Abstract
We investigated the molecular nature of the interaction between the functionally selective M1 muscarinic acetylcholine receptor (mAChR) agonist xanomeline and the human M1 mAChR expressed in Chinese hamster ovary (CHO) cells. In contrast to the non-subtype-selective agonist carbachol, xanomeline demonstrated M1 mAChR binding that was resistant to extensive washout, resulting in a significant reduction in apparentN-[3H]methylscopolamine saturation binding affinity in intact cells. Functional assays, using both M1mAChR-mediated phosphoinositide hydrolysis and activation of neuronal nitric oxide synthase, confirmed that this persistent binding resulted in elevated basal levels of system activity. Furthermore, this phenomenon could be reversed by the addition of the antagonist atropine. However, pharmacological analysis of the inhibition by atropine of xanomeline-mediated functional responses indicated a possible element of noncompetitive behavior that was not evident in several kinetic and equilibrium binding experimental paradigms. Taken together, our findings indicate for the first time a novel mode of interaction between an mAChR agonist and the M1 mAChR, which may involve unusually avid binding of xanomeline to the receptor. This yields a fraction of added agonist that is retained at the level of the receptor compartment to persistently bind to and activate the receptor subsequent to washout. The results of the current study suggest that elucidation of the mechanism or mechanisms of interaction of xanomeline with the M1 mAChR is particularly important in relation to the potential therapeutic use of this agent in the treatment of Alzheimer’s disease.
A significant portion of the ongoing research efforts into the treatment of Alzheimer’s disease has been devoted to the attempted restoration of central cholinergic function (Giacobini, 1992). This is primarily due to the impressive body of evidence pointing to a consistent loss of presynaptic cholinergic markers, such as choline acetyltransferase and M2 mAChRs, associated with the disorder (Mashet al., 1985; Quirion et al., 1989). Although the functional properties of postsynaptic M1 mAChRs, which are involved in learning and memory, also have been shown to be affected during the progression of the disorder (Ferrari-Dileo et al., 1995; Ladner et al., 1995), overall postsynaptic receptor numbers are not (Ferrari-Dileo et al., 1995).
Consequently, many of the pharmacotherapeutic agents spawned as a result of the cholinergic hypothesis of dementia have consisted of either acetylcholinesterase inhibitors or directly acting mAChR agonists (Giacobini, 1992) in an attempt to exploit any residual M1 mAChR functionality. Xanomeline (Fig.1) is a novel member of the latter class of drugs and has been identified as both highly potent and functionally M1 selective (Shannon et al., 1994). Clinical studies that used this agent have revealed a significant improvement in the cognitive function of patients with Alzheimer’s disease (Bodick et al., 1997). Because xanomeline represents one of the most potent M1 mAChR agonists known to date (Shannon et al., 1994), the current study was undertaken to delineate further the molecular mechanisms behind the potency and functional selectivity of this compound. In particular, the possibility of a unique mode of mAChR activation by xanomeline, which is not common to classic mAChR ligands, was investigated both functionally and in binding studies. Our results demonstrate that xanomeline is able to bind to and activate the M1mAChR in a persistent manner and that this mode of interaction is different from that used by conventional agonists such as CCh. A preliminary report of the persistent M1mAChR-activating ability of xanomeline was made recently (Christopoulos and El-Fakahany, 1997).
Experimental Procedures
Materials.
[3H]NMS (84.5 Ci/mmol) and l-[14C]citrulline (50 mCi/mmol) were purchased from DuPont-New England Nuclear (Boston, MA).myo-[3H]Inositol (107 mCi/mmol) andl-[3H]arginine (64 Ci/mmol) were purchased from Amersham Life Science (Arlington Heights, IL). [14C]Inositol-1-phosphate (300 mCi/mmol) was purchased from American Radiolabeled Chemicals (St. Louis, MO). DMEM was purchased from GIBCO BRL (Gaithersburg, MD). Geneticin and hygromycin were purchased from Calbiochem (La Jolla, CA). Bovine calf serum was purchased from Hyclone (Logan, UT). Xanomeline tartrate was a generous gift from Lilly Research Laboratories (Indianapolis, IN). All other reagents were purchased from Sigma Chemical (St. Louis, MO).
Cell culture.
CHO cells, stably expressing the human M1 mAChR (CHO hm1 cells), were kindly provided by Dr. M. Brann (University of Vermont Medical School, Burlington, VT) and were grown for 4 days at 37° in DMEM supplemented with 10% bovine calf serum and 50 μg/ml geneticin in a humidified atmosphere consisting of 5% CO2/95% air. CHO cells stably expressing both the M1 mAChR and nNOS (CHO hm1/nNOS cells) were grown under similar conditions but in the additional presence of hygromycin (50 μg/ml). The cDNA encoding nNOS was kindly provided by Drs. S. H. Snyder and D. R. Bredt (The Johns Hopkins University School of Medicine, Baltimore, MD). Cells were used 4 days after subculturing and were harvested by trypsinization followed by centrifugation (300 × g, 3 min) and resuspension of the pellet in HEPES buffer (110 mm NaCl, 5.4 mm KCl, 1.8 mm CaCl2, 1 mm MgSO4, 25 mmglucose, 50 mm HEPES, 58 mm sucrose; pH 7.4 ± 0.02; 340 ± 5 mOsm), repeated twice.
