Bisphenol A Analogues Suppress Spheroid Attachment on Human Endometrial Epithelial Cells through Modulation of Steroid Hormone Receptors Signaling Pathway

Bisphenol A (BPA) is a well-known endocrine disruptor, widely used in various consumer products and ubiquitously found in air, water, food, dust, and sewage leachates. Recently, several countries have restricted the use of BPA and replaced them with bisphenol S (BPS) and bisphenol F (BPF), which have a similar chemical structure to BPA. Compared to BPA, both BPS and BPF have weaker estrogenic effects, but their effects on human reproductive function including endometrial receptivity and embryo implantation still remain largely unknown. We used an in vitro spheroid (blastocyst surrogate) co-culture assay to investigate the effects of BPA, BPS, and BPF on spheroid attachment on human endometrial epithelial cells, and further delineated their role on steroid hormone receptor expression. We also used transcriptomics to investigate the effects of BPA, BPS, and BPF on the transcriptome of human endometrial cells. We found that bisphenol treatment in human endometrial Ishikawa cells altered estrogen receptor alpha (ERα) signaling and upregulated progesterone receptors (PR). Bisphenols suppressed spheroid attachment onto Ishikawa cells, which was reversed by the downregulation of PR through PR siRNA. Overall, we found that bisphenol compounds can affect human endometrial epithelial cell receptivity through the modulation of steroid hormone receptor function leading to impaired embryo implantation.


Introduction
Bisphenol A (BPA), a well-known endocrine disruptor, is widely used in consumer products including plastics, paper bags, baby bottles, food cans, dental sealants, and thermal receipts [1]. Bisphenol A is ubiquitously found in air, dust, sewage leachates, and water. As a consequence, humans can be exposed to BPA through diet, inhalation of dust, and dermal contact [2]. As early as in the 1930s, BPA was found to have estrogenic effects on the female reproduction system [3]. It has been shown to bind to estrogen receptors and regulate gene expression [4]. In the past two decades, many studies have reported on the adverse effects of BPA, including reproduction, development, metabolic diseases, and the immune system in humans and laboratory animals [5][6][7]. As a result, several countries including Norway, Denmark, Germany, France, and the US have restricted the use of BPA in consumer products. They have been replaced with substitutes including bisphenol S (BPS) and bisphenol F (BPF), which have similar chemical structures to BPA. Ishikawa cells (3 × 10 5 ) were seeded on a 6-well plate in MEM containing 10% FBS and supplements. The culture medium was changed to phenol red-free MEM without FBS to eliminate any potential estrogenic effects due to phenol red or hormones in FBS. After 24 h, cells were treated with 1-100 µM of BPA, BPF, or BPS for a further 24 h. Cells were trypsinized, stained with trypan blue, and counted using a hemocytometer.

Transfection and Luciferase Assay
Ishikawa cells were seeded onto 12-well plates and cultured to 80-90% confluency before transfection. Cells were transfected with 1 µg luciferase reporter plasmid (3xERE-TATA-Luc) or control plasmid (pGL2-TATA-Luc) together with 0.1 µg internal control plasmid (pRL-TK) using lipofectamine 2000. Transfected cells were treated with different concentrations of bisphenols with or without steroid receptor antagonists. After 24 h, firefly and renilla luciferase activity were measured using Dual-Glo luciferase assay. The ratio of firefly to renilla luminescence in each well was calculated and compared with the empty vector controls.

RNA Extraction, RT-PCR, and Real-Time PCR
Ishikawa cells (3 × 10 5 ) were seeded on 6-well plates in MEM with 10% FBS. Before treatment, the culture medium was changed to phenol red-free MEM without FBS for 24 h. Cells were treated with 10 µM BPA, BPS, or BPF with or without steroid hormone antagonists for 24 h. Total RNA was extracted and reverse transcribed using TaqMan reagents. The resulting cDNA was amplified using the TaqMan real-time PCR system, and the expression was normalized with the 18S internal control.

