Abstract
The present report describes in vitro studies demonstrating that the heterocyclic substituted pyridine compound (±)-2-(3-pyridinyl)-1-azabicyclo[2.2.2]octane (RJR-2429) is extremely potent in activating human muscle nicotine ACh receptor (nAChR) (EC50 = 59 ± 17 nM;Emax = 110 ± 09% vs.nicotine). RJR-2429 is markedly less potent in activating nAChRs in the clonal cell line PC12, with EC50 = 1100 ± 230 nM andEmax = 85 ± 20% when compared with nicotine. The activation of a putative α3β4-containing nAChR in PC12 by RJR-2429 reveals a potency intermediate between nicotine and epibatidine (EC50 of 20,000 nM for nicotine and 30 nM for epibatidine). Dose-response curves for agonist-induced ileum contraction indicate that RJR-2429 is equipotent with nicotine, having an EC30 of approximately 2 μM. RJR-2429 binds with high affinity to α4β2 receptor subtype (Ki = 1.0 ± 0.3 nM), and chronic exposure results in significant up-regulation of the high-affinity [3H]nicotine binding sites. In addition, RJR-2429 does not activate nAChRs present in rat thalamic preparations but is a potent inhibitor of this receptor subtype. It antagonizes nicotine-stimulated ion flux in thalamic synaptosomes with an IC50 of 154 ± 37 nM. It also is a potent partial agonist at nAChRs mediating dopamine release from rat synaptosomal preparations (EC50 = 2 ± 1 nM;Emax = 40%; epibatidine and nicotine have EC50 values of 0.4 and 100 nM, respectively). A model for the structure-activity profile of RJR-2429, nicotine and epibatidine was derived by molecular forcefield and quantum mechanics calculations and may provide important clues for the development of ligands selective for nAChR subtypes as probes in the life sciences or as potential therapeutic tools.
Numerous studies have suggested that nicotinic cholinergic pharmacology plays a role in cognitive function both in animals and in humans (Calderon-Gonzalez, 1993; Davies and Maloney, 1976; Decker et al., 1994; Jarvik, 1991; Jones et al., 1992; Josephet al., 1994; Levin and Torry, 1994; Newhouse et al., 1994; Sahakian et al., 1989), in movement disorders (Devor and Isenberg, 1989; Emerich et al., 1991;Hughes and McHugh, 1994; Janson et al., 1994; McConvilleet al., 1991, 1992; Silver et al., 1994) and in certain peripheral disorders (Calkins, 1989). The potential therapeutic benefit of nicotinic ligands in a variety of neurodegenerative pathologies involving the CNS (Smith and Giacobini, 1992) has energized research efforts to develop nicotinic ACh receptor (nAChR) subtype-selective ligands (Baron, 1994; de Fiebre et al., 1995; Freedman et al., 1994; Garvey et al., 1994a,b; Lin et al., 1994; Abreo et al., 1996;Arneric et al., 1995; Cosford et al., 1996). In particular, there has been a concerted effort to develop nicotinic compounds with selectivity for CNS nAChRs as potential pharmaceutical tools in the management of these disorders (e.g., ABT-418 from Abbott laboratories, Decker et al., 1994, and Arnericet al., 1995; RJR-2403 from RJR, Bencherif et al., 1996, and Lippiello et al., 1996; SIB-1508Y from SIBIA, Cosford et al., 1996). The characterization of other novel nicotinic ligands such as epibatidine, that show a marked increase in potency at nAChRs, has provided additional support for the development of potent, selective ligands at individual nAChR subtypes.
