INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In acetylsalicylic acid (ASA)-sensitive asthmatic patients ingestion of aspirin or other nonsteroidal anti-inflammatory drugs (NSAIDs) induces, within 30 to 90 min, severe asthmatic attacks, which are usually difficult to reverse (1). Moreover, up to 90% of ASA-sensitive asthmatics challenged with provocative doses of aspirin, in addition to bronchospasm, demonstrate nasal responses heralded by rhinorrhea and nasal obstruction (2). ASA sensitivity may be present in 5 to 20% of the asthmatic population (1, 2) and usually is associated with a typical clinical syndrome; i.e., "aspirin triad," involving bronchial asthma, chronic rhinosinusitis with nasal polyps, and intolerance to ASA and other NSAIDs. Recurrent polyposis, reflecting chronic hypertrophic inflammation of the upper airway mucosa, is a consistent finding in these patients.

Aspirin sensitivity is not an immunologically mediated reaction and the most plausible basic mechanism underlying sensitivity to aspirin seems to involve interaction of the drug with arachidonic acid (AA) metabolism (3). The fatty acid AA is released from cell membranes by phospholipases and is then metabolized by the cyclooxygenase or lipoxygenase enzyme. Eicosanoids generated on both pathways may induce development of acute reaction owing to their capability to induce bronchial smooth muscle contraction (prostaglandin F₂ alpha [PGF₂], PGD₂, leukotriene C₄ [LTC₄]); they may also increase vascular permeability (PGE₂, PGD₂, LTC₄) and glandular secretion (LTC₄). Eicosanoids may also activate other cells, promoting mast cell or eosinophil degranulation and chemotaxis, thus actively perpetuating chronic inflammation. Aspirin in pharmacological doses irreversibly acetylates and inactivates cyclooxygenase resulting in inhibition of prostaglandin, thromboxane, and prostacyclin synthesis. It has been proposed by Szczeklik and coworkers (4) that in sensitive individuals aspirin, by inhibition of cyclooxygenase, leads to altered metabolism of AA, resulting in bronchospasm and rhinorrhea. It has been shown that basal peptidoleukotriene concentrations in urine of ASA-sensitive asthmatics are elevated and further increase after ASA challenge (5). It has also been demonstrated that oral or nasal challenge with aspirin in ASA-sensitive patients induces significant peptidoleukotriene release into the nasal lavages, concomitant with nasal symptoms (6). Because ASA-induced reactions both in the lower and upper airways can be inhibited or significantly decreased by peptidoleukotriene synthesis inhibitors or receptor antagonists, one may suggest a crucial role for these leukotrienes in the pathophysiology of ASA-induced reaction (7). Although the cellular source of peptidoleukotrienes released during ASA-induced reaction has not been identified, ASA-induced reactions involve activation of mast cells and eosinophils (8). However, previous studies do not provide information as to which cells are the target cells for aspirin to initiate the adverse reaction. In vitro studies of basophils (9), platelets, and peripheral blood leukocytes (10, 11) yielded conflicting results and could not demonstrate specific and distinct cell activation in ASA-sensitive patients. Thus, target cells for aspirin to induce adverse reactions and the source of putative mediators resulting in asthmatic or nasal symptoms of hypersensitivity still remain unknown.

In ASA-sensitive patients, adverse symptoms can be evoked not only by oral but also by bronchial or intranasal challenges with soluble aspirin, suggesting that ASA sensitivity may be a local phenomenon confined to the respiratory airway mucosa (6). Airway epithelial cells (ECs), which form the first point of contact for external stimuli in the airways, are capable of generating lipoxygenase and cyclooxygenase metabolites of AA under baseline conditions or in response to various stimuli. Culture of human respiratory epithelium may provide an efficient source of eicosanoids for in vitro quantification, allowing for assessment of their modulation by endogenous and exogenous agents. This study was designed to investigate the production of AA derivatives (eicosanoids) from cultured nasal ECs in subjects with and without aspirin sensitivity. We aimed to test the hypothesis that local production of eicosanoids in the respiratory epithelium in patients with ASA sensitivity differs from that of patients tolerant to aspirin, and that aspirin challenge in vitro may have a differential effect on AA acid metabolism in ASA-sensitive and ASA-tolerant patients.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