Assay of nNOS activity.
The activity of nNOS in CHO hm1/nNOS cells was assayed by quantifying the conversion ofl-[3H]arginine tol-[3H]citrulline according to the method of Bredt and Snyder (1989) with modifications (Wang et al., 1994). Subsequent to harvesting and washout, CHO hm1/nNOS cells (∼5 × 105/tube) were incubated in 0.3 ml of HEPES buffer for 15 min at 37°, after whichl-[3H]arginine (0.06 μCi/tube) was added to initiate the reaction. At the same time, increasing concentrations of either CCh or xanomeline were added in the absence or presence of atropine (30 nm) or pirenzepine (150 nm). Antagonists were incubated with the cells for 20 min before the addition of agonist. After 1 hr at 37°, the reaction was stopped with an excess of l-arginine (5 mm) and EDTA (4 mm).l-[3H]Citrulline was separated froml-[3H]arginine using ion exchange chromatography (DOWEX AG50W-X8 resin), and the amount of radioactivity, expressed as dpm, was determined via liquid scintillation counting. In other experiments, the residual effects of xanomeline pretreatment on basal nNOS activity were examined in terms of time dependence, concentration dependence, and reversibility by atropine (10 μm), as outlined in Results. In these latter experiments, the l-[3H]arginine was added subsequent to all washout procedures.
Assay of PI hydrolysis.
CHO hm1 cells were suspended in buffer and loaded withmyo-[3H]inositol (8 μCi/ml) for 1 hr at 37°. Labeled cells were washed with HEPES buffer containing 10 mm LiCl, distributed to assay tubes (∼5 × 105/tube), and allowed to incubate for 15 min at 37°. Concentration-response curves for the stimulation of PI hydrolysis by either CCh or xanomeline were constructed in the absence or presence of 30 nm atropine. The antagonist was equilibrated with the cells for 20 min before the addition of agonist. To assess agonist-induced desensitization, experiments also were conducted on cells that had been pretreated for 24 hr at 37° with either 1 mm CCh or 1 μm xanomeline before harvesting. With a third experimental protocol, we examined basal PI activity in cells that had been pretreated with either 1 mmCCh or 1 μm xanomeline for 1 hr at 37° before harvesting and washing. In all instances, the reaction was allowed to proceed for 1 hr after the addition of agonist or vehicle control before being stopped with chloroform/methanol (2:1). Tubes were centrifuged (450 × g; 15 min), and total inositol phosphates were separated by ion exchange chromatography on DOWEX AG1-X8 resin, with [14C]inositol-1-phosphate as a recovery standard. The amount of radioactivity (dpm) in each sample then was determined by liquid scintillation counting.
Saturation binding experiments.
All radioligand binding experiments used ∼105 cells/assay tube in a total volume of 1 ml. For the saturation binding assays, three different protocols were used: (1) to determine radioligand binding parameters, cells were incubated with increasing concentrations of [3H]NMS (0.02–2 nm) for 1 hr at 37°; and (2) to assess persistent agonist binding, CHO hm1 cells were incubated for 1 hr at 37° in the absence or presence of either 1 mm CCh or 1 μm xanomeline, followed by centrifugation and resuspension in HEPES buffer (three times), as described above. Subsequently, the cells were incubated with a saturating concentration (2 nm) of [3H]NMS for 1 hr at 37°; (3) to test for pseudoirreversible agonist binding, CHO hm1 cells were incubated for 2.5 hr at 37° with either 2 nm[3H]NMS alone or in combination with either 100 μm CCh or 100 nm xanomeline, with variations in the order of ligand addition. Agonists were added together with, 30 min before, or 30 min after the radioligand. In all instances, nonspecific binding was defined using 10 μm atropine. Additional saturation experiments were conducted on cells that had been pretreated for 1 hr at 37° with 1 μm xanomeline before washout, as above. Protein determinations were performed according to the method of Bradford (1976).
Competition binding experiments.
CHO hm1 cells were incubated with a fixed concentration of [3H]NMS (0.2, 0.4, or 2 nm, as described in Results) in the absence or presence of either CCh (0.1 μm to 10 mm), xanomeline (0.1 nm to 20 μm), or atropine (0.1 nm to 2 μm) for 1 hr at 37°. Additional atropine competition experiments were conducted in the presence of both [3H]NMS and a fixed concentration of xanomeline (30 or 100 nm). Nonspecific binding was defined as above.
Dissociation kinetic experiments.
