Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Cystic fibrosis (CF) is the most common lethal recessive genetic disorder in Caucasians. The disease affects a number of organs, including the pancreas and lung, causing abnormal fluid and electrolyte transport in exocrine epithelia (1). The basic defect is a change in a single gene coding for a 170 kD protein, the cystic fibrosis transmembrane conductance regulator (CFTR) (2, 3). The protein contains two similar units, each including a membrane-spanning domain and a nucleotide-binding domain linked by a single regulatory domain. The most common mutation (70% of the mutant alleles) is the deletion of a phenylalanine at position 508 of CFTR (F508 CFTR) in the first nucleotide-binding domain (4).

Recent studies on the mechanism by which CF mutations affect CFTR function have shown that F508 CFTR is abnormally retained in the endoplasmic reticulum, reducing its delivery to the plasma membrane (5, 6). The defective transport and processing of F508 CFTR is not fully understood, but seems to involve incomplete glycosylation of the protein (6), rendering it more sensitive to degradation by intracellular proteases (7). While it has been established that CFTR is a cAMP-regulated chloride channel (8), it seems to have other functions, such as the regulation of high-conductance chloride channels (9) and the transport of small molecules across the membrane (10).

CF is often associated with impaired pulmonary function due to chronic airway infection and inflammation (11). CF patients are very sensitive to infection because there are more receptors for P. aeruginosa and S. aureus in the apical membrane of CF epithelia (12). Their increased sensitivity to inflammatory stimuli might also help to make pulmonary infections chronic, with high levels of neutrophil-derived elastase in the pulmonary fluid of CF patients (13), and increased production of IL8 (14) and 5-lipoxygenase derivatives (15) in cells from F508 patients.

We have shown that a tracheal epithelial cell line from a CF patient bearing the F508 mutation releases 3-fold more arachidonic acid (AA) and produce more eicosanoids in response to bradykinin than do control cells. This increased sensitivity was not due to a clonal selection or to individual variation as the same hypersensitivity was found in epithelial tracheal cells in primary culture and skin fibroblasts from other patients bearing the same mutation (16). This increased production of lipid mediators is not due to a difference in the number of bradykinin receptors, but seems to involve dysregulation of the transduction pathway coupling the bradykinin receptor to the stimulation of phospholipase A₂ (PLA₂).

The bradykinin receptor belongs to the family of the seven-transmembrane domain receptors coupled to heterotrimeric G proteins. The bradykinin receptor in tracheal epithelial cells lies on the apical membrane (17) and is coupled to the stimulation of phosphatidylinositol hydrolysis (18). Bradykinin also stimulates a PLA₂, releasing free AA from membrane phospholipids and resulting in the synthesis of PGE₂ (16, 18). The PLA₂ responsible for this AA release from tracheal epithelial cells is unknown. Two forms of PLA₂ are involved in the stimulation of eicosanoid synthesis in mammalian cells. A 14 kD secreted form, whose synthesis is induced by cytokines and growth factors, is responsible for the long-term stimulation of eicosanoid production by membrane receptors (for review, see 19). An 85 kD cytosolic form (cPLA₂) is rapidly activated in response to stimulation of G protein-coupled receptors, cytokine receptors and growth factor receptors, and is responsible for the short-term stimulation of eicosanoid production (for review, see 20). The way in which cPLA₂ is stimulated is still poorly understood, but seems to involve its translocation by a rise in the intracellular concentration of calcium ions (21), and its phosphorylation by protein kinases (22, 23).

We have examined the mechanism by which the F508 CFTR mutation affects AA release by comparing the coupling of bradykinin receptors to the activation of PLC and PLA₂ in a F508 tracheal epithelial cell line, a double heterozygote cell line (S549N, N1303K), and in a control cell line. The results suggest that the F508 mutation of CFTR increases the ability of the PLA₂ to hydrolyze membrane phospholipids, while the double heterozygous form does not. This increase seems to involve an increase of Gq/11 expression, leading to greater stimulation of PLC and MAP kinase and increased anchoring of cPLA₂ to the membrane.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Materials

Dulbecco's modified Eagle's medium (DMEM), Ham's F12 medium, penicillin, streptomycin, glutamine, and trypsin-EDTA were from Gibco BRL (Lergy-Poutoise, France). Ultroser G was from Biosepra (IBF, Villeneuvela Lagarenne, France). [³H]Arachidonic acid (3,000 Ci/mmol), L 1-stearoyl-2-[1-¹⁴C] arach idonyl-phosphatidylcholine (55 mCi/mmol), the ECL (Enhanced Chemi-luminescence) detection system, and [³H]D-myo-inositol 1,4,5 triphosphate assay system, [³²P] ATP (1,600 Ci/mmol) and radiographic films (hyperfilm MP) were purchased from Amersham (Les Ulis, France). Thin layer chromatography silica gel plates and nitrocellulose (BA83) were from Schleicher and Schuell (Ceratabo Aubervillies, Switzerland). Phosphocellulose filter paper (P81) was from Whatman. The antisera against G subunits were purchased from Du Pont New England Nuclear (Du Pont de Nemours, Les Ulis, France). Blotting grade goat anti-rabbit IgG (H+L) horseradish peroxidase conjugate was from Bio-Rad (Ivory Suseine, France). Monoclonal mouse anti-MAP kinase (p42 and p44 MAP kinase) antibody was from Zymed (TEBU, Le Perray, France). Rabbit anti-mouse peroxidase-conjugated monoclonal antibody was from Biosys (Compiegne, France). Monoclonal anti-human cPLA₂ was from Santa-Cruz (TEBU, Le Perray, France). Monoclonal mouse anti-phosphotyrosine antibody and the MAP kinase substrate peptide were from Upstate Biotechnology, Inc. (Euromedex, Souffel Weyershein, France). Fura-2/AM was from Molecular Probes (Interchim, Montflaçon, France). Thapsigargin and PD098059 were from Calbiochem (France Biochem, Meridon, France). 1-stearoyl, 2-arachidonyl L--phosphatidylcholine, potato acid phosphatase (type III) and all other reagents were from Sigma (Saint-Quentin, France).

