Inactivation of Chk2 and Mus81 Leads to Impaired Lymphocytes Development, Reduced Genomic Instability, and Suppression of Cancer

Chk2 is an effector kinase important for the activation of cell cycle checkpoints, p53, and apoptosis in response to DNA damage. Mus81 is required for the restart of stalled replication forks and for genomic integrity. Mus81Δex3-4/Δex3-4 mice have increased cancer susceptibility that is exacerbated by p53 inactivation. In this study, we demonstrate that Chk2 inactivation impairs the development of Mus81Δex3-4/Δex3-4 lymphoid cells in a cell-autonomous manner. Importantly, in contrast to its predicted tumor suppressor function, loss of Chk2 promotes mitotic catastrophe and cell death, and it results in suppressed oncogenic transformation and tumor development in Mus81Δex3-4/Δex3-4 background. Thus, our data indicate that an important role for Chk2 is maintaining lymphocyte development and that dual inactivation of Chk2 and Mus81 remarkably inhibits cancer.


Introduction
DNA damage response is a result of the coordinated actions of DNA damage signaling and repair pathways, cell cycle checkpoints, and apoptosis [1]. Highlighting the importance of the damage signaling and repair mechanisms, mutations of genes such as ATM, BRCA1 and NBS1, involved in these mechanisms, are associated with increased DNA damage sensitivity, genomic instability, cancer predisposition, immunodeficiency, and developmental defects [2].
Mammalian Mus81 with its partners Eme1 or Eme2 form a heterodimeric structure-specific endonuclease that preferentially cleaves 39 Flaps and replication fork intermediates [3]. This endonuclease has been shown to facilitate restart of stalled DNA replication forks by generating DNA double-strand breaks (DSBs) [4]. Mus81 also interacts with other DNA damage repair proteins including Rad54, Blm, as well as SLX4 [5][6][7]. Interestingly, Cds1, the yeast homolog of the mammalian Serine/Threonine kinase Chk2, was reported to phosphorylate and release mus81 from chromatin, presumably to prevent it from cleaving stalled replication forks (RFs) [8].
Three strains of Mus81 mutant mice have been reported. In addition to Mus81 inactivation, Mus81 Dex1-10 mice have been reported to also display decreased expression of Fibulin-4 gene. Several of these homozygous mutant mice developed cardiovascular complications and died before reaching weaning age [9]. The phenotypes of these mice have been attributed to the decreased expression of Fibulin-4 [9]. Mus81 Dex9-12 mice have been also reported [10]. These mice displayed increased sensitivity to interstrand crosslinking (ICL) agents including MMC. Genomic instability was reported to be increased in homozygous Mus81  MEFs expressing the human papillomavirus type 16 E6 that promotes degradation of p53. While these mutant mice were viable, they showed no increased of tumorigenesis when monitored for a period of 15 months [10]. The Mus81 Dex3-4 mutant mice and cells that we have generated were also highly sensitive to MMC [11]. Mice homozygous for the Mus81 Dex3-4 mutation showed no expression of Mus81 protein, and displayed elevated levels of spontaneous genomic instability and cancer predisposition [11]. While the cause for the lack of tumorigenesis in Mus81 Dex9-12 mutant mice is still not clear, inactivation of p53 in Mus81 Dex3-4/Dex3-4 mice rescued their MMC hypersensitivity and exacerbated their genomic instability and tumorigenesis [12].
Inactivation of MUS81 in human cells also resulted in hypersensitivity to ICL agents and elevated levels of genomic instability [13]. Importantly, MUS81 expression was found significantly decreased in human hepatocellular carcinomas, and this reduced expression correlates with a poor prognosis for patients with this cancer [14]. Moreover, a variant MUS81 allele (rs545500) was recently associated with increased risk for breast cancer [15].
CHK2 plays important roles in the DNA damage response, the signaling of the ATM-CHK2-P53 pathway and in cell cycle checkpoints including G2/M checkpoint [16,17]. CHK2 phosphorylates a number of substrates including p53, CDC25A, CDC25C, BRCA1, E2F1, and MDC1. A role for CHK2 in cancer is supported by its rare germline or somatic mutations in certain human familial cancers and in a number of tumors and by its central role in oncogene-induced senescence [18,19]. Interestingly, mounting evidence also supports the benefit of CHK2 inhibition in promoting tumor killing in response to genotoxic drugs [16].
Given the importance of Chk2 and Mus81 in DNA damage signaling and repair respectively, we have examined the effect of their dual inactivation on lymphoid cell differentiation, DNA damage response and cancer.

