Complete Prevention but Stimulus-dependent Reversion of Morphine Tolerance by the Glycine/NMDA Receptor Antagonist (+)-HA966 in Neuropathic Rats

All experiments were conducted in accordance with the European Communities Council Directive (86/609/EEC) as well as with Ministry of Agriculture regulations. In addition, we adhered to the recommendations of the Committee for Research and Ethical Issues of the International Association for the Study of Pain Ethical Guidelines. 25In particular, the duration of the experiments was as short as possible, and the number of animals used was kept to a minimum.

Male Sprague-Dawley rats (Charles River, Saint-Aubin Les-Elbeuf, France; n = 115) that weighed 175–200 g on arrival were used. The rats were housed five in a cage on a 12-h light/12-h dark cycle. The ambient temperature was kept at 22°C, and the rats had free access to standard laboratory food and tap water. The animals were allowed to habituate to the housing facilities for at least 1 week before the experiments were begun.

The unilateral peripheral mononeuropathy was produced on the right hind limb according to the method described by Bennett and Xie 11and Attal et al.  12The animals were anesthetized with sodium pentobarbital (50 mg/kg intraperitoneally). The common sciatic nerve was exposed by blunt dissection at the level of the midthigh, and four loose ligatures (5-0 chromic catgut, approximately 1-mm spacing) were placed around the nerve, taking care not to interrupt the epineural circulation. To minimize the discomfort and possible painful mechanical stimulation, the rats were housed in large cages with sawdust bedding after the surgery. The neuropathic rats were able to eat and drink unaided.

All experiments were conducted in a quiet room between 8:00 AM and 12:00 PM. The animals were randomly assigned in groups of five (mechanical test) or 10 (thermal test) for a given series of tests and were not acclimatized to the test situations beforehand. The experimenter was unaware of the drug combinations used. Each animal received drugs only once and was used in only one experiment. Testing sessions lasted for approximately 2 h, and at the end of an experiment, rats were killed by an overdose of pentobarbital.

In mechanical tests, the antinociceptive action was determined by measuring the vocalization threshold elicited by pressure on the nerve-injured hind paw, using the Ugo Basile analgesymeter (Comerio, Italy). This instrument generates a linearly increasing mechanical force applied by a dome-shaped plastic tip (diameter = 1 mm) on the dorsal surface of the paw. The tip was positioned between the third and fourth metatarsus (into the sciatic nerve territory), and force was applied until the rat squeaked. This centrally integrated response is especially sensitive to analgesic compounds, particularly in this model of neuropathy. 13,15,16,26For each rat, a control threshold (mean of two consecutive stable thresholds expressed in grams) was determined before injecting the drugs. After drug administration, the vocalization thresholds were measured every 10 min until they had returned to the level of the control values.

Thermal nociception was tested by measuring the struggle latency elicited by immersion of the nerve-injured hind paw into a 46°C hot-water bath (Ministat MHUB 11; Bioblock Scientific, Illkirch, France) as described elsewhere. 12,15–17For each rat, a control latency (mean of two consecutive trials 20 min apart, expressed in seconds) was determined before injecting the drugs. After drug administration, struggle latencies were measured every 20 min until they had returned to baseline. The long interval (20 min) between successive measurements was necessary, because in this model, abnormal reactions lasting for more than 15 min after a thermal stimulus have been reported. 12,27 

The following drugs were used: morphine hydrochloride (Meram, Paris, France), naloxone hydrochloride (Narcan, DuPont Pharma, Paris, France), (+)-HA966 [(+)-(1-hydroxy-3-aminopyrrolodine-2-one; Tocris Cookson, Bristol, United Kingdom], and saline (0.9% NaCl). Morphine, naloxone, and (+)-HA966 were diluted in saline and injected in a volume of 1 ml/kg. The control rats received the same volume of saline.