Preparation of Human Recombinant Cytosolic Phospholipase A₂

The coding sequence of the cPLA₂ cDNA was cloned by RT-PCR from U937 cells. NheI and BglII restriction sites were added to the primers corresponding to the 5' and 3' ends respectively. The PCR product was digested and inserted into the XbaI and BglII sites of the PVL1393 vector. This cPLA₂-PVL1393 plasmid and the linear transfection module from Invitrogen were used to obtain recombinant virus from Sf9 cell cultures. The cytosol of Sf9 cells was prepared 69 h after infection. The specific activities in the cytosol was 18-24 nmol/min/mg proteins, and were consistent with values previously published (24, 25).

Cell Culture

Tracheal epithelial cells from cystic fibrosis and normal human fetuses have been immortalized and characterized (26). Three cell lines were studied: a control cell line (NT-1), a CF cell line carrying the homozygous mutation F508 of CFTR (CFT-2) and a CF cell line carrying a double heterozygous mutation S549N and N1303K (CFT-1). The three cell lines were grown to confluency in 1:1 DMEM/ Ham's F12 medium, supplemented with 2% Ultroser G, 100 IU/ml penicillin and 100 µg/ml streptomycin at 37°C in 5% CO₂. Confluent cells were then deprived of serum by incubation with medium containing 0.2% bovine serum albumin for 24 h before experiment.

Cytosol and Membrane Preparation

The cells were washed twice with phosphate-buffered saline (PBS) and scraped off into buffer A (40 mM Tris/HCl, pH 7.4, 0.25 M sucrose, 1 mM PMSF, 1 µg/ml leupeptin), buffer B (40 mM Tris/HCl, pH 7.4, 0.25 M sucrose, 1 mM EDTA, 1 mM PMSF, 1 µg/ml leupeptin), or buffer C (40 mM Tris/HCl, pH 7.4, 0.25 mM sucrose, 10 µM CaCl₂, 1 mM PMSF, 1 µg/ml leupeptin) and disrupted by sonication. The broken cells were centrifuged at 100,000 × g for 1 h at 4°C, and the cytosol and pellet (resuspended in buffer A or buffer B) were stored at 80°C at protein concentrations of 0.5- 1 mg/ml.

MAP kinase activation was studied in confluent cells starved for 24 h and stimulated with 1 µM bradykinin or 5% fetal calf serum for 5 min. The cells were washed three times with ice-cold PBS, scraped off into 1 ml, 25 mM Hepes buffer, pH 7.4, containing 5 mM EDTA, 50 mM NaF, 100 µM sodium orthovanadate, 1 mM PMSF and 10 µg/ml leupeptin and homogenized by passing them 30 times through a 25-gauge needle. They were centrifuged at 100,000 × g for 20 min at 4°C and the supernatant, containing MAP kinases, was aliquoted and stored at 80°C until use.

Phospholipase A₂ Activity

The substrate was prepared by drying L 1-stearoyl-2-[1- ¹⁴C] arachidonyl-phosphatidylcholine (40,000 cpm/nmol) in chloroform/methanol under a stream of nitrogen, resuspending it in 10 µl ethanol plus 10 µl diethyl ether and dispersing in 100 mM Tris/HCl, pH 8.5 by sonication. The substrate (4 µM) was added to the assay mixture, which contained 100 mM Tris/HCl, pH 8.5, 5 mM CaCl₂, 0.1% bovine serum albumin and cytosol or membrane proteins (50 µg) prepared with buffer A, B or C, in a final volume of 250 µl. The mixture was incubated at 37°C for 30 min and the reaction was stopped by adding chloroform/methanol. The lipids were extracted by the Bligh and Dyer procedure (27) and resolved by thin-layer chromatography on silica gel plates using chloroform/methanol/water (65:25:4 v/v/v) as solvent. The spots were visualized with iodine vapor and the spots corresponding to phosphatidylcholine and free fatty acids were scraped off. Their radioactivity was determined by liquid scintillation counting.

