Regulation of Baboon Fetal Ovarian Development by Placental Estrogen: Onset of Puberty Is Delayed in Offspring Deprived of Estrogen In Utero

ABSTRACT

Using the baboon as a model for studies of human reproductive biology, we previously showed that placental estrogen regulates fetal ovarian follicle development. In this study, offspring of baboons untreated or treated in utero with the aromatase inhibitor letrozole (estradiol reduced >95%) or letrozole and estradiol were reared to adulthood to determine whether estrogen programming of the fetal ovary impacted puberty and reproduction in adulthood. All offspring exhibited normal growth and blood pressure/chemistries. Puberty onset in untreated baboons (43.2 ± 1.4 mo) was delayed (P < 0.01) in animals of letrozole-treated mothers (49.0 ± 1.2 mo) and normal in offspring of mothers treated with letrozole and estradiol (42.7 ± 0.8 mo). During the first 2 yr postmenarche, menstrual cycles in estrogen-suppressed animals (43.2 ± 1.3 days) were longer (P < 0.05) than in untreated baboons (38.3 ± 0.5 days) or those treated with letrozole and estrogen (39.6 ± 0.8 days). Moreover, in estrogen-suppressed offspring, serum levels of estradiol were lower and follicle-stimulating hormone greater (P < 0.05) in the follicular and luteal phases, and the elevation in luteal-phase progesterone extended (P < 0.02). Thus, puberty onset was delayed and menstrual cycles prolonged and associated with altered serum hormone levels in baboon offspring that developed in an intrauterine environment in which estradiol levels were suppressed. Because puberty and follicle development, as shown previously, were normal in baboons treated in utero with letrozole and estradiol, we propose that fetal ovarian development and timely onset of puberty in the primate is programmed by fetal exposure to placental estrogen.

Fetal ovarian development and timely onset of puberty in the primate is programmed by fetal exposure to estrogen.

INTRODUCTION

Development of the pool of ovarian follicles for reproductive function in adulthood is established in utero [1], but few studies have addressed this aspect of reproductive biology. Using our nonhuman primate baboon model for studies of human reproduction [2, 3], we showed that placental estrogen, the levels of which increase with advancing gestation in humans and nonhuman primates [4], regulates fetal ovarian follicle development [5, 6]. Thus, in baboons in which estrogen was suppressed by treatment with the aromatase inhibitor letrozole during the second half of gestation, the number of primordial follicles formed in the fetal ovary was reduced by more than 50% [6], and the majority of follicles that developed contained unhealthy oocytes [7]. Additional studies showed that this critical action of estrogen most likely is elicited directly on the fetal ovary, which expresses estrogen receptor (ER) α/β [8, 9], and not indirectly via fetal pituitary gonadotrophin support, fetal ovarian follicle-stimulating hormone (FSH) receptor expression [10], or fetal pituitary prolactin secretion [11]. Collectively, these observations, plus the fact that fetal ovarian folliculogenesis and oocyte integrity were normal in baboons treated with letrozole and estradiol, support our hypothesis that placental estrogen programs fetal ovarian development and regulates formation of the stockpile of healthy follicles critical for long-term survival [5, 6] and presumably reproductive function in adulthood.

The importance of estrogen in programming primate fetal ovarian development is heightened by studies showing that polymorphisms of the ERα gene in women are associated with premature ovarian failure [12], and that reproductive function was disrupted in women born prematurely at 22–27 wk, and thus not exposed to the increasing levels of estrogen of advancing pregnancy [13]. Despite these important human epidemiologic reports, because of the difficulty in conducting in vivo studies during human pregnancy, almost nothing is known about the impact of alterations of the hormonal milieu (e.g., estrogen) on fetal ovarian development and reproductive function in adulthood. However, recent study has shown that the human fetal ovary during the period of follicle formation in the second trimester contains the machinery to detect and respond to estrogen. It has been proposed that estrogen may play an important role underpinning fetal primordial follicle formation in the human [14] and, as our laboratories have documented, in the baboon [5, 6]. Exposure of rodents and domestic animals to endocrine disruptors that interfere with estrogen action at the level of the receptor and/or signaling pathways in utero, or during the perinatal period, is associated with impairment of reproductive function in adulthood (for review see [15]).

