Obesity (2008) 16 10, 2374–2378. doi:10.1038/oby.2008.346
- 1Department of Endocrinology, Newcastle University Teaching Hospitals, Newcastle upon Tyne, UK
- 2Northern Institute for Cancer Research, Newcastle University, Newcastle upon Tyne, UK
- 3School of Clinical Medical Sciences, Newcastle University, Newcastle upon Tyne, UK
- 4Department of Diabetes and Endocrinology, Queen Elizabeth Hospital, Gateshead, UK
Correspondence: Christopher P.F. Redfern (firstname.lastname@example.org)
Received 21 September 2007; Accepted 10 March 2008; Published online 24 July 2008.
Clinical similarities between Cushing's syndrome and obesity/metabolic syndrome have led to speculation of a role for glucocorticoids (GCs) in the etiopathogenesis of obesity. People with idiopathic obesity have normal circulating cortisol concentrations. However, there may be considerable interindividual variation in GC sensitivity. The objective of this study was to determine whether enhanced GC sensitivity in the absence of GC excess was a characteristic of obese people with cushingoid features. We studied 12 obese subjects with cushingoid features in the absence of Cushing's syndrome and six slim control participants. Data recorded included BMI, waist-to-hip ratio, blood pressure, glucose and insulin response to 75 g oral glucose challenge, and low-dose (0.25 mg) overnight dexamethasone (DEX) suppression test (ODST-0.25 mg). To study GC-sensitivity in vitro, we performed dose–response studies of DEX-induced suppression of interleukin-6 (IL-6) secretion in skin fibroblast cultures. Seven obese subjects were normosensitive and five subjects hypersensitive to GCs in vitro. ODST-0.25 mg resulted in a median suppression of cortisol from baseline of 32% in normosensitive and 60% in hypersensitive obese subjects (P < 0.004). No other clinical or biochemical measures were discriminatory between these two groups. These data from two independent measures of GC sensitivity suggest that enhanced GC sensitivity may characterize a substantial proportion of obese people with cushingoid appearance.
The role of glucocorticoids (GCs) in promoting and maintaining positive energy balance, a function that is exaggerated in Cushing's syndrome, has been suspected in the etiopathogenesis of simple obesity. People with idiopathic obesity have normal circulating cortisol concentrations that are independent of fat stores but an augmented cortisol turnover rate (1,2) and enhanced hypothalamic–pituitary–adrenal (HPA) axis reactivity (3,4,5,6). There is considerable interindividual variation in GC sensitivity in health (7,8). This is also evident in some patients who develop Cushing's syndrome on low-dose GC therapy for conditions such as asthma, whereas others appear to be resistant to much higher doses. Thus, it has been speculated that obesity with Cushingoid appearance is a state of enhanced responsiveness to GCs. The aim of this study was to test the hypothesis that adiposity was associated with enhanced GC-sensitivity in vitro and in vivo in obese individuals with a cushingoid appearance.
Participants and Methods
We studied 12 obese subjects (10 women) aged 18–53 (median 29.5) years who had cushingoid features such as central obesity and skin striae on clinical assessment, suggestive of enhanced GC activity; none had Cushing's syndrome (at least two normal 24-h urinary free cortisol measurements) or a history of treatment with steroids, hypothyroidism, or any major illness. For comparative analysis, we studied six healthy control participants (four women) aged 33–49 (median 40) years with BMI <25 kg/m2. All participants gave informed signed consent and the local research ethics committee approved the study. Data recorded included age, gender, weight, height, BMI, waist circumference, hip circumference, waist-to-hip ratio, blood pressure, and bioimpedance analysis of body composition. A 75-g oral glucose tolerance test was performed after an overnight fast with measurements of fasting and 2-h glucose and insulin levels. Insulin resistance was computed by homeostasis model assessment (9). An overnight dexamethasone (DEX) suppression test (ODST) was performed on a subsequent day with 0.25 mg DEX administered at 2300 hours and serum samples collected at 0900 hours the following morning. Cortisol, corticosteroid-binding globulin, prolactin, adiponectin, resistin, leptin, interleukin-6 (IL-6), testosterone, sex hormone binding globulin, luteinizing hormone, follicle-stimulating hormone, and estradiol concentrations were measured at baseline (0900 hours fasting sample of oral glucose tolerance test); cortisol, prolactin, adiponectin, resistin, leptin, and DEX concentrations were also measured following ODST-0.25 mg. Free cortisol index, calculated as total cortisol (nmol/l)/corticosteroid-binding globulin (mg/l), was used as a surrogate for serum-free cortisol (10). The proportional change in concentrations before and after ODST-0.25 mg was used in analyses. Study participants were not on estrogenic or cytochome P450-metabolized drug therapies and gonadal hormone profiles in menstruating women subjects indicated that they were in the follicular phase of the menstrual cycle when biochemical tests were carried out.
