Jacob Bell
New Member
MILES HERKENHAM*t, ALLISON B. LYNN*, MARK D. LITTLE*, M. ROSS JOHNSONt, LAWRENCE S. MELVIN§, BRIAN R. DE COSTA¶, AND KENNER C. RICE¶
*Unit on Functional Neuroanatomy, Building 36, Room 2D-15, National Institute of Mental Health, Bethesda, MD 20892; tGlaxo Inc., Research Triangle Park, NC 27709; ¢Central Research, Pfizer Inc., Groton, CT 06340; and 1Laboratory of Medicinal Chemistry, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD 20892
ABSTRACT [3H]CP 55,940, a radiolabeled synthetic can-nabinoid, which is 10-100 times more potent in vivo than 4i9-tetrahydrocannabinol, was used to characterize and localize a specific cannabinoid receptor in brain sections. The potencies of a series of natural and synthetic cannabinoids competitors of [3H]CP 55,940 binding,correlated closely with their relative potencies in several biological assays, suggesting that the receptor characterized in our in vitro assay is the same receptor that mediates behavioral and pharmacological effects of can¬nabWoids, including human subjective experience. Autoradi¬ography of cannabinoid receptors in brain sections from sev¬eral mammalian species, including human, reveals a unique and conserved distribution; binding is most dense in outflow nuclei of the basal ganglia-the substantia nigra pars reticulate and globus pallidus-and in the hippocampus and cerebellum. Generally high densities in forebrain and cerebellum implicate roles for cannabinoids in cognition and movement. Sparse densities W lower brainstem areas controlling cardiovascular and respiratory functions may explain why high doses of 4,9-tetrahydrocannabinol are not lethal.
Marihuana (Cannabis sativa) is one of the oldest and most widely used drugs in the world (1, 2). The major psychoactive ingredient of the marihuana plant is 09-tetrahydrocannabinol (A9-THC) (3). A9-MC and other natural and synthetic can-nabinoids produce characteristic motor, cognitive, and an-algesic effects (4, 5). Early reports showing cannabinoid-like activity of 9/i-hydroxyhexahydrocannabinol (p-HHC) (6-8) inspired the synthesis of several distinct cannabinoids for studies of their potential use as analgesics (9). The synthetic cannabinoids share physicochemical properties with the nat¬ural cannabinoids and produce many behavioral and physi¬ological effects characteristic of Q9-THC but are 5-1000 times more potent and show high enantioselectivity. One of these, CP 55,940, was tritiated and used to identify and fully characterize a unique cannabinoid receptor in membranes from rat brain (10). In this study we characterize and validate the binding of [3H]CP 55,940 in slide-mounted brain sections and use the same assay conditions to autoradiographically visualize the distribution of cannabinoid receptors.
METHODS
[3H]CP 55,940 is a bicyclic molecule that is one of a series of synthetic cannabinoids whose structure and biological activity have been documented (9-12) (Fig. 1). It was custom radiolabeled at DuPont/NEN by tritium reduction of CP 60,106 (10). The product was purified by thin layer chroma tography on silica gel, eluting with ethylacetate/hexane [1:9(vol/vol)], and the band comigrating with unlabeled CP payment. This article must therefore be hereby marked "advertisement'' in accordance with 18 U.S.C. §1734 solely to indicate this fact.
~B-HHC (the original synthetic cannabinoid from which the CP compounds were derived), and CP 55,940. (Lower) Compet¬itive inhibition of 1 nM [3H]CP 55,940 binding in whole rat brain sausage sections by various synthetic and natural cannabinoids at the concentrations indicated. The data are normalized to specific binding (total minus nonspecific binding) in the absence of competitors. Nonspecific binding was determined by addition of 10 µM CP 55,244 [the most potent cannabinoid in the CP series (9)] and typically represented 10-20% of total binding at both 1 and 10 nM [3H]CP 55,940. Data points represent means of eight determinations. ACD, tricyclic; AC, bicyclic ring nomenclature of Johnson and Melvin (9).
55,940 was extracted, giving a radiochemical yield of 15% and a specific activity of 79 Ci/mmol (1 Ci = 37 GBq). Optimi-zation and competition studies were carried out with slide-mounted sections cut from unfixed frozen rat brains. Incu-bations were in plastic cytomailers(CMS), each containing eight 30-µm-thick "sausage" sections on four gelatin-coated slides in 5 ml of solution (13). The sausage sections were prepared by combining and mincing three whole rat brains to achieve relative homogeneity of receptor and protein content, then placing the paste into a tube, and freezing it to produce a cylindrical sausage that can be cryostat-cut to make sections of uniform composition and size (14). Accu¬racy of dilutions was checked in mock-incubations in which either [3H]CP 55,940 or [3H]A9-THC (provided by the Na¬tional Institute on Drug Abuse) were substituted for unla¬beled drug, and the solutions were assayed for radioactivity.
To determine binding kinetics, two concentrations of [3H]CP 55,940 (1 and 10 nM) were each competitively inhibited by 6-12 concentrations of unlabeled drug. Competitive inhibition curves were subjected to binding surface analysis, which is a comput¬erized iterative curve-fitting program for determining best-fit parameter estimates (Kd, K;, and B) according to 1- or 2-site competitive binding models (15-17). Determination of fmol bound per section was by liquid scintillation counting of the section-laden slide fragments placed overnight in detergent fluor. Some sausage sections were analyzed for protein content by the method of Lowry et al. (18) and found to have 456 ± 26 µg of protein per section.