It is well established that reproductive function in adulthood is preceded by the process of puberty, a transition period during which gonadotrophin-releasing hormone (GnRH) neurons in the hypothalamus become less sensitive to estradiol negative feedback, GnRH secretion and pulsatility are increased, and the pituitary-gonadal axis activated [16–19]. Moreover, it appears that kisspeptin neurons in the arcuate and periventricular regions of the hypothalamus express ERα, regulate GnRH secretion, and play a major role in regulating the process of puberty [20, 21]. Although the onset and progression of puberty in female mice were altered by ablation of ERα in kisspeptin neurons [22], administration of an aromatase inhibitor during the first half of ovine pregnancy had no effect upon the timing of the onset of puberty in offspring [23]. However, maturation of the neuroendocrine-ovarian axis in rodents occurs after birth, suggesting that the regulatory mechanisms (e.g., the role of the hormones of pregnancy, per se [24, 25]) are unlike those in the human [14], while endocrine interactions between the mother, placenta, and fetus in large domestic animals differ from that in primates [26], thereby limiting translation of findings to the human. Therefore, in the current study, we employed the baboon as a nonhuman primate translational model in which offspring of mothers untreated or treated in utero with letrozole or letrozole and estradiol were reared to adulthood to test the hypothesis that the timing of the onset of puberty and normal reproductive function in adulthood is programmed by fetal exposure to placental estrogen in utero.

MATERIALS AND METHODS

Animals

Female baboons (Papio anubis) weighing 10–15 kg were housed individually in stainless steel cages in air-conditioned quarters and fed Purina monkey chow (Ralston Purina, St. Louis, MO), fresh fruit and vegetables, and water ad libitum. Females were paired with males for 5 days at the anticipated time of ovulation, and pregnancy was subsequently confirmed by palpation and ultrasonography. Fetal sex was determined by chromosomal analysis of fetal cells in amniotic fluid at Day 80 of gestation (term = Day 184). Pregnant baboons with a female fetus were either untreated (n = 9) or treated (n = 5) with a highly specific aromatase inhibitor, letrozole (4,4-[1,2,3-triazol-1yl-methylene] bis-benzonitrite; Novartis Pharma AG, Basel, Switzerland), 115 μg/kg body weight (BW)/day, administered s.c. on Days 100–175, as described previously [27]. Additional animals were injected with letrozole (115 μg/kg BW/day) plus estradiol benzoate (beginning at 50 μg/kg BW/day and increasing to a maximum of 150 μg/kg BW/day) on Days 100–175 (n = 4) or Days 60–175 (n = 2) of gestation. Blood samples (2–3 ml) were obtained at 1- to 4-day intervals between Days 85 and delivery via a maternal saphenous vein after sedation with an i.m. injection of ketamine-HCl (10 mg/kg BW; Parke-Davis, Detroit, MI). Thirteen of the baboons delivered spontaneously: 7 were untreated (mean ± SE day of delivery, 180 ± 6), 1 treated with letrozole (Day 169), and 5 treated with letrozole and estradiol (Day 170 ± 5), and all of these neonates were raised by the mother. The remaining seven animals were delivered by Cesarean section on Days 173 and 175 (untreated; n = 2), Day 171 ± 4 (letrozole-treated; n = 4), and Day 174 (letrozole and estradiol-treated; n = 1), as described previously [27]. Briefly, baboons were sedated with ketamine, anesthetized with isoflurane, and, after obtaining maternal and umbilical blood samples, the placenta and the fetus were delivered. After clearing the pharyngeal cavity of mucous, neonates were kept warm and attended to, but with no further treatment, and three of these neonates returned to the mother within 3–4 h after recovery from anesthesia. Four neonates (untreated, n = 1; letrozole-treated, n = 3) were placed in a heated baby isolette and fed Similac and water ad libitum by staff at 3- to 5-h intervals. By age 3 mo, these neonates were housed in a small stainless steel cage, placed in a room with adult baboons, and fed liquid from a suspended water bottle and Purina Primate Chow soaked in water.

