Novel Copy Number Variation in Autism

Background: Autism is a neurodevelopmental disorder that affects approximately 1 in 150 individuals and is characterized by deficits in reciprocal social interaction, communication and patterns of repetitive behaviors and restricted interests. Twin and family studies indicate high heritability, but evidence supports a highly complex architecture for the underlying genetic etiology. Recent discoveries point to Copy Number Variation (CNV) as an important class of rare variation that may cause or increase risk for autism. Objectives: The objectives of this study were to screen autism probands and parents for evidence of CNV to identify loci that cause or confer susceptibility for autism. Methods: A series of 49 autism probands ascertained at Vanderbilt were screened on the Affymetrix 6.0 SNP platform. Raw genotype data was analyzed using multiple algorithms, including Partek, dChip, and AffyConsole, to maximize sensitivity and specificity of CNV detection. Variants predicted by two or more algorithms were prioritized for experimental validation. For a subset of probands, parents were genotyped on the Affymetric 6.0 arrays to determine whether predicted CNV was inherited or de novo. Parental samples were included in validation experiments, which primarily involved qPCR on the Roche LightCycler 480 using SybrGreen and/or Universal Probe Library (UPL) assays. Results: Analyses of the Affymetrix 6.0 data resulted in 3,265 CNV predictions in all, and of these 670 (20.5%) were called by two and 351 (10.7%) by three algorithms. Among CNV predictions, 1170 (36%) identified regions containing annotated genes. The majority were gains (1994, 61%), with only 39% predicted to be losses (1271). Only 670 (20.5%) and 351 (10.7%) CNVs were detected by two or three algorithms, respectively. Mean and median sizes for this group were 171kb and 73kb, respectively. Prioritized variants were validated using qPCR, and of those detected by two or more algorithms, all were confirmed and all found to be inherited. Validated CNVs identified loci previously implicated in mental retardation and/or point generally to neuronal cell surface molecules with a range of specific functions including cell adhesion, migration, synaptic vesicle release and cellular signaling. Conclusions: We have conducted a genome-wide SNP-based screen for CNV in a small cohort of autism probands, and as expected, have identified a number of inherited variants very likely to cause or contribute to risk in these families. The fact that most CNV in autism are inherited makes their interpretation more difficult. Possible explanations include reduced penetrance in the parents or that a greater overall mutation burden in probands who contain multiple CNVs. Lending support for specific CNVs is previously documented involvement of a gene in dominantly inherited MR or CNVs previously identified in autism in both de novo and inherited states. DOCK8 is an example of one that fits both of these scenarios. We would predict that several of the variants identified will point to genes that contain rare mutations in non-CNV-containing subjects. In conclusion, the findings reported here further underscore the important role that CNV plays in the genetic etiology of autism.