Developmental exposure to bisphenol A alters expression and DNA ...

Molecular and Cellular Endocrinology 417 (2015) 191e199

Contents lists available at ScienceDirect

Molecular and Cellular Endocrinology journal homepage: www.elsevier.com/locate/mce

Developmental exposure to bisphenol A alters expression and DNA methylation of Fkbp5, an important regulator of the stress response €lle Rüegg c, d, * Efthymia Kitraki a, Ivan Nalvarte b, Ali Alavian-Ghavanini c, d, Joe a

School of Health Sciences, National and Kapodistrian University of Athens, Thivon 2str, Athens 11527, Greece Department of Biosciences and Nutrition, Karolinska Institutet, 141 83 Huddinge, Sweden c Department of Clinical Neuroscience, Karolinska Institutet, Center for Molecular Medicine (CMM) L8:00, 171 76 Stockholm, Sweden d €lje, Sweden €derta Swedish Toxicology Science Research Center (Swetox), Forskargatan 20, 15136 So b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 3 July 2015 Received in revised form 9 September 2015 Accepted 24 September 2015 Available online 30 September 2015

Bisphenol A (BPA), an abundant endocrine disruptor, affects stress-responsiveness and related behaviors in children. In rats, perinatal BPA exposure modifies stress response in pubertal offspring via unknown mechanisms. Here we examined possible epigenetic modifications in the glucocorticoid receptor gene and its regulator Fkbp5 in hypothalamus and hippocampus of exposed offspring. We found increased DNA methylation of Fkbp5 and reduced protein levels in the hippocampus of exposed male rats. Similar effects were obtained in a male hippocampal cell line when exposed to BPA during differentiation. The estrogen receptor (ER) antagonist ICI 182,780 or ERb knock-down affected Fkbp5 expression and methylation similarly to BPA. Further, BPA's effect on Fkbp5 was abolished upon knock-down of ERb, suggesting a role for this receptor in mediating BPA's effects on Fkbp5. These data demonstrate that developmental BPA exposure modifies Fkbp5 methylation and expression in male rats, which may be related to its impact on stress responsiveness. © 2015 The Authors. Published by Elsevier Ireland Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Keywords: Bisphenol A DNA methylation Fkbp5 Hippocampus HT22 cells Estrogen receptor beta

1. Introduction Bisphenol A (BPA) is a widely used component of plastics and resins with endocrine disruptive features, exhibiting agonistic properties on both estrogen receptor (ER) isoforms, ERa and ERb, and antagonistic properties on androgen receptor (Delfosse et al., 2014). Human exposure to this chemical is extensive since BPA is abundant in a vast number of consumer products, including toys, drinking bottles, food containers and dental sealants. Up to 95% of the human population has detectable BPA levels in their bodies and there is increasing concern for its higher bioaccumulation in developing organisms (Calafat et al., 2008). The reference dose of 50 mg BPA/kg of bw/day, previously determined as the safe daily human exposure (Vandenberg et al., 2012), has recently been reduced to 4 mg/kg of bw/day (http://www.efsa.europa.eu/en/ topics/topic/bisphenol.htm) due to increasing evidence for adverse effects at lower exposures, especially impacting brain and

€derta €lje, Sweden. * Corresponding author. Forskargatan 20, 151 36 So E-mail addresses: [email protected] (E. Kitraki), [email protected] (I. Nalvarte), [email protected] (A. Alavian-Ghavanini), joelle.ruegg@ swetox.se (J. Rüegg).

prostate development in fetuses and children (http://www.niehs. nih.gov/health/topics/agents/sya-bpa/). Early life exposure to BPA affects a variety of developmental functions (Kundakovic and Champagne, 2011), including neuronal differentiation and migration (Wolstenholme et al., 2012). In the rodent brain, BPA modifies sexual and social behavior, impairs cognition and increases anxiety and depression-like behavior (Wolstenholme et al., 2011). In our previous studies, we showed that perinatal exposure of rats to a low BPA dose alters their basal and stress-induced HippocampalHypothalamic Pituitary Adrenal (HHPA) axis activity and related behaviors at mid-puberty in a sexually dimorphic manner (Poimenova et al., 2010; Panagiotidou et al., 2014): BPA-exposed females exhibited increased basal corticosterone and reduced hypothalamic glucocorticoid receptors (GR) levels, as well as anxiety-like behavior but less efficient late stress responses. BPAexposed males, on the other hand, showed a heightened stress response compared to untreated counterparts. In humans, higher pre- and postnatal BPA levels have been associated with increased anxiety and depressive behavior in children, expressed differently between boys and girls (Braun et al., 2009; Braun et al., 2011; Harley et al., 2013; Evans et al., 2014). BPA crosses

http://dx.doi.org/10.1016/j.mce.2015.09.028 0303-7207/© 2015 The Authors. Published by Elsevier Ireland Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

192

E. Kitraki et al. / Molecular and Cellular Endocrinology 417 (2015) 191e199

List of abbreviations ANOVA BPA bw ChIP: DCC dex DMEM E2 EDTA EGTA ERa ERb ERE-luc ERs

analysis of variance bisphenol A body weight chromatin immunoprecipitation dextran coated charcoal dexamethasone Dulbecco's modified Eagle's medium Estradiol ethylenediaminetetraacetic acid ethylene glycol tetraacetic acid estrogen receptor alpha estrogen receptor beta estrogen responsive element-luciferase reporter estrogen receptors

