1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. References


Serum cortisol concentrations fluctuate in a circadian fashion, and glucocorticoids exert strong effects on adipose tissue and induce obesity through the glucocorticoid receptor.

Design and Methods:

To examine the impact of physiologic levels of circulating cortisol on subcutaneous adipose tissue, 25 overweight and obese subjects were employed, and their serum levels of morning (AM) and evening (PM) cortisol, AM/PM cortisol ratios, and 24-h urinary-free cortisol (UFC) were compared with their clinical parameters, serum cytokine levels, and mRNA expression of 93 receptor action-regulating and 93 glucocorticoid-responsive genes in abdominal subcutaneous fat.

Results and Conclusions:

AM cortisol levels did not correlate with mRNA expression of the all genes examined, whereas PM cortisol levels, AM/PM cortisol ratios, and 24-h UFC were associated with distinct sets of these genes. Body mass index did not significantly correlate with the four cortisol parameters employed. These results suggest that physiologic levels of AM serum cortisol do not solely represent biological effects of circulating cortisol on the expression of glucocorticoid-related genes in subcutaneous adipose tissue, whereas PM levels, amplitude, and net amounts of the diurnally fluctuating serum cortisol have distinct effects. Through the genes identified in this study, glucocorticoids appear to influence intermediary metabolism, energy balance, inflammation, and local circadian rythmicity in subcutaneous fat. Our results may also explain in part the development of metabolic abnormality and obesity in subjects under stress or patients with melancholic/atypical depression who demonstrate elevated levels of PM serum cortisol.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. References

Glucocorticoids, steroid hormones secreted from the adrenal cortices, have strong and diverse regulatory actions on many aspects of human physiology (1, 2). In humans, cortisol is a major endogenous glucocorticoid secreted into circulation in a circadian fashion where it reaches zenith in the early morning and nadir in the evening (3, 4). This daily fluctuation of cortisol is created under the strong influence of the circadian CLOCK center located in the suprachiasmatic nucleus (SCN) of brain hypothalamus to the corticotropin-releasing hormone (CRH)-secreting neurons (4). These neurons distribute in the paraventricular nucleus of the same brain area and function as a central component of the hypothalamic-pituitary-adrenal (HPA) axis (1). The corticotropin-releasing hormone stimulates the secretion of the adrenocorticotropic hormone (ACTH) from the anterior pituitary gland, which in turn increases production and release of cortisol from the adrenal cortex (1, 4). The circadian fluctuation of circulating cortisol is an essential component of normal human physiology and pathophysiology, but its exact effects and underlying molecular mechanisms have not been fully characterized as yet (4).

Most of the diverse actions of glucocorticoids are mediated by an intracellular receptor molecule, the glucocorticoid receptor (GR), which belongs to the superfamily of steroid/sterol/thyroid/retinoid/orphan nuclear hormone receptors, and functions as a hormone-dependent transcription factor (5). In some organs and tissues including fat and brain, the mineralocorticoid receptor (MR) may also function as their receptor due to the absence of 11β-hydroxysteroid dehydrogenase type 2 (11BHSD2), an enzyme that converts active cortisol into inactive cortisone, and thus, protects MR from high levels of circulating glucocorticoids (6). Since these hormones are essential for every aspect of human physiology, the transcriptional activity of these receptors is under tight regulation by multiple cellular signaling pathways to meet with needs of local organs and tissues through several distinct regulatory mechanisms, such as differential isoform expression, post-translational modifications (e.g., phosphorylation and acetylation), and direct protein-protein interaction with other transcription factors (2).

Over half of the adult population in the United States and other Western countries are overweight or obese. Obesity decreases life expectancy by increasing the incidence of obesity-associated metabolic complications, such as insulin resistance/overt diabetes mellitus, hyperlipidemia, and hypertension, and subsequently by developing atherosclerosis and resultant ischemic heart diseases and cerebrovascular accidents (7). Recent evidence indicates that chronic, low-grade inflammation observed in adipose tissue of obese subjects, represented by local production of proinflammatory cytokines and chemokines, including the tumor necrosis factor α (TNFα), interleukin (IL)-6, IL-18, reactive oxygen species, and nitrogen intermediates, is crucial for developing metabolic syndrome in these subjects (8).

