Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Prostaglandins (PGs) and thromboxanes are potent modulators of biological function, particularly inflammation. They are produced via a complex enzyme cascade regulated by two principal enzymes: phospholipase A₂ (PLA₂) and cyclooxygenase (COX). PLA₂ mobilizes arachidonic acid from cellular phospholipids, which is then metabolized by COX to PGH₂ (Hamberg et al., 1974), the common substrate for a number of synthetase and isomerases essential to the formation of a variety of prostanoids.

PLA₂ exists in both calcium-dependent and -independent isoforms. The extracellular, or secretory, PLA₂ (sPLA₂) requires millimolar levels of calcium for activation, whereas the intracellular, or cytosolic, PLA₂ (cPLA₂) requires only nanomolar levels of calcium (for a review, see Mukherjee et al., 1994). Thus, sPLA₂ is activated continuously by the levels of calcium found in the extracellular environment, whereas cPLA₂ is activated via increases in intracellular calcium elicited by inflammatory mediators such as bradykinin (Burch and Axelrod, 1987; Slivka and Insel, 1988; Ricupero et al., 1993).

Similar to PLA₂, COX exists in multiple isoforms. A constitutive isoform (COX-1) is thought to be responsible for the "housekeeping" functions of the enzyme, whereas the inducible (COX-2) isoform (Xie et al., 1991; Mitchell et al., 1995) is thought to mediate inflammatory events. COX-2 is expressed in vitro in response to a number of proinflammatory mediators, including interleukin (IL)-1 (Lee et al., 1992; Hempel et al., 1994; Mitchell et al., 1994; Newman et al., 1994; Croxtall et al., 1996; Newton et al., 1996; Vadas et al., 1996; Belvisi et al., 1997), and in vivo at the site of inflammation (Vane et al., 1994; Chan et al., 1995).

Inflammation is orchestrated by mediators that are released in tissues in chronological order. For instance, the formations of amines (e.g., histamine) and kinins (e.g., bradykinin) often are early events that are followed by a later phase involving the infiltration of activated leukocytes and the release of cytokines (e.g., IL-1). In chronic inflammatory diseases such as rheumatoid arthritis or asthma, early- and late-phase inflammatory mediators may be released in synchronized cycles. Thus, activation of PLA₂ (e.g., with bradykinin) and expression of COX-2 (e.g., by IL-1) may occur simultaneously at the same site of inflammation. Therefore, to further understand how PLA₂ and COX-2 interact to produce prostanoids during inflammatory events, we investigated the effect of stimulating endogenous arachidonic acid release with bradykinin or A23187 or of directly supplying exogenous substrate on PGE₂ release from the human pulmonary cell line A549 induced with IL-1 to express COX-2 (Mitchell et al., 1994, 1997). We addressed the roles of cPLA₂ and sPLA₂ in these responses by using "specific" inhibitors of these enzymes. Furthermore, we used the B₁ agonist Des-Arg¹⁰-kallidin and the B₂ antagonist Hoe140 to establish which bradykinin receptor is responsible for prostanoid release in these cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Culture of Human Pulmonary Cell Line A549. The human pulmonary cell line A549 (Giard et al.,1973) was purchased from American Type Culture Collection (Rockville, MD). Cells were cultured onto either 6- or 96-well plates in Dulbecco's modified Eagle's medium containing 2 mM calcium (unless otherwise stated) supplemented with 10% fetal calf serum, 2 mM L-glutamine, 100 µg/ml streptomycin, 100 units/ml penicillin, and 2.5 µg/ml amphotericin B, an antifungal agent. Cells were treated with IL-1 or, in some experiments, medium controls at confluence.

