The Effect of Rosiglitazone on the Liver: Decreased Gluconeogenesis in Patients with Type 2 Diabetes

Abstract

Aims/Hypothesis: Diabetic hyperglycemia results from insulin resistance of peripheral tissues and glucose overproduction due to increased gluconeogenesis (GNG). Thiazolidinediones have been shown to improve glycemic control and increase peripheral insulin sensitivity. Whether chronic thiazolidinedione treatment is associated with a decrease in GNG has not been determined.

Materials and Methods: We studied 26 diet-treated type 2 diabetic patients randomly assigned to rosiglitazone (RSG; 8 mg/d; n = 13) or placebo (n = 13) for 12 wk. At baseline and 12 wk, we measured endogenous glucose production (by [³H]glucose infusion) and GNG (by the [²H]₂O technique) after a 15-h fast. Peripheral insulin sensitivity was evaluated by a two-step (240 and 960 pmol/min/m⁻²) euglycemic insulin clamp.

Results: Compared with placebo, RSG reduced fasting plasma glucose (9.7 ± 0.7 to 7.4 ± 0.3 mmol/liter; P < 0.001), fasting fractional GNG (−15 ± 4%; P = 0.002), and fasting GNG flux (−3.9 ± 1.2 μmol/min/kg fat-free mass; P = 0.004), with no effect on glycogenolytic flux. Changes in GNG flux and fasting glucose were tightly correlated (r = 0.83; P < 0.0001). During both clamp steps, RSG enhanced insulin-mediated glucose clearance (by 26% and 31%; P = 0.01 and P < 0.02, respectively). In a subgroup of patients studied with magnetic resonance imaging, the reduction in GNG flux was correlated (r = 0.65; P < 0.02) with the reduction in visceral fat area.

Conclusion/Interpretation: RSG increases peripheral tissue insulin sensitivity and decreases endogenous glucose release via an inhibition of gluconeogenesis.

THE ANTIDIABETIC EFFECT of the thiazolidinediones (TZDs) is well established, although the target tissues and the mechanisms underlying the pharmacological activity of these drugs are still incompletely understood. TZDs act by binding to and activating peroxisome proliferator-activated receptors γ (PPARγ) (1), which predominantly are expressed in adipose tissue, but have also been found in muscle, liver, pancreas, heart, and spleen (2–4). Over the past several years, research has focused on the effects of TZDs on skeletal muscle and adipose tissue. The effects of TZDs on fat are well known: adipocyte differentiation (5), fat redistribution (6, 7), decreased lipolysis, and adipogenesis (8). TZD treatment also has a significant effect on muscle to improve its insulin sensitivity (9, 10). Because PPARγ are mainly expressed in adipose tissue, the effect of TZDs on other organs, such as the liver, is still debated and has not been extensively investigated. Although TZDs improve hepatic insulin sensitivity (9, 10), fasting endogenous glucose production (EGP) has been reported to be both reduced (10) and unchanged compared with basal values (9), although the hepatic insulin resistance index [basal EGP × fasting plasma insulin (FPI) concentration] consistently declines. In perfused rat livers as well as isolated hepatocytes, TZDs acutely inhibit the rate of glucose release, mainly as a result of reduced gluconeogenesis (GNG) from lactate (11–14). Recently, it was shown that in rats, TZDs promote the inactivation of liver phosphoenolpyruvate carboxykinase (PEPCK), glucose-6-phosphatase, and pyruvate carboxylase, i.e. key enzymes of the GNG pathway (15). Moreover, TZD treatment significantly reduces nonesterified fatty acid (NEFA) release from adipose tissue (9, 10), and this should provide an inhibitory signal for GNG. Thus, it has been postulated that TZD treatment in vivo should result in a reduction of gluconeogenesis. The goal of this study was to directly measure GNG and glycogenolysis in vivo in type 2 diabetic patients before and after treatment with rosiglitazone (RSG), a potent PPARγ agonist (16) used for the treatment of type 2 diabetes (17).

