53BP1 loss rescues embryonic lethality but not genomic instability of BRCA1 total knockout mice

BRCA1 is critical for DNA double-strand break (DSB) repair by homologous recombination (HR). BRCA1 deficient mice are embryonic lethal. Previous studies have shown that 53BP1 knockout (KO) rescues embryonic lethality of BRCA1 hypomorphic mutant mice by restoring HR. Here, we show that 53BP1 KO can partially rescue embryonic lethality of BRCA1 total KO mice, but HR is not restored in BRCA1-53BP1 double knockout (DKO) mice. As a result, BRCA1-53BP1 DKO cells are extremely sensitive to PARP inhibitors (PARPi). In addition to HR deficiency, BRCA1-53BP1 DKO cells have elevated microhomology-mediated end joining (MMEJ) activity and G2/M cell cycle checkpoint defects, causing severe genomic instability in these cells. Interestingly, BRCA1-53BP1 DKO mice rapidly develop thymic lymphoma that is 100% penetrant, which is not observed in any BRCA1 mutant mice rescued by 53BP1 KO. Taken together, our study reveals that 53BP1 KO can partially rescue embryonic lethality caused by complete BRCA1 loss without rescuing HR-related defects. This finding suggests that loss of 53BP1 can support the development of cancers with silenced BRCA1 expression without causing PARPi resistance.


