Adipose tissue is an important source of angiotensinogen (ATG), and hypertension is commonly associated with android obesity. Therefore, we tested the hypothesis that androgens may control ATG gene expression and secretion in rat fat cells. In intact male rats, ATG mRNA expression (Northern blot and co-reverse transcription-polymerase chain reaction analysis) and protein secretion were significantly higher in deep intra-abdominal (perirenal and epididymal) than in subcutaneous adipocytes. After castration, ATG mRNA was reduced almost 50% in the three fat deposits, with parallel changes in ATG protein secretion. Conversely, testosterone treatment fully restored the ATG mRNA decrease after castration, whatever the anatomical origin of the adipocytes. Finally, a 24-h in vitro exposure of perirenal fat cells or differentiated preadipocytes from castrated rats to testosterone or dihydrotestosterone (10 nM free hormone concentration) increased ATG mRNA expression by 50–100%, an effect that was prevented by the anti-androgen cyproterone acetate. These data, demonstrating both in vivo and in vitro androgen induction of ATG mRNA expression in rat adipocytes, add further weight to the hypothesis of a link between adipose tissue ATG production, androgens, and android obesity-related hypertension.
white adipose tissueis an important source of angiotensinogen (ATG) (4, 5, 15, 17,38, 39, 45), the unique substrate of renin in the renin angiotensin system (RAS) and the precursor of angiotensin II (ANG II) (for a review, see Ref. 34). ANG II plays a crucial role in blood pressure regulation and in fluid volume balance (4). Various tissues express their own local renin-angiotensin system (RAS), which is regulated independently of the systemic RAS and may affect local organ function (4, 26). Adipose tissue has also been shown to be a major source of ANG II (40), which elicits a variety of physiological effects after its binding to specific membranous receptors, the ANG II receptors (21, 40). Although the physiological function of the adipocyte RAS has not yet been delineated, the role played by the adipocyte-derived ANG II has been assigned to the control of adipose tissue growth and development. As a matter of fact, when bound to the type 1 ANG II receptor, ANG II stimulates the production of prostacyclin (PGI2), which in turn triggers preadipocytes to differentiate into adipocytes and enhances the terminal differentiation of preadipocytes (10). Moreover, ANG II increases lipid synthesis and storage in adipocytes (23).
Little is known about the regulation of the RAS in adipose tissue. Although various biological signals, including circulating estrogens, triiodothyronine, cytokines, and ANG II, are known to promote hepatic ATG synthesis (25, 27, 36, 44), glucocorticoids and insulin are the only hormonal factors that were reported to modulate ATG expression in adipocytes (1, 2, 32).
Epidemiological studies have clearly demonstrated a close correlation between body weight and increased risk of cardiovascular diseases (19). Although most obese patients have high blood pressure, the mechanisms linking obesity to hypertension are still poorly understood. The discovery that adipose tissue expresses and secretes ATG, together with the finding that obese patients have abnormally high circulating ATG levels (47), has led to the suggestion of the possible involvement of the adipocyte-derived ATG in the pathogenesis of obesity-related hypertension (17). In this respect, a precise knowledge of the hormonal control of ATG synthesis and secretion from adipocytes appears of considerable interest.
The aim of the present study was to determine whether androgens elicit in vivo and in vitro any regulatory effect on ATG gene expression and secretion in rat adipose tissue. Androgens were chosen because1) adipocytes express androgen receptors (11,13); and 2) hypertension, like fat mass distribution, conforms to a sexual dimorphism illustrated by the finding that the risk of developing hypertension is much lower in women before menopause than in men but after menopause is the same as in men (3, 22). Furthermore, because the metabolic activities and androgen receptor densities of adipocytes are variable according to the anatomical origin of these cells (12, 13), this study was extended to a comparison of ATG expression and secretion by rat adipocytes from various locations.
Animals and experimental protocols.
