Frequency of acute myeloid leukaemia-associated mouse chromosome 2 deletions in X-ray exposed immature haematopoietic progenitors and stem cells.

Exposure to ionising radiation can lead to an increased risk of cancer, particularly leukaemia. In radiation-induced acute myeloid leukaemia (rAML), a partial hemizygous deletion of mouse chromosome 2 is a common feature in several susceptible strains. The deletion is an early event detectable 24h after exposure in bone marrow cells using cytogenetic techniques. Expanding clones of bone marrow cells with chromosome 2 deletions can be detected less than a year after exposure to ionising radiation in around half of the irradiated mice. Ultimately, 15-25% of exposed animals develop AML. It is generally assumed that leukaemia originates in an early progenitor cell or haematopoietic stem cell, but it is unknown whether the original chromosome damage occurs at a similar frequency in committed progenitors and stem cells. In this study, we monitored the frequency of chromosome 2 deletions in immature bone marrow cells (Lin(-)) and haematopoietic stem cells/multipotent progenitor cells (LSK) by several techniques, fluorescent in situ hybridisation (FISH) and through use of a reporter gene model, flow cytometry and colony forming units in spleen (CFU-S) following ex vivo or in vivo exposure. We showed that partial chromosome 2 deletions are present in the LSK subpopulation, but cannot be detected in Lin(-) cells and CFU-S12 cells. Furthermore, we transplanted irradiated Lin(-) or LSK cells into host animals to determine whether specific irradiated cell populations acquire an increased proliferative advantage compared to unirradiated cells. Interestingly, the irradiated LSK subpopulation containing cells carrying chromosome 2 deletions does not appear to repopulate as well as the unirradiated population, suggesting that the chromosomal deletion does not provide an advantage for growth and in vivo repopulation, at least at early stages following occurrence.

Introduction

Exposure to ionising radiation (IR) occurs from a mixture of natural and occupational sources [1], medical treatment [2] as well as accidental exposure [3]. Epidemiological studies, including the cohort studies from survivors from the Hiroshima and Nagasaki atomic bombings, show exposure leads to an increased risk of cancer [4,5].

Acute myeloid leukaemia (AML) is one of the most common malignancies seen to occur in human populations exposed to IR [2]. A dose-dependent increase in AML incidence has been documented in the atomic bomb blast survivors [5] and it is known to occur following radiotherapy [6]. The crucial initial event for leukaemia is believed to be DNA damage with DNA double-strand breaks potentially leading to chromosomal aberrations such as deletions or fusion genes characteristic of specific leukaemias when misrepaired [7]. Through the loss of genetic material due to chromosome deletions, for example, loss-of-function mutations can occur and increase the likelihood of progression to cancer [8]. Even though tumours are generally thought to evolve from a mutation in a single cell, the “cell of origin” and the cell type sustaining the tumour, “cancer stem cells (CSC)” may not be the same [9,10]. Defining both the cell of origin and the CSC in tumours will provide crucial information to understand the molecular mechanisms behind individual cancer types, or indeed within cancer types. Furthermore, more accurate risk estimates can be obtained if the chromosomal aberrations in the cells are early biomarkers of late effects such as cancers.

Leukaemia is the prevailing neoplastic disorder of the haematopoietic system. It is the consequence of an acquired modification in the genome of a target cell, most probably a stem cell of the bone marrow compartment following exposure to a carcinogen, such as ionising radiation (IR). Nevertheless, leukaemia may also arise from more committed progenitors caused by mutations and/or selective expression of genes that enhance their self-renewal capabilities [11]. For example, it was demonstrated that leukaemia stem cells can be generated from committed progenitors by expressing a fusion protein encoded by a specific translocation [12].

