Department of Obstetrics and Gynecology, Wisconsin Primate Research Center, University of Wisconsin-Madison, 1223 Capitol Court, Madison, WI 53715. E-mail: email@example.com
Objective: The aim of this study was to investigate the effects of early-gestational androgen excess on adult body fat distribution in female rhesus monkeys.
Research Methods and Procedures: Six midreproductive-aged, adult female rhesus monkeys that were exposed to androgen excess started during the first one-third of gestation were pair-matched to control females by age, body weight, and body mass index. Body composition was determined using somatometrics, DXA, and computed tomography.
Results: Total abdominal and intra-abdominal fat depots are increased in adult female rhesus monkeys exposed to prenatal androgen excess.
Discussion: Early gestational androgen excess in female rhesus monkeys causes a preferential accumulation of total abdominal and intra-abdominal fat during adulthood. Fat accumulation in these regions is independent of total body adiposity, occurring throughout the spectrum of body mass index in these animals. This study establishes alterations in abdominal adiposity as another consequence of prenatal androgen excess in female rhesus monkeys that may contribute to the impaired insulin secretion observed in these animals during adulthood.
Pre- or perinatal androgen excess in nonprimate mammals causes species-dependent changes in female body composition, with increased lean tissue mass and abdominal adiposity as frequent outcomes. For example, perinatal androgen excess in female rats, induced by a single 1-mg subcutaneous (SC)1 injection of testosterone propionate (TP) within 3 hours of birth, alters both lean and fat mass, increasing skeletal muscle mass along with preferential accumulation of abdominal adipose tissue (1). Prenatal androgen excess in heifers, produced by SC implants in the mother releasing 38 mg TP per day into the maternal circulation from day 60 until the end of gestation, exaggerates postnatal gain in body weight and adipose tissue (2). In ewes, similar prenatal androgen excess, produced by SC implants in the mother releasing 9 mg TP per day into the maternal circulation from days 40 to 60 of gestation until 3 weeks before lambing, increases total body skeletal muscle mass but does not alter adipose tissue accumulation (3). Apparently, female perinatal androgen excess differentially enhances lean and/or fat mass in rats, cattle, and sheep. While such alterations in body composition may be species dependent, age at onset or relative magnitude of the androgen excess may play an additional role in determining the phenotypic changes induced.
The consequences of prenatal androgen excess on female body composition in primates seem similar to those in rats and heifers, enhancing total body mass and adipose tissue accumulation. Initial studies by Goy and Robinson reported an increase in body weight at menarche in female rhesus monkeys after prenatal androgen excess (4). Kemnitz et al. demonstrated a small, but nonsignificant, increase in abdominal skinfold thickness and abdominal circumference in adult prenatally androgenized female rhesus monkeys (5). In a separate group of similarly androgenized female rhesus monkeys, we have shown a significant increase in abdominal skinfold thickness compared with controls of similar age and body mass index (BMI) (6). Whereas these alterations in somatometric measures suggest that prenatal androgen excess alters body composition in female rhesus monkeys, somatometric measurements alone are limited in accurately quantifying differential deposition of body fat in discrete body regions, such as the abdomen (7).
DXA, on the other hand, provides a more accurate and precise means of estimating body composition, including regional and total body lean and fat mass [rhesus (7), humans (8)(9)] by quantifying fat content in adipose tissue, as well as in nonadipose tissue, such as muscle and liver (10)(11). There are, however, limitations to DXA's ability to estimate the precise placement of fat within the body cavity. For instance, DXA yields measures of lean and fat mass in specific regions of interest, including the abdomen, but it is unable to differentiate intra-abdominal fat from subcutaneous abdominal fat (10)(11).
Whereas DXA can be combined with somatometric measures to estimate intra-abdominal fat content (11), a more accurate measure of intra-abdominal fat is obtained when DXA is combined with computed tomography (CT) (8)(9). CT also quantifies fat content, but unlike DXA, detects fat that is stored only within adipose tissue (10). In addition, CT clearly delineates between subcutaneous abdominal and intra-abdominal tissue, and combined with DXA, provides a more accurate determination of abdominal body fat distribution than either measure alone (10).
