Abstract:Drug-induced convulsions/seizures are a serious adverse event in the clinic; however, precise convulsion/seizure risk assessment remains challenging, particularly in the early phase of drug development. There is a demand for reliable in vitro assay systems that can predict in vivo convulsions. Here, we investigated the usefulness of in vitro microelectrode array (MEA) assays using rat primary neurons by comparing them with an in vivo convulsion study for 14 reference drugs known to cause seizure/convulsion: Paroxetine, 4-aminopyridine, pentylenetetrazol, strychnine, amoxapine, fluvoxamine, linopirdine, theophylline, pilocarpine, tramadol, bupropion, diphenhydramine, kainic acid, and veratridine. For the in vitro assay, a culture condition was established that demonstrated stable firing rates and drug responses of neurons and ensured a balanced expression of excitatory and inhibitory neuronal markers. Cultured rat primary cortical neurons were tested by MEA for electrophysiological changes induced by drugs. An MEA parameter, network burst frequency (NBF), was increased in a concentration-dependent manner by some of the drugs tested, demonstrating reproducibility. Rats were intraperitoneally or intravenously given the drugs at doses that did and did not cause convulsions. Concentrations of the drugs in cerebrospinal fluid (CSF), obtained just after convulsion onset, were measured by LC-MS/MS. The NBF threshold for seizure could be used to predict the CSF concentrations of some drugs in rats with convulsions. Thus, an in vitro MEA assay using rat primary neurons can predict the risk of in vivo convulsions, which could be useful for drug screening during the early stage of drug candidate selection.