点击显示 收起
School of Medical Sciences, Institute of Medical Sciences, University of Aberdeen, Aberdeen, Scotland (M.R.P., G.L.B., A.T., L.A.S., R.G.P., R.A.R.)
Organon Research, Newhouse, Lanarkshire, Scotland (M.E., R.G., A.M., L.M., G.G., G.W., P.W., J.M., F.T., P.C.)
Department of Pharmacology, University of Melbourne, Parkville, Australia (A.C.)
Abstract
We investigated the pharmacology of three novel compounds, Org 27569 (5-chloro-3-ethyl-1H-indole-2-carboxylic acid [2-(4-piperidin-1-yl-phenyl)-ethyl]-amide), Org 27759 (3-ethyl-5-fluoro-1H-indole-2-carboxylic acid [2-94-dimethylamino-phenyl)-ethyl]-amide), and Org 29647 (5-chloro-3-ethyl-1H-indole-2-carboxylic acid (1-benzyl-pyrrolidin-3-yl)-amide, 2-enedioic acid salt), at the cannabinoid CB1 receptor. In equilibrium binding assays, the Org compounds significantly increased the binding of the CB1 receptor agonist [3H]CP 55,940 [(1R,3R,4R)-3-[2-hydroxy-4-(1,1-dimethylheptyl)phenyl]-4-(3-hydroxypropyl)cyclohexan-1-ol], indicative of a positively cooperative allosteric effect. The same compounds caused a significant, but incomplete, decrease in the specific binding of the CB1 receptor inverse agonist [3H]SR 141716A [N-(piperidin-1-yl)-5-(4-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboximide hydrochloride], indicative of a limited negative binding cooperativity. Analysis of the data according to an allosteric ternary complex model revealed that the estimated affinity of each Org compound was not significantly different when the radioligand was [3H]CP 55,940 or [3H]SR 141716A. However, the estimated cooperatively factor for the interaction between modulator and radioligand was greater than 1 when determined against [3H]CP 55,940 and less than 1 when determined against [3H]SR 141716A. [3H]CP 55,940 dissociation kinetic studies also validated the allosteric nature of the Org compounds, because they all significantly decreased radioligand dissociation. These data suggest that the Org compounds bind allosterically to the CB1 receptor and elicit a conformational change that increases agonist affinity for the orthosteric binding site. In contrast to the binding assays, however, the Org compounds behaved as insurmountable antagonists of receptor function; in the reporter gene assay, the guanosine 5'-O-(3-[35S]thio)triphosphate binding assay and the mouse vas deferens assay they elicited a significant reduction in the Emax value for CB1 receptor agonists. The data presented clearly demonstrate, for the first time, that the cannabinoid CB1 receptor contains an allosteric binding site that can be recognized by synthetic small molecule ligands.
Mammalian tissues express at least two types of cannabinoid receptor, CB1 and CB2, both G protein-coupled (for review, see Howlett et al., 2002). CB1 receptors are found predominantly at central and peripheral nerve terminals where they mediate inhibition of transmitter release. Endogenous ligands for these receptors also exist. These "endocannabinoids" are all eicosanoids, prominent examples including arachidonoylethanolamide (anandamide) and 2-arachidonoyl glycerol, both of which are synthesized on demand, removed from their sites of action by tissue uptake processes and metabolized by intracellular enzymes (Pertwee and Ross, 2002). There is ample evidence for a role of the endocannabinoid system in a number of physiological processes, including cardiovascular regulation, appetite control, learning, and memory and pain processing (for review, see Ross, 2003; Di Marzo et al., 2004). Within the brain, the distribution of CB1 receptors is heterogeneous, accounting for several well documented pharmacological properties of CB1 receptor agonists. For example, the cerebral cortex, hippocampus, lateral caudate-putamen, substantia nigra, pars reticulata, globus pallidus, entopeduncular nucleus, and the molecular layer of the cerebellum are all populated with particularly high concentrations of CB1 receptors. CB1 receptor ligands have therapeutic potential in a range of disorders, including obesity, nicotine addiction, pain, inflammation, gastrointestinal disorders, multiple sclerosis, psychosis, schizophrenia, and osteoporosis.
The traditional view of G protein-coupled receptor (GPCR) signaling posits that agonists initiate signaling by binding to a site on the GPCR that has specifically evolved to recognize the endogenous agonist for that receptor; this site has been defined as the orthosteric binding site (Neubig et al., 2003). It is now recognized, however, that GPCRs may also contain allosteric binding sites for endogenous and/or synthetic ligands (Christopoulos and Kenakin, 2002). Allosteric sites are topographically distinct from the orthosteric site; hence, the structural features that determine ligand binding to allosteric sites are different from those of orthosteric ligands. In contrast to the direct effects on receptor function that are mediated by orthosteric ligands, allosteric drugs act by modulating receptor activity through conformational changes in the receptor that are transmitted from the allosteric to the orthosteric site and/or to effector coupling sites (Christopoulos, 2002).
The simplest allosteric interaction can be described by the allosteric ternary complex model (TCM; Fig. 1), where the effect of an allosteric ligand is to change the affinity of an orthosteric ligand for the receptor, and vice versa (Ehlert, 1988; Lazareno and Birdsall, 1995). From a mechanistic standpoint, allosteric modulation of affinity reflects a conformational change in the receptor that may lead to an alteration in the dissociation kinetics of a preformed orthosteric ligand-receptor complex; this effect on dissociation kinetics cannot occur if interacting ligands recognize the same binding site and is thus diagnostic of an allosteric effect (Kostenis and Mohr, 1996; Christopoulos, 2000). More recently, data have been presented to suggest that allosteric modulators may affect the efficacy of orthosteric ligands in addition to, or independently of, effects on orthosteric ligand binding, necessitating the development of more complex models of allosteric modulation (Christopoulos and Kenakin, 2002, and references therein).
