Abstract
Aims
Vasoconstriction in human coronary artery can be mediated via activation of both 5-HT2 and 5-HT1B-receptors. Coronary vasoconstriction is a rare, but potential adverse effect of the antimigraine drug sumatriptan. In order to investigate the receptor population involved we compared the vasoconstrictor effects of sumatriptan (a mixed 5-HT1B/1D-receptor agonist) with those of L-775,606 (a selective 5-HT1D-receptor agonist) and 5-HT (the endogenous ligand) in human isolated coronary arteries.
Methods
Coronary arteries were obtained from human hearts removed prior to transplant surgery. Several endothelium denuded ring segments (4 mm in length) were obtained from each artery and mounted for isometric tension recording. Each segment was first exposed to 45 mm KCl and then to 5-HT (1 nm-100 μm ). Concentration-effect curves to L-775,606 (1-(3-(5-(1,2,4-triazol-4-yl)-1H-indol-3-yl)propyl)-4-(2-(3-fluorophenyl)ethyl)piperazine) and sumatriptan were then performed in a consecutive and random manner. The response to repeated application of 5-HT was obtained in separate segments.
Results
Twenty-five segments from seven different coronary arteries were studied. Concentration-effect curves were fitted to the data using nonlinear regression analysis. The maximum contraction for L-775,606 was significantly less than that for sumatriptan with Emax values (% relative to 45 mm KCl=100%) of 30.1±4.22 and 41.5±2.7, respectively. L-775,606 was significantly (30-fold) less potent than sumatriptan in causing contraction compared with sumatriptan (EC50 values were 6.0 μm and 0.2 μm, respectively). For comparison the Emax value for 5-HT was 77.2% and the EC50 value was 0.2 μm.
Conclusions
The selective 5-HT1D-receptor agonist L-775,606 has less propensity towards vasoconstriction in human isolated coronary artery (endothelium-denuded) than was mixed 5-HT1B/1D-receptor agonist sumatriptan. The contractions produced were at concentrations where L-775,606 would be expected to occupy 5-HT1B-receptors.
Keywords: 5-HT-receptor agonists, human coronary arteries, vasoconstriction
Introduction
Coronary side-effects, including isolated cases of cardiac ischaemia [1, 2] and myocardial infarction [3, 4], have been reported following sumatriptan administration. Although serious cardiac events following sumatriptan administration are rare, such cases are of concern as it is possible that drugs of this class will be given unknowingly to migraine patients with concomitant silent ischaemic heart disease. It should be noted that the more common drug-related side-effect of ‘chest symptoms’ (including feelings of chest tightness, heaviness or pressure) experienced by some migraineurs is not cardiac in origin but may be related to changes in oesophageal motility or in the pulmonary vasculature [5, 6].
The antimigraine activity of 5-HT1B/1D-receptor agonists is believed to be mediated through vasoconstriction of intracranial, extracerebral blood vessels and/or inhibition of the activation of trigeminal sensory nerves [7]. There is substantial evidence obtained using a variety of techniques (e.g. pharmacological characterization, RT-PCR, immunohistochemistry) to support the view that vasoconstrictor responses in human blood vessels (including coronary artery) are 5-HT1B-receptor mediated [8–10]. In contrast, inhibition of transmitter release from the terminals of trigeminal nerves is 5-HT1D-receptor mediated [11–13]. Thus a selective 5-HT1D-receptor agonist may be effective in treating migraine headache but have reduced vasoconstrictor effects.
L-775,606 (1-(3-(5-(1,2,4-triazol-4-yl)-1H-indol-3-yl) propyl)-4-(2-(3-fluorophenyl)ethyl)piperazine) is a selective 5-HT1D-receptor agonist. In cell lines expressing recombinant human 5-HT1D- or 5-HT1B-receptors, L-775,606 has been shown to have greater than 360-fold functional selectivity for 5-HT1D-receptors compared with 5-HT1B-receptors [14]. The aim of the present study was to compare the vasoconstrictor effects of L-775,606 on human isolated coronary artery with those of 5-HT and sumatriptan.
The collection of human material for these studies was approved by the University of Cambridge and Addenbrooke’s Hospital Ethics Committee.
