Infant avoidance training alters cellular activation patterns in prefronto-limbic circuits during adult avoidance learning: II. Cellular imaging of neurons expressing the activity-regulated cytoskeleton-associated protein (Arc/Arg3.1).

Positive and negative feedback learning is essential to optimize behavioral performance. We used the two-way active avoidance (TWA) task as an experimental paradigm for negative feedback learning with the aim to test the hypothesis that neuronal ensembles activate the activity-regulated cytoskeletal (Arc/Arg3.1) protein during different phases of avoidance learning and during retrieval. A variety of studies in humans and other animals revealed that the ability of aversive feedback learning emerges postnatally. Our previous findings demonstrated that rats, which as infants are not capable to learn an active avoidance strategy, show improved avoidance learning as adults. Based on these findings, we further tested the hypothesis that specific neuronal ensembles are "tagged" during infant TWA training and then reactivated during adult re-exposure to the same learning task. Using cellular imaging by immunocytochemical detection of Arc/Arg3.1, we observed that, compared to the untrained control group, (1) only in the dentate gyrus the density of Arc/Arg3.1-expressing neurons was elevated during the acquisition phase of TWA learning, and (2) this increase in Arc/Arg3.1-expressing neurons was not specific for the TWA learning task. With respect to the effects of infant TWA training we found that compared to the naïve non-pretrained group (a) the infant pretraining group displayed a higher density of Arc/Arg3.1-expressing neurons in the anterior cingulate cortex during acquisition on training day 1, and (b) the infant pretraining group displayed elevated density of Arc/Arg3.1-expressing neurons in the dentate gyrus during retrieval on test day 5. Correlation analysis for the acquisition phase revealed for the ACd that the animals which showed the highest number of avoidances and the fastest escape latencies displayed the highest density of Arc/Arg3.1-expressing neurons. Taken together, we are the first to use the synaptic plasticity protein Arc/Arg3.1 to label neuronal ensembles which are involved in different phases of active avoidance learning and whose activity patterns are changing in response to previous learning experience during infancy. Our results indicate (1) that, despite the inability to learn an active avoidance response in infancy, lasting memory traces are formed encoding the subtasks that are learned in infancy (e.g., the association of the CS and UCS, escape strategy), which are encoded in the infant brain by neuronal ensembles, which alter their synaptic connectivity via activation of specific synaptic plasticity proteins such as Arc/Arg3.1 and Egr1, and (2) that during adult training these memories can be retrieved by reactivating these neuronal ensembles and their synaptic circuits and thereby accelerate learning.