Physiologically based kinetic (PBK) models to give insight into dose-, species-, matrix- and interindividual human variation-dependent effects on bioactivation and detoxification of methyleugenol

Methyleugenol, which occurs naturally in various herbs such as tarragon, basil, nutmeg and allspice, is added to food either directly as a flavoring substance or as a constituent of added essential oils (Smith et al., 2002). The interest in the risk of methyleugenol as a food constituent came from its widespread use in a variety of foods and beverages as well as its structural resemblance to the known carcinogen safrole (Johnson et al., 2000). In addition, methyleugenol has been reported to be DNA reactive and carcinogenic, inducing malignant tumors in multiple tissues of rats and mice as well as inducing unscheduled DNA synthesis in rat liver (Ding et al., 2011; NTP, 2000; Smith et al., 2002). The safety of human exposure to methyleugenol at low dietary intake levels has been assessed several times (Hall and Oser, 1965; NTP, 2000; SCF, 2001; Smith et al., 2002) without reaching a scientific agreement on how to translate the carcinogenicity data of rodent animal experiments obtained at high levels of exposure to the relevant human situation. A recent evaluation, performed by the Joint FAO/WHO Expert Committee on Food Additives (JECFA) in 2008, has indicated that although evidence of carcinogenicity to rodents given high doses of methyleugenol exists, further research is needed to assess the potential risk to human health at relevant dietary exposure resulting from the presence of methyleugenol in foods and essential oils and its use as flavoring agent (JECFA, 2008). Predicting the cancer risk in humans at relevant dietary intake levels requires extrapolation of the animal carcinogenicity data taking in consideration dose, species, and interindividual variation. Furthermore, it implies extrapolation from rat or mouse studies with high dose levels of the pure compound to the human situation in which exposure at low dose levels occurs within the context of a complex food matrix. The aim of the present PhD project was to obtain quantitative insight into the consequences of dose- and species-dependent effects and of interindividual differences and matrix effects for the bioactivation and detoxification of methyleugenol by using physiologically based kinetic (PBK) modeling. The first chapter of this thesis presents background information to the topic. In chapter 2, a physiologically based kinetic (PBK) model for the alkenylbenzene methyleugenol in rat was defined based on in vitro metabolic parameters determined using relevant tissue fractions, in silico derived partition coefficients (Payne and Kenny, 2002 and reference therin), and physiological parameters (Brown et al., 1997) derived from the literature. The model was based on the model previously developed for the related alkenylbenzene estragole and consists of eight compartments including liver, lung, and kidney as metabolizing compartments, and separate compartments for fat, arterial blood, venous blood, richly perfused and slowly perfused tissues (Punt et al., 2008). Evaluation of the model was performed by comparing the PBK predicted concentration of methyleugenol in the venous compartment to methyleugenol plasma levels reported in the literature, by comparing the PBK predicted dose-dependent % of formation of 2-hydroxy-4,5-dimethoxyallylbenzene, 3-hydroxy-4-methoxyallylbenzene, and 1′- hydroxymethyleugenol glucuronide to the corresponding % of metabolites excreted in urine reported in the literature, which were demonstrated to be in the same order of magnitude (Solheim and Scheline, 1976). With the model obtained the relative extent of bioactivation and detoxification of methyleugenol at different oral doses was examined. At low doses, formation of3-(3,4-dimethoxyphenyl)-2-propen-1-olandmethyleugenol-2′,3′-oxideleadingto detoxification appear to be the major metabolic pathways, occurring in the liver. At high doses, the model reveals a relative increase in the formation of the proximate carcinogenic metabolite 1′- hydroxymethyleugenol, occurring in the liver. This relative increase in formation of 1′- hydroxymethyleugenol leads to a relative increase in formation of 1′-hydroxymethyleugenol glucuronide, 1′-oxomethyleugenol, and 1′-sulfooxymethyleugenol the latter being the ultimate carcinogenic metabolite of methyleugenol. These results indicate that the relative importance of different metabolic pathways of methyleugenol may vary in a dose-dependent way, leading to a relative increase in bioactiviation of methyleugenol at higher doses. In subsequent studies described in chapter 3 a physiologically based kinetic (PBK) model for methyleugenol in human based on in vitro and in silico derived parameters was identified based on the model previously developed for the related alkenylbenzene estragole. The model consists of six compartments including liver as metabolizing compartment, and separate compartments for fat, arterial blood, venous blood, richly perfused and slowly perfused tissues (Punt et al., 2009). With the model obtained, bioactivation and detoxification of methyleugenol at different dose