Endothelium-dependent relaxation has been reported to be impaired in the blood vessels of hypertensive rats (1-4). The impairment of endothelium-dependent relaxation involves different mechanisms in different forms of hypertension both in animals (5,6) and in humans (7). The impairment of relaxation can be attributed either to the decreased release of endothelium-derived relaxing factor (EDRF) or to increased release of endothelium-derived contracting factor (EDCF). EDCF has been regarded as a product of the cyclooxygenase pathway of the arachidonic acid cascade, because the impaired relaxation can be restored in the presence of indomethacin (8-11). Most of the relevant experiments have been performed on acetylcholine (ACh)-induced relaxation and in purine nucleotide-induced relaxation (8,12) in preparations contracted by agents such as adrenergic stimulants.
Contractions induced by adrenergic stimulants are decreased by the endothelium (13-18) due to the spontaneous release of EDRF or the release of EDRF by stimulation of the adrenoceptors of the vascular endothelium. In the latter case, stimulation of the α2-adrenoceptors is believed to be more effective in releasing EDRF than is stimulation of α1-adrenoceptors (13,15,19). We have observed that the decrease in the norepinephrine (NE)-induced contraction is less prominent in aorta of strokeprone spontaneously hypertensive rats (SHRSP) (17). In addition, the possibility of changes in the adrenoreceptors in hypertensive rats has been reported (20). In the present experiments, the differences in the endothelium-dependent relaxation induced by α2-agonists in aorta from SHRSP and normotensive Wistar Kyoto rats (WKY) were studied.
SHRSP and WKY aged 16 weeks were used. They were originally developed by Dr. K. Okamoto (21) and were successively bred in our animal facility. The blood pressure (BP) of the rats was measured by the tail-cuff method. Animals were handled in accordance with the Kinki University guidelines for the ethical use of animals (February 1994). The rats were killed under anesthesia with ethylether by cutting of the carotid artery. The aorta was excised from the thoracic cavity and immersed in a modified Tyrode's solution of the following composition (in mM): NaCl 137, KCl 5.4, CaCl2 2.0, MgCl2 1.0, NaHCO3 11.9, NaH2PO4 0.4, glucose 5.6, and Ca(II)-EDTA 0.026 equilibrated with a gas mixture of 95% O2/5% CO2 at 37 °C. High-K Tyrode's solution was made by replacing the NaCl with equimolar KCl.
Each aorta was cleaned of fat and connective tissue, and ring preparations 1 mm long were made from the thoracic portion. In some rings, the luminal surface was scrubbed with rubbers to remove the endothelium. A complete removal of the endothelium was confirmed by a loss of relaxation by ACh 10-5M. Two tungsten wires (30 μm in diameter) were inserted into the lumen, and the preparations were mounted with one tungsten wire in an organ bath filled with the modified Tyrode's solution. The other tungsten wire was connected to a force-displacement transducer so that the changes in tension could be measured isometrically.
The preparations were equilibrated for at least 1 h under a stretch tension of 8 mN. This stretch-tension value was chosen because the maximum endothelium-dependent relaxation was observed at this value in both rat strains (22). We then subjected the preparation to two successive high-K-induced contractions by changing the solution from the modified Tyrode's to a high-K Tyrode's solution containing 50 mM K+. These procedures were required to obtain constant values of contractions in the subsequent experiments (17,18). Endothelium-dependent relaxation was induced by applying α2-agonist clonidine or UK-14304 (23) or ACh to the preparation, which was contracted in the presence of ONO-11113, a stable thromboxane A2 analogue (24). The concentration of ONO-11113 chosen to induce precontraction was 10-9M. This dose caused 64 ± 4% (n = 8) of the maximum contraction in the WKY preparations.
A modified sandwich bioassay technique was used to evaluate endothelium-derived factors (25). We inserted two pieces of endothelium-intact sheets (1 × 1 mm) into the endothelium-denuded ring preparations so that the intimal side of the sheets met the luminal surface of the ring.
The drugs used were clonidine (Sigma, St. Louis, MO, U.S.A.), 5-bromo-6-(2-imidazolin-2-ylamino)-quinoxaline (UK-14304, Pfizer, Osaka), (+)-9, 11-Epithia-11, 12-methano-TXA2 (ONO-11113, ONO, Osaka), ACh hydrochloride (Sigma), NG-nitro-L-arginine (L-NNA, Sigma), prazosin hydrochloride (Sigma), propranolol hydrochloride (Sigma), yohimbine hydrochloride (Sigma), indomethacin (Wako, Osaka, Japan), and Ca(II)-EDTA (Wako).
