| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
research-article |
Reproductive Sciences Program,3 Departments of Psychology,4 Pediatrics,5 Obstetrics and Gynecology,6 Molecular and Integrative Physiology,7 and Ecology and Evolutionary Biology,8 University of Michigan, Ann Arbor, Michigan 48109
ABSTRACT
As steroids and steroid-like compounds accumulate in the environment, it has become important to understand how low-dose exposure affects reproductive function. Ovary-intact sheep were used in a multigenerational study, to determine whether chronic exposure to low levels of estrogen disrupts reproductive function and behavior. We assessed parameters of reproductive performance in control and postnatally estradiol-treated females (Generation 1, G1), and their offspring (Generation 2, G2). In the G1 animals, 17beta-estradiol (E) was administered continuously from 4 wk of age at two doses via subcutaneous implants (ultralow E [<1 pg/ml in circulation, n = 8] or low E [1–3 pg/ml, n = 8]). Both doses delayed puberty; low E also produced pronounced prepubertal and seasonal anestrus hypogonadotropism, and delayed the onset of the second breeding season. All G1 animals conceived and produced offspring (G2), the treatment of which resulted from continuous maternal exposure during pregnancy and lactation. Behavioral observations of G2 females revealed that low prenatal E modestly masculinized play behavior and increased the frequency of attempts to displace competitors relative to ultralow E and control animals. The timing and magnitude of the LH surge also differed in prepubertal low prenatal E females relative to the controls, although these differences were not evident when retested at one year of age. These findings support the hypothesis that chronic exposure to physiologic amounts of exogenous estrogens has multigenerational effects on behavior and neuroendocrine function. Despite these disruptive steroid actions, ovarian cyclicity and fertility are not invariably compromised, pointing to an impressive resiliency of the reproductive axis to insult by exogenous estrogenic compounds.
behavior, environment, estradiol, neuroendocrinology, puberty
The sensitivity of the developing reproductive neuroendocrine system to insult by exogenous agents has been demonstrated in many species [1–3]. The sheep, which is a precocial species, is a practical agromedical model to assess the potential consequences of prenatal and postnatal exposure to exogenous steroids. During fetal life, irreversible functional abnormalities shaped by sex steroid exposure can develop, some of which are not manifest until sexual maturity. In females, the effects of prenatal exposure to excess androgens, estrogens, and sex steroid-like compounds include changes and defects in anatomy, physiology, and behavior [2, 4–11]. Among these disruptions, the feedback mechanisms that regulate neuroendocrine function during the ovarian cycle are particularly sensitive, and the severities of these effects vary depending on the timing, dose, and duration of prenatal steroid exposure [1, 4, 7, 10, 12].
Postnatal exposure to steroids and steroid-like compounds, outside of the recognized critical period of development, can also disrupt reproductive function. Exposure to estrogenic forage, such as subterranean clover (Trifolium subterraneum), has been shown to cause reduced conception rates, increased embryonic loss, and impaired ovarian function in cattle and sheep [13]. In some regions, these reproductive dysfunctions have resulted in substantial reductions in agricultural productivity [14, 15].
Chronic exposure of female sheep to low levels of exogenous estradiol (E) after birth delays the onset of reproductive cycles [16]. Furthermore, the possibility that the impact of postnatal steroids is related to prenatal steroid exposure is raised by the finding that the incidence of ovulatory failure in intact female sheep exposed in utero to androgens increases with age [17, 18]. It is possible that cyclic increases in estradiol during adulthood exacerbate the aberrant programming actions of prenatal steroid exposure. Understanding the pathophysiologic consequences of exposure to prenatal and postnatal substances requires that the role of each type of exposure be tested independently.
The aim of this multigenerational study in female sheep was to assess the effects of chronic exposure to physiologically relevant doses of estradiol before and after birth on reproductive development and function. We used a Silastic capsule inserted subcutaneously that was designed to produce very low doses of estradiol (<1 to 3 pg/ml in the circulation of an agonadal individual), since the sheep should be particularly sensitive to such low levels. In this species, compared to others, circulating estradiol is very low, normally ranging from less than 1 pg/ml during the mid-luteal phase to approximately 10 pg/ml 2–3 days prior to the pre-ovulatory LH surge [19]. Low estradiol levels are more potent in sheep than in other species due to the low binding affinity of sex hormone-binding globulin (SHBG) for estradiol in this species [20, 21].
