Purpose and methods: Regional overexpression of the multidrug transporter P-glycoprotein (P-gp) in epileptic brain tissue may lower target site concentrations of antiepileptic drugs and thus contribute to pharmacoresistance in epilepsy. We used the P-gp substrate R-[(11)C]verapamil and positron emission tomography (PET) to test for differences in P-gp activity between epileptogenic and nonepileptogenic brain regions of patients with drug-resistant unilateral temporal lobe epilepsy (n = 7). We compared R-[(11)C]verapamil kinetics in homologous brain volumes of interest (VOIs) located ipsilateral and contralateral to the seizure focus.
Results: Among different VOIs, radioactivity was highest in the choroid plexus. The hippocampal VOI could not be used for data analysis because it was contaminated by spill-in of radioactivity from the adjacent choroid plexus. In several other temporal lobe regions that are known to be involved in seizure generation and propagation ipsilateral influx rate constants K(1) and efflux rate constants k(2) of R-[(11)C]verapamil were descriptively increased as compared to the contralateral side. Parameter asymmetries were most prominent in parahippocampal and ambient gyrus (K(1), range: -3.8% to +22.3%; k(2), range: -2.3% to +43.9%), amygdala (K(1), range: -20.6% to +31.3%; k(2), range: -18.0% to +38.9%), medial anterior temporal lobe (K(1), range: -8.3% to +14.5%; k(2), range: -14.5% to +31.0%) and lateral anterior temporal lobe (K(1), range: -20.7% to +16.8%; k(2), range: -24.4% to +22.6%). In contrast to temporal lobe VOIs, asymmetries were minimal in a region presumably not involved in epileptogenesis located outside the temporal lobe (superior parietal gyrus, K(1), range: -3.7% to +4.5%; k(2), range: -4.2% to +5.8%). In 5 of 7 patients, ipsilateral efflux (k(2)) increases were more pronounced than ipsilateral influx (K(1)) increases, which resulted in ipsilateral reductions (10%-26%) of R-[(11)C]verapamil distribution volumes (DV). However, for none of the examined brain regions, any of the differences in K(1), k(2) and DV between the epileptogenic and the nonepileptogenic hemisphere reached statistical significance (p > 0.05, Wilcoxon matched pairs test).
Conclusions: Even though we failed to detect statistically significant differences in R-[(11)C]verapamil model parameters between epileptogenic and nonepileptogenic brain regions, it cannot be excluded from our pilot data in a small sample size of patients that regionally enhanced P-gp activity might contribute to drug resistance in some patients with temporal lobe epilepsy.