Using in vitro data to derive acceptable exposure levels for environmental chemicals: A case study on p,p'-DDE...

Current acceptable exposure levels are largely based on animal models, which are costly, time-consuming, and may poorly predict adverse outcomes in humans. Alternative testing methods are needed to adequately tackle the large number of environmental chemicals.

To evaluate a method integrating human in vitro data and computational modeling to calculate acceptable exposure levels through a case study on early-life p,p'-dichlorodiphenyldichloroethylene (p,p'-DDE) and developmental obesogenicity.

We reviewed in vitro studies of p,p'-DDE and obesogenicity-related endpoints to select points of departure (PODs). Nominal PODs were converted into lipid-based cellular concentrations using a dynamic mass-balance model. Cellular concentrations were converted into tolerable daily intakes and biomonitoring equivalents in pregnant individuals using a toxicokinetic model and uncertainty factors. We compared estimated biomonitoring equivalents to maternal and cord plasma levels measured in epidemiological studies reporting associations between early-life p,p'-DDE exposure and child adiposity.

We estimated PODs for phenotypic (181,897 ng/g lipids) and transcriptomic (14,405 ng/g lipids) endpoints. Application of the toxicokinetic model and uncertainty factors led to tolerable daily intakes of 2.50-8.65 ng/kg/day (phenotypic) and 0.198-0.685 ng/kg/day (transcriptomic). Corresponding biomonitoring equivalents were 54.4-188 ng/g lipids (phenotypic) and 4.31-14.9 ng/g lipids (transcriptomic). Mean/median concentrations measured in epidemiological studies of p,p'-DDE exposure and child adiposity were mostly within or above the range of concentrations produced using the phenotypic in vitro POD.

This study adds to a growing body of literature on the potential of in vitro data combined with computational modeling for chemical risk assessment, while also identifying challenges to regulatory adoption.