Chronic bronchitis, a disease mainly of cigarette smokers, shares many clinical features with cystic fibrosis, a disease of altered ion transport, suggesting that the negative effects of cigarette smoke on mucociliary clearance may be mediated through alterations in ion transport. We tested the hypothesis that cigarette smoke extract would inhibit chloride secretion in human bronchial epithelial cells. In agreement with studies in canine trachea, cigarette smoke extract inhibited net chloride secretion without affecting sodium transport. We performed microelectrode impalements and impedance analysis studies to investigate the physiological mechanisms of this inhibition. These data demonstrated that cigarette smoke extract caused an acute increase in membrane resistances in conjunction with apical membrane hyperpolarization, an effect consistent with inhibition of an apical membrane anion conductance. After this acute phase, both membrane resistances decreased while membrane potentials continued to hyperpolarize, indicating that cigarette smoke extract also inhibited the basolateral entry of chloride into the cell. Furthermore, cigarette smoke extract caused an increase in mucin secretion. Therefore, the ion transport phenotype of human bronchial epithelial cells exposed to cigarette smoke extract is similar to that of cystic fibrosis epithelia in which there is sodium absorption out of proportion to chloride secretion in the setting of increased mucus secretion.
- chronic obstructive pulmonary disease
- cystic fibrosis
- cystic fibrosis transmembrane conductance regulator
- airway surface liquid
- mucociliary clearance
the mucociliary apparatus consists of the epithelial lining of the airways, cilia, a periciliary fluid layer (sol), and a mucus layer (gel). Together the periciliary fluid layer and the mucus layer make up the airway surface liquid (ASL) (38). Mucociliary clearance (MCC) comprises the organized movement of the ASL from the distal airway toward the pharynx. MCC is an innate lung defense mechanism that traps pollutants and inhaled particulate material and removes them from the lungs. Maintenance of normal MCC depends on proper ciliary beating and the biological properties of the ASL. Airway epithelial cells actively participate in the modulation and modification of ASL. In particular, the volume of the periciliary fluid layer (22) and the viscoelastic properties of the mucus (32) are affected by vectorial ion transport.
The hypothesis that alterations in ion transport lead to clinically significant alterations in MCC is strongly supported by observations of patients with cystic fibrosis (CF). In CF, mutations in the CFTR result in diminished apical membrane anion conductance and elevated sodium absorption (17). Together these changes in electrolyte transport cause a decrease in MCC that eventually results in chronic mucopurulent bronchitis and chronic infection of the lower airways (1). Chronic bronchitis, a disease seen mainly in cigarette smokers, shares with CF the presence of goblet cell hyperplasia, mucus hypersecretion, neutrophilic infiltration of the airway lumen, and the presence of bacterial pathogens in the lower airways (25). These similarities suggest that cigarette smoke negatively affects MCC, an effect that has been observed both clinically and experimentally (27, 28, 36).
Cigarette smoke is a complex mixture of biologically active chemicals (3) that affects more than one component of the mucociliary apparatus (9). Histological studies of cilia from the airways of smokers have shown both decreased numbers of ciliated cells (38) and increased numbers of ciliary ultrastructural abnormalities (34). In vitro studies have demonstrated that cigarette smoke upregulates the production of airway mucin (30). Furthermore, in vivo and in vitro studies with canine epithelia demonstrate that cigarette smoke acutely decreases chloride secretion (39). The purpose of the studies reported here was to test the hypothesis that cigarette smoke extract (CSE) would decrease chloride secretion in human bronchial epithelial cells (HBECs). To test this hypothesis we performed short-circuit current (Isc) measurements and unidirectional ion flux studies to assess active ion transport. In addition, microelectrode impalements and impedance analysis studies were performed to gain some insight into the mechanism of action of CSE-mediated inhibition of chloride secretion in HBECs.
MATERIALS AND METHODS
The use of HBECs was approved by the Institutional Review Board of the University of Pittsburgh. HBECs purchased from Cambrex (East Rutherford, NJ) were grown in bronchial epithelial cell growth medium (BEGM; Clonetics) supplemented with bovine pituitary extract (52 μg/ml), hydrocortisone (0.5 μg/ml), human recombinant epidermal growth factor (0.5 ng/ml), epinephrine (0.5 μg/ml), transferrin (10 μg/ml), insulin (5 μg/ml), retinoic acid (0.1 μg/ml), triiodothyronine (6.5 μg/ml), gentamicin (50 μg/ml), and amphotericin B (50 μg/ml). Medium was changed every 48 h until cells were 90% confluent. Cells were then passaged and seeded on polystyrene Snapwell inserts (Costar) in differentiation medium containing 50% DMEM in BEGM with the same supplements as above but without amphotericin B or triiodothyronine and with a final retinoic acid concentration of 50 nM (all-trans retinoic acid). Cells on inserts were kept submerged for the first 7 days in culture, after which time they were exposed to an apical air interface for the remainder of the culture period. Cells were used between days 14 and 21 after establishment of the apical air interface. At all stages of culture, cells were maintained at 37°C in 5% CO2 in an air incubator.
