|Home | About | Journals | Submit | Contact Us | Français|
Prenatal androgens masculinize postnatal reproductive neuroendocrine function and behavior in sheep. Testosterone treatment of pregnant ewes during midgestation masculinizes sexual behavior and luteinizing hormone secretion in female lambs, presumably in part via aromatization and estrogen receptor (ESR) binding in the brain. We hypothesized that male and female sheep also differ in the number and distribution of ESR-containing neurons. If so, ESR expression should be sensitive to prenatal hormones delivered exogenously or in situ. ESR alpha (ESR1) was compared by immunocytochemistry in male and female lambs at the end of gestation, as well as in fetal females exposed prenatally to testosterone or dihydrotestosterone. ESR1-positive neurons were abundant in the posteromedial bed nucleus of the stria terminalis (BSTpm), medial preoptic area (MPOA), posterior medial amygdaloid nucleus (MeP), amygdalohippocampal area (AHi), ventromedial hypothalamic nuclei (VMH), and arcuate hypothalamic nuclei (ARC). In females, the ARC had the largest number of stained cells (mean ± SEM, 475.6 ± 57.4 cells/0.173 mm2), while staining intensity was greatest in the MPOA (mean ± SEM gray level, 31.3 ± 5.3). The mean ± SEM integrated gray level (IGL) was high in the ARC (0.63 ± 0.13) and in the MPOA (0.51 ± 0.08). The mean ± SEM IGL was low in the MeP (0.31 ± 0.10) and in the BSTpm (0.21 ± 0.06), while it was intermediate in the AHi (0.36 ± 0.10) and in the VMH (0.37 ± 0.07). ESR immunostaining was not significantly different in male and female fetal lambs, nor in females fetuses exposed prenatally to androgens (P > 0.05). However, ESR1 staining was significantly increased in the ARC, MPOA, and AHi of adult rams vs. adult ewes. These results suggest that brain ESR immunoreactivity in fetal lambs is unlikely to account for postnatal sex differences in reproductive function. Instead, sex differences in ESR emerge postnatally.
Exposure to gonadal steroid hormones during the prenatal critical period has profound organizational effects on reproductive function and behavior in sheep . For instance, androgens from the developing testes cause male lambs to experience earlier pubertal onset of tonic pulsatile secretion of luteinizing hormone (LH) . Conversely, female sheep demonstrate a preovulatory surge of LH and gonadotropin-releasing hormone (GnRH) in response to rising levels of ovarian estradiol; the LH surge is abolished in males . During the autumn breeding season, rams demonstrate intermale aggression and masculine sexual behavior (courtship of estrous females, mounting, and copulation), while ewes show receptive sexual behaviors during estrus . Similar to the development of sex differences in other longer-gestation species (guinea pigs, primates, and humans [4, 5]), the critical period for sexual differentiation in sheep occurs during the middle of gestation . The testes of the fetal ram produce testosterone by 30 days of gestation  (145–150 days is term). Circulating androgens reach 1 ng/ml in the male fetus by 64 days of gestation  and fall to 0.3 ng/ml near the end of gestation .
As in guinea pigs and primates , experimental manipulation of the prenatal hormonal environment in sheep can reverse the normal process of sexual differentiation. Systemic treatment of pregnant ewes with testosterone masculinizes sexual behavior and alters reproductive function in their female lambs [6, 9–15]. These findings parallel many virilizing symptoms of congenital adrenal hyperplasia in young girls , suggesting that the hormonal control of phenotypic sexual differentiation is similar in sheep and humans.
In males and females, reproductive neuroendocrine function and sexual behavior are exquisitely sensitive to gonadal steroid hormone action in the brain. Steroid negative feedback modifies secretion of GnRH from the hypothalamus [17, 18]. Steroids facilitate the sex-specific expression of mating and aggressive behavior as well . In large part, the actions of gonadal steroids are thought to be mediated in males and females via binding to estrogen receptors (ESRs) in the limbic system . In particular, circulating androgens from the testes of the male can be aromatized locally in the brain to estradiol . Considering the substantial sex differences in steroid-dependent physiology and behavior, it is reasonable to expect that male and female sheep may differ in the number and distribution of ESR-containing neurons. Thus, ESR expression should be sensitive to prenatal hormone treatments. To evaluate organizational sex differences, we measured immunoreactivity for ESR alpha (ESR1) in male and female fetal lambs at the end of gestation (145 days), which is after the critical period for sexual differentiation [6, 20, 21] but before postnatal activational effects of steroids. In addition, to determine if prenatal androgens modify ESR in females, we compared ESR in fetal females exposed to testosterone or dihydrotestosterone (DHT) during the prenatal critical period (30–90 days of gestation). ESR1 staining in adult rams and ewes was evaluated to determine if sex differences in ESR are manifest postnatally under the combined organizational and activational influence of steroids.
