INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Prostaglandins (PGs) represent a family of oxygenated arachidonic acid (AA) metabolites that exert potent paracrine and autocrine biologic effects via G-protein-coupled receptors (1). The first committed step in the biosynthesis of PGs is carried out by prostaglandin H₂ synthase, which exists in the two isoforms of cyclooxygenase (COX)-1 and COX-2, encoded by different genes (2, 3). Both isozymes catalyze two separate reactions, a COX-mediated reaction in which two molecules of oxygen are added to arachidonate to form PGG₂, and a peroxidase reaction in which PGG₂ is reduced to PGH₂ (4). COX-1 is constitutively expressed in most cell types, whereas COX-2 expression is rapidly induced by an immediate early gene as a response to proinflammatory stimuli (5).

The lung represents a rich source of prostanoids, which are powerful modulators of vascular and airway tone, affect mucous secretion, and downregulate pulmonary fibroblast proliferation (6).

Investigation aimed at identifying the cellular source of lung prostanoids showed expression of COX-2 in pulmonary epithelial-like type II A549 cells (7), lung fibroblasts (8), and bronchial epithelial cells (9). We and others have also shown that proinflammatory cytokines can induce COX-2 expression in human bronchial smooth-muscle cells (HBSMC) (10, 11), in which PGE₂ represents the major metabolite of COX-2.

Recently, a regulatory role for PGE₂ in the expression of COX-2 has been reported by independent groups; in porcine aortic smooth muscle cells, Karim and coworkers (12) demonstrated that endogenous and exogenous PGE₂ exerts an inhibitory effect mediated by stimulation of protein kinase A, and Minghetti and colleagues showed that lipopolysaccharide (LPS)-induced prostanoid production and COX-2 expression were specifically increased by PGE₂ and 11-deoxy-16,16-dm PGE₂, a selective agonist at the EP₂ receptor, which is coupled to the activation of adenylate cyclase (13). It is therefore likely that prostanoid action is tissue- and cell-specific in enhancing or decreasing cytokine-driven COX-2 expression.

In light of these observations, we studied the effect of endogenous and exogenous PGE₂ on the interleukin (IL)-1- induced expression of COX-2 in HBSMC.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

COX-2 Induction in IL-1-Treated Human Airway Smooth Muscle Cells

HBSMC were obtained from segments of human bronchi and examined with an antibody to smooth-muscle -actin (1:400 final dilution Sigma, St. Louis, MO) to confirm the prevalence of smooth-muscle- type cells as previously described (10).

Cells were grown to confluence in Petri dishes (60 mm) in the presence of minimal essential medium (MEM) containing 10% (vol/ vol) fetal calf serum (FCS), and 100 U/ml penicillin, and 100 µg/ml streptomycin. At 48 h before the experiment, the medium was changed to 5 ml of MEM with 0.4% FCS containing antibiotics, in order to synchronize the cells' activity. Samples were then treated with IL-1 (50 U/ml) and the study test compounds. Incubations were conducted for different time intervals and, at the end of the exposure, the culture medium was removed, and the cell plates were washed twice with 2 ml of phosphate-buffered saline (PBS) and incubated for 30 min with 400 µl of lysis-buffer (Tris-HCl, 20 mM, pH 7.5; 3-[(3-cholamidopropyl)- dimethyl-ammonio]-1-propansulfonate, 16 mM; dithiothreitol, 0.5 mM; ethylendiamminetetraacetic acid [EDTA], 1 mM; benzamidine HCl, 1 mM; leupeptin, 1 µg/ml; and soybean trypsin inhibitor, 10 µg/ml).

Cells were subsequently harvested and kept at 80° C until Western blot analysis.

Western Blot Analysis

The protein content of the different samples was determined with a micro-bicinchoninic acid (BCA) assay (Pierce, Rockford, IL), using bovine serum albumin (BSA) as a standard. Cell lysates (30 µg of protein) were mixed with Laemmli reagent (final concentration: 1% [wt/ vol] sodium dodecyl sulfate; 10% [vol/vol] glycerol; and 0.5% [wt/vol] bromophenol blue) under reducing conditions (4% [vol/vol]; -mercaptoethanol) and heated for 5 min at 85° C (14).

