INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: prostaglandin; thromboxane; cyclooxygenase inhibitors; asthma

Prostaglandins and thromboxanes are known to modify airway tone, as well as inflammatory responses, during episodes of allergic inflammation such as in asthma. Nonselective inhibition of prostaglandin and thromboxane synthesis by indomethacin increased production of interleukin-5 and interleukin-13 and airway hyperresponsiveness in allergic mice (1). In contrast, selective blockade of prostaglandin D₂ (PGD₂) function by targeted deletion of the gene encoding its receptor in mice reduced airway hyperresponsiveness and pulmonary accumulation of eosinophils and lymphocytes (2). Inhibition of thromboxane A₂ (TxA₂) activity by specific inhibitors or receptor antagonists also suppressed the late asthmatic response and airway hyperresponsiveness in guinea-pigs and mice (3, 4). PGE₂ has some protective effect on physiologic responses to allergen and allergen-induced eosinophil influx into guinea-pig lungs (5).

Cyclooxygenase is one of the key enzymes catalyzing formation of prostaglandins and thromboxanes. It is found in at least two isoforms: cyclooxygenase-1 (COX-1), which is expressed constitutively and ubiquitously in most cells and organs as a housekeeping enzyme; and cyclooxygenase-2 (COX-2), which is inducible in endothelial cells, monocytes, and some other cells in response to proinflammatory cytokines such as tumor necrosis factor- and interleukin-1 or to growth factors such as epidermal growth factor (6). Expression of COX-2, which is closely related to inflammation, has been observed, for example, in synovial tissue from patients with rheumatoid arthritis (7). However, there are few data on the expression of COX-2 under conditions of allergen-induced airway inflammation. In the one known study, Gavett and coworkers (8) showed that repeated challenges with allergen upregulated the level of COX-2 protein and caused a corresponding increase in PGE₂ in bronchoalveolar lavage fluid (BALF) in murine lungs. They also showed that COX-2-deficient mice developed more prominent airway inflammation and IgE secretion than mice without the deficiency (8). However, their study left several questions unanswered. First, they did not detect induction of COX-2 messenger RNA (mRNA) in their model, and thus did not convincingly settle the question of whether COX-2 is upregulated at the mRNA level. Second, although the COX pathway may produce either anti-inflammatory PGE₂ or proinflammatory PGD₂ and TxA₂, no effort was made to identify the prostanoid species synthesized by the COX-2 pathway. In the present study, we demonstrated an upregulated COX-2 protein in lungs from allergen-sensitized and -challenged guinea pigs. Using this model, we attempted to elucidate the transition in the expression of COX-2 mRNA, identify the prostanoid species produced by the upregulated COX-2, and examine the pathophysiologic role of COX-2 in allergic airway inflammation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animal Model

The allergic airway inflammation model was prepared by exposing male Hartley guinea pigs to aerosolized 7% ovalbumin solution (Sigma, St. Louis, MO; OA group) or saline (saline group) for 3 min on Days 0, 7, and 14 (9). Days 0 and 7 are termed "sensitization" and Day 14 is termed "challenge". Lungs were harvested from the OA group, the saline group, untreated animals (control group), and animals intravenously injected with lipopolysaccharide (LPS, 2 mg/kg, Sigma; LPS group).

Northern Blot Analysis

RNA extraction and Northern blot analysis was performed as described previously (10). Complementary DNA (cDNA) probes specific to guinea-pig COX-1, COX-2, or glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were generated by reverse-transcription polymerase chain reaction (RT-PCR) using guinea-pig lung RNA.

Western Blot Analysis

Lung homogenates, a guinea-pig platelet lysate (COX-1 control), and recombinant human COX-2 (Oxford Biomedical Research, Oxford, MI) were subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (7.5% Tris-HCl gel) and Western blot analysis (10). Rabbit antiserum to murine COX-1 (1:1,000 dilution; Cayman Chemical Co., Ann Arbor, MI) or human COX-2 (1:100 dilution; Santa Cruz Biotechnology Inc., Santa Cruz, CA) were used as the primary antibodies and an antibody to rabbit IgG (1:1,000 dilution; Amersham, Arlington Heights, IL) was used as the secondary antibody.