The purpose of this study was to examine body composition in prenatally androgenized and control adult female rhesus monkeys using a combination of DXA and CT, to determine whether prenatal androgen excess alters body fat distribution in these animals during adulthood, as suggested by somatometric measures (4)(5)(6). Determination of increased abdominal body fat mass in prenatally androgenized females compared with controls could provide a putative mechanism to explain how prenatal androgen excess adversely affects insulin-glucose metabolism in adult primates (12).
Research Methods and Procedures
Twelve midreproductive-aged, adult female rhesus monkeys (Macaca mulatta) were maintained at the Wisconsin Primate Research Center (WPRC) according to standard protocol (4)(13). Animals were fed Purina monkey chow (product no. 5038; Ralston Purina, Inc., St. Louis, MO) with occasional supplementation of fresh fruits and bread. This formulation of monkey chow provides 70% of calories as carbohydrate, 13% as fat, and 17% as protein.
Prenatally androgenized female rhesus monkeys were developed as previously reported (4). Briefly, six pregnant rhesus monkeys carrying female fetuses were injected with 5 (n = 1) or10 mg (n = 5) TP SC for 15 to 80 consecutive days (total gestation length, 165 days). The TP injections were initiated between days 40 and 44 of gestation. The timing, duration, dose, and dam age at the onset of TP injections are presented in Table 1. The early-treated prenatally androgenized females produced by this in utero androgen excess exhibited external genital masculinization, obliteration of the external vaginal orifice, and masculinized behavior (14). These female fetuses experienced fetal male levels of testosterone during sexual differentiation and major organogenesis (15)(16). The control group for this study was comprised of six females that were not exposed to prenatal androgen treatment, but similar to the prenatally androgenized females, were maintained according to standard WPRC protocol. Each prenatally androgenized female was pair-matched to a control female by age, body weight, and BMI (see below) before somatometric measurements and DXA/CT determination of body composition were obtained Tables 2 and 3). The range of body weight and BMI for the cohort of prenatally androgenized females examined in this study was similar to the range exhibited by our entire population of prenatally androgenized females (n = 28), as well as the cohorts of control and prenatally androgenized females described previously (6)(12). These monkeys have not been used in previously described studies of DXA- or CT-assessed body composition at WPRC (7)(17).
Table 1. TP exposure characteristics of prenatally androgenized females and age of dam at the start of TP exposure
TP dose (mg/d)
Start of TP injection (d)
Duration of TP injection (d)
Dam age at start of TP injection (year)
Table 2. Age and somatometric characteristics of the study animals
Prenatally androgenized females (n = 6)
Control females (n = 6)
All prenatally androgenized females were pair-matched with control females for age, body weight, and BMI.
The somatometric measurements were performed immediately before DXA scans, as previously validated for rhesus monkeys (7). The animals were anesthetized with ketamine HCl (7 mg/kg, intramuscularly) and xylazine (0.6 mg/kg, intramuscularly), and were then assessed for body weight, body length, body and limb circumference, and skinfold thickness. All measurements were performed with the animal supine except for abdominal and chest circumferences; for these, the animals were in a right lateral recumbency. Body length (crown-rump and crown-heel) was measured using a calibrated rule with a fixed headrest. Skinfold thickness was determined using a Lange caliper at 1) 5 cm above and below the umbilicus (as previously reported) (7), 2) the left triceps, and 3) the left pectoral. Circumferences of the abdomen, chest, upper arm, and upper leg were measured with a cloth tape measure to the nearest 0.1 cm. BMI was calculated as body weight (kilograms) divided by the square of crown-rump length (meters squared) (18).
DXA and CT
DXA is a highly precise and repeatable means of quantifying lean and fat mass in rhesus monkeys, with a precision (CV%) for total body composition and regional analysis of <4% (7). In this study, total body and abdominal lean and fat mass were measured with the animals in a supine position using DXA scanner model DPX-L (GE/Lunar Corp., Madison, WI) with pediatric software (version 1.5e) as previously reported (7). The abdomen, encompassing the entire length of the lumbar vertebral column, was defined as a region of interest. Each scan lasted ∼15 minutes. Nonabdominal fat mass was calculated by subtracting abdominal fat mass from total body fat mass.
