Defining ATM-Independent Functions of the Mre11 Complex with a Novel Mouse Model Free

The maintenance of genome stability and suppression of malignancy depend on the DNA damage response (DDR), a network of pathways comprising signal transduction, cell-cycle regulation, and DNA repair (1). The Mre11 complex, composed of Mre11, Rad50, and Nbs1, plays a central role in the DDR. In addition to sensing DNA double-strand breaks (DSB) and promoting DNA repair, the complex is required for Ataxia-telangiectasia mutated (ATM) kinase activation and signaling to govern DNA damage checkpoints and apoptosis (2).

Ataxia-telangiectasia (A-T) and Nijmegen breakage syndrome (NBS) are two distinct, but related single-gene disorders caused by mutation in ATM and NBS1, respectively. A-T and NBS patients share several clinical and cellular phenotypes characterized by immunodeficiency, sterility, radiosensitivity, and cancer predisposition (3, 4). These phenotypic similarities highlight the functional relationship between the Mre11 complex and ATM.

ATM and the Mre11 complex are involved in specialized DSB repair mechanisms, including V(D)J recombination and class switching (2, 3, 5). Most of the studies concerning the role of the Mre11 complex in these processes rely on effects observed upon deletion of one of its subunits (refs. 6, 7). However, it is conceivable that abnormal cell-cycle progression, proliferation, and increased mortality consequent to Mre11 complex ablation have impeded an appropriate characterization of its influence, as well as its relationship with ATM, in the repair of physiologic breaks.

Despite phenotypic similarities between the Mre11 complex and ATM mutants, and the requirement for the Mre11 complex to activate ATM, several lines of evidence indicate that the Mre11 complex exerts ATM-independent functions. First, ATM-null mice are viable, whereas loss of any subunit of the Mre11 complex is lethal in cultured cells and in vivo (4). Second, ATM deficiency is synthetically lethal with Mre11 complex hypomorphic mutations (Nbs1^ΔB and Mre11^ATLD1; refs. 8–10). We consider two nonexclusive possibilities for the embryonic lethality observed. First, reduction of checkpoint functions ensuing from ATM deficiency may be lethal in combination with DNA repair defects associated with Nbs1 hypomorphism. Second, the Mre11 complex may influence ATM-independent DDR signaling. In this scenario, a decrement in ATR or DNAPKcs activation may be incompatible with the viability of ATM-deficient embryos.

Recent data suggest that the Mre11 complex may influence the activation of ATR (11, 12). These data predict roles for the Mre11 complex in the cellular response to DNA replication stress, which is mediated predominantly by ATR, in addition to its role in activating ATM in response to DNA damage (13, 14). However, the significance of Mre11 complex–dependent regulation of ATR has not been assessed in vivo. Here, we crossed a conditional ATM mutant, ATM^flox (15), with Nbs1^ΔB/ΔB mice to circumvent the synthetic embryonic lethality of Atm^−/− Nbs1^ΔB/ΔB double mutants and create a context in which the ATM-independent functions of the Mre11 complex could be assessed in vivo.

We used a vav^cre mouse (16) to ablate ATM in hematopoietic stem cells of Nbs1^ΔB/ΔB Atm^flox/− mice. Nbs1^ΔB/ΔB Atm^{−/− VAV} mice exhibited defects in lymphoid development, with impaired developmental progression during class switching (CSR), reflective of a severe defect in DSB repair during CSR. Double-mutant mice also displayed increased spontaneous chromosomal instability and completely penetrant lymphomagenesis by 8 months of age. These data support the view that the DNA repair functions of the Mre11 complex are acutely required in the absence of ATM. Conversely, the DDR signaling phenotype of Nbs1^ΔB/ΔB Atm^−/− mice is not consistent with the Mre11 complex exerting a strong influence on ATR activation in vivo. These data suggest that synergy of Mre11 complex hypomorphism and ATM deficiency primarily reflects the importance of the Mre11 complex as an effector of DNA repair.

Animal use protocols were approved by the Institutional Animal Care and Use Committee of the Memorial Sloan-Kettering Cancer Center. ATM^flox and vav^cre mice were obtained from F. Alt (Howard Hughes Medical Institute, The Children's Hospital, Boston) and M. Stadfeld (NYU School of Medicine, New York), respectively. Nbs1^ΔB mouse was described (8). Mice were housed in ventilated rack caging in a pathogen-free facility and genotyped by PCR (details upon request).