Adult male Sprague-Dawley rats were kept under controlled lighting conditions (light: 6 AM, dark: 8 PM) and had free access to a standard diet and water. Animals (150 g) were castrated (CAST) or sham-operated (controls or SHAM) under pentobarbital sodium anesthesia (40 mg/kg ip) as previously described (29, 30) and were killed by decapitation 2 wk later. At the time of death, body weights of SHAM and CAST rats were 340 ± 38 and 310 ± 26 g, respectively. In some experiments and 5 days after the operation, one-half of the castrated rats received one injection of testosterone propionate (0.5 mg/100 g body wt) every other day for 10 days, while the other one-half (CAST) received the vehicle (polyethylene glycol) only.
Isolated adipocyte and preadipocyte preparation and culture.
Mature adipocytes were obtained from subcutaneous (sc), perirenal (peri) and epidydimal (epi) fat pads by use of the collagenase digestion procedure described by Rodbell (35). Briefly, fat pads were removed aseptically, minced, and digested at 37°C with vigorous shaking in DMEM-Ham's-F12 (8 ml/g tissue, Sigma, St. Louis, MO) containing 2% (wt/vol) BSA (Sigma), 0.5–1 mg/ml collagenase (Worthington Biochemical, Lakewood, NJ), 0.1 mg/ml streptomycin, and 100 U/ml penicillin. After 30 min of incubation, the digests were filtered through a 250-μm sterile nylon mesh cone and centrifuged at 200 g for 3 min at 25°C. Floating adipocytes were washed three times with DMEM-Ham's-F12 containing 1% wt/vol BSA, 0.1 mg/ml streptomycin, and 100 U/ml penicillin. Aliquots (1–2 ml) of the resulting adipocyte suspension were rapidly diluted in a standard acid guanidinium isothiocyanate solution (7), centrifuged at 3,000g for 3 min at 4°C to remove lipids, and kept at −80°C. The infranatant containing the stroma vascular fraction was successively filtered through 150- and 25-μm nylon screens, and the filtrate was centrifuged at 600 g for 10 min. After two washes, the preadipocytes were plated at a density of 1–2 × 104 cells/cm2 in cell culture dishes with DMEM-8% fetal bovine serum (FBS), streptomycin (0.1 mg/ml), and penicillin (100 IU/ml). After 12 h, cells were washed and maintained under the same conditions for 3 days. Medium was changed every day. Then, cells were allowed to differentiate in DMEM-Ham's-F12 supplemented with insulin (5 μg/ml), transferrin (10 μg/ml), T3 (2 nM), and antibiotics (0.1 mg/ml streptomycin and 100 IU/ml penicillin). After 6 days, >70% of the cultured cells appeared under the microscope as lipid-filled differentiated adipose cells.
Total RNA was extracted and purified following the guanidinium isothiocyanate procedure described by Chomczynski and Sacchi (7). The yield and quality of extracted RNA were assessed by the 260-/280-nm optical density ratio and by electrophoresis under denaturing conditions on 1% agarose gel. RNA preparations were stored at −80°C.
Adipocyte total RNA (0.1 μg) was denatured for 10 min at 72°C. RT was then performed in a final volume of 10 μl for 60 min at 42°C and stopped by 5 min at 95°C. Composition of the reaction mixture was as follows: 50 mM Tris · HCl, 75 mM KCl, 3 mM MgCl2, 10 mM dithiothreitol, 0.5 mM of each dNTP (Pharmacia Biotech, Pitscataway, NJ), 62.5 mU RNAguard (Pharmacia Biotech), 50 ng random hexamers, and 100 IU Superscript II RNase H-reverse transcriptase (Life Technologies, Grand Island, NY). Controls without reverse transcriptase were systematically performed to detect an eventual genomic DNA contamination.
To verify that cyclophilin (CYC) mRNA did not vary as a function of hormonal status (gonadectomy with or without hormonal substitution) or anatomical locations, co-reverse transcription (coRT) with 0.1 μg of an exogenous mRNA (pBR322 transcript) used as a standard was performed as described previously (43).
Reverse transcripts were amplified as follows. The optimal annealing temperature was determined for each primer pair. The standard amplification mixture contained 50 mM KCl, 10 mM Tris · HCl (pH 9.0), 1.5 mM MgCl2, 200 μM of each dNTP (Pharmacia Biotech), 2 μl of RT products, 2 IU Taqpolymerase (Pharmacia Biotech), and primers (Pharmacia Biotech), at concentrations specified below, in a 50-μl final volume.
