Abstracts

Rationale: The comparative antiepileptic properties of furosemide and bumetanide and their mechanisms of action in the intact adult brain remain unclear. Since rodents metabolize these loop diuretics distinctly differently than humans, in vivo data from rodent seizure models can be difficult to interpret. We addressed these issues by testing the effects of furosemide, bumetanide, and several bumetanide prodrugs on acute cortical seizure models in a species of primate (Macaca nemestrina) that is thought to metabolize loop diuretics similarly to humans. Methods: Acute epileptic foci were created in hand sensorimotor cortex by placement of 0.5 mm-2 pledgets of gelfoam that had been soaked in either bicuculline or 4-AP (100 µm). Stimulating electrodes were placed on untreated cortex to elicit afterdischarge activity. The effects of treatments on the epileptiform activity were quantified using software that counted interspike-intervals (ISI), spike height, and intensity and duration of afterdischarge activity. Results: Bumetanide (2 mg/kg IV and 10 mg/kg oral) blocked bicuculline spiking, 4-AP ictal bursting, and stimulation-evoked afterdischarges. Sodium thiopental also blocked epileptiform activity in these models. Furosemide was not effective in blocking epileptiform activity at the same dose as bumetanide, but was effective at 5X its dose (10 mg/kg IV). Several amide and ester prodrugs of bumetanide were as efficacious as bumetanide while eliciting little or no diuresis. Fosphenytoin (20 mg/kg IV) suppressed 4-AP bursting and afterdischarges but not bicuculline spiking. Levetiracetam (Keppra) (60 mg/kg IV) suppressed 4-AP bursting and bicuculline spiking, but not afterdischarges. Neither AED was as efficacious as bumetanide in any model. Conclusions: Blockade of epileptiform activity by sodium pentothal demonstrates that transmembrane chloride gradients were not disrupted in these models. Bumetanide and furosemide blockade of epileptiform activity in the presence of bicuculline suggests that their antiepileptic effects are independent of effects on GABA-A signaling. Furosemide was less efficacious than bumetanide in blocking epileptiform activity. Since bumetanide is a more potent and specific antagonist of NKCC1 with lesser affinity for KCC2 than furosemide, both loop diuretics appear to mediate their antiepileptic effects through antagonism of NKCC1. The ability of bumetanide prodrugs to block epileptiform activity without eliciting diuresis suggests that the antiepileptic effects of loop diuretics are not related to systemic diuretic effects. It has previously been suggested that furosemide blocks epileptiform activity through nonsynaptic mechanisms since it is effective in a variety of in vitro seizure models involving a spectrum of different synaptic mechanisms, including the zero-calcium model in which chemical synaptic transmission is absent (Hochman et al., Science, 1995). We hypothesize that both furosemide and bumetanide reduce hypersynchronous neuronal activity in adult cortex through nonsynaptic mechanisms related to NKCC1 antagonism.