Subjects and Methods

Subjects

We studied 26 subjects with type 2 diabetes, who were of Mexican-American or non-Hispanic white ethnicity. All subjects were asked about the presence of diabetes in first-degree relatives. The subjects underwent a 75-g oral glucose tolerance test. On a separate day, EGP was measured by the [3-³H]glucose infusion technique, and peripheral insulin sensitivity was determined by the euglycemic hyperinsulinemic clamp (2). The characteristics of the study population are shown in Table 1. Patients who were previously treated with insulin or TZDs were excluded. Patients were in good general health without evidence of cardiac, hepatic, renal, or other chronic diseases, as determined by history, examination, and routine blood chemistry. Patients were not taking any medication known to affect glucose tolerance. No subject was involved in strenuous physical activity, and body weight had been stable (±2 kg) for at least 3 months before the study. All studies were carried out at the Clinical Research Center of University of Texas Health Science Center (San Antonio, TX). The study protocol was approved by the institutional review board of University of Texas Health Science Center, and informed written consent was obtained from each patient before participation. The data on muscle glucose uptake in patients with type 2 diabetes are part of a larger dataset that has been previously published (10) and are used in this study to specifically investigate the relationship between EGP and GNG.

Study design

The design of the study has been described previously (10). Briefly, the study had a double-blind, placebo-controlled, parallel design. In patients who were taking sulfonylureas, the medication was discontinued 6 wk before the study. During wk 7, all subjects received a measurement of lean body mass and fat mass, using an iv bolus of [³H₂]O, and a euglycemic hyperinsulinemic clamp study combined with [³H]glucose infusion (to measure EGP and peripheral tissue sensitivity to insulin) and [²H]₂O ingestion (to measure the separate contributions of gluconeogenesis and glycogenolysis to EGP). After completion of these studies, patients were allocated at random to receive RSG (8 mg/d) or placebo at breakfast every day for 12 wk. Patients returned to the Clinical Research Center every 2 wk for follow-up visits, and during the last week of treatment, all metabolic studies were repeated. In a subgroup of subjects (six patients treated with RSG and six with placebo), quantitation of abdominal sc and visceral fat area at L4–L5 was performed using magnetic resonance imaging (MRI).

Lean and fat masses

On the morning of the study, subjects were admitted to the Clinical Research Center at 0800 h and received an oral glucose tolerance test (75 g) with measurement of plasma glucose, NEFA, and insulin concentrations every 15 min from −30 to 0 min and from 0–120 min. At time zero, a 100-μCi iv bolus of [³H₂]O was given, and plasma [³H₂]O radioactivity was determined 90, 105, and 120 min later. Lean and fat body masses were calculated as described previously (10).

Euglycemic hyperinsulinemic clamp

Subjects were admitted to the Clinical Research Center at 0700 h after an approximately 13-h overnight fast, and a spontaneously voided urine sample was obtained. The subjects were asked not to change their habitual diet regimen, to eat their last meal between 1800–1900 h, and not to eat or drink anything after the last meal. At 2200 h on the evening before the study, all subjects drank [²H₂]O [Isotech, Miamisburg, OH; 5 g/kg fat-free mass (FFM)] in fractionated doses over a period of 2 h. A baseline blood sample was taken in the morning of the day before the study for the determination of [²H₂]O enrichment. On arrival at the Clinical Research Center, a polyethylene cannula was inserted into an antecubital vein for the infusion of all test substances. A second catheter was inserted retrogradely into an ipsilateral wrist vein on the dorsum of the hand for blood sampling, and the hand was kept in a heated box at 65 C. A primed ([20 μCi] × [fasting glycemia/5])-continuous (0.20 μCi/min) infusion of 3-[³H]glucose (DuPont NEN Life Science Products, Boston, MA) was initiated and continued until the end of the study. During the last 30 min of the basal equilibration period (150–180 min), plasma samples were taken at 5- and 10-min intervals for the determination of plasma glucose and insulin concentrations and [³H]glucose specific activity. After the basal equilibration period, insulin was administered as a primed-continuous infusion at the rate of 240 pmol/min/m⁻² for 120 min and then at a rate of 960 pmol/min/m⁻² for another 120 min, as previously described (10). The plasma glucose concentration was measured every 5 min after the start of the insulin infusion, and a variable infusion of 20% glucose was adjusted based on the negative feedback principle to maintain the plasma glucose level at approximately 5 mmol/liter with a coefficient of variation less than 5%. Plasma samples were collected every 15 min from 0–90 min and from 120–210 min and every 5–10 min from 90–120 min and from 210–240 min for the determination of plasma glucose and insulin concentrations and [³H]glucose specific activity. Plasma and urine samples for the determination of gluconeogenesis were taken before starting the [³H]glucose infusion, at the end of the basal tracer equilibration period, and at the end of the first clamp period.

