Using in silico-in vitro to in vivo Extrapolation (IS-IVIVE) to Predict the Oral Dose Required to Activate the Aryl Hydrocarbon Receptor (AhR)

Authors: Lawless M, Fraczkiewicz R, DiBella J
Keywords: admet predictor, AhR, HTPK simulation module, IVIVE, live stream learning, oral dose
Software: ADMET Predictor®
Division: PBPK

Introduction

  • Activation of AhR can have toxic effects on mammalian hosts
  • Activation of AhR depends on the in vivo, steady state, unbound, plasma concentration of the compound and its relationship to the
    AC50 value and the Emax value in order to exhibit toxicity
  • We used IS-IVIVE to predict the oral dose needed to achieve this steady state plasma concentration1
  • AhR active compounds (AC50 < 100 μM) from a TOX21 assay were downloaded from PubChem2
  • 1,063 substances were active
  • In vitro fraction unbound in plasma (Fup) and human hepatocyte intrinsic clearance data was extracted from the httk R-package3
  • 1,425 records contained Fup values
  • 789 records were from EPA publications4,5
  •  The remaining data came from various publications
  • 1,020 records contained human hepatocyte CLint values
  • 828 records were from EPA publications4,5
  • The final data set consisted of 160 compounds that were in all three data sets
  • The data set contains agrochemicals, pharmaceuticals, and industrial chemicals
  • All other input properties, e.g., solubility, logP, pKa, were predicted from machine learning models1

By Michael Lawless, Robert Fraczkiewicz, John DiBella

Canceled: SOT 59th Annual Meeting & ToxExpo March 15-19. 2020 in Anaheim, California