Endocrinology Unit, Centre for Cardiovascular Science, Queen's Medical Research Institute, University of Edinburgh, 47 Little France Crescent, Edinburgh EH16 4TJ, UK. E-mail: Nik.Morton@ed.ac.uk
Objectives: In ideopathic obesity, there is evidence that enhanced cortisol regeneration within abdominal subcutaneous adipose tissue may contribute to adiposity and metabolic disease. Whether the cortisol regenerating enzyme, 11β-hydroxysteroid dehydrogenase type 1 (11βHSD1), or glucocorticoid receptor (GRα) levels are altered in other adipose depots remains uncertain. Our objective was to determine the association between 11βHSD1 and GRα mRNA levels in four distinct adipose depots and measures of obesity and the metabolic syndrome.
Research Methods and Procedures: Adipose tissue biopsies were collected from subcutaneous (abdominal, thigh, gluteal) and intra-abdominal (omental) adipose depots from 21 women. 11βHSD1 and GRα mRNA levels were measured by real-time polymerase chain reaction. Body composition, fat distribution, fat cell size, and blood lipid, glucose, and insulin levels were measured.
Results: 11βHSD1 mRNA was highest in abdominal subcutaneous (p < 0.001) and omental (p < 0.001) depots and was positively correlated with BMI and visceral adiposity in all depots. Omental 11βHSD1 correlated with percent body fat (R = 0.462, p < 0.05), fat cell size (R = 0.72, p < 0.001), and plasma triglycerides (R = 0.46, p < 0.05). Conversely, GRα mRNA was highest in omental fat (p < 0.001). GRα mRNA was negatively correlated with BMI in the abdominal subcutaneous (R = −0.589, p < 0.05) and omental depots (R = −0.627, p < 0.05). Omental GRα mRNA was inversely associated with visceral adiposity (R = −0.507, p < 0.05), fat cell size (R = −0.52, p < 0.01), and triglycerides (R = −0.50, p < 0.05).
Discussion: Obesity was associated with elevated 11βHSD1 mRNA in all adipose compartments. GRα mRNA is reduced in the omental depot with obesity. The novel correlation of 11βHSD1 with omental fat cell size, independent of obesity, suggests that intracellular cortisol regeneration is a strong predictor of hypertrophy in the omentum.
Visceral obesity is associated with an increased risk for type 2 diabetes, hyperlipidemia, and hypertension (the metabolic syndrome) (1)(2). In contrast, comparable amounts of fat stored preferentially in gluteal or femoral depots (lower body obesity) showed lower risk of morbidity and mortality from metabolic abnormalities (3). Exposure to high circulating glucocorticoid (GC)1 levels, as found in Cushing's syndrome, causes a metabolic disease that resembles features of idiopathic metabolic syndrome including pronounced visceral obesity (4). However, idiopathic obesity is not associated with high circulating GC levels (4). Rather, it seems that intracellular generation of active from inactive GCs by the enzyme 11β-hydroxysteroid dehydrogenase type 1 (11βHSD1) is aberrantly elevated in adipose tissue of obese individuals (4). Thus, we and others have shown that 11βHSD1 mRNA and activity are elevated in abdominal subcutaneous adipose tissue from obese compared with non-obese individuals (5)(6)(7)(8)(9)(10) and in adipose depots in monogenic obesity in rodents (11)(12). In vivo microdialysis confirmed increased abdominal subcutaneous adipose regeneration of cortisol from cortisone in human obesity (13). Given the key association of intra-abdominal (visceral) adipose tissue in metabolic and cardiovascular risk, it has been hypothesized that increased 11βHSD1 in the visceral, rather than subcutaneous, adipose depot causes the adverse metabolic consequences of idiopathic obesity and the metabolic syndrome (14). This hypothesis has been supported by the observation that transgenic mice overexpressing 11βHSD1 in adipose tissue develop visceral obesity