Our approach was to investigate the effects of chronic estradiol beginning shortly after birth (postnatal E study) on: 1) tonic gonadotropin secretion, 2) the timing of puberty, 3) the seasonality of reproduction, and 4) reproductive success. As part of the assessment of reproductive success, the investigation was continued into pregnancy and lactation under chronic low-level estradiol exposure. This provided the opportunity for a second study focusing on a successive generation with exposure to estradiol resulting from maternal treatment. This prenatal-lactational E (referred to hereinafter as prenatal E) study focused on testing the effects of chronic estradiol exposure in utero and during suckling on: 1) the timing of puberty, 2) behavior, and 3) the dynamics of the LH surge system.
Generation 1: Postnatal Estrogen (G1)
Female Suffolk lambs born 21 March (± 1.0 day SEM) were studied outdoors in their natural environment at the Reproductive Sciences Program Sheep Research Facility (Ann Arbor 42° 18‘N latitude). After weaning, the lambs had access ad libitum to alfalfa hay and pellets, pasture, and water, and were continuously housed with mature, vasectomized rams. Body weights were obtained weekly until all animals reached 50 kg (26 wk of age), a weight typical of an adult sheep.
Sixteen of the 21 lambs were treated chronically with low levels of estradiol from 4 wk of age (mid-April). The remaining five females served as untreated controls. Estradiol treatment was by s.c. implantation in the axillary region of a small Silastic capsule (i.d. 0.33 cm; o.d. 0.46 cm; Dow-Corning Corp., Midland, MI) that contained a packed column of crystalline 17beta-estradiol (Sigma, St. Louis, MO) and that was sealed with Silastic adhesive Type A (Dow Corning). Of the sixteen E-treated females, eight received a capsule that contained a 10-mm column of E and a separate capsule that contained a 3-mm column of E, for a total of 13-mm (postnatal low E group). The 13-mm implant has been shown to produce blood levels of 1–3 pg/ml E in similarly treated animals [22, 23]. The remaining eight treated animals received only an implant that contained a 3-mm column of E (postnatal ultralow E group). Since the steroid levels released from this type of implant are directly proportional to the implant surface area, a single 3-mm column produces a chronic increase of less than 1 pg/ml, a concentration that approaches the limits of detection of currently available radio-immunoassays for ovine estradiol. All implants were incubated in tap water for 24 h and then in 70% ethanol for 20–40 minutes before insertion, to avoid an initial, transient increase in the circulating level of E. The estradiol implants were replaced and inspected after one year. Based on visual inspections, there were no instances of implant leakage or exhaustion of steroid content, indicating continuous and consistent steroid release throughout the treatment period.
Sexual Maturity and Reproduction
The onset and maintenance of reproductive cycles at puberty (Year 1) and for the subsequent breeding season (Year 2) were determined by measuring serum progesterone in samples collected twice-weekly. Serum LH was also measured in samples collected twice-weekly from 10 to 68 wk. This sampling schedule included the pre-pubertal and pubertal periods, the first breeding season, and the beginning of the first postpubertal anestrus season. Additional samples were collected for measurement of serum LH for 6 wk (early June through July) in the middle of the first postpubertal anestrus season, to assess differences in gonadotropin release among the treatment groups during a time when young females once again become hypersensitive to the estrogen negative feedback of GnRH secretion [24]. All blood samples (3–5 ml) were collected by jugular venipuncture and stored at 4°C for 24 h before centrifugation. Serum was harvested and then stored frozen (–20°C) until assayed for LH and progesterone. At the end of the second breeding season, in early January, all 21 females were mated with a mature, intact ram. The E implants remained in place throughout pregnancy and lactation. Measurements of serum progesterone were continued throughout pregnancy.
