Department of Biomedical Sciences, IMS Building, Foresterhill, University of Aberdeen, Aberdeen AB25 2ZD.
1The dose-related inhibition of the twitch responses of the myenteric plexus-longitudinal muscle preparation of the guinea-pig small intestine by cannabinoid (CB) agonists, (+)-WIN 55212 and CP 55940 during stimulation at 0.1 Hz with supramaximal voltage was confirmed. These agonists inhibited acetylcholine (ACh) release in the presence of physostigmine (7.7 μM) thus indicating a prejunctional site of action.
2Inhibition of twitch responses and ACh release by CB agonists was reversed by the CB1-selective cannabinoid receptor antagonist, SR141716A. Dose-response curves to (+)-WIN 55212 and CP 55940 were shifted to the right, with no reduction of maximal response, by pretreatment with SR141716A (31.6–1000 nM), but not its vehicle, Tween 80 (1 μM). However, at very high concentrations (25–400 μM), Tween 80 itself caused a dose-related inhibition of the twitch response which was significantly reduced in the presence of SR141716A (1 μM). The opioid receptor antagonist, naloxone (1 μM) had no significant effect on the inhibition by CP 55940 of the twitch response.
3(+)-WIN 55212, CP 55940 and Tween 80 (50 μM) had no effect on responses to exogenous ACh, confirming that their actions were prejunctional. SR141716A (1 μM) did not increase the sensitivity of the longitudinal muscle to either ACh or histamine, but inhibited the responses to high doses of ACh.
4The (−)-enantiomer of WIN 55212, was approximately 300 times less active than the (+) enantiomer in inhibiting the twitch response, had no CB1 antagonist activity against the active isomer and did not inhibit the release of ACh in the presence of physostigmine.
5The dissociation constant (KD) values for SR 141716A against the inhibitory effect of (+)-WIN 55212 and CP 55940 on the twitch response were 12.07 nM (95% confidence intervals 8.55 and 20.83) and 6.44 nM (95% confidence intervals 4.70 and 10.24), respectively. In experiments in which the release of ACh was inhibited by (+)-WIN 55212, the KD values were 9.21 nM and 10.53 nM at SR141716A concentrations of 31.6 nM and 100 nM, respectively. The KD values for the antagonism by naloxone of the inhibition of the twitch responses and the inhibition of ACh release by normorphine in this preparation were found to be 2.38±0.69 nM and 2.00±0.9 nM, respectively.
6During maximal inhibition of ACh release by (+)-WIN 55212, the addition of normorphine (400 nM) caused a further significant decrease in ACh output.
7SR141716A alone produced a significant increase in ACh release in both the absence and presence of exogenous cannabinoid drugs, hence we conclude that it has a presynaptic site of action. We also conclude that SR141716A acts either by antagonizing the effect of an endogenous CB1 receptor agonist or by having an inverse agonist effect at these receptors.
Together with the mouse isolated vas deferens, the myenteric plexus-longitudinal muscle preparation of the guinea-pig small intestine has been used as a functional in vitro model in which to study the mode of action of psychotropic cannabinoid agonists (Pertwee et al., 1992; 1996b). Such drugs show high potency and stereoselectivity in their inhibition of the evoked contractile responses of these preparations to electrical stimulation, an action which is reversed by SR141716A (Coutts et al., 1995; Pertwee et al., 1996b), a selective antagonist of the cannabinoid CB1 receptor (Rinaldi-Carmona et al., 1994). The mechanism by which cannabinoids depress the electrically-evoked contractions of the longitudinal muscle of the small intestine is deemed to be presynaptic since concentrations of Δ9-tetrahydrocannabinol which depress the twitch have no effect on contractile responses to exogenous acetylcholine (ACh) (Gill et al., 1970; Layman & Milton, 1971; Roth, 1978), and depress the spontaneous release of ACh from segments of whole ileum (Layman & Milton, 1971) or innervated ileal longitudinal muscle (Paton et al., 1972).