CHO hm1 cells were incubated with a fixed concentration of either 0.2 nm[3H]NMS alone or 0.4 nm[3H]NMS together with 30 nmxanomeline for 1 hr at 37°. After this period, 200 nmunlabeled NMS was added, alone or in combination with 1 μm xanomeline, to inhibit radioligand reassociation. Subsequently, the amount of radioactivity was measured at various time intervals to determine the [3H]NMS dissociation rate. Nonspecific binding was defined as above.
For all binding experiments, incubation was terminated by filtration through Whatman GF/C filters (Clifton, NJ), positioned on a Brandell Cell Harvester (Montreal, Quebec, Canada). Filters were washed three times with 4-ml aliquots of ice-cold saline and dried before radioactivity (dpm) was measured using liquid scintillation counting.
Data analysis.
Datasets of total and nonspecific binding, derived from each complete saturation binding assay, were simultaneously analyzed via nonlinear regression with Microsoft Excel 5.0 to derive individual estimates of Bmax(total receptor density) and KD
(radioligand receptor equilibrium dissociation constant). Competition binding isotherms were analyzed via nonlinear regression using Prism 2.01 (GraphPAD Software, San Diego, CA) to derive estimates ofnH
(slope factor) and IC50 (midpoint location/potency parameter). Assuming simple competition, the data were refitted according to both one- and two-site mass action binding models, and the better model was determined by an extra-sum-of-squares test using Prism. IC50 values were converted toKI
values (competitor receptor equilibrium dissociation constant) according to the equation:
In the functional assays, individual agonist concentration-response curve data, in the absence and presence of antagonist, were fitted to the four-parameter logistic function (using Prism):
Data shown are the mean ± standard error. Comparisons between mean values were made by paired or unpaired t tests or one-way ANOVA, as appropriate. Unless otherwise stated, values ofp < 0.05 were taken as significant.
Results
Characterization of the concentration-response relationship for xanomeline and CCh in the activation of nNOS and PI hydrolysis.
Initial experiments established the concentration-response characteristics of xanomeline- and CCh-mediated responses under standard assay conditions. In terms of nNOS activation, quantified as the conversion of l-[3H]arginine tol-[3H]citrulline, it may be seen from Fig. 2 that xanomeline displayed an intrinsic activity similar to that of the full agonist, CCh. In addition, both agents exhibited a very steep concentration-response relationship, as demonstrated by the large curve slopes, with xanomeline being significantly more potent (p< 0.05). Similarly, xanomeline was significantly (p < 0.05) more potent than CCh in mediating PI hydrolysis (Fig. 3), with comparable intrinsic activity.
Pharmacological analysis of the antagonism by atropine and pirenzepine of CCh- and xanomeline-mediated activation of nNOS.
In subsequent experiments, concentration-response curves to either CCh- or xanomeline-mediated nNOS stimulation were established in the absence and presence of the nonselective antagonist atropine (30 nm) or the M1-selective antagonist pirenzepine (150 nm). Time course experiments indicated no significant effect of the antagonist on the initial rate of agonist-induced l-[3H]citrulline accumulation (data not shown). The results of the competition experiments are summarized in Fig. 4 and Table 1. The addition of either antagonist did not result in a significant alteration of agonist concentration-response curve parameters (p > 0.05), allowing the relative changes in agonist potency to be assessed in terms of a competitive model of pharmacological interaction. Based on the results shown in Fig. 4 and Table 1, atropine and pirenzepine were able to shift the CCh concentration-response curves by factors of ∼26- and ∼18-fold respectively, whereas for xanomeline, the corresponding shifts were ∼20- and ∼14-fold, respectively.
Pharmacological analysis of the antagonism by atropine of CCh- and xanomeline-mediated PI hydrolysis.
Similar experiments were conducted using PI hydrolysis as the biological response and atropine (30 nm) as the antagonist. In contrast to thel-[3H]citrulline assay, significant differences were noted in the ability of atropine to inhibit responses mediated by xanomeline compared with those of CCh. Although the antagonism of the CCh-induced PI response by atropine was compatible with a competitive interaction, characterized by a geometric mean curve shift of ∼21-fold, the presence of atropine resulted in a significant (p < 0.05) alteration of control xanomeline concentration-response curve slope in addition to midpoint location (Fig. 5, Table2). The steepening of the concentration-response curve is indicative of an interaction that is not simply competitive. The midpoint location parameters of the xanomeline concentration-response curves were used to derive a geometric mean shift of ∼14-fold, although it should be noted that this value may be considered relatively arbitrary because it will be influenced by the lack of parallelism between the two curves. With this caveat in mind, a significant difference (p < 0.05) was found between the degree of shift induced by atropine of the concentration-response curve to either agonist.
Persistent receptor activation by xanomeline; concentration and time dependence.
