N-Methyl-d-Aspartate Receptor Antagonists Potentiate the Antinociceptive Effects of Morphine in Squirrel Monkeys

Materials and Methods

Animals.

Five adult male squirrel monkeys (Saimiri sciureus) weighing between 0.70 and 0.95 kg were housed individually or in pairs in a colony room with a 12-h light/dark cycle. All monkeys had continuous access to water and were maintained on a high-protein monkey diet and given fresh fruit and nuts daily. All of the monkeys had previous experience with the shock titration procedure and had received various opioid compounds, but had not received drugs for at least 30 days before the start of the present experiment.

Apparatus.

During experimental sessions, each monkey sat in a Plexiglas chair and was held in place by a waist support with its tail secured by a small stock (Dykstra, 1985). The tail was coated with a noncorrosive electrode paste (EKG Sol; Graphic Controls Corporation, Buffalo, NY) to provide a low resistance electrical contact. Electric shock (110 V AC, 60 Hz) was delivered through two hinged brass plates that rested on a shaved portion of the tail.

Each chair was enclosed within a ventilated sound-attenuating chamber and was illuminated by a 10-W white houselight during experimental sessions. A lever was mounted on the right side of the front panel, 8.5 cm above the waist plate and 4.0 cm from the right side wall. During experimental sessions, presses on the lever with a downward force of 0.15 N produced an audible click and were recorded as responses. White noise was presented continuously both inside the chamber and throughout the experimental room. Experimental events, including control of shock intensity, were controlled using software and hardware from Med Associates (St. Albans, VT) through a microcomputer located in the adjacent room.

Behavioral Procedure.

A shock titration procedure nearly identical to that described by Dykstra and Massie (1988) was used. In each session, periods during which an FR 5 schedule of shock titration was in effect alternated with periods of blackout. Each FR 5 titration period began with the illumination of the houselight and presentation of 0.01 mA shock. Shock intensity increased from 0.01 to 2.0 mA in 30 increments. Completion of the FR 5 requirement at a given shock intensity initiated a timeout during which shock was off and the houselight remained illuminated. After the 15-s timeout the shock resumed at the next lower intensity. If a monkey failed to complete the FR 5 during 15 s at a given shock intensity, the intensity increased by one increment and the response requirement was reset to 5. The FR 5 titration periods usually lasted 15 min. An FR 5 period terminated automatically, however, if the shock intensity rose to the peak intensity of 2.0 mA and the FR 5 requirement was not completed during any of five consecutive 15-s periods. During the blackouts that separated the FR 5 titration periods, the chamber was dark, no shock was delivered, and lever presses had no programmed consequences. Blackouts lasted 20 min. Each session began with an FR 5 titration period and ended after completion of five FR 5 periods.

Pharmacological Procedure.

Dose-effect curves for dizocilpine and LY235959 were first obtained using a cumulative dosing procedure. Under this procedure, vehicle (saline) was injected 20 min before the first FR 5 period. On completion of the first FR 5 period (at the onset of the first blackout), the lowest dose was injected and its effects were assessed in the following FR period. At the onset of each subsequent blackout period, an amount of drug that increased the cumulative dose by 0.5 log unit was injected. Injections continued in this manner for three or four dose increments.

Time-effect curves for dizocilpine, LY235959, (+)-HA-966, morphine, and various morphine/NMDA receptor antagonist combinations were obtained by administering the dose or dose combination to monkeys 20 min before the first FR period. On completion of the first and all subsequent FR 5 periods (at the onset of each blackout), vehicle (saline) was injected.

Combinations of dizocilpine with morphine used both the cumulative dosing and time course procedures described above. For these dose combinations, saline or a single dose of morphine (0.3, 1.0, or 1.7 mg/kg) was administered 20 min before the start of the first FR 5 period. On completion of the first FR 5 period (at the onset of the first blackout), saline was injected and its effects were assessed in the following FR period. At the onset of each subsequent blackout period (prior to components 3, 4, and 5), dizocilpine (0.003–0.03 mg/kg) was injected using the cumulative dosing procedure described above.

The behavior of monkeys was assessed daily, Monday through Friday. Drugs were generally administered on Tuesdays and Fridays. The behavior of monkeys following saline vehicle injections was determined periodically throughout the course of the study.