Autoradiography was performed on 25-p.m-thick brain sec-tions of rat (male Sprague-Dawley, n = 12), guinea pig (male Hartley, n = 4), dog (beagle, n = 2), rhesus monkey (n = 1), and human (dying of nonneurological disorders, n = 3). Sections were incubated in 10 nM [3H]CP 55,940 by using optimized conditions, washed, dried, and exposed to tritium-sensitive film (LKB or Amersham) for 3-4 weeks before developing.
RESULTS
Assay conditions yielding 80-90% specific binding were as follows: incubation at 37°C for 2 hr in 50 mM Tris-HCl (pH 7.4) containing 5% (wt/vol) bovine serum albumin (BSA) and 1-10 nM [3H]CP 55,940 and washing at 0°C for 4 hr in the same buffer with 1% BSA. By using these optimized conditions, the sausage studies showed that binding was saturable and that competitive inhibition curves were best-fit by a single-site kinetic model: the affinity (Kd) of [3H]CP 55,940 was 15 ± 3 nM and the capacity (Bin whole brain was 0.9 pmol/mg of protein. Similar parameters were obtained if 1% BSA was used in the incubation, but variability was greater. Binding of 1 nM [3H]CP 55,940 was completely blocked by 10 µM A9-THC, which showed inhibition in a dose-dependent fashion (Fig. 1).
Inhibition by other natural and synthetic cannabinoids was also shown (Fig. 1; for K; values, see Table 1). The data from the section-binding assay were in close agreement with data from a centrifugation assay using mem¬branes from rat cortex (10). The Bwas similar in the two studies (though ours was derived from whole brain), but the Kd in our assay was about 100-fold higher. The low affinity in sections relative to that in membranes appears to reflect differences in the nature of the assays. In both assays the addition of guanine nucleotides converted the receptor to a low-affinity state. In sections the nonhydrolyzable GTP analog, guanosine 5'-W,y-imido]triphosphate, at 10 µM in¬hibited binding of 10 nM [3H]CP 55,940 by 94%, and the GDP analog, guanosine 5'-[$-thio]diphosphate, at 10 µM inhibited binding by 79%. Finally, in both assays there was a similar rank order of drug potencies.
For several cannabinoids, inhibition constants (K; values) and relative biological potencies are given in Table 1. Highly significant correlations exist between the K; values and potencies of the drugs in tests of dog ataxia and human subjective experience, the two most reliable markers of cannabinoid activity (4, 5). Correlations with potencies in the other tests suggest that the measured effects were similarly receptor-mediated. Enantioselectivity was striking; the (-) and (+) forms of CP 55,244 differed by more than 10,000-fold in vitro, a separation predicted by the rigid structure of the molecule (9) and by potencies in vivo. Natural cannabinoids lacking psychoactive properties, such as cannabidiol, showed extremely low potency at the receptor, and all tested noncannabinoid drugs had no potency (Table 1).
Autoradiography showed that in all species very dense binding was found in the globus pallidus, substantia nigra pars reticulata (SNr), and the molecular layers of the cere-bellum and hippocampal dentate gyros (Figs. 2 and 3). Dense binding was also found in the cerebral cortex, other parts of the hippocampal formation, and striatum. In rat, rhesus monkey, and human, the SNr contained the highest level of binding (Fig. 3). In dog, the cerebellar molecular layer was most dense (Fig. 2H). In guinea pig and dog, the hippocampal formation had selectively dense binding (Fig. 2 E and F). Neocortex in all species had moderate binding across fields, with peaks in superficial and deep layers. Very low and homogeneous binding characterized the thalamus and most of the brainstem, including all of the monoamine-containing cell groups, reticular formation, primary sensory, viscero¬motor and cranial motor nuclei, and the area postrema. The exceptions-hypothalamus, basal amygdala, central gray,nucleus of the solitary tract, and laminae I-III and X of the spinal cord-showed slightly higher but still sparse binding (Figs. 2 and 3).
Quantitative autoradiography confirmed the very high numbers of receptors, exceeding 1 pmol/mg of protein in densely labeled areas (data not shown). Cannabinoid recep¬tor density was far in excess of densities of neuropeptide receptors and was similar to levels of cortical benzodiazepine (21), striatal dopamine (22, 23), and whole-brain glutamate receptors (24).
DISCUSSION
Previous attempts to characterize the cannabinoid receptor were unsuccessful for several reasons (for discussion, see ref. 10). Cannabinoids are extremely hydrophobic and adhere to filters (see ref. 10) and other surfaces (25). The section assay circumvents some of these problems; in addition, BSA ap¬pears to act as a carrier to keep cannabinoids in solution without appreciably affecting binding kinetics. The low non-specific binding and absence of binding in white matter indicates that the autoradiographic patterns are not affected by ligand lipophilia. Other obstacles were the use of O8-[3H]THC (26) or O9-[3H]THC (27), which bind with low affinity and have low specific activities, or the use of 5'-[3H]trimethylammonium-O8-THC (20), which does not act like a cannabinoid in most animal tests and which has low affinity for the presently described receptor (Table 1). In contrast, [3H]CP 55,940 has high specific activity, high af¬finity, and biological activity similar to that of O9-THC.