Sodium dodecyl sulfate-polyrylamide gel electrophoresis (SDS-PAGE) was performed with 10% and 3% (wt/vol) acrylamide as separating and stacking gels, respectively. Protein transfer onto nitrocellulose membranes (Amersham International plc, Little Chalfont, Buckinghamshire, UK) was performed at 200 mA for 2 h. Membranes were saturated in blocking buffer containing 5% skim milk and incubated with a specific anti-COX-2 polyclonal antibody (15) for 2 h at RT. After washing with buffer (Tris-HCl, 0.42%, pH 7.4, containing 0.1% SDS, 0.5% deoxycholic acid, and 1% Nonidet P40), membranes were further incubated with an antirabbit IgG conjugated to horseradish peroxidase (Bio-Rad, Hercules, CA) at a dilution 1:2,000 for 20 min at RT. Excess secondary antibody was eliminated by washing in buffer. Chemiluminescence substrates were used to reveal positive bands according to the manufacturer's instructions, and bands were visualized after exposure to Hyperfilm enhanced chemiluminescence (ECL) (Amersham). Rabbit polyclonal antibody was obtained against the eicosapeptide (C)-NASSSRSGLDDINPTVLLK (COX-2 peptide), which is present only in the carboxy-terminal region (amino acids 580 to 598) of human COX-2 (16). Quantitation was performed by scanning the blot into Photoshop Adobe and performing densitometry with NIH Image Software.

Northern Blot Analysis

Total RNA was isolated using TRIzol reagent (Life Technologies, Inc., Grand Island, NY) and was quantitated spectrophotometrically. Total RNA (10 µg) was separated on a 1% agarose gel containing formaldehyde, visualized by ethidium bromide staining, and transferred overnight to Hybond N plus membranes (Amersham) through capillary blotting. The membranes were probed with ³²P-labeled complementary DNA (cDNA) probes (COX-2 cDNA, with glucose-3-phosphate dehydrogenase [GAPDH] cDNA as an internal control) in Rapid-Hyb hybridization buffer (Amersham). The resulting blots were then subjected to autoradiography with NEF reflection autoradiography films (NEN Life Science Products, Boston, MA). Densitometric analysis was performed by scanning the blots into Photoshop and performing densitometry with NIH Image Software.

COX-2 Activity in HBSMC: Formation of PGE₂

Monolayers of HBSMC were grown to confluence in Petri dishes (60 mm); at 48 h before the experiment for measuring formation of PGE₂, the medium was changed to 2 ml of MEM containing 0.4% FCS and antibiotics. Immediately before the experiment, cells were washed twice with 2 ml of sterile PBS and incubated with 2 ml of FCS-free MEM containing antibiotics. After treatment with IL-1 and test compounds, incubations were conducted for different time intervals. At the end of the incubation period, the medium was collected and stored at 80° C for enzyme immunoassay (EIA) of PGE₂. Solid-phase EIA was performed as previously described (10).

The standard curve and the results of quantitative determination of PGE₂ in biologic samples were determined with a linear log-logit transformation.

Adenylate Cyclase Activity

The standard assay mixture, at the final volume of 100 µl, contained Tris-HCl buffer, 10 mM, pH 8; [8-¹⁴C]adenosine triphosphate (50 dpm/pmol); 0.5 mM [8-³H]cyclic adenosine monophosphate (cAMP) (approximately 360 dpm-nmol); 2 mM MgCl₂; 2 mM creatine phosphate; 17 U-ml creatine phosphokinase, and 10 µM guanosine triphosphate. PGE₂ (0.1 to 1 µM) or forskolin (10 µM) were tested in the presence of 3-isobutyl-methylxanthine, an inhibitor of phosphodiesterase. The incubation was performed on membrane preparations (3 µg protein/sample). [8-³H]cAMP was included in the assay mixture for correction of column loss and of the possible effect of phosphodiesterases. [8-¹⁴C, 8-³H]cAMP was isolated and detected according to the method of Salomon and colleagues (17). Protein content was determined in the supernatant, using a micro-BCA assay (Pierce) with BSA as standard.

