Content uploaded by Kenner C Rice
Author content
All content in this area was uploaded by Kenner C Rice on Apr 06, 2015
Content may be subject to copyright.
Vol. 265, No. 1
Printed in U.S.A.
ABSTRACT
218
0022-3565/93/2651-0218$03.OO/O
THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTiCS
Copyright C 1993 by The American Society for Pharmacology and Experimental Therapeutics
Cannabinoid Structure-Activity Relationships: Correlation of
Receptor Binding and in Vivo Activities1
DAVID A. COMPTON, KENNER C. RICE, BRIAN R. DE COSTA, RAJ K. RAZDAN, LAWRENCE S. MELVIN,
M. ROSS JOHNSON and BILLY R. MARTIN
Department of Pharmacology and Toxicology, Medical College of Virginia, Virginia Commonwealth University, Richmond, Wrginia (D.R.C., B.R.M.);
Laboratory of Medicinal Chemistry, Nationallnstitute of Diabetes and DigestWe and Kidney Diseases, National Institutes of Health,
Bethesda, Maryland (K.C.R., B.R.D.); Organix, Inc., Woburn, Massachusetts (R.K.R.); Pfizer Inc., Groton, Connecticut (L.S.M.) and Glaxo Inc.,
Research Triangle Park, North Carolina (M.R.J.)
Accepted for publication November 23, 1992
Although a receptor exists for cannabinoid drugs, it is uncertain
which pharmacological actions this receptor mediates. This
structure-activity relationship investigation was initiated to deter-
mine which effects might correspond to binding affinity for the
cannabinoid receptor, as well as to explore the binding require-
ments of this site. The ability of nearly 60 cannabinoids to
displace [3HJCP-55,940 {(-)-3-[2-hydroxy-4-(1 ,1 -dimethylhep-
tyl)phenylj-4-[3-hydroxy propyl] cyclohexan-1 -oI’ was deter-
mined before establishing correlations between receptor affinity
and in vivo pharmacological potency. Analysis of [3H]CP-55,940
binding indicated a Hill coefficient of 0.97, a B, of 499 pM (3.3
pmol/mg of protein) and an apparent Kd of 924 pM. Closer
inspection indicated the binding assay exhibited “zone B” char-
actenstics, and use of correction equations indicated a true Kd
for CP-55,940 of 675 pM. The structure-activity relationship
indicated the importance of side chain structure to high-affinity
binding, with the most potent analogs (Kl < 10 nM) possessing
either a dimethylheptyl side-chain, a similarly complex branched
side chain or a halogen substituent at the 5’ position. Compar-
ative analysis of Km values to in viva potency in a mouse model
indicated a high degree of correlation between parameters for
the depression of spontaneous locomotor activity (r = 0.91) and
for the production of antinociception (r = 0.90), hypothermia (r
= 0.89) and catalepsy (r = 0.85). Similarly high correlations were
demonstrated between binding affinity and in viva potency in
both the rat drug discrimination model (r = 0.81) and for psycho-
tomimetic activity in humans (r = 0.88). Thus, these studies
appear to indicate that the requirements for activation of the
cannabinoid receptor are similar across different species, and
that receptor binding is sufficient to mediate many of the known
pharmacological effects of cannabinoids.
In the time since its isolation from plant material (Gaoni
and Mechoulam, 1964), 9-THC has come to be recognized as
the principle active component of marihuana. Despite the fact
that pharmacological evidence such as SAR suggested the ex-
istence of a receptor through which 9-THC and related can-
nabimimetic drugs might be producing their effects (Binder
and Franke, 1982), the first attempt to show the existence of
such a site by use of radioligand binding techniques failed to
provide convincing evidence that a receptor existed in brain
tissue, although modest information was obtained for binding
to cultured cells (Harris et at., 1978). The first successful
identification of a cannabinoid binding site in brain tissue was
Received for publication July 8, 1992
1 This work was supported by National Institutes on Drug Abuse Grant DA
03672 and by a Pharmaceutical Manufacturers Association Foundation Starter
Grant.
accomplished by use of a water-soluble THC analog, although
binding affinity at this site proved not to be well correlated
with the known behavioral effects of multiple cannabinoids
(Nye et at., 1985, 1989). Subsequently, the 5’-trimethylammon-
ium derivative used as the radioligand was found to be devoid
of the typical pharmacological profile expected of a cannabi-
mimetic (Compton and Martin, 1990). Despite the lack of
evidence for a binding site, a correlation of the analgesic prop-
erties of a series of potent nonclassical cannabinoids to their
ability to inhibit adenylyl cyclase activity in vitro, led to the
formulation of a cannabinoid receptor model (Howlett et at.,
1988). However, it was not until the development of a centrif-
ugation assay using the extremely potent analgesic CP-55,940
as the radioligand that a specific saturable cannabinoid binding
site was shown to exist in brain tissue (Devane et at., 1988).
Unlike the 5’-trimethylammonium radioligand, CP-55,940 has
ABBREVIATIONS: THC, tetrahydrocannabinol; SAR, structure-activity relationship; CBD, cannabidiol; DMHP, dimethylheptyl pyran; TMA, trimethy-
lammonium; DMH, dimethylheptyl; BSA, bovine serum albumin; PEI, polyethylenimine; Ki, dissociation constant determined from inhibition of
radioligand binding; SA, spontaneous activity; TF, tail-flick; AT, rectal temperature; RI, ring immobility; HHC, hexahydrocannabinol; CP-56,667, (+)-
CP-55,940 or (+)-3-[2-hydroxy-4-(1 ,1 -dimethylheptyl)phenyl]-4-[3-hydroxypropyljcyclohexan-1-ol; CP-55,940, (-)-3-[2-hydroxy-4-(1 ,1-dimethylhep-
tyl)phenylj-4-[3-hydroxy propyljcyclohexan-1 -ol
by on May 2, 2009 jpet.aspetjournals.orgDownloaded from
TABLE 1
Structures of bicyclic analogs
Bcyclic Analogs
Ito XIV and XVI
OH Bacyclc Analog
T XV
R,
Mog
Ill
IV
V
VI (CP-47,497)
VII
VIII
IX
X
Xl (Ci-aOH)
XII
XIII
XIV (CP-55,940r
XV (CP-56,667r
XVI
A,
H
H
H
H
H
H
H
H
H
H
-CH3
-CH3
-(CH2)OH
-(CH2)OH
-(CH2),OH
-(CH2)OH
CannabinoidSAR 219
1993
been shown to possess a profile of pharmacological activity
similar to that of 9-THC (Little et aL, 1988). Localization of
the [3HJCP-55,940 binding site using autoradiographic tech-
niques revealed significant similarities among rat, monkey and
human brain tissue, whereas limited SAR suggests some cor-
relation between in vitro displacement potency and in vivo
behavioral potency (Herkenham et at., 1990). These studies
contributed in part to the evaluation of a previously cloned rat
brain protein, which was subsequently expressed in various cell
lines as a fully functional cannabinoid receptor capable of
inhibiting adenylyl cyclase activity (Matsuda et aL, 1990), al-
though no radioligand binding data were presented in these cell
lines. Since then the human cannabinoid receptor has also been
cloned (Gerard et at., 1991).
