Modulation of NMDA Receptor Function by Ketamine and Magnesium: Part I : Anesthesia & Analgesia

Secondary Logo

Journal Logo

Anesthetic Pharmacology: Research Report

Modulation of NMDA Receptor Function by Ketamine and Magnesium: Part I

Liu, Hong-Tao MD†,; Hollmann, Markus W. MD†§,; Liu, Wei-Hua MD†,; Hoenemann, Christian W. MD†‡,; Durieux, Marcel E. MD, PhD*†

Author Information
Anesthesia & Analgesia 92(5):p 1173-1181, May 2001. | DOI: 10.1097/00000539-200105000-00019
  • Free


Glutamate, its receptors, and its roles in physiology and pathophysiology of the central nervous system (CNS) have attracted increasing attention over the past decade. Glutamate signals through various receptor types. Those receptors selectively responsive to N-methyl-d-aspartate (NMDA) have been studied in most detail. NMDA receptors are ionotropic receptors (ligand-gated ion channels), and unique in that channel activation requires binding of glutamate with glycine as an obligatory coagonist (1). NMDA receptors mediate neuronal signaling and regulate neuronal gene expression, and therefore perform critical roles in CNS functioning. However, excessive stimulation of these receptors can induce neuronal damage and death, and is assumed to be a final common pathway in the pathogenesis of many neurological diseases (2).

Because of its role in CNS function and neuronal injury, and particularly its involvement in pain processing, neuronal plasticity, and generation of central sensitization (“wind-up”) after nociceptive stimuli (3), NMDA receptor signaling is important in anesthesia. These events appear to be relevant not only to chronic pain, but also determine, in part, duration and intensity of postoperative pain (3). Therefore, blocking these processes by inhibiting NMDA receptor signaling promises to be useful in preventing development of prolonged pain states. NMDA receptor antagonists can prevent the induction of central sensitization attributed to peripheral nociceptive stimulation and abolish the hypersensitivity once it is established (3). As a result, they reduce manifestations of experimental neuropathic pain (4), wind-up (5), and spontaneous pain (6) in clinical studies.

No selective NMDA receptor antagonists are available for clinical use. However, several compounds approved for use in humans for other indications have significant NMDA receptor-blocking properties (7). Of these, ketamine and Mg2+ are of most interest. Unfortunately, clinical benefits may be limited by dose-related side effects. Because both compounds affect NMDA signaling, we hypothesized that a combination of ketamine and Mg2+ might be clinically useful, and possibly more effective than either compound alone. In an initial clinical study, we demonstrated that the two compounds indeed potentiate each other: the administration of both Mg2+ and ketamine reduces postoperative morphine consumption more than does each compound alone (8).

However, in that study, we did not investigate whether the potentiating effect indeed results from interactions between the compounds at the NMDA receptor. This is more easily addressed in an in vitro system. In the present report we therefore determined the effects of ketamine or Mg2+ on NMDA receptors expressed recombinantly in Xenopus oocytes, and tested the hypothesis that the combined application of the compounds is associated with super-additive inhibition of NMDA signaling. As reported in our companion article, we investigated these interactions in the presence of volatile anesthetics.


The study protocol was approved by the Animal Care and Use Committee at the University of Virginia. Oocytes were obtained and defolliculated as we described previously (9). The NMDA receptor consists of NR1 and NR2 subunits. The rat NR1, NR2A, and NR2B complementary DNA (cDNA) were obtained from Dr. P. J. Whiting (Merck, Sharp & Dohme Research Laboratories, Essex, UK). Three nanograms of cDNA in a 9.2-nL volume was injected into the nucleus of the oocyte, using an automated microinjector (Nanoject; Drummond Scientific, Broomall, PA). To minimize formation of homomeric NR1 receptors, an NR1/NR2 cDNA weight ratio of approximately 1:5 was used. The cells were then cultured in modified Barth’s solution at 18°C for 7 to 10 days before study.