Statistical Analysis

Analysis of variance and Student's t test for unpaired data were used to evaluate differences among control and treated samples. A value of p  0.05 was considered significant. Unless stated otherwise, results are presented as the mean ± SE of n different experiments.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In HBSMC, treatment with IL-1 (50 IU/ml) induced a significant increase in COX-2 expression, as previously shown (10). Exogenous PGE₂ (1 µM), coadministered with IL-1, induced, after 18 h of incubation (Figure 1A), a significant increase in protein expression over that induced by IL-1 per se. Treatment with flurbiprofen, a nonselective COX inhibitor, significantly blunted the production of PGE₂ (Figure 1B) and the expression of COX-2, suggesting a contribution of endogenous PGE₂ to the observed effect of IL-1 (Figure 1A). Moreover, the effect of flurbiprofen was overcome when cells were treated with exogenous PGE₂ (data not shown), indicating the existence of a causal relation between drug modulation of COX-2 expression and inhibition of COX activity (Figures 1A and 1B).

View larger version (18K): [in this window] [in a new window]   Figure 1.   Induction of COX-2 protein by IL-1 in HBSMC: effect of exogenous PGE2 and of COX inhibition by (flurbiprofen). (A) PGE2 (1 µM) and flurbiprofen (Flurbi, 10 µM) were added at the same time as IL-1, and incubation was done for 18 h. SDS-PAGE was performed on 30 µg of protein, and COX-2 was detected with a polyclonal antibody raised against a COX-2-specific peptide. (B) Samples of the culture medium of cells treated as in A were withdrawn and analyzed for PGE2 content as described in METHODS. Results are expressed as mean ± SEM (n = 7). *p < 0.05 versus IL-1 alone.

Time-course studies of IL-1-induced COX-2 expression showed that PGE₂ enhanced, while flurbiprofen blunted, enzyme expression at every time point studied (12 to 48 h) (Figure 2A). None of the previous treatments was able to modify the expression of COX-1 (Figure 2B).

View larger version (55K): [in this window] [in a new window]   Figure 2.   Time-dependent induction of COX-2 protein by IL-1 in HBSMC: effect of exogenous PGE2 and of COX inhibition (by flurbiprofen). PGE2 (1 µM) and flurbiprofen (10 µM) were added at the same time as IL-1, and incubation was done for 12, 18, 24, and 48 h. SDS-PAGE was performed on 30 µg of protein for each sample, and COX-2 was detected with a polyclonal antibody raised against a COX-2-specific peptide. (B) Expression of COX-1 protein in HBSMC. SDS-PAGE was performed on 30 µg of protein, and COX-1 was detected with a polyclonal antibody raised against a COX-1-specific peptide. Results are representative of two different experiments with the same results.

Administration of exogenous AA (10 µM), resulting in significant formation of PGE₂ (15.13 ± 3.4 [mean ± SE] ng/ml, versus 0.45 ± 0.21 ng/ml in the absence of exogenous AA n = 3), led to upregulation of COX-2 expression induced by IL-1, but this effect was not statistically prevented in the presence of flurbiprofen (data not shown).

Determination of COX-2 mRNA expression showed increased amounts of transcript in the presence of PGE₂ at 6 h, but not at 3 h, after coadministration of IL-1 and PGE₂ (Figure 3).

View larger version (44K): [in this window] [in a new window]   Figure 3.   Induction of COX-2 mRNA by IL-1 in HBSMC: effect of exogenous PGE2. PGE2 (1 µM) was added at the same time as IL-1, and incubation was done for 1, 3, and 6 h. Total RNA was extracted and separated, and Northern blotting was performed as described in METHODS. Results are representative of two different experiments.

PGE₂ per se did not induce COX-2 protein expression (Figure 4, lane 4). Administration of PGE₂ at 6 h after IL-1 caused an upregulation of COX-2 enzyme expression similar to that observed after coadministration of PGE₂ and IL-1 (Figure 4, lanes 1, 2, and 3), suggesting that the effect of the prostanoid involves mechanisms taking place at times after the initial activation induced by IL-1. In order to evaluate the role of transcription and translation mechanisms contributing to the PGE₂-dependent increase in COX-2 levels, we studied the effect of cycloheximide (CHX, 10 µM), a translation inhibitor, as well as that of dexamethasone (DEX, 1 µM), a reported inhibitor of cytokine-induced COX-2 transcription (10, 18). Concomitant treatment with CHX and PGE₂ at 6 h after the cytokine totally abolished the upregulation induced by PGE₂, and its effect did not differ from the effect of CHX alone (Figure 4, lanes 5 and 6 ). A similar pattern was observed with the transcription inhibitor DEX (Figure 4, lanes 7 and 8). The same results were confirmed in the presence of actinomycin D (5 µg/ml), a more general inhibitor of transcription (data not shown).

