Ketamine (Ketalar; Parke-Davis, Morris Plains, NJ) has been shown to inhibit muscarinic signaling with a median inhibitory concentration (IC50) of 5.7 microM. Whereas Ketalar is a racemic mixture, recent interest has focused on clinical use of the S(+) ketamine isomer, which is three times as potent an analgesic as the R(-) isomer yet seems to be associated with fewer psychoactive side effects. Therefore, the authors studied the effects of S(+) and R(-) ketamine and the preservative benzethonium chloride on muscarinic signaling.
Rat ml muscarinic acetylcholine receptors were expressed recombinantly in Xenopus laevis oocytes. Ca2(+)-activated Cl- currents in response to 10(-7) M acetyl-beta-methylcholine were determined by two-electrode voltage clamping in the presence of various concentrations of ketamine and benzethonium. Concentration-inhibition curves were constructed and used for algebraic and isobolographic analysis.
The IC50. was 125 +/- 33 microM for S(+) ketamine, and 91 +/- 19 microM for R(-) ketamine. This difference was not statistically significant, indicating that muscarinic inhibition by ketamine is not stereoselective. The R(-)/S(+) mixture had an IC50 of 48 +/- 1 microM, and thus the stereoisomers interact synergistically. When appropriate concentrations of benzethonium were added, an IC50 of 15 +/- 2 microM resulted.
The muscarinic inhibitory action of ketamine isomers is not stereoselective. Because S(+) ketamine is a significantly more potent analgesic, it should have less muscarinic inhibitory action than R(-) ketamine when used in clinically equivalent doses. A significant fraction of the muscarinic inhibitory action of Ketalar is due to the preservative benzethonium. If reconstituted with a different preservative, Ketalar might be a less potent muscarinic antagonist.
Phencyclidines have been shown to interact with muscarinic acetylcholine receptors, [1–5] and previously we showed that the phencyclidine derivative ketamine also has muscarinic antagonist activity.  Although ketamine's primary site of action is the N-methyl-D-aspartate receptor, [7,8] secondary actions of the compound might be relevant to the unique anesthetic state it induces. In particular, the inhibitory actions of ketamine on muscarinic signaling may have clinical relevance: In addition to explaining some of the compound's side effects (such as the excess sympathetic tone and bronchodilation observed after ketamine administration), they may be relevant to its anesthetic properties. Recent studies have shown that central muscarinic systems play profound roles in learning, maintenance of consciousness, and pain perception,  and that inhibition of these pathways results in states resembling sleep.
Our previous study  used Ketalar, the commercial form of ketamine available from Parke-Davis (Morris Plains, NJ). It is a racemic mixture of the two stereoisomers of ketamine, S(+) and R(-), and it also contains the preservative benzethonium chloride (BCl). Recently, interest has developed in using S(+) ketamine for clinical purposes because it is approximately three times as potent an analgesic as R(-) yet is associated with fewer of the troublesome psychoactive side effects. [10,11] Thus we thought that a study of the muscarinic inhibitory actions of the two ketamine isomers would be appropriate. We considered two possible outcomes of the investigation. If muscarinic inhibitory potencies of the ketamine isomers were found to correlate with their anesthetic potencies, this would suggest that muscarinic antagonism plays a role in the anesthetic action of the compound. If so, anesthetic action and muscarinic side effects probably could not be dissociated. In contrast, if muscarinic inhibitory potency and anesthetic potency did not correlate, this would suggest the possibility that the ketamine structure could be modified to obtain anesthetics with fewer muscarinic side effects.
Thus we studied the effects of S(+) and R(-) ketamine and that of the preservative BCl on the functioning of m1 muscarinic acetylcholine receptors recombinantly expressed in Xenopus oocytes. The specific goals of the study were 1) to determine if inhibition of muscarinic acetylcholine signaling by ketamine isomers is stereoselective, and 2) if so, to determine if muscarinic inhibitory potencies correlate with those observed for anesthetic action.
