GABAergic connections among interneurons play an important role in regulating the overall excitability of neuronal networks. Here we report that GABAergic inhibition of hippocampal interneurons shows several unusual pharmacological properties, implying that the GABA receptor content is more diverse than in pyramidal neurons.

Whole-cell voltage-clamp recordings were obtained from CA1 stratum radiatum interneurons or pyramidal neurons in guinea-pig hippocampal slices (animals killed by Schedule 1 method). We recorded evoked IPSCs and responses to iontophoretically applied GABA or the GABA_C agonist CACA in the presence of AMPA, kainate, NMDA, group III mGluR and GABA_B receptor antagonists. Under these conditions, extracellular stimulation evoked a monosynaptic IPSC in both interneurons and pyramidal cells, which was completely blocked by the GABA_A antagonist SR95531 (10 µM). A dose-response curve for picrotoxin showed that IPSCs in interneurons were blocked with two binding sites (IC₅₀ = 1.1 and 50.5 µM). In contrast to IPSCs in interneurons, 100 µM picrotoxin completely abolished IPSCs in pyramidal neurons. In the continued presence of 100 µM picrotoxin, current-voltage relationships for IPSCs and for responses to iontophoretically applied GABA or CACA showed identical reversal potentials. Interestingly, both the GABA_A antagonist bicuculline (10 µM) and the GABA_C antagonist TPMPA (200 µM) reversibly reduced all three currents. Thus GABA receptor-mediated currents in hippocampal interneurons have a picrotoxin-resistant component, with features of both GABA_A and GABA_C receptors.

Pentobarbitone (100 µM) increased the amplitude of both the synaptic IPSC and the GABA-evoked current recorded in 100 µM picrotoxin. Furthermore pentobarbitone as well as zolpidem (0.2 µM) significantly slowed the decay of IPSCs, which is characteristic of GABA_A but not GABA_C receptors.

GABA receptors in hippocampal interneurons thus represent a pharmacologically diverse population, with pharmacological features that do not correspond to recognised subunit combinations. They may represent novel pharmacological targets that could alter network excitability.