Glucocorticoid Regulation of P450 Aromatase Activity in Human Adipose Tissue: Gender and Site Differences

The distinct gender-specific patterns of fat distribution in men and women (android and gynoid) suggest a role for sex steroids. In keeping with these observations, it has been suggested that estrogens can promote preadipocyte cell proliferation and/or differentiation. The enzyme aromatase P450 is responsible for the conversion of androgen precursor steroids to estrogens and may, therefore, have a role in regulating adipose tissue mass and its distribution. We have investigated the glucocorticoid regulation of aromatase expression in human adipose tissue, specifically to define any site- and gender-specific differences. Abdominal subcutaneous (Sc) and omental (Om) adipose tissue was obtained from male and female patients undergoing elective surgery. After collagenase digestion, preadipocytes were cultured in serum-free medium, for 6–10 d, until confluent with either cortisol (10⁻⁶m, 10⁻⁷m) or insulin (500 nm) or a combination of both treatments. Adipocytes were studied in suspension cultures. Aromatase activity was assessed using tritiated [1β-³H]-androstenedione as substrate. In Sc preadipocytes, basal aromatase activity increased in females from 11.5 ± 1.4 (mean ± sem) to 28.0 ± 1.8 pmol/mg·h (n = 17, P < 0.05) with 10⁻⁶m cortisol. By contrast, in males, aromatase activity was inhibited by 10⁻⁶m cortisol (19.4 ± 2.4 pmol/mg·h vs. 7.5 ± 1.3, n = 9, P < 0.01; men vs. women, P < 0.005). These data were endorsed through Western blot analysis using an in-house antihuman aromatase antibody, which recognized a specific 55-kDa species. Aromatase activity was less at Om sites in preadipocytes, increasing in females from 1.1 ± 0.2 to 3.2 ± 0.7 pmol/mg·h with 10⁻⁶m cortisol (P < 0.05) and in males from 2.6 ± 0.1 pmol/mg·h to 7.8 ± 0.3 pmol/mg·h after cortisol (men vs. women, P < 0.001). Cortisol-induced aromatase activity in Om adipocytes from postmenopausal females was higher than that in premenopausal females (P < 0.001). Insulin had no independent effect on aromatase expression, but coincubation of preadipocytes with cortisol and insulin eliminated both gender- and site-specific differences. In conclusion, in women, but not men, cortisol increased aromatase activity at Sc sites, and this may facilitate predilection for Sc adiposity in females. The observed site-, gender-, and menopausal-specific differences in the glucocorticoid regulation of this enzyme may contribute to the gender- and menopausal-specific patterns of fat distribution.

BODY FAT IS distributed in a distinct gender-specific pattern, with men accumulating upper body fat (central or male-pattern obesity) and women gluteofemoral fat (1). Male pattern distribution of adipose tissue has been associated with premature mortality attributable to an increased prevalence of type 2 diabetes, dyslipidemia, and coronary heart disease (2, 3). Women are relatively protected against these conditions until the menopause, which is often accompanied by an increase in abdominal fat (3). The molecular mechanisms underlying the regulation of adipose tissue distribution remains poorly understood, though this is likely to be regulated by hormonal factors.

The gender- and depot-specific regulation of adipose tissue mass is achieved through the hormonal regulation of a number of processes affecting adipocyte number and size. Sex steroids, glucocorticoids, and insulin are all known to regulate adipose tissue mass though their effects on proliferation and or differentiation of preadipocytes in addition to their effects on lipid metabolism in mature adipocytes (4–7). Previous studies have determined that there are differences in the number of glucocorticoids receptors between subcutaneous (Sc) and omental (Om) adipose tissue (8). Furthermore, depot-specific differences in expression of the 11β−hydroxysteroid dehydrogenase type 1 (11β-HSD1) isoenzyme, a key enzyme in the prereceptor metabolism of glucocorticoids, has been described, with Om adipose tissue demonstrating higher activity (9). These findings may explain the predilection of glucocorticoids to promote visceral obesity, but the direct effects of glucocorticoids do not explain the gender-specific differences in fat distribution.

