INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

THE CYSTIC FIBROSIS TRANSMEMBRANE CONDUCTANCE REGULATOR (CFTR) is a cAMP-regulated Cl channel and is responsible for regulating ion transport in many tissues where it is expressed, including the airways and airway submucosal glands, sweat glands, gastrointestinal tract, pancreatic and bile ducts, and the reproductive tract (58, 74). Defective CFTR function disrupts ion transport, which is believed to contribute to the clinical sequelae characteristic of cystic fibrosis (CF) (49, 58, 67, 74). CFTR regulation in vivo is accomplished through ATP binding and hydrolysis at nucleotide binding domain-1 and -2 and regulation of the phosphorylation status of the regulatory domain (10, 11, 36). This process is complex and depends on cAMP production through surface receptors that couple to adenylyl cyclase, cAMP degradation by phosphodiesterase (PDE) activity (40, 41), domain-domain interactions within the CFTR (55), protein-protein interactions between the CFTR and associated regulatory proteins (31, 32, 54, 64), and direct phosphorylation and dephosphorylation of the regulatory domain by protein kinase A (PKA) and phosphatases (6, 11), respectively. Receptor-based pathways that signal through cAMP are therefore relatively accessible candidates for promoting activation of the CFTR in vivo. This has been established for ₂-adrenergic receptors, which are commonly stimulated in assays to detect CFTR-dependent Cl secretion as measured by changes in the nasal potential difference (PD) in mice and in humans (44).

Adenosine (Ado) is a Cl secretagogue that signals through P₁ purinergic receptors, and recent studies have demonstrated that A_2B Ado receptors (ARs) can tightly couple to CFTR through PKA. In Calu-3 cells, part of this coupling is due to compartmentalized signaling of CFTR with adenylyl cyclase and PKAII through A kinase anchoring protein interactions (34, 71). T84 cells also express A_2B-ARs, where they have been shown to mediate neutrophil-stimulated Cl secretion (48, 70). Studies (2-4) have also suggested a role for phospholipase A₂ (PLA₂) in Ado-activated Cl secretion in T84 cells. The nature of this role and whether similar signaling contributes to CFTR activation in other CFTR-expressing cells and tissues, however, are unknown.

In this report, we describe CFTR regulation by A₂-ARs in vitro and in vivo and evaluate the contribution of cytoplasmic PLA₂ (cPLA₂) to A₂ receptor signaling in airway cells and the murine airway. Our findings suggest that A₂ receptors maximize CFTR activation by signaling through both adenylyl cyclase-PKA and cPLA₂ and that components of these receptor signaling pathways can activate common disease-causing CFTR mutations.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cell culture. All cell lines were purchased from the American Type Culture Collection (Manassas, VA). COS-7 cells were grown in DMEM plus 10% fetal bovine serum (FBS) and 1% penicillin plus streptomycin. IB-3-1 cells were grown in LHC medium plus 3% FBS and 1% penicillin plus streptomycin. Calu-3 cells were grown in MEM plus 10% FBS and 1% penicillin plus streptomycin supplemented with nonessential amino acids. To study polarized Calu-3 cells at an air-liquid interface, polyester Transwell-Clear Costar filters [0.4-µm pore diameter; 6-mm insert diameter for arachidonic acid (AA) release and Ussing chamber experiments and 24-mm insert diameter for immunocytochemistry experiments; Fisher Scientific, Pittsburgh, PA] were coated with human placental collagen matrix (Becton Dickinson, Franklin Lakes, NJ) at a concentration of 5 µg/cm² overnight and then seeded at ~1 × 10⁶ cells/cm². Once the filters were confluent (~1 wk), the medium was removed from the apical surface and the cells were fed only on the basolateral surface. After 48 h, resistance was checked (~1,000-2,000  · cm for the 6-mm filters and ~400-600  · cm for the 24-mm filters), and at 72-96 h the cells were studied as described in Transepithelial short-circuit currents. The primary human nasal cells were isolated and grown as explant cultures on Vectabond-treated glass coverslips as previously described (15) and were studied <1 wk after being seeded.

Transient CFTR expression. CFTR was transiently expressed in COS-7 cells with a vaccinia-based expression system as previously described (16, 17). Cells grown on Vectabond-treated glass coverslips were infected with vaccinia containing the T7 polymerase (generous gift of Dr. B. Moss, National Institutes of Health, Bethesda, MD) at a multiplicity of infection of 10 for 30 min. Wild-type (WT), F508, or G551D CFTR under control of the T7 promoter in the pTM-1 vector was then introduced into the cells in complex with N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium (DOTAP)-propane-1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE; 20 µg DOTAP-DOPE and 5 µg pTM-1 CFTR/5 × 10⁵ cells) for 4 h. These CFTR plasmids in the pTM-1 vector were the generous gift of Dr. S. Cheng (Genzyme, Cambridge, MA). Cells were then washed in PBS, returned to DMEM plus 10% FBS, and studied 18-24 h postinfection (for WT CFTR and G551D CFTR) or after being grown at 29°C for 48 h (F508 CFTR).

