Perinatal exposure to environmentally relevant levels of bisphenol A decreases fertility and fecundity in CD-1 mice.

When analyzed by HPLC with ultraviolet (UV) detection, the retention time of our BPA stock (43 min) was identical to that expected for authentic BPA ( Zalko et al. 2003 ). Based on UV absorbance at 269 nm, BPA purity was close to 100% ( Figure 1A ). We detected minor impurities at lower wavelengths that eluted from the column with retention times of 10 min (major peak) and 22–23 min (minor peak), respectively. ESI-MS analysis detected the molecular ion of BPA (negatively charged) at m / z 227, consistent with its molecular mass of 228. We observed no other ion in the mass spectrum, indicating that interfering compounds were not present or were undetectable. The MS/MS analysis confirmed the structure of BPA, with a fragment ion detected at m / z 212, corresponding to the loss of one methyl group, as well as a fragment ion at m / z 133, diagnostic of the elimination of the phenol moiety. The NMR analysis showed two doublets at 7.01 and 6.65 ppm and the singlet at 1.57 ppm, characteristic of BPA. The singlet at 1.28 ppm did not belong to BPA; we concluded that it was associated with the presence of an impurity, which we estimated by integration measurements to account for 3 ± 2% of the sample. In addition, in the 1 H NMR spectrum, the signals between 0.8 and 2.5 ppm corresponding to an impurity ( Figure 1B ) strongly suggested that this impurity was not a phenolic compound. The degree of purity determined for the BPA used for these experiments (97 ± 2%) was comparable to that reported by the manufacturer (≥ 99%). We observed no statistical differences in the dam weights by treatment groups. The percentage of dead pups and the sex ratio were comparable in all groups ( Table 1 ). Additionally, when mated at 2 months of age, the time to first delivery did not differ by treatment group ( Table 1 ). The cumulative number of live pups born to BPA 25 ng/kg mice was significantly lower compared with controls starting at week 16 (mean ± SE, 49.74 ± 4.21 pups vs. 67.48 ± 4.38 pups; p = 0.037) and persisted until the end of the experiment (67.00 ± 5.70 pups vs. 81.90 ± 5.45 pups; p < 0.001) ( Figure 2 ). BPA 25 μg/kg mice also had significantly lower cumulative number of pups per dam compared with control mice. We observed a statistically significant difference between this group and controls beginning at week 16 (49.85 ± 3.69 pups vs. 67.48 ± 4.38 pups; p = 0.037) that persisted until the end of the experiment (58.25 ± 4.77 pups vs. 81.90 ± 5.45 pups; p < 0.001). The cumulative number of pups born to mice exposed to DES was also significantly lower than in the controls, beginning at week 25 (64.55 ± 6.44 pups vs. 80.00 ± 5.19 pups) and persisting until week 32 (65.70 ± 6.66 pups vs. 81.90 ± 5.45 pups; p < 0.01). The cumulative numbers of pups did not differ between the control and BPA 250 ng/kg groups at any time point. However, the BPA 250 ng/kg group differed from the BPA 25 ng/kg and BPA 25 μg/kg groups [77.89 ± 6.64 pups vs. 57.00 ± 4.89 pups ( p = 0.042) and 56.20 ± 4.19 pups ( p = 0.042), respectively] starting at week 21 and persisting until the end of the study period [88.22 ± 7.95 pups vs. 67.00 ± 7.70 pups ( p < 0.001) and 58.25 ± 4.77 pups ( p < 0.001), respectively]. Likewise, we observed a statistically significant difference between BPA 250 ng/kg and DES groups beginning at week 27 (86.11 ± 7.39 pups vs. 65.45 ± 6.62 pups; p = 0.037) and persisting until the end (88.22 ± 7.95 pups vs. 65.70 ± 6.66 pups; p < 0.004). We assessed fertility by the total number of litters per dam for each treatment group at the end of the 32-week period of forced breeding. Dams in the BPA 25 μg/kg group had significantly fewer pregnancies than the control dams ( p = 0.024) ( Figure 3A ). We observed a trend toward a decreased number of pregnancies in the BPA 25 ng/kg group (5.74 ± 0.48 pregnancies) and DES group (5.5 ± 0.44 pregnancies) relative to controls (6.37 ± 0.39 pregnancies). The proportion of dams with six or more litters was significantly lower in the BPA 25 μg/kg group (35%) than in the controls (76%) ( Figure 3B ), as was the proportion of mice with four or more litters in the BPA 25 μg/kg group (60%) compared with the controls (95%; p = 0.012). We created scatterplots to depict the number of pups born at each delivery for each treatment group, determined a best-fit line for each treatment group, and compared the slopes and correlation coefficients ( R ) of these lines ( Figure 4 ). The difference in correlation coefficients indicated that the variation in fecundity within each treatment group that is attributed to aging differed for the BPA 25 μg/kg group ( R = 0.60) compared to the control ( R = 0.36; p = 0.025) and BPA 250 ng/kg ( R = 0.30; p = 0.005) groups. The differences in slope indicated that the decline in number of pups delivered over time was hastened in the BPA 25 μg/kg group compared with the control ( p = 0.012) and BPA 250 ng/kg ( p = 0.003) groups.