the placenta and although its metabolic clearance is high, its actions can be chronic and potentially engage epigenetic modifications. Indeed, BPA induces alterations in DNA methylation in various species, organs, and model systems upon different exposures (Kundakovic and Champagne, 2011). The HHPA axis is an important regulator of the stress response and its dysfunction is correlated to several neuropsychiatric disorders including anxiety and depression. Glucocorticoids act as downstream effectors of the axis (Holsboer and Ising, 2010). By activating GR in the hippocampus and hypothalamus, glucocorticoids exert a negative feedback on the HHPA axis towards termination of stress response and resilience (Smith and Vale, 2006). GR function depends on a large complex of transcriptional coregulators, chaperones and co-chaperones. One of them, FKBP51 (the protein product of the Fkbp5 gene), reduces hormone binding affinity and nuclear translocation of GR (Riggs et al., 2003; Touma et al., 2011). Fkbp5 is itself a GR target and glucocorticoids induce its expression as part of an intracellular ultra-short negative feedback loop for GR activity (Vermeer et al., 2003). Recent evidence indicates the sensitivity of Fkbp5 to environmental factors and epigenetic changes, thus highlighting the importance of this co-regulator in stress related disorders (Schmidt et al., 2012). Chronic exposure to glucocorticoids persistently changes Fkbp5 expression by altering DNA methylation of Fkbp5 gene in the mouse hippocampus and hypothalamus (Lee et al., 2010; Yang et al., 2012; Wochnik et al., 2005). Interestingly, DNA methylation changes in the human FKBP5 gene are also found in patients with post-traumatic stress disorder (Klengel et al., 2013a) and bipolar disorder (Fries et al., 2014). Based on the above, we herein examined whether developmental exposure to BPA may lead to epigenetic alterations in genes encoding important mediators of the stress response, such as the glucocorticoid receptor and its regulator Fkbp5. Therefore we first investigated DNA methylation changes in the regulatory regions of the aforementioned genes in the hypothalamus and hippocampus of BPA-exposed rats. The detected changes in Fkbp5 methylation in the hippocampus of male rats, which coincided with lower FKBP51 levels, led us to further examine the molecular basis of this BPA effect in a murine hippocampal cell line of male origin. Specifically, the involvement of estrogen receptors (ERs) in mediating BPA's effects in hippocampal neurons was analyzed by inhibiting ER using either ICI 182780 (ICI) or shRNA-mediated knock-down. Our results suggest an involvement of ERb in BPA's epigenetic effects on Fkbp5.

FBS Fkbp5 FKBP51 GAPDH GR GRE HEPES HHPA ICI LSD Nr3c1 PBS PMSF SDS shRNA ssDNA

fetal bovine serum FK506 binding protein 5 FK506 binding protein 51 glyceraldehyde 3-phosphate dehydrogenase glucocorticoid receptor glucocorticoid responsive element 4-(2-Hydroxyethyl)-1-piperazineethanesulfonic acid hippocampal-hypothalamic pituitary adrenal ICI 182780 least square differences nuclear receptor subfamily 3, group C, member 1 phosphate-buffered saline phenylmethanesulfonyl fluoride sodium dodecyl sulfate small hairpin RNA single-stranded DNA

2. Materials and methods 2.1. Chemicals Bisphenol A and dexamethasone were purchased from SigmaeAldrich (St. Louis, Missouri, USA), ICI 182780 from AstraZeneca (London, UK), cell culture reagents and Lipofectamine 2000 from Life Technologies (Carlsbad, CA, USA). Pre-designed shRNA against ERa and ERb and control shRNA were obtained from SigmaeAldrich. The luciferase reporter construct 3 ERE-luc has been published (Legler et al., 1999). pRL-TK for normalisation of luciferase activity was purchased from Promega (Madison, WI, USA). Antibodies and primers used are listed in Supplemental material Tables S1 and S2. 2.2. Animals Animal tissues used here were obtained in a previous study described elsewhere (Panagiotidou et al., 2014) and the protocol was approved by the Ethical Committee of the School of Health Sciences, National and Kapodistrian University of Athens, Greece. In brief, female Wistar rat breeders received BPA (40 mg BPA/kg bw/ day) or the vehicle (water, 1% in ethanol) orally via impregnated cornflakes throughout pregnancy and lactation. The offspring (BPAexposed or unexposed controls) were left to grow. At mid-puberty (postnatal day 46) the offspring were killed by decapitation either at basal conditions or two hours following a 15-min swimming stress. 2.3. Cell culture and treatments The hippocampal cell line HT22, deriving from male mouse primary hippocampal cells (Liu et al., 2009), was cultured in Dulbecco's modified Eagle's medium (DMEM), supplied with 10% fetal bovine serum (FBS) and 0.5 mg/ml penicillin/streptomycin under standard conditions (37 C, 5% CO2). For differentiation, HT22 cells were seeded into six-well plates (2 105 cells/well) in phenol red-free DMEM containing 5% dextran coated charcoal (DCC)-treated FBS and allowed to settle. Medium was changed to phenol red-free Neurobasal medium containing N2 supplement and 100 mM dibutyrylcAMP, and different concentrations of BPA and/or 10 nM E2 or ICI 182,780 (ICI). After two days, medium was changed to phenol red-free Neurobasal medium without dibutyrylcAMP, BPA, and ICI, and allowed to grow for another 3 days. 1 mM dexamethasone was added 16 h before