E2 regulates adipose tissue mass by increasing proliferation of preadipocytes and the size of mature adipocytes through its effects on lipolysis and lipogenesis (4, 7). In humans, a significant proportion of sex steroids are synthesized in adipose tissue, from the precursor steroids, dehydroepiandrosterone and androstenedione (10–12). Thus, the local concentration of E2 within adipose tissue is largely determined by the expression of the enzyme p450 aromatase, responsible for the conversion of androstenedione to estrone and T to E2 (10, 11). Therefore, gender- and fat depot-specific differences in the activity of aromatase may underpin depot- and gender-specific patterns of fat distribution mediated via local estrogen biosynthesis. Indeed, previous in vitro studies have revealed regional differences in aromatase expression in women, among Sc abdominal, thigh, and buttock adipose tissue, as well as in other tissues such as bone and brain (13, 14). Tissue-specific aromatase expression is mainly achieved by alternative expression of different promoters (I.4 and I.3 for adipose tissue), activated by hormones, growth factors, and cytokines (15–18). Characterization of the aromatase P450 gene upstream of the adipose-specific promoter, exon I.4, has shown that it contains a glucocorticoid response element, indicating the potential for glucocorticoid regulation of P450 aromatase (15). Glucocorticoids, therefore, may influence sex steroid metabolism by regulating aromatase expression in adipose tissue in a site- and gender-specific pattern.

The aims of this study were, therefore, to examine aromatase expression in male and female (pre- and postmenopausal) Sc and Om adipose tissue and investigate whether there were gender- and site-specific differences in its regulation by glucocorticoids.

Materials and Methods

Tissue culture

Sc abdominal adipose tissue was obtained from 26 subjects: 7 premenopausal women [age, 39.0 ± 4.6 (mean ± sd); weight, 77.6 ± 16.4 kg], 10 postmenopausal women (age, 57.7 ± 6.9; weight, 66.5 ± 11.2 kg), and 9 men (age, 48.8 ± 17.3; weight, 78.2 ± 12.2 kg). Om adipose tissue was obtained from 21 subjects: 6 premenopausal women (age, 39.0 ± 5.3; weight, 73.0 ± 13.7 kg), 6 postmenopausal women (age, 58.7 ± 6.0; weight, 69.1 ± 11.7 kg), and 9 men (age, 56.4 ± 25.9; weight, 73.6 ± 8.9 kg). All patients were either undergoing elective surgery or cosmetic surgery in accordance with guidelines of the local hospital ethical committee. Subjects on endocrine therapy (e.g. steroids, HRT, T₄) or antihypertensive therapy were excluded.

Fresh abdominal Sc adipose tissue (1- to 2-g wet weight) was collected. Tissue was initially washed with 1× HBSS containing penicillin (100 U/ml) and streptomycin (100 μg/ml). Visible blood vessels and connective tissue were removed, and the tissue was finely chopped. All adipose tissue was digested with the same batch of collagenase class 1 (2 mg/ml, Worthington Biochemical Corp., Reading, UK) in 1× HBSS (Life Technologies, Inc., Paisley, UK), for 1 h at 37 C in a water bath, and shaken at 100 cycles/min at 37 C (16). The disrupted tissue was filtered through a double-layered cotton mesh, and preadipocytes and adipocytes were separated by centrifugation at 360 × g for 5 min.

Preadipocyte cell isolation

Preadipocyte cells were resuspended in erythrocyte lysis buffer (0.154 m NH₄Cl, 10 mm KHCO₃, 0.1 mm EDTA) for 10 min, to remove erythrocyte contamination, and centrifuged at 360 × g for 5 min. Pellets of the preadipocyte cells were plated in tissue-culture dishes in DMEM/nutrient mix F-12 (DMEM:F12) (Life Technologies, Inc.) medium supplemented with 15% FBS. After 20 h, the medium was changed. Preadipocyte cells were cultured in serum-free conditions in DMEM:F12 containing transferrin (5 μg/ml), penicillin (100 U/ml), and streptomycin (100 μg/ml) for 6–10 d, until confluent, in concentrations of cortisol (10^−6--10⁻⁷m) and/or insulin (500 nm).