Materials and Methods
The study protocol was approved by the Animal Research Committee at the University of Virginia. Female Xenopus laevis toads were obtained from Xenopus I (Ann Arbor, MI), housed in an established colony, and fed regular frog brittle twice weekly. For removal of oocytes, an animal was immersed in an 8-mM solution of tricaine until unresponsive to a painful stimulus (toe pinching). A 1-cm-long incision was made in a lower abdominal quadrant, and a lobule of ovarian tissue, containing approximately 200 oocytes, was removed and placed in modified Barth's solution (containing 88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO3, 0.41 mM CaCl2, 0.82 mM MgSO4, 0.3 mM Ca2(NO3)2, 15 mM HEPES, and 0.1 mg/ml gentamicin, with the pH adjusted to 7.6). The wound was closed in two layers, and the animal was allowed to recover in a separate tank overnight.
The ovarian tissue was washed immediately with modified Barth's solution and dissected into small clusters of 20–50 oocytes. Mature oocytes (Dumont stage V and VI) were isolated, and the follicular layer was removed manually from each oocyte using microforceps.
The rat m1 muscarinic acetylcholine receptor was a gift of T. I. Bonner (National Institutes of Mental Health, Bethesda, MD). Its complementary DNA (cDNA) consists of a 2.8-kilobasepair fragment in a commercial vector (pGEM1; Promega, Madison, WI). The construct was linearized by digestion with the nuclease HindIII, and complementary RNA (cRNA) was transcribed in vitro using the bacteriophage RNA polymerase T7. A capping analog (sup 7m GpppG) was included in the reaction to generate capped transcripts because these are translated more efficiently in the oocyte. The resulting cRNA was quantified by spectroscopy and diluted in water to a concentration of 0.1 mg/ml. Each oocyte was injected with 5 ng cRNA using an automated microinjector (Nanoject; Drummond Scientific, Broomall, PA). Adequacy of injection was confirmed by noting the slight increase in cell size during injection. The cells were cultured in modified Barth's solution at 18 degrees Celsius for 3 days before study.
A single defolliculated oocyte was placed in a perfusable recording chamber (3-ml volume) filled with Tyrode's solution (containing 150 mM NaCl, 5 mM KCl, 1 mM MgCl2, 2 mM CaCl2, 10 mM dextrose, 10 mM HEPES, with the pH adjusted to 7.4), with appropriate concentrations of ketamine isomers or BCl. Microelectrodes were pulled in one stage from capillary glass (BBL with fiber; World Precision Instruments, Sarasota, FL) on a micropipette puller (model 700C; David Kopf Instruments, Tujunga, CA). Tips were broken to a diameter of approximately 10 micro meter, providing a resistance of 1–3 M Ohm. Electrodes were filled with 3 M KCl.
The cell was voltage clamped using a two-microelectrode oocyte voltage clamp amplifier (OC725C; Warner Corp., New Haven, CT) connected to a data acquisition and analysis system running on an IBM-compatible personal computer. The acquisition system consisted of a DAS-8 A/D conversion board (Keithley-Metrabyte, Taunton, MA), and analysis was performed with OoClamp software.  Holding potential was -70 mV. Cells that did not show a stable holding current of less than 1 micro Amphere during a 1-min equilibration period (fewer than 5% of cells tested) were excluded from analysis. Membrane current was sampled at 125 Hz and recorded for 5 s before and 55 s after application of the test compound. Acetyl-beta-methylcholine (10 sup -7 M) was applied to the oocyte in a 30-micro liter aliquot for 1 or 2 s using a micropipettor positioned approximately 5 mm from the oocyte. Each oocyte received only a single agonist application.
Intracellular Ca2+ release in the Xenopus oocyte leads to opening of endogenous Ca2+-activated Cl sup - channels and a resulting inward Cl sup - current (ICl(Ca)), [6,13] which can be used as a measure of agonist-induced Ca2+ release. Responses were quantified by integrating the current trace by quadrature and are reported as charge movements in units of microCoulombs (micro Coulomb) because this reflects Ca2+ release better than peak current does. 
