Ikaros is a critical target during simultaneous exposure to X-rays and N-ethyl-N-nitrosourea in mouse T-cell lymphomagenesis

Authors

  • Shinobu Hirano,

    1. Radiobiology for Children's Health Program, Research Center for Radiation Protection, National Institute of Radiological Sciences, Inage-Ku, Chiba, Japan
    2. Department of Molecular Pathogenesis, Juntendo University, Bunkyo-Ku, Tokyo, Japan
    Search for more papers by this author
  • Shizuko Kakinuma,

    Corresponding author
    1. Radiobiology for Children's Health Program, Research Center for Radiation Protection, National Institute of Radiological Sciences, Inage-Ku, Chiba, Japan
    • Radiobiology for Children's Health Program, Research Center for Radiation Protection, National Institute of Radiological Sciences, 4-9-1 Anagawa, Inage-Ku, Chiba 263-8555, Japan
    Search for more papers by this author
    • Tel.: +81-432063160, Fax: +81-432514138

  • Yoshiko Amasaki,

    1. Radiobiology for Children's Health Program, Research Center for Radiation Protection, National Institute of Radiological Sciences, Inage-Ku, Chiba, Japan
    Search for more papers by this author
  • Mayumi Nishimura,

    1. Radiobiology for Children's Health Program, Research Center for Radiation Protection, National Institute of Radiological Sciences, Inage-Ku, Chiba, Japan
    Search for more papers by this author
  • Tatsuhiko Imaoka,

    1. Radiobiology for Children's Health Program, Research Center for Radiation Protection, National Institute of Radiological Sciences, Inage-Ku, Chiba, Japan
    Search for more papers by this author
  • Shinji Fujimoto,

    1. Department of Immunology, Field of Regeneration Control, Institute for Frontier Medical Sciences, Kyoto University, Sakyo-Ku, Kyoto, Japan
    Search for more papers by this author
  • Okio Hino,

    1. Department of Molecular Pathogenesis, Juntendo University, Bunkyo-Ku, Tokyo, Japan
    Search for more papers by this author
  • Yoshiya Shimada

    1. Radiobiology for Children's Health Program, Research Center for Radiation Protection, National Institute of Radiological Sciences, Inage-Ku, Chiba, Japan
    Search for more papers by this author

Abstract

Cancer risk associated with radiation exposure is considered the result of concurrent exposure to other natural and manmade carcinogens. Available data on the molecular characteristics of cancer after simultaneous exposure to radiation and chemicals are insufficient. In our study, we used a mouse thymic lymphoma (TL) model that was synergistically induced by simultaneous exposure to X-rays and N-ethyl-N-nitrosourea (ENU) at subcarcinogenic doses and analyzed the mutation frequency and spectrum of the TL-associated genes Ikaros, Notch1, p53 and Kras. We found that the point mutation frequency in Ikaros was significantly increased to 47% for simultaneous exposure compared to 13 and 0% for X-ray and ENU exposure alone, respectively. These mutations were mostly G:C > A:T at non-CpG sites and T:A > C:G, both of which are characteristic of ENU mutagenesis. About half of the point mutations were accompanied by loss of heterozygosity (LOH), typical of X-irradiation. The remaining half did not include LOH, which suggests that they were dominant-negative mutations. In Notch1, the frequency of abnormalities was high (>58%) regardless of the treatment, suggesting that Notch1 aberration may be important for T-cell lymphomagenesis. The p53 and Kras mutation frequencies were low for all treatments (<23%). Importantly, the frequency of TLs containing mutations in multiple genes, especially both Ikaros and Notch1, increased after simultaneous exposure. Thus, after simultaneous exposure, Ikaros is a critical target and is inactivated by ENU-induced point mutations and/or X-ray-induced LOH in T-cell lymphomagenesis. Furthermore, concomitant alterations of multiple tumor-associated genes may contribute to enhanced lymphomagenesis after simultaneous exposure.

A multitude of natural and manmade chemicals with cancer-initiating and cancer-promoting potential are present in the environment. Many human cancers show considerable dependence on lifestyle, such as the use of tobacco, and dietary factors.1 Moreover, radiation carcinogenesis in humans is considered to result from the combined effect of radiation exposure and environmental carcinogens. There have been many attempts to estimate the risk of such combined exposure in animal and cell culture models.2, 3 There is, however, almost no information about the molecular mechanism that leads to carcinogenesis following combined exposure. Murine thymic lymphoma (TL) can be reproducibly induced by exposure to ionizing radiation and chemical carcinogens such as N-ethyl-N-nitrosourea (ENU), and this model is considered useful in the search for and characterization of genes involved in the development of human T-cell acute lymphoblastic leukemia (T-ALL).

