Effects of Maternal Diet and Exposure to Bisphenol A on Sexually Dimorphic Responses in Conceptuses and Offspring

Authors


Author’s address (for correspondence): Cheryl S. Rosenfeld, Departments of Biomedical Sciences and Bond Life Sciences Center, University of Missouri, Columbia, MO 65211, USA. E-mail: rosenfeldc@missouri.edu

Contents

Whereas sexual differentiation is considered as the onset of differentiation of the male or female gonads, mounting evidence indicates that sex differences in developmental programming are established as early as the zygotic stage. Genetic and epigenetic differences between the sexes might govern how each responds to shifts in their early environment, including in the uterus or culture dish, as in the case of in vitro cultured pre-implantational embryos. Even if no differences are evident between the sexes at birth, divergent conceptus responses to surrounding changes, such as maternal diet and exposure to endocrine disrupting compounds (EDC), such as bisphenol A (BPA), might predispose one sex over the other to later adult-onset diseases, otherwise termed developmental origin of health and disease (DOHaD). Overall, males subjected to less than optimal in utero conditions tend to be at greater risk for various diseases, including neurobehavioural disorders. As the placenta is the primary nutrient acquisition and communication organ between the dam and foetus, its ability to adapt rapidly to environmental shifts might buffer the conceptus against environmental insults. The placenta of one sex over the other might possess greater ability to respond to environmental fluctuations. In utero environmental changes, including maternal nutrient excess or reduction or exposure to the EDC, BPA, might govern sex-dependent behavioural alterations. In sum, this review examines the evidence to date that male and female zygotes and conceptuses diverge in their responses to shifting environmental conditions and whether these contrasting sexually dimorphic responses underpin later DOHaD outcomes, namely neurobehavioural changes.

Introduction

The periconception through parturition period might dramatically influence the future health of a woman’s offspring, including their risk for stroke, coronary heart disease, Type 2 diabetes, obesity and hypertension (Barker 1997). This in utero programming, leading to disease later in life, is now generally referred to as either the ‘developmental origins of health and disease’ (DOHaD) or the ‘fetal origin of adult disease’ (FOAD) (Barker 1997). Even prior to sexual differentiation, the two sexes might demonstrate opposing responses to their surrounding environmental changes. The ability of each sex to adapt quickly to these challenges might dictate their risk for later diseases. These environmental fluctuations can include factors ranging from maternal diet, stress and exposure to endocrine disrupting compounds (EDC). For illustration purposes, we will focus on how maternal diet/nutrient composition and exposure to the EDC, such as bisphenol A (BPA), might impact in a sex-dependent manner the developing pre-implantational embryo and conceptus and how these alterations might lead to adult-onset diseases, specifically neurobehavioural disruptions. Furthermore, it should be noted that there are a range of other pervasive environmental chemicals, such as methoxychlor, polychlorinated biphenyls, polybrominated diphenyl ethers, organochlorine pesticides, di-(2-ethylhexyl)-phthalate (DEHP), vinclozolin (Mahoney and Padmanabhan 2010; Hamlin and Guillette 2011; Kang et al. 2011; Petro et al. 2012), that may independently and synergistically act to disrupt these above processes. For brevity and illustration purposes, this review will focus on the effects of BPA. All animal experiments reported herein were approved by the University of Missouri Animal Care and Use Committee (ACUC) committee and performed in accordance with NIH Animal Care and Use Guidelines.

Effects of Maternal Diet/Nutrients and Exposure to Bisphenol A on Male and Female Pre-implantational Embryos

