Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

In the past decade, there has been growing concern about the effects of hormone-disrupting chemicals to various aspects of animal and human health (Roy et al., 1997; Sonnenschein and Soto, 1998). These chemicals accumulate to high concentrations in the human body due to their high lipid-solubility (Rivero-Rodriguez et al., 1997) and are capable of mimicking or antagonizing endogenous hormones, perturbing the normal physiology of life (Roy et al., 1997). Among these chemicals, bisphenol A is one of the major compounds that we are exposed to in daily life; it contaminates many consumer products, such as foods and beverages, in containers lined with polycarbonate (Sonnenschein and Soto, 1998). Bisphenol A also leaches into the human body from dental sealants and composite fillings and can be detected in saliva (up to 950 µg/h) of patients with treated teeth (Sonnenschein and Soto, 1998). Indeed, bisphenol A absorption through the skin was shown to produce extensive damage to lung, kidney, liver, spleen, and pancreas (Sax, 1975). A study of tissue distribution of bisphenol A in rats found it predominantly in the lung (Yoo et al., 2000). To date, evidence has been provided that bisphenol A mimics genomic effects of estrogens by binding cytosolic estrogen receptors (ER and ER) and causes developmental and reproductive damage and carcinogenesis by interfering with normal endocrine function (Morrissey et al., 1987; Roy et al., 1997). Besides these genomic effects, xenoestrogens and estrogens are widely recognized to have rapid nongenomic effects via specific outer membrane receptors (Nadal et al., 2000). The interaction, for example, leads to changes in the kinetics of ion channels, such as K⁺ (White et al., 1995; Liu et al., 1998), L-type Ca²⁺ (Nakajima et al., 1995; Ruehlmann et al., 1998), and Cl channels (Valverde et al., 1993; Zhang et al., 1994), in various kinds of cells. However, less attention has been paid to the effects of hormone-disrupting chemicals on the epithelial ion transport system. Active transcellular Cl transport across the epithelial cells, followed by Na⁺ and water movement through the paracellular pathway, are important in the formation of low-viscosity mucus, thereby maintaining a conductive and aseptic environment in the lung (Quinton, 1990). Therefore, dysfunction of Cl transport results in production of thick mucous plugs and consequently causes respiratory disorders. In the present study, we focused especially on the nongenomic effects of bisphenol A on transepithelial Cl secretion in human airway epithelial Calu-3 cells. This cell line is known to retain a phenotype analogous to that of submucosal serous airway cells, expressing abundant cystic fibrosis transmembrane conductance regulator (CFTR) Cl channels on the apical membrane (Haws et al., 1994; Shen et al., 1994; Ito et al., 2000). Herein, we demonstrate that bisphenol A disturbs electrogenic Cl secretion through down-regulation of the basolateral K⁺ channels that provide a driving force for apical Cl export.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Culture. Calu-3 human airway cells purchased frozen (80°C) from American Type Culture Collection (Manassas, VA) were grown in a 1:1 mixture of Dulbecco's modified Eagle's medium and Ham's F-12 (Invitrogen, Carlsbad, CA) containing 10% fetal bovine serum (Invitrogen), 100 µg/ml streptomycin, and 100 U/ml penicillin (Invitrogen). The cells were maintained in culture flasks (T₇₅) at 37°C in an atmosphere of 5% CO₂ in air. When 80 to 90% confluent, cells were detached with a solution of phosphate-buffered saline, 0.04% EDTA, and 0.25% trypsin. The collected cells were passaged with a 1:4 dilution of the same solution and plated at 10⁶ cells/cm² on snapwell inserts (0.4-µm pore size, 12-mm diameter, polyester; Costar, Cambridge, MA). The membrane filters on the inserts had been coated overnight with 0.2 mg/ml human placental collagen type VI (Sigma-Aldrich, St. Louis, MO). One day after seeding the cells on the filters, the medium remaining on the apical side was removed to establish an air interface, which markedly improves the differentiation of human airway epithelia in a well polarized fashion. The cells were fed by replacement of the basolateral medium every 48 h. Every experiment was conducted after 7 to 13 days in culture.

