2018 Volume 43 Issue 5 Pages 321-327
Bisphenol AF (BPAF) is now recognized as one of the replacements for bisphenol A (BPA). Although considerable experimental evidence suggests that BPA is an endocrine-disrupting chemical, the toxicological profile of BPAF has been investigated in less detail than that of BPA, even at the in vitro level. BPAF has been established as an activator of estrogen receptor α (ERα) in many cell lines; however, controversy surrounds its effects on the other isoform, ERβ (i.e., whether it functions as a stimulator). Five human ERβ isoforms have been cloned and characterized. Of these, we focused on the interactions between BPAF and the two isoforms, ERβ1 and ERβ2. We demonstrated that i) BPAF functioned as a stimulator of ERβ1 (and ERα), which is transiently expressed in the two types of human breast cancer cells (MDA-MB-231 and SK-BR-3 cells) (EC50 values for ERβ: 6.87 nM and 2.58 nM, respectively, and EC50 values for ERα: 24.7 nM and 181 nM, respectively), ii) the stimulation of ERβ1 by BPAF (1-25 nM) was abrogated by PHTPP (an ERβ selective antagonist), and iii) the expression of ERβ1 and ERβ2 was not modulated by BPAF at nanomolar concentrations up to 25 nM. These results indicate that BPAF activates not only human ERα, but also the ERβ1 isoform in breast cancer cells, and exhibits higher activation potency for ERβ1.
Estrogens including 17β-estradiol (E2) are essential for the development and maintenance of the reproductive organs. Estrogens generally utilize two estrogen receptors (ERs), ERα (ESR1) and ERβ (ESR2), to induce a number of physiological effects as ligands (Shanle and Xu, 2011). Endocrine-disrupting chemicals (EDCs) are compounds that interfere with hormone biosynthesis, metabolism, or signaling. In interactions between EDCs and estrogen signaling, many EDCs are known to exhibit estrogenic/anti- estrogenic activities, which are, at least in part, mediated by the two ERs (Shanle and Xu, 2011). EDCs have been shown to modulate normal estrogen signaling in two manners: 1) a direct/indirect interaction with ERs (as ligands) and 2) the up-/down-regulation of ERs. The physiological functions of ERβ have not yet been examined in as much detail as those of ERα, including the effects of EDCs on the five ERβ isoforms: ERβ1, ERβ2 (also called ERβcx), ERβ3 (restricted to the testis), ERβ4, and ERβ5 (Moore et al., 1998; Leung et al., 2006; Sugiyama et al., 2010).
Bisphenol A (BPA), a monomer used in polycarbonate plastic and polystyrene resins, is regarded as an endocrine-disrupting chemical that perturbs estrogen signaling by acting as a ligand for ERα and ERβ (Rochester, 2013). Due to the toxic effects of high levels of BPA, it is being replaced with analogs that are structurally similar, such as bisphenol B (BPB), bisphenol S (BPS), and bisphenol AF (BPAF). Among these bisphenols, the abnormal nature of BPAF (a fluorinated derivative of BPA: –CF3) (See Fig. 1A) has been reported; it has been shown to more strongly accumulate in hepatocytes than parent BPA even though the hallmark of lipophilicity (e.g., Pow) of BPAF is not as high as that of BPA (HSDB, 2001; NTP, 2008; Waidyanatha et al., 2015). Furthermore, studies on BPA and BPAF demonstrated that the activity of the latter for “ERβ” depended on the cell type employed (Matsushima et al., 2010; Li et al., 2012). There is currently no experimental evidence to show whether BPAF affects the transcriptional activity of ERβ in the human breast cancer cell lines, MDA-MB-231 and SK-BR-3. Therefore, we herein evaluated the effects of BPAF on ERβ activity in these two breast cancer cell lines transfected with human ERβ cDNA.