Causative genes of human lymphoid leukemia, including IKZF1, NOTCH1, TP53 and KRAS, are also involved in the induction of mouse TL. Ikaros is a Krüppel-type zinc-finger DNA-binding protein and hematopoietic cell-specific transcription factor.4, 5 IKZF1 is associated with some subtypes of human T-ALL and B-cell acute lymphoblastic leukemia (B-ALL). Current studies show that genetic alteration of IKZF1 is a less frequent event in human T-ALL compared to B-ALL.6, 7 However, loss of Ikaros promotes the development of T-cell lymphoma/leukemia in mice.8 Therefore, Ikaros may act as a tumor suppressor in T-ALL and B-ALL. We previously studied Ikaros alterations in TLs induced by X-rays or ENU exposure alone, and we defined mutation characteristics for each exposure, including unusual alternative-splicing, null, frameshift and point mutations.9–11 Interestingly, point mutations in Ikaros in X-ray-induced TL accompany loss of heterozygosity (LOH), whereas those in ENU-induced TL do not. The former are thus considered loss-of-function mutations, whereas the latter are gain-of-function or dominant-negative mutations.

Notch1 is a cell surface receptor that is essential for T-cell development.12 More than 50% of patients with T-ALL have NOTCH1-activating mutations in the heterodimerization (HD) and the polypeptide rich in proline, glutamate, serine and threonine (PEST) domains.13 Truncated Notch1 exhibits transforming activity and functions as an oncogene.14 Moreover, in mouse TL, Notch1 mutations, specifically insertions/deletions in the HD and PEST domains15 and deletions within the 5′-end region,16, 17 are frequently detected.

The loss of p53 results in the development of lymphoma in mice,18 and a corresponding high rate of TP53 mutation is found in human T-ALL.19 Studies using the mutant KrasG12D mouse model implicate hyperactive Ras in TL pathogenesis.20 Elevated levels of Ras signaling have been observed in ∼50% of patients with T-ALL.21

Here, to clarify the molecular mechanism of carcinogenesis induced by simultaneous exposure to radiation and a chemical carcinogen, we analyzed the tumor suppressor genes Ikaros and p53 and the oncogenes Notch1 and Kras in mouse TLs induced by simultaneous exposure to X-rays and ENU. We focused our attention on their mutation frequencies and on the spectrum of mutations induced in these genes. We then asked whether specific genes are mutated or if selective mutation types are increased in TLs after simultaneous exposure.

Abbreviations

B-ALL: B-cell acute lymphoblastic leukemia; ENU: N-ethyl-N-nitrosourea; GAPDH: glyceraldehyde-3-phosphate dehydrogenase; gpt: guanine phosphoribosyltransferase; HD: heterodimerization; Hes1: hairy enhancer of split-1; LOH: loss of heterozygosity; Mgmt: O6-methylguanine-DNA methyltransferase; PEST: polypeptide rich in proline, glutamate, serine and threonine; Rag: recombination activating gene; RBP-Jk: recombining binding protein suppressor of hairless; RSS: recombination signal-like sequence; RT-PCR: reverse transcriptase polymerase chain reaction; T-ALL: T-cell acute lymphoblastic leukemia; TL: thymic lymphoma

Material and Methods

Mice and tumor induction

B6C3F1 mice were purchased from Charles River (Kanagawa, Japan). The method of induction of TL by X-rays and ENU has been described.9, 22 In brief, beginning at 4 weeks of age, female B6C3F1 mice were exposed weekly to whole-body X-rays at a dose of 0.8, 1.0 or 1.2 Gy for four consecutive weeks or were given ENU in drinking water for four consecutive weeks at a dose of 100 or 200 ppm as a single agent, or they were given both treatments simultaneously (Fig. 1a). The mice were observed daily until moribund and then were killed under ether anesthesia and autopsied. Thymic tissues were then removed, weighed and prepared for molecular analyses. In total, 97 TLs (31 X-ray, 13 ENU and 53 simultaneous) were examined (Supporting Information Table S1), and each TL was given a number. All experiments with mice were conducted according to the legal regulations in Japan and were in compliance with the guidelines for the care of laboratory animals of the National Institute of Radiological Sciences.