Measurement of Short-Circuit Current (I_sc) and Apical Cl Conductance. The filter inserts where cells had grown confluent were mounted in modified Ussing chambers (EasyMount Chamber; Physiologic Instruments, San Diego, CA) with physiological saline solution composed of 140 mM NaCl, 5 mM KCl, 1 mM MgCl₂, 2 mM CaCl₂, 10 mM glucose, and 10 mM Hepes (pH adjusted to 7.4 at 37°C). The bathing solutions were kept at 37°C and bubbled with 21% O₂/79% N₂. The monolayers were continuously open-circuited to monitor transepithelial potential differences by a high-impedance millivoltmeter functioning as a voltage clamp with automatic fluid- resistance compensation (VCC MC2; Physiologic Instruments). Every 20 s, we applied a 2-µA pulse for 0.5 s under open-circuit conditions through pipette-shaped passing-current electrodes made of Ag wire filled with a solution of 3% (w/v) agarose in 3 M KCl solution to cause voltage deflections ( potential difference). This procedure enabled us to calculate transepithelial conductance (G_t) by Ohm's law (G_t = 2 µA/ potential difference). When the I_sc was measured, the potential difference was clamped to 0 mV by the amplifier. The solutions in the apical and basolateral chambers were symmetrical to measure active ion currents through the transcellular pathways. A positive I_sc was defined as a net flow of cations from the apical to the basolateral side. Sustained Cl transport produced by cAMP-related agents, such as isoproterenol, forskolin, or 8-bromo-cAMP, was evaluated by measuring the I_sc reduction for 30 min after bilateral application of 100 µM 5'-nitro-2-(3-phenylpropylamino) benzoate (NPPB; a Cl-channel blocker); this is called the NPPB-sensitive I_sc. It is well established that 1-ethyl-2-benzimidazolinone (1-EBIO) activates charybdotoxin-sensitive Ca²⁺-activated K⁺ (K_Ca) channels, causing sustained Cl secretion (Devor et al., 1999). The inhibitory effects of 100 nM charybdotoxin on the sustained I_sc in response to 1-EBIO were almost saturated in 10 min. To assess the K_Ca channel-dependent Cl secretion, we measured I_sc reduction for 10 min after basolateral application of charybdotoxin (100 nM); this is called the charybdotoxin-sensitive I_sc. To analyze apical membrane conductance (G_Ap), the basolateral membrane was permeabilized with 100 µM nystatin applied to the basolateral solution (Ito et al., 2000). The NPPB-sensitive component of G_Ap (NPPB-sensitive G_Ap) reflects the apical Cl permeability through CFTR in Calu-3 cells (Ito et al., 2000, 2001a) because the only Cl channels detected on the apical membrane of Calu-3 cells are CFTR channels (Haws et al., 1994). Values of the NPPB-sensitive G_Ap in the sustained component of isoproterenol-induced I_sc were obtained by measuring the decrease in G_Ap during 30-min exposure to NPPB (100 µM, bilateral) applied 30 min after isoproterenol treatment.

Measurement of Apical Membrane Cl Current and Basolateral Membrane K+ Current. To assess apical membrane Cl current (I_Cl), the basolateral membrane was permeabilized with nystatin (100 µM) for more than 30 min, and an apical-to-basolateral Cl concentration gradient was established. This procedure avoids the complexities associated with basolateral ion transporters and permits analyses of apical membrane Cl conductance. (Devor et al., 1999). Apical NaCl was replaced by equimolar Na-gluconate, and CaCl₂ was increased to 4 mM to compensate for the Ca²⁺-buffering capacity of the gluconate (Devor et al., 1999). The basolateral membrane K⁺ current (I_K) was estimated after permeabilization of the apical membrane with nystatin (50 µM) for more than 30 min and establishment of an apical-to-basolateral K⁺ concentration gradient (Devor et al., 1999). Apical NaCl was replaced by equimolar K-gluconate, whereas basolateral NaCl was substituted with equimolar Na-gluconate. Cl was removed from these solutions (Wong et al., 1990).

Chemicals. Isoproterenol, forskolin, 8-bromo-cAMP, 1,2-bis-(o-amino-phenoxy)-ethane-N,N,N',N'-tetraacetic acid tetra-(acetoxymethyl)-ester (BAPTA-AM), 17-estradiol, tamoxifen, NPPB, and phlorizin were obtained from Sigma-Aldrich. Clotrimazole was purchased from Calbiochem (San Diego, CA), and 1-EBIO was from Sigma-Aldrich. Charybdotoxin and bisphenol A were obtained from Peptide Institute, Inc. (Osaka, Japan) and Wako Chemical (Tokyo, Japan), respectively. Isoproterenol, 8-bromo-cAMP, and charybdotoxin were dissolved in distilled water. All other drugs were dissolved in dimethyl sulfoxide (DMSO). Nystatin stock solution (100 mM) was made and sonicated for 30 s just before use.

