Abstracts

Rationale: Cortical input to the dentate gyrus (DG) is anatomically separated into two components: the lateral perforant path (LPP) and the medial perforant path (MPP). In temporal lobe epilepsy (TLE), the cortical cells that give rise to the MPP become hyperexcitable and hypersynchronous (Kumar et al. Neurosci. 27(6):1239). However, it is unclear how a pathological MPP would rearrange DG output because few studies have examined LPP or MPP mediated synaptic currents or potentials in individual GCs or fast-spiking inhibitory interneurons (FS INs) in the DG. Here we compared evoked EPSCs (eEPSCs) and synaptically driven spikes between LPP and MPP synapses onto GCs and FS INs.Methods: Whole-cell voltage-clamp and current clamp recordings were made from GCs and FS INs in hippocampal slices from 14-20 day old C57B6 mice (-60 mV; bicarbonate-buffered standard ACSF, K-gluconate based intracellular solution). To activate LPP or MMP fibers, a small bipolar stimulating pipette (<5 µm I.D., filled with ACSF) was placed in either the outer third (LPP) or middle third (MPP) of the molecular layer, and constant current stimulus trains were applied (20 pulses, 100 µs, 0.1-1.0 mA, 10 Hz). At each stimulus intensity, 10 trials were run in voltage clamp to record EPSCs, followed by 10 trials in current clamp to measure the EPSP-driven spike probability caused by the same stimulus. Each recording included both MPP and LPP measurements. Recordings in GCs were made in the presence of bicuculline (5 µM) and CGP 55845 (1 µM) to block inhibitory components that were recruited directly by the stimulus. This was not the case in fast-spiking INs where such recruitment was not observed.Results: Input-output relationships were constructed by plotting spike probability versus the corresponding mean EPSC amplitudes, and were fit using a sigmoidal function whose midpoint A1/2 indicates the strength of EPSC-to-spike coupling. In FS INs the MPP was more strongly coupled to spiking than the LPP (MPP: A1/2 = 268 ± 23 pA; LPP A1/2 = 324 ± 26 pA; mean ± SEM; Paired t-test, n=7, p ≤ 0.004). In contrast, for GCs, there was no significant difference between LPP and MPP midpoints (MPP: A1/2 = 288 ± 16 pA; LPP A1/2 = 267 ± 9 pA; mean ± SEM; n=6, p = 0.49). Furthermore, eEPSCs arising from the LPP and MPP exhibited different kinetics in FS INs, but not in GCs. In FS INs, weighted decay time constants (±SEM) were 8 ± 0.2 ms for MPP and 13 ± 1 ms for LPP (n = 20, p ≤ 0.001). In GCs the values were 7 ± 2 ms for MPP and 8 ± 2 ms for LPP (n = 20, p = 0.4).Conclusions: These data suggest that signals entering the DG via the MPP recruit more feedforward inhibition compared to signals arriving via the LPP. It is known that FS INs are prone to synchronization and are able to synchronize large populations of excitatory principle cells. Therefore, in TLE it is possible that a hypersynchronized drive onto FS INs yields a more synchronized DG output.