Enhancing the applicability and predictability of the embryonic stem cell test for developmental toxicity

Abstract

Within the full risk assessment of a chemical, developmental toxicity testing is one of the endpoints that require the highest percentage of experimental animals. With the high number of experimental animals, cost and time involved in in vivo developmental toxicity testing there is an urgent need for in vitro methods for these toxicological endpoints. Unfortunately, until now the Embryonic Stem cell Test (EST), one of the few well established alternative system for developmental toxicity testing that does not require the use of any animals, has not been accepted for implementation into regulatory settings. One of the reasons why this is the case for the ESTc is that the predictivity and applicability domain of the assay, in terms of the chemical classes and biological mechanisms for which it can make a accurate prediction has not been satisfactory defined. Furthermore, the assay currently only uses cardiomyocyte differentiation to determine the potential embryotoxic effect of compounds which could lead to false-negative results for compounds only affecting other differentiation lineages, hereby reducing the predictive value of the assay. The first aim of the thesis was to assess the use of category approaches in evaluating the applicability domain of the EST by performing in vitro – in vivo comparisons within chemical classes. Furthermore, we assessed how incorporating data on kinetic properties could enhance these in vitro – in vivo extrapolations. The second aim of this thesis was to investigate if incorporating an osteoblast differentiation route into an EST test panel could improve the overall predictive value of the method. The results of the thesis show that the category approach represents a useful method to evaluate the applicability domain of alternative assays. For the ESTc to be implemented within a risk assessment setting the in vitro effect concentrations need to be extrapolated to in vivo doses. We showed that the category approach leads to far more favorable in vivo effective dose estimates based on the in vitro – in vivo correlations compared to studies were a set of diverse chemicals are used. This supports the use of the category approach as it allows a more reliable extrapolation of in vitro concentration to in vivo doses that could be used in risk assessment. Furthermore, we showed that the in vitro - in vivo extrapolations could be enhanced by the use of a PBK model for chemical classes that show diverse kinetic properties. With regard to the second aim of the thesis, we found that the results on the osteoblast differentiation assay developed and described in this thesis show that although there is a large overlap for most compounds in their effect on different differentiation routes, there are some compounds, for example TCDD, that can only be identified as embryotoxicants by investigating very specific differentiation endpoints. This shows that an EST test panel incorporating multiple differentiation endpoints, including osteoblast differentiation, can improve its predictive value in determining the embryotoxicity of compounds compared to an assay that only assess the effect of compounds on a single differentiation pathway.