Abstracts

Rationale: Brain injury may cause lesions that are associated with increased risk of acquiring epilepsy.  However, the mechanism by which lesions contribute to epileptogenesis is not completely understood.  Neurons that have been experimentally deafferented have been shown to sprout axons; thus, a lesion that disrupts a neuron's connections may cause it to grow new axonal collaterals.  If new axons cannot cross the lesion site, they may grow and make synapses locally, forming excessive recurrent connections and increasing local network excitability.  In this work, we investigate this hypothesis of post-injury epileptogenesis both experimentally and computationally. Methods: We previously reported organotypic hippocampal culture model of post-injury epileptogenesis.  In this model, a latent period after injury (slicing) is followed by the appearance of spontaneous interictal and then ictal-like activity.   Here, we use organotypic cultures derived from hippocampal slices, isolated hippocampal subregion CA3, entorhinal cortex, or slices containing both hippocampal and entorhinal cortical regions.   We placed slices onto tungsten microwires glued to the bottom of tissue-culture plates and maintained cultures on a rocking plate in an incubator.    Spontaneous electrical activity in cultures was recorded for 30 minutes every 1 to 2 days.  We tracked axon sprouting from neurons in hippocampal slice cultures by phase contrast microscopy.  We developed a new computational model of an epileptic network that is based on Wilson and Cowan (1972) population models and system based modeling of Wendling et al., 2000 and 2002.  We extended these models by developing system-level implementation of depolarization block and short-term synaptic depression.  Our model consists of system-level blocks representing separate excitatory and inhibitory populations in a sub-region such as CA3 or CA1.  Blocks can be connected together into a larger system representing a whole hippocampus or hippocampal-cortical network. Results: Time course of axon sprouting in organotypic hippocampal cultures was closely correlated with the increase in duration of spontaneous paraxysmal electrical activity.  In experiments where two cultures were placed side by side, axons crossed from one slice to the other, and spontaneous electrical activity in both slices became synchronized.  This data suggest that sprouting axons play an important role in development of the spontaneous epileptiform activity in this model.  We found that cumulative duration of spontaneous epileptiform activity in isolated hippocampal slice cultures was significantly longer than in the isolated cortical slice cultures.   Sub-region cultures of CA3 developed epileptiform activity earlier than hippocampal slice cultures, while the latter developed epileptiform activity earlier than the combined hippocampal-cortical cultures.   We hypothesized that the axons sprouted in the more confined networks were more likely to form local recurrent connections than axons sprouted in less isolated networks, and that local recurrent connections were primarily responsible for appearance of spontaneous epileptiform discharges.  We incorporated different types of connectivity and network organization into the computational model and found that computational results support our experimental data. Conclusions: We used experimental and computational methods to demonstrate how injury can lead to epileptogenesis through network confinement and axon sprouting.   Our results may be useful in interpreting epileptogenic potential of different brain injuries. Funding: This work was supported in part by NIH/NINDS R33 NS088358.