Exome Sequencing of Extended Families with Autism Reveals Genes Shared Across Neurodevelopmental and Neuropsychiatric Disorders

Background: Autism spectrum disorders (ASDs) encompass a constellation of devastating neurodevelopmental conditions and studies to date demonstrate that the underlying etiology is extremely heterogeneous. With the advent of whole exome sequencing (WES), studies have been rapidly identifying genetic causes of many Mendelian disorders. Thus far, autism exome studies have primarily focused either on simplex families to discover de novoalterations or consanguineous families that carry recessive mutations.

Objectives: Our study performs exome sequencing in extended, multiplex families with affected cousin pairs to identify potential new ASD loci. We hypothesized that identical by descent (IBD) filtering in large, multiplex pedigrees would permit us to isolate genes contributing to ASD susceptibility. These extended families are likely to carry rare and partially penetrant alterations that are inherited to affected individuals from common ancestors.

Methods: We performed WES on at least two affected cousins across 40 multiplex ASD families to identify rare, segregating mutations that are incompletely penetrant. A total of 164 individuals were captured with the Agilent SureSelect Human All Exon kit, sequenced on the Illumina HiSeq 2000, and the resulting data processed and annotated with BWA, GATK, and SeattleSeq. Variants were filtered to those in IBD regions delineated by SNP genotyping data. Initial analyses focused on novel and rare (MAF < 0.05) variants predicted to be detrimental, either by altering amino acids or splicing patterns. Validation was performed by Sanger sequencing and genotyping on the Infinium HumanExome BeadChip.

Results: Following exome sequencing, each extended family had changes at approximately 90,000 locations. Variants were filtered to those in identity by descent (IBD) regions delineated by SNP genotyping data. Initial analyses focused on novel and rare (MAF < 0.05) variants predicted to be detrimental, either by altering amino acids or splicing patterns. In accordance with a dominant model of inheritance, exome sequencing identified 742 heterozygous changes and 3 potentially X-linked alterations; 502 variants were validated either by Sanger sequencing or genotyping on the Infinium HumanExome BeadChip. We identified numerous potentially damaging, ASD associated risk variants in genes previously unrelated to autism. A subset of these genes has been implicated in other neurological disorders including depression (SLIT3), epilepsy (CLCN2), mental retardation (AP4M1, CEP290), schizophrenia (WDR60), and Tourette syndrome (OFCC1). This reinforces the theory that there are shared genetic components across distinct neurological disorders. We also found additional alterations in previously reported autism candidate genes (NRXN2, PRICKLE1, and SYN1) including three genes with alterations in multiple families (CSMD1, FAT1, and STXBP5), thereby expanding the evidence that these genes are involved in autism etiology. Compiling a list of ASD candidate genes from the literature, we determined that variants occurred in ASD candidate genes 1.65 times more frequently than in random genes captured by exome sequencing (p=8.55x10^-5).

Conclusions: By studying these unique pedigrees, we have identified novel DNA variations related to ASDs, demonstrated that exome sequencing in extended families is a powerful tool for ASD candidate gene discovery, and provided further evidence of an underlying genetic component to a wide range of neurodevelopmental and neuropsychiatric diseases.