Thiopurines are extensively used as immunosuppressants and in the treatment of childhood cancers, even though there is concern about therapy-induced leukemias and myelodysplastic syndromes resulting from thiopurine use. Following metabolic activation, thiopurines are incorporated into DNA and invoke mismatch repair (MMR). Recognition of 6-thioguanine (6-thioG) in DNA by key MMR proteins results in cell death rather than repair. There are suggestions that homologous recombination (HR) is involved downstream of MMR following thiopurine treatment, but the precise role of HR is poorly understood. In this study, we demonstrate that cells deficient in RAD51D (a RAD51 paralogue) are extremely sensitive to 6-thioG. This sensitivity is almost completely rescued by the deletion of Mlh1, which suggests that HR is involved in the repair of the 6-thioG–induced recombinogenic lesions generated by MMR. Furthermore, 6-thioG induces chromosome aberrations in the Rad51d-deficient cells. Interestingly, Rad51d-deficient cells show a striking increase in the frequency of triradial and quadriradial chromosomes in response to 6-thioG therapy. The presence of these chromatid exchange–type aberrations indicates that the deficiency in RAD51D-dependent HR results in profound chromosomal damage precipitated by the processing of 6-thioG by MMR. The radials are notable as an important source of chromosomal translocations, which are the most common class of mutations found in hematologic malignancies. This study thus suggests that HR insufficiency could be a potential risk factor for the development of secondary cancers that result from long-term use of thiopurines in patients. Mol Cancer Res; 9(2); 206–14. ©2011 AACR.

This article is featured in Highlights of This Issue, p. 131

Antimetabolite thiopurines are extensively used as immunosuppressants and in the treatment of childhood cancers. However, there is increasing concern about therapy-induced leukemias and myelodysplastic syndromes, particularly maintenance therapies, in patients treated with thiopurines. For example, the relative risk of developing leukemia is 5.3 for heart and/or lung transplant patients (1). Data from a large French cohort of patients (19,486) with inflammatory bowel disease identified a relative risk of developing lymphoproliferative disorders as 5.2 for patients who were treated with thiopurines compared with those who were not (2). Despite the success stories of treating childhood leukemias with thiopurines as part of the regimen, there is a clear risk of secondary neoplasms associated with the duration and intensity of maintenance therapy (3).

Although DNA is acknowledged as the ultimate target, the molecular mechanisms by which thiopurines exert their antiproliferative effects via DNA damage are poorly understood. Thiopurines are metabolized to form 6-thioguanine (6-thioG), which is ultimately misincorporated into DNA opposite thymine (1). The cytotoxic DNA damage that 6-thioG induces is intriguing in that it invokes 2 DNA repair responses, mismatch repair (MMR) and homologous recombination (HR), with differing consequences. Biochemically, base:base mismatches are recognized by 2 heterodimeric MMR protein complexes, MSH2:MSH6 (also known as MutSα) and MLH1:PMS2 (also known as MutLα), which recruit exonuclease 1 (Exo1). Processing of 6-thioG:T and the structurally similar O6-methyl guanine (O6-meG) adduct by MutSα, MutLα, and Exo1 does not lead to repair but instead triggers cell death (4, 5). Therefore, loss of MMR by genetic or epigenetic means can promote a “tolerance” to DNA damage and resistance to DNA-damaging agents (6–8). There has been debate whether the induction of apoptosis results from MMR's “futile” attempt at repair or the involvement of MMR proteins in DNA damage checkpoint signaling. Data from studying mutants of MSH2 and MSH6 suggested that the mismatch correction and damage-sensing functions of MMR are distinct (9, 10). Yet, more recent articles clearly implicate the involvement of the nuclease function of Exo1 downstream of MutSα/MutLα because Exo1-deficient mice and cells, including cells expressing a nuclease-deficient Exo1, are resistant to killing by 6-thioG and agents that cause O6-meG lesions (4, 5).

The role of HR in the cellular response to thiopurines is poorly understood, but there is cytogenetic and genetic evidence that HR is involved in the processing of 6-thioG and O6-meG lesions, regardless of the precise damage signaling and/or futile cycling roles of MMR. For example, 6-thioG induces sister chromatid exchanges (SCE), a strong indicator of HR events, that are dependent on the presence of functional MMR (11). We recently reported the generation of a unique mouse model combining MMR and HR defects by deletion of Mlh1 and Rad51d (12). RAD51D is a member of the RAD51 protein family that plays an important role in HR and has been identified to interact as part of a complex with MSH2 in a proteomics screen of RAD51D interacting partners (13). We demonstrated that the sensitivity of Rad51d-deficient cells to MNNG, an alkylating agent that causes O6-meG lesions, was partially dependent on the presence of MMR. Loss of Mlh1 not only alleviated sensitivity of Rad51d-deficient cells to MNNG but also reduced γ-H2AX induction, G2/M cell-cycle arrest, and apoptosis. In addition to RAD51D, other HR proteins including XRCC2, XRCC3, and BRCA2 have been implicated in protection against O6-meG lesions (14, 15). Cells deficient in the FancG protein are sensitized to 6-thioG (16), and recently, it was reported that BRCA2-deficient cells are sensitive to 6-thioG (17), thus implicating RAD51-associated HR in the cellular response to thiopurines.

In this study, we examined the role of RAD51D in repair of 6-thioG–induced DNA damage and found that HR is indeed required in the cellular response to 6-thioG–induced damage downstream of MMR. Rad51d-deficient mouse embryonic fibroblasts (MEF) are hypersensitive to 6-thioG and this sensitivity is almost completely dependent on the presence of MMR. Rad51d-deficient cells show a heightened G2 arrest accompanied by an increase in the aneuploid cell population and multinucleation. In addition, there is a substantial increase in the chromosomal aberrations, especially radials in the Rad51d-deficient cells upon 6-thioG treatment. These findings have implications for understanding the mechanisms of therapy-induced cancers caused by therapeutic thiopurines.

