MS and PAH conceived and designed the experiments. MS performed the experiments. MS, TJH, and PAH analyzed the data. TJH and PAH contributed reagents/materials/analysis tools. EF assisted in some methods used in studies. MS, TJH, and PAH wrote the paper.
The authors have declared that no competing interests exist.
Estrogen plays an essential role in the growth and maturation of the mammalian oocyte, and recent studies suggest that it also influences follicle formation in the neonatal ovary. In the course of studies designed to assess the effect of the estrogenic chemical bisphenol A (BPA) on mammalian oogenesis, we uncovered an estrogenic effect at an even earlier stage of oocyte development—at the onset of meiosis in the fetal ovary. Pregnant mice were treated with low, environmentally relevant doses of BPA during mid-gestation to assess the effect of BPA on the developing ovary. Oocytes from exposed female fetuses displayed gross aberrations in meiotic prophase, including synaptic defects and increased levels of recombination. In the mature female, these aberrations were translated into an increase in aneuploid eggs and embryos. Surprisingly, we observed the same constellation of meiotic defects in fetal ovaries of mice homozygous for a targeted disruption of ERβ, one of the two known estrogen receptors. This, coupled with the finding that BPA exposure elicited no additional effects in ERβ null females, suggests that BPA exerts its effect on the early oocyte by interfering with the actions of ERβ. Together, our results show that BPA can influence early meiotic events and, importantly, indicate that the oocyte itself may be directly responsive to estrogen during early oogenesis. This raises concern that brief exposures during fetal development to substances that mimic or antagonize the effects of estrogen may adversely influence oocyte development in the exposed female fetus.
The potential effects on reproduction of chemicals with hormone-like activity is a growing concern. One estrogenic chemical, bisphenol A (BPA), has received considerable attention because low-dose exposures have been reported to induce a variety of reproductive effects in rodents. In the course of studies to assess the effects of BPA on the mouse oocyte, we have uncovered a novel “grandmaternal” effect: exposure to BPA during pregnancy disturbs oocyte development in unborn female fetuses. When these fetuses reach adulthood, the perturbations are translated into an increase in chromosomally abnormal eggs and embryos. Thus, low-dose BPA exposure during pregnancy has multigenerational consequences; it increases the likelihood of chromosomally abnormal grandchildren. Our studies also provide mechanistic insight, and, surprisingly, suggest that BPA acts in the fetal ovary not by mimicking the actions of estrogen but by interfering with the function of one of the known estrogen receptors. Thus, our data suggest that estrogen plays a far earlier role in oocyte development than previously suspected and, importantly, raise the possibility that a variety of substances—both synthetic and naturally occurring—that mimic the actions of estrogen or act as estrogen antagonists may affect early oocyte development.
The link between exposure to synthetic chemicals that mimic the actions of endogenous hormones and risks to human health is a growing concern. As early as 1970, Herbst and Scully reported vaginal clear-cell adenocarcinoma in six 14- to 21-y-old women exposed in utero to the synthetic estrogenic drug diethylstilbestrol (DES) [
Bisphenol A (BPA) was formulated around the same time as DES, but, because it was considered a less potent estrogen, it was never used clinically. We are, however, exposed to BPA daily; it is a component of polycarbonate plastics, resins lining food/beverage containers, and additives in a variety of consumer products. Over 6 billion pounds are produced worldwide annually, and several studies have reported levels of BPA in human tissues in the parts per billion range [
Short-term exposure to environmentally relevant doses of BPA has been linked to a variety of reproductive effects in laboratory rodents, including reduced sperm production, alterations in prostate development, and increased susceptibility to prostate carcinogenesis in the male [
Our laboratory is interested in the possible effects of BPA on the genetic quality of gametes. Low-dose BPA exposure in vivo during the final stages of oocyte growth [
During fetal development, germ cells in both sexes undergo massive mitotic proliferation. Subsequently, germ cells in the testis enter mitotic arrest and remain quiescent until after birth, while those in the ovary initiate meiosis. The prophase events of female meiosis (i.e., pairing, synapsis, and recombination between homologous chromosomes) occur during fetal development. By the time of birth, oocytes have entered a protracted period of meiotic arrest, where they remain until just prior to ovulation. Resumption and completion of the first meiotic division occurs only after an extensive period of follicle growth in the adult ovary, and this occurs weeks, years, or even decades (depending on the species) after the initiation of meiotic arrest.
