Here, we have investigated the in vitro pharmacology of a muscarinic agonist, (3R,4R)-3-(3-hexylsulfanyl-pyrazin-2-yloxy)-1-aza-bicyclo[2.2.1]heptane (WAY-132983), and we demonstrated its activity in several models of pain. WAY-132983 had a similar affinity for the five muscarinic receptors (9.4–29.0 nM); however, in calcium mobilization studies it demonstrated moderate selectivity for M1 (IC50 = 6.6 nM; Emax = 65% of 10 μM carbachol-stimulation) over the M3 (IC50 = 23 nM; Emax = 41%) and M5 receptors (IC50 = 300 nM; Emax = 18%). WAY-132983 also activated the M4 receptor, fully inhibiting forskolin-induced increase in cAMP levels (IC50 = 10.5 nM); at the M2 receptor its potency was reduced by 5-fold (IC50 = 49.8 nM). In vivo, WAY-132983 demonstrated good systemic bioavailability and high brain penetration (>20-fold over plasma levels). In addition, WAY-1329823 produced potent and efficacious antihyperalgesic and antiallodynic effects in rodent models of chemical irritant, chronic inflammatory, neuropathic, and incisional pain. It is noteworthy that efficacy in these models was observed at doses that did not produce analgesia or ataxia. Furthermore, a series of antagonist studies demonstrated that the in vivo activity of WAY-132983 is mediated through activation of muscarinic receptors primarily through the M4 receptor. The data presented herein suggest that muscarinic agonists, such as WAY-132983, may have a broad therapeutic efficacy for the treatment of pain.
Acetylcholine is a major neurotransmitter involved in many physiological processes. Its effects are mediated in part through the five muscarinic acetylcholine receptors (mAChRs; M1–M5), which are expressed throughout the CNS and elsewhere (Caulfield and Birdsall, 1998). The M1,M3, and M5 receptors are coupled to Gq, and they activate phospholipase C, leading to the formation of inositol phosphates, which cause an increase in intracellular calcium. The M2 and M4 receptors couple through the inhibitory G proteins, Gi/o, and they reduce levels of cAMP through inhibition of adenylate cyclases.
The tissue distribution of the mAChRs suggests a potential role for these receptors in pain signaling. For example, mRNA for M2,M3, and M4 have been identified in the spinal cord and dorsal root ganglion of rat (Wei et al., 1994; Tata et al., 2000). Furthermore, mAChRs are also expressed in central pain processing areas such as the thalamus, periaqueductal gray, and rostral ventrolateral medulla. A major site of action for cholinomimetics in nociceptive processing is the spinal cord, where they mimic the release of acetylcholine (ACh) from the spinal cholinergic nerves. Noxious pain stimuli are known to increase ACh in the cerebrospinal fluid of anesthetized sheep; similarly, increased ACh levels are seen during the postoperative period, i.e., during pain states (Bouaziz et al., 1995; Eisenach et al., 1996). The increase in withdrawal latency to a noxious heat stimulus (Li et al., 2002) and the antiallodynic action of intrathecal ACh or neostigmine are inhibited by the GABAB receptor antagonist CGP55845 (Li et al., 2002; Chen and Pan, 2003). This is consistent with a muscarinic-receptor evoked release of GABA, which is then able to decrease glutamate release via presynaptic GABAB receptors (Baba et al., 1998). Furthermore, ACh can also inhibit glutamate release directly through presynaptic M2 receptors (Li et al., 2002).
Due to the high degree of homology between the mAChRs, especially in the ligand binding pocket, there is a paucity of selective ligands. Therefore, the identity of the mAChR involved in ACh-mediated antinociception is unclear. For example, the M1/M3 subtypes have been implicated (Naguib and Yaksh, 1997; Honda et al., 2000), as have M1 and/or M2 (Iwamoto and Marion, 1993). In contrast, Sheardown et al. (1997) argue that the M1 receptor is not necessary for antinociception consistent with a lack of M1 expression in the spinal cord (Wei et al., 1994; Höglund and Baghdoyan, 1997).
More recently, receptor knockout mice have provided insight into the roles of the individual mAChR subtypes in antinociception. Oxotremorine produced dose-dependent analgesic effects in both the hot-plate and tail-flick assays in wild type mice. However, the potency of the effects of oxotremorine was markedly reduced by deletion of the M2 receptor, although a maximal effect was still observed, presumably mediated via the M4 receptor, because oxotremorine is ineffective in the M2/M4 double knockout animal (Gomeza et al., 1999; Duttaroy et al., 2002). This was the first definitive evidence for the specific roles of these two receptor subtypes in muscarinic receptor-mediated analgesia.
In our article, we fully describe the in vitro and pharmacokinetic evaluation of WAY-132983, a muscarinic agonist that has been shown to have positive effects in animal models of cognition (Sabb et al., 1999; Bartolomeo et al., 2000). WAY-132983 demonstrates linear pharmacokinetic characteristics and substantial CNS penetration; therefore, it is an ideal pharmacological tool for investigating the in vivo roles of mAChRs. We demonstrate for the first time the antinociceptive effects of the muscarinic agonist WAY-132983 in models of chemical-induced, neuropathic, inflammatory, and postsurgical pain. Antagonist studies were also performed in pain models to verify the selectivity of WAY-132983. In so doing, we have further characterized the role of muscarinic receptors in pain, and our novel demonstration of activity in models of neuropathic and visceral pain extends the therapeutic potential of muscarinic agonists for the treatment of pain.
Materials and Methods
CHO-K1 cells stably expressing hM1–5 were cultured in Dulbecco's modified Eagle's medium (Gibco, Carlsbad, CA), containing 10% fetal bovine serum (Sigma-Aldrich, St. Louis, MO), 100 units of Pen/Strep (Gibco), and 500 mg/ml Geneticin (G418; Gibco).
