Computational modeling of serum-binding proteins and clearance in extrapolations across life stages and species for endocrine active compounds.

One measure of the potency of compounds that lead to the effects through ligand-dependent gene transcription is the relative affinity for the critical receptor. Endocrine active compounds that are presumed to act principally through binding to the estrogen receptor (e.g., estradiol, genistein, bisphenol A, and octylphenol) comprise one class of such compounds. For making simple comparisons, receptor-binding affinity has been equated to in vivo potency, which consequently defines the dose-response characteristics for the compound. Direct extrapolation of in vitro estimated affinities to the corresponding in vivo system and to specific species or life stages (e.g., neonatal, pregnancy) can be misleading. Accurate comparison of the potency of endocrine active compounds requires characterization of biochemical and pharmacokinetic factors that affect their free concentration. Quantitative in vitro and in vivo models were developed for integrating pharmacokinetics factors (e.g., serum protein and receptor-binding affinities, clearance) that affect potency. Data for parameterizing these models for several estrogenic compounds were evaluated and the models exercised. While simulations of adult human or rat sera were generally successful, difficulties in describing early life stages were identified. Exogenous compounds were predicted to be largely ineffective at competing estradiol off serum-binding proteins, suggesting this was unlikely to be physiologically significant. Discrepancies were identified between relative potencies based upon modeling in vitro receptor-binding activity versus in vivo activity in the presence of clearance and serum-binding proteins. The examples illustrate the utility of this approach for integrating available experimental data from in vitro and in vivo studies to estimate the relative potency of these compounds.