p53-Binding protein 1 (53BP1) encodes a critical checkpoint protein that localizes to sites of DNA double-strand breaks (DSBs) and participates in DSB repair. Mice that are 53bp1 deficient or hemizygous have an increased incidence of lymphoid malignancies. However, 53BP1 abnormalities in primary human tumors have not been described. By combining high-density single nucleotide polymorphism (HD SNP) array data and gene expression profiles, we found 9 of 63 newly diagnosed human diffuse large B-cell lymphomas (DLBCLs) with single copy loss of the chromosome 15q15 region including the 53BP1 locus; these nine tumors also had significantly lower levels of 53BP1 transcripts. 53BP1 single copy loss found with the HD SNP array platform was subsequently confirmed by fluorescence in situ hybridization. These studies highlight the role of 53BP1 copy loss in primary human DLBCLs and the value of integrative analyses in detecting this genetic lesion in human tumors.
DNA-damage response mechanisms are crucial for genomic stability and defects in DNA-damage repair have been associated with cancer. A major form of DNA damage, DNA double-strand breaks (DSBs), results from exogenous and endogenous causes such as ionizing irradiation or free-radial accumulation. Developmentally programmed DNA breaks also occur in normal B and T lymphocytes during the assembly of immunoglobulin and cell surface antigen receptors. In activated mature B cells, specific DSBs are introduced into the immunoglobulin heavy chain (IgH) locus during IgH class-switch recombination (CSR) which involves the exchange of IgH constant region exons (Chaudhuri and Alt, 2004; Manis et al., 2004).
One of the major pathways by which DNA DSBs are repaired is non-homologous end joining (NHEJ). The introduction of DSBs activates phosphatidylinositol 3-kinase-like proteins that phosphorylate specific residues in DNA-damage response proteins including the histone variant, H2AX, and the p53-binding protein 1 (53BP1). Phosphorylated H2AX (γ-H2AX) and 53BP1 accumulate at presumptive sites of DNA damage, forming foci that flank DSBs in preparation for NHEJ (Mochan et al., 2004; Adams and Carpenter, 2006; Franco et al., 2006a).
Although 53BP1 was originally identified as a p53-interacting protein, its accumulation at DNA-damage foci is independent of p53 function. Recent studies indicate that 53BP1 is required for IgH CSR in normal activated B cells (Manis et al., 2004). 53BP1 has also been implicated in precancerous lesions in which aberrant cellular proliferation triggers DNA replication stress, the formation of DNA DSBs and the localization of 53BP1 to discrete nuclear foci (Gorgoulis et al., 2005). In these studies, the progression of preneoplastic lesions to overt carcinomas was associated with reduced expression of 53BP1 in some tumors and less evidence of apoptosis. These observations suggested a model in which decreased expression of 53BP1 and ineffective DNA DSB repair was followed by loss of cell-cycle checkpoints and malignant transformation (Gorgoulis et al., 2005).
Murine models of 53bp1 deficiency confirm the role of this DNA-damage response protein in tumor suppression. In initial murine models, 53bp1−/− animals had an increased incidence of early-onset thymic lymphomas (Ward et al., 2003b). In these studies, there was a close association between 53bp1 protein abundance and gene copy number, with 53bp1+/− cells from hemizygous animals expressing intermediate levels of the damage response protein (Ward et al., 2003b). With additional long-term follow-up, both 53bp1−/− and 53bp1+/− mice had an increased incidence of late-onset tumors (Ward et al., 2005). Furthermore, in a p53 null background, the loss of one or both copies of 53bp1 dramatically increased both the incidence and rapidity of onset of lymphoid malignancies including thymic and B-cell tumors (Ward et al., 2005). In an additional murine model of 53bp1 and p53 deficiency, only 53bp1−/−/p53−/− animals had an increased incidence of T- and B-cell lymphoma (Morales et al., 2006). Of note, when normal 53bp1−/− B cells were activated for CSR, these lymphocytes exhibited dramatically increased genomic instability and frequent IgH chromosomal breaks and translocations (Franco et al., 2006a, 2006b).