Immunoblot Analysis

Membrane proteins (100 µg) in buffer A were separated by 10% SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose by electroblotting for Western blotting analysis of G subunits. The blots were washed for 1 h in blocking buffer (Tris-buffered saline, pH 7.6, 5% nonfat dried milk, 0.1% Tween 20) at room temperature and then incubated for 2 h in the same buffer containing: RM1 directed against Gs; AS/7 directed against Gi1 and Gi2; EC/2 directed against Gi3 and Go; or QL directed against Gq and G11. The blots were washed thoroughly and incubated for 1 h with peroxidase conjugated goat anti-rabbit antibody. Immunostained proteins were visualized using the ECL detection system and quantified by densitometry.

Cytosolic or membrane proteins (100 µg) were separated by 7.5% SDS-PAGE for the Western blotting analysis of cPLA₂. The blots were incubated for 3 h with anti-cPLA₂ mouse monoclonal antibody (1 µg/ml) and revealed with peroxidase-conjugated anti-mouse antibody as above. For the phosphatase treatment, 250 ng partially purified recombinant human cPLA₂ were incubated with 5 µg acid phosphatase in 100 mM Hepes, pH 6, containing 2 mM MgCl₂, 1 mM dithiothreitol, 1 mM PMSF, 10 µg/ml leupeptin at 30°C for 60 min.

The mobility shift of MAP kinases was assessed by separating cytosolic proteins (100 µg) on 8-15% continuous SDS-PAGE gradient, followed by immunoblotting. The blots were incubated with mouse monoclonal anti-MAP kinase antibody (0.25 µg/ml) for 2.5 h, or with mouse monoclonal anti-phosphotyrosine (1 µg/ml) for 3 h, and specific reactions revealed with peroxidase-conjugated anti-mouse antibody, as described above.

Assay of Inositol-(1,4,5) Triphosphate

Confluent cells starved for 24 h were suspended in 1:1 DMEM/ Ham's F12 medium supplemented with 0.2% bovine serum albumin, 25 mM Hepes, pH 7.5 and 10 mM LiCl at 7.5 × 10⁶ cells/ml. They were incubated for 30 min at 37°C and then stimulated with 1 µM bradykinin for the indicated time under gentle agitation. The reaction was stopped by adding 0.2 volume of ice-cold 20% perchloric acid and incubation for 20 min at 4°C. The precipitate was removed by centrifugation at 2,000 × g for 15 min at 4°C. The pH of the supernatant was then adjusted to 7.5 with 10 M KOH and the centrifugation repeated. The resulting supernatant was stored at 20°C. The extracts were centrifuged at 2,000 × g for 7 min prior to assaying inositol- (1,4,5) triphosphate using a competitive protein binding assay kit (Amersham, France).

AA Release from Intact Cells

See reference 16. Confluent cells were incubated with 0.5 µCi/ml [³H]arachidonic acid in DMEM/Ham's F12 medium for 15 min, washed with PBS plus 0.2% bovine fatty acid-free serum albumin, then with PBS, and then incubated in fresh medium supplemented with 0.2% bovine fatty acid-free serum albumin, plus agonists or vehicle. The supernatants were removed at the indicated times and centrifuged at 10,000 × g for 10 min to remove cell debris. Under these conditions, the total incorporation, the distribution of phospholipids among the different clones, and the basal release of AA were similar in the cell lines (16). The cells were washed and scraped off. The radioactivity of cells and media was quantified by scintillation counting.

MAP Kinase Assay

The kinase activity was measured by incubating 1-3 µg cytosolic proteins for 10 min at 30°C with 0.5 µCi [³²P]ATP and 1 mM of a synthetic substrate peptide (APRTPGGRR) corresponding to amino acids 95-98 of bovine myelin basic protein (28), in a buffer containing 80 mM -glycerophosphate, pH 7.4, 20 mM EGTA, 15 mM MgCl₂, 1 mM PMSF, 50 µg/ml aprotinin, 4 µg/ml leupeptin, 10 µg/ml antipaïn, 1 mM trypsin inhibitor, 1 mM benzamidine, 10 µg/ml pepstatin and 1 mM of the tyrosine phosphatase inhibitor sodium orthovanadate, in a final volume of 25 µl. The reaction was stopped by adding 3% trichloracetic acid for 10 min on ice. After centrifugation at 10,000 × g allowing separation of phosphorylated endogenous proteins from the specific substrate peptide, the supernatant samples were then spotted onto a 1 cm² piece of Whatman P81 phosphocellulose filter paper and washed twice in 30% acetic acid containing 3 mM ATP once for 2 h, then once overnight. After this procedure, the filters were dried with ether/alcohol (1:1 v/v) mixture. The filter-adsorbed radioactivity was quantified by liquid scintillation counting. All results were corrected for the radioactivity measured in control incubation (performed in the absence of the substrate peptide).

Fura-2/AM Loading and Ca²⁺ Imaging

See reference 29. Confluent cells grown on glass coverslips and deprived of serum for 24 h were bathed in 2 ml of Ca²⁺ buffer (10 mM Hepes buffered at pH 7.4 with Tris base, 10 mM glucose, 130 mM NaCl, 5 mM KCl, 1 mM CaCl₂) and were incubated for 35 min at 25°C with 2 µM Fura-2/AM in the presence of 1 mg/ml bovine serum albumin. Cells were then washed twice with Ca²⁺ buffer and incubated for 10 min at 25°C to facilitate hydrolysis of intracellular Fura-2/AM. Microfluorimetric measurements were performed at 25°C. Cells were perfused with Ca²⁺ buffer containing 1 µM thapsigargin or 1 µM bradykinin for 5 min or 1 µM A23187 for 15 min. Fluorescence images at 360 and 380 nm were recorded every 0.5 second. Data for single cells are expressed as the fluorescence ratio F360/F380, calculated after substracting respective backgrounds. Ca²⁺ imaging was described by Sauvadet (29).