The present experiments were undertaken to investigate whether changes in evoked responses of the innervated longitudinal muscle strip due to cannabinoid agonists were accompanied by parallel changes in evoked release of ACh from the myenteric plexus. Similar experiments were performed to determine whether the antagonism of these CB1 effects by the selective antagonist, SR141716A, was also due to corresponding changes in ACh release. In the absence of exogenous cannabinoid drugs, it was observed (Coutts et al., 1995; Pertwee et al., 1996b) that this antagonist was able to increase the contractile responses to electrical stimulation. In these experiments we have sought to demonstrate corresponding changes in ACh release. We have also been able to substantiate earlier findings of lack of postjunctional effects of cannabinoid agonists and extend these to include the effects of SR141716A, and of high concentrations of the vehicle, Tween 80, which are also inhibitory in this preparation (Pertwee et al., 1996b). Further demonstration of the receptor-mediated mode of action of cannabinoids was sought by comparing the effects of enantiomers of the agonist WIN 55212 on both electrically-evoked twitches and ACh release.
The presynaptic action of morphine and opioid drugs, acting through μ opioid receptors and clonidine, acting via α2-adrenoceptors in the myenteric plexus-longitudinal muscle preparation of the guinea-pig small intestine is well documented. These drugs reduce the twitch responses to low frequency stimulation by reducing the release of ACh (Greenberg et al., 1970). It was important, therefore, to differentiate between the site of action of cannabinoid drugs and μ and α2-adrenoceptors. We have already shown (Pertwee et al., 1996b) that dose-response curves for the inhibitory effect of normorphine and clonidine on evoked contractile responses are unaffected by the CB1 antagonist SR141716A. Further experiments were designed to investigate interactions between the inhibitory effects of opioid and cannabinoid drugs on the evoked contractile responses and ACh release from this preparation. The evoked release of ACh is highly dependent on the frequently of stimulation (Cowie et al., 1968). Previous studies have shown that inhibitory drugs which act presynaptically in this tissue are more effective at low than at high frequencies of stimulation (Greenberg et al., 1970; Kilbinger, 1977). Therefore in all the experiments described in the present study the frequency of stimulation was 0.1 Hz.
Segments of small intestine were dissected from male albino Dunkin-Hartley guinea-pigs (282-820 g) and set up for field stimulation by the method of Paton & Zar (1968). For the recording of contractile responses to electrical stimulation, tissues were mounted in 4 ml organ baths under an initial tension of 0.5 g with the method described by Pertwee et al. (1996b). The baths contained Krebs solution which was kept at 37°C and bubbled with 95% O2 and 5% CO2. The composition of the Krebs solution was (mM): NaCl 118.2, KCl 4.75, KH2PO4 1.19, NaHCO3 25.0, CaCl2.6H2O 2.54, MgSO4.7H2O 1.19 and glucose 11.0. Isometric contractions were evoked by continuous stimulation at 0.1 Hz with pulses of 110% maximal voltage and 0.5 ms pulse duration from Grass 88 stimulators through platinum electrodes attached to the upper and lower ends of each bath. Contractions were recorded via Dynanometer UFl transducers (Ether) linked to a pen oscillograph (Grass Polygraph 7D). For cumulative concentration-response curves for the effect of cannabinoid agonists on the twitch response, a 20 min interval was left between consecutive additions of drug.