Murine embryo fibroblasts (MEF) were generated, cultured, and immortalized as described (17). SV40-immortalized Nbs1^ΔB/ΔB Atm^−/− MEFs were generated from Nbs1^ΔB/ΔB Atm^flox/− mice by infection with cre-green fluorescent protein vector-based lentivirus. Lentiviral production, concentration, and titering were carried out using previously described methods (18, 19). A total of 5 × 10⁴ cells were resuspended in 3 mL of DMEM supplemented with 10% cosmic calf serum (CCS) containing 5 μg/mL polybrene and cre-lentivirus at a multiplicity of infection of 10 followed by clonal selection. Positive clones were identified by PCR, and suppression of ATM gene product expression was confirmed by Western blot. Atm^−/− genotyping was carried out with the following primers: ATMgF86723 (5′ ATCAAATGTAAAGGCGGCTTC 3′) and ATM BAC7 (5′ GCCCATCCCGTCCACAATATCTCTGC 3′). The 903-bp product corresponding to ATM deletion was detected by agarose gel electophoresis.

Murine splenic B cells were isolated using CD43 Microbeads (Miltenyi Biotec) and cultured at a density of 1 × 10⁶ cells/mL in RPMI 1640 supplemented with 15% (v/v) FBS, 100 U/mL of penicillin, and 0.1 mg/mL of streptomycin, 1% (v/v) glutamine, and 10 mmol/L β-mercaptoethanol. Stimulation for CSR was achieved with 1 μg/mL anti-CD40 antibody (eBioscience) and 12.5 μg/mL IL4 (R&D Systems).

Lymphocyte populations were analyzed by flow cytometry in single-cell suspensions from thymus and bone marrow. Cells were depleted of red blood cells by hypotonic lysis and maintained in PBS. Labeled antibodies specific for CD45R/B220 (phycoerythrin), IgM (FITC), CD43 (FITC), Cd11b (phycoerythrin), Gr-1 (FITC), TER119 (FITC), CD8a (FITC), and CD4 (phycoerythrin) were from BD Biosciences PharMingen. Dead cells were excluded by 4′,6-diamidino-2-phenylindole (DAPI) staining. The data were collected on an LSR-Fortessa flow cytometer (Becton Dickinson) and were analyzed with FlowJo software (TreeStar).

For flow cytometric analyses of CSR, cells were resuspended in PBS with 2% BSA and stained with APC-conjugated anti-IgG1 antibody (X56; BD Pharmingen).

Cell proliferation by SNARF labeling was performed as described (20).

Germline transcripts were analyzed as described (21).

For the analysis of switch junctions, genomic DNA was prepared from B cells stimulated with αCD40/IL4 for 72 hours. Sμ–Sγ1 junction DNA was amplified by PCR (38 cycles) using Sμ and Sγ1 primers (22). The PCR products, which spanned from 1 to 2 kb, were gel-extracted, cloned, sequenced, and analyzed as previously described (20).

Immuno-FISH was carried out as previously described (23).

We detected the 3′ end of the IgH locus using BAC199 and the 5′ end using BAC207 (a gift of Fred Alt lab, Howard Hughes Medical Institute, The Children's Hospital, Boston). BAC 97L3 and BAC 307011 that mark either side of the common fragile site (CFS) were used for Fra8E1.

Chromosome spread preparation from splenocytes and SV40-immortalized MEFs were performed as previously described (17). Briefly, cells were incubated in 50 ng/mL colcemid (KaryoMAX; GIBCO), harvested, hypnotically swelled with 0.075 mol/L KCl for 15 minutes at 37°C, fixed, washed in ice-cold 3:1 methanol:acetic acid, and dropped on slides. Slides were stained with 5% Giemsa (Sigma) for 10 minutes and rinsed with distilled water, and coverslips were mounted with Permount (Fisher). Images were captured using an Olympus IX60 microscope and imaged with a Hammamatsu CCD camera. More than 50 spreads were scored for each sample. Two-color FISH was performed as described (24).

