Neuromuscular relaxants such as pancuronium bind to M2 and M3 muscarinic receptors as antagonists. Blockade of muscarinic receptors in atria of the M2 subtype mediates tachycardia. In the lung, blockade of M2 receptors on parasympathetic nerves potentiates vagally induced bronchospasm, whereas blockade of M3 receptors on bronchial smooth muscle inhibits bronchospasm. The current study was designed to quantify the affinity of a series of neuromuscular relaxants for the M2 and M3 muscarinic receptors, which were individually stably transfected in Chinese hamster ovary cell lines.
Competitive radioligand binding assays determined the relative binding affinities of the neuromuscular relaxants pancuronium, succinylcholine, mivacurium, doxacurium, atracurium, rocuronium, gallamine, and pipecuronium for the muscarinic receptor in the presence of a muscarinic receptor antagonist (3H-QNB) in membranes prepared from cells individually expressing either the M2 or M3 muscarinic receptor.
All muscle relaxants evaluated displaced 3H-QNB from muscarinic receptors. The relative order of potency for the M2 muscarinic receptor (highest to lowest) was pancuronium, gallamine, rocuronium, atracurium, pipecuronium, doxacurium, mivacurium, and succinylcholine. The relative order of potency for the M3 muscarinic receptor (highest to lowest) was pancuronium, atracurium, pipecuronium, rocuronium, mivacurium, gallamine, succinylcholine, and doxacurium.
All neuromuscular relaxants studied had affinities for the M2 and M3 muscarinic receptor, but only pancuronium and gallamine had affinities within the range of concentrations achieved with clinical use. The high affinities of gallamine and pancuronium for the M2 muscarinic receptor are consistent with a mechanism of M2 receptor blockade in relaxant-induced tachycardia.
MUSCARINIC cholinergic receptors found on postganglionic neurons, cardiac muscle, smooth muscle, and glands mediate important physiologic functions such as heart rate, airway caliber, and salivation. Molecular cloning studies have identified five different subtypes of muscarinic receptors encoded by five distinct genes (m1, m2, m3, m4, and m5), although pharmacologic techniques only distinguish between the M1, M2, M3, and M sub 4 muscarinic receptors. 
Neuromuscular blockers commonly used in anesthetic practice interact with muscarinic and nicotinic receptors. [2,3]Agents that are antagonists for the M2muscarinic receptor should increase heart rate because the heart expresses a relatively pure population of M sub 2 receptors. Tachycardia results from blockade of M2muscarinic receptors on both cardiac myocytes and vagal efferent fibers. Similarly, M2selective muscarinic antagonists should potentiate vagally induced bronchoconstriction, [3,5]because M2receptor activation also inhibits the release of acetylcholine from postganglionic parasympathetic fibers in the airway. Conversely, neuromuscular blockers that selectively inhibit the M3muscarinic receptor on the smooth muscle in the airway should inhibit vagally induced bronchoconstriction, because M3receptor activation mediates smooth muscle contraction in the airway. Therefore, M sub 2 -selective muscarinic antagonist activity that occurs at doses of neuromuscular relaxants in the clinical range is undesirable because of the side effects of tachycardia and bronchospasm, whereas M3-selective muscarinic receptor antagonist activity that occurs in the dose of the neuromuscular relaxant used clinically is desirable as it results in decreases in airway tone and salivary secretions. Gallamine, a neuromuscular blocker no longer used because of a high prevalence of undesirable side effects, was later discovered to be a highly selective M2muscarinic receptor antagonist. 
Because lung and most other tissues express a mixed population of muscarinic receptors, and because the differences in affinity for even subtype-selective muscarinic antagonists are often minimal, the estimation of the affinity of the neuromuscular blockers for M2and M3receptors cannot be calculated with any degree of precision in tissues coexpressing multiple muscarinic receptor subtypes. Pure populations of muscarinic receptors in a tissue provide a unique way of obtaining pharmacologic profiles of each receptor subtype. Until recently, heart tissue was thought to express a pure population of M2muscarinic receptors, but molecular and functional evidence exists for the expression of M1muscarinic receptors also. Many peripheral tissues, such as smooth muscle of airway and gut, that express M3muscarinic receptors also express M2muscarinic receptors; therefore, we used two separate cell lines, one expressing the M2and the other expressing the M3muscarinic receptor, to characterize a series of commonly used neuromuscular blocking drugs regarding their affinities at the M2and M3muscarinic receptor.
