Downregulation of the cancer susceptibility protein WRAP53β in epithelial ovarian cancer leads to defective DNA repair and poor clinical outcome

Alterations in the scaffold protein WRAP53β have previously been linked to carcinogenesis and, in particular, associated with an increased risk for epithelial ovarian cancer. Here, we investigated the pathogenic impact and prognostic significance of WRAP53β in connection with epithelial ovarian cancer and examined the underlying mechanisms. We find that reduced expression of WRAP53β in ovarian tumors correlated with attenuated DNA damage response and poor patient survival. Furthermore, in ovarian cancer cell lines, WRAP53β was rapidly recruited to DNA double-strand breaks, where it orchestrated the recruitment of repair factors involved in homologous recombination and non-homologous end joining, including RNF168, 53BP1, BRCA1 and RAD51. Mechanistically, WRAP53β accomplishes this by facilitating the necessary ubiquitinylation at DNA breaks. Finally, we demonstrate that loss of WRAP53β significantly impairs the repair of DNA double-strand breaks, resulting in their accumulation. Our findings establish WRAP53β as a regulator of homologous recombination and non-homologous end joining repair in ovarian cancer cells, suggesting that loss of this protein contributes to the development and/or progression of ovarian tumors. Moreover, our current observations identify the nuclear levels of WRAP53β as a promising biomarker for the survival of patients with ovarian cancer.

Accounting for 2% of all cancers in women, but ranking fifth among the causes of all cancer-related deaths in women, ovarian cancer is associated with the highest mortality rate among gynecological malignancies. 1 Its poor prognosis is primarily due to late diagnosis, since the symptoms do not usually appear until the disease has spread outside the ovaries. Most ovarian cancers are epithelial and treatment usually includes cytoreductive surgery (debulking) followed by chemotherapy (platinum-based drugs). Unfortunately, the majority of patients who respond to primary chemotherapy later experience relapse.
Alteration of the DNA damage response is one major factor in the onset and/or progression of ovarian cancer. For example, repair by homologous recombination (HR) is defective in approximately half of all ovarian tumors due to inactivation of genes encoding proteins involved in this pathway, such as BRCA1 and BRCA2. 2 Since HR is involved in repairing DNA lesions caused by platinum-based chemotherapeutics and poly ADP-ribose polymerase (PARP) inhibitors, HR-deficient tumors are hypersensitive to these drugs, which help prolong patient survival, 3,4 although perhaps for not more than 5 years. [5][6][7] When DNA double-strand breaks arise, the high-fidelity HR pathway and error-prone non-homologous end joining (NHEJ) pathway compete to repair them. Inactivation of HR enhances NHEJ repair and overstimulation of this error-prone pathway was recently shown to contribute to the hypersensitivity of HRdeficient ovarian tumors to PARP inhibitors. 8,9 Consistent with this observation, inactivation of NHEJ through inhibition of DNA-PK/Ku80 or deletion of 53BP1 abrogates the cytotoxicity and genomic instability induced by PARP inhibitors leading to drug resistance. 8,9 Thus, impairment of NHEJ in HR-deficient tumors could result in resistance to treatment and reduce patient survival.
Previously, we identified the gene WRAP53 (WD40-encoding RNA antisense to p53) and showed that its product (referred to as WRAP53α) regulates expression and function of the tumor suppressor p53. 10 WRAP53 also encodes a scaffold protein designated WRAP53β (alias WRAP53, WDR79 or TCAB1), which is present both in the cytoplasm and nucleus, where it is highly enriched in the nuclear organelles Cajal bodies. WRAP53β clearly plays a central role in the maintenance and localization of factors involved in splicing and telomere elongation to the Cajal body [11][12][13] and was also recently shown to control the repair of DNA doublestrand breaks by both the HR and NHEJ pathways through targeting the critical ubiquitin ligase RNF8 to these lesions. 14 Loss of WRAP53β function is associated with various disorders, including dyskeratosis congenita, which is caused by germline mutations in WRAP53β and characterized by bone marrow failure and predisposition for cancer 15 and spinal muscular atrophy, a neurodegenerative disorder that is the leading genetic reason of infant mortality worldwide. 13 Moreover, single nucleotide polymorphisms in WRAP53 or altered expression of the protein are correlated with an elevated risk of developing a variety of sporadic tumors, including ovarian, breast, head and neck cancers. [16][17][18][19][20] Nonetheless, the exact involvement of WRAP53β in carcinogenesis remains unclear.