Twenty-five patients with nasal polyps were studied, 10 of whom were aspirin-sensitive asthmatics with chronic rhinosinusitis, and 15 were ASA-tolerant patients with chronic rhinosinusitis. The diagnosis of aspirin sensitivity was made on the basis of positive history of bronchial and nasal reaction to ASA or other NSAIDs, and confirmed by positive oral or inhalation challenge with lysine aspirin as previously described (12). Characteristics of groups of patients are shown in Table 1. Ten ASA-sensitive patients (four men, six women, 30 to 57 yr of age, five of whom were atopic) suffered from bronchial asthma, chronic rhinosinusitis, and recurrent polyposis. All patients were receiving inhaled steroids and were using regular intranasal steroids in therapeutic dose of 200 µg budesonide per day. Intranasal steroids were stopped at least 4 wk before the surgery and since this point patients were asked to use, on demand, intranasal decongestant to relieve nasal symptoms. Of 15 ASA-tolerant patients (five men, 10 women, 33 to 72 yr of age, eight of whom were atopic), all had chronic rhinosinusitis and polyposis; two suffered from bronchial asthma. None of them was receiving systemic steroids, two were taking steroids in inhalations, and five of them were on intranasal steroids which were stopped 4 wk before surgery. Atopy was defined by the presence of at least one positive skin prick test (wheal > 3 mm with flare) out of a battery of 14 inhalant allergens tested.

Nasal polyps were obtained during elective nasal surgery performed for reasons unrelated to the goals of this project. The study was approved by the local regional ethical committee and informed consent was obtained from patients. Patients did not receive any oral medications, including NSAIDs, within a week before the polypectomy.

The Technique of Nasal Polypectomy and Tissue Handling

Nasal polypectomies were performed under local anesthesia. Before surgery, 2% tetracaine HCl and 0.25% phenylephrine HCl were applied topically to the turbinates, nasal septum, and middle meatus. After decongesting, the polyp was injected with 2 to 4 ml of 1% lidocaine with 1:100,000 epinephrine. Then, the polyp was removed using gentle traction and the snare technique. Polyps were placed into Ham's F-12 medium supplemented with antibiotics and immediately transported to the laboratory in a thermoinsulation container (temperature 4 to 6° C).

Materials

Plastic culture dishes (24-well) were purchased from Nunc (Roskilde, Denmark), Vitrogen 100 from Collagen Biomaterials (Palo Alto, CA), Ham's F-12 medium from Biochrom AG (Berlin, Germany), Epithelial Growth Factor and Endothelial Cell Growth Supplement from Becton Dickinson (Heidelberg, Germany), aspirin (Aspisol) was a kind gift from Bayer (Warsaw, Poland), and calcium ionophore A23187 was supplied by Calbiochem (La Jolla, CA). All other reagents were purchased from Sigma (St. Louis, MO).

EC Culture and Experimental Design

Human nasal polyp ECs were grown from nasal polyps on a collagen matrix (Vitrogen 100) in Ham's medium supplemented with growth factors according to the method described previously (13) with modifications. The culture system and the modifications are described in detail elsewhere (14). The purity of ECs as assessed by immunocytochemistry with anti-human cytokeratin (anti-CK-1) antibody (Dako, Glostrup, Denmark) and by electron transmission microscopy was consistently above 98%. Each culture consisted of six to 16 wells on one or two plates (maximum 8 wells per plate). Human nasal ECs were cultured in serum-free hormone-supplemented Ham's F-12 medium in accordance with the previously described conditions; confluence as high as 90% was reached (usually 5 to 6 d). Then epithelial cultures were exposed to Ham's F-12 medium supplemented with hormones and 10% fetal bovine serum (FBS) (vol/vol) overnight. After this preincubation medium was removed and cultures were stabilized with Ham's F-12 medium for 1 h.