Chk2 Deficiency Does Not Affect Embryonic Development of Mus81 Dex3-4/Dex3-4 Mice
In contrast to the female specific embryonic lethality of Mus81 Dex3-4/Dex3-4 mice in p53 deficient background [12], Mus81 Dex3-4/Dex3-4 Chk2 2/2 mice were viable and born at the expected Mendelian ratio (Table S1). We next examined the fertility of Mus81 Dex3-4/Dex3-4 Chk2 2/2 mice and have also further assessed the viability of double mutant females. Interbreeding of Mus81 Dex3-4/Dex3-4 Chk2 2/2 mice as well as their breeding (males and females) to mice from different genotypes resulted in normal size litters compared to control mice (P,0.05; Table S1). In addition, interbreeding of double mutant mice or heterozygous compound mice resulted in the expected ratio of males and females (P,0.05; Table S2). Double mutant males and females were indistinguishable from their wildtype (WT) and single mutant littermates. Examination of the weight of 6 to 8 week old and 4 to 8 month old double mutant males and females indicated no significant differences compared to WT mice and single mutant controls ( Figure S1A).
Collectively, these data indicate that Mus81 Dex3-4/Dex3-4 Chk2 2/2 males and females were viable and fertile, and their gross morphology was indistinguishable from their WT and single mutant littermates.
To examine the impaired homeostasis of peripheral lymphocytes and the imbalance of the ratio CD4 + to CD8 + peripheral Tcells in double mutant mice, we examined the level of spontaneous and activation induced death of these cells. Spontaneous cell death was found significantly increased in both CD4 + and CD8 + naïve splenocytes from double mutant mice compared to controls ( Figure S3A and S3B). In addition, LPS activation of double mutant B-cells also resulted in elevated levels of cell death compared to B-cells from single mutants or WT mice ( Figure S7C). While the number of T-cells and B-cells were reduced in spleen of double mutant mice, the number of macrophages in these spleens remained similar to single mutants and WT controls.
Taken together these data demonstrate a specific role for Chk2 in maintaining homeostasis of peripheral T-and B-cells deficient for Mus81.

