You are here:
Combining In Vivo, In Vitro And Toxicokinetics Data In Readacross: A Case Study Using Caffeine (SOT)
Citation:
Chavan, S., R. Judson, G. Patlewicz, R. Stark, P. Russell, AND P. Carmichael. Combining In Vivo, In Vitro And Toxicokinetics Data In Readacross: A Case Study Using Caffeine (SOT). Presented at SOT 2017, Baltimore, MD, March 12 - 16, 2017. https://doi.org/10.23645/epacomptox.5178604
Impact/Purpose:
poster prsentation at the SOT 2017 annual meeting
Description:
Readacross can be used to fill data gaps in chemical safety assessment. Readacross starts with the identification and evaluation of source analogs, including assessment of the physicochemical and mechanistic similarity of source analogs. We describe an approach to quantitative readacross prediction of point of departure (POD) using physicochemical, in vitro and in vivo data, with caffeine as an example target. Thirteen analogs were identified on the basis of structural similarity. Only theophylline, theobromine, paraxanthine were associated with any in vitro or in vivo experimental data. Relevant physicochemical properties (e.g. density, MP, BP, etc.) were all within a 1 log unit. Mechanistic in vitro assay information was used to determine bioactivity similarity. Caffeine and its analogs were active in three assays related to blockade of adenosine receptor A1 and A2a, demonstrating mechanistic similarity. Caffeine was as or more potent in vitro than any of the source analogs. In vivo POD data, specifically lowest effect levels (LELs) were available for 2 of the analogs; theophylline and theobromine. The LELs from rat subchronic studies were 75 and 250 mg/kg/day respectively. The LELs in rat chronic studies were 7.5 and 250 mg/kg/day, respectively. The lowest LEL value, 7.5 mg/kg/day, was taken as the readacross POD, and provides a conservative prediction of the LEL for caffeine (100 and 49 mg/kg/day in subchronic and chronic studies). A highthroughput toxicokinetics (HTTK) model was used as an in vitro to in vivo extrapolation approach for adenosine receptor activity (at 1 uM) which was the most representative assay, and yielded a POD for bioactivity of 0.1 mg/kg/day, significantly below the in vivo readacross POD. This value includes both human population variability and uncertainty in the kinetics, which is one reason for the low value. Additionally, this is the predicted dose at which desired activity occurs, as opposed to a higher dose where toxicity would be seen. The HTTK model is helpful in extrapolating from in vitro data, and can help to define a “safe” range for a target chemical. The readacross predictions derived for caffeine provide a proof of principle of this approach for using structure, physicochemical and in vitro data to help select readacross analogs. This abstract does not necessarily represent U.S. EPA policy
URLs/Downloads:
DOI: Combining In Vivo, In Vitro And Toxicokinetics Data In Readacross: A Case Study Using Caffeine (SOT)