The structure-activity profile suggests that the receptor defined by the binding of [3H]CP 55,940 is the same receptor that mediates all of the behavioral and pharmacological effects of cannabinoids listed in Table 1, including the sub¬jective experience termed the human "high". All other tested psychoactive drugs, neurotransmitters, steroids, and ei¬cosanoids at 10 µM concentrations failed to bind to this receptor (Table 1). There was no compelling evidence for receptor subtypes from the present analysis.
The overall central nervous system distribution, although not similar to any known drug or neurotransmitter receptor pattern, resembles autoradiographic distributions of second messengers (28, 29). These mapping similarities, the very high abundance of the cannabinoid receptor, and the pro-found inhibition of binding by guanine nucleotides suggest that the cannabinoid receptor is closely associated with second messenger systems. Total inhibition of binding by the GTP analog indicates that the receptor is functionally and strongly coupled to a guanine nucleotide-binding regulatory (G) protein in our assay. It also indicates that the ligand is an agonist and that there are multiple affinity states of the receptor, as found with the other major receptor classes coupled to adenylate cyclase by G proteins (30).
Dense binding in the basal ganglia and cerebellum suggests cannabinoid involvement in movement control. Cannabi¬noids depress motor functions with a characteristic stimula¬tory component (4, 5). Dog shows a static ataxia (Table 1) and has high receptor levels in cerebellum and relatively low levels in SNr (Fig. 2 F and H). Human shows much less motor depression (3-5) and lower relative densities in cere¬bellum (Fig. 3), suggesting cerebellar mediation of the motor impairments in animals.
Accounts of cannabis use in humans stress the loosening of associations, fragmentation of thought, and confusion on attempting to remember recent occurrences (5, 31). The most consistent effect of Q9-THC on performance is disruption of selective aspects of short-term memory tasks, similar to that found in monkeys and patients with damage to limbic cortical areas (31-33). These cognitive effects may be mediated by receptors in the cerebral cortex. The hippocampal cortex "gates" information during memory consolidation and codes spatial and temporal relations among stimuli and responses (34, 35). Q9-THC causes memory "intrusions" (36), impairs temporal aspects of performance (37), and suppresses hip¬pocampal electrical activity (38).
The presence of cannabinoid receptors in the ventromedial striatum suggests an association with dopamine circuits thought to mediate reward (39-41). However, reinforcing properties of cannabinoids have been difficult to demonstrate in animals (42, 43). Moreover, cannabinoid receptors in the basal ganglia are not localized on dopamine neurons (44).There are virtually no reports of fatal cannabis overdose in humans (1, 4, 5). The safety reflects the paucity of receptors in medullary nuclei that mediate respiratory and cardiovascular functions.
Anticonvulsant and antiemetic effects of cannabinoids
have therapeutic value (4, 5). The localization of cannabinoid receptors in motor areas suggests additional therapeutic applications. Cannabinoids exacerbate hypokinesia in Park¬inson disease but are beneficial for some forms of dystonia, tremor, and spasticity (4, 5, 45-47). The development of an antagonist could provide additional therapeutic uses of value. The receptor binding assay will be helpful in this regard, and it can be used also to screen drugs that have greater potency or bind irreversibly to aid in the identification of the receptor gene and the putative endogenous ligand.
We thank Drs. J. Atack, J. Hill, and J. Johannessen for providing human and dog tissues, Dr. R. Rothman for pharmacokinetics advice, and Drs. A. Howlett and W. Devane for sharing unpublished
data.
References:
1.Harris, L. S., Dewey, W. L. & Razdan, R. K. (1977) in Handbook of Experimental Pharmacology, ed. Martin, W. R. (Springer, New York), pp. 371-429.
2.Mechoulam, R. (1986) in Cannabinoids as Therapeutic Agents, ed. Mechoulam, R. (CRC, Boca Raton, FL), pp. 1-19.
3.Razdan, R. K. (1986) Pharmacol. Rev. 38, 75-149.
4.Dewey, W. L. (1986) Pharmacol. Rev. 38, 151-178.
5.Hollister, L. E. (1986) Pharmacol. Rev. 38, 1-20.
6.Wilson, R. S. & May, E. L. (1975) J. Med. Chem. 18, 700-703.
7.Wilson, R. S., May, E. L., Martin, B. R. & Dewey, W. L. (1976) J. Med. Chem. 19, 1165-1167.
8.Wilson, R. S., May, E. L. & Dewey, W. L. (1979) J. Med. Chem. 22,886-888.
9.Johnson, M. R. & Melvin, L. S. (1986) in Cannabinoids as Therapeutic Agents, ed. Mechoulam, R. (CRC, Boca Raton, FL), pp. 121-145.
10.Devane, W. A., Dysarz, F. A. I., Johnson, M. R., Melvin, L. S. & Howlett, A. C. (1988) Mol. Pharmacol. 34, 605-613.