In additional analyses, the consequences of complete or partial 53bp1 deficiency on DSB repair were also examined in p53+/+ murine embryonic fibroblasts (Ward et al., 2005). Following irradiation, the numbers of γ-H2AX foci, chromosome breaks and aneuploid cells were highest in 53bp1−/− cells, lowest in 53bp1+/+ cells and intermediate in cells with 53bp1 single copy loss (Ward et al., 2005). Taken together, these data suggested that 53bp1 hemizygosity and the associated reduction in 53bp1 protein levels were sufficient to compromise DNA-damage repair and promote lymphoid tumor development (Ward et al., 2005).
Although in vitro functional analyses and murine genetic models strongly implicated 53bp1 copy loss in lymphomagenesis, there have been no descriptions of 53BP1 homo- or hemizygosity in primary human tumors including lymphoid malignancies. Herein, we report 53BP1 hemizygosity in a subset of human primary diffuse large B-cell lymphomas (DLBCLs) evaluated using integrated high-density single nucleotide polymorphism (HD SNP) array data and transcriptional profiles.
53BP1 copy loss and decreased expression in a subset of DLBCLs
Integrative analysis of the primary DLBCLs series revealed significantly decreased copy numbers of chromosome 15 band q15 and lower transcript levels of genes mapping to this region, including 53BP1. Using the previously described dChipSNP algorithm (Lin et al., 2004; Zhao et al., 2004; Garraway et al., 2005), seven of the 63 tumors were predicted to have only one copy of the 53BP1 gene, detected with the SNP_A-1659281 (rs10518815) and SNP_A-1657986 (rs555252) probes.
The recently developed genomic identification of significant targets in cancer (GISTIC) method confirmed the statistical significance of the copy number loss at chromosome 15q15, position 26.99–56.57 Mb (Figure 1a). The peak of statistical significance was found within the q15 band at 41.27–43.03 Mb, which includes 53BP1 (position 41.49–41.59 Mb) (Figures 1a and b and Supplementary Information). GISTIC identified nine tumors as having 53BP1 copy loss, including all seven cases predicted by the dChipSNP and two additional tumors.
Given the known role of 53bp1 copy loss in murine B- and T-cell lymphomas, we compared 53BP1 transcript abundance in the primary human DLBCLs with predicted 53BP1 copy loss (n=9) and normal copy numbers (n=54) (Figure 2). Tumors with predicted 53BP1 copy loss had significantly lower 53BP1 transcript levels (P=0.007) (Figure 2).
Recent genome-wide analyses of copy-number variations in normal populations revealed significant deletion variants including a region of 15q15 (Iafrate et al., 2004; Conrad et al., 2006; McCarroll et al., 2006; Redon et al., 2006). For this reason, we asked whether the region of copy loss in primary DLBCLs was within the described 15q15 region of normal variation, and found a partial overlap. However, the latest mapping information (Redon et al., 2006) indicates that the 53BP1 locus is not located within this overlap region.
53BP1 FISH analysis
To confirm the predicted 53BP1 copy loss in a cohort of DLBCLs, we analysed the 53BP1 locus by fluorescence in situ hybridization (FISH). Two primary tumors with predicted 53BP1 copy loss had available frozen tissue for the requisite touch preparations. The first DLBCL had only one signal with 53BP1 probe, whereas the tumor had two signals for the control probe at 15p11.2 (Figure 3a). The second tumor also had only one 53BP1 signal and either two or three control probe signals (Figure 3b). In contrast, all of the five DLBCLs with predicted normal 53BP1 copy numbers had two signals for both the 53BP1 and the control probes (data not shown). Taken together, these FISH data confirm the 53BP1 copy numbers in primary DLBCLs assessed by HD SNP arrays.
Analysis of remaining 53BP1 allele
We next assessed the integrity of the remaining 53BP1 allele in the nine DLBCLs with loss of one copy of 53BP1 by performing genomic PCR amplification and direct sequencing of all 53BP1 coding regions including exon–intron junctions. In seven tumors, the remaining 53BP1 allele had a normal sequence; two DLBCLs exhibited two previously described SNPs, which resulted in amino-acid changes, D353E and K1136Q (NCBI SNP database (dbSNP; http://www.ncbi.nlm.nih.gov/SNP)).