For CYC and pBR322 coamplification, the appropriate primers (see below) (43) were used at 0.25 μM. For ATG and CYC coamplification, the primers were used at 0.25 μM and 0.2 μM, respectively.
Amplification reactions were conducted for 25–27 cycles under the following conditions: initial denaturation step at 94°C for 4 min, then for 30 s at 94°C, followed by a 30-s annealing step at 57°C and a 45-s elongation step at 72°C. A final elongation step was performed for 7 min at 72°C. Thermal cycling was carried on a GeneAmp PCR System 2400 (PE Applied Biosystems, Norwalk, CT).
Primers used for the rat ATG gene were sense 5′-TGAGGCAAGAGGTGTAGC-3′ and antisense 5′-AAGATGGCGGGGGTGAAG-3′, and PCR generated a 336-bp cDNA fragment. Rat CYC gene-specific primers were sense 5′-GGGAAGGTGAAAGAAGGCAT-3′ and antisense 5′-GAGAGCAGAGATTACAGGGT-3′. The CYC expected cDNA product size was 210 bp.
PCR products (10 μl) were analyzed on a 1.5% agarose gel stained with ethidium bromide, photographed, and analyzed by Bio-Imager software. Semiquantitative data were expressed as the ratio of the ATG/CYC signals.
Assay for ATG protein secretion.
Aliquots (0.5–1 ml) of the adipocyte suspension were rapidly dispensed and incubated for 24 h at 37°C under 5% CO2-95% air atmosphere in 5–10 ml of DMEM-Ham's-F12 medium containing BSA (1.5%), vitamin E (4 mg/ml), penicillin (100 U/ml), streptomycin (0.1 mg/ml), and antiproteolytic agents [100 μM phenylmethylsulfonyl fluoride (PMSF), 2 μM leupeptin, 25 μg/ml aprotinin]. Culture medium was collected 24 h later for ATG assay in the presence of 10 μM captopril and was stored at −80°C. In parallel and to check possible cellular lysis and unviability, lacticodehydrogenase (LDH) activity released into the culture medium and cellular glucose uptake were determined but showed no differences between incubation sets. ATG was indirectly determined by following the generation of ANG I in the presence of an excess of hog renin (EC188.8.131.52, 50 μIU/50 μl; 60 min at 37°C; Sigma) and a protease inhibitor (PMSF) to protect generated ANG I. ANG I was then measured by direct radioimmunoassay (REN-CT2, CIS Bio International, Gif sur Yvette, France). The inter- and intra-assay coefficients of variation were <10%. The cross-reactivity of the ANG I antibody was <0.01% for ANG II and III. The sensitivity of the assay was 0.15 ng/ml.
Northern blot analysis.
Total RNA (10 μg) was electrophoresed through 1% agarose and 12.5% formaldehyde denaturating gels and transferred to nylon membrane (Hybond, Amersham) in NaOH (0.05 N) for 3 h and cross-linked to this membrane (1 h at 80°C). The blots were prehybridized at 68°C for 3 h in Church solution (0.5 M sodium phosphate, 1 mM EDTA, and 7% SDS) (8) and then hybridized with [32P]-labeled cDNA probes labeled by random priming (2–5 × 108 dpm/μg) (Megaprime, Amersham) at 68°C overnight. After hybridization, blots were washed twice in 2× standard sodium citrate (SSC; 30 mM Na3-citrate, 300 mM NaCl, pH 7) and 0.1% SDS at room temperature for 15 min and once at 45°C for 30 min before being exposed to X-ray film at −80°C. The autoradiographic signal densities were quantitated by an optical densitometer (Helena Laboratory, St Leu la Forêt, France). To quantitate total loaded RNA, the [32P]-labeled ATG cDNA probe was stripped from the membrane after a boiling solution of 0.5% (wt/vol) SDS was poured and then allowed to cool to room temperature. The blots were then incubated in 4 mM NaOH at 45°C for 30 min and washed in 0.1× SSC, 0.1% (wt/vol) SDS and 2 mM Tris · HCl (pH 7.5) for an additional 15 min. Membranes were then prehybridized and rehybridized with a [32P]-labeled ribosomal acidic protein (RP) cDNA probe (28) under the conditions specified above.
