- Email: [email protected]
glycine site receptor in the ventroposterolateral nucleus of the thalamus
Pain 84 (2000) 213±224 www.elsevier.nl/locate/pain
Modulation of nociceptive transmission by NMDA/glycine site receptor in the ventroposterolateral nucleus of the thalamus Fabio Bordi*, Mauro Quartaroli Pharmacology Department, GlaxoWellcome Medicine Research Centre, 37100 Verona, Italy Received 12 March 1999; received in revised form 22 July 1999; accepted 3 August 1999
Abstract NMDA-type glutamate receptors are involved in the generation and maintenance of altered pain states. In the present study, we examined the effect of an NMDA-glycine site antagonist, GV196771A [E-4,6-dichloro-3-(2-oxo-1-phenyl-pyrrolidin-3-ylidenemethyl)-1H-indole-2carboxylic acid sodium salt], on responses to noxious stimuli both in normal rats and during peripheral mononeuropathy induced by chronic constriction injury (CCI) of the sciatic nerve. In one series of experiments, activity of nociceptive neurons in the ventroposterolateral (VPL) nucleus of the thalamus was recorded in response to pressure stimuli to the contralateral hindpaw. Intravenous injection (iv) of the glycine antagonist had no effect on these cells in normal rats. When tested in rats with CCI induced 2±3 weeks previously, however, GV196771A (0.125, 0.5 and 2.0 mg/kg) blocked responses to noxious stimulation in a dose-dependent and reversible manner. Morphine (0.5 mg/kg, iv) and the NMDA channel blocker MK801 (0.1 mg/kg, iv) suppressed noxious stimulus-evoked activity of VPL neurons in both normal and CCI-treated rats. MK801 also decreased the responses of non-nociceptive neurons to brush stimulation in both sets of animals, in contrast to the glycine antagonist which did not alter the responses of these cells. Similar results were obtained from a series of behavior experiments in which the latency for paw withdrawal from heat stimulation was measured in normal and CCI-treated rats. GV196771A (3 and 10 mg/kg) injected orally, reduced the hyperalgesic response in the treated rats but did not change the withdrawal latency in normal rats. Taken together, these ®ndings suggest that block of the NMDA receptor decreases nociceptive transmission in the thalamus and can modulate hyperalgesic states. GV196771A and glycine antagonists in general may represent innovative and safe agents for the treatment of neuropathic pain. q 2000 International Association for the Study of Pain. Published by Elsevier Science B.V. Keywords: NMDA; Glycine antagonist; Thalamus; Sensory transmission; CCI; Hot plate
1. Introduction The N-methyl-D-aspartate (NMDA) glutamate receptor subtype has an important role in mediating nociceptive processing. Several results from behavioral studies suggest that NMDA receptor antagonists have antinociceptive effects in a number of models of nociception, both in animals and man (Murray et al., 1991; Kristensen et al., 1992; Nasstrom et al., 1992; Mao et al., 1992; Ren et al., 1992; Ren and Dubner, 1993). Activation of spinal NMDA receptors has been shown to mediate these effects. Spinal NMDA receptors are critical for pain perception and transmission (Aanonsen et al., 1990; Woolf and Thompson, 1991; Dubner and Ruda, 1992), while NMDA antagonists reverse the increased responsiveness of spinal neurons induced by acute arthritis (Neugebauer et al., 1993), intra* Corresponding author. Pharmacology Department, GlaxoWellcome Research, Via Fleming 4, 37100 Verona, Italy. Tel.: 139-045-921-8845; fax: 139-045-921-8153. E-mail address: [email protected] (F. Bordi)
dermal injection of capsaicin (Dougherty et al., 1992), intraplantar injection of formalin (Haley et al., 1990), application of the irritant mustard oil (Woolf and Thompson, 1991), and experimental peripheral neuropathy (Laird and Bennett, 1992). The extent to which these effects are conserved throughout ascending sensory pathways is not known, however. Few studies have examined the contribution of NMDA receptors to the processing of the nociceptive input or in chronic pain models in areas of the CNS other than the spinal cord. Only some studies have characterized the response of thalamic neurons to peripheral stimuli showing the involvement of NMDA and non-NMDA receptors in somatosensory and nociceptive transmission (Salt, 1987; Eaton and Salt, 1990; Dougherty et al., 1996). The spinothalamic tract originates in the dorsal horn of the spinal cord and terminates in the ventroposterolateral (VPL) nucleus of the thalamus. VPL receives inputs from wide dynamic range neurons in the spinal cord, which encode the strength of noxious and non-noxious stimuli
0304-3959/00/$20.00 q 2000 International Association for the Study of Pain. Published by Elsevier Science B.V. PII: S 0304-395 9(99)00205-5
214
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
(Mendell, 1966; see Willis and Coggeshall, 1991, for a review). VPL neurons have receptive ®elds on a restricted area of the contralateral skin, making them well suited for the sensory-discriminative aspects of pain (Peschanski et al., 1983; Yokota et al., 1988; Martin et al., 1990; Willis and Westlund, 1997). Like their spinal counterparts, VPL nociceptive neurons sensitizes to mechanical and thermal stimuli after peripheral in¯ammation (Guilbaud et al., 1986, 1987) and peripheral neuropathy (Guilbaud et al., 1990). Recent evidence, showing that VPL is a major site of convergence of somatic and visceral inputs conveyed through the dorsal column (Al-Chaer et al., 1996), make this nucleus an important station for pain processing. The NMDA receptor complex belongs to the family of ligand-gated ion channels, but with unique features. This receptor is highly permeable to Ca 21, is under voltagedependent regulation of extracellular Mg 21, and requires the simultaneous presence of two agonists to be activated: glutamate itself and glycine (Johnson and Ascher, 1987; Kleckner and Dingledine, 1988; Hollmann and Heinemann, 1994). The use of agents that act at the glycine modulatory site appears to be more attractive clinically because these agents have fewer adverse CNS side effects (see Danysz and Parsons, 1998, for in-depth review). Glycine antagonists have been reported to be effective in reducing nociceptive responses to formalin injection in spinal cord neurons (Dickenson and Aydar, 1991; Millan and Seguin, 1994) and in reversing thermal hyperalgesia evoked by in¯ammation or peripheral neuropathy in behavioral experiments (Mao et al., 1992; Laird et al., 1996). The aim of the present study was to examine the role of the NMDA receptor in the processing of nociceptive stimuli, using an NMDA channel blocker, MK801, and the novel NMDA/glycine site antagonist GV196771A (Giacobbe et al., 1998; Quartaroli et al., 1999). To test the effects of these two different NMDA antagonists in a model of chronic pain, a chronic constrictive injury (CCI) was experimentally produced around the sciatic nerve in rats. Electrophysiological recordings were made of nociceptive responses of wide dynamic range neurons in the VPL nucleus of the thalamus of anesthetized rats. Behavioral procedures examined the analgesic properties of the GV196771A in a acute model of nociception, the hot plate test, and in a chronic model of nociception, the paw withdrawal to thermal stimulation in CCI animals.
2. Methods 2.1. Electrophysiological methods 2.1.1. Animals and surgical preparation Male Sprague±Dawley rats (n 63) weighing 300±400 g were anesthetized with urethane in three injections separated each by about 10 min (1.5 g/kg, ip, ®nal concentration) and placed in a Kopf stereotaxic frame adjusted so that
the surface of the skull was level between lambda and bregma. Body temperature was regulated at 37 ^ 18 C by means of a heating pad. Depth of anesthesia was monitored during the experiment by assessing responsivity to tail pinch and urethane was supplemented as needed. In most cases no extra anesthetic was necessary as the period of anesthesia lasted 4±5 h, determined by previous experiments (Bordi et al., 1997). Stereotaxic coordinates were measured from bregma and calculated using a brain atlas (Paxinos and Watson, 1986). A small hole was made over the cortex above VPL, and a stainless steel electrode (3±4 MV impedance at 1000 Hz; Frederick Haer & Co., Bowdoinham, ME) was inserted into VPL using an electronic micropositioner (Kopf Inst., Tujunga, CA). Neurons were identi®ed by spontaneous activity or by a light brush stimulus of the contralateral hindpaw using a wooden probe or brush. Neurons responding to the search stimulus were selected and their responses tested to a graded pressure stimulus applied to the receptive ®eld. Data were collected for 2 s before the stimulus onset, during the 5 s of stimulus presentation, and for an additional 5 s at the end of stimulus offset. A tail vein was cannulated for intravenous (iv) injections. Drugs were tested only once in each animal, at a single dose. Twenty-six rats used for the electrophysiological studies had an experimental neuropathy induced by previous ligature of the sciatic nerve. 2.1.2. Sciatic nerve ligation surgery Chronic constriction injury (CCI) was induced in rats under pentobarbital anesthesia (50 mg/kg), and the left common sciatic nerve was exposed. Proximal to the sciatic trifurcation, about 10 mm of nerve was freed of adhering tissue and 4 ligatures (3.0 chromic gut) were tied loosely around it at intervals of 1 mm (Bennett and Xie, 1988). As previously shown the animals developed hyperalgesia within 14 days (Quartaroli et al., 1999). Behavioral and electrophysiological experiments were performed 14±21 days after ligature. 2.1.3. Pain stimulation and recording A computer-controlled air cylinder was used to administer the pressure stimulus using an adaptation of the method described by Martin et al. (1996). Air pressure was controlled by a computer-driven Picospritzer II (General Valve, Fair®eld, NJ) and was measured on-line in the air cylinder with a pressure transducer (Spectramed BV, Bithoven, The Netherlands). The stimulus was 5 s in duration, rising continuously from zero to a peak of about 4.2 kg/ cm 2, applied to the probe itself, which was about 2 mm 2 in cross-section. The relation between air pressure and stimulus force was measured empirically. To determine the threshold to pain, a behavioral experiment was carried out. Ten rats were anesthetized with a dose of urethane (1 g/kg, ip) that reduced motor tone but did not suppress the animal's re¯exes. At this dose, the animals permitted insertion of the hind paw into the pressure device
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
but would also withdraw their limb in response to noxious stimulation applied to their hind paw by the air cylinder. Pressure stimuli were presented 8 times at 2 min intervals. The mean pressure at which the behavioral response occurred was calculated for each rat. The mean pressure eliciting limb withdrawal was 3:1 ^ 0:15 kg/cm 2. This level of stimulation was reached after about 3 s from stimulus onset; therefore the last 2 s of the pressure stimulus were above the pain threshold. Once a neuron was isolated and found responsive to light brushing, its baseline responses to the noxious pressure stimulus were determined by applying the stimulus 10 times at 1 min intervals. The drug or vehicle was then injected intravenously, and the stimulus was applied again to examine the effects of the treatment. Recovery from the effects of the drug were measured for the next hour, during which responses were tested every 2±3 min. Neuronal potentials were ampli®ed by an a.c. ampli®er (Fintronic model WDR 420, Derby, CT), band pass ®ltered (400±8000 Hz), and displayed on a dual-trace storage oscilloscope (Tektronix 5111A, Beaverton, OR) (Bordi and
215
LeDoux, 1994). A window discriminator was used to convert single action potentials to logic pulses, which were recorded (Digidata 1200, Axon Inst., Foster City, CA), displayed and stored to disk using customized software. Peristimulus histograms (PSTHs) and raster plots of each response were displayed at the end of each stimulus presentation on the computer screen. To obtain stimulus ®ring rates before and after drug treatment, ®ring rates for each neuron were measured at 9 levels of pressure (from 0 to 4.2 kg/cm 2). These mean ®ring rates and pressures were subjected to linear regression. Only neurons that exhibited stimulus±response functions with a correlation coef®cient of at least 0.5 were classi®ed as wide dynamic range (WDR) neurons and included in the study. In some cases, different pressure stimulus waveforms were delivered to determine how the responses of a WDR neuron were correlated to stimulus intensity (Fig. 1A). Non-nociceptive mechanoresponsive neurons (n 13) were also encountered in VPL. Because these neurons are silent or low rates or spontaneous activity, as described (Martin et al., 1996), a search stimulus (light brush) was
Fig. 1. (A) Example of PSTHs for a single VPL neuron showing correlation of the response with stimulus magnitude. For each stimulus waveform, the level of the applied pressure is shown in register with the PSTH (top), mean response of the 10 stimulus repetitions (middle), and raster plot showing the 10 trials (bottom). The stimulus±response function (pressure vs. ®ring frequency), together with the corresponding correlation coef®cient is shown to the right. A single action potential is shown in the inset at the top right. (B) Example of a typical response to the graded mechanical pressure used throughout this study. Top, level of the applied pressure; right, stimulus-response function of the cell. (C) Mean pre-drug stimulus±response function for all nociceptive wide dynamic range neurons (n 24). Dotted lines indicate mean pressure eliciting limb withdrawal in behavioral experiments (see Section 2).
