| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Department of Endocrinology, William Harvey Research Institute, Barts and the London, Queen Mary’s School of Medicine, University of London, London, UK;
Department of Endocrine Neurobiology, Institute of Experimental Medicine, Hungarian Academy of Sciences, Budapest, Hungary, and Tupper Research Institute and Department of Medicine, Division of Endocrinology, Diabetes, Metabolism and Molecular Medicine, Tufts-New England Medical Center, Boston, Massachusetts, USA;
Division of Endocrinology, Diabetes and Clinical Nutrition, University Hospital, Basel, Switzerland;
Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, U.S. National Institutes of Health, Bethesda, Maryland, USA;
|| Service d’Endocrinologie, Diabetologie et Metabolism, Department of Internal Medicine, Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland; and
¶ Department of Internal Medicine, University of Ancona, Ancona, Italy
2Correspondence: Department of Endocrinology, John Vane Science Centre, Barts and the London, Queen Mary’s School of Medicine, London EC1M 6BQ, UK. E-mail: m.korbonits{at}qmul.ac.uk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: obesity • insulin resistance • lipid metabolism
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. These manifestations contribute to the increased cardiovascular risk observed in patients with Cushing’s syndrome, and to reduced life expectancy and quality of life. Symptoms may persist even when the excess glucocorticoid exposure is normalized (2
, 3)
. Obesity in Cushing’s syndrome is characterized by increased appetite (4)
and body weight, reduced lean mass, and a characteristically distributed fat accumulation with a shift from the more inert subcutaneous fat to an accumulation of the metabolically more active visceral fat (5)
. Glucocorticoid hormones and drugs cause abnormal metabolism in the liver with insulin resistance and hepatic steatosis, accelerated protein breakdown in the skeletal muscle, cardiac effects resulting indirectly from hypertension, left ventricular hypertrophy, atherosclerosis, and myocardial ischemia. However, the cardiac abnormalities are not fully explicable by the secondary changes resulting from insulin resistance and hypertension, suggesting an additional direct effect of excess cortisol on the myocardium (6
, 7)
.
Adenosine monophosphate-dependent kinase (AMPK) is a sensor of cellular energy status working as a cellular "fuel gauge" (8)
. AMPK is activated by decreases in the energy state of a cell and, once activated, AMPK switches off anabolic pathways such as fatty acid, triglycerides, and cholesterol synthesis, as well as protein synthesis and transcription, and switches on catabolic pathways, including glycolysis and fatty acid oxidation. Recent studies have shown that AMPK plays an important role in nutritional sensing and feeding behavior regulation in the hypothalamus (9
10
11
12
13)
. Several metabolic hormones (leptin, ghrelin, and cannabinoids) show tissue-dependent regulatory effects on AMPK activity (9
10
11
, 13)
.
Because many of the changes seen in glucocorticoid excess correspond to metabolic steps regulated by AMPK, we hypothesized that AMPK activity would be increased in the hypothalamus and decreased in adipose tissue. As glucocorticoids have been shown to stimulate endocannabinoids in the hypothalamus (14
, 15)
, and we have previously shown that endocannabinoids stimulate hypothalamic AMPK activity (13)
, we further hypothesized that glucocorticoid-induced changes in hypothalamic AMPK would be associated with changes in hypothalamic endocannabinoid content.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 17)
was used in our studies. The animals were divided into 3 groups (n=6–8).
In the first group, animals were adrenalectomized bilaterally and corticosterone pellets were implanted under the skin (16)
. The pellets, containing 100 mg of corticosterone, were made using commercially available molds, as described previously (16)
. After surgery, animals were provided with a bottle of 0.5% NaCl and a second bottle containing 30% sucrose to drink ad libitum until day 14, when they were killed (16
, 17)
.
Animals in the second group were sham-adrenalectomized bilaterally and implanted with placebo pellets (containing cholesterol instead of corticosterone). Animals were offered 0.5% NaCl and 30% sucrose to drink ad libitum for 14 days, then were killed.
Animals in the third group were sham-adrenalectomized bilaterally and implanted with placebo pellets. After drinking only 0.5% NaCl, without sucrose, for 14 days, they were killed.
Corticosteroid treatment in rats without sucrose administration results in marked stress and catabolism with only limited features of the metabolic syndrome and therefore does not provide a good model for human Cushing’s syndrome (16)
. We therefore provided the placebo-implanted rats with a sucrose solution to drink, and used this second group as our principal control for corticosterone administration. We also used the third group, the placebo-implanted, saline-drinking rats, as the appropriate control for the possible influence of sucrose on tissue AMPK activity.