Developmental plasticity has been a central theme in embryology ( Gilbert and Epel 2009 ), and it is currently being explored in medicine and public health because of growing evidence from studies in animal models and humans that various adult-onset diseases, such as obesity, diabetes, and cancer, may have their origins during fetal life ( Bateson et al. 2004 ; Newbold et al. 2008 ; Soto and Sonnenschein 2010 ). Exposure to environmental endocrine disruptors during fetal development has been postulated to contribute to declining conception rates and increased incidence of female reproductive disorders such as oocyte aneuploidy, polycystic ovarian syndrome, and altered cyclicity, as well as endometriosis, uterine fibroids, fetal growth retardation, and pregnancy loss ( Crain et al. 2008 ). Perinatal exposure to environmentally relevant doses of BPA has resulted in early vaginal opening and altered estrous cyclicity, overexpression of ERα and progesterone receptor in the endometrium, and an increased number of ovarian antral follicles in mice ( Markey et al. 2003 , 2005 ). All of these studies suggest the likelihood of functional consequences such as altered reproductive outcomes in BPA-exposed females. In the present study, perinatal exposure to BPA resulted in a decline in reproductive capacity in a forced-breeding protocol. BPA 25 μg/kg mice showed a significant decrease in the number of pregnancies, indicating a decline in fertility. We also observed a significant decrease in the cumulative number of pups born to BPA 25 ng/kg, BPA 25 μg/kg, and DES dams relative to controls. The BPA 25 μg/kg group showed a progressive decline in the number of pups delivered, which suggests the possibility of an accelerated reproductive decline. These effects are consistent with previously reported alterations in the ovary, uterus, hypothalamus, and pituitary, the estrogen target tissues that are essential for successful delivery of live pups. Therefore, all these BPA-associated alterations undoubtedly contribute to the diminished reproductive outcome observed here. The forced breeding protocol does not allow for detailed evaluation of fertilization, implantation, and pregnancy maintenance, including the potential for early pregnancy loss in these animals. The protocol records successful pregnancies and provides information about litter size and composition. Therefore, it is impossible to pinpoint a single cause of the reduced fertility observed with perinatal BPA exposure. The overall pattern observed is suggestive of accelerated reproductive decline, which could have involved changes in the function of any or several of the potential BPA target tissues, including alterations at the hypothalamus/pituitary level linked to aging that could have interfered with the preovulatory LH surge and the number of ova ovulated. This study stresses the need to identify the primary targets and the precise sequence of events leading to compromised reproduction. In contrast to the significant effects on reproductive capacity revealed over time by this forced breeding experiment, neither the time to first delivery nor the number of pups in the first litter was affected by BPA exposure. Standard toxicological studies on BPA and other endocrine disruptors have routinely analyzed the effects on first pregnancy alone, with no follow-up on potential effects in subsequent pregnancies. As a result, they failed to identify changes in reproductive functions that manifest during subsequent pregnancies, leading to an incomplete evaluation of the effects of BPA exposure on fertility. This was pointed out by Honma et al. (2002) , who acknowledged that although DES did not affect reproduction in their experiments, it did affect reproductive parameters in prior forced breeding assays. The forced breeding protocol we used in the present study was adapted from McLachlan et al. (1982) , who noted a dose-related decrease in reproductive capacity ranging from a decline of the cumulative number of pups in