cannulated in a retrograde manner and kept in a warming device (65°C) to provide arterialized venous blood for sampling. To maximally stimulate glucose uptake and suppress hepatic glucose production, regular insulin (Humalin; Eli Lilly, Indianapolis, IN) was administered at a rate of 200 mU · m⁻² · min^{− 1}, producing a mean steady-state insulin concentration of 3,480 ± 138 pmol/l, which is maximally effective for stimulating glucose uptake into skeletal muscle (43). Serum glucose was clamped at 5.0 mmol/l for at least 3 h, and maximal glucose uptake for each individual was calculated from the mean glucose infusion rate over the final three 20-min intervals. Whole-body glucose uptake was calculated based on the glucose infusion rate corrected for changes in the glucose pool size, assuming a distribution volume of 19% body weight and a pool fraction of 0.65. Glucose uptake was normalized per kilogram lean body mass (excluding bone mass) determined by DEXA to yield the GDR per kilogram of lean body mass.

The OGTT was performed the day before the hyperinsulinemic-euglycemic clamp in all phases. At study phases 2 and 4, the OGTT was performed after 3 days of dexamethasone, and clamp study was carried out after 4 days of dexamethasone.

FFAs and leptin.

Nonesterified FFA levels were measured using an enzymatic colorimetric assay (NEFA C test kit; Wako Chemicals, Richmond, VA) (44). FFA levels were measured at baseline and 30, 60, 90, 120, and 180 min into the euglycemic-hyperinsulinemic clamp procedure to assess sensitivity of FFAs to suppression by insulin. FFA suppressibility was measured as (baseline level − 30-min level)/baseline level. Serum for leptin determination was obtained from blood drawn at 8:00–9:00 a.m. after an overnight fast. Quantification of serum leptin was determined using a radioimmunoassay and immunopurified rabbit antibodies raised against recombinant human leptin (Linco, St. Louis, MO). Methodology and characteristics of the assay have previously been described in detail (42).

Other assays and statistical analysis.

Plasma glucose was measured by the glucose oxidase method using a glucose analyzer (YSI 2300; Yellow Springs Instruments, Yellow Springs, OH). Serum insulin levels were measured using a microparticle enzyme immunoassay kit (Abbott Diagnostics, Chicago, IL). LDL cholesterol was calculated using the Friedewald equation from the total and HDL cholesterol levels measured by a colorimetric oxidase technique, and triglycerides were measured by a colorimetric glycerophosphate technique, using Vitos autoanalyzers (Johnson and Johnson, Rochester, NY).

Data analysis was performed using Statistica for Windows, version 5.1, 1997 edition (Statsoft, Tulsa, OK) and the SAS program version 6.10 (SAS Institute, Cary, NC). Data are expressed as means ± SE except as indicated in Table 1. Statistical significance was detected using repeated-measures ANOVA, and Tukey’s analysis was used to assess post hoc for group-by-group differences. Differences were accepted as significant at P < 0.05.

RESULTS

Effects on glucose tolerance and insulin sensitivity.

Administration of high-dose dexamethasone (2 mg b.i.d.) for 3–4 days had dramatic effects to diminish both glucose tolerance and insulin sensitivity. In the presence of dexamethasone, fasting and 2-h glucose and insulin levels after a 75-g oral glucose challenge were significantly increased (Table 2), as were the areas under the respective glucose and insulin curves (Figs. 1A and B) when compared with the baseline study. Of the 10 subjects, 4 fulfilled diagnostic criteria for diabetes, and an additional 5 showed evidence of impaired glucose tolerance, while taking dexamethasone. One reason for these effects on glucose tolerance was a significant 34% reduction in maximal insulin-stimulated GDR in dexamethasone-treated individuals compared with baseline values (Fig. 2).

Individuals were treated with troglitazone alone (400 mg/day) for 4–6 weeks, and the metabolic effects of the thiazolidinedione are also shown in Table 2. Insulin and glucose levels after oral glucose challenge did not change significantly from baseline after troglitazone alone. Even so, troglitazone increased sensitivity as manifested by a 20% increase in GDR over baseline (P < 0.001), as depicted in Fig. 2. Subjects gained an average of 1.7 kg during the 4–6 weeks while taking this medication. There were no other untoward effects, and no elevations in hepatic transaminases were noted.

After studies were completed following 4–6 weeks of troglitazone, dexamethasone was reintroduced (2 mg b.i.d. for 4 days) while the thiazolidinedione was continued. When data in subjects treated with the combination of dexamethasone and troglitazone was compared with dexamethasone alone, troglitazone pretreatment was found to dramatically block the deleterious effects of dexamethasone on glucose tolerance (Fig. 1, Table 2). Furthermore, troglitazone completely prevented glucocorticoid-induced insulin resistance, since maximal GDRs in subjects treated with the combination of dexamethasone and troglitazone were similar to baseline values (Fig. 2). However, the mean GDR was not increased above baseline in subjects taking both drugs, as was observed with troglitazone alone.