In all animals, we assessed weight and intake of chow and sucrose daily. On day 14, blood samples for plasma corticosterone, ACTH, insulin, leptin, ghrelin, adiponectin, glucose, total cholesterol, and triglycerides were taken. Samples of tissues (hypothalamus; liver; heart; visceral and subcutaneous white adipose tissue; brown adipose tissue; red and white skeletal muscle, both from the gastrocnemius muscle; and heart) were immediately frozen on powdered dry ice and stored at –80°C. The brain was dissected and the hypothalamus removed as a 2-mm-thick coronal section from –2 to –4 from the Bregma and from the supraoptic decussation to the dorsal end of the lower part of the third ventricle. The weight of the visceral (mesenteric, perirenal, and epididymal) and subcutaneous (inguinal) fat was measured.
Ex vivo differentiated human adipocytes
The differentiation of human mesenchymal stem cells (MSC) from bone marrow into adipocytes and their culture has been described previously (18
, 19)
. Nucleated cells were isolated from the aspirate by Ficoll density gradient centrifugation (Histopaque1; Sigma, Buchs, Switzerland). The cells were expanded in Dulbecco modified Eagle medium (DMEM), supplemented with 10% fetal calf serum (FCS) and 5 ng/ml basal fibroblast growth factor (Invitrogen AG, Basel, Switzerland) until they reached confluence. Adipogenic differentiation was induced by adding DMEM/F12 containing 3% FCS (Invitrogen), 1 µM dexamethasone (Sigma), 0.1 mM L-ascorbic acid (Sigma), 250 µM 3-isobutyl-1-methylxanthine (Sigma), 5 µM transferrin (Calbiochem, La Jolla, CA, USA), 0.2 nM 3,3',5-triiodo-L-thyronine (Sigma) and 100 nM insulin (Actrapid; Novo Nordisk, Kuesnacht, Switzerland). The medium was changed every 3 days. After 15–18 days of differentiation, medium was removed by washing with PBS. Adipocytes were kept for 48 h in basal medium (FCS 3%) before experiments.
All cell treatments were performed with dexamethasone 1 µM diluted in ethanol (19
, 20)
with time-response curves, and the controls were treated with the identical medium without dexamethasone. Dexamethasone was always prepared freshly and kept covered, as it is light sensitive. Metformin, 0.1, 1, and 10 mM, was added at the time of cell treatment.
Cell culture of human hepatoma (HepG2) cell lines
The human hepatoma HepG2 cells were maintained in a humidified atmosphere of 95% air and 5% carbon dioxide at 37°C in a 1x DMEM medium containing L-glutamine (Life Technologies, Inc., Paisley, UK), 1000 mg/L glucose, pyruvate, 10% FCS (Life Technologies, Inc.) and 1% penicillin and streptomycin (Sigma). All cell treatments were performed with dexamethasone 1 µM diluted in ethanol, as described above.
Primary hypothalamic rat cultures
Primary hypothalamic neuronal cell cultures were performed as described (21)
. Briefly, hypothalami were obtained from 18-day-old rat fetuses, and cells were dispersed mechanically, plated at a density of 500 live cells/mm2 in 6-well plates coated with 5 µg/ml poly-D-lysine (Sigma), and grown in NeurobasalTM medium with 0.04% B27 supplement (Life Technologies, Inc.) containing 500 µM glutamine and 25 µM glutamate (Sigma). Forty-eight hours after plating, 2 µM araC (cytosine β-D-arabinofuranoside; Sigma) was added to prevent proliferation of non-neuronal cells. Half of the medium was then changed every fourth day, and all experiments were performed in 3- to 4-week-old culture. Experiments were performed after 6 h of starvation of the cells in DMEM (Life Technologies, Inc.) at 5.5 mM glucose with no FCS. Cells were treated for 6–24 h with 1 µM dexamethasone and 1 mM metformin (Sigma).
Measurement of AMPK activity
AMPK activity assay has been previously described (13
, 22)
. Proteins were extracted using appropriate lysis buffer (13
, 22)
, and protein content was determined using bicinchoninic acid assay (Pierce Biotechnology, Rockford, IL, USA). AMPK was immunoprecipitated using 
1 and 2 AMPK antibodies, and activity was measured using 32P– incorporation into the AMPK substrate SAMS (amino acid sequence: HMRSAMSGLHLVKRR). Samples were assayed in duplicate, and each sample also was assayed without the addition of the substrate SAMS as a negative control.
