| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
Unitat de Diabetologia, Endocrinologia i Nutricio, University Hospital of Girona, Dr. Josep Trueta, 17007 Girona, Spain; Laboratoire de la Fédération d’Endocrinologie, Hôpital de l’Antiquaille, Hospices Civils de Lyon, and INSERM, U-329, 69321 Lyon, France; and Unitat d’Endocrinologia, Hospital of Tarragona Joan XXIII, Facultat Medicina, Institut d’Estudis Avançats, Universitat Rovira i Virgili, 43007 Tarragona, Spain
Address all correspondence and requests for reprints to: J. M. Fernandez-Real, M.D., Ph.D., Unitat de Diabetologia, Endocrinologia i Nutricio, University Hospital of Girona Dr. Josep Trueta, Carretera de França s/n, 17007 Girona, Spain. E-mail: endocrino{at}htrueta.scs.es.
Abstract
It has been suggested that a low grade inflammatory state could predispose for developing insulin resistance and contribute to the development of obesity and type 2 diabetes. Corticosteroid-binding globulin (CBG), the main plasma protein transport for cortisol, has been shown to be negatively regulated by insulin and IL-6, at least in vitro, suggesting that insulin resistance and inflammation may both contribute to decreasing CBG levels. In the present study we measured CBG concentrations in a human healthy population and investigated the relationships of CBG with anthropometric and biochemical markers for inflammation and/or insulin resistance.
The data showed that the mean serum CBG level was significantly lower in males (n = 151) than in females (n = 113; 32.5 ± 9.1 vs. 39.2 ± 13.9 mg/liter; P < 0.0001). In both sexes serum CBG levels were correlated negatively with age (r = -0.12; P = 0.04), body mass index (r = -0.31; P < 0.0001), waist to hip ratio (WHR; r = -0.39; P < 0.0001), systolic (r = -0.15; P < 0.01) and diastolic (r = -0.15; P = 0.01) blood pressures, and HOMA, an index of insulin resistance (r = -0.12; P = 0.04). In addition, the CBG concentration was negatively associated with serum IL-6 concentrations (r = -0.23; P = 0.017) and with the soluble fraction of TNF
receptors, soluble TNF receptor 1 (sTNFR1; r = -0.35; P < 0.0001), and sTNFR2 (r = -0.56; P < 0.0001) in women. A stepwise regression analysis using CBG as an independent variable showed that sex (P < 0.00001), body mass index (P = 0.0002), and HOMA (P = 0.0005), but not systolic blood pressure, diastolic blood pressure, IL-6, sTNFR1, or sTNFR2, constituted significant independent factors that explained 21% of the CBG variance (14%, 2%, and 5%, respectively).
In a subsample of 120 men and 68 women, fasting serum free cortisol (calculated as the ratio fasting cortisol/CBG) was significantly associated with WHR (r = 0.24; P = 0.001), systolic (r = 0.18; P = 0.01) and diastolic (r = 0.19; P = 0.007) blood pressures, and HOMA value (r = 0.20; P = 0.005), but not with BMI or age.
BMI (P < 0.0001), free cortisol (P = 0.003), and CBG (P = 0.009), but not WHR and age, contributed to 20%, 6%, and 8%, respectively, of HOMA variance in women in a multiple regression analysis. In this model only BMI (P < 0.0001) independently contributed to HOMA variance in men.
These findings support the hypothesis that the CBG level is an interesting indicator for both insulin resistance and low grade inflammation. Whether the decrease in CBG levels is genetic by nature or directly associated to increased insulin and/or IL-6 merits further investigation. Nevertheless, because CBG has been shown to be expressed by the adipose tissue, decreased CBG could create locally increased cortisol disposal, with no change in circulating cortisol, and facilitate fat accumulation, insulin resistance, and type 2 diabetes.
SEVERAL LINES OF evidence suggest that obesity and insulin resistance per se constitute a low grade inflammatory state (1). A positive association between anthropometric indexes of obesity and plasma IL-6 levels has been described in men and postmenopausal women (2, 3, 4). Elevated C-reactive protein levels, an acute phase reactant, have been consistently shown in overweight and obese adults as well as in young adults (5). Our current working hypothesis is that some subjects may have a particular inflammatory genotype/phenotype that could predispose them for developing insulin resistance and type 2 diabetes (1). In this aim the early identification of inflammatory markers could be helpful for preventing complications associated with obesity and insulin resistance and could open new strategies for treatment.
