Article pubs.acs.org/est
Occurrence and Maternal Transfer of Chlorinated Bisphenol A and Nonylphenol in Pregnant Women and Their Matching Embryos Mo Chen,† Zhanlan Fan,† Fanrong Zhao,† Fumei Gao,† Di Mu,† Yuyin Zhou,† Huan Shen,‡ and Jianying Hu*,† †
Laboratory for Earth Surface Processes, College of Urban and Environmental Sciences, Peking University, Beijing 100871, China Reproductive Medical Center, Peking University People’s Hospital, Peking University, Beijing 100044, China
‡
S Supporting Information *
ABSTRACT: Prenatal exposure has recently raised concerns over the health risks of endocrine disruptors; however, little is known about their extent and the mechanisms of maternal transfer in the embryo stage. In this study, bisphenol A (BPA), nonylphenol (NP), and their six chlorinated derivatives were quantified in decidua samples from 25 pregnant women and their matching embryos, which were collected as chorionic villi samples. Monochloro-BPA (MCBPA), dichloro-BPA (DCBPA), monochloro-NP (MCNP), and dichloro-NP (DCNP) were detected in over 70% of the decidua or chorionic villi samples, while BPA, NP, trichloro-BPA (TCBPA), and tetrachloro-BPA (TeCBPA) were detected in less than half. The geometric mean (GM) concentrations of MCBPA, DCBPA, NP, MCNP, and DCNP in chorionic villi samples were 0.13, 0.17, 5.33, 4.52, and 2.44 ng/g dw, respectively, higher than those in maternal decidua samples, which were 0.10, 0.12, 3.27, 1.85, and 0.74 ng/g dw, respectively, while the GM concentration of BPA was lower in chorionic villi samples (0.09 ng/g dw) than in maternal decidua (0.10 ng/g dw). The ratios of the average lipid-normalized concentrations of chemicals in chorionic villi to those in maternal decidua (EMR) were calculated to be 1.53 for MCNP and 2.38 for DCNP, while those of BPA, MCBPA, DCBPA, and NP were lower than 1 (0.39−0.97). Such obvious difference in maternal transfer is probably due to their different affinities to plasma proteins, as exemplified by the correlation between EMR and the binding affinities to T4 transport proteins (TTR). This is the first report on the occurrence and maternal transfer of chlorinated derivatives of BPA and NP in human embryos and decidua.
■
blood,16−18 consistent with what was observed in animal experiments. Ex vivo, in vivo, and epidemiological investigations all found that BPA and NP can relatively easily cross the placental membrane, and all of these studies focused on maternal transfer across the placental membrane. In fact, embryogenesis during the first 8 weeks of pregnancy is a key stage for neurulation, development of the nervous and circulatory systems and the heart; thus, toxic exposures in the embryonic period can be the cause of major congenital malformations because the precursors of the major organ systems are then developing.19,20 At this stage, a mature placental barrier has not formed to come into effect, and therefore, the embryo may be directly exposed to pollutants without any barriers such as the placenta.12 However, the data gap in this period still needs to be filled. Because both NP and BPA are chemically active and reported as ubiquitous in various environmental matrices,5,6,21−23 BPA and NP can generate various chlorinated byproducts by reacting with hypochlorite in the disinfection
INTRODUCTION Bisphenol A (BPA) and nonylphenol (NP) are classified as typical endocrine disruptors due to their ability to mimic estrogens.1,2 BPA is a worldwide high-volume chemical used in the production of epoxy resins and polycarbonate plastics, and NP in the environment mainly originates from the degradation of nonylphenolethoxylates (NPEOs), common nonionic surfactants used in industrial applications and daily life.3,4 Exposure to humans could occur by direct contact with materials containing BPA and NP or by intake of food and drinks that come in contact with them.5,6 Many studies have documented the occurrence of BPA and NP in human specimens such as urine, colostrum, placenta, and adipose tissue.7−11 Because the human fetus