Effective mucociliary clearance (MCC) depends in part on adequate airway surface liquid (ASL) volume to maintain an appropriate periciliary fluid height that allows normal ciliary activity. Apically expressed large-conductance, Ca2+-activated, and voltage-dependent K+ (BK) channels provide an electrochemical gradient for Cl− secretion and thus play an important role for adequate airway hydration. Here we show that IFN-γ decreases ATP-mediated apical BK activation in normal human airway epithelial cells cultured at the air-liquid interface. IFN-γ decreased mRNA levels of KCNMA1 but did not affect total protein levels. Because IFN-γ upregulates dual oxidase (DUOX)2 and therefore H2O2 production, we hypothesized that BK inactivation could be mediated by BK oxidation. However, DUOX2 knockdown did not affect the IFN-γ effect on BK activity. IFN-γ changed mRNA levels of the BK β-modulatory proteins KCNMB2 (increased) and KCNMB4 (decreased) as well as leucine-rich repeat-containing protein (LRRC)26 (decreased). Mallotoxin, a BK opener only in the absence of LRRC26, showed that BK channels lost their association with LRRC26 after IFN-γ treatment. Finally, IFN-γ caused a decrease in ciliary beating frequency that was immediately rescued by apical fluid addition, suggesting that it was due to ASL volume depletion. These data were confirmed with direct ASL measurements using meniscus scanning. Overexpression of KCNMA1, the pore-forming subunit of BK, overcame the reduction of ASL volume induced by IFN-γ. Key experiments were repeated in cystic fibrosis cells and showed the same results. Therefore, IFN-γ induces mucociliary dysfunction through BK inactivation.
- BK channel
- ciliary activity
- airway surface liquid
- dual oxidase
ifn-γ is a pleiotropic cytokine that not only plays a major role in immune defense, but also contributes to inflammation. IFN-γ is involved in the pathogenesis of asthma (16–18), is increased in airway epithelia of patients with chronic obstructive pulmonary disease (29, 42), and, if overexpressed in murine lungs, results in emphysema (44).
Mucociliary clearance (MCC) is an important component of the innate defense system of conducting airways. Appropriate ion transport is critical for hydration of the airway and is therefore essential for MCC. The balance between Cl− secretion and Na+ absorption has proven to be important to keep an adequate water supply in the airway lumen (6, 38, 45, 46). Apical Cl− secretion occurs through the cystic fibrosis transmembrane conductance regulator (CFTR) and Ca2+-dependent Cl− channels, whereas Na+ absorption depends mainly on epithelial Na+ channels (ENaC). IFN-γ downregulates CFTR and decreases amiloride-dependent Na+ transport but seems to increase Ca2+-activated Cl− channel currents when measured in Ussing chambers (10). IFN-γ levels might be low in CF (24), but a clinical trial administering IFN-γ to patients with CF did not show clinical benefits (25).
We have recently shown that large-conductance, Ca2+-activated, and voltage-dependent K+ (BK) channels are important for adequate water supply to the apical surface (15, 22). The pore-forming structure of the BK channel is a homotetramer of the α-subunit (KCNMA1, Slo1, Ca1.1), which is encoded by a single gene (Slo1, KCNMA1). Each monomer can be bound to one of several β-modulatory proteins (KCNMB1-B4) or to leucine-rich repeat-containing proteins (LRRCs), which have now been accepted as a new family of γ-regulatory proteins of BK (47). In addition, BK activity is subjected to multiple regulation mechanisms at different levels, from mRNA splicing to posttranslational oxidation (36, 37, 52).
In normal human bronchial epithelial (NHBE) cells, apical BK channels open in response to ATP (22), an important paracellular signal (20, 39, 40). By secreting K+, apical BK channels provide a driving force for apical Cl− secretion. Simultaneous apical secretion of both Cl− and K+ will result in the transfer of ions and water directly to the airway lumen. This BK function is critical because inhibition of BK channels reduces airway surface liquid (ASL) volume significantly (22). Thus apical K+ secretion is an important contributor to the regulation of water movement into the airway lumen.
