Abstract
Isoprostanes are generated nonenzymatically during free radical-mediated lipid peroxidation, and are used clinically and experimentally as markers of oxidative stress. However, their biological effects are poorly understood. We examined the effects of seven different 8-isoprostanes in human and canine airway smooth muscles. In large order airways (carina) of the human, several isoprostanes evoked powerful contractions, with 8-iso-prostaglandin (PG) E2, 8-iso-PGF1α, and 8-iso-PGF2α being the most efficacious (and with logEC50 values of 7.0, 5.9, and 6.2 μM, respectively). These contractions were sensitive to the prostanoid TP receptor antagonist ICI 192,605 (0.1–1 μM), but not the EP prostanoid receptor antagonist AH-6809 (50 μM), or the leukotriene receptor antagonists monteleukast or ICI 198,615 (both 1 μM). Qualitatively similar results were obtained in small order human airways (<2 mm o.d.), except that the isoprostanes were generally slightly less potent. None of the isoprostanes had any marked excitatory effect in canine airways. In carbachol-preconstricted tissues (pretreated with ICI 192,605 to block any potential contraction), several isoprostanes completely relaxed canine airways: 8-iso-PGE1, 8-iso-PGE2, and 8-iso-PGF3α were the most potent, with logIC50 values of 6.9, 6.9, and 5.7, respectively. Only 8-iso-PGF3α relaxed human airways (logIC50 = 4.9). Our results show that several 8-isoprostanes are highly biologically active in human and canine airways, evoking both excitatory and/or inhibitory effects, and that these effects are compound, species, and tissue dependent.
The airways are continually exposed to a variety of free radicals and reactive oxygen species in inspired air (e.g., ozone) and liberated by inflammatory cells (e.g., peroxide, superoxide, hydroxyl radical). These agents can alter airway function, ranging from contraction in human airways (Rabe et al., 1995) to relaxation in canine airways (Gao and Vanhoutte, 1992; Janssen et al., 2000b), but the underlying mechanisms are as yet unclear.
It is now recognized that nonenzymatic peroxidation of arachidonic acid by free radicals and reactive oxygen species can give rise to isoprostanes (Morrow et al., 1990; Practico et al., 1995). Isoprostanes differ structurally from prostaglandins (PGs) by thecis-orientation at the cyclopentane ring junction compared with the trans-orientation in the classical prostanoids (Fig. 1).
8-Isoprostanes are present in substantial amounts even in normal plasma or urine (in which their levels can be several orders of magnitude higher than those of cyclooxygenase-derived PGs; Morrow et al., 1990), but they are further elevated in many states in which oxidative stress is a prominent feature. For example, they are elevated in smokers (Morrow et al., 1995; Pratico et al., 1995; Delanty et al., 1996;Reilly et al., 1996; Chiabrando et al., 1998; Pratico et al., 1998a); in patients with asthma (Montuschi et al., 1999a), chronic obstructive pulmonary disease (Pratico et al., 1998b), interstitial lung disease (Montuschi et al., 1998), cystic fibrosis (Montuschi et al., 1999b), or acute chest syndrome (Klings et al., 1999); during exposure to allergen (Dworski et al., 1999), ozone (Hazbun et al.,1993), or hyperoxia (Vacchiano and Tempel, 1994); and during ventilated ischemia (Becker et al., 1998).
Despite their prevalence in these disease states, the biological effects of 8-isoprostanes in airways are very poorly understood. 8-iso-PGF2α is a potent stimulant of vascular (Takahashi et al., 1992; Kang et al., 1993; Kromer and Tippins, 1996; Zhang et al., 1996; John and Valentin, 1997; Oliveira et al., 2000), intestinal (Elmhurst et al., 1997), and uterine (Crankshaw, 1995) smooth muscles where its effects are sensitive to selective prostanoid TP receptor antagonists. Consequently, the effects of 8-iso-PGF2α are generally, although not exclusively, thought to be mediated by action at TP receptors (Kromer and Tippins, 1999). Structure-activity studies with 8-isoprostanes in human umbilical artery (HUA) have revealed that E-ring compounds are more potent than F-ring compounds, doubly unsaturated compounds are more potent than singly unsaturated compounds, and the α-configuration is more potent than the β-configuration (Oliveira et al., 2000) (Fig. 1).