To date, the main principle underlying GPCR-based drug discovery has invariably been the optimization of lead molecules toward the orthosteric site on the GPCR to obtain selectivity of action. However, allosteric modulators offer advantages not readily available with orthosteric ligands. For example, allosteric drugs that affect affinity (but not efficacy) would be advantageous in that they would boost or dampen the effect of the endogenous ligand without disrupting the inherent spatial and temporal patterns of physiological signaling (May and Christopoulos, 2003). CB1 receptor pharmacology is the subject of intense academic and commercial research effort directed at the treatment of obesity (Horvath, 2003), pain (Pertwee, 2001; Iversen and Chapman, 2002), inflammation (Rice et al., 2002), osteoporosis (Idris et al., 2005), cancer (Bifulco and Di Marzo, 2002), multiple sclerosis (Pertwee, 2002), and cardiovascular disorders (Randall et al., 2002). There is ample evidence that the levels of endocannabinoids are altered in these pathophysiological situations. Allosteric enhancers of the cannabinoid CB1 receptor would thus offer the prospect of producing clinically useful compounds that do not display the CNS side effects that are characteristic of direct receptor agonism. In this investigation, we provide the first evidence for an allosteric binding site on the cannabinoid CB1 receptor, as revealed by the pharmacology and mode of action of a novel class of synthetic small molecules (Fig. 2).
Materials and Methods
Materials
(+)-WIN 55,212-2 (WIN 55,212) and CP 55,940 were obtained from Tocris Cookson, Inc. (Bristol, UK), and SR 141716A was from the National Institute on Drug Abuse (Bethesda, MD). Bovine serum albumin (BSA), cell culture media, dithiothreitol, nonenzymatic cell dissociation solution, GDP, Gpp(NH)p, GTPS, G418 (Geneticin), L-glutamine, Krebs' salts, penicillin/streptomycin, Tris buffer, and Triton X-100 were all obtained from Sigma-Aldrich. [3H]CP 55,940 (128 Ci/mmol) and [35S]GTPS (1250 Ci/mmol) were obtained from GE Healthcare (Little Chalfont, Buckinghamshire, UK). [3H]SR 141716A (44 Ci/mmol) was obtained from the National Institute on Drug Abuse. Three novel compounds were from Organon Research (Newhouse, Lanarkshire, Scotland): Org 27569-0, Org 27759-0, and Org 29647-1.
Radioligand Binding Assays
Mouse Brain Membrane Preparation. Whole brains from adult male MF1 mice were suspended in centrifugation buffer (320 mM sucrose, 2 mM EDTA, and 5 mM MgCl2), and the tissues were homogenized with an Ultra-Turrex homogenizer. Tissue homogenates were centrifuged at 1600g for 10 min, and the resulting supernatant was collected. This pellet was resuspended in centrifugation buffer centrifuged as described above, and the supernatant was collected. Supernatants were combined before undergoing further centrifugation at 28,000g for 20 min. The supernatant was discarded, and the pellet was resuspended in buffer A (50 mM Tris, 2 mM EDTA, 5 mM MgCl2 at pH 7.0) and incubated at 37°C for 10 min. After the incubation, the suspension was centrifuged for 20 min at 23,000g. After resuspending the pellet in buffer A, the suspension was incubated for 40 min at room temperature before a final centrifugation for 15 min at 11,000g. The final pellet was resuspended in buffer B (50 mM Tris, 1 mM EDTA, and 3 mM MgCl2), and the final protein concentration, determined by Bio-Rad DC kit, was 1 mg/ml. All centrifugation procedures were carried out at 4°C. Prepared brain membranes were stored at -80°C and defrosted on the day of the experiment.
Equilibrium Binding Assays. Binding assays were performed with the CB1 receptor agonist [3H]CP 55,940 (0.7 nM) and the CB1 receptor antagonist [3H]SR 141716A (1.2 nM), 1 mg/ml BSA and 50 mM Tris buffer containing 0.1 mM EDTA and 0.5 mM MgCl2, pH 7.4, in a total assay volume of 500 μl. Binding was initiated by the addition of mouse brain membranes (30 μg). Assays were carried out at 37°C for 60 min before termination by addition of ice-cold wash buffer (50 mM Tris buffer and 1 mg/ml BSA) and vacuum filtration using a 24-well sampling manifold (Brandel cell harvester) and Whatman GF/B glass-fiber filters (Whatman, Maidstone, UK) that had been soaked in wash buffer at 4°C for 24 h. Each reaction tube was washed five times with a 4-ml aliquot of buffer. The filters were oven-dried for 60 min and then placed in 5 ml of scintillation fluid (Ultima Gold XR; PerkinElmer Life and Analytical Sciences, Boston, MA), and radioactivity was quantitated by liquid scintillation spectrometry. Specific binding was defined as the difference between the binding that occurred in the presence and absence of 1 μM concentrations of the corresponding unlabeled ligand and was 70 to 80% of the total binding.
Dissociation Kinetics. Dissociation kinetic assays were performed with the CB1 receptor agonist [3H]CP 55,940 (0.7 nM), 1 mg/ml BSA, and 50 mM Tris buffer containing 0.1 mM EDTA and 0.5 mM MgCl2, pH 7.4, total assay volume 500 μl. We used the "isotopic dilution" method to measure the dissociation rate constant for [3H]CP 55,940 from brain membranes (Christopoulos, 2000). [3H]CP 55,940 (0.7 nM) was incubated with mouse brain membranes (30 μg) for 60 min at 25°C. Dissociation was initiated by the addition of 1 μM unlabeled ligand in the presence and absence of test compounds. Dissociation times of 0.5 to 120 min at 25°C were used. To determine the nonspecific binding, experiments were also performed in the presence of a 1 μM concentration of the unlabeled ligand. Binding was terminated by addition of ice-cold wash buffer (50 mM Tris buffer and 1 mg/ml BSA) and vacuum filtration using a 24-well sampling manifold (Brandel cell harvester) and Whatman GF/B glass-fiber filters that had been soaked in wash buffer at 4°C for 24 h. Each reaction tube was washed five times with a 4-ml aliquot of buffer. The filters were oven-dried for 60 min and then placed in 5 ml of scintillation fluid (Ultima Gold XR), and radioactivity was quantitated by liquid scintillation spectrometry. Specific binding was defined as the difference between the binding that occurred in the presence and absence of 1 μM unlabeled ligand and was 70 to 85% of the total binding.