In light of these findings, the possibility of other unique features of the interaction between xanomeline and M1 mAChR was investigated further. Incubation of CHO hm1/nNOS cells with increasing concentrations of xanomeline (10–100 nm) for 1 hr at 37°, before harvesting and extensive washout, resulted in a progressive increase in basal conversion of l-[3H]arginine tol-[3H]citrulline in the absence of further addition of agonist, as shown in Fig.6A. Although a trend is evident, a one-way ANOVA detected a significant difference (p < 0.05) above basal only for the two highest preincubation concentrations of xanomeline (100 nmand 1 μm). Fig. 6B shows the results of pretreating the cells with 1 μm xanomeline for various times (10–30 min), where differences over basal levels of activity after washing were observed for 10 min of agonist pretreatment, the earliest time point measured.
Effects of CCh and xanomeline pretreatment on PI hydrolysis.
Additional functional experiments were undertaken to assess whether the persistent activating effects of xanomeline outlined above were due to an artifact particular to the choice of biological assay that was used. Thus, the ability of M1 mAChRs to mediate PI hydrolysis was exploited as a second functional experimental paradigm. CHO hm1 cells were pretreated with either 1 mm CCh or 1 μm xanomeline for 1 hr at 37° before washout and harvesting, as above (Fig. 7A). In contrast to CCh, xanomeline pretreatment, followed by washout, resulted in a persistently elevated level of basal PI activity, being ∼20-fold greater than control cells pretreated in parallel with vehicle. Xanomeline thus displays a novel, persistent, receptor-activating ability in at least two common functional indicators of M1 mAChR activation.
In addition, the effects of long term xanomeline pretreatment of CHO hm1 cells were examined (Fig. 7B). Cells were preincubated with either 1 mm CCh or 1 μm xanomeline for 24 hr at 37° and then washed and harvested. Subsequently, full concentration-response curves to CCh were established. In this manner, any changes in lower curve asymptote, upper curve asymptote, midpoint location, and slope may be interpreted as effects of long term agonist treatment on basal activity, maximal drug-mediated activity, drug potency, and sensitivity, respectively. As shown in Fig. 7B, prolonged treatment with either agonist resulted in a marked reduction in the ability of CCh to stimulate PI hydrolysis. A one-way ANOVA revealed a significant (p < 0.05) decrease in the slope, midpoint location, and maximal asymptote of the curves generated in the agonist-pretreated cells in comparison to the control cells. The degree of reduction in potency and sensitivity of the assay system to CCh, in response to pretreatment with either agonist, was similar (p > 0.05; Fig. 7B), whereas the maximal effects of CCh varied with the agonist used. Of particular note is the fact that basal PI levels were not significantly different (p > 0.05) across the groups, indicating that the persistent activating effects of xanomeline are lost on prolonged (24-hr) treatment.
Effect of atropine on xanomeline-pretreated cells.
We demonstrated previously that coincubation of CHO hm1/nNOS cells with both xanomeline (1 μm) and the antagonist atropine (10 μm) during the 1- hr pretreatment phase abolished the ability of xanomeline after its washout to enhance basal nNOS activity, suggesting a receptor-specific mechanism (Christopoulos and El-Fakahany, 1997). In the current study, this interaction was investigated further. Cells were allowed to incubate for 1 hr with 1 μm xanomeline in the absence of atropine and subsequently were washed extensively and assayed for basall-[3H]citrulline formation in the absence or presence of atropine (10 μm). As shown in Fig.8, the subsequent addition of atropine to xanomeline-pretreated cells resulted in a significant (p < 0.05) reduction in the ability of xanomeline to persistently activate the receptor, suggesting an apparent competitive interplay between the two agents at the M1 mAChR. In a parallel series of experiments, cells were preincubated with xanomeline for 1 hr, followed by the addition of atropine for an additional 1 hr. After extensive washout, the persistent xanomeline effect was reestablished (Fig. 8, third bar).
Effects of xanomeline pretreatment on the saturation binding of [3H]NMS.
