ACVP, Atlanta, GA for 9 November 2014
Integrating pathology and toxicology with genetics using MAGIC
Mice
John E. French, Ph.D.
Principal, TOXGEN/TOXicoGENetics
Adjunct Professor, Center for Pharmacogenetics & Individualized Therapy,
Eshelman School of Pharmacy, UNC-Chapel Hill
Synopsis
Individual responses to environmental or drug exposure-related adverse outcome
and/or associated disease are likely based upon differences in exposure,
individual differences in toxicokinetics/dynamics, and/or genetic-epigenetic
variations that influence and modify outcomes revealed within the population of
individuals at risk. Presently, preclinical toxicology largely depends on
homozygous inbred strains or outbred stocks created from a limited number of
individual breeders that reduced their genetic and epigenetic variation and, thus,
provide a homogenous response to drugs or toxicants. This may confound study
interpretation of interindividual variation and across species extrapolation.
To link environmental exposure to benzene and susceptibility or resistance to
toxicity we performed a genome-wide association study (GWAS) using a new
population based mouse resource – the Diversity Outbred mice created from the
Collaborative Cross (CC), a population of multi-parental advanced generation
intercross (MAGIC) inbred lines1-4. Selected generations of breeders from the CC
AIRILS were used to also create the diversity outbred (DO) mice using a
randomized breeding protocols. Each DO mouse is genetically different from
every other DO mouse from generation to generation and represent a significant
degree of genetic diversity in the mouse genome that is equal or greater than
that of human populations 6,7. The CC AIRILs, from which the DO have been
created, are a critical tool for validating candidate genes identified in the QTLs
using molecular biology and reverse genetics approaches5,8-13. Also,
determination of a variable range of response and the genetic basis can aid in
improving the extrapolation of results from rodent models to human hazard
identification and risk assessment14.
To test for genome-wide associations based upon individual responses to
benzene-induced genotoxicity, we exposed these mice to benzene (0 (air), 1, 10
or 100 ppm; 75 mice/exposure group) by inhalation, 6 hr/day for 28 days (5
days/week) in two independent cohorts (300 mice each). Samples of pre- and
post-exposure peripheral blood or post-exposure to bone marrow were evaluated
Figure 1. The distribution of benzene induced chromosomal damage (micronucleated reticulocytes, MNRETS/1000RETS) in different cohorts of male DO mice in peripheral blood (A) and bone marrow (B). Each
data point represents the response for single DO mouse. The observed chromosomal damage is similar in
magnitude for each exposure group C. The concentration dependent increase in DNA damage (MNRET/1000RET; (x ± SEM) in bone marrow.
for micronucleus (MN) frequency using flow cytometry. CD71 expression in the
reticulocyte population was highly variable among individual mice pre- and postexposure. Micronuclei showed a significant trend in response to benzene
exposure (p < 0.0001) in both experiments.
Micronuclei in bone marrow were significantly higher in both the 10-ppm (p =
0.013) and the 100-ppm (p < 0.0001) groups than in controls (Figure 1).
Moreover, the MN response to benzene exposure was highly variable, with both
low and high responders within each exposure group of each cohort. Linkage
analysis and mapping of benzene-induced genotoxicity and hematotoxicity
revealed both resistance and susceptibility alleles associated with quantitative
trait loci on chromosomes 10, 2, and 19, respectively. The Chr 10 QTL is
associated with a copy number variant of a mouse phenol sulfotransferase gene
that may affect resistance to benzene toxicity. Humans have a similar phenol
sulfotransferase that shows copy number variation in the human population.
Additional studies are required for functional proof of the candidate gene
associated with resistance to benzene genotoxicity. Benchmark dose models
based upon susceptible subpopulation mean and estimated variance for
genotoxicity compared to resistant subpopulations increased the BMDL (95%
confidence level) approximately 600 fold compared to the default safety factor of
100 used in accounting for between species and within human population
variation. We conclude that this population-based model has the potential to
model human population exposures and provide experimental evidence for
benzene induced bone marrow toxicity under controlled conditions.
Determination of the variable range of response and the genetic and epigenetic
basis can aid in improving the extrapolation of results from rodent models to
human hazard identification and risk assessment
Figure 4. A benchmark dose model approach based on categorization of individual DO mice into
quartiles relative to their chromosomal damage response (low to high response for bone marrow MNRET/1000 RET). A point of departure for the benchmark concentration (BMC) and the benchmark
concentration 95% confidence lower bound (BMCL) for the most resistant and the most susceptible
quartiles is estimated. A. Haplotypes, means (x), standard deviations (SD), and numbers (n) of mice for
each quartile and exposure group. B. BMC and BMCL estimates derived for all quartiles combined and
selected combinations using 0, 1, and 10-ppm benzene exposure. Adding Q4 individuals to Q2 and Q3
increases the sample size and the interval is narrow and BMCL increases. C. Benchmark dose models
based on the individuals in the resistant quartile (Q1). D. Benchmark dose model based on the
individuals in the most susceptible quartile (Q4). An exponential model provided adequate fit to the 0,
1, and 10 ppm data and the BMC and BMCL were calculated for a 10% increase above the control
mean as a conservative measure. In the resistant quartile, BMC = 0.550 ppm and BMCL = 0.205 ppm.