Although it is now well established that a receptor exists for
the cannabinoids, it is unclear whether this receptor is respon-
sible for all of the centrally mediated actions of the cannabi-
noids. This SAR investigation was initiated to determine
whether affinity for this receptor correlates with any or all of
the pharmacological effects of the cannabinoids. The ability of
nearly 60 cannabinoids to displace [3HJCP-55,940 from its
binding site in a membrane preparation was determined via a
filtration assay in order to evaluate the structural requirements
for receptor recognition. Additionally, all of these compounds
have been evaluated previously in one or more models for in
vivo potency. Thus, binding data were also determined in order
to establish whether a significant correlation existed between
the affinities of these drugs for the binding site and their in
vivo pharmacological activities.
Materials and Methods
Animals. Male Sprague-Dawley rats (150-200 g) obtained from
Dominion Laboratories (Dublin, VA) were maintained on a 14:10-hr
light/dark cycle, and received food and water ad libitum.
Drugs and chemicals. [3H]CP-55,940 was prepared by tnitium
reduction, in the presence of a Pd catalyst, of an analog possessing a
double bond located between the C2 and C3 carbons of the dimethyl-
heptyl side chain (the R substituent depicted in table 1). 8-THC, .
THC, ‘-THC, 11-OH-8-THC, 11-OH-9-THC, CBD, 1’,2’-DMHP
(6.-1Os), 5’-OH-9-THC and 11-OH-5’-TMA-8-THC were obtained
from the National Institutes on Drug Abuse. 1-O-Carbamoyl-9-THC
was provided by the National Cancer Institute; nabilone by Eli Lilly
and Company (Indianapolis, IN); 5’-Bn-8-THC, 5’-I-8-THC, 5’-F3-
8-THC, 5’-F-8-THC, 2-I-5-THC and 11-F-8-THC by Dr. Alexan-
dros Makriyannis (University of Connecticut, Stoma, CT); 0,2-pro-
pamo-8-THC, 9a-OH-HHC, 93-OH-HHC and 0,10-methano-8-THC
by Dr. Patricia H. Reggio (Kennesaw State College, Marietta, GA) and
(+)- and (-)-11-0H-8-THC-DMH (HU-210 and HU-211, respec-
tively) by Dr. Rafael Mechoulam (Hebrew University, Jerusalem, Is-
rael).
-THC-DMH, 8fl-0H-9”-THC, 1-0-methyl-9”-THC, 8-(N-
morpholino)-amino-9”-THC, 3-(norpentyl)-3-propyl-9”-THC, 1-
O-(6-aminohexyl)-8-THC, 1-0-methyl-8-THC, 1-0-(2-morpholinoe-
thyl)-8-THC, 1-0-methyl-9-THC, 1-0-biphenylmethy1-8-THC, 1-
O-biphenylmethy1-9”-THC, 1-0-(4-phthalimidobutyl)-8-THC, 1-0-
methyl-3-norpentyl-3-propyl-9”-THC, 1-O-(4-aminobutyl)-8-THC,
9-nor-9-carboxy-8-THC acid, 1-0-(3-aminopropyl)-8-THC, 9-non-9-
carboxy--THC acid and abnormal-CBD (where the Cl phenolic
hydroxy and C3 pentyl side chain are transposed) were synthesized at
Organix, Inc. (Woburn, MA).
Analogs I through XV (depicted in table 1), which includes CP-
55,940 and its (+) isomer CP-56,667 as well as t,d-nantradol were
synthesized at Central Research, Pfizer Inc.
Membrane preparation. The methods for tissue preparation were
__________________ ____________ A2
-C(CH3)CH3
-C(CH3)CH2CH3
-C(CH3)(CH2)CH3
-C(CH3)S(CH2),CH3
-C(CH3)(CH2)4CH3
-C(CH3),(CH2)SCH3
-C(CH3)(CH2MCH3
-C(CH3MCH2)TCH3
-C(CH3)(CH,)CH3
-C(CH3)(CH2)CH3
-C(CH3)(CH2)SCH3
-C(CH3)(CH2)SCH3
H
-C(CH3MCH2)SCH3
-C(CH3MCH2)SCH3
-C(CH3)(CH2)SCH3
sb Previously reported as (-)-AC and (+)-ACisomers (Little et al., 1988). The
numbering system of these nondassical cannaheeds is proded for comparison
to that given in Figure 2.
those described by Devane et at. (1988). After decapitation and the
rapid removal of the brain, the cortex was dissected free using visual
landmarks following reflection of cortical material from the midline.
The cortex was immersed in 30 ml of ice-cold centrifugation solution
(320 mM sucrose, 2 mM Tnis. EDTA, 5 mM MgCl2). The process was
repeated until the cortices of five rats were combined. The cortical
material was homogenized with a Kontes Potten-Elvehjem glass-Teflon
grinding system (Fisher Scientific, Springfield, NJ). The homogenate
was centrifuged at 1600 X g for 10 mm, the supernatant saved and
combined with the two subsequent supernatants obtained from washing
(and 1600 x g centrifugation) of the P1 pellet. The combined supenna-
tant fractions were centrifuged at 39,000 x g for 15 mm. The P2 pellet
was resuspended in 50 ml of buffer A (50 mM Tnis. HC1, 2 mM Tnis.
EDTA, 5 mM MgCI2, pH 7.0), incubated for 10 mm at 37”C, then
centrifuged at 23,000 x g for 10 mm. The P2 membrane was resuspended
in 50 ml of buffer A, incubated again except at 30#{176}Cfor 40 mm, then
centrifuged at 11,000 x g for 15 mm. The final wash-treated P2 pellet
was resuspended in assay buffer B (50 mM Tris . HCI, 1 mM Tris.
EDTA, 3 mM MgC12, pH 7.4) to a protein concentration of approxi-
mately 2 mg/ml. The membrane preparation was divided into four
equal aliquots and quickly frozen in a bath solution of dry ice and 2-
methylbutane (Sigma Chemical Co., St. Louis, MO), then stored at
-80”C for no more than 2 weeks. Before performing a binding assay,
an aliquot of frozen membrane was rapidly thawed and protein values
determined by the method ofBradford (1976) using Coomassie brilliant
blue dye (Bio-Rad, Richmond, CA) and BSA standards (fatty acid free,
Sigma Chemical Co.) prepared in assay buffer.
Binding assay. The methods for nadioligand binding were essen-
tially those described by Devane et at. (1988) with the exception that
bound and free drug were separated by filtration rather than centrifu-
gation. Binding was initiated by the addition of 150 zg of P2 membrane
to test tubes containing [3HJCP-55,940 (79 Ci/mmol), a cannabinoid
analog (for displacement studies) and a sufficient quantity of buffer C
(50 mM Tris.HC1, 1 mM Tris.EDTA, 3 mM MgCl2, 5 mg/ml BSA) to
bring the total incubation volume to 1 ml. The concentration of [3H]
CP-55,940 in displacement studies was 400 pM, whereas that in satu-
ration studies varied from 25 to 2500 pM. Similar results were obtained
with preliminary saturation studies in which the [3H]CP-55,940 con-
centrations varied from 25 to 5000 pM. Nonspecific binding was deter-
by on May 2, 2009 jpet.aspetjournals.orgDownloaded from
0 100 200 300 400 500 600
Bound (fmol)
220 Compton et al.
Vol. 265
mined by the addition of 1 iM unlabeled CP-55,940. CP-55,940 and all
cannabinoid analogs were prepared by suspension in buffer C from a 1
mg/ml ethanolic stock without evaporation of the ethanol (final con-
centration of no more than 0.4%).
