Mutations in the Nijmegen Breakage Syndrome Gene in Medulloblastomas

Medulloblastoma is an invasive embryonal tumor of the cerebellum and the most common malignant brain tumor in children (1–3). Although the majority of medulloblastomas occur sporadically, some manifest within familial cancer syndromes, including the nevoid basal cell carcinoma (Gorlin) syndrome associated with germ-line PTCH mutations (4) and Turcot syndrome type 2 caused by germ-line APC mutations (5). Nijmegen breakage syndrome (NBS) is a rare autosomal recessive disease that presents with clinical features such as microcephaly, mental and growth retardation, immunodeficiency, radiosensitivity, and increased risk for cancer, particularly B-cell non-Hodgkin's lymphoma (6–8). The gene mutated in this syndrome is NBS1, located at chromosome 8q21 (9). The most frequently reported NBS1 mutation, identified in 90% of NBS patients, particularly those from Slavic populations primarily from Poland, Ukraine, and the Czech Republic, is a 5-nucleotide deletion in exon 6 (657del5; ref. 10).

There are two reports on NBS patients with medulloblastoma. Distel et al. (11) reported a 7-year-old boy with medulloblastoma who suffered from severe side effects during and after postoperative radiotherapy and chemotherapy. This patient had a germ-line homozygous mutation (657del5) of the NBS1 gene (11). Bakhshi et al. (12) described a 3-year-old medulloblastoma patient with microcephaly, facial dysmorphism, and growth retardation, who also suffered from severe side effects after craniospinal irradiation. This patient had two NBS1 germ-line mutations (675del5 and 1142delC). These observations raised the question of whether NBS1 mutations may play a role in the pathogenesis of sporadic medulloblastomas. In the present study, we therefore screened 42 medulloblastomas for NBS1 mutations.

The NBS1 protein is one substrate of the ataxia telangiectasia mutated (ATM) kinase, which signals all three cell cycle checkpoints after DNA double-strand break damage (13). Thus, NBS1 plays important roles in ATM-dependent DNA damage response by forming a complex with MRE11 and RAD50 (13–15). NBS1 also acts upstream, promoting ATM activation, ATM recruitment to breaks, and ATM accessibility to substrates (16). NBS1 binds directly to MDM2, a negative regulator of TP53, and colocalizes to sites of DNA damage following γ-irradiation (17), and MDM2 overexpression inhibits DNA double-strand break repair associated with NBS1 (17). In NBS1 mutant/mutant mice expressing a NH₂-terminally truncated NBS1 at low levels, p53 deficiency greatly facilitated the development of thymic lymphomas (13). Thus, there may be functional interactions between NBS1 and the TP53 pathways. We therefore also screened medulloblastomas for TP53 mutations, which reportedly occur in ∼5% to 10% of cases (1, 18), as well as alterations of other genes in the TP53 pathway (MDM2 amplification, p14^ARF homozygous deletion, and promoter methylation).

Tumor samples. Of 42 cases analyzed in this study, 37 were diagnosed at the Department of Neuropathology, University Hospital Zurich. Of these, 2 cases had microcephaly. Five additional cases were patients with medulloblastoma and microcephaly enrolled in a randomized, prospective, multicenter trial (HIT 2000) conducted by the German Society of Pediatric Hematology and Oncology. The mean age of all cases was 13.8 ± 12.4 years (range, 1-60 years). Thirteen (31%) patients were adults (ages ≥18 years); 30 were males and 12 were females. Tumors were fixed in buffered formalin and embedded in paraffin. Genomic DNA was extracted from paraffin sections as described previously (19).