11.Howlett, A. C., Johnson, M. R., Melvin, L. S. & Milne, G. M.(1988) Mol. Pharmacol. 33, 297-302.
12.Little, P. J., Compton, D. R., Johnson, M. R. & Martin, B. R.(1988) J. Pharmacol. Exp. Ther. 247, 1046-1051.
13.Herkenham, M. (1988) in Molecular Neuroanatomy, eds. van Leeuwen, F., Bugs, R. M., Pool, C. W. & Pach, O. (Elsevier, Amsterdam), pp. 111-120.
14.Rothman, R. B., Schumacher, U. K. & Pert, C. B. (1983) Neu¬ropeptides 3, 493-499.
15.Rothman, R. B. (1986) Alcohol Drug Res. 6, 309-325.
16.McGonigle, P., Neve, K. A. & Molinoff, P. B. (1986) Mol. Phar¬macol. 30, 329-337.
17.Rothman, R. B., Long, J. B., Bykov, V., Jacobson, A. E., Rice, K. C. & Holaday, J. W. (1988) J. Pharmacol. Exp. Ther. 247, 405-416.
18.Lowry, O. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J.
(1951) J. Biol. Chem. 193, 265-275.
19.Burstein, S. H., Hull, K., Hunter, S. A. & Latman, V. (1988) FASEB J. 2, 3022-3026.
20.Nye, J. S., Seltzman, H. H., Pitt, C. G. & Snyder, S. H. (1985) J. Pharmacol. Exp. Ther. 234, 784-791.
21.Zezula, J., Comes, R., Probst, A. & Palacios, J. M. (1988) Neuro¬science 25, 771-795.
22.Boyson, S. J., McGonigle, P. & Molinoff, P. B. (1986) J. Neurosci. 6, 3177-3188.
23.Richfield, E. K., Young, A. B. & Penney, J. B. (1986) Brain Res. 383, 121-128.
24.M. Greenamyre, J. T., Young, A. B. & Penney, J. B. (1984) J. Neu-rosci. 4, 2133-2144.
25.Garrett, E. R. & Hunt, C. A. (1974) J. Pharm. Sci. 63, 1056-1064.
26.Harris, L. S., Carchman, R. A. & Martin, B. R. (1978) Life Sci. 22, 1131-1138.
27.Roth, S. H. & Williams, P. J. (1979) J. Pharm. Pharmacol. 31, 224-230.
28.U. Worley, P. F., Baraban, J. M., De Souza, E. B. & Snyder, S. H.
(1986) Proc. Natl. Acad. Sci. USA 83, 4053-4057.
29.Worley, P. F., Baraban, J. M. & Snyder, S. H. (1986) J. Neurosci. 6, 199-207.
30.Dohlman, H. G., Caron, M. G. & Lefkowitz, R. J. (1987) Biochem¬istry 26, 2257-2264.
31.Miller, L. L. (1984) in The Cannabinoids: Chemical, Pharmaco¬logic, and Therapeutic Aspects, eds. Agurell, S., Dewey, W. L. & Willette, R. E. (Academic, New York), pp. 21-46.
32.Miller, L. L. & Braconnier, R. J. (1983) Psycho!. Bull. 93, 441-456.
33.Aigner, T. G. (1988) Psychopharmacology 95, 507-511.
34.Douglas, R. J. (1967) Psycho!. Bull. 67, 416-442.
35.Eichenbaum, H. & Cohen, N. J. (1988) Trends Neurosci. 11, 244-248.
36.Hooker, W. D. & Jones, R. T. (1987) Psychopharmacology 91, 20-24.
37. Schulze, G. E., McMillan, D. E., Bailey, J. R., Scallet, A., All, S. F., Slikker, W. J. & Paule, M. G. (1988) J. Pharmacol. Exp. Ther. 245, 178-186.
38.Campbell, K. A., Foster, T. C., Hampson, R. E. & Deadwyler, S. A. (1986) J. Pharmacol. Exp. Ther. 239, 941-945.
39.Kornetsky, C. (1985) NIDA Res. Monogr. 62, 30-50.
Q. Di Chiara, G. & Imperato, A. (1988) Proc. Natl. Acad. Sci. USA 85, 5274-5278.
41.Ng Cheong Ton, J. M., Gerhardt, G. A., Fnedemann, M., Etgen, A. M., Rose, G. M., Sharpless, N. S. & Gardner, E. L. (1988) Brain Res. 451, 59-68.
42.Stark, P. & Dews, P. B. (1980) J. Pharmacol. Exp. Ther. 214, 124-130.
43.Gardner, E. L., Paredes, W., Smith, D., Donner, A., Milling, C., Cohen, D. & Morrison, D. (1988) Psychopharmacology %,142-144.
44.Herkenham, M., Lynn, A. B., de Costa, B. & Richfield, E. K.
(1989) Soc. Neurosci. Abstr. 15, 905.
45.Marsden, C. D. (1981) in Disorders of Movement, Current Status of Modern Therapy, ed. Barbeau, A. (Lippincott, Philadelphia), Vol. 8, pp. 81-104.
46.Q. Petro, D. J. & Ellenberger, C. E. (1981) J. Clin. Pharmacol. 21, 413s-416s.