It is well known that chronic, excessive amounts of circulating glucocorticoids cause obesity and associated metabolic abnormalities (1). These conditions are typically observed in Cushing's syndrome, which is characterized by chronic elevation of serum cortisol and the absence of its circadian rhythmicity (9). Glucocorticoids induce these manifestations by strongly influencing the expression/activities of molecules involved in adipocyte metabolism (lipid turnover, uptake of fatty acids, and differentiation) and the inflammatory pathways associated with obesity (10). In addition to excess amounts of circulating glucocorticoids, increased sensitivity to glucocorticoids in adipose tissue leads to “Cushingoid” obesity and metabolic abnormalities, as evidenced by the mice over-expressing 11BHSD1, which converts inactive cortisone to active cortisol (11). Furthermore, subjects under chronic stress or patients with melancholic/atypical depression demonstrate fluttering of daily fluctuation of serum cortisol with elevation of its evening levels and frequently develop obesity, metabolic syndrome, and subsequent cardiovascular complications (12, 13). Nonetheless, the physiologic impact of circulating cortisol and its circadian rhythmicity on the expression of glucocorticoid-responsive genes has not been fully characterized in adipose tissue of obese subjects. We do not know either the expression pattern of GR action-regulating genes, which influence sensitivity of these tissues to glucocorticoids.

Therefore, we examined differential mRNA expression of known glucocorticoid-responsive and GR action-regulating genes in subcutaneous fat of overweight and obese subjects. Our results suggest that parameters for serum cortisol concentrations influence mRNA expression of distinct sets of glucocorticoid-related genes, which likely affects intermediary and energy metabolism, inflammation, and circadian rhythms in subcutaneous adipose tissue of obese subjects.

Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. References


The study cohort is represented by a subset of 285 adult volunteers with overweight or obesity with or without stable medical co-morbidities and stable body weight enrolled in a study approved by the institutional review board of the National Institute of Digestive Diabetic and Kidney Diseases (protocol number: NCT00428987). All subjects involved in this study gave written informed consent prior to admission to the Metabolic Clinical Research Unit in the Hatfield Clinical Research Center of the National Institute of Health for assessment of hormonal, metabolic, cognitive, and behavioral traits. Subjects enrolled in this analysis included 25 healthy adults (9 females and 16 males) with no medications, dietary, or vitamin supplements. Their ages ranged from 19 to 51 years (34.6 ±9.90: mean ± SD), with body mass indices (BMI) between 25 and 51 (30.4 ±5.46: mean ± SD). Based on the international classification of adult underweight, overweight and obesity by the World Health Organization (http://apps., 13 of our 25 subjects were classified as overweight (or pre-obese) (25-29.99 kg/m2), 9 as obese class I (30-34.99 kg/m2), 1 as obese class II (35-39.99 kg/m2), and 2 as obese class III (>40 kg/m2).

Acquisition of clinical parameters

A medical history, physical examination, and the following tests were performed on all subjects. Peripheral venous blood samples were obtained in the morning (08:00 h) after overnight fasting (at least 8 h) for glucose, insulin, creatinine, and cortisol and again in the evening (20:00 h) for serum cortisol after taking regular meal. Urine was collected for 24 h to measure urinary-free cortisol (UFC) and creatinine levels. All measurements for plasma glucose, insulin, cortisol and creatinine, and UFC and creatinine were performed by the Department of Laboratory Medicine, the Hatfield Clinical Research Center, National Institutes of Health. The BMI and homeostatic model assessment (HOMA) index for estimation of subjects' obesity and insulin resistance were calculated as previously reported (14). Insulin sensitivity index was calculated using the MinMod Millennium (6.02) program by using the results obtained in the frequently sampled intravenous glucose tolerance test (FSIVGTT), performed in the morning after 12-h overnight fasting by following the method previously described (15). Each subject performed a maximal graded exercise test on an upright or recumbent cycle ergometer (Model Lode 906900 or Lode 929900, Lode B.V., Groningen, The Netherlands) to determine VO2max, according to the standard protocol. Regional fat percentages were measured by using the dual energy X-ray absorptiometry (DXA) in the total body scanner (Lunar iDXA, GE Healthcare, Madison, WI) with the default thickness mode. Data analysis was performed by using the GE Encore 11.10 software (GE Healthcare Life Sciences, Madison, WI).

