Hazard assessment is a vital part of the lead compound development in a variety of industries, but the costs of performing a full hazard workup can be significant, particularly if little is known about the compound or if it has a complex metabolism. The field of predictive toxicology has been growing and developing rapidly in the last few decades, particularly in the development of ever-more-sophisticated in vitro assays that more effectively recapitulate the properties of human organ systems.
During the Workshop Session “Shifting Currents in Predictive Toxicology and Safety Evaluation with In Vitro and Alternative Approaches” at the SOT 58th Annual Meeting and ToxExpo, Dr. Kim Boekelheide and Dr. Edward LeCluyse provided a thoughtful and comprehensive survey of the state-of-the-science for in vitro kidney and liver models, respectively. Their presentations effectively summarized the most recent advancements in these organ system models and turned an eye toward the future of the science, highlighting the remaining challenges in the endeavor to simulate human tissues. Dr. Boekelheide emphasized the importance of recapitulating the structural milieu of native tissues through coculturing and agarose-based hydrogels and showed early characterizations of the 3D liver model his group is developing. Dr. LeCluyse delivered a more general exploration of the various 3D tissues available and their strengths and weaknesses. Dr. William Gwinn and Dr. Susanne Ramm, however, presented work on methods that eschewed convoluted tissue culture techniques altogether.
Dr. Gwinn led the Workshop with a presentation of the National Toxicology Program’s recent investigations into the use of transcriptomics for developing benchmark doses for chemicals. The aim was to see if the RNA transcripts produced by certain mouse organs in response to acute chemical exposure were sufficiently predictive to determine the same benchmark dose that would be otherwise determined with a two-year bioassay. In short: Do changes in cell behavior after five days of exposure predict long-term health effects? Overall, the benchmark doses for rat liver and kidney toxicity determined by a five-day benchmark dose study were within five- to tenfold of those determined with a two-year bioassay. Dr. Gwinn noted that there were still plenty of limits on the ways that this new technique could be used but that it could be helpful for providing estimates of apical benchmark doses for poorly characterized chemicals.
Dr. Ramm’s approach used genomic analysis as well, but combined it with high content imaging techniques and machine learning software to build network “maps” of mechanisms of toxicity for different compounds that will help researchers predict and classify potential drug- and chemical-induced kidney injuries. The maps generated by this particular approach conveyed an impressive amount of information in surprisingly comprehensible ways and were an inspiring way of visualizing data. The maps also were highly functional and identified several possible biomarkers for kidney injury, including the Heme Oxygenase 1, sequestosome 1, changes in cell and nucleus texture, and chromatin restructuring. Though Dr. Ramm acknowledged that there were still quite a few limits on what could be predicted using this system, it was clear that the technique will inform the development of future cell-based prediction systems.
Indeed, the session provided an exciting look at the cutting edge of predictive toxicology and safety evaluation technologies, and we can be sure to see increased adoption of them in years to come.
This blog was prepared by an SOT Reporter. SOT Reporters are SOT members who volunteer to write about sessions and events they attend during the SOT Annual Meeting and ToxExpo. If you are interested in participating in the SOT Reporter program in the future, please email Giuliana Macaluso.