216
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
used regularly during the electrode penetration. Once isolated, the neuron's response to light touch stimulation was examined. Tapping the skin within the receptive ®eld of the neuron produced a reliable activation of mechanoresponsive neurons. These neurons were de®ned as non-nociceptive because they did not respond to noxious pressure stimulus with a greater response than to the light brush stimulus. Usually they also habituated rapidly to the pressure stimulus. Responses were measured to light taps every 1±2 min for 6±10 times before drug infusion and the average calculated. The effect of the drug was examined 5±15 min post injection and in some cases for 1±2 h afterwards, to monitor recovery of the response. The average of the postinjection response was compared to pre-injection control. Only one cell was tested in a single animal. 2.1.4. Data analysis For each unit, the mean spontaneous ®ring rate was calculated during the 2 s interval immediately before each cutaneous stimulation. Mean evoked activity was calculated during the 5 s of pressure stimulus. The average of all units for each drug condition was used to construct the time course of the effect of the drug as fraction of the predrug response. The effects of drug treatment on stimulus± response functions of nociceptive neurons were determined by calculating the average rate of ®ring of the cell at 9 pressure levels before and after drug injection. A custom software program was used to construct PSTHs and individual raster plots as shown in Fig. 1A. Responses were compared using ANOVA measurements followed by Newman±Keuls post-hoc comparisons. Statistical differences for non-nociceptive cells were established using the Student's t test for unpaired samples. 2.1.5. Histology At the end of the experiment a small lesion was made at the recording site by applying an anodal current of 40± 50 mA for 5 s. Animals were then perfused with 10% buffered formalin, 5% potassium ferrocyanide, and 5% potassium ferricyanide (Bordi and LeDoux, 1994). The brains were frozen and cut on a sliding microtone (50 mm sections), and mounted sections were stained with cresyl violet. The position of the recording location was con®rmed microscopically. 2.2. Behavioral methods 2.2.1. Thermal hot plate test in naive animals Male Sprague±Dawley rats (n 44) weighing 250±350 g were used. They were housed for at least a week upon their arrival in groups of 2±3 per cage with food and water until the beginning of the experiment. The hot plate test was performed on an electrically heated and thermostatically controlled copper surface, set to a temperature of 55 ^ 0:28C. The animals were con®ned to the hot plate by a transparent observation chamber and the latency to the
response, consisting of licking of the hind paw, was measured (pre-test). A cut-off period of 30 s was used to avoid tissue damage. The animals were then treated orally with GV196771A (10 mg/kg) or vehicle and after 1 h the test was repeated. The effect of morphine (5 mg/kg, ip) and MK801 (0.125 mg/kg and 0.25 mg/kg, ip) was measured 30 min post injection. Statistics consisted in ANOVA measures followed by Dunnett's test. 2.2.2. Paw withdrawal in CCI animals Male Sprague±Dawley rats (n 30) weighing 300±350 g were used. They were housed two per cage with food and water for 14±21 days after CCI surgery. Paw withdrawal latencies to radiant heat were determined for both hindpaws as previously described (Bennett and Xie, 1988; Quartaroli et al., 1999). Brie¯y, the animals were placed in acrylic cages on a glass plate. After 15±20 min of accommodation, thermal hyperalgesia was tested using a commercially available analgesimeter (Plantar test, Ugo Basile, Comerio, Italy). A heat stimulus (50 W, 8 V) was applied directly onto the plantar surface of each hind paw and the withdrawal latency was determined. Four latency measurements were taken for each paw and averaged. GV196771A (1± 10 mg/kg) or vehicle was administered orally to animals 14±21 days after ligation and hyperalgesia tested 1 h post treatment. Statistical analysis was done comparing withdrawal latency time between unoperated vs. ligated paw for each group condition. Differential scores (latency of the unoperated paw minus latency of ligated paw) were also computed for each group and ANOVA was performed followed by Dunnett's test. 2.3. Drug treatment GV196771A [E-4,6-dichloro-3-(2-oxo-1-phenyl-pyrrolidin-3-ylidenemethyl)-1H-indole-2-carboxylic acid sodium salt] (synthesized by Medicinal Chemistry Dept., GlaxoWellcome, Verona) for the electrophysiological experiments was dissolved in DMSO (®nal concentration 1±2%) and diluted in saline. The vehicle control was 2% DMSO in saline. MK801 (RBI, Natick, MA) and morphine (Sigma, St. Louis, MO) were dissolved in saline. Drugs were administered in a volume of 1 ml/kg through the lateral vein. In the behavioral experiments, GV196771A was prepared as a stock solution of 1 mg/ml in 0.5% methylcellulose (methocel); further dilutions were made in 0.5% methocel. It was administered orally in a volume of 10 ml/kg. Morphine and MK801 were diluted in saline and injected ip in a volume of 2 ml/kg. 3. Results A total of 50 cells isolated in the VPL responded to noxious stimulation of the contralateral hindpaw and an additional 14 cells were activated by innocuous tactile stimulation. Cells were identi®ed on the basis of sponta-
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
neous activity or in response to cutaneous search stimuli (light brushing of the skin). Of these cells, 26 nociceptive and 6 non-nociceptive cells were found in animals that had an experimental neuropathy induced by CCI of the sciatic nerve. The recording sites of all units were located in the VPL region, as previously reported (Angel and Clarke, 1975; Cropper and Eisenman, 1986; Roberts and Wells, 1990; Martin et al., 1996; Sherman et al., 1997). 3.1. Effects of drug treatment on nociceptive neurons in normal animals The 24 VPL nociceptive neurons in normal animals had an average spontaneous background activity of 3.8^0.7 Hz. Their ®ring rate increased up to 20:9 ^ 1:2 Hz at the maximum pressure strength of the graded mechanical stimulus applied to their receptive ®eld (Fig. 1B,C). Responses were graded according to the intensity of the pressure stimulus, con®rming that these VPL neurons were wide dynamic range cells (Fig. 1A). The correlation coef®cient of response vs. stimulus intensity before drug injection was on average 0:86 ^ 0:18. The effects of intravenous morphine (0.5 mg/kg) on the responses of these neurons to the pressure stimulus was investigated in 4 neurons. Responses of a representative VPL neuron are shown in Fig. 2A. Morphine produced a marked depression of the stimulus-evoked response. The effect of morphine was statistically signi®cant at all stimu-
217
lus levels except for the two lowest pressures, suggesting that the effect was speci®c for nociceptive responses. The effect of morphine persisted for at least 1 h (Fig. 3C). The effect of NMDA receptor blockade on nociceptive function was tested using the NMDA blocker MK801. As shown in Fig. 3, MK801 has a potent antinociceptive action. At 0.1 mg/kg MK801 (n 5) was as effective a blocker as morphine (0.5 mg/kg), signi®cantly reducing the response at all pressures above 1 kg/cm 2. At 0.01 mg/kg (n 5), on the other hand, MK801 was effective only at higher levels of pressure stimulation (Fig. 3A). Two neurons were also tested at 1.0 mg/kg MK801, but in one case the animal died and in the other all activity was blocked (data not shown). There were dose-dependent differences in the length of time that MK801 was effective (Fig. 3C). The selective antagonist of the glycine site on NMDA receptors, GV196771A, was tested on 6 nociceptive cells in 6 normal rats. In contrast to the effects of morphine and MK801, GV196771A did not reduce the responsiveness of nociceptive cells even at high doses (2.0 mg/kg, Fig. 3B). Responses of cells after treatment with GV196771A were indistinguishable from control at all times after injection (data not shown). 3.2. Effects of drug treatment on non-nociceptive neurons of normal animals A total of 8 non-nociceptive neurons were identi®ed by
Fig. 2. Effects of morphine on nociceptive VPL neurons. (A) Evoked ®ring of a representative neuron before and after administration of morphine (0.5 mg/kg, iv). Top, level of applied pressure in register with the PSTHs below. Middle, average of predrug ®ring rate histograms and raster plot of the 10 individual trials. Bottom, PSTHs and raster plots after morphine injection. (B) Mean stimulus±response function after the administration of vehicle (®lled circles, n 4) or morphine (®lled triangles, n 4). Asterisks indicate statistical difference from vehicle at different times (P , 0:01, ANOVA followed by Newman±Keuls tests).
218
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
Fig. 3. Effects of MK801 and GV196771A on nociceptive VPL neurons. (A) Mean stimulus±response function after injection of vehicle (®lled circles, n 4), 0.01 mg/kg MK801 (open triangles, n 5), or 0.1 mg/kg MK801 (open squares, n 5). (B) Mean stimulus±response function after administration of vehicle (®lled circles, n 4) or GV196771A (®lled diamonds, n 6). No difference was found at any level of pressure applied to the contralateral paw. (C) Evoked ®ring over time before and after administration of the vehicle (®lled circles, n 4) or MK801 (0.01 mg/kg, open triangles, n 5); 0.1 mg/kg (open squares, n 5). Morphine (0.5 mg/kg, ®lled triangles, n 4) is shown for comparison. White arrow indicates injection time. * P , 0:01.
search stimulation. They all had no spontaneous activity and responded to light touch of the contralateral hindpaw. After intravenous injection of MK801 (0.1 mg/kg, n 5) or GV196771A (2.0 mg/kg, n 3) the response to nonnoxious stimulation was examined again. MK801 reduced the response to cutaneous stimulation as determined by counting the number of spikes elicited by brushing the skin (a light touch) (Fig. 4). The average decrease for the 5 cells was 44% (T5 21:7, P , 0:01). The glycine antagonist GV196771A, on the other hand, failed to suppress ®ring evoked by non-noxious stimulation; responses after treatment were 96% of the pre-drug response (T3 1:92 ns). 3.3. Behavioral effects of hot plate test The effect of GV196771A on acute nociceptive responses
Fig. 4. Activity of non-nociceptive neurons after the intravenous administration of MK801 (0.1 mg/kg, n 5) or GV196771A (2.0 mg/kg, n 3), expressed as percent of pre-drug control value. MK801, but not GV196771A, signi®cantly reduces response to light brush stimulation (t 21:7, P , 0:01).
was also tested by measuring the latency of hindpaw licking in response to heating the paw (the hot plate test; see Section 2). As shown in Fig. 5, GV196771A infused orally (10 mg/ kg, n 10) did not change the response, whereas morphine (5 mg/kg, ip, n 9) produced a signi®cant increase in the latency of the hindpaw licking response compared to control (n 10), as previously described (Yaksh, 1981; Cicero et al., 1996). Some animals were given either 0.125 mg/kg (n 6) or 0.250 mg/kg (n 9) of MK801. While the lower dose did not have any effect, the higher dose of MK801 signi®cantly increased the time of the occurrence of the ®rst hindpaw licking (data not shown). Psychomimetic effects, such as locomotor activity, stereotypic sniffing and circling, were elicited by this dose however, as described (Hoffman, 1992; Murata and Kawasaki, 1993). These effects confounded the response and may have masked the possible anti-nociceptive activity.
Fig. 5. Mean time of occurrence of ®rst hindpaw licking to hot plate stimulation (558C) after injection of vehicle (n 10), GV196771A (10 mg/kg, p.o., n 10), or morphine (5 mg/kg, ip, n 9). Asterisks denote cases signi®cantly (P , 0:01) different from vehicle control (Dunnett's test).
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
3.4. Effects of drug treatment on nociceptive neurons in CCI animals A total of 26 nociceptive neurons were recorded in the VPL of animals with experimental neuropathy induced by loose constriction of the contralateral sciatic nerve. Their average spontaneous activity was unchanged compared to control animals (4:2 ^ 0:4 Hz vs. 3:8 ^ 0:7 Hz). A major
219
difference in the response of cells in CCI animals was the presence of a long afterdischarge following the pressure stimulus (Fig. 6A). This afterdischarge was present in 61% (16/26) of the neurons, compared to only 8% (2/24) in control animals. In some neurons in CCI rats, more than 1 min was required between stimuli for the activity to return to baseline, and in some cases the interval between successive stimulus presentations was increased to 2±3 min. Even
Fig. 6. (A) Responses of a nociceptive VPL neuron in a CCI animal to graded pressure stimuli. Top, pressure applied to the paw. Middle, mean response of the 10 stimulus trials. Bottom, raster plot of the 10 individual trials. (B) Mean predrug stimulus±response function for all nociceptive neurons recorded in CCI rats (n 26). (C) Effect of morphine in a representative neuron. Top, pressure applied to the paw. Average PSTHs and individual raster plots before (middle) and after (bottom) injection of morphine (0.5 mg/kg, iv). (D) Mean stimulus±response function after the administration of vehicle (®lled circles, n 4) or morphine (®lled triangles, n 3). * P , 0:01.