Mouse Isolated Vas Deferens Assay. Vasa deferentia were obtained from albino MF1 mice weighing 26 to 40 g. Each tissue was mounted in a 4-ml organ bath at an initial tension of 0.5g. The baths contained Mg2+-free Krebs' solution that was kept at 35°C and bubbled with 95% O2 and 5% CO2. The composition of the Krebs' solution was 118.2 mM NaCl, 4.75 mM KCl, 1. 19 mM KH2PO4, 25.0 mM NaHCO3, 11.0 mM glucose, and 2.54 mM CaCl2·6H2O. Isometric contractions were evoked by stimulation with 0.5-s trains of three pulses of 110% maximal voltage (train frequency, 0.1 Hz; pulse duration, 0.5 ms) through a platinum electrode attached to the upper end and a stainless steel electrode attached to the lower end of each bath. Stimuli were generated by a Grass S48 stimulator and then amplified (Med-Lab channel attenuator) and divided to yield separate outputs to eight organ baths (Med-Lab StimuSplitter). Contractions were monitored by computer using a data recording and analysis system (MacLab) that was linked via preamplifiers (Macbridge) to UF1 transducers. After placement in an organ bath, each tissue was subjected to a stimulation-free period of 15 min and then stimulated for 10 min. Tissues were then subjected to alternate periods of stimulation (5 min) and rest (10 min) until consistent twitch amplitudes were obtained. This equilibration procedure was followed by a stimulation-free period of 10 min. Tissues were then stimulated for 10 min after which the stimulator was switched off, and the Org compound or its vehicle was added. Thirty minutes later, the first addition of the nonselective CB1/CB2 receptor agonist WIN 55,212 was made. Additions of WIN 55,212 were made cumulatively at 5-min intervals without washout, the tissues being stimulated for the final 2 min of exposure to each concentration of this agonist. Org compounds were dissolved in DMSO, and WIN 55,212 was dissolved in a solution consisting of 50% DMSO and 50% saline. By themselves, these vehicles did not inhibit the twitch response. Drug additions were made in a volume of 10 μl.
The measured response was WIN 55,212-induced inhibition of electrically-evoked contractions of the mouse isolated vas deferens. The Org compounds (or their vehicle, DMSO) were added 30 min before WIN 55,212 that was then added cumulatively using a 15-min dose cycle. Thus, we constructed WIN 55,212 concentration-response curves in the presence of DMSO or the Org compounds.
The degree of inhibition of evoked contractions induced by WIN 55,212 was calculated in percentage terms by comparing the amplitude of the twitch response after each addition of WIN 55,212 with its amplitude immediately before the first addition of this agonist.
[35S]GTPS Binding Assay. Mouse brain membranes (5 μg of protein) were preincubated for 30 min at 30°C with adenosine deaminase (0.5 U/ml). The membranes were then incubated with the agonist with vehicle or modulator for 60 min at 30°C in assay buffer (50 mM Tris-HCl, 50 mM Tris-base, 5 mM MgCl2, 1 mM EDTA, 100 mM NaCl, 1 mM dithiothreitol, and 0.1% BSA) in the presence of 0.1 nM [35S]GTPS and 30 μM GDP, in a final volume of 500 μl. Binding was initiated by the addition of [35S]GTPS. Nonspecific binding was measured in the presence of 30 μM GTPS. The reaction was terminated by rapid vacuum filtration (50 mM Tris-HCl, 50 mM Tris-base, and 0.1% BSA) using a 24-well sampling manifold (Brandel cell harvester) and GF/B filters that had been soaked in buffer (50 mM Tris-HCl, 50 mM Tris-base, and 0.1% BSA) for at least 24 h. Each reaction tube was washed six times with a 1.2-ml aliquot of ice-cold wash buffer. The filters were oven-dried for at least 60 min and then placed in 5 ml of scintillation fluid (Ultima Gold XR). Radioactivity was quantified by liquid scintillation spectrometry.
CB1 Reporter Gene Assay
Test compounds were all prepared in glass vials and were dissolved in DMSO to give a concentration of 10 mM. The solutions were then diluted in nutrient mix F-12 media with 3% BSA to give a range of concentrations from 0.1 mM to 0.1 nM (10x stocks) for addition to the assay (BSA final concentration, 0.3%). The DMSO concentration did not exceed 0.1% by volume. For all aqueous solutions, ultrapure water (Milli-Q quality) was used.
Chinese hamster ovary cells, expressing the human CB1 receptor were transfected with a luciferase reporter gene under the control of seven AP1 repeats (Stratagene, La Jolla, CA). Single cell-derived clones were maintained in Ham's F-12 nutrient mix (1:1) with 10% FetaClone II, 50 U/50 μg/ml penicillin/streptomycin, 1 μg/ml fungizone, 0.2 μg/ml geneticin (disulfate salt), and 0.5 μg/ml hygromycin at 37°C/5% CO2. The day before the assay, the cells were harvested using trypsin-EDTA and then pelleted by centrifugation (5 min, 20°C spin at 600g). The pellet was resuspended in phenol red/serum-free Dulbecco's modified Eagle's medium/Ham's F-12 nutrient mix containing 50 units of penicillin, 50 μg/ml streptomycin, and 1 μg/ml Fungizone. The cells were then seeded into white 96-well plates (CulturPlates; PerkinElmer Life and Analytical Sciences) at a density of 3 x 104 cells per well (100-μl final volume). Cells were incubated overnight (approximately 18 h at 37°C, 5% CO2, 95% air) before assay.