View larger version (38K): [in this window] [in a new window]   Figure 4.   Induction of COX-2 by IL-1 in HBSMC: effect of CHX and DEX in the absence and presence of PGE2. CHX (10 µM) or DEX (1 µM) were added to cells 6 h after IL-1, either alone or together with PGE2 (1 µM). Incubation was done for 12 additional hours, to a total of 18 h after IL-1 treatment. SDS-PAGE was performed on 30 µg of protein for each sample, and COX-2 was detected with a polyclonal antibody raised against a COX-2-specific peptide. Data are expressed as mean ± SEM (n = 3). n.s. = not statistically different.

In light of these data, we investigated the signal-transduction mechanism downstream of the interaction of PGE₂ with its putative receptor, resulting in enhancement of the transcription machinery. At least four different subtypes of PGE₂ receptors have been cloned and expressed (1); in particular, EP₂ and EP₄ receptors have been found to couple with adenylate cyclase, leading to an increase in intracellular cAMP.

We studied the effect of PGE₂ on adenylate cyclase activity in HBSMC membranes, as well as the effect of forskolin, a direct activator of adenylate cyclase, on IL-1-dependent COX-2 expression. The results showed that PGE₂ increased adenylate cyclase activity in HBSMC preparations in a concentration-dependent manner (Figure 5A); forskolin (10 µM) also resulted in a marked increase in adenylate cyclase activity, as well as COX-2 expression, suggesting the existence of a causal relationship between these events (Figures 5A and 5B). Neither forskolin per se nor PGE₂ had any effect on COX-2 expression. The same results were obtained with salbutamol (30 µM), a bronchodilator that acts by increasing cAMP (Figure 5C), and with an analogue of cAMP (8-Br-cAMP, 1 mM; data not shown).

View larger version (26K): [in this window] [in a new window]   Figure 5.   Induction of COX-2 by IL-1 and adenylate cyclase activity in HBSMC: effect of exogenous PGE2, forskolin, and salbutamol. (A) Adenylate cyclase activity was evaluated in cellular membranes as described in METHODS. (B) PGE2 (0.1 to 1 µM) and forskolin (10 µM) were added at the same time as IL-1, and incubation was done for 18 h. SDS-PAGE was performed on 30 µg of protein for each sample, and COX-2 was detected with a polyclonal antibody raised against a COX-2-specific peptide. (C ) Salbutamol (30 µM) was added at the same time as IL-1, and incubation was done for 10 h, after which Western blotting was performed. COX-2 data are representative of four different determinations, and the results of adenylate cyclase activity are expressed as mean ± SEM (n = 4). *p < 0.05 versus IL-1 alone.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The induction of COX-2 protein by several proinflammatory cytokines represents an important mechanism controlling the overall production of prostanoids and the evolution of the inflammatory response. The induction of COX-2 appears to occur at a transcriptional level, through transcription-factor activation. IL-1 has been associated with induction of nuclear factorB (NF-B) and increased COX-2 transcription (19), on the basis of the existence of two putative NF-B-binding motifs on human COX-2 gene (20). Different transcription factors such as NF-IL-6 have also been reported to be involved in COX-2 expression and may interact synergistically with NF-B (21).

A negative feedback exerted by COX end products on the expression of COX-2 may provide an important autocrine regulatory mechanism. In porcine aortic smooth-muscle cells, endogenous and exogenous PGE₂ inhibit the expression of COX-2 following activation by fibroblast growth factor-2 (12), supporting a role for endogenous PGs in limiting the expression of COX-2 in vascular smooth-muscle cells. On the other hand, PGE₂ has been shown to upregulate LPS-induced COX-2 expression and thromboxane production in microglial cells (13).

We have previously shown that HBSMC express COX-2 in response to IL-1 stimulation, and synthesize PGE₂ as their main COX metabolite (10). The reported upregulation, by exogenous PGE₂, of IL-1-dependent COX-2 expression in these cells is very likely to be of physiologic relevance; in fact, experiments conducted in our laboratory have shown the ability of human bronchial strips to release amounts of PGE₂ similar to those used in our study (unpublished observation).

