Interleukin-1β Differentially Regulates β2 Adrenoreceptor and Prostaglandin E2-mediated cAMP Accumulation and Chloride Efflux from Calu-3 Bronchial Epithelial Cells

Here we tested the effect of interleukin-1β, a pro-inflammatory cytokine, on cAMP accumulation and chloride efflux in Calu-3 airway epithelial cells in response to ligands binding to adenylyl cyclase-coupled receptors such as the β2 adrenoreceptor and EP prostanoid receptors. Interleukin-1β significantly increased isoprenaline-induced cAMP accumulation by increasing β2 adrenoreceptor numbers via a protein kinase A-dependent mechanism. In contrast, interleukin-1β significantly impaired prostaglandin E2-induced cAMP accumulation by induction of cyclo-oxygenase-2, prostaglandin E2 production, and a resulting down-regulation of adenylyl cyclase. The cAMP changes were all mirrored by alterations in chloride efflux assessed using the fluorescent chloride probe N-(ethoxycarbonylmethyl)-6-methoxyquinolinium bromide with interleukin-1β increasing chloride efflux in response to isoprenaline and reducing the response to prostaglandin E2. Studies with glibenclamide confirmed that chloride efflux was via the cystic fibrosis transmembrane conductance regulator. Calu-3 expresses EP4 receptors, but not EP2, and receptor expression is reduced by interleukin-1β. Collectively, these results provide mechanistic insight into how interleukin-1β can differentially regulate cAMP generation and chloride efflux in response to different adenylyl cyclase-coupled ligands in the same cell. These findings have important implications for diseases involving inflammation and abnormal ion flux such as cystic fibrosis.

Cystic fibrosis (CF) 1 is a multisystem disorder and the most commonly occurring lethal autosomal recessive hereditary disorder in Caucasian populations. It is caused by mutations in a 230-kb gene located on chromosome seven that codes for CFTR (1). CFTR functions in the main as a cAMP-regulated chloride channel but also regulates the activity of other ion channels and transporters such as apical Enac, basolateral Na/K ATPase (2,3), and the outwardly rectifying chloride channel (4), and can transport bicarbonate (5). Progression of lung disease due to chronic infection is the main cause of morbidity and mortality in CF (6). Alterations in ion flux in cystic fibrosis lungs is thought to be central in the pathogenesis of lung disease, although the mechanism is debated. Several hypotheses have been suggested with the biochemical defect being linked to impaired antibacterial activity of ␤-defensins (7), mucociliary clearance (8), and altered sialisation of bacterial receptors (9).
Although defective CFTR predisposes to inflammation, the reverse situation, namely whether inflammation can impair CFTR function, has not been extensively studied. This would be important as it would set up a cycle that would further damage the defense properties of the bronchial epithelium. Several pieces of evidence would be consistent with this hypothesis.
Residual chloride efflux declines as lung function falls (10), there is considerable variation in clinical outcome among patients with the same genotype (11), and variable chloride conductance in rectal epithelium has been observed between monozygotic and dizygotic twins (12). Collectively, these studies raise the intriguing possibility that the airway environment can modify the underlying biochemistry.
The airways in CF are characterized by neutrophil-dominated inflammation with high levels of pro-inflammatory cytokines such as IL-1␤, IL-6, IL-8, and tumor necrosis factor ␣ in sputum, serum, and bronchial lavage samples (13). IL-1␤ is particularly interesting as studies from our own group and others have shown that it can reduce cAMP accumulation in response to ligands that bind to adenylyl cyclase-linked receptors in other biological systems (14). Few studies have looked at the functional consequences of these changes. Here we have shown that IL-1␤ alters cAMP-mediated chloride efflux in Calu-3 airway epithelial cells. IL-1␤ had differential effects depending on whether cells were stimulated with an EP prostanoid or ␤ adrenoreceptor receptor ligand. IL-1␤ reduced chloride efflux in response to PGE 2 but enhanced the response to isoprenaline, a ␤ 2 adrenoreceptor agonist.