Corticosteroid-binding globulin (CBG) is the major blood transport protein for cortisol in humans. The few data in the literature have suggested that CBG is a negative acute phase reactant (6, 7). Indeed, dramatically low CBG levels have been reported in patients with septic shock (8) and severe burn injuries (9). In addition, low CBG levels correlated negatively to increased IL-6 in burn patients (9). These data are consistent with the negative effect of IL-6 on CBG synthesis by a human hepatoma-derived (HepG2) cell line (10, 11). In addition, the inhibitory effect of IL-6 infusion on serum CBG in man strongly supports the concept that IL-6 is a negative regulator of CBG (12). In contrast, TNF has no effect on CBG mRNA and secretion in HepG2 cell lines (11). On the other hand, using the HepG2 cell line model insulin has been shown to be a potent negative regulator of CBG secretion and of CBG mRNA as well as for SHBG, the main transport system for sex steroid hormones (13).
We, therefore, studied the relationship of serum CBG concentration with the usual parameters of a proinflammatory phenotype and with parameters of insulin resistance in an apparently healthy population from Catalonia, Spain.
Subjects and Methods
Subjects
Two hundred sixty-four subjects (113 women and 151 men) of Caucasian origin were evaluated between July 1 and September 15, 1997, as a part of an on-going epidemiological study in Catalonia. None of the subjects was taking any medication or had evidence of metabolic disease, although some were obese. Body weight had been stable for at least 3 months before the study, and the subjects were normotensive.
Anthropometric measurements
All subjects were evaluated for body mass index (BMI), calculated as weight (in kilograms) divided by height (in meters) squared, and the waist to hip ratio (WHR). Each subject’s waist was measured with a soft tape midway between the lowest rib and the iliac crest. Hip circumference was measured at the widest part of the gluteal region.
Inclusion criteria were 1) BMI below 40 kg/m2, 2) absence of any systemic disease, 3) absence of any infections in the previous month, and 4) serum glucose lower than 6.6 mmol/liter. Smokers were defined as any person consuming at least one cigarette per day in the previous 6 months. Resting blood pressure was measured after the subject had been in a sitting position for a minimum of 15 min. Using a mercury sphygmomanometer, blood pressure was measured three times in the right arm by the same investigator. The mean of three measurements was used for this study. None of the subjects was taking any medication (including glucocorticoids or estrogens) or had any evidence of metabolic disease other than obesity. Liver disease and thyroid dysfunction were specifically excluded by biochemical work-up. All women had regular menstrual cycles. The protocol was approved by the hospital ethics committee, and informed consent was obtained from each subject.
Analytical methods
The serum glucose concentration was measured in duplicate by the glucose oxidase method. The serum insulin level was measured in duplicate by monoclonal immunoradiometric assay (Medgenix Diagnostics, Fleunes, Belgium). The lowest limit of detection was 4.0 mU/liter. The intraassay coefficient of variation was 5.2% at a concentration of 10 mU/liter. The interassay coefficient of variation was 6.9% at 14 mU/liter. The fasting insulin resistance index (HOMA) was calculated using the formula: HOMA = fasting glucose (mmol/liter) x fasting insulin (mU/liter)/22.5. In our experience (14), HOMA fairly correlates with the insulin sensitivity index calculated using the minimal model approach (r = 0.79; P < 0.0001).
Total serum cholesterol was measured through the reaction of cholesterol esterase/cholesterol oxidase/peroxidase. Total serum triglycerides were measured through the reaction of glycerol-phosphate-oxidase and peroxidase. Serum uric acid (Beckman, Fullerton, CA) was determined by routine laboratory tests.
The plasma CBG concentration was measured by RIA (Radim, KP31, Angleur, Liege, Belgium). Intra- and interassay coefficients of variation were 3.6% and 7.5%, respectively. Serum cortisol was measured by RIA (8, 9). The ratio of cortisol/CBG was used to calculate free cortisol.
The MEDGENIX soluble TNF receptor 1 (sTNFR-1) and sTNFR-2 EASIA (Biosource Technologies, Inc., Fleunes, Belgium) are solid phase enzyme-amplified sensitivity immunoassays performed on a microtiter plate. The minimum detectable concentration was 0.1 ng/ml and was defined as the sTNFR1 or sTNFR2 concentration corresponding to the average OD of 20 replicates of the zero standard ± 2 SD. The intra- and interassay coefficients of variation were less than 7% and less than 9%. The sTNFR1 EASIA does not cross-react with sTNFR-2. TNF does not interfere with the assay.
Serum IL-6 was measured using an immunoassay (MEDGENIX IL-6 EASIA, Biosource Technologies, Inc.), with coefficients of variation lower than 6%.
