Altered Prefrontal Functional Connectivity and Anatomy in Mice Lacking Autism-Associated Gene Shank3

Background: Mutations in postsynaptic scaffolding protein Shank3 have been strongly implicated in 22q13 deletion syndrome (Phelan–McDermid syndrome) as well as other non-syndromic forms of autism spectrum disorder (ASD). A number of recent reports have investigated basal ganglia function in mouse lines harboring Shank3 mutations, revealing specific striatal circuit impairments underlying self-injurious repetitive grooming exhibited by these mutants. However, Shank3 mutations in humans are often associated with intellectual disability and cognitive impairments, leading to the hypothesis that functional alterations in this gene can affect extra-striatal brain substrates, including higher order neocortical regions.

Objectives: Our work probes the hypothesis that Shank3 mutations impair function and anatomy of extra-striatal neocortical substrates, including higher-order cognitive substrates. To this aim, we used high resolution structural MRI and in vivo resting-state fMRI (rsfMRI) to map regional gray matter volume and brain-wide functional connectivity in male Shank3B^-/- mutants.

Methods: Adult male Shank3B^-/- (n=11) and control wild type littermates (n=10) were imaged at 7T under shallow halothane anesthesia (0.75%) using a single-shot EPI sequence as previously described (Liska et al., 2017, Cerebral Cortex). The protocol preserves large-scale functional connectivity in rodents (Gozzi & Schwarz, 2016, NeuroImage). We calculated global and local brain connectivity maps for all subjects and mapped voxelwise inter-group differences in both measures. We also employed post-mortem voxel based morphometry (VBM) and automated anatomical labeling on high resolution images obtained with a FLASH sequence with an isotropic voxel size of 70µm. The same brains underwent DTI imaging using 80 directions, a 120x120x240 voxel size, and b=3000, d=6 and D=13 ms as recently described (Liska et al., 2017, Cerebral Cortex).

Results: Whole-brain rsfMRI connectivity mapping revealed large foci of reduced global and local connectivity in prefrontal and anterior cingulate hub regions of the mouse brain (A-B). This effect was associated with diminished long-range connectivity in midline integrative areas of the mouse default mode network (C), as well as impaired striato-cortical coupling (D). A correlation between impaired ultrasound vocalization and decreased prefrontal connectivity was also observed. No overt genotype-dependent alterations in white macrostructure or fiber-organization were observed, as documented by DTI-based mapping and tractographic analyses (G). In keeping with the observed functional alterations, prefrontal and parietal regions exhibiting impaired connectivity also showed prominent reductions in gray matter volume (E-F), suggesting of a hypotrophic maturation of these areas. Cortical regions of Shank3^-/- mutants also exhibited idiosyncratic variability in homotopic connectivity.

Conclusions: We show that Shank3B deletion leads to cortical volumetric loss and reduced local and long-range functional connectivity within prefrontal and posterior connectivity hub region, an effect that is associated with social communication deficits. These findings suggest a pivotal role for Shank3B in the establishment of functional and neuroanatomical coupling between integrative higher order neocortical areas, and suggest that mutations in this gene may predispose to autism via a selective trophic and functional downregulation of prefrontal circuits.