Analysis of Results. Concentration-response curves in the present study were obtained using a computer program Cricket Graph version 1.5.3 for Macintosh (Computer Associates International, Inc., Islandia, NY). All data are expressed as means ± S.E. with the number of experiments used (n). Statistical difference was determined by Student's t test or one-way analysis of variance. A value of p < 0.05 was considered to indicate statistical significance.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effects of Bisphenol A on Cl Secretion Stimulated by cAMP-Related Agents. Every experiment in the present study using Calu-3 cells was conducted in the apical presence of the Na⁺-glucose cotransport inhibitor phlorizin (200 µM). This allowed us to observe transepithelial Cl movement by eliminating the Na⁺ transport component in subsequent measurements (Mizuno et al., 2000; Ito et al., 2001a,b). Under this condition, application of isoproterenol (10 nM) from the basolateral face led to a rapid and biphasic increase in I_sc (Fig. 1A), a transiently increased component (isoproterenol-induced I_sc = 15.2 ± 1.2 µA/cm²; n = 7) followed by a sustained one. This response was exclusively composed of transepithelial Cl current sensitive to NPPB (Mizuno et al., 2000; Ito et al., 2001a). The NPPB-sensitive component in the sustained I_sc 30 min after the addition of isoproterenol (NPPB-sensitive I_sc) was 6.8 ± 0.5 µA/cm² (n = 7). However, cells exhibited suppressed isoproterenol-induced responses due to the 30-min pretreatment with 100 µM bisphenol A (isoproterenol-induced I_sc = 3.6 ± 1.2 µA/cm²; NPPB-sensitive I_sc = 2.8 ± 0.8 µA/cm²; n = 4; p < 0.001; Fig. 1A). The concentration dependence of the inhibition using isoproterenol-induced I_sc is demonstrated in Fig. 1B. Although estrogen (17-estradiol) and anti-estrogen (tamoxifen) mimicked the inhibitory effects of bisphenol A, higher concentrations were necessary to exert equal effects. As shown in Fig. 1B, the IC₅₀ values of the isoproterenol-induced I_sc were approximately 60 µM for bisphenol A and 1000 µM for 17-estradiol and tamoxifen. To determine the target sites of bisphenol A, changes in apical conductance were followed after exposure to isoproterenol in the presence and absence of bisphenol A. As shown in Fig. 2A, isoproterenol (10 nM) elicited an increase in the conductance of the basolaterally permeabilized monolayer. Under this condition, this change following nystatin permeabilization indicates the G_Ap (see Materials and Methods above). However, no discernible difference was made by bilateral pretreatment with 100 µM bisphenol A. The NPPB-sensitive component of the G_Ap (NPPB-sensitive G_Ap) measured after 30 min exposure to isoproterenol (1843 ± 213 µS/cm²; n = 4) was unaffected by bisphenol A (1905 ± 245 µS/cm²; n = 4). In addition, the NPPB-sensitive G_Ap was also insensitive to 1 mM 17-estradiol (1623 ± 140 µS/cm²; n = 8) and 1 mM tamoxifen (1705 ± 272 µS/cm²; n = 4). This suggests that apical CFTR Cl channels may not be involved in the attenuation of isoproterenol-induced I_sc produced by these three estrogenic compounds. Figure 3, A and B, show that the effects of bisphenol A on isoproterenol were also simulated in responses after stimulation with forskolin (10 µM) and 8-bromo-cAMP (1 mM).