47. Clifford, D. B. (1983) Ann. Neurol. 13, 669-671.
Source including Charts, Graphs and Figures: Cannabinoid receptor localization in brain
*Unit on Functional Neuroanatomy, Building 36, Room 2D-15, National Institute of Mental Health, Bethesda, MD 20892; tGlaxo Inc., Research Triangle Park, NC 27709; ¢Central Research, Pfizer Inc., Groton, CT 06340; and 1Laboratory of Medicinal Chemistry, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD 20892
ABSTRACT [3H]CP 55,940, a radiolabeled synthetic can-nabinoid, which is 10-100 times more potent in vivo than 4i9-tetrahydrocannabinol, was used to characterize and localize a specific cannabinoid receptor in brain sections. The potencies of a series of natural and synthetic cannabinoids competitors of [3H]CP 55,940 binding,correlated closely with their relative potencies in several biological assays, suggesting that the receptor characterized in our in vitro assay is the same receptor that mediates behavioral and pharmacological effects of can¬nabWoids, including human subjective experience. Autoradi¬ography of cannabinoid receptors in brain sections from sev¬eral mammalian species, including human, reveals a unique and conserved distribution; binding is most dense in outflow nuclei of the basal ganglia-the substantia nigra pars reticulate and globus pallidus-and in the hippocampus and cerebellum. Generally high densities in forebrain and cerebellum implicate roles for cannabinoids in cognition and movement. Sparse densities W lower brainstem areas controlling cardiovascular and respiratory functions may explain why high doses of 4,9-tetrahydrocannabinol are not lethal.
Marihuana (Cannabis sativa) is one of the oldest and most widely used drugs in the world (1, 2). The major psychoactive ingredient of the marihuana plant is 09-tetrahydrocannabinol (A9-THC) (3). A9-MC and other natural and synthetic can-nabinoids produce characteristic motor, cognitive, and an-algesic effects (4, 5). Early reports showing cannabinoid-like activity of 9/i-hydroxyhexahydrocannabinol (p-HHC) (6-8) inspired the synthesis of several distinct cannabinoids for studies of their potential use as analgesics (9). The synthetic cannabinoids share physicochemical properties with the nat¬ural cannabinoids and produce many behavioral and physi¬ological effects characteristic of Q9-THC but are 5-1000 times more potent and show high enantioselectivity. One of these, CP 55,940, was tritiated and used to identify and fully characterize a unique cannabinoid receptor in membranes from rat brain (10). In this study we characterize and validate the binding of [3H]CP 55,940 in slide-mounted brain sections and use the same assay conditions to autoradiographically visualize the distribution of cannabinoid receptors.
METHODS
[3H]CP 55,940 is a bicyclic molecule that is one of a series of synthetic cannabinoids whose structure and biological activity have been documented (9-12) (Fig. 1). It was custom radiolabeled at DuPont/NEN by tritium reduction of CP 60,106 (10). The product was purified by thin layer chroma tography on silica gel, eluting with ethylacetate/hexane [1:9(vol/vol)], and the band comigrating with unlabeled CP payment. This article must therefore be hereby marked "advertisement'' in accordance with 18 U.S.C. §1734 solely to indicate this fact.
~B-HHC (the original synthetic cannabinoid from which the CP compounds were derived), and CP 55,940. (Lower) Compet¬itive inhibition of 1 nM [3H]CP 55,940 binding in whole rat brain sausage sections by various synthetic and natural cannabinoids at the concentrations indicated. The data are normalized to specific binding (total minus nonspecific binding) in the absence of competitors. Nonspecific binding was determined by addition of 10 µM CP 55,244 [the most potent cannabinoid in the CP series (9)] and typically represented 10-20% of total binding at both 1 and 10 nM [3H]CP 55,940. Data points represent means of eight determinations. ACD, tricyclic; AC, bicyclic ring nomenclature of Johnson and Melvin (9).
55,940 was extracted, giving a radiochemical yield of 15% and a specific activity of 79 Ci/mmol (1 Ci = 37 GBq). Optimi-zation and competition studies were carried out with slide-mounted sections cut from unfixed frozen rat brains. Incu-bations were in plastic cytomailers(CMS), each containing eight 30-µm-thick "sausage" sections on four gelatin-coated slides in 5 ml of solution (13). The sausage sections were prepared by combining and mincing three whole rat brains to achieve relative homogeneity of receptor and protein content, then placing the paste into a tube, and freezing it to produce a cylindrical sausage that can be cryostat-cut to make sections of uniform composition and size (14). Accu¬racy of dilutions was checked in mock-incubations in which either [3H]CP 55,940 or [3H]A9-THC (provided by the Na¬tional Institute on Drug Abuse) were substituted for unla¬beled drug, and the solutions were assayed for radioactivity.
To determine binding kinetics, two concentrations of [3H]CP 55,940 (1 and 10 nM) were each competitively inhibited by 6-12 concentrations of unlabeled drug. Competitive inhibition curves were subjected to binding surface analysis, which is a comput¬erized iterative curve-fitting program for determining best-fit parameter estimates (Kd, K;, and B) according to 1- or 2-site competitive binding models (15-17). Determination of fmol bound per section was by liquid scintillation counting of the section-laden slide fragments placed overnight in detergent fluor. Some sausage sections were analyzed for protein content by the method of Lowry et al. (18) and found to have 456 ± 26 µg of protein per section.