Since the K1136Q SNP is within the 53BP1 domain thought to be required for γ-H2AX binding in vitro (Ward et al., 2003a), we compared the ability of 53BP1-K1136Q and wild-type 53BP1 to associate with γ-H2AX and form irradiation-induced foci in 293T cells. The respective wild-type and K1136Q 53BP1 proteins performed similarly in these assays (Supplementary Information).
Genetic analysis of DLBCLs with 53BP1 copy loss
Since 53bp1−/− mice develop lymphoid malignancies with frequent chromosomal translocations (Ward et al., 2005; Morales et al., 2006), we compared the frequencies of t(14;18) and t(3;__) in primary DLBCLs with and without 53BP1 copy loss. Although three tumors with 53BP1 copy loss had either t(14;18) or t(3;…), these frequencies were similar to those of DLBCLs with normal 53BP1 copy numbers (data not shown).
53BP1 copy loss in DLBCL subtypes
Primary DLBCLs were previously assigned to one of three tumor subtypes, HR, OxPhos or BCR, on the basis of comprehensive transcriptional profiles (Monti et al., 2005). HR tumors exhibit a prominent host inflammatory/immune response and resemble T-cell/histiocyte-rich LBCLs, whereas OxPhos tumors have increased expression of genes involved in oxidative phosphorylation and more common structural abnormalities of intrinsic and extrinsic apoptotic pathway components (Monti et al., 2005; Takahashi et al., 2006). In contrast, BCR DLBCLs express higher levels of multiple BCR signaling cascade components, DNA-damage response proteins such as H2AX and B-cell transcription factors including BCL6; these tumors also have more frequent BCL6 translocations (Monti et al., 2005; Takahashi et al., 2006). Recent studies also indicate that BCR DLBCLs exhibit unique reliance upon BCL6 signaling and sensitivity to BCL6 peptide inhibitors (Polo et al., 2007).
Given the known role of BCL6 in modulating DNA-damage responses in normal and malignant germinal-center (GC) B cells (Phan and Dalla-Favera, 2004; Phan et al., 2005), we asked whether 53BP1 copy loss was more common in BCR DLBCLs. Twenty percent of BCR tumors (7/35) exhibited 53BP1 copy loss whereas only 1/15 (7%) OxPhos and 1/13 (8%) HR DLBCLs had this abnormality. The seven BCR tumors with 53BP1 copy loss also had significantly less abundant 53BP1 transcripts (P=0.02) (data not shown).
Herein, we identify single copy loss of 53BP1 as a frequent genetic abnormality in DLBCLs. Although previous murine genetic models strongly implicated 53bp1 copy loss in lymphomagenesis, the current study is the first report of 53BP1 hemizygosity in primary human lymphomas. In conjunction with murine analyses of 53bp1+/− cells, which have reduced levels of the damage response protein and increased chromosomal breakage and aneuploidy (Ward et al., 2003b, 2005), the current human studies suggest that 53BP1 haploinsufficiency may be a predisposing event in certain DLBCLs. Like other ‘caretaker’ genes involved in DNA repair, 53BP1 likely functions in a dose-dependent manner with haploinsufficiency leading to increased genomic instability, additional somatic mutations and lymphoid transformation (Fodde and Smits, 2002).
The frequent 53BP1 copy loss in BCR DLBCLs suggests that aberrant DNA-damage responses may play a particularly important role in these tumors, potentially in association with programmed DNA breaks of the immunoglobulin locus. As noted, BCR DLBCLs are uniquely reliant upon BCL6 (Polo et al., 2007), a transcription repressor which targets p53 and inactivates the p53-independent DNA-damage response pathway via Miz-1-dependent transcriptional repression of CDKN1A (Phan and Dalla-Favera, 2004; Phan et al., 2005; Wu and Jelinek, 2005). In normal and malignant GC B cells, BCL6 likely limits the p53-dependent and p53-independent DNA-damage responses associated with somatic hypermutation and CSR (Phan and Dalla-Favera, 2004; Phan et al., 2005; Wu and Jelinek, 2005). For these reasons, BCL6-dependent BCR lymphomas may be particularly sensitive to 53BP1 copy loss.