Figure 1.

Incidence and latency period of TL. (a) Overview of TL induction by X-rays and ENU. (b) Incidence of TL in mice after treatment. Asterisks indicate significant differences for the incidence in simultaneous treatment versus the corresponding treatment with the same X-ray dose at *p < 0.001 or ENU concentration at **p < 0.001 based on the χ2 test. The number of mice per treatment is shown in Supporting Information Table S1. (c) Kaplan–Meier plot for TL-free survival. Asterisks indicate significant differences for the latency periods in combined treatment versus the corresponding treatment with the same X-ray dose at *p < 0.001 or ENU concentration at **p < 0.001 based on the log-rank test.

Reverse transcriptase polymerase chain reaction and Ikaros LOH analysis

Total RNA extracted from lymphoma cells was reverse transcribed to obtain cDNA as described.10 The cDNAs of Ikaros and glyceraldehyde-3-phosphate dehydrogenase (Gapdh) were amplified using the primers and PCR conditions described in Supporting Information Table S2. For the LOH analysis on chromosome 11 (including Ikaros), genomic DNA was extracted by the standard method and amplified by PCR using seven polymorphic markers (Supporting Information Table S3). PCR products were analyzed using a luminescent image analyzer LAS-3000 (Fujifilm, Tokyo, Japan).

Mutation analysis of Ikaros, Notch1, p53 and Kras

Mutations in Ikaros, Notch1 (HD and PEST domains), p53 and Kras were identified in cDNA. PCR primers for Ikaros were designed to cover exons 1–7, which encode the entire Ikaros gene product. Genomic DNA for Ikaros was also sequenced to determine mutations in both its coding region and those at the splice donor and acceptor sites. For Notch1, primers were designed to amplify exons 25–27 and exon 34, where the HD and PEST domains are located, respectively. The primer set for p53 targeted exons 5–9, which contain the DNA-binding domain. For Kras, the region between exons 1 and 3 was analyzed. Supporting Information Table S2 describes the primer sequences and PCR conditions for these reactions. PCR products were directly sequenced using a Big Dye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA) and an ABI PRISM 3100 Genetic Analyzer (Applied Biosystems) or were sequenced after TA cloning using a TOPO TA cloning kit (Invitrogen, Carlsbad, CA).

Determination of rearranged Notch1 fragments and sequencing

The DNA fragments containing rearrangements in the 5′-end region of Notch1 were isolated by the AccuPrime High Fidelity PCR System (Invitrogen). We used primers covering exons 1b and 2 of Notch1, including intronic sequences surrounding these exons (Supporting Information Table S2).16 The precise location of each rearrangement was determined by sequencing after TA cloning (Invitrogen).

Real-time RT-PCR analysis of enhancer of split-1 (Hes1) expression

The Hes1 transcript level was measured with a Mx3000P real-time PCR system (Stratagene, La Jolla, CA). PCR was performed using SYBR Premix Ex Taq (Takara Shuzo, Shiga, Japan). The Hes1 transcript level was normalized to the Gapdh transcript quantified using TaqMan Rodent GAPDH Control Reagents (Applied Biosystems). Data were analyzed with MxPro software, version 4.10 (Stratagene). Primer pairs are listed in Supporting Information Table S2.

Western blotting

Western blotting was performed as described10 with the following minor modifications: after transfer of protein from the gel, the membrane was blocked and probed with primary antibodies against Ikaros,23 activated Notch1 (Val1744; Cell Signaling Technology, Danvers, MA), p53 (Santa Cruz Biotechnology, Santa Cruz, CA) and β-actin (Santa Cruz Biotechnology). Horseradish peroxidase-conjugated anti-rabbit IgG (Cell Signaling Technology) or anti-goat IgG (Santa Cruz Biotechnology) was used as the secondary antibody. Signals were developed using ECL plus Western Blotting Detection Reagents (GE Healthcare, Little Chalfont, Buckinghamshire, UK) and analyzed using the LAS-3000.