Autoradiography was performed on 25-p.m-thick brain sec-tions of rat (male Sprague-Dawley, n = 12), guinea pig (male Hartley, n = 4), dog (beagle, n = 2), rhesus monkey (n = 1), and human (dying of nonneurological disorders, n = 3). Sections were incubated in 10 nM [3H]CP 55,940 by using optimized conditions, washed, dried, and exposed to tritium-sensitive film (LKB or Amersham) for 3-4 weeks before developing.
RESULTS
Assay conditions yielding 80-90% specific binding were as follows: incubation at 37°C for 2 hr in 50 mM Tris-HCl (pH 7.4) containing 5% (wt/vol) bovine serum albumin (BSA) and 1-10 nM [3H]CP 55,940 and washing at 0°C for 4 hr in the same buffer with 1% BSA. By using these optimized conditions, the sausage studies showed that binding was saturable and that competitive inhibition curves were best-fit by a single-site kinetic model: the affinity (Kd) of [3H]CP 55,940 was 15 ± 3 nM and the capacity (Bin whole brain was 0.9 pmol/mg of protein. Similar parameters were obtained if 1% BSA was used in the incubation, but variability was greater. Binding of 1 nM [3H]CP 55,940 was completely blocked by 10 µM A9-THC, which showed inhibition in a dose-dependent fashion (Fig. 1).
Inhibition by other natural and synthetic cannabinoids was also shown (Fig. 1; for K; values, see Table 1). The data from the section-binding assay were in close agreement with data from a centrifugation assay using mem¬branes from rat cortex (10). The Bwas similar in the two studies (though ours was derived from whole brain), but the Kd in our assay was about 100-fold higher. The low affinity in sections relative to that in membranes appears to reflect differences in the nature of the assays. In both assays the addition of guanine nucleotides converted the receptor to a low-affinity state. In sections the nonhydrolyzable GTP analog, guanosine 5'-W,y-imido]triphosphate, at 10 µM in¬hibited binding of 10 nM [3H]CP 55,940 by 94%, and the GDP analog, guanosine 5'-[$-thio]diphosphate, at 10 µM inhibited binding by 79%. Finally, in both assays there was a similar rank order of drug potencies.
For several cannabinoids, inhibition constants (K; values) and relative biological potencies are given in Table 1. Highly significant correlations exist between the K; values and potencies of the drugs in tests of dog ataxia and human subjective experience, the two most reliable markers of cannabinoid activity (4, 5). Correlations with potencies in the other tests suggest that the measured effects were similarly receptor-mediated. Enantioselectivity was striking; the (-) and (+) forms of CP 55,244 differed by more than 10,000-fold in vitro, a separation predicted by the rigid structure of the molecule (9) and by potencies in vivo. Natural cannabinoids lacking psychoactive properties, such as cannabidiol, showed extremely low potency at the receptor, and all tested noncannabinoid drugs had no potency (Table 1).
Autoradiography showed that in all species very dense binding was found in the globus pallidus, substantia nigra pars reticulata (SNr), and the molecular layers of the cere-bellum and hippocampal dentate gyros (Figs. 2 and 3). Dense binding was also found in the cerebral cortex, other parts of the hippocampal formation, and striatum. In rat, rhesus monkey, and human, the SNr contained the highest level of binding (Fig. 3). In dog, the cerebellar molecular layer was most dense (Fig. 2H). In guinea pig and dog, the hippocampal formation had selectively dense binding (Fig. 2 E and F). Neocortex in all species had moderate binding across fields, with peaks in superficial and deep layers. Very low and homogeneous binding characterized the thalamus and most of the brainstem, including all of the monoamine-containing cell groups, reticular formation, primary sensory, viscero¬motor and cranial motor nuclei, and the area postrema. The exceptions-hypothalamus, basal amygdala, central gray,nucleus of the solitary tract, and laminae I-III and X of the spinal cord-showed slightly higher but still sparse binding (Figs. 2 and 3).
Quantitative autoradiography confirmed the very high numbers of receptors, exceeding 1 pmol/mg of protein in densely labeled areas (data not shown). Cannabinoid recep¬tor density was far in excess of densities of neuropeptide receptors and was similar to levels of cortical benzodiazepine (21), striatal dopamine (22, 23), and whole-brain glutamate receptors (24).
DISCUSSION
Previous attempts to characterize the cannabinoid receptor were unsuccessful for several reasons (for discussion, see ref. 10). Cannabinoids are extremely hydrophobic and adhere to filters (see ref. 10) and other surfaces (25). The section assay circumvents some of these problems; in addition, BSA ap¬pears to act as a carrier to keep cannabinoids in solution without appreciably affecting binding kinetics. The low non-specific binding and absence of binding in white matter indicates that the autoradiographic patterns are not affected by ligand lipophilia. Other obstacles were the use of O8-[3H]THC (26) or O9-[3H]THC (27), which bind with low affinity and have low specific activities, or the use of 5'-[3H]trimethylammonium-O8-THC (20), which does not act like a cannabinoid in most animal tests and which has low affinity for the presently described receptor (Table 1). In contrast, [3H]CP 55,940 has high specific activity, high af¬finity, and biological activity similar to that of O9-THC.