Cell lines

MEF cell lines Rad51d+/+Trp53−/−, Rad51d−/−Trp53−/−, Rad51d−/−Trp53−/−Mlh1−/−, and Mlh1−/−Trp53−/− were generated and characterized as described previously (12). Because Rad51d−/− MEFs with wild-type p53 fail to proliferate, all comparisons between MLH1 and RAD51D status were made with Trp53−/ cells. The cell lines were cultured in Dulbecco's modified Eagle's medium (DMEM; Cellgro Mediatech) supplemented with 10% newborn calf serum (NCS), 2 mmol/L glutamine, and 100 units/mL penicillin–streptomycin (Hyclone).

Colony survival assays

For colony survival assays, MEFs were seeded onto 100-mm dishes at 800 cells per plate in DMEM (Cellgro Mediatech) supplemented with 10% FBS, 2 mmol/L glutamine, and 100 units/mL penicillin–streptomycin (Hyclone). Cells were treated with 6-thioG (Sigma-Aldrich) at the indicated concentrations. After 24-hour treatment, 6-thioG was removed and fresh medium was added. During the experiment, the medium was replaced once after 5 days. Cells were grown for 8 to 10 days, fixed with 100% methanol, and stained with Giemsa (Sigma-Aldrich). Colonies containing 50 or more cells were counted. Three independent experiments were performed, and each experiment was performed in triplicates. Error bars represent the standard error of the mean. The LC50 values were calculated from a semi-log plot generated by nonlinear regression analysis, using GraphPad Prism 4.0 (GraphPad Software, Inc.).

Cell-cycle analysis

For cell-cycle analysis, 1.0 × 105 cells were seeded onto 100-mm dishes and the medium containing 0.1 μmol/L 6-thioG was added to the treatment plates 42 hours after seeding. The medium was replaced 6 hours posttreatment with drug-free medium. At the indicated time points posttreatment, all cells were collected, washed with PBS, and gently resuspended in propidium iodide (PI; 50 μg/mL) staining solution containing 4 mmol/L sodium citrate (pH 7.8), 30 units/mL RNase A (Sigma), and 0.1% Triton X-100. After an incubation period of 10 minutes at 37°C, NaCl was added to a final concentration of 0.15 mol/L (18). Cell-cycle analyses were performed using a Beckman Coulter FC 500 cytometer (Beckman Coulter), and data were quantified using ModFit LT software version 3.1 (Verity Software House).

Mitotic index

Cells were treated with 0.1 μmol/L 6-thioG for 6 hours and then incubated in medium containing 200 ng/mL nocodazole for 6 hours before harvesting at 48- and 72-hour time points posttreatment. The cells were fixed in 70% ethanol overnight (−20°C) and then permeabilized with 0.25% Triton X-100 and stained with phospho-histone H3 antibody Alexa Fluor 488 conjugate (Cell Signaling), followed by PI (10 μg/mL PI and 100 μg/mL RNaseA in PBS) staining. The cells were then analyzed using a Beckman Coulter FC 500 cytometer and data was quantified using CXP analysis software (Beckman Coulter).

Metaphase spreads and nuclear phenotypes

Giemsa-stained metaphase spreads were prepared as described (19). In brief, cells were treated with 0.1 μmol/L 6-thioG for 6 hours and mitotic cells were harvested at 24-, 48-, and 72-hour time points following a 3-hour incubation with 20 ng/mL colcemid (Roche Applied Science). The cells were incubated in hypotonic solution [5% (v/v) NCS, 2 mg/mL KCl, and 2 mg/mL sodium citrate in distilled water] for 15 minutes at 37°C twice, followed by 2 rounds of fixation in fresh ice-cold Carnoy's fixative [3:1 (v/v) mixture of methanol and acetic acid]. To prepare chromosome spreads, fixed cells were dropped onto the slides, air dried, stained with 4% Giemsa (Sigma), washed in water, and mounted using Permount (ThermoFisher Scientific). The slides were observed on a Nikon Eclipse E600 Microscope, using 100 × objective, and the images were captured using a Q-imaging CCD camera. Chromosomal aberrations were scored from a minimum of 1,000 chromosomes.

SCE analysis

The SCE analysis was performed as described (19). In brief, 5 × 105 cells were seeded onto 100-mm dishes and 24 hours later, medium containing 0.1 μmol/L 6-thioG and 10 μmol/L bromodeoxyuridine (BrdU; Sigma-Aldrich) was added to the cells. After 6 hours of 6-thioG treatment, fresh medium containing only 10 μmol/L BrdU was added. The cells were harvested 48 hours later following a 3-hour incubation with 20 ng/mL colcemid (Roche), and metaphase spreads were prepared as described previously. The spreads were then differentially stained by incubation in 10 μg/mL Hoechst in 1× Sorenson's buffer, rinsed in 2× SSC, exposed to black light for 45 minutes at 65°C, and subsequently rinsed with 1× Sorenson's buffer. Staining, mounting, and the observations were performed as described for the metaphase spreads.