Heterogeneity of nAChR subtypes expressed in the vertebrate autonomic nervous system and CNS is at least partly based on the diversity of expression of receptor subunits (Goldman et al., 1987;Boulter et al., 1987; Nef et al., 1988; Deneriset al., 1988, 1991; Couturier et al., 1990;Elogoyhen et al., 1994; for reviews, see Deneris et al., 1991; Lukas and Bencherif, 1992; Sargent, 1993). The potential for these subunits to form diverse receptor subtypes (Alkondon and Albuquerque, 1993; Alkondon et al., 1994) has proved to be the main challenge in the development of target-selective nicotinic ligands. We have developed and studied a number of nicotinic compounds to identify potential candidates that exhibit such selectivity and have recently reported the development of a leading candidate (RJR-2403) as a ligand with CNS selectivity. Functional muscle nAChRs have been shown to be expressed in TE671/RD, a cell line of human origin (Luther et al., 1989; Lukas, 1989). Another model system widely used to study PNS nicotinic pharmacology is the PC12 cell line (a continuous clonal cell line of neural crest origin derived from a tumor of the rat adrenal medulla), which expresses putative ganglionic-type nAChRs. These cells have been shown to exhibit α3, α5, β2, β4and α7 subunit mRNA (Whiting et al., 1991;Rogers et al., 1992). In the present study, we report that one of these compounds, a heterocyclic substituted pyridine derivative (±)-2-(3-pyridinyl)-1-azabicyclo[2.2.2]octane (RJR-2429), is extremely potent at certain nAChRs and shows between a 100-fold and a 1000-fold selectivity over epibatidine. In addition, this derivative is very potent at antagonizing certain CNS receptors (putative α4β2). Therefore, RJR-2429 may have a potential use in at least two regards: as an additional pharmacological tool to assess nAChR heterogeneity and as a potent agonist and/or antagonist of specific nAChRs.
Materials and Methods
Synthesis and Structure
(±)-2-(3-Pyridinyl)-1-azabicyclo[2.2.2]octane difumarate was prepared from the reaction of N-diphenylmethylene)-3-(aminomethyl)pyridine with tetrahydropyran-4-methanol methanesulfonate in the presence of n-butyl lithium/diisopropylamine/THF at -78°C under nitrogen, followed by heating of the intermediate 1-amino-1-(3-pyridyl)-2-(4-tetrahydropyrano)-ethane with HBr at 100°C under pressure. The compound was utilized in pharmacological assays as the racemic difumarate salt (fig. 1).
Model Cell Systems
Clonal cells.
The M10 cell line is a mouse fibroblast clone stably transfected with chick cDNA corresponding to the subunits of the α4β2 nicotinic receptor (kindly provided by Dr. Paul Whiting; Whiting et al., 1991). Cells of the TE671/RD human clone (McAllister et al., 1977) and the PC12 rat pheochromocytoma were maintained in proliferative growth phase in DMEM (GIBCO/BRL, Gaithersburg, MD) supplemented with 10% horse serum, 5% fetal calf serum (Atlanta Biologicals, Norcross, GA) and antibiotics (penicillin/streptomycin) according to routine protocols (Bencherif and Lukas, 1991a, 1993). M10 cells were grown in DMEM containing 10% fetal calf serum according to standard protocols (Bencherif and Lukas, 1991b) except that the antibiotic geneticin (1 mg/ml) was routinely added to the medium to eliminate any revertants (Whiting et al., 1991). Our studies indicate that in the presence of dexamethasone, M10 cells exhibit a high-affinity nicotine binding site with binding characteristics similar to that expressed in mammalian brain.
Rat brain synaptosomes.
Female Sprague-Dawley rats (100–200 g) were killed by decapitation after anesthesia with 70% CO2. Striatal or thalamic tissue was rapidly dissected out and homogenized in 0.32 M sucrose containing 5 mM HEPES, pH 7.4 (7.5 ml per striatum), using a glass/glass homogenizer. The tissue was then centrifuged for 1000 × g for 10 min, and the pellet was discarded. The supernatant was centrifuged at 12,000 ×g for 20 min. The resultant pellet was resuspended in perfusion buffer (128 mM NaCl, 1.2 mM KH2PO4, 2.4 mM KCl, 3.2 mM CaCl2, 1.2 mM MgSO4, 25 mM HEPES, 1 mM ascorbic acid, 0.01 mM pargyline HCl and 10 mM glucose, pH 7.4) and centrifuged for 15 min at 25,000 × g. The final pellet was resuspended in perfusion buffer, 1 ml per striatum. For binding experiments, tissue was frozen until required and then thawed and centrifuged at 48,000 × g. The pellet was resuspended in perfusion buffer.
Functional Studies
Ion flux in TE671/RD and PC12 cells.
