Abstracts

Learning and memory deficits are common in temporal lobe epilepsy and are largely driven by pathological changes to temporal lobe circuits, especially those in the hippocampus, that support these processes. However, seizure-associated learning and memory deficits have been observed in non-epileptic humans and animals and may occur in the absence of temporal lobe pathology. Shortening seizure duration using closed-loop stimulation can improve learning and memory in epilepsy, but seizure-induced cognitive deficits are still poorly characterized. Seizures can cause retrograde amnesia and impair learning processes in the immediate post-ictal period. The time course of these deficits is complex and poorly characterized as are the underlying mechanisms. Investigating these processes in epilepsy is challenging because seizures occur spontaneously making it difficult to probe learning and memory at precise timepoints in relation to seizures. Studies inducing seizures electrically or optogenetically get around this issue, but artificially drive activity in the brain that may interfere with intrinsic activity. Here we propose a platform to induce seizures by chemogenetically increasing cell excitability rather than artificially driving cells to fire.

Rbp4-cre mice, expressing cre in dentate granule cells (DGCs) and layer V cortical neurons, are injected with a cre-dependent virus expressing the excitatory DREADD hM3Dq bilaterally in the dorsal dentate gyrus (DG) resulting in specific expression of hM3Dq in DGCs. Mice are then implanted with EEG electrodes or a GCaMP-coated GRIN lens. EEG mice are recorded continuously for one week to establish a seizure-free baseline then are injected with increasing doses of Clozapine-N-Oxide (CNO) ranging from 0.1-3mg/kg with a 48-hour washout time between injections. Seizure activity is extracted from EEG and confirmed visually with video. A separate cohort of mice are implanted with GRIN lenses targeted to either the DG or CA1 of the hippocampus then imaged using miniscopes to assess coding during behavior.

Mice given low dose CNO (0.3-0.5 mg/kg) had one to three seizures with mild behavioral components (behavioral arrest, nodding, mild clonus). At medium doses of CNO (1-3mg/kg) ,mice had multiple seizures over a three to six hour timespan (~1/hr) with more severe behavioral components (clonus and rearing). Miniscope imaging revealed two distinct phases of seizure activity, one characterized by synchronous firing of cells, behavioral arrest, and nodding, and the other characterized by a spreading wave during which hyperlocomotion, rearing, and clonus were often observed.

Specifically increasing DGC excitability using hM3Dq induces seizures in non-epileptic mice. By controlling the dose and timing of CNO administration, seizures can be precisely induced to study the time course of seizure-associated learning and memory deficits and investigate mechanisms to mitigate this dysfunction. We are currently using this platform to study spatial coding deficits associated with seizures.