Abstract
There are many types of biomarkers; the two common ones are biomarkers of exposure and biomarkers of effect. The utility of a biomarker for estimating exposures or predicting risks depends on the strength of the correlation between biomarker concentrations and exposure/effects. In the current study, a combined exposure and physiologically-based pharmacokinetic/pharmacodynamic (PBPK/PD) model of carbaryl was used to demonstrate the use of computational modeling for providing insight into the selection of biomarkers for different purposes. The Cumulative and Aggregate Risk Evaluation System (CARES) was used to generate exposure profiles, including magnitude and timing, for use as inputs to the PBPK/PD model. The PBPK/PD model was then used to predict blood concentrations of carbaryl and urine concentrations of its principal metabolite, 1-naphthol (1-N), as biomarkers of exposure. The PBPK/PD model also predicted acetylcholinesterase (AChE) inhibition in red blood cells (RBC) as a biomarker of effect. The correlations of these simulated biomarker concentrations with intake doses or brain AChE inhibition (as a surrogate of effects) were analyzed using a linear regression model. Results showed that 1-N in urine is a better biomarker of exposure than carbaryl in blood, and that 1-N in urine is correlated with the dose averaged over the last 2 days of the simulation. They also showed that RBC AChE inhibition is an appropriate biomarker of effect. This computational approach can be applied to a wide variety of chemicals to facilitate quantitative analysis of biomarker utility.