Association of Oxytocin Receptor Gene Variants with Social Phenotypes of Autism Spectrum Disorder

Background: Several lines of evidence indicate that oxytocin, signaling through its receptor (OXTR), is important in a wide range of social behaviors.  In attempts to determine whether genetic variations in the oxytocin signaling system contribute to autism susceptibility, four recent reports have indicated association of common genetic polymorphisms in the OXTR gene with autism spectrum disorder (ASD) (Wu et al, 2005, Biol Psych; Jacob et al, 2007, Neurosci Letters; Lerer et al, 2007, Mol Psych; Yrigollen et al, 2008, Biol Psych).  Each of these studies involved relatively small sample sizes (57 to 195 families) and, where it has been examined, failed to identify association of OXTR polymorphisms with ASD and measures of social behavior in individuals with ASD.
Objectives: To attempt to replicate association of common polymorphisms in the OXTR gene with ASD in a larger sample, and to examine association of genetic variants in OXTR with phenotype scores of the ADI-R, ADOS, and SRS.

Methods: The sample consisted of 664 families with at least one offspring with a diagnosed ASD.  Additional phenotype information was available for a subset of 334 families from the Autism Genetics Resource Exchange (AGRE) Consortium.  Genotypes were determined at 9 markers in the OXTR gene by ABI Taqman assays. Family Based Association Test (FBAT) analysis was performed on ASD diagnosis as well as phenotype scores of the ADI-R, ADOS, and SRS.
Results: The OXTR 3′ UTR marker rs1042778 G allele was associated with ASD diagnosis (P=0.018), replicating precisely a previous report of significant association (Lerer et al, 2007).  In contrast to previous reports, association of the rs1042778 G allele was also observed for the ADI-R social total domain (P=0.019) and for each of 4 social sub-domains (P<0.05) in the 334-family AGRE sample.  However, the rs1042778 G allele was not significantly associated with ADI-R total scores of communication (P=0.076) and behavior (P=0.067).  No other marker was associated with ASD diagnosis in the entire 664-family sample.  The intron 3 marker rs2268493 T allele, although not significant for association with ASD (P=0.073), was associated with the autism cut-off on the ADI-R (P=0.046).  Like the rs1042778 G allele, the rs2268493 T allele was associated with the ADI-R social total score (P=0.014) and 3 of 4 social sub-domain scores (P<0.05), but was not associated with ADI-R total scores for the communication and behavior domains.  No OXTR marker was associated with phenotype scores on the SRS, and no consistent pattern was observed for any phenotype scores on the ADOS modules.

Conclusions: These results replicate the association of OXTR variants with ASD in a large study sample and suggest association of OXTR genetic polymorphisms with social aspects of ASD.  These results should be interpreted with extreme caution because none of the nominally significant associations would survive an appropriate correction for multiple comparisons.  However, the precise replication of previous results in an independent cohort and the biological plausibility of participation of the oxytocin signaling system in ASD suggest that functional polymorphisms of OXTR may contribute to ASD risk in a subset of families.