It has been suggested that methamphetamine (METH)-induced neurotoxicity requires the activation of both dopamine (DA) and glutamate (GLU) systems. To investigate the possibility that METH-induced increases in extracellular GLU, as measured by in vivo microdialysis [Nash and Yamamoto (1992) Brain Res., 581:237-243], arise from neuronal stores, postembedding immunogold electron microscopy was used to measure the density of presynaptic GLU immunoreactivity within the striatum, the shell of the nucleus accumbens, and the motor cortex. Rats were treated with METH (5 mg/kg), or an equivalent volume of saline (SAL), every 2 h for a total of four injections. No ultrastructural evidence of terminal degeneration was observed. Significant decreases in the density of nerve terminal GLU immunolabeling occurred 12 h following METH administration within the primary motor cortex and the ventrolateral caudate/putamen, and a trend towards depletion was seen within the dorsolateral caudate/putamen. Although GLU immunolabeling within the shell of the nucleus accumbens was unaffected, DA content was decreased in all regions examined 1 week following METH treatment. The lack of degeneration, coupled with a partial recovery of DA levels, suggests that moderate doses of METH may inhibit DA biosynthesis without widespread terminal loss. Furthermore, METH administration results in a decrease in presynaptic GLU that correlates both temporally and anatomically with delayed GLU overflow, suggesting that neuronally derived GLU may play a role in METH-induced neurotoxicity. However, there does appear to be a dissociation between DA loss and altered GLU immunocytochemistry within the nucleus accumbens.