Our group and others have shown that dual oxidase 1 and 2 (DUOX1 and DUOX2) are the main H2O2-generating enzymes in NHBE cells (7, 12, 33). Production of H2O2 is required as part of the lactoperoxidase defense system against bacteria and viruses to oxidize thiocyanate to hypothiocyanite. However, H2O2 is also related to inflammatory processes and is increased in smokers (27). IFN-γ (1 to 100 ng/ml) upregulates DUOX2, thereby increasing basal as well as ATP-stimulated H2O2 production (11, 13).
Here we examined the effect of IFN-γ on BK activity. We found that IFN-γ decreases apical BK activity by a mechanism independent of DUOX2-mediated H2O2 production. With a reduction of apical BK activity, IFN-γ caused a decrease in ciliary beating frequency (CBF) that was immediately rescued by apical fluid addition, suggesting that it was attributable to ASL volume depletion. This was confirmed by ASL meniscus-scanning measurements. Overexpression of KCNMA1 overcame the effect of IFN-γ on CBF. IFN-γ also decreased MCC in CF cells.
MATERIALS AND METHODS
Chemicals and solutions.
All media and Hanks' balanced salt solution were purchased from Gibco, Life Technologies (Grand Island, NY). Recombinant human IFN-γ with an ED50 in the range of 0.5–3.0 ng/ml and specific activity in the range of 3.3 × 105 to 2 × 106 U/mg was obtained from Invitrogen, Life Technologies (Grand Island, NY). This specific activity translates into concentrations of ∼11.6 U/ml for the 10 ng/ml dose used in most of our experiments. Unless stated otherwise, all other materials were obtained from Sigma Aldrich (St. Louis, MO).
Air-liquid interface cell culture.
CF lungs homozygous for ΔF508 were donated by transplant recipients and collected according to protocols approved by the local Institutional Review Board. Normal human airways were obtained from organ donors whose lungs were rejected for transplant. IRB-approved consent for research with these tissues was obtained by the Life Alliance Organ Recovery Agency of the University of Miami and conformed to the declaration of Helsinki. Airway epithelial cells were isolated and dedifferentiated through expansion. Passage 1 cells were redifferentiated at an air-liquid interface (ALI) on collagen-coated, 24-mm Transwell-clear, 6.5-mm or 12-mm Snapwell filters (Costar Corning, Corning, NY) for about 4 wk (at which time cultures were ciliated and secreted mucus) as previously described (8).
Total RNA was extracted using an RNeasy Plus Mini Kit (Qiagen, Valencia, CA) and reverse transcribed into cDNA using the iScript cDNA synthesis kit (Bio-Rad, Hercules, CA) according to the manufacturer's instructions. Quantitative real-time PCR (qPCR) was performed using a TaqMan Universal PCR Master Mix (Applied Biosystems, Foster City, CA) with the TaqMan Gene Expression Assays (Applied Biosystems) indicated in Table 1. Each sample was analyzed in triplicate. The difference in the threshold cycle between the targeted gene and GAPDH (ΔCt) was used as the relative level of expression. All data were expressed as 1,000× mRNA/GAPDH mRNA.
Western blot and biotinylation experiments.
KCNMA1 protein expression was determined as described previously (22). ALI-cultured cells were dissolved with 90°C preheated 2% SDS 50 mM Tris, 5 mM EDTA, pH 8.3 buffer containing antiproteases (ThermoScientific, Rockford, IL) and sonicated, and protein was quantified by BCA assay (Pierce Biotechnology, Rockford, IL). DTT-reduced samples (20 μg of protein) were loaded onto 7.5% Bio-Rad polyacrylamide gels using a 25 mM Tris, 192 mM glycine, 0.4% SDS, pH 8.3 running buffer and transferred overnight to Immobilon-P membranes (Millipore, Billerica, MA). Membranes were blocked with 5% nonfat milk in Tris-buffered saline, pH 7.4, 0.05% Tween-20 for 1 h. For detection of the BK α-subunit, blots were incubated with 2 μg/ml of a rabbit anti-human BKCa (α-subunit) antibody (Sigma-Aldrich cat. no. P4872) for 1.5 h at room temperature and developed with horseradish peroxidase-conjugated anti-rabbit IgG antibodies (KPL, Gaithersburg, MD) and LumiGLO (KPL). To normalize for protein loading, membranes were reprobed with rabbit anti-β-actin (1:100) (Sigma-Aldrich) after stripping with Restore Western Blot Stripping Buffer (Pierce Biotechnology). Band intensities were quantified in the linear range using a Bio-Rad Chemidoc XRS.