Values are mean ± SEM; the number of animals is shown in parentheses. Data were analyzed by analysis of variance (ANOVA) or Student's t test; p < 0.05 was considered significant.
BP of the rats
The systolic BP (SBP) of WKY and SHRSP aged 16 weeks was 132 ± 1.2 mm Hg (n = 40) and 242 ± 1.7 mm Hg (n = 45), respectively. The SBP of the SHRSP was significantly higher than that of the WKY.
Relaxation and contraction by α2-agonists
The endothelium-intact aortic ring preparations showed stable contraction in the presence of 10-9M ONO-11113. The contraction amplitude was 4.8 ± 0.5 mN (n = 32) and 5.7 ± 0.4 mN (n = 33) in the WKY and SHRSP preparations, respectively. Application of clonidine or UK-14304 induced relaxation in a dose-dependent manner (Fig. 1). The relaxation induced by UK-14304 10-6M in the WKY preparations was 38.0 ± 2.5% (n = 5). The relaxation in the presence of prazosin (10-6M) and propranolol (10-6M) was 41.6 ± 4.8% (n = 5), not different from that in the absence of these drugs. The following experiments were performed in the presence of these drugs.
In the WKY preparations, the relaxing response to clonidine was observed from 10-7M and the maximum relaxation was attained at 10-5M. The maximum relaxation was 25.7 ± 3.9% (n = 8) of the 10-9M ONO-11113-induced contraction. In the SHRSP preparations, relaxation was observed at concentrations ≥10-7M, and the maximum relaxation was attained at 10-4M. The maximum relaxation in the SHRSP preparations was 12.5 ± 3.3% (n = 6) of the ONO-11113-induced contraction, a significantly lesser relaxation than that observed in the WKY aorta.
Similar results were obtained with UK-14304 (Fig. 1). The threshold concentration for the relaxation was 10-8M in preparations from both WKY and SHRSP. The maximum relaxation was achieved between 10-6 and 10-5M. The relaxation at 10-5M was 37.4 ± 4.3% (n = 9) and 8.6 ± 2.0% (n = 10) of the 10-9M ONO-11113-induced contraction, in the WKY and SHRSP preparations, respectively.
In the presence of L-NNA 10-4M, the contraction induced by 10-9M ONO-11113 was 6.7 ± 0.4 mN (n = 8) in WKY and 6.7 ± 0.6 mN (n = 8) in SHRSP. The relaxation induced by clonidine (data not shown) and by UK-14304 was blocked by the treatment with L-NNA in both the WKY and SHRSP preparations (Fig. 2). When UK-14304 was applied during a stable contraction induced by 10-9M ONO-11113, UK-14304 10-7 -10-5M did not induced relaxation in an endothelium-denuded preparation.
To test a permissive role of nitric oxide (NO) (26), we applied UK-14304 to endothelium-denuded preparations in the presence of sodium nitroprusside (SNP), an NO donor. In the presence of SNP at a subthreshold concentration (17) of 10-11M, UK-14304 10-6M did not relax the preparation (-1.5 ± 1.5%, n = 4). An increase in SNP concentration from 10-10 to 10-8M relaxed the preparation but did not permit UK-14304 to evoke relaxation.
To evaluate the relaxing effect of substances released from endothelium, we used a recipient-donor experiment with a modified sandwich method. Endothelium-denuded rings (recipient) made from WKY aorta were contracted by 10-9M ONO-11113. UK-14304 did not change the ONO-11113-induced contraction of the recipient ring. However, UK-14304 relaxed the endothelium-denuded aorta ring when endothelium-intact sheets (donor) were included in lumen of the ring. We used UK-14304 10-6M to induce relaxation because we could obtain near-maximum relaxation with this concentration. The magnitude of relaxation is expressed as percent of ONO-11113 contraction. The relaxation was 8.2 ± 1.1% (n = 10) and 2.6 ± 0.6% (n = 9) in the endothelium-denuded WKY rings with endothelium-intact sheets from WKY and SHRSP, respectively (Fig. 3). The relaxation with the SHRSP sheets was less than that with WKY sheets. The relaxation in the sandwich experiment was less than that in the endothelium-intact rings, but is understandable when the endothelial area of the sheets is taken into consideration. L-NNA 10-4M abolished the relaxation in sandwich experiment with both WKY and SHRSP endothelium donor sheets.