Results

TL incidence and latency period

We first evaluated TL incidence and latency induced by simultaneous exposure to X-rays and ENU. Mice were exposed to a subcarcinogenic dose of X-rays (0.8, 1.0 or 1.2 Gy) or ENU (100 or 200 ppm) or simultaneously to X-rays and ENU (0.8 Gy + 100 ppm; 1.0 Gy + 100 ppm; 0.8 Gy + 200 ppm or 1.0 Gy + 200 ppm; Fig. 1a). The mice exposed simultaneously developed a higher incidence of TL, which was significantly increased to 70–80% compared to that induced by exposure to X-rays (<36%) or ENU (<28%) alone (p < 0.001; Fig. 1b and Supporting Information Table S1). Furthermore, TL incidence after simultaneous exposure was much higher than the sum of the incidences from the separate exposures, indicating a synergistic effect.

The Kaplan–Meier plot of TL-free survival is shown in Figure 1c, and average latency of TL development is shown in Supporting Information Table S1. The latency periods for simultaneous exposure were significantly shorter than those for single exposure (p < 0.001). These data indicated that simultaneous exposure to X-rays and ENU accelerated the onset of TL.

Characteristics of Ikaros alteration in simultaneous exposure-induced TL

Point mutations, altered expression, alternative splicing and LOH of Ikaros occur in mouse TL.9–11 Our analyses showed that the Ikaros alteration frequency increased significantly after simultaneous exposure (50.9%) compared to X-ray exposure (25.8%) or ENU exposure (0%) alone (p < 0.05 and p < 0.001, respectively; Table 1 and Supporting Information Table S4). In particular, the Ikaros point mutation frequency in TLs increased after simultaneous exposure (47.2%) compared to X-ray exposure (12.9%) or ENU exposure (0%) alone (p < 0.01; Table 1).

Table 1. Summary of Ikaros, Notch1, p53 and Kras alterations in X-ray-, ENU- and simultaneous exposure-induced TLs
inline image

TLs induced by simultaneous exposure showed an increased frequency of base substitutions, especially G:C > A:T at non-CpG sites and T:A > C:G compared to X-ray or ENU exposure alone (p < 0.05; Fig. 2a). Interestingly, G:C > A:T mutations detected after X-ray exposure were located at CpG sites.

Figure 2.

Characteristics of Ikaros alterations in TLs. (a) Base substitutions in Ikaros shown as a percentage of all TLs examined. X-rays: 0.8, 1.0 or 1.2 Gy; ENU: 100 or 200 ppm; Simultaneous: the combination of 0.8 or 1.0 Gy and 100 or 200 ppm. *p < 0.05. 1Versus X-ray treatment; 2versus ENU treatment. (b) Distribution of Ikaros point mutations. Schematic diagrams of the Ikaros isoform (upper) and amino acid sequence of F2 and F3 (codons 140–195; lower) are shown. Sequences of zinc fingers are boxed. Linker regions are underlined. Amino acids shaded in black represent cysteines and histidines, and those shaded in gray represent hydrophobic residues. The circles and squares indicate individual mutations in TLs induced by the indicated treatments. (c) Schematic diagram of the different mutation patterns in each allele of Ikaros found in three TLs, each of which led to inactivation of this gene. Ik-1 and Ik-2 are the Ikaros isoforms expressed in normal thymocytes, whereas Ik-4 is the shortened isoform generated by alternative splicing in TLs. Abbreviations: F: finger; Ex: exon; Ins: insertion; FS: frame shift.

Next, the LOH status of Ikaros was examined using the seven microsatellite markers on chromosome 11 (Supporting Information Table S3). LOH analysis revealed that point mutations accompanied LOH in 40% (10/25) of simultaneous exposure-induced TLs (Table 1).

The Ikaros protein contains four Cys2His2 zinc fingers (F1–F4) near its N-terminus that contribute to sequence-specific DNA binding (Fig. 2b). The Cys2His2 motif is characterized by conserved cysteines, histidines and hydrophobic residues, which stabilize the three-dimensional structure of the protein.24 Of the point mutations that occurred after simultaneous exposure, 75% (20/25) were located in the DNA-binding domain. Others were located in the activation domain or dimerization domain. All point mutations were of missense or nonsense type (Supporting Information Table S4).