Analytical methods

The plasma glucose concentration was determined by the glucose oxidase method (Beckman II Glucose Analyzer, Beckman Coulter, Fullerton, CA). The plasma insulin concentration was measured by RIA (Diagnostic Products Corp., Los Angeles, CA). The hemoglobin A_1c concentration was measured by affinity chromatography (Biochemical Methodology, Drower 4350, Isolab, Akron, OH). Plasma NEFA were measured spectrophotometrically (Wako Biochemicals, Neuss, Germany). [³H]Glucose specific activity was measured on barium hydroxide/zinc sulfate deproteinized plasma samples (Somogyi’s procedure).

The pattern of ²H incorporation into plasma glucose after [²H₂]O ingestion was determined according to the method developed by Landau and recently modified (18, 19). Briefly, the fraction of glucose produced via GNG from all precursors can be quantified from the ratio of ²H enrichment of carbon 5 (C5) to that of water. ²H enrichment at C5 was obtained by converting glucose to xylose by the removal of carbon in position 6 after purification by HPLC. The C5 group was cleaved by oxidation with periodic acid and formaldehyde, collected by distillation, and incubated with ammonia to form a molecule of hexamethylenetetramine. Enrichment of hexamethylenetetramine obtained from C5 was determined by gas chromatography-mass spectrometry (GCMS) by monitoring peaks of masses 140 and 141. The precision and accuracy of C5 have been reported previously (20).

Water enrichment of the body water pool was monitored by reacting a sample of urine with calcium carbide (CaC₂) to obtain acetylene (C₂H₂), and the enrichment of C₂H₂ was then determined by GCMS by monitoring peaks of masses 26 and 27 (21). All samples were run through the GCMS processing in duplicate or triplicate.

Data analysis

FFM was measured using the [³H₂]O technique (22). Subcutaneous and visceral fat areas were quantitated by MRI at the L4–L5 level, as previously described (23). All glucose fluxes were expressed per kilogram of FFM, because this normalization has been shown to minimize differences due to sex, obesity, and age (24). During the last 30 min of the basal tracer equilibration period, both the plasma glucose concentration and [³H]glucose specific activity were stable in all subjects. Therefore, total EGP was calculated as the ratio of the [³H]glucose infusion rate to the plasma [³H]glucose specific activity (mean of five determinations). During the euglycemic clamp, the total glucose rate of appearance (Ra) was calculated using the Steele equation. EGP was then obtained as the difference between Ra and the exogenous glucose infusion rate. The tracer-determined rate of glucose disappearance provided a measure of insulin-mediated total body glucose disposal.

