CDK5RAP3, a New BRCA2 Partner That Regulates DNA Repair, Is Associated with Breast Cancer Survival

Simple Summary BRCA2 is an essential gene for DNA repair by homologous recombination and is often mutated in families at risk of breast and ovarian cancer. In this study we identified CDK5RAP3 tumor suppressor as a new BRCA2-interacting protein. CDK5RAP3 negatively regulates DNA repair of double-strand breaks, providing a new mechanism of DNA damage resistance. Consistently, gene expression data analysis showed CDK5RAP3 overexpression in breast cancer is associated with poorer survival. Finally, we identified common genetic variations in the CDK5RAP3 locus as potentially associated with breast and ovarian cancer risk in a large cohort of BRCA1 and BRCA2 mutation carriers. Abstract BRCA2 is essential for homologous recombination DNA repair. BRCA2 mutations lead to genome instability and increased risk of breast and ovarian cancer. Similarly, mutations in BRCA2-interacting proteins are also known to modulate sensitivity to DNA damage agents and are established cancer risk factors. Here we identify the tumor suppressor CDK5RAP3 as a novel BRCA2 helical domain-interacting protein. CDK5RAP3 depletion induced DNA damage resistance, homologous recombination and single-strand annealing upregulation, and reduced spontaneous and DNA damage-induced genomic instability, suggesting that CDK5RAP3 negatively regulates double-strand break repair in the S-phase. Consistent with this cellular phenotype, analysis of transcriptomic data revealed an association between low CDK5RAP3 tumor expression and poor survival of breast cancer patients. Finally, we identified common genetic variations in the CDK5RAP3 locus as potentially associated with breast and ovarian cancer risk in BRCA1 and BRCA2 mutation carriers. Our results uncover CDK5RAP3 as a critical player in DNA repair and breast cancer outcomes.

CDK5 regulatory subunit-associated protein 3 (CDK5RAP3, also known as C53, LZAP or IC53) is a tumor suppressor that activates p53, induces genotoxic-dependent apoptosis, inhibits the G2/M checkpoint via CHK1 and CDK1, and reduces NF-κB activity, cellular invasion, and tumor growth [16][17][18][19]. Low levels of CDK5RAP3 are associated with a subset of head and neck squamous cell carcinomas (HNSCCs) and hepatocellular carcinomas (HCCs) as well as with poor survival [18,20]. In this study, a screening for physical interactors of BRCA2 identified CDK5RAP3, which emerged as a critical regulator of DNA DSB repair that is associated with breast cancer outcomes.

Western Blot
U2OS, HeLa or HEK cells were seeded on 6 well plates, 24 h later were treated with hydroxyurea, MMC or left untreated, and 24 h later cells were lyzed and SDS-PAGE and blotting with indicated antibodies performed as previously described [21]. The uncropped western blots are shown in Figure S5.

Analysis of BRCA2 Interactions
Yeast two hybrid screens were carried out as previously described [22]. To confirm CDK5RAP3 interaction with BRCA2, HEK 293T cells expressing an empty vector or Flag-BRCA2 were treated with double thymidine block (10 mM) followed by a 9 h release to enrich S-phase cells, or left untreated. Cells were lysed and subjected to immunoprecipitation (IP) with anti-Flag M2 antibody. Immunoprecipitates and whole cell lysates were immunoblotted with CDK5RAP3, BRCA2 or actin antibodies. In other experiments, HEK 293T cells were left untreated or synchronized to S-phase by treatment with 2 mM hydroxyurea for 18 h. Cells were lysed and subjected to IP with anti-IgG or anti-BRCA2 antibody, followed by immunoblotting with CDK5RAP3, BRCA2, geminin or actin antibodies.

