Androgen Effects on Adipose Tissue Architecture and Function in Nonhuman Primates

Abstract

The differential association of hypoandrogenism in men and hyperandrogenism in women with insulin resistance and obesity suggests that androgens may exert sex-specific effects on adipose and other tissues, although the underlying mechanisms remain poorly understood. Moreover, recent studies also suggest that rodents and humans may respond differently to androgen imbalance. To achieve better insight into clinically relevant sex-specific mechanisms of androgen action, we used nonhuman primates to investigate the direct effects of gonadectomy and hormone replacement on white adipose tissue. We also employed a novel ex vivo approach that provides a convenient framework for understanding of adipose tissue physiology under a controlled tissue culture environment. In vivo androgen deprivation of males did not result in overt obesity or insulin resistance but did induce the appearance of very small, multilocular white adipocytes. Testosterone replacement restored normal cell size and a unilocular phenotype and stimulated adipogenic gene transcription and improved insulin sensitivity of male adipose tissue. Ex vivo studies demonstrated sex-specific effects of androgens on adipocyte function. Female adipose tissue treated with androgens displayed elevated basal but reduced insulin-dependent fatty acid uptake. Androgen-stimulated basal uptake was greater in adipose tissue of ovariectomized females than in adipose tissue of intact females and ovariectomized females replaced with estrogen and progesterone in vivo. Collectively, these data demonstrate that androgens are essential for normal adipogenesis in males and can impair essential adipocyte functions in females, thus strengthening the experimental basis for sex-specific effects of androgens in adipose tissue.

Although altered levels of androgens are associated with various metabolic abnormalities in humans, recent studies also suggest that androgens exert opposite effects on metabolically active tissues in men and women (1, 2). Androgen deficiency in men is linked to insulin resistance and obesity, and treatment of hypogonadal men with testosterone improves insulin sensitivity and reduces fat content (1–7). In contrast, the androgen excess that occurs in women with polycystic ovary syndrome (PCOS) correlates with insulin resistance and obesity (1, 8). Sex-specific effects of androgens in white adipose tissue (WAT) and other metabolically active tissues may also explain differences in body fat distribution (9) and insulin sensitivity in male and females (10–14).

The role of androgens in regulation of body fat in males has been explored using various castration and transgenic rodent models, producing conflicting results. Castration in mice does not result in obesity (14), and as a recent report showed, mice can lose fat mass in response to gonadectomy (15). Furthermore, WAT-specific androgen receptor knockout mice display no changes in body weight and adiposity (16). In rats, castration reduces adiposity (17), although an increase in adiposity has also been reported (18). Although human studies consistently demonstrate a positive correlation between hypogonadism, increased body fat, and insulin resistance (1–7, 19, 20), it is currently unknown whether the observed metabolic changes are direct effects of androgen deficiency or are secondary effects of aging or changes in lifestyle. Moreover, species-specific differences in the expression of genes involved in essential metabolic functions may also explain different effects of androgen deficiency on adiposity in rodents and humans. For example, clathrin heavy-chain, isoform 22, an integral component of the glucose transporter 4 (Glut4)-mediated glucose uptake system in WAT and muscle (21); the agouti protein involved in the development of insulin sensitivity in WAT (22); acetyl-coenzyme A carboxylase-2, the enzyme required for de novo lipogenesis (23); and natriuretic peptide-dependent lipolysis (24) are all essential for metabolic function in humans but not in rodents.

There are, therefore, significant conceptual gaps in our understanding of how sex regulates androgen action in human WAT. Achieving this insight will rely on the ability to develop new animal models using human-related species as well as new WAT culture systems that can provide a consistent framework for understanding physiologically relevant processes that take place in vivo. To circumvent possible differences in androgen response between rodents and primates, and to rule out the influences of secondary confounding factors, such as diet and lifestyle, we took advantage of the availability of nonhuman primate (NHP) models representing altered steroid hormone status obtained from unrelated studies, thus maximizing the utility of these valuable research animals. We also employed a novel ex vivo methodology that enables study of the essential properties and regulation of primary WAT under controlled culture conditions. The studies reported here demonstrate specific sex-dependent actions of androgens on adipocyte biology that are of potential relevance to androgen-imbalance syndromes in humans.