Responses are reported as means +/- SEM. Concentration-inhibition curves were fitted to the Hill equation, and Hill coefficients and median inhibitory concentration (IC50) values and their 95% confidence intervals were calculated. These were used for algebraic (fractional) and isobolographic analysis of drug interaction, using standard techniques. [15–17] Briefly, the most commonly used form of isobolographic analysis assumes that the effect of a combination of drugs without synergistic interaction is the sum of the effects of its constituents [E sub (a + b)= E sub (a)+ E sub (b)]. [17,18] Therefore, we plotted the deviation of the IC50of a mixture of two compounds from the line of additivity (i.e., a straight line joining the two single-drug IC50S) together with the 95% confidence interval. To determine if the difference between the IC50of the mixture and the line of additivity was significant, we performed a t test. Algebraic analysis was performed as follows. At IC50for the mixture, we determined, for each compound present, the ratio (rc) of its concentration in the mixture to its IC50when used alone. The rcvalues for all compounds present in the mixture were then added (Sigma rc). [17,18] If the interaction is additive, Sigma rcshould be 1. If Sigma rcis smaller than 1, then potentiation (or superadditivity) of inhibition is present. If Sigma rcis larger than 1, then antagonism (or subadditivity) of inhibition is present. R, the inverse of Sigma rc, commonly is used to describe the strength of synergistic interactions. Probability values less than 0.05 were considered significant.
Molecular biology reagents were obtained from Promega (Madison, WI). All other chemicals were obtained from Sigma (St. Louis, MO). Ketamine isomers were supplied by Parke-Davis GmbH (Berlin, Germany).
Expression of Muscarinic Acetylcholine Receptors in Oocytes
Whereas control oocytes were unresponsive to 10 sup -7 M acetyl-beta-methylcholine (MCh; data not shown), oocytes injected with cRNA encoding the m1 muscarinic acetylcholine receptor responded to MCh with a transient inward current (Figure 1(A)). This current developed within several seconds after agonist application, and membrane permeability returned to baseline after approximately 30 s. Average response sizes were 5.4 +/- 0.5 micro Coulomb. These findings are similar to those in our previous reports, [6,13] in which we showed these induced currents to be ICl(Ca).
Ketamine Stereoisomers Synergistically Inhibit Muscarinic Signaling
We tested the ability of S(+) and R(-) ketamine to inhibit I sub Cl(Ca) induced by 10 sup -7 M MCh in oocytes expressing m1 receptors. Oocyte responses were depressed in the presence of either compound. Representative traces are shown in Figure 1(B and C), and Figure 2(A and B) show concentration-inhibition curves, obtained by fitting the experimental data to the Hill equation. Calculated IC50S were 125 +/- 33 micro Meter for S(+) and 91 +/- 19 micro Meter for R(-) ketamine (Table 1). Hill coefficients were 1.1 for each drug. The IC50values were not significantly different, indicating that the muscarinic inhibitory action of ketamine on muscarinic receptor signaling is not stereoselective.
In our previous study,  we observed inhibition of muscarinic signaling by Ketalar with a calculated IC50of 6 micro Meter. Because the IC50values from the individual stereoisomers were so different from that obtained with the commercial preparation, we hypothesized that a synergistic inhibitory action might exist between S(+) and R(-) ketamine. Therefore, we determined the inhibitory potency of an equimolar mixture of S(+) and (R-) ketamine. A representative trace of inhibition by this mixture is shown in Figure 1(D), and the concentration-inhibition relation is shown in Figure 2(C). The calculated IC50was 48 +/- 1 micro Meter, with a Hill coefficient of 1.5. This suggests modest synergistic activity between the stereoisomers, which could be confirmed by isobolographic analysis (Figure 2(D)). Algebraic analysis revealed a Sigma rcvalue of 0.45 and an R value of 2.2, which was significantly different (P = 0.025) from additivity (Table 1).
Nonetheless, the calculated IC50of the reproduced racemic mixture was still nearly a factor of 10 removed from that observed with Ketalar. Therefore, we studied the influence of the preservative in the mixture.