Laser Microirradiation Experiments
Laser microirradiation experiments were reported earlier [23]. Briefly, pmCherry CDK5RAP3-transfected U2OS cells were sorted twice (FacsJazz, Becton Dickinson, Madrid, Spain) for stable transfection. Cells were plated, stimulated with 10 µM BrdU for 24 h, and before microscope analysis pre-treated 10 min at 37 • C with 5 µg/mL Hoescht 33342 (Life Technologies, Barcelona, Spain). Images were taken with an Olympus Fluoview 1000 confocal microscope (Olympus, Barcelona, Spain). For laser microirradiation studies, cells were microirradiated with 5 s pulses of spots or lines with the 405 nm laser at full power and images were taken every 5 s for up to 5 min after localized DNA damage. For time course experiments, Cherry-CDK5RAP3 fluorescence was quantified from at least 30 cells and averaged.

Survival and Clonogenic Assays
Survival experiments were reported before [24]. Briefly, U2OS or HeLa cells were transfected twice with luciferase, BRCA2 or CDK5RAP3 siRNA at a final concentration of 20-40 nM in 6-multiwell plates. 24 h later, 1 × 10 5 cells were seeded in duplicate in 6-multiwell plates (or 35 mm dish plates for irradiation) and left untreated for 24 h more. Cultures were then exposed to mitomycin C (MMC), olaparib or γ-irradiated at the indicated doses. Three to four days after the treatment, cells were rinsed with PBS, harvested by trypsinization and counted. Survival is reported as the percentage relative to untreated controls. For clonogenic assays, cells were plated and treated with MMC or γ-irradiated at the indicated doses, left to grow for 10-14 days, and Giemsa stained; then, colonies were finally counted.

In Vitro Flow Cytometric Micronuclei (FCM) Assay
FCM assay was reported previously [21]. Briefly, 250,000 cells/well were plated in 6-well plates. Cells were then treated with ICL inducer DEB (0.05 or 0.1 µg/mL) for 3 days. After treatment, trypsinized cells were centrifuged at 800 rpm for 8 min. Supernatant was removed and cells were resuspended by vortexing. 25 µL of EMA solution (0.125 mg/mL in PBS with 2% of FBS) was added to cell suspension. For the photo-activation step, plates were placed on ice under the visible light from a light bulb located 30 cm above cells for 20 min.
After the photo-activation step (EMA covalently binds the DNA of cells with disrupted membrane integrity-dead/dying cells), plates were protected from light exposure for the remaining steps of the staining procedure. 100 µL of cold PBS with 2% FBS were added, and cells were centrifuged at 800 rpm for 8 min. Supernatant was discarded and cells were resuspended by vortexing. Cell suspension was kept in room temperature for 20 min before cell lysis. Cells were then lysed in a two-step procedure. 100 µL of lysis solution 1 (0.584 mg/mL NaCl, 1 mg/mL sodium citrate, 0.3 µL/mL IGEPAL, 1 mg/mL RNase A, 0.2 µM Sytox Green) were added. Samples were incubated at room temperature for 1 h. After the incubation, 100 µL of lysis solution 2 (85.6 mg/mL sacarose, 15 mg/mL citric acid and 0.2 µM Sytox Green) were added into the samples and immediately vortexed. Samples were incubated at room temperature for 30 min and were stored into a cold chamber overnight until samples were measured. Flow cytometry analysis was performed with a FACSCalibur (Becton Dickinson). At least 20,000 EMA negative-Sytox Green positive events (divided nuclei) were gated per sample. To exclude events that are not MN or nucleus from live cells, we used different gates for analysis that included forward scatter, side scatter, Sytox Green and EMA fluorescence (see Supplementary Figure S2A for details). Micronuclei (MN) frequency was expressed as the number of MN per thousand nuclei obtained in the G plot. Percentage of cells arrested in the G2/M phase was obtained in the H plot.