To date, there have been only a few studies on the excitatory effects of 8-iso-PGF2α in the airways (Kang et al., 1993; Kawikova et al., 1996; Okazawa et al., 1997). The effects of other 8-isoprostanes have not been studied; therefore, structure-activity relationships are unknown for these tissues. Consequently, we set out to investigate the effects of a series of isoprostanes in human and canine airways and discovered some predominantly relaxant effects of these compounds that have not been reported previously. These data have been presented in abstract form (Janssen et al., 2000a).
Materials and Methods
Tissue Collection and Preparation.
Segments of donor (i.e., nondiseased) human main-stem bronchi were obtained from the Lung Transplant Program, Toronto, Ontario, Canada (n = 11). The overlying connective tissue, vasculature, and thicker portions of the epithelium were removed and the smooth muscle was then cut into strips (≈1 mm wide) parallel to the muscle fibers. This preparation is referred to as human large airways.
Portions of human lungs that had been resected at St. Joseph's Hospital, Hamilton, Ontario, Canada, and which had been judged by the pathologist to be macroscopically normal were also obtained (n = 20). From these, small order airways (0.5 to 2 mm o.d.) were carefully removed and cut into ring segments 4 to 5 mm long. This preparation is referred to as human small airways.
Adult mongrel dogs were euthanized with pentobarbital sodium (100 mg · kg−1) and the tracheae and lungs were excised (n = 15). After removal of the overlying connective tissue, vasculature and epithelium, the trachealis was cut into strips (≈1 mm wide). Lobes of lung were pinned out, the overlying parenchyma and pulmonary vasculature was removed, and ring segments (≈4–5 mm long) of 5th to 6th order bronchi (2–6 mm o.d.) were excised.
Isometric Contractions.
Strips and ring segments of smooth muscle were mounted vertically in 3-ml organ baths using silk (Ethicon 4-0) tied to either end of the strip; one end was fastened to a Grass FT03 force transducer while the other was anchored. Isometric tension was digitized and recorded using an on-line program (DigiMed system integrator; MicroMed, Louisville, KY). Tissues were bathed in physiological salt solution (PSS) containing indomethacin (10 μM), bubbled with 95% O2, 5% CO2, and maintained at 37°C. Preload tension was 1.25 g (determined previously to allow maximal responses). Tissues were first equilibrated for 1 to 2 h, during which canine tissues were also challenged with 60 mM KCl (for standardization of the data); human tissues were not challenged with KCl during this period because the resultant contractions were not always easily reversible. After the equilibration period, airway tissues were exposed to increasing concentrations of individual isoprostanes (10-fold increments) and the peak contractile response to each addition was recorded.
To examine the inhibitory effects of the isoprostanes, airway tissues were pretreated for 20 to 30 min with the prostanoid TP receptor antagonist ICI 192,605 (10−6 M) to block any potential excitatory effects of the isoprostanes (see below), and then precontracted with carbachol (10−6 M for canine tissues, and 10−5 M for human tissues). Once cholinergic tone had stabilized, the tissues were exposed to increasing concentrations of individual isoprostanes or to PGE2 (10-fold increments) and the relaxant response to each was recorded.
We examined the sensitivity of isoprostane-evoked contractions (8-iso-PGE2 was used because it was the most potent and efficacious) in human airways to various receptor antagonists in two ways. In one set of experiments, after precontracting the tissues with 10−5 M 8-iso-PGE2 and contractile tone had stabilized, ICI 192,605 (10−8, 10−7, 10−6 M) was added, which caused a progressive reversal of tone. In other experiments, tissues were pretreated with the leukotriene receptor antagonists MK-476 (monteleukast) or ICI 198,615 (both 1 μM), or the EP receptor antagonist AH-6809 (6-isopropoxy-9-oxoxanthene-2-carboxylic acid, 50 μM) for 20 min, after which the 8-iso-PGE2dose-response relationship was reexamined.
Drugs and Chemicals.
