Bisphenol A exposure alters placentation and causes preeclampsia‐like features in pregnant mice involved in reprogramming of DNA methylation of WNT2

ABSTRACT

Preeclampsia leads to adverse outcomes for pregnant women. Bisphenol A (BPA) is an environmental endocrine disruptor and has been shown to be positively associated with increased risk of preeclampsia in human studies. We investigated whether BPA exposure causes preeclampsia‐like features in pregnant mice and the associated underlying mechanisms. Experiments were performed in animal models and cell cultures. In pregnant mice, BPA‐exposed mice exhibited preeclampsia‐like features including hypertension, disruption of the circulation, and the placental angiogenesis biomarkers fms‐related tyrosine kinase 1 and placenta growth factor, and glomerular atrophy; urinary protein was not affected. These preeclampsia‐like features correlated with increased retention of smooth muscle cells and reduced vessel areas at the junctional zone of the placenta. In addition, there were disrupted expression of invasion‐related genes including increased tissue inhibitors of metalloproteinases, decreased metalloproteinases, and Wnt family member WJVT2/β‐catenin, which correlated with increased DNA methylation in its promoter region and upregulation of DNA methyltransferase (Dnmt)l. BPA exposure impeded the interaction between the human cytotrophoblast cell line, HTR‐8/SVneo, and endothelium cells. BPA exposure down‐regulated WNT2 expression, and elevated the DNA methylation of WNT2; these results were consistent with in vivo observations. Inhibition of DNMT in HTR‐8/SVneo cells resulted in reduced DNA methylation and increased expression of WNT2. Taken together, these data demonstrate that BPA exposure alters trophoblast cell invasion and causes abnormal placental vessel remodeling, both of which lead to the development of preeclampsia‐like features in pregnant mice. Our results suggest that this phenomenon involves the epigenetic reprogramming and down‐regulation of WNT2 mediated by DNMT1.—Ye, Y., Tang, Y., Xiong, Y., Feng, L., Li, X. Bisphenol A exposure alters placentation and causes preeclampsia‐like features in pregnant mice involved in reprogramming of DNA methylation of WNT2. FASEB J. 33,2732–2742 (2019). www.fasebj.org

ABBREVIATIONS

5-aza-dc

5-aza-2′-deoxycytidine;

5‐hmc

5‐hydrox‐ymethylcytosine;

5‐mc

5‐methylcytosine;

α‐SMA

α‐smooth muscle actin;

BPA

bisphenol A

CCK

cell‐counting kit

CpG

5′‐C‐phosphate‐G‐3′;

DNMT

DNA methylation transferase;

Fltl

fms‐related tyrosine kinase 1;

H&E

hematoxylin and eosin;

IHC

immunohistochemistry;

MMP

metalloproteinase;

MTC

Masson's trichrome;

P1GF

placenta growth factor;

SBP

systolic blood pressure;

TET

ten‐eleven translocation;

TIMP

tissue inhibitor of metalloproteinase;

WNT

Wnt family member

Preeclampsia is a pregnancy‐specific disease characterized by new‐onset hypertension and proteinuria after wk 20 of gestation and remains one of the major causes of adverse pregnancy and birth outcomes worldwide. Despite the severe consequences of preeclampsia for maternal and fetal health, the pathophysiological mechanisms are unclear. However, inadequate trophoblast invasion and remodeling of maternal spiral arteries early in pregnancy are thought to play vital roles in preeclampsia (1).

During early pregnancy in humans, trophoblasts must anchor and invade the maternal decidualized endometrium. The invasive trophoblasts destroy the maternal spiral arterial wall, replacing the endothelial cells and generating low‐resistance, large‐diameter vessels that promote uteroplacental blood flow to sustain fetal growth. Poor endovascular trophoblast cell invasion diminishes the remodeling and diameter of spiral arteries and allows for the retention of smooth muscle, which increases the vessel's contractility and vasoconstriction in preeclampsia. It is well established that uteroplacental cytoactive factors such as matrix metalloproteinases (MMPs), tissue inhibitors of MMPs (TIMPs) (2), canonical Wnt family member‐2 / β‐catenin (WNT2 / β‐catenin) signaling (3), and imbalanced proangiogenic/antiangiogenic factors of soluble fms‐related tyrosine kinase 1 (Flt1) and placenta growth factor P1GF (4) are associated with decreased vascular remodeling and trophoblast invasion of the myometrium.

Among risk factors for preeclampsia, environmental and genetic ones are considered equally important (5). Maternal exposure to the environmental chemical BPA may be an important risk factor for abnormal placentation (6–8) and subsequent pregnancy complications such as preeclampsia (9,10).

Bisphenol A (BPA), an endocrine‐disrupting chemical (11), is manmade, produced at high volumes, and widely used as a synthetic plasticizer (12). Humans are exposed to BPA primarily through the diet as a result of leaching from food and beverage containers. BPA is detectable in the urine, serum, and placental tissues of pregnant women (11). Epidemiologic studies have reported a positive association between maternal exposure to BPA and the risk of preeclampsia (9,10,13). BPA exposure also affects placental morphology and angiogenesis during early pregnancy in mice (7), and the WNT2 signaling pathway may be involved (7). Moreover, the placentas of BPA‐exposed mice exhibit altered global DNA methylation levels (6,14). Epigenetic regulation is vital during placental development, and abnormal epigenetic regulation in the placenta is associated with preeclampsia (15). Although mounting data suggest a positive correlation between BPA exposure and preeclampsia (6, 7, 9, 10, 13), evidence confirming a causal relationship is currently unavailable.