Experiments were performed at room temperature. A single defolliculated cell was placed in a continuous-flow recording chamber (0.5-mL volume) and superfused with 2 mL/min Mg2+- and Ca2+-free Ringer’s solution (containing in mM: NaCl 96, KCl 2, BaCl2 1.8, and HEPES 10, pH adjusted to 7.5). 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 μm. Electrodes were filled with 3 M of KCl and tip resistances were 1–3 MΩ. The cell was voltage clamped by using a two-microelectrode oocyte voltage clamp amplifier (OC725A; Warner Corporation, 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-8A/D conversion board (Keithley-Metrabyte, Taunton, MA) and analysis was performed with a custom-written program (OoClamp software). All measurements were performed at a holding potential of −70 mV. Cells that did not show a stable holding current of <1 μA during a 5-min equilibration period (<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 compounds. Responses were quantified by measurement of steady-state current (cf.Fig. 1) and reported as μA.

Figure 1:
NR1/NR2A and NR1/NR2B glutamate receptors expressed recombinantly in Xenopus oocytes. Example traces of NMDA receptor responses induced by glutamate (Glu)/glycine (Gly) (half-maximal effect concentration [EC50]) in oocytes expressing NR1/NR2A (A) or NR1/NR2B receptors (B). C, Glutamate, in the presence of 10 μM of glycine, evokes inward currents in a concentration-dependent manner. EC50 is 3.2 ± 0.1 × 10 6 M for NR1/NR2A and 4.9 ± 0.7 × 10 6 M for NR1/NR2B receptors. D, Glycine, in the presence of glutamate at EC50, evokes inward currents in a concentration-dependent manner. EC50 is 1.5 ± 0.5 × 10 7 M for NR1/NR2A and 7.8 ± 3.5 × 10 8 M for NR1/NR2B receptors.

l-Glutamic acid and l-glycine, which act as agonist and obligatory coagonist on native receptors as well as on recombinantly expressed NMDA glutamate receptors (10), were diluted in Mg2+- and Ca2+-free Ringer’s solution to the required concentrations and were superfused (2 mL/min) over the oocyte for 30 s. Ketamine racemate, its isomers S(+)- and R(−)-ketamine and Mg2+ were also diluted in Mg2+- and Ca2+-free Ringer’s solution to the required concentrations, and oocytes were superfused (2 mL/min) for 3 min before agonist application.

Results were reported as mean ± sem. Because variability between batches of oocytes is common, responses were at times normalized to control responses. At least 10 oocytes were used for each data point, and oocytes from at least 2 frogs were studied for each experiment. Statistical tests used are indicated in Results. P < 0.05 was considered significant. Concentration-response curves were fit to the following logistic function, derived from the Hill equation: y = ymin + (ymax − ymin) (1 − xn/[x50n + xn]), where ymax and ymin are the maximal and minimal responses obtained, n is the Hill coefficient, and x50 is the half-maximal effect concentration (EC50 for agonist) or the half-maximal inhibitory effect concentration (IC50 for antagonist). Ninety-five percent confidence intervals for the isobologram were calculated from the sem of the IC50 values.

Isobolographic analysis is a nonmechanistic method of characterizing the effect resulting from the administration of two compounds, by using equieffective concentrations of individual drugs and combinations of these. The application of each drug alone is used to determine the isobolar points on the axes, (a,0) and (0,b). Pure “additivity” is represented by the isobole of additivity, which is based on the equation x/a + y/b = 1, and results in a straight line connecting the axial points (11,12). If the actual measured concentration of the combination of both drugs decreases below the isobolar plot of the 95% confidence interval, super-additivity is suggested. Although the method is a simplification of reality, in complex systems as the one under study (involving multiple binding sites, multiple agonists, and multiple antagonists), it is often preferred over modeling, because of the excessive number of assumptions required for the latter approach.

R is the reciprocal value of the sum of fraction, which is determined as follows:MATH

A sum of fraction <1 means super-additivity, whereas that >1 suggests subadditivity and a sum of fraction = 1 shows pure additivity.

Molecular biology reagents were obtained from Promega (Madison, WI) and other chemicals were obtained from Sigma (St. Louis, MO).