220
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
then, the activity in some cells did not recover completely, as illustrated in the raster plots of Fig. 6C where the interstimulus interval was 3 min. The average responses of all 26 nociceptive CCI-neurons are summarized in Fig. 6B, which shows a strong correlation (r 0:89 ^ 0:14) between response and stimulus magnitudes. This relationship was very similar to the one observed for nociceptive cells in normal animals (compare Figs. 1C and 6B). As in normal animals, morphine (0.5 mg/kg, iv, n 3)
dramatically reduced the response of the neurons to pressure stimulation (Fig. 6C,D). Morphine was also effective in blocking the sensitization phenomena, as shown in a representative neuron in Fig. 6C. Similar to normal animals, MK801 (0.1 mg/kg, n 3) signi®cantly suppressed nociceptive activity in CCI animals (data not shown, see Fig. 3 for MK801 effect on normal animals). The time course of this effect was also similar to normal animals, with about 70% recovery after 1 h (Fig. 7C ). In contrast to its lack of effect in normal animals, the antagonist GV196771A produced a large, dose-dependent depression of nociceptive responses in CCI animals (Fig. 7). At the highest dose (2.0 mg/kg, n 6), blockade was complete, just as for morphine, but even at 0.5 mg/kg (n 8), a signi®cant reduction was observed (Fig. 7B). The lowest dose (0.125 mg/kg, n 3), did not have a significant effect, however (Fig. 7B). Fig. 7C shows that maximal effects occurred within 15 min after drug injection and had largely recovered (,70%) after 1 h. The action of GV196771A at 2.0 mg/kg on evoked ®ring over time was very similar to that of MK801 (0.1 mg/kg); the striking difference was that GV196771A had this effect only in CCI animals whereas MK801 blocked nociceptive responses in normal animals as well. 3.5. Effects of drug treatment on non-nociceptive neurons of CCI animals Six neurons responding to light touch cutaneous stimulation were isolated to test the effects of MK801 or GV196771A on non-nociceptive mechanosensitive neurons. Injection of MK801 (0.1 mg/kg) signi®cantly depressed responsiveness (n 3, mean 50% of pre-drug control, T3 3:9, P , 0:05), whereas GV196771A injection (2.0 mg/kg) had no effect (n 3, mean 113% of control, T3 1:7, ns). Fig. 8 shows responding to the light mechanical stimulation after MK801 or GV196771A intravenous infusion. 3.6. Behavioral effects of heat stimulation in CCI animals
Fig. 7. Decrease in the responsiveness of VPL neurons in CCI animals to noxious pressure stimuli after administration of GV196771A. (A) Effect of 2.0 mg/kg GV196771A on nociceptive responses of a representative cell. (B) Mean stimulus±response functions after intravenous administration of vehicle (®lled circles, n 4) or GV196771A (0.125 mg/kg, ®lled triangles, n 3; 0.5 mg/kg, ®lled squares, n 8; 2.0 mg/kg, ®lled diamonds, n 6). (C) Evoked ®ring over time before and after administration of the vehicle (®lled circles, n 4) or GV196771A (0.125 mg/kg, ®lled triangles, n 3; 0.5 mg/kg, ®lled squares, n 8; 2.0 mg/kg, ®lled diamonds, n 6). Effect of the NMDA blocker MK801 (0.1 mg/kg, open squares, n 3) is shown for comparison. * P , 0:01.
Prior to surgery, no difference was detected between withdrawal latencies of left vs. right paws in response to heat, both for vehicle- and GV196771A-treated animals (data not shown). Two to three weeks following surgery, vehicle-treated animals showed a decrease in thermally induced paw withdrawal in the operated paw. GV196771A reduced the latency difference between operated and control paws in a dose-dependent manner, measured 1 h after oral administration of vehicle (n 10) or GV196771A (1 mg/kg, n 10; 3 mg/kg, n 7; 10 mg/ kg, n 8). GV196771A increased the withdrawal latency for the ligated paw, without affecting the unoperated paw. Fig. 9A shows the effects of the highest dose of GV196771A (10 mg/kg, n 8). Fig. 9B shows the differential scores between the ligated and the unoperated paw. The two highest doses of GV196771A (3 and 10 mg/kg) exhibited a
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
221
tive responses in VPL neurons of the rat thalamus and reduces noxious responses as measured behaviorally. The effects of GV196771A were potent, dose-dependent and reversible. GV196771A was also highly selective for neuropathic nociceptive stimulation, because it did not alter the ®ring either of nociceptive neurons in normal rats or of nonnociceptive neurons responding to light touch. In contrast, the NMDA antagonist MK801 reduced nociceptive responses in both CCI and normal animals. Furthermore, MK801 was less selective because it also signi®cantly depressed ®ring of non-nociceptive neurons, both in normal and CCI-treated animals. Morphine also inhibited noxious stimulus-evoked responses in VPL neurons of normal rats, con®rming earlier results (Hill and Pepper, 1978; Benoist et al., 1983; Jurna et al., 1996). In our study, the amount of reduction of the response in VPL neurons (,80%) is quite similar to that reported by Jurna et al. (1996) after iv injection of morphine. Interestingly, morphine is capable of producing a maximum depression of only , 50±60% of control when injected by intrathecal administration, suggesting that spinal cord alone cannot achieve maximum analgesia (Jurna et al., 1996). The effects of MK801 on nociceptive neurons were quite similar to those of morphine, both in magnitude and duration. MK801 too, almost completely inhibited noxious stimulus-evoked response. Morphine also had a strong antihyperalgesic effect on VPL nociceptive neurons after neuropathy induced by CCI. This result is consistent with a previous report that showed depression of thalamic neuronal responses to pinch stimulation of the ligated hindpaw after morphine injection (Guilbaud et al., 1991). The glycine antagonist GV196771A at the highest dose produced effects very similar to those of morphine (see Figs. 7 and 8), Fig. 8. Activity of non-nociceptive neurons in CCI-operated animals after intravenous administration of MK801 (n 3) or GV196771A (n 3). (A) Firing rate histogram of a representative mechanosensitive non-nociceptive neuron before (top) and after (bottom) administration of 2.0 mg/kg GV196771A. The glycine antagonist does not reduce the responses of the neuron to tap stimuli (s) delivered every minute. (B) Responses of a nonnociceptive neuron are reduced after administration of MK801 (0.1 mg/kg). (C) Mean of all experiments expressed as percentage of pre-drug control responses. MK801, but not GV196771A, signi®cantly reduces responses to light brush stimulation (t 3:9, P , 0:05).
signi®cant difference vs. vehicle control (P , 0:05 and P , 0:01, respectively), indicating that the drug effectively reduces the behavioral noxious response. No statistical difference was found at the lowest dose (1 mg/kg). 4. Discussion The present study demonstrates that in a neuropathic pain state induced experimentally by chronic constriction of the sciatic nerve (CCI), the NMDA/glycine site antagonist GV196771A inhibits stimulus-evoked activity of nocicep-
Fig. 9. Dose-dependent effect of GV196771A on thermal hyperalgesia in CCI rats. (A) Withdrawal latency of unoperated (open histograms) and ligated (®lled histograms) hindpaw to radiant heat stimulus 1 h after oral administration of vehicle (n 10) or GV196771A (10 mg/kg, n 8). The glycine antagonist signi®cantly increases the latency of response for the ligated paw, but not for the unoperated paw. (B) Results expressed as changes in withdrawal latency between unoperated vs. ligated paw. GV196771A at 3 mg/kg (n 7) and 10 mg/kg (n 8), but not at 1 mg/ kg (n 10), show statistical differences compared to vehicle (# P , 0:05 and * P , 0:01, respectively).
222
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
suggesting a potent action of this agent on chronic pain, in agreement with behavioral results (Quartaroli et al., 1999). NMDA receptor antagonists reduce nociceptive-induced activity in the thalamus of both primates (Dougherty et al., 1996) and rats (Salt, 1987; Eaton and Salt, 1990). These antagonists also reduced non-nociceptive activity such as light innocuous brushing or touching of the skin, or movement of vibrissae by air jets. Our results with the NMDA channel blocker MK801 con®rm these ®ndings. The inhibitory effects of MK801 (0.1 mg/kg, iv) in both nociceptive and non-nociceptive mechanosensitive cells were similar as were the times for recovery from effects of the drug. The glycine antagonist GV196771A, on the other hand, did not affect either nociceptive or non-nociceptive transmission in normal rats. Although there are no published reports of the effects of glycine antagonists on nociception in the thalamus, D-serine, a glycine agonist, potentiates NMDA receptor-mediated responses in the ventrobasal thalamus, indicating that the glycine site is not normally saturated in these cells (Salt, 1989). This ®nding suggests that the different effects of MK801 vs. GV196771A are not the result of saturation of the glycine site, but rather represent distinct effects of glycine site vs. NMDA channel antagonists (see below). In contrast to the effects on thalamic neurons, NMDA receptor antagonists do not reduce responses of dorsal horn cells to innocuous stimuli, although they selectively inhibit responses to noxious stimuli (Davies and Watkins, 1983; Dickenson and Sullivan, 1990; Dougherty et al., 1992). Acute nociception, as measured by behavioral tests such as the hot plate model, is thought to be mediated principally in the spinal cord (Davies and Watkins, 1983). However, neither NMDA receptor antagonists nor glycine antagonists are effective in behavioral acute nociceptive models (Yaksh, 1989; Nasstrom et al., 1993; Nishiyama et al., 1998), suggesting that NMDA receptors in dorsal horn neurons do not play a major role in these responses. AMPA receptors, which mediate the monosynaptic excitation of dorsal horn neurons, have been implicated in the modulation of acute noxious stimulus (Dougherty et al., 1992; Hunter and Singh, 1994). In the present study, MK801 was effective in the hot plate test only at a dose (0.25 mg/kg, ip) that causes evident motor dysfunction, as previously reported with other NMDA antagonists (Hoffman, 1992; Murata and Kawasaki, 1993). Glycine antagonists, on the other hand, appear to have negligible motor behavior impairments (Chiamulera et al., 1990; Danysz et al., 1994; Kretschmer et al., 1997; Bordi et al., 1999). NMDA receptors play a key role in chronic pain states and hyperalgesia (Dickenson, 1990; Coderre, 1993), and NMDA receptor antagonists attenuate hyperalgesic responses (see Section 1). Thalamic neurons are involved in abnormal behavioral responses caused by peripheral injury and in¯ammation (Guilbaud et al., 1986; 1987, 1990). In the present study we observed marked afterdischarge and sensitization responses in VPL neurons in
rats with CCI. Spontaneous activity was not increased in these cells, but ®ring often continued for many minutes after a single stimulus. MK801 and GV196771A were both able to inhibit responses to noxious stimuli and to block sensitization. The behavioral results in rats with CCI con®rm the electrophysiological ®ndings and extend our previous evidence showing that the GV196771A blocks the development of hyperalgesia following CCI in rats and antagonizes the second phase of formalin-induced hyperalgesia in mice (Quartaroli et al., 1997, 1999). The ef®cacy of selective glycine antagonists in behavioral models of chronic nociception has been demonstrated previously (Vaccarino et al., 1993; Hunter and Singh, 1994; Laird et al., 1996). Although the site of action of GV196771A and MK801 was not investigated in this study, our data have shown that block of the NMDA receptor decreases nociceptive transmission in the thalamus and, together with spinal neurons, thalamic neurons may also be involved in chronic pain states and contribute to NMDA activation. Anatomical evidence for the presence of NMDA receptors in ascending sensory afferents to the thalamus or descending efferents from the somatosensory cortex to the thalamus is abundant (De Biasi et al., 1994; Salt and Eaton, 1996, for a review). The recent ®nding that NMDA receptor blockade in the thalamus inhibits behavioral hyperalgesic responses (Kolhekar et al., 1997) suggests that supraspinal and thalamic receptors may participate in the development and maintenance of central hyperalgesia following peripheral injury. Glycine site antagonists are attractive as potential therapeutic agents because they have few known side effects. The paucity of side effects is at ®rst counter intuitive, because the functional characteristics of these antagonists resemble those of NMDA antagonists. Occupancy of the glycine site is thought to be required for ef®cient activation of the NMDA channel by glutamate (Kleckner and Dingledine, 1988; Corsi et al., 1996). One possibility, suggested by Danysz and Parsons (1998) is that glycine and NMDA antagonists (or channel blockers, such as MK801) might have different selectivity for NMDA subtypes. Alternatively, glycine antagonists might induce receptor desensitization (Parsons et al., 1993) which could differentiate between various forms of NMDA activation. According to this hypothesis, NMDA receptor desensitization by glycine antagonists could block the transient physiological activation of glutamate receptors without activation of their longterm neurotoxic effects. These possibilities are currently under investigation using GV196771A. In conclusion, the present ®ndings suggest that NMDA receptors are involved in the processing of nociceptive information with different neuronal mechanisms depending on the type of pain involved. The glycine site of these receptors is important in modulating neuropathic pain and represents a potentially effective and safe target for the treatment of hyperalgesia.
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
Acknowledgements The authors wish to thank Mr Giorgio Tarter for technical assistance and Professor Eric Frank (University of Pittsburgh) for providing the software analysis and for his valuable comments on the manuscript.