The results of the functional assays indicated a mechanism of interaction of xanomeline that displayed elements of both competitive and noncompetitive, or at least surmountable and insurmountable, behavior. Radioligand binding studies were undertaken to probe more directly the nature of the interaction between xanomeline, the M1 mAChR, and classic muscarinic ligands. Initially, CHO hm1 cells were subjected to the same pretreatment schedule as used in the functional assays (i.e., 1-hr incubation at 37° with either 1 mm CCh or 1 μm xanomeline, followed by extensive washout). The subsequent ability of a saturating concentration (2 nm) of the classic antagonist [3H]NMS to bind to the M1 mAChR in agonist-pretreated cells was compared, in parallel, with that in vehicle-pretreated cells. Fig.9A shows that pretreatment with xanomeline resulted in a significant reduction (p < 0.05) in the ability of [3H]NMS to bind to the M1mAChR. In contrast, pretreatment with CCh resulted in radioligand binding that was indistinguishable from that observed in controls. The effects on [3H]NMS binding observed in this type of study, however, may have been due to either persistent competition between xanomeline and the radioligand or xanomeline-induced receptor internalization. To distinguish between the two possibilities, complete saturation binding isotherms were established for [3H]NMS in control and xanomeline-pretreated cells. As indicated in Fig. 9B, xanomeline caused a significant (p < 0.05) reduction in radioligand affinity (pKD = 9.86 ± 0.20 for vehicle versus 9.58 ± 0.12 for xanomeline-pretreated cells) with no significant alteration (p > 0.05) in maximal cell-surface receptor density (Bmax = 979 ± 157 fmol/mg protein for vehicle versus 915 ± 130 fmol/mg protein for xanomeline-pretreated cells; four experiments). Thus, xanomeline pretreatment resulted in a 2-fold displacement of the [3H]NMS concentration-occupancy curve. If a competitive interaction is assumed, then an approximate calculation of the residual xanomeline concentration in the receptor compartment may be made using the following linear metameter of the Schild equation (Arunlakshana and Schild, 1959): fold-shift = 1 + [X]/KX , where [X] is the residual concentration of xanomeline, and KX is the xanomeline-receptor dissociation constant. Using a value of 30 nm for KX , as determined in the competition binding assays (see below), yields a value of 30 nm for [X].
Inhibition by CCh, xanomeline, and atropine of [3H]NMS binding.
The concentration-occupancy relationship of xanomeline and its interaction with atropine at the M1 mAChR were studied indirectly by using a variety of [3H]NMS competition binding assays, conducted over 1 hr at 37° in CHO hm1 cells. Under standard assay conditions (0.2 nm [3H]NMS, approximate KD concentration), xanomeline was a potent inhibitor of radioligand binding (Fig.10A). Nonlinear regression analysis yielded parameter estimates that did not seem to deviate from simple competition between the two ligands at a single binding site, with a calculated xanomeline affinity of ∼30 nm (Table3). For the sake of comparison, the effects of CCh (Fig. 10A, Table 3) and atropine (Fig. 10B) also were examined under similar conditions. Unlike xanomeline, CCh displayed a shallow concentration-occupancy relationship, with the midpoint slope significantly less than unity (p < 0.05), and seemed to recognize both high (6 μm) and low (178 μm) affinity states of the receptor. In contrast, atropine seemed to interact with a single binding site characterized by a dissociation constant of ∼3 nm. Identical patterns of radioligand inhibition were observed in experiments using cell homogenates (data not shown), indicating that agonist-induced desensitization was not a contributing factor to the competition binding curve profile in the intact cells.
A common feature of noncompetitive interactions between multiple ligands at a receptor, however, is that they may not be detected at low concentrations of radioligand but tend to become more evident as higher concentrations of the latter are used (Lee and El-Fakahany, 1991). Thus, competition experiments also were conducted between xanomeline and 2 nm [3H]NMS (Fig. 10A, Table3). Even in the presence of a 10-fold-higher concentration of [3H]NMS, the binding of xanomeline did not seem to deviate from competitive, surmountable behavior, displaying a characteristic parallel, dextral shift in its competition binding isotherm. Neither the calculated affinity nor the slope factor for the interaction of xanomeline with the receptor under these conditions was significantly different (p > 0.05) from those determined in the presence of 0.2 nm[3H]NMS (Table 3).
The findings in the functional assays with atropine, however, prompted further competition binding assays, in which the ability of the antagonist to inhibit the binding of varying concentrations of [3H]NMS was studied in the concomitant presence of increasing concentrations of xanomeline (Fig. 10B, Table 3). If the ligands were acting noncompetitively in a concentration-dependent manner, then deviations should be noted in the binding profile of atropine in the presence of xanomeline. However, as may be seen in Table 3, the apparent affinity of atropine for the M1 mAChR did not change significantly (p > 0.05) under any of the assay conditions, provided the occupancy of the receptor by both the radioligand and xanomeline were accounted for (see Experimental Procedures). These findings are in contrast to the suggestion of noncompetitivity for the interaction between xanomeline and atropine at the M1 mAChR in the functional assays.
Effects of xanomeline on [3H]NMS dissociation.
Noncompetitive interactions can alter receptor conformation such that the dissociation characteristics of the radioligand also would change (Lazareno and Birdsall, 1995; Kostenis and Mohr, 1996). Therefore, radioligand kinetic experiments were undertaken in CHO hm1 cells to determine the dissociation rate constant (koff) of [3H]NMS in the absence and presence of a high concentration (1 μm) of xanomeline. After a 1-hr radioligand/receptor equilibration period at 37°, dissociation was promoted by isotopic dilution with 200 nm of unlabeled NMS. In each instance, radioligand dissociation was better described by a monoexponential model, as determined by Prism. As may be seen in Fig.11A, the presence of xanomeline had no significant effect (p > 0.05) on [3H]NMS dissociation, with thekoff values being 0.25 ± 0.03 and 0.23 ± 0.05 min−1 in the absence and presence of xanomeline, respectively. To discount the possibility that any effects of xanomeline on the dissociation characteristics of [3H]NMS were not evident due to slow xanomeline kinetics at [3H]NMS-occupied receptors, additional experiments were conducted using cells that were preequilibrated with both [3H]NMS (0.4 nm) and xanomeline (30 nm) before promotion of radioligand dissociation. Even under these conditions, the presence of xanomeline (1 μm final concentration) resulted in no significant (p > 0.05) alteration of radioligand dissociation rate (koff = 0.34 ± 0.05 min−1), providing further evidence for a competitive interaction at the classic ligand binding site.
