An essential function for NBS1 in the prevention of ataxia and cerebellar defects

538 VOLUME 11 | NUMBER 5 | MAY 2005 NATURE MEDICINE

An essential function for NBS1 in the prevention of ataxia and cerebellar defects Pierre-Olivier Frappart1, Wei-Min Tong1, Ilja Demuth2, Ivan Radovanovic3, Zdenko Herceg1, Adriano Aguzzi3, Martin Digweed2 & Zhao-Qi Wang1

Nijmegen breakage syndrome (NBS), ataxia telangiectasia and ataxia telangiectasia–like disorder (ATLD) show overlapping phenotypes such as growth retardation, microcephaly, cerebellar developmental defects and ataxia. However, the molecular pathogenesis of these neurological defects remains elusive. Here we show that inactivation of the Nbn gene (also known as Nbs1) in mouse neural tissues results in a combination of the neurological anomalies characteristic of NBS, ataxia telangiectasia and ATLD, including microcephaly, growth retardation, cerebellar defects and ataxia. Loss of Nbn causes proliferation arrest of granule cell progenitors and apoptosis of postmitotic neurons in the cerebellum. Furthermore, Nbn-deficient neuroprogenitors show proliferation defects (but not increased apoptosis) and contain more chromosomal breaks, which are accompanied by ataxia telangiectasia mutated protein (ATM)-mediated p53 activation. Notably, depletion of p53 substantially rescues the neurological defects of Nbn mutant mice. This study gives insight into the physiological function of NBS1 (the Nbn gene product) and the function of the DNA damage response in the neurological anomalies of NBS, ataxia telangiectasia and ATLD.

In humans, NBS (caused by a hypomorphism in NBS1, known as Nbn in mice), ataxia telangiectasia (caused by a mutation in ATM) and ATLD (caused by a hypomorphism in MRE11A) constitute a sub-group of chromosome instability syndromes and show very similar cellular defects such as radiosensitivity, chromosome instability and cell cycle checkpoint defects. These genetic disorders also share clini-cal features, including immunodeficiency, growth retardation, radio-sensitivity, chromosome instability and cancer predisposition1–5. But cerebellar degeneration and ataxia are further obligatory characteristics of individuals with ataxia telangiectasia and ATLD. Characterization of genes involved in these disorders illustrates the functional rela-tionship between NBS1, meiotic recombination 11 protein (Mre11) and ATM. NBS1, also known as nibrin or p95, forms a complex with Mre11 and Rad50 (the MRN complex), which interacts with ATM in the DNA damage response and in the maintenance of genomic integ-rity6. Although NBS, ataxia telangiectasia and ATLD share many clinical and cellular characteristics, they show distinct neurological anomalies. Despite the extensive biochemical characterization of the MRN pro-teins and numerous studies using cell lines derived from individuals with NBS and ATLD, the physiological function of these molecules remains elusive.

Many attempts have been made to investigate the molecular patho-genesis of these disorders, and several knockout mouse models have been generated. Whereas null mutation of Nbn and Mre11 causes embryonic lethality7–9, hypomorphic mutations are compatible with survival10–12; however, these models do not reproduce the major

neuronal anomalies seen in NBS and ATLD. Although ATM knockout mouse neuroprogenitors show genomic instability13,14, ATM null mice show only mild motor dysfunction without obvious histological cerebellar defects15–17. These data suggest that there is a missing link in the molecular pathways underlying these genetic disorders.

RESULTSGeneration of mice carrying neural-specific disruption of NbnTo gain insight into the function of MRN in these neurological defects, we specifically inactivated Nbn in the mouse central nervous system (CNS). We have previously shown that the disruption of exon 6 results in Nbn null mutation, leading to early embryonic lethality7. Thus, we first used gene targeting to engineer a mouse strain in which exon 6 was deleted (Nbn+/∆6) (Fig. 1a–c). Nbn∆6/∆6 embryos did not survive, consistent with our previous report7, suggesting that the deletion of exon 6 caused a null mutation of Nbn. We next generated mice carrying the Nbn allele modified by flanking exon 6 with two loxP sites (Nbn+/F6) (Fig. 1a,d). NbnF6/F6 homozygous mice were healthy and phenotypically normal. To disrupt the Nbn gene in the CNS, these mice were crossed with Nbn+/– mice7 and with nestin-Cre transgenic mice18. ‘Nbn-CNS-del’ mice consisted of two genotypes, NbnF6/–Cre+ and NbnF6/F6Cre+, which were phenotypically identical. The control group (Nbn-CNS-ctr) was composed of eight genotypes: Nbn+/+Cre–, Nbn+/+Cre+, Nbn+/F6Cre–, Nbn+/F6Cre+, Nbn+/–Cre–, Nbn+/–Cre+, NbnF6/–Cre–, NbnF6/F6Cre–, all of which carried at least one functional Nbn allele and appeared pheno-typically normal (data not shown). Southern blotting of different parts

1International Agency for Research on Cancer (IARC), 150 cours Albert Thomas, 69008 Lyon, France. 2Institute of Human Genetics, Charité-Campus-Virchow, Humboldt Universitaet zu Berlin, Augustenburger Platz 1, D-13353 Berlin, Germany. 3Institut für Neuropathologie, UniversitätsSpital Zürich, Schmelzbergstr, 12 CH-809 Zürich, Switzerland. Correspondence should be addressed to Z.-Q.W. ([email protected]).

