Stable dynamics of synaptic and cellular networks in the hippocampus

Abstract

During learning, experience-dependent activity triggers synaptic plasticity, modifying synaptic strength. Most excitatory inputs form synapses on dendritic spines, and changes in synaptic strength result in structural modifications of spines. Thus, memory traces formed during learning have been hypothesized to be stored at synapses and therefore spines. Formation and consolidation of declarative, episodic memories require the hippocampus, and the Schaffer collateral synapse between areas CA3 and CA1 is considered one of the prototypical small glutamatergic synapses in the central nervous system. The formation of new memories implies a dynamic restructuration of the network, which could conflict with the necessity for memory stability required to perform a task, at least on a short timescale. In CA1, the lifetime of dendritic spines has been followed for days, to determine the capacity for synapses to store memory traces. However, no consensus has been found between studies. So far only spine morphology has been investigated over multiple days in the hippocampus to assess spine lifetime and chronic analyses of synaptic function have been missing. Therefore, we are limited in our understanding of how structural changes relate to synaptic function in vivo. On a first project, I combined chronic two-photon calcium imaging of ipsilateral CA1 spines with repeated optogenetic activation of presynaptic contralateral CA3 pyramidal neurons in the awake mouse. Using this approach, I induced local, synaptically evoked calcium responses at individual spines and assessed the stability of these functionally identified synapses over more than two weeks. These responding spines tend to form functional clusters of strongly connected spines and are more stable than non-responding spines on the same dendrites, suggesting that strong synaptic connectivity is associated with spine persistence. Taken together, this work suggests that spine lifetime in the hippocampus is related to synaptic weight, which may determine long-term synaptic connectivity. Dynamic restructuration of the network has also been observed through the gradual remapping of spatial representation in the hippocampus. While the stability of the behavior of an animal navigating an environment has been shown, the representation of the environment is drifting over days. The cellular mechanism subtending this formation and elimination of place cells remains unclear. As the CA3 subdivision of the hippocampus exhibits a spatial representational drift, how this phenomenon accelerates the spatial code remapping in CA1 remains unclear. In this second project, I combined chronic two-photon calcium imaging of ipsilateral CA1 spines with optogenetic activation of presynaptic contralateral CA3 pyramidal neurons in the mouse performing a spatial navigation task. Using this approach, I was able to induce a subset of place cells in CA1, whose place fields tend to form close to the reward zone rather than to the stimulation zone. Additionally, I show that this perturbation induced a change in the overall population code of already existing place cells, although only temporarily. This work suggests that, while the network is in a stable state, a modification of the inputs from presynaptic partners led to a remapping of the environment in CA1, suggesting that the network in CA1 is gradually pushed to a new state by presynaptic partners.