Effects of order of ligand addition on attainment of [3H]NMS equilibrium binding.
If the persistent mechanism of xanomeline binding involved interaction with a significant subset of attachment points on the receptor shared by classic muscarinic ligands, then this would be expected to result in significantly different degrees of reduced radioligand binding at equilibrium, dependent on the order of ligand addition. This latter possibility was assessed by allowing CHO hm1 cells to incubate with 2 nm [3H]NMS for 2.5 hr at 37° in the absence or presence of either 100 μm CCh or 100 nm xanomeline, added together with, 30 min before, or 30 min after the radioligand. Fig. 11B summarizes the results of these experiments, in which it may be seen that the control level of [3H]NMS binding (598 ± 28 fmol/mg protein) was reduced to a similar extent by either agonist regardless of the order of addition.
Discussion
In agreement with previous reports (Shannon et al., 1994), the current study has identified xanomeline as a potent M1 agonist in functional assays of nNOS activation and PI hydrolysis (Figs. 2 and 3). Furthermore, the very steep slopes associated with both CCh and xanomeline-mediated nNOS activation (Fig. 2, Table 1) are indicative of an exquisite degree of sensitivity of the transduction pathways within the CHO hm1/nNOS cells in the generation of the this response in comparison to agonist-mediated PI hydrolysis (Fig. 3, Table 1). This is consistent with the known, nonlinear dependence of the former biochemical process on the latter (Wang et al., 1994). However, the persistent nature of the interaction of xanomeline with the mAChR, as well as the reduced potency of atropine and pirenzepine in functionally antagonizing the effects of xanomeline compared with CCh, are novel findings that have significant implications in terms of understanding alternate modes of mAChR activation and therapeutic use of xanomeline. The results of the current study investigating these phenomena may be summarized as follows: (1) a fraction of added xanomeline, unlike CCh, is resistant to extensive washout and persists within the receptor compartment to exert an agonistic effect; (2) this effect seems to be concentration dependent and is manifested within a 10-min contact time of agent with receptor before washout; (3) the phenomenon transcends the type of experimental assay used; (4) the effect is susceptible to reversal by atropine but is reestablished on removal of atropine; (5) whereas the interaction of xanomeline with the M1mAChR seems to be largely competitive, certain elements of this interaction are not completely compatible with simple, surmountable competition.
Persistent increases in basal M1 mAChR-mediated nNOS activity and PI hydrolysis were observed after short term incubation of CHO cells with xanomeline (Figs. 6A and 7A). However, increasing the pretreatment time to 24 hr resulted in a return of basal PI hydrolysis to control levels (Fig. 7B). This implies that despite being able to activate the receptor in a persistent manner, the xanomeline effect eventually succumbs to the desensitization machinery of the intact cell, a feature common to the majority of agonist-induced responses in physiological systems (Nathanson, 1989; Lohse, 1993). The results shown in Fig. 7B also indicate that xanomeline induced a significantly lower degree of system desensitization than did CCh, under conditions of approximate receptor-occupancy equivalence with respect to pretreatment concentration (∼20 × IC50 value, Fig. 10A). The ability of agonists to induce desensitization may be correlated with their efficacy (Huet al., 1991; January et al., 1997); hence, our findings with xanomeline tend to support previous claims that this agent is actually a partial agonist (Shannon et al., 1994). In addition, the competition binding experiments also support this conclusion because CCh, but not xanomeline, yielded a biphasic competition binding curve (Fig. 10A). As with the ability to induce desensitization, deviations from monophasic competition binding at a single receptor subtype also may be related to the efficacy of the agonist, being less pronounced with partial agonists in comparison to full agonists (Zhu et al., 1994). The loss of the persistent activating effect of xanomeline after prolonged (24-hr) treatment may simply be a reflection of the fact that given sufficient time, partial agonists also may induce desensitization of their own agonistic effects, especially if used in high concentrations (Huet al., 1991).