References Aanonsen LM, Lei S, Wilcox GL. Excitatory amino acid receptors and nociceptive neurotransmission in rat spinal cord. Pain 1990;41:309± 321. Al-Chaer ED, Lawand NB, Westlund KN, Willis WD. Visceral nociceptive input into the ventral posterolateral nucleus of the thalamus: a new function for the dorsal column pathway. J Neurophysiol 1996;76:2661±2674. Angel A, Clarke KA. An analysis of the representation of the forelimb in the ventrobasal thalamic complex of the albino rat. J Physiol (London) 1975;249:399±423. Bennett GJ, Xie YK. A peripheral mononeuropathy in rat that produces disorders of pain sensation like those seen in man, see comments. Pain 1988;33:87±107. Benoist JM, Kayser V, Gautron M, Guilbaud G. Low dose of morphine strongly depresses responses of speci®c nociceptive neurones in the ventrobasal complex of the rat. Pain 1983;15:333±344. Bordi F, LeDoux JE. Response properties of single units in areas of rat auditory thalamus that project to the amygdala: II. Cells receiving convergent auditory and somatosensory inputs and cells antidromically activated by amygdala stimulation. Exp Brain Res 1994;98:275±286. Bordi F, Terron A, Reggiani A. The neuroprotective glycine antagonist GV150526 does not produce neuronal vacuolization or cognitive de®cits in rats. Eur J Pharmacol 1999;378:153±160. Bordi F, Pietra C, Ziviani L, Reggiani A. The glycine antagonist GV150526 protects somatosensory evoked potentials and reduces the infarct area in the MCAo model of focal ischemia in the rat. Exp Neurol 1997;145:425±433. Chiamulera C, Costa S, Reggiani A. Effect of NMDA- and strychnineinsensitive glycine site antagonists on NMDA-mediated convulsions and learning. Psychopharmacology (Berlin) 1990;102:551±552. Cicero TJ, Nock B, Meyer ER. Gender-related differences in the antinociceptive properties of morphine. J Pharmacol Exp Ther 1996;279:767± 773. Coderre TJ. The role of excitatory amino acid receptors and intracellular messengers in persistent nociception after tissue injury in rats. Mol Neurobiol 1993;7:229±246. Corsi M, Fina P, Trist DG. Co-agonism in drug-receptor interaction: illustrated by the NMDA receptors. Trends Pharmacol Sci 1996;17:220± 222. Cropper EC, Eisenman JS. Localization of responses in the somatosensory thalamus of the rat. Brain Res Bull 1986;16:83±91. Danysz W, Parsons CG. Glycine and NMDA receptors Ð Physiological signi®cance and possible therapeutic applications. Phrmacol Rev 1998;50:597±664. Danysz W, Essmann U, Bresink I, Wilke R. Glutamate antagonists have different effects on spontaneous locomotor activity in rats. Pharmacol Biochem Behav 1994;48:111±118. Davies J, Watkins JC. Role of excitatory amino acid receptors in mono- and polysynaptic excitation in the cat spinal cord. Exp Brain Res 1983;49:280±290. De Biasi S, Amadeo A, Sprea®co R, Rustioni A. Enrichment of glutamate immunoreactivity in lemniscal terminals in the ventropostero lateral thalamic nucleus of the rat: an immunogold and WGA-HRP study. Anat Rec 1994;240:131±140.
223
Dickenson AH. A cure for wind up: NMDA receptor antagonists as potential analgesics. Trends Pharmacol Sci 1990;11:307±309. Dickenson AH, Aydar E. Antagonism at the glycine site on the NMDA receptor reduces spinal nociception in the rat. Neurosci Lett 1991;121:263±266. Dickenson AH, Sullivan AF. Differential effects of excitatory amino acid antagonists on dorsal horn nociceptive neurones in the rat. Brain Res 1990;506:31±39. Dougherty PM, Palecek J, Paleckova V, Sorkin LS, Willis WD. The role of NMDA and non-NMDA excitatory amino acid receptors in the excitation of primate spinothalamic tract neurons by mechanical, chemical, thermal, and electrical stimuli. J Neurosci 1992;12:3025±3041. Dougherty PM, Li YJ, Lenz FA, Rowland L, Mittman S. Evidence that excitatory amino acids mediate afferent input to the primate somatosensory thalamus. Brain Res 1996;728:267±273. Dubner R, Ruda MA. Activity-dependent neuronal plasticity following tissue injury and in¯ammation. Trends Neurosci 1992;15:96±103. Eaton SA, Salt TE. Thalamic NMDA receptors and nociceptive sensory synaptic transmission. Neurosci Lett 1990;110:297±302. Giacobbe SA, Baraldi D, Di Fabio R. Unusual synthesis of new glycine antagonists via sequential aldol condensation-lactonization-elimination reaction. Bioorg Med Chem Lett 1998;8:1689±1692. Guilbaud G, Benoist JM, Gautron M. Electrophysiological evidence that morphine can exert an antinociceptive effect in a neuropathic state: a study in the ventrobasal thalamus of rats after moderate ligation of one sciatic nerve. Brain Res 1991;551:346±350. Guilbaud G, Benoist JM, Jazat F, Gautron M. Neuronal responsiveness in the ventrobasal thalamic complex of rats with an experimental peripheral mononeuropathy. J Neurophysiol 1990;64:1537±1554. Guilbaud G, Neil A, Benoist JM, Kayser V, Gautron M. Thresholds and encoding of neuronal responses to mechanical stimuli in the ventrobasal thalamus during carrageenin-induced hyperalgesic in¯ammation in the rat. Exp Brain Res 1987;68:311±318. Guilbaud G, Peschanski M, Briand A, Gautron M. The organization of spinal pathways to ventrobasal thalamus in an experimental model of pain (the arthritic rat). An electrophysiological study. Pain 1986;26:301±312. Haley JE, Sullivan AF, Dickenson AH. Evidence for spinal N-methyl-Daspartate receptor involvement in prolonged chemical nociception in the rat. Brain Res 1990;518:218±226. Hill RG, Pepper CM. Selective effects of morphine on the nociceptive responses of thalamic neurones in the rat. Br J Pharmacol 1978;64:137±143. Hoffman DC. Typical and atypical neuroleptics antagonize MK-801induced locomotion and stereotypy in rats. J Neural Transm Gen Sect 1992;89:1±10. Hollmann M, Heinemann S. Cloned glutamate receptors. Annu Rev Neurosci 1994;17:31±108. Hunter JC, Singh L. Role of excitatory amino acid receptors in the mediation of the nociceptive response to formalin in the rat. Neurosci Lett 1994;174:217±221. Johnson JW, Ascher P. Glycine potentiates the NMDA response in cultured mouse brain neurons. Nature 1987;325:529±531. Jurna I, Baldauf J, Fleischer W. Depression by morphine-6-glucuronide of nociceptive activity in rat thalamus neurons: comparison with morphine. Brain Res 1996;722:132±138. Kleckner NW, Dingledine R. Requirement for glycine in activation of NMDA-receptors expressed in Xenopus oocytes. Science 1988;241: 835±837. Kolhekar R, Murphy S, Gebhart GF. Thalamic NMDA receptors modulate in¯ammation-produced hyperalgesia in the rat. Pain 1997;71:31±40. Kretschmer BD, Kratzer U, Breithecker K, Koch M. ACEA 1021, a glycine site antagonist with minor psychotomimetic and amnestic effects in rats. Eur J Pharmacol 1997;331:109±116. Kristensen JD, Svensson B, Gordh Jr T. The NMDA-receptor antagonist CPP abolishes neurogenic `wind-up pain' after intrathecal administration in humans, see comments. Pain 1992;51:249±253.
224
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
Laird JM, Bennett GJ. Dorsal root potentials and afferent input to the spinal cord in rats with an experimental peripheral neuropathy. Brain Res 1992;584:181±190. Laird JM, Mason GS, Webb J, Hill RG, Hargreaves RJ. Effects of a partial agonist and a full antagonist acting at the glycine site of the NMDA receptor on in¯ammation-induced mechanical hyperalgesia in rats. Br J Pharmacol 1996;117:1487±1492. Mao J, Price DD, Mayer DJ, Lu J, Hayes RL. Intrathecal MK-801 and local nerve anesthesia synergistically reduce nociceptive behaviors in rats with experimental peripheral mononeuropathy. Brain Res 1992;576:254±262. Martin RJ, Apkarian AV, Hodge Jr CJ. Ventrolateral and dorsolateral ascending spinal cord pathway in¯uence on thalamic nociception in cat. J Neurophysiol 1990;64:1400±1412. Martin WJ, Hohmann AG, Walker JM. Suppression of noxious stimulusevoked activity in the ventral posterolateral nucleus of the thalamus by a cannabinoid agonist: correlation between electrophysiological and antinociceptive effects. J Neurosci 1996;16:6601±6611. Mendell LM. Physiological properties of unmyelinated ®ber projection to the spinal cord. Exp Neurol 1966;16:316±332. Millan MJ, Seguin L. Chemically-diverse ligands at the glycine B site coupled to N-methly-D- asparate (NMDA) receptors selectively block the late phase of formalin- induced pain in mice. Neurosci Lett 1994;178:139±143. Murata S, Kawasaki K. Common and uncommon behavioural effects of antagonists for different modulatory sites in the NMDA receptor/channel complex. Eur J Pharmacol 1993;239:9±15. Murray CW, Cowan A, Larson AA. Neurokinin and NMDA antagonists (but not a kainic acid antagonist) are antinociceptive in the mouse formalin model. Pain 1991;44:179±185. Nasstrom J, Karlsson U, Post C. Antinociceptive actions of different classes of excitatory amino acid receptor antagonists in mice. Eur J Pharmacol 1992;212:21±29. Nasstrom J, Karlsson U, Berge OG. Systemic or intracerebroventricular injection of NMDA receptor antagonists attenuates the antinociceptive activity of intrathecally administered NMDA receptor antagonists. Brain Res 1993;623:47±55. Neugebauer V, Lucke T, Schaible HG. N-methyl-D-aspartate (NMDA) and non-NMDA receptor antagonists block the hyperexcitability of dorsal horn neurons during development of acute arthritis in rat's knee joint. J Neurophysiol 1993;70:1365±1377. Nishiyama T, Yaksh TL, Weber E. Effects of intrathecal NMDA and nonNMDA antagonists on acute thermal nociception and their interaction with morphine. Anesthesiology 1998;89:715±722. Parsons CG, Zong X, Lux HD. Whole cell and single channel analysis of the kinetics of glycine- sensitive N-methyl-D-aspartate receptor desensitization. Br J Pharmacol 1993;109:213±221. Paxinos G, Watson C. The rat brain in stereotaxic coordinates, 2nd. Academic Press, 1986. Peschanski M, Mantyh PW, Besson JM. Spinal afferents to the ventrobasal thalamic complex in the rat: an anatomical study using wheat-germ
agglutinin conjugated to horseradish peroxidase. Brain Res 1983;278:240±244. Quartaroli M, Arban R, Bettelini L, Corsi M, DiFabio R, Giacobbe S, Pecunioso A, Reggiani A, Trist D, Ratti E. Antinociceptive activity of a novel NMDA receptor glycine antagonist: GV196771A. Soc Neurosci Abst 1997;23:943. Quartaroli M, Carignani C, DalForno G, Mugnaini M, Ugolini A, Arban R, Bettelini L, Maraia F, Belardetti F, Trist D, Ratti E, DiFabio R, Corsi M. Potent antihyperalgesic activity without tolerance produced by the glycine site antagonist of the NMDA receptor GV196771A. J Pharmacol Exp Ther 1999;290:158±169. Ren K, Dubner R. NMDA receptor antagonists attenuate mechanical hyperalgesia in rats with unilateral in¯ammation of the hindpaw. Neurosci Lett 1993;163:22±26. Ren K, Williams GM, Hylden JL, Ruda MA, Dubner R. The intrathecal administration of excitatory amino acid receptor antagonists selectively attenuated carrageenan-induced behavioral hyperalgesia in rats. Eur J Pharmacol 1992;219:235±243. Roberts WA, Wells J. Extensive dual innervation and mutual inhibition by forelimb and hindlimb inputs to ventroposterolateral nucleus projection neurons in the rat. Somatosens Mot Res 1990;7:85±95. Salt TE. Excitatory amino acid receptors and synaptic transmission in the rat ventrobasal thalamus. J Physiol (London) 1987;391:499±510. Salt TE. Modulation of NMDA receptor-mediated responses by glycine and D-serine in the rat thalamus in vivo. Brain Res 1989;481:403±406. Salt TE, Eaton SA. Functions of ionotropic and metabotropic glutamate receptors in sensory transmission in the mammalian thalamus. Prog Neurobiol 1996;48:55±72. Sherman SE, Luo L, Dostrovsky JO. Altered receptive ®elds and sensory modalities of rat VPL thalamic neurons during spinal strychnineinduced allodynia. J Neurophysiol 1997;78:2296±2308. Vaccarino AL, Marek P, Kest B, Weber E, Keana JF, Liebeskind JC. NMDA receptor antagonists MK-801 and ACEA-1011, prevent the development of tonic pain following subcutaneous formalin. Brain Res 1993;615:331±334. Willis WD, Coggeshall RE. Sensory mechanisms of the spinal cord, Plenum, 1991. Willis WD, Westlund KN. Neuroanatomy of the pain system and of the pathways that modulate pain. J Clin Neurophysiol 1997;14:2±31. Woolf CJ, Thompson SW. The induction and maintenance of central sensitization is dependent on N-methyl-D-aspartic acid receptor activation; implications for the treatment of post-injury pain hypersensitivity states. Pain 1991;44:293±299. Yaksh TL. Spinal opiate analgesia: characteristics and principles of action. Pain 1981;11:293±346. Yaksh TL. Behavioral and autonomic correlates of the tactile evoked allodynia produced by spinal glycine inhibition: effects of modulatory receptor systems and excitatory amino acid antagonists. Pain 1989;37:111±123. Yokota T, Asato F, Koyama N, Masuda T, Taguchi H. Nociceptive body representation in nucleus ventralis posterolateralis of cat thalamus. J Neurophysiol 1988;6:1714±1727.