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Effects of bisphenol A on Isc in response to the -adrenergic agonist isoproterenol in Calu-3. To block transepithelial Na+ transport, every experiment was conducted in the presence of a Na+-glucose transport blocker, phlorizin (200 µM). A, isoproterenol (10 nM) applied to the basolateral solution 30 min after adding 0.05% DMSO (vehicle) produced a biphasically increased Isc that is markedly reduced by the bilateral pretreatment with bisphenol A (100 µM). 5'-Nitro-2-(3-phenylpropylamino) benzoate (NPPB; 100 µM) was applied to the bilateral solutions 30 min after addition of isoproterenol to evaluate the sustained Cl transport. B, inhibitory effects of bisphenol A at various concentrations on the isoproterenol (10 nM)-induced increase in Isc (isoproterenol-induced Isc) compared with those of 17-estradiol and tamoxifen. Data are means ± S.E. (n = 4-11).

View larger version (33K): [in this window] [in a new window]   Fig. 2.   Effects of bisphenol A, 17-estradiol (17-ES) and tamoxifen (TAM) on apical Cl conductance. A, conductance of the monolayer was followed after permeabilization of the basolateral membrane with 100 µM nystatin. After exposure to bisphenol A (100 µM, bilateral) or its vehicle (0.05% DMSO) for 30 min, isoproterenol (ISO) was added to the basolateral solution. Subsequently, NPPB (100 µM, bilateral) was applied 30 min after application of ISO to measure the NPPB-sensitive component of apical conductance (NPPB-sensitive GAp). B, the NPPB-sensitive GAp was measured in cells pretreated with bisphenol A (100 µM), 17-ES (1 mM), or TAM (1 mM) in the same time course as A. Note that ISO markedly potentiated the conductance that was unaffected by bisphenol A, 17-ES, or TAM. Data are means ± S.E. (n = 4-8). *, significantly different from the control (ISO-untreated condition) with p < 0.0001.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Effects of bisphenol A on the forskolin- and 8-bromo cAMP-induced Isc. Every experiment was performed in the presence of phlorizin (200 µM). Forskolin (10 µM) (A) or 8-bromo cAMP (1 mM) (B) was applied to the basolateral solution pretreated with bisphenol A (100 µM) or 0.05% DMSO (vehicle) in the bilateral solutions. NPPB (100 µM) was applied in the bilateral solutions 30 min after addition of forskolin or 8-bromo cAMP. Bisphenol A prevented Isc in response to forskolin and 8-bromo cAMP. Data are means ± S.E. (n = 4-11).

Separate Measurements of Apical I_Cl and Basolateral I_K. Sustained electrogenic Cl secretion requires the driving force produced by basolateral membrane K⁺ conductance (MacVinish et al., 1998). From the results in Fig. 2, we speculated that basolateral K⁺ conductance is related to the inhibitory effects of bisphenol A on cAMP-dependent Cl secretion. Thus, after permeabilization of the basolateral or apical membrane with nystatin, apical I_Cl or basolateral I_K was measured under the establishment of transepithelial Cl or K⁺ gradients, respectively. Consequently, 8-bromo-cAMP (1 mM, basolateral) stimulated inward sustained I_Cl, which was insensitive to bisphenol A (100 µM; Fig. 4A), whereas the cAMP-dependent outward sustained I_K was abolished by the presence of bisphenol A (Fig. 4B). The peak values of the stimulated I_K (15.0 ± 3.6 µA/cm²; n = 7) were suppressed to 2.7 ± 0.3 µA/cm² (n = 5; p < 0.05) by bisphenol A (Fig. 5). This current was unaffected by the intracellular Ca²⁺ chelator BAPTA-AM (10 µM, bilateral; 14.2 ± 3.2 µA/cm²; n = 4) or charybdotoxin (100 nM, basolateral; 16.5 ± 2.9 µA/cm²; n = 4) but was very sensitive to clotrimazole (30 µM, basolateral; 1.3 ± 0.4 µA/cm²; n = 6, p < 0.01). In light of the fact that clotrimazole is a dual blocker of the Ca²⁺ (K_Ca)-activated and cAMP-activated K⁺ (K_cAMP) channels (Devor et al., 1999), Calu-3 monolayer seems likely to possess the K_cAMP channel on the basolateral membrane to generate sustained Cl secretion, such as colonic epithelia (MacVinish et al., 1998).