Statistical Analysis

The results were expressed as means ± SEM of n experiments and were analyzed for statistical significance using the non-parametric Mann-Whitney test.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Cytosolic PLA₂ Activity Specific for AA in Epithelial Cystic Fibrosis Cell Lines

We have shown that an epithelial F508 cell line (CFT-2) releases more AA in response to bradykinin than do the control epithelial cell line (NT-1), or a CF epithelial cell line bearing the double heterozygous mutation S549N and N1303K (CFT-1). The involvement of CFTR in this increased sensitivity was checked using the effect of low temperature, which allows normal processing of F508 CFTR and the appearance of cAMP-regulated Cl channels in the plasma membrane (30). CFT-2 (F508) cells were incubated at 28°C for 48 h; this reduced the stimulation of AA release by bradykinin to the value in normal cells (Table 1).

As it is generally agreed that PLA₂ is the major enzyme involved in AA release, we compared the basal PLA₂ activities of the three cell lines by preparing cell homogenates in the presence of 1 mM EDTA to prevent non-specific Ca²⁺-induced PLA₂ translocation to the membranes during homogenization (21). Over 80% of the PLA₂ activity was in the cytosol of resting cells. This cytosolic PLA₂ activity was Ca²⁺-dependent and highly specific for 2 arachidonyl phospholipids (data not shown). The PLA₂ activity was highest in the cytosol of CFT-2 (F508) cells, and lowest in the cytosol from control cells NT-1 (CFT-2: 48.8 ± 3.8; CFT-1: 14.3 ± 2.2; NT-1: 4.2 ± 1.5 pmol/min/ mg of proteins). The low PLA₂ activity associated with the membranes was similar in all three cell lines in the presence of EDTA (Figure 1). When the homogenates were prepared in the presence of 10 µM Ca²⁺, a concentration known to translocate cPLA₂ to membranes (20), the PLA₂ activity decreased in cytosol (unshown results) and increased in membranes (Figure 1), suggesting that the translocated activity was due to Ca²⁺-sensitive cytosolic PLA₂. Membranes from CFT-2 (F508 cells) contained more PLA₂ activity than membranes from NT-1 and CFT-1 cells when prepared in the presence of exogenous Ca²⁺.

View larger version (15K): [in this window] [in a new window]   Figure 1.   Increased cPLA2 activity in membranes of resting cells by exogenous calcium. Confluent cells were scraped off and disrupted in buffer containing 10 µM calcium (solid bars) or 1 mM EDTA (open bars). The cytosol and membranes were separated and PLA2 activity was measured using L 1-stearoyl-2-[1-14C] arachidonyl-phosphatidylcholine as substrate. Specific activities are expressed in pmol substrate hydrolyzed/min/mg proteins. Data are means ± SEM of 4-6 independent experiments performed in triplicate. Asterisks: *P < 0.05 and **P < 0.02 indicate statistical differences when compared with CFT-2 cells in Ca2+ buffer (CFT-1, P = 0.0275; NT-1, P = 0.0143). NT-1 cells (P < 0.05), CFT-2 cells (P < 0.0001) and CFT-1 cells (P < 0.02) in Ca2+ buffer are significantly different from the same cells extracted in EDTA buffer, using the Mann-Whitney test.

No secreted PLA₂ activity was detected in culture media under control conditions, or after incubation with 1 µM bradykinin for 5-120 min, or with recombinant 10 ng/ml IL1/ for 24 h (data not shown).

Upregulation of cPLA₂ Associated with the Membrane of CFT-2 Cells after Stimulation by Bradykinin

The best candidate for causing G protein-coupled receptor-mediated AA release is the 85 kD cPLA₂ (21, 31). This cytosolic PLA₂ can hydrolyze membrane phospholipids because it is associated with the cell membranes. We therefore compared the ability of bradykinin stimulation to cause PLA₂ to associate with cell membranes. The binding of cPLA₂ to membrane allows its access to membrane phospholipids and hydrolysis. We have shown previously that the stimulation of AA release by confluent cells is maximal after 2 min, and plateaus up to 15 min (16). Cells were therefore stimulated with 1 µM bradykinin for 5 min, disrupted without added Ca²⁺, or with EDTA, and separated into cytosol and membrane fractions. The membrane-bound PLA₂ activity in CFT-2 cells stimulated by bradykinin was increased 2-fold, but remained at the resting level in CFT-1 and NT-1 cells (Figure 2). The PLA₂ activity remaining in cytosol was not significantly increased by pretreatment with bradykinin in any of the 3 cell lines (data not shown). PLA₂ activity remained associated with the membranes of bradykinin-stimulated cells, even with EDTA in the homogenization buffer, indicating that cPLA₂ was strongly associated with a membrane component once translocation had occurred.