At 8–12 mo of age, all baby baboons were weaned and placed in a specially constructed large primate cage in groups of 4–6, which permitted animal interaction/socialization and free roaming. At approximately 2 yr of age, animals were pair-housed in stainless-steel cages. At monthly intervals between 12 and 30 mo of age, animals were anesthetized with ketamine (10 mg/kg BW), weighed, and a saphenous vein blood sample (<1–2 ml) obtained for subsequent analysis of serum estradiol and for electrolytes and blood chemistries using iSTAT system (Abbott Laboratories, Abbott Park, IL).

Between ages 2 and 4 yr, blood pressure was also determined in representative ketamine-sedated animals from each study group. At 30 mo of age and in anticipation of puberty onset, the blood sampling protocol for analysis of serum steroids (e.g., estradiol) was implemented every 2 wk. All baboons were cared for and used strictly in accordance with U.S. Department of Agriculture regulations and the National Institutes of Health Guide for the Care and Use of Laboratory Animals (publication 85-23, revised 1996). The Institutional Animal Care and Use Committees of Eastern Virginia Medical School and the University of Maryland School of Medicine approved the experimental protocol employed in this study.

Onset of Puberty

Baboon offspring were observed daily for onset of puberty [28, 29], defined as a change in perineal sex skin appearance, color, and tumescence (swelling), indicative of increase in and/or response to estradiol [30]. Onset of menarche was assessed by subsequent appearance of menstrual bleeding and completion of puberty by onset of regular menstrual cycles, as determined by daily assessment of perineal swelling/detumescence and detection of vaginal blood. In addition, for each baboon, the length of 15 consecutive menstrual cycles expressed as the number of days from completion of menses to onset of next mense, was determined over the next 18–24 mo. Moreover, in three untreated offspring and four treated in utero with letrozole, blood samples (1–2 ml) were obtained daily during one or two of these menstrual cycles for determination of serum levels of estradiol, progesterone, LH, and FSH.

Radioimmunoassays

Serum samples were stored at −20°C until measurement of estradiol, testosterone, and progesterone by radioimmunoassay using an automated chemiluminescent immunoassay system (Immulite; Siemens Healthcare Diagnostics, Deerfield, IL), as described previously for estradiol [27]. The Immulite assays for estradiol, testosterone, and progesterone were previously shown to be highly specific and exhibited minimal (<1%) cross-reactivity with potential competing steroid hormones, estrone, cortisol, cortisone, DHAS, androstenedione, and corticosterone [7, 27, 31]. The intra-assay and interassay coefficients of variation and limit of detection were 6.9%, 7.3%, and 15 pg/ml, respectively, for estradiol, and 7.6%, 7.9%, and 90 pg/ml, respectively, for progesterone [27]. Serum LH and FSH levels were measured by RIA using antibodies/reagents and procedures established for the baboon, as described previously [10, 32]. The baboon FSH and LH RIA had intra-assay coefficients of variation of 5.4% and 6.2%, respectively, interassay coefficients of variation of 6.9% and 4.5%, respectively, and a minimal detectable dose of 0.009 ng/tube [10, 32].

Statistics

Data are expressed as an overall mean (± SEM) and were analyzed using Student t-tests for independent observations, linear regression, repeated measures ANOVA, or one-way ANOVA with post hoc comparisons of the means by the Student-Newman-Keuls multiple comparison tests or by the Kruskal-Wallis nonparametric ANOVA and the Dunn multiple comparison tests in instances where the Bartlett statistic indicated that standard deviations between groups were significantly different.

