Mechanism of the Hypotensive Action of Anandamide in Anesthetized Rats
We studied the effects of the endogenous cannabinoid ligand anandamide on blood pressure, single unit activity of barosensitive neurons in the rostral ventrolateral medulla, and postganglionic splanchnic sympathetic nerve discharge in urethane-anesthetized rats. In rats with an intact baroreflex, an intravenous bolus of 4 mg/kg anandamide caused a triphasic blood pressure response: transient hypotension, followed by a brief pressor and more prolonged depressor phase. Anandamide evoked a “primary” increase in neuronal firing coincident with its pressor effect and a “secondary,” baroreflex-mediated rise coincident with its depressor effect at both sites. Pretreatment of rats with phentolamine or trimethaphan did not inhibit either the pressor response or the primary increase in splanchnic nerve discharge elicited by anandamide. In barodenervated rats, electrical stimulation of the rostral ventrolateral medulla increased blood pressure and splanchnic nerve discharge. Anandamide treatment blunted the rise in blood pressure without affecting the increase in splanchnic nerve discharge. Anandamide did not affect the rise in blood pressure in response to an intravenous bolus dose of phenylephrine. The results indicate that (1) the brief pressor response to anandamide is not sympathetically mediated, and (2) the prolonged hypotensive response to anandamide is not initiated in the central nervous system, in ganglia, or at postsynaptic adrenergic receptors but is due to a presynaptic action that inhibits norepinephrine release from sympathetic nerve terminals in the heart and vasculature.
The widespread use of marijuana as a recreational drug has naturally focused attention on the psychoactive properties of Δ9-THC and related cannabinoids. However, Δ9-THC also potently affects cardiovascular variables. In the early 1970s, several studies documented the hypotensive property of Δ9-THC in experimental animals,1 2 3 4 5 and hypotension in response to Δ9-THC has also been reported in humans.6 A key to our understanding of how cannabinoids lower BP was the observation that cervical spinal cord transection abolished the hypotensive action of Δ9-THC even though normal BP levels were maintained by chronic infusion of norepinephrine.5 This suggested that cannabinoids lower BP by interfering with sympathetic outflow to the vasculature and heart, although it remained unclear whether this inhibition occurs in the central nervous system, in sympathetic ganglia, or at the level of sympathetic nerve terminals. A possible central mechanism was supported by the reported decrease in the firing rate of cardiac postganglionic sympathetic neurons in response to Δ9-THC.5 However, in the same study, intracerebroventricular Δ9-THC did not lower BP. A central mechanism of action was also proposed based on the results of cross-circulation experiments in dogs, in which Δ9-THC lowered BP when administered into the cerebral vasculature of recipient dogs whose head was vascularly isolated from the rest of the body.3 However, in the same dogs, Δ9-THC caused similar hypotension when injected into the systemic circulation of the recipient dog, which indicates a peripheral site of action. In both of the above studies, the vasoconstrictor response to exogenous norepinephrine was maintained or even increased after Δ9-THC, indicating that the sympathoinhibitory effect is not due to decreased postsynaptic sensitivity to the neurotransmitter. Thus, the question of whether the hypotensive action of cannabinoids is due to a central sympathoinhibitory action or an action at peripheral sympathetic nerve terminals remains unsettled.
Interest in the biological actions of cannabinoids has been rekindled by the recent identification of specific cannabinoid receptors, the CB1 receptor expressed predominantly in brain7 and the CB2 receptor expressed in lymphoid tissue,8 and by the discovery of an endogenous cannabinoid ligand, anandamide, which binds to cannabinoid receptors and shares many of the neurobehavioral effects of Δ9-THC.9 Not surprisingly, anandamide was also found to have cardiovascular effects similar to those of Δ9-THC. In anesthetized rats, both Δ9-THC and anandamide cause a brief pressor and more prolonged depressor response, and a novel antagonist selective for the CB1 receptor10 was able to inhibit the depressor but not the pressor response to both agents.11 Furthermore, the prolonged hypotensive component in the effect of anandamide was inhibited by cervical cord transection or by α-adrenergic receptor blockade despite the maintained vasodilator capacity of peripheral blood vessels.11 This suggests that similar to Δ9-THC, the hypotensive response to anandamide is due to inhibition of sympathetic tone. We designed the present study to distinguish between possible central and peripheral sites of action for this effect. The results are incompatible with an inhibitory action of anandamide on central neurons generating sympathetic tone, on ganglionic transmission, or on postsynaptic vascular responses to α-receptor stimulation. Rather, they suggest that anandamide lowers BP by inhibiting transmitter release from postganglionic sympathetic nerve terminals in the heart and vasculature.