Methods to masculinize female lambs in utero were similar to those reported previously . Beginning on Day 30 of gestation, pregnant Suffolk ewes (n = 75) of known conception dates received biweekly 1-ml injections i.m. of cottonseed oil, testosterone (100 mg/ml), or DHT (100 mg/ml). Testosterone- and DHT-treated ewes received androgen (200 mg/wk) for 9 wk until Day 90. This duration spans the critical period for sexual differentiation of the external genitalia and reproductive behavior in their fetal lambs . The resulting fetal lambs included control (n = 5), testosterone-treated (n = 7), and DHT-treated (n = 6) females, as well as control males (n = 6). Each experimental group included singletons and twin lambs. Control males and females were obtained from mixed-sex pairs; sheep do not usually form vascular connections between twin placentae , comparable to freemartinism in cattle. In a subsequent experiment, brains were obtained from adult ewes (n = 4) and adult rams (n = 4) during the nonbreeding season to determine if sex differences in ESR are present in adulthood. All investigations were approved by the Institutional Animal Care and Use Committee of the University of Michigan and were conducted in accord with the National Research Council publication Guide for Care and Use of Laboratory Animals (copyright 1996, National Academy of Science).
Brains were obtained from 24 fetal lambs at 145 days of gestation. Pregnant ewes were initially anesthetized with ketamine (4–6 mg/kg) and diazepam (0.2–0.3 mg/kg) i.v. and were subsequently maintained under inhalant anesthesia (1.5% halothane in a 2:1 oxygen:nitrous oxide mixture). Lambs were obtained via laparotomy. After clamping the umbilical cord, fetal lambs were deeply anesthetized with sodium pentobarbital and perfused transcardially with 150 ml of 0.1 M PBS containing 0.1% sodium nitrite for vasodilation, followed by 250 ml of 0.1 M phosphate buffer (PB) containing 4% paraformaldehyde (pH 7.3) at room temperature. Brains from adult rams and ewes were obtained as described by Moenter et al. . Sheep were euthanized with a barbiturate overdose and immediately decapitated. Heads were perfused via the carotid arteries with 6 L of PB containing 4% paraformaldehyde, 0.1% sodium nitrite for vasodilation, and heparin (1 U/ml) to prevent clotting.
Brains from fetal lambs and adult sheep were removed, postfixed overnight in the perfusion fixative, and cryoprotected in 30% sucrose with 37.5% ethylene glycol and 1.25% polyvinylpyrrolidone in PB at 4°C. Before freezing on dry ice, brains were divided sagitally and blocked from the nucleus accumbens rostrally to the mammillary bodies caudally and at the dorsal limit of the lateral ventricles. Coronal sections (50 μm) were cut on a freezing microtome (K400; Microm International, Walldorf, Germany) and collected in PB with 0.01% sodium azide as a preservative. Sections were stored at 4°C until processed for immunocytochemistry.
Every fourth serial section was immunostained for ESR1. Brains from lambs in different experimental groups were stained at one time. Likewise, sections from adult rams and ewes were processed together. All incubations and washes were performed at room temperature with gentle agitation. Sections were incubated overnight in primary antiserum with 4% normal donkey serum and 0.3% Triton X-100 in PB. The monoclonal mouse anti-ESR1 antibody (1D5 clone; Invitrogen Co., Carlsbad, CA) was used at 1:1000. The 1D5 ESR1 antibody has previously been characterized in sheep brain [25, 26]. The next day, sections were exposed to biotinylated donkey anti-mouse secondary antibody (1:400; Jackson Immunoresearch, Malvern, PA), followed by avidin-biotin-horseradish peroxidase complex (1:800, Vectastain ABC Elite Kit; Vector Labs, Burlingame, CA), each for 1 h with extensive washes in between with PB between. The horseradish peroxidase was visualized with NiCl-enhanced 3,3′ diaminobenzidine (DAB) as the chromagen to produce a blue-black reaction product over labeled neurons and fibers. The DAB reaction was terminated by extensive washes in PB. Sections were mounted on gelatin-coated slides, dehydrated, and coverslipped with Permount (Fisher Scientific, Pittsburgh, PA).
