| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
and -
Agonists on 11ß-Hydroxysteroid Dehydrogenase Type 1 in Subcutaneous Adipose Tissue in MenEndocrinology Unit (D.J.W., R.H.S., N.Z.M.H., R.A., B.R.W.), Centre for Cardiovascular Science, University of Edinburgh, Queen’s Medical Research Institute, Edinburgh EH16 4TJ, Scotland, United Kingdom; and Oxford Centre for Diabetes, Endocrinology, and Metabolism (G.D.T., F.K.), University of Oxford, Churchill Hospital, Oxford OX3 7LJ, United Kingdom
Address all correspondence and requests for reprints to: Prof. Brian R. Walker, University of Edinburgh, Endocrinology Unit, Centre for Cardiovascular Science, Queen’s Medical Research Institute, 47 Little France Crescent, Edinburgh EH16 4TJ, Scotland, United Kingdom. E-mail: B.Walker{at}ed.ac.uk.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
(PPAR) and PPAR
agonists down-regulate 11ß-hydroxysteroid dehydrogenase type 1 (11ß-HSD1) mRNA and activity in liver and adipose tissue, respectively, and PPAR agonists reduce ACTH secretion from corticotrope cells.
Objective: Our objective was to test whether PPAR agonists alter cortisol secretion and peripheral regeneration by 11ß-HSD1 in humans and whether reduced cortisol action contributes to metabolic effects of PPAR agonists.
Design and Setting: Three randomized placebo-controlled crossover studies were conducted at a clinical research facility.
Patients and Participants: Healthy men and patients with type 2 diabetes participated.
Interventions, Outcome Measures, and Results: In nine healthy men, 7 d of PPAR agonist (fenofibrate) or PPAR agonist (rosiglitazone) had no effect on cortisol secretion, hepatic cortisol generation after oral cortisone administration, or tracer kinetics during 9,11,12,12-[2H]4-cortisol infusion, although rosiglitazone marginally reduced cortisol generation in sc adipose tissue measured by in vivo microdialysis. In 12 healthy men, 4–5 wk of rosiglitazone increased insulin sensitivity during insulin infusion but did not change 11ß-HSD1 mRNA or activity in sc adipose tissue, and insulin sensitization was unaffected by glucocorticoid blockade with a combination of metyrapone and RU38486. In 12 men with type 2 diabetes 12 wk of rosiglitazone reduced arteriovenous cortisone extraction across abdominal sc adipose tissue and reduced 11ß-HSD1 mRNA in sc adipose tissue but increased plasma cortisol concentrations.
Conclusions: Neither PPAR nor PPAR agonists down-regulate 11ß-HSD1 or cortisol secretion acutely in humans. The early insulin-sensitizing effect of rosiglitazone is not dependent on reducing intracellular glucocorticoid concentrations. Reduced adipose 11ß-HSD1 expression and increased plasma cortisol during longer therapy with rosiglitazone probably reflect indirect effects, e.g. mediated by changes in body fat.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
agonists) and thiazolidinediones (PPAR agonists) (1, 2). PPAR agonists act principally in the liver and alter lipid metabolism and lipoprotein turnover. PPAR agonists act principally in the adipose tissue and alter adipokine secretion, increase insulin sensitivity, promote adipocyte differentiation, and alter body fat distribution in favor of sc rather than visceral fat accumulation. In several respects, the actions of PPAR agonists are the inverse of those of glucocorticoids. In excess, glucocorticoids inhibit fatty acid ß-oxidation, promote an atherogenic lipid profile, decrease insulin sensitivity, and alter fat distribution in favor of visceral fat accumulation (3). Recent data suggest that the effects of PPAR agonists might be mediated, at least in part, by a reduction in glucocorticoid receptor (GR) activation.
The availability of ligand for GR is determined not only by the circulating concentration of glucocorticoids but also by the rate of local intracellular generation of glucocorticoid by the enzyme 11ß-hydroxysteroid dehydrogenase type 1 (11ß-HSD1) (4). Administration of PPAR agonists, including fenofibrate, for 5–7 d down-regulated 11ß-HSD1 mRNA levels and activity in the liver of both mice and hamsters by about 50–80% (5). PPAR agonists, including rosiglitazone, reduced 11ß-HSD1 mRNA levels and activity in murine 3T3-L1 adipocyte-derived cells, in epididymal adipose tissue of db/db mice after 11 d of in vivo administration (6), and in adipose tissue in rats after 3 wk of in vivo administration (7). Reducing 11ß-HSD1 activity by a similar magnitude has been achieved with selective enzyme inhibitors and results in enhanced insulin sensitivity, weight loss, and an atheroprotective lipid profile in several mouse models of diabetes and obesity (8, 9). Moreover, 11ß-HSD1 knockout mice have improved insulin sensitivity and an atheroprotective lipid profile and accumulate less fat in visceral adipose tissue on a high-fat diet (10, 11). However, the extent to which a reduction in 11ß-HSD1 activity, and thereby in intracellular glucocorticoid concentrations, accounts for the beneficial metabolic effects of PPAR and/or PPAR agonists is unknown. In one study, rosiglitazone had no effect on liver or adipose 11ß-HSD1 activity in Zucker rats (12), although it did alter other pathways of intrahepatic glucocorticoid metabolism (13). In another study, improvements in insulin sensitivity and serum triglycerides induced by rosiglitazone were similar in rats that had been adrenalectomized or sham-operated (14), suggesting that the availability of glucocorticoids does not determine these effects of PPAR agonists.