For the overexpression experiments, protein was extracted with 150 mM NaCl, 50 mM HEPES, pH 7.5, 1.5 mM MgCl2, 10 mM sodium pyrophosphate, 20 mM NaF, 1 mM EDTA, 5 mM EGTA, 10% (vol/vol) glycerol, and 1% Triton X-100 (lysis buffer) in the presence of complete protease inhibitor mixture (Roche Applied Science, Indianapolis, IN). Samples were sonicated and centrifuged at 10,000 g for 2 min to eliminate debris. DTT-reduced samples in Laemmli sample buffer were heated to less than 60°C and loaded onto the gel. HA-antibodies or the monoclonal anti-BK α-antibody L6/60 (Neurolab; Antibodies, Davis, CA) were used for blotting.
Surface biotinylation was accomplished by using a Pierce Cell Surface Protein Isolation Kit (Thermo Fisher Scientific, Waltham, MA). Briefly, cells cultured on 24-mm filters were surface labeled on ice with 1 mg/ml EZ-Link sulfo-N-hydroxysuccinimidebiotin in Dulbecco's PBS (DPBS) with gentle shaking. Reaction was stopped by adding 50 μl quenching solution, washed with TBS, and lysed with the lysis buffer, as indicated before, in the presence of a complete protease inhibitor mixture (Roche Applied Science). Biotinylated proteins were captured with a streptavidin matrix and then washed and released by using sample buffer (Bio-Rad).
Ussing chamber experiments.
Fully differentiated NHBE cells grown on Snapwell filters (1.13 cm2) were mounted in Ussing chambers (EasyMount Chamber) connected to a VCC MC6 voltage clamp unit (both from Physiologic Instruments, San Diego, CA). Solutions were maintained at 37°C by heated water jackets. To monitor short-circuit current (Isc), transepithelial membrane potential was clamped to 0 mV. Transepithelial resistance was calculated according to Ohm's law by determining the change of current induced by 1-mV pulse applied at 1-min intervals. Only cultures with >200 Ω/cm2 were used for the experiments. Signals were digitized and recorded with DAQplot software (VVI Software, College Station, PA) via a LabJack A/D converter (LabJack, Lakewood, CO). All experiments were conducted with culture and date-matched filters.
Transepithelial currents in response to amiloride, forskolin, or ATP were measured on symmetrical Krebs-Henseleit solutions (118 mM NaCl, 25 mM NaHCO3, 4.7 mM KCl, 1.2 mM MgSO4, 1.2 mM NaH2PO4, 1.2 mM CaCl2, 5.5 M glucose, pH 7.35 bubbled with 95% O2-5% CO2).
Apical K+ currents were measured 30 min after basolateral permeabilization using 20 μM amphotericin B, 10 μM valinomycin, and 10 μM nigericin and in the presence of apically applied 10 μM amiloride. A K+ gradient was created with a basolateral solution containing: 142 mM K+ gluconate, 1 mM MgSO4, 4 mM Ca2+ gluconate, 10 mM glucose, and 10 mM HEPES, pH 7.4 and a similar apical solution in which 137 mM K+ gluconate was replaced by Na+ gluconate. The protocol was described by Namkung et al. (26) and used by us previously (22).
pLKO.1-puro plasmids (pLKO) or those containing a DUOX2 shRNA (TRCN0000045963) were used (11, 31). Viral yield was quantified using a p24 ELISA assay (Perkin Elmer, Boston, MA). Nondifferentiated NHBE cells were infected with shRNA-encoding lentiviruses, selected with 0.5–1.0 μg/ml puromycin and then redifferentiated at the ALI. Control cultures were grown in identical conditions on different filters simultaneously, infected with vector only. When beating cilia became apparent, cells were used for qPCR, Western blots, Ussing chamber experiments, and CBF measurements.
Overexpression of KCNMA1 in NHBE cells.