View larger version (18K): [in this window] [in a new window]   Figure 2.   Increased cPLA2 activity associated with the membranes of CFT-2 (F508) cells in response to bradykinin. Confluent cells were incubated with 1 µM bradykinin or vehicle for 5 min. The cells were scraped off in buffer A (without EDTA), solid bars, or in buffer B (1 mM EDTA), open bars. PLA2 activity was measured in membranes and expressed as % of the activity in unstimulated cells. Data are means ± SEM of 3-5 independent experiments performed in triplicate. Asterisk: *P < 0.05 indicates statistical difference when compared with unstimulated cells in the same buffer. Stimulated activities of NT-1 cells (P < 0.05) and CFT-1 cells (P < 0.01) extracted in buffer A are significantly different from stimulated CFT-2 cells extracted in the same buffer, using the Mann-Whitney test.

Immunoblot experiments using an anti-cPLA₂ antibody were used to see if the increase in PLA₂ activity in CFT-2 membranes was due to an increase in cPLA₂. The anti-cPLA₂ antibody revealed two bands in the cytosol from tracheal epithelial cells: a 103 kD band comigrating with the cPLA₂ from U937 human macrophages and the recombinant human cPLA₂ expressed in Sf9 insect cells, and a 100 kD band comigrating with a lower band, which increased in blots prepared with recombinant cPLA₂ pretreated with phosphatase (Figure 3).

View larger version (34K): [in this window] [in a new window]   View larger version (19K): [in this window] [in a new window]   Figure 3.   Immunoblot analysis of cytosolic and membrane associated cPLA2. Confluent cells were incubated with 1 µM bradykinin (Bk) or vehicle (C) for 5 min. The cells were then scraped off in buffer A (without EDTA). Membrane or cytosol proteins (100 µg) were separated on 7.5% SDS-PAGE, transfered to nitrocellulose and blotted with a specific anti-cPLA2 antibody. (A) Representative blot of cytosols showing the phosphorylated (103 kD, P-cPLA2) and unphosphorylated forms (100 kD, cPLA2). Cytosol from U937 cells and recombinant human cPLA2 treated (+P) with acid phosphatase or not (P) were used as migration standards. (B) Representative blot showing the effect of bradykinin on the membrane cPLA2 content. (C) Densitometric quantification of blots of membrane cPLA2 from separate experiments: P-cPLA2, open bars, cPLA2, solid bars. Data are means of two to three independent experiments. Asterisk: *P < 0.05 indicates statistical differences when compared with unstimulated cells using the Mann-Whitney test.

The 103 kD band representing the phosphorylated form of cPLA₂ in the cytosol was of the same intensity as the unphosphorylated form which migrated at 100 kD. In contrast with the specific activities under resting conditions, CFT-2 cells did not have larger amounts of either phosphorylated or unphosphorylated forms of cPLA₂ than NT-1 cells or CFT-1 cells in cytosol (Figure 3A). Bradykinin stimulation caused only CFT-2 cells to have more higher phosphorylated cPLA₂ in the cytosol (Figure 3A). A Western blot of membrane proteins showed a main band of 103 kD in all three cell lines. The intensity of the 103 kD band increased in the membrane fraction of F508 (CFT-2) epithelial cells when the cells were stimulated with bradykinin, but was unchanged in CFT-1 and control cells (Figure 3B and 3C). Therefore, the increased stimulation of AA release after bradykinin treatment was parallel to an increased PLA₂ activity in membranes and to more phosphorylated cPLA₂ in the cytosol and membranes of CFT-2 cells than CFT-1 or control cells.

G Protein Content and PI Hydrolysis

The bradykinin receptor interacts with Pertussis toxin-sensitive and -insensitive G proteins (32, 33) to induce the hydrolysis of phosphatidylinositol diphosphate via the activation of phospholipase C, and stimulation of the serine/ threonine protein kinase cascades.

We first examined the concentrations of the G subunits involved in the regulation of AA release. Immunoblotting experiments with antisera specific for Gs, Gi1/ Gi2, Gi3/Go and Gq/G11 showed that the 3 cell lines had the same amounts of Gs migrating as 3 bands between 45 and 52 kD (Figure 4A). They also had the same concentrations of Gi1/Gi2 (40 kD) (Figure 4B), but no Gi3/Go was detected in these cell lines (Figure 4C). By contrast, the concentration of Gq was significantly higher in CFT-2 cells (45.7 ± 1.7) than in control cells (33.8 ± 4.9) (P < 0.02), or CFT-1 cells (35.6 ± 2.0) (P < 0.005) (Figure 4D). This was confirmed using another control cell line which had the same concentration of Gq/G11 as NT-1 cells (results not shown).

View larger version (34K): [in this window] [in a new window]   Figure 4.   Immunoblot analysis of G-protein  subunits in NT-1 (control), CFT-1 (non-F508) and CFT-2 (F508) epithelial cell lines. Membrane proteins (100 µg) were separated on 10% SDS-PAGE, transferred to nitrocellulose, blotted with specific antisera and quantified by densitometry: (A) Gs: NT-1, 104.3 ± 0.8; CFT-2, 105.4 ± 11.0; CFT-1, 104.7 ± 0.7; (B) Gi1 and Gi2: NT-1, 125.9 ± 4.1; CFT-2, 138.2 ± 9.2; CFT-1, 120.7 ± 4.3; (C) Gi3 and Go; (D) Gq: NT-1, 33.8 ± 4.9; CFT-2, 45.7 ± 1.7; CFT-1, 35.6 ± 2.0. The results are the means ± SD of three independent experiments.