Functional Expression of NMDA Receptors in Xenopus Oocytes

Whereas uninjected oocytes were unresponsive to glutamate/glycine application, oocytes injected with NR1/NR2A or NR1/NR2B receptor cDNA responded to the application of glutamate (103–109 M in combination with glycine 105 M) with inward currents. These currents had the typical appearance described previously for NMDA receptor signaling in this model (13), and consisted of a rapid initial component that decayed over several seconds to a relatively stable steady-state level (Fig. 1A, B). Amplitude of the steady-state component was used as endpoint and determined at the time just before agonist washout. The shape of currents induced in oocytes expressing NR1/NR2B receptors was distinct from that evoked in those expressing NR1/NR2A (compare Fig. 1, A and B). The initial peak current typical for NR1/NR2A receptors was much less prominent in NR1/NR2B receptors. We did not investigate this further. To verify that the glutamate/glycine response was indeed mediated by glutamate receptors of the NMDA subtype, we tested the ability of the selective agonist NMDA to activate the receptors. In both NR1/NR2A- and NR1/NR2B-expressing oocytes, NMDA induced currents indistinguishable from those activated by glutamate (data not shown). Glutamate, as the physiologic agonist, was used for the remainder of the study.

Glutamate signaling was concentration-dependent (Fig. 1C, Table 1). EC50 values were not different between the subtypes (P = 0.76, t-test); Emax, however, was more for the NR1/NR2A subtype (P = 0.03), which is consistent with previous findings reported in this model (13,14).

Table 1:
Pharmacologic Variables Describing Interactions Between Ketamine and Mg2+

As anticipated, currents evoked by stimulation of the NMDA receptors with glycine in the presence of glutamate at EC50 were dependent on glycine concentration (Fig. 1D, Table 1). EC50s were not different between subtypes (P = 0.26, t-test), but Emax was significantly more for the NR1/NR2A subtype (P < 0.001, t-test). Responses obtained at large concentrations of agonist (glutamate concentration 103 M) or coagonist (glycine 0.1 mM) were similar (3.2 ± 0.6 and 2.9 ± 0.5, respectively, on NR1/NR2A; 2.1 ± 0.6 and 1.9 ± 0.4, respectively, on NR1/NR2B).

Together, these findings indicate that the receptors behave appropriately after recombinant expression.

Interactions of Mg2+ with NMDA Receptor Function

Mg2+(15) acts as an NMDA receptor blocker, and determines its effect, in addition to providing required data for subsequent combined Mg2+/ketamine studies; therefore, it also helped to confirm appropriate function of expressed receptors. We stimulated NMDA receptors with glutamate/glycine at EC50 (3.2 μM/150 nM for NR1/NR2A and 4.9 μM/78 nM for NR1/NR2B) in the presence of various concentrations of Mg2+ (102 to 106 M). Mg2+ was superfused for 3 min after a 5-min stabilization period. We observed a concentration-dependent inhibition of receptor responses (Fig. 2A). The inhibition curves for the two receptor subtypes were virtually superimposed, and curve fitting to the Hill equation revealed very similar IC50s for Mg2+ (Table 1). The largest Mg2+ concentration tested was 10 mM; at this concentration, NR1/NR2A and NR1/NR2B receptor responses were inhibited by 81.2% and 77%, respectively.

Figure 2:
Noncompetitive inhibition of NMDA receptor signaling by Mg2+. A, Mg2+ inhibits glutamate/glycine (half-maximal effect concentration)-induced NMDA responses in a concentration-dependent manner. Half-maximal inhibitory effect concentration is 4.2 ± 1.2 × 10-4 M for NR1/NR2A and 6.3 ± 2.4 × 10 4 M for NR1/NR2B receptors. B and C, Mg2+ acts as an allosteric antagonist. Mg2+ at half-maximal inhibitory effect concentration decreases maximal effect (Emax, indicated as horizontal lines) significantly from 3.4 ± 0.1 μA to 1.9 ± 0.2 μA at NR1/NR2A (B, P < 0.001) and from 2.5 ± 0.3 μA to 1.6 ± 0.1 μA at NR1/NR2B (C, P = 0.021), without affecting half-maximal effect concentration (3.2 ± 0.1 × 10 6 M versus 3.3 ± 0.3 × 10 6 M at NR1/NR2A (B) and 4.9 ± 0.7 × 10 6 M versus 3.7 ± 1.7 × 10 6 M at NR1/NR2B [C]).