Interestingly, in each of three simultaneous exposure-induced TLs, both Ikaros alleles were inactivated by a different mutation pattern (Figs. 2b and 2c). TL no. 90 had a point mutation in F3 of the DNA-binding domain (codon 188) of one allele and a frameshift insertion in the activation domain of the other allele. On one allele TL nos. 57 and 70 had point mutations in the activation domain and dimerization domain, respectively, while both TLs expressed alternative mRNA Ik-4, which lacks exons 3 and 5, from the other allele.

Frequent abnormality of Notch1

Deletions in the 5′-region and mutations in the HD and PEST domains of Notch1 were found frequently in mouse TLs. The frequency of Notch1 abnormality was high in all treatment groups 58.1, 84.6 and 73.6% after X-ray, ENU and simultaneous exposure, respectively (Table 1 and Supporting Information Table S4). The frequency of 5′-region deletions was 32.3, 46.2 and 34.0% after X-ray, ENU and simultaneous exposure, respectively (Table 1). Examples of deletions are shown in Figure 3a. Consistent with a previous study,16 most of the breakpoints in the rearrangements were located at hotspots and accompanied a recombination signal-like sequence (RSS) at both ends (Fig. 3a-1). These data revealed that rearrangements in the 5′-region in all treatment groups occurred through a recombination activating gene (Rag)-dependent mechanism.17 TL no. 40 was an exception, as the breakpoints deviated from the hotspot position; guanine was added to the deletion junction as a nontemplated nucleotide (Fig. 3a-2). In this case, there was no RSS at the breakpoints, and thus, deletion may have occurred through a Rag-independent mechanism.17

Figure 3.

Notch1 abnormality in TL. (a) Examples of Notch1 5′-end region deletion. The domain organization of the Notch1 receptor (upper) and corresponding genomic DNA with exons (lower) are shown. Abbreviations: EGF-like: epidermal growth factor-like; LNR: Lin-12/Notch repeats; HD: heterodimerization domain; TM: transmembrane domain; ICN: intracellular Notch; RAM: RBP-Jk-associated molecule domain; ANK: ankyrin repeats; TAD: transactivation domain; PEST: PEST domain. (1) Rearrangement with a RSS at both ends. Numerals indicate positions of hotspot breakpoints (GenBank accession number AB100603). Dotted lines designate deleted regions. Lowercase letters indicate small insertions; underlined letters indicate P-nucleotides; letters without underlining indicate N-nucleotides. (2) Deletion with a nontemplated nucleotide addition. Numerals indicate the end positions of the retained DNA. The lowercase “g” indicates a nontemplated nucleotide in a deleted region (dotted line). (b) Hotspots for mutations in the PEST domain. The bar indicates the Notch1 cDNA upstream of the PEST domain. Bold numerals indicate amino acid residues, and the numerals below them correspond to the nucleotide positions in Notch1 cDNA at which mutations occurred (GenBank accession number NM_008714.3). Mutation hotspots at codons 2361 and 2398 are boxed. The circles, triangles and squares indicate individual mutations in TLs induced by the indicated treatments.

Mutations in the HD and PEST domains were found frequently after all treatments: 51.6, 69.2 and 62.3% after X-ray, ENU and simultaneous exposure, respectively (Table 1). Most mutations were located in the PEST domain, and they were primarily nucleotide insertions and/or deletions that led to frameshift mutations resulting in premature stop codons and PEST domain deletions (Supporting Information Table S4). Single-base substitutions were relatively rare, and all were nonsense mutations. Interestingly, mutations in simultaneous exposure-induced TLs were detected frequently at codons 2361 and 2398 in the PEST domain (Fig. 3b). After X-ray exposure, mutations were also frequently found at codon 2361. Codons 2361 and 2398 were thus considered hotspots, consistent with an earlier report.15 The HD domain had a few mutations, all of which were located at residues conserved in human, mouse, Xenopus, zebrafish and fly (data not shown).25, 26

To examine when 5′-region deletion occurred during lymphomagenesis, we analyzed the nucleotide sequences around the deletion junctions and their clonality (Supporting Information Table S5). More than half of the TLs had only one sequence alteration, independent of the treatment, suggesting high clonality. Thus, these deletions occurred before the onset of malignant proliferation. The remaining TLs had two to four sequence alterations, indicating that deletion occurred during the expansion of the lymphoma. TL no. 42 had the greatest variety of deletions; there were 24 separate sequence alterations, suggesting that each deletion formed independently more than once as malignant cells proliferated.