In the current investigation, we find that attenuated expression of WRAP53β contributes to the progression of and is associated with altered DNA damage response in epithelial ovarian cancer. In this context, we demonstrate that WRAP53β participates in DNA repair in epithelial ovarian cancer cell lines by targeting factors involved in the HR and NHEJ pathways to such DNA lesions and that loss of this protein eliminates repair of DNA double-strand breaks. In summary, we establish a distinct role for this protein in the DNA damage response and repair in ovarian cancer cells and propose that WRAP53β thereby acts as a tumor suppressor in connection with epithelial ovarian cancer.

Results
Reduced levels of WRAP53β mRNA and protein correlate with shorter survival in ovarian cancer patients. Kaplan-Meier analysis of WRAP53β mRNA levels in epithelial ovarian cancer cohort I revealed that lower levels were associated with shortened progression-free and overall survival (Figures 1a and b). After confirming the specificity of the WRAP53β antibody by immunohistochemistry and western blotting of tumor cells either expressing or lacking this protein (Figures 1c and d) immunohistochemical analysis of tumor samples (cohort II) revealed tumor-specific and nuclear expression of WRAP53β of varying degrees ( Figure 1e). Kaplan-Meier analysis of the nuclear levels revealed that lower intensity of nuclear staining for WRAP53β was correlated with reduced survival of the patients with ovarian cancer (Figure 1f). For statistical comparison, the tumors were grouped into those exhibiting low (combined score 0-2) and high (score 3) nuclear staining for WRAP53β ( Figure 1g).
Examination of the relationship between nuclear expression of WRAP53β and clinical variables revealed significant correlations between the levels of expression and stage (P = 0.009) and differentiation grade (P = 0.049) of the tumors (Table 1), but not with age, histology or levels of p53 (data not shown). Multivariate analysis including nuclear expression of WRAP53β and stage, differentiation grade, histology and p53 expression of the tumors revealed that low WRAP53β expression was associated with a 4-fold higher risk of dying from ovarian cancer and demonstrated that WRAP53β is an independent marker of survival in patients with ovarian cancer (HR = 4.20, 95% CI = 1.00-17.61, P = 0.05; Table 2).
Together, these findings suggest that nuclear expression of WRAP53β correlates with the progression of epithelial ovarian cancer and might serve as a prognostic marker for this type of tumor.
WRAP53β and the DNA damage response show a positive correlation in ovarian tumors. To gain a deeper understanding of the underlying molecular processes associated with WRAP53β expression in epithelial ovarian cancer, we performed gene set enrichment analysis (GSEA) of two independent cohorts, both of which are included in the Kaplan-Meier analysis presented in Figures 1a and b. This showed in cohort III an association between high levels of WRAP53β and of a number of processes involved in the DNA damage response, including DNA repair, chromatin architecture, histone modification and histone binding (n = 241; Figure 2a). GSEA of cohort IV (n = 403) confirmed this association ( Figure 2b). Together, these observations suggest that WRAP53β plays an important role in the DNA damage response in epithelial ovarian cancer and that attenuation of this function may contribute to tumor formation, progression and therapeutic response.
WRAP53β regulates repair of DNA double-strand breaks in ovarian cancer cell lines. To further explore whether WRAP53β is involved in DNA repair of ovarian tumors, we studied the behavior of this protein following DNA damage in the ovarian cancer cell lines A2780 and SKOV-3. One hallmark of DNA repair proteins is their accumulation at the sites of damage, often forming discrete foci. Following exposure of the ovarian cancer cell lines A2780 and SKOV-3 to ionizing radiation (IR), WRAP53β was rapidly recruited to sites of DNA damage in these cells ( Figure 3a). Moreover, the IR-induced foci formed by WRAP53β clearly overlapped with the foci containing Ser139-phosphorylated histone H2AX (referred to as γH2AX), a marker of DNA damage. Furthermore, the WRAP53β staining was specific, since it was abolished by siRNA knockdown of WRAP53β ( Figure 3a). Thus, WRAP53β is recruited to sites of DNA damage in ovarian cancer cell lines. The possibility that the intracellular distribution of WRAP53β, which is located both in the nucleus and cytoplasm, 13 is altered by irradiation was examined using A2780 and SKOV-3 cells. The rabbit α-WRAP53-C2 antibody, which detects both the cytoplasmic and nuclear forms, was used, since the mouse α-WDR79 clone 1F12 antibody only recognizes nuclear WRAP53β. The lack of change in intracellular distribution upon irradiation ( Figure 3b) indicates that the WRAP53β protein recruited to the sites of DNA damage originates from the nuclear pool.
