BPA Directly Decreases GnRH Neuronal Activity via Noncanonical Pathway

Abstract

Peripheral feedback of gonadal estrogen to the hypothalamus is critical for reproduction. Bisphenol A (BPA), an environmental pollutant with estrogenic actions, can disrupt this feedback and lead to infertility in both humans and animals. GnRH neurons are essential for reproduction, serving as an important link between brain, pituitary, and gonads. Because GnRH neurons express several receptors that bind estrogen, they are potential targets for endocrine disruptors. However, to date, direct effects of BPA on GnRH neurons have not been shown. This study investigated the effects of BPA on GnRH neuronal activity using an explant model in which large numbers of primary GnRH neurons are maintained and express many of the receptors found in vivo. Because oscillations in intracellular calcium have been shown to correlate with electrical activity in GnRH neurons, calcium imaging was used to assay the effects of BPA. Exposure to 50μM BPA significantly decreased GnRH calcium activity. Blockage of γ-aminobutyric acid ergic and glutamatergic input did not abrogate the inhibitory BPA effect, suggesting direct regulation of GnRH neurons by BPA. In addition to estrogen receptor-β, single-cell RT-PCR analysis confirmed that GnRH neurons express G protein-coupled receptor 30 (G protein-coupled estrogen receptor 1) and estrogen-related receptor-γ, all potential targets for BPA. Perturbation studies of the signaling pathway revealed that the BPA-mediated inhibition of GnRH neuronal activity occurred independent of estrogen receptors, GPER, or estrogen-related receptor-γ, via a noncanonical pathway. These results provide the first evidence of a direct effect of BPA on GnRH neurons.

Endocrine-disrupting chemicals (EDCs) are environmental pollutants that interfere with the endocrine system. Some EDCs such as bisphenol A (BPA) have estrogenic properties and, as such, can hamper the proper action of estradiol, a key regulator of reproductive function. Exposure of mice and rats to estrogenic EDCs during development impacts pubertal onset and induces reproductive failure in adults (1). GnRH neurons, the last neuronal output controlling fertility, exhibit detrimental consequences of estrogenic EDCs. Both in vitro and in vivo studies (pre- and postnatal exposure) indicate that EDCs can alter levels of GnRH gene expression, GnRH release, number of GnRH neurons and their morphology (reviewed in Ref. 1). However, it is unclear whether these changes are due to direct or indirect effects of BPA on GnRH neurons.

Currently, 5 estrogen receptors (ERs) have been identified: classical ERs, ERα and ERβ), membrane-bound ERs (mERs) (mERα and mERβ), STX-sensitive membrane ER, G protein-coupled estrogen receptor 1 (GPER), and the plasma membrane-associated, estrogen receptor ER-X; and they are widely distributed in the central nervous system (2). The literature also describes 3 estrogen-related receptors (ERRs) (ERRα, ERRβ, and ERRγ), which are nuclear receptor transcription factors with constitutive activities (3). ERRs display a low affinity for estrogen. However, they can interact with estrogen signaling by modulating estrogen response elements (4, 5) and can enhance estrogen responsiveness of tissues with low levels of ERs (6). Although the endogenous ligands for ERRs are unknown, ERRγ has a high affinity for BPA (7, 8). Like ERs, they are widely expressed throughout the brain during fetal and adult life (9).

GnRH neurons are known to express 3 ERs, ERβ (10), G protein-coupled receptor 30 (GPER) (11, 12), and STX-sensitive membrane ER (13). Although BPA can bind to some of these receptors, a direct effect of BPA on GnRH neurons has not been investigated. Thus, the objectives in this study were to determine whether acute BPA affects GnRH neuronal activity and if so, which cellular mechanisms are involved. Using calcium imaging and explants containing primary GnRH cells, our data show that BPA can induce a rapid inhibition of GnRH neuronal activity via indirect mechanisms partially dependent on ERs and GPER and direct mechanisms independent of ERs, GPER and ERRγ.

Materials and Methods

Nasal explants

All procedures were approved by National Institute of Neurological Disorder and Stroke, Animal Care and Use Committee and performed in accordance with National Institutes of Health guidelines.