Reverse transcriptase-polymerase chain reaction (RT-PCR)
Total RNA from adipose tissues (100 mg) was extracted with QIAzol reagent (Qiagen, Crawley, UK) and from liver and heart tissues (30 mg) using the Promega SV isolation kit (Promega). cDNA was synthesized from 1 µg RNA in a total volume of 25 µl using random primer (Roche, Burgess Hill, UK) and reverse transcriptase (Invitrogen). Levels of gene expression were quantified using real-time PCR (ABI PRISM 7900) with the human- and rat-specific assay-by-design primer and probe sets by Applied Biosystems (ABI, Warrington, UK) for rat-specific primers for sterol regulatory element-binding protein-1 (SREBP1c), fatty acid synthase (FAS), phosphoenolpyruvate carboxykinase (PEPCK)-1c, G6P, hormone sensitive lipase (HSL), AMPK 1, and AMPK 2 (assay codes available on request). Control reactions for RT (containing RNA but no RT enzyme) and PCR (containing PCR mixture but no cDNA) were run together with samples. All gene expression assays have FAM TM reporter dye at the 5' end of the TaqMan® MGB probe and a nonfluorescent quencher at the 3' end of the probe. The TaqMan MGB probes and primers have been premixed (20x) to a concentration of 18 µM for each primer and 5 µM for the probe (ABI) in a final volume of 10 µl. All the reactions were obtained in a duplex PCR reaction with β-actin (ACTB) as endogenous control (VIC® MGB Probe; ABI) at these conditions: 5 µl TaqMan Universal Master Mix (ABI), 0.5 µl 20x Assay Mix primer, 0.35 µl 20x Assay Mix ACTB, and 3.5 µl Tris-EDTA (TE). Reaction was run at 50°C for 2 min, 95°C for 10 min, and then at 40 cycles at 95°C for 15 s and 60°C for 1 min. Data were analyzed using the standard curve method. The relative quantities of target transcripts were calculated from triplicate samples after normalization of the data against the housekeeping gene β-actin.
Western blotting
Western blotting was performed running 20–40 µg protein on a 10 or 12% sodium dodecyl sulfate (SDS) gel (Invitrogen) and transferring it to a nitrocellulose membrane (pore size 0.45 µm; Whatman GmbH, Dassel, Germany). Membranes were incubated with the primary antiphosphorylated-LKB1 (pLKB1) antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA), which recognizes the LKB1 phosphorylated at Ser-431, at a concentration of 1:500; with the anti-Ca2+/calmodulin-dependent protein kinase kinase (CaMKK) antibody (1:200, Santa Cruz); and with the anti-CaMKKβ antibody (1:200, Santa Cruz) in 5% nonfat dry milk in Tris-buffered saline-Tween or in 5% BSA in Tris-buffered saline-Tween for CaMKKβ, overnight at 4°C. Goat anti-rabbit IgG IRDye 800CW, 1:10000, (Li-COR Bioscience, Lincoln, NE, USA) was used as secondary antibody for pLKB1 and CaMKK, and donkey anti-goat IgG IRDye 800CW, 1:10000, (Li-COR) as secondary antibody for CaMKKβ. After stripping the membrane, the glyceraldehyde 3-phosphate dehydrogenase (primary antibody 1:200, Santa Cruz; secondary goat anti-rabbit antibody 1:10,000, Li-COR) antibody was used to normalize for equal loading. Band intensities were detected with Li-COR, and densitometry values were calculated with the Image J program (U.S. National Institutes of Health, Bethesda, MD, USA).
Basal and insulin-mediated glucose uptake in human adipocytes
Basal and insulin-mediated glucose uptake was performed as described previously (23)
. On day 1, the differentiation medium of the human adipocytes was removed. Cells were washed 3x in warm PBS and kept in DMEM/F12 containing 5 mM glucose and 3% FBS. Supplements were added on day 2. On day 3, at t = 0 min, 100 nM insulin was added to half of the cells. At t = 20 min, 1 µC 2-deoxy-D-[3H(G)] deoxy-glucose (Perkin Elmer, Boston, MA, USA) was added to all wells. After another 15 min, the cells were washed 3x in ice-cold PBS and lysed in 0.1% SDS. Radioactivity was measured in a scintillation counter.
Lipid staining
Rat liver samples were stained with Oil-Red-O. Staining of sections was scored for droplet distribution and for staining intensity in positive cells as follows: 1) distribution: universal (all or almost all cells positive), score 2; zonal (groups of cells with intervening negative areas), score 1; negative (only the odd cell or no cells positive), score 0; 2) staining intensity in positive cells: strong (mixture of large and small globules of fat per cell), score 2; weak (only small globules of fat), score 1. The analysis was performed by adding the scores of 1 and 2.