Fluorescence-based halide efflux measurements. To study CFTR activation, we measured halide efflux in COS-7 cells transiently expressing WT, F508, or G551D CFTR or in Calu-3 cells with the halide-quenched dye 6-methoxy-N-(3-sulfopropyl)quinolinium (SPQ; Molecular Probes, Eugene, OR) as previously described (16, 17). Briefly, COS-7 cells were seeded on Vectabond-treated glass coverslips, and Calu-3 cells were seeded on Vectabond-treated coverslips coated with 5 µg/cm² of human placental extracellular matrix. Cells were grown until ~80% confluent. Immediately before study, the cells were hypotonically loaded with 10 mM SPQ for 10 min and placed in a quenching NaI buffer. The cells were then placed in a specially designed perfusion chamber and studied at 23°C. The fluorescence of individual cells was measured with a Zeiss inverted microscope (excitation 340 nm, emission >410 nm), a PTI imaging system, and a Hamamatsu camera. Baseline fluorescence was measured in isotonic NaI buffer (16, 17), and cells were then perfused with isotonic dequench buffer (NaNO₃ replaced NaI) at the indicated time point (generally 200 s). The perfusate was then switched to dequench buffer plus agonist and requenched at the end of the experiment. Fluorescence was normalized for each cell to its baseline value, and increases are shown as percent increase in fluorescence above basal (quenched) fluorescence. Unless otherwise specified, the means ± SE of all cells were included in the curves shown.

Protein detection. To detect the A_2B-AR, cells were lysed with radioimmunoprecipitation assay buffer (150 mM NaCl, 1% Nonidet P-40, 0.05% sodium deoxycholate, 0.1% SDS, 50 mM Tris-Cl, and 10 mg/ml of phenylmethylsulfonyl fluoride, pH 8.0). For the COS-7 cells, lysates were immunoprecipitated with an isoform-specific polyclonal rabbit anti-A_2B-AR-antibody (Alpha Diagnostics, San Antonio, TX) raised against a 16-amino acid sequence corresponding to the third extracellular domain of human brain A_2B-AR cDNA as previously described (17). Briefly, anti-A_2B-AR antibody was linked to protein A agarose beads and incubated with cell lysates for 2 h. Precleared cell lysates (beads without primary antibody) were used as the negative control. The beads were then washed three times with PBS plus 0.1% Tween, and the immunoprecipitates were released with sample buffer incubated at 37°C for 10 min. For the Calu-3 and IB-3-1 cells, lysates were Western blotted without immunoprecipitation. Proteins were separated by SDS-PAGE with precast 12% gels (Novex gels; Invitrogen, Carlsbad, CA) and electrophoretically blotted onto polyvinylidene difluoride membranes. The membranes were then blocked with 1% BSA in PBS for 30 min, washed three times with PBS plus 0.1% Tween, and probed with anti-A_2B-AR antibody (1:1,000 dilution) for 2 h. Negative control membranes were blotted with a fivefold excess (by weight) of A_2B-AR-specific neutralizing peptide that was added with the primary antibody (Alpha Diagnostics). The membranes were washed three times and incubated with secondary antibody (1:1,000 dilution of goat anti-rabbit antibody conjugated to alkaline phosphatase; Southern Biotechnology Associates, Birmingham, AL) for 2 h. Membranes were washed three times and developed with 5-bromo-4-chloro-3-indolyl phosphate (BCIP) and 4-nitro blue tetrazolium chloride (NBT) in carbonate buffer (pH 9.8). Similar techniques were used to identify cPLA₂ in cell lysates using an affinity-purified mouse monoclonal IgG2b antibody raised against the amino-terminal domain of the human cPLA₂ cDNA (Santa Cruz Biotechnology, Santa Cruz, CA). This antibody was used to identify cPLA₂ protein in cell lysate immunoprecipitates (Calu-3 and IB-3-1 cells; anti-cPLA₂ linked to precleared protein G beads; Roche Diagnostics, Indianapolis, IN) or cell lysates alone (COS-7 cells). Proteins were separated by SDS-PAGE with 8% precast gels, and after being blotted onto polyvinylidene difluoride membranes, the membranes were blocked and washed as above and probed with primary antibody (1:1,000). The secondary antibody was a goat anti-mouse IgG alkaline phosphatase conjugate (1:1,000 dilution; Southern Biotechnology Associates), and blots were developed with NBT-BCIP as above.

Immunofluorescence studies. Nasal explant cultures from human surgical specimens were grown on Vectabond-treated coverslips coated with 5 µg/cm² of human placental extracellular matrix as previously described (15). At 5 days of age, cells were fixed in 4% formaldehyde in PBS (20 min; pH 7.4), washed twice with PBS, and then permeabilized with 0.05% surfactant Triton X-100 (Pierce Endogen, Rockford, IL). The slides were then pretreated with 1% BSA in PBS to block nonspecific protein-binding sites. A_2B-AR antigen was detected with the polyclonal rabbit anti-A_2B-AR antibody (1:25 dilution) in 1% BSA for 1 h. Cells were then washed three times with PBS over 15 min and incubated with secondary antibody (goat anti-rabbit FITC conjugate, 1:100 dilution in 1% BSA) for 1 h. The cells were mounted with 4',6-diamidino-2-phenylindole Vectashield mounting fluid (Vector Laboratories). For the studies of A_2B-AR localization in polarized Calu-3 cells, the A_2B-AR antigen was identified with the filter fold technique as previously described (7). Briefly, high-resistance monolayers were fixed with 3% transmission electron microscopy-grade formaldehyde (Touismis, Rockville, MD) for 45 min at room temperature and stained without detergent permeabilization. The fixed monolayers were washed three times with PBS over 15 min, blocked with nonimmune goat serum for 30 min (1:25 dilution; Vector Laboratories), and incubated with primary rabbit anti-A_2B-AR antibody (1:100 dilution in nonimmune goat serum) for 1 h at room temperature. Neutralizing peptide at five times (by weight) the concentration of primary antibody was used for negative control filters. Goat anti-rabbit Alexa fluorochromes (1:80 dilution; Molecular Probes) were used to identify the primary antibody. Nuclei were identified with Hoechst 33258 staining (20 µg/ml for 4 min). The filters were folded with the apical side exposed and mounted in 0.1% p-phenylenediamine (Sigma, St. Louis, MO) in PBS-glycerol (1:9 dilution). Digital confocal images were captured and analyzed with an Olympus IX70 inverted reflective fluorescent light microscope at 623 nm excitation with UplanAPO ×100 or U-APO/340 ×40 objectives, a Photometric Sensys digital camera, and IPLab Spectrum software supplemented with Power Microtome extension software (Signalytics, Fairfax, VA).