Effects on FFAs.

No significant changes in fasting FFA levels, triglyceride levels, or lipid panels were detected (Table 2). However, significant effects were observed in the suppression of FFAs during the hyperinsulinemic-euglycemic clamp, an index of insulin’s antilipolytic effect on adipose tissue. FFA suppression by insulin was decreased 22% by dexamethasone and did not change significantly with troglitazone alone, when compared with baseline values. However, troglitazone completely prevented dexamethasone from impairing FFA suppressibility, as shown in Fig. 3.

We found no correlation between fasting FFA levels and either GDR within treatment subgroups or the changes in GDR induced by dexamethasone and troglitazone (data not shown). However, when data were analyzed before and after treatment with dexamethasone, we found that the effect of this drug on glucose uptake was in proportion to its influence on FFA suppressibility (Fig. 4A). In contrast, the ability of troglitazone to increase glucose disposal above baseline was not correlated with FFA suppressibility effects (Fig. 4B). Even so, the improvement in GDR from subjects treated with troglitazone plus dexamethasone over dexamethasone alone was highly correlated with changes in FFA suppressibility (Fig. 4C). Thus, the pattern of alterations in FFA suppressibility was related to the level of glucocorticoid-induced insulin resistance and its reversal by troglitazone, but it was unrelated to the increase in insulin sensitivity resulting from the thiazolidinedione per se.

Effects on leptin.

Fasting leptin levels were increased 2.2-fold by dexamethasone and were not altered by troglitazone alone, when compared with baseline values (Fig. 5). Troglitazone pretreatment, however, blocked the doubling of leptin values by dexamethasone and returned values to baseline. Fasting leptin levels were not correlated with GDR, fasting or 2-h OGTT glucose or insulin levels, or FFA levels (data not shown).

DISCUSSION

We have assessed the effects of the glucocorticoid dexamethasone and the thiazolidinedione troglitazone and for the first time examined interaction between these drugs with respect to multiple metabolic parameters in humans. Our findings provide confirmation of previous studies demonstrating that glucocorticoids induce peripheral insulin resistance (2–4) in healthy humans, and they show a pronounced effect of steroids to decrease maximal insulin responsiveness, which was not clearly evident in a previous report (2). This reflects impaired stimulation of glucose uptake in skeletal muscle, which has been shown to result from defects in the recruitment of GLUT4 glucose transporters to the cell surface in animal (12, 13) and cell (14–17) models. Peripheral insulin resistance, together with insulin resistance at the level of the liver (2–4), leads to impaired glucose tolerance, as manifested in the current data by higher glucose and insulin concentrations after an oral glucose challenge. Glucocorticoids can also increase lipolysis and circulating FFA concentrations, and previous authors have implicated this effect in the pathogenesis of glucocorticoid-induced insulin resistance in skeletal muscle (5,34,35). Our data support this idea, since dexamethasone decreased insulin’s ability to suppress FFA levels, and this effect was well correlated with the reduction in maximally stimulated glucose uptake rates. Although dexamethasone did not alter fasting FFA levels in our subjects, the impairment in FFA suppressibility would predictably result in increased FFA concentrations after meals. Thus, chronic elevations in postprandial FFA profiles could contribute to insulin resistance in skeletal muscle (27–29). A relationship between insulin’s ability to suppress FFA and to stimulate muscle glucose uptake has been previously demonstrated by Reaven and colleagues (45,46) in a sample of healthy individuals.