Cell Treatments to Assess Release of PGE₂ from A549 Exposed to Bradykinin, Des-Arg¹⁰-Kallidin, Arachidonic Acid, or Calcium Ionophore A23187. Cells were exposed to IL-1 (10 ng/ml) or medium for 24 h. Medium was then removed for analysis of PGE₂ by radioimmunoassay (Mitchell et al., 1993, 1994), and fresh medium was added alone or containing either bradykinin (1 pM to 100 µM), the B₁-specific agonist Des-Arg¹⁰-kallidin (1 pM to 100 µM), A23187 (100 nM to 300 µM), or arachidonic acid (100 nM to 300 µM). After 15 min, the medium was removed for analysis of PGE₂. In subsequent experiments, cells were pretreated for 30 min (24 h after IL-1) with either the B₂ antagonist Hoe140 (Hock et al., 1991), the mixed COX-1/COX-2 inhibitor indomethacin (0.01-10 µM) (Meade et al., 1993; Mitchell et al., 1993), the COX-2-selective inhibitor L-745,337 (0.01-10 µM) (Chan et al., 1995), or the PLA₂ inhibitors arachidonate trifluoromethyl ketone (AACOCF₃), palmitoyl trifluoromethyl ketone (PACOCF₃), aristolochic acid, or 12-ep-scalaradial. Cells then were stimulated for 15 min with equieffective concentrations of bradykinin (1 µM), A23187 (10 µM), and arachidonic acid (30 µM) for PGE₂ release, measured by radioimmunoassay as above.

Western Blot Analysis. Western blot analysis was performed as described previously (Mitchell et al., 1993,1994). Briefly, cells were grown to confluence on 6-well plates and were treated for 24 h with either vehicle or IL-1 (10 ng/ml). After 24 h, the medium was removed, and the cells were washed with Hanks' balanced salt solution. The cells then were incubated with an extraction buffer [50 mM Tris, 10 mM ethylenediaminetetraacetic acid, 1% Triton X-100 (v/v), 1 mM phenylmethylsulfonyl fluoride, 50 µM pepstatin A, and 0.2 mM leupeptin]. The resulting cell extract was boiled with gel loading buffer [50 mM Tris, 10% sodium dodecyl sulfate (SDS), 10% glycerol, 10% 2-mercaptoethanol, and 2 mg/ml bromophenol blue] in a ratio of 1:1. Approximately 10 µg of protein, determined by Bradford protein assay, was loaded onto a 4% SDS stacking/7.5% SDS separating gel. After electrophoretic separation (1.5 h at 125-200 V), the samples were transferred to nitrocellulose (1 h at 0.3 A; Bio-Rad, Hercules, CA) and primed with a specific polyclonal antibody raised against murine COX-2 (Chan et al., 1995). The blot then was incubated with a secondary antibody linked to horseradish peroxidase and visualized using enhanced chemiluminescence (ECL; Amersham, Arlington Heights, IL).

Cell Viability. At the end of each treatment, the cell viability was assessed by the mitochondrial-dependent reduction in 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide to formazan (Mitchell et al., 1994). None of the cell treatments outlined above had an effect on cell viability unless otherwise stated.

Purified COX-2 Activity. As described previously (Mitchell et al., 1997), ovine purified COX-2 was obtained from Cayman Chemical Co. (Ann Arbor, MI). Pure COX-2 and the cofactors glutathione (5 mM), epinephrine (5 mM), and hematin (1 µM) were dissolved in 50 mM Tris buffer, pH 7.5. Hematin was first dissolved in a concentrated stock of 100 mM in 1 M NaOH before being further diluted in Tris buffer. Enzyme reactions were carried out in individual wells of 96-well plates with a final reaction volume of 200 µl. Different concentrations of the PLA₂ inhibitors were added to the plate, followed by the addition of 10 units of enzyme (180 µl). The plates were incubated at 37°C for 30 min before arachidonic acid (30 µM) was added for an additional 15 min. The reaction was stopped by heating the plate to 100°C for 5 min. The 96-well plate was then centrifuged at 10,000g for 10 min, and appropriate samples were removed for the measurement of PGE₂ by radioimmunoassay.

Materials. Bradykinin, Des-Arg¹⁰-kallidin, and Hoe140 were purchased from Scientific Marketing Associates (Barnet, Hertfordshire, UK). IL-1 was purchased from Boehringer-Mannheim (Lewes, East Sussex, UK). COX-2 antibody and L-745,337 were a kind gift from Dr. Ian Rodger (Merck Frosst, Point du Claire, Quebec, Canada). The PLA₂ inhibitors arachidonate trifluoromethyl ketone (AACOCF₃), palmitoyl trifluoromethyl ketone (PACOCF₃), aristolochic acid, and 12 epi-scalaradial were purchased from Calbiochem Novabiochem (San Diego, CA).