Mechanistic studies showed that the effect was mediated by IL-1␤ having differential effects on cAMP accumulation to the two agents. The down-regulation of cAMP accumulation and chloride efflux in response to prostaglandin E 2 (PGE 2 ) was mediated by an autocrine loop involving induction of cyclooxygenase 2 (COX-2) and endogenous PGE 2 production.
In contrast, the up-regulation of isoprenaline-induced cAMP accumulation was prostanoid-independent, was at least in part mediated by protein kinase A, and involved increased expression of ␤ 2 adrenoreceptors. These studies provide important mechanistic and functional insights into how different adenylyl cyclase-coupled receptors can be differentially regulated by cytokines in the same cell type. These effects may play an important role in the pathophysiology of CF lung disease and may have relevance to its therapy. Similar biochemical processes may be important in diseases of other mucosa such as the gut.

MATERIALS AND METHODS
Experiments were performed on cultured Calu-3 cells, a transformed bronchial epithelial cell line originally derived from a 25-year-old male in 1975 (obtained from LGC Ltd., Teddington, UK). This cell line shows cAMP-dependent chloride flux through CFTR that is expressed in high levels (15,16). Cells were incubated in 24-well plates (at passage 3-7) in 10% fetal calf serum-Eagle's minimum essential medium, supplemented with non-essential amino acids and sodium pyruvate, in humidified 5% CO 2 -95% air at 37°C, and the medium was changed every 2-3 days. Cells were cultured to confluence (typically 7-10 days) and growth arrested in fetal calf serum-free Eagle's minimum essential medium for 24 h prior to each experiment. Cells were then treated with either 200 pM IL-1␤ or 1 M PGE 2 for a further 24 h. Appropriate controls were performed in all experiments; cells were treated with vehicle-only for the same times. Inhibitors were added to the treated cells at the same time as the cytokines and mediators in certain experiments. To examine the role of cyclo-oxygenase isoforms, we used 1 M indomethacin, a broad spectrum COX inhibitor, and 1 M NS-398, a selective COX-2 inhibitor. The role of protein kinase A was examined using the specific and selective inhibitor H-89.
Cyclic AMP Accumulation-Immediately before each experiment, medium was removed and the cells washed twice with sterile phosphate-buffered saline. This was then replaced by 500 l of serum-free Eagle's minimum essential medium containing 1 mM isobutylmethylxanthine to prevent cAMP breakdown by phosphodiesterases. Isoprenaline, PGE 2, forskolin, and the selective EP 2 (ONO-AE1 259) or EP 4 (ONO-AE1 329) agonist was then added to control and treated cells and incubated for 10 min. This time was chosen based on our previous studies. Experiments were terminated by the addition of 100 l of ice-cold 30% trichloroacetic acid. The resulting solution was transferred to centrifuge tubes kept at 4°C, and the trichloroacetic acid removed by amine-freon extraction. The cAMP content of the extract was deter-mined using a protein binding assay (17). The bound 3 H cAMP was measured using the Tri-Carb 2100TR liquid scintillation analyzer (Packard Bioscience LTD., Pangbourne, Berkshire, UK). PGE 2 Assay-Cell supernatants were stored at Ϫ20°C until the determination of PGE 2 content by radioimmunoassay as previously described (17).
␤ 2 Adrenoreceptor Assay-␤ 2 adrenoreceptor numbers were measured using a radioligand binding assay (18). Cells were incubated with increasing concentrations of [ 3 H]CGP12177 (doubling concentrations from 0.0625 to 8 nM), a ␤ 2 adrenoreceptor ligand, in serum-free medium for 1 h. At the end of this time supernatant was removed and the cells washed twice in sterile phosphate-buffered saline. Cells were then lysed with 0.5 M sodium hydroxide, and the bound [ 3 H]CGP12177 was measured using the Tri-Carb 2100TR liquid scintillation analyzer. These results were taken to indicate the total binding of [ 3 H]CGP12177. The procedure was repeated with 0.1 M ICI 118551, a potent specific ␤ 2 adrenoreceptor antagonist, added to all the ascending concentrations of [ 3 H]CGP12177 to specifically occupy all available ␤ 2 adrenoreceptor sites. The resulting measurements of [ 3 H]CGP12177 binding are therefore representative of nonspecific binding. Specific binding was calculated from the difference between the total and nonspecific binding at each concentration of [ 3 H]CGP12177. Protein content was measured using the method of Bradford (19).