The experiment involved incubation of cells in separate wells with calcium ionophore (10⁵ M), lysine aspirin (for dose-response experiments 2, 20, and 200 µM, and all other 200 µM) or with Ham's F-12 medium alone (as a control). Three sequential incubations with medium, lysine aspirin, and calcium ionophore were performed for 5, 15, and 60 min one after another. After each stimulation period the supernatants were collected and immediately replaced by the same volume of the fresh medium containing stimulant, lysine ASA, or the medium alone. All collected supernatants were stored in 70° C for eicosanoids measurement. Our preliminary dose-response experiments demonstrated that 10⁵ M of calcium ionophore is an optimal concentration for stimulation of AA metabolism in nasal polyps ECs as assessed by the release of incorporated ³H AA (data not shown), and this concentration was used in the present study.

All experiments were conducted with the same design, thus for eicosanoids measurement 5-, 15-, and 60-min fractions were available. Because of limited availability of enzyme-linked immunosorbent assays (ELISA) and radioimmunoassays (RIA) for eicosanoids, and based on the preliminary time-course experiments which demonstrated maximal generation of PGE₂ and 15-hydroxyeicosatetraenoic acid (15-HETE), during the third collection period we decided to perform eicosanoids measurements only in supernatants collected after 60 min incubation with medium alone, ASA, or calcium ionophore. After the last collection of the supernatant, cells were exposed to 0.1% trypsin- ethylenediaminetetraacetic acid (EDTA) solution for 30 min in incubator (37° C, 5% CO₂), aspirated, homogenized, and used for double-stranded deoxyribonucleic acid (dsDNA) measurement.

Eicosanoids Analysis

Two enzyme-linked immunoassays were used for PGE₂ and PGF₂ quantification (purchased from Assay Designs, Ann Arbor, MI). 15-HETE, 12-HETE, 5-HETE, LTC₄/D₄/E₄, and PGD₂ were measured using ³H RIA kits (from Perseptive Biosystem, Cambridge, MA and Amersham, Little Chalfat, UK, respectively). Kits were used following the original manuals. RIA tests were performed on unextracted aliquots of each sample and the measurements were performed in duplicates. Sensitivity of immunoassays was as follows: for PGE₂, 36.2 pg/ml; for PGF₂, 4.62 pg/ml; PGD₂, 3.1 pg/500 µl; 15-HETE, 2.5 pg/ 100 µl; 12-HETE, 10.3 pg/100 µl; 5-HETE, 4.88 pg/100 µl; and peptidoleukotrienes, 7.63 pg/100 µl. The concentrations of PGE₂, PGF₂, and 15-HETE measured in undiluted supernatants were well above the threshold of ELISA and RIA assays sensitivity so they could be read from the middle portion of the standard curve.

Results from each well were calculated per amount of dsDNA in cells measured by the method described by Labarca and Paigen (15). In brief, homogenized cells were mixed with Hoechst 33258 compound stock and the solution (containing finally 0.1 µg/ml of Hoechst) was transferred into the luminescence spectrometer cell (LS-50B; Perkin-Elmer, Branchburg, NJ). The excitation wavelength was 356 nm and the emission wavelength was 458 nm. Emission measurements were done in triplicates. The standard curve was plotted using fluorescence values against dsDNA standard concentrations. Concentrations of dsDNA were counted from the standard curve using the computerized optimalization system. Only results with coefficient of determination (R²) greater than 0.97 were used for dsDNA quantification.

Statistical Analysis

Results of eicosanoid measurement are expressed as mean ± SEM in ng/µg dsDNA. The Wilcoxon matched pairs test was used for analysis within the group and the Mann-Whitney U test was used for data analysis between groups of patients.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

As mentioned previously, supernatants, including generation of eicosanoids during 60 min, were found to be the most appropriate for the assessment of the eicosanoids generation, and the data represent only generation measured in supernatants harvested at this time point. The time-course including 5-min, 15-min, and 60-min incubation periods is presented only for ASA challenge and 15-HETE measurement.