In a mouse model of radiation-induced AML (rAML), an early event detectable 24 h after exposure is an interstitial deletion of one copy of chromosome 2 (del2) [13-15]. At 12–15 months after exposure, 50% of mice carry an expanding chromosome 2-aberrant clone [14] and eventually up to 25% of the animals develop rAML [16]. Approximately 80–90% of cases carry the partial del2 [13,17,18] and a small number of mice also carry Flt3-internal tandem duplications a common mutation in human leukaemias [19,20] as was recently reported [21]. Molecular mapping of leukaemic cell genomes in mice with rAML identified a minimal deleted region on chromosome 2, in which a potential tumour suppressor gene, Sfpi1 was located [22,23]. The gene encodes the transcription factor PU.1, an essential and important transcription factor in haematopoiesis [24,25]. Further work has shown that PU.1 acts as a tumour suppressor in haematopoietic and myeloid cell development [26,27]. In approximately 70% of cases of rAML, the remaining copy of the Sfpi1 gene has a point mutation in the DNA sequence coding for the DNA-binding domain of the protein [28-30]. It is currently not known at what time after irradiation this point mutation occurs. Del2 and/or point mutation of PU.1 in human AML has been less frequently reported [31,32], although heterozygous point mutations in Sfpi1/PU.1 have been found in 7% of cases. In these cases, it was found that the mutation leads to a reduced or aberrant function of the transcription factor [33]. PU.1 has further been shown to be of importance in certain types of AML [34,35].

Here we have conducted experiments to quantify del2 after X-irradiation using genetically modified CBA/H mice, that express green fluorescent protein (GFP) as a marker for Sfpi1 expression (Olme et al. manuscript submitted) [36,37]. Immature bone marrow cells (Lin−) and haematopoietic stem cells/multipotent progenitors (LSK and CFU-S12) were investigated for the frequency of Sfpi1 loss at several early (7 days) and late (10 months) time points following ex vivo or in vivo exposure as well as assessing their ability to repopulate the bone marrow compartment with a view to have a better definition of the cell of origin of mouse rAML.

Materials and methods

Mice

C57BL/6 GFP expressing mice from Nutt et al. [37] and Dakic et al. [36] were re-derived onto an AML-sensitive CBA/H background at MRC (Harwell, Oxon, UK) for at least 10 generations. The mice have an IRES-GFP cassette inserted in the 3′ untranslated region of the Sfpi1 gene. GFP is under the control of the Sfpi1 promoter, where transcription produces a bicistronic mRNA resulting in PU.1 protein and GFP. Both CBA/H Sfpi1/GFP homozygous (Sfpi1) and CBA/H Sfpi1/GFP heterozygous (Sfpi1) animals were utilised in this study along with wild-type CBA/H mice as flow cytometry controls or host animals for repopulation and colony forming unit-spleen assays. All animals were bred and handled according to UK Home Office Animals (Scientific Procedures) Act 1986 and with guidance from the local ethical review committee on animal experiments. Details of the original GFP construct can be found in [36,37].

Irradiation of cells and mice

Cells were irradiated in vitro, either immediately post-isolation as total bone marrow cells (BMC) or as LSK or Lin− cells following 24 h culture in Stemspan (Stem Cell Technologies), with 250 kVp X rays at a dose rate of 0.5 Gy/min, 13 Amp (AGO, Reading UK).

Mice were whole body irradiated with 250 kVp X rays at a dose rate of 0.887 Gy/min, 11 Amp (MRC Radiation and Genome Stability Unit, Harwell, Oxon UK) at 3 Gy for harvest of in vivo irradiated donor cells, or 8.5 Gy to ablate host animals for CFU-S12 assay.