Statistical Analysis. Results are shown as the mean ± S.E.M. from n determinations. Where appropriate, data were analyzed by Kruskal-Wallis nonparametric analysis of variance test followed by Dunn's test for multiple comparisons. All treatments were compared with control values, and p < .05 was considered to be significant. Apparent pK_b was determined using the following equation: Apparent pK_b = log(dose ratio  1)  log[antagonist].

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effect of IL-1 on COX-2 Expression in A549 over 24 h. As we (Mitchell et al., 1994, 1997) and others (Newman et al., 1994; Croxtall et al., 1996; Newton et al., 1996) have previously shown, A549 released undetectable (<0.2 ng/ml) levels of PGE₂ and expressed undetectable levels of COX-2 protein (Fig. 1) when cultured under "control" culture conditions, measured after a 24-h period. However, when A549 cells were cultured in the presence of IL-1 (10 ng/ml), cells released 12.8 ± 1.5 ng/ml (n = 9) PGE₂ in 24 h and expressed COX-2 protein (Fig. 1).

View larger version (36K): [in this window] [in a new window]   Fig. 1.   Representative Western blot analysis of extracts from A549 with a specific antibody raised against COX-2 that recognizes a band of approximately 70 kDa. Each lane was loaded with approximately 10 µg of protein. Left, untreated A549 after 24 h. Right, A549 treated with IL-1 for 24 h.

Release of PGE₂ by Cells Pretreated with IL-1 followed by Stimulation with Bradykinin, Des-Arg¹⁰-Kallidin, A23187, or Arachidonic Acid. In contrast to results obtained with COX-2-expressing cells incubated for 24 h (as above), A549 cells, pretreated with IL-1 for 24 h, released undetectable levels of PGE₂ when exposed to medium alone for 15 min (Fig. 2). However, COX-2-expressing cells released PGE₂ in a concentration-dependent manner when exposed to bradykinin (Fig. 2), A23187 (Fig. 3), or arachidonic acid (Fig. 3) with E_max values for 15 min in excess of those obtained with IL-1 alone for 24 h. The B₁-selective agonist Des-Arg¹⁰-kallidin had no significant effect on PGE₂ release from IL-1-stimulated A549 cells (n = 9; Fig. 2). The release of PGE₂ elicited by bradykinin was inhibited by the B₂ antagonist Hoe140 (1 µM), given 30 min before bradykinin treatment, with an apparent pK_b of 7.8 ± 0.2. Arachidonic acid also elicited PGE₂ release from IL-1-treated cells in a concentration-dependent manner; although due to toxic effects, it was not possible to achieve a maximal response (Fig. 3b; n = 9). Bradykinin (Fig. 2), Des-Arg¹⁰-kallidin, A23187, or arachidonic acid did not release detectable PGE₂ from A549 cultured without IL-1.

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Release of PGE2 from A549 cells pretreated with IL-1 (10 ng/ml) for 24 h followed by a 15-min exposure to bradykinin. A, synergistic relationship between IL-1 (10 ng/ml) and bradykinin (1 µM) for the release of PGE2 from A549 cells. Similar data were obtained when cells were stimulated with A23187 or arachidonic acid. B, concentration-dependent release of PGE2 by bradykinin in cells pretreated with IL-1 (). , release of PGE2 by cells pretreated with IL-1 and stimulated with the B1-selective agonist Des-Arg10-kallidin. Data represent the mean ± S.E.M. for nine determinations performed on 3 experimental days.

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Demonstrates the release of PGE2 from A549 cells pretreated with IL-1 (10 ng/ml) for 24 h followed by a 15-min exposure to (A) the calcium ionophore A23187 or (B) exogenous arachidonic acid for 15 min. Data represent the mean ± S.E.M. for six determinations performed on 3 experimental days.