mice exposed prenatally to 10 ng DES/kg BW/day to a high frequency of sterility at the highest dose of 100 μg DES/kg BW/day compared with nonexposed controls. In our study, exposure to 10 ng DES/kg BW/day resulted in a significant decline in the cumulative number of pups born to the treated mice relative to controls. Of interest, the intermediate BPA dose, 250 ng/kg, did not produce significant effects in the forced breeding regimen, whereas both lower and higher doses did. Natural hormones often show nonmonotonic dose–response curves ( Amara and Dannies 1983 ; Sonnenschein et al. 1989 ; Vandenberg et al. 2006 ; Wadia et al. 2007 ), which have also been reported for several endocrine disruptors ( Myers and Hessler 2007 ). These curves are believed to represent a composite of effects over diverse targets ( Vandenberg et al. 2006 ). Nonmonotonic dose–response curves induced by BPA have been documented in human prostate cancer cell line proliferation ( Wetherill et al. 2002 ), rat pituitary and cerebellar cortex cells ( Wozniak et al. 2005 ), and production of adiponectin by human adipocytes ( Hugo et al. 2008 ). Our data now add reproductive capacity to the list of nonmonotonic responses to BPA exposure. However, it is important to note that although the reproductive capacity of the BPA 250 ng/kg females did not appear to be affected in our forced breeding regimen, it is possible that other experimental protocols would reveal evidence of impaired reproductive function in this treatment group. Indeed, we have observed alterations in estrous cyclicity and antral follicle number ( Markey et al. 2003 ), mammary gland development ( Munoz de Toro et al. 2005 ), and the hypothalamus of females exposed to 250 ng BPA/kg BW/day ( Rubin et al. 2006 ). The forced breeding protocol used here eliminates regular estrous cycles and spontaneous preovulatory LH surges during the 32 weeks of breeding; therefore, it might mask relevant reproductive deficits in this treatment group. In an earlier study, CD-1 mice exposed to a single bolus injection of a low dose of BPA (25 μg/kg BW) showed significant levels of BPA residues (4% of the dose) in fetuses, confirming the passage of BPA through the placental barrier ( Zalko et al. 2003 ). The finding with regard to the purity of the BPA used in the present study and all our previous studies strongly indicates that it is BPA and not a contaminant that is responsible for the observed effects in exposed offspring. To date, most reports on BPA toxicity rely on the manufacturer’s claim of the purity of the chemical. In the present study we independently assessed and confirmed the purity of the commercial BPA used. In addition, the analysis indicated that the potential for any impurity to correspond to another phenolic compound is negligible. This finding strongly suggests that the effects observed are due to BPA. Numerous responses to BPA cited above and responses observed in the present study occur with BPA levels that are below the current U.S. Environmental Protection Agency reference dose for this chemical (50 μg/kg BW/day) ( Welshons et al. 2003 ; IRIS-USEPA 2002 ). In fact, in the present study we used doses that were 2- and 2,000-fold below the current safe dose of BPA. Some of the effects observed at low doses are not found at much higher doses ( Myers et al. 2009 ). Therefore, the current practice in regulatory toxicology of testing chemicals at very high doses might not reliably predict the hazards posed by low doses of endocrine-disrupting chemicals such as BPA. The current rise in the incidence of infertility cases as well as the advanced onset of puberty in girls may be significantly influenced by the exposure to multiple hormonally active chemicals in our environment. Our findings of reproductive effects in mice after perinatal exposure to low doses of BPA suggest the possibility that environmental exposures to low BPA doses might affect human reproductive health as well and support the need to strictly regulate its use.