develops in the uterus, which is a relatively closed environment, maternal transfer is considered to be the most important exposure pathway for the fetus.12,13 A monkey experiment demonstrated that BPA can be maternally transferred to the fetus and detected in fetal serum, liver, brain, placenta, and umbilical cord tissue.14 Transplacental transfer of NP has been observed in a dual ex vivo recirculating model of placental perfusion.15 Transplacental transfer of BPA and NP was also observed in humans, and their levels and detection frequency in cord blood were both lower than in maternal © 2015 American Chemical Society
Received: Revised: Accepted: Published: 970
August 26, 2015 December 5, 2015 December 22, 2015 December 22, 2015 DOI: 10.1021/acs.est.5b04130 Environ. Sci. Technol. 2016, 50, 970−977
Article
Environmental Science & Technology
centrifuged for 10 min at 3000 rpm, and plasma samples were transferred and stored in pretreated 8 mL glass vials. All subjects were told to avoid food for 8 h before the blood sample collection. Decidua and chorionic villi sample collection was performed by hospital medical staff and assisted by our study team. The collected samples were immediately transferred and stored in pretreated 15 mL glass vials. All samples were subsequently kept at −80 °C until analysis. All subjects signed their consent to participate and filled and returned a brief questionnaire about demographic information including age, body mass index (BMI), and days after last menstruation. No subject reported any known occupational exposure to BPA and NP. This study was approved by the Human Ethics Committee of the Peking University Peoples’ Hospital (no. 2011-33). Sample Preparation and Analysis. Decidua and chorionic villi samples were freeze-dried and ground into fine powder. In this study, only free BPA, NP, and their chlorinated derivatives were directly extracted and measured. After that, 10 mg to 100 mg of sample was accurately weighed and spiked with 20 μL of methanol containing a mixture of internal standards (10 ng/mL for each). After equilibration for 30 min at room temperature, 3 mL of ACN was then added to the sample. The mixture was shaken for 20 min on an orbital shaker and centrifuged at 4000 rpm for 10 min, and the supernatant was transferred to a 15 mL glass vial (each vial was accurately weighed before use). Extraction from the residue was repeated twice more, and the organic layers were combined and concentrated to 1 mL. Lipid content were determined according to a method reported previously with minor modification.33 Briefly, an aliquot of 0.5 mL of the extract was evaporated to dryness and constant weight. The total extractable lipid was then measured gravimetrically. The remaining extract was transferred to a 1.5 mL centrifuge tube. After storage at −20 °C overnight, the extracts were centrifuged immediately at 10000g and −10 °C for 5 min (Thermal NK320) for the separation of organic solvent and solidified lipid. The supernatant was transferred to an 8 mL glass vial, and then 2 mL n-hexane was added and shaken for 10 min. The extraction was repeated twice, and each time, the n-hexane layer containing lipid was discarded. The ACN layer was saved, evaporated to dryness, and redissolved in 1 mL n-hexane. A cleanup procedure was then performed according to previously reported literature with minor modification.34 The extracted samples were loaded onto Sep-Pak Vac NH2 (6 cm3, 1 g) cartridges, which were preconditioned by 10 mL MeOH/ acetone (50:50, v/v), and 10 mL n-hexane. After being washed by 5 mL DCM/n-hexane (50/50, v/v), a volume of 5 mL MeOH/acetone (50:50, v/v) was used to elute the analytes from the NH2 cartridges. The extracts were evaporated to dryness under a gentle stream of nitrogen and then redissolved in 1 mL ACN for dansylation. Plasma samples were thawed at room temperature and prepared immediately for analysis. Each plasma sample (0.5 mL) was transferred into a glass centrifuge tube and spiked with 20 μL of MeOH containing a mixture of internal standards (10 ng/mL). D16-BPA was used as the internal standard for the quantification of BPA, and D12-DCBPA was used as the internal standard for MCBPA, DCBPA, TCBPA, and TeCBPA. 4-n-NP was used as the internal standard for the determination of NP, MCNP, and DCNP. After equilibration for 30 min at room temperature, 2 mL MTBE was then added to the sample. The mixture was shaken for 20 min on an orbital shaker and then