Introduction
Cells in our body constantly encounter all kinds of DNA damage that are repaired by different DNA damage repair pathways. DNA double-strand break (DSB) is one of the most deleterious forms of DNA damage. Homologous recombination (HR) is an important pathway for DSB repair that requires loading of key enzyme RAD51 to DSBs for repair using sister chromatids as templates [1]. HR deficiency impairs the maintenance of genomic stability and often leads to tumorigenesis [2]. BRCA1 is a key protein for HR whose loss or mutation leads to HR deficiency and genomic instability [3]. BRCA1 mutations are frequently identified in familial breast and ovarian cancers [4]. BRCA1 loss or mutation leads to PARP inhibitor (PARPi) sensitivity, which has been exploited to treat BRCA1 mutant cancers [5]. Besides tightly linked to tumorigenesis, HR is also important for embryonic development. BRCA1 deficient mice are embryonic lethal [6], but the exact role of BRCA1 in embryonic development is not clear.
In the past decades, studies have provided mechanistic insights into how BRCA1 promotes HR. BRCA1 interacts with and recruits PALB2 to DSBs, which is essential for BRCA2/RAD51 loading [7][8][9]. Besides HR, DSB can be repaired by nonhomologous end joining (NHEJ). BRCA1 also functions in DSB repair pathway choices between HR and NHEJ [10,11]. BRCA1 removes 53BP1 from DSB ends and promotes DNA end resection, a prerequisite for HR. It is believed that 53BP1 KO lifts the barrier on DNA end resection and restores HR in the absence of intact BRCA1 [12][13][14][15].
The identification of BRCA1's function in DSB repair pathway choice originates from the finding that 53BP1 KO can fully rescue the embryonic lethality of Brca1 Δ11/Δ11 and Brca1 Δ2/Δ2 mice [12][13][14][15]. The HR deficiency and genomic instability are also fully rescued so that both Brca1 Δ11/Δ11 ; Trp53bp1 −/− mice and Brca1 Δ2/Δ2 ;Trp53bp1 −/− mice have normal life span and no elevated tumor incidences than wild-type (WT) mice [12][13][14][15]. However, both Brca1 Δ11/Δ11 and Brca1 Δ2/Δ2 mice are Brca1 hypomorphic mutant but not total KO mice. In these mice, truncated BRCA1Δ11 and BRCA1Δ2 proteins are produced, both of which retain the BRCT domain that localizes the protein to DSBs and the coiled coil domain that interacts with PALB2 (Fig. 1a). A recent study has shown that 53BP1 KO can also fully rescue embryonic lethality of Brca1 ΔC/ΔC mice [16]. Interestingly, HR deficiency and genomic instability are only mildly a Domain structures of BRCA1 WT, Δ11, Δ2, ΔC, and Δ5-13 proteins. CC coiled coil. b Summary of mice at 3 weeks or embryos at 13.5 dpc of indicated genotypes. c Western blotting analyses of BRCA1, 53BP1, and BARD1 protein in MEFs with indicated genotypes. α-tubulin was used as loading control. Asterisk, nonspecific band. Arrowhead, BRCA1 isoforms or degraded BRCA1 proteins. d Immunofluorescence staining analyses of BRCA1, 53BP1, and BARD1 foci in MEFs with indicated genotypes 6 h after 10 Gy IR exposure. γH2AX marks sites of DNA damage. Hoechst 33342 marks nucleus. Scale bars, 10 μm. e Quantification of BRCA1, 53BP1, and BARD1 foci. Error bars represent SEM from three independent experiments. *p < 0.05; **p < 0.01. rescued in Brca1 ΔC/ΔC ;Trp53bp1 −/− mice and these mice have shorter life span [16]. Theoretically, a truncated BRCA1ΔC protein that deletes both the BRCT and the coiled coil domain is produced in Brca1 ΔC/ΔC mice, but no truncated protein could be detected, suggesting that Brca1 ΔC/ΔC mice are close to BRCA1 total KO mice. However, it is still possible that BRCA1 is not completely absent in Brca1 ΔC/ΔC mice, in which the truncated BRCA1ΔC proteins might express at low levels and contribute to the viability of Brca1 ΔC/ΔC ;Trp53bp1 −/− mice. Therefore, it remains to be studied if 53BP1 KO can rescue the lethality of bona fide BRCA1 total KO mice.
In this study, we report that 53BP1 KO can partially rescue embryonic lethality of bona fide BRCA1 total KO mice, but HR deficiency is not restored. BRCA1-53BP1 double knockout (DKO) mice have severe genomic instability and G2/M cell cycle checkpoint defects and develop a unique type of thymic lymphoma that is 100% penetrant. This study reveals that HR deficiency is compatible with embryonic development, but HR is important for preventing tumorigenesis in adults.
In 1-month-old Brca1 −/− ;Trp53bp1 −/− mice when lymphomas were not developed or 3-month-old Brca1 −/− ; Fig. 2 HR efficiency is not restored in BRCA1-53BP1 DKO cells. a Schematic representation of the DR-GFP reporter for analyzing efficiency of DSB repair by HR. b Summary of HR efficiency obtained using the DR-GFP reporter in MEFs with indicated genotypes. c Immunofluorescence staining analyses of RAD51 foci in MEFs with indicated genotypes 6 h after 10 Gy IR exposure. γH2AX marks sites of DNA damage. Hoechst 33342 marks nucleus. Scale bars, 10 μm. d Quantification of RAD51 foci. e Western blotting analyses of BRCA1 and 53BP1 protein in WT and 53BP1 KO U2OS DR-GFP cells after control or BRCA1 knockdown. α-tubulin was used as loading control. f Summary of HR efficiency in WT and 53BP1 KO U2OS DR-GFP cells after control or BRCA1 knockdown. g MEFs with indicated genotypes were labeled with 10 μM BrdU and 24 h later cells were treated with 10 Gy IR and recovered for 1 h. Representative BrdU foci were shown. Hoechst 33342 marks nucleus. Scale bars, 10 μm. h Quantification of BrdU foci. i Immunofluorescence staining analyses of RPA2 foci in MEFs with indicated genotypes 6 h after 10 Gy IR exposure. γH2AX marks site of DNA damage. Hoechst 33342 marks nucleus. Scale bars, 10 μm. j Quantification of RPA2 foci. k Western blotting analyses of RPA2 and RPA2 phosphorylation (S4/8) in MEFs with indicated genotypes 1 h after 10 Gy IR exposure. α-tubulin was used as loading control. Error bars represent SEM from three independent experiments. *p < 0.05; **p < 0.01; ***p < 0.001; ns not significant.
Trp53bp1 −/− mice when lymphomas were not developed in most mice, T lymphocytes in the thymus and spleen as well as B lymphocytes in the spleen were indistinguishable from those in WT mice (Figs. 4e, S1, and S2). However, in Brca1 −/− ;Trp53bp1 −/− mice that were viable with no obvious symptoms at 5 months, early stage thymic lymphomas containing CD4 and CD8 double positive lymphocytes were frequently observed (Figs. S1 and S2). Histological examinations revealed that cell organization of the thymus was completely disrupted (Fig. S3). These observations suggested that thymic lymphomas in Brca1 −/− ; Trp53bp1 −/− mice initiated by expansion of CD4 and CD8 double positive T lymphocytes. In agreement with this observation, the T lymphocytes in the spleen were predominantly double positive for CD4 and CD8 (Fig. S2). In addition, CD19 positive B lymphocytes were dramatically reduced in the spleen (Fig. S2). Cell organization of the spleen was also severely disrupted (Fig. S3). These observations together suggested that the lymphoma cells had massively infiltrated the spleen and replaced the original cells in the spleen. The infiltration of the lymphoma cells was also found in other organs such as kidney, despite that the overall structure of the kidney was not altered (Fig. S3).