Modulation of nociceptive transmission by NMDA/glycine site receptor in the ventroposterolateral nucleus of the thalamus Fabio Bordi*, Mauro Quartaroli Pharmacology Department, GlaxoWellcome Medicine Research Centre, 37100 Verona, Italy Received 12 March 1999; received in revised form 22 July 1999; accepted 3 August 1999
Abstract NMDA-type glutamate receptors are involved in the generation and maintenance of altered pain states. In the present study, we examined the effect of an NMDA-glycine site antagonist, GV196771A [E-4,6-dichloro-3-(2-oxo-1-phenyl-pyrrolidin-3-ylidenemethyl)-1H-indole-2carboxylic acid sodium salt], on responses to noxious stimuli both in normal rats and during peripheral mononeuropathy induced by chronic constriction injury (CCI) of the sciatic nerve. In one series of experiments, activity of nociceptive neurons in the ventroposterolateral (VPL) nucleus of the thalamus was recorded in response to pressure stimuli to the contralateral hindpaw. Intravenous injection (iv) of the glycine antagonist had no effect on these cells in normal rats. When tested in rats with CCI induced 2±3 weeks previously, however, GV196771A (0.125, 0.5 and 2.0 mg/kg) blocked responses to noxious stimulation in a dose-dependent and reversible manner. Morphine (0.5 mg/kg, iv) and the NMDA channel blocker MK801 (0.1 mg/kg, iv) suppressed noxious stimulus-evoked activity of VPL neurons in both normal and CCI-treated rats. MK801 also decreased the responses of non-nociceptive neurons to brush stimulation in both sets of animals, in contrast to the glycine antagonist which did not alter the responses of these cells. Similar results were obtained from a series of behavior experiments in which the latency for paw withdrawal from heat stimulation was measured in normal and CCI-treated rats. GV196771A (3 and 10 mg/kg) injected orally, reduced the hyperalgesic response in the treated rats but did not change the withdrawal latency in normal rats. Taken together, these ®ndings suggest that block of the NMDA receptor decreases nociceptive transmission in the thalamus and can modulate hyperalgesic states. GV196771A and glycine antagonists in general may represent innovative and safe agents for the treatment of neuropathic pain. q 2000 International Association for the Study of Pain. Published by Elsevier Science B.V. Keywords: NMDA; Glycine antagonist; Thalamus; Sensory transmission; CCI; Hot plate
1. Introduction The N-methyl-D-aspartate (NMDA) glutamate receptor subtype has an important role in mediating nociceptive processing. Several results from behavioral studies suggest that NMDA receptor antagonists have antinociceptive effects in a number of models of nociception, both in animals and man (Murray et al., 1991; Kristensen et al., 1992; Nasstrom et al., 1992; Mao et al., 1992; Ren et al., 1992; Ren and Dubner, 1993). Activation of spinal NMDA receptors has been shown to mediate these effects. Spinal NMDA receptors are critical for pain perception and transmission (Aanonsen et al., 1990; Woolf and Thompson, 1991; Dubner and Ruda, 1992), while NMDA antagonists reverse the increased responsiveness of spinal neurons induced by acute arthritis (Neugebauer et al., 1993), intra* Corresponding author. Pharmacology Department, GlaxoWellcome Research, Via Fleming 4, 37100 Verona, Italy. Tel.: 139-045-921-8845; fax: 139-045-921-8153. E-mail address: [email protected] (F. Bordi)
dermal injection of capsaicin (Dougherty et al., 1992), intraplantar injection of formalin (Haley et al., 1990), application of the irritant mustard oil (Woolf and Thompson, 1991), and experimental peripheral neuropathy (Laird and Bennett, 1992). The extent to which these effects are conserved throughout ascending sensory pathways is not known, however. Few studies have examined the contribution of NMDA receptors to the processing of the nociceptive input or in chronic pain models in areas of the CNS other than the spinal cord. Only some studies have characterized the response of thalamic neurons to peripheral stimuli showing the involvement of NMDA and non-NMDA receptors in somatosensory and nociceptive transmission (Salt, 1987; Eaton and Salt, 1990; Dougherty et al., 1996). The spinothalamic tract originates in the dorsal horn of the spinal cord and terminates in the ventroposterolateral (VPL) nucleus of the thalamus. VPL receives inputs from wide dynamic range neurons in the spinal cord, which encode the strength of noxious and non-noxious stimuli
0304-3959/00/$20.00 q 2000 International Association for the Study of Pain. Published by Elsevier Science B.V. PII: S 0304-395 9(99)00205-5
214
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
(Mendell, 1966; see Willis and Coggeshall, 1991, for a review). VPL neurons have receptive ®elds on a restricted area of the contralateral skin, making them well suited for the sensory-discriminative aspects of pain (Peschanski et al., 1983; Yokota et al., 1988; Martin et al., 1990; Willis and Westlund, 1997). Like their spinal counterparts, VPL nociceptive neurons sensitizes to mechanical and thermal stimuli after peripheral in¯ammation (Guilbaud et al., 1986, 1987) and peripheral neuropathy (Guilbaud et al., 1990). Recent evidence, showing that VPL is a major site of convergence of somatic and visceral inputs conveyed through the dorsal column (Al-Chaer et al., 1996), make this nucleus an important station for pain processing. The NMDA receptor complex belongs to the family of ligand-gated ion channels, but with unique features. This receptor is highly permeable to Ca 21, is under voltagedependent regulation of extracellular Mg 21, and requires the simultaneous presence of two agonists to be activated: glutamate itself and glycine (Johnson and Ascher, 1987; Kleckner and Dingledine, 1988; Hollmann and Heinemann, 1994). The use of agents that act at the glycine modulatory site appears to be more attractive clinically because these agents have fewer adverse CNS side effects (see Danysz and Parsons, 1998, for in-depth review). Glycine antagonists have been reported to be effective in reducing nociceptive responses to formalin injection in spinal cord neurons (Dickenson and Aydar, 1991; Millan and Seguin, 1994) and in reversing thermal hyperalgesia evoked by in¯ammation or peripheral neuropathy in behavioral experiments (Mao et al., 1992; Laird et al., 1996). The aim of the present study was to examine the role of the NMDA receptor in the processing of nociceptive stimuli, using an NMDA channel blocker, MK801, and the novel NMDA/glycine site antagonist GV196771A (Giacobbe et al., 1998; Quartaroli et al., 1999). To test the effects of these two different NMDA antagonists in a model of chronic pain, a chronic constrictive injury (CCI) was experimentally produced around the sciatic nerve in rats. Electrophysiological recordings were made of nociceptive responses of wide dynamic range neurons in the VPL nucleus of the thalamus of anesthetized rats. Behavioral procedures examined the analgesic properties of the GV196771A in a acute model of nociception, the hot plate test, and in a chronic model of nociception, the paw withdrawal to thermal stimulation in CCI animals.
2. Methods 2.1. Electrophysiological methods 2.1.1. Animals and surgical preparation Male Sprague±Dawley rats (n 63) weighing 300±400 g were anesthetized with urethane in three injections separated each by about 10 min (1.5 g/kg, ip, ®nal concentration) and placed in a Kopf stereotaxic frame adjusted so that
the surface of the skull was level between lambda and bregma. Body temperature was regulated at 37 ^ 18 C by means of a heating pad. Depth of anesthesia was monitored during the experiment by assessing responsivity to tail pinch and urethane was supplemented as needed. In most cases no extra anesthetic was necessary as the period of anesthesia lasted 4±5 h, determined by previous experiments (Bordi et al., 1997). Stereotaxic coordinates were measured from bregma and calculated using a brain atlas (Paxinos and Watson, 1986). A small hole was made over the cortex above VPL, and a stainless steel electrode (3±4 MV impedance at 1000 Hz; Frederick Haer & Co., Bowdoinham, ME) was inserted into VPL using an electronic micropositioner (Kopf Inst., Tujunga, CA). Neurons were identi®ed by spontaneous activity or by a light brush stimulus of the contralateral hindpaw using a wooden probe or brush. Neurons responding to the search stimulus were selected and their responses tested to a graded pressure stimulus applied to the receptive ®eld. Data were collected for 2 s before the stimulus onset, during the 5 s of stimulus presentation, and for an additional 5 s at the end of stimulus offset. A tail vein was cannulated for intravenous (iv) injections. Drugs were tested only once in each animal, at a single dose. Twenty-six rats used for the electrophysiological studies had an experimental neuropathy induced by previous ligature of the sciatic nerve. 2.1.2. Sciatic nerve ligation surgery Chronic constriction injury (CCI) was induced in rats under pentobarbital anesthesia (50 mg/kg), and the left common sciatic nerve was exposed. Proximal to the sciatic trifurcation, about 10 mm of nerve was freed of adhering tissue and 4 ligatures (3.0 chromic gut) were tied loosely around it at intervals of 1 mm (Bennett and Xie, 1988). As previously shown the animals developed hyperalgesia within 14 days (Quartaroli et al., 1999). Behavioral and electrophysiological experiments were performed 14±21 days after ligature. 2.1.3. Pain stimulation and recording A computer-controlled air cylinder was used to administer the pressure stimulus using an adaptation of the method described by Martin et al. (1996). Air pressure was controlled by a computer-driven Picospritzer II (General Valve, Fair®eld, NJ) and was measured on-line in the air cylinder with a pressure transducer (Spectramed BV, Bithoven, The Netherlands). The stimulus was 5 s in duration, rising continuously from zero to a peak of about 4.2 kg/ cm 2, applied to the probe itself, which was about 2 mm 2 in cross-section. The relation between air pressure and stimulus force was measured empirically. To determine the threshold to pain, a behavioral experiment was carried out. Ten rats were anesthetized with a dose of urethane (1 g/kg, ip) that reduced motor tone but did not suppress the animal's re¯exes. At this dose, the animals permitted insertion of the hind paw into the pressure device
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
but would also withdraw their limb in response to noxious stimulation applied to their hind paw by the air cylinder. Pressure stimuli were presented 8 times at 2 min intervals. The mean pressure at which the behavioral response occurred was calculated for each rat. The mean pressure eliciting limb withdrawal was 3:1 ^ 0:15 kg/cm 2. This level of stimulation was reached after about 3 s from stimulus onset; therefore the last 2 s of the pressure stimulus were above the pain threshold. Once a neuron was isolated and found responsive to light brushing, its baseline responses to the noxious pressure stimulus were determined by applying the stimulus 10 times at 1 min intervals. The drug or vehicle was then injected intravenously, and the stimulus was applied again to examine the effects of the treatment. Recovery from the effects of the drug were measured for the next hour, during which responses were tested every 2±3 min. Neuronal potentials were ampli®ed by an a.c. ampli®er (Fintronic model WDR 420, Derby, CT), band pass ®ltered (400±8000 Hz), and displayed on a dual-trace storage oscilloscope (Tektronix 5111A, Beaverton, OR) (Bordi and
215
LeDoux, 1994). A window discriminator was used to convert single action potentials to logic pulses, which were recorded (Digidata 1200, Axon Inst., Foster City, CA), displayed and stored to disk using customized software. Peristimulus histograms (PSTHs) and raster plots of each response were displayed at the end of each stimulus presentation on the computer screen. To obtain stimulus ®ring rates before and after drug treatment, ®ring rates for each neuron were measured at 9 levels of pressure (from 0 to 4.2 kg/cm 2). These mean ®ring rates and pressures were subjected to linear regression. Only neurons that exhibited stimulus±response functions with a correlation coef®cient of at least 0.5 were classi®ed as wide dynamic range (WDR) neurons and included in the study. In some cases, different pressure stimulus waveforms were delivered to determine how the responses of a WDR neuron were correlated to stimulus intensity (Fig. 1A). Non-nociceptive mechanoresponsive neurons (n 13) were also encountered in VPL. Because these neurons are silent or low rates or spontaneous activity, as described (Martin et al., 1996), a search stimulus (light brush) was
Fig. 1. (A) Example of PSTHs for a single VPL neuron showing correlation of the response with stimulus magnitude. For each stimulus waveform, the level of the applied pressure is shown in register with the PSTH (top), mean response of the 10 stimulus repetitions (middle), and raster plot showing the 10 trials (bottom). The stimulus±response function (pressure vs. ®ring frequency), together with the corresponding correlation coef®cient is shown to the right. A single action potential is shown in the inset at the top right. (B) Example of a typical response to the graded mechanical pressure used throughout this study. Top, level of the applied pressure; right, stimulus-response function of the cell. (C) Mean pre-drug stimulus±response function for all nociceptive wide dynamic range neurons (n 24). Dotted lines indicate mean pressure eliciting limb withdrawal in behavioral experiments (see Section 2).