Pretreatment with flurbiprofen, a nonspecific COX inhibitor, significantly suppressed endogenous PGE₂ production as well as COX-2 expression induced by IL-1, suggesting that PGE₂ itself may play a role in the expression of COX-2 following cytokine stimulation. Since HBSMC constitutively express COX-1 (10), a nonspecific COX inhibitor was used in order to inhibit overall PGE₂ production, which might contribute to the modulation of COX-2 expression.

Interestingly, the administration of exogenous AA (10 µM), resulting in significant formation of PGE₂, led to the upregulation of COX-2 expression induced by IL-1. This effect was not statistically prevented in the presence of flurbiprofen, indicating that AA-dependent COX-2 overexpression was linked to an additional, direct mechanism of action (i.e., protein kinase C activation) (22).

The observations that: (1) PGE₂ was ineffective in the absence of IL-1; (2) COX-2 mRNA analysis showed increased amounts of transcript at 6 h after IL-1 and PGE₂ treatment; and (3) administration of PGE₂, either concomitantly or 6 h after cytokine stimulation, led to identical COX-2 upregulation suggest that the effect of PGE₂ was linked to modifications that lay downstream of the initial cytokine activation, as expected from the coordinated sequence of events leading to PGH₂ synthase-mediated prostaglandin synthesis. The substantial reduction of protein expression observed with CHX suggests that the effect of PGE₂ requires de novo protein synthesis; moreover, an effect of PGE₂ on protein stabilization can be ruled out on the basis of the identical protein expression observed in the preesence and absence of PGE₂ in CHX-treated cells.

COX-2 mRNA has a long 3' untranslated region containing several different polyadenylation signals and multiple AUUUA instability sequences that mediate rapid degradation of the transcript (23) and increased COX-2 protein expression may result from stabilization of COX-2 mRNA. Nevertheless, in HBSMC, increased COX-2 protein synthesis appears to be the result of increased DNA transcription rather than of an increased mRNA half-life, as indicated by the effect of DEX or actinomycin D coadministered with PGE₂ at 6 h after IL-1. The significant reduction of COX-2 expression in DEX-treated cells, when compared with the effect of CHX, is in agreement with the reported NF-kB-dependent overtranscription of COX-2 DNA in the presence of translational blocking agents (24).

Previous results have shown that PGs and forskolin (a direct adenylate cyclase activator) can induce COX-2 mRNA in osteoblastic cells (25, 26), in accord with the presence of a cAMP response element on the COX-2 gene (20, 27). On the other hand, PGs of the E and I series, as well as forskolin and 8 Br-cAMP, inhibited LPS-induced COX-2 and inducible nitric oxide synthase expression in murine macrophages (28). In HBSMC, PGE₂ was able to stimulate adenylate cyclase activity in a concentration-dependent manner, suggesting the activation of EP₂ or EP₄ receptor subtypes. A causal relationship between increased cAMP and increased COX-2 expression was supported by the marked upregulation of IL-1-induced COX-2 expression induced by forskolin. Moreover, achievement of the same results in the presence of salbutamol, a bronchodilator acting on cAMP levels through a nonprostanoid receptor, significantly adds to the mechanistic as well as the clinical implications of the observed effect.

In conclusion, the results of our study indicate that end-products of COX-2, and specifically PGE₂, positively modulate the expression of the enzyme responsible for their biosynthesis in HBSMC. This effect may be of clinical relevance in light of the hypothesized antiinflammatory role of PGE₂ in human airways (29, 30), contributing to the modulation of bronchial reactivity in the presence of an inflammatory event. Indeed, drugs that inhibit COX enzyme activity are of no use in the treatment of inflammatory diseases such as asthma, if indeed they do not trigger airway hyperreactivity. Furthermore, PGE₂ has been reported to inhibit the expression of genes encoding inflammatory cytokines such as IL-1 and TNF (31, 32), and to modulate hypoxia-dependent increases in TNF- synthesis in human alveolar macrophages (33).

A self-amplifying loop, based on increased PGE₂ production leading to increased COX-2 expression that, in turn, increases PGE₂ production, may render the airway smooth-muscle cell an active participant in biochemical modulation of the inflammatory response in human airways.