Assay of Chloride Flux-Chloride flux was measured using the fluorescent chloride probe MQAE (20). Cells were grown to confluence in 96-well flat-bottomed and black-walled plates (Costar, UK). Once confluent, half the cells were treated with serum-free medium plus IL-1␤ and half with serum-free medium plus vehicle alone.
After 4 h, 5 l of 8 mM MQAE was added to each well to bring the total well volume to 100 l. 20 h later the medium was removed, and each well was washed three times with chloride-containing buffer, tipping onto blotting paper as before. The plates were incubated with 100 l of chloride-containing buffer (to induce chloride channel activation before initiating chloride efflux) for 10 min at room temperature. During experiments with glibenclamide (200 M), glibenclamide was added to selected wells during this period. The plates were then read on a Wallac Victor fluorescence plate reader (LKB Wallac, Turku, Finland) Both agonists caused a concentration-dependent increase in cAMP accumulation; this was significantly reduced by pretreatment with 200 pM IL-1␤ for 24 h when cells were stimulated with PGE 2 (B). In contrast cAMP accumulation in response to ISO was significantly enhanced (A). Data expressed as mean Ϯ S.E. (n ϭ 12 for each group). ***, p Ͻ 0.0001; *, p Ͻ 0.05.
to give a baseline reading before agonists were added (excitation wavelength 360 nm, emission wavelength 460 nm). The buffer was then removed as before and replaced by 100 l of chloride-free buffer Ϯ agonists to be tested. The plate was then read every minute for 16 min at the same settings. Results are expressed as change in fluorescence from baseline against time. Increasing fluorescence represents increasing chloride efflux from cells.
RNA Isolation and Reverse Transcriptase Polymerase Chain Reaction-Cells were grown to confluence in 6-well plates, and growth was arrested for 24 h and then treated for 0

FIG. 3. The effects of the broad spectrum COX inhibitor indomethacin and the COX-2-selective inhibitor NS-398 on changes in cAMP accumulation induced by IL-1␤.
A, COX inhibitors have no effect on the increase in cAMP accumulation in response to isoprenaline. B, the reduction in PGE 2 -induced cAMP accumulation is abolished by both agents. These results suggest that the reduced cAMP response to PGE 2 is caused by induction of COX-2 and production of endogenous PGE 2 . In contrast, the enhanced response to ISO appears to be COX independent. Data expressed as mean Ϯ S.E. (n ϭ 8 -16 for each group). ***, p Ͻ 0.0001; **, p Ͻ 0.001. g/ml ethidium bromide and visualized using ultraviolet illumination and the GeneGenius gel documentation and analysis system (Syngene).
Western Blot Analysis-Western blotting for COX-2 and ␤ 2 adrenoreceptor was performed as previously described (17).
Statistical Analyses-Data are expressed as means and S.E. from n determinations. Statistical analysis was performed with the software program GraphPad Prism 4. For experiments measuring cAMP and PGE 2 release, unpaired two-tailed Student's t test was used to determine the significance of differences between means. For experiments measuring chloride efflux, representative experiments are shown. Three time points were chosen for subsequent analysis as summary statistics. Statistical significance was determined by efflux for each condition at these time points. Unpaired two-tailed Student's t tests were used to determine significance between these sets of observations. In assays measuring ␤ 2 adrenoreceptor numbers, B max (a measure of receptor number) and the dissociation constant K d were calculated using non-linear regression (GraphPad Prism 4). In all experiments p values Ͻ0.05 were accepted as statistically significant. PGE 2 also caused a concentration-dependent increase in cAMP production in control cells significant from 0.01 M. IL-1␤ (200 pM) pretreatment for 24 h markedly attenuated the response to all PGE 2 concentrations (Fig. 1B). We next went on to determine the role of prostanoids in the effects of IL-1␤.