Mature adipocyte isolation

After centrifugation, the upper layer of mature adipocytes was removed from the collagenase-dispersed preparation and washed in DMEM:F12 twice and centrifuged at 360 × g for 2 min. Adipocytes were maintained for 20 h at 37 C, in a 5% CO₂ incubator, in DMEM:F12 containing 1% BSA (protein fraction V; insulin removed and lipid adsorbed; First Links, Dudley, UK), penicillin (100 U/ml), and streptomycin (100 μg/ml), before being plated into tissue-culture dishes in DMEM:F12 containing 1% BSA. Adipocytes were maintained in medium containing varying concentrations of cortisol (10⁻⁶–10⁻⁷m) and/or insulin (500 nm; Sigma, Poole, UK; receiving identical treatment regimens and conditions as preadipocyte cells). Cyclodextran water-soluble cortisol (Sigma) was used for all experiments. Viability of adipocytes was assessed using trypan blue (Sigma)

Enzyme assays

Aromatase activity was measured, using a tritiated water release assay, by incubating intact preadipocytes and adipocytes in serum-free medium (4 h and 14 h, respectively) with 20 pmol [1β-³H]-androst-4-ene-3, 17 dione (Perkin Elmer Life Sciences, Wellesley, MA; specific activity, 24.3 Ci/mmol). Each experiment was carried out in triplicate, and activity was expressed as mean ± se pmol, as previously reported (17, 18).

Protein assay

Protein in each incubate was determined using the Bradford method (19) (Bio-Rad Laboratories, Inc., Preston, UK). As adipocytes were maintained in medium containing 1% BSA, separate culture tissue dishes were maintained in the absence of BSA to assess protein levels.

Synthesis of a human aromatase antibody

Synthetic peptides corresponding to amino acids 379–398 (DDVID GYPVK KGTNI ILNIG) and 480–499 (PDETK NMLEM IFTPR NSDRC) of the P450 aromatase sequence were used as immunogens (Binding Site, Birmingham, UK) (20). These sequences were synthesized as eight-branched multiantigenic peptides, which were emulsified with Freund’s complete adjutant, and used to immunize a single sheep. An IgG fraction was prepared from the immune serum, by ammonium sulfate precipitation and ion exchange chromatography.

Immunohistochemistry

Five-micron-thick formalin-fixed sections of normal adipose tissue, placenta, ovary, and liver were cut and placed on coated glass slides (Fro-Tissuer pen, Binding Site, Birmingham, UK). After dewaxing, slides were treated with methanol-hydrogen peroxide (1:1000) to block endogenous peroxidase activity. After washing in PBS (0.05 m; pH 7.6), slides were incubated with polyclonal antibody to human aromatase P450 (1:500; placenta, 1:500; ovary, 1:125; and liver, 1:250), in 10% normal swine serum, for 1 h at room temperature. Control sections included: 1) omission of primary antibody; and 2) use of primary antibody preabsorbed with the immunizing peptides. The secondary antibodies, the biotinylated universal antibody (1:100), was added to sections for 30 min. The Streptavidin biotinylated peroxidase complex (ABC, 1:100; Binding Site) was added to sections for 30 min. Slides were developed using 3,3′-diaminobenzidine and counter-stained with Mayer’s hematoxylin.

Immunoblotting

Western analysis was performed using a modification of a previously described method (21). In brief, proteins in each sample were determined using the Bradford method (Bio-Rad Laboratories, Inc.) and were separated by SDS-PAGE using 10% gels (19, 21). All samples (10–20 μg) were heated for 5 min at 95 C in sample buffer containing dithiothreitol (100 nm; Sigma). Proteins were transferred from the polyacrylamide gels to polyvinylidene difluoride membranes by electroblotting, using a vertical transfer apparatus at 425 Amp for 3 h. Membranes were blocked (1 h, 25 C) in PBS with Tween [PBS-T (PBS + 0.05% Tween 20); Sigma] containing 10% (wt/vol) nonfat milk powder (Marvel; Premier Brands Ltd, Stafford, Staffs, UK). Filters were incubated overnight at 4 C with a polyclonal aromatase primary antibody raised in sheep. The primary antibody was diluted 1:250 in PBS-T containing 0.25% BSA (fatty acid absorbed BSA; First Links UK Ltd) The antisheep secondary antibody (horseradish peroxidase-conjugated; Calbiochem, Nottingham, UK) was diluted 1:50,000 in PBS-T (0.05% T). Specific protein expression was detected as a 55-kDa band, using a sensitive chemiluminescent assay (ECL; Amersham Pharmacia Biotech, Little Chalfont, UK), after exposure to x-ray film for 1–20 min. Autoradiographs were quantified by densitometry using the Windows UVP: Gelbase/Gelblot program (UVP Ltd, Cambridge, UK). Immunizing peptides used to synthesize the aromatase antibody were used to absorb aromatase antibodies.