Microarrays
Total RNA was extracted as previously described [27], and RNA quality was analyzed by an Agilent 2100 Bioanalyzer (Agilent Technologies, Inc., Santa Clara, CA, USA). Microarray analysis was performed using a GeneChip Human Transcriptome Array 2.0 (Affymetrix) by The Centre for PanorOmic Sciences (CPOS) at Li Ka Shing Faculty of Medicine, University of Hong Kong. Per-chip normalization was performed using the robust multi-chip average (RMA) algorithm based on the expression values of all genes. The normalized expression values of all genes were then statistically analyzed by one-way ANOVA with p-value set at 0.01 or less. All differentially expressed genes (>2-fold and p < 0.01) were presented in a Venn diagram as upregulated and downregulated genes. Un- supervised clustering was employed to analyze differences in the gene expression profile between treatment groups based on the normalized Microarray data of all genes.

Spheroid-Endometrial Cells Attachment Assay
Human choriocarcinoma Jeg-3 cells were trypsinized, and 3 mL of the cell suspension at 1 × 10 5 cell/mL was transferred to 6-well plates in 1% BSA containing DMEM/F12 medium, and then rotated at 88 rpm for 16-18 h to generate spheroids. Spheroids with size ranging from 60 to 200 µm were selected and transferred onto a confluent monolayer of endometrial Ishikawa cells in a 12-well plate under a light microscope. Before co-culture, Ishikawa cells were starved in phenol red-free MEM medium supplemented with 5% csFBS for 24 h, and then treated with bisphenols for another 24 h. Spheroids on endometrial Ishikawa cells were co-cultured for 1 h at 37 • C in a humidified atmosphere with 5% CO 2 . Undetached spheroids were removed by centrifugation at 140 rpm for 10 min. The number of attached spheroids was counted, and the attachment rate was expressed as the percentage of attached spheroids over the total number of spheroids transferred (% adhesion).

Effect of Bisphenols, BPA, BPS, and BPF on Cell Proliferation and Cell Viability
Human endometrial epithelial Ishikawa cells were exposed to 0.01-100 µM BPA, BPS, or BPF for 1 to 3 days. Cell proliferation was evaluated and compared with the control (0.1% DMSO). The high concentration (100 µM) BPA, BPS, and BPF, but not the lower concentrations, reduced cell proliferation from 1 to 3 days ( Figure 1A). The reduction in proliferation was more prominent (p < 0.05) with 100 µM BPA on days 2 and 3 compared to BPF and BPS. Cell viability assessed by trypan blue staining showed 100 µM BPA, BPS, and BPF reduced the viability of Ishikawa cells after the 1 day treatment ( Figure 1B). The decrease in cell viability was more prominent in BPA and BPF than in BPS. In all experiments, 5 µM MTX was used as the positive control, which significantly (p < 0.05) reduced proliferation and viability of Ishikawa cells.

Effect of BPA, BPS, and BPF on Spheroid Attachment Rate
We used Jeg-3 spheroids as an embryo surrogate in the in vitro model to study embryoendometrium interactions. We showed that 10 and 100 µM BPA and 100 µM BPS significantly (p < 0.05) reduced spheroid attachment rate, whereas BPF had no suppressive effects on spheroid attachment even at 100 µM ( Figure 1C). Again, 5 µM MTX was used as the positive control, which significantly (p < 0.05) suppressed spheroid attachment.

Effect of BPA, BPS, and BPF on the Expression of Bisphenol Receptors in Ishikawa Cells
Estrogen receptors (ERα and ERβ) and GPR30 receptor were selected to study the effect of BPA, BPS, and BPF on regulating the expression of endometrial genes. We found that ERα, ERβ, and GPR30 proteins could be detected in Ishikawa cells by Western blotting (Figure 2A). Semi-quantitative analysis confirmed BPA, BPS, and BPF at 100 µM significantly downregulated ERα protein expression, with BPA showing the most significant