process of water-supply systems and in food-contact paper (FCP) bleaching.24,25 Chlorinated BPA and NP, including monochloro-BPA (MCBPA), dichloro-BPA (DCBPA), monochloro-NP (MCNP), and dichloro-NP (DCNP), have been widely detected in drinking water nationwide and in bleached FCPs.21,26 Although in vivo data are still limited for chlorinated BPAs and NPs, stronger binding affinities with estrogen receptors (ERs), thyroid hormone receptors (TRs), peroxisome proliferator-activated receptor γ (PPARγ), and pregnane X receptor (PXR) have been reported.27−29 Although many studies have documented the occurrence of chlorinated derivatives in human specimens such as urine, colostrum, milk, placenta, and adipose tissue,11,30−32 knowledge about human prenatal exposure is lacking. Before 8 gestational weeks, the embryo is involved in the chorionic villi, which covers the maternal decidual membrane face, and the decidual tissues participate in the exchanges of nutrition, gas, and waste during gestation as the maternal interface to the embryo.12 Thus, these valuable samples gave us a chance to assess the maternal transfer ability at the embryonic stage. In this study, BPA, NP, and their chlorinated derivatives were analyzed in plasma and decidual tissues from 25 recruited pregnant women (within 8 weeks after fertilization) and their matched embryos using a highly sensitive dansylation UPLC− ESI−MS/MS method. The maternal-embryo transfer potentials of seven target compounds were assessed by calculating the concentration ratios between chorionic villi and decidual tissue (EMRs). The results in this study fill the data gap regarding human embryo exposure and maternal transfer of BPA and NP and their chlorinated derivatives.
■
MATERIALS AND METHODS Standards and Reagents. Chlorinated BPA and NP derivatives, including 4-chloro-BPA (MCBPA), dichloro-BPA (DCBPA, a mixture of 2,6-dechloro-BPA and 2,6′-dichloroBPA (1:0.25)), trichloro-BPA (TCBPA), 2-chloro-NP (MCNP), and 2,6-dichloro-NP (DCNP), were synthesized in our previous study.21 4-NP was purchased as a technical-grade product from Hayashi (Tokyo, Japan, CAS 84852-15-3). BPA was obtained from Kanto Chemical Co. (Tokyo, Japan). Tetrachloro-BPA (TeCBPA) was purchased from TCI Corp. (Tokyo, Japan). Surrogate standards including 4-n-NP and D16BPA were supplied by C/D/N Isotope (Montreal, Canada). D12-2,2′-dichloro-BPA (D12-DCBPA) was supplied by @ rtMolecule (Poitiers, France). The purity of all commercial and synthesized compounds was 95% or higher. Dansyl chloride (DNS) and 4-(dimethylamino)pyridine (DMAP) were obtained from Sigma-Aldrich (St. Louis, MO). Sep-Pak Vac NH2 (6 cm3, 1 g) and silica (6 cm3, 1 g) solid-phase extraction (SPE) cartridges were purchased from Waters (Milford, MA). Methanol (MeOH), acetonitrile (ACN), nhexane, dichloromethane (DCM), acetone, and methyl tertbutyl ether (MTBE) were all HPLC grade and purchased from Fisher Chemical Co. (Beijing, China). Ultrapure water was prepared using a Milli-Q RC apparatus (Millipore, Bedford, MA). Subjects and Sample Collection. Human decidua, chorionic villi samples and plasma samples were collected from 25 women in early pregnancy (first trimester) who came to the Peking University People’s Hospital (Beijing, China) to induce abortion from January to November 2014. Blood samples were collected before the operation through venipuncture by registered nurses with vacuum tubes and 971
DOI: 10.1021/acs.est.5b04130 Environ. Sci. Technol. 2016, 50, 970−977
Article
Environmental Science & Technology