To check quality and quantity of loaded RNA, parallel gels were run and stained with ethidium bromide (0.3 mg/ml) with 2 μg of total RNA to visualize ribosomal RNAs (28S and 18S). Specific rat ATG and RP probes were prepared by RT-PCR from adipocyte total RNA. Probes were then extracted by the phenol-chloroform extraction procedure.
Cell numbers were calculated from the mean fat cell size and the lipid content of a known volume of the fat cell suspension as previously described (20). LDH activity and glucose concentrations in the incubation medium were assessed as previously described (29).
Results are expressed as means ± SE of at least three individual experiments. One-way ANOVA was performed for differences in ATG mRNA levels between experimental groups. If the one-way ANOVA revealed a significant relationship, a Student's t-test was performed to detect significance (P < 0.05).
In the present study, ATG mRNA expression was investigated using either Northern blot analysis in the in vivo experiments or coRT-PCR in the in vitro experiments, where Northern blot could not be applied because of the limited number of cells. As shown in Fig.1, A-C, coRT-PCR of CYC performed in the presence of exogenous pBR322 mRNA as control under the conditions previously described (43) demonstrated that CYC mRNA does not vary between the various adipose tissue localizations or as a function of hormonal status (gonadectomy with or without hormonal substitution). Thus CYC was chosen as an internal standard for the subsequent coRT-PCR experiments. For Northern blot experiments, the RP, which does not vary between the various anatomical localizations of adipose tissue (Fig. 1, B,D, and E), was used as internal standard.
ATG mRNA levels were compared by Northern blot (Fig.2 A) and semiquantitative coRT-PCR (Fig. 2 B) in isolated adipocytes from intact male rat subcutaneous, epididymal, and perirenal fat deposits. Semiquantitative data were expressed either as the ATG/RP or the ATG/CYC densitometric signal ratios. As can be seen in Fig.2 A, Northern blot analysis revealed that ATG mRNA levels are significantly higher in perirenal and epididymal than in subcutaneous adipocytes (1.59 ± 0.24- and 1.91 ± 0.24-fold increases). These findings were confirmed by coRT-PCR analysis (Fig.2 B), showing again higher ATG mRNA levels in perirenal and epididymal than in subcutaneous adipocytes considered as a reference (1.33 ± 0.1- and 1.82 ± 0.24-fold increases). To determine whether these ATG mRNA site specificities were accompanied by parallel changes in ATG protein production, perirenal, epididymal, and subcutaneous adipocytes were maintained in primary culture for 24 h, and accumulation of secreted ATG was then assayed in the culture medium. Figure 2 C shows significantly higher ATG protein release from perirenal and epididymal than from subcutaneous adipocytes chosen as the reference (2.73 ± 0.62- and 2.00 ± 0.18-fold increase in adipocyte primary culture media, respectively).
We next examined the influence of male rat gonadal status on ATG expression in adipocytes from perirenal, epididymal, and subcutaneous adipose tissues. As can be seen in Fig.3 A, after castration, ATG mRNA levels were reduced by almost one-half in adipocytes from the three regions (61, 47, and 53% of SHAM values in subcutaneous, perirenal, and epididymal adipocytes, respectively). These mRNA changes were accompanied by parallel decreases in the amounts of ATG protein secreted by these cells: (−38, −68, and −65%, respectively; Fig.3 B). Moreover, testosterone treatment for 10 days fully restored the ATG mRNA decrease caused by castration in adipocytes whatever their anatomical origin (Fig. 3 A).
To verify that the effects of castration and androgen substitutive treatment on ATG are related to androgens, isolated perirenal adipocytes from castrated males were kept in primary culture for 24 h in the absence or presence of testosterone or dihydrotestosterone (DHT), a metabolite of testosterone that is not metabolized to estrogens. The androgen concentration tested was 100 nM, which corresponds to only 10 nM free hormone concentration because of the presence of albumin in the culture medium. Samples of the culture media were then removed for measurements of ATG secretion. Adipocyte total RNA was also extracted at the end of the 24-h incubation period, and ATG mRNA levels were measured by semiquantitative coRT-PCR. As shown in Fig. 4 A, a 24-h exposure to testosterone or DHT resulted in a significant increase in ATG mRNA levels in adipocytes from castrated rats (+58 ± 28 and +104 ± 38%, respectively) but not from SHAM rats (data not shown). However, in these adipocytes, exposure to androgens failed to modify the ATG protein secretion (data not shown).
