Bisphenol A (BPA), a high volume production chemical compound attracts growing attention as a health-relevant xenobiotic in humans. It can directly bind to hormone receptors, enzymes, and ion channels to become biologically active. In this study we show that BPA acts as a potent blocker of voltage-activated Ca2+ channels. We determined the mechanisms of block and the structural elements of BPA essential for its action. Macroscopic Ba2+/ Ca2+ currents through native L-, N-, P/Q-, T-type Ca2+ channels in rat endocrine GH3 cells, mouse dorsal root ganglion neurons or cardiac myocytes, and recombinant human R-type Ca2+ channels expressed in human embryonic kidney (HEK) 293 cells were rapidly and reversibly inhibited by BPA with similar potency (EC50 values: 26–35 μM). Pharmacological and biophysical analysis of R-type Ca2+ channels revealed that BPA interacts with the extracellular part of the channel protein. Its action does not require intracellular signaling pathways, is neither voltage- nor use-dependent, and does not affect channel gating. This indicates that BPA interacts with the channel in its resting state by directly binding to an external site outside the pore-forming region. Structure-effect analyses of various phenolic and bisphenolic compounds revealed that 1) a double-alkylated (R-C(CH3)2-R, R-C(CH3)(CH2CH3)-R), or double-trifluoromethylated sp3-hybridized carbon atom between the two aromatic rings and 2) the two aromatic moieties in angulated orientation are optimal for BPA’s effectiveness. Since BPA highly pollutes the environment and is incorporated into the human organism, our data may provide a basis for future studies relevant for human health and development.
Bisphenol A (BPA) is a chemical that is extensively used (>3.8 million tons/yr worldwide) to produce polycarbonate plastic and epoxy resins. Both synthetics are manufactured into highly diverse mass products such as optical media (CD and DVD), protective coatings inside metal food containers, baby bottles, thermal paper, composites, and sealants in dentistry and medical tubing (for review see Vandenberg et al., 2007; Dekant and Völkel, 2008). BPA is an environmental pollutant that is incorporated into living organisms (for review see von Goetz et al., 2010). Common routes of BPA exposure in humans are oral intake, respiration, and dermal absorption (Braunrath, 2005; Biedermann et al., 2010; Loganathan and Kannan, 2011). Detectable levels of BPA were found in over 90% of people living in industrialized countries due most likely to chronic exposure (Calafat et al., 2005; Calafat et al., 2008; see also He et al., 2009). BPA’s adverse effects on human health and the ecosphere are being increasingly recognized (for review see Chapin et al., 2008).
Numerous studies have shown that BPA influences a wide range of physiologic functions (for review see Rubin, 2011). It is very well documented that BPA and various related compounds bind to hormone receptors and influence multiple endocrine pathways (Matsushima et al., 2008; Okada et al., 2008; Swedenborg et al., 2009; Riu et al., 2011; Soriano et al., 2012). Beyond this, recent studies provide evidence that BPA can directly interact with biologically active proteins such as enzymes and ion channels (Hiroi et al., 2006; Asano et al., 2010; Hashimoto et al., 2012; O'Reilly et al., 2012; Pandey and Deshpande, 2012). Protein disulfide isomerase (PDI) has been isolated as a binding protein of BPA in the rat brain. BPA binds to different domains of protein disulfide isomerase with KD values in the range of 10−6 to 10−3 molar (Hashimoto et al., 2012). A rapid and reversible increase of large conductance Ca2+/voltage-sensitive K+ (Maxi-K) channel activity by BPA (10–100 μM) has been shown in human and canine coronary smooth muscle cells as well as in AD-293-cells expressing the recombinant form of the channel (Asano et al., 2010). Furthermore, inhibitory effects of BPA in the micromolar range have been reported for mouse neuronal (Wang et al., 2011) and human cardiac Na+ channels (O'Reilly et al., 2012). In the latter case it has been shown that BPA directly binds to the channel protein at the local anesthetic receptor site.
In the present study we have characterized the interaction of BPA and several of its related compounds with voltage-activated Ca2+ channels. This family of channels can be divided into subtypes that are expressed in different cells of the body (for review see Catterall et al., 2005). The different types of voltage-activated Ca2+ channels play key roles in various physiologic and pathophysiological processes such as excitation-contraction coupling, synaptic transmission, hormone release, gene expression, and cell death and differentiation (for review see Catterall, 2011).