BPAF-mediated stimulation of transcriptional activity mediated by ERs in MDA-MB-231 cells. (A) Chemical structure of BPAF. (B) MDA-MB-231 cells were transiently transfected with an ERE-luciferase reporter plasmid. After transfection, cells were exposed to four bisphenols: BPA, BPB, BPS, and BPAF, at 25 µM. After 24 hr, cells were harvested and assayed for luciferase activity, and all transfections were normalized for efficiency using the internal Renilla control plasmid. Data are expressed as a fold induction from the vehicle-treated control (Ctl.) (indicated as 1), as the mean ± S.E. (n = 3). *Significantly different (P < 0.05) from the vehicle-treated control.
BPA [2,2-bis(4-hydroxy phenyl)propane] (purity: > 99%) was purchased from Wako Pure Chemical Industries (Osaka, Japan). BPAF [1,1,1,3,3,3-hexafluoro-2,2-bis(4-hydroxy phenyl)propane] (purity: > 98%), BPB [2,2-bis(4-hydroxyphenyl)butane] (purity: > 98%), and BPS [bis(4-hydroxyphenyl)sulfone] (purity: > 98%) were purchased from Tokyo Chemical Industry (Tokyo, Japan). PHTPP (purity: > 99%) was purchased from Tocris Biosciences (Ellisville, MO, USA). All other reagents were of the highest grade commercially available.
The conditions/methods of cell cultures were described previously (Takeda et al., 2012, 2013). Briefly, the human breast cancer cell lines, MDA-MB-231 and SK-BR-3 (obtained from the American Type Culture Collection, Rockville, MD, USA) were routinely grown in phenol red-containing minimum essential medium α (MEMα) (Invitrogen, Carlsbad, CA, USA), supplemented with 10 mM HEPES [4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid], 5% fetal bovine serum, penicillin (100 U/mL), and streptomycin (100 µg/mL) in a humidified incubator, within an atmosphere of 5% CO2 at 37°C. Prior to chemical treatments, culture medium was changed to phenol red-free MEMα (Invitrogen) supplemented with 10 mM HEPES, 5% dextran-coated charcoal-treated fetal bovine serum, penicillin (100 U/mL), and streptomycin (100 µg/mL). Cultures of approximately subconfluence in a 100-mm Petri dish were used to seed for experiments on mRNA expression as well as the transfection analysis (See the dual-luciferase reporter assay).
Experiments were performed as described previously (Takeda et al., 2013; Okazaki et al., 2017). In brief, 24 hr prior to transfection, MDA-MB-231 and SK-BR-3 cells were seeded (5 × 104 cells/well) on 24-well plates containing MEMα. The transfection of each expression plasmid was performed using Lipofectamine® LTX with PLUS™ reagent (Invitrogen) according to the manufacturer’s instructions. DNA mixtures containing 300 ng of the (estrogen-responsive element, ERE)3-Luc plasmid were co-transfected with 2 ng of the Renilla luciferase reporter plasmid (pRL-CMV) in plates together with an expression plasmid carrying human ERα, ERβ1, or ERβ2 cDNA (100 ng). Cells were washed with phosphate-buffered saline 24 hr post-transfection and changed to phenol red-free MEMα supplemented with 5% dextran-coated charcoal-treated serum, followed by the respective chemical treatments being tested. Cell extracts were then prepared using 100 µL of passive lysis buffer (Promega, Madison, WI, USA), and 20 µL was then applied to the firefly luciferase and Renilla luciferase assays (Dual-Luciferase Reporter Assay System, Promega). The ratio of firefly luciferase activity (expressed from reporter plasmids) to Renilla luciferase activity (expressed from pRL-CMV) in each sample served as a measure of normalized luciferase activity.
Total RNA was prepared from MDA-MB-231 cells using the RNeasy kit (Qiagen, Inc., Hilden, Germany) and purified using RNeasy/QIAamp columns (Qiagen, Inc.). In the real-time RT-PCR analysis of ERβ1, ERβ2, and β-actin, cDNA was prepared via RT of total RNA using the ReverTra Ace® qPCR RT kit (Toyobo Co., Ltd., Osaka, Japan). Real-time quantitative RT-PCR assays were performed with FastStart Essential DNA Green Master (Roche Applied Science, Indianapolis, IN, USA). The primers used for PCR on human ERβ1, ERβ2, and β-actin were from previous studies (Leung et al., 2006; Takeda et al., 2013). Target mRNA levels were normalized to the corresponding β-actin mRNA levels. Quantification cycle (Cq) values were assessed in order to compare expression between ERβ1 and ERβ2 (Fig. 2A).