Immunomagnetic cell separation and cell sorting

To obtain Lin− or Lin− Sca-1+ c-Kit+ (LSK) cells, BMC were flushed from femora and tibias of 8 donor mice either exposed to 3 Gy X-rays 7–9 days beforehand or from unirradiated controls. Lin− cells were selected using the Mouse Hematopoietic Progenitor Enrichment Kit (Stem Cell Technologies, Grenoble) according to the manufacturer's instruction. Lin− cells were further sorted into LSK by staining with the following antibodies: c-Kit-PE/Allophycocyanin (APC) (BD Bioscience, Oxford) and Sca-1-PE (Biolegend, CA, USA). Unstained and single stained samples were included as controls. Samples were incubated in PBS/3% FBS for 45–60 min on ice. Following 2× washes in PBS/1%FBS 7-aminoactinomycin D (7-AAD) was added (0.25 μg/sample) to eliminate dead cells on sorting (MoFlo cell sorter (Dako Cytomation, Denmark) at Jenner Institute, Oxford). Cells were gated as 7-AAD− Sca-1+ c-Kit+ (Fig. 3B) and sorted into Stemspan Serum-free expansion media (SFEM) media (Stem Cell Technologies) on ice. These cells were used in repopulation assays and suspension cultures.

Fig. 3

(A) BMC were isolated from 8 male CBA Sfpi1 mice exposed to a total body dose of 3 Gy X-rays 7–9 days previously. The cells were pooled, Lin− cells were selected as described and stained with antibodies for Sca-1 and c-Kit. The cells were sorted using a MoFlow cell sorter. Sorted cells were then either cultured in suspension cultures for 9 days, then chromosome preparations were made and analysed by BAC-FISH. Alternatively, cells were injected directly after sorting into irradiated host animals. (B) Sorting gates for isolation of LSK cells. A region was set around the major population to exclude debris and dead cells were excluded by gating on the 7-AAD negative population. A further gate was set on the Sca-1+ c-Kit+ cells. (C) BMC were isolated from unirradiated mice. After overnight culture in Stemspan media, the cultures were exposed to a 3 Gy X-rays dose ex vivo. Cells were grown in suspension for 9 days and then prepared and analysed as in Fig. 3A for chromosome 2 deletions.

Suspension cultures for expansion of Lin− Sca-1+ c-Kit+ (LSK) or lineage depleted (Lin−) cells

300 (control) and 3000 (irradiated) LSK (Fig. 3A) or Lin− (Fig. 1) cells were cultured in 35 mm Petri dishes in 2 mL SFEM (Stem Cell technologies) with the addition of recombinant murine stem cell factor 50 ng/mL (rmSCF), 100 ng/mL recombinant human interleukin-11 (rhIL-11) and 100 ng/mL recombinant human Flt3 Ligand (rhFlt3L) (all cytokines from Stem Cell Technologies), 40 μg/mL low density lipoprotein (LDL) (Sigma, UK), 100 U/mL Penicillin (Fisher Scientific, UK) and 100 μg/mL Streptomycin (Fisher Scientific). Growth was monitored by visually checking the cells under an EVOSXL microscope (AMG, USA). If the cell density was higher than 5 × 105 cells/mL, the culture was diluted 1:2 in fresh media. After 7–9 days, cells were harvested and the culture dish rinsed with Iscove's modified Dulbecco's medium (IMDM). A cell scraper was used to gently remove any cells that had adhered to the culture dish.

Fig. 1

Bone marrow cells were isolated from Sfpi1 mice and Lineage depleted. Lin− cells were put into suspension cultures, irradiated in vitro with a 2 Gy X-rays dose after overnight culture and then cultured for a total of 7 days. Cultures were harvested, chromosome preparations made and analysed by BAC-FISH for deletions on chromosome 2.

Repopulation assay

200–300 LSK cells from Sfpi1 mice together with 2 × 105 CBA wild-type bone marrow “helper cells” were resuspended in a volume of 120–200 μL of IMDM (Life Technologies, UK) and injected via the tail vein into 8.5 Gy ablated wild type (non-GFP) CBA/H mice (Fig. 3A).

The analysis of repopulation was performed using flow cytometry. Blood samples from the tail vein of host mice were acquired and the red blood cells removed with lysis buffer (20.5 g NH4 Cl, 2.5 g NaH CO3, 0.093 g EDTA/litre ddH20) in a 1:20 dilution and incubating for 5 min at room temperature. Samples were washed once after lysis with a small amount of PBS/1% FBS and then resuspended in PBS. All samples were analysed by flow cytometry using the FACS Calibur (BD Biosciences). Positive repopulation was defined as over 1% GFP expression in a sample.

