Introduction

Thiazolidinediones are oral antidiabetic drugs which increase insulin sensitivity, thereby reducing hyperglycaemia in type 2 diabetes. Although much is known about the molecular mechanisms of thiazolidinedione action, questions remain about their physiological targets. Thiazolidinediones bind to and activate the gamma isoform of peroxisome proliferator-activated receptor (PPARγ), which regulates the expression of a variety of genes involved in metabolism in multiple insulin-sensitive tissues [1, 2]. Several lines of evidence have led to the idea that the thiazolidinedione-mediated increase in insulin sensitivity is secondary to the direct effect of this drug on adipose tissue. When administered in vivo to animal models of obesity and diabetes, thiazolidinediones augment fat mass due to an increase in fatty acid storage and a stimulated adipogenesis, as observed by metabolic studies [3, 4]. A concomitant decrease in circulating NEFAs usually occurs [2]. The thiazolidinedione lowering of plasma NEFAs and triacylglycerol has therefore been considered as the major factor responsible for the insulin-sensitising effect of these drugs in rodents [2, 5]. The situation looks more complex in humans in which either a decrease [6, 7], a trend towards a decrease [8] or no effect [9] on plasma NEFAs has been reported depending on the type and concentration of drugs, the duration of treatment and the nutritional background, i.e. fasting or postprandial.

Plasma NEFA levels represent a balance between their release from triacylglycerol stores in adipose tissue and their clearance into tissues that need energy. Since thiazolidinediones seem to have little or no effect on basal rates of lipolysis in rodents and humans [7], the decrease in plasma NEFA is probably due to increased fatty acid clearance, via oxidation and/or esterification. Recent data clearly demonstrated that 2 months of treatment with thiazolidinedione of patients with type 2 diabetes mellitus induced a coordinated upregulation of genes involved in fatty acid uptake, β-oxidation, electron transport and oxidative phosphorylation in adipose tissue [10]. However, it is worth noting that 30% or more of the lipolytic NEFAs are immediately re-esterified into newly synthesised triacylglycerol in adipose tissue in rats and humans [11, 12]. This apparent ‘futile cycle’ of simultaneous lipolysis and re-esterification creates an important mechanism for energy homeostasis [13]. Glycerol-3-phosphate (G3P) is required as a substrate for fatty acid re-esterification into triacylglycerol [14]. Under lipolytic conditions during fasting or stress, G3P cannot arise from glucose transport and metabolism. Glycerol phosphorylation could be involved in G3P production, although glycerol kinase (GYK) activity is very weak in adipose tissue under physiological conditions [15]. Glyceroneogenesis, the pathway described by Hanson, Reshef and their colleagues more than 35 years ago, actually allows G3P synthesis from pyruvate, lactate or certain amino acids [16]. Lactate is the most probable physiological substrate for G3P synthesis because of the large amount of this metabolite in plasma during fasting [17]. This pathway occurs also in liver, in addition to gluconeogenesis [18]. Recent data identified glyceroneogenesis as a thiazolidinedione target in cultured rat adipocytes and adipose tissue, which was the result of induction of its key enzyme, the cytosolic isoform of phosphoenolpyruvate carboxykinase (PEPCK-C) [19]. In cultured explants from human s.c. adipose tissue, the prototype thiazolidinedione rosiglitazone increased the expression of the gene endoing PEPCK-C (PCK1) and glyceroneogenesis [20, 21]. It was also shown that pioglitazone-treated type 2 diabetes patients presented an increase in adipose tissue PCK1 mRNA [22]. However, the respective involvement of various adipocyte proteins involved in lipid metabolism in response to thiazolidinedione in diabetic and non-diabetic humans and rodents remained unclear.

Collectively, the available data led us to postulate that induction of PEPCK-C/glyceroneogenesis is the key mechanism by which thiazolidinediones suppress fatty acid release. This hypothesis needed to be tested in a single coordinated study that addressed each of the known pathways affecting fatty acid release. Thus the aim of this work was: (1) to quantify the effect of thiazolidinediones on the expression of genes related to fatty acid metabolism in a rodent model of obesity with insulin resistance and dyslipidaemia (the Zucker rat) and in s.c. adipose tissue from normal subjects and from type 2 diabetes humans; and (2) to analyse the metabolic consequences of the observed changes. The results of this study support the hypothesis.

