Incorporating Population Variability and Susceptible Subpopulations into Dosimetry for High-Throughput Toxicity Testing

TOXICOLOGICAL SCIENCES, 142(1), 2014, 210–224 doi: 10.1093/toxsci/kfu169 Advance Access Publication Date: August 21, 2014 Incorporating Population Variability and Susceptible Subpopulations into Dosimetry for High-Throughput Toxicity Testing *The Hamner Institutes for Health Sciences, Research Triangle Park, North Carolina 27709-2137, † United States Environmental Protection Agency, Office of Research and Development, National Center for Computational Toxicology, Research Triangle Park, North Carolina 27711 and ‡ Simcyp Limited (a Certara company), Blades Enterprise Centre, John Street, Sheffield S2 4SU, UK 1 To whom correspondence should be addressed at The Hamner Institutes for Health Sciences, 6 Davis Drive, PO Box 12137, Research Triangle Park, NC 27709. Fax: (919) 558-1300. E-mail: . 2 Present address: United States Environmental Protection Agency, Office of Research and Development, National Center for Computational Toxicology, Research Triangle Park, NC 27711. ABSTRACT Momentum is growing worldwide to use in vitro high-throughput screening (HTS) to evaluate human health effects of chemicals. However, the integration of dosimetry into HTS assays and incorporation of population variability will be essential before its application in a risk assessment context. Previously, we employed in vitro hepatic metabolic clearance and plasma protein binding data with in vitro in vivo extrapolation (IVIVE) modeling to estimate oral equivalent doses, or daily oral chemical doses required to achieve steady-state blood concentrations (Css ) equivalent to media concentrations having a defined effect in an in vitro HTS assay. In this study, hepatic clearance rates of selected ToxCast chemicals were measured in vitro for 13 cytochrome P450 and five uridine 5 -diphospho-glucuronysyltransferase isozymes using recombinantly expressed enzymes. The isozyme-specific clearance rates were then incorporated into an IVIVE model that captures known differences in isozyme expression across several life stages and ethnic populations. Comparison of the median Css for a healthy population against the median or the upper 95th percentile for more sensitive populations revealed differences of 1.3- to 4.3-fold or 3.1- to 13.1-fold, respectively. Such values may be used to derive chemical-specific human toxicokinetic adjustment factors. The IVIVE model was also used to estimate subpopulation-specific oral equivalent doses that were directly compared with subpopulation-specific exposure estimates. This study successfully combines isozyme and physiologic differences to quantitate subpopulation pharmacokinetic variability. Incorporation of these values with dosimetry and in vitro bioactivities provides a viable approach that could be employed within a high-throughput risk assessment framework. Key words: In vitro in vivo extrapolation; population variability; toxicokinetics; reaction phenotyping; dosimetry; risk assessment Over the past five years, high-throughput in vitro screening (HTS) has been advocated as the future of toxicity testing (Andersen and Krewski, 2009; Collins et al., 2008; NRC, 2007). The ability to screen thousands of chemicals across hundreds of molecular targets and pathways provides an efficient, economical, and humane alternative to the current use of high-dose animal-based  C The Author 2014. Published by Oxford University Press on behalf of the Society of Toxicology. All rights reserved. For permissions, please email: 210 Barbara A. Wetmore*,1 , Brittany Allen*, Harvey J. Clewell, III*, Timothy Parker*, John F. Wambaugh† , Lisa M. Almond‡ , Mark A. Sochaski*, and Russell S. Thomas*,2 WETMORE ET AL. FIG. 1. Schematic representation of approach to incorporate population-specific dosimetry with high-throughput in vitro screening data. Isozyme-specific clearance rates for each chemical were measured across 18 recombinantly expressed cytochrome P450s and UGTs. Using a population-based in vitro in vivo extrapolation model (IVIVE), steady state plasma concentrations (Css ) were predicted by incorporating experimentally measured plasma protein binding and isozymespecific clearance data with population-specific physiology and isozyme ontogenies scaled to capture known abundances across different life stages and ethnic populations. Population-specific Css values were predicted; in addition, Monte Carlo simulations captured variability within each population, with median, lower, and upper 95th percentile values obtained as output. Using reverse dosimetry, population-specific plasma Css were incorporated with chemical in vitro bioactivity concentrations as measured in ToxCast to estimate population-specific oral equivalents. These values can then be compared against population-specific exposure estimates. cantly impact clearance rates. Life stage-based differences in the abundances of CYPs 1A2, 2C8, 2C9, 2C18, 2C19, 2D6, 2E1, 3A4, and UGTs 1A1, 1A4, 1A6, 1A9, and 2B7 have been documented and extensively characterized from fetal and pediatric life stages to the adult (Cresteil, 1998; Hines, 2007; Johnson et al., 2006; Kearns et al., 2003; Leakey et al., 1987; Miyagi and Collier, 2007; Salem et al., 2013; Yokoi, 2009). Genetic polymorphisms of CYPs 2B6, 2C9, 2C19, and 2D6 have also been well characterized, resulting in differing drug metabolizing capabilities for up to 20% of the Asian population for CYP 2C19 substrates and 10% of the Caucasian population for CYP2D6 substrates (Hiratsuka, 2012; McGraw and Waller, 2012). Incorporation of these enzyme differences in the measurement of xenobiotic clearance rates would provide key information in determining the range and extent of inter-individual PK variability. In this study, we outline an approach to link in vitro HTS results with subpopulation-specific dosimetry and exposure estimates. For subpopulation dosimetry, in vitro intrinsic clearance of a subset of nine ToxCast chemicals was determined across a panel of recombinant CYP (13) and UGT (5) isozymes and used in conjunction with population databases describing the genetic, ethnic, and life-stage (ontogeny)-related differences in enzyme abundance and physiology (Fig. 1). This approach allowed the estimation of oral equivalent dose values for each of the previously published ToxCast assays across a range of life stages and ethnic populations. The oral equivalent doses for each chemicalassay combination were compared with the corresponding exposure estimates for each population. The approach also provides the ability to estimate chemical-specific human toxicokinetic adjustment factors (HKAF ) for use in risk assessment (IPCS, 2005) (Lipscomb et al., 2004). studies. Although these attributes make HTS desirable within a toxicity testing framework, several key considerations need to be addressed before it can be realistically considered within the context of risk assessment. First, the chemical concentration in a well that elicits in vitro biological activity may be differen (...truncated)