Influence of low-dose radiation on leukaemia development in a genetically predisposed mouse model

Leukaemia development in the case of low-dose radiation exposure in a genetically predisposed mouse model

Ionising radiation is a known risk factor for the development of childhood leukaemia, but the significance of weak ionising radiation in the low-dose range remains unclear. From 2003 to 2007, the German Childhood Cancer Registry in Mainz carried out a study entitled “Childhood cancer in the vicinity of nuclear power plants” (the KiKK study) on behalf of the Federal Office for Radiation Protection and demonstrated a correlation between the proximity of residence to a nuclear power plant and the risk of developing leukaemia. However, this correlation cannot be explained by current scientific knowledge of the effects of radiation. For the purposes of radiation protection, it is necessary to investigate whether people with a genetic predisposition are subject to a particular risk with respect to ionising radiation.

In children who develop leukaemia, it is known that there is often a genetic predisposition and that just one more genetic mutation is likely to be sufficient for disease development. One of the most common genetic predispositions in children is the translocation ETV6‐RUNX1, which occurs while the child is still in the womb. The aim of this research project was to verify by experiment in the mouse model whether genetic factors affect a person’s sensitivity to the harmful influence of ionising radiation.

Objective

This project aimed to use the Sca1‐ETV6‐RUNX1 mouse model, which carries the most frequently described preleukaemic gene translocation in children, ETV6‐RUNX1, to test whether exposure to low-dose radiation can modify mutation events in cancer genes and therefore has an influence on leukaemia development. The development of leukaemia following exposure to radiation was to be observed and the developing leukaemias were to be characterised phenotypically.

Genetic alterations in leukaemias occurring in the mice were to be recorded following radiation exposure, compared with leukaemias in sham-exposed mice, and examined for radiation-induced mutation patterns. The results were to be compared with published, secondary genomic alterations in murine and human ETV6‐RUNX1‐positive pB‐ALL in order to detect genomic alterations as a result of low-dose radiation that are specifically associated with ETV6‐RUNX1 predisposition.

Methods

First of all, online literature searches (in the PubMed database) were conducted in relation to the current state of science, and the planned experimental procedure was reviewed.

This was followed by one-off (or sham) irradiation of genetically predisposed Sca1‐ETV6‐RUNX1 mice (in each case, n=30 per treatment) with defined doses down to the low-dose range (2 Gy, 0.5 Gy, 50 mGy). The occurrence of leukaemias was recorded in the animals up to an age of two years, and any tumours that occurred underwent phenotypic and genotypic analysis.

Implementation

Groups of 30 Sca1‐ETV6‐RUNX1 mice each were irradiated with one-off doses of 2 Gy, 0.5 Gy and 50 mGy respectively using a closed, calibrated gamma radiation source (Cs-137, Gammacell 1000 Elite, nominal activity 12.8 TBq) or sham irradiated as a control (0 Gy). The animals were regularly tested for the presence of leukaemia through blood controls up to an age of two years. Animals that showed signs of leukaemia were killed and underwent the usual necropsy procedures. Tissue infiltration, tumour cell count and tumour clonality in haematopoietic organs were analysed macroscopically, histologically, by flow cytometry, and using biomolecular techniques.

Tumour DNA from leukaemic tissue (bone marrow, lymph nodes or spleen) in the diseased mice was isolated and purified in order to carry out whole exome sequencing of leukaemic cells. DNA was extracted from the tip of the respective mouse’s tail as reference germline material. The exome libraries were prepared using the Agilent SureSelectXT Mouse All Exon Kit. Library sequencing was carried out on the NextSeq550 platform (Illumina).

For data analysis, Fastq data was initially generated using Bcl2Fastq 1.8.4 (Illumina). The sequencing data obtained from the mouse reference genome (GRCm38.71) was aligned using version 0.7.4 of BWA. Conversion steps were carried out with Samtools and were followed by the removal of PCR duplicates using Picard. GATK 2.4.9 was used for sequence comparison by means of a local procedure for small insertions/deletions (“indels”, < 50 bp) and for the calling of somatic single nucleotide variants (SNVs), as well as for annotation and recalibration. The identified variants were annotated using the Variant Effect Predictor (VEP) with the help of the Ensembl database (v70).

Scores were also added from the software tools SIFT and PolyPhen, which predict potential loss of function in the affected genes/proteins. Genes that are known to have a causal relationship with cancer were identified using the gene list of the Cancer Gene Consensus (CGC, COSMIC dataset). The exome data underwent standardised alignment in order to analyse the mutation signatures. Following several preprocessing steps, single nucleotide variants were determined for the pB-ALL samples from the irradiated Sca1‐ETV6‐RUNX1 mice as well as for control samples (infection-triggered pB‐ALL samples from unirradiated sca1‐ETV6‐RUNX1 mice) using Platypus version 0.8.1 and standard parameters. Mutation signature analyses were then carried out using version 1.6.169 of the R/Bioconductor package MutationalPatterns. The analyses were performed in COSMIC release v93 using version 3.2 of the mutational signatures.

Results

Using a Sca1‐ETV6‐RUNX1 mouse model, it has now been possible to investigate the effect of ionising radiation in the medium (0.5 and 2 Gy) and low (0.05 Gy) dose range. This mouse model carries the gene translocation ETV6‐RUNX1, a relatively common genetic alteration that is known to increase the risk of leukaemia in children as well as in mice. In the medium dose range, some 10% of 30 animals developed leukaemia. Precursor B cell acute lymphoblastic leukaemia (pB‐ALL) occurred in Sca1‐ETV6‐RUNX1 mice that were exposed to a single dose of at least 0.5 Gy using a gamma radiation source (Cs-137; 0.5 Gy, n=3/30; 2 Gy, n=4/30) at an age of four weeks. Exposure-related somatic mutations in these cases of pB‐ALL affected (1) hotspot regions in known cancer genes (Jak1, Jak3, Ptpn11, Kras), (2) genes that were also mutated in human ETV6‐RUNX1‐positive pB‐ALL (Atm, Sh2b3, Ptpn11, Kras), (3) ALL predisposition genes (Sh2B3, Ptpn11), and (4) other known cancer genes. The small number of tumours and somatic single nucleotide variants meant it was not possible to identify any specific radiation-induced mutation signatures.

In a group of 30 mice, no leukaemia was triggered in the low-dose range (0.05 Gy).

In the future, larger groups or mouse models with a higher rate of tumour formation could be used in conjunction with whole genome sequencing and complementary omics analyses to generate larger datasets and to provide a comprehensive picture of specific t(12;21)-associated secondary genomic alterations as a result of radiation exposure.