Rescue of neuropsychiatric phenotypes in a mouse model of 16p11.2 duplication syndrome by genetic correction of an epilepsy network hub

Abstract

Neuropsychiatric disorders (NPDs) share genetic etiology and are frequently co-morbid with epilepsy, but the biological basis of this shared risk remains poorly understood. The 16p11.2 microduplication (16p11.2dup/+) is a highly pleiotropic copy number variant (CNV) conferring risk for multiple NPDs including autism spectrum disorder, schizophrenia and intellectual disability, and is associated with a high prevalence of seizures. We used a mouse model of the 16p11.2 duplication (16p11.2dup/+) to uncover molecular and circuit properties associated with this broad phenotypic spectrum, and examined genes within the locus capable of phenotype reversal. Quantitative proteomics of cortical membranes revealed alterations to synaptic protein networks and products of diverse NPD risk genes in 16p11.2dup/+ mice. Network analysis identified an epilepsy-associated protein subnetwork, which was dysregulated in 16p11.2dup/+ mice and proteomic datasets from human NPDs. We investigated circuit properties in 16p11.2dup/+ mice and found they exhibited hypersynchronous activity and enhanced network glutamate release, which increased susceptibility to seizures. We hypothesized that a regulator of the synaptic and epilepsy-associated protein network could have an important impact on pathophysiology. Human brain co-expression and interactome analysis revealed PRRT2 as a major hub in the dysregulated epilepsy subnetwork. Remarkably, restoring Prrt2 copy number to wild-type levels rescued aberrant circuit properties, seizure susceptibility and social interaction deficits in 16p11.2dup/+ mice. We show that proteomics and network biology can identify important disease hubs in multigenic CNVs, and reveal molecular and circuit phenotypes which may be relevant to the complex symptomatology of 16p11.2 duplication carriers.

iPSC-derived models of PACS1 syndrome reveal transcriptional and functional deficits in neuron activity

“…Traces were processed in Matlab (vR2022a) using code developed by Marc Dos Santos. [91][92][93] In brief, data were normalized to a rolling median background of 25 seconds. Peaks were detected using the findpeaks function (MinPeakDistance = 3 seconds, MinPeakHeight = Δ0.15 F/F0, MinPeakProminence Δ0.08 F/F0, MaxPeakWidth = 10 seconds).…”