Materials, subjects and methods

Materials

DMEM was from Invitrogen (Cercy-Pontoise, France). Rosiglitazone (BRL 49653) was a gift from GlaxoSmithKline (Brentford, Middlesex, UK) and commercially available from Alexis Biochemicals (Coger; Paris, France). All other products were purchased from Sigma (L’Isle d’Abeau Chesnes, France).

Animals and treatments

Male Zucker fatty rats and their lean littermates were purchased from Charles River laboratories (L’Arbresle, France). Rats were kept on a 12-h light–dark cycle at constant room temperature. Conventional laboratory diet and tap water were provided freely until 4 h before the animals were killed, when food was withdrawn. At the time of experiment, eight 8-week-old rats were used per treatment group. Rats were given a dose of 5 mg kg−1 day−1 rosiglitazone (maleate) or vehicle (0.5% methylcellulose) via oral gavage in the morning. After 4 days of treatment, animals were fasted for 4 h, then killed. Plasma was obtained via the jugular vein and aliquots were stored at −20°C for further glucose (Accu-Check; Roche, Meylan, France), insulin (Ultrasensitive Rat Insulin ELISA; EuroBio, Courtaboeuf, France), triacylglycerol (Vitros Chemistry System; Ortho-Clinicals Diagnostics, Raritan, NJ, USA), NEFA (Free Fatty Acids Half Micro Test; Roche) and glycerol (Glycerol UV-Method; R-Biopharm, Saint-Didier au Mont d’Or, France) determinations. Samples of s.c., periepididymal and retroperitoneal adipose tissues were dissected and frozen at −80°C for further analyses. For s.c. samples, a region of the flank was arbitrarily defined, allowing an easy weight comparison from rat to rat. The protocol for the animal studies was conducted according to the French Guidelines for the Care and Use of Experimental Animals.

Diabetic patients

Twenty-four patients with type 2 diabetes, treated with diet alone, participated in the study. Metabolic parameters of the patients have been previously published [8] and are included, with permission, in the Electronic Supplementary Material (ESM) Table 1.

Human adipose tissue explants

Explants of s.c. adipose tissue were obtained from eight women undergoing elective surgery. They were aged 45.7 ± 4.4 years (mean±SEM) and had a BMI of 30.3 ± 1.4 kg/m2. None of the subjects suffered from known metabolic or malignant diseases, nor were they taking medications known to alter adipocyte metabolism. The study was performed according the Declaration of Helsinki. All the patients gave informed written consent.

Adipose tissue specimens from the s.c. region were obtained within 15 min after the onset of surgery. About 5 g of tissue were cut in about 20 mg fragments and cultured in DMEM containing 12.5 mmol/l glucose, 200 IU/ml penicillin, 50 mg/l streptomycin and 10% FCS at 37°C in 10% CO2. After a 1-h preincubation, rosiglitazone was added or not for different times. Explants were frozen before further studies.

RNA analysis

Total RNA was extracted from rat adipose tissues by the method of Chomczynski and Sacchi [23] and from human adipose tissue explants using the Rneasy total RNA kit from Qiagen (Courtaboeuf, France).

Total RNA (1.25 μg) was reverse transcribed using a High Capacity cDNA Archive kit from Applied Biosystems (Courtaboeuf, France). cDNAs were amplified in an ABI prism 7900 HT (Applied Biosystems) using SYBR green, and were analysed with the SDS 2.1 real-time detection system software. Ribosomal 18S RNA (Rnr1) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA were used to normalise cDNA. Quantification of mRNA was carried out by comparison of the number of cycles required to reach reference and target threshold values (δ−δCt method). Sequences of the sense and antisense oligonucleotides corresponding to the different tested genes are given in Table 1.

Table 1 Primer sequences of RNA tested in real-time RT-PCR
Full size table

Enzyme assays

Freshly obtained adipose tissues were homogenised in ice-cold 10 mmol/l Tris-HCl buffer, pH 7.4, containing 250 mmol/l sucrose, 1 mmol/l EDTA and 1 mmol/l dithiothreitol. After centrifugation of the homogenate at 800 × g for 10 min, the supernatant was decanted and centrifuged at 10,000 × g for 15 min to pellet the mitochondrial fraction. This fraction was washed and resuspended in the presence of 0.5% Triton X-100. The cytosolic fraction was obtained by centrifugation at 24,000 × g for 45 min. The relative enrichment in mitochondrial vs cytosolic fraction was monitored using the mitochondrial-specific marker, citrate synthase. Citrate synthase activity in the cytosolic fraction was one-tenth of that obtained in the mitochondrial fraction (2,146 ± 215 nmol min−1 mg−1; n = 11). PEPCK activity was assayed in the cytosolic fraction according to the method of Chang and Lane [24]. Citrate synthase activity was measured by spectrophotometry in the presence of 5,5′-dithiobis-(2-nitro-benzoate) as previously described [25].