Materials and Methods

Animal procedures

WAT samples used in this study were obtained as excess tissue from animals studied under protocols previously approved by the Institutional Animal Care and Use Committee of the Oregon National Primate Research Center. Animals were fed ad libitum standard monkey chow that provides 13% of calories from fat, 69% of calories from carbohydrates, and 18% of calories from proteins and contained sufficient vitamins and minerals for normal growth. Animal groups included ovary-intact adult female rhesus macaques (n = 4), ovariectomized adult female rhesus macaques with placebo (n = 4) or estrogen and progesterone replacement using SILASTIC (Dow Corning, Midland, MI) capsules (25) containing crystalline estradiol for 14 d and then estradiol and progesterone for 7–14 d (n = 4), juvenile female rhesus macaques (n = 4), and castrated (after puberty) adult male Japanese macaques that were treated with placebo (n = 4) or testosterone implants (n = 5). Bilateral ovariectomy, bilateral orchiectomy, and implantation of SILASTIC capsules containing steroid hormones were performed by expert surgical personnel at Oregon National Primate Research Center according to well-accepted veterinary surgical procedures under sterile conditions and appropriate anesthesia with postoperative pain control. SILASTIC capsules were implanted in the periscapular region. Animals were euthanized according to procedures recommended by the Panel on Euthanasia of the American Veterinary Association. The night before necropsy, food was withheld. Before necropsy, animals were sedated with ketamine in the home cage, transported to the necropsy suite, treated with pentobarbital (25 mg/kg), and exsanguinated by severance of the descending aorta. Animals used in the present study were adults (6–12 yr; mean age ± sd = 9.0 ± 2.6 yr; representing the typical reproductive age in macaques) or juveniles (mean age ± sd = 1.0 ± 0.2 yr). The intact adult female animals used in this study displayed regular breeding behavior and regular menstrual cycles. All adult females had experienced at least one recorded pregnancy.

Serum hormone assays

Assays for estradiol, progesterone, and testosterone were performed using a Roche Diagnostics (Indianapolis, IN) Cobas-C411 assay instrument. The assay sensitivity ranges were 5–4250 pg/ml for estradiol, 0.035–59 ng/ml for progesterone, and 0.025–15 ng/ml for testosterone. Dihydrotestosterone (DHT) was measured by an extraction-chromatography RIA. Briefly, samples diluted in ethanol were dried under forced air and extracted with 300 μl 0.1% gel-PBS and 5 ml diethyl ether, dried under forced air, and redissolved in 200 ml column solvent (hexane/benzene/methanol 62:20:13). The sample was then added to a 1 × 6-cm, all-glass column containing 1 g Sephadex LH-20 for separation of different steroids. Each fraction was then dried under a forced-air stream before immunoassay. Hormone values were corrected for extraction-chromatography losses determined by radioactive tracer recovery at the same time with sample extraction. The DHT assay sensitivity range was 5–750 pg/tube. Interassay and intraassay coefficients of variation, based on quality controls provided by Roche did not exceed 10%. The overall inter- and intraassay variation for DHT extraction-chromatography RIA did not exceed 15%. Serum insulin and glucose levels and homeostasis model assessment for insulin resistance values were determined at fasting as described (26).

Ex vivo hormone treatments of WAT explants

At necropsy, approximately 2 g retroperitoneal WAT were dissected and collected in 50-ml tubes filled with 20 ml M199 medium (Invitrogen, Grand Island, NY) (pH 7.4) at room temperature. We had previously determined that cold media inhibited subsequent insulin responsiveness of WAT (data not shown), possibly due to a lipid-phase transition. Because WAT can acidify incubation media, the presence of phenol red was essential for monitoring pH. Hormone treatment of WAT was started within 30 min of necropsy. WAT was dissected nonsterilely on a bench, using small surgical scissors. To prevent spontaneous adipocyte lysis, it is essential to avoid contact with glass and to minimize exposure to air. For Western blotting studies, 100 ± 10-mg WAT explants were placed into a 12-well culture dish containing 2 ml incubation medium [M199 medium, 0.1% fatty acid-free BSA (Sigma-Aldrich, St. Louis, MO), and 20 mm HEPES (pH7.4) supplemented with penicillin, streptomycin, and fungizone] and incubated at 37 C free floating for 2 h in an atmosphere of 5% CO₂ at 37 C in the presence of the indicated concentrations of human insulin (Sigma-Aldrich). For imaging studies, smaller, approximately 2-mm WAT explants were incubated free floating in plastic eight-well chambers (Lab-Tek II chambered no. 1.5, German coverglass system; Nalge Nunc International, Rochester, NY) filled with 0.4 ml incubation medium. WAT explants were either incubated with insulin, as described above, or first incubated with androgens before insulin treatment. DHT was prepared and stored at −20 C as a 1 mm master stock solution in ethanol. Before each experiment, the master stock solution was diluted in ethanol to 10 μm and then further diluted to incubation medium to a final concentration of 10 nm. Where indicated, 200 nm of the androgen receptor antagonist bicalutamide was added to the well (prepared from a 200 μm stock solution in dimethylsulfoxide). Medium was replaced every 24 h.