To assess the effect of BPA exposure during the fetal stages of oogenesis, we implanted time-release BPA pellets (designed to leach a low, environmentally relevant dose of 20 μg/kg body weight/day, as used in our previous studies of BPA exposure in young adult females [
To analyze meiotic prophase, ovaries were isolated from female fetuses at 18.5 d of gestation and meiotic preparations made as described previously [
We focused subsequent analyses on pachytene oocytes, the stage at which synapsis between homologous chromosomes is complete and the sites of exchange between homologs become detectable as MLH1-positive foci (reviewed in [
(A) Frequency of synaptic abnormalities in 402 pachytene cells from ten placebo mice and 648 cells from ten BPA-exposed females.
(B) Pachytene oocyte from placebo-exposed female immunolabeled with SCP3 (red) and MLH1 (green) and showing normal synapsis.
(C and D) Pachytene oocytes from BPA-exposed females showing incomplete synapsis of a single pair of chromosomes (arrow) (C) and abnormal end-to-end associations involving multiple SCs (arrows) (D).
Pachytene oocytes from exposed females also displayed striking aberrations in recombination, as assessed by the number and distribution of MLH1 foci along the SCs. A total of 124 cells were analyzed from nine placebo-exposed females, with a pooled mean of 26.0 ± 2.3 foci per cell; MLH1 foci counts were significantly elevated in BPA-exposed females, with a mean of 29.0 ± 3.7 foci per cell from the analysis of 155 cells from ten females (
Recombination is regulated by crossover interference, a mechanism that ensures at least one exchange per chromosome pair and controls the proximity of multiple exchanges on a given chromosome (reviewed in [
Distribution of MLH1 Foci in Pachytene Cells from Placebo and BPA-Exposed Females
(A) Distribution of MLH1 foci. Proportion of SCs with zero, one, two, or three MLH1 foci (exchanges) in pachytene cells from placebo and BPA-exposed females.
(B–E) Air-dried chromosome preparations from BPA-exposed females. (B) Diakinesis cell containing a bivalent with three chiasmata (arrow). (C–D) Hyperploid metaphase II eggs with (C) 21 chromosomes and (D) 20 chromosomes plus one prematurely separated sister chromatid (arrow). (E) Hyperploid blastomere with 41 chromosomes from two-cell embryo.
Previous studies have demonstrated that MLH1 foci at pachytene accurately reflect the sites of exchange [
Analysis of Metaphase I Oocytes from Placebo and BPA-Exposed Mice
Analysis of Chiasmata Distribution at Metaphase I in Placebo and BPA-Exposed Females
In humans, aberrations in recombination are associated with meiotic nondisjunction. Both differences in the number of exchanges and their placement along the length of the chromosome (i.e., too close to the centromere or too close to the telomere) have been reported to play a role in the genesis of human trisomy (reviewed in [
To assess meiotic nondisjunction, ten pregnant females implanted with placebo and 16 implanted with BPA pellets were allowed to go to term and their offspring (17 placebo and 24 BPA-exposed females) were fostered at birth as described above. At 4–5 wk of age, these females were used either as oocyte or embryo donors for analysis of air-dried chromosome preparations from metaphase II–arrested eggs and two-cell embryos, respectively.