After incubation at 30#{176}Cfor 1 hr, binding was terminated by addition
of 2 ml of ice-cold buffer D (50 mM Tnis . HC1, 1 mg/mi BSA) and
vacuum filtration through pretreated filters in a 12-well sampling
manifold (Millipore, Bedford, MA). Reaction vessels were washed once
with 2 ml of ice-cold buffer D, and the filters washed twice with 4 ml
of ice-cold buffer D. Filters were placed into 20-mi plastic scintillation
vials (Packard, Downer’s Grove, IL) with 1 ml of distilled water and
10 ml of Budget-Solve (RPI Corp., Mount Prospect, IL). After shaking
for 1 hr, the radioactivity present was determined by liquid scintillation
spectrometry.
Assay conditions were performed in triplicate, and results represent
the combined data of three to six independent experiments. All assays
were performed in siliconized test tubes, which were prepared by air
drying (12 hr) the inverted borosilicate tubes after two rinses with a
0.1% solution of AquaSil (Pierce, Rockford, IL). The GF/C glass-fiber
filters (2.4 cm, Baxter, McGaw Park, IL) were pretreated in a 0.1%
solution of pH 7.4 PEI (Sigma Chemical Co.) for at least 6 hr.
Data analysis. The Bmu and Kd values obtained from Scatchard
analysis (Scatchard, 1951; Rosenthal, 1967) values were determined via
the KELL package of binding analysis programs for the Macintosh
computer (Biosoft, Milltown, NJ). The software represents a modifi-
cation of the original descriptions of LIGAND (Munson and Rodbard,
1980) and Equilibrium Binding Data Analysis (McPherson, 1983).
Displacement ICo values were originally determined by unweighted
least-squares linear regression of log concentration-percent displace-
ment data and then converted to Kz values using reported methods
(Chengand Prusoff, 1973). Statistical evaluation ofparallelism between
displacement isotherms was performed using ALLFIT (De Lean et aL,
1987), a program for the simultaneous curve fitting of a family of
sigmoidal curves. Statistical analysis and generation of the Pearson
product-moment coefficients for the correlation studies was performed
on the Macintosh computer using the StatView 512+ statistical package
(Brainpower, Inc., Agoura Hills, CA).
Results
Assay characteristics. Because this filtration assay repre-
sented a modification of the centrifugation procedure, all in-
cubation conditions were evaluated before initiation of satura-
tion analysis. The optimal incubation conditions were found to
be similar to those previously described (Devane et al., 1988).
Thus, the results for characterization are not shown. Unlike
the previously published method, nonspecific binding was de-
fined as the binding observed in the presence of 1 MM CP-
55,940. Specific binding averaged 90% in the displacement
studies where radioligand concentration was 400 pM. The
protein dependency of the assay was determined using the
standard assay conditions for all preparations and incubations
with [3HJCP-55,940 having been arbitrarily held at 400 pM.
Similar to the previously published assay, binding was found
to be linear from 20 to 50 g of protein per tube, but there was
a high degree of variability in the binding at these low protein
levels. However, variability in response was minimized at a
higher protein range, and binding was linear (r = 0.95) from 20
to 200 ofprotein. There was a suggestion of a deviation from
linearity above 175 tg of protein, so subsequent assays were
conducted using 150 tg of protein per tube. The time course of
binding was evaluated under standard conditions using 400 pM
[3H]CP-55,940. Binding increased from 10 to 50 mm, and
maximal binding was maintained through 3 hr suggesting no
degradation of radioligand or receptor during this period. Thus,
subsequent assays were conducted with an incubation time of
1 hr. The temperature dependence of binding was determined
under otherwise standard conditions using 400 pM [3HJCP-
55,940. Total binding at 25 and 35#{176}Cwas 66 and 71%, respec-
tively, ofthat determined at 30#{176}C,whereas at 40#{176}Ctotal binding
decreased tremendously and was highly variable. Thus, the
incubation temperature of 30”C was used in subsequent assays.
Treatment of filters with PEI reduced nonspecific binding, as
did washing filters with two 4-ml rinses of buffer. Total binding
was greater than 10% of total radioactivity added, a value
normally considered the maximum for Scatchard analysis.
However, when protein concentrations were reduced to 50 zg
and total binding maintained at less than 10% of total radio-
activity added, then the Kd value derived from saturation analy-
sis was similar to that found using 150 g of protein. Unfortu-
nately, binding at 50 g of protein proved highly variable. The
impact of this phenomenon (>10% of total radioactivity bound)
on saturation and displacement analysis is addressed under
“Discussion.”
Scatchard analysis. Independent saturation analyses con-
ducted over a period of 1 year have resulted in reproducible
results. An example of the Scatchard analysis of a single
independent experiment is shown in figure 1 with the saturation
curve shown as an inset. Computer analysis of saturation data
(n = 5, simultaneous analysis with LIGAND) indicated a Kd of
924 ± 140 pM and a B,,,55 of 499 ± 60 pM (3.3 ± 0.6 pmol/mg
protein) and a Hill coefficient of 0.97 ± 0.01.
Because cannabinoids bind to albumin (Garrett and Hunt,
1974; Haque and Poddar, 1984), BSA can be included in buffers
to act as the cannabinoid carrier while in an aqueous medium.
To evaluate the potential interference of BSA binding on
receptor binding, the quantity of BSA present in the assay
buffer (buffer C) was reduced from 5 mg/ml to 2.5 and 1.0 mgI
ml in order to determine whether there was an effect on
Scatchard analysis. At a concentration of 2.5 mg/ml of BSA
the EBDA/LIGAND analysis (n = 2) indicated a Kd of 1160 ±
260 pM, a Bmax of 705 ± 117 pM and a Hill coefficient of 0.99
± 0.01. At a concentration of 1.0 mg/ml of BSA the EBDA/
LIGAND analysis (n = 2) indicated a Kd of 932 ± 102 pM, a
I
Fig. 1. Scatchard analysis of [3H]CP-55,940 receptor binding. This
saturation experiment (inset) and resulting Scatchard analysis is repre-
sentative of the binding obtained with rat cortical tissue. In this particular
experiment the Kd was found to be 799 pM and the B equal to 4.0
pmol/mg of protein.
by on May 2, 2009 jpet.aspetjournals.orgDownloaded from
11-OH-Et8-THC-DMH
1993
Cannabinold SAR 221
Bmax of 661 ± 49 pM and a Hill coefficient of 0.98 ± 0.01. Thus,
displacement assays were performed using the 5 mg/ml concen-
tration of BSA.
Drug displacement. A primary focus of this investigation
was the determination of the SAR for binding to the cannabi-
noid receptor by evaluating analogs with as much structural
diversity and widely varying potencies as possible. A series of
16 nonclassical bicyclic cannabinoids were evaluated in dis-
placement studies, and as these analogs were previously iden-
tified by roman numerals (Compton et at., 1992), the same
designation was used in these studies. The structure of CP-
55,940 and related bicyclic cannabinoids are shown in table 1.