Confocal microscopy

Minor modifications were made to our previously described protocol (26). BODIPY-500/510 C₁, C₁₂ (BODIPY-C12; Invitrogen) was prepared in advance by diluting a 2.5 mm methanol stock solution in incubation medium to a final concentration of 10 μm. To allow BODIPY-C12 to bind BSA, the 10 μm BODIPY-C12 solution was incubated for 15 min, protected from light, in a 37 C water bath. During this incubation step, 2 μl ethidium homodimer (LIVE/DEAD Viability/Cytotoxicity Kit; Invitrogen) was mixed into each well containing WAT explants and incubation continued for 15 min. This approach identifies dead cells (adipocytes and nonadipocytes) that exhibit red fluorescent staining of nuclei. Eight-well chambers containing WAT explants free floating in 0.4 ml incubation medium (with ethidium homodimer and hormones) were removed from the incubator, and 100 μl 10 μm BODIPY-C12 solution was added to each well, mixed by repeated pipetting, and the chambers were incubated for an additional 10 min at 37 C. Labeling reactions were placed on ice, and WAT explants were brought to the bottom of the chamber with 8 × 8-mm squares of light stainless steel mesh (0.4 mm; TWP, Inc., Berkeley, CA). Medium was removed by aspiration, and tissue was washed three times with ice-cold incubation medium. WAT explants were fixed at room temperature with 4% paraformaldehyde in PBS for 20 min, washed four times with PBS, and stored in PBS at 4 C, protected from light, for up to 48 h before confocal analysis. Image recording and cell size determination were conducted using a Leica SP5 AOBS spectral confocal system as described (26).

Western immunoblotting

After ex vivo incubation with hormones, 100-mg WAT explants were removed from the dish with tweezers, rapidly placed on top of the clean 0.4-mm stainless steel mesh sheet, and washed three times with ice-cold PBS. Excess liquid was removed by repeated blotting of the bottom of the sheet with a paper towel, WAT explants were transferred to 1.5-ml microcentrifuge tubes, snap-frozen on dry ice, and stored at −80 C for further analysis. Tubes containing frozen WAT explants were placed on ice, and 200 μl ice-cold sodium dodecyl sulfate (SDS)-lysis buffer [50 mm Tris-Cl (pH 6.8), 1% SDS, 10% glycerol, complete cocktail of protease inhibitors (Roche Diagnostics), and 2 mm sodium vanadate] were added to the tubes. Tissue was homogenized using a hand-held pestle motor homogenizer, boiled for 5 min, and centrifuged for 10 min to separate lipids (top layer) and cell membranes (pellet). Liquid interface (170 μl) was transferred to fresh 1.5-ml tubes, supplemented with 0.1% bromophenol blue and 100 mm dithiothreitol, and boiled for 4 min, and 20–40 μg proteins was loaded onto precast 10% SDS-PAGE gels (Bio-Rad, Hercules, CA). Protein concentrations were determined before bromophenol blue and dithiothreitol were added to the tubes, using a detergent-compatible protein assay kit (Bio-Rad). Western blotting was performed using primary antibodies to Akt (Cell Signaling Technologies, Danvers, MA), Ser473-Akt (Invitrogen), and androgen receptor (Santa Cruz Biotechnology, Santa Cruz, CA) at a 1:500 dilution and antibodies to actin (Santa Cruz Biotechnology) at a 1:5000 dilution. Secondary horseradish peroxidase-conjugated antibodies were used at a 1:5000 dilution. For quantification of in vivo androgen receptor levels, WAT tissue obtained at necropsy was snap-frozen in liquid nitrogen and processed as described above.

RNA isolation and quantitative RT-PCR (qRT-PCR) analysis of gene expression

RNA was prepared from 100 mg whole WAT using an RNeasy lipid kit (QIAGEN, Inc., Germantown, MD). Equal amounts of input RNA were used for all PCR. RNA was reverse transcribed using random-hexamer primers and treated with deoxyribonuclease. qRT-PCR was performed using SYBR Green-I reagent, transcript-specific primers, and actin primers as an endogenous control (Table 1). The 2^−ΔΔCT relative quantification method was used to calculate fold difference in transcript levels between samples, and amplification efficiencies for all primer pairs were verified to be equivalent over a range of template concentrations. qRT-PCR was performed using an ABI7900 thermocycler (Applied Biosystems, Inc., Foster City, CA).

Table 1.