View larger version (19K): [in this window] [in a new window]   Fig. 4.   A, effects of bisphenol A (100 µM, bilateral) on the 8-bromo-cAMP (1 mM)-induced ICl following establishment of a basolateral-to-apical Cl gradient and permeabilization of the basolateral membrane with nystatin (100 µM). A negative ICl represents an absorptive Cl flow from the apical to the basolateral side. B, effects of bisphenol A (100 µM, bilateral) and its vehicle (0.05% DMSO) on the 8-bromo-cAMP (1 mM)-induced IK following establishment of a apical-to-basolateral K+ gradient in the monolayer apically permeabilized with nystatin (50 µM). A positive Isc represents a net flow of K+ from the apical to the basolateral side. Note that bisphenol A inhibited the IK without affecting the ICl. Data are means ± S.E. (n = 4-7).

View larger version (30K): [in this window] [in a new window]   Fig. 5.   Effects of bisphenol A (100 µM) on the 8-bromo-cAMP (1 mM, basolateral)-induced peak increase in IK (IK) following establishment of an apical-to-basolateral K+ gradient in the monolayer apically permeabilized with nystatin (50 µM). The IK values were measured in the presence of BAPTA-AM (10 µM, bilateral), charybdotoxin (100 nM, basolateral), clotrimazole (30 µM, bilateral), bisphenol A (100 µM, bilateral), and bilateral 0.05% DMSO (control) that were applied 30 min before addition of 8-bromo-cAMP. Data are means ± S.E. (n = 4-7). Significantly different from the control with p < 0.05 (*) and p < 0.01 (**).

Asymmetry of the Effects of Bisphenol A. To locate the site where bisphenol A affects the basolateral K_cAMP channel, it was unilaterally applied. As shown in Fig. 6A, the peak and NPPB-sensitive I_sc after the addition of isoproterenol were reduced by bisphenol A (100 µM) applied to either side of the monolayer. Figure 6B shows the isoproterenol (10 nM)-induced increase in I_sc (isoproterenol-induced I_sc) in the apical or basolateral presence of bisphenol A at various concentrations (10-1000 µM). The effects of bisphenol A were remarkable in the apical (IC₅₀ = 90 µM) rather than the basolateral application (IC₅₀ = 310 µM). These data suggest that the functional sites of bisphenol A for the K_cAMP channel face the cytosolic side.

View larger version (24K): [in this window] [in a new window]   Fig. 6.   Differences between the bisphenol A-produced effects in apical and basolateral applications. Experiments were conducted in the presence of phlorizin (200 µM). A, isoproterenol (10 nM) was applied 30 min after adding bisphenol A either in the apical or basolateral solutions, resulting in diminution of the transient and sustained actions of isoproterenol. Subsequently, 5'-nitro-2-(3-phenylpropylamino) benzoate (NPPB, 100 µM) was applied in the bilateral solutions 30 min after addition of isoproterenol to evaluate the sustained Cl transport. B, inhibitory effects of bisphenol A at various concentrations applied to either apical or basolateral side on the isoproterenol-induced increase in Isc (Isoproterenol-induced Isc). Note that apical bisphenol A has more effect on isoproterenol-induced actions than basolateral bisphenol A does. Data are means ± S.E. (n = 4). Significant differences from the corresponding values with basolateral bisphenol A are expressed with * (p < 0.05), ** (p < 0.01), and *** (p < 0.001).

Effects of Bisphenol A on 1-EBIO-Induced I_sc. 1-EBIO is a pharmacological agent for generating sustained Cl secretion by directly activating the intermediate conductance (10-31 pS) inward-rectifying K_Ca (hIK1) channel (Ishii et al., 1997) without affecting the cytosolic Ca²⁺ concentration in Calu-3 cells (Devor et al., 1999; Singh et al., 2000a). This channel activity was virtually eliminated by charybdotoxin and clotrimazole (Syme et al., 2000). As shown in Fig. 7A, bilateral application of 1-EBIO (500 µM) elicited an initial peak I_sc (13.3 ± 1.5 µA/cm²; n = 13) followed by a sustained one (11.3 ± 1.0 µA/cm², 30 min after the stimulation). The 1-EBIO-stimulated I_sc was significantly affected by bilateral pretreatment with 100 µM bisphenol A (the transient I_sc was 6.1 ± 0.9 µA/cm², p < 0.01; the sustained I_sc was 4.7 ± 0.5 µA/cm², n = 4, p < 0.001; Fig. 7A). charybdotoxin-sensitive I_sc measured 30 min after addition of 1-EBIO were 7.3 ± 0.7 µA/cm² (n = 13) in the control (vehicle) and 2.0 ± 0.7 µA/cm² (n = 4, p < 0.01) in the bilateral presence of 100 µM bisphenol A. The IC₅₀ of the effects of bisphenol A on charybdotoxin-sensitive I_sc is approximately 80 µM (Fig. 7B). The asymmetry of the effects of bisphenol A on 1-EBIO-induced responses demonstrated in Fig. 7B is similar to the results in Fig. 6. Charybdotoxin-sensitive I_sc was suppressed more strongly by bisphenol A applied from the apical (IC₅₀ = 130 µM) than from the basolateral side (IC₅₀ = 500 µM). In contrast, 17-estradiol (300-1000 µM) and tamoxifen (300-1000 µM) were without effect on the charybdotoxin-sensitive component of the 1-EBIO-stimulated I_sc, albeit at the highest concentrations (1 mM).

