Application of generalized concentration addition to predict mixture effects of glucocorticoid receptor ligands

Environmental exposures often occur in complex mixtures and at low concentrations. Generalized concentration addition (GCA) is a method used to estimate the joint effect of receptor ligands that vary in efficacy. GCA models have been successfully applied to mixtures of aryl hydrocarbon receptor (AhR) and peroxisome proliferator-activated receptor gamma (PPARγ) ligands, each of which can be modeled as a receptor with a single binding site. Here, we evaluated whether GCA could be applied to homodimer nuclear receptors, which have two binding sites, to predict the combined effect of full glucocorticoid receptor (GR) agonists with partial agonists. We measured transcriptional activation of GR using a cell-based bioassay. Individual concentration-response curves for dexamethasone (full agonist), prednisolone (full agonist), and medroxyprogesterone 17-acetate (partial agonist) were generated and applied in three additivity models, GCA, effect summation (ES), and relative potency factor (RPF), to generate response surfaces. GCA and RPF yielded adequate predictions of the experimental data for two full agonists. However, GCA fit experimental data significantly better than ES and RPF for all other binary mixtures. This work extends the application of GCA to homodimer nuclear receptors and improves prediction accuracy of mixture effects of GR agonists.