Single-stranded conformational polymorphism analysis and direct DNA sequencing for NBS1 mutations. Single-stranded conformational polymorphism (SSCP) analysis was carried out to prescreen for mutations in all 16 exons of the NBS1 gene. Primers for SSCP-PCR for NBS1 were designed with software Primer 5 (Premier Biosoft International; Table 1). PCR was done in a total volume of 10 μL, consisting of 1 μL DNA solution (∼100 ng/μL), 0.5 units Platinum Taq DNA polymerase (Invitrogen), 0.1 μCi [α-³³P]dCTP (ICN Biomedicals; specific activity, 3,000 Ci/mmol), 1 to 4 mmol/L MgCl₂, 0.1 to 0.2 mmol/L of each deoxynucleotide triphosphate, 0.2 to 0.4 μmol/L of each primer, 10 mmol/L Tris-HCl (pH 8.3), and 50 mmol/L KCl in a thermal cycler (Biometra) with an initial denaturing step at 95°C for 5 min followed by 37 to 40 cycles of denaturation at 95°C for 50 s, annealing at 45°C to 54°C for 60 s, extension at 72°C for 60 s, and a final extension at 72°C for 5 min. After PCR amplification, 10 μL PCR products were mixed with 20 μL loading buffer (0.02 N NaOH, 95% formamide, 20 mmol/L EDTA, 0.05% xylene cyanol and bromophenol blue), denatured at 95°C for 10 min, and quenched on ice. Then, 5.5 μL above mixture was loaded onto a 12.5% polyacrylamide nondenaturing gel containing 10% glycerol. Electrophoresis was done at 45 W for 3.5 to 4.5 h at room temperature with cooling by fan. Gels were dried at 80°C and autoradiography was done for 24 to 36 h. Samples exhibiting mobility shifts in SSCP analyses were subsequently reamplified using the same primers as for SSCP and sequenced using the Big Dye Terminator cycle sequencing kit (ABI PRISM; Applied Biosystems) in an ABI 3100 PRISM DNA sequencer (Applied Biosystems).

SSCP analysis and direct DNA sequencing for TP53 mutations. Prescreening for mutations in exons 5 to 8 of the TP53 gene by PCR-SSCP analysis was carried out as described previously (19). Samples that showed mobility shifts were further analyzed by direct DNA sequencing as described above.

Differential PCR for p14^ARF homozygous deletion. To screen for p14^ARF homozygous deletion, differential PCR was carried out with a GAPDH sequence as reference as described previously (20). Sequences of primers were as follows: 5′-GAGTGAGGGTTTTCGTGGTT-3′ (forward) and 5′-GCCTTTCCTACCTGGTCTTC-3′ (reverse), which cover exon 1β of the p14^ARF gene, and 5′-AACGTGTCAGTGGTGGACCTG-3′ (forward) and 5′-AGTGGGTGTCGCTGTTGAAGT-3′ (reverse) for the GAPDH sequence. PCR was done in a total volume of 10 μL, consisting of 1 μL DNA solution (concentration, ∼100 ng/μL), 0.5 units Platinum Taq DNA polymerase, 1.5 mmol/L MgCl₂, 0.2 mmol/L of each deoxynucleotide triphosphate, 0.25 μmol/L primers for p14^ARF, 0.045 μmol/L primers for GAPDH, 10 mmol/L Tris-HCl (pH 8.3), and 50 mmol/L KCl in the T3 thermal cycler with an initial denaturing step at 95°C for 5 min followed by 30 cycles of denaturation at 95°C for 50 s, annealing at 60°C for 60 s, extension at 72°C for 60 s, and a final extension for 3 min at 72°C. Electrophoresis was done with the PCR products on 8% acrylamide gels and the gels were photographed with a DC120 Zoom Digital Camera (Kodak). The density of each PCR fragment was estimated using Kodak Digital Science ID Image Analysis Software (Kodak). Samples presenting <20% of the control signals were considered as having p14^ARF homozygous deletion as described previously (20).

Methylation-specific PCR for p14^ARF promoter methylation. To screen for p14^ARF promoter methylation, methylation-specific PCR was carried out as described previously (20). The sequences of primers were as follows: 5′-GTGTTAAAGGGCGGCGTAGC-3′ (forward) and 5′-AAAACCCTCACTCGCGACGA-3′ (reverse) for methylated PCR and 5′-TTTTTGGTGTTAAAGGGTGGTGTAGT-3′ (forward) and 5′-CACAAAAACCCTCACTCACAACAA-3′ (reverse) for unmethylated PCR. The sodium bisulfite modification was done using the EZ DNA modification kit (Zymo Research) according to the manufacturer's protocol. Methylated and unmethylated PCR were done in a 10 μL volume, containing 2 μL bisulfite-modified DNA, PCR buffer [10 mmol/L Tris (pH 8.3) and 50 mmol/L KCl], MgCl₂ (1.5 mmol/L for methylated PCR and 2.0 mmol/L for unmethylated PCR), 0.3 mmol/L of each deoxynucleotide triphosphate, 0.5 μmol/L of each primer, and 0.5 units Platinum Taq DNA polymerase in the Biometra T3 thermal cycler with initial denaturing at 95°C for 10 min followed by 40 cycles of denaturing at 95°C for 45 s and annealing for 45 s at 62°C for methylated reaction and 60°C for unmethylated reaction, with extension at 72°C for 45 s. A final extension at 72°C was added for 10 min after the last cycle. Positive controls (methylated DNA) and unmethylated DNA (blood samples from healthy individuals) were treated with sodium bisulfite as described above and were included in each experiment. Electrophoresis was done with the amplified products on 4% agarose gels, with visualization by ethidium bromide.