We sought to elucidate the mechanisms underlying BPA’s interaction with voltage-activated Ca2+ channels expressed in GH3 pituitary tumor cells, mouse DRG neurons, mouse cardiac myocytes, and with voltage-activated recombinant human R-type Ca2+ channels stably expressed in human embryonic kidney (HEK) 293 cells. The blocking mechanism of BPA was analyzed in detail in recombinant human R-type Ca2+ channels. We aimed at characterizing state-dependence of the block and the site where BPA binds to the channel protein. Furthermore, we intended to determine which structural features of the BPA molecule are basic for its ability to block Ca2+ channels. For that purpose we compared the degree of Ca2+ channel block induced by BPA-related compounds to that of BPA.
Material and Methods
Cells and Cell Culture.
Endocrine rat pituitary tumor cells (GH3 line; DSMZ, Braunschweig, Germany) were grown on 100 mm culture dishes (Becton Labware, Franklin Lakes, NJ) in Ham F10 medium supplemented with 15% horse serum, 2.5% fetal calf serum, 1% penicillin-streptomycin, and 2 mM glutamine. Two to seven days before electrophysiological recordings, cells were harvested with trypsin or accutase (PAA, Pasching, Austria) and plated on poly-l-lysine–coated (Sigma-Aldrich, Taufkirchen, Germany) 6-mm glass coverslips in P.neural medium (P.Glia, Rheinbach, Germany) and maintained in humidified atmosphere (37°C). Cells were used for electrophysiological recordings within the following 48 hours.
Dorsal root ganglion (DRG) neurons were excised from decapitated 4- to 14-day-old male or female (ratio ∼1:1) mice (C57BL/6) and incubated for 25 ± 5 minute in extracellular solution consisting of 124 mM NaCl, 2.5 mM KCl, 26 mM NaHCO3, 1 mM NaH2PO4, 1.3 mM MgSO4, 2 mM CaCl2, 10 mM d-glucose, pH 7.35, gassed with carbogen (95% O2, 5% CO2) containing 0.04% collagenase and 0.025% trypsin type I (Roche, Mannheim, Germany) at 37°C. Digestion was stopped by resuspending the tissue in culture medium: Dulbecco’s modified Eagle’s medium (DMEM)/F12 (1:1) (Invitrogen, Mannheim, Germany) supplemented with 10% fetal bovine serum, penicillin (100 IU/ml) and streptomycin (100 IU/ml), and 2 mM l-glutamine (all Invitrogen). Tissue was triturated mechanically by vigorously shaking and centrifuged at 160g for 5 minutes after filtration. Pellet was resuspended in culture medium and cells were plated on poly-l-lysine–coated glass coverslips and kept in humidified atmosphere (37°C, 95% air, 5% CO2). Cells of medium size were used for electrophysiological recordings within the following 12 hours. All animal experiments were conducted in accordance with the guidelines of the Animal Care Committee of the University of Bonn.
Cardiac myocytes were isolated from male or female (ratio ∼1:1) mice (C57BL/6) as described previously (Linz and Meyer, 2000). Briefly, hearts were removed and cardiopleged in cold Tyrode solution containing 135 mM NaCl, 4 mM KCl, 1 mM MgCl2(H2O)6, 2 mM HEPES, 2.6 mM EGTA, pH 7.4. The heart was rapidly attached to a Langendorff apparatus and perfused with the Tyrode solution described above (37°C) followed by a high-potassium solution (4 mM NaCl, 10 mM KCl, 1 mM MgCl2, 0.2 mM CaCl2, 130 mM K-glutamate). For digestion 200 IU/ml trypsin type I (Sigma) and 0.4 mg/ml collagenase type II (Sigma-Aldrich) were added. The ventricles were cut down, chunked and allowed to settle in oxygenated Tyrode solution containing 135 mM NaCl, 4 mM KCl, 1 mM MgCl2, 2 mM HEPES, 1.8 mM CaCl2, 1 mM BSA, and 16.6 mg/l trypsin inhibitor. The solution was filtered through a 125-μm nylon mesh, centrifuged briefly, and the pellet was resuspended in fresh Tyrode solution. Cells were plated on poly-l-lysine–coated glass coverslips and recordings were performed 1–8 hours after plating.