Effect of BPA, BPS, and BPF on Spheroid Attachment Rate
We used Jeg-3 spheroids as an embryo surrogate in the in vitro model to study embryo-endometrium interactions. We showed that 10 and 100 μM BPA and 100 μM BPS The number of attached spheroids over the total spheroids added is shown on the graph. * denotes p < 0.05 compared with DMSO control, and # denotes p < 0.05 compared with 100 µM BPA. effect of BPA, BPS, and BPF on regulating the expression of endometrial genes. We found that ERα, ERβ, and GPR30 proteins could be detected in Ishikawa cells by Western blotting ( Figure 2A). Semi-quantitative analysis confirmed BPA, BPS, and BPF at 100 μM significantly downregulated ERα protein expression, with BPA showing the most significant decrease compared with BPF and BPS. Meanwhile, 10 and 100 μM BPA and BPS, and 1 and 10 μM BPF reduced ERβ expression. We also found that 1, 10, and 100 μM BPA, BPS, and BPF significantly (p < 0.05) reduced GPR30 protein expression ( Figure 2A).

Effect of BPA, BPS, and BPF on the Regulation of Estrogen Responsive Element (ERE) Reporter Expression in Transfected Ishikawa Cells
We investigated the effects of BPA, BPS, and BPF on the classical (ERα and ERβ) and non-classical (GPR3) estrogen receptor signaling pathways. The 3xERE-TATA-Luc vector was first transfected into Ishikawa cells. We found 0.1 to 10 µM BPA and 1 to 10 µM BPF and BPS significantly (p < 0.05) activated luciferase expression ( Figure 2B). Moreover, 1 and 10 µM BPA and BPF had higher transactivation activity on luciferase expression than BPS. The activation of ERE by 1 µM BPA was higher than that of 1 µM BPF. However, the transactivation activity of 10 nM estrogen was comparable to that of 10 µM BPA, BPS, and BPF, suggesting the estrogenic potency of BPA was 1000-fold less than estradiol. To further confirm the effects on estrogen receptor signaling pathway by bisphenols, estrogen receptor antagonist (ICI 182,780), estrogen receptor α-specific antagonist (MPP dihydrochloride), estrogen receptor β-specific antagonist (PHTPP), and GPR30 antagonist (G15) were used to study luciferase expression in transfected human endometrial Ishikawa cells. We found that ICI 182,780 and MPP, but not PHTPP and G15, antagonized the effects of BPA, BPS, and BPF on the estrogen receptor signaling pathway in transfected Ishikawa cells ( Figure 2C).

Effect of BPA, BPS, and BPF on Progesterone Receptor Expression and Spheroid Attacment on Transfected Ishikawa Cells
We next examined if bisphenols also regulated progesterone receptor (PR) expression in Ishikawa cells. Progesterone receptor has two isoforms, PR-A and PR-B. We found that 0.1-100 µM BPA and 1-100 µM BPS and BPF upregulated PR-A and PR-B expressions in Ishikawa cells ( Figure 3A). Similarly, the positive control diethylstilbestrol (DES, 10 nM), an estrogen agonist, was also able to upregulate PR expression in Ishikawa cells. The levels of PR upregulation were comparable between 10 nM DES and 1 µM BPA, BPS, and BPF, suggesting they have a strong potency (~100-fold) at higher concentrations ( Figure 3A). We used PR siRNA to study the functional role of PR upregulation by bisphenol on spheroid attachment. Transfection of non-target siRNA did not change the expression of PR proteins. The strongly induced expressions of PR-A and PR-B by DES, BPA, BPS, and BPF were abolished by PR siRNA, but not by non-target siRNA, in the transfected Ishikawa cells ( Figure 3B). No change in the house-keeping gene (β-actin) was found in all samples tested. Importantly, transfection of PR siRNA in Ishikawa cells reversed the suppressive effect on spheroid attachment by bisphenols and DES, but not MTX ( Figure 3C).