BRCA1-53BP1 DKO lymphomas have unique characteristics of genomic instability
Genomic instability often leads to aneuploidy. In contrast to Trp53 KO mice [24], no aneuploidy was observed in lymphomas from Brca1 −/− ;Trp53bp1 −/− mice (Fig. 5a), suggesting that p53 signaling was normal. To provide insights into genomic features of these lymphomas, we performed whole-genome sequencing (WGS) of five individual lymphoma samples and their corresponding liver samples as germ line control. Consistent with cytological studies, no global chromosome number changes were identified in any of the lymphoma samples (Fig. 5b). Each lymphoma harbored a few fragment gain or loss, but no common altered regions across all samples were found (Fig. 5b). Unlike lymphomas from Atm KO mice, trisomy chromosome 15 was not observed (Fig. 5b).
Many single nucleotide polymorphisms (SNPs), small insertions or deletions (Indels), and structural variations (SVs) were identified in each lymphoma (Fig. 5c), reflecting genomic instability in these tumors. No genomic alternation was identified at Atm or Trp53 gene loci except one SNP with unknown pathogenic property in Trp53 gene (Fig. 5d), further suggesting that ATM and p53 signaling is intact. Unlike lymphomas from Atm KO mice, there were no alterations of Myc oncogene either (Fig. 5d). However, similar to lymphomas from Atm KO or Trp53 KO mice, in which deletions of tumor suppressor gene Pten and mutations/amplifications of Notch1 oncogene are often identified, the majority (80%) of the sequenced lymphomas from Brca1 −/− ;Trp53bp1 −/− mice contained SVs at Pten and/or Notch1 locus (Fig. 5d). This suggests that although genomic instability can originate from different signaling defects at different developmental stages of T cells, loss of functional PTEN and NOTCH1 is often selected to drive thymic lymphoma development.
SVs, including intra-chromosomal large deletions/inversions and inter-chromosomal translocations, are deleterious genomic alterations that are frequently identified in tumors. Different numbers of sVs were found in five lymphoma samples, but no common altered fragments were identified (Fig. 5e). In lymphomas from Atm KO mice, interchromosomal translocations between chromosome 12 and 14 are most frequently observed. However, they were not present in lymphomas from Brca1 −/− ;Trp53bp1 −/− mice (Fig. 5f). The above analyses together reveal that genomic instability in Brca1 −/− ;Trp53bp1 −/− mice likely leads to a unique type of thymic lymphoma with intact ATM and p53 signaling.

