31409
Maternal Serum Biomarkers and Autistic Traits in Expectant Mothers and Their Children
The prenatal endocrine environment may be a significant factor in the development of autism, as indicated by recent findings. Several steroid hormones, including androgens, are elevated in the fetal circulation of males with autism (Baron-Cohen et al., 2015). Furthermore, three separate epidemiological studies have associated autism likelihood in children with maternal polycystic ovaries syndrome (PCOS), a condition associated with high androgens (Kosidou et al., 2016; Berni et al., 2018; Cherskov et al., 2018). In contrast, low levels of estriol in maternal serum (2nd trimester) have been shown to increase the likelihood of autism in the offspring (Windham et al., 2016). Therefore, the overall ratio of androgens to estrogens may be a more suitable prenatal biomarker for the development of autism and related traits.
Objectives:
To investigate: 1. prenatal steroidogenesis in association with autistic traits in expectant mothers and their child; and 2. the ratio of testosterone to estradiol, as a proxy of prenatal aromatisation.
Methods:
Maternal serum samples were collected from n=126 neurotypical, pregnant women, taking part in the Cambridge Ultrasound Siblings and Parents (CUSP) study, corresponding to the first trimester of their pregnancy (weeks 10 to 14). Concentrations of the following steroids and peptides were measured: Testosterone (T), Estradiol (E2), Dehydroepiandrosterone sulphate (DHEAS), Progesterone (P), Insulin-like Growth Factor 1 (IGF1), sex hormone-binding globulin (SHBG). Samples were analysed on a DiaSorin Liaison XL automated immunoassay analyser using a one-step competitive chemiluminescence immunoassay for each hormone and two monoclonal antibodies for each peptide. The bioactive forms of testosterone and estradiol were calculated by standardising the levels according to SHBG levels. The aromatisation ratio was measured by dividing estradiol by testosterone concentrations. Expectant mothers were asked to complete the Autism Spectrum Quotient (AQ) to measure autistic traits and the Quantitative Checklist for Autism in Toddlers (Q-CHAT) online when their infant was between 18- and 20-months old. Hormone values were analysed in conjunction with autistic traits via Pearson’s Correlation Coefficient. Correction of significance thresholds for multiple comparisons was performed by application of the False Discovery Rate (FDR). Latent factors driving the variance in steroid levels were calculated via the Bartlett method.
Results:
None of the analysed hormones, nor the aromatisation ratio, were significantly associated with autistic traits in the mother, in univariate analyses and following correction for multiple comparisons. However, pairwise correlations of steroid hormones revealed significant interdependence between them (e.g. estradiol to DHEAS: Pearson’s β=0.629, p<0.0001 & estradiol to testosterone: Pearson’s β=0.466, p<0.0001). Latent factor analysis, identified three separate factors driving most of the variance in the observed steroid levels. The predicted values of the latent factor that regulated DHEAS, bioactive estradiol and bioactive testosterone were significantly correlated to maternal AQ Scores (Pearson’s β=0.2515, FDR-adjusted q=0.026).
At the time of submission, Q-CHAT scores of the infants were still being collected, with their association to biomarkers to be completed and presented in May 2019.
Conclusions:
A latent steroidogenic factor affecting both androgens (DHEAS, testosterone) and estrogens (estradiol) is positively associated with autistic traits in expectant mothers.