GC-sensitivity in vitro
We developed a novel bioassay for determining GC-sensitivity in vitro by quantifying DEX-induced suppression of unstimulated, natural state IL-6 secretion in cultured human skin fibroblasts (HSFs). Forearm skin biopsies were taken with a 3 mm Stiefel disposable punch. Primary cultures were set up on Matrigel (BD Biosciences, Oxford, UK) bathed in MEM-Alpha with Glutamax-I (50%), fetal bovine serum (50%), penicillin (5,000 units), streptomycin (5,000 g), and amphotericin 0.1% (all from Invitrogen, Paisley, UK) at 37 °C in an atmosphere of 5% CO2. HSFs were propagated in 90% DMEM with Nutrient Mixture F-12 with Glutamax-I (Invitrogen) and 10% fetal bovine serum and passaged by trypsinisation. HSFs harvested at Passage 5–8 were grown to full confluence in 96-well culture plates. The wells were exposed to DEX in doses ranging from 10–11 to 10–4 M for 1 h followed by rinsing and incubation in serum-free media overnight for 12 h. Cell supernatants were aspirated and analyzed in duplicate for IL-6 using a human IL-6 ELISA (Bender MedSystems, Vienna, Austria) as per the manufacturer's instructions. Control solutions containing high and low concentrations of IL-6 were used in every batch. A coefficient of variation <20% was accepted for reproducibility of the assay (11), and the overall mean coefficient of variation was 4.58% in this study. Under these conditions, DEX produced a clear dose–response decrement in the IL-6 content of culture supernatants and the IC50 (median inhibitory concentration) was computed by four-parameter nonlinear regression sigmoidal dose–response with variable slope curve fitting. The lower-bound value of the 97% confidence interval of the median IC50 in controls was used as the cutoff for classifying obese subjects as normosensitive (IC50 4.4 nmol/l) or hypersensitive (IC50 < 4.4 nmol/l) to GCs. Statistical analyses were performed with Prism v4 (GraphPad Software, San Diego, CA), SPSS v11 (SPSS, Chicago, IL), and Systat v10 (Systat Software, San Jose, CA). We used conservative, nonparametric statistics (Mann–Whitney U-test, asymptotic two-tailed significance) for comparative analysis. For testing the association of cortisol response to ODST-0.25 mg and in vitro GC sensitivity, stepwise backward multiple regression analysis was performed with an F-ratio probability of 0.15 as the criterion for removal or inclusion of explanatory variables in the model. Pearson's correlation coefficient was determined as a measure of correlation between variables. Receiver operating characteristic analysis was used for defining sensitivity and specificity of cortisol response to ODST-0.25 mg in relation to GC-sensitivity in vitro. A P value <0.05 was accepted as significant.
GC-sensitivity in vitro
GC-sensitivity in vitro was represented by the IC50 of DEX-induced inhibition of IL-6 secretion in HSFs (Figure 1a). The median IC50 for controls was 9.7 nmol/l (97% confidence interval, 4.4–22.3 nmol/l). Seven obese subjects were normosensitive (obese normosensitive, ONS; Figure 1b) with median IC50 12.6 nmol/l (compared to controls, P = 0.09). The remaining five obese subjects were hypersensitive (obese hypersensitive, OHS); the median IC50 was 0.5 nmol/l and significantly different from control participants (P < 0.006) and ONS subjects (P < 0.004). There were no significant differences in clinical measures including the degree and distribution of adiposity, blood pressure, fasting and 2-h glucose during oral glucose tolerance test, homeostasis model assessment of insulin resistance, and lipid profiles between ONS and OHS subjects (Table 1).