Immunofluorescence

The cells were plated at a density of 5,000 cells per coverslip 24 hours prior to treatment with 0.1 μmol/L 6-thioG. The medium was changed 6 hours posttreatment. After the treatment, the coverslips were washed with PBS and incubated for 30 seconds in microtubule stabilization buffer (130 mmol/L HEPES; 4.2 mmol/L MgCl2; 10 mmol/L EGTA, pH 6.9). The cells were fixed with 4% formaldehyde (ThermoFisher Scientific) and permeabilized with 0.2% Triton X-100. Fixed cells were incubated with β-tubulin antibody (Abcam) overnight at 4°C and Alexa Fluor 488–conjugated anti-rabbit secondary IgG (Molecular Probes) for 1 hour at room temperature to visualize microtubules. Nuclei were stained with DAPI (4′,6-diamidino-2-phenylindole), and the cells were mounted. The coverslips were examined with an Olympus microscope (IX81), using a 40× objective, and the images were captured using Hamamatsu OrcaR2 camera. At least 500 cells were counted per sample.

Statistical analysis

Calculations of the mean, SD, and standard error were performed using Microsoft Excel. Statistical analysis for comparison of each set of experimental means was performed using GraphPad Instat 3.0 (GraphPad Software, Inc.). Statistical analysis of proportions was performed using Sigmaplot 11 (Systat Software Inc.).

Absence of RAD51D enhances the sensitivity of MMR-proficient cells to 6-thioG

We previously developed a Rad51d- and Mlh1-deficient MEF cell line to study the combined defects of MMR and HR (12). In this study, we first confirmed that the absence of Mlh1 alleviated the sensitivity of Rad51d-proficient cells to 6-thioG by 10.1-fold (Fig. 1A), which agreed with previous studies (20). Next, the role of RAD51D in response to 6-thioG in a wild-type mismatch repair background was evaluated. Rad51d−/−Mlh1+/+ cells were approximately 4.9-fold more sensitive to 6-thioG, with an LC50 value of 0.065 μmol/L, than 0.32 μmol/L for Rad51d+/+Mlh1+/+ cells (Fig. 1B). The next comparison done was to see whether loss of MMR alleviated the sensitivity of the HR-deficient cells. Indeed, the Rad51d−/−Mlh1−/− cells were dramatically resistant compared with Rad51d−/−Mlh1+/+ cells by 36.4-fold (Fig. 1C). In other words, when one compares sensitivity in the Mlh1-deficient background, the Rad51d-deficient cells were only 1.4-fold more sensitive to 6-thioG than Rad51d-proficient cells, with an LC50 value of 2.38 μmol/L compared with 3.23 μmol/L (Supplementary Fig. S1). This compellingly shows that the hypersensitivity of HR-deficient cells is downstream of MMR and highly dependent on functional MMR. We also observed an increased induction of the sub-G1 population in the Rad51d-deficient cells in response to 6-thioG, indicating the role of RAD51D in repairing lesions that otherwise induce apoptosis (Supplementary Fig. S2B).

Figure 1.

Rad51d-deficient MEFs are extremely sensitive to 6-thioG in an MMR-proficient background. Colony survival assays of MEFs deficient in Rad51d, Mlh1, or both, treated with 6-thioG at the indicated doses. A, pair-wise comparison of Mlh1-deficient (♦) and Mlh1-proficient (▪) MEFs. B, pair-wise comparison of Rad51d-proficient (▪) and Rad51d-deficient (▴) MEFs. C, pair-wise comparison of Rad51d-deficient Mlh1-proficient (▴) and Rad51d-deficient Mlh1-deficient (▾) MEFs. Survival is plotted as percentage compared with untreated control. Each data point represents the mean of 3 independent experiments, each performed in triplicate. Error bars, SEMs.

Figure 1.

Rad51d-deficient MEFs are extremely sensitive to 6-thioG in an MMR-proficient background. Colony survival assays of MEFs deficient in Rad51d, Mlh1, or both, treated with 6-thioG at the indicated doses. A, pair-wise comparison of Mlh1-deficient (♦) and Mlh1-proficient (▪) MEFs. B, pair-wise comparison of Rad51d-proficient (▪) and Rad51d-deficient (▴) MEFs. C, pair-wise comparison of Rad51d-deficient Mlh1-proficient (▴) and Rad51d-deficient Mlh1-deficient (▾) MEFs. Survival is plotted as percentage compared with untreated control. Each data point represents the mean of 3 independent experiments, each performed in triplicate. Error bars, SEMs.

Close modal

6-thioG–induced G2 arrest is heightened in Rad51d-deficient cells

Previous studies have shown that 6-thioG induces a G2 arrest in MMR-proficient cells (21). Here, we examined the influence of HR on the 6-thioG–induced G2 arrest by examining cell-cycle progression in the Rad51d+/+Mlh1+/+, Rad51d−/−Mlh1+/+, and Rad51d−/−Mlh1−/− cell lines. Following treatment with 0.1 μmol/L 6-thioG, an increase in the 4n population was observed in both Rad51d+/+ Mlh1+/+ and Rad51d−/−Mlh1+/+ cell lines 48 hours postremoval of 6-thioG, which becomes much more pronounced at the 72-hour time point (Fig. 2 and Supplementary Fig. S2A). Specifically, we observed that the Rad51d+/+Mlh1+/+ cells show 30.9% ± 1.1% of cells in G2/M whereas Rad51d−/−Mlh1+/+ cells show 65.1% ± 4.8% in G2/M (Fig. 2A and B), which shows that the G2 arrest in MMR proficient cells is heightened by HR deficiency. Rad51d−/−Mlh1−/− cells show no accumulation in a G2/M population, which confirms that functional MMR is necessary for the cell-cycle arrest in response to 6-thioG (Fig. 2C). Note that there is no observable arrest 24 hours following treatment in any of the cell lines used.

Figure 2.