Transepithelial short-circuit currents. Calu-3 cells grown as monolayers at an air-liquid interface were mounted in modified Ussing chambers (Jim's Instruments, Iowa City, IA) and initially bathed on both sides with identical Ringer solutions containing (in mM) 115 NaCl, 25 NaHCO₃, 2.4 KH₂PO₄, 1.24 K₂HPO₄, 1.2 CaCl₂, 1.2 MgCl₂, and 10 D-glucose (pH 7.4). Bath solutions were vigorously stirred and gassed with 5% CO₂. Solutions and chambers were maintained at 37°C. Short-circuit current (I_sc) measurements were obtained with an epithelial voltage clamp (University of Iowa Bioengineering, Iowa City, IA). A 3-mV pulse of 1 s duration was imposed every 100 s to monitor resistance, which was calculated with Ohm's law. To measure stimulated I_sc, the mucosal bathing solution was changed to a low Cl solution containing (in mM) 1.2 NaCl, 115 sodium gluconate, and all other components as above plus 100 µM amiloride. Increasing concentrations of Ado were added to the mucosal or serosal bathing solutions (8 min of observation at each Ado concentration). After cells were stimulated with 100 µM Ado, 200 µM glibenclamide was added to the mucosal bathing solution, effectively blocking the stimulated I_sc (>90%).

Murine nasal PD measurements. Cftr(+) and cftr(/) mice (CFTR^unc mice, C57BL6J genetic background) were studied by a conventional nasal PD protocol (29, 69). The cftr(/) mice carried two copies of the human cftr cDNA, which contains a stop codon at position 489 (S489X). Cftr(+) mice included cftr(+/+) and cftr(+/) mice, which have been shown to have similar nasal ion transport characteristics (42). Genotypes were verified by PCR and dental enamel characteristics. Both male and female mice ~16-40 wk of age were studied. Mice were anesthetized with a cocktail consisting of ketamine (100 mg/ml, 82.5 µl), acepromazine (10 mg/ml, 30 µl), and xylazine (100 mg/ml, 15 µl) administered by intraperitoneal injection (0.1 ml/g body wt). The mouse tail was gently abraded, placed in lactated Ringer solution, and connected through a calomel cell to a high-impedance voltage follower (VF-1; World Precision Instruments, Sarasota, FL). An exploring bridge was established by connecting a Ag-AgCl electrode (wire) bridge to a syringe that pumped solutions at a rate of 180 µl/h. After ~5 min, mice were appropriately somnolent to permit cannulation of the nostril with a PE-10 cannula pulled to a tip diameter of ~0.15 mm. The solutions perfused included Ringer lactate plus amiloride (100 µM; solution 1); a low Cl concentration ([Cl]) solution containing 2.4 mM K₂HPO₄, 0.4 mM KH₂PO₄, 115 mM sodium gluconate, 25 mM NaHCO₃, 1.24 mM calcium gluconate₂, and 100 µM amiloride (solution 2); and solution 3 (solution 2 plus agonist as described in text). Each superperfused condition was studied for 6 min (total of ~18 min per protocol per mouse).

AA release. Calu-3 cells, grown to confluence on 35-mm culture dishes coated with 5 µg/ml of human placental collagen or as high-resistance monolayers at an air-liquid interface were washed in PBS and loaded with 1 µCi/ml of [³H]AA overnight (Moravek Biochemical, La Brea, CA). The plates and filters were then washed five times with PBS and placed in MEM plus 10% FBS with and without 100 µM Ado (750-µl volume for the 35-mm plates; 150-µl apical volume and 300-µl basolateral volume for the cells on filters). AA release was quantified for 20 min (cells on plates) and for 20 and 40 min (cells on filters). The cells were lysed (1 N NaOH), and the effluxed AA from each plate was quantified by scintillation counting and normalized to the percent of total number of counts in a manner similar to that previously described (13).

cAMP measurements. Cellular cAMP was measured with an ELISA-based detection kit as previously described (17) (Cayman Chemical, Ann Arbor, MI). Briefly, cells grown on 35-mm dishes (~7 × 10⁶ cells/dish) were stimulated with agonist for 10 min, and the cellular cAMP was extracted with ice-cold ethanol. The supernatants were vacuum dried and resuspended in phosphate buffer, and the cAMP levels were quantified per the manufacturer's directions. For all experiments, papaverine (100 µM; nonspecific nonxanthine PDE inhibitor) was included to improve cAMP detection. Xanthine-based inhibitors were avoided because these commonly interact with Ado receptors (39).

Materials. Ado hemisodium salt, chlorpromazine (CPZ), H-89 Cl, forskolin, zaprinast, and AA were purchased from Calbiochem (San Diego, CA); albuterol (Alb; salbutamol), theophylline, 8-cyclopentyl-1,3-dipropylxanthine (DPCPX), 8-phenyltheophylline, rolipram, and milrinone were purchased from Sigma; papaverine HCl was purchased from Research Biochemicals International (Natick, MA).

