High throughput modeling of the effects of mixtures of ToxCast chemicals on steroid hormone cycles in women
Résumé
Exposure to mixtures of chemicals is an increasing toxicological concern. The availability of exposure data for thousands of chemicals through ExpoCast project, together with the ToxCast results for the hundreds of high throughput in vitro assays, offers the opportunity to explore the toxicity of the chemical mixtures in realistic scenarios. We used computer modeling to predict the size of potential effects of random mixtures of aromatase inhibitors on women's menstrual cycle. We had previously investigated the impact of mixtures on steroidogenesis by a systems biology model for aromatase inhibition in adult female rats. In current work, to consider a larger number of events involved to hormonal balance disruption, we adapted a mathematical model of the hypothalamus-pituitary-ovarian control of estradiol and progesterone concentrations in blood. We used the model (coupled to a pharmacokinetic model of intake and disposition) to predict the effects of a million of chemical mixtures sampled by Monte Carlo simulations. To simulate a realistic exposure scenario, the exposures were also sampled from statistical distributions provided by the ExpoCast database (see illustrated work-flow). We find that a sizable fraction of the mixtures generated led to more than 20% inhibition of estradiol production. In contrast, exposures to chemicals considered individually almost never reach such effect sizes. Those results are discussed in light of the approximations and assumptions made, but demonstrate the possibility to address large scale mixture questions in a predictive toxicology framework, suitable for high throughput risk assessment of endocrine perturbation.