Colony forming unit – spleen day 12 (CFU-S 12) assay

BMC (2 × 106 for unirradiated control and in vivo irradiated donor cells, and 2 × 107 for in vitro irradiated cells) were injected into the tail vein of host wild type CBA/H mice radioablated with 8.5 Gy X rays (Fig. 2A and B). On day 12 after injection, the mice were euthanised with a rising concentration of CO2 and the spleens removed. Visible, unfixed and unstained colonies were counted and dissected using a scalpel under a dissecting microscope. Colonies were added to 1.5 mL microfuge tubes containing PBS, disaggregated using bow spring scissors and passed through a 19 G then 21 G needle. Cells were then washed once in PBS/1%FBS and then resuspended in PBS for further analysis by flow cytometry and FISH (Fig. 2C and D).

Fig. 2

Analysis of chromosome 2 deletions using the CFU-S12 assay. (A) CBA Sfpi1 or Sfpi1 mice were irradiated with a whole body dose of 3 Gy X-rays. After 7 days or 10 months, total bone marrow was isolated and the cells were injected into ablated hosts. After 12 days, the mice were sacrificed and individual spleen colonies dissected. The colonies were disaggregated and individual cell suspensions from each colony were analysed for their GFP expression by flow cytometry and by BAC-FISH. (B) BMC was isolated after 7 days as in (A) from unirradiated mice and the cells irradiated with a 3 Gy X-rays ex vivo. The cells were then injected into ablated hosts and the CFU-S12 assay performed and analysed as previously described. (C) A gate was drawn around the major cell population on a forward scatter side scatter plot and the GFP expression analysed in each dissected colony. The GFP fluorescence of control colony is shaded in grey and a representative colony from an Sfpi1 mouse is shown in white. (D) Representative BAC-FISH image of a metaphase spread (centre) and interphase cell (top right) from a CFU-S12 colony showing no deletion (in both cells Sfpi1/PU.1 is retained). Sfpi1/PU.1 is labelled with Spectrum-orange (in red) and the two chromosome 2 specific markers by Spectrum green (in green).

Chromosome preparations and fluorescent in situ hybridisation (FISH)

Chromosome preparations from spleen colony cells and Lin− cells were produced as previously described [15] except culture conditions for Lin− cells were as follows: cells were incubated overnight at 37 °C/5% CO2 in 1–2 mL Stemspan SFEM (1.5–2 × 106 cells/mL) with 0.6 μL/mL Colcemid (Invitrogen, UK) added and then fixed as for spleen. Bacterial artificial chromosome-FISH (BAC-FISH) was carried out as previously described [15].

Statistical analysis

Z test for proportions was performed to analyse the data shown in Tables 2 and 4.

Results

Detection of chromosome 2 deletion in lineage depleted cells

The initiating event for murine Acute Myeloid Leukaemia is believed to occur within haematopoietic stem cells (HSC) or in a more committed myeloid progenitor cell [10,11]. Since HSC and progenitors are rare cell types [38,39] we started by looking at the Lin− population and the frequency of del2 was assessed by BAC-FISH (Fig. 1). No del2 were observed in the samples analysed and del2 has to be a rare event (less than 1/100, Table 1), i.e. below the detectable sensitivity of the assay. The upper limit of the 95% confidence interval (CI) at 0.0444 furthermore indicates that the true value of del2 in the Lin-subpopulation is low.

Table 1

Chromosome 2 interstitial deletions in Lin− cells, sham exposed (control) or exposed to a 2 Gy X-rays dose (irradiated), detected by BAC-FISH.