HEK293 cells stably expressing human α1E and β3 Ca2+ channel-subunits were kindly provided by T. Schneider, University of Cologne (Nakashima et al., 1998). Cells were maintained in DMEM supplemented with penicillin (100 IU/ml), streptomycin (100 μg/ml) (Invitrogen, Darmstadt, Germany), and 10% fetal bovine serum. Geneticin (0.5 mg/ml) and hygromycin (0.2 mg/ml) were used for selection of α1E- and β3-subunit expression, respectively. Cells were plated on poly-l-lysine–coated glass coverslips and used within 24–48 hours after plating for recordings.
Ca2+ and/or Ba2+ currents were recorded in the whole-cell configuration of the patch-clamp technique using an EPC9 patch-clamp amplifier and the PULSE software (HEKA Electronic, Lambrecht, Germany). Data were filtered and digitized at 3 and 10 kHz, respectively, and stored on hard disk. For recordings of tail currents the sampling rate was 100 kHz. In ventricular cardiac myocytes Cav 1.2 L-type Ca2+ channels constitute the major pathway for high voltage–activated (HVA) Ca2+ currents (Bers and Perez-Reyes, 1999). Ba2+ (2 mM) currents through these channels were elicited by voltage ramps (200-millisecond) from –50 mV to +70 mV. In mouse DRG neurons Ca2+ currents through T-type Ca2+ channels were evoked by depolarizations (100-millisecond) from –70 mV to –20 mV using Ca2+ (10 mM) as charge carrier. Currents through N-type Ca2+ channels were elicited by voltage ramps (100-millisecond) from –70 mV to +50 mV using Ba2+ (1 mM) as charge carrier. Currents were recorded in the presence of nifedipine (10 μM) and ω-agatoxin-TK (0.2 μM) to eliminate contribution from L- and P/Q-type Ca2+ channels, respectively. The internal solution of the recording electrodes contained: 160 mM cesium-aspartate, 10 mM EGTA, 2 mM MgATP, 20 mM phosphocreatine, 0.2 mM Na2GTP, 20 mM HEPES (pH 7.3, 290–300 mOsmol/l). Filled electrodes had resistances between 1.5 and 4 MΩ. The extracellular solution contained: 160 mM TEA-Br, 3 mM KCl, BaCl2, or CaCl2, 1 mM NaHCO3, 1 mM MgCl2, 10 mM HEPES, and 4 mM glucose (pH 7.4, 300–310 mOsmol/l). The final concentrations of Ba2+ and Ca2+ are indicated throughout the text. All recordings were performed at room temperature (20–23°C).
The recording chamber was continuously perfused at a rate of 1 ml/min. The bath volume was exchanged every 20 seconds. Drugs were applied in external solution using a fast-pressure–application system (DAD-VM Superfusion System; ALA Scientific Instruments, Farmingdale, NY). The tip of the application pipette (diameter 100 μm) was positioned within 100 μm off the cells and solution exchange was obtained within ∼20 milliseconds.
Bisphenol A and the related compounds were purchased from Alfa Aesar (Karlsruhe, Germany) or Sigma-Aldrich (Taufkirchen, Germany). Stock solutions of the drugs were prepared in ethanol at a concentration of 200 mM and stored at room temperature. All toxins were obtained from Tocris Biosciences (Wiesbaden, Germany) and stock solutions were prepared in distilled water. Application solutions were freshly prepared from stock solution. The highest final concentration of the solvent ethanol was 0.15% which did not affect voltage-activated Ca2+ channels. All other chemicals were obtained from Sigma-Aldrich.
Data Analysis and Statistics.