DNA Repair Assays
HR, SSA and NHEJ assays were performed essentially as described [25][26][27]. For the HR assay, the U2OS DR-GFP stable cell line was used. For SSA and NHEJ assays, U2OS and HeLa cell lines were transiently transfected with SA-GFP or EJ5-GFP plasmids, selected with puromycin, and clonal cell lines obtained from limited dilution. Stable cell lines were confirmed to have specific GFP fluorescence upon I-SCEI but not upon empty vector transfection. For DNA repair assays, I-SCEI-endonuclease-expressing vector, empty vector or GFP-expressing vector were transfected in U2OS or HeLa cell lines for DNA repair assays. Cells were transfected with NZE-GFP, pCBAS or pCAGGS for 16-24 h with the indicated plasmids with Lipofectamine 2000 (Invitrogen) and, 24 h, later collected and analyzed by cell cytometry.

Analysis of Publicly Available Cancer Transcriptomic Datasets
To analyze the effect of CDK5RAP3 expression on the prognostics of ovarian and breast cancer patients, we generated Kaplan-Meier survival curves of these cancers with low or high expression of CDK5RAP3 by using Kaplan-Meier Plotter (www.kmplot.com accessed on 1 October 2017) [28]. For TCGA analysis, pre-processed and normalized gene expression data were downloaded from the corresponding TCGA repository (http: //tcga-data.nci.nih.gov/tcga/tcgaHome2.jsp (accessed on 3 July 2021)). Expression profiles were clustered using the PAM50 signature. The signatures were compiled from the corresponding publications and corresponded to up-regulated resistance-associated genes. The CDK5RAP3 signature correlations were computed using the PCC and the average Z-score value per gene set. The profiles were clustered using the R software package cluster.

CDK5RAP3 Interacts with BRCA2 and Localizes to DNA Damage
We screened for BRCA2 interactors by applying the yeast two-hybrid system and five baits based on Pfam-predicted domains and PONDR-predicted disordered regions [22], which overall covered most of the functionally important BRCA2 domains/regions ( Figure 1A). The helical domain baits identified two potential interactors: CDK5RAP3 and LRRC45. The tower domain baits identified two additional potential interactors: SIPA1L1 and TACC3. Of these, we focused on CDK5RAP3, owing to its previously reported role as a tumor suppressor in hepatocarcinoma [18]. The CDK5RAP3-BRCA2 interaction was confirmed by coimmunoprecipitation (co-IP) in HEK293T cells. Cells were transfected with Flag-tagged BRCA2 and lysates were immunoprecipitated with anti-Flag antibody. Untransfected cells were immunoprecipitated with anti-BRCA2 antibody. As seen in Figure 1B,C, CDK5RAP3 interacted with BRCA2 in cells overexpressing Flag-BRCA2, as well as with endogenous BRCA2, only when cells were synchronized in the S-phase with double thymidine block or hydroxyurea treatment. These findings identified CDK5RAP3 as a novel physical interactor with BRCA2 and suggested that the association occurs during the S-phase of the cell cycle.  Once the CDK5RAP3-BRCA2 interaction was confirmed, we checked whether CDK5RAP3 relocalized to the site of DSBs by laser microirradiation experiments and confocal microscopy. As seen in Figure 1D,E, CDK5RAP3 relocalized to DNA damage shortly after microirradiation, and nuclear distribution was also enhanced after DNA damage. Depletion of BRCA2 resulted in reduced relocation of CDK5RAP3 to DNA damage, suggesting that the role of CDK5RAP3 in DNA damage repair is functionally connected to BRCA2 ( Figure 1F,G).