With the growing incidence of diabetic pregnant mothers (Pampfer 2000; Doblado and Moley 2007), it has become essential to understand how varying glucose concentrations impact pre-implantational embryo development (Diamond et al. 1991; Fraser et al. 2007; Wyman et al. 2008; Pantaleon et al. 2010). Culture of pre-implantation mouse embryos in media containing glucose concentrations spanning 15–27 mm (2–3 times that of serum from normal mice) suppresses blastocyst development and cleavage rates (Diamond et al. 1991; Fraser et al. 2007; Pantaleon et al. 2010); whereas an excessive concentration (52 mm) leads to higher rates of post-implantational resorptions following embryo transfer into surrogate dams (Wyman et al. 2008). Increasing glucose concentrations in the culture media appears to favour the development of bovine male in vitro cultured embryos (Bredbacka and Bredbacka 1996), and thus, greater glucose concentrations might ultimately lead to a skew of pre-implanational sex ratio toward males (Gutierrez-Adan et al. 2001; Larson et al. 2001; Kimura et al. 2005). Conversely, another study suggests that much higher glucose concentrations provided during in vitro culture decreases the percentage of males in both mice and bovine (Jimenez et al. 2003); whereas a separate study indicates pregnant mice rendered diabetic with streptozocin treatment birth more male than female offspring (Machado et al. 2001). Pregnant diabetic women have been proposed to give birth to a higher percentage of daughters compared to non-diabetic pregnant mothers (Rjasanowski et al. 1998), although these results remain controversial (James 2006). To test whether glucose concentrations approximating that which occurs in diabetic mothers affect pre-implantational embryo development in a sex-dependent manner and offspring outcomes in transferred embryos, mouse zygotes were cultured to the blastocyst stage in the presence of low and high glucose concentrations, 0.2 and 28.0 mm, respectively. These concentrations of glucose were chosen because the former is routinely used in culturing of mouse embryos (Summers et al. 1995; Biggers and McGinnis 2001; Amarnath et al. 2011) and the latter has been previously used in mouse embryo culture to approximate serum glucose concentrations observed in diabetic pregnant women (Diamond et al. 1991; Fraser et al. 2007; Pantaleon et al. 2010). In vitro culture in medium with a glucose concentration approximating that of diabetic serum reduced total and trophectoderm (TE) cell numbers in blastocysts (Fig. 1) and their development to term when transferred into surrogate dams (Tables 1 and 2), but these detrimental effects were not sex-specific, and equal number of male and female offspring were born in each group (Table 3) (Bermejo-Alvarez et al. 2012).

Figure 1.

 Representative confocal images of blastocyst trophectoderm (TE) and total cell staining and PCR sexing procedure. (a) 10 μm Z-stack section of a blastocyst. Total cell number were determined based on DAPI nuclei staining (right image); TE cells were detected by antiCDX2 (middle image) and composite overlay with CDX2 positive cells (arrows) correlating with TE cells and DAPI positive/CDX2 negative cells representing inner cell mass (ICM) (left image). (b) Example of PCR sexing analysis of blastocysts. Female embryos exhibited only one band corresponding to Rn18S sequence; whereas males also amplify a Y chromosome-specific product (DYzEms3). Adapted from (Bermejo-Alvarez, 2012)

Table 1.   Effect of glucose concentration during mouse IVC on embryo development. There were no differences (p < 0.05).
Glucose (mm)No. of oocytes% of cleaved
Mean ± SEM (no.)
% of blastocysts
Mean ± SEM (no.)
0.218591.6 ± 1.9 (169)84.4 ± 2.9 (153)
2823395.1 ± 2.1 (221)77.9 ± 4.2 (188)
Table 2.   Mean ± SEM effect of glucose concentration during mouse IVC on embryo cell number according to sex. Two-way anova detected a significant association between glucose concentration and both total cells or TE (p < 0.05), but there were no significant differences between sexes or in ICM cell numbers.
Glucose (mm)SexTotal no. of cellsTrophectodermInner cell mass
  1. a–dWithin a column, means without a common superscript differed (p < 0.05).

0.2Male76.3 ± 4.6a60.8 ± 4c15.4 ± 1.2
0.2Female76.3 ± 4a61.9 ± 4.8c14.4 ± 1.4
28Male61.1 ± 3.8b45.8 ± 3.1d15.3 ± 1.5
28Female54.8 ± 3.9b38.6 ± 3.6d16.1 ± 1.3
Table 3.   Effect of glucose concentration during mouse IVC on survival to term and litter sex ratio.
Glucose (mm)Embryo transfersEmbryos transferred% of survival to term mean ± SEM (no.)% of males (no.)
  1. a,bWithin a column, means without a common superscript differed (p < 0.05).

0.21010074 ± 4 (74)a52.7 (39)
281212055.8 ± 7.1 (67)b52.2 (35)

The reproductive tract milieu might also contain varying amounts of EDC, including BPA at reported concentrations of 1–2 ng/ml in ovarian follicular fluid (Ikezuki et al. 2002), and these chemicals might impact pre-implantational embryos. Yet, only a few studies have analysed the effects of BPA on pre-implantational embryos. One of these studies suggests that in vivo developmental exposure to BPA (100 mg/kg body weight/day) leads to suppression of mouse pre-implantational embryo progression, transport and implantation into a receptive surrogate uterus (Xiao et al. 2011). In contrast, another study suggests that only a high dose of BPA (100 μm) in the culture media inhibits in vitro progression to the blastocyst stage in treated mouse embryos, but those subjected to lower concentrations of BPA, spanning 1 and 3 nm, exhibit similar cleavage rates as controls (Takai et al. 2000). Future efforts must thus be directed at reconciling these apparent differences between in vivo versus in vitro exposure to BPA, determining the dose range where BPA hinders pre-implantational embryo development, and whether these effects occur in an epigenetic and sex-dependent manner.