BRCA1-53BP1 DKO cells have increased microhomology-mediated end joining
SVs occur by ligating two broken fragments at distant genomic loci through canonical NHEJ (cNHEJ) or microhomology-mediated end joining (MMEJ) using homologous sequences at breakpoints. Examination of sequence patterns at variation junctions can infer which pathways are utilized for ligation of the two joints. Precise junctions are likely ligated through cNHEJ, while junctions with shared nucleotides between breakpoints (microhomologies) are likely ligated through MMEJ (Fig. 6a) [25,26]. It has been shown that >80% junctions of SVs in Brca2 mutant murine brain tumors have no or very short microhomologies (0-1 bp), suggesting that cNHEJ are mainly responsible for the ligation [26]. Interestingly, in lymphomas from Brca1 −/− ;Trp53bp1 −/− mice, only 20% junctions had no or very short microhomologies (0-1 bp), and most junctions had longer microhomologies (Fig. 6b). In particular, more than 50% junctions had microhomologies of 2-3 bp (Fig. 6b). This observation suggests that SVs in these lymphomas are predominantly ligated through MMEJ.
The above in vivo study prompts that Brca1 −/− ; Trp53bp1 −/− cells might have increased MMEJ activity. We tested this possibility using U2OS cells stably integrated with an MMEJ reporter. Consistent with role of 53BP1 in suppressing DNA end resection, 53BP1 KO significantly increased the efficiency of MMEJ that requires DNA end resection (Fig. 6c-e). The MMEJ efficiency was also higher in cells deficient for both BRCA1 and 53BP1 than in WT cells (Fig. 6c-e), which is likely the same in Brca1 −/− ; Trp53bp1 −/− cells. This idea was further supported by the observation that the expression levels of POLQ and LIG3, two key proteins in MMEJ pathway, were significantly increased in Brca1 −/− ;Trp53bp1 −/− MEFs (Fig. 6f, g). Besides HR and MMEJ, single-strand annealing (SSA) is another repair pathway that requires DNA end resection. Using U2OS cells stably integrated with an SSA reporter, we have shown before that although 53BP1 KO significantly increases the efficiency of SSA, further depletion of BRCA1 in 53BP1 KO cells reduces the SSA efficiency to a similar level as in WT cells [27]. Therefore, it is likely that Brca1 −/− ;Trp53bp1 −/− cells do not have altered SSA activity. In agreement with this idea, the expression level of RAD52, the key proteins in SSA pathway, was not changed in Brca1 −/− ;Trp53bp1 −/− MEFs (Fig. 6h). Therefore, MMEJ, but not SSA, is specifically activated in these cells.
ATM-dependent CHK2 activation is the major signaling pathway to activate G2/M cell cycle checkpoint after DNA damage. However, no defect in IR-induced CHK2 phosphorylation was observed in Brca1 −/− ;Trp53bp1 −/− MEFs (Fig. 7b). IR-induced phosphorylation of ATM substrate KAP1 was normal as well (Fig. 7b). These data are highly consistent with the above observations that the lymphoma and T cells in Brca1 −/− ;Trp53bp1 −/− mice had no defects similar to those in Atm KO mice (Fig. 4c-f). ATRdependent CHK1 activation also contributes to DNA damage-induced G2/M cell cycle checkpoints [28]. Interestingly, both Trp53bp1 −/− MEFs and Brca1 −/− ; Trp53bp1 −/− MEFs had defects in IR-induced CHK1 phosphorylation (Fig. 7b). This observation suggests that defective ATR-CHK1 signaling is likely the major reason behind the G2/M cell cycle checkpoint defect, which might be one of causes for lymphoma development.

53BP1 expression is decreased in cancers with silenced BRCA1 expression
BRCA1 is frequently mutated in familial breast and ovarian cancers. Previous studies have shown that 53BP1 loss is associated with triple-negative and BRCA1 mutated breast cancers [29]. Besides mutations, BRCA1 is silenced through promoter hypermethylation in some breast or ovarian cancers, causing diminished BRCA1 expression level similar to the BRCA1 status in Brca1 −/− mice. To test if 53BP1 loss can also promote the growth of tumors with diminished BRCA1 expression, we ranked breast cancer samples in TCGA database based on BRCA1 expression levels and compared 53BP1 expression levels between BRCA1-low (lowest 10%) and BRCA1-high (highest 10%) samples. Interestingly, 53BP1 expression was significantly lower in BRCA1-low than in BRCA1-high samples (Fig. S4A). Similar phenomenon was observed in ovarian cancer samples in TCGA database (Fig. S4B). We went on to test if the positive correlation between BRCA1 and 53BP1 expression levels in breast cancer samples are influenced by prognostic molecular profiling factors. We first separated breast cancer samples into two groups based on estrogen receptor (ER) expression status (positive or negative) and ranked the samples in each group based on BRCA1 expression levels. 53BP1 expression was significantly lower in BRCA1-low than in BRCA1-high samples in both groups (Fig. S4C), suggesting that ER expression did not affect the positive correlation between BRCA1 and 53BP1 expression levels. Similarly, progesterone receptor (PR) or HER2 expression did not have any effect either (Fig. S4D, E). This correlation was also significant for triple-negative breast cancer samples, which were negative for ER, PR, and HER2 (Fig. S4F). Therefore, the positive correlation between BRCA1 and 53BP1 expression levels was independent of the above prognostic molecular profiling factors. These results suggest that 53BP1 loss might promote the development of cancers with silenced BRCA1 expression.