The analysis of the whole-cell recordings was carried out offline using PulseFit (HEKA, Germany) or IGOR software (Wavemetrics, Lake Oswego, OR). Data fitting and statistical analysis were performed using PRISM 5.0 (GraphPad Software Inc., San Diego, CA). EC50 values and Hill slopes were determined by fitting data points to a logistic function. Charge-voltage relationships were fitted by an equation in the form y = Gmax * (Vtest – Vrev) / (1 + exp((V1/2 – Vtest) / k)), with Gmax = maximum conductance, V1/2 = half-maximal activation, Vtest = test potential, Vrev = reversal potential, and k = slope factor. Data for m∞ and h∞ were fitted to a Boltzmann function Y = Min + (Max – Min) / (1 + exp((V0.5 – Vtest) / k)). Time course for recovery from inactivation was fitted by a double exponential function. The data in the manuscript are presented as mean ± standard error unless stated otherwise. Statistical analysis was performed using Students t test or one-way analysis of variance and Tukey’s post-test. Differences with P value <0.05 were regarded as significant and levels of significance are indicated by asterisks (*** P<0.001, ** P<0.01, * P<0.05).
Inhibition of High Voltage–Activated Ca2+ Channels by BPA in GH3 Cells.
GH3 cells used in our experiments express essentially HVA Ca2+ channels. BPA was applied extracellularly to single cells at micromolar concentrations using a fast application system (see Materials and Methods). Fig. 1A shows the effect of BPA on Ba2+ currents (5 mM) elicited by depolarizing voltage steps (50-millisecond) to 0 mV at 0.1 Hz for three different concentrations. Each concentration (starting with 10 μM) was applied for 40 seconds during which maximal inhibition was achieved. The current inhibition was almost fully reversible and currents recovered completely within about 1 minute upon washout of BPA. Similar results were obtained with Ca2+ (10 mM) as charge carrier. Note that internal perfusion with BPA even at high concentrations (100 μM) was ineffective (unpublished data). Fig. 1B illustrates the time course of Ca2+ channel block by 100 μM BPA and of its washout when brief depolarizing steps (15-millisecond) to 0 mV were applied at 1 Hz. The finding that the amount of block by BPA did not vary with different stimulation frequencies (0.1 and 1 Hz) implies that the block is not use-dependent (for a more detailed analysis see R-type Ca2+ channels below).
In further experiments the effect of BPA (70 μM) was studied at different depolarizing voltages. The charge transferred by an individual Ba2+ current was determined from the area under the current curves. Fig. 1C shows the charge-voltage relationship obtained in the presence and absence of BPA. While charge transfer was significantly reduced there was no significant shift in the charge-voltage curve in the presence of BPA. The mean value of the block at different membrane potentials was calculated from corresponding data points and plotted against voltage (fractional block, see inset). The degree of block did not vary significantly within the voltage range tested (P = 0.92; n = 3).
To analyze the blocking action of BPA in more detail total currents were pharmacologically dissected using nifedepine and/or ω-conotoxins. As illustrated in Fig. 1D1 the application of ω-conotoxin GVIA (1.5 μM) was ineffective, indicating that N-type channels did not contribute to the total current. Application of nifedepine (10 μM), which blocks L-type Ca2+ channels, resulted in a 52 ± 3.9% (n = 6) reduction of the total current. Block of P/Q-type Ca2+ channels by ω-conotoxin MVIIC (1.5 μM) reduced the total current by 30 ± 0.9% (n = 7). A small current fraction (probably R-type) of 17 ± 2.0% (n = 3) of the total current was resistant to both nifedepine and ω-conotoxin MVIIC. Figure 1D2 illustrates the contribution of the different HVA Ca2+ channel types to the total current.
In the presence of BPA (70 μM) the total Ba2+ current was inhibited by 74.2 ± 2.6% (n = 7; Fig. 1E1). Figure 1E2 shows that inhibition of the current components remaining after nifedepine and/or ω-conotoxin MVIIC application did not significantly differ from inhibition of the total current by BPA (P > 0.7 between the four columns). These findings indicate that BPA does not discriminate between the different HVA Ca2+ channel types and blocks L-, P/Q-, and probably R-type Ca2+ channels to the same extent.
Figure 1F illustrates the concentration-effect relationship for total Ba2+ currents and currents in the presence of nifedepine. The potency of BPA to inhibit currents is similar for both curves (see Table 1). This supports our findings that BPA does not distinguish between the different HVA channel types.
Inhibition of High and Low Voltage–Activated Ca2+ Channels by BPA in DRG Neurons.