Eicosanoids Generation by Unstimulated ECs

Unstimulated ECs from ASA-sensitive patients generated significantly less PGE₂ than cells from ASA-tolerant patients (0.8 ± 0.3 ng/µg dsDNA and 2.4 ± 0.5 ng/µg dsDNA; p < 0.05; for ASARS and ATRS patients, respectively). One ASA-tolerant patient had approximately 10 times higher PGE₂ level in the supernatant (31.1 ng/µg dsDNA) than the mean PGE₂ value for the rest of the group, and his data were not included in the analysis. Inclusion of the outlier patient (not shown) did not affect statistical analysis of the results. PGF₂ was present in similar concentration in supernatants from ASARS and ATRS patients (0.5 ± 0.2 ng/µg dsDNA versus 0.4 ± 0.2 ng/µg dsDNA). The mean PGE₂/PGF₂ ratio was only 3.45 in ASARS patients compared with 11.4 in ATRS patients. There were also no differences between spontaneous generation of 15-HETE in cultures from both groups (1.4 ± 0.4 ng/µg dsDNA and 2.3 ± 0.5 ng/µg dsDNA; not significant [NS]) (Figure 1). No release of PGD₂ or peptidoleukotrienes LTC₄/D₄/E₄ could be detected.

View larger version (12K): [in this window] [in a new window]   Figure 1.   Generation of eicosanoids by unstimulated nasal polyp ECs from ASA-tolerant patients and ASA-sensitive patients. Bars represent the mean for each group. Asterisk indicates p = 0.01, Wilcoxon matched pairs test. PGD2, 5-HETE, 12-HETE, and LTC4/ D4/E4 were not detected.

Effect of ASA and Calcium Ionophore on PGE₂ Generation

Incubation of ECs with 2 µM, 20 µM, and 200 µM of aspirin resulted in a dose-dependent decrease in PGE₂ concentration in cell supernatants in both groups of patients (Figure 2). In separate experiments (Figure 3A) incubation with one concentration (200 µM) of ASA decreased synthesis of PGE₂ by 35% (from 2.4 ± 0.5 ng/µg dsDNA to 1.8 ± 0.4 ng/µg dsDNA) in the ATRS group (n = 7) and by 47% (from 0.8 ± 0.3 ng/µg dsDNA to 0.4 ± 0.2 ng/µg dsDNA) in ASARS patients (n = 8). After aspirin challenge PGE₂ generation was still significantly lower in ASARS patients compared with ATRS patients.

View larger version (24K): [in this window] [in a new window]   Figure 2.   Dose-dependent inhibition of PGE2 generation by ASA in ECs from ASA-tolerant (n = 4) and ASA-sensitive (n = 3) patients. Data are presented as median values.

View larger version (18K): [in this window] [in a new window]   View larger version (18K): [in this window] [in a new window]   Figure 3.   (A) Effect of 200 µM of aspirin and (B) effect of stimulation with calcium ionophore (105 M) on PGE2 generation by ECs from ASA-tolerant and ASA-sensitive patients. Each point represents an individual patient. Bars represent the mean for each group.

Calcium ionophore (10⁵ M) increased PGE₂ generation by ECs from both groups of subjects: in the ATRS group from 2.4 ± 0.5 ng/µg dsDNA to 4.2 ± 0.5 ng/µg dsDNA (p < 0.05) and in the ASARS group from 0.8 ± 0.3 ng/µg dsDNA to 1.6 ± 0.5 ng/µg dsDNA (p < 0.05) (Figure 3b). Although the relative magnitude of the increase was similar in both groups (mean 160.3% and 166.7% for ATRS and ASARS patients, respectively), stimulated PGE₂ generation was still significantly lower in ASARS patients compared with ATRS patients (p < 0.05).

