An association analysis of microsatellite markers across the Prader-Willi/Angelman critical region on chromosome 15 (q11-13) and autism spectrum disorder.

Autism (OMIM 209850) is a neurodevelopmental disorder with a significant genetic component of a complex nature. Cytogenetic abnormalities in the Prader-Willi/Angelman syndrome critical region (PWACR) on chromosome 15 (q11-13) have been described in several individuals with autism. We have examined five microsatellite markers spread across the 4 Mb PWACR for linkage disequilibrium (LD) in 148 families with autism spectrum disorder (ASD) and a subset of 82 families with autism using the extended transmission disequilibrium test (ETDT). The markers examined were D15S11, D15S128, D15S1506, GABRB3, and D15S1002. In addition we have examined the microsatellite D15S822 for hemizygous deletion status in our sample as it had been previously reported to be increased in autism. We found no significant LD with any of the markers tested either in the ASD or autism families when looking at paternal and maternal meioses combined. However, as there are known imprinted genes in the region, including possibly GABRB3, we also examined for LD in paternal and maternal meioses separately. Examining paternal transmissions only, we found marginal evidence for LD with a protective allele at marker D15S11 in the ASD families (Chi-sq 7 df, P = 0.05) and marginal evidence for risk alleles at markers D15S1506 (Chi-sq 13.7, 6 df, P = 0.06), GABRB3 (Chi-sq 15.9, 8 df, P = 0.11) and D15S1002 (Chi-sq 17.7, 9 df, P = 0.08) in the autism only families. The allele responsible for the association with GABRB3 is the 191 allele which was previously reported to be overtransmitted. Hemizygous deletion of the microsatellite D15S822 was found in 3 out of 340 independent chromosomes in our sample; a rate of 0.8%. This is not significantly different to the frequency in the general population. In conclusion, our results did not rule out the involvement of this chromosomal region, but provided further evidence, albeit very limited, to implicate GABRB3. Further more systematic work in larger samples is required and confirmation that GABRB3 is imprinted is desirable. (c) 2005 Wiley-Liss, Inc.