Systems biology approaches to explain and predict developmental neurotoxicology

Abstract

The UKN1 test models early neural development. The major endpoint is the cellular transcriptome. This assay has been used to develop classifiers and toxicity indices for various compound classes. A major shortcoming of this type of tests is so-called "phenotypic anchoring", i.e. an assay endpoint that the transcriptome changes can be anchored to. To approach this issue, the capacity of cells in culture to self-organize to tissue-like aggregates was used. The neuroepithelial cells used in the UKN1 test system organize to neural tubes during in vivo mammalian development. The two-dimensional correlate of this, which is observed in cell cultures, is the organization to neural rosettes. We have established and characterized an assay that quantifies rosette formation, and its disturbance by developmental toxicants. This assay was then used as phenotypic anchoring to separate normal transcriptome changes from those indicating an adverse effect. As an example for a developmental toxicant, valproic acid was examined in closer detail. A mathematical model was developed to predict the response of genes for any given concentration or time of challenge. A sensitivity analysis of the model suggested that modulation of the wnt pathway during a specific time window was linked to developmental toxicity of the drug sodium valproate.