This recruitment of WRAP53β to DNA breaks indicates its direct involvement in DNA repair, and, indeed in control cells expressing WRAP53β, the majority of γH2AX foci rapidly formed in response to IR was resolved 24 h later reflecting efficient DNA repair (Figures 3c and d). In contrast, in cells depleted of WRAP53β recovery from DNA damage was severely delayed and a significant number of γH2AX foci remained even 24 h after IR. Western blotting confirmed that the level of γH2AX in these WRAP53β-deficient cells remained elevated 24 h post IR ( Figure 3e). Together, these findings  Knockdown of WRAP53β impairs recruitment of DNA repair factors to DNA breaks in ovarian cancer cell lines. HR inactivation contributes to ovarian tumorigenesis, and we observed reduced accumulation of key factors involved in HR and NHEJ, including BRCA1 (HR), RAD51 (HR) and 53BP1 (NHEJ) at DNA double-strand breaks in irradiated A2780 cells depleted of WRAP53β. At the same time, the upstream proteins γH2AX and MDC1 still formed foci (Figures 4a and b). Thus, loss of WRAP53β leads to defective accumulation of critical factors mediating HR and NHEJ to sites of DNA damage in ovarian cancer cells.
To explore the underlying mechanism, we monitored ubiquitinylation at the sites of DNA damage known to be important for the local accumulation of BRCA1, 53BP1 and RAD51 but not γH2AX and MDC1 at such site. 14,21-24 For detection of ubiquitinylation at the sites of DNA damage, we used the FK2 antibody, which binds to ubiquitin chains on mono-and polyubiquitinated proteins, but not free ubiquitin. Indeed, knockdown of WRAP53β reduced the accumulation of both the E3 ligase RNF168, which catalyzes this ubiquitinylation, and conjugated ubiquitin at DNA double-strand breaks (Figures 4a and b). This impaired accumulation at DNA damage sites was not due to altered levels of these factors (Figure 4c). Altogether, our results demonstrate that WRAP53β regulates the repair of DNA double-strand breaks in ovarian cancer cells and that loss of this protein leads to defects in both HR and NHEJ.

Discussion
Here, we demonstrate for the first time that low nuclear expression of the scaffolding protein WRAP53β correlates with aggressiveness and poor prognosis of epithelial ovarian cancer. A similar observation was recently reported, where loss of nuclear WRAP53β is associated with reduced survival and enhanced radioresistance in patients with head and neck cancer. 20 This correlation was observed only for WRAP53β in the nucleus and not in the cytoplasm, 20 emphasizing that the subcellular localization of this protein should be taken into account when predicting outcome. In addition, single nucleotide polymorphisms in the WRAP53 gene are associated with poor outcome in epithelial ovarian cancer 18,19 and inherited mutations in this same gene cause dyskeratosis congenita, which is associated with a dramatic elevation in risk for developing cancer. 15 Since inactivation of WRAP53β by mutations in both alleles is required for development of this disease, this protein appears to be a bona fide tumor suppressor.
Reductions in the levels of both WRAP53β mRNA and protein were associated with shorter survival in patients with epithelial ovarian cancer. This suggests that downregulation of WRAP53β in such tumors occurs at the transcriptional or posttranscriptional level rather than post-translationally, although this remains to be determined.
We have shown that ovarian tumors expressing low levels of WRAP53β exhibit downregulation of key factors involved in the DNA damage response, indicating impaired DNA repair. Indeed, in ovarian cancer cell lines WRAP53β rapidly accumulates at DNA breaks, where it orchestrates the accumulation of DNA repair proteins involved in both HR and NHEJ, including RNF168, 53BP1, BRCA1 and RAD51. WRAP53β achieves this recruitment by promoting ubiquitinylation at the sites of DNA damage, in agreement with our recent findings that WRAP53β serves as a scaffold for complex formation between the E3 ligase RNF8 and the anchoring protein MDC1. 14 Accordingly, knockdown of WRAP53β impairs the repair of DNA double-strand breaks by both HR and NHEJ resulting in their accumulation. These observations suggest that attenuated expression of this protein contributes to genomic instability and carcinogenesis.