Statistics. Descriptive statistics (means ± SE) and tests of statistical significance were performed with SigmaStat software (Jandel, CA). Paired and unpaired t-tests were used for samples with continuous data (cAMP levels, AA levels, stimulated PD measurements, and I_sc), and ² analysis was used to compare the number of F508 CFTR-expressing cells responding to different agonists (Ado, forskolin, DPCPX, and control). An -level of 0.05 was considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Human airway cells express A_2B-ARs. With a polyclonal anti-A_2B-AR antibody, Calu-3 and IB-3-1 cell lysates were probed for A_2B receptor expression through Western blotting. As a positive control, COS-7 cells, which our laboratory previously demonstrated to endogenously express A_2B-ARs (17), were also evaluated. Figure 1 shows that in all three cell types, a specific ~40-kDa band was identified. Using the same anti-A_2B-AR antibody, we investigated A_2B receptor localization in primary human nasal airway cells (Fig. 2A) and polarized Calu-3 cells (an airway cell line with a serous phenotype and high levels of CFTR expression; Fig. 2B) (37, 53). In both cell types, the antibody detected plasma membrane-localized A_2B-AR. Staining was predominantly along the apical surface in Calu-3 cells. Functional studies in Ussing chambers demonstrated that Ado added to either the apical or basolateral surface briskly activated I_sc (Fig. 2C, left). Ado was a more potent stimulus when added to the apical membrane than to the basolateral membrane, and I_sc stimulated from either membrane was sensitive to apical glibenclamide blockade (~90% inhibition of stimulated I_sc produced by 100 µM Ado; P < 0.05). Although the range of concentrations capable of activating I_sc was most consistent with A₂ receptor stimulation, these experiments do not exclude the possibility that additional AR subtypes may contribute to Ado-stimulated I_sc. Because HCO transport, in addition to Cl transport, may contribute to I_sc in Calu-3 monolayers (22, 66), further functional studies of CFTR activity were performed with an SPQ-based halide efflux assay (15-17).

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Identification of A2B adenosine (Ado) receptors (ARs). Characteristic 40-kDa band was identified (+) in 3 cell types. MW, molecular mass. Negative controls () were precleared beads with cell lysates (COS-7 cell immunoprecipitates, no primary antibody conjugated to beads) or with addition of neutralizing peptide during Western blotting (Calu-3 and IB-3-1 cells, 5-fold excess of neutralizing peptide added during incubation with primary antibody).

View larger version (39K): [in this window] [in a new window]   Fig. 2.   Immunofluorescence of primary human bronchial epithelial cells and Calu-3 cell monolayers studied with immunofluorescence (A and B, respectively) and in Ussing chambers (C). Cells were prepared and studied as described in METHODS. A, left: A2B-AR antigen is identified with a membrane-staining pattern (green). A, right: negative control cultures (5-fold excess of neutralizing peptide during incubation with primary antibody) eliminated membrane staining. B, top: A2B-AR antigen in Calu-3 monolayers (green) shows predominantly apical membrane staining. B, bottom: addition of neutralizing peptide (as in A) to negative control cultures eliminated membrane staining. C: effects of apical and basolateral addition of Ado on short-circuit current (Isc) in Calu-3 cells. Cells were grown at an air-liquid interface and studied in Ussing chambers. C, left: representative Isc tracings. Top: Calu-3 cells were initially cultured in lactated Ringer solution followed by (from left to right) 1) mucosal low-Cl concentration ([Cl]) buffer + amiloride (100 µM), 2) addition of mucosal 1 µM Ado, 3) addition of basolateral 1 µM Ado, and 4) blockade with mucosal 200 µM glibenclamide. Bottom: same experiment as above, except basolateral Ado (1 µM; 2) was added before mucosal Ado (1 µM; 3). Right: Ado was a strong stimulus when added to either membrane. Apical: P = 0.002 for 0.1 µM compared with 1.0 µM and 1.0 µM compared with 10 µM. Basolateral: P < 0.03 for 0.10 µM compared with 1.0 µM; P < 0.001 for 1.0 µM compared with 10 µM; and P < 0.001 for 10 µM compared with 100 µM. Apical-stimulated Isc was greater than basolateral-stimulated Isc for 0.1 and 10 µM (*P < 0.02) and 1.0 µM (P < 0.001). Approximately 90% of the stimulated Isc (100 µM, either membrane) was blocked by glibenclamide (200 µM added to the mucosal compartment; data not shown).

Ado stimulates PLA₂, adenylyl cyclase, and CFTR activation. In previous studies of CFTR regulation by A_2B receptors in COS-7 cells, our laboratory (17) showed that A_2B receptor signaling was a potent stimulus, accomplishing strong activation of CFTR despite only modest effects on cellular [cAMP] compared with forskolin. These results were similar to those reported by Barrett and colleagues (3, 4) in T84 cells. In subsequent studies by Barrett and Bigby (2), Ado-activated I_sc in T84 monolayers was found to be associated with AA mobilization and sensitivity to PLA₂ inhibition. To determine whether PLA₂ signaling might contribute to A₂ receptor activation of CFTR in airway and COS-7 cells, we evaluated cells for cPLA₂ expression and cPLA₂ activity after Ado stimulation. Figure 3 shows that in Calu-3, IB-3-1, and COS-7 cell lysates, a monoclonal anti-cPLA₂ antibody detected an ~110-kDa protein consistent with cPLA₂.

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Identification of cytoplasmic phospholipase A2 (cPLA2) in 3 cell types. Cell lysates were immunoprecipitated with anti-cPLA2 antibody (Calu-3 and IB-3-1 cells) and separated by SDS-PAGE. cPLA2 antigen (~110 kDa) was detected with monoclonal anti-cPLA2 antibody (arrow). Control conditions were precleared beads from cell lysates without primary antibody (Calu-3 and IB-3-1 cells). COS-7 cell lysates were separated by SDS-PAGE and Western blotted with anti-cPLA2 antibody. Controls received no primary antibody during blotting.