Ten microliters of a 10x stock of test compound was added directly to the relevant wells, and the plates were incubated at 37°C for 5 h. Then, under subdued light, 100 μl of LucLite reagent (reconstituted per instructions from PerkinElmer Life and Analytical Sciences) was added to each well. The plates were covered with Top Seal and then incubated at room temperature for 5 min before counting on the PerkinElmer Life and Analytical Sciences TopCount (single photon counting, 0.01-min count time, no count delay). For modulator studies, the cells were preincubated for 5 min with the test compound and then agonist. Incubations and luciferase assays were then performed as detailed above.
Data Analysis
Radioligand competition binding isotherms for the interaction between orthosteric ligands were analyzed using Prism 4.01 (Graph-Pad Software Inc., San Diego, CA) according to the following logistic function:
(1)
where Y denotes the percentage of specific binding; top and bottom denote the maximal and minimal asymptotes, respectively; [I] denotes the inhibitor concentration; and IC50 denotes the inhibitor potency (midpoint location) parameter.
The interaction between each radioligand and the test modulators was analyzed according to the following allosteric ternary complex model of interaction:
(2)
where Y denotes the fractional specific binding, [A] and [B] denote the concentration of orthosteric and allosteric ligands, respectively, KA and KB denote their respective equilibrium dissociation constants, and denotes a cooperativity factor that governs the magnitude and direction of the allosteric interaction between the two ligands when they both occupy the receptor. For this analysis, values of [A] and KA were fixed as constants (the latter determined from previous saturation binding assays), whereas values of and KB were determined by nonlinear regression.
For the dissociation kinetic experiments, the decay in radioligand binding over time was analyzed according to the following two-phase exponential decay model:
(3)
where Span1 and Span2 denote the percentage of each phase; plateau denotes the minimal asymptotic value; and k1 and k2 denote the rate constants for the components defined by Span1 and Span2, respectively. An extra-sum-of-squares (F test) was used to determine whether the data were significantly better fitted to this model compared with a simpler model characterized by a single Span and k value.
Concentration-response data were fitted empirically to the following four-parameter logistic equation:
(4)
where Y denotes effect; [A] the concentration of agonist; nH the Hill slope, pEC50 the negative logarithm of the midpoint location parameter; and Emax and basal the upper and lower asymptotes, respectively. An extra-sum-of-squares (F test) was used to determine whether the data were significantly better fitted assuming separate values for the basal and nH parameters across each data set in the absence relative to presence of test compound, compared with common values for each parameter. In practice, all affinity (KB), potency (EC50), and cooperativity () parameters were estimated as logarithms (Christopoulos, 1998).
Operational Model-Fitting. The functional interaction between test compounds and the orthosteric agonist WIN 55,212 in the mouse isolated vas deferens was fitted to the following operational model of allosterism (see Appendix for derivation):
(5)
where E denotes the effect, denotes an empirical proportionality constant that quantifies the change in stimulus imparted to the receptor by the agonist in the presence of modulator, Em denotes the maximum possible effect, is an operational measure of orthosteric ligand efficacy, and n is a logistic slope factor that governs the shape of the stimulus-response function. [A], KA, and are as defined above.
Statistics
Values have been expressed as means and variability as S.E.M. or as 95% confidence limits. Mean values have been compared using Student's unpaired t test or analysis of variance followed by Dunnett's test or the Newman-Keuls test. A P value <0.05 was considered significant.
Results
Equilibrium Binding Assays. In equilibrium binding experiments with the orthosteric agonist probe [3H]CP 55,940, unlabeled CP 55,940 competed with the radioligand according to a simple one-site competition model, characterized by a pIC50 value of 9.54 ± 0.09 (n = 5). In contrast, the test compounds all produced a significant, but saturable, increase in the level of specific [3H]CP 55,940 binding (Fig. 3A), the most marked effect being observed with Org 27569.
When the inverse agonist [3H]SR 141716A was used as the orthosteric probe, a different profile of activity was observed. As expected, unlabeled SR 141716A displayed simple mass-action behavior, competing with the radioligand and yielding a pIC50 value of 9.22 ± 0.02 (n = 4). Unlike their effects on agonist binding, the test compounds (at concentrations up to their solubility limit) also produced a significant decrease in the equilibrium binding of [3H]SR 141716A (Fig. 3B). However, the reduction in specific radioligand binding mediated by the test compounds seemed saturable and incomplete, which is at odds with the expectations of simple competitive behavior.
Both agonist and antagonist radiolabel binding data were fitted to the allosteric TCM (Fig. 1; eq. 2, and the resulting parameters are shown in Table 1 where it can be seen that there is no significant difference in the estimated pKB value for each compound at the putative allosteric site on the CB1 receptor when tested against [3H]CP 55,940 compared with when [3H]SR 141716A was used as the orthosteric probe. In contrast, the [3H]CP 55,940 data were characterized by values greater than 1, indicative of a positive allosteric modulation of the binding of radiolabeled agonist, whereas those of the [3H]SR 141716A data set were less than 1, indicating a negative allosteric modulation of the binding of the inverse agonist.
Dissociation Kinetic Assays. To validate that the test compounds were indeed allosteric modulators of the CB1 receptor, [3H]CP 55,940 dissociation kinetic experiments were performed. In the absence of test compound, the dissociation of [3H]CP 55,940 was significantly (P < 0.0001; F test) biphasic (Fig. 4). The dissociation rate constants and the percentages for the two phases are listed in Table 2. When this experiment was performed in the presence of each of the test compounds, a concentration-dependent reduction was observed in each of the slow and fast radioligand dissociation rate constants (Fig. 4, A–C; Table 2), with no significant effect on the percentage of either phase. This finding is consistent with an alteration of receptor conformation that reduces orthosteric ligand dissociation without affecting G protein coupling status. In contrast, the addition of the nonhydrolysable GTP analog Gpp(NH)p (40 μM), which is known to uncouple complexes between receptors and G-protein, did not have a significant effect on [3H]CP 55,940 dissociation rates, but it did significantly reduce the percentage of the slowly dissociating component of binding, from 59.03 ± 3.03 to 16.10 ± 2.93%
Mouse Vas Deferens. The effects of the modulators on CB1 receptor function were initially investigated using the mouse vas deferens as an in vitro model of CB1 receptor activity. As shown in Fig. 5, A to C, the compounds by themselves neither inhibited nor enhanced electrically evoked contractions at the concentrations that were used (up to 1 μM), suggesting that they are neither allosteric agonists nor inverse agonists. This is in contrast to the effect of the established CB1 receptor inverse agonist SR 141716A, which significantly increases electrically evoked contractions under the same assay conditions (Fig. 5D). The functional effects of the compounds on the inverse agonism displayed by SR 141716A in the mouse vas deferens are the subject of ongoing investigation.