PSS was composed as follows: 116 mM NaCl, 4.2 mM KCl, 2.5 mM CaCl2, 1.6 mM NaH2PO4, 1.2 mM MgSO4, 22 mM NaHCO3, 11 mMd-glucose, 0.01 mM indomethacin, bubbled to maintain pH at 7.4.
The isoprostanes, PGE2, and AH-6809 were obtained from Cayman Chemicals (Ann Arbor, MI). ICI 192,605 {4(Z)-6-[(2,4,5cis)2-(2-chlorophenyl)-4-(2-hydroxy phenyl)1,3-dioxan-5-yl]hexenoic acid} and ICI 198,615 {[1-([2-methoxy-4{[(phenylsulfonyl) amino] carbonyl} phenyl] methyl)-1H-indazol-6-yl] carbamic acid cyclopentyl ester} were gifts from Zeneca (Alderley Park, UK). MK-476 was a gift from Merck Frosst Canada (Dorval, Quebec). All other chemicals were obtained from Sigma (Oakville, Ontario, Canada). All drugs were made as stock solutions in ethanol except for MK-476 (aqueous), 8-iso-PGF2α (ethyl acetate), ICI 192,605 (dimethylsulphoxide), and ICI 198,615 (dimethylsulphoxide). Immediately before experiments, appropriate serial dilutions of drugs were made into PSS.
Statistics.
Magnitudes of the contractions in the canine airways were expressed relative to the KCl-induced contraction; those in the human airways are given as absolute values. Magnitudes of isoprostane-evoked relaxations in human and canine airways were corrected for vehicle-related effects, as described previously (Oliveira et al., 2000), and expressed relative to the carbachol-induced contraction (100%). Concentration-effect curves were constructed and EC50 or IC50 values derived, as described previously (Oliveira et al., 2000). ICI 192,605-triggered relaxations were expressed as a percentage reversal of isoprostane-induced tone. Data are reported as mean ± S.E., and compared using an unpaired two-tailed Student's t test, with P values <.05 being considered significant.
Results
Effects of Isoprostanes on Human Airways.
All the isoprostanes tested produced concentration-dependent contractions of smooth muscle from human large and small airways with the exception of 8-iso-PGF3α, which produced relaxation in large airways; representative traces are shown in Fig.2. Mean concentration-effect relationships for the isoprostanes in the human airway smooth muscle preparations are given in Fig. 3, and EC50 values for these are found in Table1. Of interest, 8-iso-PGE2 was considerably more potent (almost a full log unit more so) and efficacious than 8-iso-PGF2α, the only isoprostane that has been studied to date in the airways.
We sought to ascertain whether the isoprostanes were acting through TP or EP prostanoid receptors (as shown elsewhere; Kawikova et al., 1996;Elmhurst et al., 1997) or through leukotriene receptors, which are primarily responsible for basal tone in human airways (Ellis and Undem, 1994; Watson et al., 1997). Tissues were pretreated with the selective leukotriene receptor antagonists MK-476 or ICI 198,615 (both 1 μM), the EP receptor antagonist AH-6809 (50 μM), or the TP receptor antagonist ICI 192,605 (1 μM; done only in the large airways). Neither the leukotriene nor the EP receptor antagonists had any significant effect on the dose-response relationship for 8-iso-PGE2 (Fig.4A), whereas the selective TP receptor antagonist caused a marked displacement of this relationship. We examined more closely the dose dependence of this inhibitory effect of ICI 192,605 in tissues preconstricted with 8-iso-PGE2 (10 μM), finding that this tone was essentially unaltered by 10−8 M ICI 192,605 and reversed ≈50% by 10−7 M ICI 192,605, even though the reported pKB value for this antagonist against TP receptors in guinea pig airways is ≈9.5 (Kawikova et al., 1996).
All 8-isoprostanes, with the exception of 8-iso-PGF3α, were without quantifiable inhibitory effect on carbachol-induced tone (in the presence of ICI 192,605) in human large airways, except when used at 10−4 M; later experiments using canine airway tissues (below) indicate relaxations at these extreme concentrations may be due to vehicle alone (0.1% ethanol). 8-iso-PGF3α, however, produced concentration-dependent relaxations with a pIC50 of 4.9 ± 0.3 (n = 8). An illustrative original tracing of the effect of 8-iso-PGF3α on carbachol-induced tone is shown in Fig. 5A; mean data for all compounds tested are shown in Fig. 5B.