Materials and methods
Drugs and Reagents
Pipecuronium and rocuronium were gifts from Organon (West Orange, NJ); doxacurium, mivacurium, and atracurium were gifts from Glaxo-Wellcome (Research Triangle Park, NC). Pancuronium and gallamine were purchased from Sigma (St. Louis, MO). The muscarinic receptor antagonist3H-QNB was purchased from Amersham Life Science (Arlington Heights, IL). All drugs were dissolved in dH2O. Chinese hamster ovary (CHO) cells stably transfected with complementary deoxyribonucleic acid encoding the rat M3muscarinic receptor were purchased from the American Type Culture Collection (Rockville, MD). Chinese hamster ovary cells stably transfected with complementary deoxyribonucleic acid encoding the human M2muscarinic receptor were provided by Norman Lee, Ph.D. (The Institute for Genome Research, Rockville, MD).
Chinese hamster ovary cells stably transfected with either the M2or M3muscarinic receptor were grown in Dulbecco's modified essential medium (DMEM) media containing 10% fetal bovine serum and antibiotic agents (100 U/ml penicillin G sodium, 100 micro gram/ml streptomycin sulfate, 0.25 micro gram/ml amphotericin B, and 100 U/ml nystatin). Cells were cultured in T500 flasks (500 cm2) at 37 [degree sign] Celsius in a humidified atmosphere of 5% CO2/95% air and harvested at confluence.
Preparation of Chinese Hamster Ovary Cell Membranes
At confluence, culture media were removed from flasks. Cells were incubated in lysis buffer (10 mM HEPES, 2 mM ethylenediaminetetraacetic acid, 100 micro Meter PMSF, pH 8.0) at 37 [degree sign] Celsius in a carbon dioxide incubator until detached (20–40 min). Lysed cells were centrifuged at 48,000 x g (Sorvall RC-5B with SS-34 rotor, Newton, CT) for 20 min at 4 [degree sign] Celsius. Cold HEPES buffer (100 mm, pH 7.4) was used to resuspend the pellet after the removal of the supernatant. The lysates were washed two additional times, and the final pellet was resuspended in 6 ml of HEPES buffer at 2–5 mg/ml and stored at -70 [degree sign] Celsius until used for radioligand binding assays.
Saturation Radioligand Binding
Forty micrograms of CHO cell membrane proteins was incubated with increasing concentrations of the radioligand3H-QNB (47 Ci/mmol; 0.05–3.00 nM) in the presence or absence of atropine (2 micro Meter) in binding buffer (40 mm KH2PO4, 160 mM K2HPO sub 4 in 50 mm NaCl, pH 7.4). All radioligand experiments were incubated for 2 h at room temperature in a final volume of 1 ml. Preliminary experiments confirmed that the 2-h incubation period was adequate to achieve equilibrium binding. All binding experiments were terminated by filtration through GF/B glass fiber filters and washed three times with 5 ml of cold 0.9% NaCl using a cell harvester (Brandell, Gaithersburg, MD). Filters were immersed in 5 ml of Econo scintillation fluid, stored overnight, and counted in a scintillation counter (Beckman LS 5000 TD; Beckman, Fullerton, CA) with an efficiency of 45–50%. Specifically bound counts were analyzed by linear regression after Scatchard transformation using the EBDA computer program to obtain the line of best fit.
Competitive Radioligand Binding Assays
Forty micrograms of CHO cell membrane protein was incubated with3H-QNB (0.18 nM) and muscle relaxants of increasing concentrations (10 sup -10 -10 sup -3 M), under conditions described previously for saturation experiments. The chosen radioligand concentration (0.18 nM) for the competition experiments was 3.6 times the equilibrium constant (Kd) of the M2muscarinic receptor and 1 x Kdof the M3muscarinic receptor. The competitive displacement of3H-QNB by increasing concentrations of muscle relaxants was analyzed by nonlinear regression. A reiterative curve-fitting program, Inplot 4.0 (Graph Pad, San Diego, CA) was used to calculate the relative binding affinity (IC50) values using a four-parameter logistic equation (log scale) with the slope factor set to -1.
The Kdof3H-QNB and the receptor numbers of the M sub 2 or M3muscarinic receptor in CHO cell membranes were calculated from radioligand saturation binding assays after Scatchard transformation. The Kdwas determined from the negative reciprocal of the slope of the line and the maximum number of binding sites (Bmax), were determined from the x-intercept. IC50, values were obtained by nonlinear regression analysis of competitive radioligand displacement curves using the values from the reiterative curve-fitting program Inplot using a four-parameter logistic equation (log scale) with the slope factor set to -1.