Published online 10 April 2005; corrected online 17 April 2005 (details online); doi:10.1038/nm1228

A R T I C L E S©

2005

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

emed

icin

e

NATURE MEDICINE VOLUME 11 | NUMBER 5 | MAY 2005 539

of the brain at postnatal day 21 (P21) showed efficient excision of the floxed Nbn allele (Fig. 1e). RT-PCR and PCR analyses further confirmed the efficient deletion of Nbn in the CNS of Nbn-CNS-del mice (Fig. 1f and Supplementary Fig. 1 online). Furthermore, we observed an effi-cient downregulation of NBS1 in neural cells as early as embryonic day (E)13.5. These data indicate that embryonic deletion of exon 6 mediated by nestin-Cre led to Nbn deficiency in the CNS.

Nbn-CNS-del mice were viable at birth and appeared indistinguish-able from their littermates. But we observed growth retardation as early as P7 in these mice, which reached only about 50% of the body weight of Nbn-CNS-ctr mice at weaning (Supplementary Fig. 1 online). All Nbn-CNS-del mice showed balance disorders, tremors, altered gait, repetitive movements and akinesis after P7 (Fig. 2a). When the mutants were suspended by the tail, Nbn-CNS-ctr mice stretched out their hind limbs, whereas Nbn-CNS-del mice kept their hind limbs in a crossed position (Fig. 2a). All of these features are criteria of cerebellar ataxia. Moreover, owing to the severe ataxia, Nbn-CNS-del animals could not complete the balance beam test (data not shown).

Macroscopic examination of brains from P7–P21 Nbn-CNS-del mice showed greatly reduced cerebella lacking foliation (Fig. 2b). Brain weight of Nbn-CNS-del mice (P7, n = 7; P14, n = 5; P21, n = 9) was also significantly reduced compared to controls (P7, n = 56; P14, n = 21; P21, n = 22) which was evident as early as P7 (Fig. 2c). Parasagittal sections of Nbn-CNS-del brains showed a considerable reduction of the cerebel-lum compared to Nbn-CNS-ctr mice (Fig. 4d). Histological analysis of P21 Nbn-CNS-del cerebella showed a marked agenesis of foliation associated with the disorganized granule cell layer and Purkinje cell layer (Fig. 2d). Immunostaining using neuron-specific markers for mature granule cells (neuronal nuclei; NeuN) and for Purkinje cells (calbindin D-28K) showed a reduction in the number of granule cells and ectopic presence of Purkinje cells lacking a lining pattern, with altered dendritic trees in Nbn-CNS-del cerebella (Fig. 2d). But owing to the general disruption of the cerebellar structure of Nbn-CNS-del mice, we could not accurately score the numerical alteration in Purkinje cells. Nbn-CNS-del cerebella also showed an alteration of Bergmann glia cells visualized after glial fibrillary acidic protein (GFAP) staining (Fig. 2d). These morphological alterations were readily detectable at P0 (data not shown) and were exacerbated in P7 Nbn-CNS-del mice (Supplementary Fig. 1 online).

In vivo and in vitro characterization of cerebellar defectsTo elucidate the cause of cerebellar developmental defects and the fate of Nbn-deficient neuroprogenitor cells, we examined cell proliferation and cell death. To assess the proliferative potential of neuronal progenitor cells, we performed in vivo labeling with bromodeoxyuridine (BrdU). We found a much lower number of BrdU-positive cells in the outer external granule layer (EGL; the major area of proliferation) in E15.5, E18.5 and P7 Nbn-CNS-del cerebella than in their Nbn-CNS-ctr coun-terparts (Fig. 2e and data not shown). Immunostaining with antibody against Ki67 further confirmed the reduced proliferation of neuronal cells in Nbn-CNS-del cerebella (Fig. 2e). We next analyzed cell death and found an increased number of cells containing pyknotic or fragmented nuclei (a feature of apoptotic cells) in the inner EGL of the Nbn-CNS-delcerebellar primordium at E15.5 and in cerebella at E18.5 and P7 (Fig. 2f). These observations suggest that postmitotic neurons are sus-ceptible to cell death once they migrate inward to the inner EGL.

To further study the cellular defects of Nbn-CNS-del neural cells, we isolated and cultured neural stem and progenitor cells from E13.5 brains and analyzed them for neurosphere formation. The total number of neurospheres formed and the number of cells in each neurosphere were significantly lower in Nbn-CNS-del brains than in Nbn-CNS-ctr brains (Fig. 3a). We also replated the cells after dissociation of the primary neu-rospheres; secondary neurospheres did not form from the Nbn-CNS-del genotype (Fig. 3b). In addition, analysis of cell death by TUNEL showed that apoptosis in Nbn-CNS-del E13.5 neurosphere cultures was comparable to that in Nbn-CNS-ctr cultures (Fig. 3c), suggesting that apoptosis did not have a major role in formation of Nbn-CNS-del neuroprogenitors from compromised neurospheres. Most notably, in contrast to Nbn-CNS-ctr, we were unable to isolate any primary Nbn-CNS-del neurospheres from newborn animals after four inde-pendent experiments using several seeding concentrations, including 0.4 × 105 cells/ml, 5 × 105 cells/ml and 10 × 105 cells/ml (data not shown). These data indicate that the lack of NBS1 reduced the proliferation of neural stem and progenitor cells without inducing apoptosis.

