FG4592 br E mail addresses jenny aasa aces su se J
E-mail addresses: [email protected] (J. Aasa), [email protected] (M. Törnqvist).
benchmark dose at the lower bound (BMDL) is compared to human intake estimates (EFSA, 2005). Although uncertainties in the extra-polation of risks from high exposure doses in animals to lower exposure doses in humans are accounted for, the derived MOE only gives a rough estimate intended for risk management purposes. For compounds where dose-response data are insufficient for BMDL derivation, the carcinogenic potency index T25 is recommended (EFSA, 2005), which has previously been applied for glycidol (EFSA, 2016).
In the present study we evaluate a multiplicative (relative) risk model for its applicability to data from carcinogenicity studies of genotoxic compounds with data for glycidol. This is the risk model that is used for quantitative extrapolation of tumor induction by ionizing radiation (BEIR, 2006). The model has previously been further developed by our research group for application to genotoxic carcinogens, aiming to obtain more reliable cancer risk estimations of genotoxic compounds (Granath et al., 1999). According to this model, the cancer risk increment is pro-portional to the internal dose (in vivo dose) of the genotoxic FG4592 and the baseline (background) cancer incidence in the target tissue of the studied species. The internal dose of the genotoxic agent is defined as the area under the concentration-time curve (AUC, expressed in e.g. mMh), reflecting the pharmacokinetics of the compound. Background cancer incidence in different sites, measured from control animals or estimated in human populations, could be anticipated to reflect background mu-tations (partly from replication errors during stem cell divisions) and their interaction with growth-promotive factors (Granath et al., 1999). From the relative risk model, an estimate of the relative cancer risk coefficient (β) is obtained, expressed as risk per internal dose. This coefficient is assumed to be approximately independent of tumor site, sex, and species (Granath et al., 1999). The relative risk could also be expressed as the doubling dose (1/β) of the carcinogen, which is the dose that doubles the lifetime cumulative hazard; this implies that it approximately leads to a doubling of lifetime risk if the background tumor incidence is low. The applicability of the relative risk model to data from carcinogenicity studies of genotoxic compounds has so far been validated for a few compounds. In these evaluations the number of tumors in exposed animals predicted with the relative risk model was shown to correlate well with observed tumor incidence in responding sites, irrespectively of tumor site or species (Fred et al., 2008; Törnqvist et al., 2008; Granath et al., 1999). The model has also been successfully evaluated and validated with data from mutagenicity tests in vitro (expressed as relative mutagenic potency) for two of these compounds (Fred et al., 2008; Granath et al., 1999). Thus, the evaluations support that the cancer risk coefficient obtained with this model is a measure of the genotoxic potency of the studied compound. These evaluations strongly indicate that application of the relative risk model for cancer risk assessment of genotoxic chemicals also facilitates the use of in vitro genotoxicity data for assessment of the relative cancer risk coefficients for such chemicals.
In the present work we have further evaluated the relative risk model for its applicability to genotoxic chemicals using data from published carcinogenicity studies of glycidol in mice and rats (Irwin et al., 1996; NTP, 1990). To obtain accurate estimates of the risk coefficient, the AUCs per exposure dose of glycidol in mouse and rat are obtained from short-term exposures performed at similar conditions as in the carcinogenicity studies. Hemoglobin (Hb) adduct levels measured in the exposed animals are used for calculation of the AUC's. Finally, the risk coefficient is transferred to human exposures.
2. Materials and methods
2.1. In vivo dosimetry of glycidol
Male and female B6C3F1 mice (approximately 10 weeks) and Sprague Dawley rats (approximately 8 weeks) were obtained from Envigo, Venray (Netherlands). The animals were housed in a controlled facility with standard diet and tap water ad libitum. The ethical Food and Chemical Toxicology 128 (2019) 54–60
application was approved by the Ethical committee on animal experi-ments, Swedish Board of Agriculture, license number S7-15. Three animals per dose group were administered glycidol (96%; Acros Organics, Geel, Belgium; CAS 556-52-5; 74.08 g/mol) (10 mL/kg), dis-solved in water, at identical dose levels as in the published 2-year carcinogenicity studies (Irwin et al., 1996; NTP, 1990): 25 and 50 mg/ kg (mice) and 37.5 and 75 mg/kg (rats), by gavage once daily for five consecutive days. Blood collection from each animal was performed at one occasion in EDTA-coated tubes three days after the final dosing day (eight days after first exposure) by sampling from the tail vein (rats) or from the orbital plexus during anesthesia with isoflurane and oxygen (mice). The blood samples were stored at −20 °C until preparation and analysis of Hb adducts, used as a biomarker for internal dose.