The question of whether the activation of cPLA₂ by bradykinin was due to increased hydrolysis of phosphatidylinositol diphosphate by PLC, or to increased stimulation of protein kinase C in F508 CF cells (CFT-2) was next addressed. We have shown previously that the two CF cell lines, CFT-1 and CFT-2, have the same number of specific binding sites for bradykinin (16); we therefore looked for a link between the increased expression of Gq/G11 in CFT-2 cells and greater stimulation of the PLC/protein kinase C pathway. PLC activation was followed by measuring the IP₃ produced by stimulation with bradykinin. Transient peaks were observed in all three cell lines. The maximal level was reached after 5 s in CFT-2 cells and after 30 s in control and CFT-1 cells. The peak decreased after 60 s in the three cell lines and reached the resting level after 2 min. The maximal concentration in F508 epithelial cells was 3 times greater than in the control cells or in S549N-N1303K CF cells (Figure 5). These results indicate that PLC activity in CFT-2 cells was stimulated by bradykinin earlier and more than in CFT-1 cells or control cells.

View larger version (22K): [in this window] [in a new window]   Figure 5.   Stimulation of phosphatidylinositol diphosphate hydrolysis by bradykinin. Cells were incubated with 1 µM bradykinin for the indicated times. The IP3 content was measured using a competitive protein binding assay. Data are representative of two separate experiments.

The link between an increased Gq/G11 content and an increased stimulation of PLC and PLA₂ in response to bradykinin is strengthened by the effect of other agonists binding to receptors coupled to Gi or Gq proteins. The stimulation of purinergic receptor by ATP which was Pertussis toxin-insensitive caused a greater stimulation of AA release from CFT-2 cells than from CFT-1 or NT-1 cells. In contrast, stimulation of receptors known to be coupled to both Gi and Gq (PAF acether, thrombin) (35, 36) did not stimulate the release of more AA from CFT-2 cells than from control or CFT-1 cells (data not shown).

Effect of Thapsigargin and Bradykinin on Intracellular Calcium and AA Release

To check whether the increased PLC activation could lead to a better stimulation of membrane cPLA₂ and AA release, we stimulated the cells with calcium ionophore A23187. Surprisingly, 1 µM A23187 had no effect on the release of radioactivity from prelabeled cells, although it increased intracellular Ca²⁺ level measured by fluorescence of Fura-2 (unshown results). TPA alone, or in combination with A23187, weakly stimulated [³H]AA release, but this was much lower than the release from CFT-2 cells induced by bradykinin (unshown results). This seems to indicate that the protein kinase C isoforms sensitive to stimulation by TPA are not involved in the bradykinin-stimulated pathway in these cells.

Incubation of cells with thapsigargin, which inhibits calcium reuptake into intracellular stores, induced a stimulation of AA release in the same range of that observed with bradykinin. The stimulation was higher in CFT-2 and CFT-1 cells when compared with NT-1 cells (Table 2). However, the intracellular calcium increase produced by thapsigargin was much higher in CFT-2 cells than the one produced by bradykinin. From these experiments, we can conclude that calcium increase results in an increase of AA release, but to a different extent; therefore intracellular calcium increase by bradykinin does not fully explain the hypersensitivity of CFT-2 cells.

Activation of MAP Kinase by Bradykinin

We next explored the involvement of the other main mechanism leading to stimulation of cPLA₂. The phosphorylation of cPLA₂ is increased during cell activation (37, 38). This phosphorylation slows the electrophoretic migration of cPLA₂ due to the phosphorylation of Ser 505 by MAP kinase (22). MAP kinase is activated by a protein kinase cascade which is switched on by either protein kinase C-dependent, or -independent pathways (39). It has been shown recently that the MAP kinase cascade might be activated by Gq/11 protein-coupled receptors via a pathway involving subunits and the Src-related tyrosine kinase Lyn (40). MAP kinase is rapidly activated in response to many extracellular signals by phosphorylation on both threonine and tyrosine residues (41), resulting in an electrophoretic shift (42). The cytosolic fractions from cells treated with bradykinin for 1 or 2 min were therefore immunoblotted using an anti-MAP kinase (anti-p42 and p44) and anti-phosphotyrosine antibodies. Positive controls were obtained by treating the cells with fetal calf serum for 5 min. Western blot experiments using anti-MAP kinase antibody revealed a mobility shift of p42 MAP kinase corresponding to the pp42 phosphorylated form in all three cell lines incubated with bradykinin or fetal calf serum (Figure 6A). The blots revealed by anti-phosphotyrosine antibody had increased labeling of pp42 in all three cell lines incubated with bradykinin or fetal calf serum (Figure 6B). Consistent with the earlier activation of bradykinin-stimulated PLC, the maximal stimulation of phosphotyrosine labeling and shift was greater and occurred earlier in CFT-2 cells (1 min) than in CFT-1 and NT-1 cells (2 min). These changes in tyrosine phosphorylation of p42 MAP kinase were confirmed by parallel changes of MAP kinase activity tested on MBP peptide (95-98) (28) (Figure 6C). However, both the shift and tyrosine phosphorylation and MAP kinase activity returned to the basal level after incubation with bradykinin for 5 min. Pretreatment of cells with a previously described MEK inhibitor, PD098059, did not inhibit AA release nor MAP kinase activation induced by bradykinin, except at very high doses (100 µM) which appeared to be cytotoxic.