Metaphases from thymic lymphomas were obtained as described (25). Spectral karyotyping was performed as per instructions (Applied Spectral Imaging, ASI). Slides were examined with a BX61 microscope (×600 magnification) from Olympus controlled by a LAMBDA 10-B Smart Shutter (Sutter Instrument). Images were captured using a LAMBDA LS light source (Sutter) and a COOL-1300QS camera (ASI) and analyzed by Case Data Manager Version 5.5 (ASI).

Thymi were harvested in 10 mL of RPMI/3% FCS from 6- to 9-week-old animals and washed twice before plating. For each genotype and treatment, 1 × 10⁶ thymocytes were plated in triplicate in 1 mL of supplemented RPMI. Twenty hours after the mock or 5 Gy ionizing radiation (IR) treatments, cells were harvested for Annexin V–FITC staining (BD Biosciences). Cells were analyzed within 3 hours after staining on an LSR-Fortessa flow cytometer (Becton Dickinson) and were analyzed with FlowJo software (TreeStar).

Western blots were carried out on 40 μg of protein extracted with NTEN (20 mmol/L Tris at pH 8, 150 mmol/L NaCl, 1 mmol/L EDTA, 0.5% NP-40) plus protease and phosphatase inhibitors. All of the antibodies were incubated overnight at 4°C in 2% milk. The antibodies used were anti-Rad50, -Nbs1, and -Mre11 polyclonal (26), ATM (Cell Signaling Technology), RPA32 pSer4/Ser8 (Bethyl Laboratories), Chk1 (C9358; Sigma), Chk1 pSer345 (Cell Signaling Technology), AID (27), GAPDH (6C5; Millipore), and Smc1 (Cell Signaling Technology).

To define ATM-independent functions of the Mre11 complex, we established Nbs1^ΔB/ΔB Atm^flox/− double-mutant mice. Cre-mediated deletion of the residual wild-type ATM allele (ATM^flox) was effected in immortalized fibroblasts in vitro via infection with cre-encoding lentivirus, and in vivo by vav^cre, the expression of which is restricted to hematopoietic stem cells (Fig. 1A and Supplementary Fig. S1A).

Although Nbs1^ΔB/ΔB Atm^−/− embryos were inviable (8), we determined that viable Nbs1^ΔB/ΔB Atm^−/− MEFs could be obtained. Immortalized Nbs1^ΔB/ΔB Atm^flox/− fibroblasts were infected with a cre-expressing lentivirus and plated for colony formation. From approximately 80 colonies, seven Nbs1^ΔB/ΔB Atm^−/− colonies were obtained, and their DDR functions were characterized.

The response to IR was examined first. The growth properties of double-mutant MEFs precluded a conventional colony-forming assay; double-mutant cells were essentially unable to form colonies even without IR, whereas they were readily obtained in the Atm^−/− single mutant (Supplementary Fig. S1B). This phenotype was not due to a defect in cell growth because both single- and double-mutant cells showed a comparable growth profile (data not shown).

As an alternative, we monitored the appearance and disappearance of nuclear foci formed by single-strand DNA-binding proteins Replication protein A (RPA) and Rad51 as an index of DNA repair capacity (Fig. 1Bi, ii, iii, and iv). Following exposure to IR (4 Gy), RPA and Rad51 foci persisted longer in Nbs1^ΔB/ΔB Atm^−/− clones than any other genotype. An average of 70% of Nbs1^ΔB/ΔB Atm^−/− cells exhibited RPA nuclear foci 5 hours after treatment, with a 10% reduction observed in Atm^−/− cells. Sixteen hours after IR treatment, RPA foci were present in virtually 100% of two independent double-mutant clones, whereas the RPA signal was reduced to 20% in WT and Nbs1^ΔB/ΔB, and 34% of Atm^−/− cells (Fig. 1C). A similar result was obtained for Rad51 foci (Fig. 1D). These data suggest that DSB repair is markedly defective in Nbs1^ΔB/ΔB Atm^−/− double mutants relative to Atm^−/− or Nbs1^ΔB/ΔB single mutants.

Vav^cre-mediated deletion of the ATM^flox allele in the bone marrow of Nbs1^ΔB/ΔB Atm^flox/− mice was carried out to examine the phenotype of Nbs1^ΔB/ΔB Atm^−/− double mutants in a physiologic setting. The bone marrow was chosen for this analysis because the Mre11 complex and ATM influence developmental stages that rely upon chromosome rearrangements initiated by programmed induction of DSBs—V(D)J recombination and immunoglobulin class switching (28, 29).

