The endothelins are three structurally related peptides named endothelin-1, endothelin-2, and endothelin-3. These peptides produce a variety of biologic responses mediated by two distinct receptors (ETA and ETB), which have been sequenced and cloned (1). In the vasculature, ETA receptors mediate constriction via membrane receptors on the smooth-muscle cells, whereas ETB receptors can mediate both vasodilation by release of relaxing factors from vascular endothelial cells and vasoconstriction through receptors located on smooth-muscle cells (2,3).
The ETB receptors on vascular endothelial cells and vascular smooth muscle have been purported to represent two different subtypes of the ETB receptor, ETB1 and ETB2, respectively. This classification was first proposed based on differences in the sensitivity of rat vascular endothelium and rabbit vascular smooth muscle to the endothelin antagonist PD 142893 (4). PD 142893 was potent at blocking ETB-mediated vasodilation in the isolated perfused rat mesentery, but only weakly blocked ETB-mediated constriction in rabbit pulmonary artery rings. In addition, Clozel et al. (5) subsequently reported that another antagonist, bosentan, was selective for inhibiting endothelial (rabbit mesentery artery rings) ETB responses over smooth muscle (rat tracheal rings) ETB responses. In contrast to the functional results, Clozel et al. (5) also reported that bosentan had a higher binding affinity for ETB2-type receptors (porcine tracheal smooth muscle) than for the ETB1 receptors (human placenta). This discrepancy between the binding and functional results with bosentan was ascribed to differences in test systems and experimental conditions.
Because tissues from different species were used to compare the endothelial and smooth-muscle activity in the previously published functional studies, it is also possible that species, as well as experimental differences, rather than receptor subtypes, are responsible for the observed selectivity of bosentan and PD 142893 for endothelial versus smooth-muscle ETB receptors. Therefore the purpose of this study was to determine whether vascular endothelial and vascular smooth-muscle ETB receptors could be quantitatively differentiated by PD 142893 in the same species (rat) by using closely matched experimental conditions. In addition, two structurally unrelated antagonists, SB 209670 (6) and the selective ETB-receptor antagonist BQ 788 (7), were also studied.
MATERIALS AND METHODS
L-Phenylephrine, acetylcholine, and 3-[(3-cholamidopropyl)dimethylammoniol)-1-propanesulfonate (CHAPS), were purchased from Sigma Chemical Co. (St. Louis, MO, U.S.A.). Sarafotoxin S6c was purchased from Peptides International (Louisville, KY, U.S.A.) and is a highly selective ETB receptor agonist (8). S6c was dissolved in 0.1% acetic acid in distilled water and then diluted with buffered 0.9% saline. PD 142893 (PD), SB 209670 (SB), and BQ 788 (BQ) were synthesized by the Department of Chemistry (Warner Lambert, Parke-Davis Pharmaceutical Research Division, Ann Arbor, MI, U.S.A.); purity >95% was verified by HPLC. PD, SB, and BQ were dissolved in buffered 0.9% saline and then added to the perfusion media.
Isolated perfused rat kidney preparation
The isolated perfused rat kidney was used to study vascular smooth-muscle cell ETB-receptor activity. Methods used were as previously published (9). In brief, male Sprague-Dawley rats (350-400 g) were anesthetized with pentobarbital (60 mg/kg, i.p.) and heparinized (100 U, i.v.). The superior mesenteric artery was cannulated and the cannula passed into the right renal artery. The kidney was immediately perfused with a constant flow (18 ± 2 ml/min) of warmed (37°C) and gassed (95% O2, 5% CO2) Krebs bicarbonate buffer at an initial perfusion pressure of 90 mm Hg. The rat was killed by exsanguination, and the kidney excised and suspended by the renal artery in a climate-controlled chamber. Changes in perfusion pressure were measured with an in-line pressure transducer. After a 15-min equilibration period, the kidney was perfused for an additional 15 min with either saline or antagonist in the buffer before being challenged with increasing bolus doses of sarafotoxin S6c (1-10,000 pmol/0.05 ml bolus) given at 5-min intervals. Each kidney was used for only one study and was treated with either vehicle or one dose of the antagonist. Time control studies verified that with these methods, a kidney preparation was hemodynamically stable for ≥2 h, and that there was no significant change in the S6c dose-response curve over this period.