Statistical analysis
Descriptive results of continuous variables are expressed as the mean ± SD. Before statistical analysis, normal distribution and homogeneity of the variances were tested. Parameters that did not fulfill these tests (IL-6, sTNFR1, and sTNFR2) were log-transformed. Comparison of proportions was made using 
2 test. Comparison of variables between groups of subjects was performed by t test. Relationships between variables were sought by linear correlation analysis (Spearman’s r) and stepwise multivariate linear regression analysis. Levels of statistical significance were set at P less than 0.05.
Results
The anthropometric and biochemical characteristics of the subjects at the time of entry into the study are shown in Table 1
. Men and women were similar in age. The mean BMI and WHR (P = 0.011 and P < 0.0001, respectively), serum glucose (P = 0.002), cholesterol (P = 0.015), and triglycerides (P < 0.0001), but not serum insulin (P = 0.43), were significantly higher in men than in women. Women had significantly higher mean serum CBG concentrations than men (P < 0.0001). Samples from 120 men and 68 women (who were not significantly different from the remaining subjects regarding anthropometric and general characteristics) were available to measure serum cortisol, which was significantly higher in men (P = 0.005). The serum free cortisol index was also significantly higher in men (P < 0.0001).
|
), WHR (r = -0.39; P < 0.0001), systolic (SBP; r = -0.15; P = 0.01) and diastolic (DBP; r = -0.15; P = 0.01) blood pressures, fasting glucose (r = -0.19; P = 0.002), fasting insulin resistance index (HOMA; r = -0.12; P = 0.04; Fig. 2), and serum uric acid (r = -0.24; P < 0.0001). Fasting triglycerides tended to be associated with CBG, but not significantly (r = -0.11; P = 0.07). However, gender significantly affected these associations. Indeed, the serum CBG concentration was negatively associated with log IL-6 (r = -0.23; P = 0.017), log sTNFR1 (r = -0.35; P < 0.0001), and log sTNFR2 (r = -0.56; P < 0.0001) in women, whereas in men only the correlation between log sTNFR2 and CBG was significant (r = -0.18; P = 0.027).
|
|
). In contrast, women with low CBG levels (n = 57) had higher BMI (P < 0.011) and WHR (P < 0.003) and were more frequently obese (BMI, 
30 kg/m2) than women with higher than median CBG levels [21% (12 of 56) vs. 7% (4 of 56); 2 = 4.68; P = 0.03] and had increased log sTNFR1 and log sTNFR2 levels (Table 3). Although the serum CBG concentration correlated significantly with BMI in women, the association between HOMA and CBG was significant when the entire population was analyzed (Fig. 2
). In addition, subjects with a CBG below the median (33.8 mg/dl) had a significantly higher HOMA value (1.75 ± 1.18 vs. 1.50 ± 0.66; P = 0.04).
|
|
In a subsample of 120 men and 68 women, fasting serum free cortisol (calculated as the fasting cortisol/CBG ratio) was significantly associated with CBG levels (r = -0.27; P = 0.0001), WHR (r = 0.24; P = 0.001), SBP (r = 0.18; P = 0.01), DBP (r = 0.19; P = 0.007), and HOMA value (r = 0.20; P = 0.005), but not wit BMI or age.
BMI (P < 0.0001), free cortisol (P = 0.003), and CBG (P = 0.009), but not WHR and age, contributed to 20%, 6%, and 8%, respectively, of HOMA variance in women in a multiple regression analysis. In contrast, in men only BMI (P < 0.0001) independently contributed to HOMA variance in men.
Discussion
In our apparently healthy population CBG was found to be negatively associated with indexes of insulin resistance such as BMI, WHR, and HOMA and with inflammatory parameters such as sTNFR1, sTNFR2, and IL-6 concentrations. Vgontzas et al. (15) have also recently reported that visceral obesity was an additional determinant of inflammatory cytokine elevation.
We found that free cortisol was significantly associated with several features of the insulin resistance syndrome as classically described and was negatively associated with CBG itself, linking at cortisol-induced CBG down-regulation (11). However, at least in women, both decreased CBG and free cortisol independently contributed to HOMA variance in a multiple linear regression analysis, suggesting that both mechanisms would be metabolically additive.