Sections were viewed under bright-field illumination with an Olympus BH-2 microscope (Tokyo, Japan). The microscope was adjusted for Kohler illumination on the 20× objective. Photomicrographs of selected brain regions, including the posteromedial bed nucleus of the stria terminalis (BSTpm), medial preoptic area (MPOA), posterior medial amygdaloid nucleus (MeP), amygdalohippocampal area (AHi), ventromedial hypothalamic nuclei (VMH), and arcuate hypothalamic nuclei (ARC) (Fig. 1), were collected using a digital camera (Camedia; Olympus). For each brain region, ESR1 immunoreactivity was measured in an area measuring 480 × 360 μm (0.173 mm2). Digital images were converted to gray scale and resized to 600 × 450 pixels using Adobe Photoshop CS2 software (Adobe Systems Inc., San Jose, CA). The same light levels and camera settings were used during acquisition and analysis of all images. In addition, to control for uneven illumination, a white reference image (no specimen) was subtracted from each experimental image. To ensure that staining was not saturated, maximum pixel intensity for all images was <256. The resulting images were evaluated using NIH Image software (version 1.63; National Institutes of Health, Bethesda, MD). Initially, each image was thresholded to eliminate background staining. The threshold was defined as 2 SDs above the median gray level for each image. Sex differences in ESR1 may include differences in the number of estrogen-responsive neurons and/or the abundance of ESR1 within individual neurons. To encompass these different possibilities, the number of objects (representing the number of ESR1-positive neurons), the mean gray level (reflecting the amount of ESR1 within individual neurons), and the integrated gray level (IGL) were determined for all pixel values above threshold. The IGL is calculated as the sum of gray level intensities for all pixel values above threshold (expressed in millimeters squared). In this manner, IGL reflects the combination of staining intensity (the mean gray level) and staining area (the number of stained pixels above threshold) within the region of interest.
To validate automated cell counting, we compared the number of ESR-positive cells in VMH identified by the NIH Image system with that counted by an observer blind to the experimental group using a microscope equipped with a drawing tube. One section from each fetal lamb was evaluated. NIH Image identified a higher mean ± SEM number of cells (256.8 ± 15.9) than that counted by eye (200.0 ± 20.3) (P < 0.05). However, there was a significant positive correlation (R2 = 0.6, P < 0.05) between cell counts obtained using these two methods.
For comparisons of the BSTpm, MPOA, MeP, VMH, ARC, and AHi in adult rams and ewes, control male and female fetal lambs, as well as females exposed to androgens in utero, were analyzed bilaterally (one section per animal through each region), and the two measurements for each region in each animal were averaged. To minimize variance, values were square root transformed before statistical comparison by Student t-test to determine sex differences in control males and females or by ANOVA to evaluate effects of prenatal androgen treatment. Statistical analyses were performed using Statview 5.0 (SAS Institute, Cary, NC) with post hoc analysis using Fisher probable least-squares difference. For all comparisons, P < 0.05 was considered statistically significant.
Figure 1 shows ESR immunoreactivity in the AHi, ARC, BSTpm, MeP, MPOA, and VMH from a representative control female fetus. The distribution of ESR1 staining in the late-gestation fetal lamb closely resembled the pattern of ESR1 immunoreactivity in adult male and female sheep previously described . ESR1 staining was concentrated in cell nuclei, with limited perinuclear cytoplasmic immunoreactivity (Fig. 1). ESR1-positive neurons were distributed throughout the ventral telencephalon and hypothalamus, including the BSTpm, MPOA, VMH, ARC, and the ventrolateral septum and anterior hypothalamus. In addition, labeled cells were present in nuclei of the corticomedial amygdala, particularly the MeP and the AHi.
