BRCA2 promotes genomic integrity and therapy resistance primarily through its role in homology-directed repair

Abstract

Tumor suppressor BRCA2 functions in homology-directed repair (HDR), the protection of stalled replication forks, and the suppression of replicative gaps, but their relative contributions to genome integrity and chemotherapy response are under scrutiny. Here, we report that mouse and human cells require a RAD51 filament stabilization motif in BRCA2 for fork protection and gap suppression but not HDR. In mice, the loss of fork protection/gap suppression does not compromise genome stability or shorten tumor latency. By contrast, HDR deficiency increases spontaneous and replication stress-induced chromosome aberrations and tumor predisposition. Unlike with HDR, fork protection/gap suppression defects are also observed in Brca2 heterozygous cells, likely due to reduced RAD51 stabilization at stalled forks/gaps. Gaps arise from PRIMPOL activity, which is associated with 5-hydroxymethyl-2'-deoxyuridine sensitivity due to the formation of SMUG1-generated abasic sites and is exacerbated by poly(ADP-ribose) polymerase (PARP) inhibition. However, HDR proficiency has the major role in mitigating sensitivity to chemotherapeutics, including PARP inhibitors.

Keywords: BRCA2; PARP inhibitors; PRIMPOL; RAD51; gap suppression; hmdU; homologous recombination; homology-directed repair; stalled fork protection.

Figure 1.. BRCA2 C-terminal RAD51-interacting site is dispensable for HDR but crucial for stalled fork protection in mouse cells.

A. Mouse and human BRCA2 TR2 peptides from the C terminus interact with RAD51 but the interaction is abrogated by mutation of a conserved serine (SA: mouse, S3214A; human, S3291A). Co-immunoprecipitation of RAD51 is observed with GST-tagged wild-type, but not SA mutant, peptides in HEK293T cells. (See also Figure S1B.) B. Domain structures of murine BRCA2 mutants. BRCA2 interacts with RAD51 at the BRC repeats and a distinct site in the C terminus that is mutated in BRCA2 S3214A (abbreviated “SA”). BRCA2 Δ27 is a truncation that deletes the C-terminal RAD51 interacting site and a DNA binding motif KKRR. C,D. HDR is not significantly affected in Brca2^SA/SA mouse cells, as assayed with an integrated DR-GFP reporter in primary ear fibroblasts (C) and Hsp90-ZsGreen gene targeting in primary MEFs (D). The diagram in (C) depicts BRCA2 nucleation of RAD51 nucleoprotein filaments to promote the strand invasion step of HDR. Statistical analysis, unpaired Student’s t test (mean ± 1 SD). ns, not significant; *, p<0.05; **, p<0.01. See Figures S1H-J for controls. E. Brca2^Δ27/Δ27 mice, but not Brca2^SA/SA mice, are sensitive to the crosslinking agent mitomycin C (MMC). The Kaplan-Meier survival curve of Brca2 mice after intraperitoneal injection of 3.5 mg/kg MMC is shown. F. Stalled fork degradation in Brca2 homozygous and heterozygous cells. Diagram shows BRCA2 and RAD51 protecting a reversed fork from MRE11-mediated nucleolytic attack. Statistical analysis, Mann-Whitney U test (median ± 95% confidence intervals). ns, not significant; ***, p<0.001; ****, p<0.0001. See Figure S2A,B for replicates and summary.

Figure 2.. BRCA2 C-terminal RAD51 interaction site suppresses PRIMPOL-mediated ssDNA gaps.

A. Impaired gap suppression in Brca2 homozygous and heterozygous Δ27 and SA primary MEFs treated with HU (A). Diagram shows PRIMPOL activity creating S1 nuclease-sensitive ssDNA gaps when BRCA2 and RAD51 are absent. B-D. BRCA2 S3291A/E-expressing cells show impaired gap suppression with HU, including hamster V-C8 (B), human mammary MCF10A (C) and human fibroblast VU423 (D) . Below (D) is the Western blot for human BRCA2 wild-type and S3291E expressed in VU423 cells from the PiggyBac-neo (PBn) vector. EV, empty vector. E. Impaired gap suppression in Brca2 homozygous and heterozygous primary MEFs treated with olaparib. F-H. BRCA2 S3291A/E-expressing cells show impaired gap suppression with olaparib, including V-C8 (F), MCF10A (G) or VU423 (H) cells. I,J. PRIMPOL knockdown suppresses gap formation in Brca2 mutant immortalized MEFs (iMEFs) treated with HU (I) or olaparib (J). K. Stalled fork degradation in Brca2^+/null primary MEFs. All statistical analyses by Mann-Whitney U test (median ± 95% confidence intervals). ns, not significant; *, p<0.05; **, p<0.01; ***, p<0.001; ****, p<0.0001. See Figures S2D,G,J and S3B,D,F,G for replicates.

Figure 3.. BRCA2 C-terminal RAD51 interaction site is dispensable for maintaining genome integrity.

