Diversity Outbred Mice Identify Exposure Thresholds and Genetic Factors That Influence Benzene-Induced Genotoxicity

Exposure to benzene, an airborne pollutant and carcinogen, is associated with multiple hematologic malignancies, including various leukemias, aplastic anemia, and myelodysplastic syndrome. Inbred and F1 hybrid mouse strains are commonly used animal models for exposure threshold studies. Because all individuals from the same inbred or F1 hybrid strain are virtually genetically identical (isogenic), using them in toxicology studies is generally regarded as beneficial, because they minimize within-group variance and reduce the numbers of animals needed per cohort. Isogenic mice, however, do not model the genetic diversity of the human population or the variability in their responses to potential toxicants. Moreover, selecting a mouse strain that responds uniquely to a toxicant can skew test results, leading to unreliable exposure threshold estimates. Due to these limitations of isogenic mouse models, there is increasing interest in performing toxicology studies using more genetically diverse mice.

The Diversity Outbred (J:DO, 009376) is a genetically heterogeneous stock that was developed from eight inbred founder strains: A/J, C57BL/6J, 129S1/SvImJ, NOD/ShiLtJ, NZO/HlLtJ, CAST/EiJ, PWK/PhJ, and WSB/EiJ. These strains were chosen for their genetic and lineage distinctiveness, and the genomes of all eight strains have been fully sequenced. Collectively, the strains contributed ~45 million single nucleotide polymorphisms (SNPs) and structural variants to the J:DO population - a level of genetic diversity that is roughly equivalent to the diversity seen in humans. Additionally, J:DO mice can be used for high resolution genetic mapping to discover genomic regions - and, ultimately, single genes -that underlie differing responses to toxicant exposure. A 2015 report in Environmental Health Perspectives (French et al. 2015) demonstrates that benzene exposure threshold estimates from J:DO mice are an order of magnitude lower than previous studies using isogenic mouse models, and identifies two sulfotransferases as candidate genes that govern resistance to benzene-induced genotoxicity.

J:DO mice show dose-dependent increases in chromosome damage after benzene exposure

French et al. exposed two cohorts of J:DO male mice to benzene via inhalation at 0, 1, 10, or 100 ppm by volume and then measured micronucleated reticulocytes (MN-RET) and erythrocytes (MN-ERC) as markers of chromosomal damage. Because the MN-RET and MN-ERC data are very similar, only the MN-RET data is reported. Prior to exposure, the mean number of MN-RET per 1,000 reticulocytes (MN-RET/1000) was 2.17, and no difference between the four exposure groups or between the two cohorts was observed. Following the 4-week benzene exposure, no change was observed in MN-RET/1000 means in the 0, 1, or 10 ppm exposure groups. However, the mean of the 100 ppm exposure group was significantly increased to 14.6 MN-RET/1000, - a 573% increase.

In bone marrow, the MN-RET/1000 means increased with increasing benzene. As before, there were no differences between the two cohorts. Modeling of the bone marrow MN-RET data using a benchmark concentration (BMC) method recommended by the U.S. Environmental Protection Agency’s Guidelines for Carcinogen Risk Assessment estimated a benzene exposure threshold necessary to increase the MN-RET/1000 mean by 10% over the 0 ppm group (BCML10) of 0.205 ppm. BMC modeling on MN-RET/1000 data from a previously published benzene inhalation study using F1 hybrid mice with the same study design estimated a BCML10 of 3.66 ppm - an order of magnitude higher than that calculated from the J:DO mice.

Sulfotransferases identified as candidate genes for resistance to chromosomal damage

SNP genotyping and linkage mapping of post-exposure MN-RET in the J:DO animals from the combined 100 ppm exposure groups (144 mice) identified a single significant locus on chromosome 10 (Chr. 10) that encompasses 21 genes. This locus accounts for 48.7% of the phenotypic variance. Lower MN-RET/1000 means associate with the CAST/EiJ allele, which acts in a dominant fashion. J:DO individuals with at least one CAST/EiJ allele on Chr. 10 are more resistant to benzene-induced chromosomal damage. Identifying CAST/EiJ specific polymorphisms in coding exons, examining genomic sequencing data, and finding CAST/EiJ-specific gene expression differences, together, reduced the list of candidate genes to two: Gm4794 (gene model 4794) and Sult3a1 (sulfotransferase 3a1). Both genes are more highly expressed in the livers of CAST/EiJ mice than the other J:DO founder strains. Gm4794 is a paralog of Sult3a1 (84% amino acid sequence identity), and contains a sulfotransferase domain and produces a protein product.

These results demonstrate the greater sensitivity of the J:DO stock compared to isogenic strains in toxicology studies and their utility for efficiently mapping and identifying genes underlying variation in toxicant responses. Such valuable information ultimately may lead to new methods and treatments to mitigate the detrimental effects of toxicant exposures. This work is an elegant example of how the J:DO can be used in translational biomedical research.