Session 5: Adverse outcome pathways

Defining adverse outcome pathways from omics-driven bioinformatics and modeling approaches
PRESENTING AUTHOR: 

Prof Roland Grafström

INSTITUTION / COMPANY : 

Institute of Environmental Medicine,
Karolinska Institutet, Sweden

POSITION: 

Professor

AUTHOR(S): 

Roland Grafström1, Pekka Kohonen1, Penny Nymark1, Vesa Hongisto2, Ola Spjuth3 and Barry Hardy4

1Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden
2Toxicology Department, Misvik Biology Corporation, Turku, Finland
3Department of Pharmaceutical Biosciences and Science for Life Laboratory, Uppsala University, Uppsala, Sweden
4Douglas Connect, Switzerland

ABSTRACT CONTENT / DETAILS: 

Knowledge integration of data from in vitro and in silico methods are increasingly being used in safety testing relative to traditional costly animal testing, including for evaluation of chemicals, nanomaterials and diverse consumer products.

Stimulating this development, the European Chemicals Agency now advocates for a flexible and extensible ”conceptual framework for evaluating chemicals safety”, built on combining results generated from new predictive tools with existing data.

The Grafström laboratory has for many years analyzed numerous formaldehyde effects in vitro as a case study for adverse outcome pathways, involving a set of complex interactions that associate with toxicity, DNA damage, mutagenicity, cell transformation and carcinogenicity.

We demonstrate under the theme of “knowledge integration supporting decision making” that:

1) toxicogenomics results serve excellently to explore modes of action and adverse outcome pathways,
2) bioinformatics-driven analysis versus large data repositories is key to generating the mechanism-based predictions, and that
3) omics-driven toxicity modeling is a complementary and useful exercise relative other methods.

We conclude on this basis that the omics-driven bioinformatics and modeling approaches overall is able to address toxicity pathways, threshold of toxicological concern ranking, coupling of in vivo pathology to in vitro data, dose response estimation of NOEL/LOEL results, as well as the grouping and read across between different agents.

Consistent and systematic warehousing of the toxicity data types and analysis now presented promises to handle future ab initio “in vitro-based only” toxicity predictions.

To this end, we foresee the application of possibly different variants of the “adverse outcome pathways framework”.