IL-1␤ Increases cAMP Formation in
IL-1␤ Induces COX-2 and Increases PGE 2 Production in Calu-3 Cells-IL-1␤ (200 pM) caused a marked increase in PGE 2 production that was time dependent and significant ( Fig.  2A). The increase in PGE 2 was abolished when cells were pretreated with the broad spectrum COX inhibitor indomethacin (1 M) or the COX-2-selective agent NS-398 (1 M) (Fig. 2B). We have previously shown that this concentration of NS-398 abolishes COX-2-mediated PGE 2 production with no effect on COX-1 in airway cells. Western blotting showed that IL-1␤ induced COX-2 with protein levels that increased from 4 h and remained increased at 24 h (Fig. 2C).
Inhibitors of Prostanoid Synthesis Inhibit the Effects of IL-1␤ on PGE 2 -induced cAMP Accumulation but Not ISO-induced cAMP Accumulation-To determine whether prostanoids were responsible for the effects of IL-1␤ on ISO-and PGE 2induced cAMP accumulation, we studied the effect of the broad spectrum COX inhibitor indomethacin and selective COX-2 inhibitor NS-398 on IL-1␤-induced changes in cAMP production. These agents (both at 1 M) had no effect on the increase in ISO-stimulated cAMP production seen after IL-1␤ ( Fig. 3A), suggesting that this effect occurs via a prostanoidindependent mechanism.
In contrast, COX inhibitors completely abolished the downregulation of PGE 2 -induced cAMP accumulation seen with IL-1␤ (Fig. 3B). This suggests that the attenuated cAMP response to exogenous PGE 2 produced by IL-1␤ occurs via induction of COX-2 and production of endogenous PGE 2 .
Exogenous PGE 2 Mimics the Effect of IL-1␤ on PGE 2 -mediated cAMP Accumulation but Not ISO-mediated cAMP Accumulation-Further evidence was sought for a prostanoid-dependent mechanism by applying exogenous PGE 2 (1 M for 24 h) to mimic the effects of IL-1␤. Exogenous PGE 2 downregulated PGE 2 -induced cAMP accumulation in a similar manner to IL-1␤ (Fig. 4A), thus providing further evidence that the effects of IL-1␤ on PGE 2 -induced cAMP accumulation are via COX-2 induction and endogenous PGE 2 production.
In contrast to the effects of IL-1␤, exogenous PGE 2 decreased ISO-induced cAMP accumulation (Fig. 4B). This provided further evidence that the effects of IL-1␤ on ISO-mediated cAMP synthesis are prostanoid independent.
IL-1␤ Reduces cAMP Accumulation in Response to Forskolin-To study the site of the effect of IL-1␤ on PGE 2 -induced cAMP accumulation, we studied the effect of forskolin, a direct adenylyl cyclase activator that bypasses G protein-coupled receptors. Forskolin concentration dependently increased cAMP accumulation in control cells significant from 0.1 M. IL-1␤ (200 pM) for 24 h significantly attenuated the response to 1 M forskolin (Fig. 5A), suggesting that adenylyl cyclase is downregulated by IL-1␤. Consistent with this, exogenous PGE 2 also decreased forskolin-induced cAMP accumulation (Fig. 5B).

IL-1␤ Increases ISO-induced cAMP Accumulation by
Increasing ␤ 2 Adrenoreceptor Numbers-To determine the mechanism whereby IL-1␤ increases ISO-induced cAMP accumulation, we studied the effects of IL-1␤ on ␤ 2 adrenoreceptor numbers. There was a consistent 3-4-fold increase in receptor number after incubation with IL-1␤ for 24 h. The mean B max for untreated cells was 26.5 (Ϯ 7.2) fmol/g protein. This rose to 76.1 (Ϯ 10.5) fmol/g protein after 24 h of treatment with IL-1␤. Representative whole cell binding data are shown in Fig.  6, A and B. Fig. 6C shows the receptor density for each condition. These observations were confirmed by Western blotting. Fig. 6D shows the time-dependent increase in ␤ 2 adrenoreceptor protein.