RESULTS

Maternal Serum Estradiol Concentrations

In untreated baboons, maternal peripheral serum estradiol levels rose from approximately 1.5 ng/ml on Days 85–120 of gestation to approximately 3.0 ng/ml by Day 175 (Fig. 1). Within 48–72 h of the onset of letrozole treatment on Day 100, maternal serum estradiol levels decreased to and remained at approximately 0.1–0.2 ng/ml. The pattern of serum estradiol in baboons treated with letrozole plus estradiol benzoate was similar to that in controls; however, levels were slightly, but not significantly, greater than normal.

FIG. 1.

Mean maternal peripheral serum estradiol levels between Days 85 and 175 of gestation in baboons with a female fetus that were untreated or treated with letrozole (115 μg/kg BW/day; s.c.) or letrozole and estradiol benzoate (115 μg/kg BW/day and 50 μg increasing to 150 μg/kg BW/day, respectively; s.c.) on Days 100–175 of gestation (term = Day 184).

Umbilical Serum Estradiol, Testosterone, and FSH

Because the majority of the animals in the current study delivered spontaneously, levels of hormones in these female fetuses at birth could not be ascertained. However, serum concentrations of estradiol, testosterone, and FSH in umbilical vein/artery measured in a contemporaneous cohort of baboons carrying female fetuses and studied as part of other experiments on the effects of estrogen on placental/fetal development are shown in Table 1. Levels of estradiol in the umbilical vein (n = 4–11/group) were increased between Day 100 (mean ± SE, 0.12 ± 0.02 ng/ml) and Days 162–175 of gestation (0.68 ± 0.26 ng/ml), reduced by 95% (P < 0.05) on Day 160–165 by treatment with letrozole (0.04 ± 0.01 ng/ml), and partially restored by letrozole and estradiol (0.14 ± 0.06 ng/ml). The restoration of maternal, but not fetal, serum estradiol in baboons treated with letrozole and estrogen presumably reflects the fact that placental estradiol is preferentially secreted into the maternal circulation during primate pregnancy [33], and that estradiol benzoate was injected into the mother.

TABLE 1.

Serum levels (ng/ml) of estradiol (E₂) and testosterone in umbilical vein and of FSH in umbilical artery at mid and late gestation in a contemporaneous group of baboons untreated or treated with letrozole with and without E₂ benzoate.*

^* Baboons with a female fetus untreated or treated daily on Days 100 (mid) to 175 (late) of gestation with letrozole or letrozole plus estradiol benzoate, as described in the legend for Figure 1.

^† From Zachos et al. [10].

^‡ P < 0.05 versus untreated late gestation (Kruskal-Wallis nonparametric ANOVA and Dunn multiple comparison tests).

^a,b Values with different letter superscripts differ at P < 0.05 (ANOVA and multiple comparison of means using Student-Newman-Keuls statistic).

In contrast, compared with untreated baboons, serum testosterone concentrations on Days 160–175 of gestation were 4-fold higher (P < 0.05) in the umbilical vein of animals treated with letrozole, and remained elevated in baboons that received letrozole and estrogen, because aromatization remained blocked by letrozole. As previously reported in another cohort of baboons [10], fetal (umbilical artery) serum levels of FSH were greater at mid than at late gestation, remained elevated in late gestation in baboons treated with letrozole, and were restored to normal in baboons treated with letrozole and estradiol.

Fetal Body and Organ Weights

Despite alterations in estradiol and fetal FSH levels, fetal ovarian wet weight, as well as fetal body weight and weight of the pituitary, liver, kidney, and other organs, were not altered in the contemporaneous group of baboon fetuses treated with letrozole with and without estradiol (Table 2).

TABLE 2.

Fetal body and organ weights in a contemporaneous group of baboons untreated or treated with letrozole with and without estradiol (E₂) benzoate.*

^* Baboons with a female fetus untreated or treated daily on Days 100–175 of gestation with letrozole or letrozole plus estradiol benzoate.

