This blog contains key takeaways from the Symposium Session “Challenges in the Development of In Vitro-In Vivo Extrapolation Models for Next-Generation Risk Assessment” that was presented during the 2023 SOT Annual Meeting and ToxExpo.
John Wambaugh and Mathieu Vinken co-chaired the session. Next-Generation risk assessments (NGRA) rely on new approach methods (NAMs), which include human in vitro and in silico approaches for hazard identification and characterization. One key aspect for risk assessment is the derivation of a point of departure (POD), which for NGRA might be derived by in vitro to in vivo extrapolation (IVIVE) from an in vitro benchmark concentration.
In the EU, the newly launched ONTOX (ontology-driven and artificial intelligence–based repeated dose toxicity testing of chemicals for next-generation risk assessment) program is developing a strategy to create innovative NAMs to predict systemic repeated dose toxicity effects, while the new program RISK-HUNT3R (Risk Assessment of Chemicals Integrating Human-Centric Next-Generation Testing Strategies Promoting the 3Rs) will develop, validate, and implement integrated approaches to lead the way toward NGRA.
In the US, the federal Tox21 and US Environmental Protection Agency (US EPA) ToxCast and ExpoCast programs also have been working to develop a working NGRA to inform chemical safety decision-making.
The IVIVE approach relies on in vitro biokinetics modeling to define the free intracellular concentration of test substances in the in vitro test conditions. Part of NGRA are generic physiologically based kinetic (PBK, also known by PBTK and PBPK) models parameterized using human in vitro inputs (clearance, absorption, and binding).
Andreas Schepky and Abdulkarim Najjar discussed a NGRA framework that was recently developed for cosmetic ingredients. This framework integrates non-animal NAMs to ensure human safety of cosmetics ingredients without generating animal data. NGRA requires the extrapolation of in vitro PODs to equivalent external exposures relevant to cosmetic exposure scenarios. PBK modeling provides a means for translating in vitro concentration response relationships to in vivo dose-response relationships in humans.
The case study presented utilized PBK modeling for reverse-dosimetry of in vitro POD on the estrogenic pathway for genistein and daidzein and considered relevant cosmetic exposure scenarios. The genistein PBK model was validated using in vivo kinetic data, which confirmed the ability to reproduce the observed kinetic parameters. The case study demonstrated the applicability and crucial role of PBK modeling in NGRA, as well as provided strategies for building valid PBK models without animal data using the read-across concept proposed by OECD.
There are challenges to improve the predictive performance of the developed models and the following presentations addressed in vitro distribution and PBK modeling.
Susana Duarte Lopes Mascarenhas Proença and Nynke Kramer pointed out that in vitro concentration-effect relationships are traditionally based on nominal concentrations, but this dose metric does not accurately reflect the concentration of the chemical causing toxicity in cells in vitro. The extent to which a chemical partitions into cells will depend on physicochemical properties and binding affinities to assay constituents such as plastic, serum protein, and cell lipids. As part of the ONTOX project, a gap analysis was performed to systematically evaluate and extend the chemical and assay applicability domain of in vitro disposition models for in vitro assays assessing molecular perturbations related to steatosis, cholestasis, developmental neurotoxicity, and tubular necrosis.
A literature review was performed to compare the chemical and assay applicability domains of in vitro disposition models published in literature. Using artificial intelligence, a database was constructed listing chemicals associated with the previously mentioned ontologies and their physicochemical and kinetic properties. The database is used to analyze the chemical space of these chemicals and assess the extent to which they fit within the chemical and assay applicability domains of in vitro disposition models.
Dr. Wambaugh presented that NG RA hinges on quantitative determination of surrogate PODs using IVIVE. Understanding in vitro disposition is critical for IVIVE since the free effective concentration might be a hundredth or a hundred times the nominal test concentration. While mathematical models exist for predicting in vitro disposition from physicochemical properties, the data for evaluating these predictions represent limited chemical structure diversity. The chemical library of the US Tox21 screening program contains thousands of diverse chemicals. The Tox21 library has already been screened in concentration-response mode for diverse bioactivities using high-throughput in vitro assays. In some cases, PODs based on nominal in vitro tested concentration have been identified. His presentation described how Tox21 has been collecting new data characterizing in vitro disposition of chemicals to assess any differences between nominal and free concentration. These data permit evaluation of a variety of mathematical models for in vitro disposition across a wider range of physicochemical properties, including key chemical classes found in commerce and the environment. Accurate prediction of in vitro disposition will enhance the predictive power of quantitative NGRA.
Iain Gardner discussed how PBK models simulate the absorption, distribution, metabolism, and excretion of a chemical through the body and are suited to extrapolate effect concentrations of chemicals from in vitro toxicity assays to human bioequivalent doses. IVIVE-PBK models have been used extensively to describe the plasma and tissue concentration profiles of drugs and chemicals. Within the RISK-HUNT3R project, different strategies have been investigated to produce IVIVE-PBK models for some compounds. This includes using only in silico estimates of model inputs compared to in vitro model inputs and a middle out approach whereby in vitro data and some human pharmacokinetic data were used to inform the PBK model input parameters. It is important to investigate approaches that include key metabolites into the PBK model. Examples of parent and metabolite IVIVE-PBK models were presented.
Sylvia Escher discussed how in the RIS-HUNT3R project, the extent of systemic absorption of (volatile) compounds in the respiratory tract, intestine, and skin is investigated by using and developing innovative in vitro models. The kinetics of relevant metabolites also are quantified, and this information is integrated into PBK models to perform IVIVE. To be able to quantify the uncertainty of the obtained prediction, RISK-HUNT3R develops the concept by using data-rich case study compounds for which reference in vivo toxicokinetic data are available. Case study results were presented.
This blog reports on the Symposium Session titled “Challenges in the Development of In Vitro-In Vivo Extrapolation Models for Next-Generation Risk Assessment” that was held during the 2023 SOT Annual Meeting and ToxExpo. An on-demand recording of this session is available for meeting registrants on the SOT Online Planner and SOT Event App.
This blog was prepared by an SOT Reporter and represents the views of the author. SOT Reporters are SOT members who volunteer to write about sessions and events in which they participate during the SOT Annual Meeting and ToxExpo. SOT does not propose or endorse any position by posting this article. If you are interested in participating in the SOT Reporter program in the future, please email SOT Headquarters.