Abstracts

Rationale: 20% of symptomatic epilepsy is post-traumatic. In infants, severe traumatic brain injury involving extensive axonal shearing leads to epilepsy in up to 60% of patients. Although we presume that sprouting of axons and dendrites in injured tissues leads to local hyper-connectivity and recurrent excitation, very little is known about the rules governing the process of neural reorganization following injury. Methods: Here, we model post-traumatic epilepsy using organotypic slices, which are prepared from neonatal mice (P5-P8) using shearing trauma (slicing). Slices become severely epileptic over the course of 3 weeks in culture. The mice used for these experiments continuously express in all neurons the calcium-sensitive protein Yellow Cameleon 3.6 (YC3.6). Brain slices were cultured using the membrane insert technique in which the sample rests on a piece of filter paper at the interface between culture media and air. The samples were imaged at intervals of 2-4 days-in-vitro (DIV) beginning the day after slicing (1 DIV). Using two-photon, targeted path scan (TPS) imaging through a dry objective (to allow the slice to remain at an interface), we recorded 15 minute epochs of spontaneous activity from at least 30 cells at a sampling rate of 25-50Hz (sampling rate with TPS depends on the path selected). These imaging sessions were repeated at consecutive weeks in vitro, as the slices underwent epileptogenesis. Results: TPS calcium imaging revealed a pattern of epileptogenesis consistent with previously published electrophysiological measurements, with epileptiform activity appearing as large, rapid, spatially correlated increases in calcium (manifest as increases in the YC3.6 ratio). Activity began at 1DIV as brief bouts of synchronous activity in subpopulations of the imaged neurons. The activity transitioned to population-wide synchronous bursting, followed by intense seizure-like activity during the second week in vitro. The seizures then became longer and less intense (lower peak calcium for individual neurons) during the third week in vitro. Correlation analyses reveal a spatial distribution of highly correlated activity that was local during the first week, uniform (local and long-distance correlations) during the second week, and preferentially local during the third week. We are currently using network analysis techniques to quantify changes in network structure revealed by calcium imaging of neural activity. Conclusions: Our results support a model in which initial sprouting following injury targets nearby neurons. We hypothesize that, when the sprouting extends to include long-distance connections, intense seizure activity leads to death of the most heavily connected neurons. This initial wave of status epilepticus-induced cell death leaves behind an epileptic network that consists of predominantly local connections. While the organotypic slice culture represents an extreme model of injury (resulting in epilepsy in 100% of samples), we propose that similar principles of neural re-wiring following traumatic injury lead to epilepsy in vivo.