The intrathecal (i.t.) administration of glutamate (10-100 nmol) caused dose-related hyperalgesia (mean ED50 of 35 nmol) when assessed in the thermal behaviour model of nociception, the hot-plate test maintained at 50 degrees C. The i.p., i.t. or intracerebroventricular (i.c.v.) injection of the nitric oxide synthase inhibitors, L-NOARG and L-NAME, did not induce any detectable effect per se, but instead, produced dose-related inhibition of glutamate-induced hyperalgesia. D-NAME, the inactive enantiomer of L-NAME, had no effect. The i.c.v. or i.t. administration of L-NIO caused graded attenuation of glutamate-induced hyperalgesia. L-arginine (3.4 mmol kg(-1), i.p.), but not D-arginine (3.4 mmol kg(-1), i.p.) significantly potentiated glutamate (10 nmol)-induced hyperalgesia, an action that was prevented by L-NOARG (137 nmol kg(-1)). The co-injection of S-nitroso-N-acetyl-D,L-penicillamine (SNAP) (0.22 micromol) or 8-bromo-cGMP (22.5 nmol) with glutamate (10 nmol), via either i.t. or i.c.v. routes, also significantly enhanced glutamate-induced hyperalgesia. The guanylate cyclase inhibitors LY 83583 (0.1-1.0 nmol) or ODQ (30-300 pmol) co-administered with glutamate, dose-dependently antagonised the glutamate-induced hyperalgesia. Collectively, these results demonstrate that the i.t. injection of glutamate into the spinal cord of mice produces dose-related hyperalgesia an effect that was largely mediated by the L-arginine-nitric oxide-cGMP pathway from both spinal and supraspinal sites.