ELECTROPHYSIOLOGICAL EVIDENCE THAT DENTATE HILAR MOSSY CELLS ARE EXCITATORY AND INNERVATE BOTH GRANULE CELLS AND INTERNEURONS

1. The hypothesis that dentate hilar ''mossy'' cells are excitatory was tested by simultaneous intracellular recording in rat hippocampal slices. Messy cells were recorded simultaneously with their potential targets, granule cells and interneurons. The gamma-aminobutyric acid-A (GABA(A)) receptor antagonist bicuculline was used in most experiments to block the normally strong inhibitory inputs to granule cells that could mask excitatory effects of messy cells. Some cells were recorded with electrodes containing the marker Neurobiotin so that their identity could be confirmed morphologically. 2. A mossy cell action potential was immediately followed by a brief depolarization in a granule cell in 20 of 1,316 pairs (1.5%) that were recorded in the presence of bicuculline. The mean amplitude of depolarizations was 1.99 +/- 0.24 (SE) mV when the postsynaptic membrane potential was -55 to -65 mV. Depolarizations could trigger an action potential if the granule cell was depolarized from its resting potential so that its membrane potential was -50 to -60 mV. These data suggest that messy cells excite granule cells monosynaptically. 3. Monosynaptic excitation of an interneuron by a messy cell was recorded in 4 of 47 (8.5%) simultaneously recorded messy cells and interneurons, also in the presence of bicuculine. The mean interneuron depolarization was 1.64 +/- 0.29 mV when the interneuron membrane potential was approximately -60 mV. When an interneuron was at its resting potential (-52 to -63 mV), action potentials were often triggered by the depolarizations. 4. Without bicuculline present, messy cells had no apparent monosynaptic effects on granule cells, as has been previously reported. However, effects that appeared to be polysynaptic were observed in 5 of 92 pairs (5.4%). Specifically, a small, brief hyperpolarization occurred in granule cells 2.5-7.3 ms after the peak of a messy cell action potential. Given the results indicating that messy cells excite interneurons, and the long latency to onset of the hyperpolarization, one possible explanation for the hyperpolarization is that messy cells excited interneurons that inhibited granule cells. 5. The results suggest that messy cells are excitatory neurons. In addition, messy cells appear to innervate both granule cells and interneurons that are located within several hundred micrometers of the messy cell soma. The only detectable effect on granule cells in this area under normal conditions appears to be disynaptic and inhibitory. However, when GABA(A)-receptor-mediated inhibition is blocked, monosynaptic excitation of granule cells by messy cells can be detected.