After the radioisotope loading period (37°C), cells were washed twice with standard ion flux medium composed of 0.13 M NaCl, 5 mM KCl, 1.8 nM CaCl2, 10 mM glucose and 15 mM HEPES, pH 7.4, and ligands were added to cells plated on micro-wells. 86Rb+ efflux was terminated by three rapid washes in 1 ml of standard ion flux medium. Washed cells were solubilized in 0.01 N NaOH/0.1% sodium dodecyl sulfate (v/w), harvested and assayed for sequestered86Rb+ using Cerenkov counting (∼40% efficiency). Levels of nonspecific ion flux were equivalent, whether defined using samples containing agonist plus 100 μM d-tubocurarine or using blank samples that contained no agonist, and specific nAChR function was defined as total experimentally determined ion flux in the presence of agonist +/− test compounds minus nonspecific ion flux.
Dopamine uptake.
The synaptosomal suspension was incubated for 10 min at 37°C to restore metabolic activity. [3H]-dopamine was added to a final concentration of 0.1 μM, and the suspension was incubated at 37°C for a further 10 min. Then 75-μl aliquots of tissue were added to 96-well microtiter plate wells containing 250 μl of perfusion buffer, harvested onto Gelman AE filters (6 mm in diameter) using an Inotech cell harvester and washed with 2 ml of perfusion buffer.
Dopamine release from striatum and 86Rb+release from thalamus.
Tissue-loaded filters were placed onto Gelman A/E filters 11 mm in diameter on an open-air support. Perfusion buffer containing various compounds as required was dripped onto the tissue through a blunted 21-gauge needle at a rate of 3.2 ml/min using a peristaltic pump. The buffer was drawn through the filter using a second pump with an off-flow rate of 4.5 ml/min. After a 10-min wash period, fractions were collected to establish the basal release, and then agonist was applied in the perfusion stream. Further fractions were collected after agonist application to re-establish the base line. Any change in the base line observed after the removal of agonist was assumed to occur in a linear fashion with time. The perfusate was collected directly into scintillation vials, and released radioactivity was quantified using conventional liquid scintillation techniques. Release of dopamine or 86Rb+ was determined in the presence of various ligands and was expressed as a percentage of the maximal activation induced by nicotine or tetramethylammonium (TMA). The latter acted as a full or nearly full agonist at various nicotinic receptor subtypes and provided more reliable base-line values than nicotine. Antagonism studies were conducted in the presence of maximal activation with TMA and increasing concentrations of RJR-2429.
Ileum contractility.
Male guinea pigs (Hartley strain) 7 to 9 weeks old were used in this paradigm. Ileum was dissected at a point 10 cm from the cecum, and a longitudinal smooth muscle strip was prepared from the isolated ileum and suspended in a 30-ml organ bath. Agonist-induced contraction of ileal tissue was measured using a Magnus test apparatus. Contractions of the strip were induced by successive application of 1 μM ACh until a uniform contraction was produced. Compounds were delivered in a 300-μl volume for 1 min and then washed. Isotonic contraction and changes in the length of the longitudinal smooth muscle strip were expressed as percent contraction compared with that of 1 μM ACh.
Ligand Binding Studies
[3H]-(S)-(−)-nicotine binding.
Cells were mechanically scraped, harvested in cold Tris buffer (5 mM, pH 7.4) and homogenized with a Polytron (Brinkmann Instruments, Westbury, NY; settings at full power for 10 sec). The homogenate was centrifuged at 40,000 × g for 10 min, the supernatant was discarded and the pellet was reconstituted in phosphate-buffered saline (pH 7.4). Standard procedures were used for ligand binding studies at 4°C (Lippiello et al., 1987), and sample aliquots were routinely reserved for determination of protein concentration (Smith et al., 1985) with bovine serum albumin as the standard. Equilibrium binding assays were conducted by incubating membrane aliquots suspended in 300 μl of assay buffer with 10 nM [3H]-(S)-(−)-nicotine (78.4 Ci/mmol; Dupont, New England Nuclear, Boston, MA). Nonspecific binding was determined in samples supplemented with 10 μM nicotine or 1 mM carbachol. Incubation was terminated by rapid filtration on a multimanifold tissue harvester (Brandel, Gaithersburg, MD) using G/C filters presoaked in 0.33% polyethyleneimine (w/v). Chronic exposure studies for determination of agonist-induced up-regulation were performed in cell cultures incubated with agonist for 24 hr before membrane preparations and ligand binding studies.
Data Analysis
Agonist dose-response profiles for activation of nAChR function for TE671/RD and PC12 cells were analyzed according to the logistic equation
Modeling of Nicotine, Epibatidine and RJR-2429
Molecular mechanics calculations.