Effects of COX-2-Selective Inhibitor L-745,337 and COX-1/COX-2 Inhibitor Indomethacin on Bradykinin-, A23187-, or Arachidonic Acid-Stimulated PGE₂ Release. In COX-2-expressing cells, the release of PGE₂ stimulated by equieffective, for PGE₂ release, concentrations of bradykinin (1 µM), A23187 (10 µM), and arachidonic acid (30 µM) was inhibited in a concentration-dependent manner by indomethacin with a similar potency for each agonist (Fig. 4; n = 6). Moreover, L-745,337 inhibited PGE₂ release stimulated by bradykinin more potently than that stimulated by A23187 or arachidonic acid (Fig. 4; n = 6).

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Inhibitory effects of indomethacin (A) or L-745,337 (B) on COX activity in IL-1-primed A549 cells stimulated with bradykinin (BK; ; 1 µM), A23187 (; 10 µM), or arachidonic acid (AA; ; 30 µM). COX was measured by the release of PGE2 and calculated as percentage release in the absence of COX inhibitor (control). Data represent the mean ± S.E.M. for six determinations carried out on 3 experimental days.

Effects of AACOCF₃ and Other PLA₂ Inhibitors on COX-2 in Intact Cells and in Purified Form. In separate experiments, the so-called cPLA₂ inhibitor AACOCF₃ (Bartoli et al., 1994) caused concentration-dependent inhibitions of PGE₂ release by A549 cells, expressing COX-2, and stimulated with equieffective (for PGE₂ release) concentrations of bradykinin (1 µM), A23187 (10 µM), or arachidonic acid (30 µM) (Fig. 5; n = 6). However, AACOCF₃ directly inhibited purified COX-2 activity at similar concentrations (Fig. 5; n = 4). By contrast, the inhibitors PACOCF₃ (calcium-independent PLA₂ inhibitor; Ackermann et al., 1995), aristolochic acid (inhibitor of snake venom, human platelet, and synovial fluid sPLA₂ as well as cPLA₂), and 12-epi-scalaradial (inhibitor of bee venom and neutrophil PLA₂ as well as cPLA₂) had no effect at concentrations up to 100 µM on COX-2 activity in whole cells stimulated with 1 µM bradykinin (control, 100%; plus 100 µM PACOCF₃, 101 ± 15%; plus 100 µM 12-epi-scalaradial, 105 ± 10%; plus 100 µM aristolic acid, 68 ± 25%), 10 µM A23187 (control, 100%; plus 100 µM PACOCF₃, 108 ± 1%; plus 100 µM 12-epi-scalaradial, 69 ± 28%; plus 100 µM aristolic acid, 95 ± 15%), or 30 µM arachidonic acid (control, 100%; plus 100 µM PACOCF₃, 92 ± 6%; plus 100 µM 12-epi-scalaradial, 72 ± 19%; plus 100 µM aristolic acid, 74 ± 22%) (n = 4-6). In addition, at concentrations of 100 µM, no effect of PACOCF₃ (102 ± 20%), 12-epi-scalaradial (75 ± 30%), or aristolic acid (81 ± 25%) was seen on PGE₂ production by purified COX-2 (n = 4).

View larger version (17K): [in this window] [in a new window]   Fig. 5.   A, effect of the cytosolic phospholipase A2 inhibitor AACOCF3 on COX-2 activity in whole cells pretreated with IL-1 for 24 h before being stimulated for 15 min with either bradykinin (BK), arachidonic acid (AA), or A23187. B, effect of AACOCF3 on purified COX-2 activity measured in the presence of 30 µM arachidonic acid. Data represent the mean ± S.E.M. for six (A) or four (B) experiments.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