Calu-3 Cells Express EP 4 but Not EP 2 Receptors; EP 4 Expression Is Reduced by IL-1␤-PGE 2 -stimulated cAMP accumulation occurs via stimulation of the G s -coupled prostanoid receptors EP 2 and EP 4 . We therefore examined the effects of agonists and antagonists at these receptors on PGE 2stimulated cAMP accumulation. Stimulation with EP 4 agonist (10 nM) led to a marked increase in cAMP accumulation after 10 min that was equivalent to that seen in response to PGE 2 stimulation. Exposure to the same concentration of EP 2 agonist did not lead to an increase in cAMP accumulation. Pretreatment of cells for 1 h with a selective EP 4 antagonist abolished cAMP accumulation in response to both PGE 2 and EP 4 agonist (Fig.  7A). These experiments were repeated using chloride efflux as the end point. The same effects were observed (data not shown). We then examined the expression of EP 2 and EP 4 receptor messenger RNA in these cells using reverse transcription PCR. Cells were treated for between 0 and 24 h with IL-1␤ prior to RNA extraction. EP 2 receptor mRNA was not detected at any time point. EP 4 receptor mRNA was present at all time points and was reduced by IL-1␤ treatment after 24 h (Fig. 7B). Fig. 7C shows densitometry data from three separate experiments.

The PKA Inhibitor H-89 Abolishes the Effect of IL-1␤ on ISO-induced cAMP Accumulation but Has No Effect on That
Caused by PGE 2 -Experiments were performed to determine whether the effects of IL-1␤ on cAMP accumulation were me- diated by protein kinase A. We used the inhibitor H-89 at a concentration of 10 M and found that it abolished the increase in cAMP accumulation due to IL-1␤ when cells were stimulated with ISO (Fig. 8A). In contrast, H-89 had no effect on the attenuated response to PGE 2 (Fig. 8B). These results suggest IL-1␤ enhances cAMP accumulation to ␤ 2 adrenoreceptor ago- nists via PKA but that the attenuated response to prostanoids is PKA independent.
IL-1␤ Increases Chloride Efflux from Cells in Response to ISO but Reduces Efflux in Response to PGE 2 -We next went on to determine whether the effects of IL-1␤ on cAMP were mirrored by effects on cAMP-mediated chloride flux. Isoprenaline and PGE 2 both caused concentration-dependent increases in chloride efflux (Fig. 9, A and B). From these time course experiments, time points of 3, 5, and 16 min were chosen for summary analysis. Chloride efflux was blocked at all summary time points by 200 M glibenclamide, confirming that it was CFTR-mediated (data not shown). IL-1␤ (200 pM) pretreatment for 24 h increased 10 M ISO-induced chloride efflux but reduced 1 M PGE 2 -mediated chloride efflux. These changes mirrored those seen in experiments where cAMP production was measured. The effects of IL-1␤ on ISO-and PGE 2 -mediated chloride efflux are shown at these points (Figs. 10 and 11).
COX Inhibitors Abolish IL-1␤-induced Changes in Chloride Efflux in Response to PGE 2 -Experiments were performed to determine whether the mechanism of the effects of IL-1␤ on chloride efflux in response to PGE 2 were the same as those seen on cAMP accumulation. The COX-2-specific inhibitor NS-398 was able to completely abolish the attenuated response at all summary time points (Fig. 12). The broad spectrum COX inhibitor indomethacin had the same effect (data not shown). Furthermore, PGE 2 pretreatment mimicked the effect of IL-1␤ (Fig. 13).