Acetylcholine release in the presence of physostigmine (7.7 μM) was measured by bioassay on a myenteric plexus-longitudinal muscle preparation with a method modified from Waterfield (1973). Stimulation period was 4 min, with 10-20 min between consecutive stimulation periods, depending on whether the experimental conditions had been altered e.g. by increasing the drug concentration or by introducing a new drug. In release experiments, each preparation was subjected to all the experimental conditions described for that experiment, except for experiments for the determination of KD values in which different preparations were used for each concentration of SR141716A. Therefore control release and that in the presence of drugs were determined in the same tissue. The effects of drugs on resting release of ACh were not determined. The values of ACh release in each experiment was the mean of triplicate samples which were collected under the same experimental conditions. Experiments showed that none of the drugs tested for an effect on output in the course of these experiments had any effect on the sensitivity of these preparations to ACh (see Results section), therefore these drugs were omitted from standard solutions of ACh. When the effects of drugs on ACh release were being measured, after determination of control output, preparations were exposed to the new drug environment for 20 min before the start of collection of samples. The responses to both unknown samples and standard solutions of ACh were abolished in the presence of atropine (100 nM). The output of ACh was calculated in pmol g−1 per pulse.
Dose-response curves for ACh and histamine were performed with a 2 min dose cycle, the bath fluid being renewed by overflow after each contraction had reached maximum. After control dose-response curves had been determined, preparations were preincubated with cannabinoid drugs for 20 min before the ACh or histamine dose-response curves were constructed. Cannabinoid drugs were replaced in the organ bath after each washout.
Acetylcholine chloride was obtained from BDH and choline chloride, clonidine HCl, physostigmine sulphate, histamine acid phosphate, naloxone HCl and Tween 80 from Sigma. (+)-WIN 55212 (mesylate (R)-(+) -[2,3-dihydro-5-methyl-3-[4-morpholino)methyl]pyrrolo-[1,2,3-de]-1,4-benzoxazin-6-yl] (1-naphthyl)methanone) was supplied by Sterling Winthrop and (–)-WIN 55212 by Research Biochemicals International. CP 55940 ((–)-3-[2-hydroxy-4-(1,1-dimethyl heptyl)phenyl]-4-(3-hydroxy propyl)cyclohexan-1-ol) was obtained from Pfizer and normorphine from Burroughs Wellcome. SR 141716A (N-(piperidin-1-yl)-5-(-4-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide hydrochloride) was a gift from Sanofi. The WIN 55212 isomers, CP 55940 and SR 141716A were each mixed with 2 parts of Tween 80 by weight and dispersed in a 0.9% aqueous solution of NaCl (saline) before addition to the organ bath. Other drugs were dissolved in saline. It was not possible to reverse the inhibitory effect of cannabinoids on either the twitch response or on ACh release by washing them out of the organ bath, therefore cumulative log dose-response curves were constructed. In control experiments, Tween 80 was added instead of cannabinoid receptor agonists or SR 141716A in concentrations corresponding to those required to disperse these agents. Since Tween 80 is a mixture of organic acids and its concentration cannot be expressed in molar terms, in this paper the ‘molarity’ given for Tween 80 refers to the molarity of the agent for which it was the vehicle control.
Analysis of data
Values are expressed as means and limits of error as s.e.mean (mean±s.e.mean). Inhibition of the electrically-evoked twitch response is expressed in percentage terms and has been calculated by comparing the amplitude of the twitch response after each addition of a twitch inhibitor with its amplitude immediately before the first addition of the inhibitor. Levels of significance were calculated between two sets of results by paired or unpaired Student's t test as appropriate. Comparison of more than two sets of values was carried out by analysis of variance followed by Scheffé's test (Super ANOVA, Abacus Concepts Inc., Berkeley). A P value < 0.05 was considered to be significant.
Dose-ratios have been determined by symmetrical (2+2) dose parallel line assays (Colquhoun, 1971) with pairs of agonist concentrations located on the steepest part of each log concentration-response curve. The limits of error are given as 95% confidence intervals.