Chromosomal instability and ATM-mediated p53 activationBecause NBS1 is important in the surveillance of chromosomal integ-rity, we performed cytogenetic analysis on primary neurosphere cells. Although all Nbn-deficient and control neurosphere cells were diploid,

a b

c

d

e f

I

I

I

Figure 1 Disruption of the Nbn gene in embryonic stem and neuronal cells in mice. (a) Gene targeting strategy. (b) Southern blotting of embryonic stem cell DNA digested with BamHI and ApaI and hybridized with the intron 4 (I4) probe. +/+, wild-type; +/T, targeted clones. (c) Southern blots of DNA from Nbn+/∆6 (+/∆6) and wild-type mouse tails after digestion with BamHI and ApaI, hybridized with the I4 probe. (d) Southern blots of DNA from Nbn+/F6 (+/F6) and wild-type (+/+) mouse tails after digestion with SacI and ApaI, hybridized with the E6 probe. (e) Southern blot analysis of deletion of the Nbn gene in different regions of the Nbn-CNS-del brain. OB, olfactory bulbs; BC, brain cortex; H, hypothalamus; C, cerebellum; BS, brainstem. Controls are tail DNA from mice with indicated Nbn genotype. (f) Efficient deletion of the Nbn mRNA in the cerebellum and cerebrum of NbnF6/–Cre+ mice. Lanes 1 and 5: NbnF6/–Cre+ cerebellum; lanes 2 and 6: NbnF6/–Cre+ cerebrum; lane 3: Nbn+/+ cerebellum; lane 4: Nbn+/+ cerebrum. Note the lack of PCR product in NbnF6/–Cre+ samples.

A R T I C L E S©

2005

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

emed

icin

e

540 VOLUME 11 | NUMBER 5 | MAY 2005 NATURE MEDICINE

the number of chromosomal breaks per metaphase was significantly higher in Nbn-CNS-del neurospheres (0.57 for Nbn-deficient neuro-spheres versus 0.14 for Nbn-CNS-ctr neurospheres, P = 0.0025), and 43% of Nbn-CNS-del metaphases contained at least one aberration, in contrast to 12% in controls (P = 0.0051; Table 1).

We next studied whether the defects in proliferation and increased cell death in Nbn-CNS-del cerebella were the result of activation of the DNA damage response in cells containing more chromosome breaks. As p53 (encoded by the Trp53 gene) regulates cell cycle progression and

apoptosis in response to DNA damage19,20, we asked whether p53 was activated by these chromosomal breaks in Nbn-CNS-del cells. Western blotting of E13.5 neurospheres showed a higher level of phosphorylation of p53 at Ser15 in Nbn-CNS-del neural cells than in Nbn-CNS-ctr cells (Fig. 4a). Consequently, the levels of p21, a downstream effector of p53 responsible for cell cycle arrest, were greatly increased in these cells (Fig. 4a). Although the expression of NBS1 was strongly reduced (Fig. 4a), there was a residual amount of NBS1 in Nbn-CNS-del neurospheres, which might be attributable to the incomplete depletion of the protein

a

e

d

f

b

c

Nbn-CNS-ctr

H&E NeuN D-28K GFAP

Per

cen

tag

e p

osi

tive

cells

in E

GL

Nu

mb

er o

f p

osi

tive

cells

in E

GL

Figure 2 Ataxia, microcephaly and cerebellar developmental defects in Nbn-CNS-del mice. (a) Nbn-CNS-del mice with severe ataxia and motor reflex defects. (b) Reduced cerebellar size and foliation (circles) in Nbn-CNS-del mice. (c) Comparison of brain weight between Nbn-CNS-del mice (gray) and Nbn-CNS-ctr mice (white) at P7, P14 and P21. **P = 0.001; ***P < 0.0001. (d) Histological analysis of P21 cerebella. Hematoxylin and eosin (H&E) staining: original magnification, ×1. Neuronal marker (NeuN) staining: original magnification, ×4. Calbindin (D-28K) staining: original magnification, ×40. GFAP staining: original magnification, ×20. PC, Purkinje cell; GCL, granular cell layer. (e) In vivo BrdU labeling (original magnification, ×40) and immunohistological staining of Ki67 (original magnification, ×20). Dotted lines and arrowheads, EGL. Upper histogram: BrdU-positive cells as a percentage of total cells (n) in the EGL. Lower histogram: average number (± s.d.) of Ki67-positive cells from six sections (n = 6) of two animals of each genotype. ***P < 0.0001. (f) Hematoxylin and eosin staining of cerebellum at E18.5 and P7 (original magnification, ×40). Apoptotic cells, characterized by their pyknotic and fragmented nuclei (arrows), are located mainly at the inner EGL. Insets: representative pyknotic cells. Histogram: percentage of total EGL cells (n) that are apoptotic, from at least four mice with more than five sections of each. ***P < 0.0001. ML, molecular layer; GCL, granule cell layer; PC, Purkinje cells, DT, dendritic trees. EGL, external germinal layer. White bars, Nbn-CNS-ctr. Black bars, Nbn-CNS-del.

A R T I C L E S©

2005

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

emed

icin

e

NATURE MEDICINE VOLUME 11 | NUMBER 5 | MAY 2005 541

after Cre-mediated deletion at this early stage (E13.5; ref. 21). To define the upstream activator of the p53 pathway and to gain insight into the involvement of ATM in the neuronal phenotype of these mice, we treated neuroprogenitors with caffeine and wortmannin at a dose that specifically inhibited the ATM kinase. We found downregulation of phosphoryla-tion of p53 at Ser15 and of p21 in Nbn-CNS-del neural cells, suggesting that the kinase involved in phosphorylation of p53 at Ser15 is probably ATM (Fig. 4b). To further verify whether the activation of p53 and p21 also occurred in vivo, we performed in situ immunostaining for p53 and p21 proteins in E18.5 cerebella. In normal control cerebellar sections, only a few cells were positive for p53, and, consequently, almost no p21-positive cells could be detected (Fig. 4c). In contrast, the number of EGL granule cells that showed p53 immunoreactivity and activation of p21 were increased in the Nbn-CNS-del cerebellar EGL layer (Fig. 4c). These results suggest that chromosomal breaks in Nbn-deficient cells activated p53-mediated cell cycle arrest and block of proliferation.