CDK5RAP3-Depleted Cells Are Resistant DNA Damage
It was previously reported that CDK5RAP3 affects sensitivity to DNA damage, as its inhibition reduced cell death and apoptosis in response to 20 µM etoposide and 2.5 Gy radiation treatment in HeLa cells [16]. Accordingly, we examined the effect of CDK5RAP3 depletion in the context of the HR pathway. In particular, we compared the DNA damage sensitivity of CDK5RAP3-depleted U2OS cells with that of BRCA2-depleted cells. Small interfering RNAs (siRNAs) were used to deplete CDK5RAP3 or BRCA2, and the expression reduction was confirmed by Western blot (Figure 2A, right panel and Figure S5). We then performed survival and clonogenic assays after treating the cells with increasing doses of an ICL inducer (MMC), a DSB inducer (ionizing radiation, IR), or a PARP inhibitor (olaparib). As expected, inhibition of BRCA2 rendered the cells highly sensitive to DNA-damaging agents ( Figure 2D). Similar results were seen in HeLa cells (data not shown). Finally, to discard the potential interference of cell growth in mutagen sensitivity, we analyzed cell cycle progression and proliferation, which were similar in CDK5RAP3-depleted cells in comparison with control cells (data not shown).

CDK5RAP3-Depleted Cells Are Resistant DNA Damage
It was previously reported that CDK5RAP3 affects sensitivity to DNA damage, as its inhibition reduced cell death and apoptosis in response to 20 µM etoposide and 2.5 Gy radiation treatment in HeLa cells [16]. Accordingly, we examined the effect of CDK5RAP3 depletion in the context of the HR pathway. In particular, we compared the DNA damage sensitivity of CDK5RAP3-depleted U2OS cells with that of BRCA2-depleted cells. Small interfering RNAs (siRNAs) were used to deplete CDK5RAP3 or BRCA2, and the expression reduction was confirmed by Western blot (Figure 2A, right panel and Figure S5). We then performed survival and clonogenic assays after treating the cells with increasing doses of an ICL inducer (MMC), a DSB inducer (ionizing radiation, IR), or a PARP inhibitor (olaparib). As expected, inhibition of BRCA2 rendered the cells highly sensitive to DNA-damaging agents (Figure 2A-D). However, CDK5RAP3-depleted cells became resistant to MMC (Figure 2A,B), IR (Figure 2A,C), and to a lesser extent, olaparib ( Figure  2D). Similar results were seen in HeLa cells (data not shown). Finally, to discard the potential interference of cell growth in mutagen sensitivity, we analyzed cell cycle progression and proliferation, which were similar in CDK5RAP3-depleted cells in comparison with control cells (data not shown).

CDK5RAP3 Regulates DSB Repair and Genomic Instability
Because BRCA2-deficient cells are HR defective [1,2], we examined whether CDK5RAP3 similarly regulates DSB repair. An in vitro HR assay was performed in U2OS cells by detecting the HR-dependent correction of an incomplete GFP expression cassette inserted in the genome ( Figure 3A left panel) in the presence or absence of CDK5RAP3 [25]. Control cells with intact HR exhibited efficient functional GFP expression ( Figure 3B, upper left panel). As expected, BRCA2 inhibition strongly abolished HR repair ( Figure 3B upper right and Figure 3C). Interestingly, CDK5RAP3 downregulation resulted in a marked increase in the percentage of fluorescent cells ( Figure 3B, upper middle and Figure 3C).    We previously noted in Western blots that BRCA2 expression was consistently increased when CDK5RAP3 was inhibited (Figure 2A, Figures S1A and S5). We wanted to know if other proteins involved in HR repair could have a similar expression profile upon CDK5RAP3 inhibition. As seen in Supplementary Figure S2B,C (and data not shown, uncropped images in Supplementary Figure S5), BRCA2, RAD51 and RPA2 expression increased in CDK5RAP3depleted U2OS and HeLa cells, consistent with a global upregulation of HR repair.
It is well known that the lack of HR repair in BRCA2-deficient cells is overcome by increasing error-prone repair through homology-directed single-strand annealing (SSA) [26]. To determine whether CDK5RAP3 plays a role in SSA, we performed another in vitro cellular assay based on an SSA-dependent correction of an incomplete GFP expression cassette ( Figure 3A left panel). Similar to the results of the HR assay, U2OS ( Figure 3) and HeLa cells (not shown) with active SSA repair exhibited functional, although less active GFP expression ( Figure 3B lower left panel). As anticipated, BRCA2 depletion resulted in a 5-to 30-fold increase in the number of fluorescent cells ( Figure 3B lower right panel  and 3D). Interestingly, CDK5RAP3 depletion also resulted in a 2-to 10-fold increase in the number of fluorescent cells ( Figure 3B lower middle panel and 3D). To further assess the role of CDK5RAP3 in DSB repair, a third GFP reporter-based cellular assay for nonhomologous end joining was performed in U2OS and HeLa cells [27]. This assay showed that depletion of CDK5RAP3 (or BRCA2) had little or no effect on non-homologous end joining (Supplementary Figure S1B-D and data not shown).
Genomic instability is a hallmark of BRCA2-deficiency upon treatment with chromosomebreaking agents, and frequently micronuclei (MN) production in cells is used to study chromosome fragility [29]. We used the in vitro flow cytometric MN (FCM) assay (see materials and methods) in the presence or absence of CDK5RAP3 [30]. Consistent with DNA damage resistance ( Figure 2) and HR and SSA DSB repair induction upon CDK5RAP3 depletion ( Figure 3A-D), U2OS and HeLa CDK5RAP3-depleted cells showed a marked reduction in spontaneous and DNA damaging-induced chromosome fragility ( Figure 3E,F and data not shown).

