Abstracts

Rationale: De novo mutations causing dysfunction of the ATP1A3 gene encoding a3 subunit of Na+/K+-ATPase in neurons result in alternating hemiplegia of childhood (AHC). AHC manifests as paroxysmal episodes of hemiplegia, dystonia, behavioral abnormalities, seizures and sudden unexpected death of epilepsy (SUDEP). The aim of this study was to characterize a novel knock-in mouse model containing the E815K mutation of the ATP1A3 gene recognized as the most severe phenotype of AHC and to investigate the efficiency of flunarizine- the most effective drug used in AHC patients. Many E815K patients have had catastrophic regression after stopping flunarizine. It is unknown whether this is because of withdrawal from its acute beneficial effect, withdrawal from its long-term neuroprotective effect or a detrimental effect of flunarizine itself. Methods: Behavioral and neurophysiological testing demonstrated E815K mice express a phenotype that bears a resemblance to humans with AHC. Results: All E815K mice had spontaneous seizures with high incidence of mortality (SUDEP), and required less electrical stimulation to reach the kindled state as compared to wild type littermates. Flunarizine reduced amygdala kindling-induced After Discharge Duration in E815K mice compared to vehicle-treated controls. E815K mice treated acutely with flunarizine had reduction in provoked seizures and shorter seizure duration as well as hemiplegic attacks induced by vestibular stimulation in the first few days of treatment compared with vehicle-treated mice. However, after withdrawal of flunarizine, this beneficial effect was lost and treated mice did neither better nor worse. Conclusions: Our data suggest that Flunarizine has acute but not chronic effects on seizure and hemiplegia duration in E815K mice suggesting that it does not have a neuroprotective effect. We conclude that our knock-in mouse model containing the E815K mutation of the ATP1A3 gene manifests clinical and neurophysiological features of AHC, and is a useful model to test potential treatments. Funding: This work was supported by The Duke Institute for Brain Sciences (DIBS) Fund number 4510874, Duke Fund numbers 4410161 and 3912247, The Duke Translational Research Institute (DTRI) Grant SOMVP-2012, and a grant from a Cure AHC Foundation (MAM).