Morphine sulfate and (+)-HA-966 were generously provided by the National Institute on Drug Abuse and LY235959 by Lilly Research Laboratories (Indianapolis, IN). Dizocilpine (MK-801) was purchased from Research Biochemicals, Inc. (Natick, MA). All drugs were dissolved in sterile saline and injected intramuscularly into the calf.

Data Analysis.

There were two dependent variables extensively analyzed in this study: MSL (mA) and response rate during shock (RR, responses/s). Response rate during timeout, a third dependent variable, was only analyzed for data from monkeys that received vehicle control injections for all five components of an experimental session.

Because the same monkeys received all drug doses within an experiment, and because the drug effects were measured over time in the same monkeys, a repeated measures analysis of variance (ANOVA) was used to analyze all of the data in this study. Both dose and time were within-subjects variables. When the data from a time course experiment were analyzed [saline control injections, morphine alone, NMDA antagonists alone, morphine plus LY235959, morphine plus (+)-HA-966], raw data were entered into the statistical analysis. When data from an experiment that used cumulative dosing were analyzed (dizocilpine and LY235959 dose-effect curves, morphine plus dizocilpine), data entered into the analysis were difference-scored. Difference scores were calculated as the difference between drug treatment and saline control injection (for drug combinations, the difference between morphine plus dizocilpine and morphine alone) for the appropriate component. This was done for two reasons. First, there were small increases in MSL over time when monkeys received saline control injections. Although this effect was small (a difference of 0.04 mA, compared with a total possible increase of 1.95 mA), it was statistically significant. Analysis of difference scores for a treatment relative to the effect of saline at that time point accounted for this small effect. Second, the analysis of difference scores simplified the analysis by replacing two within-subject variables (dose and time) with one.

Saline and Morphine Control Data.

For each monkey, data from the 7 most recent saline control days were averaged and the average for a monkey was used in the statistical analysis. For morphine/LY235959 and morphine/(+)-HA-966 combinations, the effects of morphine alone were assessed before and after each dose combination. Thus there were six morphine-alone curves for each monkey in each of these experiments. Since there was little variability in the effectiveness of morphine for a given monkey, these morphine curves were averaged and the average for a monkey represented that monkey in the statistical analysis. The results of the statistical analysis did not differ, however, when only one morphine-alone curve was used in place of the averaged morphine curve (data not shown).

Contrasts.

When a post hoc analysis of a factor in the repeated-measures ANOVA was conducted, the analysis made the following comparisons: for time, all time points versus time 1; for multiple doses of a drug tested alone, all doses versus saline control; and for dose combinations, all combinations versus the effect of morphine or LY235959 alone.

All data were analyzed using SAS for Windows version 6.1 (SAS Institute, Cary, NC).

Results

Control Performance.

To examine the effects of saline injections on MSL, a repeated-measures ANOVA was performed using time as the within-subject variable. Small increases in MSL over time were observed when monkeys were injected with saline. Post hoc analysis of the repeated-measures ANOVA main effect of time (F4,16 = 9.68, P = 0.0004) revealed a trend toward significance, with MSL at component 5 significantly higher than MSL at component 1 (0.09 versus 0.05 mA, F1,4 = 12.25,P = 0.0249). However, this difference in mean MSL is small relative to the magnitude of increase possible in this procedure.

To compare response rates during shock and during the postshock timeouts, a repeated-measures ANOVA was performed using time and shock condition (shock or postshock timeout) as within-subject variables. The ANOVA revealed a significant main effect of shock condition (F1,4 = 10.17, P = 0.0333). Response rates during shock were significantly higher than response rates during the postshock timeouts (data not shown). The ANOVA did not reveal a significant main effect of time (F4,16 = 1.39,P = 0.2814) or a significant shock condition by time interaction (F4,16 = 1.52, P = 0.2439).

Effects of Morphine on MSL and RR.

Figure1 shows the time course of the antinociceptive effects of morphine in five monkeys. Morphine (0.3, 1.0, 1.7, and 3.0 mg/kg) dose- and time-dependently increased MSL in all five monkeys tested. Average MSLs are presented in Table1. Modest increases in MSL occurred following administration of 1.7 mg/kg morphine, whereas 3.0 mg/kg morphine produced maximal increases in all five monkeys tested. For three of the five monkeys that received 3.0 mg/kg morphine, the session was terminated after the second component, and naltrexone was administered to prevent possible adverse effects of this high dose of morphine. The remaining two monkeys were injected with naltrexone following the fourth and fifth components. As a result, data from the 3.0 mg/kg dose of morphine was not included in the statistical analysis. The results of a repeated-measures ANOVA, presented in Table2, revealed a main effect of dose, a main effect of time, but no dose × time interaction.