Before the behavioral testing of the effect of an acute injection of morphine, rats were pretreated for 4 days with two consecutive subcutaneous injections of the following combinations of drugs: (1) saline and saline; (2) saline and morphine (10 mg/kg); (3) (+)-HA966 (2.5 mg/kg) and morphine (10 mg/kg); (4) (+)-HA966 (2.5 mg/kg) and saline. This morphine pretreatment has been shown to reliably induce tolerance to the antinociceptive effects of acute morphine in neuropathic rats. 18,19The 2.5-mg/kg dose of (+)-HA966 was chosen because it alone produced no antinociception but potentiated a single injection of morphine in neuropathic rats. 16The drugs were given twice daily (at 9:30 AM and 5:30 PM) for 4 days, beginning on day 12 after the surgery. The effect of an acute injection of morphine (1 mg/kg intravenously) was tested at 17 h after the last pretreatment injection, on day 16 after the surgery. At this time, the abnormal pain behavior in this model is at a stable maximum. 11,12Morphine was injected into the lateral tail vein. The dose of acute morphine (1 mg/kg intravenously) used in this study has repeatedly been shown to produce antinociceptive effects against both mechanical 13,15,16and noxious thermal 15–17stimuli. In an additional group of rats pretreated with (+)-HA966 and morphine, naloxone (0.1 mg/kg) was coinjected with acute morphine at a dose that has been shown to prevent the effect of 1 mg/kg morphine. 13,19 

In this experiment, morphine tolerance was induced by two daily injections of morphine (10 mg/kg subcutaneously) on postoperative days 12–15 after the same pretreatment paradigm as in experiment 1. At 17 h after the last pretreatment injection, on postoperative day 16, we examined the effect of an acute injection of the following combinations of drugs: (1) (+)-HA966 (2.5 mg/kg subcutaneously) and morphine (1 mg/kg intravenously); (2) saline (subcutaneously) and morphine (1 mg/kg intravenously); or (3) (+)-HA966 (2.5 mg/kg subcutaneously) and saline (intravenously). (+)-HA966 was administered 20 min before morphine. The latter of the drug combinations [(+)-HA966 2.5 mg/kg and saline) serves to examine whether (+)-HA966, at this dose, exhibits any analgesic properties in morphine-tolerant rats.

Data are expressed as mean ± SD. The overall effects of various treatments (areas under the curve [AUCs]) were calculated by use of the trapezoidal rule. Student t  test was used to determine the difference between two means. With three or more means, analysis of variance was used. The observed significances were then confirmed with Tukey test. The statistical procedures were performed with a statistical computer program (Statgraphics Plus; Manugistics, Rockville, MD). The observed differences were regarded as significant when P  values were < 0.05.

In agreement with numerous previous studies, 15,16the vocalization threshold of the nerve-injured hind paw was decreased on postoperative day 16 (216 ± 34 g vs.  the preconstriction value 330 ± 58 g;P < 0.001; n = 54).

As also reported previously, 15–17the struggle latency after immersion of the nerve-injured hind paw into a 46°C hot water bath was decreased (4.9 ± 1.2 s vs.  the preconstriction value 8.1 ± 1.2 s;P < 0.001; n = 61).

The mean vocalization thresholds and struggle latencies in the various pretreatment groups did not differ from the vehicle-pretreated group (table 1), which indicates that the pretreatments by themselves were devoid of effects on the pain-related behavior.

In the mechanical test, morphine produced an effect that peaked at 30 min (410 ± 30 g) and lasted for 80 min (fig. 1). In the thermal test, morphine also produced antinociceptive effects that peaked at 40 min (8.3 ± 1.2 s) and lasted for 60 min (fig. 2).

As previously observed, 18,19the intensity and duration of the effect of morphine in this group compared with the saline-pretreated rats was either strongly reduced, as in the mechanical test (AUC decreased to one fourth of that of the saline-pretreated group;figs. 1B and 1E), or absent, as in the thermal test (figs. 2B and 2E). This indicates that tolerance has developed to the antinociceptive effect of morphine.

When (+)-HA966 was administered in combination with morphine for 4 consecutive days, it prevented the development of tolerance to the antinociceptive effect of morphine on day 5. In the mechanical test, an intravenous injection of morphine resulted in antinociception that peaked at 30 min (450 ± 52 g) and lasted for 80 min (fig. 1C). Likewise, in the thermal test, morphine produced an effect that peaked at 40 min (7.6 ± 2.6 s) and lasted for 60 min (fig. 2C). The effect of morphine in the morphine and (+)-HA966–pretreated rats did not differ from the effect of morphine in the saline-pretreated rats (figs. 1E and 2E). Naloxone completely blocked the effect of morphine in both the mechanical (fig. 1C) and the thermal (fig. 2C) test.