	Cells counted	Retained	Deleted	PU.1 loss (%)	Upper limit of the 95% confidence interval
Lin− control	100	100	0	0
Lin− irradiated (in vitro)	100	100	0	0	0.0444

Assessment of del2 in progenitor subpopulations using in vivo clonogenic CFU-S12 assay

The CFU-S12 assay was used to assess whether del2 could be detected in an in vivo, short-term clonal repopulation assay (Fig. 2A) [40-42]. BMC were isolated from CBA Sfpi1 and Sfpi1 mice exposed to an optimal rAML-inducing dose of 3 Gy X rays 7 days or 10 months previously. Alternatively, BMC were taken from unirradiated mice and exposed to 3 Gy X rays ex vivo (Fig. 2B). The irradiated cells were then injected into ablated hosts. 12 days after transplant, the mice were euthanised, and the individual spleen colonies dissected out. The frequency of del2 was assessed using flow cytometry (Fig. 2C) by analysing the GFP expression levels in individual colonies (Olme et al., manuscript submitted) and by BAC-FISH (Fig. 2D). In the colonies from mice exposed 7 days earlier or BMC exposed ex vivo, no deletions were found (0/54) and similarly no deletions were observed in colonies from mice exposed 10 months (more accurately 40 weeks) earlier (0/17) (Table 2). For these results, the upper CI was 0.0794 and 0.2163 for the 7 days and 10 months time points respectively.

Table 2

Chromosome 2 interstitial deletions in CFUS-12 (sham exposed (control) or exposed to a 3 Gy X-rays dose (irradiated)) detected by BAC-FISH and/or flow cytometry.

	Flow cytometry GFPlow/negative	PU.1 loss (%)	Upper limit of the 95% confidence interval
Control	0/10	0
Irradiated (5 in vivo, 49 in vitro) (7 days before)	0/54	0	0.0794
Irradiated in vivo (10 months before)	0/17 (of which 15 were confirmed by FISHa)	0	0.2163

FISH were performed by scoring interphases due to metaphases being rare in the chromosome preparations.

Detection of del2 in LSK cells

The Lin-, Sca-1+, c-Kit+ (LSK) cell population is enriched for HSC, approximately 10% are HSC, the remainder being progenitor cells [43,44]. To analyse this cell population we used two different protocols. In the first protocol cells were derived from mice exposed to 3 Gy X rays in vivo 7–9 days before. BMC were isolated and Lin− cells were selected (Fig. 3A). The Lin− cells were then stained with fluorescently tagged antibodies against the surface markers Sca-1 and c-Kit, and sorted using flow cytometry to acquire the LSK population (Fig. 3B). LSK cells were cultured in suspension for 9 days. In the second protocol, the LSK population was acquired from unirradiated mice and the cells exposed ex vivo after overnight culture in suspension (Fig. 3C). After culturing cells in suspension for 9 days, chromosome preparations were made and the frequency of del2 determined by BAC-FISH. The frequency of del2 was found to be approximately 6% in LSK cells exposed either in vivo or ex vivo (Table 3). This is significantly elevated compared to unirradiated LSK cells (p = 0.0064 and 0.0068 respectively).

Table 3

Chromosome 2 interstitial deletions in LSK cells sham exposed (control) or exposed to a 3 Gy X-rays dose (irradiated) detected by BAC-FISH.

	Metaphase scored	Retained	Deleted	PU.1 loss (%)
LSK control	67	66	1	1.5
LSK irradiated (in vivo)	100	94	6	6**
LSK irradiated (in vitro)	32	30	2	6.2***

p = 0.0064.

p = 0.0068 for differences between irradiated and control populations.