harvesting. For ER knock-down studies, HT22 cells were seeded into sixwell plates (2 105 cells/well) in phenol red-free DMEM containing 5% DCC-treated FBS. The next day, mission shRNA was transfected using Lipofectamine 2000 according to the manufacturer's instructions using 2.5 mg plasmid DNA and 4 ml reagent per well. After 24 h, medium was changed to phenol red-free Neurobasal medium containing N2 supplement and 100 mM dibutyrylcAMP to induce differentiation. 2.4. Western blot analyses Hippocampi were collected and immediately frozen in liquid nitrogen. The tissues (3e4 per animal group) were homogenized in NP-40 lysis buffer (137 mM NaCl, 20 mM TriseHCl pH 8.0, 1% NP-40 (v/v), 10% glycerol (v/v), 48 mM NaF, 2 mM Na3VO4, and protease inhibitor mix). Similarly, whole cell extracts were prepared from treated or untreated cells homogenized in RIPA buffer (150 mM NaCl, 50 mM Tris-HCl, pH 8.0, 1% NP-40, 0.5% sodium deoxycholate, 2 mM Na3VO4, and protease inhibitor mix). 50 mg or 30 mg total protein of the tissue extracts or whole cell extracts, respectively, were electrophoresed using 4%e12% polyacrylamide gels (Invitrogen Co., Carlsbad, CA, USA). Proteins were transferred to nitrocellulose membranes (Schleicher &Schuell, Dassel, Germany). The membranes were incubated overnight at 4 C with anti-FKBP51 antibody (Supplemental material, Table S1) in TBS buffer containing 0.05% Tween-20 and 5% fat-free milk. Incubation with antiGAPDH or b-actin antibody (Supplemental material, Table S1) was used for normalisation of protein loading. The signal was visualized by chemiluminescence (ECL, Cell Signaling Technology, Danvers, Massachusetts, USA) on Fuji Medical X-ray film. Quantification of the signal was done on scanned films using the Image J v.1.41 software. 2.5. Gene expression analyses RNA was isolated using Trizol (LifeTechnologies, Carlsbad, CA, USA) according to the manufacturer's instructions. 1 mg of total RNA was treated with DNaseI and reverse transcribed using iScript (BioRad, California, USA). Real-time PCR was performed in a 15 ml reaction with 1 ml cDNA using SYBR® Select (LifeTechnologies, Carlsbad, CA, USA) and primers for Fkbp5, GR, ERs, and 36B4 (Supplemental material, Table S2) on an ABI 7500 Fast Real Time PCR System (Thermo Scientific, Waltham, Massachusetts, USA). Relative gene expression was assessed using the DDCt method. 36B4 expression was chosen as most robust normalizer, compared to GAPDH, b-actin, and 18S RNA. 2.6. DNA methylation analyses Genomic DNA was extracted from cells using GenElute™ (SigmaeAldrich, St. Louis, Missouri, USA) and from rat brain tissues using the AllPrep DNA/RNA/Protein Mini Kit (Qiagen, Hilden, Germany) following manufacturer's instructions. Bisulfite conversion was performed using EZ Methylation Gold Kit (Zymo Research, Irvine, California, USA) and Nr3c1 promoter, Fkbp5 intron 5 and intron 1was amplified by nested PCR (for primer sequences (Supplemental material, Table S1). Pyrosequencing was performed with 25 ml of the resulting PCR product in a Pyromark Q96 ID using Pyromark Q96 Gold reagents (Qiagen, Hilden, Germany) according to the manufacturer's instructions. 2.7. Luciferase assays HT22 cells were seeded into 24-well plates (5 104 cells/well)

193

in phenol red-free DMEM containing 5% DCC-treated FBS. The next day, pRL-TK and 3 ERE-luc were transfected using Lipofectamine 2000 according to the manufacturer's instructions using a total of 0.5 mg plasmid DNA and 1 ml reagent per well. 9 h after transfection, fresh medium was added containing 10 nM E2 and/or 1 mM BPA. The next day, luciferase reporter assays were performed using the Dual-Luciferase® Reporter Assay System (Promega, Madison, WI, USA) according to the manufacturer's protocol: cells were lysed in passive lysis buffer and firefly and renilla luciferase activity measured in 96-well plates using a Tecan infinite F200 luminometer (Tecan). Firefly luciferase activity was normalized to renilla luciferase activity. 2.8. Chromatin immunoprecipitation (ChIP) ChIP assays were performed as described (Cortazar et al., 2011) 2e5 mg of the respective antibody was used (Supplemental material, Table S1). DNA was purified using the QIAquick PCR purification kit (Qiagen, Hilden, Germany) and the isolated DNA fragments were analysed by qPCR using Rotor-Gene SYBR Green PCR Kit on a Rotor-Gene RG-3000 (Qiagen, Hilden, Germany). 2.9. Statistical analyses To evaluate Western blot data, the mean ratio of FKBP51 to GAPDH or b-actin for each analysed sample was used. Two-way analysis of variance (ANOVA) was used to examine the effect of BPA treatment and stress on hippocampal FKBP51 levels. The LSD posthoc test was used for group comparisons. Unpaired student's ttest was used for methylation, gene expression, and ChIP analyses unless stated otherwise. Significance was accepted for P < 0.05. 3. Results 3.1. Perinatal BPA exposure affects Fkbp5 methylation and FKBP51 levels in male rats We have previously shown that perinatal exposure to a dose of BPA close to the range of human exposure leads to prolonged sexspecific alterations in stress response and HHPA-axis function in rats, affecting among others GR levels (Poimenova et al., 2010; Panagiotidou et al., 2014). Here, we set out to investigate if perinatal BPA exposure leads to DNA methylation changes in the GR gene (Nr3c1) and/or in its regulator Fkbp5. We used the hippocampi and hypothalami from 46-day old rats whose mothers had been exposed to 40 mg BPA/kg bw/day during gestation and lactation, a dose belonging to the low-dose exposure range for BPA, repeatedly reported to exert adverse effects in experimental animals (Vandenberg et al., 2012), and investigated a region in the Nr3c1 promoter (Weaver et al., 2004) and the intron 5 of Fkbp5. No significant difference was found in the Nr3c1 promoter in either brain areas (Table 1), whereas in male hippocampi, we found a significant increase for methylation at Fkbp5 intron 5, an important regulatory region that is differentially methylated upon corticoid treatment (Lee et al., 2010) (Fig. 1a, Table 1). We also investigated methylation at Fkbp5 intron 1, another regulatory region of this gene (Lee et al., 2010). No significant differences were found in this region (Table 1). Next, we examined if FKBP51 levels are affected in male hippocampi. At basal conditions FKBP51 protein levels were significantly lower in the hippocampus of BPA-exposed male rats compared to the controls (p ¼ 0.045) (Fig. 1b and Supplemental material, Fig. S1a). Two hours following acute swimming stress, FKBP51 levels increased compared to baseline in BPA-treated animals (p ¼ 0.016), whereas they did not change significantly in untreated controls (Fig. 1b and Supplemental material, Fig. S1a).