View larger version (9K): [in this window] [in a new window]   Fig. 4.   Ado stimulates arachidonic acid (AA) release from Calu-3 cells. Cells were loaded with [3H]AA overnight, washed 5 times, and then studied. Values are means ± SE. A: AA release from cells grown on 35-mm plastic dishes stimulated with Ado (100 µM) for 20 min compared with cells in unstimulated dishes (n = 10 dishes/condition). *P < 0.025. B: apical release of AA from Calu-3 cells grown on permeable supports at an air-liquid interface and stimulated with 100 µM Ado compared with cells on unstimulated filters (n = 6 filters/condition). Cells were stimulated for two 20-min time points. Release over the second 20 min is shown. For the entire 40-min period, Ado stimulated AA release by ~20% over the unstimulated condition (P = 0.06). *P = 0.025. C: basolateral release of AA for same 20-min time point. Total no. of counts (released to medium and retained in cells) for the stimulated and unstimulated conditions was not different for cells grown on dishes (A) or on filters (B and C). P = 0.08 for Ado-stimulated compared with control cells. P = 0.06 for entire 40-min time period.

View larger version (35K): [in this window] [in a new window]   Fig. 5.   Ado activation of halide efflux in cystic fibrosis transmembrane conductance regulator (CFTR)-expressing COS-7 cells (A and B) and Calu-3 cells (C-E). Values are means ± SE in %increase in fluorescence over basal (NaI-quenched fluorescence); n > 40 cells/condition. A: COS-7 cells expressing CFTR were stimulated with Ado (arrows), with activation seen at the higher concentration. Same experiment was performed in the presence of 25 µM AA, from 200 s, with shift of CFTR activation to 0.2 µM Ado. Treatment of cells with chlorpromazine (CPZ; 50 µM) from 200 s blocked CFTR activation by 2.0 µM Ado. AA at 25 and 100 µM (arrows) failed to activate halide efflux. B: COS-7 cells expressing CFTR were stimulated with forskolin (Forsk) alone (20 µM; arrow) or with addition of 25 µM AA (200 s) to forskolin stimulation of CFTR (20 µM). Treatment with CPZ (50 µM) had no effect on forskolin activation. T7 controls (no CFTR) stimulated with AA (25 µM at 200 s) and forskolin (20 µM, arrow) failed to activate halide efflux. C: Calu-3 cells stimulated with Ado (2 µM). Treatment with CPZ (50 µM from 200 s) or H-89 (5 µM for 4 h) blocked Ado-activated halide efflux. AA alone (100 µM; arrow) failed to activate halide efflux. D: Calu-3 cells stimulated with albuterol (Alb; 0.2 µM) activated halide efflux. Treatment with CPZ (50 µM from 200 s) had no effect on Alb activation, whereas treatment with H-89 Cl (5 µM for 4 h) blocked activation of halide efflux. E: Calu-3 cells treated with CPZ (50 µM) from 200 s. High-concentration Ado (100 µM) partially overcame CPZ blockade.

View larger version (40K): [in this window] [in a new window]   Fig. 6.   Ado stimulates cAMP production in COS-7, IB-3-1, and Calu-3 cells. cAMP levels were measured by ELISA as described (see METHODS). Cells were stimulated for 10 min before extraction, n = 4-12 cultures/condition. Concentrations were 10 µM Ado, 50 µM CPZ, and 100 µM AA. Papaverine (100 µM) was included in all conditions to enhance cAMP detection as previously described (17). Values are means ± SE for each condition. A: COS-7 cells. *P < 0.001 compared with control (100 µM papaverine). B: IB-3-1 cells. *P < 0.02 compared with control. C: Calu-3 cells. *P < 0.001 compared with control. P < 0.02 compared with AA alone. D: Calu-3 cells cAMP dose-response curve. Ado (0.1, 1.0, and 10 µM) stimulation of cAMP production was not inhibited by treatment with 100 µM CPZ. *P < 0.02 for Ado alone compared with papaverine control (100 µM). P < 0.01 for Ado + CPZ compared with papaverine control.

Ado activation of CFTR-dependent Cl transport in vivo. The results given in Ado stimulates PLA₂, adenylyl cyclase, and CFTR activation provide a framework in which to investigate Ado-activated Cl transport in vivo. For these studies, we investigated ion transport across the murine nasal mucosa using the nasal PD, a well-established bioelectric assay of CFTR activity in vivo. Figure 7 shows examples of nasal PD tracings from a cftr(+) (Fig. 7A) and a cftr(/) (Fig. 7B) mouse. In the cftr(+) mouse, Ado stimulated further hyperpolarization (Fig. 7A, arrow #3), consistent with Cl conductance. In the cftr(/) mouse, depolarization continued during perfusion with a low [Cl] solution and a low [Cl] solution plus Ado (500 µM). Figure 8A summarizes a comparison of Ado (100 µM)-, Alb (100 µM)-, isoproterenol (Iso; 100 µM)-, and forskolin (10 µM)-stimulated Cl secretion in cftr(+) mice. Ado was a potent Cl secretagogue, producing further polarization in 10 of 12 mice studied (P < 0.005 compared with the no-agonist or Iso control mice). Alb was also a strong agonist (similar to forskolin), producing further polarization in six of eight mice studied (P < 0.05 for Alb and forskolin compared with Iso and control animals). In contrast, Iso was less predictable, producing further hyperpolarization in only 5 of 14 mice. Both AR stimulation with Ado (500 µM) and ₂-receptor stimulation with the more specific ₂-receptor agonist Alb (500 µM) failed to activate Cl conductance in cftr(/) mice (Fig. 8B). These studies confirm that both receptors stimulate CFTR-dependent Cl transport. Ado-stimulated Cl conductance in cftr(+) mice was sensitive to A₂ receptor blockade with 8-phenyltheophylline (P < 0.02), indicating that Ado activates Cl conductance through A₂ receptor signaling (Fig. 8C).

