Dopaminergic modulation of the persistence of one-trial hippocampus-dependent memory

Abstract

The persistence of new memory traces in the hippocampus, encoded following appropriate activation of glutamatergic receptors and the induction of synaptic plasticity, can be influenced by heterosynaptic activation of neuromodulatory brain systems. We therefore investigated the effects of a hippocampus-specific blockade of dopamine D1/D5 receptors on the persistence of spatial memory encoded in one trial using a delayed matching-to-place (DMP) task in a watermaze in which rats learn a new escape location each day. A within-subjects design was used such that both short (20 min) and long (6 h) retention intervals, and both drug (SCH23390, a D1/D5 receptor antagonist) and vehicle (aCSF) infusions were tested on different days in the same animals. Bilateral intrahippocampal infusion of SCH23390 (5 microg in 1 microL per side) prior to trial 1 (encoding) caused a differential impairment as a function of memory delay-with no effect during trial 2 (memory retrieval) after a 20-min interval, but a block of memory at 6 h. Further experiments revealed that infusion of SCH23390 immediately after trial 1 had no effect on retention 6 h later, and the poor memory seen at long retention intervals when the drug was present at encoding was not due to a state-dependent failure of retrieval. These results suggest that activation of D1/D5 receptors during memory encoding is necessary for the formation of a persistent memory trace in the hippocampus. The complementary effects of D1/D5 receptor blockade on the persistence of LTP and the duration of memory are consistent with the idea that changes in synaptic strength underlie memory.

Figure 1.

Experimental design. (A) Varied platform location and memory delay across days. All experiments used the delayed matching-to-place (DMP) protocol in the watermaze, in which four trials are given each day; the platform position is fixed within a day, but moved to different locations between days. Swim paths are circuitous in trial 1 of each day, but the focus is on whether memory traces formed during this trial persist with any detectable strength until trial 2, which is scheduled after a variable delay. The point of trials 3 and 4 is to sustain the learned win-stay strategy that this protocol engenders. (B) Training conditions. There were three experiments, each preceded by a common set of 8 d of pretraining. Drugs were infused into the hippocampus at various times in relation to trials 1 and 2 as described.

Figure 2.

Pretraining. Acquisition of delayed matching-to-place. Absolute escape latency across the 8 d of pretraining during Experiment 1. Note the gradual acquisition of an effective win-stay strategy characterized by, over the last 2 d (shaded), a striking reduction in escape latency between trials 1 and 2 of each day. Escape latency in trial 1 remains stable at ca 60 sec, reflecting the novel daily location of the hidden platform. Means ± 1 SEM.

Figure 3.

Experiment 1. (A) Drug injections prior to trial 1—impact of memory delay. Normalized path length averaged across days during the drug-infusion phase, showing performance across the four trials of each day (n = 23). A delay-dependent deficit is apparent, with the condition in which the animals received SCH23390 and a long memory delay resulting in no memory in trial 2 of each day. The syringes’ symbols and dotted lines indicate the time of drug infusion. (B) Savings in path length between trials 1 and 2. The change in performance between trials 1 and 2 (i.e., savings) is plotted in terms of absolute path length. These data show exactly the same pattern as in the normalized scores. (C) Swim speed. The drug did transiently decrease swim speeds, but the effect is only apparent in trial 1 of each day, scheduled 15 min after the intrahippocampal drug infusion. Means ± 1 SEM.

Figure 4.

Experiment 2. (A) Drug injections after trial 1 (6-h memory delay). Normalized path length averaged across days during the drug-infusion phase, showing performance across the four trials of each day (n = 8). No impairment is observed in the SCH23390 condition with the drug infusion occurring 15 min after trial 1. (B) Savings in path length between trials 1 and 2. The data show an equivalent level of absolute savings in path length in both conditions. Means ± 1 SEM.

Figure 5.

Experiment 3. (A) Drug injection prior to trials 1 and 2 (6-h memory delay). Normalized path length averaged across days during the drug-infusion phase, showing performance across the four trials of each day (n = 8). A clear interaction is apparent to the effect that memory in trial 2 is impaired when the drug is infused on two occasions each day—before both trials 1 and 2. Conversely, the group receiving aCSF initially, but drug later, showed good memory. The impairment in Experiment 1 cannot be due to state dependency. (B) Savings in path length between trials 1 and 2. The same pattern is apparent in the absolute path-length data, with only the group having SCH23390 present at the time of encoding showing a memory deficit. Means ± 1 SEM.

Figure 6.

Histology. (A) Photomicrographs showing representative cannulae tracks into the dorsal hippocampus on each side of the brain. (B) Plots of the locations of the tips of the cannulae (n = 39 per hippocampus) as revealed in Nissl-stained brain sections. Plots of the locations of the tips of the cannulae (n = 39 per hippocampus) as revealed in Nissl-stained brain sections. Infusion sites are marked on the appropriate section of a stereotaxic brain atlas (reprinted with permission from Elsevier © 1998, Paxinos and Watson 1998 [Figs. 35–39]).