RT-PCR primers used in the present study

Gene	Sequence of 5′ primer	Sequence of 3′ primer	Accession no.	5′ primer	3′ primer
Actin	CGGCATCGTCACCAACTG	GGCACACGCAGCTCATTG	M10277	18mer	18mer
ATGL	GTGTCAGACGGCGAGAATG	TGGAGGGAGGGAGGGATG	AY894804	19mer	18mer
C/EBPα	GTGGAGACGCAGCAGAAG	TTCCAAGGCACAAGGTTATC	NM_004364	18mer	20mer
C/EBPβ	CTCGCAGGTCAAGAGCAAG	GCAGTGGCCGGAGGAGGCGAGC	NM_005194	19mer	22mer
C/EBPδ	AGCGCAACAACATCGCCGTG	GTCGGGTCTGAGGTATGGGTC	NM_005195	20mer	21mer
CIDEA	GGCAGGTTCACGTGTGGATA	GAAACACAGTGTTTGGCTCAAGA	NM_001279	20mer	23mer
FASN	CTGGCTCAGCACCTCTATCC	CAGGTTGTCCCTGTGATCCT	NM_004104	20mer	20mer
GLUT1	CCATACTCATGACCATCGCGCTAG	CAAAGAAGGCCACAAAGCCAAAG	NM_006516	24mer	23mer
GLUT4	CTTCGAGACAGCAGGGGTAG	AGGAGCAGAGCCACAGTCAT	M91463	20mer	20mer
HSL	GAGCGGATCACACAGAACCT	CCAGAGACGATAGCACTTCCA	NM_005357	20mer	21mer
PGC1a	CTGTGTCACCACCCAAATCCTTAT	TGTGTCGAGAAAAGGACCTTGA	NM_013261	24mer	22mer
PPARδ	CTGCAGATGGGCTGTGACGG	GTCTCGATGTCGTGGATCAC	NM_006238	20mer	20mer
PPARγ	AGCCTCATGAAGAGCCTTCCA	TCCGGAAGAAACCCTTGCA	NM_005037	21mer	19mer
UCP1	GGTGTCCTGGGAACAATCAC	GAGAGGCGGAGCTGATTTGCC	NM_021833	20mer	21mer
Leptin	GTGCGGATTCTTGTGGCTTT	GGAATGAAGTCCAAACCGGTG	NM_000230	20mer	20mer
SREBP1C	GGAGCCATGGATTGCACTTT	TCAAATAGGCCAGGGAAGTCA	NM_001005291	20mer	21mer
PRDM16	GAAACTTTATTGCAATAGTGAGATGA	CCGTCCACGATCTGCATGT	NM_022114	26mer	19mer

Data analysis

Differences between groups were determined by Student's t test and by two-way ANOVA followed by Bonferroni post hoc pairwise comparisons using Prism version 4 (GraphPad Software, Inc., San Diego, CA). Intracellular fluorescence associated with BODIPY-C12 was determined as previously described (26).

Results

The effects of androgen depletion on body weight and insulin sensitivity in males

To mimic a hypogonadal state in males, we used adult macaques that were castrated for 6 months and then treated for an additional 4 months either with placebo [castrated males (CAS) group, Fig. 1A] or physiological concentrations of testosterone [testosterone-treated CAS (TEST) group, Fig. 1A]. Testosterone replacement restored normal androgen levels typically observed in male rhesus macaques and related primate species (27–30) (Fig. 1, B and C). Serum insulin, glucose, and triglyceride levels and insulin sensitivity were normal and not statistically different between the two groups (Fig. 1, D–F, and data not shown). CAS displayed a slight reduction in body weight, presumably due to muscle loss, although no significant differences in weight were seen between the two groups (Fig. 1G). This finding is consistent with previous reports that gonadectomy results primarily in a loss of lean mass but has no significant effect on fat mass in male NHP (31, 32). Thus, androgen depletion did not induce overt obesity or insulin resistance in adult male macaques.

Fig. 1.

Testosterone deficiency does not cause systemic insulin resistance in NHP males. A, Androgen depletion and replacement strategies in macaques. Adult males were castrated and 6 months later treated for an additional 5 months with either placebo (CAS) or testosterone (TEST). B–E, Serum levels of testosterone (B), DHT (C), fasting insulin (D), and fasting glucose (E) were determined at the day of necropsy. The upper and lower limits of serum testosterone and DHT observed in intact male macaques and related species (27–30) are indicated by dotted lines. F, Homeostasis model assessment for insulin resistance was determined as described (26). G, Body weight changes in response to castration and testosterone replacement. Animals were weighed before castration (initial weight) and 11 months after castration. The graphs are means ± sem; n = 5 (CAS), and n = 4 (TEST). *, P < 0.01; **, P < 0.05, t test.

In vivo effects of androgen depletion on adipocyte morphology in males

Confocal microscopy studies revealed that castration had a striking effect on the morphological appearance of WAT. WAT of CAS (CAS-WAT) consisted of significantly smaller adipocytes than WAT of TEST (TEST-WAT; Fig. 2A) and was also enriched in abnormally small adipocytes less than 40 μm in diameter (Fig. 2B). A subpopulation of adipocytes from CAS also had an abnormal multilocular appearance and fragmented lipid droplets (Fig. 3A, upper panels, and B, arrowheads). In contrast to adipocytes of CAS, adipocytes of TEST were morphologically homogeneous, larger in size, and had a typical unilocular appearance (Figs. 2 and 3A, lower panels, arrowheads). Multilocular adipocytes were not observed in TEST-WAT. Thus, testosterone replacement appeared to eliminate a subpopulation of multilocular adipocytes, stimulating cell enlargement and a unilocular phenotype.