Measurement of serum cytokines

Serum levels of IL-1β, -2, -6, -8, -10, and -12 p70, interferon γ (IFNγ), granulocyte/macrophage colony-stimulating factor (GM-CSF), TNFα, and visfatin were quantified in sera of subjects obtained in the morning by using the Human ProInflammatory 9-Plex Ultra-Sensitive Kit (Meso Scale Discovery, Gaithersburg, MD; IL-1β, -2, -6, -8, -10, -12 p70, IFNγ, GM-CSF, and TNFα) and the Human Visfatin/PBEf1/NAMPT ELISA kit (ADIPO BIOSCIENCE, Santa Clara, CA).

Subcutaneous fat biopsy, purification of total RNA, and reverse transcription

Subcutaneous adipose tissues (3-5 g) were obtained from the abdominal region by aspirating with a 16-gauge needle under local anesthesia. They were separated into aliquots of 500 mg, frozen quickly in liquid nitrogen, and were stored at −80°C until further use. Total RNA was purified from frozen adipose tissues by using TRIzol® reagent (Invitrogen, Carlsbad, CA) and was further cleaned up with the RNeasy Mini kit (Qiagen, Valencia, CA), according to the protocols of manufacturers. After treatment with DNase I, purified total RNA was converted to first strand cDNA using the TaqMan reverse transcription kit (Applied Biosystems, Foster City, CA).

RT2 Profile Custom PCR Arrays for mRNA expression of GR action-regulating and glucocorticoid-responsive genes

Gene expression profiling of adipose tissues was performed by using the RT2 Profile Custom PCR Arrays for measuring quantitatively mRNA expression of 186 genes: 93 known GR action-regulating genes and 93 known glucocorticoid-responsive genes (SA Bioscience, Frederick, MD) in a 96-well format. In each plate design, we also inserted three housekeeping genes: ribosomal protein large P0 (RPLP0), glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and β-actin (ACTB). Lists of the genes measured are shown in Supplementary information Table S1A and S1B. We ran reaction for each subject sample by using these two plates and cDNA converted from 380 ng of total RNA purified from the sample.

Statistical analyses

The threshold cycle (Ct) values obtained in the real-time PCR using the RT2 Profile Custom PCR Arrays were multiplied by −1 to convert to “expression” values in the log2 format. The Ct values of measured genes in each sample were normalized by subtracting median Ct values of all genes obtained in the same plate. We performed a statistical test for examining the association between each pair of clinical covariates or gene expressions. We used the Spearman rank method for calculating the correlation r and the permutation method for computing the exact P value. The data values were permuted 10,000 times and the Spearman correlation was calculated. This gives an estimate of the null distribution. The two-sided P value was computed as the proportion of times where the absolute null correlation values were higher than the absolute r. P values were adjusted for multiple comparisons by using the Benjamini Hochberg false discovery rate (FDR) method (16).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. References