In this group, the effect of morphine in the mechanical test peaked at 30 min (373 ± 50 g) and lasted up to 60 min (fig. 1D). In the thermal test, the effect of morphine peaked at 40 min (9.8 ± 3.2 s) and lasted for 60 min (fig. 2D). The effect of morphine against both mechanical (fig. 1E) and thermal (fig. 2E) stimuli was similar to the effect observed in the saline-pretreated group, indicating that pretreatment with (+)-HA966 on days 1–4 does not modify the effect of morphine on day 5.

As in experiment 1, the acute injection of morphine in morphine-pretreated rats resulted in very weak overall effects (AUCs;figs. 3A and 3B) compared with the effect of morphine in saline-pretreated rats (figs. 1E and 2E). This indicates that tolerance has developed to the antinociceptive effect of morphine.

In the mechanical test, acute concomitant injection of (+)-HA966 with morphine in morphine-pretreated rats produced an effect (F^9,60= 4.4, analysis of variance;P < 0.001) that peaked at 30 min (251 ± 35 g) and lasted for 50 min. The overall effect (AUC) of the combination of (+)-HA966 and morphine did not differ from the overall effect of morphine alone (fig. 3A;P = 0.12), indicating that the NMDA/glycine receptor antagonist is unable to inhibit established morphine tolerance. In the thermal test, combined administration of (+)-HA966 and morphine resulted in an antinociceptive effect (F^7,64= 2.8., analysis of variance;P < 0.05) that peaked (9.4 ± 2.7 s) at and lasted for 40 min. The overall effect (AUC) of the combination of (+)-HA966 and morphine was increased compared with the overall effect of morphine alone (fig. 3B;P < 0.05), indicating that against a noxious hot stimulus, (+)-HA966 inhibits established morphine tolerance. The effect of the combination of (+)-HA966 and morphine in morphine-pretreated rats did not differ from the effect of morphine alone in rats pretreated with (+)-HA966 and morphine (P = 0.44).

(+)-HA966 alone was devoid of any effect in the mechanical test (fig. 3A) as well as in the thermal test (fig. 3B).

This study confirms our previous findings, showing that morphine tolerance develops rapidly in neuropathic rats in the mechanical test, 18,19and extends this observation to the noxious thermal test. Although 1 mg/kg of intravenous morphine was completely ineffective against the thermal stimulus, it remained effective against the mechanical stimulus; however, it was reduced to approximately one fourth–fold after the morphine tolerance-inducing pretreatment. This may reflect previously observed differences in the effect of morphine depending on the stimulus modality. We have previously shown that the mechanical test exhibits a higher sensitivity to morphine than the thermal test. 13,15 

Experiment 1 found that the glycine/NMDA receptor antagonist (+)-HA966 prevents the development of morphine tolerance in neuropathic rats. The result is in accordance with previous studies performed in uninjured healthy rats that showed that NMDA receptor antagonist in general, 1,3,4and antagonist acting at the glycine site in particular, 8,28,29prevents morphine tolerance in rodents. To our knowledge, this is the first study to demonstrate an inhibitory effect of a NMDA receptor antagonist on the development of morphine tolerance in a model of neuropathic pain. However, it must be noted that these results have been obtained using one drug dose, and the uncertainty of the applicability of these observations to other doses must be underlined. The effect of morphine in the (+)-HA966–pretreated group did not differ from the effect of morphine in the saline-pretreated group, indicating that the repeated administration of (+)-HA966 on days 1–4 does not modulate the antinociceptive effect of morphine on day 5. The observation minimizes the possibility that a residue of (+)-HA966 injected on days 1–4 may have unmasked an effect of morphine in the morphine-tolerant rats rather than prevented the development of morphine tolerance. This is of particular importance because of the previously demonstrated synergy between (+)-HA966 and morphine in nociceptive tests. 16,24The (+)-HA966–induced prevention of morphine tolerance was completely reversed by naloxone, suggesting that opioid receptors mediate not only the acute effect of (+)-HA966 and morphine interaction as previously observed, 16but also events after chronic concomitant exposure of the μ-opioid and NMDA receptor.