Effects of COX Inhibitors on Prostaglandin Synthesis in Lung Slices

The harvested lungs were cut into slices and submerged in a 6-well plate with RPMI 1640 medium containing either a nonselective COX inhibitor indomethacin (1 µM, Sigma), selective COX-2 inhibitors (NS-398 [1 µM, Taisho Pharmaceutical Co., Tokyo, Japan], or JTE-522 [1 µM, Japan Tobacco Inc., Tokyo, Japan]), or vehicle. For the measurement of stimulated prostanoid release, lung slices were harvested 6 h after experimented manipulation, spiked with prewarmed medium containing A23187 (final concentration 10 µM, Sigma), and incubated for 30 min at 37° C. For the measurement of spontaneous prostanoid release, lung slices were harvested 0.5 h after OA exposure, preincubated in RPMI 1640 containing arachidonic acid (50 µM) for 1.5 h, and then incubated for 0.25 to 22 h in arachidonic acid-free medium. The concentrations of PGD₂, PGE₂, 6-keto PGF₁ (a stable metabolite of prostacyclin), and TxB₂ (a stable metabolite of TxA₂) in the supernatant were determined using an enzyme-linked immunosorbent assay (Cayman Chemical Co.).

Effects of NS-398 in Airway Physiology and Inflammation

NS-398 (1 mg/kg) or vehicle was injected intraperitoneally at 0.5, 8, and 17 h after the OA challenge (Day 14). Respiratory resistance (Rresp) was measured in mechanically ventilated animals with a constant-mass plethysmograph during intravenous delivery of histamine (11, 12); measurements were made 18 h after the Day 14 OA challenge. BAL was performed with 20 ml of saline at the completion of the physiologic measurements.

Statistical Analysis

Values are presented as mean ± SEM. Statistical significance of changes in COX-isoform mRNA levels, airway responsiveness, and prostanoid release were judged with one-way or repeated measurement analysis of variance (ANOVA) followed by Scheffé's post hoc test. The BAL cell counts were compared by Mann-Whitney U-test. p Values less than 0.05 were considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We examined the expression of isoform-specific COX mRNA under normal and inflamed conditions in guinea-pig lungs. This study required preparing a cDNA probe with the guinea-pig COX-1-specific nucleotide sequence (154 base pairs [bp], Gene Bank accession number AB054840), which is 87.0%, 85.7%, and 85.1% homologous to the sequence in murine, rat, and human COX-1 cDNA, respectively, but only 64.3% homologous to the sequence in guinea-pig COX-2 cDNA. Northern blot analysis using this probe demonstrated constitutive expression of COX-1 mRNA (2.3 kb) in control guinea-pig lungs (Figure 1, control group). No significant differences were observed among control, saline, and OA groups, and mRNA levels were unchanged over a period of 72 h after exposure to OA or saline (Figures 1 and 2).

View larger version (55K): [in this window] [in a new window]   Figure 1.   Representative data from Northern blot analysis for COX-1 and COX-2 mRNA expression in guinea-pig lungs. (A) Lungs were harvested 1 or 2 h after the intravenous injection of LPS (2 mg/kg, LPS group). (B) Lungs were harvested 1 h after the final challenge from animals sensitized and challenged with aerosols of saline (saline group) or 7% OA in solution (OA group). Lungs were also harvested from untreated guinea pigs (control group).

View larger version (20K): [in this window] [in a new window]   Figure 2.   Kinetics of pulmonary COX-1 mRNA (A) and COX-2 mRNA (B) expression. Lungs were harvested 1 to 72 h after the final exposure from saline-sensitized and -challenged animals (open columns, n = 6 for each time point), or OA-sensitized and -challenged animals (closed columns, n = 6 for each time point). Lungs from untreated guinea pigs were also harvested (control, n = 6). Pulmonary levels of COX-1 or COX-2 mRNA were standardized with levels of GAPDH mRNA. The mean value of the COX/GAPDH mRNA ratio was designated as 1.0. Mean ± SEM. *p < 0.01 compared with saline group.