Regulation of 11ß-HSD1 transcription has been studied extensively in many species (reviewed in Ref. 4). Notably, regulation is tissue specific and species specific. Only one report has examined whether PPAR agonists down-regulate 11ß-HSD1 in humans and found no effect, although this was restricted to measurement of adipose tissue mRNA after 12 wk of pioglitazone treatment (15). Because 11ß-HSD1 expression is increased in adipose tissue in obese humans (16, 17), and 11ß-HSD1 in obese mice is more sensitive to PPAR agonists (6), it is possible that obese subjects might respond differently from lean subjects.
In addition, PPAR is expressed in pituitary corticotrope cells (18) and thiazolidinediones have been reported to lower ACTH and cortisol secretion in patients with Cushing’s disease (19). Effects of PPAR agonists on the physiological control of cortisol secretion have not been investigated.
Here, we report experiments in lean and obese men that investigate the acute and longer-term effects of PPAR agonists on cortisol secretion and 11ß-HSD1 activity in liver and sc adipose tissue, quantify the effects in vivo, and test whether insulin-sensitizing effects of PPAR agonists depend upon changing glucocorticoid action.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Objective. This experiment was conducted after 7 d drug administration to study effects of PPAR agonists over a similar time course as was employed in animals (5, 6) and to examine direct acute effects of PPAR activation rather than indirect longer-term effects (mediated for example by subsequent body fat redistribution).
Participants. Nine male healthy volunteers were recruited by advertisement. Inclusion criteria were age 20–70 yr; body mass index between 20 and 35 kg/m2; normal thyroid, renal, and hepatic function on biochemical screening; alcohol intake less than 28 U per week; no chronic disease (including dyslipidemia or glucose intolerance); and no chronic medication or glucocorticoid therapy in the previous 6 months. Local ethical approval and written informed consent were obtained.
Study design and protocol. A three-phase double-blind randomized balanced crossover study compared 7-d treatments with oral rosiglitazone (Avandia, 4 mg daily), fenofibrate (Lipantil Micro, 200 mg daily), and placebo (one daily tablet). Phases were separated by a 2-wk washout. At the end of each treatment phase, volunteers attended on 2 consecutive days for the following measurements.
A 24-h urine collection was completed for measurement of cortisol metabolites by gas chromatography/mass spectrometry. After overnight fast, subjects attended the Clinical Research Facility at 0800 h. An antecubital vein was cannulated in each arm for infusion and sampling. After cleaning with Betadine (SSL, Knutsford, UK) and injection of local anesthetic (5 ml lignocaine 1%; Braun, Sheffield, UK), a microdialysis cannula (with a 20-kDa-permeable membrane; CMA Microdialysis, Solna, Sweden) was inserted in abdominal sc fat, approximately 10 cm lateral to the umbilicus on each side, as previously described (20, 21). At time zero (
0830 h) a primed iv infusion of cortisol [20% 9,11,12,12-[2H]4-cortisol (D4-cortisol; Cambridge Isotopes, Philadelphia, PA) and 80% hydrocortisone-21-succinate] was commenced (3.6 mg priming bolus and then continuous 1.74 mg/h infusion) (20, 21, 22, 23, 24). At the same time, an intra-adipose infusion of 1,2-[3H]2-cortisone (67 nmol/liter; Amersham, Poole, UK) was commenced at a rate of 0.3 µl/min (20, 21). Volunteers were given 200 ml of water to drink every hour to encourage regular bladder emptying. Blood samples were obtained for steroid analysis at intervals indicated in the figures, and plasma was separated promptly and stored at –80 C. Urine was collected at baseline and then every hour for analysis of tracer steroid metabolism, the volume recorded, and aliquots stored at –20 C. The effluent from the microdialysis cannula was collected in microvials that were changed every hour, and dialysate was stored at –80 C. Infusions and sampling continued for 300 min.