In competition experiments (Fig. 2, B and C) with glutamate (109 M to 103 M), Mg2+ at IC50 (420 and 630 μM for NR1/NR2A and NR1/NR2B, respectively) decreased Emax significantly (Table 1) for both NR1/NR2A (P < 0.001, t-test) and NR1/NR2B (P = 0.021, t-test). In contrast, EC50 was unchanged for both NR1/NR2A (P = 0.921, t-test) and NR1/NR2B (P = 0.513, t-test). Thus, inhibition by Mg2+ is noncompetitive, and not significantly influenced by the NR2 subunits.

Interactions of Ketamine and Its Stereoisomers with NMDA Receptor Function

Ketamine inhibited functioning of NR1/NR2A and NR1/NR2B receptors, activated by glutamate/glycine (3.2 μM/150 nM for NR1/NR2A and 4.9 μM/78 nM for NR1/NR2B), to a similar degree (Fig. 3, A and B;Table 1), consistent with an action on the NR1 subunit. Compared with the isomers, racemic ketamine showed a nonsignificant, approximately two-fold more inhibitory potency than did R(−)-ketamine, but a significant (P = 0.007, one-way analysis of variance, Student-Newman-Keuls correction), nearly four-fold less potency than S(+)-ketamine. This is in agreement with data reported by other investigators (16).

Figure 3:
Noncompetitive inhibition of NMDA receptor signaling by ketamine. Ketamine racemate and its stereoisomers show a concentration-dependent inhibition of NR1/NR2A (A) and NR1/NR2B receptors (B). Calculated half-maximal inhibitory effect concentrations for NR1/NR2A (A) are 4.1 ± 2.5 × 10 6 M (S[+]-isomer), 25 ± 4.1 × 10 6 M (R[−]-isomer), and 14 ± 8.5 × 10 6 M (racemic mixture); for NR1/NR2B (B) 3.0 ± 0.3 × 10 6 M (S[+]-isomer), 26 ± 2.4 × 10 6 M (R[−]-isomer), and 18 ± 7.2 × 10 6 M (racemic mixture). Competition experiments with glutamate demonstrate that ketamine racemate, like Mg2+, inhibits NR1/NR2A (C) and NR1/NR2B (D) receptors in a noncompetitive manner. Emax is reduced significantly from 3.4 ± 0.1 μA to 1.6 ± 0.1 μA at NR1/NR2A (C) and from 2.5 ± 0.3 μA to 1.2 ± 0.1 μA at NR1/NR2B (D), without a significant shift in half-maximal effect concentration (3.2 ± 0.1 × 10 6 M versus 4.5 ± 3.0 × 10 6 M at NR1/NR2A [C] and 4.9 ± 0.7 × 10 6 M versus 2.8 ± 2.5 × 10 6 M at NR1/NR2B [D]).

In competition experiments with glutamate (Fig. 3, C and D;Table 1), racemic ketamine at IC50 (1.4 ± 0.9 × 105 M on NR1/NR2A and 1.8 ± 0.7 × 105 M on NR1/NR2B) was found to inhibit NR1/NR2A and NR1/NR2B receptors noncompetitively. It reduced Emax significantly (P < 0.001 for NR1/NR2A, t-test; and P = 0.001 for NR1/NR2B, t-test). In contrast, it did not shift EC50 significantly (P = 0.647 for NR1/NR2A, t-test; and P = 0.432 for NR1/NR2B, t-test).

Because S(+)-ketamine is the more relevant isomer, we used it for the remainder of this study and for the investigations reported in the companion article.

Combined Application of Mg2+ and S(+)-Ketamine Inhibits NMDA Signaling in a Super-Additive Manner

