the following sections we review evidence for the role of prenatal androgens and sex chromosome effects.
Testosterone and Autism Spectrum Conditions
The first research to address the potential role of fetal testosterone (fT) exposure in ASC risk was performed in the context of the Cambridge Antenatal Predictors of Child Development Project. For more than 10 years, this project has been following a sample of typically developing children whose prenatal hormone exposure was measured in amniotic fluid. Within this sample, fT levels predicted behaviors that, in the extreme, are diagnostic symptoms for autism. fT levels were inversely correlated with eye contact at 12 months, vocabulary development at 18 and 24 months, and social functioning/empathy at 48 months and 6– 9 years. fT levels were positively correlated with restricted interests at 48 months and autistic traits at 18– 24 months and 6– 9 years [4]. Similarly, girls exposed to unusually high testosterone levels (as a result of congenital adrenal hyperplasia) showed more autism-like traits than their unaffected sisters [9].
These studies suggest that high levels of testosterone could produce a behavioral profile similar to that seen in autism but do not directly investigate whether individuals with autism show signs of high testosterone exposure. Evidence relevant to the latter includes the following: (1) the finding that low ratios of second to fourth digit length (2D:4D, a potential index of fT exposure) are associated with ASC [10]; (2) women with ASC show an increased rate of medical conditions and behavioral traits related to testosterone [11, 12]; (3) disrupted postnatal sex hormone profiles have been reported in individuals with ASC [13], although the exact nature of this change varies slightly from study to study; (4) polymorphisms in genes controlling the synthesis, metabolism, transport, and signaling of sex hormones have shown associations with ASC [14– 16], and (5) the timing of puberty is altered in females with ASC in a direction consistent with high testosterone [12].
Thus, early testosterone exposure appears to be a strong candidate for a contributory role in the pathogenesis of ASC. However, there is an important theoretical limitation to this theory as it currently stands. It is a top-down theory that takes, as one of its starting points, cognitive differences between the sexes. Advancement of this theory to the next level requires an understanding of how prenatal androgens alter neurodevelopmental trajectories to produce both typical sex differences and autistic symptoms. One important component of this work will be to understand how prenatal androgens interact with the genetic networks implicated in ASC. Testosterone can interact with other chemicals implicated in autism, including serotonin, GABA, oxytocin, and BDNF [4]; testosterone has recently been reported to regulate retinoic acid-related orphan receptor alpha (RORA), a proposed candidate gene for ASC [17].
Another important component will be to test whether individual variation in fT predicts variation in neuroimaging phenotypes relevant to ASC. At present, 3 published studies have addressed this issue. Kallai et al. [18] found that the 2D:4D ratio was related to volume differences in subregions of the hippocampus in 40 adult women. These investigators did not find a relationship between 2D:4D and amygdala volume. The second study, which was performed in the context of the Cambridge Antenatal Predictors of Child Development Project, found that fT levels were positively associated with increasing rightward asymmetry in the thickness of one subsection of the corpus callosum, the isthmus [19]. The isthmus connects the left and right posterior parietal and superior temporal cortices, which are integral for language and visuospatial ability and are known to be sexually dimorphic in lateralization, structure, and function. However, the relevance of these phenotypes (hippocampal volume or asymmetry of the isthmus) to ASC is unclear. The third study, also performed in the Cambridge project, assessed localized effects of fT exposure in juvenile males using voxel-based morphometry [20]. fT was positively associated with gray matter volume in the right temporoparietal junction/posterior superior temporal sulcus, an area that is relatively larger in males. fT was negatively associated with gray matter volume in the planum temporale/parietal operculum and posterior lateral orbitofrontal cortex, areas that are relatively larger in females. These areas are involved in language processing and social cognition and may be altered in children with autism.
Finally, careful work in animals can produce detailed pathophysiological theories that can then be tested in human populations. For example, Pfaff et al. [21] recently theorized that high fT levels increase arousal-related inputs to the amygdala, creating greater sensitivity to early life stresses and producing social avoidance as a compensatory strategy.
The X Chromosome and Autism Spectrum Conditions
One of the best-understood mechanisms producing biased sex ratios in a phenotype is the phenomenon of X-linked inheritance. However, the majority of linkage and association studies of ASC carried out to date have failed to find regions of interest on the X chromosome. Mutations associated with ASC have been reported at Xp22.33, Xq13, Xq28, and Xp22.1 affecting the NLGN4, NLGN3, MECP2, and PTCHD1 genes, respectively [22– 24]. In addition, 2 recent genome-wide studies of copy number variation (CNV) in individuals with ASC identified mutations affecting the X chromosome, but this represented a very small subset of cases. Other large CNV scans as well as a recent exome sequencing study have reported no findings on the X chromosome and 3 independent mutation screening studies of NLGN4 and NLGN3 have concluded that mutations in these genes are very rarely associated with ASC [see 4 for review]. In summary, at present it appears that there are X-linked causes of ASC, but these represent a relatively small percentage of cases.
Even if X-linked inheritance does not play a major role in the biased sex ratio seen in ASC, the sex chromosomes exhibit numerous interesting complexities that could play a role in sex differences in ASC. For example, although females have 2 X chromosomes, only one of these is generally active. X chromosome inactivation (the process by which one X is suppressed while the other remains active) acts to negate the ‘dosage’ difference in X chromosome genes between males and females. Inactivation does not involve the same X chromosome in all cells, meaning that females are all epigenetic mosaics. Mosaicism could potentially act as a protective factor if X chromosomes carrying deleterious mutations were preferentially inactivated. Gong et al. [25] tested this hypothesis and found no evidence for skewed X chromosome inactivation in a large sample of individuals with and without ASC.
Up to 20% of the genes on the human X chromosome may continue to be expressed from the ‘inactive’ X, although a recent gene expression study found that only 5% of X-linked genes actually achieve significantly higher expression levels in females compared to males [26]. These genes could play a role in sex ratios if the nonsilenced genes were protective. Girls with Turner syndrome (TS), which is characterized by the XO karyotype and who are haploinsufficient for these gene products, are at an increased risk for