Decreased Airway Mucosal Prostaglandin E2 Production during Airway Obstruction ill an Animal Model of Asthma 1 , 2

PETER R. GRAY, FREDERIK J. DERKSEN, RICHARD V. BROADSTONE, N. EDWARD ROBINSON, and MARC PETERS-GOLDEN

Introduction

Airway responsiveness is the term used to describe the ease with which airways narrow in response to a variety of stimuli. Hyperresponsiveness to such stimuli is considered a cardinal characteristic of human asthma, and also of heaves, a naturally occurring respiratory disorder of horses and ponies that has similarities to human asthma (1). Reaves has an allergic etiology and is precipitated by housing susceptible animals in a barn and exposing them to dusty/moldy hay. Removalfrom the barn results in remission of clinical signs (2-4). The disease is thus characterized by episodes of acute airway obstruction followed by periods of disease remission. During these episodes of airway obstruction, affected animals demonstrate airway hyperreactivity to histamine, methacholine, and citric acid, which wanes during periods of disease remission (2, 4-6). Studies performed in vitro using tracheal smooth muscle collected from control horses and horses with heaves suggest that epithelial tissue from control, but not affected, animals produces a relaxant factor that modulates smooth muscle reactivity (7, 8). Airway epithelium produces substances that can modulate the responsiveness of underlying airway smooth muscle to contractile agonists (9-11). In some species, this epithelium-derived relaxant factor appears to be a cyclooxygenase metabolite, most probably prostaglandin E 2 (PGE2 ) (12-14), which is released by intact airway epithelium in response to a variety of stimuli (15-17). PGE2 is a bronchodilator in normal and asthmatic humans (18, 19),and can reverse bronchoconstriction caused by PGF2 u administration (20). In vitro, it relaxes isolated human (21), canine (12),and equine (22) tracheal smooth muscle, possibly by acting presynaptically to modulate acetylcholine release or postsynaptically to alter the responsiveness of the muscle to acetylcholine (23-25). Decreased produc586

SUMMARY Heaves is a respiratory disorder of horses and ponies characterized by bouts of acute airway obstruction and airway hyperresponslveness. We measured prostaglandin E2 (PGE2 ) and 15hydroxyelcosatetraenoic acid (15-HETE) production In vitro In tracheal epithelium obtained from six affected horses at the time of acute airway obstruction as compared with six matched control horses. Strips of epithelium and subepithelial tissue were prepared and stimulated with A23187, histamine, and bradykinin. The PGE2 and 15-HETEin media from strips was quantitated by radioimmunoassay. 15-HETE above the limits of accurate detection was found In epithelial strips of only two principal animals and In none of the control horses, and the amount of 15-HETE was not Increased when strips were stimulated. Epithelial strips from affected horses tended to produce less PGE2 than did strips from control horses, and there was a significant correlation between epithelial PGE2 production and the time taken for affected animals to develop airway obstruction. Subepithelial tissue strips from control horses produced significantly more PGE2 In response to A23187and bradykinin than did strips from affected horses. We conclude that equine tracheal opithellum Is not a significant source of 15-HETE. Airway mucosal PGE2 production Is reduced In horses with heaves, which suggests that a relative decrease In this bronchorelaxant substance may be a factor In the pathogenesis of this model of asthma. AM REV RESPIR DIS 1992;146:586-591

tion of PGE 2 by tracheal epithelium has been associated with endotoxin-induced hyperreactivity in guinea-pig isolated trachea (14). PGE2 is the major prostanoid released by tracheal epithelium from normal horses (26), but there are no reports documenting epithelial PGE 2 production in horses with heaves. One possible explanation for the hyperresponsiveness and airway obstruction seen with affected animals is that there is decreased epithelial PGE2 production in these animals. To test this hypothesis, in the present study we measured PGE 2 release by airwayepithelial and subepithelial tissue obtained from affected horses during disease episodes, and from control horses housed under the same conditions. Recently, we reported that plasma 15hydroxyeicosatetraenoic acid (IS-RETE) concentrations are increased in ponies with heaves relative to controls, that the production of 15-HETE increases during the occurrence of airway obstruction, and that the lung is the source of this increase (27). These findings are of interest because IS-RETE concentrations are-increased in human asthma (28) and because IS-RETE is a potent secretagogue (29), causes neutrophil chemotaxis (30),

and can activate or inhibit the 5-lipoxygenase enzyme depending on the cell type (31,32). The pulmonary epithelium is a potential source of 15-HETE in human airways (33). Therefore, in the present experiment we also measured the 15HETE production of airway epithelium obtained from horses with heaves and from control horses. Methods Six horses with a history of consistently developingairway obstruction and hyperresponsiveness when housed in a barn (principals) were obtained and paired with six horses with no history of heaves or evidence of respiratory disease. Each pair was housed, fed, and