Author Summary
Failure to repair DNA damage has been associated with a number of human syndromes, neurodegenerative diseases, immunodeficiency, and cancer. In addition, radiotherapy and many cancer chemotherapeutic drugs induce DNA damage, thus allowing the killing of tumors. Recent data indicated Mus81's role in maintaining genomic integrity and suppressing cancer. Furthermore, inactivation of p53, the most frequently inactivated tumor suppressor in cancer, leads to synergistic tumorigenesis in Mus81 mutant mice. As Chk2 is important for p53 activation, we have examined the effect of its inactivation on the phenotypes associated with Mus81 loss of function. We report that Chk2 is essential for the development of lymphoid cells deficient for Mus81. Chk2 inactivation increased spontaneous cell death of Mus81 deficient cells and impaired the development of T and B-cell lineages. Chk2 inactivation also reduced the frequency of Mus81deficient cells that carry elevated levels of spontaneous genomic instability. Importantly, inactivation of Chk2 protected Mus81 mutant mice from developing spontaneous tumorigenesis. These data indicate potential therapeutic benefits for the inactivation of Chk2 and Mus81.
We next examined the level of expression of TCRb expression in thymocytes from the four genotypes using FACS analysis and antipan TCRb chain constant region, anti-TCRVb4, anti-TCRVb5 and anti-TCRVb17a. Consistent with the normal number of CD4 2 CD8 2 thymocyte population in Mus81 Dex3-4/Dex3-4 Chk2 2/2 mice, double mutant thymocytes displayed no significant difference in the level of TCRb expression or in the percentages of thymocytes expressing these TCRVb compared to controls ( Figure S4A and S4B). Consistent with these data, no difference was observed in TCRb expression between Mus81 Dex3-4/Dex3-4 Chk2 2/2 and control splenocytes ( Figure S4C).
To further evaluate the cause for the depletion of BM cells in the absence of Mus81 and Chk2 we examined the cell death level of B220 + BM cells from double mutant mice and control littermates using Propidium Iodide (PI) staining and FACS analysis. Remarkably, the level of spontaneous cell death of B220 + BM cells was significantly higher in Mus81 Dex3-4/Dex3-4 Chk2 2/2 mice compared to single mutants (P,0.05), and WT (P,0.01) controls ( Figure 2D).
These data indicate that in contrast to the inactivation of either Mus81 or Chk2, their dual inactivation significantly reduces the pool of B-cell precursors. Our data also identify increased cell death as likely to contribute to the defective differentiation of Bcell lineage in Mus81 Dex3-4/Dex3-4 Chk2 2/2 mice. The earlier differentiation defect of the B-cell lineage compared to T-cell lineage of Mus81 Dex3-4/Dex3-4 Chk2 2/2 mice is likely to contribute to the more pronounced imbalance of the ratio B-to T-cells in the periphery of these mice.
Therefore, in contrast to p53 inactivation [12], inactivation of Chk2 failed to rescue MMC hypersensitivity of Mus81 Dex3-4/Dex3-4 T-cells. Given the high in vivo MMC sensitivity of Mus81 Dex3-4/Dex3-4 mice [11] and Mus81 Dex9-12 mice [10], we assessed the effect of Chk2 inactivation on the MMC hypersensitivity of Mus81 mutant mice. Cohorts of mice from the four genotypes received intraperitoneal (i.p) injection of MMC (12.5 mg/kg) and were monitored for survival ( Figure 3E). In contrast to the results of the in vitro BM colony forming assay suggestive of equivalent MMC sensitivity of double mutant and Mus81 Dex3-4/Dex3-4 BM cells, double mutant mice displayed significantly higher sensitivity to MMC (mean survival = 6 days) than Mus81 Dex3-4/Dex3-4 mice (mean survival = 9 days).