The method of determination of dissociation constant (KD) has been described previously (Pertwee et al., 1996b). KD values were calculated from the slope (1/KD) of the best fit straight line of the plot of (x–1) against B, constrained to pass through the origin (Tallarida et al., 1979) and calculated by linear regression analysis with GraphPad InPlot (GraphPad Software, San Diego). The equation for this graph is (x–1) = B/KD, where × (dose-ratio) is the ratio of concentrations of agonist which produce the same degree of inhibition in the presence to that in the absence of a given concentration, B, of antagonist. The value of the Schild slope for the interaction between SR141716A and WIN 55212 or CP 55940 was obtained from the best fit straight line of a plot of log (x–1) against –log B. The equation for this graph, log (x–1) = (log B)–log KD, predicts a slope of 1 for all receptor-mediated interactions between agonists and antagonists that are competitive and reversible. Values for the Schild slope have been calculated by linear regression analysis by use of GraphPad InPlot.
Effects of cannabinoid agonists on the twitch response
In the present experiments the cannabinoid agonists (+)-WIN 55212 and CP 55940 produced dose-related inhibition of evoked twitch responses. Dose-response curves for these drugs were unaffected by the opioid receptor antagonist, naloxone (1 μM) as illustrated for CP 55940 in Figure 1a. The results for the effect of naloxone on the inhibition by (+)-WIN 55212 are not shown. The vehicle for SR141716A and for cannabinoid agonists, Tween 80, in a concentration equivalent to that used to disperse 1000 nM cannabinoid drug, had no significant effect on dose-response curves to cannabinoid agonists as shown for (+)-WIN 55212 (Figure 1b). However, these curves were shifted in parallel to the right when the CB1-selective cannabinoid receptor antagonist, SR141716A (100 nM), was present in the organ bath (n=8) (Figure 2) producing a dose-ratio of 30.54. In the pooled results from 27 experiments, (+)-WIN 55212 (135 nM) caused 67.33±2.19% inhibition of the twitch which was reversed in a dose-related fashion by the addition of increasing concentrations of SR141716A (Figure 3a).
At high concentrations (25-400 μM), which are well in excess of those required to disperse cannabinoid drugs, Tween 80 caused a dose-dependent inhibition of the twitch response which was significantly reduced in the presence of the cannabinoid antagonist SR141716A (1 μM) (Figure 3b).
The effects of cannabinoid drugs on the sensitivity to ACh and histamine
The responses to ACh added to the organ bath were unaffected by the presence of (+)-WIN 55212 (200 nM) (Figure 4a), a concentration which causes near maximal inhibition of the evoked contractile response to electrical stimulation. A similar lack of effect was found with CP 55940 (200 nM) in another 6 preparations (not shown). Tween 80 (50 μM) also had no effect on the dose-response curves to ACh (Figure 4b).
In the presence of SR 141716A (1 μM), the responses to relatively low doses (< 1.4 μM) of ACh were not significantly affected compared with controls (Figure 4c). However, the responses to high doses of ACh (1.4 and 4.14 μM) were significantly reduced by 20.5% and 26.2%, respectively. The concentration-response curve to histamine was unaffected by SR141716A (1 μM) (Figure 4d).
The inhibitory effects of cannabinoid agonists on the evoked release of ACh from the myenteric plexus and reversal with SR141716A
CP 55940 and (+)-WIN 55212 inhibited the release of ACh from this preparation in a concentration-dependent manner. In a concentration of 200 nM, which caused near maximal inhibition of the twitch response in the absence of physostig-mine (Figure 1a), CP 55940 reduced the ACh release from this preparation by 34.7% from 144.1±26.2 to 90.0±15.9 pmol g–1 per pulse (P<0.005; paired t test, n=14). Tween 80 (1 μM) had no significant effect on ACh release (see below). The inhibition due to CP 55940 was reversed by the additional presence of SR141716A (1 μM) (Figure 5a). That these changes were not artifacts due to time was demonstrated by the failure of the output to alter significantly in the presence of Tween 80 (equivalent to 1 μM SR141716A) (Figure 5b). (+)-WIN 55212 was also shown to inhibit the release of ACh during stimulation at 0.1 Hz (Figure 2c). This effect was dose related, was maximal at a concentration of 200 nM and was shifted in parallel to the right in the presence of the CB1, antagonist SR141716A (Figure 2b). In 4 experiments, (+)-WIN 55212 (200 nM) inhibited the ACh output by 35.1±4.19% from 105.5±22.7 to 65.7±10.6 pmol g–1 per pulse (P=0.01; paired t test). This inhibitory response was not increased further when the concentration of (+)-WIN 55212 was doubled, but was significantly reversed (P < 0.05; paired t test) in the presence of 141716A (1 μM) when the output increased to 94.7±17.5 pmol g−1 per pulse.
