A physiologically based biokinetic (PBBK) model for the alkenylbenzene safrole in humans was developed based on in vitro- and in silico-derived kinetic parameters. With the model obtained, the time- and dose-dependent formation of the proximate and ultimate carcinogenic metabolites, 1-hydroxysafrole and 1-sulfooxysafrole in human liver were estimated and compared with previously predicted levels of these metabolites in rat liver. In addition, Monte Carlo simulations were performed to predict interindividual variation in the formation of these metabolites in the overall population. For the evaluation of the model performance, a comparison was made between the predicted total amount of urinary metabolites of safrole and the reported total levels of metabolites in the urine of humans exposed to safrole, which adequately matched. The model results revealed no dose-dependent shifts in safrole metabolism and no relative increase in bioactivation at dose levels up to 100mg/kg body weight/day. Species differences were mainly observed in the detoxification pathways of 1-hydroxysafrole, with the formation of 1-oxosafrole being a main detoxification pathway of 1-hydroxysafrole in humans but a minor pathway in rats, and glucuronidation of 1-hydroxysafrole being less important in humans than in rats. The formation of 1-sulfooxysafrole was predicted to vary 4- to 17-fold in the population (fold difference between the 95th and median, and 95th and 5th percentile, respectively), with the median being three to five times higher in human than in rat liver. Comparison of the PBBK results for safrole with those previously obtained for the related alkenylbenzenes estragole and methyleugenol revealed that differences in 1-sulfooxy metabolite formation are limited, being only twofold to fivefold.

A physiologically based biokinetic (PBBK) model for alkenylbenzene safrole in rats was developed using in vitro metabolic parameters determined using relevant tissue fractions. The performance of the model was evaluated by comparison of the predicted levels of 1,2-dihydroxy-4-allylbenzene and 1'-hydroxysafrole glucuronide to levels of these metabolites reported in the literature to be excreted in the urine of rats exposed to safrole and by comparison of the predicted amount of total urinary safrole metabolites to the reported levels of safrole metabolites in the urine of safrole exposed rats. These comparisons revealed that the predictions adequately match observed experimental values. Next, the model was used to predict the relative extent of bioactivation and detoxification of safrole at different oral doses. At low as well as high doses, P450 mediated oxidation of safrole mainly occurs in the liver in which 1,2-dihydroxy-4-allylbenzene was predicted to be the major P450 metabolite of safrole. A dose dependent shift in P450 mediated oxidation leading to a relative increase in bioactivation at high doses was not observed. Comparison of the results obtained for safrole with the results previously obtained with PBBK models for the related alkenylbenzenes estragole and methyleugenol revealed that the overall differences in bioactivation of the three alkenylbenzenes to their ultimate carcinogenic 1'-sulfooxy metabolites are limited. This is in line with the generally less than 4-fold difference in their level of DNA binding in in vitro and in vivo studies and their almost similar BMDL(10) values (lower confidence limit of the benchmark dose that gives 10% increase in tumor incidence over background level) obtained in in vivo carcinogenicity studies. It is concluded that in spite of differences in the rates of specific metabolic conversions, overall the levels of bioactivation of the three alkenylbenzenes are comparable which is in line with their comparable carcinogenic potential.

No abstract