Interactions Dynamics of 16p11.2 Genes Across the Developing Human Brain

Background: Autism spectrum disorders (ASD) has a strong genetic component. Several Copy Number Variants (CNVs) have been firmly implicated in ASD. Recurrent CNV deletions and duplications of ~600 kb at chromosome 16p11.2 confer high susceptibility to ASD in up to 1% of ASD patients. Moreover, both de novodeletion and duplications of that CNV have been recurrently identified in ASD patients. Functional studies have shown that reciprocal deletion or duplication of 16p11.2 results in brain overgrowth or reduced brain volume in mice, respectively, further supported using zebra fish model. Recently, induced pluripotent stem cell technology is also starting to relate CNV deletions to cellular phenotypes in humans. Several binary protein-protein interactions (PPI) have already been mapped to 16p11.2 genes. However, the dynamic of interaction cold/hot spots of 16p11.2 genes throughput human brain development is still vastly unknown. 

Objectives: We focused our study on the pattern of protein interactions and co-expression of the 33 proteins encoded by 16p11.2 CNV genes. We aimed to identify particular co-expression patterns of interacting protein pairs in order to determine subsets of interactions occurring at specific stages of brain development. We present a temporal-spatial assessment of 16p11.2 genes interactions throughout early human brain developments, from early development (8 post conception weeks) to young adult (40 yrs).

Methods: We used the spatio-temporal gene expression dataset from BrainSpan project, comprising various brain regions across different stages of brain development to identify the time-space interaction cold/hotspots within the PPI network. Tissue samples were divided into four anatomic regions: prefrontal cortex (FC), temporal and parietal regions (TP), sensory-motor regions (SM), and subcortical regions (SC); and five stages of brain development: early fetal, mid-late fetal, infancy to childhood, adolescence and young adult. It resulted in 20 different time-space categories. Spearman correlation coefficient was calculated for each gene pair in each category. Several rigorous controls have been chosen for the analysis (such as randomly selected genomic regions with the same number of genes and PPIs, common CNVs from the database of Genomic Variants or extracted from the 1000 Genomes dataset). The statistical significance of co-expression enrichment of PPIs in specific category when comparing to a control is calculated using Fisher’s exact test.

Results: We first observed that interactions from 16p11.2 genes are distinctly clustered in three general temporal regions: fetal, childhood/adolescence, and young adult, based on their Spearrman correlations patterns. The separation of the interaction groups in 16p11.2 could indicate a different functional relevance of specific PPI subset in precise time and brain region. Furthermore, when comparing to the controls, in early fetal, the interacting pairs involving 16p11.2 protein products showed significantly depleted co-expression (coldspots), specifically at FC and TP regions. On the contrary, in adolescence, the interacting pairs showed significantly more co-expression (enrichments or hotspots), specifically at SM and TP regions. These patterns were not observed when the same analysis was not restricted to physical interacting pairs in 16p11.2. 

Conclusions:  

This analysis would provide valuable information about expression patterns of the shared and unique PPIs in different disorders detected by cross-disorder interactome mapping.