Dissociation constant of SR141716A at CB1, receptors as measured in the myenteric plexus
The dissociation constant was determined for the inhibition of the twitch response by CP55940 and its antagonism by SR141716A. Dose-response curves to CP55940 were constructed in the absence of any antagonist and after pretreatment of preparations with 31.6, 100, 316 or 1000 nM SR141716A. These concentrations of antagonist produced dose-ratios of 5.17, 27.81, 33.15 and 80.04, respectively. Experiments carried out with (+)-WIN 55212 as agonist on the twitch response and SR141716A concentrations of 10, 31.6 and 100 nM produced dose-ratios of 1.93, 4.12 and 17.14, respectively. KD values of 12.07 nM and 6.45 nM were obtained for the antagonism of CP 55940 and (+)-WIN 55212, respectively (Table 1). A similar series of experiments was carried out for the inhibition of ACh release in the presence of physostigmine by (+)-WIN 55212 and its antagonism by SR141716A at concentrations of 31.6 nM, 100 nM and 400 nM. The resulting dose-ratios were 4.43±1.94 (n=3), 10.5±4.00 (n=4) and 9.55±3.52 (n=5), respectively. As the Schild plot (Figure 6) generated from the release data gave a slope significantly less than 1, it would appear that at high concentrations SR141716A may not be acting in a simple competitive manner on the release of ACh. Therefore the KD values given in Table 1 were calculated from the dose-ratios obtained with 31.6 nM and 100 nM only.
Table 1. Dissociation constants (KD) of SR141716A determined in the presence of CP 55940 AND (+)-WIN 55212 on both the contractile response and the release of ACh from the myenteric plexus-longitudinal muscle preparation
KD values were determined by the Tallarida method (mean values and 95% confidence intervals) with at least three of the following concentrations of SR141716A: 10, 31.6, 100, 316.2 and 1000 nM. Slope is that of a Schild plot. P values indicate the levels of significant difference from a slope of unity.
* Determined on contractile response. (n=5–8 at each concentration of SR141716A).
** Determined on release of ACh (n=3 or 4 at each concentration of SR141716A).
τ values calculated from mean (DR–1) at SR141716A concentrations of 31.6 nM and 100 nM, respectively.
8.55 and 20.83
4.70 and 10.24
Dissociation constant of opioid receptors for naloxone
The dissociation constant for the effect of naxolone in antagonizing the inhibitory effects of normorphine on the twitch response was compared with that for its antagonism of the inhibition due to normorphine of the release of ACh. In this series of experiments, cumulative dose-response curves for the inhibitory effect of normorphine (50–3200 nM) on the twitch response were constructed before and after the addition of naloxone (final bath concentration 10 nM) to the organ bath. The maximum inhibition of the twitch was 70.15±9.58% and 68.7±6.04% (n=5) in the absence and presence of naloxone and the corresponding IC50 values were 89.0 nM and 652.3 nM, respectively. In these experiments, naloxone produced a mean dose-ratio of 6.91±1.61. The mean value for the dissociation constant determined from these data was 2.38±0.69 nM. The dissociation constant for the antagonism by naloxone (10 nM) of the inhibition of ACh release from the myenteric plexus by normorphine (50–4050 nM) was found to be 2.00±0.90 nM (mean± s.e.mean for 3 experiments). The maximum inhibition due to normorphine on ACh release was 53.1±4.7% in the absence and 53.9±3.0% in the presence of naloxone, and the mean dose-ratio was 7.93 ±2.16. When the maximum inhibition of output in each experiment was taken as 100% maximum inhibition, the mean IC50 values for the inhibitory effect of normorphine on ACh release in the absence and presence of naloxone were 158.3 nM and 1021 nM, respectively.