(Receivedin originalform August 19, 1991 and in revised form March 16, 1992) 1 From the Pulmonary Laboratory, Department of Large Animal Clinical Sciences, Michigan State University, East Lansing, and the University -of Michigan and VA Medical Centers, Ann Arbor, Michigan. 2 Correspondence and requests for reprints should be addressed to Frederik Derksen, DY.M., Ph.D., Department of Large Animal Clinical Sciences, VeterinaryClinical Center, Michigan State University, East Lansing, MI 48824-1314.

587

TRACHEAL MUCOSAL PGE, PRODUCTION IN HEAVES

transported together to ensure the same environmental exposures.

Experimental Protocol Baseline pulmonary function was measured while horses were kept at pasture with no exposure to hay, straw, or a barn environment (Period A). They were then housed in a barn, bedded on straw, and fed hay. When the principal horse developed labored breathing and hypoxemia (Pa021ess than 75 mm Hg), pulmonary function measurements wererepeated on both horses. The principal horse was then euthanized and tracheal epithelium obtained for in vitro studies. The control horse was kept in the barn for a further 48 h, and pulmonary function tests were repeated. The horse was then euthanized and tracheal epithelium was collected. Pulmonary Function Measurements A facemask was placed over the horse's face and sealed against the face using a rubber shroud. A pneumotachograph (Fleisch No. 5; Dynasciences, Blue Bell, PA) with associated pressure transducer (Model DP45-22; Validyne Corp., Northridge, CA) was attached to the mask. The pneumotachograph transducer system produced a signal proportional to flow (V) that was electronically integrated to give tidal volume (VT). Before each experiment this system was calibrated by forcing a known volume of air through the pneumotachograph using a 2-L syringe (Super syringe; Hamilton Syringe Co., Warminster, PA). An esophageal balloon (lO-cm length, 3.5-cm circumference, OJ>6-cm wall thickness) was sealed over the end of a polypropylene catheter (0.3-cm internal diameter, 0.44-cm external diameter, 140-cm length) that had several spirally arranged holes in the portion covered by the balloon. The balloon volume was adjusted to 0.5-ml with room air, and the catheter was inserted via the nares until the balloon was positioned in the midthoracic esophagus. The external end of the catheter was then connected to a pressure transducer (Model DP45-34; Validyne)for measurement of esophageal pressure.Transpulmonary pressure (PL) was defined as the pressure difference between atmospheric and esophageal pressure. PL, VT,and Vwererecorded on lightsensitive paper (VR12; Electronics for Medicine, White Plains, NY). Dynamic compliance (Cdyn), pulmonary resistance (RL),minute ventilation (VE), and frequency (f) were calculated by a pulmonary function computer (Model 6; BuxcoElectronics, Sharon, CT) and data logger (Model DL12;Buxco). Blood samples werecollected from the carotid artery and analyzed for Pa02 and Pac02 using an automated blood gas analyzer (Model ABL3; Radiometer, Copenhagen, Denmark). Collection and Processing of Tracheal Epithelium Strips of tracheal epithelium were obtained using a previously described method (26). Horses were euthanized with pentobarbital (100mg/kg administered intravenously) and

tissues were collected and processing begun within 20 min of administering the pentobarbital. Twelve-centimeter lengths of trachea wereplaced in ice-cold balanced salt solution (NaCI118.5 mM, KC14.7 mM, KH 2P04 1.2 mM, glucose 11.0mM, NaHC03 24.9 mM), which was bubbled with 02/C0 2 (95:5) and contained antibiotics (penicillin G 100U/ml, streptomycin 0.1 mg/ml, amphotericin 0.25 ug/ml [antibiotic, antimycotic; Sigma Chemical Co., St. Louis, MOJ). Tracheal segments were kept in this solution and bubbled with 02/C02 until ready for further processing. Each segment was divided in half longitudinally, pinned to a dissecting board with the mucosal surface uppermost, and rinsed with balanced salt solution to remove any blood or mucus and to keep the mucosal surface moist. A transverse incision was made through the epithelial layer into the underlying elastic lamina, taking care to remain superficial to the deeper submucosal tissues. The epithelium was teased back using Allis forceps, then grasped using half-curved iris dressing forceps(Jarit, Hawthorne, NY) and pulled gently downwards, producing strips of epithelium 150to 200-J.1m thick, 1.5-cmwide, and 12-cm long. The strips were placed in chilled Dulbecco's MEM/F-12 HAM medium (Sigma) containing 10070 newborn calf serum (Sigma) and antibiotics and kept on ice.