Fig. 2.

Testosterone deficiency induces smaller adipocyte size in male WAT. A, Average adipocyte diameters in WAT of CAS and TEST. Adipocytes were stained with BODIPY-C12, and their diameters were determined as previously described (26). Typically, 100 adipocytes were averaged to calculate mean adipocyte diameter in individual animals. B, Distribution of adipocyte diameters in WAT of CAS and TEST. The graphs are means ± sem; n = 5 (CAS), and n = 4 (TEST). *, P < 0.05, t test.

Fig. 3.

Testosterone deficiency induces a multilocular phenotype in male WAT. Insulin-stimulated WAT explants from CAS (A, upper panels, and B, both panels) and TEST (A, lower panels) were labeled for 10 min with green fluorescent BODIPY-C12 (lipid droplets) and red fluorescent wheat germ agglutinin (blood vessels) and analyzed by confocal microscopy as described in Materials and Methods. Arrowheads indicate small multilocular adipocytes in CAS (upper panels) and small unilocular adipocytes in TEST (lower panels). Scale bar, 50 μm.

Androgen depletion and expression of BAT markers in male WAT

One potential explanation for the appearance of multilocular adipocytes is androgen depletion-induced trans-differentiation of white (unilocular) adipocytes into brown (multilocular) adipocytes. To evaluate this possibility, we compared the expression of brown adipose tissue (BAT)-specific genes in CAS-WAT and TEST-WAT. The expression levels of UCP1 and PRDM16 (33) were below the detection limit, and the expression of CIDEA, another gene that is up-regulated in BAT (34), was not statistically different between the two groups (Fig. 4A). We conclude that the mutilocular phenotype induced by androgen depletion does not result from the trans-differentiation of WAT to BAT.

Fig. 4.

Androgens potentiate adipogenic gene expression and insulin sensitivity in male WAT. A, Comparative gene expression in retroperitoneal WAT of CAS and TEST; the graph shows the fold change in mRNA levels in the TEST group in comparison with the CAS group (see Materials and Methods for details). The graph is mean ± sem; n = 5 (CAS), and n = 4 (TEST). *, P < 0.01, t test. ND, Nondetected. B, WAT explants from CAS and TEST were incubated with 0–10 nm insulin for 2 h, tissue was lysed in the presence of phosphatase inhibitors, and protein extracts were analyzed by Western blotting using antibodies to Akt and phospho-Ser473-Akt. The levels of Akt phosphorylation were normalized to total Akt levels. Graphs are means ± sem; n = 3 (CAS), and n = 3 (TEST). *, P < 0.05, t test.

Effects of testosterone on WAT gene expression and adipocyte insulin sensitivity in males

Given that CAS-WAT showed no enrichment in BAT markers, we considered the possibility that androgens may coordinate genes that regulate adipogenesis in males. Consistent with this scenario, the expression level of the genes encoding the transcription factor C/EBP-α, but not C/EBP-β or C/EBP-δ, was more than 2-fold higher, and the level of leptin gene expression was more than 5-fold higher in the TEST group than in the CAS group (Fig. 4A). Serum leptin levels, however, were similar in CAS and TEST groups (1.1 ± 0.15 and 0.9 ± 0.12 ng/ml, respectively), indicating a possible posttranscriptional control of leptin production or the involvement of other fat depots in androgen-independent leptin production. The expression levels of other adipogenic genes, including fatty acid synthase, adipose triglyceride lipase, Glut 4, and sterol response element binding protein 1C were also higher in TEST-WAT, although statistical significance between the two groups was not achieved (Fig. 4A). The later observation may reflect the heterogeneity of WAT with respect to its differentiation state. The levels of the mRNA encoding the inflammatory cytokines TNF-α and IL-6 were similar in the two groups of males (data not shown), which suggests that there was not extensive macrophage infiltration as a result of androgen deficiency. Given that testosterone potentiates adipogenic gene expression and a normal morphological phenotype of WAT, it may also drive the development of the efficient response of WAT to insulin. To examine this, CAS-WAT and TEST-WAT explants were pretreated with various doses of insulin ex vivo, and the level of Akt phosphorylation at Ser473 in WAT extracts was quantified by Western blotting. As expected, WAT-TEST exhibited greater insulin-stimulated Akt phosphorylation than CAS-WAT (Fig. 4B), although this difference was statistically significant only at the highest insulin concentration tested. These data suggest that androgens are likely to drive the adipogenic transcriptional program required for the development of a mature unilocular, insulin-sensitive phenotype in male WAT.