Comparison of the responses to (+)- and (–)-isomers of the cannabinoid agonist WIN 55212
Dose-response curves for the (–)-isomer of WIN 55212 in inhibiting the twitch response were compared with those for the (+)-isomer. The responses to Tween 80, in concentrations equivalent to those used to disperse (–)-WIN 55212 are also shown (Figure 7a). (–)-WIN 55212 was approximately 300 times less potent than (+)-WIN 55212 on the twitch response, though slightly more active than the vehicle alone. In 4 experiments, dose-response curves to (+)-WIN 55212 were constructed in the presence of (–)-WIN 55212 (1 μM) and were not different from control curves (Figure 7b). These results indicate that (–)-WIN 55212 is also devoid of cannabinoid receptor antagonist activity.
When the effects of (–)-WIN 55212 were measured on the release of ACh from the myenteric plexus, in sequential treatments of the same preparations, the output of ACh was not changed in the presence of (–)-WIN 55212 (1 μM) or its vehicle, Tween 80, but was significantly reduced by subsequent exposure of the preparation to the active isomer, (+)-WIN 55212 (200 nM) (Figure 7c).
Effects ofSR141716A on the evoked release of ACh in the absence of exogenous cannabinoids
In 9 experiments, the output of ACh in the presence of Tween 80 (1 μM) was not significantly different from the control output. However, subsequent addition of SR141716A, giving a final bath concentration of 1 μM, resulted in the ACh output being increased by 25.34±9.59% (Figure 8a) compared with the control output (P < 0.05).
A rise in ACh release due to SR141716A could also be demonstrated in the presence of concentrations of normorphine (100 nM) which produced significant inhibition of ACh release (data not shown; n=5).
The effect of normorphine on the evoked release of ACh in the presence of maximal activation of cannabinoid receptors
In 3 experiments, the output of ACh was determined before and after treatment of preparations with a supramaximal dose of (+)-WIN 55212 (1215 nM) and again after subsequent exposure of the preparations to normorphine (400 nM) in the presence of the cannabinoid. (+)-WIN 55212 alone significantly inhibited the output of ACh by 35.6±0.94%. The output was further inhibited to produce a final inhibition of 65.58±0.59% on exposure to normorphine (Figure 8b, P < 0.005; paired t test).
That cannabinoid agonists act presynaptically by reducing ACh release in the myenteric plexus-longitudinal muscle preparation is now quite clear. The present experiments show that, in doses which inhibit the evoked contractile responses at low frequency, these agonists produce a corresponding reduction in ACh release. These results support earlier findings for the effects of Δ9-tetrahydrocannabinol on resting ileal preparations (Layman & Milton, 1971; Paton et al., 1972). In all experiments in which the release of ACh was determined, the rate of release was not significantly changed by time or by incubation with the vehicle, Tween 80 alone. Similar presynaptic CB1-receptor activation has been shown to inhibit the release of [14C]-ACh from hippocampal slices (Gifford & Ashby, 1996), [3H]-noradrenaline from peripheral sympathetic nerves (Ishac et al., 1996) and evoked [3H]-noradrenaline and [3H]-dopamine from guinea-pig superfused retinal discs (Schlicker et al., 1996).