Generation 2: Prenatal Estrogen (G2)
Lambs were born 5 June (± 1.05 days SEM) to G1 ewes at the Reproductive Sciences Program Sheep Research Facility. The experimental group of G2 lambs was based on the treatment of their G1 mothers; the G2 females received no additional E treatment. Female lambs born to control mothers experienced no exogenous E during gestation and served as controls (n = 5). Eight female lambs were born to ultralow E mothers, and five female lambs were born to low postnatal E mothers. Twelve males were also borne to the G1 ewes and raised with the G2 females until 8 wk of age. All of the males were castrated before 2 wk of age. Body measurements were taken within six days after birth, and body weights were obtained at least weekly through 5 wk of age and each month thereafter until all the females reached 30 kg (approximately 12 wk of age). Females that weigh 30 kg or more are considered to have sufficient energy reserves to support the initiation of puberty [25]. All experimental procedures were performed according to NIH guidelines and were approved by the University Committee on the Use and Care of Animals.
Behavior is sensitive to modification by the organizational actions of sex steroids during development and can be used to assess the influence of estrogen exposure during maturation [2, 9, 26–28]. G2 female lambs were monitored by focal analysis to quantify sexually dimorphic behaviors, which included association patterns, play, and competitive displacement behaviors. Focal analyses were initiated at 2 wk of age, and consisted of 10-min observational periods during which all the behaviors of one individual were recorded. Each female lamb was observed at least twice-weekly for 12 wk (from 2 to 14 wk of age), to produce a minimum of 20 min of focal data every week, with equal observational time allocated across individuals. After 14 wk of age, the lambs were used for other aspects of the research (see below), and frequent behavioral observations were discontinued.
The behaviors exhibited by an individual were grouped into one of three categories: association, play, and displacement, and were analyzed separately. The behaviors were recorded by noting the date, time, focal animal, and other individuals with whom the focal animal was interacting. All behavior was recorded as events; the only durations noted were for associations during focal observations. Behaviors recorded included: 1) associations (while eating, resting, standing, following, and sniffing within 1 m of the associating animal), 2) nursing interactions with the ewe and sibling (if one existed), 3) play (jumps, runs [with or without group], head-butting, mounts, displacements from atop a high location), and 4) displacements (while eating hay or pellets, from in front of fan, from sleeping location). The individual attempting to displace a competitor was recorded as initiating the displacement. Successfully displacing another lamb was considered a win, while being displaced or failing to displace was considered a loss. Beginning at 6 wk of age, the displacement behavior among all the lambs (non-focal analysis) was recorded for a 1-h interval 2–3 times per week at the time the animals were fed (approximately 0700 h each day). Little displacement at the feeders was recorded prior to 8 wk of age.
The female lambs within each treatment group were allowed to interact freely with one another, as well as with male siblings (castrated by 2 wk of age) and their mothers until 8 wk of age, at which stage weaning took place and the mothers and male lambs were removed. To permit easy identification during observation, all lambs were numerically dyed on their sides with nyanzol-D dye (American Color and Chemical Corp., Charlotte, NC). Based on behavioral similarities between prenatal ultralow E lambs and controls, we concluded that the prenatal ultralow E treatment did not affect neural development and, thus, this group was excluded from subsequent studies of reproductive neuroendocrine function in G2 lambs (described below).
Neuroendocrine Function and Sexual Maturity
The stimulatory feedback action of estradiol on LH secretion was tested in G2 females (low prenatal E and control) at 14 wk of age using the following paradigm. Blood samples (3 ml) were collected every 2 h for 6 h before and 36 h after implantation of four 30-mm Silastic estradiol implants. This surge induction protocol is designed to produce follicular phase levels of circulating E (12–15 pg/ml), and is sufficient to induce a LH surge in normal female lambs [22]. Based on the difference in LH surge dynamics in low prenatal E and controls during this prepubertal surge induction experiment, the same study was repeated after puberty during the subsequent non-breeding season; the sampling frequency was increased to one per hour.