The objectives of the present study are to use an in vivo mouse model and an in vitro human cell model to establish a causal relationship between maternal BPA exposure and the risk of preeclampsia, and illuminate the underlying mechanisms. We hypothesize that BPA exposure induces preeclampsia‐like features in pregnant mice and the underlying mechanisms involve abnormal placental vessel remodeling, altered trophoblast cell invasion, and changes in relevant signaling pathways. To test our hypothesis, mice were exposed to BPA from embryonic day (E)7 to El7 at gradient concentrations from 0 to 400 µM. Preeclampsia‐like features were characterized by several measurements, including systolic blood pressure (SBP), urinary protein levels, renal and placental morphology, and the circulating/placental Flt‐1/PlGF ratio. Using both mice placenta tissues and a human cytotrophoblast cell line, HTR‐8/SVneo, we examined the placental vessel remodeling and trophoblast invasion‐related signaling pathways, including MMPs and WNT2, in response to BPA exposure. In addition, we further explored the role of methylation in WNT2 expression under BPA exposure.

MATERIALS AND METHODS

Chemicals

BPA (99% purity) was purchased from (239658; MilliporeSigma, Burlington, MA, USA), dissolved in 100% DMSO, and stored in glass containers at 4°C. The stock solution was appropriately diluted with drinking water or culture medium containing 0.1% DMSO.

Animals and experimental design

Eight‐week‐old ICR (CD‐1) mice were purchased from Jiesijie Laboratory (Shanghai, China). During the experiments, the mice were housed at 21–22°C in a 12‐h dark‐light cycle in a controlled room with free access to food and water. The mouse cages were polypropylene plastic, drinking bottles were glass, and the net cover and drinking pipes were stainless steel. Female and male mice were mated in a 2:1 ratio, and time was recorded as E0 when the vaginal plug appeared. Pregnant mice were randomly divided into 5 groups (n = 8 per group) at E7 and exposed to BPA through free access to drinking water containing 0, 0.4, 4, 40, or 400 µM BPA with reference to previous reports (16). Nonpregnant female mice were exposed to 400 µM BPA for 10 d. All animals were kept in the same housing conditions, and exposure levels of BPA were confirmed in all animals through blood level measurements. The route and timing of BPA exposure were carefully considered. First, although previous animal experiments involved a variety of BPA exposure patterns (8,17,18), we administrated BPA orally to the mice, instead of injection and subcutaneous implantation, because humans are exposed to BPA commonly by ingestion (19). We chose free access to drinking water instead of gastric gavage to alleviate the stress response in the mice. Second, because the timing of BPA exposure is important to target organ toxicity and mechanism of action and the fact that BPA exposure in early pregnancy (E0–E7) can lead to embryonic absorption in mice (8), we selected E7 to El7, a period during which implantation completes and placental development continues. This period is also equivalent to placental vascular remodeling in humans.

Urine samples were collected with methylbenzene‐treated, BPA‐free centrifuge tubes (Corning, Corning, NY, USA) from E16 to E17 (24 h total), and urine volumes were recorded. Samples were separated into aliquots and stored at −80°C until further analysis. Placental and kidney samples were collected after the mice were euthanized. A portion of placental and kidney tissues was fixed with 4% paraformaldehyde, and the other portion was snap frozen in liquid nitrogen and then stored at — 80°C. Maternal weight, litter size, fetal weight, and placental weight were examined. After we determined that the effective dose of BPA was 4 µM, 2 additional groups of mice were treated with 0 or 4 µM BPA and euthanized at E13, and the placentas were collected to determine the effects of BPA exposure on early placental development. All animal procedures were performed in accordance with the guidelines issued by Fudan University for the care and use of laboratory animals.

Measurement of urinary BPA and protein levels

Total free BPA concentrations in 24‐h urine samples were measured by liquid chromatography‐mass spectrometry (Acquity TOD; Waters, Milford, MA, USA) at Weipu Microspectrum Technology (Shanghai, China), as previously described (20). The limit of detection was 1 µg/L. Protein concentrations in 24‐h urine samples were quantified with a Bradford Protein Assay Kit (Beyotime Biotechnology, Beijing, China).

Blood pressure measurement in mice

Indirect SBP was measured in conscious mice by tail cuff plethysmography (BP2000; Visitech System, Apex, NC, USA). To prevent stress, all mice were habituated for 3 d (E13‐E15) before starting the measurement at E16. The mice were prewarmed to 30°C for 20 min in the detection room, arterial pressure was measured several times between 9:00 am and 12:00 pm, and pressure values were considered acceptable when similar values were obtained in ≥10 consecutive measurements.