While DXA was able to quantify total abdominal fat mass, CT distinguished intra-abdominal (visceral) from subcutaneous fat. Therefore, similar to methods used in humans (9)(10), the product of a single CT slice at the L2–L3 vertebral interspace and the DXA-determined total abdominal fat was used to estimate visceral fat mass for the study animals. The CT scans were performed and analyzed as previously described in humans (9)(10) using a GE High Speed Advantage CT (spiral) scanner (GE Medical Systems, Waukesha, WI) at 120 kVP and 200 MAS with a scan time of 1 s and slice thickness of 0.5 cm. The images were stored on DAT tape until analysis using a Sun workstation (Sun Microsystems, Inc., Mountainview, CA). Image data were converted to Hounsfield units that were most representative of adipose tissue. The lower and upper limits of adipose tissue density were −117.5 ± 4.1 and −26.3 ± 4.0 Hounsfield units (mean ± SEM), respectively. Visceral adipose was identified by removing the CT image exterior to the innermost abdominal muscle wall using a mouse-driven cursor (9)(10). The percent of intra-abdominal adiposity in the L2–L3 CT slice was determined by dividing the intra-abdominal adipose area by the total adipose area in that slice. This percentage was highly correlated with the percent of intra-abdominal adipose for the entire abdominal region of interest in these study animals (r2 = 0.73, p ≤ 0.0004), and this relationship did not differ between the two female groups (p = not significant). Visceral fat was calculated as the product of total abdominal fat determined by DXA and the percent intra-abdominal adiposity in the L2–L3 CT slice.
Normality of variables was confirmed using a Lilliefors test (two-tailed). All variables were compared using analysis of covariance (ACOVA; JMP Version 3.2.5; SAS Institute, Inc., Cary, NC) with type of female (prenatally androgenized or control) as a factor, BMI and total body fat as a covariate, and measures of body composition as dependent variables. The relationships between somatometric and CT/DXA data were assessed using simple linear regressions (both female groups combined). A p ≤ 0.05 value was considered significant. All data are expressed as mean ± SEM.
Tables 2 and 3 present the age and somatometric characteristics of the study animals. The factors used to pair-match the prenatally androgenized and control females (age, body weight, and BMI) were the same in both groups of females. BMI increased from ∼28 to 50 kg/m2, and body weight increased from ∼6.5 to 12.2 kg from the least to the most adipose of the pair-matched study animals. Body length (p = not significant), skinfold thickness (p = not significant), and body circumference (p = not significant) were not different between prenatally androgenized and control female rhesus monkeys.
Figures 1 and 2 illustrate the DXA/CT measures of lean and fat mass. Total body fat mass increased from ∼0.3 to 3.5 kg from the least to the most adipose of the pair-matched study animals. Total lean body mass (p = not significant), total body fat mass (p = not significant), nonabdominal fat mass (p = not significant), and abdominal lean mass (p = not significant) were similar in both groups of females (Figure 1A1B1C1D). Whereas there was a sizeable increase in abdominal fat in both prenatally androgenized (644 to 1684 g) and control females (100 to 1712 g) from the leanest to the most adipose of the pair-matched study animals, prenatally androgenized females exhibited 24% greater mean abdominal fat mass (p ≤ 0.04; Figure 2A) and 38% greater mean visceral fat mass (p ≤ 0.01; Figure 2B) than control females. Furthermore, five of six prenatally androgenized females had greater total abdominal fat mass than their pair-matched controls, and all of the prenatally androgenized females exhibited greater visceral fat mass.
Correlations between Abdominal and Total-Body Somatometric and DXA/CT Measures
To examine how closely abdominal and total body somatometric measures reflect DXA/CT measures of body fat distribution, a series of correlations were performed. BMI (r2 = 0.90, p ≤ 0.001) and body weight (r2 = 0.93, p ≤ 0.001) were highly positively correlated with DXA-determined total body fat. Abdominal circumference was similarly positively correlated with DXA-determined total abdominal fat mass (r2 = 0.87, p ≤ 0.001) and DXA/CT-determined visceral fat mass (r2 = 0.73, p ≤ 0.001). Lesser but significant correlations between somatometric and DXA/CT measures included abdominal skinfold thickness and either DXA-determined total abdominal fat mass (r2 = 0.53, p ≤ 0.01) or DXA/CT-determined visceral fat mass (r2 = 0.37, p ≤ 0.01).