Endocannabinoid content
Whole hypothalamus was homogenized and extracted with a chloroform/methanol method, and extracts were dried under nitrogen stream. Levels of anandamide, 2-arachidonylglycerol (2-AG) and 1-AG were quantified by liquid chromatography/inline mass spectrophotometry, as described (24)
. The amount of anandamide and 2-AG in the samples was determined by using inverse linear regression of standard curves. Values of 2-AG are expressed as femtomole or picomole per milligram wet tissue.
Hormone and biochemical measurements in the rat model
Corticosterone and ACTH were measured as described previously (25)
. Insulin, leptin, total (both acylated and unacylated) ghrelin, and adiponectin (Linco, St. Louis, MO, USA) and glucose, triglycerides, and cholesterol (Instrumental Laboratory, Lexington, MA, USA) were measured using commercial assays. The homeostasis model assessment (HOMA) index was calculated by multiplying insulin (µU/ml) and blood glucose (mmol/L), divided by 22.5.
Statistical analysis
Student’s t test or 1-way analysis of variance with least square difference for post hoc comparisons was applied for normally distributed data, whereas the Mann-Whitney U and Kruskal-Wallis tests were used for non-normally distributed data. Correlations were carried out using Spearman’s rank correlations. P < 0.05 was considered to indicate statistical significance. Data are shown as means ± SE unless stated otherwise.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 17)
. Biochemical and hormonal parameters confirmed that glucocorticoid-treated rats had the expected hormonal and metabolic profile of Cushing’s syndrome. In glucocorticoid-treated rats, corticosterone levels were higher (P=0.001) and ACTH levels were lower (P=0.007), compared with the 2 control groups (Fig. 1
A, B). Insulin concentrations were higher in corticosterone-treated rats compared with the 2 other groups (P=0.002), with no difference in glucose concentrations (Fig. 1C, D
). The HOMA-index for corticosteroid-treated rats (2.2±0.3) indicates insulin resistance, whereas the index was normal in the 2 control groups (0.9±0.07 and 0.8±0.1). Plasma leptin levels were higher in the corticosterone-treated rats compared with the 2 other groups (P=0.005, Fig. 1E
). In contrast, we found no difference in either adiponectin or ghrelin levels between the 3 groups (data not shown).
|
Corticosteroid-treated animals had significantly higher plasma total cholesterol levels compared with the 2 control groups (P=0.003, Fig. 1F
). Triglyceride levels were higher in both groups of sucrose-drinking animals treated with either placebo or corticosterone (Fig. 1G
). The more pronounced effect of corticosterone per se on hypercholesterolemia compared with hypertriglyceridemia is concordant with data reported in Cushing’s syndrome in humans (26)
. Total calorie intake was higher in the 2 sucrose-drinking groups compared with saline-drinking animals (P<0.001), and there was a trend for higher calorie intake in corticosterone-treated animals drinking sucrose compared with sucrose alone (P=0.07, Fig. 1H
). Caloric efficiency (grams gained/calorie ingested), in accordance to the marked catabolism of corticosteroid-treated animals, was lowest in the corticosterone-treated animals (P<0.001).
Adipose tissue
Rat model
Visceral white adipose fat depot weight was higher in corticosterone-treated rats (P=0.01) compared with the other 2 groups, whereas the inguinal fat depot, as a measure of subcutaneous fat, was similar between corticosteroid-treated rats and sucrose-drinking control rats (Fig. 2
A, B). In the visceral fat of corticosterone-treated rats, AMPK activity was significantly lower compared with the 2 control groups (P=0.008). In contrast, in subcutaneous fat there was no difference in AMPK activity between the 3 animal groups (P=0.16, Fig. 2C, D
). PEPCK and HSL mRNA expression were higher in the visceral adipose tissue of corticosteroid-treated animals compared with the other 2 groups (P=0.01, 0.006), and FAS and SREBP1c mRNA were higher in corticosteroid-treated rats compared with saline-drinking rats (P=0.03, 0.02, Fig. 2E
). By contrast, we found no difference between these gene products in subcutaneous adipose tissue except for HSL (P=0.003), suggesting an influence of corticosteroids on gluconeogenesis and lipogenesis primarily in visceral but not in subcutaneous fat tissue. The predominant AMPK subunit in adipose tissue, AMPK1, showed no difference in visceral adipose tissue mRNA expression between the groups (Fig. 2F
), whereas the expression level of the 2-subunit was negligible. We also studied the expression of the AMPK-regulating enzymes pLKB1 as well as CaMKK and CaMKKβ with Western blotting. The visceral fat tissue of glucocorticoid-treated animals showed a reduced expression of CaMKKβ (69.2±5.4% of saline control, P<0.05, Fig. 2G
), compatible with the reduced activity of AMPK in this tissue. No significant differences were detected for CaMKK or LKB1.
