Abstracts

Rationale: Seizures induce excitatory shifts in the reversal potential for GABA_A receptor-mediated responses (E_GABA) (Khalilov et al., 2003). GABA-mediated excitation precludes the effect of GABAergic anticonvulsants such as phenobarbital, so seizure-induced shifts in E_GABA may contribute to the intractability of electroencephalographic neonatal seizures. However there are conflicting reports regarding both the efficacy of GABAergic anticonvulsants and the anticonvulsant efficacy of therapeutic manipulation of E_GABA in experimental models of neonatal seizures. Methods: We used the whole hippocampal preparation to avoid potential alterations in NKCC1 activity related to axonotomy that occurs diffusely during slice preparation. Recurrent seizures were induced by low magnesium perfusate or by desinhibition. Simultaneous extracellular field potentials of multiple unit activity and two-photon fluorescence Ca²⁺ imaging from hundreds of neurons with single-cell resolution were performed to determine whether the responses to GABA_A receptor modulators are excitatory or inhibitory. Results: Seizures altered postsynaptic GABA_A receptor-mediated responses. The time elapsed since the first seizure and the cumulative seizure duration and power were strongly correlated with the degree of GABA_A receptor-mediated excitation assayed by the frequency of action potentials, intracellular Ca²⁺ transients, and reduction in phenobarbital efficacy. These changes were all blocked when shifts in E_GABA were prevented by inhibition of neuronal NKCC1-mediated inward chloride transport with the diuretic bumetanide. Conclusions: Our results clarify conflicts in the literature regarding the efficacy of GABAergic anticonvulsants by demonstrating the importance of two co-variables: time elapsed since first seizure, and the quantity of seizure activity. The results also support the synergy between phenobarbital and bumetanide as anticonvulsants in neonatal seizures.