Assessment of placenta and kidney E17 morphology

Fixed mice kidneys were embedded in paraffin, sliced into 5‐µm–thick serial sections, and stained with hematoxylin and eosin (H&E) or periodic acid‐Schiff (PAS) to assess glomerular morphology. Fixed mice placentas were stained with H&E or Masson's trichrome (MTC) to assess placental morphology. Placental sections were immunostained with antibodies against the cell proliferation marker Ki67, α‐smooth muscle actin (α‐SMA), and the trophoblast biomarker pan‐cytokeratin (all from Abeam, Cambridge, United Kingdom) overnight at 4°C to examine the cell proliferation of placental cells and maternal vascular remodeling, followed by washes and an incubation with the corresponding horseradish peroxidase–, FITC 488‐(green)–, and AlexaFluor 647–conjugated (red) secondary antibodies (Wuhang Google Biotechnology, Wuhan, China; http://www.servicebio.cn/index.php). Pictures were taken with a positive microscope (Nikon, Tokyo, Japan) or confocal microscope (Leica Microsystems, Buffalo Grove, IL, USA) as necessary from 4 phenomena of the top, bottom, left, and right sides of each placenta. The results of the immunohistochemistry (IHC) were analyzed by 2 individuals blind to sample location by calculating an immunoreactive score using methods Ogawa et al. (21). The intensity was scored as follows: 0, negative; 1, weak; 2, moderate; and 3, intense. The frequency of positive cells was defined as follows: 0, none; 1, 1–10%; 2, 11–50%; 3, 51–80%; and 4, 80–100%. Scores were calculated by multiplying the intensity by the frequency. The results of immunofluorescence were analyzed using Image‐Pro Plus 6.0 software (Media Cybernetics, Rockville, MD, USA).

Measuring gene expression in E13 placental tissues

Quantitative RT‐PCR

Total RNA was extracted from mouse placentas with an RNeasy Mini Kit (Qiagen, Hilden, Germany). The cDNA templates were synthesized from 1 µg total RNA by reverse transcription using the Reverse Transcription (RT) Reagent Kit (Qiagen). Amplification reactions were conducted in a 7900 Real‐Time PCR System (Thermo Fisher Scientific Waltham, MA, USA) using a Quanti‐Nova Sybr Green PCR Kit (Qiagen). Gene expression levels were normalized to GAPDH, set relative to controls, and fold changes were calculated by 2^−ΔΔCt . The primers used to amplify MMP2, MMP9, TIMP1, TIMP2, WNT2, and β‐catenin mRNAs are listed in the Supplemental Table S1.

PCR array

The Mouse Epigenetic Chromatin Modification Enzymes RT² Profiler PCR Array profiles the expression of 84 key genes. These 84 genes encode enzymes known or predicted to modify genomic DNA and histones that regulate chromatin accessibility and therefore gene expression. Using this PCR Array, we profiled the 84 genes including DNA (cytosine‐5)‐methyltransferase (DNMT)1, ‐3a, and ‐3b, and ten‐eleven translocation (TET)1, ‐2, and ‐3 in placental tissues at E13. PCR arrays were performed in duplicate, and verification experiments were performed by Western blot in triplicate. Data obtained from the PCR Array were analyzed with the official Qiagen website. Dissociation curves and hierarchical clustering are shown in Supplemental Figs. S1 and S2.

Bisulfite sequencing for measuring DNA methylation

Genomic DNA (1000 ng) was harvested from each placenta and modified by bisulfate conversion with the EpiTect Fast DNA Bisulfite Kit (Qiagen) according to the manufacturer's protocol. Bisulfate‐labeled DNA products were amplified by PCR reaction using Hot Star Taq DNA polymerase (Takara, Tokyo, Japan). PCR products were confirmed by agarose gel electrophoresis and further purified with a QiaQuick PCR Purification Kit (Qiagen). Purified PCR products were cloned with the PMD19‐T Vector Cloning Kit (Takara). Ten individual colonies from each sample were amplified and sent to Biosune Technology Corp. (Shanghai, China) for sequencing. The 5′‐cytosine‐phosphate‐guanine‐3′ (CpG) islands in each gene were identified with Methyl Primer Express software v.2.0 (Thermo Fisher Scientific). The details of the primers are listed in Supplemental Table S2.

Measuring protein expression in E13 placental tissues

Western blot analysis

Total proteins were extracted with RIPA lysis buffer supplemented with PMSF (Beyotime Biotechnology) per a standard protocol. Western blot analysis was performed as described by Suzuki et αl. (22) with modifications, in brief, proteins (10–20 µg) were loaded on 10 or 7.5% Tris‐glycine gels (Beyotime Biotechnology) and transferred to PVDF membranes (Milli‐poreSigma). Membranes were incubated with primary antibodies (all from Abcam unless otherwise noted) against Fltl (ab32152), P1GF (abl96666), DNMT1 (abl3537), DNMT3a (ab27725), DNMT3b (ab2601‐S), TET1 (abl57004), TET‐2 (abl35087), TET‐3 (abl53724), WNT2 (27214‐1‐AP; Proteintech, Rosemont, IL, USA), and β‐catenin (ab6302) at 4°C overnight followed by a 1‐h incubation at room temperature with the corresponding secondary antibody (Wuhang Google Biotechnology). Glyceraldehyde dehydrogenase protein (6004‐1‐Ig; Proteintech) levels were used as the control for equal protein loading. Gray values representing protein levels were analyzed with Image‐J software (National Institutes of Health, Bethesda, MD, USA; https://imagei.nih.gov/ii/index.html).

IHC staining

IHC staining was used to examine 5‐methylcytosine (5‐mc) (A‐1014‐100) and 5‐hydroxymefhylcytosine (5‐hmc) (A‐1018‐100; both from EpiGentek, Farmingdale, NY, USA) levels in placental tissues, as previously described.

ELISA

Levels of 5‐mc in placental extracts were further determined with MefhylFlash Global DNA Methylation (5‐mc) ELISA Kits (P‐1030‐48; EpiGentek), according to the manufacturer's protocol. The detection limit of the kit is as low as 0.05% of methylated DNA from 100 ng of input DNA with no cross‐reactivity toward unmethylated cytosines and no cross‐reactivity toward hydroxymethylcytosine in the sample DNA.