Trp53 mutation rescues neuronal defects of Nbn-CNS-del miceTo further test whether activation of p53 is responsible for the Nbn dele-tion–specific cerebellar defects, we introduced Trp53 null alleles into Nbn-CNS-del mice. Whereas Nbn-CNS-del mice in a Trp53+/+ back-ground showed growth retardation and progressive ataxia as in Nbn-CNS-del mutant mice (Fig. 2a), the Trp53 null mutation fully corrected the growth retarda-tion and corrected the motor coordination defects of Nbn-CNS-del mice to a large extent (Fig. 5a,b). In a balance beam assay, control mice (Nbn-CNS-ctr and Trp53–/–) were all proficient at walking along the beam, whereas Nbn-CNS-del Trp53+/+ mice were unable to walk along the beam, as expected (Fig. 5b). Notably, Nbn-CNS-del Trp53–/– mice could walk along the balance beam with fewer missteps than the control group (Fig. 5b), indicating greatly improved coordinat-ing function. Additionally, heterozygous muta-tion of Trp53 also improved motor dysfunction of Nbn-CNS-del mice, as they completed the walk along the beam, albeit with a considerable amount of missteps (Fig. 5b).

The brain weight and cerebellar morphol-ogy of Nbn-CNS-del Trp53–/– mice was very similar to that of control mice (Fig. 5c and data not shown). Histological analysis showed that Trp53 heterozygosity had a marginal effect on the granule cell numbers and partially rescued

the foliation structure, as evidenced by the Purkinje cell lining (Fig. 5c), implying a certain degree of rescue, which may be responsible for the improved motor function in these mice. Notably, Trp53 homozygous mutation rescued cerebellar defects: the appearance of the major folia formation, distinct structures of the molecular layer, granule cell layer (GCL) and Purkinje cell layer (Fig. 5c) were all comparable to those in the control cerebella, although granule cellularity was still reduced, and the lining of Purkinje cells was still inappropriate (Fig. 5c). These data indicate that Nbn-CNS-del cerebellar developmental defects are mainly caused by a p53-dependent pathway. Thus, p53 is a key player in the neurological phenotype of Nbn-CNS-del mice. To date, none of the Nbn-CNS-del Trp53–/– double mutant mice developed specific cere-bellar tumors, probably because of early onset of p53 null–associated tumorigeneses such as sarcoma and lymphoma (data not shown).

DISCUSSIONWhereas the microcephaly of Nbn-CNS-del mutant mice resembles the neurological phenotype of individuals with NBS, the cerebellar defects and ataxia in these mice are reminiscent of ataxia telangiectasia and ATLD2,3,22. The absence of ataxia and cerebellar defects in individuals with NBS may be a result of the putative expression of a truncated C-terminal NBS1 in most neural tissues of individuals with NBS23, which allows formation of

a b

c

n = 7 n = 9 n = 7 n = 9 n = 2 n = 4

Nu

mb

er o

fn

euro

sph

eres

/ml

Nu

mb

er o

fn

euro

sph

eres

/ml

Nu

mb

er o

f ce

llsin

neu

rosp

her

es

Per

cen

tag

e T

UN

EL

-p

osi

tive

cel

ls

Figure 3 Characterization of Nbn-deficient neural cells in vitro. (a) The number of neurospheres derived from 5 × 105 cells/ml of E13.5 Nbn-CNS-del brain preparations. Insets show representative neurospheres after 7 d in culture. The number of cells was counted from all neurospheres. **P = 0.0016, ***P < 0.0001. (b) Self-renewal assay of neural stem cells after replating cells at 5 × 105 cells/ml from primary neurosphere cells. ***: P < 0.0001. (c) TUNEL analysis of neuroprogenitor cells of E13.5 Nbn-CNS-del brains. We scored 500 cells for each neurosphere cell population after 7 d in culture. White bars, Nbn-CNS-ctr mice. Gray bars, Nbn-CNS-del mice. n, number of mice analyzed of a given genotype.

Table 1 Cytogenetic analysis of E13.5 Nbn-CNS-del neurosphere cellsCell populations Genotype Metaphase

analyzedNumber of

chromosomesNumber of

breaksAverage number

of breaks per metaphase

Number of metaphase with

breaks

Controls

NS-56 NbnF6/F6Cre− 11 39.4 ± 0.8 2 0.18 2 (18%)

NS-82 Nbn+/F6Cre− 45 39.5 ± 0.7 4 0.09 4 (9%)

NS-89 Nbn+/F6Cre− 42 39.6 ± 0.8 6 0.14 4 (10%)

Nbn-deficient

NS-54 NbnF6/−Cre+ 49 39.2 ± 1.4 31 0.63 24 (49%)

NS-61 NbnF6/−Cre+ 23 38.1 ± 2.0 18 0.64 12 (43%)

NS-62 NbnF6/−Cre+ 23 38.2 ± 1.6 14 0.61 12 (52%)

NS-74 NbnF6/−Cre+ 42 38.8 ± 1.5 16 0.38 12 (29%)

Only chromatid breaks were scored and shown.

