Respiratory tissues can be damaged by the exposure of airway epithelial cells to reactive oxygen species that generate oxidative stress. We studied the effects of the hydroxyl radical ·OH, for which there is no natural intra- or extracellular scavenger, on a Ca2+-activated chloride channel (CACC) that participates in Cl− secretion in the apical membrane of airway epithelial cells. We identified and characterized CACC in cell-attached and in inside-out excised membrane patches from the apical membrane of cultured nonciliated human nasal epithelial cells. In these cells, the CACC was outwardly rectified, Ca2+/calmodulin-kinase II, and voltage dependent. The channel was activated in cell-attached and inside-out patches in a bath solution containing millimolar [Ca2+] and ran down quickly. The channel was reversibly or irreversibly activated by exposure of the internal surface of the membrane to ·OH, which depended on the concentration and the duration of exposure to H2O2. CACC activity evoked by oxidative stress was inhibited by 1,3-dimethyl-2-thiurea, an antioxidant that scavenges hydroxyl radicals, and by the reduced form of glutathione. The oxidized SH residues could be close to the Ca2+/calmodulin kinase site. The reversible or irreversible activation of CACC after a period of oxidative stress without change in [Ca2+] is a new observation. CACC play a direct role in mucus production by goblet cells and may thus contribute to the pathogenesis of asthma.
- endogenous chloride channels
- outwardly rectifying chloride channel
- cystic fibrosis transmembrane conductance regulator
- airway epithelial cells
the human respiratory tract is exposed to oxidant stress upon inhalation of atmospheric pollutants and microorganisms. Protection of the airway surface depends on both mucociliary clearance and the production of antioxidants. Atmospheric particulate matter (PM2.5) and diesel exhaust particles generate reactive oxygen species (ROS) in human airway epithelial cells (2–4, 6, 22) and ·OH in the lungs of mice (13). Extracellular and intracellular ROS activate kinases and transcription factors, which can provoke inflammation of the airway epithelium (22).
There is little information concerning the effects of ROS on ion transport in airway epithelia. We have demonstrated that intracellular application of ·OH inhibited an outwardly rectifying Cl− channel (ORCC) and reversibly activated a Ca2+-activated, nonselective cation channel (16, 17). These effects could lead to sustained inflammation of the airway. It is also reported that extracellular oxidant stress stimulated anion secretion from human airway epithelia, which the investigators suggested to result from ROS in the extracellular environment, triggering an immune response (7).
To further evaluate the pathophysiological mechanisms involved in airway inflammation, we have investigated how Cl− channels of human nasal epithelial cells from outgrowth cultures respond to incidents of acute oxidant stress. The apical membrane of airway epithelial cells contains a variety of Cl− channels, including ORCC, the cystic fibrosis transmembrane conductance regulator (CFTR), and the Ca2+-activated Cl− channel (CACC) (10). In this study, we recorded the activity of CACC in the apical membrane of nonciliated epithelial cells. These channels were activated by intracellular [Ca2+] and by phosphorylation of Ca2+/calmodulin (CaM) protein kinase. We show that intracellular hydroxyl radicals directly increase CACC activity. This could be one of the responses of airway epithelial cells to oxidant stress.
Culture of human nasal epithelial cells.
Profs. A. Coste, P. Herman, and J. Soudant generously provided human nasal mucosa from turbinectomies (20 patients) with informed consent and institutional approval (ethical permission was obtained from the Assistance Publique-Hopitaux de Paris). Most of the patients were on medication (corticotherapy and/or antibiotic treatments). There were no ciliated cells in 60% of the human turbinate samples. Nasal epithelial cells were cultured by cell outgrowth with a modified explant cell culture technique (8, 31). Nasal mucosa from human turbinate samples was dissected into small sections (1–2 mm2). Three pieces of mucosa were transferred to each 35-mm-diameter Falcon Primaria culture dish, which had been coated with 200 μl of 1/10 diluted extracellular matrix product (BTI, Biomatrix I; Clinisciences, Montrouge, France). The explants were cultured in DMEM/Ham's F-12 without HEPES (1:1; GIBCO, Cergy-Pontoise, France) with 2% Ultroser G (GIBCO), 50 U/ml penicillin, 50 μg/ml streptomycin, and 50 μg/ml gentamicin (only added to the medium at the beginning of the culture for 2 days) and incubated at 37°C in a humidified atmosphere of 5% CO2 in air. After 5–15 days in culture, explants were surrounded by a coherent outgrowth of epithelial cells with tight junctions resulting from both cell migration and cell proliferation. The periphery of the outgrowth and the proximal part of the explant contained ciliated and nonciliated cells. Ciliated cells seen at the beginning of the culture migrated with the proliferating nonciliated cells. Ciliary beating and presumptive mucous granules could be seen. The number of ciliated cells decreased with the time of culture. This method of cell culture avoided cell passage and enzyme alteration of ion channels and allowed cell-attached or excised inside-out patch-clamp configurations on the apical cell membrane.