At 2300 h, subjects ingested a 1-mg dexamethasone tablet (Organon, Cambridge, UK). After overnight fast, they attended at 0800 h the following morning. An iv cannula was sited for blood sampling, and they were administered a 25-mg cortisone acetate tablet (Beacon Pharmaceuticals, Tunbridge Wells, UK). Blood samples were obtained at intervals indicated in the figures, and plasma was stored at –80 C for analysis of cortisol.
Role of glucocorticoids in mediating metabolic effects of 5 wk administration of rosiglitazone
Objective.
Results from the 7-d administration of rosiglitazone suggested that any effect on 11ß-HSD1 is limited to adipose tissue, as judged by a marginal change in microdialysis measurement. To validate this observation, we aimed to assess changes in adipose 11ß-HSD1 mRNA. In addition, we aimed to test whether reductions in glucocorticoid signaling mediate the insulin-sensitizing effects of rosiglitazone; drug was administered for 5 wk to ensure that insulin sensitization would be measurable (25). In animals, the most striking change in adipose 11ß-HSD1 mRNA was described after PPAR agonist administration to obese db/db mice (6), so that in this study participants with a wide range of body mass indices were recruited to establish whether obesity modifies the effects of rosiglitazone on 11ß-HSD1 in humans.
Participants. Twelve male healthy volunteers were recruited by advertisement. Inclusion criteria were age 20–70 yr; body mass index between 20 and 40 kg/m2; normal thyroid, renal, and hepatic function on biochemical screening; alcohol intake less than 28 U per week; no chronic disease (including dyslipidemia or glucose intolerance); and no chronic medication or glucocorticoid therapy in the previous 6 months. Local ethical approval and written informed consent were obtained.
Study design and protocol. A two-phase double-blind randomized balanced crossover study compared 5 wk treatment with oral rosiglitazone 4 mg daily vs. placebo. Phases were separated by 2 wk washout. Subjects attended for measurements on the last day of the fourth week and again on the last day of the fifth week of each phase. On one of these visits, they received a glucocorticoid blockade combination of tablets and on the other occasion placebo tablets in double-blind randomized order.
Glucocorticoid blockade was achieved by administration of the GR antagonist RU38486 (mifepristone; Exelgyn, Henley-on-Thames, UK) and, to prevent rebound hypercortisolemia, the cortisol biosynthesis inhibitor metyrapone (Metopirone; Alliance Pharmaceuticals, Chipenham, UK). Subjects took 400 mg RU38486 and 1 g metyrapone at 2300 h the previous evening and again at 0800 h on the morning of the study. This dose of RU38486 has previously been shown to be sufficient to reduce plasma triglycerides (26) and achieve substantial drug levels in adipose tissue (27).
On each study day, subjects attended the Clinical Research Facility at 0800 h after overnight fast. Cannulas were inserted in both antecubital fossae for infusion and sampling. Abdominal sc adipose tissue was biopsied under local anesthesia (1% lignocaine) using a 14-gauge needle. Samples of approximately 500 mg were immediately frozen in liquid nitrogen and stored at –80 C. An incremental low-dose insulin infusion was performed by infusing saline for 1 h and then insulin at 0.01 U/kg body weight·h for 1 h, 0.033 U/kg·h for 1 h, and 0.1 U/kg·h for 1 h (28). Venous blood was taken every 60 min and plasma stored at –80 C for measurement of glucose, insulin, free fatty acids, and glycerol. Blood glucose was monitored with a bedside capillary glucose meter and 10% dextrose was infused if necessary to maintain venous blood glucose between 4.0 and 4.5 mmol/liter.
Effects of 12 wk administration of rosiglitazone on sc adipose 11ß-HSD1
Objective. Results from the administration of rosiglitazone for 7 d or 5 wk did not show potent effects on 11ß-HSD1. Finally, to test whether longer-term effects of rosiglitazone lead indirectly to changes in adipose 11ß-HSD1, we aimed to test the effects of drug administration for 12 wk on 11ß-HSD1 mRNA in vitro and enzyme activity in vivo.
Participants. Twelve men with type 2 diabetes treated with diet alone were recruited from the clinic; other data from this cohort have been reported previously (29). Inclusion criteria were age 30–70 yr, fasting plasma glucose 7–12 mmol/liter, and body mass index more than 24 kg/m2. Exclusion criteria were previous treatment with oral hypoglycemic agents; current medication known to affect glucose metabolism; comorbidity including cardiac, hepatic, or renal disease; or microvascular complications of diabetes as determined by history, clinical examination, and routine blood investigations. Local ethical approval and written informed consent were obtained.