An exponential model provided an adequate fit to the 0, 1, 10 ppm data. For a 10% increase above the
control mean, BMC = 0.257 and BMCL <0.001 ppm (≈1 ppb). Red = maximum likelihood estimates
fit of exponential model of continuous benchmark responses; Green = 95% confidence interval for
each benchmark response.
Based on the variation in interindividual differences in benzene metabolism and
genotoxicity, it is probable that the most susceptible genotypes are likely to show
a greater association with increased lymphohematopoietic cancer, etc. In
anticipation of the use of genetically diverse mouse models, the parent inbred
male and female mice were placed on study under NTP study specifications to
determine the genetic background differences in regard to survival, observed
tumor phenotypes, and prevalence. The variance observed in the inbred strains
are likely to demonstrate the range of phenotypes observed in the diversityoutbred population. However, the homozygous inbred mice phenotypes cannot
account for gene-gene interactions affecting epistasis, dominance, or recessive
characteristics in the randomly breeding population. Under the conditions of
these studies, we observed significance differences in strain and sex dependent
survival, tissue specific non-neoplastic and neoplastic phenotypes, and
prevalence. These studies provide a reference database of lesions and their
frequencies that may assist in study interpretation in the genetically diverse
outbred mice.
The results of this DO mice study describe a link between environmental
exposure and susceptibility-resistance to benzene induced hematotoxicity and
chromosomal damage. The CC MAGIC mice, from which the DO have been
created, are also a robust tool for linkage mapping as well as validating
candidate genes using molecular biology and reverse genetics approaches.
In summary, the GWAS described here demonstrates a link between
environmental exposure and susceptibility-resistance to benzene induced
hematotoxicity and genotoxicity through quantitative-trait analysis in a new
experimental population-based mouse model. These genetically diverse models
can aid to improve the extrapolation of predictive toxicity results from rodent
models to identify human hazards and improve risk assessment for
pharmaceuticals and environmental exposures.
References cited:
1. Aylor DL, Valdar W, Foulds-Mathes W, et al. Genetic analysis of complex traits in the
emerging Collaborative Cross. Genome Res 2011;21:1213-22.
2. Chesler EJ, Miller DR, Branstetter LR, et al. The Collaborative Cross at Oak Ridge
National Laboratory: developing a powerful resource for systems genetics. Mamm
Genome 2008;19:382-9.
3. Churchill GA, Airey DC, Allayee H, et al. The Collaborative Cross, a community
resource for the genetic analysis of complex traits. Nat Genet 2004;36:1133-7.
4. Threadgill DW, Miller DR, Churchill GA, de Villena FP. The collaborative cross: a
recombinant inbred mouse population for the systems genetic era. Ilar J 2011;52:2431.
5. Svenson KL, Gatti DM, Valdar W, et al. High-resolution genetic mapping using the
Mouse Diversity outbred population. Genetics 2012;190:437-47.
6. Frazer KA, Eskin E, Kang HM, et al. A sequence-based variation map of 8.27 million
SNPs in inbred mouse strains. Nature 2007;448:1050-3.
7. Kirby A, Kang HM, Wade CM, et al. Fine mapping in 94 inbred mouse strains using a
high-density haplotype resource. Genetics 2010;185:1081-95.
8. Gatti D, Maki A, Chesler EJ, et al. Genome-level analysis of genetic regulation of
liver gene expression networks. Hepatology 2007;46:548-57.
9. Harrill AH, Ross PK, Gatti DM, Threadgill DW, Rusyn I. Population-based discovery
of toxicogenomics biomarkers for hepatotoxicity using a laboratory strain diversity
panel. Toxicol Sci 2009;110:235-43.
10. Harrill AH, Watkins PB, Su S, et al. Mouse population-guided resequencing reveals
that variants in CD44 contribute to acetaminophen-induced liver injury in humans.
Genome Res 2009;19:1507-15.
11. Kelada SN, Aylor DL, Peck BC, et al. Genetic analysis of hematological parameters
in incipient lines of the collaborative cross. G3 (Bethesda) 2012;2:157-65.
12. Mathes WF, Aylor DL, Miller DR, et al. Architecture of energy balance traits in
emerging lines of the Collaborative Cross. Am J Physiol Endocrinol Metab
2011;300:E1124-34.
13. Philip VM, Sokoloff G, Ackert-Bicknell CL, et al. Genetic analysis in the Collaborative
Cross breeding population. Genome Res 2011;21:1223-38.
14. Rusyn I, Gatti DM, Wiltshire T, Kleeberger SR, Threadgill DW. Toxicogenetics:
population-based testing of drug and chemical safety in mouse models.
Pharmacogenomics 2010;11:1127-36.
.