A R T I C L E S©

2005

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

emed

icin

e

542 VOLUME 11 | NUMBER 5 | MAY 2005 NATURE MEDICINE

the MRN complex and almost normal cerebellar development and func-tion in humans. Consistent with this hypothesis, we have recently shown that hypomorphic NBS1657del5 (the prominent mutation in human NBS patients) mutant cDNA can rescue the viability of Nbn-null mutant cells24. A possible explanation for Nbn-CNS-del mice showing ataxia and cerebel-lar defects is that the reduced number of granule cells induces an ectopic localization of Purkinje cells and impairs development of their dendritic trees, leading to malfunction of these neurons25. It is also conceivable that, as is the case in ATLD cells that contain low levels of NBS1, Nbn deficiency can cause downregulation of Mre11 (ref. 2), which may affect Purkinje cell function. Indeed, we have found lower amounts of NBS1 partners such as Mre11 and Rad50 in Nbn-deficient cells (Supplementary Fig. 1 online).

Endogenous DNA damage occurs frequently in neural stem cells and progenitor cells, and an inability to repair them results in apoptosis26,27. Consistent with the role of NBS1 in double-strand break (DSB) repair, we have found that Nbn deficiency leads to increased apoptosis in postmitotic neurons in the cerebellum. In this regard, it is noteworthy that inactiva-tion of DNA ligase IV, XRCC4 and Brca1 leads to abnormal neurogenesis in mice, owing to apoptosis dependent on ATM and p53 (referred to as ATM/p53-dependent apoptosis) in postmitotic neurons28–32. Another notable finding in this study is that granule progenitors cells at the outer EGL in Nbn-CNS-del cerebella show a low proliferative activity. Both in vivo and in vitro proliferation assays show a reduced proliferative capacity of neural stem cells or progenitor cells in Nbn-CNS-del brains. The mech-anism underlying this phenotype may be explained by a high basal level of chromosomal aberrations in Nbn-deficient neural cells which activates the DNA damage checkpoint response and leading to cell cycle arrest33. In this regard, we have observed a high basal level of chromosomal breaks in Nbn-CNS-del neural cells associated with p53 hyperphosphorylation and p21 induction in neuroprogenitor cells in vitro and associated with accumulation of p53 and p21 in granule progenitor cells in vivo. Thus, the developmental defect of Nbn-null cerebella is probably the result of p53-mediated proliferation arrest in progenitor cells and apoptosis in

postmitotic neurons20. The causal role of p53 activation in the cerebellar developmental defects of Nbn-CNS-del mice is well supported by the observation that a Trp53 null mutation rescues the Nbn-CNS-del cerebel-lar phenotype. The activation of the p53 pathway seems to be dependent upon ATM, as ATM inhibitors such as caffeine and wortmannin decrease phosphorylation of p53 at Ser15 and downregulate p21 in Nbn-deficient neural cells. The fact that ATM still activates p53 function in the absence of NBS1 is notable, given recent studies showing that NBS1 regulates ATM function for its downstream effectors34,35. Taken together, these data strongly suggest that Nbn deficiency leads to spontaneous genomic insta-bility and activation of the ATM/p53-mediated DNA damage response, leading to cell cycle progression delay in neural stem cells and progenitor cells in vivo and in vitro and apoptosis in postmitotic neurons.

Although DNA damage response–induced apoptosis in postmitotic neurons is known to be responsible for the neurological developmental defects in mice lacking DSB molecules26,27, proliferation defects have not been shown for the neurological phenotypes. This is the first in vivo evidence showing that NBS1, a DNA DSB repair molecule, acts in both proliferation of early neuronal progenitors and apoptosis of postmitotic neurons. Thus, specific cerebellar defects and the ataxia phenotype in Nbn-CNS-del mice distinguish NBS1 from other DSB repair molecules with regard to physiological function. Because of the multifunctional nature of MRN, we speculate that, like other DSB repair molecules such as ligase IV and XRCC4, the major function of NBS1 in nonproliferat-ing cells is DSB repair; however, both damage signaling (in the form of cell cycle control, for instance) through activation and interaction with ATM, as well as DNA repair activity are required in proliferating cells to prevent accumulation of genome damage.

Notably, the cerebellum seems to be a major target in the current model. Neurons in the cerebellum are postmitotic, highly active cells that live as long as the organism does. To have such a long lifespan in the face of a high level of genotoxic stress, neuronal cells presumably have elaborate DNA damage response and repair mechanisms that keep

a

c

b

p53-ser15-p p53-ser15-p

Figure 4 ATM-mediated p53 activation in Nbn-CNS-del neural cells. (a) Western blot showing efficient downregulation of NBS1 in E13.5 Nbn-CNS-del neural cells. Two cell populations of each indicated genotype are shown; probes are shown at right. p53ser15-p, phosphorylation of p53 at Ser15. (b) Western blot of ATM-mediated p53 activation. Caffeine (5 mM) and wortmannin (20 µM) were added to neurosphere cultures. Probes are shown at right. (c) Representative immunohistological analysis of p53 and p21 in E18.5 cerebella. Insets show the areas containing p53- and p21-positive cells. Original magnification, ×40.

