Session 6: Alternative methods to support read across

Use of alternative methods to support read across - experiences form the Detective project
OpenTox Euro 2014 speaker: Sylvia Escher

Sylvia Escher


Fraunhofer Institute of Toxicology and Experimental Medicine, Germany


Sylvia Escher, Jan Hengstler

Wed, 24. Sep. 2014

Read across is a well know approach in regulatory risk assessment, where toxicity data from one/many data rich source compound(s) is/are used to predict the toxicity of one/many data poor target compound(s). Quantitative predictions such as no observed adverse effect level (NOAEL), as well a qualitative predictions e.g. mutagenic potential are possible.

For read across the chemical as well as biological similarity between the source and target compounds has to be clearly demonstrated. Chemical similarity is often based on structural similarity and the evaluation of relevant physico-chemical properties (being similar or following a consistent trend). Biological similarity comprises same mode of action in vivo, same toxikokinetics e.g. absorption, distribution, metabolism and excretion. Very often these data on mode of action are not available.

This presentation will introduce the Detective read across case study. In the Detective project, we aim to evaluate, whether and how mechanistic data e.g. from in vivo /in vitro omic experiments can be used to support read across. Detective is one out of the six project of the SEURAT cluster, which evaluates new strategies to replace in vivo repeated dose toxicity testing.

Valproic acid (VPA), one of the gold compounds in the Detective project, was selected for this pilot read across case study. It is known to be of concern to induce microvesicular steatosis in the liver. The strategy of the read across use case will be presented, the selection of source and target compounds as well as promising results from the first in vitro experiments.