Discussion
Domains of BRCA1 determine the ability of 53BP1 KO to rescue HR-related defects In this study, we have demonstrated that 53BP1 KO can partially rescue embryonic lethality of BRCA1 total KO mice that are homozygous for a bona fide Brca1 null allele. Unlike Brca1 Δ11/Δ11 ;Trp53bp1 −/− and Brca1 Δ2/Δ2 ; Trp53bp1 −/− mice but similar to Brca1 ΔC/ΔC ;Trp53bp1 −/− mice, our study in Brca1 −/− ;Trp53bp1 −/− mice suggests that 53BP1 KO can partially rescue embryonic lethality but not HR deficiency, genomic instability, or PARPi sensitivity caused by complete BRCA1 loss. As a result of severe genomic instability, all Brca1 −/− ;Trp53bp1 −/− mice die of thymic lymphoma within 7 months (Fig. 8). Both BRCA1Δ11 and BRCA1Δ2 proteins have BRCT domains that facilitate their localization to DNA damage sites and coiled coil domains that interact with PALB2, but these domains are absent in BRCA1ΔC proteins or the remaining peptides in Brca1 −/− cells, suggesting that these domains are important for 53BP1 KO to rescue HR-related deficiency.
Although not examined, it is likely that BARD1 levels are also diminished in Brca1 ΔC/ΔC ;Trp53bp1 −/− cells, given the dramatic reduction of BRCA1ΔC protein levels [16]. Since 53BP1 KO rescues HR-related deficiency in Brca1 Δ2/Δ2 ; Trp53bp1 −/− cells, but fails to do so in Brca1 ΔC/ΔC ; Trp53bp1 −/− or Brca1 −/− ;Trp53bp1 −/− cells, it is likely that the BRCT and/or coiled coil domains of BRCA1 are required for 53BP1 KO to rescue HR-related deficiency in the absence of BARD1. BARD1 can also localize to early DNA damage sites independent of BRCA1's BRCT domains [30]. Recent studies have shown that BARD1 can directly interact with RAD51 [31]. It will be interesting to investigate if restoration of BARD1 protein level can alleviate the requirement of BRCA1's BRCT and/or coiled coil domain for 53BP1 KO to rescue HR-related deficiency.

HR deficiency can be partially uncoupled from embryonic lethality
Since BRCA1 is critical for HR, it is generally believed that HR deficiency is the major cause of embryonic lethality of Brca1 −/− embryos. Although unable to restore HR deficiency, 53BP1 KO partially rescued embryonic lethality of Brca1 −/− embryos and no abnormality was observed in the viable Brca1 −/− ;Trp53bp1 −/− mice until they die of thymic lymphoma. Together with the previous study using Brca1 ΔC/ΔC ;Trp53bp1 −/− mice [16], our study has suggested that HR deficiency can be partly compatible with embryonic development in some BRCA1 mutant mice. In agreement with this idea, BRCA1 BRCT mutant mice (Brca1 S1598F/S1598F ) can survive, have HR deficiency, but appear normal before tumor onset [32]. Similar to BRCA1 mutants, KO mice of several Fanconi anemia complementation group proteins are viable but have HR defects [33]. Therefore, HR deficiency can be partially uncoupled with embryonic lethality in mice.
A previous study has shown that RNF168 can also load PALB2 to DNA damage sites independent of BRCA1 [34]. A recent report has also suggested that RNF168-dependent PALB2 loading is required for the viability of BRCA1 haploinsufficient mice [35]. Since 53BP1 KO has lifted the barrier for DNA end resection in Brca1 −/− cells, it is possible that RNF168 can load PALB2 to DNA damage sites in Brca1 −/− ;Trp53bp1 −/− MEFs and mildly rescue the HR defects in Brca1 −/− cells. Although we did not directly test this idea, we speculate that even if RNF168 can load PALB2 to damaged sites in these cells, this pathway has limited contribution to HR in these cells, given the severe HR deficiency in Brca1 −/− ;Trp53bp1 −/− MEFs.

