Model and in vitro based prediction of human hepatotoxicity

Abstract

Toxicological studies can be designed in different ways, of which the extremes can be considered as the frog's and the eagle's perspective (Rahnenführer and Leist, 2015). The frog's perspective design focusses on a specific hypothesis with high local resolution; a danger with this approach is that conclusions are typically biased towards prior knowledge. Since the advent of omics technologies the opposite extreme has become popular, the eagle's type study perspective; it can help to generate new and unbiased hypotheses. However, the link to adverse effects is often unclear and because of the large observational distance the picture may become so blurry that it is of no use. One way to bridge this gap is a combination of spatio-temporal mathematical modeling and gene regulatory network analysis. In the present study we demonstrate how key mechanisms of hepatotoxicity can be identified by this technique. Based on concentration and time-resolved genome-wide analysis of 148 compounds in 3D-systems of primary human hepatocytes, possibilities and limitations of predicting human hepatotoxicity by modelling and gene regulatory network analysis will be illustrated.