Because the FPI is a strong inhibitory stimulus for EGP (25), an index of insulin resistance of basal EGP [IR_GP; in units of micromoles per minute per kilogram FFM (kg_ffm) per picomolar concentration] was calculated as the product of fasting EGP and FPI. Experimental validation for the use of this index has been published (26, 27). Peripheral insulin sensitivity was calculated as the mean rate of glucose disappearance during each clamp step divided by the average plasma glucose [i.e. the glucose metabolic clearance rate (MCR)], divided by the average plasma insulin concentration (MCR/I; in units of milliliters per minute per kg_ffm per nanomolar concentration).

The percent contribution of GNG to plasma glucose was calculated as the ratio of the enrichments in C5/[²H₂]O (19). Gluconeogenic flux was calculated by multiplying the percent GNG by EGP. The glycogenolytic flux was obtained as the difference between EGP and the gluconeogenic flux. During the two-step euglycemic insulin clamp, [²H₂]O enrichment is not altered by the infused fluid (28). In contrast, C5 enrichment is diluted by the exogenous glucose. Therefore, the C5/[²H₂]O ratio gives the contribution of GNG to the total (endogenous plus exogenous) concentration of glucose in the plasma. By applying the standard precursor-product relationship, gluconeogenic flux was calculated by multiplying the C5/[²H₂]O ratio by the total Ra at any time point during the clamp (28).

Data are given as the mean ± se. For each measured variable, the effect of treatment was tested by regressing the changes of the variable (over 12 wk) against the baseline value (as a continuous variable) and the group (RSG vs. placebo) and calculating the interaction term between the two independent variables. Placebo-adjusted differences (mean ± se) were calculated with the use of contrasts. Pre- vs. posttreatment significances were calculated using the paired t test. Simple and multiple regression analyses were used to estimate associations among continuous variables in the whole dataset.

Results

At baseline, the RSG- and placebo-treated groups were well matched for age, gender, duration of diabetes, and BMI. After 12 wk of treatment, BMI and fat mass increased in the RSG group (P < 0.05) and remained unchanged in the placebo group (Table 1). Fasting plasma glucose and hemoglobin A_1c decreased in the RSG group, and the decrease was significantly different from that in the placebo group. Both fasting plasma NEFA and insulin concentrations declined in the RSG group, but these changes fell short of statistical significance (Table 1).

Fasting state

At baseline, fractional GNGs from all precursors were similar in the placebo and RSG groups (Table 2). After 12 wk of treatment, fractional GNG decreased significantly with RSG (placebo-adjusted change, −15 ± 4%; P = 0.002), as did total GNG flux (placebo-adjusted change, −3.9 ± 1.2 μmol/min/kg_ffm; P = 0.004), whereas no significant change occurred in glycogenolytic flux. Consequently, EGP was significantly changed by RSG (Fig. 1) and the hepatic insulin resistance index (IR_GP = EGP × FPI) was reduced only in the RSG group (Table 2). The changes in GNG flux were positively correlated with the changes in fasting plasma glucose (Fig. 2). Fasting glucose clearance was significantly stimulated by RSG (placebo-adjusted change, +0.4 ± 0.1 ml/min/kg_ffm; P < 0.001).

In the subgroup of subjects in whom abdominal fat distribution was measured by MRI, RSG treatment was associated with increased sc fat and reduced visceral fat (the latter change was not statistically different from placebo). In the whole dataset, a direct correlation was found between the change in GNG flux and the change in visceral fat (Fig. 3).

Euglycemic clamp

In both treatment groups, the lower insulin infusion rate caused a 60% suppression of EGP and a 30% inhibition of fractional GNG; GNG flux was responsible for virtually all EGP under these conditions (Table 2). RSG treatment was associated with a significant improvement in peripheral insulin sensitivity (as the MCR/I; placebo-adjusted change, +7 ± 3 ml/min/kg_ffm/nm; P = 0.01). At the higher insulin infusion rate, EGP was similarly suppressed (by ∼90%) in the two groups, whereas peripheral insulin sensitivity was significantly improved by RSG (placebo-adjusted change, +4 ± 2 ml/min/kg_ffm/nm; P < 0.02).