Metabolic studies

Adipose tissue (about 200 mg) was cut into pieces of about 20 mg and incubated in 6-well plates in a humidified atmosphere of 10% CO2 at 37°C in 1.5 ml of glucose-free DMEM containing 3% (w/v) fatty acid-free BSA and 5 mmol/l pyruvate. [1-14C]pyruvate (2 × 104 Bq/ml) was used at an isotopic dilution of 1/250. After 2 h, fragments were rinsed and frozen in liquid nitrogen before lipid extraction, according the simplified method of Bligh and Dyer [26]. The subsequent [1-14C]pyruvate incorporation was estimated by counting the radioactivity associated with the fragments.

Protein analysis

Protein concentrations were determined using the Bradford method with BSA as standard [27].

Image quantification and data analysis

Quantitative results were obtained by densitometry in an image analysis system. The non-parametric Mann–Whitney U test for pairwise comparisons was applied due to the small number of experiments. Analyses were performed using the StatView 4.01 Non-FPU (Abacus Concepts, 1992–1993; Berkeley, CA, USA) statistical package. A value of p < 0.05 was considered statistically significant.

Results

Effect of rosiglitazone administration to fa/fa rats on gene expression in adipose tissues

The obese (fa/fa) Zucker rat is frequently used as a model of the insulin-resistant or prediabetic state. These rats exhibit impaired glucose tolerance, hyperphagia, hyperinsulinaemia and hyperlipidaemia associated with an increase in body and adipose tissue weight (Table 2). Although not overly hyperglycaemic, the fa/fa rat exhibits an abnormal glucose profile after feeding when compared with its lean counterpart, which is associated with a small but consistent and significantly elevated glycated haemoglobin level.

Table 2 Effect of rosiglitazone treatment on metabolic parameters
Full size table

In this model, we explored the expression of two genes involved in G3P synthesis, Pck1 and Gyk, in s.c., periepididymal and retroperitoneal adipose tissues and compared their expression with that of other adipocyte genes, such as Fabp4 coding for the adipocyte lipid-binding protein (AP2) and Lpl coding for lipoprotein lipase (LPL). As previously observed [28], Lpl and Fabp4 mRNAs were induced about 3.5- and 7.0-fold, respectively, in s.c. adipose tissue from Zucker fatty rats, when compared with their lean counterparts (Fig. 1a). Pck1 expression was also induced about 3.5-fold while that of Gyk showed a 60% decrease and that of Pck2 coding for mitochondrial PEPCK (PEPCK-M) was not modified. In the other two adipose tissue depots, no change was observed in the expression of these different genes (Fig. 1b,c).

Fig. 1
figure 1

Effect of rosiglitazone (Rosi) treatment on gene expression in adipose tissues. Subcutaneous (a), periepididymal (b) and retroperitoneal (c) adipose tissues were collected from rats treated with or without rosiglitazone (5 mg kg−1 day−1) for 4 days. Pck1 (black bars), Pck2 (white bars), Gyk (hatched bars), Fabp4 (grey bars) and Lpl (vertical hatched bars) mRNA concentrations were analysed by real-time RT-PCR. Data were normalised to Rnr1 rRNA (18S) and expressed as a percent of the value obtained with lean rats. Data points in triplicate are means±SEM from four rats in each group. **p < 0.01 vs Fa/? rats; ***p < 0.001 vs Fa/? rats; §p < 0.05 vs fa/fa rats; §§p < 0.01 vs fa/fa rats; §§§p < 0.001 vs fa/fa rats

Full size image

Treating fatty Zucker male rats with rosiglitazone for 4 days significantly lowered the elevated serum lipid levels (Table 2). As expected, serum NEFA levels decreased from 192 ± 29 to 119 ± 35 μmol/l after treatment (p < 0.05) and serum triacylglycerols decreased from 1,350 ± 161 to 880 ± 91 μmol/l (p < 0.05). Rosiglitazone did not have any impact on weight gain in these 4-day-treated rats. In s.c. adipose tissue from fatty Zucker rats, rosiglitazone treatment resulted in a coordinated increase of about twofold in Pck1, Gyk and Fabp4 mRNAs and a trend towards an increase for Lpl (Fig. 1a). For periepididymal and retroperitoneal adipose tissue, expression of only Pck1 and Gyk either increased or showed a trend towards an increase (Fig. 1b,c). Altogether these results show that Pck1 is clearly the most constantly responsive to rosiglitazone whatever the adipose tissue localisation. In contrast, expression of the liver Pck1 remained non-significantly altered by rosiglitazone treatment when compared with basal values obtained with untreated obese rats (147 ± 14 vs 107 ± 21% of the lean rat level).