A R T I C L E S©

2005

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

emed

icin

e

NATURE MEDICINE VOLUME 11 | NUMBER 5 | MAY 2005 543

them functional. Moreover, the hyperproliferation of granule neurons during embryonic life (with the first wave of proliferation at E15.5) and during early postnatal life (with the second wave from P7–P14) may also require an efficient DNA damage response27. In this regard, the extreme sensitivity of the cerebellum to chronic DNA damage as well as the key role of p53 in medulloblastoma development in mutants with defective DSB repair are demonstrated by double mutant ligase IV-p53 and PARP-1–p53 mice36,37. Finally, despite efficient deletion of NBS1 in neurons in other parts of the brain, the function of the MRN complex in the cerebellum might not be compensated for by other DSB repair mechanisms.

The present study suggests a model in which Nbn inactivation results in an accumulation of chromosomal instability which activates ATM- and p53-mediated DNA damage response, leading to cell cycle progression delay in pro-genitor cells and apoptosis in postmitotic neu-rons. In neurospheres, p53 expression could be induced by stress under culture conditions, but the observed p53 activation seems to be specifi-cally associated with the Nbn mutant genotype. Moreover, we cannot rule out the possibility that Nbn mutant neural cells have abnormal ATM function, leading to altered p53 activation. The Nbn-CNS-del mice show a combination of neu-rological phenotypes of NBS, ataxia telangiecta-sia and ATLD, including microcephaly, growth retardation, cerebellar developmental defects and ataxia. These findings are notable because they strongly suggest that the cerebellar degenera-tion observed in ataxia telangiectasia and ATLD patients is probably the result of an altered DNA damage response mediated by ATM. Thus, this model provides new insights into the role of NBS1 and the DNA damage response in the neu-rological anomalies of NBS patients, as well as the crucial role of NBS1 in cerebellar development.

METHODSGeneration of mice carrying the floxed Nbn allele. To perform gene targeting in embryonic stem cells, we constructed a targeting vector (pTVFloxM-Nbn1) using the 8-kb sequences encompassing Nbn exons 3–8; it contained two loxP sites flanking the neo cassette and another loxP that was inserted into intron 6. After gene targeting in E14.1 embry-onic stem cells, we identified nine targeted clones (Nbn+/T) by PCR and Southern blotting, eight of which contained all three loxP sites in the locus (data not shown). Transient expression of Cre in these Nbn+/T embryonic stem cells generated embryonic stem cell clones carrying the allele either with floxed exon 6 (F6) or deleted exon 6 (∆6). We used two independent Nbn+/F6 clones and one Nbn+/∆6 clone to generate Nbn+/F6 and Nbn+/∆6 mice, respectively. All animal experiments were approved by and performed in accordance with the guidelines of the International Agency for Research on Cancer’s Animal Care and Use Committee.

Genotyping of embryonic stem cells and mice by PCR and Southern blotting. For PCR geno-typing, two primers were used to genotype the

NbnF6 allele: EX6 (5′-CAGGGCGACATGAAAGAAAAC-3′) and loxPtestR (5′- AATACAGTGACTCCTGGAGG-3′). For the Nbn∆6 allele, primers intron5F (5′-ATAAGACAGTCACCACTGCG-3′ and loxPtestR were used. For Southern blotting, we digested genomic DNA with BamHI and ApaI, and hybridized blots with a probe corresponding to intron 4 (I4) of the Nbn gene. To genotype the NbnF6 allele by Southern blotting, we digested genomic DNA with SacI and ApaI, and hybridized blots with a probe corresponding to exon 6 (E6) of the Nbn gene.

Generation of Trp53-deficient Nbn-CNS-del mice. The Trp53–/– and Nbn+/F6NesCre+ mice were interbred to generate Nbn+/F6 Cre+Trp53+/– mice, which were intercrossed to generate NbnF6/F6 Cre+Trp53+/+, NbnF6/F6 Cre+Trp53+/– or NbnF6/F6Cre+Trp53–/– mice.

Figure 5 p53 deficiency rescues growth retardation and neurogenesis defects of Nbn-CNS-del mice. (a) Null mutation of p53 rescues the growth retardation of Nbn-CNS-del mice. (b) Balance beam test shows that p53 mutations partly, but substantially, corrected the motor coordination defects of Nbn-CNS-del mice. ND, not determined; mice could not complete the walk along the beam. (c) Macroscopic and microscopic analyses of p53 mutant Nbn-CNS-del cerebella. ML, molecular layer; GCL, granule cell layer; PC, Purkinje cells. Scale bar at top, 0.5 cm. Original magnification in upper panel of H&E, ×2; lower panel of H&E and NeuN, ×4; D-28K, ×20.

a

c

bControl (n = 6)

A R T I C L E S©

2005

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

emed

icin

e

544 VOLUME 11 | NUMBER 5 | MAY 2005 NATURE MEDICINE

Balance beam tests. Following the previously described protocol38, we recorded the number of hindfoot missteps while the mouse walked once along the beam.

Histological and immunohistochemical analysis. Histological analysis was carried out on 3-µm sections stained with hematoxylin and eosin (H&E) or immunostained as described previously37. Antibodies included those specific for: human neuronal nuclei (NeuN, Chemicon), calbindin D-28K (Sigma-Aldrich), glial fibrillary acid protein (GFAP, Dakocytomation), human Ki67-MM1 (Novocastra), BrdU (Sigma-Aldrich), p53 (CM5, Novocastra) and p21 (Santa Cruz). For the in vivo proliferation assay, we injected newborn mice or pregnant females intraperitoneally with 50 µg BrdU/g body weight (Sigma-Aldrich). Embryos or brains were removed 6 h after injection and fixed in 4% buffered paraformaldehyde. The TUNEL assay was per-formed using the In Situ Cell Death Detection Kit, Fluorescein (Roche) according to the manufacturer’s instruction; 500 cells were scored for each group.

