PATCH-CLAMP RECORDINGS REVEAL POWERFUL GABAERGIC INHIBITION IN DENTATE HILAR NEURONS

Whole-cell and cell-attached patch-clamp recordings were used to investigate the nature of GABA(A) receptor-mediated inhibition in the adult rat dentate gyrus in standard 400-mu m-thick horizontal slices. In the presence of the glutamate receptor antagonists D-2-amino-5-phosphonovaleric acid and 6-cyano-7-nitroquinoxaline-2,3-dione, whole-cell voltage-clamp experiments with chloride-filled electrodes ([Cl-](in) = [Cl-](out)) revealed a high degree of spontaneous activity(10-60 Hz) in all hilar neurons (HNs) recorded with access resistances lower than 20 M Omega. The events were inward at negative holding potentials, reversed at around the Cl- equilibrium potential, and were completely abolished by the specific antagonists of the GABA(A) receptor channel picrotoxin and SR-95531 in a reversible manner, indicating that they were spontaneous inhibitory postsynaptic currents (sIPSCs) mediated by GABA(A) receptors. The majority of the slPSCs were TTX-insensitive miniature currents resulting from the action potential-independent release of GABA. The 10-90% rise times and the monoexponential decay time constants of the sIPSCs were significantly longer in HNs than those found in neighboring granule cells (GCs). Furthermore, the decay time constant of the hilar slPSCs was not voltage dependent, contrary to the voltage dependency of the decay time constant of the sIPSCs recorded from GCs. As HNs have longer electrotonic length than GCs do, dendritic filtering may contribute to the kinetic differences. Nonstationary fluctuation analysis showed that whereas the number of channels open at the peak of individual sIPSCs was similar, the single-channel conductances significantly differed between the two cell groups. The 21% smaller single-channel conductance and the existence of electrotonically close GABAergic synapses on HNs indicate that dendritic filtering alone cannot explain the differences between HNs and GCs. The distinct subunit composition of the GABA(A) receptor channels in HNs and GCs may also be responsible for the altered kinetics of IPSCs in HNs. However, the subunit specific benzodiazepine agonist zolpidem (3 mu M) prolonged the monoexponential decay time constants in both HNs and GCs. Thus, differences between the GABA(A) receptors of the two cell types are not due to a simple all-or-none presence/absence of the alpha 5 subunit. In order to determine the effect of the activation of GABA(A) receptors on the resting membrane potential in HNs and GCs in a nonintrusive way, we used single potassium channels as transmembrane voltage sensors by measuring the change in their conductance in cell-attached recordings in response to the GABA(A) agonist muscimol. GABA(A) receptor activation resulted in a strong peak depolarization (about 16 mV) in GCs but induced only small (about 4 mV) depolarizations in HNs. These results reveal for the first time that spontaneous activation of GABA(A) receptors takes place in HNs with a high frequency. Thus, while significant differences exist in the way GABAergic inhibition operates in the two neighboring neuronal population, it is highly unlikely that a general lack of inhibition can explain the extreme vulnerability of HNs to excitotoxic insults.