194


Table 1 Percentage DNA methylation of Nr3c1 (glucocorticoid receptor) and Fkbp5 in hypothalamus and hippocampus assessed with bisulfite-pyrosequencing, BPA vs. control. Values that differ significantly (p < 0.05) between BPA and control treated are in bold. Gene

Sex

Hypothalamic Nr3c1 Hypothalamic Fkbp5 intron 5 Hippocampal Nr3c1 Hippocampal Fkbp5 intron 5 Hippocampal Fkbp5 intron 1

CpG position

m f m f m f m f m f

1

2

3

4

5

6

7

1.67/2.40 1.96/2.49 14.30/15.63 14.75/14.68 0.30/0.42 0.74/0.15 20.97/14.38 13.64/18.07 85.61/85.94 87.55/85.75

1.20/1.68 1.43/1.87 12.26/13.13 11.05/11.96 1.04/1.26 1.40/0.99 17.01/12.58 15.73/18.19 93.77/93.87 95.61/92.98

0.50/0.00 0.17/2.03 20.60/19.49 19.11/19.25 1.61/0.80 1.77/1.41 12.08/6.36 5.96/9.01 51.99/48.94 59.80/58.00

1.47/1.73 1.63/1.72 14.89/18.56 19.21/19.43 1.41/1.62 1.59/1.26 18.40/10.88 11.07/14.69

1.22/1.36 1.35/2.26 10.15/9.69 10.45/10.40 1.99/1.32 2.63/2.44 18.28/14.00 15.21/18.78

1.27/1.59 1.36/1.57 18.60/19.12 20.07/18.68 1.62/1.51 1.69/1.72 9.21/6.20 5.62/8.11

1.80/2.04 2.00/2.08

Thus, perinatal BPA exposure leads to sexually dimorphic changes in hippocampal Fkbp5 methylation approximately one month after exposure has finished. At basal conditions, methylation changes coincide with reduced FKBP51 levels in the hippocampus of male offspring. However, following stress, FKBP51 levels are increased to the levels of untreated controls.

3.2. BPA treatment of differentiating HT22 cells induces changes in Fkbp5 expression and methylation To address the mechanisms underlying the expression and methylation changes of Fkbp5 in BPA-exposed male rats, we employed a cell model. We chose HT22 cells, a murine hippocampal cell line derived from male mouse primary hippocampal cells that can be differentiated from proliferating mitotic cells to cholinergic neurons displaying neurite outgrowth (Liu et al., 2009). To mimic developmental exposure in animals, these cells were treated with increasing amounts of BPA during the 2-days differentiation

a 30

control BPA

% methylation

*

*

20

* **

** 10

7

6

pG C

5

pG C

4

pG C

3

pG C

pG C

pG C

C

pG

1

2

0

b Fkbp51/ Gapdh ratio

2.0

† 1.5

control BPA

1.0

*

0.5 0.0

basal

post stress

Fig. 1. Perinatal exposure to BPA induces expression and DNA methylation changes of FKBP51 in rat hippocampus. a: DNA methylation at Fkbp5 intron 5 in male rat hippocampus assessed by bisulfite-pyrosequencing. b: Quantification of western blot analyses of FKBP51 protein levels in male rat hippocampus at basal conditions and 2 h after a 15 min-swimming stress. All bars represent the means of at least 3 animals, analysed in 3 independent experiments. *p < 0.05, **p < 0.01 BPA-treated vs. untreated, yp < 0.05 basal vs. stressed.

1.49/1.90 1.42/1.95 15.22/9.33 9.50/11.94

procedure and subsequently kept without BPA for another 3 days (Fig. 2a) before analysis. BPA treatment decreased Fkbp5 gene expression in a dose-dependent fashion with significant changes at doses as low as 10 nM (Fig. 2b). This was also reflected at the protein level showing a significant decrease upon BPA treatment (Fig. 2c and Supplemental material, Fig. S1b). We then analysed Fkbp5 methylation upon treatment with 10 and 1000 nM BPA at the region in intron 5 that is homologous to the regions investigated in the rats (Fig. 2d). We focused on the CpGs 5 and 6, corresponding to CpGs 6 and 7 in the rat, which are part of a ~200 bp region conserved between mouse and rat containing a GRE. BPA treatment led to an increase in methylation at these two CpGs, which reached significance at CpG 5 for both doses (Fig. 2e). Thus, BPA exposure of HT22 cells during differentiation leads to increased DNA methylation and decreased expression of Fkbp5 3 days after BPA treatment was stopped, similarly to the changes observed in vivo. 3.3. BPA treatment in HT22 cells changes Fkbp5 inducibility upon glucocorticoid treatment As part of the negative feedback regulation of the HPA-axis, GR, activated by elevated glucocorticoid levels during stress, increases Fkbp5 expression by binding to GREs in regulatory regions of the gene, among others the one in intron 5 (Hubler and Scammell, 2004). We thus analyzed if Fkbp5 inducibility by glucocorticoids changes upon BPA treatment in the cell model. To this end, we differentiated HT22 cells in the absence and presence of 10 and 1000 nM BPA as described above. Additionally, cells were treated with a single dose of dexamethasone (dex), a stable synthetic glucocorticoid, 16e18 h before Fkbp5 expression was assessed on day 3 (Fig. 2a). We used 1 mM dex, a concentration commonly used in cell experiments, e.g. (Klengel et al., 2013b), that induces Fkbp5 expression (Nehme et al., 2009). As expected, Fkbp5 expression increased upon dex-treatment in the absence of BPA (Fig. 3, left grey vs. black bar). Similar inducibility with 10 and 100 nM dex were observed (data not shown). In cells exposed to BPA during differentiation, basal Fkbp5 was decreased significantly whereas upon dex-stimulation, the increase of Fkbp5 expression was even higher than in the absence of BPA (Fig. 3, middle and right grey vs. black bars). Thus, BPA treatment during HT22 differentiation changes inducibility of Fkbp5 in response to glucocorticoids, similarly to the rat model. 3.4. The estrogen receptor inhibitor ICI 182 780 affects Fkbp5 expression and methylation similarly to BPA Next, we investigated if ERs are involved in the effects induced by BPA. Both receptors are present in HT22 cells (Supplemental