Naturally occurring analogs chosen for displacement studies
included 9-THC and CBD (fig. 2). Many analogs evaluated
possessed a side chain other than the typical straight chain
pentyl observed for 9-THC. A common side chain modification
that often results in a more potent cannabinoid is the substi-
tution of a DMH side chain (Razdan, 1986; Little et al., 1989).
One example of such an analog chosen for evaluation in these
studies was 11-OH-8-THC-DMH (fig. 2). Other analogs in-
cluded were derivatives ofthe exocyclic cannabinoid 9”-THC
(fig. 2).
The calculated KI values obtained from experimentally de-
termined IC values for 59 different compounds are listed in
table 2. The Ki values obtained ranged from 728 pM to more
than 4 tM. The maximum observed displacement averaged
approximately 90% for all compounds with Ki values of less
than 1 tiM, and varied from a 64 to 86% (for CBD and (+)-11-
OH-z8-THC-DMH, respectively) for compounds with Ki values
greater than 1 zM. Ki values could not be obtained for 16 of
the compounds, all of which failed to displace at least 50% at
10 M drug. The displacement curves for a representative
sample of drugs are shown in figure 3. Statistical analysis
(ALLFIT; P < .05) indicated that the displacement curves of
most analogs were parallel to that of 9-THC and CP-55,940.
This suggested competitive inhibition of radioligand binding
(rather than noncompetitive inhibition) is an indication that
receptor recognition occurs in a similar fashion for most of the
cannabinoid drugs, suggesting binding at a single site on the
receptor. However, not all displacement curves could be shown
to be parallel to that of 9-THC, which is described in more
detail later (see under “Discussion”).
Correlation studies. The compounds showing a measurable
affinity for the cannabinoid receptor were compared to their
abilities to produce pharmacological effects in the mouse (fig.
4) as well as the rat and human (fig. 5). All ED values were
converted to units of micromoles per kilogram before conduct-
ing correlation studies. Human potency values were previously
expressed relative to 9-THC (=1.0), and (as with ED values)
the smaller the number, the more potent the compound. Be-
cause CBD has long been recognized to be devoid of the
psychoactive properties found with 9-THC and demonstrated
to have a profile of pharmacological activity different from that
of \9-THC (Compton et aL, 1990), this compound was not
included in any of the correlation analyses. Similarly, seven
other relatively weak analogs were not included in any corre-
lation analysis because these compounds possessed ED50 values
in the mouse that were greater than 100 mg/kg (the highest
dose evaluated) and data were not available in either the rat or
the human.
The potency of 29 analogs was previously determined in
multiple behavioral procedures in the mouse (Martin et aL,
1987; Little et at., 1988, 1989; Compton et aL, 1989, 1990, 1992;
Compton and Martin, 1990; Reggio et al., 1990, 1991; Charalam-
bous et aL, 1991). Not all compounds were active in all four
evaluations which included measures of SA, antinociception
via a TF assay, hypothermia via RT evaluation and a RI or
catalepsy measure. The log ED (tmol/kg) value was plotted
against the corresponding log Ki (nM) value for each analog
(fig. 4). The solid line in each plot represents the linear regres-
sion of that data set. The correlation coefficients for SA, TF,
RT and RI evaluations were 0.91, 0.90, 0.89 and 0.85, respec-
tively. All correlations were statistically significant at the P <
.01 level (two-tailed) based on the Pearson product-moment
values. The dashed line in each plot depicts the linear regression
that was obtained when the KI value for each drug was plotted
against the mean ED value obtained from averaging the
potencies obtained in all four mouse procedures. Although there
is a noticeable divergence between the mean behavioral potency
and the actual potency of drugs in the locomotor (SA) and
antinociception (TF) paradigms, it is also obvious that there is
little difference in the temperature (RT) and immobility (RI)
paradigms. It is clear that for general discussions of the mouse
model, the single mean ED value (not given) is a much less
cumbersome, yet still a relatively accurate, indicator of drug
potency.
Besides establishing correlations between receptor affinity
and general pharmacological activity in the mouse model, it
was of great interest to determine whether correlations also
existed between receptor affinity and potency in the rat drug
Fig. 2. Structures of cannabinoids. 9-THC is numbered
according to the dibenzopyran nomenclature, and related
analogs given in this report (including 8-THC derivatives)
have been named similarly. In contrast, CBD is typically
numbered according to the monoterpinoid nomenclature.
Also shown is the extremely potent analog 11-OH-5-THC-
DMH, as well as the exocyclic cannabinoid analog
THC.
by on May 2, 2009 jpet.aspetjournals.orgDownloaded from
-4
222 Compton et aI.
Vol. 265
TABLE 2
K, values of cannabinoid analogs
Analog K, ±S.E. Analog K, ±S.E.
nM nM
(-)-1 1-OH-8-THC-DMH 0.728 0.1 13 8-THC
126 16
CP-55,940 [(-)-AC, XIV] 0.924
0.140 IX 163 35
I-Nantradol
1 .06 0.24 9a-OH-HHC 171 8
XVI 1 .55 0.85 911-THC 236 55
1’,2’-DMHP (6a-1Oa)
1.56 0.08
X
381 28
8-THC-DMH 1 .59 0.24 IV 735 14
VII 4.73 1 .34 8-OH-911-THC 820 85
XI
5’-Br-8-THC
6.15
7.63
1 .99
1 .41
1-O-methyI-911-THC
8-(N-morpholino)-amino-i9’1-THC
827 71
869 71
XII
5’-I-8-THC
VI
5’-Fr-THC
7.70
7.77
9.54
19.9
2.05
2.40
0.35
0.9
3-(NorpentyI)-3-propyl-911-THC
1-O-(6-Aminohexyl)-i8-THC
(+)-11-OH-8-THC-DMH
1 1-OH-5’-TMA-8-THC
968 35
1730 150
1990 1430
21 10 460
Nabilone
VIII
22.3
28.5
6.6
3.3
1-O-MethyI-t8-THC
1-O-(2-MorphoIinoethyl)-8-THC
2170 110
3040 460
11-OH--THC 38.4 0.8 d-Nantradol 3100 550
9-THC
40.7 1 .7
III
3760 70
1 1-OH-8-THC 54.9 10.2 CBD
4350 390
5’-F--THC
CP-56,667; (+)-AC, XV
57.0
61.7
2.3
5.0
1-O-Methyl-9-THC
O,2-Propano-z8-THC
>10,000
>10,000
5’-OH-9-THC
87.6 6.4 I >10,000
2-I--THC 89.0 15.5 II >10,000
1 1-F-’-THC 107 28 XIII >10,000
93-OH-HHC
V
124
126
11
9
1-O-biphenylmethyI-8-THC
1-O-biphenylmethyl-19’1-THC
>10,000
>10,000
1-O-(4-PhthalimicIobutyI)--THC
>10,000
1-O-Methy3-(norpentri)-
3-PropyI-9’1-THC
1-O-(4-Aminobutyl)--THC
9-Nor-9-carboxy-8-THC (acid)
1-O-(3-AminopropyI)-6-THC
9-Nor-9-Carboxy-9-THC (acid)
O,10-Methano-9-THC
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
1-O-Carbamoyl-9-THC >10,000
Abnormal-CBD (abn-CBD) >10,000
100
80
.40
.
