Purpose: Traditional methods of preclinical screening have predicted the effects of a putative antiepileptic drug (AED) against human absence seizures by testing its efficacy against clonic seizures in the high-dose pentylenetetrazole (PTZ) model. This high-dose PTZ model correctly predicted the efficacy of ethosuximide (ESM), benzodiazepines, and valproate (VPA) and the lack of efficacy of phenytoin (PHT) and carbamazepine (CBZ). However, the high-dose PTZ model erred in predictions for (a) phenobarbital (PB) (PTZ: efficacy; human: nonefficacy); (b) lamotrigine (LTG) (PTZ: nonefficacy; human: efficacy); (c) vigabatrin (VGB) (PTZ: nonefficacy; human: proabsence effect); and (d) tiagabine (TGB) (PTZ: efficacy; human: possible proabsence). It also appears to have erred in predictions for gabapentin (GBP) (PTZ: efficacy) and topiramate (TPM) (PTZ: efficacy). Because the lh/lh genetic model of absence seizures correctly predicted effects of ESM, clonazepam, VPA, PHT, CBZ, and PB against human absence seizures, we performed this study to test the predictive utility of the lh/lh model for LTG, VGB, TGB, GBP, and TPM.
Methods: Bipolar recording electrodes were implanted bilaterally into frontal neocortex of 8-week-old male lh/lh mice. With the exception of VGB, vehicle or drugs were administered intraperitoneally (i.p.) on alternating days, and an EEG was used to record effects on seizure frequency. With VGB, vehicle was administered i.p. on day 1, and gradually increasing doses of VGB were administered on successive days. Drug and vehicle effects were compared in corresponding 15-min epochs of the 150-min period after administration.
Results: LTG (4.8-144 micromol/kg) significantly (p < 0.04) reduced seizure frequency (by 65%) compared with vehicle. In contrast, VGB (0.35-11 mmol/kg) and TGB (0.27-27 micromol/kg) significantly increased seizure frequency (300-700%) and seizure duration (1,700-1,800%; p < 0.001). GBP (18 micromol/kg to 1.8 mmol/kg) and TPM (8.9-295 micromol/kg) had no significant effect on seizure frequency.
Conclusions: In contrast to the high-dose PTZ model, the lh/lh model correctly predicted the antiabsence effect of LTG, the possible proabsence effects of VGB and TGB, and the lack of effect of GBP and TPM. The lh/lh model appears to be superior to the high-dose PTZ model in predicting efficacy of putative AEDs against human absence seizures.