Methods
Patients
Coronary arteries (left descending, left circumflex or right coronary arteries) were obtained from seven patients (six males, one female) undergoing orthotopic cardiac transplant. The age of these patients ranged from 43 to 55 years. Reasons for cardiac transplant included dilated cardiomyopathy, cardiomyopathy secondary to viral myocarditis, ischaemic heart disease and left ventricular failure secondary to aortic valve disease.
Preparation of arterial segments
Arteries were dissected from the heart immediately following explantation. Five out of the seven arteries tested were kept overnight (4° C) in physiological salt solution control prior to use and these arteries were allowed to equilibrate at 37° C before commencing the experiment. The arteries were prepared by removal of excess connective tissue and then the endothelium was removed mechanically. Each artery was cut into segments (each segment was 4 mm in length) with 3,4,5 or 6 segments being prepared from each artery. Individual ring segments were mounted for isometric tension recording in tissue baths containing physiological salt solution (for composition see below) aerated with 95% O2/5% CO2 maintained at pH 7.4, 37° C. The segments were stretched in a stepwise manner and the maximum length-tension curve was determined by measuring the response to 45 mm KCl and determining the point at which two different levels of stretch produced comparable responses to KCl. The segments were then allowed to equilibrate to their own basal tension level.
Experimental procedure
For all segments from the same artery, contractions to 45 mm KCl were first elicited and this response was used as a reference. To test for 5-HT responsiveness a cumulative concentration-effect curve to 5-HT was then obtained in all segments (Curve 1). The segments were then randomly allocated to one of three groups and the order of agonist application is shown in Table 1.
Table 1.
Order of agonist application.
For each agonist, the response to each concentration was allowed to plateau (up to 10 min) before addition of the next concentration. If no response was observed after 3 min, the next concentration of the agonist was added to the organ bath. Further additions of agonists were suspended on attaining the maximum contractile response (i.e. where addition of a higher agonist concentration no longer produced a further increase in tension). Following each concentration-effect curve the segments were washed several times in control solution until baseline tone was re-established.
Drugs and solutions
5-HT was obtained from Sigma, UK. L-775,606 (maleate salt) and sumatriptan (succinate salt) were synthesized by Merck Sharp and Dohme Research Laboratories. The physiological salt solution had the following composition (mm ): NaCl 90, KCl 5, MgSO4.7H2O 0.5, Na2HPO41.0, NaCO3 45, glucose 10, CaCl2 2.5; Na pyruvate 5, fumarate 5.
Analysis of data
Responses to 5-HT, sumatriptan and L-775,606 were measured in g tension and expressed as a percentage of the response to 45 mm KCl. Inter-agonist comparisons were made by pooling data obtained in curves 2 and 3. All analyses were performed using BMDP statistical software (University of California Press, Los Angeles, Vax/VMS release 7.0, 1993).
Comparison of overall effects of the agonists using analysis of variance (anova)
anovas were made on mean, standard error of mean, area of curve (AOC), linear slope values and all produced essentially similar results. Maximum and AOC, the variables judged to best describe the concentration-effect curves, are reported here. anovas were performed to test for: (a) desensitization of responses to 5-HT obtained in curves 1, 2 and 3; (b) differences in agonist evoked responses across arteries obtained from different patients; (c) differences between the effects of 5-HT, sumatriptan and L-775,606 (curves 2 and 3).
Comparison of concentration-effect curve parameters (Emax and EC50 values) using nonlinear regression analysis
For each agonist, concentration-effect curves were fitted to the mean data, using weighted least squares nonlinear regression analysis and the equation E=Emax/(1+(EC50/agonist concentration)(nH ), where Emax is the maximal contraction evoked by each agonist (relative to 45 mm KCl), EC50 is the half maximally effective concentration and nH is the Hill slope. The weighting was the reciprocal of the error estimate for each mean value. In all cases a good fit of the model was obtained and asymptotic ‘t’ tests were used to contrast the Emax and EC50 values obtained for the agonists.