The development of T and B cells was examined. Five-week-old double-mutant mice showed approximately 10-fold reduction in the cellularity of the thymus compared with Atm^{−/− VAV} (Fig. 2Ai and Supplementary Fig. S2A). Flow cytometric analysis also revealed significant alterations in the developmental distribution of Nbs1^ΔB/ΔB Atm^{−/− VAV} thymocytes. The stages of T-cell development can be differentiated by the individual presence or coincidence of CD4 and CD8 surface receptors; double negative (DN) are most primitive, double positive (DP) intermediate, and single positive (SP) is most mature. A 3-fold accumulation in the percentage of DN cells (15%) was observed in Nbs1^ΔB/ΔB Atm^{−/− VAV} thymus compared with Atm^{−/− VAV} (6%). This result was accompanied by a concomitant reduction in the percentage (average reduction of 1.2-fold) and the cell number of DP thymocytes, whereas no variation in the percentage of SP cells was evident (Fig. 2B and Supplementary Fig. S2A). These data suggest a temporal effect on T-cell development at the transition from the DN to DP stages.

ATM inhibition in Mre11^ATLD1/ATLD1 lymphocytes leads to the accumulation of unrepaired DSB ends induced during V(D)J recombination (30). We reasoned that the developmental delays observed may reflect an analogous block in V(D)J recombination in both T and B lineages. Thymocytes were stained for CD25 and CD44 surface markers to pinpoint the DN stage at which maturation defect occurs. DN T-cell maturation is further subdivided into four stages of differentiation (DN1, DN2, DN3, and DN4) based on CD44 and CD25 expression. At this early developmental stage, the progression through DN2 to DN4 depends upon successful V(D)J recombination events (31). We found that ATM deletion led to an accumulation at the DN3 stage in Nbs1^ΔB/ΔB cells (39% vs. 13% of Atm^{−/− VAV}) with a concomitant attrition of the DN4 population (44% vs. 80%, double vs. single mutant, respectively; ref. 32; Fig. 2C).

Similarly, flow cytometric analysis of bone marrow revealed a lag in B-cell development at the transition from pro- to pre-B cells, a point at which immunoglobulin gene rearrangement is initiated; a 2-fold reduction in the number of pre-B cells in Nbs1^ΔB/ΔB Atm^{−/− VAV} was observed (Fig. 2D). In contrast, no variation in cellularity or the relative numbers of erythroid and myeloid cells was observed in any of the mutants (Fig. 2Aii and Supplementary Fig. S2B and S2C). These data suggest an additive effect of Nbs1^ΔB/ΔB Atm^{−/− VAV} mutants in resolution of programmed DSBs.

To further examine the effect of Nbs1^ΔB/ΔB Atm^{−/− VAV} on programmed gene rearrangement, CSR was analyzed in double-mutant splenocytes. B cells isolated from WT, single-, and double-mutant spleens were stimulated ex vivo with antisera recognizing CD40 (αCD40) in the presence or absence of IL4. Coadministration of IL4 induces CSR to IgG1 in this setting (33). Analysis of cell surface IgG1 was monitored by FACS following stimulation with αCD40 and IL4. WT splenocytes exhibited 18.3% IgG1 positivity at 4 days after stimulation, compared with 13.5% of Nbs1^ΔB/ΔB splenocytes. In contrast, Atm^{−/− VAV} and Nbs1^ΔB/ΔB Atm^{−/− VAV} cultures were markedly CSR defective, exhibiting approximately 4.5% IgG1 positivity (Fig. 3A). These differences were not attributable to variation in proliferation rates, apoptotic index, or to defects in switch region transcription and AID induction (ref. 28; Supplementary Fig. S3A–S3D).