Molecular mechanics calculations were performed using the universal force field (UFF;Rappé et al., 1992) with electrostatic potential fit atomic charges derived from the AM1 Hamiltonian in MOPAC (Dewaret al., 1985). S(−)-RJR-2429, R-(+)-epibatidine and (S)-(−)-nicotine were analyzed as the protonated forms; in the case of nicotine, the hydrogen of protonation orients cisoid to the inter-ring bond. (S)-(−)-RJR-2429 was chosen over R-(+)-RJR-2429 because root mean square (RMS)-based alignment indicates that the former possesses a higher degree of complementarity to R-(+)-epibatidine and cytisine (data not shown). A first-order approximation based on molecular similarity was used to establish the relative activities of the RJR-2429 isomers. Conformational flexibility of the analogs was assessed using the torsion driver technique, rotating the bond connecting the ring systems through 360° in steps of 4.5°. Each resulting conformer was minimized via the method of conjugate gradients, the variable torsion being held constant using a harmonic potential (force constant 1000 kcal/deg).
Calculation of active conformations.
The nicotinic receptor agonist cytisine was used as a conformational template because its rigid shape defines a plausible relationship between the planar aromatic moiety presenting the H-bond acceptor and the cationic nitrogen. Active conformation hypotheses were generated using graph theory-based alignment tools in the Cerius2 suite of molecular modeling tools (V2.1; Molecular Simulations Inc., 1996). For a given molecule, ten alignments are performed, followed by minimization via conjugate gradients. The alignment said to be the active conformation is the one that possesses the greatest steric and pharmacophore overlap with cytisine (based on RMS difference).
Semiempirical calculations.
Active conformations were further minimized with the semi-empirical quantum mechanics package MOPAC (V6.0; Stewart, 1990). The AM1 Hamiltonian (Dewar et al., 1985) was used in all calculations. Further keyword details are given in the legend of Table 1.
Materials
Unless specifically mentioned above, reagents were purchased from Sigma Chemical Co. (St. Louis, MO) and were of the highest available grade. Radiolabeled ligands were purchased from New England Nuclear (Boston, MA).
Results
Receptor binding.
The binding affinity of RJR-2429 to CNS receptors was determined in membrane preparations from rat brain cortex or from clonal cells transfected with the predominant high-affinity binding protein in the CNS (α4β2; M10 cells). RJR-2429 displaced [3H]-(S)-(−)-nicotine with aKi of 1.0 ± 0.3 nM. Similar results were obtained in both preparations. Analysis of the competition binding curves is consistent with RJR-2429 interacting with a single population of high-affinity binding sites. For comparison, epibatidine was extremely potent in displacing high-affinity [3H]-S(−)-nicotine binding, with aKi of 0.05 nM. As with nicotine, chronic exposure of M10 cells to RJR-2429 produced significant up-regulation of high-affinity [3H]-nicotine binding sites (1.5- to 2.5-fold; fig 2). By comparison, nicotine resulted in a 2- to 3-fold increase in nAChR density. Furthermore, RJR-2429 does not interact with muscarinic receptors, as evidenced by the lack of inhibition of high-affinity [3H]-quinuclydinyl benzilate (3H-QNB) binding in TE671/RD cells, which express abundant muscarinic receptors (Bencherif and Lukas, 1991b). By contrast, atropine inhibited3H-QNB with an IC50 of 10 nM (data not shown).
Receptor function (peripheral subtypes).
To evaluate the effects of RJR-2429 at peripheral nAChRs, we measured the extent of its interaction with nAChR subtypes from preparations exhibiting muscle or ganglionic receptors. The ability of RJR-2429 to interact with ganglionic-type receptors was determined in PC12 cells, which express multiple nAChR subunit genes (α3, α5, β2, β4 and α7) and a functional pharmacologically identified α3β4-containing receptor (Lukas, 1989; see “Discussion”). RJR-2429 was more potent than nicotine in activating nAChRs in PC12 cells (EC50 = 1100 ± 230 nM for RJR-2429 vs. 30,000 nM for nicotine;Emax = 85 ± 20% of nicotine; fig. 3, A and B) but much less potent than epibatidine (EC50 = 30 nM; Sullivan et al., 1996). The interaction of RJR-2429 with muscle nAChRs was evaluated in TE671/RD cells, which are known to express a human muscle nAChR subtype (Luther et al., 1989). RJR-2429 elicited a concentration-dependent activation of muscle receptors with an EC50 of 59 ± 17 nM and efficacy of 110 ± 9% (fig. 3, A and B). This potency is much greater than that of nicotine (80,000 nM). Dose-response curves for ileum contraction revealed a potency very similar to that of nicotine (EC30 of 1.5 μMvs. 2 μM for nicotine; data not shown). The similarity in potencies of nicotine and epibatidine in inducing ileum contraction and the 10-fold difference between their respective potencies at ganglion-type receptors in PC12 cells suggest distinct nAChRs in both preparations.