Results
A total of 25 arterial segments from seven donors were studied. Some of the arteries were stored overnight prior to use. It was anticipated that this would not effect the outcome of the experiment since the arteries were equilibrated at 37° C before starting the experiment and comparisons of the effects of the agonists were made using a ‘within-segments’ design. The response to 45 mm KCl was 5.36±0.67 g. Desensitization to the 5-HT1B/1D-receptor agonists was tested in a two-way anova using the 5-HT data obtained in curves 1, 2 and 3. No statistically significant desensitization effect was found.
To test for interpatient variability in 5-HT responsiveness a two-way mixed design anova was made in which the grouping factor was artery (patient) and the within subject factor was agonist concentration. A significant artery (patient) effect (Chi-square=51.6, d.f.=6, P < 0.0001) and interaction (with agonist concentration) effect (Chi square=188.6, d.f.=54, P < 0.0001) were detected, showing that interarterial differences go beyond random variation and confirms the value of within artery agonist comparisons.
Post anova ‘t’ tests (paired comparisons) showed that contractile responses to both sumatriptan and L-775,606 were significantly lower than those for 5-HT (maximums: P = 0.023 and P < 0.0002, respectively). For L-775,606 the area under the concentration-effect curve (AOC) was also significantly lower than that for 5-HT (P < 0.0004, for comparison sumatriptan vs AOC: P = 0.0522) and significantly lower than the AOC for sumatriptan (P = 0.0412).
Comparison of concentration-effect curves parameters (Emax and EC50 values)
Concentration-effect curves for 5-HT, sumatriptan and L-775,606 (using data obtained in curves 2 and 3) are shown in Figure 1 and the Emax and EC50 values are summarized in Table 2. Asymptotic ‘t’ tests showed that the Emax values for both sumatriptan and L-775,606 were significantly lower than that for 5-HT (for 5-HT vs sumatriptan t = 8.82, d.f.=12, P < 0.00001; the difference between the Emax values was 35.7% with 95% confidence interval (CI) 22.9,48.5% and for 5-HT vs L-775,606 t = 9.10, d.f.=13, P < 0.00001; the difference between the Emax values was 47.0% with 95% CI 30.9,63.1%). The Emax value for L-775,606 was significantly less than that for sumatriptan (t = 2.25, d.f.=13, P < 0.021; the difference between the Emax values was 11.3% with 95% CI −4.21,26.8%). The EC50 value for sumatriptan was not significantly different from the value for 5-HT (t = 0.24, d.f.=12, P = 0.41; the difference between the EC50 values was 0.03 with 95% CI −0.35,0.41) whereas the EC50 value for L-775,606 was significantly different from the EC50 values for 5-HT (t = 8.18, d.f.=13, P < 0.00001; the difference between the EC50 values was 1.52 with 95% CI 0.99,2.06). The EC50 value for L-775,606 was significantly different from the value for sumatriptan (t = 7.50, d.f.=13, P < 0.00001; the difference between the EC50 values was 1.49 with 95% CI 0.89,2.09) showing that L-775,606 was 30-fold less potent in causing contraction than sumatriptan.
Figure 1.
Concentration-effect curves to 5-HT (closed squares), sumatriptan (closed circles) and L-775,606 (open diamonds) obtained in isolated segments of endothelium-denuded human coronary artery. Points represent mean data expressed relative to the contraction evoked by 45 mm KCl. Vertical bars signify ±s.e. mean. Curves were fitted to the data using nonlinear regression analysis.
Table 2.
Predicted Emax (% 45 m‘cf6 m KCl) and pEC50 values obtained from curve fitting to the overall mean data (obtained from Curves 2 and 3) using nonlinear regression analysis. The values shown represent mean with 95% confidence intervals (CI) (a total of 41 concentration-effect curves were performed on 25 segments obtained from arteries from seven different patients). Significantly different (P < 0.03) from 5-HT (*) and from 5-HT and sumatriptan (**).
Discussion
In the present study, all arteries tested responded to 5-HT and, similarly to previous reports, there was significant variability in 5-HT responsiveness between arteries obtained from different patients [15]. These findings support the value of using a crossover design, with respect to order of agonist application, enabling the direct (within-segment) comparison of the vasoconstrictor effects of sumatriptan and L-775,606 on the same segment of human isolated coronary artery.