Effect of Chk2 Inactivation on MMC Hypersensitivity of
As BM failure might contribute to the elevated MMC sensitivity of Mus81 Dex3-4/Dex3-4 Chk2 2/2 mice, we examined the in vivo effects  of MMC on BM cell populations from Mus81 Dex3-4/Dex3-4 Chk2 2/2 mice, single mutants and WT littermates. Mice were either left untreated or subjected to i.p injection of MMC (12.5 mg/kg) and the level of death of B220 + BM cells was examined three days later using PI staining. Increased cell death was observed in B220 + BM cells from both untreated and MMC treated double mutant mice compared to control littermates ( Figure 3F).
Increased Mitotic Catastrophe of Mus81 Dex3-4/Dex3-4 Chk2 2/2 Cells Mitotic catastrophe is an abnormal mitosis that triggers the death of cells with damaged DNA as they enter mitosis [22]. Mitotic catastrophe is also triggered by agents that affect the stability of microtubules as well as by mitotic failure that results from impaired cell cycle checkpoints. Chk2, which is important for G2/M checkpoint, has been reported to negatively regulate mitotic catastrophe of DNA damaged cells [23]. As Mus81 Dex3-4/Dex3-4 Chk2 2/2 cells displayed increased levels of spontaneous cell death and these mutant cells and mice also displayed elevated MMC-induced cell death, we have examined whether cell death of Mus81 Dex3-4/Dex3-4 Chk2 2/2 cells is triggered by mitotic catastrophe. Cells undergoing mitotic catastrophe typically display micronuclei, increased frequency of giant cells, multilobed nuclei, and nuclear bridging [22].
Loss or reduction of the expression of survivin [24], an inhibitor of apoptosis protein (IAP), has been shown to lead to cell death by mitotic catastrophe [25]. We have therefore performed Western blot analyses to examine the levels of survivin in untreated or 18 hr and 24 hr post-MMC treatment of LPS activated B-cells from Mus81 Dex3-4/Dex3-4 Chk2 2/2 mice, single mutants and WT controls. We observed a lower level of survivin in MMC treated Mus81 Dex3-4/Dex3-4 Chk2 2/2 B-cells compared to single mutants and WT controls ( Figure S8). We propose that decreased survivin level in Mus81 Dex3-4/Dex3-4 Chk2 2/2 cells could contribute to their mitotic catastrophe.
Increased mitotic catastrophe would trigger cell death of Mus81 Dex3-4/Dex3-4 Chk2 2/2 cells and therefore is likely to contribute to the impaired homeostasis of BM cells, thymocytes and splenocytes in double mutant mice.
Impaired of Activation of p53 and Defective G2/M Checkpoint in Mus81 Dex3-4/Dex3-4 Chk2 2/2 Cells p53 plays a central role in DNA damage responses and its phosphorylation by Chk2 on Serine 20 is important for its stability and activation [26]. Therefore, we examined the level of p53 expression and activation in double mutants and control B-cells from the four genotypes. While LPS activated double mutant Bcells displayed increased cell death levels ( Figure 5A), the basal level of p53 in these activated B-cells remained similarly low compared to single mutants and WT B-cells ( Figure 5B). MMC treatment of WT and Mus81 Dex3-4/Dex3-4 activated B-cells increased p53 expression levels ( Figure 5B). However, double mutant and Chk2 2/2 cells failed to increase their level of p53 in response to MMC treatment ( Figure 5B). In accordance with the low level of p53 in untreated and MMC treated double mutant cells, the level of Serine 15-p53 (a substrate for ATM), and the levels of Bax and p21 (p53 downstream targets) were low to undetectable in double mutant cells under untreated conditions and were not induced in response to MMC ( Figure 5C). In contrast, the expression levels of Serine 15-p53, Bax and p21 in Mus81 Dex3-4/Dex3-4 and WT cells were significantly increased in response to MMC ( Figure 5C). These data indicate that despite the lack of p53 activation in double mutant B-cells, these cells displayed elevated level of spontaneous and MMC-induced death, supporting that this death is p53-independent. Consistent with these data, mitotic catastrophe has been shown to take place in a p53-independent manner [27].
Chk2 is required for the enforcement of the G2/M cell cycle checkpoint following DNA damage [16,18]. It has been shown that loss of Chk2 expression or inhibition of its kinase activity sensitizes cells to death during mitosis [23]. Therefore, we have examined G2/M cell cycle checkpoint in untreated and MMCtreated double mutant and control activated B-cells ( Figure 5D). While the fraction of mitotic cells (positive for phospho-Histone H3) was reduced 18 hr and 24 hr post MMC treatment in WT and Mus81 Dex3-4/Dex3-4 B-cells, both Chk2 2/2 and Mus81 Dex3-4/Dex3-4 Chk2 2/2 B-cells failed to activate G2/M checkpoint in response to MMC treatment as indicated by their increased fraction of phospho-Histone H3 positive cells compared to untreated controls ( Figure 5D).
These data indicate that despite MMC-induced DNA damage, double mutant cells are allowed to enter mitosis, due to their impaired G2/M checkpoint, therefore resulting in mitotic catastrophe and cell death.
The increased genomic instability of MMC treated

Inactivation of Chk2 Suppresses Tumorigenesis of Mus81 Dex3-4/Dex3-4 Mutants
CHK2 is mutated in certain human familial cancers and in a number of other tumors [18]. Tumorigenesis was also observed in a mouse model for CHK2 del1100C, a mutation associated with increased cancer risk in humans [31]. In addition, inactivation of Chk2 increased cancer risk of Brca1, Nbs1 and Mre11 mutant mice [30]. Stracker et al. demonstrated that inactivation of Chk2 in DNA-PKcs deficient background did not predispose these mice for cancer, and proposed that Chk2 suppresses the oncogenic potential of DNA damage arising in the S and G2 but not G1 phases of the cell cycle [30].
The remarkable mitigation of tumorigenesis of Mus81 Dex3-4/Dex3-4 mice by Chk2 inactivation parallels their reduced level of spontaneous genomic instability, failure of G2/M checkpoint and increased mitotic catastrophe. Our finding is consistent with a beneficial effect of CHK2 inhibition on tumor responses to genotoxic chemotherapeutic drugs [18].