When tested against the CB1 agonist WIN 55,212, an interesting effect was observed in the presence of modulator. In particular, the maximal agonist effect (Emax) of WIN 55,212 to inhibit the electrically evoked contractions of the mouse vas deferens was significantly reduced by each of Org 27569, Org 27759, and Org 29647, indicative of an insurmountable mode of antagonism (Table 3); with the exception of 1 μM Org27569 and Org 27759, the compounds did not significantly alter the pEC50 values for WIN 55,212 (Table 3).
To investigate whether the effects of the compounds were unique to cannabinoid receptors, we conducted experiments with the 2 adrenoceptor agonist clonidine. In the presence of DMSO, clonidine inhibited electrically evoked contractions of the mouse vas deferens with a pEC50 of 9.00 ± 0.14 and Emax of 83% (95% confidence limits, 77.9–88.1). These values were not significantly different in the presence of 1 μM Org 27569, with the pEC50 and Emax being 8.66 ± 0.11 and 77% (95% confidence limits, 70.6–82.7), respectively.
To gain further insight into the apparent dichotomy between modulators' effects on agonist affinity, on the one hand, and efficacy, on the other, the data were globally fitted to an operational model of allosterism (see Appendix). This analysis was based on the following assumptions. First, WIN 55,212 was assumed to be a full agonist at the CB1 receptor (Pertwee, 1997): the Emax value of approximately 80% of the maximal tissue inhibition presumably represents the maximal inhibition that can be mediated by CB1 receptor activation. Second, the G protein-coupled receptor active state, which is promoted by agonist binding, was assumed to be extremely transient, as expected for intact cells (Ehlert and Rathbun, 1990), such that the affinity estimate for WIN 55,212 reflected its KA value for the uncoupled form of the receptor.
The results of this analysis are shown in Fig. 6 and Table 4 and allow us to reconcile our observations with a mechanism whereby the binding cooperativity ( value) between each modulator and WIN 55,212 is positive, consistent with the enhancement of agonist binding observed in the [3H]CP 55,940 binding experiments, but the signaling efficiency of the ARB ternary complex (Fig. 1) is significantly impaired ( values tend toward or equal 0). Together, these two mechanisms can account for an insurmountable antagonism of receptor function with minimal effect on full agonist potency.
[35S]GTPS Binding Assay. The insurmountable antagonism mediated by the modulators in the intact mouse vas deferens functional assay is also evident when receptor activity is monitored using a more proximal measure of receptor-effector coupling. In the presence of vehicle, CP 55,940 produced a stimulation of [35S]GTPS binding that was 78.65% (95% confidence limits, 66.7–90.6) above basal binding with a pEC50 value of 7.93 ± 0.14 (Fig. 7A). In the presence of 100 nM Org 29647, the curve was shifted to the right, although there was no statistically significant effect on Emax or pEC50, with these values being 73.70% (95% confidence limits, 36.8–110) and 7.54 ± 0.48, respectively. However, when the concentration of Org 29647 was increased to 1 μM, the ability of CP 55,950 to stimulate [35S]GTPS turnover was completely abrogated.
In the presence of vehicle the endogenous cannabinoid anandamide produced a stimulation of [35S]GTPS binding that was 61.1% (95% confidence limits, 53.8–68.3) above basal binding with a pEC50 value of 6.62 ± 0.15 (Fig. 7B). In the presence of 100 nM Org 27569, the curve was shifted to the right, although there was no statistically significant effect on Emax of 59.8% (95% confidence limits, 49.9–69.6). However, when the concentration of Org 27569 was increased to 1 μM, the ability of anandamide to stimulate [35S]GTPS turnover was significantly attenuated, the Emax being 17.0% (95% confidence limits, 10.2–23.7).
Human CB1 Receptor Reporter Gene Assay. To confirm that the effect of the modulators was not species-specific, additional functional assays were performed using cloned human CB1 receptors. In the luciferase reporter assay, the Org compounds all behaved as insurmountable antagonists of CP 55,940 (Fig. 8), causing a concentration-dependent reduction in the agonist Emax value (Table 5).
Discussion
The Org compounds that we have identified display a number of characteristics commonly associated with allosteric modulators, including 1) either enhancement or inhibition of orthosteric CB1 ligand binding depending on the nature of the orthosteric probe, 2) a slowing of the dissociation rate constant(s) for [3H]CP 55,940 from the occupied CB1 receptor, and 3) a noncompetitive inhibition of CB1 receptor orthosteric agonist efficacy.