In the whole dataset, the decrement in fasting glucose and the increase in fat mass were positively related (r = 0.56; P = 0.002). The changes in GNG flux and peripheral insulin sensitivity also were positively correlated (r = 0.50; P < 0.01). Both changes were directly related to the observed changes in the fasting plasma glucose concentration; in bivariate regression, only the changes in GNG flux were significantly associated with changes in glycemia (r = 0.83; P < 0.0001). In multivariate analysis, the main determinants of the changes in fasting plasma glucose were the changes in gluconeogenic flux (partial r = 0.80; P < 0.001) and changes in kg of fat (partial r = −0.57; P < 0.01), whereas glycogenolysis, glucose disposal, and changes in NEFA and insulin were not significant.

Discussion

Diabetic hyperglycemia results from impaired insulin action on skeletal muscle and excessive release of endogenous glucose due to increased GNG (28). TZDs have been shown to improve both peripheral and hepatic insulin sensitivities (9, 10) and to modulate the expression of the same gene differentially in insulin target tissues. For example, TZDs up-regulate PEPCK in adipose tissue (29), thereby promoting lipogenesis, but down-regulate PEPCK in the liver, thereby inhibiting GNG (15). Recently, it was found that gene regulation by TZDs in the liver involves not only PEPCK, but also two other enzymes involved in the gluconeogenic pathway (15), glucose-6-phosphatase and pyruvate carboxylase. Moreover, it has been shown that in the liver, TZDs can up-regulate the expression of glucokinase and genes involved in lipogenesis (15). Coupled with a reduction in circulating NEFA, TZD treatment in vivo has been postulated to lower plasma glucose levels by decreasing GNG and increasing adipogenesis and glycolysis, but this postulate has not been verified in humans.

In the present placebo-controlled study, RSG treatment significantly decreased fasting GNG as both fractional contribution to EGP and total flux. In particular, GNG returned to values comparable to those in lean, glucose-tolerant subjects (30). Of note, RSG had no effect on glycogenolysis. The RSG-induced decrement in fasting plasma glucose concentration was the combined result of a modest drop in EGP and a 20% increase in fasting glucose clearance. Because fasting plasma insulin concentrations were not appreciably affected in this group of patients, the hepatic insulin resistance index improved with RSG.

During low-dose insulin infusion, EGP was suppressed by 60% on the average; this effect was due to a 30% fall in GNG flux and complete inhibition of glycogenolysis (28, 31). Although RSG did not potentiate this effect of insulin significantly, it potentiated insulin-mediated glucose clearance by 20%. Similar results were observed with high-dose insulin infusion. Thus, under conditions of hyperinsulinemia and euglycemia, RSG improved the insulin sensitivity of glucose uptake, whereas under conditions of fasting hyperglycemia, it improved hepatic insulin sensitivity (i.e. reduction of the hepatic insulin resistance index, EGP × FPI). Although the increases in hepatic and peripheral insulin sensitivities were well correlated, the RSG-induced decrement in GNG flux was the strongest determinant of the drug’s antihyperglycemic effect after the 12 wk of therapy. In a recent study (32), hepatic insulin sensitivity was shown to be improved in type 2 diabetic patients early after the institution of TZD treatment, at a time when body weight, fat distribution, and peripheral insulin sensitivity were still unchanged. Taken together with the current results, these findings document that one of the first actions of TZD is to lower EGP and GNG via an improvement of hepatic insulin sensitivity, whereas TZD-induced skeletal muscle sensitization is a later phenomenon. The mechanisms by which TZDs act on muscle to enhance insulin sensitivity are an area of active investigation (33, 34). The high molecular weight form of adiponectin, an adipocyte-derived cytokine, correlates closely with TZD-induced improvements of peripheral (35) and hepatic (7, 32) insulin sensitivity. The reduction in hepatic fat content after pioglitazone treatment in type 2 diabetics also correlates closely with increased splanchnic (hepatic) glucose uptake after glucose ingestion (36). Thus, adiponectin is a prime insulin-sensitizing candidate that mediates the cross-talk among adipose tissue, muscle, and liver.