Because some of the endothelin antagonists are noted for their high degree of protein binding (10), we deendothelialized the isolated perfused rat kidney to determine whether the endothelium limited accessibility of the antagonist to the underlying vascular smooth muscle. In a separate set of experiments, the kidneys were perfused with the detergent CHAPS (5 mg/ml Krebs buffer, for 30 s) according to previously published methods (11). The kidneys were then constricted with an 80% effective concentration of phenylephrine (0.5 μM) and tested for an endothelium-dependent vasodilator response to acetylcholine (300 ng, bolus). CHAPS perfused kidneys exhibiting a normal vasoconstriction response to phenylephrine (40 ± 9 mm Hg), and a loss of endothelium-dependent vasodilation compared with previously established nondenuded controls (−38 ± 4 mm Hg) were accepted as endothelium-denuded preparations.
Isolated perfused rat mesentery preparation
The isolated perfused rat mesentery was used to study vascular endothelial cell ETB-receptor activity. Methods used were as previously published (4,12). In brief, male Sprague-Dawley rats (350-400 g) were anesthetized with pentobarbital (60 mg/kg, i.p.) and heparinized (100 U, i.v.). The abdomen was opened and the pyloric, colonic, and ileocecal branches of the superior mesenteric artery were ligated. The superior mesenteric artery was cannulated and immediately perfused with a constant flow (10 ml/min) of warmed (37°C) and gassed (95% O2, 5% CO2) Krebs bicarbonate buffer at an initial perfusion pressure of 35 mm Hg. The rat was killed by exsanguination, and the mesenteric vasculature with adhering small intestine was excised and suspended by the mesenteric artery in a climate-controlled chamber. Changes in perfusion pressure were measured with an in-line pressure transducer. After a 15-min equilibration period, the mesentery was perfused for an additional 15 min with either saline or antagonist in the buffer. The mesentery was then constricted with 6 μM phenylephrine, increasing perfusion pressure to 128 ± 6 mm Hg. Functional vascular endothelium was confirmed by a depressor response equal to previously established controls in response to 100-pmol bolus of acetylcholine (−63 ± 8 mm Hg change). After perfusion pressure returned to preacetylcholine levels, the mesentery was challenged with increasing bolus doses of S6c (1-10,000 pmol/0.05 ml bolus) given at 5-min intervals. Each mesentery was used for only one study, and was treated with either vehicle or one dose of the antagonist. Time control studies were done and verified that by using these methods, a mesentery preparation was hemodynamically stable for ≥2 h, and that there was no significant change in the acetylcholine, phenylephrine, or S6c dose-response curve over this period.
Results are expressed in mm Hg as mean change in pressure from baseline ± SEM from four to six preparations per treatment group. A pA2 value, an index of antagonist potency, was determined by analysis of the Schild regression for each antagonist in vascular endothelial and vascular smooth-muscle preparations. The -log of the antagonist concentration was plotted against the corresponding log of the shift in the S6c median effective dose (ED50) minus 1. If the resulting line has a slope of unity, indicating competitive antagonism, then the X-intercept of the line equals the pA2 value (13). An analysis of covariance was used to determine whether the slopes were different from unity (14). The pA2 values also were compared by using the analysis of covariance to determine whether the antagonist potency was the same for vascular smooth muscle and vascular endothelial cell ETB receptors.