Thirty male Sprague-Dawley rats weighing 250 to 300 g were used. The rats were tracheotomized after ether induction and mechanically ventilated with 100% oxygen throughout the experiment. Surgery was done with rats under halothane (1.3% to 15%) anesthesia. A femoral artery and vein were cannulated for arterial BP monitoring and drug administration, respectively. After surgery, halothane was discontinued and replaced by 1.2 g/kg urethane IV for maintenance of higher resting BP. Muscle relaxation was achieved by administration of 1 mg/kg pancuronium IV, followed by supplemental doses of 0.3 mg/kg as required. The temperature of the rats was maintained at 37° to 38°C by an electrical heating blanket and rectal probe.
Baroreceptor denervation was achieved by bilateral sectioning of the aortic depressor, carotid sinus, glossopharyngeal, superior laryngeal, and cervical vagus nerves. Denervation was regarded to be complete if the intravenous bolus injection of phenylephrine did not evoke any change in postganglionic sympathetic nerve activity.
Recording of Postganglionic Sympathetic Nerve Activity
Rats were mounted prone in a stereotaxic frame (David Kopf Instruments) with the bite bar 3.5 mm below the interaural line. The left splanchnic nerve was isolated through a retroperitoneal approach. The segment distal to the suprarenal ganglion (postganglionic) was placed on a bipolar silver hook electrode for recording, and the area was covered with Wacker Sil-Gel. sSND was recorded with a 100- to 3000-Hz band-pass filter and a 60-Hz notch filter and stored on an FM tape recorder for analysis. The zero level of sSND was taken as the residual noise after administration of the ganglionic blocker trimethaphan camsylate (30 mg/kg IV) at the end of the experiment. sSND was normalized by defining the mean resting level of sSND at the beginning of the experiment as 100 sSND units. The postganglionic nature of the recorded neuronal activity was verified by the disappearance of electrical activity after administration of 30 mg/kg trimethaphan camsylate IV.
For electrical stimulation of the RVLM, a fine-tipped monopolar metal electrode (1 to 2 MΩ impedance) was lowered into the medulla oblongata with the tip located at 3.2 to 3.8 mm posterior to the atlanto-occipital suture, 1.7 mm lateral from midline, and 8.7 to 8.9 mm ventral from the cerebellar surface with the bite bar set at −3.5 mm. Stimulation parameters were 1- to 10-second trains at 40 Hz, 0.5-millisecond pulse duration, and 200 μA intensity. The localization of the electrode was verified by the increase in sSND observed on stimulation.
Recording From Barosensitive Neurons in the RVLM
Recording from barosensitive neurons in the RVLM was done in rats that had not been previously debuffered. A preliminary mapping of the location of the facial motor nucleus was used as a guide for directing the single-barreled glass recording electrodes to the RVLM.12 Within the RVLM, presympathetic vasomotor neurons were identified by their barosensitivity and the absence of respiratory phase locking. Baroreceptor activation was achieved by raising the upper body BP with an aortic cuff.
Anandamide (arachidonyl ethanolamide) was synthesized and kindly provided by Dr Raj Razdan. It was dissolved at 10 mg/mL in a vehicle containing ethanol/emulphor/saline (1:1:18). Injection of up to 0.5 mL of this vehicle caused no change in BP or sSND. Urethane, phenylephrine HCl, and phentolamine HCl were from Sigma Chemical Co, trimethaphan camsylate from Roche Laboratories, and pancuronium bromide from Astra USA.
Data are expressed as mean±SE. The effects of RVLM stimulation on BP and splanchnic nerve discharge were assessed by ANOVA followed by Tukey's post hoc test.
Effects of Anandamide on BP, sSND, and SUA in RVLM
The effects of a bolus intravenous dose of 4 mg/kg anandamide on arterial BP, postganglionic sympathetic nerve activity, and SUA of barosensitive neurons in the RVLM were tested in baroreceptor-intact rats. The results from a typical experiment are illustrated in Fig 1⇓. The aortic cuff was first briefly occluded for verification of the baroreflex-induced inhibition of SUA in the RVLM and the parallel inhibition of sSND. Anandamide (4 mg/kg IV) was then administered as a bolus and elicited the typical triphasic BP response documented earlier11 : transient hypotension (phase I), followed by a brief pressor (phase II) and more prolonged depressor phase (phase III). SUA in a barosensitive neuron in the RVLM showed a transient sharp rise preceding the pressor phase, followed by a more prolonged rise during the phase III hypotension; there were parallel changes in sSND. Similar changes were noted in three additional experiments: SUA increased by +103±28% during the pressor phase and by +92±25% during phase III hypotension, and the corresponding changes in sSND were +53±13% and +24±7%, respectively.