Chk2 Inactivation Suppresses Oncogenic Transformation of Mus81 Dex3-4/Dex3-4 Cells
Dysregulation of oncogenes, similar to tumor suppressors, plays critical roles in human cancer. Chk2 is activated by oncogenes such as Ras, and is important for mediating oncogene-induced senescence, a safeguard against cancer development [32]. In addition, Chk2 inactivation has been shown to promote oncogenic transformation [19].
We therefore investigated the effect of Chk2 inactivation on transformation of Mus81 Dex3-4/Dex3-4 cells and examined oncogenic transformation of double mutant primary MEFs, single mutants and WT controls. Cells were infected with a retrovirus expressing E1A and Ras and transformed colonies were counted three weeks later ( Figure 7B and 7C). Consistent with previous studies [19], the number of E1A-Ras transformed colonies formed by Chk2 2/2 MEFs was elevated compared to WT controls (P,0.05). Strikingly, while Mus81 Dex3-4/Dex3-4 MEFs were able to form E1A-Ras transformed colonies at a similar frequency compared to WT MEFs (P = 0.36), no E1A-Ras transformed colonies were obtained from Mus81 Dex3-4/Dex3-4 Chk2 2/2 MEFs in three independent experiments.
To confirm that the lack of transformation of Mus81 Dex3-4/Dex3-4 Chk2 2/2 MEFs is due to the loss of Mus81 and Chk2 and not due to the inability of these MEFs to be infected with pBabe E1A-HRasV12 retrovirus, MEFs from the four genotypes were infected with this retrovirus and genomic DNA of these cells was prepared 7 days post puromycin selection and PCR analysis was performed to assess the presence of E1A-HRasV12 in the extracted genomic DNA. These analyses demonstrated that despite integration of These data demonstrate that similar to the suppression of spontaneous tumorigenesis of Mus81 Dex3-4/Dex3-4 mice, Chk2 inactivation also suppresses oncogenic transformation of Mus81 Dex3-4/Dex3-4 cells.
Collectively our data demonstrate an important role for Chk2 in maintaining homeostasis of BM cells, thymocytes and splenocytes deficient for Mus81. Remarkably, inactivation of Chk2 also reduced spontaneous genomic instability associated with Mus81 mutation, and significantly protected Mus81 mutants from tumorigenesis and oncogenic transformation. We also report increased mitotic catastrophe in double mutant cells, likely due to the inactivation of G2/M checkpoint associated with loss of Chk2 and increased spontaneous DNA damage associated with loss of Mus81. The phenotypes observed in Mus81 Dex3-4/Dex3-4 Chk2 2/2 mice are p53-independent. In contrast to Mus81 Dex3-4/Dex3-4 Chk2 2/2 cells, Mus81 Dex3-4/Dex3-4 p53 2/2 cells retain proficient G2/M checkpoint that prevents their entry to mitosis in the presence of damaged DNA. Therefore, we propose that G2/M checkpoint failure and increased mitotic catastrophe are the mechanisms that result in reduced spontaneous genomic instability, tumorigenesis and oncogenic transformation.
Our data provide in vivo evidence that inhibition of Chk2 can have remarkable inhibitory effects on tumorigenesis. While pharmacological inhibitors of CHK2 are being considered for cancer therapy [18], our data suggest that the therapeutic effect of such inhibitors is likely to depend on the genetic background of the tumors. Namely, while Chk2 inactivation promotes tumorigenesis of Brca1, Nbs1 and Mre11 mutant strains, it remarkably suppresses cancer in Mus81 Dex3-4/Dex3-4 mutant background. Our preclinical data highly support the potential therapeutic value for CHK2 or MUS81 inhibitors for cancer patients with MUS81 or CHK2 mutations, respectively.

Cell Cycle Analysis
Peripheral T-cells stimulated with anti-CD3 and IL-2 (50 U/mL) for 48 h, were treated with 0.5 mg/ml MMC (Sigma) for 18 hr. Cells were washed three times with PBS and cultured for an additional 18 hr. Cells were fixed in 70% ethanol, and DNA was stained with 5 mg/mL of PI (Sigma). Cells at the G1, S, and G2/M phases of the cell cycle were determined using FLOWJO analysis software.