Cellular Membrane Preparation
Confluent 245-cm2 dishes of cells were washed two times with ice-cold phosphate-buffered saline. Cells were scraped in 10 ml of ice-cold buffer (20 mM HEPES, pH 7.5, and 10 mM EDTA), homogenized in a Dounce homogenizer, and pelleted at 32,000g. Cell pellets were resuspended in buffer (20 mM HEPES, pH 7.5, 100 mM NaCl, and 1 mM MgCl), homogenized, aliquoted, and frozen at –80°C. Protein concentrations were determined using Bio-Rad Protein Assay (Bio-Rad, Hercules, CA) reagents as per manufacturer's instructions.
Cellular Radioligand Binding
Test compounds and radioligand [3H]N-methylscopolamine (NMS; PerkinElmer Life and Analytical Sciences, Boston MA) (final 0.32 nM for competition and 1 pM–300 nM for saturation) were diluted to 100× concentrations in assay buffer (20 mM HEPES, 100 mM NaCl, and 1 mM MgCl2, pH 7.5) before being added to the reaction tubes (1:100). hM1–5 CHO-K1 frozen membrane homogenates were suspended in assay buffer, and then they were added to the reaction tubes (26 μg of protein/reaction tube) with a final assay volume of 500 μl/tube. Nonspecific binding was determined by a saturating concentration of atropine (10 μM; Sigma-Aldrich). The reactions were vortexed, and then they were incubated while shaking for 60 min at 30°C. The binding reaction was terminated by vacuum filtration (Brandel Inc., Gaithersburg, MD) through Whatman glass fiber filters (GF/B paper; Whatman, Brentford, UK) presoaked in 0.15% polyethylenimine/H2O (Sigma-Aldrich) followed by four washes with ice-cold wash buffer (50 mM Tris-HCl and 0.9% NaCl, pH 7.4, at 4°C). Filter circles were removed to 7-ml glass vials, mixed with 5 ml of scintillation cocktail (OptiFluor; PerkinElmer Life and Analytical Sciences), and incubated at room temperature for 8 h. The vials were counted, and dpms were determined using a liquid scintillation counter (PerkinElmer Life and Analytical Sciences).
Tissue Membrane Preparation
Rat cortical membranes were purchased from Analytical Biological Services (Wilmington, DE). They were prepared from Sprague-Dawley male rat cortex homogenates that were pelleted at 20,000 rpm (SS-34 rotor; Sorvall, Newton, CT) for 15 min at 4°C. They were resuspended and pelleted two more times before being resuspended in 50 mM Tris (3 mg/ml), pH 7.4, and stored at –80°C.
Rat cardiac membranes were prepared by mincing Sprague-Dawley rat hearts in homogenizing buffer (50 mM Tris and 250 mM sucrose, pH 7.4), and then homogenized with a Polytron (2 × 10sat setting 8; Kinematica, Basel, Switzerland) and centrifuged at 1500 rpm (SS-34 rotor) for 10 min at 10°C. The supernatant was then spun at 18,500 rpm (SS-34 rotor) for 30 min at 4°C. The pellets were then resuspended in binding buffer (50 mM Tris and 2 mM MgCl2,pH 7.5), aliquoted (1.2 mg/ml), and stored at –80°C.
Tissue Radioligand Binding
Test compounds and radioligand [3H]NMS (PerkinElmer Life and Analytical Sciences) (final 1.4 nM for competition and 100 pM–100 nM for saturation) were diluted to 10× concentrations in assay buffer (cortical: 20 mM HEPES, 100 mM NaCl, and 1 mM MgCl2,pH 7.5; cardiac: 50 mM Tris and 2 mM MgCl2, pH 7.5) before being added to the reaction tubes (1:10). Frozen membrane homogenates were suspended in assay buffer, and then they were added to the 96-well flat bottom plates (30 μg of protein/well, Corning 3912; Corning Glassworks, Corning, NY) with a final assay volume of 150 μl/well. Nonspecific binding was determined by a saturating concentration of atropine (10 μM; Sigma-Aldrich). The reactions were vortexed and incubated for 60 min at 37°C. The binding reaction was terminated by centrifugation of the pates at 10,000 rpm (SH3000 rotor; Sorvall) for 10 min at 4°C followed by aspiration of the supernatant. The wells were washed with 250 μl of cold 50 mM Tris and repelleted. The pellets were solubilized with 25 μl of 0.25 M NaOH and shaken for 10 min. Scintillation cocktail was added (100 μlof Microscint 20; PerkinElmer Life and Analytical Sciences), and then the pellets were shaken for another 10 min. The plates were counted, and dpms were determined using a TopCount liquid scintillation counter (PerkinElmer Life and Analytical Sciences).
cAMP Inhibition Assays
CHO-K1 cells stably expressing hM4 or hM2 were cultured in T-175 flasks, as described above, and then cells were harvested by washing two times with phosphate-buffered saline, followed by addition of 5 ml of cell dissociation solution (Mediatech, Herndon, VA). After 3- to 5-min room temperature incubation, cells were removed, mixed with 10 ml of Krebs' assay buffer (118 mM NaCl, 5 mM KCl, 1.2 mM KH2PO4, 25 mM NaHCO3, 11.1 mM glucose, 1.2 mM MgSO4, and 2.4 mMCaCl2), and pelleted. Cell pellets were suspended in Krebs' buffer and counted. Compounds were serially diluted in Krebs' buffer containing 30 mM forskolin. Compounds were combined with 1.5 × 104 cells in 96-well plates (Corning 3912; Corning Glassworks) and incubated at 37°C for 30 min. cAMP levels were determined using the HitHunter kit (DiscoveRx, Fremont, CA) as per the manufacturer's instructions. Chemiluminescence was counted on a Wallac Victor V (PerkinElmer Life and Analytical Sciences) after a 3-h incubation.