In DRG neurons N-type Ca2+ channels are predominant (Fig. 2A2), as indicated by the finding that 78 ± 0.6% (n = 3) of the total Ba2+ currents were blocked by the application of ω-conotoxin GVIA (500 nM) (Fig. 2A1). Application of nifedipine (10 µM) reduced the current by 15 ± 1.7%. Application of agatoxin-TK (200 nM), a P/Q-type Ca2+-channel blocker, was ineffective. A small current fraction (∼6%) was resistant to the blockers applied and was presumably through R-type Ca2+ channels (Fig. 2A2). Dose-dependent inhibition of N- (+R-) type channels by BPA was studied in the presence of nifedipine and agatoxin TK (Fig. 2B). The EC50 of 35 ± 1.3 μM was comparable to the EC50 values obtained for other HVA Ca2+-channel types in GH3 cells. In DRG neurons the effect of BPA was also investigated on low voltage–activated (T-type) Ca2+ channels. BPA also reduced this type of Ca2+ channel with a potency comparable to that obtained for HVA Ca2+ channel types (Fig. 2C; Table 1).
Inhibition of L-Type Ca2+ Channels by BPA in Cardiac Myocytes.
In ventricular cardiac myocytes HVA Ca2+ currents are through Cav 1.2 L-type Ca2+ channels (Bers and Perez-Reyes, 1999). BPA inhibited these Ca2+ channels in a concentration-dependent manner with an EC50 of 35 ± 1.3 μM (Fig. 2D). As in endocrine and neuronal cells, the time course of block and of its washout was fast and occurred within seconds. Together these findings suggest that BPA inhibits the different classes of Ca2+ channels found in different tissues with similar potency (for summary see Table 1).
Biophysical and Pharmacological Characterization of BPA’s Action on Human R-Type Ca2+ Channels Expressed in HEK293 Cells.
A detailed analysis of the blocking mechanisms of BPA was performed on human R-type Ca2+ channels in HEK293 cells stably expressing α1E and β3 Ca2+-channel-subunits. Figure 3A shows a family of Ba2+ (15 mM) currents elicited by depolarizing steps (100-millisecond) to voltages between –10 and +40 mV. BPA (35 μM; middle panel) reversibly inhibited currents and nearly full recovery from current inhibition was observed within 2 minutes upon washout (Fig. 3A, right panel). Small reductions in current amplitudes after washout were most likely due to current rundown during internal perfusion. In some cells full recovery could be obtained. To test for use-dependence of the block, depolarizing steps (15-millisecond) to 0 mV were applied at 0.1, 1, and 10 Hz. The amount of block obtained by BPA (35 µM) at each frequency did not vary significantly (P = 0.88; unpublished data; see also Fig. 1, A and B). This was also the case when the frequency was changed to 10 Hz after block had reached its equilibrium at 0.1 or 1 Hz.
Figure 3B illustrates the charge transferred by Ba2+ ions through R-type Ca2+ channels at different depolarizing voltages for control and in the presence of 35 μM BPA (C-V relationship). Inhibition of Ba2+ currents through R-type Ca2+ channels by BPA was concentration-dependent and occurred with an EC50 of 26 ± 1 μM (Fig. 3C).
To test whether binding of BPA is state-dependent, holding potential was shifted from –70 mV to –100 mV where almost all of the channels are in the resting state (Fig. 4A). As we did not observe any significant change in potency (Fig. 3C; EC50 32 ± 1 μM; P = 0.66), this finding strongly suggests that BPA binds to the channel in its resting state.
As illustrated in Fig. 3A current activation and inactivation kinetics were hardly affected by BPA. In further experiments we determined steady state activation and inactivation under control conditions and in the presence of 35 μM BPA (Fig. 4A) and found that neither the steady state activation nor the inactivation curves were significantly affected by BPA. Taken together these findings provide additional evidence that BPA binds to and stabilizes the resting state of the channel. Further support comes from the observation that neither recovery from inactivation (Fig. 4B; P = 0.824) nor deactivation kinetics (Fig. 4C; P = 0.73) were significantly changed by BPA.
