Systems toxicology and predictive modeling of male developmental toxicity

The ToxCast paradigm for predictive toxicology utilizes in vitro data from automated high-throughput screening (HTS), coupled with in silico models for data mining and systems modeling. This chapter highlights progress with an HTS-driven approach for developmental toxicity, using the male reproductive tract development as an example. A systems toxicology model built from ToxCast data revealed an aggregate phenotypic hierarchy reflecting human testicular dysgenesis syndrome. Bioactivity profiling of 54 chemicals across assays identified 156 candidate molecular targets in a network of biological processes for steroidogenic and xenobiotic metabolism, developmental regulation, angiogenesis and inflammatory pathways, and neural activity. Cell agent-based modeling was applied to recapitulate key events during genital tubercle (GT) development and dysmorphogenesis (hypospadias). Capturing sonic hedgehog, fibroblast growth factor 10, and androgen signaling pathways, as well as cell-autonomous programming, cellular biomechanics, and microphysiological influences underlying urethral closure in a simulated “virtual GT” provided a means to translate the systems toxicology model into a spatially scaled dynamical systems and a probabilistic prediction of toxicity.