Study design and protocol. A two-phase double-blind randomized balanced crossover study compared 12 wk of treatment with oral rosiglitazone 4 mg twice daily vs. placebo. At the end of each phase, patients attended the clinical research unit after a 10-h overnight fast for investigation. Vigorous exercise and alcohol were avoided for 24 h before each study.
Arteriovenous sampling was employed to assess sc adipose 11ß-HSD1 activity. Arterialized blood was obtained from a vein draining a heated hand. Venous blood was obtained simultaneously from the superficial epigastric vein, as described previously (30). Two sets of blood samples were taken 30 min apart for measurement of cortisol and cortisone (31). The mean of these two results was used to calculate the arteriovenous difference in concentration. Fractional extraction was calculated as (arteriovenous concentration difference)/(arterial concentration).
Abdominal sc adipose tissue was biopsied under local anesthesia (1% lignocaine) using a 12-gauge needle. Samples were immediately frozen in liquid nitrogen and stored at –70 C for later mRNA quantification.
Laboratory assays
Deuterated cortisol and its metabolites. Isotopomers of D4-cortisol, 9,12,12-[2H]3-cortisone (D3-cortisone), and 9,12,12-[2H]3-cortisol (D3-cortisol) were measured as described previously (22) in plasma and urine by gas chromatography/mass spectrometry after formation of methoxime-trimethylsilyl derivatives. Epi-cortisol was added to plasma, and epi-cortisol and epi-5ß-tetrahydrocortisol added to urine as internal standards. Analysis of isotopomers was performed on a Finnigan Voyager gas chromatography/mass spectrometry system with a CP-Sil 5CB (25 m x 0.25 mm internal diameter x 0.25 µm; J&W Scientific, Folsom, CA). Enrichments were calculated from peak areas. Results for D4-cortisol were corrected for isotopic interference from endogenous mass+4 cortisol and mass+1 D3-cortisol. Results for D3-cortisol were corrected for isotopic interference from endogenous mass+3 cortisol.
Cortisol kinetics were calculated as described previously (20, 21), using the mean of four measurements obtained from each subject in steady state after 240–300 min of D4-cortisol infusion. Rates of appearance of endogenous cortisol were calculated as [(rate of D4-cortisol infusion)/(D4-cortisol:cortisol ratio)] – (rate of infusion of cortisol). Clearance of D4-cortisol was calculated as (infusion rate)/(concentration). Rate of appearance of D3-cortisol was calculated as (rate of D4-cortisol infusion)/(D4-cortisol:D3-cortisol ratio). The difference in the metabolism of D3- and D4-cortisol is that D3-cortisol is reversibly interconverted with D3-cortisone, whereas D4-cortisol cannot be regenerated from D3-cortisone; as a result, the dilution of D4-cortisol by D3-cortisol reflects the rate of 11ß-HSD1 reductase activity. Urinary excretion rates of deuterated steroids were calculated for each hour as (concentration of deuterated steroid) x (volume of urine).
Endogenous urinary steroids.
Urinary excretion of cortisol and its metabolites [5ß-tetrahydrocortisol (5ß-THF), 5-THF, 5ß-tetrahydrocortisone (THE), cortols, cortolones, and cortisone] was measured in 24-h samples by electron impact gas chromatography/mass spectrometry as previously described (32). Total cortisol excretion was calculated from the sum of 5ß-THF, 5-THF, and THE. The balance between 11ß-HSD activities in all tissues was assessed as the ratio of (5ß-THF + 5-THF)/THE. Renal 11ß-HSD type 2 activity was assessed as urinary cortisol/cortisone ratio (32). The balance of 5- and 5ß-reductases was assessed by the ratio 5ß-THF/5-THF. Relative 5- and 5ß-reduction of cortisol was also assessed by 5-THF/cortisol, 5ß-THF/cortisol, and 5ß-THE/cortisone ratios.
Microdialysis. Steroids were extracted from the dialysis effluent and separated by thin layer chromatography as previously described (20). Tritiated steroids were quantified by liquid scintillation counting (to <2% error), and background counts were subtracted; counts in the sample were always at least 5-fold higher than background. 11ß-HSD1 reductase activity was calculated as (dpm cortisol)/(dpm cortisol + dpm cortisone) x 100%.
Analyses in adipose biopsies. After RNA extraction from whole adipose tissue, mRNAs were amplified by real-time PCR as previously described (17, 33), quantified against a standard curve of serial dilution of a pooled sample, and expressed as a ratio against either cyclophilin A (12-wk study) or 18S (4- to 5-wk study) as housekeeping genes.