Nasal regions were cultured as previously described (14, reviewed in Ref. 15). Briefly, embryos were obtained from (embryonic day 11.5)-timed pregnant NIH Swiss mice. Nasal pits were isolated under aseptic conditions in Gey's balanced salt solution (Life Technologies, Inc) enriched with glucose (Sigma Chemical Co). Explants were adhered onto coverslips by a plasma (Cocalico Biologicals)/thrombin (Sigma) clot and maintained at 37°C in a defined serum-free medium (SFM) in a humidified atmosphere with 5% CO₂. On culture day 3, fresh medium containing fluorodeoxyuridine (80μM; Sigma) was applied for 3 days to inhibit proliferation of dividing olfactory neurons and nonneuronal explant tissue. On culture day 6, and every 2 days afterward, the medium was changed with fresh SFM. Explants were used between 6 and 10 days in vitro (div) for calcium imaging.

Calcium imaging

Experiments were performed as previously described (16). Briefly, Calcium Green-1 AM (Life Technologies) was dissolved at 2.7mM in 80% dimethyl sulfoxide and 20% pluronic F-127 (Life Technologies), diluted at 13.5μM in SFM and aliquoted. Explants were incubated in warm loading solution for 20 minutes at 37°C in a 5% CO₂ humidified incubator. After 2 10-minute washes in fresh SFM, explants were mounted in a perfusion chamber (Warner Instruments) and continuously perfused. Calcium Green-1-loaded cells were visualized using an inverted Nikon microscope, through a ×20 fluorescence objective and a charge-coupled device camera (QImaging) connected to a computer. Time lapse was piloted by imaging software (iVision; Scanalytics, Inc) and pictures acquired every 2 seconds for up to 33 minutes. Excitation wavelengths were provided via a medium-width excitation bandpass filter at 465–495 nm, and emission was monitored through a 40 nm bandpass centered on 535 nm. All recordings were terminated with a 40mM KCl stimulation to ensure the viability of the cells. The changes of fluorescence over time were measured in single GnRH neurons a posteriori with iVision and analyzed with MATLAB (Mathworks) as previously described (17). Briefly, calcium oscillations were monitored as a reflection of neuronal activity (17). A calcium elevation had to be greater than the mean of the 5 previous and 5 next points plus a minimal value (which represented small fluctuations in baseline) to be considered as a calcium oscillation or peak. The frequency of calcium oscillations was expressed in peaks per minute. The phenotype of the cells analyzed was confirmed immunocytochemically using GnRH specific antibodies as previously described (Figure 1A) (18).

Figure 1.

Calcium imaging of GnRH neurons. A, GnRH cells were identified by their bipolar morphology (bright field) (A₁), loaded with fluorescent calcium-sensitive dye (fluorescence) (A₂), and their identity verified postimaging by immunocytochemistry (immunofluorescence) (A₃). Arrows indicate identical cells in all fields. B, Representative recording showing spontaneous baseline oscillations in intracellular calcium levels in a single GnRH neuron during 15 minutes in SFM (y-scale = optical density, OD units).

PCR on single cells and explants

Poly(A) amplified cDNA libraries have been created from individual GnRH cells maintained in explants (19). The phenotype and the quality of each single-cell cDNA pool was confirmed by PCR for GnRH, L19, and β-tubulin (20). The same method was used to generate cDNA libraries from whole explants (19). Primers were then designed in the 3′-untranslated region of the genes ERα, GPER, and ERRγ, within the 300 base pairs before the polyadenylation site. All designed primers were screened using BLAST to ensure specificity of binding.

For each reaction, 30.5-μL H₂O, 5-μL 10× PCR buffer, 4-μL 25mM MgCl₂, 5-μL deoxynucleotide mix (25 μL of each 100mM deoxynucleotide and 900-μL H₂O), 2-μL 6.25μM forward primer, 2-μL 6.25μM reverse primer, and 0.5-μL AmpliTaq Gold were added to 1-μL template cDNA. PCR was performed as following: initial 10-minute denaturation (94°C); 40 cycles with denaturation 30 seconds (94°C); annealing 30 seconds (as mention for each primer set) and extension 2 minutes (72°C); followed by a 10-minute post-elongation at 72°C. Amplified products were run on a 1.5% agarose gel. Specific bands of the predicted size were observed in the control total brain lane, whereas no bands were seen in water (negative control). The sequences of the primers were: ERα (annealing 62°C) [5′-tttttcggggagaggcacag-3′][5′-ttacaggggcttgagcatcc-3′], GPER (annealing 60°C) [5′-ctctgggatgctcctctcac-3′][5′-agctcgctgttcccagtatg-3′], and EERγ (annealing 62°C) [5′-gtgcaaagatcgtgaatgga-3′][5′-catttgctggagcacacagt-3′].

