NMDAR-LTD is usually associated with spine shrinkage and loss. The changes in synaptic strength and dendritic spines during LTD can be maintained for prolonged periods of time. Long-lasting structural and functional modification of synapses is an essential cellular substrate for information storage in the brain (Hofer, Mrsic-Flogel et al. 2009, Fu, Yu et al. 2012). Not only is it central to normal brain development and function, it has also been shown to play an important role in the pathophysiology of brain disorders, including psychiatric disorders. The mechanisms underlying the structural and functional plasticity in LTD overlap partially. For instance, they both require NMDA receptors, calcineurine and actin depolymerization, but AMPA receptor endocytosis and protein phosphatase 1 are only involved in synaptic depression. Despite intensive study of synaptic plasticity, however, the molecular mechanisms underlying spine remodeling associated with LTD, especially long-term maintenance of changes in spines, are still largely unclear. During the current review period, we have systematically investigated the role of miRNAs in LTD and spine remodeling associated with it. We used next-generation deep sequencing to identify miRNAs differentially expressed in LTD, along with a bioinformatics tool (which we developed in-house) to reveal genes and cellular processes enriched by them. Remarkably, we found that LTD induction leads to a global change in miRNA transcriptomes with potentially pleiotropic effects on cellular processes, many of which modulate the properties and functions of synapses. To evaluate the physiological consequence of miRNA expression changes in LTD, we tested the effects of two differentially expressed miRNAs, miR-191 and miR-135, on synaptic and spine plasticity using electrophysiology in hippocampal slices and time-lapse imaging of live neurons in primary cultures. We found that both miRNAs are necessary for LTD induction, persistence of spine shrinkage and delayed spine elimination, but neither is required for induction of spine shrinkage. The roles of miR-191 and miR-135 in LTD maintenance, however, are unclear as LTD is not induced in cells when their expression is perturbed. Future studies characterizing other putative LTD miRNAs are needed to determine the role of miRNAs in LTD maintenance. While investigating the mechanisms by which miR-191 and miR-135 regulate spine remodeling, we found that their target genes tropomodulin 2 (for miR-191) and complexin-1 and -2 (for miR-135) mediate their functions in spine plasticity. Intriguingly, we found that: 1) translation is required for long-lasting spine shrinkage. This is a surprising finding given that smaller spines contain relatively fewer AMPA receptors and postsynaptic density proteins;2) actin depolymerization is regulated in a sustained, protein synthesis-dependent manner to support prolonged spine plasticity. Actin depolymerization is a well-known mechanism to trigger cell remodeling. Our study, however, shows that actin cytoskeleton needs to be continuously modified beyond the initial phase of spine restructuring to maintain the morphological change of cells;3) AMPA receptor exocytosis is reduced by LTD induction. AMPA receptor exocytosis is a predominant mechanism for LTP, and is not thought to be altered in LTD. However, our data show that AMPA receptor exocytosis is suppressed by LTD induction to support long-lasting spine remodeling.