Differential PCR for MDM2 amplification. To screen for MDM2 gene amplification, differential PCR was carried out with the dopamine receptor sequence as reference as described previously (21). The sequences of primers were 5′-GAGGGCTTTGATGTTCCTGA-3′ (forward) and 5′-GCTACTAGAAGTTGATGGC-3′ (reverse) for MDM2 and 5′-CCACTGAATCTGTCCTGGTATG-3′ (forward) and 5′-GTGTGGCATAGTAGTTGTAGTGG-3′ (reverse) for dopamine receptor. PCR was done in a total volume of 10 μL, consisting of 1 μL DNA solution (∼100 ng/μL), 0.5 units Platinum Taq DNA polymerase, 1.5 mmol/L MgCl₂, 0.2 mmol/L of each deoxynucleotide triphosphate, 0.4 μmol/L primers for MDM2, 0.5 μmol/L primers for dopamine receptor, 10 mmol/L Tris-HCl (pH 8.3), and 50 mmol/L KCl in the Biometra T3 thermal cycler with an initial denaturing step at 95°C for 5 min followed by 30 cycles of denaturation at 95°C for 50 s, annealing at 55°C for 60 s, polymerization at 72°C for 60 s, and a final extension at 72°C for 3 min. Electrophoresis was done with PCR products on 8% acrylamide gels and the gels were photographed with a DC 120 Zoom Digital Camera. The density of each PCR fragment was estimated using Kodak Digital Science ID Image Analysis Software. Samples with a MDM2/dopamine receptor ratio of >2.5 were considered to show MDM2 amplification as described previously (21).

Statistical analysis. The Fisher's exact test was carried out to analyze the significance of the associations between TP53 mutations and NBS1 mutations and between NBS1 mutations and microcephaly. The Stata 10.0 t test was carried out to compare the mean age of patients with and without NBS1 mutations. Statistical analyses were carried out using software Stata 9.1 (StataCorp).

NBS1 mutations. SSCP followed by direct sequencing in all 16 NBS1 exons revealed a total of 15 miscoding NBS1 mutations in 7 of 42 (17%) medulloblastomas (Table 2; Fig. 1). Three of the tumors showed more than two mutations. NBS1 miscoding mutations were located in exons 2, 4, 8, 10, and 14 and introns 2, 6, 7, and 9 (Table 2; Fig. 2). Ten mutations were missense point mutations and 5 were splicing mutations. All, except one mutation (14 of 15; 93%), were G:C→A:T transitions. Two medulloblastomas also contained a silent mutation (codon 181, TTG→CTG, Leu→Leu in case 691; codon 313, GCG→GCC, Ala→Ala in case 315).

In two cases with NBS1 mutations (cases 342 and 691), adjacent nontumorous brain tissue was available. No NBS1 mutations were found in DNA from normal tissue, indicating that the NBS1 mutations in these medulloblastomas were somatic (Fig. 1).

In one patient (case 692), two separate tumor samples taken by surgical intervention at ages 14 and 16 years were analyzed. The first biopsy showed one NBS1 mutation, whereas the second contained the same plus three additional NBS1 mutations and a TP53 mutation (Table 2).

Histologically, two tumors (cases 315 and 692) with NBS1 mutations were desmoplastic/nodular medulloblastomas, according to the new WHO classification (1), whereas other tumors with NBS1 mutations were classic medulloblastomas.

All of the patients with NBS1 mutations were children ages ≤14 years at the first diagnosis. The mean age of patients with NBS1 mutation was 5.7 ± 4.5 years at diagnosis, significantly lower than patients without mutation (15.4 ± 13.0 years; P = 0.0305).

