Using chemical structure information to develop predictive models for in vitro toxicokinetic parameters to inform high-throughput risk-assessment

Abstract

The toxicokinetic (TK) parameters fraction of the chemical unbound to plasma proteins and metabolic clearance are critical for relating exposure and internal dose when building in vitrobased risk assessment models. However, experimental toxicokinetic studies have only been carried out on limited chemicals of environmental interest (~1000 chemicals with TK data relative to tens of thousands of chemicals of interest). This work evaluated the utility of chemical structure information to predict TK parameters in silico; development of cluster-based read-across and quantitative structure-activity relationship models of fraction unbound or fub (regression) and intrinsic clearance or Cl int (classification and regression) using a dataset of 1487 chemicals; utilization of predicted TK parameters to estimate uncertainty in steady-state plasma concentration (C ss ); and subsequent in vitro-in vivo extrapolation analyses to derive bioactivity-exposure ratio (BER) plot to compare human oral equivalent doses and exposure predictions using androgen and estrogen receptor activity data for 233 chemicals as an example dataset. The results demonstrate that fub is structurally more predictable than Cl int . The model with the highest observed performance for fub had an external test set RMSE/σ=0.62 and R 2 =0.61, for Cl int classification had an external test set accuracy = 65.9%, and for intrinsic clearance regression had an external test set RMSE/σ=0.90 and R 2 =0.20. This relatively low performance is in part due to the large uncertainty in the underlying Cl int data. We show that C ss is relatively insensitive to uncertainty in Cl int . The models were benchmarked against the ADMET Predictor software. Finally, the BER analysis allowed identification of 14 out of 136 chemicals for further risk assessment demonstrating the utility of these models in aiding risk-based chemical prioritization.

Comprehensive interpretation of in vitro micronucleus test results for 292 chemicals: from hazard identification to risk assessment application

“…These data were unavailable for many of the compounds tested in this study; therefore, in silico predictions were used to obtain predictions for these parameters (Supplemental File 1). Previous work has demonstrated that ADMET Predictor can provide reliable estimates of in vitro parameters for HTTK modeling to derive stable estimates of C ss (Pradeep et al 2020 ). Thus, ADMET Predictor version 9.5 was used to provide F up percentage (hum_fup%) and human liver microsomal clearance (CYP_HLM_CLint) by CASRN (CAS number).…”

IVIVE: Facilitating the Use of In Vitro Toxicity Data in Risk Assessment and Decision Making

“…However, new efforts are examining systematic approaches to metals and PK models that generally allow for chemical transformation, including cycling [ 255 ]. In addition, for PK models, the rapid growth in informatics has allowed the development of many approaches relating chemical structure features to important properties, including in vitro PK measurements [ 41 , 78 , 84 ]. Quantitative structure–property relationships (QSPRs) are rapidly developing, and both new models and consensus predictors based on multiple models should be expected [ 79 , 256 ].…”

Artificial Intelligence-Based Toxicity Prediction of Environmental Chemicals: Future Directions for Chemical Management Applications

“…In the study of prediction of changes in cell activity, high-throughput cellular phenotype images or chemical response signatures were learned to predict cellular phenotypes or activity patterns that could be linked to in vivo outcomes. , Also, recently, toxicokinetics prediction studies to extrapolate in vivo toxicity based on in vitro HTS data are being conducted. In these studies, two key parameters related to toxicokinetics, intrinsic clearance rate (Cl int ) and the fraction of the chemical unbound in plasma (fub), were predicted. , However, studies using data other than chemical structure information have not been explored much, so more efforts are needed for research using these data to predict the toxicity of chemicals.…”

“…In these studies, two key parameters related to toxicokinetics, intrinsic clearance rate (Cl int ) and the fraction of the chemical unbound in plasma (fub), were predicted. 96,97 However, studies using data other than chemical structure information have not been explored much, so more efforts are needed for research using these data to predict the toxicity of chemicals.…”