Culture of HTR‐8/SVneo cells and treatment conditions

HTR‐8/SVneo cells, derived from human first‐trimester extravillous trophoblast cells, were cultured in DMEM/F12 (GE Healthcare, Chicago, IL, USA) supplemented with 10% charcoal‐stripped fetal calf serum (Thermo Fisher Scientific) and BPA (0, 0.1,1,10, or 100 µM) for 14 d to establish an in vitro exposure model. This gradient concentration was chosen according to previous reports (23). During the 14 d of incubation, cells were passaged every 3 d in the same proportions. The viabilities of HTR‐8 / SVneo cells that had been treated with BPA for 14 d were measured with a Cell‐Counting Kit‐8 (CCK‐8; Dojin, Tokyo, Japan) according to the manufacturer's instructions. In brief, after 12 d of BPA treatment, cells were trypsinized and planted in 96‐well plates (5000 cells/well) for 48 h. Absorption of 450 nm was determined after incubation with CCK‐8 solution (10 µl/well) for 40 min.

Cell lysates were harvested, and the levels of WNT‐2 and β‐catenin proteins were measured by Western blot analysis. We further explored the role of DNA methylation in regulating WNT2 expression by treating HTR‐8/SVneo cells with 5 εM 5‐aza‐2′‐deoxycytidine (5‐aza‐dc; MilliporeSigma), a DNA methylation inhibitor, for 48 h. The expression of the DNA methylation enzymes (DNMT1, ‐3a, and ‐3b) WNT‐2 and β‐catenin were analyzed by Western blot analysis to evaluate the effects of 5‐aza‐dc. The methylation of the WNT2 in these cells was examined as previously described. We did not examine the methylation of the β‐catenin DNA in HTR‐8/SVneo cells, because the human gene does not contain CpG islands.

Interaction between HTR‐8/SVneo and HUVECs

To study the ability of trophoblasts to invade and replace endothelial cells, we evaluated the interaction between HTR‐8 / SVneo cells and HUVECs. HUVECs were cultured in Endothelial Cell Medium (ScienCell, Carlsbad, CA, USA) supplemented with 5% fetal bovine serum and 1% endothelial cell growth factors. Interactions between HTR‐8/SVneo cells and HUVECs were studied with an in vitro model of 3‐dimensional dual cell coculture. In this analysis, HUVECs, labeled with a green tracker (Thermo Fisher Scientific), were cultured on growth factor‐reduced Matrigel‐coated (Corning) 96‐well plates (1 × 10⁴ cells/well) for 2 h, to form tube‐like structures. HTR‐8/SVneo cells were pre‐treated with different concentrations of BPA for 14 d and then labeled with a red tracker (Thermo Fisher Scientific). HTR‐8/ SVneo cells (1 × 10⁴ cells/well) were then added to the HUVEC tube structures for 6 h. Interactions were analyzed by ratios of red/green fluorescence as measured by Image‐Pro Plus 6.0.

Statistical analysis

All data were obtained from at least 3 biologic replicates, except for the PCR array analysis (performed in duplicate). Data are expressed as means ± sem. Differences between control and treatment groups were determined with a Student's t test or ANOVA. Values of P < 0.05 were considered statistically significant.

RESULTS

BPA exposure induced preeclampsia‐like features in pregnant mice

BPA exposure was confirmed by quantifying 24‐h urinary BPA levels (Fig. 1A ), and it was found that levels increased in a BPA dose‐dependent manner. Statistically significant urinary BPA levels were recorded in mice exposed to BPA at doses ≥4 µM compared with levels in control mice that were not exposed to BPA. Pregnant mice exposed to ≥4 µM BPA displayed significantly elevated SBP (4 µM BPA vs. control: 131.5 ± 7.66 vs. 104.9 ± 3.78 mmHg) (Fig. 1B ), but nonpregnant mice were not affected by exposure to 400 µM BPA. Twenty‐four‐hour total urine protein levels at El 7 in BPA‐exposed groups were not significantly different from that of the control group (Fig. 1C ). As indicated, a dose of 4 µM BPA was the lowest dose to elevate blood pressure in the present study and was thus the dose chosen for further analysis. A histopathologic analysis of renal tissues collected from control and BPA (4 µM)‐treated pregnant mice revealed no significant changes in the control group (Fig. 1D ). Conversely, in the BPA‐exposed group, glomerular atrophy, narrowed glomerular cystic cavities, and glomerular capillary loops occupying the whole glomerulus were observed in various degrees (Fig. 1D ). Placental and circulating Flt1 (Fig. 1E, G ) levels and the Fltl :P1GF ratio (Fig. 1F, I ) were significantly elevated in response to BPA exposure (4 µM) although changes in P1GF levels were not statistically significant. Maternal body weight, litter size, mean fetal weight, mean placental weight, and placenta/fetus weight ratios did not show significant variation in response to BPA (Fig. 1J‐N ). Preterm birth was not observed in either the BPA‐exposed group or the control group.

Figure 1.