60
20
0
-10 -8 -6
Log Drug Concentration
Fig. 3. Displacement of bound [3H]CP-55,940 by various cannabinoid
analogs.
discrimination model as well as in previously reported psy-
choactivity values in humans A correlation analysis was per-
formed using analogs for which drug discrimination data in the
rat model was available (Razdan, 1986; Balster and Prescott,
1992). These analogs were 1-nantradol, z8-THC-DMH, CP-
47497 (or VI), nabilone, 11-OH-8-THC, 11-OH-9-THC,
THC and 9-THC. The linear regression of the log values of
Ki vs. rat drug discrimination potency is shown in the upper
panel of figure 5. The correlation coefficient was 0.81, which
was statistically significant at the P < .01 level (two-tailed).
Despite the potential for species differences, and because of the
similarity between the cloned rat (Matsuda et at., 1990) and
human cannabinoid receptors (Gerard et at., 1991), a correlation
analysis was performed on the analogs in this study for which
data on the behavioral potency in humans was available (Raz-
dan, 1986). These analogs were l-nantradol, DMHP, nabilone,
11-OH-8-THC, 11-OH-z9-THC, 8-THC and 9-THC. The
correlation coefficient of the regression of the log values of KI
vs. human potency data (fig. 5) was 0.88, which also proved to
be statistically significant at the P < .01 level (two-tailed).
Discussion
The results presented in this manuscript clearly indicate that
behavioral potency of cannabinoids in the mouse can be pre-
dicted by establishing the affinity of the cannabinoid for the
receptor labeled by [3HJCP-55,940. The method chosen for
demonstrating this correlation of in vitro receptor affinity to
in vivo agonist potency was linear regression analysis of log-log
data, as previously used for both dopamine (Seeman et at., 1978;
Seeman, 1980) and opiate receptors (Loh et aL, 1978), rather
than via nongraphic statistical correlations which generally
only establish simple rank ordering (Stahl et aL, 1977; Furch-
gott, 1978). Drugs with Ki values in the 0.8 to 5.0 iM range
were either inactive at i.v. doses up to 100 mg/kg, showed
by on May 2, 2009 jpet.aspetjournals.orgDownloaded from
3
2
-1
-2
3
2
-1
-2
*2
C
.! i
I-
0
-1
a -2
3
C
a
C#{176}
a
E
0
-1 0 1 2 3 4
1993
CannabinoidSAR 223
3
2
Ri
3
2
0
-2
-1 0 1 2 3 4
Log
Fig. 4. Linear correlations of log EDse values (mol/kg) for cannabinoids
and the log Kt values given in table 2. The potency of these analogs
administered i.v. in the mouse on four pharmacological measures have
been reported previously (see under “Results”), and include reduction of
SA, antinociception TF procedure, hypothermia as measured via AT and
immobility (or catalepsy) in the RI measure. The dashed line represents
the linear regression that would be obtained if the mean EDse value of
the four measures was plotted instead of that shown for each individual
behavior.
minimal activity or only showed activity in a limited number
of evaluations. Drugs with 1<1 values in the 1 to 100 pM range
proved potent in the production of hypoactivity, hypothermia,
antinociception and catalepsy. Similarly, drugs with Ki values
in the 100 to 800 nM range also generally proved moderately
potent in the production of all four mouse measures, although
there were minor exceptions. Although it would seem highly
probable that these four distinct pharmacological effects are
mediated by different neuronal pathways/mechanisms, the po-
tencies of cannabinoids across all four results were very similar.
This point was demonstrated by the fact that the mean or
average potency of the cannabinoids could be correlated with
receptor affinity just as well as potency in any individual
behavioral evaluation. However, it must be cautioned that this
generalization may not necessarily apply to behavioral effects
of cannabinoids not reported here. For example, it should not
be concluded that cannabinoid receptor affinity is predictive of
Log K
Fig. 5. Linear correlation of log KI values to the log EDse values (mol/
kg) for cannabinoids in the rat drug discrimination procedure, and log Ki
values to the relative potencies of analogs to 9-THC. The scale chosen
is that used in figure 3. In vivo data for rats and humans were obtained
from previously published results (see under “Results”).
other behavioral effects of cannabinoids such as anxiolytic,
anticonvulsant and antiemetic actions until appropriate studies
have been conducted. Lastly, it is interesting to note that
behavioral results in the mouse model were highly correlated
with receptor affinity, despite the fact that 1(1 values were
generated by using rat brain tissue and by using only cortical
tissue (although it is likely that some ofthe behaviors evaluated
are dependent upon neural paths lying outside of the cortex).
These findings suggest that any drug with affinity for the
cannabinoid receptor will apparently produce the whole spec-
trum of cannabinoid pharmacological effects described here.
These data are consistent with the fact that development of a
cannabinoid with selectivity of pharmacological action has been
difficult to attain. It may be difficult to design drugs which act
via this receptor while producing only one pharmacological
effect, such as analgesia or sedation or antiemesis.
Although the mouse model of pharmacological activity has
been used to determine whether a drug possesses cannabimi-
metic properties, it has not generally been considered to be
highly predictive of psychoactivity in humans. However, the
rat-drug discrimination model has been widely considered in-
dicative of drugs which would produce cannabimimetic actions
in man (Balster and Prescott, 1992). Based upon the com-
pounds presented here, this study provides evidence that in
vitro cannabinoid receptor affinity is predictive of in vivo
potency in the rat drug discrimination model. This correlation
suggests that those properties of cannabinoids by which dis-
crimination is possible are produced via activation of the recep-
tor labeled by [3H]CP-55,940.
A strong positive correlation was also obtained between
binding affinity and psychoactivity in humans. This suggests
that in vitro cannabinoid receptor affinity is predictive of
psychoactivity in humans. Although this conclusion is based
upon a limited number of compounds, it seems likely that
potency at the cannabinoid receptor in one species is indicative
of potency in another species, and activation of this receptor is
responsible for the production of psychoactivity in humans as
well as the behavioral effects in rodents. These correlations
by on May 2, 2009 jpet.aspetjournals.orgDownloaded from
224 Comptonetal.
VoL 265
suggests that either rodent model may be used to determine
whether novel cannabinoid compounds might possess the psy-
choactive properties for which 9-THC is abused by humans.
This study represents the first comprehensive report on the
SAR of cannabinoid receptor binding. These data support the
SAR of cannabinoids as described by in vivo pharmacological
results. Affinity for the cannabinoid receptor is greatly in-
creased by altering the side chain. The structures of 8-THC
and 9-THC include a pentyl side chain at the C3 position.
However, increasing the side chain to a seven-carbon atom
length and branching the side chain by adding two methyl
groups at the first carbon of the backbone, will produce a DMH
derivative with a greatly increased affinity. In fact, ofthe nearly
60 compounds evaluated, there are 12 compounds with Ki
values of less than 10 nM, and eight of these contain a DMH
substituent. Two of the other drugs also possess side chains
that are branched, long carbon backbones. One is levonantradol
and the other (VII) possesses a dimethyloctyl side chain. Only
two of the 12 drugs possess the typical pentyl side chain.