The onset and maintenance of reproductive cycles for the first breeding season were compared in G2 low E females and controls by measuring progesterone in serum samples collected twice-weekly from 18 to 44 wk of age. This sampling period included all progestogenic cycles in both experimental groups during the first breeding season.
Circulating concentrations of LH were measured in duplicate 10–200-µl aliquots of serum using a previously validated RIA developed by Niswender et al. [29]. Assay sensitivity, defined as two SD from the buffer control, averaged 0.74 ± 0.08 ng/ml NIH LH-S12 for 200 µl of serum (n = 14 assays). The intra-assay coefficients of variation (CV) calculated from six replicates of three standard sera binding 29%, 50%, and 83% of buffer controls averaged 15.3%, 5.6%, and 5.5%, respectively. The interassay CV for the same standards were 19.5%, 7.4%, and 11.7%, respectively. Concentrations of progesterone were measured in duplicate 100-µl aliquots of serum using a commercially available RIA kit (Coat-A-Count P4; Diagnostic Products Corp., Los Angeles, CA). The validation of this assay for use in sheep has been described elsewhere [30]. The sensitivity of this assay averaged 0.02 ± 0.01 ng/ml for 100 µl of serum (n = 7 assays). The intra-assay CV using two quality control pools were 4.5% and 4.4%; the interassay CV were 5.4% and 9.4%.
For G1 females, the mean prepubertal and anestrus LH concentrations in twice-weekly samples, the ages at onset and offset of reproductive cycles, the numbers of reproductive cycles per breeding season, and the mean lengths of individual reproductive cycles were compared using ANOVA. Undetectable LH concentrations were assigned to the limit of assay sensitivity.
Differences in play behavior for G2 lambs were analyzed by multivariate analysis. In addition, the total number of play events was totaled across all behaviors and was compared between groups using ANOVA. The patterns of play partner and general association frequencies were compared using Chi square analysis. Group differences in displacement behavior at the feeder were analyzed by multivariate analysis of the number of initiations, number of wins, number of losses, and the total number of displacement interactions. Subsequently, we examined the group differences in percentage wins and losses using a Chi square analysis. Displacement at locations other than the feeder occurred too rarely to produce sufficient data for analysis. For the G2 control and low E females, the mean ages at onset and offset of reproductive cycles and the lengths of the reproductive cycles were compared using the Student t-test.
The onset of the LH surge induced in G2 females by acute estradiol treatment was identified using criteria described previously [31, 32]. Briefly, the surge onset was defined as: 1) an increase of circulating LH above baseline by two SD and lasting at least 8 h, and 2) peak concentrations of LH exceeding twice the average levels during pre-estradiol periods. For females that exhibited LH surge, the rate of increase from the surge onset to peak was also calculated. The resulting slopes were ranked and compared with the Mann-Whitney test. The time from E administration to the onset of the LH surge, the time elapsed between surge onset and peak, the time from E administration to surge peak, and the peak amplitude of the LH surge were analyzed in a multivariate analysis of variance model. P < 0.05 was considered to be statistically significant in all analyses.
Generation 1: Postnatal Estrogen
Female lambs treated chronically with low E exhibited prolonged prepubertal hypogonadotropism (from 10 to 23 wk of age) relative to ultralow E and control females (Fig. 1a). During that time, the mean (± SEM) LH concentrations in low-E females (1.4 ± 0.1 ng/ml) were lower than in ultralow E (3.0 ± 0.4 ng/ml) and control (2.3 ± 0.2 ng/ml) females (Fig. 1c). Based on the circulating concentrations of progesterone, there was a dose-dependent delay in the onset of reproductive cycles for both low and ultralow E groups relative to controls during the first breeding season (Fig. 2, top). Control animals initiated reproductive cycles at 26.8 ± 1.1 wk of age, while ultralow and low E animals began at 29.1 ± 0.3 and 31.3 ± 0.5 wk, respectively. Despite delaying puberty, chronic postnatal E had no effect on the numbers of reproductive cycles after puberty (control = 9.4 ± 0.7, ultralow = 8.8 ± 0.4, low = 8.0 ± 0.6) (Fig. 3), nor did it affect the ages at seasonal offset of reproductive cycles (control = 47.8 ± 1.0 wk, ultralow = 49.1 ± 0.9 wk, low = 49.0 ± 0.9 wk) or the lengths of the individual reproductive cycles (control 15.8 ± 0.4 days, ultralow E 16.5 ± 0.1 days, low E 16.0 ± 0.3 days) during the first breeding season.