In human isolated coronary artery 5-HT-evoked vasoconstriction can be mediated via activation of both 5-HT1B- and 5-HT2-receptors [10, 16, 17]. The maximum contractile response evoked by both sumatriptan and L-775,606 was lower than that evoked by 5-HT which most likely reflects the relatively weak affinity of both of these drugs for human 5-HT2-receptors [14]. The concentration-effect curve parameters for 5-HT and sumatriptan were in keeping with those previously reported [15]. Previously Nilsson et al. 1999 [18] have demonstrated the presence of 5-HT1B-receptor immunoreactivity on the endothelial layer in normal coronary artery, however, the physiological role of these receptors is not clear. In vitro studies using isolated coronary arteries have shown that 5-HT receptor agonists produce similar effects in endothelium-denuded arterial segments obtained from diseased hearts [15] and in endothelium intact segments obtained from donor (normal) hearts [19]. However, it is possible that the in vivo situation may be different since the blood vessel would only be exposed to 5-HT1B/1D-receptor agonists from the ‘blood-side’ and the endothelium may play a greater role in influencing coronary artery reactivity. Indeed, coronary angiographic studies have shown that in normal subjects 5-HT produces a vasodilator response but in patients with coronary artery disease this response is lost and vasoconstriction is seen [20, 21]. 5-HT1B/1D-receptor agonists are contraindicated in patients with known or suspected coronary artery disease.
The most important finding of the present study was the weak effect of L-775,606 on coronary artery reactivity. L-775,606 was markedly less potent than sumatriptan in causing contraction of human isolated coronary artery with the EC50 value for L-775,606 being 30-fold higher than the EC50 value for sumatriptan and in addition the maximum contraction was significantly lower than that of sumatriptan. L-775,606 is a selective 5-HT1D-receptor agonist. In cell lines expressing human recombinant receptors L-775,606 is 360-fold times more potent in activating 5-HT1D-receptors compared to 5-HT1B-receptors (EC50values of 1.2 nm and 440 nm, respectively [14]). For comparison, sumatriptan shows 16-fold selectivity for 5-HT1D-over 5-HT1B-receptors with EC50 values of 8.3 and 140 nm, respectively [14]. The weak potency of L-775,606 in causing coronary vasoconstriction is consistent with the activation of the 5-HT1B-receptor subtype and furthermore immunohistochemical studies in human coronary artery have shown the selective expression of 5-HT1B-receptor protein, but not 5-HT1D-receptor protein in this artery [10].
In clinical trials it has been shown that clinically relevant oral doses of sumatriptan (100 mg) and rizatriptan (10 mg) produce plasma levels of 47.4 ng ml−1 and 19.8 ng ml−1 (equivalent to 160 nm and 56 nm, respectively [22]). These plasma levels would be expected to activate peripheral 5-HT1D-receptors located on trigeminal nerve terminals surrounding blood vessels within the dura mater [11]. The EC50 at recombinant human 5-HT1D-receptors is approximately 8 nm for both drugs [23]). Since L-775,606 has eightfold higher potency at the 5-HT1D-receptor the predicted therapeutic plasma levels of free L-775,606 would be between 9 and 20 nm (assuming the antimigraine effect of sumatriptan and rizatriptan is mediated via 5-HT1D-receptor activation). Thus for L-775,606 there is greater than a 300-fold window between the EC50value (6 μm ) in coronary artery and the predicted therapeutic plasma levels of L-775,606. This is a substantial improvement over existing triptans. Recently, Maassen VanDenBrink et al. [19] have made similar comparisons for a series of triptans between EC50 values in isolated coronary artery and therapeutically relevant plasma levels (Cmax ), showing some apparent differences between triptans. However, small differences in EC50 values (obtained as part of a logarithmic dose–response relationship) can produce large differences in the EC50/Cmax ratio. For example, in the study by Maassen VanDenBrink et al. [19] the EC50 for zolmitriptan was 478 nm (equivalent to pEC50=6.3), giving a EC50/Cmax ratio of 0.02 whereas Nilsson et al. [18] reported an EC50 value of 251 nm (pEC50=6.5) for zolmitriptan giving a EC50/Cmax ratio of 0.1 (compared with a ratio of 0.18 for sumatriptan). Therefore, apparent differences between triptans currently used as antimigraine drugs should be viewed with caution, since the differences may be small and all drugs will carry similar cardiac risk liability. Selective 5-HT1D-receptor agonists, such as L-775,606 may offer a real advantage since vasoconstrictor effects in human isolated coronary artery are small in magnitude and the window between coronary vasoconstriction and predicted therapeutic plasma levels is greater than 300-fold (EC50/Cmax ratio=<0.003). The efficacy of selective 5-HT1D-receptor agonists as antimigraine agents needs to be established.