The guinea-pig COX-2 cDNA hybridized with the expected 4.4-kb transcript in lung RNA obtained from guinea-pigs that had received LPS intravenously but not in lung RNA from untreated animals (Figure 1). Robust expression of COX-2 mRNA was also observed in the lungs of sensitized animals that were harvested within 1 h after OA challenge (16 ± 4-fold increase compared with control/saline groups, mean ± SEM, p < 0.01, Figures 1 and 2). The amount of COX-2 mRNA in lungs of the OA group rapidly returned to the baseline value, and no significant difference was found between saline and OA groups at 3 to 72 h after the exposure. This transient induction of expression of lung COX-2 mRNA was not observed in OA-sensitized saline-challenged animals nor saline-sensitized OA-challenged animals (n = 3 for each group, data not shown). Because changes in the expression of COX-1 and COX-2 mRNA did not differ between the control and saline groups, subsequent analysis of levels of COX protein and prostaglandin synthesis was conducted only for the control and OA groups.

Expression of both COX-1 and COX-2 protein in control lungs was observed in Western blot analysis (Figure 3). Six hours after animals were exposed to OA, COX-2 protein was elevated by 2.5 ± 0.3-fold, whereas COX-1 protein remained unchanged.

View larger version (41K): [in this window] [in a new window]   Figure 3.   Western blot analysis of COX-1 and COX-2 proteins in guinea-pig lungs. Lungs were harvested 6 h after the final OA challenge from sensitized animals (OA group, n = 3) or untreated animals (control group, n = 2). Guinea-pig platelet lysate and human recombinant COX-2 were loaded as positive control for COX-1 and COX-2 proteins, respectively. The intensity of each band, determined by densitometry, is shown under the photographs. Mean intensity for the control group was designated as 1.0.

To determine which prostanoid species were synthesized as the result of COX-2 action, we first examined the effects of the COX-2 inhibitors NS-398 (1 µM) and JTE-522 (1 µM) on A23187-triggered release of PGD₂, PGE₂, TxB₂, and 6-keto-PGF₁ from lung slices harvested 6 h after the OA challenge. In control lungs, although the nonselective COX inhibitor, indomethacin (1 µM), significantly suppressed A23187-triggered production of PGD₂, PGE₂, TxB₂, and 6-keto-PGF₁, the selective COX-2 inhibitors NS-398 and JTE-522 had little effect (Table 1). In OA-treated lungs, production of PGE₂ was suppressed not only by indomethacin (66% ± 4%), but also by NS-398 (46% ± 6%) and JTE-522 (45% ± 7%) (Table 1). NS-398 and JTE-522 also suppressed pulmonary PGD₂ synthesis (34% ± 6% by NS-398; 20% ± 7% by JTE-522) in the OA group, but had no significant effects on the synthesis of TxB₂ and 6-keto-PGF₁ (Table 1). The spontaneous release of PGD₂ and PGE₂ in the 24 h after allergen exposure was also greater in the OA group (n = 6) than in the control group (n = 4) (For PGD₂: 1.30 ± 0.25 pg/mg lung/24 h for OA lungs versus 0.08 ± 0.03 pg/mg lung/24 h for control lungs. For PGE₂: 3.93 ± 0.68 pg/mg lung/24 h for OA lungs versus 0.48 ± 0.18 pg/mg lung/24 h for control lungs, p < 0.05, Figure 4). NS-398 significantly inhibited the OA-induced release of PGD₂ and PGE₂ (72 ± 12% for PGD₂, 97 ± 1% for PGE₂, p < 0.05, Figure 4).

View larger version (13K): [in this window] [in a new window]   Figure 4.   Spontaneous release of PGD2 (A) and PGE2 (B) from guinea-pig lungs ex vivo. Lungs were harvested 30 min after the final exposure to OA or saline and preincubated for 1.5 h in the presence of arachidonic acid (50 µM). Spontaneous prostaglandin release was then measured in arachidonic acid-free medium in the presence or absence of NS-398 (1 µM). Open triangles: lungs from OA-exposed animals incubated without NS-398 (n = 6); closed triangles: lungs from OA-exposed animals incubated in the presence of NS-398 (n = 6); open circles: lungs from untreated animals incubated without NS-398 (n = 4); closed circles: lungs from untreated animals incubated in the presence of NS-398 (n = 4). Mean ± SEM. *p < 0.05 compared with lungs from untreated animals. p < 0.05 compared with lungs from OA-exposed animals incubated with NS-398.