Sex-specific effects of androgens on functional properties of WAT ex vivo

The well-known association of PCOS and hyperandrogenism with insulin resistance and obesity (1, 8, 35) suggests that androgens may exert opposite effects on WAT physiology in females and males. Because female WAT is normally exposed to cyclical fluctuations of estradiol and progesterone, cross talk between androgens and ovarian hormones may influence adipose function. To examine this, we established organotypic cultures (explants) of WAT from several experimental groups of animals, including intact adult females (INT), intact juvenile females (JUV), ovariectomized adult females (OVX), OVX that were replaced with estradiol and progesterone in vivo (EP), and androgen-depleted CAS. WAT explants were incubated ex vivo with 10 nm DHT or vehicle control for 48 h. Under these experimental conditions, DHT has been shown to inhibit glucose uptake and insulin signaling in in vitro-differentiated female adipocytes (36). Although limited WAT availability did not allow the direct analysis of glucose uptake and insulin signaling, use of the fluorescent nonesterified free fatty acid (FA) tracer BODIPY-C12 allowed us to quantify insulin response based on the fluorescence emitted from single adipocytes (26). This method was originally used to demonstrate that the translocation of Glut-4 and the FA transporter-1 to the cell surface are insulin dependent and correlated with increased BODIPY-C12 uptake in 3T3-L1 and primary mouse adipocytes (37). More recently, it has been demonstrated that FA transporter-1 is the insulin-dependent energy sensor in adipocytes (38).

Figure 5A depicts basal and insulin-dependent FA uptake in WAT explants from different experimental groups of animals that were incubated under control (without DHT) conditions ex vivo. Among the five experimental groups, CAS-WAT demonstrated the lowest insulin response, and OVX-WAT had the highest insulin response. INT-WAT and JUV-WAT displayed a lower insulin response than OVX-WAT. Although the levels of serum estradiol and progesterone were significantly higher in the EP group than in the INT and OVX groups (Fig. 5, C and D), insulin-dependent FA uptake in EP-WAT was statistically undistinguishable from that in INT-WAT and OVX-WAT (Fig. 5A).

Fig. 5.

Sex- and ovary-specific effects of androgens in WAT. Retroperitoneal WAT explants from CAS (n = 4), INT (n = 4), OVX (n = 4), EP (n = 4), and JUV (n = 4) were preincubated for 48 h with alcohol vehicle (A) or 10 nm DHT (B) and then treated for 2 h with buffer alone (basal) or 10 nm insulin. BODIPY-C12 uptake and data analysis were performed as described in Materials and Methods and (26). A, Insulin-dependent FA uptake values were normalized to basal FA uptake values. B, Basal and insulin-dependent FA uptake values were normalized to control values shown in (A). Graphs are means ±sem, n = 4 per group. Data were analyzed by two-way ANOVA followed by Bonferroni posttests; *, P < 0.05; **, P < 0.01. C and D, Serum levels of estradiol (C) and progesterone (D) in INT, EP, and OVX were determined at necropsy. The graphs show means ± sem; n = 4. *, P < 0.05, t test. F, Females; M, males.

To examine androgen effects on adipocytes ex vivo, we exposed WAT explants to DHT and then assayed basal and insulin-dependent FA uptake. This study revealed a pattern that clearly correlated with the presence or absence of insulin, sex, and previous ovarian status. Although insulin-dependent FA uptake in (male) CAS-WAT was not significantly affected by DHT, female WAT from all experimental groups showed a decrease in FA uptake (Fig. 5B, black bars), confirming the previously reported inhibitory effect of androgens on insulin action in female adipocytes (36). Under basal conditions, however, DHT-induced FA uptake and the extent of this induction depended on sex and ovarian status. Although CAS-WAT and INT-WAT showed only a slight increase in FA uptake, OVX-WAT displayed an almost 3-fold increase in basal FA uptake in response to DHT treatment (Fig. 5B, gray bars). JUV, whose estrogen and progesterone levels were similar to that of adult OVX females (data not shown), also exhibited elevated basal FA uptake, whereas in vivo estrogen and progesterone replacement reduced basal FA uptake (Fig. 5B). These data are consistent with previous studies that female sex hormones can reduce lipid storage and visceral adiposity in females (reviewed in Ref. 39).