Furthermore, when the same experiments were repeated in the presence of the androgen receptor antagonist cyproterone (1 μM), the ATG mRNA overexpression caused by testosterone or DHT could no longer be observed. It is important to note that cyproterone alone failed to affect ATG mRNA (Fig. 4 A).
Because a previous report showed that dexamethasone stimulates ATG mRNA expression in the Ob 1771 preadipocyte cell line and in rat epididymal adipose explants (1), we also tested the influence of 24-h dexamethasone exposure on ATG mRNA in perirenal adipocytes from castrated rats. Under these conditions, ATG mRNA levels were increased by 85%, an effect that was slightly but insignificantly amplified when dexamethasone was combined with testosterone (+91%; Fig.4 A). Moreover, addition of cyproterone had no significant influence on the ATG mRNA increase induced by the combination of dexamethasone plus testosterone (Fig. 4 A). Surprisingly, when these experiments were repeated using in vitro differentiated preadipocytes, ATG mRNA expression still remained increased (by a factor of 2) after exposure to androgens, but it was dramatically increased (+550%) in the presence of dexamethasone (Fig.4 B). Finally, when isolated perirenal adipocytes from normal or ovariectomized female rats were investigated under the same experimental conditions as above, here again, a 70% increase in ATG mRNA levels was observed after 24 h of exposure to testosterone (data not shown).
Various data from this laboratory have shown that adipocytes from subcutaneous fat deposits differ in many respects from those in deep abdominal deposits (29, 31). The present study provides an additional example of such site specificities by showing higher ATG expression and secretion in deep intra-abdominal than in subcutaneous adipocytes. Although we checked mRNA recovery during extraction, occurrence of tissue-specific differences in mRNA degradation or efficiency of mRNA extraction cannot be excluded. However, the recoveries of total RNA from the three studied regions were equivalent in efficiency, and the results obtained with the present semiquantitative coRT-PCR were concordant with those resulting from Northern blot analysis. Moreover, CYC mRNA expression was not affected by the adipose tissue anatomical locations and hormonal status, and RP mRNA expression was not affected by the adipose tissue anatomical locations. They were therefore chosen as “internal standards” for the present experiments.
So far, no evidence has been provided for a significant contribution of adipose tissue ATG to the circulating ATG pool that is involved in blood pressure regulation. In humans, however, obesity-associated hypertension is linked predominantly to deep intra-abdominal obesity (metabolic syndrome) (3) with a clear prevalence in males. Regional fat distribution differs between men and women; subcutaneous fat depots on hips and thighs are considered typically female, whereas excess fat in men is stored predominantly in the abdominal regions (16). In rats, overfeeding enhances expression and secretion of ATG from mature adipocytes and induces a rise in blood pressure (17). Furthermore, ATG gene expression in epididymal adipose tissue is increased in spontaneously hypertensive rats (SHR) (46), and in obese and diabetic mice (17). All these observations, together with the present finding of higher ATG mRNA levels in deep intra-abdominal than in subcutaneous fat, strongly suggest that ATG production from deep intra-abdominal adipocytes contributes to the association of obesity and hypertension (17). This is further supported by two recent observations, one showing the existence of a strong positive correlation between leptin and ATG plasma levels in healthy young men (41) and the other revealing a similar correlation between body mass index and plasma ATG in obese subjects (47).
From these observations, however, it cannot be excluded that secretion of ATG by adipose tissue, and especially by deep fat deposits, influences the RAS activity without affecting circulating ATG levels. In fact, another role that may be assigned to adipocyte ATG production could be local, because recent studies have reported the presence of all of the RAS components in adipose tissue (24, 41). Thus it seems likely that the local release of ANG II contributes to the modulation of adipose tissue blood flow and consequently to the regulation of adipose tissue metabolism. Moreover, ATG expression increases dramatically during the preadipocyte-adipocyte conversion of 3T3-L1 cells (39), and ANG II enhances the fat cell synthesis of PGI2, which is a powerful stimulator of the adipoconversion process (10). In rats, oral treatment with an ANG II receptor antagonist (losartan) reduces fat mass and adipocyte size independently of changes in food intake (9). Thus locally released ANG II appears to play a role in the regulation of adipose tissue development. Therefore, the site-related specificities in adipocyte ATG expression described herein could well account for the site-related differences observed in preadipocyte-adipocyte conversion capacities (31).