Four cortisol parameters we employed and their mutual relationship

To examine the impact of circulating cortisol on the mRNA expression of known GR action-regulating and glucocorticoid-responsive genes in subcutaneous fat, we employed four parameters associated with serum cortisol concentrations: (1) AM serum cortisol levels, (2) PM serum cortisol levels, (3) AM/PM serum cortisol ratios, and (4) 24-h UFC. AM/PM serum cortisol ratios represent the amplitude of diurnally fluctuating serum cortisol, while 24-h UFC is an indicator for net amounts of cortisol produced and available during 24 h (17). All subjects demonstrated normal AM and PM serum cortisol concentrations and 24-h UFC levels. Mutual association between these four cortisol parameters is shown in Table 1. AM serum cortisol levels demonstrated strong correlation with 24-h UFC (r = 0.637, P < 0.01) and weak correlation with AM/PM cortisol ratios (r = 0.513, P = 0.047). PM serum cortisol levels did not show correlation with 24-h UFC (r = −0.110, P = 0.710) and AM serum cortisol (r = 0.134, P = 0.742) levels but were negatively and strongly associated with the AM/PM serum cortisol ratios (r = −0.779, P < 0.01). These results indicate that AM serum cortisol levels are major contributors for determining the net amounts of cortisol produced during 24 h, whereas PM serum cortisol levels strongly influence the amplitude of diurnal serum cortisol fluctuations. PM serum cortisol levels appear to be independent and unique parameters for circulating cortisol, as they did not show correlation with AM serum cortisol levels and 24-h UFC.

Table 1. Correlations among the four cortisol parameters employed in this study
 AM cortisol (μg/dl)PM cortisol (μg/dl)AM/PM cortisol ratio24-h UFC (μg/24 h)
  • Numbers indicate r values (P values) obtained in the correlation analysis.

  • *


  • **

    P<0.01, compared to the two conditions indicated.

AM cortisol (μg/dl)1 (N/A)0.134 (0.742)0.443 (0.201)0.637 (<0.01)**
PM cortisol (μg/dl) 1 (N/A)−0.799 (<0.01)**−0.110 (0.710)
AM/PM cortisol ratio  1 (N/A)0.513 (0.047)*
24-h UFC (μg/24 h)   1 (N/A)

Association of the four cortisol parameters to clinical covariates for obesity, glucose metabolism, and blood pressure

We next compared the four cortisol parameters we employed with the clinical covariates that represent degree of obesity and obesity-associated complications, such as glucose intolerance/insulin resistance and hypertension. Among the clinical covariates shown in Table 2, we only found moderate association of 24-h UFC with the levels of measured VO2 (r = 0.562, P = 0.030), which indicate aerobic activity and potency of cardiovascular system (18). We did not find any correlation of AM and PM serum cortisol levels, AM/PM serum cortisol ratios, and 24-h UFC with BMI, regional fat percentages, plasma fasting glucose levels, fasting serum insulin levels, HOMA index, insulin sensitivity index, and systolic and diastolic blood pressures. Among these clinical covariates, BMI was correlated with regional fat percentage (r = 0.676, P = 0.002), fasting serum insulin levels (r = 0.530, P = 0.016), HOMA index (r = 0.541, P = 0.027), and measured VO2 (r = −0.543, P = 0.030), but not with systolic and diastolic blood pressures. Regional fat percentages correlated not only with BMI, but also with fasting serum insulin levels (r = 0.510, P = 0.033), HOMA index (r = 0.541, P = 0.032), and measured VO2 (r = −0.676, P = 0.007). Some indicators for glucose intolerance/insulin resistance showed association with each other, whereas systolic and diastolic blood pressures demonstrated strong correlation (Supplementary information Table S2A and S2B). Taken together, these results suggest that the degree of overweight or obesity is independent of the parameters of circulating cortisol in our subjects. Obesity, particularly a visceral type, appears to be a risk factor for glucose intolerance/insulin resistance and reduction of aerobic activity and potency of cardiovascular system in our subjects, as reported previously (19).

