Abstracts

Rationale: Epilepsy and premature mortality are common and prominent features among the multiple clinical manifestations of Leigh syndrome (LS) and other mitochondrial disorders. Recent pharmacological and pathological studies have led to the hypothesis that mitochondrial epilepsy may arise from impaired neuronal network oscillations caused by higher susceptibility of GABAergic interneurons (than excitatory neurons) to activity disturbance and cell death, following mitochondrial respiratory chain dysfunction. However, this hypothesis has not yet been tested directly.Leigh Syndrome (LS), or subacute necrotizing encephalopathy, is a debilitating, progressive, and neurodegenerative mitochondrial disorder of childhood. It typically presents with multi-systemic clinical symptoms which result in disability and ultimately death by 3 years of age. Genetic studies have shown that LS is strongly associated with loss-of-function mutations in NDUFS4, the gene that encodes a subunit of the protein complex I in the mitochondrial electron transport chain. Mouse models of LS, generated by global or CNS-specific Knock-out (KO) of Ndufs4, recapitulate several key clinical features of the disease in humans, including spontaneous seizures and premature deaths.We examined the contribution of excitatory and inhibitory neurons to the development of epilepsy and sudden death phenotypes in the mouse model of LS. Methods: Control mice and mice with Ndufs4 knocked out selectively in either GABAergic neurons (Gad2-specific KO mice) or in Vglut2-positive glutamatergic neurons (Vglut2-specific KO mice) were generated by crossing the Ndufs4 floxed mice with the Gad2Cre or Vgut2Cre driver mice respectively. Long-term video monitoring, thermal seizure test, and video-EEG assessment of the mice were conducted as described in our previous work. Results: All Gad2-specific KO (not control) mice exhibited spontaneous behavioral seizures and died prematurely. Series of myoclonic seizures, often preceding and following a generalized tonic-clonic seizure, were observed beginning on postnatal day (P) 32. Premature deaths occurred starting at P49 and none of the mice survived past P82. All witnessed deaths occurred immediately following a Racine 5 generalized spontaneous seizure, suggesting they are precipitated by a seizure. Combined video-EEG recordings revealed generalized interictal epileptiform spikes during resting behavior or sleep as well as spontaneous and thermal seizures, marked by high-voltage spike and wave EEG discharges closely associated with hypermotor behaviors lasting 31 ± 9 seconds. In a striking contrast, Vglut2-specific KO mice did not show any behavioral or electrographic sign of predisposition to spontaneous or thermal seizures. Interestingly however, they succumb to non-epilepsy related premature death, starting at ~P35, and none of the mice surviving past P60. Conclusions: These results suggest that the impact of Ndufs4 KO in GABAergic neurons (not in excitatory neurons) play a critical role in the development of epilepsy in LS. In addition, a portion of mortality in LS may arise from sudden unexpected death in epilepsy. Funding: Ellenbogen Chair Research Funds, UW Department of Neurological SurgeryRyan Murphy SUDEP Research Funds