Effects of Isoprostanes on Canine Airways.
8-Isoprostanes had no appreciable excitatory effects on canine airways at rest, even though these same tissues produced marked responses to 60 mM KCl (Fig. 6). In contrast, the compounds produced concentration-dependent relaxations of carbachol-induced tone in canine trachea and intraparenchymal bronchi. In general, the responses to the E-ring isoprostanes were comparable with those of the prostanoid PGE2, whereas those of the F-ring isoprostanes were comparable with vehicle alone (ethanol), with the exception of 8-iso-PGF3α, which was intermediate between the two extremes. Concentration-effect curves are shown in Fig.7 and logIC50values are given in Table 1.
Discussion
Previous studies of the effects of 8-isoprostanes on airway smooth muscle have been limited to the study of the excitatory effects of only one isoprostane, 8-iso-PGF2α, under the assumption that this isoprostane is the predominant form generated during free radical attack of cell membranes. In human large airways, the effects of 8-iso-PGF2α were shown to be mediated by TP receptors, although some heterogeneity in the receptor population involved could not be excluded (Kawikova et al., 1996).
In this study, we found several other isoprostanes can also cause contraction of airway smooth muscle. The 8-isoprostanes we tested generally had a higher contractile potency in human large airways compared with human small airways (Table 1). The two exceptions are 8-iso-PGE1, which had highly variable potency in both preparations, and 8-iso-PGF3α, which had powerful relaxant effects in human large airways that most likely masked any excitatory effects. The order of potency of the compounds in human small airways is identical with that found in HUA, where their actions are mediated via TP receptors (Oliveira et al., 2000). TP receptors are present in abundance in human airway smooth muscle (Armour et al., 1989) and our finding that ICI 192,605 rapidly and completely reversed 8-isoprostane-induced contractions suggests that TP receptors make a major contribution to the compound's actions in these tissues. The concentrations of ICI 192,605 required to achieve this inhibition, however, were orders of magnitude greater than those found to be effective against TP receptors in guinea pig airways (pKB ≈ 9.5; Kawikova et al., 1996). Although 8-isoprostanes have been shown to contract smooth muscle via excitatory EP receptors (Elmhurst et al., 1997), we found that these contractions were not sensitive to an antagonist of (excitatory) EP1 receptors, suggesting these are not involved. Similarly, although leukotrienes are primarily responsible for basal tone in human airways (Ellis and Undem, 1994; Watson et al., 1997), we found that the contractions were insensitive to two structurally different leukotriene receptor antagonists (Fig. 4).
The predominantly relaxant effects of 8-iso-PGF3α on human large airways represent a novel finding for this class of compounds in any smooth muscle: we found that this compound abolished both carbachol-induced tone and “basal” tone in these preparations. The failure of other 8-isoprostanes to relax human large airways suggests a high degree of structural specificity of the receptor involved. Human airway smooth muscle expresses inhibitory EP2 and IP receptors (Norel et al., 1999) but we do not know whether either of these receptors mediate the inhibitory effect of 8-iso-PGF3α.
TP receptors are sparse in canine bronchus and virtually absent from canine trachea (Coleman et al., 1994), and the potent inhibitory actions of PGE2 (Table 1) also suggest that excitatory EP receptors are operationally insignificant in canine airways. This may explain why the 8-isoprostanes were devoid of excitatory effects in this species. The rank order of potency for relaxation was identical between canine trachea and bronchus except for the reversal of the low-potency compounds 8-iso PGF2α and 8-iso-PGF2β. Structural requirements for this TP receptor antagonist-insensitive mechanism showed both similarities and differences to the TP receptor-mediated contraction of HUA. In both airway smooth muscle and HUA, E-ring compounds were more potent than F-ring compounds, and among the E-ring compounds, doubly unsaturated compounds were more potent than singly unsaturated ones (Oliveira et al., 2000). However, whereas 8-iso-PGF2α was the most potent F-ring compound and 8-iso-PGF3α was devoid of activity in HUA (Oliveira et al., 2000), in canine airways 8-iso-PGF3α was 4 to 20 times more potent than 8-iso-PGF2α, which was unable to fully relax the trachea (Fig. 7). The relaxation mechanism in canine airways also appears to be different from that in human airways where the E-ring compounds were devoid of relaxant effects when TP receptors were blocked. Canine airway smooth muscle expresses inhibitory EP receptors (Coleman et al., 1994), although we are not aware that these have been further characterized. The inhibitory actions of the 8-isoprostanes may be mediated through EP receptors. A more complete and diverse repertoire of pharmacological tools is necessary to clarify these questions.