Ketamine and Mg2+ reduce postoperative morphine requirements more effectively when combined (8), and this interaction is hypothesized to take place at the NMDA receptor. To test this hypothesis, we investigated whether the combined administration of both compounds demonstrated super-additive inhibitory effects on NMDA signaling. We therefore first determined on both NR1/NR2A and NR1/NR2B receptors IC50 values for Mg2+ and S(+)-ketamine, and then for both compounds mixed in a ratio reflecting their IC50s (140:1 Mg (2+)/ketamine). To minimize experimental variability, each set of IC50s was obtained in oocytes from the same frog. EC50 values for S(+)-ketamine and Mg2+ are therefore slightly (and nonsignificantly) different from those reported above, which were obtained in separate experiments. Glutamate/glycine (3.2 μM/150 nM for NR1/NR2A and 4.9 μM/78 nM for NR1/NR2B) was used as agonist (Fig. 4, A and B;Table 1). When compounds were studied in combination on NR1/NR2A receptors, IC50 was decreased significantly for Mg2+ (P < 0.001, t-test, a 97% decrease) and S(+)-ketamine (P < 0.001, t-test, a 97% decrease). Similar results were obtained for NR1/NR2B: IC50 was reduced significantly for Mg2+ (P = 0.003, t-test, a 96% decrease) and S(+)-ketamine (P < 0.001, t-test, a 94% decrease) when compounds were studied in combination.

Figure 4:
Super-additive interactions between ketamine and Mg2+ at NMDA receptors. Mg2+ and S(+)-ketamine induce super-additive inhibition of NR1/NR2A (A) and NR1/NR2B (B) receptors. The half-maximal inhibitory effect concentrations (IC50s) for NR1/NR2A (A) are 606 ± 105 × 10 6 M (Mg2+) and 4.0 ± 0.2 × 10 6 M (S[+]-ketamine) when studied in isolation versus 17 ± 1.7 × 10 6 M (Mg2+) and 0.12 ± 0.01 × 10 6 M (S[+]-ketamine) for the combination. The IC50s for NR1/NR2B (B) are 602 ± 148 × 10 6 M (Mg2+) and 2.9 ± 0.3 × 10 6 M (S[+]-ketamine) in isolation versus 27 ± 1.8 × 10 6 M (Mg2+) and 0.19 ± 0.01 × 10 6 M (S[+]-ketamine) for the combination. C and D, Isobolographic plots confirm the super-additive effects of Mg2+/S(+)-ketamine combination on NR1/NR2A (C) and NR1/NR2B receptors (D). Measured IC50s for the combinations are outside the 95% confidence interval (sum of fraction 0.057, NR1/NR2A [C] and sum of fraction 0.108, NR1/NR2B [D]) for purely additive action, indicating a super-additive effect of the compounds. R (C and D) is the reciprocal of sum of fraction.

To make these differences more clear, we performed isobolographic analysis (Fig. 4, C and D). Because measured IC50s for the combinations are outside the 95% confidence interval for purely additive action (sum of fraction for NR1/NR2A: 0.057; sum of fraction for NR1/NR2B: 0.108), our hypothesis that S(+)-ketamine and Mg2+ exhibit super-additive interactions at recombinantly expressed NR1/NR2A and NR1/NR2B NMDA glutamate receptors was confirmed.


Our findings show that ketamine and Mg2+ inhibit responses of recombinantly expressed NR1/NR2A and NR1/NR2B glutamate receptors, and that combinations of the compounds act in a super-additive manner. These findings may explain, in part, why combinations of ketamine and Mg2+ are more effective analgesics than either compound alone. We previously studied patients undergoing abdominal hysterectomy under general anesthesia and pretreated them with placebo, ketamine (0.15 mg/kg, IV), MgSO4 (30 mg/kg, IV), or both compounds. Whereas postoperative pain scores were similar in all groups, requirements for morphine were decreased by either ketamine or Mg2+ (to approximately 50% of control), and decreased much more (to approximately 20% of control) in patients receiving both drugs (8).

We chose NR1/NR2A and NR1/NR2B combinations for study. In adult mouse brain preparations, NR1, NR2A, and NR2B subunits are widely distributed in the brain, whereas NR2C is found predominantly in the cerebellum and NR2D is detected only in small levels in the thalamus and brainstem (17). Thus, the NR1/NR2A and NR1/NR2B combinations are probably most relevant physiologically. Each of the recombinantly expressed, heteromeric NR1/NR2 subunit receptor combinations has distinctive subunit-dependent biophysical and pharmacological signatures, including affinities for specific agonists and antagonists (18–21). In this context, our data are consistent with previous findings showing that channel blockers like ketamine and Mg2+ did not differentiate between NR1/NR2A and NR1/NR2B combinations expressed in oocytes (22).