The combined analyses of HD SNP array data and transcriptional profiles underscores the value of whole-genome platforms in identifying previously unrecognized genetic lesions such as 53BP1 copy loss in DLBCL. Since decreased 53BP1 expression has been associated with the progression of preneoplastic lesions to overt carcinomas in certain solid tumors (Gorgoulis et al., 2005), it is possible that 53BP1 copy loss may be seen in other malignancies analysed with similar platforms.
Materials and methods
Sixty-three primary DLBCLs from a previously described series (Monti et al., 2005) had additional high-quality DNA available for analysis. Since paired normal DNAs from the DLBCL patients were unavailable, DNAs from a series of 180 normal individuals from the International HapMap Project (http://www.hapmap.org/) were used as reference samples.
High-density SNP arrays and detection of chromosomal copy number alterations
Genomic DNAs were extracted from frozen tissues and used for hybridization according to the manufacturer's protocols (Affymetrix Inc., Santa Clara, CA, USA). In brief, DNA was digested with restriction endonucleases (XbaI or HindIII), ligated to the adaptors and amplified by PCR. Thereafter, amplified DNA was fractionated, labeled with biotin and hybridized to the Gene Chip Mapping 100K set (Affymetrix Inc.), which contains a total of 1 15 593 SNPs with a median/average distance of 8.5/23.6 kb between SNPs.
The scanned data from the SNP arrays were processed with dChipSNP software (Li and Wong, 2001; Lin et al., 2004). After normalization, the intensity of each SNP in each tumor sample was rescaled using the average intensity of the same SNP across the 180 normal samples (Li and Wong, 2001). A genome-wide view of inferred copy number was generated by a Hidden Markov Model with dChipSNP as described previously (Lin et al., 2004; Zhao et al., 2004; Garraway et al., 2005).
A newly developed method GISTIC (Beroukhim et al., submitted) was also used for the detection of copy number alterations. The central feature of GISTIC is that it computes separate scores for amplifications and deletions for each SNP, taking into account both the frequency and average amplitude of these copy number changes. GISTIC assesses the statistical significance of each of these scores (under the assumption that the alterations occur independently of each other within a given sample), using false discovery rate (FDR) q-values (Benjamini and Hochberg, 1995; Storey, 2002) to correct for multiple hypothesis testing. For the application of GISTIC, the normalized raw data were first smoothed by means of a segmentation algorithm (Hupe et al., 2004). Regions of significant amplification/deletion were defined as contiguous sets of SNPs whose scores had FDR-corrected q-values below a specified threshold (FDR⩽0.25) (Beroukhim et al., submitted). A tumor was identified as having copy loss in a specific region if there were less than 1.3 calculated copies of the corresponding SNPs.
Integrated analysis of transcriptional profiles and HD SNP array data
Transcriptional profiling of the 63 primary DLBCLs was reported previously (Monti et al., 2005). To analyse the association between copy number abnormalities and gene transcript abundance, tumor samples were sorted into two classes, those with copy gain (or loss) and those with normal copy numbers. Thereafter, gene expression-based differential analysis was performed to identify genes within the altered region that were significantly up- (or down-) regulated.
Fluorescence in situ hybridization
Air-dried touch preparations were obtained from seven DLBCL frozen tissue samples including two cases with predicted 53BP1 copy loss and five cases with predicted normal 53BP1 copy numbers by HD SNP analysis. FISH was performed on sample nuclei using a BAC clone (PR11-355D13, BACPAC Resources Center at Children's Hospital Oakland, CA, USA) that spanned the 53BP1 locus at 15q15. The 53BP1 probe was labeled with SpectrumGreen and co-hybridized to nuclei with a chromosome 15 satellite III DNA probe at 15p11.2 (D15Z1, Abbott Molecular/Vysis Inc., Des Plaines, IL, USA) labeled with SpectrumOrange as a control probe. In each case, at least 30 cells with robust, discrete FISH signals were analysed by fluorescence microscopy after nuclear counterstaining with DAPI (4′,6-diamidine-2-phenylindole). BCL2 and BCL6 rearrangements were detected by FISH as described previously (Monti et al., 2005; Takahashi et al., 2006).