CDK5RAP3 Expression Is Associated with Poor Breast and Ovarian Cancer Survival
Given that CDK5RAP3 interacts with BRCA2, whose mutations considerably increase the risk of breast and ovarian cancer, and that CDK5RAP3 downregulation leads to DNA damage resistance, we hypothesized that CDK5RAP3 expression may be associated with poorer survival of breast and ovarian cancer patients. To test this hypothesis, we used the Kaplan-Meyer Plotter cancer transcriptomic database [28]. Data from 3,554 breast cancer patients revealed a significant association between low CDK5RAP3 expression and poor overall survival (HR = 1.61; p = 3.6 × 10 −4 ) and relapse-free survival (HR = 1.64; p < 10 −16 , Figure 4A,B). A similar but milder association was observed in patients with high-grade serous ovarian cancer (data not shown). Additional analysis also showed an association between poor survival rates and low CDK5RAP3 expression in basal, luminal A and luminal B breast cancer patients (Supplementary Figure S3).
The above data show that lack of CDK5RAP3 leads to increased resistance to DNAdamaging agents in vitro and, in parallel, that low CDK5RAP3 expression is associated with poor breast and ovarian cancer prognosis. To further assess the prediction that low CDK5RAP3 levels promote tumor chemoresistance, we computed the expression correlation between CDK5RAP3 and previously defined signatures of chemoresistance and/or DNA damage resistance using breast cancer data from The Cancer Genome Atlas [31]. The resistance signatures corresponded to sets of genes whose products were involved in epithelial-tomesenchymal transition [32,33], induced by interferon signaling [34], by radiation [35], or associated with poor prognosis triple-negative breast cancer [36]. Remarkably, all analyzed signatures showed a significant negative correlation with CDK5RAP3 ( Figure 4B), specially for triple negative breast cancer; that is, low expression of CDK5RAP3 was associated with relatively higher expression of the defined signatures, which, in turn, further endorses a functional link between CDK5RAP3 downregulation and chemoresistance.