|
|
|
During a 6-wk sampling period in the ensuing anestrus season, the circulating levels of LH were again suppressed in low E females (0.9 ± 0.1 ng/ml) compared to ultralow E (2.2 ± 0.2 ng/ml) and control females (1.9 ± 0.2 ng/ml) (Fig. 1d).
Unlike puberty, the onset of reproductive cycles during the second breeding season was delayed in low E females (81.6 ± 0.8 wk) compared to control (78.4 ± 0.6) and ultra low E (79.0 ± 0.4) females (Fig. 2, bottom). The controls and ultra low E groups did not differ. Postnatal E did not impair the abilities of E-treated females to conceive and carry a pregnancy during the second breeding season (Fig. 3). Pregnancies were confirmed by progesterone measured in samples collected twice-weekly after the first ovulatory cycle in the presence of an intact ram. Due to simultaneous breeding of all treatment groups and the delay in the onset of cycles in low E females, this group exhibited fewer cycles before pregnancy (6.4 ± 0.3) than the ultralow E (7.6 ± 0.2) or control (7.5 ± 0.3) group (Fig. 3). Since all of the females in each group became pregnant, the offset of cycles (seasonal anestrus) could not be assessed as it was during the pubertal year.
Generation 2: Prenatal Estrogen
Females born to mothers that received low-dose E throughout gestation and lactation exhibited altered competitive behavior with other female lambs, as measured by displacement at feeders between 8 and 14 wk of age (Fig. 4). The mean number of play behavior bouts and patterns of non-play association between 2–14 wk did not differ significantly among the groups. Data for male lambs are not presented because the males were only present for 2 wk at the start of the food displacement behavior observations (from 6–8 wk of age), and the data produced were insufficient for analysis. Prenatal low E females were involved in a higher mean number of total displacement interactions (59.8 ± 14.8) than control (8.7 ± 4.9) and prenatal ultralow E (13.4 ± 2.7) females. The initiation of aggressive displacement to reach hay or alfalfa pellets was also higher in prenatal low E females than in control and prenatal ultralow E females (31.4 ± 8.0, 4.9 ± 2.8, and 7.0 ± 1.6, respectively). Despite attempting to displace other females more frequently, prenatal low E females were not different from the other groups with respect to the percentage of successful displacements (wins) (50.9 ± 3.1%, 42.8 ± 14.4%, and 48.3 ± 9.1%, respectively). Because of the large difference in the number of total interactions between the prenatal low E group and the other two groups, most of the interactions used to determine the prenatal E win-loss ratio were between individuals within the prenatal low E group; that is, low E females preferentially engaged in displacements with other low E females. As a result, treatment group win-loss ratio is a poor indicator of displacement success for interactions between individuals from different groups.
|
Although the mean number of play bouts did not differ between the female treatment groups or their male sibs (between 2 and 8 wk when males were present), a treatment effect on the preferred play partner emerged. Male lambs spent 34% of their time playing with other male lambs, whereas females played less frequently with animals from the same treatment group (control females = 5.7%, ultralow E females = 13.5%, and low E females = 2.7%; Chi-square = 49.6, P < 0.001 compared with males; no difference among the female groups). In addition, low E females tended to play more with males than did other female treatment groups (low E = 48.6%, ultralow E and control = 35.9%; Chi-square = 3.304, P = 0.07). In the 6 wk after the males were removed, there were no treatment differences among the females in terms of number of play bouts or partner preference patterns.