Immunocytochemistry

Primary antibodies were: rabbit GnRH (SW-1, 1:3000) (Figure 1A; see also figure 5 below) (21), mouse monoclonal anti-GnRH (F1D3C5, 1:4000) (see figure 4 below; gift from Dr A. Karande, Department of Biochemistry, Indian Institute of Science, Bangalore, India) (22), chicken anti-green fluorescent protein (GFP) (1:1000; Abcam, see figure 4 below), rabbit anti-GPER (1:1000; Abcam, see figure 4 bellow), and mouse monoclonal anti-ERRγ (1:150; R&D Systems, see figure 5 below). Fresh frozen adult GnRH:GFP female sections (20 μm) (23) or explants were fixed in 4% formaldehyde (30 min at room temperature). After washes in PBS, tissues were incubated in a blocking solution (10% normal horse serum + 0.3% Triton X-100) for 1 hour, washed several time in PBS, and incubated in primary antibody (SW-1, GPER, ERRγ; 1–2 nights in primary 4°C). The next day, tissues were washed in PBS, incubated in donkey antirabbit (for SW-1 and GPER) Alexa Fluor 568 (1:1000, 1 h; Molecular Probes) or for ERRγ, the signal was amplified by incubation in donkey antimouse-biotinylated (1:500; Jackson ImmunoResearch) before Av555 (1:2000; Invitrogen). The tissue was fixed (4% formaldehyde; 10 min), and incubated in the second primary antibody (F1D3C5, GFP, and SW-1; 48 h at 4°C). The second antigen-antibody complex was visualized with donkey antimouse (for F1D3C5) Alexa Fluor 488 (Molecular Probes), donkey antichicken (for GFP) Alexa Fluor 488 (1:1000, 1 h; Molecular Probes), or donkey antirabbit (for SW-1) Alexa Fluor 488 (1:1000; Molecular Probes), depending on the species of the primary used (see Table 1).

Table 1.

Antibody Table

Peptide/Protein Target	Name of Antibody	Manufacturer, Catalog Number, and/or Name of Individual Providing the Antibody	Species Raised in; Monoclonal or Polyclonal	Dilution Used
GnRH	SW-1	Susan Wray	Rabbit; polyclonal	1:3000
GnRH	F1D3C5	A. Karande	Mouse; monoclonal	1:4000
GPR30	Anti-GPR30	Abcam, ab39742	Rabbit; polyclonal	1:1000
ERRγ	Anti-ERRγ	R&D Systems, PP-H6812-00	Mouse; monoclonal	1:150
GFP	Anti-GFP	Abcam, ab92456	Chicken; polyclonal	1:1000

Drugs

BPA and 17β-estradiol (E₂) were purchased from Sigma. (−)-bicuculline chloride (A-type γ-aminobutyric acid [GABA] receptor antagonist), D-(−)-2-amino-5-phosphonopentanoic acid (AP5) (N-methyl-D-aspartate receptor antagonist), 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) (α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid/kainate receptor antagonist), ICI 182780 (ICI) (ERα/β antagonist), and GSK 4716 (GSK) (ERRγ agonist) were purchased from Tocris. G15 (GPER antagonist) was purchased from Cayman Chemicals.

Statistical analysis

Paired Student's t tests were used to compare frequencies of calcium oscillations between 2 recording periods among a pool of cells. Unpaired t tests were used to compare the magnitudes of inhibition between 2 groups of cells. ANOVA tests, followed by Dunnett's multiple comparison test, were used to compare the magnitudes of inhibition between a control group and multiple groups of cells. P < .05 was chosen for significance. In Results and figures, frequencies of calcium oscillations are expressed as mean ± SEM, and n and N represent the number of cells and explants recorded, respectively.

Results

BPA reduces frequency of spontaneous calcium oscillations in GnRH neurons

According to the Environmental Protection Agency, 50 μg/kg·d of BPA is the current Lowest Observed Adverse Effect Level (LOAEL). For in vitro studies, circulating levels of BPA at the LOAEL are estimated around 21.9μM (24, 25). The effects of acute exposure to BPA (0.5μM and 50μM; 5–10 min) on GnRH neuronal activity were assessed using calcium imaging to monitor changes in the frequency of intracellular calcium oscillations, known to be congruent with bursts of action potentials (17). Both of these low doses significantly reduced the frequency of calcium oscillations in GnRH neurons (P < .05, paired Student's t test; Table 2, lines a and b) (Figure 2A). The frequency of calcium oscillations in GnRH neurons did not change using the same paradigm with SFM (P > .05, paired Student's t test) (Table 2, line c). Because the dose of 50μM BPA had a more robust effect than the 0.5μM dose (∼44 times lower than the estimated circulating levels of BPA at the LOAEL), it was chosen as the concentration to use for further investigations.