195

a Change to medium w/o BPA/ICI 2 days Differentiation +/- BPA, +/- ICI, +/- E2

Harvest 2 days

12-16h Dex treatment

b relative expression

1.5

1.0

*** 0.5

***

***

***

B PA

B PA

uM

uM 1

10

B PA

B PA

nM

nM

1

10

10 0

l ro

nM

nt co

B PA

0.0

c Fkbp51/actin ratio

1.5

1.0

*

0.5

1

uM

co

nt

ro

l

B PA

0.0

d rFkbp5 intron 5

369 bp

mFkbp5 intron 5

368 bp

e

% methylation

70

*

*

no BPA 10 nM BPA 1 uM BPA

60 50 40

6 pG C

C

pG

5

30

Fig. 2. BPA treatment during differentiation induces expression and DNA methylation changes of Fkbp5 in HT22 cells. a: Design of the cell experiments. Dexamethasone (Dex) treatment was carried out for the experiments in Fig. 3. Estradiol (E2) treatment was carried out for the experiments in Fig. 4. b: Fkbp5 gene expression in HT22 cells after treatment with a range of BPA doses during cell differentiation. Expression was assessed by real-time PCR 3 days after BPA treatment was stopped. c: Quantification of western blot analyses of FKBP51 protein levels in HT22 after treatment with 1 mM BPA. d: Schematic view of the rat and mouse Fkbp5 intron 5. Vertical lines represent the analysed CpGs, the shaded bar depicts the GRE. e: DNA methylation at Fkbp5 intron 5 after treatment with 2 different doses of BPA during cell differentiation, assessed by bisulfite-pyrosequencing. All bars represent the means of at least 3 independent experiments and error bars the standard deviation. *p < 0.05, ***p < 0.001, BPA-treated vs. untreated.

material, Fig. S2a,b). We tested if estradiol (E2) treatment during differentiation affects Fkbp5 expression, and if inhibition of ERs with ICI 182 780 (ICI) changes the effects of BPA. HT22 cells were differentiated as above in the absence or presence of 10 nM E2 or 1 mM ICI, which are commonly used doses of the hormone and antagonist (e.g. (Dauvois et al., 1992)). Gene expression was assessed 3 days after the differentiation was finished and the substances had been washed out (Fig. 2a). ICI treatment led to a decrease in Fkbp5 expression (Fig. 4a, left grey vs. black bar), similar to that observed upon BPA treatment (Fig. 4a, middle black bars). In accordance, FKBP51 protein levels were decreased by both BPA and ICI treatment and their combination (Fig. 4b, Supplemental Fig. S1c). E2, on the other hand, did not affect Fkbp5 expression by itself, but reverted the decrease induced by ICI (Fig. 4a, bars to the right). We also tested if these treatments affect Fkbp5 expression through direct transcriptional activation. Acute treatment with BPA

or E2 in the absence or presence of ICI for 6 h did not induce any differences in Fkbp5 expression (Fig. 4c). On the other hand, E2 readily activated an ERE-luciferase reporter construct transiently transfected into HT22 cells (Fig. 4d), demonstrating that ERE-driven transcription can be activated in these cells. BPA had no effect on luciferase activity by itself but reverted E2 induced transcription (Fig. 4d). As ICI treatment had similar effects to BPA on Fkbp5 transcription, we also investigated its effects on DNA methylation at CpG 5 and 6. As shown in Fig. 4e, ICI increased Fkbp5 methylation to a similar extent as BPA. In the presence of both ICI and BPA during differentiation, Fkbp5 methylation increased even further, in particular for 1 mM BPA at CpG6 (Fig. 4e, middle and right grey vs. black bars). Thus, the ER inhibitor ICI affects Fkbp5 methylation and expression similarly to BPA, suggesting an involvement of the ERs in BPA's effects, however, not via a classical ERE-driven transcriptional mechanism.

196

E. Kitraki et al. / Molecular and Cellular Endocrinology 417 (2015) 191e199 ###

*

###

15

*

10

control treatment 1 uM dex

###

5 2.0 1.5 1.0 0.5 0.0

***

B PA

***

1

uM

B PA nM 10

co

nt

ro

l

relative expression

20

Fig. 3. BPA treatment during differentiation changes inducibility of Fkbp5 expression by Dex. Fkbp5 gene expression in HT22 cells after exposure to BPA during cell differentiation, and an overnight treatment with Dexamethasone (dex) assessed by realtime PCR. All bars represent the means of at least 3 independent experiments and error bars the standard deviation. *p < 0.05, ***p < 0.005, BPA-treated vs. untreated, ### p < 0.001 Dex-treated vs. untreated.

3.5. ERb is involved in the effects of BPA on Fkbp5 To better understand the role of the ERs in regulating Fkbp5 transcription, we investigated if ERa and/or ERb is recruited to the regulatory region of Fkbp5 at intron 5. To this end, we conducted chromatin immunoprecipitation assays in HT22 cells, analyzing ERa and ERb binding to the differentially methylated region in intron 5 upon treatment with BPA and ICI. As shown in Fig. 5a, ERb was significantly enriched at this region, both in the absence and presence of E2. On the other hand, no significant enrichment could be found for ERa with two different antibodies (data not shown). Upon ICI treatment, ERb binding to intron 5 was lost (Fig. 5a), supporting a specific enrichment at this region. Following BPA treatment, ERb recruitment decreased significantly compared to control (Fig. 5a). We then tested a direct involvement of ERb in mediating BPA's effect on Fkbp5. Using shRNA, we knocked-down ERb before induction of the differentiation process (Fig. 5b). Knock-down of ERb lead to a significant decrease in Fkbp5 expression (Fig. 5c black bars, left panel) and to a significant increase in Fkbp5 methylation (Fig. 5c black bars, right panel). BPA treatment during differentiation decreased expression and increased methylation when cells were transfected with control shRNA. However, no further effect was observed when ERb was knocked-down (Fig. 5c grey bars). Together, these findings implicate that ERb is involved in the effects of BPA on Fkbp5 transcription and methylation.