The permanent physiological and behavioral consequences of prenatal androgen excess in female rhesus monkeys include impaired insulin action and secretion (6)(12), increased ovarian and adrenal androgen secretion (6)(19), luteinizing hormone hypersecretion (20), and masculinized behavior (14)(21). A limited number of studies of body composition in female rhesus monkeys exposed to prenatal androgen excess also have found altered body composition, based on somatometric measures, such as body weight, circumference, and skinfold measures (4)(5)(6). None have measured body fat distribution using more accurate means such as CT and DXA (4)(5)(6). This study used CT and DXA to measure body fat distribution in early-treated prenatally androgenized female rhesus monkeys in adulthood and identified increased abdominal adiposity as another important consequence of prenatal androgen excess.
The two groups of study animals were pair-matched by body weight and BMI before DXA/CT assessment of body composition. DXA confirmed the appropriateness of the matching, because DXA determined that total lean and total fat mass were similar in both female groups and both were highly correlated with body weight and BMI. Whereas BMI nearly doubled from the least adipose of the pair-matched study animals to the most adipose pair, abdominal fat more than quadrupled. This finding is not surprising because obesity in adult male and female rhesus monkeys is normally accompanied by preferential accumulation of abdominal fat (22). This study shows that preferential accumulation of abdominal fat in female rhesus monkeys is further enhanced by prenatal androgen excess, independent of obesity.
After adjustment for BMI- and DXA-determined total body fat mass, total abdominal fat mass and visceral fat mass in prenatally androgenized females are greater than expected by 24% and 38%, respectively, compared with control females. In other words, preferential abdominal adipose accumulation is not found only in the most adipose of the prenatally androgenized females but is found through the entire range of adiposity in these animals, from the least adipose to the most obese (22). While there is a preferential abdominal adipose accumulation in prenatally androgenized females, there does not seem to be a similar decrease in adiposity in nonabdominal adipose stores. By excluding the abdominal region of interest from analysis, the nonabdominal fat mass is nearly identical between the two groups of study animals.
Increased abdominal fat mass exhibited by our female rhesus monkeys exposed to early-gestation androgen excess is consistent with altered body fat distribution induced by pre- or perinatal androgen excess in nonprimate mammals. For example, prenatal androgen excess in heifers starting during the first one-third of gestation causes increased subcutaneous fat (2). Androgen excess at birth in female rats also causes increased accumulation of abdominal fat mass (1). Pre- or perinatal androgen excess in females alters body fat distribution with enhanced abdominal adipose tissue accumulation as one of the consequences.
One putative factor contributing to enhanced preferential abdominal and visceral fat accumulation in prenatally androgenized female rhesus monkeys may be altered adipocyte lipolytic activity (2)(23). Lipolysis is stimulated through binding of catecholamines to β-adrenergic receptors, causing downstream activation of hormone-sensitive lipase (23)(24). Prenatal androgen excess in female rhesus monkeys may impart a resistance to catecholamine induced lipolysis. In heifers, similar androgen excess causes adipocyte resistance to β-adrenergic-stimulation, leading to increased adipose mass accumulation (2). In hyperandrogenic women, resistance to catecholamine-stimulated lipolysis also seems to be associated with upper body obesity and abdominal adiposity (23). Similar to humans (23), any lipolytic resistance to catecholamines in prenatally androgenized female rhesus monkeys would be further enhanced by hyperinsulinemia. Therefore, as proposed in prenatally androgenized heifers (2) and hyperandrogenic women (23), lipolytic resistance to catecholamines in prenatally androgenized female rhesus monkeys may enhance adipocyte fat storage, leading to preferential accumulation of abdominal adiposity.
Because body fat distribution in humans is sexually dimorphic, another factor contributing to enhanced accumulation of abdominal adiposity in prenatally androgenized female rhesus monkeys may be a masculinized pattern of visceral fat accumulation. Men exhibit preferential accumulation of visceral fat compared with women matched for body weight or BMI, and this preferential accumulation occurs whether individuals are lean or obese (25)(26). In female rhesus monkeys, several sexual dimorphic traits are masculinized after exposure to prenatal androgen excess, including behavior (4)(21), external genital development (14), and steroid negative feedback on luteinizing hormone release (27). If similar to humans (25)(26), the accumulation of visceral adiposity may also be sexually dimorphic in rhesus monkeys and could be masculinized in females after prenatal androgen excess.