In our study, we measured steady-state rather than peak currents for several reasons: 1) peak currents may be missed by slow agonist equilibration, 2) limits to voltage control in the oocyte might make peak current measurements inaccurate, 3) measuring steady-state currents allows sufficient time for open-channel block by Mg2+ and ketamine, 4) there is concern that after oocyte expression the peak current may be contaminated with Ca2+-activated Cl current (23). We used Ba2+ as charge carrier to minimize this possibility, but in addition felt safer not using the peak current for analysis. 5) NR1/NR2B receptors did not show appreciable peak current in our model.

Several potential limitations of our model should be considered. First, the oocyte membrane contains an endogenous Ca2+-activated Cl channel, and it is conceivable that this channel was activated by bivalent cation influx during NMDA receptor activation. For this reason, we used Ba2+ rather than Ca2+ as our charge carrier, because Ba2+ has little, if any, effect on the Ca2+-activated Cl channel (23). Second, our experiments were performed at room temperature, whereas the expressed receptor is derived from a homeothermic animal (rat, which is 99% homologous for the NR1 subunit and 95% for the NR2A subunit with the human receptor) (24). Although in theory this might influence its behavior, we thought it more important to keep the cell membrane in its normal state of fluidity. Other investigators have expressed NMDA receptors in oocytes, and generally have found them to function appropriately (23). However, some differences between channels expressed in oocytes and those expressed in mammalian cells have been reported, particularly when NR1 subunits are expressed alone (15). Third, we coexpressed NR1 subunits with NR2 subunits to enhance currents through expressed receptors and to provide a more normal receptor configuration (19). However, we could not determine expression levels or stoichiometry for the subunit combinations. Thus, it is conceivable that the ratio of NR1 to NR2 varied during experimentation. To minimize formation of homomeric NR1 receptors, we used a cDNA weight ratio of 1:5 between NR1 and NR2 subunits.

Despite these caveats, we feel that our expression system provides a reasonable model of NMDA receptor expression in a neuron, and this was confirmed by our glutamate and glycine concentration-response studies, effectiveness of NMDA in activating the receptors, and noncompetitive inhibition by ketamine and Mg2+.

It should also be noted that, because of study design, concentrations of agonists and inhibitors in our study may be different from those attained in vivo. Our main interest focuses on the pharmacological aspect of NMDA receptor functioning. Therefore, we used EC50 or IC50, rather than concentrations occurring in the physiologic setting. Glutamate concentrations attained in the synaptic cleft may be large enough to saturate glutamate receptors, but are only present for very short periods. Baseline glycine concentrations are saturating in vivo. (25) However, because the effects of ketamine and Mg2+ were noncompetitive, this should not influence interpretation of our results.

Ketamine plasma concentrations are approximately 108 ng/mL (360 nM) after an IV bolus of 0.25 mg/kg (26). This is similar to the IC50 we determined when combined with Mg2+, suggesting that a significant NMDA blocking effect can be obtained with small ketamine doses in vivo, especially in combination with Mg2+. A comparison of Mg2+ concentrations in plasma (0.76–0.96 mM) (27) or cerebrospinal fluid (0.95–1.13 mM) (27)in vivo with the IC50 determined in our experiments suggests that, in nondepolarized cells, NMDA receptors are largely blocked by Mg2+. However, this will not necessarily be true in depolarized cells, because Mg2+ block is voltage-dependent.

The IC50 for Mg2+ as determined in our study is comparable with that obtained for Mg2+ block of NMDA-evoked acetylcholine release in neurons (28), although other investigators, using the Xenopus oocyte model, have reported an increased sensitivity of NMDA receptors to Mg2+(29). The IC50 for ketamine on NMDA receptors as obtained in our study is somewhat more than that reported by other investigators using the same model (0.4–1 μM) (29). Indeed, values obtained in our study are closer to those determined by patch clamp experiments in cultured neurons (30) or using in vitro CNS preparations (31).