Effect of BPA, BPS, and BPF on the Transcriptome of Treated Ishikawa Cells
To investigate if BPA, BPS, and BPF induce similar molecular changes in Ishikawa cells, we performed a transcriptomic analysis on bisphenol-treated Ishikawa cells. Total RNA of Ishikawa cells treated with 10 µM BPA, BPS, or BPF for 24 h were collected and analyzed by Microarray. Differentially expressed genes were selected based on statistical analyses (p < 0.05, one-way ANOVA) and on fold change (1.5-fold up-/downregulated) compared with the DMSO control. There was a total of 667 genes with p < 0.05, but only 43 genes had at least a 1.5-fold change in expression ( Table 1). The clustering analysis in Figure 4A shows the high similarity gene expressions among BPA, BPS, and BPF treatment groups compared with the DMSO control. There were 35 upregulated and eight downregulated genes in the bisphenol treatments compared with the control ( Figure 4B). The gene most upregulated by BPA, BPF, and BPS was progesterone receptor (PR or PGR), with 4.58-, 4.18-, and 3.74-fold increases, respectively. Of 35 upregulated genes, 24 were induced by all bisphenols, suggesting similar gene activation. Two genes (POLR3G and EPB41L2) were specifically upregulated (>1.5-fold) by BPA, and five genes (RNU7-24P, STC2, DNAJB9, and two novel transcripts) were specifically upregulated by BPF. Three genes (LGR5, GPR110 and SLC2A12) out of eight downregulated genes were suppressed by all bisphenols. Two genes (PIH1D2 and ERP27) were specifically downregulated by BPS, and two genes (MEOX1 and SOX4) were specifically suppressed by both BPA and BPF. The large overlap of genes regulated by all three bisphenols indicate that BPA, BPS, and BPF have similar effects on Ishikawa cells. Cells 2021, 10, x 8 of 1

Gene Upregulation Is Mediated by ERs and Not GPR30
We validated the changes in the transcript levels by qPCR. We also further investigated the effect of ER antagonists (ICI 182,780 and G15) on bisphenol-regulated gene expression in Ishikawa cells. Based on the Microarray data (Table 1), six genes were selected, including the top three genes (PGR, THBS1, and NPPC) and three genes related to endometrial receptivity (ANO1, TGFA and OLFM1). Real-time PCR analysis confirmed the upregulation of these genes by BPA, BPS, and BPF in Ishikawa cells ( Figure 4C). Interestingly, the fold changes were found to be higher than in the qPCR analysis. Putative estrogen responsive elements (ERE) were found in the upstream promoter regions of these genes. We found that 4, 1, 8, 1, 2, and 11 were putative ERE-sites at the promoter regions of the TGFA, THBS1, ANO1, PGR, NPPC, and OLFM1 genes, respectively, as determined by Dragon ERE Finder online (http://datam.i2r.a-star.edu.sg/ereV3/, 8 July 2016) (data not shown).
We further tested if the transcript expressions induced by bisphenols could be nullified by ER antagonists (estrogen receptor antagonist ICI 182,780 and GPR30 antagonist G15) in treated Ishikawa cell. We found the upregulation of six genes (TGFA, THBS1, ANO1, PGR, NPPC and OLFM1) by 10 µM BPA, BPS, and BPF was reversed by ICI 182,780 but not G15 ( Figure 4D). However, G15 partially suppressed the upregulation of OLFM1 by BPS, which was still significantly higher than in the control group.   #, ## and ### denote significant differences from the DMSO control at p < 0.05, <0.01 and <0.005, respectively. *, ** and *** denote significant differences from the same bisphenol control at p < 0.05, <0.01 and <0.005, respectively.