In the absence of drug-induced precontraction, UK-14304 did not contract endothelium-intact WKY rings. When NO synthesis was inhibited by L-NNA 10-4M, UK-14304 contracted the rings concentration dependently. The threshold was 3 × 10-8M, and maximal contraction was attained at 10-5M in both endothelium-intact and endothelium-denuded preparations. There was no remarkable difference in the contraction amplitude between endothelium-intact and endothelium-denuded preparations; e.g., the 10-6M UK-14304-induced contraction was 45.8 ± 10.3% (n = 8) and 60.6 ± 16.0% (n = 7) of the high-K 50 mM Tyrode's solution-induced contraction in endothelium-intact and endothelium-denuded WKY preparations, respectively, and 42.4 ± 4.7% (n = 5) and 60.1 ± 7.0% (n = 7) of the high-K Tyrode's solution-induced contractions in endothelium-intact and endothelium-denuded SHRSP preparations, respectively. ANOVA showed no difference among these values. Therefore, there was no evidence to support induction of endothelium-dependent contraction by UK-14304.
ACh >10-8M, induced dose-dependent relaxation in both the WKY and SHRSP preparations precontracted in the presence of 10-9M ONO-11113 (Fig. 4). The maximum relaxation was observed at 10-4M in both the WKY and SHRSP preparations. The maximum relaxation was 74.2 ± 3.3% (n = 7) and 47.6 ± 2.1% (n = 6) of the ONO-11113-induced contraction in WKY and SHRSP. The maximum relaxation was less in the SHRSP preparations. The relaxation induced by ACh was blocked by L-NNA (10-4M) in both WKY and SHRSP (Fig. 4). ACh did not induce relaxation in endothelium-denuded preparations.
Effect of indomethacin
The amplitude of contraction induced by 10-9M ONO-11113 was 4.3 ± 0.7 mN (n = 6) and 5.5 ± 0.2 mN (n = 8) in WKY and SHRSP preparations, respectively. The endothelium-dependent relaxation induced by ACh was potentiated by indomethacin (10-5M) in the SHRSP preparations, but not in the WKY preparations (Fig. 5). However, the endothelium-dependent relaxation induced by UK-14304 was not influenced by indomethacin in either WKY or SHRSP preparations (Fig. 6).
We analyzed drug-receptor action by fitting a concentration-response curve of ACh- or UK-14304-induced relaxation. These curves were fitted with Hill coefficients of 0.9-1.2 in both WKY and SHRSP, with and without indomethacin, with correlation coefficients of 0.99.
Endothelial cells in an unstimulated state release EDRF spontaneously (27). This tonic basal release of EDRF may decrease the contractions induced by vasoconstrictors. Stimulation of α2-adrenoceptors of the endothelium has been reported to cause the release of EDRF (4,19,28-31). In our previous study endothelium decreased NE-induced contraction of aortas from both WKY and SHRSP (32). Both α1- and α2-adrenoceptors in the endothelium may be involved in EDRF release, since both phenylephrine- and clonidine-induced contractions were decreased by the endothelium, although the latter was decreased to a greater degree (18). Similar results have been reported in other vascular preparations (13,15,19). We also observed that the decrease in contractions caused by the endothelium was impaired in aorta from SHRSP (32). The decrease in the release of NO is believed to be a cause of the impairment, since the decrease was blocked by L-NNA, a blocker of the synthesis of NO (31,33).
In the present study, clonidine and UK-14304, an α2-adrenoceptor agonist (23), could induce relaxation in aorta contracted by ONO-11113, a stable thromboxane analogue (24). We previously reported that relaxation of a preparation by clonidine or guanabenz was not attenuated in the presence of prazosin and propranolol (18), which indicates that relaxation is caused by the stimulation of α2-adrenoceptors. We have now shown that the relaxation induced by UK-14304 was also unaffected by prazosin or propranolol, indicating that α1- or β-adrenoceptors are not involved in the relaxation. Although both clonidine and UK-14304 are α2-adrenoceptor agonists, the maximum relaxation induced by clonidine was less than that induced by UK-14304, possibly because clonidine is a partial α2-adrenoceptor agonist whereas UK-14304 is a full agonist (34).