Phospho-Histone H3 Staining
Peripheral B-cells activated with LPS (10 mg/ml) for 48 hr, were treated with 0.1 mg/ml of MMC (Calbiochem) for 18 hr. Cells fixed and permeabilized with ice cold methanol were stained with anti-phospho Histone-H3 ser10 FITC (Cell Signaling). Cells were analysed using FACS Calibur (Becton Dickinson) and results were analysed using FLOWJO analysis software.

CFSE Staining
Peripheral T-cells were stained with 5 mM CFSE and then activated for two days with plate bound anti-CD3 in the presence of IL-2. Untreated and 0.1 mg/ml MMC treated cells were grown for an additional 72 hr and analyzed by flow cytometry.

MEFs and Retrovirus Infection
Primary MEFs (3610 5 ) were cultured in DMEM plus 10% FCS in the presence of pBabe E1A-HRasV12 retrovirus and 8 mg/ml of polybrene. At day three post-infection, Puromycin selection (2 mg/ ml) was carried out for three weeks. Colonies on the plates were fixed with ice cold methanol, stained with 0.5% crystal violet and counted. To confirm that MEFs have integrated E1A-HRasV12, genomic DNA was prepared from infected cells post 7 days of puromycin selection. PCR was then performed on MEFs genomic DNA using specific primers for HRasV12 (F: CGGAATA-TAAGCTGGTGGTG and R: CGGTATCCAGGATGTC-CAAC). Eme2 primers (F: ACGGCTTCCCTACCAGCACA and R: AGTGGCTGCTACTCGGCTTCA) were used as controls for the PCR.

Chromosomal Aberrations Analysis
LPS (10 mg/ml) activated splenocytes were cultured for 48 hr in the presence or absence of MMC (40 ng/ml) and metaphase spreads prepared as previously described [11]. Chromosomal aberrations were determined for a minimum of 60 metaphase spreads per cell type. In mFISH, all 21 chromosomes are each painted in a different color using combinatorial labeling and mFISH probe kit (MetaSystems) as previously described [34].
T-cells were activated with anti-CD3 (10 mg/mL) + IL2 (50 U/ mL) for 48 hr in the presence or the absence of MMC (40 ng/ml). Metaphase spreads were prepared as described [11]. Chromosomal aberrations were determined for a minimum of 100 metaphase spreads.
Metaphase spreads of primary MEFs were similarly prepared and chromosomal aberrations were determined for a minimum of 50 metaphase spreads.
Percent aberrations are calculated as aberrations per metaphase then percentages are determined. Breaks/Fragments include chromosome breaks, chromatid breaks, and chromosome fragments. Structural aberrations include chromosome fusions (such as end to end fusions, Robertsonian fusion like configurations).

Cell Death Assays
Cell death was assessed using 7AAD, PI or Annexin V-PI staining and FACS analysis as described [12].

Bone Marrow Transplantation
Bone marrow cells were harvested from the femur of 8 week old mice. 1610 6 cells were transplanted to 6 week old Rag1 2/2 mice by tail vein injection. The transplanted mice were scarified and analysed 8 weeks post-transplantation.

In Vivo Sensitivity to MMC-Induced Damage
Mice (n = 10) for each genotype were injected (i.p) with 12.5 mg MMC/Kg of body mass and observed for two weeks post treatment. Mice were sacrificed when they became moribund and the day of sacrifice was counted as day of death.

Statistical Analysis
The two tailed unpaired student's t test was used for statistical analysis except for survival curves where Log Rank test was employed (Prism 5, GraphPad Software).

Ethics Statement
All experiments were performed in compliance with Ontario Cancer Institute animal care committee guidelines. Chk2 -/mice were exposed to IR (4 Gy) and 12hr later cell death was examined using FACS analysis 7-AAD. Data presented is normalized to UT cells and bar graphs show means 6 SEM. At least three independent experiments using one mouse per group were performed. ns: not statistically significant.