Discussion
In this study, we found low doses of BPA, BPS, and BPF did not affect cell viability or spheroid attachment on human endometrial epithelial cells. High doses of BPA, BPS, and BPF suppressed ERα, ERβ, and GPR30 expressions and induced PR expression in human endometrial epithelial cells. We found that BPA, BPS, and BPF acted through ERα to regulate the downstream signaling pathway. Treatment with PR siRNA nullified the suppressive effects of BPA, BPS, and BPF on spheroid attachment onto endometrial Ishikawa cells. Moreover, BPA, BPS, and BPF modulated similar subsets of genes in Ishikawa cells that control endometrial receptivity.
We first tested the effects of bisphenols on the viability of Ishikawa cells. High concentrations of bisphenol reduced the viability of Ishikawa cells, with BPA having the highest cytotoxicity among the three bisphenols tested. Bisphenols can act via the ER by binding to estrogen response element (ERE) to regulate gene expressions [28]. Studies showed that BPA can regulate the expression of some ERE responsive genes in vitro [29] through classical and non-classical estrogen signaling pathways in specific cell types [30]. Moreover, BPA mainly acts as an antagonist on estrogen receptors to exert its effect [31,32]. However, BPA can also disrupt the endocrine system through other hormone receptors, including thyroid hormone receptor [33,34] and androgen receptor [32,35].
The membrane estrogen receptor GPR30 was recently identified as a novel receptor for BPA [24,36]. It was shown that GPR30 can potentially be activated through non-classical estrogen pathways [24]. It contributes to estrogen physiology and pathophysiology in different contexts such as endometrium, pregnancy decidua, and implantation [37,38]. The estrogenic potencies of BPS and BPF have been compared with BPA in several cell lines, including MCF-7 cells and MELN cells [17]. Interestingly, the potencies of BPS and BPF on gene activation were found to be in the same order of magnitude as BPA.
Another method to investigate the estrogenic activity of bisphenols is to transfect ERE-TATA luciferase reporter plasmid into Ishikawa cells, which will be exposed to the bisphenols. Estrogen response element (ERE), a specific DNA sequence in the regulatory regions of some genes, is required for classical estrogen receptor binding and for regulating the expression of estrogen responsive genes [28]. In transfected Ishikawa cells, bisphenols activated estrogen responsive genes, as demonstrated by increased luciferase signals mediated by ERE binding, with BPA having higher potency than BPF and BPS. The gene activation by bisphenols in Ishikawa cells was mainly through ERα, because the increased luciferase signal was nullified only by ICI 182,780 (ERs antagonist) and MPP dihydrochloride (ERα-specific antagonist) but not PHTPP (ERβ-specific antagonist) or G15 (GPR30 antagonist). Similarly, BPA-induced gene activation in stromal cells of the uterus was inhibited by ICI 182,780 [39]. Induced cell proliferation by BPA in mouse Sertoli TM4 cells was shown to involve both GPR30 and ERα/β [40]. In human breast cancer cells, GPR30 was found to be necessary for BPA-induced activation of Erk1/2, cell proliferation and migration, and transcriptional regulation of genes (c-fos, EGR-1, and CTGF) independent of ERα/β-mediated signaling [36,41]. However, the effects and activity of BPA depend on the cell type and potential receptors [42]. As we showed, gene activation induced by bisphenols in Ishikawa cells involve ERE binding mediated through ERα but not ERβ or GPR30.
The progesterone receptor is a strong estrogen responsive gene, which has been found to be regulated by ERα [43]. Similar to our findings, other researchers showed that PR mRNA and protein expression was upregulated in BPA-treated Ishikawa cells [44]. In addition, BPA was reported to increase PR protein expression in the uterus of pregnant mice [45]. The progesterone-PR signaling pathway was found to be required and indispensable for the establishment and maintenance of pregnancy [46,47]. In mice, the expression of PR in luminal epithelial cells is increased from pregnancy days 2-4 but is extinguished on day 5 during the window of receptivity [48]. In humans, the downregulation of PR in endometrial epithelial cells was observed during pregnancy [49]. Therefore, BPA-induced expression of PR may disturb the normal dynamic expression of PR in the uterus, resulting in the aberrant activation of the PR signaling pathway, leading to compromised implantation and pregnancy.