In equilibrium binding assays, the Org compounds significantly increased the binding of the CB1 receptor agonist [3H]CP 55,940, indicative of a positively cooperative allosteric effect. In contrast, the same compounds caused a significant, but incomplete, decrease in the specific binding of the CB1 receptor inverse agonist [3H]SR 141716A, indicative of a limited negative binding cooperativity (Christopoulos and Kenakin, 2002). The most common mechanistic description of inverse agonism is based on the premise that GPCRs exist in at least two conformational states, an inactive R and an active R state (Pertwee 2005). In this model, agonists have higher affinity for the R state and shift the equilibrium toward R resulting in G protein activation and an increase in GDP/GTP exchange. In contrast, inverse agonists bind preferentially to the R state, resulting in a decrease in constitutive activity. We used [3H]CP 55,940 to selectively label the R state, most likely present as the complexed RG state in our membrane-based assays, and [3H]SR 141716A to label the R state of the receptor. Analysis of the data according to the allosteric TCM revealed that the estimated pKB values for the Org compounds were not significantly different when the radioligand was [3H]CP 55,940, or [3H]SR 141716A (Table 1). Because these affinity values represent the dissociation constant for the modulator binding to the allosteric site on the free (nonorthosteric ligand-bound) receptor, they should indeed be independent of the orthosteric radioligand that is used (Lazareno and Birdsall, 1995). In addition, because our orthosteric radioligands labeled specific receptor states, our analysis indicates that under the current assay conditions, the binding of the allosteric modulators is not influence appreciably by the G protein coupling status of the receptor. The differences in the effects of each modulator on the binding of the orthosteric ligands could thus be adequately described and quantified by the estimated values (Table 1), which were all greater than 1 when determined against the binding of [3H]CP 55,940 and less than 1 when determined against the binding of [3H]SR141716A. Cooperativity is known to be influenced by the nature of the ligand binding to the orthosteric site, and it is thus not surprising that the values are different for each modulator as well as between each modulator and either an agonist or antagonist probe (May and Christopoulos, 2003).
It is noteworthy that dissociation kinetic studies using [3H]CP 55,940 as the probe also validated the allosteric nature of the Org compounds, because they all significantly decreased the dissociation rate constants for the radioligand. An alteration of the dissociation of a preformed orthosteric ligand-receptor complex by a second ligand unambiguously demonstrates an allosteric interaction (Kostenis and Mohr, 1996). It is noteworthy that [3H]CP 55,940 displayed biphasic dissociation kinetics, reflecting the presence of at least two receptor conformations (Devlin and Christopoulos, 2002). The rapidly dissociating component of [3H]CP 55,940 binding probably represents binding to the low-affinity uncoupled receptors (R), whereas the slowly dissociating component possibly reflects binding to high-affinity activated (R) receptors. Irrespective of the actual nature of the receptor states detected in the dissociation assays, the Org compounds significantly decreased the dissociation rate constant for both the slow and fast components of agonist binding without significantly affecting the proportion of each state. The compounds therefore modulate the dissociation of [3H]CP 55,940 from both the R and R receptor conformations. A reduction in the dissociation rate constant is in line with the positive cooperativity observed in the equilibrium binding assay of [3H]CP 55,940. Together, these data suggest that the Org compounds bind allosterically to the CB1 receptor and elicit a conformational change that increases agonist affinity for the orthosteric binding site.
Despite the positive allosteric enhancement of agonist binding observed in the radioligand binding assays, perhaps the most striking observation from our study was the dissimilitude between the effect on binding affinity and the effect of the same modulators on agonist function (i.e., efficacy). In contrast to the binding assays, the Org compounds did not behave as allosteric enhancers but rather as insurmountable antagonists of receptor function; in the reporter gene assay, the [35S]GTPS binding assay and the mouse vas deferens assay they elicited a significant reduction in the Emax value for CB1 receptor agonists. The effects of the modulators on function were specifically manifested at the receptor when it was dually occupied by both orthosteric and allosteric ligand, because there was no evidence of agonist or inverse agonist properties of the modulators when tested alone (Fig. 5).
There are a number of examples in the recent literature of allosteric modulators that have profound inhibitory effects on orthosteric ligand efficacy but minimal effects on orthosteric ligand binding (Litschig et al., 1999; Zahn et al., 2002; Watson et al., 2005), but our study is the first to reveal a striking difference between the magnitude and direction of allosteric modulation of orthosteric affinity, on the one hand, and efficacy, on the other. Numerous models have been proposed to specifically account for such behavior at the molecular level (Hall, 2000; Christopoulos and Kenakin, 2002; Parmentier et al., 2002). An essential feature of these models is that they contain multiple receptor states that are differentially stabilized by both orthosteric and allosteric ligands. As a consequence, the type of probe and the nature of the assay used can bias the detection of one set of receptor states over another and hence lead to a situation such as that described in our study where different observations are made depending on the experimental endpoint.
These findings have highly significant implications for drug discovery based on allosteric modulators, because they highlight the possibility of misclassification of novel ligands if they are not tested in as full a spectrum of biological assays as possible. In this regard, the operational model of allosterism, as presented in our study, can prove particularly useful. Like the molecular models described in the preceding paragraph, the operational model we have used is able to accommodate effects of allosteric modulators on both binding and function. However, the operational model has the additional advantage of containing fewer parameters than the molecular models; to our knowledge, this is the first application of such a model to derive quantitative functional parameter estimates for allosteric modulators (Table 4). The operational model parameters KA, KB, and have the same mechanistic meaning as they would in any molecular model of allosterism, and their determination can lead to useful information for structure-activity studies. Although the parameter is used only as an empirical proportionality constant, its value can also provide useful insights. For example, values significantly different to either 1 or 0 would be indicative of a ternary complex (ARB) species that retains the ability to signal in the presence of modulator, but in a modified way.
One point of divergence between the functional and the binding data concerns the apparent affinity of the modulators for the mouse CB1 receptor; in the vas deferens bioassay, the estimated affinity of the Org compounds was higher than that estimated in the binding assay (Table 1 versus 4). At the moment, the basis for this discrepancy between functional and binding affinity estimates remains unknown. However, a previous study by Pedder et al. (1991) revealed a striking divergence in the affinity estimates of the allosteric modulator, gallamine, at M2 muscarinic receptors depending on the assay conditions used to determine this value. Indeed, differences in the affinity of the modulator spanned almost two orders of magnitude. It is also worth noting that these same variations in assay conditions had minimal effects on orthosteric ligand affinity estimates (Pedder et al., 1991). Because allosteric sites are, by their very nature, topographically distinct from the orthosteric site on a GPCR, they can be sensitive to ionic changes under conditions where orthosteric ligands are minimally affected. Such a scenario could explain the discrepancy between the pKB values obtained from our equilibrium binding (Table 1) and those obtained in the functional assay in the mouse vas deferens (Fig. 6). Together, our findings highlight the prospect of manipulating structure-activity to produce compounds that are either allosteric enhancers or allosteric antagonists of the CB1 receptor. Endocannabinoid levels are significantly increased in models of inflammatory pain and CB1 receptor agonists are highly effective analgesics in these models (Pertwee 2001; Iversen and Chapman, 2002; Rice et al., 2002). Furthermore, inhibitors of fatty acid amide hydrolase, the enzyme responsible for the rapid intracellular hydrolysis of anandamide, are antinociceptive in models of inflammatory pain. These compounds do not display the CNS side effects that are characteristic of direct CB1 receptor agonism (Cravatt and Lichtman, 2004). Allosteric modulators of the cannabinoid CB1 receptor would also offer the prospect of producing clinically useful compounds that do not display CNS side effects.