Effect of ASA and Calcium Ionophore on PGF₂ Generation

Incubation of ECs with 2 µM, 20 µM, and 200 µM of ASA resulted in dose-dependent inhibition of PGF₂ generation in both groups of patients (data not shown). Cells' incubation with 200 µM of aspirin caused a significant decrease of PGF₂ synthesis to mean 0.2 ± 0.06 ng/µg dsDNA in ATRS patients (p < 0.05) and to 0.2 ± 0.05 ng/µg dsDNA in ASARS patients (NS) (Figure 4A). The mean PGF₂ levels after ASA were not significantly different between ATRS and ASARS patients. However, the mean PGE₂/PGF₂ ratio during incubation with aspirin was 10.3 in ATRS and only 2.1 in ASARS patients. Five of six ATRS patients had PGE₂/PGF₂ ratio above 8.5 and one 3.8, whereas all ASARS patients revealed PGE₂/ PGF₂ ratio below 3. 

View larger version (18K): [in this window] [in a new window]   View larger version (18K): [in this window] [in a new window]   Figure 4.   (A) Effect of 200 µM of aspirin and (B) effect of stimulation with calcium ionophore (105 M) on PGF2 generation by ECs from ASA-tolerant and ASA-sensitive patients. Each point represents an individual patient. Bars represent the mean for each group.

Calcium ionophore induced an increase in PGF₂ generation from 0.4 ± 0.2 ng/µg dsDNA to 1.1 ± 0.4 ng/µg dsDNA in ATRS patients and from 0.5 ± 0.2 ng/µg dsDNA to 2.4 ± 1.6 ng/µg dsDNA in ASARS patients; the mean of the calcium ionophore-stimulated PGF₂ generation was not significantly different between groups (Figure 4B).

Effect of ASA and Calcium Ionophore on 15-HETE Generation

Incubation of ECs with ASA (200 µM) did not change the mean of 15-HETE generation in ATRS patients (n = 6). On the contrary, aspirin (200 µM) significantly enhanced generation of 15-HETE in cultured epithelial cells from ASARS patients (n = 7) from the mean 1.4 ± 0.4 ng/µg dsDNA to 3.7 ± 0.6 ng/µg dsDNA; p < 0.05; mean increase +359%. There was also a statistically significant difference in 15-HETE concentration after exposure to ASA between the two groups of patients (2.1 ± 0.4 ng/µg dsDNA versus 3.7 ± 0.6 ng/µg dsDNA; p < 0.05) (Figure 5A). These differential responses were also evident when responses from individual patients were analyzed. In the ATRS group, four of six patients responded to aspirin with a marked decrease in 15-HETE generation ranging from 23 to 56% inhibition from the control; in one patient no change occurred, and in another an increase in 15-HETE was observed after ASA. On the contrary, all but one ASA-sensitive patient responded to ASA with increase in 15-HETE release, which in individual patients ranged from 44% to 1,673% of the unstimulated control. Only one patient demonstrated a decrease in 15-HETE release by 32% after ASA challenge.

View larger version (20K): [in this window] [in a new window]   View larger version (21K): [in this window] [in a new window]   Figure 5.   (A) Effect of incubation with ASA (200 µM) and (B) effect of stimulation with calcium ionophore (105 M) on 15-HETE generation by ECs from ASA-tolerant and ASA-sensitive patients. Each point represents an individual patient. Bars represent the mean for each group.

In separate experiments when three doses of ASA were tested, ASA induced increase in 15-HETE generation only in ASA-sensitive patients. Although the increase was not linearly dose-dependent for 2, 20, and 200 µM of aspirin, the mean 15-HETE generation after exposure to 200 µM of ASA was higher than after 20 µM or 2 µM (+162%, +4%, and 8% of control generation, respectively) only in the ASA-sensitive group (data not shown). The time-course of the 15-HETE generation in response to 200 µM of ASA demonstrated that during 5 and 15 min of incubation there was no significant increase in the 15-HETE concentration in the supernatants of either group. Significant enhancement of 15-HETE concentration was noticed in ASA-sensitive patients only after 60 min incubation with aspirin (Figure 6). Calcium ionophore increased 15-HETE generation in both groups (to 5.7 ± 1.4 ng/µg dsDNA and to 2.7 ± 0.8 ng/µg dsDNA in ASA-tolerant and in ASA-sensitive patients, NS; Figure 5B).