View larger version (19K): [in this window] [in a new window]   Fig. 7.   Examples of nasal potential differences (PDs) in cftr(+) and cftr(/) mice. Mice underwent a standard nasal PD protocol as described in METHODS. Solution 1, lactated Ringer + 100 µM amiloride; solution 2, low [Cl] solution + 100 µM amiloride; solution 3, solution 2 + agonist. Upward deflections are hyperpolarization (lumen negative), conventionally taken to represent Cl secretion. A: cftr(+) mouse with 100 µM Ado included in solution 3. B: cftr(/) mouse with 500 µM Ado included in solution 3. A depolarizing phenotype throughout the entire protocol is seen.

View larger version (11K): [in this window] [in a new window]   Fig. 8.   Stimulated Cl transport in cftr(+) (A and C) and cftr(/) (B) mice. Change in PD after the switch to solution 3 is shown. A: cftr(+) mice stimulated with Ado (100 µM; n = 12), Alb (100 µM; n = 8), Forsk (10 µM; n = 10), isoproterenol (Iso; 100 µM; n = 14), and low [Cl] control (no agonist; n = 18) included in solution 3 are shown. *P < 0.005 for Ado compared with ISO or control. P < 0.025 for FORSK compared with ISO or control; P < 0.05 for Alb compared with ISO or control. B: cftr(/) mice stimulated with Ado (500 µM; n = 10) or Alb (500 µM; n = 10). A depolarizing phenotype is seen with both agonists that is similar to that of cftr(/) control mice (low-[Cl] alone in solution 3; n = 14). C: cftr(+) mice stimulated with 50 µM Ado in solution 3 (n = 8) and 50 µM Ado (solution 3) + 100 µM 8-phenyltheophylline (8-PT) in solutions 2 and 3 (n = 10). *P < 0.02 for change in PD after stimulation with solution 3 (+ Ado) compared with solution 2 (no agonist).

View larger version (11K): [in this window] [in a new window]   Fig. 9.   cPLA2 inhibition blocks Ado-stimulated Cl transport. Nasal PDs were performed in cftr(+) mice as described in Fig. 7 (n = 10-12 mice/condition). Mice were stimulated with Ado (500 µM) or Alb (500 µM) in solution 3 in the presence and absence of CPZ (100 µM, included in solutions 2 and 3). *P < 0.02 for Ado + CPZ compared with Ado alone, Alb alone, or Alb + CPZ.

A_2B-ARs activate F508 CFTR and G551D CFTR. The results discussed in Ado activation of CFTR-dependent Cl transport in vivo show that Ado and A₂ receptor signaling potently activate CFTR-dependent Cl transport in vitro and in vivo. To determine whether A_2B receptor signaling could be used to improve the activity of common disease-causing CFTR mutations, we transiently expressed F508 CFTR (class II mutation) and G551D CFTR (class III mutation) in COS-7 cells. Figure 10 shows that after localization to the cell membrane (growth at 29°C for 48 h), F508 CFTR was activated by A_2B receptor stimulation (10 µM Ado) in a fashion similar to direct adenylyl cyclase activation with forskolin (20 µM). Halide efflux was stronger than that produced by stimulation with DPCPX, an agent that can acutely stimulate F508 CFTR-dependent halide efflux in F508 CFTR-expressing cells [and is currently in clinical trials as an activator of F508 CFTR in CF patients (24, 30); P < 0.001 comparing the proportion of responding cells stimulated by Ado with DPCPX]. We did not test the ability of A_2B receptor stimulation to activate F508 CFTR after prolonged treatment with DPCPX. Activation of F508 CFTR was accomplished despite only modest effects of Ado on [cAMP] compared with forskolin (17).

View larger version (31K): [in this window] [in a new window]   Fig. 10.   A2B receptor activation of halide efflux in F508 CFTR-expressing cells. COS-7 cells expressing F508 CFTR were studied after growth at 29°C for 48 h. Cells were perfused with NaI quenching buffer from 0 to 200 s, NO3 dequenching buffer from 200 to 500 s, and NO3 buffer + agonist from 500 s (arrow). Nos. in parentheses, total no. of responding cells/total no. of screened cells per condition (as described in METHODS). No. of responding cells in each condition was compared by the 2 test. Values are means ± SE of responding cells in each stimulated condition for Forsk (20 µM) and Ado (10 µM) or means ± SE of nonresponding cells for the 8-cyclopentyl-1,3-dipropylxanthine (DPCPX; 30 nM)-stimulated cells and NO3 control cells. Concentration of DPCPX used has previously been shown to maximally induce acute 36Cl release from F508 CFTR-expressing cells (24, 30). *P < 0.001 for Ado or Forsk compared with DPCPX or NO3 conditions.