BPAF-mediated up-regulation of ERβ1 and ERβ2 expression in MDA-MB-231 cells. (A) Real-time RT-PCR analyses of ERβ1 and ERβ2 in MDA-MB-231 cells (basal expression levels). Quantification cycle (Cq) values for ERβ1 and ERβ2 are indicated. Data are expressed as the mean ± S.E. (n = 3). (B) Real-time RT-PCR analyses of ERβ1 and ERβ2 in MDA-MB-231 cells 48 hr after the treatment with vehicle (control) and BPAF (25 µM). Data are expressed as a fold induction from the vehicle-treated control (Ctl.) (indicated as 1), as the mean ± S.E. (n = 3). *Significantly different (P < 0.05) from the vehicle-treated control.
Differences were considered to be significant when the P value was calculated as less than 0.05. Significant differences between two groups were calculated by the Student’s t-test. A data analysis of differences among multiple groups was performed using Dunnett’s test. Calculations were performed using Statview 5.0J software (SAS Institute Inc., Cary, NC, USA).
We recently reported that BPAF (Fig. 1A) “up-regulates” the expression of ERβ at markedly higher concentrations (i.e., 25 µM), and this is associated with the suppression of estrogen signaling in MCF-7 cells (Okazaki et al., 2017). MCF-7 cells express both isoforms of ERs, and intermolecular interactions between ERα and ERβ, which may result in changes in estrogen signaling, have been reported (Powell and Xu, 2008; Shanle and Xu, 2011; Takeda et al., 2013; Takeda, 2014). In order to clarify/investigate the actions of BPAF on the ERβ isoform, we herein focused on another human breast cancer cell line, MDA-MB-231, which only expresses ERβ transcripts (an ERα-negative cell line) (Weigel and deConinck, 1993; Takeda et al., 2013). A previous study reported that ERβ dispatches signals to its downstream pathways, even in the absence of its ligands, which is different from unliganded ERα (Vivar et al., 2010). If this is the case for the result described above, and if BPAF also stimulates the expression of ERβ in MDA-MB-231 cells, ERE-mediated transcription may be activated. As shown in Fig. 1B, BPAF at 25 µM exhibited the strongest activation effects on ERE-mediated transcription among the other representative bisphenols tested (> 1.5-fold), followed by BPA, a parent compound of BPAF (~1.25-fold), indicating that the highest transcription activity by BPAF is mediated by ERβ.