View larger version (14K): [in this window] [in a new window]   Figure 6.   Time-course of 15-HETE generation by ECs from ATRS (n = 3) and ASARS (n = 4) patients after incubation with 200 µM of ASA. Data are presented as median values of a percentage of 15-HETE generated by control cells incubated with medium.

We did not find any measurable generation of PGD₂, 5-HETE, or peptido-LTs in supernatants of unstimulated or calcium ionophore-stimulated EC cultures, or in ECs cultures exposed to ASA in either group of subjects.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

It has been widely accepted that the mechanism of adverse reactions induced by aspirin and other NSAIDs in the upper and lower airways is related to inhibition of cyclooxygenase (4). However, there is little direct experimental evidence to explain the relationship between inhibition of cyclooxygenase and development of respiratory symptoms in ASA-sensitive patients. The major obstacle to identify putative defect in AA metabolism in ASA-sensitive patients is the fact that the target cell for aspirin in this adverse reaction has not been identified, and thus, no biochemical abnormality could be confirmed at the cellular level. Testing the hypothesis that the airway ECs are the target cells for aspirin in sensitive patients, we are the first to demonstrate that AA metabolism in nasal polyp ECs from ASA-sensitive patients differs significantly from that of ASA-tolerant patients. Unstimulated ECs from aspirin-sensitive patients generated significantly less PGE₂ than cells from aspirin-tolerant patients, and this difference was evident also after stimulation of cells with calcium ionophore or after inhibition with aspirin. Such a difference may reflect either a decreased overall metabolism of AA in ECs or a decreased release of prostaglandins into the extracellular compartment in ASA-sensitive patients. Alternatively, an impairment of the metabolic pathway responsible specifically for PGE₂ production may be conceivable.

Prostaglandins in the airway ECs are generated by PGH synthase which may exist as two isoforms: cyclooxygenase-1 (COX-1) is constitutively expressed in almost all types of cells, and COX-2 is induced by several types of stimuli including proinflammatory cytokines and growth factors. We have recently reported that nasal polyp ECs, cultured in the system described in this study, demonstrated little expression of COX-1 messenger RNA (mRNA), but a significant COX-2 mRNA signal is detected which can be further enhanced by stimulation with lipopolysaccharide (16). Decreased capacity of ECs from ASA-sensitive patients to release PGE₂ is further confirmed by the effect of stimulation with calcium ionophore.

Taken together, these data point to the conclusion that ECs from ASA-sensitive patients demonstrate intrinsic defect in their capacity to generate PGE₂. This impairment seems to be selective, i.e., limited to one prostaglandin (PGE₂), because the mean basal levels of PGF₂ were similar in both groups and calcium ionophore-stimulated release of PGF₂ was almost twofold higher in ASARS versus ATRS. However, it is not clear whether this defect is limited to the upper airway ECs or may be generalized to others, e.g., inflammatory cells. In vitro studies demonstrated that PGE₂ concentrations in nasal polyp specimens, in platelets, and in nasal or bronchial lavages were not significantly different between ATRS and ASARS patients (17). Only Yamashita (20) observed decreased generation of PGE₂ in nasal polyps extracts from ASARS patients. These data suggest that diminished production of PGE₂ may be rather a local phenomenon limited to the airway nasal polyp ECs. Our observation is in line with a recent study demonstrating a significant downregulation of COX-2 mRNA expression in nasal polyps from ASA-sensitive asthmatics (21).