The bath solution contained (mM) 120 NaCl, 25 NaHCO3, 3.3 KH2PO4, 0.8 K2HPO4, 1.2 MgCl2, 1.2 CaCl2, and 10 glucose. The pH of this solution was 7.3–7.4 when gassed with a mixture of 95% O2-5% CO2 at 37°C.
CSE was generated in a fume hood by placing a cigarette (research cigarette 2RF; University of Kentucky) in a holder attached by rubber tubing to a variable-speed pump (Fig. 1). The pump outflow track was attached by rubber tubing to a glass straw submerged in 10 ml of warm, gassed bath solution. The solution was contained in a glass test tube with a rubber stopper that was vented to the atmosphere with a second glass straw. The smoke of eight cigarettes was passed through 10 ml of warmed, gassed bath solution to obtain 100% CSE. A final bath concentration of 20% (0.16 cigarettes/ml) CSE was used for all experiments, except for the initial dose-response experiment (5–40%). For the experiments examining the effect of CSE on mucin secretion, bath solution was replaced with cell culture medium.
Amiloride was added to the apical solution at 5 or 10 μM. Forskolin was added to both the apical and basolateral solutions at a final concentration of 0.6 or 2 μM. UTP was added to the apical bath at a final concentration of 30 μM. ATP was added to the apical medium at a final concentration of 100 μM. Bumetanide was added to the basolateral bath at a final concentration of 20 μM.
Short circuit (Isc) measurements.
Snapwell inserts were mounted in an Ussing chamber (Costar), and the monolayers were continuously short-circuited after fluid resistance compensation with automatic voltage clamps (558C-5; Iowa Bioengineering). The Isc was digitized at 60 samples/s, and data were stored on a computer hard drive with Gould hardware and software. Transepithelial resistance (Rt) was measured every 60 s with a 2-mV, 2-s, bipolar pulse, and the resistance was calculated by Ohm's law (V = IR, where V is potential, I is current, and R is resistance).
To examine the dose-response relationship of CSE on Isc inhibition, Isc was allowed to stabilize and then 0.25, 0.5, 1, or 2 ml of apical bath solution was removed and replaced with an equal volume of 100% CSE to obtain 5%, 10%, 20%, and 40% CSE, respectively. After Isc had stabilized, amiloride (10 μM) was added to the apical bath. Isc was allowed to stabilize again, after which UTP was added to the apical bath. Isc was read at the peak current seen after UTP and at the plateau current, ∼10 min after UTP stimulation. When Isc had reached a stable plateau, forskolin (0.6 μM) was added to both the apical and basolateral baths. Isc were taken at the peak current after forskolin stimulation and again at the plateau ∼10 min after forskolin stimulation. Once Isc had stabilized after forskolin, bumetanide was added to the basolateral bath. On the basis of the results of these experiments, we designed paired experiments to study the effect of 20% CSE on forskolin-stimulated Isc in the absence of UTP. For these experiments, Isc was allowed to stabilize and then amiloride (5 μM) was added to the apical bath. After an additional 10 min, forskolin (2 μM) was added to both the apical and basolateral baths. Isc was allowed to stabilize again, after which 1 ml of apical bath solution was removed and replaced with an equal volume of 100% CSE containing appropriate stimuli. Finally, bumetanide (20 μM) was added to the basolateral bath. In a third set of experiments designed to examine the time and stimulus dependence of CSE-mediated effects, Isc was allowed to stabilize and 1 ml of apical bath solution was removed and replaced with 100% CSE. Filters were incubated for 5, 10, 15, or 30 min, after which amiloride (10 μM) was added to the apical bath and forskolin (0.6 μM) was added to both the apical and basolateral baths.
Unidirectional ion fluxes.