MMEJ potentially contribute to the survival of BRCA1-53BP1 DKO embryos
Besides HR, MMEJ and SSA are two repair pathways that also require DNA end resection. Reporter assays suggest that the repair efficiency of MMEJ, but not SSA, is activated in Brca1 −/− ;Trp53bp1 −/− cells. These observations are consistent with previous reports that HR-deficient tumor cells have increased POLQ expression and rely on MMEJ for survival [36][37][38]. Previous studies have revealed that although either FANCD2 KO or POLQ KO mice are viable, FANCD2 and POLQ DKO mice are largely embryonic lethal, suggesting that the loss of both HR and MMEJ leads to embryonic lethality. Since Brca1 −/− ;Trp53bp1 −/− cells have severe HR deficiency, elevated MMEJ activity in these cells should reflect the significant contribution of this pathway to the survival of these embryos. As MMEJ is highly mutagenic, it is possible that Brca1 −/− ;Trp53bp1 −/− embryos survive at the cost of having increased genomic instability and eventually develop thymic lymphoma. Fig. 8 Working model: 53BP1 loss partially rescue embryonic lethality of BRCA1 total knockout mice without restoring HR. BRCA1 is critical for DSB repair by homologous recombination (HR). In WT mice, BRCA1 promotes DNA end resection by removing 53BP1 from DSB ends and facilitates HR repair (left panel). 53BP1 fails to be removed in BRCA1 total KO mice (Brca1 −/− ) and occupies DSB ends. DNA end resection is inhibited, which results in HR deficiency and embryonic lethality of BRCA1 total KO mice (middle panel). Loss of 53BP1 lifts the barrier for DNA end resection in BRCA1 total KO mice. HR is not restored but microhomologymediated end joining (MMEJ) is increased. Brca1 −/− ;Trp53bp1 −/− mice are partially viable, but they have severe genomic instability and develop thymic lymphomas (right panel). BRCA1 is also required for protecting replication forks after replication stress [39]. However, 53BP1 KO rescued embryonic lethality but not replication fork protection defects caused by different BRCA1 mutants in previous studies [15,16,39]. Therefore, it is unlikely that 53BP1 KO rescued replication fork protection defects in Brca1 −/− ; Trp53bp1 −/− cells.
Although Brca1 −/− ;Trp53bp1 −/− mice are born at reduced Mendelian ratio, Brca1 −/− ;Trp53bp1 −/− embryos at 13.5 dpc can be obtained with normal Mendelian ratio. This suggests that some mice die during late embryonic development, when hematopoietic organs are developed and hematopoietic stem cells transit from fetal liver to fetal bone marrow [40]. Recent studies have shown that BRCA1 is important for hematopoiesis [41,42]. Therefore, hematopoietic defects might account for the death of some Brca1 −/− ;Trp53bp1 −/− embryos during late embryonic development. The detailed mechanism requires further investigation.

BRCA1-53BP1 DKO mice develop a unique type of thymic lymphoma
Although Brca1 −/− ;Trp53bp1 −/− mice are viable, they have severe genomic instability and eventually develop thymic lymphoma within 7 months. These lymphomas are distinct from those from Atm or Trp53 KO mice because BRCA1, ATM, and p53 have different roles in DNA damage repair and signaling. ATM has a minor role in HR that is distinct from BRCA1. Unlike BRCA1, ATM also functions in NHEJ that is required for V(D)J development in lymphocytes. This explains why Atm KO mice, but not in Brca1 −/− ;Trp53bp1 −/− mice, have defective TCRβ rearrangement.
TP53 mutations are frequently found in BRCA1associated human breast cancers [43,44]. Therefore, the absence of Trp53 mutations or SVs in lymphomas from Brca1 −/− ;Trp53bp1 −/− mice is surprising. TP53 mutations often occur late in development to offer growth advantage of tumors that originate from other genetic alterations [45]. It is possible that the intrinsic G2/M cell-cycle checkpoint defects in Brca1 −/− ;Trp53bp1 −/− cells alleviate the requirement for Trp53 mutations and offer growth advantage to allow rapid development of thymic lymphomas. Therefore, the combination of defective HR, elevated MMEJ, and compromised cell cycle checkpoint leads to the development of a unique type of thymic lymphoma in Brca1 −/− ;Trp53bp1 −/− mice.