Overexpression of hSlo1 was accomplished by infection of nondifferentiated NHBE cells using pCDH-EF1-MCS-T2A-Puro lentiviruses (SBI, Mountain View, CA) encoding a human Slo1 sequence kindly provided by Dr. Hoshi (University of Pennsylvania) (U11058) (52) or empty virus. In some experiments, we also overexpressed the double mutant hSlo C911A: C430A, resistant to H2O2 (52), also provided by Dr. Hoshi. Infected NHBE cells were selected with puromycin and grown until differentiation (∼20 days). When beating cilia became apparent, cells were used for qPCR, Western blots, Ussing chamber experiments, and CBF measurements.
Measurement of CBF.
For these experiments, regular washing of the apical surface with DPBS was suspended for the duration of the experiments. Cells were imaged on a Zeiss Axiovert 200, equipped with differential interference contrast optics. CBF was measured as previously described (30). All measurements were done within a 0.5-cm radius from the center of each 24-mm filters (no influence of possible fluid menisci at the edges of the cultures). Because adequate ASL volume maintains an appropriate periciliary fluid height that allows for normal ciliary activity, we measured CBF before and after apical fluid addition. CBF reduction that is fully rescued upon fluid addition to the apical compartment indicates ASL volume depletion.
Measurement of ASL by meniscus refraction scanning.
To confirm the indirect assessment of ASL with CBF, cells grown on 6.5-mm Snapwell filters (Costar Corning) were imaged with an Epson flatbed scanner and images analyzed with a custom algorithm (14) developed in ImageJ (NIH, Bethesda, MD).
GraphPad Prism version 5.0b for Mac OS X (GraphPad Software, San Diego, CA) was used to analyze all results. Student's t-test or Mann-Whitney tests were used to compare two groups as appropriate. One-way ANOVA was used for comparisons of more than two groups followed by a posttest for linear trend analysis when different concentrations of IFN-γ or H2O2 were used. P < 0.05 was accepted as significant. All data are presented as means ± SE.
IFN-γ effect on transepithelial ion transport.
We first examined the effect of IFN-γ on total transepithelial ion transport in fully differentiated NHBE cells grown at the ALI (Fig. 1). Data were obtained from at least four different lung donors. Basal Isc was not significantly different after 48-h treatment with 10 ng/ml IFN-γ (6.64 ± 0.323 μA/cm2, n = 6 for IFN-γ, vs. 5.30 ± 0.542 μA/cm2, n = 6 for control; P > 0.05). Resistance was also not different (496 ± 75.4 Ω/cm2, n = 8 for IFN-γ, vs. 622 ± 36.2 Ω/cm2, n = 7 for control; P > 0.05). As shown in Fig. 1A, the addition of amiloride decreased Isc, indicating the presence of amiloride-sensitive Na+ channels. This current was reduced by IFN-γ treatment. Response to forskolin, indicative of cAMP-mediated CFTR activity, was also significantly reduced by IFN-γ treatment (Fig. 1B). ATP triggered a dual-phase response in our cells (a representative trace is shown in Fig. 1C, left), a shorter response with a maximum reached within 100 s after ATP addition (Fig. 1C, middle), and a second maximum embedded in a more prolonged response (Fig. 1C, right). IFN-γ-treated cells responded to ATP with a higher increase of Isc, which is thought to be at least in part due to Cl− release through Ca2+-activated Cl− channels.
These results mirror the effect of IFN-γ on ion transport in NHBE cells previously described in not fully differentiated cells (10). The decrease of amiloride-dependent currents and the increase of ATP-mediated transepithelial currents should favor apical hydration. On the other hand, the decrease of cAMP-dependent currents (CFTR) favors the opposite. We previously demonstrated that apical BK activity is another important contributor to the balance between absorption and secretion of water in the airway epithelium (22). We therefore explored the effect of IFN-γ on apical, ATP-induced K+ secretion via BK.
Effect of IFN-γ on BK expression and apical, ATP-induced K+ secretion.