Epithelium-on Strips Strips of epithelium weretransferred to a fresh petri dish and trimmed to approximately the same size (15 x 10 mm) using a graph-paper template attached to the base of the petri dish. The surface area of each strip was recorded, and it was sutured to a cylindrical glass support. Each mounted strip was transferred to a wellof a 24-wellculture plate (Gibco, Grand Island, NY), and 1.5ml of M199(Sigma) containing L-glutamine (Sigma), 10% newborn calf serum, and antibiotics wereadded to each well. The culture plate was then placed in an incubator at 37° C until the strips were stimulated. Epithelium-off (Control) Strips These consisted of strips made from tracheal segments that had had the epithelial layer removed by first scraping across the mucosal surface with the flat edge of a scalpel blade and then rubbing firmly with a gauze swab. This method has previously been shown to adequately remove the epithelial layer (26). The resultant denuded strip was handled in the same manner as the intact epithelial strip. Agonist Stimulation Stock solutions of calcium ionophore (A23187), histamine, and bradykinin (Sigma) were prepared at a concentration such that a 25-J.1L aliquot added to 975 J.1L of media would produce the desired final agonist concentration. Histamine and bradykinin were used at a final concentration of 50 J.1M, and A23187 at a concentration of 5 J.1M, previously shown to be optimal. Histamine and bradykinin were dissolved in distilled water,

and A23187was dissolved in dimethylsulfoxide. Unstimulated controls had distilled water only added to the media. The final dimethylsulfoxide concentration in the medium did not exceed 0.5%, a concentration previously shown not to affect eicosanoid production by cultured alveolar epithelial cells (34). Agonists were prepared immediately before their addition to the tissues and kept on ice. Each agonist was evaluated in duplicate strips, and stimulation was performed approximately 4 h after the horse was euthanized. The culture plates containing the strips were removed from the incubator and the media aspirated and replaced with 975 J.1L of proteinfree M199 solution. Agonists were added to their respective strips and the plates returned to the incubator for 45 min. At the end of the stimulation period, media from each pair of strips were pooled and centrifuged, and the supernatant was divided into 3OO-J.1L aliquots and stored in polypropylene tubes at - 70° C. After the media were collected, each strip was placed in a dessicator for 24 hand then weighed to allow eicosanoid production to be expressed as pg per mg dry weight of tissue.

Quantification of PGE2 and 15-HETE Release by Radioimmunoassay Samples from each pair of horses were stored at - 70° C until tissues from all the horses had been collected and processed. Samples from the respective epithelium-on and epithelium-off strips were then analyzed as a single batch. PGE2antiserum was obtained from Advanced Magnetics (Cambridge, MA), tritiated PGE 2 from Dupont-New England Nuclear (Boston, MA), and authentic PGE2 standard from Cayman Chemical Co. (Ann Arbor, MI). Separation of bound from free trace was achieved with dextran-coated charcoal. The cross-reactivity of the antiserum at 50% normalized percent bound (B/B o) was PGE2 100%, PGEt 50%, PGA 2 6.0%, PGF 2a 1.3%, all other eicosanoids 1.0% or less. When necessary, samples were diluted with medium to produce values on the linear portion of the standard curve. The lower limit of accurate quantification for the assay was 12pg per tube. 15-HETE was measured using a commercial radioimmunoassay kit, obtained from Amersham (Arlington Heights, IL). The antiserum cross-reactivity at 50% B/B o was 15-HETE 100%, 15-HPETE 41%, 5,15-diHETE 2%, 13-HOOD 0.4%, all other eicosanoids 0.1% or less. The lower limit of accurate quantitation for the assay was 8 pg per tube. Statistical Analysis Pulmonary function values and eicosanoid release by epithelium-on and epithelium-off strips from principal or control groups were analyzed using Student's t test for paired values. Pulmonary function values and epithelium-on and epithelium-off PGE2 release for principal and control horses were compared using Student's t test for unpaired values. Correlations between PGE2 release