To enhance the recording of Cl− channel currents, bath and pipette solutions were nominally free of monovalent cations. The bath solution contained (in mM) 145 N-methyl-d-glucamine (NMDG)-Cl−, 1.4 MgCl2, 1 CaCl2, 10 N-tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid (TES), and 300 mosmol/kgH2O. The patch pipettes were filled with solution containing (in mM) 145 NMDG-Cl−, 2 MgCl2, 2 CaCl2, 10 TES, and 300 mosmol/kgH2O. The low NMDG-Cl− bath solution consisted of (in mM) 15 or 5 NMDG-Cl−, 1.4 MgCl2, 1 CaCl2, 10 TES, and osmolality 300 mosmol/kgH2O adjusted with sucrose. The pH of all solutions was 7.4. Solutions were prepared with Ultrapure Water (Milli-Q System; Millipore, St-Quentin en Yvelines, France). The free Ca2+ concentration in the bath solution was adjusted with Ca2+ and EGTA (Sigma) (19). In recording currents in the cell-attached configuration, the bath solution also sometimes contained 10 μM forskolin (Sigma). Forskolin was dissolved in DMSO, and the final bath concentration of solvent was ≤0.1%. For excised, inside-out patch-clamp recording, bath solutions also contained either 150 U/ml catalytic subunit of cAMP-dependent protein kinase (PKA; Promega) and 200 μM MgATP (Sigma) or no additives. Chemicals employed were of the highest grade of purity available (Merck, Darmstadt, Germany; Prolabo, Fontenay-sous-bois, France; Sigma Aldrich). CaM human brain (CaM), Ca2+/CaM kinase II, rat brain (CaMK), and Ca2+/CaM kinase II inhibitor peptide 281-309 (CaMKI) were obtained from Calbiochem (Merck Biosciences, Nottingham, UK). The reduced form of glutathione was obtained from Sigma.
Ion channel currents were recorded with the patch-clamp technique in cell-attached and inside-out membrane patch configurations performed on the apical membrane of human nasal nonciliated cells from the periphery of the cell outgrowth layer with a high rate of tight junctions. This method avoided enzymatic treatment of the cells. The bath solution surrounding cells or excised membrane patches could be changed with a gravity-driven, multibarrel perfusion system (7 reservoirs) placed within 100 μm of the pipette patch and delivering 34 μl/min. Solution changes were achieved within 10 s by manual switching between reservoirs. Recording pipettes were pulled in two stages from borosilicate glass capillary tubes (GC 150–7.5; Clark Electromedical Instruments, Reading, UK). Pipettes were coated with two layers of Sylgard 184 (Dow Corning Europe, Brussels, Belgium) and fire-polished and had a resistance of 15–20 MΩ when filled with the pipette solution. The reference Ag-AgCl electrode was connected to the bath via an NMDG-Cl− agar bridge. Single-channel currents were recorded with an Axopatch 200B amplifier (Axon Instruments, Dipsi Industrie, Chatillon-sous-Bagneux, France), filtered using a four-pole Bessel filter at 1 kHz and recorded on digital audiotape (DAT DTR, 1205; Biologic, Claix, France). Current recordings were converted using an analog-to-digital interface (DMA 100 OEM, Card Lab Master, Biologic) coupled to a computer running appropriate software (pCLAMP, v.6; Axon Instruments, Foster City, CA). Currents were digitized at 20 kHz. Recording sequences (30 or 60 s or several minutes) were chosen upon replay of DAT cassettes then transferred to storage media (ZIP disks; Iomega, Ropy, UT) or to a printer (Dash IV model XL; Astro-Med, Trappes, France) for long sequences. Experiments were performed at room temperature (21–23°C).