Table 2.

Frequencies of Calcium Oscillations in GnRH Neurons

#	Paradigms	Period 1	Period 2	Period 3	Period 4	n	N
a	SFM/BPA	1.62 ± 0.05	1.27 ± 0.04a			305	10
b	SFM/BPA (0.5μM)	1.85 ± 0. 10	1.58 ± 0.12a			71	2
c	SFM/SFM	1.66 ± 0.09	1.64 ± 0.09			54	2
d	SFM/E2	1.99 ± 0.09	1.42 ± 0.09a			77	3
e	SFM/E2/+BPA	1.50 ± 0.07	1.29 ± 0.08a	0.97 ± 0.08a		67	4
f	SFM/E2/E2/+BPA	1.35 ± 0.16	1.22 ± 0.16a	1.21 ± 0.21	0.81 ± 0.14a	15	1
g	SFM/AAB/+BPA	1.74 ± 0.06	1.09 ± 0.06a	0.91 ± 0.05a		172	4
h	SFM/AAB/AAB	1.85 ± 0.07	1.09 ± 0.08a	1.17 ± 0.09		105	4
i	SFM/AAB/+E2	1.81 ± 0.05	1.27 ± 0.06a	1.23 ± 0.06		163	3
j	SFM/ICI/+BPA	2.00 ± 0.07	2.06 ± 0.07	1.78 ± 0.08a		128	3
k	SFM/G15/+BPA	1.76 ± 0.06	1.63 ± 0.05a	1.33 ± 0.05a		198	6
l	SFM/ICI + G15/ + E2	1.75 ± 0.08	1.59 ± 0.08a	1.54 ± 0.08		101	4
m	SFM/ICI + G15/ + BPA	1.70 ± 0.07	1.44 ± 0.06a	1.31 ± 0.06a		121	7
n	SFM/AAB/AAB + ICI/ + BPA	1.86 ± 0.10	1.16 ± 0.10a	1.28 ± 0.09	0.99 ± 0.09a	55	3
o	SFM/AAB/AAB + G15/ + BPA	1.87 ± 0.12	1.34 ± 0.11a	1.44 ± 0.11	1.16 ± 0.09a	47	4
p	SFM/AAB/AAB + ICI + G15/ + BPA	2.02 ± 0 .10	1.32 ± 0.10a	1.38 ± 0.08	1.08 ± 0.07a	67	3
q	SFM/GSK	1.97 ± 0.08	1.64 ± 0.09a			94	3
r	SFM/AAB + ICI + G15/ + GSK/BPA	2.28 ± 0.07	1.61 ± 0.08a	0.74 ± 0.07a	0.32 ± 0.04a	104	3

Unless stated, BPA (50μM), E₂ (10nM), AAB (BIC [20μM], CNQX [10μM], AP5 [10μM]), ICI (1μM), G15 (100nM), and GSK (100μM). Periods 1–4 are expressed in peaks per minute. Data are expressed in mean ± SEM; n cells; N explants.

^aSignificant difference compared with the previous period (Student's t test; P < .05).

Figure 2.

BPA inhibits GnRH neuronal activity. A, Representative recording showing spontaneous baseline oscillations in intracellular calcium levels in 2 GnRH neurons during a 10-minute superfusion in control media (SFM) and subsequent decrease in activity during a 10-minute BPA (50μM) exposure (y-scale = Arbitrary units, AU). The bar graph indicates the results for all cells (*, P < .05). B, E₂ (10nM) also decreased spontaneous baseline oscillations in intracellular calcium levels during a 10-minute application. However, subsequent application of BPA (50μM) further decreased GnRH neuronal activity (y-scale = Arbitrary units, AU, summarized in bar graph; *, P < .05).