An important factor in endothelium-dependent relaxation is NO, and NO synthesis can be blocked by compounds such as N-monomethyl-L-arginine (L-NMMA), or L-NNA (35). In the present experiment, the relaxation induced by α2-adrenoceptor stimulation could be blocked by L-NNA, suggesting that the relaxation is mediated by NO. Involvement of smooth muscle α2-adrenoceptors in the relaxation can be excluded, since stimulation of these receptors in uncontracted endothelium-denuded preparations causes contractions in rat aorta (18,36). In the present study, UK-14304 did not relax the ONO-11113-induced contraction in endothelium-denuded preparations. Therefore, we conclude that the relaxation is due to NO released by stimulation of the endothelium α2-adrenoceptors.
A permissive role of NO in UK-14304-induced relaxation was recently reported in endothelium-denuded rat cerebral artery (26). However, in the present study, no remarkable relaxation was observed in the endothelium-denuded preparations in the presence of SNP, an NO donor. Therefore, NO itself plays a main role in the UK-14304-induced relaxation in rat aorta.
The relaxation induced by α2-adrenoceptor stimulation was impaired in the SHRSP preparations. The impairment could be caused either by the decreased release of EDRF (NO) or the increased release of EDCF (37). EDCF, which is involved in the impairment of the endothelium-dependent relaxation induced by ACh or purine nucleotides in SHR, is believed to be a product of the cyclooxygenase pathway of the arachidonic acid cascade, since the impaired relaxation was improved by indomethacin (2,8-12). In the present study, the impaired ACh-induced relaxation of SHRSP aorta was improved by indomethacin, confirming the involvement of EDCF in the impairment, as previously reported (8-11). However, the impaired relaxation by α2-adrenoceptor stimulation in the SHRSP preparation was not improved by the treatment with indomethacin. The Hill coefficient values in ACh- and UK-14304-induced relaxation were ≈1 for both WKY and SHRSP preparations. Therefore, receptor kinetics changes were not involved in the impaired relaxation.
UK-14304 contracted the aortas in a concentration-dependent manner under NO synthesis inhibition. There was no remarkable difference between endothelium- intact and endothelium-denuded preparations in the UK-14304-induced contractions. No endothelium-dependent contraction was induced by UK-14304, which suggests that relaxation by α2-agonists in WKY and SHRSP preparations is not affected by EDCF.
The relaxation induced by UK-14304 is abolished by L-NNA; therefore, the relaxation is mediated mainly by NO. We observed that SNP, an NO releaser (38), relaxed the endothelium-denuded SHRSP and WKY aorta rings to the same degree (17). Therefore, relaxing response to NO is not different between WKY and SHRSP. The results obtained in our modified sandwich experiment clearly show that the change in endothelium-derived NO is responsible for the decreased relaxation in SHRSP. Although we did not measure the amount of NO or cyclic GMP, our results indicate that the impairment of α2-adrenoceptor-induced endothelium-dependent relaxation in the SHRSP preparations is due to the reduction of NO release from endothelium.
The reduced release of NO with or without a change in the α2-adrenoceptors of the endothelium may be the cause of the impairment of relaxation. The impairment of relaxation induced by stimulation with α2-adrenoceptor is greater than that induced by stimulation with ACh. α2-Adrenoceptor agonists caused a greater degree of impairment in relaxation than did ACh as calculated by comparison of the relaxation in SHRSP and WKY (77, 51, and 36% impairment induced in the relaxation by UK-14304, clonidine, and ACh, respectively). In smooth muscle of tail artery, the α1-adrenoceptor response was comparable in normotensive and hypertensive rats (39), and the α2-adrenoceptor-mediated response was enhanced in hypertensive rats (20,40). If changes in density or affinity of the endothelium α2-adrenoceptor are involved, they should be decreased, in contrast to the results reported for arterial smooth muscle. The mechanisms involved in the impaired relaxation in the aorta of SHRSP remain to be investigated (41). In endothelium, α2-adrenoceptor agonists stimulated Gi protein (42), whereas stimulation of M3 receptors (43) and purinoceptors were mediated by phospholipase C through another G protein (41). These pathways may be modulated differently in hypertension, although the detailed mechanisms remain unclarified.