All data are presented as mean +/- SE unless otherwise indicated. The CHO M2and CHO M3cells were first characterized for the level of expression of muscarinic receptors and for receptor affinity for the antagonist3H-QNB and the agonist carbachol. Saturation of specific binding was achieved in the membranes of the CHO M2or CHO M3cells over the range of3H-QNB used (0.01–3.00 mM; n = 3;Figure 1(A and B)). Hill coefficients were 1.03 +/- 0.14 and 0.92 +/- 0.04 for M2and M3muscarinic receptors, respectively, which was indicative of a single class of binding sites in both CHO M2and CHO M3cells (n = 3). Scatchard transformation of saturation binding showed that the CHO M2membranes contained 681 +/- 7 fmol/mg protein of receptor (Bmax) with an affinity (-log Kd) of 10.3 +/- 0.1 and that CHO M3membranes contained 580 +/- 67 fmol/mg protein of receptor with affinity (-log Kd) of 9.7 +/- 0.007 (n = 3;Figure 1(A and 1)). Forty micrograms of membrane proteins and 0.18 nM3H-QNB yielded specific binding rates of 92% for the M2muscarinic receptor and 89% for the M3muscarinic receptor.
In membranes prepared from cells expressing either the M2or the M3muscarinic receptor, carbachol displaced3H-QNB with two affinity sites, which is expected for agonists that bind to G protein-coupled receptors (data not shown). These two affinity sites represent the receptor associated with the G protein (high affinity) and receptor disassociated from the G protein (low affinity). Chinese hamster ovary cells expressing either the M2or the M3muscarinic receptor showed a biphasic displacement curve for the competitive binding of carbachol and3H-QNB, with -log Ki affinities of 4.77 and 6.68 for CHO M2and 4.23 and 6.66 for CHO M sub 3, respectively (n = 3).
In competitive binding experiments using the muscle relaxants and3H-QNB, the IC50values of each muscle relaxant at the M sub 2 and M3muscarinic receptors were determined by nonlinear regression analysis (Table 1). At least three separate competitive displacements were performed for each muscle relaxant. All muscle relaxants tested (pancuronium, gallamine, succinylcholine, atracurium, mivacurium, pipecuronium, doxacurium, and rocuronium) displaced3H-QNB from M2and M3muscarinic receptors in a dose-dependent manner (Figure 2and Figure 3). There was a wide variation, however, in the affinity of muscle relaxants for both the M2and M3muscarinic receptors. Among the muscle relaxants evaluated, pancuronium had the highest affinity for the M2receptor, with an IC50of 0.28 micro Meter. Newer-generation muscle relaxants (mivacurium, doxacurium, rocuronium, and pipecuronium) had lower affinities for the M sub 2 muscarinic receptor (Table 1). The relative order of potency for the M2muscarinic receptor (highest to lowest) was pancuronium, gallamine, rocuronium, atracurium, pipecuronium, doxacurium, mivacurium, and succinylcholine.
In general, the affinity of the muscle relaxants for the M sub 3 muscarinic receptor was lower than the affinity for the M2muscarinic receptor (Table 1). Pancuronium had the highest affinity for the M3muscarinic receptor (IC50= 1.2 micro Meter) among the relaxants evaluated. The newer-generation muscle relaxants (mivacurium, doxacurium, rocuronium, and pipecuronium) had much lower affinities for the M3muscarinic receptor than pancuronium, with IC sub 50 values ranging from 1.4–145.0 micro Meter. The relative order of potency for the M3muscarinic receptor (highest to lowest) was pancuronium, atracurium, pipecuronium, rocuronium, mivacurium, gallamine, succinylcholine, and doxacurium.
The affinities of neuromuscular blocking agents for the M2and M3muscarinic receptors were characterized in an effort to relate the affinities of the neuromuscular relaxants for muscarinic receptors with their known clinical side effects. This study performed in cells with pure populations of either the M2or M3muscarinic receptor allowed affinity measurements without the competing effect of additional muscarinic receptor sub-types in the same cell. This eliminated a potential difficulty, seen in previous studies of native tissues in which multiple muscarinic receptor subtypes were present, confounding individual subtype affinity measurements. 
All neuromuscular blocking agents evaluated in the current study had measurable affinities for both the M2and M3muscarinic receptors, but there was a wide variation in their affinities for both receptors. In general, the affinity measurements determined in the CHO M2cells were similar to those reported in cardiac membranes, which are thought to predominantly express M2muscarinic receptors. Less similarity was found between the affinities of muscle relaxants measured in the CHO M3cells and those reported in the literature for tissues such as ileum. Although ileum was originally thought to contain a relatively pure M3muscarinic population, a mixture of M2and M3muscarinic receptors is now known to exist, with a predominance of the M2subtype. In several studies, no differences were found in the affinities of muscle relaxants for the muscarinic receptor in heart or ileum, which is likely due to the predominance of the M2subtype in both tissues. [2,11]
The clinical side effects induced by neuromuscular blocking agents are thought to be caused by their interaction with muscarinic receptors in the heart and airway. Muscle relaxants can interact with the muscarinic receptor in two ways. They can bind directly to the ligand binding site with agonist or antagonist properties, or, alternatively, muscle relaxants can interact with another region of the receptor, which causes a change in the characteristics of the binding site thus causing a decrease in ligand binding (a negative allosteric effect, e.g., gallamine)or an increase in ligand binding (a positive allosteric effect, e.g., alcuronium). In the current study, no attempt was made to distinguish between competitive antagonism versus negative allosteric effects of the muscle relaxants, as the clinical effect of either interaction is the same; the muscle relaxant would decrease the binding of acetylcholine to the muscarinic receptor.