Effects of Maternal Diet and Exposure to Bisphenol A on Placental Function

The placenta transfers nutrients from the mother to the foetus and it acts as the primary communication organ for both participants (Jansson and Powell 2006). How the placenta responds to shifts in the in utero environment might thus either minimize or exacerbate the risk for later adult diseases. For example, the effects of maternal stress on placental function might selectively increase the risk for neurobehavioural disorders in males (Mueller and Bale 2008). It has also been postulated that placental ability to uptake O2 and nutrients might govern the later risk for cardiovascular disease (Thornburg et al. 2010). Thus, the placenta is seemingly the most logical organ to monitor how fluctuations in the in utero environment, including maternal diet, impact the developing conceptus. Yet, only select studies have analysed the impact of altering maternal diet on the full-range of placental gene responses (Gheorghe et al. 2006; Gallou-Kabani et al. 2010; Mao et al. 2010). The 2006 study determined that in mouse dams maintained on a 50% reduced protein content but otherwise isocaloric diet, for a short period during gestation (10.5–17.5 days post-coitus), leads to seemingly harmful effects on placental gene expression (Gheorghe et al. 2006). Yet, this study did not explore the possibility that the placenta of sons versus daughters might differentially respond to the protein-restricted maternal diet.

Only a few studies have defined individual transcripts that are contrastingly expressed in male versus female human placentae (Brown et al. 1987; Steier et al. 2004; Lehavi et al. 2005), and there has only been one paper in which microarray analyses were used to determine the global gene transcripts that are sexually dimorphic in the placenta with the placenta of daughters demonstrating greater number of gene changes compared to sons (Sood et al. 2006).

Before our study was published, there had not been an in-depth exploration on the interaction between maternal diet and expression of sexually dimorphic placenta gene transcripts that might begin to explain the varying sensitivities of male and female foetuses to what a mother eats or potentially other physiologically stressors that she may be subjected to during pregnancy (Mao et al. 2010). We examined how maternal diets (low fat, LF, very high fat, VHF, and chow-based or control (C) diets) influence the full range of placental gene expression profile in male and female conceptuses at 12.5 day post-coitus (dpc), a time characterized by completion of the morphological development of the placenta (Mueller and Bale 2008), yet the gonads are in the early stage of formation. Our studies provide conclusive evidence that placental gene expression is strongly impacted by maternal diet and foetal sex (Fig. 2a,b). Expression of 1972 genes in the placenta was changed more than two-fold (p < 0.05) in comparisons across diet in at least one of the three groups, and select gene expression patterns were confirmed with quantitative real-time PCR (qRT-PCR) (Mao et al. 2010). Importantly, each maternal diet provided a unique signature of sexually dimorphic genes, with expression overall greater in genes (651 of 700) from female placenta than those of males. Many genes in the placenta considered synonymous with regulating renal function were altered by maternal diet, including those regulating ion balance and chemoreception. Our data also revealed that the placenta expressed the majority of the known olfactory receptor genes (Olfr), with several of these transcripts influenced by diet and foetal sex. The placenta of these olfactory receptor genes might be a means by which it perceives slight nutrient changes and adapts accordingly. These combined studies demonstrate that maternal diet and foetal sex shape the placental gene transcript profile, with placentae of females seemingly more adaptable and ready to respond to changes in the mother’s diet.

Figure 2.

 (a) Conditional heat map analysis based on maternal diet effects on placental gene expression. The gene expression pattern in the placentae from dams on the chow (C) diet is distinct from those dams maintained on the defined diets [low fat (LF) and very high fat (VHF)]. While there were overt differences in caloric density between the two defined diets, placental gene expression pattern of LF and VHF conceptuses resembled each other more than those of the C diet group. Based on the gene tree clustering of 1972 genes whose expression was altered more than twofold with p < 0.05, the gene expression pattern of placentae from dams on the LF and VHF diets was distinguished. (b) Conditional heat map analysis based on foetal sex effects on placental gene expression. The placental gene expression pattern from male conceptuses clustered separately from the gene expression pattern of the placentae from females based on two-fold differences, which includes all total genes and spans all dietary groups (p < 0.05). The majority of the placental samples were from samples implanted in the right uterine horn with the notable exceptions from the left uterine horn indicated with an asterisk. As these samples clustered with the others, it implies that uterine implantation site does not mediate placental gene expression patterns. Adapted from (Mao et al. 2010).

