Abstract
Synaptic scaling is an important form of homeostatic plasticity that bidirectionally adjusts postsynaptic strengths to keep neuronal activity close to a homeostatic set point. Synaptic scaling is defective in a number of monogenic rodent models of Autism Spectrum Disorders (ASDs), including loss of function of the multidomain synaptic scaffolding protein Shank3. In work that I contributed to, our lab established that phosphorylation and dephosphorylation of Shank3 at two key sites, S1586/S1615, blocks synaptic scaling up and down, respectively (Chapter 2, Wu et al., 2022). This study left unresolved what molecular mechanisms downstream of Shank3 dephosphorylation are necessary for the expression of synaptic scaling up. To first characterize my synaptic scaling up protocol, I examined the correlations between intensities of surface AMPAR subunit, GluA2, and presynaptic marker, VGluT1. I find that both control and TTX-induced scaling up conditions have high correlations between these two proteins and that scaling up increases the intensities of these two proteins, although to a greater magnitude for sGluA2. I then investigated whether Homer1, a widely studied interactor of Shank3 with relevance to synaptic scaling, was involved. I found that phosphorylated Shank3 had reduced binding ability to long-form Homer1. This led me to question whether phosphomimetic Shank3 would undergo changes in its colocalization with endogenous Homer1 compared with wildtype Shank3. Moreover, presence of Shank3 phosphomimetic mutant causes loss in association with long-form Homer1. Previous literature shows that group I mGluR agonist-independent activity resulting from Homer1a-mGluR binding is required for scaling down. I therefore hypothesized that the complementary pathway, consisting of long-form Homer1 binding to mGluR that causes a switch to agonist-dependent activity, is required for synaptic scaling up. Indeed, I find that agonist-dependent activity of these mGluRs is required for synaptic upscaling. I then wondered whether this mGluR activity was sufficient to enable synaptic scaling up on its own. I found that the ability to undergo scaling up in the presence of phosphomimetic Shank3 was rescued by enhancement of mGluR5 activity using a positive allosteric modulator. Finally, I addressed whether specific molecular targets of the downstream group I mGluR canonical pathway are involved in synaptic scaling and found that PLC and PKC activity are each required for synaptic scaling up. However, depletion of intracellular calcium stores did not affect synaptic upscaling. Overall, this work strongly supports the novel idea that synaptic upscaling requires interaction between Shank3 and long-form Homer1 that leads to a switch from constitutive to agonist-dependent activity of group I mGluRs, thus providing a model that is complementary to the prevailing understanding of synaptic downscaling.