Growth and Development of Offspring

All offspring, untreated and treated in utero with letrozole with and without estradiol, grew normally throughout Years 1–4 of postnatal life, as assessed by an almost linear increase in body weight (Fig. 2), and exhibited normal arterial blood pressure and basal fasting serum chemistry analytes and glucose, as well as blood hematocrit, hemoglobin, pH, and gas levels (Table 3). In addition, parameters in Table 3 appeared to be similar in baboons delivered spontaneously or by Cesarean section and nurtured by human staff or baboon mothers (data not shown).

FIG. 2.

Mean (± SE) BW at postnatal ages 1–4 yr and at onset of puberty of female baboon offspring born to mothers untreated (n = 9) or treated in utero with letrozole (n = 5) or letrozole plus estradiol benzoate (n = 6), as described in legend for Figure 1.

TABLE 3.

Mean arterial blood pressure (MABP) and blood chemistry analytes in female baboon offspring untreated or treated in utero with letrozole with and without estradiol (E₂) benzoate.*

^* Values represent means (± SE) obtained at monthly intervals between 12 and 30 mo of postnatal age.

^†BUN, blood urea nitrogen; HB, hemoglobin; Hct, hematocrit; pCO₂, partial pressure of carbon-dioxide; TCO₂, total carbon dioxide.

Onset of Puberty

Onset of puberty in young baboons born to untreated mothers occurred at 43.2 ± 1.4 mo of age (mean ± SE; Fig. 3). In the two untreated baboon offspring delivered by Cesarean section, puberty onset (36.3 and 44.6 mo) was similar to that in animals delivered vaginally (range, 38.6–49.9 mo). Importantly, puberty onset in baboons born to mothers treated in utero with letrozole was significantly delayed (49.0 ± 1.2 mo of age; P < 0.01) by approximately 6 mo. The single animal in this group that was delivered spontaneously exhibited onset of puberty at 46.2 mo. Moreover, onset of puberty in offspring born to mothers treated with letrozole and estradiol (42.7 ± 0.8 mo) was similar to that in untreated animals, and thus also different (P < 0.05) from that in baboons treated only with letrozole. Onset of menarche in untreated baboons (45.4 ± 1.4 mo) was also delayed (P = 0.06) in letrozole-treated baboons (49.8 ± 1.1 mo) and normal in animals treated with letrozole and estradiol (45.1 ± 0.9 mo). As shown in Figure 2, the difference in onset of puberty in estrogen-suppressed animals was not due to a significant difference in body weight. Thus, body weight at onset of puberty in untreated baboons (9.5 ± 0.3 kg) was slightly, but not significantly, higher in animals treated in utero with letrozole (10.8 ± 0.6 kg) or letrozole plus estradiol (10.6 ± 0.5 kg).

FIG. 3.

Age (mo) at onset of puberty in female baboons born to mothers untreated (n = 9) or treated in utero with letrozole (n = 5) or letrozole plus estradiol benzoate (n = 6). Black circles = baboons that were delivered spontaneously; gray circles = baboons that were delivered by Cesarean section. *Mean (± SE) age of puberty onset in letrozole-treated animals differs from that in animals untreated (P < 0.01) or treated with letrozole plus estrogen (P < 0.05; ANOVA and Student-Newman-Keuls multiple statistic). Bars = 95% confidence limits for the mean age to onset of puberty.