Snapwell inserts were mounted in an Ussing chamber, and the monolayers were continuously short-circuited after fluid resistance compensation with automatic voltage clamps as described above. Fifteen minutes after the filters were mounted, amiloride was added to the apical solution. Ten minutes after addition of amiloride, 1 ml of apical bath solution was removed and replaced with either an equal volume of new apical bath solution with amiloride or an equal volume of 100% CSE plus amiloride. Forskolin (2 μM) was then added to the apical and basolateral baths. Immediately after application of forskolin, 36Cl (5 μCi) was added to either the apical (mucosal) or the basolateral (serosal) bath. When Isc had stabilized, two 0.1-ml samples were taken from the labeled bath. After an additional 2 min, two 0.4-ml samples were taken from the unlabeled bath and 0.8 ml of fresh, appropriate unlabeled bath solution was added back to maintain a constant volume. This time was considered time 0, and duplicate 0.4-ml samples were taken at 15-min intervals for the next 30 min. Isotope activities were determined in a liquid scintillation counter. All samples were weighed, and these volumes were used to correct the chamber volume and to calculate the serosal to mucosal chloride flux (J) and the mucosal to serosal chloride flux (J) with standard equations (2). Isc readings were taken at time 0 and at 30 min for conversion to microequivalents of Isc.
Microelectrodes were fabricated from borosilicate glass (BF100-50-10; Sutter Instrument) on a Sutter P-2000 laser puller and back-filled with 1 M KCl immediately before use. Microelectrodes ranged in resistance from 30 to 50 MΩ. The microelectrode was attached to channel A of a Duo 773 Electrometer (World Precision Instruments, Sarasota, FL). The electrode pairs used to measure transepithelial potential (Vt) were attached to a an automatic voltage clamp (558C-5; University of Iowa Bioengineering). Apical membrane potential (Va) and Vt were digitized at 60 samples/s, and data were stored on a computer hard drive with Gould hardware and software. Snapwell filters were mounted and studied in a modified Costar horizontal chamber that allowed for the impalement of the cells with a microelectrode from the apical side. Before impalement, both the apical and basolateral surfaces were continuously perfused with warm (37°C), gassed (95% O2-5% CO2) bath solution kept under pressure with 95% O2-5% CO2. The apical membrane potential was allowed to stabilize after impalement, and the cells were sequentially perfused with solution containing amiloride; amiloride and forskolin; and amiloride, forskolin, and CSE. Solution changes were made with an automated valve system (AutoMate Scientific) controlled from outside of the Faraday cage in which the chamber was housed. Vt and Va were recorded from cells maintained under open-circuit conditions. An impalement was considered successful if the electrode remained impaled throughout all conditions tested and the potential recorded from the microelectrode returned to 0 mV on retraction of the microelectrode. By convention of the electrode arrangement, basolateral membrane potential (Vb) was calculated as Va + Vt. A 50-μA, 2-s bipolar transepithelial pulse was passed every 20 s, and Rt and the apical fractional resistance (fRa) were calculated from ΔVt and the ΔVa-to-ΔVt ratio, respectively (5).
Impedance analysis studies were performed as previously described (21, 26, 33). Briefly, Snapwell filters were mounted in modified vertical Ussing chambers as described above. The voltage-measuring and current-passing electrodes were positioned in the chamber, and then the chamber was placed in a jacketed heating manifold at 37°C. The electrodes were then connected to a custom, low-noise, high-bandwidth voltage clamp designed for impedance studies by Dr. W. van Driessche (Laboratory of Physiology, Campres Gasthuisberg, K. U. Leuven, Leuven, Belgium) (33). Two digital signal processing boards (DSP) were connected to the voltage clamp and to a personal computer. One DSP was used to record transepithelial conductance (Gt) and Isc, and one DSP was used to record total capacitance (CT) and impedance spectra while the Isc, Gt, and CT signals were interrupted. HBEC filters were amiloride inhibited and then forskolin stimulated, and Isc was allowed to stabilize. One milliliter of apical bath solution was then removed and replaced with one milliliter of 100% CSE plus amiloride and forskolin. This was considered time 0, and impedance spectra were taken at 1-min intervals thereafter. Apical membrane resistance (Ra), basolateral membrane resistance (Rb), apical membrane capacitance (Ca), and basolateral membrane capacitance (Cb) were calculated with the equations that define the equivalent electrical circuit shown in Fig. 2. Paracellular resistance (Rp) was estimated from the y-intercept of Gt vs. Isc plots and was assumed to remain constant. In support of this assumption, CSE had no effect on the J. Furthermore, CSE had no effect on Rt, which in large measure is determined by Rp.