Receptor function (CNS subtypes).
Other studies were performed to evaluate the agonistic properties of RJR-2429 on CNS nAChR subtypes. We compared its potency and efficacy to those of nicotine and epibatidine for activation of nAChR subtypes expressed in the thalamus and the striatum. For reference, we also evaluated the effects of tetramethylammonium (TMA), which acts as a full agonist at these receptor subtypes.
To assess the ability of RJR-2429 to activate nAChR expressed in the striatum, we evaluated its ability to release dopamine from rat striatal synaptosomes, which contain a putative α4α3β2 nAChR; Clarke and Reuben, 1996). Our results indicate that RJR-2429 is very potent (EC50 = 2.4 ± 0.9 nM) in activating these receptors but is only a partial agonist with efficacy about 40% of that of nicotine (fig 4). For comparison, nicotine has an EC50 of 100 ± 28 nM and is a full agonist. The results also showed that RJR-2429 does not cause any significant activation of 86Rb+ release from rat thalamic synaptosome preparations (EC50 > 100,000 nM for RJR-2429 vs. 591 ± 120 nM for nicotine;Emax < 10% vs. nicotine, which was nearly a full agonist at these receptors, having anEmax of 87 ± 22% vs. TMA; table 1). The binding and functional results indicate that RJR-2429 binds with high affinity to the predominant brain nAChR but does not result in any notable activation of the putative α4β2 receptor found in thalamic preparations. These results raised the possibility that RJR-2429 had antagonistic properties at this nAChR subtype. This was tested by adding increasing concentrations of RJR-2429 to a maximally activating concentration of the agonist TMA or nicotine. The results indicated that after stimulation with 100 μM TMA, RJR-2429 inhibits nAChR function in thalamic preparations with an IC50 of 154 ± 37 nM (fig. 5). By contrast, RJR-2429 was less potent in inhibition of agonist-stimulated dopamine release from striatal preparations, with an IC50 of 2 ± 0.19 μM. The ratio of potencies for activation of thalamic and striatal nAChRs is greater than 50,000 for RJR-2429, in contrast to epibatidine and nicotine, which have selectivity ratios of 2.5 and 5.9, respectively (table 2).
Modeling analysis.
Results of the torsion analysis are given in figure 6. The lower panel shows the resulting potential energy curves; (R)-(+)-epibatidine exhibits the highest degree of conformational flexibility, 38.8% of the found conformations lying within 2 kcal of the global minimum, followed by S(−)-nicotine (23.8%) and (S)-(−)-RJR-2429 (20.0%). Epibatidine possesses potential minima near −153.0, −68.0 and 116.5°, whereas nicotine possesses minima at −123.2 and 58.7° and RJR-2429 possess minima at −135.9 and 45.1°. Only the latter two analogs possess significant rotational barriers (−86.1, 21.9, 89.4 and 166.35° for RJR-2429 and −32.7, 4.3 and 165.4° for nicotine; epibatidine had minor barriers located at −107.8, 7.3 and 71.3°). Nitrogen-to-nitrogen (N-to-N) distance maps from the torsion analysis are shown in the top panel of figure 6. Epibatidine can accommodate a wide range of distances (4.68–5.72 Å, a 1.04-Å range), whereas nicotine and RJR-2429 cover a narrower range of N-to-N distances (4.32–4.96 Å, a 0.64-Å range). Active conformations are compared in figure 6. Table 2 tabulates the active conformation N-to-N distances and inter-ring torsion. Note that the active conformations are located within the global-minima basins, shown in figure 6 as large dots.