Effects of Anandamide on BP and sSND in Barodenervated Rats
To distinguish between the primary effects of anandamide and secondary, baroreceptor reflex–mediated changes, we analyzed the effects of anandamide in barodenervated rats. First, we tested the effects of electrical stimulation of the RVLM. Electrical stimulation caused sharp, brief increases in BP and sSND that were reproducible on repeated stimulation (Fig 2⇓). Anandamide (4 mg/kg IV) was then administered and elicited a brief pressor response (phase II) followed by a more prolonged depressor phase (phase III). The absence of the transient hypotension (phase I) in these barodenervated rats was expected, because earlier studies demonstrated that this initial component, which is associated with profound bradycardia, is vagally mediated and can be eliminated by methyl-atropine or cervical vagotomy.11 sSND showed the transient increase preceding the pressor response, but the secondary, more prolonged increase was absent (Fig 2⇓). Note, however, that there is no indication of a decrease in sSND during the depressor phase, and similar observations were made in five additional rats, in which the mean change in sSND was +3±2 U (P>.4, basal=100 U). Electrical stimulation of the RVLM was repeated during and after the depressor response to anandamide. When applied during the hypotensive phase, RVLM stimulation caused a much smaller increase in BP than under control conditions, whereas the stimulation-induced increase in sSND remained unchanged. The inhibition of the pressor response to RVLM stimulation by anandamide was reversible, as indicated by the increase in pressor responses as BP returned toward control levels. These experiments therefore indicate that anandamide inhibits the pressor response to RVLM stimulation without inhibiting sSND, as also illustrated by the cumulative responses recorded in six experiments (Fig 3⇓). The reversibility of the anandamide-induced inhibition of the pressor response to RVLM stimulation is illustrated in Fig 4⇓.
To test whether anandamide influences target-organ sensitivity to sympathetic stimuli, in five experiments we tested the pressor response to a bolus dose of 10 μg/kg phenylephrine IV. Phenylephrine increased mean arterial BP by 27±8 mm Hg under control conditions and by 32±9 mm Hg when administered during the hypotensive response to anandamide (P>.2). We also examined whether the anandamide-induced transient rise in SUA in the RVLM is related to its brief pressor effect. As tested in three experiments, neither the transient rise in sSND nor the brief pressor response to anandamide was affected by pretreatment of the rats with 2 mg/kg phentolamine (sSND: +67±17% versus +68±22%; BP: +27±4 versus +39±1 mm Hg, before and after phentolamine, respectively).
Results of an earlier study demonstrated that the endogenous cannabinoid ligand anandamide causes complex yet highly reproducible changes in BP and heart rate in anesthetized rats and suggested different underlying mechanisms for the various components of this effect.11 The initial profound but transient bradycardia as well as associated hypotension is thought to be vagally mediated, as indicated by the absence of this component after atropine treatment or cervical vagotomy.11 This is confirmed by the present observation that the initial brief drop in BP does not occur in barodenervated rats in which the cervical vagi have been transected. Unlike this vagal response, which is unique to anandamide, the subsequent brief pressor response has been observed after administration of both Δ9-THC1 13 and anandamide.11 The mechanism of this effect is unclear. The finding that this pressor component is not inhibited but rather enhanced after α-receptor blockade or cervical cord transection suggests a peripheral mechanism that does not involve the sympathetic nervous system.11 The present findings confirm the lack of sympathetic nervous system involvement in the pressor response to anandamide. Anandamide caused a brief rise in sSND in both baroreceptor-intact and barodenervated rats, which indicates a direct rather than reflex-induced effect. This increase in neuronal activity preceded the pressor response, suggesting that it may have caused it. However, the finding that blocking of ganglionic transmission or α-adrenergic receptors failed to alter either effect clearly dissociates them. Why the increase in RVLM activity and the parallel increase in sSND do not result in a pressor response is not clear, although it is possible that a centrally triggered increase in sympathetic outflow is counteracted by an additional action of anandamide at the sympathetic nerve terminal to decrease transmitter release (see below). Whether the peripherally induced pressor response to anandamide is receptor mediated is not known, but an involvement of CB1 receptors has been ruled out.11
From a practical point of view, the prolonged hypotensive response to cannabinoids is clearly the most interesting, as it might be exploited in the treatment of hypertension. Indeed, such thoughts had been encouraged in the past by a reported dissociation of neurobehavioral and hypotensive effects in the case of certain cannabinoid analogues.13 The depressor and pressor components may be also dissociable, as indicated by the involvement of CB1 receptors in the former but not the latter effect of anandamide and Δ9-THC.11 As for the mechanism underlying the depressor response, sympathoinhibition was suggested by the observation that phentolamine treatment or cervical cord transection greatly attenuated this component of the anandamide response,11 although the site of the sympathoinhibitory effect has not been clarified. The present findings clearly argue against a centrally induced sympathoinhibitory effect. The sympathetic premotor neurons in the RVLM represent an obligatory common pathway for centrally elicited sympathomodulatory effects. In baroreceptor intact rats, the hypotensive response to anandamide was accompanied by an increase in the activity of these neurons, which was abolished by barodenervation, without any evidence for an anandamide-induced decrease in neuronal activity. Sympathetic postganglionic nerve activity closely paralleled the activity of RVLM neurons; ie, activity was increased in baroreceptor-intact rats and remained unchanged during the hypotensive response to anandamide in barodenervated rats. This confirms the absence of a central inhibitory action of anandamide and also suggests the lack of a significant effect on ganglionic transmission. The lack of effect of anandamide on ganglionic transmission is more clearly indicated by the finding that anandamide did not affect the increase in sSND triggered by RVLM stimulation. This is in agreement with the reported lack of effect of Δ9-THC on transmission through various sympathetic ganglia.3 5 14 The failure of anandamide to decrease sSND is in contrast to the reported decrease in the activity of cardiac sympathetic neurons by Δ9-THC.5 The reason for this difference is not clear. It is possible that despite the remarkable similarity in the mechanisms involved in the hypotensive effects of Δ9-THC and anandamide,11 a potential central site of action may contribute to the very prolonged (>1 hour) hypotensive action of Δ9-THC but not to the much shorter effect (<10 minutes) of anandamide. However, neither Δ9-THC5 nor anandamide11 produced hypotension when administered into a cerebral ventricle or the cisterna magna, which reduces the likelihood of a central site of action for both.