View larger version (25K): [in this window] [in a new window]   Fig. 7.   Actions of 1-EBIO (500 µM, bilateral) in the presence and absence of bisphenol A. Experiments were carried out using cells pretreated with phlorizin (200 µM). A, the 1-EBIO-induced Isc responses were markedly inhibited by bisphenol A (100 µM, bilateral) applied 30 min before 1-EBIO stimulation. Charybdotoxin (100 nM, basolateral) was added 30 min after 1-EBIO stimulation to evaluate the charybdotoxin-sensitive component of the responses. B, inhibitory effects of bisphenol A, 17-estradiol, and tamoxifen at various concentrations on the charybdotoxin-sensitive component of 1-EBIO-generated sustained Isc (charybdotoxin-sensitive Isc). Apical (Api) or bilateral (Bil) BPA was more effective on the charybdotoxin-sensitive Isc than basolateral (Baso) bisphenol A. 17-Estradiol and tamoxifen were without effect on charybdotoxin-sensitive component of the 1-EBIO-stimulated Isc Data are means ± S.E. (n = 4-13). Significant differences from the corresponding values with basolateral bisphenol A are expressed with * (p < 0.05) and ** (p < 0.01). Significant differences from the corresponding values with apical bisphenol A are expressed with # (p < 0.05) and ## (p < 0.005).

Effects of Long-Term Incubation of Bisphenol A on Isoproterenol- and 1-EBIO-Induced I_sc. To further examine the genomic effects of bisphenol A on Cl transport, I_sc in response to isoproterenol or 1-EBIO was evaluated after long-term incubation (12, 24, 48, and 72 h) with bisphenol A at lower concentrations (1 µM) than those required to exert its nongenomic effects. No distinction was seen between these responses in the presence and absence of bisphenol A (Fig. 8).

View larger version (45K): [in this window] [in a new window]   Fig. 8.   Effects of long exposure to bisphenol A (1 µM) on Isc in response to ISO or 1-EBIO. After exposure to bisphenol A (1 µM) for 12 to 72 h, the peak increment of Isc just after application of 10 nM ISO (A), NPPB-sensitive Isc 30 min after addition of ISO (B), and charybdotoxin (ChTx)-sensitive Isc 30 min after addition of 1-EBIO (C) is shown. No distinction was made between these responses in the presence and absence of bisphenol A. Data are means ± S.E. (n = 4-5). *, p < 0.05.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The CFTR is a major Cl secretory pathway that greatly contributes to regulating the amount of liquid on the respiratory tract. In cystic fibrosis patients, genetically defective responses of the CFTR Cl channel to cAMP-mediated signals result in airway dehydration, defective mucociliary clearance, and consequent bacterial infection (Quinton, 1990). The present study provided evidence that short-term application of the ubiquitous xenoestrogen bisphenol A blocks cAMP-mediated Cl secretion in Calu-3 cells. It has been previously established that bisphenol A exhibits estrogenicity approximately 2,000 to 15,000 times less potent than 17-estradiol in vitro (Krishnan et al., 1993; Gaido et al., 1997). Nevertheless, our data revealed that the potency of bisphenol A on Cl secretion in response to isoproterenol was much higher than that of 17-estradiol. Additionally, the antiestrogen tamoxifen mimicked the 17-estradiol actions with equivalent potency, suggesting that the effects of bisphenol A were not mediated by classical estrogen receptor-dependent (genomic) mechanisms. Because the genomic effects of estrogen involve a complex process, inclusive of ligand-receptor binding in the cytosol, targeted gene expression, and protein synthesis, it may take hours for the onset of hormone actions (Falkenstein et al., 2000; Li and Hay, 2000). In contrast, the nongenomic effects, which are acute in onset, require membrane estrogenic receptors without involvement of nuclear estrogen receptors and gene expression (Falkenstein et al., 2000). Thus, the inhibition of bisphenol A on Cl secretion are attributable to nongenomic mechanisms.