To evaluate the effect of IFN-γ on BK activity, we performed experiments in Ussing chambers with basolateral permeabilization, a basolateral-to-apical K+ chemical gradient, by substituting gluconate for Cl− and amiloride addition to inhibit Na+ absorption (22, 26). In these conditions, apical ATP generates a current that corresponds to K+ secretion. This current is decreased by the BK inhibitors paxilline and iberiotoxin and in cells in which BK function was suppressed using shRNA (22).
We found that IFN-γ decreased BK activity in response to ATP (Fig. 2A) and that this effect is dose dependent from 1 to 100 ng/ml of IFN-γ (Fig. 2B). Real-time quantification showed a statistically significant decrease of KCNMA1 mRNA levels in IFN-γ-treated samples (Fig. 2C; P < 0.05). However, total protein expression normalized to β-actin was not significantly changed (n = 3, Fig. 2D). Thus, although IFN-γ decreased mRNA levels of KCNMA1, it did not affect total protein levels, suggesting that posttranslational regulatory mechanisms or changes in regulatory proteins must be involved in the functional reduction of BK channel activity.
Effect of IFN-γ on CBF and ASL volume.
It has been established that IFN-γ exerts important effects on ion transport in NHBE cells. However, the overall impact of these changes on MCC remains unclear. We previously showed that impairment of BK activity results in ASL volume depletion (22). Figure 3 shows that IFN-γ significantly decreases CBF in NHBE cells cultured at the ALI compared with control cells when measured after 48 h of IFN-γ treatment. The CBF decrease was immediately rescued by the addition of apical fluid, indicating that CBF was low because of ASL volume depletion and not because of an intrinsic problem with cilia after IFN-γ treatment. The IFN-γ effect is dose dependent (Fig. 3A). These findings were confirmed by direct ASL volume estimations using optical meniscus scanning (Fig. 3B). These data indicate that the overall effect of IFN-γ on airway homeostasis is ASL volume depletion.
Effect of KCNMA1 overexpression.
To assess the relevance of BK impairment in IFN-γ-mediated ASL depletion, we overexpressed KCNMA1 (Slo-1) protein tagged with HA in NHBE cells. We reasoned that even a significantly reduced BK function after IFN-γ treatment could be overcome by the combined activity of many BK channels, no matter the mechanism of reduced BK channel function. As expected, KCNMA1 mRNA levels (1,000× relative to GAPDH mRNA) were augmented (P < 0.0001) in overexpressing cells (4.04 ± 0.558, n = 6) compared with cells infected with empty vector (0.26 ± 0.027, n = 6). IFN-γ treatment did not decrease but increased KCNMA1 mRNA in overexpressing cells (5.61 ± 0.423, n = 6 vs. 4.04 ± 0.558, n = 6 in nontreated, P < 0.05), whereas the vector-infected cells responded with a decrease in KCNMA1 mRNA (0.10 ± 0.007, n = 6 vs. 0.26 ± 0.027, n = 6, P < 0.001). Overexpression of KCNMA1 protein (Slo-1) was confirmed by Western blot using HA antibodies (not shown) or by the L6/60 antibody (Fig. 4A). Total KCNMA1 protein expression was not changed by IFN-γ treatment of either vector or overexpressing cells (Fig. 4A, P > 0.05 for 3 independent experiments). KCNMA1 overexpression was accompanied by enhanced K+ currents in response to ATP stimulation. IFN-γ decreased BK activity in both vector and overexpressing cells by a similar fraction (Fig. 4B); however, overall K+ current in overexpressing cells was still high after IFN-γ treatment. Concurrently, 48 h of IFN-γ treatment did not decrease CBF in overexpressing cells (Fig. 4C), indicating that the presence of high BK activity through KCNMA1 overexpression can counteract the functional effects of IFN-γ on MCC.
Effect of DUOX2 knockdown on IFN-γ-induced impairment of BK activity.
Because IFN-γ upregulates DUOX2 expression and activity (11) and because H2O2 has been reported to oxidize cysteines and reduce BK activity (52), we measured the effect of H2O2 in NHBE and in KCNMA1-overexpressing cells (Fig. 5).
