NMDA NR2B-containing receptor subtype are antinociceptive in preclinical models and lack effects on motor function, suggesting these compounds will have a better side-effect profile in man, compared to ketamine, a non subunit selective antagonist. Pharmaceutical companies are therefore developing selective NMDA NR2B-containing receptor subtype compounds. The activity of these selective NR2B compounds has been reported to be mediated only in the brain (Chizh et al. 2001). If so, preclinical testing cannot rely solely on withdrawal reflex tests. The aim of this study was to identify the site of action of the selective NR2B compound Ro-25,6981 in modulating noxious heat responses of WDR neurones in rats with inflammation. Male Wistar rats (180-225g) with an intraplantar injection of CFA into one hind paw (100μl/paw); under brief isoflurane (2-3% in O₂) anaesthesia), were anaesthetised with isoflurane (2-3% in O₂). The trachea, two jugular veins and a carotid artery were cannulated for administration of isoflurane, α-chloralose, compounds and measurement of blood pressure, respectively. L4-L6 spinal cord was exposed in all animals for electrophysiological recording; in some animals T12 spinal cord was exposed and the cord transected. Once surgery was completed animals were transferred to α-chloralose anaesthesia (300mg/kg bolus then 100mg/kg/h infusion). Extracellular recordings of isolated neurones were made using tungsten microelectrodes (1.8-2.2Ohms); a peltier device was used to apply noxious heat (50-55°C) to the peripheral receptive field every 4min and responses recorded. Increasing doses of Ro-25,6981 and ketamine were administered to a maximum dose of 100μmol/kg and 120μmol/kg, respectively. Experiments were terminated with an overdose of pentobarbitone. ED₅₀ values were calculated using XL fit software. Ro-25,6981 (3.125-50μmol/kg; i.v.) inhibited noxious heat responses in animals with an intact spinal cord, with an ED₅₀ value of 17.9μmol/kg (n=6); responses were reduced from baseline values of 25.3±3.4 to 10.3±2.5 spikes/s, after administration of a mean dose of 17.7±7.1μmol/kg. In contrast, Ro-25,6981 (6.25-100μmol/kg; iv) had no effect (n=6) in animals with a transected spinal cord. Ketamine was active; the ED₅₀ values were 68 and 60μmol/kg in the intact and transected spinal cord preparation, respectively. The present results show that the spinal cord needs to be intact to observe an antinociceptive effect of Ro-25,6981 suggesting that the brain plays a key role in mediating the antinociception of NR2B-containing receptor antagonists in rats with inflammation. This is in contrast to the non-subunit selective antagonist ketamine, which inhibited WDR responses with similar potency in both preparations.