It is currently axiomatic that the driving force of apical Cl export is produced by the basolateral K⁺ conductance (Moon et al., 1997; Singh et al., 2001). In epithelial cells, sustained Cl secretion is generated by cAMP-related agents through simultaneous activation of both apical CFTR and basolateral K_cAMP which are phosphorylated by protein kinase A. However, bisphenol A, 17-estradiol, and tamoxifen seem unlikely to affect cAMP/protein kinase A-dependent cascades because the isoproterenol-stimulated NPPB-sensitive G_Ap reflecting CFTR-mediated anion conductance was intact even in the presence of these three estrogenic compounds. That is, at least in Calu-3 cells, the observed inhibitory effects of bisphenol A, 17-estradiol, and tamoxifen are localized to the basolateral membrane, and the CFTR is not a target of these inhibitory actions. This result conflicts with a report by Singh et al. (2000b) using colon epithelial T₈₄ cells possessing a heterologous Cl exit pathway on the apical membrane. Focusing on the effects of bisphenol A, separate measurement of apical I_Cl and basolateral I_K revealed that bisphenol A attenuates basolateral I_K in response to 8-bromo-cAMP, while not affecting apical I_Cl. Considering the data above, it is more likely that the effects of bisphenol A on cAMP-dependent Cl secretion were due to direct inhibition of basolateral K⁺ channels in Calu-3 cells. Actually, 17-estradiol and tamoxifen have been shown to directly inhibit or activate various kinds of K⁺ channels through nongenomic mechanisms (Liu et al., 1998; Okabe et al., 1999; Valverde et al., 1999; Li and Hay, 2000), whereas the effects of bisphenol A on the K⁺ channels have never been documented.

In the single Calu-3 cell, an hIK1 channel has been detected as a K⁺ channel using the excised patch-clamp technique (Devor et al., 1999). However, it is difficult to identify K⁺ channels expressed on the basolateral membrane in well polarized cells forming a monolayer. Moon et al. (1997) has reported that the basolateral membrane in Calu-3 contains K_Ca channels, but no K_cAMP channel, because it responds to thapsigargin but not to forskolin. In contrast, our previous data (Ito et al., 2001b) and those of Devor et al. (1999) have shown that forskolin generated sustained I_sc. These forskolin-induced responses are insensitive to charybdotoxin (an hIK1 blocker) and BAPTA-AM (an intracellular Ca²⁺ chelator) in the Calu-3 cell monolayer (Ito et al., 2001b). To characterize the cAMP-mediated basolateral K⁺ current, we used clotrimazole, which inhibits both hIK1 and K_cAMP channels, with K_i values of 0.27 and 6.7 µM, respectively (Devor et al., 1999; Syme et al., 2000). Consequently, the 8-bromo-cAMP-induced I_K of the monolayer was sensitive to clotrimazole (30 µM) and bisphenol A but insensitive to charybdotoxin or BAPTA-AM. These observations suggest that 8-bromo-cAMP activates the K_cAMP channel without involvement of the hIK1 channel on the basolateral membrane of the Calu-3 cell monolayer and that the K_cAMP channel is an effective target of bisphenol A in the cAMP-generated Cl secretion.

Several lines of evidence demonstrated that 1-EBIO directly activates hIK1 channels in the presence of resting levels of cytosolic Ca²⁺, causing sustained Cl secretion (Olesen et al., 1994; Devor et al., 1999). In the present study, bisphenol A reduced the charybdotoxin-sensitive component of the 1-EBIO-elicited I_sc. Taken together, in the light of the insensitivity of apical CFTR to bisphenol A, these observations indicate that bisphenol A is a dual K⁺ (K_Ca/K_cAMP) channel blocker like clotrimazole.