In conclusion, this investigation has provided the first evidence for an allosteric binding site on the cannabinoid CB1 receptor. This site can be recognized by synthetic small molecules and offers the prospect of a new avenue of CB1 receptor research that will complement the existing portfolio of selective direct agonists and competitive antagonists.
Appendix
An Operational Model of Allosteric Modulation
In the allosteric TCM (Fig. 1), the orthosteric ligand, A, is distributed over two receptor-bound species, AR and ARB. From this model, the concentration of the AR species in the presence of modulator, B, is as follows:
(6)
where [R]T denotes the total concentration of all receptor species (bound and unbound). Likewise, the concentration of the ARB ternary complex is described as:
(7)
In addition to effects on ligand binding affinity, allosteric modulators may also affect the efficacy of orthosteric agonists. To derive a functional model of allosterism that accounts for this possibility, it is first necessary to describe mathematically the effect of an allosteric modulator on the stimulus imparted by an orthosteric agonist to the cell, and the effect of stimulus-response coupling on the final observed response. In accordance with the conventions of classic receptor theory (Furchgott, 1966; Ehlert, 1988), the stimulus (S) is equal to the product of the concentration of orthosteric agonist-occupied receptors and the agonist intrinsic efficacy. In the presence of an allosteric modulator, the stimulus may thus be expressed as:
(8)
where denotes the intrinsic efficacy of the orthosteric ligand and is a coupling factor that describes the ability of an allosteric modulator to alter the signaling capacity of the ternary complex. Values of < 1 thus denote an attenuation in signaling, a situation that may be classed as a form of noncompetitive antagonism of functional responses, values of = 1 denote no change in the signaling capacity of the receptor in the presence of modulator, and values of > 1 denote an increased capacity of the receptor to signal in the presence of modulator. This model assumes that the allosteric modulator does not itself impart any stimulus to the receptor. Substitution of eqs. 6 and 7 into eq. 8, followed by simplification, results in the following general form of the allosteric stimulus equation:
(9)
To derive the relationship between agonist concentration and final tissue response, an allowance must be made for stimulus-response coupling. Because most relationships between occupancy and response are nonlinear in nature, eq. 9 must be processed through a nonlinear function to derive an equation describing tissue response. One appropriate general function is the logistic equation, because it has been shown to closely model the shape of agonist concentration-response curves in many systems (Kenakin and Beek, 1982; Black and Leff, 1983). Thus, the tissue response (E), expressed as a fraction of the maximal response (Em), may be described as a function of the stimulus as follows:
(10)
where the parameter KS denotes a constant that governs the efficiency of stimulus-response coupling, and n denotes a logistic slope factor. Substituting eq. 9 into eq. 10, and defining = [R]T/KS as an operational measure of efficacy, gives the following operational model of agonism in the presence of an allosteric modulator:
(11)
Footnotes
A.C. is a Senior Research Fellow of the National Health and Medical Research Council of Australia, M.P. is funded by the National Institute on Drug Abuse, and G.B. is a Biotechnology and Biological Sciences Research Council/Organon Case award student. A.T. and L.S. are funded by GW Pharmaceuticals (Salsbury, Wiltshire, UK).
ABBREVIATIONS: CB, cannabinoid; GPCR, G protein-coupled receptor; TCM, ternary complex model; CNS, central nervous system; WIN 55,212, (+)-WIN 55,212-2 [(R)-(+)-[2,3-dihydro-5-methyl-3-(4-morpholinylmethyl)pyrrolo[1,2,3-de]-1,4-benzoxazin-6-yl]-1-naphthalenylmethanone]; CP 55,940, (1R,3R,4R)-3-[2-hydroxy-4-(1,1-dimethylheptyl)phenyl]-4-(3-hydroxypropyl)cyclohexan-1-ol; SR 141716A, N-(piperidin-1-yl)-5-(4-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboximide hydrochloride; BSA, bovine serum albumin; DTT, dithiothreitol; Gpp(NH)p, guanosine 5'-(,-imido)triphosphate; GTPS, guanosine 5'-O-(3-thio)triphosphate; Org 27569-0, 5-chloro-3-ethyl-1H-indole-2-carboxylic acid [2-(4-piperidin-1-yl-phenyl)-ethyl]-amide; Org 27759-0, 3-ethyl-5-fluoro-1H-indole-2-carboxylic acid [2-94-dimethylamino-phenyl)-ethyl]-amide; Org 29647-1, 5-chloro-3-ethyl-1H-indole-2-carboxylic acid (1-benzyl-pyrrolidin-3-yl)-amide, 2-enedioic acid salt; DMSO, dimethyl sulfoxide.
References
Bifulco M and Di Marzo V (2002) Targeting the endocannabinoid system in cancer therapy: a call for further research Nat Med 8: 547-550.
Black JW and Leff P (1983) Operational models of pharmacological agonism. Proc R Soc Lond B Biol Sci 220: 141-162.
Christopoulos A (1998) Assessing the distribution of parameters in models of ligand-receptor interaction: to log or not to log. Trends Pharmacol Sci 19: 351-357.