Interestingly, both of these cannabinoids are halogenated (Br
or I) at the 5’ (terminal) end of the side chain. It should also
be pointed out that some restrictions do apply to the approach
of altering the carbon length of the side chain to increase
receptor affinity. Within the series of nonclassical bicyclic
cannabinoids, increasing the dimethyl side chain carbon back-
bone to 9, 10 or 1 1 carbons as in analogs VIII, IX and X,
respectively, diminishes receptor affinity. Similarly, decreasing
the backbone length to 6, 5, 4, 3 or 2 carbons, or removal of the
side chain as in analogs V, IV, III, II, I and XIII, respectively,
also reduces receptor affinity. Receptor affinity is reduced from
236 nM for 9”-THC to 968 nM when the pentyl side chain
is shortened to a propyl side chain. Thus, these data specify
the requirements for cannabinoid side chain length and/or
branching for receptor affinity.
Affinity for the cannabinoid receptor is increased by halo-
genation of the 5’ end of the side chain. Either bromination or
iodination produced analogs in the 8-THC series with KI
values of approximately 8 nM. Both the 5’-fluoro and 5’-
trifluoro analogs were somewhat less potent. The only other
drugs in this study with substituents at the 5’ position were 5’-
OH-9-THC and 11-OH-5’-TMA-z8-THC. Although the 5’-
OH analog showed good affinity, the trimethylammonium an-
alog possessed very weak affinity for the cannabinoid receptor
labeled by [3HJCP-55,940. The finding that the trimethylam-
monium analog lacks strong affinity for the CP-55,940 binding
site is consistent with the binding profile and pharmacological
activity of the parent compound 5’-TMA-8-THC (Nye et aL,
1985; Compton and Martin, 1990).
These binding data also confirm the enantiospecificity antic-
ipated based upon in vivo pharmacological results. The affinity
ratio between 1,d-nantradol was greater than 2900, and that
between (+)- and (-)-11-OH-8-THC-DMH was greater than
2700, although that between CP-55,940 and CP-56,667 was a
value of only 67. These findings are also consistent with the
observations that these weak/inactive stereoisomers have not
proven to be antagonists (Little et aL, 1988).
The expected profile of a cannabinoid antagonist would be a
compound with high receptor affinity and no in vivo agonistic
activity. Such a compound would be expected to block the
agonistic actions of a cannabinoid by competitively binding to
the receptor. Besides the inactive stereoisomers just discussed,
several other compounds have previously been found to be
inactive in in vivo pharmacological evaluations. Almost all of
these inactive analogs have been evaluated for antagonist prop-
erties and, like the inactive stereoisomers mentioned above,
have not been found to be capable of blocking or reducing the
in vivo effects of 9-THC. The primary reason now appears to
be due to the fact that these inactive analogs simply do not
bind to the receptor, as evidenced by the fact that most do not
displace the radioligand even at 10-tiM concentrations. How-
ever, some inactive compounds were found to possess measur-
able binding affinities, but those drugs had affinities in the 0.8
to 4.0 M range, and drugs with such weak affinity would not
necessarily be expected to be effective antagonists at the doses
normally used. Interestingly, the carboxylic acid metabolite of
9-THC has been reported to antagonize some effects of -
THC (Burstein et aL, 1987), but it does not appear to bind to
the cannabinoid receptor. Similarly, the acid metabolite of
THC also does not bind to the cannabinoid receptor, which
presumably would explain its inability to antagonize the effects
of 9-THC (unpublished results).
The K,, value of 924 pM obtained here was somewhat larger
than the 133 pM value obtained initially by Devane et at.
(1988), but very close to the 950 pM value reported subse-
quently by the same laboratory using a filtration assay (Hous-
ton et at., 1991). However, the Bmax value obtained in this
investigation was larger than either the 1.85 or the 0.94 pmol/
mg of protein values reported previously (Devane et at., 1988;
Houston et aL, 1991). However, use of traditional Scatchard
methods or saturation analysis of radioligand binding data
requires (among other conditions) that total binding be less
than 10% of the total radioactivity added, otherwise the ap-
proximation that “free” radioligand be set equal to “total radi-
oligand added” is unacceptable. This condition can occur if
there is a very high concentration of receptor present under
the assay conditions. However, if receptor concentration cannot
be assumed to be negligible, then classical conditions are vio-
lated and the traditional approach to determining K,, and Bmaz
are inappropriate. Specifically, if the ratio Bmax/Kd 5 greater
than [1/10] then the assay shifts from “zone A” behavior, where
Scatchard analysis is acceptable, to “zone B” behavior, where
alternate equations describing ligand binding must be used
(Goldstein, 1949; Boeynaems and Dumont, 1980). In this report
the Bmaz/Kd ratio equals 499/924 (or 0.54), thus corrected
equations for saturation and displacement studies should have
been used. Interestingly, although the previously described
binding assay by Devane et at. (1988) exhibited the identical
characteristic (a Bm.x/Kd ratio of 92/133 = 0.69), the autoradi-
ographic study by Herkenham et aL (1990) with a 15-nM K,,
value did not (a Bmax/Kd ratio of 82/15,000 = 0.006). Addi-
tionally, although these binding data were gathered under zone
B conditions, the curvilinear characteristics would not have
been readily observed with experimental data since Bm.jKd
ratios 1.0 are only slightly different from idealized curves
(Chang et at., 1975). With use of the proper binding formulas,
the Scatchard representation of the data is slightly convex
upward (Boeynaems and Dumont, 1980), but linear graphical
representations have been developed (Riggs et at., 1970; Boey-
naems and Dumont, 1980), and the mathematical relationship
between the apparent dissociation constant and receptor con-
centration described (Chang et al., 1975). Because the deviation
from accepted (or ideal) conditions is minimal, the Bm.,, value
can be set equal to that found by computer analysis (499 pM).
However, the true value of the Kd must be calculated as the
by on May 2, 2009 jpet.aspetjournals.orgDownloaded from
1993
Cannabinoid SAR 225
apparent K,, (924 pM) minus one-half the Bmax (Boeynaems
and Dumont, 1980; Chang et al., 1975), which means the true
K,, is equal to 675 pM. The Ki values of displacing drugs were
calculated in table 2 using the Cheng-Prussoff formula (Cheng
and Prusoff, 1973) based upon a K,, value of 675 pM and a
corrected free radioligand concentration of 280 pM. No further
corrections to the Ki values were necessary because, under
conditions where the Bmax/Kd ratio <2.0, the exact displacement
equations for zone B conditions (Chang et al., 1975) revert to
more traditional analysis methods for competitive inhibitors.
The competitive nature of the binding of these analogs was
suggested by statistical analysis (ALLFIT; P < .05), which
indicated that the displacement curves of almost all analogs
were parallel to that of 9-THC (and CP-55,940). This suggests
that receptor recognition occurs in a similar fashion for most
of the cannabinoid drugs, and that the actual site on the
receptor to which drugs bind is identical. However, not all
displacement curves could be shown to be parallel to that of
i9-THC. Due to the limited number of data points that could
be generated for many analogs with KI values greater than 1
M, the ALLFIT program was unable to fit sigmoid curves for
subsequent comparisons. Therefore, the nature of the binding
of these compounds to the cannabinoid receptor is unknown.