View larger version (29K): [in this window] [in a new window]   View larger version (39K): [in this window] [in a new window]   Figure 6.   MAP kinase activation by bradykinin. Confluent cells were incubated with 1 µM bradykinin for 1 min (Bk1'), 2 min (Bk2') or with 5% fetal calf serum (FCS) or vehicle for 5 min. Cell extracts were resolved by 8-15% continuous SDS-PAGE followed by immunoblotting with an anti-MAP kinase antibody (A), or an anti-phosphotyrosine antibody (B). Arrows indicate the positions of unphosphorylated (p42) and phosphorylated (pp42) forms of MAP kinase. (C) Confluent cells were incubated with 1 µM bradykinin for 1 min (black bars), 2 min (white bars), 5 min (dark grey bars), or with 5% fetal calf serum (light grey bars) or vehicle for 5 min. MAP kinase activities are expressed as % of unstimulated cells. Asterisks: *P < 0.05 indicates statistical difference when compared with bradykinin 1 min for CFT-2 cells (NT-1, P = 0.0339; CFT-1, P = 0.0339), using the Mann-Whitney test.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

We have previously shown that F508 CF cells from different patients are more sensitive to bradykinin, and respond by releasing AA (16). Surprisingly, neither the tracheal epithelial cells nor the skin fibroblasts from patients bearing a double heterozygous mutation of CFTR showed this dysregulation. Immortalized epithelial cell lines having the F508 mutation of CFTR (CFT-2) and a double heterozygous mutation of CFTR (CFT-1) were used as models to study this dysregulation of AA release. These 2 cell lines had similar defects in chloride transport (26) and the same number of specific binding sites for bradykinin (16). The effect of the F508 mutation on AA release is not due to altered function of the protein itself, since incubating the cells at 28°C for 48 h, which restores the delivery of CFTR to the plasma membrane (30), lowered the AA release by bradykinin to the value in control cells (Table 1). This indirect evidence supports the hypothesis that both the wild type CFTR and normally processed F508 CFTR can inhibit the AA release induced by bradykinin.

The two types of PLA₂ involved in AA release by stimulated cells are a 85 kD cytosolic form and a 14 kD secreted form (for review, see 19 and 20). No secreted PLA₂ activity was detected in media from resting cells or after stimulation with bradykinin. But there is cytosolic PLA₂ in all three cell lines, which have some of the characteristics of the cPLA₂ cloned in macrophages (21) and identified in several cell types. It was specific for AA, was translocated to the membrane by Ca²⁺, and was recognized by an antibody directed against human cPLA₂.

The PLA₂ activity in the cytosol of resting CFT-2 and CFT-1 cells was much higher than in control cells. However, in contrast with the PLA₂ activity, the amount of immunoreactive cPLA₂ was the same in the cytosols of the three cell lines (Figure 3). This difference might be due to another, unidentified, cytosolic PLA₂ in the cytosol of CFT-1 and CFT-2 cells, which cannot account for the increased bradykinin-induced AA release that occurs in CFT-2 cells but not in CFT-1 cells. Another possibility is that cPLA₂-specific activity is enhanced in CFT-2 cells by activation of an upstream regulator without increased expression of the enzyme. For example, in vitro studies of phosphorylation of cPLA₂ by PKC increases the activity but has no effect on the mobility shift of cPLA₂ (23).

The increase in PLA₂ activity and immunoreactive cPLA₂ in response to bradykinin occurred only in the membranes of CFT-2 cells, and thus seemed to explain the greater AA release from these cells. Incubation of whole cells with bradykinin also increased the PLA₂ activity and the amount of cPLA₂ associated with the membrane fraction of CFT-2 cells to the same extent (compare Figures 2 and 3). Similar increases in particulate PLA₂ activity have been shown in endothelial cells stimulated with bradykinin (43).

EDTA decreased the association of cPLA₂ with the membrane of resting cells, but not with membranes of cells stimulated with bradykinin (Figure 1 and 2). Hence, stimulation by bradykinin triggers an increase in intracellular free Ca²⁺ which stimulates the association of cPLA₂ with the membrane, but binding also involves an interaction with a membrane component that anchors cPLA₂ which is maintained when Ca²⁺ is removed from the medium. No protein-protein interaction has yet been described between cPLA₂ and a membrane component. Further experiments are needed to examine this point.