Our findings indicate that the WRAP53β recruited to DNA lesions originates from the nuclear pool alone. This may explain why lower levels of nuclear, but not of cytoplasmic WRAP53β are associated with poor prognosis and   20 as well as with altered DNA repair and poor prognosis in patients with ovarian cancer. Precancerous lesions are characterized by activation of the DNA damage response (often due to replication stress), which is believed to eliminate hazardous cells. At an early stage in the development of cancer, this defense is lost by inactivation of factors involved in the DNA damage response, which contributes to progression to carcinoma. 25,26 At this early stage, the p53 protein is still active and promotes removal of dangerous cells through growth arrest or apoptosis. Subsequent inactivation of p53, often occurring at a later stage in tumor development, results in survival of damaged cells, which augment tumor progression and aggressiveness.
In line with this model, we find that the levels of WRAP53β are higher in ovarian cancer cells than nonmalignant tubal cells indicating activation of the DNA damage response in these cells. Our findings further demonstrate that subsequent downregulation of WRAP53β in ovarian cancer cells impairs their damage response and drives tumor progression. Moreover, patients whose tumors exhibited both a low level of nuclear WRAP53β and positive/high p53 expression, indicative of mutation, suffered a higher rate of mortality compared to those with both high-nuclear WRAP53β and no expression of p53 (HR = 4.71, 95% CI = 1.15-19.33, P = 0.032).
Although the mutational status of p53 needs to be verified by sequencing, these data indicate that inactivation of p53 in WRAP53β-deficient cells contributes further to tumor progression and aggressiveness.
Our own findings and those of others reveal that appropriate expression of p53 is dependent on WRAP53α, which also is encoded by the WRAP53 gene, and, moreover, that p53 activity in response to DNA damage is abrogated when WRAP53α is downregulated. 10,27 Several lines of evidence indicate that WRAP53α and WRAP53β act independently and that neither WRAP53β transcripts nor protein are involved in regulating p53. 10,28 However, in tumors containing reduced levels of WRAP53β transcripts, such as ovarian cancer, WRAP53α, which is transcribed from the same locus, might also be downregulated resulting in inactivation of p53.
Still, the involvement of WRAP53β in the repair of DNA double-strand breaks is independent of WRAP53α-mediated regulation of p53, since this also occurs in SKOV-3, H1299 and HeLa cells, which contain no or inactive p53 (Figures 3 and  4). 14 Nonetheless, it remains to be determined whether the parallel actions of WRAP53α and WRAP53β are required for a complete DNA damage response that protects against tumor development and/or progression.
We have also established that WRAP53β is an upstream regulator of BRCA1. Since these proteins act in the same   [29][30][31] The demonstration that functional NHEJ contributes to the cytotoxicity of such inhibitors suggests that HR-deficient cancers with diminished NHEJ will be relatively resistant. This line of reasoning might explain why loss of WRAP53β, which impairs both HR and NHEJ, shortens the survival of epithelial ovarian cancer patients. Further investigations on the contribution of WRAP53β to the response of ovarian cancer to treatment may reveal whether its downregulation leads to drug resistance, thereby helping to design individualized treatment.
