Use of human pluripotent stem cells and their progeny to develop in vitro models for neurotoxicity testing

Abstract

At present, only a minority of commercially available compounds and chemicals have been tested for neurotoxicity (NT) and developmental NT (DNT), although some estimations suggest that almost one third of all chemicals may cause adverse neurological effects. Thus, there is a need for a better and faster testing of all 'daily-use' chemicals. Human pluripotent stem cells (hPSCs) have been shown to be a suitable tool for NT/DNT testing, as they can be differentiated into diverse neuronal cell types in the culture plate and thereby recapitulate crucial processes of nervous system development and function. Within the scope of this doctoral thesis we have established three in vitro systems based on hPSCs, which provide insight into different concepts of in vitro (developmental) neurotoxicity testing.As first step, we have established an hPSC-based 3-D neurosphere system. Long-term exposure to non-cytotoxic concentrations of the DNT gold standard methylmercury or well-defined polyethylene nanoparticles induced changes in the expression profile of a select set of neuronal marker genes. Our data suggest that the system has the potential to detect long-term DNT effects of nanoparticles on neural differentiation.In the second step, neurally differentiating hPSCs have been exposed to six histone deacetylase inhibitors (HDACi) and six mercurials. We used bioinformatics tools to establish a transcriptome-based classifier to discriminate between these two different groups. The validations by a 'leave-one out' approach and with legacy data sets showed a correct prediction of HDACi under conditions relevant for DNT. These findings indicate that this approach is a suitable tool for DNT assessment to classify groups of compounds according to the toxicity-inducing changes on the transcriptome.As a final approach, we have established a differentiation protocol to produce hPSC-derived dorsal root ganglia-like cells, which meet the needs for toxicity testing. A broad range of more than 30 compounds was tested by quantifying neurite growth and viability as functional endpoints. A comparison with a similar test of central neurons showed that specific peripheral neurotoxicants were correctly detected in the new test system, but not in another test used for comparison. The results of this comparison suggest the importance of using the correct target cell type for neurotoxicity testing. The findings contribute to new concepts of in vitro (developmental) neurotoxicity test system development and highlighted critical points that have to be considered in test system development.