Among control female lambs, ESR1 staining was abundant in the ARC and in the MPOA. The number of objects and the mean gray level are given in Table 1; Figures 2 and and33 show IGL in fetal (Fig. 2) and adult (Fig. 3) sheep, reflecting overall ESR1 staining. While ARC had the largest number of stained cells (mean ± SEM, 475.6 ± 57.4 cells/0.173 mm2), staining intensity was greatest in the MPOA (mean ± SEM gray level, 31.3 ± 5.3) (Table 1). As a result, IGL was similar in the two regions (0.63 ± 0.13 for the ARC and 0.51 ± 0.08 for the MPOA) (Fig. 2). By contrast, the mean ± SEM IGL was low in the MeP (0.31 ± 0.10) and in the BSTpm (0.21 ± 0.06) was low, reflecting the limited number of stained neurons and the lower staining intensity (measured as the mean gray level) in these regions. The mean ± SEM ESR1 immunoreactivity (measured as the IGL) was intermediate between these extremes in the AHi (0.36 ± 0.10) and in the VMH (0.37 ± 0.07).
The distribution and intensity of ESR1-positive neurons were similar in control male lambs. There were no statistically significant sex differences in the number of stained cells, the mean gray level, or the IGL in any brain region (P > 0.05). Likewise, there were no differences in ESR1 immunoreactivity among fetal females exposed prenatally to testosterone, DHT, or vehicle (Fig. 2) (P > 0.05).
We also compared ESR1 staining in adult rams and ewes to determine if sex differences in ESR protein are manifest postnatally. Because tissues from adults and fetal lambs were not stained in parallel, it is impossible to perform statistical comparison of results from the two experiments. Nonetheless, although ESR1 staining was not significantly different in fetal males and females, sex differences in ESR1 staining were evident in adults (Table 1). Specifically, males had higher mean ± SEM numbers of ESR-stained cells than females in the MPOA (197.8 ± 39.0 vs. 96.5 ± 17.7 cells/0.173 mm2) and in the AHi (153.5 ± 42.1 vs. 30.0 ± 11.9 cells/0.173 mm2) (P < 0.05 for both). In the ARC, the mean ± SEM gray level was higher in males (29.4 ± 0.7) than in females (14.7 ± 4.4) (P < 0.05). Likewise, the mean ± SEM IGL was higher in males (0.5 ± 0.1) than in females (0.2 ± 0.1) (P < 0.05) (Fig. 3).
The present study demonstrates that there is little evidence for sex differences in ESR1 immunoreactivity in late-gestation fetal lambs, despite significant sex differences in postnatal estrogen-sensitive social behavior and neuroendocrine function. At 145 days of gestation, the distribution of ESR immunostaining in male and female lambs resembles that in adult males and females as reported by Lehman et al. . Similar to other mammalian species, ESR1-positive cells in sheep are widely distributed throughout the brain but are concentrated within a group of interconnected brain nuclei that regulate sexual behavior and gonadotropin secretion. As measured by the number of immunopositive cells, the mean gray level, or the IGL, there were no differences between male and female fetal lambs or between control females and those exposed prenatally to testosterone or DHT. Although the prenatal organizational actions of gonadal steroids masculinize behavior, including nonreproductive behavior in neonates (urination posture and play behavior [28–30]), the pattern of ESR1 staining in brain regions regulating reproductive function in late gestation is not sexually dimorphic. However, we observed selective sex differences in ESR staining in adult male and female sheep, with greater ESR immunoreactivity in the MPOA, AHi, and ARC of rams. Specifically, the mean gray level and the IGL of males exceeded those of females in the ARC, while males had significantly more labeled cells than females in the MPOA and in the AHi. These sex differences in ESR1 may contribute to sex-specific responses to estrogen in adult sheep [2, 15].