Cell-cycle progression of Rad51d-deficient and -proficient MEFs in response to 6-thioG in Mlh1-proficient and -deficient backgrounds. Cell-cycle progression was followed for untreated cells and those treated with 0.1 μmol/L 6-thioG for 6 hours. Percentage of the cells in the G1, S, and G2/M phases in the diploid population were calculated using MODFIT-LT software and shown for (A) Rad51d+/+Mlh1+/+, (B) Rad51d−/−Mlh1+/+, and (C) Rad51d−/−Mlh1−/−. Each data point represents the mean of 3 independent experiments. Error bars, SEMs. Statistical significance was determined by paired t test. *, P < 0.05; **, P < 0.01.

Figure 2.

Cell-cycle progression of Rad51d-deficient and -proficient MEFs in response to 6-thioG in Mlh1-proficient and -deficient backgrounds. Cell-cycle progression was followed for untreated cells and those treated with 0.1 μmol/L 6-thioG for 6 hours. Percentage of the cells in the G1, S, and G2/M phases in the diploid population were calculated using MODFIT-LT software and shown for (A) Rad51d+/+Mlh1+/+, (B) Rad51d−/−Mlh1+/+, and (C) Rad51d−/−Mlh1−/−. Each data point represents the mean of 3 independent experiments. Error bars, SEMs. Statistical significance was determined by paired t test. *, P < 0.05; **, P < 0.01.

Close modal

To confirm that the arrest was in G2 and determine whether the G2 checkpoint response was altered by the failure to complete HR, the mitotic index of the Rad51d−/− Mlh1+/+ and Rad51d+/+Mlh1+/+ cells was determined by phospho-histone H3 (Ser 10) staining (Fig. 3). Nocodazole trapping was used to quantify the progression of G2 cells into mitosis. At 48 hours, the mitotic index of the Rad51d-proficient cells treated with 6-thioG was reduced by 3.6% compared with the untreated cells (Fig. 3A). However at 48 hours, the mitotic index of the Rad51d-deficient cells was reduced by 5.9% compared with untreated cells (Fig. 3C). In other words, the data demonstrate that 6-thioG treatment induces a G2 arrest which is more pronounced in the Rad51d-deficient cells. At 72 hours, the reduction in mitotic index caused by 6-thioG treatment was less than that seen at 48 hours in both genotypes. Specifically, the Rad51d-proficient cells were reduced by 1.7% whereas the Rad51d-deficient cells were reduced by 3.3% (Fig. 3B and D). These data suggest that the G2 arrest does not persist up to 72 hours and the cells eventually progress into mitosis, although there are dire consequences in the HR-deficient cells (see the next section). Several conclusions can be drawn from these data. First, it confirms that the initial arrest is G2 and not mitotic. Second, the absence of HR in the Rad51d-deficient cells heightened the G2 arrest induced by 6-thioG, suggesting that HR intermediates cause the G2 arrest. Third, the HR intermediates are MMR dependent, that is, MMR is generating the substrate that activates HR. Fourth, the G2 arrest observed does not depend on RAD51D because the arrest is heightened in the Rad51d-deficient cells.

Figure 3.

Rad51d-deficient cells display a G2 arrest after treatment with 6-thioG. Following a 6-hour treatment with 0.1 μmol/L 6-thioG, the cells were harvested at the 48- and 72-hour time points posttreatment for Rad51d-proficient cells (A and B) and Rad51d-deficient cells (C and D). Nocodazole (Nz) was added 6 hours before harvesting the cells and the mitotic index was calculated. Each data point represents the mean of 3 independent experiments. Error bars, SEMs. Statistical significance was determined by ANOVA followed by Bonferroni's multiple comparison's test. **, P < 0.01; ***, P < 0.001.

Figure 3.

Rad51d-deficient cells display a G2 arrest after treatment with 6-thioG. Following a 6-hour treatment with 0.1 μmol/L 6-thioG, the cells were harvested at the 48- and 72-hour time points posttreatment for Rad51d-proficient cells (A and B) and Rad51d-deficient cells (C and D). Nocodazole (Nz) was added 6 hours before harvesting the cells and the mitotic index was calculated. Each data point represents the mean of 3 independent experiments. Error bars, SEMs. Statistical significance was determined by ANOVA followed by Bonferroni's multiple comparison's test. **, P < 0.01; ***, P < 0.001.

Close modal

The cell-cycle data also revealed what appeared to be a substantial induction of cells with more than 4n DNA content, so the number of cells with the more than 4n DNA content was quantified (Supplementary Fig. S2A). At the 72-hour time point, Rad51d+/+Mlh1+/+ cells show 1.8-fold increase in the number of cells with more than 4n DNA content whereas Rad51d−/−Mlh1+/+ cells show a 3.1-fold induction in the aneuploid population (Fig. 4). Rad51d−/− Mlh1−/− cells did not show an increase in the aneuploid population, confirming the MMR dependence of the damage. The apparent induction of aneuploidy compelled us to examine the nuclear integrity and chromosomal instability in the different genotypes.

Figure 4.

6-thioG induces the appearance of cells with higher 4n DNA content in Rad51d-deficient cells. Rad51d+/+Mlh1+/+, Rad51d−/−Mlh1+/+, and Rad51d−/−Mlh1−/− cells were treated with 0.1 μmol/L 6-thioG for 6 hours. Cells were harvested at 24, 48, and 72 hours posttreatment. The number of cells containing more than 4n DNA content was quantified using the MODFIT-LT software, and fold induction was calculated by normalizing the data of the treated population to that of the untreated controls. The data represent the mean of 3 independent experiments. Error bars, SEMs. Statistical significance was determined by ANOVA followed by Bonferroni's multiple comparison's test. *, P < 0.05; **, P < 0.01.