Because CBG secretion has been shown to be negatively regulated by both insulin (13) and IL-6 (10, 11), it is tempting to propose that the CBG concentration is an index of insulin resistance and inflammation. On the other hand, constitutive low CBG levels might also contribute to insulin resistance by increasing cortisol biodisposal to target cells, including muscular cells. Interestingly, this paradigm was different in men and women. Indeed, we found sexual dimorphism for CBG levels, with women having higher CBG levels than men, who also had a weaker correlation of CBG with inflammatory indexes than women. A gender difference in CBG levels was previously reported by De Moor et al. (16), who also showed a bimodal distribution of CBG values in women, but not in men. Estrogens are well known positive regulators of CBG by an unknown mechanism that could implicate increased CBG gene transcription and/or altered glycosylation of the CBG molecule (7). Therefore, the gender difference in CBG levels remains unexplained. De Moor et al. (16) reported a statistical excess of low CBG values in the relatives of subjects with low CBG, suggesting that the CBG concentration was indeed a sex-linked trait (16). A gender difference in the inflammatory response has been well documented in rodents (17, 18). In humans some gender difference in the associations of IL-6 levels have been suggested by Straub et al. (4), who made the interesting observation that serum IL-6 level was an independent predictor of blood pressure in women, but not in men. Similar findings in the sex-linked relationships between IL-6 and blood pressure and between IL-6 and CBG were found in this study.
In our population insulin resistance (HOMA index) was negatively associated with serum CBG levels independently of body fatness in both sexes, and CBG levels were negatively associated with IL-6 levels only in women. Men are more insulin resistant than women, so the relationship of CBG is expected to be stronger with insulin resistance in men than in women. These results extend our previous observation showing that the CBG level was negatively associated with insulin secretion in obese and glucose-tolerant subjects who also had lower CBG levels than obese and glucose-intolerant subjects (19). We have also shown that both insulin secretion and insulin sensitivity independently contributed to changes in CBG after glucose-induced insulin stimulation (20). In addition, a polymorphism of the IL-6 gene promoter, which is linked to increased IL-6 levels and insulin resistance, was associated with low CBG levels (21).
A consensus responsive element for IL-6 has been identified within the CBG gene promoter (7). The prolonged suppression of CBG levels during and after IL-6 infusion in normal men (12) and the suppressive effect of IL-6 on CBG secretion by HepG2 cell lines (10, 11) are strong arguments that IL-6 is a negative regulator of CBG. A negative relationship of circulating CBG and IL-6 levels has been reported in burn patients (9), who had increased CBG levels and decreased IL-6 levels during a low fat diet, leading to increased survival after sepsis. Because a low fat diet is associated with improved insulin sensitivity (22), these data and the findings of the present study strongly suggest that CBG could part of a network of signals regulating insulin resistance and inflammation.
On the other hand, CBG is a member of the SERPIN (serine protease inhibitors) family (6, 7, 23) and is a substrate for elastase that is expressed at the surface of neutrophils. Therefore, the variability in circulating CBG might be linked to CBG cleavage by activated neutrophils. This mechanism in part disrupts the steroid-binding site from CBG-bound sites (23). Increased peripheral white blood cell count and neutrophils are usually found in both obesity (24) and insulin resistance (21), which might facilitate serine protease availability and CBG cleavage. This mechanism is likely to contribute to decreased serum CBG levels in obesity and insulin resistance.
We have measured circulating sTNFR concentrations because TNF is one proinflammatory cytokine implicated in the phenotypic expression of obesity and insulin resistance (1). It has been shown that sTNFR levels remain elevated for longer periods of time after the administration of TNF and are a valuable indicator of the effects of TNF (25, 26, 27, 28). The association between CBG and sTNFRs could be incidental rather than causal, as TNF had no effect on CBG secretion by HepG2 cell lines (11), but it still suggests that both sTNFRs and CBG are a part of a mild chronic inflammation syndrome.
A CBG gene mutation that decreased CBG levels by 50% was recently described in a 43-yr-old woman with complaints of chronic fatigue and obesity (29). These clinical findings have been recently confirmed in a large kindred as the likely phenotype of genetic deficiency in circulating CBG (30). These observations suggest that low CBG might facilitate the development of obesity by a mechanism that remains to be explained. However, it is interesting that in conditions of extremely low CBG, free urinary cortisol remains normal, suggesting that in normal circumstances, the hypothalamic-pituitary-adrenal axis is physiologically regulated by negative feedback of free (CBG-unbound) cortisol, and this overcomes CBG deficiency to maintain cortisol production within the normal range. It remains to be demonstrated whether a low grade inflammatory state could be associated with impaired adrenal function that deviates from normal feedback regulation.
Rosmond et al. (31) found that stress-related cortisol secretion, an index derived from self-reported stress and salivary cortisol levels, was associated with several components of the insulin resistance syndrome. Interestingly, this association was stronger in subjects with a flattened diurnal cortisol curve; a flattened CBG curve was observed in obesity-associated insulin resistance (20). As cortisol suppresses CBG (11), this could be an additional mechanism that would explain the findings of this study. In this respect we have previously studied CBG at various time intervals and found that CBG remained essentially unchanged during a 3-h period in normal subjects. We also reported that the associations among CBG, insulin secretion, and insulin sensitivity differed between lean and obese subjects, suggesting that CBG was independently associated with glucose metabolism (20).