and the phenotype of the metabolic syndrome (12)(15). Conversely, global 11βHSD1 knockout mice are protected from the metabolic consequences of dietary obesity, at least in part, through adipose tissue insulin sensitization and fat redistribution away from the visceral depot (16)(17). However, Tomlinson et al. (18) reported a negative association between omental 11βHSD1 activity and BMI in primary omental human adipocytes cultured in vitro, although there was no correlation in whole adipose biopsies. A further study could not detect increased portal vein cortisol (19) or increased splanchnic cortisol production rates (20) in obesity, suggesting that if cortisol generation is indeed elevated in visceral fat, it is further metabolized or is not released at higher rates from adipocytes. Together, these data have engendered some confusion as to whether visceral and subcutaneous 11βHSD1 is elevated in human obesity. Mechanistically, GCs induce adipose tissue expansion by stimulating preadipocyte differentiation (21) and lipoprotein lipase-mediated triglyceride accumulation (22)(23). A key factor in this process is tissue levels of the intracellular glucocorticoid receptor α (GRα). In contrast to the plethora of studies (5)(6)(7)(8)(9)(10)(14) of 11βHSD1 in adipose tissue and metabolic syndrome, curiously few studies have examined associations between GR and such parameters. Genetic studies strongly support the role of the GC receptor in determining body composition and fat distribution (24)(25). Higher levels of GR found in omental compared with subcutaneous fat (26) would further be expected to have a more pronounced effect on GC action in the omental depot. However, our previous study showed no correlation between adipose tissue GR mRNA in abdominal subcutaneous adipose tissue and obesity (10), whereas Kannisto et al. (9) reported an inverse correlation between subcutaneous adipose GRα mRNA and BMI. Here we hypothesized that high adipose 11βHSD1, but not GR, mRNA levels predicts obesity in visceral and subcutaneous fat depots in women.
Research Methods and Procedures
Informed, written consent was obtained from 21 women undergoing elective, laparoscopic tubal ligation surgery; the study was approved by the Mayo Clinic Institutional Review Board. Tubal ligation surgery was done routinely in the follicular phase of the menstrual cycle to eliminate the risk of pregnancy. Before surgery, body composition was assessed by measuring weight, height, body fatness (% fat) using DXA (DPX-IQ; Lunar Radiation, Madison, WI), and abdominal fat distribution using a single sliced computerized tomography scan at the L2–L3 level (27)(28). Visceral and subcutaneous adipose tissue areas were calculated as previously described (28). Although 8 of 21 subjects were taking oral contraceptives, there was no effect on our variables by Student's t test, and the data were pooled.
Fasting plasma triglycerides, glucose, and insulin levels were assayed as previously described (29). Homeostasis model of assessment insulin resistance index (HOMA-IR) was calculated by the following equation: fasting insulin (μU/mL) × fasting glucose (mM)/22.5 (30).
Adipose Tissue Biopsies
Subcutaneous fat from abdominal (n = 13), thigh (n = 16), and gluteal (n = 18) regions were collected just before surgery and omental (n = 21) adipose biopsies were obtained intraoperatively. The tissue was washed to remove blood and snap frozen in liquid nitrogen and stored at −80 °C. Biopsies were homogenized in 1 to 2 mL Trizol (Gibco BRL, Paisley, United Kingdom). RNA was extracted using RNAid+ binding matrix (Anachem, Luton, UK) and eluted in diethylpyrocarbonate-treated water, containing 10 mM dithiothreitoland 400 U/mL RNasin (Promega, Southampton, UK). Total RNA was quantified using a spectrophotometer at A260.
