Abstracts

Rationale: Temporal lobe epilepsy (TLE) is among the most common forms of epilepsy in adults, with a significant proportion of patients experiencing drug-resistant seizures. TLE is often associated with mesial temporal sclerosis (MTS), in which there is extensive cell loss in hippocampal areas CA1 and CA3 while the dentate gyrus (DG) and CA2 remain relatively intact. A prior study found that CA2 neurons in tissue from TLE patients show interictal-like firing and receive aberrant perisomatic excitatory synapses from DG granule cell (GC) mossy fibers (Wittner et al. Brain. 2009;132:3032–3046). We thus explored the hypothesis that pathophysiological changes to CA2 pyramidal neuron (PN) excitability or synaptic connectivity may be associated with TLE by examining CA2 properties in a mouse model of TLE.

Methods: We used a mouse model in which one dose of pilocarpine (PILO) induced acute status epilepticus, leading to a pattern of hippocampal neurodegeneration reminiscent of MTS and recurring spontaneous seizures. We performed electrophysiological recordings from PNs in acute hippocampal slices from PILO mice in the chronic phase of epilepsy as well as age-matched controls. In some experiments we used Cre-expressing mouse lines to selectively express a light-activated excitatory channel in CA2 PNs or DG GCs. We also performed immunohistochemistry to examine CA2 interneuron (IN) populations following PILO.

Results: We found that in healthy tissue DG GC mossy fibers made direct but relatively weak excitatory synapses onto CA2 PNs. CA2 PNs, like those in CA3, both directly excited other CA2 PNs via a recurrent CA2-CA2 PN circuit and indirectly inhibited other CA2 PNs in a feedforward manner. The CA2 and CA3 subfields also form reciprocal excitatory and feedforward inhibitory circuits. These recurrent and reciprocal circuits constitute an auto-associative network in which INs crucially control CA2/CA3 population excitability. Recordings in slices from PILO mice revealed a significant increase in excitatory synaptic input to CA2 PNs from DG GC mossy fibers, with little change in CA2-CA2 recurrent excitation or in excitatory input from CA3. In contrast, we saw a significant and widespread reduction in feedforward inhibition recorded in CA2 PNs upon stimulation of the CA3 Schaffer collaterals, CA2 recurrent collaterals, or DG GC mossy fibers. Notably, fast inhibition mediated by GABA_A receptors showed a greater reduction than slow GABA_B-mediated inhibition. We also saw increased CA2 PN input resistance and thus increased intrinsic excitability, leading to a higher firing rate upon direct current injection. Finally, we observed a significant decrease in the density of pro-cholecystokinin-expressing INs and relative survival of parvalbumin-expressing INs in PILO mice.

Conclusions: These data reveal a shift in the inhibitory-excitatory balance of the CA2 network towards net excitation and provide support for the hypothesis that CA2 may contribute to the generation of epileptiform activity in TLE.

Funding: Please list any funding that was received in support of this abstract.: Supported by 1F31NS113466-01.