passing through the entire analytical procedure to determine the background. The linearity of the method was measured over the established working concentration range of 0.005−100 ng/mL, with three replicated working standards for each concentration level. The interday precisions were calculated based on the means of three spiked samples at three different levels (0.02, 0.2, and 2 ng/mL for BPA, D16-BPA, NP, 4-n-NP, and DCNP and 0.005, 0.1, and 2 ng/mL for MCBPA, DCBPA, D12-DCBPA, TCBPA, TeCBPA, and MCNP in plasma samples; 0.02, 0.2, and 2 ng/g for BPA, D16-BPA, MCBPA, DCBPA, and D12-DCBPA and 0.2, 0.5, and 2 ng/g for TCBPA, TeCBPA, NP, 4-n-NP, MCNP, and DCNP in decidua and chorionic villi samples) during 5 days. Identification of the target analytes was accomplished by comparing the retention time (within 2%) and the ratio (within 20%) of the two selected precursor ion-produced ion transitions with those of standards. To automatically correct for the losses of analytes during sample preparation and the matrix-induced change in ionization, and to compensate for the variations in instrument response from injection to injection, we used available internal standards (D16-BPA for BPA; D122,2′-DCBPA for MCBPA, DCBPA, TCBPA, and TeCBPA; 4n-NP for NP, MCNP, and DCNP) throughout the study. Limits of quantification (LOQs) of all target analytes except for BPA and NP were based on the peak-to-peak noise of the baseline near the analyte peak obtained by analyzing field samples (spiked with each analyte at 1 ng/mL) using 10 as the minimal value of signal-to-noise. The LOQs of BPA and NP were calculated as 10 times the standard deviation of the blank after subtraction of the blank mean. Standard injections were done every three sample injections, and ACN injections were done after each standard injection to monitor background contaminations. The extent of the signal suppression and enhancement in UPLC−ESI−MS/MS analysis was evaluated by adding 20 μL derivatized ACN spiked with 0.05 ng of each analyte into 180 μL of derivatized sample extracts. The matrix effects were calculated according to the following equation.
centrifuged at 4000 rpm for 10 min, and the MTBE layer was transferred to an 8 mL glass vial. Extraction from the residue was repeated twice more, and the organic layers were combined, concentrated to near dryness under a gentle stream of nitrogen, and redissolved in 1 mL of ACN for dansylation. A previously developed dansylation method for determination of fluorotelomer alcohols (FTOH) was used for dansylation of BPA, NP, and their chlorinated derivatives in this study with some modification.35 Briefly, 200 μL of 30 mg/ mL DNS and DMAP in DCM was added to the extract solution. The mixed solution was shaken vigorously for 1 min. The resulting mixture was kept at 65 °C for 60 min and then transferred to a 15 mL centrifuge tube, and then 1 mL of ultrapure water and 4 mL of n-hexane were added. After being vigorously shaken for 10 min, the organic layer was transferred to a clean glass vial. This procedure was repeated three times, and a total of 12 mL n-hexane was collected. The n-hexane extraction solution was loaded onto a silica cartridge (6 mL, 1 g, Waters), which was conditioned with 10 mL acetone, 10 mL DCM and 10 mL n-hexane. After the column was rinsed with 5 mL n-hexane/DCM (1:1, v/v), the analytes were eluted with 7 mL DCM/acetone (9:1; v/v). The solution was evaporated to dryness and reconstituted with 100 μL of ACN prior to UPLC−MS/MS analysis. UPLC−ESI−MS/MS Analysis. The UPLC apparatus was an Acquity Ultra Performance LC (Waters, Milford, MA). Target analytes were separated using a Waters Acquity UPLC BEH C18 column (100 × 2.1 mm × 1.7 μm). The column was maintained at 40 °C and a flow rate of 0.3 mL/min, and the injection volume was 5 μL. ACN and water containing 0.1% formic acid were chosen as the mobile phases. Gradient conditions were initiated with 60% ACN followed by a linear increase to 75% ACN in 0.5 min. After being increased to 80% in 6 min, ACN was increased to 95% in 0.5 min and then to 100% in 2 min and kept isocratic