Benzethonium Chloride Is a Muscarinic Antagonist
Benzethonium chloride is present at concentrations of not more than 0.1 mg/ml in the stock solution of Ketalar, which contains 10 mg/ml racemic ketamine. Because BCl shows structural similarity to synthetic muscarinic antagonists, we hypothesized that the compound might have an inhibitory action on muscarinic signaling. Therefore, we determined the ability of BCl to inhibit ICl(Ca) in response to 10 sup -7 M MCh in oocytes expressing m1 muscarinic receptors. A representative trace is shown in Figure 1(E), and the concentration-inhibition relation is presented in Figure 3. The calculated IC50for the compound is 0.88 +/- 0.25 micro Meter (Table 1), making it a more potent muscarinic antagonist than ketamine. The Hill coefficient for the concentration-inhibition relation was 0.9.
Benzethonium Chloride Acts Synergistically with Ketamine
However, at a concentration of racemic ketamine of approximately 50 micro Meter (IC50), only 0.3 micro Meter BCl will be present in the commercial preparation, making significant inhibition of muscarinic signaling by the compound itself unlikely. Therefore, we hypothesized that a synergistic interaction between ketamine and BCl might explain the IC50of 6 micro Meter observed with Ketalar. To test this hypothesis, we reproduced the commercial preparation by mixing equimolar concentrations of S(+) and R(-) ketamine with the appropriate concentration of BCl and tested the resulting mixture for muscarinic inhibitory ability. A representative trace of its inhibitory activity is shown in Figure 1(F), and the concentration-response relation is shown in Figure 4(A). The reproduced mixture inhibited muscarinic signaling with a calculated IC50of 15 +/- 2 micro Meter and a Hill coefficient of 0.6. This IC50value is close to that previously observed with the commercial mixture.
As this suggests a synergistic interaction between ketamine and BCl, we subjected our data to isobolographic analysis, as shown in Figure 4(B). This analysis confirmed that ketamine and BCl act synergistically. As indicated in Table 1, Sigma rcwas 0.40, and R was 2.5, similarly indicating synergistic interaction (P = 0.016).
Finally, to ascertain that our previous findings with Ketalar were reproducible, we determined its muscarinic inhibitory activity. A representative trace is shown in Figure 1(G), and the concentration-inhibition relation is shown in Figure 4(C). Ketalar inhibited MCh-induced ICl(Ca) with an IC50of 7.2 +/- 3.8 micro Meter, which is remarkably similar to our previously reported findings and not significantly different from the data obtained with our reproduced mixture. The Hill coefficient was 0.7. If the concentration of BCl in Ketalar stock solution was assumed to be 0.1 mg/ml, Sigma rcwas 0.20, and R was 5.0, which is significantly different from additivity (P < 0.001;Table 1). This is a conservative estimate of the synergism; if the concentration of BCl was actually less than 0.1 mg/ml, a greater degree of synergism would be present.
The main findings from this study are 1) the inhibition of muscarinic signaling by ketamine is not stereoselective, 2) the isomers act synergistically in their effect on muscarinic signaling, 3) BCl is a muscarinic antagonist, and 4) BCl acts synergistically with ketamine in its effect on muscarinic signaling.
Muscarinic signaling has been a subject of significant interest recently, particularly as it has become clear that muscarinic pathways in the central nervous systems are relevant to several anesthetic states, such as consciousness, learning, and pain perception.  In addition, the cloning of the muscarinic receptors  and continuing advances in molecular biology have facilitated the study of muscarinic signaling pathways. All five subtypes of muscarinic receptors are represented in the brain and usually in mixed populations.  Because different subtypes might have different anesthetic sensitivities and because no completely selective agonists and antagonists are available, direct study of brain tissue is difficult. For this reason, we chose to express one specific subtype of receptor in an environment that is easily studied. Choice of subtype was guided by brain distribution: the m1 receptor represents 34–40% of muscarinic receptor protein in cortex and 47% in hippocampus, making it the most prominent subtype in these areas.  Because of the relevance of these areas to anesthetic action, we chose to study the m1 receptor. Our findings, therefore, do not preclude the possibility that ketamine might act in a stereoselective manner at these other receptors.