RNA integrity was verified by agarose gel electrophoresis. Oligo dT-primed cDNA was synthesized from 0.5 μg of RNA samples using the Promega Reverse Transcription System (Promega, Madison, WI). GRα and 11βHSD1 mRNA levels were quantified by real-time polymerase chain reaction primer-probe sets using the ABI PRISM 7700/7900 Sequence Detection System (Applied Biosystems, Foster City, CA). The primers and probes used are as follows: 11βHSD1, 5′-GGAATATTCAGTGTCCAGGGTCAA-3′ (forward), 5′-TGATCTCCAGGGCACATTCCT-3′ (reverse) and 5′-6-FAM-ACATTGACAACCTTCGCTGGGAGG-TAMRA-3′ (probe); GRα, 5′-CATTGTCAAGAGGGAAGGAAACTC-3′ (forward), 5′-ATTTTCAACCACTTCATGCATAGAA-3′ (reverse), and 5′-6-FAM-TTGTCAGTTGATAAAACCGCTGCCAGTTCT-TAMRA-3′ (probe). Levels of GRα or 11βHSD1 mRNA are reported relative to human cyclophilin A RNA (Applied Biosystems, Cheshire, United Kingdom) as previously optimized (10) and are expressed in arbitrary units. Standard curves for primer-probe sets and real-time polymerase chain reaction analysis were performed as previously described (10). Fat cell size (mean diameter of mature adipocytes in micrometers) was determined using the AdCount (Biomedical Imaging Resource, Rochester, MN) approach as previously described (31).
Repeated-measures ANOVA followed by Tukey post hoc test were performed to determine differences in transcript levels in paired adipose compartments. Pearson correlation was performed to examine the relationships between anthropometric and metabolic parameters with 11βHSD1 or GRα mRNA levels in different fat depots according to a priori hypotheses to minimize multiple interdependent variables. Multiple linear regression was performed to test whether relationships between fat cell sizes or triglycerides with transcript levels are independent of obesity. Data were tested for normality and were normalized by log transformation where appropriate.
Data are means ± standard error unless otherwise stated. Differences were considered significant at p < 0.05. All statistical analyses were performed using Sigma Stat 3.1 software (Systat Software Inc., San Jose, CA).
Volunteers were white women 35 ± 1 years of age with a mean BMI of 32.7 ± 1.5 kg/m2 (Table 1), indicating an obese group. Their median ± interquartile range fasting insulin was 15.6 ± 19.6 μU/mL, and HOMA-IR as 3.3 ± 4.6 (range, 0.42 to 47). Median fasting glucose and mean triglyceride concentrations averaged 4.9 ± 0.44 mM and 148 ± 19 mg/dL, respectively (Table 1).
Table 1. Anthropometric and metabolic characteristics of study participants
Women (n = 21)
VAT, visceral adipose tissue; SAT, subcutaneous adipose tissue. Data are presented as mean ± standard error. For glucose, data are presented as median ± interquartile range. Range values in parentheses.
35.3 ± 1.4 (20 to 44)
32.7 ± 1.5 (20 to 42)
Body fat (%)
46 ± 2 (23 to 56)
VAT area (cm2)
96 ± 16.3 (23 to 236)
SAT area (cm2)
326 ± 35 (59 to 577)
0.28 ± 0.03 (0.11 to 0.64)
Fasting glucose (mM)
4.9 ± 0.44 (4.1 to 7.0)
Fasting insulin (μU/mL)
16.5 ± 3.5 (2.2 to 64.3)
Fasting triglyceride (mg/dL)
146.7 ± 18.7 (38 to 278)
Comparison of 11βHSD1 and GRα mRNA Levels in Multiple Fat Depots
11βHSD1 mRNA levels were greater in subcutaneous abdominal and omental adipose tissues than in other subcutaneous tissues (gluteal, thigh; Figure 1A). 11βHSD1 mRNA levels were positively correlated in the three subcutaneous depots: abdominal subcutaneous vs. thigh (R = 0.83, p < 0.005) and abdominal vs. gluteal (R = 0.86, p < 0.005). There were no associations between abdominal subcutaneous and omental 11βHSD1 mRNA levels. GRα mRNA expression levels were highest in omental and lowest in thigh subcutaneous adipose tissue (Figure 1B) but were not correlated between any compartments. No correlation was found between 11βHSD1 and GRα mRNA levels in any depot (data not shown).
