Dose-response modeling of cytochrome P450 induction in rats by octamethylcyclotetrasiloxane

Abstract

Inhalation of octamethylcyclotetrasiloxane (D4) induces CYP2B1/2 protein and causes liver enlargement. We have developed a pharmacodynamic (PD) extension to a physiologically based pharmacokinetic (PBPK) model to characterize these dose-response behaviors. The PD model stimulates interactions of D4 with a putative receptor, leading to increased production of cytochrome P450 2B1/2. Induction was modeled with a Hill equation with dissociation constant, Kd, and Hill coefficient, N. Both a 1- and a 5-compartment liver model were evaluated. The PBPK model provided excellent simulations of tissue D4 and hepatic CYP2B1/2 protein concentrations following 6 h/day, 5-day inhalation exposures to 0, 1, 7, 30, 70, 150, 300, 500, 700, or 900 ppm D4. Either the 1- or 5-compartment liver model could accurately simulate increases in CYP2B1/2 protein in the liver. With a 1-compartment liver, Kd and N were 0.67 μM (free liver concentration) and 1.9, respectively. The 5-compartment model used higher N-values (∼ 4.0) and varied Kd between compartments. The fitted 5-compartment model parameters were Kd = 0.67 μM in the midzonal compartment with geometric differences in Kd between compartments of 2.9. On the basis of unbound (free) plasma concentrations, D4 appeared to be a higher potency inducer than phenobarbital (PB). Dose-response curves for increased liver weights had N ≃ 1.0 and Kd ≃ 3.4 μM, very different values from those for enzyme induction. Exposure concentration leading to a 0.1% increase in CYP2B1/2 protein predicted by the 1- and 5-compartment models were 2.1 ppm and 5.1 ppm, respectively. The 1- and 5-compartment liver models provided very similar fits to the whole liver induction data, excluding the lowest dose, but the 5-compartment liver model had the additional advantage of simultaneously describing the regional induction of CYP2B1/2.