Chronic exposure of the mother to E altered the overall pattern of the LH surge in response to acute treatment with exogenous E in G2 prepubertal female offspring (14 wk of age; Wilk Lambda F = 8.0, df = 4, 4, P = 0.04) (Fig. 5). Despite apparent alterations in the general pattern of LH release, none of the commonly measured LH surge parameters were independently significant (respective P values: surge onset, 0.05; time of peak, 0.06; peak concentration, 0.12; time from surge onset to peak, 0.12). All five control females exhibited LH surges, which were initiated at 14.8 ± 0.5 h (mean ± SEM) following the insertion of E implants. Peak LH levels were achieved at 18.8 ± 1.0 h following E implantation, with a mean peak amplitude of 29.5 ± 6.6 ng/ml LH and 4.0 ± 0.6 h elapsed between surge onset and peak. Four of the five prenatal low E females initiated LH surges at 16.0 ± 0.0 h after insertion of E implants, and reached a peak on average 4.2 h later than the controls at 23.0 ± 1.7 h, with 7.0 ± 1.7 h between surge onset and peak. The fifth prenatal low E female did not produce a LH surge and these data were excluded from the analysis. The mean (± SEM) peak surge amplitude was 15.7 ± 2.2 ng/ml LH, roughly half the mean value observed for the control females. In addition, the rate of increase from the initiation of the surge to its peak was markedly reduced in prenatal low E females (Fig. 6). The rates of increase (slopes) from surge onset to peak in prenatal low E-treated animals (1.3 to 2.6 ng LH/ml/h) were below the lowest extreme of the range in the control animals (3.3 to 8.7 ng LH/ml/h).
|
|
When the response of the LH surge mechanism was retested in the same G2 females after puberty (at approximately one year of age), the characteristics of the surge did not differ between groups (Wilk Lambda F = 0.5, df = 5, 4, P = 0.62). Prenatal low E females exhibited LH surge onset 10.4 ± 0.7 h after E implantation (5 out of 5 individuals), compared to 9.2 ± 0.2 h for the controls (5 out of 5). LH surge peak levels were attained at 14.0 ± 0.8 h after E insertion in prenatal low E females, and at 12.8 ± 0.5 h in controls, with mean (± SEM) peak amplitudes of 192.2 ± 17.4 ng/ml LH and 209.8 ± 34.7 ng/ml LH, respectively. The time elapsed between onset of the surge and its peak in prenatal low E females (3.6 ± 1.1 h) was also similar to that in control females (3.6 ± 0.4 h). In contrast to the prepubertal response of LH surge, there was considerable overlap in the ranges of rate of change in circulating LH from surge onset to peak between the groups (Fig. 6). The rate of increase in circulating LH from surge onset to peak ranged from 21.6 to 180.3 ng/ml/h in control females and from 35.2 to 62.3 ng/ml/h in prenatal low E females.
Chronically elevated prenatal E had no significant effect on the onset or offset of reproductive cycles during the first breeding season. Low E female lambs initiated progestogenic cycles at a mean age of 24.8 ± 0.8 wk, while the offset of cycles occurred at 34.9 ± 0.8 wk, both of which were similar to the values observed for the controls (onset, 22.3 ± 1.2 wk; offset, 36.8 ± 1.0 wk). The mean (± SEM) length of the reproductive cycles was also similar between prenatal low E females (2.3 ± 0.1 wk) and controls (2.3 ± 0.0 wk).
The results of the present study indicate that either chronic postnatal (G1) or chronic prenatal (G2) exposure to physiologically relevant doses of estradiol can alter reproductive function in female sheep. A delay in the age at onset of the first reproductive cycles for postnatal E-treated female lambs is consistent with our previous observation [16], though the magnitude of the delay was substantially smaller in the present study. Since the delay was on the order of several weeks, and due to variation in the age at offset, there was no significant difference in the average number of cycles during the first breeding season. The greater delay reported by Foster et al. [16] may be due in part to a slower growth rate compared to that of the females used in the present study. The diet used previously was restricted to pasture grazing, and it produced growth that was only marginally adequate to allow the attainment of puberty by the onset of the breeding season. In contrast, the females in the present study had unlimited access to a high energy food supply of alfalfa pellets and hay, and by their first breeding season their weights averaged more than twice that of the lambs used in our earlier study [16]. Such observations raise the possibility of increased susceptibility to steroid-induced reproductive dysfunction in individuals that experience suboptimal nutritional conditions. If this interaction of dietary constraints and exogenous chemical exposure is operative and applicable to other species, it could be of particular importance in human and wildlife populations that face limited food availability.