Because analysis of eggs is limited to a single cell, aneuploidy levels are usually estimated by doubling the frequency of hyperploidy to avoid artifacts introduced by chromosome loss. Typical aneuploidy levels for eggs in the laboratory mouse are between 0.5% and 1.0% [
Aneuploidy Analysis
To assess aneuploidy in embryos from exposed females, we superovulated 4- to 5-wk-old females, mated them with wild-type males, and analyzed two-cell embryos. The level of hyperploidy in embryos closely matched the level in eggs (0/13 or 0.0% in placebo versus 4/19 or 21.1% in BPA-exposed;
BPA is considered a “weak” estrogen due to its low binding affinity for the known estrogen receptors; however, its ability to act as a highly potent estrogen mimic at very low concentrations has been demonstrated (reviewed in [
Although the data from αERKO females paralleled that of wild-type females (unpublished data), βERKO mice yielded a surprising meiotic phenotype that did not fit our expectation: Pachytene oocytes from unexposed ERβ −/− females exhibited virtually identical defects to BPA-exposed wild-type females. Specifically, we observed similar levels of synaptic aberrations (57.0% of pachytene cells from unexposed ERβ −/− females, as compared to the 52.0% level in BPA-exposed wild-type females in
(A) For unexposed animals, there was no difference in mean number of MLH1 foci/cell between wild-type (26.3 ± 3.0) and heterozygous (25.8 ± 2.8) females, but unexposed mutants (28.7 ± 3.2) were highly significantly increased over wild type (
(B) Among exposed animals, the mean values for the three genotypes were virtually identical, but all had highly significantly elevated means over that of unexposed wild-type animals (28.6 ± 3.5,
Previous studies in our laboratory of female mice exposed as young adults [
In contrast, the studies described herein reveal an effect of BPA on meiotic chromosome segregation by a second, and completely independent, mechanism, that is, by disturbing synapsis and recombination between homologs in the fetal ovary. The finding that unexposed ERβ-null females exhibit a similar phenotype—and that the phenotype cannot be enhanced by BPA exposure—suggests that BPA exerts its effects on the fetal ovary by interfering with ERβ-mediated cellular responses. Knockouts of key meiotic genes involved either in synapsis (e.g., [
Defects in synapsis and altered levels of recombination have been correlated with increased aneuploidy in a variety of eukaryotic species. In addition, in humans, subtle changes in the placement of exchanges are correlated with meiotic nondisjunction; indeed, aberrant recombination is the only known molecular correlate of meiotic aneuploidy (reviewed in [
While these are worrying possibilities, the implications of our findings are actually much broader: synaptic and recombination defects typically result in the loss of a significant proportion of oocytes prior to sexual maturation [
Further, because oocytes in the fetal ovary are not yet enclosed in primordial follicles, our findings raise the intriguing possibility that, during the earliest stages of oogenesis, the oocyte is directly responsive to estrogen and to chemicals that can bind ERβ. Although localization studies of ERβ in the adult mouse ovary demonstrate the presence of the receptor largely in granulosa cells [
An obvious and important question is whether the effects observed in mice can be translated to humans. BPA levels in the parts per billion range have been reported in human serum and amniotic fluid [
All wild-type mice used in the study were on the C57BL/6 inbred strain background. They were housed in ventilated rack caging in a pathogen-free facility, with drinking water provided in glass water bottles and mouse chow (Purina 5010,
For exposures, BPA or placebo pellets (Innovative Research of America, www.innovrsrch.com) were implanted according to manufacturer guidelines in pregnant females at 11.5 d of gestation. Pellets were designed to release 400 ng of BPA daily, with doses calculated assuming an average weight of 20 g for sexually matured females. This dose was chosen based on our previous studies of BPA exposure in young adult females [
To obtain prophase oocytes, pregnant females were killed at 18.5 d of gestation and fetal ovarian tissues prepared as described previously [
Synaptonemal complex preparations were made [
Statistical evaluations of possible between-group differences in the mean numbers of MLH1 foci or chiasmata were carried out using standard
We thank John Nilson and Fred vom Saal for discussions and helpful comments on the manuscript.
bisphenol A
diethylstilbestrol
synaptonemal complex