α2-Agonist-induced endothelium-dependent relaxation was impaired in the SHRSP aorta, due to the decrease in NO release and not to the increased release of EDCF. The α2-adrenoceptors of the endothelium and/or the synthesis of NO through the stimulation of the receptors may be altered in SHRSP.
1. Konishi M, Su C. Role of endothelium in dilator responses of spontaneously hypertensive rat arteries. Hypertension
2. Luscher TF, Vanhoutte PM. Endothelium-dependent contractions to acetylcholine in the aorta of the spontaneously hypertensive rat. Hypertension
3. Sunano S, Osugi S, Shimamura K. Blood pressure and impairment of endothelium-dependent relaxation in spontaneously hypertensive rats
4. Sunano S, Osugi S, Kaneko K, Yamamoto K, Shimamura K. Effects of chronic treatment with SQ29852 on spontaneous smooth muscle tone and endothelium-dependent relaxation in aorta of stroke-prone spontaneously hypertensive rats
. J Cardiovasc Pharmacol
5. Luscher TF, Vanhoutte PM. Endothelium-dependent contractions to acetylcholine in the aorta of the spontaneously hypertensive rats
6. Luscher TF, Raij L, Vanhoutte PM. Endothelium-dependent vascular responses in normotensive and hypertensive Dahl rats. Hypertension
7. Taddei S, Virdis A, Mattei P, Salvetti A. Vasodilation to acetylcholine in primary and secondary forms of human hypertension. Hypertension
8. Kato T, Iwama Y, Okumura K, Hashimoto H, Ito T, Satake T. Prostaglandin H2
may be the endothelium-derived contracting factor released by acetylcholine in the aorta of the rat. Hypertension
9. Diedrich D, Yang Z, Buhler FR, Luscher T. Impaired endothelium-dependent relaxations in hypertensive resistance arteries involve cyclooxygenase pathway. Am J Physiol
10. Fu-Xiang D, Skopec J, Diedrich A, Diedrich D. Prostaglandin H2
and thromboxane A2
are contractile factors in intrarenal arteries of spontaneously hypertensive rats
11. Li J, Bukoski RD. Endothelium-dependent relaxation of hypertensive resistance arteries is not impaired under all conditions. Circ Res
12. Mombouli J-V, Vanhoutte PM. Purinergic endothelium-dependent and -independent contractions in rat aorta. Hypertension
13. Miller VM, Vanhoutte PM. Endothelial α2
-adrenoceptors in canine pulmonary and systemic blood vessels. Eur J Pharmacol
14. White RE, Carrier GO. α1
- and α2
-adrenoceptor agonist-induced contraction in rat mesenteric artery upon removal of endothelium. Eur J Pharmacol
15. Bullock GR, Taylor SG, Westone AH. Influence of the vascular endothelium on agonist-induced contractions and relaxations in rat aorta. Br J Pharmacol
16. Godfraind T, Alosachie I. Influence of endothelium and cyclic GMP on alpha-adrenoceptors. In: Vanhoutte PM, ed. Vasodilatation: vascular smooth muscle, peptide, autonomic nerves, and endothelium
. New York: Raven Press, 1988:437-42.