Anandamide significantly reduced the pressor response to RVLM stimulation without reducing postganglionic sympathetic nerve activity (see Fig 2⇑) or the pressor response to direct stimulation of vascular α-receptors. The only mechanism that can account for this response pattern is presynaptic inhibition of norepinephrine release from peripheral sympathetic nerve terminals. Such a mechanism is compatible with the dependence of the hypotensive effects of anandamide and Δ9-THC on intact sympathetic innervation and vascular vasoconstrictor tone.5 11 In the absence of evidence for a central site of action, presynaptic inhibition of norepinephrine release may also account for the reported ability of Δ9-THC to inhibit baroreflex-mediated vasoconstriction in rats.3 Δ9-THC inhibits electrical stimulation–induced norepinephrine release15 and twitch response16 in mouse vas deferens. More recently, we found that both Δ9-THC and anandamide inhibit electrical stimulation–induced but not tyramine-induced norepinephrine release from rat atria and vas deferens, and these effects are potently inhibited by a selective CB1 receptor antagonist. Furthermore, CB1 receptor mRNA was found to be present in sympathetic ganglia.17 Together, these observations provide strong evidence for the presence of presynaptic CB1 receptors on postganglionic sympathetic nerve terminals and suggest that activation of these receptors is responsible for the pronounced hypotensive response to anandamide and Δ9-THC. These observations do not exclude the possibility that certain synthetic cannabinoids that cause much more pronounced and longer lasting hypotension than anandamide may have additional (eg, central) sites of hypotensive action.
Selected Abbreviations and Acronyms
|RVLM||=||rostral ventrolateral medulla|
|sSND||=||splanchnic sympathetic nerve discharge|
|SUA||=||single unit activity|
This work was supported by National Institutes of Health (NIH) grants HL-49938 to G.K. and HL-28785 to P.G.G. K.D.L. was supported by a predoctoral fellowship from NIH training grant DA-07027. We thank Dr Raj Razdan (Organix Inc, Woburn, Mass) for the synthesis of anandamide used in this study.
Reprint requests to Karoly Varga, MD, PhD, Department of Pharmacology and Toxicology, Virginia Commonwealth University, Box 980613, Richmond, VA 23298. E-mail firstname.lastname@example.org.
- Received April 16, 1996.
- Revision received June 5, 1996.
- Revision received June 19, 1996.
Dewey WL, Harris LS, Howes JF, Kennedy JS, Granchelli FE, Pars HG, Razdan RK. Pharmacology of some marijuana constituents and two heterocyclic analogs. Nature. 1970;226:1255-1257.
Hardman HF, Domino EF, Seevers MH. General pharmacological actions of some synthetic tetrahydrocannabinol derivatives. Pharmacol Rev. 1971;23:295-315.
Benovitz NL, Jones RT. Cardiovascular effects of prolonged delta-9-tetrahydrocannabinol ingestion. Clin Pharmacol Ther. 1975;25:440-446.
Devane WA, Hanus L, Breuer A, Pertwee RG, Stevenson LA, Griffin G, Gibson G, Mandelbaum A, Etinger A, Mechoulam R. Isolation and structure of a brain constituent that binds to the cannabinoid receptor. Science. 1992;258:1946-1949.
Allen AM, Guyenet PG. α2-Adrenoceptor-mediated modulation of bulbospinal barosensitive cells of rat rostral medulla. Am J Physiol. 1993;34:R1065-R1075.