Mutations of p53 and Kras

Mutation analyses revealed that the mutation frequency of p53 was low in all treatment groups 16.1, 7.7 and 18.9% after X-ray, ENU and simultaneous exposure, respectively (Table 1 and Supporting Information Table S4). The mutation frequency of Kras was also low in all treatment groups 12.9, 23.1 and 11.3% after X-ray, ENU and simultaneous exposure, respectively (Table 1 and Supporting Information Table S4).

Concomitant alterations of multiple tumor-associated genes in simultaneous exposure-induced TLs

Studies have indicated a profound effect of simultaneous alterations of oncogenes and/or tumor suppressor genes in carcinogenesis.27, 28 We examined the incidence of TLs harboring mutations in one or more of the following genes: Ikaros, Notch1, p53 and Kras. X-ray- or ENU-induced TLs most frequently harbored only one mutated gene (48.4 and 84.6%, respectively; Supporting Information Table S6). In simultaneous exposure-induced TLs, however, the incidence of TLs harboring two or more mutated genes was significantly higher (52.8%) than with X-ray exposure (29.0%) or ENU exposure (15.4%) alone (p < 0.05). Examining the combinations of mutated genes, we found that the incidence of TL with both Ikaros and Notch1 alterations was significantly higher after simultaneous exposure (43.4%) than after X-ray exposure (19.4%) or ENU exposure (0.0%) alone (p < 0.05; Table 2).

Table 2. Combinations of mutated genes in TLs
inline image

Analysis of Hes1 transcription as a measure of Notch signaling

Ikaros plays a crucial role in vitro in repressing the transcriptional response to Notch signaling in T-cell development.29 To repress transcription of the Notch1 target gene Hes1, Ikaros and recombining binding protein suppressor of hairless (RBP-Jk) compete for common recognition sites in the Hes1 promoter in normal thymocytes. Furthermore, cooperative mutations causing a reduction in Ikaros activity and an increase in Notch1 activation promote T-cell leukemogenesis.30, 31

To evaluate the activity of Notch signaling in cells carrying alterations in Ikaros and/or Notch1, we measured the Hes1 transcript level in TL and normal thymus mRNAs by real-time RT-PCR. All TLs were classified into four categories: (i) both Ikaros and Notch1 were mutated (IK-mut/NOTCH-mut); (ii) only Ikaros was mutated (IK-mut); (iii) only Notch1 was mutated (NOTCH-mut) and (iv) neither Ikaros nor Notch1 was mutated (no-mut). The NOTCH-mut group displayed higher levels of Hes1 transcripts than did the no-mut group (2.1-fold, p < 0.01; Fig. 4). Furthermore, the IK-mut/NOTCH-mut group exhibited much higher Hes1 levels compared to the NOTCH-mut, IK-mut and no-mut groups (1.5-, 2.2- and 3.3-fold, respectively; Fig. 4).

Figure 4.

Expression of Hes1 transcripts in TLs harboring Ikaros and/or Notch1 mutations. Hes1 expression in TLs was compared to that in normal thymus. Transcript levels were measured using real-time RT-PCR of TL and normal thymus mRNAs, normalized with Gapdh and transformed into logarithmic values. Horizontal bars represent median values. p-values were calculated using the two-tailed Mann-Whitney test. *p < 0.05; **p < 0.01; ***p < 0.001.

Discussion

Our study revealed that the synergistic effect of simultaneous exposure to X-rays and ENU on T-cell lymphomagenesis could be ascribed to an increase in the mutation frequency of Ikaros, and in particular to an increase in point mutations. The high frequency of Notch1 abnormalities, either through deletions or insertions but not point mutations, (but not point mutations) was independent of the treatment, suggesting that it was a gene that was not specifically targeted by the action of ENU or X-rays. Notch signals are required for normal T-cell development, and their deregulation disrupts cell cycle progression and apoptosis, leading to transformation.32 The mutation frequencies of p53 and Kras were low for all treatments, suggesting that mutations in these genes were not a major event in lymphomagenesis after simultaneous exposure. This high selectivity for point mutations in Ikaros may be associated with Ikaros functions during early stages of thymocyte differentiation, such as β-selection, which promotes the clonal expansion and further differentiation of immature T cells.33 During the next round of cell division, DNA adducts induced by ENU might be mismatched during DNA replication, such that the repairable DNA lesion becomes fixed in Ikaros as a point mutation.