After our studies were published (Mao et al. 2010), another group examined the impact of a high-fat diet (HFD) on expression of imprinted genes and DNA methylation patterns in the female and male mouse placenta (Gallou-Kabani et al. 2010). For all but one of the imprinted genes analysed, the placentae of males demonstrated decreased gene expression compared to those of females with the exception being Ascl2, where placental expression in males was greater than that of females (Gallou-Kabani et al. 2010). No clear pattern emerged, though, on the overall direction of imprinted gene expression based on maternal diet. Exposure to a maternal HFD led to significant DNA hypomethylation in the placenta of females compared to control females, and this difference across maternal diet groups was not evident in their male siblings.

In humans and rodents, environmentally relevant concentrations of BPA can easily be transferred across the placenta (Balakrishnan et al. 2010; Morck et al. 2010; Nishikawa et al. 2010), and BPA can be detected in amniotic fluid of humans with a reported concentration range of 2.80–5.62 ng/ml to 0.5 μg/l (Ikezuki et al. 2002; Yamada et al. 2002; Engel et al. 2006). Treatment of pregnant mice with BPA (0.2 mg/kg body weight/day) from 8.5 to 12.5 dpc leads to inappropriate expression of select imprinted genes in the placenta (Kang et al. 2011). Another study that treated pregnant mice with 10 mg BPA/kg body weight led to dramatic placental changes, including decreased region in the labyrinthine zone, narrowing of the intervillous spaces, degeneration of the trophoblastic giant cells and spongiotrophoblastic layers of the conceptuses with resulting reduced survival of offspring (Tachibana et al. 2007). Gestational exposure from 6.5 days post-coitus (dpc) to 17.5 dpc to 2 μg/kg BPA also alters the expression of several nuclear and non-nuclear receptors, and the observed effects occur in a sex-dependent manner (Imanishi et al. 2003). Another means to address the effects of BPA on placental cells has been to use an in vitro-based approach with stable cell cultures (Benachour and Aris 2009; Avissar-Whiting et al. 2010; Morice et al. 2011). Culture of three human immortalized cytotrophoblast cell lines in the presence of BPA (25 ng/μl) altered the expression pattern of several microRNAs (miRNAs), including miR-146a and correspondingly led to slower cellular proliferation and greater vulnerability to DNA damage (Avissar-Whiting et al. 2010). A second study that used the JEG-3 human trophoblastic cell line also demonstrated that exposure of these cells to BPA (up to 10−6 m) inhibited cell proliferation and increased apoptosis (Morice et al. 2011). Similar pro-apoptotic and toxic effects have been described in freshly isolated, primary human placental cells exposed to BPA (0.02 μg/ml) (Benachour and Aris 2009). Thus, BPA exposure might lead to less number of viable placenta cells that are available to uptake essential nutrients from the mother and compromise placental communication between the mother and conceptus.

Effects of Maternal Diet and Exposure to Bisphenol A on Later Offspring Behaviours

A variety of neurobehavioural patterns are programmed during in utero development, particularly those that are shaped by early exposure to testosterone and oestrogen (Watson and Adkins-Regan 1989a,b; Bowers et al. 2010; Konkle and McCarthy 2011). Rat females exposed during gestation to a protein-restricted diet exhibit increased anxiety-like behaviours but also show learning impairment relative to controls (Reyes-Castro et al. 2012a). Furthermore, rat males exposed to a similar protein-restricted diet regiment exhibit decreased anxiety compared to control sons (Reyes-Castro et al. 2012b). Another study, though, reported that a maternal protein-restricted diet leads to increased anxiety-like behaviours in male and female offspring (Watkins et al. 2008). Female but not male infants from non-human primate mothers maintained a HFD exhibits increased anxiety-like responses (Sullivan et al. 2010). Sons but not daughters of rat dams on high saturated fat or trans-fat diet are more anxious than control sons (Bilbo and Tsang 2010). Some of these offspring behavioural changes might be due to increased lipid peroxidation in the brain, decreased expression of BDNF and impaired synaptic connections (Tozuka et al. 2010). Maternal lipid-enriched diet might also induce DNA hypomethylation of select genes in the brain of her offspring (Tozuka et al. 2010).