Mucin secretion assays.
Mucin secretion assays were performed as previously described (15). Briefly, the apical surface of the HBECs was washed with minimal medium to remove mucin accumulated during the differentiation period. One milliliter of CSE (0.16 cigarettes/ml, equal to 20% CSE) or ATP (100 μM in differentiation medium) was added to the apical surface and allowed to incubate for 30 min at 37°C. Supernatants were then removed and analyzed by enzyme-linked lectin assay (ELLA) with a comparison to a standard curve generated with mucin from a healthy smoker according to the methods of Goswami et al. (12). ELLA was performed by coating 96 well plates with Ulex europaeus agglutinin type 1 at 1.25 μg/ml in 0.1 M bicarbonate-carbonate buffer, pH 9.5. The plates were sealed and incubated overnight at 4°C. The next day, the plates were washed three times with 10 mM PBS (pH 7.4)-0.05% Tween 20-0.05% gelatin. Residual binding sites were then blocked with 10 mM PBS (pH 7.4)-0.1% Tween 20. If plates had been stored, they were washed an additional two times before use. Samples or standards (100 μl) diluted in 10 mM PBS (pH 7.4) were added to duplicate wells and incubated for 2 h at room temperature. Plates were washed four times and tapped dry, and then horseradish peroxidase-labeled wheat germ agglutinin (WGA-HRP, 100 μl) was added to each well. Plates were washed six times and tapped dry. Bound WGA-HRP was detected by adding o-phenylenediamine dihydrochloride solution (100 μl) and then stopping the reaction with 4 M H2SO4 (100 μl). Absorbance was read at 490 nm, and mucin concentrations were calculated with SoftMax Pro v3.0 software (Molecular Devices, Sunnyvale, CA). Data are reported as percentage of mucin secreted compared with vehicle control.
Unless otherwise stated, values are means ± SE. Statistical comparisons were made with Prism 4 (GraphPad Software, San Diego, CA). Paired experiments were compared with two-tailed t-tests. All other comparisons were made with one-way ANOVA followed by Dunn's posttest analysis for significance between groups. Significance was defined as a P value ≤0.05.
A total of 143 HBEC filters from two patients were used in the completion of these studies. At baseline, these HBECs had a mean Rt of 755 ± 26.7 Ω·cm2 and a mean Isc of 21 ± 0.5 μA/cm2 (n = 140 for Isc because the 3 filters used for microelectrode impalement studies were not studied under Isc conditions). There was no significant difference between the Rt and Isc of control and CSE-exposed filters when the two groups were compared before the addition of either CSE or stimulants.
In an initial set of experiments, control filters (n = 8) were compared with filters that were exposed to 5%, 10%, 20%, or 40% CSE (n = 6, 5, 7, and 5, respectively). After exposure to CSE, filters were amiloride inhibited and then serially stimulated with UTP and forskolin. The addition of CSE to resting HBECs caused a small, statistically significant decrease in Isc (6–15% for CSE vs. 4% for vehicle) without significantly affecting amiloride-sensitive Isc (P > 0.05 by ANOVA). In contrast, CSE inhibited both purinergically stimulated and cAMP-stimulated Isc in a dose-dependent fashion (Fig. 3). The UTP-stimulated peak Isc was decreased by 22%, 34%, 53%, and 84% by 5%, 10%, 20%, and 40% CSE, respectively, and the UTP plateau Isc was decreased by 64%, 54%, 105%, and 150%, where a decrease of >100% represents Isc falling below prestimulus levels. The forskolin-stimulated peak Isc was increased by 11% by 5% CSE (P > 0.05) but then decreased by 50%, 71%, and 78% by 10%, 20%, and 40% CSE, respectively (P < 0.01). Despite the increase in mean forskolin-stimulated peak Isc with 5% CSE, forskolin-stimulated plateau Isc was decreased by 34%, 72%, 86%, and 88% by 5%, 10%, 20%, and 40% CSE, respectively. Mean values for Isc under the various conditions are provided in Table 1. The use of a final concentration of 20% CSE in the apical bath provided a maximum inhibition of anion secretion without affecting amiloride-sensitive Isc, and therefore we chose to proceed with the use of 20% CSE for the remainder of the protocols.