To estimate total 11ß-HSD1 protein, activity was measured in the dehydrogenase direction, which is the more stable in vitro, as previously described (16). Briefly, 250 mg tissue was homogenized in KREBS buffer, and 750 µg/ml total protein was incubated with 2 mM nicotinamide adenine dinucleotide phosphate, 0.2% glucose, and 100 nM cortisol (of which 10 nM 1,2,6,7-[3H]4-cortisol) and incubated at 37 C for 30 h. Aliquots were withdrawn at intervals and conversion to 1,2,6,7-[3H]4-cortisone was measured by HPLC with online scintillation detection.
Other assays. Insulin was measured by enzyme immunoassay (Eurogenetics Tasah Corp. UK Ltd., Hampton, UK). Electrolytes were measured with a Vitras 950 (Ortho Diagnostics, Raritan, NJ). Glucose and lipid profile were determined on a Cobas Mira Plus (Roche, Mannheim, Germany), using enzymatic colorimetric kits from Roche for triglycerides and total cholesterol. Free fatty acids (Wako, Neuss, Germany) and glycerol (Sigma, Poole, UK) were measured by enzymatic colorimetric techniques on a Cobas Fara. Cortisone was measured by RIA (Immunovation, Southampton, UK). RU38486 was measured by gas chromatography/mass spectrometry. ELISA kits were used to measure cortisol (MP Biomedicals, Illkirch, France), ACTH (Biomerica Inc., Newport Beach, CA), dehydroepiandrosterone, and androstenedione (DRG Instruments GmbH, Marburg, Germany).
Statistical analyses
Sample sizes were calculated to give more than 80% power to detect to P < 0.05 a similar magnitude of difference in the 11ß-HSD1 parameter, as has been reported previously in rodent studies. Results are shown as mean ± SEM. P < 0.05 was considered statistically significant. Single measurements in paired groups were compared by Student’s t tests or Wilcoxon signed rank tests. For single measurements in multiple paired groups, comparison was by repeated-measures ANOVA with post hoc Fisher’s least squares difference (LSD) tests as appropriate. For time courses in multiple paired groups after a single intervention (i.e. rosiglitazone or fenofibrate vs. placebo), placebo data were subtracted from drug-treatment data and the difference from zero tested by repeated-measures ANOVA; to test variability in response to drugs, this difference was also correlated with age and body mass index by Pearson correlation. For dual simultaneous interventions (i.e. with rosiglitazone and glucocorticoid blockade), effects of each intervention were tested by factorial ANOVA with post hoc Fisher’s LSD tests as appropriate; to test variability in response to intervention, median age and body mass index were used to define subgroups, and differences between subgroups were tested in a factorial ANOVA model.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Participants were men aged 40 ± 2 yr (range, 30–52 yr) with body mass indices of 27.3 ± 1.5 kg/m2 (range, 19.8–33.7 kg/m2). In fasting plasma (Table 1
), fenofibrate lowered total cholesterol compared with placebo, and both total cholesterol and triglycerides were lower on fenofibrate than on rosiglitazone. There were no significant changes in glucose or insulin levels.
|
). Neither drug had statistically significant effects on urinary excretion of endogenous cortisol and its metabolites in urine (Table 1
), although rosiglitazone tended to increase excretion, particularly of 5-reduced cortisol metabolites. Neither drug affected whole-body metabolism of D4-cortisol during tracer infusion (Table 1).
|
. Rosiglitazone had a marginal, but statistically significant, effect to reduce conversion of cortisone to cortisol during microdialysis in adipose tissue (Fig. 1A). Fenofibrate did not alter cortisol generation in adipose tissue, and neither drug altered first-pass liver metabolism of cortisone to cortisol (Fig. 1B). Responses to fenofibrate did not correlate with age or body mass index. Responses to rosiglitazone did not correlate with age, although rosiglitazone tended to increase urinary steroid excretion among obese subjects (P < 0.05 for correlation of body mass index with changes in excretion of endogenous 5ß-THF, D4–5ß-THF, D3–5ß-THF, and D3-cortisone).
Role of glucocorticoids in mediating metabolic effects of 5 wk administration of rosiglitazone
Participants were men aged 41 ± 2.4 yr (range, 25–52 yr) with body mass indices of 29.6 ± 1.7 kg/m2 (range, 20.3–39.6 kg/m2). Effects of rosiglitazone and glucocorticoid blockade on baseline biochemical measurements are shown in Table 2, effects in adipose tissue are shown in Fig. 2, and results from insulin infusions are shown in Fig. 3. Body weight was not different on placebo vs. rosiglitazone (93.8 ± 5.5 vs. 93.7 ± 5.7 kg, respectively).
|
|
|
). In the absence of glucocorticoid blockade, rosiglitazone increased baseline serum triglycerides (Table 2) and lowered free fatty acids, glycerol, and C-peptide concentrations before and during incremental insulin infusion (Fig. 3) but had no effect on plasma cortisol or ACTH concentrations. Rosiglitazone did not alter insulin or glucose concentrations achieved, or dextrose infusion rates, during low-dose insulin infusion.