A. Spontaneous and HU-induced chromosome aberrations are substantially increased in HDR-deficient Brca2^Δ27/Δ27 primary MEFs but only mildly in Brca2 mutants proficient in HDR. Three repeats were performed for each genotype. Statistical analysis, Mann-Whitney U test (mean ± 1 SD). ns, not significant; *, p<0.05, **, p<0.01, ***, p<0.001, ****, p<0.0001. B. Breakdown of chromosomal aberrations in (A). Most are breaks/gaps, although acentrics, exchanges, and radials are also observed, especially in Brca2^Δ27/Δ27 MEFs, for which representative aberrations are shown. C. Elevated micronuclei are observed in Brca2^Δ27/Δ27 and especially Brca2^Δ27/null blood cells but not in Brca2^SA/SA or Brca2^SA/null erythrocytes. Normochromic erythrocytes (NCEs) are negative for both CD71 and propidium iodide (PI) but become PI positive when a micronucleus is present (MN-NCE). (See also Figure S4B-D.) Statistical analysis, unpaired Student’s t test (mean ± 1 SD). ns, not significant; ****, p<0.0001. D. Testis weights are reduced in Brca2^Δ27/Δ27 and especially Brca2^Δ27/null mice but not in Brca2^SA/SA or Brca2^SA/null mice. (See body and testis weights in Figure S4A,E.) Statistical analysis, unpaired Student’s t test (mean ± 1 SD). ns, not significance; *, p<0.05; ****, p<0.0001. Inset images are representative cross sections of stage IV H&E-stained tubules. Additional testis images are in Figure S4F. Scale bar, 100 μm. E. Brca2^Δ27/Δ27 mice, but not other Brca2 genotypes, are tumor prone. Kaplan-Meier tumor-free survival curves were analyzed using the log-rank (Mantel-Cox) test. Tumor types are in Figure S4G.

Figure 4.. Genotoxin sensitivity is greatest in the absence of HDR, but is observed for hmdU in fork protection/gap suppression defective cells.

A-F. Genotoxin sensitivity assays of immortalized MEFs. Cells were treated with various concentrations of HU (A), cisplatin (B), olaparib (C), hmdU (D), olaparib with a fixed concentration of hmdU (E), or hmdU with a fixed concentration of olaparib (F) for 6 days, and their survival was plotted. In all conditions, HDR-deficient Brca2^Δ27/Δ27 MEFs display the greatest sensitivity, while MEFs with only fork protection/gap suppression defects show milder or no sensitivity, except in the case of hmdU alone, in which the sensitivities of all Brca2 mutants are more similar at higher doses (D) and are exacerbated by olaparib (F). Statistical analysis, two-way ANOVA test (mean ± SEM). ns, not significant; *, p<0.05; **, p<0.01; ***, p<0.001; ****, p<0.0001. G. Western blots for human BRCA2 wild-type and S3291E expressed in Brca2^Δ27/Δ27 immortalized MEFs. H. Human BRCA2 S3291E fully rescues olaparib sensitivity of Brca2^Δ27/Δ27 immortalized MEFs but only partially rescues sensitivity to hmdU and combined treatment. Statistical analysis, unpaired Student’s t test (mean ± 1 SD). ns, not significant; **, p<0.01; ****, p<0.0001.

Figure 5.. hmdU sensitivity arises from PRIMPOL activity in HDR-proficient cells, but not in HDR-deficient cells.

A. Stalled fork protection is restored in Brca2 mutants with SMARCAL1, but not PRIMPOL, knockdown. Statistical analysis, Mann-Whitney U test (median ± 95% confidence intervals). ns, not significant; *, p<0.05; **, p<0.01; ****, p<0.0001. B. Gap suppression is restored in Brca2 mutants with PRIMPOL, but not SMARCAL1, knockdown. Statistical analysis, Mann-Whitney U test (median ± 95% confidence intervals). ns, not significant; ****, p<0.0001. C. hmdU sensitivity is associated with PRIMPOL activity in HDR-proficient cells, but not in HDR-deficient cells, while olaparib sensitivity is not affected by either PRIMPOL or SMARCAL1. hmdU sensitivity is rescued by SMUG1 knockdown in all mutants. Statistical analysis, unpaired Student’s t test (mean ± 1 SD). ns, not significant; *, p<0.05; ***, p<0.001; ****, p<0.0001.

Figure 6.. BRCA2 function during replication stress.

HDR is required when replication forks collapse, leading to DSBs and ssDNA. However, when forks stall, BRCA2 protects reversed forks, which can restart when replication stress is relieved, and prevents or protects PRIMPOL-mediated ssDNA gaps. Loss of the BRCA2 C-terminal RAD51-interacting site, which stabilizes RAD51 filaments, leads to fork degradation and ssDNA gap formation. Gap suppression and HDR are important for hmdU resistance, while HDR plays the major role in olaparib resistance and tumor suppression.