Statistical analysis

Statistical analysis was undertaken, using unpaired t tests for analysis of male vs. female treatment and by ANOVA for comparison of control vs. treatments, for both male and female subjects. Data are presented as mean ± sem.

Results

Regulation of aromatase: effect of site

Sc vs. Om preadipocyte cells.

In Sc preadipocytes from men (n = 9), basal aromatase activity (19.4 ± 2.4 pmol/mg·h) was significantly higher than in Om preadipocytes (2.6 ± 0.14 pmol/mg·h, P < 0.001; Fig. 1, A and B). Incubation of Sc preadipocytes from men with either 10⁻⁶m cortisol (7.5 ± 1.3 pmol/mg·h), or cortisol + insulin (7.1 ± 0.9 pmol/mg·h), revealed aromatase activity comparable with that observed in Om preadipocytes after incubation with cortisol alone (7.8 ± 0.3 pmol/mg·h) or cortisol + insulin (8.8 ± 0.43 pmol/mg·h; Fig. 1, A and B).

In Sc preadipocytes from women, basal aromatase activity (11.5 ± 1.4 pmol/mg·h) was significantly higher than Om preadipocyte basal aromatase activity from women (1.1 ± 0.2 pmol/mg·h; P < 0.001; Fig. 1, A and B). Treated Sc preadipocytes from women with cortisol (10⁻⁶m) alone (28.0 ± 1.8 pmol/mg·h) or cortisol + insulin (30.5 ± 1.7 pmol/mg·h; Fig. 1A) compared with Om preadipocytes from women also treated with cortisol alone (3.2 ± 0.7 pmol/mg·h; P < 0.001) or cortisol + insulin (4.2 ± 0.8 pmol/mg·h; P < 0.001; Fig. 1, A and B), revealed a significantly higher aromatase activity in preadipocytes from the Sc depot.

Regulation of aromatase expression: effect of gender

Sc preadipocyte cells.

Aromatase activity was higher in Sc (compared with Om) adipose tissue in both men and women, except under a certain treatment regimen, as discussed below. There was a significant difference in basal aromatase activity in preadipocyte cells from males (19.4 ± 2.4 pmol/mg·h), compared with females (11.5 ± 1.4; P < 0.05) (Fig. 1, A and B). Additionally, incubation with either cortisol alone or with insulin inhibited aromatase activity in males and stimulated activity in females, demonstrating clear gender differences (P < 0.001; Fig. 1, A and B).

Om preadipocyte cells.

There was a significant difference in basal aromatase activity in preadipocyte cells from males (2.6 ± 0.1 pmol/mg·h), compared with females (1.1 ± 0.2 pmol/mg·h; P < 0.001) (Fig. 1, A and B). Gender differences were also observed with incubation of cortisol (10⁻⁶m; male, 7.8 ± 0.3 pmol/mg·h; female, 3.2 ± 0.7 pmol/mg·h) insulin alone, or in combination (10⁻⁶m + insulin, 500 nm; male, 8.8 ± 0.4 pmol/mg·h; female, 4.2 ± 0.8 pmol/mg·h; P < 0.001; Fig. 1, A and B).

Sc vs. Om adipocytes.

Adipocytes were assessed to be more than 99% viable, as previously described (18), with aromatase activity in adipocytes consistently observed; but activity was at least 10-fold less than that observed in abdominal Sc and Om preadipocyte cells. In contrast to Sc preadipocyte cells, cortisol and insulin stimulated aromatase activity in both sexes and at both sites (10⁻⁶m + 500 nm insulin; P < 0.05). Gender differences were noted in Om adipocytes, both basally and in cells treated with insulin or cortisol alone (P < 0.05) (Fig. 2, A and B). An additive response was observed between insulin (500 nm) and glucocorticoids (10⁻⁶m) in Sc and Om adipocytes from males and females, which eliminated the gender- and site-specific response in aromatase regulation (Fig. 2, A and B). No change in additive response with insulin in combination with cortisol was observed in adipose cells from either gender, or by alteration in menopausal status in preadipocyte cells or adipocytes (Figs. 1 and 3³).

Regulation of aromatase expression: effect of menopausal state

Preadipocyte cells and adipocytes.