BPA‐exposed pregnant mice exhibited preeclampsia‐like features and impaired placental angiogenesis. A) Twenty‐four‐hour total urinary free BPA levels rose at E17 with increased dose of BPA (n = 3). *P < 0.05, ***P < 0.001 vs. 0 µM BPA. B) SBP was elevated at E16 after BPA exposure (n = 8). **P < 0.01, ***P < 0.001 vs. 0 µM BPA. C) Twenty‐four‐hour total urinary protein level at E17 was not affected vs. 0 µM BPA (n = 5). D) Renal morphology assessed at E17 showed constricted glomerular (H&E) and thickened basement membrane (PAS). Scale bars, 20 µm. E) BPA exposure (4 µM) altered expression of the Fit and P1GF proteins in mouse placenta at E17 (n = 3). F) Gray values were quantified (n = 3). *P < 0.05, 4 µM vs. 0 µM BPA. G, H) BPA exposure (4 µM) altered serum levels of sFlt‐1 (G) and P1GF (H) at El7, as quantified by ELISA (n = 3). *P < 0.05 vs. 0 µM BPA. I) The sFlt‐1 :P1GF ratio was increased after BPA exposure (n = 3). *P < 0.05 vs. 0 µM BPA. J–N) BPA exposure (4 µM) did not alter the maternal weight (J), litter size (K), mean fetal weight (L), mean placental weight (M), or placentalfetal weight ratio (N) (n = 8). *P < 0.05 vs. 0 <M BPA. O) Placenta morphology assessed at E17 showed abnormal lesions in the placental labyrinth (arrowheads; H&E) and collagen deposition (arrowheads; MTC). Scale bars, 50 µm. P) An analysis of the implantation site at E13 showed elevated retention of smooth muscle cells in decidual vessels immunostained for α‐SMA. Scale bars, 100 µm. Q) Vessel area was measured in placental cross‐sections at E13 and was reduced by BPA exposure. Four quadrants in each placenta were estimated (n = 5). *P < 0.05 vs. 0 µM BPA. R) BPA exposure did not affect cell proliferation in the placenta of El3 embryo sections, as visualized by Ki67 staining. Scale bars, 100 µm. S) Scores of Ki67 IHC were quantified vs. 0 µM BPA.

Impaired placental structure and reduced placental vessel remodeling

Placentas from BPA‐exposed mice, but not control mice, exhibited abnormal lesions upon H&E staining (Fig. 1O ) that were confirmed to be fibrous matrix deposits by MTC staining (Fig. 1O ). At the maternal side of junctional zone, α‐SMA‐labeled vascular smooth muscle cells were retained, and the number of pan‐ cytokeratin–labeled trophoblast cells was decreased in vessels (Fig. 1P ); moreover, the vessel area was significantly reduced (68.26 ± 1.20 vs. 164.4 ± 28 µm²) (Fig. 1Q ). Placental cell proliferation was not different between control animals and 4 µM BPA‐treated mice, as evidenced by immunostaining for the cell proliferation marker Ki67 (Fig. 1R, S ).

BPA exposure altered the signaling pathways of MMPs and WNT2

In 4 µM‐BPA‐exposed mice, the trophoblast invasion‐related gene profile was disrupted; there were increased levels of transcripts for the invasion‐inhibiting genes TIMP1 and ‐2 and decreased levels of transcripts for the invasion‐mediating genes MMP2 and ‐9 (Fig. 2A ), WNT2 (Fig. 2B ), and β‐catenin (Fig. 2C ). There were also decreased levels of WNT‐2 and β‐catenin proteins (Fig. 2D ) in mouse placental extracts of BPA‐exposed vs. non‐exposed mice, which implies that BPA plays a role in the suppression of trophoblast cell invasion.

Figure 2.

BPA altered trophoblast cell invasion–related signaling pathways, DNMT1 expression, and DNA methylation of the WNT2 gene in mouse placenta at El3. A) BPA exposure (4 µM) altered the expression of the MMP2 and ‐9 and TIMP1 and ‐2 genes in the placenta at El3 (n = 3). *P < 0.05 vs. 0 µM BPA. B, C) BPA exposure reduced the expression of WNT2 (B) and β‐catenin (C) genes in the placenta at El3 (n = 3). *P < 0.05 vs. 0 µM BPA. D) Reduced levels of the WNT2 and β‐catenin proteins in the placenta at E13 (n = 3). E, F) Altered levels of DNMT and TETproteins in the placenta at E13 (n = 3). G) Altered immunostaining for 5‐mc and 5‐hmc in the placenta at E13. Scale bars, bottom: 200 µm; top right: 100 µm. H, I) Scores for 5‐mc and 5‐hmc immunohistochemistry were quantified (n = 5). ***P < 0.001 vs. 0 µM BPA. J) BPA elevated total 5‐mc levels in the placenta at El3, as quantified by ELISA (n = 5). *P < 0.05 vs. 0 µM BPA. K) BPA elevated the DNA methylation in promoter of the mouse WNT2 gene of 22 CpG sites. (Each circle shows the mean methylation of an individual CpG site; open circle, 0% methylation; rightmost circle and percentage, mean total methylation per sample) (n = 5). L) Total percentage of DNA methylation of CpG site of WNT2 gene was increased after BPA exposure (n = 5). *P < 0.05 vs. 0 µM BPA.

We then examined whether BPA exposure regulates these signaling pathways through epigenetic mechanisms. In the initial screening, using a PCR array analysis of 84 genes related to epigenetic modifications in mouse placenta, BPA exposure up‐regulated the expression of 8 genes and down‐regulated (≥ 2‐fold) the expression of 2 genes (Supplemental Table S3) which involve histone methylation and acetylation, protein phosphorylation, and DNA methylation. Among the 10 regulated genes, the change in DNMT1 expression was the most remarkable (fold changes = 2.36) (Supplemental Table S3). We further confirmed that BPA exposure significantly elevated the levels of DNMT1 protein in placenta, whereas levels of DNMT3a and ‐3b (Fig. 2E ), hydroxymethyl transferases (TET1, ‐2, and ‐3) (Fig. 2F ), and 5‐hmc (Fig. 2G, I ) were not affected. Furthermore, elevated levels of 5‐mc (Fig. 2G, H, J ) indicated an elevated methylation level in the placental tissues of BPA‐exposed mice.

