Abstracts

Rationale: Mesiotemporal lobe Epilepsy (MTLE), the most frequent form of focal epilepsy, is often drug-resistant. Recent studies proposed enriching the epileptic focus with GABA-releasing engineered cells to prevent seizures. However, ex vivo data from animal models and MTLE patients suggest that, due to changes in chloride homeostasis, GABA_A receptor activation is depolarizing and partly responsible for focal seizure initiation. To understand how these two apparently contradictory aspects of GABA inhibition coexist in MTLE, we used an established mouse model of MTLE presenting hippocampal sclerosis and recurrent focal seizures 30-40 days after a unilateral injection of kainate (KA) in the dorsal hippocampus. Methods: In this model, we first examined NKCC1 and KCC2 chloride co-transporter expression by western blot and tested in vivo whether NKCC1 chloride transporter blockade suppressed hippocampal focal seizures. We assessed protein expression levels of GABA_A, GABA_B, NMDA, and AMPA receptor subunits and integrated these data in a computational model of hippocampal pyramidal neurons. We then compared numerical simulations with in vivo intracellular recordings from MTLE mice hippocampus and evaluated both in silico and in vivo (intrahippocampal infusions) the effects of GABA_A receptor agonists (muscimol, isoguvacine) or antagonists (picrotoxin, gabazine) on focal hippocampal seizures. Results: In mice injected with KA in the dorsal hippocampus, we showed 30-40 days post-injection that expression of the chloride transporter NKCC1 was increased, whereas chloride transporter KCC2 expression was decreased, compared to saline controls. Furthermore, focal seizures were suppressed by systemic injection of two blockers of NKCC1 transporter. In addition, levels of α1, α3 and γ2 GABA_A receptor subunits were increased, while levels of NR1, NR2A and NR2B NMDA receptor subunits, GluR1 and GluR2 AMPA receptor subunits, and R1 and R2 GABA_B receptor subunits were decreased. In vivo intracellular recordings of hippocampal neurons, as well as biosimulation integrating these different changes support the notion that GABA is depolarizing in the MTLE mouse model. Paradoxically, intrahippocampal injections of GABA_A receptor agonists (muscimol and isoguvacine), but not antagonists, suppressed focal seizures. Conclusions: Our data demonstrate that despite a depolarizing effect of GABA in MTLE, local pharmacological activation of GABA_A receptor remains an effective antiepileptic strategy by restoring physiological intracellular chloride concentrations.