levels could be investigated. The outcomes of this human model were compared with those of the PBK model for methyleugenol in male rat. The results obtained reveal that formation of 1′-hydroxymethyleugenol glucuronide, a major metabolic pathway in male rat liver, appears to represent a minor metabolic pathway in human liver whereas in human liver a significantly higher formation of 1′-oxomethyleugenol compared with male rat liver is observed. Furthermore, formation of 1′-sulfooxymethyleugenol, which readily undergoes desulfonation to a reactive carbo-cation that can form DNA or protein adducts, is predicted to be the same in the liver of both human and male rat at oral doses of 0.0034 up to 300 mg/(kg bw). Altogether it was concluded that despite a significant difference in especially the metabolic pathways of the proximate carcinogenic metabolite 1′-hydroxymethyleugenol between human and male rat, the influence of species differences on the ultimate overall bioactivation of methyleugenol to 1′-sulfooxymethyleugenol appears to be negligible. Moreover, the PBK model predicted the formation of 1′-sulfooxymethyleugenol in the liver of human and rat to be linear from doses as high as the benchmark dose (BMD10) down to as low as the virtual safe dose (VSD). This shows that kinetic data do not provide a reason to argue against linear extrapolation from the rat tumor data to the human situation. Another aim of the present PhD study was to study the effect of the basil constituent nevadensin on the bioactivation and genotoxicity of herb based methyleugenol. The results presented in chapter 4 show that nevadensin is able to inhibit DNA adduct formation in HepG2 cells exposed to the proximate carcinogen 1′-hydroxymethyleugenol in the presence of this flavonoid. This inhibition occurs at the level of sulfotransferase (SULT)-mediated bioactivation of 1′-hydroxymethyleugenol. In order to investigate possible in vivo implications the SULT inhibition by nevadensin was integrated into the male rat and human PBK models for bioactivation and detoxification of methyleugenol. The results thus obtained reveal that coadministration of methyleugenol with nevadensin may reduce the levels of bioactivation of 1′- hydroxymethyleugenol to the DNA reactive metabolite, without reducing its detoxification via glucuronidation or oxidation. This effect may be significant even at realistic low dose human exposure levels. The results obtained point at a potential reduction of the cancer risk when methyleugenol exposure occurs by oral intake within a relevant food matrix containing SULT inhibitors compared to what is observed in rodent bioassays upon exposure to pure methyleugenol dosed by gavage. Besides dose-dependent effects, species differences effects, and matrix effects on the bioactivation of methyleugenol the effect of interindividual variation on methyleugenol detoxification and bioactivation was investigated in chapter 5. To this end we predicted the level of formation of the ultimate carcinogenic metabolite 1′-sulfooxymethyleugenol in the human population by taking the variability in key bioactivation and detoxification reactions into account using Monte Carlo simulations. Insight in the variation in relevant metabolic routes was obtained by determining kinetic constants for the metabolic reactions by specific isoenzymes or by measuring the kinetic constants in incubations with a range of individual human liver fractions. The results of the study indicate that formation of 1′-sulfooxymethyleugenol is predominantly affected by i) P450 1A2 catalyzed bioactivation of methyleugenol to 1′- hydroxymethyleugenol ii) P450 2B6 catalyzed epoxidation of methyleugenol and iii) the apparent kinetic constants for detoxification of 1′-hydroxymethyleugenol via oxidation and iv) the apparent kinetic constants for bioactivation of 1′-hydroxymethyleugenol to 1′- sulfooxymethyleugenol. Based on the Monte Carlo simulation a chemical-specific adjustment factor (CSAF) for intraspecies variation could be derived which is defined as the 95th or 99th percentile divided by the 50th percentile of the predicted distribution of the formation of 1′- sulfooxymethyleugenol in the liver. The obtained CSAF value at the 95th percentile was 3.7 indicating that the default uncertainty factor of 3.16 for human variability in kinetics (WHO, 1999) may adequately protect 95% of the population. While protecting 99% of the population requires a larger uncertainty factor of 5.8. Altogether, the results shown in this thesis reveal that integrating in vitro metabolic parameters within a framework of a PBK model provides a good method to evaluate the occurrence of dose-dependent effects, species differences, and human variability in detoxification and bioactivation of a genotoxic carcinogen. Moreover, the results presented in this thesis show the possible protective effect of the basil constituent nevadensin on SULT catalysed bioactivation and DNA adduct formation of methyleugenol in vitro. Upon validation of these effects in vivo, it may turn out that rodent carcinogenicity data on methyleugenol substantially overestimate the risks posed when humans are exposed to methyleugenol within a nevadensin containing food matrix.