Isolated perfused rat kidney and mesentery models
Administration of S6c in the isolated perfused rat kidney produced only dose-dependent vasoconstriction (Fig. 1). Vasodilation was not demonstrated in this model in response to any of the endothelin isoforms or S6c (9). The lack of a detectable ETB receptor-mediated vasodilation makes this a suitable model for studying vascular smooth-muscle cell ETB-receptor activity. S6c increased perfusion pressure, starting at 10 pmol. A maximal increase of 166 ± 23 mm Hg from baseline occurred at 300 pmol. In the isolated perfused rat mesentery, S6c produced only dose-dependent vasodilation. Vasoconstriction was not demonstrated in this model in response to ETB-receptor activation by S6c (4). The lack of a detectable ETB receptor-mediated vasoconstriction makes this a suitable model for studying vascular endothelial cell ETB-receptor activity. S6c decreased perfusion pressure starting at 10 pmol. A maximal decrease of 40 ± 5 mm Hg from baseline occurred at 100 pmol (Fig. 1). Doses >100 pmol produced progressively smaller responses, which could reflect changes in endothelial signaling, unmasking of a S6c-mediated vasoconstriction that competes with the vasodilation, or receptor desensitization (15,16). In the absence of antagonist, the ED50 values for S6c dose-response curves in the isolated perfused rat kidney and isolated perfused rat mesentery ETB-receptor assays were similar, 24 ± 4 and 22 ± 3 pmol, respectively.
All three antagonists, PD, SB, and BQ, produced parallel concentration-dependent rightward shifts in the S6c dose-response curves in the isolated perfused rat kidney (Figs. 2-4). PD and SB also produced parallel concentration-dependent rightward shifts in the S6c dose-response curves in the isolated perfused rat mesentery (Figs. 5 and 6). In contrast to PD and SB, BQ in the isolated perfused mesentery produced only a slight shift in the S6c dose-response curve at 0.001 μM, and at the higher concentrations (0.003 and 0.01 μM) produced an insurmountable antagonism (Fig. 7).
The slopes of the Schild regressions (Fig. 8) were not significantly different from one (unity), indicating competitive antagonism for PD, SB, and BQ in the isolated perfused kidney (Table 1). PD and SB also produced slopes not significantly different from unity in the isolated perfused mesentery, whereas BQ appeared to be an insurmountable antagonist in the isolated perfused mesentery.
Analysis of covariance indicated that the Schild-derived pA2 values for PD and SB were significantly greater for inhibiting S6c-mediated vasodilation (endothelial cell) in the rat mesentery compared with S6c-mediated vasoconstriction (smooth muscle) in the rat kidney. Furthermore, the antagonists differed in their relative potency between the two assays. The ratio of antagonist potency between the vascular endothelial and vascular smooth-muscle ETB receptors was 13-fold for PD and 50-fold for SB (Table 1). Because maximal vasodilation could be achieved at only one effective concentration of BQ in the mesentery, it was not possible quantitatively to compare the antagonist activity of BQ in the two assays.
Effect of prolonged exposure and denudation on antagonist activity
Two additional sets of experiments were done in the isolated perfused rat kidney to explore the possibility that the differences in antagonist potency between the vascular endothelial and vascular smooth-muscle ETB receptors were due to the endothelium inhibiting accessibility of the antagonist to the vascular smooth muscle. In the first set of experiments, SB at 0.03 μM was perfused for an additional 15 min (30-min total equilibration period) before being challenged with S6c (Fig. 9). SB was just as potent against S6c-induced contraction in the isolated perfused rat kidney with a 30-min equilibration period (pA2 = 8.1) as it was after the 15-min equilibration period (pA2 = 8.0).
