A Cellular and Molecular Study Investigating the Fetal Androgen Theory of Autism Using a Human Stem Cell Model

Background:  Autism spectrum conditions (autism) are associated with elevated fetal steroidogenic levels, suggesting a role for fetal steroids such as testosterone in the development of the autism during fetal development. Fetal testosterone is an androgen that ‘masculinizes’ the brain and reproductive system, reduces apoptosis, and exerts epigenetic influence on gene expression. Fetal testosterone is produced in two-fold quantities in male compared to female fetuses, and may be an important contributor to the early development of autism, given the established male biased sex ratio on the prevalence of autism.

Objectives:  To determine the effect of fetal testosterone on cellular and molecular pathways in neurons derived from an individual with autism, and how this differs from its effect in neurons derived from a typically developing individual.

Methods:  We used a human induced pluripotent stem cell (hiPSC) model of autism. Keratinocytes from typical individuals, and individuals with autism were reprogrammed into iPSCs, and then differentiated into cortical neurons using the dual SMAD signalling inhibition strategy. Testosterone was administered to these neurons at an early stage of development at the physiological and supra-physiological levels. Quantitative real time PCR was performed to measure gene expression levels. Immunocytochemistry followed by high throughput screening measured cell counts based on neuronal markers expression.

Results:  Preliminary data indicates that neurons from individuals with autism develop differentially, while being more responsive to testosterone than neurons from typical individuals. The androgen receptor (AR) and some of its putative downstream genes such as gonadotropin-releasing hormone (GnRH) and p38 (a class of mitogen-activated protein kinases) were differentially expressed in autism neurons compared to typically developing neurons, predicting differential structural and functional outcomes of neurons in the two groups (Fig1a). Genes indicating neuron cell fates such as T-box brain 1 (TBR1) and BRN2 were also differentially expressed in autism neurons, demonstrating altered characteristics (Fig1b). Long term testosterone administration does not seem to regulate TBR1 and BRN2 expression, and cells positive for TBR1 seemed to increase with time especially in autism neurons, while BRN2 positive cells increased in both (Fig2).

Conclusions:  These data suggest that the autism phenotype develops in utero at very early stages of brain development, and that neurons of individuals with autism are programmed, at a molecular level, to react differently to sex steroids such as fetal testosterone. The iPSC model has the power to distinguish genetic predisposition to fetal androgen sensitivity, environmental exposure to fetal androgen level, and interactions mediated via epigenetic influences of fetal androgens on the autistic genome. The current experiment casts light on the first two factors, and future work will test the latter hypothesis using RNA-sequencing.