There are five splicing variants of ERβ: ERβ1 (wild-type), ERβ2 (also called ERβcx), ERβ3 (a form restricted to the testis), ERβ4, and ERβ5 (Moore et al., 1998; Leung et al., 2006; Sugiyama et al., 2010). Although the physiological activities of the ERβ2/ERβ4/ERβ5 isoforms in breast cancer cells currently remain unclear, ERβ1 has been shown to exhibit transcriptional activity through ERE, and ERβ2 itself cannot stimulate ERE-driven activity, even in the presence of E2 (Leung et al., 2006). We initially studied the basal expression profiles of ERβs in MDA-MB-231 cells. A real-time RT-PCR analysis revealed that the expression of ERβ2 was approximately 8-fold stronger than that of ERβ1 (Fig. 2A). The expression signals of ERβ4 and β5 were not detected (data not shown). The expression profiles of the ERβ isoforms obtained in this study were consistent with those reported by Leung et al. (2006). ERβ1 and ERβ2 expression levels were then investigated after exposure to BPAF (25 µM). As shown in Fig. 2B, the expression of two ERβ isoforms was up-regulated, and the positive modulation of ERβ2 was significant (4.9-fold, P < 0.05). In contrast to ERβ1 and ERβ2 isoforms, the expression signals of ERβ4 and ERβ5 were not obtained due to below detection limits (data not shown), and the expression of ERα was not significantly modulated by 25 µM BPAF (relative expression; 1.53 ± 0.42, p = 0.39) when compared to vehicle-treated control. Moreover, the expression of ERβ1 and ERβ2 was not affected by BPAF at concentrations ranging between 1 and 25 nM (Fig. 3). Although the ERβ2 isoform was shown to have undetectable affinity for E2 (Leung et al., 2006), no experimental evidence to show that BPAF modifies the function of ERβ2 was obtained. We utilized MDA- MB-231 cells transfected with an expression plasmid carrying human ERβ2 cDNA. Similar to E2, no modulative effects by BPAF on ERβ2/ERE up to 25 µM were observed (Fig. 4), suggesting that BPAF activates ERE-driven transcription possibly via ERβ1 (See Fig. 1B) in MDA-MB-231 cells. In subsequent experiments, we focused on the interplay between BPAF and the ERβ1 isoform from the standpoint of transcriptional activation of the receptor after its interaction with ERβ1.
Effects of BPAF on ERβ1 and ERβ2 expression in MDA-MB-231 cells. Real-time RT-PCR analyses of (A) ERβ1 and (B) ERβ2 in MDA-MB-231 cells 48 hr after the treatment with vehicle (Control: Ctl.) and BPAF (1, 5, and 25 nM). Data are expressed as a fold induction from the Ctl. (indicated as 1), as the mean ± S.E. (n = 6).
Effects of BPAF on ERβ2-mediated transcriptional activity in MDA-MB-231 cells. MDA-MB-231 cells were transiently transfected with an ERE-luciferase reporter plasmid in combination with ERβ2 cDNA. After transfection, cells were treated with BPAF (5, 10, and 25 µM). After 24 hr, cells were harvested and assayed for luciferase activity, and all transfections were normalized for efficiency using the internal Renilla control plasmid. Data are expressed as a fold induction from the vehicle-treated control (Ctl.) (indicated as 1), as the mean ± S.E. (n = 3). *Significantly different (P < 0.05) from the vehicle-treated control.
Although BPAF has been recognized as a stimulator of ERα in many cancer cell lines that were established from different organs (Matsushima et al., 2010; Li et al., 2012), a focus on the other isoform, ERβ, proposed a largely unaddressed question as to whether BPAF is also a “stimulator” of ERβ because BPAF not only exhibits agonistic potential against ERβ, but also lacks modulation potential, which, for example, was observed in human cervical cancer HeLa cells (Matsushima et al., 2010; Li et al., 2012). In the present study, we utilized two types of human breast cancer cell lines, SK-BR-3 and MDA-MB-231 cells, the former of which is entirely negative for both ERs and the latter is known as ERα (–)/ERβ (very low) (Okazaki et al., 2017). As shown in Figs. 4A and B, in the MDA-MB-231 and SK-BR-3 cell lines transfected with human ERβ1 cDNA, BPAF activated ERβ1-mediated transcriptional activity in a concentration-dependent manner, giving EC50 values of 6.87 and 2.58 nM, respectively. ERα activity was positively stimulated by BPAF in both cell lines (EC50 values: 24.7 and 181 nM), although ERα activity at higher concentrations (higher than 15 µM) slightly decreased in MDA-MB-231 cells (Figs. 5C and 5D). It is important to note that the activation of ERβ1 by BPAF was observed at markedly lower concentrations that those for ERα, and this concentration is relevant to environmental levels (Yang et al., 2014).