Troglitazone, the first available antidiabetic medication in the thiazolidinedione class, has been shown to enhance peripheral insulin sensitivity in both genetic and acquired models of insulin resistance in rodents (18,47,48), and in diabetic (19,49, 50) and obese nondiabetic individuals (21). In the current study, troglitazone similarly led to a significant increase in insulin-stimulated glucose uptake rates. However, we did not observe significant alterations in glucose tolerance or post–glucose challenge insulin concentrations compared with baseline. Clinical trials with troglitazone (19,20) and other thiazolidinediones have demonstrated that up to 3 months of treatment is necessary to achieve the full therapeutic effects of the drug. Our subjects were treated for 4–6 weeks with a submaximal dose (400 mg/day as opposed to 600 mg/day), and the majority received 4 weeks of treatment. Thus, it is likely that a longer course of treatment would have resulted in more pronounced effects on insulin sensitivity and a reduction in insulin concentrations during OGTTs (21). In addition, troglitazone did not alter fasting FFA levels or FFA suppressibility in the current study. Therefore, unlike dexamethsone, the effect of troglitazone on insulin sensitivity was not related to any alteration in FFA suppressibility or fasting FFA levels, suggesting that FFAs were not involved in the enhancement of insulin action in muscle. Thus, our data do not support a common notion that troglitazone’s insulin-sensitizing effect on muscle is entirely dependent on an antilipolytic effect in adipose tissue that lowers circulating FFA concentrations. This notion is based on the observations that troglitazone is generally found to decrease circulating FFA concentrations (21) and that FFA infusions induce insulin resistance (27–29). However, troglitazone has been shown to increase muscle insulin action in transgenic mice without adipose tissue (33). Furthermore, our data are in agreement with Maggs et al. (20), who found that troglitazone administration at doses of 200 and 400 mg/day increased insulin sensitivity without any effects on fasting FFA levels. Thus, the current data and these previous results indicate that troglitazone can increase insulin sensitivity independent of an effect in adipose tissue that lowers circulating FFA. By inference, troglitazone could have direct effects to enhance insulin action in muscle.

This study is the first to examine interactions between glucocorticoids and thiazolidinediones in humans. The data demonstrate that troglitazone is able to completely reverse glucocorticoid-induced changes in glucose tolerance, insulin sensitivity, and FFA suppressibility. Dexamethasone led to marked elevations in glucose and insulin concentrations, and these values were returned to baseline levels when individuals were pretreated with troglitazone. Additionally, troglitazone normalized maximally stimulated GDR in subjects receiving dexamethasone; however, the mean value in subjects treated with both drugs was similar to baseline, in contrast to the increase above baseline observed in subjects treated with troglitazone alone. The decline in FFA suppressibility induced by dexamethasone was also completely prevented by troglitazone. Furthermore, the increase in FFA suppressibility in subjects given both drugs compared with dexamethasone alone was correlated with the increment in maximal GDR. Thus, changes in FFA suppressibility are related to both the induction of insulin resistance by dexamethasone and its reversal by troglitazone. The data suggest that although troglitazone can increase insulin sensitivity independent of any effects on FFA as discussed above, its ability to reverse glucocorticoid-induced insulin resistance may be linked to an antilipolytic effect that would lower postprandial FFA levels. There have been several publications in animal models that are in agreement with the current data. First, thiazolidinediones have been shown to ameliorate glucocorticoid-induced insulin resistance in rats (51) and mice (52). Second, Mokuda and Sakamoto (35) have shown that glucocorticoid-induced insulin resistance in rat hindquarter is dependent on FFAs and that troglitazone improves insulin sensitivity by antagonizing these FFA effects.

Interaction between dexamethasone and troglitazone was also observed in the regulation of fasting leptin levels. Dexamethasone increased leptin above baseline, and although troglitazone alone had no effect, the thiazolidinedione was able to completely block the glucocorticoid-induced rise in leptin. Glucocorticoids are known to increase leptin levels (36,54), which has led to speculation that glucocorticoids produce leptin resistance (55). In contrast, troglitazone has been reported to decrease leptin in rats (38, 56); however, the current and previous data (57,58) demonstrate the absence of an effect in humans. The interaction between glucocorticoids and thiazolidinediones is a novel observation with respect to leptin and is analogous to effects of insulin and IGF-1 to antagonize glucocorticoid-induced increments in leptin secretion in cultured rat adipose tissue (59). It is unlikely that modulation of leptin by dexamethasone and troglitazone directly mediates changes in glucose tolerance or insulin sensitivity, because leptin levels showed no correlation with GDR, fasting or 2-h OGTT glucose or insulin levels, or FFA levels.

The mechanism underlying antagonism between thiazolidinediones and glucocorticoids for these multiple metabolic parameters remains unknown. Because both drugs are ligands for nuclear receptors, it is reasonable to speculate that the antagonism probably occurs at the level of regulated gene transcription. Consistent with this hypothesis, glucocorticoids have also been shown to have potent effects on the PPAR system of nuclear receptors (60). Thiazolidinediones have also been found to antagonize other glucocorticoid actions in cultured cells, including regulation of phosphoenolpyruvate carboxykinase expression (61), gene expression in osteoblasts (62), and downregulation of insulin receptor substrate-1 (63). Regardless of the underlying mechanism, the most immediate significance of the current data translates into the clinical management of diabetic patients with endogenous or exogenous hypercortisolism. Control of hyperglycemia in these patients is extremely problematic and cannot be achieved even with very high insulin doses. By reversing glucocorticoid-induced insulin resistance, the use of troglitazone or other thiazolidinediones could allow clinicians to achieve acceptable therapeutic targets for glycemia. This hypothesis is worthy of further study.