Three of 7 patients with NBS1 mutations and 4 of 35 cases without NBS1 mutations had microcephaly (P = 0.077).

Common NBS1 single nucleotide polymorphisms were identified at similar frequencies as reported previously in healthy individuals (22), including L34L (27%), E185Q (33%), D399D (32%), and P672P (23%). Intronic single nucleotide polymorphisms, which have been reported previously (23–25), were also observed: IVS9+18C/T (30%), IVS9+91C/A (30%), IVS13-7A/G (23%), and IVS14-30A/T (27%). In addition, a rare single nucleotide polymorphism in intron 13 (IVS14-61A/T) was found in one medulloblastoma.

TP53 mutations. Of 42 cases of medulloblastomas analyzed, 5 (12%) cases contained a TP53 mutation. Four mutations were missense point mutations and one was a splicing mutation (Table 2). Of 5 medulloblastomas with TP53 mutation, 4 (80%) contained NBS1 mutations. Of 7 medulloblastomas with NBS1 mutations, 4 (57%) showed concomitant presence of TP53 mutation. There was a significant association between TP53 mutations and NBS1 mutations (P = 0.001).

Alterations in the p14^ARF and MDM2 genes p14^ARF. Promoter methylation was detected by methylation-specific PCR in 3 of 42 (7%) medulloblastomas analyzed. No p14^ARF homozygous deletion or MDM2 amplification was found in any of the medulloblastomas.

The most frequent genetic alteration in medulloblastomas is isochromosome 17q, which occurs in up to 50% of cases (26, 27), whereas loss of heterozygosity on 17p13.3 is observed in 30% to 50% of medulloblastomas (1, 28, 29). Accordingly, comparative genomic hybridization revealed gain of 17q and loss of 17p as most frequent chromosomal imbalance (30–32). Mutations in the PTCH gene are found in 8% to 20% of cases, particularly in the desmoplastic variant (33–35). PTCH is an inhibitor of the Hedgehog signaling pathway, which is a major mitogenic factor in the development of the embryonic external granular cell layer of the cerebellum, from which most medulloblastomas are considered to originate (36). TP53 mutations are infrequent (5-10% of cases; refs. 1, 18), but most mouse models of medulloblastoma with high penetrance involve loss of p53 expression in addition to other knockout genes, including Ptch (37), Rb (38), PARP-1 (39), and Lig4 (40). Other pathway involved in the development of medulloblastomas are Wnt signaling, with APC, β-catenin, or AXIN1 mutations in 4% to 10% of cases (41–45). Although infrequent, amplification of genes such as c-Myc, Gli1, Pax5, PDGF, SPARK, and Notch2 has also been found in medulloblastomas (36). Our present data provide the first evidence that somatic NBS1 mutations are involved in the development of medulloblastomas. All of the NBS1 mutations were found in medulloblastomas in children.

The main functional domains of NBS1 in DNA damage responses comprise the FHA domain (amino acids 24-100, corresponding to exons 2-3), the breast cancer COOH-terminal (BRCT) domain (residues 105-190, exons 4-5), and the MRE11-binding domain (residues 601-700, binding sites 665-693, exons 13-14), which play important functional roles in cell survival after exposure to irradiation (46, 47), and ATM phosphorylation sites (Ser²⁷⁸ in exon 7 and Ser³⁴³ in exon 9; ref. 46). The MRE11-NBS1-RAD50 complex plays an important role in DNA damage-induced checkpoint control and DNA repair (47, 48). Both FHA and BRCT domains are essential for DNA damage responses, including IRIF formation, S-phase checkpoint activation, and nuclear focus formation after irradiation, and play a crucial role in cell survival after radiotherapy (47, 49). NBS1 also functions as a downstream mediator of ATM function (47, 48).