In Sc preadipocyte cells and adipocytes from pre- and postmenopausal females, there was no significant difference in aromatase expression, either in the basal state or after incubation with cortisol alone or in combination with insulin (Fig. 3, A and B). Basal aromatase activity was significantly higher in Om preadipocytes from postmenopausal women, compared with premenopausal women (1.9 ± 0.3 pmol/mg·h vs. 0.5 ± 0.1 pmol/mg·h; P < 0.01). Cortisol (10⁻⁶m) significantly increased aromatase activity in adipocytes from both pre- (1.4 ± 0.2 pmol/mg·h; P < 0.05) and postmenopausal (5.9 ± 1.6 pmol/mg·h; P < 0.05) females. Menopausal status did affect the stimulatory response, postmenopausal females showing significantly increased aromatase activity, compared with premenopausal females in Om adipocytes (P < 0.001) (Fig. 3B). In Sc adipocytes from pre- and postmenopausal females, there was no significant trend identified between the two groups (Fig. 3A). No significant gender differences were observed, with age, for either Sc or Om preadipocyte cell aromatase activity (Figs. 4A and 5A⁵). Additionally, no significant gender differences were observed with adiposity, for Sc or Om preadipocyte cell aromatase activity (Figs. 4B and 5B⁵).

Determination of specificity of P450 aromatase antibody within human tissues.

Aromatase immunoreactivity was analyzed in normal placenta, to confirm the specificity of the antibody sections with the antihuman aromatase antibody, and revealed intense staining within the syncytiotrophoblasts (Fig. 6A). No staining was observed when the antibody was preabsorbed with the immunizing peptides at a dilution of 1:500 (Fig. 6B). Human ovary was similarly assessed for the expression of the aromatase immunoreactive protein, which was localized to the granulosa cells surrounding the developing oocyte (Fig. 6C). No staining was noted from the ovary sections where the primary antibody was preabsorbed with the immunizing peptides at a dilution of 1:500 (Fig. 6D). Adult liver was used as a negative control for the aromatase immunoreactive protein (Fig. 6, E and F).

Expression of aromatase P450-aromatase protein within human tissues by Western analysis.

Western blot analysis confirmed an increase in aromatase immunoreactivity in Sc preadipocytes from females treated with cortisol in combination with insulin, compared with control (Fig. 7) [cortisol (F, 10⁻⁷m) + insulin (Ins, 500 nm), 2.1 ± 0.3 vs. 1.0; P < 0.01] and a reduction in Sc preadipocytes from males treated with cortisol in combination with insulin compared with control [(F, 10⁻⁷m) + (Ins, 500 nm), 0.2 ± 0.1 vs. control 1.0; P < 0.05]. Western analysis demonstrated an increase in aromatase expression in Om preadipocyte cells from males and females treated with cortisol and insulin [male, control (1.0 ± 0.0), F (10⁻⁷m) + Ins (500 nm), 2.4 ± 0.15, P < 0.01; female, control (1.0 ± 0.0), F (10⁻⁷m) + Ins (500 nm), 2.90 ± 0.65, P < 0.05]. No immunoreactive proteins were observed when Western analysis was carried out using primary antibody preabsorbed with the immunizing peptide (data not shown).

Discussion

The distinct gender-specific patterns of body fat distribution suggest that sex steroids may play a role in regulating this process. Sex steroids may regulate fat mass by increasing preadipocyte proliferation, differentiation, and adipocyte size through effects on lipolysis and lipogenesis (4). In vivo and in vitro studies suggest that administration of estrogen is associated with a redistribution of adipose tissue from Sc to the Om depot, which may be mediated through alterations in adipose tissue preadipocyte cell proliferation, differentiation, lipoprotein lipase activity, and adipocyte cytokine release (4, 5, 22, 23). However, studies examining the relationship of circulating levels of sex steroids to adipose tissue mass and distribution have proven inconclusive (14). The extragonadal sites for estrogen production possess several features unique from the ovaries, principally that estrogen in bone, brain, and breast adipose tissue is probably only active at a local paracrine or intracrine level (24). Thus, although the combined levels of E2 synthesized in extragonadal sites may represent relatively low systemic levels, the local tissue concentrations are thought to be high, and may have major local biological influence (24).