In summary, our present findings indicate that nuclear levels of WRAP53β are a promising biomarker for prediction of the clinical outcome of epithelial ovarian cancer hopefully contributing to novel treatment strategies and improved survival. Moreover, our observations establish altered DNA repair as a cause of WRAP53β-associated ovarian cancer and suggest that defects in DNA repair may contribute to other forms of WRAP53β-related cancer as well. (c) A2780 cells were treated as in (a) and then subjected to western blotting for WRAP53β, γH2AX, MDC1, RNF168, 53BP1, RAD51 and β-actin. We could not assess the protein levels of BRCA1 due to a lack of antibodies that work for western blotting

Materials and Methods
Characterization of patients WRAP53β mRNA: By using microarray data on overall and progression-free survival for 1581 patients (cohort I), WRAP53β expression was assessed using the KM-plotter meta-analysis software (2015 version; http://kmplot.com 32 ) and the JetSet best probe (44563_at). Gene Expression Omnibus (GEO) IDs: GSE14764, GSE15622, GSE19829, GSE3149, GSE9891, GSE18520, GSE26712 and TCGA (The Cancer Genome Atlas). The median expression value was used as a threshold for survival analysis. Patients whose tumors exhibited WRAP53β mRNA levels above this threshold were classified as high expressers, and those with WRAP53β mRNA levels below this threshold was classified as low expressers. The study cohort involved here has been described in detail previously. [33][34][35] Statistical analysis. Kaplan-Meier analysis and the log-rank test were applied to relate overall and progression-free survival to WRAP53β expression. Pearson's chi-square test was used to explore potential associations between WRAP53β expression and clinicopathological parameters. Both uni-and multivariable Cox regression analysis were used to estimate hazard ratios for death from ovarian cancer in relationship to WRAP53β expression, with adjustment for the stage, differentiation grade, histology and p53 levels of the tumors. All calculations were performed using the SPSS version 19.0 software (SPSS Inc, Chicago, IL, USA) and P-values o0.05 were considered statistically significant.
Gene set enrichment analysis. Gene set enrichment was analyzed using GSEA software (http://www.broadinstitute.org/gsea/index.jsp) as described previously. 36 In these analyses, additional cohorts (III and IV) were used, both of which are included in cohort I. Cohort III originally consisted of 285 cases of epithelial ovarian cancer, fallopian tube and primary peritoneal cancers, as described previously. 37 In the present case, patients with potential tumors of low malignancy and those who received neoadjuvant chemotherapy were excluded, leaving a final total of 241 cases. Cohort IV consisted of 566 patients with high-grade serous ovarian cancer characterized in connection with TCGA project described previously. 2 The present analysis was restricted to 403 of these samples profiled on the Affymetrix U133A platform. Expression data were downloaded from the GEO website (http://www.ncbi. nlm.nih.gov/gds/), GEO (http://www.ncbi.nlm.nih.gov/geo) or the TCGA data portal (https://tcga-data.nci.nih.gov/tcga/tcga-Home2.jsp). The R package 'Affy' (http://www.bioconductor.org) was used to normalize the CEL files with the RMA procedure. 38 For WRAP53β, normalized gene expression values were extracted from each data set and used without further modification. From each cohort III and IV, the 50 tumors expressing the highest levels of WRAP53β mRNA and the 50 tumors not expressing or expressing the lowest levels of WRAP53β mRNA were selected for comparison by GSEA. Arrays were compared on the basis of the signal-to-noise ratio using the gene set C.5 (all) v 2.5.
Immunofluorescence microscopy. Cells grown on sterilized coverslips were fixed with 4% paraformaldehyde for 15 min at room temperature. They were then permeabilized with 0.1% Triton X-100 (Sigma-Aldrich) for 5 min at room temperature, followed by 30 min of blocking in blocking buffer (2% BSA, 5% glycerol, 0.2% Tween-20 and 0.1% NaN 3 ). The coverslips were subsequently incubated for 1 h in primary antibody followed by 40 min in secondary antibody, both diluted in blocking buffer, and finally mounted with VECTASHIELD mounting medium containing DAPI (4′,6-diamidino-2-phenylindole, Vector Laboratories, Bionordika, Stockholm, Sweden). Images were acquired with a Zeiss Axioplan 2 microscope (Zeiss, Stockholm, Sweden) equipped with an AxioCam HRm camera (Zeiss) using 40 or 63 oil immersion lenses and processed using AxioVision Release 4.7 or with a LSM700 confocal microscope (Zeiss), mounted on Zeiss Axio observer.Z1 equipped with Plan-Apochromat 63x/1.4 oil immersion lenses, and processed using Zen 2012 Black.
Preextraction: To visualize IR-induced foci formed by WRAP53β and MDC1, the cells were first washed with PBS and then incubated for 3 min at room temperature with cytoskeleton buffer (CSK; 10 mM pipes (pH 7.0), 100 mM NaCl, 300 mM sucrose, 3 mM MgCl 2 and 0.7% Triton X-100) and thereafter for another 3 min with the same CSK buffer supplemented with 0.3 mg/ml RNase A (CSK+R). Following these treatments, the cells were washed once again with PBS and then fixed in 4% paraformaldehyde.