SNPs in CDK5RAP3 May Be Associated with Cancer Risk in BRCA1/2 Mutation Carriers
Common genetic variation (represented by single nucleotide polymorphisms (SNPs)) at loci encoding for functional interactors of BRCA1 and/or BRCA2 modify breast and/or ovarian cancer risk in carriers of germline BRCA1 and/or BRCA2 mutations [37,38]. Since the CDK5RAP3 gene product physically interacts with BRCA2 and functions in the DNA damage response, we conducted an association study using data from the Consortium of Investigators of Modifiers of BRCA1/2, CIMBA [39] (Figure 4C). Combined results of the iCOGS and Oncoarray studies identified potential associations with cancer risk in these settings (minor allele frequency (MAF) > 0.01, p values < 10 −3 , see supplementary results and methods): BRCA1 mutation carriers, breast cancer risk, imputed rs536194633 HR = 0.90; and ovarian cancer risk, genotyped rs2905855 HR = 1.07, imputed rs201764131 HR = 0.91, and imputed rs6503964 HR = 0.93; BRCA2 mutation carriers, breast cancer risk, imputed rs77563750 HR = 1.11 and rs17617360 HR = 0.90. Additional variants showed potential associations with ovarian cancer risk in BRCA2 mutation carriers, but their MAFs were < 0.01 (imputation r 2 values > 0.88; the complete locus results are shown in Table S1). Bioinformatic analysis of publicly available data suggested an enhancer role on genes at the corresponding genomic region by some of the depicted variants (Supplementary Figure  S4, Table S2 and Supplementary Methods). Data were divided at the best cutoff value into high and low expressing groups (for RFS, n = 1745 for CDK5RAP3 low, n = 1809 for CDK5RAP3 high; p < 10-16 with log-rank analysis; for OS, n = 678 for CDK5RAP3 low, n = 439 for CDK5RAP3 high; p = 0.00036 with log-rank analysis). Data obtained from the Kaplan-Meier Plotter (http://www.kmplot.com). (B) Unsupervised clustering of TCGA breast cancer gene expression data of CDK5RAP3 and defined signatures of chemoresistance and/or DNA damage resistance (two sets of epithelial-to-mesenchymal (EMT)-related genes are shown, as indicated in the text). The Pearson correlation coefficients (PCCs), corresponding p values, and tumor subtypes according to PAM50 classification are also shown. The dashed box highlights the cases with relative lower expression of CDK5RAP3 and higher for the signatures. (C) SNPs linked to CDK5RAP3 may influence breast/ovarian cancer risk in BRCA1/2 mutation carriers. Association plot for variants within the genomic region 100 kb on