Both groups of patients studied by us apparently were not matched with regard to the presence or absence of bronchial asthma, reflecting lower prevalence of bronchial asthma in the population of ASA-tolerant patients with nasal polyposis subjected to routine polypectomy. Because we could not exclude that the presence or absence of asthma might affect the difference in the PGE₂ concentrations between ASA-sensitive and ASA-tolerant patients, we analyzed the individual values for PGE₂ taking into account asthmatic status of the patients. In one of two asthmatic and ASA-tolerant patients (in the second patient PGE₂ was not measured), the basal PGE₂ level was 4.45 ng/µg dsDNA. This value was 5 times higher than the mean value for asthmatic but ASA-sensitive patients (0.8 ± 0.3 ng/µg dsDNA) and close to the mean value for ASA-tolerant nonasthmatic patients (2.5 ± 0.5 ng/µg dsDNA), suggesting that the lower concentrations of PGE₂ in ASA-sensitive patients are not related to their asthmatic status, but rather to the presence of ASA sensitivity.

Our observation seems to provide new insight into the mechanism of ASA-induced hypersensitivity reactions. It has been proposed that endogenous PGE₂ has important protective and anti-inflammatory functions in the airways. Schafer and coworkers (22) demonstrated that exogenous PGE₂ has an inhibitory effect on the release of peptidoleukotrienes in human bronchial biopsy specimens. It has been suggested that ASA-sensitive patients have reduced overreliance on PGE₂. Inhalation or ingestion of PGE₂ (or its analogue misoprostol) inhibits ASA-induced respiratory symptoms (23).

The earlier study of Szczeklik and coworkers (24) demonstrated that prostaglandin synthesis in nasal polyp specimens from ASA-sensitive patients is more susceptible to inhibition by aspirin as compared with ASA-tolerant patients. Although our experiments demonstrated similar relative degree of prostaglandin inhibition by ASA in sensitive and tolerant patients, postaspirin concentrations of PGE₂ were significantly lower in ASA-sensitive patients. It is possible that in ASA-sensitive patients aspirin may drive PGE₂ synthesis to lower levels resulting in activation of inflammatory cells including mast cells and eosinophils and increased generation of peptidoleukotrienes. It can be assumed that similar mechanisms operate in the lower and upper airway mucosa, including nasal polyps of ASA-sensitive patients. However, because we did not measure eicosanoid generation by airway ECs cultured from nonpolypous nasal or bronchial mucosa, we cannot exclude that abnormalities of AA metabolism observed in this study in ASA-sensitive patients are not generalized to other organs (tissues), but are specific for nasal polyp ECs.

Chronic rhinosinusitis in ASA-sensitive patients is usually more severe, and involves more sinuses as documented by the semiquantitative computer tomography scoring system, when compared with ASA-tolerant patients (25). Because PGE₂ has significant anti-inflammatory activity, including inhibitory effect on eosinophil chemotaxis and activation, one may speculate that an intrinsic defect in local generation of PGE₂ may contribute to more severe eosinophilic inflammation in ASA-sensitive patients. Interestingly, a decreased PGE₂ synthesis has been attributed to enhanced inflammatory and fibrogenic responses in the lower airways of patients with idiopathic pulmonary fibrosis (26). AA metabolism in nasal polyps and corresponding nasal mucosa has been intensively studied (27, 28). However, because all these studies used either biopsies or tissue explants containing infiltrating inflammatory cells, it is not possible to compare them to our results reflecting pure culture of ECs.

The second important observation in our study is the differential effect of ASA challenge on lipoxygenase metabolism, namely on acid 15-HETE generation in ASA-sensitive and ASA-tolerant patients. Although generation of 15-HETE by unstimulated nasal polyps ECs was similar in groups, incubation with aspirin caused statistically significant enhancement of 15-HETE generation in ASA-sensitive patients and did not change (in fact, in four of six cases inhibited) 15-HETE synthesis in the ASA-tolerant group. Only one patient in the ASA-sensitive group did not respond to ASA with enhancement of 15-HETE generation, and the relative enhancement ranged from 44% to 1,673% in individual patients. On the contrary, four of six ASA-tolerant patients responded to aspirin with marked decrease in 15-HETE generation (23 to 56% inhibition from control); in one other patient no change was observed, and only in one patient an increase in 15-HETE was observed after ASA. The differential response to ASA was confirmed, although with less consistency, in the dose-response experiments demonstrating progressive enhancement of 15-HETE generation with increased dose of ASA in aspirin-sensitive patients and also dose-dependent decrease in 15-HETE release in ASA-tolerant patients (data not shown).