The same primary DLBCL genomic DNAs were used for HD SNP arrays and 53BP1 sequence analyses. The 53BP1 exons and exon–intron junctions were amplified by PCR using the tumor genomic DNA as a template. Primer sequences and conditions for PCR amplification are available upon request. All PCR products were purified by gel extraction and sequenced with an automated sequencer. To assess our sequencing results, we used NCBI sequences of 53BP1 cDNA (NM_005657) and genome (NT_010194.16) as references.
Adams MM, Carpenter PB . (2006). Tying the loose ends together in DNA double strand break repair with 53BP1. Cell Div 1: 19.
Benjamini Y, Hochberg Y . (1995). Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. J R Statist Soc B 57: 289–300.
Beroukhim R, Getz G, Nghiemphu L, Barretina J, Hsueh T, Linhart D et al. Genome-wide analysis of chromosomal aberrations in glioma. (Submitted).
Chaudhuri J, Alt FW . (2004). Class-switch recombination: interplay of transcription, DNA deamination and DNA repair. Nat Rev Immunol 4: 541–552.
Conrad DF, Andrews TD, Carter NP, Hurles ME, Pritchard JK . (2006). A high-resolution survey of deletion polymorphism in the human genome. Nat Genet 38: 75–81.
Fodde R, Smits R . (2002). Cancer biology A matter of dosage. Science 298: 761–763.
Franco S, Alt FW, Manis JP . (2006a). Pathways that suppress programmed DNA breaks from progressing to chromosomal breaks and translocations. DNA Repair 5: 1030–1041.
Franco S, Gostissa M, Zha S, Lombard DB, Murphy MM, Zarrin AA et al. (2006b). H2AX prevents DNA breaks from progressing to chromosome breaks and translocations. Mol Cell 21: 201–214.
Garraway LA, Widlund HR, Rubin MA, Getz G, Berger AJ, Ramaswamy S et al. (2005). Integrative genomic analyses identify MITF as a lineage survival oncogene amplified in malignant melanoma. Nature 436: 117–122.
Gorgoulis VG, Vassiliou LV, Karakaidos P, Zacharatos P, Kotsinas A, Liloglou T et al. (2005). Activation of the DNA damage checkpoint and genomic instability in human precancerous lesions. Nature 434: 907–913.
Hupe P, Stransky N, Thiery JP, Radvanyi F, Barillot E . (2004). Analysis of array CGH data: from signal ratio to gain and loss of DNA regions. Bioinformatics 20: 3413–3422.
Iafrate AJ, Feuk L, Rivera MN, Listewnik ML, Donahoe PK, Qi Y et al. (2004). Detection of large-scale variation in the human genome. Nat Genet 36: 949–951.
Li C, Wong WH . (2001). Model-based analysis of oligonucleotide arrays: expression index computation and outlier detection. Proc Natl Acad Sci USA 98: 31–36.
Lin M, Wei LJ, Sellers WR, Lieberfarb M, Wong WH, Li C . (2004). dChipSNP: significance curve and clustering of SNP-array-based loss-of-heterozygosity data. Bioinformatics 20: 1233–1240.
Manis JP, Morales JC, Xia Z, Kutok JL, Alt FW, Carpenter PB . (2004). 53BP1 links DNA damage-response pathways to immunoglobulin heavy chain class-switch recombination. Nat Immunol 5: 481–487.
McCarroll SA, Hadnott TN, Perry GH, Sabeti PC, Zody MC, Barrett JC et al. (2006). Common deletion polymorphisms in the human genome. Nat Genet 38: 86–92.
Mochan TA, Venere M, DiTullio Jr RA, Halazonetis TD . (2004). 53BP1, an activator of ATM in response to DNA damage. DNA Repair 3: 945–952.