BPA uses a different pathway than estradiol to inhibit GnRH neurons

To begin to unravel the signaling used by BPA to alter GnRH neuronal activity, the effect of BPA was compared with the effect of estrogen (E₂, 10nM) (26). Acute exposure (10 min) of GnRH neurons to E₂ induced a decrease in the frequency of calcium oscillations (P < .05, paired Student's t test) (Table 2, line d). The magnitude of the inhibition evoked by E₂ was similar to the magnitude of inhibition evoked by BPA (in peaks/min; E₂: −0.38 ± 0.04 [n = 159; N = 4]; BPA: −0.35 ± 0.04 [n = 305; N = 10]; P > .05, unpaired Student's t test). However, using a cut-off of −0.9 peaks/min (∼2 SD), approximately 16% of GnRH neurons were inhibited by BPA, whereas approximately 27% were inhibited by E₂, suggesting 2 different pathways. Thus, E₂ (10 min) followed by E₂+BPA (10 min) was examined. Although GnRH neurons were inhibited by E₂ (P < .05, paired Student's t test) (Table 2, line e), the subsequent application of E₂+BPA depressed the frequency of calcium oscillations further (P < .05, paired Student's t test; Table 2, line e) (Figure 2B).

To ensure the inhibition was due to BPA rather than to a longer exposure to E₂ and the initiation of genomic action from the first 10 minutes in E₂, the experiment was repeated with E₂ for 10+10 minutes followed by E₂+BPA (10 min). Although the first 10-minute application of E₂ consistently inhibited the frequency of calcium oscillations (P < .05, paired Student's t test) (Table 2, line f), the second 10-minute application of E₂ had no further effect on the frequency of calcium oscillations (P > .05, paired Student's t test) (Table 2, line f). In contrast, the application of E₂+BPA, after 2 10-minute E₂ periods, still showed significant inhibition of GnRH neuronal activity (P < .05, paired Student's t test) (Table 2, line f). In addition, the magnitude of the inhibition was significantly greater with E₂+BPA than after E₂ alone, demonstrating additive effects of E₂ and BPA (in peaks/min; E₂+BPA: −0.54 ± 0.08 [n = 82; N = 5]; E₂: −0.38 ± 0.04 [n = 159; N = 4]; P < .05, unpaired Student's t test). These data indicate that BPA targets a different pathway than E₂ to inhibit GnRH neurons.

BPA inhibits GnRH neurons through a direct mechanism

Two different pathways are required to explain the additive effects of BPA and E₂ on the neuronal activity of GnRH neurons. Two hypotheses are possible: the targets are 2 different receptors 1) within the same cell, or 2) on 2 different cell types. Previous work has shown that an intricate cellular network develops in our model system and that GnRH neurons receive multiple inputs (16, 27, 29). Therefore, the effect of BPA was assessed in the presence of an amino acid blocker (AAB) cocktail known to isolate GnRH neurons from GABAergic and glutamatergic inputs (AAB, bicuculline [20μM] + AP5 [10μM] + CNQX [10μM]). The removal of inputs with 5-minute AAB treatment reduced the frequency of calcium oscillations (P < .05, paired Student's t test) (Table 2, line g) but did not prevent subsequent inhibition by 5-minute BPA application (P < .05, paired Student's t test; Table 2, line g) (Figure 3A). In contrast, when E₂ was applied in presence of AAB, no further effects on the frequency of oscillations was seen (P > .05, paired Student's t test; Table 2, line i) (Figure 3B). To check that the observed BPA inhibition of GnRH neurons was not due to the longer exposure to AAB, the paradigm was repeated with only AAB for 5+5 minutes. During the first 5-minute, the full AAB inhibition was reached (P < .05, paired Student's t test) (Table 2, line h). The frequency of calcium oscillations did not change during the second 5-minute application (P > .05, paired Student's t test) (Table 2, line h). Taken together, these data indicate that E₂ has an indirect effect on GnRH neurons and support a direct effect of BPA on GnRH neurons.

Figure 3.

BPA inhibits GnRH neuronal activity directly. A, Representative calcium imaging recording showing spontaneous baseline oscillations in intracellular calcium levels of 2 GnRH neurons during 5 minutes of SFM, 5 minutes of AAB (bicuculline [20μM] + AP5 [10μM] + CNQX [10μM]) and 5-minute coapplication of AAB and BPA (50μM) (y-scale = Arbitrary units, AU). The bar graph shows that application of AAB decreases GnRH neuronal activity, whereas coapplication with BPA further inhibits internal calcium oscillations in cells, supporting a direct effect of BPA (*, P < .05). B, In contrast, the inhibitory effect of E₂ was abrogated when applied in presence of AAB, indicating that E₂ has an indirect effect on GnRH neurons (NS; P > .05).

