Triage of Developmental and Reproductive Toxicants using an In vitro to In Vivo Extrapolation (IVIVE)-Toxicokinetic Computational modeling Application

1 PROJECT SUMMARY 2 3 A significant shift in the safety assessment of environmental chemicals is ongoing. This shift includes a reduction, 4 or in some cases elimination, of traditional toxicity testing in animals with the implementation of new approach 5 methodologies (NAMs). Although these emerging NAMs provide an attractive means for identifying teratogens 6 and other developmental toxicants, the absence of a quantitative framework for relating in vitro effects to 7 equivalent in vivo human exposures limits their use for public health. In Phase I, we successfully demonstrated 8 the importance of physiologically based pharmacokinetic (PBPK) models in extrapolating and assessing 9 developmental toxicity in vitro. The combination of these tools demonstrated that they could improve the safety 10 assessment of drugs and chemicals without animal testing The focus of this Phase II SBRI is to specifically 11 develop and validate New Approach Methodologies (NAMs) that combine in vitro reproductive assays 12 with a web browser-based graphical interfaces, DRIIVE (Developmental and Reproductive In Vitro to In 13 vivo Extrapolation), that includes a three-tiered PBPK model to predict human equivalent developmental 14 toxicity dose levels (HEDs) of drugs or chemicals that would correspond with in vitro concentrations associated 15 with adverse outcomes in in vitro developmental assay. In specific aim 1, we will design DRIIVE further to 16 provide researchers and regulators with a powerful resource to predict the pharmacokinetics of different 17 substances during pregnancy as well as human equivalent exposure that correspond with in vitro concentrations 18 associated with adverse outcomes in in vitro developmental assay. The application's scope will allow users to 19 choose from three different pregnancy models, parameter modification features, quantitative structure-activity 20 relationship (QSAR) models (predicting intrinsic clearance, unbound fraction in plasma, skin permeability, oral 21 absorption rate, partition coefficients), and advanced visualization tools. We will conduct software validation with 22 documented evidence that DRIIVE meets users' needs and functions according to its intended use. The criteria 23 for success will be assessed by running simulations in DRIIVE for 10 chemicals and demonstrating the 24 reproducibility of the software by comparing with published results (article or report). In specific aim 2, we will 25 select 50 to 100 chemicals to run DRIIVE to conduct a retrospective chemical space analysis. This analysis will 26 establish whether there are chemical classes or chemical regions for which the PBPK model predict 27 concentrations reliably within the Absolute Average Fold Error (AAFE) metric ≤2. The criteria for success will be 28 the identification of different clusters of compounds, based on the 50 to 100 compounds selected, for which 29 DRIIVE predictions give AAFE≤2. The analysis will also cluster an unknown test chemical to one of the chemical 30 clusters to inform the user if it's the chemical is within chemical space used for development and validation of 31 DRIIVE. This Phase II SBIR will provide the necessary validation required to develop an innovative 32 friendly user interface to provide a developmental and reproductive toxicity (DART) in vitro to in vivo 33 extrapolation (IVIVE) computational model to our customers. This developmental toxicity product fits into 34 the existing NAM products we offer that reduce reliance on animal models. 35