To further examine the ability of CSE to inhibit cAMP-mediated anion secretion, we performed two additional sets of Isc experiments. In the first set of experiments, we studied the effect of CSE added to amiloride-inhibited, forskolin-stimulated HBECs (n = 6). At baseline HBECs had a mean Isc of 28.0 ± 1.54 μA/cm2, which decreased to 6.7 ± 0.75 μA/cm2 with the addition of amiloride and then increased to 29.3 ± 1.72 μA/cm2 with the addition of forskolin. The addition of CSE to the apical bath reduced Isc in a biphasic manner to a final value of 19.8 ± 1.64 μA/cm2 (P < 0.05), 55% of the bumetanide-sensitive Isc, without changing Rt. A representative tracing of the effect of CSE on the forskolin-stimulated Isc vs. time is shown in Fig. 4.
Next we examined the time and stimulus dependence of CSE-mediated inhibition of anion secretion by adding 20% CSE to the apical bath of HBECs before the addition of amiloride and forskolin. As in the dose-response protocol, the addition of CSE to resting HBECs had no effect on amiloride-sensitive Isc. Forskolin-stimulated peak Isc was decreased by 55%, 39%, 47%, and 61% by 5-, 10-, 15-, and 30-min preincubation with CSE, respectively, and forskolin-stimulated plateau Isc was inhibited by 42%, 32%, 38%, and 50%. These experiments demonstrated that CSE added 5 min before the addition of amiloride and forskolin was able to inhibit forskolin-stimulated anion secretion, but no additional inhibition was gained by longer incubation times (Fig. 5).
Chloride flux studies.
To determine whether CSE affected chloride or bicarbonate secretion, we performed unidirectional ion flux measurements with 36Cl. Cells were mounted in modified Ussing chambers and studied under standard Isc conditions as described above. Ten minutes after addition of amiloride, 1 ml of apical solution was removed and replaced with 1 ml of either new apical solution (n = 11) or 100% CSE (n = 12) containing 5 μM amiloride and 2 μM forskolin. Immediately after application of forskolin, 36Cl was added to either the mucosal (n = 11) or the serosal (n = 12) bath solution. Two 15-min flux periods were measured after the addition of forskolin and 36Cl to the amiloride-inhibited filters. In the absence of CSE, the Isc of −1.01 ± 0.04 μeq·cm−2·h−1 was equal to chloride secretion as measured by a net chloride flux (J) of −1.12 ± 0.21 μeq·cm−2·h−1. Addition of CSE to amiloride-inhibited, forskolin-stimulated HBECs caused a 73% reduction in J(P < 0.05 by t-test), and this effect was due to a decrease in J. Notably, there was only a 50% reduction in Isc despite the 73% inhibition in J. Transepithelial flux data are summarized in Table 2.
Microelectrode impalement studies.
To evaluate the effects of CSE on Va, Vt, and fRa, we performed single-barrel microelectrode impalements (n = 3). By microelectrode impalement, unstimulated HBECs had a mean Va of −34.4 ± 3.20 mV and a mean Vt of −11.9 ± 1.43 mV. The addition of amiloride resulted in a 24.3-mV hyperpolarization of the apical membrane and a 6.5-mV depolarization of Vt. The addition of forskolin after amiloride resulted in a 20.8-mV depolarization of the apical membrane and a 6.1-mV hyperpolarization of Vt, so that after amiloride inhibition and forskolin stimulation Va and Vt were close to baseline values. Mean fRa in amiloride-inhibited, forskolin-stimulated HBECs was 0.29 ± 0.009. Addition of CSE to amiloride-inhibited, forskolin-stimulated HBECs resulted in a gradual hyperpolarization of the apical membrane and a gradual depolarization of Vt that persisted over the length of the impalement. Furthermore, the addition of CSE caused a gradual rise in fRa similar to that observed in the impedance analysis studies described below. Microelectrode impalement data are summarized in Table 3. A tracing from a successful impalement is shown in Fig. 6.
Impedance analysis studies.
To gain insight into CSE-mediated changes in membrane resistances and capacitances, we performed impedance analysis studies (n = 6). HBECs were studied under Isc conditions, and impedance spectra were taken every minute beginning after Isc had reached a stable plateau under amiloride-inhibited, forskolin-stimulated conditions. After the first five spectra had been obtained, 1 ml of apical bath was removed and replaced with 100% CSE containing appropriate concentrations of amiloride and forskolin. Amiloride-inhibited, forskolin-stimulated HBECs had a mean Isc of 19.6 ± 1.92 μA/cm2. The addition of CSE to the apical membrane resulted in a biphasic inhibition of Isc consistent with previous measurements made under standard Isc conditions.