Figure 4.

6-thioG induces the appearance of cells with higher 4n DNA content in Rad51d-deficient cells. Rad51d+/+Mlh1+/+, Rad51d−/−Mlh1+/+, and Rad51d−/−Mlh1−/− cells were treated with 0.1 μmol/L 6-thioG for 6 hours. Cells were harvested at 24, 48, and 72 hours posttreatment. The number of cells containing more than 4n DNA content was quantified using the MODFIT-LT software, and fold induction was calculated by normalizing the data of the treated population to that of the untreated controls. The data represent the mean of 3 independent experiments. Error bars, SEMs. Statistical significance was determined by ANOVA followed by Bonferroni's multiple comparison's test. *, P < 0.05; **, P < 0.01.

Close modal

6-thioG induced multinucleation in the absence of HR

We examined the cells by DAPI and β-tubulin staining to directly visualize nuclear integrity. Interestingly, exposure to 6-thioG caused multinucleation in Rad51d-proficient and -deficient cells in the MMR-proficient background (Fig. 5; see also representative images in Supplementary Fig. S3). The percentages of multinucleated cells were the same in untreated and treated cells at 24 hours posttreatment, although the basal level of multinucleated cells in the Rad51d-deficient cells was 2.3-fold higher (P < 0.05). At 48 hours, Rad51d-proficient cells show a 3.3% increase in multinucleated cells compared with untreated cells whereas Rad51d-deficient cells show a 7.6% increase. By 72 hours, the Rad51d-proficient cells show a 6.8% increase whereas the Rad51d-deficient cells show an 18.0% increase in multinucleated cells compared with the untreated cells (Fig. 5A and B). In the MMR-deficient background, Rad51d-deficient cells show only a 2.0% increase in multinucleated cells compared with the untreated cells, demonstrating the dependence on MMR (Fig. 5C; P > 0.05).

Figure 5.

6-thioG treatment induces increased multinucleation in Mlh1-proficient and Rad51d-deficient cells. Rad51d+/+Mlh1+/+ (A), Rad51d−/−Mlh1+/+ (B), and Rad51d−/−Mlh1−/− (C) cells were treated with 0.1 μmol/L 6-thioG for 6 hours and harvested at 24, 48, and 72 hours posttreatment. Percentages of multinucleated cells were calculated from at least 500 cells per sample. Error bars represent the 95% CIs. Statistical significance was determined by z test. *, P < 0.05; ***, P < 0.001.

Figure 5.

6-thioG treatment induces increased multinucleation in Mlh1-proficient and Rad51d-deficient cells. Rad51d+/+Mlh1+/+ (A), Rad51d−/−Mlh1+/+ (B), and Rad51d−/−Mlh1−/− (C) cells were treated with 0.1 μmol/L 6-thioG for 6 hours and harvested at 24, 48, and 72 hours posttreatment. Percentages of multinucleated cells were calculated from at least 500 cells per sample. Error bars represent the 95% CIs. Statistical significance was determined by z test. *, P < 0.05; ***, P < 0.001.

Close modal

Increased 6-thioG induced chromosomal aberrations in the absence of HR

We next examined the influence of HR on chromosomal instability following 6-thioG treatment focusing on MMR-proficient cells, because it has been shown that 6-thioG–induced chromosome aberrations are not observed in MMR-deficient cells (11). In Rad51d-proficient cells, there was a 3.1-fold increase in the number of total aberrations 24 hours posttreatment and a 3.5-fold increase by 48 hours (Fig. 6A). In Rad51d-deficient cells, there was a 3.5-fold increase 24 hours posttreatment but a much larger (9.4-fold) increase at 48 hours posttreatment (Fig. 6B). The fold inductions are calculated by comparing the treated samples with the corresponding untreated control for that genotype. It is important to note that Rad51d-deficient cells have an approximately 5-fold higher background level of total aberrations, so the 9.4-fold increase seen by 48 hours represents an extraordinary level of chromosomal instability induced by the low (0.1 μmol/L) dose of 6-thioG (note different y-axes in Fig. 6A and B). By 72 hours, the fold induction of all chromosomal aberrations seen in the Rad51d-deficient cells remained at 9.5, but there was a dramatic increase in the number of complex exchange–type aberrations. Of particular note was a huge induction of radials (25-fold) in Rad51d-deficient cells treated with 6-thioG compared with a 1.3-fold induction in the Rad51d-proficient cells (Fig. 6C and D). Moreover, there were complex, presumptive interchromosomal intermediates evident only in the Rad51d-deficient cells that were not quantified (examples are shown in Supplementary Fig. S4). Collectively, these data suggest that chromosomal fusions were occurring in the absence of HR.

Figure 6.

Rad51d-deficient cells display increased chromosomal instability after 6-thioG treatment. Total number of aberrations (gaps, breaks, triradials, quadriradials, fusions, and giant marker chromosomes) per chromosome × 10−3 as detected by Giemsa staining of metaphase spreads from (A) Rad51d+/+Mlh1+/+ and (B) Rad51d−/−Mlh1+/+ MEF cell lines treated with 0.1 μmol/L 6-TG for 6 hours. The aberration frequencies were calculated 24, 48, and 72 hours posttreatment. Note different y-axis scales. Triradial and quadriradial aberrations per chromosome × 10−3 as detected by Giemsa staining of metaphase spreads from (C) Rad51d-proficient and (D) Rad51d-deficient MEF cell lines. Error bars represent the 95% CIs. Statistical significance was determined by z test. **, P < 0.01; ***, P < 0.001.