Single-channel data were analyzed using pCLAMP software (v.6). Channels were identified and characterized according to their ionic selectivity with respect to a NMDG-Cl− concentration gradient (15 mM in the bath vs. 145 mM NMDG-Cl− in the pipette) and their single-channel conductance. Unitary current reversal potential and conductance values were estimated from the linear portion of current-voltage (I-V) relationships. Channel amplitude was calculated from Gaussian fits to amplitude/distribution histograms constructed from single-channel recordings. The probability of a channel being open (Po) was measured from 30- to 60-s stable and representative recordings. To calculate Po, digitized single-channel data were subjected to event detection (pCLAMP). Po was calculated as the fraction of the specified recording time spent by the channel in the open state.
Treatment with ·OH.
To test the effects of ·OH on ion channels, the cytoplasmic side of an inside-out membrane patch was superfused simultaneously with two 145 mM NMDG-Cl− bath solutions, one containing 10−5 M Fe(SO4)2(NH4)2 and the other containing 10−2 M H2O2. The mixture of these solutions produces ·OH according to the Fenton reaction: H2O2 + Fe2+ → Fe3+ + ·OH + HO−. The brief burst of ·OH generation was verified by electron spin resonance spectroscopy (17). The local superfusion system placed within 100 μm of the pipette patch delivered 30 μl/min into 1 ml of 145 mM NMDG/Cl− bath solution. This system yielded in 1-min superfusion an estimated final H2O2 concentration of 0.2 mM and Fe2+ concentration of 0.2 μM. During the reaction, all H2O2 was converted to ·OH very rapidly (10−9 s) and disappeared so that for continuous generation of ·OH, both H2O2 and Fe2+ must be supplied continuously. We generated a dose-response relationship between H2O2 (20, 40, 100, 200, 400 μM and 0.5, 1, 2, 5, 8, 10 mM) and one concentration of Fe2+ (10 μM) and tested the effect on channel activity. We also tested hydroxyl radical toxicity on airway cells. The metabolic function of human nasal epithelial cells was determined using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide reduction assay (26). This was not altered in a medium that contained either 0.2–0.5 mM H2O2 alone or this combined together with Fe2+ (10–250 μM) for a 4-h incubation period (n = 16 experiments). Similar experiments were conducted with the human bronchial cell line 16HBE14o−. The cultured human nasal epithelium, which contains many mucous-secreting cells, was protected by the mucus layer against the deleterious effect of ·OH that was observed in 16HBE14o− cells with 0.2 mM H2O2. The exposure of excised membrane patches to ·OH was very short (1–5 min) by comparison.
The superfusion of NMDG-Cl− bath solution containing 10−2 M 1,3-dimethyl-2-thiurea (DMTU) was used as an antioxidant to scavenge hydroxyl radicals created in front of the patch pipette.
Results are reported as means ± SE. Significance was tested at P = 0.05 using the Kruskal-Wallis nonparametric test.
Ca2+/CaM-dependent, protein kinase II-dependent CACC in apical membranes of nonciliated human nasal epithelial cells.