588

GRAY, DERKSEN, BROADSTONE, ROBINSON, AND PETERS·GOLDEN

TABLE 1 PULMONARY FUNCTION MEASUREMENTS IN PRINCIPAL AND CONTROL HORSES AT PASTURE (PERIOD A) AND DURING AIRWAY OBSTRUCTION (PERIOD B)· Period A Horses

~Ppl

Period B ~Ppl

RL

Principal (n = 6) 7.6 ± 0.3t 1.23 ± 0.18

0.91 ± 0.09

12.0 ± 1.8

94.8 ± 2.7

38.2 ± 0.9

Control (n = 6) 6.3 ± 0.4

0.63 ± 0.11

17.2 ± 1.05 95.3 ± 2.8

40.5 ± 1.1

1.37 ± 0.28

RL

36.1 ± 4.9 t:t: 0.31 ± 0.06t:t: 2.43 ± 0.26t:t: 17.6 ± 2.0t 60.6 ± 3.8t:t: 45.9 ± 1.4t:t: 6.7 ± 0.4

1.26 ± 0.15

Definition of abbreviations: Pao2 = arterial oxygen tension (mm Hg); Paco 2 = arterial carbon dioxide tension (mm Hg); Cdyn (cm H20/Us); ~Ppl = change in pleural pressure (cm H20); f = frequency (min-I). • Values are mean :I: SEM. t Significant difference from Period A value (p < 0.05). :j: Significant difference from control value (p < 0.05).

and time taken for principal horses to develop airway obstruction were performed using the Pearson correlation procedure. Significance was set at p < 0.05.

change in pleural pressure, and Pac02 were significantly increased, and Cdyn and Pa02 significantly decreased in principals relative to controls.

Results

15-HETE Release by Tracheal Epithelium Quantifiable amounts of IS-RETE were detected in epithelium-on strip supernatants from only two principal horses (table 2). IS-RETE did not increase when strips werestimulated withA23I87, histamine, or bradykinin.

Pulmonary Function Measurements In the principal horses, pulmonary resistance, respiratory frequency, change in pleural pressure during tidal breathing, and Pac0 2 were significantly increased, and dynamic compliance and Pa02 significantly decreased at Period B relative to Period A (table 1). No equivalent changes were present in the control horses. There were no significant differences in any of the pulmonary parameters in the control horses at either of the Period B measurement periods, so these values werecombined to provide a single Period B value for each parameter. At Period A, the change in pleural pressure was significantly greater in the principals relative to the controls. At Period B, RL,

PGE2 Release by Tracheal Epithelium Unstimulated epithelium-on strips from both principal and control horses produced measurable quantities of PGE 2. There was a significant increase over controllevels in PGE 2 release from epithelium-on strips from both groups in response to stimulation with A23I87, hista-

TABLE 2 IMMUNOREACTIVE 15·HETE RELEASE BY EPITHELIAL STRIPS FROM PRINCIPAL ANO CONTROL HORSES· Horse No. Principal 1 2 3 4 5

6 Control 1 2 3 4

5 6

Unstimulated

A23187

Bradykinin

Histamine

NO 21

NO 11

NO 21

NO 19

NO 12 NO NO

NO 11 NO NO

NO NO NO NO

NO NO NO NO

NO NO

NO NO

NO NO

NO NO

NO NO NO NO

35 NO NO 8

20 NO NO

14 NO NO NO

9

• Tissues were stimulated with calcium ionophore (A23187), bradykinin, and histamine. Values expressed as pg/100 III supernatant; 8 pg = lower limit of accurate quantification for assay; NO '" < 8 pg/100 Ill.

0.87 ± 0.10

13.9 ± 2.2

= dynamic compliance (Ucm

95.1 ± 1.7

H2 0 ); RL

42.6 ± 1.0

= pulmonary resistance

mine, and bradykinin. Stimulated and unstimulated strips from principal horses produced less PGE 2than did strips from controls, but the difference was not significant (figure 1). Expressing PGE2 production per strip, per mm" of epithelium, or per mg dry weight of epithelial tissue did not alter the interpretation of the data. In principal horses, there was a significant correlation betweenthe time to develop airway obstruction 'and the amount of PGE 2 released in response to A23I87 (r = 0.92, p = 0.009), and bradykinin (r = 0.86, p = 0.03) (figure 2), but not histamine (r = 0.74, p = 0.09). Epithelium-off strips released PGE 2, but the amounts were S to 10070 of those measured for epithelium-on strips. There was a significant increase in PGE 2 release in response to A23I87, histamine, and bradykinin in the controls, but not the principals. Epitheliumoff strips from principal horses produced less PGE 2 than did strips from control horses in response to A23I87, bradykinin, and histamine, and this difference was significant for A23I87 and bradykinin (figure 3). Tissue from only two of the six principals produced PGE2 in response to stimulation, whereas all six controls were able to produce PGE l • Discussion