To exert its action BPA has to be present for a certain time period (preapplication time) before current activation. When preapplied for 100 milliseconds BPA (70 μM) started to become effective and approached its maximal inhibitory effect after 10 seconds of preapplication (Fig. 5A). When block was brought to equilibrium by prolonged preapplication (3 minutes) of BPA (35 μM) before a long-lasting depolarizing step (1-second) was applied, we observed only a reduction in current amplitude and no change in the rate of current decay (Fig. 5, C and D). These findings demonstrate that BPA is not blocking open channels. Further support for these results came from the observation that BPA did not affect either current amplitude or current kinetics when applied during depolarizing pulses with channels in the open state (Fig. 5B). Even concentrations as high as 300 μM did not produce any current inhibition.
To test for a possible involvement of intracellular signaling pathways in BPA’s action, modulators of G proteins and protein kinases A and C were applied (Fig. 6). Irreversible activation or inhibition of G proteins by adding guanosine 5′-O-(3-thiotriphosphate) (20 μM) or GDPβS (1 mM) to the internal perfusion solution had no significant effect on the blocking action of BPA. Neither was BPA’s action affected by the inhibition of protein kinases A and C by H-89 (10 μM) and GÖ-6983 (10 μM), respectively. These modulators were applied extracellularly for up to 10 minutes. It is noteworthy that GÖ-6983 itself inhibited Ca2+ channels to about 77.6 ± 1.8% (n = 3) after 4–5 minute. Similar effects of GÖ-6983 have been previously described for L-type Ca2+ channels (Welling et al., 2005). BPA was always applied after the inhibitory effect of GÖ-6983 had reached its maximum.
Several reports have suggested or demonstrated that hormones are able to modulate Ca2+ channels by direct interaction (for review see Boonyaratanakornkit and Edwards, 2007). In particular 17β-estradiol has been shown to rapidly and reversibly reduce cardiovascular L-type Ca2+ channels at micromolar concentrations (Jiang et al., 1992; Nakajima et al., 1995; Meyer et al., 1998). Since we found that BPA inhibited L-type Ca2+ channels in cardiomyocytes (Fig. 2D) we wondered if it shares a common binding site on Ca2+ channels with E2. Indeed E2 (100 μM) also reduced R-type Ca2+ channels rapidly and reversibly by 40.7 ± 2.3% (Fig. 7A). At this concentration the E2-mediated inhibition was near saturation (Fig. 7B) but was significantly smaller than that observed with BPA (more than 80%) at the same concentration (Fig. 7, A and B). Similar results were obtained with total currents in GH3 cells (unpublished data).
In case of a common binding site one would expect that under these conditions the combination of the two compounds is less effective than BPA alone. This, however, was not observed (Fig. 7, A and B). On the contrary, the combination of the two compounds was slightly but significantly more effective (BPA, 81.5 ± 1.3%, n = 14; BPA + E2, 85.6. ± 1.3%, n = 12; P < 0.05).
Effect of BPA and Related Compounds on Ca2+ Channels in GH3 Cells.
To study the structure-activity relationship of bisphenol A and related substances, compounds with 1) different bridging structures between the two phenolic rings (e.g., BPAF, BPF, 4,4′-HBP), 2) different aromatic substitution patterns (4′-CP, 2,2′-DPP, TMBPA), and 3) different sterically demanding structure moieties (4′-TBP, 4′-TAP, BPM, BPP) were investigated in GH3 cells (Fig. 8). Table 2 summarizes the effect of the various phenolic and bisphenolic compounds on HVA Ba2+ currents in GH3 cells in relation to BPA.
TBBA, one of the most abundantly produced halogenated flame retardants, effectively inhibited Ba2+ currents at 100 μM concentration. As shown in Fig. 8A, TBBA was slightly but significantly less effective in blocking Ca2+ channels than BPA (Table 2). In contrast, BPS, another high production monomer, had no significant effect at this high concentration (Table 2).
Comparing the effects of BPA, 4′-CP, 2,2′-DPP, TBBA, and TMBPA, the aromatic substitution pattern strongly influences the ability to inhibit Ca2+ channels. The chemical structure of BPA is composed of two methyl groups and two phenol moieties on the central sp3-hybridized carbon atom. 4′-CP, which lacks one phenol-hydroxyl group, was about half as active as BPA. 2,2′-DPP, lacking both of the phenol-hydroxyl groups, was ineffective (<20%). Furthermore, four aromatic methyl groups at meta-position (TMBPA) decrease the activity, whereas TBBA with four bromines as substituents was almost as active as BPA. Obviously, an aromatic substituent like bromine with a positive mesomeric effect and a predominant negative inductive effect supports the ability of meta-substituted compound to inhibit Ca2+ channels by reducing the electron density of π-systems.