We further investigated global transcriptomic changes in bisphenol-treated Ishikawa cells. The gene profiles of Ishikawa cells exposed to estrogen, DES, and BPA have been previously reported [50][51][52]. In this study, we investigated the transcriptomic changes due to BPA, BPF, and BPS in Ishikawa cell. Comparing a similar study by Naciff [52] and another study on Ishikawa cells exposed to BPA and DES [49], we identified the upregulation of TGFA, THBS1, PGR, and OLFM1 genes.
It was reported that OLFM1 (olfactomedin 1) is a negative factor for embryo implantation or pregnancy. A higher expression of OLFM1 was found in endometrial tissues from patients with unexplained recurrent spontaneous abortion [53]. An in vitro study found the spheroid attachment onto endometrial epithelial cells was suppressed by OLFM1 [54,55]. The expression of transforming growth factor alpha (TGFA) in human endometrial epithelial cells varies with menstrual cycle stage, with high expression levels together with high serum E2 levels in the late follicular and luteal stages [56]. However, the role of TGFA in pregnancy or embryo implantation is unknown. Thrombospondin 1 (THBS1) is an adhesive glycoprotein that mediates cell-cell and cell-matrix interactions and is an inhibitor of angiogenesis [57]. The expression of THBS1 is higher in the receptive phase than in the pre-receptive phase of human endometrial tissues. It was also shown to be highly expressed in receptive endometrial RL95-2 cells compared with non-receptive HEC1-A cells [58]. Moreover, the expression of THBS1 in Ishikawa cells was shown to be regulated by progesterone [59]. Decreased THBS1 expression in decidua macrophages was associated with unexplained recurrent spontaneous abortion [60]. Anoctamin 1 (ANO1) is calcium-activated chloride channel protein and was shown to be involved in myometrial contractility in human and murine myometrial tissue [61]. In mouse ovarian granulosa cells, estradiol production was enhanced by the inhibition or knockdown of ANO1 [62]. The role of ANO1 in endometrial function and pregnancy outcome still needs further investigation. Natriuretic peptide C (NPPC) is highly expressed in the uterus and placenta of mouse and human [63], and its receptor is also found in the uterus during pregnancy [64]. The expression of uterine NPPC was induced by estradiol in a mouse model [65]. An increase in myometrial NPPC expression was found to be associated with pregnancy complications (intrauterine growth retardation) [66], and an increased secreted NPPC level in the serum was found in women with complicated pregnancy [67]. However, the role of NPPC in endometrial receptivity and embryo implantation remains obscure.
Using different ER inhibitors (ICI 182,780 and G15), we found bisphenols acted through nuclear ER receptors, but not membrane ER receptors, to induce the expression of the six selected genes. This was confirmed by our luciferase reporter experiments, which showed that ICI 182,780 and MPP (ERα antagonist) nullified bisphenol-induced luciferase activity in the transfected Ishikawa cells. Similarly, bisphenol-induced luciferase activity has also been demonstrated in human breast cancer cells [68].
Several models have been established to study human embryo-endometrium interactions in vitro, including spheroids co-cultured with endometrial cell monolayer [69]. The receptivity of endometrial cells is critical for embryo implantation and pregnancy [70]. Previous studies identified genes that are changed in the human receptive endometrium [71]. With advances in single-cell sequencing techniques, the transcriptomic changes in the human endometrium at the single-cell level have now been reported [72].
In summary, we compared the effects of BPA, BPF, and BPS on endometrial Ishikawa cell toxicity, viability, spheroid attachment, and on the involvement of the ER signaling pathway. At high concentrations, BPA, BPS, and BPF downregulated ERα and stimulated PR expression. A similar finding was observed in human prostate cancer LNCaP cells, and the activation of EGFR/ERK/p53 signaling pathway was demonstrated after high doses of BPA exposure [73]. Knockdown of PR by siRNA reversed the suppressive effect of bisphenols on spheroid attachment, suggesting physiological doses of bisphenols may not affect human reproductive function in vivo. The transgenerational effects of bisphenols on human reproductive function warrant further investigation.