1.13.T-01 - Development of an In Vitro to In Vivo Extrapolation Approach To Assess Endocrine Activity in Fish

Abstract

---

The U.S. EPA is exploring New Approach Methods, like in vitro-to-in vivo extrapolation (IVIVE), to increase the pace and reduce the cost and animal usage of endocrine disruption screening for hundreds of pesticide chemicals. Toxicokinetic (TK) models are a key component of mammalian IVIVE approaches, but similar approaches for fish are limited. The objective here is to leverage IVIVE approaches used in human health assessments as a model for predicting endocrine activity in fish. Chemicals with in vivo fecundity lowest observed effect concentrations (LOEC) in fish were collected as benchmark reference values. Endocrine-related in vitro bioactivity points of departure (POD) from high-throughput assays (e.g. estrogen receptor [ER]) were collected for chemicals having corresponding in vivo fecundity LOECs. Empirical in vitro TK data (fraction unbound in blood [fUB] and intrinsic metabolic clearance [Clint]) were also collected for those same chemicals. Exposure and Safety Estimation Suite was then used to compile Quantitative Structure Activity Relationship (QSAR)-modeled fUB and whole-body biotransformation rates (kB), and to extrapolate empirical in vitro Clint to whole-body kB values for a comparison of empirical fUB and kB values to their QSAR-modeled counterparts. Lastly, reverse TK (rTK) modeling utilized in vitro PODs as free blood concentrations to evaluate the concordance between predicted exposure water concentrations and in vivo fecundity LOECs in fish. Both the empirically derived fUB and IVIVE (using empirical Clint) kB values exhibited positive and significant correlations to the QSAR-modeled values (p<0.05). The rTK-predicted exposure concentrations (using the most sensitive PODs) were significantly correlated to the in vivo fecundity LOECs (n=19, p=0.009, r2=0.30), and this relationship strengthened when constraining the dataset to chemicals with positive ER agonist scores (n=7, p=0.004, r2=0.80). The significant relationships between the empirical in vitro TK data and QSAR-modeled TK parameters indicate QSAR-modeled values are likely sufficient for TK model parameterization for this limited dataset. When constraining the dataset to chemicals with positive ER agonist scores, a strong relationship was observed between the rTK model-predicted exposure concentrations and empirical in vivo fecundity LOECs, showing promise as a viable tool for predicting in vivo fecundity effects in fish. This abstract does not necessarily reflect US EPA policy.

Author(s)

---

Presenter: Jaylen Sims, Oak Ridge Institute for Science and Education (ORISE),

Authors:
Jaylen Sims Oak Ridge Institute for Science and Education &#x28;ORISE&#x29; United States
Jonathan Haselman U.S. Environmental Protection Agency (US EPA) United States
Scott Lynn U.S. Environmental Protection Agency (US EPA) United States