To examine DSB rejoining directly, we performed a two-color FISH analysis with probes specific for sequences upstream of the Igh variable domain (5′ Igh, labeled for green signal) and sequences immediately downstream of the Igh constant region exons (3′ Igh, labeled for red signal; ref. 24). In this experiment, unbroken chromosomes 12 exhibit closely spaced green and red signals, whereas unresolved or improperly resolved DSBs appear as separated green and red “split signals” (Fig. 3Di, ii, iii) manifest in three outcomes: liberation of a small telomeric fragment (green in Fig. 3Diii); translocation between chromosome 12 telomeric fragment (that contains the 5′ Igh green probe) and another chromosome (Fig. 3Dii); translocation of a centric chromosome 12 (carrying the red signal) with another centric chromosome to form a dicentric chromosome (red signal between two fused chromosomes in Fig. 3Diii). Whereas a split Igh signal was rarely detected in WT and AID^−/− control B cells (less than 4%), split signals, indicative of IgH locus breaks, were observed in 49% of Nbs1^ΔB/ΔB Atm^{−/− VAV} cells, a significantly higher level than Atm^{−/− VAV} (23.7%) or Nbs1^ΔB/ΔB (8.8%; Fig. 3B and C and Supplementary Fig. S3E).

This apparent defect in DSB rejoining was accompanied by increased usage of microhomology at the residual switch region junctions that did form. The majority (15%) of the junctions in the double mutant had microhomology length of 10 nucleotides or more, roughly 2-fold higher than all other genotypes (Fig. 3E and Supplementary Fig. S3F). These data clearly indicate an additive defect in the repair of DSBs induced to initiate immunoglobulin class switch recombination in Nbs1^ΔB/ΔB Atm^{−/− VAV} mice.

In addition to defects in rejoining DSBs during CSR, Nbs1^ΔB/ΔB Atm^{−/− VAV} cells exhibited gross defects in the rejoining of spontaneous DSBs. Splenocytes stimulated to proliferate with αCD40 exhibited pronounced karyotypic instability. Fifty-two percent of Nbs1^ΔB/ΔB Atm^{−/− VAV} exhibited chromosomal aberrations compared with 17% in the Atm^{−/− VAV} single mutant (Fig. 4A, gray bars, Ai and B). A similar metaphase spreads pattern was also observed in MEFs (Supplementary Fig. S4C). Coadministration of αCD40 and IL4 further increased the number of aberrations observed in both single and Nbs1^ΔB/ΔB Atm^{−/− VAV} double mutants (Fig. 4A, black bars; Supplementary Fig. S4A).

In each genotype, the spectrum of spontaneous aberrations observed was suggestive of defects in the response to DSBs arising during DNA replication, but their abundance was markedly increased in Nbs1^ΔB/ΔB Atm^{−/− VAV} cells. The vast majority (over 90%) were either chromatid breaks and fragments or chromatid fusions and exchanges (Fig. 4C and Supplementary Fig. S4B). Chromosome fragility was associated with a 3-fold increased staining of 53BP1, a marker for DNA breaks, in double-mutant cells (7%) compared with single mutants (average of 2.2%; Supplementary Fig. S4D). Consistent with the gross level of instability observed karyotypically, 15% of Nbs1^ΔB/ΔB Atm^{−/− VAV} cells exhibited micronuclei, indicative of cell division in the presence of broken chromosomes (Fig. 4D). These data clearly indicate that ATM deficiency combined with Mre11 complex hypomorphism results in severe DSB repair deficiency.

An alternative and nonexclusive interpretation for the phenotypic synergy observed was based on potential role of the Mre11 complex in the activation of ATR (11, 12). In this scenario, the severity of the phenotype observed would reflect concomitant impairment of both the ATM and ATR arms of the DDR network.

To address this possibility, we assessed aphidicolin- and Camptothecin-induced phosphorylation of Chk1 on Ser345, both of which depend on ATR (34). Use of ATR-45, an ATR inhibitor (12), provided evidence that Chk1 phosphorylation was dependent on ATR after aphidicolin- and CPT treatment (Fig. 5A and Supplementary Fig. S5A). In WT cells, Chk1 phosphorylation was evident at 1 hour and decreased 5 hours after treatment. Similar levels of Chk1-Ser345 signal were observed in Nbs1^ΔB/ΔB and Atm^{−/− VAV} cells. However, a moderate but reproducible reduction in Ser345 phosphorylation was detected in Nbs1^ΔB/ΔB Atm^{−/− VAV} (Fig. 5A and B). A second ATR-dependent phosphorylation, RPA32 Ser4/Ser8 (35), was also compromised in Nbs1^ΔB/ΔB Atm^{−/− VAV} (Fig. 5B). The modest reduction of Chk1 phosphorylation observed in the double mutants would appear to argue against a primary role for the Mre11 complex in regulating ATR. Further supporting this interpretation, we found that Atm^{−/− VAV} and Nbs1^ΔB/ΔB Atm^{−/− VAV} cells exhibited a comparable defect in the maintenance of the G₂–M checkpoint, which is governed cooperatively by ATM and ATR (Supplementary Fig. S5B; ref. 36).