Calcium Mobilization Assays
CHO-K1 cells stably expressing hM1,hM3,orhM5 were plated at a density of 60,000 cells/well in 96-well plates (Corning 3603; Corning Glassworks) 18 to 24 h before the assay. One hour before the assay, the media were removed and replaced with Ca2+ 3 Assay Dye (Molecular Devices, Sunnyvale, CA) dissolved in assay buffer [Hanks' balanced salt solution (Gibco) with CaCl2 and MgSO4 without phenol red), 20 mM HEPES (Gibco), and 0.25 mM probenecid (Sigma-Aldrich)]. At the time of the assay, drugs were added, and fluorescence was measured with a FLEX station (Molecular Devices).
In Vivo Pharmacokinetic Studies and CNS Permeability of WAY-132983
Male Sprague-Dawley rats (200–350 g) precannulated with jugular and femoral veins cannulas were purchased from Charles River (Raleigh, NC). They were fasted overnight before dosing, and they were fed 4 h postdose while water was ad libitum. Three groups of rats (n = 3) were dosed with WAY-132983 at 0.1, 0.3, or 1.0 mg/kg i.p. in 0.5% methyl cellulose plus 2% Tween 80 in distilled water (Sigma-Aldrich). A separate group of rats (n = 3) was dosed by bolus intravenous injection at 1 mg/kg in dimethyl sulfoxide/80% polyethylene glycol. All dosage formulations were sonicated for 15 min before administration. Blood was collected predose and at predetermined times over a 24-h period. We drew 400 μl of blood at each time point, and animals were given 1 ml of donor blood at 2- and 8-h time points to replace the sampled blood and to prevent any physiological effects associated with volume depletion. The collection tubes contained EDTA as an anticoagulant. Plasma was harvested by centrifugation at 14,000 rpm for 10 min at 4°C, and it was stored at –80°C before analysis.
To study the CNS permeability of the compound, rats were administered a single dose of 0.1, 0.3, or 1.0 mg/kg i.p. of WAY-132983 in 0.5% methyl cellulose plus 2% Tween 80 in distilled water. Blood and brain samples were collected at 0.5, 1, 3, and 5 h (n = 3 per time point for a total of 12 samples/dose) after drug administration. Plasma was harvested from blood as described above. The brain samples were weighed, diluted 5-fold with water, and homogenized by Polytron.
WAY-132983 was extracted from the plasma or brain samples (50-μl aliquots each) by protein precipitation with acetonitrile (400 μl) and by vortexing for 5 min. The mixture was centrifuged at 3400 rpm for 10 min at 10°C. An aliquot of the supernatant (20 μl) was analyzed by liquid chromatography/tandem mass spectrometry consisting of a Hewlett Packard LC system (Hewlett Packard, Palo Alto, CA) coupled with positive electrospray tandem mass spectrometry (Sciex API 4000; PerkinElmerSciex Instruments, Boston, MA). Separation was achieved on a 30 × 2-mm Luna C18 column (Phenomenex, Torrance, CA) by elution with mobile phase consisting of 0.1% formic acid in water (A) and 0.1% formic acid in acetonitrile (B). The mobile phase was pumped at a flow rate of 1 ml/min, and it used a 2-min linear gradient (between 1 and 3 min) from 95% A, 5% B to 5% A, 95% B, out of a total run time of 5 min. Multiple reaction monitoring mode was used to monitor the compound with m/z transition of 308.36 to 279.1. The limits of quantification were 0.5 and 0.1 ng/ml for plasma and brain, respectively. The measured brain concentrations (nanograms per milliliter) were converted to nanograms per gram of original tissues using the total volume of the homogenate and the original brain weights. WAY-132983 concentrations in plasma and brain samples were determined by interpolation from standard curves prepared by spiking respective matrices with known concentrations of the compound and by analyzing them as described above for the samples.
In Vivo Studies
Compounds. WAY-132983 was obtained from Dr. Annmarie Sabb (Wyeth Research, Princeton, NJ), and gabapentin and celecoxib were purchased from Toronto Research Chemicals (Toronto, ON, Canada). Morphine and scopolamine were purchased from Sigma-Aldrich. MT-3 was purchased from Peptides International Inc. (Louisville, KY). For all in vivo testing, doses were calculated as the free base molecular weight of the compound. WAY-132983, gabapentin, and celecoxib were suspended in 2% Tween 80/0.5% methylcellulose in sterile water. Morphine and MT-3 were dissolved in sterile water, and scopolamine was dissolved in sterile saline. WAY-132983, gabapentin, and morphine (postincisional) were administered i.p. for all assays, with the exception of the visceral pain model in which WAY-132983 was administered s.c. This is to avoid issues arising with i.p. dosing of PPQ. Morphine and scopolamine were administered s.c., celecoxib was administered p.o., and MT-3 was administered i.t.
Subjects. The Wyeth Institutional Animal Care and Use Committee approved all animal procedures according to the guidelines of the Office of Laboratory Animal Welfare. Male CD-1 mice (20–25 g; Charles River; Kingston/Stoneridge, NY) were used for the visceral pain study. Male Sprague-Dawley rats (125–150 g; Harlan, Indianapolis, IN) were used for the acute, inflammatory, neuropathic, postincisional, and ataxia studies. Male Sprague-Dawley rats (200–250 g; Charles River) were used for the prostaglandin (PGE2) study. For all studies animals were maintained in climate-controlled room on a 12-h light/dark cycle with food and water available ad libitum. All assays were performed in a randomized manner by individuals blinded to the experimental condition.