The potential for added estrogenic exposure arising from alfalfa being the primary food source in the present study was considered. Diets that consist largely of alfalfa have been shown to contain sufficient quantities of phytoestrogens, particularly coumestrol, to affect uterine weight [33] and sexual development [34] in the rat. In the present study, it is possible that, as a result of diet, the total exogenous estrogen load was slightly higher than that produced by the E implants alone. However the phytoestrogen content of alfalfa is considered to be very low when compared to other estrogenic forage species [13], and any dietary estrogen load would be similar in E-treated and control animals. At this time, the effects of combining a highly estrogenic food source with another prolonged low-dose exogenous steroid insult in a large, long-lived species remain to be studied.
In contrast to postnatal E females, puberty was not delayed in either prenatal ultra low E or prenatal low E females relative to controls. Reproductive function in prenatal E females, as reflected by the ability to produce repeated progestogenic cycles, also appeared to be normal. Behavior was not monitored in postnatal E females, but was clearly altered in the prenatal low E treatment group. The lack of more dramatic effects in prenatal low E females and the absence of any observed effect from prenatal ultralow E treatment could be attributable to the transplacental pharmacokinetics of estradiol, which likely resulted in prenatal individuals being subjected to even lower exposures than their mothers. In subhuman primates, only small fractions of exogenous estradiol added to the maternal circulation have been found in fetal plasma, predominately as the less-potent estradiol metabolites, estrone and estrone sulfate [35]. Due to the effects of placental metabolism, it is also possible that the E exposure of second-generation females was higher during suckling than gestation.
Based on the alterations in sexually dimorphic behavior, prenatal E could have affected central nervous system function. It is also possible that altered maternal behavior played a role in the behavioral differences noted in Generation 2 females. Although no behavioral monitoring of Generation 1 females occurred, studies in the rat have demonstrated that variations in maternal care can influence behavioral and physiological parameters in offspring [36].
One of the most notable behavioral differences arising from elevated prenatal and lactational E was increased displacement behavior. This type of change is typical of female sheep that attain a higher rank within a group [37, 38]. In social species, such as sheep, higher rank usually results in females gaining access to better or more food, and thereby allows the production of larger, healthier and perhaps more offspring, as has been documented in other ruminant species [39, 40] and primates [41]. However, there may be an overall trade-off, in that prenatal exposure resulting in mild masculinization may cause an early decrease in fertility, as evidenced by studies in the androgenized female rat [42].
The behavioral masculinization of the low E group was far from complete, as demonstrated by play behavior and favored association patterns. The low E females tended to play more often with males than did females from other treatment groups. This suggests that low E may have caused some masculinization of play, consistent with earlier studies showing that play behavior is influenced by prenatal androgens [26–28]. Association patterns during non-play and non-displacement periods did not differ between the treatment groups but were different between the sexes. Before weaning, females were found with siblings and their mother more often than with any other individuals, whereas males were more often with other males after 4 wk of age, as previously reported [38, 43].
The behavioral data suggest that masculinization of the low E group increased over time. From 2–8 wk, when play was most frequent, the low E group showed female play patterns, except for the trend toward higher rates of play with male lambs than occurred with the control and ultralow E female groups. Displacement behavior at the feeding sites increased between 8–14 wk after weaning occurred, and the males were removed. Low E females demonstrated much higher rates of displacement behavior than the other two groups of females. If males had remained in the group, the females might not have demonstrated as great an increase, as the males would likely have won more often at this behavior because of their greater size and growth rate. Interestingly, much of this displacement behavior occurred within the low E female group, as is typically seen for males that are beginning the process of establishing male hierarchy among their peers. The long-term effect of such mild masculinization on later behavior is currently being pursued.