We wondered whether the rosiglitazone-induced increases in Pck1 and Gyk mRNAs were linked to variations in NEFA and glycerol release from incubated s.c. adipose tissue. As shown in Fig. 2a, glycerol release remained unchanged while NEFA output was drastically reduced by twofold, showing that glyceroneogenesis was stimulated, whereas glycerol phosphorylation was not. We thus decided to focus on glyceroneogenesis and monitored incorporation of radiolabelled pyruvate into triacylglycerols. [1-14C]pyruvate was chosen because, in contrast to C2- or C3-labelled molecules, the C1 carbon of pyruvate is conserved in G3P synthesis and is thus a specific marker of glyceroneogenesis [29]. Rosiglitazone indeed induced pyruvate incorporation into triacylglycerol from 42.4 ± 6.4 to 104.4 ± 9.7 nmol h−1 g−1 tissue (Fig. 2b). Results very similar to those shown with s.c. adipose tissue were obtained with periepididymal and retroperitoneal adipose tissue (not shown). To determine whether the observed reduction in NEFA output was related to variations in PEPCK-C activity, we monitored its activity in the various localisations of adipose tissue and in response to rosiglitazone treatment. PEPCK-C activities were similar in the s.c., periepididymal and retroperitoneal adipose tissue and were increased 1.5- to 2-fold by rosiglitazone in Zucker fatty rats (Fig. 2c).

Fig. 2
figure 2

Effect of rosiglitazone on metabolic parameters and on PEPCK-C activity in obese Zucker rat (fa/fa) adipose tissues. Glycerol and NEFA release in the incubation medium (a) and glyceroneogenic activity (b) were monitored from s.c. adipose tissue (AT) from control (black bars) and rosiglitazone-treated (white bars) fa/fa rats. (c) PEPCK activity was measured in s.c. (black bars), periepididymal (white bars) and retroperitoneal (hatched bars) adipose tissues from control and rosiglitazone-treated fa/fa rats. Results are means±SEM from eight rats in each group. *p < 0.05 vs fa/fa rats; **p < 0.01 vs fa/fa rats

Full size image

Effect of rosiglitazone on PCK1 expression in type 2 diabetes patients

In order to explore whether PCK1 responded to rosiglitazone in insulin-resistant humans as in rodents, we used the previously described cohort of type 2 diabetes patients in whom metabolic studies were conducted [8]. In this double-blind, placebo-controlled crossover study the effect of rosiglitazone was evaluated. The expression of PCK1 mRNA showed a strong and significant 2.5-fold upregulation (150% above placebo) following the 12-week thiazolidinedione treatment (Fig. 3). In contrast, PCK2, soluble glycerol-3-phosphate dehydrogenase (GPD1), pyruvate kinase dehydrogenase, isoenzyme 4 (PDK4), retinoid X receptor α (RXRA) and cyclophilin mRNAs remained unaffected. Among the genes described above in our rodent model, only FABP4 was previously shown to be upregulated in the same cohort [8].

Fig. 3
figure 3

Effect of rosiglitazone treatment on gene expression in s.c. adipose tissue from type 2 diabetic patients. Effect of the 12-week treatment with rosiglitazone or placebo on the indicated mRNAs. Data were normalised to cyclophilin (CYCLO) mRNA and expressed as a percent of variations from placebo. Results are means ± SEM obtained from the 24 subjects. ***p < 0.001 vs placebo

Full size image

Rosiglitazone regulation of gene expression in ex vivo cultured human adipose tissue explants

To explore further the regulation of these genes in humans, we cultured s.c. adipose tissue explants obtained from healthy overweight women having elective plastic surgery. Since culture of adipose tissue is preceded by several steps (surgery, transfer and dissection) which could alter adipocyte gene expression [30], we verified the expression of TNF and VEGFA coding, respectively, for tumour necrosis factor alpha and vascular endothelial growth factor, which are markers for hypoxia and inflammation, respectively, in our experimental conditions. Both genes remained stably expressed for as long as 3 days in culture, as was GAPDH (data not shown).