Table 2. Correlation between four cortisol parameters employed in this study and clinical covariates/serum cytokine levels
 Correlation r valueAdjusted P value
 AM cortisolPM cortisolAM/PM cortisol ratio24-h UFCAM cortisolPM cortisolAM/PM cortisol ratio24-h UFC
  • *


  • BMI: body mass index, IL: interleukin, INFγ: interferon γ, GM-CSF: granulocyte/macrophage colony-stimulating factor, HOMA index: homeostatic model assessment index, TNFα: tumor necrosis factor α, UFC: urinary-free cortisol

Regional fat percentage−0.0820.008−0.103−0.3080.8850.9980.8330.349
Fasting plasma glucose−0.423−0.04320.143−0.2960.1280.1730.6810.349
Fasting serum insulin−0.185−0.1960.005−0.3070.7150.8300.9770.349
HOMA index−0.209−0.2350.030−0.3430.6690.8300.9770.281
Insulin sensitivity index0.476−0.0010.3520.4570.1280.9950.3660.094
Measured VO20.332−0.0140.1590.5620.3230.9950.6810.024*
Systolic blood pressure0.354−0.1730.3650.2550.3210.8300.3540.499
Diastolic blood pressure−0.002−0.2160.148−0.1590.9920.8300.6810.710
IL-12 p70−0.0590.092−0.176−0.2290.8850.8930.6810.574

Association of four serum cortisol parameters with mRNA expression of some GR action- regulating and glucocorticoid-responsive genes

To determine gene expression profiles of known glucocorticoid-related genes in adipose tissue of overweight and obese subjects, we quantified mRNA expression of 186 genes—93 GR action-regulating genes and 93 glucocorticoid-responsive genes—in subcutaneous fat biopsied from our 25 subjects. The genes belonging to these two groups are listed in Supplementary information Table S2A and S2B, respectively. We compared their mRNA expression with the four cortisol parameters we employed in this study.

In comparison with the levels of AM serum cortisol, none of the GR action-regulating and glucocorticoid-responsive genes showed correlation with this cortisol parameter (Table 3A). On the other hand, PM serum cortisol levels correlated with mRNA expression of three GR action-regulating genes [14-3-3η, G protein γ2 (GNG2) and estrogen-related receptor β (ERRB)] and four glucocorticoid-responsive genes [glucose-6-phosphatase catalytic subunit (G6PC), dihydrolipoamide branched chain transacylase E2 (DBT), serum/glucocorticoid-regulated kinase 1 (SGK1), and von Hippel-Lindou tumor suppressor (VHL)] (Table 3B). Smad3, cytochrome p450, family 3, subfamily A, polypeptide 5 (CYP3A5), and TNFα were just outside of statistical significance in this comparison. AM/PM serum cortisol ratios correlated with mRNA expression of six GR action-regulating genes [nuclear factor of κB (NFκB) p50 component (NFKB2), signal transducer and stimulator of transcription 5A (STAT5A), calreticulin (CALR), 14-3-3η, 11β-hydroxysteroid dehydrogenase 1 (11BHSD1), and circadian locomotor output cycles kaput (CLOCK)] and seven glucocorticoid-responsive genes [SGK1, dual-specificity protein phosphatase 1 (DUSP1), G6PC, uncoupling protein 2 (UCP2), insulin-induced gene 2 (INSIG2), TNFα, and annexin A1 (ANXA1)] (Table 3C). Smad6 and fat mass and obesity associated (FTO) were just outside of statistical significance. Twenty-four-hour UFC correlated with three GR action-regulating genes [p38 mitogen-activated protein kinase (MAPK14), protein arginine methyltransferase 1 (PRMT1), and 11BHSD1] and three glucocorticoid-responsive genes [plasminogen activator inhibitor 1 (SERPINE1), interleukin-6 (IL-6), and ANXA1 (Table 3D)]. STAT5A and phosphoenolpyruvate carboxykinase 1 (PEPCK) were just outside of statistical significance. We examined correlation between clinical covariates and mRNA expression of the genes that had multiple hits—11βHSD1, 14-3-3η, ANXA1, G6PC, and SGK1—but none of these genes showed significant association (data not shown).