Experiment 2 showed that a single injection of (+)-HA966 at 2.5 mg/kg reverses established morphine tolerance in the thermal test. Although we cannot completely exclude that the effect of (+)-HA966 and morphine in the morphine-pretreated rats reflects a simple synergic antinociceptive effect of the two drugs, this seems unlikely. We have previously shown that a single injection of low doses of morphine and (+)-HA966 results in profound antinociceptive effects that are more pronounced in mechanical than in thermal tests. 16Thus, it would be difficult to explain why the current study demonstrates an effect of combined administration of (+)-HA966 and morphine in morphine-pretreated rats in the thermal but not the mechanical test if it was the result of a simple synergy between the two drug systems. The result contrasts with studies in the uninjured healthy rat that demonstrated that a single injection of different noncompetitive NMDA receptor antagonist was devoid of effects on established morphine tolerance. 5,6,30However, more recent studies have shown a pattern of gradual reversal of tolerance by treatment with the NMDA receptor antagonist over a period of 2–4 days. 4,7,31,32Our study differs from previous studies in several aspects. First, we used a noxious thermal test to assess analgesia that differs from the often-used tail-flick or hot-plate test. In comparison to these tests, immersing the hind paw into a hot-water bath may stimulate a larger skin area and activate a greater number of receptors. Second, we used a glycine/NMDA receptor antagonist, whereas previous studies have used noncompetitive or competitive antagonists. The pharmacologic differences among NMDA receptors in different regions of the central nervous system may explain divergent results between studies using different antagonists. It was recently demonstrated that, although (+)-HA966 exhibits a preference of binding to the spinal cord, competitive antagonists conversely has a higher relative potency in the cerebral cortex. 33Furthermore, the study found that the generation of a chronic constriction injury had no effect on the density of NMDA receptors in the spinal cord, nor on the relative potencies of the different antagonist. The relative spinal cord selectivity of (+)-HA966 compared with other antagonists is interesting because the spinal cord is probably a major site of action of NMDA receptor antagonists in attenuating morphine tolerance. 34Finally, our study is performed in neuropathic rats, in which the spinal NMDA receptors are supposed to be already in an activated state before any drug pretreatment. 9Increased spontaneous activity of the injured nerves is believed to result in sustained activation of the NMDA receptor and to be partly responsible for the phenomenon of central sensitization, resulting in allodynia and hyperalgesia. 35Woolf and Thompson 35showed that both competitive and noncompetitive NMDA antagonists cannot only prevent but also reverse the central sensitization state once it is established, an observation that parallels our results, demonstrating that (+)-HA966 prevents and reverses tolerance in the thermal test. Our results join those of previous studies, suggesting that the reorganization of the nervous system during continued opioid exposure leading to a reduction in opioid responsiveness is mediated mainly by the activation of the NMDA receptor. Because the NMDA receptor seems to be implicated in the events leading to both neuropathic pain and morphine tolerance, it has been suggested that the development of these phenomena shares common intracellular events, of which an important step is an increased intracellular calcium concentration. 36Thus, one might expect a relatively high level of intracellular calcium because of prolonged NMDA receptor activation in the animal exposed to both morphine tolerance and nerve injury. It may therefore seem unexpected that the glycine/NMDA receptor antagonist is able to reverse the adaptive processes, leading to morphine tolerance in the nerve-injured rat and not in the uninjured healthy rats. 5,6 

(+)-HA966 reversed established morphine tolerance in the thermal but not in the mechanical test. This may suggest that the expression of morphine tolerance in the neuropathic pain state is related to the modality and/or the intensity of the nociceptive stimulus. Although the analgesymeter used in the mechanical test generates a punctiform stimulus, the thermal stimulus is applied to a much larger surface. Phenomena such as spatial summation and spatial convergence are major features of the central encoding of noxious heat messages. 37The test-depending ability to reverse morphine tolerance may also rely on a difference in afferent fibers that are stimulated in the mechanical and the thermal test of the nerve-injured animal. 38–40Finally, primary afferent fibers contain several neuropeptides and excitatory amino acids, and the spinal release of putative neurotransmitters, such as substance P, seems to vary in animals subjected to different nociceptive stimuli. 41Further studies are needed to confirm if the ability of NMDA receptor antagonists to reverse established morphine tolerance vary with the applied stimuli.

Whatever the precise mechanisms underlying the effect of (+)-HA966 on morphine tolerance, the design of the present study mimics the clinical situation in which morphine administration is started only after the manifestation of a painful disorder. By adding an NMDA receptor antagonist, more efficient antinociception and inhibition of morphine tolerance is obtained in animal studies. However, the clinical utility of these antagonists is diminished by their adverse side effects. Antagonists acting at the glycine site of the NMDA receptor complex represent a subclass that may have the potential of exerting beneficial therapeutic effects without side effects, properties that merit further exploration also in clinical trials.