Mice
Brca1 flox/flox mice (NCI Mouse Repository, 01XB8, mixed background of FVB and 129) were mated with Ddx4-cre mice (The Jackson Laboratory, 006954, obtained in FVB background and backcrossed to C57 background) to obtain Brca1 Δ5−13/+ mice. They were mated with Trp53bp1 −/− mice (The Jackson Laboratory, 006495, mixed background of C57 and 129) for two generations to obtain Brca1 Δ5−13/+ ; Trp53bp1 −/− mice. These mice were subsequently intercrossed to generate mice used in this study. All mice studies were approved by the Zhejiang University Animal Care and Use Committee.
To isolate MEFs, 13.5 dpc mouse embryos were minced with surgical blades and then dissociated in 1× TrpLE (GIBCO) at 37°C for 30 min in a 15 ml pipette with occasional shaking. Liberated cells were plated and cultured. MEFs at passage 2 were transfected with SV40 Tantigen plasmids for immortalization.
To generate ES cells, 3.5 dpc blastocysts were isolated from naturally mated females by flushing the uterus with M2 medium (Sigma) and were cultured on irradiated MEF feeders in ES cell medium. Seven days later, blastocyst outgrowth was trypsinized and cells were plated on irradiated MEF feeders for expansion. Feeders were gradually removed and ES cells were cultured in gelatin-coated plates for experiments.

Generation of 53BP1 KO cells
DR-GFP U2OS and MMEJ-EGFP U2OS cell lines were gifts from Jeremy Stark (City of Hope) and Xiaohua Wu (Scripps Research Institute), respectively. 53BP1 KO was generated in these cell lines by CRISPR/Cas9 technology. 53BP1 guide RNA (5′-CATAGCAGAACAGTCCAGCA-3′) was cloned into the PX459 V2.0 (gifts from Feng Zhang, Addgene 62988) plasmids. Twenty-four hours after transfection of the PX459 V2.0-53BP1 guide RNA, cells were cultured in the presence of puromycin (2 μg/ml) for 48 h and plated to isolate individual clones. KO cells were validated by western blotting.

Western blotting
Cells were lysed with RIPA buffer (50 mM Tris-HCl, 150 mM NaCl, 5 mM EDTA, 1% Nonidet P-40, 0.1% SDS, 0.5% sodium deoxycholate) containing 1 μg/ml of phosphatase inhibitor on ice for 15 min. Lysates were centrifuged at 16900 × g for 5 min to remove cellular debris. Samples were boiled after the addition of loading buffer. Proteins were separated by SDS-PAGE, transferred to PVDF membranes. The membrane was blocked with 5% milk for 10 min and blotted with indicated antibodies overnight, washed three times with TBST buffer, and blotted with HRP-linked goat anti-rabbit or goat anti-mouse secondary antibodies (Jackson ImmunoResearch) for 1 h. After three washes with TBST buffer, blots were analyzed by enhanced chemiluminescence system.

Immunofluorescent staining
MEFs cultured on coverslips were treated with 10 Gy IR and recovered for 6 h under the same culture condition. Cells were then washed with PBS once, pre-extracted with 0.5% Triton X-100 for 1 min, and fixed with 4% paraformaldehyde for 10 min at room temperature. Cells were then incubated with primary antibody for 1 h at room temperature. After washing with PBS for twice, cells were incubated with Alexa Fluor 488/594 labeled secondary antibody (Jackson ImmunoResearch) for 30 min at room temperature. Cells were subsequently stained with Hoechst 33342 and visualized using a fluorescence microscope (Eclipse Ti2; Nikon).

Single-stranded DNA detection at DNA damage sites
MEFs were grown in culture medium containing BrdU (10 μM) for 24 h. Cells were then treated with 10 Gy IR and recovered for 1 h under the same culture condition. Cells were subjected to immunofluorescent staining using an anti-BrdU antibody.