IL-1␤ Has No Effect on Cell
Proliferation-To make sure that none of the effects in these studies was an artifact due to alterations in cell numbers, we performed 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assays of cells after 24 h of treatment with IL-1␤ at the concentration used in other experiments. There was no significant difference between treated and control cell populations (data not shown). DISCUSSION The main findings of these studies are that IL-1␤ has differential effects on ISO-and PGE 2 -mediated cAMP accumulation and chloride efflux in Calu-3 cells. IL-1␤ enhanced ISO-mediated cAMP accumulation via an increase in ␤ 2 adrenoreceptor numbers. In contrast, IL-1␤ reduced PGE 2 -mediated cAMP accumulation by induction of COX-2, production of endogenous PGE 2 , down-regulation of adenylyl cyclase, and down-regulation of EP 4 receptor expression. These findings are novel in showing that IL-1␤ can differentially effect the responses to different adenylyl cyclase-coupled receptor ligands in the same cell. Furthermore, these findings have important functional consequences. The effects of IL-1␤ on ISO-and PGE 2 -mediated chloride flux mirrored its effects on cAMP accumulation with the response to ISO being enhanced and the response to PGE 2 being attenuated. These findings have implications for the understanding of the pathophysiology of cystic fibrosis lung disease and may also be relevant for disease processes at other mucosal surfaces such as the gut. Our results suggest that pro-inflammatory cytokines such as IL-1␤ present in CF airways may reduce PGE 2 -mediated chloride flux. This might further impair defective chloride flux in cystic fibrosis and exacerbate the abnormalities in airway surface liquid composition that occur as a result of CFTR dysfunction. This effect could also cause abnormal fluid movement in non-cystic fibro-FIG. 8. The PKA inhibitor H-89 (10 M) abolishes the enhanced accumulation of cAMP in response to ISO seen with IL-1␤ (A). In contrast, H-89 has no effect on the attenuated response to PGE 2 (B). These experiments suggest that IL-1␤ increases the cAMP response to ISO via PKA. In contrast, its effect on PGE 2 -stimulated cAMP accumulation is PKA independent. Each data set is the mean Ϯ S.E. of eight observations. ***, p Ͻ 0.0001; **, p Ͻ 0.001.

FIG. 9. Isoprenaline (A) and PGE 2 (B) cause a concentrationdependent increase in chloride efflux from Calu-3 cells.
Each data point is the mean Ϯ S.E. of eight observations. sis bronchiectasis that might worsen the disease.
It is well established that CF lungs contain high levels of pro-inflammatory cytokines such as IL-1␤ (13). As IL-1␤ can regulate adenylyl cyclase function in some biological systems, we hypothesized that pro-inflammatory cytokines such as IL-1␤ would alter cAMP accumulation in response to ligands binding to adenylyl cyclase-coupled receptors and, therefore, CFTR function. The CFTR channel is controlled by receptors that couple to guanine nucleotide-binding proteins and adenylyl cyclase leading to production of cAMP (15,16). Cyclic AMP then activates protein kinase A, which phosphorylates the regulatory domain of CFTR. Airway epithelial cells express several adenylyl cyclase-coupled receptors, including ␤ 2 adrenoreceptors and EP prostanoid receptors whose activation promotes CFTR-mediated chloride flux (21). Although there is some evidence that cytokines can affect the function of components of this signaling pathway in airway cells, including epithelial cells (22) and airway smooth muscle cells (14), little is known about the functional consequences of these effects, particularly on CFTR-mediated chloride efflux in airway epithelial cells.
It is generally accepted that the main physiological regulator of CFTR is cAMP. However, surprisingly little information exists about the pathways that lead to physiologically important cAMP production in airway epithelial cells. Prostanoid receptors and adrenoreceptors are likely to be the main candidate sources of signal transduction as stimulation of these receptors causes marked changes in cAMP. Adrenergic innervation of the lung is relatively sparse (23), suggesting that circulating catecholamines may activate ␤ 2 adrenoreceptors on epithelial cells. As PGE 2 is produced locally in large quantities in the lung (24), it is likely that its actions at EP 2 and EP 4 prostanoid receptors are important in increasing cAMP-mediated chloride flux.