View larger version (32K): [in this window] [in a new window]   Fig. 11.   Activation of halide efflux in G551D CFTR-expressing cells. A: G551D CFTR-expressing cells compared with wild-type (WT) CFTR-expressing cells. COS-7 cells transiently expressing WT or G551D CFTR were studied with 6-methoxy-N-(3-sulfopropyl)quinolinium (SPQ) as described in METHODS; n 40 cells/condition. G551D CFTR-expressing cells failed to activate halide efflux when stimulated with Ado (200 µM) + papaverine (PAP; 200 µM), Ado + rolipram (ROL; 20 µM) in combination, or Forsk (20 µM; arrow). WT CFTR-expressing cells stimulated with Forsk (CFTR + Forsk; 20 µM) are shown for comparison. B: G551D CFTR-expressing COS-7 cells stimulated with Ado (200 µM) and AA (100 µM) combined with a series of phosphodiesterase inhibitors (PDEis). Cells were perfused with NO3 buffer + AA from 200 s and Ado + PDEi was added (arrows). Isotype-specific PDEis were studied at concentrations ~20-fold over the inhibition constant of their respective PDEs. Cells were also studied with Ado + AA alone (i.e., no PDEi). Values are means ± SE of all cells studied in each condition (n > 40). Concentrations of PDEis were (in µM) 20 milrinone (MIL), 20 ROL, 200 PAP, 5 ZAP, and 200 theophylline (THEO).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this report, we investigated the cellular pathways utilized by Ado to activate CFTR-dependent ion transport. Using a series of experimental systems, including single cells, polarized cell monolayers, and in vivo measurements in cftr(+) and cftr(/) mice, we demonstrated that Ado, through A₂-ARs, activates CFTR-dependent halide transport. These receptors are expressed in human airway cells and signal in part through cAMP and PKA. We also demonstrated the dependence of this signaling pathway on PLA₂ activity in vitro and in vivo. Finally, we showed that the two most common disease-associated CFTR mutations can be activated by A_2B receptors alone (F508 CFTR) or by using messenger pathways stimulated by A_2B receptors (G551D CFTR). These results help to establish the physiological role of Ado-stimulated Cl secretion in vivo and may identify new targets for improving the function of mutant CFTR molecules.

Ado is a ubiquitous signaling molecule, serving protean functions ranging from regulation of neurotransmission to cardiovascular tone and inflammation in many organ systems. Our results, together with those of other investigators, indicate an important role for Ado as a regulator of CFTR. Ado and A₂ receptors are potent activators of Cl secretion in primary human airway cell monolayers, in addition to canine and gerbil airway cells and T84 intestinal cells (28, 46, 59, 70). In each of these systems, Cl secretion has been accomplished in the setting of low cAMP levels. Part of this high efficiency is due to compartmentalization of A_2B receptors, transduction proteins, and CFTR (34, 71). Our results suggest that in addition to spatial compartmentalization, efficient transduction between A₂ receptors and CFTR involves cPLA₂ activation and AA signaling. A₂ receptors, including A_2A and A_2B subtypes, classically signal through stimulation of adenylyl cyclase, cAMP, and, ultimately, PKA (56). Although A_2A receptors appear to hold more strictly to this observation, A_2B receptors have been shown to stimulate additional signaling pathways such as phospholipase C, direct regulation of calcium channels, and, in our experience, PLA₂ (26). These observations are of particular interest because recent studies (27) have demonstrated abnormal lipid profiles, including elevated AA levels, in murine CF tissues. Our findings demonstrate that lipid signaling may be important to the regulation of CFTR activity by Ado in vitro and in vivo. The present results therefore set the stage for future studies designed to test the relationship between membrane lipid composition, A₂ receptors, and CFTR activity.

There are two general classes of PLA₂ enzymes: cPLA₂ (three subtypes) and secretory PLA₂ (seven subtypes) (33). Both are expressed in many tissues including the lung and leukocytes. cPLA₂ translocates to cell membranes after activation and primarily releases AA from membrane lipids. cPLA₂ has been shown to play an important immunologic role in human airway epithelial cells, governing AA release from epithelia after immunostimulation (76). Studies (50) have also observed dysregulation of PLA₂ in the CF airway, with increased AA release seen after bradykinin exposure in F508 CFTR-expressing airway cells compared with normal control cells. AA is the parent molecule of two important inflammatory signaling cascades including 1) the cyclooxygenase (COX) pathway, which leads to the production of prostanoids and thromboxanes and 2) the lipoxygenase pathway in which AA is metabolized to 5-hydroperoxyeicosatetraenoic acid, 5-hydroxyeicosatetraenoic acid, and the leukotrienes A₄-E₄. Clinical studies suggest that products of AA metabolism, including proinflammatory species, have direct relevance to CF-related inflammation. The COX-1 inhibitor ibuprofen is routinely used in the pediatric CF population to reduce lung inflammation and slow the progression of CF lung disease (45). Glucocorticoids, which suppress COX-2 expression, have also been studied in CF clinical trials, attenuating the decline in pulmonary function in pediatric CF patients (25). The relevance of AA signaling pathways to CFTR activity and ion transport, however, has not been previously demonstrated. Our results suggest that a buildup in AA may have beneficial effects on CFTR activity in the context of its activation by surface receptors.

The mechanism by which A₂ receptors activate CFTR-dependent Cl transport is complex. CFTR activation by A₂ receptors could be blocked by either PKA or cPLA₂ inhibition (Fig. 5). Production of cAMP was not reduced by PLA₂ inhibition in any of the three cell lines (Fig. 6). Although 10 min of high-dose (100 µM) AA exposure could stimulate some cAMP production in the two airway cell lines (primarily Calu-3 cells, likely through the actions of distal metabolites such as PGI, PGE₂, PGF₂ or PGD) (73), it failed to augment cAMP production by Ado or acutely activate CFTR when used alone (COS-7 or Calu-3 cells; Fig. 5). These results suggest that the additional effects of AA on CFTR-dependent halide transport were not through influences on total cellular cAMP. Rather, A₂ receptors appear to activate both adenylyl cyclase and cPLA₂ in parallel, with each signaling pathway contributing to the maximization of CFTR activity. PLA₂ signaling was not necessary to activate CFTR through receptors, however, because ₂-adrenergic stimulation activated CFTR-dependent Cl transport in vitro and in vivo despite cPLA₂ inhibition (Figs. 5 and 9).