|
Human adipocytes
We also were able to show a decrease in AMPK activity in human adipocytes treated with dexamethasone 1 µM for 24 h compared with vehicle-treated adipocytes (P=0.04, Fig. 2H
), suggesting a direct cellular effect of corticosteroids on adipocyte AMPK activity. The dose-response curve with metformin 0.1 to 10 mM showed that metformin at doses of 1 and 10 mM significantly induced AMPK activity in human adipocytes compared with control. The inhibitory effect of dexamethasone on AMPK activity was antagonized by coadministration of metformin 10 mM, which increased AMPK activity to 224 ± 14% compared with dexamethasone treatment alone (P<0.01, Fig. 2I
). We studied the downstream enzyme FAS in response to metformin and dexamethasone. Similar to in vivo rat visceral adipose tissue, dexamethasone 1 µM for 24 h significantly stimulated FAS expression (174±30% of control, P<0.05, Fig. 2J
). Metformin (10 mM) treatment on its own inhibited FAS expression (63±11%), and this inhibition was abolished by dexamethasone (metformin vs. dexamethasone+metformin, P<0.001).
We furthermore studied the effect of dexamethasone on glucose uptake in human adipocytes (Fig. 2K
). Basal glucose uptake was inhibited by dexamethasone (77±1% of control, P<0.05) and stimulated by metformin (231±10% of control, P<0.001), whereas coadministration of dexamethasone+metformin decreased basal glucose uptake compared with metformin alone (P<0.05). Similar results were obtained when insulin-induced glucose uptake was studied. Dexamethasone inhibited (89±0.7% of insulin alone, P<0.05), whereas metformin further stimulated, insulin-induced glucose uptake (176±2% of insulin alone, P<0.001). Coadministration of dexamethasone+metformin decreased insulin-stimulated glucose uptake compared with metformin alone (P<0.05).
Liver
To confirm the metabolic effects of corticosteroids on rat liver, we first performed lipid staining of sections of rat livers. Corticosterone-treated rats showed a higher liver lipid content compared with placebo-treated rats drinking either sucrose or saline (P=0.002, Fig. 3
A). Higher FAS mRNA was found in corticosteroid-treated rats compared with the 2 control groups (P<0.001), whereas expression of PEPCK, SREBP1c, and glucose-6-phosphatase (G6P) was not significantly different (Fig. 3B
).
|
However, contrary to our initial expectations, AMPK activity in the liver was significantly higher in corticosterone-treated rats compared with the other 2 groups (P=0.01, Fig. 3C
). To investigate whether the observed effect was direct or indirect, we used a human hepatoma cell line. Here, similar to the rodent in vivo data, AMPK activity was increased following dexamethasone 1 µM treatment after 6 h compared with control cell cultures (P<0.001, Fig. 3D
), whereas at 24 h no significant changes were observed.
AMPK1-subunit mRNA expression was lower in the corticosteroid-treated animals compared with the saline-drinking rats (P=0.01), whereas no difference was observed between corticosteroid-treated and sucrose-drinking animals in AMPK2-subunit expression between the 3 groups (Fig. 3E
).
Hypothalamus
Humans with Cushing’s syndrome demonstrate increased appetite (4)
, and we therefore studied the influence of corticosteroids on AMPK activity in the hypothalamus. We observed a higher AMPK activity in corticosterone-treated rats compared with sucrose-drinking control rats (P<0.001). The lower AMPK activity in sucrose-drinking compared with saline-drinking control rats confirmed the inhibitory effect of sucrose on hypothalamic AMPK activity. Corticosterone administration counteracted the effects of sucrose and increased hypothalamic AMPK activity to levels comparable to saline-drinking animals (Fig. 4
A). These are compatible with an increase in appetite and, as noted above, we observed an increase in total calorie intake in corticosterone-treated compared with sucrose-drinking rats (Fig. 1H
).
|
To explore possible mechanisms as to how corticosteroids increase hypothalamic AMPK activity, we measured the endocannabinoid content in the 3 groups of animals. Similar to the pattern of the AMPK activity, the endocannabinoid 2-AG content was higher in corticosterone-treated rats than in sucrose-drinking control rats (P=0.01 Fig. 4B
). Anandamide levels were also higher in corticosterone-treated rats than in sucrose-drinking control rats (P=0.04, data not shown).
Primary rat hypothalamic cell cultures showed a significant increase of the AMPK activity after treatment with dexamethasone 1 µM for 6 h (P=0.02, Fig. 4C
). Addition of 1 mM metformin inhibited the effect of dexamethasone and decreased AMPK activity to 26 ± 2% of the level seen with dexamethasone alone (P<0.01, Fig. 4D
).