In conclusion, a decreased serum CBG concentration is commonly found associated with insulin resistance and some markers of inflammation within an apparently healthy population. As shown for low serum albumin in the Atherosclerosis Risk in Communities Study (32), we suggest that low CBG could be an interesting index for the development of type 2 diabetes as well as for the incidence of cardiovascular disease, as previously reported in postmenopausal women (33).
Acknowledgments
Footnotes
Abbreviations: BMI, Body mass index; CBG, corticosteroid-binding globulin; DBP, diastolic blood pressure; EASIA, enzyme-amplified sensitivity immunoassay; HOMA, fasting insulin resistance index; SBP, systolic blood pressure; sTNFR1 or 2, soluble TNF receptor 1 or 2; WHR, waist to hip ratio.
Received November 19, 2001.
Accepted July 9, 2002.
References
, in vivo. J Clin Endocrinol Metab 82:4196–4200This article has been cited by other articles:
 |
|
 |
|
  D. L. Batista, N. Courcoutsakis, J. Riar, M. F. Keil, and C. A. Stratakis Severe Obesity Confounds the Interpretation of Low-Dose Dexamethasone Test Combined with the Administration of Ovine Corticotrophin-Releasing Hormone in Childhood Cushing Syndrome J. Clin. Endocrinol. Metab., November 1, 2008; 93(11): 4323 - 4330. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Qi and B. Rodrigues Glucocorticoids produce whole body insulin resistance with changes in cardiac metabolism Am J Physiol Endocrinol Metab, March 1, 2007; 292(3): E654 - E667. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. H. Petersen, T. K. Andreassen, T. Breiderhoff, J. H. Brasen, H. Schulz, V. Gross, H.-J. Grone, A. Nykjaer, and T. E. Willnow Hyporesponsiveness to glucocorticoids in mice genetically deficient for the corticosteroid binding globulin. Mol. Cell. Biol., October 1, 2006; 26(19): 7236 - 7245. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Guyonnet-Duperat, N. Geverink, G. S. Plastow, G. Evans, O. Ousova, C. Croisetiere, A. Foury, E. Richard, P. Mormede, and M.-P. Moisan Functional Implication of an Arg307Gly Substitution in Corticosteroid-Binding Globulin, a Candidate Gene for a Quantitative Trait Locus Associated With Cortisol Variability and Obesity in Pig Genetics, August 1, 2006; 173(4): 2143 - 2149. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. T. Ho, H. Al-Musalhi, M. J. Chapman, T. Quach, P. D. Thomas, C. J. Bagley, J. G. Lewis, and D. J. Torpy Septic Shock and Sepsis: A Comparison of Total and Free Plasma Cortisol Levels J. Clin. Endocrinol. Metab., January 1, 2006; 91(1): 105 - 114. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. M. Kurina, L. A. Weiss, S. W. Graves, R. Parry, G. H. Williams, M. Abney, and C. Ober Sex Differences in the Genetic Basis of Morning Serum Cortisol Levels: Genome-Wide Screen Identifies Two Novel Loci Specific to Women J. Clin. Endocrinol. Metab., August 1, 2005; 90(8): 4747 - 4752. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Fernandez-Real, A. Lopez-Bermejo, M. Broch, J. Vendrell, C. Richart, and W. Ricart Circulating Soluble CD14 Monocyte Receptor Is Associated with Increased Alanine Aminotransferase Clin. Chem., August 1, 2004; 50(8): 1456 - 1458. [Full Text] [PDF]
|
|
|
|
 |
|
 O. Ousova, V. Guyonnet-Duperat, N. Iannuccelli, J.-P. Bidanel, D. Milan, C. Genet, B. Llamas, M. Yerle, J. Gellin, P. Chardon, et al. Corticosteroid Binding Globulin: A New Target for Cortisol-Driven Obesity Mol. Endocrinol., July 1, 2004; 18(7): 1687 - 1696. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Ten and N. Maclaren Insulin Resistance Syndrome in Children J. Clin. Endocrinol. Metab., June 1, 2004; 89(6): 2526 - 2539. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Fernandez-Real and W. Ricart Insulin Resistance and Chronic Cardiovascular Inflammatory Syndrome Endocr. Rev., June 1, 2003; 24(3): 278 - 301. [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 |