Christopoulos A (2000) Quantification of allosteric interactions at G protein-coupled Receptors using radioligand binding assays, in Current Protocols in Pharmacology (Enna SJ ed) pp 1.22.1-1.22.40. John Wiley & Sons, New York.
Christopoulos A (2002) Allosteric binding sites on cell-surface receptors: novel targets for drug discovery. Nat Rev Drug Discov 1: 198-210.
Christopoulos A and Kenakin T (2002) G protein-coupled receptor allosterism and complexing. Pharmacol Rev 54: 323-374.
Cravatt BF and Lichtman AH (2004) The endogenous cannabinoid system and its role in nociceptive behavior. J Neurobiol 61: 149-160.
Devlin MG and Christopoulos A (2002) Modulation of cannabinoid agonist binding by 5-HT in the rat cerebellum. J Neurochem 80: 1095-1102.
Di Marzo V, Bifulco M, and De Petrocellis L (2004) The endocannabinoid system and its therapeutic exploitation. Nat Rev Drug Discov 3: 771-784.
Ehlert FJ (1988) Estimation of the affinities of allosteric ligands using radioligand binding and pharmacological null methods. Mol Pharmacol 33: 187-194.
Ehlert FJ and Rathbun BE (1990) Signaling through the muscarinic receptor-adenylate cyclase system of the heart is buffered against GTP over a range of concentrations. Mol Pharmacol 38: 148-158.
Furchgott RF (1966) The use of -haloalkylamines in the differentiation of receptors and in the determination of dissociation constants of receptor-agonist complexes. Adv Drug Res 3: 21-55.
Hall DA (2000) Modeling the functional effects of allosteric modulators at pharmacological receptors: an extension of the two-state model of receptor activation. Mol Pharmacol 58: 1412-1423.
Horvath TL (2003) Endocannabinoids and the regulation of body fat: the smoke is clearing. J Clin Investig 112: 323-326.
Howlett AC, Barth F, Bonner TI, Cabral G, Casellas P, Devane WA, Felder CC, Herkenham M, Mackie K, Martin BR, et al. (2002) International Union of Pharmacology. XXVII. Classification of cannabinoid receptors. Pharmacol Rev 54: 161-202.
Idris AI, van't Hof RJ, Greig IR, Ridge SA, Ross RA, and Ralston SH (2005) Regulation of bone mass, bone loss and osteoclast activity. Nat Med 11: 774-779.
Iversen L and Chapman V (2002) Cannabinoids: a real prospect for pain relief Curr Opin Pharmacol 2: 50-55.
Kenakin T and Beek D (1982) In vitro studies on the cardiac activity of prenalterol with reference to use in congestive heart failure. J Pharmacol Exp Ther 220: 77-85.
Kostenis E and Mohr K (1996) Two-point kinetic experiments to quantify allosteric effects on radioligand dissociation. Trends Pharmacol Sci 17: 280-283.
Lazareno S and Birdsall NJM (1995) Detection, quantitation and verification of allosteric interactions of agents with labeled and unlabeled ligands at G protein-coupled receptors: interactions of strychnine and acetylcholine at muscarinic receptors. Mol Pharmacol 48: 362-378.
Litschig S, Gasparini F, Rueegg D, Stoehr N, Flor PJ, Vranesic I, Prezeau L, Pin JP, Thomsen C, and Kuhn R (1999) CPCCOEt, a noncompetitive metabotropic glutamate receptor 1 antagonist, inhibits receptor signaling without affecting glutamate binding. Mol Pharmacol 55: 453-461.
May LT and Christopoulos A (2003) Allosteric modulators of G-protein-coupled receptors. Curr Opin Pharmacol 3: 551-556.
Neubig RR, Spedding M, Kenakin T, and Christopoulos A (2003) International Union of Pharmacology Committee on Receptor Nomenclature and Drug Classification. XXXVIII. Update on terms and symbols in quantitative pharmacology. Pharmacol Rev 55: 597-606.
Parmentier ML, Prezeau L, Bockaert J, and Pin JP (2002) A model for the functioning of family 3 GPCRs. Trends Pharmacol Sci 23: 268-274.
Pedder EK, Eveleigh P, Poyner D, Hulme EC, and Birdsall NJM (1991) Modulation of the structure-binding relationships of antagonists for muscarinic acetylcholine receptor subtypes. Br J Pharmacol 103: 1561-1567.
Pertwee RG (1997) Pharmacology of cannabinoid CB1 and CB2 receptors. Pharmacol Ther 74: 129-180.
Pertwee RG (2001) Cannabinoid receptors and pain. Prog Neurobiol 63: 569-611.
Pertwee RG (2002) Cannabinoids and multiple sclerosis. Pharmacol Ther 95: 165-174.
Pertwee RG (2005) Inverse agonism and neutral antagonism at cannabinoid CBI receptors. Life Sci 76: 1307-1324.
Pertwee RG and Ross RA (2002) Cannabinoid receptors and their ligands. Prostaglandins Leukot Essent Fatty Acids 66: 101-121.
Randall MD, Harris D, Kendall DA, and Ralevic V (2002) Cardiovascular effects of cannabinoids. Pharmacol Ther 95: 191-202.
Rice ASC, Farquhar-Smith WP, and Nagy I (2002) Endocannabinoids and pain: spinal and peripheral analgesia in inflammation and neuropathy. Prostaglandins Leukot Essent Fatty Acids 66: 243-256.
Ross RA (2003) Anandamide and vanilloid TRPV1 receptors. Br J Pharmacol 140: 790-801.
Watson C, Jenkinson S, Kazmierski W, and Kenakin T (2005) The CCR5 receptor-based mechanism of action of 873140, a potent allosteric noncompetitive HIV entry inhibitor. Mol Pharmacol 67: 1268-1282.
Zahn K, Eckstein N, Trankle C, Sadee W, and Mohr K (2002) Allosteric modulation of muscarinic receptor signaling: alcuronium-induced conversion of pilocarpine from an agonist into an antagonist. J Pharmacol Exp Ther 301: 720-728.