Skeletal and heart muscle
We found no difference in total AMPK activity in either white or dark muscle between the 3 groups. We therefore analyzed 1 or 2 AMPK activity separately, but again no changes were detected (data not shown). In the heart, AMPK activity was lower in corticosterone-treated rats compared with the other 2 groups (P<0.001, Fig. 5
).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Corticosterone-treated animals had significantly increased corticosterone and decreased ACTH levels, became markedly hyperinsulinemic and hyperleptinemic, showed high triglyceride and cholesterol levels, and developed visceral obesity. In rats, it is well known that chronic stressors associated with corticosterone excess usually decrease chow intake (16)
. This result is in contrast to humans, who generally demonstrate weight gain and increased appetite in response to excess glucocorticoids or chronic stress (27)
. To counteract the effect of glucocorticoids on catabolism, in the model of Cushing’s syndrome established by Dallman et al. (16
, 17)
, 30% sucrose is added to the diet of corticosterone-treated rats. In this model, considered one of the best animal models to study chronic excess glucocorticoid action and the consequent insulin resistance, we found a significantly increased appetite in the glucocorticoid-treated rats compared with saline-treated rats and a strong trend compared with sucrose-drinking rats. The palatability of the sucrose could be part of the increased total calorie intake in the sucrose-treated animals (16)
.
Typically, with excess glucocorticoid activity, there is a detrimental shift from the more inert subcutaneous fat to accumulation of the metabolically more disadvantageous visceral, intra-abdominal fat (5)
. Glucocorticoid effects on adipose cell differentiation and lipid accumulation are more pronounced in visceral than in subcutaneous fat, suggesting that glucocorticoids may play a pivotal role in the pathogenesis of central obesity (28)
. The higher level of local production of active glucocorticoids from inactive metabolites by levels of visceral adipose tissue 11β-hydroxydehydrogenase-1 in obesity and data from the tissue-specific 11β-hydroxydehydrogenase-1 knockout mice also support an important role of glucocorticoids in visceral adiposity (29)
. Here we show several changes demonstrated in response to glucocorticoids by visceral but not by subcutaneous fat. AMPK activation in adipose tissue inhibits lipogenesis and lipolysis and stimulates lipid oxidation. Thus, inhibition of AMPK leads to increased lipid stores in association with enhanced lipolysis, leading to the release of free fatty acids (30)
. AMPK activity in our experiment was significantly decreased in visceral but not in subcutaneous adipose tissue compared with controls. Therefore, we suggest that this is at least one of the mechanisms for the prominent effect of corticosteroids in accelerating lipid deposition in visceral adipose tissue, as these effects correspond to the effects of AMPK on lipid-regulating enzymes. To study the mechanism of the AMPK activation, the expression of the upstream kinases LKB1, CaMKK, and CaMKKβ was determined. Visceral adipose tissue showed decreased expression of CaMKKβ, and we speculate that this could explain, at least in part, the reduced AMPK activity measured. Our data suggest that corticosteroids, via a decrease in AMPK activity, increase the gene expression of fatty-acid synthesis enzymes (FAS, SREBP1c, and PEPCK) and HSL, which increases the release of lipids into the circulation. This change does not exclude an additional direct glucocorticoid effect via the nuclear glucocorticoid receptor on the expression of some of these enzymes (31
, 32)
. Similar to the in vivo rat data, in vitro incubation of human adipocytes with dexamethasone led to a fall in AMPK activity, indicative of a direct effect of glucocorticoids on human adipocyte AMPK. AMPK has a well-described effect on glucose uptake, and we observed that dexamethasone reduced both basal and insulin-induced glucose uptake in human adipocytes. The 1-subunit is the predominant subtype in adipose tissue, but there was no change in its mRNA expression with glucocorticoid treatment. Thus, these data suggest that the phosphorylation of AMPK, but not the total level of AMPK, is inhibited by treatment with glucocorticoids.