The structure-activity profile suggests that the receptor defined by the binding of [3H]CP 55,940 is the same receptor that mediates all of the behavioral and pharmacological effects of cannabinoids listed in Table 1, including the sub¬jective experience termed the human "high". All other tested psychoactive drugs, neurotransmitters, steroids, and ei¬cosanoids at 10 µM concentrations failed to bind to this receptor (Table 1). There was no compelling evidence for receptor subtypes from the present analysis.
The overall central nervous system distribution, although not similar to any known drug or neurotransmitter receptor pattern, resembles autoradiographic distributions of second messengers (28, 29). These mapping similarities, the very high abundance of the cannabinoid receptor, and the pro-found inhibition of binding by guanine nucleotides suggest that the cannabinoid receptor is closely associated with second messenger systems. Total inhibition of binding by the GTP analog indicates that the receptor is functionally and strongly coupled to a guanine nucleotide-binding regulatory (G) protein in our assay. It also indicates that the ligand is an agonist and that there are multiple affinity states of the receptor, as found with the other major receptor classes coupled to adenylate cyclase by G proteins (30).
Dense binding in the basal ganglia and cerebellum suggests cannabinoid involvement in movement control. Cannabi¬noids depress motor functions with a characteristic stimula¬tory component (4, 5). Dog shows a static ataxia (Table 1) and has high receptor levels in cerebellum and relatively low levels in SNr (Fig. 2 F and H). Human shows much less motor depression (3-5) and lower relative densities in cere¬bellum (Fig. 3), suggesting cerebellar mediation of the motor impairments in animals.
Accounts of cannabis use in humans stress the loosening of associations, fragmentation of thought, and confusion on attempting to remember recent occurrences (5, 31). The most consistent effect of Q9-THC on performance is disruption of selective aspects of short-term memory tasks, similar to that found in monkeys and patients with damage to limbic cortical areas (31-33). These cognitive effects may be mediated by receptors in the cerebral cortex. The hippocampal cortex "gates" information during memory consolidation and codes spatial and temporal relations among stimuli and responses (34, 35). Q9-THC causes memory "intrusions" (36), impairs temporal aspects of performance (37), and suppresses hip¬pocampal electrical activity (38).
The presence of cannabinoid receptors in the ventromedial striatum suggests an association with dopamine circuits thought to mediate reward (39-41). However, reinforcing properties of cannabinoids have been difficult to demonstrate in animals (42, 43). Moreover, cannabinoid receptors in the basal ganglia are not localized on dopamine neurons (44).There are virtually no reports of fatal cannabis overdose in humans (1, 4, 5). The safety reflects the paucity of receptors in medullary nuclei that mediate respiratory and cardiovascular functions.
Anticonvulsant and antiemetic effects of cannabinoids
have therapeutic value (4, 5). The localization of cannabinoid receptors in motor areas suggests additional therapeutic applications. Cannabinoids exacerbate hypokinesia in Park¬inson disease but are beneficial for some forms of dystonia, tremor, and spasticity (4, 5, 45-47). The development of an antagonist could provide additional therapeutic uses of value. The receptor binding assay will be helpful in this regard, and it can be used also to screen drugs that have greater potency or bind irreversibly to aid in the identification of the receptor gene and the putative endogenous ligand.
We thank Drs. J. Atack, J. Hill, and J. Johannessen for providing human and dog tissues, Dr. R. Rothman for pharmacokinetics advice, and Drs. A. Howlett and W. Devane for sharing unpublished
data.
References:
1.Harris, L. S., Dewey, W. L. & Razdan, R. K. (1977) in Handbook of Experimental Pharmacology, ed. Martin, W. R. (Springer, New York), pp. 371-429.
2.Mechoulam, R. (1986) in Cannabinoids as Therapeutic Agents, ed. Mechoulam, R. (CRC, Boca Raton, FL), pp. 1-19.
3.Razdan, R. K. (1986) Pharmacol. Rev. 38, 75-149.
4.Dewey, W. L. (1986) Pharmacol. Rev. 38, 151-178.
5.Hollister, L. E. (1986) Pharmacol. Rev. 38, 1-20.
6.Wilson, R. S. & May, E. L. (1975) J. Med. Chem. 18, 700-703.
7.Wilson, R. S., May, E. L., Martin, B. R. & Dewey, W. L. (1976) J. Med. Chem. 19, 1165-1167.
8.Wilson, R. S., May, E. L. & Dewey, W. L. (1979) J. Med. Chem. 22,886-888.
9.Johnson, M. R. & Melvin, L. S. (1986) in Cannabinoids as Therapeutic Agents, ed. Mechoulam, R. (CRC, Boca Raton, FL), pp. 121-145.
10.Devane, W. A., Dysarz, F. A. I., Johnson, M. R., Melvin, L. S. & Howlett, A. C. (1988) Mol. Pharmacol. 34, 605-613.
11.Howlett, A. C., Johnson, M. R., Melvin, L. S. & Milne, G. M.(1988) Mol. Pharmacol. 33, 297-302.