Membrane resistance measurements.
Amiloride-inhibited, forskolin-stimulated HBECs had a mean Ra of 373 ± 16.4 Ω·cm2, a mean Rb of 1,749 ± 128.4 Ω·cm2, and a mean fRa of 0.17 ± 0.009. After the addition of 20% CSE to the apical bath, both Ra and Rb increased rapidly to peaks of 923 ± 73.2 and 3,535 ± 635.5 Ω·cm2, respectively. Similar to the effect seen by microelectrode impalement, the addition of CSE increased fRa by 10% to 0.27 ± 0.016. Ra then gradually decreased over 15 min before leveling off at ∼400 Ω·cm2. Rb paralleled Ra, but Rb continued to fall throughout the remainder of the measurements, reaching a minimum value of 759 ± 52.1 Ω·cm2. Therefore, fRa continued to rise, reaching a maximum value of 0.55 ± 0.050, a change of ∼40% over the amiloride-inhibited, forskolin-stimulated value at 30 min. The changes in apical and basolateral resistances in relation to change in Isc as measured by impedance analysis are shown in Fig. 7.
Membrane capacitance measurements.
Amiloride-inhibited, forskolin-stimulated HBECs had a mean Ca of 3.5 ± 0.08 μF/cm2 and a mean Cb of 22.2 ± 0.68 μF/cm2. Concurrent with the increase in Ra caused by CSE, there was a transient increase in Ca to a maximum of 4.4 ± 0.21 μF/cm2. Conversely, the increase in Rb was associated with a transient decrease in Cb to a minimum of 11.8 ± 0.39 μF/cm2. Both Ca and Cb returned to near the amiloride-inhibited, forskolin-stimulated levels by 20 min after the addition of CSE.
Mucin secretion studies.
To evaluate the possibility that the increase in Ca represented a fusion of mucin granules at the apical membrane, we compared the ability of CSE to stimulate mucin secretion in HBECs with that of differentiation media (vehicle) and 100 μM ATP (positive control). These experiments demonstrated that CSE induced significantly more mucin secretion in HBE cells than vehicle, though less than that stimulated by ATP (Fig. 8).
The studies presented here examined the effects of CSE on active ion transport in HBECs. The results demonstrate that CSE inhibits anion secretion. Initial evidence to support this conclusion was obtained by exposing resting HBECs to CSE. The resting Isc of HBECs reflects the sum of sodium absorption and anion secretion (7). When CSE is added to resting HBECs, there is a small, statistically significant decrease in Isc, which could be either a decrease in cation absorption or a decrease in anion secretion. However, this decrease in Isc did not result in a significant decrease in amiloride-sensitive Isc, and there was a trend toward increased amiloride-sensitive Isc at lower doses of CSE. Together these findings support the hypothesis that the decrease in Isc seen when CSE is added to unstimulated HBECs is due to an inhibition of basal anion secretion, which in unstimulated HBECs can be attributed to net bicarbonate secretion (7). The ability of CSE to inhibit anion secretion was not limited to unstimulated HBECs, such that CSE added to the apical bath solution of amiloride-inhibited, forskolin-stimulated HBECs significantly inhibited Isc in a concentration-dependent fashion. The effect of CSE on forskolin-stimulated Isc appeared to be biphasic.1
There was an initial, rapid decrease in Isc followed by a brief recovery and a sustained, slower inhibition of Isc. Under Isc conditions, Rt did not change significantly over the measurement period. The final change in Isc was measured at the plateau after addition of CSE. According to our impedance analysis experiments, both the apical and basolateral conductances have returned to forskolin-stimulated levels by this time. Therefore, any change in Rt at this point could reflect nonspecific cytotoxicity or damage to tight junctions. In contrast, the addition of bumetanide to filters exposed to CSE after amiloride and forskolin increased mean Rt, although the change did not reach statistical significance.
To test the hypothesis that the inhibition of the forskolin-stimulated Isc was due to effects on chloride secretion, we performed unidirectional chloride flux studies. These studies demonstrated that CSE caused a decrease in the J that was due solely to a decrease in J in amiloride-inhibited, forskolin-stimulated HBECs. The ability of CSE to inhibit forskolin-stimulated Isc without affecting net sodium absorption, J, or Rt suggests that the inhibition of forskolin-stimulated Isc was due to a specific effect on ion transport and not due to loss of epithelial integrity or nonspecific cytotoxicity. Rather, these results demonstrated that CSE had a specific effect on J that resulted in the inhibition of chloride secretion. These data are in agreement with studies performed in canine trachea that showed inhibition of chloride secretion with minimal effects on sodium transport (39). Furthermore, our measurement of Vt during microelectrode impalement experiments demonstrated that CSE caused a mean 4-mV change (−11.6 ± 2.9 mV to −7.71 ± 1.90 mV) in Vt, which is consistent with an inhibition of chloride secretion and is the same change that was recorded when tracheal Vt was measured in vivo in dogs that inhaled a single cigarette (39).
