Systems biology perspectives on cerebellar long-term depression

Long-term depression (LTD) at parallel fiber-Purkinje cell (PF-PC) synapses is thought to be the cellular correlate of cerebellar associative learning. The molecular processes are, in brief, phosphorylation of AMPA-type glutamate receptors (AMPARs) and their subsequent removal from the surface of the PF-PC synapse. In order to elucidate the fundamental mechanisms for cerebellar LTD and further the understanding of its computational role, we have investigated its systems biology and proposed the following hypotheses, some of which have already been experimentally verified: (1) due to the mitogen-activated protein kinase (MAPK)-protein kinase C (PKC) positive feedback loop, phosphorylation of AMPARs is an all-or-none event; (2) the inositol 1,4,5-triphosphate receptor detects concurrent PF and climbing fiber inputs, forming the cellular basis for associative learning, and (3) the local concentration of nitric oxide in the PC dendrite reflects the relevance of a given context, enabling context-dependent selection of learning modules within the cerebellum. In this review, we first introduce theoretical studies on cerebellar LTD, mainly focusing on our own published work, followed by a discussion of the effects of stochasticity, localization, diffusion, and scaffolding. Neurons embody two features that are apparently contradictory, yet necessary for synaptic memory: stability and plasticity. We will also present models for explaining how neurons solve this dilemma. In the final section, we propose a conceptual model in which a cascade of excitable dynamics with different time scales, i.e., Ca(2+)-induced Ca(2+) release, the MAPK-PKC positive feedback loop, and protein kinase Mzeta (PKMzeta)-induced PKMzeta synthesis, provides a mechanism for stable memory that is still amenable to modifications.