In addition, the double-methylated sp3-hybridized carbon atom present in BPA and TBBA, which bridges the two phenol rings, is optimal for the ability of the compounds to inhibit Ca2+ channels. In fact, successive removal of the methyl groups revealed less effective bisphenols (BPE, BPF). Remarkably, BPAF, where the methyl groups are replaced with trifluoromethyl moieties was twice as potent as BPA (EC50 = 13 ± 1.15 µM, data not shown). BPAB, which has a bulky bridging structure, (R-C(CH3)(CH2CH3)-R), was almost equally effective compared with BPA. Furthermore, sp2-hybridization of this carbon atom or introduction of different bridging structures led to bisphenols that were ineffective (4′4- HBP, BPS). It is interesting to note, that biphenol [(1,1'-biphenyl)-4,4'-diol] did not show any activity, which again confirms that a sp3-hybridized carbon atom bridging the aromatic rings is essential for the blocking effect on Ca2+ channels.
Apparently, two aromatic ring moieties in angulated orientation are required for Ca2+ channel block. The blocking activity of compounds lacking one of the aromatic rings (4′-TBP; 4′-TAP) was much less pronounced than that of BPA (Fig. 8B; Table 2). Finally, the ability to inhibit Ca2+ channels by sterically demanding bisphenols like BPM strongly depends on the spatial orientation of all aromatic rings. In BPM, which blocked voltage-activated Ca2+ channels by ∼70%, these rings are arranged in line. In case the molecule is angulated as in BPP, where one of the aromatic rings is rotated almost through 90° to form a right-angled structure, inhibition of voltage-activated Ca2+ channels was negligible.
In this study we present evidence that BPA interacts with voltage-activated Ca2+ channels as inhibitory ligand. BPA’s efficacy as inhibitor is comparable to that of polyvalent cations such as cadmium, cobalt, and manganese (Carbone and Swandulla, 1989). Pharmacological experiments with specific organic blockers (nifedepine and ω-conotoxins) for the different Ca2+ channel types showed that BPA affects all subtypes studied here (L-, N-, P/Q-, R-, T-type) to the same extent. Detailed analysis of biophysical properties on human R-type channels revealed that channel kinetics (activation, inactivation, and deactivation) and steady-state characteristics (activation and inactivation) were not significantly altered by BPA.
In summary, our biophysical and pharmacological analysis strongly suggests that BPA exerts its action by binding to the channels in their resting state. This is supported by the findings that 1) the amount of block is independent of the frequency and duration of current activation, 2) the current kinetics are not altered, 3) there is no shift in the steady state inactivation curve. Furthermore, our findings suggest that the binding site is located at the extracellular part of the pore-forming subunit. BPA as a highly lipophilic substance might reach a binding site at the transmembranal part of the channel protein. However, when applied intracellularly BPA was ineffective.
Direct evidence that BPA interacts with a specific binding site comes from the analysis of structurally related phenol and bisphenol derivatives. To be effective the molecules have to meet certain structural requirements. From the binding motive of a double-methylated or double-trifluoromethylated sp3-hybridized carbon atom flanked by two phenol moieties in angulated orientation, one can strongly assume a specific binding site at the various Ca2+ channels that interacts with effective compounds. The ability of BPA and related compounds to specifically interact with proteins has already been demonstrated for certain hormone receptors (Matsushima et al., 2007). Similar findings were described for BPA estrogen-related receptor-γ (ERR-γ) interactions. Whereas BPA (IC50 = 9.78 ± 0.87 nM) showed a high binding affinity to ERR-γ, BPF was approx. (IC50 = 131 ± 17.9 nM) 16-fold less potent, which clearly demonstrated the importance of the double-methylated central carbon atom for ERR-γ interactions. (Matsushima et al., 2008; Okada et al., 2008). In BPAF the two methyl groups are replaced by two electron-rich trifluoromethyl moieties, which reduced the binding affinity to the ERR-γ considerably by a factor of 35. In contrast, the inhibitory effect on Ca2+ channels was more pronounced with BPAF compared with BPA. Furthermore, 2,2′-DPP lacking both of the phenol-hydroxyl groups did not show any ERR-γ binding compared with 4′-CP, which was as active as BPA. Although there are some matches in the compound’s activity profiles for ERR-γ and Ca2+ channels, we do not assume that the two binding sites for BPA possess major structural similarities on the two proteins.