We observed super-additive interactions of ketamine and Mg2+ at the NMDA receptor. This is in contrast to a previous study by Harrison and Simmonds (32), who reported additive interactions between the compounds. This discrepancy likely results from the different models used. Harrison and Simmonds used a rather indirect measurement of NMDA signaling, stimulating in one part of the brain and observing very slow and long-lasting electrophysiologic effects in another part. The investigators almost certainly stimulated multiple NMDA receptor subtypes at the same time. Thus, it is not surprising that their findings cannot be immediately reconciled with ours. Several whole-animal studies of ketamine-Mg2+ interaction have been performed, with divergent results. In Mg2+-deficient rats, subadditive interactions were shown when ketamine sensitivity was determined (33), whereas in mice, Mg2+ potentiated ketamine anesthesia in a super-additive manner (34).

In summary, ketamine and Mg2+ exert several beneficial effects because of their inhibition of NMDA signaling. However, the administration of these drugs in the clinical setting is associated with dose-dependent side effects. Our data suggest that the combined administration of the compounds, with resulting super-additive interactions at the NMDA receptor, may allow more profound clinical effects to be obtained without exceeding safe doses. This may be useful in prevention or treatment of pain in several settings. In addition, ketamine—and possibly Mg2+—is a neuroprotectant. Mg2+ protects cerebellar neurons against glutamate toxicity (35) and ketamine is under intense investigation for its protective properties. Thus, it might be appropriate to study the effects of combinations of these drugs in the setting of neuronal compromise.

We thank Prof. Dr. med. Eike Martin (Ruprecht-Karls-Universität Heidelberg, Germany) and Prof. Dr. med. Hugo Van Aken (Westfälische-Wilhelms-Universität Muenster, Germany) for their support. Sincere thanks to Dagmar Westerling, MD, PhD, for a preliminary analysis of the results.