Dimethyl sulfoxide (DMSO; CAS no. 67-68-5), DES (CAS no. 56-53-1), and BPA (4,4′-dihydroxydiphenyl dimethylmethane; CAS no. 80-05-7; product no. 23965-8, lot no. 03105ES; purity ≥ 99%) were purchased from Sigma Chemical Company (Saint Louis, MO, USA). We further analyzed the purity of the BPA stock by HPLC using a Spectra P1000 pump (Thermo Separation Products, Les Ulis, France) associated with a 250 mm × 4.6 mm (5 μm) Zorbax SB-C 18 column protected by a Kromasil C 18 guard precolumn (Agilent; Interchim, Montluçon, France). Mobile phases and analytical conditions were as previously described ( Jaeg et al. 2004 ). The system was coupled to a diode area detector (UV6000LP Spectra System; wavelength range, 220–360 nm; Thermo Separation Products). Next, the mass spectrometric (MS) analysis of BPA was performed on a quadruple ion trap LCQ mass spectrometer (Thermo Electron, Les Ulis, France) equipped with an electrospray ionization (ESI) source. Finally, the nuclear magnetic resonance (NMR) spectrum of BPA was obtained at 300K on a Bruker Avance DRX-600 spectrometer (Bruker, Wissembourg, France) operating at 600.13 MHz and equipped with a 5-mm H, C, N inverse triple-resonance TXI cryoprobe attached to a cryoplatform (the preamplifier cooling unit). BPA (1 mg) was dissolved in 600 μL deuterated methanol (CD 3 OD), and one-dimensional spectrum was acquired using a standard pulse sequence for 1 H NMR. Sixteen free induction decays were collected with a spectral width of 12 ppm into 64,000 data points. CD-1 female mice (12 weeks of age) and proven breeder male mice (all from Charles River Laboratories, Wilmington, MA, USA) were maintained in temperature- and light-controlled (14/10-hr light/dark cycle) conditions at the at Tufts University Human Nutrition and Research Center animal facility (approved by the Association for Assessment and Accreditation of Laboratory Animal Care International). All experimental procedures were approved by the Tufts University–New England Medical Center Institutional Animal Care and Use Committee. The animals were treated humanely and with regard for alleviation of suffering. Cages, water, and bedding all tested negligible for estrogenicity by the E-SCREEN assay ( Soto et al. 1992 ). Water was supplied ad libitum from glass bottles. Food (Teklad 2018; Harlan, Indianapolis, IN, USA) was supplied ad libitum . The estrogenicity of food lots used was measured by the E-SCREEN assay and found to be negligible (< 20 pmol estrogen equivalents per gram of feed). The estrogenic activity of the feed was also tested independently by W.V. Welshons in his laboratory ( Welshons et al. 1990 ). Mice were allowed to acclimatize for 5 days before being paired to mate. The morning on which a vaginal plug was detected was considered gestational day (GD) 1. On the evening of GD8, dams were implanted subcutaneously with Alzet osmotic pumps (model 1004; Alza Corp., Palo Alto, CA, USA) following the manufacturer recommendations. Pumps were designed to deliver vehicle alone (50% DMSO in water), the positive control (DES), or one of three doses of BPA. These pumps continued to release at a constant rate (0.11 μL/hr) until day 16 of lactation. The actual delivered dose of BPA decreased as pregnancy progressed because the weight of the mother at GD6 was used to calculate BPA dose, and body weights (BW) increased from this point throughout pregnancy. BPA exposure groups were 0 (control), 25 ng, 250 ng, and 25 μg BPA/kg BW/day. DES was administered at 10 ng/kg BW/day, which was the lowest effective dose tested in a forced breeding regimen ( McLachlan et al. 1982 ). The dams delivered naturally, and the F 1 litters were culled to eight pups on the day after birth, keeping at least two males. Litters were weaned on postnatal day 21. At 8 weeks of age, one F 1 female mouse from each treatment group was randomly chosen from each litter and individually housed with a nonexposed young male of proven fertility for a period of 32 weeks. A total of 18–21 dams in each group delivered litters (control, n = 21; BPA 25 ng/kg, n = 19; BPA 250 ng/kg, n = 18; BPA 25 μg/kg, n = 20; DES, n = 20). All the breeding cages were examined daily. On the day of each delivery, F 2 pups were separated from the mother, and their number, sex, and weight were recorded. The pups were sacrificed by decapitation, and the F 1 dam was immediately returned to the breeding cage. The cumulative number of live pups delivered, time to first litter, number of deliveries (fertility) per dam, number of neonates per litter (fecundity), and number of mice having litters were recorded. We used SPSS statistical software (version 15.0; SPSS Inc., Chicago, IL, USA) for all statistical analyses. The Kolmogorov-Smirnov test was used to assess differences in the cumulative number of pups between treatment groups. Differences in the total number of pregnancies were assessed using analysis of variance (ANOVA) and Dunnett’s post hoc test. Fisher’s exact test was used to analyze differences in the proportions of female mice in each treatment group with four or more pregnancies. All these results are presented as mean ± SE or percentages. To study fecundity, scatter plots and trend lines of the number of pups versus delivery number were produced for each treatment group. The slope and correlation coefficient ( R ) of the trend lines were each compared across treatment groups using t -tests. For all statistical tests, results were considered significant at p < 0.05.