In the second set of experiments, the effects of denuding the endothelium on antagonist activity were determined. In the absence of antagonist, denuding the kidney increased S6c potency about twofold, and the ED50 value for the S6c dose-response curve shifted from 24 ± 4 pmol in the endothelium intact kidney to 11 ± 1 pmol in the denuded kidney. This increase in S6c potency may be attributed to the endothelium no longer limiting accessibility of the agonist to the underlying vascular smooth muscle in the denuded kidney, or removal of a S6c-induced endothelium-derived vasodilator component that may have been competing with the S6c vasoconstrictor effect. Maximal S6c-dependent vasoconstriction was reduced from a 166 ± 23 mm Hg increase in perfusion pressure from baseline in the endothelium-intact kidney to 75 ± 7 mm Hg in the denuded kidney. Because the vasoconstriction response to phenylephrine in the denuded kidney was not affected, it is unlikely that this reduction in the maximal S6c vasoconstriction response was due to damaging the vascular smooth muscle with CHAPS. In the isolated perfused rat kidney, S6c induces the release of prostanoids via ETB receptors, independent of its direct vasoconstrictor effect (9). The source of these prostanoids within the rat kidney has not been determined. However, this release of prostaglandins has been thought to contribute to the vasoconstrictor effect of S6c in the rat kidney. While removing the endothelium with CHAPS, we could have also removed cells that were a source of the prostaglandins, which could explain the observed reduction in the S6c pressor response.
All three antagonists, PD, SB, and BQ, produced parallel rightward shifts in the S6c dose-response curves in the isolated denuded rat kidney (Fig. 10). Removing the endothelium increased the potency of all three antagonists for the vascular smooth-muscle ETB receptor (Table 2). The antagonist potency of PD and SB were increased approximately threefold and fourfold, respectively. The antagonist potency of BQ in the denuded kidney for the vascular smooth-muscle ETB receptor was increased fivefold. Even though antagonist potency increased with the endothelium removed in the vascular smooth-muscle ETB-receptor assay, PD and SB were still significantly more potent for inhibiting vascular endothelial ETB receptors compared with vascular smooth-muscle ETB receptors. The ratio of antagonist potency between the vascular endothelial and vascular smooth-muscle (endothelial denuded) ETB receptors was fourfold for PD and 13-fold for SB.
The purpose of this study was to determine whether vascular endothelial cell and vascular smooth-muscle cell ETB receptors could be quantitatively differentiated by ETB-receptor antagonists in the same species by using closely matched experimental conditions. The results indicate that all three antagonists, PD, SB, and BQ, were more potent at inhibiting vascular endothelial cell ETB-receptor activity than vascular smooth-muscle cell ETB receptor activity. More important, the antagonists differed in their relative potency between the vascular endothelial and vascular smooth-muscle ETB receptors. BQ appeared to be a noncompetitive antagonist in rat vascular endothelium yet competitive in rat vascular smooth muscle.
Schild analysis indicated competitive antagonism of S6c by PD and SB in vascular endothelial cells and vascular smooth muscle. The slopes of the regression lines were not significantly different from unity. These results are based on the assumption that equilibrium conditions prevailed. It is, nevertheless, possible to have competitive antagonism in nonequilibrium states. One relevant situation would be insufficient equilibration time for the antagonist in a tissue in which diffusion is rate limiting. This situation is possible in the vasculature, where endothelial receptors would be less limited by diffusion than would smooth-muscle receptors. However, because SB was just as potent against S6c-induced contraction in the isolated perfused rat kidney with a 30-min equilibration period (pA2 = 8.1), as it was after the 15-min equilibration period (pA2 = 8.0), it is unlikely that insufficient equilibration time could account for the observed differences. In the denuded isolated perfused rat kidney preparation, removing the endothelium did increase the antagonist potency fivefold for BQ and threefold to fourfold for PD and SB. Although it is unlikely that insufficient equilibration time explains this increase in potency, the isolated perfused kidney is a complex model, and unknown factors such as protein binding could influence the accessibility of the antagonist to the vascular smooth-muscle receptor.