Table 3. Genes significantly correlated with serum cortisol and 24-h UFC levels
Gene nameSymbolGenbank accession #r valueP value
A. AM serum cortisol levels
GR action-regulating genes    
Glucocorticoid-responsive genes    
B. PM serum cortisol levels    
GR action-regulating genes    
G-protein γ2GNG2NM_053064−0.5740.032
Estrogen-related receptor βERRBNM_004452−0.5930.032
Glucocorticoid-responsive genes    
Glucose-6 phosphatase, catalytic subunitG6PCNM_0001510.6530.009
Dihydrolipoamide branched chain transacylase E2DBPNM_0019180.5890.028
Serum/glucocorticoid-regulated kinase 1SGK1NM_0056270.5970.032
von Hippel-Lindau tumor suppressorVHLNM_0005510.5640.032
Cytochrome P450, family 3, subfamily A, polypeptide 5CYP3A5NM_0007770.5440.054
Tumor necrosis factor αTNFαNM_0005940.5410.054
C. AM/PM serum cortisol ratios    
GR action-regulating genes    
Nuclear factor of κB p50 componentNFKB2NM_0530640.5690.018
Signal transducer and stimulator of transcription 5ASTAT5ANM_0031520.6640.026
11β-Hydroxysteroid dehydrogenase 111BHSD1NM_181755−0.7000.039
Circadian locomotor output cycles kaputCLOCKNM_005902−0.4730.045
Smad 6SMAD6NM_005585−0.4590.052
Glucocorticoid-responsive genes    
Serum/glucocorticoid-regulated kinase 1SGK1NM_0056270.736<0.001
Dual specificity protein phosphatase 1DUSP1NM_0044170.5130.008
Glucose-6 phosphatase, catalytic subunitG6PCNM_0001510.5190.010
Uncoupling protein 2UCP2NM_0033550.5210.012
Insulin-induced gene 2INSIG2NM_0007770.4200.025
Tumor necrosis factor αTNFαNM_0005940.4080.034
Annexin A1ANXA1NM_0007000.3440.037
Fat mass and obesity associatedFTONM_0010804320.4740.051
D. 24-h UFC levels    
GR action-regulating genes    
p38 Mitogen-activated protein kinaseMAPK14NM_0013150.6060.021
Protein arginine methyltransferase 1PRMT1NM_0015360.4690.025
11β-Hydroxysteroid dehydrogenase type 111BHSD1NM_181755−0.7450.026
Signal transducer and activator of transcription 5ASTAT5NM_0031520.5730.057
Glucocorticoid-responsive genes    
Plasminogen activator inhibitor 1SERPINE1NM_000602−0.5390.018
Annexin A1ANXA1NM_000700−0.4380.034
Phosphoenolpyruvate carboxykinase 1PEPCKNM_0025910.4940.053

In above comparisons between the four cortisol parameters and mRNA expression of glucocorticoid-related genes, we found several genes functional in the immune system, such as ANXA1, DUSP1, IL-6, NFKB2, SERPINE1, SGK1, STAT5A, and TNFα, while obesity is associated with local adipocyte inflammation (20). Thus, we measured in our subjects serum concentrations of the cytokines and bioactive molecules, IL-1β, IL-2, IL-6, IL-8, IL-10, IL-12 p70, INFγ, GM-CSF, TNFα, and visfatin, and compared their levels with the four cortisol parameters (Table 2): We did not find any significant correlation in the comparisons. We further compared concentrations of these cytokines to the clinical covariates, but again no correlation was observed. These results suggest that serum levels of these cytokines do not represent their production/concentrations and/or expected inflammation in local adipose tissues, possibly jammed by those derived from other organs and tissues.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. References

To examine the impact of physiologic levels of circulating cortisol on the functions of adipose tissue in overweight and obese subjects, we measured mRNA expression of GR action-regulating and glucocorticoid-responsive genes in their abdominal subcutaneous fat and compared their expression with the four parameters associated with circulating cortisol. Since GR and MR share numerous upstream regulatory molecules, many of the GR action-regulating genes that we measured in this study may also influence glucocorticoid activity in adipocytes through modulation of MR actions (5).