1. Johnson JW, Ascher P. Glycine potentiates the NMDA response in cultured mouse brain neurons. Nature 1987; 325: 529–31.
2. Lipton SA, Rosenberg PA. Excitatory amino acids as a final common pathway for neurologic disorders. N Engl J Med 1994; 330: 613–22.
3. Woolf CJ, Thompson WN. The induction and maintenance of central sensitization is dependent on N-methyl- d -aspartate acid receptor activation: implications for the treatment of post-injury pain hypersensitivity states. Pain 1991; 44: 293–9.
4. Haley JE, Sullivan AF, Dickenson AH. Evidence for spinal N-methyl- d -aspartate receptor involvement in prolonged chemical nociception in the rat. Brain Res 1990; 518: 218–26.
5. Davies SN, Lodge D. Evidence for involvement of N- methylaspartate receptors in ‘wind-up’ of class 2 neurones in the dorsal horn of the rat. Brain Res 1987; 424: 402–6.
6. Mathisen LC, Skjelbred P, Skoglund LA, Oye I. Effect of ketamine, an NMDA receptor inhibitor, in acute and chronic orofacial pain. Pain 1995; 61: 215–20.
7. Lodge D, Johnson KM. Noncompetitive excitatory amino acid receptor antagonists. Trends Pharmacol Sci 1990; 11: 81–6.
8. Lo B, Hoenemann CW, Durieux ME. Preemptive analgesia: ketamine and magnesium reduce postoperative morphine requirements after abdominal hysterectomy [abstract]. Anesthesiology 1998; 89: A1163.
9. Nietgen GW, Chan CK, Durieux ME. Inhibition of lysophosphatidate signaling by lidocaine and bupivacaine. Anesthesiology 1997; 86: 1112–9.
10. Williams K, Chao J, Kashiwagi K, et al. Activation of N-methyl- d -aspartate receptors by glycine: role of an aspartate residue in the M3-M4 loop of the NR1 subunit. Mol Pharmacol 1996; 50: 701–8.
11. Tallarida RJ, Porreca F, Cowan A. Statistical analysis of drug-drug and site-site interactions with isobolograms. Life Sci 1989; 45: 947–61.
12. Tallarida RJ. Statistical analysis of drug combinations for synergism. Pain 1992; 49: 93–7.
13. Verdoorn TA, Kleckner NW, Dingledine R. Rat brain N-methyl- d -aspartate receptors expressed in Xenopus oocytes. Science 1987; 238: 1114–6.
14. Stone TW. CNS neurotransmitters and neuromodulators: glutamate. Boca Raton, FL: CRC Press, 1995: 98–104.
15. Sucher NJ, Awobuluyi M, Choi Y-B, Lipton SA. NMDA receptors: from genes to channels. Trends Pharmacol Sci 1996; 17: 348–55.
16. Ebert B, Mikkelsen S, Thorkildsen C, Borgbjerg FM. Norketamine, the main metabolite of ketamine, is a non-competitive NMDA receptor antagonist in the rat cortex and spinal cord. Eur J Pharmacol 1997; 333: 99–104.
17. Mori H, Mishina M. Structure and function of the NMDA receptor channel. Neuropharmacology 1995; 34: 1219–37.
18. Moriyoshi K, Masu M, Ishii T, et al. Molecular cloning and characterization of the rat NMDA receptor. Nature 1991; 354: 31–7.
19. Monyer H, Sprengel R, Schoepfer R, et al. Heteromeric NMDA receptors: molecular and functional distinction of subtypes. Science 1992; 256: 1217–21.
20. Buller AL, Larson HC, Schneider BE, et al. The molecular basis of NMDA receptor subtypes: native receptor diversity is predicted by subunit composition. J Neurosci 1994; 14: 5471–84.
21. Yamazaki M, Mori H, Araki K, et al. Cloning, expression and modulation of a mouse NMDA receptor subunit. FEBS Lett 1992; 300: 39–45.
22. Yamakura T, Mori H, Masaki H, et al. Different sensitivities of NMDA receptor channel subtypes to non-competitive antagonists. Neuroreport 1993; 4: 687–90.
23. Leonard JP, Kelso SR. Apparent desensitization of NMDA responses in Xenopus oocytes involves calcium-dependent chloride current. Neuron 1990; 4: 53–60.
24. Le BB, Wafford KA, Kemp JA, et al. Cloning, functional coexpression, and pharmacological characterisation of human cDNAs encoding NMDA receptor NR1 and NR2A subunits. J Neurochem 1994; 62: 2091–8.
25. Huettner JE. Competitive antagonism of glycine at the N-methyl- d -aspartate (NMDA) receptor. Biochem Pharmacol 1991; 41: 9–16.
26. Ireland SJ, Livingston A. Effect of biliary excretion on ketamine anaesthesia in the rat. Br J Anaesth 1980; 52: 23–8.
27. Fuchs-Buder T, Tramer MR, Tassonyi E. Cerebrospinal fluid passage of intravenous magnesium sulfate in neurosurgical patients. J Neurosurg Anesthesiol 1997; 9: 324–8.
28. Nicolas C, Fage D, Carter C. NMDA receptors with different sensitivities to magnesium and ifenprodil control the release of [14C]acetylcholine and [3H]spermidine from rat striatal slices in vitro. J Neurochem 1994; 62: 1835–9.
29. Ferrer-Montiel AV, Merino JM, Planells-Cases R, et al. Structural determinants of the blocker binding site in glutamate and NMDA receptor channels. Neuropharmacology 1998; 37: 139–47.
30. MacDonald JF, Milikovic Z, Pennefaher P. Use-dependent block of excitatory amino acid currents in cultured neurons by ketamine. J Neurophysiol 1987; 58: 251–66.
31. Yamamura T, Harada K, Okamura A, Kemmotsu O. Is the site of action of ketamine anesthesia the N-methyl- d -aspartate receptor? Anesthesiology 1990; 72: 704–10.
32. Harrison NL, Simmonds MA. Quantitative studies on some antagonists of N-methyl- d -aspartate in slices of rat cerebral cortex. Br J Pharmacol 1985; 84: 381–91.
33. Orser B, Smith D, Henderson S, Gelb A. Magnesium deficiency increases ketamine sensitivity in rats. Can J Anaesth 1997; 44: 883–90.
34. Irifune M, Shimizu T, Nomoto M, Fukuda T. Ketamine-induced anesthesia involves the N-methyl- d -aspartate receptor-channel complex in mice. Brain Res 1992; 569: 1–9.
35. Cox JA, Lysko PG, Henneberry RC. Excitatory amino acid neurotoxicity at the N-methyl- d -aspartate receptor in cultured neurons: role of the voltage-dependent magnesium block. Brain Res 1989; 499: 267–72.
© 2001 International Anesthesia Research Society