The effects of morphine on RR are also presented in Fig. 1 and Table 1. Only the 3.0 mg/kg dose of morphine decreased response rates. The results of a repeated-measures ANOVA for RR, presented in Table 2, did not reveal a main effect of time, dose, or a dose × time interaction when 3.0 mg/kg morphine was excluded from the analysis.

Effects of NMDA Receptor Antagonists Alone on MSL and RR: Cumulative Dosing.

The effects of cumulatively administered dizocilpine and LY235959 on MSL are presented in Fig.2. Only the highest dose of dizocilpine (0.1 mg/kg) increased MSL relative to saline. The average MSL following a cumulative dose of 0.1 mg/kg dizocilpine was 0.42 (0.11) mA. This increase in MSL is greater than the increase in MSL over time observed when monkeys were injected with saline. The effects of each cumulative dose of dizocilpine were subtracted from the effects of saline control injections for the equivalent time point, and these difference scores were subjected to a repeated-measures ANOVA. The repeated-measures ANOVA for these difference scores was significant (F3,12 = 9.85, P = 0.0015). Post hoc contrasts revealed that only the difference score for 0.1 mg/kg dizocilpine and saline at time 4 was significantly different from the difference score for saline and saline at time 1 (F1,4 = 9.83, P = 0.0350). In contrast to the increase in MSL observed following injections with dizocilpine, no dose of LY235959 tested increased MSL relative to saline. In addition, neither dizocilpine nor LY235959 altered RR when administered cumulatively to monkeys in the shock titration procedure.

Effects of NMDA Receptor Antagonists Alone on MSL and RR: Time Course Analysis.

The effects of various doses of dizocilpine (0.03 mg/kg) and LY235959 (0.1, 0.3, and 1.0 mg/kg) on MSL and RR were measured over time, as were the effects of (+)-HA-966 (10, 30, and 56 mg/kg). Table 2 presents the results of repeated-measures analyses performed for each drug and for each measure (MSL and RR). The effects of each of these three NMDA receptor antagonists on MSL and RR are also presented in Table 1. Dizocilpine, LY235959, and (+)-HA-966 did not significantly increase MSL or alter RR relative to saline control injections.

Effects of Dizocilpine in Combination with Morphine on MSL and RR.

Figure 3 shows the antinociceptive effect of morphine alone and in combination with dizocilpine. Doses of morphine from 0.3 to 1.7 mg/kg produced either no antinociceptive effect or a modest increase in antinociception when administered alone. For example, the mean MSL for monkeys treated with 1.7 mg/kg morphine was 0.14, 0.21, 0.30, 0.30, and 0.36 mA in components 1, 2, 3, 4, and 5, respectively. Dizocilpine (0.003–0.03 mg/kg) dose-dependently potentiated the modest antinociceptive effect of morphine (Fig. 3). The mean MSL for monkeys treated with 1.7 mg/kg morphine and 0.03 mg/kg dizocilpine was 1.14 (0.25) mA. A repeated-measures ANOVA was performed on MSL difference scores (morphine plus dizocilpine–morphine alone), and the results of that analysis are presented in Table 3. There was a significant morphine dose × dizocilpine dose interaction (F9,36 = 4.50, P = 0.0005). In contrast to the effects on MSL, no dose of morphine alone or in combination with dizocilpine decreased the mean rate of responding (Fig. 3, Table 3), demonstrating that the potentiation of the effect of morphine on MSL is not due to motor impairment.