We chose Calu-3 cells for our studies as they have previously been shown to possess both CFTR-mediated chloride flux and ␤ 2 adrenoreceptors (15,16). Our experiments have shown that these cells express EP 4 prostanoid receptors. These cells thus provide a useful model system to study regulation of these pathways in airway epithelial cells. Available CF airway epithelial cell lines were not used, as they do not have sufficient cAMP-mediated chloride flux to be able to measure a further reduction by IL-1␤. This is a property of these cell lines and does not represent the properties of CF airway epithelial cells in vivo, which do show residual chloride flux. Cells responded to both isoprenaline and PGE 2 with rapid accumulation of cAMP. We found that IL-1␤ enhanced ISO-induced cAMP accumulation but paradoxically reduced PGE 2 -induced cAMP accumulation. The magnitude of these effects differed, with the attenuation of PGE 2 -induced cAMP accumulation being the more marked. We then performed mechanistic studies to characterize these effects further. Recent studies in human airway smooth muscle (25) and airway epithelial (26) cells have shown that the effects of IL-1␤ on the release of other cytokines and mediators can occur via the action of protein kinase A. We therefore went on to see whether a similar mechanism could underlie the effects of IL-1␤ on Calu-3 cells. Studies with the PKA inhibitor H-89 suggested that PKA plays an important role in the effect of IL-1␤ on ISO-induced cAMP accumulation but that its effect on prostanoid-induced cAMP accumulation is PKA independent. We measured ␤ 2 adrenoreceptor numbers using a ligand binding assay. We found a large increase in ␤ 2 adrenoreceptor numbers after 24 h of pretreatment with IL-1␤, suggesting that this was at least partly responsible for the increased ISO-induced cAMP accumulation. These findings were then confirmed by Western blotting. This result is similar to studies in BEAS2-B human airway epithelial cells where IL-1␤ caused a 2.5-fold increase in receptor number (22) and to a recent study in fetal guinea pig lung homogenates (27). Our studies extend previous studies to determine whether the upregulation of ␤ 2 adrenoreceptor numbers and cAMP accumulation has any functional significance. To do this we used MQAE, a fluorescent chloride probe. We found that ISO caused a marked dose-dependent increase in chloride efflux from the cells. Studies with glibenclamide, a CFTR inhibitor, showed that ISO-induced chloride efflux was CFTR-mediated.
Treatment with drugs that act at the ␤ 2 adrenoreceptor (e.g. Salbutamol and Salmeterol) is common in cystic fibrosis, largely in patients who show symptoms of reversible airways obstruction. Use of these drugs has been shown to lead to both short and long term improvements in lung function (28,29). This improvement is mediated by increases in cellular cAMP, which in turn excerpts beneficial effects via relaxation of airway smooth muscle and possibly increased activation of residual CFTR function in epithelial cells. The effects of IL-1␤ on ␤ 2 adrenoreceptor number and adenylyl cyclase function would be likely to increase the effectiveness of these agents on residual CFTR function in CF.
In marked contrast to the effects of IL-1␤ on ISO-induced cAMP accumulation, it attenuated PGE 2 -induced cAMP accumulation and chloride efflux. We hypothesized that endogenous prostanoids produced in response to IL-1␤ might be responsible for this effect. Cyclo-oxygenase is the enzyme that converts arachidonic acid into PGH 2 , which is subsequently converted to prostanoids by specific synthases. COX exists in three isoforms (30). COX-1 produces physiological levels of prostanoids (31), whereas COX-2 is induced in inflammatory conditions (32). The recently identified COX-3 is a splice variant of COX-1, whose function is unclear (33). We found that IL-1␤ induced COX-2 and caused an increase in PGE 2 release. We measured PGE 2 as it is the most abundant arachidonic acid metabolite in human airway epithelium (24). Interestingly, recent research has demonstrated that CF patients have abnormally high levels of arachidonic acid and reduced docosahexanoic acid, a fatty acid with anti-inflammatory properties (34). To determine whether IL-1␤-induced PGE 2 production was responsible for down-regulation of PGE 2 -induced cAMP accumulation by chronic desensitization, we used indomethacin, a broad spectrum COX inhibitor, and NS-398, a specific COX-2 inhibitor. We have previously shown that the concentration of NS-398 used inhibits 90% of COX-2-mediated prostanoid production, with no effect on COX-1 (35). Consistent with a major role for COX-2 in this process, both agents abolished the effects of IL-1␤ on PGE 2 -induced cAMP production and chloride efflux.