In 98 cell-attached patches (Fig. 1) and 74 inside-out patches (Fig. 2), we recorded channel currents characterized by a strong outward rectification and a high rate of channel activation with depolarization. The unitary I-V relationship obtained from these channels in cell-attached patches in symmetrical NMDG/Cl− solutions was rectified over the range −140 to +140 mV (see Fig. 4, n = 23). The mean conductance of the channels was 5.2 ± 0.5 pS at +120 mV and 2.8 ± 0.3 pS at −100 mV. The reversal potential was at or close to 0 mV in symmetrical 145 mM NMDG/Cl− solutions in 21 out of 23 cell-attached patches. The I-V relationships from cell-attached and inside-out patches in symmetrical 145 mM were similar (see Fig. 4, n = 23 and n = 15, respectively) with the same rectification and the same conductance (5.3 pS at +120 mV and 1.9 pS at −140 mV in inside-out patches). Under asymmetric conditions, when intracellular [NMDG/Cl−] was reduced to 15 mM (Figs. 3A and 4), a negative reversal potential was measured (−52 ± 5 mV). This value is close to the −56 mV predicted by the Goldman-Hodgkin-Katz voltage equation for a perfectly anion-selective channel. In asymmetric Cl− solutions, the single-channel conductance was significantly increased (11 pS at +100 mV, P < 0.05), and the recordings and the magnitude of the channel currents between ±100 mV were improved (Fig. 3A) compared with 145 mM NMDG/Cl− symmetrical conditions. When intracellular [NMDG/Cl−] was reduced to 5 mM, only a negative current was observed between −140 and −60 mV. These experimental conditions favor the recording of anionic channel currents, and separate experiments showed that these channels were more permeant to I− and Br− than Cl− (data not shown) and thus represent an anion-selective channel.
In cell-attached membrane patches, chloride channel activity was not sustained but declined and disappeared within several minutes of the onset of recording. This rundown of channel activity was independent of the type of bathing solution whether this contained forskolin and ATP (n = 57), ATP alone (n = 4), or no additives to the NMDG/Cl− bath solution (n = 41). Activity could be regained with a second application of forskolin following a period of washing (n = 21).
In cell-attached or inside-out patches in symmetrical or asymmetric NMDG/Cl− bath solution, the chloride channel activity was voltage dependent, increasing with depolarization of the membrane (Figs. 1, 2, and 3A). In inside-out patches containing only one channel at hyperpolarized membrane potential under asymmetric Cl− conditions (15 mM NMDG/Cl− bath solutions), Po increased slowly with depolarization from 0.15 ± 0.03 at −100 mV to 0.67 ± 0.05 at +100 mV (n = 3, Fig. 3B). In inside-out patches, although these low-conductance Cl− channels could be activated by membrane depolarization, this effect gradually declined, with rundown being complete 3–5 min after excision.
Figure 5A shows single-channel current recording from an excised inside-out membrane patch containing two channels. At +100 mV, the open channel currents were outwardly directed. Reducing [Ca2+] in the bath solution from 10−3 to 10−7 M decreased Po from 0.56 ± 0.06 to 0.05 ± 0.01 (n = 3, P < 0.05), and channel activity ceased completely at 10−8 M [Ca2+] (Fig. 5B). This effect was rapidly reversible, and recovery of channel activity was obtained by increasing [Ca2+] in the bath from 10−8 to 10−3 M (n = 5). These results suggest that these channels belong to the CACC family.
In some experiments, single-channel current recordings from inside-out patches were held at +140 mV in solutions containing 10−3 M Ca2+ until channel opening disappeared 2–5 min after excision. Then, the addition of CaM alone (5 or 20 μg/ml), ATP alone (500 μM), or CaM and ATP had no effect. But the addition of CaM kinase II (0.08 μg/ml) in the presence of Ca2+, CaM, and ATP led to channel activation (Fig. 6A). These Ca2+/CaM kinase II-activated channels showed outward rectification, and their conductance was similar to that shown in Fig. 2. This effect of CaM kinase II was blocked by the application of 3 nM CaM kinase II inhibitory peptide (281-309) (Fig. 6B). These observations were reproduced in three patches.
In summary, these experiments show that voltage- and Ca2+-activated chloride channels (CACC) with outwardly rectified low conductance could be activated by Ca2+/CaM kinase II in the apical membrane of nonciliated human nasal epithelial cells.
Effects of exposure of the cytoplasmic face of the CACC to ·OH.