Recently, Valverde et al. (1999) reported that 17-estradiol is capable of binding and activating the -subunit of the Maxi-K⁺ channel, which is a category of large conductance (15-250 pS) K_Ca channels. The -unit constitutes the channel pore, and the -unit constitutes the regulatory site susceptible to intracellular Ca²⁺. However, the present study demonstrated that 1-EBIO-induced responses are insensitive to 17-estradiol and tamoxifen, indicating that, unlike the Maxi-K⁺ channel, hIK1 channels have no effective binding sites for these two estrogenic compounds but do for bisphenol A.

Estrogen receptors have been characterized not only in the cytoplasm but also on the plasma membrane (Pappas et al., 1995). Nadal et al. (2000) has demonstrated that a membrane estrogen receptor facing the extracellular side is involved in the nongenomic reactions to estrogens and xenoestrogens and is unrelated to cytosolic/nuclear estrogen receptors. However, the results of the present study suggest that the effective sites for bisphenol A on the basolateral hIK1 and K_cAMP channel may be facing the cytosol because the inhibitory effects of bisphenol A on the K_cAMP- and K_Ca-mediated responses were remarkable when it was applied to the apical rather than the basolateral membrane. Namely, if the site were on the external side, basolateral bisphenol A would be more effective on the 1-EBIO and isoproterenol-induced I_sc. In well polarized epithelial Calu-3 cells, there seem to be differences in distribution of ion-transporting proteins and drug-permeability between the apical and basolateral membranes (Mizuno et al., 2000). Thus, the present result indicates that the apical membrane of Calu-3 cells may be more permeable to bisphenol A than the basolateral membrane. Furthermore, one of the considerable implications of the present study is the possible existence of inner surface proteins on the two basolateral K⁺ channels that function as nongenomic receptors for bisphenol A with higher affinity for it than for 17-estradiol and tamoxifen. Unfortunately, however, whether the intracellular effects of bisphenol A on K⁺ channels are direct or indirect is still elusive. Further experiments will unravel the mechanisms of the inhibition of K⁺ channels by bisphenol A.

Under physiological conditions, bisphenol A might not modulate K⁺ channel activity with consequent attenuation of Cl secretion since micromolar concentrations are necessary for bisphenol A to inhibit the basolateral K⁺ channels. However, environmental hormone-disrupting chemicals, with their high-lipid solubility, have been reported to accumulate to high concentrations in human tissue, especially in the lung (Sax, 1975; Yoo et al., 2000). Thus, chronic exposure to bisphenol A and its resultant accumulation might affect mucous clearance in the airway. In cystic fibrosis and chronic obstructive pulmonary diseases, this may increase mucus congestion, although common bisphenol A concentrations in human lung tissue at various ages are still unknown and should be clarified in the future.

In addition to membrane estrogen receptor-mediated nongenomic actions, estrogenic compounds could directly interact with the cell membrane, leading to alteration of membrane fluidity (Falkenstein et al., 2000). However, the changes in membrane fluidity are unlikely to be involved in bisphenol A inhibition of K⁺ channels because 17-estradiol and tamoxifen, despite their high lipophilicity, had smaller or no effect, albeit at much higher concentrations (1 mM). This argues against nonspecific effects of bisphenol A on the Cl secretion and rather for specific interactions with inner surface proteins of K⁺ channels.

In contrast to the nongenomic actions caused by estrogenic compounds, genomic ones are effective at submicromolar concentrations. In uterine and fetal rat epithelium, long-term incubation with estrogen influences CFTR expression via the genomic pathway (Rochwerger and Buchwald, 1993; Sweezey et al., 1997). On the other hand, no discernible difference in CFTR mRNA expression was noted between controls and pancreatic epithelial cells incubated 20 h with estrogen (Sweezey et al., 1996). We tested the effects of long-term incubation with bisphenol A at 1 µM but saw no significant effect on isoproterenol- and 1-EBIO-induced I_sc. These data suggests that bisphenol A may be without genomic effects on Cl transporters at least in Calu-3 cells.

Overall, bisphenol A interrupted CFTR-mediated Cl secretion by blocking two types of basolateral K⁺ channels (hIK1 and K_cAMP) without affecting the CFTR-mediated Cl current. The inhibitory effect may be mediated by neither the genomic nor the conventional nongenomic estrogenic receptors that have been reported; it suggests the presence of inner binding sites for bisphenol A unrelated to estrogen receptors on the two K⁺ channels.