After confirming that BPA exposure regulated the expression of methylation enzymes in the mouse placenta, we examined whether WNT2 expression was affected by BPA exposure through methylation reprogramming. The mean methylation level of WNT2 was increased by ∼50% in the BPA‐exposed group compared with that in the control group (15.12 ± 1.68 vs. 7.88 ± 2.35%) (Fig. 2K, L ), and the main affected CpG sites were CpG1, ‐2, ‐3, and ‐4 (Fig. 2K ). The methylation of β‐catenin DNA was not affected by BPA exposure (Supplemental Fig. S3A, B).

BPA treatments and WNT2 signaling in HTR‐8/SVneo cells

BPA treatment did not alter viability of HTR‐8/SVneo cells until an exposure level of 100 µM was reached, as evidenced by the CCK‐8 test (100 µM BPA vs. the control: 14.63 ± 2.63 vs. 99.98 ± 21.87%) (Fig. 3A ), which indicates possible cytotoxicity.

Figure 3.

BPA decreased WNT2 and β‐catenin expression by up‐regulating DNA methylation in HTR‐8/SVneo cells; BPA impaired HTR‐8/SVneo/endothelial cell interaction. A) The viability HTR‐8/SVneo cells treated with BPA was detected with a cell‐counting assay. Data were from 3 independent experiments. *P < 0.05 vs. corresponding 0 µM BPA. B) BPA exposure altered expression of DNMT proteins in HTR‐8/SVneo cells. C) Gray values were qualified from 3 independent experiments. **P < 0.01, ***P < 0.001 vs. corresponding 0 µM BPA. D) BPA elevated the DNA methylation of 23 CpG sites in the promoter region of the human WNT2 gene in HTR‐8/SVneo cells. (Each circle shows the mean methylation of an individual CpG site; open circle, 0% methylation; rightmost circle and percentage, mean of total methylation per sample). E) The total percentage (%) of 23 methylated CpG sites of human WNT2 gene. Data were from 3 independent experiments. **P< 0.01, ***P< 0.001 vs. corresponding 0 µM BPA. F) BPA reduced levels of WNT2 and β‐catenin proteins in HTR‐8/SVneo. G, H) Gray values were qualified. Data were from 3 independent experiments. **P< 0.01, ***P< 0.001 vs. corresponding 0 µM BPA. I) Treatment with 5‐aza‐dc altered levels of DNMT, WNT2, and β‐catenin proteins in HTR‐8/SVneo cells. J, K) Gray values were qualified from 3 independent experiments. *P < 0.05, **P < 0.01 vs. corresponding 0 µM BPA. L) Treatment with 5‐aza‐dc reduced the DNA methylation of 23CpG sites in the promoter region of the human WNT2 gene in HTR/SVneo‐8 cells. Each circle shows the mean methylation level of an individual CpG site. M) The total percentage of methylation levels in 23 CpG sites of the human WNT2 gene was analyzed. Data are from 3 independent experiments. **P < 0.01 vs. corresponding control. N) BPA exposure attenuated trophoblast‐endothelial cell interaction: HTR‐8/SVneo cells (red) and HUEVC cells (green). Scale bars, 200 µm. O) Quantification of trophoblast‐endothelial cell interaction by measuring the red and the green fluorescence portion. Data were from 3 independent experiments. ***P < 0.001 vs. corresponding control.

In HTR‐8/SVneo cells, exposure to >1 µM BPA up‐regulated DNMT1 expression (Fig. 3B, C ), whereas the expression of DNMT3a and ‐3b were not affected (Fig. 3B ). The mean methylation level of the WNT2 gene was elevated in cells treated with ≥0.l µM BPA (76.70 ±1.5 vs. 44.5 ± 1.4%) (Fig. 3D, E ), and the main affected CpG sites were CpGl, ‐2, ‐8, ‐9, ‐10, ‐11, and ‐14 (Fig. 3D ). Meanwhile, levels of the WNT‐2 and β‐catenin proteins were decreased in cells treated with ≥10 µM BPA (Fig. 3F–H ). Overall, BPA exposure increased DNA methylation and reduced the expression of WNT‐2 and β‐catenin in HTR‐8 /SVneo cells, consistent with the results obtained in mice.

We treated HTR‐8/SVneo cells with the DNA methyl transferase inhibitor 5‐aza‐dc to establish a direct role for DNMT1 in regulating WNT2 and β‐catenin expression. Inhibition of DNA methyl transferases effectively reduced the methylation of the WNT2 gene (5‐aza‐dc vs. controls: 83.9 vs. 66.5%) (Fig. 3L, M) and the main affected CpG sites were CpGl, ‐2, ‐4, ‐9, ‐10, and ‐14 (Fig. 3L ). As expected, levels of the WNT2 and β‐catenin proteins increased (Fig. 3I–K ).

BPA impeded interaction between HTR‐8/SVneo and HUVECs

BPA exposure impaired the interaction between HTR‐8 / SVneo and HUVECs at ≥ 1 µM BPA as revealed by a lower ratio of HTR‐8/SVneo/HUVECs in the tube‐like structures (1 µM BPA vs. control: 23.06 ± 2.4 vs. 44.67 ± 2.54%) (Fig. 3N, O ). Results were consistent with those obtained in mice.

