Imaging the N-methyl-D-aspartate receptors (NMDARs) in the living human brain by positron emission tomography (PET) or single photon emission computed tomography (SPECT) would provide useful information on the role of these receptors in ischemia and in various neurological disorders such as degenerative diseases, epilepsy or schizophrenia. To assess NMDAR radiotracer development and to propose perspectives, we overviewed the PET and SPECT candidate radioligands developed until now. Labelled molecules of interest were classified in three groups according to their binding site: intrachannel pore site blockers, glycine site inhibitors and NR2B selective subunit antagonists.

Flumazenil is a specific antagonist of the central benzodiazepine receptor (CBZR). Labelled with 11C, this compound is the reference radioligand for positron emission tomography (PET) study of the CBZR in humans and primates. The time-course of [11C]-flumazenil radioactivity and its main acid metabolite [11C] Ro 15-3890 were reconstructed from the time-course of total radioactivity in plasma after administration with high or low SRA in primates and humans, applying an extraction procedure validated by TLC. The measured pharmacokinetics of [11C]-flumazenil (T1/2 beta = 45.1 +/- 12.3 min, T1/2 alpha = 1.5 +/- 1.5 min; K = 0.14 +/- 0.14 min-1; Vd area = 44.0 +/- 17.0 l; Clp = 40.0 +/- 8.5 l/h) exhibited a very rapid distribution phase followed by fast elimination, with a large volume of distribution; these results were confirmed by HPLC determinations and agree with previously published data on unlabelled flumazenil. Pharmacokinetics of [11C] Ro 15-3890 acid metabolite show that high drug concentrations in the blood are promptly achieved (kf = 0.13 +/- 0.004 min-1), with a very rapid elimination half-life (T1/2m = 4.47 +/- 1.31 min) comparable to that of [11C]-flumazenil. The percentage metabolization of parent compound to the acid [11C] Ro 15-3890 was constant from the 15th minute and was significantly higher in man compared to the monkey. This percentage was increased by prior eating. The other putative metabolites, i.e. labelled [11C] Ro 15-4965 and unlabelled Ro 15-5528, were never observed at detectable concentrations with TLC and HPLC in rabbit, baboon and human blood samples. This pharmacokinetic study of plasma flumazenil may be useful to implement a dynamic method of CBZR quantification using PET and for analysis of pharmacokinetics in brain tissue.