In these experiments, weshowed that unlike human tracheal epithelium, equine tracheal epithelium is not an important source of IS-RETE. When detected, amounts of IS-RETE produced wereless than one tenth of those measured for PGE 2 production. Thus, these results suggestthat the tracheal epithelium is not the source of the increase in pulmonary IS-RETE production in ponies with heaves (27). Because the percentage of total airway surface area contributed by the trachea is small, this conclusion can-

589

TRACHEAL MUCOSAL PGE, PRODUCTION IN HEAVES

*

600 r

.-

*

I

Io.:l

~

Fig. 1. Immunoreactive PGE 2 production by epithelium-on strips from principal and control horses. Tissues were stimulated with A23187, bradykinin, or histamine. Values are mean ± SEM. Asterisk indicates significant difference from unstimulated value (p < 0.05).

o

o

400

...-t

--~ N

~

*

L

200

~ Unstimulated

Bradykinin

A23187

Histamine

Agonist ~ Control

not be extended to include all airway epithelium. Given the marked variation in cellular composition of epithelium throughout the lung, airway epithelium obtained from other lung regions may well have differing eicosanoid profiles. Several other cell types may have been the source of the 15-HETE produced in ponies with heaves (27). Holtzman and colleagues examined the ability of a wide range of human cell types to produce 15HETE (35). Of the cells examined, airway epithelial cells and eosinophils were major sources, producing approximately 5,000 and 3,000pmol/Io" cells, respectively. Eosinophil numbers are not consistently increased in bronchoalveolar lavage fluid or peripheral blood collected from horses with heaves(3), but increased eosinophil numbers are present in histologic sections of lung tissues from affected horses, especially around arteries, veins, and bronchioles (36). Therefore, it is possible that eosinophils are the source of 15-HETE reported in horses with heaves.

500

D

Principal

In these experiments, we also demonstrated that epithelial and subepithelial tissue from horses exhibiting airway obstruction produced less PGE2 than the same tissues from matched control horses. The decreased PGE2 production by epithelium-on strips was not statistically significant, due largely to a marked variation in PGE 2 production among individual animals. The difference in PGE2 production was more evident in the epithelium-off strips inasmuch as strips from controls produced significantly more PGE2 than did those from principals when stimulated with A23187 and bradykinin. The variability in PGE 2 production by the epithelium-on strips was accompanied by variability in the time taken for the principals to develop airway obstruction when placed in the barn. A similar variability in the time taken for principals to develop airway obstruction is routinely observed in our laboratory. If the ability of epithelial tissue to release PGE2 was an important factor in determining the onset of airway obstruction, then we would have expected to see some correlation between epithelial PGE 2 production and the time taken for airway ob-

struction to develop. Such a correlation existed between PGE 2 production in response to A23187and bradykinin and the number of days that principals took to develop airway obstruction. Epithelial strips that produced the greatest amount of PGE 2 tended to come from horses that took the longest time to develop clinical symptoms. These data support the theory that PGE2 produced by the airway mucosa could oppose the development of airway obstruction and airway hyperresponsiveness 0 bserved in this animal model (2). The PGE 2 produced by epithelial and subepithelial tissue may serve severalpurposes. An inflammatory response is present in the airway lumen in ponies with heaves (3), and decreased PGE2 production by the airway epithelium may be important in this process. PGE 2 can inhibit polymorphonuclear cell chemotaxis and modulate release of other lipid mediators such as leukotriene By (37) and reduce production of inflammatory mediators such as tumor necrosis factor, oxygen metabolites, and interleukin 1in vitro (38, 39). It may act similarly in vivo to modulate inflammatory responses. In contrast, PGE 2 released into the subepithelial tissues would be able to interact closely with nerve endings, airway smooth muscle, blood vessels, and nerve fibers supplying airway smooth muscle and submucosal glands, since many of these structures are located in, or pass through, the subepithelial layer. It is important to note, however, that PGE 2 in the alveolar lumen and interstitium is not compartmentalized and that PGE 2 is moved rapidly from the alveolar lumen to the interstitium and vascular space (40); therefore, PGE2 produced by airway epithelium might well influence events both in the airway lumen and in the subepithelial tissues. The amounts of

400

:::J

:::J

a ~

300

150 r

OJ

I

.e: N

.-

Io.:l

200

W

oCL. 100

Time (days) c

•

A23187 Bradykinin

Fig. 2. Correlation between the time taken for principal horses to develop airway obstruction (days) and PGE 2 production by epithelium-on strips in response to A23187 and bradykinin.