Although the mean MSL for 1.7 mg/kg morphine was increased in all monkeys when combined with 0.03 mg/kg dizocilpine, the increase was modest in two monkeys. The difference scores for these two monkeys (0.15 mA for M290 and 0.55 mA for M540) were low compared with the other three monkeys (0.70, 1.10, and 1.40 mA). Furthermore, MSLs following this drug combination in these two monkeys (M290 and M540) were less than 1.0 mA, whereas the three other monkeys had MSLs that were greater than 1.3 mA and approached maximal antinociceptive effectiveness in this procedure. To begin to address whether these individual differences were related to time course variables not accounted for with the cumulative dosing procedure, monkeys M290 and M540 were administered 1.7 mg/kg morphine in combination with 0.03 mg/kg dizocilpine 20 min before the start of the first session. The effect of these drugs on MSL and RR was measured over the five components of the experimental session. Both monkeys showed substantial increases in MSL following administration of 1.7 mg/kg morphine with 0.03 mg/kg dizocilpine, but these increases were not apparent until the second component (60–70 min after administration). The MSLs for monkey M290 were 0.10, 1.40, and 1.80 for components 1, 2, and 3, respectively, after which the experiment was terminated. For monkey M540, MSL increased in a similar manner from 0.25 mA in the first component to 1.20, 1.50, 1.20, and 1.20 mA in components 2, 3, 4, and 5. Although monkey M290 did not respond during the third component, both monkeys responded within the range of control response rates during the second component, and monkey M540 responded normally throughout five components of the session. Following these results, cumulative dosing was abandoned and time course analyses were conducted for all subsequent morphine/NMDA antagonist combinations.

Effects of LY235959 in Combination with Morphine on MSL and RR.

Figure 4 shows the effects of the competitive NMDA receptor antagonist LY235959 on the antinociceptive effect of 1.0 mg/kg morphine. When administered alone 20 min before the start of the first component, 1.0 mg/kg morphine produced minimal increases in MSL. For example, the mean MSL for monkeys administered 1.0 mg/kg morphine was 0.06, 0.11, 0.14, 0.16, and 0.19 in components 1, 2, 3, 4, and 5, respectively (Table4). LY235959, when administered with morphine 20 min before the start of the first component, dose-dependently increased MSL in all five monkeys tested for the duration of the experimental session. For example, the mean MSL values at component 5 (165–175 min postinjection) were 0.19, 0.19, 0.70, and 1.23 for morphine alone, morphine plus 0.1 mg/kg LY235959, morphine plus 0.3 mg/kg LY235959, and morphine plus 1.0 mg/kg LY235959, respectively. The results of a repeated-measures ANOVA of the MSL values are presented in Table 3. There were highly significant main effects of the LY235959 dose (0, 0.1, 0.3, and 1.0), time (components 1, 2, 3, 4, and 5), and a highly significant dose × time interaction. In contrast, response rate was unchanged over the five components of the experimental session following administration of 1.0 mg/kg morphine alone or in combination with LY235959 (Fig. 4, Table 3, and Table 4). These data demonstrate that the potentiation of the antinociceptive effect of morphine is not due to motor impairment.

Figure 5 shows the effects of 0.3 mg/kg LY235959 administered alone and in combination with 0.3, 1.0, and 1.7 mg/kg morphine is a subset of monkeys (n = 3). The results of a repeated-measures ANOVA are presented in Table 3. These data demonstrate that the interaction between morphine and LY235959 is dependent on both the dose of LY235959 (Fig. 4) and the dose of morphine (Fig. 5). Once again, response rate was not altered in monkeys that received 0.3 mg/kg LY235959 alone or in combination with morphine, demonstrating that this antinociceptive effect is not due to motor impairment.

Effects of (+)-HA-966 in Combination with Morphine on MSL and RR.

A subset of monkeys (n = 3) were administered 1.0 mg/kg morphine alone or in combination with the glycine-site antagonist (+)-HA-966. Figure 6 shows that (+)-HA-966 dose- and time-dependently increased the antinociceptive effect of 1.0 mg/kg morphine. For example, the mean MSL values following administration of 1.0 mg/kg morphine alone were 0.05, 0.09, 0.11, 0.13, and 0.14 mA during components 1, 2, 3, 4, and 5. Following 1.0 mg/kg morphine plus 56 mg/kg (+)-HA-966, mean MSL values were 0.08, 0.47, 0.80, 1.07, and 1.07 during components 1, 2, 3, 4, and 5 (Table 4). The mean MSL values at time 5 following morphine alone, morphine plus 10 mg/kg (+)-HA-966, morphine plus 30 mg/kg (+)-HA-966, and morphine plus 56 mg/kg (+)-HA-966 were 0.14, 0.09, 0.37, and 1.07 mA, respectively. The results of the repeated-measures ANOVA revealed a significant (+)-HA-966 dose × time interaction when MSL was the dependent variable; however, as with the noncompetitive antagonist, dizocilpine, and the competitive antagonist, LY235959, there was no such interaction when response rate was the dependent measure (Table3).