Metaphase spreads analysis
Cells were treated with or without PARPi olaparib (1 μM) overnight and arrested with colcemid (0.2 μg/ml) for 2 h. Cells were collected, resuspended in 75 mM KCl and incubated at room temperature for 8 min. Cells were then fixed with methanol-acetic acid (3:1 (v/v)) and dropped onto glass slides. DNA were stained with 5% Giemsa for 5 min. Chromosome numbers and chromosomal aberrations were counted. Twenty metaphases were analyzed for cells of each genotype.
Plasmid HR reporter assay DR-GFP plasmids were gifts from Maria Jasin (Memorial Sloan Kettering Cancer Center). MEFs were transfected with DR-GFP plasmids or plasmids expressing an intact GFP protein. Forty-eight hours after transfection, cells were infected with adenovirus expressing I-SceI. After 24 h, GFP positive cells were analyzed by flow cytometry using a Beckman CytoFLEX LX Analyzer. HR efficiency were normalized by transfection efficiency marked by GFP only cells.
Integrated HR and MMEJ reporter assay DR-GFP U2OS and MMEJ-EGFP U2OS cell lines were transfected with indicated siRNAs and infected with adenovirus expressing I-SceI. Cells were recovered for 48 h after infection and were analyzed by flow cytometry.

Cell viability assay
A total of 5 × 10 4 ES cells or 3 × 10 3 U2OS cells were plated in each well of 6-well plates. After 24 h, cells were treated with the indicated concentration of PARPi olaparib and medium was changed every day. After incubation for 7 days, the surviving cells were assessed with CCK8 reagent (DOJINDO) following the manufacturer's instructions.

Lymphocyte analysis
Single-cell suspensions were prepared from thymus, lymphoma or spleen from mice of the indicated genotypes. A total of 1 × 10 5 cells were stained using fluorescenceconjugated antibodies as indicated. Lymphocyte populations were analyzed by flow cytometry.
G2/M cell-cycle checkpoint analysis MEF cells were exposed to 5 Gy ionizing radiation or without treatment. After 1 h of recovery, cells were fixed with cold 70% ethanol and stained with anti-phospho-Histone H3 S10 antibody. Cells were incubated with FITCconjugated goat-anti-rabbit secondary antibody. Cells were then treated with RNase A and then incubated with propidium iodide. Samples were analyzed by flow cytometry.

Hematoxylin and eosin (H&E) staining
Tissues were fixed in 4% paraformaldehyde for 24 h, dehydrated, and embedded in paraffin. Paraffin sections of 5 μm were stained with H&E.

WGS and bioinformatics analysis
40× paired-end 150 bp WGS was performed for lymphoma and corresponding liver samples using Illumina NovaSeq. Raw sequencing data were aligned to GRCm38 reference genome by Burrows-Wheeler Aligner mem algorithm (BWA-mem v 0.7.12-r1039) with default parameters [49]. After alignment, reads were sorted by samtools (version 1.1) and duplicated reads were removed by Picard (v 2.0.1) [50] ("Picard Toolkit." 2019. Broad Institute, GitHub Repository.). To identify SVs with breakpoint sequence pattern, a local-assembly tool SvABA (v 1.1.0) was applied with default parameters [51]. Somatic SVs identified were then filtered using SvABA default filters with an addition condition: no secondary alignments are associated with this contig fragment (SUBN = 0). SV joint sequences were obtained from the filtered SvAVA output. To analyze the microhomology status near SV breakpoints, the length of each microhomology was extracted and the distribution of different length groups (0-1, 2-3, 4-5, 6-7 nt, and more than 7 nt) was calculated. SNPs and small indels were called independently by GATK toolkit (v 4.0.11.0) and SvABA [52]. Variants that were identified in both GATK and SvABA, as well as with more than five supporting reads, were kept as true SNPs/indels. Circos (v 0.69.6) was applied to generate the Circos plots [53]. GATK toolkit was also utilized to calculate copy ratios in tumor samples with their germ line controls. To plot the genome wide copy ratio coverage, bin size was set at 1000 bp. To examine the relationship between BRCA1 and 53BP1 expression, gene expression of breast cancer and ovarian cancer tumors were downloaded from TCGA Data Portal and then log2transformed. Their corresponding prognostic molecular profiling factors were also obtained from TCGA Data Portal. Unless specified, two-tailed unpaired Student's t test was used to evaluate statistical significance for all experiments. The WGS data has been deposited in NCBI SRA database (PRJNA587820).
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons. org/licenses/by/4.0/.