Figure 6.

Rad51d-deficient cells display increased chromosomal instability after 6-thioG treatment. Total number of aberrations (gaps, breaks, triradials, quadriradials, fusions, and giant marker chromosomes) per chromosome × 10−3 as detected by Giemsa staining of metaphase spreads from (A) Rad51d+/+Mlh1+/+ and (B) Rad51d−/−Mlh1+/+ MEF cell lines treated with 0.1 μmol/L 6-TG for 6 hours. The aberration frequencies were calculated 24, 48, and 72 hours posttreatment. Note different y-axis scales. Triradial and quadriradial aberrations per chromosome × 10−3 as detected by Giemsa staining of metaphase spreads from (C) Rad51d-proficient and (D) Rad51d-deficient MEF cell lines. Error bars represent the 95% CIs. Statistical significance was determined by z test. **, P < 0.01; ***, P < 0.001.

Close modal

6-thioG has been shown to increase SCEs, presumptive HR events, in MMR-proficient cells (11). We observed a 1.9-fold increase in SCEs in Rad51d-proficient cells after treatment with 0.1 μmol/L 6-thioG. However, there was no detectable increase in SCEs in Rad51d-deficient cells, demonstrating that the absence of RAD51D has prevented the induction of SCEs by MMR recognition of 6-thioG (Supplementary Fig. S5) and suggesting that HR is indeed defective in Rad51d-deficient cells.

Here, we report that Rad51d-deficient cells are hypersensitive to 6-thioG and that this sensitivity is rescued by deletion of Mlh1. The observations in this study have several important implications for understanding the role of HR in response to the DNA damage caused by thiopurines. The data clearly place HR downstream of MMR and also demonstrate an almost exclusive dependence on MMR for the genotoxic effects of thiopurines. It has been demonstrated that 6-thioG induces a G2 arrest in MMR-proficient cells (22). Rad51d-deficient cells display more G2 accumulation in response to 6-thioG, suggesting that the G2 checkpoint machinery is intact. However, the G2 arrest was transient, and by 72 hours, both Rad51d-deficient and -proficient cells show progression into mitosis. HeLa cells exposed to low dose of MNNG have been previously demonstrated to undergo a transient G2 arrest before entering into mitosis (23). Our results show that the status of RAD51D-dependent HR does not affect the activation of the G2 checkpoint or exit from the G2 arrest. However, dramatic chromosomal instability occurs when the HR-deficient cells enter into mitosis following 6-thioG treatment.

The induction of chromosomal aberrations and SCEs in MMR- and HR-proficient cells provides strong evidence that MMR is causing a direct DNA intermediate that must be resolved by a DSB repair process. In the absence of HR, the chromosomal instability induced by 6-thioG dramatically worsens and appears to invoke a different DSB process. Recent studies suggest that in HR-deficient cells, alternative end-joining pathways tend to join the DNA breaks and produce complex chromosome rearrangements (24). We did not observe an appreciable change in the nuclear localization of 2 representative NHEJ participants, Ku70 and XRCC4 (data not shown), which implies that classical NHEJ is not responsible. However, we did see a statistically significant increase in XRCC1 foci at 72 hours (data not shown), which was the same time point at which the radial induction occurred (Fig. 6). This suggests that an alternative NHEJ pathway is likely triggering the observed fusion events in the absence of RAD51D. The basis of the alternative NHEJ process that results in the formation of radials remains under active investigation.

6-thioG treatment induces centrosome amplification in MMR-proficient cells, which gives rise to aneuploid and multinucleated cells (25). Rad51d-deficient cells also displayed an increase in the aneuploid population and multinucleated cells in response to 6-thioG. Aneuploidy and multinucleation are associated with various leukemias and lymphomas, for example, a cell type associated with the pathology of Hodgkin's lymphoma known as Reed–Sternberg (RS) cells, which coincidently are MMR proficient (26). RS cells are giant multinucleated cells that display hyperdiploid DNA content and the DNA is unequally partitioned among the nuclei. RS cells also show telomere dysfunction and centrosome amplification (27, 28). Coincidently, we have shown that telomere dysfunction and centrosome amplification are associated with Rad51d deficiency (19, 29). It is therefore tempting to speculate that RAD51D insufficiency can drive telomere and chromosomal instability associated with the progression of Hodgkin's cells to RS cells. Recently, XRCC2 (another RAD51 paralogue and also part of a protein complex that includes RAD51D) has been implicated in B-cell development (30). Specifically, the study directly implicated XRCC2 in suppressing genomic instability in B cells caused by activation-induced cytidine deaminase. This article strongly suggests that RAD51-dependent HR is required to suppress the development of B-cell lymphomas. We believe the tumor-suppressing activity of RAD51-dependent HR more broadly extends to therapy-induced malignancies.