This study demonstrates that when COX-2 is induced in intact cells, a secondary inflammatory stimulus is required to achieve maximum prostanoid release. Thus, a "two-component" system is required for prostanoid release in cells expressing COX-2 (Fig. 6). However, we and others (Mitchell et al., 1995) have demonstrated, in a variety of cell types, that induction of COX-2 stimulates the release of PGE₂ (usually measured over a 24-h period) without the need for a secondary agonist. Here, we confirm that after 24 h, IL-1 induces COX-2 and releases PGE₂; however, this release proved to be very low, in fact undetectable, when measured over a 15-min period. This contrasts with the large amounts of PGE₂ released when cells were treated with IL-1 followed with secondary stimuli such as bradykinin, A23187, or exogenous arachidonic acid. The "low" level of PGE₂ release by cells stimulated with IL-1 alone may well reflect the basal levels of arachidonic acid present in cells. Alternatively, IL-1 may be acting on both components, inducing COX-2 and partially activating PLA₂ (Fig. 6). In addition, cytokines are able to induce cPLA₂ and sPLA₂ (Kol et al., 1997; Newton et al., 1997; Pruzanski et al., 1998) Thus, IL-1 in this study may also be up-regulating PLA₂ enzymes. Nevertheless, in our experiments, we observed a strict synergism between IL-1 pretreatment and stimulation with either bradykinin, A23187, or arachidonic acid on PGE₂ release from A549 cells.

View larger version (22K): [in this window] [in a new window]   Fig. 6.   Release of PGs by a two-component model. PG release is relatively low when cells are stimulated to either induce COX-2 (e.g., with IL-1) (component 1) or activate PLA2 (e.g., with bradykinin) (component 2). However, PGs are released in relatively high proportions when both COX-2 is induced and PLA2 is activated. The two components are essentially related to amount of active COX-2 present and concentration of arachidonic acid available to it.

Bradykinin is known to increase intracellular calcium in the A549 (Levesque et al., 1997) from a resting level to approximately 800 nM. In addition, A23187 has been shown to elicit similar increases in intracellular Ca⁺⁺ in a number of in vitro systems (Vargaftig et al., 1980; Bradford et al., 1983). As such, bradykinin or A23187 will activate the cytosolic, intracellular form of PLA₂, which requires an increase in calcium in the nanomolar range for activation and translocation to the nuclear and plasma (Glover et al., 1995, Schievella et al., 1995) membranes. Thus, bradykinin and A23187 are likely to be releasing PGE₂ in our cells by activation of a calcium-dependent PLA₂, either cPLA₂ or sPLA₂. Indeed, it should be noted that at high concentrations, A23187 may stimulate elevations in intracellular calcium that are sufficiently high to activate both cPLA₂ and sPLA₂. However, because the E_max for PGE₂ production by cells stimulated with bradykinin and A23187 were very similar, we may conclude that in each case, common levels of arachidonic acid were available to COX-2. Moreover, we found that the release of PGE₂ stimulated either by bradykinin or A23187 from IL-1-pretreated cells, was inhibited in a concentration-dependent fashion by the cPLA₂ inhibitor AACOCF₃ (Bartoli et al., 1994; Riendeau et al., 1994). The observations are in keeping with others (Croxtall et al., 1996) showing that AACOCF₃ inhibits arachidonic acid release by primed A549 cells. However, we also found that the release of PGE₂ stimulated by arachidonic acid was similarly inhibited by AACOCF₃. Experiments performed in the presence of exogenous arachidonic acid should remove in part the requirement for PLA₂. It must be noted that arachidonic acid may also activate cells leading to liberation of substrate from intracellular stores. However, we found that AACOCF₃ directly inhibited COX-2 in purified form at similar concentrations to those required for inhibition in whole cells. Thus, we show, for the first time, that AACOCF₃ is a COX-2 inhibitor, an observation that renders this drug inaccurate as a tool to study the role of cPLA₂ in prostanoid release in our system. In addition, the commercially available inhibitors of other forms of PLA₂ (aristolochic acid, PACOCF₃, or 12-epi-scalaradial) (Dennis, 1997) were relatively inactive as inhibitors of PGE₂ release by intact cells or formation by purified COX-2. Thus, we are not able to comprehensively address which PLA₂ (or indeed which lipase) is responsible for arachidonic acid release in our cells. However, it is clear that substrate liberation is an essential component of prostanoid release by cells expressing COX-2.