Effects of BPAF on ERβ1- and ERα-mediated transcriptional activity in breast cancer cells. (A and C) MDA-MB-231 cells and (B and D) SK-BR-3 cells and were transiently transfected with an ERE- luciferase reporter plasmid in combination with ERβ1 or ERα cDNAs. After transfection, cells were treated with BPAF ranging between 25 pM and 25 µM. After 24 hr, cells were harvested and assayed for luciferase activity, and all transfections were normalized for efficiency using the internal Renilla control plasmid; the expression of ERβ1 and ERα was increased approximately 6000-fold and 90000-fold by cDNA introduction, respectively, in these breast cancer cells. Data are expressed as a fold induction from the vehicle-treated control (Ctl.) (indicated as 1), as the mean ± S.E. (n = 3). *Significantly different (P < 0.05) from the vehicle- treated control.
In order to investigate whether BPAF directly activates the transcriptional activity of ERβ1, PHTPP, a highly selective ERβ antagonist, was introduced into the reaction system of ERβ1/ERE along with BPAF in MDA-MB-231 cells. As shown in Fig. 6, the concentration-dependent stimulation of ERβ1 activity was reproducibly demonstrated up to 25 nM (See also Fig. 5A), and BPAF-stimulated activity was effectively abrogated by the presence of 1 µM PHTPP, a selective full antagonist of the ERβ isoform (Compton et al., 2004), to control levels. It is important to note that when we focused on the control system (Ctl.), no observable modulation was detected by PHTPP alone, suggesting that PHTPP inhibited ERβ1 transcriptional activity stimulated by BPAF only. As described above, since the expression of ERβ1 was not up-regulated by BPAF in the concentration range of 25 nM or less (Fig. 3A), BPAF appears to function as an activator of ERβ1-mediated transcription in the nanomolar concentration range. We recently reported that BPAF has the potential to up-regulate ERβ at the mRNA/protein level in human breast cancer cells at relatively “higher concentrations” (25 µM) (Okazaki et al., 2017). In addition, BPAF has been shown to exhibit abnormal behavior; although its log Pow (i.e., n-octanol/water: 2.82) is smaller than that of BPA (3.32), its clearance from hepatocytes is markedly slower than BPA, indicating that BPAF accumulates in cells, possibly leading to unwanted toxicological outcomes (HSDB, 2001; NTP, 2008; Waidyanatha et al., 2015). These findings suggest that BPAF acts as an endocrine disruptor underlying the dual activation of ERβ1, i) a direct activator for the ERβ1 ligand at low concentrations (nM order) and ii) the induction of ERβ at high concentrations (µM order). Since bisphenols are detected in human biological samples (Yang et al., 2014), BPAF is able to evoke its unwanted effects on breast tissues through interactions with ERβ rather than ERα. Although we mainly focused on the interaction between BPAF and ERβ1 (genomic effects), one of the alternatively spliced transcript variants of the ESR2 gene, to comprehensively understand the biological (toxicological) profiles of BPAF, further studies that consider the involvement of the G protein-coupled estrogen receptor pathway, which has recently been reported to be involved in BPAF-mediated non-genomic estrogenic effects in breast cancer SK-BR-3 cells (Cao et al., 2017), are needed.
Effects of BPAF on ERβ1-mediated transcriptional activity in MDA-MB-231 cells. MDA-MB-231 cells were transiently transfected with an ERE-luciferase reporter plasmid in combination with ERβ1 cDNA. After transfection, cells were treated with BPAF at concentrations ranging between 1 nM and 25 nM in the presence (+) or absence (–) of 1 µM PHTPP. In the system without PHTPP, an equivalent volume of vehicle was added. After 24 hr, cells were harvested and assayed for luciferase activity, and all transfections were normalized for efficiency using the internal Renilla control plasmid. Data are expressed as a fold induction from the vehicle-treated control (Ctl.) (indicated as 1), as the mean ± S.E. (n = 3). *Significantly different (P < 0.05) from the vehicle-treated control.
This research was supported by EXTEND2010 grants from the Ministry of the Environment, Japan (to H.A.). This study was also supported in part by Grants-in-Aid for Scientific Research (C) [25460182 and 17K08402, (to S.T.)] from the Japan Society for the Promotion of Science (JSPS) KAKENHI.
The authors declare that there is no conflict of interest.