Fig. 3. Immunoreactive PGE 2-production by epithelium-off strips from principal and control horses. Tissues were stimulated with A23187, bradykinin, or histamine. Values are mean ± SEM. Asterisk indicates significant difference from unstimulated value (p < 0.05).Dagger indicates significant difference from corresponding principal value (p < 0.05).

*T

I

*

~ .I ~~ · 2:'_"~~_r:~~--Unstimulated

Bradykinin

A23187

Agonist ~ Control

D

Principal

Histamine

590

PGE2 generated by subepithelial tissue in our experiment were small relative to that produced by the epithelium, but PGE 2 in concentrations as low as 10-9 M is able to modulate cholinergic neurotransmission (23, 25). We have previously reported that there is a major cholinergic component to the airway obstruction in ponies with heaves, inasmuch as atropine administered intravenously or by aerosol significantly reduces pulmonary resistance (6). In vitro experiments utilizing isolated tracheal smooth muscle have shown that compared with controls, muscle from affected animals is hypo responsive to exogenous acetylcholine but hyperresponsiveto electrical field stimulation (7). One possible explanation for these observations is that there is defective regulation of cholinergic neurotransmission in affected animals, such that nerve stimulation results in increased acetylcholine release from the parasympathetic nerves. If modulation of cholinergic innervation of airway smooth muscle by PGE 2 is important in the pathogenesis of heaves in horses, the subepithelial tissue is one logical site for such an interaction to occur. Potential cellular sources for the PGE 2 released by subepithelial tissue include vascular endothelial cells (41) and fibroblasts (42). Indeed, Korn (42) reported that PGE 2 production by fibroblasts could be altered by exposure to mononuclear cell products and that this alteration could persist through multiple cell generations. Electrical field stimulation-induced airway smooth muscle contraction in vitro can cause epithelial-dependent PGE 2 release (43). Therefore, the bronchospasm of affected horses in vivo could have resulted in increased PGE2 production by tissues derived from principal horses in vitro. In fact, we found that stimulated and unstimulated strips from principal horses produced less PGE2 than strips from controls. This finding argues against the importance of in vivo bronchoconstriction by itself affecting PGE2 production by tissues in vitro. Nonetheless, horses with sustained bronchoconstriction induced with agents such as methacholine before being killed may have been another appropriate control. In summary, we have shown that equine tracheal epithelium is not an important source of 15-HETE. In heaves, a disease of horses and ponies characterized by airway obstruction and hyperresponsiveness, there was decreased PGE2 production by subepithelial tissue, and possibly epithelial tissue from affect-

GRAY, DERKSEN, BROADSTONE, ROBINSON, AND PETERS·GOLDEN

ed horses. In addition, PGE 2 production and the time taken to develop clinical signs were correlated, with tissues from affected horses that developed airway obstruction most rapidly tending to produce the least PGE 2 • Although decreased epithelial PGE 2 production has been associated with airway hyperresponsiveness in endotoxin-treated guinea pigs (14), we can find no previous report documenting decreased PGE 2 production by airway mucosa in a spontaneously occurring asthma-like condition. Weconclude from these data that diminished airway mucosal PGE 2 production may play a role in the pathogenesis of equine heaves. ~eferences