In conclusion, L-775,606 (a selective 5-HT1D-receptor agonist) was markedly less potent than sumatriptan (a mixed 5-HT1B/1D-receptor agonist) in causing contraction of human isolated coronary artery. This observation is consistent with the suggestion that vasoconstriction in human coronary artery is mediated via the 5-HT1B-receptor subtype and with the view that selective 5-HT1D-receptor agonists are likely to have reduced vasoconstrictor properties in man.
Acknowledgments
We thank the cardiac surgeons and technicians at Papworth Hospital for collection of human material, Dr A.P. Davenport for his advice in the preparation of this manuscript and Mr Richard Jelley, Austin Reeve and Drs Francine Sternfeld and Tamy Ladduwahetty for the synthesis of L-775,606.
References
- 1.Willett F, Curzen N, Adams J, Armitage M. Coronary vasospasm induced by subcutaneous sumatriptan. Br Med J. 1992;304:1415. doi: 10.1136/bmj.304.6839.1415. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Lloyd DK, Simmons V. Adverse reactions associated with sumatriptan. Lancet. 1993;341:1092. [PubMed] [Google Scholar]
- 3.O’Connor P, Gladstone P. Oral sumatriptan-associated myocardial infarction. Neurol. 1995;45:2274–2276. doi: 10.1212/wnl.45.12.2274. [DOI] [PubMed] [Google Scholar]
- 4.Ottervanger JP, Paalman JJA, Boxma GL, Stricker BHC. Transmural myocardial infarction with sumatriptan. Lancet. 1993;341:861–862. doi: 10.1016/0140-6736(93)93064-8. [DOI] [PubMed] [Google Scholar]
- 5.Houghton LA, Foster JM, Whorwell PJ, Morris J, Fowler P. Is chest pain after sumatriptan oesophageal in origin. Lancet. 1994;344:985–986. doi: 10.1016/s0140-6736(94)91642-x. [DOI] [PubMed] [Google Scholar]
- 6.MacIntyre PD, Bhargava B, Hogg KJ, Gemmill JD, Hillis WS. Effect of subcutaneous sumatriptan, a selective 5HT1 agonist, on the systemic, pulmonary, and coronary circulation. Circulation. 1993;87:401–405. doi: 10.1161/01.cir.87.2.401. [DOI] [PubMed] [Google Scholar]
- 7.Humphrey PPA, Feniuk W. Mode of action of the antimigraine drug sumatriptan. Trends Pharmacol Sci. 1991;12:444–446. doi: 10.1016/0165-6147(91)90630-b. [DOI] [PubMed] [Google Scholar]
- 8.Hamel E, Fan E, Linville D, Ting V, Villemure J-G, Chia L-S. Expression of mRNA for the serotonin 5-hydroxytryptamine1Dβ receptor subtype in human ‘qland bovine cerebral arteries. Mol Pharmacol. 1993;44:242–246. [PubMed] [Google Scholar]
- 9.Ullmer C, Schmuck K, Kalkman HO, Lubbert H. Expression of serotonin receptor mRNAs in blood vessels. FEBS Lett. 1995;370:215–221. doi: 10.1016/0014-5793(95)00828-w. [DOI] [PubMed] [Google Scholar]
- 10.Longmore J, Razzaque Z, Shaw D, et al. Comparison of the vasoconstrictor effects of rizatriptan and sumatriptan in human isolated cranial arteries: immunohistological demonstration of the involvement of 5-HT1B-receptors. Br J Clin Pharmacol. 1998;46:577–582. doi: 10.1046/j.1365-2125.1998.00821.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Longmore J, Shaw D, Smith D, et al. Differential distribution of 5-HT1B- and 5-HT1D-immunoreactivity within the human trigemino-cerebrovascular system: implications for the discovery of new anti-migraine drugs. Cephalalgia. 1997;17:833–842. doi: 10.1046/j.1468-2982.1997.1708833.x. [DOI] [PubMed] [Google Scholar]