Previous studies have investigated the regulation of aromatase by cytokines, growth factors, and hormones in human adipose tissue and bone, as well as studies conducted on the aromatase knockout mice (13, 25–27). In humans, the link between aromatase and adipose tissue distribution is suggested by studies where patients lack the P450 aromatase gene. Female patients with point mutations in the aromatase gene (28) failed to develop breast adipose tissue, whereas affected males developed a eunuchoid skeleton. Conversely, the syndrome of apparent aromatase excess, because of a different point mutation of the aromatase gene, leads to development of a feminine body habitus in men (29). Furthermore, in the aromatase knockout mouse, there is an associated loss of Sc adipose tissue and accumulation of visceral fat (30).

Most studies to date that analyze P450 aromatase expression and regulation in adipose tissue have used tissue derived only from females (14). In the present study, we report the gender- and site-specific regulation of estrogen metabolism by glucocorticoids in human abdominal adipose tissue. Aromatase expression in Sc preadipocyte cells from male subjects was inhibited by cortisol; whereas, in females, cortisol stimulated aromatase activity. Changes in protein expression were confirmed using an in-house antibody against human P450 aromatase. Furthermore, this effect was site-specific, with Om preadipocyte cells demonstrating lower aromatase expression, compared with Sc preadipocyte cells, which has also been previously observed (31). Aromatase activity at Om sites was, however, increased by coincubation with increasing doses of glucocorticoid and insulin in both sexes.

Previous in vivo studies by Longcope and colleagues (24, 32), measuring the urinary excretion of tritiated estrone, support the observed gender-specific differences in total aromatase activity, but these experiments did not address the question of regulation. However, support for a gender-specific pattern of regulation of abdominal adipose tissue metabolism has been provided by studies examining the role of sex steroids and glucocorticoids in regulating lipolysis and leptin production (23, 33). Adipose tissue mass may also be regulated by glucocorticoids in a site- and gender-specific pattern, through their prereceptor metabolism by the enzyme 11β-HSD or by differences in the expression of the GRs (9, 34). Preliminary data of 11β-HSD1 dehydrogenase and reductase activity suggest no significant gender difference (data not shown). However, 11β-HSD1 isoenzyme is increased in Om fat, compared with Sc fat. This would potentially increase glucocorticoid concentration at this site and suggests that other factors may be important for the induction of P450 aromatase in Sc adipose tissue. The relative expression in aromatase activity may result from known gender-specific differences at the level of the receptor. GR mRNA expression has been shown to be increased in female Sc adipose tissue, compared with male Sc adipose tissue (8, 35, 36). There is also evidence that aromatase is directly affected by GR, indicating additional regulatory influence by glucocorticoids at the receptor level (37).

Investigation of the influence of age on aromatase activity determined that, at least in Om adipose tissue, menopausal status altered the glucocorticoid regulation of aromatase expression. Adipose tissue from postmenopausal (but not premenopausal) women was stimulated by cortisol in Om preadipocyte cells. Both in vitro and in vivo studies, however, suggest that Om fat has a relatively low aromatase activity. However, although Om fat only accounts for 6–20% of total adipose tissue, previous data suggest that it is more deleterious to health and most strongly associated with insulin resistance and, as such, at this site, aromatase expression may have a greater impact on health (3, 38). Further enhanced aromatase activity in postmenopausal female Om preadipocytes may arise because of an intrinsic cell alteration in estrogen metabolism and aromatase expression after the menopause, as fat becomes the main source for estrogen synthesis. Whereas Sc aromatase metabolism is maintained in the postmenopausal females, there is an increased requirement for additional estrogen production to maintain circulating estrogen concentration. This increased aromatase activity in Om preadipocytes from postmenopausal females may also be associated with a detrimental effect on increasing adiposity locally, through proliferation and differentiation of preadipocytes. In addition, the trends observed with age and weight in this present study largely support previous studies showing that aromatase activity increases with age and adiposity (25).