Discussion

In this study, the acute administration of morphine produced a maximal increase in MSL that was dose- and time-dependent. Dizocilpine, LY235959, and (+)-HA-966 produced little to no increase in MSL across the range of doses tested. However, when these drugs were combined with doses of morphine that alone produced little or no increase in MSL, MSL was greatly increased relative to treatment with morphine alone. Several other laboratories have demonstrated similar effects in both rodent models of antinociception (Mao et al., 1996; Lutfy et al., 1999) and in human experimental and clinical pain (Sethna et al., 1998;Caruso, 2000; Katz, 2000).

The NMDA receptor antagonists used in this study represented several chemically and functionally distinct classes of drugs. We chose to investigate a range of classes of NMDA receptor antagonist rather than multiple examples of a single class to provide convergent evidence for a role of the NMDA receptor in any observed effect. Because an NMDA receptor channel blocker (dizocilpine), a competitive NMDA receptor antagonist (LY235959), and a glycine-site antagonist [(+)-HA-966] all effectively increased MSL relative to morphine alone, this suggests that blockading activity at the NMDA receptor plays an important role in the potentiation of the antinociceptive effect of morphine in this procedure.

A second major finding from this study is that the increases in MSL observed following morphine/NMDA receptor antagonist combinations occurred without disruptions in RR during shock. That is, following a morphine/NMDA receptor antagonist combination, monkeys actively titrated the shock intensity at significantly higher mA values than following treatment with saline or morphine alone, yet the rate at which they responded on the lever to titrate the shock was no different than following treatment with saline or morphine alone. Thus, the increases in MSL that followed morphine/NMDA receptor antagonist combinations cannot be attributed to the effect of these drugs or drug combinations on motor performance.

One strength of the shock titration procedure is its ability to differentiate the antinociceptive effect of a drug from changes in motor function that can confound the interpretation of latency increases in other antinociception assays. Traditionally, inferences about the contribution of motor effects within antinociception assays have been drawn from analyses of potency differences. For example, the noncompetitive NMDA receptor antagonist dizocilpine was equipotent in increasing squirrel monkey tail-withdrawal latencies from warm water and impairing motor function as assessed with an observer-scored rating scale (Rupniak et al., 1993). In contrast, France et al. (1989)demonstrated maximal increases in tail-withdrawal latencies using a warm water tail-withdrawal procedure when rhesus monkeys were tested following administration of several noncompetitive NMDA receptor antagonists. Also, the doses for antinociception were 3- to 10-fold lower than the doses for anesthesia. Unfortunately, these data do not critically test the hypothesis that motor effects are causally related to antinociceptive efficacy. For example, it is not known to what extent the ataxia produced by dizocilpine in the Rupniak study was related to the increases observed in squirrel monkey tail-withdrawal latencies. Similarly, potency differences in anesthetic and analgesic effects only demonstrate that increases in tail-withdrawal latency are not due to anesthesia, but leave open the possibility that other motor effects may be causally related to the increase in tail-withdrawal latencies. In the shock titration procedure, however, RR is inextricably linked to the antinociceptive measure MSL, and inference regarding the relationship between potentially unrelated motor tasks and the measure of antinociception is not confounded.

It is important to note that motor impairments were sometimes observed in monkeys that received selected morphine-NMDA antagonist combinations. These effects can be described as periods of inactivity and difficulty maneuvering and maintaining balance on a perch. Still, data from the titration procedure clearly indicate that response rate was unaltered under these conditions.

Here, however, we have a separation of the goals of theoretical basic science research and those of clinical practice for the treatment of pain. Although these data clearly demonstrate that NMDA receptor antagonists can potentiate morphine analgesia despite gross motor effects, it is the clinical question of tolerability that will determine the utility of such drug combination for the treatment of chronic pain. Some data already suggest that the combination of morphine with dextromethorphan, a noncompetitive NMDA receptor antagonist, is an effective analgesic with a favorable side-effect profile (Caruso, 2000; Katz, 2000).