Association of Obesity with Elevated 11βHSD1 but Reduced GRα mRNA in Omental Fat
A strong positive association between BMI and omental adipose 11βHSD1 mRNA levels was observed (R = 0.570; Figure 2A; Table 2; p < 0.01). Moreover, visceral fat tissue [visceral adipose tissue (VAT)] area by computed tomography was correlated with increased 11βHSD1 mRNA levels in the omentum (Figure 2B; Table 2). Omental 11βHSD1 mRNA levels were consistently and positively associated with general adiposity (as defined by percent of body fat; R = 0.462, p < 0.05) and with subcutaneous/peripheral adiposity (Table 2). In contrast, obesity was associated with decreased GRα mRNA levels in the omental depot (R = −0.627, p < 0.001; Figure 2D; Table 2). Visceral adiposity (Figure 2E) was negatively correlated with omental GRα mRNA levels (VAT; R = −0.507, p < 0.05).
Table 2. Pearson correlation (R) of 11βHSD1 and GRα mRNA levels in multiple adipose tissue depots with body composition, metabolic parameters, and fat cell sizes
11β HSD1, 11β-hydroxysteroid dehydrogenase type 1; GRα, glucocorticoid receptor α; VAT, visceral adipose tissue; SAT, subcutaneous adipose tissue; FSC, fat cell size. Percent body fat was measured using DXA. VAT and SAT were measured by computed tomography scan. FCS was determined by measuring the mean adipocyte diameter. Plasma glucose values, abdominal FCS, and 11βHSD1 mRNA levels in thigh and gluteal regions and GRα mRNA levels in thigh showed inhomogeneity of variance and were log-transformed before analysis.
Association of Omental Fat Cell Hypertrophy with Increased 11βHSD1 but Reduced GRα
Fat cells were smaller in the omental depot than in subcutaneous adipose tissue (SAT) (omental, 84 ± 4 μm; vs. abdominal subcutaneous, thigh, gluteal, 104 ± 3, 109 ± 3, and 106 ± 2 μm, respectively; p < 0.0001). Omental fat cell size was positively correlated (R = 0.72, p < 0.001) with 11βHSD1 mRNA levels (Figure 2C; Table 2). In contrast, GRα transcript levels were negatively correlated with omental fat cell size (Figure 2F; Table 2). To test for independent effects, multiple regression analyses were performed to understand the effects of obesity (BMI) and fat cell size on 11βHSD1 or GRα mRNA. Omental fat cell size was strongly and independently correlated with 11βHSD1 (standardized β coefficient, 1.2; p < 0.05), whereas GRα was not associated with fat cell size independently of obesity.
Association of Obesity with Increased 11βHSD1 mRNA Levels in Subcutaneous Depots but Reduced GRα mRNA in the Abdominal Subcutaneous Fat
Because reporting on glucocorticoid action has been largely in the abdominal subcutaneous depot, we extended our subcutaneous depots to gluteal and thigh to test whether the expected relationships held in these distinct compartments. Positive associations with BMI and 11βHSD1 mRNA levels were detected in all subcutaneous adipose compartments examined (e.g., abdominal subcutaneous adipose: R = 0.584, p < 0.05; Table 2). Thigh adipose 11βHSD1 mRNA correlated positively with SAT, and this relationship showed a consistent trend in the other depots (Table 2). In contrast, abdominal subcutaneous GRα mRNA levels were inversely correlated with BMI (R = −0.589, p < 0.05; Table 2). No significant associations were observed between SAT and abdominal subcutaneous GRα mRNA levels (Figure 3B) or GRα mRNA levels in any other subcutaneous depots (Table 2).
Abdominal Subcutaneous Hypertrophy is Associated with Reduced GRα mRNA Levels
Although multiple regression analysis with small numbers of abdominal subcutaneous biopsies should be interpreted cautiously, our data nevertheless indicate the following. Surprisingly, in subcutaneous abdominal adipose tissue, 11βHSD1 mRNA levels were not associated with fat cell size (Figure 3C). In contrast, abdominal subcutaneous GRα transcript levels were negatively correlated with fat cell size (Figure 3D). Multiple regression analysis confirmed that GRα mRNA levels were independently and negatively correlated with abdominal subcutaneous fat cell size (standardized β coefficient, −0.014; p < 0.05). Although GRα levels were negatively correlated with fat cell size in the thigh, this relationship did not hold when tested by multiple regression analysis.