The simplest explanation to account for the persistent activating effects of xanomeline relates to the lipophilic nature of the compound. The aliphatic hexyloxy side chain (Fig. 1) may be predicted to contribute to the ability of this compound to partition indiscriminately into biological membranes, resulting in the development of a substantial depot of agonist that can demonstrate relative resistance to washing. However, three experimental observations point to a more specific receptor-mediated mechanism of interaction that can account for the effects of xanomeline. First, our previous demonstration that coincubation of xanomeline with atropine in the pretreatment phase prevents the ability of the former agent to display persistent receptor activation (Christopoulos and El-Fakahany, 1997). This suggests that a prerequisite for the establishment of the phenomenon is the ability of xanomeline to gain access to regions on the receptor that are also used by atropine. Second, the reappearance of the residual agonistic activity of xanomeline, after washout of atropine (Fig. 8), suggests that the persistent binding of xanomeline occurs in regions of close proximity to the receptor. Third, the lipophilicity of this compound may actually result in a significant degree of its depletion at the level of the receptor compartment. Functionally, this would be manifested as an increased potency of a competitive antagonist in the inhibition of xanomeline-mediated responses compared with responses elicited by a hydrophilic agonist, such as CCh. As shown in Figs. 4 and 5 and Tables 1 and 2, however, in neither type of functional assay did atropine or pirenzepine seem to be better able to inhibit the responses to xanomeline relative to those of CCh. In fact, the opposite was observed. This supports the hypotheses that the interaction between xanomeline and atropine involves a receptor-mediated mechanism and that this mechanism is different in nature from that between atropine and CCh.
Although the above considerations point to a receptor-specific mechanism to account for the persistent actions of xanomeline, aspects of the functional assays suggest that this mechanism may be noncompetitive in nature when assessed in terms of the interaction between xanomeline and atropine. This was particularly evident in the PI assay, in which significant differences in xanomeline concentration-response curve shape were detected in the absence and presence of atropine. Furthermore, the antagonist exhibited a reduced potency in inhibiting xanomeline-mediated responses in comparison to CCh-mediated responses (Fig. 5). Theoretically, the magnitude of the shift that a competitive antagonist is able to produce on an agonist concentration-response curve should be independent of the potency and efficacy of the agonists used, provided these agonists activate the response via the same receptor (Kenakin, 1997). The reduced potency of atropine in inhibiting xanomeline-mediated responses, relative to CCh-mediated responses, cannot be attributed to M1-independent receptor-activating properties of xanomeline because no responses to this agent were observed in nontransfected CHO cells (data not shown). Thus, it is possible that the persistent nature of xanomeline binding introduces additional steric or conformational effects that confound the ability of antagonists to bind. However, the unambiguous assignment of a noncompetitive mode of interaction between agonist/antagonist pairs would require a more rigorous pharmacological analysis, using a range of antagonist concentrations, than that undertaken in the current study. Furthermore, it is quite possible that the differential abilities of CCh and xanomeline to desensitize the M1 mAChR also may contribute to the apparent noncompetitive interaction observed in the functional assays, not by altering antagonist binding properties but instead by changing the agonist concentration-response relationship at the higher agonist concentrations that are required to overcome the antagonism by atropine.
A number of radioligand binding paradigms were used to assess further the mechanism or mechanisms of interaction of xanomeline with the M1 mAChR. Increasing the [3H]NMS concentration 10-fold resulted in a parallel, dextral shift of the xanomeline competition curve (Fig. 10A), with complete inhibition of radioligand binding still achievable. Such a profile is predicted for competitive interactions, in contrast to noncompetitive interactions that deviate from this behavior when progressively higher radioligand concentrations are used (Lee and El-Fakahany, 1991). Furthermore, the affinity of xanomeline, calculated with the assumption of competition, was not significantly different in either instance, demonstrating the independence of this characteristic drug/receptor parameter from radioligand concentration. Ligand-induced conformational changes also may be detected as alterations of radioligand dissociation rate in the presence of a modulating agent (Lee and El-Fakahany, 1991; Lazareno and Birdsall, 1995; Kostenis and Mohr, 1996). Again, this possibility was not substantiated in the dissociation kinetic experiments (Fig. 11A), in which the control radioligand dissociation rate was unaltered by the presence of xanomeline. If the mechanism of interaction involved an essentially pseudoirreversible steric phenomenon, such as a persistent “capping” of the classic ligand binding site, then the order of ligand addition would be expected to result in differing degrees of steady state binding (Proška and Tucek, 1994). This was not observed; the combination of both xanomeline and [3H]NMS resulted in similar levels of steady state binding, regardless of the order of addition (Fig. 11B). Finally, if the interaction were noncompetitive only between the specific combination of atropine and xanomeline, then the [3H]NMS/xanomeline/atropine competition experiments (Fig. 10B) should have provided some evidence for this phenomenon, yet none was noted (Table 3).
