Screening Methods for Antiepileptic Drugs

Introduction & rationale

• Epilepsy = chronic disorder of recurrent unprovoked seizures (paradoxical abnormal, excessive, hypersynchronous neuronal discharge); drug therapy is empirical and does not correct the underlying pathology.
• Seizures reflect an ion-channel imbalance — reduced inhibitory (GABA-mediated) or increased excitatory (glutamate-mediated) drive — which the in-vitro mechanistic assays directly interrogate.
• The driving problem — despite >20 marketed antiseizure drugs, 20–30% of epilepsy is drug-resistant and ~15% of GTCS / ~25% of absence stay uncontrolled — the rationale for ever-better preclinical models, especially of pharmacoresistant epilepsy.
• Central caveat — most classical screens are models of seizures, not of epilepsy — they evoke an acute provoked seizure in a normal brain; only genetic and chronic post-status models approach true (spontaneous, recurrent) epilepsy.
• Purpose of a screening cascade: detect anticonvulsant activity → define the spectrum (which human seizure type) → probe mechanism → flag disease-modifying / antiepileptogenic potential → identify activity in drug-resistant models → assess neurotoxicity / therapeutic index. No single test predicts the full profile, so a battery is mandatory.
• "Ideal" model benchmark — spontaneous seizures, human-resembling seizure type, EEG correlates, age-dependency of onset; no model meets all four — genetic models come closest (the 2002 NIH/NINDS/AES workshop urged more genetic and injury-induced models for pharmacoresistant, disease-modifying targets).