Apically applied H2O2 (500 μM) caused a slight but significant decrease of baseline Isc that was reversible by 200 nM paxilline (not shown) and more pronounced in KCNMA1 overexpressing cells (2.0 ± 0.436 μA/cm2, n = 3) than in vector cells (0.66 ± 0.258 μA/cm2, n = 4), consistent with apical BK channel opening. To explore the effect of H2O2 on ATP-induced currents, ATP was added after 3–4 min of apical H2O2 (100–500 μM). The effect of H2O2 on ATP-induced BK currents in vector cells was dose dependent; 100 μM H2O2 caused an increase, whereas higher doses caused a decreased response (Fig. 5A). However, in overexpressing cells, 100 μM H2O2 did not stimulate ATP-induced BK currents, and the currents showed a lower sensitivity to H2O2 compared with vector-only-infected NHBE cells (Fig. 5A).
The effect of DUOX2 knockdown on apical BK activity was next examined. Specific DUOX2 shRNA was used to decrease DUOX2 mRNA levels and H2O2 release as described before (11). Knockdown was successful in diminishing basal levels of DUOX2 mRNA and DUOX2 mRNA upregulation by IFN-γ (Fig. 6A). DUOX2 knockdown was specific because basal DUOX1 expression was not affected (Fig. 6B). Interestingly, IFN-γ decreased DUOX1 in DUOX2 knockdown cells (Fig. 6B). DUOX2 knockdown influenced neither KCNMA1 mRNA at baseline nor its decrease during IFN-γ exposure (Fig. 6C). Finally, we found that DUOX2 knockdown did not increase BK activity at baseline and did not rescue the decrease in BK activity upon IFN-γ exposure (Fig. 6D). Therefore, DUOX2-mediated H2O2 production is not responsible for IFN-γ-induced decreases in BK channel activity. The results in Fig. 6D actually suggest that some DUOX2-mediated H2O2 production could be necessary for normal BK activity.
Although these results indicate that DUOX2 upregulation by IFN-γ was not the cause of decrease of BK activity, they do not exclude the possibility that other sources of H2O2 could oxidize BK. We therefore expressed the double mutant hSlo C911A:C430A of the BK α-subunit in NHBE cells (37, 52). Both mutant and wild-type proteins were detectable in cells (Fig. 7A) and equally able to reach the surface (Fig. 7B). However, the expression of the H2O2-resistant mutant did not significantly change the impairment of BK activity by IFN-γ: all vector, wild-type overexpressed, or mutant-expressing cell cultures were susceptible to the IFN-γ effect on ATP-induced activity (Fig. 7C). These results corroborate the idea that neither DUOX2 nor other H2O2 sources are responsible for the IFN-γ effect on ATP-induced BK activity.
Effect of IFN-γ on BK β-modulatory proteins.
The results indicated that the effect of IFN-γ on BK activity is mediated through posttranslational modifications. Therefore, modulatory proteins might play a role, and we measured the effect of IFN-γ on mRNA levels of the β-modulatory proteins KCNMB1-B4 (Fig. 8). KCNMB1 was not detected in the samples, and KCNMB3 did not show a statistically significant difference for any analysis. However, IFN-γ consistently changed the pattern of expression of KCNMB2 (increase, Fig. 8A) and KCNMB4 (decrease, Fig. 8B) mRNA. The pattern was not affected by KCNMA1 overexpression. β-Modulatory proteins have complex effects on BK activity. KCNMB2 has a negative effect on the activity of BK channels (41, 43, 50). In contrast, KCNMB4 increases BK sensitivity at low Ca2+ concentrations (2). Interestingly, both KCNMB2 (50) and KCNMB4 (34) reduce apical BK localization in other tissues. Therefore, we measured surface expression of the pore-forming KCNMA1 in overexpressing cells in the presence or absence of IFN-γ treatment and did not find decreases in the levels of KCNMA1 captured by surface biotinylation (Fig. 8C), a result that suggests that surface availability of the channel has not been compromised.
Effect of IFN-γ on the γ-modulatory protein LRRC26.