NS-398 treatment did not modify the enhanced airway responsiveness to histamine observed after allergen sensitization and challenge (Figure 5), but it did significantly reduce the number of eosinophils (25 ± 8 × 10⁴/ml versus 56 ± 17 × 10⁴/ml, p < 0.05) and neutrophils (4 ± 2 × 10⁴/ml versus 28 ± 4 × 10⁴/ml, p < 0.05) in BALF (Figure 6).

View larger version (11K): [in this window] [in a new window]   Figure 5.   Effect of NS-398 on airway responsiveness to histamine. NS-398 (1 mg/kg) or vehicle was injected intraperitoneally at 0.5, 8, and 17 h after the final OA challenge. Percent change of respiratory resistance (%Rresp) in response to intravenous injection of histamine was measured at 18 h. Closed circles: animals treated with NS-398 (n = 7); open circles: animals treated with vehicle alone (n = 8). Mean ± SEM.

View larger version (15K): [in this window] [in a new window]   Figure 6.   Effect of NS-398 treatment on airway inflammation. NS-398 (1 mg/kg) or vehicle was administered at 0.5, 8, and 17 h after OA challenge. Inflammatory cells in fluid from BAL performed at 18 h were counted. Closed bars: animals pretreated with NS-398 (n = 10); open bars: animals treated with vehicle alone (n = 8). Mean ± SEM. *p < 0.05 compared with animals pretreated with vehicle alone.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We demonstrated the constitutive expression of COX-1 mRNA and a transient but marked induction of COX-2 mRNA after allergen exposure, followed by an increase in concentrations of pulmonary COX-2 protein in the lungs of ovalbumin-sensitized and -challenged guinea-pigs. Our findings showing the constitutive expression of COX-1 mRNA confirmed the observation in murine and rat lungs (8, 13, 14), whereas the findings showing an induction of COX-2 are consistent with the observations in patients with asthma (15, 16) demonstrating an increased number of COX-2-positive cells in the asthmatic airway. Although Gavett and coworkers (8) reported an increase in COX-2 protein in murine asthmatic lungs, they did not find induction of COX-2 mRNA and speculated that the increase in COX-2 protein occurred by posttranscriptional mechanisms. However, because Gavett and coworkers examined COX-2 mRNA expression only at 24 h after the allergen exposure, it is likely that they missed upregulation of COX-2 mRNA level during an earlier stage of allergen inflammation.

A variety of lung cells such as macrophages, neutrophils, eosinophils, mast cells, vascular endothelial cells, bronchial and alveolar epithelial cells, bronchial and vascular smooth muscle cells, and fibroblasts are capable of expressing COX-2 mRNA and protein (6, 15, 17, 18). In asthmatic human airways, eosinophils, bronchial epithelial cells, mast cells, and macrophages exhibit COX-2 immunoreactivity (19). In our system, however, the observed increase in COX-2 mRNA expression at 1 h of OA exposure cannot be explained by the influx of COX-2-positive eosinophils, because airway recruitment of eosinophils requires at least 3 h after allergen challenge (9). We speculate that the COX-2 gene expression is induced de novo in resident cells of the guinea-pig lungs. Immunohistochemical analysis, which was not feasible with commercially available antibodies, is needed when an appropriate antibody for guinea-pig COX-2 is available to confirm this supposition.

The mechanism of COX-2 induction in allergen-exposed guinea pigs remains unknown. LPS can be a strong inducer of the COX-2 gene in macrophages (22), and thus we must be careful to exclude the possibility that trace amounts of LPS had contaminated the OA aerosols, leading to induction of COX-2 mRNA in the animal lungs. In our experiments, OA exposure upregulated levels of COX-2 mRNA only in the sensitized animals, suggesting that an allergic mechanism is crucial for the induction of COX-2. Furthermore, the kinetics of COX-2 mRNA expression induced by allergen exposure differed significantly from that induced by LPS exposure. Upregulation of pulmonary level of COX-2 mRNA after LPS administration was delayed and persisted for more than 4 h in guinea-pigs (Figure 1A and unpublished data) and rats (14). These findings allow us to exclude the possibility that contaminated LPS induced expression of COX-2 in our model.