The MOPAC minimized conformations differed very little from the starting structures. This result is particularly encouraging, because it indicates that the active conformations hypothesized using molecular mechanics are likely to be in or very near local minima. In the case of epibatidine, the cationic nitrogen resisted direct overlap with the cationic sites of the other ligands because of a disfavored interaction between the pyridyl C2 hydrogen and the azabicycloheptane ring, which led to a greater N-to-N distance. Salient differences between analogs in the quantum mechanical descriptors include molecular dipole, N-to-N distance and atomic charge.
Discussion
The quest for receptor-selective nAChR ligands has motivated the development of compounds that share structural similarities with nicotine. One such compound, RJR-2429, has shown very strong potency at striatal nAChRs and muscle-derived nAChRs, while exhibiting lower potency at ganglionic nAChRs derived from the adrenal gland and showing no agonist properties at nAChRs derived from thalamic tissue.
These results indicate that functional receptors expressed in ganglia can be distinguished pharmacologically from those expressed in thalamic and striatal preparations (Grady et al., 1992). The greater efficacy of cytisine in activating nAChRs in PC12 cells (Emax = 100% vs. nicotine) compared with activation of ion flux from thalamic synaptosomes (Emax = 17%; Marks et al., 1993) supports the hypothesis that different nAChR subtypes mediate these effects. Our finding that RJR-2429 both binds with high affinity to the α4β2 receptor subtype and antagonizes agonist-stimulated ion flux from thalamic synaptosomes is reminiscent of the effects of cytisine at nAChRs in thalamic synaptosomes (Markset al., 1993) and at α4β2 nAChRs expressed in frog oocytes (Papke, 1993; Papke and Heinemann, 1994). These findings are consistent with an identity between nAChRs mediating ion flux in thalamus and the α4β2 nAChR subtype, as suggested previously (Marks et al., 1993). However, as for nicotine, after chronic exposure (1 day) to RJR-2429, we observed a 1.5- to 2.5-fold increase in high-affinity [3H]-nicotine binding sites, a result that supports the notion that up-regulation can occur in the absence of detectable activation.
The partial agonist properties, coupled with the strong potency of RJR-2429 in inducing dopamine release from striatal preparations, make this compound a potentially useful ligand for studying the behavioral effect of nicotinic receptor-mediated dopamine release or as a therapeutic candidate for treatment of Parkinson’s disease, provided that reduction in muscle activity is achieved. This profile contrasts with that of the 5-substituted nicotine analog SIB-1508Y (Cosfordet al., 1996), which shows enhanced efficacy (163 ± 28% of that of nicotine) but a reduced potency (maximal activation occurred at about 300 μM vs. 10 μM for nicotine). Compared with nicotine, SIB-1508Y has a similar potency but reduced efficacy at the α4β2 subtype. These data together indicate that intra-CNS selectivity is achievable and that ligands can be found that activate the release of dopamine and/or other neurotransmitters. The significance of this selectivity rests on the determination of the role of various neurotransmitters in health and disease.
The potency of RJR-2429 in activating ganglion-like nAChRs in PC12 cells and in inducing ileum contraction contrasts with the differential effect of RJR-2403 (Bencherif et al., 1996), which showed no detectable activation of the nAChRs in PC12 cells while maintaining activation of receptors in ileal preparations (data not shown). As suggested previously, our results support the idea that the nAChR subtype(s) mediating GI effects differ(s) from the putative α3β4-containing ganglionic receptor subtype expressed in chromaffin cells (Bencherif et al., 1996). Given the efficacy of cytisine and the sensitivity to neuronal α-bungarotoxin, the pharmacological evidence suggests that the phenotype of nAChRs involved in guinea pig ileum contraction may resemble the α3β2 combination in frog oocytes and may share some similarities with the subtype that underlies neurotransmitter release in some regions of the CNS (Luetje and Patrick, 1991; Papke, 1993; Smith et al., 1993). However, the much greater potency of RJR-2429 compared with nicotine in eliciting dopamine release suggests that this receptor subtype is different from that present in ileal preparations (where both nicotine and RJR-2429 exhibit similar potency).