Glucocorticoid hormones and drugs are also known to cause abnormal metabolism in the liver, and we have shown previously that Cushing’s syndrome in humans is associated with increased hepatic fat accumulation (33)
. Accordingly, lipid staining revealed a higher fat content in the livers of corticosteroid-treated rats compared with controls. Unexpectedly, hepatic AMPK activity was significantly higher in corticosteroid-treated rats compared with sucrose-treated controls. On the other hand, previous reports in rats and mice subcutaneously injected with dexamethasone for 5 days, as well as in primary cultured hepatocytes, have also shown significant up-regulation of hepatic AMPK expression by glucocorticoids (34)
, therefore supporting our findings. Acute hepatic overexpression of constitutively-active AMPK leads to a significant accumulation of lipids within the liver (35)
, as gluconeogenesis is suppressed and lipid oxidation is facilitated (36)
; the resultant shift to lipid oxidation as a source of fuel induces a net flux of fatty acids into the liver (35)
. As we also have shown here, the decreased AMPK activity in visceral fat and the possible consequent increase in genes involved in fat metabolism could lead to a shift to enhanced lipogenesis as well as lipolysis from adipose tissue, resulting in an efflux of fatty acids from the adipose depots to the liver and thus to increased hepatic lipid uptake. The increased flux of centripetal hepatic lipid delivery may account for the hepatic steatosis characteristic of Cushing’s syndrome (33)
. In the model of liver-specific acute overexpression of constitutively active AMPK, total adipose tissue fell as lipids were transferred to the liver (35)
, but in Cushing’s syndrome this decrease would be counteracted by the fall in adipose tissue AMPK and thus increase local adipose tissue lipogenesis. Hypercortisolism is known to lead to increased lipolysis (37
, 38)
. Therefore, glucocorticoid agents have a dual effect on lipogenesis and lipolysis in adipose tissue, but, as shown from the increasing visceral fat depot mass, the effect on lipogenesis prevails.
On the other hand, corticosteroid excess leads to a state of insulin resistance, as was indeed seen in this model, and which is also characteristic of Cushing’s syndrome (26)
. Insulin resistance could be due to the increase in circulating free fatty acids consequent to increased fat tissue lipolysis and would be itself responsible for the increased liver lipid accumulation. Glucocorticoids also have been shown to impair insulin signaling via reduced cellular content of insulin receptor substrate (IRS-1) and reduction in tyrosine phosphorylation in IRS-1 (39
40
41)
. In agreement with previous findings in primary cultured rat hepatocytes (34)
, we found that dexamethasone treatment significantly increased AMPK activity in a human hepatoma cell line, supporting our in vivo data.
Obesity in Cushing’s disease results in part from increased food intake due to central nervous system effects (4
, 42
43
44)
, a complex process that is controlled by both hypothalamic and peripheral factors, and AMPK has been shown to play an important role (9
10
11
12
13)
. Administration of glucose inhibits hypothalamic AMPK activity (11)
, and we therefore analyzed 3 groups of animals, with the saline-drinking group being an appropriate control for the influence of sucrose on hypothalamic AMPK activity. Sucrose inhibited hypothalamic AMPK activity, as shown by a significantly lower hypothalamic AMPK activity in sucrose-drinking compared with saline-drinking control rats. Corticosterone administration counteracted the sucrose effects and increased hypothalamic AMPK activity to levels comparable to saline-drinking animals. This increase in hypothalamic AMPK activity would lead to an increased total calorie intake. Because of their marked catabolism, the experimental rats did not increase their net weight, despite the increase in adipose tissue weight.
Hypothalamic slices treated in vitro with dexamethasone showed an increase in endocannabinoid levels via a rapid, membrane receptor-dependent mechanism (14
, 15)
, whereas chronic glucocorticoid treatment has been reported to increase the 2-AG content of the rat amygdala (45)
. Here we show that glucocorticoid treatment increases hypothalamic 2-AG content in vivo. As it has been shown that cannabinoids have a direct effect on hypothalamic AMPK activity (13)
, this relation provides a possible mechanism for the orexigenic effect of glucocorticoids. We therefore suggest that glucocorticoids increase appetite via endocannabinoid mediation in the hypothalamus, although the involvement of other mediators, such as neuropeptide Y, agouti-related peptide, and pro-opiomelanocortin, is also possible (46
, 47)
.
Metformin is a mainstay of therapy in the treatment of type 2 diabetes. The ability of metformin to lower blood glucose requires LKB1 and AMPK signals (48
, 49)
. In our study, metformin reversed the effects of corticosteroids on AMPK in vitro both in primary hypothalamic cell culture as well as in adipocytes, suggesting that metformin and glucocorticoids influence the AMPK signaling pathway in opposite ways and that the metformin effect is able to override the cortisol effect. In addition, metformin treatment was able to reverse the effects of dexamethasone on 2 AMPK-regulated processes in human adipocytes: FAS mRNA expression and glucose uptake. These data suggest that metformin treatment could be beneficial in patients with Cushing’s syndrome in the treatment of their metabolic complications. The opposite effect of metformin on hypothalamic AMPK compared with peripheral tissues has recently been shown (50)
. Glucocorticoid agents, similarly to metformin and to other hormones such as leptin and ghrelin, influence AMPK activity in a tissue-specific manner. The underlying mechanisms for explaining the tissue specificity are to date not known, but this influence could be due to differential regulation or expression of either the AMPK kinases or AMPK phosphatases. This tissue-specific regulation of AMPK by glucocorticoids could be also a novel mechanism which might explain the glucocorticoid sensitivity of different tissues.