1. LowellFC. Observations on heaves. An asthmalike syndrome in the horse. J Allergy 1964; 35:322-30. 2. Derksen FJ, Robinson NE, Armstrong PJ, Stick JA, Slocombe RF. Airway reactivity in ponies with recurrent airway obstruction (heaves).J Appl Physiol 1985; 58:598-604. 3. Derksen FJ, Scott JS, Miller DC, Slocombe RF, Robinson NE. Bronchoalveolar lavage in ponies with recurrent airway obstruction. Am Rev Respir Dis 1985; 132:1066-70. 4. GrayPR, DerksenFJ, Robinson NE, CarpenterDeyo LJ, Johnson HG, Roth RA. The role of cyclooxygenaseproducts in the acute airway obstruction and airway hyperreactivity of ponies with heaves. Am Rev Respir Dis 1989; 140:154-60. 5. Armstrong PJ, Derksen FJ, Slocombe RF, Robinson NE. Airway responses to aerosolized methacholine and citric acid in ponies with recurrent airway obstruction (heaves). Am Rev Respir Dis 1986; 133:357-61. 6. Broadstone RV, Scott JS, Derksen FJ, Robinson NE. Effects of atropine in ponies with recurrent airway obstruction. J Appl Physiol 1988; 65:2720-5. 7. Broadstone RV, LeBlanc PH, Derksen FJ, Robinson NE. In vitro responses of airway smooth muscle from horses with recurrent airway obstruction. Pulmonary Pharmacology 1991; (4):191-202. 8. Broadstone RV, Robinson NE, Gray PR, Derksen FJ. Lack of a mucosal-derivedprostanoid relaxing factor in horses with heaves. FASEB J 1991; 5:AI244. 9. Flavahan NA, Aarhus LL, Rimele TJ, Vanhoutte PM. Respiratory epithelium inhibits bronchial smooth muscle tone. J Appl Physiol 1985; 58:834-8. 10. Morrison KJ, Gao Y,Vanhoutte PM. Epithelial modulation of airway smooth muscle. Am J Physiol 1990; 258:L254-62. 11. Tschirhart E, Frossard N, Bertrand C, Landry Y. Arachidonic acid metabolites and airway epithelium-dependent relaxant factor. J Pharmacol Exp Ther 1987; 243:310-5. 12. Barnett K, Jacoby DB, Nadel JA, Lazarus SC. The effects of epithelial cell supernatant on contractions of isolated canine tracheal smooth muscle. Am Rev Respir Dis 1988; 138:780-3. 13. Butler GB, Adler KB,Evans IN, Morgan DW, Szarek JL. Modulation of rabbit airway smooth muscle responsiveness by respiratory epithelium. Am Rev Respir Dis 1987; 135:1099-104. 14. Folkerts G, Engels F, Nijkamp FP. Endotoxininduced hyperreactivity of the guinea-pig isolated

trachea coincides with decreased prostaglandin E 2 production by the epithelial layer. Br J Pharmacol 1989; 96:388-94. 15. Jacoby DB, Ueki IF, Widdicombe JH, Loegering DA, Gleich GJ, Nadel JA. Human eosinophil major basic protein stimulates prostaglandin E2 production by dog tracheal epithelium. Am Rev Respir Dis 1987; 135:A316. 16. Kasura JW, Terada L, Douglas JG. Regulation of respiratory epithelium prostaglandin production by granulocytes and reduced oxygenmolecules. Clin Res 1987; 35:631A. 17. Widdicombe JH, Ueki IF, Emery D, Margolskee D, Yergey J, Nadel JA. Release of cyclooxygenase products from primary cultures oftracheal epithelia from dog and human. Am J Physiol 1989; 257:L361-5. 18. Cuthbert MF. Bronchodilator activity of aerosols of prostaglandins E 1 and E 2 in asthmatic subjects. Proc Roy Soc Med 1971; 64:15-6. 19. Kawakami Y, Uchiyama K, Irie T, Murao M. Evaluation of aerosols of prostaglandins E 1 and E2 as bronchodilators. Eur J Clin Pharmacol1973; 6:127-32. 20. Smith AP, Cuthbert MF, Dunlop LS. Effects of inhaled prostaglandins E h E2 and F 2u on the airway resistance of healthy and asthmatic man. Clin Sci 1975; 48:421-30. 21. Sweatman WJF, Collier HOJ. Effects of prostaglandins on human bronchial muscle. Nature 1968; 217:69. 22. Gill KK, Kroeger EA. Effect of indomethacin on neural and myogenic components in equine airwaysmooth muscle. J Pharmacol Exp Ther 1990; 252:358-64. 23. Daniel EE, Davis C, Sharma V. Effects of endogenous and exogenous prostaglandin in neurotransmission in canine trachea. Can J Physiol Pharmacol 1987; 65:1433-41. 24. Bethel RA, McClure CL. Cyclooxygenase inhibitors increase canine tracheal muscle response to parasympathetic stimuli in situ. J Appl Physiol 1990; 68:2597-603. 25. Walters EH, O'Byrne PM, Fabbri LM, Graf PD, Holtzman MJ, Nadel JA. Control of neurotransmission by prostaglandins in canine trachealis smooth muscle. J Appl Physiol1984; 57:129-34. 26. Gray PR, Derksen FJ, Robinson NE, PetersGolden ML. Epithelial strips- An alternative method for examining arachidonate metabolism in equine tracheal epithelium. Am J Respir Cell Mol BioI 1992; 6:29-36. 27. Gray PR, Derksen FJ, Broadstone RV, Robinson NE, Johnson HG, Olson NC. Increased pulmonary production of immunoreactive 15hydroxyeicostatetraenoic acid in an animal model of asthma. Am Rev Respir Dis 1992; 145:1092-7. 28. Hamberg M, Hedqvist P, Radegran K. Identification of 15-hydroxy-5,8,1l,13-eicosatetraenoic acid (I5-HETE) as a major metabolite of arachidonic acid in human lung. Acta Physiol Scand 1980; 110:219-21. 29. Marom Z, Shelhamer JH, Sun F, Kaliner M. Human airway monohydroxyeicosatetraenoic acid generation and mucus release. J Clin Invest 1983; 72:122-7. 30. Johnson HG, McNee ML, Sun FF. 15hydroxyeicostatetraenoic acid is a potent inflammatory mediator and agonist of canine tracheal mucus secretion. Am Rev Respir Dis 1985; 131:917-22. 31. Vanderhoek JY, Thre NS, Bailey JM, Goldstein AL, Pluznik DH. New role for 15hydroxyeicosatetraenoic acid. Activator of leukotriene biosynthesis in PT-18 mast/basophile cells. J BioI Chern 1982; 257:12191-5. 32. Goetzl EJ. Selective feed-back inhibition of