Acute Pain Models. The effect of WAY-132983 on acute analgesia was investigated using the tail-flick and hot-plate assays as described previously (Valenzano et al., 2005). For the tail-flick assay, rats were placed on the apparatus (Ugo Basile, Comerio, Italy), and an infrared beam was focused onto the tail, 5 cm from the tip. The latency to tail-flick was assessed, and the cut-off was set at 30 s and the intensity was set to 35%. For the hot-plate assay, rats were placed on a metal plate maintained at 52°C (Ugo Basile). The latency to exhibit a nocifensive response, defined as hind paw lift, flutter, licking, or escape behavior, was measured, and the cut-off was set at 30 s to avoid tissue damage. Dosing and testing regimes were the same for both assays. Baseline latencies were determined, and 1 h later, the rats received a single dose of 0.1, 0.3, or 1.0 mg/kg WAY-132983, 10 mg/kg morphine (the positive control), or vehicle. Latency to exhibit a nocifensive response was again determined at 1, 3, and 5 h postdrug administration. Latency was determined once for each animal at each time point (n = 10/group).
Chemical-Induced Models. The ability of WAY-132983 to attenuate acute visceral (abdominal) pain was assessed after an i.p. injection of 2 mg/kg para-phenylquinone (PPQ; dissolved in 4% ethanol in distilled water; Sigma-Aldrich) (Siegmund et al., 1957). WAY-132983 at 0.1, 0.3, or 1.0 mg/kg or vehicle (n = 10/group) was pretreated 60 min before PPQ administration. During testing, after PPQ administration, mice were individually placed in a Plexiglas cage, and the total number of abdominal constrictions was recorded for 1-min periods, starting at 5 and 10 min after PPQ injection.
PGE2-induced thermal hyperalgesia was assessed using a warmwater tail withdrawal assay. For all test sessions, thermal thresholds were assessed by submerging the terminal 10 cm of the tail into water warmed to 38, 42, 46, 50, or 54°C. The latency in seconds (maximum 20-s cut-off) for the rat to remove the tail from the water was recorded. After the assessment of baseline sensitivity, thermal hyperalgesia was produced by injection of 0.1 mg of PGE2 (50 μl) into the distal 1 cm of the tail. Animals were tested 30 min after injection of PGE2. For compound testing, 0.1, 0.3, or 1.0 mg/kg WAY-132983 (n = 8/group) was administered 30 min before injection of PGE2 (i.e., drug effects were evaluated 60 min postdrug). For each test session a temperature effect curve was generated. The temperature that produced half-maximal (10-s) tail withdrawal latency (T10) was calculated from each temperature-effect curve. The T10 was determined by interpolation from a line drawn between the point above and the point below 10 s on the temperature-effect curve. Results are presented as the percentage of blockade of PGE2-induced T10 thermal hypersensitivity calculated as follows:
Inflammatory Model. The ability of WAY-132983 to reverse hyperalgesia associated with inflammation was investigated using the Freund's complete adjuvant (FCA) model. For this assay, hind paw-withdrawal thresholds (PWTs) to a noxious mechanical stimulus were determined using an analgesimeter (model 7200; Ugo Basile). Cut-off was set at 250 g, and the endpoint was taken as complete paw-withdrawal. PWT was determined once for each rat at each time point. Baseline PWT was determined, and the rats were anesthetized with isofluorane (2% in oxygen), and they received an intraplantar injection of 50% FCA (50 μl, diluted in saline) to the left hind paw. Twenty-four hours after FCA injection, predrug PWTs were measured, and the rats received a single dose of 0.1, 0.3, or 1.0 mg/kg WAY-132983, 30 mg/kg celecoxib (the positive control), or vehicle (n = 10/group). PWT was again determined 1, 3, and 5 h postdrug administration. Percentage of reversal of hyperalgesia for each rat was calculated according to the following equation:
Neuropathic Model. Surgical procedures were performed under 4% isoflurane/O2 anesthesia, delivered via nose cone, and anesthesia was maintained at 2.5% for the duration of the surgery. After induction of anesthesia, the incision site was shaved and prepared in a sterile manner. Spinal nerve ligation (SNL) surgery was performed as described previously (Valenzano et al., 2005) with the exception that nerve injury was produced by tight ligation of the left L5 spinal nerve only (LaBuda and Little, 2005). Tactile thresholds were assessed using a series of calibrated von Frey monofilaments (Stoelting, Wood Dale, IL). Assessment of tactile allodynia was measured as the hind paw-withdrawal threshold that produced a 50% likelihood of a withdrawal using the up-down method, as described previously (Valenzano et al., 2005). Thresholds were evaluated before surgery, and they were reassessed 3 to 4 weeks after SNL surgery. On test day, rats were administered 0.1, 0.3, or 1.0 mg/kg WAY-132983, 100 mg/kg gabapentin (the positive control), or vehicle, and tactile thresholds were assessed 1, 3, and 5 h after administration (n = 8/group). Percentage of reversal of allodynia for each rat was calculated according to the following equation:
Incisional Model of Postsurgical Pain. Postsurgical pain was induced using an incisional pain model, as described previously (Valenzano et al., 2005). In brief, a 1-cm longitudinal incision was made through the skin and fascia of the left plantar aspect of the foot. Tactile thresholds were assessed using a series of calibrated von Frey monofilaments (Stoelting). Assessment of tactile allodynia was measured as the hind paw-withdrawal threshold that produced a 50% likelihood of a withdrawal using the up-down method. Thresholds were evaluated before surgery, and they were reassessed 24 h after surgery. On test day, rats were administered 0.1, 0.3, or 1.0 mg/kg WAY-132983, 5.6 mg/kg morphine (the positive control), or vehicle, and tactile thresholds were assessed 0.5, 1.7, 3, and 5 h after administration (n = 8/group). Percentage of reversal of allodynia for each rat was calculated according to the following equation:
Ataxia Model. To examine the potential effects of WAY-132983 on motor performance, rats were tested on an accelerating rotarod (Ugo Basile). Rats were initially trained to walk on the rotarod at 16 rpm for 120 s. Rats that did not meet criteria after five attempts were excluded from further testing. One hour after training, 0.1, 0.3, or 1.0 mg/kg WAY-132983 (n = 6/group) was administered, and 60 min later, rats were placed on the rotarod (accelerating from 4 to 40 rpm over the course of 5 min). The latency to fall off the rotarod was recorded, and the value used as an indication of motor coordination.