DISCUSSION

To our knowledge, this study is the first to demonstrate preeclampsia‐like features in a mouse model after prenatal oral exposure to BPA. The development of preeclampsia occurs in 2 stages: it is initiated by reduced placental perfusion resulting from insufficient spiral artery remodeling and then impaired placentation causes systemic pathophysiological changes in the maternal circulation. This study showed changes in maternal circulation including elevated SBP, increased Flt1/PlGF ratios, and kidney damage; however, urinary protein levels were not affected in mice exposed to BPA. BPA exposure reduced the invasion of trophoblasts in vessels of placental sections in mice. The mouse is not an ideal model in which to study placental vascular remodeling, as the trophoblast cells invade only the maternal side of the junctional zone. Therefore, we further investigated the trophoblast invasion after BPA exposure by using a human placental line. We found that BPA exposure inhibited the ability of HTR‐8/SVneo cells to invade and replace endothelium, indicating that BPA exposure in humans may lead to poor uterine artery modification by trophoblasts. Moreover, BPA exposure in a human cytotrophoblast cell line negatively affected cytotrophoblast invasion. This suggests a possible molecular mechanism by which BPA exposure down‐regulates WNT2 expression by inducing hypermethylation of the WNT2 gene via DNMT1.

Physiologic cytotrophoblast invasion must be tightly regulated to ensure that the depth of invasion proceeds to the appropriate extent. Shallow invasion and insufficient vessel remodeling are characteristics of preeclampsia. Impaired trophoblast invasion and placental vessel remodeling were observed in our BPA exposure–induced preeclampsia‐like model. Immunofluorescence staining of cytotrophoblasts and smooth muscle cells in mice placentas revealed that BPA exposure inhibited the trophoblast invasion, leading to vascular smooth muscle retention and subsequent reduction in decidual vascular areas, consistent with previous findings that BPA exposure impairs placental labyrinth and spongiotrophoblast layers in mice (7,24). BPA did not alter the proliferation of placental cells in the BPA‐induced preeclampsia model in our study. In contrast, at much higher exposure levels (50mg/kg/d) than the 4 µM (corresponding to 182.5 µg/kg/d) used in the present study, BPA exposure induced placental cell degeneration and necrosis in mice (7), implying that high concentrations of BPA exert obvious toxic effects on placental cells.

Accumulating evidence has suggested a role for BPA in defective biologic processes in trophoblasts in vitro. In a study by Spagnoletti et al. (25), l0^–5 µM BPA inhibited HTR‐8/SVneo cell migration and invasion but did not affect trophoblast/HUVEC interactions after 48 h. However, in the present study, 1 µM BPA impaired trophoblast/HUVEC interactions after 14 d of exposure, indicating that either chronic or higher dose exposures may be needed to observe a BPA‐related effect. Moreover, another study showed that 24‐h exposure to 10 µM BPA reduced the viability of a trophoblast cell line, BeWo (26), but we did not observe reduced HTR‐8/SVneo cell viability at doses <100 µM BPA. In primary human trophoblasts, a 24‐h exposure to 0.02 µg/ml BPA increased cell apoptosis and necrosis (27). These results reveal different levels of susceptibility to BPA exposure in these cells. We speculate that this discrepancy is related to the different levels of syncytialization among the cells. Results from one recent study consistently showed that BPA exposure (<50 µM) did not alter the cell viability of HTR‐8/SVneo cells (24). However, the finding of enhanced cell invasion and migration after BPA exposure is inconsistent with the findings reported by Spagnoletti et al. (25). We suspect that the use of different culture conditions account for this discrepancy. The evidence, including our results, confirms the adverse effects of BPA on trophoblast invasion and suggest that this is the underlying biologic mechanism by which BPA exposure induces preeclampsia‐like features in mice. This notion is supported by both systematic and molecular examinations of our model. In the present study, pregnant mice treated with BPA developed comprehensive diagnostic criteria for preeclampsia including elevated SBP, increased Flt1:PlGF ratios, and impaired kidney glomerular morphology, but proteinuria remained unaffected. Although proteinuria is a common sign and risk factor of kidney impairment, the level of albuminuria does not completely reflect the severity of renal disease. For example, when proteinuria is less frequent, it is possible that the renal lesion is improved or worsened by glomerular fibrosis and reduced filtration protein, which was observed in the current study. Thus, proteinuria is not essential for the diagnosis of preeclampsia (28).

Nonpregnant female mice treated with BPA did not develop preeclampsia‐like features in our study. In contrast, Saura et al. (29) reported elevated blood pressure in nonpregnant mice treated with 0.4 µM BPA for 30 d. We speculate that the difference between our findings is related to the duration of exposure, implying that short‐term BPA exposure in a nonpregnant state is insufficient to cause the endothelial dysfunction and vascular constriction that lead to hypertension, whereas short‐term BPA exposure during pregnancy causes elevated blood pressure, probably because of the presence of the placenta.

These data reveal pregnancy‐specific effects of BPA on preeclampsia‐like features in mice, strongly suggesting that the effects are mediated by placental alterations. Indeed, we observed unbalanced placental angiogenesis and alterations in the expression of antiangiogenesis factors in BPA‐exposed pregnant mice, as manifested by the elevated circulating and placental Flt1:PlGF ratio and impaired placental vessel remodeling. In addition, BPA exposure is associated with many other pregnancy complications related to abnormal placentation, including spontaneous preterm birth (30), small size for gestational age (31), and loss of pregnancy (32). Further studies are warranted to evaluate the effects of BPA on other placental disorders, including preterm birth, recurrent pregnancy loss, and fetal outcomes, such as intrauterine growth restriction, even though we did not observe these phenomena in our study.