In this study, the acute administration of NMDA receptor antagonists generally did not alter MSL or RR. Only the highest dose of dizocilpine (0.1 mg/kg) increased MSL and this effect was small (mean MSL = 0.42 mA). LY235959 and (+)-HA-966 did not increase MSL when administered alone at any dose tested. It is also important to note that RR was not altered by any dose of an NMDA receptor antagonist examined in this study. Given that the profound motor effects produced by NMDA receptor antagonists should interfere with responding at some dose, it is possible that the antinociceptive effects of NMDA receptor antagonists might be revealed in this procedure if higher doses were examined.

The doses of NMDA receptor antagonists selected for this study were based on the effects of these drugs in operant procedures described in the literature and on pilot work in our own laboratory. First, a dose of LY235959 higher than those presented here (3.0 mg/kg) was examined in two monkeys using this procedure (data not shown). This dose did not increase MSL, but did produce profound sedation that lasted over 36 h. On the basis of this pilot work, the dose range for LY235959 was restricted to doses below 3.0 mg/kg. The highest dose of dizocilpine tested in this study was 0.1 mg/kg. Slightly higher doses (0.17 mg/kg) have been shown to suppress response rates in squirrel monkeys responding in a drug discrimination procedure (Wiley et al., 1997). In fact, 0.1 mg/kg dizocilpine reduced response rates from 0.301 responses/s to 0.097 responses/s, while increasing MSL from 0.08 to 0.6 mA for one monkey in the present study. We have also noted during pilot experiments that ketamine and 1-(1-phenylcyclohexyl)piperidine(phencyclidine), two noncompetitive NMDA receptor antagonists, display steep dose-response functions with no-effect doses followed by doses that produce maximal increases in MSL and abolish responding (unpublished observations).

Thus, there were limitations on the doses of NMDA receptor antagonists suitable for testing in this procedure. It is a specious argument, however, that NMDA receptor antagonists generally lack antinociceptive effects. There is an extensive published literature documenting antinociceptive effects with the noncompetitive NMDA receptor antagonist ketamine, using a variety of procedures, experimental subjects, and with both experimental and clinical pain (Sadove et al., 1971; France et al., 1989; Park et al., 1995). More recently, attention has focused on dextromethorphan, another noncompetitive NMDA receptor antagonist that has demonstrated efficacy, at sufficient doses, for the relief of experimental and clinical pain (Price et al., 1994; Caruso, 2000). Indeed, there is electrophysiological evidence that NMDA receptor antagonists can reduce the response of spinothalamic tract neurons to the presentation of noxious thermal, mechanical, chemical, and electrical stimuli (Dougherty et al., 1992). There is also electrophysiological and behavioral evidence indicating that the facilitated neuronal response (Dougherty et al., 1992) and subjective ratings of pain (Price et al., 1994), characterized by central sensitization are prevented with an NMDA receptor antagonist.

Although there are contradictory findings, studies of morphine tolerance have demonstrated a lack of antinociceptive efficacy for NMDA receptor antagonists when administered alone and no potentiation of morphine antinociception (Trujillo and Akil, 1991; Tiseo and Inturrisi, 1993; Allen and Dykstra, 2000). These studies tend to use methods that may not be ideal for demonstrating these effects of NMDA receptor antagonists. For example, LY235959 did not potentiate the antinociceptive effects of morphine when tested with the rat warm-water tail-withdrawal procedure; however, morphine was administered in a cumulative fashion. Thus, the effect of LY235959 on low doses of morphine was only assessed at a single time point (Allen and Dykstra, 2000). A thorough time-dependent analysis of the effects of LY235959 in combination with morphine is necessary to address this difference.

From an empirical perspective, prevention of tolerance in the absence of acute antinociceptive effects of NMDA receptor antagonists is important for demonstrating that NMDA receptor antagonists prevent the development of opioid tolerance rather than simply enhancing the effectiveness of an opioid through an additive interaction with its own acute antinociceptive effects. From a treatment perspective, however, an acute interaction in addition to a role in preventing tolerance is a benefit. For example, research shows that the magnitude of tolerance that develops to the effects of an opioid is related to the amount of the drug administered chronically, with greater tolerance resulting from higher maintenance doses (Fernandes et al., 1982; Schuh et al., 1996; Allen and Dykstra, 2000). A drug combination that increases the effectiveness of lower doses of morphine might be expected to produce less tolerance relative to an equianalgesic dose of morphine alone independently of its potential to block the mechanism of tolerance. Thus, the results from this and other studies suggest that NMDA antagonists/opioid combinations have promise as analgesic agents for the long-term treatment of pain.