216
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
used regularly during the electrode penetration. Once isolated, the neuron's response to light touch stimulation was examined. Tapping the skin within the receptive ®eld of the neuron produced a reliable activation of mechanoresponsive neurons. These neurons were de®ned as non-nociceptive because they did not respond to noxious pressure stimulus with a greater response than to the light brush stimulus. Usually they also habituated rapidly to the pressure stimulus. Responses were measured to light taps every 1±2 min for 6±10 times before drug infusion and the average calculated. The effect of the drug was examined 5±15 min post injection and in some cases for 1±2 h afterwards, to monitor recovery of the response. The average of the postinjection response was compared to pre-injection control. Only one cell was tested in a single animal. 2.1.4. Data analysis For each unit, the mean spontaneous ®ring rate was calculated during the 2 s interval immediately before each cutaneous stimulation. Mean evoked activity was calculated during the 5 s of pressure stimulus. The average of all units for each drug condition was used to construct the time course of the effect of the drug as fraction of the predrug response. The effects of drug treatment on stimulus± response functions of nociceptive neurons were determined by calculating the average rate of ®ring of the cell at 9 pressure levels before and after drug injection. A custom software program was used to construct PSTHs and individual raster plots as shown in Fig. 1A. Responses were compared using ANOVA measurements followed by Newman±Keuls post-hoc comparisons. Statistical differences for non-nociceptive cells were established using the Student's t test for unpaired samples. 2.1.5. Histology At the end of the experiment a small lesion was made at the recording site by applying an anodal current of 40± 50 mA for 5 s. Animals were then perfused with 10% buffered formalin, 5% potassium ferrocyanide, and 5% potassium ferricyanide (Bordi and LeDoux, 1994). The brains were frozen and cut on a sliding microtone (50 mm sections), and mounted sections were stained with cresyl violet. The position of the recording location was con®rmed microscopically. 2.2. Behavioral methods 2.2.1. Thermal hot plate test in naive animals Male Sprague±Dawley rats (n 44) weighing 250±350 g were used. They were housed for at least a week upon their arrival in groups of 2±3 per cage with food and water until the beginning of the experiment. The hot plate test was performed on an electrically heated and thermostatically controlled copper surface, set to a temperature of 55 ^ 0:28C. The animals were con®ned to the hot plate by a transparent observation chamber and the latency to the
response, consisting of licking of the hind paw, was measured (pre-test). A cut-off period of 30 s was used to avoid tissue damage. The animals were then treated orally with GV196771A (10 mg/kg) or vehicle and after 1 h the test was repeated. The effect of morphine (5 mg/kg, ip) and MK801 (0.125 mg/kg and 0.25 mg/kg, ip) was measured 30 min post injection. Statistics consisted in ANOVA measures followed by Dunnett's test. 2.2.2. Paw withdrawal in CCI animals Male Sprague±Dawley rats (n 30) weighing 300±350 g were used. They were housed two per cage with food and water for 14±21 days after CCI surgery. Paw withdrawal latencies to radiant heat were determined for both hindpaws as previously described (Bennett and Xie, 1988; Quartaroli et al., 1999). Brie¯y, the animals were placed in acrylic cages on a glass plate. After 15±20 min of accommodation, thermal hyperalgesia was tested using a commercially available analgesimeter (Plantar test, Ugo Basile, Comerio, Italy). A heat stimulus (50 W, 8 V) was applied directly onto the plantar surface of each hind paw and the withdrawal latency was determined. Four latency measurements were taken for each paw and averaged. GV196771A (1± 10 mg/kg) or vehicle was administered orally to animals 14±21 days after ligation and hyperalgesia tested 1 h post treatment. Statistical analysis was done comparing withdrawal latency time between unoperated vs. ligated paw for each group condition. Differential scores (latency of the unoperated paw minus latency of ligated paw) were also computed for each group and ANOVA was performed followed by Dunnett's test. 2.3. Drug treatment GV196771A [E-4,6-dichloro-3-(2-oxo-1-phenyl-pyrrolidin-3-ylidenemethyl)-1H-indole-2-carboxylic acid sodium salt] (synthesized by Medicinal Chemistry Dept., GlaxoWellcome, Verona) for the electrophysiological experiments was dissolved in DMSO (®nal concentration 1±2%) and diluted in saline. The vehicle control was 2% DMSO in saline. MK801 (RBI, Natick, MA) and morphine (Sigma, St. Louis, MO) were dissolved in saline. Drugs were administered in a volume of 1 ml/kg through the lateral vein. In the behavioral experiments, GV196771A was prepared as a stock solution of 1 mg/ml in 0.5% methylcellulose (methocel); further dilutions were made in 0.5% methocel. It was administered orally in a volume of 10 ml/kg. Morphine and MK801 were diluted in saline and injected ip in a volume of 2 ml/kg. 3. Results A total of 50 cells isolated in the VPL responded to noxious stimulation of the contralateral hindpaw and an additional 14 cells were activated by innocuous tactile stimulation. Cells were identi®ed on the basis of sponta-
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
neous activity or in response to cutaneous search stimuli (light brushing of the skin). Of these cells, 26 nociceptive and 6 non-nociceptive cells were found in animals that had an experimental neuropathy induced by CCI of the sciatic nerve. The recording sites of all units were located in the VPL region, as previously reported (Angel and Clarke, 1975; Cropper and Eisenman, 1986; Roberts and Wells, 1990; Martin et al., 1996; Sherman et al., 1997). 3.1. Effects of drug treatment on nociceptive neurons in normal animals The 24 VPL nociceptive neurons in normal animals had an average spontaneous background activity of 3.8^0.7 Hz. Their ®ring rate increased up to 20:9 ^ 1:2 Hz at the maximum pressure strength of the graded mechanical stimulus applied to their receptive ®eld (Fig. 1B,C). Responses were graded according to the intensity of the pressure stimulus, con®rming that these VPL neurons were wide dynamic range cells (Fig. 1A). The correlation coef®cient of response vs. stimulus intensity before drug injection was on average 0:86 ^ 0:18. The effects of intravenous morphine (0.5 mg/kg) on the responses of these neurons to the pressure stimulus was investigated in 4 neurons. Responses of a representative VPL neuron are shown in Fig. 2A. Morphine produced a marked depression of the stimulus-evoked response. The effect of morphine was statistically signi®cant at all stimu-
217
lus levels except for the two lowest pressures, suggesting that the effect was speci®c for nociceptive responses. The effect of morphine persisted for at least 1 h (Fig. 3C). The effect of NMDA receptor blockade on nociceptive function was tested using the NMDA blocker MK801. As shown in Fig. 3, MK801 has a potent antinociceptive action. At 0.1 mg/kg MK801 (n 5) was as effective a blocker as morphine (0.5 mg/kg), signi®cantly reducing the response at all pressures above 1 kg/cm 2. At 0.01 mg/kg (n 5), on the other hand, MK801 was effective only at higher levels of pressure stimulation (Fig. 3A). Two neurons were also tested at 1.0 mg/kg MK801, but in one case the animal died and in the other all activity was blocked (data not shown). There were dose-dependent differences in the length of time that MK801 was effective (Fig. 3C). The selective antagonist of the glycine site on NMDA receptors, GV196771A, was tested on 6 nociceptive cells in 6 normal rats. In contrast to the effects of morphine and MK801, GV196771A did not reduce the responsiveness of nociceptive cells even at high doses (2.0 mg/kg, Fig. 3B). Responses of cells after treatment with GV196771A were indistinguishable from control at all times after injection (data not shown). 3.2. Effects of drug treatment on non-nociceptive neurons of normal animals A total of 8 non-nociceptive neurons were identi®ed by
Fig. 2. Effects of morphine on nociceptive VPL neurons. (A) Evoked ®ring of a representative neuron before and after administration of morphine (0.5 mg/kg, iv). Top, level of applied pressure in register with the PSTHs below. Middle, average of predrug ®ring rate histograms and raster plot of the 10 individual trials. Bottom, PSTHs and raster plots after morphine injection. (B) Mean stimulus±response function after the administration of vehicle (®lled circles, n 4) or morphine (®lled triangles, n 4). Asterisks indicate statistical difference from vehicle at different times (P , 0:01, ANOVA followed by Newman±Keuls tests).
218
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
Fig. 3. Effects of MK801 and GV196771A on nociceptive VPL neurons. (A) Mean stimulus±response function after injection of vehicle (®lled circles, n 4), 0.01 mg/kg MK801 (open triangles, n 5), or 0.1 mg/kg MK801 (open squares, n 5). (B) Mean stimulus±response function after administration of vehicle (®lled circles, n 4) or GV196771A (®lled diamonds, n 6). No difference was found at any level of pressure applied to the contralateral paw. (C) Evoked ®ring over time before and after administration of the vehicle (®lled circles, n 4) or MK801 (0.01 mg/kg, open triangles, n 5); 0.1 mg/kg (open squares, n 5). Morphine (0.5 mg/kg, ®lled triangles, n 4) is shown for comparison. White arrow indicates injection time. * P , 0:01.
search stimulation. They all had no spontaneous activity and responded to light touch of the contralateral hindpaw. After intravenous injection of MK801 (0.1 mg/kg, n 5) or GV196771A (2.0 mg/kg, n 3) the response to nonnoxious stimulation was examined again. MK801 reduced the response to cutaneous stimulation as determined by counting the number of spikes elicited by brushing the skin (a light touch) (Fig. 4). The average decrease for the 5 cells was 44% (T5 21:7, P , 0:01). The glycine antagonist GV196771A, on the other hand, failed to suppress ®ring evoked by non-noxious stimulation; responses after treatment were 96% of the pre-drug response (T3 1:92 ns). 3.3. Behavioral effects of hot plate test The effect of GV196771A on acute nociceptive responses
Fig. 4. Activity of non-nociceptive neurons after the intravenous administration of MK801 (0.1 mg/kg, n 5) or GV196771A (2.0 mg/kg, n 3), expressed as percent of pre-drug control value. MK801, but not GV196771A, signi®cantly reduces response to light brush stimulation (t 21:7, P , 0:01).
was also tested by measuring the latency of hindpaw licking in response to heating the paw (the hot plate test; see Section 2). As shown in Fig. 5, GV196771A infused orally (10 mg/ kg, n 10) did not change the response, whereas morphine (5 mg/kg, ip, n 9) produced a signi®cant increase in the latency of the hindpaw licking response compared to control (n 10), as previously described (Yaksh, 1981; Cicero et al., 1996). Some animals were given either 0.125 mg/kg (n 6) or 0.250 mg/kg (n 9) of MK801. While the lower dose did not have any effect, the higher dose of MK801 signi®cantly increased the time of the occurrence of the ®rst hindpaw licking (data not shown). Psychomimetic effects, such as locomotor activity, stereotypic sniffing and circling, were elicited by this dose however, as described (Hoffman, 1992; Murata and Kawasaki, 1993). These effects confounded the response and may have masked the possible anti-nociceptive activity.
Fig. 5. Mean time of occurrence of ®rst hindpaw licking to hot plate stimulation (558C) after injection of vehicle (n 10), GV196771A (10 mg/kg, p.o., n 10), or morphine (5 mg/kg, ip, n 9). Asterisks denote cases signi®cantly (P , 0:01) different from vehicle control (Dunnett's test).
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
3.4. Effects of drug treatment on nociceptive neurons in CCI animals A total of 26 nociceptive neurons were recorded in the VPL of animals with experimental neuropathy induced by loose constriction of the contralateral sciatic nerve. Their average spontaneous activity was unchanged compared to control animals (4:2 ^ 0:4 Hz vs. 3:8 ^ 0:7 Hz). A major
219
difference in the response of cells in CCI animals was the presence of a long afterdischarge following the pressure stimulus (Fig. 6A). This afterdischarge was present in 61% (16/26) of the neurons, compared to only 8% (2/24) in control animals. In some neurons in CCI rats, more than 1 min was required between stimuli for the activity to return to baseline, and in some cases the interval between successive stimulus presentations was increased to 2±3 min. Even
Fig. 6. (A) Responses of a nociceptive VPL neuron in a CCI animal to graded pressure stimuli. Top, pressure applied to the paw. Middle, mean response of the 10 stimulus trials. Bottom, raster plot of the 10 individual trials. (B) Mean predrug stimulus±response function for all nociceptive neurons recorded in CCI rats (n 26). (C) Effect of morphine in a representative neuron. Top, pressure applied to the paw. Average PSTHs and individual raster plots before (middle) and after (bottom) injection of morphine (0.5 mg/kg, iv). (D) Mean stimulus±response function after the administration of vehicle (®lled circles, n 4) or morphine (®lled triangles, n 3). * P , 0:01.