The molecular evidence of interest in our study was the observation of dysregulated signaling pathways critical for trophoblast invasion after BPA exposure. WNTs are a family of secreted glycoproteins with diverse, vital roles in development and human trophoblast invasion (33, 34), partially by inducing MMP2 and ‐9 expression (35). Reduced WNT2 expression impairs trophoblast invasion and correlates to an increased risk of preeclampsia (3). In our mouse model of BPA‐induced preeclampsia‐like features, we observed reduced expression of WNT2, β‐catenin, MMP2, and MMP9 and elevated expression of the TIMP1 and ‐2 mRNAs in placentas, strongly suggesting that BPA inhibits trophoblast invasion through canonical WNT signaling. WNT2 expression is down‐regulated by the reprogramming of DNA methylation in the human placenta (36). In our study, we identified a negative correlation between DNA methylation and expression of WNT2 in mouse placenta and HTR‐8/SVneo cells in response to BPA exposure. Epigenetic modifications, regulated by corresponding enzymes, include DNA methylation, histone modification, and phosphorylation. According to the results of a PCR array analysis of epigenetic modification enzymes, the expression of DNMT1 was altered to the greatest extent in the BPA‐induced preeclampsia‐like mouse model, whereas the levels of DNMT3a and ‐3b and TET1, ‐2, and ‐3 were not affected. This finding indicates a role for DNMT1‐mediated methylation reprogramming, which is manifested in the placentas of women with preeclampsia (37). Using HTR8/ SVneo cells, we confirmed that BPA exposure down‐regulates WNT signaling by increasing DNA methylation in the promoter region of the WNT2 gene through DNMT1. In a previous study, BPA was shown to inhibit the WNT/β‐catenin pathway in neural stem cells (38). However, BPA has also been reported to activate β‐catenin in trophoblasts of the labyrinthine and spongiotrophoblast layers of the mouse placenta after exposure to 0.5 mg/kg per day from E1 to Ell (17). A possible explanation for this difference from our findings may be the different exposure and observation windows.

BPA has been shown to activate estrogen receptors (39), producing effects similar to those induced by natural estrogens (40,41). A relationship between estrogen receptors and metabolites and preeclampsia has been described, but little is known about the contribution of these estrogen receptors or estrogen metabolites to preeclampsia mediated by BPA exposure (42–45). It has been demonstrated that women with preeclampsia have lower estrogen and disrupted estrogen profiles (43); we speculate, therefore, that it is unlikely that this estrogen disruption is the underlying mechanism in BPA‐induced preeclampsia. However, BPA could alter estrogen‐dependent normal physiology by disrupting hormonal homeostasis (41, 46). More studies are warranted to further explore this aspect.

In the present study, we found that exposure to 4 µM BPA significantly induced development of preeclampsia‐like features in pregnant mice, which corresponds to ∼2500 ng of BPA in a 24‐h urine sample. This value is similar to the exposure level found in women with preeclampsia (1.5 ng/ml) (9), corresponding to the total urine BPA of ∼2250–3000 ng/24 h in 1500–2000 ml of total urine volume for 24 h. Thus, the levels of BPA we tested in this study represent the normal exposure in pregnant women. Considering that the major method of exposure to BPA is through diet, a change in behavior, such as switching to BPA‐free products, could reduce BPA exposure and therefore risk of preeclampsia. Understanding the role of BPA exposure in preeclampsia and placentation not only provides insight into disease prevention but also helps to answer the questions about the reproductive toxicities of BPA.

The strength of this study is the comprehensive evaluation of our observations. However, there are some limitations to our study. First, total BPA intake could not be assessed, as daily water intake was not measured, thus we could not calculate the exact BPA exposure level. However, 24‐h urinary BPA concentrations were measured to evaluate exposure level and we found that urinary BPA levels rose progressively with increasing doses of BPA, confirming that our methods were sufficient to distinguish BPA exposure levels. Second, we were not able to analyze the long‐term effects of BPA exposure on pregnancy outcomes in our mouse model, which warrants future research.

In summary, BPA exposure during pregnancy impeded trophoblast invasion and placental spiral artery remodeling, which led to preeclampsia‐like features in mice. This occurs through mechanisms involved in down‐regulation of MMPs and up‐regulation of TIMPs, as well as reduced WNT signaling through increased DNA methylation, mediated by DNMT1, in the promoter region of the WNT2 gene. Our study provides new evidence for a causal relationship between BPA exposure and preeclampsia. Therefore, BPA exposure should be reduced or avoided during pregnancy to reduce the risk of preeclampsia and improve maternal and infant outcomes.

AUTHOR CONTRIBUTIONS

Y. Ye designed and performed the research, analyzed data, and wrote the manuscript; Y Tang analyzed the data; Y. Xiong designed the research; L. Feng was involved in study design and contributed significantly to the construction of this manuscript and discussion; and X. Li designed the research and revised the manuscript.

Supporting information

Supplementary Material 1

Ye, Y. , Tang, Y. , Xiong, Y. , Feng, L. , Li, X. Bisphenol A exposure alters placentation and causes preeclampsia-like features in pregnant mice involved in reprogramming of DNA methylation of WNT2. FASEB J. 33,2732–2742 (2019). www.fasebj.org