After glucocorticoid blockade, plasma concentrations of RU38486 were undetectable in the placebo phase and 6.2 ± 0.6 µmol/liter at 0830 h on the morning of study, comparable with previous studies (27). Glucocorticoid blockade lowered plasma cortisol (Fig. 3A) and increased serum ACTH, dehydroepiandrosterone, and androstenedione levels (Table 2); these effects were not modified by rosiglitazone. In the absence of rosiglitazone, glucocorticoid blockade had no effect on biochemical responses to insulin infusion (Fig. 3) or on adipose 11ß-HSD1 activity but tended to increase intra-adipose leptin mRNA levels (P = 0.06; Fig. 2A). In the presence of rosiglitazone, glucocorticoid blockade decreased serum triglycerides (Table 2), increased intra-adipose leptin mRNA levels (P < 0.01; Fig. 2A), and decreased adipose 11ß-HSD1 activity (Fig. 2B) but did not modify the effects of rosiglitazone on biochemical responses to insulin infusion (Fig. 3).
To address the possibility that rosiglitazone affects 11ß-HSD1 only in obese men, participants were divided into those with body mass indices less than 27 kg/m2 (n = 5; mean, 23.9 ± 1.0 kg/m2) or more than 27 kg/m2 (n = 7; mean, 33.7 ± 1.6 kg/m2). Rosiglitazone did not lower 11ß-HSD1 mRNA or activity in either group (data not shown). Similarly, there was no difference in effects of rosiglitazone on 11ß-HSD1 mRNA or activity in older (>42 yr) or younger (<42 yr) participants.
Effects of 12 wk administration of rosiglitazone on sc adipose 11ß-HSD1
Participants included 12 men with diet-controlled type 2 diabetes, who were a subset of 24 patients described previously (29). They were aged 51 ± 10 yr (range, 31–69 yr) with body mass indices of 33 ± 6 kg/m2 (range, 25–40 kg/m2). As previously reported (29), rosiglitazone decreased glycosylated hemoglobin (from a mean of 7.4 to 7.0%) and improved insulin sensitivity while increasing body mass index and total plasma cholesterol (data not shown).
Concentrations of cortisol and cortisone in arterial samples and in samples obtained from veins draining sc abdominal adipose tissue are shown in Table 3. With both placebo and rosiglitazone, there was measurable extraction of cortisone across the sc adipose tissue, but cortisol production was not measurable. There was no correlation between cortisone extraction and body mass index or age. Rosiglitazone decreased fractional cortisone extraction and also increased systemic plasma cortisol concentrations.
|
). The change in mRNA levels was not correlated with the change in cortisone fractional extraction, degree of insulin sensitization, or rosiglitazone-induced weight gain and was independent of body weight and age of participants (data not shown). |
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
agonists on lipid metabolism. For rosiglitazone, the results were more complex. Consistent with the high expression of PPAR in adipose tissue, rosiglitazone was reported to reduce 11ß-HSD1 expression principally in adipose tissue in rodents (6, 7). Although after 7 d of rosiglitazone therapy in healthy men there was reduced conversion of cortisone to cortisol in sc adipose tissue during microdialysis, this effect was of small magnitude, did not influence whole-body rates of cortisone to cortisol conversion, and was observed only after 4–5 h of cortisone perfusion (Fig. 1A), which is not typical of the differences in initial rate of cortisol generation observed in previous microdialysis experiments (20). Moreover, 5 wk of rosiglitazone therapy did not change 11ß-HSD1 mRNA or enzyme activity in adipose tissue, and with respect to lowering C-peptide concentrations and enhancing insulin-mediated suppression of serum free fatty acids, rosiglitazone was equally effective in the presence or absence of glucocorticoid blockade. These data are consistent with previous experiments in rats showing that metabolic responses to rosiglitazone are not influenced by adrenalectomy (14) and suggest that, contrary to inferences from the literature (6, 7), changes in intra-adipose cortisol regeneration or GR activation do not mediate acute effects of PPAR agonists in humans.