H2O2 is converted to short-lived ·OH when mixed with Fe2+ (17). The simultaneous application of H2O2 (10−2 M) and Fe2+ (10−5 M) produced ·OH continuously in front of the patch pipette (for several minutes), and this increased Po to 0.26 ± 0.05 and 0.67 ± 0.06 at +140 mV (Fig. 7, B and C, n = 17). The effect of ·OH was sustained following return to control bath solution that contained either 10−3 M or 10−8 M [Ca2+] (Fig. 7C). The magnitude of the effect of ·OH was similar to that which had been evoked by Ca2+/CaM kinase/ATP. The effects of ·OH on CACC were also tested in solutions containing 10−8 M [Ca2+]. Po increased from 0 to 0.50 after a 1-min exposure to ·OH at a membrane potential of +140 mV (n = 3). The I-V relationship of the ·OH-activated channel was the same as that of the Ca2+-activated channel with equivalent rectification and conductance. Channel activity evoked after brief exposure to ·OH (1 min, 10−5 M Fe2+ and 10−2 M H2O2) was reversibly inhibited by the simultaneous application of the antioxidant DMTU (Fig. 8C). The simultaneous application of Fe2+ (10−5 M) and variable H2O2 concentrations (0.02, 0.04, 0.1, 0.2, 0.4, 0.5, 1, 2, 5, 8, and 10 mM in the superfusion system) produced increasing quantities of hydroxyl radicals and a dose-dependent effect of ·OH on channel activity (Table 1 and Fig. 9). The increase in Po was reversible spontaneously without wash out when we applied from 0.02 to 1 mM H2O2 and 10 μM Fe2+. The dose-response effect of ·OH is shown in Table 1. The effect of ·OH on the CACC activity was high and sustained when we applied 2–8 mM H2O2 (Table 1 and Fig. 9) and decreased upon either the wash out of H2O2 (Fig. 9D) or the addition of 3 nM CaMKI (Fig. 9E). It was inhibited by 1.5 mM GSH (Fig. 9F). Variability in the duration of exposure to ·OH before channel activation, magnitude of Po, and duration of channel activity, probably reflected the different redox status of different cells (Table 1). The effect of the highest concentration of ·OH (10 mM H2O2) on the channel activity was sustained for several minutes and irreversible following return to control bath solution with 10−3 M or 10−8 M [Ca2+] (see Fig. 7C). In this case, reversibility could only be evoked with 1.5 mM GSH (3 patches).
Effects of extracellular forskolin and ·OH on CACC activity in cell-attached patches.
Exposure of intact cells to forskolin (10 μM) elicited, after a short delay (1 min), a marked and reversible increase in channel Po in 21 cell-attached patches tested. The application of ·OH to intact cells also elicited, after a short delay (1–2 min), a reversible increase in channel Po in cell-attached patches (Po = 0.57 ± 0.05, n = 7).
This study shows for the first time that a CACC can be both reversibly and irreversibly activated by exposure to ·OH. The same channel was activated upon exposure of intact cells to ·OH. These observations have pathophysiological implications for human nasal epithelial cells. The reversible and irreversible activation of CACC in excised patches by oxidative stress without change in intracellular [Ca2+] is a new observation. The reversibility depended on the quantity of hydroxyl radicals produced according to the H2O2 concentrations superfused. Because of its low oxidizing potential, H2O2 is not by itself usually reactive enough with organic molecules. Nevertheless, H2O2 has the ability to generate highly reactive ·OH through its interaction with redox-active transitional metals. PM2.5, which is incriminated in respiratory disorders, is characterized by high Fe and Cu contents. The suspension of PM in water (100–1,000 μg/ml) released hydroxyl radicals in the absence or in the presence of H2O2 (3). Therefore, metals adsorbed on PM become bioavailable and generate ·OH with or without H2O2 in ranges comparable with those of this study (3).
Here, we show that applying H2O2 simultaneously with Fe2+ to the cytoplasmic face of excised inside-out membrane patches of human nasal cells produced a dose-dependent increase in Po of CACC. This suggests that ·OH could be targeting exposed cysteine SH residues. The redox condition of these SH groups changes with the addition of GSH. These SH residues could be located near the Ca2+/CaM kinase site on the channel because CaMKI also decreased channel activity.
