Abstract
To assess the biological risks of low-dose/low-dose-rate long-time radiation exposure, we have been conducting the multigenerational accumulation of de-novo mutations. We have succeeded in detecting spontaneous and radiation-induced mutations in the control and exposed mice, respectively, by using the whole-genome sequencing (WGS) technology. We have already identified over 4000 de-novo mutations in a total of 42 offspring mice compared to the original common pair of mice. It took less than 2 years from the first mating of the original pair to the completion of WGS; thus, it is very quick and cost-effective compared to the specific-locus test which is one of the representative risk assessment methods. According to the number of detected mutations, the precision and statistical power to detect de-novo mutations are extremely higher than any conventional risk assessment methods in the mouse model. We have detected a significant increase in the 20 mGy day−1–induced base substitutions, which is the lowest dose rate ever detected in radiation-induced base substitutions. The mouse mutagenesis model based on the WGS technology, therefore, should provide a powerful tool to the risk assessment system, particularly on the low-dose/low-dose-rate long radiation exposures.
INTRODUCTION
It has been revealed since the beginning of the 20th century that high-dose acute radiation exposure of several Gy or more causes various biological effects, including tumorigenesis and mutation induction. On the other hand, the biological effects of low dose and, in particular, long-term exposure at low dose rates have long been unknown. Due to the lack of risk data for low-dose/low-dose-rate long-time exposures, the practical recommendation encompassing the radiation protection standard is based on the ‘As Low As Reasonably Achievable (ALARA)’ principle by adopting the linear non-threshold (LNT) model. The basic concept of the ALARA principle dates back to the 1950s and remains in the latest ICRP recommendation (ICRP, 2007). The LNT model must have been proposed as a null working hypothesis, to begin with, that should be tested. Nevertheless, for more than half a century, it has continued to be applied to assessing the effects of low-dose and low-dose-rate long-term exposure without being verified. We hereby show the efficacy of the previously proposed expanded TRIO analysis with whole-genome sequencing (WGS) (Gondo, 2022) as a new risk assessment system for radiation exposures to directly test the LNT model.
C57BL/6J AS AN IDEAL MODEL STRAIN
The mouse has many advantages as a model, having many genetic markers, mutant strains, and established tools with a small body size and a short life cycle enough to study genetics for more than 100 years since the rediscovery of Mendel’s law in 1900. Among hundreds of established inbred mouse strains, C57BL/6J has long been widely used as a standard strain not only for genetics but also for modelling human diseases (reviewed by Gondo, 2008, 2013) as well as mutagenesis studies (e.g. the Megamouse Project by Russell, 1951). The C57BL/6J genome was chosen as the second mammalian genome to be completely sequenced by the Mouse Genome Sequencing Consortium (2002), just after the completion of the first phase of the Human Genome Project (International Human Genome Sequencing Consortium, 2001). Thereby, human as well as mouse reference genomes are currently available. Unlike the human genome which is heterogeneous having many polymorphic loci, the inbred strain like C57BL/6J is mostly homozygous giving rise to a very uniform haploid status. Accordingly, the mouse reference genome sequences are ideal for detecting very rare de-novo mutations with high precision. We chose C57BL/6JJcl (substrain of C57BL/6J) for our mutagenesis study on the gamma-ray risk assessment adopting the expanded TRIO analysis (Gondo, 2022).
EXPANDED TRIO ANALYSIS
In fruit flies (Drosophila melanogaster), balancer chromosomes have been available to accumulate de-novo recessive lethal/deleterious mutations in each generation without being lost through recombination, meiosis, segregation, and fertilisation to detect rare mutations as summarised by Crow and Temin (1964). For instance, Muller-5 and In(2LR)SM1 balancer chromosomes cover roughly 20% and 40% of the fly genome, respectively. In the mouse, however, no such balancer chromosomes have been available. To effectively accumulate de-novo mutation in the mouse, Gondo (2022) proposed the expanded TRIO analysis. Briefly, one original Generation-0 (G0) pair of C57BL/6JJcl produces several G1 mouse pairs. One G1 pair was mated for several generations with no irradiations as a negative pedigree. Another G1 pair was independently mated in parallel with a defined gamma-ray exposure as an experimental pedigree. We have been conducting an expanded TRIO analysis at Tokai University (TOKAI) starting from two G1 pairs: one G1 pedigree has been exposed to ∼0.15 mGy day−1 with 22Na gamma ray along with another G1 pedigree as a negative control. We also set up another experiment at the Institute for Environmental Sciences (IES) on a larger scale. Starting from an independent G0 pair with four G1 pairs, we have kept mating each G1 pedigree with either 0.05, 1, and 20 mGy−1 or no exposure to 137Cs gamma ray.