The point mutations in Ikaros that were significantly increased after simultaneous exposure were G:C > A:T at non-CpG sites and T:A > C:G. Using guanine phosphoribosyltransferase (gpt)-delta transgenic mice harboring the gpt reporter transgene as a mutagenesis test system, we previously demonstrated that ENU induced G:C > A:T mutations at CpG sites in thymocytes.22 We also reported a frequent T:A > C:G transition in Ikaros, which accompanied LOH at Ikaros loci in TLs induced by X-rays but not in those induced by ENU.10, 11 In our study, 80% of T:A > C:G transitions did not accompany LOH (Supporting Information Table S4). Taken together, these results suggest that G:C > A:T mutations at non-CpG sites and T:A > C:G mutations in Ikaros may be induced by the direct action of ENU.

There are a few possible explanations for the enhancement of ENU-induced point mutations upon X-irradiation. A direct mechanism may involve impaired activation of the DNA repair pathway associated with ENU-induced DNA damage. ENU produces ethylated base damage leading to G:C > A:T and T:A > C:G transitions. Base damage of guanine is repaired by O6-methylguanine-DNA methyltransferase (Mgmt), which recognizes and directly binds O6-ethylguanine and removes the ethyl residue, thereby preventing G:C > A:T transitions.34 Expression of Mgmt remained unchanged in the thymuses of mice exposed to X-rays (1.0 Gy), whereas it was dramatically increased after ENU treatment (200 ppm; unpublished data). Thus, Mgmt expression induced by ENU might be repressed by simultaneous exposure to X-rays, leading to enhancement of the mutation induction. This mechanism is supported by the fact that the frequency of G:C > A:T mutations increased after simultaneous exposure; however, this increase is not observed after exposure to ENU alone.11

An indirect mechanism may also play a role in the clonal expansion of cells susceptible to ENU during the recovery and proliferation phases of surviving thymocytes after irradiation. The major effect of radiation is to induce cell death, which may be followed by the regeneration of the surviving cells.35 Fractionated whole-body irradiation, that is, once weekly for four consecutive weeks, causes a repeated cycle of regeneration.36 Rapidly dividing cells may permanently fix the ENU-induced DNA lesions as point mutations during DNA replication before repair of the lesion can occur. Cells with these ENU-induced mutations may have undergone clonal expansion during repeated exposure to X-rays. Seyama et al.35 demonstrated that X-ray exposure 5 days before but not 30 days before ENU treatment enhances lymphomagenesis. We also reported that 1.0 Gy irradiation followed by ENU (200 ppm) enhanced mutant frequency and accelerated clonal expansion of mutated cells in the thymuses of gpt-delta mice.22 Therefore, coexposure to X-rays may accelerate ENU lymphomagenesis by promoting the induction of mutations and/or by giving existing mutants a chance to expand.

In our study, 60% (15/25) of Ikaros point mutations were not accompanied by LOH after simultaneous exposure. Because Ikaros forms a heterodimeric complex, Ikaros mutants containing lesions in the DNA-binding domain would affect wild-type Ikaros in a dominant-negative manner.8 Most of the point mutations without LOH changed amino acid residues located in the N-terminal zinc fingers, which are expected to affect DNA binding (10/15; Fig. 2b) and thus should function as dominant-negative mutations. Concerning the locations of these 10 point mutations, codons 147, 175 and 195 correspond to cysteines or histidines, which are involved in zinc coordination. Those substitutions would be expected to disrupt the conserved Cys2His2 motif.24 IkPlastic/+ mice, with a point mutation that affects the histidine at codon 191, show dominant-negative effects on transcription.37 A mutation at codon 166, which is next to a histidine, could affect the three-dimensional structure of the zinc finger. Codon 160 is a hydrophobic residue that is important for the structure of the zinc finger.24 Phosphomimetic substitutions at specific linker residues abolish Ikaros DNA binding and pericentromeric localization.38 Although codon 172 in the linker region was not one of those specified residues, its mutation might affect DNA binding, because each residue within a canonical linker can contribute to the stability of the protein–DNA complex.39 Codons 184 and 186, located at the −1 and +2 position of the α-helix, respectively, are important for sequence-specific DNA interactions of Ikaros.40, 41 Codon 180, which is located in the β-sheet, and codon 184 are critical for both Ikaros DNA binding and pericentromeric targeting in 3T3 fibroblasts.42