RT-PCR. Total RNA was isolated from the brain and cerebellum using Trizol reagent (Invitrogen). To measure Nbn, we used prim-ers Nbn102F (5′-ATGTGGAAGCTGCTCCCGGCC-3′) and Nbn648R (5′-CTGTTTCTTAGATTCAACTGC-3′); we used primers hprt-F (5′-GCTGGTGAAAAGGACCTCT-3′) and hprt-R (5′-CACAGGACTAGAACACCTGC-3′) as internal controls. Total RNA (2 µg) was used for reverse transcription using 200U of Superscript RnaseH Reverse Transcriptase (Invitrogen), and 2 µl of the reverse transcriptase reaction mixture was subsequently subjected to PCR amplification.

Western blotting analysis. Western blotting was carried out essentially as described previously7. We separated protein extracts (30 µg) by electrophoresis on 10% sodium dodecyl sulfate gel and transferred it onto a nylon membrane. We used antibodies specific for: NBS1 (1:500, NBS1-15CR)7, actin (1:20,000; MP Biochemicals), p53 (1:500, Ab-1, EMD Biosciences), p53-Ser15 phosphorylated (1:1,000; Cell Signaling Technology), P21 (1 µg/ml, BD Biosciences Pharmingen), Mre11 (1:5,000; Novus Biologicals) and Rad50 (1 µg/ml, Upstate Biotechnology).

Neurosphere cultures. The brains of E13.5 and newborn animals were disaggregated using a lysis solution consisting of 30 U/ml papain (Sigma-Aldrich), 240 µg/ml D,L-cysteine (Sigma-Aldrich), 400 µg/ml DNase I (Roche) in DMEM-NutMix/F12 (Invitrogen). After 1 h incubation at 37 °C, the lysis solution was neutralized with an inhibitor solution: 0.1125% ovomucoid trypsin inhibitor (Worthington Biochemicals), 0.0525% BSA (Sigma-Aldrich), 400 µg/ml DNase I (Roche) in L-15 medium (Invitrogen). We resuspended the cells and plated them in Dulbecco Modified Eagle Medium–NutMix/F12, supplemented with B27 (Invitrogen), 10 ng/ml EGF and 20 ng/ml basic fibroblast growth factor (Peprotech). The neuro-spheres were treated with 5 mM caffeine (Sigma-Aldrich) and 20 µM wortmannin (Sigma-Aldrich) for 4 h before protein extraction.

Cytogenetic analysis. Preparation of metaphase spreads, telomere staining and chromosome analysis were performed as described previously37.

Note: Supplementary information is available on the Nature Medicine website.

ACKNOWLEDGMENTSWe thank D. Galendo for her excellent assistance in the maintenance of the animal colonies, and C. Carreira, G. Hildebrand, N. Lyandrat and S. Roche for their excellent technical assistance. We also thank Y. Shiloh, M. Tommasino and E. F. Wagner for critical reading of the manuscript. We are grateful to J. Cheney and F. Corry for editing the manuscript. P.-O. Frappart is a recipient of a fellowship from the Comité Départmental de la Ligue Nationale contre le Cancer de Haute-Savoie (2000–2002), the Association de la Recherche pour le Cancer (ARC) (2002–2003) and the Ligue Nationale contre le Cancer (2004). This work was supported by the Deutsche Forschungsgemeinschaft (SFB577).

COMPETING INTERESTS STATEMENTThe authors declare that they have no competing financial interests.

Received 26 July 2004; accepted 25 February 2005Published online at http://www.nature.com/naturemedicine/

1. van der Burgt, I., Chrzanowska, K.H., Smeets, D. & Weemaes, C. Nijmegen breakage syndrome. J. Med. Genet. 33, 153–156 (1996).

2. Stewart, G.S. et al. The DNA double-strand break repair gene hMRE11 is mutated in

individuals with an ataxia-telangiectasia-like disorder. Cell 99, 577–587 (1999).3. Tauchi, H., Matsuura, S., Kobayashi, J., Sakamoto, S. & Komatsu, K. Nijmegen breakage

syndrome gene, NBS1, and molecular links to factors for genome stability. Oncogene 21, 8967–8980 (2002).

4. Shiloh, Y. ATM and related protein kinases: safeguarding genome integrity. Nat. Rev. Cancer. 3, 155–168 (2003).

5. Chun, H.H. & Gatti, R.A. Ataxia-telangiectasia, an evolving phenotype. DNA. Repair. (Amst). 3, 1187–1196 (2004).

6. D’Amours, D. & Jackson, S.P. The Mre11 complex: at the crossroads of dna repair and checkpoint signalling. Nat. Rev. Mol. Cell. Biol. 3, 317–327 (2002).

7. Dumon-Jones, V. et al. Nbn heterozygosity renders mice susceptible to tumor formation and ionizing radiation-induced tumorigenesis. Cancer. Res. 63, 7263–7269 (2003).

8. Zhu, J., Petersen, S., Tessarollo, L. & Nussenzweig, A. Targeted disruption of the Nijmegen breakage syndrome gene NBS1 leads to early embryonic lethality in mice. Curr. Biol. 11, 105–109 (2001).