Antagonism of the effects of cannabinoid agonists on the contractile response or on the evoked release of ACh is demonstrated by the parallel, rightward shift of the agonist dose-response curves by pretreatment of the preparation with SR141716A, the CB1 selective antagonist. The shifts of the dose-response curves to (+)-WIN 55212 and CP 55940 on twitch responses are in line with those previously published (Pertwee et al., 1996b) as are the resulting values for dissociation constants. The exception to this is the failure of the highest dose of SR141716A to cause a further parallel rightward shift of the inhibitory dose-response curve for the effect of (+)-WIN 55212 on the release of acetylcholine. Consequently, the slope for the Schild plot appears to be less than unity. This suggests that the effects of (+)-WIN 55212 in this preparation may not be totally explained in terms of a straightforward competitive surmountable interaction with CB1 receptors and care must be taken in the interpretation of these findings. It is possible that part of the response to higher concentrations of (+)-WIN 55212, which are required to produce inhibition in the presence of 400 nM SR141716A, may be mediated via a different population of receptors which are poorly antagonized by SR141716A e.g. CB2 receptors, as (+)-WIN 55212 has been shown to bind preferentially to CB2 rather than CB1, receptors (Showalter et al., 1996).
In experiments in which KD values were determined on the contractile response, the responses in the presence of cannabinoid agonists were compared with those before the addition of the agonist but after equilibration with SR141716A. These responses were not the same as those of the naive preparation and this poses some difficulty. However, when determined on the twitch responses, the KD values (see Table 1) generated from this analysis gave apparent pA2 values of 8.19 for (+)-WIN 55212-2 and 7.94 for CP 55940 which compare favourably with the pA2 values found for the antagonism of WIN 55212-2 by SR141716A in the mouse isolated vas deferens (Rinaldi-Carmona et al., 1994) and cannabinoid-mediated inhibition of dopamine release in the retina (Schlicker et al., 1996). The KD value of 2.38 nM (pA2 of 8.62) found for the antagonism of normorphine by naloxone agrees well with that previously published in the literature (Lord et al., 1977).
In addition, none of the cannabinoid drugs tested showed postjunctional activity, apart from SR141716A which reduced the responses to very high doses of ACh. Since this drug increases electrically-evoked contractile responses, any effect which reduces smooth muscle contractions is unlikely to be related to its effects on the contractile responses to electrical stimulation.
Unlike the twitch response, which can be virtually abolished by CB agonists, the inhibition of ACh release reaches a plateau at 35-40% inhibition. There are several possible explanations for this discrepancy. Firstly, the abolition of the twitch response may leave a residual release of ACh from the nerve terminals which is subthreshold for excitation of the smooth muscle in the absence of physostigmine but is still detectable by bioassay when physostigmine is present. Secondly, ACh is released from sites in the myenteric plexus other than from the terminal varicosities of the final motorneurones, for example from the preganglionic terminals in the myenteric ganglia. These sources may provide a significant contribution to the ACh measured in the presence of physostigmine and may be insensitive to CB1 receptor activation. Thirdly, the inhibitory effects of CB agonists on ACh release may be partially precluded by the autoinhibition of acetylcholine release in the presence of anticholinesterases (Ten Berge et al., 1996). Finally, there is electrophysiological evidence (North & Nishi, 1974) that after pretreatment with physostigmine (10 μM), fast excitatory postsynaptic potentials recorded in S neurones of guinea-pig myenteric ganglia are augmented in amplitude and the duration prolonged, resulting in repetitive firing. This may lead to further release of ACh which does not occur in the absence of physostigmine and is not sensitive to inhibition by cannabinoids.
The ability of normophine to reduce significantly the residual release of ACh after maximal activation of cannabinoid receptors confirms the distinction between opioid and cannabinoid receptors. This is supported by the finding that dose-response curves to cannabinoid agonists are unaffected by the presence of naloxone in the organ bath, and that dose-response curves to the inhibitory effects of normorphine and also clonidine are not antagonized by SR141716A (Pertwee et al., 1996b).
The difference in potencies of the stereoisomers of WIN 55212 on the contractile responses and the complementary effects on ACh release is consistent with a specific receptormediated mechanism of action of cannabinoids.