In contrast with the lack of acute effects, after 12 wk of rosiglitazone administration, there was evidence of decreased 11ß-HSD1 expression and activity in vitro and in vivo in abdominal sc adipose tissue. 11ß-HSD1 mRNA levels were reduced by approximately 40%, and arteriovenous extraction of cortisone was reduced by approximately 50%. The magnitude of cortisol regeneration in adipose tissue in humans remains uncertain. In earlier studies, Katz et al. (31) were unable to measure net cortisol generation across human sc adipose tissue, although they did detect net cortisone extraction. We repeated these measurements in our 12-wk study and found the same (Table 2). This apparent paradox also occurs in the splanchnic circulation, where hepatic vein cortisol concentrations are not higher than arterial concentrations, even though tracer studies reveal substantial cortisol production in the splanchnic circulation and cortisone concentrations are substantially lower in hepatic vein than artery (23, 24). The most likely explanation is that cortisol, but not cortisone, is metabolized by other enzymes in adipose tissue, resulting in removal of cortisol in excess of removal of cortisone, as occurs across the liver (24). This interpretation is consistent with the expression of 5-reductase (which metabolizes cortisol but not cortisone) but not 5ß-reductase (which metabolizes cortisol and cortisone) in adipose tissue (34). Tracer studies with arteriovenous sampling across the sc adipose tissue bed will be required to quantify the rate of cortisol generation. However, because no other enzymes that metabolize cortisone are known to be expressed in adipose tissue, it is reasonable to infer that the reduced cortisone extraction rate reflects a fall in sc adipose 11ß-HSD1 activity after 12 wk of rosiglitazone therapy.
For practical reasons, the longer-term study was feasible only in patients with type 2 diabetes who are motivated to participate in a 24-wk crossover study. The short-term studies were conducted in nondiabetic subjects with a wide range of body mass index. As a result, direct comparison between the studies should be cautious. Although type 2 diabetes per se does not alter adipose 11ß-HSD1 activity (35), it is conceivable that patients with diabetes respond differently to PPAR agonists, particularly because the insulin-sensitizing effect of these agents is greatest among the most insulin-resistant patients (1). However, in the study of 5 wk exposure to rosiglitazone, we did not find any down-regulation of adipose 11ß-HSD1 mRNA even in the most obese participants, who were hyperinsulinemic and exhibited the greatest insulin-sensitizing response (data not shown). Moreover, variations in body mass index did not predict the response of 11ß-HSD1 to PPAR agonists in the 7-d study. A more likely explanation for long-term down-regulation of adipose 11ß-HSD1 by rosiglitazone in diabetic patients is that it reflects indirect effects, mediated for example by changes in body fat distribution, which were not apparent in the shorter-term studies.
In Zucker rats, we showed that 21 d of rosiglitazone administration reversed the up-regulation of 5-reductase that occurs in obesity (13). In humans, however, we did not find effects of rosiglitazone on urinary excretion of 5-reduced metabolites during D4-cortisol infusion, and there was a nonsignificant trend for an increase, rather than decrease, in urinary excretion of endogenous 5-reduction (Table 1).
The protocol for glucocorticoid blockade, using a combination of RU38486 and metyrapone, was based upon previous work demonstrating effective inhibition of glucocorticoid negative feedback and lowering of serum triglycerides (26) and basal hepatic glucose output (36). We achieved similar serum concentrations of RU38486, increased ACTH and adrenal androgen levels, consistent with successful GR antagonism. In addition, metyrapone prevented any compensatory hypercortisolemia, and indeed plasma cortisol concentrations were lower with glucocorticoid blockade. No detailed evaluation of metabolic responses to this form of glucocorticoid blockade has been published previously, but on the basis of previous studies with similarly acute administration of GR agonists, we anticipated that glucocorticoid blockade would decrease serum triglycerides (26), free fatty acids (37), ketone bodies (38), glycerol (39), and C-peptide and increase glucose infusion rate during low-dose insulin infusion. Effects on intra-adipose mRNA levels were more difficult to predict, because relatively little research has been conducted in vivo in humans. Furthermore, we predicted that these effects would be smaller during rosiglitazone administration, if rosiglitazone reduced intra-adipose cortisol concentrations. However, the only effects of glucocorticoid blockade that we observed were a reduction in triglycerides, an increase in intra-adipose leptin mRNA levels, and a reduction in adipose 11ß-HSD1 activity, all paradoxically more obvious during rosiglitazone treatment. The lack of effect of glucocorticoid blockade alone makes the incremental effects during rosiglitazone therapy more difficult to interpret, although if anything, it appears that glucocorticoid action is increased, rather than decreased, by rosiglitazone. Different results might have been obtained if it were possible to administer glucocorticoid blockade for a longer period, but these results call into question the acute physiological role of glucocorticoids in modifying fatty acid metabolism.
The effects of rosiglitazone on 11ß-HSD1 were limited to abdominal sc adipose tissue. 11ß-HSD1 is likely to be important also in visceral adipose tissue, where we were unable to make direct measurements. It is unlikely that visceral adipose tissue 11ß-HSD1 was altered in the 7-d study (because we would have expected changes in D3-cortisol generation) (23), or in the 5-wk study (because we would have expected changes in responses to glucocorticoid blockade). However, additional studies will be required to clarify the effects of long-term PPAR agonist administration on visceral adipose 11ß-HSD1.