A large number of Hodgkin's lymphoma patients develop secondary tumors after chemotherapy (31, 32). Therefore, HR insufficiency could be a risk factor for Hodgkin's lymphoma patients and an additional important consideration of treatment options for these patients because HR status may predispose these patients to the development of secondary cancers resulting from chemotherapy. Although many patients have benefited from thiopurines, there is growing concern for long-term survivors of an increased risk of iatrogenic (therapy-induced) cancer. While the manuscript was in preparation, a study reported that BRCA2-deficient cells are sensitized to 6-thioG (17), which agrees with the findings here in Rad51d-deficient cells. One implication is that HR-defective tumors might be successfully treated with thiopurines. However, there are a number of important caveats that might suggest healthy caution before treating BRCA2 carriers with thiopurines. First, the chromosomal instability induced by 6-thioG treatment in the HR-defective cells was impressive in its extent and complexity. The chromosomal instability observed is highly likely to result from unresolved HR intermediates and the outcomes appear to be chromosomal fusions. The fact that MMR proficiency promotes cell death following thiopurine treatment sets up a strong selection pressure for loss of MMR, certainly in vitro (33). In fact, it has been observed that although microsatellite instability (MSI) associated with MMR deficiency is not associated with the development of de novo hematologic malignancies, the frequency of MMR defects in therapy-related leukemia is high (34). Mlh1 polymorphism has been associated with an increased risk of secondary cancers in Hodgkin's lymphoma patients after methylating chemotherapy (32). This suggests that thiopurine treatment indeed puts a selection for the outgrowth of MMR-defective cells in patients. There is a possibility that HR insufficiency will not only play a role in increasing the selection pressure on these cells to select for MMR-deficiency but also play an important role during the progression of MSI-positive lymphomas by increasing the overall chromosome instability characterizing these cancers. This provides indirect evidence that proficiency in MMR and insufficiencies in HR might predispose such patients for thiopurine-induced leukemias and lymphomas. It also suggests that variations in HR capacity might explain susceptibility to thiopurine-induced cancer.

No potential conflicts of interests were disclosed.