Glucocorticoid (GC)-sensitivity in vitro and in vivo. (a) GC- sensitivity in vitro. Dose–response curves showing fall in IL-6 concentration (y-axis; percentage of baseline value) plotted against increasing dexamethasone (DEX) concentrations (x-axis; log scale); thick solid line, median dose–response curve for control participants (CON), flanked by 95% confidence intervals of the curve (gray shaded area) for illustration (97% confidence interval of the median was used for determining the reference range); thin solid lines, dose–response curves for obese-hypersensitive (OHS) subjects; broken lines, dose–response curves for obese-normosensitive (ONS) subjects; IC50 is marked by dotted horizontal line. (b) Median inhibitory concentration. IC50 values for DEX-induced interleukin-6 (IL-6) suppression in vitro (y-axis; log scale), grouped by participant type. (c) GC-sensitivity in vivo. Percent suppression in cortisol concentration from baseline after overnight 0.25 mg DEX suppression test (y-axis), grouped by participant type. (d) Correlation of in vivo and in vitro GC sensitivity. Linear regression of cortisol response to overnight 0.25 mg DEX suppression test (y-axis) against IC50 values of DEX concentration (x-axis; log scale) confirmed significant correlation (Pearson's r = 0.6, P < 0.009). Open circles, female; filled circles, male; horizontal lines, median value.Full table (55K)
Overnight 0.25 mg DEX suppression test (ODST-0.25 mg)
There were no significant differences between the groups in serum DEX concentrations after ODST-0.25 mg (Table 1). Although there were no significant differences in baseline serum total cortisol concentrations or free cortisol index, the median-cortisol response to ODST-0.25 mg was 24.1% in control participants, 32.4% in ONS and 60.0% in OHS subjects (Figure 1c). The cortisol responses in OHS subjects were significantly different compared to ONS subjects (P < 0.004). Stepwise backward multiple regression performed to further test the relationship of cortisol response (dependent variable) to in vitro GC sensitivity in obese subjects with age, BMI, waist-to-hip ratio, baseline cortisol, corticosteroid-binding globulin, and serum DEX concentration after ODST-0.25 mg as additional explanatory variables confirmed a significant association (P < 0.001; Table 2). The IC50 of DEX-induced inhibition of IL-6 secretion in HSFs correlated significantly with cortisol response to ODST-0.25 mg (Pearson's r = 0.6, P < 0.009; Figure 1d). On receiver operating characteristic analysis, a cutoff of 50% suppression in cortisol from baseline in response to ODST-0.25 mg gave a sensitivity of 100% and specificity of 92% for distinguishing OHS from ONS subjects.Full table (27K)
Insulin, prolactin, adipokines, and gonadal hormones
Insulin resistance was greater in obese subjects (1.55) than in controls (0.5; P < 0.04), but there was no significant difference between the ONS (1.8) and OHS (1.1) subgroups (P > 0.51). Median basal concentrations in control participants compared to obese subjects of leptin were 7.68 ng/ml vs. 118.06 ng/ml (P < 0.001), resistin 3.60 ng/ml vs. 7.51 ng/ml (P < 0.025), and IL-6 0.70 pg/ml vs. 1.98 pg/ml (P < 0.002). However, there were no significant differences between ONS and OHS subjects in leptin (99.61 ng/ml vs. 140.24 ng/ml, P > 0.46), resistin (8.24 ng/ml vs. 6.78 ng/ml, P > 0.68), or IL-6 (1.70 pg/ml vs. 2.18 pg/ml, P > 0.93). There were no significant differences in basal concentrations of prolactin and adiponectin between any of the groups and there were no significant differences in the responses of leptin, resistin, adiponectin, IL-6, or prolactin to ODST-0.25 mg across the groups. After taking gender into account, there were no differences in serum concentrations of luteinizing hormone, follicle-stimulating hormone, sex hormone binding globulin, total and calculated bioavailable (free) testosterone, and estradiol between groups.
Obesity with cushingoid appearance, by sharing some clinical features of Cushing's syndrome, raises the possibility of hypersensitivity to normal circulating levels of endogenous GCs. Various bioassays have been described for measuring cellular GC sensitivity (8,12). Most rely on GC-induced suppression of cell proliferation or mitogen-stimulated cytokine release in peripheral blood cells, which may be affected by exposure to circulating GCs in vivo before harvesting, or by the stimulation process itself. We therefore developed a bioassay based on GC-induced suppression of IL-6 secretion in HSFs to measure GC sensitivity. This is unaffected by in vivo circulating GCs and reflects a stable, heritable property of the HSFs derived from each participant. We studied a sample of obese subjects who were preselected on clinical suspicion of GC hypersensitivity (cushingoid appearance with no evidence of GC excess) of which two-fifths manifested evidence of tissue GC hypersensitivity. Furthermore, the hypersensitive obese subjects were indistinguishable from normosensitive obese subjects by routine objective clinical measures such as body weight, BMI, waist circumference, or waist-to-hip ratio in this small sample. Thus, conventional clinical features alone are inadequate at identifying GC-hypersensitive individuals.