This was not the situation with all compounds of weak potency.
Analysis was possible for the comparison of 9-THC with CBD
and all (+)-isomers. Results indicated that displacement curves
for CBD and XV (or CP-56,667 the isomer of CP-55,940) were
not parallel to that of 9-THC. This indicates binding of a
noncompetitive nature, and suggests that the pharmacological
actions of these two drugs may not be by mechanisms identical
to that of 9-THC. Interestingly, displacement curves for (+)-
11-OH-8-THC-DMH and dextronantradol were parallel to
that of 9-THC, which suggests binding on the receptor at the
same site as that for 9-THC.
In conclusion, data presented here establish a high degree of
correlation between the ability of a wide variety of cannabinoids
to bind to the receptor labeled by [3H]CP-55,940 and their
ability to produce in vivo pharmacological effects. These cor-
relations suggest a lack of species differences in terms of
receptor SAR, despite the fact that the pharmacological ef-
fect(s) measured between each species do not necessarily appear
to be related to one another. Additionally, these correlations
were established using a set of cannabinoids incorporating a
wide degree of structural diversity, and this set includes natural
cannabinoids, cannabinoid metabolites, dimethylheptyl (or re-
lated) side chain analogs, nonclassical bicyclic cannabinoids,
halogenated analogs and other synthetic analogs including ster-
eoisomers. Thus, in the process of establishing these correla-
tions, data presented here further enhance the body of knowl-
edge concerning the structural requirements for binding to the
cannabinoid receptor. However, evidence is presented suggest-
ing caution be used in analyzing radioligand binding data from
assays of tissue homogenates, because zone B binding behavior
may be encountered. This appears to be due to the fact that
the density of cannabinoid receptors is extremely high, espe-
cially compared to most other receptor systems. In fact, it has
been pointed out by others that the density of cannabinoid
receptors is close to values typically observed for receptor
systems activated by amino acid neurotransmitters. It is also
clear that the use of BSA in aqueous media does not signifi-
cantly alter binding of cannabinoids to the receptor, despite
that fact that cannabinoids bind to serum proteins, which
actually allows the drugs to be placed into solution. Lastly, data
presented here suggests that a single cannabinoid receptor
exists to which almost all cannabinoids bind at a single recog-
nition site. However, some cannabinoids such as CBD may
produce pharmacological actions either by interacting at this
receptor at a different recognition site, or by another receptor
mechanism altogether. No evidence is presented here which
would suggest which is likely to occur. In time, more selective
ligands may emerge which possess specific pharmacological
actions (analgesia, psychoactivity, etc., rather than the entire
spectrum of cannabinoid effects), or possess antagonist prop-
erties. However, finding such drugs will require continued com-
parison of in vitro binding and in vivo behavioral actions.
Further analysis of cannabinoid receptor SAR using ligands
with structures different from that of the bicyclic nature of CP-
55,940 may also prove useful, as data presented here indicates
some cannabinoids may bind to this receptor in a fashion
different from that of 9-THC. Examples of other cannabinoid
receptor ligands might be the aminoalkylindole WIN-55,212-
2(+)-[2,3-dihydro-5-methyl-3-[(4-morpholinyl)methyl]pyr-
rolo[1,2,3-de]-1,4-benzoxazin-6-yl](l-napthalenyl)methane or
possibly even a novel analog related to either the halogenated
or DMH cannabinoids presented here. However, data presented
here suggests the search for a cannabinoid antagonist or one of
selective action may prove to be very difficult.
Acknowledgments
The authors gratefully thank D. Troy Bridgen and Xin Wei for technical
assistance. The authors are also grateful to Drs. Richard Rothman and Allyn
Howlett for helpful suggestions regarding the initial cannabinoid binding studies.
References
BALSTER, R. L. AND PRESCOVF, W. R.: 9-Tetrahydrocannabinol discrimination
in rats as a model for cannabis intoxication. Neurosci. Biobehav. Rev. 16: 55-
62, 1992.
BINDER, M. AND FRANKE, I.: Is there a THC receptor? In Neuroreceptors, ed. by
F. Hucho, pp. 151-161, Walter de Gruyter Press, Berlin, 1982.
BOEYNAEMS, J. M. AND DUMONT, J. E.: Concepts of zones A, B and C. In
Outlines of Receptor Theory, ed. by J. M. Boenaems and J. E. Dumont, pp.
17-23, Elsevier/North-Holland Biomedical Press, Amsterdam, 1980.
BRADFORD, M. M.: A rapid and sensitive method for the quantitation of micro-
gram quantities of protein utilizing the principle of protein-dye binding. Anal.
Biochem. 72: 248-254, 1976.
BURSTEIN, S., HUNTER, S. A., LATHAM, V. AND RENZULLI, L.: A major metabolite
of .‘-tetrahythocannabinol reduces its cataleptic effect in mice. Expenientia
43: 402-403, 1987.
CHANG, K.-J., JACOBS, S. AND CUATRECASAS, P.: Quantitative aspects of hor-
mone-receptor interactions of high affinity. Effect of receptor concentration
and measurement of dissociation constants of labeled and unlabeled hormones.
Biochim. Biophys. Acta 406: 294-303, 1975.
CHARALAMBOUS, A., LIN, S., MARCINIAK, G., BANIJAMALI, A. R., FRIEND, F. L.,
COMPTON, D. R., MARTIN, B. R. AND MAKRIYANNIS, A.: Pharmacological
evaluation of halogenated .8-THC analogs. Pharmacol. Biochem. Behav. 40:
509-512, 1991.
CHENG, Y..C. AND PRUSOFF, W. H.: Relationship between the inhibition con-
stant (K,) and the concentration of inhibitor which causes 50 percent inhibition
(I,) on an enzymatic reaction. Biochem. Pharmacol. 22: 3099-3108, 1973.
COMPTON, D. R., JOHNSON, M. R., MELVIN, L. S. AND MARTIN, B. R.: Phar-
macological evaluation ofa series of bicyclic cannabinoid analogs: Classification
as cannabimimetic agents. J. Pharmacol. Exp. Ther. 260: 201-209, 1992.
COMPTON, D. R., LITTLE, P. J., MARTIN, B. R., GILMAN, J. W., SAHA, J. K.,
JORAPUR, V. S., SARD, H. P. AND RAZDAN, R. K.: Synthesis and pharmacolog-
ical evaluation of amino, azido, and nitrogen mustard analogues of lO-substi-
tuted cannabidiol and 11- or 12-substituted 8-tetrahydrocannabinol. J. Med.
Chem. 33: 1437-1443, 1990.
COMPTON, D. R., LIrFLE, P. J., MARTIN, B. R., SAHA, J. K., GILMAN, J., SARD,
H. AND RAZDAN, R. K.: Synthesis and pharmacological evaluation of mercapto-
and thioacetyl- analogues of cannabidiol and 8-tetrahydrocannabinol. Eur. J.
Med. Chem. 24: 293-298, 1989.
CoMrroN, D. R. AND MARTIN, B. R.: Pharmacological evaluation ofwater soluble
cannabinoids and related analogs. Life Sci. 46: 1575-1585, 1990.
DE LEAN, A., MUNSON, P. J. AND RODBARD, D.: Simultaneous analysis of families
of siginoidal curves: Application to bioassay, radioligand assay, and physiolog-
ical dose-response curves. Am. J. Physiol. 235: E97-E102, 1987.