This research was supported in part by a grant from the NIH to the Center for Colon Cancer Research at USC (P20 RR17698), from the NIH/NCI to M.D. Wyatt (R01 CA100450), and from the American Cancer Society to D.L. Pittman (RSG-03-158-01-GMC).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Karran
P
. 
Thiopurines, DNA damage, DNA repair and therapy-related cancer
.
Br Med Bull
2006
;
79–80
:
153
70
.
2.
Beaugerie
L
,
Brousse
N
,
Bouvier
AM
,
Colombel
JF
,
Lemann
M
,
Cosnes
J
, et al
Lymphoproliferative disorders in patients receiving thiopurines for inflammatory bowel disease: a prospective observational cohort study
.
Lancet
2009
;
374
:
1617
25
.
3.
Schmiegelow
K
,
Al-Modhwahi
I
,
Andersen
MK
,
Behrendtz
M
,
Forestier
E
,
Hasle
H
, et al
Methotrexate/6-mercaptopurine maintenance therapy influences the risk of a second malignant neoplasm after childhood acute lymphoblastic leukemia: results from the NOPHO ALL-92 study
.
Blood
2009
;
113
:
6077
84
.
4.
Klapacz
J
,
Meira
LB
,
Luchetti
DG
,
Calvo
JA
,
Bronson
RT
,
Edelmann
W
, et al
O6-methylguanine-induced cell death involves exonuclease 1 as well as DNA mismatch recognition in vivo
.
Proc Natl Acad Sci U S A
2009
;
106
:
576
81
.
5.
Schaetzlein
S
,
Kodandaramireddy
NR
,
Ju
Z
,
Lechel
A
,
Stepczynska
A
,
Lilli
DR
, et al
Exonuclease-1 deletion impairs DNA damage signaling and prolongs lifespan of telomere-dysfunctional mice
.
Cell
2007
;
130
:
863
77
.
6.
Karran
P
. 
Mechanisms of tolerance to DNA damaging therapeutic drugs
.
Carcinogenesis
2001
;
22
:
1931
7
.
7.
O'Brien
V
,
Brown
R
. 
Signalling cell cycle arrest and cell death through the MMR System
.
Carcinogenesis
2006
;
27
:
682
92
.
8.
Stojic
L
,
Brun
R
,
Jiricny
J
. 
Mismatch repair and DNA damage signalling
.
DNA Repair
2004
;
3
:
1091
101
.
9.
Lin
DP
,
Wang
Y
,
Scherer
SJ
,
Clark
AB
,
Yang
K
,
Avdievich
E
, et al
An Msh2 point mutation uncouples DNA mismatch repair and apoptosis
.
Cancer Res
2004
;
64
:
517
22
.
10.
Yang
G
,
Scherer
SJ
,
Shell
SS
,
Yang
K
,
Kim
M
,
Lipkin
M
, et al
Dominant effects of an Msh6 missense mutation on DNA repair and cancer susceptibility
.
Cancer Cell
2004
;
6
:
139
50
.
11.
Armstrong
MJ
,
Galloway
SM
. 
Mismatch repair provokes chromosome aberrations in hamster cells treated with methylating agents or 6-thioguanine, but not with ethylating agents
.
Mutat Res
1997
;
373
:
167
78
.
12.
Rajesh
P
,
Rajesh
C
,
Wyatt
MD
,
Pittman
DL
. 
RAD51D protects against MLH1-dependent cytotoxic responses to O(6)-methylguanine
.
DNA Repair
2010
;
9
:
458
67
.
13.
Rajesh
C
,
Gruver
AM
,
Basrur
V
,
Pittman
DL
. 
The interaction profile of homologous recombination repair proteins RAD51C, RAD51D and XRCC2 as determined by proteomic analysis
.
Proteomics
2009
;
9
:
4071
86
.
14.
Mojas
N
,
Lopes
M
,
Jiricny
J
. 
Mismatch repair-dependent processing of methylation damage gives rise to persistent single-stranded gaps in newly replicated DNA
.
Genes Dev
2007
;
21
:
3342
55
.
15.
Roos
WP
,
Nikolova
T
,
Quiros
S
,
Naumann
SC
,
Kiedron
O
,
Zdzienicka
MZ
, et al
Brca2/Xrcc2 dependent HR, but not NHEJ, is required for protection against O(6)-methylguanine triggered apoptosis, DSBs and chromosomal aberrations by a process leading to SCEs
.
DNA Repair
2009
;
8
:
72
86
.
16.
Tebbs
RS
,
Hinz
JM
,
Yamada
NA
,
Wilson
JB
,
Salazar
EP
,
Thomas
CB
, et al
New insights into the Fanconi anemia pathway from an isogenic FancG hamster CHO mutant
.
DNA Repair
2005
;
4
:
11
22
.
17.
Issaeva
N
,
Thomas
HD
,
Djurenovic
T
,
Jaspers
JE
,
Stoimenov
I
,
Kyle
S
, et al
6-Thioguanine selectively kills BRCA2-defective tumors and overcomes PARP inhibitor resistance
.
Cancer Res
2010
;
70
:
6268
76
.
18.
Tate
EH
,
Wilder
ME
,
Cram
LS
,
Wharton
W
. 
A method for staining 3T3 cell nuclei with propidium iodide in hypotonic solution
.
Cytometry
1983
;
4
:
211
5
.
19.
Smiraldo
PG
,
Gruver
AM
,
Osborn
JC
,
Pittman
DL
. 
Extensive chromosomal instability in Rad51d-deficient mouse cells
.
Cancer Res
2005
;
65
:
2089
96
.
20.
Buermeyer
AB
,
Wilson-Van Patten
C
,
Baker
SM
,
Liskay
RM
. 
The human MLH1 cDNA complements DNA mismatch repair defects in Mlh1-deficient mouse embryonic fibroblasts
.
Cancer Res
1999
;
59
:
538
41
.
21.
Yamane
K
,
Schupp
JE
,
Kinsella
TJ
. 
BRCA1 activates a G2-M cell cycle checkpoint following 6-thioguanine-induced DNA mismatch damage
.
Cancer Res
2007
;
67
:
6286
92
.
22.
Yan
T
,
Desai
AB
,
Jacobberger
JW
,
Sramkoski
RM
,
Loh
T
,
Kinsella
TJ
. 
CHK1 and CHK2 are differentially involved in mismatch repair-mediated 6-thioguanine-induced cell cycle checkpoint responses
.
Mol Cancer Ther
2004
;
3
:
1147
57
.
23.
Schroering
AG
,
Kothandapani
A
,
Patrick
SM
,
Kaliyaperumal
S
,
Sharma
VP
,
Williams
KJ
. 
Prolonged cell cycle response of HeLa cells to low-level alkylation exposure
.
Cancer Res
2009
;
69
:
6307
14
.
24.
Bunting
SF
,
Callen
E
,
Wong
N
,
Chen
HT
,
Polato
F
,
Gunn
A
, et al
53BP1 inhibits homologous recombination in Brca1-deficient cells by blocking resection of DNA breaks
.
Cell
2010
;
141
:
243
54
.
25.
Robinson
HM
,
Black
EJ
,
Brown
R
,
Gillespie
DA
. 
DNA mismatch repair and Chk1-dependent centrosome amplification in response to DNA alkylation damage
.
Cell Cycle
2007
;
6
:
982
92
.
26.
Re
D
,
Benenson
L
,
Wickenhauser
C
,
Starostik
P
,
Staratschek-Jox
A
,
Muller-Hermelink
HK
, et al
Proficient mismatch repair protein expression in Hodgkin and Reed Sternberg cells
.
Int J Cancer
2002
;
97
:
205
10
.
27.
Knecht
H
,
Sawan
B
,
Lichtensztejn
D
,
Lemieux
B
,
Wellinger
RJ
,
Mai
S
. 
The 3D nuclear organization of telomeres marks the transition from Hodgkin to Reed-Sternberg cells
.
Leukemia
2009
;
23
565
73
.
28.
Martin-Subero
JI
,
Knippschild
U
,
Harder
L
,
Barth
TFE
,
Riemke
J
,
Grohmann
S
, et al
Segmental chromosomal aberrations and centrosome amplifications: pathogenetic mechanisms in Hodgkin and Reed-Sternberg cells of classical Hodgkin's lymphoma?
Leukemia
2003
;
17
:
2214
9
.
29.
Tarsounas
M
,
Munoz
P
,
Claas
A
,
Smiraldo
PG
,
Pittman
DL
,
Blasco
MA
, et al
Telomere maintenance requires the RAD51D recombination/repair protein
.
Cell
2004
;
117
:
337
47
.
30.
Hasham
MG
,
Donghia
NM
,
Coffey
E
,
Maynard
J
,
Snow
KJ
,
Ames
J
, et al
Widespread genomic breaks generated by activation-induced cytidine deaminase are prevented by homologous recombination
.
Nat Immunol
2010
;
11
:
820
6
.
31.
Allan
JM
,
Travis
LB
. 
Mechanisms of therapy-related carcinogenesis
.
Nat Rev Cancer
2005
;
5
:
943
55
.
32.
Worrillow
LJ
,
Smith
AG
,
Scott
K
,
Andersson
M
,
Ashcroft
AJ
,
Dores
GM
, et al
Polymorphic MLH1 and risk of cancer after methylating chemotherapy for Hodgkin lymphoma
.
J Med Genet
2008
;
45
:
142
6
.
33.
Karran
P
,
Offman
J
,
Bignami
M
. 
Human mismatch repair, drug-induced DNA damage, and secondary cancer
.
Biochimie
2003
;
85
:
1149
60
.
34.
Karran
P
,
Attard
N
. 
Thiopurines in current medical practice: molecular mechanisms and contributions to therapy-related cancer
.
Nat Rev Cancer
2008
;
8
:
24
36
.