Antagonist Studies. The nonselective muscarinic antagonist scopolamine was evaluated in combination with WAY-132983 in the PPQ and FCA model to determine whether a nonselective muscarinic receptor antagonist could block the activity of WAY-132983. In both assays, 1 mg/kg scopolamine was administered 15 min before 1 mg/kg WAY-132983, and behavioral testing was conducted 1 h following WAY-132983 administration as described previously. In addition, the selective M4 toxin MT-3 was evaluated in the FCA model. WAY-132983 (1 mg/kg) was administered 1 h before testing, and 3 μg of MT-3 was administered 15 min before behavioral testing.
Analysis of Results.Ki,EC50, and IC50 values were determined using GraphPad Prism (GraphPad Software Inc., San Diego, CA). The concentration-time data were used to estimate the pharmacokinetic parameters by noncompartmental approaches using WinNonlin Professional 4.1 (Pharsight, Mountain View, CA).
Statistical significance was determined on untransformed data using a one-way (PPQ, PGE2, and rotarod) or a repeated measures analysis of variance (acute pain, FCA, SNL, and postincisional) using a customized SAS-Excel application (SAS Institute, Cary, NC). Significant main effects were analyzed further by subsequent least significant difference analysis. The level of significance was set at p < 0.05. Data are shown as mean ± S.E.M.
WAY-132983 Binds with Equal Affinity to Each of the Muscarinic Receptors. The in vitro characterization of WAY-132983 was assessed in CHO-K1 cells stably expressing one of the human muscarinic receptors (M1–M5). All binding studies were conducted with the tritiated nonselective muscarinic antagonist [3H]NMS.
Saturation binding analysis on CHO-K1 membranes revealed that although each receptor is expressed at different levels (301–1116 fmol/mg protein), their affinities are relatively the same for [3H]NMS (0.24–0.74 nM; Table 1). The affinity of WAY-132983 was determined by competition binding with 0.32 nM [3H]NMS, a concentration that approximates the Kd of NMS for the receptors. Competition binding demonstrates that WAY-132983 has equal affinity for all of the muscarinic receptors (Table 2).
Competition binding results obtained with a rat cortical membrane preparation, which expresses primarily the M1 receptor (481 fmol/mg tissue), demonstrated WAY-132983 to have a similar affinity to the cloned M1 (Ki = 14.7 nM). However, an affinity of 175.8 nM was obtained in binding studies conducted on a rat cardiac membrane preparation, which primarily expresses the M2 receptor although the level of receptor expression of receptor was similar (234 fmol/mg tissue) (Table 2). WAY-132983 was also evaluated for its activity at a large selection of receptors, ion channels, enzymes, and transporters (Table 3 for a list; NovaScreen, Hanover, MD). WAY-132983 was inactive (<35% inhibition of binding) in all assays except cloned M1 and a muscarinic nonselective central assay.
WAY-132983 Is an M1/M4-Preferring Agonist. M1,M3, and M5 are Gq-linked seven transmembrane receptors; therefore, a calcium mobilization study was used to assess the efficacy of WAY-132983 at these receptors. WAY-132983 preferentially activates the M1 receptor (Fig. 1A), producing partial agonist response of 65% of the maximal response of carbachol. However, WAY-132983 exhibited lower efficacy at the M3 and M5 receptors (41 and 18%, respectively). WAY-132983 also exhibits greater potency at the M1 receptor than the M3 or M5 (EC50 values: M1 = 6.6 ± 0.8 nM; M3 = 23 ± 9.4 nM; and M5 = 300 ± 160 nM).
Effects of WAY-132983 at the M2 and M4 were assessed using a cAMP assay. In cells expressing the human M4 receptor, WAY-132983 stimulation resulted in complete inhibition of the forskolin response at the M4 (IC50 = 10.49 ± 1.23 nM), whereas its effects at the M2 were less potent (IC50 = 49.8 ± 7.2 nM).
WAY-132983 Has Good Systemic Bioavailability Following Intraperitoneal Administration. The pharmacokinetics of WAY-132983 was characterized in rats after i.v., p.o., and i.p. administrations. Samples were collected over 24 h, but the compound could only be quantified up to 8 h. The compound was essentially unquantifiable after oral administration of 0.1 mg/kg. After i.v. administration, the pharmacokinetic profile (Fig. 2) was characterized by a very high clearance, (190 ± 12.3 ml/min/kg), high volume of distribution (22.8 ± 1.7 l/kg), and terminal plasma half-life of 2.1 ± 0.2 h (all values are mean ± S.E.M.) (Table 4).
After p.o. administration, plasma concentrations of WAY-132983 were detectable after 0.3 and 1 mg/kg oral dosing. Both dose groups demonstrated rapid absorption with maximal concentrations reached within 2 h after administration (Table 5). The overall exposure was fairly low with consequently low oral bioavailability as well (6.9 and 11%, respectively). There was insufficient data to characterize the terminal phase and to calculate a half-life.