A finding consistent with a competitive mode of interaction between residual xanomeline and other ligands at the classic binding site on the mAChR also was observed in the [3H]NMS saturation binding experiments (Fig. 9), in which pretreatment with xanomeline, followed by washout, affected apparent radioligand binding affinity but not maximal receptor density. As calculated from Fig. 9B, pretreatment with 1 μm xanomeline, followed by extensive washout, resulted in the equivalent of ∼3.3% of the added xanomeline displaying a tenacious binding ability. Furthermore, the calculated residual concentration of 30 nm is sufficient to account for the elevation of basal response to near-maximal levels of receptor activation (Figs. 6A and 8), as determined by inspection of the complete xanomeline concentration-response curves in thel-[3H]citrulline assay (Figs. 2 and4). At this point, it is worth noting that the residual concentration and, indeed, calculated affinity constants for xanomeline determined in this study are, quite possibly, composite measures, with their true values being dictated by the relative contributions of both reversible and pseudoirreversible binding components. At the very least, however, they may still be used as empirical descriptors of the overall binding phenomenon.
While attempting to construct an explanatory scheme that encompasses our experimental findings, it is evident that much of the behavior exhibited by xanomeline is reminiscent of observations made with the β2-adrenoceptor agonist salmeterol (Colemanet al., 1996). The latter compound has been studied extensively due to its ability to display persistent binding and an extremely long duration of action (Nials et al., 1993). To explain the findings with salmeterol, an “exosite” model has been postulated, in which the lipophilic phenylalkyloxalkyl “tail” of the compound attaches itself to a complementary region on the receptor, distinct from the classic agonist binding site, in a pseudoirreversible manner (Green et al., 1996), whereas the saligenin “head” of the compound interacts in a reversible manner with the classic site to activate the receptor. However, the study of the interaction between salmeterol or xanomeline with competitive antagonists also points to important differences between the two agonists. Radioligand binding experiments with salmeterol have suggested a pseudoirreversible mechanism of attachment, whereas functional experiments were consistent with surmountable competition (Nials et al., 1993; Clark et al., 1996). Conversely, xanomeline seemed competitive in binding assays, as reported herein. In addition, coincubation of salmeterol with competitive antagonists did not prevent the establishment of a persistent receptor-activating effect of the agonist after antagonist washout (Clark et al., 1996). This latter finding is in contrast to our observations with xanomeline and atropine coincubation followed by washout (Fig. 8), in which the residual agonistic effects of xanomeline reappeared. Thus, although our data suggest that xanomeline may be interacting with an exosite on the M1 mAChR, the mechanism of binding most likely involves regions of the receptor in a closer spatial proximity to attachment points used by classic antagonists than does the mechanism for salmeterol at the β2-adrenoceptor. A possible scheme for xanomeline binding to the M1mAChR is depicted in Fig. 12. Salient features of this model include the ability of a large fraction of added xanomeline to interact in a readily reversible manner with the classic agonist binding site, while a minor fraction is able to interact in an orientation that allows persistent binding, yet does not preclude access to the classic site. Competitive antagonists would be expected to interfere predominantly with binding at the classic site but may still sterically hinder interaction at the accessory site. Alternatively, the second binding site may represent a point of attachment only, with the activating effect achievable only on agonist dissociation from the exosite to rebind at the classic site. A third possibility is that the interaction of xanomeline may involve a two-step binding process, in which the agonist first associates with the classic binding site before forming a tighter association with an accessory site. Each of these suggested schemes accommodates the observed effects of atropine. An additional scheme that takes into account the possibility of multiple receptor conformational states (Kenakin, 1997) is that the persistent activating effects of xanomeline are manifested only when the receptor is not “held” in an inactive conformation by antagonists. By proposing such models, we by no means reject the possibility of alternative explanations, nor do we suggest that the orientations of xanomeline depicted in Fig. 12 represent those achieved in the interaction of this agent with the receptor. Obviously, further detailed investigation is warranted. One approach may entail the use of receptor solubilization because the removal of the receptor from its native lipid environment may shed some light on the contribution of the latter to the mechanism of action of xanomeline.
Regardless of the molecular nature of the interaction of xanomeline with the M1 mAChR, however, the results reported in the current study point to a novel mode of receptor activation that outlasts the presence of free agonist in the receptor compartment. These findings might have significant clinical implications, especially in the treatment of Alzheimer’s disease, because the therapeutic and/or side effects of xanomeline may not necessarily be predicted by the measurement of free drug concentration in the cerebrospinal fluid.
Footnotes
-
Send reprint requests to: Prof. Esam E. El-Fakahany, Division of Neuroscience Research in Psychiatry, Box 392, Mayo Memorial, University of Minnesota Medical School, Minneapolis, MN 55455. E-mail: elfak001{at}maroon.tc.umn.edu
-
This work was supported by National Institutes of Health Grant NS25743.
- Abbreviations:
- mAChR
- muscarinic acetylcholine receptor
- CHO
- Chinese hamster ovary
- CCh
- carbachol
- NMS
- N-methylscopolamine
- HEPES
- 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
- PI
- phosphoinositide
- nNOS
- neuronal nitric oxide synthase
- DMEM
- Dulbecco’s modified Eagle’s medium
- dpm
- disintegrations per minute
- ANOVA
- analysis of variance
- Received November 14, 1997.
- Accepted February 23, 1998.
- The American Society for Pharmacology and Experimental Therapeutics