The second part of the study examined the intracellular mechanisms triggered by the bradykinin receptor leading to the better association of cPLA₂ with the membrane of CFT-2 cells. Bradykinin binds to a G protein-coupled receptor in tracheal epithelial cells, which increases intracellular calcium by stimulating phosphatidylinositol-specific PLC (19). The magnitude of PLC stimulation is often related to the number of specific binding sites on the cell surface (44). But CFT-2 cells have an increased IP₃ production in response to bradykinin (Figure 5), although they have the same number of bradykinin receptors as CFT-1 cells (16). The receptor-mediated stimulation of PLC and PLA₂ is known to be mediated by two main types of heterotrimeric G protein: Pertussis toxin-sensitive Gi proteins, which stimulate PLC through the subunits, and Pertussis toxin-insensitive Gq/11 proteins, which stimulates PLC by activating the  subunit (45, 46). The concentrations of the Gs and Gi1/Gi2 subunits in the three cell lines were similar, but CFT-2 cells contained 30% more Gq/11 than did CFT-1 and NT-1 cells (Figure 4). Two lines of evidence suggest that this increased Gq/11 content may be responsible for the increased stimulation of IP3 production by bradykinin. First, in CFT-1 cells, which have no increased Gq/11 content, the increase in IP₃ in response to bradykinin is smaller than in CFT-2 cells. Second, the release of AA from CFT-2 cells is increased by a purinergic receptor coupled to Pertussis toxin-insensitive G proteins, but not by thrombin or platelet-activating factor receptors, which are coupled to both Pertussis toxin-sensitive and insensitive G proteins. However, the mechanism by which the defective processing of F508 CFTR induces the increased expression and efficiency of Gq/11-PLC stimulation is still unclear.

The increased stimulation of IP₃ production in CFT-2 cells by bradykinin could explain the increase in membrane-associated cPLA₂. This would be in agreement with the poor stimulation of AA release and the lack of increased cPLA₂ translocation in CFT-1 cells. However, when cells were homogenized in the presence of Ca²⁺, there was more PLA₂ activity associated with the membrane in CFT-2 cells than in either CFT-1 cells or control cells. This suggests that the absence of CFTR also results in the better association of cPLA₂ with a particular membrane component.

The other main mechanism involved in the receptor-mediated activation of cPLA₂ is its phorphorylation by protein kinases. We have examined the roles of the two main kinases believed to activate cPLA₂, protein kinase C (23) and MAP kinase (22). The results indicate that conventional PKCs have no influence on the differential stimulation of cPLA₂ by bradykinin. TPA alone, or together with the calcium ionophore, did not increase AA release by CFT-2 or CFT-1 cells. This result does not rule out the involvement of TPA-insensitive atypical PKCs  and  in bradykinin-stimulated PLA₂. By contrast, there was an increased shift of MAP kinase in bradykinin-treated CFT-2 cells, corresponding to pp42 and an increase in phosphotyrosine labeling of pp42 as early as 1 min. This early stimulation was confirmed by a significant increase of MAP kinase activity after 1 min (Figure 6C). These increases were slower and smaller in CFT-1 and control cells than in CFT-2 cells. The activation of MAP kinase was transient in all three cell lines and returned to the control value after incubation with bradykinin for 5 min. The time-course and intensity of MAP kinase phosphorylation reflecting MAP kinase activation (39) followed IP₃ production. This might suggest that activation of Gq/11 stimulates PLC and activates MAP kinases by a mechanism independent from PKC stimulation. Such a PKC-independent mechanism of MAP kinase stimulation has been recently shown to occur through stimulation of a Src-related tyrosine kinase by a Gq/G11-dependent pathway (40).

In conclusion, the F508 mutation of CFTR causes greater stimulation of cPLA₂ by a mechanism involving its phosphorylation, its membrane translocation, and its strong association with a still unknown membrane component. Many CF mutations of CFTR disturb the processing of the protein, which may not reach the plasma membrane. This is the case for F508 mutation, and CFTR is retained in the endoplasmic reticulum and rapidly degraded by proteases (6, 7). The presence of normal CFTR in the plasma membrane of NT-1 cells seems to correlate negatively with the irreversible association of cPLA₂ induced by bradykinin. This is confirmed by the effect of low temperature, which restores normal processing of F508 CFTR to the plasma membrane and decreases the release of AA by bradykinin. This might explain why bradykinin, although it has functional receptors able to stimulate phosphatidylinositol hydrolysis and to activate MAP kinase, is not good at stimulating AA release from membrane phospholipids of control cell lines. The heterozygous mutation of CFTR, with mutations on another part of the protein, might be processed normally and have another phenotype which does not include abnormal eicosanoid production.

The decreased amount of F508 CFTR in plasma membrane is caused by a defective processing of the mutated protein (6). This defective processing seems to induce various cellular responses (47, 48), which might include the stimulation of pathways activating cPLA₂. CFTR normally processed to the cell membrane seems to inhibit the production of lipid mediators in response to bradykinin. The mechanism of this inhibition is still unclear, but might involve an annexin-like domain in CFTR, which could inhibit the interaction of cPLA₂ with the membrane (49). The increased AA release by F508 CF cells may contribute to the greater sensitivity of these patients to the inflammatory reaction accompanies pulmonary infection. This defect, together with the recently described change in bacterial receptors (12), might be part of the increased sensitivity of CF patients to pulmonary infection.