Serum Hormone Levels Before and after Puberty Onset

The absolute levels and patterns of serum estradiol before and after onset of puberty were similar (repeated-measures ANOVA) in offspring born to mothers untreated or treated with letrozole or letrozole plus estradiol. Therefore, values were combined and overall mean values presented in Figure 4. In baboon offspring, serum estradiol levels prior to postnatal age 30 mo were negligible and often below the limit of assay sensitivity (i.e., 15 pg/ml; Fig. 4). At approximately 4 wk prior to the onset of puberty, serum estradiol levels began to increase, after which levels increased from 35 to >40 pg/ml 8–12 wk postpuberty, after which time menstrual cycles had been initiated. Thus, the overall mean level of serum estradiol for the period 4 wk before through 6 wk after puberty onset exceeded the overall respective mean value for the preceding 12-wk period (i.e., 18–6 wk before onset of puberty) in all offspring (41.3 ± 2.9 pg/ml vs. 24.6 ± 0.7 pg/ml, respectively; P < 0.05; Fig. 4), as well as in the individual treatment groups: untreated offspring, 34.4 ± 3.0 pg/ml versus 24.8 ± 1.1 pg/ml; treated with letrozole, 49.6 ± 2.8 pg/ml versus 25.8 ± 1.7 pg/ml; or treated with letrozole plus estradiol, 46.1 ± 8.8 pg/ml versus 23.2 ± 1.6 pg/ml (data not shown).

FIG. 4.

Mean (± SE) serum estradiol levels in blood samples obtained at 2-wk intervals before and after onset of puberty (normalized to Week 0) in baboon offspring untreated or treated in utero with letrozole with and without estradiol. Because the absolute levels and patterns of estradiol were similar in the three treatment groups (repeated-measures ANOVA), values were combined and overall mean values presented. *The overall mean level of serum estradiol for the period 4 wk before through 6 wk after puberty onset exceeded the overall respective mean value for the preceding 12-wk period (P < 0.05; unpaired t-test with Welch correction).

Characterization of Menstrual Cycles after Menarche

During the first 2 yr postmenarche, untreated baboons exhibited fairly consistent menstrual cycles lasting approximately 35–42 days and which, overall, averaged 38.3 ± 0.5 days (Fig. 5). In contrast, in letrozole-treated baboons, cycle lengths were more variable (range, 36–52 days) and overall mean length of 43.2 ± 1.3 days exceeded that in untreated animals (P < 0.05). In baboon offspring treated in utero with letrozole and estrogen, menstrual cycle length (range, 34–48 days; mean, 39.6 ± 0.8 days) was similar to that in untreated animals.

FIG. 5.

Overall mean (± SE) length (days) of menstrual cycles in baboon offspring untreated (n = 9) or treated in utero with letrozole (n = 5) or letrozole plus estradiol benzoate (n = 5). The length of 15 consecutive menstrual cycles was determined over the next 18–24 mo after onset of menarche in each baboon. The mean length (days from completion of menses to onset of next mense) of each of these 15 cycles was ascertained, and an overall mean cycle length calculated. Values with different lowercase letters differ at P < 0.05 (Kruskal-Wallis nonparametric ANOVA and the Dunn multiple comparison test).

Serum estrogen, progesterone, FSH, and LH levels were assessed after the onset of puberty in five menstrual cycles of three untreated baboon offspring, and in seven menstrual cycles of four baboon offspring treated prenatally with letrozole, and the mean values are shown in Figure 6. In letrozole-treated baboons, overall mean serum estradiol levels (Fig. 6A) on Days −20 to −11 (relative to estradiol surge at Day 0) of the follicular phase (32 ± 3 pg/ml) and Days 3–14 of the luteal phase (55 ± 4 pg/ml) were 47% and 25% lower, respectively, than in untreated animals (60 ± 2 pg/ml; and 72 ± 4 pg/ml; P < 0.05). In contrast, although serum levels of FSH declined during the follicular phase in baboons untreated and deprived of estrogen in utero (P < 0.001; linear regression; Fig. 6C), the mean level of FSH in letrozole-treated animals on Days −20 to −5 of the follicular phase (2.43 ± 0.13 ng/ml) and Days 5–12 of the luteal phase (2.12 ± 0.13) exceeded that in untreated animals (1.71 ± 0.14 and 1.46 ± 0.13 ng/ml, respectively; P < 0.05). Moreover, in estrogen-suppressed animals, serum progesterone levels (Fig. 6B) were maintained longer (P < 0.02) and at a higher level (P < 0.05) on Days 15–17 of the luteal phase than in untreated animals. However, the absolute level of the midcycle LH peak, as well as mean daily serum LH levels in the follicular and luteal phases, were similar in untreated and letrozole-treated animals (Fig. 6D).

