Toward the Replacement of Animal Experiments through the Bioinformatics-driven Analysis of ‘Omics’ Data from Human Cell Cultures

Abstract

This paper outlines the work for which Roland Grafström and Pekka Kohonen were awarded the 2014 Lush Science Prize. The research activities of the Grafström laboratory have, for many years, covered cancer biology studies, as well as the development and application of toxicity-predictive in vitro models to determine chemical safety. Through the integration of in silico analyses of diverse types of genomics data (transcriptomic and proteomic), their efforts have proved to fit well into the recently-developed Adverse Outcome Pathway paradigm. Genomics analysis within state-of-the-art cancer biology research and Toxicology in the 21st Century concepts share many technological tools. A key category within the Three Rs paradigm is the Replacement of animals in toxicity testing with alternative methods, such as bioinformatics-driven analyses of data obtained from human cell cultures exposed to diverse toxicants. This work was recently expanded within the pan-European SEURAT-1 project (Safety Evaluation Ultimately Replacing Animal Testing), to replace repeat-dose toxicity testing with data-rich analyses of sophisticated cell culture models. The aims and objectives of the SEURAT project have been to guide the application, analysis, interpretation and storage of ‘omics’ technology-derived data within the service-oriented sub-project, ToxBank. Particularly addressing the Lush Science Prize focus on the relevance of toxicity pathways, a ‘data warehouse’ that is under continuous expansion, coupled with the development of novel data storage and management methods for toxicology, serve to address data integration across multiple ‘omics’ technologies. The prize winners’ guiding principles and concepts for modern knowledge management of toxicological data are summarised. The translation of basic discovery results ranged from chemical-testing and material-testing data, to information relevant to human health and environmental safety.

A transcriptomics data-driven gene space accurately predicts liver cytopathology and drug-induced liver injury

“…4d). Regulatory agencies such as the European Chemicals Agency and the United States Environmental Protection Agency are increasingly advocating for the inclusion of trancriptomics data and new approach methodologies in chemicals risk evaluation12111824. Thus, the demonstration of this mostly expected outcome fills the important role of implying broad applicability of the PTGS concept also outside of drug discovery studies.…”

“…Systems biology-based assays relying upon quantitative mechanistic information are increasingly envisaged as cornerstones of future safety evaluation of drugs and chemicals17891011. Accordingly, various modelling approaches have analysed ‘omics data sets to generate biomarker signatures or to characterize mechanisms of toxicity at a system-wide level, but suffer from the high dimensionality of omics data relative to sample number as well as problems in scaling across experimental systems (for example, hepatocyte cultures to liver) or species (for example, rat to human)111213141516.…”

“…Accordingly, various modelling approaches have analysed ‘omics data sets to generate biomarker signatures or to characterize mechanisms of toxicity at a system-wide level, but suffer from the high dimensionality of omics data relative to sample number as well as problems in scaling across experimental systems (for example, hepatocyte cultures to liver) or species (for example, rat to human)111213141516. Complicating predictive biomarker discovery, different genes and gene families will unlikely have similar dose response, so dose-dependent transitions will influence the classification of (toxic) modes-of-action1718.…”

“…The US Broad Institute Connectivity Map (CMap) data set has thousands of gene expression profiles of mostly FDA approved drugs and has been used to connect small molecules, genes and diseases (‘connectivity mapping’) to define biologically similar compounds, including for the purpose of identifying toxic modes of action11152526272829. The US National Cancer Institute (NCI) 60 tumour cell line screen includes results on GI 50 (50% growth inhibition), total growth inhibition (TGI), and LC 50 (50% lethal concentration) for many compounds tested in the major CMap cell lines30.…”

Transcriptomics in Toxicogenomics, Part I: Experimental Design, Technologies, Publicly Available Data, and Regulatory Aspects

“…This has led to the establishment of several prominent programs in the EU and worldwide aiming to generate omics-based big data. These could be used in systems biology-based approaches to build predictive models, simultaneously shifting the focus from in vivo animal testing towards in vitro methodologies and in silico modeling [8]. Studying gene expression changes in response to a foreign substance or a compound allows the building of a detailed picture of the organisms molecular levels of alteration and consequently understanding the MOA of the possible toxicant [7].…”