4. Discussion The HHPA axis is essential for the organism to cope with stress. Dysfunction of this system has been associated with a number of psychiatric illnesses such as major depression and post-traumatic stress disorder. While major research efforts have been made to identify genetic components involved in psychiatric diseases, much less is known about environmental factors contributing to these disorders and the underlying mechanisms. In this study, we demonstrate both in an in vivo and in an in vitro model that BPA, an abundant chemical that disrupts estrogen signaling, can induce lasting changes in the regulation of Fkbp5, an important regulator of the HHPA axis. We show that perinatal exposure to a low, human-relevant, BPA dose (Vandenberg et al., 2012) leads to epigenetic changes in the Fkbp5 gene in the hippocampus of male rats about one month after BPA exposure had finished. Increased methylation coincided with reduced FKBP51 levels at basal conditions that imply an increased alertness towards threatening insults. Notably, upon stress, FKBP51

levels of these animals increase, suggesting that BPA-treated male rats become equally efficient as controls to terminate an acute stress response. This competency complies well with previous results where BPA-treated males exhibited similar behavioral coping and corticosterone responses to stress as the untreated animals (Panagiotidou et al., 2014). The finding that a short-term stress can modify FKBP51 levels despite the BPA-induced methylation changes, suggests an enhanced expression plasticity of this regulator that was also observed in the HT22 cell line deriving from male mouse hippocampus. Our findings in male rats are in good agreement with the results obtained in HT22 cells and demonstrate that this cell line is a valid model to study the mechanisms underlying the effects of BPA exposure on Fkbp5 regulation. In the cell line, BPA exposure during differentiation lead to decreased Fkbp5 expression and increased methylation at the corresponding CpGs in intron 5. Of note, these changes were not due to acute effects of BPA as they were observed 3 days after BPA washout. Similarly to the rat model, BPA's effect on Fkbp5/FKBP51 levels was larger than on Fkbp5 methylation. This is at least partly due to the fact that the detection methods do not have the same sensitivity, which makes comparison of the results difficult. Further, other factors and/or genomic regions not investigated here, but involved in Fkbp5 regulation, might be affected by BPA. Interestingly, dexamethasone treatment, simulating the stressful situation in the rat model, increased Fkbp5 expression to a greater extend in BPA-treated than in the untreated cells. The reason for decreased basal expression of Fkbp5 but increased inducibility of the gene by glucocorticoids despite an increase in DNA methylation is not solved. One possibility is that methylation changes at the responsive element under study affect mineralocorticoid receptor binding, responsible for mediating the effects of basal glucocorticoid levels, but not GR binding that follows the increase of hormone levels. Although sharing the DNA recognition sequence, the two receptors are modulated in their activity by diverse co-activators and co-repressors that might be affected differently by the methylation changes. Further studies are necessary to understand the detailed implications of the DNA methylation changes induced by BPA. The ER inhibitor ICI affected Fkbp5 expression and methylation during HT22 differentiation similarly to BPA. Furthermore, ERb, but not ERa, bound to the differentially methylated region at intron 5, and this binding was disrupted by BPA and ICI. These results imply a function of ERb in mediating BPA's effects on Fkbp5, which was further supported by the fact that ERb knock-down abolished the effects of BPA on Fkbp5 expression and DNA methylation. Notably, however, the effects found here cannot be ascribed to BPA's estrogenic properties as i) they were observed three days after BPA had been washed out of the culture and ii) BPA did not display agonistic properties in the HT22 cells on an ERE-driven transcription, in contrast to E2 (Fig. 4c). Further, there is an agreement in the literature that BPA does not induce the same conformational changes as the natural ligands when binding to ERs (Wetherill et al., 2007; Delfosse et al., 2012), hence presumably attracting a different set of co-regulatory proteins than in the presence of natural ligand. Accordingly, we propose that BPA affects Fkbp5 transcriptional regulation by interfering with ERb binding to the regulatory region of intron 5, where ERb controls DNA methylation, a function of ERb that we have described previously (Ruegg et al., 2011). A tentative model is depicted in Supplemental material, Fig. S3. We could not see any effect of BPA on ER or GR protein expression in our cell model (Supplemental material, Fig. S2b and c). Further, ERb levels were not changed in the hippocampi of the treated rats (Supplemental material, Fig. S2d). However, in mice it was shown that BPA exposure leads to decreased ERb levels in brain areas other than the hippocampus (Cao et al., 2014). Thus in other regions BPA


b

a $

1.5

control treatment 10 uM ICI

Fkbp51/actin ratio

1.0

**

#

# #

0.5

***


1.0

*

#

0.5

#

0.0

B PA

l ro nt

1

uM

co

10

1

10

uM

nM

nM

B PA

B PA

l ro nt co

E2

0.0

d

0.004 0.002

uM 1

m at re

B PA

t en

B PA uM

B PA

0.000

10

nM

co

nt

E2

ro

+

lt

1

10

co

nM

nt

ro

l

E2

0.0

0.006

uM

0.5

$$$

0.008

1

1.0

0.010

nM

1.5

relative luciferase activity

no ICI 1 uM ICI

E2

c

10

relative expression

1.5

relative Fkbp5 expression

197

CpG 6

CpG 5

e

50

80

#

70

*

#

*

60

% methylation


40 35 30

1

nM 10

uM

B PA

l ro nt

B PA 1

nM 10

uM

B PA

l ro nt co

B PA

25

50

co

% methylation

#

45

Fig. 4. ICI treatment during differentiation induces similar changes in expression and DNA methylation of Fkbp5 as BPA. a: Fkbp5 gene expression in HT22 cells after treatment with BPA and E2 alone or in combination with ICI during cell differentiation, assessed by real-time PCR. **p < 0.005, ***p < 0.001 BPA-treated vs. untreated control, #p < 0.05 ICI-treated vs. untreated control, $ p < 0.05 E2þICI treated vs. ICI treated only. b: Quantification of western blot analyses of FKBP51 protein levels in HT22 after treatment with BPA or ICI alone or in combination. *p < 0.05 BPA-treated vs. untreated control, #p < 0.05 ICI-treated vs. untreated control. c: Fkbp5 gene expression in HT22 cells immediately after a 6 h-treatment with BPA and E2 alone or in combination with ICI, assessed by real-time PCR. d: ER-activity in HT22 cells upon direct induction with E2 and BPA. HT22 cells were transfected with plasmids encoding for the reporter gene 3 ERE-luc and TK-renilla, and treated with E2, BPA or the combination of both. After 16 h, firefly luciferase activity was measured and normalised against renilla luciferase activity. $$$p < 0.001 E2 vs. all other treatments. e: DNA methylation at Fkbp5 intron 5 after treatment with BPA or ICI alone or in combination during cell differentiation, assessed by bisulfite-pyrosequencing. *p < 0.05 BPA-treated vs. untreated control, #p < 0.05 ICI-treated vs. untreated control. All bars represent the means of at least 3 independent experiments and error bars the standard deviation.