Treatment of explants with 1 μmol/l rosiglitazone caused a progressive increase in the abundance of the PCK1 transcript that reached a maximum of 9-fold at 14 h and decreased thereafter (Fig. 4a). A late and modest twofold increase was observed for FABP4 mRNA at 24 h, while the expression of PCK2, GYK and LPL remained unchanged whatever the time of treatment (Fig. 4a). To examine a possible effect of rosiglitazone on PCK1 mRNA stability, we first treated explants with rosiglitazone for 5 h, then with the transcription inhibitor 5,6-dichloro-1β-d-ribofuranosyl benzimidazole (DRB) in the presence or absence of rosiglitazone and monitored PCK1 mRNA concentration over time. No change in PCK1 mRNA half-life was observed under these conditions (Fig. 4b). Calculated half-life was about 5 h in these conditions, in agreement with what had already been observed in 3T3-F442A adipocytes [31]. While DRB had only a mild effect on basal PCK1 mRNA, it inhibited rosiglitazone induction (Fig. 4c). Besides, the protein synthesis inhibitor puromycin affected neither basal nor rosiglitazone-stimulated PCK1 mRNA, showing that this gene was a primary target for the drug (Fig. 4c). Altogether these results strongly suggested that, as expected from what was already observed in the 3T3-F442A rodent adipocyte cell line, rosiglitazone directly induced PCK1 transcription in human adipose tissue [32].

Fig. 4
figure 4

Effect of rosiglitazone (Rosi) on gene expression in human s.c. adipose tissue. a PCK1 (black bars), PCK2 (white bars), GYK (hatched bars), GPD1 (grey bars), FABP4 (vertical hatched bars) and LPL (dotted bars) mRNAs from human s.c. explants were analysed by real-time RT-PCR. b Human adipose tissue explants were first treated (white triangles) or not (black triangles) for 5 h with rosiglitazone at 1 μmol/l. The horizontal and vertical bars indicate how half-life was evaluated. Then DRB (25 μg/ml) was added and RNA isolated as indicated. c Human adipose tissue explants were treated for 5 h with rosiglitazone at 1 μmol/l in the absence or presence of DRB (25 μg/ml) or puromycin (10−5 mol/l). cont, control. Transcripts were analysed in triplicate by real-time RT-PCR. Throughout data were normalised to GAPDH values and expressed as a per cent of control. Data points are means ± SEM from six individuals at 5 and 24 h and four individuals at 14 h. *p < 0.05 vs control; ***p < 0.001 vs control

Full size image

Additionally, PEPCK-C protein, PEPCK activity and incorporation of [1-14C]pyruvate into lipids were significantly increased about 95, 60 and 50%, respectively, at 18 h of rosiglitazone treatment (ESM Fig. 1).

Discussion

Among the pathways that thiazolidinediones alter in adipose tissue, fatty acid re-esterification has been described as an important one. Thiazolidinediones induce G3P synthesis through glyceroneogenesis and glycerol phosphorylation [33]. We showed that glyceroneogenesis was the main contributor pathway for fatty acid re-esterification in rat adipose tissue in both basal and thiazolidinedione-stimulated conditions [19]. This result was reinforced by elegant in vivo experiments from Chen et al. [34], in which mass isotopomer analyses in mice confirmed the major contribution of glyceroneogenesis to thiazolidinedione-induced triacylglycerol synthesis, with a very minor role of glycerol phosphorylation. One important observation of our present work is that the administration of rosiglitazone to insulin-resistant rats induced an increase in glyceroneogenesis, which was accompanied by an improvement in dyslipidaemia, as shown by the large decrease in plasma triacylglycerol. Similar observations were previously reported in humans, in whom thiazolidinediones induced a reduction in postprandial plasma triacylglycerol [8, 35]. As expected, a strong correlation was also shown between the glyceroneogenic flux of the tissue and the release of NEFA in the incubation medium. As a consequence, we observed a large decrease in plasma NEFA concentration in agreement with the role that glyceroneogenesis is playing in G3P synthesis required for fatty acid esterification after a high-fat or high-protein carbohydrate-free diet and for re-esterification during fasting [36, 37]. According to the minor role of glycerol phosphorylation in this process, we could not detect any variation in glycerol release whether rats were treated with rosiglitazone or not, in agreement with our previous studies, in which either a very mild decrease or no variation in glycerol release was observed [18]. We can thus easily attribute the improvement in plasma dyslipidaemia to the early induction of fatty acid re-esterification linked to glyceroneogenesis in response to thiazolidinediones.