Our demonstration of COX-2-dependent prostanoid synthesis was based on the response to two COX-2 selective inhibitors NS-398 and JTE-522. These compounds are highly selective for purified COX-2, and their selectivity has been confirmed in cell culture systems (23). In our previous study (10), we showed that 1 µM NS-398 inhibited 88% of PGE₂ synthesis in COX-2-dominant A549 cells without affecting PGE₂ synthesis in COX-1-dominant bronchial smooth muscle cells. However, COX-2 "selective" inhibitors at higher concentrations or under different conditions also inhibit COX-1 activity. Wakitani and colleagues (25) demonstrated that NS-398 did not inhibit COX-1 activity at 1 µM but inhibited 90% of COX-1 activity at 100 µM. JTE-522, which showed no inhibitory effect on the purified COX-1 enzyme at 100 µM, still inhibited COX-1-dependent release of TxA₂ from human platelets at 10 µM (25). The ex vivo system in our study was thus designed to control the concentration of NS-398 or JTE-522 carefully at the condition highly specific for COX-2 (1 µM). Using this system, we identified PGD₂ and PGE₂ as the major COX-2 products synthesized and released in the airway during allergic inflammation. However, we can not exclude the possibility that COX-2-dependent production of other prostanoid species such as TxA₂ and prostacyclin occurred in a limited compartment of the lung.

Our findings of the effects of NS-398 on BAL cell counts in guinea pigs were in sharp contrast to the previous studies that examined the impact of cyclooxygenase inhibition on allergic inflammation in mice. Gavett and coworkers demonstrated that allergen-exposed COX-1- or COX-2-deficient mice exhibited lower PGE₂ concentrations and more eosinophils in BALF than did wild-type mice (8). Administration of a nonselective COX inhibitor, indomethacin, throughout the allergen sensitization and challenge period also decreased PGE₂ synthesis and increased interstitial eosinophilia. Our protocol, which inhibited cyclooxygenase only after the final allergen challenge and did not manipulate the enzyme activity during the allergen sensitization period, also demonstrated a decrease in PGE₂ but suppressed airway inflammation. We speculate that PGE₂ or other prostanoid species such as prostacyclin acts against the immunologic responses during allergen sensitization, but COX-2-derived prostaglandins, especially PGD₂, act as proinflammatory mediators at the site of established allergen inflammation.

PGD₂ exhibits its biologic activities via two types of receptors, prostaglandin D receptor (DP) and chemoattractant receptor-homologous molecule expressed on Th₂ cells (CRTH₂), both of which may be associated with eosinophilic inflammation. Arimura and associates demonstrated that a specific DP antagonist, S-5751, reduced the number of eosinophils in BALF from OA-exposed guinea pigs (26). Data obtained from DP-deficient mice also showed that PGD₂ and DP are essential for accumulation of inflammatory cells and production of T helper cell, type (Th₂) cytokines in lungs during allergen-induced inflammation. The predominant expression of DP in airway epithelial cells suggests a hypothesis that PGD₂ activates airway epithelial cells to produce eosinophilic cytokines and chemokines. In contrast, CRTH₂ is exclusively expressed on eosinophils, basophils, and Th₂ lymphocytes and directly mediates PGD₂-induced eosinophil migration (27).

In conclusion, transient but robust expression of COX-2 mRNA was induced after an allergen provocation in lungs of sensitized guinea-pigs and was followed by an increase in the level of COX-2 protein and its enzymatic activity. In contrast to nonselective inhibition of prostaglandin synthesis by indomethacin, COX-2-selective inhibitors blocked the synthesis of PGD₂ and PGE₂ but not of prostacyclin or TxA₂. In addition, a COX-2 selective inhibitor significantly reduced allergic airway inflammation. These data indicate that COX-2 selective inhibitors may provide a significant, but different, physiologic consequence in the allergic airway inflammation from nonselective COX inhibitors such as indomethacin or specific prostanoid inhibitors such as TxA₂ synthase inhibitors and TxA₂ receptor antagonists.