Research in rodent models has had a pivotal role in establishing the first principles of brain sexual differentiation and the distribution of steroid receptor-containing neurons . From pioneering studies of rats in the 1960s , we understand that the female is the default phenotype in mammals and that males are masculinized and defeminized under the influence of testicular steroids during a critical period for sexual differentiation. Furthermore, investigators have accumulated a detailed map of androgen receptor- and ESR-containing neurons in the rodent brain by in vivo autoradiography, receptor binding, immunocytochemistry, and in situ hybridization . Through complementary investigations of physiology and behavior, it is possible to relate certain circuits of steroid receptor-containing neurons to distinct aspects of steroid-dependent behaviors . In this regard, the six brain regions evaluated in the present study have dense interconnections and abundant populations of ESR- and androgen receptor-containing neurons, and all are associated with steroid-dependent aspects of sexual behavior and neuroendocrine function . In mammals, including sheep [34, 35], the MPOA and the VMH are central to steroid-responsive expression of sexual behavior and regulation of reproductive neuroendocrine function [36, 37]. Together with the ARC, these midline hypothalamic nuclei also contribute to the hormonal regulation of GnRH from the median eminence . The MPOA, VMH, and ARC receive afferent input from other steroid-sensitive brain regions, including the MeP, BSTpm, and AHi . In addition to transducing circulating hormonal stimuli, the MeP, BSTpm, and AHi are involved in processing chemosensory stimuli from conspecifics for social behavior [35, 37]. Thus, any sex differences in ESR-containing neurons in these regions would have potential to influence patterns of masculine and feminine sexual behavior.
Sex differences in ESR-containing neurons have been previously reported in rodents. As measured by autoradiography , receptor binding [41–43], in situ hybridization , or immunocytochemistry [45, 46], the VMH consistently demonstrates more ESR in adult females than in male rats, voles, and guinea pigs. Several studies [40–43, 45, 46] also report a bias toward females for ESR in the MPOA, BST, Me, or hypothalamus. The sex difference in ESR in the MPOA and in the VMH develops during the first few weeks of rodent postnatal life [47–50], which encompasses the critical period for sexual differentiation. Our observations of a sex difference favoring ESR1 immunoreactivity in the AHi, MPOA, and ARC from adult rams stand in contrast to these findings in rodents. However, it is important to acknowledge the potential for circulating hormone levels to influence the levels of ESR1 immunostaining. To our knowledge, staining with the 1D5 ESR1 antibody has not been compared under different hormonal conditions (e.g., ovariectomized vs. ovariectomized estradiol-treated). Future studies should compare staining for ESR1 in gonadectomized adult male and female sheep.
In a comparison of adult male and female sheep during the fall breeding season, Scott et al.  found increased ESR mRNA in the VMH of females. Variations in measuring techniques (in situ hybridization vs. immunocytochemistry) or in circulating hormone concentrations between the breeding and nonbreeding season most likely account for the differences between their study and the present study. In this regard, circulating steroid levels in ewes and rams are considerably higher during the breeding season vs. nonbreeding season (5 pg/ml vs. undetectable for estradiol in ewes  and 10 vs. 5 ng/ml for testosterone in rams ). Furthermore, the present study evaluated different aspects of ESR1 staining, namely, intensity, number of objects, and IGL (reflecting both intensity and area). Sex differences in ESR1 may reflect variations in the number of estrogen-responsive neurons, alterations in the abundance of ESR1 within individual neurons (reflected in the mean gray level), or elements of both (as represented by the IGL). Although sex differences in ESR1 immunoreactivity reported herein reached statistical significance for the IGL and the mean gray level only in the ARC, and for number of objects in the AHi and in the MPOA, most measures of ESR1 staining also favored immunoreactivity in males. Females showed a slight increase in the mean gray level in the AHi, but this was not significant. A study  of ESR binding in gonad-intact male and female sheep during the nonbreeding season found a sex difference in the amygdala, with occupied and unoccupied ESR in males exceeding that in females. Furthermore, men show increased ESR1 expression in the MPOA compared with that in women , similar to our results in sheep. This suggests that ESR in males and females may develop according to different patterns when brain sexual differentiation occurs prenatally vs. postnatally.
As a model for sexual differentiation, sheep, guinea pigs, and primates present a significant contrast with the concepts developed from rodent studies. In particular, the prenatal critical period and the extended prepubertal juvenile phase in sheep and primates offer potential for understanding aspects of human sexual differentiation [56, 57]. Some aspects of steroid-dependent behavior and steroid receptor-containing neurons are similar to the rodent model. For example, the brain regions sampled for ESR1 in the present study are the same as those with large numbers of ESR1-positive neurons in rodents [27, 51]. The intracellular distribution of ESR1 immunoreactivity is also similar: staining is concentrated in the cell nucleus, with faint labeling extending into the adjacent cytoplasm. Thus, the absence of sex differences in ESR1 immunostaining among late-gestation fetal lambs stands in contrast to what has been reported in the neonatal rat [47–50]. However, measurement of ESR in neonatal rats is during the critical period, while the critical period has already passed in late-gestation fetal lambs. Another consideration is the difference in the hormonal milieu of the late-gestation fetal lamb and that of the early postnatal rat. In this regard, male and female fetal lambs are each exposed to maternal estrogens and progestins via transplacental transport . This could account for the absence of sex differences in ESR from fetal lambs in the present study.