These data suggest that enhanced release of 15-HETE after ASA challenge is an abnormality specific to ASA-sensitive epithelium and may be causally related to the mechanism of ASA-induced reaction. 15-HETE has potent proinflammatory activity: it stimulates the release of mediators from mast cells (29) and induces the release of mucous glycoprotein from human airways in culture. If 15-HETE is inhaled before antigen challenge, it is capable of enhancing the early bronchoconstrictor response to inhaled allergen in atopic asthmatics (30). Additionally, 15-HETE can also be converted to lipoxins, substances that contract airway smooth muscle (31), and elevated levels of 15-HETE were detected in bronchial tissue taken from asthmatic patients (32). These data indicate that 15-HETE should be considered as a potential mediator of asthmatic and inflammatory reactions in the airway. If aspirin can exert a similar effect on 15-HETE generation in both upper and lower epithelium after in vivo challenge, an enhanced generation of 15-HETE may provide a specific mechanism contributing to the development of adverse symptoms in the airways.

On the other hand, 15-HETE has been reported to produce in certain systems anti-inflammatory activity (e.g., 5-lipoxygenase inhibition), and lipoxins derived from 15-HETE metabolism may have anti-inflammatory activity as well. Thus, the real meaning of increased 15-HETE generation by stimulated airway ECs remains unclear.

The mechanism of increased generation of 15-HETE by epithelial cells after in vitro incubation with aspirin is not known. The major source of 15-HETE in airways seems to be 15-lipoxygenase (15-LO) (33) and abundant expression of this enzyme has been detected in the bronchial mucosal ECs of patients with bronchial asthma and chronic bronchitis (34), as well as in the nasal mucosa. Thus, it is possible that this enzyme was a source of 15-HETE in ECs cultured from nasal polyps. However, it is difficult to explain how aspirin challenge in vitro could enhance 15-LO activity leading to increased generation of 15-HETE after in vitro ASA challenge.

One may speculate that the activity of 15-LO in ASA-sensitive patients is controlled by endogenous PGE₂, and that removal of PGE₂ production by aspirin results in enhancement of its activity. A similar mechanism has been previously postulated to explain enhancement of peptidoleukotriene release during ASA challenge in ASA-sensitive patients (35). Because this hypothesis has little experimental support, another mechanism may be considered. It has been demonstrated that, similarly to 15-LO which generates (S)-stereoisomer of 15-HETE, both COX-1 and COX-2 are capable of synthesizing 15-(R)-HETE, and that aspirin treatment, while inhibiting prostaglandin generation, stimulates formation of 15-HETE by COX-2 but not by COX-1 (36). Because we demonstrated that COX-2 mRNA is present in cultured human nasal polyp ECs (16), we hypothesize that aspirin enhances generation of 15-HETE in ASA-sensitive patients by activation of lipoxygenase activity of COX-2. The lack of ASA-induced enhancement of 15-HETE in ASA-tolerant patients could result from structural and/or functional differences of the enzyme between ASA-sensitive and ASA-tolerant patients. Although the immunoassay used in our study was, according to the manufacturer, S-stereoisomer-specific, the cross-reactivity of the antibody with 15-(R)-HETE is not known, thus the percentage share of R and S isomers is not known.

In conclusion, our study provides the first evidence for significant abnormalities in AA metabolism in the airway ECs of patients suffering from aspirin sensitivity. We detected abnormally low spontaneous generation of PGE₂ in ECs from ASA-sensitive patients, and differential response to stimulation with aspirin resulting in enhancement of 15-HETE production. These data suggest that airway ECs may be involved in the pathogenesis of either acute response to aspirin or development of chronic inflammation in the respiratory mucosa of ASA-sensitive patients.