Although the primary focus of these experiments was on cortisol metabolism, we also assessed cortisol concentrations in blood, knowing that PPAR agonists might influence ACTH secretion (18). Total cortisol metabolite excretion, endogenous cortisol production during D4-cortisol infusion, and morning plasma cortisol after 7 d or 5 wk of treatment were unaffected by rosiglitazone. However, after 12 wk exposure, rosiglitazone increased systemic cortisol concentrations. These observations have not been reported previously to our knowledge. It seems most likely that the redistribution of fat after long-term rosiglitazone therapy results in alterations in the hypothalamic-pituitary-adrenal axis, reversing the factors determining the lower plasma cortisol that is characteristic of obesity (40).
There is evidence from 11ß-HSD1 knockout mice that loss of GR activation leads to up-regulation of PPAR expression (11), potentially increasing sensitivity to PPAR agonists. Thus, PPAR agonists and 11ß-HSD1 inhibitors may yet have synergistic effects if administered concurrently. However, we conclude that, in contrast with findings in rodents, neither PPAR nor PPAR agonists acutely regulate cortisol secretion and metabolism in humans, and the direct insulin-sensitizing effect of PPAR agonists appears to be independent of glucocorticoid action.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure Statement: D.J.W., R.H.S., G.D.T., N.Z.M.H., R.A., and F.K. have nothing to declare. B.R.W. has consulted recently for Astra-Zeneca, Merck, Johnson & Johnson, Syrrx, Incyte, and Vitae and has received lecture fees from Abbott and Bristol-Myers Squibb.
First Published Online February 27, 2007
1 As joint first authors, D.J.W. and R.H.S. made equal contributions. 
Abbreviations: D3-cortisol, 9,12,12-[2H]3-Cortisol; D3-cortisone, 9,12,12-[2H]3-cortisone; D4-cortisol, 9,11,12,12-[2H]4-cortisol; GR, glucocorticoid receptor; 11ß-HSD1, 11ß-hydroxysteroid dehydrogenase type 1; LSD, least squares difference; PPAR, peroxisome proliferator-activated receptor; THE, 5ß-tetrahydrocortisone; THF, tetrahydrocortisol.
Received December 8, 2006.
Accepted February 15, 2007.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
and human metabolic disease. J Clin Invest 116:581–589[CrossRef][Medline] in energy metabolism and vascular homeostasis. J Clin Invest 116:571–580[CrossRef][Medline] agonists reduce 11ß-hydroxysteroid dehydrogenase type 1 in the liver. Biochem Biophys Res Commun 279:330–336[CrossRef][Medline] ligands inhibit adipocyte 11ß-hydroxysteroid dehydrogenase type 1 expression and activity. J Biol Chem 276:12629–12635 activation mediates adipose depot-specific effects on gene expression and lipoprotein lipase activity: mechanisms for modulation of postprandial lipemia and differential adipose accretion. Diabetes 52:291–299 agonism on adipose tissue remodeling, insulin sensitivity, and lipemia in absence of glucocorticoids. Am J Physiol Regul Integr Comp Physiol 287:R1116–R1123 target genes related to lipid storage in vivo. Diabetes Care 27:1660–1667 receptor is a novel therapeutic target for ACTH-secreting pituitary adenomas. Nat Med 8:1281–1287[CrossRef][Medline] ligand rosiglitazone in Cushing’s disease. Eur J Endocrinol 151:173–178[Abstract]This article has been cited by other articles:
 |
|
 |
|
  D. E W Livingstone, S. L Grassick, G. L Currie, B. R Walker, and R. Andrew Dysregulation of glucocorticoid metabolism in murine obesity: comparable effects of leptin resistance and deficiency J. Endocrinol., May 1, 2009; 201(2): 211 - 218. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. H. Stimson, J. Andersson, R. Andrew, D. N. Redhead, F. Karpe, P. C. Hayes, T. Olsson, and B. R. Walker Cortisol Release From Adipose Tissue by 11{beta}-Hydroxysteroid Dehydrogenase Type 1 in Humans Diabetes, January 1, 2009; 58(1): 46 - 53. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
), fenofibrate (
), or rosiglitazone (
). By repeated-measures ANOVA, cortisol generation in adipose tissue was lower after rosiglitazone than during placebo or fenofibrate therapy (P < 0.01), although post hoc Fisher’s LSD tests did not reveal significant differences at any individual time point. There were no differences in first-pass liver metabolism of cortisone to cortisol.