Aphidicolin treatment causes instability of CFS. The frequency of that outcome is heavily dependent upon ATR and Chk1 (37, 38). Therefore, we assessed CFS stability in Nbs1^ΔB/ΔB Atm^{−/− VAV} cells as an index of ATR function. Fra8E1 (ortholog of human FRA16D; ref. 39) was analyzed by a two-color FISH assay analogous to that used for CSR (Fig. 3Ci) using DNA probes adjacent to the Fra8E1 fragile site. Chromosome breakage was indicated by the appearance of separated green and red signals (Fig. 5Ci and 5Cii). Breakage of Fra8E1 was elevated relative to WT or Nbs1^ΔB/ΔB in untreated cells; Atm^{−/− VAV}, Nbs1^ΔB/ΔB Atm^{−/− VAV}, and WT cells treated with ATR inhibitor each exhibited approximately 10% breakage. Aphidicolin treatment doubled the frequency of breakage in Nbs1^ΔB/ΔB, Nbs1^ΔB/ΔB Atm^{−/− VAV}, and ATR inhibited WT cells, but had no effect on Atm^{−/− VAV}, consistent with the view that fragile site induction by aphidicolin is ATM independent (ref. 37; Fig. 5C). Aphidicolin induced breakage at other loci as well (Supplementary Fig. S5C). Collectively, these data suggest that CFS stability is compromised in Nbs1^ΔB/ΔB Atm^{−/− VAV} cells. The observation that aphidicolin induced fragile site breakage in Nbs1^ΔB/ΔB suggests that the Mre11 complex may also contribute to its stability in an ATM-proficient setting. Our data raise the possibility that the Mre11 complex–dependent protection of CFS is attributable to a function independent on both ATM and ATR.

The high degree of spontaneous genomic instability in Nbs1^ΔB/ΔB Atm^{−/− VAV} hematopoietic cells was correlated with increased risk of lymphomagenesis. Cohorts of 25 mice per each genotype were aged and monitored for malignancy. No malignancy was observed in Nbs1^ΔB/ΔB, whereas 50% of the Atm^{−/− VAV} cohort succumbed to thymic lymphoma within 12 months. Lymphomagenesis was completely penetrant in Nbs1^ΔB/ΔB Atm^{−/− VAV} mice, with tumors observed by 6 months of age in 91% of the mouse cohort (Fig. 6A). The increased risk of malignancy in Nbs1^ΔB/ΔB Atm^−/−VAV was not attributable to a differential effect on apoptosis, as Atm^{−/− VAV} and Nbs1^ΔB/ΔB Atm^{−/− VAV} thymocytes exhibited comparable apoptotic defects (Fig. 6B)

Thymic lymphomas arising in Atm^−/−VAV and Nbs1^ΔB/ΔB Atm^−/−VAV were histologically similar (Supplementary Fig. S6), but spectral karyotype analysis (SKY) revealed a highly complex genome rearrangements in Nbs1^ΔB/ΔB Atm^{−/− VAV}, many of which were clonal in the tumor (Fig. 6C). This outcome is consistent with the observation of gross chromosomal instability in Nbs1^ΔB/ΔB Atm^{−/− VAV} cells.

Hypomorphic mutations of the Mre11 complex confer embyronic lethality in the context of ATM deficiency (8, 10). These observations indicate that in addition to being required for ATM activation, the Mre11 complex specifies ATM-independent functions. That these functions are essential to embryonic viability suggests that the Mre11 complex normally mitigates pathologic outcomes of ATM deficiency.