Developmental exposure to BPA also disturbs programming and elaboration of a variety of adult behaviours, and some of these traits are sexually dimorphic (Palanza et al. 1999, 2002; Patisaul and Bateman 2008; Jasarevic et al. 2011; Xu et al. 2011). At sexual maturity, male deer mice (Peromyscus maniculatus bairdii) display enhanced spatial navigational ability and thereby expand their home territory with the increased likelihood of locating prospective breeding partners that are widely disseminated throughout the ecology. We have recently determined that developmental exposure of deer mouse males to either 50 mg BPA/kg feed weight or 0.1 μg ethinyl estradiol (EE)/kg feed weight through the maternal diet prevents the elaboration of essential spatial navigational abilities when the animals reach sexual maturity (Fig. 3) (Jasarevic et al. 2011). These disrupted behaviours are likely due to a failure in the BPA- and EE-exposed males to learn directional and positional intra-maze cues that aide in navigating a large-scale space (Fig. 4). In contrast to the feminized responses observed in EE-exposed males, female deer mice exposed developmentally to EE apparently become masculinized in their spatial navigational abilities, almost approximating those of control males (Fig. 3). The EE-treated females quickly located the escape hole and adopted a search strategy similar to control males (Fig. 4). Yet, no comparable response was observed in female deer mice developmentally exposed to BPA, suggesting that this dose of BPA was either insufficient in activating the oestrogen receptor pathway or BPA does not act solely by engaging oestrogen receptors. Males exposed during development also showed more anxiety-like and correspondingly decreased exploratory behaviours than control males (Jasarevic et al. 2011). In a mate-choice experiment, control and BPA developmentally exposed females seemingly sense the compromised state of the BPA-exposed males and instead preferred the control males on a 2 : 1 basis (Jasarevic et al. 2011). In these experiments with the male and female deer mice offspring, there were no overt differences in general or outward phenotype, including neuromuscular, olfactory, auditory and visual senses between the males exposed during development to BPA or EE compared to control males. The primary difference was in select and seemingly subtle behaviours in the males that are vital for reproductive success, which would likely be overlooked without specifically analysing for these traits. Epigenetic alterations in the developing brain following in utero exposure to BPA (20 μg/kg body weight) (Yaoi et al. 2008) might account for these observed sex differences in spatial abilities and anxiety behaviour in these exposed animals.

Figure 3.

 Effects of early developmental exposure to bisphenol A and ethinyl estradiol on spatial learning and memory of adult male and female deer mice (Peromyscus maniculatus) in the Barnes maze. Latency, that is, time required to escape the maze across days of training for males (a) and females (c) (Mean ± SEM). Number of escape errors across days of training for males (b) and females (d) (Mean ± SEM, *p < 0.01). Adapted from (Jasarevic et al. 2011).

Figure 4.

 Effects of early developmental exposure to bisphenol A (BPA) and ethinyl estradiol (EE) on spatial search strategy of adult male and female deer mice (Peromyscus maniculatus) in the Barnes maze. (a) Examples of composite images from single animals tracked from entry to escape illustrating different spatial strategies used to exit the maze. (b) Distribution of different spatial strategies according to sex, diet exposure and day of training. During the initial training period (day 1), most animals navigated by using a random strategy (black), followed by a serial search strategy (yellow). The most efficient spatial search strategy (green) emerged when the animals began to use directional and positional intra-maze cues. By day 3 of training, control males used more efficient strategies than control females and EE- and BPA-exposed males, who in turn did not differ on any day (p < 0.0002). EE-exposed females used more efficient strategies than control and BPA-exposed females on all days except day 2. Adapted from (Jasarevic et al. 2011).

Conclusions

The fertilization through parturition period is thought to be a time when the developing embryo/conceptus is nourished and supported by the mother’s reproductive tract secretions and underlying uterine tissue. Yet, the developing conceptus is also at the mercy of the in utero environment. Components in the maternal diet and exposure to BPA might dramatically shift the trajectory of conceptus development. Even if the offspring are born seemingly healthy, these abnormalities, including neurobehavioural disturbances, might manifest with age. These alterations might occur in a sex-dependent manner and equate to profound cognitive and memory deficiencies that hamper the overall ability of the animal or human to thrive and chances for reproduction. While there is increasing concern to minimize exposure of the neonate and infant to EDC, including BPA, and meanwhile optimize nutrients provided during the postnatal period, there is a need to understand better the impact of maternal diet and exposure to these chemicals on her unborn offspring. Furthermore, pregnant women should be educated on adjusting their diet and reducing their exposure to EDC, including BPA, to prevent later disease risk in their offspring.

Acknowledgements

Dr Jiude Mao, Dr Pablo Bermejo-Alvarez, Eldin Jasarevic, Paizlee T Sieli, Dr David Geary and Dr R. Michael Roberts contributed to work described herein. Funding for these studies was provided by National Institutes of Health Challenge Grant RC1 ES018195 (to C.S.R.) and NIH Grant HD 044042 (to R.M.R. and C.S.R.).

Author Contributions

The manuscript was written by CSR.

Conflicts of interest statement

The author declares no conflicts of interest.

Ancillary