After i.p. administration, plasma concentrations of WAY-132983 were detectable following all three i.p. dosings. The profiles after the three doses, 0.1, 0.3, and 1 mg/kg, were comparable, and they were consistent with rapid absorption with maximum concentration achieved within 1 h after administration. The terminal half-life was comparable with that after i.v. administration. The exposure was good, resulting in good bioavailability that ranges between 84 and ∼100%.
WAY-132983 Has High Penetration into the CNS. The plasma and brain exposures of WAY-132983 were determined over a 5-h sampling period following single i.p. doses of 0.1, 0.3, and 1 mg/kg (Fig. 3). At all doses examined, WAY-132983 was detectable in both the plasma and brain samples with similar concentration-time profiles. At all doses, the concentration of WAY-132983 in the brain was higher than the corresponding levels in the plasma, and this resulted in high brain-to-plasma exposure ratios that ranges between 24 and 33 (Table 5).
WAY-132983 Does Not Affect Acute Nociception. WAY-132983 was evaluated for its effect on acute nociception in both a tail-flick and hot-plate assay. WAY-132983 did not affect latency in the tail-flick or hot-plate assays at 1, 3, or 5 h following administration of doses up to 1.0 mg/kg, whereas morphine (10 mg/kg s.c.) produced a significant increase in latency to tail-flick at 1 and 3 h postadministration, and it exhibited a nocifensive response in the hot-plate assay at 1, 3, and 5 h postadministration (p < 0.05) (data not shown).
WAY-132983 Blocks Visceral Pain and Thermal Hyperalgesia Associated with Chemical Irritants. WAY-132983 (0.3 and 1.0 mg/kg) produced a statistically significant blockade of PPQ-induced abdominal constrictions (F3,36 = 250.12; p < 0.05) (Fig. 4A). The maximum percentage of blockade (95%) was achieved following the 1.0 mg/kg dose.
Injection of PGE2 into the tail resulted in thermal hyperalgesia as assessed using a warm-water tail withdrawal assay (Fig. 4B). WAY-132983 (0.1, 0.3, and 1.0 mg/kg) produced a statistically significant blockade of PGE2-induced thermal hyperalgesia (F2,21 = 19.87; p < 0.05). The maximum percentage of reversal (86%) was achieved following the 1.0 mg/kg dose.
WAY-132983 Reverses Mechanical Hyperalgesia Associated with Inflammation. Intraplantar injection of 50 μl of FCA into the hind paw resulted in the development of mechanical hyperalgesia as indicated by a decreased paw-withdrawal threshold to a noxious mechanical stimulus (Fig. 5). WAY-132983 (0.3 and 1.0 mg/kg) produced statistically significant antihyperalgesia 1 and 3 h postadministration (F20,220 = 12.07; p < 0.05). The maximum percentage of reversal (56%) was achieved 1 h after the 1.0 mg/kg dose. Celecoxib (30 mg/kg p.o.) also produced a statistically significant reversal of hyperalgesia 1, 3, and 5 h postadministration (Fig. 5).
WAY-132983 Reverses Tactile Allodynia Associated with Nerve Injury. Spinal nerve ligation resulted in the development of tactile allodynia as indicated by a decreased response threshold to a non-noxious tactile stimulus (von Frey fiber), 3 to 4 weeks postsurgery (Fig. 6). No differences in responses were seen between 3 and 4 weeks postsurgery. WAY-132983 (0.1, 0.3, and 1.0 mg/kg) produced statistically significant antiallodynia 1 h postadministration (F16,140 = 1.62; p < 0.05). The maximum percentage of reversal (84%) was achieved 1 h after the 1.0 mg/kg dose. Gabapentin (100 mg/kg i.p.) also produced a statistically significant reversal of hyperalgesia 1 h postadministration (Fig. 6).
WAY-132983 Reverses Tactile Allodynia Associated with Surgical Incision. Incision of the plantar surface of the hind paw resulted in the development of tactile allodynia as indicated by a decreased response threshold to a non-noxious tactile stimulus (von Frey withdrawal threshold) 24 h postincision (Fig. 7). WAY-132983 (0.3 and 1.0 mg/kg) produced statistically significant antiallodynia 1 h postadministration (F20,155 = 4.73; p < 0.05). The maximum percentage of reversal (72%) was achieved 1 h after the 1.0 mg/kg dose. Morphine (5.6 mg/kg i.p.) also produced a statistically significant reversal of hyperalgesia 1 h postadministration (Fig. 7).
WAY-132983 Exerts Its Activity in Vivo through Muscarinic Receptors. WAY-132983 (1.0 mg/kg) significantly blocked PPQ-induced abdominal constrictions (F4,49 = 30.81; p < 0.05; 87% blockade of pain response) (Fig. 8A), and its activity was significantly decreased by pretreatment with the nonselective muscarinic antagonist scopolamine at 1 mg/kg (p < 0.0547% blockade of pain response). Scopolamine (1 mg/kg) treatment alone did not significantly block PPQ-induced abdominal constrictions.
Similar effects were seen in the FCA model. WAY-132983 (1.0 mg/kg) significantly reversed FCA-induced mechanical hyperalgesia [F4,20 = 15.5; p < 0.05; 78% reversal of pain response (Fig. 8B)], and pretreatment with the nonselective muscarinic antagonist scopolamine at 1 mg/kg significantly blocked this effect (p < 0.05; 17% reversal of pain response). Scopolamine (1 mg/kg) treatment alone did not significantly reverse FCA-induced mechanical hyperalgesia.