FIG. 6.

Mean (± SE) serum levels of estradiol [E₂] (A), progesterone (B), FSH (C), and LH (D) during the menstrual cycles, as described in the legend for Figure 5, of baboons born to mothers untreated (n = 3 baboons; 5 menstrual cycles) or treated in utero with letrozole (n = 4 animals; 7 menstrual cycles). In letrozole-treated baboons, overall mean serum E₂ levels on Days −20 to −11 (i.e., relative to E₂ surge at Day 0) of the follicular phase (i.e., prior to onset of increase in estrogen production) and Days 3 to 14 of the luteal phase were lower than in untreated animals (P < 0.05; Student t-test). In contrast, serum FSH levels were greater on Days −20 to −5 of the follicular phase and Days 5–12 of the luteal phase than in untreated baboons (P < 0.05). Serum progesterone levels in letrozole-treated animals on Days 15–17 of the luteal phase (i.e., during period of luteal regression in untreated baboons) were greater than in untreated baboons (P < 0.05; Student t-test).

DISCUSSION

The current study is the first to show that the onset of puberty and menarche was delayed in baboon offspring that developed in an intrauterine environment in which levels of estradiol were suppressed throughout the second half of gestation by treatment with the aromatase inhibitor letrozole. Because puberty onset and menarche were restored to normal in baboons treated with letrozole and estradiol, we propose that estrogen in utero programs the fetus for the timely onset of puberty, and thus reproductive function in adulthood in the primate.

The current study also shows that, once menstrual cyclicity had been attained in baboon offspring deprived of estrogen in utero, menstrual cycle length was increased and the profiles of steroid and gonadotrophin hormone secretion were significantly different from those in baboons exposed prenatally to the normal elevation in estrogen. Because we previously showed that the pool of follicles that developed in utero in estrogen-suppressed baboon fetuses was 50% lower than in untreated baboons [5–7], we propose that the subnormal levels of estradiol and elevations in FSH during the follicular and luteal phase in offspring of letrozole-treated mothers were the consequence of the reduction in the stockpile of healthy follicles. Moreover, although LH levels and length of the luteal phase were not different, the elevation in serum progesterone levels typical of the luteal phase was sustained for approximately 2 days longer in offspring of letrozole-treated mothers. These findings provide further evidence of the disruption in ovarian function in baboon offspring deprived in utero of estrogen, and suggest that luteal development, function, and/or demise had been altered. Finally, it is possible that baboon offspring deprived of estrogen during pregnancy may be at risk for early depletion of the follicle pool (i.e., early menopause); thus the syndrome of premature menopause may have its origin in utero, and reflect availability of estrogen. Indeed, polymorphisms of the ERα gene in women are associated with premature ovarian failure [12].

A decline in the follicle stockpile and estrogen may also have impacted puberty onset, because there were fewer follicles available and/or have disrupted ovarian responsivity to signals from the hypothalamic-pituitary axis and/or the physiologic set point of the hypothalamic-pituitary-ovarian axis via estrogen feedback on the brain/pituitary. Indeed, the decline in pituitary FSH expression and serum FSH levels that occurs with normal advancing baboon [10] and human [34] gestation was prevented in estrogen-suppressed fetal baboons. This clearly establishes a role for estrogen negative feedback on the neural circuits controlling GnRH/gonadotrophin release during fetal development, as originally proposed by Kaplan and Grumbach [34]. It is also possible that estrogen may be acting to control development of the transsynaptic and neural-glial cell populations/communication and production of critical factors (e.g., kisspeptin, excitatory amino acids, and growth factors) that are important to activation of GnRH neurons controlling the initiation of puberty [35–38]. Indeed, based on studies in mice with a specific ablation of ERα in kisspeptin neurons, Mayer et al. [22] have proposed that puberty onset and progression reflect the regulation by estradiol of the temporal expression of kisspeptin in two neuronal populations. While the mechanisms by which estradiol regulates puberty onset in the primate remain to be determined, we suggest that the letrozole-treated pregnant baboon provides an excellent in vivo model to explore these important possibilities.