In this study, we examined the ATM-independent functions of the Mre11 complex using a conditional allele of ATM (ATM^flox) in combination with Nbs1^ΔB which models the canonical Nbs1 allele inherited in NBS patients (8, 40). Expression of cre in Nbs1^ΔB/ΔB Atm^flox/− MEFs revealed that viable Nbs1^ΔB/ΔB Atm^−/− cells could be obtained in vitro. Nbs1^ΔB/ΔB Atm^−/− cells exhibited severe cellular phenotypes, such as impaired DNA repair, extensive chromosome instability, and reduced activation of ATR, the replication-stress checkpoint kinase. Ablation of ATM in hematopoietic lineages in vivo was effected by crossing Nbs1^ΔB/ΔB Atm^flox/− mice to hematopoietic stem cell–specific vav^cre mice. Hematopoietic cells were viable in the Nbs1^ΔB/ΔB Atm^−/−VAV mice, but they exhibited pronounced defects in lymphoid development, impaired class switching, and completely penetrant early onset lymphomagenesis.

Previous studies analyzing the relationship between ATM and the Mre11 complex have been carried out in cultured cells with siRNA or chemically mediated suppression of Mre11 complex and ATM functions (41, 42). The mouse model presented here allowed us to examine that relationship in vivo in the context of hematopoietic development in a physiologic context that also affords the opportunity to assess tumor risk.

Collectively, the Nbs1^ΔB/ΔB Atm^−/− phenotypes appear primarily attributable to defects in DSB repair. The severe defect in the joining of AID-induced DSBs during CSR strongly suggests that non-homologous end joining (NHEJ) is impaired in Nbs1^ΔB/ΔB Atm^{−/− VAV} cells (Fig. 3D). Recent data regarding the influence of 53BP1 and Rif1 on DSB resection suggest a possible mechanistic basis for this outcome. ATM-dependent phosphorylation of 53BP1 induces the recruitment of Rif1 to DSBs and to dysfunctional telomeres. This event likely underlies the inhibitory effect of 53BP1 on DSB resection. Because resected DSB ends are poor substrates for NHEJ, inhibition of resection by 53BP1 and Rif1 effectively promotes NHEJ (43). The effect of ATM deficiency on CSR may partially reflect the failure to inhibit resection of AID-induced DSBs. Similarly, the Mre11 complex has been implicated in promoting NHEJ at dysfunctional telomeres (44), as well as in NHEJ during CSR and VDJ recombination (Fig. 3A; refs. 5, 7, 30, 45). Although in those contexts, defects in NHEJ-mediated rejoining must at least partially reflect a decrement in ATM activation associated with Mre11 complex hypomorphism, the additive defects in Nbs1^ΔB/ΔB Atm^−/− cells reveal a direct (i.e., ATM-independent) role for the Mre11 complex in NHEJ.

On the other hand, residual NHEJ functions are evident in Nbs1^ΔB/ΔB Atm^−/− cells as indicated by the development of mature B and T cells, and the frequent occurrence of radial structures that require NHEJ to form (Figs. 2, 4B; Supplementary Fig. S4C). Because the Mre11 complex is involved in DSB repair mediated by homology directed repair (HDR; ref. 46), the coincident impairment of NHEJ and homologous recombination (HR) in Nbs1^ΔB/ΔB Atm^−/− cells may also account for the phenotypic severity observed. Following IR treatment, persistent RPA and Rad51 nuclear foci (Fig. 1B and C) are observed in Nbs1^ΔB/ΔB Atm^−/− compared with Atm^−/−. As RPA and Rad51 assemblies are intermediates in the HR process, their persistence indicates that the completion of HDR is impaired or delayed in double-mutant cells. We propose that combined defects in DSB repair DDR signaling associated with ATM deficiency underlie the phenotype of Nbs1^ΔBΔB Atm^{−/− VAV} mutants.

Defects in the response to replication stress were also evident in Nbs1^ΔB/ΔB Atm^{−/− VAV} cells, as inferred from expression of common fragile sites induced by aphidicolin treatment. The suppression of fragile sites' instability is strongly dependent on ATR (37). Consequently, the simplest explanation for this result is that ATR function is impaired in Nbs1^ΔB/ΔB Atm^{−/− VAV} cells. However, the effect on Chk1 and RPA phosphorylation in double mutants was extremely mild, arguing against altered ATR functions as the underlying basis of CFS expression. Consequently, we favor the possibility that Nbs1 plays a more direct role in maintaining stability at common fragile sites, or that the stability of CFS depends both on ATR and on ATM in mouse cells.