Warner et al. (4) concluded that there were at least two endothelin ETB-receptor subtypes, which he termed ETB1 (vascular endothelial cell) and ETB2 (vascular smooth muscle). The conclusion was largely based on results with PD, which lacked a significant inhibitory effect on ETB-mediated contraction in denuded rabbit pulmonary artery rings yet was quite potent at inhibiting ETB-mediated vasodilation in isolated perfused rat mesentery. In our study, PD was a potent (pA2 = 8.0) vascular smooth-muscle ETB antagonist in the denuded rat kidney, albeit fourfold less potent compared with its efficacy as an inhibitor of vascular endothelial ETB activity. It is possible that the differences between the studies are related to differences in preparations (organ perfusion vs. tissue bath) and to species differences (rabbit vs. rat) used to compare endothelial cell and smooth-muscle ETB receptor activity. However, in a study by Ohlstein et al. (6) with an in vivo rat blood pressure model for the ETB1-receptor study, and denuded rabbit pulmonary artery rings for the ETB2 receptor, Ohlstein reported results with SB similar to those obtained in our study. In Ohlstein's study, SB inhibited both ETB1 and ETB2 receptors. In our study, SB also inhibited both ETB1 and ETB2 receptors, but in addition, we were able to show a 13-fold greater selectivity for vascular endothelial ETB receptors. The only published pA2 values for PD, SB, and BQ were for ETB2-receptor activity with the denuded rabbit pulmonary artery rings (17,18). The pA2 values for PD, SB, and BQ in the rabbit pulmonary artery against S6c were 6.3, 7.7, and 6.2, respectively. Comparing the pA2 values for PD, SB, and BQ in the denuded rabbit pulmonary artery rings with the values in the denuded isolated perfused rat kidney (pA2 = 8, 8.6, and 8.4, respectively), the antagonists were 50-fold, eightfold, and 158-fold less potent in the rabbit. Significant antagonist-potency differences appear to exist between rat and rabbit with respect to the vascular smooth-muscle ETB receptor. The 50-fold difference in antagonist potency between rat and rabbit with PD could explain the difference in results between Warner's study and ours with respect to the efficacy of PD as an inhibitor of vascular smooth-muscle ETB activity. If species differences can significantly affect antagonist-potency relations, then we also must be sensitive to other factors such as the test systems and experimental conditions when comparing receptors.
If the differences between endothelial cell and smooth-muscle ETB receptors for PD and SB truly reflect differences in affinity between receptor subtypes, then receptor-binding studies, where affinity can be directly measured, should also demonstrate these differences. A direct comparison of receptor-binding affinity between rat vascular endothelial cell and rat vascular smooth-muscle cell ETB receptors would be best. Unfortunately, our attempt to culture commercially available or explanted rat vascular endothelial cells was unsuccessful. However, in a recent study by Flynn et al. (19), neither PD nor SB had significant differences in Ki values between human umbilical vein endothelial cells and human aortic smooth-muscle ETB receptors. It is possible that differences in receptor expression due to the experimental conditions (cell culture expression vs. native receptor expression in isolated perfused organs) may be responsible for this discrepancy. In addition, despite their high degree of homology, rat and human cloned ETB receptors have been shown to differ significantly with respect to both agonist and antagonist binding (20) and as such, human ETB receptor-binding profiles may not be predictive of rat ETB-receptor binding. Evidence supporting post-transcriptional differences in ETB receptors was recently published by Mizuguchi et al. (21). PD-sensitive vasodilator responses and PD-resistant smooth-muscle contractile responses to equipotent doses of S6c were completely absent in ETB-receptor gene knockout mice. These results indicate that it is possible to have two pharmacologically heterogeneous responses mediated by receptors derived from the same receptor gene.
The possible existence of ETB-receptor subtypes could have important implications for endothelin drug-discovery programs, not only in the discovery of new antagonists but also for interpreting the results from studies that have used existing antagonists. We conclude that in the rat vasculature, by using closely matched experimental conditions, PD, SB, and BQ were significantly more potent at inhibiting vascular endothelial cell ETB-receptor activity than vascular smooth-muscle cell ETB receptor activity. Although the pharmacologic data strongly suggest the presence of endothelin ETB-receptor subtypes, so far there is no supporting molecular evidence. Whether these differences between vascular endothelial cell and vascular smooth-muscle ETB receptors is due to receptor subtypes or some other posttranscriptional difference, the fact still remains that it takes more antagonist to functionally block the vascular smooth-muscle than the endothelial cell ETB receptors.