Figure 12 is a summary model that provides three possible mechanisms by which cPLA₂ and AA could contribute to A₂ receptor activation of CFTR. It is possible that Ado and/or AA exerts its effects in part by activating K⁺ channels and increasing the driving force for halide efflux. AA has been shown to have both stimulatory and inhibitory effects on epithelial K⁺ channels, and modulation of basolateral K⁺ channel activity can strongly influence transepithelial Cl transport (20, 21). In related experiments, Ado-activated halide efflux in CFTR-expressing COS-7 cells was not reduced by two K⁺ channel blockers, including barium and tetraethylammonium (10 mM each) (18, 19). These results suggest that the stimulatory effects of Ado on CFTR seen in our studies were independent of K⁺ channel activation. The observation that AA augmented the stimulation of CFTR by Ado and forskolin (Figs. 5 and 11), but failed to stimulate CFTR alone in vitro (Fig. 5), suggests that AA may exert a permissive effect on CFTR activation. Membrane-derived fatty acids, including AA and its metabolites, have been demonstrated to interact with ion channels and influence their activation by other signaling molecules, including nucleotide sensitivity of ATP-activated K⁺ channels (5, 65), light-sensitive transient receptor potential, and transient receptor potential-like channels (14), small Cl channels in rabbit parietal cells (61), the nicotinic acetylcholine receptor (9), and calcium channels as part of oscillating calcium currents (51). AA does interact with CFTR and alter its Cl channel activity. Cytoplasmic AA produces a flickery block of CFTR in inside-out membrane patches after heterologous expression (1, 35, 47). Membrane lipids (including AA) can have differential effects on the function of other channels, stimulating channel activity on one surface and inhibiting channel activity on the other (52). We speculate that AA or one of its metabolites may interact with an external portion of the channel or with an unidentified regulatory factor that secondarily influences CFTR Cl channel function. Alternatively, products of AA metabolism could accelerate local cAMP production and, in this way, increase CFTR channel activity. This process would be difficult to detect with total cellular assays to quantitate [cAMP]. The effect would need to be quite dramatic, however, because G551D CFTR, which has previously been shown to be poorly responsive to powerful phosphorylating stimuli, appears to be activated by costimulation with AA (Fig. 11B, Refs. 38, 58, and 77). This occurs despite no detectable change in [cAMP] when combined with Ado (Fig. 6A) but is most pronounced when combined with PDE inhibition. Single-channel-based studies will be necessary to help clarify the mechanism by which cPLA₂ and AA promote CFTR activity. Finally, our results do not exclude the possibility that additional processes may be involved because CPZ has been shown to have effects on other cell signaling pathways in addition to PLA₂ (8, 12, 23, 43, 57).

View larger version (16K): [in this window] [in a new window]   Fig. 12.   Three potential mechanisms by which A2 receptors signal CFTR activation. Ado stimulates A2 adenosine receptors (A2-AR), which activate cPLA2 and adenylyl cyclase. AA or products of AA metabolism, when combined with cAMP, helped maximize CFTR transport of Cl. The mechanism could involve effects on the electrochemical driving force for ion transport through CFTR (i.e., activation of K+ channels) through direct effects of lipid species with CFTR or associated factors (direct effects) or possibly by increasing the rate of cAMP production in the vicinity of CFTR (accelerated cAMP production). COX, cyclooxygenase; LO, lipoxygenase.

The results shown in Figs. 7-9 indicate that Ado is a potent Cl secretagogue in vivo, activating CFTR-dependent Cl transport across the murine nasal mucosa more predictably than Iso (when studied at 100 µM). This is of relevance to human nasal PD protocols, which typically use ₂-receptor stimulation with Iso to aid in the detection of CFTR-dependent Cl transport. Our results suggest that comparisons of these two receptor pathways in human subjects may improve the sensitivity of the nasal PD, which could have implications for clinical studies that use this assay to measure low-level CFTR function. The importance of PLA₂ signaling to Ado activation of Cl transport was also demonstrated in vivo because Ado- but not Alb-stimulated Cl conductance was sensitive to PLA₂ inhibition.

A₂ receptor signaling may also be useful in improving the activity of disease-related CFTR mutations. Using an in vitro system, we have demonstrated that CFTR mutations from three subclasses (classes II, III, and IV) can be activated by A_2B receptors (17) (Figs. 10 and 11). Although the expression levels of the mutant CFTRs were far higher than those found in vivo, the results shown in Figs. 10 and 11 support the notion that mutant CFTRs, when localized to the cell membrane, can be activated by endogenous receptor-based signaling pathways. Current approaches to increase mutant CFTR activity, including improvement of CFTR biosynthesis [e.g., suppression of premature stop mutations with aminoglycoside antibiotics (7, 15, 75)], increase in trafficking of F508 CFTR to the cell membrane [e.g., treatment with butyrate compounds (60)], or treatment of surface mutant CFTRs with activating compounds [e.g., genistein or PDE inhibitors (38, 68)] may be complemented by the use of A₂ receptor signaling pathways. Our findings also suggest that common mutant CFTRs should be at least partially responsive to this signaling pathway when available at the cell membrane.

In summary, our studies demonstrate that A_2B-ARs are expressed in CF and normal airway cells, localizing predominantly to the apical membrane of polarized Calu-3 cells. A₂ receptors mediate activation of CFTR-dependent Cl transport in vitro and in vivo, utilizing cPLA₂ and AA in addition to cAMP and PKA. Ado and A₂-ARs compare favorably with other agents as activators of CFTR-dependent Cl conductance, stimulating Cl secretion better than the ₂-adrenergic receptor agonist Iso in mice and activating F508 CFTR similar to forskolin in vitro. Our studies therefore provide a rationale for the investigation of the effects of Ado and A₂ receptor signaling on measured Cl secretion in humans.