Incorporating human interindividual biotransformation variance in health risk assessment.

The protection of sensitive individuals within a population dictates that measures other than central tendencies be employed to estimate risk. The refinement of human health risk assessments for chemicals metabolized by the liver to reflect data on human variability can be accomplished through (1) the characterization of enzyme expression in large banks of human liver samples, (2) the employment of appropriate techniques for the quantification and extrapolation of metabolic rates derived in vitro, and (3) the judicious application of physiologically based pharmacokinetic (PBPK) modeling. While in vitro measurements of specific biochemical reactions from multiple human samples can yield qualitatively valuable data on human variance, such measures must be put into the perspective of the intact human to yield the most valuable predictions of metabolic differences among humans. For quantitative metabolism data to be the most valuable in risk assessment, they must be tied to human anatomy and physiology, and the impact of their variance evaluated under real exposure scenarios. For chemicals metabolized in the liver, the concentration of parent chemical in the liver represents the substrate concentration in the Michaelis Menten description of metabolism. Metabolic constants derived in vitro may be extrapolated to the intact liver, when appropriate conditions are met. Metabolic capacity Vmax; the maximal rate of the reaction) can be scaled directly to the concentration of enzyme (or enzyme fraction) contained in the liver. Several environmental, genetic and lifestyle factors can influence the concentration of cytochrome P450 forms (CYP) in the liver by affecting either (1) the extent to which the CYP forms are expressed in the endoplasmic reticulum of the cell (isolated as the microsomal fraction from tissue homogenates), or (2) the expression of microsomal protein in intact liver tissue. Biochemically sound measures of the hepatic distribution of xenobiotic metabolizing enzymes among humans, based on expression of the enzymes within microsomal protein and the distribution of microsomal protein among intact livers, can be combined with metabolic constants derived in vitro to generate values consistent with those employed in PBPK models. When completed, the distribution (and bounds) of Vmax values can be estimated and included in PBPK models. Exercising such models under plausible exposure scenarios will demonstrate the extent to which human interindividual enzyme variance can influence parameters (i.e., the detoxication of a toxic chemical through metabolism) that may influence risk. In this article, we describe a methodology and conditions which must exist for such an approach to be successful.