<i>In vitro</i> and preclinical assessment of drug interactions between fluoroquinolones and a non-steroidal anti-inflammatory drug predicting risk of seizure

Seizures are a common reason for drugs to fail during development, but they are difficult to predict preclinically. Enoxacin, a fluoroquinolone, rarely causes convulsions as a monotherapy, but convulsions have been seen after combination therapy with fenbufen, a non-steroidal anti-inflammatory drug. The interaction between the drugs is thought to result from inhibition of GABAA receptors, which are not their primary targets. Here we show that, among 15 marketed fluoroquinolones and one active metabolite, six, including enoxacin, inhibited GABA-evoked depolarization in cells expressing human GABAA receptors when administered with felbinac, the active metabolite of fenbufen. Within these six except for a prodrug possessed a piperazinyl group at the seventh position of the quinolone ring. We also administered enoxacin or norfloxacin plus felbinac to rats and determined the cerebrospinal fluid concentrations of each drug when convulsions occurred. As we previously reported, an increase in network burst frequency recorded from primary cultured rat cortical neurons on microelectrode arrays is a risk marker for seizures, so we tested whether this assay could predict seizures induced by drug interactions between fluoroquinolones and felbinac. When co-administered with felbinac, only those fluoroquinolones that inhibited GABA currents in patch-clamp tests increased network burst frequency. Principal component analysis using 17 microelectrode array parameters supported that the mechanism of action was due to GABA antagonism in rodent neurons. Thus, the microelectrode array assay predicted seizure risks from the combination of enoxacin and the active metabolite of fenbufen and identified other fluoroquinolones with seizure risk potential.