Although ODST-1 mg is often used as a screening test in clinical investigation of Cushing's syndrome, in people without Cushing's syndrome, it results in too much suppression of the HPA axis to allow detection of individual differences in feedback sensitivity. However, ODST-0.25 mg results in a broad range of cortisol concentrations, giving a good insight into feedback sensitivity (7) and low-dose DEX can distinguish a finely regulated HPA axis from a poorly regulated one (13). Earlier population studies employing low-dose overnight DEX suppression tests have reported considerable variation in GC sensitivity of the HPA axis, but have been unable to reliably identify enhanced GC sensitivity in obesity. By selecting subjects with morphological features suggestive of GC excess in the absence of true Cushing's syndrome, we have shown that there is a clear difference in regulation of cortisol concentration in vivo between normosensitive and hypersensitive obese people, thus maintaining circulating cortisol levels within normal limits. This autoregulation is also reflected by GC sensitivity at a cellular level. Whereas measurement of serum-free cortisol is technically difficult, expensive, and not widely available, free cortisol index has been shown to correlate strongly with measured free cortisol (10). Moreover, percent-cortisol response (proportional change from baseline) to ODST, as used in this study, is unaffected by whether total cortisol or free cortisol index is used. The results of this study on a small sample of obese individuals suggest that percent-cortisol response to ODST-0.25 mg may be used to identify those who are likely to have underlying enhanced GC sensitivity. However, this conclusion will need independent replication in larger cohorts before warranting incorporation as a test for GC sensitivity into standard clinical practice.
Leptin, an adipokine hormone involved in negative feedback regulation of food intake, is inducible by GCs both in normal weight and obese individuals (14). Conversely, it can directly inhibit GC secretion in the adrenal gland (15). Thus, it has been suggested that feedback loops exist at the tissue level regulating cortisol concentrations within the intraindividual physiological range, resulting in reduced negative feedback tone higher up in the HPA axis (16). Basal leptin, resistin, and IL-6 concentrations were significantly higher overall in obese subjects compared to control participants, but there was no statistical difference between ONS and OHS subjects in our small sample. Similarly, although obese subjects had greater insulin resistance overall, there was no statistical difference between ONS and OHS subjects. This suggests that insulin resistance is more a function of obesity per se rather than GC sensitivity, but study in a larger sample is warranted. A previous study has demonstrated evidence of variation in cortisol response to ODST-0.25 mg depending on the phase of the menstrual cycle in women (17). However, this is unlikely to have influenced the outcome of this study as the women were assessed in the same phase of the menstrual cycle. A reduction in the sensitivity of the HPA axis to GC feedback inhibition has been demonstrated in 70-year-old compared to 26-year-old subjects (18). Although the median age of control participants was 10 years greater than of obese subjects in our study, there was no difference in age between the ONS and OHS subgroups. Moreover, intraindividual GC sensitivity has been shown to be constant with aging (7).
In conclusion, we suggest that enhanced sensitivity to GCs may characterize excess adiposity in a proportion of obese people with cushingoid appearance. As clinical assessment alone may not have sufficient discriminatory value, an ODST-0.25 mg may be used as a screening test to identify people with GC hypersensitivity for future research.
The authors declared no conflict of interest.