9. Xiao, Y. & Weaver, D.T. Conditional gene targeted deletion by Cre recombinase dem-onstrates the requirement for the double-strand break repair Mre11 protein in murine embryonic stem cells. Nucleic. Acids. Res. 25, 2985–2991 (1997).

10. Kang, J., Bronson, R.T. & Xu, Y. Targeted disruption of NBS1 reveals its roles in mouse development and DNA repair. Embo. J. 21, 1447–1455 (2002).

11. Williams, B.R. et al. A murine model of Nijmegen breakage syndrome. Curr. Biol. 12, 648–653 (2002).

12. Theunissen, J.W. et al. Checkpoint failure and chromosomal instability without lymphoma-genesis in Mre11(ATLD1/ATLD1) mice. Mol. Cell. 12, 1511–1523 (2003).

13. McConnell, M.J. et al. Failed clearance of aneuploid embryonic neural progenitor cells leads to excess aneuploidy in the Atm-deficient but not the Trp53-deficient adult cerebral cortex. J. Neurosci. 24, 8090–8096 (2004).

14. Allen, D.M. et al. Ataxia telangiectasia mutated is essential during adult neurogenesis. Genes. Dev. 15, 554–566 (2001).

15. Barlow, C. et al. Atm-deficient mice: a paradigm of ataxia telangiectasia. Cell 86, 159–171 (1996).

16. Xu, Y. et al. Targeted disruption of ATM leads to growth retardation, chromosomal frag-mentation during meiosis, immune defects, and thymic lymphoma. Genes. Dev. 10, 2411–2422 (1996).

17. Borghesani, P.R. et al. Abnormal development of Purkinje cells and lymphocytes in Atm mutant mice. Proc. Natl. Acad. Sci. USA. 97, 3336–3341 (2000).

18. Betz, U.A., Vosshenrich, C.A., Rajewsky, K. & Muller, W. Bypass of lethality with mosaic mice generated by Cre-loxP-mediated recombination. Curr. Biol. 6, 1307–1316 (1996).

19. Fei, P. & El-Deiry, W.S. P53 and radiation responses. Oncogene 22, 5774–5783 (2003).

20. Slee, E.A., O’Connor, D.J. & Lu, X. To die or not to die: how does p53 decide? Oncogene 23, 2809–2818 (2004).

21. Groszer, M. et al. Negative regulation of neural stem/progenitor cell proliferation by the Pten tumor suppressor gene in vivo. Science 294, 2186–2189 (2001).

22. Shiloh, Y. & Kastan, M.B. ATM: genome stability, neuronal development, and cancer cross paths. Adv. Cancer. Res. 83, 209–254 (2001).

23. Maser, R.S., Zinkel, R. & Petrini, J.H. An alternative mode of translation permits produc-tion of a variant NBS1 protein from the common Nijmegen breakage syndrome allele. Nat. Genet. 27, 417–421 (2001).

24. Demuth, I. et al. An inducible null mutant murine model of Nijmegen breakage syndrome proves the essential function of NBS1 in chromosomal stability and cell viability. Hum. Mol. Genet. 13, 2385–2397 (2004).

25. Goldowitz, D. & Hamre, K. The cells and molecules that make a cerebellum. Trends. Neurosci. 21, 375–382 (1998).

26. Lee, Y. & McKinnon, P.J. ATM dependent apoptosis in the nervous system. Apoptosis 5, 523–529 (2000).

27. Abner, C.W. & McKinnon, P.J. The DNA double-strand break response in the nervous system. DNA. Repair. (Amst). 3, 1141–1147 (2004).

28. Xu, X. et al. Genetic interactions between tumor suppressors Brca1 and p53 in apoptosis, cell cycle and tumorigenesis. Nat. Genet. 28, 266–271 (2001).

29. Gao, Y. et al. A critical role for DNA end-joining proteins in both lymphogenesis and neurogenesis. Cell 95, 891–902 (1998).

30. Gao, Y. et al. Interplay of p53 and DNA-repair protein XRCC4 in tumorigenesis, genomic stability and development. Nature 404, 897–900 (2000).

31. Frank, K.M. et al. Late embryonic lethality and impaired V(D)J recombination in mice lacking DNA ligase IV. Nature 396, 173–177 (1998).

32. Frank, K.M. et al. DNA ligase IV deficiency in mice leads to defective neurogenesis and embryonic lethality via the p53 pathway. Mol. Cell. 5, 993–1002 (2000).

33. Lakin, N.D. & Jackson, S.P. Regulation of p53 in response to DNA damage. Oncogene 18, 7644–7655 (1999).

34. Lee, J.H. & Paull, T.T. Direct activation of the ATM protein kinase by the Mre11/Rad50/Nbs1 complex. Science 304, 93–96 (2004).

35. Stracker, T.H., Theunissen, J.W., Morales, M. & Petrini, J.H. The Mre11 complex and the metabolism of chromosome breaks: the importance of communicating and holding things together. DNA. Repair. (Amst). 3, 845–854 (2004).

36. Lee, Y. & McKinnon, P.J. DNA ligase IV suppresses medulloblastoma formation. Cancer. Res. 62, 6395–6399 (2002).

37. Tong, W.M. et al. Null mutation of DNA strand break-binding molecule poly(ADP-ribose) polymerase causes medulloblastomas in p53(–/–) mice. Am. J. Pathol. 162, 343–352 (2003).

38. Herson, P.S. et al. A mouse model of episodic ataxia type-1. Nat. Neurosci. 6, 378–383 (2003).

A R T I C L E S©

2005

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

emed

icin

e