Interaction Between Human Sexual Dimorphism and ASD Neurobiology.

Background: The 4:1 male to female sex bias is one of the most consistent and striking observations in autism spectrum disorder (ASD). One explanation for this sex bias is the existence of a female protective effect (FPE), in which a greater burden of ASD risk factors are required for a diagnosis of ASD in females than in males. Direct observation of de novo ASD risk factors in genomic analysis supports this hypothesis, however indirect assessment of sibling ASD recurrence risk in epidemiological studies finds little supporting evidence for the FPE. Understanding the nature and mechanism of this protection may hold great potential for therapeutic strategies.

Objectives: To assess the role and potential mechanism of female protection in ASD

Methods: To explain the discordant genomic and epidemiological evidence for the FPE we developed a simulation of ASD risk in families to estimate the power to detect a difference in the burden of de novo mutations vs. sibling recurrence risk. Should the FPE exist, it must act through sexually dimorphic processes, including gene expression. We therefore also compared gene expression data from male and female brain samples in the BrainSpan dataset ranging from mid-fetal to adult developmental stages to identify such sex differences in neurobiology.

Results: Under a quantitative model of ASD risk (Gaugler et al.2014), in which 50% of ASD risk in the population comes from unique environmental exposure, 47% comes from common inherited genetic variants and 3% comes from rare de novo mutations, we estimated the power to detect the FPE. Considering de novo mutations we achieve 80% power with a sample of about 500 ASD families, consistent with genomic literature. Furthermore, by combining exome and CNV data for over 5,500 ASD cases we consistently observe an increased burden of ASD risk factors in females, to a similar extent as predicted by the simulation. In contrast, the power to detect a significant difference in sibling recurrence rate in 10,000 ASD families remains below 30%. The FPE hypothesis is therefore also consistent with the epidemiologic literature. To explore the nature of the FPE, we assessed differential gene expression using RNA-Seq from over 1,200 region- and age-specific samples from males and females. We find similar developmental trajectories of gene expression in both sexes, with the exception of genes specific to microglia that are enriched in males during mid-late fetal development. These genes overlap with those observed in co-expression modules previously observed to be upregulated in ASD brains.

Conclusions: The FPE is the leading hypothesis of the underlying mechanism of ASD sex bias, with strong supporting evidence from genomic studies. The existing large-scale epidemiological analyses remain under-powered to corroborate this finding. Analysis of human transcriptome data identifies higher expression of microglial genes during male fetal development and finds that these genes overlap with those over-expressed in the post mortem ASD brain. This cellular and molecular pathway may sensitize males to ASD risk factors, and may account for the apparent female protective effect in ASD.