Kinins such as bradykinin and kallidin act via two distinct receptor subtypes, B₁ and B₂. In this study, we show that the release of PGE₂ stimulated by bradykinin was inhibited by a B₂ antagonist. Moreover, the B₁-selective agonist Des-Arg¹⁰-kallidin had no effect on prostanoid release. Thus, in A549 cells, bradykinin appears to be acting via its B₂ receptors. Interestingly, recently, the expression of B₂ receptors has been shown to increase in response to proinflammatory cytokines (Haddad et al., 1997). This, together with COX-2 induction, may help to explain why bradykinin is particularly active in and important for a variety of inflammatory responses.

We found that bradykinin and A23187 elicit the release of similar levels of PGE₂. Because A23187 acts independently of receptors, these observations suggest that either COX-2 or PLA₂/arachidonic acid could be rate limiting. However, we found that cells stimulated with arachidonic acid released far greater levels of PGE₂ than cells stimulated with either bradykinin or A23187. The way in which arachidonic acid stimulates the release of PGE₂ from cells is not straightforward. For example, in addition to increasing substrate levels available to COX, arachidonic acid can stimulate cells (Damron et al., 1993). Indeed, we may expect that a portion of PGE₂ formed by A549 cells stimulated with arachidonic acid originates from endogenous stores. Nevertheless, exogenous arachidonic acid was able to produce a much higher release of PGE₂ than either bradykinin or A23187, suggesting that COX-2 is not the rate-limiting factor in prostanoid release, even when PLA₂ is stimulated.

The absence of detectable COX-1 (Mitchell et al., 1994) and the expression of COX-2 protein in A549 cells after IL-1 stimulation (Mitchell et al., 1994, 1997; this study) suggest that PGE₂ production is mediated via COX-2. However, because COX-1 still may be present in small amounts and COX-1/COX-2 are located on distinct cellular membranes (Morita et al., 1995), it remains possible that bradykinin or A23187 can access different subcellular stores of arachidonic acid and hence preferentially supply one COX isoform. To further address this idea, we looked at the effect of the COX-2-selective inhibitor L-745,337 (Chan et al., 1995) and the mixed COX-1/COX-2 inhibitor indomethacin (Meade et al., 1993; Mitchell et al., 1993) on PGE₂ release stimulated by bradykinin, A23187, or arachidonic acid. Chan et al. (1995) demonstrated that indomethacin was approximately equipotent as a inhibitor of COX-1 and COX-2, whereas L-745,337 had a similar potency as indomethacin against COX-2 but was 500-fold less potent against COX-1, with an IC₅₀ value of >10 µM. As such, it is possible to use these two nonsteroidal anti-inflammatory agents to establish pharmacologically which COX isoform is responsible for the production of PGE₂ from A549. Both indomethacin and L-745,337 inhibited the release of PGE₂ stimulated by bradykinin, A23187, or arachidonic acid with similar potencies, suggestive of COX-2 being responsible for the prostanoid formation. However, L-745,337 was slightly more potent as an inhibitor of PGE₂ release stimulated by bradykinin than by either A23187 or arachidonic acid, which may support the idea of different pools of arachidonic acid being available to COX-2 (Reddy et al., 1996).

Prostanoids are known to have a pivotal role in the inflammatory process. Indeed, the therapeutic benefits of nonsteroidal anti-inflammatory drugs are attributable to the inhibition of prostaglandin production (Vane, 1971). The "discovery" of a second isoform of COX, COX-2, has propagated research into the regulation of each isoform and their individual roles in overall prostanoid biosynthesis. Traditionally, it has been considered that prostanoid release was a one-component system. The only component required for COX-1-expressing cells was stimulation of PLA₂ with agonists such as bradykinin, and for inflammatory cells, the component was induction of COX-2. The data presented here suggest that cells expressing the inflammatory isoform of COX-2, like those expressing COX-1, require a secondary stimulus to activate cPLA₂ with subsequent release of prostanoids, thus establishing the requirement for a two-component system in prostanoid release at the site of inflammation. These observations may help to explain how the perpetuation of chronic inflammation occurs in diseases such as asthma and rheumatoid arthritis.