WAY-132983 Exerts Its Activity in Vivo in Part through the M4 Receptor. WAY-132983 (1.0 mg/kg) significantly reversed FCA-induced mechanical hyperalgesia (F4,20 = 24.28; p < 0.05; 72% reversal of pain response) (Fig. 8C), and this effect was significantly blocked by i.t. pretreatment with the M4-selective toxin MT-3 at 3 μg(p < 0.05; 20% reversal of pain response). MT-3 (3 μg) treatment alone was without effect.
WAY-132983 Does Not Cause Ataxia. Ataxia is a common centrally mediated side effect that can confound the interpretation of behavioral assays. Rats were tested for motor function using the rotarod assay. WAY-132983 did not affect rotarod performance at 1 h postadministration of doses up to 1.0 mg/kg (p > 0.05) (Fig. 9).
In vitro studies indicate that WAY-132983 is highly selective for the muscarinic receptor family as evidenced by a broad panel of selectivity assays. Functional studies indicate WAY-132983 preferentially activates the M1 and M4 receptors over the other muscarinic receptors. Although WAY-132983 is a partial agonist at the M1 receptor (65% maximal response), it has both a greater efficacy and potency at this subtype than the M3 and M5. We also present data demonstrating that WAY-132983 has a limited degree of selectivity for the M4 receptor over the M2 receptor. Accordingly, Spalding et al. (2002) demonstrated by using a β-galactosidase reporter gene (R-SAT) assay that WAY-132983 is an M1- and M4-preferring agonist despite having activity at the other mAChRs, albeit with lower potency and efficacy. Interestingly, in cell lines stably expressing human muscarinic receptors, WAY-132983 has equal affinity for all five receptors. A similar affinity was determined using rat cortical membranes (primarily expressing M1). However, in a rat cardiac membrane preparation (primarily M2), a somewhat lower affinity was determined, suggesting that WAY-132983 may be more selective over M2 in rats.
Identifying the physiological role(s) of mAChRs in vivo has been hampered by limited number of agonists with suitable pharmacokinetic properties. Systemic exposure of WAY-132983 in rats can be achieved after low dose i.p. administration, which provides greater (9-fold) exposure than p.o. administration. It is noteworthy that even at extremely low circulating plasma levels, substantial quantities of WAY-132983 were found in brain (brain/plasma ratio >20-fold). The high CNS levels of WAY-132983 with low circulating levels may be beneficial in avoiding the side effects observed with other muscarinic agents. WAY-132983 did cause an increase in salivation (≥10 mg/kg) and hypothermia (≥3 mg/kg), but it did not produce chromodacryorrhea, lacrimation, or tremors (Bartolomeo et al., 2000).
There are several possible reasons why the levels are found to be higher in the brain than the plasma. First, the physical-chemical properties of WAY-132983 may enhance CNS exposure. In fact, WAY-132983 is reasonably lipophilic (cLogP = 4.3), and its molecular size and polar characteristics would allow it to easily cross lipid barriers such as the blood-brain barrier (Sabb et al., 1999). Second, WAY-132983 could be a substrate for an uptake but not for an efflux transporter in the brain, concentrating the compound in the brain relative to the plasma, although there is no specific evidence of this at this time. Third, WAY-132983 may bind to specific proteins or lipids specific in brain, thereby sequestering WAY-132983 within the CNS. Regardless of the mechanism responsible for the substantial CNS levels, WAY-132983 offers a useful pharmacokinetic and pharmacological profile, demonstrating good systemic availability, tissue distribution, and high CNS penetration. This property is important to further understand the role of mAChRs in vivo.
In this study, we demonstrate the in vivo efficacy of WAY-132983 in models of chemical-induced, neuropathic, inflammatory, and postsurgical pain. Although WAY-132983 was not active in models of acute pain, several groups have described muscarinic agonist activity in these models. The reason for this discrepancy is not apparent but it may reflect the lower affinity and potency of WAY-132983 for M2, which accounts for approximately 90% of [3H]N-methylscopolamine binding in the spinal cord (Wei et al., 1994; Duttaroy et al., 2002).
WAY-132983 demonstrated efficacy in two models of chemical-induced pain. Injection of PGE2 into the tail produces thermal hyperalgesia, and WAY-132983 blocked this response by more than 80%. In a model of visceral pain, the intraperitoneal injection of chemical irritants, such as PPQ, mimics gastrointestinal tract pain. Injection of PPQ produces an abdominal constricting behavior in mice, and WAY-132983 almost completely blocked this response. This is the first demonstration of the efficacy of muscarinic agonists in visceral pain, for which therapeutic approaches are limited.
To broaden the therapeutic potential of muscarinic agonists, we investigated the effects of WAY-132983 in a model of inflammatory pain and demonstrated that it reversed FCA induced mechanical hyperalgesia although effects on inflammation per se were not examined. Efficacy of muscarinic agonists in models of inflammatory pain is consistent with the proposed “cholinergic anti-inflammatory pathway” in which the CNS is in direct communication with peripheral tissues, such as the spleen (Borovikova et al., 2000).
Muscarinic agonists are also likely to be effective in a number of neuropathic pain states. In the SNL model of neuropathic pain, we demonstrate that WAY-132983 fully reversed the tactile allodynia, a clinical hallmark of neuropathic pain. Although the effects of muscarinic agonists in the SNL model have not been reported previously, our observations are consistent with the reported effects of cholinesterase inhibitors in the SNL model (Lavand'homme et al., 1998; Chiari et al., 1999; Hwang et al., 1999). Furthermore, cholinomimetics reverse the tactile allodynia seen in streptozotocin-induced diabetic rats (Chen et al., 2001), and an increase in mAChR expression levels (Chen and Pan 2003) makes it tempting to speculate that muscarinic agonists may be effective in the treatment of diabetic neuropathy.