It is well known that chronic TZD treatment results in adipocyte differentiation (5), fat redistribution, in particular, from visceral to sc adipose tissue (6, 7), decreased lipolysis (10), and adipogenesis (8). In our patients, RSG treatment was associated with an increase in total fat mass and sc abdominal adipose area that paralleled the improvement in glycemia. Although this is a typical finding with TZD treatment (6, 37), antihyperglycemic efficacy is not strictly conditional on fat mass expansion (38). Rather, weight gain appears to be a distinct pharmacological effect that parallels the improvements in insulin sensitivity and glycemic control. In very hyperglycemic patients, some weight gain may be consequent upon the reduction of glycosuria (39). What does appear to be critical for TZD action is improved insulin sensitivity of adipose tissue (10, 40), resulting in reduced release of NEFA, particularly from the visceral fat depots (41), and increased NEFA uptake by sc adipocytes (33). Excess visceral fat has been associated with increased GNG flux and worse glycemic control in diabetic patients (42), and an increase in systemic NEFA availability has been related to the increase in both GNG and EGP (43). It has been recently shown that with visceral fat accumulation, the release of NEFA and gluconeogenic substrates into the portal circulation is increased (44); according to the portal theory (45), this should lead to an increase in GNG flux. One could therefore speculate that TZD-induced changes in abdominal fat distribution and the flushing out of intrahepatic lipids (36, 46, 47) are key for the drug’s effect on liver metabolism. It should be noted that in the present study the RSG-induced decrease in circulating NEFA did not reach statistical significance (probably due to the relatively small sample size and NEFA variability).

In summary, the liver is a main target of TZD action, and the reduction in GNG flux was the main determinant of RSG’s antihyperglycemic effect after 12 wk of treatment. The efficacy of PPARγ agonists can be attributed to 1) the improvement in hepatic insulin sensitivity during fasting, resulting in decreased hepatic glucose production; 2) the improvement in muscle insulin sensitivity during conditions of hyperinsulinemia, resulting in increased glucose uptake; and 3) improvement of adipose insulin sensitivity, resulting in decreased lipolysis and NEFA release.

Acknowledgments

We thank Magda Ortiz, Dianne Frantz, Socorro Mejorado, Janet Shapiro, John Kinkaid, John King, and Norma Diaz for their assistance in performing the insulin clamp studies, and S. Frascerra, Ph.D.; S. Baldi, Ph.D.; and E. Buzzigoli, B.S., for their technical assistance with the measurement of GNG. We thank L. Albarado and E. Chapa for their secretarial assistance.

This work was supported by National Institutes of Health Grant DK-24092, General Clinical Research Center Grant M01-RR-01346, a Veterans Administration Merit Award, and funds from the Veterans Administration Medical Research Service.

Abbreviations:

 
• C5,

  Carbon 5;

 
• EGP,

  endogenous glucose production;

 
• FFM,

  fat-free mass;

 
• FPI,

  fasting plasma insulin;

 
• GCMS,

  gas chromatography-mass spectrometry;

 
• GNG,

  gluconeogenesis;

 
• IR_GP,

  index of insulin resistance of basal endogenous glucose production;

 
• kg_ffm,

  kilogram of fat-free mass;

 
• MCR,

  glucose metabolic clearance rate;

 
• MRI,

  magnetic resonance imaging;

 
• NEFA,

  nonesterified fatty acid;

 
• PEPCK,

  phosphoenolpyruvate carboxykinase;

 
• PPARγ,

  peroxisome proliferator-activated receptor γ;

 
• Ra,

  glucose rate of appearance;

 
• RSG,

  rosiglitazone;

 
• TZD,

  thiazolidinedione.

References