220
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
then, the activity in some cells did not recover completely, as illustrated in the raster plots of Fig. 6C where the interstimulus interval was 3 min. The average responses of all 26 nociceptive CCI-neurons are summarized in Fig. 6B, which shows a strong correlation (r 0:89 ^ 0:14) between response and stimulus magnitudes. This relationship was very similar to the one observed for nociceptive cells in normal animals (compare Figs. 1C and 6B). As in normal animals, morphine (0.5 mg/kg, iv, n 3)
dramatically reduced the response of the neurons to pressure stimulation (Fig. 6C,D). Morphine was also effective in blocking the sensitization phenomena, as shown in a representative neuron in Fig. 6C. Similar to normal animals, MK801 (0.1 mg/kg, n 3) signi®cantly suppressed nociceptive activity in CCI animals (data not shown, see Fig. 3 for MK801 effect on normal animals). The time course of this effect was also similar to normal animals, with about 70% recovery after 1 h (Fig. 7C ). In contrast to its lack of effect in normal animals, the antagonist GV196771A produced a large, dose-dependent depression of nociceptive responses in CCI animals (Fig. 7). At the highest dose (2.0 mg/kg, n 6), blockade was complete, just as for morphine, but even at 0.5 mg/kg (n 8), a signi®cant reduction was observed (Fig. 7B). The lowest dose (0.125 mg/kg, n 3), did not have a significant effect, however (Fig. 7B). Fig. 7C shows that maximal effects occurred within 15 min after drug injection and had largely recovered (,70%) after 1 h. The action of GV196771A at 2.0 mg/kg on evoked ®ring over time was very similar to that of MK801 (0.1 mg/kg); the striking difference was that GV196771A had this effect only in CCI animals whereas MK801 blocked nociceptive responses in normal animals as well. 3.5. Effects of drug treatment on non-nociceptive neurons of CCI animals Six neurons responding to light touch cutaneous stimulation were isolated to test the effects of MK801 or GV196771A on non-nociceptive mechanosensitive neurons. Injection of MK801 (0.1 mg/kg) signi®cantly depressed responsiveness (n 3, mean 50% of pre-drug control, T3 3:9, P , 0:05), whereas GV196771A injection (2.0 mg/kg) had no effect (n 3, mean 113% of control, T3 1:7, ns). Fig. 8 shows responding to the light mechanical stimulation after MK801 or GV196771A intravenous infusion. 3.6. Behavioral effects of heat stimulation in CCI animals
Fig. 7. Decrease in the responsiveness of VPL neurons in CCI animals to noxious pressure stimuli after administration of GV196771A. (A) Effect of 2.0 mg/kg GV196771A on nociceptive responses of a representative cell. (B) Mean stimulus±response functions after intravenous administration of vehicle (®lled circles, n 4) or GV196771A (0.125 mg/kg, ®lled triangles, n 3; 0.5 mg/kg, ®lled squares, n 8; 2.0 mg/kg, ®lled diamonds, n 6). (C) Evoked ®ring over time before and after administration of the vehicle (®lled circles, n 4) or GV196771A (0.125 mg/kg, ®lled triangles, n 3; 0.5 mg/kg, ®lled squares, n 8; 2.0 mg/kg, ®lled diamonds, n 6). Effect of the NMDA blocker MK801 (0.1 mg/kg, open squares, n 3) is shown for comparison. * P , 0:01.
Prior to surgery, no difference was detected between withdrawal latencies of left vs. right paws in response to heat, both for vehicle- and GV196771A-treated animals (data not shown). Two to three weeks following surgery, vehicle-treated animals showed a decrease in thermally induced paw withdrawal in the operated paw. GV196771A reduced the latency difference between operated and control paws in a dose-dependent manner, measured 1 h after oral administration of vehicle (n 10) or GV196771A (1 mg/kg, n 10; 3 mg/kg, n 7; 10 mg/ kg, n 8). GV196771A increased the withdrawal latency for the ligated paw, without affecting the unoperated paw. Fig. 9A shows the effects of the highest dose of GV196771A (10 mg/kg, n 8). Fig. 9B shows the differential scores between the ligated and the unoperated paw. The two highest doses of GV196771A (3 and 10 mg/kg) exhibited a
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
221
tive responses in VPL neurons of the rat thalamus and reduces noxious responses as measured behaviorally. The effects of GV196771A were potent, dose-dependent and reversible. GV196771A was also highly selective for neuropathic nociceptive stimulation, because it did not alter the ®ring either of nociceptive neurons in normal rats or of nonnociceptive neurons responding to light touch. In contrast, the NMDA antagonist MK801 reduced nociceptive responses in both CCI and normal animals. Furthermore, MK801 was less selective because it also signi®cantly depressed ®ring of non-nociceptive neurons, both in normal and CCI-treated animals. Morphine also inhibited noxious stimulus-evoked responses in VPL neurons of normal rats, con®rming earlier results (Hill and Pepper, 1978; Benoist et al., 1983; Jurna et al., 1996). In our study, the amount of reduction of the response in VPL neurons (,80%) is quite similar to that reported by Jurna et al. (1996) after iv injection of morphine. Interestingly, morphine is capable of producing a maximum depression of only , 50±60% of control when injected by intrathecal administration, suggesting that spinal cord alone cannot achieve maximum analgesia (Jurna et al., 1996). The effects of MK801 on nociceptive neurons were quite similar to those of morphine, both in magnitude and duration. MK801 too, almost completely inhibited noxious stimulus-evoked response. Morphine also had a strong antihyperalgesic effect on VPL nociceptive neurons after neuropathy induced by CCI. This result is consistent with a previous report that showed depression of thalamic neuronal responses to pinch stimulation of the ligated hindpaw after morphine injection (Guilbaud et al., 1991). The glycine antagonist GV196771A at the highest dose produced effects very similar to those of morphine (see Figs. 7 and 8), Fig. 8. Activity of non-nociceptive neurons in CCI-operated animals after intravenous administration of MK801 (n 3) or GV196771A (n 3). (A) Firing rate histogram of a representative mechanosensitive non-nociceptive neuron before (top) and after (bottom) administration of 2.0 mg/kg GV196771A. The glycine antagonist does not reduce the responses of the neuron to tap stimuli (s) delivered every minute. (B) Responses of a nonnociceptive neuron are reduced after administration of MK801 (0.1 mg/kg). (C) Mean of all experiments expressed as percentage of pre-drug control responses. MK801, but not GV196771A, signi®cantly reduces responses to light brush stimulation (t 3:9, P , 0:05).
signi®cant difference vs. vehicle control (P , 0:05 and P , 0:01, respectively), indicating that the drug effectively reduces the behavioral noxious response. No statistical difference was found at the lowest dose (1 mg/kg). 4. Discussion The present study demonstrates that in a neuropathic pain state induced experimentally by chronic constriction of the sciatic nerve (CCI), the NMDA/glycine site antagonist GV196771A inhibits stimulus-evoked activity of nocicep-
Fig. 9. Dose-dependent effect of GV196771A on thermal hyperalgesia in CCI rats. (A) Withdrawal latency of unoperated (open histograms) and ligated (®lled histograms) hindpaw to radiant heat stimulus 1 h after oral administration of vehicle (n 10) or GV196771A (10 mg/kg, n 8). The glycine antagonist signi®cantly increases the latency of response for the ligated paw, but not for the unoperated paw. (B) Results expressed as changes in withdrawal latency between unoperated vs. ligated paw. GV196771A at 3 mg/kg (n 7) and 10 mg/kg (n 8), but not at 1 mg/ kg (n 10), show statistical differences compared to vehicle (# P , 0:05 and * P , 0:01, respectively).
222
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
suggesting a potent action of this agent on chronic pain, in agreement with behavioral results (Quartaroli et al., 1999). NMDA receptor antagonists reduce nociceptive-induced activity in the thalamus of both primates (Dougherty et al., 1996) and rats (Salt, 1987; Eaton and Salt, 1990). These antagonists also reduced non-nociceptive activity such as light innocuous brushing or touching of the skin, or movement of vibrissae by air jets. Our results with the NMDA channel blocker MK801 con®rm these ®ndings. The inhibitory effects of MK801 (0.1 mg/kg, iv) in both nociceptive and non-nociceptive mechanosensitive cells were similar as were the times for recovery from effects of the drug. The glycine antagonist GV196771A, on the other hand, did not affect either nociceptive or non-nociceptive transmission in normal rats. Although there are no published reports of the effects of glycine antagonists on nociception in the thalamus, D-serine, a glycine agonist, potentiates NMDA receptor-mediated responses in the ventrobasal thalamus, indicating that the glycine site is not normally saturated in these cells (Salt, 1989). This ®nding suggests that the different effects of MK801 vs. GV196771A are not the result of saturation of the glycine site, but rather represent distinct effects of glycine site vs. NMDA channel antagonists (see below). In contrast to the effects on thalamic neurons, NMDA receptor antagonists do not reduce responses of dorsal horn cells to innocuous stimuli, although they selectively inhibit responses to noxious stimuli (Davies and Watkins, 1983; Dickenson and Sullivan, 1990; Dougherty et al., 1992). Acute nociception, as measured by behavioral tests such as the hot plate model, is thought to be mediated principally in the spinal cord (Davies and Watkins, 1983). However, neither NMDA receptor antagonists nor glycine antagonists are effective in behavioral acute nociceptive models (Yaksh, 1989; Nasstrom et al., 1993; Nishiyama et al., 1998), suggesting that NMDA receptors in dorsal horn neurons do not play a major role in these responses. AMPA receptors, which mediate the monosynaptic excitation of dorsal horn neurons, have been implicated in the modulation of acute noxious stimulus (Dougherty et al., 1992; Hunter and Singh, 1994). In the present study, MK801 was effective in the hot plate test only at a dose (0.25 mg/kg, ip) that causes evident motor dysfunction, as previously reported with other NMDA antagonists (Hoffman, 1992; Murata and Kawasaki, 1993). Glycine antagonists, on the other hand, appear to have negligible motor behavior impairments (Chiamulera et al., 1990; Danysz et al., 1994; Kretschmer et al., 1997; Bordi et al., 1999). NMDA receptors play a key role in chronic pain states and hyperalgesia (Dickenson, 1990; Coderre, 1993), and NMDA receptor antagonists attenuate hyperalgesic responses (see Section 1). Thalamic neurons are involved in abnormal behavioral responses caused by peripheral injury and in¯ammation (Guilbaud et al., 1986; 1987, 1990). In the present study we observed marked afterdischarge and sensitization responses in VPL neurons in
rats with CCI. Spontaneous activity was not increased in these cells, but ®ring often continued for many minutes after a single stimulus. MK801 and GV196771A were both able to inhibit responses to noxious stimuli and to block sensitization. The behavioral results in rats with CCI con®rm the electrophysiological ®ndings and extend our previous evidence showing that the GV196771A blocks the development of hyperalgesia following CCI in rats and antagonizes the second phase of formalin-induced hyperalgesia in mice (Quartaroli et al., 1997, 1999). The ef®cacy of selective glycine antagonists in behavioral models of chronic nociception has been demonstrated previously (Vaccarino et al., 1993; Hunter and Singh, 1994; Laird et al., 1996). Although the site of action of GV196771A and MK801 was not investigated in this study, our data have shown that block of the NMDA receptor decreases nociceptive transmission in the thalamus and, together with spinal neurons, thalamic neurons may also be involved in chronic pain states and contribute to NMDA activation. Anatomical evidence for the presence of NMDA receptors in ascending sensory afferents to the thalamus or descending efferents from the somatosensory cortex to the thalamus is abundant (De Biasi et al., 1994; Salt and Eaton, 1996, for a review). The recent ®nding that NMDA receptor blockade in the thalamus inhibits behavioral hyperalgesic responses (Kolhekar et al., 1997) suggests that supraspinal and thalamic receptors may participate in the development and maintenance of central hyperalgesia following peripheral injury. Glycine site antagonists are attractive as potential therapeutic agents because they have few known side effects. The paucity of side effects is at ®rst counter intuitive, because the functional characteristics of these antagonists resemble those of NMDA antagonists. Occupancy of the glycine site is thought to be required for ef®cient activation of the NMDA channel by glutamate (Kleckner and Dingledine, 1988; Corsi et al., 1996). One possibility, suggested by Danysz and Parsons (1998) is that glycine and NMDA antagonists (or channel blockers, such as MK801) might have different selectivity for NMDA subtypes. Alternatively, glycine antagonists might induce receptor desensitization (Parsons et al., 1993) which could differentiate between various forms of NMDA activation. According to this hypothesis, NMDA receptor desensitization by glycine antagonists could block the transient physiological activation of glutamate receptors without activation of their longterm neurotoxic effects. These possibilities are currently under investigation using GV196771A. In conclusion, the present ®ndings suggest that NMDA receptors are involved in the processing of nociceptive information with different neuronal mechanisms depending on the type of pain involved. The glycine site of these receptors is important in modulating neuropathic pain and represents a potentially effective and safe target for the treatment of hyperalgesia.
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
Acknowledgements The authors wish to thank Mr Giorgio Tarter for technical assistance and Professor Eric Frank (University of Pittsburgh) for providing the software analysis and for his valuable comments on the manuscript.