Discussion
Tumor formation and progression result from an imbalance in the expression of oncogenes and tumor suppressors. Oncogenes promote cell growth, and their overexpression or constitutive activation is highly deleterious. In contrast, tumor suppressors are essential for the regulation of uncontrolled cell division and maintenance of homeostasis, and normal cell functions may be compromised when they are downregulated or when their activity is impaired. DNA damage response pathways act as tumor suppressors by maintaining genomic stability and thus preventing oncogenic mutations. Similarly, cell cycle checkpoint pathways also act as tumor suppressors by controlling apoptosis and cell cycle progression, thereby allowing time for DNA repair.
CDK5RAP3 was first suggested to be a tumor suppressor when Wang and co-workers demonstrated functional interactions with p14 ARF and p53 [17]. Shortly thereafter, the same authors reported that CDK5RAP3 inhibits NF-κB signaling, providing additional evidence for its role in preventing tumor formation. Cells lacking CDK5RAP3 are more prone to cell invasion and exhibit enhanced xenograft tumor growth [18]. Interestingly, one-third of HNSCCs express remarkably low levels of CDK5RAP3, and low CDK5RAP3 levels were associated with reduced survival in HCC patients [20].
In this study, we link for the first time CDK5RAP3 function with DNA repair by demonstrating that it physically interacts with BRCA2 and its absence leads to increased HR and SSA repair pathways. This is in contrast with BRCA2, whose depletion or mutation leads to HR abrogation compensated with induced SSA [26,40,41]. In addition, the observed DNA damage resistance and reduced genomic instability in CDK5RAP3-depleted cells may contribute to increased tumor aggressiveness and progression in breast cancer patients (as seen in Km plot analysis).
By yeast two hybrid experiments, we found that the BRCA2 helical domain interacted with CDK5RAP3 ( Figure 1A), and this binding was further validated in co-IP experiments ( Figure 1B,C). Some pathogenic missense variants have been identified in BRCA2 that map to the helical domain. They have been classified as pathogenic or not based on its capacity to disrupt HR pathway, but there are no reports on their role in SSA activity [4,42]. Thus, it would be interesting to address their capacity to maintain CDK5RAP3 interaction, in addition to SSA activity. HR induction by lack of CDK5RAP3 may be explained by BRCA2 expression induction (Figure 2A right and Supplementary Figure S2B,C), which could be produced through gene expression regulation at the transcriptional level or through protein stability mechanisms. However, how SSA activity increases in the absence of CDK5RAP3 remains elusive. Further experiments are needed to unravel the mechanisms by which CDK5RAP3 and BRCA2 regulate the HR and SSA pathways.
Our data connecting CDK5RAP3 to the BRCA2 pathway suggest a potential therapeutic approach for breast cancer. Blocking SSA activity is synthetically lethal in BRCA2deficient tumors [43][44][45]. Similarly, because CDK5RAP3 deficiency could make tumors dependent on HR or SSA repair pathways, it is conceivable that inhibiting HR, SSA or both combined would also be synthetically lethal in tumor cells with low CDK5RAP3 expression. Indeed, HR inhibitors targeting RAD51 and SSA inhibitors targeting RAD52 or XPF have recently been identified and are under development for clinical applications [46][47][48][49][50].
The functional connection between CDK5RAP3 and breast cancer may be extended to cancer etiology, as suggested by the results of the association study between common genetic variants and cancer development in BRCA1/2 mutation carriers. The biochemical interaction with BRCA2 and the functional link to the DNA damage response support this hypothesis. In addition, functional annotation of the strongest risk variation indicates potential alterations of regulatory mechanisms. However, further genetic analyses combined with assays measuring the impact of specific variants are warranted in future studies.
In summary, in addition to the link to cancer risk, our study identifies CDK5RAP3 expression in breast tumors as a new biomarker for poor prognosis and potential target for therapeutic intervention. Lack of CDK5RAP3 targets HR and SSA, two crucial tumorsuppressor DSB repair pathways, which can be added to its link with cell cycle arrest and p53 [17]. Interestingly, both BRCA2-deficient and p53-deficient mice have a carcinogenic phenotype independently. However, when both genes are simultaneously abrogated, tumor progression is extremely enhanced, and tumor cells acquire resistance to DNA-damaging drugs, a phenotype that resembles our findings related to CDK5RAP3 depletion [51][52][53][54]. CDK5RAP3 downregulation may therefore lead to the worst possible scenario for tumor progression because it confers resistance to DNA damage agents, reduces genome instability, and increases tumor aggressiveness.

Conclusions
CDK5RAP3 is a novel BRCA2-interacting protein that negatively regulates homologous recombination and single-strand annealing. CDK5RAP3 expression is linked to breast cancer survival and its genetic variations are associated with BRCA1/2 mutation carriers. CDK5RAP3 thus emerges as a potential prognostic biomarker for breast cancer.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/cancers14020353/s1, Figure S1: CDK5RAP3 regulation in non homologous end joining. Figure S2: Genomic instability by flow cytometric micronucleus (FCM) assay and CDK5RAP3 role in HR-related protein expression. Figure S3: Low CDK5RAP3 expression association with survival curves for different breast cancer subtypes. Figure S4: Functional annotation of the 17q21.32 region showing positions of candidate variants in relation to RefSeq annotated genes. Figure S5: Uncropped Western blots from Figure 2A Table S1: Results of the combined iCOGS and Oncoarray studies at the CDK5RAP3 locus (imputed and genotyped variants) in BRCA1 and BRCA2 mutation carriers. Table S2: Functional annotation of potentially associated variants at the CDK5RAP3 locus.

Data Availability Statement:
The data presented in this study are available in this article and supplementary materials.