17. Osugi S, Shimamura K, Sunano S. Decreased modulation by endothelium of noradrenaline-induced contraction in aorta from stroke-prone spontaneously hypertensive rats
. Arch Int Pharmacodyn
18. Kaneko K, Sunano S (1993) Involvement of α-adrenoceptors in the endothelium-dependent depression of noradrenaline-induced contraction in rat aorta. Eur J Pharmacol
19. Egleme C, Godfraind T, Miller RC. Enhanced responsiveness of rat isolated aorta to clonidine after removal of the endothelium cells. Br J Pharmacol
20. Medgett IC, Hicks PE, Langer SZ. Smooth muscle alpha-2 adrenoceptors mediate vasconstrictor responses to exogenous norepinephrine and to sympathetic stimulation to a greater extent in spontaneously hypertensive than in Wistar Kyoto rat tail arteries. J Pharmacol Exp Ther
21. Okamoto K, Yamori Y, Nagaoka A. Establishment of stroke-prone spontaneously hypertensive rats
(SHR). Circ Res
22. Sunano S, Sasaki F, Osugi S, Shimamura K. Comparison of endothelium-dependent and -independent tension oscillation in aortae of stroke-prone spontaneously hypertensive rats
and Wistar Kyoto rats. J Smooth Muscle Res
23. Cambridge D. UK14,304, a potent and selective α2
-agonist for the characterization of α-adrenoceptor subtypes. Eur J Pharmacol
24. Toda N, Nakajima M, Okamura T, Miyazaki M. Interactions of thromboxane A2
analogs and prostaglandins in isolated dog arteries. J Cardiovasc Pharmacol
25. Furchgott RF, Zawadzki JV. The obligatory role of endothelial cells in the relaxation of arterial smooth muscle by acetylcholine. Nature
26. Bryan RM, Steenberg ML, Eichler MY, Johnson TD, Swafford MWG, Suresh MS. Permissive role of NO in α2
-adrenoceptor-mediated dilations in rat cerebral arteries. Am J Physiol
27. Martin W, Furchgott RF, Villani GM, Jothianandan D. Depression of contractile responses in rat aorta by spontaneously released endothelium-derived relaxing factor
. J Pharmacol Exp Ther
28. Cooks TM, Angus JA. Endothelium-dependent relaxation of coronary arteries by noradrenaline and serotonin. Nature
29. Murakami K, Karaki H, Urakawa N. Role of endothelium in the contraction induced by norepinephrine and clonidine in rat aorta. Jpn J Pharmacol
30. Urabe M, Kawasaki H, Takasaki K. Effect of endothelium removal on the vasoconstrictor response to neurally released 5-hydroxytryptamine and noradrenaline in the rat isolated mesenteric and femoral arteries. Br J Pharmacol
31. Vo PA, Reid JJ, Rand MJ. Attenuation of vasoconstriction by endogenous nitric oxide in rat caudal artery. Br J Pharmacol
32. Kaneko K, Yamamoto K, Sasaki F, Sunano S. Effects of nitric oxide (NO) synthesis inhibitor in aorta of SHRSP. Jpn Heart J
33. Moore PK, al-Swayeh OA, Chong NWS, Evans RA, Gibson A. L-NG
-nitro arginine (L-NOARG), a novel, L-arginine-reversible inhibitor of endothelium-dependent vasodilation in vitro. Br J Pharmacol
34. Grant JA, Scrutton MC. Interaction of selective α-adrenoceptor agonists and antagonists with human and rabbit blood platelets. Br J Pharmacol
35. Moncada S, Palmer RMJ, Higgs EA. Nitric oxide: physiology, pathophysiology and pharmacology. Pharmacol Rev
36. Carrier GO, White RE. Enhancement of alpha-1 and alpha-2 agonist-induced vasoconstriction by removal of endothelium in rat aorta. J Pharmacol Exp Ther
37. Luscher TF, Vanhoutte PM. Mechanisms of altered endothelium-dependent responses in hypertensive blood vessels. In: Vanhoutte PM, ed. Relaxing and contracting factors
. Clifton, NJ: Humana Press, 1988:495-509.
38. Butler AR, Flitney FW, Williams DLH. NO, nitrosonium ions, nitroxide ions, nitrosothiols and iron-nitrosyls in biology: a chemist's perspective. Trends Pharmacol Sci
39. Bhalla RC, Aqel MB, Sharma RV. Alpha 1-adrenoceptor-mediated responses in the vascular smooth muscle of spontaneously hypertensive rats
. J Hypertens
40. Weiss RJ, Webb RC, Smith CB. Comparison of alpha2
adrenoceptors on arterial smooth muscle and brain homogenates from spontaneously hypertensive and Wistar-Kyoto rats. J Hypertens
41. Pearson PJ, Vanhoutte PM. Vasodilator and vasoconstrictor substances produced by the endothelium. Rev Physiol Biochem Pharmacol
42. Liao JK, Homcy CJ. The release of endothelium-derived relaxing factor
via α-adrenergic receptor activation is specifically mediated by Giα2
. J Biol Chem
43. Boulanger CM, Morrison KJ, Vanhoutte PM. Mediation by M3
-muscarinic receptors of both endothelium-dependent contraction and relaxation to acetylcholine in the aorta of the spontaneously hypertensive rat. Br J Pharmacol