Characteristics of the adipoconversion process and fat cell metabolism differ not only according to the anatomical origin of the cells but also according to gender. A sexual dimorphism also exists in the pattern of high blood pressure development, where ANG II is particularly involved (6, 18). These observations led us to test the hypothesis that ATG expression could be regulated by androgens in rat adipose tissue. As revealed by the present study, adipocyte ATG expression and secretion are, in fact, controlled by androgens, because ATG expression decreased after castration and was fully restored by testosterone treatment whatever the adipocyte's anatomical origin. Moreover, as shown by our in vitro experiments, these effects seem to be mediated through the adipocyte androgen receptors, because cyproterone acetate, a potent antagonist of these receptors, prevented the positive in vitro influence of androgens on ATG mRNA expression. Induced transcription rate or increased transcript stability may account for this modulation.
Dexamethasone has been reported to potently stimulate ATG gene expression in explants of epididymal adipose tissue ex vivo from normal rats (1). In the present study, however, we could not observe such a potent effect on isolated adipocytes from castrated rats, as the mean stimulation reached was +85%. Moreover, no additive effects between testosterone and dexamethasone could be observed. In contrast, dexamethasone stimulation was clearly observed on in vitro differentiated rat preadipocytes in which the androgen effects were still present. In these in vitro experiments, we observed that the levels of ATG mRNA in control adipocytes declined by 50–60% between the experiment starting point and the end of incubation (data not shown). The finding that the increase in ATG mRNA expression caused by androgens in vitro never exceeded +100% does not allow a firm statement that the effects of androgens are solely related to transcriptional activation of the ATG gene expression or to an increase in the stability of ATG mRNA. Moreover, for still unknown reasons, we were unable to observe any parallelism between the in vitro effects of androgens on adipocyte ATG mRNA content and those on ATG secretion. Such discrepancies, which have also been reported for ATG modulation by fatty acids (37) or by insulin (2), could likely be related to an impairment of the translational and/or posttranslational process.
Despite a recent report showing an inverse correlation between serum ATG and androgen levels in boys (33), various experimental studies, mostly in rats, suggest an important regulatory role of androgens on both ATG expression and blood pressure control. Ganten et al. (18) have shown that castration prevents the gender-specific increase in blood pressure and that testosterone increases and androgen antagonists decrease blood pressure. Ellison et al. (14) demonstrated that kidney ATG mRNA is under the control of androgens in castrated rats. Renal and hepatic ATG mRNA levels have also been reported to be androgen dependent in SHR of both sexes (6). In contrast, Klett et al. (26) concluded that androgen-dependent ATG synthesis plays a minor role, if any, in the sexual dimorphism of blood pressure in SHR, because these authors failed to observe any alteration in ATG mRNA in liver, heart, kidney, or adrenal gland after testosterone stimulation. In that study, however, adipose tissue was not investigated. Most of these observations, taken together with our finding that androgens enhance adipose tissue ATG expression especially in deep intra-abdominal adipocytes, strongly suggest that ATG released from deep intra-abdominal fat deposits, which are particularly expended in android obesity, intervenes in the pathophysiology of the android obesity-related hypertension.
In conclusion, the present study provides clear evidence for an androgen upregulation of ATG expression in adipose tissue, especially in deep fat deposits. This finding, together with the present observation of higher production of ATG by intra-abdominal than by subcutaneous adipocytes, adds further weight to the hypothesis of a link between adipose tissue ATG production, androgens, and android obesity-related hypertension.
This work was supported by the Université RenéDescartes Paris V and the Ligue départementale des Yvelines contre le cancer.
Address for reprint requests and other correspondence: Y. Giudicelli, Service de biologie, Hôpital de Poissy, F78303 Poissy, France (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2000 the American Physiological Society