1. Whole cell patch-clamp recordings were made from CA1 stratum oriens
inhibitory neurons of rat hippocampal slices in vitro to determine th
eir contribution to the epileptiform activity elicited by elevating th
e extracellular potassium ion concentration ([K+](o)) from 3.5 to 8.5
mm. 2. Under current-clamp conditions, spontaneous action potential ac
tivity in inhibitory neurons normally occurs in a sustained repetitive
firing mode paced by nonsynaptic, intrinsic mechanisms. On elevation
of [K+](o) to 8.5 mm the pattern of activity is altered such that clus
ters of action potentials occur interrupted by periods of silence with
out an appreciable afterhyperpolarization (AHP). In addition, elevatio
n of [K+](o) caused a large reduction in the action potential AHP ampl
itude and duration concomitant with a 20-mV shift in the reversal pote
ntial of the AHP. 3. In voltage clamp a small persistent inward curren
t was observed after the introduction of elevated potassium concomitan
t with an increase in the frequency of spontaneous excitatory postsyna
ptic currents (EPSCs) in all interneurons studied. After a short perio
d of time (similar to 1 min) temporal summation of synchronously occur
ring EPSCs contributed a periodic inward current (PIC; 10-40 pA, 0.8 H
z) that persisted for the duration of the [K+](o) elevation. Analysis
of the charge transfer associated with the PIC suggests that they comp
rise the temporal summation of similar to 35 EPSCs. This PIC was synch
ronous with the extracellular field potential recorded from the CA1 py
ramidal neuron layer. 4. The PIC was responsible for the clustering of
action potential activity because blockade of EPSC activity by the lp
ha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor
antagonist 6,7-dinitroquinoxaline-2,3-dione (DNQX) abolished PICs and
reverted action potential activity to single sustained firing, despite
the continued application of 8.5 mm [K+](o). Antagonists of N-methyl-
D-aspartate receptors were without effect on either the PICs or the ac
tion potential activity. 5. Addition of the metabotropic glutamate rec
eptor (mGluR) antagonist (+)-2-methyl-4-carboxyphenylglycine (MCPG) re
versibly abolished the PIC without affecting the increase in EPSC freq
uency. 6. Recordings from CA3 pyramidal neurons in 8.5 mm [K+](o) demo
nstrated that interictal activity occurred at a frequency identical to
the PICs observed in interneurons. Interictal activity in CA3 pyramid
al neurons was attenuated but never abolished by MCPG, suggesting a ro
le for mGluR receptors in the maintenance of interictal activity in ar
ea CA3. 7. Flurries of synchronous inhibitory postsynaptic currents (I
PSCs) were observed on a small number of CA1 pyramidal neurons on elev
ation of [K+]o, consistent with these neurons being the synaptic targe
t of stratum oriens/alveus interneurons. Unexpectedly, a rundown of th
is IPSC activity was observed during the transition to CA1 pyramidal n
euron epileptiform activity. 8. In CA1 pyramidal neurons, current-volt
age relationships measured at the peak response to gamma-aminobutyric
acid (GABA) were identical in both normal or elevated [K+](o). However
, subsequent current-voltage relationships determined during the fade
of the GABA response revealed a positive shift in the reversal potenti
al in both control and 8.5 mm [K+](o) conditions. The mechanism for th
is positive shift in reversal potential is at present un known but may
occur due to the intracellular accumulation of Cl- after sustained GA
BA receptor activation. 9. These experiments suggest that the ongoing
intrinsic action potential activity of the CA1 interneuron population
is ''overridden'' during electrographic activity resulting from [K+](o
) elevation. The resulting activity is paced by interictal events occu
rring in the CA3 pyramidal neuron population. A role for mGluRs in the
interictal activity of the CA3 pyramidal neuron population is also de
scribed. After the establishment of epileptiform activity, synaptic in
hibition onto CA1 pyramidal neurons is severly attenuated. This result
s not from the [K+](o) elevation per se but possibly from an activity-
dependent Cl- redistribution across the CA1 pyramidal neurons after su
stained IPSC activity.