591

TRACHEAL MUCOSAL PGE. PRODUCTION IN HEAVES

the 5-lipoxygenation of arachidonic acid in human 'l-lymphocytes. Biochem Biophys Res Commun 1981; 101:344-50. 33. Hunter JA, Finkbeiner WE, Nadel JA, Goetz! EJ, Holtzman MJ. Predominant generation of 15lipoxygenase metabolites of arachidonic acid by epithelial cells from human trachea. Proc Natl Acad Sci USA 1985; 82:4633-7. 34. Lipchik RJ, Chanucey JB, Paine R, Simon RH, Peters-Golden M. Arachidonate metabolism increases as rat alveolar type II cells differentiate in vitro. Am J Physiol 1990; 259 (3):L73-80. 35. Holtzman MJ, Pentland A, Baensiger NL, Hanborough JR. Heterogenicity of cellular expression of arachidonate 15-lipoxygenase:implications for biological activity. Biochim Biophys Acta 1989;

1003:204-8. 36. Thurlbeck WM, LowellFC. Heaves in horses. Am Rev Respir Dis 1964; 89:82-8. 37. Brigham KL, Serafin W, zaaerr A, Blair I, Meyrick B, Oates JA. Prostaglandin E 2 attenuation of sheep lung responses to endotoxin. J Appl Physiol 1988; 64:2568-74. 38. Kunkel SL, Chensue SW, Phan SH. Prostaglandins as endogenous mediators of interleukin 1production. J Immunol1986; 136:186-92. 39. Kunkel SL, Wiggins RC, Chensue SW, Larrick J. Regulation of macrophage tumor necrosis factor production by prostaglandin E 2 • Biochem Biophys Res Commun 1986; 137:404-10. 40. Pitt BR, Moalli R, Man SFP, Gillis CN. Alveolar transfer of prostaglandin E 2 in isolated per-

fused guinea pig lungs. J Appl Physiol 1985; 59:691-7. 41. Gerritsen ME, Chell CD. Arachidonic acid and prostaglandin endoperoxide metabolism in isolated rabbit and coronary microvessels and isolated and cultivated coronary microvessel endothelial cells. J Clin Invest 1983; 72:1658-71. 42. Korn JH. Fibroblast prostaglandin E2 synthesis. J Clin Invest 1983; 71:1240-6. 43. Ullman A, Ciabattoni G, Lofdahl G, Linden A, SvedmyrN, Skoogh BE. Epithelial-derivedPGE 2 inhibits the contractile response to cholinergic stimulation in isolated ferret trachea. Pulmonary Pharmacology 1990; 3:155-160.