We found that AM serum cortisol levels did not correlate with mRNA expression of any of the genes examined, while PM serum cortisol levels, AM/PM serum cortisol ratios, and 24-h UFC were associated with expression of distinct gene sets. These results suggest that physiologic levels of AM serum cortisol alone do not represent biological effects of circulating cortisol on glucocorticoid-related gene expression in subcutaneous fat. Our results also suggest that PM serum cortisol levels, AM/PM cortisol ratios (amplitude of diurnally fluctuating serum cortisol), and 24-h UFC (production of cortisol during 24 h) have distinct biological effects in subcutaneous fat. PM serum cortisol levels and AM/PM serum cortisol ratios shared one GR action-regulated gene (14-3-3η) and three glucocorticoid-responsive genes (G6PC, SGK1, and TNFα), whereas both AM/PM serum cortisol ratios and 24-h UFC had two GR action-regulating gene (11BHSD1 and STAT5A) and one glucocorticoid-regulated gene (ANXA1) (Table 4), consistent with strong and weak association between the former two and the latter two parameters, respectively.

Table 4. Genes whose mRNA expression was correlated with four cortisol parameters employed in this study
Gene nameSymbolPM CortisolAM/PM Cortisol Ratio24-h UFC
GR action-regulating genes    
11β-Hydroxysteroid dehydrogenase 111BHSD1 
Circadian locomotor output cycles kaputCLOCK  
Estrogen-related receptor βERRB  
G-protein γ2GNG2  
Nuclear factor of κB p50 componentNFKB2  
p38 Mitogen-activated protein kinaseMAPK14  
Protein arginine methyltransferase 1PRMT1  
Signal transducer and activator of transcription 5ASTAT5A 
Samd 3SMAD3  
Smad 6SMAD6  
Glucocorticoid-responsive genes    
Annexin A1ANXA1 
Cytochrome p450, family 3, subfamily A, polypeptide 5CYP3A5  
Dihydrolipoamide-branched chain transacylase E2DBP  
Dual-specificity protein phosphatase 1DUSP1  
Fat mass and obesity associatedFTO  
Glucose-6 phosphatase, catalytic subunitG6PC 
Insulin-induced gene 2INSIG2  
Plasminogen activator inhibitor 1SERPINE1  
Phosphoenolpyruvate carboxykinase 1PEPCK  
Serum/glucocorticoid-regulated kinase 1SGK1 
Tumor necrosis factor αTNFα 
Uncoupling protein 2UCP2  
von Hippel–Lindau tumor suppressorVHL  

We found that many glucocorticoid-responsive genes correlated with levels of PM serum cortisol, whereas AM serum cortisol had no such genes. These results suggest that PM cortisol levels are more important than the AM levels for regulation of glucocorticoid-responsive genes at least in their physiologic levels, in part underlying the previous clinical observation that elevation of PM serum cortisol levels was associated with development of metabolic syndrome and cardiovascular complications (12, 13). We recently found that peripheral CLOCK acetylates GR and suppresses GR transcriptional activity (21, 22). Furthermore, CLOCK-mediated acetylation of GR is much lower in the evening than in the morning in human peripheral blood mononuclear cells, and thus, tissue sensitivity to glucocorticoids is elevated in this day period (23, 24). It appears that this regulatory system may underlie in part the current finding that PM serum cortisol levels have stronger biologic effects than AM serum cortisol levels for the regulation of glucocorticoid-related gene expression by changing tissue glucocorticoid sensitivity. Among the genes associated with PM serum cortisol levels, G6PC encodes a catalytic component of the glucose-6-phosphatase, a rate-limiting enzyme for glycolysis (25). Although PM serum cortisol levels did not correlate with any clinical covariates for glucose intolerance/insulin resistance, this gene may be important for developing the metabolic complications associated with elevated PM serum cortisol. In addition to G6PC, PM serum cortisol levels correlated with mRNA expression of DBT, the gene encoding a component of the protein complex, which catalyzes branched-chain amino acids, SGK1, a well-known glucocorticoid-responsive gene that mediates cellular stress response and development of hypertension and diabetic nephropathy, and VHL, an E3 ubiquitine protein ligase whose C. elegance homolog VHL-1 was recently shown to reduce life span through the hypoxia-inducible factor-1 in this organism (26-28). PM serum cortisol levels may influence these biological pathways in subcutaneous fat by changing expression of these genes.