Although excessive androgen exposure in early ovine pregnancy impairs ovarian function in adulthood [39, 40], the present study also shows that the delay in the onset of puberty in estrogen-suppressed baboons was not due to exposure of the fetus in utero to an increase in androgen levels secondary to inhibition of aromatization by letrozole treatment. Thus, fetal testosterone levels were equally elevated in utero in both letrozole- and letrozole plus estradiol-treated baboons, but only the animals concomitantly treated with letrozole plus estrogen achieved puberty onset comparable to that in untreated estrogen-replete offspring. Although the level of estradiol in letrozole and estradiol-treated baboons was not restored to normal, this level of estradiol was sufficient to overcome the adverse effects of estrogen deprivation on the timely onset of puberty, as well as on fetal ovarian follicle development [6] and pituitary FSH secretion/expression [10, 11]. It is unlikely, but remains to be established, that the adverse effect of estrogen suppression on puberty onset are due to an alteration in estrogen-dependent events in the mother that could affect the fetus. Consistent with this suggestion, we have shown that maternal weight and uteroplacental and fetal blood flow dynamics and distribution were not altered by letrozole administration in baboons [41].

Although puberty onset is influenced by weight/growth patterns, as shown in humans [42, 43] and animal models (for review see [28]), the delay in onset of puberty in baboon offspring that occurred in the absence of estrogen was not associated with a change in rate of growth or body weight. Moreover, puberty onset in baboons was not preceded by any postnatal change in adrenal cortisol levels (Pepe, Lynch, and Albrecht, unpublished results), and was thus not influenced by differences in ability of adolescents to respond to potential stressors.

Finally, our study showing that puberty onset in the primate is programmed by fetal exposure to estrogen in utero takes on heightened significance and relevance to human fetal development, considering the increasing incidence of premature birth in the United States [44], which would deprive the developing fetus of the elevated levels of estrogen typical of advancing gestation. Indeed, a recent epidemiologic study showed that reproduction function in women born prematurely at 22–27 wk was lower than in women born at term [13]. Moreover, exposure to endocrine disruptors, which interfere with estrogen action at the level of the receptor and/or signaling pathways in utero and/or during the perinatal period, confers greater risk for impairment of reproductive function in adulthood (for review see [15]). In addition, recent study has shown that the human fetal ovary expresses the metabolic machinery to respond to estrogen during the second half of gestation, when the critical process of primordial follicle formation is occurring [14]. Finally, although numerous experimental studies of the intrauterine hormonal milieu have focused on the role of undernutrition on puberty onset and reproductive function in adulthood in rodents and sheep [45–47], our findings clearly document the important physiologic role played in pregnancy by a naturally produced endogenous hormone in programming fetal reproductive capacity in adulthood.

In summary, the results of the current study are the first to establish that the onset of puberty and menarche was delayed, length of menstrual cycle increased, and hormone profiles exhibited during the menstrual cycle altered in baboon offspring that developed in an intrauterine environment in which levels of estradiol were suppressed during the second half of gestation. Because the onset of puberty/menarche and follicle development were restored to normal in baboons treated in utero with letrozole and estrogen, we propose that fetal ovarian development and timely onset of puberty in the primate is programmed by fetal exposure to placental estrogen in utero. The likely origin of the delay in puberty onset and alteration in hormonogenesis during the menstrual cycle is the impaired formation of the pool of healthy follicles in the fetal ovary of estrogen-deprived baboons.