The species differences in airway responses to 8-isoprostanes observed in the present study parallel the different responses evoked by free radicals and reactive oxygen species in these tissues. For example, hydrogen peroxide evokes contractions in human airway smooth muscle (Rabe et al., 1995) but relaxations in canine airway smooth muscle (Gao and Vanhoutte, 1992; Janssen et al., 2000b). It is our hypothesis that such exposure to peroxide would generate a variety of isoprostanes in various proportions. It might otherwise be hard to predict the response to such a complex mixture of autacoids. However, it is worth pointing out that in the dog, most isoprostanes evoke large relaxations but none exhibit any appreciable excitatory activity (Figs. 6 and 7), consistent with the overall relaxant effect of hydrogen peroxide in this tissue (Gao and Vanhoutte, 1992). In the human, on the other hand, several isoprostanes are powerful constricting agents and only one is moderately inhibitory (8-iso-PGF3α, and only at very high concentrations; Figs. 2-5), which would account for the observed bronchoconstrictor response to peroxide (Rabe et al., 1995). Thus, isoprostanes may be the mediators of the effects of free radicals. It is as yet unknown whether the various reactive oxygen species (peroxide, superoxide, hydroxyl radical, ozone) produce similar or different proportions of the various isoprostane isomers. Thus, it will be important to ascertain which isoprostanes are produced during oxidative stress in the lungs. It has already been shown that the overall levels of isoprostanes are increased substantially after inhalation of cigarette smoke (Morrow et al., 1995; Pratico et al., 1995; Delanty et al., 1996; Reilly et al., 1996; Chiabrando et al., 1998; Pratico et al., 1998a) or other noxious stimuli (Hazbun et al., 1993; Vacchiano and Tempel, 1994; Becker et al., 1998; Dworski et al., 1999), and in patients with airway-related disease (Montuschi et al., 1998, 1999a,b; Pratico et al., 1998b; Klings et al., 1999).
Several important issues remain unanswered regarding these potentially clinically relevant compounds. Their effects on other smooth muscle tissues, particularly those that are also regularly exposed to free radicals and reactive oxygen species, such as the pulmonary vasculature, need to be investigated. The mechanisms underlying these responses, the receptors and second messengers, must be identified. More importantly, although pharmacological tools that can be used to block their excitatory effects or mimic their inhibitory effects need to be identified or developed, these might prove highly useful in the treatment of diseases in which oxidative stress is a prominent feature.
In conclusion, we provide evidence for both excitatory and inhibitory actions of 8-isoprostanes in airway smooth muscles. Furthermore, we found substantial isoprostane-, species-, and tissue-related differences in these actions. Inhibitory effects were not sensitive to TP receptor antagonists and their structural requirements were different from those of excitatory effects. The receptor(s) mediating the inhibitory effects of 8-isoprostanes remain uncharacterized. These findings have important implications with respect to bronchoconstriction in the context of asthma and many other breathing-related disorders.
Footnotes
-
Send reprint requests to: Dr. L. J. Janssen, Department of Medicine, McMaster University, 50 Charlton Ave. East, Hamilton, Ontario, Canada, L8N 4A6. E-mail:janssenl{at}fhs.csu.mcmaster.ca
-
↵1 This study was supported by an operating grant and a Scientist Award from the Medical Research Council of Canada (to L.J.J.).
- Abbreviations:
- PG
- prostaglandin
- HUA
- human umbilical artery
- PSS
- physiological salt solution
- Received April 4, 2000.
- Accepted July 11, 2000.
- The American Society for Pharmacology and Experimental Therapeutics