- Dunkelman SS, Fairhurst B, Plager J, Waterhouse C. Cortisol metabolism in obesity. J Clin Endocrinol Metab 1964;24: 832–841. | PubMed | ChemPort |
- Hautanen A, Adlercreutz H. Altered adrenocorticotropin and cortisol secretion in abdominal obesity: implications for the insulin resistance syndrome. J Intern Med 1993;234: 461–469. | PubMed | ChemPort |
- Marin P, Darin N, Amemiya T et al. Cortisol secretion in relation to body fat distribution in obese premenopausal women. Metabolism 1992;41: 882–886. | Article | PubMed | ChemPort |
- Weaver JU, Kopelman PG, McLoughlin L, Forsling ML, Grossman A. Hyperactivity of the hypothalamo-pituitary-adrenal axis in obesity: a study of ACTH, AVP, beta-lipotrophin and cortisol responses to insulin-induced hypoglycaemia. Clin Endocrinol (Oxf) 1993;39: 345–350. | Article | PubMed | ChemPort |
- Rosmond R, Dallman MF, Bjorntorp P. Stress-related cortisol secretion in men: relationships with abdominal obesity and endocrine, metabolic and hemodynamic abnormalities. J Clin Endocrinol Metab 1998;83: 1853–1859. | Article | PubMed | ISI | ChemPort |
- Pasquali R, Ambrosi B, Armanini D et al. Cortisol and ACTH response to oral dexamethasone in obesity and effects of sex, body fat distribution, and dexamethasone concentrations: a dose-response study. J Clin Endocrinol Metab 2002;87: 166–175. | Article | PubMed | ChemPort |
- Huizenga NA, Koper JW, de Lange P et al. Interperson variability but intraperson stability of baseline plasma cortisol concentrations, and its relation to feedback sensitivity of the hypothalamo-pituitary-adrenal axis to a low dose of dexamethasone in elderly individuals. J Clin Endocrinol Metab 1998;83: 47–54. | Article | PubMed | ChemPort |
- Chriguer RS, Elias LL, da Silva IM Jr, Vieira JG, Moreira AC, de Castro M. Glucocorticoid sensitivity in young healthy individuals: in vitro and in vivo studies. J Clin Endocrinol Metab 2005;90: 5978–5984. | Article | PubMed | ChemPort |
- Levy JC, Matthews DR, Hermans MP. Correct homeostasis model assessment (HOMA) evaluation uses the computer program. Diabetes Care 1998;21: 2191–2192. | Article | PubMed | ISI | ChemPort |
- le Roux CW, Sivakumaran S, Alaghband-Zadeh J, Dhillo W, Kong WM, Wheeler MJ. Free cortisol index as a surrogate marker for serum free cortisol. Ann Clin Biochem 2002;39: 406–408. | Article | PubMed | ChemPort |
- Rosner B. Fundamentals of biostatistics. 2006; Thomson-Brooks/Cole: Belmont, CA.
- Vermeer H, Hendriks-Stegeman BI, Verrijn Stuart AA, van Buul-Offers SC, Jansen M. A comparison of in vitro bioassays to determine cellular glucocorticoid sensitivity. Eur J Endocrinol 2004;150: 41–47. | Article | PubMed | ChemPort |
- Rosmond R, Bjorntorp P. The hypothalamic-pituitary-adrenal axis activity as a predictor of cardiovascular disease, type 2 diabetes and stroke. J Intern Med 2000;247: 188–197. | Article | PubMed | ChemPort |
- Dagogo-Jack S, Selke G, Melson AK, Newcomer JW. Robust leptin secretory responses to dexamethasone in obese subjects. J Clin Endocrinol Metab 1997;82: 3230–3233. | Article | PubMed | ChemPort |
- Bornstein SR, Uhlmann K, Haidan A, Ehrhart-Bornstein M, Scherbaum WA. Evidence for a novel peripheral action of leptin as a metabolic signal to the adrenal gland: leptin inhibits cortisol release directly. Diabetes 1997;46: 1235–1238. | Article | PubMed | ISI | ChemPort |
- Syed AA, Weaver JU. Glucocorticoid sensitivity: the hypothalamic- pituitary-adrenal-tissue axis. Obes Res 2005;13: 1131–1133. | Article | PubMed | ChemPort |
- Altemus M, Redwine L, Leong YM et al. Reduced sensitivity to glucocorticoid feedback and reduced glucocorticoid receptor mRNA expression in the luteal phase of the menstrual cycle. Neuropsychopharmacology 1997;17: 100–109. | Article | PubMed | ISI | ChemPort |
- Wilkinson CW, Peskind ER, Raskind MA. Decreased hypothalamic- pituitary-adrenal axis sensitivity to cortisol feedback inhibition in human aging. Neuroendocrinology 1997;65: 79–90. | Article | PubMed | ChemPort |
Funding for this study was provided by Gateshead NHS Research and Development Fund and the Queen Elizabeth Hospital Diabetes Charitable Fund. We are grateful to Lorna Ingoe (Queen Elizabeth Hospital and Newcastle University) for assisting in the conduct of clinical assessments, John Barker (Queen Elizabeth Hospital) for clinical biochemistry services, and Penny Lovat (Newcastle University) for helpful discussions on tissue culture technique.