Evidence that anandamide and EDHF act via different mechanisms in rat isolated mesenteric arteries


Department of Pharmacology, University of Bristol, University Walk, Bristol, BS8 1TD


The endogenous cannabinoid, anandamide, has been suggested as an endothelium-derived hyperpolarizing factor (EDHF). We found that anandamide-evoked relaxation in isolated segments of rat mesenteric artery was associated with smooth muscle hyperpolarization. However, although anandamide-evoked relaxation was inhibited by either charybdotoxin (ChTX) or iberiotoxin, inhibition of the relaxation to EDHF required a combination of ChTX and apamin. The relaxations induced by either anandamide or EDHF were not inhibited by the cannabinoid receptor (CB1) antagonist SRI41716A, or mimicked by selective CB1 agonists. Thus, anandamide appears to cause smooth muscle relaxation via a CB1 receptor-independent mechanism and cannabinoid receptor activation apparently does not contribute to EDHF-mediated relaxation in this resistance artery.

British Journal of Pharmacology (1997) 121, 1509–1511; doi:10.1038/sj.bjp.0701361


Recently, it was shown that endothelium-derived hyperpolarizing factor (EDHF)-induced relaxation of the rat isolated mesenteric bed could be inhibited by the CB1 cannabinoid receptor antagonist SR141716A, and that anandamide, an endogenous cannabinoid, may be an EDHF in mesenteric arteries (Randall et al., 1996). In the present study, the mechanism underlying anandamide-evoked relaxation in rat single, isolated mesenteric arteries was examined and compared to EDHF-mediated relaxation.


Male Wistar or Sprague-Dawley rats (250–350 g,) were stunned and killed by cervical dislocation. Segments (2 mm in length) of third order branches of the superior mesenteric artery (D100 315±12 μm; n=24) were removed and mounted in a Mulvany-Halpern myograph, under normalized tension, for simultaneous recording of smooth muscle membrane potential and tone as previously described (Waldron & Garland, 1994). The tissues were maintained in a static bath and concentration-response curves were constructed by cumulative addition of anandamide to arterial segments preconstricted with phenylephrine (1–3 μM). All experiments were carried out in endothelium-intact arteries plus indome-thacin (2.8 μM) and the nitric oxide (NO) synthase inhibitors NG-nitro-L-arginine and NG-nitro-L-arginine methyl ester (both 100 μM). In the presence of the potassium channel inhibitors, the concentration of phenylephrine was adjusted to ensure a similar level of tone to control conditions. Data are expressed as mean±s.e.mean and the significance of differences between means calculated by Student's t test. All drugs were supplied by Sigma except for anandamide (CalBiochem), 3-(1, 1 -dimethylheptyl)-0-11-hydroxy-Δ8-tetrahydrocannabinol (HU210; Tocris), R(+)-[2, 3-dihydro-5-methyl-3-[(morpholinyl)methyl]pyrrolo [1, 2, 3-de] -1, 4-benz-oxazon-yl] - (1-napthalenyl)methanone mesylate (WIN 55, 212-2; RBI) and N-piperidino-5-(4-chlorophenyl) -1-(2,-dichloophenyl)-4-methyl-3-pyrazole-carb-oxamide (SR141716A; gift from Sanofi). Stock solutions of anandamide, HU210, WIN55, 2212-2 and SR141716A were prepared in ethanol.


The resting membrane potential of smooth muscle cells in the mesenteric artery was–59.3±3.7 mV (n=14 cells from 5 tissues). Anandamide (1-300 nM) caused endothelium-independent repolarization and relaxation of arterial segments precontracted with phenylephrine (mean contraction and depolarization 15.4±3.4 mN (n=16) and 28.0±4.3 mV (n=4); Figure 1a). The maximal repolarization and relaxation to anandamide were 89.0±6.1% (n=4) and 93.0±5.6% (n=26), respectively, with pD2 values of 7.99±0.09 (n=4) and 8.0±0.04 (n=26), respectively. Reproducible repolarization and relaxation required frequent washing for at least 60 min between dose-response curves. Increasing the extracellular potassium concentration to 60 nM abolished relaxation to both anandamide and EDHF (n=6).

Figure 1.

(a) Representative traces showing simultaneous recording of anandamide (1-100nM)-evoked changes in smooth muscle membrane potential and tension in tissues stimulated with phenylephrine, (b and c) Mean concentration-response curves for (b) anandamide-and (c) acetylcholine-evoked relaxation of arterial segments in the presence and absence of potassium channel inhibitors. All values are means of 3 to 6 experiments with s.e.means shown by vertical lines *P<0.01 compared to control values.