We also demonstrate that WAY-132983 significantly and dose-dependently reversed tactile allodynia in an incisional model of postsurgical pain. Our observation is consistent with Prado and Segalla (2004) who reported that intrathecal administration of bethanecol produced antinociception in this model. Interestingly, the activities of bethanecol and the nicotinic agonist dimethylphenylpiperazinium were both sensitive to atropine. This may be due to the ability of the nicotinic agonist to stimulate the release of spinal ACh, which can then act on mAChRs. However, other mechanisms, including a nicotinic receptor-evoked release of norepinephrine, are indeed possible (Li and Eisenach, 2002).
The in vivo effects of WAY-132983 are mediated through the activation of muscarinic receptors as indicated by the fact that pretreatment with the nonselective muscarinic receptor antagonist scopolamine was able to block the in vivo effects of WAY-132983 in both the PPQ and FCA models. Subtype selectivity was addressed using the highly selective M4 toxin MT-3 (Ellis et al., 1999). MT-3 pretreatment alone did not alter FCA-induced mechanical hyperalgesia, but it almost completely blocked WAY-132983-induced antihyperalgesic effects in the FCA model. These data suggest that the majority of antihyperalgesic activity of WAY-132983 may be mediated through the M4 receptor. The in vitro data demonstrate that WAY-132983 is selective for the muscarinic system; therefore, the partial blockade by either antagonist may be due to insufficient exposure of the ligand to the site of action, differences in the pharmacokinetic profiles of WAY-132983 and the antagonist, or the involvement of other muscarinic receptors as discussed below.
Centrally expressed mAChRs, particularly M1, are thought to be involved in the supraspinal mechanisms of analgesia. Intracerebroventricular injections of antisense oligonucleotides designed against the M1 receptor prevent the antinociceptive effects of a subcutaneous administration of physostigmine or oxotremorine (Ghelardini et al., 2000). It is likely that activation of supraspinal M1 receptors ultimately leads to activation of the inhibitory descending noradrenergic and serotonergic pathways that activate spinally expressed M2 and M4 receptors. Furthermore, activation of muscarinic receptors, particularly the M1, in the thalamus may affect the emotional component of pain (Di-Cheng et al., 1988; Hart et al., 2004).
A role for the M4 receptor has been shown through receptor knockout mice. In the M2/M4 double knockout animal, oxotremorine had no activity, demonstrating the pivotal role of these subtypes in muscarinic-induced antinociception (Duttaroy et al., 2002). Although oxotremorine was less potent in the M2 knockout, it was still able to maximally inhibit pain stimuli presumably through activation of the M4 alone (Duttaroy et al., 2002). However, due to the limited selectivity of WAY-132983 against M2, we cannot rule out an involvement of this subtype. Further characterization of the effects of WAY-132983 in mAChR null mice is an attractive means to unequivocally address this issue.
The antinociceptive effects observed following activation of spinal mAChRs (particularly M2 and M4) are thought to mediated in part through an increased release of inhibitory transmitters and a decrease in the release of excitatory transmitters. The ACh-stimulated reduction of miniature excitatory postsynaptic potentials in rat spinal cord was blocked by the GABAB receptor antagonist CGP55845. This is consistent with a muscarinic-receptor evoked release of GABA, which then decreases glutamate release via presynaptic GABAB receptors (Baba et al., 1998; Iyadomi et al., 2000; Li et al., 2002). Accordingly, the antiallodynic action of intrathecal ACh or neostigmine is inhibited by CGP55845 (Li et al., 2002; Chen and Pan, 2003). ACh can also reduce currents in the presence of tetrodotoxin, suggesting a direct presynaptic effect of muscarinic agonists in inhibiting glutamate release. In rats, mAChR activation also increases the release of the inhibitory transmitter glycine, from spinal cord interneurons (Wang et al., 2006).
In summary, WAY-132983 is a potent, efficacious, and bioavailable muscarinic agonist with high brain penetration. As such, it constitutes an ideal tool for investigating the role of mAChRs in pain. We therefore undertook an extensive characterization of WAY-132983 in models of acute, chemical-induced, inflammatory, neuropathic, and postsurgical pain. Our results demonstrate that WAY-132983 is efficacious in these in vivo models and is effective in multiple species, against multiple stimulus modalities. Our observations suggest that muscarinic agonists may have a broad efficacy against a range of clinically relevant pain states.
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
ABBREVIATIONS: mAChR, muscarinic acetylcholine receptor; CNS, central nervous system; ACh, acetylcholine; WAY-132983, (3R,4R)-3-(3-hexylsulfanyl-pyrazin-2-yloxy)-1-aza-bicyclo[2.2.1]heptane; CHO, Chinese hamster ovary; h, human; r, rat; NMS, N-methylscopolamine; PGE2, prostaglandin E2; PPQ, para-phenylquinone; T10, temperature that produced half-maximal (10-s) tail withdrawal latency; FCA, Freund's complete adjuvant; PWT, paw-withdrawal threshold; SNL, spinal nerve ligation; BL, baseline latency; Pre, predrug reading; CGP 54626A, [S-(R*,R)]-[3-[[1-(3,4-dichlorophenyl)ethyl]amino]-2-hydroxypropyl][3,4-3H]-(cyclohexylmethyl)phosphinic acid; AUC, area under the curve; CGP55845, 3-N[1-(S)-(3,4-dichlorophenyl)ethyl]amino-2(S)-hydroxypropyl-p-benzyl-phosphinic acid; MT-3, muscarinic toxin-3.
- Received December 13, 2006.
- Accepted June 20, 2007.
- The American Society for Pharmacology and Experimental Therapeutics