Growth of irradiated and unirradiated LSK cells in vivo assessed using bone marrow repopulation assay

To further assess the effect that del2 and exposure to ionising radiation have on haematopoietic progenitor populations, we isolated LSK cells from Sfpi1 mice exposed to 3 Gy X rays 7–9 days before, transplanted the cells into ablated wild type CBA host mice (Fig. 3A) and monitored the repopulation ability of the injected cells. Approximately 4 months after the injection of LSK cells, blood samples were taken from the tail vein and GFP expression within peripheral white blood cells assessed by flow cytometry. Only cells derived from the donor transplant express GFP making it possible to determine the frequency of reconstitution of control unirradiated cells or cells exposed to X-rays and to compare these frequencies of reconstitution. Results showed that irradiated LSK cells from donor animals (Sfpi1) had a significantly reduced capability to repopulate host animals (non-GFP) (1/11 mice, 9%), compared to unirradiated control LSK cells (3/5 mice, 60%) (Table 4). Mice were considered to have some level of donor cell repopulation if 1% of the cells in peripheral blood samples were GFP+ (of donor origin).

Table 4

Long-term reconstitution of host animals injected with control (unirradiated) or irradiated (3 Gy X-rays dose) donor LSK cells.

	Number of mice reconstituted	%
LSK control	3/5	60
LSK irradiated (in vivo)	1/11*	9

p = 0.014 for difference between irradiated and unirradiated controls.

Discussion

The importance of the interstitial deletion of mouse chromosome 2 in rAML is demonstrated by its frequent occurrence in this disease model [13,17,18] and by its early occurrence after exposure to ionising radiation [14,15]. While recent results have suggested a leukaemic stem cell in the model [45], the level to which del2 occurs in early progenitor cells and haematopoietic stem cells is not yet known. Defining the cell of origin for leukaemia is important to understanding disease mechanisms and identifying potential therapeutic interventions [10]. Even though the initial event is detectable at an early stage, leukaemic progression is a step-wise process, evidenced by the heterogeneity of the disease [46]. Recent work has shown that the cell of origin in leukaemia defines the gene expression, epigenetic state and drug response of leukaemic cells [47] and furthermore that the initial mutation in an AML may co-operate with random mutations already existing before the initiating event, although only specific mutations may be able to drive leukaemogenesis in this way. [48].

From our results, it seems that del2 is rare in Lin −. Previous work from our group found that in BMC the frequency of the deletion in vivo is approximately 15% at 7 days after exposure to ionising radiation and then approximately 8% after a month [15]. The cell type in which the deletion is present cannot be determined from these studies since they were derived from FISH studies on unfractionated cells. In light of this, the <1/100 frequency of deletions in the Lin − fraction could appear to be unexpectedly low. It is possible that subsequent culturing for 7 days inadvertently selects for cells that are not damaged. One solution might be to try and perform FISH on Lin− without culturing, although this would be technically challenging due to the reduced number of cells available for FISH analysis.

The subpopulations of cells giving rise to the colonies in the CFU-S12 assay are mainly multipotent progenitors (MPP) and HSC compared to CFU at day eight (CFU-S8) were oligo-lineage progenitors (MEP/CMP) giving rise to most colonies [49-51]. In our hands, the deletion is also rare (0/54) in the cell type or types giving rise to the spleen colonies at day 12 (i.e. HSC/MPP). The frequency of Sfpi1 loss was assessed at 7 days post X-rays exposure, as this is the peak incidence of del2 in BMC [15] and thus being theoretically optimum for the detection of del2 events in spleen colonies. We furthermore investigated the frequency of the deletion 10 months after radiation exposure and similarly found no deletions. The selection of this time point was based upon a number of factors. Experiments carried out in our laboratory have shown that AML presentation starts from 38 weeks following exposure to 3 Gy X rays (Olme et al., submitted). Moreover, experiments by Riches [52] showed an increase in AML induction in recipient CBA/H mice transplanted with donor cells irradiated with 3 Gy X-rays 10 months beforehand. Additionally, we expected to observe an increase in the number of MPP carrying a del2 at this much later time point as it has been shown that chromosome 2 aberrant clones are selected for entry into the proliferating bone marrow cell compartment and that these clones in general have a higher proliferative potential (14). Our data (Table 2) did not confirm our hypothesis, which may suggest that these cells have nevertheless an impairment in the ability to form a spleen colony after being re-introduced in the blood system. This is in line with previous published data where no spleen colonies out of 82 had the deletion in CFU-S at day 11 [53].

