To keep or not to keep: The molecular mechanisms of activity-dependent synapse refinement

Abstract

To establish efficient neural networks, neurons selectively strengthen active synapses and eliminate inactive ones through a neuronal activity-dependent developmental process known as synaptic refinement. Aberrant synaptic refinement underlies several neuropsychiatric disorders. However, the molecular mechanisms that detect neuronal activity and accordingly determine synaptic connectivity remain to be elucidated. Active synapses undergo synapse maturation, a process regulated by the transmembrane protein, SIRPα. In the presence of synaptic activity, SIRPα’s ectodomain is cleaved and released to strengthen synapses. I determined how SIRPα detects synaptic activity to function at and strengthen active synapses. I showed that SIRPα tyrosine-phosphorylation, which is driven by neuronal activity, drives SIRPα cleavage, which in turn, stabilizes synapses in an activity-dependent manner. This work identified a new role for SIRPα phosphorylation and established a novel mechanism through which neuronal activity regulates SIRPα function to promote proper refinement. Inactive synapses are eliminated. We demonstrated that inactive synapses in the brain are eliminated only when there are active connections to compete with. This suggests that active connections send ‘punishment’ signals that trigger ‘elimination’ signals within inactive synapses and instruct them to leave. However, the molecular signals that determine inactive synapse elimination are unknown. Here, we identified the tyrosine kinases, Pyk2 and JAK2 as determinants of inactive synapse elimination. I showed that Pyk2 is necessary for functional and structural synaptic elimination, and is activated at inactive synapses, but only in the presence of other active ones. JAK2, a Pyk2-interacting kinase activated at inactive synapses, is necessary and sufficient for synaptic elimination. This work uncovered key mechanisms that dictate the elimination of inactive synapses. Finally, I reasoned that ‘punishment’ signals that drive the elimination of inactive synapses may target other active synapses as well. To prevent the elimination of active synapses, I proposed that a ‘stabilization’ signal signals at and protects active synapses. I hypothesized that signals that strengthen active synapses, i.e. SIRPα, may also stabilize them. Indeed, we showed that SIRPα suppresses Pyk2/JAK2 activation and prevents synaptic elimination. Together, this work established the molecular bridges that link neuronal activity to synaptic connectivity, thus, furthering our understanding of brain development and function.