14 August 2002

by Michele Ramsay micheler@mail.saimr.wits.ac.za
 

Autism is a pervasive developmental disorder characterized by impaired social interaction, communication and repetitive behaviour that is diagnosed in childhood and persists throughout life. Familial clustering and significantly higher concordance rates between monozygotic twins when compared with dizygotic twins suggest a large genetic component to susceptibility to autism. There is a continuous severity spectrum and autism most likely results from the complex interaction of many genetic factors (perhaps 10 to 15) and the environment in a cumulative but not necessarily additive effect. Many candidate gene studies have failed to reveal causative mutations and genome screens have pointed to several chromosomal regions that are likely to be involved in autism.

Yu et al. studied variation at polymorphic markers on chromosomes 7, 8, 15 and 16 in 105 families with two or more affected siblings with autism who were diagnosed according to strict diagnostic criteria. They were surprised to identify many instances of incompatible inheritance and showed that these were often due to inheritance of a null allele from an apparently homozygous parent. On further investigation they identified four chromosomal regions with deletions ranging from 5kb to over 260kb in 12 of the families. One deletion was characterized by a 192kb deletion on chromosome 7, including the absence of marker D8S272. It was found in five of the families with autism and was also shown to be polymorphic in each of the control groups, ranging in frequency from 0.026 to 0.082 in the random controls. The other seven families showed deletions in three chromosomal regions, two on chromosome 7 (characterized by markers D7S630 and D7S517 respectively) and one on chromosome 8 (D8S264). Each family had a unique deletion and none was correlated with the presence of repeat elements in the DNA sequence, as is the case with many other deletion syndromes. These specific deletions were not observed in control families and unrelated individuals. Rather than postulate candidate genes in each of these regions, the authors suggest that a more likely hypothesis is that allelic variants at other loci predispose an individual to autism through a mechanism that induces errors during meiosis. This is interesting in view of the increasing number of reports of translocations and rearrangements associated with autism and other behavioural disorders.

As the deletion mutations were present in some unaffected individuals, the proposed susceptibility allele (or alleles) might not segregate with the deletion mutation and might be neither sufficient nor necessary for the development of autism. Only seven of the 105 families studied showed detectable unique deletions, but only four chromosomes were investigated. Subsequent studies will no doubt screen the whole genome with a denser distribution of polymorphic markers to determine just how common deletions and chromosomal rearrangements are in individuals with autism and possibly other behavioural disorders.

At present the experimental design of the majority of association studies would not detect large deletions and one cannot help but wonder just how common they are? If they are so common, what is their role in health and disease in human populations?