Therefore, the effect of cigarette smoke on ion transport in the airways can be observed in different model systems using different methods for cigarette smoke exposure. In the literature, there is a wide range of techniques for the generation and application of CSE for in vitro use (14, 18, 19, 29, 35). We used a water-soluble extract from filtered cigarettes and acutely exposed HBECs mounted in Ussing chambers. Therefore, our CSE did not contain the volatile fraction of cigarette smoke, which is cytotoxic in alveolar type II cells (14) and not required for effects on ion transport, as demonstrated by Welsh's experiments (39) in which he recapitulated inhibition of chloride secretion by washing the residue from a filter through which a cigarette was smoked. Furthermore, the dose of CSE that we used (0.16 cigarettes/ml = 20%) was determined by a dose-response relationship to provide a robust response within the time frame required for studies in a Ussing chamber. Measurements of lactate dehydrogenase were performed and demonstrated that cytotoxicity was insignificant over a 30-min exposure (data not shown).
The mechanism through which CSE inhibits chloride secretion is incompletely understood, but the appearance of a biphasic inhibition and the complex nature of CSE (3) raise the possibility that CSE affects more than one aspect of the chloride secretory apparatus. The microelectrode studies demonstrated that the baseline Va of HBECs was −34.2 mV, a value similar to, but slightly more hyperpolarized than, that measured in human nasal epithelia (4, 40) and slightly less hyperpolarized than that of Calu-3 cells, a cell line derived from a lung adenocarcinoma (31). Amiloride caused the expected apical membrane hyperpolarization due to inhibition of the apical membrane sodium conductance, and forskolin caused the expected apical membrane depolarization due to activation of an apical membrane anion conductance. CSE caused hyperpolarization of both the apical and basolateral membrane potentials. Apical membrane hyperpolarization may be due to a decrease in apical membrane anion conductance, an increase in basolateral membrane potassium conductance, or both. In this case, if hyperpolarization of Va were due solely to an increased basolateral membrane potassium conductance, one would expect to measure this as a decrease in Rb. However, impedance analysis demonstrated an acute increase in both Ra and Rb, suggesting that the initial hyperpolarization of Va is due to a decrease in apical membrane anion conductance. Therefore, the initial, rapid inhibition of chloride current by CSE appears to be the result of an inhibition of the cAMP-dependent apical membrane anion conductance. This inhibition is likely due to an effect on the CFTR, the dominant cAMP-activated anion channel (13). The ability of CSE to inhibit both UTP- and forskolin-stimulated Isc in HBE cells suggests that there may be a common apical anion conductance stimulated by different chloride secretory agonists, but another plausible explanation is that CSE contains elements capable of inhibiting both a cAMP-stimulated and a purinergically activated apical anion conductance. However, our results do not directly address this question, because our time course experiments demonstrate that cells pretreated with CSE are already in the sustained phase of inhibition, during which the apical and basolateral conductances have been restored. Therefore, the observed effect is most likely due to inhibition of the basolateral entry of chloride on the Na-K-2Cl cotransporter (NKCC)-1.
