Physiologically-based pharmacokinetic modeling of simple and complex mixtures of gasoline and the gasoline components N-hexane, benzene, toluene, ethylbenzene, and xylene

Abstract

Gasoline consists of hundred of hydrocarbon components, some of the principal ones being n-hexane, benzene, toluene, ethylbenzene and xylenes. Various components of gasoline cause different toxicological effects, including depression of the central nervous system, irritation, and cancer. The assessment of the dose response of different components in various mixtures of the components or in complex mixtures such as gasoline is, however, confounded by the fact that the components are capable of interacting within the mixture. One of the best understood interactions occurs during metabolism, a pharmacokinetic process that results in clearance of the parent chemical from the body and production of metabolites that may be more or less toxic than the parent. The studies described herein were intended to clarify the extent of these interactions, and to develop a quantitative description of the pharmacokinetics of gasoline and mixtures of some of its components. A simple mixture can be described as a well-defined mixture of two or more chemicals. The interactions of a simple mixture of the gasoline components benzene, toluene, ethylbenzene, and xylene (BTEX) have been previously described. However, the significance of the interactions in a real-world context had not been explored. Using established physiologically-based pharmacokinetic (PBPK) models, we studied the effect of exposing humans at the Threshold Limit Values and Permissible Exposure Limits, and included the effect of exercise. Depending on the accepted measure of toxicity, the PBPK model indicates that over-exposure to these chemicals can occur under a variety of real world scenarios. If a simple mixture is well-defined, a complex mixture is a mixture of many compounds that are not present in specified amounts. Existing approaches for building PBPK models rely on building single chemical PBPK models for each component and then mathematically joining them in a biologically relevant way. This is only practical for mixtures that have a limited number of components. A method was therefore developed to cope with the hundreds of components in gasoline. Specific components were handled in the conventional manner. To account for mixture interactions between remaining components and the specifically identified ones, the remaining components were assumed to act like a single additional chemical with properties that were averaged over the range of components in the group. These models were developed using in vivo pharmacokinetic data from rat studies. Rats were placed in sealed chambers and the rate of disappearance of chemicals as the animals absorbed them was determined. Existing chamber designs were improved by developing a method for measuring carbon dioxide in the chamber. While not used in the studies herein, an additional method that allows simultaneous sampling of blood was also developed. The pharmacokinetic data supported the development of a PBPK model for gasoline, with BTEX, n-hexane and an additional lumped component. A winter blend and summer blend of gasoline were analyzed and a good pharmacokinetic description of the mixture and the interactions between its components is presented. To assess the ability of the approach to deal with variable composition mixtures, volatile fractions of gasoline were also studied. The model successfully described the pharmacokinetics of various fractions of gasoline that were weighted by volatility. These pharmacokinetic descriptions of gasoline were limited to the pharmacokinetics of the parent chemicals and did not address metabolites. To take a closer look at some metabolite issues, the chemical n-hexane was explored in more depth, n-hexane is metabolized to several intermediates. The downstream intermediate hexanedione is responsible for n-hexane's neurotoxicological effects. Counterintuitively, more hexanedione is formed in rats acutely exposed to 1000 ppm than exposed to 3000 ppm, i.e., the chemical has a negative dose-response curve when hexanedione is the response metric. A PBPK model was developed, based on data from previous studies, to determine the underlying cause of this effect. It was found that, if inhibition of downstream oxidative metabolic steps is assumed to be caused by n-hexane and other metabolites in blood, hexanedione production will be substantially depressed during exposure. After exposure ends, a large increase in hexanedione formation can occur as the inhibition is released. A reasonable correlation between the area-under-the-curve of hexanedione in the body and neurological outcomes was found. These studies thus evaluate several types of mixtures that occur with this group of compounds. The initial study assessed interactions in a simple mixture of toluene, ethylbenzene, and xylene, three important constituents of gasoline. The next studies assessed interactions on a broader scale, in a complex mixture of gasoline components. Finally, interactions between a chemical and two of its metabolites was investigated with n-hexane. While all of these mixtures exhibit similar pharmacokinetic interaction properties, we found that significant impacts on the pharmacokinetics of each element of each mixture can occur.