Although previous studies in in vitro and in culture cells suggest the existence of an Nbs1-dependent ATR activation pathway (11, 12, 47), our data do not support the interpretation that the Mre11 complex exerts a strong influence on ATR activity in ATM-proficient cells. In this regard, the use of an in vivo genetic system that allowed the derivation of primary cells provided a sensitive setting for the assessment of the Mre1 complex influence on ATR activation. It is also conceivable that because the product of the Nbs1^ΔB allele retains the RPA-interacting domain that appears to influence ATR activation (11), the influence of the Mre11 complex on ATR may not be fully revealed in Nbs1^ΔB/ΔB cells.

The development of aggressive T-cell lymphoma in Nbs1^ΔB/ΔB Atm^{−/− VAV} mice demonstrates that Nbs1 contributes substantially to suppressing the oncogenic potential of ATM deficiency. Despite the fact that genome instability per se is not sufficient to elicit the onset of lymphomagenesis in Mre11 complex hypomorphic mice, it is sufficient to promote the penetrance of an initial mutation such as p53 heterozygosity (9). Indeed, recent data have highlighted a role of the Mre11 complex acting as a barrier to oncogene-driven breast tumorigenesis (48). We propose that the higher rate of genome instability exhibited by double-mutant cells, combined with reduced DDR signaling, underlies the basis for the observed increase in tumor predisposition. Supporting the idea that reduced ATR activity in Nbs1^ΔB/ΔB Atm^{−/− VAV} mice accelerates lymphomagenesis, ATR^+/− mice have been reported to show increase in tumor incidence (49). In addition, we found that Nbs1^ΔB/ΔB Atm^{−/− VAV} cells show a synergistic increase in spontaneously arising and CSR-associated chromosomal aberrations (Figs. 3B and 4A).

The development of an animal model in which the ATM-independent functions of the Mre11 complex can be analyzed offers a novel perspective for analyzing relationships among the components of the DDR network. The profound defects associated with coincident inhibition of the ATM and ATR arms of the DDR (50) support the idea that the simultaneous inhibition of both DNA damage signaling protein kinases could be exploited to improve the efficacy of clastogenic therapies.

C.H. Bassing has expert testimony in Regeneron Pharmaceuticals. No potential conflicts of interest were disclosed by the other authors.

Conception and design: A. Balestrini, L. Nicolas, J. Chaudhuri, J.H.J. Petrini

Development of methodology: A. Balestrini, L. Nicolas, J.H.J. Petrini

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A. Balestrini, L. Nicolas, K. Yang-lott, O.A. Guryanova, J.H.J. Petrini

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A. Balestrini, L. Nicolas, K. Yang-lott, O.A. Guryanova, R.L. Levine, J. Chaudhuri, J.H.J. Petrini

Writing, review, and/or revision of the manuscript: A. Balestrini, L. Nicolas, J. Chaudhuri, J.H.J. Petrini

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A. Balestrini, J.H.J. Petrini

Study supervision: C.H. Bassing, J. Chaudhuri, J.H.J. Petrini

The authors thank Fred Alt for ATM^flox mice and genotyping, Matthias Stadtfeld and Thomas Graf for vav^cre mice, Monica Gostissa and Ryan L. Ragland for reagents and technical advice, and members of the Petrini laboratory and Thomas J. Kelly for providing helpful comments and suggestions throughout the course of this study.

This work was supported by the Geoffrey Beene Center at the Memorial Sloan Kettering Cancer Center (MSKCC) and NIH grant GM59413 (to J.H.J. Petrini), grants from the NIH (1RO1AI072194) and the Starr Cancer Research Foundation (to J. Chaudhuri), Department of Pathology and Laboratory Medicine of the Children's Hospital of Philadelphia Research Institute, a Leukemia and Lymphoma Society Scholar Award, and the NIH R01 grants CA125195 and CA136470 (to C.H. Bassing), NIH/NCI 1K99CA178191 (to O.A. Guryanova), and American-Italian Cancer Foundation fellowship and EMBO fellowship (EMBO ALTF 43-2011; to A. Balestrini). R.L. Levine is a Leukemia and Lymphoma Society Scholar.

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.