When to learn and when to forget : NMDA normalization in hippocampal neurons-- activity-dependent, temporal and spatial properties.

Abstract

Synaptic plasticity is the substrate of a vast variety of learning mechanisms. However, the molecular pathways and physiological patterns that regulate it are poorly understood. In the first part of this thesis, we focus on the physiological determinants of pre-synaptic plasticity. We find that a complete blockade of activity succeeds in inducing only a transient enhancement of plasticity. A permanent enhancement of synaptic plasticity is achieved by selectively reducing the NMDA-R mediated Ca2+ flux associated with uncorrelated activity, via adjustment of the voltage-dependent Mg2+ block of NMDA receptors. NR2B-containing NMDARs are up-regulated by this treatment, and this is found to be an important contributor to plasticity enhancement. Thus, the quality, but not the quantity, of activity is the important parameter to manipulate to obtain a permanent enhancement of intrinsic plasticity. In the second part, we study the relationship between activity patterns and postsynaptic NMDA-R regulation by using high-precision iontophoresis. We identified a new homeostatic mechanism of NMDA-R regulation which we have thus termed "NMDA Normalization" to differentiate it from prior uses of "NMDA homeostasis." Through this novel Ca++ dependent mechanism, we show that the neuron, by opposing NMDA-R functional expression counter to activity changes, causes the average charge transfer through NMDA-Rs to remain constant. We elucidate the activity-dependent, temporal, and spatial characteristics of this process. We propose an explanatory hypothesis. The reduction of uncorrelated activity used in the first part reduced the NMDA-R mediated average charge transfer. Through normalization, the cell increased functional NMDA-R expression to normalize NMDA-R mediated average charge transfer. However, when a burst of activity arrives in this new condition, in which an identical burst of glutamate release and depolarization now meets a larger number of very weakly activated NMDA-Rs, it is able to induce a disproportionately larger peak Ca++ flux. This increase in the maximal Ca++ flux, due to the combination of reduction of uncorrelated activity and NMDA-R normalization, is responsible in great part for the enhancement of synaptic plasticity. Our findings propose a clear strategy for the development of compounds to restore, and potentially enhance, synaptic plasticity, and thus with likelihood learning and memory.