DEVANE, W. A., DYSARZ, F. A., III, JOHNSON, M. R., MELVIN, L. S. AND
by on May 2, 2009 jpet.aspetjournals.orgDownloaded from
226 Compton St al
HowLrrr, A. C.: Determination and characterization ofa cannabinoid receptor
in rat brain. Mol. Phaninacol. 34: 605-613, 1988.
FURCHGOTT, R. F.: Pharmacological characterization of receptors: its relation to
radioligand-binding studies. Fed. Proc. 37: 115-120, 1978.
GA0NI, Y. AND MECHOULAM, R.: Isolation, structure, and partial synthesis of an
active constituent of hashish. J. Am. Chem. Soc. 86: 1646-1647, 1964.
GARRETT, E. R. AND HUNT, C. A.: Physicochemical properties, solubility, and
protein binding of -tetrahydrocannabinol. J. Pharm. Sci. 63: 1056-1064,
1974.
GERARD, C. M., MOLLEREAU, C., VASSART, G. AND PARMENTIER, M.: Molecular
cloning of a human cannabinoid receptor which is also expressed in testis.
Biochem. J. 279: 129-134, 1991.
GOLDSTEIN, A.: The interactions of drugs and plasma proteins. Pharmacol. Rev.
1: 102-165, 1949.
HAQUE, S. J. AND PODDAR, M. K.: Interactions of cannabinoids with bovine
serum albumin. Biosci. Rep. 4: 239-43, 1984.
HARRIs, L. S., CARCHMAN, R. A. AND MARTIN, B. R.: Evidence for the existence
ofspecific cannabinoid binding sites. Life Sci. 22: 1131-1138, 1978.
HERKENHAM, M., LYNN, A. B., LITFLE, M. D., JoHNsoN, M. R., MELVIN, L. S.,
DECOSTA, B. R. AND RICE, K. C.: Cannabinoid receptor localization in the
brain. Proc. Natl. Acad. Sci. U.S.A. 87: 1932-1936, 1990.
HOUSTON, D. B., EVANS, D. M., HowLrrr, A. C. AND MELVIN, L. S.: [3HJCP-
55.940 binding to the cannabinoid receptor in brain. Biotech Update (Du Pont)
6: 21-27, 1991.
HowLrrr, A. C., JOHNSON, M. R., MELVIN, L. S. AND MILNE, G. M.: Nonclas-
sical cannabinoid analgetics inhibit adenylate cyclase: development of a can-
nabinoid receptor model. Mol. Pharmacol. 33: 297-302, 1988.
LITTLE, P. J., C0MPT0N, D. R., JoHNsoN, M. R., MELVIN, L. S. AND MARTIN,
B. R.: Pharmacology and steneoselectivity of structurally novel cannabinoids
in mice. J. Pharmacol. Exp. Ther. 247: 1046-1051, 1988.
LImE, P. J., COMPTON, D. R., MECHOULAM, R. AND MARTIN, B. R.: Stereo-
chemical effects of 1 1-OH-dimethylheptyl-8-tetrahydrocannabinol. Pharma-
col. Biochem. Behav. 32: 661-666, 1989.
LOH, H. H., Law, P. Y., OSTWALD, T., CHO, T. M. AND WAY, E. L.: Possible
involvement of cerebroside sulfate in opiate receptor binding. Fed. Proc. 37:
1978.
MARTIN, B. R., COMPTON, D. R., LITTLE, P. J., MARTIN, T. J. AND BEARDSLEY,
P. M.: Pharmacological evaluation of agonist and antagonist activity of can-
nabinoids. Natl. Inst. Drug Abuse Res. Monogr. Ser. 79: 108-122, 1987.
MAT5UDA, L. A., L0LAIT, S. J., BROWNSTEIN, M. J., YOUNG, A. C. AND BONNER,
Vol. 265
T. I.: Structure of a cannabinoid receptor and functional expression of the
cloned cDNA. Nature 346: 561-564, 1990.
MCPHERSON, G. A.: A practical computer-based approach to the analysis of
radioligand binding experiments. Comput. Programs Biomed. 17: 107-114,
1983.
MUNSON, P. J. AND RODBARD, D.: LIGAND: A versatile computerized approach
for characterization of ligand-binding systems. Anal. Biochem. 107: 220-239,
1980.
Nvz, J. S., SELTZMAN, H. H., Prrr, C. G. AND SNYDER, S. S.: High-affinity
cannabinoid binding sites in brain membranes labeled with [3H)-5’-trimethy-
1anmonium--tetrahydrocannabinol.J. Pharmacol. Exp. Ther. 234: 784-791,
1985.
Nvs, J. S., SNOWMAN, A. M., VOGLMAIER, S. AND SNYDER, S. H.: High-affinity
cannabinoid binding site: Regulation by ions, ascorbic acid, and nucleotides. J.
Neurochem. 52: 1892-7, 1989.
RAZDAN, R. K.: Structure-activity relationships in cannabinoids. Pharmacol.
Rev. 38: 75-149, 1986.
REGGIO, P. H., McGAUGHEY, G. B., ODEAR, D. F., SELTZMAN, H. H., C0MPT0N,
D. R. AND MARliN, B. R.: A national search for the separation of psychoactivity
and analgesia in cannabinoids. Pharmacol. Biochem. Behav. 40: 479-486,
1991.
REGGIO, P. H., SELTZMAN, H. H., CoMrroN, D. R., PREsco’rr, W. R. AND
MARTIN, B. R.: An investigation of the role of the phenolic hydroxyl in
cannabinoid activity. Mol. PharmacoL 38: 854-862, 1990.
RIGGS, A. D., SUZUKI, H. AND BoURGEoIS, S.: lac Repressor-operator interaction.
I. Equilibrium studies. J. Mol. Biol. 48: 67-83, 1970.
ROSENTHAL, H. E.: A graphic method for the determination and presentation of
binding parameters in a complex system. Anal. Biochem. 20: 525-532, 1967.
SCATCHARD, G.: The attractions of proteins for small molecules and ions. Ann.
N.Y. Aced. Sci. 51: 1951.
SEEMAN, P.: Brain dopamine receptors. Pharmacol. Rev. 32: 229-313, 1980.
SEEMAN, P., TEDESCO, J. L., LEE, T., CHAU-WONG, M., MULLER, P., BOWLES,
J., WHITAKER, P. M., MCMANUS, C., TFrFLER, M., WEINREICH, P., FRIEND,
W. C. AND BROWN, G. M.: Dopamine receptors in the central nervous system.
Fed. Proc. 37: 130-136, 1978.
STAHL, K. D., VAN BEvER, W., JANSSEN, P. AND SIMoN, E. J.: Receptor affinity
and pharmacologic potency of a series of narcotic analgesic, anti-diarrheal and
neuroleptic drugs. Eur. J. Pharmacol. 46: 199-205, 1977.
Send reprint requests to: David R. Compton, Ph.D., Department of Phanma-
cology and Toxicology, Medical College of Virginia, Virginia Commonwealth
University, Richmond, VA 23298-0613.
by on May 2, 2009 jpet.aspetjournals.orgDownloaded from