for 2 min. Mass spectrometry was performed using a Premier XE tandem quadrupole mass spectrometer equipped with a Z-Spray ionization (ESI) source and operated in the positive ion (PI) mode. The most abundant multiselected reaction monitoring (MRM) transitions, cone voltages, and collision energies are summarized in Table S1. Common MS parameters were as follows: capillary voltage, 3.2 kV (ESI+); source temperature, 120 °C; desolvation temperature, 450 °C; source gas flow, 50 L/h; and desolvation gas flow, 800 L/h. Quantification and Quality Control. All analytical procedures were checked for precision, reproducibility, blank contamination, linearity, and matrix effects. The recovery experiments were conducted by spiking a known concentration (0.05, 0.5, and 2 ng/g) of all target analytes into sample matrices, with subsequent passage through the entire analytical procedure (n = 3). To prevent possible specimen contamination, we washed all glassware materials used for sample processing, storage, and pretreatment with detergent, rinsed with abundant Milli-Q water, sonicated in an acetone bath for 1 h, and subsequently dried in a muffle furnace at 450 °C for at least 6 h. A field blank (n = 3) was conducted by adding 1 mL of pure water into an identical glass vial at the spot, which was kept and analyzed along with subject samples. In the analysis of plasma, procedure blanks (n = 3) were prepared by substitution of 0.5 mL Milli-Q water for samples and subjecting them to the entire analytical procedure, while in the analysis of decidua and chorionic villi, a total of four procedure blanks were analyzed. Procedure blanks were prepared from 3 mL ACN, followed by
⎛ peak area − peak area unspiked standard addition matrix effect = ⎜⎜ peak area working standard ⎝ ⎞ − 1⎟⎟ × 100% ⎠
Statistical Analysis. Data analysis was performed with SPSS, version 18.0. For log-normally distributed data, any concentration below the LOQ was substituted with a value equal to LOQ divided by 2 for the calculation and statistical analysis of the geometric mean (GM). The correlation between the concentrations of different analytes was identified by Pearson’s correlation coefficients. The statistical level of significance is p < 0.05.
■
RESULTS AND DISCUSSION Method Validation. It has been reported that ACN coupled with n-hexane, which can precipitate the proteins but extract less lipid, has been used to extract BPA and its chlorinated derivatives from human adipose samples.31 In the present study, a similar liquid−liquid extraction (LLE) method was adopted. Because there are strong matrix effects when the extracts of decidua and chorionic villi samples are analyzed by only a LLE cleanup method, additional steps are necessary to 972
DOI: 10.1021/acs.est.5b04130 Environ. Sci. Technol. 2016, 50, 970−977
Article
Environmental Science & Technology
Table 1. Limits of Quantification (LOQs) and Recoveries of Target Analytes in Decidua, Chorionic Villi Samples, and Plasma decidua and chorionic villi
plasma
recovery (%) ± RSD
recovery (%) ± RSD
analyte
LOQs (ng/g)
0.05 ng/g
0.5 ng/g
2 ng/g
LOQs (ng/mL)
0.05 ng/mL
0.5 ng/mL
2 ng/mL
BPA D16-BPA MCBPA DCBPA D12-DCBPA TCBPA TeCBPA NP 4-n-NP MCNP DCNP
0.03 − 0.02 0.04 − 0.3 0.3 1.92 − 0.15 0.24
106 ± 6 98 ± 11 108 ± 8 102 ± 11 97 ± 8 95 ± 7 95 ± 3 103 ± 8 99 ± 8 97 ± 2 87 ± 5
90 ± 12 85 ± 1 92 ± 20 91 ± 17 87 ± 9 98 ± 11 90 ± 9 90 ± 7 99 ± 2 103 ± 8 98 ± 15
82 ± 6 96 ± 2 72 ± 14 90 ± 18 93 ± 15 79 ± 17 87 ± 17 107 ± 16 104 ± 19 98 ± 3 97 ± 19
0.04 − 0.001 0.001 − 0.005 0.005 1.2 − 0.004 0.03
103 ± 9 84 ± 2 108 ± 1 95 ± 9 90 ± 12 86 ± 9 99 ± 1 91 ± 5 108 ± 10 96 ± 1 83 ± 5
101 ± 2 103 ± 17 118 ± 5 118 ± 7 102 ± 5 98 ± 5 93 ± 14 106 ± 2 89 ± 16 99 ± 15 84 ± 9
115 ± 5 108 ± 16 98 ± 7 93 ± 17 103 ± 11 101 ± 2 93 ± 3 111 ± 12 102 ± 3 96 ± 7 94 ± 13
BPA, bisphenol A; NP, nonylphenol; MCBPA, 4-chloro-BPA; DCBPA, dichloro-BPA (mixture of 2,6-dichloro-BPA and 2,6′-dichloro-BPA (1:0.25)); TCBPA, trichloro-BPA; TeCBPA, tetrachloro-BPA; MCNP, 2-chloro-NP; DCNP, 2,6-dichloro-NP.
