Letter to “Chepelev et al. Establishing a quantitative framework for regulatory interpretation of genetic toxicity dose–response data: Margin of exposure case study of 48 compounds with both in vivo mutagenicity and carcinogenicity dose–response data”

To the Editor We read with great interest the recent article by Chepelev et al. (2023) in Environmental and Molecular Mutagenesis. The study expands upon previous important work investigating the relationship between carcinogenic and genotoxic potency of several carcinogens (Soeteman-Hernandez et al., 2016). We were surprised to find sodium dichromate (i.e., hexavalent chromium [Cr(VI)]) among the 48 chemicals investigated by Chepelev et al. despite our previous publication highlighting issues with the inclusion of Cr(VI) in SoetemanHernandez et al. (2016), including the choice of genotoxic endpoint and data indicating that the genotoxic potency of Cr(VI) is likely lower (i.e., higher benchmark dose) than the carcinogenic potency (Thompson et al., 2016). Cr(VI) is a site of contact carcinogen in rodents following oral exposure, inducing tumors in the oral cavity of rats and small intestine of mice (NTP, 2008). Transgenic rodent mutation assays in both rats and mice in target tissues are negative, as are micronucleus assays in the crypts of the mouse small intestine (Thompson et al., 2021). Multiple in vivo blood and bone marrow micronucleus studies in rodents are negative following oral exposure, with the notable exception of results in the am3-C57BL/6 strain of mice (NTP, 2007; Thompson et al., 2021), which SoetemanHernandez et al. (2016) and now Chepelev et al. use as a benchmark for in vivo genotoxic potency of Cr(VI). As discussed in Thompson et al. (2016), we were unable to find any micronucleus studies in am3-C57BL/6 mice aside from Cr(VI) thereby calling into question the validity of using results from this strain as a benchmark for the genotoxic potency of Cr(VI), especially when there are several negative in vivo micronucleus assays (Thompson et al., 2021). Notably, NTP (2007) reported only the combined results of two am3-C57BL/6 assays as positive, even though their website indicates that one study was negative. Soeteman-Hernandez et al. (2016) and Chepelev et al. appear to have only modeled the positive am3-C57BL/6 assay. Moreover, in vivo genotoxicity data published before 2016 and since indicate that the tumors in NTP (2008) are not the result of a genotoxic mode of action (Thompson et al., 2021), but rather sustained intestinal injury in mice (Bhat et al., 2020) and unknown mechanisms in rats. Chepelev et al. also highlight Cr(VI) as having margin of exposure (MOE) values ≤10,000; however, the exposure estimate the authors used was conservatively set to 0.1 mg/kg/day due to the supposed lack of exposure data. However, most chromium in biota is likely trivalent chromium due to the presence of reducing agents in gastric fluid and cells. Most chromium in groundwater sources is Cr(VI) (Seidel & Corwin, 2013) and US Environmental Protection Agency's (EPA) own environmental monitoring data indicate median and 95th percentile Cr(VI) levels of 0.001 and 0.003 ppm, respectively, resulting in daily exposures on the order of 3E-5 to 9E-5 mg/kg/day. Adjusting the dichromate benchmark dose lower confidence limit (BMDL) for intestinal tumors in mice listed in Table 1 of Chepelev et al. to Cr(VI) results in MOE values above 10,000, as does our previous analyses on the oral tumors in rats (Thompson et al., 2018). Notably, Health Canada (2016) lists estimated daily exposures to Cr(VI) as 0.065 μg/kg/day, which also results in MOEs >10,000. The MOE calculations for Cr(VI) based on genotoxicity by Chepelev et al. are also questionable based on the exposure estimate and the BMDL used for genotoxicity potency (see above). Notably, several regulatory bodies have developed threshold-based oral toxicity values and/or water standards for Cr(VI) using data described in the review articles cited herein (FSCJ, 2019; Health Canada, 2016; TCEQ, 2016; WHO, 2020). Thompson et al. (2016) previously raised issues related to the inclusion of Cr(VI) in analyses by Soeteman-Hernandez et al. (2016), and do so again regarding the important analyses in Chepelev et al. (2023).