Molecular modeling was employed to explore possible determinants of specificity for muscle subtype nAChR activation. Although it is beyond the scope of the current manuscript to estimate relative potencies of both enantiomers of RJR-2429, the (S)-(−) enantiomer was chosen because it superimposes with cytisine to a greater extent than (R)-(+)-RJR-2429. Preliminary conformational analysis indicates that RJR-2429 and nicotine have very similar torsion and N-to-N distance profiles. Epibatidine, on the other hand, is remarkably more flexible and adopts a wider range of N-to-N distances. These observations indicate that N-to-N distance and conformational flexibility by themselves do not confer muscle specificity.
A re-evaluation of the nicotinic pharmacophore based on the larger N-to-N distance postulated for epibatidine has been put forth by Glennon and co-workers (Glennon et al., 1994). Our results indicate that although the conformational space of epibatidine does contain larger N-to-N distances, a degree of caution must be exercised when inferring such parameters with flexible molecules. For example, epibatidine’s two deepest minimal-energy wells possess very different N-to-N distances (4.7 Å at −68°; 5.45 Å at 116.5°). It is well established that pharmacophores for individual subtypes differ (Karlin, 1993), and an additional consideration is that different allosteric states (open, closed or desensitized; Galzi et al., 1996) of a given subtype may also possess pharmacophore variation. On the basis of the present data, one might surmise that the open-α1β1γδ pharmacophore may require a smaller N-to-N distance than the open-α4β2or open-α3β4 pharmacophore.
We have evaluated the hypothesis that cytisine, because of its rigidity, potency and high affinity at various nAChR subtypes, may serve as a template for aligning flexible ligands, thus allowing proper presentation of hydrogen bond acceptor and cationic nitrogen relative to overall molecular geometry. Using a combination of graph theory and manual alignment, we present the active conformation hypothesis shown in figure 7. Comparison of these models shows an inverse relationship between functional activity at α1β1γδ and steric hindrance at the cationic nitrogen. Another salient feature is the correlation between hydrophobic bulk to the side of the cationic nitrogen pointing away from the hydrogen bond acceptor and functional activity at α1β1γδ, (vertical lines through the models help in visualization). Studies of the active conformations at the semi-empirical level of theory indicate a number of interesting correlations. Binding to desensitized α4β2correlates with charge on the hydrogen(s) of protonation and, to a lesser extent, with the cationic nitrogen. Rubidium efflux EC50 for this subtype correlates with molecular dipole and weakly with charge on the cationic nitrogen. The EC50 for α3β4 rubidium efflux correlates moderately with N-to-N distance. Finally, the charge of the cationic nitrogen was the only descriptor found to correlate with α1β1γδ activation. The theoretical studies indicate that although the simple calculations described above may yield a first-order approximation of what contributes to subtype specificity, the actual factors are likely to be a less solvent combination of conformational, electrostatic and spatial determinants.
The potency of RJR-2429 at muscle receptors exceeds that reported for other known or recently characterized ligands. The rank order of potency for activation of nAChR (in μM) is [RJR-2429] = 0.06 > [epibatidine] = 0.09 > [isoarecolone] = 0.1 > [suberyldicholine] = 0.2 > [ACh] = 5 > [dimethylacetylpiperazinium] = 10 > [succinyldicholine] = [carbachol] = 30 > [nicotine] = 100 > [cytisine] (See “Results,” data not shown, and Lukas, 1989). These data indicate that RJR-2429 exhibits a window of selectivity for activation of at least four putative subtypes (α4α3β2? (striatal) > α1β1γδ (muscle) > α3β4 (ganglion-like) ≫ α4β2 (thalamic); relative ratio: 1/20/400/antagonist), whereas nicotine and epibatidine show the same selectivity profile (α4α3β2> α4β2 > α3β4> α1β1γδ; relative ratio: 1/6/200/300 for nicotine and 1/2/75/225 for epibatidine). RJR-2429 is most potent at striatal nAChR and the most potent activator of muscle nAChR subtype. In addition, RJR-2429 shows a marked receptor selectivity between neuronal nicotinic receptors in striatum and thalamus, which makes it a potentially useful ligand for unraveling the role of nAChRs in CNS function.
Footnotes
-
Send reprint requests to: Merouane Bencherif, Pharmacology Division, R.J. Reynolds Research & Development, Winston-Salem, NC 27102.
- Abbreviations:
- nAChR
- nicotinic acetylcholine receptor
- DMEM
- Dulbecco’s modified Eagle’s medium
- Received July 28, 1997.
- Accepted November 7, 1997.
- The American Society for Pharmacology and Experimental Therapeutics