The CB1 antagonist SR141716A (5 μM) did not modify anandamide-induced relaxation; (maximum relaxation, 94.1±3.2%; pD2 value, 8.04±0.19 (n=4)). Similarly, SR141716A did not alter EDHF-mediated relaxation to either acetylcholine (maximum, control 97.4± 1.9%; test 94.5± 3.8%; n=5) or A23187 (maximum, control 95.1±1.79%; test 92.2±1.5%; n=5). In addition, HU210 (0.01–3 μM) and WIN 55,212-2 (0.01–3 μM), CB1 agonists structurally unrelated to anandamide, caused maximum relaxant responses of only 24.9±6.8% (n=5) and 0% (n=15), respectively.

Anandamide-induced relaxations were not altered by apamin (50 nM; n=6), but were significantly inhibited by either charybdotoxin (ChTX; 100 nM; n=4; P<0.01) or iberiotoxin (100 and 300 nM; Figure 1b). In contrast, EDHF-mediated relaxation was significantly attenuated by apamin (maximum relaxation reduced to 33±4.7% (n=3; P<0.05)) and, although ChTX alone was without effect (n=4), exposure of arterial segments to a combination of apamin and ChTX abolished EDHF-mediated relaxation and repolarization (n=4; P<0.01; Figure 1c).


The clear correlation between relaxation and repolarization to anandamide, and the sensitivity of these responses to increases in the extracellular potassium concentration, clearly indicate a link between the two parameters. This is very similar to the link between EDHF and smooth muscle relaxation in this vascular bed (Garland & McPherson, 1992). However, relaxation to anandamide appears to be mediated by potassium channels which are different from those opened by EDHF and does not appear to be mediated through activation of CB1 cannabinoid receptors.

Anandamide evoked a concentration-dependent relaxation in depolarized and precontracted arterial segments which was mirrored by repolarization of the smooth muscle cell membrane potential. These responses were abolished by 60 mM extracellular potassium, a possible indication that potassium channel opening is required for relaxation to anandamide. However, 25 mM potassium, which has been shown to abolish EDHF-mediated relaxation in the same vessels (Adeagbo & Triggle, 1993; Waldron & Garland, 1994), only slightly inhibited relaxation to anandamide (data not shown).

Anandamide has been shown to cause relaxation via activation of cannabinoid CB1 receptors (Randall et al., 1996). However, in the present study, the selective CB1 antagonist SR141716A did not alter anandamide-evoked relaxation of segments of mesenteric artery. Also, HU210 and WIN 55,212-2, selective CB1 agonists which are structurally unrelated to anandamide, were either much less effective or ineffective as relaxants. These observations indicate that anandamide evokes arterial repolarization and relaxation independently of CB1 receptors. Anandamide is structurally similar to arachidonic acid, and might therefore interact directly with potassium channels or an associated protein (Clark et al., 1997).

SR141716A had no effect on EDHF-induced relaxations to either acetylcholine or A23187, providing further evidence that cannabinoid receptors are not involved in the actions of EDHF. These observations clearly contrast with SR141716A-induced block of dilatation to both carbachol and A23187 in the rat isolated, perfused mesenteric bed found previously (Randall et al., 1996). However, in that study, around 40–50% and 60–70%, respectively of the carbachol and A23187-induced dilatation persisted in the presence of the antagonist. Thus, as in the present investigation, a large component of the EDHF response appeared to be independent of CB1 receptor activation.

Relaxation to anandamide was unaffected by apamin, an inhibitor of small-conductance calcium-activated potassium channels, but almost completely inhibited by either ChTX or iberiotoxin. This finding suggests that activation of large-conductance calcium-activated potassium channels may underlie relaxation to anandamide. In contrast, ChTX did not affect EDHF-mediated relaxation, whereas apamin caused a significant reduction in the response and a combination of apamin and ChTX together abolished EDHF-evoked relaxation (as previously described in other arteries: Corriu et al., 1996; Zygmunt & Högestatt, 1996).

In conclusion, although anandamide-evoked relaxation in rat isolated mesenteric arteries appears to be associated with repolarization of the smooth muscle membrane potential, the characteristics of this response differ markedly from EDHF-mediated relaxation in these arteries. This suggests that anandamide is not an EDHF in this tissue.


This work was supported by a Wellcome Trust programme grant and an MRC project grant.