References Aanonsen LM, Lei S, Wilcox GL. Excitatory amino acid receptors and nociceptive neurotransmission in rat spinal cord. Pain 1990;41:309± 321. Al-Chaer ED, Lawand NB, Westlund KN, Willis WD. Visceral nociceptive input into the ventral posterolateral nucleus of the thalamus: a new function for the dorsal column pathway. J Neurophysiol 1996;76:2661±2674. Angel A, Clarke KA. An analysis of the representation of the forelimb in the ventrobasal thalamic complex of the albino rat. J Physiol (London) 1975;249:399±423. Bennett GJ, Xie YK. A peripheral mononeuropathy in rat that produces disorders of pain sensation like those seen in man, see comments. Pain 1988;33:87±107. Benoist JM, Kayser V, Gautron M, Guilbaud G. Low dose of morphine strongly depresses responses of speci®c nociceptive neurones in the ventrobasal complex of the rat. Pain 1983;15:333±344. Bordi F, LeDoux JE. Response properties of single units in areas of rat auditory thalamus that project to the amygdala: II. Cells receiving convergent auditory and somatosensory inputs and cells antidromically activated by amygdala stimulation. Exp Brain Res 1994;98:275±286. Bordi F, Terron A, Reggiani A. The neuroprotective glycine antagonist GV150526 does not produce neuronal vacuolization or cognitive de®cits in rats. Eur J Pharmacol 1999;378:153±160. Bordi F, Pietra C, Ziviani L, Reggiani A. The glycine antagonist GV150526 protects somatosensory evoked potentials and reduces the infarct area in the MCAo model of focal ischemia in the rat. Exp Neurol 1997;145:425±433. Chiamulera C, Costa S, Reggiani A. Effect of NMDA- and strychnineinsensitive glycine site antagonists on NMDA-mediated convulsions and learning. Psychopharmacology (Berlin) 1990;102:551±552. Cicero TJ, Nock B, Meyer ER. Gender-related differences in the antinociceptive properties of morphine. J Pharmacol Exp Ther 1996;279:767± 773. Coderre TJ. The role of excitatory amino acid receptors and intracellular messengers in persistent nociception after tissue injury in rats. Mol Neurobiol 1993;7:229±246. Corsi M, Fina P, Trist DG. Co-agonism in drug-receptor interaction: illustrated by the NMDA receptors. Trends Pharmacol Sci 1996;17:220± 222. Cropper EC, Eisenman JS. Localization of responses in the somatosensory thalamus of the rat. Brain Res Bull 1986;16:83±91. Danysz W, Parsons CG. Glycine and NMDA receptors Ð Physiological signi®cance and possible therapeutic applications. Phrmacol Rev 1998;50:597±664. Danysz W, Essmann U, Bresink I, Wilke R. Glutamate antagonists have different effects on spontaneous locomotor activity in rats. Pharmacol Biochem Behav 1994;48:111±118. Davies J, Watkins JC. Role of excitatory amino acid receptors in mono- and polysynaptic excitation in the cat spinal cord. Exp Brain Res 1983;49:280±290. De Biasi S, Amadeo A, Sprea®co R, Rustioni A. Enrichment of glutamate immunoreactivity in lemniscal terminals in the ventropostero lateral thalamic nucleus of the rat: an immunogold and WGA-HRP study. Anat Rec 1994;240:131±140.
223
Dickenson AH. A cure for wind up: NMDA receptor antagonists as potential analgesics. Trends Pharmacol Sci 1990;11:307±309. Dickenson AH, Aydar E. Antagonism at the glycine site on the NMDA receptor reduces spinal nociception in the rat. Neurosci Lett 1991;121:263±266. Dickenson AH, Sullivan AF. Differential effects of excitatory amino acid antagonists on dorsal horn nociceptive neurones in the rat. Brain Res 1990;506:31±39. Dougherty PM, Palecek J, Paleckova V, Sorkin LS, Willis WD. The role of NMDA and non-NMDA excitatory amino acid receptors in the excitation of primate spinothalamic tract neurons by mechanical, chemical, thermal, and electrical stimuli. J Neurosci 1992;12:3025±3041. Dougherty PM, Li YJ, Lenz FA, Rowland L, Mittman S. Evidence that excitatory amino acids mediate afferent input to the primate somatosensory thalamus. Brain Res 1996;728:267±273. Dubner R, Ruda MA. Activity-dependent neuronal plasticity following tissue injury and in¯ammation. Trends Neurosci 1992;15:96±103. Eaton SA, Salt TE. Thalamic NMDA receptors and nociceptive sensory synaptic transmission. Neurosci Lett 1990;110:297±302. Giacobbe SA, Baraldi D, Di Fabio R. Unusual synthesis of new glycine antagonists via sequential aldol condensation-lactonization-elimination reaction. Bioorg Med Chem Lett 1998;8:1689±1692. Guilbaud G, Benoist JM, Gautron M. Electrophysiological evidence that morphine can exert an antinociceptive effect in a neuropathic state: a study in the ventrobasal thalamus of rats after moderate ligation of one sciatic nerve. Brain Res 1991;551:346±350. Guilbaud G, Benoist JM, Jazat F, Gautron M. Neuronal responsiveness in the ventrobasal thalamic complex of rats with an experimental peripheral mononeuropathy. J Neurophysiol 1990;64:1537±1554. Guilbaud G, Neil A, Benoist JM, Kayser V, Gautron M. Thresholds and encoding of neuronal responses to mechanical stimuli in the ventrobasal thalamus during carrageenin-induced hyperalgesic in¯ammation in the rat. Exp Brain Res 1987;68:311±318. Guilbaud G, Peschanski M, Briand A, Gautron M. The organization of spinal pathways to ventrobasal thalamus in an experimental model of pain (the arthritic rat). An electrophysiological study. Pain 1986;26:301±312. Haley JE, Sullivan AF, Dickenson AH. Evidence for spinal N-methyl-Daspartate receptor involvement in prolonged chemical nociception in the rat. Brain Res 1990;518:218±226. Hill RG, Pepper CM. Selective effects of morphine on the nociceptive responses of thalamic neurones in the rat. Br J Pharmacol 1978;64:137±143. Hoffman DC. Typical and atypical neuroleptics antagonize MK-801induced locomotion and stereotypy in rats. J Neural Transm Gen Sect 1992;89:1±10. Hollmann M, Heinemann S. Cloned glutamate receptors. Annu Rev Neurosci 1994;17:31±108. Hunter JC, Singh L. Role of excitatory amino acid receptors in the mediation of the nociceptive response to formalin in the rat. Neurosci Lett 1994;174:217±221. Johnson JW, Ascher P. Glycine potentiates the NMDA response in cultured mouse brain neurons. Nature 1987;325:529±531. Jurna I, Baldauf J, Fleischer W. Depression by morphine-6-glucuronide of nociceptive activity in rat thalamus neurons: comparison with morphine. Brain Res 1996;722:132±138. Kleckner NW, Dingledine R. Requirement for glycine in activation of NMDA-receptors expressed in Xenopus oocytes. Science 1988;241: 835±837. Kolhekar R, Murphy S, Gebhart GF. Thalamic NMDA receptors modulate in¯ammation-produced hyperalgesia in the rat. Pain 1997;71:31±40. Kretschmer BD, Kratzer U, Breithecker K, Koch M. ACEA 1021, a glycine site antagonist with minor psychotomimetic and amnestic effects in rats. Eur J Pharmacol 1997;331:109±116. Kristensen JD, Svensson B, Gordh Jr T. The NMDA-receptor antagonist CPP abolishes neurogenic `wind-up pain' after intrathecal administration in humans, see comments. Pain 1992;51:249±253.
224
F. Bordi, M. Quartaroli / Pain 84 (2000) 213±224
Laird JM, Bennett GJ. Dorsal root potentials and afferent input to the spinal cord in rats with an experimental peripheral neuropathy. Brain Res 1992;584:181±190. Laird JM, Mason GS, Webb J, Hill RG, Hargreaves RJ. Effects of a partial agonist and a full antagonist acting at the glycine site of the NMDA receptor on in¯ammation-induced mechanical hyperalgesia in rats. Br J Pharmacol 1996;117:1487±1492. Mao J, Price DD, Mayer DJ, Lu J, Hayes RL. Intrathecal MK-801 and local nerve anesthesia synergistically reduce nociceptive behaviors in rats with experimental peripheral mononeuropathy. Brain Res 1992;576:254±262. Martin RJ, Apkarian AV, Hodge Jr CJ. Ventrolateral and dorsolateral ascending spinal cord pathway in¯uence on thalamic nociception in cat. J Neurophysiol 1990;64:1400±1412. Martin WJ, Hohmann AG, Walker JM. Suppression of noxious stimulusevoked activity in the ventral posterolateral nucleus of the thalamus by a cannabinoid agonist: correlation between electrophysiological and antinociceptive effects. J Neurosci 1996;16:6601±6611. Mendell LM. Physiological properties of unmyelinated ®ber projection to the spinal cord. Exp Neurol 1966;16:316±332. Millan MJ, Seguin L. Chemically-diverse ligands at the glycine B site coupled to N-methly-D- asparate (NMDA) receptors selectively block the late phase of formalin- induced pain in mice. Neurosci Lett 1994;178:139±143. Murata S, Kawasaki K. Common and uncommon behavioural effects of antagonists for different modulatory sites in the NMDA receptor/channel complex. Eur J Pharmacol 1993;239:9±15. Murray CW, Cowan A, Larson AA. Neurokinin and NMDA antagonists (but not a kainic acid antagonist) are antinociceptive in the mouse formalin model. Pain 1991;44:179±185. Nasstrom J, Karlsson U, Post C. Antinociceptive actions of different classes of excitatory amino acid receptor antagonists in mice. Eur J Pharmacol 1992;212:21±29. Nasstrom J, Karlsson U, Berge OG. Systemic or intracerebroventricular injection of NMDA receptor antagonists attenuates the antinociceptive activity of intrathecally administered NMDA receptor antagonists. Brain Res 1993;623:47±55. Neugebauer V, Lucke T, Schaible HG. N-methyl-D-aspartate (NMDA) and non-NMDA receptor antagonists block the hyperexcitability of dorsal horn neurons during development of acute arthritis in rat's knee joint. J Neurophysiol 1993;70:1365±1377. Nishiyama T, Yaksh TL, Weber E. Effects of intrathecal NMDA and nonNMDA antagonists on acute thermal nociception and their interaction with morphine. Anesthesiology 1998;89:715±722. Parsons CG, Zong X, Lux HD. Whole cell and single channel analysis of the kinetics of glycine- sensitive N-methyl-D-aspartate receptor desensitization. Br J Pharmacol 1993;109:213±221. Paxinos G, Watson C. The rat brain in stereotaxic coordinates, 2nd. Academic Press, 1986. Peschanski M, Mantyh PW, Besson JM. Spinal afferents to the ventrobasal thalamic complex in the rat: an anatomical study using wheat-germ
agglutinin conjugated to horseradish peroxidase. Brain Res 1983;278:240±244. Quartaroli M, Arban R, Bettelini L, Corsi M, DiFabio R, Giacobbe S, Pecunioso A, Reggiani A, Trist D, Ratti E. Antinociceptive activity of a novel NMDA receptor glycine antagonist: GV196771A. Soc Neurosci Abst 1997;23:943. Quartaroli M, Carignani C, DalForno G, Mugnaini M, Ugolini A, Arban R, Bettelini L, Maraia F, Belardetti F, Trist D, Ratti E, DiFabio R, Corsi M. Potent antihyperalgesic activity without tolerance produced by the glycine site antagonist of the NMDA receptor GV196771A. J Pharmacol Exp Ther 1999;290:158±169. Ren K, Dubner R. NMDA receptor antagonists attenuate mechanical hyperalgesia in rats with unilateral in¯ammation of the hindpaw. Neurosci Lett 1993;163:22±26. Ren K, Williams GM, Hylden JL, Ruda MA, Dubner R. The intrathecal administration of excitatory amino acid receptor antagonists selectively attenuated carrageenan-induced behavioral hyperalgesia in rats. Eur J Pharmacol 1992;219:235±243. Roberts WA, Wells J. Extensive dual innervation and mutual inhibition by forelimb and hindlimb inputs to ventroposterolateral nucleus projection neurons in the rat. Somatosens Mot Res 1990;7:85±95. Salt TE. Excitatory amino acid receptors and synaptic transmission in the rat ventrobasal thalamus. J Physiol (London) 1987;391:499±510. Salt TE. Modulation of NMDA receptor-mediated responses by glycine and D-serine in the rat thalamus in vivo. Brain Res 1989;481:403±406. Salt TE, Eaton SA. Functions of ionotropic and metabotropic glutamate receptors in sensory transmission in the mammalian thalamus. Prog Neurobiol 1996;48:55±72. Sherman SE, Luo L, Dostrovsky JO. Altered receptive ®elds and sensory modalities of rat VPL thalamic neurons during spinal strychnineinduced allodynia. J Neurophysiol 1997;78:2296±2308. Vaccarino AL, Marek P, Kest B, Weber E, Keana JF, Liebeskind JC. NMDA receptor antagonists MK-801 and ACEA-1011, prevent the development of tonic pain following subcutaneous formalin. Brain Res 1993;615:331±334. Willis WD, Coggeshall RE. Sensory mechanisms of the spinal cord, Plenum, 1991. Willis WD, Westlund KN. Neuroanatomy of the pain system and of the pathways that modulate pain. J Clin Neurophysiol 1997;14:2±31. Woolf CJ, Thompson SW. The induction and maintenance of central sensitization is dependent on N-methyl-D-aspartic acid receptor activation; implications for the treatment of post-injury pain hypersensitivity states. Pain 1991;44:293±299. Yaksh TL. Spinal opiate analgesia: characteristics and principles of action. Pain 1981;11:293±346. Yaksh TL. Behavioral and autonomic correlates of the tactile evoked allodynia produced by spinal glycine inhibition: effects of modulatory receptor systems and excitatory amino acid antagonists. Pain 1989;37:111±123. Yokota T, Asato F, Koyama N, Masuda T, Taguchi H. Nociceptive body representation in nucleus ventralis posterolateralis of cat thalamus. J Neurophysiol 1988;6:1714±1727.