Abstracts

Rationale: Sudden Unexpected Death in Epilepsy (SUDEP) is defined as the sudden, unexpected, and unexplained death of a person with epilepsy and accounts for up to 17% of all epilepsy-related deaths and 50% for patients who are refractory to treatment. While the mechanisms underlying SUDEP are not fully understood, there is increasing evidence that apnea is the primary cause. Our work using preclinical models of epilepsy shows that seizure-induced apnea (SIA) occurs during the tonic phase, and minutes before terminal asystole. We have previously observed SIA is not impacted by forebrain inhibition; thus, we hypothesize that overactive brainstem neural circuitry produces SIA.

Methods: We used a rapid kindling protocol on C57Bl/6J (WT) mice, which produces seizures (Fig. 1) on our modified Racine scale that display, wild running/jumping (stage 6), wild running/jumping and SIA (6a), tonic extension and SIA (stage 7), and death from SIA (stage 8). Apnea duration, seizure threshold and duration, EEG power amplitude, heart rate, and breathing rate were assessed for each condition. Using the same rapid kindling protocol, TRAP2 (Fos-iCre/tdTomato) mice (Fig. 2A) were kindled until reaching their first seizure behavior stage 6 or 7, and then 30 minutes later injected with 4-hydroxytamoxifen (4-OHT) to observe the activated neuronal populations in mice experiencing seizures. In a subset of these experiments, we performed in situ hybridization to assess the subtypes of activated neurons.

Results: In wild-type mice that were kindled for 3 days, 39% of mice died from SUDEP, with 100% of deaths occurring after a tonic seizure with apnea. We also observed apnea in a subset of stage 6 seizures that did not correspond with a tonic phase. Out of 83 seizures scored as 6 based on behavior, 44.6% had a pause in breathing > 1.5 seconds, which we categorized as 6a with an average apnea of 3.37±0.35 seconds. 18.9% of 6a seizures had an apnea > 5 seconds. In kindled TRAP2 mice, the most notable brainstem structure where activated neurons were observed was the periaqueductal gray (PAG), a brainstem structure that has an important role in breathing integration. We found more activated neurons in the PAG of mice that had seizures with apnea versus seizures without (Fig. 2B). We examined the cellular phenotype of these activated neurons and observed the majority to be glutamatergic (Fig 2C).

Conclusions: Our model reliably produces SIA and SUDEP within three days of kindling and can be used to further investigate differences between seizures with and without apnea. The histology results suggest excitatory neurons in the PAG are activated during seizures to produce the apnea that leads to SUDEP. The PAG represents a potential target for intervention to prevent SIA and SUDEP. Future studies examining the mechanisms by which the PAG produces SIA will be enlightening.

Funding: R01NS133139 and supplement 3R01NS133139-01S1