In both Sc and Om adipocytes, we demonstrated significant site- and gender-specific differences in basal aromatase activity, which were eliminated when cells were treated with cortisol (10⁻⁶m) in combination with insulin. It was important to demonstrate that our aromatase activity findings in adipocyte samples were not attributable to preadipocyte cell contamination; several approaches were therefore undertaken to safeguard from this occurrence. For both this study and our previous studies, adipocytes were purified from preadipocyte cells by washes and centrifugation (18). Further, we were unable to detect aromatase activity in cultured preadipocyte cells when less than 30% confluent (data not shown), suggesting that it is unlikely that contamination with a few preadipocyte cells would explain the observed aromatase activity in adipocytes (18). The gender-specific regulation of aromatase activity identified in the preadipocyte cell fraction also seems to be different from that observed in the adipocyte fraction. This is further reinforced by our previous data from a cortisol dose response study in which differences in relative expression of aromatase between the sexes were observed in adipocytes (18). Finally, photographs were taken of each adipocyte sample for assessment of cell viability; this failed to demonstrate the presence of any preadipocyte cells. As such, we believe our data accurately reflect adipocyte aromatase activity.

Previous reports suggest that insulin plays an important role in adipocyte proliferation and differentiation (18, 36). This study indicates that insulin alone has no effect on aromatase expression, whereas insulin in combination with cortisol (10⁻⁶m) demonstrated an additive effect on aromatase activity, eliminating site- and gender-specific difference in mature adipocyte aromatase activity. This suggests that relatively high doses of insulin in combination with cortisol increase aromatase activity, to promote estrogen production and, at a paracrine level, may further promote proliferation and differentiation of preadipocytes, enhancing central adiposity (18, 36). The loss of site- and gender-specific effects on aromatase activity in the mature adipocytes, as a result of the cortisol and insulin combination, may result from changes in GR expression or glucocorticoid prereceptor signaling. To date, no study has addressed the influence of insulin on GR regulation in adipose tissue. However, studies on glucocorticoid regulation of insulin receptor in lymphocytes have shown that glucocorticoids directly regulate insulin receptor gene transcription, indicating a potential for cortisol-induced regulation of insulin action in adipose tissue (38). Assessment of in vivo studies of glucocorticosteroid treatment has shown down-regulation of GR expression in isolated mature adipocytes (39). Furthermore, no gender-specific GR expression was identified after treatment (39). Additionally, glucocorticoid prereceptor signaling analysis of insulin, in combination with cortisol, in myocyte cells has shown that 11βHSD1 activity is enhanced more than with cortisol alone (40). Therefore, as 11βHSD1 may maintain tissue sensitivity to glucocorticoids by modulating the accessibility to functional GR, as shown in previous in vitro studies using cultured human preadipocyte cells and myoblast cells, insulin may, at the prereceptor level, influence glucocorticoid regulation, to eliminate the observed gender- and site-specific differences in mature adipocytes (9, 41, 42).

In conclusion, we have demonstrated that aromatase activity is higher in Sc preadipocyte cells than in Om preadipocyte cells. Second, cortisol regulates aromatase expression, and hence estrogen production, in a gender-specific manner, with aromatase activity being increased in female, and reduced in male, Sc preadipocyte cells. The pattern of aromatase regulation is also altered by menopausal status in Om preadipocyte cells. In mature adipocytes, aromatase activity is stimulated by insulin in combination with cortisol, eliminating gender- and site-specific differences. These findings suggest that the differences in glucocorticoid regulation of aromatase may form part of a gender-specific and menopausal mechanism to account for the characteristic patterns of increased Sc observed in women, and increased Om fat mass in postmenopausal women and men.

Acknowledgements

We thank Dr. A. R. Bradwell and Dr. P. Stubbs, from the Binding Site, for the synthesized aromatase antibody. We also thank Miss Keely Jenner and Mrs. Jane Starczynski, who kindly performed immunocytochemistry on ovarian tissue. We thank all the operative surgeons and theater staff at both the University Hospitals Trust and the Womens Hospital Trust; and Mr. Levick, at the Priory Hospital Edgbaston, who kindly provided fat samples for the studies.

This work was supported by the British Diabetic Association (Grant RD96/0001268), Eli Lilly \|[amp ]\| Co. Industries (Surrey, UK), and the MRC. Dr. A. Anwar was supported by a Wellcome research fellowship (no. 051133/171). Professor P. M. Stewart is an MRC Senior Clinical Fellow.

Abbreviations:

 
• 11β-HSD,

  11β-Hydroxysteroid dehydrogenase;

 
• F,

  cortisol (in references to Fig. 7);

 
• Ins,

  insulin (in references to Fig. 7);

 
• Om,

  omental;

 
• PBS-T,

  PBS + 0.05% Tween 20;

 
• Sc,

  subcutaneous.