In the current study, pancuronium and gallamine had the highest affinities for the M2muscarinic receptor, whereas pancuronium and atracurium had the highest affinities for the M3muscarinic receptor. The affinities of pancuronium and gallamine for the M2receptor measured in the current study in a stably transfected cell line are in accordance with previous studies performed using rat heart membranes that predominantly express M2muscarinic receptors. [2,14]Rocuronium, atracurium, and pipecuronium had lower affinities for the M2muscarinic receptor than did pancuronium and gallamine. Data from the current study are also consistent with a study in rat atria in which rocuronium had a higher affinity for muscarinic receptors than pipecuronium (Ki = 13.6 and 27 micro Meter, respectively). Atracurium in this study had an affinity for the M2receptor similar to the affinity reported in rat atria. One expected clinical consequence of blockade of the cardiac M2muscarinic receptor is tachycardia, which occurs clinically with gallamine and pancuronium but is not reported with the newer-generation muscle relaxants. The results of our study are consistent with this mechanism of muscle relaxant-induced tachycardia in that gallamine and pancuronium were found to have the highest affinities for the M2muscarinic receptor, with affinity values within the range of serum concentrations achieved with clinical use. Pancuronium and gallamine had IC50affinity values for the M2muscarinic receptor of 0.28 and 0.59 micro Meter, respectively, which are 14- and 137-fold lower, respectively, than the concentrations of these muscle relaxants achieved with routine clinical use, suggesting high M2muscarinic receptor blockade with these agents. The muscle relaxant with the next highest affinity for M2muscarinic receptors was rocuronium, which had an affinity IC50value of 3.0 micro Meter-a value only 1.6-fold lower than the concentration achieved with clinical use of rocuronium. The other muscle relaxants evaluated had affinities substantially lower than the serum concentrations achieved with normal clinical use.
Okanlami et al. from our laboratory previously evaluated the interaction of the nondepolarizing muscle relaxants pancuronium, mivacurium, pipecuronium, and doxacurium at M2and M sub 3 receptors in guinea pig heart and lung in vivo and found that the rank order of potency of the affinities (highest to lowest) for the M sub 2 muscarinic receptor was similar to that seen in the current study-pancuronium, pipecuronium, mivacurium, and doxacurium-but they were unable to establish a rank order of potency for the M3muscarinic receptor.
The affinities of pancuronium and gallamine for the M3muscarinic receptor measured in the current study in cells that express only M3muscarinic receptors were lower than that reported for the rat ileum. This discrepancy is likely explained by the measurement of affinity values in a tissue with a mixed population of muscarinic receptor subtypes (rat ileum), as opposed to the current study in which affinity was measured in cells expressing only the M3muscarinic receptor subtype.
Muscarinic receptor control of airway tone is complicated. Airway smooth muscle expresses both M2and M3muscarinic receptors, whereas parasympathetic postganglionic nerves express M2muscarinic receptors. The net effect of neuromuscular relaxant-induced effects on airways depends on the relative blockade of M2and M3muscarinic receptors. Because M3muscarinic receptor activation is associated with initiation of airway smooth muscle contraction, agents that are potent antagonists at the M3muscarinic receptor should inhibit bronchoconstriction despite the M2muscarinic receptor blockade and the increased release of acetylcholine from parasympathetic postganglionic nerves. Therefore, pancuronium, which is more potent than gallamine as an M2antagonist, should not and clinically does not appear to be associated with bronchoconstriction because pancuronium is a potent M3muscarinic receptor antagonist at doses in the clinical range.
The current study of relative affinities of neuromuscular relaxants for muscarinic receptors is not a direct analysis of the clinical side effects of these drugs; however, the results suggest a possible role of muscarinic receptors in the mechanism of muscle relaxant-induced tachycardia by drugs such as gallamine and pancuronium due to M2muscarinic receptor blockade. Moreover, the simultaneous blockade of M2and M3muscarinic receptors in the airway by relaxants would also explain a relative lack of clinical effect on airway tone, because blockade of the M3receptor on the airway muscle prevents constriction by the increased release of acetylcholine by blockade of M2muscarinic receptors on parasympathetic nerves.