טכניון מכון טכנולוגי לישראל
הטכניון מכון טכנולוגי לישראל - בית הספר ללימודי מוסמכים  
Ph.D Thesis
Ph.D StudentHazan Liran
SubjectShaping of Synaptic Size Distributions by Activity
Dependent and Independent Processes
DepartmentDepartment of Medicine
Supervisor Professor Noam Ziv


Abstract

Synaptic diversity is generally believed to be driven by activity-dependent modifications to synaptic connections, a phenomenon known as synaptic plasticity. This diversity is manifested in broad, heavy tailed distributions of synaptic sizes, better approximated by lognormal than normal distributions. Such distributions have been repeatedly suggested to arise from myriad forms of synaptic plasticity. Conversely, recent studies suggest that these might arise from activity independent processes, namely spontaneous size fluctuations driven by innate dynamics and continuous turnover of synaptic molecules. Thus, the relative contribution of activity dependent and independent processes in shaping these synaptic size distributions remains unclear.

Here we set out to study relationships between activity levels, synaptic size fluctuations, size diversity and distributions as well as relationships between changes in synaptic sizes and changes in important aspects of network function. To that end, rat cortical networks, raised on multielectrode arrays, were exposed from the moment of plating to pharmacological agents that block network activity and excitatory synaptic transmission. Synaptic sizes were determined by confocal imaging of synapses expressing a fluorescently-tagged variant of the PSD protein PSD-95 (PSD-95:EGFP). We found that in networks with no history of activity, size distributions of glutamatergic synapses were approximately log-normal and stable. Chronic silencing was associated with somewhat reduced size fluctuations, larger synaptic sizes and broader size distributions, yet distribution shapes in silenced and active networks were practically identical when scaled by mean synaptic size.

Further analysis revealed that silencing seemed to weaken constraints on synaptic size distributions- excessive growth, on the one hand, and synaptic loss, on the other. Yet, size fluctuations in chronically silenced networks were sufficient to generate stable, lognormal-like size distributions. To explore the source of these fluctuations and their relationships with innate molecular dynamics, we used a previously published mesoscopic model which ties these two levels in order to generate potential explanations which were then tested experimentally. Two such explanations- slower molecular exchange kinetics or increased synaptic protein levels were not supported by subsequent experiments. However, a third explanation- a reduction in total synapse numbers- was fully supported by experimental observations found in silenced networks.

To determine whether the observed changes in synaptic sizes and numbers impact network function, we compared these to changes in activity propagation pathways in the same networks. To that end, networks were stimulated periodically from multiple locations, followed by recordings of spatiotemporal propagation patterns of evoked activity. The covariance of changes in synaptic sizes/numbers and propagation pathways was then quantified. We found weak but statistically significant covariance between some structural and functional measures, indicating that changes in synaptic sizes detected by imaging-based approaches are at least partially reflected in physiologically relevant aspects of network function.

In summary, our findings indicate that synaptic size distributions are primarily shaped by activity-independent processes, with activity levels mainly setting their scale, rather than their shape. Furthermore, our findings integrate phenomena typically explored separately- synaptic molecule dynamics, size fluctuations, size distribution shape, distribution scale and activity-dependent synapse formation- into a unified view of these phenomena as tightly interconnected processes.