1. Long-term depression (LTD) is an activity-dependent reduction in the strength of synaptic transmission that can persist for hours. It is a neural model for processes underlying learning and memory, such as extinction and forgetting. LTD of excitatory postsynaptic potentials (EPSPs) in cells of the CA1 region of hippocampal slices can be induced in an anti-Hebbian paradigm, i.e., by conditioning stimuli that activate the postsynaptic neuron in the absence of evoked synaptic transmission in the test pathway. Past work showed that LTD was not produced consistently in a pharmacologically untreated slice, but it could be induced more reliably when the conditioning stimuli were applied during block of evoked transmitter release. We have now defined further the conditions in which LTD can be obtained using postsynaptic conditioning by investigating 1) whether intracellular conditioning is effective, 2) the requirement for extracellular Ca2+, and 3) the consequences of selective block of glutamate ionotropic receptor subtypes during the conditioning procedure. 2. Intracellular recordings were made from CA1 pyramidal neurons. Test shocks were applied to the stratum radiatum except during conditioning, and the depolarizing slopes and amplitudes of evoked EPSPs were measured. The conditioning procedure activated the postsynaptic neuron either antidromically (via trains of shocks at 100 Hz applied to the axons in the alveus) or intracellularly (via depolarizing pulses of 1.5-3.5 nA). During conditioning, postsynaptic potentials (PSPs) evoked by the conditioning stimuli either were transiently blocked by bathing slices for 5 min in artificial cerebrospinal fluid (CSF) containing a high [Mg2+] or were reduced by glutamate antagonists. 3. When slices were bathed in CSF containing 25 mM Mg2+ and 2 mM Ca2+, evoked PSPs were transiently abolished; conditioning, either by antidromic or intracellular stimulation, always evoked a significant LTD. During the LTD produced by antidromic stimulation, the mean EPSP slope was 52.6 +/- 11.4% (mean +/-SE) of its control at 30-35 min after conditioning (n = 7). The LTD produced by intracellular conditioning was of similar magnitude: the mean EPSP slope was 57.2 +/- 11.6% of its control at 30-35 min postconditioning (n = 7). When slices were bathed in CSF containing 25 mM Mg2+ and 2 mM Ca2+ without conditioning stimuli, there was no LTD (mean EPSP slope 109 +/- 8.1% of its control at 30-35 min after reperfusion with CSF; n = 5). 4. However, when antidromic conditioning stimuli were applied in CSF with 15 mM Mg2+ but no added Ca2+, LTD was not induced (mean EPSP slope 102.7 +/- 2.8% of its control at 30-35 min postconditioning; n = 6), showing that extracellular Ca2+ is needed for the induction of LTD. 5. An activity-dependent increase in cytosolic [Ca2+] occurred when the slice was bathed in CSF containing 25 mM Mg2+ and 2 mM Ca2+, as shown by the fluorescent Ca2+ indicator FURA-2 in the postsynaptic cell. A smaller but significant activity-dependent increase in cytosolic [Ca2+]was found when the slice was bathed in CSF containing 15 mM Mg2+ but no added Ca2+ 6. LTD was also produced after intracellular conditioning stimuli applied when evoked PSPs in the slice had been transiently reduced by 5 min of application of the selective glutamate receptor antagonist 6-cyano-7-nitroquinoxaline-3,3-dione alone (5-10 muM) or together with D-2-amino-5-phosphonopentanoic acid. During LTD, the mean EPSP slope was 70 +/- 5.7% of its control at 40-45 min after conditioning (n = 5). Recovery from the drugs' effects took longer than from the high-[Mg2+] solutions: in the absence of intracellular conditioning, the mean EPSP slope recovered to 101.3 +/- 5.5% of its control at 40-45 min after reperfusion with normal medium (n = 10). 7. We conclude that a significant nonassociative LTD of EPSPs can be reliably produced by postsynaptic depolarization and firing when there is a pharmacological reduction of synaptic transmission during the conditioning procedure. We have shown that extracellular Ca2+ is necessary for the postsynaptic induction of LTD. We propose that the importance of the pharmacological reduction of synaptic transmission during conditioning lies mainly in the reduction of PSPs evoked by the conditioning stimuli. This could have two effects: a reduction in recurrent inhibitory postsynaptic potentials would decrease the membrane conductance, thereby increasing the space constant of the postsynaptic cell and facilitating the entry of critical amounts of Ca2+ into the dendrites during conditioning. A reduction in EPSPs would reduce the probability of the conditioning stimuli eliciting long-term potentiation and facilitate the unopposed production of LTD.
