Abstracts

Rationale: Traumatic brain injury is the leading cause of acquired epilepsy and is the major cause of disability and death in the U.S. Currently available antiepileptic therapy is ineffective for prevention of post-traumatic epileptogenesis (PTE) suggesting different pathological mechanisms. Electroencephalography (EEG) remains the gold standard in clinical and pre-clinical epilepsy research and has a potential to identify electrographic biomarkers of post-traumatic epileptogenesis. Yet, the types, patterns, frequency and relevance of various electrographic signatures remain to be characterized in pre-clinical models. Mounting evidence in the field demonstrates contributing and causative roles for non-neuronal cells, such as astrocytes, glial cells carrying out critical homeostatic functions in the brain. Here we characterized the dynamic changes in EEG frequency bands over time from TBI to onset of post-traumatic seizures and investigated the downstream effects of TBI on astrocytes.

Methods: We characterized a model of PTE induced by repeated diffuse TBI in male C57Bl6/J mice (8-16 weeks old). As controls, we used sham-injured mice and naive (no sham-injury, only EEG electrode implantation) mice. We utilized 1-3 channel EEG acquisition using intra-cranial screw electrodes. Mice were recorded chronically via uninterrupted, 24/7 video-EEG data acquisition for 2-4 months after TBI. The EEG data were examined using manual and semi-automated algorithms for identification of spontaneous electro-clinical seizures and quantitative spectral analysis. At the end of the recording, all mice were sacrificed for cardiac perfusion and immunohistochemistry.

Results: In our model, 25% of mice developed chronic, spontaneous, unprovoked seizures. These mice demonstrated dynamic spectral changes in the slow (delta, theta) and high frequency range (ripples and fast ripples) early and chronically after TBI. Comparison of manual data analysis and use of the machine learning algorithm revealed higher accuracy of the algorithm in a shorter time frame. Immunohistochemistry against key astrocyte proteins including Glt1, Kir4.1, Glutamine synthetase and Connexin 43 revealed an atypical response of astrocytes to diffuse TBI, characterized by lack of these critical homeostatic proteins in the cortex. The cortical area covered by atypical astrocytes was significantly increased in mice that developed PTE.

Conclusions: Post-traumatic seizure activity can vary from animal to animal and can cluster, emphasizing the importance of continuous, 24/7 video-EEG data acquisition. The use of quantitative, semi-automated algorithms are useful tools in determining the acute and chronic alterations of the EEG frequency spectrum over time and dependence of these alterations on circadian cycle of the animals. Presence of pathologic high frequency oscillations is a promising EEG signature for prediction of epileptic activity, although better characterization, spatial and temporal identification and upstream/downstream effects are yet to be detailed for data translation.

Funding: Please list any funding that was received in support of this abstract.: R01NS105807, Citizens United For Research In Epilepsy (CURE).