Figure 1. UPLC−MS/MS MRM chromatograms of the ions selected for identification of analytes detected in samples of decidual tissues and chorionic villi. BPA, bisphenol A; NP, nonylphenol; MCBPA, 4-chloro-BPA; DCBPA, dichloro-BPA (mixture of 2,6-dichloro-BPA and 2,6′-dichloroBPA (1:0.25)); TCBPA, trichloro-BPA; TeCBPA, tetrachloro-BPA; MCNP, 2-chloro-NP; DCNP, 2,6-dichloro-NP.
eliminate potential interferences. In this study, freezing was used for the removal of lipid, which was precipitated after having been kept at −20 °C overnight and immediately centrifuged at −10 °C. Although the matrix effects were greatly improved, they were still as high as 68% (Table S2) when DCBPA and DCNP were analyzed. Thus, NH2 SPE cartridges were further tested for sample cleanup. Although 20:80 (v/v) MeOH/acetone, 50:50 (v/v) MeOH/acetone, and 80:20 (v/v) MeOH/acetone were used to elute the target chemicals from NH2 cartridges, 20:80 (v/v) MeOH/acetone was found to be the most effective for elution of all eight analytes with the lowest matrix interferences. Finally, a method using LLE coupled with freezing-lipid centrifugation and NH2 SPE was applied to prepare decidua and chorionic villi samples, with signal suppression rates less than 19% in all cases. A recovery test was conducted by spiking each target compound at different concentrations, followed by the entire pretreatment procedure (n = 3). The method recoveries of 8 target analytes were 83−108% with a relative standard deviation (RSD) of 1−12% for 0.05 ng/mL, 84−118% with a RSD of 1−
20% for 0.5 ng/mL, and 72−115% with a RSD of 2−19% for 2 ng/mL (Table 1). A previously developed dansylation LC-MS/ MS method, which demonstrated high sensitivity and low matrix effects in the analysis of BPA, NP and their chlorinated derivatives in food-contact papers (FCPs),26 was applied to the analysis of the same target analytes in decidua, chorionic villi, and plasma samples. As displayed in Table 1, the LOQs in plasma samples were 0.004−0.03 ng/mL for chlorinated NPs and 0.001−0.005 ng/mL for chlorinated BPAs, which are significantly lower than the LOQs of chlorinated BPAs in serum (0.05 ng/mL) reported in a previous paper.30 The method applied to the analysis of decidua and chorionic villi samples also showed high sensitivity, with LOQs of 0.02−0.3 ng/g, which are significantly lower than those reported in placental tissue samples (1.0−2.0 ng/g) or in adipose samples (0.5−3.0 ng/mL).31,32 BPA and NP were detected in the procedural blanks (0.02 ± 0.01 ng/mL for BPA and 0.82 ± 0.40 ng/mL for NP in plasma samples; 0.59 ± 0.01 ng/g for BPA and 5.88 ± 0.64 ng/g for NP in decidua and chorionic villi samples), and their LOQs were 0.04 ng/mL for BPA and 1.20 ng/mL for NP 973
DOI: 10.1021/acs.est.5b04130 Environ. Sci. Technol. 2016, 50, 970−977
Article
Environmental Science & Technology Table 2. Concentrations (ng/g) of Analytes Detected in Decidua and Chorionic Villi (n = 25) decidua
chorionic villi
analytes
detection frequency
mean (SD)
GM
range
detection frequency
mean (SD)
GM
range
BPA MCBPA DCBPA TCBPA TeCBPA NP MCNP DCNP
40% 76% 72% 20% 12% 40% 92% 84%
1.30 (2.26) 0.46 (0.83) 0.43 (0.76) 0.17 (0.27) 0.10 (0.26) 12.27 (29.20) 5.4 (12.51) 2.2 (3.29)
0.10 0.10 0.12 0.19 0.17 3.27 1.85 0.74