The anabolic effect of chronic hypercortisolism on adipose tissue contrasts sharply to its catabolic effect on muscle leading to accelerated protein breakdown (4)
. While glucose-induced insulin resistance resulted in decreased 2 AMPK activity on dark skeletal muscle (51)
, in our model of corticosteroid-induced insulin resistance there was no effect of corticosteroids on skeletal muscle total or 2 AMPK activity, neither on slow-twitch nor on fast-twitch skeletal muscle, suggesting that AMPK is not modulated by glucocorticoids in skeletal muscle in this experimental setting. However, in cardiac muscle, corticosterone induced a significant decrease in AMPK activity. Clinically, chronic hypercortisolism causes hypertension, left ventricular hypertrophy, and myocardial ischemia (52)
. Nevertheless, the myocardial abnormalities are not fully explained by the secondary changes due to ischemia and hypertension and suggest a direct detrimental effect of excess cortisol on the myocardium (6
, 7)
. Activation of AMPK during ischemia (53)
also lowers malonyl-CoA and thus increases ATP generation via fatty acid oxidation during reperfusion (54)
. Recent results using mice expressing a dominant-negative AMPK mutant in the heart suggest that the presence of AMPK protects cardiac ATP levels, and reduces infarct size and damage to myocytes, during ischemia (55)
. In our study, the decrease in myocardial AMPK activity induced by a glucocorticoid suggests that the deleterious effect of corticosteroids on the heart could, at least in part, be mediated by the decrease in AMPK activity.
Corticosteroids provide the link between stress, the metabolic syndrome, and mortality (27)
and may also mediate the adverse effects of low birth weight in programming obesity and insulin resistance in later life (56)
. Support from these data for a link between the metabolic syndrome, obesity, and AMPK could aid in the development of new treatments for these highly prevalent disorders. A recognized link between AMPK and cortisol could improve the development of safer forms of corticosteroid therapy and improve therapeutic possibilities for the treatment of Cushing’s syndrome. Data from Europe (57
58
59)
suggest that 0.5–0.9% of the population is receiving chronic (>3 months) glucocorticoid treatment. Therefore, together with patients with pituitary, adrenal, or ectopic Cushing’s syndrome, more than half a million people in the UK and more than 2.5 million people in the USA are currently exposed to long-term glucocorticoid effects and would thus be candidates for the prevention of glucocorticoid-induced metabolic changes.
Our demonstration of the discordant effects of corticosteroids on hypothalamic, adipose, hepatic, and cardiac tissues suggests the importance of tissue-specific pharmacological modulation of AMPK activity. Such pharmacological agents might not only play a role in prevention or reversal of the deleterious metabolic consequences of excessive glucocorticoid activity, similar, for example, to the bisphosphonate role in the prevention of glucocorticoid-induced osteoporosis, but may also be relevant to the much bigger problem of the metabolic syndrome affecting more than 30% of the population of developed societies.
In summary, our data suggest that glucocorticoid-induced changes in AMPK constitute a novel mechanism that could explain the increase in appetite, the deposition of lipids in visceral adipose and hepatic tissue, as well as the cardiac changes characteristic of glucocorticoid excess. Our data suggest that metformin treatment could be effective in preventing at least some of the metabolic complications of chronic glucocorticoid excess.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication July 17, 2007. Accepted for publication December 6, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-aminobutyric acid inputs to hypothalamic magnocellular neurons. Endocrinology 146,4292-4301This article has been cited by other articles:
|
|
 |
|
  S. Palomba, A. Falbo, F. Zullo, and F. Orio Jr. Evidence-Based and Potential Benefits of Metformin in the Polycystic Ovary Syndrome: A Comprehensive Review Endocr. Rev., February 1, 2009; 30(1): 1 - 50. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Kola, M. Christ-Crain, F. Lolli, G. Arnaldi, G. Giacchetti, M. Boscaro, A. B. Grossman, and M. Korbonits Changes in Adenosine 5'-Monophosphate-Activated Protein Kinase as a Mechanism of Visceral Obesity in Cushing's Syndrome J. Clin. Endocrinol. Metab., December 1, 2008; 93(12): 4969 - 4973. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Shimizu, H. Arima, M. Watanabe, M. Goto, R. Banno, I. Sato, N. Ozaki, H. Nagasaki, and Y. Oiso Glucocorticoids Increase Neuropeptide Y and Agouti-Related Peptide Gene Expression via Adenosine Monophosphate-Activated Protein Kinase Signaling in the Arcuate Nucleus of Rats Endocrinology, September 1, 2008; 149(9): 4544 - 4553. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