The role of the androgen receptor in mediating androgen effects in WAT

The classical androgen receptor is expressed in male and female WAT of humans and rodents (40, 41). We determined whether the effect of DHT on FA uptake in female WAT was mediated by the classical androgen receptor, using the androgen receptor antagonist bicalutamide (BIC). BIC pretreatment of WAT explants reversed the effects of DHT on both basal and insulin-stimulated FA uptake (Fig. 6A), whereas BIC alone had no effect of FA uptake (data not shown). These results support a classical androgen receptor-mediated genomic response as the basis for the observed effects of DHT on FA uptake in female adipocytes and suggest that locally synthesized androgens (42, 43), at least in lean females, do not significantly contribute to FA uptake in WAT. As shown in Fig. 6B, the androgen receptor is expressed in female NHP WAT, and the apparent level was increased by ex vivo DHT treatment, presumably representing androgen-induced stabilization of the receptor. We also verified that androgen receptors are expressed in both male and female NHP WAT and determined the effects of castration and testosterone replacement in males and ovariectomy in females on receptor levels. In females, ovariectomy had no significant effect on androgen receptor expression in WAT (Fig. 6C). and in males, the androgen receptor was expressed at similar levels in CAS-WAT and TEST-WAT (Fig. 6D). Thus, it appears that androgen and estrogen deficiency in males and females, respectively, are not associated with altered androgen receptor levels in WAT, whereas androgen excess in females may increase expression of androgen receptors in WAT. The later ex vivo-based observation should to be interpreted with caution, because it needs to be further verified in vivo. Collectively, these findings suggest that the classical androgen receptor is the essential mediator of androgen action in WAT.

Fig. 6.

Androgens exert their action via the androgen receptor expressed in WAT. A, Reversal of DHT effects on FA uptake in females with an androgen receptor antagonist. WAT explants from an adult OVX were incubated for 48 h with buffer or 10 nm DHT alone or together with 200 nm BIC and then treated for 2 h with 10 nm insulin or medium alone (basal). Values are means of intracellular fluorescence ± sem; n = 30 adipocytes (see Ref. 26 for experimental details). **, P < 0.01; *, P < 0.05, t test. BIC treatment experiments were repeated with two independent animals and yielded similar results. B, Apparent androgen receptor (AR) levels in female WAT can be increased by ex vivo DHT treatment. Retroperitoneal WAT explants were incubated for 48 h with or without DHT and lysed, and androgen receptor levels were assessed by Western blot. The experiment was repeated twice. Androgen receptor levels normalized to actin were determined by Western blot using freshly isolated retroperitoneal WAT from INT and OVX (C) and from TEST and CAS (D). The graphs are means ± sem; n = 6 (INT), n = 5 (OVX), n = 4 (TEST), and n = 3 (CAS). F.U., Fluorescent units.

Discussion

The NHP model employed in the present studies furthers our understanding of the sex-specific role of androgens in controlling WAT function. Moreover, the utility of organotypic cultures of WAT provides a novel ex vivo approach to study WAT in a physiologically relevant setting. Because a significant portion of the U.S. population suffers from various forms of metabolic complications related to androgen imbalance, including hypoandrogenism men (1–7) and hyperandrogenism in women (1, 8), the present study also sheds light on the mechanisms of obesity associated with androgen imbalance in humans.

Effects of hypoandrogenism on WAT in males

A principal finding emerging from the present study, illustrated in Fig. 7, is that androgen deprivation induced the appearance of small multilocular adipocytes and that testosterone replacement restored normal adipocyte size and a unilocular phenotype in male WAT. Although overt systemic insulin resistance was not observed in CAS, androgen-deficient WAT exhibited clear evidence of altered morphological organization, impaired insulin response, and reduction in expression of essential genes involved in energy use and adipogenesis. Testosterone replacement of CAS restored a normal unilocular WAT phenotype and stimulated the expression of C/EBP-α, a key transcription factor required for the terminal step of adipogenesis, consistent with previous studies that adipocytes derived from C/EBP-α-deficient mice are more insulin resistant and accumulate less lipids than adipocytes from control mice (44). Furthermore, the multilocular phenotype and reduced adipocyte size observed in CAS-WAT agree with previous reports that the suppression of C/EBP-α in vivo leads to lipodystrophy (45, 46) and multilocular WAT in mice (47). C/EBP-α has been shown to activate leptin gene expression in WAT, and C/EBP-α knockout mice exhibited decreased leptin levels and WAT atrophy (46, 48–50). Lipodystrophy of WAT can also occur in HIV patients treated with protease inhibitors, resulting in reduced levels of the transcriptional factor SREBP1c (51) and in transgenic mice that overexpress a dominant-negative form of SREBP1c (52). The latter also displayed altered WAT differentiation, including the simultaneous presence of small and large adipocytes. These data suggest that testosterone is an essential regulator of adipogenic transcriptional factors required for the development of mature WAT in males.

Fig. 7.

Regulation of adipogenesis by androgens in male WAT. Initiation of the adipogenic program in preadipocytes requires activation of the transcriptional factors peroxisome proliferator-activated receptor-γ (not shown), C/EBPβ, and C/EBP-δ. PPAR-γ is further activated by transcription factors C/EBP-α and SREBP1c. The appearance of C/EBP-α stimulates the development of insulin (INS) sensitivity and promotes further up-regulation of lipogenic genes in adipocytes (44). Androgens can inhibit the recruitment of mature adipocytes from preadipocytes in vitro through an androgen receptor-mediated nuclear translocation of β-catenin and activation of downstream Wnt signaling (53, 55). We propose that androgens may potentiate C/EBP-α and SREBP1c gene expression, stimulating the development of the mature adipogenic phenotype and insulin sensitivity. It is also possible that androgen deficiency promotes the transition of mature adipocytes to a less-differentiated state.

