Amyloid Precursor Protein-Binding Protein (APBA2) Is An Autism Candidate Gene

Background: Copy number variation (CNV) and resequencing analyses have implicated several autism candidate genes related to the structure and function of the synapse, including NRXN1, NLGN3 and NLGN4, SHANK3, and CNTNAP2. We have previously reported an autistic proband and affected brother with two maternally-inherited microduplications located in 15q13.1 and 15q13.3. Duplication and deletion of an overlapping interval within 15q13.1 have been reported in patients with schizophrenia and mental retardation, respectively. The amyloid precursor protein-binding protein A2 (APBA2) gene is located within the 15q13.1 region and encodes a neuronal adaptor protein essential to synaptic transmission that interacts directly with NRXN1 at the presynaptic membrane. We hypothesize that APBA2 may be an additional synaptic protein associated with autism.

Objectives: To screen APBA2 for rare nucleotide variation in autistic individuals and controls.

Methods: PCR-based Sanger sequencing of genomic DNA was used to screen APBA2 in 372 autism samples and 372 control individuals. Autism samples were acquired from the Autism Genetics Resource Exchange (AGRE) and controls from the NIMH Genetics Initiative control sample set. Family members were acquired to determine inheritance patterns and segregation of the autism phenotype with identified variants. Bioinformatics approaches were used to predict effects of variants on protein function.

Results: We identified eight novel autism-specific non-synonymous coding variants in APBA2, seven of which are predicted to affect protein function by at least one prediction program and/or alter residues that are conserved across all 28 species examined. Four variants occur in the MUNC18-interacting domain or PDZ domains of APBA2, which are essential to its neuronal function. We identified two non-synonymous variants specific to controls. These data suggest a trend of increased mutation burden in our autism cohort compared to controls (Fisher exact test p = 0.11). All autism-specific variants were inherited from unaffected parents. Of particular interest were two variants that were identified in the same proband: (1) a paternally-inherited heterozygous 6-bp deletion in exon three that causes deletion of two amino acid residues and substitution of a third, and (2) a maternally-inherited heterozygous missense mutation in exon 11. Both variants were also observed in an affected sibling. These results suggest compound heterozygous mutations of APBA2 in both siblings with autism.

Conclusions: This work represents the first sequence-level evaluation of APBA2 as an autism candidate. Although all eight non-synonymous variants are inherited, we hypothesize that rare variation of APBA2 may underlie risk for autism. The co-occurrence of two non-synonymous mutations in both affected siblings in a single family, each transmitted from a different unaffected parent, may indicate a causative role for APBA2 mutations in this isolated case. Functional studies to evaluate the effects of these mutations on the interaction of APBA2 with MUNC18-1 and NRXN1 are necessary. Taken together, the observation of rare CNVs involving APBA2 in autism, schizophrenia, and mental retardation, the observation of a trend towards increased mutation burden in autistic individuals, and the occurrence of compound heterozygous mutations in a single family provide strong support for APBA2 as a candidate gene for autism.