12.Little, P. J., Compton, D. R., Johnson, M. R. & Martin, B. R.(1988) J. Pharmacol. Exp. Ther. 247, 1046-1051.
13.Herkenham, M. (1988) in Molecular Neuroanatomy, eds. van Leeuwen, F., Bugs, R. M., Pool, C. W. & Pach, O. (Elsevier, Amsterdam), pp. 111-120.
14.Rothman, R. B., Schumacher, U. K. & Pert, C. B. (1983) Neu¬ropeptides 3, 493-499.
15.Rothman, R. B. (1986) Alcohol Drug Res. 6, 309-325.
16.McGonigle, P., Neve, K. A. & Molinoff, P. B. (1986) Mol. Phar¬macol. 30, 329-337.
17.Rothman, R. B., Long, J. B., Bykov, V., Jacobson, A. E., Rice, K. C. & Holaday, J. W. (1988) J. Pharmacol. Exp. Ther. 247, 405-416.
18.Lowry, O. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J.
(1951) J. Biol. Chem. 193, 265-275.
19.Burstein, S. H., Hull, K., Hunter, S. A. & Latman, V. (1988) FASEB J. 2, 3022-3026.
20.Nye, J. S., Seltzman, H. H., Pitt, C. G. & Snyder, S. H. (1985) J. Pharmacol. Exp. Ther. 234, 784-791.
21.Zezula, J., Comes, R., Probst, A. & Palacios, J. M. (1988) Neuro¬science 25, 771-795.
22.Boyson, S. J., McGonigle, P. & Molinoff, P. B. (1986) J. Neurosci. 6, 3177-3188.
23.Richfield, E. K., Young, A. B. & Penney, J. B. (1986) Brain Res. 383, 121-128.
24.M. Greenamyre, J. T., Young, A. B. & Penney, J. B. (1984) J. Neu-rosci. 4, 2133-2144.
25.Garrett, E. R. & Hunt, C. A. (1974) J. Pharm. Sci. 63, 1056-1064.
26.Harris, L. S., Carchman, R. A. & Martin, B. R. (1978) Life Sci. 22, 1131-1138.
27.Roth, S. H. & Williams, P. J. (1979) J. Pharm. Pharmacol. 31, 224-230.
28.U. Worley, P. F., Baraban, J. M., De Souza, E. B. & Snyder, S. H.
(1986) Proc. Natl. Acad. Sci. USA 83, 4053-4057.
29.Worley, P. F., Baraban, J. M. & Snyder, S. H. (1986) J. Neurosci. 6, 199-207.
30.Dohlman, H. G., Caron, M. G. & Lefkowitz, R. J. (1987) Biochem¬istry 26, 2257-2264.
31.Miller, L. L. (1984) in The Cannabinoids: Chemical, Pharmaco¬logic, and Therapeutic Aspects, eds. Agurell, S., Dewey, W. L. & Willette, R. E. (Academic, New York), pp. 21-46.
32.Miller, L. L. & Braconnier, R. J. (1983) Psycho!. Bull. 93, 441-456.
33.Aigner, T. G. (1988) Psychopharmacology 95, 507-511.
34.Douglas, R. J. (1967) Psycho!. Bull. 67, 416-442.
35.Eichenbaum, H. & Cohen, N. J. (1988) Trends Neurosci. 11, 244-248.
36.Hooker, W. D. & Jones, R. T. (1987) Psychopharmacology 91, 20-24.
37. Schulze, G. E., McMillan, D. E., Bailey, J. R., Scallet, A., All, S. F., Slikker, W. J. & Paule, M. G. (1988) J. Pharmacol. Exp. Ther. 245, 178-186.
38.Campbell, K. A., Foster, T. C., Hampson, R. E. & Deadwyler, S. A. (1986) J. Pharmacol. Exp. Ther. 239, 941-945.
39.Kornetsky, C. (1985) NIDA Res. Monogr. 62, 30-50.
Q. Di Chiara, G. & Imperato, A. (1988) Proc. Natl. Acad. Sci. USA 85, 5274-5278.
41.Ng Cheong Ton, J. M., Gerhardt, G. A., Fnedemann, M., Etgen, A. M., Rose, G. M., Sharpless, N. S. & Gardner, E. L. (1988) Brain Res. 451, 59-68.
42.Stark, P. & Dews, P. B. (1980) J. Pharmacol. Exp. Ther. 214, 124-130.
43.Gardner, E. L., Paredes, W., Smith, D., Donner, A., Milling, C., Cohen, D. & Morrison, D. (1988) Psychopharmacology %,142-144.
44.Herkenham, M., Lynn, A. B., de Costa, B. & Richfield, E. K.
(1989) Soc. Neurosci. Abstr. 15, 905.
45.Marsden, C. D. (1981) in Disorders of Movement, Current Status of Modern Therapy, ed. Barbeau, A. (Lippincott, Philadelphia), Vol. 8, pp. 81-104.
46.Q. Petro, D. J. & Ellenberger, C. E. (1981) J. Clin. Pharmacol. 21, 413s-416s.
47. Clifford, D. B. (1983) Ann. Neurol. 13, 669-671.
Source including Charts, Graphs and Figures: Cannabinoid receptor localization in brain