To assess the possible site of the desensitization, we performed studies using forskolin, a direct activator of adenylyl cyclase, to stimulate cAMP accumulation. We found that IL-1␤ reduced forskolin-induced cAMP accumulation, suggesting that IL-1␤ was causing down-regulation of adenylyl cyclase. Consistent with this, exogenous PGE 2 also reduced forskolininduced cAMP accumulation. In addition to its effects on adenylyl cyclase, we have also shown that IL-1␤ reduces the expression of EP 4 prostanoid receptors. Presumably the reason that IL-1␤ did not cause a similar reduction in ISO-induced cAMP accumulation but paradoxically increased it was that the increased ␤ 2 adrenoreceptor numbers were sufficient to overcome the reduced adenylyl cyclase activity.
Data regarding the effect of other cytokines on CFTRmediated chloride flux in epithelial cells are sparse. Interferon-␥ reduced cAMP-dependent chloride efflux in primary cultures of human bronchial epithelial cells stimulated with the cAMP analogue 8-(4-chlorophenylthio) (CPT)-cAMP (36). In bile duct epithelium, a combination of IL-6, interferon-␥ 16 , . These data suggest that the reduced cAMP responses to PGE 2 seen with IL-1␤ treatment lead to reduced chloride efflux and that this effect is mediated by the same COX-2-dependent mechanism. Data expressed as mean Ϯ S.E. (n ϭ 16 for each group). ***, p Ͻ 0.0001.
FIG. 14. A schematic diagram summarizing an important conclusion of this study. IL-1␤ sets up a positive feedback loop whereby autocrine production of PGE 2 down-regulates subsequent cAMP production and chloride efflux in response to endogenous PGE 2 . It is likely that this would lead to further abnormalities in airway surface liquid composition and thus predispose the airway to further infection and production of pro-inflammatory cytokines. tumor necrosis factor ␣, and IL-1␤ reduced chloride efflux in response to forskolin, but individual cytokines had no effect (37). It is likely that the actions of IL-1␤ in attenuating PGE 2 -mediated cAMP accumulation and chloride flux would play a role in the pathogenesis of cystic fibrosis lung disease. The abnormal airway surface liquid due to CFTR dysfunction in the CF lung leads to infection and inflammation. This inflammatory response with increased levels of IL-1␤ would down-regulate adenylyl cyclase and the action of endogenous PGE 2 to increase cAMP and activate CFTR. This would further impair electrolyte abnormalities, predisposing to further infection and inflammation by setting up a positive feedback mechanism (Fig. 14). Strategies that break this loop would be useful therapeutically. Interestingly, clinical trials have shown the broad spectrum COX inhibitor ibuprofen can delay CF lung disease progression (38), consistent with these observations. The effect of IL-1␤ in regulating chloride flux may also be important in other mucosal sites. For example, CFTR is important in electrolyte secretion in the gut, and similar effects on colonic mucosa may be relevant to the pathogenesis of inflammatory and infective bowel disorders.
In conclusion, we have shown that IL-1␤ differentially regulates cAMP accumulation and cAMP-mediated chloride efflux in response to adrenoreceptor or EP receptor ligands in Calu-3 human bronchial epithelial cells. This is likely to be an important phenomenon with relevance to the pathophysiology and treatment of cystic fibrosis lung disease.