RESULTS
The abovementioned expanded TRIO analysis amplifies the number of de-novo mutations by accumulating the rare event of de-novo mutations per generation. The larger the number of generations, the greater the difference between the exposed and control groups. The accumulated de-novo mutations may be analysed in any generation after G2 (Gondo, 2022). Therefore, we preliminarily attempted the WGS analysis and mutation detection to confirm that the experimental system is proceeding as planned, while still in the process of irradiation mating for successive generations. The extracted genomic DNA samples from descendant mice in the exposed pedigree(s) as well as in the corresponding control pedigree were subjected to the WGS together with the genomic DNAs from the G0 pair. Since all the mice in each set of the experiment, all the descendant mice both in the exposed and control pedigrees origin from the same G0 pair, we may subject the G0 parental genome only once, making the analysis quick and cost-effective. Uchimura et al. (2015)’s bioinformatics pipeline effectively identifies de-novo mutations from the WGS dataset, covering ∼80% of the mouse genome. We extracted G4 mice from TOKAI and IES pedigrees except for 20 mGy day−1 (IES1) and succeeded in identifying the de-novo mutations. For the precise and quick evaluation of the detection system, we primarily focused on autosomal base substitutions in the preliminary mutation detection. In the TOKAI experiment, 24.6 (SE = 1.20) and 22.5 (SE = 0.60) base substitutions mice−1 generation−1 were detected with 0 and 0.15 mGy day−1 exposures, respectively. In the IES1 experiment, we have identified 25.0 (SE = 0.86), 24.8 (SE = 0.71), and 25.0 (SE = 0.86) base substitutions mouse−1 generation−1 with 0, 0.05, and 1 mGy day−1 exposures, respectively. No statistically significant differences between exposed and control pedigrees were identified in either TOKAI or IES1 experiments. Contrarily, the 20 mGy day−1 exposure at IES showed a direct radiation effect on the G2 females that were found to be infertile. We, therefore, mated the G2 males from 20 mGy day−1 exposed pedigree to the unexposed G2 females. The G3 mice from the alternative G2 mating together with unexposed G3 mice were then subjected to WGS. In these G3 mice (IES2), we have found 24.6 (SE = 0.78) and 32.9 (SE = 1.91) base substitutions mouse−1 generation−1 with 0 and 20 mGy day−1 exposures, respectively. The detected number of base substitutions in the 20 mGy day−1 exposed mouse) was significantly increased to that in the unexposed control mice (p < 0.005), which is the lowest dose rate ever detected for radiation-induced germline base substitutions. In this study, we have detected a total of 4058 de-novo base substitutions encompassing X chromosomes from 42 mice in less than 2 years.
DISCUSSION
The expanded TRIO analysis with WGS demonstrated that the method is efficient in assessing the radiation effects on the next generation, particularly of the low-dose-rate long-time exposure. Despite the mouse experiment being conducted at two completely independent research institutions (TOKAI and IES), highly comparable results were obtained from the negative controls, proving the robustness and high reproducibility of the assessment system. It now becomes feasible and realistic to directly conduct experimental studies and discuss the low-dose-rate long-time exposure effects on the next generation based on large-scale datasets under various conditions, e.g. female vs male, age effect, diet, and other environmental factors, by setting appropriate negative controls. The key to further fruitful elucidation is the multidisciplinary challenges encompassing radiobiology, genetics, genomics, molecular biology, biobanking, statistics, mathematical modelling, informatics, databases, AI-assisted data mining, WEB-based information and data sharing, and so on. Toki et al. (2023) and Bando et al. (2023) have also been conducting further statistical analyses, mathematical modelling, and comparisons to the historical SLT analysis by Russell and Kelly (1982). Further analysis is in progress including the detection of other types of mutations, e.g. insertions/deletions, inversions, translocations, and copy-number variations. The mutational signature analysis (Alexandrov et al., 2020) is another issue to be done. Unlike the Megamouse Project by Russell and his colleagues, the expanded TRIO and WGS analysis is much quicker and smaller so that it is feasible to re-examine and/or validate the experimental results. The WGS data per se may be easily re-examined by anyone, anytime, and anywhere because all the WGS data will be open and freely available to the public via the Sequence Read Archive (SRA: https://www.ddbj.nig.ac.jp/dra/index-e.html). Most of the mouse bodies and extracted genomic DNA samples have also been archived in deep freezers, and the biobank may also be internationally open to the public if the research communities need it with the necessary funds. Most importantly, studies must be expanded to examine how the mouse model and data apply to humans. Intriguingly, Yeager et al. (2021) did not detect transgenerational effects of radiation exposure in humans despite the large-scale TRIO analysis with WGS in the Chernobyl cohort, concluding the necessity of further human studies. Altogether, it now seems feasible to test the LNT model with concrete experimental data to prime evidence-based scientific discussions. Not only science communities but also the public are expecting to see the next ICRP recommendation if the ALARA principle be rewritten at last after 50 years and more.
Footnotes
ACKNOWLEDGEMENTS
The authors appreciate the members of the 195th committee ‘Utilization and Biological Effects of Radiation’ of the University-Industry Research Cooperation Society Applied Scientific Linkage and Collaboration organised by the Japan Society for the Promotion of Science for their discussions, support, and encouragement to conduct this research.
FUNDING
This work was partly supported by Research Project on the Health Effects of Radiation organised by the Ministry of the Environment, Japan; Ministry of Education, Culture, Sports, Science, and Technology (KAKENHI 221S0002 to Y.G.); and the Japanese Society for the Promotion of Science (KAKENHI 16H06279 to Y.G., 17H00789 to Y.G., and 21K19842 to Y.M.). A part of this work was performed by using the facilities of the Medical Science College Office, Tokai University.