We also measured mRNA levels of LRRC26, which has been named a BK γ-modulatory subunit in addition to the already known evolutionarily related β-modulatory proteins; LRRC26 shifts the voltage necessary for BK activation to less negative values and reduces the Ca2+ requirement for opening the channel (48). In these conditions, BK channels can be activated at more physiological voltage and calcium levels encountered in nonexcitatory cells. LRRC26 also changes the sensitivity of BK channels to pharmacological openers. In particular, it decreases the sensitivity of the channel to the BK opener mallotoxin (49). Thus the presence of LRRC26 can be identified by a low response of BK to mallotoxin (1). We found that IFN-γ reduced LRRC26 mRNA expression (Fig. 9A) while increasing the sensitivity to mallotoxin (Fig. 9B). On the other hand, both control and IFN-γ-treated cells had a similar response to NS11021, another known BK opener (3) (Fig. 9C). This last result also indicates that similar amounts of protein are at the apical surface in native NHBE cells. Overall, the data suggest that IFN-γ-induced LRRC26 downregulation contributes to decreased BK activity in NHBE cells.
Effect of IFN-γ on CF bronchial epithelial cell.
Finally, we addressed the effect of IFN-γ on CF bronchial epithelial (CFBE) cells. CFBE cells constitute a natural model to assess the effect of IFN-γ in the absence of functional CFTR. IFN-γ decreased forskolin-induced responses in NHBE cells. By using CFBE cells, we addressed the effects of IFN-γ on other ion channels, particularly BK, because it has been postulated that ENaC activity is decreased by IFN-γ in favor of hydration (Fig. 1A and Ref. 10).
After 30 h of IFN-γ treatment (10 ng/ml), BK currents in response to ATP were inhibited 60% comparable to normal cells (n = 6 from 3 different ΔF508 homozygous donors). Mallotoxin increased BK currents in IFN-γ-treated cells but not in controls, and NS11021 opened BK under both circumstances, similar to NHBE cells. CBF was consistently reduced (n = 5 from 3 different ΔF508 homozygous donors), and CBF was restored by apical fluid addition (Fig. 10A). These results were confirmed by direct ASL volume estimations using optical meniscus scanning (Fig. 10B). Even though IFN-γ has been described to increase Ca2+-dependent Cl− currents in CFBE cells to a higher level than in NHBE cells (10), these data suggest that increased Ca2+-dependent Cl− currents seen in Ussing chambers do not offset IFN-γ-mediated depressed BK activity on ASL volume. In fact, the effect was even more pronounced in CFBE than in NHBE cells (compare Fig. 10 with Fig. 3).
The data presented here suggest that the observed IFN-γ-mediated decrease in MCC is attributable to decreased apical BK activity and possibly mediated by regulation of BK accessory proteins but not by increased H2O2 production attributable to upregulation of DUOX2 (Fig. 11). Galietta et al. (10) previously explored the effect of IFN-γ on airway epithelial cell ion transport. They found that IFN-γ decreased cAMP-dependent Isc, reduced amiloride-dependent Na+ absorption, and increased Isc responses to UTP. Our results corroborate these findings (Fig. 1). Even though these data suggest that CFTR dysfunction upon IFN-γ stimulation should be compensated for by increased Cl− secretion through Ca2+-dependent Cl− channels (9, 10), our physiological experiments do not support this hypothesis.
In addition, there are important differences in experimental design between Galietta et al. (10) and our studies. First, the IFN-γ concentration used by us was 100-fold lower for most experiments and closer to what can be expected locally under stress conditions. We usually used 10 ng/ml, which according to the manufacturer was equivalent to ∼11.6 U/ml vs. 1,000 U/ml used in the previous study (10). Our high concentration (100 ng/ml) would be equivalent to ∼116 U/ml. Although levels of pg/ml have been reported in bronchoalveolar lavage fluid and serum of healthy individuals (19), whole blood IFN-γ in response to stimulation with specific Mycobacterium tuberculosis antigens can rise to 25 ng/ml in vitro (32). Second, Galietta et al. (10) conducted studies that were in mostly undifferentiated NHBE cells, only 2 wk after establishing ALI conditions in cells that have been passaged between three to six times (10). Expression and function of BK channels under these conditions are unknown. On the other hand, our studies were conducted with fully differentiated epithelial cells (4 wk in ALI culture), which more closely represent the native airway epithelium.