Accumulation of abdominal fat in normal rhesus monkeys leads to metabolic dysfunction, including hypertriglyceridemia, glucose intolerance, hyperinsulinemia, pancreatic β-cell exhaustion, and diabetes mellitus (22)(28)(29)(30)(31). Some of these metabolic disturbances are present in prenatally androgenized female rhesus monkeys. For example, adiposity-induced insulin resistance occurs in prenatally androgenized females with a body mass below that which causes similar insulin resistance in normal controls (32). Impaired pancreatic β-cell function occurs in these females after early-gestation androgen excess, and this impairment is shown to be independent of total body fat mass or BMI (6)(12). Therefore, this study offers insight into an additional consequence of prenatal androgen excess, namely enhanced preferential accumulation of abdominal fat, which may contribute to impaired pancreatic β-cell function in these animals (6)(12).
Whether prenatal androgen excess causes increased visceral fat through resistance to catecholamine-stimulated lipolysis (2)(23) or some other unknown mechanism, a preferential increase in visceral adipose in these animals probably enhances free fatty acid release into the portal circulation (33)(34)(35). In humans, increased androgens have been shown to mobilize the release of free fatty acids by stimulating hormone sensitive lipase (9)(36)(37). Therefore, stimulation of hormone sensitive lipase with enhanced free fatty acid release may occur in prenatally androgenized females as a consequence of their elevated ovarian (6) and adrenal (19) androgen concentrations during adulthood. Such physiological consequences of both increased visceral adiposity and hyperandrogenism in prenatally androgenized female rhesus monkeys could have profound detrimental effects on glucose-insulin homeodynamics.
Increased free fatty acid concentrations have also been implicated in the development of metabolic disturbances leading to impaired insulin action and secretion in humans (38). Experimentally increased circulating free fatty acid concentrations inhibit insulin-stimulated glucose uptake and use, while promoting endogenous glucose production (39). Such altered glucose uptake and use seem to result from free fatty acid disruption of the insulin signaling cascade (40), whereas the enhanced endogenous glucose production represents free fatty acid stimulation of gluconeogenesis (41). Prenatal androgen excess results in the elevation of circulating concentrations of free fatty acids in adult female rhesus monkeys (42). Enhanced visceral abdominal adiposity in prenatally androgenized female rhesus monkeys may promote increased circulating concentrations of free fatty acids and that could contribute to the disturbed metabolic function exhibited by these animals, particularly the enhanced adiposity-dependent insulin resistance (32), and impaired insulin action and secretion (6)(12).
In conclusion, DXA combined with CT is useful in quantifying subtle, yet significant, differences in body fat distribution in rhesus monkeys. Using these tools, prenatal androgen excess enhances the preferential accumulation of abdominal adiposity in female rhesus monkeys, possibly from altered lipolytic activity and masculinized visceral fat accumulation. Enhanced preferential accumulation of abdominal adiposity in female rhesus monkeys is another consequence of metabolic importance after prenatal androgen excess, and it may be inextricably linked to impaired insulin action and secretion in these females (6)(12)(32).
This work was supported, in part, by National Institute of Health Grants RR00167, AG11915, RR14093, and RR13635. This is publication number 42-016 of the WPRC. The authors thank S. G. Eisele, K. M. Boehm, and the Animal Care Staff of WPRC for management and maintenance of the animals; L. Mason for assistance with the DXA scans; M. A. Pozniak, F. Kelcz, T. Lanoway, and the Department of Radiology, UW-Madison for their expertise in performing CT scans; C. O'Rourke for veterinary care; and J. E. Reed and T. J. Welch at the Mayo Clinic for assistance with the analysis of CT scans.
Nonstandard abbreviations: SC, subcutaneous; TP, testosterone propionate; BMI, body mass index; CT, computed tomography; WPRC, Wisconsin Primate Research Center.