Muscarinic receptors have been expressed frequently in the Xenopus oocyte model system by ourselves [6,13] and by other investigators, [19,21] and it is well established that they function similarly in oocytes and in native cells. Although the receptors are expressed at a lower temperature than they are normally exposed to, this has not been shown to result in significant changes in kinetics or function. Therefore, we are confident that our results can be extrapolated to cells with endogenously expressed muscarinic receptors. Obviously, they cannot be extrapolated to the other muscarinic receptor subtypes.
Our finding that the muscarinic antagonistic properties of ketamine are not stereoselective is important because it makes muscarinic inhibition unlikely to be essential in ketamine's mode of anesthetic action, which is strongly stereoselective: S(+) ketamine is approximately three or four times as potent an anesthetic as R(-) ketamine, whereas R(-) ketamine exhibits more psychic side effects. [10,11]
Two other implications follow from the lack of stereospecificity observed in the present study. First, it suggests a benefit of using the S(+) isomer as a clinical anesthetic: when used in equipotent concentrations, it would exhibit less muscarinic inhibition than the R(-) isomer does. Second, our findings suggest that (nonstereoselective) muscarinic inhibition by ketamine is mediated by a different part of the molecular structure than (stereoselective) anesthetic action is. If muscarinic inhibition is not essential to the anesthetic action of ketamine, it might be possible to develop compounds that retain ketamine's anesthetic activity yet lack muscarinic inhibitory action. The observed Hill coefficients indicate that the interaction is unlikely to be cooperative in nature.
Whereas ketamine's primary mode of anesthetic action is considered to be competitive antagonism at the N-methyl-D-aspartate receptor, [7,8] this does not preclude the relevance of secondary activities. A recent report suggests that the direct cardiac depressant effects of ketamine might be due to interaction with cardiac Ca2+ channels.  The compound was shown to decrease contractile force and transsarcolemmal Ca2+ current. As in our study, the S(+) and R(-) isomers were of similar potency. However, this issue cannot be considered resolved because other investigators found the S(+) isomer to induce less cardiac depression than the R(-) isomer, a result attributed to increased availability of catecholamines.  Suppression of neutrophil function by ketamine has been shown to be nonstereoselective.  It appears, therefore, that at least some secondary effects of ketamine can result from nonstereoselective interactions.
Our finding that BCl is a muscarinic antagonist by itself was surprising. To our knowledge, an action on muscarinic signaling for this compound has not been suggested before. The compound has been used primarily as a topical antiseptic or anti-infective agent, and also as a preservative, as with Ketalar. There are, however, similarities in chemical structure between BCl and several synthetic antimuscarinic agents, such as eucatropine, tridihexetyl chloride, and tropicamide: All consist of a modified ring structure separated by a carbon chain of variable length from a modified choline group. Hence, in retrospect, it is not surprising that BCl would have significant antimuscarinic activity. Of interest is the finding that it acts synergistically with ketamine. In the absence of a synergistic interaction, the concentration of BCl in Ketalar would be too low to be clinically relevant as an antimuscarinic compound. However, our data indicate that the presence of such low concentrations of BCl potentiates the antimuscarinic action of ketamine by approximately three times. This suggests that reformulating ketamine with a different preservative might reduce its antimuscarinic activity. Interestingly, whereas ketamine marketed in the United States and Germany, among other countries, is formulated with BCl, commercial ketamine marketed in France, although produced by the same company, contains chlorobutanol as a preservative.  We are not aware of studies determining the muscarinic antagonist properties of this formulation.
Our study shows that the muscarinic inhibitory effects of ketamine are not stereoselective, which makes it unlikely that muscarinic inhibition plays an important role in the anesthetic properties of ketamine. However, the synergistic interaction between the ketamine stereoisomers and the preservative BCl may explain some of the antimuscarinic side effects of the commercial compound.