Since BPA and several of its related compounds are environmental pollutants that can be incorporated into the human organism, exposure to these chemicals may cause serious health problems. Urinary levels of BPA have recently been associated with chronic diseases, including heart disease, diabetes, as well as neurobehavioral changes in toddlers (Lang et al., 2008; Melzer et al., 2010; Sathyanarayana et al., 2011).
The Environmental Protection Agency considers a safe level of exposure to be 50 μg/kg of body weight per day (U.S. EPA, 2010). Recent experiments with primates, however, indicate that human exposure may be much higher than previously assumed (Taylor et al., 2011). Under certain conditions exposure to BPA may exceed putative safe levels. This could be the case in manufacturing facilities. Indeed recent studies have shown that workers in epoxy resin factories can have approximately 1000 times higher urinary BPA concentrations compared with control groups (He et al., 2009; Melzer et al., 2010; Wang et al., 2012). The highest levels measured were almost up to 10 μM (Wang et al., 2012), which is within the effective concentration range of Ca2+ channels block by BPA. High exposure to BPA has also been reported for patients, particularly premature infants undergoing intensive medical care treatment (Calafat et al., 2009).
In general the incorporation of the lipid-soluble compound BPA in certain body compartments is not very well studied and it is conceivable that BPA could accumulate to micromolar concentrations in the human body (see above). This may apply also to highly lipophilic, halogenated BPA derivatives such as tetrabromobisphenol A (TBBA) one of today’s most abundantly used brominated flame retardants. Indeed, it has been shown that in mice fed with 100 µg TBBA/kg body weight TBBA accumulates over-proportionally in the striatum compared with other brain regions (Nakajima et al., 2009).
The toxicity of BPA has been evaluated with respect to development, reproduction, and cancer (see, e.g., Xu et al., 2010; Fernandez et al., 2012). In this context, a recent Food and Drug Administration update points to “some concern about the potential effect (of BPA) on the brain, behavior and prostate gland in fetuses, infants, and young children” (for review see Wolstenholme et al., 2011). The BPA concentration administered in toxicological studies is expected to result in micromolar concentrations in different compartments of the body and should critically depend on BPA metabolism. Exact values in humans, however, are currently not available.
Given the fact that BPA and its related compounds are ubiquitously contaminating our environment, we expect that further studies will surface that medical diseases such as cardiovascular, respiratory, and metabolically caused disorders can at least partly be attributed to the effect of BPA on voltage-activated Ca2+ channels. Our findings that certain BPA-related compounds are less or even non-effective on voltage-activated Ca2+ channels may provide a key to finding molecules that could substitute for BPA in large-scale plastic production.
The authors thank T. Schneider for providing the α1Ε Ca2+-channel expressing cell line, and M. Zweyer and H. Bock for excellent technical assistance.
Participated in research design: Swandulla, Hans, Meyer.
Conducted experiments: Deutschmann, Swandulla, Hans.
Performed data analysis: Deutschmann, Hans, Swandulla, Häberlein, Meyer.
Wrote or contributed to the writing of the manuscript: Swandulla, Hans, Häberlein, Meyer, Deutschmann.
- 4,4′-(2,2-propanediyl)diphenol (also bisphenol A)
- concentration-effect relationship
- dorsal root ganglion
- Dulbecco′s modified Eagle′s medium
- estrogen-related receptor-γ
- human embryonic kidney
- high voltage–activated
- low voltage–activated
- protein disulfide isomerase
- 4,4′-(1-methylethylidene)bis-2,6-dibromophenol (also tetrabromobisphenol A)
- Received July 26, 2012.
- Accepted November 29, 2012.
- Copyright © 2013 by The American Society for Pharmacology and Experimental Therapeutics