In the present study, the competency of the pre-ovulatory LH surge system was assessed in prenatal low E G2 females before puberty, and a difference was noted relative to the controls in the rate of increase from the initiation of the LH surge to its peak, as well as the timing of the surge and the amount of LH secreted. The effects of prenatal E on the dynamics of the LH surge may reflect a decrease in slope of the GnRH surge or a delayed increase in GnRH receptors stemming from reduced pituitary sensitivity to estradiol and GnRH. Considering that a very small increase in GnRH is required for the generation of the surge [44], it is unlikely that a change in the dynamics of the GnRH surge could have contributed to this slope change. By one year of age, when the females were postpubertal, differences in the E-induced surge were no longer evident. The amplitude of the LH surge in prenatal low E females was markedly higher than when tested prior to puberty. However, this also occurred in controls and was consistent with previous findings [45] that the magnitude of the LH surge increases progressively during maturation. If the effects of prenatal E involve changes in the organization of the neural network that controls GnRH secretion, one would expect the defect noted prepubertally to persist or become exacerbated rather than be corrected with time. As abnormalities in the LH surge were not evident at a later age, it may be that the controls of the surge mechanism were not altered by prenatal E exposure, and the earlier defects were perhaps due to the activational effects of exogenous E. Previous studies have determined that upregulation of GnRH receptors and consequent increases in pituitary sensitivity to GnRH play a role in the generation of the surge [46, 47], and these mechanisms may have responded differently to the acute estradiol treatment of prenatal E females. If the reproductive neuroendocrine system of prenatal E females was further challenged during postnatal development by chronic exposure to E or a suboptimal plane of nutrition, the relatively minor alterations in the surge mechanism might ultimately disrupt ovulation and potentially result in lower fecundity and shorter reproductive lifespan.
Whether our prenatal exposure using the levels of E tested in the present study affects successful maintenance of reproductive cycles (including the LH surge system) at older ages remains unknown. The overall reproductive competencies of prenatal E-treated females beyond one year of age and the roles that further steroid exposure (either endogenous or exogenous) play were not studied. Similarly, the ways in which behaviors, such as the increased frequency of displacement in prenatal low E females, affect social interactions and reproductive functions in later life remain to be determined. However, if these results can be extrapolated to other species, delayed onset of reproductive cycles (and therefore, delayed conception) could alter the timing of birth in seasonally breeding species. This type of delay could have a pronounced effect on juvenile survival in wildlife populations, while, in agriculturally important species, these delays could potentially decrease productivity. Exposure to exogenous compounds with estrogenic activity, either at higher doses or for longer duration, is likely to produce more severe reproductive consequences. Nonetheless, the ability of female sheep to exhibit normal reproductive cycles and produce seemingly healthy offspring when chronically exposed to exogenous E offers some evidence that female reproductive function is relatively resilient to low-dose estrogen disruption. Additional studies of animal models for wild and domesticated species in real-life settings exposed to extraneous yet physiologically relevant levels of sex steroids are clearly warranted.
ACKNOWLEDGMENTS
We thank Mr. Douglas D. Doop for expert technical advice and assistance, Mr. Gary R. McCalla of the Sheep Research Core Facility for conscientious animal care, and the Reproductive Sciences Program Assays and Reagents Core Facility for standardization and providing hormone RIA reagents.
FOOTNOTES
1Supported by research grants from the NIH (HD-18394 and HD-44232). Preliminary reports presented in part at the 36th Annual Meeting of the Society for the Study of Reproduction, Cincinnati, OH, July 19–22, 2003 (Abstract 381); and the 37th Annual Meeting of the Society for the Study of Reproduction, Vancouver, British Columbia, Canada, August 1–4, 2004 (Abstract 47). 
Correspondence: 2 Douglas L. Foster, Room 1101, 300 North Ingalls Building, University of Michigan, Ann Arbor, Michigan 48109-0404. FAX: 734 936 8620; e-mail: dlfoster{at}umich.edu
Received: 17 April 2006.
First decision: 18 May 2006.
Accepted: 14 August 2006.
REFERENCES
and ß in neonatal rat testis: identification of gonocytes as targets of estrogen exposure. Biol Reprod 2003; 68:867–880
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