Monti S, Savage KJ, Kutok JL, Feuerhake F, Kurtin P, Mihm M et al. (2005). Molecular profiling of diffuse large B-cell lymphoma identifies robust subtypes including one characterized by host inflammatory response. Blood 105: 1851–1861.
Morales JC, Franco S, Murphy MM, Bassing CH, Mills KD, Adams MM et al. (2006). 53BP1 and p53 synergize to suppress genomic instability and lymphomagenesis. Proc Natl Acad Sci USA 103: 3310–3315.
Phan RT, Dalla-Favera R . (2004). The BCL6 proto-oncogene suppresses p53 expression in germinal-centre B cells. Nature 432: 635–639.
Phan RT, Saito M, Basso K, Niu H, Dalla-Favera R . (2005). BCL6 interacts with the transcription factor Miz-1 to suppress the cyclin-dependent kinase inhibitor p21 and cell cycle arrest in germinal center B cells. Nat Immunol 6: 1054–1060.
Polo JM, Juszczynski P, Monti S, Cerchietti L, Ye K, Greally JM et al. (2007). A transcriptional signature with differential expression of BCL6 target genes accurately identifies BCL6-dependent diffuse large B-cell lymphomas. Proc Natl Acad Sci USA 104: 3207–3212.
Redon R, Ishikawa S, Fitch KR, Feuk L, Perry GH, Andrews TD et al. (2006). Global variation in copy number in the human genome. Nature 444: 444–454.
Storey JD . (2002). A direct approach to false discovery rates. J Roy Statisti Soc B 64: 479–498.
Takahashi H, Feuerhake F, Kutok JL, Monti S, Dal Cin P, Neuberg D et al. (2006). FAS death domain deletions and cellular FADD-like interleukin 1beta converting enzyme inhibitory protein (long) overexpression: alternative mechanisms for deregulating the extrinsic apoptotic pathway in diffuse large B-cell lymphoma subtypes. Clin Cancer Res 12: 3265–3271.
Ward IM, Difilippantonio S, Minn K, Mueller MD, Molina JR, Yu X et al. (2005). 53BP1 cooperates with p53 and functions as a haploinsufficient tumor suppressor in mice. Mol Cell Biol 25: 10079–10086.
Ward IM, Minn K, Jorda KG, Chen J . (2003a). Accumulation of checkpoint protein 53BP1 at DNA breaks involves its binding to phosphorylated histone H2AX. J Biol Chem 278: 19579–19582.
Ward IM, Minn K, van Deursen J, Chen J . (2003b). p53 Binding protein 53BP1 is required for DNA damage responses and tumor suppression in mice. Mol Cell Biol 23: 2556–2563.
Wu X, Jelinek DF . (2005). A-Miz-ing BCL6. Nat Immunol 6: 964–966.
Zhao X, Li C, Paez JG, Chin K, Janne PA, Chen TH et al. (2004). An integrated view of copy number and allelic alterations in the cancer genome using single nucleotide polymorphism arrays. Cancer Res 64: 3060–3071.
This study was supported by grants from National Cancer Institute PO1 CA92625 (KT, SM, JM, JA, FA, TG, MS) and Department of Defense PC040638 (RB).
About this article
Cite this article
Takeyama, K., Monti, S., Manis, J. et al. Integrative analysis reveals 53BP1 copy loss and decreased expression in a subset of human diffuse large B-cell lymphomas. Oncogene 27, 318–322 (2008). https://doi.org/10.1038/sj.onc.1210650
- DNA damage
- genetic abnormalities
- high-density single nucleotide polymorphism array
Cancer Treatment Reviews (2020)
Mechanism for IL-15–Driven B Cell Chronic Lymphocytic Leukemia Cycling: Roles for AKT and STAT5 in Modulating Cyclin D2 and DNA Damage Response Proteins
The Journal of Immunology (2019)
Characterization of genomic imbalances in diffuse large B-cell lymphoma by detailed SNP-chip analysis
International Journal of Cancer (2015)