Mirror Effects of 4 Neurodevelopmental CNVs on General Functional Connectivity and Implications for Idiopathic Autism

Background: Copy number variants (CNVs) at the 16p11.2 and 22q11.2 loci are among the most frequent risk factors for autism and schizophrenia. Almost nothing is known, however, on the alterations of functional connectivity (FC) associated with these genomic variants. fMRI studies of ASD (Autism Spectrum Disorder) have resulted in heterogeneous findings yet there is a paucity of research into homogeneous groups of individuals that share the same genetic ASD risk factors.

Objectives: Our aims are : 1) To characterize the functional impact on brain connectivity of 4 CNVs at the 16p11.2 and 22q11.2 loci. 2) To investigate whether these CNV-associated patterns are observed in a heterogeneous group of idiopathic autism at a level consistent with the risk conferred by each genetic variant. 3) To explore the similarity of spatial patterns of FC alterations associated with each CNV.

Methods: To this mean, we have analyzed rs-fMRI data from n=105 deletion and duplication carriers and 128 controls from the Simons VIP dataset and UCLA. We then correlated the pattern of connectivity of each CNV to patterns observed in n= 276 autism and n=290 control subjects from the ABIDE dataset. fMRI data were preprocessed with NIAK pipeline and FC was estimated as the pairwise correlation between the average time series of 64 brain regions. A linear model was fitted to each connection with genetic status, subject motion, sex and site as explanatory factors. The effect of the 4 genetic groups were then tested by post-hoc contrasts. Each contrast was controlled for false discovery rate (FDR) across all connections at q < 5%. The 16p11.2 and 22q11.2 connectivity patterns were compared to the autism cohort using spatial correlation of the individual connectome with the CNV derived pattern of connectivity alterations.

Results: The 16pDel is associated with an overall increase in connectivity while the 22qDel is associated with an overall decrease in connectivity compared to control. An enrichment of their individual profiles is reported in an idiopathic ASD sample in 4 seed regions (Putamen, Thalamus, perigenual anterior Cingulate, and dorsomedial Prefrontal Cortex). Finally, a functional covariance pattern between 16pDel and 22qDel is observed in the Fronto-parietal, Limbic, and Anterior DMN networks.

Conclusions: Deletions and reciprocal duplications are associated with a mirror effect at the global and individual connection level for both the 16p11.2 and 22q11.2 CNV loci. In addition, FC alterations exhibit opposing relationships for DEL (and DUP) carriers at both loci: 16p11.2 DEL is characterized by overall over-connectivity while the reverse is true for 22q11.2 DEL carriers. Striato-striatal and striato-cortical over-connectivity in the DEL carriers is consistent with previous reports of aberrant functional connectivity in ASD. Despite the opposing pattern of FC alteration exhibited by DEL carriers of each locus, we find an enrichment of both in an idiopathic autism sample. This is likely conferred by the similar relative distribution of FC alterations (convergence of covariance connectivity patterns between 16p11.2 and 22q11.2 deletions).