Multiple estrogen targets are detected in explants

BPA can bind to ERα and ERβ, the G protein-coupled receptor 30 (GPER) and ERRγ. Transcripts for ERβ but not ERα have already been reported in GnRH neurons maintained in explants (40% and 60% of 7 and 14 div, respectively) (10). The presence of transcripts for GPER was assessed in GnRH neurons by single-cell PCR. Data showed that GnRH neurons expressed GPER (3/5) (Figure 4A) at 7 div. GPER protein expression in GnRH neurons was confirmed using double-label immunofluorescence in female adult mice and explants (Figure 4, B and C).

Figure 4.

GPER is expressed in GnRH neurons but not a target of BPA. A, PCR on cDNAs from single GnRH cells showed bands of appropriate size for GPER (left, GnRH transcript right). B₁–B₃, GnRH neurons in female adult GnRH:GFP mice express GPER protein on their cell membrane. C₁–C₃, Primary GnRH neurons in explants also express GPER protein in their cell membrane. D, Blocking ERβ (ICI [1μM]) and GPER (G15 [100nM]) in presence of AAB did not prevent the inhibitory effect of BPA (50μM) (left; representative recording showing oscillations in intracellular calcium levels in 2 GnRH neurons [y-scale = Arbitrary units, AU], right; bar graph indicates the results for all cells [*, P < .05]). Scale bars, 5 μm (B and C).

BPA acts on ER and GPER but not in GnRH neurons

The activation of ERβ and GPER by BPA was tested pharmacologically. BPA was applied after either ICI (1μM) (26) or G15 (100nM), 2 specific antagonists for ERα/β and GPER, respectively. Neither ICI nor G15 alone prevented the inhibition evoked by BPA (for either condition; P < .05, paired Student's t test) (Table 2, lines j and 1k). In addition, either drug alone did not decrease the degree of inhibition produced by BPA (in peaks/min; BPA: −0.35 ± 0.04; ICI+BPA: −0.28 ± 0.04; G15+BPA: −0.30 ± 0.04; P > .05, ANOVA and Dunnett's multiple comparison test). Although ICI+G15 were able to prevent the inhibition evoked by E₂ (P > .05, paired Student's t test) (Table 2, line l), this combination did not block BPA inhibition (P < .05, paired Student's t test) (Table 2, line m). However, the combination ICI+G15 did decrease the magnitude of BPA inhibition (−0.12 ± 0.05 peaks/min; P < .05). To determine whether this response was due to direct or indirect action, the paradigm was repeated in presence of AAB.

During the coapplication of AAB+ICI or AAB+G15, BPA still decreased the frequency of calcium oscillations in GnRH neurons (P < .05, paired Student's t test) (Table 2, lines n and o). Even the cocktail AAB+ICI+G15 did not prevent BPA inhibition of GnRH neuronal activity (P < .05, paired Student's t test; Table 2, line p) (Figure 4D). Neither ICI, G15, nor ICI+G15 decreased the degree of inhibition of AAB+BPA (in peaks/min; AAB+BPA: −0.18 ± 0.04; AAB+ICI+BPA: −0.29 ± 0.06; AAB+G15+BPA: −0.28 ± 0.07; AAB+ICI+G15+BPA: −0.30 ± 0.07; P > .05, ANOVA and Dunnett's multiple comparison test). These data indicated that BPA targeted different receptors than ERs and GPER in GnRH neurons.

ERRγ is another potential target of BPA

BPA has been shown to strongly bind to ERRγ (30). The presence of transcripts for ERRγ was assessed in GnRH neurons by single-cell PCR. Data showed that GnRH neurons expressed ERRγ at 7 div (3/7) (Figure 5A). ERRγ protein expression in GnRH neurons was confirmed using double-label immunofluorescence in explants (Figure 5B). To determine whether binding of BPA to ERRγ could have an impact on GnRH neuronal activity, GnRH neurons were exposed to GSK (100μM), an ERRγ agonist. The frequency of calcium oscillations in GnRH neurons decreased (P < .05, paired Student's t test; Table 2, line q) (Figure 5C). The magnitude of inhibition was comparable between BPA and GSK (in peaks/min; BPA: −0.35 ± 0.04; GSK: −0.33 ± 0.07; P > .05, unpaired Student's t test).