might not affect DNA binding of ERb but rather its protein levels. Ultimately, however, this will lead to the same result, a lack of ERb binding to intron 5 of Fkbp5 and thus an increase in DNA methylation. Interestingly, BPA seems to affect ERb expression in the rodent brain in a sexual dimorphic manner ((Cao et al., 2014) and data not shown). This might explain why we could not detect any

methylation changes in female rats at the investigated regions of Nr3c1 and Fkbp5. Sexually dimorphic effects of BPA in brain function have been reported in several previous in vivo studies (Cao et al., 2014; Cao et al., 2013; Chen et al., 2014; Gioiosa et al., 2013; Jasarevic et al., 2013; Wolstenholme et al., 2011; Xu et al., 2015). Furthermore, the few epidemiological studies linking BPA exposure to neuropsychiatric outcomes in children also show differences

198


a

ERb recruitment #

#

**

0.100

IgG anti-ERb

% input

0.075

*

0.050 0.025

E2 nM

IC I 10

uM 1

1

uM

co

nt

ro

l

B PA

0.000

b ERb expression relative expression

1.5

1.0

0.5

###

sh

sh

co

ER

nt

be

ro

l

ta

0.0

c Fkbp5 methylation

Fkbp5 expression

#

1.0

*

0.8

#

% methylation

0.6 0.4

*

60

50

0.2

sh

nt co

ER

l ro

b sh

ER

l ro nt co sh

b

40

0.0

sh

relative expression

control 1 uM BPA

70

1.2

Fig. 5. ERb is involved in the effects of BPA on Fkbp5 expression and methylation. a: ERb binding after BPA, ICI, or E2-treatment in HT22 cells assessed by chromatin immunoprecipitation. ERb was precipitated using an anti-ERb antibody and the region at Fkbp5 intron 5 was amplified from co-precipitated DNA using qRT-PCR and normalised to input sample. *p < 0.05, **p < 0.005 enrichment ERb antibody vs. IgG, #p < 0.05 ERb enrichment in BPA- or ICI-treated vs. untreated cells. b: ERb gene expression in HT22 cells transfected with 2 different shRNA sequences against ERb or a control shRNA, assessed by real-time PCR. ###p < 0.001 shERb vs. shcontrol c: Fkbp5 gene expression (left panel) and methylation (right panel) upon BPA treatment during differentiation in HT22 cells transfected with 2 different shRNA sequences against ERb or a control shRNA. *p < 0.05 BPA-treated vs. untreated control, #p < 0.05 shERb vs. shcontrol. All bars represent the means and error bars the standard deviation.

between girls and boys (Braun et al., 2011; Harley et al., 2013; Evans et al., 2014). This demonstrates the intricate interaction between BPA and the endogenous sex hormones and consequently the importance to investigate its effects on both sexes.

that requires further studies in the in vivo setting. The BPA-induced changes in hippocampal Fkbp5 confer a link between environmental chemicals and stress-related disorders. Financial disclosures

5. Conclusion We demonstrate here that perinatal exposure of rats to a low BPA dose alters Fkbp5 expression, methylation pattern and inducibility by stress in the hippocampus of male offspring. The observed alterations in Fkbp5 were also detected in differentiating hippocampal neurons of male origin. In the cell model, the mechanism implicates ERb in the regulation of the epigenetic impact, a finding

The authors have no competing financial interests. Acknowledgments €lle Rüegg was supported by the Swiss National Science Dr. Joe Foundation (PZ00P3_126319) and the Swedish Research Council Formas.


Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.mce.2015.09.028. References Braun, J.M., et al., 2009. Prenatal bisphenol A exposure and early childhood behavior. Environ. Health Perspect. 117 (12), 1945e1952. Braun, J.M., et al., 2011. Impact of early-life bisphenol A exposure on behavior and executive function in children. Pediatrics 128 (5), 873e882. Calafat, A.M., et al., 2008. Exposure of the U.S. population to bisphenol A and 4tertiary-octylphenol: 2003e2004. Environ. Health Perspect. 116 (1), 39e44. Cao, J., et al., 2013. Prenatal bisphenol A exposure alters sex-specific estrogen receptor expression in the neonatal rat hypothalamus and amygdala. Toxicol. Sci. 133 (1), 157e173. Cao, J., et al., 2014. Sex-specific Esr2 mRNA expression in the rat hypothalamus and amygdala is altered by neonatal bisphenol A exposure. Reproduction 147 (4), 537e554. Chen, F., et al., 2014. Sex differences in the adult HPA axis and affective behaviors are altered by perinatal exposure to a low dose of bisphenol A. Brain Res. 1571, 12e24. Cortazar, D., et al., 2011 Feb 17. Embryonic lethal phenotype reveals a function of TDG in maintaining epigenetic stability. Nature 470 (7334), 419e423. Dauvois, S., et al., 1992. Antiestrogen ICI 164,384 reduces cellular estrogen receptor content by increasing its turnover. Proc. Natl. Acad. Sci. U. S. A. 89 (9), 4037e4041. Delfosse, V., et al., 2012. Structural and mechanistic insights into bisphenols action provide guidelines for risk assessment and discovery of bisphenol A substitutes. Proc. Natl. Acad. Sci. U. S. A. 109 (37), 14930e14935. Delfosse, V., et al., 2014. Nuclear receptor profiling of bisphenol-A and its halogenated analogues. Vitam. Horm. 94, 229e251. Evans, S.F., et al., 2014. Prenatal bisphenol A exposure and maternally reported behavior in boys and girls. Neurotoxicology 45C, 91e99. Fries, G.R., et al., 2014 Oct 31. Hypothalamic-pituitary-adrenal axis dysfunction and illness progression in bipolar disorder. Int. J. Neuropsychopharmacol. 18 (1). Gioiosa, L., et al., 2013. The effects of bisphenol A on emotional behavior depend upon the timing of exposure, age and gender in mice. Horm. Behav. 63 (4), 598e605. Harley, K.G., et al., 2013. Prenatal and early childhood bisphenol A concentrations and behavior in school-aged children. Environ. Res. 126, 43e50. Holsboer, F., Ising, M., 2010. Stress hormone regulation: biological role and translation into therapy. Annu. Rev. Psychol. 61, 81e109. C1-11. Hubler, T.R., Scammell, J.G., 2004. Intronic hormone response elements mediate regulation of FKBP5 by progestins and glucocorticoids. Cell Stress Chaperones 9 (3), 243e252. Jasarevic, E., et al., 2013. Sex and dose-dependent effects of developmental exposure to bisphenol A on anxiety and spatial learning in deer mice (Peromyscus maniculatus bairdii) offspring. Horm. Behav. 63 (1), 180e189. Klengel, T., et al., 2013. Allele-specific FKBP5 DNA demethylation mediates genechildhood trauma interactions. Nat. Neurosci. 16 (1), 33e41. Klengel, T., et al., 2013. Allele-specific FKBP5 DNA demethylation mediates genechildhood trauma interactions. Nat. Neurosci. 16 (1), 33e41. Kundakovic, M., Champagne, F.A., 2011. Epigenetic perspective on the

199

developmental effects of bisphenol A. Brain Behav. Immun. 25 (6), 1084e1093. Lee, R.S., et al., 2010. Chronic corticosterone exposure increases expression and decreases deoxyribonucleic acid methylation of Fkbp5 in mice. Endocrinology 151 (9), 4332e4343. Legler, J., et al., 1999. Development of a stably transfected estrogen receptormediated luciferase reporter gene assay in the human T47D breast cancer cell line. Toxicol. Sci. 48 (1), 55e66. Liu, J., Li, L., Suo, W.Z., 2009. HT22 hippocampal neuronal cell line possesses functional cholinergic properties. Life Sci. 84 (9e10), 267e271. Nehme, A., et al., 2009. Glucocorticoids with different chemical structures but similar glucocorticoid receptor potency regulate subsets of common and unique genes in human trabecular meshwork cells. BMC Med. Genomics 2, 58. Panagiotidou, E., et al., 2014. Perinatal exposure to low-dose bisphenol A affects the neuroendocrine stress response in rats. J. Endocrinol. 220 (3), 207e218. Poimenova, A., et al., 2010. Corticosterone-regulated actions in the rat brain are affected by perinatal exposure to low dose of bisphenol A. Neuroscience 167 (3), 741e749. Riggs, D.L., et al., 2003. The Hsp90-binding peptidylprolyl isomerase FKBP52 potentiates glucocorticoid signaling in vivo. EMBO J. 22 (5), 1158e1167. Ruegg, J., et al., 2011. Epigenetic regulation of glucose transporter 4 by estrogen receptor beta. Mol. Endocrinol. 25 (12), 2017e2028. Schmidt, M.V., et al., 2012. The prospect of FKBP51 as a drug target. ChemMedChem 7 (8), 1351e1359. Smith, S.M., Vale, W.W., 2006. The role of the hypothalamic-pituitary-adrenal axis in neuroendocrine responses to stress. Dialogues Clin. Neurosci. 8 (4), 383e395. Touma, C., et al., 2011. FK506 binding protein 5 shapes stress responsiveness: modulation of neuroendocrine reactivity and coping behavior. Biol. Psychiatry 70 (10), 928e936. Vandenberg, L.N., et al., 2012. Hormones and endocrine-disrupting chemicals: lowdose effects and nonmonotonic dose responses. Endocr. Rev. 33 (3), 378e455. Vermeer, H., et al., 2003. Glucocorticoid-induced increase in lymphocytic FKBP51 messenger ribonucleic acid expression: a potential marker for glucocorticoid sensitivity, potency, and bioavailability. J. Clin. Endocrinol. Metab. 88 (1), 277e284. Weaver, I.C., et al., 2004. Epigenetic programming by maternal behavior. Nat. Neurosci. 7 (8), 847e854. Wetherill, Y.B., et al., 2007. In vitro molecular mechanisms of bisphenol A action. Reprod. Toxicol. 24 (2), 178e198. Wochnik, G.M., et al., 2005. FK506-binding proteins 51 and 52 differentially regulate dynein interaction and nuclear translocation of the glucocorticoid receptor in mammalian cells. J. Biol. Chem. 280 (6), 4609e4616. Wolstenholme, J.T., et al., 2011. Gestational exposure to low dose bisphenol A alters social behavior in juvenile mice. PLoS One 6 (9), e25448. Wolstenholme, J.T., et al., 2012. Gestational exposure to bisphenol a produces transgenerational changes in behaviors and gene expression. Endocrinology 153 (8), 3828e3838. Wolstenholme, J.T., Rissman, E.F., Connelly, J.J., 2011. The role of Bisphenol A in shaping the brain, epigenome and behavior. Horm. Behav. 59 (3), 296e305. Xu, X., et al., 2015. Sex-specific effects of long-term exposure to bisphenol-A on anxiety- and depression-like behaviors in adult mice. Chemosphere 120, 258e266. Yang, X., et al., 2012. Glucocorticoid-induced loss of DNA methylation in nonneuronal cells and potential involvement of DNMT1 in epigenetic regulation of Fkbp5. Biochem. Biophys. Res. Commun. 420 (3), 570e575.