An effect of ROS on a Cl− current has also been described for transepithelial anion secretion across monolayers of human submucosal gland serous cells (7). They suggest that CFTR is the major anion conduction pathway mediating this response and that this plays an important role in keeping the airways clear from damaging radicals that could potentially initiate tissue destruction. H2O2 has also been described as an essential second messenger mediating the activation of volume-sensitive ORCC in HTC and HeLa cell lines (29). ROS also increase mucus secretion from rodent respiratory epithelial cells through a mechanism involving cyclooxygenase metabolism of arachidonic acid with production of PGF2a (1). Our results suggest that when the intracellular protective mechanisms against oxidants are overloaded in human nonciliated nasal epithelial cells, CACC located on the apical membrane are activated by ROS.
This is the first description of a CACC in nonciliated human nasal epithelial cells. CACC have also been reported in rat lacrimal gland secretory cells (23), in normal and cystic fibrosis human nasal epithelia mounted in Ussing chambers (5), in a distal nephron A6 cell line (24), and in guinea pig hepatocytes (18). In all cases, the low conductance of the channels or current showed strong outward rectification, and channel Po was increased by depolarization and [Ca2+]. We show that this channel was also regulated by Ca2+ acting via a CaM kinase II-dependent mechanism. Cl− channel activation by Ca2+ mediated by multifunctional Ca2+/CaM-dependent protein kinase has also been found in human airway cell lines (30).
There are 12 genes in the CACC gene family (11). The expression of the human (h) CACC2 and hCACC3, but not hCACC1, was demonstrated in human native nasal tissues by RT-PCR (21). On the other hand, the porcine pCACC1 gene, which has 78% amino acid sequence identity with hCACC1, contains a unique A-kinase consensus site on the cytoplasmic loop between the putative transmembrane TM3 and TM4 domains, and expression in an epithelial cell line revealed a CACC that could be activated by cAMP (20). The CACC described in the present study was also activated by cAMP. That cAMP activates basolateral K+ channels, which could generate the driving force for CACC-mediated Cl− secretion in epithelia (21), cannot account for activation of CACC in excised membrane patches.
Notwithstanding that forskolin, PKA, and ATP were added together or separately in the bath solution applied to intact cells and inside-out membrane patches, we failed to record CFTR. Immunohistochemical studies have localized CFTR to the apical domain of ciliated cells in epithelia from human fetal airway, human nasal polyps, and human adult turbinate mucosae (12, 15, 28). It is therefore perhaps not surprising that we have not found CFTR-like ion channels in the apical membrane of nonciliated human nasal epithelial cells. However, it should be noted that extracellular nucleotides can regulate anion transport in airway epithelia (25). This effect is mediated by plasma membrane P2Y receptors, activation of phospholipase C, the generation of d-myo-inositol 1,4,5-trisphosphate, and mobilization of Ca2+ from internal stores. Apical CACC are then stimulated via an increase of intracellular [Ca2+] (9, 14, 27). This phenomenon may have been observed in cell-attached membrane patches in the present study since ATP was present in the bath solution at the beginning of some experiments.
In conclusion, we have shown that CACC were reversibly or irreversibly activated by intracellular exposure to ·OH without change in intracellular [Ca2+]. Reversibility depended on the H2O2 concentration and the duration of exposure to ·OH. This was blocked by the antioxidant DMTU after a brief exposure to ·OH generated by the highest H2O2 concentration. These results suggest a pathophysiological role for ·OH that could lead to sustained inflammation of the airway.
This research was supported by Grants 2001-2004 from Université Paris VII, Denis Diderot, EA DRED 1553, and Caisse d'Assurance Maladie des Professions Libérales de Province, Paris, France.
We thank Prof. Philippe Herman and coworkers, Service d'Otorhinolaryngologie (ORL), CHU Lariboisière, Paris, Prof. André Coste and coworkers, Service ORL and Institut National de la Santé et de la Recherche Médicale, Faculté de Médecine, Paris, and Prof. Jacques Soudant and coworkers, Service ORL, CHU Pitié Salpetrière, Paris, France, for the gift of human turbinate samples. We thank Olja Kacanski for technical assistance. We also thank Dr. Ian Findlay, Centre National de la Recherche Scientifique, Faculté des Sciences, Tours, and Dr. Jacques Teulon, Centre National de la Recherche Scientifique, Centre de Recherche Biomedicale des Cordeliers, Paris, for help, suggestions, and fruitful discussion of the manuscript.
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