Files
Abstract
A physiologically based pharmacokinetic (PBPK) model was experimentally parameterized and constructed for the conazole fungicides triadimefon and triadimenol. In vitro metabolic parameters for reduction of triadimefon to its primary metabolite, triadimenol, were measured in male and female Sprague-Dawley rats and CD-1 mice. Partition coefficients for triadimefon and triadimenol were measure in vitro in Sprague-Dawley rat tissues. A pharmacokinetic study of triadimefon and triadimenol disposition was performed following intravenous exposure to 50 mg/kg triadimefon in male Sprague-Dawley rats. Measured in vitro metabolic parameters and partition coefficients were incorporated into a PBPK model for triadimefon and triadimenol, and model simulations were compared to the pharmacokinetic data. The model could not adequately predict the complex distribution of both parent and metabolite during the clearance phase. Two possible explanations for this behavior were explored using alternate PBPK models: blood and tissue binding of triadimefon and triadimenol, and reverse metabolism of triadimenol to triadimefon. While the model with blood and tissue binding provided the best simulations of pharmacokinetic data, the individual binding parameters for each tissue were fit to the single pharmacokinetic data set, and the model lacked parsimony. The model with bidirectional metabolism (i.e. triadimefon reduction to triadimenol, and triadimenol oxidation to triadimefon) provided an improved fit relative to the original model, as well as a probable explanation supported by the available literature for the observed behavior. All three models were extrapolated to humans using human metabolic parameters, and human equivalent doses were calculated for dosimetrics from simulation of rat no observed adverse effects level (NOAEL) oral exposure. Comparison to oral reference dose for triadimefon in humans indicated that the value was sufficiently protective of human health.Finally, three methods of partition coefficient determination were compared for triadimefon and triadimenol: in vitro measurement, calculation from area under the curve for chemical concentration in in vivo pharmacokinetic data, and calculation by algorithm incorporating chemical- and tissue-specific information. The reverse metabolism model was employed to illustrate the differences between these methods. It was found that the algorithm method may over-estimate partition coefficient values, while in vitro and in vivo methods provided similar outcomes.