- 12.Cumberbatch MJ, Hill RG, Hargreaves RJ. Rizatriptan has central antinociceptive effects against durally evoked responses. Eur J Pharmacol. 1997;328:37–40. doi: 10.1016/s0014-2999(97)83024-5. [DOI] [PubMed] [Google Scholar]
- 13.Williamson DJ, Shepheard SL, Hill RG, Hargreaves RJ. The novel anti-migraine agent rizatriptan inhibits neurogenic dural vasodilation and extravasation. Eur J Pharmacol. 1997;328:61–64. doi: 10.1016/s0014-2999(97)83028-2. [DOI] [PubMed] [Google Scholar]
- 14.MacLeod AM, Street LJ, Reeve AJ, et al. Selective, orally active 5-HT1D receptor agonists as potential antimigraine agents. J Med Chem. 1997;40:3501–3503. doi: 10.1021/jm9704560. [DOI] [PubMed] [Google Scholar]
- 15.Longmore J, Boulanger CM, Desta B, Schofield WN, Hill RG, Taylor AA. 5-HT1D-receptor agonists and human coronary artery reactivity in vitro: crossover comparisons of 5-HT and sumatriptan with MK-462 and L-741,519. Br J Clin Pharmacol. 1996;42:431–441. doi: 10.1046/j.1365-2125.1996.04217.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Connor HE, Feniuk W, Humphrey PPA. 5-Hydroxytryptamine contracts human coronary arteries predominantly via 5-HT2 receptor activation. Eur J Pharmacol. 1989;161:91–94. doi: 10.1016/0014-2999(89)90184-2. [DOI] [PubMed] [Google Scholar]
- 17.Kaumann AJ, Frenken M, Posival H, Brown AM. Variable participation of 5-HT1-like receptors and 5-HT2 receptors in serotonin-induced contraction of human isolated coronary arteries. Circulation. 1994;90:1141–1153. doi: 10.1161/01.cir.90.3.1141. [DOI] [PubMed] [Google Scholar]
- 18.Nilsson T, Longmore J, Shaw D, et al. Characterisation of 5-HT receptors in human coronary arteries by molecular and pharmacological techniques. Eur J Pharmacol. 1999;372:49–56. doi: 10.1016/s0014-2999(99)00114-4. [DOI] [PubMed] [Google Scholar]
- 19.Maassen VanDenBrink A, Reekers M, Bax WA, et al. Coronary side-effect potential of current and prospective antimigraine drugs. Circulation. 1998;98:25–30. doi: 10.1161/01.cir.98.1.25. [DOI] [PubMed] [Google Scholar]
- 20.Golino P, Piscioni F, Willerson JT, et al. Divergent effects of serotonin on coronary artery dimensions and blood flow in patients with coronary atherosclerosis and control patients. N Engl J Med. 1991;324:641–648. doi: 10.1056/NEJM199103073241001. [DOI] [PubMed] [Google Scholar]
- 21.McFadden EP, Clarke JG, Davies GJ, et al. Effect of intracoronary serotonin on coronary vessels in patients with stable angina and patients with variant angina. N Engl J Med. 1991;324:648–654. doi: 10.1056/NEJM199103073241002. [DOI] [PubMed] [Google Scholar]
- 22.Sciberras DG, Polvino WJ, Gertz BJ, et al. Initial human experience with MK-462 (rizatriptan): a novel 5-HT1D agonist. Br J Clin Pharmacol. 1997;43:49–54. doi: 10.1111/j.1365-2125.1997.tb00032.x. [DOI] [PubMed] [Google Scholar]
- 23.Beer MS, Middlemiss DN, Stanton JA, et al. An in vitro pharmacological profile of rizatriptan. Headache. 1997;37:223. [Google Scholar]