Associations of Depot-specific Glucocorticoid Action with Metabolic Parameters
Pearson correlation was used to examine interrelationships between transcript levels and metabolic parameters (Table 2). Higher omental and thigh 11βHSD1 levels were associated with increased triglycerides levels (omental: R = 0.46, p < 0.05; thigh: R = 0.57, p < 0.01). This effect was not independent of obesity as confirmed by multiple linear regression analysis. There were no significant associations between fasting glucose, insulin, or HOMA-IR and 11βHSD1 transcript levels. In contrast, there was a negative correlation between GRα mRNA levels and plasma triglycerides levels selectively in the omental compartment (R = −0.50, p < 0.05). Moreover, there was a trend for inverse associations between omental GRα and fasting plasma insulin levels (R = −0.48, p = 0.052).
This study examined relationships between two major determinants of tissue glucocorticoid action in multiple adipose tissue depots in lean and obese women. The key findings were 1) 11βHSD1 mRNA is most highly expressed in subcutaneous abdominal and omental tissues, whereas GRα mRNA is highest in omental adipose tissue; 2) within individuals, there are positive associations between 11βHSD1 mRNA levels in subcutaneous adipose depots but not omental fat, whereas GRα mRNA levels do not correlate between depots; 3) between individuals, 11βHSD1 mRNA levels in all adipose depots, including omentum, are positively associated with BMI and local fat mass, whereas for GRα mRNA the main relationships were negative associations in the omentum; 4) increased fat cell size strongly associates with increased 11βHSD1 but reduced GRα in the omentum.
Here we extend our (4)(6)(8)(10) and others (5)(7)(9) previous observations on abdominal subcutaneous adipose to show a positive relationship between BMI and 11βHSD1 mRNA levels in all four adipose compartments examined. This implies an up-regulation of glucocorticoid production in multiple adipose tissue compartments, including the visceral fat depot. Increased visceral fat is a better predictor of metabolic abnormalities compared with upper-body subcutaneous fat (1)(2) and is independently correlated with increased morbidity and mortality (32). In contrast, lower-body obesity may even show a protective effect with respect to metabolic abnormalities (33). In our cohort, omental 11βHSD1 mRNA levels were also most strongly associated with increased general adiposity (% body fat). Kannisto et al. (9) previously reported a positive association between percent body fat and 11βHSD1 mRNA levels in subcutaneous in obese male and female monozygotic twins. In our study, subcutaneous adipose expression of 11βHSD1 mRNA did not correlate with percent body fat. Differences between these studies may be caused by the distinct populations studied and/or our limited power to detect this interaction because 11βHSD1 did correlate with other markers of generalized obesity such as BMI. Although in this study we did not measure 11βHSD1 activity levels because of limited sizes of biopsies, we expect, from previous studies (7)(10), that activity and mRNA levels are correlated in abdominal subcutaneous adipose tissue biopsies.
GRα transcript levels were highest in omental (visceral) fat, as previously reported in humans (26) and mice (12). The major new finding is that in omental and abdominal subcutaneous adipose, GRα is inversely associated with adiposity and fat cell size. A previous study also reported lower GRα mRNA levels in subcutaneous adipose tissue in severely obese women (34). Moreover, omental GRα mRNA levels were negatively associated with plasma triglyceride and, albeit as a trend, insulin levels. This finding is perhaps counterintuitive because GR polymorphisms causing increased glucocorticoid sensitivity lead to increased visceral fat accumulation (25). Our data suggest that reduction of GR levels to the degree we observe in obesity does not sufficiently compensate for the hypertrophy caused by increased ligand regeneration (i.e., 11βHSD1). This notion is supported by data from transgenic animal models. Thus, fat-specific 11βHSD1 overexpressing mice (12) have increased visceral adiposity, and 11βHSD1 null mice have reduced visceral adiposity (16), despite unaltered GR levels. In contrast in humans, to our knowledge, this is the first study describing a consistent opposing change in 11βHSD1 and GRα mRNA in multiple adipose compartments with obesity. However, because the two transcripts were not correlated in any depot, consistent with previous observations in subcutaneous adipose tissue (9), it seems unlikely that GRα and 11βHSD1 directly regulate each other. Naturally, there are clearly limitations with association studies, which do not test causality, and further direct study is needed to elucidate possible mechanisms involved.