1. Takeshi S, Yanagisawa M, Masaki T. Molecular characterization of endothelin receptors. Trends Pharmacol Sci
2. Panek RL, Major TC, Hingorani GP, Doherty AM, Taylor DG, Rapundalo ST. Endothelin and structurally related analogs distinguish between endothelin receptor subtypes
. Biochem Biophys Res Commun
3. Takayanagi R, Kitazumi K, Takasaki C, et al. Presence of non-selective type of endothelin receptor on vascular endothelium and its linkage to vasodilation. FEBS Lett
4. Warner TD, Allcock GH, Corder R, Vane JR. Use of the endothelin antagonists BQ-123 and PD 142893 to reveal three endothelin receptor mediating smooth muscle contraction and the release of EDRF. Br J Pharmacol
5. Clozel M, Volker B, Gray GA, et al. Pharmacological characterization of bosentan: a new potent orally active nonpeptide endothelin receptor antagonist. J Pharmacol Exp Ther
6. Ohlstein EH, Nambi P, Douglas SA, et al. SB 209670, a rationally designed potent nonpeptide endothelin receptor antagonist. Proc Natl Acad Sci U S A
7. Ishikawa K, Ihara M, Noguchi K, et al. Biochemical and pharmacological profile of a potent and selective endothelin B- receptor antagonist, BQ-788. Proc Natl Acad Sci U S A
8. Williams DL, Jones KL, Pettibone DJ, Lis EV, Clineschmidt BV. Sarafotoxin S6c: an agonist which distinguishes between endothelin receptor subtypes
. Biochem Biophys Res Commun
9. Warner TD, Battistini B, Allcock GH, Vane JR. Endothelin ETA
receptors mediate vasoconstriction and prostanoid release in the isolated kidney of the rat. Eur J Pharmacol
10. Wu-Wong JR, Chiou WJ, Hoffman DJ, Winn M, von Geldern TW, Opgenorth TJ. Endothelins and endothelin receptor antagonists: binding to plasma proteins. Life Sci
11. Dominque S, Billing A, Krieger J-P, et al. Endothelium-dependent relaxation in the isolated rat kidney: impairment by cyclosporin A. J Cardiovasc Pharmacol
12. Douglas SA, Hiley CR. Endothelium-dependent vascular activities of endothelin-like peptides in the isolated superior mesenteric arterial bed of the rat. Br J Pharmacol
13. Arunlakshana O, Schild HO. Some quantitative uses of drug antagonists. Br J Pharmacol
14. Hubert JJ, Bohidar NR, Peace KE. In: Peace KE. ed. Assessment of pharmacological activity. Biopharmaceutical statistics for drug development.
New York: Marcel Dekker, 1988:120-4.
15. Le Monnier de Gouville AC, Lippton H, Cohen G, Icilio C, Hyman A. Vasodilator activity of endothelin-1 and endothelin-3: rapid development of cross-tachyphylaxis and dependence on the rate of endothelin administration. J Pharmacol Exp Ther
16. Takanari K, Heistad D, Faraci F. Dilatation of the basilar artery in response to selective activation of endothelin B receptors in vivo. J Pharmacol Exp Ther
17. Doherty AM, Cody WL, He JX, et al. In vitro and in vivo studies with a series of hexapeptide endothelin antagonists. J Cardiovasc Pharmacol
18. Hay DWP, Luttmann MA, Beck G, Ohlstein EH. Comparison of endothelinB
) receptors in rabbit isolated pulmonary artery and bronchus. Br J Pharmacol
19. Flynn MA, Haleen SJ, Welch KM, Reynolds EE. Endothelin B receptors on human endothelial cells and smooth muscle cells are pharmacologically equivalent. J Cardiovasc Pharmacol
20. Reynolds EE, Hwang O, Flynn MA, et al. Pharmacological differences between rat and human endothelin B receptors. Biochem Biophys Res Commun
21. Mizuguchi T, Nishiyama M, Moroi K, et al. Analysis of two pharmacologically predicted endothelin B receptor subtypes
by using the endothelin B receptor gene knockout mouse. Br J Pharmacol