We were able to assess normal tissues from two cases with NBS1 mutations, showing that these NBS1 mutations are somatic, but it cannot be ruled out that the mutations are also present in normal tissues and constitute previously unidentified single nucleotide polymorphisms in remaining five cases. However, in the present study, three NBS1 miscoding mutations in exon 2 (codon 26, GTT→ATT, Val→Ile; codon 57, CTG→ATG, Leu→Met) and exon 4 (codon 148, ACT→ATT, Thr→Ile) were located in the FHA and BRCT domains; another mutation in exon 14 (codon 711, CAT→TAT, His→Tyr) was close to the MRE11-binding site. Furthermore, three NBS1 mutations in exon 10 (codon 383, GAA→AAA, Glu→Lys; codon 407, TGC→TAC, Cys→Tyr; codon 427, CCC→CTC, Pro→Leu) and three in exon 8 (codon 308, GCA→ACA, Ala→Thr; codon 311, GGA→AGA, Gly→Arg; codon 319, ACA→ATA, Thr→Ile) were located close to the domain of ATM phosphorylation sites. One mutation (codon 148, ACT→ATT, Thr→Ile) was located in the BRCT domain, which facilitates the interaction between NBS1 and BRCA1, forming a BRCA1-associated genome surveillance complex that is responsible for recognition and repair of aberrant DNA (50). BRCT domains are required for optimal chromatin association of the MRE11-NBS1-RAD50 complex, irradiation-induced phosphorylation of NBS1, and S-phase checkpoint activation (47). The BRCT domain is a protein-protein interaction domain (47), in which any amino acid change may interfere with the interaction. Thus, the NBS1 mutations identified in the present study are likely to play significant roles in the pathogenesis of medulloblastomas. It is notable that any of the NBS1 mutations detected in medulloblastomas in the present study are novel mutations and have not been reported previously as germ-line mutations in NBS patients (10, 12, 51). The absence of a common mutation in exon 6 (657del5) reported previously in NBS patients (10) in any of the medulloblastomas in the present study suggests the importance of screening the whole NBS1 gene in human neoplasms.

NBS1 missense and splicing mutations have been reported in sporadic cancers of gastrointestinal cancer (52), breast cancer (53, 54), and acute lymphoblastic leukemia (55, 56). Several studies have shown that missense and splicing NBS1 mutations in sporadic tumors have functional consequences for the NBS1 protein. Heterozygous NBS1 splicing mutation (IVS11+2insT), which was detected in sporadic gastric, colorectal, and lung cancer, led to the loss of the MRE11- and ATM-binding sites at the COOH terminus with several functional abrogations (defective in crucial binding to MRE11, MDC1, BRCA1, and wild-type NBS1) and caused impaired ATM phosphorylation in response to low-dose irradiation in a heterozygous state (52). In breast cancer cells with a NBS1 missense mutation (R215W), levels of NBS1/p95 protein and radiation-induced phosphorylation of Nbs1/p95 (Ser³⁴³) were reduced to 70% and 60% of wild-type, respectively (53). Missense NBS1 mutations (L150F and I171V) were associated with chromosomal instability in sporadic breast cancer (54) and aplastic anemia (56).

The present study also notably shows the copresence of mutations in the NBS1 and TP53 genes. Of 5 medulloblastomas with a TP53 mutation, 4 (80%) contained NBS1 mutations, whereas TP53 mutations were copresent in 4 of 7 (57%) medulloblastomas with NBS1 mutations. There was a significant association between TP53 mutations and NBS1 mutations (P = 0.001). This finding provides evidence that simultaneous disruptions of NBS1 and TP53 functions may constitute a novel genetic pathway in the pathogenesis of a subset of medulloblastomas. Simultaneous occurrence of NBS1 and TP53 mutations may facilitate the development of medulloblastomas, as has been observed in NBS1 mutant/mutant mice expressing a NH₂-terminally truncated NBS1 at low levels, in which p53 deficiency greatly facilitated the development of thymic lymphomas (13). Alternatively, TP53 mutations may be late events acquired after NBS1 mutations. In the present study, in one patient (case 692), two tumor samples taken by surgical intervention at ages 14 and 16 years were analyzed. The first biopsy showed one NBS1 mutation, whereas the second contained the same plus three additional NBS1 mutations and a TP53 mutation. The initial NBS1 mutation may have caused genomic instability that led to additional alterations in the NBS1 and TP53 genes.

In conclusion, we provide evidence of medulloblastomas characterized by NBS1 mutations typically associated with mutational inactivation of the TP53 gene. It remains to be shown in a large clinical trial whether this subset of medulloblastomas differs with respect to response to therapy or other clinical parameters.

No potential conflicts of interest were disclosed.

We thank Wiebke Treulieb for clinical data of medulloblastoma patients in HIT study. Dr. Jian Huang was supported by an IARC Postdoctoral Fellowship.