Interestingly, mutations in three TLs were complicated, in that both Ikaros alleles were inactivated by the occurrence of point mutations and other mutation patterns (TL nos. 57, 70 and 90; Fig. 2c). These multiple mutations have not been detected in Ikaros alleles following X-ray or ENU exposure alone. These data indicate that point mutation of one Ikaros allele together with another mutation in the other allele could lead to the loss of Ikaros function without LOH. In contrast, 40% (10/25) of Ikaros point mutations were accompanied by LOH in simultaneous exposure-induced TL. These point mutations could result in the loss of function based on the “two-hit” mechanism.43

We thus propose that ENU, when accompanied by X-irradiation, efficiently induces point mutations in Ikaros. Point mutations without LOH may yield a dominant-negative form of Ikaros. The remaining point mutations would require the loss of wild-type Ikaros by LOH as a second hit for complete loss of tumor suppressor function. Because LOH is reportedly stimulated by irradiation,9, 10 the latter mechanism may be important for the stimulation of lymphomagenesis by simultaneous exposure.

The incidence of TLs harboring two or more mutated genes, especially Ikaros and Notch1, was significantly higher after simultaneous exposure than after single exposure, suggesting that cancer-associated genes cooperate to advance tumorigenesis. Cooperative mutations in proto-oncogenes and tumor suppressor genes occur in many forms of cancer.27, 28 The model of cooperative mutations causing a reduction in Ikaros activity and an increase in Notch1 activation fits this paradigm in the genesis of T-ALL.30, 44 However, most Notch1 mutations found in TL do not generate signals of sufficient strength to initiate leukemia development and require other leukemogenic factors.45 Furthermore, Ikaros undergoes a genetic change earlier than Notch1 during lymphomagenesis after irradiation, suggesting that activated Notch1 may contribute to expansion of late preleukemic cells.46 Thus, Ikaros alteration rather than Notch1 abnormality is the critical, and possibly the first, mutation in the multistep evolution of simultaneous exposure-induced TL. There are, however, other important genes in this process that remain to be studied. Bcl11b and Pten, for example, are also altered in early preleukemic cells and in TLs after irradiation, suggesting that they may be key factors of lymphomagenesis.46

Importantly, Ikaros and Notch1 intersect in the regulation of normal T-cell development. Additionally, leukemogenesis induced by Ikaros deficiency is similar in phenotype to that induced by constitutively active Notch1.33, 47, 48 Kathrein et al.41 contended that Ikaros deficiency results in de-repression of Hes1 in vivo and calculated that activating mutations of Notch1 in Ikaros-deficient cells further increase the expression of Hes1. In our study, the Hes1 level in TLs with mutations in both Ikaros and Notch1 was increased significantly compared to TLs harboring either mutation alone, suggesting that cooperative mutations advance Notch signaling via Hes1.

In conclusion, Ikaros is a critical target of T-cell lymphomagenesis upon simultaneous exposure to X-rays and ENU. Moreover, concomitant alterations of multiple tumor-associated genes, especially Ikaros and Notch1, may synergistically contribute to lymphomagenesis. This is the first report to delineate the target gene and the cooperative effect of multiple gene mutations in tumorigenesis after simultaneous exposure to X-rays and ENU. These results may help to further explain features of human carcinogenesis, such as the enhanced rate of secondary cancers caused by radiotherapy in patients who smoke after their treatment.49 Our findings may also contribute to the risk assessment of human carcinogenesis after exposure to multiple environmental carcinogens.

Acknowledgements

The authors thank all laboratory members for their encouragement throughout this work, and the Laboratory Animal Science Section in the National Institute of Radiological Sciences for animal management. This project was funded under grants from the Long-Range Research Initiative of the Japan Chemical Industry Association to Y.S. and S.K. (2003CC03 to 2005CC03 and 2006CC03 to 2010CC03). This project was funded, in part, under a grant-in-aid from the Ministry of Education, Culture, Sports, Science, and Technology of Japan to S.K. (18510053) and Y.S. (21610029) and a grant of “Ground-Based Research Announcement for Space Utilization” promoted by the Japan Space Forum to Y.S.

Ancillary