Figure 5.

ERRγ is present in GnRH neurons but not involved in the inhibitory BPA effect. A, PCR on single GnRH cells revealed expression of ERRγ in a subset of GnRH neurons. B₁–B₃, Primary GnRH cells were immunopositive for ERRγ. C, Exposure of GnRH neurons to GSK (100μM), an ERRγ agonist decreased the frequency of calcium oscillations in GnRH neurons comparable with BPA (left; representative recording showing oscillations in intracellular calcium levels in a GnRH neuron [y-scale = Arbitrary units, AU] and right; bar graph indicates the results for all cells [*, P < .05]). D, Even when blocking GABA/glutamatergic inputs as well as ERs and GPER, GSK persisted to inhibit GnRH neuronal activity. However, subsequent application of BPA decreased the frequency of calcium oscillations in GnRH neurons further (left; representative recording showing oscillations in intracellular calcium levels in 2 GnRH neurons [y-scale = Arbitrary units, AU] and right; bar graph indicates the results for all cells [*, P < .05]). Scale bars, 5 μm (B).

When repeating the same paradigm without GABA/glutamatergic inputs and blocking ERs and GPER, the frequency of calcium oscillations in GnRH neurons still decreased upon application of GSK (P < .05, paired Student's t test) (Table 2, line r). Notably, subsequent application of BPA decreased the frequency of calcium oscillations in GnRH neurons further (P < .05, paired Student's t test) (Table 2, line r) (Figure 5D). The magnitude of the inhibition with BPA was similar after AAB+ICI+G15 and AAB+ICI+G15+GSK (in peaks/min; AAB+ICI+G15+BPA: −0.30 ± 0.07; ICI+G15+GSK+BPA: −0.42 ± 0.04; P > .05, unpaired Student's t test). To ensure that ERRγ were saturated with GSK at 100μM, and that BPA was not having only additive effects, the paradigm was repeated with 150μM GSK. BPA still reduced the frequency of calcium oscillations. The magnitude of inhibition after AAB+ICI+G15+GSK 150μM did not increase compared with 100μM GSK (P > .05, unpaired Student's t test). These data suggest that ERRγ can directly and rapidly modulate GnRH neuronal activity and that BPA inhibits GnRH neuronal activity via another mechanism.

Discussion

BPA is one of the many pollutants found in today's environment in industrialized countries. BPA exposure in the great majority of people occurs as a byproduct of plastics, dental sealants, and food can linings (24, 31, 32). BPA belongs to the family of EDCs, ie, molecules that can highjack the endocrine system. Literature reports the ability of BPA to bind to all ERs: both ERα and ERβ (33, 34), the membrane receptor for ERα (mERα) (35), ERRγ (36), and GPER (37). Estrogen being a key regulator of the reproductive system, sub- or infertility has been the most documented effect of BPA. Although the literature mostly describes how timed exposure to BPA around birth might have a long lasting impact on the hypothalamic-pituitary-gonadal axis and hamper fertility in adults, cellular mechanisms remain unclear. We report here that acute exposure to BPA has a direct effect on GnRH neuronal activity. The BPA-mediated inhibition of GnRH neurons is complex due to the fact that BPA acts on multiple receptors expressed by GnRH neurons including ERs, GPER as well as ERRγ. However, under all conditions BPA inhibited GnRH neuronal activity.

The literature estimating the extent of the human exposure to BPA flourishes. Although recent data suggest that the daily human exposure is higher than previously assumed (38), the consensus around a tolerable daily intake value of 0.05 mg/kg·d still prevails based on adult renal clearance (39). However, controversy remains regarding the safety of this dose during embryonic development. BPA crosses the placental barrier (40, 41) and can reach up to 50 ng/g in the liver of human fetuses (42). Numerous studies reported long-term effects of BPA in adults after a prenatal or early postnatal exposure. With a dissociation constant of approximately 0.1nM for estrogen, and a relative binding affinity of ERα and ERβ approximately 10 000 times lower for BPA than for estrogen (33), it can be assumed that a micromolar range of BPA is necessary in embryos to exert its effects through classical ERs. In contrast, BPA can bind ERRγ, displacing 4-hydroxytamoxifen at an EC₅₀ value of approximately 10nM. In this study, both 0.5μM and 50μM BPA inhibited GnRH neurons, ie, doses lower than, and slightly above, 22μM, the estimated circulating levels following the exposure at the LOAEL. Due to its consistent and robust effect, 50μM BPA was used to investigate the signaling pathways involved in the BPA inhibition.