Occasionally, the control of ACh output from myenteric plexus-longitudinal muscle preparations is found to be considerably lower than normal despite the sex, weight range and experimental conditions being identical to other experiments (see Figure 5). In terms of sensitivity to cannabinoid drugs, these preparations behave in the same way as those showing ‘normal’ levels of ACh release. The reason for this observed difference in ACh output remains obscure.
SR141716A has previously been shown to produce a slight, though significant and dose-related increase in twitch tension when added to a naive preparation during stimulation at 0.1 Hz compared with responses in the presence of the equivalent concentration of Tween 80 (Coutts et al., 1995; Pertwee et al., 1996b). The augmenting action of SR141716A on the twitch response correlates well with the significant increase in ACh release when acetylcholinesterase has been completely inhibited, therefore the increased twitch induced by SR141716A in the absence of a cholinesterase inhibitor is unlikely to result from a reduction in ACh metabolism. Pretreatment of the preparation with normorphine (100 nM) did not prevent this augmenting effect which appeared to result in a physiological partial reversal of the inhibitory response to normorphine on both the twitch response and on ACh release.
This reverse-agonist effect of SR141716A alone is not without precedent. It has been shown that treatment with SR141716A can increase forskolin-induced stimulation of cyclic AMP production in cells transfected with cannabinoid receptors (Felder et al., 1995), decrease spontaneous firing of rat substantia nigra pars reticulata neurones (Tersigni & Rosenberg, 1996), increase locomotor activity in mice (Compton et al., 1996) and produce signs of increased arousal in rats (Santucci et al., 1996). In more comparable experimental situations, SR141716A has been found to augment electrically evoked twitch responses of mouse isolated vas deferens (Pertwee et al., 1996a) and urinary bladder (Pertwee & Fernando, 1996) and facilitate the release of radiolabeled neurotransmitter from superfused retinal discs (Schlicker et al., 1996) and hippocampal slices (Gifford & Ashby, 1996). It may be postulated that the increase in twitch response and evoked release by ACh by SR141716A is due to antagonism of an endogenously-released CB1 receptor agonist. Endogenous cannabinoid ligands which have been found include arachidonoylethanolamide (anandamide) and 2-arachidonoylglycerol. Such ligands have been detected in brain (Devane et al., 1992; Sugiura et al., 1995) and the latter has also been found in the canine intestinal tract (Mechoulam et al., 1995). However, these ligands have not been identified in the myenteric plexus-longitudinal muscle preparation of the guinea-pig to date.
Alternatively, the augmenting effect of SR141716A in this tissue may be explained in terms of the 2-state model of receptors such as proposed by Leff (1995). This assumes that the receptor can exist in two states, one active, which can couple to the signal transducer mechanism, and one inactive, which cannot. According to this hypothesis (see Pertwee et al., 1996b), the receptor states coexist in equilibrium, even in the absence of any agonist. Preferential binding of SR141716A to the receptor in the inactive state would be expected to produce a shift in equilibrium away from the active state and hence a reduction in the number of precoupled receptors and an attenuation in inhibitory activity. If this hypothesis were true, then SR141716A would act as an inverse agonist rather than a pure antagonist. Gifford & Ashby (1996) arrived at similar conclusions in a recent study in which SR141716A was shown to enhance the electrically-evoked release of [14C]-ACh from hippocampal slices of rat brain.
In conclusion, the findings of the present study, taken together with the evidence of CB receptor localization on nerve terminals (Herkenham, 1995) and the effects of CB1 receptor activation on transmitter release in the central nervous system (Gifford & Ashby, 1996) and other peripheral neural tissues (Ishac et al., 1996; Schlicker et al., 1996; Pertwee et al., 1996b), reinforce the hypothesis that cannabinoids perform a neuro-modulatory role. Since SR1417167A is such a widely used and invaluable tool in the characterization of cannabinoid receptors in both the central and peripheral nervous system and in cultured cells, it is essential that its mechanism of action is clarified in the near future in order to facilitate the interpretation of experimental findings.
This work was supported by grant 043325 from the Wellcome Trust.