Perhaps due to the species-specific timing of the critical period (midgestation in sheep vs. perinatal in rodents), the relative role of androgens vs. estrogens is one of the most important features of sexual differentiation that differs in rodents vs. long-gestation species. In the rodent model, masculinization of the male brain is largely under the influence of estrogens produced by local aromatization of testosterone . Much evidence has shown that estrogen treatment of neonatal male or female rat pups induces masculinization comparable to that induced by testosterone, yet DHT is largely without effect . Conversely, androgens have a much more important role in sexual differentiation of the brain in sheep, guinea pigs, and primates . Prenatal exposure to testosterone abolishes the preovulatory GnRH/LH surge in sheep, while DHT is without effect . Accordingly, we infer that estrogen exposure during the critical period is essential to defeminize the surge in sheep, as it is in rats . However, prenatal DHT masculinizes control of tonic pulsatile LH secretion at puberty in young lambs . Likewise, female lambs exposed prenatally to DHT seem to be less attractive to adult rams, although they still show proceptive behavior . The relative roles of DHT and estradiol are slightly different in guinea pigs, although prenatal androgens have a substantial role in sexual differentiation as well . Although prenatal androgen treatment, whether through testosterone or DHT, masculinizes adult reproductive physiology and behavior, the results of the present study suggest that this is not accompanied by prenatal remodeling of the principal populations of ESR1-containing neurons.
In addition to ESR1, there is also potential for sex differences in the beta form of the estrogen receptor (ESR2) and in androgen receptors (AR). At the present time, immunocytochemistry for ESR2 or AR in sheep brain is problematic. As described in a recent report, ESR2 can only be visualized by high temperature antigen retrieval , which is incompatible with the tissue preparation in the present study. However, as with ESR1, female sheep had more ESR2-containing neurons in VMH as measured by in situ hybridization . Similar findings have been reported in rats [60, 61]. By contrast, sex differences in AR immunoreactivity favor males, with increased levels of AR mRNA and protein in MPOA and BST of male sheep , hamsters , rats , and mice . Similar sex differences in AR were not identified in human brain .
The prolonged phase of juvenile development in sheep is useful to reveal organizational effects of steroids on nonreproductive behavior. Although we did not observe significant differences in ESR staining in male and female fetal lambs, sex differences are evident in urination and play behavior even during the first 10 wk of life. Compared with females, male lambs spend more time with each other than with their mother, engaging in play aggressive behavior such as head-butting and mounting [28, 29]. By contrast, female lambs show greater preference for staying near their mother and display low levels of aggressive play . Sex differences in play behavior have also been reported in rhesus monkeys . Like adult rams and ewes, neonatal females squat while voiding in a continuous stream, but males stand and void in spurts . In this regard, urination posture is determined centrally and not in response to genital morphology. Female lambs masculinized with testosterone during the latter part of the critical period have feminine genitalia but adopt a male urination posture . Despite these early sex differences, we perhaps should not expect that organizational effects of steroids will be accompanied by sex differences in ESR-positive neurons. By definition, organizational sex differences do not require continued exposure to gonadal steroids. Thus, steroid receptors are not central to the expression of sex differences before puberty.
Although sex differences in ESR were not evident in fetal lambs, adult rams demonstrated increased ESR immunostaining in the MPOA, AHi, and ARC relative to that in adult ewes. These sex differences could contribute to masculine sexual behavior (MPOA) , responsiveness to sexually relevant chemosensory cues (AHi and MPOA) [35, 69], or regulation of gonadotropin secretion, possibly through kisspeptin (ARC) . When male and female fetal lambs are exposed to high levels of maternal estrogens, sex differences in ESR may not be evident prenatally. However, our data suggest that sex differences may emerge postnatally, perhaps under the combined actions of organizational and activational steroids.
We thank Ms. Cortney Ballard and Ms. Morgan Cross for assistance with histologic processing.
1Supported by NIH PO1 HD044232.