AM/PM serum cortisol ratios correlated with mRNA expression of many glucocorticoid-responsive genes, such as SGK1, DUSP1, G6PC, UCP2, INSIG2, TNFα, and ANXA1. We explained biological effects of SGK1 and G6PC above. DUSP1, also called as the mitogen-activated protein kinase (MAPK) phosphatase 1 (MPK-1), acts as a non-receptor-type protein-tyrosine phosphatase and plays important roles in the regulation of inflammation, cell proliferation, and stress response by antagonizing to phosphorylation and resulting regulation by MAPKs (29). TNFα is a proinflammatory cytokine important for the development of insulin resistance in adipose tissue (20). UCP2 plays a central role in heat generation and subsequent energy consumption by uncoupling the oxidative phosphorylation and ATP synthesis in mitochondria (30). INSIG2 involves in carbohydrate metabolism by blocking proteomic processing of the sterol regulatory element-binding proteins (SREBPs) (31). ANXA1 is a Ca2+-dependent phospholipid-binding protein with anti-inflammatory activity (32). Changes in AM/PM serum cortisol ratios may influence above-indicated biologic pathways, such as inflammation, glucose and fatty acid metabolism, and energy balance, by altering expression of these genes.

We found that PM serum cortisol levels, AM/PM serum cortisol ratios, and 24-h UFC correlated with distinct sets of GR action-regulating genes. Among them, NFKB2 is a major component of the heterodimeric NFκB transcription factor, which plays critical roles in the regulation of proinflammatory response and represses GR-induced transcriptional activity with several different mechanisms (33). STAT5A is also a proinflammatory transcription factor, transducing the effects of many cytokines, including IL-2, IL-3, IL-7 GM-CSF, erythropoietin and thrombopoietin, and growth hormone and prolactin (34). STAT5A represses the transcriptional activity of GR through direct protein-protein interaction (34). 14-3-3η regulates several signaling pathways including those of insulin and growth factors, apoptosis, and cell cycle by altering stability and subcellular localization of their key molecules through binding to their phophorylated serine or threonine residues (35). It modulates GR transcriptional activity by physically interacting with this receptor through its C-terminal portion (36). GNG2, together with its molecular partner G-protein β, participates in many intracellular signaling pathways including those associated with ion channels, phospholipase C, MAP kinases, and adenyl cyclases (37). The G-protein βγ complex physically interacts with GR and represses its transcriptional activity in the nucleus (38). 11BHSD1 whose mRNA expression was found to correlate negatively with AM/PM serum cortisol ratios and 24-h UFC produces active cortisol from inactive cortisone in local tissues (6). MAPK14 is a serine/threonine kinase essential for many signaling pathways including those of immune and stress response, cell growth, cell cycle, and transcriptional regulation (39). It phosphoarylates GR and modulates the transcriptional activity of this receptor (40). Association between expression of these GR action-regulating genes and serum and urinary cortisol parameters suggests the presence of another regulatory loop between their encoding molecules and the GR through glucocorticoids. Finally, we found that the expression of CLOCK negatively correlated with AM/PM serum cortisol ratios. This result suggests that amplitude of diurnally fluctuating serum cortisol potentially influences the activity of the local circadian CLOCK system residing in subcutaneous fat by changing the expression of this key transcription factor. This potential regulatory mechanism might help synchronizing the circadian rhythms of the peripheral CLOCK system to those of the central CLOCK via circulating cortisol (4).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. References