We have previously shown that a functional promoter polymorphism of the MAOA gene is associated with brain structure volumes in children with idiopathic autism . Given the phenotypic overlap with autism, we now tested for a similar association in children with the fragile X syndrome (FXS). We found, as in idiopathic autism, that the low-expressing allele was associated with increased cerebral cortical gray and white matter volumes in the FXS children, both those with and without autism (Figure 1). These associations were not observed in older typically developing children assessed in our previous study (Figure 2). We thus demonstrate that this modifying genetic effect on an endophenotypic trait is found in two neurodevelopmental disorders, and we demonstrate the utility of FXS and autism as a system for testing relationships amongst specific genes within the heterogeneity of ASDs.
The plausibility of the MAOA polymorphism producing these brain effects is supported by multiple lines of research. Serotonin influences numerous aspects of brain development, as well as ongoing synaptic activity throughout life [34, 35]. Perturbations of the serotonin system have been implicated in a variety of neuropsychiatric disorders, with particularly compelling effects in autism [36–40]. MAOA is the primary enzyme responsible for degrading synaptic monoamine neurotransmitters, including serotonin, and levels of these neurotransmitters in the brain vary in association with levels of the MAOA protein [41–44].
Alterations in expression of MAOA are, in turn, associated with neuropsychiatric disorders. Two null mutations cause severe phenotypes. Norrie’s disease is characterized by mental retardation, autistic behavior, and motor hyperactivity, and is caused by an X-chromosomal deletion that includes MAOA [45, 46], while Brunner’s syndrome includes violent and criminal behavior in the phenotype and is due to an MAOA stop mutation [47, 48]. The MAOA promoter polymorphism, by contrast, is associated with more moderate expression effects. The polymorphism is a 30 base pair VNTR that is 1.2 kb upstream of exon 1 . When present as 3.5 or 4 copies, the repeat is associated with increased MAOA expression, while 2, 3 or 5 copies are associated with decreased expression [33, 50]. The repeat has also been associated with numerous less severe but more common neuropsychiatric phenotypes [51–54]. Of relevance to autism, the low activity allele has been associated with increased severity of a range of social and behavioral difficulties, including sensory behaviors, arousal regulation, aggression, social communication skills [54, 55], a lower IQ , and, through our previous work, cerebral cortical enlargement . This association with brain structure is noteworthy because increased head circumference and enlargement of the cerebral cortex are highly replicable biological correlates of autism [16, 56].
The current findings are consistent with a more general relationship between serotonin and brain structure suggested by our research. In addition to MAOA, we have also tested a functional promoter polymorphism of the serotonin transporter gene (SERT) . Just as with MAOA, we found that an allele associated with decreased expression of SERT is associated with increased cerebral cortical gray and white matter volumes in idiopathic autism. For both genes, decreased expression is associated with increased synaptic and central nervous system (CNS) serotonin levels [41, 42, 58]. Thus a picture begins to emerge of serotonin system genetic variation influencing brain structure where increased serotonin expression is associated with increased cerebral cortical volumes in children with neurodevelopmental disorders, particularly autism and FXS.
The findings also show how studying single gene disorders can help to clarify the specificity of genetic effects found in more complex, heterogeneous disorders. Autism has a highly variable phenotype, suggesting an array of interacting genetic factors that influence the expression of disease. FXS is a monogenic disorder in which some children have phenotypic features of autism while others do not. Through our examination, we find that the MAOA polymorphism produces the same effects on brain structure in three groups of affected children: those with idiopathic autism, those with FXS and autism, and those with FXS but no autism. We do not find any effect of the polymorphism on brain structure in typically developing children, though we do note that this is not an ideal comparison group because they are older than the patients and their images were acquired through a different scan protocol.
Furthermore, though an understanding of why the MAOA polymorphism would produce brain effects in some children but not others is not known, studies of mice support this type of interaction. Mice of different genetic backgrounds display markedly different patterns of social behavior, with some strains exhibiting social deficits resembling those seen in autism [59, 60]. The same genetic defect placed on these different backgrounds can produce strikingly different phenotypes. Moy et al., for example, reported that targeted disruption of Fmr1 produced sociability deficits in FVB/129 mice but not in C57BL/6 J mice . More closely related to our study, Page et al. used mice to examine interactive effects of two autism-related genes on brain structure . Mice haploinsufficient for either the SERT or PTEN gene had enlarged brains as well as deficits in social interaction. After crossing, mice haploinsufficient for both genes had more substantial brain enlargement and more severe sociability deficits.