Although we have not specifically isolated HSC with long-term capacity for reconstitution of the haematopoietic system, they are present amongst the cells injected for CFU-S12 and the LSK population (they represent roughly 5–10% of LSK and 50% of CFU-S12 population (11)). These cells have been shown to contain lower levels of reactive oxygen species than their more mature progeny [54] and are more resistant to IR-induced cell killing [55,56]. They might be less prone to del2 although this would be difficult to confirm with this model.

However, our results also show that del2 can be readily induced in LSK cells and del2 are also detectable in unirradiated cells at a low percentage (15). This implies that del2 does not give significant survival/growth disadvantage to LSK cells at this early time point after exposure to IR but may provide them specifically with a selective advantage for proliferation at a later stage.

It suggests that the cell of origin may be of this phenotype rather than a more defined myeloid progenitor in the present model. The statistical certainty of the value of the zero incidence of del2 in the CFU-S12 population is not as strong as for the Lin− population. That is, should more colonies have been analysed, del2 may have been detected in the CFU-S12 subpopulation. However, del2 appears to be more rare in the CFU-S population than the LSK population, making the latter more likely to contain the cell of origin.

Poor haematopoietic reconstitution by LSK cells exposed to IR has previously been reported [57]. Furthermore, it has been reported that unirradiated bone marrow cells can outcompete irradiated cells when they are in direct competition, even if the irradiated cells are at a ratio of 9:1 [58] Our results also show that irradiated cells have a reduced “fitness”, which has been suggested to be caused by reactive oxygen species [58,59]. The number of LSK cells injected into the mice (200–300) would contain a theoretical minimum of 4 and a maximum of 20 cells capable of at least partial reconstitution of host animals, since previously published data show that about 1 out of 15–50 LSK cells are capable of reconstitution [60,61]. It is possible that a few of the injected LSK sorted from irradiated mouse bone marrow carried a del2 (Table 4). However, our current data does not show the frequency of del2 in long-term reconstituting HSCs and to assess this, further experiments would need to be performed. Although in low number, they could have actively participated to the repopulation; nevertheless this was not the case, which suggests that the general reduced fitness is not overcome by the early chromosome interstitial deletion event and that the cells containing the deletion, at this early stage, do not seem to have a proliferative advantage. This furthermore indicates that other factors, such as reduced fitness, are important in leukaemic development and has to be taken into account when studying the progression of the disease. Co-operation with other mutations after irradiation exposure may also be important [58] as well as the microenvironment in the bone marrow [62].

In conclusion, we have investigated the presence of interstitial chromosome 2 deletions in several subpopulations of bone marrow cells from a mouse strain susceptible to rAML and found that del2 is detectable at a higher frequency in the LSK subpopulation than in populations containing more committed progenitors MEPs and CMPs or primitive HSC. Nevertheless, it should be noted that the growth periods used to allow cell division/multiplication allow for selection and loss of the cells carrying the deletion. According to our results, it seems that the cell of origin in this model of rAML is more likely to be an early progenitor. We suggest that further characterisation of the frequency of Sfpi1 deletions in cells with an HSC- (e.g. L− S+ K+ Thy-1lo Flt3−) and MPP-phenotype (e.g. LSK Thy-1lo Flt3+) should be performed to confirm these results. Further characterisation of the interaction between the chromosome 2 interstitial deletions and other mutations, and the effect of ionising radiation exposure on the proliferative capacity (in the presence and absence of del2) of these specific cells is required.

Conflicts of interest

The authors report no conflict of interest. The authors alone are responsible for the content and writing of the paper. This report is work commissioned by the National Institute for Health Research. The views expressed in this publication are those of the authors and not necessary those of the NHS, the National Institute for Health Research or the Department of Health.