Indeed, PCK1 mRNA was strongly induced in s.c., periepididymal and retroperitoneal adipose tissues at 4 days of rosiglitazone treatment, while expression of the gene encoding the mitochondrial isoform of PEPCK, PEPCK-M, remained unaffected. Expression of the Fabp4 and Gyk genes was also augmented. Similar results were obtained previously in rodent adipose tissue [38]. These changes in gene expression are reminiscent of what was shown in 3T3-F442A adipocytes, in which, however, Pck1 is by far the earliest and directly responsive gene [32].

As a consequence of these rapid and profound changes in adipose tissue gene expression and metabolism, blood NEFA concentration decreased. This decrease would lead to a diminished fatty acid flux towards the liver, resulting in a reduction in triacylglycerol synthesis and export to the blood as VLDL. Indeed, fatty acids are taken up and used by the liver in proportion to their delivery rate and hepatic VLDL synthesis and secretion is directly linked to substrate availability [39]. We noticed the absence of a rosiglitazone effect on liver PEPCK-C, in agreement with the results from Way et al. [2], who showed that a 24-h treatment of Zucker diabetic fatty rats with a PPARγ agonist had no effect on liver PCK1 mRNA, whereas at 7 days of exposure an 80% decrease was observed [2]. These observations agree with the expected glucose-lowering role of thiazolidinediones, which would increase adipose tissue re-esterification in the short term but would later on decrease gluconeogenesis.

The second important observation from our present study is that rosiglitazone administration to type 2 diabetes patients for 12 weeks induced PCK1 mRNA, but not PCK2 mRNA, in s.c. adipose tissue at a much higher incremental level than that of any other gene tested in the same cohort [8, 40, 41]. Similar results were obtained by Bogacka et al. in a cohort of pioglitazone-treated type 2 diabetes patients [22]. However, our data showed no effect of the drug on GPD1 mRNA, in contrast to the 30% increase observed in the Bogacka study. These results reinforced our view that PCK1 is the major thiazolidinedione target gene in human adipose tissue, as in rodents. Indeed, there was little or no induction in GYK expression in adipose tissue in either cohort [22, 42].

The third important observation from our study is that in adipose tissue explants from humans only PCK1, and to a lesser extent FABP4, were acutely induced by rosiglitazone. Guan et al. showed an induction of GYK mRNA in primary human adipocytes [38]. However, in the study of Guan et al., primary cultures were treated for 10 days with ciglitazone, which could correspond to a differentiation effect due to the long-term treatment. The situation is quite different from what we show in adipose tissue explants treated with rosiglitazone for 24 h at most and in which a differentiation effect of the drug is unlikely to occur. Previous data from our laboratory showed that adipose tissue explants, from lean (BMI < 27 kg/m2) but not overweight subjects, exposed for 3 days to rosiglitazone resulted in an increase in GYK activity [21]. These data are in agreement with the present study in which rosiglitazone action was monitored in short-term treated (5–24 h) adipose tissue explants from overweight subjects (mean BMI = 30.3 kg/m2) of a different cohort.

The demonstration that PPARγ ligands acutely regulate PCK1 expression in human adipose tissue agrees with what has been shown earlier by our laboratory and others in rodent adipocytes or adipose tissue [32, 43, 44]. We show additionally here that PCK1 transcription is the affected step in human adipose tissue and that no post-transcriptionnal effect occurs in response to the thiazolidinedione. The demonstration that rosiglitazone action does not require protein synthesis strongly suggests that the whole machinery for stimulating transcription of this gene is present and functional in human adipocytes. This does not appear to be the case for the GYK which, at least in 3T3-L1 adipocytes, requires complex events to be activated at the transcriptional level, thus explaining the delayed thiazolidinedione induction [45]. Therefore, it is not surprising that the PCK1 and the associated metabolic pathway, i.e. glyceroneogenesis, are acute rosiglitazone targets. We propose that this pathway is the initially affected step in response to thiazolidinedione and that the early occurring stimulation of fatty acid re-esterification is the major effect of this type of drug in human adipose tissue, as in rodents.