ICCVAM Biennial Report 2014-2015

ICCVAM Biennial Report 2014-2015

Immunotoxicity Testing: NIEHS Activities

Development of an Open-source Bayesian Network for Identification of Potential Skin Sensitizers

NICEATM and other scientists within the National Institute of Environmental Health Sciences (NIEHS) collaborated with scientists at Procter & Gamble (P&G) to develop an integrated testing strategy that can identify potential skin sensitizers and characterize skin sensitization potency without conducting animal tests.
The P&G-developed integrated testing strategy uses a Bayesian network to analyze all available relevant substance information, including non-animal tests, in silico models, and other information such as chemical structure and solubility, to produce a numerical probability of skin sensitization potency. Using the available information, the Bayesian network can identify the subsequent test that will best inform the skin sensitization potency prediction of a substance. The calculated probability could be used to make a hazard labeling decision without animal testing.
The software used by P&G for these analyses is patented, which could limit regulatory implementation of this approach because it is not fully transparent. Accordingly, NIEHS and P&G collaborated to develop an integrated testing strategy that uses open-source software and publicly available data. The code and data used to generate the predictions is available for other researchers to test, verify, and build upon.
The open-source integrated testing strategy was described in a short communication to the journal ALTEX. Subsequent revisions of the approach improved overall accuracy and enabled consideration of variability in reference data. A poster describing improvements to the strategy was presented at the 2015 Society of Toxicology Annual Meeting (Pirone et al.). Other improvements to simplify model inputs and refine potency estimates are described in a paper published in Archives of Toxicology.

Development of QSAR Models to Identify Potential Skin Sensitizers

NICEATM is collaborating with scientists at the University of North Carolina at Chapel Hill to develop quantitative structure-activity relationship (QSAR) models to support identification of potential human skin sensitizers without using animals. Two papers published in 2015 in Toxicology and Applied Pharmacology described QSAR models developed for skin penetration and to predict results of animal tests to identify sensitizers. More recent efforts have focused on developing QSAR models to predict human skin sensitization; a manuscript describing these models is in preparation.

Collaboration with Cosmetics Europe Skin Tolerance Task Force

NICEATM is collaborating with the Cosmetics Europe Skin Tolerance Task Force to evaluate skin sensitization integrated approaches to testing and assessment (IATAs) that have been submitted to the Organisation for Economic Co-operation and Development. NICEATM has evaluated seven IATAs against a set of previously untested chemicals with in vitro and in silico data provided by Cosmetics Europe. Manuscripts describing the datasets and the outcome of the IATA analyses are in preparation.

Participation on Management Teams for Validation Studies of In Vitro Test Methods

NICEATM scientists participated on management teams for studies to validate two in vitro methods identifying potential skin sensitizers: the IL-8 luciferase skin sensitization assay and the Vitrigel-SST assay. These studies were coordinated by the Japanese Center for the Validation of Alternative Methods. Laboratory work on the studies has been completed and the study results are undergoing peer review.