Another hallmark of the present study is the finding that CAS-WAT consisted of abnormally small adipocytes that displayed reduced insulin sensitivity. We have previously demonstrated that adipocytes with cell diameters over 100 μm are insulin resistant compared with smaller adipocytes (26). The adipocytes in androgen-deficient males were, however, even smaller than those in androgen-replaced animals (Fig. 2) or insulin-sensitive adipocytes in intact animals (26). Although we do not currently have an obvious mechanism that would explain the very small adipocytes present in the androgen-deficient animals, the current data, in conjunction with our previous studies, suggest that there may be an optimal size for insulin sensitivity. An alternative explanation is that the decreased insulin response is associated with the concurrent multilocular phenotype rather that the small size per se. Assessing these alternatives will require a more detailed analysis of individual small unilocular vs. multilocular cells.

Androgen deficiency in the males in this study was not associated with systemic insulin resistance. Because these animals were fed monkey chow, which is much lower in saturated fats than the typical Western-style diet, we propose that a Western-style, high-fat diet may be an essential factor that can accelerate the development of obesity and insulin resistance in androgen-depleted males. Although androgens can inhibit adipocyte differentiation in vitro (53–55), whether androgen deficiency can increase de novo adipocyte differentiation in vivo is unknown. The multilocular adipocytes we observed in CAS-WAT may represent newly differentiated adipocytes, or may result from the dedifferentiation of mature adipocytes (Fig. 7). Our data do not allow us to distinguish between these two possibilities.

Sex-specific effects of androgens on insulin response in WAT

We exploited our previously described organotypic culture of NHP WAT (26) to assess sex differences in androgen responsiveness of WAT ex vivo. The reduced insulin-stimulated FA uptake observed in DHT-treated female WAT is consistent with previous studies showing that androgens inhibit glucose uptake and insulin sensitivity of in vitro-differentiated human female adipocytes (36). Although ex vivo DHT treatment did not improve insulin sensitivity of male CAS-WAT, in vivo testosterone replacement of androgen-deficient males did potentiate insulin-induced Akt phosphorylation in male WAT (Fig. 4B), suggesting that androgens regulate male adipose insulin sensitivity. Taken together, the present findings also suggest the opposite regulation of insulin sensitivity by androgens in WAT of males and females, although underlying mechanisms that control sex-specific responses to insulin remain to be identified.

Effects of ovarian hormones on androgen response in WAT

Previous ovarian status played a significant role in determining androgen sensitivity of female WAT with respect to basal FA uptake. That androgens significantly increased basal FA uptake in OVX-WAT, in comparison with INT-WAT, suggests that ovarian hormones can decrease lipid uptake in WAT and may reduce the lipotoxic effects of FA in adipocytes. The elevated basal FA uptake seen in the presence of DHT (present study) and the impaired lipolysis seen in hyperandrogenic females (54, 56–58) may contribute to WAT hypertrophy in women with PCOS. This finding is consistent with in vivo studies that the levels of estrogens inversely correlate with visceral fat mass (39). The loss of the protective role of ovarian hormones may also explain menopausal weight gain (59). The analysis of protein expression showed no signs of up-regulation of FA transporters (Supplemental Fig. 1, published on The Endocrine Society's Journals Online web site at http://endo.endojournals.org) or significant changes in phosphorylation of Akt (data not shown) in androgen-treated female WAT, suggesting the involvement of a novel sex-specific, possibly estrogen-dependent, mechanism that controls FA uptake in female adipocytes. Hyperandrogenism is a risk factor for the development of obesity and metabolic syndrome in women with PCOS, although the molecular mechanisms are largely unknown (1, 8, 35). Considering the lack of efficient treatments for PCOS, estrogen replacement alone or in combination with antiandrogens may help to reduce or eliminate metabolic abnormalities associated with this disease, which affects up to 10% of women of reproductive age.

Study limitations

Our study does have certain limitations. The male animals were Japanese macaques, whereas the female animals were Rhesus macaques. This was dictated by the nature of the animals from which appropriate tissues were available, but these species are very closely related, and there is no evidence that basic physiology differs between them. Similarly, although we were able to evaluate both previous in vivo manipulation and acute effects of androgens on male WAT, we were able to assess only acute effects of androgen on female WAT, although we were able to compare the effects of varying ovarian hormone status. Neither of these aspects, however, is expected to materially affect our conclusions on the sex-specific effects of androgens on WAT. Finally, because the ex vivo cultures of female WAT were not supplemented with female sex hormones, we could assess only the effects of previous steroidal status on measured parameters. Although this ex vivo approach does not completely reconstitute the in vivo environment, it nevertheless provides a foundation for future studies of WAT physiology using more physiological ex vivo environments.