Many studies trying to explain the adverse effects of BPA on fertility in adults observed gross abnormalities induced by a precisely timed exposure to BPA during the perinatal period. BPA affects almost every aspect of the reproductive system: circulating hormones (43–46), GnRH mRNA (47–49), GnRH pulsatility (50), puberty onset (50, 51), sperm count and prostate gland in males (44), ovaries and reproductive tract in females (45, 52–55), estrous cycles (50, 54), and number of litters and sizes (56). To date, only a few studies have reported how these abnormalities may occur, implicating sexually dimorphic cell populations of the preoptic area, anteroventral periventricular, arcuate paraventricular nuclei (36, 46, 48, 49, 57, 59, 60, reviewed in Refs. 58, 61), and epigenetics (62, reviewed in Ref. 63). However, the direct effects of BPA on GnRH neurons were unknown.

Over the last decades, GnRH neurons maintained in the explant model used here have been found to express a wide range of receptors (64) as well as patterns of spontaneous activity (65) similar to that seen by GnRH neurons in vivo. One of the main advantages of this model, is that large numbers of primary GnRH cells can be assayed and manipulated under defined conditions. Therefore this model system was used to assess acute effects of BPA on GnRH neurons and to decipher the mechanisms involved.

BPA rapidly inhibited the frequency of calcium oscillations in GnRH neurons after only 5–10 minutes. Literature reports BPA effects within the same or shorter time scale on intracellular calcium in oocytes (66) and a GH3 pituitary tumor cell subline (67), on levels of phosphorylation in cultured cerebellar neurons (68) and spine morphology in cultured hippocampal neurons (69), ruling out genomic actions. Estradiol (E₂) also inhibited the frequency of calcium oscillations in GnRH neurons after only 5–10 minutes. Notably, BPA-induced inhibition in GnRH neurons was measurable after E₂-induced inhibition, suggesting a BPA specific mechanism. In contrast, in both oocytes and GH3, BPA mimicked the effects of E₂. However, BPA stopped oscillations in oocytes whereas it initiated oscillations in GH3 (66, 67). In primary cultured cerebellar neurons BPA again acted like E₂, ie, activating a pathway that should normally be silent. However, when coapplied with E₂, BPA interfered with the endogenous ligand to silence the pathway that should have normally been activated (68). Taken together, these data indicate complex tissue-specific actions of BPA, rather than highlight an unique cellular pathway (25, 34, 70).

GnRH neurons maintained in explants establish an intricate cellular network and the GnRH neuronal activity relies, at least partially, upon cell-to-cell communication (29). A major component of this network is GABA/glutamatergic inputs. However, isolating GnRH neurons from these inputs did not prevent the BPA-induced inhibition. In fact, BPA was still effective after blocking ERs and GPER. These data provide the first evidence of a direct action of BPA on GnRH neurons. Notably, although BPA-induced inhibition persisted without inputs, the E₂-induced inhibition did not. This indicates that BPA was able to act on targets that E₂ would not activate normally. E₂ binds poorly to ERRγ compared with BPA (30). Both mRNA and protein for ERRγ were found in GnRH neurons. To date, only transcriptional effects have been described for ERRγ (30, 71). Application of GSK, an ERRγ agonist, produced a rapid (within minutes) decrease in the frequency of calcium oscillations in GnRH, suggesting, in fact, a nontranscriptional action. In vivo the binding of ERRγ to BPA is so potent it has been linked to the accumulation of BPA in the placenta, an organ that contains a very large quantity of ERRγ (72). Our results add ERRγ as a new functional pathway in GnRH neurons. However, the caveat that remains unanswered is the persistence of a BPA effect after GSK with or without AAB/ICI/G15, indicating another pathway for BPA to act on GnRH neurons using mechanism(s), independent of ERs, GPER and ERRγ. The literature reports other targets for BPA such as voltage-gated sodium channels (73), potassium channels (74, 75), and calcium channels (28), all present in GnRH neurons. The fact that some GnRH neurons showed a fast recovery from the BPA inhibition during the washout period is consistent with a direct action of BPA on voltage-gated channels. Together, the complexity by which BPA can exert its deleterious effects upon fertility expands with studies. Although exposure to BPA is usually chronic, this study underscores the importance of understanding how changes are triggered at a cellular level.