The overall effect of IFN-γ on ASL is probably the sum of its influence on Na+, Cl−, and apical K+ transport. Our data support the hypothesis that the IFN-γ-mediated decrease in BK activity is at least partially responsible for mucociliary dysfunction because overexpression of KCNMA1 (Slo1) prevents the decrease of CBF and therefore MCC caused by IFN-γ.
Our results also indicate that the impairment of BK activity is caused by events unrelated to KCNMA1 transcription and translation because changes in expression of the α-unit of the BK protein were not observed. Because DUOX2 expression increases in response to IFN-γ and our own results (Fig. 5C) indicated that high concentrations of H2O2 decrease ATP-induced BK activity, we tested the hypothesis that BK activity was impaired by increased oxidative stress through DUOX2-mediated H2O2 production. We found that DUOX2 knockdown did not prevent IFN-γ-induced inactivation of the BK response to ATP. Our data also reject the possibility that other sources of H2O2 were responsible for the decrease of BK activity because the ATP-induced response of a KCNMA1 H2O2-resistant mutant was also affected to a similar degree by IFN-γ. On the contrary, the overall results suggested that small amounts of H2O2 favor BK opening because DUOX2 knockdown decreased the BK response to ATP. Indeed, much of the data that support BK inactivation by H2O2 were obtained from patch-clamp experiments with the cytosolic compartment directly exposed to H2O2 (35). These experiments included patch excision and consequent rundown (52). In contrast, low concentrations of H2O2 (50 μM) were able to increase BK currents in cell-attached patches, a process mediated by H2O2 signaling (51).
The present results, however, do not exclude the possibility that other oxidative or nitration processes contribute to BK inactivation. For example, IFN-γ has been shown to increase nitric oxide production in differentiated NHBE cells (28). Although nitric oxide by itself has been reported to increase BK activity (4, 23), peroxynitrite produced by the reaction of nitric oxide with reactive oxygen species has been reported to impair BK activity (5, 21).
We also considered the effect of the modulatory subunits on BK activity. IFN-γ increases mRNA expression of KCNMB2 and decreases KCNMB4 and LRRC26 mRNA. The functional data evaluating the response to the BK opener mallotoxin support an effect of LRRC26 downregulation on BK activity. Although our results discard the hypothesis that KCNMB2 and KCNMB4 changes in mRNA expression translate into reduced apical expression of BK, we could not rule out the possibility of KCNMB2 directly decreasing BK activity. Further investigations are needed to fully assess the relevance of these changes on the functional downregulation of BK on IFN-γ exposure.
In conclusion, two lines of evidence support the importance of BK activity on MCC. First, KCNMA1 overexpression in NHBE cells overcomes the effect of IFN-γ on MCC. Second, CFBE cells that lack functional CFTR respond with a decrease in ASL volume and MCC to the presence of IFN-γ. Therefore, at least part of the overall MCC response to IFN-γ occurs through BK inactivation. Our results show the importance of BK on ion transport and illustrate how mediators of inflammation can impair BK activity.
This work was supported in part by grants from the NIH (HL-60644 and HL-89399 to M. Salathe) and FAMRI (CIA 103027 to M. Salathe and CIA 123060 to G. E. Conner).
No conflicts of interest related to this work, financial or otherwise, are declared by the authors.
Author contributions: D.M., M. Srinivasan, G.E.C., and M. Salathe conception and design of research; D.M., M. Srinivasan, S.T.S., P.I., N.B., J.S.D., and M. Salathe performed experiments; D.M., M. Srinivasan, S.T.S., P.I., N.B., J.S.D., and M. Salathe analyzed data; D.M., S.T.S., P.I., N.B., J.S.D., G.E.C., and M. Salathe interpreted results of experiments; D.M. and S.T.S. prepared figures; D.M. drafted manuscript; D.M., M. Srinivasan, S.T.S., P.I., N.B., J.S.D., G.E.C., and M. Salathe edited and revised manuscript; D.M., M. Srinivasan, S.T.S., P.I., N.B., J.S.D., G.E.C., and M. Salathe approved final version of manuscript.
Meniscus scanning software was kindly provided by Mike Myerburg and BK channel cDNA by Toshinori Hoshi.
- Copyright © 2014 the American Physiological Society