The initial rapid increase in both Ra and Rb is consistent with the rapid decrease in Isc and an inhibition of CFTR chloride conductance. However, the apical membrane conductance (Ga = 1/Ra) returns to control values after ∼20 min. Despite this reactivation of Ga, chloride secretion remains inhibited. Therefore, an additional mechanism must exist through which chloride secretion is inhibited by CSE. One hypothesis to explain the sustained decrease in Isc is that transient effects on membrane resistance are only responsible for the rapid decrease in Isc and that the sustained decrease in Isc is secondary to inactivation of the NKCC-1. This hypothesis is supported by the observation that simultaneous activation of an apical membrane anion conductance and a basolateral membrane potassium conductance would be expected to increase Isc (8) and the evidence suggesting that chloride uptake via the NKCC-1 is critical in epithelial chloride and fluid secretion (10). Our impedance analysis studies showed that CSE-exposed HBECs demonstrated an acute drop in Cb. At face value, a decrease in membrane capacitance suggests loss of membrane area; however, an alternate explanation is that the decrease in Cb represents closure of the lateral intracellular space because of cell swelling rather than a true loss of basolateral membrane surface area (33), and this explanation would be more consistent with the rather large observed increase in Rb of 102%. Cell swelling is a potent downregulator of the NKCC-1 (20). Further support of this hypothesis comes from our unidirectional ion flux studies, which demonstrated that the inhibition of chloride secretion caused by CSE was due to a decrease in the J, a result that is consistent with inhibition of the NKCC-1. Although the inhibition of J is consistent with inhibition of the NKCC-1, it could also be the result of inhibition of an apical chloride conductance or a basolateral potassium conductance. The former explanation is supported by the results of the impedance analysis, where both the apical and basolateral membrane conductances have returned to forskolin-stimulated levels but chloride secretion is still inhibited. Therefore, our impedance analysis and unidirectional chloride flux studies support the hypothesis that the NKCC-1 is inactivated by CSE as a result of cell swelling and that this inhibition is responsible for the sustained phase of Isc inhibition.
The unidirectional chloride flux studies also demonstrated that the forskolin-stimulated Isc in amiloride-inhibited HBECs is almost completely attributable to chloride secretion, but subject to modification by CSE. CSE inhibited ∼50% of the forskolin-stimulated Isc within 20 min, but flux analysis demonstrated that nearly 80% of J is inhibited. If chloride were the only ion whose conductance was increased by forskolin stimulation of HBECs, one would expect CSE to decrease Isc by the same 80% as J. The discrepancy between Isc and J implies that another ion is actively transported after forskolin stimulation and the addition of CSE. By convention, this Isc must be the net serosal to mucosal movement of an anion or the net mucosal to serosal movement of a cation. Sodium is unlikely to be the responsible ion because in our experiments sodium transport was inhibited by the presence of amiloride in the bath solution. Therefore, the remainder of the Isc is likely to be due to either potassium absorption or bicarbonate secretion. The latter hypothesis is supported by evidence for bicarbonate secretion in response to forskolin stimulation in other HBEC studies (7).
Cigarette smoking results in a variety of changes to the airways including stimulation of airway inflammation (11), induction of goblet cell hyperplasia (24), and stimulation of mucus overproduction (23). Impedance analysis of HBECs demonstrated that Ca increases transiently with the application of CSE. Our data on total mucin secretion demonstrate that exposure of HBECs to CSE results in a nearly twofold increase in mucus secretion compared with vehicle control. These data support the hypothesis that at least part of the increase in Ca reflects an increase in apical membrane surface area due to fusion of mucin-containing secretory granules. Mucus hypersecretion in response to cigarette smoke has been demonstrated experimentally in rats and dogs (6, 16) and is clinically observed in people who smoke, even when asymptomatic (38). Clinically, smokers are observed to have increased mucus production and decreased MCC compared with nonsmokers. However, in vitro studies of the airway mucus produced by smokers have suggested better mucus hydration and lower viscosity as measured by a magnetic microrheometer (23). In addition, in vivo studies of MCC in animal models of acute smoke exposure have shown increased MCC (6), but studies of MCC in animal models of chronic smoke exposure have shown decreased MCC (37). Therefore, the mechanisms through which smoking results in clinically apparent inhibition of MCC remain incompletely characterized. One hypothesis to explain the ability of cigarette smoke-induced reductions in MCC is that in addition to mucus overproduction smoking also results in reduced periciliary fluid volume. The studies presented here and previous studies in canine trachea (39) lend support to the hypothesis that the mucus produced in response to cigarette smoke is inefficiently cleared secondary to a decrease in chloride secretion concomitant with sustained sodium absorption and that together these cause a reduction in the periciliary fluid layer volume and reduced MCC.
This work was funded by the Novartis Institutes for Biomedical Research (A. D. Jackson, P. A. Kemp, H. Danahay), National Institutes of Health Grants F32-HL-74575 (J. L. Kreindler) and R01-DK-54774 (R. J. Bridges), Cystic Fibrosis Foundation Research Development Program, and the University of Pittsburgh.
↵1 In a series of Isc experiments the effect of a 20-min exposure to CSE on amiloride-inhibited, forskolin-stimulated Isc was not reversible after a 20-fold exchange of bath solutions containing appropriate stimuli.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society