The volunteers who participated in our study had normal glucose and triglyceride levels but impaired insulin sensitivity. Increased plasma triglycerides were associated with increased 11βHSD1 in both thigh and omental regions but with reduced GRα mRNA levels selectively in the omental compartment. Neither 11βHSD1 nor GRα mRNA levels were associated with fasting glucose, which is consistent with our previous report (7). However, there seemed to be a trend for a negative association of omental and thigh GRα mRNA with fasting insulin levels that would be interesting to confirm in larger cohorts.
Visceral fat adipocytes are more resistant to insulin's anti-lipolytic effects compared with subcutaneous adipocytes (35). Transgenic mice overexpressing 11βHSD1 in fat develop central obesity attributed to adipocyte hypertrophy, particularly in the mesenteric depot (12). To our knowledge, this is the first study that addressed the relationship of 11βHSD1 or GRα mRNA levels with regional fat cell size. Consistent with previous studies (36)(37)(38), we found larger adipocytes in the abdominal subcutaneous compared with omental depot, indicating higher fat storage in this depot in women. We showed a strong positive correlation between 11βHSD1 and adipocyte size in the omental depot, independent of obesity. Despite the negative association of GRα mRNA with omental fat cell size, this was not independent of obesity. Although 11βHSD1 was equally high in the abdominal subcutaneous and omental depot, we did not find an association between 11βHSD1 mRNA levels and abdominal fat cell size, despite the strong positive correlation with BMI. This may seem counterintuitive. The reasons for this might be our limited power to detect a possible correlation because we had fewer abdominal subcutaneous compared with omental depot biopsies. Alternatively, 11βHSD1 is not a strong predictor of subcutaneous hypertrophy in our cohort. This is supported by the lack of correlation between 11βHSD1 and fat cell size in the thigh and gluteal depots. In contrast to findings in the omental depot, GRα mRNA was negatively associated with fat cell size in the abdominal subcutaneous depot, independent of obesity. It would seem that obesity is associated with a down-regulation of GRα that is expected to limit the hypertrophic effects of GCs in adipose tissue (21)(22)(23). However, our data suggest that this GR down-regulation cannot entirely counteract increased ligand levels, particularly in the omental depot. It will be important to determine whether and how altered GR levels can indeed reach a functionally limiting state with respect to adipose hypertrophy in obesity.
In conclusion, 11βHSD1 mRNA is a strong predictor of omental fat hypertrophy, whereas GR mRNA is negatively associated with obesity.
This work was supported by the Wellcome Trust (PhD studentship awarded to ZM; Programme Grant to J.R.S., B.R.W., and K.E.C.; Intermediate Fellowship to N.M.M.), DK45343 (P.I.M. Jensen), DK50456, and RR-0585 from the U.S. Public Health Service and by the Mayo Foundation. We are grateful to the staff of the Genetics Core Laboratory, Wellcome Trust CRF, Western General Hospital, Edinburgh, for assistance in the real-time polymerase chain reaction assays and to members of the Endocrinology Unit for helpful discussions.
Nonstandard abbreviations: GC, glucocorticoid; 11βHSD1, 11β-hydroxysteroid dehydrogenase type 1; GRα, glucocorticoid receptor α; HOMA-IR, homeostasis model of assessment insulin resistance index; VAT, visceral adipose tissue; SAT, subcutaneous adipose tissue.
The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.